* [PATCH for-next] io_uring: optimise io_req_init() sqe flags checks
@ 2021-09-15 11:03 Pavel Begunkov
2021-09-16 1:54 ` Jens Axboe
0 siblings, 1 reply; 2+ messages in thread
From: Pavel Begunkov @ 2021-09-15 11:03 UTC (permalink / raw)
To: Jens Axboe, io-uring
IOSQE_IO_DRAIN is quite marginal and we don't care too much about
IOSQE_BUFFER_SELECT. Save to ifs and hide both of them under
SQE_VALID_FLAGS check. Now we first check whether it uses a "safe"
subset, i.e. without DRAIN and BUFFER_SELECT, and only if it's not
true we test the rest of the flags.
Signed-off-by: Pavel Begunkov <[email protected]>
---
fs/io_uring.c | 29 +++++++++++++++++------------
1 file changed, 17 insertions(+), 12 deletions(-)
diff --git a/fs/io_uring.c b/fs/io_uring.c
index f4dcdda9d4ae..511fb8052ae9 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -103,9 +103,11 @@
#define IORING_MAX_REG_BUFFERS (1U << 14)
-#define SQE_VALID_FLAGS (IOSQE_FIXED_FILE|IOSQE_IO_DRAIN|IOSQE_IO_LINK| \
- IOSQE_IO_HARDLINK | IOSQE_ASYNC | \
- IOSQE_BUFFER_SELECT)
+#define SQE_COMMON_FLAGS (IOSQE_FIXED_FILE | IOSQE_IO_LINK | \
+ IOSQE_IO_HARDLINK | IOSQE_ASYNC)
+
+#define SQE_VALID_FLAGS (SQE_COMMON_FLAGS|IOSQE_BUFFER_SELECT|IOSQE_IO_DRAIN)
+
#define IO_REQ_CLEAN_FLAGS (REQ_F_BUFFER_SELECTED | REQ_F_NEED_CLEANUP | \
REQ_F_POLLED | REQ_F_INFLIGHT | REQ_F_CREDS)
@@ -7044,20 +7046,21 @@ static int io_init_req(struct io_ring_ctx *ctx, struct io_kiocb *req,
req->fixed_rsrc_refs = NULL;
req->task = current;
- /* enforce forwards compatibility on users */
- if (unlikely(sqe_flags & ~SQE_VALID_FLAGS))
- return -EINVAL;
if (unlikely(req->opcode >= IORING_OP_LAST))
return -EINVAL;
+ if (unlikely(sqe_flags & ~SQE_COMMON_FLAGS)) {
+ /* enforce forwards compatibility on users */
+ if (sqe_flags & ~SQE_VALID_FLAGS)
+ return -EINVAL;
+ if ((sqe_flags & IOSQE_BUFFER_SELECT) &&
+ !io_op_defs[req->opcode].buffer_select)
+ return -EOPNOTSUPP;
+ if (sqe_flags & IOSQE_IO_DRAIN)
+ ctx->drain_active = true;
+ }
if (!io_check_restriction(ctx, req, sqe_flags))
return -EACCES;
- if ((sqe_flags & IOSQE_BUFFER_SELECT) &&
- !io_op_defs[req->opcode].buffer_select)
- return -EOPNOTSUPP;
- if (unlikely(sqe_flags & IOSQE_IO_DRAIN))
- ctx->drain_active = true;
-
personality = READ_ONCE(sqe->personality);
if (personality) {
req->creds = xa_load(&ctx->personalities, personality);
@@ -10917,6 +10920,8 @@ static int __init io_uring_init(void)
/* should fit into one byte */
BUILD_BUG_ON(SQE_VALID_FLAGS >= (1 << 8));
+ BUILD_BUG_ON(SQE_COMMON_FLAGS >= (1 << 8));
+ BUILD_BUG_ON((SQE_VALID_FLAGS | SQE_COMMON_FLAGS) != SQE_VALID_FLAGS);
BUILD_BUG_ON(ARRAY_SIZE(io_op_defs) != IORING_OP_LAST);
BUILD_BUG_ON(__REQ_F_LAST_BIT > 8 * sizeof(int));
--
2.33.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH for-next] io_uring: optimise io_req_init() sqe flags checks
2021-09-15 11:03 [PATCH for-next] io_uring: optimise io_req_init() sqe flags checks Pavel Begunkov
@ 2021-09-16 1:54 ` Jens Axboe
0 siblings, 0 replies; 2+ messages in thread
From: Jens Axboe @ 2021-09-16 1:54 UTC (permalink / raw)
To: Pavel Begunkov, io-uring
On 9/15/21 5:03 AM, Pavel Begunkov wrote:
> IOSQE_IO_DRAIN is quite marginal and we don't care too much about
> IOSQE_BUFFER_SELECT. Save to ifs and hide both of them under
> SQE_VALID_FLAGS check. Now we first check whether it uses a "safe"
> subset, i.e. without DRAIN and BUFFER_SELECT, and only if it's not
> true we test the rest of the flags.
Applied, thanks.
--
Jens Axboe
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-09-16 1:54 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-09-15 11:03 [PATCH for-next] io_uring: optimise io_req_init() sqe flags checks Pavel Begunkov
2021-09-16 1:54 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox