public inbox for io-uring@vger.kernel.org
 help / color / mirror / Atom feed
From: Pavel Begunkov <asml.silence@gmail.com>
To: io-uring@vger.kernel.org
Cc: asml.silence@gmail.com
Subject: [RFC 13/16] io_uring: refactor rings_size()
Date: Thu,  6 Nov 2025 17:01:52 +0000	[thread overview]
Message-ID: <e1698a6bf952053026183bc2150a05822f844615.1762447538.git.asml.silence@gmail.com> (raw)
In-Reply-To: <cover.1762447538.git.asml.silence@gmail.com>

Refactor rings_size() to facilitate further changes. Move entries checks
to the beginning, which localises offset calculation. And instead of
special casing IORING_SETUP_NO_SQARRAY, reverse the check and just
continue offset calculation. This way it assign cq_comp_size only once
at the end.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 io_uring/io_uring.c | 31 +++++++++++++++----------------
 1 file changed, 15 insertions(+), 16 deletions(-)

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 4f38a0b587fd..ff52d9f110ce 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -2761,6 +2761,15 @@ int rings_size(unsigned int flags, unsigned int sq_entries,
 {
 	size_t cqe_size, off, sqe_size;
 
+	if (flags & IORING_SETUP_CQE_MIXED) {
+		if (cq_entries < 2)
+			return -EOVERFLOW;
+	}
+	if (flags & IORING_SETUP_SQE_MIXED) {
+		if (sq_entries < 2)
+			return -EOVERFLOW;
+	}
+
 	dims->sq_array_offset = SIZE_MAX;
 
 	sqe_size = sizeof(struct io_uring_sqe);
@@ -2786,29 +2795,19 @@ int rings_size(unsigned int flags, unsigned int sq_entries,
 	if (off == SIZE_MAX)
 		return -EOVERFLOW;
 
-	if (flags & IORING_SETUP_CQE_MIXED) {
-		if (cq_entries < 2)
-			return -EOVERFLOW;
-	}
-	if (flags & IORING_SETUP_SQE_MIXED) {
-		if (sq_entries < 2)
-			return -EOVERFLOW;
-	}
-
 #ifdef CONFIG_SMP
 	off = ALIGN(off, SMP_CACHE_BYTES);
 	if (off == 0)
 		return -EOVERFLOW;
 #endif
 
-	if (flags & IORING_SETUP_NO_SQARRAY) {
-		dims->cq_comp_size = off;
-		return 0;
-	}
+	if (!(flags & IORING_SETUP_NO_SQARRAY)) {
+		dims->sq_array_offset = off;
 
-	dims->sq_array_offset = off;
-	if (check_add_overflow(off, dims->sq_array_size, &off))
-		return -EOVERFLOW;
+		off = size_add(off, dims->sq_array_size);
+		if (off == SIZE_MAX)
+			return -EOVERFLOW;
+	}
 
 	dims->cq_comp_size = off;
 	return 0;
-- 
2.49.0


  parent reply	other threads:[~2025-11-06 17:02 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-06 17:01 [RFC 00/16] Introduce ring flexible placement Pavel Begunkov
2025-11-06 17:01 ` [RFC 01/16] io_uring: add helper calculating region byte size Pavel Begunkov
2025-11-06 17:01 ` [RFC 02/16] io_uring: pass sq entires in the params struct Pavel Begunkov
2025-11-06 17:01 ` [RFC 03/16] io_uring: use mem_is_zero to check ring params Pavel Begunkov
2025-11-06 17:01 ` [RFC 04/16] io_uring: move flags check to io_uring_sanitise_params Pavel Begunkov
2025-11-06 17:01 ` [RFC 05/16] io_uring: introduce struct io_ctx_config Pavel Begunkov
2025-11-06 17:01 ` [RFC 06/16] io_uring: split out config init helper Pavel Begunkov
2025-11-06 17:01 ` [RFC 07/16] io_uring: add structure keeping ring offsets Pavel Begunkov
2025-11-06 17:01 ` [RFC 08/16] io_uring: pre-calculate scq offsets Pavel Begunkov
2025-11-06 17:01 ` [RFC 09/16] io_uring: inroduce helper for setting user offset Pavel Begunkov
2025-11-06 17:01 ` [RFC 10/16] io_uring: separate cqe array from headers Pavel Begunkov
2025-11-06 17:01 ` [RFC 11/16] io_uring/region: introduce io_region_slice Pavel Begunkov
2025-11-06 17:01 ` [RFC 12/16] io_uring: convert pointer init to io_region_slice Pavel Begunkov
2025-11-06 17:01 ` Pavel Begunkov [this message]
2025-11-06 17:01 ` [RFC 14/16] io_uring: extract io_create_mem_region Pavel Begunkov
2025-11-06 17:01 ` [RFC 15/16] io_uring: allow creating mem region at setup Pavel Begunkov
2025-11-06 17:01 ` [RFC 16/16] io_uring: introduce SCQ placement Pavel Begunkov

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=e1698a6bf952053026183bc2150a05822f844615.1762447538.git.asml.silence@gmail.com \
    --to=asml.silence@gmail.com \
    --cc=io-uring@vger.kernel.org \
    /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: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox