diff options
Diffstat (limited to 'Documentation')
| -rw-r--r-- | Documentation/ABI/testing/sysfs-gpio | 26 | ||||
| -rw-r--r-- | Documentation/DMA-mapping.txt | 2 | ||||
| -rw-r--r-- | Documentation/cpusets.txt | 18 | ||||
| -rw-r--r-- | Documentation/feature-removal-schedule.txt | 8 | ||||
| -rw-r--r-- | Documentation/filesystems/Locking | 15 | ||||
| -rw-r--r-- | Documentation/filesystems/proc.txt | 2 | ||||
| -rw-r--r-- | Documentation/ioctl/cdrom.txt | 4 | ||||
| -rw-r--r-- | Documentation/networking/can.txt | 44 | ||||
| -rw-r--r-- | Documentation/networking/phonet.txt | 175 | ||||
| -rw-r--r-- | Documentation/sysctl/kernel.txt | 7 | ||||
| -rw-r--r-- | Documentation/usb/anchors.txt | 12 |
11 files changed, 284 insertions, 29 deletions
diff --git a/Documentation/ABI/testing/sysfs-gpio b/Documentation/ABI/testing/sysfs-gpio new file mode 100644 index 00000000000..8aab8092ad3 --- /dev/null +++ b/Documentation/ABI/testing/sysfs-gpio @@ -0,0 +1,26 @@ +What: /sys/class/gpio/ +Date: July 2008 +KernelVersion: 2.6.27 +Contact: David Brownell <dbrownell@users.sourceforge.net> +Description: + + As a Kconfig option, individual GPIO signals may be accessed from + userspace. GPIOs are only made available to userspace by an explicit + "export" operation. If a given GPIO is not claimed for use by + kernel code, it may be exported by userspace (and unexported later). + Kernel code may export it for complete or partial access. + + GPIOs are identified as they are inside the kernel, using integers in + the range 0..INT_MAX. See Documentation/gpio.txt for more information. + + /sys/class/gpio + /export ... asks the kernel to export a GPIO to userspace + /unexport ... to return a GPIO to the kernel + /gpioN ... for each exported GPIO #N + /value ... always readable, writes fail for input GPIOs + /direction ... r/w as: in, out (default low); write: high, low + /gpiochipN ... for each gpiochip; #N is its first GPIO + /base ... (r/o) same as N + /label ... (r/o) descriptive, not necessarily unique + /ngpio ... (r/o) number of GPIOs; numbered N to N + (ngpio - 1) + diff --git a/Documentation/DMA-mapping.txt b/Documentation/DMA-mapping.txt index b463ecd0c7c..c74fec8c235 100644 --- a/Documentation/DMA-mapping.txt +++ b/Documentation/DMA-mapping.txt @@ -740,7 +740,7 @@ failure can be determined by: dma_addr_t dma_handle; dma_handle = pci_map_single(pdev, addr, size, direction); - if (pci_dma_mapping_error(dma_handle)) { + if (pci_dma_mapping_error(pdev, dma_handle)) { /* * reduce current DMA mapping usage, * delay and try again later or diff --git a/Documentation/cpusets.txt b/Documentation/cpusets.txt index 1f5a924d1e5..47e568a9370 100644 --- a/Documentation/cpusets.txt +++ b/Documentation/cpusets.txt @@ -635,14 +635,16 @@ prior 'mems' setting, will not be moved. There is an exception to the above. If hotplug functionality is used to remove all the CPUs that are currently assigned to a cpuset, -then the kernel will automatically update the cpus_allowed of all -tasks attached to CPUs in that cpuset to allow all CPUs. When memory -hotplug functionality for removing Memory Nodes is available, a -similar exception is expected to apply there as well. In general, -the kernel prefers to violate cpuset placement, over starving a task -that has had all its allowed CPUs or Memory Nodes taken offline. User -code should reconfigure cpusets to only refer to online CPUs and Memory -Nodes when using hotplug to add or remove such resources. +then all the tasks in that cpuset will be moved to the nearest ancestor +with non-empty cpus. But the moving of some (or all) tasks might fail if +cpuset is bound with another cgroup subsystem which has some restrictions +on task attaching. In this failing case, those tasks will stay +in the original cpuset, and the kernel will automatically update +their cpus_allowed to allow all online CPUs. When memory hotplug +functionality for removing Memory Nodes is available, a similar exception +is expected to apply there as well. In general, the kernel prefers to +violate cpuset placement, over starving a task that has had all +its allowed CPUs or Memory Nodes taken offline. There is a second exception to the above. GFP_ATOMIC requests are kernel internal allocations that must be satisfied, immediately. diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index c93fcdec246..d0f22fac55d 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -340,3 +340,11 @@ Why: Accounting can now be enabled/disabled without kernel recompilation. controlled by a kernel/module/sysfs/sysctl parameter. Who: Krzysztof Piotr Oledzki <ole@ans.pl> +--------------------------- + +What: ide-scsi (BLK_DEV_IDESCSI) +When: 2.6.29 +Why: The 2.6 kernel supports direct writing to ide CD drives, which + eliminates the need for ide-scsi. The new method is more + efficient in every way. +Who: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> diff --git a/Documentation/filesystems/Locking b/Documentation/filesystems/Locking index 680fb566b92..8362860e21a 100644 --- a/Documentation/filesystems/Locking +++ b/Documentation/filesystems/Locking @@ -144,8 +144,8 @@ prototypes: void (*kill_sb) (struct super_block *); locking rules: may block BKL -get_sb yes yes -kill_sb yes yes +get_sb yes no +kill_sb yes no ->get_sb() returns error or 0 with locked superblock attached to the vfsmount (exclusive on ->s_umount). @@ -409,12 +409,12 @@ ioctl: yes (see below) unlocked_ioctl: no (see below) compat_ioctl: no mmap: no -open: maybe (see below) +open: no flush: no release: no fsync: no (see below) aio_fsync: no -fasync: yes (see below) +fasync: no lock: yes readv: no writev: no @@ -431,13 +431,6 @@ For many filesystems, it is probably safe to acquire the inode semaphore. Note some filesystems (i.e. remote ones) provide no protection for i_size so you will need to use the BKL. -->open() locking is in-transit: big lock partially moved into the methods. -The only exception is ->open() in the instances of file_operations that never -end up in ->i_fop/->proc_fops, i.e. ones that belong to character devices -(chrdev_open() takes lock before replacing ->f_op and calling the secondary -method. As soon as we fix the handling of module reference counters all -instances of ->open() will be called without the BKL. - Note: ext2_release() was *the* source of contention on fs-intensive loads and dropping BKL on ->release() helps to get rid of that (we still grab BKL for cases when we close a file that had been opened r/w, but that diff --git a/Documentation/filesystems/proc.txt b/Documentation/filesystems/proc.txt index 394eb2cc1c3..f566ad9bcb7 100644 --- a/Documentation/filesystems/proc.txt +++ b/Documentation/filesystems/proc.txt @@ -2413,6 +2413,8 @@ The following 4 memory types are supported: - (bit 1) anonymous shared memory - (bit 2) file-backed private memory - (bit 3) file-backed shared memory + - (bit 4) ELF header pages in file-backed private memory areas (it is + effective only if the bit 2 is cleared) Note that MMIO pages such as frame buffer are never dumped and vDSO pages are always dumped regardless of the bitmask status. diff --git a/Documentation/ioctl/cdrom.txt b/Documentation/ioctl/cdrom.txt index 62d4af44ec4..59df81c8da2 100644 --- a/Documentation/ioctl/cdrom.txt +++ b/Documentation/ioctl/cdrom.txt @@ -271,14 +271,14 @@ CDROMCLOSETRAY pendant of CDROMEJECT usage: - ioctl(fd, CDROMEJECT, 0); + ioctl(fd, CDROMCLOSETRAY, 0); inputs: none outputs: none error returns: - ENOSYS cd drive not capable of ejecting + ENOSYS cd drive not capable of closing the tray EBUSY other processes are accessing drive, or door is locked notes: diff --git a/Documentation/networking/can.txt b/Documentation/networking/can.txt index 297ba7b1cca..2035bc4932f 100644 --- a/Documentation/networking/can.txt +++ b/Documentation/networking/can.txt @@ -35,8 +35,9 @@ This file contains 6.1 general settings 6.2 local loopback of sent frames 6.3 CAN controller hardware filters - 6.4 currently supported CAN hardware - 6.5 todo + 6.4 The virtual CAN driver (vcan) + 6.5 currently supported CAN hardware + 6.6 todo 7 Credits @@ -584,7 +585,42 @@ solution for a couple of reasons: @133MHz with four SJA1000 CAN controllers from 2002 under heavy bus load without any problems ... - 6.4 currently supported CAN hardware (September 2007) + 6.4 The virtual CAN driver (vcan) + + Similar to the network loopback devices, vcan offers a virtual local + CAN interface. A full qualified address on CAN consists of + + - a unique CAN Identifier (CAN ID) + - the CAN bus this CAN ID is transmitted on (e.g. can0) + + so in common use cases more than one virtual CAN interface is needed. + + The virtual CAN interfaces allow the transmission and reception of CAN + frames without real CAN controller hardware. Virtual CAN network + devices are usually named 'vcanX', like vcan0 vcan1 vcan2 ... + When compiled as a module the virtual CAN driver module is called vcan.ko + + Since Linux Kernel version 2.6.24 the vcan driver supports the Kernel + netlink interface to create vcan network devices. The creation and + removal of vcan network devices can be managed with the ip(8) tool: + + - Create a virtual CAN network interface: + ip link add type vcan + + - Create a virtual CAN network interface with a specific name 'vcan42': + ip link add dev vcan42 type vcan + + - Remove a (virtual CAN) network interface 'vcan42': + ip link del vcan42 + + The tool 'vcan' from the SocketCAN SVN repository on BerliOS is obsolete. + + Virtual CAN network device creation in older Kernels: + In Linux Kernel versions < 2.6.24 the vcan driver creates 4 vcan + netdevices at module load time by default. This value can be changed + with the module parameter 'numdev'. E.g. 'modprobe vcan numdev=8' + + 6.5 currently supported CAN hardware On the project website http://developer.berlios.de/projects/socketcan there are different drivers available: @@ -603,7 +639,7 @@ solution for a couple of reasons: Please check the Mailing Lists on the berlios OSS project website. - 6.5 todo (September 2007) + 6.6 todo The configuration interface for CAN network drivers is still an open issue that has not been finalized in the socketcan project. Also the diff --git a/Documentation/networking/phonet.txt b/Documentation/networking/phonet.txt new file mode 100644 index 00000000000..0e6e592f4f5 --- /dev/null +++ b/Documentation/networking/phonet.txt @@ -0,0 +1,175 @@ +Linux Phonet protocol family +============================ + +Introduction +------------ + +Phonet is a packet protocol used by Nokia cellular modems for both IPC +and RPC. With the Linux Phonet socket family, Linux host processes can +receive and send messages from/to the modem, or any other external +device attached to the modem. The modem takes care of routing. + +Phonet packets can be exchanged through various hardware connections +depending on the device, such as: + - USB with the CDC Phonet interface, + - infrared, + - Bluetooth, + - an RS232 serial port (with a dedicated "FBUS" line discipline), + - the SSI bus with some TI OMAP processors. + + +Packets format +-------------- + +Phonet packets have a common header as follows: + + struct phonethdr { + uint8_t pn_media; /* Media type (link-layer identifier) */ + uint8_t pn_rdev; /* Receiver device ID */ + uint8_t pn_sdev; /* Sender device ID */ + uint8_t pn_res; /* Resource ID or function */ + uint16_t pn_length; /* Big-endian message byte length (minus 6) */ + uint8_t pn_robj; /* Receiver object ID */ + uint8_t pn_sobj; /* Sender object ID */ + }; + +On Linux, the link-layer header includes the pn_media byte (see below). +The next 7 bytes are part of the network-layer header. + +The device ID is split: the 6 higher-order bits consitute the device +address, while the 2 lower-order bits are used for multiplexing, as are +the 8-bit object identifiers. As such, Phonet can be considered as a +network layer with 6 bits of address space and 10 bits for transport +protocol (much like port numbers in IP world). + +The modem always has address number zero. All other device have a their +own 6-bit address. + + +Link layer +---------- + +Phonet links are always point-to-point links. The link layer header +consists of a single Phonet media type byte. It uniquely identifies the +link through which the packet is transmitted, from the modem's +perspective. Each Phonet network device shall prepend and set the media +type byte as appropriate. For convenience, a common phonet_header_ops +link-layer header operations structure is provided. It sets the +media type according to the network device hardware address. + +Linux Phonet network interfaces support a dedicated link layer packets +type (ETH_P_PHONET) which is out of the Ethernet type range. They can +only send and receive Phonet packets. + +The virtual TUN tunnel device driver can also be used for Phonet. This +requires IFF_TUN mode, _without_ the IFF_NO_PI flag. In this case, +there is no link-layer header, so there is no Phonet media type byte. + +Note that Phonet interfaces are not allowed to re-order packets, so +only the (default) Linux FIFO qdisc should be used with them. + + +Network layer +------------- + +The Phonet socket address family maps the Phonet packet header: + + struct sockaddr_pn { + sa_family_t spn_family; /* AF_PHONET */ + uint8_t spn_obj; /* Object ID */ + uint8_t spn_dev; /* Device ID */ + uint8_t spn_resource; /* Resource or function */ + uint8_t spn_zero[...]; /* Padding */ + }; + +The resource field is only used when sending and receiving; +It is ignored by bind() and getsockname(). + + +Low-level datagram protocol +--------------------------- + +Applications