- 12 Apr, 2016 40 commits
-
-
Aaro Koskinen authored
commit 4c11e554 upstream. Commit 1f330c32 ("drivers/firmware/broadcom/bcm47xx_nvram.c: use __ioread32_copy() instead of open-coding") switched to use a generic copy function, but failed to notice that the header pointer is updated between the two copies, resulting in bogus data being copied in the latter one. Fix by keeping the old header pointer. The patch fixes totally broken networking on WRT54GL router (both LAN and WLAN interfaces fail to probe). Fixes: 1f330c32 ("drivers/firmware/broadcom/bcm47xx_nvram.c: use __ioread32_copy() instead of open-coding") Signed-off-by:
Aaro Koskinen <aaro.koskinen@iki.fi> Reviewed-by:
Stephen Boyd <sboyd@codeaurora.org> Cc: Rafal Milecki <zajec5@gmail.com> Cc: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Luis R. Rodriguez authored
commit b0f84ac3 upstream. All architectures now need ioremap_uc(), ia64 seems defines this already through its ioremap_nocache() and it already ensures it *only* uses UC. This is needed since v4.3 to complete an allyesconfig compile on ia64, there were others archs that needed this, and this one seems to have fallen through the cracks. Signed-off-by:
Luis R. Rodriguez <mcgrof@kernel.org> Reported-by:
kbuild test robot <fengguang.wu@intel.com> Acked-by:
Tony Luck <tony.luck@intel.com> Signed-off-by:
-
Johannes Weiner authored
commit b6e6edcf upstream. Setting the original memory.limit_in_bytes hardlimit is subject to a race condition when the desired value is below the current usage. The code tries a few times to first reclaim and then see if the usage has dropped to where we would like it to be, but there is no locking, and the workload is free to continue making new charges up to the old limit. Thus, attempting to shrink a workload relies on pure luck and hope that the workload happens to cooperate. To fix this in the cgroup2 memory.max knob, do it the other way round: set the limit first, then try enforcement. And if reclaim is not able to succeed, trigger OOM kills in the group. Keep going until the new limit is met, we run out of OOM victims and there's only unreclaimable memory left, or the task writing to memory.max is killed. This allows users to shrink groups reliably, and the behavior is consistent with what happens when new charges are attempted in excess of memory.max. Signed-off-by:
Johannes Weiner <hannes@cmpxchg.org> Acked-by:
Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by:
-
Johannes Weiner authored
commit 588083bb upstream. When setting memory.high below usage, nothing happens until the next charge comes along, and then it will only reclaim its own charge and not the now potentially huge excess of the new memory.high. This can cause groups to stay in excess of their memory.high indefinitely. To fix that, when shrinking memory.high, kick off a reclaim cycle that goes after the delta. Signed-off-by:
-
Eric Wheeler authored
commit f8b11260 upstream. When bch_cache_set_alloc() fails to kzalloc the cache_set, the asyncronous closure handling tries to dereference a cache_set that hadn't yet been allocated inside of cache_set_flush() which is called by __cache_set_unregister() during cleanup. This appears to happen only during an OOM condition on bcache_register. Signed-off-by:
Eric Wheeler <bcache@linux.ewheeler.net> Signed-off-by:
-
Eric Wheeler authored
commit 07cc6ef8 upstream. The bch_writeback_thread might BUG_ON in read_dirty() if dc->sb==BDEV_STATE_DIRTY and bch_sectors_dirty_init has not yet completed its related initialization. This patch downs the dc->writeback_lock until after initialization is complete, thus preventing bch_writeback_thread from proceeding prematurely. See this thread: http://thread.gmane.org/gmane.linux.kernel.bcache.devel/3453Signed-off-by:
Marc MERLIN <marc@merlins.org> Signed-off-by:
Jens Axboe <axboe@fb.com> Signed-off-by:
-
Eric Wheeler authored
