On 8/11/21 7:35 AM, Bruno Goncalves wrote: > Hello, > > We've hit the issue below twice during our tests on kernel 5.14.0-rc5, > the issues happened only on aarch64 when using xfstests [1]. More logs > are available at [2]. > > The commits we hit this are: > Commit: 39a7b1209b44 - Merge branch 'io_uring-bio-cache.3' into for-next > Commit: 4a23f3da70f0 - Merge branch 'io_uring-bio-cache.3' into for-next > > [ 1984.069676] XFS (sda4): Mounting V5 Filesystem > [ 1984.099258] XFS (sda4): Ending clean mount > [ 1984.104770] xfs filesystem being mounted at /mnt/xfstests/mnt1 > supports timestamps until 2038 (0x7fffffff) > [ 1984.185598] run fstests generic/006 at 2021-08-10 18:02:37 > [ 1985.746695] XFS (sda4): Unmounting Filesystem > [ 1986.007696] Unable to handle kernel NULL pointer dereference at > virtual address 0000000000000020 > [ 1986.016515] Mem abort info: > [ 1986.019298] ESR = 0x96000004 > [ 1986.022368] EC = 0x25: DABT (current EL), IL = 32 bits > [ 1986.027671] SET = 0, FnV = 0 > [ 1986.030732] EA = 0, S1PTW = 0 > [ 1986.033863] FSC = 0x04: level 0 translation fault > [ 1986.038730] Data abort info: > [ 1986.041609] ISV = 0, ISS = 0x00000004 > [ 1986.045434] CM = 0, WnR = 0 > [ 1986.048390] user pgtable: 4k pages, 48-bit VAs, pgdp=000000881c5e2000 > [ 1986.054831] [0000000000000020] pgd=0000000000000000, p4d=0000000000000000 > [ 1986.061625] Internal error: Oops: 96000004 [#1] SMP > [ 1986.066496] Modules linked in: dm_log_writes dm_flakey rfkill > mlx5_ib ib_uverbs ib_core sunrpc coresight_etm4x i2c_smbus > coresight_replicator coresight_tpiu coresight_tmc joydev mlx5_core > mlxfw psample tls acpi_ipmi ipmi_ssif ipmi_devintf ipmi_msghandler > thunderx2_pmu coresight_funnel coresight cppc_cpufreq vfat fat fuse > zram ip_tables xfs ast crct10dif_ce i2c_algo_bit drm_vram_helper > ghash_ce drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops > cec drm_ttm_helper ttm drm gpio_xlp i2c_xlp9xx uas usb_storage > aes_neon_bs > [ 1986.113226] CPU: 115 PID: 246060 Comm: xfs_repair Not tainted 5.14.0-rc5 #1 > [ 1986.120179] Hardware name: HPE Apollo 70 /C01_APACHE_MB > , BIOS L50_5.13_1.16 07/29/2020 > [ 1986.129908] pstate: 80400009 (Nzcv daif +PAN -UAO -TCO BTYPE=--) > [ 1986.135905] pc : __blkdev_direct_IO+0x45c/0x4dc > [ 1986.140438] lr : __blkdev_direct_IO+0x44c/0x4dc > [ 1986.144959] sp : ffff80003aa03ba0 > [ 1986.148262] x29: ffff80003aa03ba0 x28: ffff8000123936c8 x27: ffff00879948cb18 > [ 1986.155391] x26: 0000000000000002 x25: ffff00879948cb18 x24: ffff000f7d92c180 > [ 1986.162519] x23: ffff000811325400 x22: 00000000ffffffff x21: 0000000000200000 > [ 1986.169646] x20: ffff80003aa03d38 x19: ffff00879948cb00 x18: 0000000000000000 > [ 1986.176774] x17: 0000000000000000 x16: 0000000000010000 x15: fffffc021e4f0002 > [ 1986.183901] x14: 0000000000000000 x13: ffff8000115b8090 x12: 00000000ffffffff > [ 1986.191029] x11: ffff80001145afe0 x10: 0000000000001c60 x9 : ffff800010401b70 > [ 1986.198156] x8 : ffff000f7d92de40 x7 : 0000000000000000 x6 : 00000008e1dbe6bc > [ 1986.205283] x5 : 0000000000001cd0 x4 : 0000000000000000 x3 : ffffffffffffefff > [ 1986.212411] x2 : 0000000000000002 x1 : ffff800011064b70 x0 : 0000000000000000 > [ 1986.219539] Call trace: > [ 1986.221975] __blkdev_direct_IO+0x45c/0x4dc > [ 1986.226149] blkdev_direct_IO+0x70/0xb0 > [ 1986.229977] generic_file_read_iter+0x8c/0x194 > [ 1986.234420] blkdev_read_iter+0x54/0x90 > [ 1986.238246] new_sync_read+0xdc/0x154 > [ 1986.241906] vfs_read+0x158/0x1e4 > [ 1986.245212] ksys_pread64+0x84/0xcc > [ 1986.248691] __arm64_sys_pread64+0x2c/0x40 > [ 1986.252779] invoke_syscall+0x50/0x120 > [ 1986.256520] el0_svc_common+0x48/0x100 > [ 1986.260259] do_el0_svc+0x34/0xa0 > [ 1986.263563] el0_svc+0x2c/0x54 > [ 1986.266616] el0t_64_sync_handler+0x1a4/0x1b0 > [ 1986.270964] el0t_64_sync+0x19c/0x1a0 > [ 1986.274621] Code: 2a0003f5 35fffda0 f85e8320 b9400a75 (b9402001) > [ 1986.280706] ---[ end trace c7734f7d18437758 ]--- Thanks for the report! That branch is not in for-next, it's a work in progress. It was briefly in my next branch by mistake, but only for a very short amount of time. I'll be pushing out a new version later today. -- Jens Axboe
