diff options
Diffstat (limited to 'arch/mips/mm')
-rw-r--r-- | arch/mips/mm/Makefile | 1 | ||||
-rw-r--r-- | arch/mips/mm/c-octeon.c | 307 | ||||
-rw-r--r-- | arch/mips/mm/c-r4k.c | 22 | ||||
-rw-r--r-- | arch/mips/mm/cache.c | 6 | ||||
-rw-r--r-- | arch/mips/mm/cex-oct.S | 70 | ||||
-rw-r--r-- | arch/mips/mm/dma-default.c | 25 | ||||
-rw-r--r-- | arch/mips/mm/fault.c | 21 | ||||
-rw-r--r-- | arch/mips/mm/tlb-r4k.c | 5 | ||||
-rw-r--r-- | arch/mips/mm/tlbex.c | 1 |
9 files changed, 426 insertions, 32 deletions
diff --git a/arch/mips/mm/Makefile b/arch/mips/mm/Makefile index 95ba32b5b72..d7ec9552229 100644 --- a/arch/mips/mm/Makefile +++ b/arch/mips/mm/Makefile @@ -27,6 +27,7 @@ obj-$(CONFIG_CPU_SB1) += c-r4k.o cerr-sb1.o cex-sb1.o tlb-r4k.o obj-$(CONFIG_CPU_TX39XX) += c-tx39.o tlb-r3k.o obj-$(CONFIG_CPU_TX49XX) += c-r4k.o cex-gen.o tlb-r4k.o obj-$(CONFIG_CPU_VR41XX) += c-r4k.o cex-gen.o tlb-r4k.o +obj-$(CONFIG_CPU_CAVIUM_OCTEON) += c-octeon.o cex-oct.o tlb-r4k.o obj-$(CONFIG_IP22_CPU_SCACHE) += sc-ip22.o obj-$(CONFIG_R5000_CPU_SCACHE) += sc-r5k.o diff --git a/arch/mips/mm/c-octeon.c b/arch/mips/mm/c-octeon.c new file mode 100644 index 00000000000..44d01a0a849 --- /dev/null +++ b/arch/mips/mm/c-octeon.c @@ -0,0 +1,307 @@ +/* + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + * + * Copyright (C) 2005-2007 Cavium Networks + */ +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/bitops.h> +#include <linux/cpu.h> +#include <linux/io.h> + +#include <asm/bcache.h> +#include <asm/bootinfo.h> +#include <asm/cacheops.h> +#include <asm/cpu-features.h> +#include <asm/page.h> +#include <asm/pgtable.h> +#include <asm/r4kcache.h> +#include <asm/system.h> +#include <asm/mmu_context.h> +#include <asm/war.h> + +#include <asm/octeon/octeon.h> + +unsigned long long cache_err_dcache[NR_CPUS]; + +/** + * Octeon automatically flushes the dcache on tlb changes, so + * from Linux's viewpoint it acts much like a physically + * tagged cache. No flushing is needed + * + */ +static void octeon_flush_data_cache_page(unsigned long addr) +{ + /* Nothing to do */ +} + +static inline void octeon_local_flush_icache(void) +{ + asm volatile ("synci 0($0)"); +} + +/* + * Flush local I-cache for the specified range. + */ +static void local_octeon_flush_icache_range(unsigned long start, + unsigned long end) +{ + octeon_local_flush_icache(); +} + +/** + * Flush caches as necessary for all cores affected by a + * vma. If no vma is supplied, all cores are flushed. + * + * @vma: VMA to flush or NULL to flush all icaches. + */ +static void octeon_flush_icache_all_cores(struct vm_area_struct *vma) +{ + extern void octeon_send_ipi_single(int cpu, unsigned int action); +#ifdef CONFIG_SMP + int cpu; + cpumask_t mask; +#endif + + mb(); + octeon_local_flush_icache(); +#ifdef CONFIG_SMP + preempt_disable(); + cpu = smp_processor_id(); + + /* + * If we have a vma structure, we only need to worry about + * cores it has been used on + */ + if (vma) + mask = vma->vm_mm->cpu_vm_mask; + else + mask = cpu_online_map; + cpu_clear(cpu, mask); + for_each_cpu_mask(cpu, mask) + octeon_send_ipi_single(cpu, SMP_ICACHE_FLUSH); + + preempt_enable(); +#endif +} + + +/** + * Called to flush the icache on all cores + */ +static void