diff options
Diffstat (limited to 'drivers')
71 files changed, 8917 insertions, 2196 deletions
diff --git a/drivers/Kconfig b/drivers/Kconfig index 283c089537b..bddf431bbb7 100644 --- a/drivers/Kconfig +++ b/drivers/Kconfig @@ -68,4 +68,6 @@ source "drivers/infiniband/Kconfig" source "drivers/sn/Kconfig" +source "drivers/edac/Kconfig" + endmenu diff --git a/drivers/Makefile b/drivers/Makefile index 7c45050ecd0..619dd964c51 100644 --- a/drivers/Makefile +++ b/drivers/Makefile @@ -63,6 +63,7 @@ obj-$(CONFIG_PHONE) += telephony/ obj-$(CONFIG_MD) += md/ obj-$(CONFIG_BT) += bluetooth/ obj-$(CONFIG_ISDN) += isdn/ +obj-$(CONFIG_EDAC) += edac/ obj-$(CONFIG_MCA) += mca/ obj-$(CONFIG_EISA) += eisa/ obj-$(CONFIG_CPU_FREQ) += cpufreq/ diff --git a/drivers/char/synclink_gt.c b/drivers/char/synclink_gt.c index 07c9be6a6bb..a85a60a93de 100644 --- a/drivers/char/synclink_gt.c +++ b/drivers/char/synclink_gt.c @@ -2630,7 +2630,7 @@ static int get_interface(struct slgt_info *info, int __user *if_mode) static int set_interface(struct slgt_info *info, int if_mode) { unsigned long flags; - unsigned char val; + unsigned short val; DBGINFO(("%s set_interface=%x)\n", info->device_name, if_mode)); spin_lock_irqsave(&info->lock,flags); diff --git a/drivers/char/tlclk.c b/drivers/char/tlclk.c index bc56df8a347..4c272189cd4 100644 --- a/drivers/char/tlclk.c +++ b/drivers/char/tlclk.c @@ -34,7 +34,6 @@ #include <linux/kernel.h> /* printk() */ #include <linux/fs.h> /* everything... */ #include <linux/errno.h> /* error codes */ -#include <linux/delay.h> /* udelay */ #include <linux/slab.h> #include <linux/ioport.h> #include <linux/interrupt.h> @@ -156,6 +155,8 @@ This directory exports the following interfaces. There operation is documented in the MCPBL0010 TPS under the Telecom Clock API section, 11.4. alarms : current_ref : +received_ref_clk3a : +received_ref_clk3b : enable_clk3a_output : enable_clk3b_output : enable_clka0_output : @@ -165,7 +166,7 @@ enable_clkb1_output : filter_select : hardware_switching : hardware_switching_mode : -interrupt_switch : +telclock_version : mode_select : refalign : reset : @@ -173,7 +174,6 @@ select_amcb1_transmit_clock : select_amcb2_transmit_clock : select_redundant_clock : select_ref_frequency : -test_mode : All sysfs interfaces are integers in hex format, i.e echo 99 > refalign has the same effect as echo 0x99 > refalign. @@ -226,7 +226,7 @@ static int tlclk_release(struct inode *inode, struct file *filp) return 0; } -ssize_t tlclk_read(struct file *filp, char __user *buf, size_t count, +static ssize_t tlclk_read(struct file *filp, char __user *buf, size_t count, loff_t *f_pos) { if (count < sizeof(struct tlclk_alarms)) @@ -242,7 +242,7 @@ ssize_t tlclk_read(struct file *filp, char __user *buf, size_t count, return sizeof(struct tlclk_alarms); } -ssize_t tlclk_write(struct file *filp, const char __user *buf, size_t count, +static ssize_t tlclk_write(struct file *filp, const char __user *buf, size_t count, loff_t *f_pos) { return 0; @@ -278,21 +278,21 @@ static ssize_t show_current_ref(struct device *d, static DEVICE_ATTR(current_ref, S_IRUGO, show_current_ref, NULL); -static ssize_t show_interrupt_switch(struct device *d, +static ssize_t show_telclock_version(struct device *d, struct device_attribute *attr, char *buf) { unsigned long ret_val; unsigned long flags; spin_lock_irqsave(&event_lock, flags); - ret_val = inb(TLCLK_REG6); + ret_val = inb(TLCLK_REG5); spin_unlock_irqrestore(&event_lock, flags); return sprintf(buf, "0x%lX\n", ret_val); } -static DEVICE_ATTR(interrupt_switch, S_IRUGO, - show_interrupt_switch, NULL); +static DEVICE_ATTR(telclock_version, S_IRUGO, + show_telclock_version, NULL); static ssize_t show_alarms(struct device *d, struct device_attribute *attr, char *buf) @@ -309,6 +309,50 @@ static ssize_t show_alarms(struct device *d, static DEVICE_ATTR(alarms, S_IRUGO, show_alarms, NULL); +static ssize_t store_received_ref_clk3a(struct device *d, + struct device_attribute *attr, const char *buf, size_t count) +{ + unsigned long tmp; + unsigned char val; + unsigned long flags; + + sscanf(buf, "%lX", &tmp); + dev_dbg(d, ": tmp = 0x%lX\n", tmp); + + val = (unsigned char)tmp; + spin_lock_irqsave(&event_lock, flags); + SET_PORT_BITS(TLCLK_REG1, 0xef, val); + spin_unlock_irqrestore(&event_lock, flags); + + return strnlen(buf, count); +} + +static DEVICE_ATTR(received_ref_clk3a, S_IWUGO, NULL, + store_received_ref_clk3a); + + +static ssize_t store_received_ref_clk3b(struct device *d, + struct device_attribute *attr, const char *buf, size_t count) +{ + unsigned long tmp; + unsigned char val; + unsigned long flags; + + sscanf(buf, "%lX", &tmp); + dev_dbg(d, ": tmp = 0x%lX\n", tmp); + + val = (unsigned char)tmp; + spin_lock_irqsave(&event_lock, flags); + SET_PORT_BITS(TLCLK_REG1, 0xef, val << 1); + spin_unlock_irqrestore(&event_lock, flags); + + return strnlen(buf, count); +} + +static DEVICE_ATTR(received_ref_clk3b, S_IWUGO, NULL, + store_received_ref_clk3b); + + static ssize_t store_enable_clk3b_output(struct device *d, struct device_attribute *attr, const char *buf, size_t count) { @@ -436,26 +480,6 @@ static ssize_t store_enable_clka0_output(struct device *d, static DEVICE_ATTR(enable_clka0_output, S_IWUGO, NULL, store_enable_clka0_output); -static ssize_t store_test_mode(struct device *d, - struct device_attribute *attr, const char *buf, size_t count) -{ - unsigned long flags; - unsigned long tmp; - unsigned char val; - - sscanf(buf, "%lX", &tmp); - dev_dbg(d, "tmp = 0x%lX\n", tmp); - - val = (unsigned char)tmp; - spin_lock_irqsave(&event_lock, flags); - SET_PORT_BITS(TLCLK_REG4, 0xfd, 2); - spin_unlock_irqrestore(&event_lock, flags); - - return strnlen(buf, count); -} - -static DEVICE_ATTR(test_mode, S_IWUGO, NULL, store_test_mode); - static ssize_t store_select_amcb2_transmit_clock(struct device *d, struct device_attribute *attr, const char *buf, size_t count) { @@ -475,7 +499,7 @@ static ssize_t store_select_amcb2_transmit_clock(struct device *d, SET_PORT_BITS(TLCLK_REG3, 0xc7, 0x38); switch (val) { case CLK_8_592MHz: - SET_PORT_BITS(TLCLK_REG0, 0xfc, 1); + SET_PORT_BITS(TLCLK_REG0, 0xfc, 2); break; case CLK_11_184MHz: SET_PORT_BITS(TLCLK_REG0, 0xfc, 0); @@ -484,7 +508,7 @@ static ssize_t store_select_amcb2_transmit_clock(struct device *d, SET_PORT_BITS(TLCLK_REG0, 0xfc, 3); break; case CLK_44_736MHz: - SET_PORT_BITS(TLCLK_REG0, 0xfc, 2); + SET_PORT_BITS(TLCLK_REG0, 0xfc, 1); break; } } else @@ -653,9 +677,7 @@ static ssize_t store_refalign (struct device *d, dev_dbg(d, "tmp = 0x%lX\n", tmp); spin_lock_irqsave(&event_lock, flags); SET_PORT_BITS(TLCLK_REG0, 0xf7, 0); - udelay(2); SET_PORT_BITS(TLCLK_REG0, 0xf7, 0x08); - udelay(2); SET_PORT_BITS(TLCLK_REG0, 0xf7, 0); spin_unlock_irqrestore(&event_lock, flags); @@ -706,15 +728,16 @@ static DEVICE_ATTR(reset, S_IWUGO, NULL, store_reset); static struct attribute *tlclk_sysfs_entries[] = { &dev_attr_current_ref.attr, - &dev_attr_interrupt_switch.attr, + &dev_attr_telclock_version.attr, &dev_attr_alarms.attr, + &dev_attr_received_ref_clk3a.attr, + &dev_attr_received_ref_clk3b.attr, &dev_attr_enable_clk3a_output.attr, &dev_attr_enable_clk3b_output.attr, &dev_attr_enable_clkb1_output.attr, &dev_attr_enable_clka1_output.attr, &dev_attr_enable_clkb0_output.attr, &dev_attr_enable_clka0_output.attr, - &dev_attr_test_mode.attr, &dev_attr_select_amcb1_transmit_clock.attr, &dev_attr_select_amcb2_transmit_clock.attr, &dev_attr_select_redundant_clock.attr, diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig new file mode 100644 index 00000000000..4819e7fc00d --- /dev/null +++ b/drivers/edac/Kconfig @@ -0,0 +1,102 @@ +# +# EDAC Kconfig +# Copyright (c) 2003 Linux Networx +# Licensed and distributed under the GPL +# +# $Id: Kconfig,v 1.4.2.7 2005/07/08 22:05:38 dsp_llnl Exp $ +# + +menu 'EDAC - error detection and reporting (RAS)' + +config EDAC + tristate "EDAC core system error reporting" + depends on X86 + default y + help + EDAC is designed to report errors in the core system. + These are low-level errors that are reported in the CPU or + supporting chipset: memory errors, cache errors, PCI errors, + thermal throttling, etc.. If unsure, select 'Y'. + + +comment "Reporting subsystems" + depends on EDAC + +config EDAC_DEBUG + bool "Debugging" + depends on EDAC + help + This turns on debugging information for the entire EDAC + sub-system. You can insert module with "debug_level=x", current + there're four debug levels (x=0,1,2,3 from low to high). + Usually you should select 'N'. + +config EDAC_MM_EDAC + tristate "Main Memory EDAC (Error Detection And Correction) reporting" + depends on EDAC + default y + help + Some systems are able to detect and correct errors in main + memory. EDAC can report statistics on memory error + detection and correction (EDAC - or commonly referred to ECC + errors). EDAC will also try to decode where these errors + occurred so that a particular failing memory module can be + replaced. If unsure, select 'Y'. + + +config EDAC_AMD76X + tristate "AMD 76x (760, 762, 768)" + depends on EDAC_MM_EDAC && PCI + help + Support for error detection and correction on the AMD 76x + series of chipsets used with the Athlon processor. + +config EDAC_E7XXX + tristate "Intel e7xxx (e7205, e7500, e7501, e7505)" + depends on EDAC_MM_EDAC && PCI + help + Support for error detection and correction on the Intel + E7205, E7500, E7501 and E7505 server chipsets. + +config EDAC_E752X + tristate "Intel e752x (e7520, e7525, e7320)" + depends on EDAC_MM_EDAC && PCI + help + Support for error detection and correction on the Intel + E7520, E7525, E7320 server chipsets. + +config EDAC_I82875P + tristate "Intel 82875p (D82875P, E7210)" + depends on EDAC_MM_EDAC && PCI + help + Support for error detection and correction on the Intel + DP82785P and E7210 server chipsets. + +config EDAC_I82860 + tristate "Intel 82860" + depends on EDAC_MM_EDAC && PCI + help + Support for error detection and correction on the Intel + 82860 chipset. + +config EDAC_R82600 + tristate "Radisys 82600 embedded chipset" + depends on EDAC_MM_EDAC + help + Support for error detection and correction on the Radisys + 82600 embedded chipset. + +choice + prompt "Error detecting method" + depends on EDAC + default EDAC_POLL + +config EDAC_POLL + bool "Poll for errors" + depends on EDAC + help + Poll the chipset periodically to detect errors. + +endchoice + +endmenu diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile new file mode 100644 index 00000000000..93137fdab4b --- /dev/null +++ b/drivers/edac/Makefile @@ -0,0 +1,18 @@ +# +# Makefile for the Linux kernel EDAC drivers. +# +# Copyright 02 Jul 2003, Linux Networx (http://lnxi.com) +# This file may be distributed under the terms of the +# GNU General Public License. +# +# $Id: Makefile,v 1.4.2.3 2005/07/08 22:05:38 dsp_llnl Exp $ + + +obj-$(CONFIG_EDAC_MM_EDAC) += edac_mc.o +obj-$(CONFIG_EDAC_AMD76X) += amd76x_edac.o +obj-$(CONFIG_EDAC_E7XXX) += e7xxx_edac.o +obj-$(CONFIG_EDAC_E752X) += e752x_edac.o +obj-$(CONFIG_EDAC_I82875P) += i82875p_edac.o +obj-$(CONFIG_EDAC_I82860) += i82860_edac.o +obj-$(CONFIG_EDAC_R82600) += r82600_edac.o + diff --git a/drivers/edac/amd76x_edac.c b/drivers/edac/amd76x_edac.c new file mode 100644 index 00000000000..2fcc8120b53 --- /dev/null +++ b/drivers/edac/amd76x_edac.c @@ -0,0 +1,356 @@ +/* + * AMD 76x Memory Controller kernel module + * (C) 2003 Linux Networx (http://lnxi.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * Written by Thayne Harbaugh + * Based on work by Dan Hollis <goemon at anime dot net> and others. + * http://www.anime.net/~goemon/linux-ecc/ + * + * $Id: edac_amd76x.c,v 1.4.2.5 2005/10/05 00:43:44 dsp_llnl Exp $ + * + */ + + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/init.h> + +#include <linux/pci.h> +#include <linux/pci_ids.h> + +#include <linux/slab.h> + +#include "edac_mc.h" + + +#define AMD76X_NR_CSROWS 8 +#define AMD76X_NR_CHANS 1 +#define AMD76X_NR_DIMMS 4 + + +/* AMD 76x register addresses - device 0 function 0 - PCI bridge */ +#define AMD76X_ECC_MODE_STATUS 0x48 /* Mode and status of ECC (32b) + * + * 31:16 reserved + * 15:14 SERR enabled: x1=ue 1x=ce + * 13 reserved + * 12 diag: disabled, enabled + * 11:10 mode: dis, EC, ECC, ECC+scrub + * 9:8 status: x1=ue 1x=ce + * 7:4 UE cs row + * 3:0 CE cs row + */ +#define AMD76X_DRAM_MODE_STATUS 0x58 /* DRAM Mode and status (32b) + * + * 31:26 clock disable 5 - 0 + * 25 SDRAM init + * 24 reserved + * 23 mode register service + * 22:21 suspend to RAM + * 20 burst refresh enable + * 19 refresh disable + * 18 reserved + * 17:16 cycles-per-refresh + * 15:8 reserved + * 7:0 x4 mode enable 7 - 0 + */ +#define AMD76X_MEM_BASE_ADDR 0xC0 /* Memory base address (8 x 32b) + * + * 31:23 chip-select base + * 22:16 reserved + * 15:7 chip-select mask + * 6:3 reserved + * 2:1 address mode + * 0 chip-select enable + */ + + +struct amd76x_error_info { + u32 ecc_mode_status; +}; + + +enum amd76x_chips { + AMD761 = 0, + AMD762 +}; + + +struct amd76x_dev_info { + const char *ctl_name; +}; + + +static const struct amd76x_dev_info amd76x_devs[] = { + [AMD761] = {.ctl_name = "AMD761"}, + [AMD762] = {.ctl_name = "AMD762"}, +}; + + +/** + * amd76x_get_error_info - fetch error information + * @mci: Memory controller + * @info: Info to fill in + * + * Fetch and store the AMD76x ECC status. Clear pending status + * on the chip so that further errors will be reported + */ + +static void amd76x_get_error_info (struct mem_ctl_info *mci, + struct amd76x_error_info *info) +{ + pci_read_config_dword(mci->pdev, AMD76X_ECC_MODE_STATUS, + &info->ecc_mode_status); + + if (info->ecc_mode_status & BIT(8)) + pci_write_bits32(mci->pdev, AMD76X_ECC_MODE_STATUS, + (u32) BIT(8), (u32) BIT(8)); + + if (info->ecc_mode_status & BIT(9)) + pci_write_bits32(mci->pdev, AMD76X_ECC_MODE_STATUS, + (u32) BIT(9), (u32) BIT(9)); +} + + +/** + * amd76x_process_error_info - Error check + * @mci: Memory controller + * @info: Previously fetched information from chip + * @handle_errors: 1 if we should do recovery + * + * Process the chip state and decide if an error has occurred. + * A return of 1 indicates an error. Also if handle_errors is true + * then attempt to handle and clean up after the error + */ + +static int amd76x_process_error_info (struct mem_ctl_info *mci, + struct amd76x_error_info *info, int handle_errors) +{ + int error_found; + u32 row; + + error_found = 0; + + /* + * Check for an uncorrectable error + */ + if (info->ecc_mode_status & BIT(8)) { + error_found = 1; + + if (handle_errors) { + row = (info->ecc_mode_status >> 4) & 0xf; + edac_mc_handle_ue(mci, + mci->csrows[row].first_page, 0, row, + mci->ctl_name); + } + } + + /* + * Check for a correctable error + */ + if (info->ecc_mode_status & BIT(9)) { + error_found = 1; + + if (handle_errors) { + row = info->ecc_mode_status & 0xf; + edac_mc_handle_ce(mci, + mci->csrows[row].first_page, 0, 0, row, 0, + mci->ctl_name); + } + } + return error_found; +} + +/** + * amd76x_check - Poll the controller + * @mci: Memory controller + * + * Called by the poll handlers this function reads the status + * from the controller and checks for errors. + */ + +static void amd76x_check(struct mem_ctl_info *mci) +{ + struct amd76x_error_info info; + debugf3("MC: " __FILE__ ": %s()\n", __func__); + amd76x_get_error_info(mci, &info); + amd76x_process_error_info(mci, &info, 1); +} + + +/** + * amd76x_probe1 - Perform set up for detected device + * @pdev; PCI device detected + * @dev_idx: Device type index + * + * We have found an AMD76x and now need to set up the memory + * controller status reporting. We configure and set up the + * memory controller reporting and claim the device. + */ + +static int amd76x_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + int index; + struct mem_ctl_info *mci = NULL; + enum edac_type ems_modes[] = { + EDAC_NONE, + EDAC_EC, + EDAC_SECDED, + EDAC_SECDED + }; + u32 ems; + u32 ems_mode; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + pci_read_config_dword(pdev, AMD76X_ECC_MODE_STATUS, &ems); + ems_mode = (ems >> 10) & 0x3; + + mci = edac_mc_alloc(0, AMD76X_NR_CSROWS, AMD76X_NR_CHANS); + + if (mci == NULL) { + rc = -ENOMEM; + goto fail; + } + + debugf0("MC: " __FILE__ ": %s(): mci = %p\n", __func__, mci); + + mci->pdev = pci_dev_get(pdev); + mci->mtype_cap = MEM_FLAG_RDDR; + + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_EC | EDAC_FLAG_SECDED; + mci->edac_cap = ems_mode ? + (EDAC_FLAG_EC | EDAC_FLAG_SECDED) : EDAC_FLAG_NONE; + + mci->mod_name = BS_MOD_STR; + mci->mod_ver = "$Revision: 1.4.2.5 $"; + mci->ctl_name = amd76x_devs[dev_idx].ctl_name; + mci->edac_check = amd76x_check; + mci->ctl_page_to_phys = NULL; + + for (index = 0; index < mci->nr_csrows; index++) { + struct csrow_info *csrow = &mci->csrows[index]; + u32 mba; + u32 mba_base; + u32 mba_mask; + u32 dms; + + /* find the DRAM Chip Select Base address and mask */ + pci_read_config_dword(mci->pdev, + AMD76X_MEM_BASE_ADDR + (index * 4), + &mba); + + if (!(mba & BIT(0))) + continue; + + mba_base = mba & 0xff800000UL; + mba_mask = ((mba & 0xff80) << 16) | 0x7fffffUL; + + pci_read_config_dword(mci->pdev, AMD76X_DRAM_MODE_STATUS, + &dms); + + csrow->first_page = mba_base >> PAGE_SHIFT; + csrow->nr_pages = (mba_mask + 1) >> PAGE_SHIFT; + csrow->last_page = csrow->first_page + csrow->nr_pages - 1; + csrow->page_mask = mba_mask >> PAGE_SHIFT; + csrow->grain = csrow->nr_pages << PAGE_SHIFT; + csrow->mtype = MEM_RDDR; + csrow->dtype = ((dms >> index) & 0x1) ? DEV_X4 : DEV_UNKNOWN; + csrow->edac_mode = ems_modes[ems_mode]; + } + + /* clear counters */ + pci_write_bits32(mci->pdev, AMD76X_ECC_MODE_STATUS, (u32) (0x3 << 8), + (u32) (0x3 << 8)); + + if (edac_mc_add_mc(mci)) { + debugf3("MC: " __FILE__ + ": %s(): failed edac_mc_add_mc()\n", __func__); + goto fail; + } + + /* get this far and it's successful */ + debugf3("MC: " __FILE__ ": %s(): success\n", __func__); + return 0; + +fail: + if (mci) { + if(mci->pdev) + pci_dev_put(mci->pdev); + edac_mc_free(mci); + } + return rc; +} + +/* returns count (>= 0), or negative on error */ +static int __devinit amd76x_init_one(struct pci_dev *pdev, + const struct pci_device_id *ent) +{ + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + /* don't need to call pci_device_enable() */ + return amd76x_probe1(pdev, ent->driver_data); +} + + +/** + * amd76x_remove_one - driver shutdown + * @pdev: PCI device being handed back + * + * Called when the driver is unloaded. Find the matching mci + * structure for the device then delete the mci and free the + * resources. + */ + +static void __devexit amd76x_remove_one(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + + debugf0(__FILE__ ": %s()\n", __func__); + + if ((mci = edac_mc_find_mci_by_pdev(pdev)) == NULL) + return; + if (edac_mc_del_mc(mci)) + return; + pci_dev_put(mci->pdev); + edac_mc_free(mci); +} + + +static const struct pci_device_id amd76x_pci_tbl[] __devinitdata = { + {PCI_VEND_DEV(AMD, FE_GATE_700C), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + AMD762}, + {PCI_VEND_DEV(AMD, FE_GATE_700E), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + AMD761}, + {0,} /* 0 terminated list. */ +}; + +MODULE_DEVICE_TABLE(pci, amd76x_pci_tbl); + + +static struct pci_driver amd76x_driver = { + .name = BS_MOD_STR, + .probe = amd76x_init_one, + .remove = __devexit_p(amd76x_remove_one), + .id_table = amd76x_pci_tbl, +}; + +static int __init amd76x_init(void) +{ + return pci_register_driver(&amd76x_driver); +} + +static void __exit amd76x_exit(void) +{ + pci_unregister_driver(&amd76x_driver); +} + +module_init(amd76x_init); +module_exit(amd76x_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Linux Networx (http://lnxi.com) Thayne Harbaugh"); +MODULE_DESCRIPTION("MC support for AMD 76x memory controllers"); diff --git a/drivers/edac/e752x_edac.c b/drivers/edac/e752x_edac.c new file mode 100644 index 00000000000..770a5a63307 --- /dev/null +++ b/drivers/edac/e752x_edac.c @@ -0,0 +1,1071 @@ +/* + * Intel e752x Memory Controller kernel module + * (C) 2004 Linux Networx (http://lnxi.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * See "enum e752x_chips" below for supported chipsets + * + * Written by Tom Zimmerman + * + * Contributors: + * Thayne Harbaugh at realmsys.com (?) + * Wang Zhenyu at intel.com + * Dave Jiang at mvista.com + * + * $Id: edac_e752x.c,v 1.5.2.11 2005/10/05 00:43:44 dsp_llnl Exp $ + * + */ + + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/init.h> + +#include <linux/pci.h> +#include <linux/pci_ids.h> + +#include <linux/slab.h> + +#include "edac_mc.h" + + +#ifndef PCI_DEVICE_ID_INTEL_7520_0 +#define PCI_DEVICE_ID_INTEL_7520_0 0x3590 +#endif /* PCI_DEVICE_ID_INTEL_7520_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7520_1_ERR +#define PCI_DEVICE_ID_INTEL_7520_1_ERR 0x3591 +#endif /* PCI_DEVICE_ID_INTEL_7520_1_ERR */ + +#ifndef PCI_DEVICE_ID_INTEL_7525_0 +#define PCI_DEVICE_ID_INTEL_7525_0 0x359E +#endif /* PCI_DEVICE_ID_INTEL_7525_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7525_1_ERR +#define PCI_DEVICE_ID_INTEL_7525_1_ERR 0x3593 +#endif /* PCI_DEVICE_ID_INTEL_7525_1_ERR */ + +#ifndef PCI_DEVICE_ID_INTEL_7320_0 +#define PCI_DEVICE_ID_INTEL_7320_0 0x3592 +#endif /* PCI_DEVICE_ID_INTEL_7320_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7320_1_ERR +#define PCI_DEVICE_ID_INTEL_7320_1_ERR 0x3593 +#endif /* PCI_DEVICE_ID_INTEL_7320_1_ERR */ + +#define E752X_NR_CSROWS 8 /* number of csrows */ + + +/* E752X register addresses - device 0 function 0 */ +#define E752X_DRB 0x60 /* DRAM row boundary register (8b) */ +#define E752X_DRA 0x70 /* DRAM row attribute register (8b) */ + /* + * 31:30 Device width row 7 + * 01=x8 10=x4 11=x8 DDR2 + * 27:26 Device width row 6 + * 23:22 Device width row 5 + * 19:20 Device width row 4 + * 15:14 Device width row 3 + * 11:10 Device width row 2 + * 7:6 Device width row 1 + * 3:2 Device width row 0 + */ +#define E752X_DRC 0x7C /* DRAM controller mode reg (32b) */ + /* FIXME:IS THIS RIGHT? */ + /* + * 22 Number channels 0=1,1=2 + * 19:18 DRB Granularity 32/64MB + */ +#define E752X_DRM 0x80 /* Dimm mapping register */ +#define E752X_DDRCSR 0x9A /* DDR control and status reg (16b) */ + /* + * 14:12 1 single A, 2 single B, 3 dual + */ +#define E752X_TOLM 0xC4 /* DRAM top of low memory reg (16b) */ +#define E752X_REMAPBASE 0xC6 /* DRAM remap base address reg (16b) */ +#define E752X_REMAPLIMIT 0xC8 /* DRAM remap limit address reg (16b) */ +#define E752X_REMAPOFFSET 0xCA /* DRAM remap limit offset reg (16b) */ + +/* E752X register addresses - device 0 function 1 */ +#define E752X_FERR_GLOBAL 0x40 /* Global first error register (32b) */ +#define E752X_NERR_GLOBAL 0x44 /* Global next error register (32b) */ +#define E752X_HI_FERR 0x50 /* Hub interface first error reg (8b) */ +#define E752X_HI_NERR 0x52 /* Hub interface next error reg (8b) */ +#define E752X_HI_ERRMASK 0x54 /* Hub interface error mask reg (8b) */ +#define E752X_HI_SMICMD 0x5A /* Hub interface SMI command reg (8b) */ +#define E752X_SYSBUS_FERR 0x60 /* System buss first error reg (16b) */ +#define E752X_SYSBUS_NERR 0x62 /* System buss next error reg (16b) */ +#define E752X_SYSBUS_ERRMASK 0x64 /* System buss error mask reg (16b) */ +#define E752X_SYSBUS_SMICMD 0x6A /* System buss SMI command reg (16b) */ +#define E752X_BUF_FERR 0x70 /* Memory buffer first error reg (8b) */ +#define E752X_BUF_NERR 0x72 /* Memory buffer next error reg (8b) */ +#define E752X_BUF_ERRMASK 0x74 /* Memory buffer error mask reg (8b) */ +#define E752X_BUF_SMICMD 0x7A /* Memory buffer SMI command reg (8b) */ +#define E752X_DRAM_FERR 0x80 /* DRAM first error register (16b) */ +#define E752X_DRAM_NERR 0x82 /* DRAM next error register (16b) */ +#define E752X_DRAM_ERRMASK 0x84 /* DRAM error mask register (8b) */ +#define E752X_DRAM_SMICMD 0x8A /* DRAM SMI command register (8b) */ +#define E752X_DRAM_RETR_ADD 0xAC /* DRAM Retry address register (32b) */ +#define E752X_DRAM_SEC1_ADD 0xA0 /* DRAM first correctable memory */ + /* error address register (32b) */ + /* + * 31 Reserved + * 30:2 CE address (64 byte block 34:6) + * 1 Reserved + * 0 HiLoCS + */ +#define E752X_DRAM_SEC2_ADD 0xC8 /* DRAM first correctable memory */ + /* error address register (32b) */ + /* + * 31 Reserved + * 30:2 CE address (64 byte block 34:6) + * 1 Reserved + * 0 HiLoCS + */ +#define E752X_DRAM_DED_ADD 0xA4 /* DRAM first uncorrectable memory */ + /* error address register (32b) */ + /* + * 31 Reserved + * 30:2 CE address (64 byte block 34:6) + * 1 Reserved + * 0 HiLoCS + */ +#define E752X_DRAM_SCRB_ADD 0xA8 /* DRAM first uncorrectable scrub memory */ + /* error address register (32b) */ + /* + * 31 Reserved + * 30:2 CE address (64 byte block 34:6) + * 1 Reserved + * 0 HiLoCS + */ +#define E752X_DRAM_SEC1_SYNDROME 0xC4 /* DRAM first correctable memory */ + /* error syndrome register (16b) */ +#define E752X_DRAM_SEC2_SYNDROME 0xC6 /* DRAM second correctable memory */ + /* error syndrome register (16b) */ +#define E752X_DEVPRES1 0xF4 /* Device Present 1 register (8b) */ + +/* ICH5R register addresses - device 30 function 0 */ +#define ICH5R_PCI_STAT 0x06 /* PCI status register (16b) */ +#define ICH5R_PCI_2ND_STAT 0x1E /* PCI status secondary reg (16b) */ +#define ICH5R_PCI_BRIDGE_CTL 0x3E /* PCI bridge control register (16b) */ + +enum e752x_chips { + E7520 = 0, + E7525 = 1, + E7320 = 2 +}; + + +struct e752x_pvt { + struct pci_dev *bridge_ck; + struct pci_dev *dev_d0f0; + struct pci_dev *dev_d0f1; + u32 tolm; + u32 remapbase; + u32 remaplimit; + int mc_symmetric; + u8 map[8]; + int map_type; + const struct e752x_dev_info *dev_info; +}; + + +struct e752x_dev_info { + u16 err_dev; + const char *ctl_name; +}; + +struct e752x_error_info { + u32 ferr_global; + u32 nerr_global; + u8 hi_ferr; + u8 hi_nerr; + u16 sysbus_ferr; + u16 sysbus_nerr; + u8 buf_ferr; + u8 buf_nerr; + u16 dram_ferr; + u16 dram_nerr; + u32 dram_sec1_add; + u32 dram_sec2_add; + u16 dram_sec1_syndrome; + u16 dram_sec2_syndrome; + u32 dram_ded_add; + u32 dram_scrb_add; + u32 dram_retr_add; +}; + +static const struct e752x_dev_info e752x_devs[] = { + [E7520] = { + .err_dev = PCI_DEVICE_ID_INTEL_7520_1_ERR, + .ctl_name = "E7520"}, + [E7525] = { + .err_dev = PCI_DEVICE_ID_INTEL_7525_1_ERR, + .ctl_name = "E7525"}, + [E7320] = { + .err_dev = PCI_DEVICE_ID_INTEL_7320_1_ERR, + .ctl_name = "E7320"}, +}; + + +static unsigned long ctl_page_to_phys(struct mem_ctl_info *mci, + unsigned long page) +{ + u32 remap; + struct e752x_pvt *pvt = (struct e752x_pvt *) mci->pvt_info; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + if (page < pvt->tolm) + return page; + if ((page >= 0x100000) && (page < pvt->remapbase)) + return page; + remap = (page - pvt->tolm) + pvt->remapbase; + if (remap < pvt->remaplimit) + return remap; + printk(KERN_ERR "Invalid page %lx - out of range\n", page); + return pvt->tolm - 1; +} + +static void do_process_ce(struct mem_ctl_info *mci, u16 error_one, + u32 sec1_add, u16 sec1_syndrome) +{ + u32 page; + int row; + int channel; + int i; + struct e752x_pvt *pvt = (struct e752x_pvt *) mci->pvt_info; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + /* convert the addr to 4k page */ + page = sec1_add >> (PAGE_SHIFT - 4); + + /* FIXME - check for -1 */ + if (pvt->mc_symmetric) { + /* chip select are bits 14 & 13 */ + row = ((page >> 1) & 3); + printk(KERN_WARNING + "Test row %d Table %d %d %d %d %d %d %d %d\n", + row, pvt->map[0], pvt->map[1], pvt->map[2], + pvt->map[3], pvt->map[4], pvt->map[5], + pvt->map[6], pvt->map[7]); + + /* test for channel remapping */ + for (i = 0; i < 8; i++) { + if (pvt->map[i] == row) + break; + } + printk(KERN_WARNING "Test computed row %d\n", i); + if (i < 8) + row = i; + else + printk(KERN_WARNING + "MC%d: row %d not found in remap table\n", + mci->mc_idx, row); + } else + row = edac_mc_find_csrow_by_page(mci, page); + /* 0 = channel A, 1 = channel B */ + channel = !(error_one & 1); + + if (!pvt->map_type) + row = 7 - row; + edac_mc_handle_ce(mci, page, 0, sec1_syndrome, row, channel, + "e752x CE"); +} + + +static inline void process_ce(struct mem_ctl_info *mci, u16 error_one, + u32 sec1_add, u16 sec1_syndrome, int *error_found, + int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_process_ce(mci, error_one, sec1_add, sec1_syndrome); +} + +static void do_process_ue(struct mem_ctl_info *mci, u16 error_one, u32 ded_add, + u32 scrb_add) +{ + u32 error_2b, block_page; + int row; + struct e752x_pvt *pvt = (struct e752x_pvt *) mci->pvt_info; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + if (error_one & 0x0202) { + error_2b = ded_add; + /* convert to 4k address */ + block_page = error_2b >> (PAGE_SHIFT - 4); + row = pvt->mc_symmetric ? + /* chip select are bits 14 & 13 */ + ((block_page >> 1) & 3) : + edac_mc_find_csrow_by_page(mci, block_page); + edac_mc_handle_ue(mci, block_page, 0, row, + "e752x UE from Read"); + } + if (error_one & 0x0404) { + error_2b = scrb_add; + /* convert to 4k address */ + block_page = error_2b >> (PAGE_SHIFT - 4); + row = pvt->mc_symmetric ? + /* chip select are bits 14 & 13 */ + ((block_page >> 1) & 3) : + edac_mc_find_csrow_by_page(mci, block_page); + edac_mc_handle_ue(mci, block_page, 0, row, + "e752x UE from Scruber"); + } +} + +static inline void process_ue(struct mem_ctl_info *mci, u16 error_one, + u32 ded_add, u32 scrb_add, int *error_found, int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_process_ue(mci, error_one, ded_add, scrb_add); +} + +static inline void process_ue_no_info_wr(struct mem_ctl_info *mci, + int *error_found, int handle_error) +{ + *error_found = 1; + + if (!handle_error) + return; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + edac_mc_handle_ue_no_info(mci, "e752x UE log memory write"); +} + +static void do_process_ded_retry(struct mem_ctl_info *mci, u16 error, + u32 retry_add) +{ + u32 error_1b, page; + int row; + struct e752x_pvt *pvt = (struct e752x_pvt *) mci->pvt_info; + + error_1b = retry_add; + page = error_1b >> (PAGE_SHIFT - 4); /* convert the addr to 4k page */ + row = pvt->mc_symmetric ? + ((page >> 1) & 3) : /* chip select are bits 14 & 13 */ + edac_mc_find_csrow_by_page(mci, page); + printk(KERN_WARNING + "MC%d: CE page 0x%lx, row %d : Memory read retry\n", + mci->mc_idx, (long unsigned int) page, row); +} + +static inline void process_ded_retry(struct mem_ctl_info *mci, u16 error, + u32 retry_add, int *error_found, int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_process_ded_retry(mci, error, retry_add); +} + +static inline void process_threshold_ce(struct mem_ctl_info *mci, u16 error, + int *error_found, int handle_error) +{ + *error_found = 1; + + if (handle_error) + printk(KERN_WARNING "MC%d: Memory threshold CE\n", + mci->mc_idx); +} + +static char *global_message[11] = { + "PCI Express C1", "PCI Express C", "PCI Express B1", + "PCI Express B", "PCI Express A1", "PCI Express A", + "DMA Controler", "HUB Interface", "System Bus", + "DRAM Controler", "Internal Buffer" +}; + +static char *fatal_message[2] = { "Non-Fatal ", "Fatal " }; + +static void do_global_error(int fatal, u32 errors) +{ + int i; + + for (i = 0; i < 11; i++) { + if (errors & (1 << i)) + printk(KERN_WARNING "%sError %s\n", + fatal_message[fatal], global_message[i]); + } +} + +static inline void global_error(int fatal, u32 errors, int *error_found, + int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_global_error(fatal, errors); +} + +static char *hub_message[7] = { + "HI Address or Command Parity", "HI Illegal Access", + "HI Internal Parity", "Out of Range Access", + "HI Data Parity", "Enhanced Config Access", + "Hub Interface Target Abort" +}; + +static void do_hub_error(int fatal, u8 errors) +{ + int i; + + for (i = 0; i < 7; i++) { + if (errors & (1 << i)) + printk(KERN_WARNING "%sError %s\n", + fatal_message[fatal], hub_message[i]); + } +} + +static inline void hub_error(int fatal, u8 errors, int *error_found, + int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_hub_error(fatal, errors); +} + +static char *membuf_message[4] = { + "Internal PMWB to DRAM parity", + "Internal PMWB to System Bus Parity", + "Internal System Bus or IO to PMWB Parity", + "Internal DRAM to PMWB Parity" +}; + +static void do_membuf_error(u8 errors) +{ + int i; + + for (i = 0; i < 4; i++) { + if (errors & (1 << i)) + printk(KERN_WARNING "Non-Fatal Error %s\n", + membuf_message[i]); + } +} + +static inline void membuf_error(u8 errors, int *error_found, int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_membuf_error(errors); +} + +#if 0 +char *sysbus_message[10] = { + "Addr or Request Parity", + "Data Strobe Glitch", + "Addr Strobe Glitch", + "Data Parity", + "Addr Above TOM", + "Non DRAM Lock Error", + "MCERR", "BINIT", + "Memory Parity", + "IO Subsystem Parity" +}; +#endif /* 0 */ + +static void do_sysbus_error(int fatal, u32 errors) +{ + int i; + + for (i = 0; i < 10; i++) { + if (errors & (1 << i)) + printk(KERN_WARNING "%sError System Bus %s\n", + fatal_message[fatal], global_message[i]); + } +} + +static inline void sysbus_error(int fatal, u32 errors, int *error_found, + int handle_error) +{ + *error_found = 1; + + if (handle_error) + do_sysbus_error(fatal, errors); +} + +static void e752x_check_hub_interface (struct e752x_error_info *info, + int *error_found, int handle_error) +{ + u8 stat8; + + //pci_read_config_byte(dev,E752X_HI_FERR,&stat8); + stat8 = info->hi_ferr; + if(stat8 & 0x7f) { /* Error, so process */ + stat8 &= 0x7f; + if(stat8 & 0x2b) + hub_error(1, stat8 & 0x2b, error_found, handle_error); + if(stat8 & 0x54) + hub_error(0, stat8 & 0x54, error_found, handle_error); + } + //pci_read_config_byte(dev,E752X_HI_NERR,&stat8); + stat8 = info->hi_nerr; + if(stat8 & 0x7f) { /* Error, so process */ + stat8 &= 0x7f; + if (stat8 & 0x2b) + hub_error(1, stat8 & 0x2b, error_found, handle_error); + if(stat8 & 0x54) + hub_error(0, stat8 & 0x54, error_found, handle_error); + } +} + +static void e752x_check_sysbus (struct e752x_error_info *info, int *error_found, + int handle_error) +{ + u32 stat32, error32; + + //pci_read_config_dword(dev,E752X_SYSBUS_FERR,&stat32); + stat32 = info->sysbus_ferr + (info->sysbus_nerr << 16); + + if (stat32 == 0) + return; /* no errors */ + + error32 = (stat32 >> 16) & 0x3ff; + stat32 = stat32 & 0x3ff; + if(stat32 & 0x083) + sysbus_error(1, stat32 & 0x083, error_found, handle_error); + if(stat32 & 0x37c) + sysbus_error(0, stat32 & 0x37c, error_found, handle_error); + if(error32 & 0x083) + sysbus_error(1, error32 & 0x083, error_found, handle_error); + if(error32 & 0x37c) + sysbus_error(0, error32 & 0x37c, error_found, handle_error); +} + +static void e752x_check_membuf (struct e752x_error_info *info, int *error_found, + int handle_error) +{ + u8 stat8; + + stat8 = info->buf_ferr; + if (stat8 & 0x0f) { /* Error, so process */ + stat8 &= 0x0f; + membuf_error(stat8, error_found, handle_error); + } + stat8 = info->buf_nerr; + if (stat8 & 0x0f) { /* Error, so process */ + stat8 &= 0x0f; + membuf_error(stat8, error_found, handle_error); + } +} + +static void e752x_check_dram (struct mem_ctl_info *mci, + struct e752x_error_info *info, int *error_found, int handle_error) +{ + u16 error_one, error_next; + + error_one = info->dram_ferr; + error_next = info->dram_nerr; + + /* decode and report errors */ + if(error_one & 0x0101) /* check first error correctable */ + process_ce(mci, error_one, info->dram_sec1_add, + info->dram_sec1_syndrome, error_found, + handle_error); + + if(error_next & 0x0101) /* check next error correctable */ + process_ce(mci, error_next, info->dram_sec2_add, + info->dram_sec2_syndrome, error_found, + handle_error); + + if(error_one & 0x4040) + process_ue_no_info_wr(mci, error_found, handle_error); + + if(error_next & 0x4040) + process_ue_no_info_wr(mci, error_found, handle_error); + + if(error_one & 0x2020) + process_ded_retry(mci, error_one, info->dram_retr_add, + error_found, handle_error); + + if(error_next & 0x2020) + process_ded_retry(mci, error_next, info->dram_retr_add, + error_found, handle_error); + + if(error_one & 0x0808) + process_threshold_ce(mci, error_one, error_found, + handle_error); + + if(error_next & 0x0808) + process_threshold_ce(mci, error_next, error_found, + handle_error); + + if(error_one & 0x0606) + process_ue(mci, error_one, info->dram_ded_add, + info->dram_scrb_add, error_found, handle_error); + + if(error_next & 0x0606) + process_ue(mci, error_next, info->dram_ded_add, + info->dram_scrb_add, error_found, handle_error); +} + +static void e752x_get_error_info (struct mem_ctl_info *mci, + struct e752x_error_info *info) +{ + struct pci_dev *dev; + struct e752x_pvt *pvt; + + memset(info, 0, sizeof(*info)); + pvt = (struct e752x_pvt *) mci->pvt_info; + dev = pvt->dev_d0f1; + + pci_read_config_dword(dev, E752X_FERR_GLOBAL, &info->ferr_global); + + if (info->ferr_global) { + pci_read_config_byte(dev, E752X_HI_FERR, &info->hi_ferr); + pci_read_config_word(dev, E752X_SYSBUS_FERR, + &info->sysbus_ferr); + pci_read_config_byte(dev, E752X_BUF_FERR, &info->buf_ferr); + pci_read_config_word(dev, E752X_DRAM_FERR, + &info->dram_ferr); + pci_read_config_dword(dev, E752X_DRAM_SEC1_ADD, + &info->dram_sec1_add); + pci_read_config_word(dev, E752X_DRAM_SEC1_SYNDROME, + &info->dram_sec1_syndrome); + pci_read_config_dword(dev, E752X_DRAM_DED_ADD, + &info->dram_ded_add); + pci_read_config_dword(dev, E752X_DRAM_SCRB_ADD, + &info->dram_scrb_add); + pci_read_config_dword(dev, E752X_DRAM_RETR_ADD, + &info->dram_retr_add); + + if (info->hi_ferr & 0x7f) + pci_write_config_byte(dev, E752X_HI_FERR, + info->hi_ferr); + + if (info->sysbus_ferr) + pci_write_config_word(dev, E752X_SYSBUS_FERR, + info->sysbus_ferr); + + if (info->buf_ferr & 0x0f) + pci_write_config_byte(dev, E752X_BUF_FERR, + info->buf_ferr); + + if (info->dram_ferr) + pci_write_bits16(pvt->bridge_ck, E752X_DRAM_FERR, + info->dram_ferr, info->dram_ferr); + + pci_write_config_dword(dev, E752X_FERR_GLOBAL, + info->ferr_global); + } + + pci_read_config_dword(dev, E752X_NERR_GLOBAL, &info->nerr_global); + + if (info->nerr_global) { + pci_read_config_byte(dev, E752X_HI_NERR, &info->hi_nerr); + pci_read_config_word(dev, E752X_SYSBUS_NERR, + &info->sysbus_nerr); + pci_read_config_byte(dev, E752X_BUF_NERR, &info->buf_nerr); + pci_read_config_word(dev, E752X_DRAM_NERR, + &info->dram_nerr); + pci_read_config_dword(dev, E752X_DRAM_SEC2_ADD, + &info->dram_sec2_add); + pci_read_config_word(dev, E752X_DRAM_SEC2_SYNDROME, + &info->dram_sec2_syndrome); + + if (info->hi_nerr & 0x7f) + pci_write_config_byte(dev, E752X_HI_NERR, + info->hi_nerr); + + if (info->sysbus_nerr) + pci_write_config_word(dev, E752X_SYSBUS_NERR, + info->sysbus_nerr); + + if (info->buf_nerr & 0x0f) + pci_write_config_byte(dev, E752X_BUF_NERR, + info->buf_nerr); + + if (info->dram_nerr) + pci_write_bits16(pvt->bridge_ck, E752X_DRAM_NERR, + info->dram_nerr, info->dram_nerr); + + pci_write_config_dword(dev, E752X_NERR_GLOBAL, + info->nerr_global); + } +} + +static int e752x_process_error_info (struct mem_ctl_info *mci, + struct e752x_error_info *info, int handle_errors) +{ + u32 error32, stat32; + int error_found; + + error_found = 0; + error32 = (info->ferr_global >> 18) & 0x3ff; + stat32 = (info->ferr_global >> 4) & 0x7ff; + + if (error32) + global_error(1, error32, &error_found, handle_errors); + + if (stat32) + global_error(0, stat32, &error_found, handle_errors); + + error32 = (info->nerr_global >> 18) & 0x3ff; + stat32 = (info->nerr_global >> 4) & 0x7ff; + + if (error32) + global_error(1, error32, &error_found, handle_errors); + + if (stat32) + global_error(0, stat32, &error_found, handle_errors); + + e752x_check_hub_interface(info, &error_found, handle_errors); + e752x_check_sysbus(info, &error_found, handle_errors); + e752x_check_membuf(info, &error_found, handle_errors); + e752x_check_dram(mci, info, &error_found, handle_errors); + return error_found; +} + +static void e752x_check(struct mem_ctl_info *mci) +{ + struct e752x_error_info info; + debugf3("MC: " __FILE__ ": %s()\n", __func__); + e752x_get_error_info(mci, &info); + e752x_process_error_info(mci, &info, 1); +} + +static int e752x_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + int index; + u16 pci_data, stat; + u32 stat32; + u16 stat16; + u8 stat8; + struct mem_ctl_info *mci = NULL; + struct e752x_pvt *pvt = NULL; + u16 ddrcsr; + u32 drc; + int drc_chan; /* Number of channels 0=1chan,1=2chan */ + int drc_drbg; /* DRB granularity 0=64mb,1=128mb */ + int drc_ddim; /* DRAM Data Integrity Mode 0=none,2=edac */ + u32 dra; + unsigned long last_cumul_size; + struct pci_dev *pres_dev; + struct pci_dev *dev = NULL; + + debugf0("MC: " __FILE__ ": %s(): mci\n", __func__); + debugf0("Starting Probe1\n"); + + /* enable device 0 function 1 */ + pci_read_config_byte(pdev, E752X_DEVPRES1, &stat8); + stat8 |= (1 << 5); + pci_write_config_byte(pdev, E752X_DEVPRES1, stat8); + + /* need to find out the number of channels */ + pci_read_config_dword(pdev, E752X_DRC, &drc); + pci_read_config_word(pdev, E752X_DDRCSR, &ddrcsr); + /* FIXME: should check >>12 or 0xf, true for all? */ + /* Dual channel = 1, Single channel = 0 */ + drc_chan = (((ddrcsr >> 12) & 3) == 3); + drc_drbg = drc_chan + 1; /* 128 in dual mode, 64 in single */ + drc_ddim = (drc >> 20) & 0x3; + + mci = edac_mc_alloc(sizeof(*pvt), E752X_NR_CSROWS, drc_chan + 1); + + if (mci == NULL) { + rc = -ENOMEM; + goto fail; + } + + debugf3("MC: " __FILE__ ": %s(): init mci\n", __func__); + + mci->mtype_cap = MEM_FLAG_RDDR; + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED | + EDAC_FLAG_S4ECD4ED; + /* FIXME - what if different memory types are in different csrows? */ + mci->mod_name = BS_MOD_STR; + mci->mod_ver = "$Revision: 1.5.2.11 $"; + mci->pdev = pdev; + + debugf3("MC: " __FILE__ ": %s(): init pvt\n", __func__); + pvt = (struct e752x_pvt *) mci->pvt_info; + pvt->dev_info = &e752x_devs[dev_idx]; + pvt->bridge_ck = pci_get_device(PCI_VENDOR_ID_INTEL, + pvt->dev_info->err_dev, + pvt->bridge_ck); + if (pvt->bridge_ck == NULL) + pvt->bridge_ck = pci_scan_single_device(pdev->bus, + PCI_DEVFN(0, 1)); + if (pvt->bridge_ck == NULL) { + printk(KERN_ERR "MC: error reporting device not found:" + "vendor %x device 0x%x (broken BIOS?)\n", + PCI_VENDOR_ID_INTEL, e752x_devs[dev_idx].err_dev); + goto fail; + } + pvt->mc_symmetric = ((ddrcsr & 0x10) != 0); + + debugf3("MC: " __FILE__ ": %s(): more mci init\n", __func__); + mci->ctl_name = pvt->dev_info->ctl_name; + mci->edac_check = e752x_check; + mci->ctl_page_to_phys = ctl_page_to_phys; + + /* find out the device types */ + pci_read_config_dword(pdev, E752X_DRA, &dra); + + /* + * The dram row boundary (DRB) reg values are boundary address for + * each DRAM row with a granularity of 64 or 128MB (single/dual + * channel operation). DRB regs are cumulative; therefore DRB7 will + * contain the total memory contained in all eight rows. + */ + for (last_cumul_size = index = 0; index < mci->nr_csrows; index++) { + u8 value; + u32 cumul_size; + /* mem_dev 0=x8, 1=x4 */ + int mem_dev = (dra >> (index * 4 + 2)) & 0x3; + struct csrow_info *csrow = &mci->csrows[index]; + + mem_dev = (mem_dev == 2); + pci_read_config_byte(mci->pdev, E752X_DRB + index, &value); + /* convert a 128 or 64 MiB DRB to a page size. */ + cumul_size = value << (25 + drc_drbg - PAGE_SHIFT); + debugf3("MC: " __FILE__ ": %s(): (%d) cumul_size 0x%x\n", + __func__, index, cumul_size); + if (cumul_size == last_cumul_size) + continue; /* not populated */ + + csrow->first_page = last_cumul_size; + csrow->last_page = cumul_size - 1; + csrow->nr_pages = cumul_size - last_cumul_size; + last_cumul_size = cumul_size; + csrow->grain = 1 << 12; /* 4KiB - resolution of CELOG */ + csrow->mtype = MEM_RDDR; /* only one type supported */ + csrow->dtype = mem_dev ? DEV_X4 : DEV_X8; + + /* + * if single channel or x8 devices then SECDED + * if dual channel and x4 then S4ECD4ED + */ + if (drc_ddim) { + if (drc_chan && mem_dev) { + csrow->edac_mode = EDAC_S4ECD4ED; + mci->edac_cap |= EDAC_FLAG_S4ECD4ED; + } else { + csrow->edac_mode = EDAC_SECDED; + mci->edac_cap |= EDAC_FLAG_SECDED; + } + } else + csrow->edac_mode = EDAC_NONE; + } + + /* Fill in the memory map table */ + { + u8 value; + u8 last = 0; + u8 row = 0; + for (index = 0; index < 8; index += 2) { + + pci_read_config_byte(mci->pdev, E752X_DRB + index, + &value); + /* test if there is a dimm in this slot */ + if (value == last) { + /* no dimm in the slot, so flag it as empty */ + pvt->map[index] = 0xff; + pvt->map[index + 1] = 0xff; + } else { /* there is a dimm in the slot */ + pvt->map[index] = row; + row++; + last = value; + /* test the next value to see if the dimm is + double sided */ + pci_read_config_byte(mci->pdev, + E752X_DRB + index + 1, + &value); + pvt->map[index + 1] = (value == last) ? + 0xff : /* the dimm is single sided, + so flag as empty */ + row; /* this is a double sided dimm + to save the next row # */ + row++; + last = value; + } + } + } + + /* set the map type. 1 = normal, 0 = reversed */ + pci_read_config_byte(mci->pdev, E752X_DRM, &stat8); + pvt->map_type = ((stat8 & 0x0f) > ((stat8 >> 4) & 0x0f)); + + mci->edac_cap |= EDAC_FLAG_NONE; + + debugf3("MC: " __FILE__ ": %s(): tolm, remapbase, remaplimit\n", + __func__); + /* load the top of low memory, remap base, and remap limit vars */ + pci_read_config_word(mci->pdev, E752X_TOLM, &pci_data); + pvt->tolm = ((u32) pci_data) << 4; + pci_read_config_word(mci->pdev, E752X_REMAPBASE, &pci_data); + pvt->remapbase = ((u32) pci_data) << 14; + pci_read_config_word(mci->pdev, E752X_REMAPLIMIT, &pci_data); + pvt->remaplimit = ((u32) pci_data) << 14; + printk("tolm = %x, remapbase = %x, remaplimit = %x\n", pvt->tolm, + pvt->remapbase, pvt->remaplimit); + + if (edac_mc_add_mc(mci)) { + debugf3("MC: " __FILE__ + ": %s(): failed edac_mc_add_mc()\n", + __func__); + goto fail; + } + + /* Walk through the PCI table and clear errors */ + switch (dev_idx) { + case E7520: + dev = pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_7520_0, NULL); + break; + case E7525: + dev = pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_7525_0, NULL); + break; + case E7320: + dev = pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_7320_0, NULL); + break; + } + + + pvt->dev_d0f0 = dev; + for (pres_dev = dev; + ((struct pci_dev *) pres_dev->global_list.next != dev); + pres_dev = (struct pci_dev *) pres_dev->global_list.next) { + pci_read_config_dword(pres_dev, PCI_COMMAND, &stat32); + stat = (u16) (stat32 >> 16); + /* clear any error bits */ + if (stat32 & ((1 << 6) + (1 << 8))) + pci_write_config_word(pres_dev, PCI_STATUS, stat); + } + /* find the error reporting device and clear errors */ + dev = pvt->dev_d0f1 = pci_dev_get(pvt->bridge_ck); + /* Turn off error disable & SMI in case the BIOS turned it on */ + pci_write_config_byte(dev, E752X_HI_ERRMASK, 0x00); + pci_write_config_byte(dev, E752X_HI_SMICMD, 0x00); + pci_write_config_word(dev, E752X_SYSBUS_ERRMASK, 0x00); + pci_write_config_word(dev, E752X_SYSBUS_SMICMD, 0x00); + pci_write_config_byte(dev, E752X_BUF_ERRMASK, 0x00); + pci_write_config_byte(dev, E752X_BUF_SMICMD, 0x00); + pci_write_config_byte(dev, E752X_DRAM_ERRMASK, 0x00); + pci_write_config_byte(dev, E752X_DRAM_SMICMD, 0x00); + /* clear other MCH errors */ + pci_read_config_dword(dev, E752X_FERR_GLOBAL, &stat32); + pci_write_config_dword(dev, E752X_FERR_GLOBAL, stat32); + pci_read_config_dword(dev, E752X_NERR_GLOBAL, &stat32); + pci_write_config_dword(dev, E752X_NERR_GLOBAL, stat32); + pci_read_config_byte(dev, E752X_HI_FERR, &stat8); + pci_write_config_byte(dev, E752X_HI_FERR, stat8); + pci_read_config_byte(dev, E752X_HI_NERR, &stat8); + pci_write_config_byte(dev, E752X_HI_NERR, stat8); + pci_read_config_dword(dev, E752X_SYSBUS_FERR, &stat32); + pci_write_config_dword(dev, E752X_SYSBUS_FERR, stat32); + pci_read_config_byte(dev, E752X_BUF_FERR, &stat8); + pci_write_config_byte(dev, E752X_BUF_FERR, stat8); + pci_read_config_byte(dev, E752X_BUF_NERR, &stat8); + pci_write_config_byte(dev, E752X_BUF_NERR, stat8); + pci_read_config_word(dev, E752X_DRAM_FERR, &stat16); + pci_write_config_word(dev, E752X_DRAM_FERR, stat16); + pci_read_config_word(dev, E752X_DRAM_NERR, &stat16); + pci_write_config_word(dev, E752X_DRAM_NERR, stat16); + + /* get this far and it's successful */ + debugf3("MC: " __FILE__ ": %s(): success\n", __func__); + return 0; + +fail: + if (mci) { + if (pvt->dev_d0f0) + pci_dev_put(pvt->dev_d0f0); + if (pvt->dev_d0f1) + pci_dev_put(pvt->dev_d0f1); + if (pvt->bridge_ck) + pci_dev_put(pvt->bridge_ck); + edac_mc_free(mci); + } + return rc; +} + +/* returns count (>= 0), or negative on error */ +static int __devinit e752x_init_one(struct pci_dev *pdev, + const struct pci_device_id *ent) +{ + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + /* wake up and enable device */ + if(pci_enable_device(pdev) < 0) + return -EIO; + return e752x_probe1(pdev, ent->driver_data); +} + + +static void __devexit e752x_remove_one(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + struct e752x_pvt *pvt; + + debugf0(__FILE__ ": %s()\n", __func__); + + if ((mci = edac_mc_find_mci_by_pdev(pdev)) == NULL) + return; + + if (edac_mc_del_mc(mci)) + return; + + pvt = (struct e752x_pvt *) mci->pvt_info; + pci_dev_put(pvt->dev_d0f0); + pci_dev_put(pvt->dev_d0f1); + pci_dev_put(pvt->bridge_ck); + edac_mc_free(mci); +} + + +static const struct pci_device_id e752x_pci_tbl[] __devinitdata = { + {PCI_VEND_DEV(INTEL, 7520_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7520}, + {PCI_VEND_DEV(INTEL, 7525_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7525}, + {PCI_VEND_DEV(INTEL, 7320_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7320}, + {0,} /* 0 terminated list. */ +}; + +MODULE_DEVICE_TABLE(pci, e752x_pci_tbl); + + +static struct pci_driver e752x_driver = { + name: BS_MOD_STR, + probe: e752x_init_one, + remove: __devexit_p(e752x_remove_one), + id_table: e752x_pci_tbl, +}; + + +static int __init e752x_init(void) +{ + int pci_rc; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + pci_rc = pci_register_driver(&e752x_driver); + return (pci_rc < 0) ? pci_rc : 0; +} + + +static void __exit e752x_exit(void) +{ + debugf3("MC: " __FILE__ ": %s()\n", __func__); + pci_unregister_driver(&e752x_driver); +} + + +module_init(e752x_init); +module_exit(e752x_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Linux Networx (http://lnxi.com) Tom Zimmerman\n"); +MODULE_DESCRIPTION("MC support for Intel e752x memory controllers"); diff --git a/drivers/edac/e7xxx_edac.c b/drivers/edac/e7xxx_edac.c new file mode 100644 index 00000000000..d5e320dfc66 --- /dev/null +++ b/drivers/edac/e7xxx_edac.c @@ -0,0 +1,558 @@ +/* + * Intel e7xxx Memory Controller kernel module + * (C) 2003 Linux Networx (http://lnxi.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * See "enum e7xxx_chips" below for supported chipsets + * + * Written by Thayne Harbaugh + * Based on work by Dan Hollis <goemon at anime dot net> and others. + * http://www.anime.net/~goemon/linux-ecc/ + * + * Contributors: + * Eric Biederman (Linux Networx) + * Tom Zimmerman (Linux Networx) + * Jim Garlick (Lawrence Livermore National Labs) + * Dave Peterson (Lawrence Livermore National Labs) + * That One Guy (Some other place) + * Wang Zhenyu (intel.com) + * + * $Id: edac_e7xxx.c,v 1.5.2.9 2005/10/05 00:43:44 dsp_llnl Exp $ + * + */ + + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/pci.h> +#include <linux/pci_ids.h> +#include <linux/slab.h> +#include "edac_mc.h" + + +#ifndef PCI_DEVICE_ID_INTEL_7205_0 +#define PCI_DEVICE_ID_INTEL_7205_0 0x255d +#endif /* PCI_DEVICE_ID_INTEL_7205_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7205_1_ERR +#define PCI_DEVICE_ID_INTEL_7205_1_ERR 0x2551 +#endif /* PCI_DEVICE_ID_INTEL_7205_1_ERR */ + +#ifndef PCI_DEVICE_ID_INTEL_7500_0 +#define PCI_DEVICE_ID_INTEL_7500_0 0x2540 +#endif /* PCI_DEVICE_ID_INTEL_7500_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7500_1_ERR +#define PCI_DEVICE_ID_INTEL_7500_1_ERR 0x2541 +#endif /* PCI_DEVICE_ID_INTEL_7500_1_ERR */ + +#ifndef PCI_DEVICE_ID_INTEL_7501_0 +#define PCI_DEVICE_ID_INTEL_7501_0 0x254c +#endif /* PCI_DEVICE_ID_INTEL_7501_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7501_1_ERR +#define PCI_DEVICE_ID_INTEL_7501_1_ERR 0x2541 +#endif /* PCI_DEVICE_ID_INTEL_7501_1_ERR */ + +#ifndef PCI_DEVICE_ID_INTEL_7505_0 +#define PCI_DEVICE_ID_INTEL_7505_0 0x2550 +#endif /* PCI_DEVICE_ID_INTEL_7505_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_7505_1_ERR +#define PCI_DEVICE_ID_INTEL_7505_1_ERR 0x2551 +#endif /* PCI_DEVICE_ID_INTEL_7505_1_ERR */ + + +#define E7XXX_NR_CSROWS 8 /* number of csrows */ +#define E7XXX_NR_DIMMS 8 /* FIXME - is this correct? */ + + +/* E7XXX register addresses - device 0 function 0 */ +#define E7XXX_DRB 0x60 /* DRAM row boundary register (8b) */ +#define E7XXX_DRA 0x70 /* DRAM row attribute register (8b) */ + /* + * 31 Device width row 7 0=x8 1=x4 + * 27 Device width row 6 + * 23 Device width row 5 + * 19 Device width row 4 + * 15 Device width row 3 + * 11 Device width row 2 + * 7 Device width row 1 + * 3 Device width row 0 + */ +#define E7XXX_DRC 0x7C /* DRAM controller mode reg (32b) */ + /* + * 22 Number channels 0=1,1=2 + * 19:18 DRB Granularity 32/64MB + */ +#define E7XXX_TOLM 0xC4 /* DRAM top of low memory reg (16b) */ +#define E7XXX_REMAPBASE 0xC6 /* DRAM remap base address reg (16b) */ +#define E7XXX_REMAPLIMIT 0xC8 /* DRAM remap limit address reg (16b) */ + +/* E7XXX register addresses - device 0 function 1 */ +#define E7XXX_DRAM_FERR 0x80 /* DRAM first error register (8b) */ +#define E7XXX_DRAM_NERR 0x82 /* DRAM next error register (8b) */ +#define E7XXX_DRAM_CELOG_ADD 0xA0 /* DRAM first correctable memory */ + /* error address register (32b) */ + /* + * 31:28 Reserved + * 27:6 CE address (4k block 33:12) + * 5:0 Reserved + */ +#define E7XXX_DRAM_UELOG_ADD 0xB0 /* DRAM first uncorrectable memory */ + /* error address register (32b) */ + /* + * 31:28 Reserved + * 27:6 CE address (4k block 33:12) + * 5:0 Reserved + */ +#define E7XXX_DRAM_CELOG_SYNDROME 0xD0 /* DRAM first correctable memory */ + /* error syndrome register (16b) */ + +enum e7xxx_chips { + E7500 = 0, + E7501, + E7505, + E7205, +}; + + +struct e7xxx_pvt { + struct pci_dev *bridge_ck; + u32 tolm; + u32 remapbase; + u32 remaplimit; + const struct e7xxx_dev_info *dev_info; +}; + + +struct e7xxx_dev_info { + u16 err_dev; + const char *ctl_name; +}; + + +struct e7xxx_error_info { + u8 dram_ferr; + u8 dram_nerr; + u32 dram_celog_add; + u16 dram_celog_syndrome; + u32 dram_uelog_add; +}; + +static const struct e7xxx_dev_info e7xxx_devs[] = { + [E7500] = { + .err_dev = PCI_DEVICE_ID_INTEL_7500_1_ERR, + .ctl_name = "E7500"}, + [E7501] = { + .err_dev = PCI_DEVICE_ID_INTEL_7501_1_ERR, + .ctl_name = "E7501"}, + [E7505] = { + .err_dev = PCI_DEVICE_ID_INTEL_7505_1_ERR, + .ctl_name = "E7505"}, + [E7205] = { + .err_dev = PCI_DEVICE_ID_INTEL_7205_1_ERR, + .ctl_name = "E7205"}, +}; + + +/* FIXME - is this valid for both SECDED and S4ECD4ED? */ +static inline int e7xxx_find_channel(u16 syndrome) +{ + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + if ((syndrome & 0xff00) == 0) + return 0; + if ((syndrome & 0x00ff) == 0) + return 1; + if ((syndrome & 0xf000) == 0 || (syndrome & 0x0f00) == 0) + return 0; + return 1; +} + + +static unsigned long +ctl_page_to_phys(struct mem_ctl_info *mci, unsigned long page) +{ + u32 remap; + struct e7xxx_pvt *pvt = (struct e7xxx_pvt *) mci->pvt_info; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + if ((page < pvt->tolm) || + ((page >= 0x100000) && (page < pvt->remapbase))) + return page; + remap = (page - pvt->tolm) + pvt->remapbase; + if (remap < pvt->remaplimit) + return remap; + printk(KERN_ERR "Invalid page %lx - out of range\n", page); + return pvt->tolm - 1; +} + + +static void process_ce(struct mem_ctl_info *mci, struct e7xxx_error_info *info) +{ + u32 error_1b, page; + u16 syndrome; + int row; + int channel; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + /* read the error address */ + error_1b = info->dram_celog_add; + /* FIXME - should use PAGE_SHIFT */ + page = error_1b >> 6; /* convert the address to 4k page */ + /* read the syndrome */ + syndrome = info->dram_celog_syndrome; + /* FIXME - check for -1 */ + row = edac_mc_find_csrow_by_page(mci, page); + /* convert syndrome to channel */ + channel = e7xxx_find_channel(syndrome); + edac_mc_handle_ce(mci, page, 0, syndrome, row, channel, + "e7xxx CE"); +} + + +static void process_ce_no_info(struct mem_ctl_info *mci) +{ + debugf3("MC: " __FILE__ ": %s()\n", __func__); + edac_mc_handle_ce_no_info(mci, "e7xxx CE log register overflow"); +} + + +static void process_ue(struct mem_ctl_info *mci, struct e7xxx_error_info *info) +{ + u32 error_2b, block_page; + int row; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + /* read the error address */ + error_2b = info->dram_uelog_add; + /* FIXME - should use PAGE_SHIFT */ + block_page = error_2b >> 6; /* convert to 4k address */ + row = edac_mc_find_csrow_by_page(mci, block_page); + edac_mc_handle_ue(mci, block_page, 0, row, "e7xxx UE"); +} + + +static void process_ue_no_info(struct mem_ctl_info *mci) +{ + debugf3("MC: " __FILE__ ": %s()\n", __func__); + edac_mc_handle_ue_no_info(mci, "e7xxx UE log register overflow"); +} + + +static void e7xxx_get_error_info (struct mem_ctl_info *mci, + struct e7xxx_error_info *info) +{ + struct e7xxx_pvt *pvt; + + pvt = (struct e7xxx_pvt *) mci->pvt_info; + pci_read_config_byte(pvt->bridge_ck, E7XXX_DRAM_FERR, + &info->dram_ferr); + pci_read_config_byte(pvt->bridge_ck, E7XXX_DRAM_NERR, + &info->dram_nerr); + + if ((info->dram_ferr & 1) || (info->dram_nerr & 1)) { + pci_read_config_dword(pvt->bridge_ck, E7XXX_DRAM_CELOG_ADD, + &info->dram_celog_add); + pci_read_config_word(pvt->bridge_ck, + E7XXX_DRAM_CELOG_SYNDROME, &info->dram_celog_syndrome); + } + + if ((info->dram_ferr & 2) || (info->dram_nerr & 2)) + pci_read_config_dword(pvt->bridge_ck, E7XXX_DRAM_UELOG_ADD, + &info->dram_uelog_add); + + if (info->dram_ferr & 3) + pci_write_bits8(pvt->bridge_ck, E7XXX_DRAM_FERR, 0x03, + 0x03); + + if (info->dram_nerr & 3) + pci_write_bits8(pvt->bridge_ck, E7XXX_DRAM_NERR, 0x03, + 0x03); +} + + +static int e7xxx_process_error_info (struct mem_ctl_info *mci, + struct e7xxx_error_info *info, int handle_errors) +{ + int error_found; + + error_found = 0; + + /* decode and report errors */ + if (info->dram_ferr & 1) { /* check first error correctable */ + error_found = 1; + + if (handle_errors) + process_ce(mci, info); + } + + if (info->dram_ferr & 2) { /* check first error uncorrectable */ + error_found = 1; + + if (handle_errors) + process_ue(mci, info); + } + + if (info->dram_nerr & 1) { /* check next error correctable */ + error_found = 1; + + if (handle_errors) { + if (info->dram_ferr & 1) + process_ce_no_info(mci); + else + process_ce(mci, info); + } + } + + if (info->dram_nerr & 2) { /* check next error uncorrectable */ + error_found = 1; + + if (handle_errors) { + if (info->dram_ferr & 2) + process_ue_no_info(mci); + else + process_ue(mci, info); + } + } + + return error_found; +} + + +static void e7xxx_check(struct mem_ctl_info *mci) +{ + struct e7xxx_error_info info; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + e7xxx_get_error_info(mci, &info); + e7xxx_process_error_info(mci, &info, 1); +} + + +static int e7xxx_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + int index; + u16 pci_data; + struct mem_ctl_info *mci = NULL; + struct e7xxx_pvt *pvt = NULL; + u32 drc; + int drc_chan = 1; /* Number of channels 0=1chan,1=2chan */ + int drc_drbg = 1; /* DRB granularity 0=32mb,1=64mb */ + int drc_ddim; /* DRAM Data Integrity Mode 0=none,2=edac */ + u32 dra; + unsigned long last_cumul_size; + + + debugf0("MC: " __FILE__ ": %s(): mci\n", __func__); + + /* need to find out the number of channels */ + pci_read_config_dword(pdev, E7XXX_DRC, &drc); + /* only e7501 can be single channel */ + if (dev_idx == E7501) { + drc_chan = ((drc >> 22) & 0x1); + drc_drbg = (drc >> 18) & 0x3; + } + drc_ddim = (drc >> 20) & 0x3; + + mci = edac_mc_alloc(sizeof(*pvt), E7XXX_NR_CSROWS, drc_chan + 1); + + if (mci == NULL) { + rc = -ENOMEM; + goto fail; + } + + debugf3("MC: " __FILE__ ": %s(): init mci\n", __func__); + + mci->mtype_cap = MEM_FLAG_RDDR; + mci->edac_ctl_cap = + EDAC_FLAG_NONE | EDAC_FLAG_SECDED | EDAC_FLAG_S4ECD4ED; + /* FIXME - what if different memory types are in different csrows? */ + mci->mod_name = BS_MOD_STR; + mci->mod_ver = "$Revision: 1.5.2.9 $"; + mci->pdev = pdev; + + debugf3("MC: " __FILE__ ": %s(): init pvt\n", __func__); + pvt = (struct e7xxx_pvt *) mci->pvt_info; + pvt->dev_info = &e7xxx_devs[dev_idx]; + pvt->bridge_ck = pci_get_device(PCI_VENDOR_ID_INTEL, + pvt->dev_info->err_dev, + pvt->bridge_ck); + if (!pvt->bridge_ck) { + printk(KERN_ERR + "MC: error reporting device not found:" + "vendor %x device 0x%x (broken BIOS?)