- 27 Jan, 2020 40 commits
-
-
Dan Carpenter authored
[ Upstream commit 9e053523 ] The etnaviv_gem_prime_get_sg_table() is supposed to return error pointers. Otherwise it can lead to a NULL dereference when it's called from drm_gem_map_dma_buf(). Fixes: 5f4a4a73 ("drm/etnaviv: fix gem_prime_get_sg_table to return new SG table") Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by:
Christian Gmeiner <christian.gmeiner@gmail.com> Signed-off-by:
Lucas Stach <l.stach@pengutronix.de> Signed-off-by:
Sasha Levin <sashal@kernel.org>
-
Magnus Karlsson authored
[ Upstream commit e6762c8b ] All the setup code in AF_XDP is protected by a mutex with the exception of the mmap code that cannot use it. To make sure that a process banging on the mmap call at the same time as another process is setting up the socket, smp_wmb() calls were added in the umem registration code and the queue creation code, so that the published structures that xsk_mmap needs would be consistent. However, the corresponding smp_rmb() calls were not added to the xsk_mmap code. This patch adds these calls. Fixes: 37b07693 ("xsk: add missing write- and data-dependency barrier") Fixes: c0c77d8f ("xsk: add user memory registration support sockopt") Signed-off-by:
Magnus Karlsson <magnus.karlsson@intel.com> Signed-off-by:
Alexei Starovoitov <ast@kernel.org> Signed-off-by:
-
Nicholas Mc Guire authored
[ Upstream commit 42c7c6ef ] devm_kasprintf() may return NULL if internal allocation failed so this assignment is not safe. Moved the error exit path and added the !NULL which then allows the devres manager to take care of cleanup. Signed-off-by:
Nicholas Mc Guire <hofrat@osadl.org> Fixes: cd2315d4 ("ipmi: kcs_bmc: don't change device name") Signed-off-by:
Corey Minyard <cminyard@mvista.com> Reviewed-by:
Haiyue Wang <haiyue.wang@linux.intel.com> Signed-off-by:
-
Steve Wise authored
[ Upstream commit c8a7eb55 ] When IPv6 support was added, the correct tos was not passed to cxgb_find_route6(). This potentially results in the wrong route entry. Fixes: 830662f6 ("RDMA/cxgb4: Add support for active and passive open connection with IPv6 address") Signed-off-by:
Steve Wise <swise@opengridcomputing.com> Signed-off-by:
Jason Gunthorpe <jgg@mellanox.com> Signed-off-by:
-
Steve Wise authored
[ Upstream commit cb3ba0bd ] import_ep() is passed the correct tos, but doesn't use it correctly. Fixes: ac8e4c69 ("cxgb4/iw_cxgb4: TOS support") Signed-off-by:
-
YueHaibing authored
[ Upstream commit 8e71fa5e ] Fixes gcc '-Wunused-but-set-variable' warning: drivers/video/fbdev/chipsfb.c: In function 'chipsfb_pci_init': drivers/video/fbdev/chipsfb.c:352:22: warning: variable 'size' set but not used [-Wunused-but-set-variable] Fixes: 8c870933 ("[PATCH] ppc32: Remove CONFIG_PMAC_PBOOK"). Signed-off-by:
YueHaibing <yuehaibing@huawei.com> Acked-by:
Michael Ellerman <mpe@ellerman.id.au> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Christophe Leroy <christophe.leroy@c-s.fr> [b.zolnierkie: minor commit summary and description fixups] Signed-off-by:
Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by:
-
Colin Ian King authored
[ Upstream commit e4228088 ] Shifting a u8 by 24 will cause the value to be promoted to an integer. If the top bit of the u8 is set then the following conversion to an unsigned long will sign extend the value causing the upper 32 bits to be set in the result. Fix this by casting the u8 value to an unsigned long before the shift. Detected by CoverityScan, CID#1309693 ("Unintended sign extension") Fixes: 9a9a54ad ("drivers/rtc: add support for Qualcomm PMIC8xxx RTC") Signed-off-by:
Colin Ian King <colin.king@canonical.com> Signed-off-by:
Alexandre Belloni <alexandre.belloni@bootlin.com> Signed-off-by:
-
Colin Ian King authored
[ Upstream commit fb0b3225 ] Shifting a u8 by 24 will cause the value to be promoted to an integer. If the top bit of the u8 is set then the following conversion to an unsigned long will sign extend the value causing the upper 32 bits to be set in the result. Fix this by casting the u8 value to an unsigned long before the shift. Detected by CoverityScan, CID#714646-714649 ("Unintended sign extension") Fixes: 2985c29c ("rtc: Add rtc support to 88PM80X PMIC") Signed-off-by:
-
Colin Ian King authored
[ Upstream commit dc9e4716 ] Shifting a u8 by 24 will cause the value to be promoted to an integer. If the top bit of the u8 is set then the following conversion to an unsigned long will sign extend the value causing the upper 32 bits to be set in the result. Fix this by casting the u8 value to an unsigned long before the shift. Detected by CoverityScan, CID#144925-144928 ("Unintended sign extension") Fixes: 008b3040 ("mfd: Add rtc support to 88pm860x") Signed-off-by:
-
Karsten Graul authored
[ Upstream commit 232dc8ef ] Commit ed75986f ("net/smc: ipv6 support for smc_diag.c") changed the value of the diag_family field. The idea was to indicate the family of the IP address in the inet_diag_sockid field. But the change makes it impossible to distinguish an inet_sock_diag response message from SMC sock_diag response. This patch restores the original behaviour and sends AF_SMC as value of the diag_family field. Fixes: ed75986f ("net/smc: ipv6 support for smc_diag.c") Reported-by:
Eugene Syromiatnikov <esyr@redhat.com> Signed-off-by:
Karsten Graul <kgraul@linux.ibm.com> Signed-off-by:
Ursula Braun <ubraun@linux.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Uwe Kleine-König authored
[ Upstream commit 3f929cad ] When the EXTENSION.WADA bit is set, register 0x19 contains a bitmap of week days, not a day of month. As Linux only handles a single alarm without repetition using day of month is more flexible, so clear this bit. (Otherwise a value depending on time.tm_wday would have to be written to register 0x19.) Also optimize setting the AIE bit to use a single register write instead of a bulk write of three registers. Fixes: ee0981be ("rtc: ds1307: Add support for Epson RX8130CE") Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by:
-
Moritz Fischer authored
[ Upstream commit 8f289805 ] Fix fixed_phy not checking GPIO if no link_update callback is registered. In the original version all users registered a link_update callback so the issue was masked. Fixes: a5597008 ("phy: fixed_phy: Add gpio to determine link up/down.") Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
Moritz Fischer <mdf@kernel.org> Signed-off-by:
-
Rakesh Pillai authored
[ Upstream commit 6e8a8991 ] The management frames transmitted are dma mapped with direction TO_DEVICE, but incorrectly mapped with direction FROM_DEVICE during tx complete and error cases. Fix the direction of dma during dma unmap of the transmitted management frames. Tested HW: WCN3990 Tested FW: WLAN.HL.2.0-01188-QCAHLSWMTPLZ-1 Fixes: 38a1390e ("ath10k: dma unmap mgmt tx buffer if wmi cmd send fails") Signed-off-by:
Rakesh Pillai <pillair@codeaurora.org> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org> Signed-off-by:
-
Niklas Cassel authored
[ Upstream commit e4f045ef ] The apcs node has #clock-cells = <0>, which means that those who references it should specify 0 arguments. The apcs reference in the cpu node incorrectly specifies an argument, remove this bogus argument. Fixes: 65afdf45 ("arm64: dts: qcom: msm8916: Add CPU frequency scaling support") Signed-off-by:
Niklas Cassel <niklas.cassel@linaro.org> Reviewed-by:
Bjorn Andersson <bjorn.andersson@linaro.org> Reviewed-by:
Amit Kucheria <amit.kucheria@linaro.org> Signed-off-by:
Andy Gross <andy.gross@linaro.org> Signed-off-by:
-
Michael Kao authored
[ Upstream commit eb9aecd9 ] The index of msr and adcpnp should match the sensor which belongs to the selected bank in the for loop. Fixes: b7cf0053 ("thermal: Add Mediatek thermal driver for mt2701.") Signed-off-by:
Michael Kao <michael.kao@mediatek.com> Signed-off-by:
Eduardo Valentin <edubezval@gmail.com> Signed-off-by:
-
Colin Ian King authored
[ Upstream commit f0c04c27 ] Shifting a u8 by 24 will cause the value to be promoted to an integer. If the top bit of the u8 is set then the following conversion to an unsigned long will sign extend the value causing the upper 32 bits to be set in the result. Fix this by casting the u8 value to an unsigned long before the shift. Detected by CoverityScan, CID#138801 ("Unintended sign extension") Fixes: edf1aaa3 ("[PATCH] RTC subsystem: DS1672 driver") Signed-off-by:
-
Paul Cercueil authored
[ Upstream commit b7e29924 ] The UDC clock is gated when the bit is cleared, not when it is set. Signed-off-by:
Paul Cercueil <paul@crapouillou.net> Tested-by:
Artur Rojek <contact@artur-rojek.eu> Fixes: 2b555a4b ("clk: ingenic: Add missing flag for UDC clock") Signed-off-by:
Stephen Boyd <sboyd@kernel.org> Signed-off-by:
-
Colin Ian King authored
[ Upstream commit 5ae89078 ] Currently the call to cdev_add is missing a check for failure. Fix this by checking for failure and exiting via a new error path that ensures the allocated comp_channel struct is kfree'd. Detected by CoverityScan, CID#1462359 ("Unchecked return value") Fixes: 9bc79bbc ("Staging: most: add MOST driver's aim-cdev module") Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by:
-
Sara Sharon authored
[ Upstream commit 608dce95 ] The hash mask is a bitmap, so we should use BIT() on the enum values. Signed-off-by:
Sara Sharon <sara.sharon@intel.com> Fixes: 43413a97 ("iwlwifi: mvm: support rss queues configuration command") Signed-off-by:
Luca Coelho <luciano.coelho@intel.com> Signed-off-by:
-
Oleksandr Andrushchenko authored
[ Upstream commit 24ded292 ] When GEM backing storage is allocated those are normal pages, so there is no point using pgprot_writecombine while mmaping. This fixes mismatch of buffer pages' memory attributes between the frontend and backend which may cause screen artifacts. Fixes: c575b7ee ("drm/xen-front: Add support for Xen PV display frontend") Signed-off-by:
Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com> Suggested-by:
Julien Grall <julien.grall@arm.com> Acked-by:
-
Vladimir Zapolskiy authored
[ Upstream commit dc141b99 ] The fixed voltage regulator on Phytec phyCORE-LPC3250 board, which supplies SD/MMC card's power, has a constant output voltage level of either 3.15V or 3.3V, the actual value depends on JP4 position, the power rail is referenced as VCC_SDIO in the board hardware manual. Fixes: d06670e9 ("arm: dts: phy3250: add SD fixed regulator") Signed-off-by:
Vladimir Zapolskiy <vz@mleia.com> Signed-off-by:
-
Vladimir Zapolskiy authored
[ Upstream commit 30fc01ba ] The originally added ARM PrimeCell PL111 clocks property misses the required "clcdclk" clock, which is the same as a clock to enable the LCD controller on NXP LPC3230 and NXP LPC3250 SoCs. Fixes: 93898eb7 ("arm: dts: lpc32xx: add clock properties to device nodes") Signed-off-by:
-
Vladimir Zapolskiy authored
[ Upstream commit 7a0790a4 ] ARM PrimeCell PL111 LCD controller is found on On NXP LPC3230 and LPC3250 SoCs variants, the original reference in compatible property to an older one ARM PrimeCell PL110 is invalid. Fixes: e04920d9 ("ARM: LPC32xx: DTS files for device tree conversion") Signed-off-by:
-
Vladimir Zapolskiy authored
[ Upstream commit 489261c4 ] After switching to a new interrupt controller scheme by separating SIC1 and SIC2 from MIC interrupt controller just one SoC keypad controller was not taken into account, fix it now: WARNING: CPU: 0 PID: 1 at kernel/irq/irqdomain.c:524 irq_domain_associate+0x50/0x1b0 error: hwirq 0x36 is too large for interrupt-controller@40008000 ... lpc32xx_keys 40050000.key: failed to get platform irq lpc32xx_keys: probe of 40050000.key failed with error -22 Fixes: 9b8ad3fb ("ARM: dts: lpc32xx: reparent SIC1 and SIC2 interrupts from MIC") Signed-off-by:
-
Vladimir Zapolskiy authored
[ Upstream commit 3e88bc38 ] NXP LPC32xx keypad controller requires a clock property to be defined. The change fixes the driver initialization problem: lpc32xx_keys 40050000.key: failed to get clock lpc32xx_keys: probe of 40050000.key failed with error -2 Fixes: 93898eb7 ("arm: dts: lpc32xx: add clock properties to device nodes") Signed-off-by:
-
