- 03 Aug, 2015 11 commits
-
-
Thomas Petazzoni authored
commit ea78b951 upstream. There was a mistake in the definition of the functions for MPP48 on Marvell Armada XP. The second function is dev(clkout), and not tclk. Signed-off-by:
Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Fixes: 463e270f ("pinctrl: mvebu: add pinctrl driver for Armada XP") Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Thomas Petazzoni authored
commit 80b3d04f upstream. The latest version of the Armada XP datasheet no longer documents the VDD cpu_pd functions, which might indicate they are not working and/or not supported. This commit ensures the pinctrl driver matches the datasheet. Signed-off-by:
-
Thomas Petazzoni authored
commit bc99357f upstream. After updating to a more recent version of the Armada XP datasheet, we realized that some of the pins documented as having a NAND-related functionality in fact did not have such functionality. This commit updates the pinctrl driver accordingly. Signed-off-by:
-
Thomas Petazzoni authored
commit 7c580311 upstream. The pinctrl_gpio_range[] array described a first bank of 32 GPIOs and a second one of 27 GPIOs. However, since there is a total of 60 MPP pins that can be muxed as GPIOs, the second bank really has 28 GPIOs. Signed-off-by:
-
Thomas Petazzoni authored
commit 27e7cd01 upstream. The pinctrl_gpio_range[] array described a first bank of 32 GPIOs and a second one of 27 GPIOs. However, since there is a total of 60 MPP pins that can be muxed as GPIOs, the second bank really has 28 GPIOs. Signed-off-by:
-
Thomas Petazzoni authored
commit 331642fb upstream. A new revision of the Marvell Armada 38x hardware datasheet unveiled that the definition of some of the PCIe functions were not correct. This commit fixes the pinctrl driver accordingly. Some PCIe functions simply do not exist, some of the PCIe functions in fact were corresponding to other functions, and some PCIe functions have been added. Note: the seemingly unrelated removal of spi(cs2) on MPP47 is related: this function is in fact implemented on MPP43, instead of a PCIe function. Signed-off-by:
-
Thomas Petazzoni authored
commit e5447d26 upstream. After updating to a more recent version of the Armada 375, we realized that some of the pins documented as having a NAND-related functionality in fact did not have such functionality. This commit updates the pinctrl driver accordingly. Signed-off-by:
-
Thomas Petazzoni authored
commit d538990e upstream. There was an incorrect space in the definition of the function of one pin in the Armada 375 pinctrl driver, which this commit fixes. Signed-off-by:
-
Thomas Petazzoni authored
commit 438881df upstream. Due to a mistake, the CS0 and CS1 SPI0 functions were incorrectly named "spi0-1" instead of just "spi0". This commit fixes that. This DT binding change does not affect any of the in-tree users. Signed-off-by:
-
Masahiro Yamada authored
commit 5cf021d5 upstream. The address for SD0_WP_CD_SEL, SD1_WP_CD_SEL is 0xf8000830, 0xf8000834, respectively. Each offset address must be prefixed with 0x. Fixes: add958ce "pinctrl: Add driver for Zynq" Signed-off-by:
Masahiro Yamada <yamada.masahiro@socionext.com> Reviewed-by:
Sören Brinkmann <soren.brinkmann@xilinx.com> Signed-off-by:
-
Masahiro Yamada authored
commit 4f652cea upstream. The offset to the mux register is missing. Fixes: add958ce "pinctrl: Add driver for Zynq" Signed-off-by:
-
- 21 Jul, 2015 29 commits
-
-
Greg Kroah-Hartman authored
-
Frodo Lai authored
commit 469d7d22 upstream. The i2c_master_recv() uses readsize to receive data from i2c but compares to size of rdbuf which is always 27. This would cause problem when the max_fingers is not 5. Change the comparison value to readsize instead. Fixes: 36874c7e ("Input: pixcir_i2c_ts - support up to 5 fingers and hardware tracking IDs:) Signed-off-by:
Frodo Lai <frodo_lai@bcmcom.com> Signed-off-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by:
-
Zhichang Yuan authored
commit 5dbb4c61 upstream. 41f8bba7 ("of/pci: Add pci_register_io_range() and pci_pio_to_address()") added support for systems with several I/O ranges described by OF bindings. It modified pci_address_to_pio() look up the io_range for a given CPU physical address, but the conversion was wrong. Fix the conversion of address to I/O port. [bhelgaas: changelog] Fixes: 41f8bba7 ("of/pci: Add pci_register_io_range() and pci_pio_to_address()") Signed-off-by:
Zhichang Yuan <yuanzhichang@hisilicon.com> Signed-off-by:
Bjorn Helgaas <bhelgaas@google.com> Acked-by:
Liviu Dudau <Liviu.Dudau@arm.com> Signed-off-by:
-
Alex Williamson authored
commit a5dd4b4b upstream. The commit referenced below deferred waiting for command completion until the start of the next command, allowing hardware to do the latching asynchronously. Unfortunately, being ready to accept a new command is the only indication we have that the previous command is completed. In cases where we need that state change to be enabled, we must still wait for completion. For instance, pciehp_reset_slot() attempts to disable anything that might generate a surprise hotplug on slots that support presence detection. If we don't wait for those settings to latch before the secondary bus reset, we negate any value in attempting to prevent the spurious hotplug. Create a base function with optional wait and helper functions so that pcie_write_cmd() turns back into the "safe" interface which waits before and after issuing a command and add pcie_write_cmd_nowait(), which eliminates the trailing wait for asynchronous completion. The following functions are returned to their previous behavior: pciehp_power_on_slot pciehp_power_off_slot pcie_disable_notification pciehp_reset_slot The rationale is that pciehp_power_on_slot() enables the link and therefore relies on completion of power-on. pciehp_power_off_slot() and pcie_disable_notification() need a wait because data structures may be freed after these calls and continued signaling from the device would be unexpected. And, of course, pciehp_reset_slot() needs to wait for the scenario outlined above. Fixes: 3461a068 ("PCI: pciehp: Wait for hotplug command completion lazily") Signed-off-by:
Alex Williamson <alex.williamson@redhat.com> Signed-off-by:
-
Yinghai Lu authored
commit 3a9ad0b4 upstream. David Ahern reported that d63e2e1f ("sparc/PCI: Clip bridge windows to fit in upstream windows") fails to boot on sparc/T5-8: pci 0000:06:00.0: reg 0x184: can't handle BAR above 4GB (bus address 0x110204000) The problem is that sparc64 assumed that dma_addr_t only needed to hold DMA addresses, i.e., bus addresses returned via the DMA API (dma_map_single(), etc.), while the PCI core assumed dma_addr_t could hold *any* bus address, including raw BAR values. On sparc64, all DMA addresses fit in 32 bits, so dma_addr_t is a 32-bit type. However, BAR values can be 64 bits wide, so they don't fit in a dma_addr_t. d63e2e1f added new checking that tripped over this mismatch. Add pci_bus_addr_t, which is wide enough to hold any PCI bus address, including both raw BAR values and DMA addresses. This will be 64 bits on 64-bit platforms and on platforms with a 64-bit dma_addr_t. Then dma_addr_t only needs to be wide enough to hold addresses from the DMA API. [bhelgaas: changelog, bugzilla, Kconfig to ensure pci_bus_addr_t is at least as wide as dma_addr_t, documentation] Fixes: d63e2e1f ("sparc/PCI: Clip bridge windows to fit in upstream windows") Fixes: 23b13bc7 ("PCI: Fail safely if we can't handle BARs larger than 4GB") Link: http://lkml.kernel.org/r/CAE9FiQU1gJY1LYrxs+ma5LCTEEe4xmtjRG0aXJ9K_Tsu+m9Wuw@mail.gmail.com Link: http://lkml.kernel.org/r/1427857069-6789-1-git-send-email-yinghai@kernel.org Link: https://bugzilla.kernel.org/show_bug.cgi?id=96231Reported-by:
David Ahern <david.ahern@oracle.com> Tested-by:
Yinghai Lu <yinghai@kernel.org> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Rafael J. Wysocki authored
commit 08249651 upstream. Refine the mechanism introduced by commit f244d8b6 ("ACPIPHP / radeon / nouveau: Fix VGA switcheroo problem related to hotplug") to propagate the ignore_hotplug setting of the device to its parent bridge in case hotplug notifications related to the graphics adapter switching are given for the bridge rather than for the device itself (they need to be ignored in both cases). Link: https://bugzilla.kernel.org/show_bug.cgi?id=61891 Link: https://bugs.freedesktop.org/show_bug.cgi?id=88927 Fixes: b440bde7 ("PCI: Add pci_ignore_hotplug() to ignore hotplug events for a device") Reported-and-tested-by:
tiagdtd-lava <tiagdtd-lava@yahoo.de> Signed-off-by:
Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by:
-
Uwe Kleine-König authored
commit e5babdf9 upstream. Since commit bd31b859 (which is in 3.2-rc1) nw_gpio_lock is a raw spinlock that needs usage of the corresponding raw functions. This fixes: drivers/mtd/maps/dc21285.c: In function 'nw_en_write': drivers/mtd/maps/dc21285.c:41:340: warning: passing argument 1 of 'spinlock_check' from incompatible pointer type spin_lock_irqsave(&nw_gpio_lock, flags); In file included from include/linux/seqlock.h:35:0, from include/linux/time.h:5, from include/linux/stat.h:18, from include/linux/module.h:10, from drivers/mtd/maps/dc21285.c:8: include/linux/spinlock.h:299:102: note: expected 'struct spinlock_t *' but argument is of type 'struct raw_spinlock_t *' static inline raw_spinlock_t *spinlock_check(spinlock_t *lock) ^ drivers/mtd/maps/dc21285.c:43:25: warning: passing argument 1 of 'spin_unlock_irqrestore' from incompatible pointer type spin_unlock_irqrestore(&nw_gpio_lock, flags); ^ In file included from include/linux/seqlock.h:35:0, from include/linux/time.h:5, from include/linux/stat.h:18, from include/linux/module.h:10, from drivers/mtd/maps/dc21285.c:8: include/linux/spinlock.h:370:91: note: expected 'struct spinlock_t *' but argument is of type 'struct raw_spinlock_t *' static inline void spin_unlock_irqrestore(spinlock_t *lock, unsigned long flags) Fixes: bd31b859 ("locking, ARM: Annotate low level hw locks as raw") Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by:
Brian Norris <computersforpeace@gmail.com> Signed-off-by:
-
Brian Norris authored
commit 073db4a5 upstream. On A MIPS 32-cores machine a BUG_ON was triggered because some acesses to mtd->usecount were done without taking mtd_table_mutex. kernel: Call Trace: kernel: [<ffffffff80401818>] __put_mtd_device+0x20/0x50 kernel: [<ffffffff804086f4>] blktrans_release+0x8c/0xd8 kernel: [<ffffffff802577e0>] __blkdev_put+0x1a8/0x200 kernel: [<ffffffff802579a4>] blkdev_close+0x1c/0x30 kernel: [<ffffffff8022006c>] __fput+0xac/0x250 kernel: [<ffffffff80171208>] task_work_run+0xd8/0x120 kernel: [<ffffffff8012c23c>] work_notifysig+0x10/0x18 kernel: kernel: Code: 2442ffff ac8202d8 000217fe <00020336> dc820128 10400003 00000000 0040f809 00000000 kernel: ---[ end trace 080fbb4579b47a73 ]--- Fixed by taking the mutex in blktrans_open and blktrans_release. Note that this locking is already suggested in include/linux/mtd/blktrans.h: struct mtd_blktrans_ops { ... /* Called with mtd_table_mutex held; no race with add/remove */ int (*open)(struct mtd_blktrans_dev *dev); void (*release)(struct mtd_blktrans_dev *dev); ... }; But we weren't following it. Originally reported by (and patched by) Zhang and Giuseppe, independently. Improved and rewritten. Reported-by:
Zhang Xingcai <zhangxingcai@huawei.com> Reported-by:
Giuseppe Cantavenera <giuseppe.cantavenera.ext@nokia.com> Tested-by:
Alexander Sverdlin <alexander.sverdlin@nokia.com> Signed-off-by:
-
Grygorii Strashko authored
commit 084609bf upstream. Setting a dev_pm_ops suspend/resume pair of callbacks but not a set of hibernation callbacks means those pm functions will not be called upon hibernation - that leads to system crash on ARM during freezing if gpio-led is used in combination with CPU led trigger. It may happen after freeze_noirq stage (GPIO is suspended) and before syscore_suspend stage (CPU led trigger is suspended) - usually when disable_nonboot_cpus() is called. Log: PM: noirq freeze of devices complete after 1.425 msecs Disabling non-boot CPUs ... ^ system may crash or stuck here with message (TI AM572x) WARNING: CPU: 0 PID: 3100 at drivers/bus/omap_l3_noc.c:148 l3_interrupt_handler+0x22c/0x370() 44000000.ocp:L3 Custom Error: MASTER MPU TARGET L4_PER1_P3 (Idle): Data Access in Supervisor mode during Functional access CPU1: shutdown ^ or here Fix this by using SIMPLE_DEV_PM_OPS, which appropriately assigns the suspend and hibernation callbacks and move led_suspend/led_resume under CONFIG_PM_SLEEP to avoid build warnings. Fixes: 73e1ab41 (leds: Convert led class driver from legacy pm ops to dev_pm_ops) Signed-off-by:
Grygorii Strashko <Grygorii.Strashko@linaro.org> Acked-by:
Jacek Anaszewski <j.anaszewski@samsung.com> Signed-off-by:
-
Liu Ying authored
commit 2fa3b4c4 upstream. The LCDIF engines embedded in i.MX6sl and i.MX6sx SoCs need the axi clock as the engine's system clock. The clock should be enabled when accessing LCDIF registers, otherwise the kernel would hang up. We should also keep the clock enabled when the engine is being active to scan out frames from memory. This patch makes sure the axi clock is enabled when accessing registers so that the kernel hang up issue can be fixed. Reported-by:
Peter Chen <peter.chen@freescale.com> Tested-by:
Liu Ying <Ying.Liu@freescale.com> Signed-off-by:
Tomi Valkeinen <tomi.valkeinen@ti.com> Signed-off-by:
-
Axel Lin authored
commit 63781394 upstream. request_any_context_irq() returns a negative value on failure. It returns either IRQC_IS_HARDIRQ or IRQC_IS_NESTED on success. So fix testing return value of request_any_context_irq(). Also fixup the return value of devm_request_any_context_irq() to make it consistent with request_any_context_irq(). Fixes: 0668d306 ("genirq: Add devm_request_any_context_irq()") Signed-off-by:
Axel Lin <axel.lin@ingics.com> Reviewed-by:
Stephen Boyd <sboyd@codeaurora.org> Link: http://lkml.kernel.org/r/1431334978.17783.4.camel@ingics.comSigned-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Bart Van Assche authored
commit a44074f1 upstream. Although it is possible to let SRP I/O continue if a reconnect results in a reduction of the number of channels, the current code does not handle this scenario correctly. Instead of making the reconnect code more complex, consider this as a reconnection failure. Signed-off-by:
Bart Van Assche <bart.vanassche@sandisk.com> Cc: Sagi Grimberg <sagig@mellanox.com> Cc: Sebastian Parschauer <sebastian.riemer@profitbricks.com> Signed-off-by:
Doug Ledford <dledford@redhat.com> Signed-off-by:
-
Bart Van Assche authored
commit c014c8cd upstream. Reception of a DREQ message only causes the state of a single channel to change. Hence move the 'connected' member variable from the target to the channel data structure. This patch avoids that following false positive warning can be reported by srp_destroy_qp(): WARNING: at drivers/infiniband/ulp/srp/ib_srp.c:617 srp_destroy_qp+0xa6/0x120 [ib_srp]() Call Trace: [<ffffffff8106e10f>] warn_slowpath_common+0x7f/0xc0 [<ffffffff8106e16a>] warn_slowpath_null+0x1a/0x20 [<ffffffffa0440226>] srp_destroy_qp+0xa6/0x120 [ib_srp] [<ffffffffa0440322>] srp_free_ch_ib+0x82/0x1e0 [ib_srp] [<ffffffffa044408b>] srp_create_target+0x7ab/0x998 [ib_srp] [<ffffffff81346f60>] dev_attr_store+0x20/0x30 [<ffffffff811dd90f>] sysfs_write_file+0xef/0x170 [<ffffffff8116d248>] vfs_write+0xc8/0x190 [<ffffffff8116d411>] sys_write+0x51/0x90 Signed-off-by:
-
Bart Van Assche authored
commit 8de9fe3a upstream. Avoid that receiving a DREQ while RDMA channels are being established causes target->qp_in_error to be reset. Signed-off-by:
-
Bart Van Assche authored
commit fb49c8bb upstream. Fix a scsi_get_host() / scsi_host_put() imbalance in the error path of srp_create_target(). See also patch "IB/srp: Avoid that I/O hangs due to a cable pull during LUN scanning" (commit ID 34aa654e). Signed-off-by:
Sagi Grimberg <sagig@mellanox.com> Cc: Sebastian Parschauer <sebastian.riemer@profitbricks.com> Signed-off-by:
-
Bart Van Assche authored
commit 535fb906 upstream. Avoid that srp_terminate_io() can get invoked while srp_queuecommand() is in progress. This patch avoids that an I/O timeout can trigger the following kernel warning: WARNING: at drivers/infiniband/ulp/srp/ib_srp.c:1447 srp_terminate_io+0xef/0x100 [ib_srp]() Call Trace: [<ffffffff814c65a2>] dump_stack+0x4e/0x68 [<ffffffff81051f71>] warn_slowpath_common+0x81/0xa0 [<ffffffff8105204a>] warn_slowpath_null+0x1a/0x20 [<ffffffffa075f51f>] srp_terminate_io+0xef/0x100 [ib_srp] [<ffffffffa07495da>] __rport_fail_io_fast+0xba/0xc0 [scsi_transport_srp] [<ffffffffa0749a90>] rport_fast_io_fail_timedout+0xe0/0xf0 [scsi_transport_srp] [<ffffffff8106e09b>] process_one_work+0x1db/0x780 [<ffffffff8106e75b>] worker_thread+0x11b/0x450 [<ffffffff81073c64>] kthread+0xe4/0x100 [<ffffffff814cf26c>] ret_from_fork+0x7c/0xb0 See also patch "scsi_transport_srp: Add transport layer error handling" (commit ID 29c17324). Signed-off-by:
-
Bart Van Assche authored
commit be34c62d upstream. Introduce the helper function srp_wait_for_queuecommand(). Move the definition of scsi_request_fn_active(). Add a comment above srp_wait_for_queuecommand() that support for scsi-mq needs to be added. This patch does not change any functionality. A second call to srp_wait_for_queuecommand() will be introduced in the next patch. Signed-off-by:
-
Ezequiel Garcia authored
commit ea6055c4 upstream. Since commit 39a6ac11 ("spi/pl022: Devicetree support w/o platform data") the 'num-cs' parameter cannot be passed through platform data when probing with devicetree. Instead, it's a required devicetree property. Fix the binding documentation so the property is properly specified. Fixes: 39a6ac11 ("spi/pl022: Devicetree support w/o platform data") Signed-off-by:
Ezequiel Garcia <ezequiel@vanguardiasur.com.ar> Signed-off-by:
Mark Brown <broonie@kernel.org> Signed-off-by:
-
Gregory CLEMENT authored
commit ce2f6ea1 upstream. The commit df59fa7f "spi: orion: support armada extended baud rates" was too optimistic for the maximum baud rate that the Armada SoCs can support. According to the hardware datasheet the maximum frequency supported by the Armada 370 SoC is tclk/4. But for the Armada XP, Armada 38x and Armada 39x SoCs the limitation is 50MHz and for the Armada 375 it is tclk/15. Currently the armada-370-spi compatible is only used by the Armada 370 and the Armada XP device tree. On Armada 370, tclk cannot be higher than 200MHz. In order to be able to handle both SoCs, we can take the minimum of 50MHz and tclk/4. A proper solution is adding a compatible string for each SoC, but it can't be done as a fix for compatibility reason (we can't modify device tree that have been already released) and it will be part of a separate patch. Fixes: df59fa7f (spi: orion: support armada extended baud rates) Reported-by:
Kostya Porotchkin <kostap@marvell.com> Signed-off-by:
Gregory CLEMENT <gregory.clement@free-electrons.com> Signed-off-by:
-
Martin Sperl authored
commit 8e76ef88 upstream. Fix a race (with some kernel configurations) where a queued master->pump_messages runs and frees dummy_tx/rx before spi_unmap_msg is running (or is finished). This results in the following messages: BUG: Bad page state in process page:db7ba030 count:0 mapcount:0 mapping: (null) index:0x0 flags: 0x200(arch_1) page dumped because: PAGE_FLAGS_CHECK_AT_PREP flag set ... Reported-by:
Noralf Trønnes <noralf@tronnes.org> Suggested-by:
Martin Sperl <kernel@martin.sperl.org> Signed-off-by:
-
Miroslav Benes authored
commit 9a1bd63c upstream. The list of loaded modules is walked through in module_kallsyms_on_each_symbol (called by kallsyms_on_each_symbol). The module_mutex lock should be acquired to prevent potential corruptions in the list. This was uncovered with new lockdep asserts in module code introduced by the commit 0be964be ("module: Sanitize RCU usage and locking") in recent next- trees. Signed-off-by:
Miroslav Benes <mbenes@suse.cz> Acked-by:
Josh Poimboeuf <jpoimboe@redhat.com> Signed-off-by:
Jiri Kosina <jkosina@suse.cz> Signed-off-by:
-
Stefan Wahren authored
commit a7068e39 upstream. The buffer for condtraints debug isn't big enough to hold the output in all cases. So fix this issue by increasing the buffer. Signed-off-by:
Stefan Wahren <stefan.wahren@i2se.com> Signed-off-by:
-
Joe Perches authored
commit c53403a3 upstream. The code should handle more than 32 bits here because "id" can be a value up to MAX77686_REGULATORS (currently 34). Convert the gpio_enabled type to DECLARE_BITMAP and use test_bit/set_bit. Fixes: 3307e902 ("regulator: max77686: Add GPIO control") Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Joe Perches <joe@perches.com> Reviewed-by:
Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by:
-
Maxime Coquelin authored
commit 921cc294 upstream. The way the mask is generated in regmap_field_init() is wrong. Indeed, a field initialized with msb = 31 and lsb = 0 provokes a shift overflow while calculating the mask field. On some 32 bits architectures, such as x86, the generated mask is 0, instead of the expected 0xffffffff. This patch uses GENMASK() to fix the problem, as this macro is already safe regarding shift overflow. Signed-off-by:
Maxime Coquelin <maxime.coquelin@st.com> Signed-off-by:
-
Arun Chandran authored
commit 15b8d2c4 upstream. In big endian mode regmap_bulk_read gives incorrect data for byte reads. This is because memcpy of a single byte from an address after full word read gives different results when endianness differs. ie. we get little-end in LE and big-end in BE. Signed-off-by:
Arun Chandran <achandran@mvista.com> Signed-off-by:
-
Vlastimil Babka authored
commit 0867a57c upstream. Since commit 077fcf11 ("mm/thp: allocate transparent hugepages on local node"), we handle THP allocations on page fault in a special way - for non-interleave memory policies, the allocation is only attempted on the node local to the current CPU, if the policy's nodemask allows the node. This is motivated by the assumption that THP benefits cannot offset the cost of remote accesses, so it's better to fallback to base pages on the local node (which might still be available, while huge pages are not due to fragmentation) than to allocate huge pages on a remote node. The nodemask check prevents us from violating e.g. MPOL_BIND policies where the local node is not among the allowed nodes. However, the current implementation can still give surprising results for the MPOL_PREFERRED policy when the preferred node is different than the current CPU's local node. In such case we should honor the preferred node and not use the local node, which is what this patch does. If hugepage allocation on the preferred node fails, we fall back to base pages and don't try other nodes, with the same motivation as is done for the local node hugepage allocations. The patch also moves the MPOL_INTERLEAVE check around to simplify the hugepage specific test. The difference can be demonstrated using in-tree transhuge-stress test on the following 2-node machine where half memory on one node was occupied to show the difference. > numactl --hardware available: 2 nodes (0-1) node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 24 25 26 27 28 29 30 31 32 33 34 35 node 0 size: 7878 MB node 0 free: 3623 MB node 1 cpus: 12 13 14 15 16 17 18 19 20 21 22 23 36 37 38 39 40 41 42 43 44 45 46 47 node 1 size: 8045 MB node 1 free: 7818 MB node distances: node 0 1 0: 10 21 1: 21 10 Before the patch: > numactl -p0 -C0 ./transhuge-stress transhuge-stress: 2.197 s/loop, 0.276 ms/page, 7249.168 MiB/s 7962 succeed, 0 failed, 1786 different pages > numactl -p0 -C12 ./transhuge-stress transhuge-stress: 2.962 s/loop, 0.372 ms/page, 5376.172 MiB/s 7962 succeed, 0 failed, 3873 different pages Number of successful THP allocations corresponds to free memory on node 0 in the first case and node 1 in the second case, i.e. -p parameter is ignored and cpu binding "wins". After the patch: > numactl -p0 -C0 ./transhuge-stress transhuge-stress: 2.183 s/loop, 0.274 ms/page, 7295.516 MiB/s 7962 succeed, 0 failed, 1760 different pages > numactl -p0 -C12 ./transhuge-stress transhuge-stress: 2.878 s/loop, 0.361 ms/page, 5533.638 MiB/s 7962 succeed, 0 failed, 1750 different pages > numactl -p1 -C0 ./transhuge-stress transhuge-stress: 4.628 s/loop, 0.581 ms/page, 3440.893 MiB/s 7962 succeed, 0 failed, 3918 different pages The -p parameter is respected regardless of cpu binding. > numactl -C0 ./transhuge-stress transhuge-stress: 2.202 s/loop, 0.277 ms/page, 7230.003 MiB/s 7962 succeed, 0 failed, 1750 different pages > numactl -C12 ./transhuge-stress transhuge-stress: 3.020 s/loop, 0.379 ms/page, 5273.324 MiB/s 7962 succeed, 0 failed, 3916 different pages Without -p parameter, hugepage restriction to CPU-local node works as before. Fixes: 077fcf11 ("mm/thp: allocate transparent hugepages on local node") Signed-off-by:
Vlastimil Babka <vbabka@suse.cz> Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by:
David Rientjes <rientjes@google.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
Larry Finger authored
commit 8a8c35fa upstream. Beginning at commit d52d3997 ("ipv6: Create percpu rt6_info"), the following INFO splat is logged: =============================== [ INFO: suspicious RCU usage. ] 4.1.0-rc7-next-20150612 #1 Not tainted ------------------------------- kernel/sched/core.c:7318 Illegal context switch in RCU-bh read-side critical section! other info that might help us debug this: rcu_scheduler_active = 1, debug_locks = 0 3 locks held by systemd/1: #0: (rtnl_mutex){+.+.+.}, at: [<ffffffff815f0c8f>] rtnetlink_rcv+0x1f/0x40 #1: (rcu_read_lock_bh){......}, at: [<ffffffff816a34e2>] ipv6_add_addr+0x62/0x540 #2: (addrconf_hash_lock){+...+.}, at: [<ffffffff816a3604>] ipv6_add_addr+0x184/0x540 stack backtrace: CPU: 0 PID: 1 Comm: systemd Not tainted 4.1.0-rc7-next-20150612 #1 Hardware name: TOSHIBA TECRA A50-A/TECRA A50-A, BIOS Version 4.20 04/17/2014 Call Trace: dump_stack+0x4c/0x6e lockdep_rcu_suspicious+0xe7/0x120 ___might_sleep+0x1d5/0x1f0 __might_sleep+0x4d/0x90 kmem_cache_alloc+0x47/0x250 create_object+0x39/0x2e0 kmemleak_alloc_percpu+0x61/0xe0 pcpu_alloc+0x370/0x630 Additional backtrace lines are truncated. In addition, the above splat is followed by several "BUG: sleeping function called from invalid context at mm/slub.c:1268" outputs. As suggested by Martin KaFai Lau, these are the clue to the fix. Routine kmemleak_alloc_percpu() always uses GFP_KERNEL for its allocations, whereas it should follow the gfp from its callers. Reviewed-by:
Catalin Marinas <catalin.marinas@arm.com> Reviewed-by:
Kamalesh Babulal <kamalesh@linux.vnet.ibm.com> Acked-by:
Martin KaFai Lau <kafai@fb.com> Signed-off-by:
Larry Finger <Larry.Finger@lwfinger.net> Cc: Martin KaFai Lau <kafai@fb.com> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by:
-
Catalin Marinas authored
commit c5f3b1a5 upstream. The kmemleak scanning thread can run for minutes. Callbacks like kmemleak_free() are allowed during this time, the race being taken care of by the object->lock spinlock. Such lock also prevents a memory block from being freed or unmapped while it is being scanned by blocking the kmemleak_free() -> ... -> __delete_object() function until the lock is released in scan_object(). When a kmemleak error occurs (e.g. it fails to allocate its metadata), kmemleak_enabled is set and __delete_object() is no longer called on freed objects. If kmemleak_scan is running at the same time, kmemleak_free() no longer waits for the object scanning to complete, allowing the corresponding memory block to be freed or unmapped (in the case of vfree()). This leads to kmemleak_scan potentially triggering a page fault. This patch separates the kmemleak_free() enabling/disabling from the overall kmemleak_enabled nob so that we can defer the disabling of the object freeing tracking until the scanning thread completed. The kmemleak_free_part() is deliberately ignored by this patch since this is only called during boot before the scanning thread started. Signed-off-by:
Vignesh Radhakrishnan <vigneshr@codeaurora.org> Tested-by:
-
Will Deacon authored
commit 6f1a6ae8 upstream. When building the kernel with a bare-metal (ELF) toolchain, the -shared option may not be passed down to collect2, resulting in silent corruption of the vDSO image (in particular, the DYNAMIC section is omitted). The effect of this corruption is that the dynamic linker fails to find the vDSO symbols and libc is instead used for the syscalls that we intended to optimise (e.g. gettimeofday). Functionally, there is no issue as the sigreturn trampoline is still intact and located by the kernel. This patch fixes the problem by explicitly passing -shared to the linker when building the vDSO. Reported-by:
Szabolcs Nagy <Szabolcs.Nagy@arm.com> Reported-by:
James Greenlaigh <james.greenhalgh@arm.com> Signed-off-by:
Will Deacon <will.deacon@arm.com> Signed-off-by:
-