octeon_flush_icache_all(void) +{ + octeon_flush_icache_all_cores(NULL); +} + + +/** + * Called to flush all memory associated with a memory + * context. + * + * @mm: Memory context to flush + */ +static void octeon_flush_cache_mm(struct mm_struct *mm) +{ + /* + * According to the R4K version of this file, CPUs without + * dcache aliases don't need to do anything here + */ +} + + +/** + * Flush a range of kernel addresses out of the icache + * + */ +static void octeon_flush_icache_range(unsigned long start, unsigned long end) +{ + octeon_flush_icache_all_cores(NULL); +} + + +/** + * Flush the icache for a trampoline. These are used for interrupt + * and exception hooking. + * + * @addr: Address to flush + */ +static void octeon_flush_cache_sigtramp(unsigned long addr) +{ + struct vm_area_struct *vma; + + vma = find_vma(current->mm, addr); + octeon_flush_icache_all_cores(vma); +} + + +/** + * Flush a range out of a vma + * + * @vma: VMA to flush + * @start: + * @end: + */ +static void octeon_flush_cache_range(struct vm_area_struct *vma, + unsigned long start, unsigned long end) +{ + if (vma->vm_flags & VM_EXEC) + octeon_flush_icache_all_cores(vma); +} + + +/** + * Flush a specific page of a vma + * + * @vma: VMA to flush page for + * @page: Page to flush + * @pfn: + */ +static void octeon_flush_cache_page(struct vm_area_struct *vma, + unsigned long page, unsigned long pfn) +{ + if (vma->vm_flags & VM_EXEC) + octeon_flush_icache_all_cores(vma); +} + + +/** + * Probe Octeon's caches + * + */ +static void __devinit probe_octeon(void) +{ + unsigned long icache_size; + unsigned long dcache_size; + unsigned int config1; + struct cpuinfo_mips *c = ¤t_cpu_data; + + switch (c->cputype) { + case CPU_CAVIUM_OCTEON: + config1 = read_c0_config1(); + c->icache.linesz = 2 << ((config1 >> 19) & 7); + c->icache.sets = 64 << ((config1 >> 22) & 7); + c->icache.ways = 1 + ((config1 >> 16) & 7); + c->icache.flags |= MIPS_CACHE_VTAG; + icache_size = + c->icache.sets * c->icache.ways * c->icache.linesz; + c->icache.waybit = ffs(icache_size / c->icache.ways) - 1; + c->dcache.linesz = 128; + if (OCTEON_IS_MODEL(OCTEON_CN3XXX)) + c->dcache.sets = 1; /* CN3XXX has one Dcache set */ + else + c->dcache.sets = 2; /* CN5XXX has two Dcache sets */ + c->dcache.ways = 64; + dcache_size = + c->dcache.sets * c->dcache.ways * c->dcache.linesz; + c->dcache.waybit = ffs(dcache_size / c->dcache.ways) - 1; + c->options |= MIPS_CPU_PREFETCH; + break; + + default: + panic("Unsupported Cavium Networks CPU type\n"); + break; + } + + /* compute a couple of other cache variables */ + c->icache.waysize = icache_size / c->icache.ways; + c->dcache.waysize = dcache_size / c->dcache.ways; + + c->icache.sets = icache_size / (c->icache.linesz * c->icache.ways); + c->dcache.sets = dcache_size / (c->dcache.linesz * c->dcache.ways); + + if (smp_processor_id() == 0) { + pr_notice("Primary instruction cache %ldkB, %s, %d way, " + "%d sets, linesize %d bytes.\n", + icache_size >> 10, + cpu_has_vtag_icache ? + "virtually tagged" : "physically tagged", + c->icache.ways, c->icache.sets, c->icache.linesz); + + pr_notice("Primary data cache %ldkB, %d-way, %d sets, " + "linesize %d bytes.