Rafael J. Wysocki authored
[ Upstream commit a1fdbfbb ] Calling rpm_put_suppliers() from pm_runtime_drop_link() is excessive as it affects all suppliers of the consumer device and not just the one pointed to by the device link being dropped. Worst case it may cause the consumer device to stop working unexpectedly. Moreover, in principle it is racy with respect to runtime PM of the consumer device. To avoid these problems drop runtime PM references on the particular supplier pointed to by the link in question only and do that after the link has been dropped from the consumer device's list of links to suppliers, which is in device_link_free(). Fixes: a0504aec ("PM / runtime: Drop usage count for suppliers at device link removal") Signed-off-by:
Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by:
-
Rafael J. Wysocki authored
[ Upstream commit e2f3cd83 ] After commit ead18c23 ("driver core: Introduce device links reference counting"), if there is a link between the given supplier and the given consumer already, device_link_add() will refcount it and return it unconditionally without updating its flags. It is possible, however, that the second (or any subsequent) caller of device_link_add() for the same consumer-supplier pair will pass DL_FLAG_PM_RUNTIME, possibly along with DL_FLAG_RPM_ACTIVE, in flags to it and the existing link may not behave as expected then. First, if DL_FLAG_PM_RUNTIME is not set in the existing link's flags at all, it needs to be set like during the original initialization of the link. Second, if DL_FLAG_RPM_ACTIVE is passed to device_link_add() in flags (in addition to DL_FLAG_PM_RUNTIME), the existing link should to be updated to reflect the "active" runtime PM configuration of the consumer-supplier pair and extra care must be taken here to avoid possible destructive races with runtime PM of the consumer. To that end, redefine the rpm_active field in struct device_link as a refcount, initialize it to 1 and make rpm_resume() (for the consumer) and device_link_add() increment it whenever they acquire a runtime PM reference on the supplier device. Accordingly, make rpm_suspend() (for the consumer) and pm_runtime_clean_up_links() decrement it and drop runtime PM references to the supplier device in a loop until rpm_active becones 1 again. Fixes: ead18c23 ("driver core: Introduce device links reference counting") Signed-off-by:
-
Rafael J. Wysocki authored
[ Upstream commit 5db25c9e ] It is incorrect to call pm_runtime_get_sync() under device_links_write_lock(), because it may end up trying to take device_links_read_lock() while resuming the target device and that will deadlock in the non-SRCU case, so avoid that by resuming the supplier device in device_link_add() before calling device_links_write_lock(). Fixes: 21d5c57b ("PM / runtime: Use device links") Fixes: baa8809f ("PM / runtime: Optimize the use of device links") Signed-off-by:
-
Rafael J. Wysocki authored
[ Upstream commit f265df55 ] After commit ead18c23 ("driver core: Introduce device links reference counting"), if there is a link between the given supplier and the given consumer already, device_link_add() will refcount it and return it unconditionally. However, if the flags passed to it on the second (or any subsequent) attempt to create a device link between the same consumer-supplier pair are not compatible with the existing link's flags, that is incorrect. First off, if the existing link is stateless and the next caller of device_link_add() for the same consumer-supplier pair wants a stateful one, or the other way around, the existing link cannot be returned, because it will not match the expected behavior, so make device_link_add() dump the stack and return NULL in that case. Moreover, if the DL_FLAG_AUTOREMOVE_CONSUMER flag is passed to device_link_add(), its caller will expect its reference to the link to be dropped automatically on consumer driver removal, which will not happen if that flag is not set in the link's flags (and analogously for DL_FLAG_AUTOREMOVE_SUPPLIER). For this reason, make device_link_add() update the existing link's flags accordingly before returning it to the caller. Fixes: ead18c23 ("driver core: Introduce device links reference counting") Signed-off-by:
-
Rafael J. Wysocki authored
[ Upstream commit c8d50986 ] Change the list walk in device_links_driver_cleanup() to a safe one to avoid use-after-free when dropping a link from the list during the walk. Also, while at it, fix device_link_add() to refuse to create stateless device links with DL_FLAG_AUTOREMOVE_SUPPLIER set, which is an invalid combination (setting that flag means that the driver core should manage the link, so it cannot be stateless), and extend the kerneldoc comment of device_link_add() to cover the DL_FLAG_AUTOREMOVE_SUPPLIER flag properly too. Fixes: 1689cac5 ("driver core: Add flag to autoremove device link on supplier unbind") Signed-off-by:
-
Corentin Labbe authored
[ Upstream commit 6e88098c ] When building without CONFIG_HW_RANDOM_PPC4XX, I hit the following build failure: drivers/crypto/amcc/crypto4xx_core.c: In function 'crypto4xx_probe': drivers/crypto/amcc/crypto4xx_core.c:1407:20: error: passing argument 1 of 'ppc4xx_trng_probe' from incompatible pointer type [-Werror=incompatible-pointer-types] In file included from drivers/crypto/amcc/crypto4xx_core.c:50:0: drivers/crypto/amcc/crypto4xx_trng.h:28:20: note: expected 'struct crypto4xx_device *' but argument is of type 'struct crypto4xx_core_device *' drivers/crypto/amcc/crypto4xx_core.c: In function 'crypto4xx_remove': drivers/crypto/amcc/crypto4xx_core.c:1434:21: error: passing argument 1 of 'ppc4xx_trng_remove' from incompatible pointer type [-Werror=incompatible-pointer-types] In file included from drivers/crypto/amcc/crypto4xx_core.c:50:0: drivers/crypto/amcc/crypto4xx_trng.h:30:20: note: expected 'struct crypto4xx_device *' but argument is of type 'struct crypto4xx_core_device *' This patch fix the needed argument of ppc4xx_trng_probe()/ppc4xx_trng_remove() in that case. Fixes: 5343e674 ("crypto4xx: integrate ppc4xx-rng into crypto4xx") Signed-off-by:
Corentin Labbe <clabbe@baylibre.com> Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by:
-
Liu Jian authored
[ Upstream commit 221a1f4a ] In uio_dev_add_attributes() error handing case, idev is used after device_unregister(), in which 'idev' has been released, touch idev cause use-after-free. Fixes: a93e7b33 ("uio: Prevent device destruction while fds are open") Signed-off-by:
Liu Jian <liujian56@huawei.com> Reviewed-by:
Hamish Martin <hamish.martin@alliedtelesis.co.nz> Signed-off-by:
-
Liu Jian authored
[ Upstream commit 1a392b3d ] 'idev' is malloced in __uio_register_device() and leak free it before leaving from the uio_get_minor() error handing case, it will cause memory leak. Fixes: a93e7b33 ("uio: Prevent device destruction while fds are open") Signed-off-by:
-
YueHaibing authored
[ Upstream commit 7dd50e20 ] There is a potential NULL pointer dereference in case alloc_ctrl_packet() fails and returns NULL. Fixes: 099dc4fb ("ipwireless: driver for PC Card 3G/UMTS modem") Signed-off-by:
-
Tony Lindgren authored
[ Upstream commit 9bd34c63 ] Commit 84badc5e ("ARM: dts: omap4: Move l4 child devices to probe them with ti-sysc") started producing a warning for pwm-omap-dmtimer: WARNING: CPU: 0 PID: 77 at drivers/bus/omap_l3_noc.c:147 l3_interrupt_handler+0x2f8/0x388 44000000.ocp:L3 Custom Error: MASTER MPU TARGET L4PER2 (Idle): Data Access in Supervisor mode during Functional access ... __pm_runtime_idle omap_dm_timer_disable pwm_omap_dmtimer_start pwm_omap_dmtimer_enable pwm_apply_state pwm_vibrator_start pwm_vibrator_play_work This is because the timer that pwm-omap-dmtimer is using is now being probed with ti-sysc interconnect target module instead of omap_device and the ti-sysc quirk for SYSC_QUIRK_LEGACY_IDLE is not fully compatible with what omap_device has been doing. We could fix this by reverting the timer changes and have the timer probe again with omap_device. Or we could add more quirk handling to ti-sysc driver. But as these options don't work nicely as longer term solutions, let's just make timers probe with ti-sysc without any quirks. To do this, all we need to do is remove quirks for timers for ti-sysc, and drop the bogus pm_runtime_irq_safe() flag for timer-ti-dm. We should not use pm_runtime_irq_safe() anyways for drivers as it will take a permanent use count on the parent device blocking the parent devices from idling and has been forcing ti-sysc driver to use a quirk flag. Note that we will move the timer data to DEBUG section later on in clean-up patches. Fixes: 84badc5e ("ARM: dts: omap4: Move l4 child devices to probe them with ti-sysc") Cc: Andy Shevchenko <andy.shevchenko@gmail.com> Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: H. Nikolaus Schaller <hns@goldelico.com> Cc: Keerthy <j-keerthy@ti.com> Cc: Ladislav Michl <ladis@linux-mips.org> Cc: Pavel Machek <pavel@ucw.cz> Cc: Sebastian Reichel <sre@kernel.org> Cc: Tero Kristo <t-kristo@ti.com> Cc: Thierry Reding <thierry.reding@gmail.com> Cc: Thomas Gleixner <tglx@linutronix.de> Reported-by:
H. Nikolaus Schaller <hns@goldelico.com> Tested-By:
Andreas Kemnade <andreas@kemnade.info> Tested-By:
Tony Lindgren <tony@atomide.com> Signed-off-by:
-
Johannes Berg authored
[ Upstream commit 1f7698ab ] The current code assigns the reference, and then goes to increment it if the toggle bit has changed. That way, we get Toggle 0 0 0 0 1 1 1 1 ID 1 1 1 1 1 2 2 2 Fix that by assigning the post-toggle ID to get Toggle 0 0 0 0 1 1 1 1 ID 1 1 1 1 2 2 2 2 Reported-by:
Danny Alexander <danny.alexander@intel.com> Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Fixes: fbe41127 ("iwlwifi: mvm: update mpdu metadata API") Signed-off-by:
-
Chen-Yu Tsai authored
[ Upstream commit 52d9bcb3 ] The GIC device node was placed out of order in the initial device tree submission. Move it so the nodes are correctly sorted by base address again. Fixes: e54be32d ("arm64: allwinner: h6: add the basical Allwinner H6 DTSI file") Signed-off-by:
Chen-Yu Tsai <wens@csie.org> Signed-off-by:
Maxime Ripard <maxime.ripard@bootlin.com> Signed-off-by:
-
wenxu authored
[ Upstream commit 6e6b904a ] Init the gre_key from tuninfo->key.tun_id and init the mark from the skb->mark, set the oif to zero in the collect metadata mode. Fixes: cfc7381b ("ip_tunnel: add collect_md mode to IPIP tunnel") Signed-off-by:
wenxu <wenxu@ucloud.cn> Signed-off-by:
-
Moni Shoua authored
[ Upstream commit 33814e5d ] The lock in qp_table might be taken from process context or from interrupt context. This may lead to a deadlock unless it is taken with IRQs disabled. Discovered by lockdep ================================ WARNING: inconsistent lock state 4.20.0-rc6 -------------------------------- inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} python/12572 [HC1[1]:SC0[0]:HE0:SE1] takes: 00000000052a4df4 (&(&table->lock)->rlock#2){?.+.}, /0x50 [mlx5_core] {HARDIRQ-ON-W} state was registered at: _raw_spin_lock+0x33/0x70 mlx5_get_rsc+0x1a/0x50 [mlx5_core] mlx5_ib_eqe_pf_action+0x493/0x1be0 [mlx5_ib] process_one_work+0x90c/0x1820 worker_thread+0x87/0xbb0 kthread+0x320/0x3e0 ret_from_fork+0x24/0x30 irq event stamp: 103928 hardirqs last enabled at (103927): [] nk+0x1a/0x1c hardirqs last disabled at (103928): [] unk+0x1a/0x1c softirqs last enabled at (103924): [] tcp_sendmsg+0x31/0x40 softirqs last disabled at (103922): [] 80 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&table->lock)->rlock#2); lock(&(&table->lock)->rlock#2); *** DEADLOCK *** Fixes: 032080ab ("IB/mlx5: Lock QP during page fault handling") Signed-off-by:
Moni Shoua <monis@mellanox.com> Reviewed-by:
Leon Romanovsky <leonro@mellanox.com> Signed-off-by:
Saeed Mahameed <saeedm@mellanox.com> Signed-off-by:
-
Mordechay Goodstein authored
[ Upstream commit b0d795a9 ] The value in txq_id can be out of array scope, validate it before accessing the array. Signed-off-by:
Mordechay Goodstein <mordechay.goodstein@intel.com> Fixes: cf961e16 ("iwlwifi: mvm: support dqa-mode agg on non-shared queue") Signed-off-by:
-
