On 4/27/17 10:38 AM, Darrick J. Wong wrote:
> On Thu, Apr 27, 2017 at 09:34:04AM -0500, Eric Sandeen wrote:
>> On 4/27/17 3:39 AM, Lukas Czerner wrote:
>>> fstrim can take really long time on big, slow device or on file system
>>> with a lots of allocation groups. Currently there is no way for the user
>>> to cancell the operation. This patch makes it possible for the user to
>>> kill fstrim pocess by adding the check for fatal_signal_pending() in
>>> xfs_trim_extents().
>>
>> With this patch, it seems that if it is killed, then nothing is copied
>> back to the user about the number of blocks trimmed before the kill.
>>
>> Is that intentional?
>>
>> I think it could be done in a way that the progress until the kill
>> does get reported, and that might be more useful?
>
> AFAICT ext4 doesn't report anything if somoene ERESTARTSYS's it...
Well we aren't ext4, are we? ;)
actually, it seems that it may report "count" as ERESTARTSYS in:
ext4_debug("trimmed %d blocks in the group %d\n",
count, group);
But yeah, although it does update the &range values, they don't
get copied out on any error.
Anyway, ok, ext4 doesn't. Is there a reason that we shouldn't?
Thanks,
-Eric
>>
>> Thanks,
>> -Eric
>>
>>> Signed-off-by: Lukas Czerner <lczerner@xxxxxxxxxx>
>>> Reported-by: Zdenek Kabelac <zkabelac@xxxxxxxxxx>
>>> ---
>>> fs/xfs/xfs_discard.c | 10 +++++++++-
>>> 1 file changed, 9 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
>>> index d796ffa..6a05d27 100644
>>> --- a/fs/xfs/xfs_discard.c
>>> +++ b/fs/xfs/xfs_discard.c
>>> @@ -132,6 +132,11 @@ xfs_trim_extents(
>>> error = xfs_btree_decrement(cur, 0, &i);
>>> if (error)
>>> goto out_del_cursor;
>>> +
>>> + if (fatal_signal_pending(current)) {
>
> ...but the process isn't going to be around to care anyway, right?
>
> --D
>
>>> + error = -ERESTARTSYS;
>>> + goto out_del_cursor;
>>> + }
>>> }
>>>
>>> out_del_cursor:
>>> @@ -196,8 +201,11 @@ xfs_ioc_trim(
>>> for (agno = start_agno; agno <= end_agno; agno++) {
>>> error = xfs_trim_extents(mp, agno, start, end, minlen,
>>> &blocks_trimmed);
>>> - if (error)
>>> + if (error) {
>>> last_error = error;
>>> + if (error == -ERESTARTSYS)
>>> + break;
>>> + }
>>> }
>>>
>>> if (last_error)
>>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
