[for-next] io_uring: clean up cqe trace points

Commit Message

Pavel Begunkov Oct. 18, 2024, 4:14 p.m. UTC

We have too many helpers posting CQEs, instead of tracing completion
events before filling in a CQE and thus having to pass all the data,
set the CQE first, pass it to the tracing helper and let it extract
everything it needs.

Signed-off-by: Pavel Begunkov <[e-mail]>
---
 include/linux/io_uring_types.h  |  5 +++++
 include/trace/events/io_uring.h | 24 +++++++++---------------
 io_uring/io_uring.c             |  4 ++--
 io_uring/io_uring.h             |  7 +++----
 4 files changed, 19 insertions(+), 21 deletions(-)

Comments

Jens Axboe Oct. 18, 2024, 6:37 p.m. UTC | #1

On Fri, 18 Oct 2024 17:14:00 +0100, Pavel Begunkov wrote:
> We have too many helpers posting CQEs, instead of tracing completion
> events before filling in a CQE and thus having to pass all the data,
> set the CQE first, pass it to the tracing helper and let it extract
> everything it needs.
> 
> 

Applied, thanks!

[1/1] io_uring: clean up cqe trace points
      commit: 525b26d3647037195dae188b6d706eed1fa9e5fb

Best regards,

Patch

diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index 4b9ba523978d..b1869d7bdab5 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -667,4 +667,9 @@  struct io_overflow_cqe {
 	struct io_uring_cqe cqe;
 };
 
+static inline bool io_ctx_cqe32(struct io_ring_ctx *ctx)
+{
+	return ctx->flags & IORING_SETUP_CQE32;
+}
+
 #endif
diff --git a/include/trace/events/io_uring.h b/include/trace/events/io_uring.h
index 412c9c210a32..fb81c533b310 100644
--- a/include/trace/events/io_uring.h
+++ b/include/trace/events/io_uring.h
@@ -315,20 +315,14 @@  TRACE_EVENT(io_uring_fail_link,
  * io_uring_complete - called when completing an SQE
  *
  * @ctx:		pointer to a ring context structure
- * @req:		pointer to a submitted request
- * @user_data:		user data associated with the request
- * @res:		result of the request
- * @cflags:		completion flags
- * @extra1:		extra 64-bit data for CQE32
- * @extra2:		extra 64-bit data for CQE32
- *
+ * @req:		(optional) pointer to a submitted request
+ * @cqe:		pointer to the filled in CQE being posted
  */
 TRACE_EVENT(io_uring_complete,
 
-	TP_PROTO(void *ctx, void *req, u64 user_data, int res, unsigned cflags,
-		 u64 extra1, u64 extra2),
+TP_PROTO(struct io_ring_ctx *ctx, void *req, struct io_uring_cqe *cqe),
 
-	TP_ARGS(ctx, req, user_data, res, cflags, extra1, extra2),
+	TP_ARGS(ctx, req, cqe),
 
 	TP_STRUCT__entry (
 		__field(  void *,	ctx		)
@@ -343,11 +337,11 @@  TRACE_EVENT(io_uring_complete,
 	TP_fast_assign(
 		__entry->ctx		= ctx;
 		__entry->req		= req;
-		__entry->user_data	= user_data;
-		__entry->res		= res;
-		__entry->cflags		= cflags;
-		__entry->extra1		= extra1;
-		__entry->extra2		= extra2;
+		__entry->user_data	= cqe->user_data;
+		__entry->res		= cqe->res;
+		__entry->cflags		= cqe->flags;
+		__entry->extra1		= io_ctx_cqe32(ctx) ? cqe->big_cqe[0] : 0;
+		__entry->extra2		= io_ctx_cqe32(ctx) ? cqe->big_cqe[1] : 0;
 	),
 
 	TP_printk("ring %p, req %p, user_data 0x%llx, result %d, cflags 0x%x "
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index e253f0176d0a..9caef6a6ca28 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -823,8 +823,6 @@  static bool io_fill_cqe_aux(struct io_ring_ctx *ctx, u64 user_data, s32 res,
 	 * the ring.
 	 */
 	if (likely(io_get_cqe(ctx, &cqe))) {
-		trace_io_uring_complete(ctx, NULL, user_data, res, cflags, 0, 0);
-
 		WRITE_ONCE(cqe->user_data, user_data);
 		WRITE_ONCE(cqe->res, res);
 		WRITE_ONCE(cqe->flags, cflags);
@@ -833,6 +831,8 @@  static bool io_fill_cqe_aux(struct io_ring_ctx *ctx, u64 user_data, s32 res,
 			WRITE_ONCE(cqe->big_cqe[0], 0);
 			WRITE_ONCE(cqe->big_cqe[1], 0);
 		}
+
+		trace_io_uring_complete(ctx, NULL, cqe);
 		return true;
 	}
 	return false;
diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h
index 70b6675941ff..9cd9a127e9ed 100644
--- a/io_uring/io_uring.h
+++ b/io_uring/io_uring.h
@@ -189,16 +189,15 @@  static __always_inline bool io_fill_cqe_req(struct io_ring_ctx *ctx,
 	if (unlikely(!io_get_cqe(ctx, &cqe)))
 		return false;
 
-	if (trace_io_uring_complete_enabled())
-		trace_io_uring_complete(req->ctx, req, req->cqe.user_data,
-					req->cqe.res, req->cqe.flags,
-					req->big_cqe.extra1, req->big_cqe.extra2);
 
 	memcpy(cqe, &req->cqe, sizeof(*cqe));
 	if (ctx->flags & IORING_SETUP_CQE32) {
 		memcpy(cqe->big_cqe, &req->big_cqe, sizeof(*cqe));
 		memset(&req->big_cqe, 0, sizeof(req->big_cqe));
 	}
+
+	if (trace_io_uring_complete_enabled())
+		trace_io_uring_complete(req->ctx, req, cqe);
 	return true;
 }