\n", + dcache_size >> 10, c->dcache.ways, + c->dcache.sets, c->dcache.linesz); + } +} + + +/** + * Setup the Octeon cache flush routines + * + */ +void __devinit octeon_cache_init(void) +{ + extern unsigned long ebase; + extern char except_vec2_octeon; + + memcpy((void *)(ebase + 0x100), &except_vec2_octeon, 0x80); + octeon_flush_cache_sigtramp(ebase + 0x100); + + probe_octeon(); + + shm_align_mask = PAGE_SIZE - 1; + + flush_cache_all = octeon_flush_icache_all; + __flush_cache_all = octeon_flush_icache_all; + flush_cache_mm = octeon_flush_cache_mm; + flush_cache_page = octeon_flush_cache_page; + flush_cache_range = octeon_flush_cache_range; + flush_cache_sigtramp = octeon_flush_cache_sigtramp; + flush_icache_all = octeon_flush_icache_all; + flush_data_cache_page = octeon_flush_data_cache_page; + flush_icache_range = octeon_flush_icache_range; + local_flush_icache_range = local_octeon_flush_icache_range; + + build_clear_page(); + build_copy_page(); +} + +/** + * Handle a cache error exception + */ + +static void cache_parity_error_octeon(int non_recoverable) +{ + unsigned long coreid = cvmx_get_core_num(); + uint64_t icache_err = read_octeon_c0_icacheerr(); + + pr_err("Cache error exception:\n"); + pr_err("cp0_errorepc == %lx\n", read_c0_errorepc()); + if (icache_err & 1) { + pr_err("CacheErr (Icache) == %llx\n", + (unsigned long long)icache_err); + write_octeon_c0_icacheerr(0); + } + if (cache_err_dcache[coreid] & 1) { + pr_err("CacheErr (Dcache) == %llx\n", + (unsigned long long)cache_err_dcache[coreid]); + cache_err_dcache[coreid] = 0; + } + + if (non_recoverable) + panic("Can't handle cache error: nested exception"); +} + +/** + * Called when the the exception is not recoverable + */ + +asmlinkage void cache_parity_error_octeon_recoverable(void) +{ + cache_parity_error_octeon(0); +} + +/** + * Called when the the exception is recoverable + */ + +asmlinkage void cache_parity_error_octeon_non_recoverable(void) +{ + cache_parity_error_octeon(1); +} + diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c index 6e99665ae86..c43f4b26a69 100644 --- a/arch/mips/mm/c-r4k.c +++ b/arch/mips/mm/c-r4k.c @@ -618,15 +618,35 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size) if (cpu_has_inclusive_pcaches) { if (size >= scache_size) r4k_blast_scache(); - else + else { + unsigned long lsize = cpu_scache_line_size(); + unsigned long almask = ~(lsize - 1); + + /* + * There is no clearly documented alignment requirement + * for the cache instruction on MIPS processors and + * some processors, among them the RM5200 and RM7000 + * QED processors will throw an address error for cache + * hit ops with insufficient alignment. Solved by + * aligning the address to cache line size. + */ + cache_op(Hit_Writeback_Inv_SD, addr & almask); + cache_op(Hit_Writeback_Inv_SD, + (addr + size - 1) & almask); blast_inv_scache_range(addr, addr + size); + } return; } if (cpu_has_safe_index_cacheops && size >= dcache_size) { r4k_blast_dcache(); } else { + unsigned long lsize = cpu_dcache_line_size(); + unsigned long almask = ~(lsize - 1); + R4600_HIT_CACHEOP_WAR_IMPL; + cache_op(Hit_Writeback_Inv_D, addr & almask); + cache_op(Hit_Writeback_Inv_D, (addr + size - 1) & almask); blast_inv_dcache_range(addr, addr + size); } diff --git a/arch/mips/mm/cache.c b/arch/mips/mm/cache.c index 