Skip to contentSkip to conversation

FAILED: patch "[PATCH] ASoC: mediatek: mt8192: Check runtime resume during probe" failed to apply to 5.15-stable tree

PATCH New [email protected] 2 participants 6 messages Jul 20 – Jul 23

Subsystems (4)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x e24d5dde56a50946020b134fa8448869093db76a
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026072022-fading-available-28ca@gregkh' --subject-prefix 'PATCH 5.15.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From e24d5dde56a50946020b134fa8448869093db76a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?C=C3=A1ssio=20Gabriel?= <[email protected]>
Date: Wed, 27 May 2026 10:55:47 -0300
Subject: [PATCH] ASoC: mediatek: mt8192: Check runtime resume during probe
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The MT8192 AFE probe enables runtime PM temporarily while reinitializing the regmap cache from hardware, but it uses pm_runtime_get_sync() without checking the return value. If runtime resume fails, probe keeps going without the device necessarily being accessible, and pm_runtime_get_sync() may leave the PM usage count incremented.

The regmap_reinit_cache() failure path also returns before dropping the temporary PM reference and before clearing pm_runtime_bypass_reg_ctl.

Use pm_runtime_resume_and_get() so resume failures do not leak a usage count, and clear the temporary bypass flag after dropping the probe PM reference on all regmap_reinit_cache() outcomes.

Fixes 125ab5d588b0 ("ASoC: mediatek: mt8192: add platform driver")
sound/soc/mediatek/mt8192/mt8192-afe-pcm.c +8 -4
index 9f5057eeeff9..db0ae44a86af 100644
@@ -2227,15 +2227,19 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 
 	/* enable clock for regcache get default value from hw */
 	afe_priv->pm_runtime_bypass_reg_ctl = true;
-	pm_runtime_get_sync(dev);
+	ret = pm_runtime_resume_and_get(dev);
+	if (ret) {
+		afe_priv->pm_runtime_bypass_reg_ctl = false;
+		return dev_err_probe(dev, ret, "failed to resume device\n");
+	}
 
 	ret = regmap_reinit_cache(afe->regmap, &mt8192_afe_regmap_config);
-	if (ret)
-		return dev_err_probe(dev, ret, "regmap_reinit_cache fail\n");
-
 	pm_runtime_put_sync(dev);
 	afe_priv->pm_runtime_bypass_reg_ctl = false;
 
+	if (ret)
+		return dev_err_probe(dev, ret, "regmap_reinit_cache fail\n");
+
 	regcache_cache_only(afe->regmap, true);
 	regcache_mark_dirty(afe->regmap);
 
maintainer
#1 lore raw

PATCH 1/5 ASoC: mediatek: mt8192-afe-pcm: Convert to devm_pm_runtime_enable()

From: AngeloGioacchino Del Regno <[email protected]>

[ Upstream commit 7aaaa22de56ce0dae15fd9f42a69a1d1a7a6e078 ]

Switch from pm_runtime_enable() to devm_pm_runtime_enable(), allowing to remove all gotos from the probe function.

Stable-dep-of: e24d5dde56a5 ("ASoC: mediatek: mt8192: Check runtime resume during probe")

---
 sound/soc/mediatek/mt8192/mt8192-afe-pcm.c | 46 ++++++++--------------
 1 file changed, 17 insertions(+), 29 deletions(-)
sound/soc/mediatek/mt8192/mt8192-afe-pcm.c +17 -29
index 31c280339c503b..7ed36a760c87a2 100644
@@ -2221,21 +2221,20 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 		return ret;
 	}
 
-	pm_runtime_enable(&pdev->dev);
-	if (!pm_runtime_enabled(&pdev->dev))
-		goto err_pm_disable;
+	ret = devm_pm_runtime_enable(&pdev->dev);
+	if (ret)
+		return ret;
 
 	/* regmap init */
 	afe->regmap = syscon_node_to_regmap(dev->parent->of_node);
 	if (IS_ERR(afe->regmap)) {
 		dev_err(dev, "could not get regmap from parent\n");
-		ret = PTR_ERR(afe->regmap);
-		goto err_pm_disable;
+		return PTR_ERR(afe->regmap);
 	}
 	ret = regmap_attach_dev(dev, afe->regmap, &mt8192_afe_regmap_config);
 	if (ret) {
 		dev_warn(dev, "regmap_attach_dev fail, ret %d\n", ret);
-		goto err_pm_disable;
+		return ret;
 	}
 
