- 04 Feb, 2015 29 commits
-
-
Tejun Heo authored
commit 29187a9e upstream. A worker_pool's forward progress is guaranteed by the fact that the last idle worker assumes the manager role to create more workers and summon the rescuers if creating workers doesn't succeed in timely manner before proceeding to execute work items. This manager role is implemented in manage_workers(), which indicates whether the worker may proceed to work item execution with its return value. This is necessary because multiple workers may contend for the manager role, and, if there already is a manager, others should proceed to work item execution. Unfortunately, the function also indicates that the worker may proceed to work item execution if need_to_create_worker() is false at the head of the function. need_to_create_worker() tests the following conditions. pending work items && !nr_running && !nr_idle The first and third conditions are protected by pool->lock and thus won't change while holding pool->lock; however, nr_running can change asynchronously as other workers block and resume and while it's likely to be zero, as someone woke this worker up in the first place, some other workers could have become runnable inbetween making it non-zero. If this happens, manage_worker() could return false even with zero nr_idle making the worker, the last idle one, proceed to execute work items. If then all workers of the pool end up blocking on a resource which can only be released by a work item which is pending on that pool, the whole pool can deadlock as there's no one to create more workers or summon the rescuers. This patch fixes the problem by removing the early exit condition from maybe_create_worker() and making manage_workers() return false iff there's already another manager, which ensures that the last worker doesn't start executing work items. We can leave the early exit condition alone and just ignore the return value but the only reason it was put there is because the manage_workers() used to perform both creations and destructions of workers and thus the function may be invoked while the pool is trying to reduce the number of workers. Now that manage_workers() is called only when more workers are needed, the only case this early exit condition is triggered is rare race conditions rendering it pointless. Tested with simulated workload and modified workqueue code which trigger the pool deadlock reliably without this patch. Signed-off-by:
Tejun Heo <tj@kernel.org> Reported-by:
Eric Sandeen <sandeen@sandeen.net> Link: http://lkml.kernel.org/g/54B019F4.8030009@sandeen.net Cc: Dave Chinner <david@fromorbit.com> Cc: Lai Jiangshan <laijs@cn.fujitsu.com> [ luis: backported to 3.16: - maybe_create_worker() is now void, 'return' instead of 'return true' ] Signed-off-by:
Luis Henriques <luis.henriques@canonical.com>
-
Yinghai Lu authored
commit 851b0936 upstream. Every PCI-PCI bridge window should fit inside an upstream bridge window because orphaned address space is unreachable from the primary side of the upstream bridge. If we inherit invalid bridge windows that overlap an upstream window from firmware, clip them to fit and update the bridge accordingly. [bhelgaas: changelog] Link: https://bugzilla.kernel.org/show_bug.cgi?id=85491Reported-by:
Marek Kordik <kordikmarek@gmail.com> Tested-by:
Yinghai Lu <yinghai@kernel.org> Signed-off-by:
Bjorn Helgaas <bhelgaas@google.com> CC: Thomas Gleixner <tglx@linutronix.de> CC: Ingo Molnar <mingo@redhat.com> CC: "H. Peter Anvin" <hpa@zytor.com> CC: x86@kernel.org Signed-off-by:
-
Yinghai Lu authored
commit 8505e729 upstream. Add pci_claim_bridge_resource() to claim a PCI-PCI bridge window. This is like regular pci_claim_resource(), except that if we fail to claim the window, we check to see if we can reduce the size of the window and try again. This is for scenarios like this: pci_bus 0000:00: root bus resource [mem 0xc0000000-0xffffffff] pci 0000:00:01.0: bridge window [mem 0xbdf00000-0xddefffff 64bit pref] pci 0000:01:00.0: reg 0x10: [mem 0xc0000000-0xcfffffff pref] The 00:01.0 window is illegal: it starts before the host bridge window, so we have to assume the [0xbdf00000-0xbfffffff] region is inaccessible. We can make it legal by clipping it to [mem 0xc0000000-0xddefffff 64bit pref]. Previously we discarded the 00:01.0 window and tried to reassign that part of the hierarchy from scratch. That is a problem because Linux doesn't always assign things optimally. For example, in this case, BIOS put the 01:00.0 device in a prefetchable window below 4GB, but after 5b285415, Linux puts the prefetchable window above 4GB where the 32-bit 01:00.0 device can't use it. Clipping the 00:01.0 window is less intrusive than completely reassigning things and is sufficient to let us use most of the BIOS configuration. Of course, it's possible that devices below 00:01.0 will no longer fit. If that's the case, we'll have to reassign things. But that's a separate problem. [bhelgaas: changelog, split into separate patch] Link: https://bugzilla.kernel.org/show_bug.cgi?id=85491Reported-by:
-
Yinghai Lu authored
commit 0f7e7aee upstream. Add pci_bus_clip_resource(). If a PCI-PCI bridge window overlaps an upstream bridge window but is not completely contained by it, this clips the downstream window so it fits inside the upstream one. No functional change (this adds the function but no callers). [bhelgaas: changelog, split into separate patch] Link: https://bugzilla.kernel.org/show_bug.cgi?id=85491Reported-by:
-
Yinghai Lu authored
commit 3f2f4dc4 upstream. pci_setup_bridge_io(), pci_setup_bridge_mmio(), and pci_setup_bridge_mmio_pref() program the windows of PCI-PCI bridges. Previously they accepted a pointer to the pci_bus of the secondary bus, then looked up the bridge leading to that bus. Pass the bridge directly, which will make it more convenient for future callers. No functional change. [bhelgaas: changelog, split into separate patch] Link: https://bugzilla.kernel.org/show_bug.cgi?id=85491Reported-by:
-
Alex Williamson authored
commit c3e59ee4 upstream. Reports against the TL-WDN4800 card indicate that PCI bus reset of this Atheros device cause system lock-ups and resets. I've also been able to confirm this behavior on multiple systems. The device never returns from reset and attempts to access config space of the device after reset result in hangs. Blacklist bus reset for the device to avoid this issue. [bhelgaas: This regression appeared in v3.14. Andreas bisected it to 425c1b22 ("PCI: Add Virtual Channel to save/restore support"), but we don't understand the mechanism by which that commit affects the reset path.] [bhelgaas: changelog, references] Link: http://lkml.kernel.org/r/20140923210318.498dacbd@dualc.maya.orgReported-by:
Andreas Hartmann <andihartmann@freenet.de> Tested-by:
Alex Williamson <alex.williamson@redhat.com> Signed-off-by:
-
Alex Williamson authored
commit f331a859 upstream. Enable a mechanism for devices to quirk that they do not behave when doing a PCI bus reset. We require a modest level of spec compliant behavior in order to do a reset, for instance the device should come out of reset without throwing errors and PCI config space should be accessible after reset. This is too much to ask for some devices. Link: http://lkml.kernel.org/r/20140923210318.498dacbd@dualc.maya.orgSigned-off-by:
-
Hans Holmberg authored
commit 7b8792bb upstream. of_get_named_gpiod_flags fails with -EPROBE_DEFER in cases where the gpio chip is available and the GPIO translation fails. This causes drivers to be re-probed erroneusly, and hides the real problem(i.e. the GPIO number being out of range). Signed-off-by:
Hans Holmberg <hans.holmberg@intel.com> Reviewed-by:
Alexandre Courbot <acourbot@nvidia.com> Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
-
Johan Hovold authored
commit 0915e6fe upstream. The gpio device attributes were never destroyed when the gpio was unexported (or on export failures). Use device_create_with_groups() to create the default device attributes of the gpio class device. Note that this also fixes the attribute-creation race with userspace for these attributes. Remove contingent attributes in export error path and on unexport. Fixes: d8f388d8 ("gpio: sysfs interface") Signed-off-by:
Johan Hovold <johan@kernel.org> Signed-off-by:
-
Johan Hovold authored
commit 121b6a79 upstream. The gpio-chip device attributes were never destroyed when the device was removed. Fix by using device_create_with_groups() to create the device attributes of the chip class device. Note that this also fixes the attribute-creation race with userspace. Fixes: d8f388d8 ("gpio: sysfs interface") Signed-off-by:
-
Alex Deucher authored
commit d8a74e18 upstream. This was accidently lost in 76a0df85. Signed-off-by:
Alex Deucher <alexander.deucher@amd.com> Signed-off-by:
-
Ahmed S. Darwish authored
commit 5e7e6e0c upstream. Recent Leaf firmware versions (>= 3.1.557) do not allow to send commands for non-existing channels. If a command is sent for a non-existing channel, the firmware crashes. Reported-by:
Christopher Storah <Christopher.Storah@invetech.com.au> Signed-off-by:
Olivier Sobrie <olivier@sobrie.be> Signed-off-by:
Ahmed S. Darwish <ahmed.darwish@valeo.com> Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by:
-
Ahmed S. Darwish authored
commit 889b77f7 upstream. Flooding the Kvaser CAN to USB dongle with multiple reads and writes in very high frequency (*), closing the CAN channel while all the transmissions are on (#), opening the device again (@), then sending a small number of packets would make the driver enter an almost infinite loop of: [....] [15959.853988] kvaser_usb 4-3:1.0 can0: cannot find free context [15959.853990] kvaser_usb 4-3:1.0 can0: cannot find free context [15959.853991] kvaser_usb 4-3:1.0 can0: cannot find free context [15959.853993] kvaser_usb 4-3:1.0 can0: cannot find free context [15959.853994] kvaser_usb 4-3:1.0 can0: cannot find free context [15959.853995] kvaser_usb 4-3:1.0 can0: cannot find free context [....] _dragging the whole system down_ in the process due to the excessive logging output. Initially, this has caused random panics in the kernel due to a buggy error recovery path. That got fixed in an earlier commit.(%) This patch aims at solving the root cause. --> 16 tx URBs and contexts are allocated per CAN channel per USB device. Such URBs are protected by: a) A simple atomic counter, up to a value of MAX_TX_URBS (16) b) A flag in each URB context, stating if it's free c) The fact that ndo_start_xmit calls are themselves protected by the networking layers higher above After grabbing one of the tx URBs, if the driver noticed that all of them are now taken, it stops the netif transmission queue. Such queue is worken up again only if an acknowedgment was received from the firmware on one of our earlier-sent frames. Meanwhile, upon channel close (#), the driver sends a CMD_STOP_CHIP to the firmware, effectively closing all further communication. In the high traffic case, the atomic counter remains at MAX_TX_URBS, and all the URB contexts remain marked as active. While opening the channel again (@), it cannot send any further frames since no more free tx URB contexts are available. Reset all tx URB contexts upon CAN channel close. (*) 50 parallel instances of `cangen0 -g 0 -ix` (#) `ifconfig can0 down` (@) `ifconfig can0 up` (%) "can: kvaser_usb: Don't free packets when tight on URBs" Signed-off-by:
-
Ahmed S. Darwish authored
commit b442723f upstream. Flooding the Kvaser CAN to USB dongle with multiple reads and writes in high frequency caused seemingly-random panics in the kernel. On further inspection, it seems the driver erroneously freed the to-be-transmitted packet upon getting tight on URBs and returning NETDEV_TX_BUSY, leading to invalid memory writes and double frees at a later point in time. Note: Finding no more URBs/transmit-contexts and returning NETDEV_TX_BUSY is a driver bug in and out of itself: it means that our start/stop queue flow control is broken. This patch only fixes the (buggy) error handling code; the root cause shall be fixed in a later commit. Acked-by:
-
Oliver Hartkopp authored
commit 9b1087aa upstream. When changing flags in the CAN drivers ctrlmode the provided new content has to be checked whether the bits are allowed to be changed. The bits that are to be changed are given as a bitfield in cm->mask. Therefore checking against cm->flags is wrong as the content can hold any kind of values. The iproute2 tool sets the bits in cm->mask and cm->flags depending on the detected command line options. To be robust against bogus user space applications additionally sanitize the provided flags with the provided mask. Cc: Wolfgang Grandegger <wg@grandegger.com> Signed-off-by:
Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by:
-
Johan Hovold authored
commit 6798acaa upstream. Move direct and indirect calls to gpiochip_remove_pin_ranges outside of spin lock as they can end up taking a mutex in pinctrl_remove_gpio_range. Note that the pin ranges are already added outside of the lock. Fixes: 9ef0d6f7 ("gpiolib: call pin removal in chip removal function") Fixes: f23f1516 ("gpiolib: provide provision to register pin ranges") Signed-off-by:
-
Johan Hovold authored
commit 00acc3dc upstream. Fix memory leak and sleep-while-atomic in gpiochip_remove. The memory leak was introduced by afa82fab ("gpio / ACPI: Move event handling registration to gpiolib irqchip helpers") that moved the release of acpi interrupt resources to gpiochip_irqchip_remove, but by then the resources are no longer accessible as the acpi_gpio_chip has already been freed by acpi_gpiochip_remove. Note that this also fixes a few potential sleep-while-atomics, which has been around since 14250520 ("gpio: add IRQ chip helpers in gpiolib") when the call to gpiochip_irqchip_remove while holding a spinlock was added (a couple of irq-domain paths can end up grabbing mutexes). Fixes: afa82fab ("gpio / ACPI: Move event handling registration to gpiolib irqchip helpers") Fixes: 14250520 ("gpio: add IRQ chip helpers in gpiolib") Signed-off-by:
-
Johan Hovold authored
commit 5539b3c9 upstream. Memory allocated and references taken by of_gpiochip_add and acpi_gpiochip_add were never released on errors in gpiochip_add (e.g. failure to find free gpio range). Fixes: 391c970c ("of/gpio: add default of_xlate function if device has a node pointer") Fixes: 664e3e5a ("gpio / ACPI: register to ACPI events automatically") Signed-off-by:
-
Jim Lin authored
commit db93facf upstream. This patch is to fix two deadlock cases. Deadlock 1: CPU #1 pinctrl_register-> pinctrl_get -> create_pinctrl (Holding lock pinctrl_maps_mutex) -> get_pinctrl_dev_from_devname (Trying to acquire lock pinctrldev_list_mutex) CPU #0 pinctrl_unregister (Holding lock pinctrldev_list_mutex) -> pinctrl_put ->> pinctrl_free -> pinctrl_dt_free_maps -> pinctrl_unregister_map (Trying to acquire lock pinctrl_maps_mutex) Simply to say CPU#1 is holding lock A and trying to acquire lock B, CPU#0 is holding lock B and trying to acquire lock A. Deadlock 2: CPU #3 pinctrl_register-> pinctrl_get -> create_pinctrl (Holding lock pinctrl_maps_mutex) -> get_pinctrl_dev_from_devname (Trying to acquire lock pinctrldev_list_mutex) CPU #2 pinctrl_unregister (Holding lock pctldev->mutex) -> pinctrl_put ->> pinctrl_free -> pinctrl_dt_free_maps -> pinctrl_unregister_map (Trying to acquire lock pinctrl_maps_mutex) CPU #0 tegra_gpio_request (Holding lock pinctrldev_list_mutex) -> pinctrl_get_device_gpio_range (Trying to acquire lock pctldev->mutex) Simply to say CPU#3 is holding lock A and trying to acquire lock D, CPU#2 is holding lock B and trying to acquire lock A, CPU#0 is holding lock D and trying to acquire lock B. Signed-off-by:
Jim Lin <jilin@nvidia.com> Signed-off-by:
-
Alex Deucher authored
commit 5615f890 upstream. This adds a quirks list to fix stability problems with certain SI boards. bug: https://bugs.freedesktop.org/show_bug.cgi?id=76490Signed-off-by:
-
Chris Wilson authored
commit 226e5ae9 upstream. If CONFIG_DEBUG_MUTEXES is set, the mutex->owner field is only cleared if the mutex debugging is enabled which introduces a race in our mutex_is_locked_by() - i.e. we may inspect the old owner value before it is acquired by the new task. This is the root cause of this error: diff --git a/kernel/locking/mutex-debug.c b/kernel/locking/mutex-debug.c index 5cf6731..3ef3736 100644 --- a/kernel/locking/mutex-debug.c +++ b/kernel/locking/mutex-debug.c @@ -80,13 +80,13 @@ void debug_mutex_unlock(struct mutex *lock) DEBUG_LOCKS_WARN_ON(lock->owner != current); DEBUG_LOCKS_WARN_ON(!lock->wait_list.prev && !lock->wait_list.next); - mutex_clear_owner(lock); } /* * __mutex_slowpath_needs_to_unlock() is explicitly 0 for debug * mutexes so that we can do it here after we've verified state. */ + mutex_clear_owner(lock); atomic_set(&lock->count, 1); } Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=87955Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Signed-off-by:
-
Chris Wilson authored
commit 48bf5b2d upstream. Like Ivybridge, we have reports that we get random hangs when flipping with multiple pipes. Extend commit 2a92d5bc Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Tue Jul 8 10:40:29 2014 +0100 drm/i915: Disable RCS flips on Ivybridge to also apply to Haswell. Reported-and-tested-by:
Scott Tsai <scottt.tw@gmail.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=87759Signed-off-by:
-
Dmitry Torokhov authored
commit 9333caea upstream. When KBC is in active multiplexing mode the touchpad on this laptop does not work. Reported-by:
Bilal Koc <koc.bilo@googlemail.com> Signed-off-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by:
-
Sam hung authored
commit 810aa091 upstream. This change allows the driver to recognize newer Elantech touchpads. Signed-off-by:
Yi ju Hong <sam.hung@emc.com.tw> Signed-off-by:
-
Alex Deucher authored
commit 4369a69e upstream. Disable dpm on certain problematic boards rather than disabling dpm for the entire chip family since most boards work fine. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1386534 https://bugzilla.kernel.org/show_bug.cgi?id=83731Signed-off-by:
-
Alex Deucher authored
commit 3a01fd36 upstream. We need to wait for the GPUVM flush to complete. There was some confusion as to how this mechanism was supposed to work. The operation is not atomic. For GPU initiated invalidations you need to read back a VM register to introduce enough latency for the update to complete. v2: drop gart changes v3: just read back rather than polling Reviewed-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
-
Alex Deucher authored
commit d474ea7e upstream. We need to wait for the GPUVM flush to complete. There was some confusion as to how this mechanism was supposed to work. The operation is not atomic. For GPU initiated invalidations you need to read back a VM register to introduce enough latency for the update to complete. v2: drop gart changes v3: just read back rather than polling Reviewed-by:
-
Alex Deucher authored
commit cbfc35b9 upstream. We need to wait for the GPUVM flush to complete. There was some confusion as to how this mechanism was supposed to work. The operation is not atomic. For GPU initiated invalidations you need to read back a VM register to introduce enough latency for the update to complete. v2: drop gart changes v3: just read back rather than polling Reviewed-by:
-
Srihari Vijayaraghavan authored
commit 148e9a71 upstream. On some laptops, keyboard needs to be reset in order to successfully detect touchpad (e.g., some Gigabyte laptop models with Elantech touchpads). Without resettin keyboard touchpad pretends to be completely dead. Based on the original patch by Mateusz Jończyk this version has been expanded to include DMI based detection & application of the fix automatically on the affected models of laptops. This has been confirmed to fix problem by three users already on three different models of laptops. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=81331Signed-off-by:
Srihari Vijayaraghavan <linux.bug.reporting@gmail.com> Acked-by:
Mateusz Jończyk <mat.jonczyk@o2.pl> Tested-by:
Guillaum Bouchard <guillaum.bouchard@gmail.com> Signed-off-by:
-
- 02 Feb, 2015 5 commits
-
-
Sasha Levin authored
commit 6ada1fc0 upstream. An unvalidated user input is multiplied by a constant, which can result in an undefined behaviour for large values. While this is validated later, we should avoid triggering undefined behaviour. Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@kernel.org> Signed-off-by:
Sasha Levin <sasha.levin@oracle.com> [jstultz: include trivial milisecond->microsecond correction noticed by Andy] Signed-off-by:
John Stultz <john.stultz@linaro.org> [ luis: backported to 3.16: adjusted context ] Signed-off-by:
-
Tobias Jakobi authored
commit 8c38d28b upstream. EXYNOS4_MCT_L_MASK is defined as 0xffffff00, so applying this bitmask produces a number outside the range 0x00 to 0xff, which always results in execution of the default switch statement. Obviously this is wrong and git history shows that the bitmask inversion was incorrectly set during a refactoring of the MCT code. Fix this by putting the inversion at the correct position again. Acked-by:
Kukjin Kim <kgene.kim@samsung.com> Reported-by:
GP Orcullo <kinsamanka@gmail.com> Reviewed-by:
Doug Anderson <dianders@chromium.org> Signed-off-by:
Tobias Jakobi <tjakobi@math.uni-bielefeld.de> Signed-off-by:
Daniel Lezcano <daniel.lezcano@linaro.org> Signed-off-by:
-
Pablo Neira Ayuso authored
commit 9ea2aa8b upstream. Make sure there is enough room for the nfnetlink header in the netlink messages that are part of the batch. There is a similar check in netlink_rcv_skb(). Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by:
-
Dan Carpenter authored
