- 10 Dec, 2014 28 commits
-
-
Paul Burton authored
commit 14fa12df upstream. The save_fp_context & restore_fp_context pointers were being assigned to the wrong variables if either: - The kernel is configured for UP & runs on a system without an FPU, since b2ead528 "MIPS: Move & rename fpu_emulator_{save,restore}_context". - The kernel is configured for EVA, since ca750649 "MIPS: kernel: signal: Prevent save/restore FPU context in user memory". This would lead to FP context being clobbered incorrectly when setting up a sigcontext, then the garbage values being saved uselessly when returning from the signal. Fix by swapping the pointer assignments appropriately. Signed-off-by:
Paul Burton <paul.burton@imgtec.com> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/8230/Signed-off-by:
Ralf Baechle <ralf@linux-mips.org> Signed-off-by:
Luis Henriques <luis.henriques@canonical.com>
-
Markos Chandras authored
commit cf0a8aa0 upstream. Make use of the Config6/FLTBP bit to set the probability of a TLBWR instruction to hit the FTLB or the VTLB. A value of 0 (which may be the default value on certain cores, such as proAptiv or P5600) means that a TLBWR instruction will never hit the VTLB which leads to performance limitations since it effectively decreases the number of available TLB slots. Signed-off-by:
Markos Chandras <markos.chandras@imgtec.com> Reviewed-by:
James Hogan <james.hogan@imgtec.com> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/8368/Signed-off-by:
-
Markos Chandras authored
commit 83fd4344 upstream. Commit de8974e3 ("MIPS: asm: r4kcache: Add EVA cache flushing functions") added cache function for EVA using the cachee instruction. However, it didn't add a case for the protected_writeback_dcache_line. mips_dsemul() calls r4k_flush_cache_sigtramp() which in turn uses the protected_writeback_dcache_line() to flush the trampoline code back to memory. This used the wrong "cache" instruction leading to random userland crashes on non-FPU cores. Signed-off-by:
-
Benjamin Herrenschmidt authored
commit 415072a0 upstream. Instead of the arch specific quirk which we are deprecating Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
Benjamin Herrenschmidt authored
commit db79afa1 upstream. A number of radeon cards have a HW limitation causing them to be unable to generate the full 64-bit of address bits for MSIs. This breaks MSIs on some platforms such as POWER machines. We used to have a powerpc specific quirk to address that on a single card, but this doesn't scale very well, this is better put under control of the drivers who know precisely what a given HW revision can do. We now have a generic quirk in the PCI code. We should set it appropriately for all radeon's from the audio driver. Signed-off-by:
Takashi Iwai <tiwai@suse.de> Reviewed-by:
Alex Deucher <alexander.deucher@amd.com> Signed-off-by:
-
Benjamin Herrenschmidt authored
commit 91ed6fd2 upstream. Some radeon ASICs don't support all 64 address bits of MSIs despite advertising support for 64-bit MSIs in their configuration space. This breaks on systems such as IBM POWER7/8, where 64-bit MSIs can be assigned with some of the high address bits set. This makes use of the newly introduced "no_64bit_msi" flag in structure pci_dev to allow the MSI allocation code to fallback to 32-bit MSIs on those adapters. Signed-off-by:
-
Benjamin Herrenschmidt authored
commit f144d149 upstream. This can be set by quirks/drivers to be used by the architecture code that assigns the MSI addresses. We additionally add verification in the core MSI code that the values assigned by the architecture do satisfy the limitation in order to fail gracefully if they don't (ie. the arch hasn't been updated to deal with that quirk yet). Signed-off-by:
Bjorn Helgaas <bhelgaas@google.com> [ luis: backported to 3.16: adjusted context ] Signed-off-by:
-
Takashi Iwai authored
commit 413cbf46 upstream. AMD/ATI HDMI controller chip models, we already have a filter to lower to 32bit DMA, but the rest are supposed to be working with 64bit although the hardware doesn't really work with 63bit but only with 40 or 48bit DMA. In this patch, we take 40bit DMA for safety for the AMD/ATI controllers as the graphics drivers does. Signed-off-by:
-
Daniel Borkmann authored
commit b5b2ffc0 upstream. While working on a different issue, I noticed an annoying use after free bug on my machine when unloading the ixgbe driver: [ 8642.318797] ixgbe 0000:02:00.1: removed PHC on p2p2 [ 8642.742716] ixgbe 0000:02:00.1: complete [ 8642.743784] BUG: unable to handle kernel paging request at ffff8807d3740a90 [ 8642.744828] IP: [<ffffffffa01c77dc>] ixgbe_remove+0xfc/0x1b0 [ixgbe] [ 8642.745886] PGD 20c6067 PUD 81c1f6067 PMD 81c15a067 PTE 80000007d3740060 [ 8642.746956] Oops: 0002 [#1] SMP DEBUG_PAGEALLOC [ 8642.748039] Modules linked in: [...] [ 8642.752929] CPU: 1 PID: 1225 Comm: rmmod Not tainted 3.18.0-rc2+ #49 [ 8642.754203] Hardware name: Supermicro X10SLM-F/X10SLM-F, BIOS 1.1b 11/01/2013 [ 8642.755505] task: ffff8807e34d3fe0 ti: ffff8807b7204000 task.ti: ffff8807b7204000 [ 8642.756831] RIP: 0010:[<ffffffffa01c77dc>] [<ffffffffa01c77dc>] ixgbe_remove+0xfc/0x1b0 [ixgbe] [...] [ 8642.774335] Stack: [ 8642.775805] ffff8807ee824098 ffff8807ee824098 ffffffffa01f3000 ffff8807ee824000 [ 8642.777326] ffff8807b7207e18 ffffffff8137720f ffff8807ee824098 ffff8807ee824098 [ 8642.778848] ffffffffa01f3068 ffff8807ee8240f8 ffff8807b7207e38 ffffffff8144180f [ 8642.780365] Call Trace: [ 8642.781869] [<ffffffff8137720f>] pci_device_remove+0x3f/0xc0 [ 8642.783395] [<ffffffff8144180f>] __device_release_driver+0x7f/0xf0 [ 8642.784876] [<ffffffff814421f8>] driver_detach+0xb8/0xc0 [ 8642.786352] [<ffffffff814414a9>] bus_remove_driver+0x59/0xe0 [ 8642.787783] [<ffffffff814429d0>] driver_unregister+0x30/0x70 [ 8642.789202] [<ffffffff81375c65>] pci_unregister_driver+0x25/0xa0 [ 8642.790657] [<ffffffffa01eb38e>] ixgbe_exit_module+0x1c/0xc8e [ixgbe] [ 8642.792064] [<ffffffff810f93a2>] SyS_delete_module+0x132/0x1c0 [ 8642.793450] [<ffffffff81012c61>] ? do_notify_resume+0x61/0xa0 [ 8642.794837] [<ffffffff816d2029>] system_call_fastpath+0x12/0x17 The issue is that test_and_set_bit() done on adapter->state is being performed *after* the netdevice has been freed via free_netdev(). When netdev is being allocated on initialization time, it allocates a private area, here struct ixgbe_adapter, that resides after the net_device structure. In ixgbe_probe(), the device init routine, we set up the adapter after alloc_etherdev_mq() on the private area and add a reference for the pci_dev as well via pci_set_drvdata(). Both in the error path of ixgbe_probe(), but also on module unload when ixgbe_remove() is being called, commit 41c62843 ("ixgbe: Fix rcu warnings induced by LER") accesses adapter after free_netdev(). The patch stores the result in a bool and thus fixes above oops on my side. Fixes: 41c62843 ("ixgbe: Fix rcu warnings induced by LER") Cc: Mark Rustad <mark.d.rustad@intel.com> Signed-off-by:
Daniel Borkmann <dborkman@redhat.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Mark Rustad authored
commit 508a8c9e upstream. In ixgbe_probe, the code at label err_dma can dereference adapter when it has a NULL value. The check is there to avoid disabling a disabled device. When adapter is NULL, treat it as if the device is enabled, because it is enabled in that case. Signed-off-by:
Mark Rustad <mark.d.rustad@intel.com> Tested-by:
Phil Schmitt <phillip.j.schmitt@intel.com> Signed-off-by:
-
Vlad Yasevich authored
commit 4556dc59 upstream. IXGBE adapter seems to require that VLAN filtering be enabled if VMDQ or SRIOV are enabled. When those functions are disabled, VLAN filtering may be disabled in promiscuous mode. Prior to commit a9b8943e ("ixgbe: remove vlan_filter_disable and enable functions") The logic was correct. However, after the commit the logic got reversed and VLAN filtered in now turned on when VMDQ/SRIOV is disabled. This patch changes the condition to enable hw vlan filtered when VMDQ or SRIOV is enabled. Fixes: a9b8943e ("ixgbe: remove vlan_filter_disable and enable functions") CC: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by:
Vladislav Yasevich <vyasevic@redhat.com> Acked-by:
Emil Tantilov <emil.s.tantilov@intel.com> Tested-by:
-
Lu Baolu authored
commit a1377e53 upstream. When system is being suspended, if host device is not allowed to do wakeup, xhci_suspend() needs to clear all root port wake on bits. Otherwise, some platforms may generate spurious wakeup, even if PCI PME# is disabled. The initial commit ff8cbf25 ("xhci: clear root port wake on bits"), which also got into stable, turned out to not work correctly and had to be reverted, and is now rewritten. Signed-off-by:
Lu Baolu <baolu.lu@linux.intel.com> Suggested-by:
Alan Stern <stern@rowland.harvard.edu> Acked-by:
Mathias Nyman <mathias.nyman@linux.intel.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by:
-
Mathias Nyman authored
commit 8e71a322 upstream. If a device is halted and reuturns a STALL, then the halted endpoint needs to be cleared both on the host and device side. The host side halt is cleared by issueing a xhci reset endpoint command. The device side is cleared with a ClearFeature(ENDPOINT_HALT) request, which should be issued by the device driver if a URB reruen -EPIPE. Previously we cleared the host side halt after the device side was cleared. To make sure the host side halt is cleared in time we want to issue the reset endpoint command immedialtely when a STALL status is encountered. Otherwise we end up not following the specs and not returning -EPIPE several times in a row when trying to transfer data to a halted endpoint. Fixes: bcef3fd5 (USB: xhci: Handle errors that cause endpoint halts.) Tested-by:
Felipe Balbi <balbi@ti.com> Signed-off-by:
-
Lu Baolu authored
commit 9b41ebd3 upstream. commit ff8cbf25 ("xhci: clear root port wake on bits if controller isn't") can cause device detection error if runtime PM is enabled, and S3 wake is disabled. Revert it. https://bugzilla.kernel.org/show_bug.cgi?id=85701 This commit got into stable and should be reverted from there as well. Signed-off-by:
Dmitry Nezhevenko <dion@inhex.net> [Mathias Nyman: reword commit message] Signed-off-by:
-
Mathias Nyman authored
commit c3492dbf upstream. A halted endpoint ring must first be reset, then move the ring dequeue pointer past the problematic TRB. If we start the ring too early after reset, but before moving the dequeue pointer we will end up executing the same problematic TRB again. As we always issue a set transfer dequeue command after a reset endpoint command we can skip starting endpoint rings at reset endpoint command completion. Without this fix we end up trying to handle the same faulty TD for contol endpoints. causing timeout, and failing testusb ctrl_out write tests. Fixes: e9df17eb (USB: xhci: Correct assumptions about number of rings per endpoint.) Tested-by:
-
Hans de Goede authored
commit 8daee135 upstream. These disks have a broken uas implementation, the tag field of the status iu-s is not set properly, so we need to fall-back to usb-storage for these. Signed-off-by:
Hans de Goede <hdegoede@redhat.com> Signed-off-by:
-
Dmitry Eremin-Solenikov authored
commit ef59a20b upstream. According to the manuals I have, XScale auxiliary register should be reached with opc_2 = 1 instead of crn = 1. cpu_xscale_proc_init correctly uses c1, c0, 1 arguments, but cpu_xscale_do_suspend and cpu_xscale_do_resume use c1, c1, 0. Correct suspend/resume functions to also use c1, c0, 1. The issue was primarily noticed thanks to qemu reporing "unsupported instruction" on the pxa suspend path. Confirmed in PXA210/250 and PXA255 XScale Core manuals and in PXA270 and PXA320 Developers Guides. Harware tested by me on tosa (pxa255). Robert confirmed on pxa270 board. Tested-by:
Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by:
Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Acked-by:
Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by:
-
Christian Sünkenberg authored
commit 18990455 upstream. Intel Multi-Flex LUNs choke on REPORT SUPPORTED OPERATION CODES resulting in sd_mod hanging for several minutes on startup. The issue was introduced with WRITE SAME discovery heuristics. Fixes: 5db44863 ("[SCSI] sd: Implement support for WRITE SAME") Signed-off-by:
Christian Sünkenberg <christian.suenkenberg@hfg-karlsruhe.de> Signed-off-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
-
Maurizio Lombardi authored
commit 01a4cc4d upstream. In some cases, the fcoe_rx_list may contains multiple instances of the same skb (the so called "shared skbs"). the bnx2fc_l2_rcv thread is a loop that extracts a skb from the list, modifies (and destroys) its content and then proceed to the next one. The problem is that if the skb is shared, the remaining instances will be corrupted. The solution is to use skb_share_check() before adding the skb to the fcoe_rx_list. [ 6286.808725] ------------[ cut here ]------------ [ 6286.808729] WARNING: at include/scsi/fc_frame.h:173 bnx2fc_l2_rcv_thread+0x425/0x450 [bnx2fc]() [ 6286.808748] Modules linked in: bnx2x(-) mdio dm_service_time bnx2fc cnic uio fcoe libfcoe 8021q garp stp mrp libfc llc scsi_transport_fc scsi_tgt sg iTCO_wdt iTCO_vendor_support coretemp kvm_intel kvm crct10dif_pclmul crc32_pclmul crc32c_intel e1000e ghash_clmulni_intel aesni_intel lrw gf128mul glue_helper ablk_helper ptp cryptd hpilo serio_raw hpwdt lpc_ich pps_core ipmi_si pcspkr mfd_core ipmi_msghandler shpchp pcc_cpufreq mperf nfsd auth_rpcgss nfs_acl lockd sunrpc dm_multipath xfs libcrc32c ata_generic pata_acpi sd_mod crc_t10dif crct10dif_common mgag200 syscopyarea sysfillrect sysimgblt i2c_algo_bit ata_piix drm_kms_helper ttm drm libata i2c_core hpsa dm_mirror dm_region_hash dm_log dm_mod [last unloaded: mdio] [ 6286.808750] CPU: 3 PID: 1304 Comm: bnx2fc_l2_threa Not tainted 3.10.0-121.el7.x86_64 #1 [ 6286.808750] Hardware name: HP ProLiant DL120 G7, BIOS J01 07/01/2013 [ 6286.808752] 0000000000000000 000000000b36e715 ffff8800deba1e00 ffffffff815ec0ba [ 6286.808753] ffff8800deba1e38 ffffffff8105dee1 ffffffffa05618c0 ffff8801e4c81888 [ 6286.808754] ffffe8ffff663868 ffff8801f402b180 ffff8801f56bc000 ffff8800deba1e48 [ 6286.808754] Call Trace: [ 6286.808759] [<ffffffff815ec0ba>] dump_stack+0x19/0x1b [ 6286.808762] [<ffffffff8105dee1>] warn_slowpath_common+0x61/0x80 [ 6286.808763] [<ffffffff8105e00a>] warn_slowpath_null+0x1a/0x20 [ 6286.808765] [<ffffffffa054f415>] bnx2fc_l2_rcv_thread+0x425/0x450 [bnx2fc] [ 6286.808767] [<ffffffffa054eff0>] ? bnx2fc_disable+0x90/0x90 [bnx2fc] [ 6286.808769] [<ffffffff81085aef>] kthread+0xcf/0xe0 [ 6286.808770] [<ffffffff81085a20>] ? kthread_create_on_node+0x140/0x140 [ 6286.808772] [<ffffffff815fc76c>] ret_from_fork+0x7c/0xb0 [ 6286.808773] [<ffffffff81085a20>] ? kthread_create_on_node+0x140/0x140 [ 6286.808774] ---[ end trace c6cdb939184ccb4e ]--- Signed-off-by:
Maurizio Lombardi <mlombard@redhat.com> Acked-by:
Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by:
-
Trond Myklebust authored
commit c6c15e1e upstream. The currect code for nfsd41_cb_get_slot() and nfsd4_cb_done() has no locking in order to guarantee atomicity, and so allows for races of the form. Task 1 Task 2 ====== ====== if (test_and_set_bit(0) != 0) { clear_bit(0) rpc_wake_up_next(queue) rpc_sleep_on(queue) return false; } This patch breaks the race condition by adding a retest of the bit after the call to rpc_sleep_on(). Signed-off-by:
Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by:
J. Bruce Fields <bfields@redhat.com> Signed-off-by:
-
Trond Myklebust authored
commit 093a1468 upstream. Both xprt_lookup_rqst() and xprt_complete_rqst() require that you take the transport lock in order to avoid races with xprt_transmit(). Signed-off-by:
Jeff Layton <jlayton@primarydata.com> Signed-off-by:
-
Christoph Hellwig authored
commit 6d0ba043 upstream. Even when security labels are disabled we support at least the same attributes as v4.1. Signed-off-by:
-
Johan Hovold authored
commit 75bcbf29 upstream. Fix reporting of overrun errors, which should only be reported once using the inserted null character. Fixes: 6b8f1ca5 ("USB: ssu100: set tty_flags in ssu100_process_packet") Signed-off-by:
Johan Hovold <johan@kernel.org> Signed-off-by:
-
Johan Hovold authored
commit 855515a6 upstream. Fix reporting of overrun errors, which are not associated with a character. Instead insert a null character and report only once. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
-
Johan Hovold authored
commit 5d1678a3 upstream. Fix handling of TTY error flags, which are not bitmasks and must specifically not be ORed together as this prevents the line discipline from recognising them. Also insert null characters when reporting overrun errors as these are not associated with the received character. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
-
Troy Clark authored
commit 204ec6e0 upstream. Add PIDs for new Matrix Orbital GTT series products. Signed-off-by:
Troy Clark <tclark@matrixorbital.ca> [johan: shorten commit message ] Signed-off-by:
-
Qipan Li authored
commit 9c4b19a0 upstream. commit 8c328a26 ("spi: sirf: Avoid duplicate code in various bits_per_word cases") is wrong in setting data width register of fifo is not right, it should use sspi->word_width >> 1 to set related bits. According to hardware spec, the mapping between register value and data width: 0 - byte 1 - WORD 2 - DWORD Fixes: 8c328a26 ("spi: sirf: Avoid duplicate code in various bits_per_word cases") is wrong in setting data width register of Signed-off-by:
Qipan Li <Qipan.Li@csr.com> Signed-off-by:
Barry Song <Baohua.Song@csr.com> Signed-off-by:
Mark Brown <broonie@kernel.org> Signed-off-by:
-
James Hogan authored
commit e6d5e7d9 upstream. Commit 79c6ab50 (clk: divider: add CLK_DIVIDER_READ_ONLY flag) in v3.16 introduced the CLK_DIVIDER_READ_ONLY flag which caused the recalc_rate() and round_rate() clock callbacks to be omitted. However using this flag has the unfortunate side effect of causing the clock recalculation code when a clock rate change is attempted to always treat it as a pass-through clock, i.e. with a fixed divide of 1, which may not be the case. Child clock rates are then recalculated using the wrong parent rate. Therefore instead of dropping the recalc_rate() and round_rate() callbacks, alter clk_divider_bestdiv() to always report the current divider as the best divider so that it is never altered. For me the read only clock was the system clock, which divided the PLL rate by 2, from which both the UART and the SPI clocks were divided. Initial setting of the UART rate set it correctly, but when the SPI clock was set, the other child clocks were miscalculated. The UART clock was recalculated using the PLL rate as the parent rate, resulting in a UART new_rate of double what it should be, and a UART which spewed forth garbage when the rate changes were propagated. Signed-off-by:
Haojian Zhuang <haojian.zhuang@gmail.com> Signed-off-by:
Michael Turquette <mturquette@linaro.org> [ luis: backported to 3.16: - drop changes to drivers/clk/rockchip/clk.c ] Signed-off-by:
-
- 04 Dec, 2014 10 commits
-
-
Charles Keepax authored
commit c1aefbdd upstream. We can only use page_address on memory that has been mapped using kmap, when the buffer passed to the SPI has been allocated by vmalloc the page has not necessarily been mapped through kmap. This means sometimes page_address will return NULL causing the pointer we pass to sg_set_buf to be invalid. As we only call page_address so that we can pass a virtual address to sg_set_buf which will then immediately call virt_to_page on it, fix this by calling sg_set_page directly rather then relying on the sg_set_buf helper. Signed-off-by:
Charles Keepax <ckeepax@opensource.wolfsonmicro.com> Signed-off-by:
-
Ben Sagal authored
commit bce4f9e7 upstream. The LEN2006 Synaptics touchpad (as found in Thinkpad E540) returns wrong min max values. touchpad-edge-detector output: > Touchpad SynPS/2 Synaptics TouchPad on /dev/input/event6 > Move one finger around the touchpad to detect the actual edges > Kernel says: x [1472..5674], y [1408..4684] > Touchpad sends: x [1264..5675], y [1171..4688] Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=88211Signed-off-by:
Binyamin Sagal <bensagal@gmail.com> Signed-off-by:
Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by:
-
Axel Lin authored
commit fbbba1f8 upstream. The mcb_device_id table is supposed to be zero-terminated. Signed-off-by:
Axel Lin <axel.lin@ingics.com> Signed-off-by:
Jonathan Cameron <jic23@kernel.org> Signed-off-by:
-
Cristina Ciocan authored
commit ccf54555 upstream. The direction field is set on 7 bits, thus we need to AND it with 0111 111 mask in order to retrieve it, that is 0x7F, not 0xCF as it is now. Fixes: ade7ef7b (staging:iio: Differential channel handling) Signed-off-by:
Cristina Ciocan <cristina.ciocan@intel.com> Signed-off-by:
-
Preston Fick authored
commit ffcfe30e upstream. Signed-off-by:
Preston Fick <pffick@gmail.com> Signed-off-by:
-
Thor Thayer authored
commit 0a8727e6 upstream. An IOCTL call that calls spi_setup() and then dw_spi_setup() will overwrite the persisted last transfer speed. On each transfer, the SPI speed is compared to the last transfer speed to determine if the clock divider registers need to be updated (did the speed change?). This bug was observed with the spidev driver using spi-config to update the max transfer speed. This fix: Don't overwrite the persisted last transaction clock speed when updating the SPI parameters in dw_spi_setup(). On the next transaction, the new speed won't match the persisted last speed and the hardware registers will be updated. On initialization, the persisted last transaction clock speed will be 0 but will be updated after the first SPI transaction. Move zeroed clock divider check into clock change test because chip->clk_div is zero on startup and would cause a divide-by-zero error. The calculation was wrong as well (can't support odd #). Reported-by:
Vlastimil Setka <setka@vsis.cz> Signed-off-by:
