linux-2.6-microblaze.git
17 years agojmr3927: do not call tc35815_killall().
Ralf Baechle [Tue, 27 Feb 2007 18:15:01 +0000 (18:15 +0000)]
jmr3927: do not call tc35815_killall().

No need to stop tc35815 before resetting the board.  This fixes the
build of tc35815 as a module.  This also means there is no caller of
tc35815_killall left, so remove that function also.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoMerge branch 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/linvil...
Jeff Garzik [Sat, 3 Mar 2007 00:40:24 +0000 (19:40 -0500)]
Merge branch 'upstream-fixes' of git://git./linux/kernel/git/linville/wireless-2.6 into upstream-fixes

17 years ago[AOE]: Add get_unaligned() calls where needed.
David S. Miller [Fri, 2 Mar 2007 02:30:08 +0000 (18:30 -0800)]
[AOE]: Add get_unaligned() calls where needed.

Based upon a report by Andrew Walrond.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[RADEON]: Fix blanking return value.
David S. Miller [Fri, 2 Mar 2007 02:29:14 +0000 (18:29 -0800)]
[RADEON]: Fix blanking return value.

If you'll recall, over a year ago, I pointed out that the current
Radeon driver erroneously returns -EINVAL for valid blanking codes,
here is a link to that thread:

http://lkml.org/lkml/2006/1/28/6

No other driver does this, and it confuses the X server into thinking
that the device does not support blanking properly.

I looked again and there is simply no reason for the Radeon driver to
return -EINVAL for FB_BLANK_NORMAL.  It claims it wants to do this in
order to convince fbcon to blank in software, right here:

if (fb_blank(info, blank))
fbcon_generic_blank(vc, info, blank);

to software blank the screen.  But it only causes that to happen
in the FB_BLANK_NORMAL case.

That makes no sense because the Radeon code does this:

val |= CRTC_DISPLAY_DIS;

in the FB_BLANK_NORMAL case so should be blanking the hardware, and
there is therefore no reason to SW blank by returning -EINVAL.

Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Antonino Daplas <adaplas@gmail.com>
17 years ago[SPARC]: Provide 'get_property()' alias for of_get_property().
David S. Miller [Fri, 2 Mar 2007 02:10:25 +0000 (18:10 -0800)]
[SPARC]: Provide 'get_property()' alias for of_get_property().

Another powerpc compatibility item, this will allow us to share
more code with them.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[SPARC]: Provide pci_device_to_OF_node() just like powerpc.
David S. Miller [Fri, 2 Mar 2007 02:09:18 +0000 (18:09 -0800)]
[SPARC]: Provide pci_device_to_OF_node() just like powerpc.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[SPARC]: Handle unresolvable resources better in of_device.c
David S. Miller [Fri, 2 Mar 2007 02:08:05 +0000 (18:08 -0800)]
[SPARC]: Handle unresolvable resources better in of_device.c

Just leave them as zero if we couldn't calculate it.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[SPARC]: Fix bus handling in build_device_resources().
David S. Miller [Thu, 1 Mar 2007 07:20:12 +0000 (23:20 -0800)]
[SPARC]: Fix bus handling in build_device_resources().

We mistakedly modify 'bus' in the innermost loop.  What
should happen is that at each register index iteration,
we start with the same 'bus'.

So preserve it's value at the top level, and use a loop
local variable 'dbus' for iteration.

This bug causes registers other than the first to be
decoded improperly.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Fri, 2 Mar 2007 03:48:21 +0000 (19:48 -0800)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  ahci: improve spurious SDB FIS handling
  ahci/pata_jmicron: match class not function number
  jmicron ATA: reimplement jmicron ATA quirk
  pata_jmicron: drop unnecessary device programming in [re]init
  libata: blacklist FUJITSU MHT2060BH for NCQ
  sata_sil24: kill unused local variable idx in sil24_fill_sg()
  libata: clear drvdata in ata_host_release(), take#2

17 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Fri, 2 Mar 2007 01:30:51 +0000 (17:30 -0800)]
Merge branch 'for-linus' of /linux/kernel/git/jikos/hid

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jikos/hid:
  HID: fix Logitech DiNovo Edge touchwheel and Logic3 /SpectraVideo middle button
  HID: add git tree information to MAINTAINERS
  HID: fix broken Logitech S510 keyboard report descriptor; make extra keys work
  HID: fix possible double-free on error path in hid parser
  HID: hid-debug.c should #include <linux/hid-debug.h>
  HID: fix bug in zeroing the last field byte in output reports
  USB HID: use CONFIG_HID_DEBUG for outputting report descriptor
  USB HID: Fix USB vendor and product IDs endianness for USB HID devices

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Fri, 2 Mar 2007 01:28:31 +0000 (17:28 -0800)]
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Fix parport_pc build.
  [SPARC64]: Update defconfig.

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 2 Mar 2007 01:27:01 +0000 (17:27 -0800)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [TCP]: Fix minisock tcp_create_openreq_child() typo.
  [TCP]: Document several sysctls.
  [NET]: Fix kfree(skb)
  [NET]: Handle disabled preemption in gfp_any()
  [BRIDGE]: Fix locking of set path cost.
  [IPV6]: /proc/net/anycast6 unbalanced inet6_dev refcnt
  [IPX]: Remove ancient changelog
  [IPX]: Remove outdated information from Kconfig
  [NET]: Revert socket.h/stat.h ifdef hacks.
  [IPV6]: anycast refcnt fix
  [XFRM] xfrm_user: Fix return values of xfrm_add_sa_expire.

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb
Linus Torvalds [Fri, 2 Mar 2007 01:25:23 +0000 (17:25 -0800)]
Merge /pub/scm/linux/kernel/git/mchehab/v4l-dvb

* master.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb:
  V4L/DVB (5305): Mark VIDIOC_DBG_S/G_REGISTER as experimental
  V4L/DVB (5271): Add VIDIOC_TRY_ENCODER_CMD and VIDIOC_ENCODER_CMD ioctls.
  V4L/DVB (5270): Add VIDIOC_G_ENC_INDEX ioctl
  V4L/DVB (5276): Cxusb: fix firmware patch for big endian systems
  V4L/DVB (5258): Cafe_ccic: fix compiler warning
  V4L/DVB (5295): Digitv: open nxt6000 i2c_gate for TDED4 tuner handling
  V4L/DVB (5304): Improve chip matching in v4l2_register
  V4L/DVB (5255): Fix cx25840 firmware loading.

17 years ago[PATCH] Fix sysfs build breakage if !CONFIG_SYSFS
Ralf Baechle [Thu, 1 Mar 2007 12:40:21 +0000 (12:40 +0000)]
[PATCH] Fix sysfs build breakage if !CONFIG_SYSFS

B0rkage introduced by dfa87c824a9a5430008acd1ed2e8111ed164fcbe.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoahci: improve spurious SDB FIS handling
Tejun Heo [Tue, 27 Feb 2007 04:24:19 +0000 (13:24 +0900)]
ahci: improve spurious SDB FIS handling

Spurious SDB FIS during NCQ might not contain spurious completions.
It could be spurious TF update or invalid async notification.  Treat
as HSM violation iff a spurious SDB FIS contains spurious completions;
otherwise, just whine once about it.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoahci/pata_jmicron: match class not function number
Tejun Heo [Mon, 26 Feb 2007 11:24:03 +0000 (20:24 +0900)]
ahci/pata_jmicron: match class not function number

Make jmiron_ata quirk update pdev->class after programming the device
and update ahci and pata_jmicron such that they match class code
instead of checking function number manually.  For ahci, it matches
for vendor and class.  For pata_jmicron, it matches vendor, device and
class as IDE class isn't as well defined as AHCI class.

This makes jmicron device matching more conventional and script
friendly.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agojmicron ATA: reimplement jmicron ATA quirk
Tejun Heo [Mon, 26 Feb 2007 11:16:13 +0000 (20:16 +0900)]
jmicron ATA: reimplement jmicron ATA quirk

Reimplement jmicron ATA quirk.

* renamed to quirk_jmicron_ata()
* quirk is invoked only for the affected controllers
* programming is stricter.  e.g. conf5 bit24 is cleared if
  unnecessary.
* code factored for readability
* JMB360 and JMB368 are programmed into proper mode

Verified on JMB360, 363 and 368.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agopata_jmicron: drop unnecessary device programming in [re]init
Tejun Heo [Mon, 26 Feb 2007 11:09:02 +0000 (20:09 +0900)]
pata_jmicron: drop unnecessary device programming in [re]init

Channel redirect and AHCI mode enable programmings are done via PCI
quirk for both probe and resume paths.  Drop duplicate and possibly
unsafe device programming from pata_jmicron().

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years ago[PATCH] Fix get_unmapped_area and fsync for hugetlb shm segments
Adam Litke [Thu, 1 Mar 2007 23:46:08 +0000 (15:46 -0800)]
[PATCH] Fix get_unmapped_area and fsync for hugetlb shm segments

This patch provides the following hugetlb-related fixes to the recent stacked
shm files changes:
 - Update is_file_hugepages() so it will reconize hugetlb shm segments.
 - get_unmapped_area must be called with the nested file struct to handle
   the sfd->file->f_ops->get_unmapped_area == NULL case.
 - The fsync f_op must be wrapped since it is specified in the hugetlbfs
   f_ops.

This is based on proposed fixes from Eric Biederman that were debugged and
tested by me.  Without it, attempting to use hugetlb shared memory segments
on powerpc (and likely ia64) will kill your box.

Signed-off-by: Adam Litke <agl@us.ibm.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Acked-by: William Irwin <bill.irwin@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agolibata: blacklist FUJITSU MHT2060BH for NCQ
Tejun Heo [Wed, 28 Feb 2007 06:21:23 +0000 (15:21 +0900)]
libata: blacklist FUJITSU MHT2060BH for NCQ

Blacklist FUJITSU MHT2060BH for NCQ.  On this drive, NCQ works iff
queue depth is equal to or less than 4.  Just turn it off.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Mike Accetta <maccetta@laurelnetworks.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agosata_sil24: kill unused local variable idx in sil24_fill_sg()
Tejun Heo [Mon, 26 Feb 2007 07:11:00 +0000 (16:11 +0900)]
sata_sil24: kill unused local variable idx in sil24_fill_sg()

Kill unused local variable idx in sil24_fill_sg().

Spotted by Jeff Garzik.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agolibata: clear drvdata in ata_host_release(), take#2
Tejun Heo [Tue, 27 Feb 2007 13:33:21 +0000 (22:33 +0900)]
libata: clear drvdata in ata_host_release(), take#2

Clearing drvdata in ->remove_one causes NULL pointer deference.  Clear
drvdata only in ata_host_release() after all resources are freed.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years ago[PATCH] VM: invalidate_inode_pages2_range() should not exit early
Trond Myklebust [Thu, 1 Mar 2007 04:13:55 +0000 (20:13 -0800)]
[PATCH] VM: invalidate_inode_pages2_range() should not exit early

Fix invalidate_inode_pages2_range() so that it does not immediately exit
just because a single page in the specified range could not be removed.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] tty_io: fix race in master pty close/slave pty close path
Aristeu Sergio Rozanski Filho [Thu, 1 Mar 2007 04:13:53 +0000 (20:13 -0800)]
[PATCH] tty_io: fix race in master pty close/slave pty close path

This patch fixes a possible race that leads to double freeing an idr index.
 When the master begin to close, release_dev() is called and then
pty_close() is called:

        if (tty->driver->close)
                tty->driver->close(tty, filp);

This is done without helding any locks other than BKL.  Inside pty_close(),
being a master close, the devpts entry will be removed:

#ifdef CONFIG_UNIX98_PTYS
                if (tty->driver == ptm_driver)
                        devpts_pty_kill(tty->index);
#endif

But devpts_pty_kill() will call get_node() that may sleep while waiting for
&devpts_root->d_inode->i_sem.  When this happens and the slave is being
opened, tty_open() just found the driver and index:

        driver = get_tty_driver(device, &index);
        if (!driver) {
                mutex_unlock(&tty_mutex);
                return -ENODEV;
        }

This part of the code is already protected under tty_mute.  The problem is
that the slave close already got an index.  Then init_dev() is called and
blocks waiting for the same &devpts_root->d_inode->i_sem.

When the master close resumes, it removes the devpts entry, and the
relation between idr index and the tty is gone.  The master then sleeps
waiting for the tty_mutex on release_dev().

Slave open resumes and found no tty for that index.  As result, a NULL tty
is returned and init_dev() doesn't flow to fast_track:

        /* check whether we're reopening an existing tty */
        if (driver->flags & TTY_DRIVER_DEVPTS_MEM) {
                tty = devpts_get_tty(idx);
                if (tty && driver->subtype == PTY_TYPE_MASTER)
                        tty = tty->link;
        } else {
                tty = driver->ttys[idx];
        }
        if (tty) goto fast_track;

The result of this, is that a new tty will be created and init_dev() returns
sucessfull. After returning, tty_mutex is dropped and master close may resume.

Master close finds it's the only use and both sides are closing, then releases
the tty and the index. At this point, the idr index is free, but slave still
has it.

Slave open then calls pty_open() and finds that tty->link->count is 0,
because there's no master and returns error.  Then tty_open() calls
release_dev() which executes without any warning, as it was a case of last
slave close when the master is already closed (master->count == 0,
slave->count == 1).  The tty is then released with the already released idr
index.

