[bug report][regression] blktests loop/004 failed

[bug report][regression] blktests loop/004 failed

[Yi Zhang]

Hi

I found the blktests loop/004[1] failed on the latest
linux-block/for-next, and it was introduced from the commit[2] from my
testing,
please help check it and let me know if you need any info/testing, thanks.

[1]
# ./check loop/004
loop/004 (combine loop direct I/O mode and a custom block size) [failed]
    runtime  2.770s  ...  5.375s
    --- tests/loop/004.out 2025-08-24 01:41:06.768628600 -0400
    +++ /root/blktests/results/nodev/loop/004.out.bad 2025-08-24
10:01:44.489825116 -0400
    @@ -1,4 +1,5 @@
     Running loop/004
     1
    -769bd186841c10e5b1106b55986206c0e87fc05a7f565fdee01b5abcaff6ae78  -
    +pwrite: No space left on device
    +e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855  -
     Test complete
[2]
47b71abd5846 loop: use vfs_getattr_nosec for accurate file size
[3] dmesg
[  279.880860] run blktests loop/004 at 2025-08-24 10:01:38
[  280.249598] scsi_debug:sdebug_driver_probe: scsi_debug: trim
poll_queues to 0. poll_q/nr_hw = (0/1)
[  280.259112] scsi host10: scsi_debug: version 0191 [20210520]
                 dev_size_mb=8, opts=0x0, submit_queues=1, statistics=0
[  280.289850] scsi 10:0:0:0: Direct-Access     Linux    scsi_debug
   0191 PQ: 0 ANSI: 7
[  280.302147] scsi 10:0:0:0: Power-on or device reset occurred
[  280.336792] sd 10:0:0:0: Attached scsi generic sg5 type 0
[  280.342300] sd 10:0:0:0: [sdf] 2048 4096-byte logical blocks: (8.39
MB/8.00 MiB)
[  280.343809] sd 10:0:0:0: [sdf] Write Protect is off
[  280.343925] sd 10:0:0:0: [sdf] Mode Sense: 73 00 10 08
[  280.347889] sd 10:0:0:0: [sdf] Write cache: enabled, read cache:
enabled, supports DPO and FUA
[  280.355174] sd 10:0:0:0: [sdf] permanent stream count = 5
[  280.359355] sd 10:0:0:0: [sdf] Preferred minimum I/O size 4096 bytes
[  280.359503] sd 10:0:0:0: [sdf] Optimal transfer size 4194304 bytes
[  280.506461] sd 10:0:0:0: [sdf] Attached SCSI disk
[  285.158954] sd 10:0:0:0: [sdf] Synchronizing SCSI cache

-- 
Best Regards,
  Yi Zhang

Re: [bug report][regression] blktests loop/004 failed

[Yu Kuai]

Hi,

在 2025/08/24 22:05, Yi Zhang 写道:
> Hi
> 
> I found the blktests loop/004[1] failed on the latest
> linux-block/for-next, and it was introduced from the commit[2] from my
> testing,
> please help check it and let me know if you need any info/testing, thanks.
> 
> [1]
> # ./check loop/004
> loop/004 (combine loop direct I/O mode and a custom block size) [failed]
>      runtime  2.770s  ...  5.375s
>      --- tests/loop/004.out 2025-08-24 01:41:06.768628600 -0400
>      +++ /root/blktests/results/nodev/loop/004.out.bad 2025-08-24
> 10:01:44.489825116 -0400
>      @@ -1,4 +1,5 @@
>       Running loop/004
>       1
>      -769bd186841c10e5b1106b55986206c0e87fc05a7f565fdee01b5abcaff6ae78  -
>      +pwrite: No space left on device
>      +e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855  -
>       Test complete

This looks like the same as:

https://lore.kernel.org/all/79ab5533-82d1-4f06-461b-689e94f738ec@huaweicloud.com/

However, I'm not sure if people will agree to change file size for block
inode returned by syscall from 0 to disk size. I'll send a patch later
today.

Thanks,
Kuai
> [2]
> 47b71abd5846 loop: use vfs_getattr_nosec for accurate file size
> [3] dmesg
> [  279.880860] run blktests loop/004 at 2025-08-24 10:01:38
> [  280.249598] scsi_debug:sdebug_driver_probe: scsi_debug: trim
> poll_queues to 0. poll_q/nr_hw = (0/1)
> [  280.259112] scsi host10: scsi_debug: version 0191 [20210520]
>                   dev_size_mb=8, opts=0x0, submit_queues=1, statistics=0
> [  280.289850] scsi 10:0:0:0: Direct-Access     Linux    scsi_debug
>     0191 PQ: 0 ANSI: 7
> [  280.302147] scsi 10:0:0:0: Power-on or device reset occurred
> [  280.336792] sd 10:0:0:0: Attached scsi generic sg5 type 0
> [  280.342300] sd 10:0:0:0: [sdf] 2048 4096-byte logical blocks: (8.39
> MB/8.00 MiB)
> [  280.343809] sd 10:0:0:0: [sdf] Write Protect is off
> [  280.343925] sd 10:0:0:0: [sdf] Mode Sense: 73 00 10 08
> [  280.347889] sd 10:0:0:0: [sdf] Write cache: enabled, read cache:
> enabled, supports DPO and FUA
> [  280.355174] sd 10:0:0:0: [sdf] permanent stream count = 5
> [  280.359355] sd 10:0:0:0: [sdf] Preferred minimum I/O size 4096 bytes
> [  280.359503] sd 10:0:0:0: [sdf] Optimal transfer size 4194304 bytes
> [  280.506461] sd 10:0:0:0: [sdf] Attached SCSI disk
> [  285.158954] sd 10:0:0:0: [sdf] Synchronizing SCSI cache
>