Thor Thayer <tthayer@opensource.altera.com> Signed-off-by:
-
Thomas Petazzoni authored
commit 2eb04ae0 upstream. There is a missing of_node_put() to decrement the device_node reference counter after a of_find_matching_node() in coherency_init(). Fixes: 501f928e ("ARM: mvebu: add a coherency_available() call") Signed-off-by:
Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Acked-by:
Gregory CLEMENT <gregory.clement@free-electrons.com> Link: https://lkml.kernel.org/r/1414423955-5933-4-git-send-email-thomas.petazzoni@free-electrons.comSigned-off-by:
Jason Cooper <jason@lakedaemon.net> Signed-off-by:
-
Gu Zheng authored
commit 835f252c upstream. https://bugzilla.kernel.org/show_bug.cgi?id=86831 Markus reported that when shutting down mysqld (with AIO support, on a ext3 formatted Harddrive) leads to a negative number of dirty pages (underrun to the counter). The negative number results in a drastic reduction of the write performance because the page cache is not used, because the kernel thinks it is still 2 ^ 32 dirty pages open. Add a warn trace in __dec_zone_state will catch this easily: static inline void __dec_zone_state(struct zone *zone, enum zone_stat_item item) { atomic_long_dec(&zone->vm_stat[item]); + WARN_ON_ONCE(item == NR_FILE_DIRTY && atomic_long_read(&zone->vm_stat[item]) < 0); atomic_long_dec(&vm_stat[item]); } [ 21.341632] ------------[ cut here ]------------ [ 21.346294] WARNING: CPU: 0 PID: 309 at include/linux/vmstat.h:242 cancel_dirty_page+0x164/0x224() [ 21.355296] Modules linked in: wutbox_cp sata_mv [ 21.359968] CPU: 0 PID: 309 Comm: kworker/0:1 Not tainted 3.14.21-WuT #80 [ 21.366793] Workqueue: events free_ioctx [ 21.370760] [<c0016a64>] (unwind_backtrace) from [<c0012f88>] (show_stack+0x20/0x24) [ 21.378562] [<c0012f88>] (show_stack) from [<c03f8ccc>] (dump_stack+0x24/0x28) [ 21.385840] [<c03f8ccc>] (dump_stack) from [<c0023ae4>] (warn_slowpath_common+0x84/0x9c) [ 21.393976] [<c0023ae4>] (warn_slowpath_common) from [<c0023bb8>] (warn_slowpath_null+0x2c/0x34) [ 21.402800] [<c0023bb8>] (warn_slowpath_null) from [<c00c0688>] (cancel_dirty_page+0x164/0x224) [ 21.411524] [<c00c0688>] (cancel_dirty_page) from [<c00c080c>] (truncate_inode_page+0x8c/0x158) [ 21.420272] [<c00c080c>] (truncate_inode_page) from [<c00c0a94>] (truncate_inode_pages_range+0x11c/0x53c) [ 21.429890] [<c00c0a94>] (truncate_inode_pages_range) from [<c00c0f6c>] (truncate_pagecache+0x88/0xac) [ 21.439252] [<c00c0f6c>] (truncate_pagecache) from [<c00c0fec>] (truncate_setsize+0x5c/0x74) [ 21.447731] [<c00c0fec>] (truncate_setsize) from [<c013b3a8>] (put_aio_ring_file.isra.14+0x34/0x90) [ 21.456826] [<c013b3a8>] (put_aio_ring_file.isra.14) from [<c013b424>] (aio_free_ring+0x20/0xcc) [ 21.465660] [<c013b424>] (aio_free_ring) from [<c013b4f4>] (free_ioctx+0x24/0x44) [ 21.473190] [<c013b4f4>] (free_ioctx) from [<c003d8d8>] (process_one_work+0x134/0x47c) [ 21.481132] [<c003d8d8>] (process_one_work) from [<c003e988>] (worker_thread+0x130/0x414) [ 21.489350] [<c003e988>] (worker_thread) from [<c00448ac>] (kthread+0xd4/0xec) [ 21.496621] [<c00448ac>] (kthread) from [<c000ec18>] (ret_from_fork+0x14/0x20) [ 21.503884] ---[ end trace 79c4bf42c038c9a1 ]--- The cause is that we set the aio ring file pages as *DIRTY* via SetPageDirty (bypasses the VFS dirty pages increment) when init, and aio fs uses *default_backing_dev_info* as the backing dev, which does not disable the dirty pages accounting capability. So truncating aio ring file will contribute to accounting dirty pages (VFS dirty pages decrement), then error occurs. The original goal is keeping these pages in memory (can not be reclaimed or swapped) in life-time via marking it dirty. But thinking more, we