- 25 Jun, 2019 18 commits
-
-
Mike Marciniszyn authored
commit cc78076a upstream. The qp priv rcd pointer doesn't match the context being used for verbs causing issues when 9B and kdeth packets are processed by different receive contexts and hence different CPUs. When running on different CPUs the following panic can occur: WARNING: CPU: 3 PID: 2584 at lib/list_debug.c:59 __list_del_entry+0xa1/0xd0 list_del corruption. prev->next should be ffff9a7ac31f7a30, but was ffff9a7c3bc89230 CPU: 3 PID: 2584 Comm: z_wr_iss Kdump: loaded Tainted: P OE ------------ 3.10.0-862.2.3.el7_lustre.x86_64 #1 Call Trace: <IRQ> [<ffffffffb7b0d78e>] dump_stack+0x19/0x1b [<ffffffffb74916d8>] __warn+0xd8/0x100 [<ffffffffb749175f>] warn_slowpath_fmt+0x5f/0x80 [<ffffffffb7768671>] __list_del_entry+0xa1/0xd0 [<ffffffffc0c7a945>] process_rcv_qp_work+0xb5/0x160 [hfi1] [<ffffffffc0c7bc2b>] handle_receive_interrupt_nodma_rtail+0x20b/0x2b0 [hfi1] [<ffffffffc0c70683>] receive_context_interrupt+0x23/0x40 [hfi1] [<ffffffffb7540a94>] __handle_irq_event_percpu+0x44/0x1c0 [<ffffffffb7540c42>] handle_irq_event_percpu+0x32/0x80 [<ffffffffb7540ccc>] handle_irq_event+0x3c/0x60 [<ffffffffb7543a1f>] handle_edge_irq+0x7f/0x150 [<ffffffffb742d504>] handle_irq+0xe4/0x1a0 [<ffffffffb7b23f7d>] do_IRQ+0x4d/0xf0 [<ffffffffb7b16362>] common_interrupt+0x162/0x162 <EOI> [<ffffffffb775a326>] ? memcpy+0x6/0x110 [<ffffffffc109210d>] ? abd_copy_from_buf_off_cb+0x1d/0x30 [zfs] [<ffffffffc10920f0>] ? abd_copy_to_buf_off_cb+0x30/0x30 [zfs] [<ffffffffc1093257>] abd_iterate_func+0x97/0x120 [zfs] [<ffffffffc10934d9>] abd_copy_from_buf_off+0x39/0x60 [zfs] [<ffffffffc109b828>] arc_write_ready+0x178/0x300 [zfs] [<ffffffffb7b11032>] ? mutex_lock+0x12/0x2f [<ffffffffb7b11032>] ? mutex_lock+0x12/0x2f [<ffffffffc1164d05>] zio_ready+0x65/0x3d0 [zfs] [<ffffffffc04d725e>] ? tsd_get_by_thread+0x2e/0x50 [spl] [<ffffffffc04d1318>] ? taskq_member+0x18/0x30 [spl] [<ffffffffc115ef22>] zio_execute+0xa2/0x100 [zfs] [<ffffffffc04d1d2c>] taskq_thread+0x2ac/0x4f0 [spl] [<ffffffffb74cee80>] ? wake_up_state+0x20/0x20 [<ffffffffc115ee80>] ? zio_taskq_member.isra.7.constprop.10+0x80/0x80 [zfs] [<ffffffffc04d1a80>] ? taskq_thread_spawn+0x60/0x60 [spl] [<ffffffffb74bae31>] kthread+0xd1/0xe0 [<ffffffffb74bad60>] ? insert_kthread_work+0x40/0x40 [<ffffffffb7b1f5f7>] ret_from_fork_nospec_begin+0x21/0x21 [<ffffffffb74bad60>] ? insert_kthread_work+0x40/0x40 Fix by reading the map entry in the same manner as the hardware so that the kdeth and verbs contexts match. Cc: <stable@vger.kernel.org> Fixes: 5190f052 ("IB/hfi1: Allow the driver to initialize QP priv struct") Reviewed-by:
Kaike Wan <kaike.wan@intel.com> Signed-off-by:
Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by:
Dennis Dalessandro <dennis.dalessandro@intel.com> Signed-off-by:
Jason Gunthorpe <jgg@mellanox.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Mike Marciniszyn authored
commit cf131a81 upstream. Heavy contention of the sde flushlist_lock can cause hard lockups at extreme scale when the flushing logic is under stress. Mitigate by replacing the item at a time copy to the local list with an O(1) list_splice_init() and using the high priority work queue to do the flushes. Fixes: 77241056 ("IB/hfi1: add driver files") Cc: <stable@vger.kernel.org> Reviewed-by:
Doug Ledford <dledford@redhat.com> Signed-off-by:
-
Mike Marciniszyn authored
commit da9de5f8 upstream. The call to sdma_progress() is called outside the wait lock. In this case, there is a race condition where sdma_progress() can return false and the sdma_engine can idle. If that happens, there will be no more sdma interrupts to cause the wakeup and the user_sdma xmit will hang. Fix by moving the lock to enclose the sdma_progress() call. Also, delete busycount. The need for this was removed by: commit bcad2913 ("IB/hfi1: Serve the most starved iowait entry first") Cc: <stable@vger.kernel.org> Fixes: 77241056 ("IB/hfi1: add driver files") Reviewed-by:
Gary Leshner <Gary.S.Leshner@intel.com> Signed-off-by:
-
Kaike Wan authored
commit 5f90677e upstream. The opcode range for fault injection from user should be validated before it is applied to the fault->opcodes[] bitmap to avoid out-of-bound error. Cc: <stable@vger.kernel.org> Fixes: a74d5307 ("IB/hfi1: Rework fault injection machinery") Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by:
-
Ronnie Sahlberg authored
commit 61cabc7b upstream. We can not hold the GlobalMid_Lock spinlock during the dfs processing in cifs_reconnect since it invokes things that may sleep and thus trigger : BUG: sleeping function called from invalid context at kernel/locking/rwsem.c:23 Thus we need to drop the spinlock during this code block. RHBZ: 1716743 Cc: stable@vger.kernel.org Signed-off-by:
Ronnie Sahlberg <lsahlber@redhat.com> Acked-by:
Pavel Shilovsky <pshilov@microsoft.com> Signed-off-by:
Steve French <stfrench@microsoft.com> Signed-off-by:
-
Ronnie Sahlberg authored
commit 487317c9 upstream. We can not depend on the tcon->open_file_lock here since in multiuser mode we may have the same file/inode open via multiple different tcons. The current code is race prone and will crash if one user deletes a file at the same time a different user opens/create the file. To avoid this we need to have a spinlock attached to the inode and not the tcon. RHBZ: 1580165 CC: Stable <stable@vger.kernel.org> Signed-off-by:
-
Mathias Nyman authored
commit b8c3b718 upstream. A USB3 device needs to be reset and re-enumarated if the port it connects to goes to a error state, with link state inactive. There is no use in trying to recover failed transactions by resetting endpoints at this stage. Tests show that in rare cases, after multiple endpoint resets of a roothub port the whole host controller might stop completely. Several retries to recover from transaction error can happen as it can take a long time before the hub thread discovers the USB3 port error and inactive link. We can't reliably detect the port error from slot or endpoint context due to a limitation in xhci, see xhci specs section 4.8.3: "There are several cases where the EP State field in the Output Endpoint Context may not reflect the current state of an endpoint" and "Software should maintain an accurate value for EP State, by tracking it with an internal variable that is driven by Events and Doorbell accesses" Same appears to be true for slot state. set a flag to the corresponding slot if a USB3 roothub port link goes inactive to prevent both queueing new URBs and resetting endpoints. Reported-by:
Rapolu Chiranjeevi <chiranjeevi.rapolu@intel.com> Tested-by:
Mathias Nyman <mathias.nyman@linux.intel.com> Signed-off-by:
-
Mathias Nyman authored
commit ddd57980 upstream. USB 3.2 capability in a host can be detected from the xHCI Supported Protocol Capability major and minor revision fields. If major is 0x3 and minor 0x20 then the host is USB 3.2 capable. For USB 3.2 capable hosts set the root hub lane count to 2. The Major Revision and Minor Revision fields contain a BCD version number. The value of the Major Revision field is JJh and the value of the Minor Revision field is MNh for version JJ.M.N, where JJ = major revision number, M - minor version number, N = sub-minor version number, e.g. version 3.1 is represented with a value of 0310h. Also fix the extra whitespace printed out when announcing regular SuperSpeed hosts. Cc: <stable@vger.kernel.org> # v4.18+ Signed-off-by:
-
Peter Chen authored
commit c19dffc0 upstream. An endpoint conflict occurs when the USB is working in device mode during an isochronous communication. When the endpointA IN direction is an isochronous IN endpoint, and the host sends an IN token to endpointA on another device, then the OUT transaction may be missed regardless the OUT endpoint number. Generally, this occurs when the device is connected to the host through a hub and other devices are connected to the same hub. The affected OUT endpoint can be either control, bulk, isochronous, or an interrupt endpoint. After the OUT endpoint is primed, if an IN token to the same endpoint number on another device is received, then the OUT endpoint may be unprimed (cannot be detected by software), which causes this endpoint to no longer respond to the host OUT token, and thus, no corresponding interrupt occurs. There is no good workaround for this issue, the only thing the software could do is numbering isochronous IN from the highest endpoint since we have observed most of device number endpoint from the lowest. Cc: <stable@vger.kernel.org> #v3.14+ Cc: Fabio Estevam <festevam@gmail.com> Cc: Greg KH <gregkh@linuxfoundation.org> Cc: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Cc: Jun Li <jun.li@nxp.com> Signed-off-by:
Peter Chen <peter.chen@nxp.com> Signed-off-by:
-
Stanley Chu authored
commit 24e2e7a1 upstream. UFS runtime suspend can be triggered after pm_runtime_enable() is invoked in ufshcd_pltfrm_init(). However if the first runtime suspend is triggered before binding ufs_hba structure to ufs device structure via platform_set_drvdata(), then UFS runtime suspend will be no longer triggered in the future because its dev->power.runtime_error was set in the first triggering and does not have any chance to be cleared. To be more clear, dev->power.runtime_error is set if hba is NULL in ufshcd_runtime_suspend() which returns -EINVAL to rpm_callback() where dev->power.runtime_error is set as -EINVAL. In this case, any future rpm_suspend() for UFS device fails because rpm_check_suspend_allowed() fails due to non-zero dev->power.runtime_error. To resolve this issue, make sure the first UFS runtime suspend get valid "hba" in ufshcd_runtime_suspend(): Enable UFS runtime PM only after hba is successfully bound to UFS device structure. Fixes: 62694735 ([SCSI] ufs: Add runtime PM support for UFS host controller driver) Cc: stable@vger.kernel.org Signed-off-by:
Stanley Chu <stanley.chu@mediatek.com> Reviewed-by:
Avri Altman <avri.altman@wdc.com> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by:
-
Ulf Hansson authored
commit 83293386 upstream. Processing of SDIO IRQs must obviously be prevented while the card is system suspended, otherwise we may end up trying to communicate with an uninitialized SDIO card. Reports throughout the years shows that this is not only a theoretical problem, but a real issue. So, let's finally fix this problem, by keeping track of the state for the card and bail out before processing the SDIO IRQ, in case the card is suspended. Cc: stable@vger.kernel.org Reported-by:
Douglas Anderson <dianders@chromium.org> Tested-by:
Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by:
-
Douglas Anderson authored
commit b4c9f938 upstream. We want SDIO drivers to be able to temporarily stop retuning when the driver knows that the SDIO card is not in a state where retuning will work (maybe because the card is asleep). We'll move the relevant functions to a place where drivers can call them. Cc: stable@vger.kernel.org #v4.18+ Signed-off-by:
Adrian Hunter <adrian.hunter@intel.com> Acked-by:
Kalle Valo <kvalo@codeaurora.org> Signed-off-by:
-
Douglas Anderson authored
commit 0a55f4ab upstream. Normally when the MMC core sees an "-EILSEQ" error returned by a host controller then it will trigger a retuning of the card. This is generally a good idea. However, if a command is expected to sometimes cause transfer errors then these transfer errors shouldn't cause a re-tuning. This re-tuning will be a needless waste of time. One example case where a transfer is expected to cause errors is when transitioning between idle (sometimes referred to as "sleep" in Broadcom code) and active state on certain Broadcom WiFi SDIO cards. Specifically if the card was already transitioning between states when the command was sent it could cause an error on the SDIO bus. Let's add an API that the SDIO function drivers can call that will temporarily disable the auto-tuning functionality. Then we can add a call to this in the Broadcom WiFi driver and any other driver that might have similar needs. NOTE: this makes the assumption that the card is already tuned well enough that it's OK to disable the auto-retuning during one of these error-prone situations. Presumably the driver code performing the error-prone transfer knows how to recover / retry from errors. ...and after we can get back to a state where transfers are no longer error-prone then we can enable the auto-retuning again. If we truly find ourselves in a case where the card needs to be retuned sometimes to handle one of these error-prone transfers then we can always try a few transfers first without auto-retuning and then re-try with auto-retuning if the first few fail. Without this change on rk3288-veyron-minnie I periodically see this in the logs of a machine just sitting there idle: dwmmc_rockchip ff0d0000.dwmmc: Successfully tuned phase to XYZ Cc: stable@vger.kernel.org #v4.18+ Signed-off-by:
