GitHub/LineageOS/G12/android_kernel_amlogic_linux-4.9.git
16 years agox86: use bitmap library for pin_programmed
Akinobu Mita [Sat, 19 Apr 2008 14:55:13 +0000 (23:55 +0900)]
x86: use bitmap library for pin_programmed

Use bitmap library for pin_programmed rather than reinvent
bitmaps.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: use MP_intsrc_info()
Akinobu Mita [Sat, 19 Apr 2008 14:55:12 +0000 (23:55 +0900)]
x86: use MP_intsrc_info()

Remove duplicate code by using MP_intsrc_info() in mpparse.c

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: use BUILD_BUG_ON() for the size of struct intel_mp_floating
Akinobu Mita [Sat, 19 Apr 2008 14:55:11 +0000 (23:55 +0900)]
x86: use BUILD_BUG_ON() for the size of struct intel_mp_floating

Use BUILD_BUG_ON() instead of compile-time error technique with
extern non-exsistent function.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86_64 ia32 ptrace: convert to compat_arch_ptrace
Roland McGrath [Tue, 22 Apr 2008 19:21:25 +0000 (12:21 -0700)]
x86_64 ia32 ptrace: convert to compat_arch_ptrace

Now that there are no more special cases in sys32_ptrace, we
can convert to using the generic compat_sys_ptrace entry point.
The sys32_ptrace function gets simpler and becomes compat_arch_ptrace.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86_64 ia32 ptrace: use compat_ptrace_request for siginfo
Roland McGrath [Tue, 22 Apr 2008 19:20:20 +0000 (12:20 -0700)]
x86_64 ia32 ptrace: use compat_ptrace_request for siginfo

This removes the special-case handling for PTRACE_GETSIGINFO
and PTRACE_SETSIGINFO from x86_64's sys32_ptrace.  The generic
compat_ptrace_request code handles these.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86 signals: lift set_fs
Roland McGrath [Sat, 19 Apr 2008 21:27:56 +0000 (14:27 -0700)]
x86 signals: lift set_fs

This lifts the set_fs(USER_DS) call for signal handler setup out of the
three places copying the same code into the one place that calls them
all.  There is no change in what it does.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86 signals: lift flags diddling code
Roland McGrath [Sat, 19 Apr 2008 21:26:54 +0000 (14:26 -0700)]
x86 signals: lift flags diddling code

This lifts the code diddling the TF and DF bits for signal handler setup
out of the several places copying the same code into the one place that
calls them all.  There is no change in what it does.

I also separated the recently-added DF bit clearing from the TF diddling.
The compiler turns them back into one instruction anyway.  The tossing
in of DF to the same line of code with no new comments was a bit more
arcane than seems wise.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: remove NexGen support
Dmitri Vorobiev [Sun, 20 Apr 2008 20:47:55 +0000 (00:47 +0400)]
x86: remove NexGen support

It is claimed that NexGen CPUs were never shipped:

   http://lkml.org/lkml/2008/4/20/179

Also, the kernel support for these chips has been broken for
a long time, the code intended to support NexGen thereby being
essentially dead.

As an outcome of the discussion that can be found using the URL
above, this patch removes the NexGen support altogether.

The changes in this patch survived a defconfig build for i386, a
couple of successful randconfig builds, as well as a runtime test,
which consisted in booting a 32-bit x86 box up to the shell prompt.

Signed-off-by: Dmitri Vorobiev <dmitri.vorobiev@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: array can become static
Dmitri Vorobiev [Sun, 20 Apr 2008 02:54:33 +0000 (06:54 +0400)]
x86: array can become static

In arch/x86/kernel/setup_64.c, the standard_io_resources array
is needlessly defined as global. This patch makes this variable
static.

This patch was successfully build-tested using the defconfig
for x86_64. Runtime test was performed by booting a 64-bit x86
box up to the shell prompt.

Signed-off-by: Dmitri Vorobiev <dmitri.vorobiev@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: remove unused function amd_init_cpu()
Dmitri Vorobiev [Sun, 20 Apr 2008 02:54:31 +0000 (06:54 +0400)]
x86: remove unused function amd_init_cpu()

There are no users for the function amd_init_cpu() defined in
arch/x86/kernel/cpu/amd.c. This patch removes this routine.

This patch was build-tested using defconfigs for i386 and x86_64,
and a few randconfig instances. Runtime tests were performed by
booting 32- and 64-bit x86 boxen up to the shell prompt.

Signed-off-by: Dmitri Vorobiev <dmitri.vorobiev@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86-64: extend MCE CPU quirk handling
Jan Beulich [Tue, 22 Apr 2008 15:22:21 +0000 (16:22 +0100)]
x86-64: extend MCE CPU quirk handling

At least on my Barcelona, I see MCE log entries after cold boot caused
by BIOS not properly clearing the respective registers. Therefore, this
patch extends the workaround to families 0x10 and 0x11 (the latter just
for completeness, I have nothing to verify this against).
At the same time, provide a way to make these entries visible via the
'mce=bootlog' command line option even on these machines.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agoi386: fix signal type for iret exception
Jan Beulich [Tue, 22 Apr 2008 15:19:25 +0000 (16:19 +0100)]
i386: fix signal type for iret exception

.. since it uses ILL_BADSTK (which is meaningless in the context of
SIGSEGV).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: fix watchdog ops for CoreDuo
Jan Beulich [Tue, 22 Apr 2008 15:28:41 +0000 (16:28 +0100)]
x86: fix watchdog ops for CoreDuo

There apparently was an unnoticed conflict between an earlier patch to
this file and mine (d1e084746b0e5806e6345ab31c5b370f8dee2b23), which
I noticed only now. I suppose a change like the one below (untested) is
needed; I didn't get any response on a confirmation request for this from
the submitter of the first patch.

The issue is the writing of the 'checkbit' member at the end of
setup_intel_arch_watchdog(), which my patch made go to intel_arch_wd_ops
rather than wd_ops.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agofix asm-x86/{posix_types,unistd}.h
Adrian Bunk [Sun, 20 Apr 2008 19:02:17 +0000 (22:02 +0300)]
fix asm-x86/{posix_types,unistd}.h

Jeff Chua reported that:

Commit e40c0fe6b0b5dd16aec3c0dad311d36b19d78fd9
(x86: cleanup duplicate includes) turned the userspace
asm-x86/posix_types.h and asm-x86/unistd.h headers into
empty files.

This patch reverts these bogus changes.

Reported-by: Jeff Chua <jeff.chua.linux@gmail.com>
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agoi386: fix asm constraint in do_IRQ()
Jan Beulich [Tue, 22 Apr 2008 15:16:50 +0000 (16:16 +0100)]
i386: fix asm constraint in do_IRQ()

Two prior changes resulted in the "ecx" clobber being lost.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agov4l/dvb: add statics to avoid multiple definitions
Harvey Harrison [Sat, 26 Apr 2008 02:03:08 +0000 (19:03 -0700)]
v4l/dvb: add statics to avoid multiple definitions

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
Linus Torvalds [Sat, 26 Apr 2008 02:05:55 +0000 (19:05 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  [PATCH] sanitize locate_fd()
  [PATCH] sanitize unshare_files/reset_files_struct
  [PATCH] sanitize handling of shared descriptor tables in failing execve()
  [PATCH] close race in unshare_files()
  [PATCH] restore sane ->umount_begin() API
  cifs: timeout dfs automounts +little fix.

16 years agomac-hid: fix build after lockdep annotation
Harvey Harrison [Sat, 26 Apr 2008 01:40:01 +0000 (18:40 -0700)]
mac-hid: fix build after lockdep annotation

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodocumentation: remove smart-config.txt
Randy Dunlap [Fri, 25 Apr 2008 23:38:44 +0000 (16:38 -0700)]
documentation: remove smart-config.txt

As requested by Sam Ravnborg: Remove Documentation/smart-config.txt.

It is outdated and has been (functionally) replaced by
Documentation/kbuild/*.txt.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoExport __locks_copy_lock() so modular lockd builds
Roland Dreier [Fri, 25 Apr 2008 22:32:51 +0000 (15:32 -0700)]
Export __locks_copy_lock() so modular lockd builds

Commit 1a747ee0 ("locks: don't call ->copy_lock methods on return of
conflicting locks") changed fs/lockd/svclock.c to call
__locks_copy_lock() instead of locks_copy_lock(), but lockd can be built
as a module and __locks_copy_lock() is not exported, which causes a
build error

    ERROR: "__locks_copy_lock" [fs/lockd/lockd.ko] undefined!

with CONFIG_LOCKD=m.

Fix this by exporting __locks_copy_lock().

Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Fri, 25 Apr 2008 19:52:16 +0000 (12:52 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/paulus/powerpc

* 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc: (49 commits)
  [POWERPC] Add zImage.iseries to arch/powerpc/boot/.gitignore
  [POWERPC] bootwrapper: fix build error on virtex405-head.S
  [POWERPC] 4xx: Fix 460GT support to not enable FPU
  [POWERPC] 4xx: Add NOR FLASH entries to Canyonlands and Glacier dts
  [POWERPC] Xilinx: of_serial support for Xilinx uart 16550.
  [POWERPC] Xilinx: boot support for Xilinx uart 16550.
  [POWERPC] celleb: Add support for PCI Express
  [POWERPC] celleb: Move miscellaneous files for Beat
  [POWERPC] celleb: Move a file for SPU on Beat
  [POWERPC] celleb: Move files for Beat mmu and iommu
  [POWERPC] celleb: Move files for Beat hvcall interfaces
  [POWERPC] celleb: Move the SCC related code for celleb
  [POWERPC] celleb: Move the files for celleb base support
  [POWERPC] celleb: Consolidate io-workarounds code
  [POWERPC] cell: Generalize io-workarounds code
  [POWERPC] Add CONFIG_PPC_PSERIES_DEBUG to enable debugging for platforms/pseries
  [POWERPC] Convert from DBG() to pr_debug() in platforms/pseries/
  [POWERPC] Register udbg console early on pseries LPAR
  [POWERPC] Mark udbg console as CON_ANYTIME, ie. callable early in boot
  [POWERPC] Set udbg_console index to 0
  ...

16 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Fri, 25 Apr 2008 19:50:00 +0000 (12:50 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [PATCH] Build fix for CONFIG_NUMA=y && CONFIG_SMP=n
  [IA64] fix bootmem regression on Altix

16 years agoMerge branch 'for-linus' of git://git.o-hand.com/linux-rpurdie-leds
Linus Torvalds [Fri, 25 Apr 2008 19:48:44 +0000 (12:48 -0700)]
Merge branch 'for-linus' of git://git.o-hand.com/linux-rpurdie-leds

* 'for-linus' of git://git.o-hand.com/linux-rpurdie-leds:
  leds: Add default-on trigger
  leds: Document the context brightness_set needs
  leds: Add new driver for the LEDs on the Freecom FSG-3
  leds: Add support to leds with readable status
  leds: enable support for blink_set() platform hook in leds-gpio
  leds: Cleanup various whitespace and code style issues
  leds: disable triggers on brightness set
  leds: Add mail LED support for "Clevo D400P"

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux...
Linus Torvalds [Fri, 25 Apr 2008 19:48:08 +0000 (12:48 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/x86/linux-2.6-x86-pat

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86-pat:
  generic: add ioremap_wc() interface wrapper
  /dev/mem: make promisc the default
  pat: cleanups
  x86: PAT use reserve free memtype in mmap of /dev/mem
  x86: PAT phys_mem_access_prot_allowed for dev/mem mmap
  x86: PAT avoid aliasing in /dev/mem read/write
  devmem: add range_is_allowed() check to mmap of /dev/mem
  x86: introduce /dev/mem restrictions with a config option

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux...
Linus Torvalds [Fri, 25 Apr 2008 19:47:56 +0000 (12:47 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mingo/linux-2.6-sched-fixes

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched-fixes:
  sched: fix share (re)distribution
  softlockup: fix NOHZ wakeup
  seqlock: livelock fix

16 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzi...
Linus Torvalds [Fri, 25 Apr 2008 19:41:55 +0000 (12:41 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  pata_bf54x: decrease count first.
  sata_mv: re-enable hotplug, update TODO list
  sata_mv: leave SError bits untouched in mv_err_intr
  sata_mv: more interrupt handling rework
  sata_mv: tidy host controller interrupt handling
  sata_mv: simplify request/response queue handling
  sata_mv: simplify freeze/thaw bit-shift calculations
  sata_mv mask all interrupt coalescing bits
  sata_mv more cosmetics
  ata_piix: add Asus Eee 701 controller to short cable list
  libata-eh set tf flags in NCQ EH result_tf
  make sata_set_spd_needed() static
  make sata_print_link_status() static
  libata-acpi.c: remove unneeded #if's
  sata_nv: make hardreset return -EAGAIN on success
  ahci: retry enabling AHCI a few times before spitting out WARN_ON()
  libata: make WARN_ON conditions in ata_sff_hsm_move() more strict
  ATA/IDE: fix platform driver hotplug/coldplug
  sata_sis: SCR accessors return -EINVAL when requested SCR isn't available
  libata: functions with definition should not be extern

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Fri, 25 Apr 2008 19:40:57 +0000 (12:40 -0700)]
Merge git://git./linux/kernel/git/gregkh/usb-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (120 commits)
  usb: don't update devnum for wusb devices
  wusb: make ep0_reinit available for modules
  wusb: devices dont use a set address
  wusb: teach choose_address() about wireless devices
  wusb: add link wusb-usb device
  wusb: add authenticathed bit to usb_dev
  USB: remove unnecessary type casting of urb->context
  usb serial: more fixes and groundwork for tty changes
  USB: replace remaining __FUNCTION__ occurrences
  USB: usbfs: export the URB_NO_INTERRUPT flag to userspace
  USB: fix compile problems in ehci-hcd
  USB: ehci: qh_completions cleanup and bugfix
  USB: cdc-acm: signedness fix
  USB: add documentation about callbacks
  USB: don't explicitly reenable root-hub status interrupts
  USB: OHCI: turn off RD when remote wakeup is disabled
  USB: HCDs use the do_remote_wakeup flag
  USB: g_file_storage: ignore bulk-out data after invalid CBW
  USB: serial: remove endpoints setting checks from core and header
  USB: serial: remove unneeded number endpoints settings
  ...

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Fri, 25 Apr 2008 19:38:14 +0000 (12:38 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (40 commits)
  Input: wacom - add support for Cintiq 20WSX
  Input: ucb1400_ts - IRQ probe fix
  Input: at32psif - update MODULE_AUTHOR with new email
  Input: mac_hid - add lockdep annotation to emumousebtn
  Input: i8042 - fix incorrect usage of strncpy and strncat
  Input: bf54x-keys - add infrastructure for keypad wakeups
  Input: add MODULE_ALIAS() to hotpluggable platform modules
  Input: drivers/char/keyboard.c - use time_after
  Input: fix ordering in joystick Makefile
  Input: wm97xx-core - support use as a wakeup source
  Input: wm97xx-core - use IRQF_SAMPLE_RANDOM
  Input: wm97xx-core - only schedule interrupt handler if not already scheduled
  Input: add Zhen Hua driver
  Input: aiptek - add support for Genius G-PEN 560 tablet
  Input: wacom - implement suspend and autosuspend
  Input: xpad - set proper buffer length for outgoing requests
  Input: omap-keypad - fix build warning
  Input: gpio_keys - irq handling cleanup
  Input: add PS/2 serio driver for AVR32 devices
  Input: put ledstate in the keyboard notifier
  ...

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Fri, 25 Apr 2008 19:36:42 +0000 (12:36 -0700)]
Merge git://git./linux/kernel/git/sam/kbuild

* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild:
  kconfig: fix broken target update-po-config
  kbuild: silence documentation GEN xml messages according to $(quiet)

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm
Linus Torvalds [Fri, 25 Apr 2008 19:33:49 +0000 (12:33 -0700)]
Merge git://git./linux/kernel/git/agk/linux-2.6-dm

* git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm: (24 commits)
  dm crypt: add documentation
  dm: remove md argument from specific_minor
  dm table: remove unused dm_create_error_table
  dm table: drop void suspend_targets return
  dm: unplug queues in threads
  dm raid1: use timer
  dm: move include files
  dm kcopyd: rename
  dm: expose macros
  dm kcopyd: remove redundant client counting
  dm kcopyd: private mempool
  dm kcopyd: per device
  dm log: make module use tracking internal
  dm log: move register functions
  dm log: clean interface
  dm kcopyd: clean interface
  dm io: clean interface
  dm io: rename error to error_bits
  dm snapshot: store pointer to target instance
  dm log: move dirty region log code into separate module
  ...

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux...
Linus Torvalds [Fri, 25 Apr 2008 19:32:10 +0000 (12:32 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/x86/linux-2.6-xen-next

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-xen-next: (52 commits)
  xen: add balloon driver
  xen: allow compilation with non-flat memory
  xen: fold xen_sysexit into xen_iret
  xen: allow set_pte_at on init_mm to be lockless
  xen: disable preemption during tlb flush
  xen pvfb: Para-virtual framebuffer, keyboard and pointer driver
  xen: Add compatibility aliases for frontend drivers
  xen: Module autoprobing support for frontend drivers
  xen blkfront: Delay wait for block devices until after the disk is added
  xen/blkfront: use bdget_disk
  xen: Make xen-blkfront write its protocol ABI to xenstore
  xen: import arch generic part of xencomm
  xen: make grant table arch portable
  xen: replace callers of alloc_vm_area()/free_vm_area() with xen_ prefixed one
  xen: make include/xen/page.h portable moving those definitions under asm dir
  xen: add resend_irq_on_evtchn() definition into events.c
  Xen: make events.c portable for ia64/xen support
  xen: move events.c to drivers/xen for IA64/Xen support
  xen: move features.c from arch/x86/xen/features.c to drivers/xen
  xen: add missing definitions in include/xen/interface/vcpu.h which ia64/xen needs
  ...

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Fri, 25 Apr 2008 19:29:55 +0000 (12:29 -0700)]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Wrap SMP IPIs with irq_enter()/irq_exit().
  [SPARC64]: Fix args to 64-bit sys_semctl() via sys_ipc().

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 25 Apr 2008 19:28:28 +0000 (12:28 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (48 commits)
  net: Fix wrong interpretation of some copy_to_user() results.
  xfrm: alg_key_len & alg_icv_len should be unsigned
  [netdrvr] tehuti: move ioctl perm check closer to function start
  ipv6: Fix typo in net/ipv6/Kconfig
  via-velocity: fix vlan receipt
  tg3: sparse cleanup
  forcedeth: realtek phy crossover detection
  ibm_newemac: Increase MDIO timeouts
  gianfar: Fix skb allocation strategy
  netxen: reduce stack usage of netxen_nic_flash_print
  smc911x: test after postfix decrement fails in smc911x_{reset,drop_pkt}
  net drivers: fix platform driver hotplug/coldplug
  forcedeth: new backoff implementation
  ehea: make things static
  phylib: Add support for board-level PHY fixups
  [netdrvr] atlx: code movement: move atl1 parameter parsing
  atlx: remove flash vendor parameter
  korina: misc cleanup
  korina: fix misplaced return statement
  WAN: Fix confusing insmod error code for C101 too.
  ...

16 years agoUpdate .gitignore files
Matthew Wilcox [Fri, 25 Apr 2008 10:21:11 +0000 (04:21 -0600)]
Update .gitignore files

Add some autogenerated files to various .gitignore files

Signed-off-by: Matthew Wilcox <willy@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge git://git.infradead.org/mtd-2.6
Linus Torvalds [Fri, 25 Apr 2008 19:25:48 +0000 (12:25 -0700)]
Merge git://git.infradead.org/mtd-2.6

* git://git.infradead.org/mtd-2.6: (82 commits)
  [MTD] m25p80: Add Support for ATMEL AT25DF641 64-Megabit SPI Flash
  [MTD] m25p80: add FAST_READ access support to M25Pxx
  [MTD] [NAND] bf5xx_nand: Avoid crash if bfin_mac is installed.
  [MTD] [NAND] at91_nand: control NCE signal
  [MTD] [NAND] AT91 hardware ECC compile fix for at91sam9263 / at91sam9260
  [MTD] [NAND] Hardware ECC controller on at91sam9263 / at91sam9260
  [JFFS2] Introduce dbg_readinode2 log level, use it to shut read_dnode() up
  [JFFS2] Fix jffs2_reserve_space() when all blocks are pending erasure.
  [JFFS2] Add erase_checking_list to hold blocks being marked.
  UBI: add a message
  [JFFS2] Return values of jffs2_block_check_erase error paths
  [MTD] Clean up AR7 partition map support
  [MTD] [NOR] Fix Intel CFI driver for collie flash
  [JFFS2] Finally remove redundant ref->__totlen field.
  [JFFS2] Honour TEST_TOTLEN macro in debugging code. ref->__totlen is going!
  [JFFS2] Add paranoia debugging for superblock counts
  [JFFS2] Fix free space leak with in-band cleanmarkers
  [JFFS2] Self-sufficient #includes in jffs2_fs_i.h: include <linux/mutex.h>
  [MTD] [NAND] Verify probe by retrying to checking the results match
  [MTD] [NAND] S3C2410 Allow ECC disable to be specified by the board
  ...

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux...
Linus Torvalds [Fri, 25 Apr 2008 19:25:03 +0000 (12:25 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mingo/linux-2.6-sched-fixes2

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched-fixes2:
  sched: use alloc_bootmem() instead of alloc_bootmem_low()

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394...
Linus Torvalds [Fri, 25 Apr 2008 19:24:06 +0000 (12:24 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  ieee1394: silence defined but not used warning in non-modular builds
  ieee1394: rawiso: requeue packet for transmission after skipped cycle

16 years agokconfig: fix broken target update-po-config
Sam Ravnborg [Fri, 25 Apr 2008 19:15:41 +0000 (21:15 +0200)]
kconfig: fix broken target update-po-config

Massimo Maiurana reported:
In the latest kernel "make update-po-config" fails because it tries
to open arch/Kconfig/Kconfig, since the ls command doesn't
distinguish between files and directories.

Cc: Massimo Maiurana <maiurana@gmail.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Fri, 25 Apr 2008 19:06:46 +0000 (12:06 -0700)]
Merge git://git./linux/kernel/git/sam/kbuild

* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild:
  kbuild: fix depmod comment
  kbuild: Add new Kbuild variable KBUILD_EXTRA_SYMBOLS
  kbuild: support loading extra symbols in modpost
  Add option to enable -Wframe-larger-than= on gcc 4.4
  kbuild: add kconfig symbols to tags output
  kbuild: fix some minor typoes
  kbuild: error out on missing MODULE_LICENSE

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux...
Linus Torvalds [Fri, 25 Apr 2008 19:03:36 +0000 (12:03 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mingo/linux-2.6-x86-fixes4

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-x86-fixes4:
  x86: harden kernel code patching
  x86: clean up text_poke()
  x86: fix text_poke()
  x86: remove set_fixmap() warning
  x86: make __set_fixmap() non-init
  x86: make clear_fixmap() available on 64-bit as well

16 years agokbuild: silence documentation GEN xml messages according to $(quiet)
Mike Frysinger [Fri, 28 Mar 2008 21:30:58 +0000 (14:30 -0700)]
kbuild: silence documentation GEN xml messages according to $(quiet)

Add rules for gen_xml and its quiet & silent variants.  This causes "make -s"
to be silent for gen_xml.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agokbuild: fix depmod comment
Uwe Kleine-König [Wed, 5 Mar 2008 14:57:04 +0000 (15:57 +0100)]
kbuild: fix depmod comment

Signed-off-by: Uwe Kleine-König <Uwe.Kleine-Koenig@digi.com>
Cc: trivial@kernel.org
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agoMerge branch 'for-linus' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Fri, 25 Apr 2008 18:45:40 +0000 (11:45 -0700)]
Merge branch 'for-linus' of git://linux-nfs.org/~bfields/linux

* 'for-linus' of git://linux-nfs.org/~bfields/linux:
  nfsd: don't allow setting ctime over v4
  Update to NFS/RDMA documentation
  locks: don't call ->copy_lock methods on return of conflicting locks
  lockd: unlock lockd locks held for a certain filesystem
  lockd: unlock lockd locks associated with a given server ip
  leases: remove unneeded variable from fcntl_setlease().
  leases: move lock allocation earlier in generic_setlease()
  leases: when unlocking, skip locking-related steps
  leases: fix a return-value mixup

16 years agokbuild: Add new Kbuild variable KBUILD_EXTRA_SYMBOLS
Richard Hacker [Thu, 28 Feb 2008 08:40:58 +0000 (09:40 +0100)]
kbuild: Add new Kbuild variable KBUILD_EXTRA_SYMBOLS

This patch adds a new (Kbuild) Makefile variable KBUILD_EXTRA_SYMBOLS.
The space separated list of file names assigned to KBUILD_EXTRA_SYMBOLS
is used when calling scripts/mod/modpost during stage 2 of the Kbuild
process for non-kernel-tree modules.

