From: Ming Lei <ming.lei@redhat.com>
To: Daejun Park <daejun7.park@samsung.com>
Cc: Keith Busch <kbusch@kernel.org>,
"Martin K. Petersen" <martin.petersen@oracle.com>,
Jens Axboe <axboe@kernel.dk>,
Bart Van Assche <bvanassche@acm.org>,
Christoph Hellwig <hch@lst.de>,
James Bottomley <James.Bottomley@hansenpartnership.com>,
Jaegeuk Kim <jaegeuk@kernel.org>,
ALIM AKHTAR <alim.akhtar@samsung.com>,
"avri.altman@wdc.com" <avri.altman@wdc.com>,
"linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>,
"linux-block@vger.kernel.org" <linux-block@vger.kernel.org>
Subject: Re: [PATCH] scsi: ufs: mark HPB support as BROKEN
Date: Thu, 28 Oct 2021 10:07:35 +0800 [thread overview]
Message-ID: <YXoF59XeZ5KS0jZj@T590> (raw)
In-Reply-To: <20211028011022epcms2p1d2b2b1c56237c7cc1cca3a612f91b520@epcms2p1>
On Thu, Oct 28, 2021 at 10:10:22AM +0900, Daejun Park wrote:
> Hi Ming,
>
> > On Wed, Oct 27, 2021 at 09:16:32AM -0700, Keith Busch wrote:
> > > On Wed, Oct 27, 2021 at 11:58:23PM +0800, Ming Lei wrote:
> > > > On Wed, Oct 27, 2021 at 11:44:04AM -0400, Martin K. Petersen wrote:
> > > > >
> > > > > Ming,
> > > > >
> > > > > > request with scsi_cmnd may be allocated by the ufshpb driver, even it
> > > > > > should be fine to call ufshcd_queuecommand() directly for this driver
> > > > > > private IO, if the tag can be reused. One example is scsi_ioctl_reset().
> > > > >
> > > > > scsi_ioctl_reset() allocates a new request, though, so that doesn't
> > > > > solve the forward progress guarantee. Whereas eh puts the saved request
> > > > > on the stack.
> > > >
> > > > What I meant is to use one totally ufshpb private command allocated from
> > > > private slab to replace the spawned request, which is sent to ufshcd_queuecommand()
> > > > directly, so forward progress is guaranteed if the blk-mq request's tag can be
> > > > reused for issuing this private command. This approach takes a bit effort,
> > > > but avoids tags reservation.
> > > >
> > > > Yeah, it is cleaner to use reserved tag for the spawned request, but we
> > > > need to know:
> > > >
> > > > 1) how many queue depth for the hba? If it is small, even 1 reservation
> > > > can affect performance.
> > > >
> > > > 2) how many inflight write buffer commands are to be supported? Or how many
> > > > is enough for obtaining expected performance? If the number is big, reserved
> > > > tags can't work.
> > >
> > > The original and clone are not dispatched to hardware concurrently, so I
> > > don't think the reserved_tags need to subtract from the generic ones.
> > > The original request already accounts for the hardware resource, so the
> > > clone doesn't need to consume another one.
> >
> > Yeah, that is why I thought the tag could be reused for the spawned(cloned)
> > request, but it needs ufshpb developer to confirm, or at least
> > ufshcd_queuecommand() can handle this situation. If that is true, it isn't
> > necessary to use reserve tags, since the current blk-mq implementation
> > requires to reserve real hardware tags space, which has to take normal
> > tags.
>
> It is true that pre-request can use the tag of READ request, but the READ
> request should wait to completion of the pre-request command. However, if
> the pre-request and the READ request are dispatched concurrently, it can
> save the time to completion of the pre-request.
>
> So I implemented as allocating new request and it has limit time to getting
> pre-request, so it doesn't cause deadlock.
The WB pre-request and the READ IO should have been issued in single
command with same tag.
If they can't be done in single command, this sw/hw interface is really bad,
one way is to follow Jens's suggestion to reserve one tag for guaranteeing
forward progress, something like the following:
ufshpb_issue_pre_req()
req = blk_get_request(cmd->device->request_queue,
REQ_OP_DRV_OUT | REQ_SYNC, BLK_MQ_REQ_NOWAIT);
if (IS_ERR(req)) {
req = blk_get_request(cmd->device->request_queue,
REQ_OP_DRV_OUT | REQ_SYNC, BLK_MQ_REQ_NOWAIT | BLK_MQ_REQ_RESERVED);
if (IS_ERR(req))
return -EAGAIN;
}
But the above isn't what Christoph complained:
The HPB support added this merge window is fundanetally flawed as it
uses blk_insert_cloned_request to insert a cloned request onto the same
queue as the one that the original request came from, leading to all
kinds of issues in blk-mq accounting (in addition to this API being
a special case for dm-mpath that should not see other users).
IMO, accounting seems fine since block layer always counts driver
private request.
thanks,
Ming
next prev parent reply other threads:[~2021-10-28 2:07 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-26 7:12 [PATCH] scsi: ufs: mark HPB support as BROKEN Christoph Hellwig
2021-10-26 7:18 ` Hannes Reinecke
2021-10-26 7:24 ` Damien Le Moal
2021-10-26 13:04 ` James Bottomley
2021-10-26 16:36 ` Bart Van Assche
2021-10-26 17:19 ` James Bottomley
2021-10-26 17:25 ` Jens Axboe
2021-10-26 18:05 ` Bart Van Assche
2021-10-26 18:10 ` Jens Axboe
2021-10-26 18:18 ` James Bottomley
2021-10-26 18:27 ` Martin K. Petersen
2021-10-26 20:10 ` Bart Van Assche
2021-10-27 5:27 ` Christoph Hellwig
2021-10-27 12:20 ` James Bottomley
2021-10-28 20:21 ` Bart Van Assche
2021-10-28 20:33 ` James Bottomley
2021-10-28 20:53 ` Bart Van Assche
[not found] ` <CGME20211028205342epcas2p40838e84438572adf052d106dc82e35ff@epcms2p6>
2021-10-28 21:14 ` Daejun Park
2021-10-27 13:16 ` Bart Van Assche
2021-10-27 14:12 ` Keith Busch
2021-10-27 14:38 ` Jens Axboe
2021-10-27 14:43 ` James Bottomley
2021-10-27 15:03 ` Ming Lei
2021-10-27 15:06 ` Jens Axboe
2021-10-27 15:16 ` Ming Lei
2021-10-27 15:44 ` Martin K. Petersen
2021-10-27 15:58 ` Ming Lei
2021-10-27 16:16 ` Keith Busch
2021-10-27 16:19 ` Jens Axboe
2021-10-28 0:42 ` Ming Lei
[not found] ` <CGME20211028004244epcas2p1f2212bf94ef861dfa6cd082c3cbb1803@epcms2p1>
2021-10-28 1:10 ` Daejun Park
2021-10-28 2:07 ` Ming Lei [this message]
2021-10-27 16:59 ` Martin K. Petersen
2021-10-27 15:35 ` Martin K. Petersen
2021-10-27 15:40 ` Jens Axboe
2021-10-27 16:16 ` Martin K. Petersen
2021-10-27 17:01 ` Bart Van Assche
2021-10-28 1:32 ` Ming Lei
[not found] ` <CGME20211026201057epcas2p4174ba542fd5abe7ec8f4469f8c60303a@epcms2p7>
2021-10-26 22:22 ` Daejun Park
2021-10-29 10:53 ` Christoph Hellwig
2021-10-29 11:39 ` James Bottomley
2021-10-29 13:35 ` Avri Altman
2021-10-29 13:44 ` James Bottomley
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YXoF59XeZ5KS0jZj@T590 \
--to=ming.lei@redhat.com \
--cc=James.Bottomley@hansenpartnership.com \
--cc=alim.akhtar@samsung.com \
--cc=avri.altman@wdc.com \
--cc=axboe@kernel.dk \
--cc=bvanassche@acm.org \
--cc=daejun7.park@samsung.com \
--cc=hch@lst.de \
--cc=jaegeuk@kernel.org \
--cc=kbusch@kernel.org \
--cc=linux-block@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).