On 25 July 2017 at 19:34, Sitsofe Wheeler <sitsofe@xxxxxxxxx> wrote:
> On 25 July 2017 at 16:29, Rebecca Cran <rebecca@xxxxxxxxxxxx> wrote:
>> On 7/25/2017 1:16 AM, Tomohiro Kusumi wrote:
>>>
>>> It seems the recent commit 2c3e17be ('filesetup: add native
>>> fallocate') breaks on Cygwin.
>>> It works on FreeBSD by the way.
>>
>>
>> I see the problem: FIO_FALLOCATE_NONE has a value of 1, so in fallocate_file
>> the 0 which is produced for platforms like Windows that don't support
>> fallocate fall through to the default case, which asserts. I'm not sure if
>> we want to renumber the fio_fallocate_mode enum to start at 0?
>
> Sigh, that was my mistake - I never ran fio on a Windows filesystem
> (only ioengine=null) to notice this and I guess every other platform I
> tried had at least posix_fallocate.
>
> I suppose other options are:
> * Just put back (modified) #ifdef guards back in around the whole case
> statement: https://github.com/axboe/fio/commit/2c3e17be4c7c9a737317ada414b98929652fec15#diff-4abbf037246dd2e450dc3f6a2ac77180L48
> * Add a case 0: /* fallocate N/A */ just above case FIO_FALLOCATE_NONE:
Hopefully this has been fixed by
https://github.com/axboe/fio/commit/d27f75b8ac13a4056cc8ef09b56de9dd184a2ad1
("init: force fallocate_mode to none when fallocate is unsupported").
--
Sitsofe | http://sucs.org/~sits/
--
To unsubscribe from this list: send the line "unsubscribe fio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