commit 3b05ac38 upstream. The app_tcp_pkt_out() function expects "*diff" to be set and ends up using uninitialized data if CONFIG_IP_VS_IPV6 is turned on. The same issue is there in app_tcp_pkt_in(). Thanks to Julian Anastasov for noticing that. Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Acked-by:
Julian Anastasov <ja@ssi.bg> Signed-off-by:
Simon Horman <horms@verge.net.au> Cc: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by:
-
Luis Henriques authored
Signed-off-by:
-
- 26 Jan, 2015 3 commits
-
-
Geert Uytterhoeven authored
commit b0ddb319 upstream. The sh73a0 INTC can't mask interrupts properly most likely due to a hardware bug. Set the .control_parent flag to delegate masking to the parent interrupt controller, like was already done for irqpin1. Without this, accessing the three-axis digital accelerometer ADXL345 on kzm9g through /dev/input/event1 causes an interrupt storm, which requires a power-cycle to recover from. This was inspired by a patch for arch/arm/boot/dts/sh73a0.dtsi from Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>. Signed-off-by:
Geert Uytterhoeven <geert+renesas@glider.be> Fixes: 341eb546 ("ARM: shmobile: INTC External IRQ pin driver on sh73a0") Signed-off-by:
Simon Horman <horms+renesas@verge.net.au> Signed-off-by:
-
Mugunthan V N authored
commit 69d2626f upstream. 64KiB is allocated for qspi dtb partition which is not sufficient, so updating the partition table size to 512KiB for device tree partition. This also aligns the QSPI partition definitions between kernel and U-Boot. Fixes: dc2dd5b8 ("ARM: dts: dra7: Add qspi device") Signed-off-by:
Mugunthan V N <mugunthanvnm@ti.com> Signed-off-by:
Tony Lindgren <tony@atomide.com> Signed-off-by:
-
David Vrabel authored
commit dbdd7476 upstream. This reverts commit 2c3fc8d2. This commit broke on x86 PV because entries in the generic SWIOTLB are indexed using (pseudo-)physical address not DMA address and these are not the same in a x86 PV guest. Signed-off-by:
David Vrabel <david.vrabel@citrix.com> Reviewed-by:
Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Signed-off-by:
-
- 23 Jan, 2015 3 commits
-
-
Chris Wilson authored
commit c4d69da1 upstream. Running igt, I was encountering the invalid TLB bug on my 845g, despite that it was using the CS workaround. Examining the w/a buffer in the error state, showed that the copy from the user batch into the workaround itself was suffering from the invalid TLB bug (the first cacheline was broken with the first two words reversed). Time to try a fresh approach. This extends the workaround to write into each page of our scratch buffer in order to overflow the TLB and evict the invalid entries. This could be refined to only do so after we update the GTT, but for simplicity, we do it before each batch. I suspect this supersedes our current workaround, but for safety keep doing both. v2: The magic number shall be 2. This doesn't conclusively prove that it is the mythical TLB bug we've been trying to workaround for so long, that it requires touching a number of pages to prevent the corruption indicates to me that it is TLB related, but the corruption (the reversed cacheline) is more subtle than a TLB bug, where we would expect it to read the wrong page entirely. Oh well, it prevents a reliable hang for me and so probably for others as well. Signed-off-by:
-
Alexander Y. Fomichev authored
commit 7ce64c79 upstream. __netdev_adjacent_dev_insert may add adjust device of different net namespace, without proper check it leads to emergence of broken sysfs links from/to devices in another namespace. Fix: rewrite netdev_adjacent_is_neigh_list macro as a function, move net_eq check into netdev_adjacent_is_neigh_list. (thanks David) related to: 4c75431aSigned-off-by:
Alexander Fomichev <git.user@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net> Cc: Miquel van Smoorenburg <mikevs@xs4all.net> Signed-off-by:
-
Alexander Y. Fomichev authored
commit 4c75431a upstream. Code manipulating sysfs symlinks on adjacent net_devices(s) currently doesn't take into account that devices potentially belong to different namespaces. This patch trying to fix an issue as follows: - check for net_ns before creating / deleting symlink. for now only netdev_adjacent_rename_links and __netdev_adjacent_dev_remove are affected, afaics __netdev_adjacent_dev_insert implies both net_devs belong to the same namespace. - Drop all existing symlinks to / from all adj_devs before switching namespace and recreate them just after. Signed-off-by:
-