have already pinned pages via elevating the page's refcount, which can already achieve the goal, so the SetPageDirty seems unnecessary. In order to fix the issue, using the __set_page_dirty_no_writeback instead of the nop .set_page_dirty, and dropped the SetPageDirty (don't manually set the dirty flags, don't disable set_page_dirty(), rely on default behaviour). With the above change, the dirty pages accounting can work well. But as we known, aio fs is an anonymous one, which should never cause any real write-back, we can ignore the dirty pages (write back) accounting by disabling the dirty pages (write back) accounting capability. So we introduce an aio private backing dev info (disabled the ACCT_DIRTY/WRITEBACK/ACCT_WB capabilities) to replace the default one. Reported-by:
Markus Königshaus <m.koenigshaus@wut.de> Signed-off-by:
Gu Zheng <guz.fnst@cn.fujitsu.com> Acked-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Benjamin LaHaise <bcrl@kvack.org> Signed-off-by:
-
Steven Capper authored
commit ded94779 upstream. For LPAE, we have the following means for encoding writable or dirty ptes: L_PTE_DIRTY L_PTE_RDONLY !pte_dirty && !pte_write 0 1 !pte_dirty && pte_write 0 1 pte_dirty && !pte_write 1 1 pte_dirty && pte_write 1 0 So we can't distinguish between writeable clean ptes and read only ptes. This can cause problems with ptes being incorrectly flagged as read only when they are writeable but not dirty. This patch renumbers L_PTE_RDONLY from AP[2] to a software bit #58, and adds additional logic to set AP[2] whenever the pte is read only or not dirty. That way we can distinguish between clean writeable ptes and read only ptes. HugeTLB pages will use this new logic automatically. We need to add some logic to Transparent HugePages to ensure that they correctly interpret the revised pgprot permissions (L_PTE_RDONLY has moved and no longer matches PMD_SECT_AP2). In the process of revising THP, the names of the PMD software bits have been prefixed with L_ to make them easier to distinguish from their hardware bit counterparts. Signed-off-by:
Steve Capper <steve.capper@linaro.org> Reviewed-by:
Will Deacon <will.deacon@arm.com> Signed-off-by:
-
Steven Capper authored
commit f2950706 upstream. Long descriptors on ARM are 64 bits, and some pte functions such as pte_dirty return a bitwise-and of a flag with the pte value. If the flag to be tested resides in the upper 32 bits of the pte, then we run into the danger of the result being dropped if downcast. For example: gather_stats(page, md, pte_dirty(*pte), 1); where pte_dirty(*pte) is downcast to an int. This patch introduces a new macro pte_isset which performs the bitwise and, then performs a double logical invert (where needed) to ensure predictable downcasting. The logical inverse pte_isclear is also introduced. Equivalent pmd functions for Transparent HugePages have also been added. Signed-off-by:
-
- 02 Dec, 2014 1 commit
-
-
David Jeffery authored
commit 92a56555 upstream. If a SIGKILL is sent to a task waiting in __nfs_iocounter_wait, it will busy-wait or soft lockup in its while loop. nfs_wait_bit_killable won't sleep, and the loop won't exit on the error return. Stop the busy-wait by breaking out of the loop when nfs_wait_bit_killable returns an error. Signed-off-by:
David Jeffery <djeffery@redhat.com> Signed-off-by:
Kamal Mostafa <kamal@canonical.com> Signed-off-by:
-
- 01 Dec, 2014 1 commit
-
-
Richard Leitner authored
commit 0224ec9e upstream. Fix the format string for serio device name generation to avoid negative device numbers when the id exceeds the maximum signed integer value. Signed-off-by:
Richard Leitner <richard.leitner@skidata.com> Signed-off-by:
-
