On 5/4/18 10:14 AM, Tetsuo Handa wrote:
> Jens Axboe wrote:
>>> The loop module ignores sysfs_create_group() failure and pretends that
>>> LOOP_SET_FD request succeeded. I guess that the author of commit
>>> ee86273062cbb310 ("loop: add some basic read-only sysfs attributes")
>>> assumed that it is not a fatal error enough to abort LOOP_SET_FD request.
>>>
>>> Do we want to abort LOOP_SET_FD request if sysfs_create_group() failed?
>>
>> Probably safer to retain that behavior.
>
> OK.
>
>>
>>>> If that's not easily done, then my next suggestion would be to
>>>> use a loop flag for it, LO_FLAGS_SYSFS_SETUP or something like that.
>>>
>>> Yes, that would be possible.
>>
>> Let's make that change.
>
> Since LO_FLAGS_* are defined in include/uapi/linux/loop.h as userspace visible
> flags, I feel that using "struct loop_device"->lo_flags for recording whether
> sysfs entry exists might be strange... Anyway, updated patch is shown below.
Hmm yes, I forgot about that, I guess that makes the flags approach
pretty much useless. Let's just go with your v1 in that case.
--
Jens Axboe
