Skip to contentSkip to conversation

io_uring: fix wrong condition to grab uring lock

PATCH Applied Hao Xu 3 participants 4 messages Oct 14, 2021

Subsystems (2)
lore
Share & tools
lore

Hao Xu

author
#0 lore raw

PATCH 1/1 io_uring: fix wrong condition to grab uring lock

Grab uring lock when we are in io-worker rather than in the original or system-wq context since we already hold it in these two situation.

---
 fs/io_uring.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/io_uring.c +1 -1
index 73135c5c6168..e2ed21c65f71 100644
@@ -2890,7 +2890,7 @@ static void kiocb_done(struct kiocb *kiocb, ssize_t ret,
 			struct io_ring_ctx *ctx = req->ctx;
 
 			req_set_fail(req);
-			if (issue_flags & IO_URING_F_NONBLOCK) {
+			if (!(issue_flags & IO_URING_F_NONBLOCK)) {
 				mutex_lock(&ctx->uring_lock);
 				__io_req_complete(req, issue_flags, ret, cflags);
 				mutex_unlock(&ctx->uring_lock);
Signature · 1 line
2.24.4
author
#1 lore raw
Quoted 3 lines · Hao Xu

Grab uring lock when we are in io-worker rather than in the original or system-wq context since we already hold it in these two situation.

Fixes b66ceaf324b3 ("io_uring: move iopoll reissue into regular IO path")
Quoted 19 lines
---
  fs/io_uring.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
fs/io_uring.c +1 -1
index 73135c5c6168..e2ed21c65f71 100644
@@ -2890,7 +2890,7 @@ static void kiocb_done(struct kiocb *kiocb, ssize_t ret,
  			struct io_ring_ctx *ctx = req->ctx;
 
  			req_set_fail(req);
-			if (issue_flags & IO_URING_F_NONBLOCK) {
+			if (!(issue_flags & IO_URING_F_NONBLOCK)) {
  				mutex_lock(&ctx->uring_lock);
  				__io_req_complete(req, issue_flags, ret, cflags);
  				mutex_unlock(&ctx->uring_lock);
#2 lore raw
In reply to Hao Xu
Quoted 2 lines · Hao Xu

Grab uring lock when we are in io-worker rather than in the original or system-wq context since we already hold it in these two situation.

Good catch, lgtm

Quoted 19 lines
---
  fs/io_uring.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
fs/io_uring.c +1 -1
index 73135c5c6168..e2ed21c65f71 100644
@@ -2890,7 +2890,7 @@ static void kiocb_done(struct kiocb *kiocb, ssize_t ret,
  			struct io_ring_ctx *ctx = req->ctx;
 
  			req_set_fail(req);
-			if (issue_flags & IO_URING_F_NONBLOCK) {
+			if (!(issue_flags & IO_URING_F_NONBLOCK)) {
  				mutex_lock(&ctx->uring_lock);
  				__io_req_complete(req, issue_flags, ret, cflags);
  				mutex_unlock(&ctx->uring_lock);
Signature · 1 line
Pavel Begunkov
maintainer Applied
#3 lore raw
In reply to Hao Xu
Quoted 4 lines · Hao Xu

Grab uring lock when we are in io-worker rather than in the original or system-wq context since we already hold it in these two situation.

Applied, thanks!

[1/1] io_uring: fix wrong condition to grab uring lock

      commit: 14cfbb7a7856f190035f8e53045bdbfa648fae41

Best regards,

Signature · 1 line
Jens Axboe