14 years ago[PATCH] I2C: missing new lines in i2c-core messages
Jean Delvare [Fri, 29 Jul 2005 19:15:29 +0000]
[PATCH] I2C: missing new lines in i2c-core messages

Two log messages lack their trailing new line in i2c-core. I'd swear I had
fixed them already, but it seems not. Bonus: improved coding style.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] I2C: use time_after in 3 chip drivers
Jean Delvare [Fri, 29 Jul 2005 19:15:12 +0000]
[PATCH] I2C: use time_after in 3 chip drivers

A few i2c drivers were not updated to use time_after() yet.

Signed-off-by: Marcelo Feitoza Parisi <marcelo@feitoza.com.br>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] I2C: Missing space in split strings
Jean Delvare [Fri, 29 Jul 2005 19:15:07 +0000]
[PATCH] I2C: Missing space in split strings

A few split string in i2c (and now hwmon) drivers lack a joining space,
causing them to display incorrectly. This trivial patch fixes that up.
Please apply, thanks.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] I2C: ds1337 - fix 12/24 hour mode bug
Ladislav Michl [Fri, 29 Jul 2005 19:15:00 +0000]
[PATCH] I2C: ds1337 - fix 12/24 hour mode bug

DS1339 manual, page 6, chapter Date and time operation:
  The DS1339 can be run in either 12-hour or 24-hour mode. Bit 6 of the
  hours register is defined as the 12-hour or 24-hour mode-select bit.
  When high, the 12-hour mode is selected.

Patch below makes ds1337 driver work as documented in manual.

Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] I2C-MPC: Restore code removed
Kumar Gala [Fri, 29 Jul 2005 19:14:40 +0000]
[PATCH] I2C-MPC: Restore code removed

I2C-MPC: Restore code removed

A previous patch to remove support for the OCP device model was way
to generious and moved some of the platform device model code, oops.

Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Add the rules about the -stable kernel releases to the Documentation directory
Greg KH [Fri, 29 Jul 2005 19:14:34 +0000]
[PATCH] Add the rules about the -stable kernel releases to the Documentation directory

This was the last agreed upon set of rules, it's probably time we actually add
them to the kernel tree to make them "official".

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] DEBUG_FS must depend on SYSFS
Adrian Bunk [Fri, 29 Jul 2005 19:14:28 +0000]
[PATCH] DEBUG_FS must depend on SYSFS

CONFIG_DEBUG_FS=y and CONFIG_SYSFS=n results in the following compile
error:

<--  snip  -->

...
  LD      vmlinux
fs/built-in.o: In function `debugfs_init':
inode.c:(.init.text+0x31be): undefined reference to `kernel_subsys'
make: *** [vmlinux] Error 1

<--  snip  -->

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] sysfs: fix sysfs_setattr
Maneesh Soni [Fri, 29 Jul 2005 19:14:19 +0000]
[PATCH] sysfs: fix sysfs_setattr

o sysfs_dirent's s_mode field should also be updated in sysfs_setattr(), else
  there could be inconsistency in the two fields. s_mode is used while
  ->readdir so as not to bring in the inode to cache.

Signed-off-by: Maneesh Soni <maneesh@in.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] sysfs: fix sysfs_chmod_file
Maneesh Soni [Fri, 29 Jul 2005 19:13:35 +0000]
[PATCH] sysfs: fix sysfs_chmod_file

o sysfs_chmod_file() must update the new iattr field in sysfs_dirent else
  the mode change will not be persistent in case of inode evacuation from
  cache.

Signed-off-by: Maneesh Soni <maneesh@in.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] stable_api_nonsense.txt fixes
Daniel Walker [Fri, 29 Jul 2005 19:14:07 +0000]
[PATCH] stable_api_nonsense.txt fixes

Signed-off-by: Daniel Walker <dwalker@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64 machine_kexec: Use standard pagetable helpers
Eric W. Biederman [Fri, 29 Jul 2005 19:25:28 +0000]
[PATCH] x86_64 machine_kexec: Use standard pagetable helpers

Use the standard hardware page table manipulation macros.
This is possible now that linux works with all 4 levels
of the page tables.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64 machine_kexec: Cleanup inline assembly.
Eric W. Biederman [Fri, 29 Jul 2005 19:02:09 +0000]
[PATCH] x86_64 machine_kexec: Cleanup inline assembly.

In an uncensored copy of code from i386 to x86_64 I wound up
with inline assembly with the wrong constraints.  Use input
constraints instead of output constraints.

So I know the assembler will do the right thing specify the size
of the operand lidtq and lgdtq instead of just lidt and lgdt.

Make load_segments use an input constraint, and delete the macro fun.
Without having to reload %cs like I do on i386 this code is noticeably
simpler.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] i386 machine_kexec: Cleanup inline assembly
Eric W. Biederman [Fri, 29 Jul 2005 19:01:18 +0000]
[PATCH] i386 machine_kexec: Cleanup inline assembly

For some reason I was telling my inline assembly that the
input argument was an output argument.

Playing in the trampoline code I have seen a couple of
instances where lgdt get the wrong size (because the
trampolines run in 16bit mode) so use lgdtl and lidtl to
be explicit.

Additionally gcc-3.3 and gcc-3.4 want's an lvalue for a
memory argument and it doesn't think an array of characters
is an lvalue so use a packed structure instead.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] reboot: remove device_suspend(PMSG_FREEZE) from kernel_kexec
Eric W. Biederman [Fri, 29 Jul 2005 18:50:57 +0000]
[PATCH] reboot: remove device_suspend(PMSG_FREEZE) from kernel_kexec

If device_suspend(PMSG_FREEZE) is not ready to be called in
kernel_restart it is definitely not ready to be called in the even more
fickle kernel_kexec.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agox86: fix new find_first_bit()
Linus Torvalds [Fri, 29 Jul 2005 15:01:22 +0000]
x86: fix new find_first_bit()

Some edge problems with the original C rewrite.

Thanks go to Cal Peake, who pinpointed the breakage to the rewrite, and
tested this fixed version.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq
Linus Torvalds [Fri, 29 Jul 2005 17:16:25 +0000]
Merge /pub/scm/linux/kernel/git/davej/cpufreq

14 years agoFix up powernow-k8 compile. (Missing definitions).
Dave Jones [Fri, 29 Jul 2005 16:56:41 +0000]
Fix up powernow-k8 compile. (Missing definitions).

From: Mark Langsdorf <mark.langsdorf@amd.com>
Signed-off-by: Dave Jones <davej@redhat.com>

14 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Fri, 29 Jul 2005 16:48:34 +0000]
Merge /pub/scm/linux/kernel/git/dtor/input

14 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm-smp
Linus Torvalds [Fri, 29 Jul 2005 16:47:08 +0000]
Merge kernel.org:/home/rmk/linux-2.6-arm-smp

14 years agoMerge head 'upstream' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev
Linus Torvalds [Fri, 29 Jul 2005 16:04:47 +0000]
Merge head 'upstream' of /linux/kernel/git/jgarzik/libata-dev

14 years ago[ARM SMP] Ensure secondary CPUs see their pen release
Russell King [Fri, 29 Jul 2005 15:36:48 +0000]
[ARM SMP] Ensure secondary CPUs see their pen release

Since the secondary CPUs will not be operating in symetric mode
while they are held in the pen, we need to ensure that the write
to pen_release is visible to them, by flushing the cache.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 29 Jul 2005 05:12:57 +0000]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

14 years ago[PATCH] selinux: Fix address length checks in connect hook
Stephen Smalley [Fri, 29 Jul 2005 04:16:21 +0000]
[PATCH] selinux: Fix address length checks in connect hook

This patch fixes the address length checks in the selinux_socket_connect
hook to be no more restrictive than the underlying ipv4 and ipv6 code;
otherwise, this hook can reject valid connect calls.  This patch is in
response to a bug report where an application was calling connect on an
INET6 socket with an address that didn't include the optional scope id and
failing due to these checks.

Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: James Morris <jmorris@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] fbdev: colormap fixes fix
Jon Smirl [Fri, 29 Jul 2005 04:16:19 +0000]
[PATCH] fbdev: colormap fixes fix

Fix a buffer overflow vunerabilty in previous cmap patch

Signed-off-by: Jon Smirl <jonsmirl@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] re-disable TSC on NUMAQ
Dave Hansen [Fri, 29 Jul 2005 04:16:18 +0000]
[PATCH] re-disable TSC on NUMAQ

Somewhere recently, the TSC got re-enabled for timekeeping on NUMAQ
machines.  However, the hardware makes these get unsynchronized quite
badly.  So badly, in fact, that the code to fix up the skew can just hang
on boot.

This patch re-disables them.  It's nicely confined to the numaq.c file.  It
would be great if this could make it into 2.6.13, I think it counts as a
bugfix.

Tested on a 16-proc 4-node NUMAQ.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] posix timers: fix normalization problem
George Anzinger [Fri, 29 Jul 2005 04:16:16 +0000]
[PATCH] posix timers: fix normalization problem

(We found this (after a customer complained) and it is in the kernel.org
kernel.  Seems that for CLOCK_MONOTONIC absolute timers and clock_nanosleep
calls both the request time and wall_to_monotonic are subtracted prior to
the normalize resulting in an overflow in the existing normalize test.
This causes the result to be shifted ~4 seconds ahead instead of ~2 seconds
back in time.)

The normalize code in posix-timers.c fails when the tv_nsec member is ~1.2
seconds negative.  This can happen on absolute timers (and
clock_nanosleeps) requested on CLOCK_MONOTONIC (both the request time and
wall_to_monotonic are subtracted resulting in the possibility of a number
close to -2 seconds.)

This fix uses the set_normalized_timespec() (which does not have an
overflow problem) to fix the problem and as a side effect makes the code
cleaner.

Signed-off-by: George Anzinger <george@mvista.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: implement hostfs syncing
Paolo 'Blaisorblade' Giarrusso [Fri, 29 Jul 2005 04:16:15 +0000]
[PATCH] uml: implement hostfs syncing

Actually implement the hostfs "sync" method.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: Clean up prink calls
Christophe Lucas [Fri, 29 Jul 2005 04:16:12 +0000]
[PATCH] uml: Clean up prink calls

printk() calls should include appropriate KERN_* constant.

Signed-off-by: Christophe Lucas <clucas@rotomalug.org>
Signed-off-by: Domen Puncer <domen@coderock.org>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: Fix typo
Bodo Stroesser [Fri, 29 Jul 2005 04:16:11 +0000]
[PATCH] uml: Fix typo

Fix a typo in wait_stub_done.

Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: Fix load average >=1
Jeff Dike [Fri, 29 Jul 2005 04:16:09 +0000]
[PATCH] uml: Fix load average >=1

update_process_times was missing its irq_enter/irq_exit wrapper.  This caused
ksoftirqd to be scheduled on every clock tick.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: Fix redundant assignment
Jeff Dike [Fri, 29 Jul 2005 04:16:08 +0000]
[PATCH] uml: Fix redundant assignment

By this point, .is_user has already been set, so this assignment is useless.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: Fix skas0 stub return
Bodo Stroesser [Fri, 29 Jul 2005 04:16:06 +0000]
[PATCH] uml: Fix skas0 stub return

It's wrong to pop a fixed number of words from stack before calling sigreturn,
as the number depends on what code is generated by the compiler for the start
of stub_segv_handler().  What we need is esp containing the address of
sigcontext.  So we explicitly load that pointer into esp.

Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: vm86 compile fix
Jeff Dike [Fri, 29 Jul 2005 04:16:05 +0000]
[PATCH] uml: vm86 compile fix

We added an include of asm/vm86.h in include/asm-i386/ptrace.h.  Since UML
includes the underlying arch's ptrace.h, it needs an asm/vm86.h in order to
build.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: avoid unnecessary pcap rebuild
Paolo 'Blaisorblade' Giarrusso [Fri, 29 Jul 2005 04:16:03 +0000]
[PATCH] uml: avoid unnecessary pcap rebuild

Just a Kbuild subtlety, not listing a target file inside targets causes it
to be rebuilt each time, and as a consequence everything depending on it is
rebuilt.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] uml: fix TT mode by reverting "use fork instead of clone"
Jeff Dike [Fri, 29 Jul 2005 04:16:01 +0000]
[PATCH] uml: fix TT mode by reverting "use fork instead of clone"

With Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>

Revert the following patch, because of miscompilation problems in different
environments leading to UML not working *at all* in TT mode; it was merged
lately in 2.6 development cycle, a little after being written, and has
caused problems to lots of people; I know it's a bit too long, but it
shouldn't have been merged in first place, so I still apply for inclusion
in the -stable tree.  Anyone using this feature currently is either using
some older kernel (some reports even used 2.6.12-rc4-mm2) or using this
patch, as included in my -bs patchset.

For now there's not yet a fix for this patch, so for now the best thing is
to drop it (which was widely reported to give a working kernel, and as such
was even merged in -stable tree).

"Convert the boot-time host ptrace testing from clone to fork.  They were
essentially doing fork anyway.  This cleans up the code a bit, and makes
valgrind a bit happier about grinding it."

URL:
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=98fdffccea6cc3fe9dba32c0fcc310bcb5d71529

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] device-mapper: fix md->lock deadlocks in core
Alasdair G Kergon [Fri, 29 Jul 2005 04:16:00 +0000]
[PATCH] device-mapper: fix md->lock deadlocks in core

This patch is an attempt to fix deadlocks discovered in the core dm.

The problems boil down to md->lock having to be held in too many places, so
I've split it into two: md->suspend_lock and md->io_lock.

suspend_lock is now held throughout dm_suspended() as well as dm_resume()
and dm_swap_table() so that these functions cannot run concurrently:
there's no requirement for that and it added complexity.

DMF_FS_LOCKED becomes redundant: DMF_SUSPENDED provides adequate
protection.

Signed-Off-By: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] device-mapper: fix deadlocks in core
Alasdair G Kergon [Fri, 29 Jul 2005 04:15:59 +0000]
[PATCH] device-mapper: fix deadlocks in core

Avoid another bdget_disk which can deadlock.

Signed-Off-By: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] device-mapper: fix deadlocks in core (prep)
Alasdair G Kergon [Fri, 29 Jul 2005 04:15:57 +0000]
[PATCH] device-mapper: fix deadlocks in core (prep)

Some code tidy-ups in preparation for the next patches.  Change
dm_table_pre/postsuspend_targets to accept NULL.  Use dm_suspended()
throughout.

Signed-Off-By: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: fix cpu_to_node setup for sparse apic_ids
Ravikiran G Thirumalai [Fri, 29 Jul 2005 04:15:55 +0000]
[PATCH] x86_64: fix cpu_to_node setup for sparse apic_ids

While booting with SMT disabled in bios, when using acpi srat to setup
cpu_to_node[], sparse apic_ids create problems.

Without this patch, intel x86_64 boxes with hyperthreading disabled in the
bios (and which rely on srat for numa setup) endup having incorrect values in
cpu_to_node[] arrays, causing sched domains to be built incorrectly etc.

Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Remove unused variable in k8-bus.c
Andi Kleen [Fri, 29 Jul 2005 04:15:54 +0000]
[PATCH] x86_64: Remove unused variable in k8-bus.c

arch/x86_64/pci/k8-bus.c: In function `fill_mp_bus_to_cpumask':
arch/x86_64/pci/k8-bus.c:32: warning: unused variable `printed'

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Use msleep in smpboot.c
Nishanth Aravamudan [Fri, 29 Jul 2005 04:15:53 +0000]
[PATCH] x86_64: Use msleep in smpboot.c

Replace schedule_timeout() with msleep() to guarantee the task delays as
expected.

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Domen Puncer <domen@coderock.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Fix gcc 4 warning in sched_find_first_bit
Jesse Millan [Fri, 29 Jul 2005 04:15:51 +0000]
[PATCH] x86_64: Fix gcc 4 warning in sched_find_first_bit

This patch eliminates the GCC4 warning on the x86_64 platform:

kernel/sched.c:1824: warning: control may reach end of non-void function
'sched_find_first_bit' being inlined.

The change follows the lead of others, i.e.  it is guaranteed that at least
one of b[0], b[1], or b[2] will have a bit set and evaluate to true.  That
being said, GCC4.0.0 notices that the code flow does not return anything if
b[0], b[1] and b[2] are not true.  Since we know better, if it's not b[0] or
b[1], it has to be b[2].

Signed-off-by: Jesse Millan <jessem@cs.pdx.edu>
Signed-off-by: Domen Puncer <domen@coderock.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Fix SRAT handling on non dual core systems
Andi Kleen [Fri, 29 Jul 2005 04:15:50 +0000]
[PATCH] x86_64: Fix SRAT handling on non dual core systems

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Switch to the interrupt stack when running a softirq in local_bh_enable()
Andi Kleen [Fri, 29 Jul 2005 04:15:49 +0000]
[PATCH] x86_64: Switch to the interrupt stack when running a softirq in local_bh_enable()

This avoids some potential stack overflows with very deep softirq callchains.
i386 does this too.

TOADD CFI annotation

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Small assembly improvements
Andi Kleen [Fri, 29 Jul 2005 04:15:48 +0000]
[PATCH] x86_64: Small assembly improvements

Save a byte here and there.  Ultimatively useless, but these things always
catch my eyes when reading the code so just fix them for now.

Also I got at least one patch fixing of them already, which gives a good
excuse.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Remove unnecessary include in fault.c
Andi Kleen [Fri, 29 Jul 2005 04:15:47 +0000]
[PATCH] x86_64: Remove unnecessary include in fault.c

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: When running cpuid4 need to run on the correct CPU
Andi Kleen [Fri, 29 Jul 2005 04:15:46 +0000]
[PATCH] x86_64: When running cpuid4 need to run on the correct CPU

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Turn BUG data into valid instruction
Andi Kleen [Fri, 29 Jul 2005 04:15:45 +0000]
[PATCH] x86_64: Turn BUG data into valid instruction

This avoids confusing the disassembler. Costs 2 bytes per BUG.

Thanks to Suresh Siddha and Jan Beulich for suggesting suitable instructions.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Icecream has no way of detecting assembler-level includes
Andreas Gruenbacher [Fri, 29 Jul 2005 04:15:44 +0000]
[PATCH] x86_64: Icecream has no way of detecting assembler-level includes

Icecream preprocesses c sources locally, and sends the result off to a remote
host for compiling.  It does not recognize includes at assembler level.  The
fix is to put the assemberincludes an a separate .s file, which will always be
assembled locally.

Signed-off-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Support more than 8 cores on AMD systems
Andi Kleen [Fri, 29 Jul 2005 04:15:42 +0000]
[PATCH] x86_64: Support more than 8 cores on AMD systems

Use physical mode instead of logical mode to address more CPUs.  This is also
used in the CPU hotplug case to avoid a race.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Remove the broadcast options that were added for cpuhotplug
Andi Kleen [Fri, 29 Jul 2005 04:15:41 +0000]
[PATCH] x86_64: Remove the broadcast options that were added for cpuhotplug

Will be obsolete with physflat.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Remove IA32_* build tools in Makefile
Andi Kleen [Fri, 29 Jul 2005 04:15:41 +0000]
[PATCH] x86_64: Remove IA32_* build tools in Makefile

Not used anymore since quite some time. Just uses -m32 instead.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Create per CPU machine check sysfs directories
Andi Kleen [Fri, 29 Jul 2005 04:15:39 +0000]
[PATCH] x86_64: Create per CPU machine check sysfs directories

This patch will create machinecheck sysdev directories per CPU.  All of the
cpus still share the same ctl banks.  When compiled with CONFIG_HOTPLUG_CPU,
it will also bring up/down sysdev directories as cpus go up/down.  I have
tested the patch along with CONFIG_HOTPLUG_CPU option on in 2.6.13-rc1 kernel.

Minor changes by AK: remove useless unload function

Signed-off-by: Jacob Shin <jacob.shin@amd.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Print a boot message for hotplug memory zones
Andi Kleen [Fri, 29 Jul 2005 04:15:39 +0000]
[PATCH] x86_64: Print a boot message for hotplug memory zones

From: Keith Manning

Print a boot message for hotplug memory zones

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Fix overflow in NUMA hash function setup
Keith Mannthey [Fri, 29 Jul 2005 04:15:38 +0000]
[PATCH] x86_64: Fix overflow in NUMA hash function setup

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Fix incorrectly defined MSR_K8_SYSCFG
Andi Kleen [Fri, 29 Jul 2005 04:15:37 +0000]
[PATCH] x86_64: Fix incorrectly defined MSR_K8_SYSCFG

Harmless because the kernel didn't use it.  Noticed by Travis Betak

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Fix some typos in system.h comments
Andi Kleen [Fri, 29 Jul 2005 04:15:36 +0000]
[PATCH] x86_64: Fix some typos in system.h comments

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Remove obsolete eat_key prototype
Andi Kleen [Fri, 29 Jul 2005 04:15:35 +0000]
[PATCH] x86_64: Remove obsolete eat_key prototype

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Fix some comments in tlbflush.h
Andi Kleen [Fri, 29 Jul 2005 04:15:35 +0000]
[PATCH] x86_64: Fix some comments in tlbflush.h

Were either outdated or misleading.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Some updates for boot-options.txt
Andi Kleen [Fri, 29 Jul 2005 04:15:34 +0000]
[PATCH] x86_64: Some updates for boot-options.txt

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Improve CONFIG_GART_IOMMU description and make it default y
Andi Kleen [Fri, 29 Jul 2005 04:15:32 +0000]
[PATCH] x86_64: Improve CONFIG_GART_IOMMU description and make it default y

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Remove unused variable in delay.c
Andi Kleen [Fri, 29 Jul 2005 04:15:31 +0000]
[PATCH] x86_64: Remove unused variable in delay.c

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Some cleanup in setup64.c
Andi Kleen [Fri, 29 Jul 2005 04:15:30 +0000]
[PATCH] x86_64: Some cleanup in setup64.c

Minor cleanup.

Move things into their include files, remove obsolete includes, fix
indentation, remove obsolete special cases etc.

I also added the per cpu section to asm-generic/sections.h and fixed
init/main.c to use it.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Clarify Booting processor ... message
Andi Kleen [Fri, 29 Jul 2005 04:15:29 +0000]
[PATCH] x86_64: Clarify Booting processor ... message

No need to print kernel addresses there and clarify what the APIC-ID is.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Minor clean up to CPU setup - use smp_processor_id instead of custom...
Andi Kleen [Fri, 29 Jul 2005 04:15:28 +0000]
[PATCH] x86_64: Minor clean up to CPU setup - use smp_processor_id instead of custom hack

Does not change any semantics because numa_add_cpu checks for CPU 0 anyways.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Move cpu_present/possible_map parsing earlier
Andi Kleen [Fri, 29 Jul 2005 04:15:27 +0000]
[PATCH] x86_64: Move cpu_present/possible_map parsing earlier

Various code needs this information now before the actual SMP bootup.  Instead
of computing it on the fly while booting the other CPUs set it up now while
initial MPtable/MADT parsing.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: i386/x86_64: remove prototypes for not existing functions in smp.h
Andi Kleen [Fri, 29 Jul 2005 04:15:26 +0000]
[PATCH] x86_64: i386/x86_64: remove prototypes for not existing functions in smp.h

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Use for_each_cpu_mask for clustered IPI flush
Andi Kleen [Fri, 29 Jul 2005 04:15:25 +0000]
[PATCH] x86_64: Use for_each_cpu_mask for clustered IPI flush

Makes it slightly more efficient.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Update defconfig
Andi Kleen [Fri, 29 Jul 2005 04:15:23 +0000]
[PATCH] x86_64: Update defconfig

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: Always ack IPIs even on errors
Andi Kleen [Fri, 29 Jul 2005 04:15:22 +0000]
[PATCH] x86_64: Always ack IPIs even on errors

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] x86_64: cpu hotplug changes kills nmi watchdog
Alexander Nyberg [Fri, 29 Jul 2005 04:15:20 +0000]
[PATCH] x86_64: cpu hotplug changes kills nmi watchdog

When the x86_64 cpu hotplug changes went in it added a check in
default_do_nmi() which kills NMI delivery on any CPU but the BSP.

The NMI watchdog is brought up quite some time before the online bit is set
in num_online_cpus so this won't work very well.  The nmi watchdogs on cpus
that are not BSP will never be reprogrammed and no NMIs.

Why was this check added? How does an offlined cpu receive an NMI?

Signed-off-by: Alexander Nyberg <alexn@telia.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Andrew Morton <akpm@osdl.org>
Cc: Zwane Mwaikambo <zwane@arm.linux.org.uk>
Cc: <ashok.raj@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agoLinux 2.6.13-rc4
Linus Torvalds [Thu, 28 Jul 2005 22:44:44 +0000]
Linux 2.6.13-rc4

14 years ago[PATCH] fix gconfig crash
Joachim Nilsson [Thu, 28 Jul 2005 22:18:03 +0000]
[PATCH] fix gconfig crash

I ran glade-2 on the glade file, fixed two missing stock icons and
cleaned up the C code that inserts the single/split/full modes. The
rest of the patch is minor cleanups only. I refrained from using all
the included xpm icons in images.c (like qconf.cc does) in favour of
using the stock Gtk+ icons instead. Oh, yes there was a "back" bug
in split mode that I also removed, oh well...

It has been tested with success by several people, including
Jesper Juhl, Randy Dunlap and myself.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agolibata: Check PCI sub-class code before disabling AHCI
Greg Felix [Thu, 28 Jul 2005 19:54:15 +0000]
libata: Check PCI sub-class code before disabling AHCI

This patch adds functionality to check the PCI sub-class code of an
AHCI capable device before disabling AHCI.  It fixes a bug where an
ICH7 sata controller is being setup by the BIOS as sub-class 1 (ide)
and the AHCI control registers weren't being initialized, thus causing
an IO error in piix_disable_ahci().

Signed-off-by: Gregory Felix <greg.felix@gmail.com>

14 years ago[ARM SMP] Fix another ARMv6 bitop problem
Russell King [Thu, 28 Jul 2005 19:36:26 +0000]
[ARM SMP] Fix another ARMv6 bitop problem

We sometimes forgot to check whether the exclusive store succeeded.
Ensure that we always check.  Also ensure that we always use the
out of line versions, since the inline versions are not SMP safe.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[ARM SMP] Ensure secondary CPUs have a clean TLB
Russell King [Thu, 28 Jul 2005 19:32:47 +0000]
[ARM SMP] Ensure secondary CPUs have a clean TLB

Since ARMv6 CPUs will not flush the TLB on context switches, it is
possible that we may end up with some global TLB entries remaining
present, eventually upsetting userspace.  Explicitly flush the
entire TLB on secondary CPUs as they startup, after we have switched
to the init_mm page tables.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[NET]: Fix busy waiting in dev_close().
David S. Miller [Thu, 28 Jul 2005 19:12:58 +0000]
[NET]: Fix busy waiting in dev_close().

If the current task has signal_pending(), the loop we have
to wait for the __LINK_STATE_RX_SCHED bit to clear becomes
a pure busy-loop.

Fixed by using msleep() instead of the hand-crafted version.

Noticed by Andrew Morton.

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

14 years agoRevert broken "statement with no effect" warning fix
Linus Torvalds [Thu, 28 Jul 2005 17:34:47 +0000]
Revert broken "statement with no effect" warning fix

It may shut up gcc, but it also incorrectly changes the semantics of the
smp_call_function() helpers.

You can fix the warning other ways if you are interested (create another
inline function that takes no arguments and returns zero), but
preferably gcc just shouldn't complain about unused return values from
statement expressions in the first place.

14 years agopowernow-k8.c: In function `query_current_values_with_pending_wait':
Dave Jones [Thu, 28 Jul 2005 16:45:10 +0000]
powernow-k8.c: In function `query_current_values_with_pending_wait':
powernow-k8.c:110: warning: `hi' may be used uninitialized in this function

Signed-off-by: Brian Gerst <bgerst@didntduck.org>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>

14 years agoHere are two possible cleanups in cpufreq.c:
Dave Jones [Thu, 28 Jul 2005 16:43:56 +0000]
Here are two possible cleanups in cpufreq.c:
* ret has no need to be unsigned in cpufreq_driver_target()
* ret has no need to be initialized in __cpufreq_governor()

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Dave Jones <davej@redhat.com>

14 years agoOpteron revision F will support higher frequencies than
Dave Jones [Thu, 28 Jul 2005 16:40:04 +0000]
Opteron revision F will support higher frequencies than
can be encoded in the current driver's 4 bit frequency
field.  This patch updates the driver to support Rev F
including 6 bit FIDs and processor ID updates.

This should apply cleanly whether or not the dual-core
bugfix I sent out last week is applied.  I'd prefer
that both get applied, of course.

Signed-off-by: David Keck <david.keck@amd.com>
Signed-off-by: Mark Langsdorf <mark.langsdorf@amd.com>
Signed-off-by: Dave Jones <davej@redhat.com>

14 years agopowernow-k8 requires that a data structure for
Dave Jones [Thu, 28 Jul 2005 16:38:21 +0000]
powernow-k8 requires that a data structure for
each core be created in the _cpu_init function
call.  The cpufreq infrastructure doesn't call
_cpu_init for the second core in each processor.
Some systems crashed when _get was called with
an odd-numbered core because it tried to
dereference a NULL pointer since the data
structure had not been created.

The attached patch solves the problem by
initializing data structures for all shared
cores in the _cpu_init function.  It should
apply to 2.6.12-rc6 and has been tested by
AMD and Sun.

Signed-off-by: Mark Langsdorf <mark.langsdorf@amd.com>
Signed-off-by: Dave Jones <davej@redhat.com>

14 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/perex/alsa
Linus Torvalds [Thu, 28 Jul 2005 15:41:58 +0000]
Merge /pub/scm/linux/kernel/git/perex/alsa

14 years ago[PATCH] speed up on find_first_bit for i386 (let compiler do the work)
Steven Rostedt [Thu, 28 Jul 2005 12:45:06 +0000]
[PATCH] speed up on find_first_bit for i386 (let compiler do the work)

Avoid using "rep scas", just let the compiler select a sequence of
regular instructions.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] alpha: fix "statement with no effect" warnings
Richard Henderson [Thu, 28 Jul 2005 08:07:41 +0000]
[PATCH] alpha: fix "statement with no effect" warnings

Apparently gcc 4.0 complains about "({ 0; });", which leads to -Werror
breakage in one of the alpha oprofile modules.

One might could argue that this is a gcc bug, in that statement-expressions
should be considered to be function-like rather than statement-like for the
purposes of this warning.  But it's just as easy to use an inline function
in the first place, side-stepping the issue.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] PCDP: if PCDP contains parity information, use it
Bjorn Helgaas [Thu, 28 Jul 2005 08:07:39 +0000]
[PATCH] PCDP: if PCDP contains parity information, use it

If the PCDP supplies parity, use it (only none/even/odd supported), and
don't append parity/stop bit arguments unless baud is present.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] e1000: no need for reboot notifier
Tony Luck [Thu, 28 Jul 2005 08:07:38 +0000]
[PATCH] e1000: no need for reboot notifier

sys_reboot() now calls device_suspend(), so it is no longer necessary for
the e1000 driver to register a reboot notifier [in fact doing so results in
e1000_suspend() getting called twice].

(akpm: we need to fast-track this.  It's causing ia64 to oops on shutdown)

Signed-off-by: Tony Luck <tony.luck@intel.com>
Cc: <cramerj@intel.com>
Cc: <john.ronciak@intel.com>
Cc: <ganesh.venkatesan@intel.com>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] SELinux: default labeling of MLS field
James Morris [Thu, 28 Jul 2005 08:07:37 +0000]
[PATCH] SELinux: default labeling of MLS field

Implement kernel labeling of the MLS (multilevel security) field of
security contexts for files which have no existing MLS field.  This is to
enable upgrades of a system from non-MLS to MLS without performing a full
filesystem relabel including all of the mountpoints, which would be quite
painful for users.

With this patch, with MLS enabled, if a file has no MLS field, the kernel
internally adds an MLS field to the in-core inode (but not to the on-disk
file).  This MLS field added is the default for the superblock, allowing
per-mountpoint control over the values via fixed policy or mount options.

This patch has been tested by enabling MLS without relabeling its
filesystem, and seems to be working correctly.

Signed-off-by: James Morris <jmorris@redhat.com>
Signed-off-by: Stephen Smalley <sds@epoch.ncsc.mil>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] cs89x0: collect tx_bytes statistics
Ian Campbell [Thu, 28 Jul 2005 08:07:34 +0000]
[PATCH] cs89x0: collect tx_bytes statistics

Signed-off-by: Ian Campbell <icampbell@arcom.com>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Fix incorrect Asus k7m irq router detection
Giancarlo Formicuccia [Thu, 28 Jul 2005 08:07:33 +0000]
[PATCH] Fix incorrect Asus k7m irq router detection

This patch:
http://marc.theaimsgroup.com/?l=bk-commits-head&m=111955644929114&w=2
uncovered a k7m bios bug, where the VT82C686A router is reported as
being "586-compatible". The two chips have different pirq mapping, so
this leads to "irq routing conflict" on many pci devices.

The suggested fix was discussed with Aleksey Gorelov, who helped me
to identify the problem as a probable bios bug.

Signed-off-by: Giancarlo Formicuccia <giancarlo.formicuccia@gmail.com>
Cc: Dave Jones <davej@codemonkey.org.uk>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] cciss per disk queue
Mike Miller [Thu, 28 Jul 2005 08:07:31 +0000]
[PATCH] cciss per disk queue

This patch adds per disk queue functionality to cciss.  Sometime back I
submitted a patch but it looks like only part of what I needed.  In the 2.6
kernel if we have more than one logical volume the driver will Oops during
rmmod.  It seems all of the queues actually point back to the same queue.
So after deleting the first volume you hit a null pointer on the second
one.

This has been tested in our labs.  There is no difference in performance,
it just fixes the Oops.

Signed-off-by: Mike Miller <mike.miller@hp.com>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: disable read prefetch/write burst on old O2Micro bridges
Daniel Ritz [Thu, 28 Jul 2005 08:07:30 +0000]
[PATCH] pcmcia: disable read prefetch/write burst on old O2Micro bridges

Older O2Micro bridges have problems with both read prefetch and write burst
depending on the combination of the chipset, bridge, cardbus card.  safest is
to disable read prefetch and write burst on those old bridges.

Signed-off-by: Daniel Ritz <daniel.ritz@gmx.ch>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] yenta: free_irq() on suspend.
Daniel Ritz [Thu, 28 Jul 2005 08:07:28 +0000]
[PATCH] yenta: free_irq() on suspend.

Resume doesn't seem to work without.

Signed-off-by: Daniel Ritz <daniel.ritz@gmx.ch>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: fix sharing IRQs and request_irq without IRQ_HANDLE_PRESENT
Dominik Brodowski [Thu, 28 Jul 2005 08:07:27 +0000]
[PATCH] pcmcia: fix sharing IRQs and request_irq without IRQ_HANDLE_PRESENT

Debugging and description from: Noah Misch <noah@cs.caltech.edu>

When a driver calls pcmcia_request_irq with IRQ_HANDLE_PRESENT unset, it looks
for an open IRQ by request_irq()ing with a dummy handler and NULL dev_info.
free_irq uses dev_info as a key for identifying the handler to free among
those sharing an IRQ, so request_irq returns -EINVAL if dev_info is NULL and
the IRQ may be shared.  That unknown error code is the -EINVAL.

It looks like only pcnet_cs and axnet_cs are affected.  Most other drivers let
pcmcia_request_irq install their interrupt handlers.  sym53c500_cs requests
its IRQ manually, but it cannot share an IRQ.

The appended patch changes pcmcia_request_irq to pass an arbitrary, unique,
non-NULL dev_info with the dummy handler.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: update documentation
Dominik Brodowski [Thu, 28 Jul 2005 08:07:26 +0000]
[PATCH] pcmcia: update documentation

Update the PCMCIA documentation to reflect some more, though older, changes.
Parts extracted from an e-mail from Randy Dunlap with his consent.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: fix many device IDs
Komuro [Thu, 28 Jul 2005 08:07:24 +0000]
[PATCH] pcmcia: fix many device IDs

If the product-id-string contains the '+' , '&' ,'_', it was not converted
properly from the /etc/pcmcia/config(pcmcia-cs config file).

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: avoid duble iounmap of one address
Dominik Brodowski [Thu, 28 Jul 2005 08:07:23 +0000]
[PATCH] pcmcia: avoid duble iounmap of one address

Avoid double iounmap of one address, and disable cis_virt if set_mem_map
failed.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: update au1000 to work with recent changes
Dominik Brodowski [Thu, 28 Jul 2005 08:07:22 +0000]
[PATCH] pcmcia: update au1000 to work with recent changes

Get the au1000 PCMCIA socket drivers to work.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: remove duplicates in orinoco_cs
Jar [Thu, 28 Jul 2005 08:07:21 +0000]
[PATCH] pcmcia: remove duplicates in orinoco_cs

Remove duplicates from the device id table.

Signed-off-by: Jarkko Raja <jar@pcuf.fi>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] pcmcia: fix comment
Pavel Roskin [Thu, 28 Jul 2005 08:07:20 +0000]
[PATCH] pcmcia: fix comment

There are two problems with the message about missing callback functions: it's
not written in correct English and it lacks newline at the end.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>