Hi,

Same crash here on 5 bookworm KVM guests after the unattended reboot into
6.1.192-1: Hetzner Cloud arm64 and amd64 (generic kernel), OVH amd64
(cloud kernel), all with ext4 on virtio-scsi. The crash came anywhere from
25 s to about 5 h after boot, so hosts that look fine may just not have hit
it yet. All of them are stable on 6.1.0-54 (6.1.190-1).

For the record, it also happens on arm64, here from a plain unlink
(reproduced on two boots, 25 s and 39 s after boot):

[   25.014698] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
[   25.030145] Internal error: Oops: 0000000096000007 [#1] SMP
[   25.056014] CPU: 0 PID: 987 Comm: syncthing Not tainted 6.1.0-55-arm64 #1  Debian 6.1.192-1 [   25.057794] Hardware name: Hetzner vServer/KVM Virtual Machine, BIOS 20171111 11/11/2017
[   25.060797] pc : ext4_mb_load_buddy_gfp+0xac/0x454 [ext4]
[   25.061941] lr : ext4_mb_load_buddy_gfp+0xa0/0x454 [ext4]
[   25.079336] Call trace:
[   25.079819]  ext4_mb_load_buddy_gfp+0xac/0x454 [ext4]
[   25.081167]  ext4_free_blocks+0x374/0x9f0 [ext4]
[   25.082021]  ext4_ext_remove_space+0x9b8/0x132c [ext4]
[   25.083095]  ext4_ext_truncate+0xd0/0x100 [ext4]
[   25.084072]  ext4_truncate+0x320/0x49c [ext4]
[   25.084930]  ext4_evict_inode+0x2f0/0x594 [ext4]
[   25.085858]  evict+0x12c/0x380
[   25.086910]  iput+0x150/0x2f4
[   25.088096]  do_unlinkat+0x1d4/0x2d0
[   25.089300]  __arm64_sys_unlinkat+0x44/0x90
[   25.090900]  invoke_syscall+0x68/0xe0
[   25.092164]  el0_svc_common.constprop.0+0xd4/0xf4
[   25.093513]  do_el0_svc+0x2c/0x9c
[   25.094613]  el0_svc+0x34/0xdc
[   25.095629]  el0t_64_sync_handler+0xf4/0x120
[   25.096826]  el0t_64_sync+0x18c/0x190
[   25.098058] Code: 940125b8 aa0003f3 b140041f 54000b48 (f9400000)
[   25.099565] ---[ end trace 0000000000000000 ]---

If the fix in MR 2168 gets an arm64 test build, we can test it on this host,
which reproduces the crash within a minute.

Regards,
Nils

Reply via email to