NeilBrown [Thu, 22 Jun 2006 21:47:28 +0000 (14:47 -0700)]
[PATCH] Fix dcache race during umount
The race is that the shrink_dcache_memory shrinker could get called while a
filesystem is being unmounted, and could try to prune a dentry belonging to
that filesystem.
If it does, then it will call in to iput on the inode while the dentry is
no longer able to be found by the umounting process. If iput takes a
while, generic_shutdown_super could get all the way though
shrink_dcache_parent and shrink_dcache_anon and invalidate_inodes without
ever waiting on this particular inode.
Eventually the superblock gets freed anyway and if the iput tried to touch
it (which some filesystems certainly do), it will lose. The promised
"Self-destruct in 5 seconds" doesn't lead to a nice day.
The race is closed by holding s_umount while calling prune_one_dentry on
someone else's dentry. As a down_read_trylock is used,
shrink_dcache_memory will no longer try to prune the dentry of a filesystem
that is being unmounted, and unmount will not be able to start until any
such active prune_one_dentry completes.
This requires that prune_dcache *knows* which filesystem (if any) it is
doing the prune on behalf of so that it can be careful of other
filesystems. shrink_dcache_memory isn't called it on behalf of any
filesystem, and so is careful of everything.
shrink_dcache_anon is now passed a super_block rather than the s_anon list
out of the superblock, so it can get the s_anon list itself, and can pass
the superblock down to prune_dcache.
If prune_dcache finds a dentry that it cannot free, it leaves it where it
is (at the tail of the list) and exits, on the assumption that some other
thread will be removing that dentry soon. To try to make sure that some
work gets done, a limited number of dnetries which are untouchable are
skipped over while choosing the dentry to work on.
I believe this race was first found by Kirill Korotaev.
Cc: Jan Blunck <jblunck@suse.de> Acked-by: Kirill Korotaev <dev@openvz.org> Cc: Olaf Hering <olh@suse.de> Acked-by: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Neil Brown <neilb@suse.de> Signed-off-by: Balbir Singh <balbir@in.ibm.com> Acked-by: David Howells <dhowells@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] avoid tasklist_lock at getrusage for multithreaded case too
Avoid taking tasklist_lock for at getrusage for the multithreaded case too.
We don't need to take the tasklist lock for thread traversal of a process
since Oleg's do-__unhash_process-under-siglock.patch and related work.
Miklos Szeredi [Thu, 22 Jun 2006 21:47:22 +0000 (14:47 -0700)]
[PATCH] remove steal_locks()
This patch removes the steal_locks() function.
steal_locks() doesn't work correctly with any filesystem that does it's own
lock management, including NFS, CIFS, etc.
In addition it has weird semantics on local filesystems in case tasks
sharing file-descriptor tables are doing POSIX locking operations in
parallel to execve().
The steal_locks() function has an effect on applications doing:
clone(CLONE_FILES)
/* in child */
lock
execve
lock
POSIX locks acquired before execve (by "child", "parent" or any further
task sharing files_struct) will after the execve be owned exclusively by
"child".
According to Chris Wright some LSB/LTP kind of suite triggers without the
stealing behavior, but there's no known real-world application that would
also fail.
Apps using NPTL are not affected, since all other threads are killed before
execve.
Apps using LinuxThreads are only affected if they
- have multiple threads during exec (LinuxThreads doesn't kill other
threads, the app may do it with pthread_kill_other_threads_np())
- rely on POSIX locks being inherited across exec
Both conditions are documented, but not their interaction.
Apps using clone() natively are affected if they
- use clone(CLONE_FILES)
- rely on POSIX locks being inherited across exec
The above scenarios are unlikely, but possible.
If the patch is vetoed, there's a plan B, that involves mostly keeping the
weird stealing semantics, but changing the way lock ownership is handled so
that network and local filesystems work consistently.
That would add more complexity though, so this solution seems to be
preferred by most people.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Cc: Matthew Wilcox <willy@debian.org> Cc: Chris Wright <chrisw@sous-sol.org> Cc: Christoph Hellwig <hch@lst.de> Cc: Steven French <sfrench@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
generic_forget_inode()
bdev_clear_inode()
/* restre the fs's ->i_mapping */
inode->i_mapping = &inode->i_data;
/* bdev's inode was freed */
destroy_inode(inode);
if (wait) {
/* dereference a freed bdev's mapping->host */
filemap_fdatawait(mapping); /* Oops */
Since __sync_single_inode() is only taking a ref-count of fs's inode, the
another process can be close() and freeing the bdev's inode while writing
fs's inode. So, __sync_signle_inode() accesses the freed ->i_mapping,
oops.
This patch takes a ref-count on the bdev's inode for the fs's inode before
setting a ->i_mapping, and the clear_inode() of the fs's inode does iput() on
the bdev's inode. So if the fs's inode is still living, bdev's inode
shouldn't be freed.
Andrew Morton [Thu, 22 Jun 2006 21:47:18 +0000 (14:47 -0700)]
[PATCH] suspend_console() warning fix
kernel/power/main.c: In function 'suspend_prepare':
kernel/power/main.c:89: warning: implicit declaration of function 'suspend_console'
kernel/power/main.c: In function 'suspend_finish':
kernel/power/main.c:137: warning: implicit declaration of function 'resume_console'
Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
David Howells [Thu, 22 Jun 2006 21:47:18 +0000 (14:47 -0700)]
[PATCH] Keys: Fix race between two instantiators of a key
Add a revocation notification method to the key type and calls it whilst
the key's semaphore is still write-locked after setting the revocation
flag.
The patch then uses this to maintain a reference on the task_struct of the
process that calls request_key() for as long as the authorisation key
remains unrevoked.
This fixes a potential race between two processes both of which have
assumed the authority to instantiate a key (one may have forked the other
for example). The problem is that there's no locking around the check for
revocation of the auth key and the use of the task_struct it points to, nor
does the auth key keep a reference on the task_struct.
Access to the "context" pointer in the auth key must thenceforth be done
with the auth key semaphore held. The revocation method is called with the
target key semaphore held write-locked and the search of the context
process's keyrings is done with the auth key semaphore read-locked.
The check for the revocation state of the auth key just prior to searching
it is done after the auth key is read-locked for the search. This ensures
that the auth key can't be revoked between the check and the search.
The revocation notification method is added so that the context task_struct
can be released as soon as instantiation happens rather than waiting for
the auth key to be destroyed, thus avoiding the unnecessary pinning of the
requesting process.
Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Michael LeMay [Thu, 22 Jun 2006 21:47:17 +0000 (14:47 -0700)]
[PATCH] selinux: add hooks for key subsystem
Introduce SELinux hooks to support the access key retention subsystem
within the kernel. Incorporate new flask headers from a modified version
of the SELinux reference policy, with support for the new security class
representing retained keys. Extend the "key_alloc" security hook with a
task parameter representing the intended ownership context for the key
being allocated. Attach security information to root's default keyrings
within the SELinux initialization routine.
Has passed David's testsuite.
Signed-off-by: Michael LeMay <mdlemay@epoch.ncsc.mil> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: James Morris <jmorris@namei.org> Acked-by: Chris Wright <chrisw@sous-sol.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Al Viro [Thu, 22 Jun 2006 21:47:14 +0000 (14:47 -0700)]
[PATCH] sparc build breakage
rd_prompt et.al. depend on CONFIG_BLK_DEV_RAM, not CONFIG_BLK_INITRD; now
that those are independent, setup.c blows with INITRD on and BLK_DEV_RAM
off.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Cc: "David S. Miller" <davem@davemloft.net> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Johannes Berg [Wed, 21 Jun 2006 13:43:44 +0000 (15:43 +0200)]
[ALSA] snd-powermac: no longer handle anything with a layout-id property
This patch removes from snd-powermac the code that check for the layout-id
and instead adds code that makes it refuse loading when a layout-id property
is present, nothing that snd-aoa should be used.
It also removes the 'toonie' codec from snd-powermac which was only ever
used on the mac mini which has a layout-id property.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Johannes Berg [Wed, 21 Jun 2006 13:42:43 +0000 (15:42 +0200)]
[ALSA] snd-aoa: add snd-aoa
This large patch adds all of snd-aoa.
Consisting of many modules, it currently replaces snd-powermac
for all layout-id based machines and handles many more (for
example new powerbooks and powermacs with digital output that
previously couldn't be used at all).
It also has support for all layout-IDs that Apple has (judging
from their Info.plist file) but not all are tested.
The driver currently has 2 known regressions over snd-powermac:
* it doesn't handle powermac 7,2 and 7,3
* it doesn't have a DRC control on snapper-based machines
I will fix those during the 2.6.18 development cycle.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Justin Sunseri reports that sound is audible on his Lenovo 3000
N100-07684JU by passing 'model=laptop-eapd' to modprobe, so this
patch adds the pci ids for his sound device to patch_analog.c .
This commit closes LP#39517. Alexey Parshin also confirmed the fix at
http://bugs.gentoo.org/137245
TODO: Mute onboard speakers when device is plugged into the headphone
jack. Muting the 'External Amplifier' mixer element while a device is
plugged into the headphone jack allows sound to be played only from the
headphone jack.
From: Daniel T Chen <crimsun@ubuntu.com> Signed-off-by: Daniel T Chen <crimsun at ubuntu.com> Signed-off-by: Daniel Drake <dsd@gentoo.org> Signed-off-by: Jaroslav Kysela <perex@suse.cz>
Jaya Kumar [Mon, 19 Jun 2006 13:06:14 +0000 (15:06 +0200)]
[ALSA] AD1888 suspend/resume fix
This patch adds a write to an undocumented register, 0x60 Extended
Codec Register Page in the AD1888 codec. It is neccessary in order
to make suspend/resume work with the AD1888.
Rodolfo Giometti [Mon, 19 Jun 2006 13:04:54 +0000 (15:04 +0200)]
[ALSA] Disable AC97 AUX and VIDEO controls for WM9705 touchscreen
This patch by Rodolfo Giometti disables the AC97 AUX and VIDEO controls
on the WM9705 when the touchscreen is selected as the AUX and VIDEO
lines are shared with the touch controller.
Changes:-
o Added AC97_HAS_NO_AUX flag
o Test for AC97_HAS_NO_AUX flag in snd_ac97_mixer_build()
o Sets AC97_HAS_NO_VIDEO and AC97_HAS_NO_AUX in patch_wolfson05() when
WM9705 touch driver is selected.
Daniel T Chen [Fri, 16 Jun 2006 14:21:54 +0000 (16:21 +0200)]
[ALSA] Add hp_only quirk for pci id [161f:2032] to via82xx
http://www.kernel.org/git/?p=linux/kernel/git/bcollins/ubuntu-dapper.git;a=commitdiff;h=eae2cc78de39502595f67b7fc1f821f5963bb8ae
UpstreamStatus: Not merged
Christian Bjalevik reports in LP#38546 that his sound chipset requires
the 'hp_only' quirk to allow him to control sound volume correctly when
headphones are inserted. This patch adds the appropriate pci id to the
via82xx ALSA driver so that the quirk is applied automatically, thereby
removing the need for users to modify /etc/modprobe.d/alsa-base (or to
unload and reload snd-via82xx with ac97_quirk=hp_only).
This patch closes LP#38546.
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com> Signed-off-by: Ben Collins <bcollins@ubuntu.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Daniel T Chen [Fri, 16 Jun 2006 10:13:00 +0000 (12:13 +0200)]
[ALSA] sound/pci/: Add hp_only quirk for Dell D800 laptops
http://www.kernel.org/git/?p=linux/kernel/git/bcollins/ubuntu-dapper.git;a=commitdiff;h=9ad787cd9670c3f3b8f3db235e84baf00a2ea526
Anders Ostling comments in Malone #41015 that his Dell D800 laptop's
volume control works correctly when the hp_only quirk is passed to
modprobe. This commit adds his hardware's sub{vendor,device} ids to
the quirk list for the intel8x0 driver.
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com> Signed-off-by: Ben Collins <bcollins@ubuntu.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Jesper Juhl [Tue, 13 Jun 2006 09:58:12 +0000 (11:58 +0200)]
[ALSA] fix potential NULL pointer deref in snd_sb8dsp_midi_interrupt()
First testing if a pointer is NULL and if it is (or might be), proceeding
with code that dereferences that same pointer is clearly a mistake.
This happens in sound/isa/sb/sb8_midi.c::snd_sb8dsp_midi_interrupt()
The patch below reworks the code so this unfortunate case doesn't happen.
Also remove some blank comments.
Found by the Coverity checker as bug #367
Patch is compile testted only due to lack of hardware.
Takashi Iwai [Tue, 13 Jun 2006 09:57:22 +0000 (11:57 +0200)]
[ALSA] hda-codec - Add SPDIF support to Thinkpad T/X/Z60
Added IEC958 (SPDIF) output support to Thinkpad T/X/Z60 with
AD1981HD codec. The spdif jack is on docking station.
Also, renamed 'IEC958 Playback Route' to 'IEC958 Playback Source'
to avoid the mixer name confliction with IEC958 switch.
Jaya Kumar [Mon, 12 Jun 2006 08:08:02 +0000 (10:08 +0200)]
[ALSA] ac97_codec - fix duplicate control creation in AC97
This patch conditions AC97 control creation by whether or not the
codec is an AD18xx codec. This fixes the case where the default control
would get created and then snd_ac97_mixer_build fails out when creation
of ad18xx specific control would get attempted. This problem was found
and debuged by Marcelo Tosatti.
Signed-off-by: Jaya Kumar <jayakumar.alsa@gmail.com> Signed-off-by: Jaroslav Kysela <perex@suse.cz>
Eric Sesterhenn [Wed, 31 May 2006 09:55:17 +0000 (11:55 +0200)]
[ALSA] NULL pointer dereference in sound/synth/emux/soundfont.c
this is about coverity id #100.
It seems the if statement is negated, since the else branch calls
remove_info() with sflist->currsf as a parameter where it gets
dereferenced.
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Alan Horstmann [Fri, 26 May 2006 15:19:34 +0000 (17:19 +0200)]
[ALSA] ice1712 - Set mpu401 info flags from _card_info
To permit use, in ice1712, of the mpu401 info flags recently added to
mpu401_uart, adds info_flags in snd_ice1712_card_info so that additional
flags can be set, if desired. Â 'MPU401_INFO_INTEGRATED' is always set with
the ice1712. Â The flags are passed on to snd_mpu401_uart_new().
_INFO_OUTPUT is set for DMX6fire mpu2.
Signed-off-by: Alan Horstmann <gineera@aspect135.co.uk> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Tue, 23 May 2006 13:46:10 +0000 (15:46 +0200)]
[ALSA] ice1724 - Add functionality for Audiotrak Prodigy 7.1 LT
This patch adds support for useable front audio channels,
user controllable headphone channel and optical output.
From: Anho Ki Signed-off-by: Matt Taylor <mwtaylor@gmail.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Tue, 23 May 2006 11:24:30 +0000 (13:24 +0200)]
[ALSA] Change an arugment of snd_mpu401_uart_new() to bit flags
Change the 5th argument of snd_mpu401_uart_new() to bit flags
instead of a boolean. The argument takes bits that consist of
MPU401_INFO_XXX flags.
The callers that used the value 1 there are replaced with
MPU401_INFO_INTEGRATED.
Jaya Kumar [Fri, 19 May 2006 10:04:22 +0000 (12:04 +0200)]
[ALSA] Single variables for cs5535audio
As per Takashi's feedback, this is a cleanup to make cs5535audio be
single device per system. The diff is against 2.6.17-rc4 with
Takashi's patch adding the module_params for index, id and enable.
Clemens Ladisch [Thu, 18 May 2006 07:35:15 +0000 (09:35 +0200)]
[ALSA] usb-audio: add workaround for CSR Bluetooth Headphones (Saitek A-250)
Some devices don't have the required class-specific endpoint descriptors.
Instead of making this case an error, this patch makes the driver guess the
endpoint attributes.
Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Takashi Iwai [Wed, 17 May 2006 15:10:35 +0000 (17:10 +0200)]
[ALSA] Fix pcm-draining of capture stream in PCM middle layer
Fix the draining of PCM capture stream in the PCM middle layer.
snd_pcm_drain() ignored capture streams, but it should change
the state to SNDRV_PCM_DRAINING.
Andreas Mohr [Wed, 17 May 2006 09:02:24 +0000 (11:02 +0200)]
[ALSA] azt3328.c: add suspend/resume support
- add suspend/resume handlers
- fix problem (private_data members not set)
Playing a file while suspending will resume correctly with this patch,
so I assume the hardware to get fully correctly reinitialized with
this patch.
Signed-off-by: Andreas Mohr <andi@lisas.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Mon, 15 May 2006 17:49:05 +0000 (19:49 +0200)]
[ALSA] Fix rwlock around snd_iprintf() in sound core
Fixed rwlock around snd_iprintf() in sound core part.
Replaced with mutex.
Also, make mutex and flags static variables with addition of
snd_card_locked() function (just for sound.c).
Takashi Iwai [Thu, 4 May 2006 12:36:08 +0000 (14:36 +0200)]
[ALSA] cmipci - Disable integrated mpu401 as default
Enable the support of mpu401 PCI port only when mpu_port=1 module
option is given, i.e. disabled as default.
It turned out that the check of integrated midi port isn't perfect
and caused hang-ups on some boards.
Alan Horstmann [Wed, 3 May 2006 15:07:29 +0000 (17:07 +0200)]
[ALSA] au88x0 - Init before create components
Change the order in vortex_probe to set the card details before creating the
components, meaning for example that card->shortname is available when
registering the midi port.
I have also added extra to card->shortname, and a line to overwrite the midi
name following snd_mpu401_uart_new.
Signed-off-by: Alan Horstmann <gineera@aspect135.co.uk> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Clemens Ladisch [Tue, 2 May 2006 14:22:12 +0000 (16:22 +0200)]
[ALSA] rawmidi: add get_port_info callback for sequencer information flags
Add a get_port_info callback to the snd_rawmidi_global_ops structure to
allow the USB MIDI driver to supply information flags for the sequencer
ports created by seq_midi.
Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Clemens Ladisch [Tue, 2 May 2006 14:08:41 +0000 (16:08 +0200)]
[ALSA] add more sequencer port type information bits
Add four new information flags SNDRV_SEQ_PORT_TYPE_HARDWARE, _SOFTWARE,
_SYNTHESIZER, _PORT for sequencer ports. This makes it easier for apps
like Rosegarden to make policy decisions based on the port type.
Signed-off-by: Clemens Ladisch <clemens@ladisch.de>