can send Phonet messages using the Phonet datagram socket +protocol from the PF_PHONET family. Each socket is bound to one of the +2^10 object IDs available, and can send and receive packets with any +other peer. + + struct sockaddr_pn addr = { .spn_family = AF_PHONET, }; + ssize_t len; + socklen_t addrlen = sizeof(addr); + int fd; + + fd = socket(PF_PHONET, SOCK_DGRAM, 0); + bind(fd, (struct sockaddr *)&addr, sizeof(addr)); + /* ... */ + + sendto(fd, msg, msglen, 0, (struct sockaddr *)&addr, sizeof(addr)); + len = recvfrom(fd, buf, sizeof(buf), 0, + (struct sockaddr *)&addr, &addrlen); + +This protocol follows the SOCK_DGRAM connection-less semantics. +However, connect() and getpeername() are not supported, as they did +not seem useful with Phonet usages (could be added easily). + + +Phonet Pipe protocol +-------------------- + +The Phonet Pipe protocol is a simple sequenced packets protocol +with end-to-end congestion control. It uses the passive listening +socket paradigm. The listening socket is bound to an unique free object +ID. Each listening socket can handle up to 255 simultaneous +connections, one per accept()'d socket. + + int lfd, cfd; + + lfd = socket(PF_PHONET, SOCK_SEQPACKET, PN_PROTO_PIPE); + listen (lfd, INT_MAX); + + /* ... */ + cfd = accept(lfd, NULL, NULL); + for (;;) + { + char buf[...]; + ssize_t len = read(cfd, buf, sizeof(buf)); + + /* ... */ + + write(cfd, msg, msglen); + } + +Connections are established between two endpoints by a "third party" +application. This means that both endpoints are passive; so connect() +is not possible. + +WARNING: +When polling a connected pipe socket for writability, there is an +intrinsic race condition whereby writability might be lost between the +polling and the writing system calls. In this case, the socket will +block until write because possible again, unless non-blocking mode +becomes enabled. + + +The pipe protocol provides two socket options at the SOL_PNPIPE level: + + PNPIPE_ENCAP accepts one integer value (int) of: + + PNPIPE_ENCAP_NONE: The socket operates normally (default). + + PNPIPE_ENCAP_IP: The socket is used as a backend for a virtual IP + interface. This requires CAP_NET_ADMIN capability. GPRS data + support on Nokia modems can use this. Note that the socket cannot + be reliably poll()'d or read() from while in this mode. + + PNPIPE_IFINDEX is a read-only integer value. It contains the + interface index of the network interface created by PNPIPE_ENCAP, + or zero if encapsulation is off. + + +Authors +------- + +Linux Phonet was initially written by Sakari Ailus. +Other contributors include Mikä Liljeberg, Andras Domokos, +Carlos Chinea and Rémi Denis-Courmont. +Copyright (C) 2008 Nokia Corporation. diff --git a/Documentation/sysctl/kernel.txt b/Documentation/sysctl/kernel.txt index 276a7e63782..e1ff0d920a5 100644 --- a/Documentation/sysctl/kernel.txt +++ b/Documentation/sysctl/kernel.txt @@ -351,9 +351,10 @@ kernel. This value defaults to SHMMAX. softlockup_thresh: -This value can be used to lower the softlockup tolerance -threshold. The default threshold is 10s. If a cpu is locked up -for 10s, the kernel complains. Valid values are 1-60s. +This value can be used to lower the softlockup tolerance threshold. The +default threshold is 60 seconds. If a cpu is locked up for 60 seconds, +the kernel complains. Valid values are 1-60 seconds. Setting this +tunable to zero will disable the softlockup detection altogether. ============================================================== diff --git a/Documentation/usb/anchors.txt b/Documentation/usb/anchors.txt index 7304bcf5a30..5e6b64c20d2 100644 --- a/Documentation/usb/anchors.txt +++ b/Documentation/usb/anchors.txt @@ -42,9 +42,21 @@ This function kills all URBs associated with an anchor. The URBs are called in the reverse temporal order they were submitted. This way no data can be reordered. +usb_unlink_anchored_urbs() +-------------------------- + +This function unlinks all URBs associated with an anchor. The URBs +are processed in the reverse temporal order they were submitted. +This is similar to usb_kill_anchored_urbs(), but it will not sleep. +Therefore no guarantee is made that the URBs have been unlinked when +the call returns. They may be unlinked later but will be unlinked in +finite time. + usb_wait_anchor_empty_timeout() ------------------------------- This function waits for all URBs associated with an anchor to finish or a timeout, whichever comes first. Its return value will tell you whether the timeout was reached. + + |