linux-2.6
18 years agoMerge branch 'upstream'
Jeff Garzik [Wed, 12 Apr 2006 20:54:16 +0000 (16:54 -0400)] 
Merge branch 'upstream'

18 years ago[PATCH] sata_sil24: enable 64bit
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: enable 64bit

Enable 64bit.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: fix on-memory structure byteorder
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: fix on-memory structure byteorder

Data structures residing on memory and fetched by the controller
should have LE ordering.  Fix it.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: don't do hardreset during driver initialization
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: don't do hardreset during driver initialization

There's no need to perform hardreset during driver initialization.
It's already done during host reset and even if the controller is in
some wacky state, we now have proper hardreset to back us up.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: reimplement hardreset
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: reimplement hardreset

Reimplement hardreset according to the datasheet.  The old hardreset
didn't reset controller status and the controller might not be ready
after reset.  Also, as SStatus is a bit flakey after hardreset,
sata_std_hardrset() didn't use to wait long enough before proceeding.

Note that as we're not depending on SStatus, DET==1 condition cannot
be used to wait for link, so use shorter timeout for no device case.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: kill 10ms sleep in softreset
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: kill 10ms sleep in softreset

Nothing, not the datasheet nor the errats, says this delay is
necessary and with the previous PORT_CS_INIT change, we know the
controller is in good state.  Kill 10ms sleep.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: put port into known state before softresetting
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: put port into known state before softresetting

Make sure the controller has no pending commands and ready for command
before issuing SRST.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: implement sil24_init_port()
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: implement sil24_init_port()

Implement sil24_init_port which performs port initialization via
PORT_CS_INIT.  To be used later by softreset and EH.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: implement loss of completion interrupt on PCI-X errta fix
Tejun Heo [Tue, 11 Apr 2006 13:32:19 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: implement loss of completion interrupt on PCI-X errta fix

SiI3124 might lose completion interrupt if completion interrupt occurs
shortly after SLOT_STAT register is read for the previous completion
interrupt if it is operating in PCI-X mode.

This currently doesn't trigger as libata never queues more than one
command, but it will with NCQ changes.  This patch implements the
workaround - turning on WoC and explicitly clearing interrupt.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: consolidate host flags into SIL24_COMMON_FLAGS
Tejun Heo [Tue, 11 Apr 2006 13:32:18 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: consolidate host flags into SIL24_COMMON_FLAGS

All sil24 controllers share the same host flags except for NPORTS.
Consolidate them into SIL24_COMMON_FLAGS.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: add more constants
Tejun Heo [Tue, 11 Apr 2006 13:32:18 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: add more constants

Add HOST_CTRL_* and more PORT_IRQ_* bits.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: rename PORT_IRQ_SDB_FIS to PORT_IRQ_SDB_NOTIFY
Tejun Heo [Tue, 11 Apr 2006 13:32:18 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: rename PORT_IRQ_SDB_FIS to PORT_IRQ_SDB_NOTIFY

Rename PORT_IRQ_SDB_FIS to more proper PORT_IRQ_SDB_NOTIFY.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: typo fix
Tejun Heo [Tue, 11 Apr 2006 13:32:18 +0000 (22:32 +0900)] 
[PATCH] sata_sil24: typo fix

Type fix.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: export ata_set_sata_spd()
Tejun Heo [Tue, 11 Apr 2006 13:32:18 +0000 (22:32 +0900)] 
[PATCH] libata: export ata_set_sata_spd()

This will be used by LLDD hardreset implementation.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: cosmetic update to ata_bus_probe()
Tejun Heo [Tue, 11 Apr 2006 13:26:29 +0000 (22:26 +0900)] 
[PATCH] libata: cosmetic update to ata_bus_probe()

Move ata_set_mode() failure handling outside of ap->ops->set_mode if
clause such that it can handle ap->ops->set_mode failures after it's
updated.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: disable failed devices only once in ata_bus_probe()
Tejun Heo [Tue, 11 Apr 2006 13:26:29 +0000 (22:26 +0900)] 
[PATCH] libata: disable failed devices only once in ata_bus_probe()

Devices which consumed all their changes used to be disabled every
iteration.  This causes unnecessary noise in the console output.
Disable once and leave alone.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: use ata_wait_register()
Tejun Heo [Tue, 11 Apr 2006 13:22:30 +0000 (22:22 +0900)] 
[PATCH] sata_sil24: use ata_wait_register()

Replace hard-coded waiting loops in sata_sil24 to ata_wait_register().

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] ahci: use ata_wait_register()
Tejun Heo [Tue, 11 Apr 2006 13:22:29 +0000 (22:22 +0900)] 
[PATCH] ahci: use ata_wait_register()

Replace ahci_poll_register() with ata_wait_register().

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: implement ata_wait_register()
Tejun Heo [Tue, 11 Apr 2006 13:22:29 +0000 (22:22 +0900)] 
[PATCH] libata: implement ata_wait_register()

As waiting for some register bits to change seems to be a common
operation shared by some controllers, implement helper function
ata_wait_register().  This function also takes care of register write
flushing.