Signed-off-by: Richard Hacker <lerichi@gmx.net>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agokbuild: support loading extra symbols in modpost
Richard Hacker [Thu, 28 Feb 2008 08:40:52 +0000 (09:40 +0100)]
kbuild: support loading extra symbols in modpost

This patch adds a new command line option -E to modpost, expecting a symbol
file as an argument which is read prior to symbol processing. -E can be
supplied multiple times for as many files as is needed.

When building kernel modules that depend on other modules not in the main
kernel tree, modpost complains about undefined symbols:
# make -C /path/to/linux/kernel M=/path/to/my/module
...
Building modules, stage 2.
....
WARNING: "rt_copy_buf" [/home/rich/osc_etl_rtw/osc_kmod.ko] undefined!
...etc

This situation occurs when modpost processes the new module's symbols. When
it finds symbols not exported by the mainline kernel, it issues this warning.

The patch adds a new command line option -e to modpost which expects a symbol
file as an argument. The symbols listed in this file are added to modpost's
symbol tables during startup. -e can be supplied as often as required.

This patch works together with the second patch. It introduces a new make
variable, KBUILD_EXTRA_SYMBOLS, which is used when calling modpost.

Signed-off-by: Richard Hacker <lerichi@gmx.net>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agoAdd option to enable -Wframe-larger-than= on gcc 4.4
Andi Kleen [Fri, 22 Feb 2008 14:15:03 +0000 (15:15 +0100)]
Add option to enable -Wframe-larger-than= on gcc 4.4

Add option to enable -Wframe-larger-than= on gcc 4.4

gcc mainline (upcoming 4.4) added a new -Wframe-larger-than=...
option to warn at build time about too large stack frames. Add a config
option to enable this warning, since this very useful for the kernel.

I choose (somewhat arbitarily) 2048 as default warning threshold for 64bit
and 1024 as default for 32bit architectures.  With some research and
fixing all the code for smaller values these defaults should be probably
lowered.

With the default allyesconfigs have some new warnings, but I think
that is all code that should be just fixed.

At some point (when gcc 4.4 is released and widely used) this should
obsolete make checkstack

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agokbuild: add kconfig symbols to tags output
Alexey Dobriyan [Fri, 22 Feb 2008 12:02:33 +0000 (15:02 +0300)]
kbuild: add kconfig symbols to tags output

Steps to reproduce:

vi -t NETFILTER

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agokbuild: fix some minor typoes
Robert P. J. Day [Mon, 18 Feb 2008 09:48:20 +0000 (04:48 -0500)]
kbuild: fix some minor typoes

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
16 years agokbuild: error out on missing MODULE_LICENSE
Sam Ravnborg [Fri, 25 Apr 2008 18:13:30 +0000 (20:13 +0200)]
kbuild: error out on missing MODULE_LICENSE

Adrian Bunk suggested a build time check for
missing MODULE_LICENSE annotation in modules.
The build time check is fatal as we really
want this fixed for all modules.
In-tree modules should all have been fixed up by now.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Adrian Bunk <bunk@kernel.org>
16 years agoDocument seq_path_root()
Jonathan Corbet [Fri, 25 Apr 2008 17:56:37 +0000 (11:56 -0600)]
Document seq_path_root()

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
16 years agox86: harden kernel code patching
Ingo Molnar [Fri, 25 Apr 2008 15:07:03 +0000 (17:07 +0200)]
x86: harden kernel code patching

Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: clean up text_poke()
Mathieu Desnoyers [Thu, 24 Apr 2008 15:03:33 +0000 (11:03 -0400)]
x86: clean up text_poke()

Clean up the codepath, remove alignment restrictions and do sanity
checking of the end result, to make sure we patched the right site.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: fix text_poke()
Jiri Slaby [Mon, 28 Apr 2008 00:51:23 +0000 (02:51 +0200)]
x86: fix text_poke()

kernel_text_address returns true even for modules which is not wanted
in text_poke. Use core_kernel_text instead.

This is a regression introduced in e587cadd8f47e202a30712e2906a65a0606d5865
which caused occasionaly crashes after suspend/resume.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
CC: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
CC: Andi Kleen <andi@firstfloor.org>
CC: pageexec@freemail.hu
CC: H. Peter Anvin <hpa@zytor.com>
CC: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: remove set_fixmap() warning
Ingo Molnar [Fri, 25 Apr 2008 16:05:57 +0000 (18:05 +0200)]
x86: remove set_fixmap() warning

set_fixmap()+clear_fixmap() is safe.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: make __set_fixmap() non-init
Ingo Molnar [Fri, 25 Apr 2008 16:28:21 +0000 (18:28 +0200)]
x86: make __set_fixmap() non-init

Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: make clear_fixmap() available on 64-bit as well
Ingo Molnar [Fri, 25 Apr 2008 16:25:25 +0000 (18:25 +0200)]
x86: make clear_fixmap() available on 64-bit as well

Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agoVarious fixes to Documentation/HOWTO
Jonathan Corbet [Fri, 25 Apr 2008 17:23:56 +0000 (11:23 -0600)]
Various fixes to Documentation/HOWTO

Fix a number of things which have gone somewhat out-of-date over the last
few months.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
16 years agonfsd: don't allow setting ctime over v4
J. Bruce Fields [Thu, 24 Apr 2008 20:59:30 +0000 (16:59 -0400)]
nfsd: don't allow setting ctime over v4

Presumably this is left over from earlier drafts of v4, which listed
TIME_METADATA as writeable.  It's read-only in rfc 3530, and shouldn't
be modifiable anyway.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
16 years agoUpdate to NFS/RDMA documentation
James Lentini [Thu, 24 Apr 2008 19:57:43 +0000 (15:57 -0400)]
Update to NFS/RDMA documentation

Update to the NFS/RDMA documentation to clarify how to configure the
exports file.

Signed-off-by: James Lentini <jlentini@netapp.com>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
16 years agolocks: don't call ->copy_lock methods on return of conflicting locks
J. Bruce Fields [Thu, 24 Apr 2008 14:08:22 +0000 (10:08 -0400)]
locks: don't call ->copy_lock methods on return of conflicting locks

The file_lock structure is used both as a heavy-weight representation of
an active lock, with pointers to reference-counted structures, etc., and
as a simple container for parameters that describe a file lock.

The conflicting lock returned from __posix_lock_file is an example of
the latter; so don't call the filesystem or lock manager callbacks when
copying to it.  This also saves the need for an unnecessary
locks_init_lock in the nfsv4 server.

Thanks to Trond for pointing out the error.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
16 years agolockd: unlock lockd locks held for a certain filesystem
Wendy Cheng [Thu, 17 Jan 2008 16:10:12 +0000 (11:10 -0500)]
lockd: unlock lockd locks held for a certain filesystem

Add /proc/fs/nfsd/unlock_filesystem, which allows e.g.:

shell> echo /mnt/sfs1 > /proc/fs/nfsd/unlock_filesystem

so that a filesystem can be unmounted before allowing a peer nfsd to
take over nfs service for the filesystem.

Signed-off-by: S. Wendy Cheng <wcheng@redhat.com>
Cc: Lon Hohberger <lhh@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
 fs/lockd/svcsubs.c          |   66 +++++++++++++++++++++++++++++++++++++++-----
 fs/nfsd/nfsctl.c            |   65 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/lockd/lockd.h |    7 ++++
 3 files changed, 131 insertions(+), 7 deletions(-)

16 years agolockd: unlock lockd locks associated with a given server ip
Wendy Cheng [Thu, 17 Jan 2008 16:10:12 +0000 (11:10 -0500)]
lockd: unlock lockd locks associated with a given server ip

For high-availability NFS service, we generally need to be able to drop
file locks held on the exported filesystem before moving clients to a
new server.  Currently the only way to do that is by shutting down lockd
entirely, which is often undesireable (for example, if you want to
continue exporting other filesystems).

This patch allows the administrator to release all locks held by clients
accessing the client through a given server ip address, by echoing that
address to a new file, /proc/fs/nfsd/unlock_ip, as in:

shell> echo 10.1.1.2 > /proc/fs/nfsd/unlock_ip

The expected sequence of events can be:
1. Tear down the IP address
2. Unexport the path
3. Write IP to /proc/fs/nfsd/unlock_ip to unlock files
4. Signal peer to begin take-over.

For now we only support IPv4 addresses and NFSv2/v3 (NFSv4 locks are not
affected).

Also, if unmounting the filesystem is required, we assume at step 3 that
clients using the given server ip are the only clients holding locks on
the given filesystem; otherwise, an additional patch is required to
allow revoking all locks held by lockd on a given filesystem.

Signed-off-by: S. Wendy Cheng <wcheng@redhat.com>
Cc: Lon Hohberger <lhh@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
 fs/lockd/svcsubs.c          |   66 +++++++++++++++++++++++++++++++++++++++-----
 fs/nfsd/nfsctl.c            |   65 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/lockd/lockd.h |    7 ++++
 3 files changed, 131 insertions(+), 7 deletions(-)

16 years agoleases: remove unneeded variable from fcntl_setlease().
David M. Richter [Wed, 23 Apr 2008 20:29:02 +0000 (16:29 -0400)]
leases: remove unneeded variable from fcntl_setlease().

fcntl_setlease() has a struct dentry* that is used only once; this patch
removes it.

Signed-off-by: David M. Richter <richterd@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
16 years agoleases: move lock allocation earlier in generic_setlease()
David M. Richter [Wed, 23 Apr 2008 20:29:01 +0000 (16:29 -0400)]
leases: move lock allocation earlier in generic_setlease()

In generic_setlease(), the struct file_lock is allocated after tests for the
presence of conflicting readers/writers is done, despite the fact that the
allocation might block; this patch moves the allocation earlier.  A subsequent
set of patches will rely on this behavior to properly serialize between a
modified __break_lease() and generic_setlease().

Signed-off-by: David M. Richter <richterd@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
16 years agoleases: when unlocking, skip locking-related steps
David M. Richter [Wed, 23 Apr 2008 20:29:00 +0000 (16:29 -0400)]
leases: when unlocking, skip locking-related steps

In generic_setlease(), we don't need to allocate a new struct file_lock
or check for readers or writers when called with F_UNLCK.

Signed-off-by: David M. Richter <richterd@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
16 years agoleases: fix a return-value mixup
David M. Richter [Wed, 23 Apr 2008 20:28:59 +0000 (16:28 -0400)]
leases: fix a return-value mixup

Fixes a return-value mixup from 85c59580b30c82aa771aa33b37217a6b6851bc14
"locks: Fix potential OOPS in generic_setlease()", in which -ENOMEM replaced
what had been intended to stay -EAGAIN in the variable "error".

Signed-off-by: David M. Richter <richterd@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
16 years agoieee1394: silence defined but not used warning in non-modular builds
Tony Breeds [Thu, 24 Apr 2008 07:02:04 +0000 (09:02 +0200)]
ieee1394: silence defined but not used warning in non-modular builds

Currently the kernel will issue the following warning:
drivers/ieee1394/raw1394.c:2938: warning: 'raw1394_id_table' defined but not used
Add #ifdef MODULE guards around the declaration.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Ditto with dv1394_id_table and video1394_id_table.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
16 years agoieee1394: rawiso: requeue packet for transmission after skipped cycle
Pieter Palmers [Wed, 19 Mar 2008 21:10:59 +0000 (22:10 +0100)]
ieee1394: rawiso: requeue packet for transmission after skipped cycle

As it seems, some host controllers have issues that can cause them to
skip cycles now and then when using large packets. I suspect that this
is due to DMA not succeeding in time. If the transmit fifo can't contain
more than one packet (big packets), the DMA should provide a new packet
each cycle (125us). I am under the impression that my current PCI
express test system can't guarantee this.

In any case, the patch tries to provide a workaround as follows:
The DMA program descriptors are modified such that when an error occurs,
the DMA engine retries the descriptor the next cycle instead of
stalling. This way no data is lost. The side effect of this is that
packets are sent with one cycle delay. This however might not be that
much of a problem for certain protocols (e.g. AM824). If they use
padding packets for e.g. rate matching they can drop one of those to
resync the streams.

The amount of skips between two userspace wakeups is counted. This
number is then propagated to userspace through the upper 16 bits of the
'dropped' parameter. This allows unmodified userspace applications due
to the following:
1) libraw simply passes this dropped parameter to the user application
2) the meaning of the dropped parameter is: if it's nonzero, something
bad has happened. The actual value of the parameter at this moment does
not have a specific meaning.

A libraw client can then retrieve the number of skipped cycles and
account for them if needed.

Signed-off-by: Pieter Palmers <pieterp@joow.be>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
16 years ago[PATCH] sanitize locate_fd()
Al Viro [Thu, 24 Apr 2008 00:38:10 +0000 (20:38 -0400)]
[PATCH] sanitize locate_fd()

* 'file' argument is unused; lose it.
* move setting flags from the caller (dupfd()) to locate_fd();
  pass cloexec flag as new argument.  Note that files_fdtable()
  that used to be in dupfd() isn't needed in the place in
  locate_fd() where the moved code ends up - we know that ->file_lock
  hadn't been dropped since the last time we calculated fdt because
  we can get there only if expand_files() returns 0 and it doesn't
  drop/reacquire in that case.
* move getting/dropping ->file_lock into locate_fd().  Now the caller
  doesn't need to do anything with files_struct *files anymore and
  we can move that inside locate_fd() as well, killing the
  struct files_struct * argument.

At that point locate_fd() is extremely similar to get_unused_fd_flags()
and the next patches will merge those two.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
16 years ago[PATCH] sanitize unshare_files/reset_files_struct
Al Viro [Tue, 22 Apr 2008 09:31:30 +0000 (05:31 -0400)]
[PATCH] sanitize unshare_files/reset_files_struct

* let unshare_files() give caller the displaced files_struct
* don't bother with grabbing reference only to drop it in the
  caller if it hadn't been shared in the first place
* in that form unshare_files() is trivially implemented via
  unshare_fd(), so we eliminate the duplicate logics in fork.c
* reset_files_struct() is not just only called for current;
  it will break the system if somebody ever calls it for anything
  else (we can't modify ->files of somebody else).  Lose the
  task_struct * argument.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
16 years ago[PATCH] sanitize handling of shared descriptor tables in failing execve()
Al Viro [Tue, 22 Apr 2008 09:11:59 +0000 (05:11 -0400)]
[PATCH] sanitize handling of shared descriptor tables in failing execve()

* unshare_files() can fail; doing it after irreversible actions is wrong
  and de_thread() is certainly irreversible.
* since we do it unconditionally anyway, we might as well do it in do_execve()
  and save ourselves the PITA in binfmt handlers, etc.
* while we are at it, binfmt_som actually leaked files_struct on failure.

As a side benefit, unshare_files(), put_files_struct() and reset_files_struct()
become unexported.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
16 years ago[PATCH] close race in unshare_files()
Al Viro [Tue, 22 Apr 2008 08:45:46 +0000 (04:45 -0400)]
[PATCH] close race in unshare_files()

updating current->files requires task_lock

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
16 years ago[PATCH] restore sane ->umount_begin() API
Al Viro [Thu, 24 Apr 2008 11:21:56 +0000 (07:21 -0400)]
[PATCH] restore sane ->umount_begin() API

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
16 years agocifs: timeout dfs automounts +little fix.
Igor Mammedov [Thu, 24 Apr 2008 08:56:07 +0000 (12:56 +0400)]
cifs: timeout dfs automounts +little fix.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
16 years agodm crypt: add documentation
Milan Broz [Thu, 24 Apr 2008 21:11:03 +0000 (22:11 +0100)]
dm crypt: add documentation

Add description of dm-crypt to device-mapper documentation.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm: remove md argument from specific_minor
Frederik Deweerdt [Thu, 24 Apr 2008 21:10:59 +0000 (22:10 +0100)]
dm: remove md argument from specific_minor

The small patch below:
- Removes the unused md argument from both specific_minor() and next_free_minor()
- Folds kmalloc + memset(0) into a single kzalloc call in alloc_dev()

This has been compile tested on x86.

Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm table: remove unused dm_create_error_table
Adrian Bunk [Thu, 24 Apr 2008 21:10:56 +0000 (22:10 +0100)]
dm table: remove unused dm_create_error_table

dm_create_error_table() was added in kernel 2.6.18 and never used...

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm table: drop void suspend_targets return
Adrian Bunk [Thu, 24 Apr 2008 21:10:51 +0000 (22:10 +0100)]
dm table: drop void suspend_targets return

void returning functions returned the return value of another void
returning function...

Spotted by sparse.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm: unplug queues in threads
Mikulas Patocka [Thu, 24 Apr 2008 21:10:47 +0000 (22:10 +0100)]
dm: unplug queues in threads

Remove an avoidable 3ms delay on some dm-raid1 and kcopyd I/O.

It is specified that any submitted bio without BIO_RW_SYNC flag may plug the
queue (i.e. block the requests from being dispatched to the physical device).

The queue is unplugged when the caller calls blk_unplug() function. Usually, the
sequence is that someone calls submit_bh to submit IO on a buffer. The IO plugs
the queue and waits (to be possibly joined with other adjacent bios). Then, when
the caller calls wait_on_buffer(), it unplugs the queue and submits the IOs to
the disk.

This was happenning:

When doing O_SYNC writes, function fsync_buffers_list() submits a list of
bios to dm_raid1, the bios are added to dm_raid1 write queue and kmirrord is
woken up.

fsync_buffers_list() calls wait_on_buffer().  That unplugs the queue, but
there are no bios on the device queue as they are still in the dm_raid1 queue.

wait_on_buffer() starts waiting until the IO is finished.

kmirrord is scheduled, kmirrord takes bios and submits them to the devices.

The submitted bio plugs the harddisk queue but there is no one to unplug it.
(The process that called wait_on_buffer() is already sleeping.)

So there is a 3ms timeout, after which the queues on the harddisks are
unplugged and requests are processed.

This 3ms timeout meant that in certain workloads (e.g. O_SYNC, 8kb writes),
dm-raid1 is 10 times slower than md raid1.

Every time we submit something asynchronously via dm_io, we must unplug the
queue actually to send the request to the device.

This patch adds an unplug call to kmirrord - while processing requests, it keeps
the queue plugged (so that adjacent bios can be merged); when it finishes
processing all the bios, it unplugs the queue to submit the bios.

It also fixes kcopyd which has the same potential problem. All kcopyd requests
are submitted with BIO_RW_SYNC.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
16 years agodm raid1: use timer
Mikulas Patocka [Thu, 24 Apr 2008 21:10:42 +0000 (22:10 +0100)]
dm raid1: use timer

This patch replaces the schedule() in the main kmirrord thread with a timer.
The schedule() could introduce an unwanted delay when work is ready to be
processed.

The code instead calls wake() when there's work to be done immediately, and
delayed_wake() after a failure to give a short delay before retrying.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm: move include files
Alasdair G Kergon [Thu, 24 Apr 2008 21:02:01 +0000 (22:02 +0100)]
dm: move include files

Publish the dm-io, dm-log and dm-kcopyd headers in include/linux.

Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm kcopyd: rename
Alasdair G Kergon [Thu, 24 Apr 2008 20:55:00 +0000 (21:55 +0100)]
dm kcopyd: rename

Rename kcopyd.[ch] to dm-kcopyd.[ch].

Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm: expose macros
Alasdair G Kergon [Thu, 24 Apr 2008 20:43:52 +0000 (21:43 +0100)]
dm: expose macros

Make dm.h macros and inlines available in include/linux/device-mapper.h

Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm kcopyd: remove redundant client counting
Mikulas Patocka [Thu, 24 Apr 2008 20:43:49 +0000 (21:43 +0100)]
dm kcopyd: remove redundant client counting

Remove client counting code that is no longer needed.

Initialization and destruction is made globally from dm_init and dm_exit and is
not based on client counts. Initialization allocates only one empty slab cache,
so there is no negative impact from performing the initialization always,
regardless of whether some client uses kcopyd or not.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm kcopyd: private mempool
Mikulas Patocka [Thu, 24 Apr 2008 20:43:46 +0000 (21:43 +0100)]
dm kcopyd: private mempool

Change the global mempool in kcopyd into a per-device mempool to avoid
deadlock possibilities.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm kcopyd: per device
Mikulas Patocka [Thu, 24 Apr 2008 20:43:44 +0000 (21:43 +0100)]
dm kcopyd: per device

Make one kcopyd thread per device.

The original shared kcopyd could deadlock.

Configuration:

16 years agodm log: make module use tracking internal
Jonathan Brassow [Thu, 24 Apr 2008 20:43:41 +0000 (21:43 +0100)]
dm log: make module use tracking internal

Remove internal module reference fields from the interface.

Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm log: move register functions
Alasdair G Kergon [Thu, 24 Apr 2008 20:43:38 +0000 (21:43 +0100)]
dm log: move register functions

Reorder a couple of functions in the file so the next patch is readable.

Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm log: clean interface
Heinz Mauelshagen [Thu, 24 Apr 2008 20:43:35 +0000 (21:43 +0100)]
dm log: clean interface

Clean up the dm-log interface to prepare for publishing it in include/linux.

Signed-off-by: Heinz Mauelshagen <hjm@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm kcopyd: clean interface
Heinz Mauelshagen [Thu, 24 Apr 2008 20:43:19 +0000 (21:43 +0100)]
dm kcopyd: clean interface

Clean up the kcopyd interface to prepare for publishing it in include/linux.

Signed-off-by: Heinz Mauelshagen <hjm@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm io: clean interface
Heinz Mauelshagen [Thu, 24 Apr 2008 20:43:17 +0000 (21:43 +0100)]
dm io: clean interface

Clean up the dm-io interface to prepare for publishing it in include/linux.

Signed-off-by: Heinz Mauelshagen <hjm@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm io: rename error to error_bits
Alasdair G Kergon [Thu, 24 Apr 2008 20:43:14 +0000 (21:43 +0100)]
dm io: rename error to error_bits

Rename 'error' to 'error_bits' for clarity.

Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm snapshot: store pointer to target instance
Mikulas Patocka [Thu, 24 Apr 2008 20:43:11 +0000 (21:43 +0100)]
dm snapshot: store pointer to target instance

Save pointer to dm_target in dm_snapshot structure.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm log: move dirty region log code into separate module
Heinz Mauelshagen [Thu, 24 Apr 2008 20:43:09 +0000 (21:43 +0100)]
dm log: move dirty region log code into separate module

Move the dirty region log code into a separate module so
other targets can share the code.

Signed-off-by: Heinz Mauelshagen <hjm@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm log: generalise name in messages
Heinz Mauelshagen [Thu, 24 Apr 2008 20:43:06 +0000 (21:43 +0100)]
dm log: generalise name in messages

Change dm-log.c messages from "mirror log" to "dirty region log" as
a new dm target wants to share this code.

Signed-off-by: Heinz Mauelshagen <hjm@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm raid1: use list_split_init
Robert P. J. Day [Thu, 24 Apr 2008 20:42:44 +0000 (21:42 +0100)]
dm raid1: use list_split_init

Use shorter list_splice_init() for brevity.

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm snapshot: reduce default memory allocation
Milan Broz [Thu, 24 Apr 2008 20:42:36 +0000 (21:42 +0100)]
dm snapshot: reduce default memory allocation

Limit the amount of memory allocated per snapshot on systems
with a large page size.  (The larger default chunk size on
these systems compensates for the smaller number of pages reserved.)

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
16 years agodm snapshot: fix chunksize sector conversion
Mikulas Patocka [Thu, 24 Apr 2008 20:41:50 +0000 (21:41 +0100)]
dm snapshot: fix chunksize sector conversion

If a snapshot has a smaller chunksize than the page size the
conversion to pages currently returns 0 instead of 1, causing:
kernel BUG in mempool_resize.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Cc: stable@kernel.org
16 years ago[SPARC64]: Wrap SMP IPIs with irq_enter()/irq_exit().
David S. Miller [Fri, 25 Apr 2008 10:11:37 +0000 (03:11 -0700)]
[SPARC64]: Wrap SMP IPIs with irq_enter()/irq_exit().

Otherwise all sorts of bad things can happen, including
spurious softlockup reports.

Other platforms have this same bug, in one form or
another, just don't see the issue because they
don't sleep as long as sparc64 can in NOHZ.

Thanks to some brilliant debugging by Peter Zijlstra.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[SPARC64]: Fix args to 64-bit sys_semctl() via sys_ipc().
David S. Miller [Fri, 25 Apr 2008 09:12:05 +0000 (02:12 -0700)]
[SPARC64]: Fix args to 64-bit sys_semctl() via sys_ipc().

Second and third arguments were swapped for whatever reason.

Reported by Tom Callaway.

Signed-off-by: David S. Miller <davem@davemloft.net>