-
jjian zhou authored
commit 20314ce3 upstream. If cmd19 timeout or response crcerr occurs during execute_tuning(), it need invoke msdc_reset_hw(). Otherwise SDIO IRQ can't be detected. Signed-off-by:
jjian zhou <jjian.zhou@mediatek.com> Signed-off-by:
Chaotian Jing <chaotian.jing@mediatek.com> Signed-off-by:
Yong Mao <yong.mao@mediatek.com> Fixes: 5215b2e9 ("mmc: mediatek: Add MMC_CAP_SDIO_IRQ support") Cc: stable@vger.kernel.org Signed-off-by:
-
jjian zhou authored
commit 8a5df8ac upstream. SDIO IRQ is triggered by low level. It need disable SDIO IRQ detected function. Otherwise the interrupt register can't be cleared. It will process the interrupt more. Signed-off-by:
-
Wolfram Sang authored
commit 97bf85b6 upstream. Our HW engineers informed us that HS400 is not working on these SoC revisions. Fixes: 0f4e2054 ("mmc: renesas_sdhi: disable HS400 on H3 ES1.x and M3-W ES1.[012]") Signed-off-by:
Wolfram Sang <wsa+renesas@sang-engineering.com> Reviewed-by:
Geert Uytterhoeven <geert+renesas@glider.be> Reviewed-by:
Fabrizio Castro <fabrizio.castro@bp.renesas.com> Reviewed-by:
Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> Cc: stable@vger.kernel.org Signed-off-by:
-
Raul E Rangel authored
commit 0f7b79a4 upstream. The O2Micro controller only supports tuning at 4-bits. So the host driver needs to change the bus width while tuning and then set it back when done. There was a bug in the original implementation in that mmc->ios.bus_width also wasn't updated. Thus setting the incorrect blocksize in sdhci_send_tuning which results in a tuning failure. Signed-off-by:
Raul E Rangel <rrangel@chromium.org> Fixes: 0086fc21 ("mmc: sdhci: Add support for O2 hardware tuning") Acked-by:
-
Miguel Ojeda authored
commit 0c97bf86 upstream. Starting with GCC 9, -Warray-bounds detects cases when memset is called starting on a member of a struct but the size to be cleared ends up writing over further members. Such a call happens in the trace code to clear, at once, all members after and including `seq` on struct trace_iterator: In function 'memset', inlined from 'ftrace_dump' at kernel/trace/trace.c:8914:3: ./include/linux/string.h:344:9: warning: '__builtin_memset' offset [8505, 8560] from the object at 'iter' is out of the bounds of referenced subobject 'seq' with type 'struct trace_seq' at offset 4368 [-Warray-bounds] 344 | return __builtin_memset(p, c, size); | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ In order to avoid GCC complaining about it, we compute the address ourselves by adding the offsetof distance instead of referring directly to the member. Since there are two places doing this clear (trace.c and trace_kdb.c), take the chance to move the workaround into a single place in the internal header. Link: http://lkml.kernel.org/r/20190523124535.GA12931@gmail.comSigned-off-by:
Miguel Ojeda <miguel.ojeda.sandonis@gmail.com> [ Removed unnecessary parenthesis around "iter" ] Signed-off-by:
Steven Rostedt (VMware) <rostedt@goodmis.org> Signed-off-by:
-
- 22 Jun, 2019 22 commits
-
-
Greg Kroah-Hartman authored
-
Eric Dumazet authored
commit b6653b36 upstream. tcp_fragment() might be called for skbs in the write queue. Memory limits might have been exceeded because tcp_sendmsg() only checks limits at full skb (64KB) boundaries. Therefore, we need to make sure tcp_fragment() wont punish applications that might have setup very low SO_SNDBUF values. Fixes: f070ef2a ("tcp: tcp_fragment() should apply sane memory limits") Signed-off-by:
Eric Dumazet <edumazet@google.com> Reported-by:
Christoph Paasch <cpaasch@apple.com> Tested-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Greg Kroah-Hartman authored
-
Andrea Arcangeli authored
commit 59ea6d06 upstream. When fixing the race conditions between the coredump and the mmap_sem holders outside the context of the process, we focused on mmget_not_zero()/get_task_mm() callers in 04f5866e ("coredump: fix race condition between mmget_not_zero()/get_task_mm() and core dumping"), but those aren't the only cases where the mmap_sem can be taken outside of the context of the process as Michal Hocko noticed while backporting that commit to older -stable kernels. If mmgrab() is called in the context of the process, but then the mm_count reference is transferred outside the context of the process, that can also be a problem if the mmap_sem has to be taken for writing through that mm_count reference. khugepaged registration calls mmgrab() in the context of the process, but the mmap_sem for writing is taken later in the context of the khugepaged kernel thread. collapse_huge_page() after taking the mmap_sem for writing doesn't modify any vma, so it's not obvious that it could cause a problem to the coredump, but it happens to modify the pmd in a way that breaks an invariant that pmd_trans_huge_lock() relies upon. collapse_huge_page() needs the mmap_sem for writing just to block concurrent page faults that call pmd_trans_huge_lock(). Specifically the invariant that "!pmd_trans_huge()" cannot become a "pmd_trans_huge()" doesn't hold while collapse_huge_page() runs. The coredump will call __get_user_pages() without mmap_sem for reading, which eventually can invoke a lockless page fault which will need a functional pmd_trans_huge_lock(). So collapse_huge_page() needs to use mmget_still_valid() to check it's not running concurrently with the coredump... as long as the coredump can invoke page faults without holding the mmap_sem for reading. This has "Fixes: khugepaged" to facilitate backporting, but in my view it's more a bug in the coredump code that will eventually have to be rewritten to stop invoking page faults without the mmap_sem for reading. So the long term plan is still to drop all mmget_still_valid(). Link: http://lkml.kernel.org/r/20190607161558.32104-1-aarcange@redhat.com Fixes: ba76149f ("thp: khugepaged") Signed-off-by:
Andrea Arcangeli <aarcange@redhat.com> Reported-by:
Michal Hocko <mhocko@suse.com> Acked-by:
Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Jann Horn <jannh@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: Mike Rapoport <rppt@linux.vnet.ibm.com> Cc: Mike Kravetz <mike.kravetz@oracle.com> Cc: Peter Xu <peterx@redhat.com> Cc: Jason Gunthorpe <jgg@mellanox.com> Cc: <stable@vger.kernel.org> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
Sagi Grimberg authored
commit 64861993 upstream. When the controller supports less queues than requested, we should make sure that queue mapping does the right thing and not assume that all queues are available. This fixes a crash when the controller supports less queues than requested. The rules are: 1. if no write queues are requested, we assign the available queues to the default queue map. The default and read queue maps share the existing queues. 2. if write queues are requested: - first make sure that read queue map gets the requested nr_io_queues count - then grant the default queue map the minimum between the requested nr_write_queues and the remaining queues. If there are no available queues to dedicate to the default queue map, fallback to (1) and share all the queues in the existing queue map. Also, provide a log indication on how we constructed the different queue maps. Reported-by:
Harris, James R <james.r.harris@intel.com> Tested-by:
Roy Shterman <roys@lightbitslabs.com> Signed-off-by:
Sagi Grimberg <sagi@grimberg.me> Signed-off-by:
-
Sagi Grimberg authored
commit f34e2589 upstream. If I/O queue connect times out, we might have freed the queue socket already, so check for that on the error path in nvme_tcp_start_queue. Signed-off-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
-
Sagi Grimberg authored
commit efb973b1 upstream. usually nvme_ prefix is for core functions. While we're cleaning up, remove redundant empty lines Signed-off-by:
Minwoo Im <minwoo.im@samsung.com> Signed-off-by:
-
Yang Shi authored
commit 7a30df49 upstream. A few new fields were added to mmu_gather to make TLB flush smarter for huge page by telling what level of page table is changed. __tlb_reset_range() is used to reset all these page table state to unchanged, which is called by TLB flush for parallel mapping changes for the same range under non-exclusive lock (i.e. read mmap_sem). Before commit dd2283f2 ("mm: mmap: zap pages with read mmap_sem in munmap"), the syscalls (e.g. MADV_DONTNEED, MADV_FREE) which may update PTEs in parallel don't remove page tables. But, the forementioned commit may do munmap() under read mmap_sem and free page tables. This may result in program hang on aarch64 reported by Jan Stancek. The problem could be reproduced by his test program with slightly modified below. ---8<--- static int map_size = 4096; static int num_iter = 500; static long threads_total; static void *distant_area; void *map_write_unmap(void *ptr) { int *fd = ptr; unsigned char *map_address; int i, j = 0; for (i = 0; i < num_iter; i++) { map_address = mmap(distant_area, (size_t) map_size, PROT_WRITE | PROT_READ, MAP_SHARED | MAP_ANONYMOUS, -1, 0); if (map_address == MAP_FAILED) { perror("mmap"); exit(1); } for (j = 0; j < map_size; j++) map_address[j] = 'b'; if (munmap(map_address, map_size) == -1) { perror("munmap"); exit(1); } } return NULL; } void *dummy(void *ptr) { return NULL; } int main(void) { pthread_t thid[2]; /* hint for mmap in map_write_unmap() */ distant_area = mmap(0, DISTANT_MMAP_SIZE, PROT_WRITE | PROT_READ, MAP_ANONYMOUS | MAP_PRIVATE, -1, 0); munmap(distant_area, (size_t)DISTANT_MMAP_SIZE); distant_area += DISTANT_MMAP_SIZE / 2; while (1) { pthread_create(&thid[0], NULL, map_write_unmap, NULL); pthread_create(&thid[1], NULL, dummy, NULL); pthread_join(thid[0], NULL); pthread_join(thid[1], NULL); } } ---8<--- The program may bring in parallel execution like below: t1 t2 munmap(map_address) downgrade_write(&mm->mmap_sem); unmap_region() tlb_gather_mmu() inc_tlb_flush_pending(tlb->mm); free_pgtables() tlb->freed_tables = 1 tlb->cleared_pmds = 1 pthread_exit() madvise(thread_stack, 8M, MADV_DONTNEED) zap_page_range() tlb_gather_mmu() inc_tlb_flush_pending(tlb->mm); tlb_finish_mmu() if (mm_tlb_flush_nested(tlb->mm)) __tlb_reset_range() __tlb_reset_range() would reset freed_tables and cleared_* bits, but this may cause inconsistency for munmap() which do free page tables. Then it may result in some architectures, e.g. aarch64, may not flush TLB completely as expected to have stale TLB entries remained. Use fullmm flush since it yields much better performance on aarch64 and non-fullmm doesn't yields significant difference on x86. The original proposed fix came from Jan Stancek who mainly debugged this issue, I just wrapped up everything together. Jan's testing results: v5.2-rc2-24-gbec7550c -------------------------- mean stddev real 37.382 2.780 user 1.420 0.078 sys 54.658 1.855 v5.2-rc2-24-gbec7550c + "mm: mmu_gather: remove __tlb_reset_range() for force flush" ---------------------------------------------------------------------------------------_ mean stddev real 37.119 2.105 user 1.548 0.087 sys 55.698 1.357 [akpm@linux-foundation.org: coding-style fixes] Link: http://lkml.kernel.org/r/1558322252-113575-1-git-send-email-yang.shi@linux.alibaba.com Fixes: dd2283f2 ("mm: mmap: zap pages with read mmap_sem in munmap") Signed-off-by:
Yang Shi <yang.shi@linux.alibaba.com> Signed-off-by:
Jan Stancek <jstancek@redhat.com> Reported-by:
Will Deacon <will.deacon@arm.com> Tested-by:
-
Tobin C. Harding authored
[ Upstream commit b9fba67b ] If a call to kobject_init_and_add() fails we should call kobject_put() otherwise we leak memory. Add call to kobject_put() in the error path of call to kobject_init_and_add(). Please note, this has the side effect that the release method is called if kobject_init_and_add() fails. Link: http://lkml.kernel.org/r/20190513033458.2824-1-tobin@kernel.orgSigned-off-by:
Tobin C. Harding <tobin@kernel.org> Reviewed-by:
Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Gang He <ghe@suse.com> Cc: Jun Piao <piaojun@huawei.com> Signed-off-by:
Sasha Levin <sashal@kernel.org>
-
Amit Cohen authored
[ Upstream commit 275e928f ] Force of 56G is not supported by hardware in Ethernet devices. This configuration fails with a bad parameter error from firmware. Add check of this case. Instead of trying to set 56G with autoneg off, return a meaningful error. Fixes: 56ade8fe ("mlxsw: spectrum: Add initial support for Spectrum ASIC") Signed-off-by:
Amit Cohen <amitc@mellanox.com> Acked-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
-
Jason Yan authored
[ Upstream commit 3b054179 ] The sas_port(phy->port) allocated in sas_ex_discover_expander() will not be deleted when the expander failed to discover. This will cause resource leak and a further issue of kernel BUG like below: [159785.843156] port-2:17:29: trying to add phy phy-2:17:29 fails: it's already part of another port [159785.852144] ------------[ cut here ]------------ [159785.856833] kernel BUG at drivers/scsi/scsi_transport_sas.c:1086! [159785.863000] Internal error: Oops - BUG: 0 [#1] SMP [159785.867866] CPU: 39 PID: 16993 Comm: kworker/u96:2 Tainted: G W OE 4.19.25-vhulk1901.1.0.h111.aarch64 #1 [159785.878458] Hardware name: Huawei Technologies Co., Ltd. Hi1620EVBCS/Hi1620EVBCS, BIOS Hi1620 CS B070 1P TA 03/21/2019 [159785.889231] Workqueue: 0000:74:02.0_disco_q sas_discover_domain [159785.895224] pstate: 40c00009 (nZcv daif +PAN +UAO) [159785.900094] pc : sas_port_add_phy+0x188/0x1b8 [159785.904524] lr : sas_port_add_phy+0x188/0x1b8 [159785.908952] sp : ffff0001120e3b80 [159785.912341] x29: ffff0001120e3b80 x28: 0000000000000000 [159785.917727] x27: ffff802ade8f5400 x26: ffff0000681b7560 [159785.923111] x25: ffff802adf11a800 x24: ffff0000680e8000 [159785.928496] x23: ffff802ade8f5728 x22: ffff802ade8f5708 [159785.933880] x21: ffff802adea2db40 x20: ffff802ade8f5400 [159785.939264] x19: ffff802adea2d800 x18: 0000000000000010 [159785.944649] x17: 00000000821bf734 x16: ffff00006714faa0 [159785.950033] x15: ffff0000e8ab4ecf x14: 7261702079646165 [159785.955417] x13: 726c612073277469 x12: ffff00006887b830 [159785.960802] x11: ffff00006773eaa0 x10: 7968702079687020 [159785.966186] x9 : 0000000000002453 x8 : 726f702072656874 [159785.971570] x7 : 6f6e6120666f2074 x6 : ffff802bcfb21290 [159785.976955] x5 : ffff802bcfb21290 x4 : 0000000000000000 [159785.982339] x3 : ffff802bcfb298c8 x2 : 337752b234c2ab00 [159785.987723] x1 : 337752b234c2ab00 x0 : 0000000000000000 [159785.993108] Process kworker/u96:2 (pid: 16993, stack limit = 0x0000000072dae094) [159786.000576] Call trace: [159786.003097] sas_port_add_phy+0x188/0x1b8 [159786.007179] sas_ex_get_linkrate.isra.5+0x134/0x140 [159786.012130] sas_ex_discover_expander+0x128/0x408 [159786.016906] sas_ex_discover_dev+0x218/0x4c8 [159786.021249] sas_ex_discover_devices+0x9c/0x1a8 [159786.025852] sas_discover_root_expander+0x134/0x160 [159786.030802] sas_discover_domain+0x1b8/0x1e8 [159786.035148] process_one_work+0x1b4/0x3f8 [159786.039230] worker_thread+0x54/0x470 [159786.042967] kthread+0x134/0x138 [159786.046269] ret_from_fork+0x10/0x18 [159786.049918] Code: 91322300 f0004402 91178042 97fe4c9b (d4210000) [159786.056083] Modules linked in: hns3_enet_ut(OE) hclge(OE) hnae3(OE) hisi_sas_test_hw(OE) hisi_sas_test_main(OE) serdes(OE) [159786.067202] ---[ end trace 03622b9e2d99e196 ]--- [159786.071893] Kernel panic - not syncing: Fatal exception [159786.077190] SMP: stopping secondary CPUs [159786.081192] Kernel Offset: disabled [159786.084753] CPU features: 0x2,a2a00a38 Fixes: 2908d778 ("[SCSI] aic94xx: new driver") Reported-by:
Jian Luo <luojian5@huawei.com> Signed-off-by:
Jason Yan <yanaijie@huawei.com> CC: John Garry <john.garry@huawei.com> Signed-off-by:
-
YueHaibing authored
[ Upstream commit 12e750bc ] If alloc_workqueue fails in alua_init, it should return -ENOMEM, otherwise it will trigger null-ptr-deref while unloading module which calls destroy_workqueue dereference wq->lock like this: BUG: KASAN: null-ptr-deref in __lock_acquire+0x6b4/0x1ee0 Read of size 8 at addr 0000000000000080 by task syz-executor.0/7045 CPU: 0 PID: 7045 Comm: syz-executor.0 Tainted: G C 5.1.0+ #28 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1ubuntu1 Call Trace: dump_stack+0xa9/0x10e __kasan_report+0x171/0x18d ? __lock_acquire+0x6b4/0x1ee0 kasan_report+0xe/0x20 __lock_acquire+0x6b4/0x1ee0 lock_acquire+0xb4/0x1b0 __mutex_lock+0xd8/0xb90 drain_workqueue+0x25/0x290 destroy_workqueue+0x1f/0x3f0 __x64_sys_delete_module+0x244/0x330 do_syscall_64+0x72/0x2a0 entry_SYSCALL_64_after_hwframe+0x49/0xbe Reported-by:
Hulk Robot <hulkci@huawei.com> Fixes: 03197b61 ("scsi_dh_alua: Use workqueue for RTPG") Signed-off-by:
YueHaibing <yuehaibing@huawei.com> Reviewed-by:
Bart Van Assche <bvanassche@acm.org> Signed-off-by:
-
Lianbo Jiang authored
[ Upstream commit 1d94f06e ] When SME is enabled, the smartpqi driver won't work on the HP DL385 G10 machine, which causes the failure of kernel boot because it fails to allocate pqi error buffer. Please refer to the kernel log: .... [ 9.431749] usbcore: registered new interface driver uas [ 9.441524] Microsemi PQI Driver (v1.1.4-130) [ 9.442956] i40e 0000:04:00.0: fw 6.70.48768 api 1.7 nvm 10.2.5 [ 9.447237] smartpqi 0000:23:00.0: Microsemi Smart Family Controller found Starting dracut initqueue hook... [ OK ] Started Show Plymouth Boot Scre[ 9.471654] Broadcom NetXtreme-C/E driver bnxt_en v1.9.1 en. [ OK ] Started Forward Password Requests to Plymouth Directory Watch. [[0;[ 9.487108] smartpqi 0000:23:00.0: failed to allocate PQI error buffer .... [ 139.050544] dracut-initqueue[949]: Warning: dracut-initqueue timeout - starting timeout scripts [ 139.589779] dracut-initqueue[949]: Warning: dracut-initqueue timeout - starting timeout scripts Basically, the fact that the coherent DMA mask value wasn't set caused the driver to fall back to SWIOTLB when SME is active. For correct operation, lets call the dma_set_mask_and_coherent() to properly set the mask for both streaming and coherent, in order to inform the kernel about the devices DMA addressing capabilities. Signed-off-by:
Lianbo Jiang <lijiang@redhat.com> Acked-by:
Don Brace <don.brace@microsemi.com> Tested-by:
-
Varun Prakash authored
[ Upstream commit cc555759 ] ip_dev_find() can return NULL so add a check for NULL pointer. Signed-off-by:
Varun Prakash <varun@chelsio.com> Signed-off-by:
-
Max Uvarov authored
[ Upstream commit 2b892649 ] PHY_INTERFACE_MODE_RGMII_RXID is less then TXID so code to set tx delay is never called. Fixes: 2a10154a ("net: phy: dp83867: Add TI dp83867 phy") Signed-off-by:
Max Uvarov <muvarov@gmail.com> Cc: Florian Fainelli <f.fainelli@gmail.com> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Max Uvarov authored
[ Upstream commit 1a97a477 ] After reset SGMII Autoneg timer is set to 2us (bits 6 and 5 are 01). That is not enough to finalize autonegatiation on some devices. Increase this timer duration to maximum supported 16ms. Signed-off-by:
-
Max Uvarov authored
[ Upstream commit 333061b9 ] For supporting 10Mps speed in SGMII mode DP83867_10M_SGMII_RATE_ADAPT bit of DP83867_10M_SGMII_CFG register has to be cleared by software. That does not affect speeds 100 and 1000 so can be done on init. Signed-off-by:
-
Russell King authored
[ Upstream commit c6787263 ] Ensure that we supply the same phy interface mode to mac_link_down() as we did for the corresponding mac_link_up() call. This ensures that MAC drivers that use the phy interface mode in these methods can depend on mac_link_down() always corresponding to a mac_link_up() call for the same interface mode. Signed-off-by:
Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by:
-
Jes Sorensen authored
[ Upstream commit 41de54c6 ] If blk_mq_init_allocated_queue() fails, make sure to free the poll stat callback struct allocated. Signed-off-by:
Jes Sorensen <jsorensen@fb.com> Signed-off-by:
Jens Axboe <axboe@kernel.dk> Signed-off-by:
-
Yoshihiro Shimoda authored
[ Upstream commit 315ca92d ] The sh_eth_close() resets the MAC and then calls phy_stop() so that mdio read access result is incorrect without any error according to kernel trace like below: ifconfig-216 [003] .n.. 109.133124: mdio_access: ee700000.ethernet-ffffffff read phy:0x01 reg:0x00 val:0xffff According to the hardware manual, the RMII mode should be set to 1 before operation the Ethernet MAC. However, the previous code was not set to 1 after the driver issued the soft_reset in sh_eth_dev_exit() so that the mdio read access result seemed incorrect. To fix the issue, this patch adds a condition and set the RMII mode register in sh_eth_dev_exit() for R-Car Gen2 and RZ/A1 SoCs. Note that when I have tried to move the sh_eth_dev_exit() calling after phy_stop() on sh_eth_close(), but it gets worse (kernel panic happened and it seems that a register is accessed while the clock is off). Signed-off-by:
Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com> Signed-off-by:
-
Sami Tolvanen authored
[ Upstream commit 1e29ab31 ] Calling sys_ni_syscall through a syscall_fn_t pointer trips indirect call Control-Flow Integrity checking due to a function type mismatch. Use SYSCALL_DEFINE0 for __arm64_sys_ni_syscall instead and remove the now unnecessary casts. Signed-off-by:
Sami Tolvanen <samitolvanen@google.com> Signed-off-by:
-
Sami Tolvanen authored
[ Upstream commit 0e358bd7 ] Although a syscall defined using SYSCALL_DEFINE0 doesn't accept parameters, use the correct function type to avoid indirect call type mismatches with Control-Flow Integrity checking. Signed-off-by:
-