Note that the condition is inverted, the wait is over when the masked
value does NOT match @val.  As we're waiting for bits to change, this
test is more powerful and allows the function to be used in more
places.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: better error message from softreset
Tejun Heo [Tue, 11 Apr 2006 13:22:29 +0000 (22:22 +0900)] 
[PATCH] sata_sil24: better error message from softreset

Improve softreset error message.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] sata_sil24: fix timeout calculation in sil24_softreset
Tejun Heo [Tue, 11 Apr 2006 13:22:29 +0000 (22:22 +0900)] 
[PATCH] sata_sil24: fix timeout calculation in sil24_softreset

sil24_softreset calculated timeout by adding ATA_TMOUT_BOOT * HZ to
jiffies; however, as ATA_TMOUT_BOOT is already in jiffies, multiplying
by HZ makes the value way off.  Fix it.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: make reset methods complain when they fail
Tejun Heo [Tue, 11 Apr 2006 13:16:45 +0000 (22:16 +0900)] 
[PATCH] libata: make reset methods complain when they fail

Make reset methods complain loud when they fail.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: kill @verbose from ata_reset_fn_t
Tejun Heo [Tue, 11 Apr 2006 13:16:45 +0000 (22:16 +0900)] 
[PATCH] libata: kill @verbose from ata_reset_fn_t

@verbose was added to ata_reset_fn_t because AHCI complained during
probing if no device was attached to the port.  However, muting
failure message isn't the correct approach.  Reset methods are
responsible for detecting no device condition and finishing
successfully.  Now that AHCI softreset is fixed, kill @verbose.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: set default cbl in probeinit
Tejun Heo [Tue, 11 Apr 2006 13:16:44 +0000 (22:16 +0900)] 
[PATCH] libata: set default cbl in probeinit

Make setting CBL type responsibility of probeinit.  This allows using
only ap->cbl == ATA_CBL_SATA test in all other parts.  Without this,
ata_down_sata_spd_limit() doesn't work during probe reset.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years agolibata: Add helper ata_shost_to_port()
Jeff Garzik [Tue, 11 Apr 2006 17:12:34 +0000 (13:12 -0400)] 
libata: Add helper ata_shost_to_port()

18 years agolibata: Fix EH merge difference between this branch and upstream.
Jeff Garzik [Tue, 11 Apr 2006 17:04:39 +0000 (13:04 -0400)] 
libata: Fix EH merge difference between this branch and upstream.

18 years agoMerge branch 'master'
Jeff Garzik [Tue, 11 Apr 2006 16:51:40 +0000 (12:51 -0400)] 
Merge branch 'master'

Conflicts:

drivers/scsi/libata-scsi.c
include/linux/libata.h

18 years ago[PATCH] __group_complete_signal: remove bogus BUG_ON
Oleg Nesterov [Tue, 11 Apr 2006 18:18:58 +0000 (22:18 +0400)] 
[PATCH] __group_complete_signal: remove bogus BUG_ON

Commit e56d090310d7625ecb43a1eeebd479f04affb48b

   [PATCH] RCU signal handling

made this BUG_ON() unsafe. This code runs under ->siglock,
while switch_exec_pids() takes tasklist_lock.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Tue, 11 Apr 2006 13:41:02 +0000 (06:41 -0700)] 
Merge git://git./linux/kernel/git/sam/kbuild

* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild:
  kconfig: fix typo in change count initialization
  kconfig: recenter menuconfig
  kconfig: revert conf behaviour change
  kconfig: fix default value for choice input
  kbuild: fix NULL dereference in scripts/mod/modpost.c
  kbuild: fix mode of checkstack.pl and other files.
  kbuild: rebuild initramfs if content of initramfs changes
  kbuild: properly pass options to hostcc when doing make O=..
  kbuild: modules_install for external modules must not remove existing modules
  kbuild: fix make dir/
  ver_linux: don't print reiser4progs version if none found
  kbuild: mips: fix sed regexp to generate asm-offset.h
  kbuild: fix building single targets with make O=.. single-target
  kbuild: use relative path to -I
  kbuild: fix unneeded rebuilds in drivers/net/chelsio after moving source tree
  kbuild: fix unneeded rebuilds in drivers/media/video after moving source tree
  kbuild: fix garbled text in modules.txt

18 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Tue, 11 Apr 2006 13:40:17 +0000 (06:40 -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:
  [IA64] Prefetch mmap_sem in ia64_do_page_fault()
  [IA64] Failure to resume after INIT in user space
  [IA64] Pass more data to the MCA/INIT notify_die hooks
  [IA64] always map VGA framebuffer UC, even if it supports WB
  [IA64] fix bug in ia64 __mutex_fastpath_trylock
  [IA64] for_each_possible_cpu: ia64
  [IA64] update HP CSR space discovery via ACPI
  [IA64] Wire up new syscalls {set,get}_robust_list
  [IA64] 'msg' may be used uninitialized in xpc_initiate_allocate()
  [IA64] Wire up new syscall sync_file_range()

18 years ago[PATCH] x86_64: inline function prefix with __always_inline in vsyscall
mao, bibo [Tue, 11 Apr 2006 10:54:54 +0000 (12:54 +0200)] 
[PATCH] x86_64: inline function prefix with __always_inline in vsyscall

In vsyscall function do_vgettimeofday(), some functions are declared as
inlined, which is a hint for gcc to compile the function inlined but it
not forced.  Sometimes compiler does not compile the function as
inlined, so here inline is replaced by __always_inline prefix.