commit 9b299728 upstream. Fix null pointer dereference by changing register_cache() to return an int instead of being void. This allows it to return -ENOMEM or -ENODEV and enables upper layers to handle the OOM case without NULL pointer issues. See this thread: http://thread.gmane.org/gmane.linux.kernel.bcache.devel/3521 Fixes this error: gargamel:/sys/block/md5/bcache# echo /dev/sdh2 > /sys/fs/bcache/register bcache: register_cache() error opening sdh2: cannot allocate memory BUG: unable to handle kernel NULL pointer dereference at 00000000000009b8 IP: [<ffffffffc05a7e8d>] cache_set_flush+0x102/0x15c [bcache] PGD 120dff067 PUD 1119a3067 PMD 0 Oops: 0000 [#1] SMP Modules linked in: veth ip6table_filter ip6_tables (...) CPU: 4 PID: 3371 Comm: kworker/4:3 Not tainted 4.4.2-amd64-i915-volpreempt-20160213bc1 #3 Hardware name: System manufacturer System Product Name/P8H67-M PRO, BIOS 3904 04/27/2013 Workqueue: events cache_set_flush [bcache] task: ffff88020d5dc280 ti: ffff88020b6f8000 task.ti: ffff88020b6f8000 RIP: 0010:[<ffffffffc05a7e8d>] [<ffffffffc05a7e8d>] cache_set_flush+0x102/0x15c [bcache] Signed-off-by:
-
Bart Van Assche authored
commit 51093254 upstream. Let the target core check task existence instead of the SRP target driver. Additionally, let the target core check the validity of the task management request instead of the ib_srpt driver. This patch fixes the following kernel crash: BUG: unable to handle kernel NULL pointer dereference at 0000000000000001 IP: [<ffffffffa0565f37>] srpt_handle_new_iu+0x6d7/0x790 [ib_srpt] Oops: 0002 [#1] SMP Call Trace: [<ffffffffa05660ce>] srpt_process_completion+0xde/0x570 [ib_srpt] [<ffffffffa056669f>] srpt_compl_thread+0x13f/0x160 [ib_srpt] [<ffffffff8109726f>] kthread+0xcf/0xe0 [<ffffffff81613cfc>] ret_from_fork+0x7c/0xb0 Signed-off-by:
Bart Van Assche <bart.vanassche@sandisk.com> Fixes: 3e4f5748 ("ib_srpt: Convert TMR path to target_submit_tmr") Tested-by:
Alex Estrin <alex.estrin@intel.com> Reviewed-by:
Christoph Hellwig <hch@lst.de> Cc: Nicholas Bellinger <nab@linux-iscsi.org> Cc: Sagi Grimberg <sagig@mellanox.com> Signed-off-by:
Doug Ledford <dledford@redhat.com> Signed-off-by:
-
Bart Van Assche authored
commit 5e4298be upstream. Avoid that discard requests with size => PAGE_SIZE fail with -EIO. Refuse discard requests if the discard size is not a multiple of the page size. Fixes: 2dbe5495 ("brd: Refuse improperly aligned discard requests") Signed-off-by:
Jan Kara <jack@suse.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Robert Elliot <elliott@hp.com> Signed-off-by:
-
Catalin Marinas authored
commit fdc69e7d upstream. The set_pte_at() function must update the hardware PTE_RDONLY bit depending on the state of the PTE_WRITE and PTE_DIRTY bits of the given entry value. However, it currently only performs this for pte_valid() entries, ignoring PTE_PROT_NONE. The side-effect is that PROT_NONE mappings would not have the PTE_RDONLY bit set. Without CONFIG_ARM64_HW_AFDBM, this is not an issue since such PROT_NONE pages are not accessible anyway. With commit 2f4b829c ("arm64: Add support for hardware updates of the access and dirty pte bits"), the ptep_set_wrprotect() function was re-written to cope with automatic hardware updates of the dirty state. As an optimisation, only PTE_RDONLY is checked to assess the "dirty" status. Since set_pte_at() does not set this bit for PROT_NONE mappings, such pages may be considered "dirty" as a result of ptep_set_wrprotect(). This patch updates the pte_valid() check to pte_present() in set_pte_at(). It also adds PTE_PROT_NONE to the swap entry bits comment. Fixes: 2f4b829c ("arm64: Add support for hardware updates of the access and dirty pte bits") Signed-off-by:
Catalin Marinas <catalin.marinas@arm.com> Reported-by:
Ganapatrao Kulkarni <gkulkarni@caviumnetworks.com> Tested-by:
Ganapatrao Kulkarni <gkulkarni@cavium.com> Signed-off-by:
-
Geert Uytterhoeven authored
commit e0a8604f upstream. pca953x_gpio_set_multiple() divides by 4 to convert from longs to bytes, which assumes a 32-bit platform, and is not correct on 64-bit platforms. Use "sizeof(...)" instead to fix this. Fixes: b4818afe ("gpio: pca953x: Add set_multiple to allow multiple bits to be set in one write.") Signed-off-by:
Geert Uytterhoeven <geert+renesas@glider.be> Acked-by:
Phil Reid <preid@electromag.com.au> Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
-
OGAWA Hirofumi authored
commit c0a2ad9b upstream. On umount path, jbd2_journal_destroy() writes latest transaction ID (->j_tail_sequence) to be used at next mount. The bug is that ->j_tail_sequence is not holding latest transaction ID in some cases. So, at next mount, there is chance to conflict with remaining (not overwritten yet) transactions. mount (id=10) write transaction (id=11) write transaction (id=12) umount (id=10) <= the bug doesn't write latest ID mount (id=10) write transaction (id=11) crash mount [recovery process] transaction (id=11) transaction (id=12) <= valid transaction ID, but old commit must not replay Like above, this bug become the cause of recovery failure, or FS corruption. So why ->j_tail_sequence doesn't point latest ID? Because if checkpoint transactions was reclaimed by memory pressure (i.e. bdev_try_to_free_page()), then ->j_tail_sequence is not updated. (And another case is, __jbd2_journal_clean_checkpoint_list() is called with empty transaction.) So in above cases, ->j_tail_sequence is not pointing latest transaction ID at umount path. Plus, REQ_FLUSH for checkpoint is not done too. So, to fix this problem with minimum changes, this patch updates ->j_tail_sequence, and issue REQ_FLUSH. (With more complex changes, some optimizations would be possible to avoid unnecessary REQ_FLUSH for example though.) BTW, journal->j_tail_sequence = ++journal->j_transaction_sequence; Increment of ->j_transaction_sequence seems to be unnecessary, but ext3 does this. Signed-off-by:
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Signed-off-by:
Theodore Ts'o <tytso@mit.edu> Signed-off-by:
-
Kamal Mostafa authored
commit 50fe6dd1 upstream. Use the local uapi headers to keep in sync with "recently" added #define's (e.g. VSS_OP_REGISTER1). Fixes: 3eb2094c ("Adding makefile for tools/hv") Signed-off-by:
Kamal Mostafa <kamal@canonical.com> Signed-off-by:
K. Y. Srinivasan <kys@microsoft.com> Signed-off-by:
-
Takashi Iwai authored
commit c64c1437 upstream. i915 get_eld ops may return an error when no encoder is connected, and currently we regard the error as fatal and skip the whole ELD handling. This ended up with the missing ELD update at unplugging. This patch fixes the issue by treating the error as the unplugged state, instead of skipping the rest. Reported-by:
Libin Yang <libin.yang@linux.intel.com> Signed-off-by:
Takashi Iwai <tiwai@suse.de> Signed-off-by:
-
Takashi Iwai authored
commit 7169701a upstream. The recent addition of on-demand i915 audio component binding in the codec driver seems leading to the unbalanced i915 power refcount, according to Intel CI tests. Typically, it gets a kernel WARNING like: WARNING: CPU: 3 PID: 173 at sound/hda/hdac_i915.c:91 snd_hdac_display_power+0xf1/0x110 [snd_hda_core]() Call Trace: [<ffffffff813fef15>] dump_stack+0x67/0x92 [<ffffffff81078a21>] warn_slowpath_common+0x81/0xc0 [<ffffffff81078b15>] warn_slowpath_null+0x15/0x20 [<ffffffffa00f77e1>] snd_hdac_display_power+0xf1/0x110 [snd_hda_core] [<ffffffffa015039d>] azx_intel_link_power+0xd/0x10 [snd_hda_intel] [<ffffffffa011e32a>] azx_link_power+0x1a/0x30 [snd_hda_codec] [<ffffffffa00f21f9>] snd_hdac_link_power+0x29/0x40 [snd_hda_core] [<ffffffffa01192a6>] hda_codec_runtime_suspend+0x76/0xa0 [snd_hda_codec] ..... The scenario is like below: - HD-audio driver and i915 driver are probed concurrently at the (almost) same time; HDA bus tries to bind with i915, but it fails because i915 initialization is still being processed. - Later on, HD-audio probes the HDMI codec, where it again tries to bind with i915. At this time, it succeeds. - At finishing the probe of HDA, it decreases the refcount as if it were already bound at the bus probe, since the component is bound now. This triggers a kernel WARNING due to the unbalance. As a workaround, in this patch, we just disable the on-demand i915 component binding in the codec driver. This essentially reverts back to the state of 4.4 kernel. We know that this is no real solution, but it's a minimalistic simple change that can be applied to 4.5.x kernel as stable. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94566Reported-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
Takashi Iwai authored
commit bd481285 upstream. We forgot to copy monitor_present value when updating the ELD information. This won't change the ELD retrieval and the jack notification behavior, but appears only in the proc output. In that sense, it's no fatal error, but a bug is a bug is a bug. Signed-off-by:
-
Takashi Iwai authored
commit 691be973 upstream. The commit [b62232d4: ALSA: hda - Limit i915 HDMI binding only for HSW and later] tried to limit the usage of i915 audio notifier to the recent Intel models and switch to the old method on pre-Haswell models. However, it assumed that the i915 component binding hasn't been done on such models, and the assumption was wrong: namely, Baytrail had already the i915 component binding due to powerwell control. Thus, the workaround wasn't applied to Baytrail. For fixing this properly, this patch introduces a new flag indicating the usage of audio notifier and codec_has_acomp() refers to this flag instead of checking the existence of audio component. Reported-by:
-
Takashi Iwai authored
commit 93a9ff15 upstream. snd_hdac_sync_audio_rate() call is mandatory only for HSW and later models, but we call the function unconditionally blindly assuming that the function doesn't do anything harmful. But since recently, the function checks the validity of the passed pin NID, and eventually spews the warning if an unexpected pin is passed. This is seen on old chips like Baytrail. The fix is to limit the call of this function again only for the chips with the proper binding. This can be identified by the same flag as the eld notifier. Reported-by:
-
Takashi Iwai authored
commit b62232d4 upstream. It turned out that the pre-HSW Intel chips are incompatible with the naive assumption we had -- the fixed mapping between the port and the HD-audio widget. This may result in the bad access, as captured by the recent patch to add a WARN_ON() for the port mapping check. As a quick workaround, disable the i915 audio component binding for all pre-Haswell models. Reported-by:
-
Takashi Iwai authored
commit 1f7c6658 upstream. Cirrus HD-audio driver may adjust GPIO pins for EAPD dynamically depending on the jack plug state. This works fine for the auto-mute mode where the speaker gets muted upon the HP jack plug. OTOH, when the auto-mute mode is off, this turns off the EAPD unexpectedly depending on the jack state, which results in the silent speaker output. This patch fixes the silent speaker output issue by setting GPIO bits constantly when the auto-mute mode is off. Reported-and-tested-by: moosotc@gmail.com Signed-off-by:
-
Aaron Plattner authored
commit 2d369c74 upstream. Vendor ID 0x10de0082 is used by a yet-to-be-named GPU chip. This chip also has the 2-ch audio swapping bug, so patch_nvhdmi is appropriate here. Signed-off-by:
Aaron Plattner <aplattner@nvidia.com> Signed-off-by:
-
Hui Wang authored
commit 6ef2f68f upstream. This Lenovo ThinkCentre AIO also uses Line2 as mic mute button and uses GPIO2 to control the mic mute led, so applying this quirk can make both the button and led work. BugLink: https://bugs.launchpad.net/bugs/1555912Signed-off-by:
Hui Wang <hui.wang@canonical.com> Signed-off-by:
-
Takashi Iwai authored
commit 4f8e4f35 upstream. The current Intel HDMI codec driver supports only three fixed ports from port B to port D. However, i915 driver may assign a DP on other ports, e.g. port A, when no eDP is used. This incompatibility is caught later at pin_nid_to_pin_index() and results in a warning message like "HDMI: pin nid 4 not registered" at each time. This patch filters out such invalid events beforehand, so that the kernel won't be too grumbling. Reported-by:
Stefan Assmann <sassmann@kpanic.de> Signed-off-by:
-
Vittorio Gambaletta (VittGam) authored
commit 4061db03 upstream. The clock measurement on the AC'97 audio card found in the IBM ThinkPad X41 will often fail, so add a quirk entry to fix it. Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=441087Signed-off-by:
Vittorio Gambaletta <linuxbugs@vittgam.net> Signed-off-by:
-
Takashi Iwai authored
commit 0ab1ace8 upstream. The commit [d507941b: ALSA: pcm: Correct PCM BUG error message] made the warning prefix back to "BUG:" due to its previous wrong prefix. But a kernel message containing "BUG:" seems taken as an Oops message wrongly by some brain-dead daemons, and it annoys users in the end. Instead of teaching daemons, change the string again to a more reasonable one. Fixes: 507941beb1e ('ALSA: pcm: Correct PCM BUG error message') Signed-off-by:
-
Takashi Iwai authored
commit 56dc66ff upstream. Just like CX20722, CX7024 codec also requires the power down at reboot in order to reduce the noise at reboot/shutdown. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=113511Signed-off-by:
-
Takashi Iwai authored
commit fc4f000b upstream. HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by:
Jiri Slaby <jslaby@suse.cz> Signed-off-by:
-
Asai Thambi SP authored
commit 008e56d2 upstream. Fail all pending requests after surprise removal of a drive. Signed-off-by:
Vignesh Gunasekaran <vgunasekaran@micron.com> Signed-off-by:
Selvan Mani <smani@micron.com> Signed-off-by:
Asai Thambi S P <asamymuthupa@micron.com> Signed-off-by:
-
Asai Thambi SP authored
commit abb0ccd1 upstream. Added timeout handler. Replaced blk_mq_end_request() with blk_mq_complete_request() to avoid double completion of a request. Signed-off-by:
Rajesh Kumar Sambandam <rsambandam@micron.com> Signed-off-by:
-
Asai Thambi SP authored
commit aae4a033 upstream. Allow device initialization to finish gracefully when it is in FTL rebuild failure state. Also, recover device out of this state after successfully secure erasing it. Signed-off-by:
-
Asai Thambi SP authored
commit 51c6570e upstream. Flush inflight IOs using fsync_bdev() when the device is safely removed. Also, block further IOs in device open function. Signed-off-by:
-
Asai Thambi SP authored
commit 59cf70e2 upstream. When FTL rebuild is in progress, alloc_disk() initializes the disk but device node will be created by add_disk() only after successful completion of FTL rebuild. So, skip deletion of device node in removal path when FTL rebuild is in progress. Signed-off-by:
-
Asai Thambi SP authored
commit 5b7e0a8a upstream. Print exact time when an internal command is interrupted. Signed-off-by:
-
Asai Thambi SP authored
commit e35b9473 upstream. Remove setting and clearing MTIP_PF_EH_ACTIVE_BIT flag in mtip_handle_tfe() as they are redundant. Also avoid waking up service thread from mtip_handle_tfe() because it is already woken up in case of taskfile error. Signed-off-by:
-
Asai Thambi SP authored
commit cfc05bd3 upstream. Service thread does not detect the need for taskfile error hanlding. Fixed the flag condition to process taskfile error. Signed-off-by:
-
Asai Thambi SP authored
commit d8a18d2d upstream. Prevent standby immediate command from being issued in remove, suspend and shutdown paths, while drive is in FTL rebuild process. Signed-off-by:
-
Tiffany Lin authored
commit 7df5ab87 upstream. In v4l2-compliance utility, test QUERYBUF required correct length value to go through each planar to check planar's length in multi-planar buffer type Signed-off-by:
Tiffany Lin <tiffany.lin@mediatek.com> Reviewed-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by:
Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@osg.samsung.com> Signed-off-by:
-
Philipp Zabel authored
commit 74dc385c upstream. During the recent vb2_buffer restructuring, the calculation of the buffer payload reported to userspace was accidentally broken for the first encoded frame, counting only the length of the headers. This patch re-adds the length of the actual frame data. Fixes: 2d700715 ("[media] media: videobuf2: Restructure vb2_buffer") Reported-by:
Michael Olbrich <m.olbrich@pengutronix.de> Signed-off-by:
Philipp Zabel <p.zabel@pengutronix.de> Tested-by:
Jan Luebbe <jlu@pengutronix.de> Signed-off-by:
-
Hans de Goede authored
commit 5c915c68 upstream. On my bttv card "Hauppauge WinTV [card=10]" capturing in YV12 fmt at max size results in a solid green rectangle being captured (all colors 0 in YUV). This turns out to be caused by max-width (924) not being a multiple of 16. We've likely never hit this problem before since normally xawtv / tvtime, etc. will prefer packed pixel formats. But when using a video card which is using xf86-video-modesetting + glamor, only planar XVideo fmts are available, and xawtv will chose a matching capture format to avoid needing to do conversion, triggering the solid green window problem. Signed-off-by:
Hans de Goede <hdegoede@redhat.com> Acked-by:
-
Hans Verkuil authored
commit b339a72e upstream. The V4L2_CID_TX_EDID_PRESENT control reports if an EDID is present. The adv7511 however still reported the EDID present after disconnecting the HDMI cable. Fix the logic regarding this control. And when the EDID is disconnected also call ADV7511_EDID_DETECT to notify the bridge driver. This was also missing. Signed-off-by:
-