\n", + PCI_VENDOR_ID_INTEL, e7xxx_devs[dev_idx].err_dev); + goto fail; + } + + debugf3("MC: " __FILE__ ": %s(): more mci init\n", __func__); + mci->ctl_name = pvt->dev_info->ctl_name; + + mci->edac_check = e7xxx_check; + mci->ctl_page_to_phys = ctl_page_to_phys; + + /* find out the device types */ + pci_read_config_dword(pdev, E7XXX_DRA, &dra); + + /* + * The dram row boundary (DRB) reg values are boundary address + * for each DRAM row with a granularity of 32 or 64MB (single/dual + * channel operation). DRB regs are cumulative; therefore DRB7 will + * contain the total memory contained in all eight rows. + */ + for (last_cumul_size = index = 0; index < mci->nr_csrows; index++) { + u8 value; + u32 cumul_size; + /* mem_dev 0=x8, 1=x4 */ + int mem_dev = (dra >> (index * 4 + 3)) & 0x1; + struct csrow_info *csrow = &mci->csrows[index]; + + pci_read_config_byte(mci->pdev, E7XXX_DRB + index, &value); + /* convert a 64 or 32 MiB DRB to a page size. */ + cumul_size = value << (25 + drc_drbg - PAGE_SHIFT); + debugf3("MC: " __FILE__ ": %s(): (%d) cumul_size 0x%x\n", + __func__, index, cumul_size); + if (cumul_size == last_cumul_size) + continue; /* not populated */ + + csrow->first_page = last_cumul_size; + csrow->last_page = cumul_size - 1; + csrow->nr_pages = cumul_size - last_cumul_size; + last_cumul_size = cumul_size; + csrow->grain = 1 << 12; /* 4KiB - resolution of CELOG */ + csrow->mtype = MEM_RDDR; /* only one type supported */ + csrow->dtype = mem_dev ? DEV_X4 : DEV_X8; + + /* + * if single channel or x8 devices then SECDED + * if dual channel and x4 then S4ECD4ED + */ + if (drc_ddim) { + if (drc_chan && mem_dev) { + csrow->edac_mode = EDAC_S4ECD4ED; + mci->edac_cap |= EDAC_FLAG_S4ECD4ED; + } else { + csrow->edac_mode = EDAC_SECDED; + mci->edac_cap |= EDAC_FLAG_SECDED; + } + } else + csrow->edac_mode = EDAC_NONE; + } + + mci->edac_cap |= EDAC_FLAG_NONE; + + debugf3("MC: " __FILE__ ": %s(): tolm, remapbase, remaplimit\n", + __func__); + /* load the top of low memory, remap base, and remap limit vars */ + pci_read_config_word(mci->pdev, E7XXX_TOLM, &pci_data); + pvt->tolm = ((u32) pci_data) << 4; + pci_read_config_word(mci->pdev, E7XXX_REMAPBASE, &pci_data); + pvt->remapbase = ((u32) pci_data) << 14; + pci_read_config_word(mci->pdev, E7XXX_REMAPLIMIT, &pci_data); + pvt->remaplimit = ((u32) pci_data) << 14; + printk("tolm = %x, remapbase = %x, remaplimit = %x\n", pvt->tolm, + pvt->remapbase, pvt->remaplimit); + + /* clear any pending errors, or initial state bits */ + pci_write_bits8(pvt->bridge_ck, E7XXX_DRAM_FERR, 0x03, 0x03); + pci_write_bits8(pvt->bridge_ck, E7XXX_DRAM_NERR, 0x03, 0x03); + + if (edac_mc_add_mc(mci) != 0) { + debugf3("MC: " __FILE__ + ": %s(): failed edac_mc_add_mc()\n", + __func__); + goto fail; + } + + /* get this far and it's successful */ + debugf3("MC: " __FILE__ ": %s(): success\n", __func__); + return 0; + +fail: + if (mci != NULL) { + if(pvt != NULL && pvt->bridge_ck) + pci_dev_put(pvt->bridge_ck); + edac_mc_free(mci); + } + + return rc; +} + +/* returns count (>= 0), or negative on error */ +static int __devinit +e7xxx_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) +{ + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + /* wake up and enable device */ + return pci_enable_device(pdev) ? + -EIO : e7xxx_probe1(pdev, ent->driver_data); +} + + +static void __devexit e7xxx_remove_one(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + struct e7xxx_pvt *pvt; + + debugf0(__FILE__ ": %s()\n", __func__); + + if (((mci = edac_mc_find_mci_by_pdev(pdev)) != 0) && + edac_mc_del_mc(mci)) { + pvt = (struct e7xxx_pvt *) mci->pvt_info; + pci_dev_put(pvt->bridge_ck); + edac_mc_free(mci); + } +} + + +static const struct pci_device_id e7xxx_pci_tbl[] __devinitdata = { + {PCI_VEND_DEV(INTEL, 7205_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7205}, + {PCI_VEND_DEV(INTEL, 7500_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7500}, + {PCI_VEND_DEV(INTEL, 7501_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7501}, + {PCI_VEND_DEV(INTEL, 7505_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + E7505}, + {0,} /* 0 terminated list. */ +}; + +MODULE_DEVICE_TABLE(pci, e7xxx_pci_tbl); + + +static struct pci_driver e7xxx_driver = { + .name = BS_MOD_STR, + .probe = e7xxx_init_one, + .remove = __devexit_p(e7xxx_remove_one), + .id_table = e7xxx_pci_tbl, +}; + + +static int __init e7xxx_init(void) +{ + return pci_register_driver(&e7xxx_driver); +} + + +static void __exit e7xxx_exit(void) +{ + pci_unregister_driver(&e7xxx_driver); +} + +module_init(e7xxx_init); +module_exit(e7xxx_exit); + + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Linux Networx (http://lnxi.com) Thayne Harbaugh et al\n" + "Based on.work by Dan Hollis et al"); +MODULE_DESCRIPTION("MC support for Intel e7xxx memory controllers"); diff --git a/drivers/edac/edac_mc.c b/drivers/edac/edac_mc.c new file mode 100644 index 00000000000..4be9bd0a126 --- /dev/null +++ b/drivers/edac/edac_mc.c @@ -0,0 +1,2209 @@ +/* + * edac_mc kernel module + * (C) 2005 Linux Networx (http://lnxi.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * Written by Thayne Harbaugh + * Based on work by Dan Hollis <goemon at anime dot net> and others. + * http://www.anime.net/~goemon/linux-ecc/ + * + * Modified by Dave Peterson and Doug Thompson + * + */ + + +#include <linux/config.h> +#include <linux/version.h> +#include <linux/module.h> +#include <linux/proc_fs.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/smp.h> +#include <linux/init.h> +#include <linux/sysctl.h> +#include <linux/highmem.h> +#include <linux/timer.h> +#include <linux/slab.h> +#include <linux/jiffies.h> +#include <linux/spinlock.h> +#include <linux/list.h> +#include <linux/sysdev.h> +#include <linux/ctype.h> + +#include <asm/uaccess.h> +#include <asm/page.h> +#include <asm/edac.h> + +#include "edac_mc.h" + +#define EDAC_MC_VERSION "edac_mc Ver: 2.0.0 " __DATE__ + +#ifdef CONFIG_EDAC_DEBUG +/* Values of 0 to 4 will generate output */ +int edac_debug_level = 1; +EXPORT_SYMBOL(edac_debug_level); +#endif + +/* EDAC Controls, setable by module parameter, and sysfs */ +static int log_ue = 1; +static int log_ce = 1; +static int panic_on_ue = 1; +static int poll_msec = 1000; + +static int check_pci_parity = 0; /* default YES check PCI parity */ +static int panic_on_pci_parity; /* default no panic on PCI Parity */ +static atomic_t pci_parity_count = ATOMIC_INIT(0); + +/* lock to memory controller's control array */ +static DECLARE_MUTEX(mem_ctls_mutex); +static struct list_head mc_devices = LIST_HEAD_INIT(mc_devices); + +/* Structure of the whitelist and blacklist arrays */ +struct edac_pci_device_list { + unsigned int vendor; /* Vendor ID */ + unsigned int device; /* Deviice ID */ +}; + + +#define MAX_LISTED_PCI_DEVICES 32 + +/* List of PCI devices (vendor-id:device-id) that should be skipped */ +static struct edac_pci_device_list pci_blacklist[MAX_LISTED_PCI_DEVICES]; +static int pci_blacklist_count; + +/* List of PCI devices (vendor-id:device-id) that should be scanned */ +static struct edac_pci_device_list pci_whitelist[MAX_LISTED_PCI_DEVICES]; +static int pci_whitelist_count ; + +/* START sysfs data and methods */ + +static const char *mem_types[] = { + [MEM_EMPTY] = "Empty", + [MEM_RESERVED] = "Reserved", + [MEM_UNKNOWN] = "Unknown", + [MEM_FPM] = "FPM", + [MEM_EDO] = "EDO", + [MEM_BEDO] = "BEDO", + [MEM_SDR] = "Unbuffered-SDR", + [MEM_RDR] = "Registered-SDR", + [MEM_DDR] = "Unbuffered-DDR", + [MEM_RDDR] = "Registered-DDR", + [MEM_RMBS] = "RMBS" +}; + +static const char *dev_types[] = { + [DEV_UNKNOWN] = "Unknown", + [DEV_X1] = "x1", + [DEV_X2] = "x2", + [DEV_X4] = "x4", + [DEV_X8] = "x8", + [DEV_X16] = "x16", + [DEV_X32] = "x32", + [DEV_X64] = "x64" +}; + +static const char *edac_caps[] = { + [EDAC_UNKNOWN] = "Unknown", + [EDAC_NONE] = "None", + [EDAC_RESERVED] = "Reserved", + [EDAC_PARITY] = "PARITY", + [EDAC_EC] = "EC", + [EDAC_SECDED] = "SECDED", + [EDAC_S2ECD2ED] = "S2ECD2ED", + [EDAC_S4ECD4ED] = "S4ECD4ED", + [EDAC_S8ECD8ED] = "S8ECD8ED", + [EDAC_S16ECD16ED] = "S16ECD16ED" +}; + + +/* sysfs object: /sys/devices/system/edac */ +static struct sysdev_class edac_class = { + set_kset_name("edac"), +}; + +/* sysfs objects: + * /sys/devices/system/edac/mc + * /sys/devices/system/edac/pci + */ +static struct kobject edac_memctrl_kobj; +static struct kobject edac_pci_kobj; + +/* + * /sys/devices/system/edac/mc; + * data structures and methods + */ +static ssize_t memctrl_string_show(void *ptr, char *buffer) +{ + char *value = (char*) ptr; + return sprintf(buffer, "%s\n", value); +} + +static ssize_t memctrl_int_show(void *ptr, char *buffer) +{ + int *value = (int*) ptr; + return sprintf(buffer, "%d\n", *value); +} + +static ssize_t memctrl_int_store(void *ptr, const char *buffer, size_t count) +{ + int *value = (int*) ptr; + + if (isdigit(*buffer)) + *value = simple_strtoul(buffer, NULL, 0); + + return count; +} + +struct memctrl_dev_attribute { + struct attribute attr; + void *value; + ssize_t (*show)(void *,char *); + ssize_t (*store)(void *, const char *, size_t); +}; + +/* Set of show/store abstract level functions for memory control object */ +static ssize_t +memctrl_dev_show(struct kobject *kobj, struct attribute *attr, char *buffer) +{ + struct memctrl_dev_attribute *memctrl_dev; + memctrl_dev = (struct memctrl_dev_attribute*)attr; + + if (memctrl_dev->show) + return memctrl_dev->show(memctrl_dev->value, buffer); + return -EIO; +} + +static ssize_t +memctrl_dev_store(struct kobject *kobj, struct attribute *attr, + const char *buffer, size_t count) +{ + struct memctrl_dev_attribute *memctrl_dev; + memctrl_dev = (struct memctrl_dev_attribute*)attr; + + if (memctrl_dev->store) + return memctrl_dev->store(memctrl_dev->value, buffer, count); + return -EIO; +} + +static struct sysfs_ops memctrlfs_ops = { + .show = memctrl_dev_show, + .store = memctrl_dev_store +}; + +#define MEMCTRL_ATTR(_name,_mode,_show,_store) \ +struct memctrl_dev_attribute attr_##_name = { \ + .attr = {.name = __stringify(_name), .mode = _mode }, \ + .value = &_name, \ + .show = _show, \ + .store = _store, \ +}; + +#define MEMCTRL_STRING_ATTR(_name,_data,_mode,_show,_store) \ +struct memctrl_dev_attribute attr_##_name = { \ + .attr = {.name = __stringify(_name), .mode = _mode }, \ + .value = _data, \ + .show = _show, \ + .store = _store, \ +}; + +/* cwrow<id> attribute f*/ +MEMCTRL_STRING_ATTR(mc_version,EDAC_MC_VERSION,S_IRUGO,memctrl_string_show,NULL); + +/* csrow<id> control files */ +MEMCTRL_ATTR(panic_on_ue,S_IRUGO|S_IWUSR,memctrl_int_show,memctrl_int_store); +MEMCTRL_ATTR(log_ue,S_IRUGO|S_IWUSR,memctrl_int_show,memctrl_int_store); +MEMCTRL_ATTR(log_ce,S_IRUGO|S_IWUSR,memctrl_int_show,memctrl_int_store); +MEMCTRL_ATTR(poll_msec,S_IRUGO|S_IWUSR,memctrl_int_show,memctrl_int_store); + + +/* Base Attributes of the memory ECC object */ +static struct memctrl_dev_attribute *memctrl_attr[] = { + &attr_panic_on_ue, + &attr_log_ue, + &attr_log_ce, + &attr_poll_msec, + &attr_mc_version, + NULL, +}; + +/* Main MC kobject release() function */ +static void edac_memctrl_master_release(struct kobject *kobj) +{ + debugf1("EDAC MC: " __FILE__ ": %s()\n", __func__); +} + +static struct kobj_type ktype_memctrl = { + .release = edac_memctrl_master_release, + .sysfs_ops = &memctrlfs_ops, + .default_attrs = (struct attribute **) memctrl_attr, +}; + + +/* Initialize the main sysfs entries for edac: + * /sys/devices/system/edac + * + * and children + * + * Return: 0 SUCCESS + * !0 FAILURE + */ +static int edac_sysfs_memctrl_setup(void) +{ + int err=0; + + debugf1("MC: " __FILE__ ": %s()\n", __func__); + + /* create the /sys/devices/system/edac directory */ + err = sysdev_class_register(&edac_class); + if (!err) { + /* Init the MC's kobject */ + memset(&edac_memctrl_kobj, 0, sizeof (edac_memctrl_kobj)); + kobject_init(&edac_memctrl_kobj); + + edac_memctrl_kobj.parent = &edac_class.kset.kobj; + edac_memctrl_kobj.ktype = &ktype_memctrl; + + /* generate sysfs "..../edac/mc" */ + err = kobject_set_name(&edac_memctrl_kobj,"mc"); + if (!err) { + /* FIXME: maybe new sysdev_create_subdir() */ + err = kobject_register(&edac_memctrl_kobj); + if (err) { + debugf1("Failed to register '.../edac/mc'\n"); + } else { + debugf1("Registered '.../edac/mc' kobject\n"); + } + } + } else { + debugf1(KERN_WARNING "__FILE__ %s() error=%d\n", __func__,err); + } + + return err; +} + +/* + * MC teardown: + * the '..../edac/mc' kobject followed by '..../edac' itself + */ +static void edac_sysfs_memctrl_teardown(void) +{ + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + /* Unregister the MC's kobject */ + kobject_unregister(&edac_memctrl_kobj); + + /* release the master edac mc kobject */ + kobject_put(&edac_memctrl_kobj); + + /* Unregister the 'edac' object */ + sysdev_class_unregister(&edac_class); +} + +/* + * /sys/devices/system/edac/pci; + * data structures and methods + */ + +struct list_control { + struct edac_pci_device_list *list; + int *count; +}; + +/* Output the list as: vendor_id:device:id<,vendor_id:device_id> */ +static ssize_t edac_pci_list_string_show(void *ptr, char *buffer) +{ + struct list_control *listctl; + struct edac_pci_device_list *list; + char *p = buffer; + int len=0; + int i; + + listctl = ptr; + list = listctl->list; + + for (i = 0; i < *(listctl->count); i++, list++ ) { + if (len > 0) + len += snprintf(p + len, (PAGE_SIZE-len), ","); + + len += snprintf(p + len, + (PAGE_SIZE-len), + "%x:%x", + list->vendor,list->device); + } + + len += snprintf(p + len,(PAGE_SIZE-len), "\n"); + + return (ssize_t) len; +} + +/** + * + * Scan string from **s to **e looking for one 'vendor:device' tuple + * where each field is a hex value + * + * return 0 if an entry is NOT found + * return 1 if an entry is found + * fill in *vendor_id and *device_id with values found + * + * In both cases, make sure *s has been moved forward toward *e + */ +static int parse_one_device(const char **s,const char **e, + unsigned int *vendor_id, unsigned int *device_id) +{ + const char *runner, *p; + + /* if null byte, we are done */ + if (!**s) { + (*s)++; /* keep *s moving */ + return 0; + } + + /* skip over newlines & whitespace */ + if ((**s == '\n') || isspace(**s)) { + (*s)++; + return 0; + } + + if (!isxdigit(**s)) { + (*s)++; + return 0; + } + + /* parse vendor_id */ + runner = *s; + while (runner < *e) { + /* scan for vendor:device delimiter */ + if (*runner == ':') { + *vendor_id = simple_strtol((char*) *s, (char**) &p, 16); + runner = p + 1; + break; + } + runner++; + } + + if (!isxdigit(*runner)) { + *s = ++runner; + return 0; + } + + /* parse device_id */ + if (runner < *e) { + *device_id = simple_strtol((char*)runner, (char**)&p, 16); + runner = p; + } + + *s = runner; + + return 1; +} + +static ssize_t edac_pci_list_string_store(void *ptr, const char *buffer, + size_t count) +{ + struct list_control *listctl; + struct edac_pci_device_list *list; + unsigned int vendor_id, device_id; + const char *s, *e; + int *index; + + s = (char*)buffer; + e = s + count; + + listctl = ptr; + list = listctl->list; + index = listctl->count; + + *index = 0; + while (*index < MAX_LISTED_PCI_DEVICES) { + + if (parse_one_device(&s,&e,&vendor_id,&device_id)) { + list[ *index ].vendor = vendor_id; + list[ *index ].device = device_id; + (*index)++; + } + + /* check for all data consume */ + if (s >= e) + break; + } + + return count; +} + +static ssize_t edac_pci_int_show(void *ptr, char *buffer) +{ + int *value = ptr; + return sprintf(buffer,"%d\n",*value); +} + +static ssize_t edac_pci_int_store(void *ptr, const char *buffer, size_t count) +{ + int *value = ptr; + + if (isdigit(*buffer)) + *value = simple_strtoul(buffer,NULL,0); + + return count; +} + +struct edac_pci_dev_attribute { + struct attribute attr; + void *value; + ssize_t (*show)(void *,char *); + ssize_t (*store)(void *, const char *,size_t); +}; + +/* Set of show/store abstract level functions for PCI Parity object */ +static ssize_t edac_pci_dev_show(struct kobject *kobj, struct attribute *attr, + char *buffer) +{ + struct edac_pci_dev_attribute *edac_pci_dev; + edac_pci_dev= (struct edac_pci_dev_attribute*)attr; + + if (edac_pci_dev->show) + return edac_pci_dev->show(edac_pci_dev->value, buffer); + return -EIO; +} + +static ssize_t edac_pci_dev_store(struct kobject *kobj, struct attribute *attr, + const char *buffer, size_t count) +{ + struct edac_pci_dev_attribute *edac_pci_dev; + edac_pci_dev= (struct edac_pci_dev_attribute*)attr; + + if (edac_pci_dev->show) + return edac_pci_dev->store(edac_pci_dev->value, buffer, count); + return -EIO; +} + +static struct sysfs_ops edac_pci_sysfs_ops = { + .show = edac_pci_dev_show, + .store = edac_pci_dev_store +}; + + +#define EDAC_PCI_ATTR(_name,_mode,_show,_store) \ +struct edac_pci_dev_attribute edac_pci_attr_##_name = { \ + .attr = {.name = __stringify(_name), .mode = _mode }, \ + .value = &_name, \ + .show = _show, \ + .store = _store, \ +}; + +#define EDAC_PCI_STRING_ATTR(_name,_data,_mode,_show,_store) \ +struct edac_pci_dev_attribute edac_pci_attr_##_name = { \ + .attr = {.name = __stringify(_name), .mode = _mode }, \ + .value = _data, \ + .show = _show, \ + .store = _store, \ +}; + +static struct list_control pci_whitelist_control = { + .list = pci_whitelist, + .count = &pci_whitelist_count +}; + +static struct list_control pci_blacklist_control = { + .list = pci_blacklist, + .count = &pci_blacklist_count +}; + +/* whitelist attribute */ +EDAC_PCI_STRING_ATTR(pci_parity_whitelist, + &pci_whitelist_control, + S_IRUGO|S_IWUSR, + edac_pci_list_string_show, + edac_pci_list_string_store); + +EDAC_PCI_STRING_ATTR(pci_parity_blacklist, + &pci_blacklist_control, + S_IRUGO|S_IWUSR, + edac_pci_list_string_show, + edac_pci_list_string_store); + +/* PCI Parity control files */ +EDAC_PCI_ATTR(check_pci_parity,S_IRUGO|S_IWUSR,edac_pci_int_show,edac_pci_int_store); +EDAC_PCI_ATTR(panic_on_pci_parity,S_IRUGO|S_IWUSR,edac_pci_int_show,edac_pci_int_store); +EDAC_PCI_ATTR(pci_parity_count,S_IRUGO,edac_pci_int_show,NULL); + +/* Base Attributes of the memory ECC object */ +static struct edac_pci_dev_attribute *edac_pci_attr[] = { + &edac_pci_attr_check_pci_parity, + &edac_pci_attr_panic_on_pci_parity, + &edac_pci_attr_pci_parity_count, + &edac_pci_attr_pci_parity_whitelist, + &edac_pci_attr_pci_parity_blacklist, + NULL, +}; + +/* No memory to release */ +static void edac_pci_release(struct kobject *kobj) +{ + debugf1("EDAC PCI: " __FILE__ ": %s()\n", __func__); +} + +static struct kobj_type ktype_edac_pci = { + .release = edac_pci_release, + .sysfs_ops = &edac_pci_sysfs_ops, + .default_attrs = (struct attribute **) edac_pci_attr, +}; + +/** + * edac_sysfs_pci_setup() + * + */ +static int edac_sysfs_pci_setup(void) +{ + int err; + + debugf1("MC: " __FILE__ ": %s()\n", __func__); + + memset(&edac_pci_kobj, 0, sizeof(edac_pci_kobj)); + + kobject_init(&edac_pci_kobj); + edac_pci_kobj.parent = &edac_class.kset.kobj; + edac_pci_kobj.ktype = &ktype_edac_pci; + + err = kobject_set_name(&edac_pci_kobj, "pci"); + if (!err) { + /* Instanstiate the csrow object */ + /* FIXME: maybe new sysdev_create_subdir() */ + err = kobject_register(&edac_pci_kobj); + if (err) + debugf1("Failed to register '.../edac/pci'\n"); + else + debugf1("Registered '.../edac/pci' kobject\n"); + } + return err; +} + + +static void edac_sysfs_pci_teardown(void) +{ + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + kobject_unregister(&edac_pci_kobj); + kobject_put(&edac_pci_kobj); +} + +/* EDAC sysfs CSROW data structures and methods */ + +/* Set of more detailed csrow<id> attribute show/store functions */ +static ssize_t csrow_ch0_dimm_label_show(struct csrow_info *csrow, char *data) +{ + ssize_t size = 0; + + if (csrow->nr_channels > 0) { + size = snprintf(data, EDAC_MC_LABEL_LEN,"%s\n", + csrow->channels[0].label); + } + return size; +} + +static ssize_t csrow_ch1_dimm_label_show(struct csrow_info *csrow, char *data) +{ + ssize_t size = 0; + + if (csrow->nr_channels > 0) { + size = snprintf(data, EDAC_MC_LABEL_LEN, "%s\n", + csrow->channels[1].label); + } + return size; +} + +static ssize_t csrow_ch0_dimm_label_store(struct csrow_info *csrow, + const char *data, size_t size) +{ + ssize_t max_size = 0; + + if (csrow->nr_channels > 0) { + max_size = min((ssize_t)size,(ssize_t)EDAC_MC_LABEL_LEN-1); + strncpy(csrow->channels[0].label, data, max_size); + csrow->channels[0].label[max_size] = '\0'; + } + return size; +} + +static ssize_t csrow_ch1_dimm_label_store(struct csrow_info *csrow, + const char *data, size_t size) +{ + ssize_t max_size = 0; + + if (csrow->nr_channels > 1) { + max_size = min((ssize_t)size,(ssize_t)EDAC_MC_LABEL_LEN-1); + strncpy(csrow->channels[1].label, data, max_size); + csrow->channels[1].label[max_size] = '\0'; + } + return max_size; +} + +static ssize_t csrow_ue_count_show(struct csrow_info *csrow, char *data) +{ + return sprintf(data,"%u\n", csrow->ue_count); +} + +static ssize_t csrow_ce_count_show(struct csrow_info *csrow, char *data) +{ + return sprintf(data,"%u\n", csrow->ce_count); +} + +static ssize_t csrow_ch0_ce_count_show(struct csrow_info *csrow, char *data) +{ + ssize_t size = 0; + + if (csrow->nr_channels > 0) { + size = sprintf(data,"%u\n", csrow->channels[0].ce_count); + } + return size; +} + +static ssize_t csrow_ch1_ce_count_show(struct csrow_info *csrow, char *data) +{ + ssize_t size = 0; + + if (csrow->nr_channels > 1) { + size = sprintf(data,"%u\n", csrow->channels[1].ce_count); + } + return size; +} + +static ssize_t csrow_size_show(struct csrow_info *csrow, char *data) +{ + return sprintf(data,"%u\n", PAGES_TO_MiB(csrow->nr_pages)); +} + +static ssize_t csrow_mem_type_show(struct csrow_info *csrow, char *data) +{ + return sprintf(data,"%s\n", mem_types[csrow->mtype]); +} + +static ssize_t csrow_dev_type_show(struct csrow_info *csrow, char *data) +{ + return sprintf(data,"%s\n", dev_types[csrow->dtype]); +} + +static ssize_t csrow_edac_mode_show(struct csrow_info *csrow, char *data) +{ + return sprintf(data,"%s\n", edac_caps[csrow->edac_mode]); +} + +struct csrowdev_attribute { + struct attribute attr; + ssize_t (*show)(struct csrow_info *,char *); + ssize_t (*store)(struct csrow_info *, const char *,size_t); +}; + +#define to_csrow(k) container_of(k, struct csrow_info, kobj) +#define to_csrowdev_attr(a) container_of(a, struct csrowdev_attribute, attr) + +/* Set of show/store higher level functions for csrow objects */ +static ssize_t csrowdev_show(struct kobject *kobj, struct attribute *attr, + char *buffer) +{ + struct csrow_info *csrow = to_csrow(kobj); + struct csrowdev_attribute *csrowdev_attr = to_csrowdev_attr(attr); + + if (csrowdev_attr->show) + return csrowdev_attr->show(csrow, buffer); + return -EIO; +} + +static ssize_t csrowdev_store(struct kobject *kobj, struct attribute *attr, + const char *buffer, size_t count) +{ + struct csrow_info *csrow = to_csrow(kobj); + struct csrowdev_attribute * csrowdev_attr = to_csrowdev_attr(attr); + + if (csrowdev_attr->store) + return csrowdev_attr->store(csrow, buffer, count); + return -EIO; +} + +static struct sysfs_ops csrowfs_ops = { + .show = csrowdev_show, + .store = csrowdev_store +}; + +#define CSROWDEV_ATTR(_name,_mode,_show,_store) \ +struct csrowdev_attribute attr_##_name = { \ + .attr = {.name = __stringify(_name), .mode = _mode }, \ + .show = _show, \ + .store = _store, \ +}; + +/* cwrow<id>/attribute files */ +CSROWDEV_ATTR(size_mb,S_IRUGO,csrow_size_show,NULL); +CSROWDEV_ATTR(dev_type,S_IRUGO,csrow_dev_type_show,NULL); +CSROWDEV_ATTR(mem_type,S_IRUGO,csrow_mem_type_show,NULL); +CSROWDEV_ATTR(edac_mode,S_IRUGO,csrow_edac_mode_show,NULL); +CSROWDEV_ATTR(ue_count,S_IRUGO,csrow_ue_count_show,NULL); +CSROWDEV_ATTR(ce_count,S_IRUGO,csrow_ce_count_show,NULL); +CSROWDEV_ATTR(ch0_ce_count,S_IRUGO,csrow_ch0_ce_count_show,NULL); +CSROWDEV_ATTR(ch1_ce_count,S_IRUGO,csrow_ch1_ce_count_show,NULL); + +/* control/attribute files */ +CSROWDEV_ATTR(ch0_dimm_label,S_IRUGO|S_IWUSR, + csrow_ch0_dimm_label_show, + csrow_ch0_dimm_label_store); +CSROWDEV_ATTR(ch1_dimm_label,S_IRUGO|S_IWUSR, + csrow_ch1_dimm_label_show, + csrow_ch1_dimm_label_store); + + +/* Attributes of the CSROW<id> object */ +static struct csrowdev_attribute *csrow_attr[] = { + &attr_dev_type, + &attr_mem_type, + &attr_edac_mode, + &attr_size_mb, + &attr_ue_count, + &attr_ce_count, + &attr_ch0_ce_count, + &attr_ch1_ce_count, + &attr_ch0_dimm_label, + &attr_ch1_dimm_label, + NULL, +}; + + +/* No memory to release */ +static void edac_csrow_instance_release(struct kobject *kobj) +{ + debugf1("EDAC MC: " __FILE__ ": %s()\n", __func__); +} + +static struct kobj_type ktype_csrow = { + .release = edac_csrow_instance_release, + .sysfs_ops = &csrowfs_ops, + .default_attrs = (struct attribute **) csrow_attr, +}; + +/* Create a CSROW object under specifed edac_mc_device */ +static int edac_create_csrow_object(struct kobject *edac_mci_kobj, + struct csrow_info *csrow, int index ) +{ + int err = 0; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + memset(&csrow->kobj, 0, sizeof(csrow->kobj)); + + /* generate ..../edac/mc/mc<id>/csrow<index> */ + + kobject_init(&csrow->kobj); + csrow->kobj.parent = edac_mci_kobj; + csrow->kobj.ktype = &ktype_csrow; + + /* name this instance of csrow<id> */ + err = kobject_set_name(&csrow->kobj,"csrow%d",index); + if (!err) { + /* Instanstiate the csrow object */ + err = kobject_register(&csrow->kobj); + if (err) + debugf0("Failed to register CSROW%d\n",index); + else + debugf0("Registered CSROW%d\n",index); + } + + return err; +} + +/* sysfs data structures and methods for the MCI kobjects */ + +static ssize_t mci_reset_counters_store(struct mem_ctl_info *mci, + const char *data, size_t count ) +{ + int row, chan; + + mci->ue_noinfo_count = 0; + mci->ce_noinfo_count = 0; + mci->ue_count = 0; + mci->ce_count = 0; + for (row = 0; row < mci->nr_csrows; row++) { + struct csrow_info *ri = &mci->csrows[row]; + + ri->ue_count = 0; + ri->ce_count = 0; + for (chan = 0; chan < ri->nr_channels; chan++) + ri->channels[chan].ce_count = 0; + } + mci->start_time = jiffies; + + return count; +} + +static ssize_t mci_ue_count_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%d\n", mci->ue_count); +} + +static ssize_t mci_ce_count_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%d\n", mci->ce_count); +} + +static ssize_t mci_ce_noinfo_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%d\n", mci->ce_noinfo_count); +} + +static ssize_t mci_ue_noinfo_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%d\n", mci->ue_noinfo_count); +} + +static ssize_t mci_seconds_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%ld\n", (jiffies - mci->start_time) / HZ); +} + +static ssize_t mci_mod_name_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%s %s\n", mci->mod_name, mci->mod_ver); +} + +static ssize_t mci_ctl_name_show(struct mem_ctl_info *mci, char *data) +{ + return sprintf(data,"%s\n", mci->ctl_name); +} + +static int mci_output_edac_cap(char *buf, unsigned long edac_cap) +{ + char *p = buf; + int bit_idx; + + for (bit_idx = 0; bit_idx < 8 * sizeof(edac_cap); bit_idx++) { + if ((edac_cap >> bit_idx) & 0x1) + p += sprintf(p, "%s ", edac_caps[bit_idx]); + } + + return p - buf; +} + +static ssize_t mci_edac_capability_show(struct mem_ctl_info *mci, char *data) +{ + char *p = data; + + p += mci_output_edac_cap(p,mci->edac_ctl_cap); + p += sprintf(p, "\n"); + + return p - data; +} + +static ssize_t mci_edac_current_capability_show(struct mem_ctl_info *mci, + char *data) +{ + char *p = data; + + p += mci_output_edac_cap(p,mci->edac_cap); + p += sprintf(p, "\n"); + + return p - data; +} + +static int mci_output_mtype_cap(char *buf, unsigned long mtype_cap) +{ + char *p = buf; + int bit_idx; + + for (bit_idx = 0; bit_idx < 8 * sizeof(mtype_cap); bit_idx++) { + if ((mtype_cap >> bit_idx) & 0x1) + p += sprintf(p, "%s ", mem_types[bit_idx]); + } + + return p - buf; +} + +static ssize_t mci_supported_mem_type_show(struct mem_ctl_info *mci, char *data) +{ + char *p = data; + + p += mci_output_mtype_cap(p,mci->mtype_cap); + p += sprintf(p, "\n"); + + return p - data; +} + +static ssize_t mci_size_mb_show(struct mem_ctl_info *mci, char *data) +{ + int total_pages, csrow_idx; + + for (total_pages = csrow_idx = 0; csrow_idx < mci->nr_csrows; + csrow_idx++) { + struct csrow_info *csrow = &mci->csrows[csrow_idx]; + + if (!csrow->nr_pages) + continue; + total_pages += csrow->nr_pages; + } + + return sprintf(data,"%u\n", PAGES_TO_MiB(total_pages)); +} + +struct mcidev_attribute { + struct attribute attr; + ssize_t (*show)(struct mem_ctl_info *,char *); + ssize_t (*store)(struct mem_ctl_info *, const char *,size_t); +}; + +#define to_mci(k) container_of(k, struct mem_ctl_info, edac_mci_kobj) +#define to_mcidev_attr(a) container_of(a, struct mcidev_attribute, attr) + +static ssize_t mcidev_show(struct kobject *kobj, struct attribute *attr, + char *buffer) +{ + struct mem_ctl_info *mem_ctl_info = to_mci(kobj); + struct mcidev_attribute * mcidev_attr = to_mcidev_attr(attr); + + if (mcidev_attr->show) + return mcidev_attr->show(mem_ctl_info, buffer); + return -EIO; +} + +static ssize_t mcidev_store(struct kobject *kobj, struct attribute *attr, + const char *buffer, size_t count) +{ + struct mem_ctl_info *mem_ctl_info = to_mci(kobj); + struct mcidev_attribute * mcidev_attr = to_mcidev_attr(attr); + + if (mcidev_attr->store) + return mcidev_attr->store(mem_ctl_info, buffer, count); + return -EIO; +} + +static struct sysfs_ops mci_ops = { + .show = mcidev_show, + .store = mcidev_store +}; + +#define MCIDEV_ATTR(_name,_mode,_show,_store) \ +struct mcidev_attribute mci_attr_##_name = { \ + .attr = {.name = __stringify(_name), .mode = _mode }, \ + .show = _show, \ + .store = _store, \ +}; + +/* Control file */ +MCIDEV_ATTR(reset_counters,S_IWUSR,NULL,mci_reset_counters_store); + +/* Attribute files */ +MCIDEV_ATTR(mc_name,S_IRUGO,mci_ctl_name_show,NULL); +MCIDEV_ATTR(module_name,S_IRUGO,mci_mod_name_show,NULL); +MCIDEV_ATTR(edac_capability,S_IRUGO,mci_edac_capability_show,NULL); +MCIDEV_ATTR(size_mb,S_IRUGO,mci_size_mb_show,NULL); +MCIDEV_ATTR(seconds_since_reset,S_IRUGO,mci_seconds_show,NULL); +MCIDEV_ATTR(ue_noinfo_count,S_IRUGO,mci_ue_noinfo_show,NULL); +MCIDEV_ATTR(ce_noinfo_count,S_IRUGO,mci_ce_noinfo_show,NULL); +MCIDEV_ATTR(ue_count,S_IRUGO,mci_ue_count_show,NULL); +MCIDEV_ATTR(ce_count,S_IRUGO,mci_ce_count_show,NULL); +MCIDEV_ATTR(edac_current_capability,S_IRUGO, + mci_edac_current_capability_show,NULL); +MCIDEV_ATTR(supported_mem_type,S_IRUGO, + mci_supported_mem_type_show,NULL); + + +static struct mcidev_attribute *mci_attr[] = { + &mci_attr_reset_counters, + &mci_attr_module_name, + &mci_attr_mc_name, + &mci_attr_edac_capability, + &mci_attr_edac_current_capability, + &mci_attr_supported_mem_type, + &mci_attr_size_mb, + &mci_attr_seconds_since_reset, + &mci_attr_ue_noinfo_count, + &mci_attr_ce_noinfo_count, + &mci_attr_ue_count, + &mci_attr_ce_count, + NULL +}; + + +/* + * Release of a MC controlling instance + */ +static void edac_mci_instance_release(struct kobject *kobj) +{ + struct mem_ctl_info *mci; + mci = container_of(kobj,struct mem_ctl_info,edac_mci_kobj); + + debugf0("MC: " __FILE__ ": %s() idx=%d calling kfree\n", + __func__, mci->mc_idx); + + kfree(mci); +} + +static struct kobj_type ktype_mci = { + .release = edac_mci_instance_release, + .sysfs_ops = &mci_ops, + .default_attrs = (struct attribute **) mci_attr, +}; + +#define EDAC_DEVICE_SYMLINK "device" + +/* + * Create a new Memory Controller kobject instance, + * mc<id> under the 'mc' directory + * + * Return: + * 0 Success + * !0 Failure + */ +static int edac_create_sysfs_mci_device(struct mem_ctl_info *mci) +{ + int i; + int err; + struct csrow_info *csrow; + struct kobject *edac_mci_kobj=&mci->edac_mci_kobj; + + debugf0("MC: " __FILE__ ": %s() idx=%d\n", __func__, mci->mc_idx); + + memset(edac_mci_kobj, 0, sizeof(*edac_mci_kobj)); + kobject_init(edac_mci_kobj); + + /* set the name of the mc<id> object */ + err = kobject_set_name(edac_mci_kobj,"mc%d",mci->mc_idx); + if (err) + return err; + + /* link to our parent the '..../edac/mc' object */ + edac_mci_kobj->parent = &edac_memctrl_kobj; + edac_mci_kobj->ktype = &ktype_mci; + + /* register the mc<id> kobject */ + err = kobject_register(edac_mci_kobj); + if (err) + return err; + + /* create a symlink for the device */ + err = sysfs_create_link(edac_mci_kobj, &mci->pdev->dev.kobj, + EDAC_DEVICE_SYMLINK); + if (err) { + kobject_unregister(edac_mci_kobj); + return err; + } + + /* Make directories for each CSROW object + * under the mc<id> kobject + */ + for (i = 0; i < mci->nr_csrows; i++) { + + csrow = &mci->csrows[i]; + + /* Only expose populated CSROWs */ + if (csrow->nr_pages > 0) { + err = edac_create_csrow_object(edac_mci_kobj,csrow,i); + if (err) + goto fail; + } + } + + /* Mark this MCI instance as having sysfs entries */ + mci->sysfs_active = MCI_SYSFS_ACTIVE; + + return 0; + + + /* CSROW error: backout what has already been registered, */ +fail: + for ( i--; i >= 0; i--) { + if (csrow->nr_pages > 0) { + kobject_unregister(&mci->csrows[i].kobj); + kobject_put(&mci->csrows[i].kobj); + } + } + + kobject_unregister(edac_mci_kobj); + kobject_put(edac_mci_kobj); + + return err; +} + +/* + * remove a Memory Controller instance + */ +static void edac_remove_sysfs_mci_device(struct mem_ctl_info *mci) +{ + int i; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + /* remove all csrow kobjects */ + for (i = 0; i < mci->nr_csrows; i++) { + if (mci->csrows[i].nr_pages > 0) { + kobject_unregister(&mci->csrows[i].kobj); + kobject_put(&mci->csrows[i].kobj); + } + } + + sysfs_remove_link(&mci->edac_mci_kobj, EDAC_DEVICE_SYMLINK); + + kobject_unregister(&mci->edac_mci_kobj); + kobject_put(&mci->edac_mci_kobj); +} + +/* END OF sysfs data and methods */ + +#ifdef CONFIG_EDAC_DEBUG + +EXPORT_SYMBOL(edac_mc_dump_channel); + +void edac_mc_dump_channel(struct channel_info *chan) +{ + debugf4("\tchannel = %p\n", chan); + debugf4("\tchannel->chan_idx = %d\n", chan->chan_idx); + debugf4("\tchannel->ce_count = %d\n", chan->ce_count); + debugf4("\tchannel->label = '%s'\n", chan->label); + debugf4("\tchannel->csrow = %p\n\n", chan->csrow); +} + + +EXPORT_SYMBOL(edac_mc_dump_csrow); + +void edac_mc_dump_csrow(struct csrow_info *csrow) +{ + debugf4("\tcsrow = %p\n", csrow); + debugf4("\tcsrow->csrow_idx = %d\n", csrow->csrow_idx); + debugf4("\tcsrow->first_page = 0x%lx\n", + csrow->first_page); + debugf4("\tcsrow->last_page = 0x%lx\n", csrow->last_page); + debugf4("\tcsrow->page_mask = 0x%lx\n", csrow->page_mask); + debugf4("\tcsrow->nr_pages = 0x%x\n", csrow->nr_pages); + debugf4("\tcsrow->nr_channels = %d\n", + csrow->nr_channels); + debugf4("\tcsrow->channels = %p\n", csrow->channels); + debugf4("\tcsrow->mci = %p\n\n", csrow->mci); +} + + +EXPORT_SYMBOL(edac_mc_dump_mci); + +void edac_mc_dump_mci(struct mem_ctl_info *mci) +{ + debugf3("\tmci = %p\n", mci); + debugf3("\tmci->mtype_cap = %lx\n", mci->mtype_cap); + debugf3("\tmci->edac_ctl_cap = %lx\n", mci->edac_ctl_cap); + debugf3("\tmci->edac_cap = %lx\n", mci->edac_cap); + debugf4("\tmci->edac_check = %p\n", mci->edac_check); + debugf3("\tmci->nr_csrows = %d, csrows = %p\n", + mci->nr_csrows, mci->csrows); + debugf3("\tpdev = %p\n", mci->pdev); + debugf3("\tmod_name:ctl_name = %s:%s\n", + mci->mod_name, mci->ctl_name); + debugf3("\tpvt_info = %p\n\n", mci->pvt_info); +} + + +#endif /* CONFIG_EDAC_DEBUG */ + +/* 'ptr' points to a possibly unaligned item X such that sizeof(X) is 'size'. + * Adjust 'ptr' so that its alignment is at least as stringent as what the + * compiler would provide for X and return the aligned result. + * + * If 'size' is a constant, the compiler will optimize this whole function + * down to either a no-op or the addition of a constant to the value of 'ptr'. + */ +static inline char * align_ptr (void *ptr, unsigned size) +{ + unsigned align, r; + + /* Here we assume that the alignment of a "long long" is the most + * stringent alignment that the compiler will ever provide by default. + * As far as I know, this is a reasonable assumption. + */ + if (size > sizeof(long)) + align = sizeof(long long); + else if (size > sizeof(int)) + align = sizeof(long); + else if (size > sizeof(short)) + align = sizeof(int); + else if (size > sizeof(char)) + align = sizeof(short); + else + return (char *) ptr; + + r = size % align; + + if (r == 0) + return (char *) ptr; + + return (char *) (((unsigned long) ptr) + align - r); +} + + +EXPORT_SYMBOL(edac_mc_alloc); + +/** + * edac_mc_alloc: Allocate a struct mem_ctl_info structure + * @size_pvt: size of private storage needed + * @nr_csrows: Number of CWROWS needed for this MC + * @nr_chans: Number of channels for the MC + * + * Everything is kmalloc'ed as one big chunk - more efficient. + * Only can be used if all structures have the same lifetime - otherwise + * you have to allocate and initialize your own structures. + * + * Use edac_mc_free() to free mc structures allocated by this function. + * + * Returns: + * NULL allocation failed + * struct mem_ctl_info pointer + */ +struct mem_ctl_info *edac_mc_alloc(unsigned sz_pvt, unsigned nr_csrows, + unsigned nr_chans) +{ + struct mem_ctl_info *mci; + struct csrow_info *csi, *csrow; + struct channel_info *chi, *chp, *chan; + void *pvt; + unsigned size; + int row, chn; + + /* Figure out the offsets of the various items from the start of an mc + * structure. We want the alignment of each item to be at least as + * stringent as what the compiler would provide if we could simply + * hardcode everything into a single struct. + */ + mci = (struct mem_ctl_info *) 0; + csi = (struct csrow_info *)align_ptr(&mci[1], sizeof(*csi)); + chi = (struct channel_info *) + align_ptr(&csi[nr_csrows], sizeof(*chi)); + pvt = align_ptr(&chi[nr_chans * nr_csrows], sz_pvt); + size = ((unsigned long) pvt) + sz_pvt; + + if ((mci = kmalloc(size, GFP_KERNEL)) == NULL) + return NULL; + + /* Adjust pointers so they point within the memory we just allocated + * rather than an imaginary chunk of memory located at address 0. + */ + csi = (struct csrow_info *) (((char *) mci) + ((unsigned long) csi)); + chi = (struct channel_info *) (((char *) mci) + ((unsigned long) chi)); + pvt = sz_pvt ? (((char *) mci) + ((unsigned long) pvt)) : NULL; + + memset(mci, 0, size); /* clear all fields */ + + mci->csrows = csi; + mci->pvt_info = pvt; + mci->nr_csrows = nr_csrows; + + for (row = 0; row < nr_csrows; row++) { + csrow = &csi[row]; + csrow->csrow_idx = row; + csrow->mci = mci; + csrow->nr_channels = nr_chans; + chp = &chi[row * nr_chans]; + csrow->channels = chp; + + for (chn = 0; chn < nr_chans; chn++) { + chan = &chp[chn]; + chan->chan_idx = chn; + chan->csrow = csrow; + } + } + + return mci; +} + + +EXPORT_SYMBOL(edac_mc_free); + +/** + * edac_mc_free: Free a previously allocated 'mci' structure + * @mci: pointer to a struct mem_ctl_info structure + * + * Free up a previously allocated mci structure + * A MCI structure can be in 2 states after being allocated + * by edac_mc_alloc(). + * 1) Allocated in a MC driver's probe, but not yet committed + * 2) Allocated and committed, by a call to edac_mc_add_mc() + * edac_mc_add_mc() is the function that adds the sysfs entries + * thus, this free function must determine which state the 'mci' + * structure is in, then either free it directly or + * perform kobject cleanup by calling edac_remove_sysfs_mci_device(). + * + * VOID Return + */ +void edac_mc_free(struct mem_ctl_info *mci) +{ + /* only if sysfs entries for this mci instance exist + * do we remove them and defer the actual kfree via + * the kobject 'release()' callback. + * + * Otherwise, do a straight kfree now. + */ + if (mci->sysfs_active == MCI_SYSFS_ACTIVE) + edac_remove_sysfs_mci_device(mci); + else + kfree(mci); +} + + + +EXPORT_SYMBOL(edac_mc_find_mci_by_pdev); + +struct mem_ctl_info *edac_mc_find_mci_by_pdev(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + struct list_head *item; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + list_for_each(item, &mc_devices) { + mci = list_entry(item, struct mem_ctl_info, link); + + if (mci->pdev == pdev) + return mci; + } + + return NULL; +} + +static int add_mc_to_global_list (struct mem_ctl_info *mci) +{ + struct list_head *item, *insert_before; + struct mem_ctl_info *p; + int i; + + if (list_empty(&mc_devices)) { + mci->mc_idx = 0; + insert_before = &mc_devices; + } else { + if (edac_mc_find_mci_by_pdev(mci->pdev)) { + printk(KERN_WARNING + "EDAC MC: %s (%s) %s %s already assigned %d\n", + mci->pdev->dev.bus_id, pci_name(mci->pdev), + mci->mod_name, mci->ctl_name, mci->mc_idx); + return 1; + } + + insert_before = NULL; + i = 0; + + list_for_each(item, &mc_devices) { + p = list_entry(item, struct mem_ctl_info, link); + + if (p->mc_idx != i) { + insert_before = item; + break; + } + + i++; + } + + mci->mc_idx = i; + + if (insert_before == NULL) + insert_before = &mc_devices; + } + + list_add_tail_rcu(&mci->link, insert_before); + return 0; +} + + + +EXPORT_SYMBOL(edac_mc_add_mc); + +/** + * edac_mc_add_mc: Insert the 'mci' structure into the mci global list + * @mci: pointer to the mci structure to be added to the list + * + * Return: + * 0 Success + * !0 Failure + */ + +/* FIXME - should a warning be printed if no error detection? correction? */ +int edac_mc_add_mc(struct mem_ctl_info *mci) +{ + int rc = 1; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); +#ifdef CONFIG_EDAC_DEBUG + if (edac_debug_level >= 3) + edac_mc_dump_mci(mci); + if (edac_debug_level >= 4) { + int i; + + for (i = 0; i < mci->nr_csrows; i++) { + int j; + edac_mc_dump_csrow(&mci->csrows[i]); + for (j = 0; j < mci->csrows[i].nr_channels; j++) + edac_mc_dump_channel(&mci->csrows[i]. + channels[j]); + } + } +#endif + down(&mem_ctls_mutex); + + if (add_mc_to_global_list(mci)) + goto finish; + + /* set load time so that error rate can be tracked */ + mci->start_time = jiffies; + + if (edac_create_sysfs_mci_device(mci)) { + printk(KERN_WARNING + "EDAC MC%d: failed to create sysfs device\n", + mci->mc_idx); + /* FIXME - should there be an error code and unwind? */ + goto finish; + } + + /* Report action taken */ + printk(KERN_INFO + "EDAC MC%d: Giving out device to %s %s: PCI %s\n", + mci->mc_idx, mci->mod_name, mci->ctl_name, + pci_name(mci->pdev)); + + + rc = 0; + +finish: + up(&mem_ctls_mutex); + return rc; +} + + + +static void complete_mc_list_del (struct rcu_head *head) +{ + struct mem_ctl_info *mci; + + mci = container_of(head, struct mem_ctl_info, rcu); + INIT_LIST_HEAD(&mci->link); + complete(&mci->complete); +} + +static void del_mc_from_global_list (struct mem_ctl_info *mci) +{ + list_del_rcu(&mci->link); + init_completion(&mci->complete); + call_rcu(&mci->rcu, complete_mc_list_del); + wait_for_completion(&mci->complete); +} + +EXPORT_SYMBOL(edac_mc_del_mc); + +/** + * edac_mc_del_mc: Remove the specified mci structure from global list + * @mci: Pointer to struct mem_ctl_info structure + * + * Returns: + * 0 Success + * 1 Failure + */ +int edac_mc_del_mc(struct mem_ctl_info *mci) +{ + int rc = 1; + + debugf0("MC%d: " __FILE__ ": %s()\n", mci->mc_idx, __func__); + down(&mem_ctls_mutex); + del_mc_from_global_list(mci); + printk(KERN_INFO + "EDAC MC%d: Removed device %d for %s %s: PCI %s\n", + mci->mc_idx, mci->mc_idx, mci->mod_name, mci->ctl_name, + pci_name(mci->pdev)); + rc = 0; + up(&mem_ctls_mutex); + + return rc; +} + + +EXPORT_SYMBOL(edac_mc_scrub_block); + +void edac_mc_scrub_block(unsigned long page, unsigned long offset, + u32 size) +{ + struct page *pg; + void *virt_addr; + unsigned long flags = 0; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + /* ECC error page was not in our memory. Ignore it. */ + if(!pfn_valid(page)) + return; + + /* Find the actual page structure then map it and fix */ + pg = pfn_to_page(page); + + if (PageHighMem(pg)) + local_irq_save(flags); + + virt_addr = kmap_atomic(pg, KM_BOUNCE_READ); + + /* Perform architecture specific atomic scrub operation */ + atomic_scrub(virt_addr + offset, size); + + /* Unmap and complete */ + kunmap_atomic(virt_addr, KM_BOUNCE_READ); + + if (PageHighMem(pg)) + local_irq_restore(flags); +} + + +/* FIXME - should return -1 */ +EXPORT_SYMBOL(edac_mc_find_csrow_by_page); + +int edac_mc_find_csrow_by_page(struct mem_ctl_info *mci, + unsigned long page) +{ + struct csrow_info *csrows = mci->csrows; + int row, i; + + debugf1("MC%d: " __FILE__ ": %s(): 0x%lx\n", mci->mc_idx, __func__, + page); + row = -1; + + for (i = 0; i < mci->nr_csrows; i++) { + struct csrow_info *csrow = &csrows[i]; + + if (csrow->nr_pages == 0) + continue; + + debugf3("MC%d: " __FILE__ + ": %s(): first(0x%lx) page(0x%lx)" + " last(0x%lx) mask(0x%lx)\n", mci->mc_idx, + __func__, csrow->first_page, page, + csrow->last_page, csrow->page_mask); + + if ((page >= csrow->first_page) && + (page <= csrow->last_page) && + ((page & csrow->page_mask) == + (csrow->first_page & csrow->page_mask))) { + row = i; + break; + } + } + + if (row == -1) + printk(KERN_ERR + "EDAC MC%d: could not look up page error address %lx\n", + mci->mc_idx, (unsigned long) page); + + return row; +} + + +EXPORT_SYMBOL(edac_mc_handle_ce); + +/* FIXME - setable log (warning/emerg) levels */ +/* FIXME - integrate with evlog: http://evlog.sourceforge.net/ */ +void edac_mc_handle_ce(struct mem_ctl_info *mci, + unsigned long page_frame_number, + unsigned long offset_in_page, + unsigned long syndrome, int row, int channel, + const char *msg) +{ + unsigned long remapped_page; + + debugf3("MC%d: " __FILE__ ": %s()\n", mci->mc_idx, __func__); + + /* FIXME - maybe make panic on INTERNAL ERROR an option */ + if (row >= mci->nr_csrows || row < 0) { + /* something is wrong */ + printk(KERN_ERR + "EDAC MC%d: INTERNAL ERROR: row out of range (%d >= %d)\n", + mci->mc_idx, row, mci->nr_csrows); + edac_mc_handle_ce_no_info(mci, "INTERNAL ERROR"); + return; + } + if (channel >= mci->csrows[row].nr_channels || channel < 0) { + /* something is wrong */ + printk(KERN_ERR + "EDAC MC%d: INTERNAL ERROR: channel out of range " + "(%d >= %d)\n", + mci->mc_idx, channel, mci->csrows[row].nr_channels); + edac_mc_handle_ce_no_info(mci, "INTERNAL ERROR"); + return; + } + + if (log_ce) + /* FIXME - put in DIMM location */ + printk(KERN_WARNING + "EDAC MC%d: CE page 0x%lx, offset 0x%lx," + " grain %d, syndrome 0x%lx, row %d, channel %d," + " label \"%s\": %s\n", mci->mc_idx, + page_frame_number, offset_in_page, + mci->csrows[row].grain, syndrome, row, channel, + mci->csrows[row].channels[channel].label, msg); + + mci->ce_count++; + mci->csrows[row].ce_count++; + mci->csrows[row].channels[channel].ce_count++; + + if (mci->scrub_mode & SCRUB_SW_SRC) { + /* + * Some MC's can remap memory so that it is still available + * at a different address when PCI devices map into memory. + * MC's that can't do this lose the memory where PCI devices + * are mapped. This mapping is MC dependant and so we call + * back into the MC driver for it to map the MC page to + * a physical (CPU) page which can then be mapped to a virtual + * page - which can then be scrubbed. + */ + remapped_page = mci->ctl_page_to_phys ? + mci->ctl_page_to_phys(mci, page_frame_number) : + page_frame_number; + + edac_mc_scrub_block(remapped_page, offset_in_page, + mci->csrows[row].grain); + } +} + + +EXPORT_SYMBOL(edac_mc_handle_ce_no_info); + +void edac_mc_handle_ce_no_info(struct mem_ctl_info *mci, + const char *msg) +{ + if (log_ce) + printk(KERN_WARNING + "EDAC MC%d: CE - no information available: %s\n", + mci->mc_idx, msg); + mci->ce_noinfo_count++; + mci->ce_count++; +} + + +EXPORT_SYMBOL(edac_mc_handle_ue); + +void edac_mc_handle_ue(struct mem_ctl_info *mci, + unsigned long page_frame_number, + unsigned long offset_in_page, int row, + const char *msg) +{ + int len = EDAC_MC_LABEL_LEN * 4; + char labels[len + 1]; + char *pos = labels; + int chan; + int chars; + + debugf3("MC%d: " __FILE__ ": %s()\n", mci->mc_idx, __func__); + + /* FIXME - maybe make panic on INTERNAL ERROR an option */ + if (row >= mci->nr_csrows || row < 0) { + /* something is wrong */ + printk(KERN_ERR + "EDAC MC%d: INTERNAL ERROR: row out of range (%d >= %d)\n", + mci->mc_idx, row, mci->nr_csrows); + edac_mc_handle_ue_no_info(mci, "INTERNAL ERROR"); + return; + } + + chars = snprintf(pos, len + 1, "%s", + mci->csrows[row].channels[0].label); + len -= chars; + pos += chars; + for (chan = 1; (chan < mci->csrows[row].nr_channels) && (len > 0); + chan++) { + chars = snprintf(pos, len + 1, ":%s", + mci->csrows[row].channels[chan].label); + len -= chars; + pos += chars; + } + + if (log_ue) + printk(KERN_EMERG + "EDAC MC%d: UE page 0x%lx, offset 0x%lx, grain %d, row %d," + " labels \"%s\": %s\n", mci->mc_idx, + page_frame_number, offset_in_page, + mci->csrows[row].grain, row, labels, msg); + + if (panic_on_ue) + panic + ("EDAC MC%d: UE page 0x%lx, offset 0x%lx, grain %d, row %d," + " labels \"%s\": %s\n", mci->mc_idx, + page_frame_number, offset_in_page, + mci->csrows[row].grain, row, labels, msg); + + mci->ue_count++; + mci->csrows[row].ue_count++; +} + + +EXPORT_SYMBOL(edac_mc_handle_ue_no_info); + +void edac_mc_handle_ue_no_info(struct mem_ctl_info *mci, + const char *msg) +{ + if (panic_on_ue) + panic("EDAC MC%d: Uncorrected Error", mci->mc_idx); + + if (log_ue) + printk(KERN_WARNING + "EDAC MC%d: UE - no information available: %s\n", + mci->mc_idx, msg); + mci->ue_noinfo_count++; + mci->ue_count++; +} + + +#ifdef CONFIG_PCI + +static u16 get_pci_parity_status(struct pci_dev *dev, int secondary) +{ + int where; + u16 status; + + where = secondary ? PCI_SEC_STATUS : PCI_STATUS; + pci_read_config_word(dev, where, &status); + + /* If we get back 0xFFFF then we must suspect that the card has been pulled but + the Linux PCI layer has not yet finished cleaning up. We don't want to report + on such devices */ + + if (status == 0xFFFF) { + u32 sanity; + pci_read_config_dword(dev, 0, &sanity); + if (sanity == 0xFFFFFFFF) + return 0; + } + status &= PCI_STATUS_DETECTED_PARITY | PCI_STATUS_SIG_SYSTEM_ERROR | + PCI_STATUS_PARITY; + + if (status) + /* reset only the bits we are interested in */ + pci_write_config_word(dev, where, status); + + return status; +} + +typedef void (*pci_parity_check_fn_t) (struct pci_dev *dev); + +/* Clear any PCI parity errors logged by this device. */ +static void edac_pci_dev_parity_clear( struct pci_dev *dev ) +{ + u8 header_type; + + get_pci_parity_status(dev, 0); + + /* read the device TYPE, looking for bridges */ + pci_read_config_byte(dev, PCI_HEADER_TYPE, &header_type); + + if ((header_type & 0x7F) == PCI_HEADER_TYPE_BRIDGE) + get_pci_parity_status(dev, 1); +} + +/* + * PCI Parity polling + * + */ +static void edac_pci_dev_parity_test(struct pci_dev *dev) +{ + u16 status; + u8 header_type; + + /* read the STATUS register on this device + */ + status = get_pci_parity_status(dev, 0); + + debugf2("PCI STATUS= 0x%04x %s\n", status, dev->dev.bus_id ); + + /* check the status reg for errors */ + if (status) { + if (status & (PCI_STATUS_SIG_SYSTEM_ERROR)) + printk(KERN_CRIT + "EDAC PCI- " + "Signaled System Error on %s\n", + pci_name (dev)); + + if (status & (PCI_STATUS_PARITY)) { + printk(KERN_CRIT + "EDAC PCI- " + "Master Data Parity Error on %s\n", + pci_name (dev)); + + atomic_inc(&pci_parity_count); + } + + if (status & (PCI_STATUS_DETECTED_PARITY)) { + printk(KERN_CRIT + "EDAC PCI- " + "Detected Parity Error on %s\n", + pci_name (dev)); + + atomic_inc(&pci_parity_count); + } + } + + /* read the device TYPE, looking for bridges */ + pci_read_config_byte(dev, PCI_HEADER_TYPE, &header_type); + + debugf2("PCI HEADER TYPE= 0x%02x %s\n", header_type, dev->dev.bus_id ); + + if ((header_type & 0x7F) == PCI_HEADER_TYPE_BRIDGE) { + /* On bridges, need to examine secondary status register */ + status = get_pci_parity_status(dev, 1); + + debugf2("PCI SEC_STATUS= 0x%04x %s\n", + status, dev->dev.bus_id ); + + /* check the secondary status reg for errors */ + if (status) { + if (status & (PCI_STATUS_SIG_SYSTEM_ERROR)) + printk(KERN_CRIT + "EDAC PCI-Bridge- " + "Signaled System Error on %s\n", + pci_name (dev)); + + if (status & (PCI_STATUS_PARITY)) { + printk(KERN_CRIT + "EDAC PCI-Bridge- " + "Master Data Parity Error on %s\n", + pci_name (dev)); + + atomic_inc(&pci_parity_count); + } + + if (status & (PCI_STATUS_DETECTED_PARITY)) { + printk(KERN_CRIT + "EDAC PCI-Bridge- " + "Detected Parity Error on %s\n", + pci_name (dev)); + + atomic_inc(&pci_parity_count); + } + } + } +} + +/* + * check_dev_on_list: Scan for a PCI device on a white/black list + * @list: an EDAC &edac_pci_device_list white/black list pointer + * @free_index: index of next free entry on the list + * @pci_dev: PCI Device pointer + * + * see if list contains the device. + * + * Returns: 0 not found + * 1 found on list + */ +static int check_dev_on_list(struct edac_pci_device_list *list, int free_index, + struct pci_dev *dev) +{ + int i; + int rc = 0; /* Assume not found */ + unsigned short vendor=dev->vendor; + unsigned short device=dev->device; + + /* Scan the list, looking for a vendor/device match + */ + for (i = 0; i < free_index; i++, list++ ) { + if ( (list->vendor == vendor ) && + (list->device == device )) { + rc = 1; + break; + } + } + + return rc; +} + +/* + * pci_dev parity list iterator + * Scan the PCI device list for one iteration, looking for SERRORs + * Master Parity ERRORS or Parity ERRORs on primary or secondary devices + */ +static inline void edac_pci_dev_parity_iterator(pci_parity_check_fn_t fn) +{ + struct pci_dev *dev=NULL; + + /* request for kernel access to the next PCI device, if any, + * and while we are looking at it have its reference count + * bumped until we are done with it + */ + while((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { + + /* if whitelist exists then it has priority, so only scan those + * devices on the whitelist + */ + if (pci_whitelist_count > 0 ) { + if (check_dev_on_list(pci_whitelist, + pci_whitelist_count, dev)) + fn(dev); + } else { + /* + * if no whitelist, then check if this devices is + * blacklisted + */ + if (!check_dev_on_list(pci_blacklist, + pci_blacklist_count, dev)) + fn(dev); + } + } +} + +static void do_pci_parity_check(void) +{ + unsigned long flags; + int before_count; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + if (!check_pci_parity) + return; + + before_count = atomic_read(&pci_parity_count); + + /* scan all PCI devices looking for a Parity Error on devices and + * bridges + */ + local_irq_save(flags); + edac_pci_dev_parity_iterator(edac_pci_dev_parity_test); + local_irq_restore(flags); + + /* Only if operator has selected panic on PCI Error */ + if (panic_on_pci_parity) { + /* If the count is different 'after' from 'before' */ + if (before_count != atomic_read(&pci_parity_count)) + panic("EDAC: PCI Parity Error"); + } +} + + +static inline void clear_pci_parity_errors(void) +{ + /* Clear any PCI bus parity errors that devices initially have logged + * in their registers. + */ + edac_pci_dev_parity_iterator(edac_pci_dev_parity_clear); +} + + +#else /* CONFIG_PCI */ + + +static inline void do_pci_parity_check(void) +{ + /* no-op */ +} + + +static inline void clear_pci_parity_errors(void) +{ + /* no-op */ +} + + +#endif /* CONFIG_PCI */ + +/* + * Iterate over all MC instances and check for ECC, et al, errors + */ +static inline void check_mc_devices (void) +{ + unsigned long flags; + struct list_head *item; + struct mem_ctl_info *mci; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + /* during poll, have interrupts off */ + local_irq_save(flags); + + list_for_each(item, &mc_devices) { + mci = list_entry(item, struct mem_ctl_info, link); + + if (mci->edac_check != NULL) + mci->edac_check(mci); + } + + local_irq_restore(flags); +} + + +/* + * Check MC status every poll_msec. + * Check PCI status every poll_msec as well. + * + * This where the work gets done for edac. + * + * SMP safe, doesn't use NMI, and auto-rate-limits. + */ +static void do_edac_check(void) +{ + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + check_mc_devices(); + + do_pci_parity_check(); +} + + +/* + * EDAC thread state information + */ +struct bs_thread_info +{ + struct task_struct *task; + struct completion *event; + char *name; + void (*run)(void); +}; + +static struct bs_thread_info bs_thread; + +/* + * edac_kernel_thread + * This the kernel thread that processes edac operations + * in a normal thread environment + */ +static int edac_kernel_thread(void *arg) +{ + struct bs_thread_info *thread = (struct bs_thread_info *) arg; + + /* detach thread */ + daemonize(thread->name); + + current->exit_signal = SIGCHLD; + allow_signal(SIGKILL); + thread->task = current; + + /* indicate to starting task we have started */ + complete(thread->event); + + /* loop forever, until we are told to stop */ + while(thread->run != NULL) { + void (*run)(void); + + /* call the function to check the memory controllers */ + run = thread->run; + if (run) + run(); + + if (signal_pending(current)) + flush_signals(current); + + /* ensure we are interruptable */ + set_current_state(TASK_INTERRUPTIBLE); + + /* goto sleep for the interval */ + schedule_timeout((HZ * poll_msec) / 1000); + try_to_freeze(); + } + + /* notify waiter that we are exiting */ + complete(thread->event); + + return 0; +} + +/* + * edac_mc_init + * module initialization entry point + */ +static int __init edac_mc_init(void) +{ + int ret; + struct completion event; + + printk(KERN_INFO "MC: " __FILE__ " version " EDAC_MC_VERSION "\n"); + + /* + * Harvest and clear any boot/initialization PCI parity errors + * + * FIXME: This only clears errors logged by devices present at time of + * module initialization. We should also do an initial clear + * of each newly hotplugged device. + */ + clear_pci_parity_errors(); + + /* perform check for first time to harvest boot leftovers */ + do_edac_check(); + + /* Create the MC sysfs entires */ + if (edac_sysfs_memctrl_setup()) { + printk(KERN_ERR "EDAC MC: Error initializing sysfs code\n"); + return -ENODEV; + } + + /* Create the PCI parity sysfs entries */ + if (edac_sysfs_pci_setup()) { + edac_sysfs_memctrl_teardown(); + printk(KERN_ERR "EDAC PCI: Error initializing sysfs code\n"); + return -ENODEV; + } + + /* Create our kernel thread */ + init_completion(&event); + bs_thread.event = &event; + bs_thread.name = "kedac"; + bs_thread.run = do_edac_check; + + /* create our kernel thread */ + ret = kernel_thread(edac_kernel_thread, &bs_thread, CLONE_KERNEL); + if (ret < 0) { + /* remove the sysfs entries */ + edac_sysfs_memctrl_teardown(); + edac_sysfs_pci_teardown(); + return -ENOMEM; + } + + /* wait for our kernel theard ack that it is up and running */ + wait_for_completion(&event); + + return 0; +} + + +/* + * edac_mc_exit() + * module exit/termination functioni + */ +static void __exit edac_mc_exit(void) +{ + struct completion event; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + init_completion(&event); + bs_thread.event = &event; + + /* As soon as ->run is set to NULL, the task could disappear, + * so we need to hold tasklist_lock until we have sent the signal + */ + read_lock(&tasklist_lock); + bs_thread.run = NULL; + send_sig(SIGKILL, bs_thread.task, 1); + read_unlock(&tasklist_lock); + wait_for_completion(&event); + + /* tear down the sysfs device */ + edac_sysfs_memctrl_teardown(); + edac_sysfs_pci_teardown(); +} + + + + +module_init(edac_mc_init); +module_exit(edac_mc_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Linux Networx (http://lnxi.com) Thayne Harbaugh et al\n" + "Based on.work by Dan Hollis et al"); +MODULE_DESCRIPTION("Core library routines for MC reporting"); + +module_param(panic_on_ue, int, 0644); +MODULE_PARM_DESC(panic_on_ue, "Panic on uncorrected error: 0=off 1=on"); +module_param(check_pci_parity, int, 0644); +MODULE_PARM_DESC(check_pci_parity, "Check for PCI bus parity errors: 0=off 1=on"); +module_param(panic_on_pci_parity, int, 0644); +MODULE_PARM_DESC(panic_on_pci_parity, "Panic on PCI Bus Parity error: 0=off 1=on"); +module_param(log_ue, int, 0644); +MODULE_PARM_DESC(log_ue, "Log uncorrectable error to console: 0=off 1=on"); +module_param(log_ce, int, 0644); +MODULE_PARM_DESC(log_ce, "Log correctable error to console: 0=off 1=on"); +module_param(poll_msec, int, 0644); +MODULE_PARM_DESC(poll_msec, "Polling period in milliseconds"); +#ifdef CONFIG_EDAC_DEBUG +module_param(edac_debug_level, int, 0644); +MODULE_PARM_DESC(edac_debug_level, "Debug level"); +#endif diff --git a/drivers/edac/edac_mc.h b/drivers/edac/edac_mc.h new file mode 100644 index 00000000000..75ecf484a43 --- /dev/null +++ b/drivers/edac/edac_mc.h @@ -0,0 +1,448 @@ +/* + * MC kernel module + * (C) 2003 Linux Networx (http://lnxi.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * Written by Thayne Harbaugh + * Based on work by Dan Hollis <goemon at anime dot net> and others. + * http://www.anime.net/~goemon/linux-ecc/ + * + * NMI handling support added by + * Dave Peterson <dsp@llnl.gov> <dave_peterson@pobox.com> + * + * $Id: edac_mc.h,v 1.4.2.10 2005/10/05 00:43:44 dsp_llnl Exp $ + * + */ + + +#ifndef _EDAC_MC_H_ +#define _EDAC_MC_H_ + + +#include <linux/config.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/module.h> +#include <linux/spinlock.h> +#include <linux/smp.h> +#include <linux/pci.h> +#include <linux/time.h> +#include <linux/nmi.h> +#include <linux/rcupdate.h> +#include <linux/completion.h> +#include <linux/kobject.h> + + +#define EDAC_MC_LABEL_LEN 31 +#define MC_PROC_NAME_MAX_LEN 7 + +#if PAGE_SHIFT < 20 +#define PAGES_TO_MiB( pages ) ( ( pages ) >> ( 20 - PAGE_SHIFT ) ) +#else /* PAGE_SHIFT > 20 */ +#define PAGES_TO_MiB( pages ) ( ( pages ) << ( PAGE_SHIFT - 20 ) ) +#endif + +#ifdef CONFIG_EDAC_DEBUG +extern int edac_debug_level; +#define edac_debug_printk(level, fmt, args...) \ +do { if (level <= edac_debug_level) printk(KERN_DEBUG fmt, ##args); } while(0) +#define debugf0( ... ) edac_debug_printk(0, __VA_ARGS__ ) +#define debugf1( ... ) edac_debug_printk(1, __VA_ARGS__ ) +#define debugf2( ... ) edac_debug_printk(2, __VA_ARGS__ ) +#define debugf3( ... ) edac_debug_printk(3, __VA_ARGS__ ) +#define debugf4( ... ) edac_debug_printk(4, __VA_ARGS__ ) +#else /* !CONFIG_EDAC_DEBUG */ +#define debugf0( ... ) +#define debugf1( ... ) +#define debugf2( ... ) +#define debugf3( ... ) +#define debugf4( ... ) +#endif /* !CONFIG_EDAC_DEBUG */ + + +#define bs_xstr(s) bs_str(s) +#define bs_str(s) #s +#define BS_MOD_STR bs_xstr(KBUILD_BASENAME) + +#define BIT(x) (1 << (x)) + +#define PCI_VEND_DEV(vend, dev) PCI_VENDOR_ID_ ## vend, PCI_DEVICE_ID_ ## vend ## _ ## dev + +/* memory devices */ +enum dev_type { + DEV_UNKNOWN = 0, + DEV_X1, + DEV_X2, + DEV_X4, + DEV_X8, + DEV_X16, + DEV_X32, /* Do these parts exist? */ + DEV_X64 /* Do these parts exist? */ +}; + +#define DEV_FLAG_UNKNOWN BIT(DEV_UNKNOWN) +#define DEV_FLAG_X1 BIT(DEV_X1) +#define DEV_FLAG_X2 BIT(DEV_X2) +#define DEV_FLAG_X4 BIT(DEV_X4) +#define DEV_FLAG_X8 BIT(DEV_X8) +#define DEV_FLAG_X16 BIT(DEV_X16) +#define DEV_FLAG_X32 BIT(DEV_X32) +#define DEV_FLAG_X64 BIT(DEV_X64) + +/* memory types */ +enum mem_type { + MEM_EMPTY = 0, /* Empty csrow */ + MEM_RESERVED, /* Reserved csrow type */ + MEM_UNKNOWN, /* Unknown csrow type */ + MEM_FPM, /* Fast page mode */ + MEM_EDO, /* Extended data out */ + MEM_BEDO, /* Burst Extended data out */ + MEM_SDR, /* Single data rate SDRAM */ + MEM_RDR, /* Registered single data rate SDRAM */ + MEM_DDR, /* Double data rate SDRAM */ + MEM_RDDR, /* Registered Double data rate SDRAM */ + MEM_RMBS /* Rambus DRAM */ +}; + +#define MEM_FLAG_EMPTY BIT(MEM_EMPTY) +#define MEM_FLAG_RESERVED BIT(MEM_RESERVED) +#define MEM_FLAG_UNKNOWN BIT(MEM_UNKNOWN) +#define MEM_FLAG_FPM BIT(MEM_FPM) +#define MEM_FLAG_EDO BIT(MEM_EDO) +#define MEM_FLAG_BEDO BIT(MEM_BEDO) +#define MEM_FLAG_SDR BIT(MEM_SDR) +#define MEM_FLAG_RDR BIT(MEM_RDR) +#define MEM_FLAG_DDR BIT(MEM_DDR) +#define MEM_FLAG_RDDR BIT(MEM_RDDR) +#define MEM_FLAG_RMBS BIT(MEM_RMBS) + + +/* chipset Error Detection and Correction capabilities and mode */ +enum edac_type { + EDAC_UNKNOWN = 0, /* Unknown if ECC is available */ + EDAC_NONE, /* Doesnt support ECC */ + EDAC_RESERVED, /* Reserved ECC type */ + EDAC_PARITY, /* Detects parity errors */ + EDAC_EC, /* Error Checking - no correction */ + EDAC_SECDED, /* Single bit error correction, Double detection */ + EDAC_S2ECD2ED, /* Chipkill x2 devices - do these exist? */ + EDAC_S4ECD4ED, /* Chipkill x4 devices */ + EDAC_S8ECD8ED, /* Chipkill x8 devices */ + EDAC_S16ECD16ED, /* Chipkill x16 devices */ +}; + +#define EDAC_FLAG_UNKNOWN BIT(EDAC_UNKNOWN) +#define EDAC_FLAG_NONE BIT(EDAC_NONE) +#define EDAC_FLAG_PARITY BIT(EDAC_PARITY) +#define EDAC_FLAG_EC BIT(EDAC_EC) +#define EDAC_FLAG_SECDED BIT(EDAC_SECDED) +#define EDAC_FLAG_S2ECD2ED BIT(EDAC_S2ECD2ED) +#define EDAC_FLAG_S4ECD4ED BIT(EDAC_S4ECD4ED) +#define EDAC_FLAG_S8ECD8ED BIT(EDAC_S8ECD8ED) +#define EDAC_FLAG_S16ECD16ED BIT(EDAC_S16ECD16ED) + + +/* scrubbing capabilities */ +enum scrub_type { + SCRUB_UNKNOWN = 0, /* Unknown if scrubber is available */ + SCRUB_NONE, /* No scrubber */ + SCRUB_SW_PROG, /* SW progressive (sequential) scrubbing */ + SCRUB_SW_SRC, /* Software scrub only errors */ + SCRUB_SW_PROG_SRC, /* Progressive software scrub from an error */ + SCRUB_SW_TUNABLE, /* Software scrub frequency is tunable */ + SCRUB_HW_PROG, /* HW progressive (sequential) scrubbing */ + SCRUB_HW_SRC, /* Hardware scrub only errors */ + SCRUB_HW_PROG_SRC, /* Progressive hardware scrub from an error */ + SCRUB_HW_TUNABLE /* Hardware scrub frequency is tunable */ +}; + +#define SCRUB_FLAG_SW_PROG BIT(SCRUB_SW_PROG) +#define SCRUB_FLAG_SW_SRC BIT(SCRUB_SW_SRC_CORR) +#define SCRUB_FLAG_SW_PROG_SRC BIT(SCRUB_SW_PROG_SRC_CORR) +#define SCRUB_FLAG_SW_TUN BIT(SCRUB_SW_SCRUB_TUNABLE) +#define SCRUB_FLAG_HW_PROG BIT(SCRUB_HW_PROG) +#define SCRUB_FLAG_HW_SRC BIT(SCRUB_HW_SRC_CORR) +#define SCRUB_FLAG_HW_PROG_SRC BIT(SCRUB_HW_PROG_SRC_CORR) +#define SCRUB_FLAG_HW_TUN BIT(SCRUB_HW_TUNABLE) + +enum mci_sysfs_status { + MCI_SYSFS_INACTIVE = 0, /* sysfs entries NOT registered */ + MCI_SYSFS_ACTIVE /* sysfs entries ARE registered */ +}; + +/* FIXME - should have notify capabilities: NMI, LOG, PROC, etc */ + +/* + * There are several things to be aware of that aren't at all obvious: + * + * + * SOCKETS, SOCKET SETS, BANKS, ROWS, CHIP-SELECT ROWS, CHANNELS, etc.. + * + * These are some of the many terms that are thrown about that don't always + * mean what people think they mean (Inconceivable!). In the interest of + * creating a common ground for discussion, terms and their definitions + * will be established. + * + * Memory devices: The individual chip on a memory stick. These devices + * commonly output 4 and 8 bits each. Grouping several + * of these in parallel provides 64 bits which is common + * for a memory stick. + * + * Memory Stick: A printed circuit board that agregates multiple + * memory devices in parallel. This is the atomic + * memory component that is purchaseable by Joe consumer + * and loaded into a memory socket. + * + * Socket: A physical connector on the motherboard that accepts + * a single memory stick. + * + * Channel: Set of memory devices on a memory stick that must be + * grouped in parallel with one or more additional + * channels from other memory sticks. This parallel + * grouping of the output from multiple channels are + * necessary for the smallest granularity of memory access. + * Some memory controllers are capable of single channel - + * which means that memory sticks can be loaded + * individually. Other memory controllers are only + * capable of dual channel - which means that memory + * sticks must be loaded as pairs (see "socket set"). + * + * Chip-select row: All of the memory devices that are selected together. + * for a single, minimum grain of memory access. + * This selects all of the parallel memory devices across + * all of the parallel channels. Common chip-select rows + * for single channel are 64 bits, for dual channel 128 + * bits. + * + * Single-Ranked stick: A Single-ranked stick has 1 chip-select row of memmory. + * Motherboards commonly drive two chip-select pins to + * a memory stick. A single-ranked stick, will occupy + * only one of those rows. The other will be unused. + * + * Double-Ranked stick: A double-ranked stick has two chip-select rows which + * access different sets of memory devices. The two + * rows cannot be accessed concurrently. + * + * Double-sided stick: DEPRECATED TERM, see Double-Ranked stick. + * A double-sided stick has two chip-select rows which + * access different sets of memory devices. The two + * rows cannot be accessed concurrently. "Double-sided" + * is irrespective of the memory devices being mounted + * on both sides of the memory stick. + * + * Socket set: All of the memory sticks that are required for for + * a single memory access or all of the memory sticks + * spanned by a chip-select row. A single socket set + * has two chip-select rows and if double-sided sticks + * are used these will occupy those chip-select rows. + * + * Bank: This term is avoided because it is unclear when + * needing to distinguish between chip-select rows and + * socket sets. + * + * Controller pages: + * + * Physical pages: + * + * Virtual pages: + * + * + * STRUCTURE ORGANIZATION AND CHOICES + * + * + * + * PS - I enjoyed writing all that about as much as you enjoyed reading it. + */ + + +struct channel_info { + int chan_idx; /* channel index */ + u32 ce_count; /* Correctable Errors for this CHANNEL */ + char label[EDAC_MC_LABEL_LEN + 1]; /* DIMM label on motherboard */ + struct csrow_info *csrow; /* the parent */ +}; + + +struct csrow_info { + unsigned long first_page; /* first page number in dimm */ + unsigned long last_page; /* last page number in dimm */ + unsigned long page_mask; /* used for interleaving - + 0UL for non intlv */ + u32 nr_pages; /* number of pages in csrow */ + u32 grain; /* granularity of reported error in bytes */ + int csrow_idx; /* the chip-select row */ + enum dev_type dtype; /* memory device type */ + u32 ue_count; /* Uncorrectable Errors for this csrow */ + u32 ce_count; /* Correctable Errors for this csrow */ + enum mem_type mtype; /* memory csrow type */ + enum edac_type edac_mode; /* EDAC mode for this csrow */ + struct mem_ctl_info *mci; /* the parent */ + + struct kobject kobj; /* sysfs kobject for this csrow */ + + /* FIXME the number of CHANNELs might need to become dynamic */ + u32 nr_channels; + struct channel_info *channels; +}; + + +struct mem_ctl_info { + struct list_head link; /* for global list of mem_ctl_info structs */ + unsigned long mtype_cap; /* memory types supported by mc */ + unsigned long edac_ctl_cap; /* Mem controller EDAC capabilities */ + unsigned long edac_cap; /* configuration capabilities - this is + closely related to edac_ctl_cap. The + difference is that the controller + may be capable of s4ecd4ed which would + be listed in edac_ctl_cap, but if + channels aren't capable of s4ecd4ed then the + edac_cap would not have that capability. */ + unsigned long scrub_cap; /* chipset scrub capabilities */ + enum scrub_type scrub_mode; /* current scrub mode */ + + enum mci_sysfs_status sysfs_active; /* status of sysfs */ + + /* pointer to edac checking routine */ + void (*edac_check) (struct mem_ctl_info * mci); + /* + * Remaps memory pages: controller pages to physical pages. + * For most MC's, this will be NULL. + */ + /* FIXME - why not send the phys page to begin with? */ + unsigned long (*ctl_page_to_phys) (struct mem_ctl_info * mci, + unsigned long page); + int mc_idx; + int nr_csrows; + struct csrow_info *csrows; + /* + * FIXME - what about controllers on other busses? - IDs must be + * unique. pdev pointer should be sufficiently unique, but + * BUS:SLOT.FUNC numbers may not be unique. + */ + struct pci_dev *pdev; + const char *mod_name; + const char *mod_ver; + const char *ctl_name; + char proc_name[MC_PROC_NAME_MAX_LEN + 1]; + void *pvt_info; + u32 ue_noinfo_count; /* Uncorrectable Errors w/o info */ + u32 ce_noinfo_count; /* Correctable Errors w/o info */ + u32 ue_count; /* Total Uncorrectable Errors for this MC */ + u32 ce_count; /* Total Correctable Errors for this MC */ + unsigned long start_time; /* mci load start time (in jiffies) */ + + /* this stuff is for safe removal of mc devices from global list while + * NMI handlers may be traversing list + */ + struct rcu_head rcu; + struct completion complete; + + /* edac sysfs device control */ + struct kobject edac_mci_kobj; +}; + + + +/* write all or some bits in a byte-register*/ +static inline void pci_write_bits8(struct pci_dev *pdev, int offset, + u8 value, u8 mask) +{ + if (mask != 0xff) { + u8 buf; + pci_read_config_byte(pdev, offset, &buf); + value &= mask; + buf &= ~mask; + value |= buf; + } + pci_write_config_byte(pdev, offset, value); +} + + +/* write all or some bits in a word-register*/ +static inline void pci_write_bits16(struct pci_dev *pdev, int offset, + u16 value, u16 mask) +{ + if (mask != 0xffff) { + u16 buf; + pci_read_config_word(pdev, offset, &buf); + value &= mask; + buf &= ~mask; + value |= buf; + } + pci_write_config_word(pdev, offset, value); +} + + +/* write all or some bits in a dword-register*/ +static inline void pci_write_bits32(struct pci_dev *pdev, int offset, + u32 value, u32 mask) +{ + if (mask != 0xffff) { + u32 buf; + pci_read_config_dword(pdev, offset, &buf); + value &= mask; + buf &= ~mask; + value |= buf; + } + pci_write_config_dword(pdev, offset, value); +} + + +#ifdef CONFIG_EDAC_DEBUG +void edac_mc_dump_channel(struct channel_info *chan); +void edac_mc_dump_mci(struct mem_ctl_info *mci); +void edac_mc_dump_csrow(struct csrow_info *csrow); +#endif /* CONFIG_EDAC_DEBUG */ + +extern int edac_mc_add_mc(struct mem_ctl_info *mci); +extern int edac_mc_del_mc(struct mem_ctl_info *mci); + +extern int edac_mc_find_csrow_by_page(struct mem_ctl_info *mci, + unsigned long page); + +extern struct mem_ctl_info *edac_mc_find_mci_by_pdev(struct pci_dev + *pdev); + +extern void edac_mc_scrub_block(unsigned long page, + unsigned long offset, u32 size); + +/* + * The no info errors are used when error overflows are reported. + * There are a limited number of error logging registers that can + * be exausted. When all registers are exhausted and an additional + * error occurs then an error overflow register records that an + * error occured and the type of error, but doesn't have any + * further information. The ce/ue versions make for cleaner + * reporting logic and function interface - reduces conditional + * statement clutter and extra function arguments. + */ +extern void edac_mc_handle_ce(struct mem_ctl_info *mci, + unsigned long page_frame_number, + unsigned long offset_in_page, + unsigned long syndrome, + int row, int channel, const char *msg); + +extern void edac_mc_handle_ce_no_info(struct mem_ctl_info *mci, + const char *msg); + +extern void edac_mc_handle_ue(struct mem_ctl_info *mci, + unsigned long page_frame_number, + unsigned long offset_in_page, + int row, const char *msg); + +extern void edac_mc_handle_ue_no_info(struct mem_ctl_info *mci, + const char *msg); + +/* + * This kmalloc's and initializes all the structures. + * Can't be used if all structures don't have the same lifetime. + */ +extern struct mem_ctl_info *edac_mc_alloc(unsigned sz_pvt, + unsigned nr_csrows, unsigned nr_chans); + +/* Free an mc previously allocated by edac_mc_alloc() */ +extern void edac_mc_free(struct mem_ctl_info *mci); + + +#endif /* _EDAC_MC_H_ */ diff --git a/drivers/edac/i82860_edac.c b/drivers/edac/i82860_edac.c new file mode 100644 index 00000000000..52596e75f9c --- /dev/null +++ b/drivers/edac/i82860_edac.c @@ -0,0 +1,299 @@ +/* + * Intel 82860 Memory Controller kernel module + * (C) 2005 Red Hat (http://www.redhat.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * Written by Ben Woodard <woodard@redhat.com> + * shamelessly copied from and based upon the edac_i82875 driver + * by Thayne Harbaugh of Linux Networx. (http://lnxi.com) + */ + + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/pci.h> +#include <linux/pci_ids.h> +#include <linux/slab.h> +#include "edac_mc.h" + + +#ifndef PCI_DEVICE_ID_INTEL_82860_0 +#define PCI_DEVICE_ID_INTEL_82860_0 0x2531 +#endif /* PCI_DEVICE_ID_INTEL_82860_0 */ + +#define I82860_MCHCFG 0x50 +#define I82860_GBA 0x60 +#define I82860_GBA_MASK 0x7FF +#define I82860_GBA_SHIFT 24 +#define I82860_ERRSTS 0xC8 +#define I82860_EAP 0xE4 +#define I82860_DERRCTL_STS 0xE2 + +enum i82860_chips { + I82860 = 0, +}; + +struct i82860_dev_info { + const char *ctl_name; +}; + +struct i82860_error_info { + u16 errsts; + u32 eap; + u16 derrsyn; + u16 errsts2; +}; + +static const struct i82860_dev_info i82860_devs[] = { + [I82860] = { + .ctl_name = "i82860"}, +}; + +static struct pci_dev *mci_pdev = NULL; /* init dev: in case that AGP code + has already registered driver */ + +static int i82860_registered = 1; + +static void i82860_get_error_info (struct mem_ctl_info *mci, + struct i82860_error_info *info) +{ + /* + * This is a mess because there is no atomic way to read all the + * registers at once and the registers can transition from CE being + * overwritten by UE. + */ + pci_read_config_word(mci->pdev, I82860_ERRSTS, &info->errsts); + pci_read_config_dword(mci->pdev, I82860_EAP, &info->eap); + pci_read_config_word(mci->pdev, I82860_DERRCTL_STS, &info->derrsyn); + pci_read_config_word(mci->pdev, I82860_ERRSTS, &info->errsts2); + + pci_write_bits16(mci->pdev, I82860_ERRSTS, 0x0003, 0x0003); + + /* + * If the error is the same for both reads then the first set of reads + * is valid. If there is a change then there is a CE no info and the + * second set of reads is valid and should be UE info. + */ + if (!(info->errsts2 & 0x0003)) + return; + if ((info->errsts ^ info->errsts2) & 0x0003) { + pci_read_config_dword(mci->pdev, I82860_EAP, &info->eap); + pci_read_config_word(mci->pdev, I82860_DERRCTL_STS, + &info->derrsyn); + } +} + +static int i82860_process_error_info (struct mem_ctl_info *mci, + struct i82860_error_info *info, int handle_errors) +{ + int row; + + if (!(info->errsts2 & 0x0003)) + return 0; + + if (!handle_errors) + return 1; + + if ((info->errsts ^ info->errsts2) & 0x0003) { + edac_mc_handle_ce_no_info(mci, "UE overwrote CE"); + info->errsts = info->errsts2; + } + + info->eap >>= PAGE_SHIFT; + row = edac_mc_find_csrow_by_page(mci, info->eap); + + if (info->errsts & 0x0002) + edac_mc_handle_ue(mci, info->eap, 0, row, "i82860 UE"); + else + edac_mc_handle_ce(mci, info->eap, 0, info->derrsyn, row, + 0, "i82860 UE"); + + return 1; +} + +static void i82860_check(struct mem_ctl_info *mci) +{ + struct i82860_error_info info; + + debugf1("MC%d: " __FILE__ ": %s()\n", mci->mc_idx, __func__); + i82860_get_error_info(mci, &info); + i82860_process_error_info(mci, &info, 1); +} + +static int i82860_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + int index; + struct mem_ctl_info *mci = NULL; + unsigned long last_cumul_size; + + u16 mchcfg_ddim; /* DRAM Data Integrity Mode 0=none,2=edac */ + + /* RDRAM has channels but these don't map onto the abstractions that + edac uses. + The device groups from the GRA registers seem to map reasonably + well onto the notion of a chip select row. + There are 16 GRA registers and since the name is associated with + the channel and the GRA registers map to physical devices so we are + going to make 1 channel for group. + */ + mci = edac_mc_alloc(0, 16, 1); + if (!mci) + return -ENOMEM; + + debugf3("MC: " __FILE__ ": %s(): init mci\n", __func__); + + mci->pdev = pdev; + mci->mtype_cap = MEM_FLAG_DDR; + + + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED; + /* I"m not sure about this but I think that all RDRAM is SECDED */ + mci->edac_cap = EDAC_FLAG_SECDED; + /* adjust FLAGS */ + + mci->mod_name = BS_MOD_STR; + mci->mod_ver = "$Revision: 1.1.2.6 $"; + mci->ctl_name = i82860_devs[dev_idx].ctl_name; + mci->edac_check = i82860_check; + mci->ctl_page_to_phys = NULL; + + pci_read_config_word(mci->pdev, I82860_MCHCFG, &mchcfg_ddim); + mchcfg_ddim = mchcfg_ddim & 0x180; + + /* + * The group row boundary (GRA) reg values are boundary address + * for each DRAM row with a granularity of 16MB. GRA regs are + * cumulative; therefore GRA15 will contain the total memory contained + * in all eight rows. + */ + for (last_cumul_size = index = 0; index < mci->nr_csrows; index++) { + u16 value; + u32 cumul_size; + struct csrow_info *csrow = &mci->csrows[index]; + + pci_read_config_word(mci->pdev, I82860_GBA + index * 2, + &value); + + cumul_size = (value & I82860_GBA_MASK) << + (I82860_GBA_SHIFT - PAGE_SHIFT); + debugf3("MC: " __FILE__ ": %s(): (%d) cumul_size 0x%x\n", + __func__, index, cumul_size); + if (cumul_size == last_cumul_size) + continue; /* not populated */ + + csrow->first_page = last_cumul_size; + csrow->last_page = cumul_size - 1; + csrow->nr_pages = cumul_size - last_cumul_size; + last_cumul_size = cumul_size; + csrow->grain = 1 << 12; /* I82860_EAP has 4KiB reolution */ + csrow->mtype = MEM_RMBS; + csrow->dtype = DEV_UNKNOWN; + csrow->edac_mode = mchcfg_ddim ? EDAC_SECDED : EDAC_NONE; + } + + /* clear counters */ + pci_write_bits16(mci->pdev, I82860_ERRSTS, 0x0003, 0x0003); + + if (edac_mc_add_mc(mci)) { + debugf3("MC: " __FILE__ + ": %s(): failed edac_mc_add_mc()\n", + __func__); + edac_mc_free(mci); + } else { + /* get this far and it's successful */ + debugf3("MC: " __FILE__ ": %s(): success\n", __func__); + rc = 0; + } + return rc; +} + +/* returns count (>= 0), or negative on error */ +static int __devinit i82860_init_one(struct pci_dev *pdev, + const struct pci_device_id *ent) +{ + int rc; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + printk(KERN_INFO "i82860 init one\n"); + if(pci_enable_device(pdev) < 0) + return -EIO; + rc = i82860_probe1(pdev, ent->driver_data); + if(rc == 0) + mci_pdev = pci_dev_get(pdev); + return rc; +} + +static void __devexit i82860_remove_one(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + + debugf0(__FILE__ ": %s()\n", __func__); + + mci = edac_mc_find_mci_by_pdev(pdev); + if ((mci != NULL) && (edac_mc_del_mc(mci) == 0)) + edac_mc_free(mci); +} + +static const struct pci_device_id i82860_pci_tbl[] __devinitdata = { + {PCI_VEND_DEV(INTEL, 82860_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + I82860}, + {0,} /* 0 terminated list. */ +}; + +MODULE_DEVICE_TABLE(pci, i82860_pci_tbl); + +static struct pci_driver i82860_driver = { + .name = BS_MOD_STR, + .probe = i82860_init_one, + .remove = __devexit_p(i82860_remove_one), + .id_table = i82860_pci_tbl, +}; + +static int __init i82860_init(void) +{ + int pci_rc; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + if ((pci_rc = pci_register_driver(&i82860_driver)) < 0) + return pci_rc; + + if (!mci_pdev) { + i82860_registered = 0; + mci_pdev = pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_82860_0, NULL); + if (mci_pdev == NULL) { + debugf0("860 pci_get_device fail\n"); + return -ENODEV; + } + pci_rc = i82860_init_one(mci_pdev, i82860_pci_tbl); + if (pci_rc < 0) { + debugf0("860 init fail\n"); + pci_dev_put(mci_pdev); + return -ENODEV; + } + } + return 0; +} + +static void __exit i82860_exit(void) +{ + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + pci_unregister_driver(&i82860_driver); + if (!i82860_registered) { + i82860_remove_one(mci_pdev); + pci_dev_put(mci_pdev); + } +} + +module_init(i82860_init); +module_exit(i82860_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR + ("Red Hat Inc. (http://www.redhat.com.com) Ben Woodard <woodard@redhat.com>"); +MODULE_DESCRIPTION("ECC support for Intel 82860 memory hub controllers"); diff --git a/drivers/edac/i82875p_edac.c b/drivers/edac/i82875p_edac.c new file mode 100644 index 00000000000..009c08fe5d6 --- /dev/null +++ b/drivers/edac/i82875p_edac.c @@ -0,0 +1,532 @@ +/* + * Intel D82875P Memory Controller kernel module + * (C) 2003 Linux Networx (http://lnxi.com) + * This file may be distributed under the terms of the + * GNU General Public License. + * + * Written by Thayne Harbaugh + * Contributors: + * Wang Zhenyu at intel.com + * + * $Id: edac_i82875p.c,v 1.5.2.11 2005/10/05 00:43:44 dsp_llnl Exp $ + * + * Note: E7210 appears same as D82875P - zhenyu.z.wang at intel.com + */ + + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/init.h> + +#include <linux/pci.h> +#include <linux/pci_ids.h> + +#include <linux/slab.h> + +#include "edac_mc.h" + + +#ifndef PCI_DEVICE_ID_INTEL_82875_0 +#define PCI_DEVICE_ID_INTEL_82875_0 0x2578 +#endif /* PCI_DEVICE_ID_INTEL_82875_0 */ + +#ifndef PCI_DEVICE_ID_INTEL_82875_6 +#define PCI_DEVICE_ID_INTEL_82875_6 0x257e +#endif /* PCI_DEVICE_ID_INTEL_82875_6 */ + + +/* four csrows in dual channel, eight in single channel */ +#define I82875P_NR_CSROWS(nr_chans) (8/(nr_chans)) + + +/* Intel 82875p register addresses - device 0 function 0 - DRAM Controller */ +#define I82875P_EAP 0x58 /* Error Address Pointer (32b) + * + * 31:12 block address + * 11:0 reserved + */ + +#define I82875P_DERRSYN 0x5c /* DRAM Error Syndrome (8b) + * + * 7:0 DRAM ECC Syndrome + */ + +#define I82875P_DES 0x5d /* DRAM Error Status (8b) + * + * 7:1 reserved + * 0 Error channel 0/1 + */ + +#define I82875P_ERRSTS 0xc8 /* Error Status Register (16b) + * + * 15:10 reserved + * 9 non-DRAM lock error (ndlock) + * 8 Sftwr Generated SMI + * 7 ECC UE + * 6 reserved + * 5 MCH detects unimplemented cycle + * 4 AGP access outside GA + * 3 Invalid AGP access + * 2 Invalid GA translation table + * 1 Unsupported AGP command + * 0 ECC CE + */ + +#define I82875P_ERRCMD 0xca /* Error Command (16b) + * + * 15:10 reserved + * 9 SERR on non-DRAM lock + * 8 SERR on ECC UE + * 7 SERR on ECC CE + * 6 target abort on high exception + * 5 detect unimplemented cyc + * 4 AGP access outside of GA + * 3 SERR on invalid AGP access + * 2 invalid translation table + * 1 SERR on unsupported AGP command + * 0 reserved + */ + + +/* Intel 82875p register addresses - device 6 function 0 - DRAM Controller */ +#define I82875P_PCICMD6 0x04 /* PCI Command Register (16b) + * + * 15:10 reserved + * 9 fast back-to-back - ro 0 + * 8 SERR enable - ro 0 + * 7 addr/data stepping - ro 0 + * 6 parity err enable - ro 0 + * 5 VGA palette snoop - ro 0 + * 4 mem wr & invalidate - ro 0 + * 3 special cycle - ro 0 + * 2 bus master - ro 0 + * 1 mem access dev6 - 0(dis),1(en) + * 0 IO access dev3 - 0(dis),1(en) + */ + +#define I82875P_BAR6 0x10 /* Mem Delays Base ADDR Reg (32b) + * + * 31:12 mem base addr [31:12] + * 11:4 address mask - ro 0 + * 3 prefetchable - ro 0(non),1(pre) + * 2:1 mem type - ro 0 + * 0 mem space - ro 0 + */ + +/* Intel 82875p MMIO register space - device 0 function 0 - MMR space */ + +#define I82875P_DRB_SHIFT 26 /* 64MiB grain */ +#define I82875P_DRB 0x00 /* DRAM Row Boundary (8b x 8) + * + * 7 reserved + * 6:0 64MiB row boundary addr + */ + +#define I82875P_DRA 0x10 /* DRAM Row Attribute (4b x 8) + * + * 7 reserved + * 6:4 row attr row 1 + * 3 reserved + * 2:0 row attr row 0 + * + * 000 = 4KiB + * 001 = 8KiB + * 010 = 16KiB + * 011 = 32KiB + */ + +#define I82875P_DRC 0x68 /* DRAM Controller Mode (32b) + * + * 31:30 reserved + * 29 init complete + * 28:23 reserved + * 22:21 nr chan 00=1,01=2 + * 20 reserved + * 19:18 Data Integ Mode 00=none,01=ecc + * 17:11 reserved + * 10:8 refresh mode + * 7 reserved + * 6:4 mode select + * 3:2 reserved + * 1:0 DRAM type 01=DDR + */ + + +enum i82875p_chips { + I82875P = 0, +}; + + +struct i82875p_pvt { + struct pci_dev *ovrfl_pdev; + void *ovrfl_window; +}; + + +struct i82875p_dev_info { + const char *ctl_name; +}; + + +struct i82875p_error_info { + u16 errsts; + u32 eap; + u8 des; + u8 derrsyn; + u16 errsts2; +}; + + +static const struct i82875p_dev_info i82875p_devs[] = { + [I82875P] = { + .ctl_name = "i82875p"}, +}; + +static struct pci_dev *mci_pdev = NULL; /* init dev: in case that AGP code + has already registered driver */ +static int i82875p_registered = 1; + +static void i82875p_get_error_info (struct mem_ctl_info *mci, + struct i82875p_error_info *info) +{ + /* + * This is a mess because there is no atomic way to read all the + * registers at once and the registers can transition from CE being + * overwritten by UE. + */ + pci_read_config_word(mci->pdev, I82875P_ERRSTS, &info->errsts); + pci_read_config_dword(mci->pdev, I82875P_EAP, &info->eap); + pci_read_config_byte(mci->pdev, I82875P_DES, &info->des); + pci_read_config_byte(mci->pdev, I82875P_DERRSYN, &info->derrsyn); + pci_read_config_word(mci->pdev, I82875P_ERRSTS, &info->errsts2); + + pci_write_bits16(mci->pdev, I82875P_ERRSTS, 0x0081, 0x0081); + + /* + * If the error is the same then we can for both reads then + * the first set of reads is valid. If there is a change then + * there is a CE no info and the second set of reads is valid + * and should be UE info. + */ + if (!(info->errsts2 & 0x0081)) + return; + if ((info->errsts ^ info->errsts2) & 0x0081) { + pci_read_config_dword(mci->pdev, I82875P_EAP, &info->eap); + pci_read_config_byte(mci->pdev, I82875P_DES, &info->des); + pci_read_config_byte(mci->pdev, I82875P_DERRSYN, + &info->derrsyn); + } +} + +static int i82875p_process_error_info (struct mem_ctl_info *mci, + struct i82875p_error_info *info, int handle_errors) +{ + int row, multi_chan; + + multi_chan = mci->csrows[0].nr_channels - 1; + + if (!(info->errsts2 & 0x0081)) + return 0; + + if (!handle_errors) + return 1; + + if ((info->errsts ^ info->errsts2) & 0x0081) { + edac_mc_handle_ce_no_info(mci, "UE overwrote CE"); + info->errsts = info->errsts2; + } + + info->eap >>= PAGE_SHIFT; + row = edac_mc_find_csrow_by_page(mci, info->eap); + + if (info->errsts & 0x0080) + edac_mc_handle_ue(mci, info->eap, 0, row, "i82875p UE"); + else + edac_mc_handle_ce(mci, info->eap, 0, info->derrsyn, row, + multi_chan ? (info->des & 0x1) : 0, + "i82875p CE"); + + return 1; +} + + +static void i82875p_check(struct mem_ctl_info *mci) +{ + struct i82875p_error_info info; + + debugf1("MC%d: " __FILE__ ": %s()\n", mci->mc_idx, __func__); + i82875p_get_error_info(mci, &info); + i82875p_process_error_info(mci, &info, 1); +} + + +#ifdef CONFIG_PROC_FS +extern int pci_proc_attach_device(struct pci_dev *); +#endif + +static int i82875p_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + int index; + struct mem_ctl_info *mci = NULL; + struct i82875p_pvt *pvt = NULL; + unsigned long last_cumul_size; + struct pci_dev *ovrfl_pdev; + void __iomem *ovrfl_window = NULL; + + u32 drc; + u32 drc_chan; /* Number of channels 0=1chan,1=2chan */ + u32 nr_chans; + u32 drc_ddim; /* DRAM Data Integrity Mode 0=none,2=edac */ + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + ovrfl_pdev = pci_find_device(PCI_VEND_DEV(INTEL, 82875_6), NULL); + + if (!ovrfl_pdev) { + /* + * Intel tells BIOS developers to hide device 6 which + * configures the overflow device access containing + * the DRBs - this is where we expose device 6. + * http://www.x86-secret.com/articles/tweak/pat/patsecrets-2.htm + */ + pci_write_bits8(pdev, 0xf4, 0x2, 0x2); + ovrfl_pdev = + pci_scan_single_device(pdev->bus, PCI_DEVFN(6, 0)); + if (!ovrfl_pdev) + goto fail; + } +#ifdef CONFIG_PROC_FS + if (!ovrfl_pdev->procent && pci_proc_attach_device(ovrfl_pdev)) { + printk(KERN_ERR "MC: " __FILE__ + ": %s(): Failed to attach overflow device\n", + __func__); + goto fail; + } +#endif /* CONFIG_PROC_FS */ + if (pci_enable_device(ovrfl_pdev)) { + printk(KERN_ERR "MC: " __FILE__ + ": %s(): Failed to enable overflow device\n", + __func__); + goto fail; + } + + if (pci_request_regions(ovrfl_pdev, pci_name(ovrfl_pdev))) { +#ifdef CORRECT_BIOS + goto fail; +#endif + } + /* cache is irrelevant for PCI bus reads/writes */ + ovrfl_window = ioremap_nocache(pci_resource_start(ovrfl_pdev, 0), + pci_resource_len(ovrfl_pdev, 0)); + + if (!ovrfl_window) { + printk(KERN_ERR "MC: " __FILE__ + ": %s(): Failed to ioremap bar6\n", __func__); + goto fail; + } + + /* need to find out the number of channels */ + drc = readl(ovrfl_window + I82875P_DRC); + drc_chan = ((drc >> 21) & 0x1); + nr_chans = drc_chan + 1; + drc_ddim = (drc >> 18) & 0x1; + + mci = edac_mc_alloc(sizeof(*pvt), I82875P_NR_CSROWS(nr_chans), + nr_chans); + + if (!mci) { + rc = -ENOMEM; + goto fail; + } + + debugf3("MC: " __FILE__ ": %s(): init mci\n", __func__); + + mci->pdev = pdev; + mci->mtype_cap = MEM_FLAG_DDR; + + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED; + mci->edac_cap = EDAC_FLAG_UNKNOWN; + /* adjust FLAGS */ + + mci->mod_name = BS_MOD_STR; + mci->mod_ver = "$Revision: 1.5.2.11 $"; + mci->ctl_name = i82875p_devs[dev_idx].ctl_name; + mci->edac_check = i82875p_check; + mci->ctl_page_to_phys = NULL; + + debugf3("MC: " __FILE__ ": %s(): init pvt\n", __func__); + + pvt = (struct i82875p_pvt *) mci->pvt_info; + pvt->ovrfl_pdev = ovrfl_pdev; + pvt->ovrfl_window = ovrfl_window; + + /* + * The dram row boundary (DRB) reg values are boundary address + * for each DRAM row with a granularity of 32 or 64MB (single/dual + * channel operation). DRB regs are cumulative; therefore DRB7 will + * contain the total memory contained in all eight rows. + */ + for (last_cumul_size = index = 0; index < mci->nr_csrows; index++) { + u8 value; + u32 cumul_size; + struct csrow_info *csrow = &mci->csrows[index]; + + value = readb(ovrfl_window + I82875P_DRB + index); + cumul_size = value << (I82875P_DRB_SHIFT - PAGE_SHIFT); + debugf3("MC: " __FILE__ ": %s(): (%d) cumul_size 0x%x\n", + __func__, index, cumul_size); + if (cumul_size == last_cumul_size) + continue; /* not populated */ + + csrow->first_page = last_cumul_size; + csrow->last_page = cumul_size - 1; + csrow->nr_pages = cumul_size - last_cumul_size; + last_cumul_size = cumul_size; + csrow->grain = 1 << 12; /* I82875P_EAP has 4KiB reolution */ + csrow->mtype = MEM_DDR; + csrow->dtype = DEV_UNKNOWN; + csrow->edac_mode = drc_ddim ? EDAC_SECDED : EDAC_NONE; + } + + /* clear counters */ + pci_write_bits16(mci->pdev, I82875P_ERRSTS, 0x0081, 0x0081); + + if (edac_mc_add_mc(mci)) { + debugf3("MC: " __FILE__ + ": %s(): failed edac_mc_add_mc()\n", __func__); + goto fail; + } + + /* get this far and it's successful */ + debugf3("MC: " __FILE__ ": %s(): success\n", __func__); + return 0; + + fail: + if (mci) + edac_mc_free(mci); + + if (ovrfl_window) + iounmap(ovrfl_window); + + if (ovrfl_pdev) { + pci_release_regions(ovrfl_pdev); + pci_disable_device(ovrfl_pdev); + } + + /* NOTE: the ovrfl proc entry and pci_dev are intentionally left */ + return rc; +} + + +/* returns count (>= 0), or negative on error */ +static int __devinit i82875p_init_one(struct pci_dev *pdev, + const struct pci_device_id *ent) +{ + int rc; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + printk(KERN_INFO "i82875p init one\n"); + if(pci_enable_device(pdev) < 0) + return -EIO; + rc = i82875p_probe1(pdev, ent->driver_data); + if (mci_pdev == NULL) + mci_pdev = pci_dev_get(pdev); + return rc; +} + + +static void __devexit i82875p_remove_one(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + struct i82875p_pvt *pvt = NULL; + + debugf0(__FILE__ ": %s()\n", __func__); + + if ((mci = edac_mc_find_mci_by_pdev(pdev)) == NULL) + return; + + pvt = (struct i82875p_pvt *) mci->pvt_info; + if (pvt->ovrfl_window) + iounmap(pvt->ovrfl_window); + + if (pvt->ovrfl_pdev) { +#ifdef CORRECT_BIOS + pci_release_regions(pvt->ovrfl_pdev); +#endif /*CORRECT_BIOS */ + pci_disable_device(pvt->ovrfl_pdev); + pci_dev_put(pvt->ovrfl_pdev); + } + + if (edac_mc_del_mc(mci)) + return; + + edac_mc_free(mci); +} + + +static const struct pci_device_id i82875p_pci_tbl[] __devinitdata = { + {PCI_VEND_DEV(INTEL, 82875_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + I82875P}, + {0,} /* 0 terminated list. */ +}; + +MODULE_DEVICE_TABLE(pci, i82875p_pci_tbl); + + +static struct pci_driver i82875p_driver = { + .name = BS_MOD_STR, + .probe = i82875p_init_one, + .remove = __devexit_p(i82875p_remove_one), + .id_table = i82875p_pci_tbl, +}; + + +static int __init i82875p_init(void) +{ + int pci_rc; + + debugf3("MC: " __FILE__ ": %s()\n", __func__); + pci_rc = pci_register_driver(&i82875p_driver); + if (pci_rc < 0) + return pci_rc; + if (mci_pdev == NULL) { + i82875p_registered = 0; + mci_pdev = + pci_get_device(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_82875_0, NULL); + if (!mci_pdev) { + debugf0("875p pci_get_device fail\n"); + return -ENODEV; + } + pci_rc = i82875p_init_one(mci_pdev, i82875p_pci_tbl); + if (pci_rc < 0) { + debugf0("875p init fail\n"); + pci_dev_put(mci_pdev); + return -ENODEV; + } + } + return 0; +} + + +static void __exit i82875p_exit(void) +{ + debugf3("MC: " __FILE__ ": %s()\n", __func__); + + pci_unregister_driver(&i82875p_driver); + if (!i82875p_registered) { + i82875p_remove_one(mci_pdev); + pci_dev_put(mci_pdev); + } +} + + +module_init(i82875p_init); +module_exit(i82875p_exit); + + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Linux Networx (http://lnxi.com) Thayne Harbaugh"); +MODULE_DESCRIPTION("MC support for Intel 82875 memory hub controllers"); diff --git a/drivers/edac/r82600_edac.c b/drivers/edac/r82600_edac.c new file mode 100644 index 00000000000..e90892831b9 --- /dev/null +++ b/drivers/edac/r82600_edac.c @@ -0,0 +1,407 @@ +/* + * Radisys 82600 Embedded chipset Memory Controller kernel module + * (C) 2005 EADS Astrium + * This file may be distributed under the terms of the + * GNU General Public License. + * + * Written by Tim Small <tim@buttersideup.com>, based on work by Thayne + * Harbaugh, Dan Hollis <goemon at anime dot net> and others. + * + * $Id: edac_r82600.c,v 1.1.2.6 2005/10/05 00:43:44 dsp_llnl Exp $ + * + * Written with reference to 82600 High Integration Dual PCI System + * Controller Data Book: + * http://www.radisys.com/files/support_downloads/007-01277-0002.82600DataBook.pdf + * references to this document given in [] + */ + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/init.h> + +#include <linux/pci.h> +#include <linux/pci_ids.h> + +#include <linux/slab.h> + +#include "edac_mc.h" + +/* Radisys say "The 82600 integrates a main memory SDRAM controller that + * supports up to four banks of memory. The four banks can support a mix of + * sizes of 64 bit wide (72 bits with ECC) Synchronous DRAM (SDRAM) DIMMs, + * each of which can be any size from 16MB to 512MB. Both registered (control + * signals buffered) and unbuffered DIMM types are supported. Mixing of + * registered and unbuffered DIMMs as well as mixing of ECC and non-ECC DIMMs + * is not allowed. The 82600 SDRAM interface operates at the same frequency as + * the CPU bus, 66MHz, 100MHz or 133MHz." + */ + +#define R82600_NR_CSROWS 4 +#define R82600_NR_CHANS 1 +#define R82600_NR_DIMMS 4 + +#define R82600_BRIDGE_ID 0x8200 + +/* Radisys 82600 register addresses - device 0 function 0 - PCI bridge */ +#define R82600_DRAMC 0x57 /* Various SDRAM related control bits + * all bits are R/W + * + * 7 SDRAM ISA Hole Enable + * 6 Flash Page Mode Enable + * 5 ECC Enable: 1=ECC 0=noECC + * 4 DRAM DIMM Type: 1= + * 3 BIOS Alias Disable + * 2 SDRAM BIOS Flash Write Enable + * 1:0 SDRAM Refresh Rate: 00=Disabled + * 01=7.8usec (256Mbit SDRAMs) + * 10=15.6us 11=125usec + */ + +#define R82600_SDRAMC 0x76 /* "SDRAM Control Register" + * More SDRAM related control bits + * all bits are R/W + * + * 15:8 Reserved. + * + * 7:5 Special SDRAM Mode Select + * + * 4 Force ECC + * + * 1=Drive ECC bits to 0 during + * write cycles (i.e. ECC test mode) + * + * 0=Normal ECC functioning + * + * 3 Enhanced Paging Enable + * + * 2 CAS# Latency 0=3clks 1=2clks + * + * 1 RAS# to CAS# Delay 0=3 1=2 + * + * 0 RAS# Precharge 0=3 1=2 + */ + +#define R82600_EAP 0x80 /* ECC Error Address Pointer Register + * + * 31 Disable Hardware Scrubbing (RW) + * 0=Scrub on corrected read + * 1=Don't scrub on corrected read + * + * 30:12 Error Address Pointer (RO) + * Upper 19 bits of error address + * + * 11:4 Syndrome Bits (RO) + * + * 3 BSERR# on multibit error (RW) + * 1=enable 0=disable + * + * 2 NMI on Single Bit Eror (RW) + * 1=NMI triggered by SBE n.b. other + * prerequeists + * 0=NMI not triggered + * + * 1 MBE (R/WC) + * read 1=MBE at EAP (see above) + * read 0=no MBE, or SBE occurred first + * write 1=Clear MBE status (must also + * clear SBE) + * write 0=NOP + * + * 1 SBE (R/WC) + * read 1=SBE at EAP (see above) + * read 0=no SBE, or MBE occurred first + * write 1=Clear SBE status (must also + * clear MBE) + * write 0=NOP + */ + +#define R82600_DRBA 0x60 /* + 0x60..0x63 SDRAM Row Boundry Address + * Registers + * + * 7:0 Address lines 30:24 - upper limit of + * each row [p57] + */ + +struct r82600_error_info { + u32 eapr; +}; + + +static unsigned int disable_hardware_scrub = 0; + + +static void r82600_get_error_info (struct mem_ctl_info *mci, + struct r82600_error_info *info) +{ + pci_read_config_dword(mci->pdev, R82600_EAP, &info->eapr); + + if (info->eapr & BIT(0)) + /* Clear error to allow next error to be reported [p.62] */ + pci_write_bits32(mci->pdev, R82600_EAP, + ((u32) BIT(0) & (u32) BIT(1)), + ((u32) BIT(0) & (u32) BIT(1))); + + if (info->eapr & BIT(1)) + /* Clear error to allow next error to be reported [p.62] */ + pci_write_bits32(mci->pdev, R82600_EAP, + ((u32) BIT(0) & (u32) BIT(1)), + ((u32) BIT(0) & (u32) BIT(1))); +} + + +static int r82600_process_error_info (struct mem_ctl_info *mci, + struct r82600_error_info *info, int handle_errors) +{ + int error_found; + u32 eapaddr, page; + u32 syndrome; + + error_found = 0; + + /* bits 30:12 store the upper 19 bits of the 32 bit error address */ + eapaddr = ((info->eapr >> 12) & 0x7FFF) << 13; + /* Syndrome in bits 11:4 [p.62] */ + syndrome = (info->eapr >> 4) & 0xFF; + + /* the R82600 reports at less than page * + * granularity (upper 19 bits only) */ + page = eapaddr >> PAGE_SHIFT; + + if (info->eapr & BIT(0)) { /* CE? */ + error_found = 1; + + if (handle_errors) + edac_mc_handle_ce( + mci, page, 0, /* not avail */ + syndrome, + edac_mc_find_csrow_by_page(mci, page), + 0, /* channel */ + mci->ctl_name); + } + + if (info->eapr & BIT(1)) { /* UE? */ + error_found = 1; + + if (handle_errors) + /* 82600 doesn't give enough info */ + edac_mc_handle_ue(mci, page, 0, + edac_mc_find_csrow_by_page(mci, page), + mci->ctl_name); + } + + return error_found; +} + +static void r82600_check(struct mem_ctl_info *mci) +{ + struct r82600_error_info info; + + debugf1("MC%d: " __FILE__ ": %s()\n", mci->mc_idx, __func__); + r82600_get_error_info(mci, &info); + r82600_process_error_info(mci, &info, 1); +} + +static int r82600_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + int index; + struct mem_ctl_info *mci = NULL; + u8 dramcr; + u32 ecc_on; + u32 reg_sdram; + u32 eapr; + u32 scrub_disabled; + u32 sdram_refresh_rate; + u32 row_high_limit_last = 0; + u32 eap_init_bits; + + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + + pci_read_config_byte(pdev, R82600_DRAMC, &dramcr); + pci_read_config_dword(pdev, R82600_EAP, &eapr); + + ecc_on = dramcr & BIT(5); + reg_sdram = dramcr & BIT(4); + scrub_disabled = eapr & BIT(31); + sdram_refresh_rate = dramcr & (BIT(0) | BIT(1)); + + debugf2("MC: " __FILE__ ": %s(): sdram refresh rate = %#0x\n", + __func__, sdram_refresh_rate); + + debugf2("MC: " __FILE__ ": %s(): DRAMC register = %#0x\n", __func__, + dramcr); + + mci = edac_mc_alloc(0, R82600_NR_CSROWS, R82600_NR_CHANS); + + if (mci == NULL) { + rc = -ENOMEM; + goto fail; + } + + debugf0("MC: " __FILE__ ": %s(): mci = %p\n", __func__, mci); + + mci->pdev = pdev; + mci->mtype_cap = MEM_FLAG_RDDR | MEM_FLAG_DDR; + + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_EC | EDAC_FLAG_SECDED; + /* FIXME try to work out if the chip leads have been * + * used for COM2 instead on this board? [MA6?] MAYBE: */ + + /* On the R82600, the pins for memory bits 72:65 - i.e. the * + * EC bits are shared with the pins for COM2 (!), so if COM2 * + * is enabled, we assume COM2 is wired up, and thus no EDAC * + * is possible. */ + mci->edac_cap = EDAC_FLAG_NONE | EDAC_FLAG_EC | EDAC_FLAG_SECDED; + if (ecc_on) { + if (scrub_disabled) + debugf3("MC: " __FILE__ ": %s(): mci = %p - " + "Scrubbing disabled! EAP: %#0x\n", __func__, + mci, eapr); + } else + mci->edac_cap = EDAC_FLAG_NONE; + + mci->mod_name = BS_MOD_STR; + mci->mod_ver = "$Revision: 1.1.2.6 $"; + mci->ctl_name = "R82600"; + mci->edac_check = r82600_check; + mci->ctl_page_to_phys = NULL; + + for (index = 0; index < mci->nr_csrows; index++) { + struct csrow_info *csrow = &mci->csrows[index]; + u8 drbar; /* sDram Row Boundry Address Register */ + u32 row_high_limit; + u32 row_base; + + /* find the DRAM Chip Select Base address and mask */ + pci_read_config_byte(mci->pdev, R82600_DRBA + index, &drbar); + + debugf1("MC%d: " __FILE__ ": %s() Row=%d DRBA = %#0x\n", + mci->mc_idx, __func__, index, drbar); + + row_high_limit = ((u32) drbar << 24); +/* row_high_limit = ((u32)drbar << 24) | 0xffffffUL; */ + + debugf1("MC%d: " __FILE__ ": %s() Row=%d, " + "Boundry Address=%#0x, Last = %#0x \n", + mci->mc_idx, __func__, index, row_high_limit, + row_high_limit_last); + + /* Empty row [p.57] */ + if (row_high_limit == row_high_limit_last) + continue; + + row_base = row_high_limit_last; + + csrow->first_page = row_base >> PAGE_SHIFT; + csrow->last_page = (row_high_limit >> PAGE_SHIFT) - 1; + csrow->nr_pages = csrow->last_page - csrow->first_page + 1; + /* Error address is top 19 bits - so granularity is * + * 14 bits */ + csrow->grain = 1 << 14; + csrow->mtype = reg_sdram ? MEM_RDDR : MEM_DDR; + /* FIXME - check that this is unknowable with this chipset */ + csrow->dtype = DEV_UNKNOWN; + + /* Mode is global on 82600 */ + csrow->edac_mode = ecc_on ? EDAC_SECDED : EDAC_NONE; + row_high_limit_last = row_high_limit; + } + + /* clear counters */ + /* FIXME should we? */ + + if (edac_mc_add_mc(mci)) { + debugf3("MC: " __FILE__ + ": %s(): failed edac_mc_add_mc()\n", __func__); + goto fail; + } + + /* get this far and it's successful */ + + /* Clear error flags to allow next error to be reported [p.62] */ + /* Test systems seem to always have the UE flag raised on boot */ + + eap_init_bits = BIT(0) & BIT(1); + if (disable_hardware_scrub) { + eap_init_bits |= BIT(31); + debugf3("MC: " __FILE__ ": %s(): Disabling Hardware Scrub " + "(scrub on error)\n", __func__); + } + + pci_write_bits32(mci->pdev, R82600_EAP, eap_init_bits, + eap_init_bits); + + debugf3("MC: " __FILE__ ": %s(): success\n", __func__); + return 0; + +fail: + if (mci) + edac_mc_free(mci); + + return rc; +} + +/* returns count (>= 0), or negative on error */ +static int __devinit r82600_init_one(struct pci_dev *pdev, + const struct pci_device_id *ent) +{ + debugf0("MC: " __FILE__ ": %s()\n", __func__); + + /* don't need to call pci_device_enable() */ + return r82600_probe1(pdev, ent->driver_data); +} + + +static void __devexit r82600_remove_one(struct pci_dev *pdev) +{ + struct mem_ctl_info *mci; + + debugf0(__FILE__ ": %s()\n", __func__); + + if (((mci = edac_mc_find_mci_by_pdev(pdev)) != NULL) && + !edac_mc_del_mc(mci)) + edac_mc_free(mci); +} + + +static const struct pci_device_id r82600_pci_tbl[] __devinitdata = { + {PCI_DEVICE(PCI_VENDOR_ID_RADISYS, R82600_BRIDGE_ID)}, + {0,} /* 0 terminated list. */ +}; + +MODULE_DEVICE_TABLE(pci, r82600_pci_tbl); + + +static struct pci_driver r82600_driver = { + .name = BS_MOD_STR, + .probe = r82600_init_one, + .remove = __devexit_p(r82600_remove_one), + .id_table = r82600_pci_tbl, +}; + + +static int __init r82600_init(void) +{ + return pci_register_driver(&r82600_driver); +} + + +static void __exit r82600_exit(void) +{ + pci_unregister_driver(&r82600_driver); +} + + +module_init(r82600_init); +module_exit(r82600_exit); + + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Tim Small <tim@buttersideup.com> - WPAD Ltd. " + "on behalf of EADS Astrium"); +MODULE_DESCRIPTION("MC support for Radisys 82600 memory controllers"); + +module_param(disable_hardware_scrub, bool, 0644); +MODULE_PARM_DESC(disable_hardware_scrub, + "If set, disable the chipset's automatic scrub for CEs"); diff --git a/drivers/md/kcopyd.c b/drivers/md/kcopyd.c index ca99979c868..8b3515f394a 100644 --- a/drivers/md/kcopyd.c +++ b/drivers/md/kcopyd.c @@ -8,6 +8,7 @@ * completion notification. */ +#include <asm/types.h> #include <asm/atomic.h> #include <linux/blkdev.h> diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index 1421941487c..626508afe1b 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -7,6 +7,7 @@ menu "Network device support" config NETDEVICES depends on NET + default y if UML bool "Network device support" ---help--- You can say N here if you don't intend to connect your Linux box to @@ -1914,6 +1915,15 @@ config E1000_NAPI If in doubt, say N. +config E1000_DISABLE_PACKET_SPLIT + bool "Disable Packet Split for PCI express adapters" + depends on E1000 + help + Say Y here if you want to use the legacy receive path for PCI express + hadware. + + If in doubt, say N. + source "drivers/net/ixp2000/Kconfig" config MYRI_SBUS diff --git a/drivers/net/b44.c b/drivers/net/b44.c index 7aa49b974dc..df9d6e80c4f 100644 --- a/drivers/net/b44.c +++ b/drivers/net/b44.c @@ -2136,7 +2136,7 @@ static int __init b44_init(void) /* Setup paramaters for syncing RX/TX DMA descriptors */ dma_desc_align_mask = ~(dma_desc_align_size - 1); - dma_desc_sync_size = max(dma_desc_align_size, sizeof(struct dma_desc)); + dma_desc_sync_size = max_t(unsigned int, dma_desc_align_size, sizeof(struct dma_desc)); return pci_module_init(&b44_driver); } diff --git a/drivers/net/cassini.c b/drivers/net/cassini.c index 1f7ca453bb4..6e295fce5c6 100644 --- a/drivers/net/cassini.c +++ b/drivers/net/cassini.c @@ -335,6 +335,30 @@ static inline void cas_mask_intr(struct cas *cp) cas_disable_irq(cp, i); } +static inline void cas_buffer_init(cas_page_t *cp) +{ + struct page *page = cp->buffer; + atomic_set((atomic_t *)&page->lru.next, 1); +} + +static inline int cas_buffer_count(cas_page_t *cp) +{ + struct page *page = cp->buffer; + return atomic_read((atomic_t *)&page->lru.next); +} + +static inline void cas_buffer_inc(cas_page_t *cp) +{ + struct page *page = cp->buffer; + atomic_inc((atomic_t *)&page->lru.next); +} + +static inline void cas_buffer_dec(cas_page_t *cp) +{ + struct page *page = cp->buffer; + atomic_dec((atomic_t *)&page->lru.next); +} + static void cas_enable_irq(struct cas *cp, const int ring) { if (ring == 0) { /* all but TX_DONE */ @@ -472,6 +496,7 @@ static int cas_page_free(struct cas *cp, cas_page_t *page) { pci_unmap_page(cp->pdev, page->dma_addr, cp->page_size, PCI_DMA_FROMDEVICE); + cas_buffer_dec(page); __free_pages(page->buffer, cp->page_order); kfree(page); return 0; @@ -501,6 +526,7 @@ static cas_page_t *cas_page_alloc(struct cas *cp, const gfp_t flags) page->buffer = alloc_pages(flags, cp->page_order); if (!page->buffer) goto page_err; + cas_buffer_init(page); page->dma_addr = pci_map_page(cp->pdev, page->buffer, 0, cp->page_size, PCI_DMA_FROMDEVICE); return page; @@ -579,7 +605,7 @@ static void cas_spare_recover(struct cas *cp, const gfp_t flags) list_for_each_safe(elem, tmp, &list) { cas_page_t *page = list_entry(elem, cas_page_t, list); - if (page_count(page->buffer) > 1) + if (cas_buffer_count(page) > 1) continue; list_del(elem); @@ -1347,7 +1373,7 @@ static inline cas_page_t *cas_page_spare(struct cas *cp, const int index) cas_page_t *page = cp->rx_pages[1][index]; cas_page_t *new; - if (page_count(page->buffer) == 1) + if (cas_buffer_count(page) == 1) return page; new = cas_page_dequeue(cp); @@ -1367,7 +1393,7 @@ static cas_page_t *cas_page_swap(struct cas *cp, const int ring, cas_page_t **page1 = cp->rx_pages[1]; /* swap if buffer is in use */ - if (page_count(page0[index]->buffer) > 1) { + if (cas_buffer_count(page0[index]) > 1) { cas_page_t *new = cas_page_spare(cp, index); if (new) { page1[index] = page0[index]; @@ -1925,8 +1951,8 @@ static void cas_tx(struct net_device *dev, struct cas *cp, u64 compwb = le64_to_cpu(cp->init_block->tx_compwb); #endif if (netif_msg_intr(cp)) - printk(KERN_DEBUG "%s: tx interrupt, status: 0x%x, %lx\n", - cp->dev->name, status, compwb); + printk(KERN_DEBUG "%s: tx interrupt, status: 0x%x, %llx\n", + cp->dev->name, status, (unsigned long long)compwb); /* process all the rings */ for (ring = 0; ring < N_TX_RINGS; ring++) { #ifdef USE_TX_COMPWB @@ -2039,6 +2065,7 @@ static int cas_rx_process_pkt(struct cas *cp, struct cas_rx_comp *rxc, skb->len += hlen - swivel; get_page(page->buffer); + cas_buffer_inc(page); frag->page = page->buffer; frag->page_offset = off; frag->size = hlen - swivel; @@ -2063,6 +2090,7 @@ static int cas_rx_process_pkt(struct cas *cp, struct cas_rx_comp *rxc, frag++; get_page(page->buffer); + cas_buffer_inc(page); frag->page = page->buffer; frag->page_offset = 0; frag->size = hlen; @@ -2225,7 +2253,7 @@ static int cas_post_rxds_ringN(struct cas *cp, int ring, int num) released = 0; while (entry != last) { /* make a new buffer if it's still in use */ - if (page_count(page[entry]->buffer) > 1) { + if (cas_buffer_count(page[entry]) > 1) { cas_page_t *new = cas_page_dequeue(cp); if (!new) { /* let the timer know that we need to diff --git a/drivers/net/e100.c b/drivers/net/e100.c index 4726722a063..bf1fd2b98bf 100644 --- a/drivers/net/e100.c +++ b/drivers/net/e100.c @@ -1,25 +1,25 @@ /******************************************************************************* - + Copyright(c) 1999 - 2005 Intel Corporation. All rights reserved. - - This program is free software; you can redistribute it and/or modify it - under the terms of the GNU General Public License as published by the Free - Software Foundation; either version 2 of the License, or (at your option) + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the Free + Software Foundation; either version 2 of the License, or (at your option) any later version. - - This program is distributed in the hope that it will be useful, but WITHOUT - ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or - FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + + This program is distributed in the hope that it will be useful, but WITHOUT + ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. - + You should have received a copy of the GNU General Public License along with - this program; if not, write to the Free Software Foundation, Inc., 59 + this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - + The full GNU General Public License is included in this distribution in the file called LICENSE. - + Contact Information: Linux NICS <linux.nics@intel.com> Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497 @@ -160,7 +160,7 @@ #define DRV_NAME "e100" #define DRV_EXT "-NAPI" -#define DRV_VERSION "3.4.14-k4"DRV_EXT +#define DRV_VERSION "3.5.10-k2"DRV_EXT #define DRV_DESCRIPTION "Intel(R) PRO/100 Network Driver" #define DRV_COPYRIGHT "Copyright(c) 1999-2005 Intel Corporation" #define PFX DRV_NAME ": " @@ -320,7 +320,7 @@ enum cuc_dump { cuc_dump_complete = 0x0000A005, cuc_dump_reset_complete = 0x0000A007, }; - + enum port { software_reset = 0x0000, selftest = 0x0001, @@ -715,10 +715,10 @@ static u16 e100_eeprom_read(struct nic *nic, u16 *addr_len, u16 addr) ctrl = (cmd_addr_data & (1 << i)) ? eecs | eedi : eecs; writeb(ctrl, &nic->csr->eeprom_ctrl_lo); e100_write_flush(nic); udelay(4); - + writeb(ctrl | eesk, &nic->csr->eeprom_ctrl_lo); e100_write_flush(nic); udelay(4); - + /* Eeprom drives a dummy zero to EEDO after receiving * complete address. Use this to adjust addr_len. */ ctrl = readb(&nic->csr->eeprom_ctrl_lo); @@ -726,7 +726,7 @@ static u16 e100_eeprom_read(struct nic *nic, u16 *addr_len, u16 addr) *addr_len -= (i - 16); i = 17; } - + data = (data << 1) | (ctrl & eedo ? 1 : 0); } @@ -1170,7 +1170,7 @@ static void e100_configure(struct nic *nic, struct cb *cb, struct sk_buff *skb) 0x00000000, 0x00000000, 0x00000000, 0x00000000, \ } -static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) +static void e100_setup_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) { /* *INDENT-OFF* */ static struct { @@ -1213,13 +1213,13 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) * driver can change the algorithm. * * INTDELAY - This loads the dead-man timer with its inital value. -* When this timer expires the interrupt is asserted, and the +* When this timer expires the interrupt is asserted, and the * timer is reset each time a new packet is received. (see * BUNDLEMAX below to set the limit on number of chained packets) * The current default is 0x600 or 1536. Experiments show that * the value should probably stay within the 0x200 - 0x1000. * -* BUNDLEMAX - +* BUNDLEMAX - * This sets the maximum number of frames that will be bundled. In * some situations, such as the TCP windowing algorithm, it may be * better to limit the growth of the bundle size than let it go as @@ -1229,7 +1229,7 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) * an interrupt for every frame received. If you do not want to put * a limit on the bundle size, set this value to xFFFF. * -* BUNDLESMALL - +* BUNDLESMALL - * This contains a bit-mask describing the minimum size frame that * will be bundled. The default masks the lower 7 bits, which means * that any frame less than 128 bytes in length will not be bundled, @@ -1244,7 +1244,7 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) * * The current default is 0xFF80, which masks out the lower 7 bits. * This means that any frame which is x7F (127) bytes or smaller -* will cause an immediate interrupt. Because this value must be a +* will cause an immediate interrupt. Because this value must be a * bit mask, there are only a few valid values that can be used. To * turn this feature off, the driver can write the value xFFFF to the * lower word of this instruction (in the same way that the other @@ -1253,7 +1253,7 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) * standard Ethernet frames are <= 2047 bytes in length. *************************************************************************/ -/* if you wish to disable the ucode functionality, while maintaining the +/* if you wish to disable the ucode functionality, while maintaining the * workarounds it provides, set the following defines to: * BUNDLESMALL 0 * BUNDLEMAX 1 @@ -1284,12 +1284,46 @@ static void e100_load_ucode(struct nic *nic, struct cb *cb, struct sk_buff *skb) for (i = 0; i < UCODE_SIZE; i++) cb->u.ucode[i] = cpu_to_le32(ucode[i]); - cb->command = cpu_to_le16(cb_ucode); + cb->command = cpu_to_le16(cb_ucode | cb_el); return; } noloaducode: - cb->command = cpu_to_le16(cb_nop); + cb->command = cpu_to_le16(cb_nop | cb_el); +} + +static inline int e100_exec_cb_wait(struct nic *nic, struct sk_buff *skb, + void (*cb_prepare)(struct nic *, struct cb *, struct sk_buff *)) +{ + int err = 0, counter = 50; + struct cb *cb = nic->cb_to_clean; + + if ((err = e100_exec_cb(nic, NULL, e100_setup_ucode))) + DPRINTK(PROBE,ERR, "ucode cmd failed with error %d\n", err); + + /* must restart cuc */ + nic->cuc_cmd = cuc_start; + + /* wait for completion */ + e100_write_flush(nic); + udelay(10); + + /* wait for possibly (ouch) 500ms */ + while (!(cb->status & cpu_to_le16(cb_complete))) { + msleep(10); + if (!--counter) break; + } + + /* ack any interupts, something could have been set */ + writeb(~0, &nic->csr->scb.stat_ack); + + /* if the command failed, or is not OK, notify and return */ + if (!counter || !(cb->status & cpu_to_le16(cb_ok))) { + DPRINTK(PROBE,ERR, "ucode load failed\n"); + err = -EPERM; + } + + return err; } static void e100_setup_iaaddr(struct nic *nic, struct cb *cb, @@ -1357,13 +1391,13 @@ static int e100_phy_init(struct nic *nic) mdio_write(netdev, nic->mii.phy_id, MII_NSC_CONG, cong); } - if((nic->mac >= mac_82550_D102) || ((nic->flags & ich) && + if((nic->mac >= mac_82550_D102) || ((nic->flags & ich) && (mdio_read(netdev, nic->mii.phy_id, MII_TPISTATUS) & 0x8000))) { /* enable/disable MDI/MDI-X auto-switching. MDI/MDI-X auto-switching is disabled for 82551ER/QM chips */ if((nic->mac == mac_82551_E) || (nic->mac == mac_82551_F) || - (nic->mac == mac_82551_10) || (nic->mii.force_media) || - !(nic->eeprom[eeprom_cnfg_mdix] & eeprom_mdix_enabled)) + (nic->mac == mac_82551_10) || (nic->mii.force_media) || + !