It does not happen in gcc compiler actually, but it possibly happens.

Signed-off-by: bibo mao <bibo.mao@intel.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Fix embarassing typo in mmconfig bus check
Andi Kleen [Tue, 11 Apr 2006 10:54:51 +0000 (12:54 +0200)] 
[PATCH] x86_64: Fix embarassing typo in mmconfig bus check

Surprising that it still worked at all with this - yes it was
tested.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386/x86-64: Remove checks for value == NULL in PCI config space access
Andi Kleen [Tue, 11 Apr 2006 10:54:48 +0000 (12:54 +0200)] 
[PATCH] i386/x86-64: Remove checks for value == NULL in PCI config space access

Nobody should pass NULL here. Could in theory make it a BUG,
but the NULL pointer oops will do as well.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386: Remove bogus special case code from AMD core parsing
Andi Kleen [Tue, 11 Apr 2006 10:54:45 +0000 (12:54 +0200)] 
[PATCH] i386: Remove bogus special case code from AMD core parsing

It's not actually needed and would break non power of two number
of cores.

Follows similar earlier x86-64 patch.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: fix sync before RDTSC on Intel cpus
Siddha, Suresh B [Tue, 11 Apr 2006 10:54:42 +0000 (12:54 +0200)] 
[PATCH] x86_64: fix sync before RDTSC on Intel cpus

Commit c818a18146997d1356a4840b0c01f1168c16c8a4 didn't do the expected
thing.  This fix will remove the additional sync(cpuid) before RDTSC on
Intel platforms..

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86_64: Remove check for canonical RIP
Andi Kleen [Tue, 11 Apr 2006 10:54:39 +0000 (12:54 +0200)] 
[PATCH] x86_64: Remove check for canonical RIP

As pointed out by Linus it is useless now because entry.S should
handle it correctly in all cases.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86-64/i386: Don't process APICs/IO-APICs in ACPI when APIC is disabled.
Andi Kleen [Tue, 11 Apr 2006 10:54:36 +0000 (12:54 +0200)] 
[PATCH] x86-64/i386: Don't process APICs/IO-APICs in ACPI when APIC is disabled.

When nolapic was passed or the local APIC was disabled
for another reason ACPI would still parse the IO-APICs
until these were explicitely disabled with noapic.

Usually this resulted in a non booting configuration unless
"nolapic noapic" was used.

I also disabled the local APIC parsing in this case, although
that's only cosmetic (suppresses a few printks)

This hopefully makes nolapic work in all cases.

Cc: Len Brown <len.brown@intel.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block
Linus Torvalds [Tue, 11 Apr 2006 13:34:02 +0000 (06:34 -0700)] 
Merge branch 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block

* 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block:
  [PATCH] vfs: add splice_write and splice_read to documentation
  [PATCH] Remove sys_ prefix of new syscalls from __NR_sys_*
  [PATCH] splice: warning fix
  [PATCH] another round of fs/pipe.c cleanups
  [PATCH] splice: comment styles
  [PATCH] splice: add Ingo as addition copyright holder
  [PATCH] splice: unlikely() optimizations
  [PATCH] splice: speedups and optimizations
  [PATCH] pipe.c/fifo.c code cleanups
  [PATCH] get rid of the PIPE_*() macros
  [PATCH] splice: speedup __generic_file_splice_read
  [PATCH] splice: add direct fd <-> fd splicing support
  [PATCH] splice: add optional input and output offsets
  [PATCH] introduce a "kernel-internal pipe object" abstraction
  [PATCH] splice: be smarter about calling do_page_cache_readahead()
  [PATCH] splice: optimize the splice buffer mapping
  [PATCH] splice: cleanup __generic_file_splice_read()
  [PATCH] splice: only call wake_up_interruptible() when we really have to
  [PATCH] splice: potential !page dereference
  [PATCH] splice: mark the io page as accessed

18 years ago[PATCH] Overrun in option-card USB driver
Matthias Urlichs [Tue, 11 Apr 2006 03:01:44 +0000 (05:01 +0200)] 
[PATCH] Overrun in option-card USB driver

Since the arrays are declared as in_urbs[N_IN_URB]
and out_urbs[N_OUT_URB], both for loops go one
over the end of the array. This fixes coverity id #555.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-Off-By: Matthias Urlichs <smurf@smurf.noris.de>
Signed-Off-By: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fbdev: Use logo with depth of 4 or less for static pseudocolor
Antonino A. Daplas [Tue, 11 Apr 2006 05:55:49 +0000 (22:55 -0700)] 
[PATCH] fbdev: Use logo with depth of 4 or less for static pseudocolor

Since the visual STATIC_PSEUDOCOLOR has a read-only colormap, use logos
with 16 colors only since these logos use the console palette.  This has a
higher likelihood that the logo will display correctly.

Signed-of-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] vesafb: Fix incorrect logo colors in x86_64
Antonino A. Daplas [Tue, 11 Apr 2006 05:55:48 +0000 (22:55 -0700)] 
[PATCH] vesafb: Fix incorrect logo colors in x86_64

Bugzilla Bug 6299:

A pixel size of 8 bits produces wrong logo colors in x86_64.

The driver has 2 methods for setting the color map, using the protected
mode interface provided by the video BIOS and directly writing to the VGA
registers.  The former is not supported in x86_64 and the latter is enabled
only in i386.

Fix by enabling the latter method in x86_64 only if supported by the BIOS.
If both methods are unsupported, change the visual of vesafb to
STATIC_PSEUDOCOLOR.

Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sparc32 vga support
Andrew Morton [Tue, 11 Apr 2006 05:55:47 +0000 (22:55 -0700)] 
[PATCH] sparc32 vga support

sparc32 lacks vga.h, so lots of fbdev drivers won't compile.  There are no
sparc32 systems with PCI slots, so it's a bit moot.

The patch gives sparc32 a copy of the sparc64 vga.h.  It fixes sparc32
allmodconfig without mucking up fbdev Kconfig and gives us wider compile
coverage.

Cc: "Antonino A. Daplas" <adaplas@pol.net>
Acked-by: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] atyfb is bust on sparc32
Andrew Morton [Tue, 11 Apr 2006 05:55:45 +0000 (22:55 -0700)] 
[PATCH] atyfb is bust on sparc32

Heaps of build errors - disable it to keep sparc32 allmodconfig happy.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] video/aty/atyfb_base.c: fix an off-by-one error
Adrian Bunk [Tue, 11 Apr 2006 05:55:45 +0000 (22:55 -0700)] 
[PATCH] video/aty/atyfb_base.c: fix an off-by-one error

Fix an obvious of-by-one error spotted by the Coverity checker.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: "Antonino A. Daplas" <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: grant delegations more frequently
NeilBrown [Tue, 11 Apr 2006 05:55:42 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: grant delegations more frequently

Keep unused openowners around for at least one lease period, to avoid the need
for as many open confirmations and to allow handing out more delegations.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: limit number of delegations handed out.
NeilBrown [Tue, 11 Apr 2006 05:55:41 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: limit number of delegations handed out.

It's very easy for the server to DOS itself by just giving out too many
delegations.

For now we just solve the problem with a dumb hard limit.  Eventually we'll
want a smarter policy.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: add missing rpciod_down()
NeilBrown [Tue, 11 Apr 2006 05:55:39 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: add missing rpciod_down()

We should be shutting down rpciod for the callback channel when we shut down
the server.

Also note that we do rpciod_up() and create the callback client *before*
setting cb_set--the cb_set only determines whether the initial null was
succesful.  So cb_set is not a reliable determiner of whether we need to clean
up, only cb_client is.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: nfsd4_probe_callback cleanup
NeilBrown [Tue, 11 Apr 2006 05:55:38 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: nfsd4_probe_callback cleanup

Some obvious cleanup.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: fix laundromat shutdown race
NeilBrown [Tue, 11 Apr 2006 05:55:37 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: fix laundromat shutdown race

We need to make sure the laundromat work doesn't reschedule itself just when
we try to cancel it.  Also, we shouldn't be waiting for it to finish running
while holding the state lock, as that's a potential deadlock.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: svcrpc: WARN() instead of returning an error from svc_take_page
J. Bruce Fields [Tue, 11 Apr 2006 05:55:36 +0000 (22:55 -0700)] 
[PATCH] knfsd: svcrpc: WARN() instead of returning an error from svc_take_page

Every caller of svc_take_page ignores its return value and assumes it
succeeded.  So just WARN() instead of returning an ignored error.  This would
have saved some time debugging a recent nfsd4 problem.

If there are still failure cases here, then the result is probably that we
overwrite an earlier part of the reply while xdr-encoding.

While the corrupted reply is a nasty bug, it would be worse to panic here and
create the possibility of a remote DOS; hence WARN() instead of BUG().

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Ingo Oeser <ioe-lkml@rameria.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: svcrpc: gss: don't call svc_take_page unnecessarily
NeilBrown [Tue, 11 Apr 2006 05:55:34 +0000 (22:55 -0700)] 
[PATCH] knfsd: svcrpc: gss: don't call svc_take_page unnecessarily

We're using svc_take_page here to get another page for the tail in case one
wasn't already allocated.  But there isn't always guaranteed to be another
page available.

Also fix a typo that made us check the tail buffer for space when we meant to
be checking the head buffer.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: fix corruption on readdir encoding with 64k pages
NeilBrown [Tue, 11 Apr 2006 05:55:33 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: fix corruption on readdir encoding with 64k pages

Fix corruption on readdir encoding with 64k pages.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: fix corruption of returned data when using 64k pages
NeilBrown [Tue, 11 Apr 2006 05:55:32 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: fix corruption of returned data when using 64k pages

In v4 we grab an extra page just for the padding of returned data.  The
formula that the rpc server uses to allocate pages for the response doesn't
take into account this extra page.

Instead of adjusting those formulae, we adopt the same solution as v2 and v3,
and put the "tail" data in the same page as the "head" data.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: remove nfsd_setuser from putrootfh
NeilBrown [Tue, 11 Apr 2006 05:55:31 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: remove nfsd_setuser from putrootfh

Since nfsd_setuser() is already called from any operation that uses the
current filehandle (because it's called from fh_verify), there's no reason to
call it from putrootfh.

Signed-off-by: Andy Adamson <andros@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd: nfsd_setuser doesn't really need to modify rqstp->rq_cred.
NeilBrown [Tue, 11 Apr 2006 05:55:30 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd: nfsd_setuser doesn't really need to modify rqstp->rq_cred.

In addition to setting the processes filesystem id's, nfsd_setuser also
modifies the value of the rq_cred which stores the id's that originally came
from the rpc call, for example to reflect root squashing.

There's no real reason to do that--the only case where rqstp->rq_cred is
actually used later on is in the NFSv4 SETCLIENTID/SETCLIENTID_CONFIRM
operations, and there the results are the opposite of what we want--those two
operations don't deal with the filesystem at all, they only record the
credentials used with the rpc call for later reference (so that we may require
the same credentials be used on later operations), and the credentials
shouldn't vary just because there was or wasn't a previous operation in the
compound that referred to some export

This fixes a bug which caused mounts from Solaris clients to fail.

Signed-off-by: Andy Adamson <andros@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd: oops exporting nonexistent directory
NeilBrown [Tue, 11 Apr 2006 05:55:27 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd: oops exporting nonexistent directory

Export a directory that does not exist:
exportfs -orw,fsid=0,insecure,no_subtree_check client:/home/NFS4

Try to mount from client with nfs4. Mount hangs (I'm not sure why -
that's another issue).

While client is hung, back on server

mkdir /home/NFS4

The server panics in dput.  I traced the problem back to svc_export_parse()
calling path_release() even though path_lookup() failed (it happens to fill in
the nameidata structure with a negative dentry - so the test after out:
succeeds).

After patching, an recreating the problem, the client mount still takes some
time before finally exiting with a message "couldn't read superblock".

Here is a simple patch to resolve this issue:

Signed-off-by: Frank Filz <ffilzlnx@us.ibm.com>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: fix acl xattr length return
NeilBrown [Tue, 11 Apr 2006 05:55:26 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: fix acl xattr length return

We should be using the length from the second vfs_getxattr, in case it
changed.  (Note: there's still a small race here; we could end up returning
-ENOMEM if the length increased between the first and second call.  I don't
know whether it's worth spending a lot of effort to fix that.)

This makes XFS ACLs usable on NFS exports, which they currently aren't, since
XFS appears to be returning a too-large value for vfs_getxattr() when it's
passed a NULL buffer.  So there's probably an XFS bug here too, though since
getxattr with a NULL buffer is usually used to decide how much memory to
allocate, it may be a fairly harmless bug in most cases.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: better nfs4acl errors
NeilBrown [Tue, 11 Apr 2006 05:55:25 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: better nfs4acl errors

We're returning -1 in a few places in the NFSv4<->POSIX acl translation code
where we could return a reasonable error.

Also allows some minor simplification elsewhere.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: nfsd4: Wrong error handling in nfs4acl
NeilBrown [Tue, 11 Apr 2006 05:55:24 +0000 (22:55 -0700)] 
[PATCH] knfsd: nfsd4: Wrong error handling in nfs4acl

this fixes coverity id #3.  Coverity detected dead code, since the == -1
comparison only returns 0 or 1 to error.  Therefore the if ( error < 0 )
statement was always false.  Seems that this was an if( error = nfs4...  )
statement some time ago, which got broken during cleanup.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fs/nfsd/nfs4state.c: make a struct static
Adrian Bunk [Tue, 11 Apr 2006 05:55:23 +0000 (22:55 -0700)] 
[PATCH] fs/nfsd/nfs4state.c: make a struct static

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Marc Eshel <eshel@almaden.ibm.com>
Cc: Andy Adamson <andros@citi.umich.edu>
Cc: J. Bruce Fields <bfields@citi.umich.edu>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: locks: flag NFSv4-owned locks
NeilBrown [Tue, 11 Apr 2006 05:55:22 +0000 (22:55 -0700)] 
[PATCH] knfsd: locks: flag NFSv4-owned locks

Use the fl_lmops field to identify which locks are ours, instead of trying to
look them up in our private hash.  This is safer and more efficient.

Earlier versions of this patch used a lock flag instead, but Trond pointed out
that adding a new flag for each lock manager wasn't going to scale well, and
suggested this approach instead; a separate patch converts lockd to using
fl_lmops in the same way.

In the NFSv4 case this looks like a bit of a hack, since the NFSv4 server
isn't currently actually defining a lock_manager_operations struct, so we end
up defining one *just* to serve as a cookie to identify our locks.

But it works, and we actually do expect to start using the
lock_manager_operations at some point anyway.

Signed-off-by: Marc Eshel <eshel@almaden.ibm.com>
Signed-off-by: Andy Adamson <andros@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Correct reserved reply space for read requests.
NeilBrown [Tue, 11 Apr 2006 05:55:20 +0000 (22:55 -0700)] 
[PATCH] knfsd: Correct reserved reply space for read requests.

NFSd makes sure there is enough space to hold the maximum possible reply
before accepting a request.  The units for this maximum is (4byte) words.
However in three places, particularly for read request, the number given is
a number of bytes.

This means too much space is reserved which is slightly wasteful.

This is the sort of patch that could uncover a deeper bug, and it is not
critical, so it would be best for it to spend a while in -mm before going
in to mainline.

(akpm: target 2.6.17-rc2, 2.6.16.3 (approx))

Discovered-by: "Eivind Sarto" <ivan@kasenna.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ISDN_DRV_GIGASET should select, not depend on CRC_CCITT
Adrian Bunk [Tue, 11 Apr 2006 05:55:19 +0000 (22:55 -0700)] 
[PATCH] ISDN_DRV_GIGASET should select, not depend on CRC_CCITT

CRC_CCITT is an internal helper function that should be select'ed.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn/gigaset/common.c: fix a memory leak
Adrian Bunk [Tue, 11 Apr 2006 05:55:19 +0000 (22:55 -0700)] 
[PATCH] isdn/gigaset/common.c: fix a memory leak

Fix a memory leak spotted by the Coverity checker if
(!try_module_get(owner)).

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/isdn/gigaset/common.c: small cleanups
Adrian Bunk [Tue, 11 Apr 2006 05:55:18 +0000 (22:55 -0700)] 
[PATCH] drivers/isdn/gigaset/common.c: small cleanups

- make the needlessly global gigaset_get_cs_by_tty() static
- remove the unused EXPORT_SYMBOL_GPL(gigaset_debugdrivers)

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: make some variables non-atomic
Tilman Schmidt [Tue, 11 Apr 2006 05:55:16 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: make some variables non-atomic

With Hansjoerg Lipp <hjlipp@web.de>

Replace some atomic_t variables in the Gigaset drivers by non-atomic ones,
using spinlocks instead to assure atomicity, as proposed in discussions on the
linux-kernel mailing list.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: add README
Tilman Schmidt [Tue, 11 Apr 2006 05:55:15 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: add README

With Hansjoerg Lipp <hjlipp@web.de>

Add a README file for the Siemens Gigaset drivers to the Documentation/isdn
directory.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: remove forward references
Tilman Schmidt [Tue, 11 Apr 2006 05:55:14 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: remove forward references

With Hansjoerg Lipp <hjlipp@web.de>

Remove four unnecessary forward function declarations and an obsolete E-mail
address from the Siemens Gigaset drivers.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: remove private version of __skb_put()
Tilman Schmidt [Tue, 11 Apr 2006 05:55:13 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: remove private version of __skb_put()

With Hansjoerg Lipp <hjlipp@web.de>

Remove the private version of __skb_put() from the Siemens Gigaset drivers.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: mutex conversion
Tilman Schmidt [Tue, 11 Apr 2006 05:55:12 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: mutex conversion

With Hansjoerg Lipp <hjlipp@web.de>

Convert the semaphores used by the Gigaset drivers to mutexes.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: eliminate from_user argument
Tilman Schmidt [Tue, 11 Apr 2006 05:55:11 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: eliminate from_user argument

With Hansjoerg Lipp <hjlipp@web.de>

Eliminate the from_user argument from a debugging function, thus easing the
job of sparse.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: uninline
Tilman Schmidt [Tue, 11 Apr 2006 05:55:09 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: uninline

With Hansjoerg Lipp <hjlipp@web.de>

Uninline a function which was slightly too big to warrant inlining.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: remove IFNULL macros
Tilman Schmidt [Tue, 11 Apr 2006 05:55:08 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: remove IFNULL macros

With Hansjoerg Lipp <hjlipp@web.de>

Remove the IFNULL debugging macros from the Gigaset drivers.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: sysfs usage
Tilman Schmidt [Tue, 11 Apr 2006 05:55:07 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: sysfs usage

With Hansjoerg Lipp <hjlipp@web.de>

Correct the way the Gigaset drivers create their sysfs entries.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: logging usage
Tilman Schmidt [Tue, 11 Apr 2006 05:55:04 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: logging usage

With Hansjoerg Lipp <hjlipp@web.de>

Improve error reporting of the Gigaset drivers, by using the
dev_err/dev_warn/dev_info macros from device.h instead of err/warn/info from
usb.h whereever possible.

Also rename the private dbg macro to gig_dbg in order to avoid confusion with
the macro of the same name in usb.h.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: timer usage
Tilman Schmidt [Tue, 11 Apr 2006 05:55:03 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: timer usage

With Hansjoerg Lipp <hjlipp@web.de>

Correct timer usage in the Gigaset drivers to take advantage of the existing
setup_timer() function, and use milliseconds as unit.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: Kconfig correction
Tilman Schmidt [Tue, 11 Apr 2006 05:55:02 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: Kconfig correction

With Hansjoerg Lipp <hjlipp@web.de>

Remove the restriction to build the Gigaset drivers as modules only.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] isdn4linux: Siemens Gigaset drivers: code cleanup
Tilman Schmidt [Tue, 11 Apr 2006 05:55:00 +0000 (22:55 -0700)] 
[PATCH] isdn4linux: Siemens Gigaset drivers: code cleanup

With Hansjoerg Lipp <hjlipp@web.de>

Source code formatting cleanups for the Siemens Gigaset drivers, such as line
length, comments, removal of unused declarations, and typo corrections.  It
does not introduce any functional changes.

Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: account background requests
Miklos Szeredi [Tue, 11 Apr 2006 05:54:59 +0000 (22:54 -0700)] 
[PATCH] fuse: account background requests

The previous patch removed limiting the number of outstanding requests.  This
patch adds a much simpler limiting, that is also compatible with file locking
operations.

A task may have at most one synchronous request allocated.  So these requests
need not be otherwise limited.

However the number of background requests (release, forget, asynchronous
reads, interrupted requests) can grow indefinitely.  This can be used by a
malicous user to cause FUSE to allocate arbitrary amounts of unswappable
kernel memory, denying service.

For this reason add a limit for the number of background requests, and block
allocations of new requests until the number goes bellow the limit.

Also use this mechanism to block all requests until the INIT reply is
received.

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: clean up request accounting
Miklos Szeredi [Tue, 11 Apr 2006 05:54:58 +0000 (22:54 -0700)] 
[PATCH] fuse: clean up request accounting

FUSE allocated most requests from a fixed size pool filled at mount time.
However in some cases (release/forget) non-pool requests were used.  File
locking operations aren't well served by the request pool, since they may
block indefinetly thus exhausting the pool.

This patch removes the request pool and always allocates requests on demand.

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: consolidate device errors
Miklos Szeredi [Tue, 11 Apr 2006 05:54:56 +0000 (22:54 -0700)] 
[PATCH] fuse: consolidate device errors

Return consistent error values for the case when the opened device file has no
mount associated yet.

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: use a per-mount spinlock
Miklos Szeredi [Tue, 11 Apr 2006 05:54:55 +0000 (22:54 -0700)] 
[PATCH] fuse: use a per-mount spinlock

Remove the global spinlock in favor of a per-mount one.

This patch is basically find & replace.  The difficult part has already been
done by the previous patch.

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: simplify locking
Miklos Szeredi [Tue, 11 Apr 2006 05:54:55 +0000 (22:54 -0700)] 
[PATCH] fuse: simplify locking

This is in preparation for removing the global spinlock in favor of a
per-mount one.

The only critical part is the interaction between fuse_dev_release() and
fuse_fill_super(): fuse_dev_release() must see the assignment to
file->private_data, otherwise it will leak the reference to fuse_conn.

This is ensured by the fput() operation, which will synchronize the assignment
with other CPU's that may do a final fput() soon after this.

Also redundant locking is removed from fuse_fill_super(), where exclusion is
already ensured by the BKL held for this function by the VFS.

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: add O_NONBLOCK support to FUSE device
Jeff Dike [Tue, 11 Apr 2006 05:54:53 +0000 (22:54 -0700)] 
[PATCH] fuse: add O_NONBLOCK support to FUSE device

I don't like duplicating the connected and list_empty tests in fuse_dev_readv,
but this seemed cleaner than adding the f_flags test to request_wait.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: add O_ASYNC support to FUSE device
Jeff Dike [Tue, 11 Apr 2006 05:54:52 +0000 (22:54 -0700)] 
[PATCH] fuse: add O_ASYNC support to FUSE device

This adds asynchronous notification to FUSE - a FUSE server can request
O_ASYNC on a /dev/fuse file descriptor and receive SIGIO when there is input
available.

