- 06 Feb, 2014 40 commits
-
-
Gregory CLEMENT authored
commit 6cf70ae9 upstream. The first variants of Armada XP SoCs (A0 stepping) have issues related to the i2c controller which prevent to use the offload mechanism and lead to a kernel hang during boot. The commit introduces a new the compatible string marvell,mv78230-a0-i2c for the i2c controller. When this compatible string is used the driver disables the offload mechanism and the kernel no more hangs on these SoCs. Signed-off-by:
Gregory CLEMENT <gregory.clement@free-electrons.com> Reported-by:
Andrew Lunn <andrew@lunn.ch> Acked-by:
Wolfram Sang <wsa@the-dreams.de> Acked-by:
Arnd Bergmann <arnd@arndb.de> Fixes: 930ab3d4 (i2c: mv64xxx: Add I2C Transaction Generator support) Signed-off-by:
Jason Cooper <jason@lakedaemon.net> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Gregory CLEMENT authored
commit 85e618a1 upstream. The first variants of Armada XP SoCs (A0 stepping) have issues related to the i2c controller which prevent to use the offload mechanism and lead to a kernel hang during boot. This commit add quirk in the mvebu platform code to check the SoC version and then update the compatible string for the i2c controller according to the revision of the SoC. Currently only some OpenBlocks AX3-4 boards are known to use an A0 revision so the check is done only for these boards. Signed-off-by:
-
Gregory CLEMENT authored
commit af8d1c63 upstream. All the mvebu SoCs have information related to their variant and revision that can be read from the PCI control register. This patch adds support for Armada XP and Armada 370. This reading of the revision and the ID are done before the PCI initialization to avoid any conflicts. Once these data are retrieved, the resources are freed to let the PCI subsystem use it. Fixes: 930ab3d4 (i2c: mv64xxx: Add I2C Transaction Generator support) Signed-off-by:
-
Takashi Iwai authored
commit b0ad4ff3 upstream. The DriveGuard chips on the new HP laptops are with a new PnP ID "HPQ6007". It should be compatible with older chips. Acked-by:
Éric Piel <eric.piel@tremplin-utc.net> Signed-off-by:
Takashi Iwai <tiwai@suse.de> Signed-off-by:
Matthew Garrett <matthew.garrett@nebula.com> Signed-off-by:
-
Minchan Kim authored
commit da4a0412 upstream. Dan and Sergey reported that there is a racy between reset and flushing of pending work so that it could make oops by freeing zram->meta in reset while zram_slot_free can access zram->meta if new request is adding during the race window. This patch moves flush after taking init_lock so it prevents new request so that it closes the race. Signed-off-by:
Minchan Kim <minchan@kernel.org> Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchan@redhat.com> Tested-by:
Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
Kent Overstreet authored
commit ef71ec00 upstream. The code that handles overlapping extents that we've just read back in from disk was depending on the behaviour of the code that handles overlapping extents as we're inserting into a btree node in the case of an insert that forced an existing extent to be split: on insert, if we had to split we'd also insert a new extent to represent the top part of the old extent - and then that new extent would get written out. The code that read the extents back in thus not bother with splitting extents - if it saw an extent that ovelapped in the middle of an older extent, it would trim the old extent to only represent the bottom part, assuming that the original insert would've inserted a new extent to represent the top part. I still haven't figured out _how_ it can happen, but I'm now pretty convinced (and testing has confirmed) that there's some kind of an obscure corner case (probably involving extent merging, and multiple overwrites in different sets) that breaks this. The fix is to change the mergesort fixup code to split extents itself when required. Signed-off-by:
Kent Overstreet <kmo@daterainc.com> Signed-off-by:
-
Eric W. Biederman authored
commit 260a459d upstream. A bug was introduced with the is_mounted helper function in commit f7a99c5b Author: Al Viro <viro@zeniv.linux.org.uk> Date: Sat Jun 9 00:59:08 2012 -0400 get rid of ->mnt_longterm it's enough to set ->mnt_ns of internal vfsmounts to something distinct from all struct mnt_namespace out there; then we can just use the check for ->mnt_ns != NULL in the fast path of mntput_no_expire() Signed-off-by:
Al Viro <viro@zeniv.linux.org.uk> The intent was to test if the real_mount(vfsmount)->mnt_ns was NULL_OR_ERR but the code is actually testing real_mount(vfsmount) and always returning true. The result is d_absolute_path returning paths it should be hiding. Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Eric W. Biederman authored
commit a8323da0 upstream. In commit 232d2d60 Author: Waiman Long <Waiman.Long@hp.com> Date: Mon Sep 9 12:18:13 2013 -0400 dcache: Translating dentry into pathname without taking rename_lock The __dentry_path locking was changed and the variable error was intended to be moved outside of the loop. Unfortunately the inner declaration of error was not removed. Resulting in a version of __dentry_path that will never return an error. Remove the problematic inner declaration of error and allow __dentry_path to return errors once again. Cc: Waiman Long <Waiman.Long@hp.com> Signed-off-by:
-
Theodore Ts'o authored
commit 09c455aa upstream. A missing cast means that when we are truncating a file which is less than 60 bytes, we don't clear the correct area of memory, and in fact we can end up truncating the next inode in the inode table, or worse yet, some other kernel data structure. Addresses-Coverity-Id: #751987 Signed-off-by:
"Theodore Ts'o" <tytso@mit.edu> Signed-off-by:
-
Tejun Heo authored
commit ecd75ad5 upstream. For some reason, some early WD drives spin up and down drives erratically when the link is put into slumber mode which can reduce the life expectancy of the device significantly. Unfortunately, we don't have full list of devices and given the nature of the issue it'd be better to err on the side of false positives than the other way around. Let's disable LPM on all WD devices which match one of the known problematic model prefixes and are SATA-I. As horkage list doesn't support matching SATA capabilities, this is implemented as two horkages - WD_BROKEN_LPM and NOLPM. The former is set for the known prefixes and sets the latter if the matched device is SATA-I. Note that this isn't optimal as this disables all LPM operations and partial link power state reportedly works fine on these; however, the way LPM is implemented in libata makes it difficult to precisely map libata LPM setting to specific link power state. Well, these devices are already fairly outdated. Let's just disable whole LPM for now. Signed-off-by:
Tejun Heo <tj@kernel.org> Reported-and-tested-by:
Nikos Barkas <levelwol@gmail.com> Reported-and-tested-by:
Ioannis Barkas <risc4all@yahoo.com> References: https://bugzilla.kernel.org/show_bug.cgi?id=57211Signed-off-by:
-
Simon Guinot authored
commit a96cc303 upstream. This patch updates the Armada 370/XP SATA node with the new compatible string "marvell,armada-370-sata". Signed-off-by:
Simon Guinot <simon.guinot@sequanux.org> Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Cc: Jason Cooper <jason@lakedaemon.net> Cc: Andrew Lunn <andrew@lunn.ch> Cc: Gregory Clement <gregory.clement@free-electrons.com> Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> Cc: Lior Amsalem <alior@marvell.com> Acked-by:
-
Lior Amsalem authored
commit 9013d64e upstream. On Armada 370/XP SoCs, once a disk is removed from a SATA port, then the re-plug events are not detected by the sata_mv driver. This patch fixes the issue by updating the PHY speed in the LP_PHY_CTL register (0x58) according to the SControl speed. Note that this fix is only applied if the compatible string "marvell,armada-370-sata" is found in the SATA DT node. Fixes: 9ae6f740 ("arm: mach-mvebu: add support for Armada 370 and Armada XP with DT") Signed-off-by:
Lior Amsalem <alior@marvell.com> Signed-off-by:
Nadav Haklai <nadavh@marvell.com> Signed-off-by:
-
Simon Guinot authored
commit b1f5c73b upstream. The sata_mv driver supports the SATA IP found in several Marvell SoCs. As some new SATA registers have been introduced with the Armada 370/XP SoCs, a way to identify them is needed. This patch introduces a new compatible string for the SATA IP found in Armada 370/XP SoCs. Signed-off-by:
-
Peter Huewe authored
commit 747d35bd upstream. Depending on the implementation strcmp might return the difference between two strings not only -1,0,1 consequently if (strcmp (a,b) == -1) might lead to taking the wrong branch -> compare with < 0 instead, which in any case is more canonical. Signed-off-by:
Peter Huewe <peterhuewe@gmx.de> Signed-off-by:
-
Peter Huewe authored
commit 85c5e0d4 upstream. The 'get_burstcount' function can in some circumstances 'return -EBUSY' which in tpm_stm_i2c_send is stored in an 'u32 burstcnt' thus converting the signed value into an unsigned value, resulting in 'burstcnt' being huge. Changing the type to u32 only does not solve the problem as the signed value is converted to an unsigned in I2C_WRITE_DATA, resulting in the same effect. Thus -> Change type of burstcnt to u32 (the return type of get_burstcount) -> Add a check for the return value of 'get_burstcount' and propagate a potential error. This makes also sense in the 'I2C_READ_DATA' case, where the there is no signed/unsigned conversion. found by coverity Signed-off-by:
-
Adrien Vergé authored
commit e7729a41 upstream. Similarly to other Apple products, MBA 1,1 needs a specific quirk. Pin 0x18 must be set to VREF_50 to have sound output. This was no longer done since commit 1a97b7f2, resulting in a mute built-in speaker. This patch corrects the regression by creating a fixup for the MBA 1,1. Fixes: 1a97b7f2 ("ALSA: hda/realtek - Remove the last static quirks for ALC882") Tested-by:
Adrien Vergé <adrienverge@gmail.com> Signed-off-by:
-
Dan Carpenter authored
commit 4c3773ed upstream. The test here is intended intended to prevent shift wrapping bugs when we do "1U << idx2". We should consider the number of bits in a u32 instead of the number of bytes. [fix another chunk similarly by tiwai] Fixes: 7bb2491b ('ALSA: Add kconfig to specify the max card numbers') Signed-off-by:
-
Takashi Iwai authored
commit 80ab8eae upstream. The PCI devices with DMA masks smaller than 32bit should enable CONFIG_ZONE_DMA. Since the recent change of page allocator, page allocations via dma_alloc_coherent() with the limited DMA mask bits may fail more frequently, ended up with no available buffers, when CONFIG_ZONE_DMA isn't enabled. With CONFIG_ZONE_DMA, the system has much more chance to obtain such pages. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=68221Signed-off-by:
-
Takashi Iwai authored
commit 43a8e50a upstream. AD1986A mic pins (0x1d and 0x1f) share the same widget for controlling the loopback volume/mute, but the generic parser didn't check it. This ended up with the duplicated controls for the same effect. This patch adds the check of the duplication for avoiding it. After this fix, there will be only one control although it affects both paths; this remaining issue should be fixed later in a different patch. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=66621Signed-off-by:
-
Takashi Iwai authored
commit ed0e0d06 upstream. The 3stack pin configs for AD1986A codec had incorrect values that resulted in broken mic and line-in. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=66621Signed-off-by:
-
Takashi Iwai authored
commit 770bd4bf upstream. The lack of comma leads to the wrong channel for an SPDIF channel. Unfortunately this wasn't caught by compiler because it's still a valid expression. Reported-by:
Alexander Aristov <aristov.alexander@gmail.com> Signed-off-by:
-
Charles Keepax authored
commit 34354792 upstream. Latest evaluation of the the device has given some patch file additions for improved performance. Signed-off-by:
Charles Keepax <ckeepax@opensource.wolfsonmicro.com> Signed-off-by:
Mark Brown <broonie@linaro.org> Signed-off-by:
-
Lars-Peter Clausen authored
commit e20970ad upstream. The driver defines ADAU1701_SEROCTL_WORD_LEN_16 as 0x10 while it should be b10, so 0x2. This patch fixes it. Reported-by:
Magnus Reftel <magnus.reftel@lockless.no> Signed-off-by:
Lars-Peter Clausen <lars@metafoo.de> Signed-off-by:
-
Krzysztof Kozlowski authored
commit 74142ffc upstream. The regmap used by max77686 MFD driver was not freed with regmap_exit() on driver exit. This lead to leak of resources. Replace regmap_init_i2c() call in driver probe with initialization of managed register map so the regmap will be properly freed by the device management code. Signed-off-by:
Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by:
Lee Jones <lee.jones@linaro.org> Signed-off-by:
-
Dongsheng Yang authored
commit ad85ace0 upstream. Currently, if we use perf kvm --guestkallsyms --guestmodules report, we can not get the perf information from perf data file. All sample are shown as unknown. Reproducing steps: # perf kvm --guestkallsyms /tmp/kallsyms --guestmodules /tmp/modules record -a sleep 1 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.624 MB perf.data.guest (~27260 samples) ] # perf kvm --guestkallsyms /tmp/kallsyms --guestmodules /tmp/modules report |grep % 100.00% [guest/6471] [unknown] [g] 0xffffffff8164f330 This bug was introduced by 207b5792 (perf kvm: Fix regression with guest machine creation). In original code, it uses perf_session__find_machine(), it means we deliver symbol to machine which has the same pid, if no machine found, deliver it to *default* guest. But if we use perf_session__findnew_machine() here, if no machine was found, new machine with pid will be built and added. Then the default guest which with pid == 0 will never get a symbol. And because the new machine initialized here has no kernel map created, the symbol delivered to it will be marked as "unknown". This patch here is to revert commit 207b5792 and fix the SEGFAULT bug in another way. Verification steps: # ./perf kvm --guestkallsyms /home/kallsyms --guestmodules /home/modules record -a sleep 1 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.651 MB perf.data.guest (~28437 samples) ] # ./perf kvm --guestkallsyms /home/kallsyms --guestmodules /home/modules report |grep % 22.64% :6471 [guest.kernel.kallsyms] [g] update_rq_clock.part.70 19.99% :6471 [guest.kernel.kallsyms] [g] d_free 18.46% :6471 [guest.kernel.kallsyms] [g] bio_phys_segments 16.25% :6471 [guest.kernel.kallsyms] [g] dequeue_task 12.78% :6471 [guest.kernel.kallsyms] [g] __switch_to 7.91% :6471 [guest.kernel.kallsyms] [g] scheduler_tick 1.75% :6471 [guest.kernel.kallsyms] [g] native_apic_mem_write 0.21% :6471 [guest.kernel.kallsyms] [g] apic_timer_interrupt Signed-off-by:
Dongsheng Yang <yangds.fnst@cn.fujitsu.com> Acked-by:
David Ahern <dsahern@gmail.com> Cc: David Ahern <dsahern@gmail.com> Link: http://lkml.kernel.org/r/1387564907-3045-1-git-send-email-yangds.fnst@cn.fujitsu.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by:
-
Chen-Yu Tsai authored
commit fa8cf57c upstream. Some GPIO users, such as fixed-regulator, request GPIO output with initial value of 1. This was ignored by sunxi driver. Signed-off-by:
Chen-Yu Tsai <wens@csie.org> Acked-by:
Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
-
Stephen Warren authored
commit 75ea799d upstream. The current MAX8907 driver has two issues related to weekday value handling: 1) The HW WEEKDAY register has range 0..6 rather than 1..7 as documented. Note that I validated the actual HW range by observing the HW register roll from 6->0 rather than 6->7->1 as would otherwise be expected. This matches Linux's tm_wday range of 0..6. When the CMOS RAM content is lost, the date returned from the device is 2007-01-01 00:00:00, which is a Monday. The WEEKDAY register reads 1 in this case. This matches the numbering in Linux's tm_wday field. Hence we should write Linux's tm_wday value to the register without modifying it. Hence, remove the +1/-1 calculations for WEEKDAY/tm_wday. 2) There's no need to make alarms match on the WEEKDAY register, since the other fields together uniquely define the alarm date/time. Ignoring the WEEKDAY value in the match isolates the driver from any incorrect value in the current time copy of the WEEKDAY register. Each change individually, or both together, solves an issue that I observed; "hwclock -r" would time out waiting for its alarm to fire if the CMOS RAM content had been lost, and hence the WEEKDAY register value mismatched what the driver expected it to be. "hwclock -w" would solve this by over-writing the HW default WEEKDAY register value with what the driver expected. Signed-off-by:
Stephen Warren <swarren@nvidia.com> Signed-off-by:
-
Sebastian Andrzej Siewior authored
commit d6a48452 upstream. In commit 85747f ("PATCH] parport: add NetMOS 9805 support") Max added the PCI ID for NetMOS 9805 based on a Debian bug report from 2k4 which was at the v2.4.26 time frame. The patch made into 2.6.14. Shortly before that patch akpm merged commit 296d3c78 ("[PATCH] Support NetMOS based PCI cards providing serial and parallel ports") which made into v2.6.9-rc1. Now we have two different entries for the same PCI id. I have here the NetMos 9805 which claims to support SPP/EPP/ECP mode. This patch takes Max's entry for titan_1284p1 (base != -1 specifies the ioport for ECP mode) and replaces akpm's entry for netmos_9805 which specified -1 (=none). Both share the same PCI-ID (my card has subsystem 0x1000 / 0x0020 so it should match PCI_ANY). While here I also drop the entry for titan_1284p2 which is the same as netmos_9815. Cc: Maximilian Attems <maks@stro.at> Signed-off-by:
Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by:
-
Heiko Carstens authored
commit 4e078146 upstream. With b8668fd0 "s390/uapi: change struct statfs[64] member types to unsigned values" the size of a couple of struct statfs64 member got incorrectly changed from 64 to 32 bit for 32 bit builds. Fix this by changing the type of couple of struct statfs64 members from unsigned long to unsigned long long. The definition of struct compat_statfs64 was correct however. Signed-off-by:
Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by:
-
Dominik Dingel authored
commit ff1f3cb4 upstream. The diagnose 500 subcode 3 contains the 32 bit subchannel id in bits 32-63 (counting from the left). As for other I/O instructions, bits 0-31 should be ignored and thus not be passed to kvm_io_bus_write_cookie(). This fixes a bug where the guest passed non-zero bits 0-31 which the host tried to interpret, leading to ioeventfd notification failures. Signed-off-by:
Dominik Dingel <dingel@linux.vnet.ibm.com> Reviewed-by:
Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
commit 743db27c upstream. The diagnose code to be used is the contents of the base register (if not zero), plus the displacement. The current code ignores the base register contents. So let's fix that... Reviewed-by:
-
Stephen Warren authored
commit 3685f19e upstream. Tegra chips have 4 or 5 identical UART modules embedded. UARTs C..E have their MODEM-control signals tied off to a static state. However UARTs A and B can optionally route those signals to/from package pins, depending on the exact pinmux configuration. When these signals are not routed to package pins, false interrupts may trigger either temporarily, or permanently, all while not showing up in the IIR; it will read as NO_INT. This will eventually lead to the UART IRQ being disabled due to unhandled interrupts. When this happens, the kernel may print e.g.: irq 68: nobody cared (try booting with the "irqpoll" option) In order to prevent this, enable UART_BUG_NOMSR. This prevents UART_IER_MSI from being enabled, which prevents the false interrupts from triggering. In practice, this is not needed under any of the following conditions: * On Tegra chips after Tegra30, since the HW bug has apparently been fixed. * On UARTs C..E since their MODEM control signals are tied to the correct static state which doesn't trigger the issue. * On UARTs A..B if the MODEM control signals are routed out to package pins, since they will then carry valid signals. However, we ignore these exceptions for now, since they are only relevant if a board actually hooks up more than a 4-wire UART, and no currently supported board does this. If we ever support a board that does, we can refine the algorithm that enables UART_BUG_NOMSR to take those exceptions into account, and/or read a flag from DT/... that indicates that the board has hooked up and pinmux'd more than a 4-wire UART. Reported-by: Olof Johansson <olof@lixom.net> # autotester Signed-off-by:
-
Jonathan Woithe authored
commit 9c5320f8 upstream. Fix the initialisation of older Quatech serial cards which are fitted with the AMCC PCI Matchmaker interface chip. Signed-off-by: Jonathan Woithe (jwoithe@just42.net) Signed-off-by:
-
Yegor Yefremov authored
commit 48c0247d upstream. Signed-off-by:
Yegor Yefremov <yegorslists@googlemail.com> Signed-off-by:
-
Sarah Sharp authored
commit f2d9b991 upstream. Commit 35773dac "usb: xhci: Link TRB must not occur within a USB payload burst" attempted to fix an issue found with USB ethernet adapters, and inadvertently broke USB storage devices. The patch attempts to ensure that transfers never span a segment, and rejects transfers that have more than 63 entries (or possibly less, if some entries cross 64KB boundaries). usb-storage limits the maximum transfer size to 120K, and we had assumed the block layer would pass a scatter-gather list of 4K entries, resulting in no more than 31 sglist entries: http://marc.info/?l=linux-usb&m=138498190419312&w=2 That assumption was wrong, since we've seen the driver reject a write that was 218 sectors long (of probably 512 bytes each): Jan 1 07:04:49 jidanni5 kernel: [ 559.624704] xhci_hcd 0000:00:14.0: Too many fragments 79, max 63 ... Jan 1 07:04:58 jidanni5 kernel: [ 568.622583] Write(10): 2a 00 00 06 85 0e 00 00 da 00 Limit the number of scatter-gather entries to half a ring segment. That should be margin enough in case some entries cross 64KB boundaries. Increase the number of TRBs per segment from 64 to 256, which should result in ring segments fitting on a 4K page. Signed-off-by:
Sarah Sharp <sarah.a.sharp@linux.intel.com> Reported-by: jidanni@jidanni.org References: http://bugs.debian.org/733907 Fixes: 35773dac ('usb: xhci: Link TRB must not occur within a USB payload burst') Signed-off-by:
-
Ben Hutchings authored
commit d6c9ea90 upstream. Currently prepare_ring() returns -ENOMEM if the urb won't fit into a single ring segment. usb_sg_wait() treats this error as a temporary condition and will keep retrying until something else goes wrong. The number of retries should be limited in usb_sg_wait(), but also prepare_ring() should not return an error code that suggests it might be worth retrying. Change it to -EINVAL. Reported-by: jidanni@jidanni.org References: http://bugs.debian.org/733907 Fixes: 35773dac ('usb: xhci: Link TRB must not occur within a USB payload burst') Signed-off-by:
Ben Hutchings <ben@decadent.org.uk> Signed-off-by:
-
Phil Pokorny authored
commit d303b1b5 upstream. Add new PCI ID to support new model "Kaveri" family. Signed-off-by:
Philip Pokorny <ppokorny@penguincomputing.com> Signed-off-by:
Guenter Roeck <linux@roeck-us.net> Signed-off-by:
-
Jean-Jacques Hiblot authored
commit 1588c51c upstream. There was a copy/paste error when reading the nwe_pulse value. Signed-off-by:
Jean-Jacques Hiblot <jjhiblot@traphandler.com> Acked-by:
Boris BREZILLON <b.brezillon@overkiz.com> Signed-off-by:
Nicolas Ferre <nicolas.ferre@atmel.com> Signed-off-by:
Kevin Hilman <khilman@linaro.org> Signed-off-by:
-
Ludovic Desroches authored
commit 0645b93f upstream. pinctrl-names property was missing from mmc nodes. Signed-off-by:
Ludovic Desroches <ludovic.desroches@atmel.com> Acked-by:
Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com> Signed-off-by:
-
Marek Roszko authored
commit 8bc661bf upstream. The uart timer will schedule a tasklet when it fires. It is possible that it can fire inside _shutdown before it is killed in the dma and pdc cleanup routines. This causes a tasklet that exists after the port is shutdown, so when the kernel finally executes it, it panics as the tty port is NULL. This is a somewhat rare condition but its possible if a program keeps on opening/closing the port. It has been observed in particular with systemd boot messages that were causing a kernel panic because of this behavior. Moving the timer deletion to the beginning of the function stops a tasklet from being scheduled unexpectedly. Signed-off-by:
Marek Roszko <mark.roszko@gmail.com> [nicolas.ferre@atmel.com: modify commit message, call setup_timer() in any case] Signed-off-by:
-
