]> pilppa.com Git - linux-2.6-omap-h63xx.git/log
linux-2.6-omap-h63xx.git
16 years agox86: mtrr_cleanup: first 1M may be covered in var mtrrs
Yinghai Lu [Sat, 4 Oct 2008 21:50:32 +0000 (14:50 -0700)]
x86: mtrr_cleanup: first 1M may be covered in var mtrrs

The first 1M is don't care when it comes to the variables MTRRs.
Cover it as WB as a heuristic approximation; this is generally what we
want to minimize the number of registers.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
16 years agox86: mtrr_cleanup: print out correct type v2
Yinghai Lu [Sun, 5 Oct 2008 02:34:18 +0000 (19:34 -0700)]
x86: mtrr_cleanup: print out correct type v2

Print out the correct type when the Write Protected (WP) type is seen.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
16 years agox86, debug: mtrr_cleanup print out var mtrr before change it
Yinghai Lu [Fri, 3 Oct 2008 06:26:59 +0000 (23:26 -0700)]
x86, debug: mtrr_cleanup print out var mtrr before change it

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: mtrr_cleanup try gran_size to less than 1M, v3
Yinghai Lu [Thu, 2 Oct 2008 22:46:20 +0000 (15:46 -0700)]
x86: mtrr_cleanup try gran_size to less than 1M, v3

J.A. Magallón reported:

 >> Also, on a 64 bit box with 4Gb, it gives this:
 >>
 >> cicely:~# cat /proc/mtrr
 >> reg00: base=0x00000000 (   0MB), size=4096MB: write-back, count=1
 >> reg01: base=0x100000000 (4096MB), size=1024MB: write-back, count=1
 >> reg02: base=0x140000000 (5120MB), size= 512MB: write-back, count=1
 >> reg03: base=0x160000000 (5632MB), size= 256MB: write-back, count=1
 >> reg04: base=0x80000000 (2048MB), size=2048MB: uncachable, count=1

boundary handling has a problem ... fix it.

Reported-by: J.A. Magallón <jamagallon@ono.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: mtrr_cleanup try gran_size to less than 1M, cleanup
J.A. Magallón [Thu, 2 Oct 2008 22:32:37 +0000 (00:32 +0200)]
x86: mtrr_cleanup try gran_size to less than 1M, cleanup

Patch below cleans up formatting, with space for big bases and sizes (64 Gb).

Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: change MTRR_SANITIZER to def_bool y
Yinghai Lu [Tue, 30 Sep 2008 23:29:40 +0000 (16:29 -0700)]
x86: change MTRR_SANITIZER to def_bool y

This option has been added in v2.6.26 as a default-disabled
feature and went through several revisions since then.

The feature fixes a wide range of MTRR setup problems that BIOSes
leave us with: slow system, slow Xorg, slow system when adding lots
of RAM, etc., so we want to enable it by default for v2.6.28.

See:

  [Bug 10508] Upgrade to 4GB of RAM messes up MTRRs
  http://bugzilla.kernel.org/show_bug.cgi?id=10508

and the test results in:

   http://lkml.org/lkml/2008/9/29/273

1. hpa
reg00: base=0xc0000000 (3072MB), size=1024MB: uncachable, count=1
reg01: base=0x13c000000 (5056MB), size=  64MB: uncachable, count=1
reg02: base=0x00000000 (   0MB), size=4096MB: write-back, count=1
reg03: base=0x100000000 (4096MB), size=1024MB: write-back, count=1
reg04: base=0xbf700000 (3063MB), size=   1MB: uncachable, count=1
reg05: base=0xbf800000 (3064MB), size=   8MB: uncachable, count=1

will get
Found optimal setting for mtrr clean up
gran_size: 1M   chunk_size: 128M        num_reg: 6      lose RAM: 0M
range0: 0000000000000000 - 00000000c0000000
Setting variable MTRR 0, base: 0MB, range: 2048MB, type WB
Setting variable MTRR 1, base: 2048MB, range: 1024MB, type WB
hole: 00000000bf700000 - 00000000c0000000
Setting variable MTRR 2, base: 3063MB, range: 1MB, type UC
Setting variable MTRR 3, base: 3064MB, range: 8MB, type UC
range0: 0000000100000000 - 0000000140000000
Setting variable MTRR 4, base: 4096MB, range: 1024MB, type WB
hole: 000000013c000000 - 0000000140000000
Setting variable MTRR 5, base: 5056MB, range: 64MB, type UC

2. Dylan Taft
reg00: base=0x00000000 (   0MB), size=4096MB: write-back, count=1
reg01: base=0x100000000 (4096MB), size= 512MB: write-back, count=1
reg02: base=0x120000000 (4608MB), size= 256MB: write-back, count=1
reg03: base=0xd0000000 (3328MB), size= 256MB: uncachable, count=1
reg04: base=0xe0000000 (3584MB), size= 512MB: uncachable, count=1
reg05: base=0xc7e00000 (3198MB), size=   2MB: uncachable, count=1
reg06: base=0xc8000000 (3200MB), size= 128MB: uncachable, count=1

will get
Found optimal setting for mtrr clean up
gran_size: 1M   chunk_size: 4M  num_reg: 6      lose RAM: 0M
range0: 0000000000000000 - 00000000c8000000
Setting variable MTRR 0, base: 0MB, range: 2048MB, type WB
Setting variable MTRR 1, base: 2048MB, range: 1024MB, type WB
Setting variable MTRR 2, base: 3072MB, range: 128MB, type WB
hole: 00000000c7e00000 - 00000000c8000000
Setting variable MTRR 3, base: 3198MB, range: 2MB, type UC
rangeX: 0000000100000000 - 0000000130000000
Setting variable MTRR 4, base: 4096MB, range: 512MB, type WB
Setting variable MTRR 5, base: 4608MB, range: 256MB, type WB

3. Gabriel
reg00: base=0xd0000000 (3328MB), size= 256MB: uncachable, count=1
reg01: base=0xe0000000 (3584MB), size= 512MB: uncachable, count=1
reg02: base=0x00000000 (   0MB), size=4096MB: write-back, count=1
reg03: base=0x100000000 (4096MB), size= 512MB: write-back, count=1
reg04: base=0x120000000 (4608MB), size= 128MB: write-back, count=1
reg05: base=0x128000000 (4736MB), size=  64MB: write-back, count=1
reg06: base=0xcf600000 (3318MB), size=   2MB: uncachable, count=1

will get
Found optimal setting for mtrr clean up
gran_size: 1M   chunk_size: 16M         num_reg: 7      lose RAM: 0M
range0: 0000000000000000 - 00000000d0000000
Setting variable MTRR 0, base: 0MB, range: 2048MB, type WB
Setting variable MTRR 1, base: 2048MB, range: 1024MB, type WB
Setting variable MTRR 2, base: 3072MB, range: 256MB, type WB
hole: 00000000cf600000 - 00000000cf800000
Setting variable MTRR 3, base: 3318MB, range: 2MB, type UC
rangeX: 0000000100000000 - 000000012c000000
Setting variable MTRR 4, base: 4096MB, range: 512MB, type WB
Setting variable MTRR 5, base: 4608MB, range: 128MB, type WB
Setting variable MTRR 6, base: 4736MB, range: 64MB, type WB

4. Mika Fischer
reg00: base=0xc0000000 (3072MB), size=1024MB: uncachable, count=1
reg01: base=0x00000000 ( 0MB), size=4096MB: write-back, count=1
reg02: base=0x100000000 (4096MB), size=1024MB: write-back, count=1
reg03: base=0xbf700000 (3063MB), size= 1MB: uncachable, count=1
reg04: base=0xbf800000 (3064MB), size= 8MB: uncachable, count=1

will get
Found optimal setting for mtrr clean up
gran_size: 1M   chunk_size: 16M         num_reg: 5      lose RAM: 0M
range0: 0000000000000000 - 00000000c0000000
Setting variable MTRR 0, base: 0MB, range: 2048MB, type WB
Setting variable MTRR 1, base: 2048MB, range: 1024MB, type WB
hole: 00000000bf700000 - 00000000c0000000
Setting variable MTRR 2, base: 3063MB, range: 1MB, type UC
Setting variable MTRR 3, base: 3064MB, range: 8MB, type UC
rangeX: 0000000100000000 - 0000000140000000
Setting variable MTRR 4, base: 4096MB, range: 1024MB, type WB