One subtlety - fuse_dev_fasync, which is called when O_ASYNC is requested,
does no locking, unlink the other methods.  I think it's unnecessary, as the
fuse_conn.fasync list is manipulated only by fasync_helper and kill_fasync,
which provide their own locking.  It would also be wrong to use the fuse_lock,
as it's a spin lock and fasync_helper can sleep.  My one concern with this is
the fuse_conn going away underneath fuse_dev_fasync - sys_fcntl takes a
reference on the file struct, so this seems not to be a problem.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: fix fuse_dev_poll() return value
Miklos Szeredi [Tue, 11 Apr 2006 05:54:50 +0000 (22:54 -0700)] 
[PATCH] fuse: fix fuse_dev_poll() return value

fuse_dev_poll() returned an error value instead of a poll mask.  Luckily (or
unluckily) -ENODEV does contain the POLLERR bit.

There's also a race if filesystem is unmounted between fuse_get_conn() and
spin_lock(), in which case this event will be missed by poll().

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fuse: fix oops in fuse_send_readpages()
Miklos Szeredi [Tue, 11 Apr 2006 05:54:49 +0000 (22:54 -0700)] 
[PATCH] fuse: fix oops in fuse_send_readpages()

During heavy parallel filesystem activity it was possible to Oops the kernel.
The reason is that read_cache_pages() could skip pages which have already been
inserted into the cache by another task.  Occasionally this may result in zero
pages actually being sent, while fuse_send_readpages() relies on at least one
page being in the request.

So check this corner case and just free the request instead of trying to send
it.

Reported and tested by Konstantin Isakov.

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: VR41XX cleanup
Alessandro Zummo [Tue, 11 Apr 2006 05:54:48 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: VR41XX cleanup

Clean up kconfig entry for the rtc-vr41xx.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: VR41XX driver
Yoichi Yuasa [Tue, 11 Apr 2006 05:54:47 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: VR41XX driver

This patch updates VR4100 series RTC driver.

* This driver supports new RTC subsystem.
* Simple set time/read time test worked fine.

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: SA1100 cleanup
Alessandro Zummo [Tue, 11 Apr 2006 05:54:46 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: SA1100 cleanup

- convert printks to dev_xxx

- remove messages in excess

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: compact error messages
Alessandro Zummo [Tue, 11 Apr 2006 05:54:45 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: compact error messages

Move registration error message from drivers to core.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: RS5C372 sysfs fix
Alessandro Zummo [Tue, 11 Apr 2006 05:54:44 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: RS5C372 sysfs fix

Fix sysfs show() return code

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: fix proc output
Alessandro Zummo [Tue, 11 Apr 2006 05:54:43 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: fix proc output

Move the "24hr: yes" proc output from drivers to rtc proc code.  This is
required because the time value in the proc output is always in 24hr mode
regardless of the driver.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: whitespaces and error messages cleanup
Alessandro Zummo [Tue, 11 Apr 2006 05:54:42 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: whitespaces and error messages cleanup

- fix whitespace

- remove some debugging in excess

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: X1205 sysfs cleanup
Alessandro Zummo [Tue, 11 Apr 2006 05:54:42 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: X1205 sysfs cleanup

Fix sysfs show() return code

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: DS1672 cleanup
Alessandro Zummo [Tue, 11 Apr 2006 05:54:41 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: DS1672 cleanup

 - removed a duplicate error message
 - bumped driver version
 - removed some debugging messages in excess
 - refined the formatting
 - adjusted copyright notice

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: DS1672 oscillator handling
Kumar Gala [Tue, 11 Apr 2006 05:54:39 +0000 (22:54 -0700)] 
[PATCH] RTC subsystem: DS1672 oscillator handling

* Always enable the oscillator when we set the time
* If the oscillator is disable when we probe the RTC report back a warning
  to the user
* Added sysfs attribute to represent the state of the oscillator

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ext3: Fix missed mutex unlock
Ananiev, Leonid I [Tue, 11 Apr 2006 05:54:38 +0000 (22:54 -0700)] 
[PATCH] ext3: Fix missed mutex unlock

Missed unlock_super()call is added in error condition code path.

Signed-off-by: Leonid Ananiev <leonid.i.ananiev@intel.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Overrun in cdrom/aztcd.c
Eric Sesterhenn [Tue, 11 Apr 2006 05:54:37 +0000 (22:54 -0700)] 
[PATCH] Overrun in cdrom/aztcd.c

This fixes coverity bug id #473.  After the for loop i==16 if we didn't find a
cdrom.  So we should check for i==16 first before checking the array element.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Wrong out of range check in drivers/char/applicom.c
Eric Sesterhenn [Tue, 11 Apr 2006 05:54:36 +0000 (22:54 -0700)] 
[PATCH] Wrong out of range check in drivers/char/applicom.c

This fixes coverity bug id #469.  The out of range check didnt work as
intended, as seen by the printk(), which states that boardno has to be 1 <=
boardno <= MAX_BOARD.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>