* Why removing REQ_FAILFAST_DRIVER in LightNVM?
@ 2017-10-06 9:19 Javier González
2017-10-06 11:59 ` Christoph Hellwig
0 siblings, 1 reply; 6+ messages in thread
From: Javier González @ 2017-10-06 9:19 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Matias Bjørling, linux-block
[-- Attachment #1: Type: text/plain, Size: 490 bytes --]
Hi Christoph,
I'm cleaning up lightnvm.c to use as much as possible the nvme helpers.
I see that in Commit: d49187e97e94 "nvme: introduce struct nvme_request"
you introduced:
rq->cmd_flags &= ~REQ_FAILFAST_DRIVER
on the lightnvm I/O path and that has propagated through the code as we
added more functionality. Can you explain why this is necessary? If I
can just remove it, it is much easier to do the cleanup.
I have tested on or HW and all works fine without it.
Thanks,
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Why removing REQ_FAILFAST_DRIVER in LightNVM?
2017-10-06 9:19 Why removing REQ_FAILFAST_DRIVER in LightNVM? Javier González
@ 2017-10-06 11:59 ` Christoph Hellwig
2017-10-06 12:01 ` Javier González
0 siblings, 1 reply; 6+ messages in thread
From: Christoph Hellwig @ 2017-10-06 11:59 UTC (permalink / raw)
To: Javier González; +Cc: Christoph Hellwig, Matias Bjørling, linux-block
On Fri, Oct 06, 2017 at 11:19:09AM +0200, Javier Gonz�lez wrote:
> on the lightnvm I/O path and that has propagated through the code as we
> added more functionality. Can you explain why this is necessary? If I
> can just remove it, it is much easier to do the cleanup.
>
> I have tested on or HW and all works fine without it.
nvme_alloc_request always sets REQ_FAILFAST_DRIVER and that srops
it to stick to the previous behavior.
If you actually want all your lightnvm requests to fail fast you can
remove it.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Why removing REQ_FAILFAST_DRIVER in LightNVM?
2017-10-06 11:59 ` Christoph Hellwig
@ 2017-10-06 12:01 ` Javier González
2017-10-06 12:06 ` Christoph Hellwig
0 siblings, 1 reply; 6+ messages in thread
From: Javier González @ 2017-10-06 12:01 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Matias Bjørling, linux-block
[-- Attachment #1: Type: text/plain, Size: 758 bytes --]
> On 6 Oct 2017, at 13.59, Christoph Hellwig <hch@infradead.org> wrote:
>
> On Fri, Oct 06, 2017 at 11:19:09AM +0200, Javier González wrote:
>> on the lightnvm I/O path and that has propagated through the code as we
>> added more functionality. Can you explain why this is necessary? If I
>> can just remove it, it is much easier to do the cleanup.
>>
>> I have tested on or HW and all works fine without it.
>
> nvme_alloc_request always sets REQ_FAILFAST_DRIVER and that srops
> it to stick to the previous behavior.
>
> If you actually want all your lightnvm requests to fail fast you can
> remove it.
I think it is good to fail fast as any other nvme I/O command and then
recover in pblk if necessary.
Thanks Christoph!
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Why removing REQ_FAILFAST_DRIVER in LightNVM?
2017-10-06 12:01 ` Javier González
@ 2017-10-06 12:06 ` Christoph Hellwig
2017-10-06 12:08 ` Javier González
0 siblings, 1 reply; 6+ messages in thread
From: Christoph Hellwig @ 2017-10-06 12:06 UTC (permalink / raw)
To: Javier González; +Cc: Christoph Hellwig, Matias Bjørling, linux-block
On Fri, Oct 06, 2017 at 02:01:46PM +0200, Javier Gonz�lez wrote:
> I think it is good to fail fast as any other nvme I/O command and then
> recover in pblk if necessary.
Note that we only do it for other nvme _passthrough_ commands - the
actual I/O commands dot not get the failfast flag.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Why removing REQ_FAILFAST_DRIVER in LightNVM?
2017-10-06 12:06 ` Christoph Hellwig
@ 2017-10-06 12:08 ` Javier González
2017-10-06 12:12 ` Javier González
0 siblings, 1 reply; 6+ messages in thread
From: Javier González @ 2017-10-06 12:08 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Matias Bjørling, linux-block
[-- Attachment #1: Type: text/plain, Size: 591 bytes --]
> On 6 Oct 2017, at 14.06, Christoph Hellwig <hch@infradead.org> wrote:
>
> On Fri, Oct 06, 2017 at 02:01:46PM +0200, Javier González wrote:
>> I think it is good to fail fast as any other nvme I/O command and then
>> recover in pblk if necessary.
>
> Note that we only do it for other nvme _passthrough_ commands - the
> actual I/O commands dot not get the failfast flag.
Yes. I was thinking of the LightNVM-specific _passthrough_ used by
liblightnvm since I'm refactoring the sync. path at the moment; the
"normal" async. path will be untouched for now.
Thanks,
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Why removing REQ_FAILFAST_DRIVER in LightNVM?
2017-10-06 12:08 ` Javier González
@ 2017-10-06 12:12 ` Javier González
0 siblings, 0 replies; 6+ messages in thread
From: Javier González @ 2017-10-06 12:12 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Matias Bjørling, linux-block
[-- Attachment #1: Type: text/plain, Size: 859 bytes --]
> On 6 Oct 2017, at 14.08, Javier González <jg@lightnvm.io> wrote:
>
>> On 6 Oct 2017, at 14.06, Christoph Hellwig <hch@infradead.org> wrote:
>>
>> On Fri, Oct 06, 2017 at 02:01:46PM +0200, Javier González wrote:
>>> I think it is good to fail fast as any other nvme I/O command and then
>>> recover in pblk if necessary.
>>
>> Note that we only do it for other nvme _passthrough_ commands - the
>> actual I/O commands dot not get the failfast flag.
>
> Yes. I was thinking of the LightNVM-specific _passthrough_ used by
> liblightnvm since I'm refactoring the sync. path at the moment; the
> "normal" async. path will be untouched for now.
In fact, for the normal I/O path the whole idea is to do this directly
in pblk, when we use struct request, and drop all these conversions in
lightnvm.c. But that will be later on...
Javier
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-10-06 12:12 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-10-06 9:19 Why removing REQ_FAILFAST_DRIVER in LightNVM? Javier González
2017-10-06 11:59 ` Christoph Hellwig
2017-10-06 12:01 ` Javier González
2017-10-06 12:06 ` Christoph Hellwig
2017-10-06 12:08 ` Javier González
2017-10-06 12:12 ` Javier González
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox