On Thu, 2017-08-10 at 11:14 +0900, Damien Le Moal wrote:
> I am currently trying different approaches for this. In the mean time, I
> would like to see the unlock change patch be applied to fix the deadlock
> problem.
Hello Damien,
That approach sounds fine to me.
Bart.
back again the (1) or (2) cases, without
>>> any change in the order. Isnt't it ?
>>
>> Hello Damien,
>>
>> Commands that get requeued are not reinserted at their original position
>> in the request queue but usually at a different position. This is what makes
&g
llo Damien,
>
> Commands that get requeued are not reinserted at their original position
> in the request queue but usually at a different position. This is what makes
> requeueing cause request reordering. Anyway, can you check whether replacing
> patch "sd_zbc: Write unlock zo
but usually at a different position. This is what makes
requeueing cause request reordering. Anyway, can you check whether replacing
patch "sd_zbc: Write unlock zone from sd_uninit_cmnd()" with the (untested)
patch below does not trigger any lockups and prevents request reordering?
Si
On 8/9/17 13:15, Bart Van Assche wrote:
> On Wed, 2017-08-09 at 13:07 +0900, Damien Le Moal wrote:
>>
>> On 8/9/17 12:57, Bart Van Assche wrote:
>>> On Wed, 2017-08-09 at 11:47 +0900, Damien Le Moal wrote:
On 8/9/17 11:19, Damien Le Moal wrote:
> On 8/9/17 00:07, Bart Van Assche wrote:
On Wed, 2017-08-09 at 13:07 +0900, Damien Le Moal wrote:
>
> On 8/9/17 12:57, Bart Van Assche wrote:
> > On Wed, 2017-08-09 at 11:47 +0900, Damien Le Moal wrote:
> > > On 8/9/17 11:19, Damien Le Moal wrote:
> > > > On 8/9/17 00:07, Bart Van Assche wrote:
> > > > > On Tue, 2017-08-08 at 13:17
On 8/9/17 12:57, Bart Van Assche wrote:
> On Wed, 2017-08-09 at 11:47 +0900, Damien Le Moal wrote:
>> On 8/9/17 11:19, Damien Le Moal wrote:
>>> On 8/9/17 00:07, Bart Van Assche wrote:
On Tue, 2017-08-08 at 13:17 +0900, Damien Le Moal wrote:
> acquired the lock completes can cause
On Wed, 2017-08-09 at 11:47 +0900, Damien Le Moal wrote:
> On 8/9/17 11:19, Damien Le Moal wrote:
> > On 8/9/17 00:07, Bart Van Assche wrote:
> > > On Tue, 2017-08-08 at 13:17 +0900, Damien Le Moal wrote:
> > > > acquired the lock completes can cause deadlocks with scsi-mq due to
> > > > potential
Bart,
On 8/9/17 11:19, Damien Le Moal wrote:
> Bart,
>
> On 8/9/17 00:07, Bart Van Assche wrote:
>> On Tue, 2017-08-08 at 13:17 +0900, Damien Le Moal wrote:
>>> Releasing the write lock of a zone when the write commnand that
>>
>>
Bart,
On 8/9/17 00:07, Bart Van Assche wrote:
> On Tue, 2017-08-08 at 13:17 +0900, Damien Le Moal wrote:
>> Releasing the write lock of a zone when the write commnand that
>
> command?
On Tue, 2017-08-08 at 13:17 +0900, Damien Le Moal wrote:
> Releasing the write lock of a zone when the write commnand that
command?
> acquired the lock completes can cause deadlocks
Releasing the write lock of a zone when the write commnand that
acquired the lock completes can cause deadlocks with scsi-mq due to
potential queue reordering if the lock owning request is requeued and
not executed.
Since sd_uninit_cmnd() is always called when a request is requeued,
call
12 matches
Mail list logo