* [PATCH 1/1] io_uring: optimise use of ctx->drain_next
@ 2020-01-17 0:57 Pavel Begunkov
2020-01-17 2:59 ` Jens Axboe
0 siblings, 1 reply; 2+ messages in thread
From: Pavel Begunkov @ 2020-01-17 0:57 UTC (permalink / raw)
To: Jens Axboe, io-uring, linux-kernel
Move setting ctx->drain_next to the only place it could be set, when it
got linked non-head requests. The same for checking it, it's interesting
only for a head of a link or a non-linked request.
No functional changes here. This removes some code from the common path
and also removes REQ_F_DRAIN_LINK flag, as it doesn't need it anymore.
Signed-off-by: Pavel Begunkov <[email protected]>
---
fs/io_uring.c | 41 +++++++++++++++++++++--------------------
1 file changed, 21 insertions(+), 20 deletions(-)
diff --git a/fs/io_uring.c b/fs/io_uring.c
index ea91f4d92fc0..2ace3f1962ff 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -506,7 +506,6 @@ struct io_kiocb {
#define REQ_F_LINK 64 /* linked sqes */
#define REQ_F_LINK_TIMEOUT 128 /* has linked timeout */
#define REQ_F_FAIL_LINK 256 /* fail rest of links */
-#define REQ_F_DRAIN_LINK 512 /* link should be fully drained */
#define REQ_F_TIMEOUT 1024 /* timeout request */
#define REQ_F_ISREG 2048 /* regular file */
#define REQ_F_MUST_PUNT 4096 /* must be punted even for NONBLOCK */
@@ -4543,12 +4542,6 @@ static void io_queue_sqe(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
int ret;
- if (unlikely(req->ctx->drain_next)) {
- req->flags |= REQ_F_IO_DRAIN;
- req->ctx->drain_next = 0;
- }
- req->ctx->drain_next = (req->flags & REQ_F_DRAIN_LINK) != 0;
-
ret = io_req_defer(req, sqe);
if (ret) {
if (ret != -EIOCBQUEUED) {
@@ -4615,8 +4608,10 @@ static bool io_submit_sqe(struct io_kiocb *req, const struct io_uring_sqe *sqe,
if (*link) {
struct io_kiocb *head = *link;
- if (sqe_flags & IOSQE_IO_DRAIN)
- head->flags |= REQ_F_DRAIN_LINK | REQ_F_IO_DRAIN;
+ if (sqe_flags & IOSQE_IO_DRAIN) {
+ head->flags |= REQ_F_IO_DRAIN;
+ ctx->drain_next = 1;
+ }
if (sqe_flags & IOSQE_IO_HARDLINK)
req->flags |= REQ_F_HARDLINK;
@@ -4640,18 +4635,24 @@ static bool io_submit_sqe(struct io_kiocb *req, const struct io_uring_sqe *sqe,
io_queue_link_head(head);
*link = NULL;
}
- } else if (sqe_flags & (IOSQE_IO_LINK|IOSQE_IO_HARDLINK)) {
- req->flags |= REQ_F_LINK;
- if (sqe_flags & IOSQE_IO_HARDLINK)
- req->flags |= REQ_F_HARDLINK;
-
- INIT_LIST_HEAD(&req->link_list);
- ret = io_req_defer_prep(req, sqe);
- if (ret)
- req->flags |= REQ_F_FAIL_LINK;
- *link = req;
} else {
- io_queue_sqe(req, sqe);
+ if (unlikely(ctx->drain_next)) {
+ req->flags |= REQ_F_IO_DRAIN;
+ req->ctx->drain_next = 0;
+ }
+ if (sqe_flags & (IOSQE_IO_LINK|IOSQE_IO_HARDLINK)) {
+ req->flags |= REQ_F_LINK;
+ if (sqe_flags & IOSQE_IO_HARDLINK)
+ req->flags |= REQ_F_HARDLINK;
+
+ INIT_LIST_HEAD(&req->link_list);
+ ret = io_req_defer_prep(req, sqe);
+ if (ret)
+ req->flags |= REQ_F_FAIL_LINK;
+ *link = req;
+ } else {
+ io_queue_sqe(req, sqe);
+ }
}
return true;
--
2.24.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH 1/1] io_uring: optimise use of ctx->drain_next
2020-01-17 0:57 [PATCH 1/1] io_uring: optimise use of ctx->drain_next Pavel Begunkov
@ 2020-01-17 2:59 ` Jens Axboe
0 siblings, 0 replies; 2+ messages in thread
From: Jens Axboe @ 2020-01-17 2:59 UTC (permalink / raw)
To: Pavel Begunkov, io-uring, linux-kernel
On 1/16/20 5:57 PM, Pavel Begunkov wrote:
> Move setting ctx->drain_next to the only place it could be set, when it
> got linked non-head requests. The same for checking it, it's interesting
> only for a head of a link or a non-linked request.
>
> No functional changes here. This removes some code from the common path
> and also removes REQ_F_DRAIN_LINK flag, as it doesn't need it anymore.
Applied, thanks.
--
Jens Axboe
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2020-01-17 2:59 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-01-17 0:57 [PATCH 1/1] io_uring: optimise use of ctx->drain_next Pavel Begunkov
2020-01-17 2:59 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox