* [PATCH] ASoC: atmel-i2s: Use devm_clk_get_enabled() for pclk
@ 2026-11-11 4:27 phucduc.bui
2026-11-11 4:35 ` sashiko-bot
0 siblings, 1 reply; 3+ messages in thread
From: phucduc.bui @ 2026-11-11 4:27 UTC (permalink / raw)
To: Mark Brown, Claudiu Beznea, Andrei Simion, Nicolas Ferre,
Alexandre Belloni
Cc: Liam Girdwood, Jaroslav Kysela, Takashi Iwai, Codrin Ciubotariu,
linux-sound, linux-arm-kernel, linux-kernel, bui duc phuc
From: bui duc phuc <phucduc.bui@gmail.com>
The peripheral clock is only acquired and enabled once during
probe. Use devm_clk_get_enabled() to combine clock acquisition
and enabling into a single call.
This also removes the need for manual clock disable handling
on probe failures and in the remove callback, as the devm API
automatically handles clock cleanup.
Signed-off-by: bui duc phuc <phucduc.bui@gmail.com>
---
sound/soc/atmel/atmel-i2s.c | 25 ++++---------------------
1 file changed, 4 insertions(+), 21 deletions(-)
diff --git a/sound/soc/atmel/atmel-i2s.c b/sound/soc/atmel/atmel-i2s.c
index e1d8af056234..a3aeb8e607f6 100644
--- a/sound/soc/atmel/atmel-i2s.c
+++ b/sound/soc/atmel/atmel-i2s.c
@@ -660,8 +660,8 @@ static int atmel_i2s_probe(struct platform_device *pdev)
if (err)
return err;
- /* Get the peripheral clock. */
- dev->pclk = devm_clk_get(&pdev->dev, "pclk");
+ /* Get and enable the peripheral clock. */
+ dev->pclk = devm_clk_get_enabled(&pdev->dev, "pclk");
if (IS_ERR(dev->pclk))
return dev_err_probe(&pdev->dev, PTR_ERR(dev->pclk),
"failed to get the peripheral clock\n");
@@ -682,11 +682,6 @@ static int atmel_i2s_probe(struct platform_device *pdev)
return err;
}
- /* Enable the peripheral clock. */
- err = clk_prepare_enable(dev->pclk);
- if (err)
- return err;
-
/* Get IP version. */
regmap_read(dev->regmap, ATMEL_I2SC_VERSION, &version);
dev_info(&pdev->dev, "hw version: %#x\n", version);
@@ -698,10 +693,8 @@ static int atmel_i2s_probe(struct platform_device *pdev)
err = devm_snd_soc_register_component(&pdev->dev,
&atmel_i2s_component,
&atmel_i2s_dai, 1);
- if (err) {
- clk_disable_unprepare(dev->pclk);
+ if (err)
return err;
- }
/* Prepare DMA config. */
dev->playback.addr = (dma_addr_t)mem->start + ATMEL_I2SC_THR;
@@ -712,28 +705,18 @@ static int atmel_i2s_probe(struct platform_device *pdev)
if (of_property_match_string(np, "dma-names", "rx-tx") == 0)
pcm_flags |= SND_DMAENGINE_PCM_FLAG_HALF_DUPLEX;
err = devm_snd_dmaengine_pcm_register(&pdev->dev, NULL, pcm_flags);
- if (err) {
- clk_disable_unprepare(dev->pclk);
+ if (err)
return err;
- }
return 0;
}
-static void atmel_i2s_remove(struct platform_device *pdev)
-{
- struct atmel_i2s_dev *dev = platform_get_drvdata(pdev);
-
- clk_disable_unprepare(dev->pclk);
-}
-
static struct platform_driver atmel_i2s_driver = {
.driver = {
.name = "atmel_i2s",
.of_match_table = atmel_i2s_dt_ids,
},
.probe = atmel_i2s_probe,
- .remove = atmel_i2s_remove,
};
module_platform_driver(atmel_i2s_driver);
--
2.43.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] ASoC: atmel-i2s: Use devm_clk_get_enabled() for pclk
2026-11-11 4:27 [PATCH] ASoC: atmel-i2s: Use devm_clk_get_enabled() for pclk phucduc.bui
@ 2026-11-11 4:35 ` sashiko-bot
2026-11-11 4:43 ` Bui Duc Phuc
0 siblings, 1 reply; 3+ messages in thread
From: sashiko-bot @ 2026-11-11 4:35 UTC (permalink / raw)
To: phucduc.bui; +Cc: linux-sound
Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] ASoC: atmel-i2s: wrong IRQ and pclk ordering in
atmel_i2s_probe()
--
commit 722d5e2f3add0fa840d1f0e68c8baa8331615c85
Author: bui duc phuc <phucduc.bui@gmail.com>
ASoC: atmel-i2s: Use devm_clk_get_enabled() for pclk
This commit replaces devm_clk_get() and clk_prepare_enable() for the
peripheral clock with devm_clk_get_enabled(), removing the manual clock
cleanup paths and the driver remove callback.
