Re: [PATCH 0/4] fix experimental verify

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 10/18/22 20:34, Shinichiro Kawasaki wrote:
On Oct 18, 2022 / 12:14, Vincent Fu wrote:

[...]

Hello Shin'ichiro,

I applied your patches to a branch on GitHub:

https://github.com/vincentkfu/fio/commits/exp-verify

All of the automated Linux and macOS tests pass but on Windows I see
failures with t0026:

https://ci.appveyor.com/project/vincentkfu/fio/builds/45105941

Would you take a look?

Thanks. All 3 failures reported "bad header rand_seed":

verify: bad header rand_seed 59864971564696990, wanted 1584616954246794908 at file t0026file offset 1019904, length 4096 (requested block: offset=1019904, length=4096)
verify: bad header rand_seed 59864971564696990, wanted 1584616954246794908 at file t0026file offset 1019904, length 4096 (requested block: offset=1019904, length=4096)
verify: bad header rand_seed 822895868, wanted 2592444842 at file t0026file offset 1019904, length 4096 (requested block: offset=1019904, length=4096)

The bug fixes made experimenal verify work again, and unveiled another failure
scenrio unique to Windows. I walked through rand_seed handling in verify path,
but could not find anything wrong, or anything os dependent. I guess integer
type sizes unique to Windows could be the cause, but not sure.

My Windows development skill is dusty now, and I don't have Windows environment.
It is difficult for me to chase it further.

Or, it would be the better to disable this test case for Windows. Is it ok to
add the requirement below to this test case?

         'requirements':     [Requirements.not_windows],


Disabling the test on Windows is a reasonable thing to do.

Vincent



[Index of Archives]     [Linux Kernel]     [Linux SCSI]     [Linux IDE]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux SCSI]

  Powered by Linux