Re: WARNING in __init_work

[Zqiang <qiang.zhang1211@gmail.com>]

On 2021/10/18 下午12:47, Christoph Hellwig wrote:
>
> On Mon, Oct 18, 2021 at 12:43:38PM +0800, Zqiang wrote:
>> This is the details of the test, Hope it helps you
> Call me stupid, but I can only find the trace and linked to unreadable
> google sites that wan't me to log in somehow and no actual details.
>
> If you have a direct link to the reproducer (an attachment would do
> it as well) I'd love to try it myself.
>
> Otherwise this commit in the block-5.15 tree should help to catch what
> I suspect is the root cause (final ref drop before unregister) earlier
> and with a better backtrace:
>
> https://git.kernel.dk/cgit/linux-block/commit/?h=block-5.15&id=a20417611b98e12a724e5c828c472ea16990b71f

I found the following calltrace

Call Trace:
[  326.460593][T27634]  dump_stack_lvl+0x8d/0xcf
[  326.461773][T27634]  should_fail+0x13c/0x160
[  326.462921][T27634]  should_failslab+0x5/0x10
[  326.464038][T27634] slab_pre_alloc_hook.constprop.100+0x4e/0xc0
[  326.466040][T27634]  kmem_cache_alloc+0x44/0x2a0
[  326.466921][T27634]  __kernfs_new_node+0x68/0x350
[  326.469602][T27634]  kernfs_new_node+0x5a/0x90
[  326.470441][T27634]  __kernfs_create_file+0x56/0x150
[  326.471386][T27634]  sysfs_add_file_mode_ns+0xe6/0x290
[  326.472358][T27634]  internal_create_group+0x186/0x4e0
[  326.473331][T27634]  internal_create_groups.part.4+0x4d/0xb0
[  326.474288][T27634]  sysfs_create_groups+0x28/0x40
[  326.474918][T27634]  device_add+0x4c3/0xc60
[  326.476286][T27634]  add_partition+0x262/0x450
[  326.476919][T27634]  bdev_disk_changed+0x3ec/0x800
[  326.477615][T27634]  loop_reread_partitions+0x2d/0x70
[  326.478515][T27634]  loop_set_status+0x274/0x320
[  326.479373][T27634]  lo_ioctl+0x392/0x920
[  326.481271][T27634]  blkdev_ioctl+0x2ff/0x370
[  326.482438][T27634]  block_ioctl+0x55/0x70
[  326.483605][T27634]  __x64_sys_ioctl+0xb6/0x100
[  326.484241][T27634]  do_syscall_64+0x34/0xb0
[  326.484843][T27634]  entry_SYSCALL_64_after_hwframe+0x44/0xae

I find in add_partition(),  if the device_add() return error, we will 
drop disk object reference count,

but i find put_device(pdev) (will call part_release())and 
put_disk(disk), both will reduce the reference of the disk object , 
however we call get_device(disk_to_dev(disk)) only once

or Did I miss something and didn't analyze it？

Thanks

Zqiang