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: mtrr_cleanup try gran_size to less than 1M
Yinghai Lu [Tue, 30 Sep 2008 01:54:12 +0000 (18:54 -0700)]
x86: mtrr_cleanup try gran_size to less than 1M

one have gran < 1M

reg00: base=0xd8000000 (3456MB), size= 128MB: uncachable, count=1
reg01: base=0xe0000000 (3584MB), size= 512MB: uncachable, count=1
reg02: base=0x00000000 (   0MB), size=4096MB: write-back, count=1
reg03: base=0x100000000 (4096MB), size= 512MB: write-back, count=1
reg04: base=0x120000000 (4608MB), size= 128MB: write-back, count=1
reg05: base=0xd7f80000 (3455MB), size= 512KB: uncachable, count=1

will get

Found optimal setting for mtrr clean up
gran_size: 512K         chunk_size: 2M  num_reg: 7      lose RAM: 0G
range0: 0000000000000000 - 00000000d8000000
Setting variable MTRR 0, base: 0GB, range: 2GB, type WB
Setting variable MTRR 1, base: 2GB, range: 1GB, type WB
Setting variable MTRR 2, base: 3GB, range: 256MB, type WB
Setting variable MTRR 3, base: 3328MB, range: 128MB, type WB
hole: 00000000d7f00000 - 00000000d7f80000
Setting variable MTRR 4, base: 3455MB, range: 512KB, type UC
rangeX: 0000000100000000 - 0000000128000000
Setting variable MTRR 5, base: 4GB, range: 512MB, type WB
Setting variable MTRR 6, base: 4608MB, range: 128MB, type WB

so start from 64k instead of 1M

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
16 years agox86: mtrr_cleanup prepare to make gran_size to less 1M
Yinghai Lu [Tue, 30 Sep 2008 01:54:11 +0000 (18:54 -0700)]
x86: mtrr_cleanup prepare to make gran_size to less 1M

make the print out right with size < 1M

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
16 years agox86: mtrr_cleanup safe to get more spare regs now
Yinghai Lu [Mon, 29 Sep 2008 20:39:17 +0000 (13:39 -0700)]
x86: mtrr_cleanup safe to get more spare regs now

Delay exit to make sure we can actually get the optimal result in as
many cases as possible.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
16 years agox86: mtrr_cleanup hole size should be less than half of chunk_size, v2
Yinghai Lu [Sun, 28 Sep 2008 03:26:06 +0000 (20:26 -0700)]
x86: mtrr_cleanup hole size should be less than half of chunk_size, v2

v2: should check with half of range0 size instead of chunk_size

So don't have silly big hole.

in hpa's case we could auto detect instead of adding mtrr_chunk_size in
command line.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: add mtrr_cleanup_debug command line
Yinghai Lu [Sat, 27 Sep 2008 07:30:06 +0000 (00:30 -0700)]
x86: add mtrr_cleanup_debug command line

add mtrr_cleanup_debug to print out more info about layout

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: mtrr_cleanup optimization, v2
Yinghai Lu [Sat, 27 Sep 2008 07:30:08 +0000 (00:30 -0700)]
x86: mtrr_cleanup optimization, v2

fix hpa's t61 with 4g ram:

   change layout from
(n - 1)*chunksize + chunk_size - NC
   to
n*chunksize - NC

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: don't need to go to chunksize to 4G
Yinghai Lu [Sat, 27 Sep 2008 07:30:07 +0000 (00:30 -0700)]
x86: don't need to go to chunksize to 4G

change back chunksize max to 2g
   otherwise will get strange layout in 2G ram system like
     0 - 4g WB, 2040M - 2048M UC, 2048M -  4G NC
   instead of
     0 - 2g WB, 2040M - 2048M UC

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agoIPoIB: Fix crash when path record fails after path flush
Roland Dreier [Thu, 25 Sep 2008 22:28:08 +0000 (15:28 -0700)]
IPoIB: Fix crash when path record fails after path flush

Commit ee1e2c82 ("IPoIB: Refresh paths instead of flushing them on SM
change events") changed how paths are flushed on an SM event.  This
change introduces a problem if the path record query triggered by
fails, causing path->ah to become NULL.  A later successful path query
will then trigger WARN_ON() in path_rec_completion(), and crash
because path->ah has already been freed, so the ipoib_put_ah() inside
the lock in path_rec_completion() may actually drop the last reference
(contrary to the comment that claims this is safe).

Fix this by updating path->ah and freeing old_ah only when the path
record query is successful.  This prevents the neighbour AH and that
path AH from getting out of sync.

This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=1194>

Reported-by: Rabah Salem <ravah@mellanox.com>
Debugged-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 24 Sep 2008 23:45:07 +0000 (16:45 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  netfilter: ip6t_{hbh,dst}: Rejects not-strict mode on rule insertion
  ath9k: disable MIB interrupts to fix interrupt storm
  [Bluetooth] Fix USB disconnect handling of btusb driver
  [Bluetooth] Fix wrong URB handling of btusb driver
  [Bluetooth] Fix I/O errors on MacBooks with Broadcom chips

16 years agoMerge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6
Linus Torvalds [Wed, 24 Sep 2008 23:43:44 +0000 (16:43 -0700)]
Merge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6

* 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
  i2c: Fix mailing lists in two MAINTAINERS entries
  i2c-dev: Return correct error code on class_create() failure
  i2c-powermac: Fix section for probe and remove functions

16 years agoMerge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 24 Sep 2008 23:39:50 +0000 (16:39 -0700)]
Merge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86: prevent stale state of c1e_mask across CPU offline/online, fix

16 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Wed, 24 Sep 2008 23:38:52 +0000 (16:38 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Fixe the definition of PTRS_PER_PGD
  [MIPS] au1000: Fix gpio direction

16 years agoMN10300: Make sched_clock() report time since boot
David Howells [Wed, 24 Sep 2008 16:48:31 +0000 (17:48 +0100)]
MN10300: Make sched_clock() report time since boot

Make sched_clock() report time since boot rather than time since last
timer interrupt.

Make sched_clock() expand and scale the 32-bit TSC value running at
IOCLK speed (~33MHz) to a 64-bit nanosecond counter, using cnt32_to_63()
acquired from the ARM arch and without using slow DIVU instructions
every call.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMN10300: Move asm-arm/cnt32_to_63.h to include/linux/
David Howells [Wed, 24 Sep 2008 16:48:26 +0000 (17:48 +0100)]
MN10300: Move asm-arm/cnt32_to_63.h to include/linux/

Move asm-arm/cnt32_to_63.h to include/linux/ so that MN10300 can make
use of it too.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Wed, 24 Sep 2008 23:21:56 +0000 (16:21 -0700)]
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6

16 years agonetfilter: ip6t_{hbh,dst}: Rejects not-strict mode on rule insertion
Yasuyuki Kozakai [Wed, 24 Sep 2008 22:53:39 +0000 (15:53 -0700)]
netfilter: ip6t_{hbh,dst}: Rejects not-strict mode on rule insertion

The current code ignores rules for internal options in HBH/DST options
header in packet processing if 'Not strict' mode is specified (which is not
implemented). Clearly it is not expected by user.

Kernel should reject HBH/DST rule insertion with 'Not strict' mode
in the first place.

Signed-off-by: Yasuyuki Kozakai <yasuyuki.kozakai@toshiba.co.jp>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh...
Linus Torvalds [Wed, 24 Sep 2008 22:33:50 +0000 (15:33 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
  9p: fix put_data error handling
  9p: use an IS_ERR test rather than a NULL test
  9p: introduce missing kfree
  9p-trans_fd: fix and clean up module init/exit paths
  9p-trans_fd: don't do fs segment mangling in p9_fd_poll()
  9p-trans_fd: clean up p9_conn_create()
  9p-trans_fd: fix trans_fd::p9_conn_destroy()
  9p: implement proper trans module refcounting and unregistration

16 years ago9p: fix put_data error handling
Eric Van Hensbergen [Wed, 24 Sep 2008 21:22:22 +0000 (16:22 -0500)]
9p: fix put_data error handling

Abhishek Kulkarni pointed out an inconsistency in the way
errors are returned from p9_put_data.  On deeper exploration it
seems the error handling for this path was completely wrong.
This patch adds checks for allocation problems and propagates
errors correctly.

Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
16 years ago9p: use an IS_ERR test rather than a NULL test
Julien Brunel [Wed, 24 Sep 2008 21:22:22 +0000 (16:22 -0500)]
9p: use an IS_ERR test rather than a NULL test

In case of error, the function p9_client_walk returns an ERR pointer, but
never returns a NULL pointer.  So a NULL test that comes after an IS_ERR
test should be deleted.

The semantic match that finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@match_bad_null_test@
expression x, E;
statement S1,S2;
@@
x = p9_client_walk(...)
... when != x = E
*  if (x != NULL)
S1 else S2
// </smpl>

Signed-off-by: Julien Brunel <brunel@diku.dk>
Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
16 years ago9p: introduce missing kfree
Julia Lawall [Wed, 24 Sep 2008 21:22:22 +0000 (16:22 -0500)]
9p: introduce missing kfree

Error handling code following a kmalloc should free the allocated data.

The semantic match that finds the problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@r exists@
local idexpression x;
statement S;
expression E;
identifier f,l;
position p1,p2;
expression *ptr != NULL;
@@

(
if ((x@p1 = \(kmalloc\|kzalloc\|kcalloc\)(...)) == NULL) S
|
x@p1 = \(kmalloc\|kzalloc\|kcalloc\)(...);
...
if (x == NULL) S
)
<... when != x
     when != if (...) { <+...x...+> }
x->f = E
...>
(
 return \(0\|<+...x...+>\|ptr\);
|
 return@p2 ...;
)

@script:python@
p1 << r.p1;
p2 << r.p2;
@@

print "* file: %s kmalloc %s return %s" % (p1[0].file,p1[0].line,p2[0].line)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
16 years ago9p-trans_fd: fix and clean up module init/exit paths
Tejun Heo [Wed, 24 Sep 2008 21:22:23 +0000 (16:22 -0500)]
9p-trans_fd: fix and clean up module init/exit paths

trans_fd leaked p9_mux_wq on module unload.  Fix it.  While at it,
collapse p9_mux_global_init() into p9_trans_fd_init().  It's easier to
follow this way and the global poll_tasks array is about to removed
anyway.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
16 years ago9p-trans_fd: don't do fs segment mangling in p9_fd_poll()
Tejun Heo [Wed, 24 Sep 2008 21:22:23 +0000 (16:22 -0500)]
9p-trans_fd: don't do fs segment mangling in p9_fd_poll()

p9_fd_poll() is never called with user pointers and f_op->poll()
doesn't expect its arguments to be from userland.  There's no need to
set kernel ds before calling f_op->poll() from p9_fd_poll().  Remove
it.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
16 years ago9p-trans_fd: clean up p9_conn_create()
Tejun Heo [Wed, 24 Sep 2008 21:22:23 +0000 (16:22 -0500)]
9p-trans_fd: clean up p9_conn_create()

* Use kzalloc() to allocate p9_conn and remove 0/NULL initializations.

* Clean up error return paths.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
16 years ago9p-trans_fd: fix trans_fd::p9_conn_destroy()
Tejun Heo [Wed, 24 Sep 2008 21:22:23 +0000 (16:22 -0500)]
9p-trans_fd: fix trans_fd::p9_conn_destroy()

p9_conn_destroy() first kills all current requests by calling
p9_conn_cancel(), then waits for the request list to be cleared by
waiting on p9_conn->equeue.  After that, polling is stopped and the
trans is destroyed.  This sequence has a few problems.

* Read and write works were never cancelled and the p9_conn can be
  destroyed while the works are running as r/w works remove requests
  from the list and dereference the p9_conn from them.

* The list emptiness wait using p9_conn->equeue wouldn't trigger
  because p9_conn_cancel() always clears all the lists and the only
  way the wait can be triggered is to have another task to issue a
  request between the slim window between p9_conn_cancel() and the
  wait, which isn't safe under the current implementation with or
  without the wait.

This patch fixes the problem by first stopping poll, which can
schedule r/w works, first and cancle r/w works which guarantees that
r/w works are not and will not run from that point and then calling
p9_conn_cancel() and do the rest of destruction.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
16 years ago9p: implement proper trans module refcounting and unregistration
Tejun Heo [Wed, 24 Sep 2008 21:22:23 +0000 (16:22 -0500)]
9p: implement proper trans module refcounting and unregistration

9p trans modules aren't refcounted nor were they unregistered
properly.  Fix it.

* Add 9p_trans_module->owner and reference the module on each trans
  instance creation and put it on destruction.

* Protect v9fs_trans_list with a spinlock.  This isn't strictly
  necessary as the list is manipulated only during module loading /
  unloading but it's a good idea to make the API safe.

* Unregister trans modules when the corresponding module is being
  unloaded.

* While at it, kill unnecessary EXPORT_SYMBOL on p9_trans_fd_init().

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
16 years agoath9k: disable MIB interrupts to fix interrupt storm
Luis R. Rodriguez [Tue, 23 Sep 2008 21:28:16 +0000 (14:28 -0700)]
ath9k: disable MIB interrupts to fix interrupt storm

Enabling the MIB interrupts has proven to cause an
interrupt storm after 7 hours of run. We will make use of the
MIB interrupt once we have ANI supported added so for now
to cure this we disable the interrupt.

The interrupt storm can be seen as follows after 7 hours of run
as reported by  Steven Noonan <steven@uplinklabs.net>:

18:28:38          sum   1106.00
18:28:39          sum   1037.62
18:28:40          sum   1069.00
18:28:41          sum   1167.00
18:28:42          sum   1155.00
18:28:43          sum   1339.00
18:28:44          sum  18355.00
18:28:45          sum  17845.45
18:28:46          sum  15285.00
18:28:47          sum  17511.00
18:28:48          sum  17568.69
18:28:49          sum  17704.04
18:28:50          sum  18566.67
18:28:51          sum  18913.13

at 18:28:44 the MIB interrupt kicked off and caused huge
latency which can be seen even on a video he submitted:

http://www.youtube.com/watch?v=4GeCx1gZMpA

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agoi2c: Fix mailing lists in two MAINTAINERS entries
Jean Delvare [Wed, 24 Sep 2008 11:39:22 +0000 (13:39 +0200)]
i2c: Fix mailing lists in two MAINTAINERS entries

Two MAINTAINER entries (I2C/SMBUS STUB DRIVER and SIS 96X I2C/SMBUS
DRIVER) were improperly pointing to the lm-sensors mailing list
instead of the i2c mailing list. Fix them.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Mark M. Hoffman <mhoffman@lightlink.com>
16 years agoi2c-dev: Return correct error code on class_create() failure
Sven Wegener [Wed, 24 Sep 2008 11:39:21 +0000 (13:39 +0200)]
i2c-dev: Return correct error code on class_create() failure

We need to convert the error pointer from class_create(), else we'll return the
successful return code from register_chrdev() on failure.

Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
16 years agoi2c-powermac: Fix section for probe and remove functions
Uwe Kleine-Koenig [Wed, 24 Sep 2008 11:39:21 +0000 (13:39 +0200)]
i2c-powermac: Fix section for probe and remove functions

__devexit for i2c_powermac_probe is obviously wrong.  In the definition
of struct platform_driver i2c_powermac_driver the remove function
i2c_powermac_remove is wrapped in __devexit_p, so it should be defined
using __devexit.

Signed-off-by: Uwe Kleine-Koenig <ukleinek@informatik.uni-freiburg.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
16 years agox86: prevent stale state of c1e_mask across CPU offline/online, fix
Marc Dionne [Wed, 24 Sep 2008 02:40:02 +0000 (22:40 -0400)]
x86: prevent stale state of c1e_mask across CPU offline/online, fix

Fix build error introduced by commit 4faac97d44ac27 ("x86: prevent stale
state of c1e_mask across CPU offline/online").

process_32.c needs to include idle.h to get the prototype for
c1e_remove_cpu()

Signed-off-by: Marc Dionne <marc.c.dionne@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years ago[MIPS] Fixe the definition of PTRS_PER_PGD
Jack Tan [Tue, 23 Sep 2008 14:52:34 +0000 (22:52 +0800)]
[MIPS] Fixe the definition of PTRS_PER_PGD

When we use > 4KB's page size the original definition is not consistent
with PGDIR_SIZE. For exeample, if we use 16KB page size the PGDIR_SHIFT is
(14-2) + 14 = 26, PGDIR_SIZE is 2^26,so the PTRS_PER_PGD should be:

2^32/2^26 = 2^6

but the original definition of PTRS_PER_PGD is 4096 (PGDIR_ORDER = 0).

So, this definition needs to be consistent with the PGDIR_SIZE.

And the new definition is consistent with the PGD init in pagetable_init().

Signed-off-by: Dajie Tan <jiankemeng@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
16 years ago[MIPS] au1000: Fix gpio direction
Bruno Randolf [Tue, 23 Sep 2008 17:48:36 +0000 (19:48 +0200)]
[MIPS] au1000: Fix gpio direction

When setting the direction of one GPIO pin we have to keep the state of the
other pins, hence use binary OR. Also gpio_direction_output() wants to set an
initial value, so add that too.

This fixes a problem with the USB power switch on mtx-1 boards.

Signed-off-by: Bruno Randolf <br1@einfach.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Tue, 23 Sep 2008 21:58:51 +0000 (14:58 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (23 commits)
  USB: revert recovery from transient errors
  usb: unusual devs patch for Nokia 5310 Music Xpress
  usb: ftdi_sio: add support for Domintell devices
  USB: drivers/usb/musb/: disable it on SuperH
  USB Serial: Sierra: Add MC8785 VID/PID
  USB: serial: add ZTE CDMA Tech id to option driver
  USB: ftdi_sio: Add 0x5050/0x0900 USB IDs (Papouch Quido USB 4/4)
  usb serial: ti_usb_3410_5052 obviously broken by firmware changes
  USB: fsl_usb2_udc: fix VDBG() format string
  USB: unusual_devs addition for RockChip MP3 player
  USB: SERIAL CP2101 add device IDs
  usb-serial: Add Siemens EF81 to PL-2303 hack triggers
  USB: fix EHCI periodic transfers
  usb: musb: fix include path
  USB: Fixing Nokia 3310c in storage mode
  usb gadget: fix omap_udc DMA regression
  USB: update of Documentation/usb/anchors.txt
  USB: fix hcd interrupt disabling
  USB: Correct Sierra Wireless USB EVDO Modem Device ID
  USB: Fix the Nokia 6300 storage-mode.
  ...

16 years agoMerge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 23 Sep 2008 21:57:36 +0000 (14:57 -0700)]
Merge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  timers: fix build error in !oneshot case
  x86: c1e_idle: don't mark TSC unstable if CPU has invariant TSC
  x86: prevent C-states hang on AMD C1E enabled machines
  clockevents: prevent mode mismatch on cpu online
  clockevents: check broadcast device not tick device
  clockevents: prevent stale tick_next_period for onlining CPUs
  x86: prevent stale state of c1e_mask across CPU offline/online
  clockevents: prevent cpu online to interfere with nohz

16 years agoMerge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 23 Sep 2008 21:57:22 +0000 (14:57 -0700)]
Merge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: fix init_hrtick() section mismatch warning

16 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Tue, 23 Sep 2008 21:56:45 +0000 (14:56 -0700)]
Merge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86: fix 27-rc crash on vsmp due to paravirt during module load
  x86, oprofile: BUG scheduling while atomic
  AMD IOMMU: protect completion wait loop with iommu lock
  AMD IOMMU: set iommu sunc flag after command queuing

16 years agoUSB: revert recovery from transient errors
Alan Stern [Mon, 22 Sep 2008 18:43:08 +0000 (14:43 -0400)]
USB: revert recovery from transient errors

This patch (as1135) essentially reverts the major parts of two earlier
patches to usbcore, because they ended up causing a regression.

Trying to recover from transient communication errors can lead to
other problems, because operations that failed during the error period
are not always retried.  The simplest example is the initial
Set-Config request sent after device enumeration; if it gets lost then
it will not be retried and the device will remain unconfigured.

This patch restores the old behavior in which any port disconnect or
port disable causes the entire device structure to be removed, fixing a
reported regression.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Frans Pop <elendil@planet.nl>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb: unusual devs patch for Nokia 5310 Music Xpress
David Almaroad [Mon, 22 Sep 2008 22:00:11 +0000 (15:00 -0700)]
usb: unusual devs patch for Nokia 5310 Music Xpress

The Nokia 5310 Music Xpress phone reports one too many sectors in
usb-storage mode.  This patch resolves that.

Signed-off-by: David Almaroad <dalmaroad@gmail.com>
Cc: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb: ftdi_sio: add support for Domintell devices
Gaetan Carlier [Mon, 22 Sep 2008 22:00:09 +0000 (15:00 -0700)]
usb: ftdi_sio: add support for Domintell devices

Support for Domintell devices (FTDI FT232BM based) : DGQG and DUSB01
module.  PIDs were missing.

Signed-off-by: Gaetan Carlier <gcpatch@gmail.com>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Cc: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: drivers/usb/musb/: disable it on SuperH
Andrew Morton [Mon, 22 Sep 2008 22:00:08 +0000 (15:00 -0700)]
USB: drivers/usb/musb/: disable it on SuperH

In file included from drivers/usb/musb/musb_core.h:59,
                 from drivers/usb/musb/musb_core.c:108:
drivers/usb/musb/musb_io.h:42: error: conflicting types for '__raw_readsl'
/usr/src/devel/arch/sh/include/asm/io.h:112: error: previous declaration of '__raw_readsl' was here
drivers/usb/musb/musb_io.h:42: error: conflicting types for '__raw_readsl'
/usr/src/devel/arch/sh/include/asm/io.h:112: error: previous declaration of '__raw_readsl' was here
drivers/usb/musb/musb_io.h:44: error: conflicting types for 'readsw'
/usr/src/devel/arch/sh/include/asm/io.h:164: error: previous definition of 'readsw' was here
drivers/usb/musb/musb_io.h:46: error: conflicting types for 'readsb'
/usr/src/devel/arch/sh/include/asm/io.h:163: error: previous definition of 'readsb' was here
drivers/usb/musb/musb_io.h:49: error: conflicting types for '__raw_writesl'
/usr/src/devel/arch/sh/include/asm/io.h:111: error: previous declaration of '__raw_writesl' was here
drivers/usb/musb/musb_io.h:49: error: conflicting types for '__raw_writesl'
/usr/src/devel/arch/sh/include/asm/io.h:111: error: previous declaration of '__raw_writesl' was here
drivers/usb/musb/musb_io.h:51: error: conflicting types for 'writesw'
/usr/src/devel/arch/sh/include/asm/io.h:164: error: previous definition of 'writesw' was here
drivers/usb/musb/musb_io.h:53: error: conflicting types for 'writesb'
/usr/src/devel/arch/sh/include/asm/io.h:163: error: previous definition of 'writesb' was here

Cc: Karsten Keil <kkeil@suse.de>
Cc: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB Serial: Sierra: Add MC8785 VID/PID
Kevin Lloyd [Wed, 17 Sep 2008 16:03:38 +0000 (09:03 -0700)]
USB Serial: Sierra: Add MC8785 VID/PID

Add another MC8785 VID/PID

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: serial: add ZTE CDMA Tech id to option driver
Otavio Salvador [Wed, 17 Sep 2008 17:40:46 +0000 (14:40 -0300)]
USB: serial: add ZTE CDMA Tech id to option driver

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: ftdi_sio: Add 0x5050/0x0900 USB IDs (Papouch Quido USB 4/4)
Jaroslav Kysela [Tue, 16 Sep 2008 13:46:50 +0000 (15:46 +0200)]
USB: ftdi_sio: Add 0x5050/0x0900 USB IDs (Papouch Quido USB 4/4)

USB: ftdi_sio: Add 0x5050/0x0900 USB IDs (Papouch Quido USB 4/4)

Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb serial: ti_usb_3410_5052 obviously broken by firmware changes
Chris Adams [Wed, 10 Sep 2008 21:11:38 +0000 (14:11 -0700)]
usb serial: ti_usb_3410_5052 obviously broken by firmware changes

While making some other changes to ti_usb_3410_5052, I noticed that the
changes made to move the firmware loading to a separate function are
broken (in ti_download_firmware(), status is set to -ENOMEM and never
changed).  This means the driver will never initialize the device
properly.  It looks like status was supposed to get the result of
ti_do_download().

Signed-off-by: Chris Adams <cmadams@hiwaay.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fsl_usb2_udc: fix VDBG() format string
Peter Korsgaard [Mon, 8 Sep 2008 08:08:59 +0000 (10:08 +0200)]
USB: fsl_usb2_udc: fix VDBG() format string

Fixes:
drivers/usb/gadget/fsl_usb2_udc.c: In function 'dr_controller_setup':
drivers/usb/gadget/fsl_usb2_udc.c:229: warning: format '%p' expects type
'void *', but argument 3 has type 'int'

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: unusual_devs addition for RockChip MP3 player
Alan Stern [Thu, 4 Sep 2008 13:56:35 +0000 (09:56 -0400)]
USB: unusual_devs addition for RockChip MP3 player

This patch (as1136) adds an unusual_devs entry for a version of the
RockChip MP3 player which can't handle the MODE SENSE command used for
write-protect detection.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: SERIAL CP2101 add device IDs
Craig Shelley [Thu, 4 Sep 2008 20:28:21 +0000 (21:28 +0100)]
USB: SERIAL CP2101 add device IDs

Signed-off-by: Craig Shelley <craig@microtron.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb-serial: Add Siemens EF81 to PL-2303 hack triggers
Andreas Bombe [Sat, 13 Sep 2008 23:58:55 +0000 (01:58 +0200)]
usb-serial: Add Siemens EF81 to PL-2303 hack triggers

This hardware needs the pl2303 hack in order to work properly :(

Signed-off-by: Andreas Bombe <aeb@debian.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix EHCI periodic transfers
David Brownell [Wed, 27 Aug 2008 06:35:04 +0000 (23:35 -0700)]
USB: fix EHCI periodic transfers

As noted by Stefan Neis <Stefan.Neis@kobil.com>, we had a recent
regression with EHCI periodic transfers, in some (seemingly not
all that common) cases.

The root cause was that the schedule activation was only loosely
coupled to the addition or removal of transfers, so two different
execution contexts could both think they had to deactivate (or
conversely activate) the schedule.  So this fix tightens that
coupling, managing it more like a refcount.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb: musb: fix include path
Felipe Balbi [Sat, 30 Aug 2008 16:42:02 +0000 (19:42 +0300)]
usb: musb: fix include path

headers were moved, fixing.

Signed-off-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: Fixing Nokia 3310c in storage mode
Filip Joelsson [Sun, 31 Aug 2008 06:22:16 +0000 (08:22 +0200)]
USB: Fixing Nokia 3310c in storage mode

I had trouble connecting my cell phone as a storage device - so I added
it to the unusual_devs.h list. I had trouble with the bcdDeviceMin and
Max values - so after some experimenting I made it pretty inclusive.

From: Filip Joelsson <filip@blueturtle.nu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb gadget: fix omap_udc DMA regression
David Brownell [Mon, 1 Sep 2008 01:04:27 +0000 (18:04 -0700)]
usb gadget: fix omap_udc DMA regression

This resolves another regression caused by the "use omap_read/write
instead of __REG" patch:  the hardware address used for DMA to/from
the UDC became wrong.  Bug noted by Russell King.

Reported-by: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: update of Documentation/usb/anchors.txt
Oliver Neukum [Tue, 2 Sep 2008 08:52:08 +0000 (10:52 +0200)]
USB: update of Documentation/usb/anchors.txt

The extended anchor API is documented

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix hcd interrupt disabling
Geoff Levand [Fri, 22 Aug 2008 21:13:00 +0000 (14:13 -0700)]
USB: fix hcd interrupt disabling

Commit de85422b94ddb23c021126815ea49414047c13dc, 'USB: fix interrupt
disabling for HCDs with shared interrupt handlers' changed usb_add_hcd()
to strip IRQF_DISABLED from irqflags prior to calling request_irq()
with the justification that such a removal was necessary for shared
interrupts to work properly.  Unfortunately, the change in that commit
unconditionally removes the IRQF_DISABLED flag, causing problems on
platforms that don't use a shared interrupt but require IRQF_DISABLED.
This change adds a check for IRQF_SHARED prior to removing the
IRQF_DISABLED flag.

Fixes the PS3 system startup hang reported with recent Fedora and
OpenSUSE kernels.

Note that this problem is hidden when CONFIG_LOCKDEP=y (ps3_defconfig),
as local_irq_enable_in_hardirq() is defined as a null statement for
that config.

CC: stable <stable@kernel.org>
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Stefan Becker <Stefan.Becker@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: Correct Sierra Wireless USB EVDO Modem Device ID
Tony Murray [Fri, 22 Aug 2008 22:30:44 +0000 (17:30 -0500)]
USB: Correct Sierra Wireless USB EVDO Modem Device ID

I was trying to figure out why my device wasn't supported by the
drivers/usb/serial/sierra.c driver, while looking throught the device
IDs I spotted what I believe to be a typo in the device IDs.  Please
apply the following patch

If you look down further, there is another HP wireless broadband card,
which has a vendor ID of 03f0, like my device.  Below is my "lsusb -v
-d 03f0:1b1d".

Bus 001 Device 005: ID 03f0:1b1d Hewlett-Packard
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               1.10
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0
  bDeviceProtocol         0
  bMaxPacketSize0        64
  idVendor           0x03f0 Hewlett-Packard
  idProduct          0x1b1d
  bcdDevice            0.01
  iManufacturer           1 HP
  iProduct                2 HP ev2200 1xEV-DO Broadband Wireless Module
  iSerial                 0
  bNumConfigurations      1
  Configuration Descriptor:
    bLength                 9
    bDescriptorType         2
    wTotalLength           67
    bNumInterfaces          1
    bConfigurationValue     1
    iConfiguration          0
    bmAttributes         0xe0
      Self Powered
      Remote Wakeup
    MaxPower                0mA
    Interface Descriptor:
      bLength                 9
      bDescriptorType         4
      bInterfaceNumber        0
      bAlternateSetting       0
      bNumEndpoints           7
      bInterfaceClass       255 Vendor Specific Class
      bInterfaceSubClass    255 Vendor Specific Subclass
      bInterfaceProtocol    255 Vendor Specific Protocol
      iInterface              3 Data Interface
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x81  EP 1 IN
        bmAttributes            3
          Transfer Type            Interrupt
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0010  1x 16 bytes
        bInterval             128
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x82  EP 2 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x02  EP 2 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x84  EP 4 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x04  EP 4 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x85  EP 5 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x05  EP 5 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0
Device Status:     0x0000
  (Bus Powered)

From: Tony Murray <murraytony@gmail.com>
Cc: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: Fix the Nokia 6300 storage-mode.
Richard Nauber [Tue, 26 Aug 2008 21:34:11 +0000 (23:34 +0200)]
USB: Fix the Nokia 6300 storage-mode.

This patch fixes that behavior:

...
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] Sense Key : 0x0 [current]
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] ASC=0x0 ASCQ=0x0
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] Sense Key : 0x0 [current]
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] ASC=0x0 ASCQ=0x0
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] Sense Key : 0x0 [current]
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] ASC=0x0 ASCQ=0x0
Aug 26 22:39:23 rnauber sd 2:0:0:0: [sda] Sense Key : 0x0 [current]
...

cat  /proc/bus/usb/devices
...
T:  Bus=04 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  3 Spd=12  MxCh= 0
D:  Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=0421 ProdID=04fa Rev= 6.01
S:  Manufacturer=Nokia
S:  Product=Nokia 6300
S:  SerialNumber=35XXXXXXXX
C:* #Ifs= 1 Cfg#= 1 Atr=c0 MxPwr=100mA
I:* If#= 0 Alt= 0 #EPs= 2 Cls=08(stor.) Sub=06 Prot=50 Driver=usb-storage
E:  Ad=81(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=01(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
...

From: Richard Nauber <RichardNauber@web.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: ehci: fix some ehci hangs and crashes
David Brownell [Tue, 26 Aug 2008 21:43:46 +0000 (14:43 -0700)]
USB: ehci: fix some ehci hangs and crashes

I noticed that the "Refactor "if (handshake()) state = HC_STATE_HALT"
patch from earlier this year perpetuated a potential problem:  it can
mark the controller as halted when it's still running (but not acting
as, perhaps wrongly, expected).

That caused some hangs and crashes, rather than more polite failure
modes of a truly halted controller.  This patch forces a true halt,
and emits a (previously missing) diagnostic.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB Serial: Sierra: Device addition & version rev
Kevin Lloyd [Tue, 26 Aug 2008 02:20:40 +0000 (19:20 -0700)]
USB Serial: Sierra: Device addition & version rev

This patch adds devices to the sierra driver and rev's the driver version.

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB Storage: Sierra: Non-configurable TRU-Install
Kevin Lloyd [Wed, 27 Aug 2008 01:30:45 +0000 (18:30 -0700)]
USB Storage: Sierra: Non-configurable TRU-Install

This patch alters the Sierra Mass Storage patch so that it is non-configurable.

Signed-off-by: Kevin Lloyd <klloyd@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes...
Linus Torvalds [Tue, 23 Sep 2008 19:15:50 +0000 (12:15 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
  PCI: fix compiler warnings in pci_get_subsys()
  PCI: Fix pcie_aspm=force

16 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Tue, 23 Sep 2008 17:21:58 +0000 (10:21 -0700)]
Merge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] kexec fails on systems with blocks of uncached memory
  [IA64] Ski simulator doesn't need check_sal_cache_flush

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Tue, 23 Sep 2008 17:21:45 +0000 (10:21 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: ASoC: maintainers - update email address for Liam Girdwood

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Tue, 23 Sep 2008 15:18:17 +0000 (08:18 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  [WATCHDOG] unlocked_ioctl changes
  [WATCHDOG] wdt285: fix sparse warnings
  [WATCHDOG] ibmasr: remove unnecessary spin_unlock()

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Tue, 23 Sep 2008 15:14:14 +0000 (08:14 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Fix missing devices due to PCI bridge test in of_create_pci_dev().
  sparc64: Fix disappearing PCI devices on e3500.

16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 23 Sep 2008 15:14:04 +0000 (08:14 -0700)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  ath9k: Fix IRQ nobody cared issue with ath9k
  wireless: zd1211rw: add device ID fix wifi dongle "trust nw-3100"
  ath9k: connectivity is lost after Group rekeying is done

16 years agoatmel_serial: update the powersave handler to match serial core
Anti Sullin [Mon, 22 Sep 2008 20:57:54 +0000 (13:57 -0700)]
atmel_serial: update the powersave handler to match serial core

This problem seems to be unnoticed so far:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=b3b708fa2780cd2b5d8266a8f0c3a1cab364d4d2

has changed the serial core behavior to not to suspend the port if the
device is enabled as a wakeup source.  If the AT91 system goes to slow
clock mode, the port should be suspended always and the clocks should be
switched off.  The patch attached updates the atmel_serial driver to match
the changes in serial core.

Also, the interrupts are disabled when the clock is disabled.  If we
disable the clock with interrupts enabled, an interrupt may get stuck.  If
this is the DBGU interrupt, this blocks the OR logic at system controller
and thus all other sysc interrupts.

Signed-off-by: Anti Sullin <anti.sullin@artecdesign.ee>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Cc: Michael Trimarchi <trimarchimichael@yahoo.it>
Cc: Andrew Victor <linux@maxim.org.za>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomemcg: check under limit at shrink_usage
Daisuke Nishimura [Mon, 22 Sep 2008 20:57:52 +0000 (13:57 -0700)]
memcg: check under limit at shrink_usage

Current memory cgroup(both in mainline and -mm) doesn't account swap
caches as memory(swap cache support is dropped temporarily now).

So try_to_free_mem_cgroup_pages doesn't reflect the count of pages that
have been moved to swap cache.

But this makes mem_cgroup_shrink_usage fail easily if most of the pages
are anon/shmem, and then shmem_getpage returns -ENOMEM and the process
will be killed.

This patch adds res_counter_check_under_limit to avoid these cases.

BTW, even if swap cache support is enabled again, if a process is moved to
another cgroup, which has been just made, between precharge and
shrink_usage in shmem_getpage, shrink_usage may fail just because there is
no pages to reclaim.

So this change would make sense anyway.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoDocumentation/sysctl/kernel.txt: fix softlockup_thresh description
Andrew Morton [Mon, 22 Sep 2008 20:57:51 +0000 (13:57 -0700)]
Documentation/sysctl/kernel.txt: fix softlockup_thresh description

- s/s/seconds/

- s/10 seconds/60 seconds/

- Mention the zero-disables-it feature.

Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomm: tiny-shmem fix lock ordering: mmap_sem vs i_mutex
Nick Piggin [Mon, 22 Sep 2008 20:57:50 +0000 (13:57 -0700)]
mm: tiny-shmem fix lock ordering: mmap_sem vs i_mutex

tiny-shmem calls do_truncate in shmem_file_setup.  do_truncate takes
i_mutex, and shmem_file_setup is called with mmap_sem held.  However
i_mutex nests outside mmap_sem.

Copy the code in shmem.c to avoid this problem.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Nick Piggin <npiggin@suse.de>
Reported-and-tested-by: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agosys_paccept: disable paccept() until API design is resolved
Michael Kerrisk [Mon, 22 Sep 2008 20:57:49 +0000 (13:57 -0700)]
sys_paccept: disable paccept() until API design is resolved

The reasons for disabling paccept() are as follows:

* The API is more complex than needed.  There is AFAICS no demonstrated
  use case that the sigset argument of this syscall serves that couldn't
  equally be served by the use of pselect/ppoll/epoll_pwait + traditional
  accept().  Roland seems to concur with this opinion
  (http://thread.gmane.org/gmane.linux.kernel/723953/focus=732255).  I
  have (more than once) asked Ulrich to explain otherwise
  (http://thread.gmane.org/gmane.linux.kernel/723952/focus=731018), but he
  does not respond, so one is left to assume that he doesn't know of such
  a case.

* The use of a sigset argument is not consistent with other I/O APIs
  that can block on a single file descriptor (e.g., read(), recv(),
  connect()).

* The behavior of paccept() when interrupted by a signal is IMO strange:
  the kernel restarts the system call if SA_RESTART was set for the
  handler.  I think that it should not do this -- that it should behave
  consistently with paccept()/ppoll()/epoll_pwait(), which never restart,
  regardless of SA_RESTART.  The reasoning here is that the very purpose
  of paccept() is to wait for a connection or a signal, and that
  restarting in the latter case is probably never useful.  (Note: Roland
  disagrees on this point, believing that rather paccept() should be
  consistent with accept() in its behavior wrt EINTR
  (http://thread.gmane.org/gmane.linux.kernel/723953/focus=732255).)

I believe that instead, a simpler API, consistent with Ulrich's other
recent additions, is preferable:

accept4(int fd, struct sockaddr *sa, socklen_t *salen, ind flags);

(This simpler API was originally proposed by Ulrich:
http://thread.gmane.org/gmane.linux.network/92072)

If this simpler API is added, then if we later decide that the sigset
argument really is required, then a suitable bit in 'flags' could be added
to indicate the presence of the sigset argument.

At this point, I am hoping we either will get a counter-argument from
Ulrich about why we really do need paccept()'s sigset argument, or that he
will resubmit the original accept4() patch.

Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: David Miller <davem@davemloft.net>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: Alan Cox <alan@redhat.com>
Cc: Ulrich Drepper <drepper@redhat.com>
Cc: Jakub Jelinek <jakub@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoDocumentation/DMA-mapping.txt: update for pci_dma_mapping_error() changes
Marin Mitov [Mon, 22 Sep 2008 20:57:46 +0000 (13:57 -0700)]
Documentation/DMA-mapping.txt: update for pci_dma_mapping_error() changes

Make the example code consistent with changed API.

Signed-off-by: Marin Mitov <mitov@ispp.bas.bg>
Acked-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agokexec: fix segmentation fault in kimage_add_entry
Jonathan Steel [Mon, 22 Sep 2008 20:57:45 +0000 (13:57 -0700)]
kexec: fix segmentation fault in kimage_add_entry

A segmentation fault can occur in kimage_add_entry in kexec.c when loading
a kernel image into memory.  The fault occurs because a page is requested
by calling kimage_alloc_page with gfp_mask GFP_KERNEL and the function may
actually return a page with gfp_mask GFP_HIGHUSER.  The high mem page is
returned because it was swapped with the kernel page due to the kernel
page being a page that will shortly be copied to.

This patch ensures that kimage_alloc_page returns a page that was created
with the correct gfp flags.

I have verified the change and fixed the whitespace damage of the original
patch.  Jonathan did a great job of tracking this down after he hit the
problem.  -- Eric

Signed-off-by: Jonathan Steel <jon.steel@esentire.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agokernel-doc: allow structs whose members are all private
Randy Dunlap [Mon, 22 Sep 2008 20:57:44 +0000 (13:57 -0700)]
kernel-doc: allow structs whose members are all private

Struct members may be marked as private by using
/* private: */
before them, as noted in Documentation/kernel-doc-nano-HOWTO.txt

Fix kernel-doc to handle structs whose members are all private;
otherwise invalid XML is generated:

xmlto: input does not validate (status 3)
linux-2.6.27-rc6-git4/Documentation/DocBook/debugobjects.xml:146: element variablelist: validity error : Element variablelist content does not follow the DTD, expecting ((title , titleabbrev?)? , varlistentry+), got ()
Document linux-2.6.27-rc6-git4/Documentation/DocBook/debugobjects.xml does not validate
make[1]: *** [Documentation/DocBook/debugobjects.html] Error 3

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Reported-by: Roland McGrath <roland@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agosmb.h: do not include linux/time.h in userspace
Kirill A. Shutemov [Mon, 22 Sep 2008 20:57:43 +0000 (13:57 -0700)]
smb.h: do not include linux/time.h in userspace

linux/time.h conflicts with time.h from glibc

It breaks building smbmount from samba.  It's regression introduced by
commit 76308da (" smb.h: uses struct timespec but didn't include
linux/time.h").

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Cc: <stable@kernel.org> [2.6.26.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoibmasr: remove unnecessary spin_unlock()
Akinobu Mita [Mon, 22 Sep 2008 20:57:41 +0000 (13:57 -0700)]
ibmasr: remove unnecessary spin_unlock()

__asr_toggle() is always called with asr_lock held.
But there is unnecessary spin_unlock() call in __asr_toggle().

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Wim Van Sebroeck <wim@iguana.be>
Acked-by: Andrey Panin <pazke@donpac.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoALSA: ASoC: maintainers - update email address for Liam Girdwood
Liam Girdwood [Thu, 18 Sep 2008 13:36:37 +0000 (14:36 +0100)]
ALSA: ASoC: maintainers - update email address for Liam Girdwood

This patch updates the maintainers email address for Liam Girdwood and
adds a URL for the ASoC website.

Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
16 years agotimers: fix build error in !oneshot case
Ingo Molnar [Tue, 23 Sep 2008 11:00:57 +0000 (13:00 +0200)]
timers: fix build error in !oneshot case

 kernel/time/tick-common.c: In function ‘tick_setup_periodic’:
 kernel/time/tick-common.c:113: error: implicit declaration of function ‘tick_broadcast_oneshot_active’

Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: c1e_idle: don't mark TSC unstable if CPU has invariant TSC
Andreas Herrmann [Thu, 18 Sep 2008 19:12:10 +0000 (21:12 +0200)]
x86: c1e_idle: don't mark TSC unstable if CPU has invariant TSC

Impact: Functional TSC is marked unstable on AMD family 0x10 and 0x11 CPUs.

This would be wrong because for those CPUs "invariant TSC" means:

   "The TSC counts at the same rate in all P-states, all C states, S0,
   or S1"

(See "Processor BIOS and Kernel Developer's Guides" for those CPUs.)

[ tglx: Changed C1E to AMD C1E in the printks to avoid confusion
with Intel C1E ]

Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agox86: prevent C-states hang on AMD C1E enabled machines
Thomas Gleixner [Mon, 22 Sep 2008 17:02:25 +0000 (19:02 +0200)]
x86: prevent C-states hang on AMD C1E enabled machines

Impact: System hang when AMD C1E machines switch into C2/C3

AMD C1E enabled systems do not work with normal ACPI C-states
even if the BIOS is advertising them. Limit the C-states to
C1 for the ACPI processor idle code.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agoclockevents: prevent mode mismatch on cpu online
Thomas Gleixner [Mon, 22 Sep 2008 17:04:02 +0000 (19:04 +0200)]
clockevents: prevent mode mismatch on cpu online

Impact: timer hang on CPU online observed on AMD C1E systems

When a CPU is brought online then the broadcast machinery can
be in the one shot state already. Check this and setup the timer
device of the new CPU in one shot mode so the broadcast code
can pick up the next_event value correctly.

Another AMD C1E oddity, as we switch to broadcast immediately and
not after the full bring up via the ACPI cpu idle code.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agoclockevents: check broadcast device not tick device
Thomas Gleixner [Mon, 22 Sep 2008 17:02:25 +0000 (19:02 +0200)]
clockevents: check broadcast device not tick device

Impact: Possible hang on CPU online observed on AMD C1E machines.

The broadcast setup code looks at the mode of the tick device to
determine whether it needs to be shut down or setup. This is wrong
when the broadcast mode is set to one shot already. This can happen
when a CPU is brought online as it goes through the periodic setup
first.

The problem went unnoticed as sane systems do not call into that code
before the switch to one shot for the clock event device happens.
The AMD C1E idle routine switches over immediately and thereby shuts
down the just setup device before the first interrupt happens.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agoclockevents: prevent stale tick_next_period for onlining CPUs
Thomas Gleixner [Mon, 22 Sep 2008 16:56:01 +0000 (18:56 +0200)]
clockevents: prevent stale tick_next_period for onlining CPUs

Impact: possible hang on CPU onlining in timer one shot mode.

The tick_next_period variable is only used during boot on nohz/highres
enabled systems, but for CPU onlining it needs to be maintained when
the per cpu clock events device operates in one shot mode.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agox86: prevent stale state of c1e_mask across CPU offline/online
Thomas Gleixner [Mon, 22 Sep 2008 16:54:29 +0000 (18:54 +0200)]
x86: prevent stale state of c1e_mask across CPU offline/online

Impact: hang which happens across CPU offline/online on AMD C1E systems.

When a CPU goes offline then the corresponding bit in the broadcast
mask is cleared. For AMD C1E enabled CPUs we do not reenable the
broadcast when the CPU comes online again as we do not clear the
corresponding bit in the c1e_mask, which keeps track which CPUs
have been switched to broadcast already. So on those !$@#& machines
we never switch back to broadcasting after a CPU offline/online cycle.

Clear the bit when the CPU plays dead.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agoclockevents: prevent cpu online to interfere with nohz
Thomas Gleixner [Mon, 22 Sep 2008 16:46:37 +0000 (18:46 +0200)]
clockevents: prevent cpu online to interfere with nohz

Impact: rare hang which can be triggered on CPU online.

tick_do_timer_cpu keeps track of the CPU which updates jiffies
via do_timer. The value -1 is used to signal, that currently no
CPU is doing this. There are two cases, where the variable can
have this state:

 boot:
    necessary for systems where the boot cpu id can be != 0

 nohz long idle sleep:
    When the CPU which did the jiffies update last goes into
    a long idle sleep it drops the update jiffies duty so
    another CPU which is not idle can pick it up and keep
    jiffies going.

Using the same value for both situations is wrong, as the CPU online
code can see the -1 state when the timer of the newly onlined CPU is
setup. The setup for a newly onlined CPU goes through periodic mode
and can pick up the do_timer duty without being aware of the nohz /
highres mode of the already running system.

Use two separate states and make them constants to avoid magic
numbers confusion.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
16 years agosched: fix init_hrtick() section mismatch warning
Rakib Mullick [Mon, 22 Sep 2008 21:55:45 +0000 (14:55 -0700)]
sched: fix init_hrtick() section mismatch warning

LD      kernel/built-in.o
WARNING: kernel/built-in.o(.text+0x326): Section mismatch in reference
from the function init_hrtick() to the variable
.cpuinit.data:hotplug_hrtick_nb.8
The function init_hrtick() references
the variable __cpuinitdata hotplug_hrtick_nb.8.
This is often because init_hrtick lacks a __cpuinitdata
annotation or the annotation of hotplug_hrtick_nb.8 is wrong.

Signed-off-by: Md.Rakib H. Mullick <rakib.mullick@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years agox86: fix 27-rc crash on vsmp due to paravirt during module load
Ravikiran G Thirumalai [Tue, 23 Sep 2008 05:58:47 +0000 (22:58 -0700)]
x86: fix 27-rc crash on vsmp due to paravirt during module load

27-rc fails to boot up if configured to use modules.

Turns out vsmp_patch was marked __init, and vsmp_patch being the
pvops 'patch' routine for vsmp, a call to vsmp_patch just turns out
to execute a code page with series of 0xcc (POISON_FREE_INITMEM -- int3).

vsmp_patch has been marked with __init ever since pvops, however,
apply_paravirt can be called during module load causing calls to
freed memory location.

Since apply_paravirt can only be called during init/module load, make
vsmp_patch with "__init_or_module"

Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
16 years ago[WATCHDOG] unlocked_ioctl changes
Wim Van Sebroeck [Thu, 18 Sep 2008 12:26:15 +0000 (12:26 +0000)]
[WATCHDOG] unlocked_ioctl changes

Fix some drivers so that they use the unlocked_ioctl call.

Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
16 years ago[WATCHDOG] wdt285: fix sparse warnings
Ben Dooks [Tue, 16 Sep 2008 10:31:01 +0000 (11:31 +0100)]
[WATCHDOG] wdt285: fix sparse warnings

The wdt285.c watchdog driver is producing a number of
sparse errors due to missing __user attributes to calls
to put_user and copy_to_user, as well as in the prototype
of watchdog_write.

wdt285.c:144:21: warning: incorrect type in argument 1 (different address spaces)
wdt285.c:144:21:    expected void [noderef] <asn:1>*to
wdt285.c:144:21:    got void *<noident>
wdt285.c:150:9: warning: incorrect type in initializer (different address spaces)
wdt285.c:150:9:    expected int const [noderef] <asn:1>*register __p
wdt285.c:150:9:    got int *<noident>
wdt285.c:159:9: warning: incorrect type in initializer (different address spaces)
wdt285.c:159:9:    expected int const [noderef] <asn:1>*register __p
wdt285.c:159:9:    got int *<noident>
wdt285.c:174:9: warning: incorrect type in initializer (different address spaces)
wdt285.c:174:9:    expected int const [noderef] <asn:1>*register __p
wdt285.c:174:9:    got int *<noident>
wdt285.c:183:12: warning: incorrect type in initializer (incompatible argument 2 (different address spaces))
wdt285.c:183:12:    expected int ( *write )( ... )
wdt285.c:183:12:    got int ( static [toplevel] *<noident> )( ... )

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
16 years ago[WATCHDOG] ibmasr: remove unnecessary spin_unlock()
Akinobu Mita [Sat, 13 Sep 2008 09:47:19 +0000 (18:47 +0900)]
[WATCHDOG] ibmasr: remove unnecessary spin_unlock()

__asr_toggle() is always called with asr_lock held.
But there is unnecessary spin_unlock() call in __asr_toggle().

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
Acked-by: Andrey Panin <pazke@donpac.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
16 years agosparc64: Fix missing devices due to PCI bridge test in of_create_pci_dev().
David S. Miller [Mon, 22 Sep 2008 22:42:24 +0000 (15:42 -0700)]
sparc64: Fix missing devices due to PCI bridge test in of_create_pci_dev().

Just like in the arch/sparc64/kernel/of_device.c code fix commit
071d7f4c3b411beae08d27656e958070c43b78b4 ("sparc64: Fix SMP bootup
with CONFIG_STACK_DEBUG or ftrace.") we have to check the OF device
node name for "pci" instead of relying upon the 'device_type' property
being there on all PCI bridges.

Tested by Meelis Roos, and confirmed to make the PCI QFE devices
reappear on the E3500 system.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Mon, 22 Sep 2008 22:29:05 +0000 (15:29 -0700)]
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6

16 years ago[Bluetooth] Fix USB disconnect handling of btusb driver
Marcel Holtmann [Mon, 22 Sep 2008 22:16:36 +0000 (00:16 +0200)]
[Bluetooth] Fix USB disconnect handling of btusb driver

The USB transport specification for Bluetooth splits the ACL and SCO
handling into two separate interfaces. In Linux it possible to probe
and disconnect these interfaces independently. So make sure that both
interfaces are tightly bound together.

This fixes the suspend regression that some people have expierenced.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
16 years ago[Bluetooth] Fix wrong URB handling of btusb driver
Marcel Holtmann [Mon, 22 Sep 2008 22:16:36 +0000 (00:16 +0200)]
[Bluetooth] Fix wrong URB handling of btusb driver

The btusb driver contains two typos that result in some buggy behavior,
but the impact is not immediately visible.

During initialization the submitting of interrupt URBs might fail and
then make sure to remove the correct flag and not one of the hci_dev
flags.

When closing down the interface make sure to kill the anchor for the
ISOC URBs and not kill the interrupt URBs twice.

Also cancel any scheduled work when closing down the interface.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
16 years ago[Bluetooth] Fix I/O errors on MacBooks with Broadcom chips
Marcel Holtmann [Mon, 22 Sep 2008 22:16:35 +0000 (00:16 +0200)]
[Bluetooth] Fix I/O errors on MacBooks with Broadcom chips

The newer MacBooks contain a Broadcom based Bluetooth chip and to make
this work properly, HCI_Reset must be send first. If HCI_Reset is not
used then a lot of I/O errors show up and its triggers packets from
non-existent ACL links.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>