DMA Engine development
 help / color / mirror / Atom feed
From: Linus Walleij <linusw@kernel.org>
To: Vinod Koul <vkoul@kernel.org>, Frank Li <Frank.Li@kernel.org>
Cc: dmaengine@vger.kernel.org, phone-devel@vger.kernel.org,
	 Linus Walleij <linusw@kernel.org>,
	sashiko-bot@kernel.org,  Frank Li <Frank.Li@nxp.com>
Subject: [PATCH v7 04/24] dmaengine: ste_dma40: Fix failed start cleanup
Date: Sun, 27 Sep 2026 11:03:17 +0200	[thread overview]
Message-ID: <20260927-dma40-fixes-v7-4-89f595e8851d@kernel.org> (raw)
In-Reply-To: <20260927-dma40-fixes-v7-0-89f595e8851d@kernel.org>

If d40_start() fails after a queued descriptor has been moved to the
active list, d40_queue_start() currently returns NULL without unwinding
the transfer state or clearing the channel busy flag.

Fix this pre-existing error path by completing the descriptor through the
normal tasklet path, clearing the busy flag, balancing the runtime PM
reference and returning an error pointer to distinguish the failure from
the no-work case. Do not free the descriptor directly, since it has
already been submitted and has a DMA cookie.

When starting the next queued transfer from the completion handler, put
the completed descriptor on the done list first. This preserves FIFO
completion order if the new transfer fails to start and prevents the
completed cookie from moving backwards.

Use done-list membership rather than the cyclic flag to identify terminal
descriptors in the tasklet. A cyclic descriptor that failed to start is
then completed and removed instead of remaining permanently at the head
of the done list.

Fixes: 7d83a854a1a4 ("dma40: remove "hardware link with previous jobs" code")
Reported-by: sashiko-bot@kernel.org
Closes: https://lore.kernel.org/dmaengine/20260618051539.15E201F000E9@smtp.kernel.org/
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Assisted-by: LLM
Link: https://lore.kernel.org/dmaengine/20260618051539.15E201F000E9@smtp.kernel.org/
Signed-off-by: Linus Walleij <linusw@kernel.org>
---
 drivers/dma/ste_dma40.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 4b138aa313c8..64f7a0ca20b9 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -1659,8 +1659,15 @@ static struct d40_desc *d40_queue_start(struct d40_chan *d40c)
 		/* Start dma job */
 		err = d40_start(d40c);
 
-		if (err)
-			return NULL;
+		if (err) {
+			d40_desc_remove(d40d);
+			d40_desc_done(d40c, d40d);
+			d40c->pending_tx++;
+			d40c->busy = false;
+			pm_runtime_put_autosuspend(d40c->base->dev);
+			tasklet_schedule(&d40c->tasklet);
+			return ERR_PTR(err);
+		}
 	}
 
 	return d40d;
@@ -1707,14 +1714,14 @@ static void dma_tc_handle(struct d40_chan *d40c)
 			return;
 		}
 
+		d40_desc_remove(d40d);
+		d40_desc_done(d40c, d40d);
+
 		if (d40_queue_start(d40c) == NULL) {
 			d40c->busy = false;
 
 			pm_runtime_put_autosuspend(d40c->base->dev);
 		}
-
-		d40_desc_remove(d40d);
-		d40_desc_done(d40c, d40d);
 	}
 
 	d40c->pending_tx += callbacks;
@@ -1728,20 +1735,22 @@ static void dma_tasklet(struct tasklet_struct *t)
 	struct d40_desc *d40d;
 	unsigned long flags;
 	bool callback_active;
+	bool from_done;
 	struct dmaengine_desc_callback cb;
 
 	spin_lock_irqsave(&d40c->lock, flags);
 
 	/* Get first entry from the done list */
 	d40d = d40_first_done(d40c);
-	if (d40d == NULL) {
+	from_done = !!d40d;
+	if (!from_done) {
 		/* Check if we have reached here for cyclic job */
 		d40d = d40_first_active_get(d40c);
 		if (d40d == NULL || !d40d->cyclic)
 			goto check_pending_tx;
 	}
 
-	if (!d40d->cyclic)
+	if (from_done)
 		dma_cookie_complete(&d40d->txd);
 
 	/*
@@ -1757,7 +1766,7 @@ static void dma_tasklet(struct tasklet_struct *t)
 	callback_active = !!(d40d->txd.flags & DMA_PREP_INTERRUPT);
 	dmaengine_desc_get_callback(&d40d->txd, &cb);
 
-	if (!d40d->cyclic) {
+	if (from_done) {
 		if (async_tx_test_ack(&d40d->txd)) {
 			d40_desc_remove(d40d);
 			d40_desc_free(d40c, d40d);

-- 
2.55.0


  parent reply	other threads:[~2026-09-27  9:03 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-27  9:03 [PATCH v7 00/24] dmaengine: ste_dma40: Fix numerous accumulated bugs Linus Walleij
2026-09-27  9:03 ` [PATCH v7 01/24] dmaengine: ste_dma40: Fix physical cyclic capability Linus Walleij
2026-09-27  9:03 ` [PATCH v7 02/24] dmaengine: ste_dma40: Fix cyclic transfer residue Linus Walleij
2026-09-27  9:03 ` [PATCH v7 03/24] dmaengine: ste_dma40: Recover coalesced cyclic callbacks Linus Walleij
2026-09-28 16:48   ` Frank Li
2026-09-27  9:03 ` Linus Walleij [this message]
2026-09-27  9:03 ` [PATCH v7 05/24] dmaengine: ste_dma40: Fix probe runtime PM disable Linus Walleij
2026-09-27  9:03 ` [PATCH v7 06/24] dmaengine: ste_dma40: Check runtime PM in IRQ Linus Walleij
2026-09-27  9:24   ` sashiko-bot
2026-09-27 17:48     ` Linus Walleij
2026-09-27  9:03 ` [PATCH v7 07/24] dmaengine: ste_dma40: Handle runtime PM resume errors Linus Walleij
2026-09-27  9:03 ` [PATCH v7 08/24] dmaengine: ste_dma40: Return IRQ_NONE when no interrupt is pending Linus Walleij
2026-09-27  9:03 ` [PATCH v7 09/24] dmaengine: ste_dma40: Init hardware before registration Linus Walleij
2026-09-28 16:49   ` Frank Li
2026-09-27  9:03 ` [PATCH v7 10/24] dmaengine: ste_dma40: Fix probe IRQ leak Linus Walleij
2026-09-28 16:50   ` Frank Li
2026-09-27  9:03 ` [PATCH v7 11/24] dmaengine: ste_dma40: Fix DMA registration unwind Linus Walleij
2026-09-28 16:57   ` Frank Li
2026-09-27  9:03 ` [PATCH v7 12/24] dmaengine: ste_dma40: Fix LCLA allocation order Linus Walleij
2026-09-27  9:03 ` [PATCH v7 13/24] dmaengine: ste_dma40: Fix probe LCLA free Linus Walleij
2026-09-27  9:03 ` [PATCH v7 14/24] dmaengine: ste_dma40: Put the LCPA SRAM node Linus Walleij
2026-09-27  9:03 ` [PATCH v7 15/24] dmaengine: ste_dma40: Fix memcpy channel parsing Linus Walleij
2026-09-27  9:03 ` [PATCH v7 16/24] dmaengine: ste_dma40: Validate disabled channel indexes Linus Walleij
2026-09-27  9:03 ` [PATCH v7 17/24] dmaengine: ste_dma40: Validate DMA specifier length Linus Walleij
2026-09-27  9:03 ` [PATCH v7 18/24] dmaengine: ste_dma40: Reject direction changes after allocation Linus Walleij
2026-09-27  9:03 ` [PATCH v7 19/24] dmaengine: ste_dma40: Fix logical channel bounds check Linus Walleij
2026-09-28 16:58   ` Frank Li
2026-09-27  9:03 ` [PATCH v7 20/24] dmaengine: ste_dma40: Fix event group bounds Linus Walleij
2026-09-28 17:00   ` Frank Li
2026-09-27  9:03 ` [PATCH v7 21/24] dmaengine: ste_dma40: Fix V4B event group mapping Linus Walleij
2026-09-30 19:36   ` Frank Li
2026-09-27  9:03 ` [PATCH v7 22/24] dmaengine: ste_dma40: Search all blocks for fixed logical channels Linus Walleij
2026-10-05 15:39   ` Vinod Koul
2026-10-05 22:26     ` Linus Walleij
2026-09-27  9:03 ` [PATCH v7 23/24] dmaengine: ste_dma40: Validate fixed physical channel indexes Linus Walleij
2026-09-27  9:03 ` [PATCH v7 24/24] dmaengine: ste_dma40: Validate memcpy configuration Linus Walleij
2026-10-05 15:43 ` [PATCH v7 00/24] dmaengine: ste_dma40: Fix numerous accumulated bugs Vinod Koul

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=20260927-dma40-fixes-v7-4-89f595e8851d@kernel.org \
    --to=linusw@kernel.org \
    --cc=Frank.Li@kernel.org \
    --cc=Frank.Li@nxp.com \
    --cc=dmaengine@vger.kernel.org \
    --cc=phone-devel@vger.kernel.org \
    --cc=sashiko-bot@kernel.org \
    --cc=vkoul@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