From: Ming Lei <ming.lei@redhat.com> To: Christoph Hellwig <hch@infradead.org> Cc: Jens Axboe <axboe@kernel.dk>, linux-block@vger.kernel.org, Jeffle Xu <jefflexu@linux.alibaba.com>, Mike Snitzer <snitzer@redhat.com>, dm-devel@redhat.com, Hannes Reinecke <hare@suse.de> Subject: Re: [PATCH V5 08/12] block: use per-task poll context to implement bio based io polling Date: Mon, 12 Apr 2021 18:37:34 +0800 [thread overview] Message-ID: <YHQi7rZ5StWUpX/r@T590> (raw) In-Reply-To: <20210412101659.GA993044@infradead.org> On Mon, Apr 12, 2021 at 11:16:59AM +0100, Christoph Hellwig wrote: > > +static int blk_bio_poll_io(struct io_context *submit_ioc, > > + struct io_context *poll_ioc) > > Givem that poll_ioc is always current->io_context there is no > need to pass it. Yeah, it is true. > > > + struct blk_bio_poll_ctx *poll_ctx = poll_ioc ? > > + poll_ioc->data : NULL; > > and it really should not be NULL here, should it? Yeah, it is NULL just in case of memory allocation failure. > > > +static int __blk_bio_poll(blk_qc_t cookie) > > +{ > > + struct io_context *poll_ioc = current->io_context; > > + pid_t pid; > > + struct task_struct *submit_task; > > + int ret; > > + > > + pid = (pid_t)cookie; > > + > > + /* io poll often share io submission context */ > > + if (likely(current->pid == pid && blk_bio_ioc_valid(current))) > > + return blk_bio_poll_io(poll_ioc, poll_ioc); > > + > > + submit_task = find_get_task_by_vpid(pid); > > + if (likely(blk_bio_ioc_valid(submit_task))) > > + ret = blk_bio_poll_io(submit_task->io_context, poll_ioc); > > + else > > + ret = 0; > > + if (likely(submit_task)) > > + put_task_struct(submit_task); > > Wouldn't it make more sense to just store the submitting context > in the bio, even if that uses more space? Having to call But where to store the submitting context in bio? pid is 32bit and we can pass it from submit_bio() perfectly, then avoid to add anything to bio. If we save the submitting context in bio, we still have to handle task exit related race, not see any benefit. So far bio is ~128byte with typical setting, and people usually hate to add more stuff into bio. > find_get_task_by_vpid in the poll context seems rather problematic. Why? We already handle submit context early exit. > > Note that this requires doing the refacoring to get rid of the separate > blk_qc_t passed up the stack I asked for earlier, but hiding all these > details seems like a really useful change anyway. That is hard to do because of race between submission and completion: 1) HIPRI could be cleared because of bio splitting, so we can't do that for this kind of bio 2) we have to make sure that the bio won't be completed when storing cookie into this bio. BIO_END_BY_POLL is added in this patch for bio based polling. You mean we may cover all normal blk-mq polling via BIO_END_BY_POLL? Thanks, Ming
WARNING: multiple messages have this Message-ID (diff)
From: Ming Lei <ming.lei@redhat.com> To: Christoph Hellwig <hch@infradead.org> Cc: Jens Axboe <axboe@kernel.dk>, Mike Snitzer <snitzer@redhat.com>, linux-block@vger.kernel.org, dm-devel@redhat.com, Jeffle Xu <jefflexu@linux.alibaba.com> Subject: Re: [dm-devel] [PATCH V5 08/12] block: use per-task poll context to implement bio based io polling Date: Mon, 12 Apr 2021 18:37:34 +0800 [thread overview] Message-ID: <YHQi7rZ5StWUpX/r@T590> (raw) In-Reply-To: <20210412101659.GA993044@infradead.org> On Mon, Apr 12, 2021 at 11:16:59AM +0100, Christoph Hellwig wrote: > > +static int blk_bio_poll_io(struct io_context *submit_ioc, > > + struct io_context *poll_ioc) > > Givem that poll_ioc is always current->io_context there is no > need to pass it. Yeah, it is true. > > > + struct blk_bio_poll_ctx *poll_ctx = poll_ioc ? > > + poll_ioc->data : NULL; > > and it really should not be NULL here, should it? Yeah, it is NULL just in case of memory allocation failure. > > > +static int __blk_bio_poll(blk_qc_t cookie) > > +{ > > + struct io_context *poll_ioc = current->io_context; > > + pid_t pid; > > + struct task_struct *submit_task; > > + int ret; > > + > > + pid = (pid_t)cookie; > > + > > + /* io poll often share io submission context */ > > + if (likely(current->pid == pid && blk_bio_ioc_valid(current))) > > + return blk_bio_poll_io(poll_ioc, poll_ioc); > > + > > + submit_task = find_get_task_by_vpid(pid); > > + if (likely(blk_bio_ioc_valid(submit_task))) > > + ret = blk_bio_poll_io(submit_task->io_context, poll_ioc); > > + else > > + ret = 0; > > + if (likely(submit_task)) > > + put_task_struct(submit_task); > > Wouldn't it make more sense to just store the submitting context > in the bio, even if that uses more space? Having to call But where to store the submitting context in bio? pid is 32bit and we can pass it from submit_bio() perfectly, then avoid to add anything to bio. If we save the submitting context in bio, we still have to handle task exit related race, not see any benefit. So far bio is ~128byte with typical setting, and people usually hate to add more stuff into bio. > find_get_task_by_vpid in the poll context seems rather problematic. Why? We already handle submit context early exit. > > Note that this requires doing the refacoring to get rid of the separate > blk_qc_t passed up the stack I asked for earlier, but hiding all these > details seems like a really useful change anyway. That is hard to do because of race between submission and completion: 1) HIPRI could be cleared because of bio splitting, so we can't do that for this kind of bio 2) we have to make sure that the bio won't be completed when storing cookie into this bio. BIO_END_BY_POLL is added in this patch for bio based polling. You mean we may cover all normal blk-mq polling via BIO_END_BY_POLL? Thanks, Ming -- dm-devel mailing list dm-devel@redhat.com https://listman.redhat.com/mailman/listinfo/dm-devel
next prev parent reply other threads:[~2021-04-12 10:37 UTC|newest] Thread overview: 92+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-04-01 2:19 [PATCH V5 00/12] block: support bio based io polling Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-01 2:19 ` [PATCH V5 01/12] block: add helper of blk_queue_poll Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-01 2:19 ` [PATCH V5 02/12] block: add one helper to free io_context Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-01 2:19 ` [PATCH V5 03/12] block: create io poll context for submission and poll task Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-12 10:19 ` Christoph Hellwig 2021-04-12 10:19 ` [dm-devel] " Christoph Hellwig 2021-04-01 2:19 ` [dm-devel] [PATCH V5 04/12] block: add req flag of REQ_POLL_CTX Ming Lei 2021-04-01 2:19 ` Ming Lei 2021-04-01 2:19 ` [dm-devel] [PATCH V5 05/12] block: add new field into 'struct bvec_iter' Ming Lei 2021-04-01 2:19 ` Ming Lei 2021-04-12 9:26 ` Christoph Hellwig 2021-04-12 9:26 ` [dm-devel] " Christoph Hellwig 2021-04-13 9:36 ` Ming Lei 2021-04-13 9:36 ` [dm-devel] " Ming Lei 2021-04-01 2:19 ` [dm-devel] [PATCH V5 06/12] block/mq: extract one helper function polling hw queue Ming Lei 2021-04-01 2:19 ` Ming Lei 2021-04-12 9:29 ` Christoph Hellwig 2021-04-12 9:29 ` [dm-devel] " Christoph Hellwig 2021-04-01 2:19 ` [dm-devel] [PATCH V5 07/12] block: prepare for supporting bio_list via other link Ming Lei 2021-04-01 2:19 ` Ming Lei 2021-04-12 10:18 ` Christoph Hellwig 2021-04-12 10:18 ` [dm-devel] " Christoph Hellwig 2021-04-12 11:37 ` Ming Lei 2021-04-12 11:37 ` [dm-devel] " Ming Lei 2021-04-01 2:19 ` [dm-devel] [PATCH V5 08/12] block: use per-task poll context to implement bio based io polling Ming Lei 2021-04-01 2:19 ` Ming Lei 2021-04-12 9:54 ` [dm-devel] " Christoph Hellwig 2021-04-12 9:54 ` Christoph Hellwig 2021-04-12 10:20 ` Ming Lei 2021-04-12 10:20 ` [dm-devel] " Ming Lei 2021-04-12 10:29 ` Christoph Hellwig 2021-04-12 10:29 ` [dm-devel] " Christoph Hellwig 2021-04-12 11:42 ` Ming Lei 2021-04-12 11:42 ` [dm-devel] " Ming Lei 2021-04-12 10:16 ` Christoph Hellwig 2021-04-12 10:16 ` [dm-devel] " Christoph Hellwig 2021-04-12 10:37 ` Ming Lei [this message] 2021-04-12 10:37 ` Ming Lei 2021-04-01 2:19 ` [PATCH V5 09/12] blk-mq: limit hw queues to be polled in each blk_poll() Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-01 2:19 ` [PATCH V5 10/12] block: add queue_to_disk() to get gendisk from request_queue Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-12 12:52 ` Jens Axboe 2021-04-12 12:52 ` [dm-devel] " Jens Axboe 2021-04-01 2:19 ` [PATCH V5 11/12] block: add poll_capable method to support bio-based IO polling Ming Lei 2021-04-01 2:19 ` [dm-devel] " Ming Lei 2021-04-12 9:38 ` Christoph Hellwig 2021-04-12 9:38 ` Christoph Hellwig 2021-04-14 8:38 ` JeffleXu 2021-04-14 8:38 ` [dm-devel] " JeffleXu 2021-04-14 11:24 ` Ming Lei 2021-04-14 11:24 ` Ming Lei 2021-04-15 1:34 ` JeffleXu 2021-04-15 1:34 ` [dm-devel] " JeffleXu 2021-04-15 7:43 ` Ming Lei 2021-04-15 7:43 ` [dm-devel] " Ming Lei 2021-04-15 9:21 ` JeffleXu 2021-04-15 9:21 ` [dm-devel] " JeffleXu 2021-04-15 10:06 ` Ming Lei 2021-04-15 10:06 ` [dm-devel] " Ming Lei 2021-04-15 11:21 ` JeffleXu 2021-04-15 11:21 ` [dm-devel] " JeffleXu 2021-04-15 13:08 ` Ming Lei 2021-04-15 13:08 ` [dm-devel] " Ming Lei 2021-04-16 8:00 ` [PATCH] block: introduce QUEUE_FLAG_POLL_CAP flag Jeffle Xu 2021-04-16 8:00 ` [dm-devel] " Jeffle Xu 2021-04-16 8:42 ` JeffleXu 2021-04-16 8:42 ` [dm-devel] " JeffleXu 2021-04-16 9:07 ` Ming Lei 2021-04-16 9:07 ` [dm-devel] " Ming Lei 2021-04-16 10:20 ` JeffleXu 2021-04-16 10:20 ` [dm-devel] " JeffleXu 2021-04-17 14:06 ` JeffleXu 2021-04-17 14:06 ` [dm-devel] " JeffleXu 2021-04-19 2:21 ` Ming Lei 2021-04-19 2:21 ` [dm-devel] " Ming Lei 2021-04-19 5:40 ` JeffleXu 2021-04-19 5:40 ` [dm-devel] " JeffleXu 2021-04-19 13:36 ` Ming Lei 2021-04-19 13:36 ` [dm-devel] " Ming Lei 2021-04-20 7:25 ` JeffleXu 2021-04-20 7:25 ` [dm-devel] " JeffleXu 2021-04-01 2:19 ` [dm-devel] [PATCH V5 12/12] dm: support IO polling for bio-based dm device Ming Lei 2021-04-01 2:19 ` Ming Lei 2021-04-09 15:39 ` [PATCH V5 00/12] block: support bio based io polling Ming Lei 2021-04-09 15:39 ` [dm-devel] " Ming Lei 2021-04-12 9:46 ` Christoph Hellwig 2021-04-12 9:46 ` Christoph Hellwig
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=YHQi7rZ5StWUpX/r@T590 \ --to=ming.lei@redhat.com \ --cc=axboe@kernel.dk \ --cc=dm-devel@redhat.com \ --cc=hare@suse.de \ --cc=hch@infradead.org \ --cc=jefflexu@linux.alibaba.com \ --cc=linux-block@vger.kernel.org \ --cc=snitzer@redhat.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.