aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2009-07-21sky2: Avoid races in sky2_downMike McCormack1-8/+8
Reset rx chain before trying to drain it. Shut interrupts off last, incase there's something to report. Signed-off-by: Mike McCormack <[email protected]> Acked-by: Stephen Hemminger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21ide-tape: Don't leak kernel stack informationMichael Buesch1-0/+1
Don't leak kernel stack information through uninitialized structure members. Signed-off-by: Michael Buesch <[email protected]> Acked-by: Borislav Petkov <[email protected]>. Signed-off-by: David S. Miller <[email protected]>
2009-07-21ide: fix memory leak when flush command is issuedMaxime Bizon1-0/+1
I'm using ide on 2.6.30.1 with xfs filesystem. I noticed a kernel memory leak after writing lots of data, the kmalloc-96 slab cache keeps growing. It seems the struct ide_cmd kmalloced by idedisk_prepare_flush is never kfreed. Commit a09485df9cda49fbde2766c86eb18a9cae585162 ("ide: move request type specific code from ide_end_drive_cmd() to callers (v3)") and f505d49ffd25ed062e76ffd17568d3937fcd338c ("ide: fix barriers support") cause this regression, cmd->rq must now be set for ide_complete_cmd to honor the IDE_TFLAG_DYN flag. Signed-off-by: Maxime Bizon <[email protected]> Acked-by: Bartlomiej Zolnierkiewicz <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21NFSv4: Fix a problem whereby a buggy server can oops the kernelTrond Myklebust2-5/+13
We just had a case in which a buggy server occasionally returns the wrong attributes during an OPEN call. While the client does catch this sort of condition in nfs4_open_done(), and causes the nfs4_atomic_open() to return -EISDIR, the logic in nfs_atomic_lookup() is broken, since it causes a fallback to an ordinary lookup instead of just returning the error. When the buggy server then returns a regular file for the fallback lookup, the VFS allows the open, and bad things start to happen, since the open file doesn't have any associated NFSv4 state. The fix is firstly to return the EISDIR/ENOTDIR errors immediately, and secondly to ensure that we are always careful when dereferencing the nfs_open_context state pointer. Signed-off-by: Trond Myklebust <[email protected]>
2009-07-21x86-64: Fix bad_srat() to clear all stateAndi Kleen1-2/+4
Need to clear both nodes and nodes_add state for start/end. Signed-off-by: Andi Kleen <[email protected]> LKML-Reference: <[email protected]> Signed-off-by: H. Peter Anvin <[email protected]> Cc: [email protected]
2009-07-21NFSv4: Fix an NFSv4 mount regressionTrond Myklebust3-22/+26
Commit 008f55d0e019943323c20a03493a2ba5672a4cc8 (nfs41: recover lease in _nfs4_lookup_root) forces the state manager to always run on mount. This is a bug in the case of NFSv4.0, which doesn't require us to send a setclientid until we want to grab file state. In any case, this is completely the wrong place to be doing state management. Moving that code into nfs4_init_session... Signed-off-by: Trond Myklebust <[email protected]>
2009-07-21NFSv4: Fix an Oops in nfs4_free_lock_stateTrond Myklebust1-1/+1
The oops http://www.kerneloops.org/raw.php?rawid=537858&msgid= appears to be due to the nfs4_lock_state->ls_state field being uninitialised. This happens if the call to nfs4_free_lock_state() is triggered at the end of nfs4_get_lock_state(). The fix is to move the initialisation of ls_state into the allocator. Signed-off-by: Trond Myklebust <[email protected]>
2009-07-21drivers/net/mlx4: Adjust constantJulia Lawall1-1/+1
The values in the advertising field are typically ADVERTISED_xxx, not SUPPORTED_xxx. Both SUPPORTED_10000baseT_Full and ADVERTISED_1000baseT_Full have the same value. The semantic match that finds this problem is as follows: (http://www.emn.fr/x-info/coccinelle/) // <smpl> @@ struct ethtool_cmd E; @@ *E.advertising = SUPPORTED_10000baseT_Full // </smpl> Signed-off-by: Julia Lawall <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21drivers/net: Move a dereference below a NULL testJulia Lawall1-1/+2
If the NULL test is necessary, then the dereference should be moved below the NULL test. The semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // <smpl> @@ type T; expression E,E1; identifier i,fld; statement S; @@ - T i = E->fld; + T i; ... when != E=E1 when != i if (E == NULL||...) S + i = E->fld; // </smpl> Signed-off-by: Julia Lawall <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21drivers/net: Move a dereference below a NULL testJulia Lawall1-3/+4