This normally would only issue a warning on idr_remove() but in case of a
customer's critical application, it's never too simple:

thread1: opens master, gets index X
thread1: begin closing master
thread2: begin opening slave with index X
thread1: finishes closing master, index X released
thread3: opens master, gets index X, just released
thread2: fails opening slave, releases index X         <----
thread4: opens master, gets index X, init_dev() then find an already in use
 and healthy tty and fails

If no more indexes are released, ptmx_open() will keep failing, as the
first free index available is X, and it will make init_dev() fail because
you're trying to "reopen a master" which isn't valid.

The patch notices when this race happens and make init_dev() fail
imediately.  The init_dev() function is called with tty_mutex held, so it's
safe to continue with tty till the end of function because release_dev()
won't make any further changes without grabbing the tty_mutex.

Without the patch, on some machines it's possible get easily idr warnings
like this one:

idr_remove called for id=15 which is not allocated.
 [<c02555b9>] idr_remove+0x139/0x170
 [<c02a1b62>] release_mem+0x182/0x230
 [<c02a28e7>] release_dev+0x4b7/0x700
 [<c02a0ea7>] tty_ldisc_enable+0x27/0x30
 [<c02a1e64>] init_dev+0x254/0x580
 [<c02a0d64>] check_tty_count+0x14/0xb0
 [<c02a4f05>] tty_open+0x1c5/0x340
 [<c02a4d40>] tty_open+0x0/0x340
 [<c017388f>] chrdev_open+0xaf/0x180
 [<c017c2ac>] open_namei+0x8c/0x760
 [<c01737e0>] chrdev_open+0x0/0x180
 [<c0167bc9>] __dentry_open+0xc9/0x210
 [<c0167e2c>] do_filp_open+0x5c/0x70
 [<c0167a91>] get_unused_fd+0x61/0xd0
 [<c0167e93>] do_sys_open+0x53/0x100
 [<c0167f97>] sys_open+0x27/0x30
 [<c010303b>] syscall_call+0x7/0xb

using this test application available on:
 http://www.ruivo.org/~aris/pty_sodomizer.c

Signed-off-by: Aristeu Sergio Rozanski Filho <aris@ruivo.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Chuck Ebbert <cebbert@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] fix memory leak in dma_declare_coherent_memory()
Yoichi Yuasa [Thu, 1 Mar 2007 04:13:51 +0000 (20:13 -0800)]
[PATCH] fix memory leak in dma_declare_coherent_memory()

When it goes to free1_out, dev->dma_mem has not been freed.

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Fix buffer overflow and races in capi debug functions
Karsten Keil [Thu, 1 Mar 2007 04:13:50 +0000 (20:13 -0800)]
[PATCH] Fix buffer overflow and races in capi debug functions

The CAPI trace debug functions were using a fixed size buffer, which can be
overflowed if wrong formatted CAPI messages were sent to the kernel capi
layer.  The code was also not protected against multiple callers.  This fix
bug 8028.

Additionally the patch make the CAPI trace functions optional.

Signed-off-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] adapt page_lock_anon_vma() to PREEMPT_RCU
Oleg Nesterov [Thu, 1 Mar 2007 04:13:49 +0000 (20:13 -0800)]
[PATCH] adapt page_lock_anon_vma() to PREEMPT_RCU

page_lock_anon_vma() uses spin_lock() to block RCU.  This doesn't work with
PREEMPT_RCU, we have to do rcu_read_lock() explicitely.  Otherwise, it is
theoretically possible that slab returns anon_vma's memory to the system
before we do spin_unlock(&anon_vma->lock).

[ Hugh points out that this only matters for PREEMPT_RCU, which isn't merged
  yet, and may never be.  Regardless, this patch is conceptually the
  right thing to do, even if it doesn't matter at this point.  - Linus ]

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Documentation: CPU load calculation description
Vassili Karpov [Thu, 1 Mar 2007 04:13:45 +0000 (20:13 -0800)]
[PATCH] Documentation: CPU load calculation description

Describes how/when the information exported to `/proc/stat' is calculated,
and possible problems with this approach.

Signed-off-by: Vassili Karpov <av1474@comtv.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] sched: fix SMT scheduler bug
Ingo Molnar [Thu, 1 Mar 2007 04:13:42 +0000 (20:13 -0800)]
[PATCH] sched: fix SMT scheduler bug

The SMT scheduler incorrectly skips kernel threads even if they are
runnable (but they are preempted by a higher-prio user-space task which got
SMT-delayed by an even higher-priority task running on a sibling CPU).

Fix this for now by only doing the SMT-nice optimization if the
to-be-delayed task is the only runnable task.  (This should cover most of
the real-life cases anyway.)

This bug has been in the SMT scheduler since 2.6.17 or so, but has only
been noticed now by the active check in the dynticks code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Michal Piotrowski <michal.k.k.piotrowski@gmail.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ps3: introduce CONFIG_PS3_ADVANCED
Geert Uytterhoeven [Thu, 1 Mar 2007 04:13:38 +0000 (20:13 -0800)]
[PATCH] ps3: introduce CONFIG_PS3_ADVANCED

ps3: Introduce CONFIG_PS3_ADVANCED, as suggested by Roman Zippel, and use
it to control questions about PS3 subsystems that may not be obvious for
the casual user.

This gets rid of the following warning on non-powerpc platforms: |
drivers/video/Kconfig:1604:warning: 'select' used by config symbol 'FB_PS3'
refer to undefined symbol 'PS3_PS3AV'

Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Acked-by: Geoff Levand <geoffrey.levand@am.sony.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ext[34]: EA block reference count racing fix
Mingming Cao [Thu, 1 Mar 2007 04:13:35 +0000 (20:13 -0800)]
[PATCH] ext[34]: EA block reference count racing fix

There are race issues around ext[34] xattr block release code.

ext[34]_xattr_release_block() checks the reference count of xattr block
(h_refcount) and frees that xattr block if it is the last one reference it.
 Unlike ext2, the check of this counter is unprotected by any lock.
ext[34]_xattr_release_block() will free the mb_cache entry before freeing
that xattr block.  There is a small window between the check for the re
h_refcount ==1 and the call to mb_cache_entry_free().  During this small
window another inode might find this xattr block from the mbcache and reuse
it, racing a refcount updates.  The xattr block will later be freed by the
first inode without notice other inode is still use it.  Later if that
block is reallocated as a datablock for other file, then more serious
problem might happen.

We need put a lock around places checking the refount as well to avoid
racing issue.  Another place need this kind of protection is in
ext3_xattr_block_set(), where it will modify the xattr block content in-
the-fly if the refcount is 1 (means it's the only inode reference it).

This will also fix another issue: the xattr block may not get freed at all
if no lock is to protect the refcount check at the release time.  It is
possible that the last two inodes could release the shared xattr block at
the same time.  But both of them think they are not the last one so only
decreased the h_refcount without freeing xattr block at all.

We need to call lock_buffer() after ext3_journal_get_write_access() to
avoid deadlock (because the later will call lock_buffer()/unlock_buffer
() as well).

Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Cc: Andreas Gruenbacher <agruen@suse.de>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: pte_mkread fix
Jeff Dike [Thu, 1 Mar 2007 04:13:33 +0000 (20:13 -0800)]
[PATCH] uml: pte_mkread fix

Fix the fact that pte_mkread set _PAGE_RW instead of _PAGE_USER (the logic is
copied from i386 in most place, so it is really as bad as you're thinking).

Thus currently page tables are more permissive than they should.

Such a change may trigger other latent bugs, so be careful with this.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: host VDSO fix
Jeff Dike [Thu, 1 Mar 2007 04:13:31 +0000 (20:13 -0800)]
[PATCH] uml: host VDSO fix

This fixes a problem seen by a number of people running UML on newer host
kernels.  init would hang with an infinite segfault loop.

It turns out that the host kernel was providing a AT_SYSINFO_EHDR of
0xffffe000, which faked UML into believing that the host VDSO page could be
reused.  However, AT_SYSINFO pointed into the middle of the address space, and
was unmapped as a result.  Because UML was providing AT_SYSINFO_EHDR and
AT_SYSINFO to its own processes, these would branch to nowhere when trying to
use the VDSO.

The fix is to also check the location of AT_SYSINFO when deciding whether to
use the host's VDSO.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: enable RAW
Allan Graves [Thu, 1 Mar 2007 04:13:29 +0000 (20:13 -0800)]
[PATCH] uml: enable RAW

Add the RAW device driver options to the UML Kconfig.char file so that you may
use them in UML.

Signed-off-by: Allan Graves<allan.graves@gmail.com>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] FRV: Missing error defs
David Howells [Thu, 1 Mar 2007 04:13:26 +0000 (20:13 -0800)]
[PATCH] FRV: Missing error defs

linux/irq.h uses EINVAL but does not #include linux/errno.h.  This results in
the compiler spitting out errors on some files.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] throttle_vm_writeout(): don't loop on GFP_NOFS and GFP_NOIO allocations
Andrew Morton [Thu, 1 Mar 2007 04:13:21 +0000 (20:13 -0800)]
[PATCH] throttle_vm_writeout(): don't loop on GFP_NOFS and GFP_NOIO allocations

throttle_vm_writeout() is designed to wait for the dirty levels to subside.
But if the caller holds IO or FS locks, we might be holding up that writeout.

So change it to take a single nap to give other devices a chance to clean some
memory, then return.

Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Kumar Gala <galak@kernel.crashing.org>
Cc: Pete Zaitcev <zaitcev@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ACPI: make bay depend on dock
Kristen Carlson Accardi [Thu, 1 Mar 2007 04:13:17 +0000 (20:13 -0800)]
[PATCH] ACPI: make bay depend on dock

Since the bay driver depends on the dock driver for proper notification,
make this driver depend on the dock driver.

Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Acked-by: Len Brown <lenb@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Bug in MM_RB debugging
David Miller [Thu, 1 Mar 2007 04:13:13 +0000 (20:13 -0800)]
[PATCH] Bug in MM_RB debugging

The code is seemingly trying to make sure that rb_next() brings us to
successive increasing vma entries.

But the two variables, prev and pend, used to perform these checks, are
never advanced.

Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Andrea Arcangeli <andrea@novell.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: add back accidentally removed error
Jeff Dike [Thu, 1 Mar 2007 04:13:11 +0000 (20:13 -0800)]
[PATCH] uml: add back accidentally removed error

In the 2.6.20 hang patch, I accidentally threw out an error message.
This puts it back.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: fix host LDT lookup initialization locking, try 2
Jeff Dike [Thu, 1 Mar 2007 04:13:06 +0000 (20:13 -0800)]
[PATCH] uml: fix host LDT lookup initialization locking, try 2

Add some locking to host_ldt_entries to prevent racing when reading LDT
information from the host.

The locking is somewhat more careful than my previous attempt.  Now, only
the check of host_ldt_entries is locked.  The lock is dropped immediately
afterwards, and if the LDT needs initializing, that (and the memory
allocations needed) proceed outside the lock.

Also fixed some style violations.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] drivers/isdn/gigaset: build asyncdata.o into the gigaset module (fix)
Tilman Schmidt [Thu, 1 Mar 2007 04:13:01 +0000 (20:13 -0800)]
[PATCH] drivers/isdn/gigaset: build asyncdata.o into the gigaset module (fix)

a) Remove #define acrobatics that have become unnecessary by the move of
   asyncdata.o into the common part.

b) Correct the rule for building the common part into the kernel when
   some or all hardware specific parts are built as modules.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Cc: Adrian Bunk <bunk@stusta.de>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] eCryptfs: no path_release() after path_lookup() error
Michael Halcrow [Thu, 1 Mar 2007 04:12:57 +0000 (20:12 -0800)]
[PATCH] eCryptfs: no path_release() after path_lookup() error

Dmitriy Monakhov wrote:
> if path_lookup() return non zero code we don't have to worry about
> 'nd' parameter, but ecryptfs_read_super does path_release(&nd) after
> path_lookup has failed, and dentry counter becomes negative

Do not do a path_release after a path_lookup error.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: Dmitriy Monakhov <dmonakhov@openvz.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] eCryptfs: remove unnecessary flush_dcache_page()
Michael Halcrow [Thu, 1 Mar 2007 04:12:52 +0000 (20:12 -0800)]
[PATCH] eCryptfs: remove unnecessary flush_dcache_page()

Remove unnecessary flush_dcache_page() call. Thanks to Dmitriy
Monakhov for pointing this out.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: Dmitriy Monakhov <dmonakhov@openvz.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] eCryptfs: set O_LARGEFILE when opening lower file
Michael Halcrow [Thu, 1 Mar 2007 04:12:47 +0000 (20:12 -0800)]
[PATCH] eCryptfs: set O_LARGEFILE when opening lower file

O_LARGEFILE should be set here when opening the lower file.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: Dmitriy Monakhov <dmonakhov@openvz.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] rtc_cmos oops fix
David Brownell [Thu, 1 Mar 2007 04:12:40 +0000 (20:12 -0800)]
[PATCH] rtc_cmos oops fix

Fix an oops on the rtc_device_unregister() path by waiting until the last
moment before nulling the rtc->ops vector.  Fix some potential oopses by
having the rtc_class_open()/rtc_class_close() interface increase the RTC's
reference count while an RTC handle is available outside the RTC framework.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] MAINTAINERS: Update email address
Antonino A. Daplas [Thu, 1 Mar 2007 04:12:37 +0000 (20:12 -0800)]
[PATCH] MAINTAINERS: Update email address

adaplas@pol.net is still alive, but is choking on the traffic.

Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] add -mm testing in SubmitChecklist
Randy Dunlap [Thu, 1 Mar 2007 04:12:35 +0000 (20:12 -0800)]
[PATCH] add -mm testing in SubmitChecklist

Add -mm testing to SubmitChecklist.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] fix section mismatch warning in lockdep
Sam Ravnborg [Thu, 1 Mar 2007 04:12:31 +0000 (20:12 -0800)]
[PATCH] fix section mismatch warning in lockdep

lockdep_init() is marked __init but used in several places
outside __init code. This causes following warnings:
$ scripts/mod/modpost kernel/lockdep.o
WARNING: kernel/built-in.o - Section mismatch: reference to .init.text:lockdep_init from .text.lockdep_init_map after 'lockdep_init_map' (at offset 0x105)
WARNING: kernel/built-in.o - Section mismatch: reference to .init.text:lockdep_init from .text.lockdep_reset_lock after 'lockdep_reset_lock' (at offset 0x35)
WARNING: kernel/built-in.o - Section mismatch: reference to .init.text:lockdep_init from .text.__lock_acquire after '__lock_acquire' (at offset 0xb2)

The warnings are less obviously due to heavy inlining by gcc - this is not
altered.

Fix the section mismatch warnings by removing the __init marking, which
seems obviously wrong.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Rename PG_checked to PG_owner_priv_1
Nick Piggin [Thu, 1 Mar 2007 04:12:27 +0000 (20:12 -0800)]
[PATCH] Rename PG_checked to PG_owner_priv_1

Rename PG_checked to PG_owner_priv_1 to reflect its availablilty as a
private flag for use by the owner/allocator of the page.  In the case of
pagecache pages (which might be considered to be owned by the mm),
filesystems may use the flag.

Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] fix the SYSCTL=n compilation
Adrian Bunk [Thu, 1 Mar 2007 04:12:23 +0000 (20:12 -0800)]
[PATCH] fix the SYSCTL=n compilation

/home/bunk/linux/kernel-2.6/linux-2.6.20-mm2/kernel/sysctl.c:1411: error: conflicting types for 'register_sysctl_table'
/home/bunk/linux/kernel-2.6/linux-2.6.20-mm2/include/linux/sysctl.h:1042: error: previous declaration of 'register_sysctl_table' was here
make[2]: *** [kernel/sysctl.o] Error 1

Caused by commit 0b4d414714f0d2f922d39424b0c5c82ad900a381.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Fix posix-cpu-timer breakage caused by stale p->last_ran value
Thomas Gleixner [Thu, 1 Mar 2007 04:12:19 +0000 (20:12 -0800)]
[PATCH] Fix posix-cpu-timer breakage caused by stale p->last_ran value

Problem description at:
http://bugzilla.kernel.org/show_bug.cgi?id=8048

Commit b18ec80396834497933d77b81ec0918519f4e2a7
    [PATCH] sched: improve migration accuracy
optimized the scheduler time calculations, but broke posix-cpu-timers.

The problem is that the p->last_ran value is not updated after a context
switch.  So a subsequent call to current_sched_time() calculates with a
stale p->last_ran value, i.e.  accounts the full time, which the task was
scheduled away.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] eCryptfs: resolve lower page unlocking problem
Michael Halcrow [Thu, 1 Mar 2007 04:12:16 +0000 (20:12 -0800)]
[PATCH] eCryptfs: resolve lower page unlocking problem

eCryptfs lower file handling code has several issues:
  - Retval from prepare_write()/commit_write() wasn't checked to equality
    to AOP_TRUNCATED_PAGE.
  - In some places page wasn't unmapped and unlocked after error.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] kernel-doc fixes for 2.6.20-git15 (non-drivers)
Randy Dunlap [Thu, 1 Mar 2007 04:12:13 +0000 (20:12 -0800)]
[PATCH] kernel-doc fixes for 2.6.20-git15 (non-drivers)

Fix kernel-doc warnings in 2.6.20-git15 (lib/, mm/, kernel/, include/).

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] kernel-doc: allow space after __attribute__
Randy Dunlap [Thu, 1 Mar 2007 04:12:10 +0000 (20:12 -0800)]
[PATCH] kernel-doc: allow space after __attribute__

Allow space(s) between "__attribute__" and "((blah))" so that
kernel-doc does not complain like:

Warning(/tester/linsrc/linux-2.6.20-git15//kernel/timer.c:939): No description found for parameter 'read_persistent_clock(void'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] update timekeeping_is_continuous comment
Daniel Walker [Thu, 1 Mar 2007 04:12:07 +0000 (20:12 -0800)]
[PATCH] update timekeeping_is_continuous comment

Signed-off-by: Daniel Walker <dwalker@mvista.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] FRV: Add some missng lazy MMU hooks for NOMMU mode
David Howells [Thu, 1 Mar 2007 04:12:03 +0000 (20:12 -0800)]
[PATCH] FRV: Add some missng lazy MMU hooks for NOMMU mode

Add some missing lazy MMU hooks for NOMMU mode.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] FRV: No ZONE_DMA
David Howells [Thu, 1 Mar 2007 04:11:58 +0000 (20:11 -0800)]
[PATCH] FRV: No ZONE_DMA

FRV does not require a ZONE_DMA, so all DMA'able pages that aren't highmem
should be in ZONE_NORMAL.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: add support for reshape of a raid6
NeilBrown [Thu, 1 Mar 2007 04:11:53 +0000 (20:11 -0800)]
[PATCH] md: add support for reshape of a raid6

i.e. one or more drives can be added and the array will re-stripe
while on-line.

Most of the interesting work was already done for raid5.  This just extends it
to raid6.

mdadm newer than 2.6 is needed for complete safety, however any version of
mdadm which support raid5 reshape will do a good enough job in almost all
cases (an 'echo repair > /sys/block/mdX/md/sync_action' is recommended after a
reshape that was aborted and had to be restarted with an such a version of
mdadm).

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: restart a (raid5) reshape that has been aborted due to a read/write error
NeilBrown [Thu, 1 Mar 2007 04:11:48 +0000 (20:11 -0800)]
[PATCH] md: restart a (raid5) reshape that has been aborted due to a read/write error

An error always aborts any resync/recovery/reshape on the understanding that
it will immediately be restarted if that still makes sense.  However a reshape
currently doesn't get restarted.  With this patch it does.

To avoid restarting when it is not possible to do work, we call into the
personality to check that a reshape is ok, and strengthen raid5_check_reshape
to fail if there are too many failed devices.

We also break some code out into a separate function: remove_and_add_spares as
the indent level for that code was getting crazy.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: clean out unplug and other queue function on md shutdown
NeilBrown [Thu, 1 Mar 2007 04:11:42 +0000 (20:11 -0800)]
[PATCH] md: clean out unplug and other queue function on md shutdown

The mddev and queue might be used for another array which does not set these,
so they need to be cleared.

Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: move warning about creating a raid array on partitions of the one device
NeilBrown [Thu, 1 Mar 2007 04:11:35 +0000 (20:11 -0800)]
[PATCH] md: move warning about creating a raid array on partitions of the one device

md tries to warn the user if they e.g.  create a raid1 using two partitions of
the same device, as this does not provide true redundancy.

However it also warns if a raid0 is created like this, and there is nothing
wrong with that.

At the place where the warning is currently printer, we don't necessarily know
what level the array will be, so move the warning from the point where the
device is added to the point where the array is started.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: RAID6: clean up CPUID and FPU enter/exit code
H. Peter Anvin [Thu, 1 Mar 2007 04:11:25 +0000 (20:11 -0800)]
[PATCH] md: RAID6: clean up CPUID and FPU enter/exit code

- Use kernel_fpu_begin() and kernel_fpu_end()
- Use boot_cpu_has() for feature testing even in userspace

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: fix raid10 recovery problem.
NeilBrown [Thu, 1 Mar 2007 04:11:18 +0000 (20:11 -0800)]
[PATCH] md: fix raid10 recovery problem.

There are two errors that can lead to recovery problems with raid10
when used in 'far' more (not the default).

Due to a '>' instead of '>=' the wrong block is located which would result in
garbage being written to some random location, quite possible outside the
range of the device, causing the newly reconstructed device to fail.

The device size calculation had some rounding errors (it didn't round when it
should) and so recovery would go a few blocks too far which would again cause
a write to a random block address and probably a device error.

The code for working with device sizes was fairly confused and spread out, so
this has been tided up a bit.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Fix failure paths in modules init in umem.c
NeilBrown [Thu, 1 Mar 2007 04:11:12 +0000 (20:11 -0800)]
[PATCH] Fix failure paths in modules init in umem.c

If register_blkdev() or alloc-disk fail in mm_init() after
pci_register_driver() succeeds, then mm_pci_driver is not unregistered
properly:

Cc: Philip Guo <pg@cs.stanford.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] drivers/video/sm501fb.c: make 4 functions static
Adrian Bunk [Thu, 1 Mar 2007 04:11:06 +0000 (20:11 -0800)]
[PATCH] drivers/video/sm501fb.c: make 4 functions static

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] mm/{,tiny-}shmem.c cleanups
Adrian Bunk [Thu, 1 Mar 2007 04:11:03 +0000 (20:11 -0800)]
[PATCH] mm/{,tiny-}shmem.c cleanups

shmem_{nopage,mmap} are no longer used in ipc/shm.c

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: "Eric W. Biederman" <ebiederm@xmission.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>
17 years ago[PATCH] make ipc/shm.c:shm_nopage() static
Adrian Bunk [Thu, 1 Mar 2007 04:11:01 +0000 (20:11 -0800)]
[PATCH] make ipc/shm.c:shm_nopage() static

shm_nopage() can become static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] arch/arm26/kernel/entry.S: remove dead code
Adrian Bunk [Thu, 1 Mar 2007 04:10:58 +0000 (20:10 -0800)]
[PATCH] arch/arm26/kernel/entry.S: remove dead code

CONFIG_ALIGNMENT_TRAP is never set on arm26.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Ian Molton <spyro@f2s.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoV4L/DVB (5305): Mark VIDIOC_DBG_S/G_REGISTER as experimental
Hans Verkuil [Sat, 24 Feb 2007 00:16:01 +0000 (21:16 -0300)]
V4L/DVB (5305): Mark VIDIOC_DBG_S/G_REGISTER as experimental

Move VIDIOC_DBG_S/G_REGISTER from the internal ioctl list to the
public ioctls, but mark it as experimental for now.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5271): Add VIDIOC_TRY_ENCODER_CMD and VIDIOC_ENCODER_CMD ioctls.
Hans Verkuil [Sun, 18 Feb 2007 17:56:22 +0000 (14:56 -0300)]
V4L/DVB (5271): Add VIDIOC_TRY_ENCODER_CMD and VIDIOC_ENCODER_CMD ioctls.

Add support for starting, stopping, pausing and resuming an MPEG (or similar
compressed stream) encoder.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5270): Add VIDIOC_G_ENC_INDEX ioctl
Hans Verkuil [Sun, 18 Feb 2007 17:05:02 +0000 (14:05 -0300)]
V4L/DVB (5270): Add VIDIOC_G_ENC_INDEX ioctl

The VIDIOC_G_ENC_INDEX ioctl can obtain the MPEG index from an MPEG
encoder.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5276): Cxusb: fix firmware patch for big endian systems
Jin-Bong lee [Wed, 21 Feb 2007 02:10:34 +0000 (23:10 -0300)]
V4L/DVB (5276): Cxusb: fix firmware patch for big endian systems

Without this patch, the device will not be detected after firmware download
on big endian systems.

Signed-off-by: Jin-Bong lee <jinbong.lee@samsung.com>
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5258): Cafe_ccic: fix compiler warning
Jean Delvare [Tue, 13 Feb 2007 22:31:45 +0000 (19:31 -0300)]
V4L/DVB (5258): Cafe_ccic: fix compiler warning

Fix the following warning:
drivers/media/video/cafe_ccic.c: In function `cafe_vidioc_reqbufs':
drivers/media/video/cafe_ccic.c:1197: warning: 'ret' might be used uninitialized in this function
Probably not a real bug, but the warning can be avoided easily.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5295): Digitv: open nxt6000 i2c_gate for TDED4 tuner handling
Michael Krufky [Thu, 22 Feb 2007 00:47:15 +0000 (21:47 -0300)]
V4L/DVB (5295): Digitv: open nxt6000 i2c_gate for TDED4 tuner handling

dvb-pll normally opens the i2c gate before attempting to communicate with
the pll, but the code for this device is not using dvb-pll.  This should
be cleaned up in the future, but for now, just open the i2c gate at the
appropriate place in order to fix this driver bug.

Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5304): Improve chip matching in v4l2_register
Hans Verkuil [Fri, 23 Feb 2007 23:55:14 +0000 (20:55 -0300)]
V4L/DVB (5304): Improve chip matching in v4l2_register

The chip matching in struct v4l2_register for VIDIOC_DBG_G/S_REGISTER
was rather primitive. It could not be extended to other busses besides
i2c and it lacked a way to.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5255): Fix cx25840 firmware loading.
Hans Verkuil [Thu, 15 Feb 2007 06:40:34 +0000 (03:40 -0300)]
V4L/DVB (5255): Fix cx25840 firmware loading.

Due to changes in the i2c handling in 2.6.20 this cx25840 bug surfaced,
causing the firmware load to fail for the ivtv driver. The correct
sequence is to first attach the i2c client, then use the client's
device to load the firmware.
Acked-by: Mike Isely <isely@pobox.com>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoHID: fix Logitech DiNovo Edge touchwheel and Logic3 /SpectraVideo middle button
Jiri Kosina [Thu, 1 Mar 2007 08:54:44 +0000 (09:54 +0100)]
HID: fix Logitech DiNovo Edge touchwheel and Logic3 /SpectraVideo middle button

Dongle shipped with Logitech DiNovo Edge (0x046d/0xc714) behaves in a weird
non-standard way - it contains multiple reports with the same usage, which
results in remapping of GenericDesktop.X and GenericDesktop.Y usages to
GenericDesktop.Z and GenericDesktop.RX respectively, thus rendering the
touchwheel unusable.

The commit 35068976916fdef82d6e69ef1f8c9a1c47732759 solved this
in a way that it didn't remap certain usages. This however breaks
(at least) middle button of Logic3 / SpectraVideo (0x1267/0x0210),
which in contrary requires the remapping.

To make both of the harware work, allow remapping of these usages again,
and introduce a quirk for Logitech DiNovo Edge "touchwheel" instead - we
disable remapping for key, abs and rel events only for this hardware.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoHID: add git tree information to MAINTAINERS
Jiri Kosina [Tue, 27 Feb 2007 16:40:09 +0000 (17:40 +0100)]
HID: add git tree information to MAINTAINERS

Update MAITAINERS entry for HID and USB HID, adding
location of HID git tree.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoHID: fix broken Logitech S510 keyboard report descriptor; make extra keys work
Jiri Kosina [Wed, 21 Feb 2007 18:27:49 +0000 (19:27 +0100)]
HID: fix broken Logitech S510 keyboard report descriptor; make extra keys work

This patch makes extra keys (F1-F12 in special mode, zooming, rotate, shuffle)
on Logitech S510 keyboard work.

Logitech S510 keyboard sends in report no. 3 keys which are far above the
logical maximum described in descriptor for given report.

This patch introduces a HID quirk for this wireless USB receiver/keyboard
in order to fix the report descriptor before it's being parsed - the logical
maximum and the number of usages is bumped up to 0x104d). The values are in the
"Reserved" area of consumer HUT, so HID_MAX_USAGE had to be changed too.

In addition to proper extracting of  the values from report descriptor, proper
HID-input mapping is introduced for them.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoHID: fix possible double-free on error path in hid parser
Jiri Kosina [Wed, 21 Feb 2007 16:18:03 +0000 (17:18 +0100)]
HID: fix possible double-free on error path in hid parser

Freeing of device->collection is properly done in hid_free_device() (as
this function is supposed to free all the device resources and could be
called from transport specific code, e.g. usb_hid_configure()).

Remove all kfree() calls preceeding the hid_free_device() call.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoHID: hid-debug.c should #include <linux/hid-debug.h>
Adrian Bunk [Wed, 21 Feb 2007 11:57:30 +0000 (12:57 +0100)]
HID: hid-debug.c should #include <linux/hid-debug.h>

Every file should include the headers containing the prototypes for
it's global functions.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoHID: fix bug in zeroing the last field byte in output reports
Jiri Kosina [Tue, 20 Feb 2007 00:33:39 +0000 (01:33 +0100)]
HID: fix bug in zeroing the last field byte in output reports

d4ae650a904612ffb7edd3f28b69b022988d2466 introduced zeroing of the
last field byte in output reports in order to make sure the unused
bits are set to 0. This is done in a wrong way, resulting in a
wrong bits being zeroed out (not properly shifted by the field offset
in the report). Fix this.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoUSB HID: use CONFIG_HID_DEBUG for outputting report descriptor
Jiri Kosina [Mon, 19 Feb 2007 13:15:59 +0000 (14:15 +0100)]
USB HID: use CONFIG_HID_DEBUG for outputting report descriptor

Report descriptor should be output when CONFIG_HID_DEBUG is defined.

This also mitigates the need for DEBUG and DEBUG_DATA defines, so let's
remove them.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years agoUSB HID: Fix USB vendor and product IDs endianness for USB HID devices
Julien BLACHE [Sun, 11 Feb 2007 17:20:25 +0000 (18:20 +0100)]
USB HID: Fix USB vendor and product IDs endianness for USB HID devices

The USB vendor and product IDs are not byteswapped appropriately, and
thus come out in the wrong endianness when fetched through the evdev
using ioctl() on big endian platforms.

Signed-off-by: Julien BLACHE <jb@jblache.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
17 years ago[SPARC64]: Fix parport_pc build.
David S. Miller [Wed, 28 Feb 2007 21:09:34 +0000 (13:09 -0800)]
[SPARC64]: Fix parport_pc build.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[TCP]: Fix minisock tcp_create_openreq_child() typo.
Arnaldo Carvalho de Melo [Wed, 28 Feb 2007 19:05:56 +0000 (11:05 -0800)]
[TCP]: Fix minisock tcp_create_openreq_child() typo.

On 2/28/07, KOVACS Krisztian <hidden@balabit.hu> wrote:
>
>   Hi,
>
>   While reading TCP minisock code I've found this suspiciously looking
> code fragment:
>
> - 8< -
> struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req, struct sk_buff *skb)
> {
>         struct sock *newsk = inet_csk_clone(sk, req, GFP_ATOMIC);
>
>         if (newsk != NULL) {
>                 const struct inet_request_sock *ireq = inet_rsk(req);
>                 struct tcp_request_sock *treq = tcp_rsk(req);
>                 struct inet_connection_sock *newicsk = inet_csk(sk);
>                 struct tcp_sock *newtp;
> - 8< -
>
>   The above code initializes newicsk to inet_csk(sk), isn't that supposed
> to be inet_csk(newsk)?  As far as I can tell this might leave
> icsk_ack.last_seg_size zero even if we do have received data.

Good catch!

David, please apply the attached patch.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[SPARC64]: Update defconfig.
David S. Miller [Wed, 28 Feb 2007 17:51:15 +0000 (09:51 -0800)]
[SPARC64]: Update defconfig.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[TCP]: Document several sysctls.
John Heffner [Tue, 27 Feb 2007 18:03:56 +0000 (10:03 -0800)]
[TCP]: Document several sysctls.

This adds documentation for tcp_moderate_rcvbuf, tcp_no_metrics_save,
tcp_base_mss, and tcp_mtu_probing.

Signed-off-by: John Heffner <jheffner@psc.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET]: Fix kfree(skb)
Patrick McHardy [Tue, 27 Feb 2007 17:57:37 +0000 (09:57 -0800)]
[NET]: Fix kfree(skb)

Signed-off-by: Patrick McHardy <kaber@trash.net>
Acked-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET]: Handle disabled preemption in gfp_any()
Patrick McHardy [Tue, 27 Feb 2007 17:56:42 +0000 (09:56 -0800)]
[NET]: Handle disabled preemption in gfp_any()

ctnetlink uses netlink_unicast from an atomic_notifier_chain
(which is called within a RCU read side critical section)
without holding further locks. netlink_unicast calls netlink_trim
with the result of gfp_any() for the gfp flags, which are passed
down to pskb_expand_header. gfp_any() only checks for softirq
context and returns GFP_KERNEL, resulting in this warning:

BUG: sleeping function called from invalid context at mm/slab.c:3032
in_atomic():1, irqs_disabled():0
no locks held by rmmod/7010.

Call Trace:
 [<ffffffff8109467f>] debug_show_held_locks+0x9/0xb
 [<ffffffff8100b0b4>] __might_sleep+0xd9/0xdb
 [<ffffffff810b5082>] __kmalloc+0x68/0x110
 [<ffffffff811ba8f2>] pskb_expand_head+0x4d/0x13b
 [<ffffffff81053147>] netlink_broadcast+0xa5/0x2e0
 [<ffffffff881cd1d7>] :nfnetlink:nfnetlink_send+0x83/0x8a
 [<ffffffff8834f6a6>] :nf_conntrack_netlink:ctnetlink_conntrack_event+0x94c/0x96a
 [<ffffffff810624d6>] notifier_call_chain+0x29/0x3e
 [<ffffffff8106251d>] atomic_notifier_call_chain+0x32/0x60
 [<ffffffff881d266d>] :nf_conntrack:destroy_conntrack+0xa5/0x1d3
 [<ffffffff881d194e>] :nf_conntrack:nf_ct_cleanup+0x8c/0x12c
 [<ffffffff881d4614>] :nf_conntrack:kill_l3proto+0x0/0x13
 [<ffffffff881d482a>] :nf_conntrack:nf_conntrack_l3proto_unregister+0x90/0x94
 [<ffffffff883551b3>] :nf_conntrack_ipv4:nf_conntrack_l3proto_ipv4_fini+0x2b/0x5d
 [<ffffffff8109d44f>] sys_delete_module+0x1b5/0x1e6
 [<ffffffff8105f245>] trace_hardirqs_on_thunk+0x35/0x37
 [<ffffffff8105911e>] system_call+0x7e/0x83

Since netlink_unicast is supposed to be callable from within RCU
read side critical sections, make gfp_any() check for in_atomic()
instead of in_softirq().

Additionally nfnetlink_send needs to use gfp_any() as well for the
call to netlink_broadcast).

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[BRIDGE]: Fix locking of set path cost.
Stephen Hemminger [Tue, 27 Feb 2007 17:55:07 +0000 (09:55 -0800)]
[BRIDGE]: Fix locking of set path cost.

This change goes with earlier change to get rid of
work queue for path cost. Now stp_set_path_cost does its own
locking. This is to allow it to call br_path_cost() which calls
ethtool interfaces (might sleep).

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPV6]: /proc/net/anycast6 unbalanced inet6_dev refcnt
David Stevens [Tue, 27 Feb 2007 00:28:56 +0000 (16:28 -0800)]
[IPV6]: /proc/net/anycast6 unbalanced inet6_dev refcnt

Reading /proc/net/anycast6 when there is no anycast address
on an interface results in an ever-increasing inet6_dev reference
count, as well as a reference to the netdevice you can't get rid of.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPX]: Remove ancient changelog
Rolf Eike Beer [Mon, 26 Feb 2007 23:48:03 +0000 (15:48 -0800)]
[IPX]: Remove ancient changelog

Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPX]: Remove outdated information from Kconfig
Rolf Eike Beer [Mon, 26 Feb 2007 23:47:16 +0000 (15:47 -0800)]
[IPX]: Remove outdated information from Kconfig

SPX was removed in early 2.5. How to connect to a Mac or the other OS isn't
hard to find out these days.

Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET]: Revert socket.h/stat.h ifdef hacks.
David S. Miller [Mon, 26 Feb 2007 23:45:15 +0000 (15:45 -0800)]
[NET]: Revert socket.h/stat.h ifdef hacks.

This reverts 57a87bb0720a5cf7a9ece49a8c8ed288398fd1bb.

As H. Peter Anvin states, this change broke klibc and it's
not very easy to fix things up without duplicating everything
into userspace.

In the longer term we should have a better solution to this
problem, but for now let's unbreak things.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPV6]: anycast refcnt fix
Michal Wrobel [Mon, 26 Feb 2007 23:36:10 +0000 (15:36 -0800)]
[IPV6]: anycast refcnt fix

This patch fixes a bug in Linux IPv6 stack which caused anycast address
to be added to a device prior DAD has been completed. This led to
incorrect reference count which resulted in infinite wait for
unregister_netdevice completion on interface removal.

Signed-off-by: Michal Wrobel <xmxwx@asn.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[XFRM] xfrm_user: Fix return values of xfrm_add_sa_expire.
David S. Miller [Mon, 26 Feb 2007 22:52:21 +0000 (14:52 -0800)]
[XFRM] xfrm_user: Fix return values of xfrm_add_sa_expire.

As noted by Kent Yoder, this function will always return an
error.  Make sure it returns zero on success.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[PATCH] x86_64/i386 irq: Fix !CONFIG_SMP compilation
Eric W. Biederman [Tue, 27 Feb 2007 07:27:41 +0000 (00:27 -0700)]
[PATCH] x86_64/i386 irq: Fix !CONFIG_SMP compilation

When removing set_native_irq I missed the fact that it was
called in a couple of places that were compiled even when
SMP support is disabled.  And since the irq_desc[].affinity
field only exists in SMP things broke.

Thanks to Simon Arlott <simon@arlott.org> for spotting this.

There are a couple of ways to fix this but the simplest one
is to just remove the assignments.  The affinity field is only
used to display a value to the user, and nothing on either i386
or x86_64 reads it or depends on it being any particlua value,
so skipping the assignment is safe.  The assignment that
is being removed is just for the initial affinity value before
the user explicitly sets it.  The irq_desc array initializes
this field to CPU_MASK_ALL so the field is initialized to
a reasonable value in the SMP case without being set.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoLinux 2.6.21-rc2
Linus Torvalds [Wed, 28 Feb 2007 04:59:12 +0000 (20:59 -0800)]
Linux 2.6.21-rc2

Too many changes for comfort since -rc1.  Some missed merges, and some
just annoyingly big fixes since.  This is not how an -rc2 should look.
Need to really calm things down!

17 years ago[SPARC64]: Fix PCI interrupts on E450 et al.
David S. Miller [Mon, 26 Feb 2007 22:55:06 +0000 (14:55 -0800)]
[SPARC64]: Fix PCI interrupts on E450 et al.

When the PCI controller OBP node lacks an interrupt-map
and interrupt-map-mask property, we need to form the
INO by hand.  The PCI swizzle logic was not doing that
properly.

This was a regression added by the of_device code.

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