1eb7c71e3d6..98ad0a82c29 100644 --- a/arch/mips/mm/cache.c +++ b/arch/mips/mm/cache.c @@ -182,6 +182,12 @@ void __devinit cpu_cache_init(void) tx39_cache_init(); } + if (cpu_has_octeon_cache) { + extern void __weak octeon_cache_init(void); + + octeon_cache_init(); + } + setup_protection_map(); } diff --git a/arch/mips/mm/cex-oct.S b/arch/mips/mm/cex-oct.S new file mode 100644 index 00000000000..3db8553fcd3 --- /dev/null +++ b/arch/mips/mm/cex-oct.S @@ -0,0 +1,70 @@ +/* + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + * + * Copyright (C) 2006 Cavium Networks + * Cache error handler + */ + +#include <asm/asm.h> +#include <asm/regdef.h> +#include <asm/mipsregs.h> +#include <asm/stackframe.h> + +/* + * Handle cache error. Indicate to the second level handler whether + * the exception is recoverable. + */ + LEAF(except_vec2_octeon) + + .set push + .set mips64r2 + .set noreorder + .set noat + + + /* due to an errata we need to read the COP0 CacheErr (Dcache) + * before any cache/DRAM access */ + + rdhwr k0, $0 /* get core_id */ + PTR_LA k1, cache_err_dcache + sll k0, k0, 3 + PTR_ADDU k1, k0, k1 /* k1 = &cache_err_dcache[core_id] */ + + dmfc0 k0, CP0_CACHEERR, 1 + sd k0, (k1) + dmtc0 $0, CP0_CACHEERR, 1 + + /* check whether this is a nested exception */ + mfc0 k1, CP0_STATUS + andi k1, k1, ST0_EXL + beqz k1, 1f + nop + j cache_parity_error_octeon_non_recoverable + nop + + /* exception is recoverable */ +1: j handle_cache_err + nop + + .set pop + END(except_vec2_octeon) + + /* We need to jump to handle_cache_err so that the previous handler + * can fit within 0x80 bytes. We also move from 0xFFFFFFFFAXXXXXXX + * space (uncached) to the 0xFFFFFFFF8XXXXXXX space (cached). */ + LEAF(handle_cache_err) + .set push + .set noreorder + .set noat + + SAVE_ALL + KMODE + jal cache_parity_error_octeon_recoverable + nop + j ret_from_exception + nop + + .set pop + END(handle_cache_err) diff --git a/arch/mips/mm/dma-default.c b/arch/mips/mm/dma-default.c index e6708b3ad34..546e6977d4f 100644 --- a/arch/mips/mm/dma-default.c +++ b/arch/mips/mm/dma-default.c @@ -111,7 +111,7 @@ EXPORT_SYMBOL(dma_alloc_coherent); void dma_free_noncoherent(struct device *dev, size_t size, void *vaddr, dma_addr_t dma_handle) { - plat_unmap_dma_mem(dma_handle); + plat_unmap_dma_mem(dev, dma_handle); free_pages((unsigned long) vaddr, get_order(size)); } @@ -122,7 +122,7 @@ void dma_free_coherent(struct device *dev, size_t size, void *vaddr, { unsigned long addr = (unsigned long) vaddr; - plat_unmap_dma_mem(dma_handle); + plat_unmap_dma_mem(dev, dma_handle); if (!plat_device_is_coherent(dev)) addr = CAC_ADDR(addr); @@ -173,7 +173,7 @@ void dma_unmap_single(struct device *dev, dma_addr_t dma_addr, size_t size, __dma_sync(dma_addr_to_virt(dma_addr), size, direction); - plat_unmap_dma_mem(dma_addr); + plat_unmap_dma_mem(dev, dma_addr); } EXPORT_SYMBOL(dma_unmap_single); @@ -229,7 +229,7 @@ void dma_unmap_page(struct device *dev, dma_addr_t dma_address, size_t size, dma_cache_wback_inv(addr, size); } - plat_unmap_dma_mem(dma_address); + plat_unmap_dma_mem(dev, dma_address); } EXPORT_SYMBOL(dma_unmap_page); @@ -249,7 +249,7 @@ void dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nhwentries, if (addr) __dma_sync(addr, sg->length, direction); } - plat_unmap_dma_mem(sg->dma_address); + plat_unmap_dma_mem(dev, sg->dma_address); } } @@ -275,6 +275,7 @@ void dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, { BUG_ON(direction == DMA_NONE); + plat_extra_sync_for_device(dev); if (!plat_device_is_coherent(dev)) { unsigned long addr; @@ -305,6 +306,7 @@ void dma_sync_single_range_for_device(struct device *dev, dma_addr_t dma_handle, { BUG_ON(direction == DMA_NONE); + plat_extra_sync_for_device(dev); if (!plat_device_is_coherent(dev)) { unsigned long addr; @@ -351,22 +353,14 @@ EXPORT_SYMBOL(dma_sync_sg_for_device); int dma_mapping_error(struct device *dev, dma_addr_t dma_addr) { - return 0; + return plat_dma_mapping_error(dev, dma_addr); } EXPORT_SYMBOL(dma_mapping_error); int dma_supported(struct device *dev, u64 mask) { - /* - * we fall back to GFP_DMA when the mask isn't all 1s, - * so we can't guarantee allocations that must be - * within a tighter range than GFP_DMA.. - */ - if (mask < DMA_BIT_MASK(24)) - return 0; - - return 1; + return plat_dma_supported(dev, mask); } EXPORT_SYMBOL(dma_supported); @@ -383,6 +377,7 @@ void dma_cache_sync(struct device *dev, void *vaddr, size_t size, { BUG_ON(direction == DMA_NONE); + plat_extra_sync_for_device(dev); if (!plat_device_is_coherent(dev)) __dma_sync((unsigned long)vaddr, size, direction); } diff --git a/arch/mips/mm/fault.c b/arch/mips/mm/fault.c index fa636fc6b7b..55767ad9f00 100644 --- a/arch/mips/mm/fault.c +++ b/arch/mips/mm/fault.c @@ -97,7 +97,6 @@ good_area: goto bad_area; } -survive: /* * If for any reason at all we couldn't handle the fault, * make sure we exit gracefully rather than endlessly redo @@ -167,21 +166,13 @@ no_context: field, regs->regs[31]); die("Oops", regs); -/* - * We ran out of memory, or some other thing happened to us that made - * us unable to handle the page fault gracefully. - */ out_of_memory: - up_read(&mm->mmap_sem); - if (is_global_init(tsk)) { - yield(); - down_read(&mm->mmap_sem); - goto survive; - } - printk("VM: killing process %s\n", tsk->comm); - if (user_mode(regs)) - do_group_exit(SIGKILL); - goto no_context; + /* + * We ran out of memory, call the OOM killer, and return the userspace + * (which will retry the fault, or kill us if we got oom-killed). + */ + pagefault_out_of_memory(); + return; do_sigbus: up_read(&mm->mmap_sem); diff --git a/arch/mips/mm/tlb-r4k.c b/arch/mips/mm/tlb-r4k.c index 5ce2fa74562..9619f66e531 100644 --- a/arch/mips/mm/tlb-r4k.c +++ b/arch/mips/mm/tlb-r4k.c @@ -478,7 +478,10 @@ void __cpuinit tlb_init(void) probe_tlb(config); write_c0_pagemask(PM_DEFAULT_MASK); write_c0_wired(0); - write_c0_framemask(0); + if (current_cpu_type() == CPU_R10000 || + current_cpu_type() == CPU_R12000 || + current_cpu_type() == CPU_R14000) + write_c0_framemask(0); temp_tlb_entry = current_cpu_data.tlbsize - 1; /* From this point on the ARC firmware is dead. */ diff --git a/arch/mips/mm/tlbex.c b/arch/mips/mm/tlbex.c index 979cf919728..42942038d0f 100644 --- a/arch/mips/mm/tlbex.c +++ b/arch/mips/mm/tlbex.c @@ -317,6 +317,7 @@ static void __cpuinit build_tlb_write_entry(u32 **p, struct uasm_label **l, case CPU_BCM3302: case CPU_BCM4710: case CPU_LOONGSON2: + case CPU_CAVIUM_OCTEON: if (m4kc_tlbp_war()) uasm_i_nop(p); tlbw(p); |