If the NULL test is necessary, then the dereferences should be moved below the NULL test. The semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // <smpl> @@ type T; expression E,E1; identifier i,fld; statement S; @@ - T i = E->fld; + T i; ... when != E=E1 when != i BUG_ON (E == NULL||...); + i = E->fld; // </smpl> Signed-off-by: Julia Lawall <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21libertas: Fix problem with broken V4 firmware on CF8381Marek Vasut2-1/+9
Firmware V4 on CF8381 reports region code shifted by 1 byte to left. The following patch checks for this and handles it properly. Signed-off-by: Marek Vasut <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21ath: add support for special 0x8000 regulatory domainLuis R. Rodriguez1-0/+17
Two users of ar9170 devices have now reported their cards have been programmed with a regulatory domain of 0x8000. This is not a valid regulatory domain as such these users were unable to use these devices. Since this doesn't seem to be a device EEPROM corruption we must treat it specially. It may have been possible the manufacturer intended to use 0x0 as the regulatory domain and that would ultimately yield to US but since we cannot get confirmationf or this we default this special case to one of our world regulatory domains, specifically 0x64. Reported-by: DavidFreeman on #linux-wireless Reported-by: Joerg Albert <[email protected]> Cc: Christian Lamparter <[email protected]>, Cc: Stephen Chen <[email protected]> Cc: David Quan <[email protected]> Cc: Tony Yang <[email protected]> Signed-off-by: Luis R. Rodriguez <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21p54spi: fix potential null deref in p54spi.cDan Carpenter1-1/+1
Fix a potential NULL dereference bug during error handling in p54spi_probe. This bug was discovered by smatch: (http://repo.or.cz/w/smatch.git). Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: Christian Lamparter <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21acer-wmi: fix rfkill conversionAlan Jenkins1-1/+1
Fix another polarity error introduced by the rfkill rewrite, this time in acer_rfkill_set(). Signed-off-by: Alan Jenkins <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21connector: maintainer/mail update.Evgeniy Polyakov5-6/+13
Signed-off-by: Evgeniy Polyakov <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21USB host CDC Phonet network interface driverRémi Denis-Courmont3-0/+470
Many Nokia handsets support a Phonet interface to the cellular modem via a vendor-specific USB interface. CDC Phonet follows the Communications Device Class model, with one control interface, and and a pair of inactive and active data alternative interface. The later has two bulk endpoint, one per direction. This was tested against Nokia E61, Nokia N95, and the existing Phonet gadget function for the Linux composite USB gadget framework. Signed-off-by: Rémi Denis-Courmont <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21inotify: use GFP_NOFS under potential memory pressureEric Paris4-12/+21
inotify can have a watchs removed under filesystem reclaim. ================================= [ INFO: inconsistent lock state ] 2.6.31-rc2 #16 --------------------------------- inconsistent {IN-RECLAIM_FS-W} -> {RECLAIM_FS-ON-W} usage. khubd/217 [HC0[0]:SC0[0]:HE1:SE1] takes: (iprune_mutex){+.+.?.}, at: [<c10ba899>] invalidate_inodes+0x20/0xe3 {IN-RECLAIM_FS-W} state was registered at: [<c10536ab>] __lock_acquire+0x2c9/0xac4 [<c1053f45>] lock_acquire+0x9f/0xc2 [<c1308872>] __mutex_lock_common+0x2d/0x323 [<c1308c00>] mutex_lock_nested+0x2e/0x36 [<c10ba6ff>] shrink_icache_memory+0x38/0x1b2 [<c108bfb6>] shrink_slab+0xe2/0x13c [<c108c3e1>] kswapd+0x3d1/0x55d [<c10449b5>] kthread+0x66/0x6b [<c1003fdf>] kernel_thread_helper+0x7/0x10 [<ffffffff>] 0xffffffff Two things are needed to fix this. First we need a method to tell fsnotify_create_event() to use GFP_NOFS and second we need to stop using one global IN_IGNORED event and allocate them one at a time. This solves current issues with multiple IN_IGNORED on a queue having tail drop problems and simplifies the allocations since we don't have to worry about two tasks opperating on the IGNORED event concurrently. Signed-off-by: Eric Paris <[email protected]>
2009-07-21fsnotify: fix inotify tail drop check with path entriesEric Paris1-0/+1
fsnotify drops new events when they are the same as the tail event on the queue to be sent to userspace. The problem is that if the event comes with a path we forget to break out of the switch statement and fall into the code path which matches on events that do not have any type of file backed information (things like IN_UNMOUNT and IN_Q_OVERFLOW). The problem is that this code thinks all such events should be dropped. Fix is to add a break. Signed-off-by: Eric Paris <[email protected]>
2009-07-21inotify: check filename before dropping repeat eventsEric Paris1-2/+7
inotify drops events if the last event on the queue is the same as the current event. But it does 2 things wrong. First it is comparing old->inode with new->inode. But after an event if put on the queue the ->inode is no longer allowed to be used. It's possible between the last event and this new event the inode could be reused and we would falsely match the inode's memory address between two differing events. The second problem is that when a file is removed fsnotify is passed the negative dentry for the removed object rather than the postive dentry from immediately before the removal. This mean the (broken) inotify tail drop code was matching the NULL ->inode of differing events. The fix is to check the file name which is stored with events when doing the tail drop instead of wrongly checking the address of the stored ->inode. Reported-by: Scott James Remnant <[email protected]> Signed-off-by: Eric Paris <[email protected]>
2009-07-21fsnotify: use def_bool in kconfig instead of letting the user chooseEric Paris3-13/+3
fsnotify doens't give the user anything. If someone chooses inotify or dnotify it should build fsnotify, if they don't select one it shouldn't be built. This patch changes fsnotify to be a def_bool=n and makes everything else select it. Also fixes the issue people complained about on lwn where gdm hung because they didn't have inotify and they didn't get the inotify build option..... Signed-off-by: Eric Paris <[email protected]>
2009-07-21inotify: fix error paths in inotify_update_watchEric Paris1-30/+49
inotify_update_watch could leave things in a horrid state on a number of error paths. We could try to remove idr entries that didn't exist, we could send an IN_IGNORED to userspace for watches that don't exist, and a bit of other stupidity. Clean these up by doing the idr addition before we put the mark on the inode since we can clean that up on error and getting off the inode's mark list is hard. Signed-off-by: Eric Paris <[email protected]>
2009-07-21inotify: do not leak inode marks in inotify_add_watchEric Paris1-4/+10
inotify_add_watch had a couple of problems. The biggest being that if inotify_add_watch was called on the same inode twice (to update or change the event mask) a refence was taken on the original inode mark by fsnotify_find_mark_entry but was not being dropped at the end of the inotify_add_watch call. Thus if inotify_rm_watch was called although the mark was removed from the inode, the refcnt wouldn't hit zero and we would leak memory. Reported-by: Catalin Marinas <[email protected]> Signed-off-by: Eric Paris <[email protected]>
2009-07-21inotify: drop user watch count when a watch is removedEric Paris1-0/+2
The inotify rewrite forgot to drop the inotify watch use cound when a watch was removed. This means that a single inotify fd can only ever register a maximum of /proc/sys/fs/max_user_watches even if some of those had been freed. Signed-off-by: Eric Paris <[email protected]>
2009-07-21macsonic, jazzsonic: fix oops on module unloadFinn Thain2-0/+2
Set the driver data before using it. Fixes an oops when doing rmmod. Signed-off-by: Finn Thain <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21macsonic: move probe function to .devinit.textUwe Kleine-König1-7/+7
A pointer to mac_sonic_probe is passed to the core via platform_driver_register and so the function must not disappear when the .init sections are discarded. Otherwise (if also having HOTPLUG=y) unbinding and binding a device to the driver via sysfs will result in an oops as does a device being registered late. Various other functions that are called by mac_sonic_probe need to move to .devinit.text, too. An alternative to this patch is using platform_driver_probe instead of platform_driver_register plus removing the pointer to the probe function from the struct platform_driver. Signed-off-by: Uwe Kleine-König <[email protected]> Tested-by: Finn Thain <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2009-07-21x86, mce: Fix set_trigger() accessorJan Beulich1-4/+2
Fix the condition checking the result of strchr() (which previously could result in an oops), and make the function return the number of bytes actively used. [ Impact: fix oops ] Signed-off-by: Jan Beulich <[email protected]> Cc: Andi Kleen <[email protected]> LKML-Reference: <[email protected]> Signed-off-by: H. Peter Anvin <[email protected]>
2009-07-21Merge branch 'fix/misc' into for-linusTakashi Iwai3-9/+6
* fix/misc: ALSA: ca0106 - Fix the max capture buffer size ALSA: OSS sequencer should be initialized after snd_seq_system_client_init ALSA: sound/isa: convert nested spin_lock_irqsave to spin_lock
2009-07-21Merge branch 'fix/hda' into for-linusTakashi Iwai2-2/+11
* fix/hda: ALSA: hda - Fix pin-setup for Sony VAIO with STAC9872 codecs ALSA: hda - Add quirk for Gateway T6834c laptop ALSA: hda_codec: Check for invalid zero connections
2009-07-21Merge branch 'fix/ctxfi' into for-linusTakashi Iwai1-2/+2
* fix/ctxfi: ALSA: ctxfi: Swapped SURROUND-SIDE channels on emu20k2
2009-07-21cfg80211: double free in __cfg80211_scan_doneChristian Lamparter1-2/+1
This patch fixes a double free corruption in __cfg80211_scan_done: ================================================ BUG kmalloc-512: Object already free ------------------------------------------------ INFO: Allocated in load_elf_binary+0x18b/0x19af age=6 INFO: Freed in load_elf_binary+0x104e/0x19af age=5 INFO: Slab 0xffffea0001bae4c0 objects=14 used=7 INFO: Object 0xffff88007e8a9918 @offset=6424 fp=0xffff88007e8a9488 Bytes b4 0xffff88007e8a9908: 00 00 00 00 00 00 00 00 5a 5a [...] Pid: 28705, comm: rmmod Tainted: P C 2.6.31-rc2-wl #1 Call Trace: [<ffffffff810da9f4>] print_trailer+0x14e/0x16e [<ffffffff810daa56>] object_err+0x42/0x61 [<ffffffff810dbcd9>] __slab_free+0x2af/0x396 [<ffffffffa0ec9694>] ? wiphy_unregister+0x92/0x142 [cfg80211] [<ffffffff810dd5e3>] kfree+0x13c/0x17a [<ffffffffa0ec9694>] ? wiphy_unregister+0x92/0x142 [cfg80211] [<ffffffffa0ec9694>] wiphy_unregister+0x92/0x142 [cfg80211] [<ffffffffa0eed163>] ieee80211_unregister_hw+0xc8/0xff [mac80211] [<ffffffffa0f3fbc8>] p54_unregister_common+0x31/0x66 [p54common] [...] FIX kmalloc-512: Object at 0xffff88007e8a9918 not freed The code path which leads to the *funny* double free: request = rdev->scan_req; dev = dev_get_by_index(&init_net, request->ifidx); /* * the driver was unloaded recently and * therefore dev_get_by_index will return NULL! */ if (!dev) goto out; [...] rdev->scan_req = NULL; /* not executed... */ [...] out: kfree(request); Signed-off-by: Christian Lamparter <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21rfkill: remove too-strict __must_checkAlan Jenkins1-1/+1
Some drivers don't need the return value of rfkill_set_hw_state(), so it should not be marked as __must_check. Signed-off-by: Alan Jenkins <[email protected]> Acked-by: Johannes Berg <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21ath9k: Tune ANI function processing on AP mode during ANI resetLuis R. Rodriguez1-0/+12
For AP mode we must tune ANI specially for 2 GHz and for 5 GHz. We mask in only the flags we want to toggle on ath9k_hw_ani_control() through the ah->ani_function bitmask, this will take care of ignoring changes during ANI reset which we were disabling before. Testedy-by: Steven Luo <[email protected]> Cc: Bennyam Malavazi <[email protected]> Cc: Jouni Malinen <[email protected]> Signed-off-by: Luis R. Rodriguez <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21rt2x00: Fix chipset detection for rt2500usbAndy Whitcroft1-1/+3
The commit below changed the semantics of rt2x00_check_rev so that it no longer checked the bottom 4 bits of the rev were non-zero. During that conversion this part of the check was not propogated to the rt2500usb initialisation. commit 358623c22c9fd837b3b1b444377037f72553dc9f Author: Ivo van Doorn <[email protected]> Date: Tue May 5 19:46:08 2009 +0200 rt2x00: Simplify rt2x00_check_rev Without this check rt73 devices are miss recognised as rt2500 devices and two drivers are loaded. Preventing the device being used. Reinstate this check. Signed-off-by: Andy Whitcroft <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21nl80211: Memory leak fixedNiko Jokinen1-2/+3
Potential memory leak via msg pointer in nl80211_get_key() function. Signed-off-by: Niko Jokinen <[email protected]> Signed-off-by: Luciano Coelho <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21rtl8187: Fix for kernel oops when unloading with LEDs enabledLarry Finger1-1/+2
When rtl8187 is unloaded and CONFIG_RTL8187_LEDS is set, the kernel may oops when the module is unloaded as the workqueue for led_on was not being cancelled. This patch fixes the problem reported in http://marc.info/?l=linux-wireless&m=124742957615781&w=2. Reported-by: Gábor Stefanik <[email protected]> Signed-off-by: Larry Finger <Larry.Finger@lwfinger> Signed-off-by: John W. Linville <[email protected]>
2009-07-21ath5k: temporarily disable crypto for AP modeBob Copeland1-0/+3
Pavel Roskin reported some issues with using AP mode without nohwcrypt=1. Most likely this is similar to the problem fixed some time ago in ath9k by 3f53dd64f192450cb331c0fecfc26ca952fb242f, "ath9k: Fix hw crypto configuration for TKIP in AP mode." That only affects TKIP but it's easiest to just disable that and WEP too until we get a proper fix in. Signed-off-by: Bob Copeland <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21mac80211: use correct address for mesh Path ErrorJavier Cardona1-3/+5
For forwarded frames, we save the precursor address in addr1 in case it needs to be used to send a Path Error. mesh_path_discard_frame, however, was using addr2 instead of addr1 to send Path Error frames, so correct that and also make the comment regarding this more clear. Signed-off-by: Andrey Yurovsky <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21mac80211_hwsim: fix use after freeJohannes Berg1-2/+2
Once the "data" pointer is freed, we can't be iterating to the next item in the list any more so we need to use list_for_each_entry_safe with a temporary variable. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21mac80211_hwsim: fix unregistrationJohannes Berg1-1/+1
If you rmmod the module while associated, frames might be transmitted during unregistration -- which will crash if the hwsim%d interface is unregistered first, so only do that after all the virtual wiphys are gone. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21rfkill: fix rfkill_set_states() to set the hw stateAlan Jenkins1-0/+4
The point of this function is to set the software and hardware state at the same time. When I tried to use it, I found it was only setting the software state. Signed-off-by: Alan Jenkins <[email protected]> Reviewed-by: Johannes Berg <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21mac80211: fix injection in monitor modePavel Roskin1-1/+1
The location of the 802.11 header is calculated incorrectly due to a wrong placement of parentheses. Found by kmemcheck. Signed-off-by: Pavel Roskin <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21rfkill: allow toggling soft state in sysfs againJohannes Berg1-8/+19
Apparently there actually _are_ tools that try to set this in sysfs even though it wasn't supposed to be used this way without claiming first. Guess what: now that I've cleaned it all up it doesn't matter and we can simply allow setting the soft-block state in sysfs. Signed-off-by: Johannes Berg <[email protected]> Tested-By: Darren Salt <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21mac80211: disable meshJohannes Berg2-0/+4
My kvm instance was complaining a lot about sleeping in atomic contexts in the mesh code, and it turns out that both mesh_path_add() and mpp_path_add() need to be able to sleep (they even use synchronize_rcu()!). I put in a might_sleep() to annotate that, but I see no way, at least right now, of actually making sure those functions are only called from process context since they are both called during TX and RX and the mesh code itself even calls them with rcu_read_lock() "held". Therefore, let's disable it completely for now. It's possible that I'm only seeing this because the hwsim's beaconing is broken and thus the peers aren't discovered right away, but it is possible that this happens even if beaconing is working, for a peer that doesn't exist or so. It should be possible to solve this by deferring the freeing of the tables to call_rcu() instead of using synchronize_rcu(), and also using atomic allocations, but maybe it makes more sense to rework the code to not call these from atomic contexts and defer more of the work to the workqueue. Right now, I can't work on either of those solutions though. Signed-off-by: Johannes Berg <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21iwlwifi: only update byte count table during aggregationReinette Chatre1-1/+2
The byte count table is only used for aggregation. Updating it in other cases caused fragmented frames to be dropped. This fixes http://www.intellinuxwireless.org/bugzilla/show_bug.cgi?id=2004 Signed-off-by: Reinette Chatre <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21iwlwifi: only show active power level via sysfsReinette Chatre2-6/+2
This changes the power_level file to adhere to the "one value per file" sysfs rule. The user will know which power level was requested as it will be the number just written to this file. It is thus not necessary to create a new sysfs file for this value. In addition it fixes a problem where powertop's parsing expects this value to be the first value in this file without any descriptions. Signed-off-by: Reinette Chatre <[email protected]> cc: [email protected] Signed-off-by: John W. Linville <[email protected]>
2009-07-21iwmc3200wifi: fix NULL pointer dereference in iwm_if_freeZhu Yi1-1/+1
The driver private data is now based on wiphy. So we should not touch the private data after wiphy_free() is called. The patch fixes the potential NULL pointer dereference by making the iwm_wdev_free() the last one on the interface removal path. Signed-off-by: Zhu Yi <[email protected]> Signed-off-by: John W. Linville <[email protected]>
2009-07-21genirq: Delegate irq affinity setting to the irq threadThomas Gleixner4-9/+48
irq_set_thread_affinity() calls set_cpus_allowed_ptr() which might sleep, but irq_set_thread_affinity() is called with desc->lock held and can be called from hard interrupt context as well. The code has another bug as it does not hold a ref on the task struct as required by set_cpus_allowed_ptr(). Just set the IRQTF_AFFINITY bit in action->thread_flags. The next time the thread runs it migrates itself. Solves all of the above problems nicely. Add kerneldoc to irq_set_thread_affinity() while at it. Signed-off-by: Thomas Gleixner <[email protected]> LKML-Reference: <new-submission>
2009-07-21jbd: fix race between write_metadata_buffer and get_write_accessdingdinghua1-9/+11
The function journal_write_metadata_buffer() calls jbd_unlock_bh_state(bh_in) too early; this could potentially allow another thread to call get_write_access on the buffer head, modify the data, and dirty it, and allowing the wrong data to be written into the journal. Fortunately, if we lose this race, the only time this will actually cause filesystem corruption is if there is a system crash or other unclean shutdown of the system before the next commit can take place. Signed-off-by: dingdinghua <[email protected]> Acked-by: "Theodore Ts'o" <[email protected]> Signed-off-by: Jan Kara <[email protected]>
2009-07-20x86: Fix movq immediate operand constraints in uaccess.hH. Peter Anvin1-2/+2
The movq instruction, generated by __put_user_asm() when used for 64-bit data, takes a sign-extended immediate ("e") not a zero-extended immediate ("Z"). Signed-off-by: H. Peter Anvin <[email protected]> Cc: Uros Bizjak <[email protected]> Cc: [email protected]
2009-07-20Input: atkbd - add force relese key quirk for Soltech TA12Jerone Young1-0/+16
Netbooks based on the Soltech TA12 do not send a key release for volume keys causing Linux to think the key is constantly being pressed forever. Added quirk data for forced release keys. BugLink: https://bugs.launchpad.net//bugs/397499 Signed-off-by: Jerone Young <[email protected]> Signed-off-by: Tim Gardner <[email protected]> Signed-off-by: Dmitry Torokhov <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.