]> err.no Git - linux-2.6/log
linux-2.6
18 years agoMerge branch 'merge'
Paul Mackerras [Sun, 23 Apr 2006 00:55:56 +0000 (10:55 +1000)]
Merge branch 'merge'

18 years agopowerpc: Fix define_machine so machine_is() works from modules
Paul Mackerras [Sun, 23 Apr 2006 00:42:04 +0000 (10:42 +1000)]
powerpc: Fix define_machine so machine_is() works from modules

machine_is() was always returning 0 when used in a module, because
we weren't exporting the machine definitions.  This was why sound
wasn't working on powermacs when CONFIG_SND_POWERMAC=m.  Original
fix from Ben Herrenschmidt, further fixed by me.

Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'for_paulus' of master.kernel.org:/pub/scm/linux/kernel/git/galak/powerpc
Paul Mackerras [Sat, 22 Apr 2006 09:47:47 +0000 (19:47 +1000)]
Merge branch 'for_paulus' of master.kernel.org:/pub/scm/linux/kernel/git/galak/powerpc

18 years agoMerge branch 'merge'
Paul Mackerras [Sat, 22 Apr 2006 09:47:24 +0000 (19:47 +1000)]
Merge branch 'merge'

18 years agoMerge branch 'misc' of master.kernel.org:/pub/scm/linux/kernel/git/galak/powerpc...
Paul Mackerras [Sat, 22 Apr 2006 09:19:06 +0000 (19:19 +1000)]
Merge branch 'misc' of master.kernel.org:/pub/scm/linux/kernel/git/galak/powerpc into merge

18 years ago[PATCH] powerpc/pseries: clear PCI failure counter if no new failures
Linas Vepstas [Wed, 19 Apr 2006 04:05:21 +0000 (21:05 -0700)]
[PATCH] powerpc/pseries: clear PCI failure counter if no new failures

The current PCI error recovery system keeps track of the number of PCI card
resets, and refuses to bring a card back up if this number is too large.
The goal of doing this was to avoid an infinite loop of resets if a card is
obviously dead.  However, if the failures are rare, but the machine has a
high uptime, this mechanism might still be triggered; this is too harsh.

This patch will avoids this problem by decrementing the fail count after an
hour.  Thus, as long as a pci card BSOD's less than 6 times an hour, it
will continue to be reset indefinitely.  If it's failure rate is greater
than that, it will be taken off-line permanently.

This patch is larger than it might otherwise be because it changes
indentation by removing a pointless while-loop.  The while loop is not
needed, as the handler is invoked once fo each event (by schedule_work());
the loop is leftover cruft from an earlier implementation.

Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Remove stale iseries global
Olof Johansson [Tue, 18 Apr 2006 16:25:53 +0000 (11:25 -0500)]
[PATCH] powerpc: Remove stale iseries global

Not even the iSeries maintainer seems to have access to this legendary
piranha simulator. It adds a bit of ugliness in the common time init
code, and if it's no longer used we might as well be done with it and
remove the bloat.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet oprofile output at boot
Olof Johansson [Wed, 12 Apr 2006 20:29:00 +0000 (15:29 -0500)]
[PATCH] powerpc: Quiet oprofile output at boot

No need to always print out which performance monitoring type is used
on the console at every boot.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet rtasd output at boot
Olof Johansson [Wed, 12 Apr 2006 20:28:13 +0000 (15:28 -0500)]
[PATCH] powerpc: Quiet rtasd output at boot

Most users won't really know the difference between a started RTAS
daemon and a missing event-scan. Move it to debug levels.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet PCI init printouts
Olof Johansson [Wed, 12 Apr 2006 20:26:59 +0000 (15:26 -0500)]
[PATCH] powerpc: Quiet PCI init printouts

Quiet some of the more debug related output from the pci probe routines.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Lack of ISA interrupts on XICS isn't dangerous
Olof Johansson [Wed, 12 Apr 2006 20:26:00 +0000 (15:26 -0500)]
[PATCH] powerpc: Lack of ISA interrupts on XICS isn't dangerous

This isn't really a dangerous thing any more; most systems lack
ISA interrupt controllers.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Less verbose mem configuration output
Olof Johansson [Wed, 12 Apr 2006 20:25:01 +0000 (15:25 -0500)]
[PATCH] powerpc: Less verbose mem configuration output

Quieten some of the debug ram config output. we already print out available
memory at KERN_INFO level.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Don't print chosen idle loop at every boot
Olof Johansson [Wed, 12 Apr 2006 20:23:22 +0000 (15:23 -0500)]
[PATCH] powerpc: Don't print chosen idle loop at every boot

No need to write out what idle loop is used on every boot.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet VETH version printk
Olof Johansson [Wed, 12 Apr 2006 20:22:29 +0000 (15:22 -0500)]
[PATCH] powerpc: Quiet VETH version printk

ibmveth_printk() is only used to print the driver version when the module
initializes, which means on all machines as long as it's compiled in.

If it's really only needed for debugging, boot with loglevel=8, or get
it from dmesg instead.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet page order output
Olof Johansson [Wed, 12 Apr 2006 20:21:06 +0000 (15:21 -0500)]
[PATCH] powerpc: Quiet page order output

No need to always print page orders.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet time init output
Olof Johansson [Wed, 12 Apr 2006 20:20:27 +0000 (15:20 -0500)]
[PATCH] powerpc: Quiet time init output

Move time_init console output to KERN_DEBUG prink level. No need to
print it at every boot.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Quiet HVSI boot output
Olof Johansson [Wed, 12 Apr 2006 20:19:50 +0000 (15:19 -0500)]
[PATCH] powerpc: Quiet HVSI boot output

There's no real need to print the number of found HVSI devices on the
console at every boot.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: remove io_page_mask
Anton Blanchard [Fri, 7 Apr 2006 05:23:03 +0000 (15:23 +1000)]
[PATCH] powerpc: remove io_page_mask

Cleanup patch which removes the io_page_mask.  It fixes the reset on
some e1000 devices which is needed for clean kexec reboots.  The legacy
devices which broke with this patch (parallel port and PC speaker) have
now been fixed in Linus' tree.

Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: clear IPIs on kdump
Haren Myneni [Thu, 6 Apr 2006 03:10:18 +0000 (21:10 -0600)]
[PATCH] powerpc: clear IPIs on kdump

In some crash scenarios, the kexec CPU is not responding to an IPI sent by
secondary CPU after init thread is forked, causing the system to drop into
xmon during kdump boot.  This problem can be reproduced each time when the
debugger is enabled and soft-reset is used to invoke kdump boot. The first
CPU sends an IPI - setting the IPI priority for all secondary cpus
(xics_cause_ipi()). But some CPUs will enter into the xmon via soft-reset,
i.e, not executing xics_ipi_action(). Hence, IPI is not cleared. When
exited from the debugger, one of these CPUs could become the primary kexec
CPU. Since the IPI is not cleared, causing this issue in kdump boot. This
patch clears and EOI IPI for kexec CPU as well before the kdump boot
started.

Signed-off-by: Haren Myneni <haren@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Disable and EOI interrupts in machine_crash_shutdown()
Michael Ellerman [Tue, 4 Apr 2006 11:43:01 +0000 (13:43 +0200)]
[PATCH] powerpc: Disable and EOI interrupts in machine_crash_shutdown()

We've seen several bugs caused by interrupt weirdness in the kdump kernel.
Panicking from an interrupt handler means we fail to EOI the interrupt, and
so the second kernel never gets that interrupt ever again. We also see hangs
on JS20 where we take interrupts in the second kernel early during boot.

This patch fixes both those problems, and although it adds more code to the
crash path I think it is the best solution.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: fix oops in alsa powermac driver
Benjamin Herrenschmidt [Fri, 21 Apr 2006 05:04:22 +0000 (15:04 +1000)]
[PATCH] powerpc: fix oops in alsa powermac driver

This fixes an oops in 2.6.16.X when loading the snd_powermac module. The
name of the requested module changed during the 2.6.16 development cycle
from i2c-keylargo to i2c-powermac.

Signed-off-by: Guido Guenther <agx@sigxcpu.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: update {g5,iseries,pseries}_defconfigs
Will Schmidt [Wed, 19 Apr 2006 21:26:22 +0000 (16:26 -0500)]
[PATCH] powerpc: update {g5,iseries,pseries}_defconfigs

Default choices for all.

Signed-off-by: Will Schmidt <willschm@us.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] ppc: Fix powersave code on arch/ppc
Becky Bruce [Tue, 18 Apr 2006 19:29:34 +0000 (14:29 -0500)]
[PATCH] ppc: Fix powersave code on arch/ppc

Fix asm_offsets.c and entry.S to work with the new power save code.
Changes in arch/powerpc needed to exist in arch/ppc as well since the
idle code is shared by both ppc and powerpc..

Signed-off-by: Becky Bruce <becky.bruce@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc/cell: remove BUILD_BUG_ON and add sys_tee to spu_syscall_table
Arnd Bergmann [Tue, 18 Apr 2006 13:24:16 +0000 (15:24 +0200)]
[PATCH] powerpc/cell: remove BUILD_BUG_ON and add sys_tee to spu_syscall_table

Every time a new syscall gets added, a BUILD_BUG_ON in
arch/powerpc/platforms/cell/spu_callbacks.c gets triggered.
Since the addition of a new syscall is rather harmless,
the error should just be removed.

While we're here, add sys_tee to the list and add a comment
to systbl.S to remind people that there is another list
on powerpc.

Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powermac: Fix i2c on keywest based chips
Benjamin Herrenschmidt [Tue, 18 Apr 2006 04:11:53 +0000 (14:11 +1000)]
[PATCH] powermac: Fix i2c on keywest based chips

The new i2c implementation for PowerMac has a regression that causes the
hardware to go out of state when probing non-existent devices. While
fixing that, I also found & fixed a couple of other corner cases. This
fixes booting with a pbbuttons version that scans the i2c bus for an LMU
controller among others. Tested on a dual G5 with thermal control (which
has heavy i2c activity) with no problem so far.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: Lower threshold for DART enablement to 1GB
Olof Johansson [Thu, 13 Apr 2006 02:52:33 +0000 (21:52 -0500)]
[PATCH] powerpc: Lower threshold for DART enablement to 1GB

Turn on the DART already at 1GB. This is needed because of crippled
devices in some systems, i.e. Airport Extreme cards, only supporting
30-bit DMA addresses.

Otherwise, users with between 1 and 2GB of memory will need to manually
enable it with iommu=force, and that's no good.

Some simple performance tests show that there's a slight impact of
enabling DART, but it's in the 1-3% range (kernel build with disk I/O
as well as over NFS).

iommu=off can still be used for those who don't want to deal with the
overhead (and don't need it for any devices).

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years ago[PATCH] powerpc: IOMMU support for honoring dma_mask
Olof Johansson [Thu, 13 Apr 2006 02:05:59 +0000 (21:05 -0500)]
[PATCH] powerpc: IOMMU support for honoring dma_mask

Some devices don't support full 32-bit DMA address space, which we currently
assume. Add the required mask-passing to the IOMMU allocators.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>
18 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Thu, 20 Apr 2006 22:26:25 +0000 (15:26 -0700)]
Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6: (21 commits)
  [PATCH] wext: Fix RtNetlink ENCODE security permissions
  [PATCH] bcm43xx: iw_priv_args names should be <16 characters
  [PATCH] bcm43xx: sysfs code cleanup
  [PATCH] bcm43xx: fix pctl slowclock limit calculation
  [PATCH] bcm43xx: fix dyn tssi2dbm memleak
  [PATCH] bcm43xx: fix config menu alignment
  [PATCH] bcm43xx wireless: fix printk format warnings
  [PATCH] softmac: report when scanning has finished
  [PATCH] softmac: fix event sending
  [PATCH] softmac: handle iw_mode properly
  [PATCH] softmac: dont send out packets while scanning
  [PATCH] softmac: return -EAGAIN from getscan while scanning
  [PATCH] bcm43xx: set trans_start on TX to prevent bogus timeouts
  [PATCH] orinoco: fix truncating commsquality RID with the latest Symbol firmware
  [PATCH] softmac: fix spinlock recursion on reassoc
  [PATCH] Revert NET_RADIO Kconfig title change
  [PATCH] wext: Fix IWENCODEEXT security permissions
  [PATCH] wireless/atmel: send WEXT scan completion events
  [PATCH] wireless/airo: clean up WEXT association and scan events
  [PATCH] softmac uses Wiress Ext.
  ...

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Thu, 20 Apr 2006 22:25:37 +0000 (15:25 -0700)]
Merge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [EBTABLES]: Clean up vmalloc usage in net/bridge/netfilter/ebtables.c
  [NET]: Add skb->truesize assertion checking.
  [TCP]: Account skb overhead in tcp_fragment
  [SUNGEM]: Marvell PHY suspend.
  [LLC]: Use pskb_trim_rcsum() in llc_fixup_skb().
  [NET]: sockfd_lookup_light() returns random error for -EBADFD

18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Thu, 20 Apr 2006 22:22:41 +0000 (15:22 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] for_each_possible_cpu
  [ARM] add_memory() build fix
  [ARM] 3483/1: ixp23xx: update defconfig to 2.6.17-rc2
  [ARM] 3482/1: ixp2000: update defconfig to 2.6.17-rc2
  [ARM] 3481/1: ep93xx: update defconfig to 2.6.17-rc2
  [ARM] 3480/1: ixp4xx: fix irq2gpio array type

18 years agoMerge branch 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/linvil...
Jeff Garzik [Thu, 20 Apr 2006 21:36:10 +0000 (17:36 -0400)]
Merge branch 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6

18 years ago[PATCH] NEx000: fix RTL8019AS base address for RBTX4938
Sergei Shtylyov [Sun, 16 Apr 2006 15:42:35 +0000 (19:42 +0400)]
[PATCH] NEx000: fix RTL8019AS base address for RBTX4938

Correct the base address of the Realtek RTL8019AS chip on the Toshiba RBTX4938
board -- this should make the driver work at least when CONFIG_PCI is enabled.

Signed-off-by: Yuri Shpilevsky <yshpilevsky@ru.mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[ARM] for_each_possible_cpu
KAMEZAWA Hiroyuki [Thu, 20 Apr 2006 20:48:02 +0000 (21:48 +0100)]
[ARM] for_each_possible_cpu

for_each_cpu() actually iterates across all possible CPUs.  We've had mistakes
in the past where people were using for_each_cpu() where they should have been
iterating across only online or present CPUs.  This is inefficient and
possibly buggy.

We're renaming for_each_cpu() to for_each_possible_cpu() to avoid this in the
future.

This patch replaces for_each_cpu with for_each_possible_cpu.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] add_memory() build fix
Andrew Morton [Thu, 20 Apr 2006 20:41:18 +0000 (21:41 +0100)]
[ARM] add_memory() build fix

This is back again.  Offending patch is x86_64-mm-hotadd-reserve.patch

arch/arm/kernel/setup.c:435: error: conflicting types for 'add_memory'
include/linux/memory_hotplug.h:102: error: previous declaration of 'add_memory' was here
arch/arm/kernel/setup.c:435: error: conflicting types for 'add_memory'
include/linux/memory_hotplug.h:102: error: previous declaration of 'add_memory' was here

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3483/1: ixp23xx: update defconfig to 2.6.17-rc2
Lennert Buytenhek [Thu, 20 Apr 2006 20:24:41 +0000 (21:24 +0100)]
[ARM] 3483/1: ixp23xx: update defconfig to 2.6.17-rc2

Patch from Lennert Buytenhek

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3482/1: ixp2000: update defconfig to 2.6.17-rc2
Lennert Buytenhek [Thu, 20 Apr 2006 20:24:40 +0000 (21:24 +0100)]
[ARM] 3482/1: ixp2000: update defconfig to 2.6.17-rc2

Patch from Lennert Buytenhek

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3481/1: ep93xx: update defconfig to 2.6.17-rc2
Lennert Buytenhek [Thu, 20 Apr 2006 20:24:39 +0000 (21:24 +0100)]
[ARM] 3481/1: ep93xx: update defconfig to 2.6.17-rc2

Patch from Lennert Buytenhek

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3480/1: ixp4xx: fix irq2gpio array type
Lennert Buytenhek [Thu, 20 Apr 2006 20:24:38 +0000 (21:24 +0100)]
[ARM] 3480/1: ixp4xx: fix irq2gpio array type

Patch from Lennert Buytenhek

The irq2gpio array was recently converted from an array of ints to an
array of chars (by patch 3368/1.)  However, this array contains elements
that are -1, and on ARM, the char type is unsigned by default, so this
patch broke the GPIO check in ixp4xx_set_irq_type.

Change the 'char' to be a 'signed char' to fix this.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years agoMerge branch 'for-linus' of git://brick.kernel.dk/data/git/linux-2.6-block
Linus Torvalds [Thu, 20 Apr 2006 15:17:04 +0000 (08:17 -0700)]
Merge branch 'for-linus' of git://brick.kernel.dk/data/git/linux-2.6-block

* 'for-linus' of git://brick.kernel.dk/data/git/linux-2.6-block:
  [PATCH] block/elevator.c: remove unused exports
  [PATCH] splice: fix smaller sized splice reads
  [PATCH] Don't inherit ->splice_pipe across forks
  [patch] cleanup: use blk_queue_stopped
  [PATCH] Document online io scheduler switching

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Thu, 20 Apr 2006 14:58:58 +0000 (07:58 -0700)]
Merge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC]: __NR_sys_splice --> __NR_splice

18 years ago[PATCH] x86_64: bring back __read_mostly support to linux-2.6.17-rc2
Eric Dumazet [Thu, 20 Apr 2006 00:36:48 +0000 (02:36 +0200)]
[PATCH] x86_64: bring back __read_mostly support to linux-2.6.17-rc2

It seems latest kernel has a wrong/missing __read_mostly implementation
for x86_64

__read_mostly macro should be declared outside of #if CONFIG_X86_VSMP block

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386/x86-64: Fix x87 information leak between processes
Andi Kleen [Thu, 20 Apr 2006 00:36:45 +0000 (02:36 +0200)]
[PATCH] i386/x86-64: Fix x87 information leak between processes

AMD K7/K8 CPUs only save/restore the FOP/FIP/FDP x87 registers in FXSAVE
when an exception is pending.  This means the value leak through
context switches and allow processes to observe some x87 instruction
state of other processes.

This was actually documented by AMD, but nobody recognized it as
being different from Intel before.

The fix first adds an optimization: instead of unconditionally
calling FNCLEX after each FXSAVE test if ES is pending and skip
it when not needed. Then do a x87 load from a kernel variable to
clear FOP/FIP/FDP.

This means other processes always will only see a constant value
defined by the kernel in their FP state.

I took some pain to make sure to chose a variable that's already
in L1 during context switch to make the overhead of this low.

Also alternative() is used to patch away the new code on CPUs
who don't need it.

Patch for both i386/x86-64.

The problem was discovered originally by Jan Beulich. Richard
Brunner provided the basic code for the workarounds, with contribution
from Jan.

This is CVE-2006-1056

Cc: richard.brunner@amd.com
Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] md: locking fix
Ingo Molnar [Thu, 20 Apr 2006 09:43:23 +0000 (02:43 -0700)]
[PATCH] md: locking fix

- fix mddev_lock() usage bugs in md_attr_show() and md_attr_store().
  [they did not anticipate the possibility of getting a signal]

- remove mddev_lock_uninterruptible() [unused]

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-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] memory_hotplug.h cleanup
Adrian Bunk [Thu, 20 Apr 2006 09:43:23 +0000 (02:43 -0700)]
[PATCH] memory_hotplug.h cleanup

We don't have to #if guard prototypes.

This also fixes a bug observed by Randy Dunlap due to a misspelled
option in the #if.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] make the OSS SOUND_VIA82CXXX option available again
Adrian Bunk [Thu, 20 Apr 2006 09:43:22 +0000 (02:43 -0700)]
[PATCH] make the OSS SOUND_VIA82CXXX option available again

There was a report of a regression in the ALSA driver for the same
hardware.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] update OBSOLETE_OSS_DRIVER schedule and dependencies
Adrian Bunk [Thu, 20 Apr 2006 09:43:21 +0000 (02:43 -0700)]
[PATCH] update OBSOLETE_OSS_DRIVER schedule and dependencies

Update the schedule for the removal of drivers depending on
OBSOLETE_OSS_DRIVER as follows:

- adjust OBSOLETE_OSS_DRIVER dependencie
- from the release of 2.6.16 till the release of 2.6.17:
  approx. two months for users to report problems with the ALSA
  drivers for the same hardware
- after the release of 2.6.17 (and before 2.6.18):
  remove the subset of drivers marked at OBSOLETE_OSS_DRIVER without
  known regressions in the ALSA drivers for the same hardware

Additionally, correct some OBSOLETE_OSS_DRIVER dependencies.
A rationale of the changes is in
  http://lkml.org/lkml/2006/1/28/135

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sonypi: correct detection of new ICH7-based laptops
Arnaud MAZIN [Thu, 20 Apr 2006 09:43:20 +0000 (02:43 -0700)]
[PATCH] sonypi: correct detection of new ICH7-based laptops

Add a test to detect the ICH7 based Core Duo SONY laptops (such as the SZ1)
as type3 models.

Signed-off-by: Arnaud MAZIN < arnaud.mazin@gmail.com>
Acked-by: Stelian Pop <stelian@poppies.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sound: fix hang in mpu401_uart.c
Jon Masters [Thu, 20 Apr 2006 09:43:20 +0000 (02:43 -0700)]
[PATCH] sound: fix hang in mpu401_uart.c

This fixes a hang in mpu401_uart.c that can occur when the mpu401 interface
is non-existent or otherwise doesn't respond to commands but we issue IO
anyway.  snd_mpu401_uart_cmd now returns an error code that is passed up
the stack so that an open() will fail immediately in such cases.

Eventually discovered after wine/cxoffice would constantly cause hard
lockups on my desktop immediately after loading (emulating Windows too
well).  Turned out that I'd recently moved my sound cards around and using
/dev/sequencer now talks to a sound card with a broken MPU.

This second version changes -EFAULT to -EIO and frees open resources on
error too.  Test booted and seems to work ok.

Signed-off-by: Jon Masters <jcm@jonmasters.org>
Cc: Jaroslav Kysela <perex@suse.cz>
Acked-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i2c-i801: Fix resume when PEC is used
Jean Delvare [Thu, 20 Apr 2006 09:43:18 +0000 (02:43 -0700)]
[PATCH] i2c-i801: Fix resume when PEC is used

Fix for bug #6395:

Fail to resume on Tecra M2 with ADM1032 and Intel 82801DBM

The BIOS of the Tecra M2 doesn't like it when it has to reboot or resume
after the i2c-i801 driver has left the SMBus in PEC mode.  The most simple
fix is to clear the PEC bit after after every transaction.  That's what
this driver was doing up to 2.6.15 (inclusive).

Thanks to Daniele Gaffuri for the very good report.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] MSI-K8T-Neo2-Fir OnboardSound and additional Soundcard
Johannes Goecke [Thu, 20 Apr 2006 09:43:17 +0000 (02:43 -0700)]
[PATCH] MSI-K8T-Neo2-Fir OnboardSound and additional Soundcard

On the MSI-K8T-NEO2 FIR ( Athlon-64, Socket 939 with VIA-K8T800- Chipset
and onboard Sound,...  ) the BIOS lets you choose "DISABLED" or "AUTO" for
the On-Board Sound Device.

If you add another PCI-Sound-Card the BIOS disables the on-board device.

So far I have a Quirk, that does set the correspondent BIT in the
PCI-registers to enable the soundcard.

But how to ensure that the code is executed ONLY on excactly this kind of
boards (not any other with similar Chipset)?

Cc: Jaroslav Kysela <perex@suse.cz>
Acked-by: Takashi Iwai <tiwai@suse.de>
Cc: Lee Revell <rlrevell@joe-job.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] xtensa: Fix TIOCGICOUNT macro
Chris Zankel [Thu, 20 Apr 2006 09:43:15 +0000 (02:43 -0700)]
[PATCH] xtensa: Fix TIOCGICOUNT macro

Remove the dependence on the async_icount structure in the TIOCGICOUNT
macro for Xtensa.  (Thanks Russell and Adrian for pointing this out)

Signed-off-by: Chris Zankel <chris@zankel.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] cs5535_gpio.c: call cdev_del() during module_exit to unmap kobject references...
Thayumanavar Sachithanantham [Thu, 20 Apr 2006 09:43:15 +0000 (02:43 -0700)]
[PATCH] cs5535_gpio.c: call cdev_del() during module_exit to unmap kobject references and other cleanups

During module unloading, cdev_del() must be called to unmap cdev related
kobject references and other cleanups(such as inode->i_cdev being set to
NULL) which prevents the OOPS upon subsequent loading, usage and unloading
of modules(as seen in the mail thread
http://marc.theaimsgroup.com/?l=linux-kernel&m=114533640609018&w=2).

Also, remove unneeded test of gpio_base.

Signed-off-by: Thayumanavar Sachithanantham <thayumk@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fix section mismatch in pm2fb.o
Darren Jenkins [Thu, 20 Apr 2006 09:43:13 +0000 (02:43 -0700)]
[PATCH] fix section mismatch in pm2fb.o

WARNING: drivers/video/pm2fb.o - Section mismatch: reference
to .init.data: from .text after 'pm2fb_set_par' (at offset 0xd5d)
WARNING: drivers/video/pm2fb.o - Section mismatch: reference
to .init.data: from .text after 'pm2fb_set_par' (at offset 0xd82)

They are caused because pm2fb_set_par() uses lowhsync and lowvsync which
are marked __devinitdata.

Signed-off-by: Darren Jenkins <darrenrjenkins@gmail.com>
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] Remove cond_resched in gather_stats()
Christoph Lameter [Thu, 20 Apr 2006 09:43:12 +0000 (02:43 -0700)]
[PATCH] Remove cond_resched in gather_stats()

gather_stats() is called with a spinlock held from check_pte_range.  We
cannot reschedule with a lock held.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] IPMI maintainer
Corey Minyard [Thu, 20 Apr 2006 09:43:12 +0000 (02:43 -0700)]
[PATCH] IPMI maintainer

Add myself as the IPMI maintainer.

Signed-off-by: Corey Minyard <minyard@acm.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] kprobes: NULL out non-relevant fields in struct kretprobe
Ananth N Mavinakayanahalli [Thu, 20 Apr 2006 09:43:11 +0000 (02:43 -0700)]
[PATCH] kprobes: NULL out non-relevant fields in struct kretprobe

In cases where a struct kretprobe's *_handler fields are non-NULL, it is
possible to cause a system crash, due to the possibility of calls ending up
in zombie functions.  Documentation clearly states that unused *_handlers
should be set to NULL, but kprobe users sometimes fail to do so.

Fix it by setting the non-relevant fields of the struct kretprobe to NULL.

Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Acked-by: Jim Keniston <jkenisto@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] block/elevator.c: remove unused exports
Adrian Bunk [Thu, 20 Apr 2006 13:45:22 +0000 (15:45 +0200)]
[PATCH] block/elevator.c: remove unused exports

This patch removes the following unused EXPORT_SYMBOL's:
- elv_requeue_request
- elv_completed_request

They are only used by the block core, hence they need not be exported.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] splice: fix smaller sized splice reads
Jens Axboe [Thu, 20 Apr 2006 11:05:48 +0000 (13:05 +0200)]
[PATCH] splice: fix smaller sized splice reads

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] Don't inherit ->splice_pipe across forks
Jens Axboe [Thu, 20 Apr 2006 11:05:33 +0000 (13:05 +0200)]
[PATCH] Don't inherit ->splice_pipe across forks

It's really task private, so clear that field on fork after copying
task structure.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[patch] cleanup: use blk_queue_stopped
Coywolf Qi Hunt [Wed, 19 Apr 2006 08:14:49 +0000 (10:14 +0200)]
[patch] cleanup: use blk_queue_stopped

This cleanup the source to use blk_queue_stopped.

Signed-off-by: Coywolf Qi Hunt <qiyong@freeforge.net>
Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] Document online io scheduler switching
Valdis Kletnieks [Wed, 19 Apr 2006 07:23:09 +0000 (09:23 +0200)]
[PATCH] Document online io scheduler switching

We added the ability to change a block device's IO elevator scheduler both
at kernel boot and on-the-fly, but we only documented the elevator= boot
parameter.  Add a quick how-to on doing it on the fly.

Signed-off-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[EBTABLES]: Clean up vmalloc usage in net/bridge/netfilter/ebtables.c
Jayachandran C [Thu, 20 Apr 2006 07:14:49 +0000 (00:14 -0700)]
[EBTABLES]: Clean up vmalloc usage in net/bridge/netfilter/ebtables.c

Make all the vmalloc calls in net/bridge/netfilter/ebtables.c follow
the standard convention.  Remove unnecessary casts, and use '*object'
instead of 'type'.

Signed-off-by: Jayachandran C. <c.jayachandran@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: Add skb->truesize assertion checking.
David S. Miller [Thu, 20 Apr 2006 07:10:50 +0000 (00:10 -0700)]
[NET]: Add skb->truesize assertion checking.

Add some sanity checking.  truesize should be at least sizeof(struct
sk_buff) plus the current packet length.  If not, then truesize is
seriously mangled and deserves a kernel log message.

Currently we'll do the check for release of stream socket buffers.

But we can add checks to more spots over time.

Incorporating ideas from Herbert Xu.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[TCP]: Account skb overhead in tcp_fragment
Herbert Xu [Thu, 20 Apr 2006 04:35:00 +0000 (21:35 -0700)]
[TCP]: Account skb overhead in tcp_fragment

Make sure that we get the full sizeof(struct sk_buff)
plus the data size accounted for in skb->truesize.

This will create invariants that will allow adding
assertion checks on skb->truesize.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] Add more prevent_tail_call()
OGAWA Hirofumi [Wed, 19 Apr 2006 21:41:39 +0000 (06:41 +0900)]
[PATCH] Add more prevent_tail_call()

Those also break userland regs like following.

   00000000 <sys_chown16>:
      0: 0f b7 44 24 0c        movzwl 0xc(%esp),%eax
      5: 83 ca ff              or     $0xffffffff,%edx
      8: 0f b7 4c 24 08        movzwl 0x8(%esp),%ecx
      d: 66 83 f8 ff           cmp    $0xffffffff,%ax
     11: 0f 44 c2              cmove  %edx,%eax
     14: 66 83 f9 ff           cmp    $0xffffffff,%cx
     18: 0f 45 d1              cmovne %ecx,%edx
     1b: 89 44 24 0c           mov    %eax,0xc(%esp)
     1f: 89 54 24 08           mov    %edx,0x8(%esp)
     23: e9 fc ff ff ff        jmp    24 <sys_chown16+0x24>

where the tailcall at the end overwrites the incoming stack-frame.

Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
[ I would _really_ like to have a way to tell gcc about calling
  conventions. The "prevent_tail_call()" macro is pretty ugly ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[SUNGEM]: Marvell PHY suspend.
Johannes Berg [Wed, 19 Apr 2006 22:42:28 +0000 (15:42 -0700)]
[SUNGEM]: Marvell PHY suspend.

In a short discussion with Benjamin Herrenschmidt he mentioned
that Marvell PHYs are powered down the same way as the other
ones we currently handle. Thus actually do that, hopefully
saving some power during suspend.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[LLC]: Use pskb_trim_rcsum() in llc_fixup_skb().
David S. Miller [Wed, 19 Apr 2006 22:37:13 +0000 (15:37 -0700)]
[LLC]: Use pskb_trim_rcsum() in llc_fixup_skb().

Kernel Bugzilla #6409

If we use plain skb_trim(), that's wrong, because if
the SKB is cloned, and it can be because we unshared
it in the caller, we have to allow reallocation.  The
pskb_trim*() family of routines is therefore the most
appropriate here.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: sockfd_lookup_light() returns random error for -EBADFD
Hua Zhong [Wed, 19 Apr 2006 22:25:02 +0000 (15:25 -0700)]
[NET]: sockfd_lookup_light() returns random error for -EBADFD

This applies to 2.6.17-rc2.

There is a missing initialization of err in sockfd_lookup_light() that
could return random error for an invalid file handle.

Signed-off-by: Hua Zhong <hzhong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC]: __NR_sys_splice --> __NR_splice
David S. Miller [Wed, 19 Apr 2006 22:00:01 +0000 (15:00 -0700)]
[SPARC]: __NR_sys_splice --> __NR_splice

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] wext: Fix RtNetlink ENCODE security permissions
Jean Tourrilhes [Fri, 14 Apr 2006 17:47:26 +0000 (10:47 -0700)]
[PATCH] wext: Fix RtNetlink ENCODE security permissions

I've just realised that the RtNetlink code does not check the
permission for SIOCGIWENCODE and SIOCGIWENCODEEXT, which means that
any user can read the encryption keys. The fix is trivial and should
go in 2.6.17 alonside the two other patch I sent you last week.

Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx: iw_priv_args names should be <16 characters
Erik Mouw [Thu, 13 Apr 2006 13:02:27 +0000 (15:02 +0200)]
[PATCH] bcm43xx: iw_priv_args names should be <16 characters

The room for the names in bcm43xx_priv_wx_args[] are IFNAMSIZ long and
IFNAMSIZ is defined as 16, so the names in bcm43xx_priv_wx_args should
be 15 characters (16 including the trailing \0). This patch fixes that
for the "set_shortpreambl", "get_shortpreambl", "set_swencryption", and
"get_swencryption" private calls. Patch is against 2.6.17-rc1.

Signed-off-by: Erik Mouw <erik@harddisk-recovery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx: sysfs code cleanup
Michael Buesch [Thu, 13 Apr 2006 00:32:58 +0000 (02:32 +0200)]
[PATCH] bcm43xx: sysfs code cleanup

This cleans up the bcm43xx sysfs code and makes it compliant
with the unwritten sysfs rules (at least I hope so).

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx: fix pctl slowclock limit calculation
Michael Buesch [Thu, 13 Apr 2006 00:30:26 +0000 (02:30 +0200)]
[PATCH] bcm43xx: fix pctl slowclock limit calculation

This fixes coverity bug:
http://marc.theaimsgroup.com/?l=linux-netdev&m=114417628413880&w=2

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx: fix dyn tssi2dbm memleak
Adrian Bunk [Thu, 13 Apr 2006 00:27:49 +0000 (02:27 +0200)]
[PATCH] bcm43xx: fix dyn tssi2dbm memleak

This patch fixes a memory leak spotted by the Coverity checker.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx: fix config menu alignment
Randy Dunlap [Tue, 11 Apr 2006 21:32:53 +0000 (14:32 -0700)]
[PATCH] bcm43xx: fix config menu alignment

Use "depends on" to make all bcm43xx driver options be listed
at the same level.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx wireless: fix printk format warnings
Randy Dunlap [Tue, 11 Apr 2006 21:31:56 +0000 (14:31 -0700)]
[PATCH] bcm43xx wireless: fix printk format warnings

Fix printk format warnings:
drivers/net/wireless/bcm43xx/bcm43xx_debugfs.c:456: warning: format ‘%u’ expects type ‘unsigned int’, but argument 3 has type ‘size_t’
drivers/net/wireless/bcm43xx/bcm43xx_debugfs.c:460: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 2 has type ‘size_t’
drivers/net/wireless/bcm43xx/bcm43xx_debugfs.c:476: warning: format ‘%u’ expects type ‘unsigned int’, but argument 3 has type ‘size_t’
drivers/net/wireless/bcm43xx/bcm43xx_debugfs.c:480: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 2 has type ‘size_t’
drivers/net/wireless/bcm43xx/bcm43xx_dma.c:200: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 2 has type ‘dma_addr_t’
drivers/net/wireless/bcm43xx/bcm43xx_dma.c:311: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 2 has type ‘dma_addr_t’
drivers/net/wireless/bcm43xx/bcm43xx_dma.c:733: warning: format ‘%08x’ expects type ‘unsigned int’, but argument 2 has type ‘dma_addr_t’

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac: report when scanning has finished
Johannes Berg [Thu, 13 Apr 2006 09:41:28 +0000 (11:41 +0200)]
[PATCH] softmac: report when scanning has finished

Make softmac report a scan event when scanning has finished, that way
userspace can wait for the event to happen instead of polling for the
results.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac: fix event sending
Johannes Berg [Thu, 13 Apr 2006 00:42:42 +0000 (02:42 +0200)]
[PATCH] softmac: fix event sending

Softmac is sending custom events to userspace already, but it
should _really_ be sending the right WEXT events instead. This
patch fixes that.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac: handle iw_mode properly
johannes@sipsolutions.net [Tue, 11 Apr 2006 08:58:08 +0000 (10:58 +0200)]
[PATCH] softmac: handle iw_mode properly

Below patch allows using iw_mode auto with softmac. bcm43xx forces managed
so this bug wasn't noticed earlier, but this was one of the problems why
zd1211 didn't work earlier.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac: dont send out packets while scanning
johannes@sipsolutions.net [Tue, 11 Apr 2006 08:58:07 +0000 (10:58 +0200)]
[PATCH] softmac: dont send out packets while scanning

Seems we forgot to stop the queue while scanning. Better do that so we
don't transmit packets all the time during background scanning.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac: return -EAGAIN from getscan while scanning
johannes@sipsolutions.net [Tue, 11 Apr 2006 08:58:06 +0000 (10:58 +0200)]
[PATCH] softmac: return -EAGAIN from getscan while scanning

Below patch was developed after discussion with Daniel Drake who
mentioned to me that wireless tools expect an EAGAIN return from getscan
so that they can wait for the scan to finish before printing out the
results.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] bcm43xx: set trans_start on TX to prevent bogus timeouts
Michael Buesch [Mon, 10 Apr 2006 00:39:54 +0000 (02:39 +0200)]
[PATCH] bcm43xx: set trans_start on TX to prevent bogus timeouts

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] orinoco: fix truncating commsquality RID with the latest Symbol firmware
Pavel Roskin [Fri, 7 Apr 2006 08:10:26 +0000 (04:10 -0400)]
[PATCH] orinoco: fix truncating commsquality RID with the latest Symbol firmware

Symbol firmware F3.91-71 has an additional word in the commsquality RID.
Extend the receiving buffer by one word to accomodate it.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac: fix spinlock recursion on reassoc
Michael Buesch [Thu, 6 Apr 2006 23:42:55 +0000 (01:42 +0200)]
[PATCH] softmac: fix spinlock recursion on reassoc

This fixes a spinlock recursion on receiving a reassoc request.

On reassoc, the softmac calls back into the driver. This results in a
driver lock recursion. This schedules the assoc workqueue, instead
of calling it directly.

Probably, we should defer the _whole_ management frame processing
to a tasklet or workqueue, because it does several callbacks into the driver.
That is dangerous.

This fix should go into linus's tree, before 2.6.17 is released, because it
is remote exploitable (DoS by crash).

Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] Revert NET_RADIO Kconfig title change
Jean Tourrilhes [Tue, 4 Apr 2006 22:59:46 +0000 (15:59 -0700)]
[PATCH] Revert NET_RADIO Kconfig title change

2.6.17-rc1 changed the title for the entry CONFIG_NET_RADIO. I
personally disagree with this change and want it reverted. Patch for
2.6.17-rc1.
Rationale : WIRELESS_EXT is an invisible option. Therefore,
the only way for a user to enable it is via NET_RADIO. Some users need
to do that for out-of-tree drivers. Therefore it should be mentionned
in the title of the option.
Rationale2 : the option just below is called "Wireless
Extension API over RtNetlink". Some users may confuse this option for
the main "Wireless Extension" option. Therefore reverting this change
help disambiguate the relation between those two options.

Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] wext: Fix IWENCODEEXT security permissions
Jean Tourrilhes [Tue, 4 Apr 2006 22:53:43 +0000 (15:53 -0700)]
[PATCH] wext: Fix IWENCODEEXT security permissions

Check the permissions when user-space try to read the
encryption parameters via SIOCGIWENCODEEXT. This is trivial and
probably should go in 2.6.17...
Bug was found by Brian Eaton <eaton.lists@gmail.com>, thanks !

Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] wireless/atmel: send WEXT scan completion events
Dan Williams [Fri, 31 Mar 2006 20:13:31 +0000 (15:13 -0500)]
[PATCH] wireless/atmel: send WEXT scan completion events

Send scan completion events to user space when a scan completes.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] wireless/airo: clean up WEXT association and scan events
Dan Williams [Fri, 31 Mar 2006 20:08:46 +0000 (15:08 -0500)]
[PATCH] wireless/airo: clean up WEXT association and scan events

Airo firmware versions >= 5.30.17 send re-association events to the
driver that are currently unrecognized, causing spurious disassociation
events to be sent to user space.  Loss of sync due to scan requests also
results in disassociation events sent to user space.  This patch traps
those two events; suppressing sync-loss on scan, and sending the correct
association event on re-association notifications.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years ago[PATCH] softmac uses Wiress Ext.
Randy Dunlap [Mon, 27 Mar 2006 22:53:41 +0000 (14:53 -0800)]
[PATCH] softmac uses Wiress Ext.

softmac uses wireless extensions, so let it SELECT that config option;
WARNING: "wireless_send_event" [net/ieee80211/softmac/ieee80211softmac.ko] undefined!

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
18 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Wed, 19 Apr 2006 18:52:30 +0000 (11:52 -0700)]
Merge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
  IB/mthca: make a function static
  IB/ipath: Fix whitespace
  IB/ipath: Make more names static
  IB/mad: Fix RMPP version check during agent registration
  IB/srp: Remove request from list when SCSI abort succeeds

18 years agoIB/mthca: make a function static
Adrian Bunk [Tue, 18 Apr 2006 22:15:35 +0000 (00:15 +0200)]
IB/mthca: make a function static

This patch makes the needlessly global mthca_update_rate() static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIB/ipath: Fix whitespace
Roland Dreier [Wed, 19 Apr 2006 18:40:12 +0000 (11:40 -0700)]
IB/ipath: Fix whitespace

Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIB/ipath: Make more names static
Roland Dreier [Wed, 19 Apr 2006 18:40:12 +0000 (11:40 -0700)]
IB/ipath: Make more names static

Make symbols that are only used in a single source file static.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIB/mad: Fix RMPP version check during agent registration
Hal Rosenstock [Thu, 13 Apr 2006 01:29:10 +0000 (21:29 -0400)]
IB/mad: Fix RMPP version check during agent registration

Only check that RMPP version is not specified when MAD class does not
support RMPP.  Just because a class is allowed to use RMPP doesn't
mean that rmpp_version needs to be set for the MAD agent to
register. Checking this was a recent change which was too pedantic.

Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoIB/srp: Remove request from list when SCSI abort succeeds
Roland Dreier [Wed, 19 Apr 2006 18:40:10 +0000 (11:40 -0700)]
IB/srp: Remove request from list when SCSI abort succeeds

If a SCSI abort succeeds, then the aborted request should to be
removed from the list of pending requests.  This fixes list corruption
after an abort occurs.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
18 years agoMerge git://git.linux-nfs.org/pub/linux/nfs-2.6
Linus Torvalds [Wed, 19 Apr 2006 17:46:59 +0000 (10:46 -0700)]
Merge git://git.linux-nfs.org/pub/linux/nfs-2.6

* git://git.linux-nfs.org/pub/linux/nfs-2.6:
  SUNRPC: Dead code in net/sunrpc/auth_gss/auth_gss.c
  NFS: remove needless check in nfs_opendir()
  NFS: nfs_show_stats; for_each_possible_cpu(), not NR_CPUS
  NFS: make 2 functions static
  NFS,SUNRPC: Fix compiler warnings if CONFIG_PROC_FS & CONFIG_SYSCTL are unset
  NFS: fix PROC_FS=n compile error
  VFS: Fix another open intent Oops
  RPCSEC_GSS: fix leak in krb5 code caused by superfluous kmalloc

18 years agoSUNRPC: Dead code in net/sunrpc/auth_gss/auth_gss.c
Eric Sesterhenn [Wed, 19 Apr 2006 17:06:49 +0000 (13:06 -0400)]
SUNRPC: Dead code in net/sunrpc/auth_gss/auth_gss.c

Hi,

the coverity checker spotted that cred is always NULL
when we jump to out_err ( there is just one case, when
we fail to allocate the memory for cred )
This is Coverity ID #79

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
18 years agoNFS: remove needless check in nfs_opendir()
Carsten Otte [Wed, 19 Apr 2006 17:06:37 +0000 (13:06 -0400)]
NFS: remove needless check in nfs_opendir()

Local variable res was initialized to 0 - no check needed here.

Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
18 years agoNFS: nfs_show_stats; for_each_possible_cpu(), not NR_CPUS
John Hawkes [Wed, 19 Apr 2006 17:06:20 +0000 (13:06 -0400)]
NFS: nfs_show_stats; for_each_possible_cpu(), not NR_CPUS

Convert a for-loop that explicitly references "NR_CPUS" into the
potentially more efficient for_each_possible_cpu() construct.

Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
18 years agoNFS: make 2 functions static
Adrian Bunk [Tue, 18 Apr 2006 17:21:50 +0000 (13:21 -0400)]
NFS: make 2 functions static

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>