> diff --git a/sound/soc/atmel/atmel-i2s.c b/sound/soc/atmel/atmel-i2s.c
> index e1d8af0562341..a3aeb8e607f61 100644
> --- a/sound/soc/atmel/atmel-i2s.c
> +++ b/sound/soc/atmel/atmel-i2s.c
> @@ -660,8 +660,8 @@ static int atmel_i2s_probe(struct platform_device *pdev)
> if (err)
> return err;
>
> - /* Get the peripheral clock. */
> - dev->pclk = devm_clk_get(&pdev->dev, "pclk");
> + /* Get and enable the peripheral clock. */
> + dev->pclk = devm_clk_get_enabled(&pdev->dev, "pclk");
> if (IS_ERR(dev->pclk))
> return dev_err_probe(&pdev->dev, PTR_ERR(dev->pclk),
> "failed to get the peripheral clock\n");
[Severity: Medium]
Can enabling dev->pclk here cause atmel_i2s_interrupt() to dereference a
NULL dev->regmap or access registers while dev->pclk is disabled during
teardown?
atmel_i2s_probe() registers atmel_i2s_interrupt() via devm_request_irq()
just before this call, while dev->regmap and dev->dev are only assigned
after devm_clk_get_optional():
sound/soc/atmel/atmel-i2s.c:atmel_i2s_probe() {
...
err = devm_request_irq(&pdev->dev, irq, atmel_i2s_interrupt, 0,
dev_name(&pdev->dev), dev);
if (err)
return err;
/* Get and enable the peripheral clock. */
dev->pclk = devm_clk_get_enabled(&pdev->dev, "pclk");
...
dev->gclk = devm_clk_get_optional(&pdev->dev, "gclk");
if (IS_ERR(dev->gclk))
return PTR_ERR(dev->gclk);
dev->dev = &pdev->dev;
dev->regmap = regmap;
...
}
Previously, clk_prepare_enable(dev->pclk) was called after dev->regmap
and dev->dev were initialized. With devm_clk_get_enabled(), the
peripheral clock is enabled while dev->regmap is still NULL, so a pending
interrupt (for example across a driver rebind, as ATMEL_I2SC_IER is never
masked on unbind) will hit a NULL pointer dereference in
regmap_read(dev->regmap, ATMEL_I2SC_SR, &sr).
In addition, because devm_clk_get_enabled() is registered after
devm_request_irq(), devres teardown on probe failure or driver unbind now
disables dev->pclk before freeing the IRQ, leaving a window where
atmel_i2s_interrupt() can read I2S registers while the peripheral clock
is gated. Initializing dev->dev and dev->regmap and calling
devm_clk_get_enabled() before devm_request_irq() would avoid both issues.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261111042746.14714-1-phucduc.bui@gmail.com?part=1
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] ASoC: atmel-i2s: Use devm_clk_get_enabled() for pclk
2026-11-11 4:35 ` sashiko-bot
@ 2026-11-11 4:43 ` Bui Duc Phuc
0 siblings, 0 replies; 3+ messages in thread
From: Bui Duc Phuc @ 2026-11-11 4:43 UTC (permalink / raw)
To: sashiko-reviews; +Cc: linux-sound
> [Severity: Medium]
> Can enabling dev->pclk here cause atmel_i2s_interrupt() to dereference a
> NULL dev->regmap or access registers while dev->pclk is disabled during
> teardown?
>
This scenario is unlikely, but it is possible.
To be safe, I'll move the devm_request_irq() call below the clock
initialization
and dev->regmap setup.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-11-11 4:43 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-11 4:27 [PATCH] ASoC: atmel-i2s: Use devm_clk_get_enabled() for pclk phucduc.bui
2026-11-11 4:35 ` sashiko-bot
2026-11-11 4:43 ` Bui Duc Phuc
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox