From: Jens Axboe <axboe@kernel.dk>
To: Christoph Hellwig <hch@infradead.org>
Cc: Shaohua Li <shli@kernel.org>, linux-kernel@vger.kernel.org
Subject: Re: [patch]blk-mq: blk_mq_tag_to_rq should handle flush request
Date: Wed, 04 Jun 2014 09:47:14 -0600 [thread overview]
Message-ID: <538F3F82.5060805@kernel.dk> (raw)
In-Reply-To: <538F3DC4.4030104@kernel.dk>
[-- Attachment #1: Type: text/plain, Size: 1312 bytes --]
On 06/04/2014 09:39 AM, Jens Axboe wrote:
> On 06/04/2014 09:31 AM, Christoph Hellwig wrote:
>> On Wed, Jun 04, 2014 at 09:02:19AM -0600, Jens Axboe wrote:
>>>> scsi_mq_find_tag only gets the scsi host, which may have multiple
>>>> queues. When called from scsi_find_tag we actually have a scsi device,
>>>> so that's not an issue, but when called from scsi_host_find_tag the
>>>> driver only provides the host.
>>>
>>> Only solution I see right now is to have the flush_rq in the shared
>>> tags, but that would potentially be a regression for multiple
>>> devices and heavy flush uses cases. I'll see if I can come up with
>>> something better, or maybe Shaohua has an idea.
>>
>> What about something like the following (untest, uncompiled, maybe
>> pseudo-code):
>>
>> struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag)
>> {
>> struct request *rq = tags->rqs[tag];
>>
>> if ((rq->cmd_flags & REQ_FLUSH_SEQ) && rq->q->flush_rq->tag == tag)
>> return rq->q->flush_rq;
>> return rq;
>
> Ah yes, that'll work, the queue is always assigned. I'll make that change.
Something like this in complete form. Compile tested only, I'll test it
on dev box. Probably doesn't matter too much, but I prefer to
potentially have the faster path (non-flush) just fall inline.
--
Jens Axboe
[-- Attachment #2: tag-to-request.patch --]
[-- Type: text/x-patch, Size: 2579 bytes --]
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 4e8e8cf00815..4e4cd6208052 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -529,15 +529,20 @@ void blk_mq_kick_requeue_list(struct request_queue *q)
}
EXPORT_SYMBOL(blk_mq_kick_requeue_list);
-struct request *blk_mq_tag_to_rq(struct blk_mq_hw_ctx *hctx, unsigned int tag)
+static inline bool is_flush_request(struct request *rq, unsigned int tag)
{
- struct request_queue *q = hctx->queue;
+ return ((rq->cmd_flags & REQ_FLUSH_SEQ) &&
+ rq->q->flush_rq->tag == tag);
+}
- if ((q->flush_rq->cmd_flags & REQ_FLUSH_SEQ) &&
- q->flush_rq->tag == tag)
- return q->flush_rq;
+struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag)
+{
+ struct request *rq = tags->rqs[tag];
+
+ if (!is_flush_request(rq, tag))
+ return rq;
- return hctx->tags->rqs[tag];
+ return rq->q->flush_rq;
}
EXPORT_SYMBOL(blk_mq_tag_to_rq);
@@ -566,7 +571,7 @@ static void blk_mq_timeout_check(void *__data, unsigned long *free_tags)
if (tag >= hctx->tags->nr_tags)
break;
- rq = blk_mq_tag_to_rq(hctx, tag++);
+ rq = blk_mq_tag_to_rq(hctx->tags, tag++);
if (rq->q != hctx->queue)
continue;
if (!test_bit(REQ_ATOM_STARTED, &rq->atomic_flags))
diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c
index abc858b3528b..74abd49fabdc 100644
--- a/drivers/block/mtip32xx/mtip32xx.c
+++ b/drivers/block/mtip32xx/mtip32xx.c
@@ -193,7 +193,9 @@ static void mtip_put_int_command(struct driver_data *dd, struct mtip_cmd *cmd)
static struct request *mtip_rq_from_tag(struct driver_data *dd,
unsigned int tag)
{
- return blk_mq_tag_to_rq(dd->queue->queue_hw_ctx[0], tag);
+ struct blk_mq_hw_ctx *hctx = dd->queue->queue_hw_ctx[0];
+
+ return blk_mq_tag_to_rq(hctx->tags, tag);
}
static struct mtip_cmd *mtip_cmd_from_tag(struct driver_data *dd,
diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h
index c15128833100..0feedebfde48 100644
--- a/include/linux/blk-mq.h
+++ b/include/linux/blk-mq.h
@@ -155,7 +155,7 @@ void blk_mq_free_request(struct request *rq);
bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
struct request *blk_mq_alloc_request(struct request_queue *q, int rw,
gfp_t gfp, bool reserved);
-struct request *blk_mq_tag_to_rq(struct blk_mq_hw_ctx *hctx, unsigned int tag);
+struct request *blk_mq_tag_to_rq(struct blk_mq_tags *tags, unsigned int tag);
struct blk_mq_hw_ctx *blk_mq_map_queue(struct request_queue *, const int ctx_index);
struct blk_mq_hw_ctx *blk_mq_alloc_single_hw_queue(struct blk_mq_tag_set *, unsigned int, int);
next prev parent reply other threads:[~2014-06-04 15:47 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-05-09 12:07 [patch]blk-mq: blk_mq_tag_to_rq should handle flush request Shaohua Li
2014-05-09 15:00 ` Christoph Hellwig
2014-05-10 4:00 ` Shaohua Li
2014-05-11 17:40 ` Jens Axboe
2014-05-30 14:09 ` Jens Axboe
2014-06-04 11:11 ` Christoph Hellwig
2014-06-04 14:15 ` Jens Axboe
2014-06-04 14:20 ` Christoph Hellwig
2014-06-04 14:54 ` Jens Axboe
2014-06-04 14:58 ` Christoph Hellwig
2014-06-04 15:02 ` Jens Axboe
2014-06-04 15:05 ` Christoph Hellwig
2014-06-04 15:08 ` Jens Axboe
2014-06-04 15:10 ` Christoph Hellwig
2014-06-04 15:11 ` Jens Axboe
2014-06-04 15:16 ` Christoph Hellwig
2014-06-04 15:19 ` Jens Axboe
2014-06-04 15:22 ` Christoph Hellwig
2014-06-04 15:28 ` Jens Axboe
2014-06-04 15:31 ` Christoph Hellwig
2014-06-04 15:39 ` Jens Axboe
2014-06-04 15:47 ` Jens Axboe [this message]
2014-06-04 16:25 ` Jens Axboe
2014-06-05 1:27 ` Shaohua Li
2014-06-05 2:05 ` Jens Axboe
2014-06-05 2:27 ` Shaohua Li
2014-06-05 2:40 ` Jens Axboe
2014-06-04 15:43 ` Ming Lei
2014-06-04 15:48 ` Jens Axboe
2014-06-04 16:00 ` Ming Lei
2014-06-04 16:09 ` Jens Axboe
2014-06-04 16:26 ` Ming Lei
2014-06-04 16:28 ` Jens Axboe
2014-06-04 16:33 ` Christoph Hellwig
2014-06-04 16:36 ` Ming Lei
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=538F3F82.5060805@kernel.dk \
--to=axboe@kernel.dk \
--cc=hch@infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=shli@kernel.org \
/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