(nic->eeprom[eeprom_cnfg_mdix] & eeprom_mdix_enabled)) mdio_write(netdev, nic->mii.phy_id, MII_NCONFIG, 0); else mdio_write(netdev, nic->mii.phy_id, MII_NCONFIG, NCONFIG_AUTO_SWITCH); @@ -1388,7 +1422,7 @@ static int e100_hw_init(struct nic *nic) return err; if((err = e100_exec_cmd(nic, ruc_load_base, 0))) return err; - if((err = e100_exec_cb(nic, NULL, e100_load_ucode))) + if ((err = e100_exec_cb_wait(nic, NULL, e100_setup_ucode))) return err; if((err = e100_exec_cb(nic, NULL, e100_configure))) return err; @@ -1493,7 +1527,7 @@ static void e100_update_stats(struct nic *nic) } } - + if(e100_exec_cmd(nic, cuc_dump_reset, 0)) DPRINTK(TX_ERR, DEBUG, "exec cuc_dump_reset failed\n"); } @@ -1542,10 +1576,10 @@ static void e100_watchdog(unsigned long data) mii_check_link(&nic->mii); /* Software generated interrupt to recover from (rare) Rx - * allocation failure. - * Unfortunately have to use a spinlock to not re-enable interrupts - * accidentally, due to hardware that shares a register between the - * interrupt mask bit and the SW Interrupt generation bit */ + * allocation failure. + * Unfortunately have to use a spinlock to not re-enable interrupts + * accidentally, due to hardware that shares a register between the + * interrupt mask bit and the SW Interrupt generation bit */ spin_lock_irq(&nic->cmd_lock); writeb(readb(&nic->csr->scb.cmd_hi) | irq_sw_gen,&nic->csr->scb.cmd_hi); spin_unlock_irq(&nic->cmd_lock); @@ -1830,7 +1864,7 @@ static void e100_rx_clean(struct nic *nic, unsigned int *work_done, struct rx *rx_to_start = NULL; /* are we already rnr? then pay attention!!! this ensures that - * the state machine progression never allows a start with a + * the state machine progression never allows a start with a * partially cleaned list, avoiding a race between hardware * and rx_to_clean when in NAPI mode */ if(RU_SUSPENDED == nic->ru_running) @@ -2066,7 +2100,7 @@ static void e100_tx_timeout(struct net_device *netdev) { struct nic *nic = netdev_priv(netdev); - /* Reset outside of interrupt context, to avoid request_irq + /* Reset outside of interrupt context, to avoid request_irq * in interrupt context */ schedule_work(&nic->tx_timeout_task); } @@ -2313,7 +2347,7 @@ static int e100_set_ringparam(struct net_device *netdev, struct param_range *rfds = &nic->params.rfds; struct param_range *cbs = &nic->params.cbs; - if ((ring->rx_mini_pending) || (ring->rx_jumbo_pending)) + if ((ring->rx_mini_pending) || (ring->rx_jumbo_pending)) return -EINVAL; if(netif_running(netdev)) @@ -2631,7 +2665,9 @@ static int __devinit e100_probe(struct pci_dev *pdev, nic->flags |= wol_magic; /* ack any pending wake events, disable PME */ - pci_enable_wake(pdev, 0, 0); + err = pci_enable_wake(pdev, 0, 0); + if (err) + DPRINTK(PROBE, ERR, "Error clearing wake event\n"); strcpy(netdev->name, "eth%d"); if((err = register_netdev(netdev))) { @@ -2682,6 +2718,7 @@ static int e100_suspend(struct pci_dev *pdev, pm_message_t state) { struct net_device *netdev = pci_get_drvdata(pdev); struct nic *nic = netdev_priv(netdev); + int retval; if(netif_running(netdev)) e100_down(nic); @@ -2689,9 +2726,14 @@ static int e100_suspend(struct pci_dev *pdev, pm_message_t state) netif_device_detach(netdev); pci_save_state(pdev); - pci_enable_wake(pdev, pci_choose_state(pdev, state), nic->flags & (wol_magic | e100_asf(nic))); + retval = pci_enable_wake(pdev, pci_choose_state(pdev, state), + nic->flags & (wol_magic | e100_asf(nic))); + if (retval) + DPRINTK(PROBE,ERR, "Error enabling wake\n"); pci_disable_device(pdev); - pci_set_power_state(pdev, pci_choose_state(pdev, state)); + retval = pci_set_power_state(pdev, pci_choose_state(pdev, state)); + if (retval) + DPRINTK(PROBE,ERR, "Error %d setting power state\n", retval); return 0; } @@ -2700,11 +2742,16 @@ static int e100_resume(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); struct nic *nic = netdev_priv(netdev); + int retval; - pci_set_power_state(pdev, PCI_D0); + retval = pci_set_power_state(pdev, PCI_D0); + if (retval) + DPRINTK(PROBE,ERR, "Error waking adapter\n"); pci_restore_state(pdev); /* ack any pending wake events, disable PME */ - pci_enable_wake(pdev, 0, 0); + retval = pci_enable_wake(pdev, 0, 0); + if (retval) + DPRINTK(PROBE,ERR, "Error clearing wake events\n"); if(e100_hw_init(nic)) DPRINTK(HW, ERR, "e100_hw_init failed\n"); @@ -2721,12 +2768,15 @@ static void e100_shutdown(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); struct nic *nic = netdev_priv(netdev); + int retval; #ifdef CONFIG_PM - pci_enable_wake(pdev, 0, nic->flags & (wol_magic | e100_asf(nic))); + retval = pci_enable_wake(pdev, 0, nic->flags & (wol_magic | e100_asf(nic))); #else - pci_enable_wake(pdev, 0, nic->flags & (wol_magic)); + retval = pci_enable_wake(pdev, 0, nic->flags & (wol_magic)); #endif + if (retval) + DPRINTK(PROBE,ERR, "Error enabling wake\n"); } @@ -2739,7 +2789,7 @@ static struct pci_driver e100_driver = { .suspend = e100_suspend, .resume = e100_resume, #endif - .shutdown = e100_shutdown, + .shutdown = e100_shutdown, }; static int __init e100_init_module(void) diff --git a/drivers/net/e1000/e1000.h b/drivers/net/e1000/e1000.h index e02e9ba2e18..27c77306193 100644 --- a/drivers/net/e1000/e1000.h +++ b/drivers/net/e1000/e1000.h @@ -72,10 +72,6 @@ #include <linux/mii.h> #include <linux/ethtool.h> #include <linux/if_vlan.h> -#ifdef CONFIG_E1000_MQ -#include <linux/cpu.h> -#include <linux/smp.h> -#endif #define BAR_0 0 #define BAR_1 1 @@ -87,6 +83,10 @@ struct e1000_adapter; #include "e1000_hw.h" +#ifdef CONFIG_E1000_MQ +#include <linux/cpu.h> +#include <linux/smp.h> +#endif #ifdef DBG #define E1000_DBG(args...) printk(KERN_DEBUG "e1000: " args) @@ -169,6 +169,13 @@ struct e1000_buffer { uint16_t next_to_watch; }; +#ifdef CONFIG_E1000_MQ +struct e1000_queue_stats { + uint64_t packets; + uint64_t bytes; +}; +#endif + struct e1000_ps_page { struct page *ps_page[PS_PAGE_BUFFERS]; }; struct e1000_ps_page_dma { uint64_t ps_page_dma[PS_PAGE_BUFFERS]; }; @@ -191,10 +198,12 @@ struct e1000_tx_ring { spinlock_t tx_lock; uint16_t tdh; uint16_t tdt; - uint64_t pkt; boolean_t last_tx_tso; +#ifdef CONFIG_E1000_MQ + struct e1000_queue_stats tx_stats; +#endif }; struct e1000_rx_ring { @@ -216,9 +225,17 @@ struct e1000_rx_ring { struct e1000_ps_page *ps_page; struct e1000_ps_page_dma *ps_page_dma; + struct sk_buff *rx_skb_top; + struct sk_buff *rx_skb_prev; + + /* cpu for rx queue */ + int cpu; + uint16_t rdh; uint16_t rdt; - uint64_t pkt; +#ifdef CONFIG_E1000_MQ + struct e1000_queue_stats rx_stats; +#endif }; #define E1000_DESC_UNUSED(R) \ @@ -251,6 +268,9 @@ struct e1000_adapter { uint16_t link_speed; uint16_t link_duplex; spinlock_t stats_lock; +#ifdef CONFIG_E1000_NAPI + spinlock_t tx_queue_lock; +#endif atomic_t irq_sem; struct work_struct tx_timeout_task; struct work_struct watchdog_task; @@ -264,6 +284,7 @@ struct e1000_adapter { #ifdef CONFIG_E1000_MQ struct e1000_tx_ring **cpu_tx_ring; /* per-cpu */ #endif + unsigned long tx_queue_len; uint32_t txd_cmd; uint32_t tx_int_delay; uint32_t tx_abs_int_delay; @@ -271,9 +292,11 @@ struct e1000_adapter { uint64_t gotcl_old; uint64_t tpt_old; uint64_t colc_old; + uint32_t tx_timeout_count; uint32_t tx_fifo_head; uint32_t tx_head_addr; uint32_t tx_fifo_size; + uint8_t tx_timeout_factor; atomic_t tx_fifo_stall; boolean_t pcix_82544; boolean_t detect_tx_hung; @@ -281,14 +304,15 @@ struct e1000_adapter { /* RX */ #ifdef CONFIG_E1000_NAPI boolean_t (*clean_rx) (struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring, - int *work_done, int work_to_do); + struct e1000_rx_ring *rx_ring, + int *work_done, int work_to_do); #else boolean_t (*clean_rx) (struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring); #endif void (*alloc_rx_buf) (struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring, + int cleaned_count); struct e1000_rx_ring *rx_ring; /* One per active queue */ #ifdef CONFIG_E1000_NAPI struct net_device *polling_netdev; /* One per active queue */ @@ -296,13 +320,15 @@ struct e1000_adapter { #ifdef CONFIG_E1000_MQ struct net_device **cpu_netdev; /* per-cpu */ struct call_async_data_struct rx_sched_call_data; - int cpu_for_queue[4]; + cpumask_t cpumask; #endif - int num_queues; + int num_tx_queues; + int num_rx_queues; uint64_t hw_csum_err; uint64_t hw_csum_good; uint64_t rx_hdr_split; + uint32_t alloc_rx_buff_failed; uint32_t rx_int_delay; uint32_t rx_abs_int_delay; boolean_t rx_csum; @@ -330,6 +356,7 @@ struct e1000_adapter { struct e1000_rx_ring test_rx_ring; + u32 *config_space; int msg_enable; #ifdef CONFIG_PCI_MSI boolean_t have_msi; diff --git a/drivers/net/e1000/e1000_ethtool.c b/drivers/net/e1000/e1000_ethtool.c index c88f1a3c1b1..5cedc81786e 100644 --- a/drivers/net/e1000/e1000_ethtool.c +++ b/drivers/net/e1000/e1000_ethtool.c @@ -80,6 +80,7 @@ static const struct e1000_stats e1000_gstrings_stats[] = { { "tx_deferred_ok", E1000_STAT(stats.dc) }, { "tx_single_coll_ok", E1000_STAT(stats.scc) }, { "tx_multi_coll_ok", E1000_STAT(stats.mcc) }, + { "tx_timeout_count", E1000_STAT(tx_timeout_count) }, { "rx_long_length_errors", E1000_STAT(stats.roc) }, { "rx_short_length_errors", E1000_STAT(stats.ruc) }, { "rx_align_errors", E1000_STAT(stats.algnerrc) }, @@ -93,9 +94,20 @@ static const struct e1000_stats e1000_gstrings_stats[] = { { "rx_csum_offload_good", E1000_STAT(hw_csum_good) }, { "rx_csum_offload_errors", E1000_STAT(hw_csum_err) }, { "rx_header_split", E1000_STAT(rx_hdr_split) }, + { "alloc_rx_buff_failed", E1000_STAT(alloc_rx_buff_failed) }, }; -#define E1000_STATS_LEN \ + +#ifdef CONFIG_E1000_MQ +#define E1000_QUEUE_STATS_LEN \ + (((struct e1000_adapter *)netdev->priv)->num_tx_queues + \ + ((struct e1000_adapter *)netdev->priv)->num_rx_queues) \ + * (sizeof(struct e1000_queue_stats) / sizeof(uint64_t)) +#else +#define E1000_QUEUE_STATS_LEN 0 +#endif +#define E1000_GLOBAL_STATS_LEN \ sizeof(e1000_gstrings_stats) / sizeof(struct e1000_stats) +#define E1000_STATS_LEN (E1000_GLOBAL_STATS_LEN + E1000_QUEUE_STATS_LEN) static const char e1000_gstrings_test[][ETH_GSTRING_LEN] = { "Register test (offline)", "Eeprom test (offline)", "Interrupt test (offline)", "Loopback test (offline)", @@ -109,7 +121,7 @@ e1000_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - if(hw->media_type == e1000_media_type_copper) { + if (hw->media_type == e1000_media_type_copper) { ecmd->supported = (SUPPORTED_10baseT_Half | SUPPORTED_10baseT_Full | @@ -121,7 +133,7 @@ e1000_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) ecmd->advertising = ADVERTISED_TP; - if(hw->autoneg == 1) { + if (hw->autoneg == 1) { ecmd->advertising |= ADVERTISED_Autoneg; /* the e1000 autoneg seems to match ethtool nicely */ @@ -132,7 +144,7 @@ e1000_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) ecmd->port = PORT_TP; ecmd->phy_address = hw->phy_addr; - if(hw->mac_type == e1000_82543) + if (hw->mac_type == e1000_82543) ecmd->transceiver = XCVR_EXTERNAL; else ecmd->transceiver = XCVR_INTERNAL; @@ -148,13 +160,13 @@ e1000_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) ecmd->port = PORT_FIBRE; - if(hw->mac_type >= e1000_82545) + if (hw->mac_type >= e1000_82545) ecmd->transceiver = XCVR_INTERNAL; else ecmd->transceiver = XCVR_EXTERNAL; } - if(netif_carrier_ok(adapter->netdev)) { + if (netif_carrier_ok(adapter->netdev)) { e1000_get_speed_and_duplex(hw, &adapter->link_speed, &adapter->link_duplex); @@ -163,7 +175,7 @@ e1000_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) /* unfortunatly FULL_DUPLEX != DUPLEX_FULL * and HALF_DUPLEX != DUPLEX_HALF */ - if(adapter->link_duplex == FULL_DUPLEX) + if (adapter->link_duplex == FULL_DUPLEX) ecmd->duplex = DUPLEX_FULL; else ecmd->duplex = DUPLEX_HALF; @@ -183,13 +195,21 @@ e1000_set_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - if(ecmd->autoneg == AUTONEG_ENABLE) { + /* When SoL/IDER sessions are active, autoneg/speed/duplex + * cannot be changed */ + if (e1000_check_phy_reset_block(hw)) { + DPRINTK(DRV, ERR, "Cannot change link characteristics " + "when SoL/IDER is active.\n"); + return -EINVAL; + } + + if (ecmd->autoneg == AUTONEG_ENABLE) { hw->autoneg = 1; - if(hw->media_type == e1000_media_type_fiber) + if (hw->media_type == e1000_media_type_fiber) hw->autoneg_advertised = ADVERTISED_1000baseT_Full | ADVERTISED_FIBRE | ADVERTISED_Autoneg; - else + else hw->autoneg_advertised = ADVERTISED_10baseT_Half | ADVERTISED_10baseT_Full | ADVERTISED_100baseT_Half | @@ -199,12 +219,12 @@ e1000_set_settings(struct net_device *netdev, struct ethtool_cmd *ecmd) ADVERTISED_TP; ecmd->advertising = hw->autoneg_advertised; } else - if(e1000_set_spd_dplx(adapter, ecmd->speed + ecmd->duplex)) + if (e1000_set_spd_dplx(adapter, ecmd->speed + ecmd->duplex)) return -EINVAL; /* reset the link */ - if(netif_running(adapter->netdev)) { + if (netif_running(adapter->netdev)) { e1000_down(adapter); e1000_reset(adapter); e1000_up(adapter); @@ -221,14 +241,14 @@ e1000_get_pauseparam(struct net_device *netdev, struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - pause->autoneg = + pause->autoneg = (adapter->fc_autoneg ? AUTONEG_ENABLE : AUTONEG_DISABLE); - - if(hw->fc == e1000_fc_rx_pause) + + if (hw->fc == e1000_fc_rx_pause) pause->rx_pause = 1; - else if(hw->fc == e1000_fc_tx_pause) + else if (hw->fc == e1000_fc_tx_pause) pause->tx_pause = 1; - else if(hw->fc == e1000_fc_full) { + else if (hw->fc == e1000_fc_full) { pause->rx_pause = 1; pause->tx_pause = 1; } @@ -240,31 +260,30 @@ e1000_set_pauseparam(struct net_device *netdev, { struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - + adapter->fc_autoneg = pause->autoneg; - if(pause->rx_pause && pause->tx_pause) + if (pause->rx_pause && pause->tx_pause) hw->fc = e1000_fc_full; - else if(pause->rx_pause && !pause->tx_pause) + else if (pause->rx_pause && !pause->tx_pause) hw->fc = e1000_fc_rx_pause; - else if(!pause->rx_pause && pause->tx_pause) + else if (!pause->rx_pause && pause->tx_pause) hw->fc = e1000_fc_tx_pause; - else if(!pause->rx_pause && !pause->tx_pause) + else if (!pause->rx_pause && !pause->tx_pause) hw->fc = e1000_fc_none; hw->original_fc = hw->fc; - if(adapter->fc_autoneg == AUTONEG_ENABLE) { - if(netif_running(adapter->netdev)) { + if (adapter->fc_autoneg == AUTONEG_ENABLE) { + if (netif_running(adapter->netdev)) { e1000_down(adapter); e1000_up(adapter); } else e1000_reset(adapter); - } - else + } else return ((hw->media_type == e1000_media_type_fiber) ? e1000_setup_link(hw) : e1000_force_mac_fc(hw)); - + return 0; } @@ -281,14 +300,14 @@ e1000_set_rx_csum(struct net_device *netdev, uint32_t data) struct e1000_adapter *adapter = netdev_priv(netdev); adapter->rx_csum = data; - if(netif_running(netdev)) { + if (netif_running(netdev)) { e1000_down(adapter); e1000_up(adapter); } else e1000_reset(adapter); return 0; } - + static uint32_t e1000_get_tx_csum(struct net_device *netdev) { @@ -300,7 +319,7 @@ e1000_set_tx_csum(struct net_device *netdev, uint32_t data) { struct e1000_adapter *adapter = netdev_priv(netdev); - if(adapter->hw.mac_type < e1000_82543) { + if (adapter->hw.mac_type < e1000_82543) { if (!data) return -EINVAL; return 0; @@ -319,8 +338,8 @@ static int e1000_set_tso(struct net_device *netdev, uint32_t data) { struct e1000_adapter *adapter = netdev_priv(netdev); - if((adapter->hw.mac_type < e1000_82544) || - (adapter->hw.mac_type == e1000_82547)) + if ((adapter->hw.mac_type < e1000_82544) || + (adapter->hw.mac_type == e1000_82547)) return data ? -EINVAL : 0; if (data) @@ -328,7 +347,7 @@ e1000_set_tso(struct net_device *netdev, uint32_t data) else netdev->features &= ~NETIF_F_TSO; return 0; -} +} #endif /* NETIF_F_TSO */ static uint32_t @@ -345,7 +364,7 @@ e1000_set_msglevel(struct net_device *netdev, uint32_t data) adapter->msg_enable = data; } -static int +static int e1000_get_regs_len(struct net_device *netdev) { #define E1000_REGS_LEN 32 @@ -381,7 +400,7 @@ e1000_get_regs(struct net_device *netdev, regs_buff[11] = E1000_READ_REG(hw, TIDV); regs_buff[12] = adapter->hw.phy_type; /* PHY type (IGP=1, M88=0) */ - if(hw->phy_type == e1000_phy_igp) { + if (hw->phy_type == e1000_phy_igp) { e1000_write_phy_reg(hw, IGP01E1000_PHY_PAGE_SELECT, IGP01E1000_PHY_AGC_A); e1000_read_phy_reg(hw, IGP01E1000_PHY_AGC_A & @@ -435,7 +454,7 @@ e1000_get_regs(struct net_device *netdev, e1000_read_phy_reg(hw, PHY_1000T_STATUS, &phy_data); regs_buff[24] = (uint32_t)phy_data; /* phy local receiver status */ regs_buff[25] = regs_buff[24]; /* phy remote receiver status */ - if(hw->mac_type >= e1000_82540 && + if (hw->mac_type >= e1000_82540 && hw->media_type == e1000_media_type_copper) { regs_buff[26] = E1000_READ_REG(hw, MANC); } @@ -459,7 +478,7 @@ e1000_get_eeprom(struct net_device *netdev, int ret_val = 0; uint16_t i; - if(eeprom->len == 0) + if (eeprom->len == 0) return -EINVAL; eeprom->magic = hw->vendor_id | (hw->device_id << 16); @@ -469,16 +488,16 @@ e1000_get_eeprom(struct net_device *netdev, eeprom_buff = kmalloc(sizeof(uint16_t) * (last_word - first_word + 1), GFP_KERNEL); - if(!eeprom_buff) + if (!eeprom_buff) return -ENOMEM; - if(hw->eeprom.type == e1000_eeprom_spi) + if (hw->eeprom.type == e1000_eeprom_spi) ret_val = e1000_read_eeprom(hw, first_word, last_word - first_word + 1, eeprom_buff); else { for (i = 0; i < last_word - first_word + 1; i++) - if((ret_val = e1000_read_eeprom(hw, first_word + i, 1, + if ((ret_val = e1000_read_eeprom(hw, first_word + i, 1, &eeprom_buff[i]))) break; } @@ -505,10 +524,10 @@ e1000_set_eeprom(struct net_device *netdev, int max_len, first_word, last_word, ret_val = 0; uint16_t i; - if(eeprom->len == 0) + if (eeprom->len == 0) return -EOPNOTSUPP; - if(eeprom->magic != (hw->vendor_id | (hw->device_id << 16))) + if (eeprom->magic != (hw->vendor_id | (hw->device_id << 16))) return -EFAULT; max_len = hw->eeprom.word_size * 2; @@ -516,19 +535,19 @@ e1000_set_eeprom(struct net_device *netdev, first_word = eeprom->offset >> 1; last_word = (eeprom->offset + eeprom->len - 1) >> 1; eeprom_buff = kmalloc(max_len, GFP_KERNEL); - if(!eeprom_buff) + if (!eeprom_buff) return -ENOMEM; ptr = (void *)eeprom_buff; - if(eeprom->offset & 1) { + if (eeprom->offset & 1) { /* need read/modify/write of first changed EEPROM word */ /* only the second byte of the word is being modified */ ret_val = e1000_read_eeprom(hw, first_word, 1, &eeprom_buff[0]); ptr++; } - if(((eeprom->offset + eeprom->len) & 1) && (ret_val == 0)) { + if (((eeprom->offset + eeprom->len) & 1) && (ret_val == 0)) { /* need read/modify/write of last changed EEPROM word */ /* only the first byte of the word is being modified */ ret_val = e1000_read_eeprom(hw, last_word, 1, @@ -547,9 +566,9 @@ e1000_set_eeprom(struct net_device *netdev, ret_val = e1000_write_eeprom(hw, first_word, last_word - first_word + 1, eeprom_buff); - /* Update the checksum over the first part of the EEPROM if needed + /* Update the checksum over the first part of the EEPROM if needed * and flush shadow RAM for 82573 conrollers */ - if((ret_val == 0) && ((first_word <= EEPROM_CHECKSUM_REG) || + if ((ret_val == 0) && ((first_word <= EEPROM_CHECKSUM_REG) || (hw->mac_type == e1000_82573))) e1000_update_eeprom_checksum(hw); @@ -567,21 +586,21 @@ e1000_get_drvinfo(struct net_device *netdev, strncpy(drvinfo->driver, e1000_driver_name, 32); strncpy(drvinfo->version, e1000_driver_version, 32); - - /* EEPROM image version # is reported as firware version # for + + /* EEPROM image version # is reported as firmware version # for * 8257{1|2|3} controllers */ e1000_read_eeprom(&adapter->hw, 5, 1, &eeprom_data); switch (adapter->hw.mac_type) { case e1000_82571: case e1000_82572: case e1000_82573: - sprintf(firmware_version, "%d.%d-%d", + sprintf(firmware_version, "%d.%d-%d", (eeprom_data & 0xF000) >> 12, (eeprom_data & 0x0FF0) >> 4, eeprom_data & 0x000F); break; default: - sprintf(firmware_version, "n/a"); + sprintf(firmware_version, "N/A"); } strncpy(drvinfo->fw_version, firmware_version, 32); @@ -613,7 +632,7 @@ e1000_get_ringparam(struct net_device *netdev, ring->rx_jumbo_pending = 0; } -static int +static int e1000_set_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring) { @@ -623,8 +642,8 @@ e1000_set_ringparam(struct net_device *netdev, struct e1000_rx_ring *rxdr, *rx_old, *rx_new; int i, err, tx_ring_size, rx_ring_size; - tx_ring_size = sizeof(struct e1000_tx_ring) * adapter->num_queues; - rx_ring_size = sizeof(struct e1000_rx_ring) * adapter->num_queues; + tx_ring_size = sizeof(struct e1000_tx_ring) * adapter->num_tx_queues; + rx_ring_size = sizeof(struct e1000_rx_ring) * adapter->num_rx_queues; if (netif_running(adapter->netdev)) e1000_down(adapter); @@ -650,25 +669,25 @@ e1000_set_ringparam(struct net_device *netdev, txdr = adapter->tx_ring; rxdr = adapter->rx_ring; - if((ring->rx_mini_pending) || (ring->rx_jumbo_pending)) + if ((ring->rx_mini_pending) || (ring->rx_jumbo_pending)) return -EINVAL; rxdr->count = max(ring->rx_pending,(uint32_t)E1000_MIN_RXD); rxdr->count = min(rxdr->count,(uint32_t)(mac_type < e1000_82544 ? E1000_MAX_RXD : E1000_MAX_82544_RXD)); - E1000_ROUNDUP(rxdr->count, REQ_RX_DESCRIPTOR_MULTIPLE); + E1000_ROUNDUP(rxdr->count, REQ_RX_DESCRIPTOR_MULTIPLE); txdr->count = max(ring->tx_pending,(uint32_t)E1000_MIN_TXD); txdr->count = min(txdr->count,(uint32_t)(mac_type < e1000_82544 ? E1000_MAX_TXD : E1000_MAX_82544_TXD)); - E1000_ROUNDUP(txdr->count, REQ_TX_DESCRIPTOR_MULTIPLE); + E1000_ROUNDUP(txdr->count, REQ_TX_DESCRIPTOR_MULTIPLE); - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_tx_queues; i++) txdr[i].count = txdr->count; + for (i = 0; i < adapter->num_rx_queues; i++) rxdr[i].count = rxdr->count; - } - if(netif_running(adapter->netdev)) { + if (netif_running(adapter->netdev)) { /* Try to get new resources before deleting old */ if ((err = e1000_setup_all_rx_resources(adapter))) goto err_setup_rx; @@ -688,7 +707,7 @@ e1000_set_ringparam(struct net_device *netdev, kfree(rx_old); adapter->rx_ring = rx_new; adapter->tx_ring = tx_new; - if((err = e1000_up(adapter))) + if ((err = e1000_up(adapter))) return err; } @@ -707,10 +726,10 @@ err_setup_rx: uint32_t pat, value; \ uint32_t test[] = \ {0x5A5A5A5A, 0xA5A5A5A5, 0x00000000, 0xFFFFFFFF}; \ - for(pat = 0; pat < sizeof(test)/sizeof(test[0]); pat++) { \ + for (pat = 0; pat < sizeof(test)/sizeof(test[0]); pat++) { \ E1000_WRITE_REG(&adapter->hw, R, (test[pat] & W)); \ value = E1000_READ_REG(&adapter->hw, R); \ - if(value != (test[pat] & W & M)) { \ + if (value != (test[pat] & W & M)) { \ DPRINTK(DRV, ERR, "pattern test reg %04X failed: got " \ "0x%08X expected 0x%08X\n", \ E1000_##R, value, (test[pat] & W & M)); \ @@ -726,7 +745,7 @@ err_setup_rx: uint32_t value; \ E1000_WRITE_REG(&adapter->hw, R, W & M); \ value = E1000_READ_REG(&adapter->hw, R); \ - if((W & M) != (value & M)) { \ + if ((W & M) != (value & M)) { \ DPRINTK(DRV, ERR, "set/check reg %04X test failed: got 0x%08X "\ "expected 0x%08X\n", E1000_##R, (value & M), (W & M)); \ *data = (adapter->hw.mac_type < e1000_82543) ? \ @@ -762,7 +781,7 @@ e1000_reg_test(struct e1000_adapter *adapter, uint64_t *data) value = (E1000_READ_REG(&adapter->hw, STATUS) & toggle); E1000_WRITE_REG(&adapter->hw, STATUS, toggle); after = E1000_READ_REG(&adapter->hw, STATUS) & toggle; - if(value != after) { + if (value != after) { DPRINTK(DRV, ERR, "failed STATUS register test got: " "0x%08X expected: 0x%08X\n", after, value); *data = 1; @@ -790,7 +809,7 @@ e1000_reg_test(struct e1000_adapter *adapter, uint64_t *data) REG_SET_AND_CHECK(RCTL, 0x06DFB3FE, 0x003FFFFB); REG_SET_AND_CHECK(TCTL, 0xFFFFFFFF, 0x00000000); - if(adapter->hw.mac_type >= e1000_82543) { + if (adapter->hw.mac_type >= e1000_82543) { REG_SET_AND_CHECK(RCTL, 0x06DFB3FE, 0xFFFFFFFF); REG_PATTERN_TEST(RDBAL, 0xFFFFFFF0, 0xFFFFFFFF); @@ -798,7 +817,7 @@ e1000_reg_test(struct e1000_adapter *adapter, uint64_t *data) REG_PATTERN_TEST(TDBAL, 0xFFFFFFF0, 0xFFFFFFFF); REG_PATTERN_TEST(TIDV, 0x0000FFFF, 0x0000FFFF); - for(i = 0; i < E1000_RAR_ENTRIES; i++) { + for (i = 0; i < E1000_RAR_ENTRIES; i++) { REG_PATTERN_TEST(RA + ((i << 1) << 2), 0xFFFFFFFF, 0xFFFFFFFF); REG_PATTERN_TEST(RA + (((i << 1) + 1) << 2), 0x8003FFFF, @@ -814,7 +833,7 @@ e1000_reg_test(struct e1000_adapter *adapter, uint64_t *data) } - for(i = 0; i < E1000_MC_TBL_SIZE; i++) + for (i = 0; i < E1000_MC_TBL_SIZE; i++) REG_PATTERN_TEST(MTA + (i << 2), 0xFFFFFFFF, 0xFFFFFFFF); *data = 0; @@ -830,8 +849,8 @@ e1000_eeprom_test(struct e1000_adapter *adapter, uint64_t *data) *data = 0; /* Read and add up the contents of the EEPROM */ - for(i = 0; i < (EEPROM_CHECKSUM_REG + 1); i++) { - if((e1000_read_eeprom(&adapter->hw, i, 1, &temp)) < 0) { + for (i = 0; i < (EEPROM_CHECKSUM_REG + 1); i++) { + if ((e1000_read_eeprom(&adapter->hw, i, 1, &temp)) < 0) { *data = 1; break; } @@ -839,7 +858,7 @@ e1000_eeprom_test(struct e1000_adapter *adapter, uint64_t *data) } /* If Checksum is not Correct return error else test passed */ - if((checksum != (uint16_t) EEPROM_SUM) && !(*data)) + if ((checksum != (uint16_t) EEPROM_SUM) && !(*data)) *data = 2; return *data; @@ -868,9 +887,9 @@ e1000_intr_test(struct e1000_adapter *adapter, uint64_t *data) *data = 0; /* Hook up test interrupt handler just for this test */ - if(!request_irq(irq, &e1000_test_intr, 0, netdev->name, netdev)) { + if (!request_irq(irq, &e1000_test_intr, 0, netdev->name, netdev)) { shared_int = FALSE; - } else if(request_irq(irq, &e1000_test_intr, SA_SHIRQ, + } else if (request_irq(irq, &e1000_test_intr, SA_SHIRQ, netdev->name, netdev)){ *data = 1; return -1; @@ -881,12 +900,12 @@ e1000_intr_test(struct e1000_adapter *adapter, uint64_t *data) msec_delay(10); /* Test each interrupt */ - for(; i < 10; i++) { + for (; i < 10; i++) { /* Interrupt to test */ mask = 1 << i; - if(!shared_int) { + if (!shared_int) { /* Disable the interrupt to be reported in * the cause register and then force the same * interrupt and see if one gets posted. If @@ -897,8 +916,8 @@ e1000_intr_test(struct e1000_adapter *adapter, uint64_t *data) E1000_WRITE_REG(&adapter->hw, IMC, mask); E1000_WRITE_REG(&adapter->hw, ICS, mask); msec_delay(10); - - if(adapter->test_icr & mask) { + + if (adapter->test_icr & mask) { *data = 3; break; } @@ -915,12 +934,12 @@ e1000_intr_test(struct e1000_adapter *adapter, uint64_t *data) E1000_WRITE_REG(&adapter->hw, ICS, mask); msec_delay(10); - if(!(adapter->test_icr & mask)) { + if (!(adapter->test_icr & mask)) { *data = 4; break; } - if(!shared_int) { + if (!shared_int) { /* Disable the other interrupts to be reported in * the cause register and then force the other * interrupts and see if any get posted. If @@ -932,7 +951,7 @@ e1000_intr_test(struct e1000_adapter *adapter, uint64_t *data) E1000_WRITE_REG(&adapter->hw, ICS, ~mask & 0x00007FFF); msec_delay(10); - if(adapter->test_icr) { + if (adapter->test_icr) { *data = 5; break; } @@ -957,40 +976,39 @@ e1000_free_desc_rings(struct e1000_adapter *adapter) struct pci_dev *pdev = adapter->pdev; int i; - if(txdr->desc && txdr->buffer_info) { - for(i = 0; i < txdr->count; i++) { - if(txdr->buffer_info[i].dma) + if (txdr->desc && txdr->buffer_info) { + for (i = 0; i < txdr->count; i++) { + if (txdr->buffer_info[i].dma) pci_unmap_single(pdev, txdr->buffer_info[i].dma, txdr->buffer_info[i].length, PCI_DMA_TODEVICE); - if(txdr->buffer_info[i].skb) + if (txdr->buffer_info[i].skb) dev_kfree_skb(txdr->buffer_info[i].skb); } } - if(rxdr->desc && rxdr->buffer_info) { - for(i = 0; i < rxdr->count; i++) { - if(rxdr->buffer_info[i].dma) + if (rxdr->desc && rxdr->buffer_info) { + for (i = 0; i < rxdr->count; i++) { + if (rxdr->buffer_info[i].dma) pci_unmap_single(pdev, rxdr->buffer_info[i].dma, rxdr->buffer_info[i].length, PCI_DMA_FROMDEVICE); - if(rxdr->buffer_info[i].skb) + if (rxdr->buffer_info[i].skb) dev_kfree_skb(rxdr->buffer_info[i].skb); } } - if(txdr->desc) { + if (txdr->desc) { pci_free_consistent(pdev, txdr->size, txdr->desc, txdr->dma); txdr->desc = NULL; } - if(rxdr->desc) { + if (rxdr->desc) { pci_free_consistent(pdev, rxdr->size, rxdr->desc, rxdr->dma); rxdr->desc = NULL; } kfree(txdr->buffer_info); txdr->buffer_info = NULL; - kfree(rxdr->buffer_info); rxdr->buffer_info = NULL; @@ -1008,11 +1026,11 @@ e1000_setup_desc_rings(struct e1000_adapter *adapter) /* Setup Tx descriptor ring and Tx buffers */ - if(!txdr->count) - txdr->count = E1000_DEFAULT_TXD; + if (!txdr->count) + txdr->count = E1000_DEFAULT_TXD; size = txdr->count * sizeof(struct e1000_buffer); - if(!(txdr->buffer_info = kmalloc(size, GFP_KERNEL))) { + if (!(txdr->buffer_info = kmalloc(size, GFP_KERNEL))) { ret_val = 1; goto err_nomem; } @@ -1020,7 +1038,7 @@ e1000_setup_desc_rings(struct e1000_adapter *adapter) txdr->size = txdr->count * sizeof(struct e1000_tx_desc); E1000_ROUNDUP(txdr->size, 4096); - if(!(txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma))) { + if (!(txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma))) { ret_val = 2; goto err_nomem; } @@ -1039,12 +1057,12 @@ e1000_setup_desc_rings(struct e1000_adapter *adapter) E1000_COLLISION_THRESHOLD << E1000_CT_SHIFT | E1000_FDX_COLLISION_DISTANCE << E1000_COLD_SHIFT); - for(i = 0; i < txdr->count; i++) { + for (i = 0; i < txdr->count; i++) { struct e1000_tx_desc *tx_desc = E1000_TX_DESC(*txdr, i); struct sk_buff *skb; unsigned int size = 1024; - if(!(skb = alloc_skb(size, GFP_KERNEL))) { + if (!(skb = alloc_skb(size, GFP_KERNEL))) { ret_val = 3; goto err_nomem; } @@ -1064,18 +1082,18 @@ e1000_setup_desc_rings(struct e1000_adapter *adapter) /* Setup Rx descriptor ring and Rx buffers */ - if(!rxdr->count) - rxdr->count = E1000_DEFAULT_RXD; + if (!rxdr->count) + rxdr->count = E1000_DEFAULT_RXD; size = rxdr->count * sizeof(struct e1000_buffer); - if(!(rxdr->buffer_info = kmalloc(size, GFP_KERNEL))) { + if (!(rxdr->buffer_info = kmalloc(size, GFP_KERNEL))) { ret_val = 4; goto err_nomem; } memset(rxdr->buffer_info, 0, size); rxdr->size = rxdr->count * sizeof(struct e1000_rx_desc); - if(!(rxdr->desc = pci_alloc_consistent(pdev, rxdr->size, &rxdr->dma))) { + if (!(rxdr->desc = pci_alloc_consistent(pdev, rxdr->size, &rxdr->dma))) { ret_val = 5; goto err_nomem; } @@ -1095,11 +1113,11 @@ e1000_setup_desc_rings(struct e1000_adapter *adapter) (adapter->hw.mc_filter_type << E1000_RCTL_MO_SHIFT); E1000_WRITE_REG(&adapter->hw, RCTL, rctl); - for(i = 0; i < rxdr->count; i++) { + for (i = 0; i < rxdr->count; i++) { struct e1000_rx_desc *rx_desc = E1000_RX_DESC(*rxdr, i); struct sk_buff *skb; - if(!(skb = alloc_skb(E1000_RXBUFFER_2048 + NET_IP_ALIGN, + if (!(skb = alloc_skb(E1000_RXBUFFER_2048 + NET_IP_ALIGN, GFP_KERNEL))) { ret_val = 6; goto err_nomem; @@ -1208,15 +1226,15 @@ e1000_nonintegrated_phy_loopback(struct e1000_adapter *adapter) /* Check Phy Configuration */ e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &phy_reg); - if(phy_reg != 0x4100) + if (phy_reg != 0x4100) return 9; e1000_read_phy_reg(&adapter->hw, M88E1000_EXT_PHY_SPEC_CTRL, &phy_reg); - if(phy_reg != 0x0070) + if (phy_reg != 0x0070) return 10; e1000_read_phy_reg(&adapter->hw, 29, &phy_reg); - if(phy_reg != 0x001A) + if (phy_reg != 0x001A) return 11; return 0; @@ -1230,7 +1248,7 @@ e1000_integrated_phy_loopback(struct e1000_adapter *adapter) adapter->hw.autoneg = FALSE; - if(adapter->hw.phy_type == e1000_phy_m88) { + if (adapter->hw.phy_type == e1000_phy_m88) { /* Auto-MDI/MDIX Off */ e1000_write_phy_reg(&adapter->hw, M88E1000_PHY_SPEC_CTRL, 0x0808); @@ -1250,14 +1268,14 @@ e1000_integrated_phy_loopback(struct e1000_adapter *adapter) E1000_CTRL_SPD_1000 |/* Force Speed to 1000 */ E1000_CTRL_FD); /* Force Duplex to FULL */ - if(adapter->hw.media_type == e1000_media_type_copper && + if (adapter->hw.media_type == e1000_media_type_copper && adapter->hw.phy_type == e1000_phy_m88) { ctrl_reg |= E1000_CTRL_ILOS; /* Invert Loss of Signal */ } else { /* Set the ILOS bit on the fiber Nic is half * duplex link is detected. */ stat_reg = E1000_READ_REG(&adapter->hw, STATUS); - if((stat_reg & E1000_STATUS_FD) == 0) + if ((stat_reg & E1000_STATUS_FD) == 0) ctrl_reg |= (E1000_CTRL_ILOS | E1000_CTRL_SLU); } @@ -1266,7 +1284,7 @@ e1000_integrated_phy_loopback(struct e1000_adapter *adapter) /* Disable the receiver on the PHY so when a cable is plugged in, the * PHY does not begin to autoneg when a cable is reconnected to the NIC. */ - if(adapter->hw.phy_type == e1000_phy_m88) + if (adapter->hw.phy_type == e1000_phy_m88) e1000_phy_disable_receiver(adapter); udelay(500); @@ -1282,14 +1300,14 @@ e1000_set_phy_loopback(struct e1000_adapter *adapter) switch (adapter->hw.mac_type) { case e1000_82543: - if(adapter->hw.media_type == e1000_media_type_copper) { + if (adapter->hw.media_type == e1000_media_type_copper) { /* Attempt to setup Loopback mode on Non-integrated PHY. * Some PHY registers get corrupted at random, so * attempt this 10 times. */ - while(e1000_nonintegrated_phy_loopback(adapter) && + while (e1000_nonintegrated_phy_loopback(adapter) && count++ < 10); - if(count < 11) + if (count < 11) return 0; } break; @@ -1327,11 +1345,11 @@ e1000_set_phy_loopback(struct e1000_adapter *adapter) static int e1000_setup_loopback_test(struct e1000_adapter *adapter) { - uint32_t rctl; struct e1000_hw *hw = &adapter->hw; + uint32_t rctl; if (hw->media_type == e1000_media_type_fiber || - hw->media_type == e1000_media_type_internal_serdes) { + hw->media_type == e1000_media_type_internal_serdes) { switch (hw->mac_type) { case e1000_82545: case e1000_82546: @@ -1362,25 +1380,25 @@ e1000_setup_loopback_test(struct e1000_adapter *adapter) static void e1000_loopback_cleanup(struct e1000_adapter *adapter) { + struct e1000_hw *hw = &adapter->hw; uint32_t rctl; uint16_t phy_reg; - struct e1000_hw *hw = &adapter->hw; - rctl = E1000_READ_REG(&adapter->hw, RCTL); + rctl = E1000_READ_REG(hw, RCTL); rctl &= ~(E1000_RCTL_LBM_TCVR | E1000_RCTL_LBM_MAC); - E1000_WRITE_REG(&adapter->hw, RCTL, rctl); + E1000_WRITE_REG(hw, RCTL, rctl); switch (hw->mac_type) { case e1000_82571: case e1000_82572: if (hw->media_type == e1000_media_type_fiber || - hw->media_type == e1000_media_type_internal_serdes){ + hw->media_type == e1000_media_type_internal_serdes) { #define E1000_SERDES_LB_OFF 0x400 E1000_WRITE_REG(hw, SCTL, E1000_SERDES_LB_OFF); msec_delay(10); break; } - /* fall thru for Cu adapters */ + /* Fall Through */ case e1000_82545: case e1000_82546: case e1000_82545_rev_3: @@ -1401,7 +1419,7 @@ static void e1000_create_lbtest_frame(struct sk_buff *skb, unsigned int frame_size) { memset(skb->data, 0xFF, frame_size); - frame_size = (frame_size % 2) ? (frame_size - 1) : frame_size; + frame_size &= ~1; memset(&skb->data[frame_size / 2], 0xAA, frame_size / 2 - 1); memset(&skb->data[frame_size / 2 + 10], 0xBE, 1); memset(&skb->data[frame_size / 2 + 12], 0xAF, 1); @@ -1410,9 +1428,9 @@ e1000_create_lbtest_frame(struct sk_buff *skb, unsigned int frame_size) static int e1000_check_lbtest_frame(struct sk_buff *skb, unsigned int frame_size) { - frame_size = (frame_size % 2) ? (frame_size - 1) : frame_size; - if(*(skb->data + 3) == 0xFF) { - if((*(skb->data + frame_size / 2 + 10) == 0xBE) && + frame_size &= ~1; + if (*(skb->data + 3) == 0xFF) { + if ((*(skb->data + frame_size / 2 + 10) == 0xBE) && (*(skb->data + frame_size / 2 + 12) == 0xAF)) { return 0; } @@ -1431,53 +1449,53 @@ e1000_run_loopback_test(struct e1000_adapter *adapter) E1000_WRITE_REG(&adapter->hw, RDT, rxdr->count - 1); - /* Calculate the loop count based on the largest descriptor ring + /* Calculate the loop count based on the largest descriptor ring * The idea is to wrap the largest ring a number of times using 64 * send/receive pairs during each loop */ - if(rxdr->count <= txdr->count) + if (rxdr->count <= txdr->count) lc = ((txdr->count / 64) * 2) + 1; else lc = ((rxdr->count / 64) * 2) + 1; k = l = 0; - for(j = 0; j <= lc; j++) { /* loop count loop */ - for(i = 0; i < 64; i++) { /* send the packets */ - e1000_create_lbtest_frame(txdr->buffer_info[i].skb, + for (j = 0; j <= lc; j++) { /* loop count loop */ + for (i = 0; i < 64; i++) { /* send the packets */ + e1000_create_lbtest_frame(txdr->buffer_info[i].skb, 1024); - pci_dma_sync_single_for_device(pdev, + pci_dma_sync_single_for_device(pdev, txdr->buffer_info[k].dma, txdr->buffer_info[k].length, PCI_DMA_TODEVICE); - if(unlikely(++k == txdr->count)) k = 0; + if (unlikely(++k == txdr->count)) k = 0; } E1000_WRITE_REG(&adapter->hw, TDT, k); msec_delay(200); time = jiffies; /* set the start time for the receive */ good_cnt = 0; do { /* receive the sent packets */ - pci_dma_sync_single_for_cpu(pdev, + pci_dma_sync_single_for_cpu(pdev, rxdr->buffer_info[l].dma, rxdr->buffer_info[l].length, PCI_DMA_FROMDEVICE); - + ret_val = e1000_check_lbtest_frame( rxdr->buffer_info[l].skb, 1024); - if(!ret_val) + if (!ret_val) good_cnt++; - if(unlikely(++l == rxdr->count)) l = 0; - /* time + 20 msecs (200 msecs on 2.4) is more than - * enough time to complete the receives, if it's + if (unlikely(++l == rxdr->count)) l = 0; + /* time + 20 msecs (200 msecs on 2.4) is more than + * enough time to complete the receives, if it's * exceeded, break and error off */ } while (good_cnt < 64 && jiffies < (time + 20)); - if(good_cnt != 64) { + if (good_cnt != 64) { ret_val = 13; /* ret_val is the same as mis-compare */ - break; + break; } - if(jiffies >= (time + 2)) { + if (jiffies >= (time + 2)) { ret_val = 14; /* error code for time out error */ break; } @@ -1488,14 +1506,25 @@ e1000_run_loopback_test(struct e1000_adapter *adapter) static int e1000_loopback_test(struct e1000_adapter *adapter, uint64_t *data) { - if((*data = e1000_setup_desc_rings(adapter))) goto err_loopback; - if((*data = e1000_setup_loopback_test(adapter))) - goto err_loopback_setup; + /* PHY loopback cannot be performed if SoL/IDER + * sessions are active */ + if (e1000_check_phy_reset_block(&adapter->hw)) { + DPRINTK(DRV, ERR, "Cannot do PHY loopback test " + "when SoL/IDER is active.\n"); + *data = 0; + goto out; + } + + if ((*data = e1000_setup_desc_rings(adapter))) + goto out; + if ((*data = e1000_setup_loopback_test(adapter))) + goto err_loopback; *data = e1000_run_loopback_test(adapter); e1000_loopback_cleanup(adapter); -err_loopback_setup: - e1000_free_desc_rings(adapter); + err_loopback: + e1000_free_desc_rings(adapter); +out: return *data; } @@ -1519,17 +1548,17 @@ e1000_link_test(struct e1000_adapter *adapter, uint64_t *data) *data = 1; } else { e1000_check_for_link(&adapter->hw); - if(adapter->hw.autoneg) /* if auto_neg is set wait for it */ + if (adapter->hw.autoneg) /* if auto_neg is set wait for it */ msec_delay(4000); - if(!(E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_LU)) { + if (!(E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_LU)) { *data = 1; } } return *data; } -static int +static int e1000_diag_test_count(struct net_device *netdev) { return E1000_TEST_LEN; @@ -1542,7 +1571,7 @@ e1000_diag_test(struct net_device *netdev, struct e1000_adapter *adapter = netdev_priv(netdev); boolean_t if_running = netif_running(netdev); - if(eth_test->flags == ETH_TEST_FL_OFFLINE) { + if (eth_test->flags == ETH_TEST_FL_OFFLINE) { /* Offline tests */ /* save speed, duplex, autoneg settings */ @@ -1552,27 +1581,27 @@ e1000_diag_test(struct net_device *netdev, /* Link test performed before hardware reset so autoneg doesn't * interfere with test result */ - if(e1000_link_test(adapter, &data[4])) + if (e1000_link_test(adapter, &data[4])) eth_test->flags |= ETH_TEST_FL_FAILED; - if(if_running) + if (if_running) e1000_down(adapter); else e1000_reset(adapter); - if(e1000_reg_test(adapter, &data[0])) + if (e1000_reg_test(adapter, &data[0])) eth_test->flags |= ETH_TEST_FL_FAILED; e1000_reset(adapter); - if(e1000_eeprom_test(adapter, &data[1])) + if (e1000_eeprom_test(adapter, &data[1])) eth_test->flags |= ETH_TEST_FL_FAILED; e1000_reset(adapter); - if(e1000_intr_test(adapter, &data[2])) + if (e1000_intr_test(adapter, &data[2])) eth_test->flags |= ETH_TEST_FL_FAILED; e1000_reset(adapter); - if(e1000_loopback_test(adapter, &data[3])) + if (e1000_loopback_test(adapter, &data[3])) eth_test->flags |= ETH_TEST_FL_FAILED; /* restore speed, duplex, autoneg settings */ @@ -1581,11 +1610,11 @@ e1000_diag_test(struct net_device *netdev, adapter->hw.autoneg = autoneg; e1000_reset(adapter); - if(if_running) + if (if_running) e1000_up(adapter); } else { /* Online tests */ - if(e1000_link_test(adapter, &data[4])) + if (e1000_link_test(adapter, &data[4])) eth_test->flags |= ETH_TEST_FL_FAILED; /* Offline tests aren't run; pass by default */ @@ -1603,7 +1632,7 @@ e1000_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - switch(adapter->hw.device_id) { + switch (adapter->hw.device_id) { case E1000_DEV_ID_82542: case E1000_DEV_ID_82543GC_FIBER: case E1000_DEV_ID_82543GC_COPPER: @@ -1617,8 +1646,9 @@ e1000_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) case E1000_DEV_ID_82546EB_FIBER: case E1000_DEV_ID_82546GB_FIBER: + case E1000_DEV_ID_82571EB_FIBER: /* Wake events only supported on port A for dual fiber */ - if(E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) { + if (E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) { wol->supported = 0; wol->wolopts = 0; return; @@ -1630,13 +1660,13 @@ e1000_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) WAKE_BCAST | WAKE_MAGIC; wol->wolopts = 0; - if(adapter->wol & E1000_WUFC_EX) + if (adapter->wol & E1000_WUFC_EX) wol->wolopts |= WAKE_UCAST; - if(adapter->wol & E1000_WUFC_MC) + if (adapter->wol & E1000_WUFC_MC) wol->wolopts |= WAKE_MCAST; - if(adapter->wol & E1000_WUFC_BC) + if (adapter->wol & E1000_WUFC_BC) wol->wolopts |= WAKE_BCAST; - if(adapter->wol & E1000_WUFC_MAG) + if (adapter->wol & E1000_WUFC_MAG) wol->wolopts |= WAKE_MAGIC; return; } @@ -1648,7 +1678,7 @@ e1000_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; - switch(adapter->hw.device_id) { + switch (adapter->hw.device_id) { case E1000_DEV_ID_82542: case E1000_DEV_ID_82543GC_FIBER: case E1000_DEV_ID_82543GC_COPPER: @@ -1660,24 +1690,25 @@ e1000_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol) case E1000_DEV_ID_82546EB_FIBER: case E1000_DEV_ID_82546GB_FIBER: + case E1000_DEV_ID_82571EB_FIBER: /* Wake events only supported on port A for dual fiber */ - if(E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) + if (E1000_READ_REG(hw, STATUS) & E1000_STATUS_FUNC_1) return wol->wolopts ? -EOPNOTSUPP : 0; /* Fall Through */ default: - if(wol->wolopts & (WAKE_PHY | WAKE_ARP | WAKE_MAGICSECURE)) + if (wol->wolopts & (WAKE_PHY | WAKE_ARP | WAKE_MAGICSECURE)) return -EOPNOTSUPP; adapter->wol = 0; - if(wol->wolopts & WAKE_UCAST) + if (wol->wolopts & WAKE_UCAST) adapter->wol |= E1000_WUFC_EX; - if(wol->wolopts & WAKE_MCAST) + if (wol->wolopts & WAKE_MCAST) adapter->wol |= E1000_WUFC_MC; - if(wol->wolopts & WAKE_BCAST) + if (wol->wolopts & WAKE_BCAST) adapter->wol |= E1000_WUFC_BC; - if(wol->wolopts & WAKE_MAGIC) + if (wol->wolopts & WAKE_MAGIC) adapter->wol |= E1000_WUFC_MAG; } @@ -1695,7 +1726,7 @@ e1000_led_blink_callback(unsigned long data) { struct e1000_adapter *adapter = (struct e1000_adapter *) data; - if(test_and_change_bit(E1000_LED_ON, &adapter->led_status)) + if (test_and_change_bit(E1000_LED_ON, &adapter->led_status)) e1000_led_off(&adapter->hw); else e1000_led_on(&adapter->hw); @@ -1708,11 +1739,11 @@ e1000_phys_id(struct net_device *netdev, uint32_t data) { struct e1000_adapter *adapter = netdev_priv(netdev); - if(!data || data > (uint32_t)(MAX_SCHEDULE_TIMEOUT / HZ)) + if (!data || data > (uint32_t)(MAX_SCHEDULE_TIMEOUT / HZ)) data = (uint32_t)(MAX_SCHEDULE_TIMEOUT / HZ); - if(adapter->hw.mac_type < e1000_82571) { - if(!adapter->blink_timer.function) { + if (adapter->hw.mac_type < e1000_82571) { + if (!adapter->blink_timer.function) { init_timer(&adapter->blink_timer); adapter->blink_timer.function = e1000_led_blink_callback; adapter->blink_timer.data = (unsigned long) adapter; @@ -1721,21 +1752,21 @@ e1000_phys_id(struct net_device *netdev, uint32_t data) mod_timer(&adapter->blink_timer, jiffies); msleep_interruptible(data * 1000); del_timer_sync(&adapter->blink_timer); - } - else if(adapter->hw.mac_type < e1000_82573) { - E1000_WRITE_REG(&adapter->hw, LEDCTL, (E1000_LEDCTL_LED2_BLINK_RATE | - E1000_LEDCTL_LED0_BLINK | E1000_LEDCTL_LED2_BLINK | - (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED0_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED1_MODE_SHIFT))); + } else if (adapter->hw.mac_type < e1000_82573) { + E1000_WRITE_REG(&adapter->hw, LEDCTL, + (E1000_LEDCTL_LED2_BLINK_RATE | + E1000_LEDCTL_LED0_BLINK | E1000_LEDCTL_LED2_BLINK | + (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED0_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED1_MODE_SHIFT))); msleep_interruptible(data * 1000); - } - else { - E1000_WRITE_REG(&adapter->hw, LEDCTL, (E1000_LEDCTL_LED2_BLINK_RATE | - E1000_LEDCTL_LED1_BLINK | E1000_LEDCTL_LED2_BLINK | - (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED1_MODE_SHIFT) | - (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED0_MODE_SHIFT))); + } else { + E1000_WRITE_REG(&adapter->hw, LEDCTL, + (E1000_LEDCTL_LED2_BLINK_RATE | + E1000_LEDCTL_LED1_BLINK | E1000_LEDCTL_LED2_BLINK | + (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED2_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LINK_ACTIVITY << E1000_LEDCTL_LED1_MODE_SHIFT) | + (E1000_LEDCTL_MODE_LED_OFF << E1000_LEDCTL_LED0_MODE_SHIFT))); msleep_interruptible(data * 1000); } @@ -1750,50 +1781,89 @@ static int e1000_nway_reset(struct net_device *netdev) { struct e1000_adapter *adapter = netdev_priv(netdev); - if(netif_running(netdev)) { + if (netif_running(netdev)) { e1000_down(adapter); e1000_up(adapter); } return 0; } -static int +static int e1000_get_stats_count(struct net_device *netdev) { return E1000_STATS_LEN; } -static void -e1000_get_ethtool_stats(struct net_device *netdev, +static void +e1000_get_ethtool_stats(struct net_device *netdev, struct ethtool_stats *stats, uint64_t *data) { struct e1000_adapter *adapter = netdev_priv(netdev); +#ifdef CONFIG_E1000_MQ + uint64_t *queue_stat; + int stat_count = sizeof(struct e1000_queue_stats) / sizeof(uint64_t); + int j, k; +#endif int i; e1000_update_stats(adapter); - for(i = 0; i < E1000_STATS_LEN; i++) { - char *p = (char *)adapter+e1000_gstrings_stats[i].stat_offset; - data[i] = (e1000_gstrings_stats[i].sizeof_stat == + for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) { + char *p = (char *)adapter+e1000_gstrings_stats[i].stat_offset; + data[i] = (e1000_gstrings_stats[i].sizeof_stat == sizeof(uint64_t)) ? *(uint64_t *)p : *(uint32_t *)p; } +#ifdef CONFIG_E1000_MQ + for (j = 0; j < adapter->num_tx_queues; j++) { + queue_stat = (uint64_t *)&adapter->tx_ring[j].tx_stats; + for (k = 0; k < stat_count; k++) + data[i + k] = queue_stat[k]; + i += k; + } + for (j = 0; j < adapter->num_rx_queues; j++) { + queue_stat = (uint64_t *)&adapter->rx_ring[j].rx_stats; + for (k = 0; k < stat_count; k++) + data[i + k] = queue_stat[k]; + i += k; + } +#endif +/* BUG_ON(i != E1000_STATS_LEN); */ } -static void +static void e1000_get_strings(struct net_device *netdev, uint32_t stringset, uint8_t *data) { +#ifdef CONFIG_E1000_MQ + struct e1000_adapter *adapter = netdev_priv(netdev); +#endif + uint8_t *p = data; int i; - switch(stringset) { + switch (stringset) { case ETH_SS_TEST: - memcpy(data, *e1000_gstrings_test, + memcpy(data, *e1000_gstrings_test, E1000_TEST_LEN*ETH_GSTRING_LEN); break; case ETH_SS_STATS: - for (i=0; i < E1000_STATS_LEN; i++) { - memcpy(data + i * ETH_GSTRING_LEN, - e1000_gstrings_stats[i].stat_string, - ETH_GSTRING_LEN); + for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) { + memcpy(p, e1000_gstrings_stats[i].stat_string, + ETH_GSTRING_LEN); + p += ETH_GSTRING_LEN; } +#ifdef CONFIG_E1000_MQ + for (i = 0; i < adapter->num_tx_queues; i++) { + sprintf(p, "tx_queue_%u_packets", i); + p += ETH_GSTRING_LEN; + sprintf(p, "tx_queue_%u_bytes", i); + p += ETH_GSTRING_LEN; + } + for (i = 0; i < adapter->num_rx_queues; i++) { + sprintf(p, "rx_queue_%u_packets", i); + p += ETH_GSTRING_LEN; + sprintf(p, "rx_queue_%u_bytes", i); + p += ETH_GSTRING_LEN; + } +#endif +/* BUG_ON(p - data != E1000_STATS_LEN * ETH_GSTRING_LEN); */ break; } } diff --git a/drivers/net/e1000/e1000_hw.c b/drivers/net/e1000/e1000_hw.c index 136fc031e4a..beeec0fbbea 100644 --- a/drivers/net/e1000/e1000_hw.c +++ b/drivers/net/e1000/e1000_hw.c @@ -318,6 +318,8 @@ e1000_set_mac_type(struct e1000_hw *hw) case E1000_DEV_ID_82546GB_FIBER: case E1000_DEV_ID_82546GB_SERDES: case E1000_DEV_ID_82546GB_PCIE: + case E1000_DEV_ID_82546GB_QUAD_COPPER: + case E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3: hw->mac_type = e1000_82546_rev_3; break; case E1000_DEV_ID_82541EI: @@ -639,6 +641,7 @@ e1000_init_hw(struct e1000_hw *hw) uint16_t cmd_mmrbc; uint16_t stat_mmrbc; uint32_t mta_size; + uint32_t ctrl_ext; DEBUGFUNC("e1000_init_hw"); @@ -735,7 +738,6 @@ e1000_init_hw(struct e1000_hw *hw) break; case e1000_82571: case e1000_82572: - ctrl |= (1 << 22); case e1000_82573: ctrl |= E1000_TXDCTL_COUNT_DESC; break; @@ -775,6 +777,15 @@ e1000_init_hw(struct e1000_hw *hw) */ e1000_clear_hw_cntrs(hw); + if (hw->device_id == E1000_DEV_ID_82546GB_QUAD_COPPER || + hw->device_id == E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3) { + ctrl_ext = E1000_READ_REG(hw, CTRL_EXT); + /* Relaxed ordering must be disabled to avoid a parity + * error crash in a PCI slot. */ + ctrl_ext |= E1000_CTRL_EXT_RO_DIS; + E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext); + } + return ret_val; } @@ -838,6 +849,11 @@ e1000_setup_link(struct e1000_hw *hw) DEBUGFUNC("e1000_setup_link"); + /* In the case of the phy reset being blocked, we already have a link. + * We do not have to set it up again. */ + if (e1000_check_phy_reset_block(hw)) + return E1000_SUCCESS; + /* Read and store word 0x0F of the EEPROM. This word contains bits * that determine the hardware's default PAUSE (flow control) mode, * a bit that determines whether the HW defaults to enabling or @@ -1584,10 +1600,10 @@ e1000_phy_setup_autoneg(struct e1000_hw *hw) if(ret_val) return ret_val; - /* Read the MII 1000Base-T Control Register (Address 9). */ - ret_val = e1000_read_phy_reg(hw, PHY_1000T_CTRL, &mii_1000t_ctrl_reg); - if(ret_val) - return ret_val; + /* Read the MII 1000Base-T Control Register (Address 9). */ + ret_val = e1000_read_phy_reg(hw, PHY_1000T_CTRL, &mii_1000t_ctrl_reg); + if(ret_val) + return ret_val; /* Need to parse both autoneg_advertised and fc and set up * the appropriate PHY registers. First we will parse for @@ -1929,14 +1945,19 @@ e1000_phy_force_speed_duplex(struct e1000_hw *hw) void e1000_config_collision_dist(struct e1000_hw *hw) { - uint32_t tctl; + uint32_t tctl, coll_dist; DEBUGFUNC("e1000_config_collision_dist"); + if (hw->mac_type < e1000_82543) + coll_dist = E1000_COLLISION_DISTANCE_82542; + else + coll_dist = E1000_COLLISION_DISTANCE; + tctl = E1000_READ_REG(hw, TCTL); tctl &= ~E1000_TCTL_COLD; - tctl |= E1000_COLLISION_DISTANCE << E1000_COLD_SHIFT; + tctl |= coll_dist << E1000_COLD_SHIFT; E1000_WRITE_REG(hw, TCTL, tctl); E1000_WRITE_FLUSH(hw); @@ -2982,6 +3003,8 @@ e1000_phy_hw_reset(struct e1000_hw *hw) if (hw->mac_type < e1000_82571) msec_delay(10); + else + udelay(100); E1000_WRITE_REG(hw, CTRL, ctrl); E1000_WRITE_FLUSH(hw); @@ -3881,17 +3904,19 @@ e1000_read_eeprom(struct e1000_hw *hw, return -E1000_ERR_EEPROM; } - /* FLASH reads without acquiring the semaphore are safe in 82573-based - * controllers. - */ - if ((e1000_is_onboard_nvm_eeprom(hw) == TRUE) || - (hw->mac_type != e1000_82573)) { - /* Prepare the EEPROM for reading */ - if(e1000_acquire_eeprom(hw) != E1000_SUCCESS) - return -E1000_ERR_EEPROM; + /* FLASH reads without acquiring the semaphore are safe */ + if (e1000_is_onboard_nvm_eeprom(hw) == TRUE && + hw->eeprom.use_eerd == FALSE) { + switch (hw->mac_type) { + default: + /* Prepare the EEPROM for reading */ + if (e1000_acquire_eeprom(hw) != E1000_SUCCESS) + return -E1000_ERR_EEPROM; + break; + } } - if(eeprom->use_eerd == TRUE) { + if (eeprom->use_eerd == TRUE) { ret_val = e1000_read_eeprom_eerd(hw, offset, words, data); if ((e1000_is_onboard_nvm_eeprom(hw) == TRUE) || (hw->mac_type != e1000_82573)) @@ -4398,7 +4423,7 @@ e1000_commit_shadow_ram(struct e1000_hw *hw) return -E1000_ERR_EEPROM; } - /* If STM opcode located in bits 15:8 of flop, reset firmware */ + /* If STM opcode located in bits 15:8 of flop, reset firmware */ if ((flop & 0xFF00) == E1000_STM_OPCODE) { E1000_WRITE_REG(hw, HICR, E1000_HICR_FW_RESET); } @@ -4406,7 +4431,7 @@ e1000_commit_shadow_ram(struct e1000_hw *hw) /* Perform the flash update */ E1000_WRITE_REG(hw, EECD, eecd | E1000_EECD_FLUPD); - for (i=0; i < attempts; i++) { + for (i=0; i < attempts; i++) { eecd = E1000_READ_REG(hw, EECD); if ((eecd & E1000_EECD_FLUPD) == 0) { break; @@ -4479,6 +4504,7 @@ e1000_read_mac_addr(struct e1000_hw * hw) hw->perm_mac_addr[i] = (uint8_t) (eeprom_data & 0x00FF); hw->perm_mac_addr[i+1] = (uint8_t) (eeprom_data >> 8); } + switch (hw->mac_type) { default: break; @@ -6720,6 +6746,12 @@ e1000_get_phy_cfg_done(struct e1000_hw *hw) break; } + /* PHY configuration from NVM just starts after EECD_AUTO_RD sets to high. + * Need to wait for PHY configuration completion before accessing NVM + * and PHY. */ + if (hw->mac_type == e1000_82573) + msec_delay(25); + return E1000_SUCCESS; } @@ -6809,7 +6841,8 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw) { uint32_t manc = 0; - if(hw->mac_type > e1000_82547_rev_2) + + if (hw->mac_type > e1000_82547_rev_2) manc = E1000_READ_REG(hw, MANC); return (manc & E1000_MANC_BLK_PHY_RST_ON_IDE) ? E1000_BLK_PHY_RESET : E1000_SUCCESS; diff --git a/drivers/net/e1000/e1000_hw.h b/drivers/net/e1000/e1000_hw.h index 7caa35748ce..f1219dd9dba 100644 --- a/drivers/net/e1000/e1000_hw.h +++ b/drivers/net/e1000/e1000_hw.h @@ -377,6 +377,7 @@ int32_t e1000_swfw_sync_acquire(struct e1000_hw *hw, uint16_t mask); void e1000_swfw_sync_release(struct e1000_hw *hw, uint16_t mask); /* Filters (multicast, vlan, receive) */ +void e1000_mc_addr_list_update(struct e1000_hw *hw, uint8_t * mc_addr_list, uint32_t mc_addr_count, uint32_t pad, uint32_t rar_used_count); uint32_t e1000_hash_mc_addr(struct e1000_hw *hw, uint8_t * mc_addr); void e1000_mta_set(struct e1000_hw *hw, uint32_t hash_value); void e1000_rar_set(struct e1000_hw *hw, uint8_t * mc_addr, uint32_t rar_index); @@ -401,7 +402,9 @@ void e1000_read_pci_cfg(struct e1000_hw *hw, uint32_t reg, uint16_t * value); void e1000_write_pci_cfg(struct e1000_hw *hw, uint32_t reg, uint16_t * value); /* Port I/O is only supported on 82544 and newer */ uint32_t e1000_io_read(struct e1000_hw *hw, unsigned long port); +uint32_t e1000_read_reg_io(struct e1000_hw *hw, uint32_t offset); void e1000_io_write(struct e1000_hw *hw, unsigned long port, uint32_t value); +void e1000_enable_pciex_master(struct e1000_hw *hw); int32_t e1000_disable_pciex_master(struct e1000_hw *hw); int32_t e1000_get_software_semaphore(struct e1000_hw *hw); void e1000_release_software_semaphore(struct e1000_hw *hw); @@ -439,6 +442,7 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw); #define E1000_DEV_ID_82546GB_FIBER 0x107A #define E1000_DEV_ID_82546GB_SERDES 0x107B #define E1000_DEV_ID_82546GB_PCIE 0x108A +#define E1000_DEV_ID_82546GB_QUAD_COPPER 0x1099 #define E1000_DEV_ID_82547EI 0x1019 #define E1000_DEV_ID_82571EB_COPPER 0x105E #define E1000_DEV_ID_82571EB_FIBER 0x105F @@ -449,6 +453,7 @@ int32_t e1000_check_phy_reset_block(struct e1000_hw *hw); #define E1000_DEV_ID_82573E 0x108B #define E1000_DEV_ID_82573E_IAMT 0x108C #define E1000_DEV_ID_82573L 0x109A +#define E1000_DEV_ID_82546GB_QUAD_COPPER_KSP3 0x10B5 #define NODE_ADDRESS_SIZE 6 @@ -897,14 +902,14 @@ struct e1000_ffvt_entry { #define E1000_TXDCTL 0x03828 /* TX Descriptor Control - RW */ #define E1000_TADV 0x0382C /* TX Interrupt Absolute Delay Val - RW */ #define E1000_TSPMT 0x03830 /* TCP Segmentation PAD & Min Threshold - RW */ -#define E1000_TARC0 0x03840 /* TX Arbitration Count (0) */ -#define E1000_TDBAL1 0x03900 /* TX Desc Base Address Low (1) - RW */ -#define E1000_TDBAH1 0x03904 /* TX Desc Base Address High (1) - RW */ -#define E1000_TDLEN1 0x03908 /* TX Desc Length (1) - RW */ -#define E1000_TDH1 0x03910 /* TX Desc Head (1) - RW */ -#define E1000_TDT1 0x03918 /* TX Desc Tail (1) - RW */ -#define E1000_TXDCTL1 0x03928 /* TX Descriptor Control (1) - RW */ -#define E1000_TARC1 0x03940 /* TX Arbitration Count (1) */ +#define E1000_TARC0 0x03840 /* TX Arbitration Count (0) */ +#define E1000_TDBAL1 0x03900 /* TX Desc Base Address Low (1) - RW */ +#define E1000_TDBAH1 0x03904 /* TX Desc Base Address High (1) - RW */ +#define E1000_TDLEN1 0x03908 /* TX Desc Length (1) - RW */ +#define E1000_TDH1 0x03910 /* TX Desc Head (1) - RW */ +#define E1000_TDT1 0x03918 /* TX Desc Tail (1) - RW */ +#define E1000_TXDCTL1 0x03928 /* TX Descriptor Control (1) - RW */ +#define E1000_TARC1 0x03940 /* TX Arbitration Count (1) */ #define E1000_CRCERRS 0x04000 /* CRC Error Count - R/clr */ #define E1000_ALGNERRC 0x04004 /* Alignment Error Count - R/clr */ #define E1000_SYMERRS 0x04008 /* Symbol Error Count - R/clr */ @@ -1497,6 +1502,7 @@ struct e1000_hw { #define E1000_CTRL_EXT_EE_RST 0x00002000 /* Reinitialize from EEPROM */ #define E1000_CTRL_EXT_IPS 0x00004000 /* Invert Power State */ #define E1000_CTRL_EXT_SPD_BYPS 0x00008000 /* Speed Select Bypass */ +#define E1000_CTRL_EXT_RO_DIS 0x00020000 /* Relaxed Ordering disable */ #define E1000_CTRL_EXT_LINK_MODE_MASK 0x00C00000 #define E1000_CTRL_EXT_LINK_MODE_GMII 0x00000000 #define E1000_CTRL_EXT_LINK_MODE_TBI 0x00C00000 @@ -1758,7 +1764,6 @@ struct e1000_hw { #define E1000_TXDCTL_FULL_TX_DESC_WB 0x01010000 /* GRAN=1, WTHRESH=1 */ #define E1000_TXDCTL_COUNT_DESC 0x00400000 /* Enable the counting of desc. still to be processed. */ - /* Transmit Configuration Word */ #define E1000_TXCW_FD 0x00000020 /* TXCW full duplex */ #define E1000_TXCW_HD 0x00000040 /* TXCW half duplex */ @@ -1954,6 +1959,23 @@ struct e1000_host_command_info { #define E1000_MDALIGN 4096 +/* PCI-Ex registers */ + +/* PCI-Ex Control Register */ +#define E1000_GCR_RXD_NO_SNOOP 0x00000001 +#define E1000_GCR_RXDSCW_NO_SNOOP 0x00000002 +#define E1000_GCR_RXDSCR_NO_SNOOP 0x00000004 +#define E1000_GCR_TXD_NO_SNOOP 0x00000008 +#define E1000_GCR_TXDSCW_NO_SNOOP 0x00000010 +#define E1000_GCR_TXDSCR_NO_SNOOP 0x00000020 + +#define PCI_EX_NO_SNOOP_ALL (E1000_GCR_RXD_NO_SNOOP | \ + E1000_GCR_RXDSCW_NO_SNOOP | \ + E1000_GCR_RXDSCR_NO_SNOOP | \ + E1000_GCR TXD_NO_SNOOP | \ + E1000_GCR_TXDSCW_NO_SNOOP | \ + E1000_GCR_TXDSCR_NO_SNOOP) + #define E1000_GCR_L1_ACT_WITHOUT_L0S_RX 0x08000000 /* Function Active and Power State to MNG */ #define E1000_FACTPS_FUNC0_POWER_STATE_MASK 0x00000003 @@ -2077,7 +2099,10 @@ struct e1000_host_command_info { /* Collision related configuration parameters */ #define E1000_COLLISION_THRESHOLD 15 #define E1000_CT_SHIFT 4 -#define E1000_COLLISION_DISTANCE 64 +/* Collision distance is a 0-based value that applies to + * half-duplex-capable hardware only. */ +#define E1000_COLLISION_DISTANCE 63 +#define E1000_COLLISION_DISTANCE_82542 64 #define E1000_FDX_COLLISION_DISTANCE E1000_COLLISION_DISTANCE #define E1000_HDX_COLLISION_DISTANCE E1000_COLLISION_DISTANCE #define E1000_COLD_SHIFT 12 diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c index 438a931fd55..31e332935e5 100644 --- a/drivers/net/e1000/e1000_main.c +++ b/drivers/net/e1000/e1000_main.c @@ -29,11 +29,71 @@ #include "e1000.h" /* Change Log - * 6.0.58 4/20/05 - * o Accepted ethtool cleanup patch from Stephen Hemminger - * 6.0.44+ 2/15/05 - * o applied Anton's patch to resolve tx hang in hardware - * o Applied Andrew Mortons patch - e1000 stops working after resume + * 6.3.9 12/16/2005 + * o incorporate fix for recycled skbs from IBM LTC + * 6.3.7 11/18/2005 + * o Honor eeprom setting for enabling/disabling Wake On Lan + * 6.3.5 11/17/2005 + * o Fix memory leak in rx ring handling for PCI Express adapters + * 6.3.4 11/8/05 + * o Patch from Jesper Juhl to remove redundant NULL checks for kfree + * 6.3.2 9/20/05 + * o Render logic that sets/resets DRV_LOAD as inline functions to + * avoid code replication. If f/w is AMT then set DRV_LOAD only when + * network interface is open. + * o Handle DRV_LOAD set/reset in cases where AMT uses VLANs. + * o Adjust PBA partioning for Jumbo frames using MTU size and not + * rx_buffer_len + * 6.3.1 9/19/05 + * o Use adapter->tx_timeout_factor in Tx Hung Detect logic + (e1000_clean_tx_irq) + * o Support for 8086:10B5 device (Quad Port) + * 6.2.14 9/15/05 + * o In AMT enabled configurations, set/reset DRV_LOAD bit on interface + * open/close + * 6.2.13 9/14/05 + * o Invoke e1000_check_mng_mode only for 8257x controllers since it + * accesses the FWSM that is not supported in other controllers + * 6.2.12 9/9/05 + * o Add support for device id E1000_DEV_ID_82546GB_QUAD_COPPER + * o set RCTL:SECRC only for controllers newer than 82543. + * o When the n/w interface comes down reset DRV_LOAD bit to notify f/w. + * This code was moved from e1000_remove to e1000_close + * 6.2.10 9/6/05 + * o Fix error in updating RDT in el1000_alloc_rx_buffers[_ps] -- one off. + * o Enable fc by default on 82573 controllers (do not read eeprom) + * o Fix rx_errors statistic not to include missed_packet_count + * o Fix rx_dropped statistic not to include missed_packet_count + (Padraig Brady) + * 6.2.9 8/30/05 + * o Remove call to update statistics from the controller ib e1000_get_stats + * 6.2.8 8/30/05 + * o Improved algorithm for rx buffer allocation/rdt update + * o Flow control watermarks relative to rx PBA size + * o Simplified 'Tx Hung' detect logic + * 6.2.7 8/17/05 + * o Report rx buffer allocation failures and tx timeout counts in stats + * 6.2.6 8/16/05 + * o Implement workaround for controller erratum -- linear non-tso packet + * following a TSO gets written back prematurely + * 6.2.5 8/15/05 + * o Set netdev->tx_queue_len based on link speed/duplex settings. + * o Fix net_stats.rx_fifo_errors <p@draigBrady.com> + * o Do not power off PHY if SoL/IDER session is active + * 6.2.4 8/10/05 + * o Fix loopback test setup/cleanup for 82571/3 controllers + * o Fix parsing of outgoing packets (e1000_transfer_dhcp_info) to treat + * all packets as raw + * o Prevent operations that will cause the PHY to be reset if SoL/IDER + * sessions are active and log a message + * 6.2.2 7/21/05 + * o used fixed size descriptors for all MTU sizes, reduces memory load + * 6.1.2 4/13/05 + * o Fixed ethtool diagnostics + * o Enabled flow control to take default eeprom settings + * o Added stats_lock around e1000_read_phy_reg commands to avoid concurrent + * calls, one from mii_ioctl and other from within update_stats while + * processing MIIREG ioctl. */ char e1000_driver_name[] = "e1000"; @@ -43,7 +103,7 @@ static char e1000_driver_string[] = "Intel(R) PRO/1000 Network Driver"; #else #define DRIVERNAPI "-NAPI" #endif -#define DRV_VERSION "6.1.16-k2"DRIVERNAPI +#define DRV_VERSION "6.3.9-k2"DRIVERNAPI char e1000_driver_version[] = DRV_VERSION; static char e1000_copyright[] = "Copyright (c) 1999-2005 Intel Corporation."; @@ -97,7 +157,9 @@ static struct pci_device_id e1000_pci_tbl[] = { INTEL_E1000_ETHERNET_DEVICE(0x108A), INTEL_E1000_ETHERNET_DEVICE(0x108B), INTEL_E1000_ETHERNET_DEVICE(0x108C), + INTEL_E1000_ETHERNET_DEVICE(0x1099), INTEL_E1000_ETHERNET_DEVICE(0x109A), + INTEL_E1000_ETHERNET_DEVICE(0x10B5), /* required last entry */ {0,} }; @@ -171,9 +233,11 @@ static boolean_t e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, struct e1000_rx_ring *rx_ring); #endif static void e1000_alloc_rx_buffers(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring, + int cleaned_count); static void e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring); + struct e1000_rx_ring *rx_ring, + int cleaned_count); static int e1000_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd); static int e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd); @@ -291,7 +355,7 @@ e1000_irq_disable(struct e1000_adapter *adapter) static inline void e1000_irq_enable(struct e1000_adapter *adapter) { - if(likely(atomic_dec_and_test(&adapter->irq_sem))) { + if (likely(atomic_dec_and_test(&adapter->irq_sem))) { E1000_WRITE_REG(&adapter->hw, IMS, IMS_ENABLE_MASK); E1000_WRITE_FLUSH(&adapter->hw); } @@ -303,23 +367,91 @@ e1000_update_mng_vlan(struct e1000_adapter *adapter) struct net_device *netdev = adapter->netdev; uint16_t vid = adapter->hw.mng_cookie.vlan_id; uint16_t old_vid = adapter->mng_vlan_id; - if(adapter->vlgrp) { - if(!adapter->vlgrp->vlan_devices[vid]) { - if(adapter->hw.mng_cookie.status & + if (adapter->vlgrp) { + if (!adapter->vlgrp->vlan_devices[vid]) { + if (adapter->hw.mng_cookie.status & E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) { e1000_vlan_rx_add_vid(netdev, vid); adapter->mng_vlan_id = vid; } else adapter->mng_vlan_id = E1000_MNG_VLAN_NONE; - - if((old_vid != (uint16_t)E1000_MNG_VLAN_NONE) && - (vid != old_vid) && + + if ((old_vid != (uint16_t)E1000_MNG_VLAN_NONE) && + (vid != old_vid) && !adapter->vlgrp->vlan_devices[old_vid]) e1000_vlan_rx_kill_vid(netdev, old_vid); } } } - + +/** + * e1000_release_hw_control - release control of the h/w to f/w + * @adapter: address of board private structure + * + * e1000_release_hw_control resets {CTRL_EXT|FWSM}:DRV_LOAD bit. + * For ASF and Pass Through versions of f/w this means that the + * driver is no longer loaded. For AMT version (only with 82573) i + * of the f/w this means that the netowrk i/f is closed. + * + **/ + +static inline void +e1000_release_hw_control(struct e1000_adapter *adapter) +{ + uint32_t ctrl_ext; + uint32_t swsm; + + /* Let firmware taken over control of h/w */ + switch (adapter->hw.mac_type) { + case e1000_82571: + case e1000_82572: + ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); + E1000_WRITE_REG(&adapter->hw, CTRL_EXT, + ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD); + break; + case e1000_82573: + swsm = E1000_READ_REG(&adapter->hw, SWSM); + E1000_WRITE_REG(&adapter->hw, SWSM, + swsm & ~E1000_SWSM_DRV_LOAD); + default: + break; + } +} + +/** + * e1000_get_hw_control - get control of the h/w from f/w + * @adapter: address of board private structure + * + * e1000_get_hw_control sets {CTRL_EXT|FWSM}:DRV_LOAD bit. + * For ASF and Pass Through versions of f/w this means that + * the driver is loaded. For AMT version (only with 82573) + * of the f/w this means that the netowrk i/f is open. + * + **/ + +static inline void +e1000_get_hw_control(struct e1000_adapter *adapter) +{ + uint32_t ctrl_ext; + uint32_t swsm; + /* Let firmware know the driver has taken over */ + switch (adapter->hw.mac_type) { + case e1000_82571: + case e1000_82572: + ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); + E1000_WRITE_REG(&adapter->hw, CTRL_EXT, + ctrl_ext | E1000_CTRL_EXT_DRV_LOAD); + break; + case e1000_82573: + swsm = E1000_READ_REG(&adapter->hw, SWSM); + E1000_WRITE_REG(&adapter->hw, SWSM, + swsm | E1000_SWSM_DRV_LOAD); + break; + default: + break; + } +} + int e1000_up(struct e1000_adapter *adapter) { @@ -329,10 +461,10 @@ e1000_up(struct e1000_adapter *adapter) /* hardware has been reset, we need to reload some things */ /* Reset the PHY if it was previously powered down */ - if(adapter->hw.media_type == e1000_media_type_copper) { + if (adapter->hw.media_type == e1000_media_type_copper) { uint16_t mii_reg; e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &mii_reg); - if(mii_reg & MII_CR_POWER_DOWN) + if (mii_reg & MII_CR_POWER_DOWN) e1000_phy_reset(&adapter->hw); } @@ -343,20 +475,26 @@ e1000_up(struct e1000_adapter *adapter) e1000_configure_tx(adapter); e1000_setup_rctl(adapter); e1000_configure_rx(adapter); - for (i = 0; i < adapter->num_queues; i++) - adapter->alloc_rx_buf(adapter, &adapter->rx_ring[i]); + /* call E1000_DESC_UNUSED which always leaves + * at least 1 descriptor unused to make sure + * next_to_use != next_to_clean */ + for (i = 0; i < adapter->num_rx_queues; i++) { + struct e1000_rx_ring *ring = &adapter->rx_ring[i]; + adapter->alloc_rx_buf(adapter, ring, + E1000_DESC_UNUSED(ring)); + } #ifdef CONFIG_PCI_MSI - if(adapter->hw.mac_type > e1000_82547_rev_2) { + if (adapter->hw.mac_type > e1000_82547_rev_2) { adapter->have_msi = TRUE; - if((err = pci_enable_msi(adapter->pdev))) { + if ((err = pci_enable_msi(adapter->pdev))) { DPRINTK(PROBE, ERR, "Unable to allocate MSI interrupt Error: %d\n", err); adapter->have_msi = FALSE; } } #endif - if((err = request_irq(adapter->pdev->irq, &e1000_intr, + if ((err = request_irq(adapter->pdev->irq, &e1000_intr, SA_SHIRQ | SA_SAMPLE_RANDOM, netdev->name, netdev))) { DPRINTK(PROBE, ERR, @@ -364,6 +502,12 @@ e1000_up(struct e1000_adapter *adapter) return err; } +#ifdef CONFIG_E1000_MQ + e1000_setup_queue_mapping(adapter); +#endif + + adapter->tx_queue_len = netdev->tx_queue_len; + mod_timer(&adapter->watchdog_timer, jiffies); #ifdef CONFIG_E1000_NAPI @@ -378,6 +522,8 @@ void e1000_down(struct e1000_adapter *adapter) { struct net_device *netdev = adapter->netdev; + boolean_t mng_mode_enabled = (adapter->hw.mac_type >= e1000_82571) && + e1000_check_mng_mode(&adapter->hw); e1000_irq_disable(adapter); #ifdef CONFIG_E1000_MQ @@ -385,7 +531,7 @@ e1000_down(struct e1000_adapter *adapter) #endif free_irq(adapter->pdev->irq, netdev); #ifdef CONFIG_PCI_MSI - if(adapter->hw.mac_type > e1000_82547_rev_2 && + if (adapter->hw.mac_type > e1000_82547_rev_2 && adapter->have_msi == TRUE) pci_disable_msi(adapter->pdev); #endif @@ -396,6 +542,7 @@ e1000_down(struct e1000_adapter *adapter) #ifdef CONFIG_E1000_NAPI netif_poll_disable(netdev); #endif + netdev->tx_queue_len = adapter->tx_queue_len; adapter->link_speed = 0; adapter->link_duplex = 0; netif_carrier_off(netdev); @@ -405,12 +552,16 @@ e1000_down(struct e1000_adapter *adapter) e1000_clean_all_tx_rings(adapter); e1000_clean_all_rx_rings(adapter); - /* If WoL is not enabled and management mode is not IAMT - * Power down the PHY so no link is implied when interface is down */ - if(!adapter->wol && adapter->hw.mac_type >= e1000_82540 && + /* Power down the PHY so no link is implied when interface is down * + * The PHY cannot be powered down if any of the following is TRUE * + * (a) WoL is enabled + * (b) AMT is active + * (c) SoL/IDER session is active */ + if (!adapter->wol && adapter->hw.mac_type >= e1000_82540 && adapter->hw.media_type == e1000_media_type_copper && - !e1000_check_mng_mode(&adapter->hw) && - !(E1000_READ_REG(&adapter->hw, MANC) & E1000_MANC_SMBUS_EN)) { + !(E1000_READ_REG(&adapter->hw, MANC) & E1000_MANC_SMBUS_EN) && + !mng_mode_enabled && + !e1000_check_phy_reset_block(&adapter->hw)) { uint16_t mii_reg; e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &mii_reg); mii_reg |= MII_CR_POWER_DOWN; @@ -422,10 +573,8 @@ e1000_down(struct e1000_adapter *adapter) void e1000_reset(struct e1000_adapter *adapter) { - struct net_device *netdev = adapter->netdev; uint32_t pba, manc; uint16_t fc_high_water_mark = E1000_FC_HIGH_DIFF; - uint16_t fc_low_water_mark = E1000_FC_LOW_DIFF; /* Repartition Pba for greater than 9k mtu * To take effect CTRL.RST is required. @@ -448,19 +597,12 @@ e1000_reset(struct e1000_adapter *adapter) break; } - if((adapter->hw.mac_type != e1000_82573) && - (adapter->rx_buffer_len > E1000_RXBUFFER_8192)) { + if ((adapter->hw.mac_type != e1000_82573) && + (adapter->netdev->mtu > E1000_RXBUFFER_8192)) pba -= 8; /* allocate more FIFO for Tx */ - /* send an XOFF when there is enough space in the - * Rx FIFO to hold one extra full size Rx packet - */ - fc_high_water_mark = netdev->mtu + ENET_HEADER_SIZE + - ETHERNET_FCS_SIZE + 1; - fc_low_water_mark = fc_high_water_mark + 8; - } - if(adapter->hw.mac_type == e1000_82547) { + if (adapter->hw.mac_type == e1000_82547) { adapter->tx_fifo_head = 0; adapter->tx_head_addr = pba << E1000_TX_HEAD_ADDR_SHIFT; adapter->tx_fifo_size = @@ -471,19 +613,21 @@ e1000_reset(struct e1000_adapter *adapter) E1000_WRITE_REG(&adapter->hw, PBA, pba); /* flow control settings */ - adapter->hw.fc_high_water = (pba << E1000_PBA_BYTES_SHIFT) - - fc_high_water_mark; - adapter->hw.fc_low_water = (pba << E1000_PBA_BYTES_SHIFT) - - fc_low_water_mark; + /* Set the FC high water mark to 90% of the FIFO size. + * Required to clear last 3 LSB */ + fc_high_water_mark = ((pba * 9216)/10) & 0xFFF8; + + adapter->hw.fc_high_water = fc_high_water_mark; + adapter->hw.fc_low_water = fc_high_water_mark - 8; adapter->hw.fc_pause_time = E1000_FC_PAUSE_TIME; adapter->hw.fc_send_xon = 1; adapter->hw.fc = adapter->hw.original_fc; /* Allow time for pending master requests to run */ e1000_reset_hw(&adapter->hw); - if(adapter->hw.mac_type >= e1000_82544) + if (adapter->hw.mac_type >= e1000_82544) E1000_WRITE_REG(&adapter->hw, WUC, 0); - if(e1000_init_hw(&adapter->hw)) + if (e1000_init_hw(&adapter->hw)) DPRINTK(PROBE, ERR, "Hardware Error\n"); e1000_update_mng_vlan(adapter); /* Enable h/w to recognize an 802.1Q VLAN Ethernet packet */ @@ -517,33 +661,31 @@ e1000_probe(struct pci_dev *pdev, struct net_device *netdev; struct e1000_adapter *adapter; unsigned long mmio_start, mmio_len; - uint32_t ctrl_ext; - uint32_t swsm; static int cards_found = 0; int i, err, pci_using_dac; uint16_t eeprom_data; uint16_t eeprom_apme_mask = E1000_EEPROM_APME; - if((err = pci_enable_device(pdev))) + if ((err = pci_enable_device(pdev))) return err; - if(!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) { + if (!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) { pci_using_dac = 1; } else { - if((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK))) { + if ((err = pci_set_dma_mask(pdev, DMA_32BIT_MASK))) { E1000_ERR("No usable DMA configuration, aborting\n"); return err; } pci_using_dac = 0; } - if((err = pci_request_regions(pdev, e1000_driver_name))) + if ((err = pci_request_regions(pdev, e1000_driver_name))) return err; pci_set_master(pdev); netdev = alloc_etherdev(sizeof(struct e1000_adapter)); - if(!netdev) { + if (!netdev) { err = -ENOMEM; goto err_alloc_etherdev; } @@ -562,15 +704,15 @@ e1000_probe(struct pci_dev *pdev, mmio_len = pci_resource_len(pdev, BAR_0); adapter->hw.hw_addr = ioremap(mmio_start, mmio_len); - if(!adapter->hw.hw_addr) { + if (!adapter->hw.hw_addr) { err = -EIO; goto err_ioremap; } - for(i = BAR_1; i <= BAR_5; i++) { - if(pci_resource_len(pdev, i) == 0) + for (i = BAR_1; i <= BAR_5; i++) { + if (pci_resource_len(pdev, i) == 0) continue; - if(pci_resource_flags(pdev, i) & IORESOURCE_IO) { + if (pci_resource_flags(pdev, i) & IORESOURCE_IO) { adapter->hw.io_base = pci_resource_start(pdev, i); break; } @@ -607,13 +749,13 @@ e1000_probe(struct pci_dev *pdev, /* setup the private structure */ - if((err = e1000_sw_init(adapter))) + if ((err = e1000_sw_init(adapter))) goto err_sw_init; - if((err = e1000_check_phy_reset_block(&adapter->hw))) + if ((err = e1000_check_phy_reset_block(&adapter->hw))) DPRINTK(PROBE, INFO, "PHY reset is blocked due to SOL/IDER session.\n"); - if(adapter->hw.mac_type >= e1000_82543) { + if (adapter->hw.mac_type >= e1000_82543) { netdev->features = NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_HW_VLAN_TX | @@ -622,16 +764,16 @@ e1000_probe(struct pci_dev *pdev, } #ifdef NETIF_F_TSO - if((adapter->hw.mac_type >= e1000_82544) && + if ((adapter->hw.mac_type >= e1000_82544) && (adapter->hw.mac_type != e1000_82547)) netdev->features |= NETIF_F_TSO; #ifdef NETIF_F_TSO_IPV6 - if(adapter->hw.mac_type > e1000_82547_rev_2) + if (adapter->hw.mac_type > e1000_82547_rev_2) netdev->features |= NETIF_F_TSO_IPV6; #endif #endif - if(pci_using_dac) + if (pci_using_dac) netdev->features |= NETIF_F_HIGHDMA; /* hard_start_xmit is safe against parallel locking */ @@ -639,14 +781,14 @@ e1000_probe(struct pci_dev *pdev, adapter->en_mng_pt = e1000_enable_mng_pass_thru(&adapter->hw); - /* before reading the EEPROM, reset the controller to + /* before reading the EEPROM, reset the controller to * put the device in a known good starting state */ - + e1000_reset_hw(&adapter->hw); /* make sure the EEPROM is good */ - if(e1000_validate_eeprom_checksum(&adapter->hw) < 0) { + if (e1000_validate_eeprom_checksum(&adapter->hw) < 0) { DPRINTK(PROBE, ERR, "The EEPROM Checksum Is Not Valid\n"); err = -EIO; goto err_eeprom; @@ -654,12 +796,12 @@ e1000_probe(struct pci_dev *pdev, /* copy the MAC address out of the EEPROM */ - if(e1000_read_mac_addr(&adapter->hw)) + if (e1000_read_mac_addr(&adapter->hw)) DPRINTK(PROBE, ERR, "EEPROM Read Error\n"); memcpy(netdev->dev_addr, adapter->hw.mac_addr, netdev->addr_len); memcpy(netdev->perm_addr, adapter->hw.mac_addr, netdev->addr_len); - if(!is_valid_ether_addr(netdev->perm_addr)) { + if (!is_valid_ether_addr(netdev->perm_addr)) { DPRINTK(PROBE, ERR, "Invalid MAC Address\n"); err = -EIO; goto err_eeprom; @@ -699,7 +841,7 @@ e1000_probe(struct pci_dev *pdev, * enable the ACPI Magic Packet filter */ - switch(adapter->hw.mac_type) { + switch (adapter->hw.mac_type) { case e1000_82542_rev2_0: case e1000_82542_rev2_1: case e1000_82543: @@ -712,8 +854,7 @@ e1000_probe(struct pci_dev *pdev, case e1000_82546: case e1000_82546_rev_3: case e1000_82571: - if((E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1) - && (adapter->hw.media_type == e1000_media_type_copper)) { + if (E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_FUNC_1){ e1000_read_eeprom(&adapter->hw, EEPROM_INIT_CONTROL3_PORT_B, 1, &eeprom_data); break; @@ -724,31 +865,42 @@ e1000_probe(struct pci_dev *pdev, EEPROM_INIT_CONTROL3_PORT_A, 1, &eeprom_data); break; } - if(eeprom_data & eeprom_apme_mask) + if (eeprom_data & eeprom_apme_mask) adapter->wol |= E1000_WUFC_MAG; + /* print bus type/speed/width info */ + { + struct e1000_hw *hw = &adapter->hw; + DPRINTK(PROBE, INFO, "(PCI%s:%s:%s) ", + ((hw->bus_type == e1000_bus_type_pcix) ? "-X" : + (hw->bus_type == e1000_bus_type_pci_express ? " Express":"")), + ((hw->bus_speed == e1000_bus_speed_2500) ? "2.5Gb/s" : + (hw->bus_speed == e1000_bus_speed_133) ? "133MHz" : + (hw->bus_speed == e1000_bus_speed_120) ? "120MHz" : + (hw->bus_speed == e1000_bus_speed_100) ? "100MHz" : + (hw->bus_speed == e1000_bus_speed_66) ? "66MHz" : "33MHz"), + ((hw->bus_width == e1000_bus_width_64) ? "64-bit" : + (hw->bus_width == e1000_bus_width_pciex_4) ? "Width x4" : + (hw->bus_width == e1000_bus_width_pciex_1) ? "Width x1" : + "32-bit")); + } + + for (i = 0; i < 6; i++) + printk("%2.2x%c", netdev->dev_addr[i], i == 5 ? '\n' : ':'); + /* reset the hardware with the new settings */ e1000_reset(adapter); - /* Let firmware know the driver has taken over */ - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext | E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm | E1000_SWSM_DRV_LOAD); - break; - default: - break; - } + /* If the controller is 82573 and f/w is AMT, do not set + * DRV_LOAD until the interface is up. For all other cases, + * let the f/w know that the h/w is now under the control + * of the driver. */ + if (adapter->hw.mac_type != e1000_82573 || + !e1000_check_mng_mode(&adapter->hw)) + e1000_get_hw_control(adapter); strcpy(netdev->name, "eth%d"); - if((err = register_netdev(netdev))) + if ((err = register_netdev(netdev))) goto err_register; DPRINTK(PROBE, INFO, "Intel(R) PRO/1000 Network Connection\n"); @@ -782,47 +934,33 @@ e1000_remove(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); struct e1000_adapter *adapter = netdev_priv(netdev); - uint32_t ctrl_ext; - uint32_t manc, swsm; + uint32_t manc; #ifdef CONFIG_E1000_NAPI int i; #endif flush_scheduled_work(); - if(adapter->hw.mac_type >= e1000_82540 && + if (adapter->hw.mac_type >= e1000_82540 && adapter->hw.media_type == e1000_media_type_copper) { manc = E1000_READ_REG(&adapter->hw, MANC); - if(manc & E1000_MANC_SMBUS_EN) { + if (manc & E1000_MANC_SMBUS_EN) { manc |= E1000_MANC_ARP_EN; E1000_WRITE_REG(&adapter->hw, MANC, manc); } } - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm & ~E1000_SWSM_DRV_LOAD); - break; - - default: - break; - } + /* Release control of h/w to f/w. If f/w is AMT enabled, this + * would have already happened in close and is redundant. */ + e1000_release_hw_control(adapter); unregister_netdev(netdev); #ifdef CONFIG_E1000_NAPI - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) __dev_put(&adapter->polling_netdev[i]); #endif - if(!e1000_check_phy_reset_block(&adapter->hw)) + if (!e1000_check_phy_reset_block(&adapter->hw)) e1000_phy_hw_reset(&adapter->hw); kfree(adapter->tx_ring); @@ -881,19 +1019,19 @@ e1000_sw_init(struct e1000_adapter *adapter) /* identify the MAC */ - if(e1000_set_mac_type(hw)) { + if (e1000_set_mac_type(hw)) { DPRINTK(PROBE, ERR, "Unknown MAC Type\n"); return -EIO; } /* initialize eeprom parameters */ - if(e1000_init_eeprom_params(hw)) { + if (e1000_init_eeprom_params(hw)) { E1000_ERR("EEPROM initialization failed\n"); return -EIO; } - switch(hw->mac_type) { + switch (hw->mac_type) { default: break; case e1000_82541: @@ -912,7 +1050,7 @@ e1000_sw_init(struct e1000_adapter *adapter) /* Copper options */ - if(hw->media_type == e1000_media_type_copper) { + if (hw->media_type == e1000_media_type_copper) { hw->mdix = AUTO_ALL_MODES; hw->disable_polarity_correction = FALSE; hw->master_slave = E1000_MASTER_SLAVE; @@ -923,15 +1061,34 @@ e1000_sw_init(struct e1000_adapter *adapter) switch (hw->mac_type) { case e1000_82571: case e1000_82572: - adapter->num_queues = 2; + /* These controllers support 2 tx queues, but with a single + * qdisc implementation, multiple tx queues aren't quite as + * interesting. If we can find a logical way of mapping + * flows to a queue, then perhaps we can up the num_tx_queue + * count back to its default. Until then, we run the risk of + * terrible performance due to SACK overload. */ + adapter->num_tx_queues = 1; + adapter->num_rx_queues = 2; break; default: - adapter->num_queues = 1; + adapter->num_tx_queues = 1; + adapter->num_rx_queues = 1; break; } - adapter->num_queues = min(adapter->num_queues, num_online_cpus()); + adapter->num_rx_queues = min(adapter->num_rx_queues, num_online_cpus()); + adapter->num_tx_queues = min(adapter->num_tx_queues, num_online_cpus()); + DPRINTK(DRV, INFO, "Multiqueue Enabled: Rx Queue count = %u %s\n", + adapter->num_rx_queues, + ((adapter->num_rx_queues == 1) + ? ((num_online_cpus() > 1) + ? "(due to unsupported feature in current adapter)" + : "(due to unsupported system configuration)") + : "")); + DPRINTK(DRV, INFO, "Multiqueue Enabled: Tx Queue count = %u\n", + adapter->num_tx_queues); #else - adapter->num_queues = 1; + adapter->num_tx_queues = 1; + adapter->num_rx_queues = 1; #endif if (e1000_alloc_queues(adapter)) { @@ -940,17 +1097,14 @@ e1000_sw_init(struct e1000_adapter *adapter) } #ifdef CONFIG_E1000_NAPI - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_rx_queues; i++) { adapter->polling_netdev[i].priv = adapter; adapter->polling_netdev[i].poll = &e1000_clean; adapter->polling_netdev[i].weight = 64; dev_hold(&adapter->polling_netdev[i]); set_bit(__LINK_STATE_START, &adapter->polling_netdev[i].state); } -#endif - -#ifdef CONFIG_E1000_MQ - e1000_setup_queue_mapping(adapter); + spin_lock_init(&adapter->tx_queue_lock); #endif atomic_set(&adapter->irq_sem, 1); @@ -973,13 +1127,13 @@ e1000_alloc_queues(struct e1000_adapter *adapter) { int size; - size = sizeof(struct e1000_tx_ring) * adapter->num_queues; + size = sizeof(struct e1000_tx_ring) * adapter->num_tx_queues; adapter->tx_ring = kmalloc(size, GFP_KERNEL); if (!adapter->tx_ring) return -ENOMEM; memset(adapter->tx_ring, 0, size); - size = sizeof(struct e1000_rx_ring) * adapter->num_queues; + size = sizeof(struct e1000_rx_ring) * adapter->num_rx_queues; adapter->rx_ring = kmalloc(size, GFP_KERNEL); if (!adapter->rx_ring) { kfree(adapter->tx_ring); @@ -988,7 +1142,7 @@ e1000_alloc_queues(struct e1000_adapter *adapter) memset(adapter->rx_ring, 0, size); #ifdef CONFIG_E1000_NAPI - size = sizeof(struct net_device) * adapter->num_queues; + size = sizeof(struct net_device) * adapter->num_rx_queues; adapter->polling_netdev = kmalloc(size, GFP_KERNEL); if (!adapter->polling_netdev) { kfree(adapter->tx_ring); @@ -998,6 +1152,14 @@ e1000_alloc_queues(struct e1000_adapter *adapter) memset(adapter->polling_netdev, 0, size); #endif +#ifdef CONFIG_E1000_MQ + adapter->rx_sched_call_data.func = e1000_rx_schedule; + adapter->rx_sched_call_data.info = adapter->netdev; + + adapter->cpu_netdev = alloc_percpu(struct net_device *); + adapter->cpu_tx_ring = alloc_percpu(struct e1000_tx_ring *); +#endif + return E1000_SUCCESS; } @@ -1017,14 +1179,15 @@ e1000_setup_queue_mapping(struct e1000_adapter *adapter) lock_cpu_hotplug(); i = 0; for_each_online_cpu(cpu) { - *per_cpu_ptr(adapter->cpu_tx_ring, cpu) = &adapter->tx_ring[i % adapter->num_queues]; + *per_cpu_ptr(adapter->cpu_tx_ring, cpu) = &adapter->tx_ring[i % adapter->num_tx_queues]; /* This is incomplete because we'd like to assign separate * physical cpus to these netdev polling structures and * avoid saturating a subset of cpus. */ - if (i < adapter->num_queues) { + if (i < adapter->num_rx_queues) { *per_cpu_ptr(adapter->cpu_netdev, cpu) = &adapter->polling_netdev[i]; - adapter->cpu_for_queue[i] = cpu; + adapter->rx_ring[i].cpu = cpu; + cpu_set(cpu, adapter->cpumask); } else *per_cpu_ptr(adapter->cpu_netdev, cpu) = NULL; @@ -1063,14 +1226,20 @@ e1000_open(struct net_device *netdev) if ((err = e1000_setup_all_rx_resources(adapter))) goto err_setup_rx; - if((err = e1000_up(adapter))) + if ((err = e1000_up(adapter))) goto err_up; adapter->mng_vlan_id = E1000_MNG_VLAN_NONE; - if((adapter->hw.mng_cookie.status & + if ((adapter->hw.mng_cookie.status & E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) { e1000_update_mng_vlan(adapter); } + /* If AMT is enabled, let the firmware know that the network + * interface is now open */ + if (adapter->hw.mac_type == e1000_82573 && + e1000_check_mng_mode(&adapter->hw)) + e1000_get_hw_control(adapter); + return E1000_SUCCESS; err_up: @@ -1105,10 +1274,17 @@ e1000_close(struct net_device *netdev) e1000_free_all_tx_resources(adapter); e1000_free_all_rx_resources(adapter); - if((adapter->hw.mng_cookie.status & + if ((adapter->hw.mng_cookie.status & E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) { e1000_vlan_rx_kill_vid(netdev, adapter->mng_vlan_id); } + + /* If AMT is enabled, let the firmware know that the network + * interface is now closed */ + if (adapter->hw.mac_type == e1000_82573 && + e1000_check_mng_mode(&adapter->hw)) + e1000_release_hw_control(adapter); + return 0; } @@ -1153,7 +1329,7 @@ e1000_setup_tx_resources(struct e1000_adapter *adapter, size = sizeof(struct e1000_buffer) * txdr->count; txdr->buffer_info = vmalloc_node(size, pcibus_to_node(pdev->bus)); - if(!txdr->buffer_info) { + if (!txdr->buffer_info) { DPRINTK(PROBE, ERR, "Unable to allocate memory for the transmit descriptor ring\n"); return -ENOMEM; @@ -1166,7 +1342,7 @@ e1000_setup_tx_resources(struct e1000_adapter *adapter, E1000_ROUNDUP(txdr->size, 4096); txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma); - if(!txdr->desc) { + if (!txdr->desc) { setup_tx_desc_die: vfree(txdr->buffer_info); DPRINTK(PROBE, ERR, @@ -1182,8 +1358,8 @@ setup_tx_desc_die: "at %p\n", txdr->size, txdr->desc); /* Try again, without freeing the previous */ txdr->desc = pci_alloc_consistent(pdev, txdr->size, &txdr->dma); - if(!txdr->desc) { /* Failed allocation, critical failure */ + if (!txdr->desc) { pci_free_consistent(pdev, txdr->size, olddesc, olddma); goto setup_tx_desc_die; } @@ -1229,7 +1405,7 @@ e1000_setup_all_tx_resources(struct e1000_adapter *adapter) { int i, err = 0; - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_tx_queues; i++) { err = e1000_setup_tx_resources(adapter, &adapter->tx_ring[i]); if (err) { DPRINTK(PROBE, ERR, @@ -1254,10 +1430,11 @@ e1000_configure_tx(struct e1000_adapter *adapter) uint64_t tdba; struct e1000_hw *hw = &adapter->hw; uint32_t tdlen, tctl, tipg, tarc; + uint32_t ipgr1, ipgr2; /* Setup the HW Tx Head and Tail descriptor pointers */ - switch (adapter->num_queues) { + switch (adapter->num_tx_queues) { case 2: tdba = adapter->tx_ring[1].dma; tdlen = adapter->tx_ring[1].count * @@ -1287,22 +1464,26 @@ e1000_configure_tx(struct e1000_adapter *adapter) /* Set the default values for the Tx Inter Packet Gap timer */ + if (hw->media_type == e1000_media_type_fiber || + hw->media_type == e1000_media_type_internal_serdes) + tipg = DEFAULT_82543_TIPG_IPGT_FIBER; + else + tipg = DEFAULT_82543_TIPG_IPGT_COPPER; + switch (hw->mac_type) { case e1000_82542_rev2_0: case e1000_82542_rev2_1: tipg = DEFAULT_82542_TIPG_IPGT; - tipg |= DEFAULT_82542_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT; - tipg |= DEFAULT_82542_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT; + ipgr1 = DEFAULT_82542_TIPG_IPGR1; + ipgr2 = DEFAULT_82542_TIPG_IPGR2; break; default: - if (hw->media_type == e1000_media_type_fiber || - hw->media_type == e1000_media_type_internal_serdes) - tipg = DEFAULT_82543_TIPG_IPGT_FIBER; - else - tipg = DEFAULT_82543_TIPG_IPGT_COPPER; - tipg |= DEFAULT_82543_TIPG_IPGR1 << E1000_TIPG_IPGR1_SHIFT; - tipg |= DEFAULT_82543_TIPG_IPGR2 << E1000_TIPG_IPGR2_SHIFT; + ipgr1 = DEFAULT_82543_TIPG_IPGR1; + ipgr2 = DEFAULT_82543_TIPG_IPGR2; + break; } + tipg |= ipgr1 << E1000_TIPG_IPGR1_SHIFT; + tipg |= ipgr2 << E1000_TIPG_IPGR2_SHIFT; E1000_WRITE_REG(hw, TIPG, tipg); /* Set the Tx Interrupt Delay register */ @@ -1378,7 +1559,7 @@ e1000_setup_rx_resources(struct e1000_adapter *adapter, size = sizeof(struct e1000_ps_page) * rxdr->count; rxdr->ps_page = kmalloc(size, GFP_KERNEL); - if(!rxdr->ps_page) { + if (!rxdr->ps_page) { vfree(rxdr->buffer_info); DPRINTK(PROBE, ERR, "Unable to allocate memory for the receive descriptor ring\n"); @@ -1388,7 +1569,7 @@ e1000_setup_rx_resources(struct e1000_adapter *adapter, size = sizeof(struct e1000_ps_page_dma) * rxdr->count; rxdr->ps_page_dma = kmalloc(size, GFP_KERNEL); - if(!rxdr->ps_page_dma) { + if (!rxdr->ps_page_dma) { vfree(rxdr->buffer_info); kfree(rxdr->ps_page); DPRINTK(PROBE, ERR, @@ -1397,7 +1578,7 @@ e1000_setup_rx_resources(struct e1000_adapter *adapter, } memset(rxdr->ps_page_dma, 0, size); - if(adapter->hw.mac_type <= e1000_82547_rev_2) + if (adapter->hw.mac_type <= e1000_82547_rev_2) desc_len = sizeof(struct e1000_rx_desc); else desc_len = sizeof(union e1000_rx_desc_packet_split); @@ -1454,6 +1635,8 @@ setup_rx_desc_die: rxdr->next_to_clean = 0; rxdr->next_to_use = 0; + rxdr->rx_skb_top = NULL; + rxdr->rx_skb_prev = NULL; return 0; } @@ -1475,7 +1658,7 @@ e1000_setup_all_rx_resources(struct e1000_adapter *adapter) { int i, err = 0; - for (i = 0; i < adapter->num_queues; i++) { + for (i = 0; i < adapter->num_rx_queues; i++) { err = e1000_setup_rx_resources(adapter, &adapter->rx_ring[i]); if (err) { DPRINTK(PROBE, ERR, @@ -1498,7 +1681,7 @@ e1000_setup_rctl(struct e1000_adapter *adapter) { uint32_t rctl, rfctl; uint32_t psrctl = 0; -#ifdef CONFIG_E1000_PACKET_SPLIT +#ifndef CONFIG_E1000_DISABLE_PACKET_SPLIT uint32_t pages = 0; #endif @@ -1510,7 +1693,10 @@ e1000_setup_rctl(struct e1000_adapter *adapter) E1000_RCTL_LBM_NO | E1000_RCTL_RDMTS_HALF | (adapter->hw.mc_filter_type << E1000_RCTL_MO_SHIFT); - if(adapter->hw.tbi_compatibility_on == 1) + if (adapter->hw.mac_type > e1000_82543) + rctl |= E1000_RCTL_SECRC; + + if (adapter->hw.tbi_compatibility_on == 1) rctl |= E1000_RCTL_SBP; else rctl &= ~E1000_RCTL_SBP; @@ -1521,32 +1707,17 @@ e1000_setup_rctl(struct e1000_adapter *adapter) rctl |= E1000_RCTL_LPE; /* Setup buffer sizes */ - if(adapter->hw.mac_type >= e1000_82571) { + if (adapter->hw.mac_type >= e1000_82571) { /* We can now specify buffers in 1K increments. * BSIZE and BSEX are ignored in this case. */ rctl |= adapter->rx_buffer_len << 0x11; } else { rctl &= ~E1000_RCTL_SZ_4096; - rctl |= E1000_RCTL_BSEX; - switch (adapter->rx_buffer_len) { - case E1000_RXBUFFER_2048: - default: - rctl |= E1000_RCTL_SZ_2048; - rctl &= ~E1000_RCTL_BSEX; - break; - case E1000_RXBUFFER_4096: - rctl |= E1000_RCTL_SZ_4096; - break; - case E1000_RXBUFFER_8192: - rctl |= E1000_RCTL_SZ_8192; - break; - case E1000_RXBUFFER_16384: - rctl |= E1000_RCTL_SZ_16384; - break; - } + rctl &= ~E1000_RCTL_BSEX; + rctl |= E1000_RCTL_SZ_2048; } -#ifdef CONFIG_E1000_PACKET_SPLIT +#ifndef CONFIG_E1000_DISABLE_PACKET_SPLIT /* 82571 and greater support packet-split where the protocol * header is placed in skb->data and the packet data is * placed in pages hanging off of skb_shinfo(skb)->nr_frags. @@ -1570,7 +1741,7 @@ e1000_setup_rctl(struct e1000_adapter *adapter) E1000_WRITE_REG(&adapter->hw, RFCTL, rfctl); rctl |= E1000_RCTL_DTYP_PS | E1000_RCTL_SECRC; - + psrctl |= adapter->rx_ps_bsize0 >> E1000_PSRCTL_BSIZE0_SHIFT; @@ -1632,22 +1803,27 @@ e1000_configure_rx(struct e1000_adapter *adapter) if (hw->mac_type >= e1000_82540) { E1000_WRITE_REG(hw, RADV, adapter->rx_abs_int_delay); - if(adapter->itr > 1) + if (adapter->itr > 1) E1000_WRITE_REG(hw, ITR, 1000000000 / (adapter->itr * 256)); } if (hw->mac_type >= e1000_82571) { - /* Reset delay timers after every interrupt */ ctrl_ext = E1000_READ_REG(hw, CTRL_EXT); + /* Reset delay timers after every interrupt */ ctrl_ext |= E1000_CTRL_EXT_CANC; +#ifdef CONFIG_E1000_NAPI + /* Auto-Mask interrupts upon ICR read. */ + ctrl_ext |= E1000_CTRL_EXT_IAME; +#endif E1000_WRITE_REG(hw, CTRL_EXT, ctrl_ext); + E1000_WRITE_REG(hw, IAM, ~0); E1000_WRITE_FLUSH(hw); } /* Setup the HW Rx Head and Tail Descriptor Pointers and * the Base and Length of the Rx Descriptor Ring */ - switch (adapter->num_queues) { + switch (adapter->num_rx_queues) { #ifdef CONFIG_E1000_MQ case 2: rdba = adapter->rx_ring[1].dma; @@ -1674,7 +1850,7 @@ e1000_configure_rx(struct e1000_adapter *adapter) } #ifdef CONFIG_E1000_MQ - if (adapter->num_queues > 1) { + if (adapter->num_rx_queues > 1) { uint32_t random[10]; get_random_bytes(&random[0], 40); @@ -1684,7 +1860,7 @@ e1000_configure_rx(struct e1000_adapter *adapter) E1000_WRITE_REG(hw, RSSIM, 0); } - switch (adapter->num_queues) { + switch (adapter->num_rx_queues) { case 2: default: reta = 0x00800080; @@ -1716,13 +1892,13 @@ e1000_configure_rx(struct e1000_adapter *adapter) /* Enable 82543 Receive Checksum Offload for TCP and UDP */ if (hw->mac_type >= e1000_82543) { rxcsum = E1000_READ_REG(hw, RXCSUM); - if(adapter->rx_csum == TRUE) { + if (adapter->rx_csum == TRUE) { rxcsum |= E1000_RXCSUM_TUOFL; /* Enable 82571 IPv4 payload checksum for UDP fragments * Must be used in conjunction with packet-split. */ - if ((hw->mac_type >= e1000_82571) && - (adapter->rx_ps_pages)) { + if ((hw->mac_type >= e1000_82571) && + (adapter->rx_ps_pages)) { rxcsum |= E1000_RXCSUM_IPPCSE; } } else { @@ -1776,7 +1952,7 @@ e1000_free_all_tx_resources(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_tx_queues; i++) e1000_free_tx_resources(adapter, &adapter->tx_ring[i]); } @@ -1784,17 +1960,15 @@ static inline void e1000_unmap_and_free_tx_resource(struct e1000_adapter *adapter, struct e1000_buffer *buffer_info) { - if(buffer_info->dma) { + if (buffer_info->dma) { pci_unmap_page(adapter->pdev, buffer_info->dma, buffer_info->length, PCI_DMA_TODEVICE); - buffer_info->dma = 0; } - if(buffer_info->skb) { + if (buffer_info->skb) dev_kfree_skb_any(buffer_info->skb); - buffer_info->skb = NULL; - } + memset(buffer_info, 0, sizeof(struct e1000_buffer)); } /** @@ -1813,7 +1987,7 @@ e1000_clean_tx_ring(struct e1000_adapter *adapter, /* Free all the Tx ring sk_buffs */ - for(i = 0; i < tx_ring->count; i++) { + for (i = 0; i < tx_ring->count; i++) { buffer_info = &tx_ring->buffer_info[i]; e1000_unmap_and_free_tx_resource(adapter, buffer_info); } @@ -1843,7 +2017,7 @@ e1000_clean_all_tx_rings(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_tx_queues; i++) e1000_clean_tx_ring(adapter, &adapter->tx_ring[i]); } @@ -1887,7 +2061,7 @@ e1000_free_all_rx_resources(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) e1000_free_rx_resources(adapter, &adapter->rx_ring[i]); } @@ -1909,12 +2083,9 @@ e1000_clean_rx_ring(struct e1000_adapter *adapter, unsigned int i, j; /* Free all the Rx ring sk_buffs */ - - for(i = 0; i < rx_ring->count; i++) { + for (i = 0; i < rx_ring->count; i++) { buffer_info = &rx_ring->buffer_info[i]; - if(buffer_info->skb) { - ps_page = &rx_ring->ps_page[i]; - ps_page_dma = &rx_ring->ps_page_dma[i]; + if (buffer_info->skb) { pci_unmap_single(pdev, buffer_info->dma, buffer_info->length, @@ -1922,19 +2093,30 @@ e1000_clean_rx_ring(struct e1000_adapter *adapter, dev_kfree_skb(buffer_info->skb); buffer_info->skb = NULL; - - for(j = 0; j < adapter->rx_ps_pages; j++) { - if(!ps_page->ps_page[j]) break; - pci_unmap_single(pdev, - ps_page_dma->ps_page_dma[j], - PAGE_SIZE, PCI_DMA_FROMDEVICE); - ps_page_dma->ps_page_dma[j] = 0; - put_page(ps_page->ps_page[j]); - ps_page->ps_page[j] = NULL; - } } + ps_page = &rx_ring->ps_page[i]; + ps_page_dma = &rx_ring->ps_page_dma[i]; + for (j = 0; j < adapter->rx_ps_pages; j++) { + if (!ps_page->ps_page[j]) break; + pci_unmap_page(pdev, + ps_page_dma->ps_page_dma[j], + PAGE_SIZE, PCI_DMA_FROMDEVICE); + ps_page_dma->ps_page_dma[j] = 0; + put_page(ps_page->ps_page[j]); + ps_page->ps_page[j] = NULL; + } + } + + /* there also may be some cached data in our adapter */ + if (rx_ring->rx_skb_top) { + dev_kfree_skb(rx_ring->rx_skb_top); + + /* rx_skb_prev will be wiped out by rx_skb_top */ + rx_ring->rx_skb_top = NULL; + rx_ring->rx_skb_prev = NULL; } + size = sizeof(struct e1000_buffer) * rx_ring->count; memset(rx_ring->buffer_info, 0, size); size = sizeof(struct e1000_ps_page) * rx_ring->count; @@ -1963,7 +2145,7 @@ e1000_clean_all_rx_rings(struct e1000_adapter *adapter) { int i; - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) e1000_clean_rx_ring(adapter, &adapter->rx_ring[i]); } @@ -1984,7 +2166,7 @@ e1000_enter_82542_rst(struct e1000_adapter *adapter) E1000_WRITE_FLUSH(&adapter->hw); mdelay(5); - if(netif_running(netdev)) + if (netif_running(netdev)) e1000_clean_all_rx_rings(adapter); } @@ -2000,12 +2182,14 @@ e1000_leave_82542_rst(struct e1000_adapter *adapter) E1000_WRITE_FLUSH(&adapter->hw); mdelay(5); - if(adapter->hw.pci_cmd_word & PCI_COMMAND_INVALIDATE) + if (adapter->hw.pci_cmd_word & PCI_COMMAND_INVALIDATE) e1000_pci_set_mwi(&adapter->hw); - if(netif_running(netdev)) { + if (netif_running(netdev)) { + /* No need to loop, because 82542 supports only 1 queue */ + struct e1000_rx_ring *ring = &adapter->rx_ring[0]; e1000_configure_rx(adapter); - e1000_alloc_rx_buffers(adapter, &adapter->rx_ring[0]); + adapter->alloc_rx_buf(adapter, ring, E1000_DESC_UNUSED(ring)); } } @@ -2023,12 +2207,12 @@ e1000_set_mac(struct net_device *netdev, void *p) struct e1000_adapter *adapter = netdev_priv(netdev); struct sockaddr *addr = p; - if(!is_valid_ether_addr(addr->sa_data)) + if (!is_valid_ether_addr(addr->sa_data)) return -EADDRNOTAVAIL; /* 82542 2.0 needs to be in reset to write receive address registers */ - if(adapter->hw.mac_type == e1000_82542_rev2_0) + if (adapter->hw.mac_type == e1000_82542_rev2_0) e1000_enter_82542_rst(adapter); memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len); @@ -2042,17 +2226,17 @@ e1000_set_mac(struct net_device *netdev, void *p) /* activate the work around */ adapter->hw.laa_is_present = 1; - /* Hold a copy of the LAA in RAR[14] This is done so that - * between the time RAR[0] gets clobbered and the time it - * gets fixed (in e1000_watchdog), the actual LAA is in one + /* Hold a copy of the LAA in RAR[14] This is done so that + * between the time RAR[0] gets clobbered and the time it + * gets fixed (in e1000_watchdog), the actual LAA is in one * of the RARs and no incoming packets directed to this port - * are dropped. Eventaully the LAA will be in RAR[0] and + * are dropped. Eventaully the LAA will be in RAR[0] and * RAR[14] */ - e1000_rar_set(&adapter->hw, adapter->hw.mac_addr, + e1000_rar_set(&adapter->hw, adapter->hw.mac_addr, E1000_RAR_ENTRIES - 1); } - if(adapter->hw.mac_type == e1000_82542_rev2_0) + if (adapter->hw.mac_type == e1000_82542_rev2_0) e1000_leave_82542_rst(adapter); return 0; @@ -2086,9 +2270,9 @@ e1000_set_multi(struct net_device *netdev) rctl = E1000_READ_REG(hw, RCTL); - if(netdev->flags & IFF_PROMISC) { + if (netdev->flags & IFF_PROMISC) { rctl |= (E1000_RCTL_UPE | E1000_RCTL_MPE); - } else if(netdev->flags & IFF_ALLMULTI) { + } else if (netdev->flags & IFF_ALLMULTI) { rctl |= E1000_RCTL_MPE; rctl &= ~E1000_RCTL_UPE; } else { @@ -2099,7 +2283,7 @@ e1000_set_multi(struct net_device *netdev) /* 82542 2.0 needs to be in reset to write receive address registers */ - if(hw->mac_type == e1000_82542_rev2_0) + if (hw->mac_type == e1000_82542_rev2_0) e1000_enter_82542_rst(adapter); /* load the first 14 multicast address into the exact filters 1-14 @@ -2109,7 +2293,7 @@ e1000_set_multi(struct net_device *netdev) */ mc_ptr = netdev->mc_list; - for(i = 1; i < rar_entries; i++) { + for (i = 1; i < rar_entries; i++) { if (mc_ptr) { e1000_rar_set(hw, mc_ptr->dmi_addr, i); mc_ptr = mc_ptr->next; @@ -2121,17 +2305,17 @@ e1000_set_multi(struct net_device *netdev) /* clear the old settings from the multicast hash table */ - for(i = 0; i < E1000_NUM_MTA_REGISTERS; i++) + for (i = 0; i < E1000_NUM_MTA_REGISTERS; i++) E1000_WRITE_REG_ARRAY(hw, MTA, i, 0); /* load any remaining addresses into the hash table */ - for(; mc_ptr; mc_ptr = mc_ptr->next) { + for (; mc_ptr; mc_ptr = mc_ptr->next) { hash_value = e1000_hash_mc_addr(hw, mc_ptr->dmi_addr); e1000_mta_set(hw, hash_value); } - if(hw->mac_type == e1000_82542_rev2_0) + if (hw->mac_type == e1000_82542_rev2_0) e1000_leave_82542_rst(adapter); } @@ -2157,8 +2341,8 @@ e1000_82547_tx_fifo_stall(unsigned long data) struct net_device *netdev = adapter->netdev; uint32_t tctl; - if(atomic_read(&adapter->tx_fifo_stall)) { - if((E1000_READ_REG(&adapter->hw, TDT) == + if (atomic_read(&adapter->tx_fifo_stall)) { + if ((E1000_READ_REG(&adapter->hw, TDT) == E1000_READ_REG(&adapter->hw, TDH)) && (E1000_READ_REG(&adapter->hw, TDFT) == E1000_READ_REG(&adapter->hw, TDFH)) && @@ -2204,24 +2388,24 @@ static void e1000_watchdog_task(struct e1000_adapter *adapter) { struct net_device *netdev = adapter->netdev; - struct e1000_tx_ring *txdr = &adapter->tx_ring[0]; + struct e1000_tx_ring *txdr = adapter->tx_ring; uint32_t link; e1000_check_for_link(&adapter->hw); if (adapter->hw.mac_type == e1000_82573) { e1000_enable_tx_pkt_filtering(&adapter->hw); - if(adapter->mng_vlan_id != adapter->hw.mng_cookie.vlan_id) + if (adapter->mng_vlan_id != adapter->hw.mng_cookie.vlan_id) e1000_update_mng_vlan(adapter); - } + } - if((adapter->hw.media_type == e1000_media_type_internal_serdes) && + if ((adapter->hw.media_type == e1000_media_type_internal_serdes) && !(E1000_READ_REG(&adapter->hw, TXCW) & E1000_TXCW_ANE)) link = !adapter->hw.serdes_link_down; else link = E1000_READ_REG(&adapter->hw, STATUS) & E1000_STATUS_LU; - if(link) { - if(!netif_carrier_ok(netdev)) { + if (link) { + if (!netif_carrier_ok(netdev)) { e1000_get_speed_and_duplex(&adapter->hw, &adapter->link_speed, &adapter->link_duplex); @@ -2231,13 +2415,28 @@ e1000_watchdog_task(struct e1000_adapter *adapter) adapter->link_duplex == FULL_DUPLEX ? "Full Duplex" : "Half Duplex"); + /* tweak tx_queue_len according to speed/duplex */ + netdev->tx_queue_len = adapter->tx_queue_len; + adapter->tx_timeout_factor = 1; + if (adapter->link_duplex == HALF_DUPLEX) { + switch (adapter->link_speed) { + case SPEED_10: + netdev->tx_queue_len = 10; + adapter->tx_timeout_factor = 8; + break; + case SPEED_100: + netdev->tx_queue_len = 100; + break; + } + } + netif_carrier_on(netdev); netif_wake_queue(netdev); mod_timer(&adapter->phy_info_timer, jiffies + 2 * HZ); adapter->smartspeed = 0; } } else { - if(netif_carrier_ok(netdev)) { + if (netif_carrier_ok(netdev)) { adapter->link_speed = 0; adapter->link_duplex = 0; DPRINTK(LINK, INFO, "NIC Link is Down\n"); @@ -2263,7 +2462,10 @@ e1000_watchdog_task(struct e1000_adapter *adapter) e1000_update_adaptive(&adapter->hw); - if (adapter->num_queues == 1 && !netif_carrier_ok(netdev)) { +#ifdef CONFIG_E1000_MQ + txdr = *per_cpu_ptr(adapter->cpu_tx_ring, smp_processor_id()); +#endif + if (!netif_carrier_ok(netdev)) { if (E1000_DESC_UNUSED(txdr) + 1 < txdr->count) { /* We've lost link, so the controller stops DMA, * but we've got queued Tx work that's never going @@ -2274,12 +2476,12 @@ e1000_watchdog_task(struct e1000_adapter *adapter) } /* Dynamic mode for Interrupt Throttle Rate (ITR) */ - if(adapter->hw.mac_type >= e1000_82540 && adapter->itr == 1) { + if (adapter->hw.mac_type >= e1000_82540 && adapter->itr == 1) { /* Symmetric Tx/Rx gets a reduced ITR=2000; Total * asymmetrical Tx or Rx gets ITR=8000; everyone * else is between 2000-8000. */ uint32_t goc = (adapter->gotcl + adapter->gorcl) / 10000; - uint32_t dif = (adapter->gotcl > adapter->gorcl ? + uint32_t dif = (adapter->gotcl > adapter->gorcl ? adapter->gotcl - adapter->gorcl : adapter->gorcl - adapter->gotcl) / 10000; uint32_t itr = goc > 0 ? (dif * 6000 / goc + 2000) : 8000; @@ -2292,7 +2494,7 @@ e1000_watchdog_task(struct e1000_adapter *adapter) /* Force detection of hung controller every watchdog period */ adapter->detect_tx_hung = TRUE; - /* With 82571 controllers, LAA may be overwritten due to controller + /* With 82571 controllers, LAA may be overwritten due to controller * reset from the other port. Set the appropriate LAA in RAR[0] */ if (adapter->hw.mac_type == e1000_82571 && adapter->hw.laa_is_present) e1000_rar_set(&adapter->hw, adapter->hw.mac_addr, 0); @@ -2314,13 +2516,14 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, { #ifdef NETIF_F_TSO struct e1000_context_desc *context_desc; + struct e1000_buffer *buffer_info; unsigned int i; uint32_t cmd_length = 0; uint16_t ipcse = 0, tucse, mss; uint8_t ipcss, ipcso, tucss, tucso, hdr_len; int err; - if(skb_shinfo(skb)->tso_size) { + if (skb_shinfo(skb)->tso_size) { if (skb_header_cloned(skb)) { err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC); if (err) @@ -2329,7 +2532,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2)); mss = skb_shinfo(skb)->tso_size; - if(skb->protocol == ntohs(ETH_P_IP)) { + if (skb->protocol == ntohs(ETH_P_IP)) { skb->nh.iph->tot_len = 0; skb->nh.iph->check = 0; skb->h.th->check = @@ -2341,7 +2544,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, cmd_length = E1000_TXD_CMD_IP; ipcse = skb->h.raw - skb->data - 1; #ifdef NETIF_F_TSO_IPV6 - } else if(skb->protocol == ntohs(ETH_P_IPV6)) { + } else if (skb->protocol == ntohs(ETH_P_IPV6)) { skb->nh.ipv6h->payload_len = 0; skb->h.th->check = ~csum_ipv6_magic(&skb->nh.ipv6h->saddr, @@ -2363,6 +2566,7 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, i = tx_ring->next_to_use; context_desc = E1000_CONTEXT_DESC(*tx_ring, i); + buffer_info = &tx_ring->buffer_info[i]; context_desc->lower_setup.ip_fields.ipcss = ipcss; context_desc->lower_setup.ip_fields.ipcso = ipcso; @@ -2374,14 +2578,16 @@ e1000_tso(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, context_desc->tcp_seg_setup.fields.hdr_len = hdr_len; context_desc->cmd_and_length = cpu_to_le32(cmd_length); + buffer_info->time_stamp = jiffies; + if (++i == tx_ring->count) i = 0; tx_ring->next_to_use = i; - return 1; + return TRUE; } #endif - return 0; + return FALSE; } static inline boolean_t @@ -2389,13 +2595,15 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, struct sk_buff *skb) { struct e1000_context_desc *context_desc; + struct e1000_buffer *buffer_info; unsigned int i; uint8_t css; - if(likely(skb->ip_summed == CHECKSUM_HW)) { + if (likely(skb->ip_summed == CHECKSUM_HW)) { css = skb->h.raw - skb->data; i = tx_ring->next_to_use; + buffer_info = &tx_ring->buffer_info[i]; context_desc = E1000_CONTEXT_DESC(*tx_ring, i); context_desc->upper_setup.tcp_fields.tucss = css; @@ -2404,6 +2612,8 @@ e1000_tx_csum(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, context_desc->tcp_seg_setup.data = 0; context_desc->cmd_and_length = cpu_to_le32(E1000_TXD_CMD_DEXT); + buffer_info->time_stamp = jiffies; + if (unlikely(++i == tx_ring->count)) i = 0; tx_ring->next_to_use = i; @@ -2429,7 +2639,7 @@ e1000_tx_map(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, i = tx_ring->next_to_use; - while(len) { + while (len) { buffer_info = &tx_ring->buffer_info[i]; size = min(len, max_per_txd); #ifdef NETIF_F_TSO @@ -2445,7 +2655,7 @@ e1000_tx_map(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, /* Workaround for premature desc write-backs * in TSO mode. Append 4-byte sentinel desc */ - if(unlikely(mss && !nr_frags && size == len && size > 8)) + if (unlikely(mss && !nr_frags && size == len && size > 8)) size -= 4; #endif /* work-around for errata 10 and it applies @@ -2453,13 +2663,13 @@ e1000_tx_map(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, * The fix is to make sure that the first descriptor of a * packet is smaller than 2048 - 16 - 16 (or 2016) bytes */ - if(unlikely((adapter->hw.bus_type == e1000_bus_type_pcix) && + if (unlikely((adapter->hw.bus_type == e1000_bus_type_pcix) && (size > 2015) && count == 0)) size = 2015; - + /* Workaround for potential 82544 hang in PCI-X. Avoid * terminating buffers within evenly-aligned dwords. */ - if(unlikely(adapter->pcix_82544 && + if (unlikely(adapter->pcix_82544 && !((unsigned long)(skb->data + offset + size - 1) & 4) && size > 4)) size -= 4; @@ -2475,29 +2685,29 @@ e1000_tx_map(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, len -= size; offset += size; count++; - if(unlikely(++i == tx_ring->count)) i = 0; + if (unlikely(++i == tx_ring->count)) i = 0; } - for(f = 0; f < nr_frags; f++) { + for (f = 0; f < nr_frags; f++) { struct skb_frag_struct *frag; frag = &skb_shinfo(skb)->frags[f]; len = frag->size; offset = frag->page_offset; - while(len) { + while (len) { buffer_info = &tx_ring->buffer_info[i]; size = min(len, max_per_txd); #ifdef NETIF_F_TSO /* Workaround for premature desc write-backs * in TSO mode. Append 4-byte sentinel desc */ - if(unlikely(mss && f == (nr_frags-1) && size == len && size > 8)) + if (unlikely(mss && f == (nr_frags-1) && size == len && size > 8)) size -= 4; #endif /* Workaround for potential 82544 hang in PCI-X. * Avoid terminating buffers within evenly-aligned * dwords. */ - if(unlikely(adapter->pcix_82544 && + if (unlikely(adapter->pcix_82544 && !((unsigned long)(frag->page+offset+size-1) & 4) && size > 4)) size -= 4; @@ -2514,7 +2724,7 @@ e1000_tx_map(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, len -= size; offset += size; count++; - if(unlikely(++i == tx_ring->count)) i = 0; + if (unlikely(++i == tx_ring->count)) i = 0; } } @@ -2534,35 +2744,35 @@ e1000_tx_queue(struct e1000_adapter *adapter, struct e1000_tx_ring *tx_ring, uint32_t txd_upper = 0, txd_lower = E1000_TXD_CMD_IFCS; unsigned int i; - if(likely(tx_flags & E1000_TX_FLAGS_TSO)) { + if (likely(tx_flags & E1000_TX_FLAGS_TSO)) { txd_lower |= E1000_TXD_CMD_DEXT | E1000_TXD_DTYP_D | E1000_TXD_CMD_TSE; txd_upper |= E1000_TXD_POPTS_TXSM << 8; - if(likely(tx_flags & E1000_TX_FLAGS_IPV4)) + if (likely(tx_flags & E1000_TX_FLAGS_IPV4)) txd_upper |= E1000_TXD_POPTS_IXSM << 8; } - if(likely(tx_flags & E1000_TX_FLAGS_CSUM)) { + if (likely(tx_flags & E1000_TX_FLAGS_CSUM)) { txd_lower |= E1000_TXD_CMD_DEXT | E1000_TXD_DTYP_D; txd_upper |= E1000_TXD_POPTS_TXSM << 8; } - if(unlikely(tx_flags & E1000_TX_FLAGS_VLAN)) { + if (unlikely(tx_flags & E1000_TX_FLAGS_VLAN)) { txd_lower |= E1000_TXD_CMD_VLE; txd_upper |= (tx_flags & E1000_TX_FLAGS_VLAN_MASK); } i = tx_ring->next_to_use; - while(count--) { + while (count--) { buffer_info = &tx_ring->buffer_info[i]; tx_desc = E1000_TX_DESC(*tx_ring, i); tx_desc->buffer_addr = cpu_to_le64(buffer_info->dma); tx_desc->lower.data = cpu_to_le32(txd_lower | buffer_info->length); tx_desc->upper.data = cpu_to_le32(txd_upper); - if(unlikely(++i == tx_ring->count)) i = 0; + if (unlikely(++i == tx_ring->count)) i = 0; } tx_desc->lower.data |= cpu_to_le32(adapter->txd_cmd); @@ -2597,20 +2807,20 @@ e1000_82547_fifo_workaround(struct e1000_adapter *adapter, struct sk_buff *skb) E1000_ROUNDUP(skb_fifo_len, E1000_FIFO_HDR); - if(adapter->link_duplex != HALF_DUPLEX) + if (adapter->link_duplex != HALF_DUPLEX) goto no_fifo_stall_required; - if(atomic_read(&adapter->tx_fifo_stall)) + if (atomic_read(&adapter->tx_fifo_stall)) return 1; - if(skb_fifo_len >= (E1000_82547_PAD_LEN + fifo_space)) { + if (skb_fifo_len >= (E1000_82547_PAD_LEN + fifo_space)) { atomic_set(&adapter->tx_fifo_stall, 1); return 1; } no_fifo_stall_required: adapter->tx_fifo_head += skb_fifo_len; - if(adapter->tx_fifo_head >= adapter->tx_fifo_size) + if (adapter->tx_fifo_head >= adapter->tx_fifo_size) adapter->tx_fifo_head -= adapter->tx_fifo_size; return 0; } @@ -2621,27 +2831,27 @@ e1000_transfer_dhcp_info(struct e1000_adapter *adapter, struct sk_buff *skb) { struct e1000_hw *hw = &adapter->hw; uint16_t length, offset; - if(vlan_tx_tag_present(skb)) { - if(!((vlan_tx_tag_get(skb) == adapter->hw.mng_cookie.vlan_id) && + if (vlan_tx_tag_present(skb)) { + if (!((vlan_tx_tag_get(skb) == adapter->hw.mng_cookie.vlan_id) && ( adapter->hw.mng_cookie.status & E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT)) ) return 0; } - if ((skb->len > MINIMUM_DHCP_PACKET_SIZE) && (!skb->protocol)) { + if ((skb->len > MINIMUM_DHCP_PACKET_SIZE) && (!skb->protocol)) { struct ethhdr *eth = (struct ethhdr *) skb->data; - if((htons(ETH_P_IP) == eth->h_proto)) { - const struct iphdr *ip = + if ((htons(ETH_P_IP) == eth->h_proto)) { + const struct iphdr *ip = (struct iphdr *)((uint8_t *)skb->data+14); - if(IPPROTO_UDP == ip->protocol) { - struct udphdr *udp = - (struct udphdr *)((uint8_t *)ip + + if (IPPROTO_UDP == ip->protocol) { + struct udphdr *udp = + (struct udphdr *)((uint8_t *)ip + (ip->ihl << 2)); - if(ntohs(udp->dest) == 67) { + if (ntohs(udp->dest) == 67) { offset = (uint8_t *)udp + 8 - skb->data; length = skb->len - offset; return e1000_mng_write_dhcp_info(hw, - (uint8_t *)udp + 8, + (uint8_t *)udp + 8, length); } } @@ -2664,7 +2874,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) unsigned int nr_frags = 0; unsigned int mss = 0; int count = 0; - int tso; + int tso; unsigned int f; len -= skb->data_len; @@ -2687,16 +2897,35 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) * 4 = ceil(buffer len/mss). To make sure we don't * overrun the FIFO, adjust the max buffer len if mss * drops. */ - if(mss) { + if (mss) { + uint8_t hdr_len; max_per_txd = min(mss << 2, max_per_txd); max_txd_pwr = fls(max_per_txd) - 1; + + /* TSO Workaround for 82571/2 Controllers -- if skb->data + * points to just header, pull a few bytes of payload from + * frags into skb->data */ + hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2)); + if (skb->data_len && (hdr_len == (skb->len - skb->data_len)) && + (adapter->hw.mac_type == e1000_82571 || + adapter->hw.mac_type == e1000_82572)) { + unsigned int pull_size; + pull_size = min((unsigned int)4, skb->data_len); + if (!__pskb_pull_tail(skb, pull_size)) { + printk(KERN_ERR "__pskb_pull_tail failed.\n"); + dev_kfree_skb_any(skb); + return -EFAULT; + } + len = skb->len - skb->data_len; + } } - if((mss) || (skb->ip_summed == CHECKSUM_HW)) + /* reserve a descriptor for the offload context */ + if ((mss) || (skb->ip_summed == CHECKSUM_HW)) count++; count++; #else - if(skb->ip_summed == CHECKSUM_HW) + if (skb->ip_summed == CHECKSUM_HW) count++; #endif @@ -2709,45 +2938,24 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) count += TXD_USE_COUNT(len, max_txd_pwr); - if(adapter->pcix_82544) + if (adapter->pcix_82544) count++; - /* work-around for errata 10 and it applies to all controllers + /* work-around for errata 10 and it applies to all controllers * in PCI-X mode, so add one more descriptor to the count */ - if(unlikely((adapter->hw.bus_type == e1000_bus_type_pcix) && + if (unlikely((adapter->hw.bus_type == e1000_bus_type_pcix) && (len > 2015))) count++; nr_frags = skb_shinfo(skb)->nr_frags; - for(f = 0; f < nr_frags; f++) + for (f = 0; f < nr_frags; f++) count += TXD_USE_COUNT(skb_shinfo(skb)->frags[f].size, max_txd_pwr); - if(adapter->pcix_82544) + if (adapter->pcix_82544) count += nr_frags; -#ifdef NETIF_F_TSO - /* TSO Workaround for 82571/2 Controllers -- if skb->data - * points to just header, pull a few bytes of payload from - * frags into skb->data */ - if (skb_shinfo(skb)->tso_size) { - uint8_t hdr_len; - hdr_len = ((skb->h.raw - skb->data) + (skb->h.th->doff << 2)); - if (skb->data_len && (hdr_len < (skb->len - skb->data_len)) && - (adapter->hw.mac_type == e1000_82571 || - adapter->hw.mac_type == e1000_82572)) { - unsigned int pull_size; - pull_size = min((unsigned int)4, skb->data_len); - if (!__pskb_pull_tail(skb, pull_size)) { - printk(KERN_ERR "__pskb_pull_tail failed.\n"); - dev_kfree_skb_any(skb); - return -EFAULT; - } - } - } -#endif - - if(adapter->hw.tx_pkt_filtering && (adapter->hw.mac_type == e1000_82573) ) + if (adapter->hw.tx_pkt_filtering && (adapter->hw.mac_type == e1000_82573) ) e1000_transfer_dhcp_info(adapter, skb); local_irq_save(flags); @@ -2765,8 +2973,8 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) return NETDEV_TX_BUSY; } - if(unlikely(adapter->hw.mac_type == e1000_82547)) { - if(unlikely(e1000_82547_fifo_workaround(adapter, skb))) { + if (unlikely(adapter->hw.mac_type == e1000_82547)) { + if (unlikely(e1000_82547_fifo_workaround(adapter, skb))) { netif_stop_queue(netdev); mod_timer(&adapter->tx_fifo_stall_timer, jiffies); spin_unlock_irqrestore(&tx_ring->tx_lock, flags); @@ -2774,13 +2982,13 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev) } } - if(unlikely(adapter->vlgrp && vlan_tx_tag_present(skb))) { + if (unlikely(adapter->vlgrp && vlan_tx_tag_present(skb))) { tx_flags |= E1000_TX_FLAGS_VLAN; tx_flags |= (vlan_tx_tag_get(skb) << E1000_TX_FLAGS_VLAN_SHIFT); } first = tx_ring->next_to_use; - + tso = e1000_tso(adapter, tx_ring, skb); if (tso < 0) { dev_kfree_skb_any(skb); @@ -2833,6 +3041,7 @@ e1000_tx_timeout_task(struct net_device *netdev) { struct e1000_adapter *adapter = netdev_priv(netdev); + adapter->tx_timeout_count++; e1000_down(adapter); e1000_up(adapter); } @@ -2850,7 +3059,7 @@ e1000_get_stats(struct net_device *netdev) { struct e1000_adapter *adapter = netdev_priv(netdev); - e1000_update_stats(adapter); + /* only return the current stats */ return &adapter->net_stats; } @@ -2868,56 +3077,57 @@ e1000_change_mtu(struct net_device *netdev, int new_mtu) struct e1000_adapter *adapter = netdev_priv(netdev); int max_frame = new_mtu + ENET_HEADER_SIZE + ETHERNET_FCS_SIZE; - if((max_frame < MINIMUM_ETHERNET_FRAME_SIZE) || - (max_frame > MAX_JUMBO_FRAME_SIZE)) { - DPRINTK(PROBE, ERR, "Invalid MTU setting\n"); - return -EINVAL; - } - -#define MAX_STD_JUMBO_FRAME_SIZE 9234 - /* might want this to be bigger enum check... */ - /* 82571 controllers limit jumbo frame size to 10500 bytes */ - if ((adapter->hw.mac_type == e1000_82571 || - adapter->hw.mac_type == e1000_82572) && - max_frame > MAX_STD_JUMBO_FRAME_SIZE) { - DPRINTK(PROBE, ERR, "MTU > 9216 bytes not supported " - "on 82571 and 82572 controllers.\n"); - return -EINVAL; - } - - if(adapter->hw.mac_type == e1000_82573 && - max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) { - DPRINTK(PROBE, ERR, "Jumbo Frames not supported " - "on 82573\n"); + if ((max_frame < MINIMUM_ETHERNET_FRAME_SIZE) || + (max_frame > MAX_JUMBO_FRAME_SIZE)) { + DPRINTK(PROBE, ERR, "Invalid MTU setting\n"); return -EINVAL; } - if(adapter->hw.mac_type > e1000_82547_rev_2) { - adapter->rx_buffer_len = max_frame; - E1000_ROUNDUP(adapter->rx_buffer_len, 1024); - } else { - if(unlikely((adapter->hw.mac_type < e1000_82543) && - (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE))) { - DPRINTK(PROBE, ERR, "Jumbo Frames not supported " - "on 82542\n"); + /* Adapter-specific max frame size limits. */ + switch (adapter->hw.mac_type) { + case e1000_82542_rev2_0: + case e1000_82542_rev2_1: + case e1000_82573: + if (max_frame > MAXIMUM_ETHERNET_FRAME_SIZE) { + DPRINTK(PROBE, ERR, "Jumbo Frames not supported.\n"); + return -EINVAL; + } + break; + case e1000_82571: + case e1000_82572: +#define MAX_STD_JUMBO_FRAME_SIZE 9234 + if (max_frame > MAX_STD_JUMBO_FRAME_SIZE) { + DPRINTK(PROBE, ERR, "MTU > 9216 not supported.\n"); return -EINVAL; - - } else { - if(max_frame <= E1000_RXBUFFER_2048) { - adapter->rx_buffer_len = E1000_RXBUFFER_2048; - } else if(max_frame <= E1000_RXBUFFER_4096) { - adapter->rx_buffer_len = E1000_RXBUFFER_4096; - } else if(max_frame <= E1000_RXBUFFER_8192) { - adapter->rx_buffer_len = E1000_RXBUFFER_8192; - } else if(max_frame <= E1000_RXBUFFER_16384) { - adapter->rx_buffer_len = E1000_RXBUFFER_16384; - } } + break; + default: + /* Capable of supporting up to MAX_JUMBO_FRAME_SIZE limit. */ + break; } + /* since the driver code now supports splitting a packet across + * multiple descriptors, most of the fifo related limitations on + * jumbo frame traffic have gone away. + * simply use 2k descriptors for everything. + * + * NOTE: dev_alloc_skb reserves 16 bytes, and typically NET_IP_ALIGN + * means we reserve 2 more, this pushes us to allocate from the next + * larger slab size + * i.e. RXBUFFER_2048 --> size-4096 slab */ + + /* recent hardware supports 1KB granularity */ + if (adapter->hw.mac_type > e1000_82547_rev_2) { + adapter->rx_buffer_len = + ((max_frame < E1000_RXBUFFER_2048) ? + max_frame : E1000_RXBUFFER_2048); + E1000_ROUNDUP(adapter->rx_buffer_len, 1024); + } else + adapter->rx_buffer_len = E1000_RXBUFFER_2048; + netdev->mtu = new_mtu; - if(netif_running(netdev)) { + if (netif_running(netdev)) { e1000_down(adapter); e1000_up(adapter); } @@ -3004,7 +3214,7 @@ e1000_update_stats(struct e1000_adapter *adapter) hw->collision_delta = E1000_READ_REG(hw, COLC); adapter->stats.colc += hw->collision_delta; - if(hw->mac_type >= e1000_82543) { + if (hw->mac_type >= e1000_82543) { adapter->stats.algnerrc += E1000_READ_REG(hw, ALGNERRC); adapter->stats.rxerrc += E1000_READ_REG(hw, RXERRC); adapter->stats.tncrs += E1000_READ_REG(hw, TNCRS); @@ -3012,7 +3222,7 @@ e1000_update_stats(struct e1000_adapter *adapter) adapter->stats.tsctc += E1000_READ_REG(hw, TSCTC); adapter->stats.tsctfc += E1000_READ_REG(hw, TSCTFC); } - if(hw->mac_type > e1000_82547_rev_2) { + if (hw->mac_type > e1000_82547_rev_2) { adapter->stats.iac += E1000_READ_REG(hw, IAC); adapter->stats.icrxoc += E1000_READ_REG(hw, ICRXOC); adapter->stats.icrxptc += E1000_READ_REG(hw, ICRXPTC); @@ -3037,12 +3247,11 @@ e1000_update_stats(struct e1000_adapter *adapter) adapter->net_stats.rx_errors = adapter->stats.rxerrc + adapter->stats.crcerrs + adapter->stats.algnerrc + - adapter->stats.rlec + adapter->stats.mpc + - adapter->stats.cexterr; + adapter->stats.rlec + adapter->stats.cexterr; + adapter->net_stats.rx_dropped = 0; adapter->net_stats.rx_length_errors = adapter->stats.rlec; adapter->net_stats.rx_crc_errors = adapter->stats.crcerrs; adapter->net_stats.rx_frame_errors = adapter->stats.algnerrc; - adapter->net_stats.rx_fifo_errors = adapter->stats.mpc; adapter->net_stats.rx_missed_errors = adapter->stats.mpc; /* Tx Errors */ @@ -3057,14 +3266,14 @@ e1000_update_stats(struct e1000_adapter *adapter) /* Phy Stats */ - if(hw->media_type == e1000_media_type_copper) { - if((adapter->link_speed == SPEED_1000) && + if (hw->media_type == e1000_media_type_copper) { + if ((adapter->link_speed == SPEED_1000) && (!e1000_read_phy_reg(hw, PHY_1000T_STATUS, &phy_tmp))) { phy_tmp &= PHY_IDLE_ERROR_COUNT_MASK; adapter->phy_stats.idle_errors += phy_tmp; } - if((hw->mac_type <= e1000_82546) && + if ((hw->mac_type <= e1000_82546) && (hw->phy_type == e1000_phy_m88) && !e1000_read_phy_reg(hw, M88E1000_RX_ERR_CNTR, &phy_tmp)) adapter->phy_stats.receive_errors += phy_tmp; @@ -3110,32 +3319,44 @@ e1000_intr(int irq, void *data, struct pt_regs *regs) struct e1000_adapter *adapter = netdev_priv(netdev); struct e1000_hw *hw = &adapter->hw; uint32_t icr = E1000_READ_REG(hw, ICR); -#if defined(CONFIG_E1000_NAPI) && defined(CONFIG_E1000_MQ) || !defined(CONFIG_E1000_NAPI) +#ifndef CONFIG_E1000_NAPI int i; +#else + /* Interrupt Auto-Mask...upon reading ICR, + * interrupts are masked. No need for the + * IMC write, but it does mean we should + * account for it ASAP. */ + if (likely(hw->mac_type >= e1000_82571)) + atomic_inc(&adapter->irq_sem); #endif - if(unlikely(!icr)) + if (unlikely(!icr)) { +#ifdef CONFIG_E1000_NAPI + if (hw->mac_type >= e1000_82571) + e1000_irq_enable(adapter); +#endif return IRQ_NONE; /* Not our interrupt */ + } - if(unlikely(icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC))) { + if (unlikely(icr & (E1000_ICR_RXSEQ | E1000_ICR_LSC))) { hw->get_link_status = 1; mod_timer(&adapter->watchdog_timer, jiffies); } #ifdef CONFIG_E1000_NAPI - atomic_inc(&adapter->irq_sem); - E1000_WRITE_REG(hw, IMC, ~0); - E1000_WRITE_FLUSH(hw); + if (unlikely(hw->mac_type < e1000_82571)) { + atomic_inc(&adapter->irq_sem); + E1000_WRITE_REG(hw, IMC, ~0); + E1000_WRITE_FLUSH(hw); + } #ifdef CONFIG_E1000_MQ if (atomic_read(&adapter->rx_sched_call_data.count) == 0) { - cpu_set(adapter->cpu_for_queue[0], - adapter->rx_sched_call_data.cpumask); - for (i = 1; i < adapter->num_queues; i++) { - cpu_set(adapter->cpu_for_queue[i], - adapter->rx_sched_call_data.cpumask); - atomic_inc(&adapter->irq_sem); - } - atomic_set(&adapter->rx_sched_call_data.count, i); + /* We must setup the cpumask once count == 0 since + * each cpu bit is cleared when the work is done. */ + adapter->rx_sched_call_data.cpumask = adapter->cpumask; + atomic_add(adapter->num_rx_queues - 1, &adapter->irq_sem); + atomic_set(&adapter->rx_sched_call_data.count, + adapter->num_rx_queues); smp_call_async_mask(&adapter->rx_sched_call_data); } else { printk("call_data.count == %u\n", atomic_read(&adapter->rx_sched_call_data.count)); @@ -3149,26 +3370,26 @@ e1000_intr(int irq, void *data, struct pt_regs *regs) #else /* if !CONFIG_E1000_NAPI */ /* Writing IMC and IMS is needed for 82547. - Due to Hub Link bus being occupied, an interrupt - de-assertion message is not able to be sent. - When an interrupt assertion message is generated later, - two messages are re-ordered and sent out. - That causes APIC to think 82547 is in de-assertion - state, while 82547 is in assertion state, resulting - in dead lock. Writing IMC forces 82547 into - de-assertion state. - */ - if(hw->mac_type == e1000_82547 || hw->mac_type == e1000_82547_rev_2){ + * Due to Hub Link bus being occupied, an interrupt + * de-assertion message is not able to be sent. + * When an interrupt assertion message is generated later, + * two messages are re-ordered and sent out. + * That causes APIC to think 82547 is in de-assertion + * state, while 82547 is in assertion state, resulting + * in dead lock. Writing IMC forces 82547 into + * de-assertion state. + */ + if (hw->mac_type == e1000_82547 || hw->mac_type == e1000_82547_rev_2) { atomic_inc(&adapter->irq_sem); E1000_WRITE_REG(hw, IMC, ~0); } - for(i = 0; i < E1000_MAX_INTR; i++) - if(unlikely(!adapter->clean_rx(adapter, adapter->rx_ring) & + for (i = 0; i < E1000_MAX_INTR; i++) + if (unlikely(!adapter->clean_rx(adapter, adapter->rx_ring) & !e1000_clean_tx_irq(adapter, adapter->tx_ring))) break; - if(hw->mac_type == e1000_82547 || hw->mac_type == e1000_82547_rev_2) + if (hw->mac_type == e1000_82547 || hw->mac_type == e1000_82547_rev_2) e1000_irq_enable(adapter); #endif /* CONFIG_E1000_NAPI */ @@ -3187,7 +3408,7 @@ e1000_clean(struct net_device *poll_dev, int *budget) { struct e1000_adapter *adapter; int work_to_do = min(*budget, poll_dev->quota); - int tx_cleaned, i = 0, work_done = 0; + int tx_cleaned = 0, i = 0, work_done = 0; /* Must NOT use netdev_priv macro here. */ adapter = poll_dev->priv; @@ -3198,19 +3419,31 @@ e1000_clean(struct net_device *poll_dev, int *budget) while (poll_dev != &adapter->polling_netdev[i]) { i++; - if (unlikely(i == adapter->num_queues)) + if (unlikely(i == adapter->num_rx_queues)) BUG(); } - tx_cleaned = e1000_clean_tx_irq(adapter, &adapter->tx_ring[i]); + if (likely(adapter->num_tx_queues == 1)) { + /* e1000_clean is called per-cpu. This lock protects + * tx_ring[0] from being cleaned by multiple cpus + * simultaneously. A failure obtaining the lock means + * tx_ring[0] is currently being cleaned anyway. */ + if (spin_trylock(&adapter->tx_queue_lock)) { + tx_cleaned = e1000_clean_tx_irq(adapter, + &adapter->tx_ring[0]); + spin_unlock(&adapter->tx_queue_lock); + } + } else + tx_cleaned = e1000_clean_tx_irq(adapter, &adapter->tx_ring[i]); + adapter->clean_rx(adapter, &adapter->rx_ring[i], &work_done, work_to_do); *budget -= work_done; poll_dev->quota -= work_done; - + /* If no Tx and not enough Rx work done, exit the polling mode */ - if((!tx_cleaned && (work_done == 0)) || + if ((!tx_cleaned && (work_done == 0)) || !netif_running(adapter->netdev)) { quit_polling: netif_rx_complete(poll_dev); @@ -3242,22 +3475,24 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter, eop_desc = E1000_TX_DESC(*tx_ring, eop); while (eop_desc->upper.data & cpu_to_le32(E1000_TXD_STAT_DD)) { - for(cleaned = FALSE; !cleaned; ) { + for (cleaned = FALSE; !cleaned; ) { tx_desc = E1000_TX_DESC(*tx_ring, i); buffer_info = &tx_ring->buffer_info[i]; cleaned = (i == eop); +#ifdef CONFIG_E1000_MQ + tx_ring->tx_stats.bytes += buffer_info->length; +#endif e1000_unmap_and_free_tx_resource(adapter, buffer_info); + memset(tx_desc, 0, sizeof(struct e1000_tx_desc)); - tx_desc->buffer_addr = 0; - tx_desc->lower.data = 0; - tx_desc->upper.data = 0; - - if(unlikely(++i == tx_ring->count)) i = 0; + if (unlikely(++i == tx_ring->count)) i = 0; } - tx_ring->pkt++; - +#ifdef CONFIG_E1000_MQ + tx_ring->tx_stats.packets++; +#endif + eop = tx_ring->buffer_info[i].next_to_watch; eop_desc = E1000_TX_DESC(*tx_ring, eop); } @@ -3266,7 +3501,7 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter, spin_lock(&tx_ring->tx_lock); - if(unlikely(cleaned && netif_queue_stopped(netdev) && + if (unlikely(cleaned && netif_queue_stopped(netdev) && netif_carrier_ok(netdev))) netif_wake_queue(netdev); @@ -3276,32 +3511,31 @@ e1000_clean_tx_irq(struct e1000_adapter *adapter, /* Detect a transmit hang in hardware, this serializes the * check with the clearing of time_stamp and movement of i */ adapter->detect_tx_hung = FALSE; - if (tx_ring->buffer_info[i].dma && - time_after(jiffies, tx_ring->buffer_info[i].time_stamp + HZ) + if (tx_ring->buffer_info[eop].dma && + time_after(jiffies, tx_ring->buffer_info[eop].time_stamp + + adapter->tx_timeout_factor * HZ) && !(E1000_READ_REG(&adapter->hw, STATUS) & - E1000_STATUS_TXOFF)) { + E1000_STATUS_TXOFF)) { /* detected Tx unit hang */ - i = tx_ring->next_to_clean; - eop = tx_ring->buffer_info[i].next_to_watch; - eop_desc = E1000_TX_DESC(*tx_ring, eop); DPRINTK(DRV, ERR, "Detected Tx Unit Hang\n" + " Tx Queue <%lu>\n" " TDH <%x>\n" " TDT <%x>\n" " next_to_use <%x>\n" " next_to_clean <%x>\n" "buffer_info[next_to_clean]\n" - " dma <%llx>\n" " time_stamp <%lx>\n" " next_to_watch <%x>\n" " jiffies <%lx>\n" " next_to_watch.status <%x>\n", + (unsigned long)((tx_ring - adapter->tx_ring) / + sizeof(struct e1000_tx_ring)), readl(adapter->hw.hw_addr + tx_ring->tdh), readl(adapter->hw.hw_addr + tx_ring->tdt), tx_ring->next_to_use, - i, - (unsigned long long)tx_ring->buffer_info[i].dma, - tx_ring->buffer_info[i].time_stamp, + tx_ring->next_to_clean, + tx_ring->buffer_info[eop].time_stamp, eop, jiffies, eop_desc->upper.fields.status); @@ -3329,21 +3563,21 @@ e1000_rx_checksum(struct e1000_adapter *adapter, skb->ip_summed = CHECKSUM_NONE; /* 82543 or newer only */ - if(unlikely(adapter->hw.mac_type < e1000_82543)) return; + if (unlikely(adapter->hw.mac_type < e1000_82543)) return; /* Ignore Checksum bit is set */ - if(unlikely(status & E1000_RXD_STAT_IXSM)) return; + if (unlikely(status & E1000_RXD_STAT_IXSM)) return; /* TCP/UDP checksum error bit is set */ - if(unlikely(errors & E1000_RXD_ERR_TCPE)) { + if (unlikely(errors & E1000_RXD_ERR_TCPE)) { /* let the stack verify checksum errors */ adapter->hw_csum_err++; return; } /* TCP/UDP Checksum has not been calculated */ - if(adapter->hw.mac_type <= e1000_82547_rev_2) { - if(!(status & E1000_RXD_STAT_TCPCS)) + if (adapter->hw.mac_type <= e1000_82547_rev_2) { + if (!(status & E1000_RXD_STAT_TCPCS)) return; } else { - if(!(status & (E1000_RXD_STAT_TCPCS | E1000_RXD_STAT_UDPCS))) + if (!(status & (E1000_RXD_STAT_TCPCS | E1000_RXD_STAT_UDPCS))) return; } /* It must be a TCP or UDP packet with a valid checksum */ @@ -3379,46 +3613,87 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, { struct net_device *netdev = adapter->netdev; struct pci_dev *pdev = adapter->pdev; - struct e1000_rx_desc *rx_desc; - struct e1000_buffer *buffer_info; - struct sk_buff *skb; + struct e1000_rx_desc *rx_desc, *next_rxd; + struct e1000_buffer *buffer_info, *next_buffer; unsigned long flags; uint32_t length; uint8_t last_byte; unsigned int i; - boolean_t cleaned = FALSE; + int cleaned_count = 0; + boolean_t cleaned = FALSE, multi_descriptor = FALSE; i = rx_ring->next_to_clean; rx_desc = E1000_RX_DESC(*rx_ring, i); + buffer_info = &rx_ring->buffer_info[i]; - while(rx_desc->status & E1000_RXD_STAT_DD) { - buffer_info = &rx_ring->buffer_info[i]; + while (rx_desc->status & E1000_RXD_STAT_DD) { + struct sk_buff *skb, *next_skb; + u8 status; #ifdef CONFIG_E1000_NAPI - if(*work_done >= work_to_do) + if (*work_done >= work_to_do) break; (*work_done)++; #endif - cleaned = TRUE; + status = rx_desc->status; + skb = buffer_info->skb; + buffer_info->skb = NULL; + + if (++i == rx_ring->count) i = 0; + next_rxd = E1000_RX_DESC(*rx_ring, i); + next_buffer = &rx_ring->buffer_info[i]; + next_skb = next_buffer->skb; + cleaned = TRUE; + cleaned_count++; pci_unmap_single(pdev, buffer_info->dma, buffer_info->length, PCI_DMA_FROMDEVICE); - skb = buffer_info->skb; length = le16_to_cpu(rx_desc->length); - if(unlikely(!(rx_desc->status & E1000_RXD_STAT_EOP))) { - /* All receives must fit into a single buffer */ - E1000_DBG("%s: Receive packet consumed multiple" - " buffers\n", netdev->name); - dev_kfree_skb_irq(skb); + skb_put(skb, length); + + if (!(status & E1000_RXD_STAT_EOP)) { + if (!rx_ring->rx_skb_top) { + rx_ring->rx_skb_top = skb; + rx_ring->rx_skb_top->len = length; + rx_ring->rx_skb_prev = skb; + } else { + if (skb_shinfo(rx_ring->rx_skb_top)->frag_list) { + rx_ring->rx_skb_prev->next = skb; + skb->prev = rx_ring->rx_skb_prev; + } else { + skb_shinfo(rx_ring->rx_skb_top)->frag_list = skb; + } + rx_ring->rx_skb_prev = skb; + rx_ring->rx_skb_top->data_len += length; + } goto next_desc; + } else { + if (rx_ring->rx_skb_top) { + if (skb_shinfo(rx_ring->rx_skb_top) + ->frag_list) { + rx_ring->rx_skb_prev->next = skb; + skb->prev = rx_ring->rx_skb_prev; + } else + skb_shinfo(rx_ring->rx_skb_top) + ->frag_list = skb; + + rx_ring->rx_skb_top->data_len += length; + rx_ring->rx_skb_top->len += + rx_ring->rx_skb_top->data_len; + + skb = rx_ring->rx_skb_top; + multi_descriptor = TRUE; + rx_ring->rx_skb_top = NULL; + rx_ring->rx_skb_prev = NULL; + } } - if(unlikely(rx_desc->errors & E1000_RXD_ERR_FRAME_ERR_MASK)) { + if (unlikely(rx_desc->errors & E1000_RXD_ERR_FRAME_ERR_MASK)) { last_byte = *(skb->data + length - 1); - if(TBI_ACCEPT(&adapter->hw, rx_desc->status, + if (TBI_ACCEPT(&adapter->hw, status, rx_desc->errors, length, last_byte)) { spin_lock_irqsave(&adapter->stats_lock, flags); e1000_tbi_adjust_stats(&adapter->hw, @@ -3433,18 +3708,41 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, } } - /* Good Receive */ - skb_put(skb, length - ETHERNET_FCS_SIZE); + /* code added for copybreak, this should improve + * performance for small packets with large amounts + * of reassembly being done in the stack */ +#define E1000_CB_LENGTH 256 + if ((length < E1000_CB_LENGTH) && + !rx_ring->rx_skb_top && + /* or maybe (status & E1000_RXD_STAT_EOP) && */ + !multi_descriptor) { + struct sk_buff *new_skb = + dev_alloc_skb(length + NET_IP_ALIGN); + if (new_skb) { + skb_reserve(new_skb, NET_IP_ALIGN); + new_skb->dev = netdev; + memcpy(new_skb->data - NET_IP_ALIGN, + skb->data - NET_IP_ALIGN, + length + NET_IP_ALIGN); + /* save the skb in buffer_info as good */ + buffer_info->skb = skb; + skb = new_skb; + skb_put(skb, length); + } + } + + /* end copybreak code */ /* Receive Checksum Offload */ e1000_rx_checksum(adapter, - (uint32_t)(rx_desc->status) | + (uint32_t)(status) | ((uint32_t)(rx_desc->errors) << 24), rx_desc->csum, skb); + skb->protocol = eth_type_trans(skb, netdev); #ifdef CONFIG_E1000_NAPI - if(unlikely(adapter->vlgrp && - (rx_desc->status & E1000_RXD_STAT_VP))) { + if (unlikely(adapter->vlgrp && + (status & E1000_RXD_STAT_VP))) { vlan_hwaccel_receive_skb(skb, adapter->vlgrp, le16_to_cpu(rx_desc->special) & E1000_RXD_SPC_VLAN_MASK); @@ -3452,8 +3750,8 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, netif_receive_skb(skb); } #else /* CONFIG_E1000_NAPI */ - if(unlikely(adapter->vlgrp && - (rx_desc->status & E1000_RXD_STAT_VP))) { + if (unlikely(adapter->vlgrp && + (status & E1000_RXD_STAT_VP))) { vlan_hwaccel_rx(skb, adapter->vlgrp, le16_to_cpu(rx_desc->special) & E1000_RXD_SPC_VLAN_MASK); @@ -3462,17 +3760,28 @@ e1000_clean_rx_irq(struct e1000_adapter *adapter, } #endif /* CONFIG_E1000_NAPI */ netdev->last_rx = jiffies; - rx_ring->pkt++; +#ifdef CONFIG_E1000_MQ + rx_ring->rx_stats.packets++; + rx_ring->rx_stats.bytes += length; +#endif next_desc: rx_desc->status = 0; - buffer_info->skb = NULL; - if(unlikely(++i == rx_ring->count)) i = 0; - rx_desc = E1000_RX_DESC(*rx_ring, i); + /* return some buffers to hardware, one at a time is too slow */ + if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) { + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); + cleaned_count = 0; + } + + rx_desc = next_rxd; + buffer_info = next_buffer; } rx_ring->next_to_clean = i; - adapter->alloc_rx_buf(adapter, rx_ring); + + cleaned_count = E1000_DESC_UNUSED(rx_ring); + if (cleaned_count) + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); return cleaned; } @@ -3492,52 +3801,59 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, struct e1000_rx_ring *rx_ring) #endif { - union e1000_rx_desc_packet_split *rx_desc; + union e1000_rx_desc_packet_split *rx_desc, *next_rxd; struct net_device *netdev = adapter->netdev; struct pci_dev *pdev = adapter->pdev; - struct e1000_buffer *buffer_info; + struct e1000_buffer *buffer_info, *next_buffer; struct e1000_ps_page *ps_page; struct e1000_ps_page_dma *ps_page_dma; - struct sk_buff *skb; + struct sk_buff *skb, *next_skb; unsigned int i, j; uint32_t length, staterr; + int cleaned_count = 0; boolean_t cleaned = FALSE; i = rx_ring->next_to_clean; rx_desc = E1000_RX_DESC_PS(*rx_ring, i); staterr = le32_to_cpu(rx_desc->wb.middle.status_error); + buffer_info = &rx_ring->buffer_info[i]; - while(staterr & E1000_RXD_STAT_DD) { - buffer_info = &rx_ring->buffer_info[i]; + while (staterr & E1000_RXD_STAT_DD) { ps_page = &rx_ring->ps_page[i]; ps_page_dma = &rx_ring->ps_page_dma[i]; #ifdef CONFIG_E1000_NAPI - if(unlikely(*work_done >= work_to_do)) + if (unlikely(*work_done >= work_to_do)) break; (*work_done)++; #endif + skb = buffer_info->skb; + + if (++i == rx_ring->count) i = 0; + next_rxd = E1000_RX_DESC_PS(*rx_ring, i); + next_buffer = &rx_ring->buffer_info[i]; + next_skb = next_buffer->skb; + cleaned = TRUE; + cleaned_count++; pci_unmap_single(pdev, buffer_info->dma, buffer_info->length, PCI_DMA_FROMDEVICE); - skb = buffer_info->skb; - - if(unlikely(!(staterr & E1000_RXD_STAT_EOP))) { + if (unlikely(!(staterr & E1000_RXD_STAT_EOP))) { E1000_DBG("%s: Packet Split buffers didn't pick up" " the full packet\n", netdev->name); dev_kfree_skb_irq(skb); goto next_desc; } - if(unlikely(staterr & E1000_RXDEXT_ERR_FRAME_ERR_MASK)) { + if (unlikely(staterr & E1000_RXDEXT_ERR_FRAME_ERR_MASK)) { dev_kfree_skb_irq(skb); goto next_desc; } length = le16_to_cpu(rx_desc->wb.middle.length0); - if(unlikely(!length)) { + if (unlikely(!length)) { E1000_DBG("%s: Last part of the packet spanning" " multiple descriptors\n", netdev->name); dev_kfree_skb_irq(skb); @@ -3547,8 +3863,8 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, /* Good Receive */ skb_put(skb, length); - for(j = 0; j < adapter->rx_ps_pages; j++) { - if(!(length = le16_to_cpu(rx_desc->wb.upper.length[j]))) + for (j = 0; j < adapter->rx_ps_pages; j++) { + if (!(length = le16_to_cpu(rx_desc->wb.upper.length[j]))) break; pci_unmap_page(pdev, ps_page_dma->ps_page_dma[j], @@ -3568,15 +3884,11 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, rx_desc->wb.lower.hi_dword.csum_ip.csum, skb); skb->protocol = eth_type_trans(skb, netdev); - if(likely(rx_desc->wb.upper.header_status & - E1000_RXDPS_HDRSTAT_HDRSP)) { + if (likely(rx_desc->wb.upper.header_status & + E1000_RXDPS_HDRSTAT_HDRSP)) adapter->rx_hdr_split++; -#ifdef HAVE_RX_ZERO_COPY - skb_shinfo(skb)->zero_copy = TRUE; -#endif - } #ifdef CONFIG_E1000_NAPI - if(unlikely(adapter->vlgrp && (staterr & E1000_RXD_STAT_VP))) { + if (unlikely(adapter->vlgrp && (staterr & E1000_RXD_STAT_VP))) { vlan_hwaccel_receive_skb(skb, adapter->vlgrp, le16_to_cpu(rx_desc->wb.middle.vlan) & E1000_RXD_SPC_VLAN_MASK); @@ -3584,7 +3896,7 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, netif_receive_skb(skb); } #else /* CONFIG_E1000_NAPI */ - if(unlikely(adapter->vlgrp && (staterr & E1000_RXD_STAT_VP))) { + if (unlikely(adapter->vlgrp && (staterr & E1000_RXD_STAT_VP))) { vlan_hwaccel_rx(skb, adapter->vlgrp, le16_to_cpu(rx_desc->wb.middle.vlan) & E1000_RXD_SPC_VLAN_MASK); @@ -3593,18 +3905,31 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter, } #endif /* CONFIG_E1000_NAPI */ netdev->last_rx = jiffies; - rx_ring->pkt++; +#ifdef CONFIG_E1000_MQ + rx_ring->rx_stats.packets++; + rx_ring->rx_stats.bytes += length; +#endif next_desc: rx_desc->wb.middle.status_error &= ~0xFF; buffer_info->skb = NULL; - if(unlikely(++i == rx_ring->count)) i = 0; - rx_desc = E1000_RX_DESC_PS(*rx_ring, i); + /* return some buffers to hardware, one at a time is too slow */ + if (unlikely(cleaned_count >= E1000_RX_BUFFER_WRITE)) { + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); + cleaned_count = 0; + } + + rx_desc = next_rxd; + buffer_info = next_buffer; + staterr = le32_to_cpu(rx_desc->wb.middle.status_error); } rx_ring->next_to_clean = i; - adapter->alloc_rx_buf(adapter, rx_ring); + + cleaned_count = E1000_DESC_UNUSED(rx_ring); + if (cleaned_count) + adapter->alloc_rx_buf(adapter, rx_ring, cleaned_count); return cleaned; } @@ -3616,7 +3941,8 @@ next_desc: static void e1000_alloc_rx_buffers(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring) + struct e1000_rx_ring *rx_ring, + int cleaned_count) { struct net_device *netdev = adapter->netdev; struct pci_dev *pdev = adapter->pdev; @@ -3629,11 +3955,18 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, i = rx_ring->next_to_use; buffer_info = &rx_ring->buffer_info[i]; - while(!buffer_info->skb) { - skb = dev_alloc_skb(bufsz); + while (cleaned_count--) { + if (!(skb = buffer_info->skb)) + skb = dev_alloc_skb(bufsz); + else { + skb_trim(skb, 0); + goto map_skb; + } + - if(unlikely(!skb)) { + if (unlikely(!skb)) { /* Better luck next round */ + adapter->alloc_rx_buff_failed++; break; } @@ -3670,6 +4003,7 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, buffer_info->skb = skb; buffer_info->length = adapter->rx_buffer_len; +map_skb: buffer_info->dma = pci_map_single(pdev, skb->data, adapter->rx_buffer_len, @@ -3695,20 +4029,23 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, rx_desc = E1000_RX_DESC(*rx_ring, i); rx_desc->buffer_addr = cpu_to_le64(buffer_info->dma); - if(unlikely((i & ~(E1000_RX_BUFFER_WRITE - 1)) == i)) { - /* Force memory writes to complete before letting h/w - * know there are new descriptors to fetch. (Only - * applicable for weak-ordered memory model archs, - * such as IA-64). */ - wmb(); - writel(i, adapter->hw.hw_addr + rx_ring->rdt); - } - - if(unlikely(++i == rx_ring->count)) i = 0; + if (unlikely(++i == rx_ring->count)) + i = 0; buffer_info = &rx_ring->buffer_info[i]; } - rx_ring->next_to_use = i; + if (likely(rx_ring->next_to_use != i)) { + rx_ring->next_to_use = i; + if (unlikely(i-- == 0)) + i = (rx_ring->count - 1); + + /* Force memory writes to complete before letting h/w + * know there are new descriptors to fetch. (Only + * applicable for weak-ordered memory model archs, + * such as IA-64). */ + wmb(); + writel(i, adapter->hw.hw_addr + rx_ring->rdt); + } } /** @@ -3718,7 +4055,8 @@ e1000_alloc_rx_buffers(struct e1000_adapter *adapter, static void e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, - struct e1000_rx_ring *rx_ring) + struct e1000_rx_ring *rx_ring, + int cleaned_count) { struct net_device *netdev = adapter->netdev; struct pci_dev *pdev = adapter->pdev; @@ -3734,16 +4072,18 @@ e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, ps_page = &rx_ring->ps_page[i]; ps_page_dma = &rx_ring->ps_page_dma[i]; - while(!buffer_info->skb) { + while (cleaned_count--) { rx_desc = E1000_RX_DESC_PS(*rx_ring, i); - for(j = 0; j < PS_PAGE_BUFFERS; j++) { + for (j = 0; j < PS_PAGE_BUFFERS; j++) { if (j < adapter->rx_ps_pages) { if (likely(!ps_page->ps_page[j])) { ps_page->ps_page[j] = alloc_page(GFP_ATOMIC); - if (unlikely(!ps_page->ps_page[j])) + if (unlikely(!ps_page->ps_page[j])) { + adapter->alloc_rx_buff_failed++; goto no_buffers; + } ps_page_dma->ps_page_dma[j] = pci_map_page(pdev, ps_page->ps_page[j], @@ -3751,7 +4091,7 @@ e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, PCI_DMA_FROMDEVICE); } /* Refresh the desc even if buffer_addrs didn't - * change because each write-back erases + * change because each write-back erases * this info. */ rx_desc->read.buffer_addr[j+1] = @@ -3762,8 +4102,10 @@ e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, skb = dev_alloc_skb(adapter->rx_ps_bsize0 + NET_IP_ALIGN); - if(unlikely(!skb)) + if (unlikely(!skb)) { + adapter->alloc_rx_buff_failed++; break; + } /* Make buffer alignment 2 beyond a 16 byte boundary * this will result in a 16 byte aligned IP header after @@ -3781,27 +4123,28 @@ e1000_alloc_rx_buffers_ps(struct e1000_adapter *adapter, rx_desc->read.buffer_addr[0] = cpu_to_le64(buffer_info->dma); - if(unlikely((i & ~(E1000_RX_BUFFER_WRITE - 1)) == i)) { - /* Force memory writes to complete before letting h/w - * know there are new descriptors to fetch. (Only - * applicable for weak-ordered memory model archs, - * such as IA-64). */ - wmb(); - /* Hardware increments by 16 bytes, but packet split - * descriptors are 32 bytes...so we increment tail - * twice as much. - */ - writel(i<<1, adapter->hw.hw_addr + rx_ring->rdt); - } - - if(unlikely(++i == rx_ring->count)) i = 0; + if (unlikely(++i == rx_ring->count)) i = 0; buffer_info = &rx_ring->buffer_info[i]; ps_page = &rx_ring->ps_page[i]; ps_page_dma = &rx_ring->ps_page_dma[i]; } no_buffers: - rx_ring->next_to_use = i; + if (likely(rx_ring->next_to_use != i)) { + rx_ring->next_to_use = i; + if (unlikely(i-- == 0)) i = (rx_ring->count - 1); + + /* Force memory writes to complete before letting h/w + * know there are new descriptors to fetch. (Only + * applicable for weak-ordered memory model archs, + * such as IA-64). */ + wmb(); + /* Hardware increments by 16 bytes, but packet split + * descriptors are 32 bytes...so we increment tail + * twice as much. + */ + writel(i<<1, adapter->hw.hw_addr + rx_ring->rdt); + } } /** @@ -3815,24 +4158,24 @@ e1000_smartspeed(struct e1000_adapter *adapter) uint16_t phy_status; uint16_t phy_ctrl; - if((adapter->hw.phy_type != e1000_phy_igp) || !adapter->hw.autoneg || + if ((adapter->hw.phy_type != e1000_phy_igp) || !adapter->hw.autoneg || !(adapter->hw.autoneg_advertised & ADVERTISE_1000_FULL)) return; - if(adapter->smartspeed == 0) { + if (adapter->smartspeed == 0) { /* If Master/Slave config fault is asserted twice, * we assume back-to-back */ e1000_read_phy_reg(&adapter->hw, PHY_1000T_STATUS, &phy_status); - if(!(phy_status & SR_1000T_MS_CONFIG_FAULT)) return; + if (!(phy_status & SR_1000T_MS_CONFIG_FAULT)) return; e1000_read_phy_reg(&adapter->hw, PHY_1000T_STATUS, &phy_status); - if(!(phy_status & SR_1000T_MS_CONFIG_FAULT)) return; + if (!(phy_status & SR_1000T_MS_CONFIG_FAULT)) return; e1000_read_phy_reg(&adapter->hw, PHY_1000T_CTRL, &phy_ctrl); - if(phy_ctrl & CR_1000T_MS_ENABLE) { + if (phy_ctrl & CR_1000T_MS_ENABLE) { phy_ctrl &= ~CR_1000T_MS_ENABLE; e1000_write_phy_reg(&adapter->hw, PHY_1000T_CTRL, phy_ctrl); adapter->smartspeed++; - if(!e1000_phy_setup_autoneg(&adapter->hw) && + if (!e1000_phy_setup_autoneg(&adapter->hw) && !e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &phy_ctrl)) { phy_ctrl |= (MII_CR_AUTO_NEG_EN | @@ -3842,12 +4185,12 @@ e1000_smartspeed(struct e1000_adapter *adapter) } } return; - } else if(adapter->smartspeed == E1000_SMARTSPEED_DOWNSHIFT) { + } else if (adapter->smartspeed == E1000_SMARTSPEED_DOWNSHIFT) { /* If still no link, perhaps using 2/3 pair cable */ e1000_read_phy_reg(&adapter->hw, PHY_1000T_CTRL, &phy_ctrl); phy_ctrl |= CR_1000T_MS_ENABLE; e1000_write_phy_reg(&adapter->hw, PHY_1000T_CTRL, phy_ctrl); - if(!e1000_phy_setup_autoneg(&adapter->hw) && + if (!e1000_phy_setup_autoneg(&adapter->hw) && !e1000_read_phy_reg(&adapter->hw, PHY_CTRL, &phy_ctrl)) { phy_ctrl |= (MII_CR_AUTO_NEG_EN | MII_CR_RESTART_AUTO_NEG); @@ -3855,7 +4198,7 @@ e1000_smartspeed(struct e1000_adapter *adapter) } } /* Restart process after E1000_SMARTSPEED_MAX iterations */ - if(adapter->smartspeed++ == E1000_SMARTSPEED_MAX) + if (adapter->smartspeed++ == E1000_SMARTSPEED_MAX) adapter->smartspeed = 0; } @@ -3896,7 +4239,7 @@ e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) uint16_t spddplx; unsigned long flags; - if(adapter->hw.media_type != e1000_media_type_copper) + if (adapter->hw.media_type != e1000_media_type_copper) return -EOPNOTSUPP; switch (cmd) { @@ -3904,10 +4247,10 @@ e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) data->phy_id = adapter->hw.phy_addr; break; case SIOCGMIIREG: - if(!capable(CAP_NET_ADMIN)) + if (!capable(CAP_NET_ADMIN)) return -EPERM; spin_lock_irqsave(&adapter->stats_lock, flags); - if(e1000_read_phy_reg(&adapter->hw, data->reg_num & 0x1F, + if (e1000_read_phy_reg(&adapter->hw, data->reg_num & 0x1F, &data->val_out)) { spin_unlock_irqrestore(&adapter->stats_lock, flags); return -EIO; @@ -3915,23 +4258,23 @@ e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) spin_unlock_irqrestore(&adapter->stats_lock, flags); break; case SIOCSMIIREG: - if(!capable(CAP_NET_ADMIN)) + if (!capable(CAP_NET_ADMIN)) return -EPERM; - if(data->reg_num & ~(0x1F)) + if (data->reg_num & ~(0x1F)) return -EFAULT; mii_reg = data->val_in; spin_lock_irqsave(&adapter->stats_lock, flags); - if(e1000_write_phy_reg(&adapter->hw, data->reg_num, + if (e1000_write_phy_reg(&adapter->hw, data->reg_num, mii_reg)) { spin_unlock_irqrestore(&adapter->stats_lock, flags); return -EIO; } - if(adapter->hw.phy_type == e1000_phy_m88) { + if (adapter->hw.phy_type == e1000_phy_m88) { switch (data->reg_num) { case PHY_CTRL: - if(mii_reg & MII_CR_POWER_DOWN) + if (mii_reg & MII_CR_POWER_DOWN) break; - if(mii_reg & MII_CR_AUTO_NEG_EN) { + if (mii_reg & MII_CR_AUTO_NEG_EN) { adapter->hw.autoneg = 1; adapter->hw.autoneg_advertised = 0x2F; } else { @@ -3946,14 +4289,14 @@ e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) HALF_DUPLEX; retval = e1000_set_spd_dplx(adapter, spddplx); - if(retval) { + if (retval) { spin_unlock_irqrestore( - &adapter->stats_lock, + &adapter->stats_lock, flags); return retval; } } - if(netif_running(adapter->netdev)) { + if (netif_running(adapter->netdev)) { e1000_down(adapter); e1000_up(adapter); } else @@ -3961,7 +4304,7 @@ e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) break; case M88E1000_PHY_SPEC_CTRL: case M88E1000_EXT_PHY_SPEC_CTRL: - if(e1000_phy_reset(&adapter->hw)) { + if (e1000_phy_reset(&adapter->hw)) { spin_unlock_irqrestore( &adapter->stats_lock, flags); return -EIO; @@ -3971,9 +4314,9 @@ e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) } else { switch (data->reg_num) { case PHY_CTRL: - if(mii_reg & MII_CR_POWER_DOWN) + if (mii_reg & MII_CR_POWER_DOWN) break; - if(netif_running(adapter->netdev)) { + if (netif_running(adapter->netdev)) { e1000_down(adapter); e1000_up(adapter); } else @@ -3995,7 +4338,7 @@ e1000_pci_set_mwi(struct e1000_hw *hw) struct e1000_adapter *adapter = hw->back; int ret_val = pci_set_mwi(adapter->pdev); - if(ret_val) + if (ret_val) DPRINTK(PROBE, ERR, "Error in setting MWI\n"); } @@ -4044,7 +4387,7 @@ e1000_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp) e1000_irq_disable(adapter); adapter->vlgrp = grp; - if(grp) { + if (grp) { /* enable VLAN tag insert/strip */ ctrl = E1000_READ_REG(&adapter->hw, CTRL); ctrl |= E1000_CTRL_VME; @@ -4066,7 +4409,7 @@ e1000_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp) rctl = E1000_READ_REG(&adapter->hw, RCTL); rctl &= ~E1000_RCTL_VFE; E1000_WRITE_REG(&adapter->hw, RCTL, rctl); - if(adapter->mng_vlan_id != (uint16_t)E1000_MNG_VLAN_NONE) { + if (adapter->mng_vlan_id != (uint16_t)E1000_MNG_VLAN_NONE) { e1000_vlan_rx_kill_vid(netdev, adapter->mng_vlan_id); adapter->mng_vlan_id = E1000_MNG_VLAN_NONE; } @@ -4080,9 +4423,10 @@ e1000_vlan_rx_add_vid(struct net_device *netdev, uint16_t vid) { struct e1000_adapter *adapter = netdev_priv(netdev); uint32_t vfta, index; - if((adapter->hw.mng_cookie.status & - E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) && - (vid == adapter->mng_vlan_id)) + + if ((adapter->hw.mng_cookie.status & + E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) && + (vid == adapter->mng_vlan_id)) return; /* add VID to filter table */ index = (vid >> 5) & 0x7F; @@ -4099,15 +4443,19 @@ e1000_vlan_rx_kill_vid(struct net_device *netdev, uint16_t vid) e1000_irq_disable(adapter); - if(adapter->vlgrp) + if (adapter->vlgrp) adapter->vlgrp->vlan_devices[vid] = NULL; e1000_irq_enable(adapter); - if((adapter->hw.mng_cookie.status & - E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) && - (vid == adapter->mng_vlan_id)) + if ((adapter->hw.mng_cookie.status & + E1000_MNG_DHCP_COOKIE_STATUS_VLAN_SUPPORT) && + (vid == adapter->mng_vlan_id)) { + /* release control to f/w */ + e1000_release_hw_control(adapter); return; + } + /* remove VID from filter table */ index = (vid >> 5) & 0x7F; vfta = E1000_READ_REG_ARRAY(&adapter->hw, VFTA, index); @@ -4120,10 +4468,10 @@ e1000_restore_vlan(struct e1000_adapter *adapter) { e1000_vlan_rx_register(adapter->netdev, adapter->vlgrp); - if(adapter->vlgrp) { + if (adapter->vlgrp) { uint16_t vid; - for(vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) { - if(!adapter->vlgrp->vlan_devices[vid]) + for (vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) { + if (!adapter->vlgrp->vlan_devices[vid]) continue; e1000_vlan_rx_add_vid(adapter->netdev, vid); } @@ -4136,13 +4484,13 @@ e1000_set_spd_dplx(struct e1000_adapter *adapter, uint16_t spddplx) adapter->hw.autoneg = 0; /* Fiber NICs only allow 1000 gbps Full duplex */ - if((adapter->hw.media_type == e1000_media_type_fiber) && + if ((adapter->hw.media_type == e1000_media_type_fiber) && spddplx != (SPEED_1000 + DUPLEX_FULL)) { DPRINTK(PROBE, ERR, "Unsupported Speed/Duplex configuration\n"); return -EINVAL; } - switch(spddplx) { + switch (spddplx) { case SPEED_10 + DUPLEX_HALF: adapter->hw.forced_speed_duplex = e1000_10_half; break; @@ -4168,35 +4516,92 @@ e1000_set_spd_dplx(struct e1000_adapter *adapter, uint16_t spddplx) } #ifdef CONFIG_PM +/* these functions save and restore 16 or 64 dwords (64-256 bytes) of config + * space versus the 64 bytes that pci_[save|restore]_state handle + */ +#define PCIE_CONFIG_SPACE_LEN 256 +#define PCI_CONFIG_SPACE_LEN 64 +static int +e1000_pci_save_state(struct e1000_adapter *adapter) +{ + struct pci_dev *dev = adapter->pdev; + int size; + int i; + if (adapter->hw.mac_type >= e1000_82571) + size = PCIE_CONFIG_SPACE_LEN; + else + size = PCI_CONFIG_SPACE_LEN; + + WARN_ON(adapter->config_space != NULL); + + adapter->config_space = kmalloc(size, GFP_KERNEL); + if (!adapter->config_space) { + DPRINTK(PROBE, ERR, "unable to allocate %d bytes\n", size); + return -ENOMEM; + } + for (i = 0; i < (size / 4); i++) + pci_read_config_dword(dev, i * 4, &adapter->config_space[i]); + return 0; +} + +static void +e1000_pci_restore_state(struct e1000_adapter *adapter) +{ + struct pci_dev *dev = adapter->pdev; + int size; + int i; + if (adapter->config_space == NULL) + return; + if (adapter->hw.mac_type >= e1000_82571) + size = PCIE_CONFIG_SPACE_LEN; + else + size = PCI_CONFIG_SPACE_LEN; + for (i = 0; i < (size / 4); i++) + pci_write_config_dword(dev, i * 4, adapter->config_space[i]); + kfree(adapter->config_space); + adapter->config_space = NULL; + return; +} +#endif /* CONFIG_PM */ + static int e1000_suspend(struct pci_dev *pdev, pm_message_t state) { struct net_device *netdev = pci_get_drvdata(pdev); struct e1000_adapter *adapter = netdev_priv(netdev); - uint32_t ctrl, ctrl_ext, rctl, manc, status, swsm; + uint32_t ctrl, ctrl_ext, rctl, manc, status; uint32_t wufc = adapter->wol; + int retval = 0; netif_device_detach(netdev); - if(netif_running(netdev)) + if (netif_running(netdev)) e1000_down(adapter); +#ifdef CONFIG_PM + /* implement our own version of pci_save_state(pdev) because pci + * express adapters have larger 256 byte config spaces */ + retval = e1000_pci_save_state(adapter); + if (retval) + return retval; +#endif + status = E1000_READ_REG(&adapter->hw, STATUS); - if(status & E1000_STATUS_LU) + if (status & E1000_STATUS_LU) wufc &= ~E1000_WUFC_LNKC; - if(wufc) { + if (wufc) { e1000_setup_rctl(adapter); e1000_set_multi(netdev); /* turn on all-multi mode if wake on multicast is enabled */ - if(adapter->wol & E1000_WUFC_MC) { + if (adapter->wol & E1000_WUFC_MC) { rctl = E1000_READ_REG(&adapter->hw, RCTL); rctl |= E1000_RCTL_MPE; E1000_WRITE_REG(&adapter->hw, RCTL, rctl); } - if(adapter->hw.mac_type >= e1000_82540) { + if (adapter->hw.mac_type >= e1000_82540) { ctrl = E1000_READ_REG(&adapter->hw, CTRL); /* advertise wake from D3Cold */ #define E1000_CTRL_ADVD3WUC 0x00100000 @@ -4207,7 +4612,7 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state) E1000_WRITE_REG(&adapter->hw, CTRL, ctrl); } - if(adapter->hw.media_type == e1000_media_type_fiber || + if (adapter->hw.media_type == e1000_media_type_fiber || adapter->hw.media_type == e1000_media_type_internal_serdes) { /* keep the laser running in D3 */ ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); @@ -4220,96 +4625,96 @@ e1000_suspend(struct pci_dev *pdev, pm_message_t state) E1000_WRITE_REG(&adapter->hw, WUC, E1000_WUC_PME_EN); E1000_WRITE_REG(&adapter->hw, WUFC, wufc); - pci_enable_wake(pdev, 3, 1); - pci_enable_wake(pdev, 4, 1); /* 4 == D3 cold */ + retval = pci_enable_wake(pdev, PCI_D3hot, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); } else { E1000_WRITE_REG(&adapter->hw, WUC, 0); E1000_WRITE_REG(&adapter->hw, WUFC, 0); - pci_enable_wake(pdev, 3, 0); - pci_enable_wake(pdev, 4, 0); /* 4 == D3 cold */ + retval = pci_enable_wake(pdev, PCI_D3hot, 0); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 0); /* 4 == D3 cold */ + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); } - pci_save_state(pdev); - - if(adapter->hw.mac_type >= e1000_82540 && + if (adapter->hw.mac_type >= e1000_82540 && adapter->hw.media_type == e1000_media_type_copper) { manc = E1000_READ_REG(&adapter->hw, MANC); - if(manc & E1000_MANC_SMBUS_EN) { + if (manc & E1000_MANC_SMBUS_EN) { manc |= E1000_MANC_ARP_EN; E1000_WRITE_REG(&adapter->hw, MANC, manc); - pci_enable_wake(pdev, 3, 1); - pci_enable_wake(pdev, 4, 1); /* 4 == D3 cold */ + retval = pci_enable_wake(pdev, PCI_D3hot, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 1); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); } } - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext & ~E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm & ~E1000_SWSM_DRV_LOAD); - break; - default: - break; - } + /* Release control of h/w to f/w. If f/w is AMT enabled, this + * would have already happened in close and is redundant. */ + e1000_release_hw_control(adapter); pci_disable_device(pdev); - pci_set_power_state(pdev, pci_choose_state(pdev, state)); + + retval = pci_set_power_state(pdev, pci_choose_state(pdev, state)); + if (retval) + DPRINTK(PROBE, ERR, "Error in setting power state\n"); return 0; } +#ifdef CONFIG_PM static int e1000_resume(struct pci_dev *pdev) { struct net_device *netdev = pci_get_drvdata(pdev); struct e1000_adapter *adapter = netdev_priv(netdev); - uint32_t manc, ret_val, swsm; - uint32_t ctrl_ext; + int retval; + uint32_t manc, ret_val; - pci_set_power_state(pdev, PCI_D0); - pci_restore_state(pdev); + retval = pci_set_power_state(pdev, PCI_D0); + if (retval) + DPRINTK(PROBE, ERR, "Error in setting power state\n"); + e1000_pci_restore_state(adapter); ret_val = pci_enable_device(pdev); pci_set_master(pdev); - pci_enable_wake(pdev, PCI_D3hot, 0); - pci_enable_wake(pdev, PCI_D3cold, 0); + retval = pci_enable_wake(pdev, PCI_D3hot, 0); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 wake\n"); + retval = pci_enable_wake(pdev, PCI_D3cold, 0); + if (retval) + DPRINTK(PROBE, ERR, "Error enabling D3 cold wake\n"); e1000_reset(adapter); E1000_WRITE_REG(&adapter->hw, WUS, ~0); - if(netif_running(netdev)) + if (netif_running(netdev)) e1000_up(adapter); netif_device_attach(netdev); - if(adapter->hw.mac_type >= e1000_82540 && + if (adapter->hw.mac_type >= e1000_82540 && adapter->hw.media_type == e1000_media_type_copper) { manc = E1000_READ_REG(&adapter->hw, MANC); manc &= ~(E1000_MANC_ARP_EN); E1000_WRITE_REG(&adapter->hw, MANC, manc); } - switch(adapter->hw.mac_type) { - case e1000_82571: - case e1000_82572: - ctrl_ext = E1000_READ_REG(&adapter->hw, CTRL_EXT); - E1000_WRITE_REG(&adapter->hw, CTRL_EXT, - ctrl_ext | E1000_CTRL_EXT_DRV_LOAD); - break; - case e1000_82573: - swsm = E1000_READ_REG(&adapter->hw, SWSM); - E1000_WRITE_REG(&adapter->hw, SWSM, - swsm | E1000_SWSM_DRV_LOAD); - break; - default: - break; - } + /* If the controller is 82573 and f/w is AMT, do not set + * DRV_LOAD until the interface is up. For all other cases, + * let the f/w know that the h/w is now under the control + * of the driver. */ + if (adapter->hw.mac_type != e1000_82573 || + !e1000_check_mng_mode(&adapter->hw)) + e1000_get_hw_control(adapter); return 0; } @@ -4327,6 +4732,9 @@ e1000_netpoll(struct net_device *netdev) disable_irq(adapter->pdev->irq); e1000_intr(adapter->pdev->irq, netdev, NULL); e1000_clean_tx_irq(adapter, adapter->tx_ring); +#ifndef CONFIG_E1000_NAPI + adapter->clean_rx(adapter, adapter->rx_ring); +#endif enable_irq(adapter->pdev->irq); } #endif diff --git a/drivers/net/e1000/e1000_osdep.h b/drivers/net/e1000/e1000_osdep.h index aac64de6143..9790db974dc 100644 --- a/drivers/net/e1000/e1000_osdep.h +++ b/drivers/net/e1000/e1000_osdep.h @@ -47,7 +47,7 @@ BUG(); \ } else { \ msleep(x); \ - } } while(0) + } } while (0) /* Some workarounds require millisecond delays and are run during interrupt * context. Most notably, when establishing link, the phy may need tweaking diff --git a/drivers/net/e1000/e1000_param.c b/drivers/net/e1000/e1000_param.c index ccbbe5ad8e0..3768d83cd57 100644 --- a/drivers/net/e1000/e1000_param.c +++ b/drivers/net/e1000/e1000_param.c @@ -177,7 +177,7 @@ E1000_PARAM(RxAbsIntDelay, "Receive Absolute Interrupt Delay"); * * Valid Range: 100-100000 (0=off, 1=dynamic) * - * Default Value: 1 + * Default Value: 8000 */ E1000_PARAM(InterruptThrottleRate, "Interrupt Throttling Rate"); @@ -227,7 +227,7 @@ static int __devinit e1000_validate_option(int *value, struct e1000_option *opt, struct e1000_adapter *adapter) { - if(*value == OPTION_UNSET) { + if (*value == OPTION_UNSET) { *value = opt->def; return 0; } @@ -244,7 +244,7 @@ e1000_validate_option(int *value, struct e1000_option *opt, } break; case range_option: - if(*value >= opt->arg.r.min && *value <= opt->arg.r.max) { + if (*value >= opt->arg.r.min && *value <= opt->arg.r.max) { DPRINTK(PROBE, INFO, "%s set to %i\n", opt->name, *value); return 0; @@ -254,10 +254,10 @@ e1000_validate_option(int *value, struct e1000_option *opt, int i; struct e1000_opt_list *ent; - for(i = 0; i < opt->arg.l.nr; i++) { + for (i = 0; i < opt->arg.l.nr; i++) { ent = &opt->arg.l.p[i]; - if(*value == ent->i) { - if(ent->str[0] != '\0') + if (*value == ent->i) { + if (ent->str[0] != '\0') DPRINTK(PROBE, INFO, "%s\n", ent->str); return 0; } @@ -291,7 +291,7 @@ void __devinit e1000_check_options(struct e1000_adapter *adapter) { int bd = adapter->bd_number; - if(bd >= E1000_MAX_NIC) { + if (bd >= E1000_MAX_NIC) { DPRINTK(PROBE, NOTICE, "Warning: no configuration for board #%i\n", bd); DPRINTK(PROBE, NOTICE, "Using defaults for all values\n"); @@ -315,12 +315,12 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_TxDescriptors > bd) { tx_ring->count = TxDescriptors[bd]; e1000_validate_option(&tx_ring->count, &opt, adapter); - E1000_ROUNDUP(tx_ring->count, + E1000_ROUNDUP(tx_ring->count, REQ_TX_DESCRIPTOR_MULTIPLE); } else { tx_ring->count = opt.def; } - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_tx_queues; i++) tx_ring[i].count = tx_ring->count; } { /* Receive Descriptor Count */ @@ -341,12 +341,12 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_RxDescriptors > bd) { rx_ring->count = RxDescriptors[bd]; e1000_validate_option(&rx_ring->count, &opt, adapter); - E1000_ROUNDUP(rx_ring->count, + E1000_ROUNDUP(rx_ring->count, REQ_RX_DESCRIPTOR_MULTIPLE); } else { rx_ring->count = opt.def; } - for (i = 0; i < adapter->num_queues; i++) + for (i = 0; i < adapter->num_rx_queues; i++) rx_ring[i].count = rx_ring->count; } { /* Checksum Offload Enable/Disable */ @@ -388,7 +388,7 @@ e1000_check_options(struct e1000_adapter *adapter) e1000_validate_option(&fc, &opt, adapter); adapter->hw.fc = adapter->hw.original_fc = fc; } else { - adapter->hw.fc = opt.def; + adapter->hw.fc = adapter->hw.original_fc = opt.def; } } { /* Transmit Interrupt Delay */ @@ -403,7 +403,7 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_TxIntDelay > bd) { adapter->tx_int_delay = TxIntDelay[bd]; - e1000_validate_option(&adapter->tx_int_delay, &opt, + e1000_validate_option(&adapter->tx_int_delay, &opt, adapter); } else { adapter->tx_int_delay = opt.def; @@ -421,7 +421,7 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_TxAbsIntDelay > bd) { adapter->tx_abs_int_delay = TxAbsIntDelay[bd]; - e1000_validate_option(&adapter->tx_abs_int_delay, &opt, + e1000_validate_option(&adapter->tx_abs_int_delay, &opt, adapter); } else { adapter->tx_abs_int_delay = opt.def; @@ -439,7 +439,7 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_RxIntDelay > bd) { adapter->rx_int_delay = RxIntDelay[bd]; - e1000_validate_option(&adapter->rx_int_delay, &opt, + e1000_validate_option(&adapter->rx_int_delay, &opt, adapter); } else { adapter->rx_int_delay = opt.def; @@ -457,7 +457,7 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_RxAbsIntDelay > bd) { adapter->rx_abs_int_delay = RxAbsIntDelay[bd]; - e1000_validate_option(&adapter->rx_abs_int_delay, &opt, + e1000_validate_option(&adapter->rx_abs_int_delay, &opt, adapter); } else { adapter->rx_abs_int_delay = opt.def; @@ -475,17 +475,17 @@ e1000_check_options(struct e1000_adapter *adapter) if (num_InterruptThrottleRate > bd) { adapter->itr = InterruptThrottleRate[bd]; - switch(adapter->itr) { + switch (adapter->itr) { case 0: - DPRINTK(PROBE, INFO, "%s turned off\n", + DPRINTK(PROBE, INFO, "%s turned off\n", opt.name); break; case 1: - DPRINTK(PROBE, INFO, "%s set to dynamic mode\n", + DPRINTK(PROBE, INFO, "%s set to dynamic mode\n", opt.name); break; default: - e1000_validate_option(&adapter->itr, &opt, + e1000_validate_option(&adapter->itr, &opt, adapter); break; } @@ -494,7 +494,7 @@ e1000_check_options(struct e1000_adapter *adapter) } } - switch(adapter->hw.media_type) { + switch (adapter->hw.media_type) { case e1000_media_type_fiber: case e1000_media_type_internal_serdes: e1000_check_fiber_options(adapter); @@ -518,17 +518,17 @@ static void __devinit e1000_check_fiber_options(struct e1000_adapter *adapter) { int bd = adapter->bd_number; - if(num_Speed > bd) { + if (num_Speed > bd) { DPRINTK(PROBE, INFO, "Speed not valid for fiber adapters, " "parameter ignored\n"); } - if(num_Duplex > bd) { + if (num_Duplex > bd) { DPRINTK(PROBE, INFO, "Duplex not valid for fiber adapters, " "parameter ignored\n"); } - if((num_AutoNeg > bd) && (AutoNeg[bd] != 0x20)) { + if ((num_AutoNeg > bd) && (AutoNeg[bd] != 0x20)) { DPRINTK(PROBE, INFO, "AutoNeg other than 1000/Full is " "not valid for fiber adapters, " "parameter ignored\n"); @@ -584,6 +584,12 @@ e1000_check_copper_options(struct e1000_adapter *adapter) .p = dplx_list }} }; + if (e1000_check_phy_reset_block(&adapter->hw)) { + DPRINTK(PROBE, INFO, + "Link active due to SoL/IDER Session. " + "Speed/Duplex/AutoNeg parameter ignored.\n"); + return; + } if (num_Duplex > bd) { dplx = Duplex[bd]; e1000_validate_option(&dplx, &opt, adapter); @@ -592,7 +598,7 @@ e1000_check_copper_options(struct e1000_adapter *adapter) } } - if((num_AutoNeg > bd) && (speed != 0 || dplx != 0)) { + if ((num_AutoNeg > bd) && (speed != 0 || dplx != 0)) { DPRINTK(PROBE, INFO, "AutoNeg specified along with Speed or Duplex, " "parameter ignored\n"); @@ -653,7 +659,7 @@ e1000_check_copper_options(struct e1000_adapter *adapter) switch (speed + dplx) { case 0: adapter->hw.autoneg = adapter->fc_autoneg = 1; - if((num_Speed > bd) && (speed != 0 || dplx != 0)) + if ((num_Speed > bd) && (speed != 0 || dplx != 0)) DPRINTK(PROBE, INFO, "Speed and duplex autonegotiation enabled\n"); break; diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c index 22c3a37bba5..40ae36b20c9 100644 --- a/drivers/net/mv643xx_eth.c +++ b/drivers/net/mv643xx_eth.c @@ -35,6 +35,8 @@ #include <linux/tcp.h> #include <linux/udp.h> #include <linux/etherdevice.h> +#include <linux/in.h> +#include <linux/ip.h> #include <linux/bitops.h> #include <linux/delay.h> @@ -55,13 +57,15 @@ /* Constants */ #define VLAN_HLEN 4 #define FCS_LEN 4 -#define WRAP NET_IP_ALIGN + ETH_HLEN + VLAN_HLEN + FCS_LEN +#define DMA_ALIGN 8 /* hw requires 8-byte alignment */ +#define HW_IP_ALIGN 2 /* hw aligns IP header */ +#define WRAP HW_IP_ALIGN + ETH_HLEN + VLAN_HLEN + FCS_LEN #define RX_SKB_SIZE ((dev->mtu + WRAP + 7) & ~0x7) -#define INT_CAUSE_UNMASK_ALL 0x0007ffff -#define INT_CAUSE_UNMASK_ALL_EXT 0x0011ffff -#define INT_CAUSE_MASK_ALL 0x00000000 -#define INT_CAUSE_MASK_ALL_EXT 0x00000000 +#define INT_UNMASK_ALL 0x0007ffff +#define INT_UNMASK_ALL_EXT 0x0011ffff +#define INT_MASK_ALL 0x00000000 +#define INT_MASK_ALL_EXT 0x00000000 #define INT_CAUSE_CHECK_BITS INT_CAUSE_UNMASK_ALL #define INT_CAUSE_CHECK_BITS_EXT INT_CAUSE_UNMASK_ALL_EXT @@ -78,8 +82,9 @@ static int eth_port_link_is_up(unsigned int eth_port_num); static void eth_port_uc_addr_get(struct net_device *dev, unsigned char *MacAddr); -static int mv643xx_eth_real_open(struct net_device *); -static int mv643xx_eth_real_stop(struct net_device *); +static void eth_port_set_multicast_list(struct net_device *); +static int mv643xx_eth_open(struct net_device *); +static int mv643xx_eth_stop(struct net_device *); static int mv643xx_eth_change_mtu(struct net_device *, int); static struct net_device_stats *mv643xx_eth_get_stats(struct net_device *); static void eth_port_init_mac_tables(unsigned int eth_port_num); @@ -124,15 +129,8 @@ static inline void mv_write(int offset, u32 data) */ static int mv643xx_eth_change_mtu(struct net_device *dev, int new_mtu) { - struct mv643xx_private *mp = netdev_priv(dev); - unsigned long flags; - - spin_lock_irqsave(&mp->lock, flags); - - if ((new_mtu > 9500) || (new_mtu < 64)) { - spin_unlock_irqrestore(&mp->lock, flags); + if ((new_mtu > 9500) || (new_mtu < 64)) return -EINVAL; - } dev->mtu = new_mtu; /* @@ -142,17 +140,13 @@ static int mv643xx_eth_change_mtu(struct net_device *dev, int new_mtu) * to memory is full, which might fail the open function. */ if (netif_running(dev)) { - if (mv643xx_eth_real_stop(dev)) - printk(KERN_ERR - "%s: Fatal error on stopping device\n", - dev->name); - if (mv643xx_eth_real_open(dev)) + mv643xx_eth_stop(dev); + if (mv643xx_eth_open(dev)) printk(KERN_ERR "%s: Fatal error on opening device\n", dev->name); } - spin_unlock_irqrestore(&mp->lock, flags); return 0; } @@ -170,15 +164,19 @@ static void mv643xx_eth_rx_task(void *data) struct mv643xx_private *mp = netdev_priv(dev); struct pkt_info pkt_info; struct sk_buff *skb; + int unaligned; if (test_and_set_bit(0, &mp->rx_task_busy)) panic("%s: Error in test_set_bit / clear_bit", dev->name); while (mp->rx_ring_skbs < (mp->rx_ring_size - 5)) { - skb = dev_alloc_skb(RX_SKB_SIZE); + skb = dev_alloc_skb(RX_SKB_SIZE + DMA_ALIGN); if (!skb) break; mp->rx_ring_skbs++; + unaligned = (u32)skb->data & (DMA_ALIGN - 1); + if (unaligned) + skb_reserve(skb, DMA_ALIGN - unaligned); pkt_info.cmd_sts = ETH_RX_ENABLE_INTERRUPT; pkt_info.byte_cnt = RX_SKB_SIZE; pkt_info.buf_ptr = dma_map_single(NULL, skb->data, RX_SKB_SIZE, @@ -189,7 +187,7 @@ static void mv643xx_eth_rx_task(void *data) "%s: Error allocating RX Ring\n", dev->name); break; } - skb_reserve(skb, 2); + skb_reserve(skb, HW_IP_ALIGN); } clear_bit(0, &mp->rx_task_busy); /* @@ -207,7 +205,7 @@ static void mv643xx_eth_rx_task(void *data) else { /* Return interrupts */ mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(mp->port_num), - INT_CAUSE_UNMASK_ALL); + INT_UNMASK_ALL); } #endif } @@ -267,6 +265,8 @@ static void mv643xx_eth_set_rx_mode(struct net_device *dev) mp->port_config &= ~(u32) MV643XX_ETH_UNICAST_PROMISCUOUS_MODE; mv_write(MV643XX_ETH_PORT_CONFIG_REG(mp->port_num), mp->port_config); + + eth_port_set_multicast_list(dev); } /* @@ -342,8 +342,6 @@ static int mv643xx_eth_free_tx_queue(struct net_device *dev, if (!(eth_int_cause_ext & (BIT0 | BIT8))) return released; - spin_lock(&mp->lock); - /* Check only queue 0 */ while (eth_tx_return_desc(mp, &pkt_info) == ETH_OK) { if (pkt_info.cmd_sts & BIT0) { @@ -351,31 +349,21 @@ static int mv643xx_eth_free_tx_queue(struct net_device *dev, stats->tx_errors++; } - /* - * If return_info is different than 0, release the skb. - * The case where return_info is not 0 is only in case - * when transmitted a scatter/gather packet, where only - * last skb releases the whole chain. - */ - if (pkt_info.return_info) { - if (skb_shinfo(pkt_info.return_info)->nr_frags) - dma_unmap_page(NULL, pkt_info.buf_ptr, - pkt_info.byte_cnt, - DMA_TO_DEVICE); - else - dma_unmap_single(NULL, pkt_info.buf_ptr, - pkt_info.byte_cnt, - DMA_TO_DEVICE); + if (pkt_info.cmd_sts & ETH_TX_FIRST_DESC) + dma_unmap_single(NULL, pkt_info.buf_ptr, + pkt_info.byte_cnt, + DMA_TO_DEVICE); + else + dma_unmap_page(NULL, pkt_info.buf_ptr, + pkt_info.byte_cnt, + DMA_TO_DEVICE); + if (pkt_info.return_info) { dev_kfree_skb_irq(pkt_info.return_info); released = 0; - } else - dma_unmap_page(NULL, pkt_info.buf_ptr, - pkt_info.byte_cnt, DMA_TO_DEVICE); + } } - spin_unlock(&mp->lock); - return released; } @@ -482,12 +470,12 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id, /* Read interrupt cause registers */ eth_int_cause = mv_read(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num)) & - INT_CAUSE_UNMASK_ALL; + INT_UNMASK_ALL; if (eth_int_cause & BIT1) eth_int_cause_ext = mv_read( MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num)) & - INT_CAUSE_UNMASK_ALL_EXT; + INT_UNMASK_ALL_EXT; #ifdef MV643XX_NAPI if (!(eth_int_cause & 0x0007fffd)) { @@ -512,9 +500,10 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id, } else { if (netif_rx_schedule_prep(dev)) { /* Mask all the interrupts */ - mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), 0); - mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG - (port_num), 0); + mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), + INT_MASK_ALL); + /* wait for previous write to complete */ + mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num)); __netif_rx_schedule(dev); } #else @@ -527,9 +516,12 @@ static irqreturn_t mv643xx_eth_int_handler(int irq, void *dev_id, * with skb's. */ #ifdef MV643XX_RX_QUEUE_FILL_ON_TASK - /* Unmask all interrupts on ethernet port */ + /* Mask all interrupts on ethernet port */ mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), - INT_CAUSE_MASK_ALL); + INT_MASK_ALL); + /* wait for previous write to take effect */ + mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num)); + queue_task(&mp->rx_task, &tq_immediate); mark_bh(IMMEDIATE_BH); #else @@ -636,56 +628,6 @@ static unsigned int eth_port_set_tx_coal(unsigned int eth_port_num, } /* - * mv643xx_eth_open - * - * This function is called when openning the network device. The function - * should initialize all the hardware, initialize cyclic Rx/Tx - * descriptors chain and buffers and allocate an IRQ to the network - * device. - * - * Input : a pointer to the network device structure - * - * Output : zero of success , nonzero if fails. - */ - -static int mv643xx_eth_open(struct net_device *dev) -{ - struct mv643xx_private *mp = netdev_priv(dev); - unsigned int port_num = mp->port_num; - int err; - - spin_lock_irq(&mp->lock); - - err = request_irq(dev->irq, mv643xx_eth_int_handler, - SA_SHIRQ | SA_SAMPLE_RANDOM, dev->name, dev); - - if (err) { - printk(KERN_ERR "Can not assign IRQ number to MV643XX_eth%d\n", - port_num); - err = -EAGAIN; - goto out; - } - - if (mv643xx_eth_real_open(dev)) { - printk("%s: Error opening interface\n", dev->name); - err = -EBUSY; - goto out_free; - } - - spin_unlock_irq(&mp->lock); - - return 0; - -out_free: - free_irq(dev->irq, dev); - -out: - spin_unlock_irq(&mp->lock); - - return err; -} - -/* * ether_init_rx_desc_ring - Curve a Rx chain desc list and buffer in memory. * * DESCRIPTION: @@ -777,28 +719,37 @@ static void ether_init_tx_desc_ring(struct mv643xx_private *mp) mp->port_tx_queue_command |= 1; } -/* Helper function for mv643xx_eth_open */ -static int mv643xx_eth_real_open(struct net_device *dev) +/* + * mv643xx_eth_open + * + * This function is called when openning the network device. The function + * should initialize all the hardware, initialize cyclic Rx/Tx + * descriptors chain and buffers and allocate an IRQ to the network + * device. + * + * Input : a pointer to the network device structure + * + * Output : zero of success , nonzero if fails. + */ + +static int mv643xx_eth_open(struct net_device *dev) { struct mv643xx_private *mp = netdev_priv(dev); unsigned int port_num = mp->port_num; unsigned int size; + int err; + + err = request_irq(dev->irq, mv643xx_eth_int_handler, + SA_SHIRQ | SA_SAMPLE_RANDOM, dev->name, dev); + if (err) { + printk(KERN_ERR "Can not assign IRQ number to MV643XX_eth%d\n", + port_num); + return -EAGAIN; + } /* Stop RX Queues */ mv_write(MV643XX_ETH_RECEIVE_QUEUE_COMMAND_REG(port_num), 0x0000ff00); - /* Clear the ethernet port interrupts */ - mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0); - mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0); - - /* Unmask RX buffer and TX end interrupt */ - mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), - INT_CAUSE_UNMASK_ALL); - - /* Unmask phy and link status changes interrupts */ - mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), - INT_CAUSE_UNMASK_ALL_EXT); - /* Set the MAC Address */ memcpy(mp->port_mac_addr, dev->dev_addr, 6); @@ -818,14 +769,15 @@ static int mv643xx_eth_real_open(struct net_device *dev) GFP_KERNEL); if (!mp->rx_skb) { printk(KERN_ERR "%s: Cannot allocate Rx skb ring\n", dev->name); - return -ENOMEM; + err = -ENOMEM; + goto out_free_irq; } mp->tx_skb = kmalloc(sizeof(*mp->tx_skb) * mp->tx_ring_size, GFP_KERNEL); if (!mp->tx_skb) { printk(KERN_ERR "%s: Cannot allocate Tx skb ring\n", dev->name); - kfree(mp->rx_skb); - return -ENOMEM; + err = -ENOMEM; + goto out_free_rx_skb; } /* Allocate TX ring */ @@ -845,9 +797,8 @@ static int mv643xx_eth_real_open(struct net_device *dev) if (!mp->p_tx_desc_area) { printk(KERN_ERR "%s: Cannot allocate Tx Ring (size %d bytes)\n", dev->name, size); - kfree(mp->rx_skb); - kfree(mp->tx_skb); - return -ENOMEM; + err = -ENOMEM; + goto out_free_tx_skb; } BUG_ON((u32) mp->p_tx_desc_area & 0xf); /* check 16-byte alignment */ memset((void *)mp->p_tx_desc_area, 0, mp->tx_desc_area_size); @@ -874,13 +825,12 @@ static int mv643xx_eth_real_open(struct net_device *dev) printk(KERN_ERR "%s: Freeing previously allocated TX queues...", dev->name); if (mp->rx_sram_size) - iounmap(mp->p_rx_desc_area); + iounmap(mp->p_tx_desc_area); else dma_free_coherent(NULL, mp->tx_desc_area_size, mp->p_tx_desc_area, mp->tx_desc_dma); - kfree(mp->rx_skb); - kfree(mp->tx_skb); - return -ENOMEM; + err = -ENOMEM; + goto out_free_tx_skb; } memset((void *)mp->p_rx_desc_area, 0, size); @@ -900,9 +850,26 @@ static int mv643xx_eth_real_open(struct net_device *dev) mp->tx_int_coal = eth_port_set_tx_coal(port_num, 133000000, MV643XX_TX_COAL); - netif_start_queue(dev); + /* Clear any pending ethernet port interrupts */ + mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0); + mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0); + + /* Unmask phy and link status changes interrupts */ + mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), + INT_UNMASK_ALL_EXT); + /* Unmask RX buffer and TX end interrupt */ + mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_UNMASK_ALL); return 0; + +out_free_tx_skb: + kfree(mp->tx_skb); +out_free_rx_skb: + kfree(mp->rx_skb); +out_free_irq: + free_irq(dev->irq, dev); + + return err; } static void mv643xx_eth_free_tx_rings(struct net_device *dev) @@ -910,14 +877,17 @@ static void mv643xx_eth_free_tx_rings(struct net_device *dev) struct mv643xx_private *mp = netdev_priv(dev); unsigned int port_num = mp->port_num; unsigned int curr; + struct sk_buff *skb; /* Stop Tx Queues */ mv_write(MV643XX_ETH_TRANSMIT_QUEUE_COMMAND_REG(port_num), 0x0000ff00); /* Free outstanding skb's on TX rings */ for (curr = 0; mp->tx_ring_skbs && curr < mp->tx_ring_size; curr++) { - if (mp->tx_skb[curr]) { - dev_kfree_skb(mp->tx_skb[curr]); + skb = mp->tx_skb[curr]; + if (skb) { + mp->tx_ring_skbs -= skb_shinfo(skb)->nr_frags; + dev_kfree_skb(skb); mp->tx_ring_skbs--; } } @@ -973,44 +943,32 @@ static void mv643xx_eth_free_rx_rings(struct net_device *dev) * Output : zero if success , nonzero if fails */ -/* Helper function for mv643xx_eth_stop */ - -static int mv643xx_eth_real_stop(struct net_device *dev) +static int mv643xx_eth_stop(struct net_device *dev) { struct mv643xx_private *mp = netdev_priv(dev); unsigned int port_num = mp->port_num; + /* Mask all interrupts on ethernet port */ + mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_MASK_ALL); + /* wait for previous write to complete */ + mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num)); + +#ifdef MV643XX_NAPI + netif_poll_disable(dev); +#endif netif_carrier_off(dev); netif_stop_queue(dev); - mv643xx_eth_free_tx_rings(dev); - mv643xx_eth_free_rx_rings(dev); - eth_port_reset(mp->port_num); - /* Disable ethernet port interrupts */ - mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0); - mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0); - - /* Mask RX buffer and TX end interrupt */ - mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), 0); - - /* Mask phy and link status changes interrupts */ - mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), 0); - - return 0; -} - -static int mv643xx_eth_stop(struct net_device *dev) -{ - struct mv643xx_private *mp = netdev_priv(dev); - - spin_lock_irq(&mp->lock); + mv643xx_eth_free_tx_rings(dev); + mv643xx_eth_free_rx_rings(dev); - mv643xx_eth_real_stop(dev); +#ifdef MV643XX_NAPI + netif_poll_enable(dev); +#endif free_irq(dev->irq, dev); - spin_unlock_irq(&mp->lock); return 0; } @@ -1022,20 +980,17 @@ static void mv643xx_tx(struct net_device *dev) struct pkt_info pkt_info; while (eth_tx_return_desc(mp, &pkt_info) == ETH_OK) { - if (pkt_info.return_info) { - if (skb_shinfo(pkt_info.return_info)->nr_frags) - dma_unmap_page(NULL, pkt_info.buf_ptr, - pkt_info.byte_cnt, - DMA_TO_DEVICE); - else - dma_unmap_single(NULL, pkt_info.buf_ptr, - pkt_info.byte_cnt, - DMA_TO_DEVICE); + if (pkt_info.cmd_sts & ETH_TX_FIRST_DESC) + dma_unmap_single(NULL, pkt_info.buf_ptr, + pkt_info.byte_cnt, + DMA_TO_DEVICE); + else + dma_unmap_page(NULL, pkt_info.buf_ptr, + pkt_info.byte_cnt, + DMA_TO_DEVICE); + if (pkt_info.return_info) dev_kfree_skb_irq(pkt_info.return_info); - } else - dma_unmap_page(NULL, pkt_info.buf_ptr, - pkt_info.byte_cnt, DMA_TO_DEVICE); } if (netif_queue_stopped(dev) && @@ -1053,14 +1008,11 @@ static int mv643xx_poll(struct net_device *dev, int *budget) struct mv643xx_private *mp = netdev_priv(dev); int done = 1, orig_budget, work_done; unsigned int port_num = mp->port_num; - unsigned long flags; #ifdef MV643XX_TX_FAST_REFILL if (++mp->tx_clean_threshold > 5) { - spin_lock_irqsave(&mp->lock, flags); mv643xx_tx(dev); mp->tx_clean_threshold = 0; - spin_unlock_irqrestore(&mp->lock, flags); } #endif @@ -1078,21 +1030,36 @@ static int mv643xx_poll(struct net_device *dev, int *budget) } if (done) { - spin_lock_irqsave(&mp->lock, flags); - __netif_rx_complete(dev); + netif_rx_complete(dev); mv_write(MV643XX_ETH_INTERRUPT_CAUSE_REG(port_num), 0); mv_write(MV643XX_ETH_INTERRUPT_CAUSE_EXTEND_REG(port_num), 0); mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), - INT_CAUSE_UNMASK_ALL); - mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), - INT_CAUSE_UNMASK_ALL_EXT); - spin_unlock_irqrestore(&mp->lock, flags); + INT_UNMASK_ALL); } return done ? 0 : 1; } #endif +/* Hardware can't handle unaligned fragments smaller than 9 bytes. + * This helper function detects that case. + */ + +static inline unsigned int has_tiny_unaligned_frags(struct sk_buff *skb) +{ + unsigned int frag; + skb_frag_t *fragp; + + for (frag = 0; frag < skb_shinfo(skb)->nr_frags; frag++) { + fragp = &skb_shinfo(skb)->frags[frag]; + if (fragp->size <= 8 && fragp->page_offset & 0x7) + return 1; + + } + return 0; +} + + /* * mv643xx_eth_start_xmit * @@ -1136,12 +1103,19 @@ static int mv643xx_eth_start_xmit(struct sk_buff *skb, struct net_device *dev) return 1; } +#ifdef MV643XX_CHECKSUM_OFFLOAD_TX + if (has_tiny_unaligned_frags(skb)) { + if ((skb_linearize(skb, GFP_ATOMIC) != 0)) { + stats->tx_dropped++; + printk(KERN_DEBUG "%s: failed to linearize tiny " + "unaligned fragment\n", dev->name); + return 1; + } + } + spin_lock_irqsave(&mp->lock, flags); - /* Update packet info data structure -- DMA owned, first last */ -#ifdef MV643XX_CHECKSUM_OFFLOAD_TX if (!skb_shinfo(skb)->nr_frags) { -linear: if (skb->ip_summed != CHECKSUM_HW) { /* Errata BTS #50, IHL must be 5 if no HW checksum */ pkt_info.cmd_sts = ETH_TX_ENABLE_INTERRUPT | @@ -1150,7 +1124,6 @@ linear: 5 << ETH_TX_IHL_SHIFT; pkt_info.l4i_chk = 0; } else { - pkt_info.cmd_sts = ETH_TX_ENABLE_INTERRUPT | ETH_TX_FIRST_DESC | ETH_TX_LAST_DESC | @@ -1158,14 +1131,16 @@ linear: ETH_GEN_IP_V_4_CHECKSUM | skb->nh.iph->ihl << ETH_TX_IHL_SHIFT; /* CPU already calculated pseudo header checksum. */ - if (skb->nh.iph->protocol == IPPROTO_UDP) { + if ((skb->protocol == ETH_P_IP) && + (skb->nh.iph->protocol == IPPROTO_UDP) ) { pkt_info.cmd_sts |= ETH_UDP_FRAME; pkt_info.l4i_chk = skb->h.uh->check; - } else if (skb->nh.iph->protocol == IPPROTO_TCP) + } else if ((skb->protocol == ETH_P_IP) && + (skb->nh.iph->protocol == IPPROTO_TCP)) pkt_info.l4i_chk = skb->h.th->check; else { printk(KERN_ERR - "%s: chksum proto != TCP or UDP\n", + "%s: chksum proto != IPv4 TCP or UDP\n", dev->name); spin_unlock_irqrestore(&mp->lock, flags); return 1; @@ -1183,26 +1158,6 @@ linear: } else { unsigned int frag; - /* Since hardware can't handle unaligned fragments smaller - * than 9 bytes, if we find any, we linearize the skb - * and start again. When I've seen it, it's always been - * the first frag (probably near the end of the page), - * but we check all frags to be safe. - */ - for (frag = 0; frag < skb_shinfo(skb)->nr_frags; frag++) { - skb_frag_t *fragp; - - fragp = &skb_shinfo(skb)->frags[frag]; - if (fragp->size <= 8 && fragp->page_offset & 0x7) { - skb_linearize(skb, GFP_ATOMIC); - printk(KERN_DEBUG "%s: unaligned tiny fragment" - "%d of %d, fixed\n", - dev->name, frag, - skb_shinfo(skb)->nr_frags); - goto linear; - } - } - /* first frag which is skb header */ pkt_info.byte_cnt = skb_headlen(skb); pkt_info.buf_ptr = dma_map_single(NULL, skb->data, @@ -1221,14 +1176,16 @@ linear: ETH_GEN_IP_V_4_CHECKSUM | skb->nh.iph->ihl << ETH_TX_IHL_SHIFT; /* CPU already calculated pseudo header checksum. */ - if (skb->nh.iph->protocol == IPPROTO_UDP) { + if ((skb->protocol == ETH_P_IP) && + (skb->nh.iph->protocol == IPPROTO_UDP)) { pkt_info.cmd_sts |= ETH_UDP_FRAME; pkt_info.l4i_chk = skb->h.uh->check; - } else if (skb->nh.iph->protocol == IPPROTO_TCP) + } else if ((skb->protocol == ETH_P_IP) && + (skb->nh.iph->protocol == IPPROTO_TCP)) pkt_info.l4i_chk = skb->h.th->check; else { printk(KERN_ERR - "%s: chksum proto != TCP or UDP\n", + "%s: chksum proto != IPv4 TCP or UDP\n", dev->name); spin_unlock_irqrestore(&mp->lock, flags); return 1; @@ -1288,6 +1245,8 @@ linear: } } #else + spin_lock_irqsave(&mp->lock, flags); + pkt_info.cmd_sts = ETH_TX_ENABLE_INTERRUPT | ETH_TX_FIRST_DESC | ETH_TX_LAST_DESC; pkt_info.l4i_chk = 0; @@ -1340,39 +1299,18 @@ static struct net_device_stats *mv643xx_eth_get_stats(struct net_device *dev) } #ifdef CONFIG_NET_POLL_CONTROLLER -static inline void mv643xx_enable_irq(struct mv643xx_private *mp) -{ - int port_num = mp->port_num; - unsigned long flags; - - spin_lock_irqsave(&mp->lock, flags); - mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), - INT_CAUSE_UNMASK_ALL); - mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), - INT_CAUSE_UNMASK_ALL_EXT); - spin_unlock_irqrestore(&mp->lock, flags); -} - -static inline void mv643xx_disable_irq(struct mv643xx_private *mp) -{ - int port_num = mp->port_num; - unsigned long flags; - - spin_lock_irqsave(&mp->lock, flags); - mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), - INT_CAUSE_MASK_ALL); - mv_write(MV643XX_ETH_INTERRUPT_EXTEND_MASK_REG(port_num), - INT_CAUSE_MASK_ALL_EXT); - spin_unlock_irqrestore(&mp->lock, flags); -} - static void mv643xx_netpoll(struct net_device *netdev) { struct mv643xx_private *mp = netdev_priv(netdev); + int port_num = mp->port_num; + + mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_MASK_ALL); + /* wait for previous write to complete */ + mv_read(MV643XX_ETH_INTERRUPT_MASK_REG(port_num)); - mv643xx_disable_irq(mp); mv643xx_eth_int_handler(netdev->irq, netdev, NULL); - mv643xx_enable_irq(mp); + + mv_write(MV643XX_ETH_INTERRUPT_MASK_REG(port_num), INT_UNMASK_ALL); } #endif @@ -1441,7 +1379,7 @@ static int mv643xx_eth_probe(struct platform_device *pdev) * Zero copy can only work if we use Discovery II memory. Else, we will * have to map the buffers to ISA memory which is only 16 MB */ - dev->features = NETIF_F_SG | NETIF_F_IP_CSUM | NETIF_F_HW_CSUM; + dev->features = NETIF_F_SG | NETIF_F_IP_CSUM; #endif #endif @@ -2054,6 +1992,196 @@ static int eth_port_uc_addr(unsigned int eth_port_num, unsigned char uc_nibble, } /* + * The entries in each table are indexed by a hash of a packet's MAC + * address. One bit in each entry determines whether the packet is + * accepted. There are 4 entries (each 8 bits wide) in each register + * of the table. The bits in each entry are defined as follows: + * 0 Accept=1, Drop=0 + * 3-1 Queue (ETH_Q0=0) + * 7-4 Reserved = 0; + */ +static void eth_port_set_filter_table_entry(int table, unsigned char entry) +{ + unsigned int table_reg; + unsigned int tbl_offset; + unsigned int reg_offset; + + tbl_offset = (entry / 4) * 4; /* Register offset of DA table entry */ + reg_offset = entry % 4; /* Entry offset within the register */ + + /* Set "accepts frame bit" at specified table entry */ + table_reg = mv_read(table + tbl_offset); + table_reg |= 0x01 << (8 * reg_offset); + mv_write(table + tbl_offset, table_reg); +} + +/* + * eth_port_mc_addr - Multicast address settings. + * + * The MV device supports multicast using two tables: + * 1) Special Multicast Table for MAC addresses of the form + * 0x01-00-5E-00-00-XX (where XX is between 0x00 and 0x_FF). + * The MAC DA[7:0] bits are used as a pointer to the Special Multicast + * Table entries in the DA-Filter table. + * 2) Other Multicast Table for multicast of another type. A CRC-8bit + * is used as an index to the Other Multicast Table entries in the + * DA-Filter table. This function calculates the CRC-8bit value. + * In either case, eth_port_set_filter_table_entry() is then called + * to set to set the actual table entry. + */ +static void eth_port_mc_addr(unsigned int eth_port_num, unsigned char *p_addr) +{ + unsigned int mac_h; + unsigned int mac_l; + unsigned char crc_result = 0; + int table; + int mac_array[48]; + int crc[8]; + int i; + + if ((p_addr[0] == 0x01) && (p_addr[1] == 0x00) && + (p_addr[2] == 0x5E) && (p_addr[3] == 0x00) && (p_addr[4] == 0x00)) { + table = MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE + (eth_port_num); + eth_port_set_filter_table_entry(table, p_addr[5]); + return; + } + + /* Calculate CRC-8 out of the given address */ + mac_h = (p_addr[0] << 8) | (p_addr[1]); + mac_l = (p_addr[2] << 24) | (p_addr[3] << 16) | + (p_addr[4] << 8) | (p_addr[5] << 0); + + for (i = 0; i < 32; i++) + mac_array[i] = (mac_l >> i) & 0x1; + for (i = 32; i < 48; i++) + mac_array[i] = (mac_h >> (i - 32)) & 0x1; + + crc[0] = mac_array[45] ^ mac_array[43] ^ mac_array[40] ^ mac_array[39] ^ + mac_array[35] ^ mac_array[34] ^ mac_array[31] ^ mac_array[30] ^ + mac_array[28] ^ mac_array[23] ^ mac_array[21] ^ mac_array[19] ^ + mac_array[18] ^ mac_array[16] ^ mac_array[14] ^ mac_array[12] ^ + mac_array[8] ^ mac_array[7] ^ mac_array[6] ^ mac_array[0]; + + crc[1] = mac_array[46] ^ mac_array[45] ^ mac_array[44] ^ mac_array[43] ^ + mac_array[41] ^ mac_array[39] ^ mac_array[36] ^ mac_array[34] ^ + mac_array[32] ^ mac_array[30] ^ mac_array[29] ^ mac_array[28] ^ + mac_array[24] ^ mac_array[23] ^ mac_array[22] ^ mac_array[21] ^ + mac_array[20] ^ mac_array[18] ^ mac_array[17] ^ mac_array[16] ^ + mac_array[15] ^ mac_array[14] ^ mac_array[13] ^ mac_array[12] ^ + mac_array[9] ^ mac_array[6] ^ mac_array[1] ^ mac_array[0]; + + crc[2] = mac_array[47] ^ mac_array[46] ^ mac_array[44] ^ mac_array[43] ^ + mac_array[42] ^ mac_array[39] ^ mac_array[37] ^ mac_array[34] ^ + mac_array[33] ^ mac_array[29] ^ mac_array[28] ^ mac_array[25] ^ + mac_array[24] ^ mac_array[22] ^ mac_array[17] ^ mac_array[15] ^ + mac_array[13] ^ mac_array[12] ^ mac_array[10] ^ mac_array[8] ^ + mac_array[6] ^ mac_array[2] ^ mac_array[1] ^ mac_array[0]; + + crc[3] = mac_array[47] ^ mac_array[45] ^ mac_array[44] ^ mac_array[43] ^ + mac_array[40] ^ mac_array[38] ^ mac_array[35] ^ mac_array[34] ^ + mac_array[30] ^ mac_array[29] ^ mac_array[26] ^ mac_array[25] ^ + mac_array[23] ^ mac_array[18] ^ mac_array[16] ^ mac_array[14] ^ + mac_array[13] ^ mac_array[11] ^ mac_array[9] ^ mac_array[7] ^ + mac_array[3] ^ mac_array[2] ^ mac_array[1]; + + crc[4] = mac_array[46] ^ mac_array[45] ^ mac_array[44] ^ mac_array[41] ^ + mac_array[39] ^ mac_array[36] ^ mac_array[35] ^ mac_array[31] ^ + mac_array[30] ^ mac_array[27] ^ mac_array[26] ^ mac_array[24] ^ + mac_array[19] ^ mac_array[17] ^ mac_array[15] ^ mac_array[14] ^ + mac_array[12] ^ mac_array[10] ^ mac_array[8] ^ mac_array[4] ^ + mac_array[3] ^ mac_array[2]; + + crc[5] = mac_array[47] ^ mac_array[46] ^ mac_array[45] ^ mac_array[42] ^ + mac_array[40] ^ mac_array[37] ^ mac_array[36] ^ mac_array[32] ^ + mac_array[31] ^ mac_array[28] ^ mac_array[27] ^ mac_array[25] ^ + mac_array[20] ^ mac_array[18] ^ mac_array[16] ^ mac_array[15] ^ + mac_array[13] ^ mac_array[11] ^ mac_array[9] ^ mac_array[5] ^ + mac_array[4] ^ mac_array[3]; + + crc[6] = mac_array[47] ^ mac_array[46] ^ mac_array[43] ^ mac_array[41] ^ + mac_array[38] ^ mac_array[37] ^ mac_array[33] ^ mac_array[32] ^ + mac_array[29] ^ mac_array[28] ^ mac_array[26] ^ mac_array[21] ^ + mac_array[19] ^ mac_array[17] ^ mac_array[16] ^ mac_array[14] ^ + mac_array[12] ^ mac_array[10] ^ mac_array[6] ^ mac_array[5] ^ + mac_array[4]; + + crc[7] = mac_array[47] ^ mac_array[44] ^ mac_array[42] ^ mac_array[39] ^ + mac_array[38] ^ mac_array[34] ^ mac_array[33] ^ mac_array[30] ^ + mac_array[29] ^ mac_array[27] ^ mac_array[22] ^ mac_array[20] ^ + mac_array[18] ^ mac_array[17] ^ mac_array[15] ^ mac_array[13] ^ + mac_array[11] ^ mac_array[7] ^ mac_array[6] ^ mac_array[5]; + + for (i = 0; i < 8; i++) + crc_result = crc_result | (crc[i] << i); + + table = MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE(eth_port_num); + eth_port_set_filter_table_entry(table, crc_result); +} + +/* + * Set the entire multicast list based on dev->mc_list. + */ +static void eth_port_set_multicast_list(struct net_device *dev) +{ + + struct dev_mc_list *mc_list; + int i; + int table_index; + struct mv643xx_private *mp = netdev_priv(dev); + unsigned int eth_port_num = mp->port_num; + + /* If the device is in promiscuous mode or in all multicast mode, + * we will fully populate both multicast tables with accept. + * This is guaranteed to yield a match on all multicast addresses... + */ + if ((dev->flags & IFF_PROMISC) || (dev->flags & IFF_ALLMULTI)) { + for (table_index = 0; table_index <= 0xFC; table_index += 4) { + /* Set all entries in DA filter special multicast + * table (Ex_dFSMT) + * Set for ETH_Q0 for now + * Bits + * 0 Accept=1, Drop=0 + * 3-1 Queue ETH_Q0=0 + * 7-4 Reserved = 0; + */ + mv_write(MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE(eth_port_num) + table_index, 0x01010101); + + /* Set all entries in DA filter other multicast + * table (Ex_dFOMT) + * Set for ETH_Q0 for now + * Bits + * 0 Accept=1, Drop=0 + * 3-1 Queue ETH_Q0=0 + * 7-4 Reserved = 0; + */ + mv_write(MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE(eth_port_num) + table_index, 0x01010101); + } + return; + } + + /* We will clear out multicast tables every time we get the list. + * Then add the entire new list... + */ + for (table_index = 0; table_index <= 0xFC; table_index += 4) { + /* Clear DA filter special multicast table (Ex_dFSMT) */ + mv_write(MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE + (eth_port_num) + table_index, 0); + + /* Clear DA filter other multicast table (Ex_dFOMT) */ + mv_write(MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE + (eth_port_num) + table_index, 0); + } + + /* Get pointer to net_device multicast list and add each one... */ + for (i = 0, mc_list = dev->mc_list; + (i < 256) && (mc_list != NULL) && (i < dev->mc_count); + i++, mc_list = mc_list->next) + if (mc_list->dmi_addrlen == 6) + eth_port_mc_addr(eth_port_num, mc_list->dmi_addr); +} + +/* * eth_port_init_mac_tables - Clear all entrance in the UC, SMC and OMC tables * * DESCRIPTION: @@ -2080,11 +2208,11 @@ static void eth_port_init_mac_tables(unsigned int eth_port_num) for (table_index = 0; table_index <= 0xFC; table_index += 4) { /* Clear DA filter special multicast table (Ex_dFSMT) */ - mv_write((MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE - (eth_port_num) + table_index), 0); + mv_write(MV643XX_ETH_DA_FILTER_SPECIAL_MULTICAST_TABLE_BASE + (eth_port_num) + table_index, 0); /* Clear DA filter other multicast table (Ex_dFOMT) */ - mv_write((MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE - (eth_port_num) + table_index), 0); + mv_write(MV643XX_ETH_DA_FILTER_OTHER_MULTICAST_TABLE_BASE + (eth_port_num) + table_index, 0); } } @@ -2489,6 +2617,7 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp, struct eth_tx_desc *current_descriptor; struct eth_tx_desc *first_descriptor; u32 command; + unsigned long flags; /* Do not process Tx ring in case of Tx ring resource error */ if (mp->tx_resource_err) @@ -2505,6 +2634,8 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp, return ETH_ERROR; } + spin_lock_irqsave(&mp->lock, flags); + mp->tx_ring_skbs++; BUG_ON(mp->tx_ring_skbs > mp->tx_ring_size); @@ -2554,11 +2685,15 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp, mp->tx_resource_err = 1; mp->tx_curr_desc_q = tx_first_desc; + spin_unlock_irqrestore(&mp->lock, flags); + return ETH_QUEUE_LAST_RESOURCE; } mp->tx_curr_desc_q = tx_next_desc; + spin_unlock_irqrestore(&mp->lock, flags); + return ETH_OK; } #else @@ -2569,11 +2704,14 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp, int tx_desc_used; struct eth_tx_desc *current_descriptor; unsigned int command_status; + unsigned long flags; /* Do not process Tx ring in case of Tx ring resource error */ if (mp->tx_resource_err) return ETH_QUEUE_FULL; + spin_lock_irqsave(&mp->lock, flags); + mp->tx_ring_skbs++; BUG_ON(mp->tx_ring_skbs > mp->tx_ring_size); @@ -2604,9 +2742,12 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp, /* Check for ring index overlap in the Tx desc ring */ if (tx_desc_curr == tx_desc_used) { mp->tx_resource_err = 1; + + spin_unlock_irqrestore(&mp->lock, flags); return ETH_QUEUE_LAST_RESOURCE; } + spin_unlock_irqrestore(&mp->lock, flags); return ETH_OK; } #endif @@ -2629,23 +2770,27 @@ static ETH_FUNC_RET_STATUS eth_port_send(struct mv643xx_private *mp, * Tx ring 'first' and 'used' indexes are updated. * * RETURN: - * ETH_ERROR in case the routine can not access Tx desc ring. - * ETH_RETRY in case there is transmission in process. - * ETH_END_OF_JOB if the routine has nothing to release. - * ETH_OK otherwise. + * ETH_OK on success + * ETH_ERROR otherwise. * */ static ETH_FUNC_RET_STATUS eth_tx_return_desc(struct mv643xx_private *mp, struct pkt_info *p_pkt_info) { int tx_desc_used; + int tx_busy_desc; + struct eth_tx_desc *p_tx_desc_used; + unsigned int command_status; + unsigned long flags; + int err = ETH_OK; + + spin_lock_irqsave(&mp->lock, flags); + #ifdef MV643XX_CHECKSUM_OFFLOAD_TX - int tx_busy_desc = mp->tx_first_desc_q; + tx_busy_desc = mp->tx_first_desc_q; #else - int tx_busy_desc = mp->tx_curr_desc_q; + tx_busy_desc = mp->tx_curr_desc_q; #endif - struct eth_tx_desc *p_tx_desc_used; - unsigned int command_status; /* Get the Tx Desc ring indexes */ tx_desc_used = mp->tx_used_desc_q; @@ -2653,22 +2798,30 @@ static ETH_FUNC_RET_STATUS eth_tx_return_desc(struct mv643xx_private *mp, p_tx_desc_used = &mp->p_tx_desc_area[tx_desc_used]; /* Sanity check */ - if (p_tx_desc_used == NULL) - return ETH_ERROR; + if (p_tx_desc_used == NULL) { + err = ETH_ERROR; + goto out; + } /* Stop release. About to overlap the current available Tx descriptor */ - if (tx_desc_used == tx_busy_desc && !mp->tx_resource_err) - return ETH_END_OF_JOB; + if (tx_desc_used == tx_busy_desc && !mp->tx_resource_err) { + err = ETH_ERROR; + goto out; + } command_status = p_tx_desc_used->cmd_sts; /* Still transmitting... */ - if (command_status & (ETH_BUFFER_OWNED_BY_DMA)) - return ETH_RETRY; + if (command_status & (ETH_BUFFER_OWNED_BY_DMA)) { + err = ETH_ERROR; + goto out; + } /* Pass the packet information to the caller */ p_pkt_info->cmd_sts = command_status; p_pkt_info->return_info = mp->tx_skb[tx_desc_used]; + p_pkt_info->buf_ptr = p_tx_desc_used->buf_ptr; + p_pkt_info->byte_cnt = p_tx_desc_used->byte_cnt; mp->tx_skb[tx_desc_used] = NULL; /* Update the next descriptor to release. */ @@ -2680,7 +2833,10 @@ static ETH_FUNC_RET_STATUS eth_tx_return_desc(struct mv643xx_private *mp, BUG_ON(mp->tx_ring_skbs == 0); mp->tx_ring_skbs--; - return ETH_OK; +out: + spin_unlock_irqrestore(&mp->lock, flags); + + return err; } /* @@ -2712,11 +2868,14 @@ static ETH_FUNC_RET_STATUS eth_port_receive(struct mv643xx_private *mp, int rx_next_curr_desc, rx_curr_desc, rx_used_desc; volatile struct eth_rx_desc *p_rx_desc; unsigned int command_status; + unsigned long flags; /* Do not process Rx ring in case of Rx ring resource error */ if (mp->rx_resource_err) return ETH_QUEUE_FULL; + spin_lock_irqsave(&mp->lock, flags); + /* Get the Rx Desc ring 'curr and 'used' indexes */ rx_curr_desc = mp->rx_curr_desc_q; rx_used_desc = mp->rx_used_desc_q; @@ -2728,8 +2887,10 @@ static ETH_FUNC_RET_STATUS eth_port_receive(struct mv643xx_private *mp, rmb(); /* Nothing to receive... */ - if (command_status & (ETH_BUFFER_OWNED_BY_DMA)) + if (command_status & (ETH_BUFFER_OWNED_BY_DMA)) { + spin_unlock_irqrestore(&mp->lock, flags); return ETH_END_OF_JOB; + } p_pkt_info->byte_cnt = (p_rx_desc->byte_cnt) - RX_BUF_OFFSET; p_pkt_info->cmd_sts = command_status; @@ -2749,6 +2910,8 @@ static ETH_FUNC_RET_STATUS eth_port_receive(struct mv643xx_private *mp, if (rx_next_curr_desc == rx_used_desc) mp->rx_resource_err = 1; + spin_unlock_irqrestore(&mp->lock, flags); + return ETH_OK; } @@ -2777,6 +2940,9 @@ static ETH_FUNC_RET_STATUS eth_rx_return_buff(struct mv643xx_private *mp, { int used_rx_desc; /* Where to return Rx resource */ volatile struct eth_rx_desc *p_used_rx_desc; + unsigned long flags; + + spin_lock_irqsave(&mp->lock, flags); /* Get 'used' Rx descriptor */ used_rx_desc = mp->rx_used_desc_q; @@ -2800,6 +2966,8 @@ static ETH_FUNC_RET_STATUS eth_rx_return_buff(struct mv643xx_private *mp, /* Any Rx return cancels the Rx resource error status */ mp->rx_resource_err = 0; + spin_unlock_irqrestore(&mp->lock, flags); + return ETH_OK; } diff --git a/drivers/net/skge.c b/drivers/net/skge.c index b538e303805..bf55a4cfb3d 100644 --- a/drivers/net/skge.c +++ b/drivers/net/skge.c @@ -3243,12 +3243,22 @@ static int __devinit skge_probe(struct pci_dev *pdev, pci_set_master(pdev); - if (!(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) + if (sizeof(dma_addr_t) > sizeof(u32) && + !(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) { using_dac = 1; - else if (!(err = pci_set_dma_mask(pdev, DMA_32BIT_MASK))) { - printk(KERN_ERR PFX "%s no usable DMA configuration\n", - pci_name(pdev)); - goto err_out_free_regions; + err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK); + if (err < 0) { + printk(KERN_ERR PFX "%s unable to obtain 64 bit DMA " + "for consistent allocations\n", pci_name(pdev)); + goto err_out_free_regions; + } + } else { + err = pci_set_dma_mask(pdev, DMA_32BIT_MASK); + if (err) { + printk(KERN_ERR PFX "%s no usable DMA configuration\n", + pci_name(pdev)); + goto err_out_free_regions; + } } #ifdef __BIG_ENDIAN diff --git a/drivers/net/sky2.c b/drivers/net/sky2.c index f5d697c0c03..f8b973a04b6 100644 --- a/drivers/net/sky2.c +++ b/drivers/net/sky2.c @@ -57,7 +57,7 @@ #include "sky2.h" #define DRV_NAME "sky2" -#define DRV_VERSION "0.11" +#define DRV_VERSION "0.13" #define PFX DRV_NAME " " /* @@ -75,6 +75,7 @@ #define RX_LE_BYTES (RX_LE_SIZE*sizeof(struct sky2_rx_le)) #define RX_MAX_PENDING (RX_LE_SIZE/2 - 2) #define RX_DEF_PENDING RX_MAX_PENDING +#define RX_SKB_ALIGN 8 #define TX_RING_SIZE 512 #define TX_DEF_PENDING (TX_RING_SIZE - 1) @@ -91,7 +92,7 @@ static const u32 default_msg = NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK | NETIF_MSG_TIMER | NETIF_MSG_TX_ERR | NETIF_MSG_RX_ERR - | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN | NETIF_MSG_INTR; + | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN; static int debug = -1; /* defaults above */ module_param(debug, int, 0); @@ -624,13 +625,16 @@ static void sky2_mac_init(struct sky2_hw *hw, unsigned port) } -static void sky2_ramset(struct sky2_hw *hw, u16 q, u32 start, size_t len) +/* Assign Ram Buffer allocation. + * start and end are in units of 4k bytes + * ram registers are in units of 64bit words + */ +static void sky2_ramset(struct sky2_hw *hw, u16 q, u8 startk, u8 endk) { - u32 end; + u32 start, end; - start /= 8; - len /= 8; - end = start + len - 1; + start = startk * 4096/8; + end = (endk * 4096/8) - 1; sky2_write8(hw, RB_ADDR(q, RB_CTRL), RB_RST_CLR); sky2_write32(hw, RB_ADDR(q, RB_START), start); @@ -639,14 +643,19 @@ static void sky2_ramset(struct sky2_hw *hw, u16 q, u32 start, size_t len) sky2_write32(hw, RB_ADDR(q, RB_RP), start); if (q == Q_R1 || q == Q_R2) { - u32 rxup, rxlo; + u32 space = (endk - startk) * 4096/8; + u32 tp = space - space/4; - rxlo = len/2; - rxup = rxlo + len/4; + /* On receive queue's set the thresholds + * give receiver priority when > 3/4 full + * send pause when down to 2K + */ + sky2_write32(hw, RB_ADDR(q, RB_RX_UTHP), tp); + sky2_write32(hw, RB_ADDR(q, RB_RX_LTHP), space/2); - /* Set thresholds on receive queue's */ - sky2_write32(hw, RB_ADDR(q, RB_RX_UTPP), rxup); - sky2_write32(hw, RB_ADDR(q, RB_RX_LTPP), rxlo); + tp = space - 2048/8; + sky2_write32(hw, RB_ADDR(q, RB_RX_UTPP), tp); + sky2_write32(hw, RB_ADDR(q, RB_RX_LTPP), space/4); } else { /* Enable store & forward on Tx queue's because * Tx FIFO is only 1K on Yukon @@ -695,9 +704,10 @@ static inline struct sky2_tx_le *get_tx_le(struct sky2_port *sky2) * This is a workaround code taken from SysKonnect sk98lin driver * to deal with chip bug on Yukon EC rev 0 in the wraparound case. */ -static inline void sky2_put_idx(struct sky2_hw *hw, unsigned q, +static void sky2_put_idx(struct sky2_hw *hw, unsigned q, u16 idx, u16 *last, u16 size) { + wmb(); if (is_ec_a1(hw) && idx < *last) { u16 hwget = sky2_read16(hw, Y2_QADDR(q, PREF_UNIT_GET_IDX)); @@ -721,6 +731,7 @@ setnew: sky2_write16(hw, Y2_QADDR(q, PREF_UNIT_PUT_IDX), idx); } *last = idx; + mmiowb(); } @@ -734,11 +745,11 @@ static inline struct sky2_rx_le *sky2_next_rx(struct sky2_port *sky2) /* Return high part of DMA address (could be 32 or 64 bit) */ static inline u32 high32(dma_addr_t a) { - return (a >> 16) >> 16; + return sizeof(a) > sizeof(u32) ? (a >> 16) >> 16 : 0; } /* Build description to hardware about buffer */ -static inline void sky2_rx_add(struct sky2_port *sky2, dma_addr_t map) +static void sky2_rx_add(struct sky2_port *sky2, dma_addr_t map) { struct sky2_rx_le *le; u32 hi = high32(map); @@ -878,13 +889,13 @@ static void sky2_vlan_rx_register(struct net_device *dev, struct vlan_group *grp struct sky2_hw *hw = sky2->hw; u16 port = sky2->port; - spin_lock(&sky2->tx_lock); + spin_lock_bh(&sky2->tx_lock); sky2_write32(hw, SK_REG(port, RX_GMF_CTRL_T), RX_VLAN_STRIP_ON); sky2_write32(hw, SK_REG(port, TX_GMF_CTRL_T), TX_VLAN_TAG_ON); sky2->vlgrp = grp; - spin_unlock(&sky2->tx_lock); + spin_unlock_bh(&sky2->tx_lock); } static void sky2_vlan_rx_kill_vid(struct net_device *dev, unsigned short vid) @@ -893,27 +904,42 @@ static void sky2_vlan_rx_kill_vid(struct net_device *dev, unsigned short vid) struct sky2_hw *hw = sky2->hw; u16 port = sky2->port; - spin_lock(&sky2->tx_lock); + spin_lock_bh(&sky2->tx_lock); sky2_write32(hw, SK_REG(port, RX_GMF_CTRL_T), RX_VLAN_STRIP_OFF); sky2_write32(hw, SK_REG(port, TX_GMF_CTRL_T), TX_VLAN_TAG_OFF); if (sky2->vlgrp) sky2->vlgrp->vlan_devices[vid] = NULL; - spin_unlock(&sky2->tx_lock); + spin_unlock_bh(&sky2->tx_lock); } #endif /* + * It appears the hardware has a bug in the FIFO logic that + * cause it to hang if the FIFO gets overrun and the receive buffer + * is not aligned. ALso alloc_skb() won't align properly if slab + * debugging is enabled. + */ +static inline struct sk_buff *sky2_alloc_skb(unsigned int size, gfp_t gfp_mask) +{ + struct sk_buff *skb; + + skb = alloc_skb(size + RX_SKB_ALIGN, gfp_mask); + if (likely(skb)) { + unsigned long p = (unsigned long) skb->data; + skb_reserve(skb, + ((p + RX_SKB_ALIGN - 1) & ~(RX_SKB_ALIGN - 1)) - p); + } + + return skb; +} + +/* * Allocate and setup receiver buffer pool. * In case of 64 bit dma, there are 2X as many list elements * available as ring entries * and need to reserve one list element so we don't wrap around. - * - * It appears the hardware has a bug in the FIFO logic that - * cause it to hang if the FIFO gets overrun and the receive buffer - * is not aligned. This means we can't use skb_reserve to align - * the IP header. */ static int sky2_rx_start(struct sky2_port *sky2) { @@ -929,7 +955,7 @@ static int sky2_rx_start(struct sky2_port *sky2) for (i = 0; i < sky2->rx_pending; i++) { struct ring_info *re = sky2->rx_ring + i; - re->skb = dev_alloc_skb(sky2->rx_bufsize); + re->skb = sky2_alloc_skb(sky2->rx_bufsize, GFP_KERNEL); if (!re->skb) goto nomem; @@ -986,19 +1012,19 @@ static int sky2_up(struct net_device *dev) sky2_mac_init(hw, port); - /* Configure RAM buffers */ - if (hw->chip_id == CHIP_ID_YUKON_FE || - (hw->chip_id == CHIP_ID_YUKON_EC && hw->chip_rev == 2)) - ramsize = 4096; - else { - u8 e0 = sky2_read8(hw, B2_E_0); - ramsize = (e0 == 0) ? (128 * 1024) : (e0 * 4096); - } + /* Determine available ram buffer space (in 4K blocks). + * Note: not sure about the FE setting below yet + */ + if (hw->chip_id == CHIP_ID_YUKON_FE) + ramsize = 4; + else + ramsize = sky2_read8(hw, B2_E_0); + + /* Give transmitter one third (rounded up) */ + rxspace = ramsize - (ramsize + 2) / 3; - /* 2/3 for Rx */ - rxspace = (2 * ramsize) / 3; sky2_ramset(hw, rxqaddr[port], 0, rxspace); - sky2_ramset(hw, txqaddr[port], rxspace, ramsize - rxspace); + sky2_ramset(hw, txqaddr[port], rxspace, ramsize); /* Make sure SyncQ is disabled */ sky2_write8(hw, RB_ADDR(port == 0 ? Q_XS1 : Q_XS2, RB_CTRL), @@ -1054,7 +1080,7 @@ static inline int tx_avail(const struct sky2_port *sky2) } /* Estimate of number of transmit list elements required */ -static inline unsigned tx_le_req(const struct sk_buff *skb) +static unsigned tx_le_req(const struct sk_buff *skb) { unsigned count; @@ -1090,6 +1116,10 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev) u16 mss; u8 ctrl; + /* No BH disabling for tx_lock here. We are running in BH disabled + * context and TX reclaim runs via poll inside of a software + * interrupt, and no related locks in IRQ processing. + */ if (!spin_trylock(&sky2->tx_lock)) return NETDEV_TX_LOCKED; @@ -1099,8 +1129,9 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev) */ if (!netif_queue_stopped(dev)) { netif_stop_queue(dev); - printk(KERN_WARNING PFX "%s: ring full when queue awake!\n", - dev->name); + if (net_ratelimit()) + printk(KERN_WARNING PFX "%s: ring full when queue awake!\n", + dev->name); } spin_unlock(&sky2->tx_lock); @@ -1199,7 +1230,7 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev) mapping = pci_map_page(hw->pdev, frag->page, frag->page_offset, frag->size, PCI_DMA_TODEVICE); - addr64 = (mapping >> 16) >> 16; + addr64 = high32(mapping); if (addr64 != sky2->tx_addr64) { le = get_tx_le(sky2); le->tx.addr = cpu_to_le32(addr64); @@ -1229,7 +1260,6 @@ static int sky2_xmit_frame(struct sk_buff *skb, struct net_device *dev) netif_stop_queue(dev); out_unlock: - mmiowb(); spin_unlock(&sky2->tx_lock); dev->trans_start = jiffies; @@ -1282,17 +1312,17 @@ static void sky2_tx_complete(struct sky2_port *sky2, u16 done) dev_kfree_skb_any(skb); } - spin_lock(&sky2->tx_lock); sky2->tx_cons = put; if (netif_queue_stopped(dev) && tx_avail(sky2) > MAX_SKB_TX_LE) netif_wake_queue(dev); - spin_unlock(&sky2->tx_lock); } /* Cleanup all untransmitted buffers, assume transmitter not running */ static void sky2_tx_clean(struct sky2_port *sky2) { + spin_lock_bh(&sky2->tx_lock); sky2_tx_complete(sky2, sky2->tx_prod); + spin_unlock_bh(&sky2->tx_lock); } /* Network shutdown */ @@ -1582,28 +1612,40 @@ out: local_irq_enable(); } + +/* Transmit timeout is only called if we are running, carries is up + * and tx queue is full (stopped). + */ static void sky2_tx_timeout(struct net_device *dev) { struct sky2_port *sky2 = netdev_priv(dev); struct sky2_hw *hw = sky2->hw; unsigned txq = txqaddr[sky2->port]; + u16 ridx; + + /* Maybe we just missed an status interrupt */ + spin_lock(&sky2->tx_lock); + ridx = sky2_read16(hw, + sky2->port == 0 ? STAT_TXA1_RIDX : STAT_TXA2_RIDX); + sky2_tx_complete(sky2, ridx); + spin_unlock(&sky2->tx_lock); + + if (!netif_queue_stopped(dev)) { + if (net_ratelimit()) + pr_info(PFX "transmit interrupt missed? recovered\n"); + return; + } if (netif_msg_timer(sky2)) printk(KERN_ERR PFX "%s: tx timeout\n", dev->name); - netif_stop_queue(dev); - sky2_write32(hw, Q_ADDR(txq, Q_CSR), BMU_STOP); - sky2_read32(hw, Q_ADDR(txq, Q_CSR)); - sky2_write32(hw, Y2_QADDR(txq, PREF_UNIT_CTRL), PREF_UNIT_RST_SET); sky2_tx_clean(sky2); sky2_qset(hw, txq); sky2_prefetch_init(hw, txq, sky2->tx_le_map, TX_RING_SIZE - 1); - - netif_wake_queue(dev); } @@ -1713,7 +1755,7 @@ static struct sk_buff *sky2_receive(struct sky2_port *sky2, } else { struct sk_buff *nskb; - nskb = dev_alloc_skb(sky2->rx_bufsize); + nskb = sky2_alloc_skb(sky2->rx_bufsize, GFP_ATOMIC); if (!nskb) goto resubmit; @@ -1745,7 +1787,7 @@ oversize: error: ++sky2->net_stats.rx_errors; - if (netif_msg_rx_err(sky2)) + if (netif_msg_rx_err(sky2) && net_ratelimit()) printk(KERN_INFO PFX "%s: rx error, status 0x%x length %d\n", sky2->netdev->name, status, length); @@ -1766,13 +1808,16 @@ error: */ #define TX_NO_STATUS 0xffff -static inline void sky2_tx_check(struct sky2_hw *hw, int port, u16 last) +static void sky2_tx_check(struct sky2_hw *hw, int port, u16 last) { if (last != TX_NO_STATUS) { struct net_device *dev = hw->dev[port]; if (dev && netif_running(dev)) { struct sky2_port *sky2 = netdev_priv(dev); + + spin_lock(&sky2->tx_lock); sky2_tx_complete(sky2, last); + spin_unlock(&sky2->tx_lock); } } } @@ -1800,7 +1845,6 @@ static int sky2_poll(struct net_device *dev0, int *budget) struct sk_buff *skb; u32 status; u16 length; - u8 op; le = hw->st_le + hw->st_idx; hw->st_idx = (hw->st_idx + 1) % STATUS_RING_SIZE; @@ -1814,10 +1858,8 @@ static int sky2_poll(struct net_device *dev0, int *budget) sky2 = netdev_priv(dev); status = le32_to_cpu(le->status); length = le16_to_cpu(le->length); - op = le->opcode & ~HW_OWNER; - le->opcode = 0; - switch (op) { + switch (le->opcode & ~HW_OWNER) { case OP_RXSTAT: skb = sky2_receive(sky2, length, status); if (!skb) @@ -1865,14 +1907,13 @@ static int sky2_poll(struct net_device *dev0, int *budget) default: if (net_ratelimit()) printk(KERN_WARNING PFX - "unknown status opcode 0x%x\n", op); + "unknown status opcode 0x%x\n", le->opcode); break; } } exit_loop: sky2_write32(hw, STAT_CTRL, SC_STAT_CLR_IRQ); - mmiowb(); sky2_tx_check(hw, 0, tx_done[0]); sky2_tx_check(hw, 1, tx_done[1]); @@ -1887,7 +1928,6 @@ exit_loop: netif_rx_complete(dev0); hw->intr_mask |= Y2_IS_STAT_BMU; sky2_write32(hw, B0_IMSK, hw->intr_mask); - mmiowb(); return 0; } else { *budget -= work_done; @@ -1900,35 +1940,42 @@ static void sky2_hw_error(struct sky2_hw *hw, unsigned port, u32 status) { struct net_device *dev = hw->dev[port]; - printk(KERN_INFO PFX "%s: hw error interrupt status 0x%x\n", - dev->name, status); + if (net_ratelimit()) + printk(KERN_INFO PFX "%s: hw error interrupt status 0x%x\n", + dev->name, status); if (status & Y2_IS_PAR_RD1) { - printk(KERN_ERR PFX "%s: ram data read parity error\n", - dev->name); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: ram data read parity error\n", + dev->name); /* Clear IRQ */ sky2_write16(hw, RAM_BUFFER(port, B3_RI_CTRL), RI_CLR_RD_PERR); } if (status & Y2_IS_PAR_WR1) { - printk(KERN_ERR PFX "%s: ram data write parity error\n", - dev->name); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: ram data write parity error\n", + dev->name); sky2_write16(hw, RAM_BUFFER(port, B3_RI_CTRL), RI_CLR_WR_PERR); } if (status & Y2_IS_PAR_MAC1) { - printk(KERN_ERR PFX "%s: MAC parity error\n", dev->name); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: MAC parity error\n", dev->name); sky2_write8(hw, SK_REG(port, TX_GMF_CTRL_T), GMF_CLI_TX_PE); } if (status & Y2_IS_PAR_RX1) { - printk(KERN_ERR PFX "%s: RX parity error\n", dev->name); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: RX parity error\n", dev->name); sky2_write32(hw, Q_ADDR(rxqaddr[port], Q_CSR), BMU_CLR_IRQ_PAR); } if (status & Y2_IS_TCP_TXA1) { - printk(KERN_ERR PFX "%s: TCP segmentation error\n", dev->name); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: TCP segmentation error\n", + dev->name); sky2_write32(hw, Q_ADDR(txqaddr[port], Q_CSR), BMU_CLR_IRQ_TCP); } } @@ -1944,8 +1991,9 @@ static void sky2_hw_intr(struct sky2_hw *hw) u16 pci_err; pci_read_config_word(hw->pdev, PCI_STATUS, &pci_err); - printk(KERN_ERR PFX "%s: pci hw error (0x%x)\n", - pci_name(hw->pdev), pci_err); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: pci hw error (0x%x)\n", + pci_name(hw->pdev), pci_err); sky2_write8(hw, B2_TST_CTRL1, TST_CFG_WRITE_ON); pci_write_config_word(hw->pdev, PCI_STATUS, @@ -1959,8 +2007,9 @@ static void sky2_hw_intr(struct sky2_hw *hw) pci_read_config_dword(hw->pdev, PEX_UNC_ERR_STAT, &pex_err); - printk(KERN_ERR PFX "%s: pci express error (0x%x)\n", - pci_name(hw->pdev), pex_err); + if (net_ratelimit()) + printk(KERN_ERR PFX "%s: pci express error (0x%x)\n", + pci_name(hw->pdev), pex_err); /* clear the interrupt */ sky2_write32(hw, B2_TST_CTRL1, TST_CFG_WRITE_ON); @@ -2250,7 +2299,7 @@ static int sky2_reset(struct sky2_hw *hw) return 0; } -static inline u32 sky2_supported_modes(const struct sky2_hw *hw) +static u32 sky2_supported_modes(const struct sky2_hw *hw) { u32 modes; if (hw->copper) { @@ -2995,7 +3044,7 @@ static __devinit struct net_device *sky2_init_netdev(struct sky2_hw *hw, return dev; } -static inline void sky2_show_addr(struct net_device *dev) +static void __devinit sky2_show_addr(struct net_device *dev) { const struct sky2_port *sky2 = netdev_priv(dev); @@ -3038,13 +3087,17 @@ static int __devinit sky2_probe(struct pci_dev *pdev, goto err_out_free_regions; } - if (sizeof(dma_addr_t) > sizeof(u32)) { - err = pci_set_dma_mask(pdev, DMA_64BIT_MASK); - if (!err) - using_dac = 1; - } + if (sizeof(dma_addr_t) > sizeof(u32) && + !(err = pci_set_dma_mask(pdev, DMA_64BIT_MASK))) { + using_dac = 1; + err = pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK); + if (err < 0) { + printk(KERN_ERR PFX "%s unable to obtain 64 bit DMA " + "for consistent allocations\n", pci_name(pdev)); + goto err_out_free_regions; + } - if (!using_dac) { + } else { err = pci_set_dma_mask(pdev, DMA_32BIT_MASK); if (err) { printk(KERN_ERR PFX "%s no usable DMA configuration\n", @@ -3052,6 +3105,7 @@ static int __devinit sky2_probe(struct pci_dev *pdev, goto err_out_free_regions; } } + #ifdef __BIG_ENDIAN /* byte swap descriptors in hardware */ { @@ -3064,14 +3118,13 @@ static int __devinit sky2_probe(struct pci_dev *pdev, #endif err = -ENOMEM; - hw = kmalloc(sizeof(*hw), GFP_KERNEL); + hw = kzalloc(sizeof(*hw), GFP_KERNEL); if (!hw) { printk(KERN_ERR PFX "%s: cannot allocate hardware struct\n", pci_name(pdev)); goto err_out_free_regions; } - memset(hw, 0, sizeof(*hw)); hw->pdev = pdev; hw->regs = ioremap_nocache(pci_resource_start(pdev, 0), 0x4000); diff --git a/drivers/net/spider_net.c b/drivers/net/spider_net.c index 0d765f1733b..1f5975a61e1 100644 --- a/drivers/net/spider_net.c +++ b/drivers/net/spider_net.c @@ -22,7 +22,6 @@ */ #include <linux/config.h> - #include <linux/compiler.h> #include <linux/crc32.h> #include <linux/delay.h> @@ -30,6 +29,7 @@ #include <linux/ethtool.h> #include <linux/firmware.h> #include <linux/if_vlan.h> +#include <linux/in.h> #include <linux/init.h> #include <linux/ioport.h> #include <linux/ip.h> @@ -43,6 +43,7 @@ #include <linux/slab.h> #include <linux/tcp.h> #include <linux/types.h> +#include <linux/vmalloc.h> #include <linux/wait.h> #include <linux/workqueue.h> #include <asm/bitops.h> @@ -108,42 +109,6 @@ spider_net_write_reg(struct spider_net_card *card, u32 reg, u32 value) writel(value, card->regs + reg); } -/** - * spider_net_write_reg_sync - writes to an SMMIO register of a card - * @card: device structure - * @reg: register to write to - * @value: value to write into the specified SMMIO register - * - * Unlike spider_net_write_reg, this will also make sure the - * data arrives on the card by reading the reg again. - */ -static void -spider_net_write_reg_sync(struct spider_net_card *card, u32 reg, u32 value) -{ - value = cpu_to_le32(value); - writel(value, card->regs + reg); - (void)readl(card->regs + reg); -} - -/** - * spider_net_rx_irq_off - switch off rx irq on this spider card - * @card: device structure - * - * switches off rx irq by masking them out in the GHIINTnMSK register - */ -static void -spider_net_rx_irq_off(struct spider_net_card *card) -{ - u32 regvalue; - unsigned long flags; - - spin_lock_irqsave(&card->intmask_lock, flags); - regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK); - regvalue &= ~SPIDER_NET_RXINT; - spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue); - spin_unlock_irqrestore(&card->intmask_lock, flags); -} - /** spider_net_write_phy - write to phy register * @netdev: adapter to be written to * @mii_id: id of MII @@ -199,60 +164,33 @@ spider_net_read_phy(struct net_device *netdev, int mii_id, int reg) } /** - * spider_net_rx_irq_on - switch on rx irq on this spider card - * @card: device structure - * - * switches on rx irq by enabling them in the GHIINTnMSK register - */ -static void -spider_net_rx_irq_on(struct spider_net_card *card) -{ - u32 regvalue; - unsigned long flags; - - spin_lock_irqsave(&card->intmask_lock, flags); - regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK); - regvalue |= SPIDER_NET_RXINT; - spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue); - spin_unlock_irqrestore(&card->intmask_lock, flags); -} - -/** - * spider_net_tx_irq_off - switch off tx irq on this spider card + * spider_net_rx_irq_off - switch off rx irq on this spider card * @card: device structure * - * switches off tx irq by masking them out in the GHIINTnMSK register + * switches off rx irq by masking them out in the GHIINTnMSK register */ static void -spider_net_tx_irq_off(struct spider_net_card *card) +spider_net_rx_irq_off(struct spider_net_card *card) { u32 regvalue; - unsigned long flags; - spin_lock_irqsave(&card->intmask_lock, flags); - regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK); - regvalue &= ~SPIDER_NET_TXINT; - spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue); - spin_unlock_irqrestore(&card->intmask_lock, flags); + regvalue = SPIDER_NET_INT0_MASK_VALUE & (~SPIDER_NET_RXINT); + spider_net_write_reg(card, SPIDER_NET_GHIINT0MSK, regvalue); } /** - * spider_net_tx_irq_on - switch on tx irq on this spider card + * spider_net_rx_irq_on - switch on rx irq on this spider card * @card: device structure * - * switches on tx irq by enabling them in the GHIINTnMSK register + * switches on rx irq by enabling them in the GHIINTnMSK register */ static void -spider_net_tx_irq_on(struct spider_net_card *card) +spider_net_rx_irq_on(struct spider_net_card *card) { u32 regvalue; - unsigned long flags; - spin_lock_irqsave(&card->intmask_lock, flags); - regvalue = spider_net_read_reg(card, SPIDER_NET_GHIINT0MSK); - regvalue |= SPIDER_NET_TXINT; - spider_net_write_reg_sync(card, SPIDER_NET_GHIINT0MSK, regvalue); - spin_unlock_irqrestore(&card->intmask_lock, flags); + regvalue = SPIDER_NET_INT0_MASK_VALUE | SPIDER_NET_RXINT; + spider_net_write_reg(card, SPIDER_NET_GHIINT0MSK, regvalue); } /** @@ -326,9 +264,8 @@ static enum spider_net_descr_status spider_net_get_descr_status(struct spider_net_descr *descr) { u32 cmd_status; - rmb(); + cmd_status = descr->dmac_cmd_status; - rmb(); cmd_status >>= SPIDER_NET_DESCR_IND_PROC_SHIFT; /* no need to mask out any bits, as cmd_status is 32 bits wide only * (and unsigned) */ @@ -349,7 +286,6 @@ spider_net_set_descr_status(struct spider_net_descr *descr, { u32 cmd_status; /* read the status */ - mb(); cmd_status = descr->dmac_cmd_status; /* clean the upper 4 bits */ cmd_status &= SPIDER_NET_DESCR_IND_PROC_MASKO; @@ -357,7 +293,6 @@ spider_net_set_descr_status(struct spider_net_descr *descr, cmd_status |= ((u32)status)<<SPIDER_NET_DESCR_IND_PROC_SHIFT; /* and write it back */ descr->dmac_cmd_status = cmd_status; - wmb(); } /** @@ -398,8 +333,9 @@ spider_net_init_chain(struct spider_net_card *card, { int i; struct spider_net_descr *descr; + dma_addr_t buf; - spin_lock_init(&card->chain_lock); + atomic_set(&card->rx_chain_refill,0); descr = start_descr; memset(descr, 0, sizeof(*descr) * no); @@ -408,14 +344,14 @@ spider_net_init_chain(struct spider_net_card *card, for (i=0; i<no; i++, descr++) { spider_net_set_descr_status(descr, SPIDER_NET_DESCR_NOT_IN_USE); - descr->bus_addr = - pci_map_single(card->pdev, descr, - SPIDER_NET_DESCR_SIZE, - PCI_DMA_BIDIRECTIONAL); + buf = pci_map_single(card->pdev, descr, + SPIDER_NET_DESCR_SIZE, + PCI_DMA_BIDIRECTIONAL); - if (descr->bus_addr == DMA_ERROR_CODE) + if (buf == DMA_ERROR_CODE) goto iommu_error; + descr->bus_addr = buf; descr->next = descr + 1; descr->prev = descr - 1; @@ -439,7 +375,8 @@ iommu_error: for (i=0; i < no; i++, descr++) if (descr->bus_addr) pci_unmap_single(card->pdev, descr->bus_addr, - SPIDER_NET_DESCR_SIZE, PCI_DMA_BIDIRECTIONAL); + SPIDER_NET_DESCR_SIZE, + PCI_DMA_BIDIRECTIONAL); return -ENOMEM; } @@ -459,7 +396,7 @@ spider_net_free_rx_chain_contents(struct spider_net_card *card) if (descr->skb) { dev_kfree_skb(descr->skb); pci_unmap_single(card->pdev, descr->buf_addr, - SPIDER_NET_MAX_MTU, + SPIDER_NET_MAX_FRAME, PCI_DMA_BIDIRECTIONAL); } descr = descr->next; @@ -480,12 +417,13 @@ static int spider_net_prepare_rx_descr(struct spider_net_card *card, struct spider_net_descr *descr) { + dma_addr_t buf; int error = 0; int offset; int bufsize; /* we need to round up the buffer size to a multiple of 128 */ - bufsize = (SPIDER_NET_MAX_MTU + SPIDER_NET_RXBUF_ALIGN - 1) & + bufsize = (SPIDER_NET_MAX_FRAME + SPIDER_NET_RXBUF_ALIGN - 1) & (~(SPIDER_NET_RXBUF_ALIGN - 1)); /* and we need to have it 128 byte aligned, therefore we allocate a @@ -493,10 +431,8 @@ spider_net_prepare_rx_descr(struct spider_net_card *card, /* allocate an skb */ descr->skb = dev_alloc_skb(bufsize + SPIDER_NET_RXBUF_ALIGN - 1); if (!descr->skb) { - if (net_ratelimit()) - if (netif_msg_rx_err(card)) - pr_err("Not enough memory to allocate " - "rx buffer\n"); + if (netif_msg_rx_err(card) && net_ratelimit()) + pr_err("Not enough memory to allocate rx buffer\n"); return -ENOMEM; } descr->buf_size = bufsize; @@ -510,12 +446,12 @@ spider_net_prepare_rx_descr(struct spider_net_card *card, if (offset) skb_reserve(descr->skb, SPIDER_NET_RXBUF_ALIGN - offset); /* io-mmu-map the skb */ - descr->buf_addr = pci_map_single(card->pdev, descr->skb->data, - SPIDER_NET_MAX_MTU, - PCI_DMA_BIDIRECTIONAL); - if (descr->buf_addr == DMA_ERROR_CODE) { + buf = pci_map_single(card->pdev, descr->skb->data, + SPIDER_NET_MAX_FRAME, PCI_DMA_BIDIRECTIONAL); + descr->buf_addr = buf; + if (buf == DMA_ERROR_CODE) { dev_kfree_skb_any(descr->skb); - if (netif_msg_rx_err(card)) + if (netif_msg_rx_err(card) && net_ratelimit()) pr_err("Could not iommu-map rx buffer\n"); spider_net_set_descr_status(descr, SPIDER_NET_DESCR_NOT_IN_USE); } else { @@ -526,10 +462,10 @@ spider_net_prepare_rx_descr(struct spider_net_card *card, } /** - * spider_net_enable_rxctails - sets RX dmac chain tail addresses + * spider_net_enable_rxchtails - sets RX dmac chain tail addresses * @card: card structure * - * spider_net_enable_rxctails sets the RX DMAC chain tail adresses in the + * spider_net_enable_rxchtails sets the RX DMAC chain tail adresses in the * chip by writing to the appropriate register. DMA is enabled in * spider_net_enable_rxdmac. */ @@ -551,6 +487,7 @@ spider_net_enable_rxchtails(struct spider_net_card *card) static void spider_net_enable_rxdmac(struct spider_net_card *card) { + wmb(); spider_net_write_reg(card, SPIDER_NET_GDADMACCNTR, SPIDER_NET_DMA_RX_VALUE); } @@ -559,32 +496,28 @@ spider_net_enable_rxdmac(struct spider_net_card *card) * spider_net_refill_rx_chain - refills descriptors/skbs in the rx chains * @card: card structure * - * refills descriptors in all chains (last used chain first): allocates skbs - * and iommu-maps them. + * refills descriptors in the rx chain: allocates skbs and iommu-maps them. */ static void spider_net_refill_rx_chain(struct spider_net_card *card) { struct spider_net_descr_chain *chain; - int count = 0; - unsigned long flags; chain = &card->rx_chain; - spin_lock_irqsave(&card->chain_lock, flags); - while (spider_net_get_descr_status(chain->head) == - SPIDER_NET_DESCR_NOT_IN_USE) { - if (spider_net_prepare_rx_descr(card, chain->head)) - break; - count++; - chain->head = chain->head->next; - } - spin_unlock_irqrestore(&card->chain_lock, flags); + /* one context doing the refill (and a second context seeing that + * and omitting it) is ok. If called by NAPI, we'll be called again + * as spider_net_decode_one_descr is called several times. If some + * interrupt calls us, the NAPI is about to clean up anyway. */ + if (atomic_inc_return(&card->rx_chain_refill) == 1) + while (spider_net_get_descr_status(chain->head) == + SPIDER_NET_DESCR_NOT_IN_USE) { + if (spider_net_prepare_rx_descr(card, chain->head)) + break; + chain->head = chain->head->next; + } - /* could be optimized, only do that, if we know the DMA processing - * has terminated */ - if (count) - spider_net_enable_rxdmac(card); + atomic_dec(&card->rx_chain_refill); } /** @@ -613,6 +546,7 @@ spider_net_alloc_rx_skbs(struct spider_net_card *card) /* this will allocate the rest of the rx buffers; if not, it's * business as usual later on */ spider_net_refill_rx_chain(card); + spider_net_enable_rxdmac(card); return 0; error: @@ -649,24 +583,30 @@ spider_net_release_tx_descr(struct spider_net_card *card, * @card: adapter structure * @brutal: if set, don't care about whether descriptor seems to be in use * - * releases the tx descriptors that spider has finished with (if non-brutal) - * or simply release tx descriptors (if brutal) + * returns 0 if the tx ring is empty, otherwise 1. + * + * spider_net_release_tx_chain releases the tx descriptors that spider has + * finished with (if non-brutal) or simply release tx descriptors (if brutal). + * If some other context is calling this function, we return 1 so that we're + * scheduled again (if we were scheduled) and will not loose initiative. */ -static void +static int spider_net_release_tx_chain(struct spider_net_card *card, int brutal) { struct spider_net_descr_chain *tx_chain = &card->tx_chain; enum spider_net_descr_status status; - spider_net_tx_irq_off(card); + if (atomic_inc_return(&card->tx_chain_release) != 1) { + atomic_dec(&card->tx_chain_release); + return 1; + } - /* no lock for chain needed, if this is only executed once at a time */ -again: for (;;) { status = spider_net_get_descr_status(tx_chain->tail); switch (status) { case SPIDER_NET_DESCR_CARDOWNED: - if (!brutal) goto out; + if (!brutal) + goto out; /* fallthrough, if we release the descriptors * brutally (then we don't care about * SPIDER_NET_DESCR_CARDOWNED) */ @@ -693,25 +633,30 @@ again: tx_chain->tail = tx_chain->tail->next; } out: + atomic_dec(&card->tx_chain_release); + netif_wake_queue(card->netdev); - if (!brutal) { - /* switch on tx irqs (while we are still in the interrupt - * handler, so we don't get an interrupt), check again - * for done descriptors. This results in fewer interrupts */ - spider_net_tx_irq_on(card); - status = spider_net_get_descr_status(tx_chain->tail); - switch (status) { - case SPIDER_NET_DESCR_RESPONSE_ERROR: - case SPIDER_NET_DESCR_PROTECTION_ERROR: - case SPIDER_NET_DESCR_FORCE_END: - case SPIDER_NET_DESCR_COMPLETE: - goto again; - default: - break; - } - } + if (status == SPIDER_NET_DESCR_CARDOWNED) + return 1; + return 0; +} +/** + * spider_net_cleanup_tx_ring - cleans up the TX ring + * @card: card structure + * + * spider_net_cleanup_tx_ring is called by the tx_timer (as we don't use + * interrupts to cleanup our TX ring) and returns sent packets to the stack + * by freeing them + */ +static void +spider_net_cleanup_tx_ring(struct spider_net_card *card) +{ + if ( (spider_net_release_tx_chain(card, 0)) && + (card->netdev->flags & IFF_UP) ) { + mod_timer(&card->tx_timer, jiffies + SPIDER_NET_TX_TIMER); + } } /** @@ -726,16 +671,22 @@ out: static u8 spider_net_get_multicast_hash(struct net_device *netdev, __u8 *addr) { - /* FIXME: an addr of 01:00:5e:00:00:01 must result in 0xa9, - * ff:ff:ff:ff:ff:ff must result in 0xfd */ u32 crc; u8 hash; + char addr_for_crc[ETH_ALEN] = { 0, }; + int i, bit; - crc = crc32_be(~0, addr, netdev->addr_len); + for (i = 0; i < ETH_ALEN * 8; i++) { + bit = (addr[i / 8] >> (i % 8)) & 1; + addr_for_crc[ETH_ALEN - 1 - i / 8] += bit << (7 - (i % 8)); + } + + crc = crc32_be(~0, addr_for_crc, netdev->addr_len); hash = (crc >> 27); hash <<= 3; hash |= crc & 7; + hash &= 0xff; return hash; } @@ -821,9 +772,11 @@ spider_net_stop(struct net_device *netdev) { struct spider_net_card *card = netdev_priv(netdev); + tasklet_kill(&card->rxram_full_tl); netif_poll_disable(netdev); netif_carrier_off(netdev); netif_stop_queue(netdev); + del_timer_sync(&card->tx_timer); /* disable/mask all interrupts */ spider_net_write_reg(card, SPIDER_NET_GHIINT0MSK, 0); @@ -872,13 +825,15 @@ spider_net_get_next_tx_descr(struct spider_net_card *card) * @skb: packet to consider * * fills out the command and status field of the descriptor structure, - * depending on hardware checksum settings. This function assumes a wmb() - * has executed before. + * depending on hardware checksum settings. */ static void spider_net_set_txdescr_cmdstat(struct spider_net_descr *descr, struct sk_buff *skb) { + /* make sure the other fields in the descriptor are written */ + wmb(); + if (skb->ip_summed != CHECKSUM_HW) { descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_NOCS; return; @@ -887,14 +842,13 @@ spider_net_set_txdescr_cmdstat(struct spider_net_descr *descr, /* is packet ip? * if yes: tcp? udp? */ if (skb->protocol == htons(ETH_P_IP)) { - if (skb->nh.iph->protocol == IPPROTO_TCP) { + if (skb->nh.iph->protocol == IPPROTO_TCP) descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_TCPCS; - } else if (skb->nh.iph->protocol == IPPROTO_UDP) { + else if (skb->nh.iph->protocol == IPPROTO_UDP) descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_UDPCS; - } else { /* the stack should checksum non-tcp and non-udp - packets on his own: NETIF_F_IP_CSUM */ + else /* the stack should checksum non-tcp and non-udp + packets on his own: NETIF_F_IP_CSUM */ descr->dmac_cmd_status = SPIDER_NET_DMAC_CMDSTAT_NOCS; - } } } @@ -914,23 +868,22 @@ spider_net_prepare_tx_descr(struct spider_net_card *card, struct spider_net_descr *descr, struct sk_buff *skb) { - descr->buf_addr = pci_map_single(card->pdev, skb->data, - skb->len, PCI_DMA_BIDIRECTIONAL); - if (descr->buf_addr == DMA_ERROR_CODE) { - if (netif_msg_tx_err(card)) + dma_addr_t buf; + + buf = pci_map_single(card->pdev, skb->data, + skb->len, PCI_DMA_BIDIRECTIONAL); + if (buf == DMA_ERROR_CODE) { + if (netif_msg_tx_err(card) && net_ratelimit()) pr_err("could not iommu-map packet (%p, %i). " "Dropping packet\n", skb->data, skb->len); return -ENOMEM; } + descr->buf_addr = buf; descr->buf_size = skb->len; descr->skb = skb; descr->data_status = 0; - /* make sure the above values are in memory before we change the - * status */ - wmb(); - spider_net_set_txdescr_cmdstat(descr,skb); return 0; @@ -972,17 +925,12 @@ spider_net_xmit(struct sk_buff *skb, struct net_device *netdev) struct spider_net_descr *descr; int result; - descr = spider_net_get_next_tx_descr(card); + spider_net_release_tx_chain(card, 0); - if (!descr) { - netif_stop_queue(netdev); + descr = spider_net_get_next_tx_descr(card); - descr = spider_net_get_next_tx_descr(card); - if (!descr) - goto error; - else - netif_start_queue(netdev); - } + if (!descr) + goto error; result = spider_net_prepare_tx_descr(card, descr, skb); if (result) @@ -990,19 +938,25 @@ spider_net_xmit(struct sk_buff *skb, struct net_device *netdev) card->tx_chain.head = card->tx_chain.head->next; - /* make sure the status from spider_net_prepare_tx_descr is in - * memory before we check out the previous descriptor */ - wmb(); - if (spider_net_get_descr_status(descr->prev) != - SPIDER_NET_DESCR_CARDOWNED) - spider_net_kick_tx_dma(card, descr); + SPIDER_NET_DESCR_CARDOWNED) { + /* make sure the current descriptor is in memory. Then + * kicking it on again makes sense, if the previous is not + * card-owned anymore. Check the previous descriptor twice + * to omit an mb() in heavy traffic cases */ + mb(); + if (spider_net_get_descr_status(descr->prev) != + SPIDER_NET_DESCR_CARDOWNED) + spider_net_kick_tx_dma(card, descr); + } + + mod_timer(&card->tx_timer, jiffies + SPIDER_NET_TX_TIMER); return NETDEV_TX_OK; error: card->netdev_stats.tx_dropped++; - return NETDEV_TX_LOCKED; + return NETDEV_TX_BUSY; } /** @@ -1027,6 +981,7 @@ spider_net_do_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) * spider_net_pass_skb_up - takes an skb from a descriptor and passes it on * @descr: descriptor to process * @card: card structure + * @napi: whether caller is in NAPI context * * returns 1 on success, 0 if no packet was passed to the stack * @@ -1035,7 +990,7 @@ spider_net_do_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd) */ static int spider_net_pass_skb_up(struct spider_net_descr *descr, - struct spider_net_card *card) + struct spider_net_card *card, int napi) { struct sk_buff *skb; struct net_device *netdev; @@ -1046,22 +1001,20 @@ spider_net_pass_skb_up(struct spider_net_descr *descr, netdev = card->netdev; - /* check for errors in the data_error flag */ - if ((data_error & SPIDER_NET_DATA_ERROR_MASK) && - netif_msg_rx_err(card)) - pr_err("error in received descriptor found, " - "data_status=x%08x, data_error=x%08x\n", - data_status, data_error); - - /* prepare skb, unmap descriptor */ - skb = descr->skb; - pci_unmap_single(card->pdev, descr->buf_addr, SPIDER_NET_MAX_MTU, + /* unmap descriptor */ + pci_unmap_single(card->pdev, descr->buf_addr, SPIDER_NET_MAX_FRAME, PCI_DMA_BIDIRECTIONAL); /* the cases we'll throw away the packet immediately */ - if (data_error & SPIDER_NET_DESTROY_RX_FLAGS) + if (data_error & SPIDER_NET_DESTROY_RX_FLAGS) { + if (netif_msg_rx_err(card)) + pr_err("error in received descriptor found, " + "data_status=x%08x, data_error=x%08x\n", + data_status, data_error); return 0; + } + skb = descr->skb; skb->dev = netdev; skb_put(skb, descr->valid_size); @@ -1073,14 +1026,14 @@ spider_net_pass_skb_up(struct spider_net_descr *descr, /* checksum offload */ if (card->options.rx_csum) { - if ( (data_status & SPIDER_NET_DATA_STATUS_CHK_MASK) && - (!(data_error & SPIDER_NET_DATA_ERROR_CHK_MASK)) ) + if ( ( (data_status & SPIDER_NET_DATA_STATUS_CKSUM_MASK) == + SPIDER_NET_DATA_STATUS_CKSUM_MASK) && + !(data_error & SPIDER_NET_DATA_ERR_CKSUM_MASK)) skb->ip_summed = CHECKSUM_UNNECESSARY; else skb->ip_summed = CHECKSUM_NONE; - } else { + } else skb->ip_summed = CHECKSUM_NONE; - } if (data_status & SPIDER_NET_VLAN_PACKET) { /* further enhancements: HW-accel VLAN @@ -1089,7 +1042,10 @@ spider_net_pass_skb_up(struct spider_net_descr *descr, } /* pass skb up to stack */ - netif_receive_skb(skb); + if (napi) + netif_receive_skb(skb); + else + netif_rx_ni(skb); /* update netdevice statistics */ card->netdev_stats.rx_packets++; @@ -1099,16 +1055,18 @@ spider_net_pass_skb_up(struct spider_net_descr *descr, } /** - * spider_net_decode_descr - processes an rx descriptor + * spider_net_decode_one_descr - processes an rx descriptor * @card: card structure + * @napi: whether caller is in NAPI context * * returns 1 if a packet has been sent to the stack, otherwise 0 * * processes an rx descriptor by iommu-unmapping the data buffer and passing - * the packet up to the stack + * the packet up to the stack. This function is called in softirq + * context, e.g. either bottom half from interrupt or NAPI polling context */ static int -spider_net_decode_one_descr(struct spider_net_card *card) +spider_net_decode_one_descr(struct spider_net_card *card, int napi) { enum spider_net_descr_status status; struct spider_net_descr *descr; @@ -1122,17 +1080,19 @@ spider_net_decode_one_descr(struct spider_net_card *card) if (status == SPIDER_NET_DESCR_CARDOWNED) { /* nothing in the descriptor yet */ - return 0; + result=0; + goto out; } if (status == SPIDER_NET_DESCR_NOT_IN_USE) { - /* not initialized yet, I bet chain->tail == chain->head - * and the ring is empty */ + /* not initialized yet, the ring must be empty */ spider_net_refill_rx_chain(card); - return 0; + spider_net_enable_rxdmac(card); + result=0; + goto out; } - /* descriptor definitively used -- move on head */ + /* descriptor definitively used -- move on tail */ chain->tail = descr->next; result = 0; @@ -1143,6 +1103,9 @@ spider_net_decode_one_descr(struct spider_net_card *card) pr_err("%s: dropping RX descriptor with state %d\n", card->netdev->name, status); card->netdev_stats.rx_dropped++; + pci_unmap_single(card->pdev, descr->buf_addr, + SPIDER_NET_MAX_FRAME, PCI_DMA_BIDIRECTIONAL); + dev_kfree_skb_irq(descr->skb); goto refill; } @@ -1155,12 +1118,13 @@ spider_net_decode_one_descr(struct spider_net_card *card) } /* ok, we've got a packet in descr */ - result = spider_net_pass_skb_up(descr, card); + result = spider_net_pass_skb_up(descr, card, napi); refill: spider_net_set_descr_status(descr, SPIDER_NET_DESCR_NOT_IN_USE); /* change the descriptor state: */ - spider_net_refill_rx_chain(card); - + if (!napi) + spider_net_refill_rx_chain(card); +out: return result; } @@ -1186,7 +1150,7 @@ spider_net_poll(struct net_device *netdev, int *budget) packets_to_do = min(*budget, netdev->quota); while (packets_to_do) { - if (spider_net_decode_one_descr(card)) { + if (spider_net_decode_one_descr(card, 1)) { packets_done++; packets_to_do--; } else { @@ -1198,6 +1162,7 @@ spider_net_poll(struct net_device *netdev, int *budget) netdev->quota -= packets_done; *budget -= packets_done; + spider_net_refill_rx_chain(card); /* if all packets are in the stack, enable interrupts and return 0 */ /* if not, return 1 */ @@ -1342,6 +1307,24 @@ spider_net_enable_txdmac(struct spider_net_card *card) } /** + * spider_net_handle_rxram_full - cleans up RX ring upon RX RAM full interrupt + * @card: card structure + * + * spider_net_handle_rxram_full empties the RX ring so that spider can put + * more packets in it and empty its RX RAM. This is called in bottom half + * context + */ +static void +spider_net_handle_rxram_full(struct spider_net_card *card) +{ + while (spider_net_decode_one_descr(card, 0)) + ; + spider_net_enable_rxchtails(card); + spider_net_enable_rxdmac(card); + netif_rx_schedule(card->netdev); +} + +/** * spider_net_handle_error_irq - handles errors raised by an interrupt * @card: card structure * @status_reg: interrupt status register 0 (GHIINT0STS) @@ -1449,17 +1432,21 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) switch (i) { case SPIDER_NET_GTMFLLINT: - if (netif_msg_intr(card)) + if (netif_msg_intr(card) && net_ratelimit()) pr_err("Spider TX RAM full\n"); show_error = 0; break; + case SPIDER_NET_GRFDFLLINT: /* fallthrough */ + case SPIDER_NET_GRFCFLLINT: /* fallthrough */ + case SPIDER_NET_GRFBFLLINT: /* fallthrough */ + case SPIDER_NET_GRFAFLLINT: /* fallthrough */ case SPIDER_NET_GRMFLLINT: - if (netif_msg_intr(card)) + if (netif_msg_intr(card) && net_ratelimit()) pr_err("Spider RX RAM full, incoming packets " - "might be discarded !\n"); - netif_rx_schedule(card->netdev); - spider_net_enable_rxchtails(card); - spider_net_enable_rxdmac(card); + "might be discarded!\n"); + spider_net_rx_irq_off(card); + tasklet_schedule(&card->rxram_full_tl); + show_error = 0; break; /* case SPIDER_NET_GTMSHTINT: problem, print a message */ @@ -1467,10 +1454,6 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) /* allrighty. tx from previous descr ok */ show_error = 0; break; - /* case SPIDER_NET_GRFDFLLINT: print a message down there */ - /* case SPIDER_NET_GRFCFLLINT: print a message down there */ - /* case SPIDER_NET_GRFBFLLINT: print a message down there */ - /* case SPIDER_NET_GRFAFLLINT: print a message down there */ /* chain end */ case SPIDER_NET_GDDDCEINT: /* fallthrough */ @@ -1482,6 +1465,7 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) "restarting DMAC %c.\n", 'D'+i-SPIDER_NET_GDDDCEINT); spider_net_refill_rx_chain(card); + spider_net_enable_rxdmac(card); show_error = 0; break; @@ -1492,6 +1476,7 @@ spider_net_handle_error_irq(struct spider_net_card *card, u32 status_reg) case SPIDER_NET_GDAINVDINT: /* could happen when rx chain is full */ spider_net_refill_rx_chain(card); + spider_net_enable_rxdmac(card); show_error = 0; break; @@ -1580,17 +1565,13 @@ spider_net_interrupt(int irq, void *ptr, struct pt_regs *regs) if (!status_reg) return IRQ_NONE; - if (status_reg & SPIDER_NET_TXINT) - spider_net_release_tx_chain(card, 0); - if (status_reg & SPIDER_NET_RXINT ) { spider_net_rx_irq_off(card); netif_rx_schedule(netdev); } - /* we do this after rx and tx processing, as we want the tx chain - * processed to see, whether we should restart tx dma processing */ - spider_net_handle_error_irq(card, status_reg); + if (status_reg & SPIDER_NET_ERRINT ) + spider_net_handle_error_irq(card, status_reg); /* clear interrupt sources */ spider_net_write_reg(card, SPIDER_NET_GHIINT0STS, status_reg); @@ -1831,34 +1812,40 @@ spider_net_setup_phy(struct spider_net_card *card) /** * spider_net_download_firmware - loads firmware into the adapter * @card: card structure - * @firmware: firmware pointer + * @firmware_ptr: pointer to firmware data * - * spider_net_download_firmware loads the firmware opened by - * spider_net_init_firmware into the adapter. + * spider_net_download_firmware loads the firmware data into the + * adapter. It assumes the length etc. to be allright. */ -static void +static int spider_net_download_firmware(struct spider_net_card *card, - const struct firmware *firmware) + u8 *firmware_ptr) { int sequencer, i; - u32 *fw_ptr = (u32 *)firmware->data; + u32 *fw_ptr = (u32 *)firmware_ptr; /* stop sequencers */ spider_net_write_reg(card, SPIDER_NET_GSINIT, SPIDER_NET_STOP_SEQ_VALUE); - for (sequencer = 0; sequencer < 6; sequencer++) { + for (sequencer = 0; sequencer < SPIDER_NET_FIRMWARE_SEQS; + sequencer++) { spider_net_write_reg(card, SPIDER_NET_GSnPRGADR + sequencer * 8, 0); - for (i = 0; i < SPIDER_NET_FIRMWARE_LEN; i++) { + for (i = 0; i < SPIDER_NET_FIRMWARE_SEQWORDS; i++) { spider_net_write_reg(card, SPIDER_NET_GSnPRGDAT + sequencer * 8, *fw_ptr); fw_ptr++; } } + if (spider_net_read_reg(card, SPIDER_NET_GSINIT)) + return -EIO; + spider_net_write_reg(card, SPIDER_NET_GSINIT, SPIDER_NET_RUN_SEQ_VALUE); + + return 0; } /** @@ -1890,31 +1877,53 @@ spider_net_download_firmware(struct spider_net_card *card, static int spider_net_init_firmware(struct spider_net_card *card) { - const struct firmware *firmware; - int err = -EIO; + struct firmware *firmware = NULL; + struct device_node *dn; + u8 *fw_prop = NULL; + int err = -ENOENT; + int fw_size; + + if (request_firmware((const struct firmware **)&firmware, + SPIDER_NET_FIRMWARE_NAME, &card->pdev->dev) == 0) { + if ( (firmware->size != SPIDER_NET_FIRMWARE_LEN) && + netif_msg_probe(card) ) { + pr_err("Incorrect size of spidernet firmware in " \ + "filesystem. Looking in host firmware...\n"); + goto try_host_fw; + } + err = spider_net_download_firmware(card, firmware->data); - if (request_firmware(&firmware, - SPIDER_NET_FIRMWARE_NAME, &card->pdev->dev) < 0) { - if (netif_msg_probe(card)) - pr_err("Couldn't read in sequencer data file %s.\n", - SPIDER_NET_FIRMWARE_NAME); - firmware = NULL; - goto out; - } + release_firmware(firmware); + if (err) + goto try_host_fw; - if (firmware->size != 6 * SPIDER_NET_FIRMWARE_LEN * sizeof(u32)) { - if (netif_msg_probe(card)) - pr_err("Invalid size of sequencer data file %s.