* [PATCH] ublk: move cold paths out of __ublk_batch_dispatch() for icache efficiency
@ 2026-03-18 1:41 Ming Lei
2026-03-23 1:04 ` Jens Axboe
0 siblings, 1 reply; 2+ messages in thread
From: Ming Lei @ 2026-03-18 1:41 UTC (permalink / raw)
To: Jens Axboe, linux-block; +Cc: Caleb Sander Mateos, Ming Lei
Mark ublk_filter_unused_tags() as noinline since it is only called from
the unlikely(needs_filter) branch. Extract the error-handling block from
__ublk_batch_dispatch() into a new noinline ublk_batch_dispatch_fail()
function to keep the hot path compact and icache-friendly. This also
makes __ublk_batch_dispatch() more readable by separating the error
recovery logic from the normal dispatch flow.
Before: __ublk_batch_dispatch is ~1419 bytes
After: __ublk_batch_dispatch is ~1090 bytes (-329 bytes, -23%)
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
drivers/block/ublk_drv.c | 70 ++++++++++++++++++++++------------------
1 file changed, 38 insertions(+), 32 deletions(-)
diff --git a/drivers/block/ublk_drv.c b/drivers/block/ublk_drv.c
index 63aeb7a76a8c..dc8720ad6115 100644
--- a/drivers/block/ublk_drv.c
+++ b/drivers/block/ublk_drv.c
@@ -1789,7 +1789,7 @@ static bool ublk_batch_prep_dispatch(struct ublk_queue *ubq,
* Filter out UBLK_BATCH_IO_UNUSED_TAG entries from tag_buf.
* Returns the new length after filtering.
*/
-static unsigned int ublk_filter_unused_tags(unsigned short *tag_buf,
+static noinline unsigned int ublk_filter_unused_tags(unsigned short *tag_buf,
unsigned int len)
{
unsigned int i, j;
@@ -1805,6 +1805,41 @@ static unsigned int ublk_filter_unused_tags(unsigned short *tag_buf,
return j;
}
+static noinline void ublk_batch_dispatch_fail(struct ublk_queue *ubq,
+ const struct ublk_batch_io_data *data,
+ unsigned short *tag_buf, size_t len, int ret)
+{
+ int i, res;
+
+ /*
+ * Undo prep state for all IOs since userspace never received them.
+ * This restores IOs to pre-prepared state so they can be cleanly
+ * re-prepared when tags are pulled from FIFO again.
+ */
+ for (i = 0; i < len; i++) {
+ struct ublk_io *io = &ubq->ios[tag_buf[i]];
+ int index = -1;
+
+ ublk_io_lock(io);
+ if (io->flags & UBLK_IO_FLAG_AUTO_BUF_REG)
+ index = io->buf.auto_reg.index;
+ io->flags &= ~(UBLK_IO_FLAG_OWNED_BY_SRV | UBLK_IO_FLAG_AUTO_BUF_REG);
+ io->flags |= UBLK_IO_FLAG_ACTIVE;
+ ublk_io_unlock(io);
+
+ if (index != -1)
+ io_buffer_unregister_bvec(data->cmd, index,
+ data->issue_flags);
+ }
+
+ res = kfifo_in_spinlocked_noirqsave(&ubq->evts_fifo,
+ tag_buf, len, &ubq->evts_lock);
+
+ pr_warn_ratelimited("%s: copy tags or post CQE failure, move back "
+ "tags(%d %zu) ret %d\n", __func__, res, len,
+ ret);
+}
+
#define MAX_NR_TAG 128
static int __ublk_batch_dispatch(struct ublk_queue *ubq,
const struct ublk_batch_io_data *data,
@@ -1848,37 +1883,8 @@ static int __ublk_batch_dispatch(struct ublk_queue *ubq,
sel.val = ublk_batch_copy_io_tags(fcmd, sel.addr, tag_buf, len * tag_sz);
ret = ublk_batch_fetch_post_cqe(fcmd, &sel, data->issue_flags);
- if (unlikely(ret < 0)) {
- int i, res;
-
- /*
- * Undo prep state for all IOs since userspace never received them.
- * This restores IOs to pre-prepared state so they can be cleanly
- * re-prepared when tags are pulled from FIFO again.
- */
- for (i = 0; i < len; i++) {
- struct ublk_io *io = &ubq->ios[tag_buf[i]];
- int index = -1;
-
- ublk_io_lock(io);
- if (io->flags & UBLK_IO_FLAG_AUTO_BUF_REG)
- index = io->buf.auto_reg.index;
- io->flags &= ~(UBLK_IO_FLAG_OWNED_BY_SRV | UBLK_IO_FLAG_AUTO_BUF_REG);
- io->flags |= UBLK_IO_FLAG_ACTIVE;
- ublk_io_unlock(io);
-
- if (index != -1)
- io_buffer_unregister_bvec(data->cmd, index,
- data->issue_flags);
- }
-
- res = kfifo_in_spinlocked_noirqsave(&ubq->evts_fifo,
- tag_buf, len, &ubq->evts_lock);
-
- pr_warn_ratelimited("%s: copy tags or post CQE failure, move back "
- "tags(%d %zu) ret %d\n", __func__, res, len,
- ret);
- }
+ if (unlikely(ret < 0))
+ ublk_batch_dispatch_fail(ubq, data, tag_buf, len, ret);
return ret;
}
--
2.47.0
^ permalink raw reply related [flat|nested] 2+ messages in thread* Re: [PATCH] ublk: move cold paths out of __ublk_batch_dispatch() for icache efficiency
2026-03-18 1:41 [PATCH] ublk: move cold paths out of __ublk_batch_dispatch() for icache efficiency Ming Lei
@ 2026-03-23 1:04 ` Jens Axboe
0 siblings, 0 replies; 2+ messages in thread
From: Jens Axboe @ 2026-03-23 1:04 UTC (permalink / raw)
To: linux-block, Ming Lei; +Cc: Caleb Sander Mateos
On Wed, 18 Mar 2026 09:41:12 +0800, Ming Lei wrote:
> Mark ublk_filter_unused_tags() as noinline since it is only called from
> the unlikely(needs_filter) branch. Extract the error-handling block from
> __ublk_batch_dispatch() into a new noinline ublk_batch_dispatch_fail()
> function to keep the hot path compact and icache-friendly. This also
> makes __ublk_batch_dispatch() more readable by separating the error
> recovery logic from the normal dispatch flow.
>
> [...]
Applied, thanks!
[1/1] ublk: move cold paths out of __ublk_batch_dispatch() for icache efficiency
commit: 24d4c90286b9a36a2b72d1e0ceeae237d427f975
Best regards,
--
Jens Axboe
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-03-23 1:04 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-18 1:41 [PATCH] ublk: move cold paths out of __ublk_batch_dispatch() for icache efficiency Ming Lei
2026-03-23 1:04 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox