Re: [V9fs-developer] KCSAN BUG report on p9_client_cb / p9_client_rpc

[jim . cromie]

 >
> > I had assumed the p9_req_put() in p9_client_cb would protect the tag,
> > but that doesn't appear to be true -- could you try this patch if this
> > is reproductible to you?
> >
>
> I applied your patch on top of my triggering case, it fixes the report  !
> you have my tested-by

I seem to have gotten ahead of my skis,
Im seeing another now, similar to 1st, differing in 2nd block

[    8.730061] Run /bin/sh as init process
[    9.027218] ==================================================================
[    9.028237] BUG: KCSAN: data-race in p9_client_cb / p9_client_rpc
[    9.029073]
[    9.029282] write to 0xffff888005e45ea0 of 4 bytes by interrupt on cpu 0:
[    9.030214]  p9_client_cb+0x1a/0x100
[    9.030735]  req_done+0xd3/0x130
[    9.031171]  vring_interrupt+0xac/0x130
[    9.031752]  __handle_irq_event_percpu+0x64/0x260
[    9.032381]  handle_irq_event+0x93/0x120
[    9.032950]  handle_edge_irq+0x123/0x400
[    9.033502]  __common_interrupt+0x3e/0xa0
[    9.034051]  common_interrupt+0x7e/0xa0
[    9.034608]  asm_common_interrupt+0x1e/0x40
[    9.035173]  native_safe_halt+0xe/0x10
[    9.035826]  default_idle+0xa/0x10
[    9.036299]  default_idle_call+0x38/0xc0
[    9.036845]  do_idle+0x1e7/0x270
[    9.037294]  cpu_startup_entry+0x19/0x20
[    9.037905]  rest_init+0xd0/0xd2
[    9.038354]  arch_call_rest_init+0xa/0x11
[    9.038922]  start_kernel+0xacb/0xadd
[    9.039444]  secondary_startup_64_no_verify+0xc2/0xcb
[    9.040140]
[    9.040369] read to 0xffff888005e45ea0 of 4 bytes by task 1 on cpu 1:
[    9.041283]  p9_client_rpc+0x185/0x860
[    9.041837]  p9_client_getattr_dotl+0x71/0x160
[    9.042645]  v9fs_inode_from_fid_dotl+0x21/0x160
[    9.043418]  v9fs_vfs_lookup.part.0+0x139/0x180
[    9.044059]  v9fs_vfs_lookup+0x32/0x40
[    9.044584]  __lookup_slow+0xc3/0x190
[    9.045095]  walk_component+0x1b8/0x270
[    9.045626]  link_path_walk.part.0.constprop.0+0x336/0x550
[    9.046425]  path_lookupat+0x59/0x340
[    9.046935]  filename_lookup+0x134/0x2a0
[    9.047484]  user_path_at_empty+0x6d/0x90
[    9.048145]  vfs_statx+0x79/0x1a0
[    9.048610]  __do_sys_newfstatat+0x1e/0x40
[    9.049173]  __x64_sys_newfstatat+0x4e/0x60
[    9.049755]  do_syscall_64+0x42/0x80
[    9.050233]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[    9.050940]
[    9.051148] Reported by Kernel Concurrency Sanitizer on:
[    9.051893] CPU: 1 PID: 1 Comm: virtme-init Not tainted
5.13.0-rc7-dd7i-00038-g4e27591489f1-dirty #126
[    9.053185] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS 1.14.0-3.fc34 04/01/2014
[    9.054358] ==================================================================




>
> > The tag is actually reclaimed in the woken up p9_client_rpc thread so
> > that would be a good match (reset in the other thread vs. read here),
> > caching the value is good enough but that is definitely not obvious...
> >
> > --
> > Dominique