\n", - SPIDER_NET_FIRMWARE_NAME); - goto out; + goto done; } - spider_net_download_firmware(card, firmware); +try_host_fw: + dn = pci_device_to_OF_node(card->pdev); + if (!dn) + goto out_err; - err = 0; -out: - release_firmware(firmware); + fw_prop = (u8 *)get_property(dn, "firmware", &fw_size); + if (!fw_prop) + goto out_err; + + if ( (fw_size != SPIDER_NET_FIRMWARE_LEN) && + netif_msg_probe(card) ) { + pr_err("Incorrect size of spidernet firmware in " \ + "host firmware\n"); + goto done; + } + err = spider_net_download_firmware(card, fw_prop); + +done: + return err; +out_err: + if (netif_msg_probe(card)) + pr_err("Couldn't find spidernet firmware in filesystem " \ + "or host firmware\n"); return err; } @@ -1934,10 +1943,11 @@ spider_net_workaround_rxramfull(struct spider_net_card *card) SPIDER_NET_CKRCTRL_RUN_VALUE); /* empty sequencer data */ - for (sequencer = 0; sequencer < 6; sequencer++) { + for (sequencer = 0; sequencer < SPIDER_NET_FIRMWARE_SEQS; + sequencer++) { spider_net_write_reg(card, SPIDER_NET_GSnPRGDAT + sequencer * 8, 0x0); - for (i = 0; i < SPIDER_NET_FIRMWARE_LEN; i++) { + for (i = 0; i < SPIDER_NET_FIRMWARE_SEQWORDS; i++) { spider_net_write_reg(card, SPIDER_NET_GSnPRGDAT + sequencer * 8, 0x0); } @@ -2061,7 +2071,15 @@ spider_net_setup_netdev(struct spider_net_card *card) SET_NETDEV_DEV(netdev, &card->pdev->dev); pci_set_drvdata(card->pdev, netdev); - spin_lock_init(&card->intmask_lock); + + atomic_set(&card->tx_chain_release,0); + card->rxram_full_tl.data = (unsigned long) card; + card->rxram_full_tl.func = + (void (*)(unsigned long)) spider_net_handle_rxram_full; + init_timer(&card->tx_timer); + card->tx_timer.function = + (void (*)(unsigned long)) spider_net_cleanup_tx_ring; + card->tx_timer.data = (unsigned long) card; netdev->irq = card->pdev->irq; card->options.rx_csum = SPIDER_NET_RX_CSUM_DEFAULT; diff --git a/drivers/net/spider_net.h b/drivers/net/spider_net.h index 22b2f234735..5922b529a04 100644 --- a/drivers/net/spider_net.h +++ b/drivers/net/spider_net.h @@ -33,25 +33,32 @@ extern struct ethtool_ops spider_net_ethtool_ops; extern char spider_net_driver_name[]; -#define SPIDER_NET_MAX_MTU 2308 +#define SPIDER_NET_MAX_FRAME 2312 +#define SPIDER_NET_MAX_MTU 2294 #define SPIDER_NET_MIN_MTU 64 #define SPIDER_NET_RXBUF_ALIGN 128 -#define SPIDER_NET_RX_DESCRIPTORS_DEFAULT 64 +#define SPIDER_NET_RX_DESCRIPTORS_DEFAULT 256 #define SPIDER_NET_RX_DESCRIPTORS_MIN 16 -#define SPIDER_NET_RX_DESCRIPTORS_MAX 256 +#define SPIDER_NET_RX_DESCRIPTORS_MAX 512 -#define SPIDER_NET_TX_DESCRIPTORS_DEFAULT 64 +#define SPIDER_NET_TX_DESCRIPTORS_DEFAULT 256 #define SPIDER_NET_TX_DESCRIPTORS_MIN 16 -#define SPIDER_NET_TX_DESCRIPTORS_MAX 256 +#define SPIDER_NET_TX_DESCRIPTORS_MAX 512 + +#define SPIDER_NET_TX_TIMER 20 #define SPIDER_NET_RX_CSUM_DEFAULT 1 -#define SPIDER_NET_WATCHDOG_TIMEOUT 5*HZ -#define SPIDER_NET_NAPI_WEIGHT 64 +#define SPIDER_NET_WATCHDOG_TIMEOUT 50*HZ +#define SPIDER_NET_NAPI_WEIGHT 64 -#define SPIDER_NET_FIRMWARE_LEN 1024 +#define SPIDER_NET_FIRMWARE_SEQS 6 +#define SPIDER_NET_FIRMWARE_SEQWORDS 1024 +#define SPIDER_NET_FIRMWARE_LEN (SPIDER_NET_FIRMWARE_SEQS * \ + SPIDER_NET_FIRMWARE_SEQWORDS * \ + sizeof(u32)) #define SPIDER_NET_FIRMWARE_NAME "spider_fw.bin" /** spider_net SMMIO registers */ @@ -142,14 +149,12 @@ extern char spider_net_driver_name[]; /** SCONFIG registers */ #define SPIDER_NET_SCONFIG_IOACTE 0x00002810 -/** hardcoded register values */ -#define SPIDER_NET_INT0_MASK_VALUE 0x3f7fe3ff -#define SPIDER_NET_INT1_MASK_VALUE 0xffffffff +/** interrupt mask registers */ +#define SPIDER_NET_INT0_MASK_VALUE 0x3f7fe2c7 +#define SPIDER_NET_INT1_MASK_VALUE 0xffff7ff7 /* no MAC aborts -> auto retransmission */ -#define SPIDER_NET_INT2_MASK_VALUE 0xfffffff1 +#define SPIDER_NET_INT2_MASK_VALUE 0xffef7ff1 -/* clear counter when interrupt sources are cleared -#define SPIDER_NET_FRAMENUM_VALUE 0x0001f001 */ /* we rely on flagged descriptor interrupts */ #define SPIDER_NET_FRAMENUM_VALUE 0x00000000 /* set this first, then the FRAMENUM_VALUE */ @@ -168,7 +173,7 @@ extern char spider_net_driver_name[]; #if 0 #define SPIDER_NET_WOL_VALUE 0x00000000 #endif -#define SPIDER_NET_IPSECINIT_VALUE 0x00f000f8 +#define SPIDER_NET_IPSECINIT_VALUE 0x6f716f71 /* pause frames: automatic, no upper retransmission count */ /* outside loopback mode: ETOMOD signal dont matter, not connected */ @@ -318,6 +323,10 @@ enum spider_net_int2_status { #define SPIDER_NET_RXINT ( (1 << SPIDER_NET_GDAFDCINT) | \ (1 << SPIDER_NET_GRMFLLINT) ) +#define SPIDER_NET_ERRINT ( 0xffffffff & \ + (~SPIDER_NET_TXINT) & \ + (~SPIDER_NET_RXINT) ) + #define SPIDER_NET_GPREXEC 0x80000000 #define SPIDER_NET_GPRDAT_MASK 0x0000ffff @@ -358,9 +367,6 @@ enum spider_net_int2_status { /* descr ready, descr is in middle of chain, get interrupt on completion */ #define SPIDER_NET_DMAC_RX_CARDOWNED 0xa0800000 -/* multicast is no problem */ -#define SPIDER_NET_DATA_ERROR_MASK 0xffffbfff - enum spider_net_descr_status { SPIDER_NET_DESCR_COMPLETE = 0x00, /* used in rx and tx */ SPIDER_NET_DESCR_RESPONSE_ERROR = 0x01, /* used in rx and tx */ @@ -373,9 +379,9 @@ enum spider_net_descr_status { struct spider_net_descr { /* as defined by the hardware */ - dma_addr_t buf_addr; + u32 buf_addr; u32 buf_size; - dma_addr_t next_descr_addr; + u32 next_descr_addr; u32 dmac_cmd_status; u32 result_size; u32 valid_size; /* all zeroes for tx */ @@ -384,7 +390,7 @@ struct spider_net_descr { /* used in the driver */ struct sk_buff *skb; - dma_addr_t bus_addr; + u32 bus_addr; struct spider_net_descr *next; struct spider_net_descr *prev; } __attribute__((aligned(32))); @@ -396,21 +402,21 @@ struct spider_net_descr_chain { }; /* descriptor data_status bits */ -#define SPIDER_NET_RXIPCHK 29 -#define SPIDER_NET_TCPUDPIPCHK 28 -#define SPIDER_NET_DATA_STATUS_CHK_MASK (1 << SPIDER_NET_RXIPCHK | \ - 1 << SPIDER_NET_TCPUDPIPCHK) - +#define SPIDER_NET_RX_IPCHK 29 +#define SPIDER_NET_RX_TCPCHK 28 #define SPIDER_NET_VLAN_PACKET 21 +#define SPIDER_NET_DATA_STATUS_CKSUM_MASK ( (1 << SPIDER_NET_RX_IPCHK) | \ + (1 << SPIDER_NET_RX_TCPCHK) ) /* descriptor data_error bits */ -#define SPIDER_NET_RXIPCHKERR 27 -#define SPIDER_NET_RXTCPCHKERR 26 -#define SPIDER_NET_DATA_ERROR_CHK_MASK (1 << SPIDER_NET_RXIPCHKERR | \ - 1 << SPIDER_NET_RXTCPCHKERR) +#define SPIDER_NET_RX_IPCHKERR 27 +#define SPIDER_NET_RX_RXTCPCHKERR 28 + +#define SPIDER_NET_DATA_ERR_CKSUM_MASK (1 << SPIDER_NET_RX_IPCHKERR) -/* the cases we don't pass the packet to the stack */ -#define SPIDER_NET_DESTROY_RX_FLAGS 0x70138000 +/* the cases we don't pass the packet to the stack. + * 701b8000 would be correct, but every packets gets that flag */ +#define SPIDER_NET_DESTROY_RX_FLAGS 0x700b8000 #define SPIDER_NET_DESCR_SIZE 32 @@ -445,13 +451,16 @@ struct spider_net_card { struct spider_net_descr_chain tx_chain; struct spider_net_descr_chain rx_chain; - spinlock_t chain_lock; + atomic_t rx_chain_refill; + atomic_t tx_chain_release; struct net_device_stats netdev_stats; struct spider_net_options options; spinlock_t intmask_lock; + struct tasklet_struct rxram_full_tl; + struct timer_list tx_timer; struct work_struct tx_timeout_task; atomic_t tx_timeout_task_counter; diff --git a/drivers/net/spider_net_ethtool.c b/drivers/net/spider_net_ethtool.c index d42e60ba74c..a5bb0b7633a 100644 --- a/drivers/net/spider_net_ethtool.c +++ b/drivers/net/spider_net_ethtool.c @@ -113,6 +113,23 @@ spider_net_ethtool_set_rx_csum(struct net_device *netdev, u32 n) return 0; } +static uint32_t +spider_net_ethtool_get_tx_csum(struct net_device *netdev) +{ + return (netdev->features & NETIF_F_HW_CSUM) != 0; +} + +static int +spider_net_ethtool_set_tx_csum(struct net_device *netdev, uint32_t data) +{ + if (data) + netdev->features |= NETIF_F_HW_CSUM; + else + netdev->features &= ~NETIF_F_HW_CSUM; + + return 0; +} + struct ethtool_ops spider_net_ethtool_ops = { .get_settings = spider_net_ethtool_get_settings, .get_drvinfo = spider_net_ethtool_get_drvinfo, @@ -122,5 +139,7 @@ struct ethtool_ops spider_net_ethtool_ops = { .nway_reset = spider_net_ethtool_nway_reset, .get_rx_csum = spider_net_ethtool_get_rx_csum, .set_rx_csum = spider_net_ethtool_set_rx_csum, + .get_tx_csum = spider_net_ethtool_get_tx_csum, + .set_tx_csum = spider_net_ethtool_set_tx_csum, }; diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c index eb86b059809..f2d1dafde08 100644 --- a/drivers/net/tg3.c +++ b/drivers/net/tg3.c @@ -69,8 +69,8 @@ #define DRV_MODULE_NAME "tg3" #define PFX DRV_MODULE_NAME ": " -#define DRV_MODULE_VERSION "3.47" -#define DRV_MODULE_RELDATE "Dec 28, 2005" +#define DRV_MODULE_VERSION "3.48" +#define DRV_MODULE_RELDATE "Jan 16, 2006" #define TG3_DEF_MAC_MODE 0 #define TG3_DEF_RX_MODE 0 @@ -1325,10 +1325,12 @@ static int tg3_set_power_state(struct tg3 *tp, int state) val &= ~((1 << 16) | (1 << 4) | (1 << 2) | (1 << 1) | 1); tw32(0x7d00, val); if (!(tp->tg3_flags & TG3_FLAG_ENABLE_ASF)) { - tg3_nvram_lock(tp); + int err; + + err = tg3_nvram_lock(tp); tg3_halt_cpu(tp, RX_CPU_BASE); - tw32_f(NVRAM_SWARB, SWARB_REQ_CLR0); - tg3_nvram_unlock(tp); + if (!err) + tg3_nvram_unlock(tp); } } @@ -4193,14 +4195,19 @@ static int tg3_nvram_lock(struct tg3 *tp) if (tp->tg3_flags & TG3_FLAG_NVRAM) { int i; - tw32(NVRAM_SWARB, SWARB_REQ_SET1); - for (i = 0; i < 8000; i++) { - if (tr32(NVRAM_SWARB) & SWARB_GNT1) - break; - udelay(20); + if (tp->nvram_lock_cnt == 0) { + tw32(NVRAM_SWARB, SWARB_REQ_SET1); + for (i = 0; i < 8000; i++) { + if (tr32(NVRAM_SWARB) & SWARB_GNT1) + break; + udelay(20); + } + if (i == 8000) { + tw32(NVRAM_SWARB, SWARB_REQ_CLR1); + return -ENODEV; + } } - if (i == 8000) - return -ENODEV; + tp->nvram_lock_cnt++; } return 0; } @@ -4208,8 +4215,12 @@ static int tg3_nvram_lock(struct tg3 *tp) /* tp->lock is held. */ static void tg3_nvram_unlock(struct tg3 *tp) { - if (tp->tg3_flags & TG3_FLAG_NVRAM) - tw32_f(NVRAM_SWARB, SWARB_REQ_CLR1); + if (tp->tg3_flags & TG3_FLAG_NVRAM) { + if (tp->nvram_lock_cnt > 0) + tp->nvram_lock_cnt--; + if (tp->nvram_lock_cnt == 0) + tw32_f(NVRAM_SWARB, SWARB_REQ_CLR1); + } } /* tp->lock is held. */ @@ -4320,8 +4331,13 @@ static int tg3_chip_reset(struct tg3 *tp) void (*write_op)(struct tg3 *, u32, u32); int i; - if (!(tp->tg3_flags2 & TG3_FLG2_SUN_570X)) + if (!(tp->tg3_flags2 & TG3_FLG2_SUN_570X)) { tg3_nvram_lock(tp); + /* No matching tg3_nvram_unlock() after this because + * chip reset below will undo the nvram lock. + */ + tp->nvram_lock_cnt = 0; + } /* * We must avoid the readl() that normally takes place. @@ -4717,6 +4733,10 @@ static int tg3_halt_cpu(struct tg3 *tp, u32 offset) (offset == RX_CPU_BASE ? "RX" : "TX")); return -ENODEV; } + + /* Clear firmware's nvram arbitration. */ + if (tp->tg3_flags & TG3_FLAG_NVRAM) + tw32(NVRAM_SWARB, SWARB_REQ_CLR0); return 0; } @@ -4736,7 +4756,7 @@ struct fw_info { static int tg3_load_firmware_cpu(struct tg3 *tp, u32 cpu_base, u32 cpu_scratch_base, int cpu_scratch_size, struct fw_info *info) { - int err, i; + int err, lock_err, i; void (*write_op)(struct tg3 *, u32, u32); if (cpu_base == TX_CPU_BASE && @@ -4755,9 +4775,10 @@ static int tg3_load_firmware_cpu(struct tg3 *tp, u32 cpu_base, u32 cpu_scratch_b /* It is possible that bootcode is still loading at this point. * Get the nvram lock first before halting the cpu. */ - tg3_nvram_lock(tp); + lock_err = tg3_nvram_lock(tp); err = tg3_halt_cpu(tp, cpu_base); - tg3_nvram_unlock(tp); + if (!lock_err) + tg3_nvram_unlock(tp); if (err) goto out; @@ -8182,7 +8203,7 @@ static void tg3_self_test(struct net_device *dev, struct ethtool_test *etest, data[1] = 1; } if (etest->flags & ETH_TEST_FL_OFFLINE) { - int irq_sync = 0; + int err, irq_sync = 0; if (netif_running(dev)) { tg3_netif_stop(tp); @@ -8192,11 +8213,12 @@ static void tg3_self_test(struct net_device *dev, struct ethtool_test *etest, tg3_full_lock(tp, irq_sync); tg3_halt(tp, RESET_KIND_SUSPEND, 1); - tg3_nvram_lock(tp); + err = tg3_nvram_lock(tp); tg3_halt_cpu(tp, RX_CPU_BASE); if (!(tp->tg3_flags2 & TG3_FLG2_5705_PLUS)) tg3_halt_cpu(tp, TX_CPU_BASE); - tg3_nvram_unlock(tp); + if (!err) + tg3_nvram_unlock(tp); if (tg3_test_registers(tp) != 0) { etest->flags |= ETH_TEST_FL_FAILED; @@ -8588,7 +8610,11 @@ static void __devinit tg3_nvram_init(struct tg3 *tp) GET_ASIC_REV(tp->pci_chip_rev_id) != ASIC_REV_5701) { tp->tg3_flags |= TG3_FLAG_NVRAM; - tg3_nvram_lock(tp); + if (tg3_nvram_lock(tp)) { + printk(KERN_WARNING PFX "%s: Cannot get nvarm lock, " + "tg3_nvram_init failed.\n", tp->dev->name); + return; + } tg3_enable_nvram_access(tp); if (GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5752) @@ -8686,7 +8712,9 @@ static int tg3_nvram_read(struct tg3 *tp, u32 offset, u32 *val) if (offset > NVRAM_ADDR_MSK) return -EINVAL; - tg3_nvram_lock(tp); + ret = tg3_nvram_lock(tp); + if (ret) + return ret; tg3_enable_nvram_access(tp); @@ -8785,10 +8813,6 @@ static int tg3_nvram_write_block_unbuffered(struct tg3 *tp, u32 offset, u32 len, offset = offset + (pagesize - page_off); - /* Nvram lock released by tg3_nvram_read() above, - * so need to get it again. - */ - tg3_nvram_lock(tp); tg3_enable_nvram_access(tp); /* @@ -8925,7 +8949,9 @@ static int tg3_nvram_write_block(struct tg3 *tp, u32 offset, u32 len, u8 *buf) else { u32 grc_mode; - tg3_nvram_lock(tp); + ret = tg3_nvram_lock(tp); + if (ret) + return ret; tg3_enable_nvram_access(tp); if ((tp->tg3_flags2 & TG3_FLG2_5750_PLUS) && diff --git a/drivers/net/tg3.h b/drivers/net/tg3.h index 890e1635996..e8243305f0e 100644 --- a/drivers/net/tg3.h +++ b/drivers/net/tg3.h @@ -2275,6 +2275,7 @@ struct tg3 { dma_addr_t stats_mapping; struct work_struct reset_task; + int nvram_lock_cnt; u32 nvram_size; u32 nvram_pagesize; u32 nvram_jedecnum; diff --git a/drivers/net/wireless/airo.c b/drivers/net/wireless/airo.c index ee866fd6957..a4c7ae94614 100644 --- a/drivers/net/wireless/airo.c +++ b/drivers/net/wireless/airo.c @@ -5668,13 +5668,13 @@ static int airo_set_freq(struct net_device *dev, int channel = fwrq->m; /* We should do a better check than that, * based on the card capability !!! */ - if((channel < 1) || (channel > 16)) { + if((channel < 1) || (channel > 14)) { printk(KERN_DEBUG "%s: New channel value of %d is invalid!\n", dev->name, fwrq->m); rc = -EINVAL; } else { readConfigRid(local, 1); /* Yes ! We can set it !!! */ - local->config.channelSet = (u16)(channel - 1); + local->config.channelSet = (u16) channel; set_bit (FLAG_COMMIT, &local->flags); } } @@ -5692,6 +5692,7 @@ static int airo_get_freq(struct net_device *dev, { struct airo_info *local = dev->priv; StatusRid status_rid; /* Card status info */ + int ch; readConfigRid(local, 1); if ((local->config.opmode & 0xFF) == MODE_STA_ESS) @@ -5699,16 +5700,14 @@ static int airo_get_freq(struct net_device *dev, else readStatusRid(local, &status_rid, 1); -#ifdef WEXT_USECHANNELS - fwrq->m = ((int)status_rid.channel) + 1; - fwrq->e = 0; -#else - { - int f = (int)status_rid.channel; - fwrq->m = frequency_list[f] * 100000; + ch = (int)status_rid.channel; + if((ch > 0) && (ch < 15)) { + fwrq->m = frequency_list[ch - 1] * 100000; fwrq->e = 1; + } else { + fwrq->m = ch; + fwrq->e = 0; } -#endif return 0; } @@ -5783,7 +5782,7 @@ static int airo_get_essid(struct net_device *dev, /* If none, we may want to get the one that was set */ /* Push it out ! */ - dwrq->length = status_rid.SSIDlen + 1; + dwrq->length = status_rid.SSIDlen; dwrq->flags = 1; /* active */ return 0; diff --git a/drivers/net/wireless/atmel.c b/drivers/net/wireless/atmel.c index f0ccfef6644..98a76f10a0f 100644 --- a/drivers/net/wireless/atmel.c +++ b/drivers/net/wireless/atmel.c @@ -1718,11 +1718,11 @@ static int atmel_get_essid(struct net_device *dev, if (priv->new_SSID_size != 0) { memcpy(extra, priv->new_SSID, priv->new_SSID_size); extra[priv->new_SSID_size] = '\0'; - dwrq->length = priv->new_SSID_size + 1; + dwrq->length = priv->new_SSID_size; } else { memcpy(extra, priv->SSID, priv->SSID_size); extra[priv->SSID_size] = '\0'; - dwrq->length = priv->SSID_size + 1; + dwrq->length = priv->SSID_size; } dwrq->flags = !priv->connect_to_any_BSS; /* active */ diff --git a/drivers/net/wireless/hostap/Kconfig b/drivers/net/wireless/hostap/Kconfig index 56f41c714d3..c8f6286dd35 100644 --- a/drivers/net/wireless/hostap/Kconfig +++ b/drivers/net/wireless/hostap/Kconfig @@ -26,11 +26,25 @@ config HOSTAP_FIRMWARE depends on HOSTAP ---help--- Configure Host AP driver to include support for firmware image - download. Current version supports only downloading to volatile, i.e., - RAM memory. Flash upgrade is not yet supported. + download. This option by itself only enables downloading to the + volatile memory, i.e. the card RAM. This option is required to + support cards that don't have firmware in flash, such as D-Link + DWL-520 rev E and D-Link DWL-650 rev P. - Firmware image downloading needs user space tool, prism2_srec. It is - available from http://hostap.epitest.fi/. + Firmware image downloading needs a user space tool, prism2_srec. + It is available from http://hostap.epitest.fi/. + +config HOSTAP_FIRMWARE_NVRAM + bool "Support for non-volatile firmware download" + depends on HOSTAP_FIRMWARE + ---help--- + Allow Host AP driver to write firmware images to the non-volatile + card memory, i.e. flash memory that survives power cycling. + Enable this option if you want to be able to change card firmware + permanently. + + Firmware image downloading needs a user space tool, prism2_srec. + It is available from http://hostap.epitest.fi/. config HOSTAP_PLX tristate "Host AP driver for Prism2/2.5/3 in PLX9052 PCI adaptors" diff --git a/drivers/net/wireless/hostap/Makefile b/drivers/net/wireless/hostap/Makefile index 353ccb93134..b8e41a702c0 100644 --- a/drivers/net/wireless/hostap/Makefile +++ b/drivers/net/wireless/hostap/Makefile @@ -1,4 +1,5 @@ -hostap-y := hostap_main.o +hostap-y := hostap_80211_rx.o hostap_80211_tx.o hostap_ap.o hostap_info.o \ + hostap_ioctl.o hostap_main.o hostap_proc.o obj-$(CONFIG_HOSTAP) += hostap.o obj-$(CONFIG_HOSTAP_CS) += hostap_cs.o diff --git a/drivers/net/wireless/hostap/hostap.h b/drivers/net/wireless/hostap/hostap.h index 5fac89b8ce3..5e63765219f 100644 --- a/drivers/net/wireless/hostap/hostap.h +++ b/drivers/net/wireless/hostap/hostap.h @@ -1,6 +1,15 @@ #ifndef HOSTAP_H #define HOSTAP_H +#include <linux/ethtool.h> + +#include "hostap_wlan.h" +#include "hostap_ap.h" + +static const long freq_list[] = { 2412, 2417, 2422, 2427, 2432, 2437, 2442, + 2447, 2452, 2457, 2462, 2467, 2472, 2484 }; +#define FREQ_COUNT (sizeof(freq_list) / sizeof(freq_list[0])) + /* hostap.c */ extern struct proc_dir_entry *hostap_proc; @@ -40,6 +49,26 @@ int prism2_update_comms_qual(struct net_device *dev); int prism2_sta_send_mgmt(local_info_t *local, u8 *dst, u16 stype, u8 *body, size_t bodylen); int prism2_sta_deauth(local_info_t *local, u16 reason); +int prism2_wds_add(local_info_t *local, u8 *remote_addr, + int rtnl_locked); +int prism2_wds_del(local_info_t *local, u8 *remote_addr, + int rtnl_locked, int do_not_remove); + + +/* hostap_ap.c */ + +int ap_control_add_mac(struct mac_restrictions *mac_restrictions, u8 *mac); +int ap_control_del_mac(struct mac_restrictions *mac_restrictions, u8 *mac); +void ap_control_flush_macs(struct mac_restrictions *mac_restrictions); +int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, u8 *mac); +void ap_control_kickall(struct ap_data *ap); +void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent, + struct ieee80211_crypt_data ***crypt); +int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[], + struct iw_quality qual[], int buf_size, + int aplist); +int prism2_ap_translate_scan(struct net_device *dev, char *buffer); +int prism2_hostapd(struct ap_data *ap, struct prism2_hostapd_param *param); /* hostap_proc.c */ @@ -54,4 +83,12 @@ void hostap_info_init(local_info_t *local); void hostap_info_process(local_info_t *local, struct sk_buff *skb); +/* hostap_ioctl.c */ + +extern const struct iw_handler_def hostap_iw_handler_def; +extern struct ethtool_ops prism2_ethtool_ops; + +int hostap_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd); + + #endif /* HOSTAP_H */ diff --git a/drivers/net/wireless/hostap/hostap_80211.h b/drivers/net/wireless/hostap/hostap_80211.h index bf506f50d72..1fc72fe511e 100644 --- a/drivers/net/wireless/hostap/hostap_80211.h +++ b/drivers/net/wireless/hostap/hostap_80211.h @@ -1,6 +1,9 @@ #ifndef HOSTAP_80211_H #define HOSTAP_80211_H +#include <linux/types.h> +#include <net/ieee80211_crypt.h> + struct hostap_ieee80211_mgmt { u16 frame_control; u16 duration; diff --git a/drivers/net/wireless/hostap/hostap_80211_rx.c b/drivers/net/wireless/hostap/hostap_80211_rx.c index 4b13b76425c..7e04dc94b3b 100644 --- a/drivers/net/wireless/hostap/hostap_80211_rx.c +++ b/drivers/net/wireless/hostap/hostap_80211_rx.c @@ -1,7 +1,18 @@ #include <linux/etherdevice.h> +#include <net/ieee80211_crypt.h> #include "hostap_80211.h" #include "hostap.h" +#include "hostap_ap.h" + +/* See IEEE 802.1H for LLC/SNAP encapsulation/decapsulation */ +/* Ethernet-II snap header (RFC1042 for most EtherTypes) */ +static unsigned char rfc1042_header[] = +{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0x00 }; +/* Bridge-Tunnel header (for EtherTypes ETH_P_AARP and ETH_P_IPX) */ +static unsigned char bridge_tunnel_header[] = +{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 }; +/* No encapsulation header if EtherType < 0x600 (=length) */ void hostap_dump_rx_80211(const char *name, struct sk_buff *skb, struct hostap_80211_rx_status *rx_stats) diff --git a/drivers/net/wireless/hostap/hostap_80211_tx.c b/drivers/net/wireless/hostap/hostap_80211_tx.c index 9d24f8a38ac..4a85e63906f 100644 --- a/drivers/net/wireless/hostap/hostap_80211_tx.c +++ b/drivers/net/wireless/hostap/hostap_80211_tx.c @@ -1,3 +1,18 @@ +#include "hostap_80211.h" +#include "hostap_common.h" +#include "hostap_wlan.h" +#include "hostap.h" +#include "hostap_ap.h" + +/* See IEEE 802.1H for LLC/SNAP encapsulation/decapsulation */ +/* Ethernet-II snap header (RFC1042 for most EtherTypes) */ +static unsigned char rfc1042_header[] = +{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0x00 }; +/* Bridge-Tunnel header (for EtherTypes ETH_P_AARP and ETH_P_IPX) */ +static unsigned char bridge_tunnel_header[] = +{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 }; +/* No encapsulation header if EtherType < 0x600 (=length) */ + void hostap_dump_tx_80211(const char *name, struct sk_buff *skb) { struct ieee80211_hdr_4addr *hdr; diff --git a/drivers/net/wireless/hostap/hostap_ap.c b/drivers/net/wireless/hostap/hostap_ap.c index 9da94ab7f05..753a1de6664 100644 --- a/drivers/net/wireless/hostap/hostap_ap.c +++ b/drivers/net/wireless/hostap/hostap_ap.c @@ -16,6 +16,14 @@ * (8802.11: 5.5) */ +#include <linux/proc_fs.h> +#include <linux/delay.h> +#include <linux/random.h> + +#include "hostap_wlan.h" +#include "hostap.h" +#include "hostap_ap.h" + static int other_ap_policy[MAX_PARM_DEVICES] = { AP_OTHER_AP_SKIP_ALL, DEF_INTS }; module_param_array(other_ap_policy, int, NULL, 0444); @@ -360,8 +368,7 @@ static int ap_control_proc_read(char *page, char **start, off_t off, } -static int ap_control_add_mac(struct mac_restrictions *mac_restrictions, - u8 *mac) +int ap_control_add_mac(struct mac_restrictions *mac_restrictions, u8 *mac) { struct mac_entry *entry; @@ -380,8 +387,7 @@ static int ap_control_add_mac(struct mac_restrictions *mac_restrictions, } -static int ap_control_del_mac(struct mac_restrictions *mac_restrictions, - u8 *mac) +int ap_control_del_mac(struct mac_restrictions *mac_restrictions, u8 *mac) { struct list_head *ptr; struct mac_entry *entry; @@ -433,7 +439,7 @@ static int ap_control_mac_deny(struct mac_restrictions *mac_restrictions, } -static void ap_control_flush_macs(struct mac_restrictions *mac_restrictions) +void ap_control_flush_macs(struct mac_restrictions *mac_restrictions) { struct list_head *ptr, *n; struct mac_entry *entry; @@ -454,8 +460,7 @@ static void ap_control_flush_macs(struct mac_restrictions *mac_restrictions) } -static int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, - u8 *mac) +int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, u8 *mac) { struct sta_info *sta; u16 resp; @@ -486,7 +491,7 @@ static int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, #endif /* PRISM2_NO_KERNEL_IEEE80211_MGMT */ -static void ap_control_kickall(struct ap_data *ap) +void ap_control_kickall(struct ap_data *ap) { struct list_head *ptr, *n; struct sta_info *sta; @@ -2321,9 +2326,9 @@ static void schedule_packet_send(local_info_t *local, struct sta_info *sta) } -static int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[], - struct iw_quality qual[], int buf_size, - int aplist) +int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[], + struct iw_quality qual[], int buf_size, + int aplist) { struct ap_data *ap = local->ap; struct list_head *ptr; @@ -2363,7 +2368,7 @@ static int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[], /* Translate our list of Access Points & Stations to a card independant * format that the Wireless Tools will understand - Jean II */ -static int prism2_ap_translate_scan(struct net_device *dev, char *buffer) +int prism2_ap_translate_scan(struct net_device *dev, char *buffer) { struct hostap_interface *iface; local_info_t *local; @@ -2608,8 +2613,7 @@ static int prism2_hostapd_sta_clear_stats(struct ap_data *ap, } -static int prism2_hostapd(struct ap_data *ap, - struct prism2_hostapd_param *param) +int prism2_hostapd(struct ap_data *ap, struct prism2_hostapd_param *param) { switch (param->cmd) { case PRISM2_HOSTAPD_FLUSH: @@ -3207,8 +3211,8 @@ void hostap_update_rates(local_info_t *local) } -static void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent, - struct ieee80211_crypt_data ***crypt) +void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent, + struct ieee80211_crypt_data ***crypt) { struct sta_info *sta; diff --git a/drivers/net/wireless/hostap/hostap_ap.h b/drivers/net/wireless/hostap/hostap_ap.h index 6d00df69c2e..2fa2452b6b0 100644 --- a/drivers/net/wireless/hostap/hostap_ap.h +++ b/drivers/net/wireless/hostap/hostap_ap.h @@ -1,6 +1,8 @@ #ifndef HOSTAP_AP_H #define HOSTAP_AP_H +#include "hostap_80211.h" + /* AP data structures for STAs */ /* maximum number of frames to buffer per STA */ diff --git a/drivers/net/wireless/hostap/hostap_common.h b/drivers/net/wireless/hostap/hostap_common.h index 6f4fa9dc308..01624005d80 100644 --- a/drivers/net/wireless/hostap/hostap_common.h +++ b/drivers/net/wireless/hostap/hostap_common.h @@ -1,6 +1,9 @@ #ifndef HOSTAP_COMMON_H #define HOSTAP_COMMON_H +#include <linux/types.h> +#include <linux/if_ether.h> + #define BIT(x) (1 << (x)) #define MAC2STR(a) (a)[0], (a)[1], (a)[2], (a)[3], (a)[4], (a)[5] diff --git a/drivers/net/wireless/hostap/hostap_config.h b/drivers/net/wireless/hostap/hostap_config.h index 7ed3425d08c..c090a5aebb5 100644 --- a/drivers/net/wireless/hostap/hostap_config.h +++ b/drivers/net/wireless/hostap/hostap_config.h @@ -21,15 +21,10 @@ #define PRISM2_DOWNLOAD_SUPPORT #endif -#ifdef PRISM2_DOWNLOAD_SUPPORT -/* Allow writing firmware images into flash, i.e., to non-volatile storage. - * Before you enable this option, you should make absolutely sure that you are - * using prism2_srec utility that comes with THIS version of the driver! - * In addition, please note that it is possible to kill your card with - * non-volatile download if you are using incorrect image. This feature has not - * been fully tested, so please be careful with it. */ -/* #define PRISM2_NON_VOLATILE_DOWNLOAD */ -#endif /* PRISM2_DOWNLOAD_SUPPORT */ +/* Allow kernel configuration to enable non-volatile download support. */ +#ifdef CONFIG_HOSTAP_FIRMWARE_NVRAM +#define PRISM2_NON_VOLATILE_DOWNLOAD +#endif /* Save low-level I/O for debugging. This should not be enabled in normal use. */ diff --git a/drivers/net/wireless/hostap/hostap_info.c b/drivers/net/wireless/hostap/hostap_info.c index 5aa998fdf1c..50f72d831cf 100644 --- a/drivers/net/wireless/hostap/hostap_info.c +++ b/drivers/net/wireless/hostap/hostap_info.c @@ -1,5 +1,8 @@ /* Host AP driver Info Frame processing (part of hostap.o module) */ +#include "hostap_wlan.h" +#include "hostap.h" +#include "hostap_ap.h" /* Called only as a tasklet (software IRQ) */ static void prism2_info_commtallies16(local_info_t *local, unsigned char *buf, diff --git a/drivers/net/wireless/hostap/hostap_ioctl.c b/drivers/net/wireless/hostap/hostap_ioctl.c index 2617d70bcda..f3e0ce1ee03 100644 --- a/drivers/net/wireless/hostap/hostap_ioctl.c +++ b/drivers/net/wireless/hostap/hostap_ioctl.c @@ -1,11 +1,13 @@ /* ioctl() (mostly Linux Wireless Extensions) routines for Host AP driver */ -#ifdef in_atomic -/* Get kernel_locked() for in_atomic() */ +#include <linux/types.h> #include <linux/smp_lock.h> -#endif #include <linux/ethtool.h> +#include <net/ieee80211_crypt.h> +#include "hostap_wlan.h" +#include "hostap.h" +#include "hostap_ap.h" static struct iw_statistics *hostap_get_wireless_stats(struct net_device *dev) { @@ -3910,7 +3912,7 @@ static void prism2_get_drvinfo(struct net_device *dev, local->sta_fw_ver & 0xff); } -static struct ethtool_ops prism2_ethtool_ops = { +struct ethtool_ops prism2_ethtool_ops = { .get_drvinfo = prism2_get_drvinfo }; @@ -3985,7 +3987,7 @@ static const iw_handler prism2_private_handler[] = (iw_handler) prism2_ioctl_priv_readmif, /* 3 */ }; -static const struct iw_handler_def hostap_iw_handler_def = +const struct iw_handler_def hostap_iw_handler_def = { .num_standard = sizeof(prism2_handler) / sizeof(iw_handler), .num_private = sizeof(prism2_private_handler) / sizeof(iw_handler), diff --git a/drivers/net/wireless/hostap/hostap_main.c b/drivers/net/wireless/hostap/hostap_main.c index 3d2ea61033b..8dd4c4446a6 100644 --- a/drivers/net/wireless/hostap/hostap_main.c +++ b/drivers/net/wireless/hostap/hostap_main.c @@ -24,6 +24,7 @@ #include <linux/kmod.h> #include <linux/rtnetlink.h> #include <linux/wireless.h> +#include <linux/etherdevice.h> #include <net/iw_handler.h> #include <net/ieee80211.h> #include <net/ieee80211_crypt.h> @@ -47,57 +48,6 @@ MODULE_VERSION(PRISM2_VERSION); #define PRISM2_MAX_MTU (PRISM2_MAX_FRAME_SIZE - (6 /* LLC */ + 8 /* WEP */)) -/* hostap.c */ -static int prism2_wds_add(local_info_t *local, u8 *remote_addr, - int rtnl_locked); -static int prism2_wds_del(local_info_t *local, u8 *remote_addr, - int rtnl_locked, int do_not_remove); - -/* hostap_ap.c */ -static int prism2_ap_get_sta_qual(local_info_t *local, struct sockaddr addr[], - struct iw_quality qual[], int buf_size, - int aplist); -static int prism2_ap_translate_scan(struct net_device *dev, char *buffer); -static int prism2_hostapd(struct ap_data *ap, - struct prism2_hostapd_param *param); -static void * ap_crypt_get_ptrs(struct ap_data *ap, u8 *addr, int permanent, - struct ieee80211_crypt_data ***crypt); -static void ap_control_kickall(struct ap_data *ap); -#ifndef PRISM2_NO_KERNEL_IEEE80211_MGMT -static int ap_control_add_mac(struct mac_restrictions *mac_restrictions, - u8 *mac); -static int ap_control_del_mac(struct mac_restrictions *mac_restrictions, - u8 *mac); -static void ap_control_flush_macs(struct mac_restrictions *mac_restrictions); -static int ap_control_kick_mac(struct ap_data *ap, struct net_device *dev, - u8 *mac); -#endif /* !PRISM2_NO_KERNEL_IEEE80211_MGMT */ - - -static const long freq_list[] = { 2412, 2417, 2422, 2427, 2432, 2437, 2442, - 2447, 2452, 2457, 2462, 2467, 2472, 2484 }; -#define FREQ_COUNT (sizeof(freq_list) / sizeof(freq_list[0])) - - -/* See IEEE 802.1H for LLC/SNAP encapsulation/decapsulation */ -/* Ethernet-II snap header (RFC1042 for most EtherTypes) */ -static unsigned char rfc1042_header[] = -{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0x00 }; -/* Bridge-Tunnel header (for EtherTypes ETH_P_AARP and ETH_P_IPX) */ -static unsigned char bridge_tunnel_header[] = -{ 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 }; -/* No encapsulation header if EtherType < 0x600 (=length) */ - - -/* FIX: these could be compiled separately and linked together to hostap.o */ -#include "hostap_ap.c" -#include "hostap_info.c" -#include "hostap_ioctl.c" -#include "hostap_proc.c" -#include "hostap_80211_rx.c" -#include "hostap_80211_tx.c" - - struct net_device * hostap_add_interface(struct local_info *local, int type, int rtnl_locked, const char *prefix, @@ -196,8 +146,8 @@ static inline int prism2_wds_special_addr(u8 *addr) } -static int prism2_wds_add(local_info_t *local, u8 *remote_addr, - int rtnl_locked) +int prism2_wds_add(local_info_t *local, u8 *remote_addr, + int rtnl_locked) { struct net_device *dev; struct list_head *ptr; @@ -258,8 +208,8 @@ static int prism2_wds_add(local_info_t *local, u8 *remote_addr, } -static int prism2_wds_del(local_info_t *local, u8 *remote_addr, - int rtnl_locked, int do_not_remove) +int prism2_wds_del(local_info_t *local, u8 *remote_addr, + int rtnl_locked, int do_not_remove) { unsigned long flags; struct list_head *ptr; diff --git a/drivers/net/wireless/hostap/hostap_proc.c b/drivers/net/wireless/hostap/hostap_proc.c index a0a4cbd4937..d1d8ce022e6 100644 --- a/drivers/net/wireless/hostap/hostap_proc.c +++ b/drivers/net/wireless/hostap/hostap_proc.c @@ -1,5 +1,12 @@ /* /proc routines for Host AP driver */ +#include <linux/types.h> +#include <linux/proc_fs.h> +#include <net/ieee80211_crypt.h> + +#include "hostap_wlan.h" +#include "hostap.h" + #define PROC_LIMIT (PAGE_SIZE - 80) diff --git a/drivers/net/wireless/hostap/hostap_wlan.h b/drivers/net/wireless/hostap/hostap_wlan.h index cfd80155949..87a54aa6f4d 100644 --- a/drivers/net/wireless/hostap/hostap_wlan.h +++ b/drivers/net/wireless/hostap/hostap_wlan.h @@ -1,6 +1,10 @@ #ifndef HOSTAP_WLAN_H #define HOSTAP_WLAN_H +#include <linux/wireless.h> +#include <linux/netdevice.h> +#include <net/iw_handler.h> + #include "hostap_config.h" #include "hostap_common.h" diff --git a/drivers/net/wireless/ipw2100.c b/drivers/net/wireless/ipw2100.c index 7518384f34d..8bf02763b5c 100644 --- a/drivers/net/wireless/ipw2100.c +++ b/drivers/net/wireless/ipw2100.c @@ -5735,70 +5735,6 @@ static struct net_device_stats *ipw2100_stats(struct net_device *dev) return &priv->ieee->stats; } -#if WIRELESS_EXT < 18 -/* Support for wpa_supplicant before WE-18, deprecated. */ - -/* following definitions must match definitions in driver_ipw.c */ - -#define IPW2100_IOCTL_WPA_SUPPLICANT SIOCIWFIRSTPRIV+30 - -#define IPW2100_CMD_SET_WPA_PARAM 1 -#define IPW2100_CMD_SET_WPA_IE 2 -#define IPW2100_CMD_SET_ENCRYPTION 3 -#define IPW2100_CMD_MLME 4 - -#define IPW2100_PARAM_WPA_ENABLED 1 -#define IPW2100_PARAM_TKIP_COUNTERMEASURES 2 -#define IPW2100_PARAM_DROP_UNENCRYPTED 3 -#define IPW2100_PARAM_PRIVACY_INVOKED 4 -#define IPW2100_PARAM_AUTH_ALGS 5 -#define IPW2100_PARAM_IEEE_802_1X 6 - -#define IPW2100_MLME_STA_DEAUTH 1 -#define IPW2100_MLME_STA_DISASSOC 2 - -#define IPW2100_CRYPT_ERR_UNKNOWN_ALG 2 -#define IPW2100_CRYPT_ERR_UNKNOWN_ADDR 3 -#define IPW2100_CRYPT_ERR_CRYPT_INIT_FAILED 4 -#define IPW2100_CRYPT_ERR_KEY_SET_FAILED 5 -#define IPW2100_CRYPT_ERR_TX_KEY_SET_FAILED 6 -#define IPW2100_CRYPT_ERR_CARD_CONF_FAILED 7 - -#define IPW2100_CRYPT_ALG_NAME_LEN 16 - -struct ipw2100_param { - u32 cmd; - u8 sta_addr[ETH_ALEN]; - union { - struct { - u8 name; - u32 value; - } wpa_param; - struct { - u32 len; - u8 reserved[32]; - u8 data[0]; - } wpa_ie; - struct { - u32 command; - u32 reason_code; - } mlme; - struct { - u8 alg[IPW2100_CRYPT_ALG_NAME_LEN]; - u8 set_tx; - u32 err; - u8 idx; - u8 seq[8]; /* sequence counter (set: RX, get: TX) */ - u16 key_len; - u8 key[0]; - } crypt; - - } u; -}; - -/* end of driver_ipw.c code */ -#endif /* WIRELESS_EXT < 18 */ - static int ipw2100_wpa_enable(struct ipw2100_priv *priv, int value) { /* This is called when wpa_supplicant loads and closes the driver @@ -5807,11 +5743,6 @@ static int ipw2100_wpa_enable(struct ipw2100_priv *priv, int value) return 0; } -#if WIRELESS_EXT < 18 -#define IW_AUTH_ALG_OPEN_SYSTEM 0x1 -#define IW_AUTH_ALG_SHARED_KEY 0x2 -#endif - static int ipw2100_wpa_set_auth_algs(struct ipw2100_priv *priv, int value) { @@ -5855,360 +5786,6 @@ void ipw2100_wpa_assoc_frame(struct ipw2100_priv *priv, ipw2100_set_wpa_ie(priv, &frame, 0); } -#if WIRELESS_EXT < 18 -static int ipw2100_wpa_set_param(struct net_device *dev, u8 name, u32 value) -{ - struct ipw2100_priv *priv = ieee80211_priv(dev); - struct ieee80211_crypt_data *crypt; - unsigned long flags; - int ret = 0; - - switch (name) { - case IPW2100_PARAM_WPA_ENABLED: - ret = ipw2100_wpa_enable(priv, value); - break; - - case IPW2100_PARAM_TKIP_COUNTERMEASURES: - crypt = priv->ieee->crypt[priv->ieee->tx_keyidx]; - if (!crypt || !crypt->ops->set_flags || !crypt->ops->get_flags) - break; - - flags = crypt->ops->get_flags(crypt->priv); - - if (value) - flags |= IEEE80211_CRYPTO_TKIP_COUNTERMEASURES; - else - flags &= ~IEEE80211_CRYPTO_TKIP_COUNTERMEASURES; - - crypt->ops->set_flags(flags, crypt->priv); - - break; - - case IPW2100_PARAM_DROP_UNENCRYPTED:{ - /* See IW_AUTH_DROP_UNENCRYPTED handling for details */ - struct ieee80211_security sec = { - .flags = SEC_ENABLED, - .enabled = value, - }; - priv->ieee->drop_unencrypted = value; - /* We only change SEC_LEVEL for open mode. Others - * are set by ipw_wpa_set_encryption. - */ - if (!value) { - sec.flags |= SEC_LEVEL; - sec.level = SEC_LEVEL_0; - } else { - sec.flags |= SEC_LEVEL; - sec.level = SEC_LEVEL_1; - } - if (priv->ieee->set_security) - priv->ieee->set_security(priv->ieee->dev, &sec); - break; - } - - case IPW2100_PARAM_PRIVACY_INVOKED: - priv->ieee->privacy_invoked = value; - break; - - case IPW2100_PARAM_AUTH_ALGS: - ret = ipw2100_wpa_set_auth_algs(priv, value); - break; - - case IPW2100_PARAM_IEEE_802_1X: - priv->ieee->ieee802_1x = value; - break; - - default: - printk(KERN_ERR DRV_NAME ": %s: Unknown WPA param: %d\n", - dev->name, name); - ret = -EOPNOTSUPP; - } - - return ret; -} - -static int ipw2100_wpa_mlme(struct net_device *dev, int command, int reason) -{ - - struct ipw2100_priv *priv = ieee80211_priv(dev); - int ret = 0; - - switch (command) { - case IPW2100_MLME_STA_DEAUTH: - // silently ignore - break; - - case IPW2100_MLME_STA_DISASSOC: - ipw2100_disassociate_bssid(priv); - break; - - default: - printk(KERN_ERR DRV_NAME ": %s: Unknown MLME request: %d\n", - dev->name, command); - ret = -EOPNOTSUPP; - } - - return ret; -} - -static int ipw2100_wpa_set_wpa_ie(struct net_device *dev, - struct ipw2100_param *param, int plen) -{ - - struct ipw2100_priv *priv = ieee80211_priv(dev); - struct ieee80211_device *ieee = priv->ieee; - u8 *buf; - - if (!ieee->wpa_enabled) - return -EOPNOTSUPP; - - if (param->u.wpa_ie.len > MAX_WPA_IE_LEN || - (param->u.wpa_ie.len && param->u.wpa_ie.data == NULL)) - return -EINVAL; - - if (param->u.wpa_ie.len) { - buf = kmalloc(param->u.wpa_ie.len, GFP_KERNEL); - if (buf == NULL) - return -ENOMEM; - - memcpy(buf, param->u.wpa_ie.data, param->u.wpa_ie.len); - - kfree(ieee->wpa_ie); - ieee->wpa_ie = buf; - ieee->wpa_ie_len = param->u.wpa_ie.len; - - } else { - kfree(ieee->wpa_ie); - ieee->wpa_ie = NULL; - ieee->wpa_ie_len = 0; - } - - ipw2100_wpa_assoc_frame(priv, ieee->wpa_ie, ieee->wpa_ie_len); - - return 0; -} - -/* implementation borrowed from hostap driver */ - -static int ipw2100_wpa_set_encryption(struct net_device *dev, - struct ipw2100_param *param, - int param_len) -{ - int ret = 0; - struct ipw2100_priv *priv = ieee80211_priv(dev); - struct ieee80211_device *ieee = priv->ieee; - struct ieee80211_crypto_ops *ops; - struct ieee80211_crypt_data **crypt; - - struct ieee80211_security sec = { - .flags = 0, - }; - - param->u.crypt.err = 0; - param->u.crypt.alg[IPW2100_CRYPT_ALG_NAME_LEN - 1] = '\0'; - - if (param_len != - (int)((char *)param->u.crypt.key - (char *)param) + - param->u.crypt.key_len) { - IPW_DEBUG_INFO("Len mismatch %d, %d\n", param_len, - param->u.crypt.key_len); - return -EINVAL; - } - if (param->sta_addr[0] == 0xff && param->sta_addr[1] == 0xff && - param->sta_addr[2] == 0xff && param->sta_addr[3] == 0xff && - param->sta_addr[4] == 0xff && param->sta_addr[5] == 0xff) { - if (param->u.crypt.idx >= WEP_KEYS) - return -EINVAL; - crypt = &ieee->crypt[param->u.crypt.idx]; - } else { - return -EINVAL; - } - - sec.flags |= SEC_ENABLED | SEC_ENCRYPT; - if (strcmp(param->u.crypt.alg, "none") == 0) { - if (crypt) { - sec.enabled = 0; - sec.encrypt = 0; - sec.level = SEC_LEVEL_0; - sec.flags |= SEC_LEVEL; - ieee80211_crypt_delayed_deinit(ieee, crypt); - } - goto done; - } - sec.enabled = 1; - sec.encrypt = 1; - - ops = ieee80211_get_crypto_ops(param->u.crypt.alg); - if (ops == NULL && strcmp(param->u.crypt.alg, "WEP") == 0) { - request_module("ieee80211_crypt_wep"); - ops = ieee80211_get_crypto_ops(param->u.crypt.alg); - } else if (ops == NULL && strcmp(param->u.crypt.alg, "TKIP") == 0) { - request_module("ieee80211_crypt_tkip"); - ops = ieee80211_get_crypto_ops(param->u.crypt.alg); - } else if (ops == NULL && strcmp(param->u.crypt.alg, "CCMP") == 0) { - request_module("ieee80211_crypt_ccmp"); - ops = ieee80211_get_crypto_ops(param->u.crypt.alg); - } - if (ops == NULL) { - IPW_DEBUG_INFO("%s: unknown crypto alg '%s'\n", - dev->name, param->u.crypt.alg); - param->u.crypt.err = IPW2100_CRYPT_ERR_UNKNOWN_ALG; - ret = -EINVAL; - goto done; - } - - if (*crypt == NULL || (*crypt)->ops != ops) { - struct ieee80211_crypt_data *new_crypt; - - ieee80211_crypt_delayed_deinit(ieee, crypt); - - new_crypt = kzalloc(sizeof(struct ieee80211_crypt_data), GFP_KERNEL); - if (new_crypt == NULL) { - ret = -ENOMEM; - goto done; - } - new_crypt->ops = ops; - if (new_crypt->ops && try_module_get(new_crypt->ops->owner)) - new_crypt->priv = - new_crypt->ops->init(param->u.crypt.idx); - - if (new_crypt->priv == NULL) { - kfree(new_crypt); - param->u.crypt.err = - IPW2100_CRYPT_ERR_CRYPT_INIT_FAILED; - ret = -EINVAL; - goto done; - } - - *crypt = new_crypt; - } - - if (param->u.crypt.key_len > 0 && (*crypt)->ops->set_key && - (*crypt)->ops->set_key(param->u.crypt.key, - param->u.crypt.key_len, param->u.crypt.seq, - (*crypt)->priv) < 0) { - IPW_DEBUG_INFO("%s: key setting failed\n", dev->name); - param->u.crypt.err = IPW2100_CRYPT_ERR_KEY_SET_FAILED; - ret = -EINVAL; - goto done; - } - - if (param->u.crypt.set_tx) { - ieee->tx_keyidx = param->u.crypt.idx; - sec.active_key = param->u.crypt.idx; - sec.flags |= SEC_ACTIVE_KEY; - } - - if (ops->name != NULL) { - - if (strcmp(ops->name, "WEP") == 0) { - memcpy(sec.keys[param->u.crypt.idx], - param->u.crypt.key, param->u.crypt.key_len); - sec.key_sizes[param->u.crypt.idx] = - param->u.crypt.key_len; - sec.flags |= (1 << param->u.crypt.idx); - sec.flags |= SEC_LEVEL; - sec.level = SEC_LEVEL_1; - } else if (strcmp(ops->name, "TKIP") == 0) { - sec.flags |= SEC_LEVEL; - sec.level = SEC_LEVEL_2; - } else if (strcmp(ops->name, "CCMP") == 0) { - sec.flags |= SEC_LEVEL; - sec.level = SEC_LEVEL_3; - } - } - done: - if (ieee->set_security) - ieee->set_security(ieee->dev, &sec); - - /* Do not reset port if card is in Managed mode since resetting will - * generate new IEEE 802.11 authentication which may end up in looping - * with IEEE 802.1X. If your hardware requires a reset after WEP - * configuration (for example... Prism2), implement the reset_port in - * the callbacks structures used to initialize the 802.11 stack. */ - if (ieee->reset_on_keychange && - ieee->iw_mode != IW_MODE_INFRA && - ieee->reset_port && ieee->reset_port(dev)) { - IPW_DEBUG_INFO("%s: reset_port failed\n", dev->name); - param->u.crypt.err = IPW2100_CRYPT_ERR_CARD_CONF_FAILED; - return -EINVAL; - } - - return ret; -} - -static int ipw2100_wpa_supplicant(struct net_device *dev, struct iw_point *p) -{ - - struct ipw2100_param *param; - int ret = 0; - - IPW_DEBUG_IOCTL("wpa_supplicant: len=%d\n", p->length); - - if (p->length < sizeof(struct ipw2100_param) || !p->pointer) - return -EINVAL; - - param = (struct ipw2100_param *)kmalloc(p->length, GFP_KERNEL); - if (param == NULL) - return -ENOMEM; - - if (copy_from_user(param, p->pointer, p->length)) { - kfree(param); - return -EFAULT; - } - - switch (param->cmd) { - - case IPW2100_CMD_SET_WPA_PARAM: - ret = ipw2100_wpa_set_param(dev, param->u.wpa_param.name, - param->u.wpa_param.value); - break; - - case IPW2100_CMD_SET_WPA_IE: - ret = ipw2100_wpa_set_wpa_ie(dev, param, p->length); - break; - - case IPW2100_CMD_SET_ENCRYPTION: - ret = ipw2100_wpa_set_encryption(dev, param, p->length); - break; - - case IPW2100_CMD_MLME: - ret = ipw2100_wpa_mlme(dev, param->u.mlme.command, - param->u.mlme.reason_code); - break; - - default: - printk(KERN_ERR DRV_NAME - ": %s: Unknown WPA supplicant request: %d\n", dev->name, - param->cmd); - ret = -EOPNOTSUPP; - - } - - if (ret == 0 && copy_to_user(p->pointer, param, p->length)) - ret = -EFAULT; - - kfree(param); - return ret; -} - -static int ipw2100_ioctl(struct net_device *dev, struct ifreq *rq, int cmd) -{ - struct iwreq *wrq = (struct iwreq *)rq; - int ret = -1; - switch (cmd) { - case IPW2100_IOCTL_WPA_SUPPLICANT: - ret = ipw2100_wpa_supplicant(dev, &wrq->u.data); - return ret; - - default: - return -EOPNOTSUPP; - } - - return -EOPNOTSUPP; -} -#endif /* WIRELESS_EXT < 18 */ - static void ipw_ethtool_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info) { @@ -6337,9 +5914,6 @@ static struct net_device *ipw2100_alloc_device(struct pci_dev *pci_dev, dev->open = ipw2100_open; dev->stop = ipw2100_close; dev->init = ipw2100_net_init; -#if WIRELESS_EXT < 18 - dev->do_ioctl = ipw2100_ioctl; -#endif dev->get_stats = ipw2100_stats; dev->ethtool_ops = &ipw2100_ethtool_ops; dev->tx_timeout = ipw2100_tx_timeout; @@ -7855,7 +7429,6 @@ static int ipw2100_wx_get_power(struct net_device *dev, return 0; } -#if WIRELESS_EXT > 17 /* * WE-18 WPA support */ @@ -8117,7 +7690,6 @@ static int ipw2100_wx_set_mlme(struct net_device *dev, } return 0; } -#endif /* WIRELESS_EXT > 17 */ /* * @@ -8350,11 +7922,7 @@ static iw_handler ipw2100_wx_handlers[] = { NULL, /* SIOCWIWTHRSPY */ ipw2100_wx_set_wap, /* SIOCSIWAP */ ipw2100_wx_get_wap, /* SIOCGIWAP */ -#if WIRELESS_EXT > 17 ipw2100_wx_set_mlme, /* SIOCSIWMLME */ -#else - NULL, /* -- hole -- */ -#endif NULL, /* SIOCGIWAPLIST -- deprecated */ ipw2100_wx_set_scan, /* SIOCSIWSCAN */ ipw2100_wx_get_scan, /* SIOCGIWSCAN */ @@ -8378,7 +7946,6 @@ static iw_handler ipw2100_wx_handlers[] = { ipw2100_wx_get_encode, /* SIOCGIWENCODE */ ipw2100_wx_set_power, /* SIOCSIWPOWER */ ipw2100_wx_get_power, /* SIOCGIWPOWER */ -#if WIRELESS_EXT > 17 NULL, /* -- hole -- */ NULL, /* -- hole -- */ ipw2100_wx_set_genie, /* SIOCSIWGENIE */ @@ -8388,7 +7955,6 @@ static iw_handler ipw2100_wx_handlers[] = { ipw2100_wx_set_encodeext, /* SIOCSIWENCODEEXT */ ipw2100_wx_get_encodeext, /* SIOCGIWENCODEEXT */ NULL, /* SIOCSIWPMKSA */ -#endif }; #define IPW2100_PRIV_SET_MONITOR SIOCIWFIRSTPRIV diff --git a/drivers/net/wireless/ipw2200.c b/drivers/net/wireless/ipw2200.c index 819be2b6b7d..4c28e332ecc 100644 --- a/drivers/net/wireless/ipw2200.c +++ b/drivers/net/wireless/ipw2200.c @@ -8936,14 +8936,12 @@ static int ipw_request_direct_scan(struct ipw_priv *priv, char *essid, IPW_DEBUG_HC("starting request direct scan!\n"); if (priv->status & (STATUS_SCANNING | STATUS_SCAN_ABORTING)) { - err = wait_event_interruptible(priv->wait_state, - !(priv-> - status & (STATUS_SCANNING | - STATUS_SCAN_ABORTING))); - if (err) { - IPW_DEBUG_HC("aborting direct scan"); - goto done; - } + /* We should not sleep here; otherwise we will block most + * of the system (for instance, we hold rtnl_lock when we + * get here). + */ + err = -EAGAIN; + goto done; } memset(&scan, 0, sizeof(scan)); diff --git a/drivers/net/wireless/prism54/isl_ioctl.c b/drivers/net/wireless/prism54/isl_ioctl.c index 135a156db25..c5cd61c7f92 100644 --- a/drivers/net/wireless/prism54/isl_ioctl.c +++ b/drivers/net/wireless/prism54/isl_ioctl.c @@ -748,7 +748,7 @@ prism54_get_essid(struct net_device *ndev, struct iw_request_info *info, if (essid->length) { dwrq->flags = 1; /* set ESSID to ON for Wireless Extensions */ /* if it is to big, trunk it */ - dwrq->length = min(IW_ESSID_MAX_SIZE, essid->length + 1); + dwrq->length = min(IW_ESSID_MAX_SIZE, essid->length); } else { dwrq->flags = 0; dwrq->length = 0; diff --git a/drivers/net/wireless/prism54/islpci_eth.c b/drivers/net/wireless/prism54/islpci_eth.c index 33d64d2ee53..a8261d8454d 100644 --- a/drivers/net/wireless/prism54/islpci_eth.c +++ b/drivers/net/wireless/prism54/islpci_eth.c @@ -177,7 +177,7 @@ islpci_eth_transmit(struct sk_buff *skb, struct net_device *ndev) #endif newskb->dev = skb->dev; - dev_kfree_skb(skb); + dev_kfree_skb_irq(skb); skb = newskb; } } diff --git a/drivers/net/wireless/ray_cs.c b/drivers/net/wireless/ray_cs.c index 319180ca7e7..7880d8c31aa 100644 --- a/drivers/net/wireless/ray_cs.c +++ b/drivers/net/wireless/ray_cs.c @@ -1256,7 +1256,7 @@ static int ray_get_essid(struct net_device *dev, extra[IW_ESSID_MAX_SIZE] = '\0'; /* Push it out ! */ - dwrq->length = strlen(extra) + 1; + dwrq->length = strlen(extra); dwrq->flags = 1; /* active */ return 0; diff --git a/drivers/net/wireless/wavelan_cs.c b/drivers/net/wireless/wavelan_cs.c index 7e2039f52c4..cf373625fc7 100644 --- a/drivers/net/wireless/wavelan_cs.c +++ b/drivers/net/wireless/wavelan_cs.c @@ -2280,7 +2280,7 @@ static int wavelan_get_essid(struct net_device *dev, extra[IW_ESSID_MAX_SIZE] = '\0'; /* Set the length */ - wrqu->data.length = strlen(extra) + 1; + wrqu->data.length = strlen(extra); return 0; } diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 605f0df0bfb..dda6099903c 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -1142,6 +1142,9 @@ static void __devinit quirk_intel_ide_combined(struct pci_dev *pdev) case 0x27c4: ich = 7; break; + case 0x2828: /* ICH8M */ + ich = 8; + break; default: /* we do not handle this PCI device */ return; @@ -1161,7 +1164,7 @@ static void __devinit quirk_intel_ide_combined(struct pci_dev *pdev) else return; /* not in combined mode */ } else { - WARN_ON((ich != 6) && (ich != 7)); + WARN_ON((ich != 6) && (ich != 7) && (ich != 8)); tmp &= 0x3; /* interesting bits 1:0 */ if (tmp & (1 << 0)) comb = (1 << 2); /* PATA port 0, SATA port 1 */ diff --git a/drivers/scsi/ahci.c b/drivers/scsi/ahci.c index d113290b5fc..19bd346951d 100644 --- a/drivers/scsi/ahci.c +++ b/drivers/scsi/ahci.c @@ -276,6 +276,16 @@ static const struct pci_device_id ahci_pci_tbl[] = { board_ahci }, /* ESB2 */ { PCI_VENDOR_ID_INTEL, 0x27c6, PCI_ANY_ID, PCI_ANY_ID, 0, 0, board_ahci }, /* ICH7-M DH */ + { PCI_VENDOR_ID_INTEL, 0x2821, PCI_ANY_ID, PCI_ANY_ID, 0, 0, + board_ahci }, /* ICH8 */ + { PCI_VENDOR_ID_INTEL, 0x2822, PCI_ANY_ID, PCI_ANY_ID, 0, 0, + board_ahci }, /* ICH8 */ + { PCI_VENDOR_ID_INTEL, 0x2824, PCI_ANY_ID, PCI_ANY_ID, 0, 0, + board_ahci }, /* ICH8 */ + { PCI_VENDOR_ID_INTEL, 0x2829, PCI_ANY_ID, PCI_ANY_ID, 0, 0, + board_ahci }, /* ICH8M */ + { PCI_VENDOR_ID_INTEL, 0x282a, PCI_ANY_ID, PCI_ANY_ID, 0, 0, + board_ahci }, /* ICH8M */ { } /* terminate list */ }; diff --git a/drivers/scsi/ata_piix.c b/drivers/scsi/ata_piix.c index 557788ec4ee..fc3ca051cee 100644 --- a/drivers/scsi/ata_piix.c +++ b/drivers/scsi/ata_piix.c @@ -157,6 +157,9 @@ static const struct pci_device_id piix_pci_tbl[] = { { 0x8086, 0x27c0, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci }, { 0x8086, 0x27c4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci }, { 0x8086, 0x2680, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci }, + { 0x8086, 0x2820, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci }, + { 0x8086, 0x2825, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci }, + { 0x8086, 0x2828, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci }, { } /* terminate list */ }; diff --git a/drivers/scsi/libata-core.c b/drivers/scsi/libata-core.c index 99bae8369ab..46c4cdbaee8 100644 --- a/drivers/scsi/libata-core.c +++ b/drivers/scsi/libata-core.c @@ -611,6 +611,10 @@ int ata_rwcmd_protocol(struct ata_queued_cmd *qc) if (dev->flags & ATA_DFLAG_PIO) { tf->protocol = ATA_PROT_PIO; index = dev->multi_count ? 0 : 8; + } else if (lba48 && (qc->ap->flags & ATA_FLAG_PIO_LBA48)) { + /* Unable to use DMA due to host limitation */ + tf->protocol = ATA_PROT_PIO; + index = dev->multi_count ? 0 : 4; } else { tf->protocol = ATA_PROT_DMA; index = 16; @@ -1051,18 +1055,22 @@ static unsigned int ata_pio_modes(const struct ata_device *adev) { u16 modes; - /* Usual case. Word 53 indicates word 88 is valid */ - if (adev->id[ATA_ID_FIELD_VALID] & (1 << 2)) { + /* Usual case. Word 53 indicates word 64 is valid */ + if (adev->id[ATA_ID_FIELD_VALID] & (1 << 1)) { modes = adev->id[ATA_ID_PIO_MODES] & 0x03; modes <<= 3; modes |= 0x7; return modes; } - /* If word 88 isn't valid then Word 51 holds the PIO timing number - for the maximum. Turn it into a mask and return it */ - modes = (2 << (adev->id[ATA_ID_OLD_PIO_MODES] & 0xFF)) - 1 ; + /* If word 64 isn't valid then Word 51 high byte holds the PIO timing + number for the maximum. Turn it into a mask and return it */ + modes = (2 << ((adev->id[ATA_ID_OLD_PIO_MODES] >> 8) & 0xFF)) - 1 ; return modes; + /* But wait.. there's more. Design your standards by committee and + you too can get a free iordy field to process. However its the + speeds not the modes that are supported... Note drivers using the + timing API will get this right anyway */ } struct ata_exec_internal_arg { @@ -1165,6 +1173,39 @@ ata_exec_internal(struct ata_port *ap, struct ata_device *dev, } /** + * ata_pio_need_iordy - check if iordy needed + * @adev: ATA device + * + * Check if the current speed of the device requires IORDY. Used + * by various controllers for chip configuration. + */ + +unsigned int ata_pio_need_iordy(const struct ata_device *adev) +{ + int pio; + int speed = adev->pio_mode - XFER_PIO_0; + + if (speed < 2) + return 0; + if (speed > 2) + return 1; + + /* If we have no drive specific rule, then PIO 2 is non IORDY */ + + if (adev->id[ATA_ID_FIELD_VALID] & 2) { /* EIDE */ + pio = adev->id[ATA_ID_EIDE_PIO]; + /* Is the speed faster than the drive allows non IORDY ? */ + if (pio) { + /* This is cycle times not frequency - watch the logic! */ + if (pio > 240) /* PIO2 is 240nS per cycle */ + return 1; + return 0; + } + } + return 0; +} + +/** * ata_dev_identify - obtain IDENTIFY x DEVICE page * @ap: port on which device we wish to probe resides * @device: device bus address, starting at zero @@ -1415,7 +1456,7 @@ void ata_dev_config(struct ata_port *ap, unsigned int i) ap->udma_mask &= ATA_UDMA5; ap->host->max_sectors = ATA_MAX_SECTORS; ap->host->hostt->max_sectors = ATA_MAX_SECTORS; - ap->device->flags |= ATA_DFLAG_LOCK_SECTORS; + ap->device[i].flags |= ATA_DFLAG_LOCK_SECTORS; } if (ap->ops->dev_config) @@ -3056,10 +3097,21 @@ static void ata_pio_data_xfer(struct ata_port *ap, unsigned char *buf, static void ata_data_xfer(struct ata_port *ap, unsigned char *buf, unsigned int buflen, int do_write) { - if (ap->flags & ATA_FLAG_MMIO) - ata_mmio_data_xfer(ap, buf, buflen, do_write); - else - ata_pio_data_xfer(ap, buf, buflen, do_write); + /* Make the crap hardware pay the costs not the good stuff */ + if (unlikely(ap->flags & ATA_FLAG_IRQ_MASK)) { + unsigned long flags; + local_irq_save(flags); + if (ap->flags & ATA_FLAG_MMIO) + ata_mmio_data_xfer(ap, buf, buflen, do_write); + else + ata_pio_data_xfer(ap, buf, buflen, do_write); + local_irq_restore(flags); + } else { + if (ap->flags & ATA_FLAG_MMIO) + ata_mmio_data_xfer(ap, buf, buflen, do_write); + else + ata_pio_data_xfer(ap, buf, buflen, do_write); + } } /** @@ -5122,6 +5174,7 @@ EXPORT_SYMBOL_GPL(ata_dev_id_string); EXPORT_SYMBOL_GPL(ata_dev_config); EXPORT_SYMBOL_GPL(ata_scsi_simulate); +EXPORT_SYMBOL_GPL(ata_pio_need_iordy); EXPORT_SYMBOL_GPL(ata_timing_compute); EXPORT_SYMBOL_GPL(ata_timing_merge); diff --git a/drivers/scsi/sata_promise.c b/drivers/scsi/sata_promise.c index 3d1ea09a06a..b0b0a69b356 100644 --- a/drivers/scsi/sata_promise.c +++ b/drivers/scsi/sata_promise.c @@ -66,6 +66,7 @@ enum { board_2037x = 0, /* FastTrak S150 TX2plus */ board_20319 = 1, /* FastTrak S150 TX4 */ board_20619 = 2, /* FastTrak TX4000 */ + board_20771 = 3, /* FastTrak TX2300 */ PDC_HAS_PATA = (1 << 1), /* PDC20375 has PATA */ @@ -190,6 +191,16 @@ static const struct ata_port_info pdc_port_info[] = { .udma_mask = 0x7f, /* udma0-6 ; FIXME */ .port_ops = &pdc_pata_ops, }, + + /* board_20771 */ + { + .sht = &pdc_ata_sht, + .host_flags = PDC_COMMON_FLAGS | ATA_FLAG_SATA, + .pio_mask = 0x1f, /* pio0-4 */ + .mwdma_mask = 0x07, /* mwdma0-2 */ + .udma_mask = 0x7f, /* udma0-6 ; FIXME */ + .port_ops = &pdc_sata_ops, + }, }; static const struct pci_device_id pdc_ata_pci_tbl[] = { @@ -226,6 +237,8 @@ static const struct pci_device_id pdc_ata_pci_tbl[] = { { PCI_VENDOR_ID_PROMISE, 0x6629, PCI_ANY_ID, PCI_ANY_ID, 0, 0, board_20619 }, + { PCI_VENDOR_ID_PROMISE, 0x3570, PCI_ANY_ID, PCI_ANY_ID, 0, 0, + board_20771 }, { } /* terminate list */ }; @@ -706,6 +719,9 @@ static int pdc_ata_init_one (struct pci_dev *pdev, const struct pci_device_id *e case board_2037x: probe_ent->n_ports = 2; break; + case board_20771: + probe_ent->n_ports = 2; + break; case board_20619: probe_ent->n_ports = 4; diff --git a/drivers/scsi/sata_svw.c b/drivers/scsi/sata_svw.c index 668373590aa..d8472563fde 100644 --- a/drivers/scsi/sata_svw.c +++ b/drivers/scsi/sata_svw.c @@ -470,6 +470,7 @@ static const struct pci_device_id k2_sata_pci_tbl[] = { { 0x1166, 0x0241, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 4 }, { 0x1166, 0x0242, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 8 }, { 0x1166, 0x024a, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 4 }, + { 0x1166, 0x024b, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 4 }, { } }; diff --git a/drivers/serial/8250.c b/drivers/serial/8250.c index d9ce8c54941..bc36edff205 100644 --- a/drivers/serial/8250.c +++ b/drivers/serial/8250.c @@ -2595,15 +2595,11 @@ static int __init serial8250_init(void) if (ret) goto out; - ret = platform_driver_register(&serial8250_isa_driver); - if (ret) - goto unreg_uart_drv; - serial8250_isa_devs = platform_device_alloc("serial8250", PLAT8250_DEV_LEGACY); if (!serial8250_isa_devs) { ret = -ENOMEM; - goto unreg_plat_drv; + goto unreg_uart_drv; } ret = platform_device_add(serial8250_isa_devs); @@ -2612,12 +2608,13 @@ static int __init serial8250_init(void) serial8250_register_ports(&serial8250_reg, &serial8250_isa_devs->dev); - goto out; + ret = platform_driver_register(&serial8250_isa_driver); + if (ret == 0) + goto out; + platform_device_del(serial8250_isa_devs); put_dev: platform_device_put(serial8250_isa_devs); - unreg_plat_drv: - platform_driver_unregister(&serial8250_isa_driver); unreg_uart_drv: uart_unregister_driver(&serial8250_reg); out: diff --git a/drivers/serial/8250_pci.c b/drivers/serial/8250_pci.c index 589fb076654..2a912153321 100644 --- a/drivers/serial/8250_pci.c +++ b/drivers/serial/8250_pci.c @@ -940,6 +940,7 @@ enum pci_board_num_t { pbn_b2_bt_2_921600, pbn_b2_bt_4_921600, + pbn_b3_2_115200, pbn_b3_4_115200, pbn_b3_8_115200, @@ -1311,6 +1312,12 @@ static struct pciserial_board pci_boards[] __devinitdata = { .uart_offset = 8, }, + [pbn_b3_2_115200] = { + .flags = FL_BASE3, + .num_ports = 2, + .base_baud = 115200, + .uart_offset = 8, + }, [pbn_b3_4_115200] = { .flags = FL_BASE3, .num_ports = 4, @@ -2272,6 +2279,9 @@ static struct pci_device_id serial_pci_tbl[] = { PCI_ANY_ID, PCI_ANY_ID, 0, 0, pbn_nec_nile4 }, + { PCI_VENDOR_ID_DCI, PCI_DEVICE_ID_DCI_PCCOM2, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, + pbn_b3_2_115200 }, { PCI_VENDOR_ID_DCI, PCI_DEVICE_ID_DCI_PCCOM4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, pbn_b3_4_115200 }, diff --git a/drivers/serial/Kconfig b/drivers/serial/Kconfig index 5e7199f7b59..9fd1925de36 100644 --- a/drivers/serial/Kconfig +++ b/drivers/serial/Kconfig @@ -301,7 +301,7 @@ config SERIAL_AT91_TTYAT depends on SERIAL_AT91=y help Say Y here if you wish to have the five internal AT91RM9200 UARTs - appear as /dev/ttyAT0-4 (major 240, minor 0-4) instead of the + appear as /dev/ttyAT0-4 (major 204, minor 154-158) instead of the normal /dev/ttyS0-4 (major 4, minor 64-68). This is necessary if you also want other UARTs, such as external 8250/16C550 compatible UARTs. diff --git a/drivers/serial/at91_serial.c b/drivers/serial/at91_serial.c index 0e206063d68..2113feb75c3 100644 --- a/drivers/serial/at91_serial.c +++ b/drivers/serial/at91_serial.c @@ -222,8 +222,6 @@ static void at91_rx_chars(struct uart_port *port, struct pt_regs *regs) while (status & (AT91_US_RXRDY)) { ch = UART_GET_CHAR(port); - if (tty->flip.count >= TTY_FLIPBUF_SIZE) - goto ignore_char; port->icount.rx++; flg = TTY_NORMAL; diff --git a/drivers/serial/suncore.c b/drivers/serial/suncore.c index 5fc4a62173d..fa4ae94243c 100644 --- a/drivers/serial/suncore.c +++ b/drivers/serial/suncore.c @@ -34,6 +34,7 @@ sunserial_console_termios(struct console *con) char *mode_prop = "ttyX-mode"; char *cd_prop = "ttyX-ignore-cd"; char *dtr_prop = "ttyX-rts-dtr-off"; + char *ssp_console_modes_prop = "ssp-console-modes"; int baud, bits, stop, cflag; char parity; int carrier = 0; @@ -43,14 +44,39 @@ sunserial_console_termios(struct console *con) if (!serial_console) return; - if (serial_console == 1) { + switch (serial_console) { + case PROMDEV_OTTYA: mode_prop[3] = 'a'; cd_prop[3] = 'a'; dtr_prop[3] = 'a'; - } else { + break; + + case PROMDEV_OTTYB: mode_prop[3] = 'b'; cd_prop[3] = 'b'; dtr_prop[3] = 'b'; + break; + + case PROMDEV_ORSC: + + nd = prom_pathtoinode("rsc"); + if (!nd) { + strcpy(mode, "115200,8,n,1,-"); + goto no_options; + } + + if (!prom_node_has_property(nd, ssp_console_modes_prop)) { + strcpy(mode, "115200,8,n,1,-"); + goto no_options; + } + + memset(mode, 0, sizeof(mode)); + prom_getstring(nd, ssp_console_modes_prop, mode, sizeof(mode)); + goto no_options; + + default: + strcpy(mode, "9600,8,n,1,-"); + goto no_options; } topnd = prom_getchild(prom_root_node); @@ -110,6 +136,10 @@ no_options: case 9600: cflag |= B9600; break; case 19200: cflag |= B19200; break; case 38400: cflag |= B38400; break; + case 57600: cflag |= B57600; break; + case 115200: cflag |= B115200; break; + case 230400: cflag |= B230400; break; + case 460800: cflag |= B460800; break; default: baud = 9600; cflag |= B9600; break; } diff --git a/drivers/serial/sunsab.c b/drivers/serial/sunsab.c index 7e773ff76c6..8bcaebcc0ad 100644 --- a/drivers/serial/sunsab.c +++ b/drivers/serial/sunsab.c @@ -897,9 +897,6 @@ static int sunsab_console_setup(struct console *con, char *options) sunserial_console_termios(con); - /* Firmware console speed is limited to 150-->38400 baud so - * this hackish cflag thing is OK. - */ switch (con->cflag & CBAUD) { case B150: baud = 150; break; case B300: baud = 300; break; @@ -910,6 +907,10 @@ static int sunsab_console_setup(struct console *con, char *options) default: case B9600: baud = 9600; break; case B19200: baud = 19200; break; case B38400: baud = 38400; break; + case B57600: baud = 57600; break; + case B115200: baud = 115200; break; + case B230400: baud = 230400; break; + case B460800: baud = 460800; break; }; /* diff --git a/drivers/video/sbuslib.c b/drivers/video/sbuslib.c index 55e6e2d60d3..a4d7cc51ce0 100644 --- a/drivers/video/sbuslib.c +++ b/drivers/video/sbuslib.c @@ -199,8 +199,7 @@ struct fbcmap32 { #define FBIOPUTCMAP32 _IOW('F', 3, struct fbcmap32) #define FBIOGETCMAP32 _IOW('F', 4, struct fbcmap32) -static int fbiogetputcmap(struct file *file, struct fb_info *info, - unsigned int cmd, unsigned long arg) +static int fbiogetputcmap(struct fb_info *info, unsigned int cmd, unsigned long arg) { struct fbcmap32 __user *argp = (void __user *)arg; struct fbcmap __user *p = compat_alloc_user_space(sizeof(*p)); @@ -236,8 +235,7 @@ struct fbcursor32 { #define FBIOSCURSOR32 _IOW('F', 24, struct fbcursor32) #define FBIOGCURSOR32 _IOW('F', 25, struct fbcursor32) -static int fbiogscursor(struct file *file, struct fb_info *info, - unsigned long arg) +static int fbiogscursor(struct fb_info *info, unsigned long arg) { struct fbcursor __user *p = compat_alloc_user_space(sizeof(*p)); struct fbcursor32 __user *argp = (void __user *)arg; @@ -263,8 +261,7 @@ static int fbiogscursor(struct file *file, struct fb_info *info, return info->fbops->fb_ioctl(info, FBIOSCURSOR, (unsigned long)p); } -long sbusfb_compat_ioctl(struct fb_info *info, unsigned int cmd, - unsigned long arg) +int sbusfb_compat_ioctl(struct fb_info *info, unsigned int cmd, unsigned long arg) { switch (cmd) { case FBIOGTYPE: diff --git a/drivers/video/sbuslib.h b/drivers/video/sbuslib.h index f753939013e..492828c3fe8 100644 --- a/drivers/video/sbuslib.h +++ b/drivers/video/sbuslib.h @@ -20,7 +20,7 @@ extern int sbusfb_mmap_helper(struct sbus_mmap_map *map, int sbusfb_ioctl_helper(unsigned long cmd, unsigned long arg, struct fb_info *info, int type, int fb_depth, unsigned long fb_size); -long sbusfb_compat_ioctl(struct fb_info *info, unsigned int cmd, +int sbusfb_compat_ioctl(struct fb_info *info, unsigned int cmd, unsigned long arg); #endif /* _SBUSLIB_H */ |