 	/* enable clock for regcache get default value from hw */
@@ -2245,7 +2244,7 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	ret = regmap_reinit_cache(afe->regmap, &mt8192_afe_regmap_config);
 	if (ret) {
 		dev_err(dev, "regmap_reinit_cache fail, ret %d\n", ret);
-		goto err_pm_disable;
+		return ret;
 	}
 
 	pm_runtime_put_sync(&pdev->dev);
@@ -2258,10 +2257,8 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	afe->memif_size = MT8192_MEMIF_NUM;
 	afe->memif = devm_kcalloc(dev, afe->memif_size, sizeof(*afe->memif),
 				  GFP_KERNEL);
-	if (!afe->memif) {
-		ret = -ENOMEM;
-		goto err_pm_disable;
-	}
+	if (!afe->memif)
+		return -ENOMEM;
 
 	for (i = 0; i < afe->memif_size; i++) {
 		afe->memif[i].data = &memif_data[i];
@@ -2275,26 +2272,22 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	afe->irqs_size = MT8192_IRQ_NUM;
 	afe->irqs = devm_kcalloc(dev, afe->irqs_size, sizeof(*afe->irqs),
 				 GFP_KERNEL);
-	if (!afe->irqs) {
-		ret = -ENOMEM;
-		goto err_pm_disable;
-	}
+	if (!afe->irqs)
+		return -ENOMEM;
 
 	for (i = 0; i < afe->irqs_size; i++)
 		afe->irqs[i].irq_data = &irq_data[i];
 
 	/* request irq */
 	irq_id = platform_get_irq(pdev, 0);
-	if (irq_id < 0) {
-		ret = irq_id;
-		goto err_pm_disable;
-	}
+	if (irq_id < 0)
+		return irq_id;
 
 	ret = devm_request_irq(dev, irq_id, mt8192_afe_irq_handler,
 			       IRQF_TRIGGER_NONE, "asys-isr", (void *)afe);
 	if (ret) {
 		dev_err(dev, "could not request_irq for Afe_ISR_Handle\n");
-		goto err_pm_disable;
+		return ret;
 	}
 
 	/* init sub_dais */
@@ -2305,7 +2298,7 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 		if (ret) {
 			dev_warn(afe->dev, "dai register i %d fail, ret %d\n",
 				 i, ret);
-			goto err_pm_disable;
+			return ret;
 		}
 	}
 
@@ -2314,7 +2307,7 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	if (ret) {
 		dev_warn(afe->dev, "mtk_afe_combine_sub_dai fail, ret %d\n",
 			 ret);
-		goto err_pm_disable;
+		return ret;
 	}
 
 	/* others */
@@ -2333,7 +2326,7 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 					      &mt8192_afe_component, NULL, 0);
 	if (ret) {
 		dev_warn(dev, "err_platform\n");
-		goto err_pm_disable;
+		return ret;
 	}
 
 	ret = devm_snd_soc_register_component(&pdev->dev,
@@ -2342,15 +2335,10 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 					      afe->num_dai_drivers);
 	if (ret) {
 		dev_warn(dev, "err_dai_component\n");
-		goto err_pm_disable;
+		return ret;
 	}
 
 	return 0;
-
-err_pm_disable:
-	pm_runtime_disable(&pdev->dev);
-
-	return ret;
 }
 
 static int mt8192_afe_pcm_dev_remove(struct platform_device *pdev)
Signature · 1 line
2.53.0
maintainer
#2 lore raw

PATCH 2/5 ASoC: mediatek: mt8192-afe-pcm: Simplify with dev_err_probe()

From: AngeloGioacchino Del Regno <[email protected]>

[ Upstream commit 324c603a4efca7d1045e0bf3477ca54970eac72c ]

Simplify the probe function by switching error prints to return dev_err_probe(), lowering the lines count; while at it, also beautify some messages and change some others' level from warn to error.

Stable-dep-of: e24d5dde56a5 ("ASoC: mediatek: mt8192: Check runtime resume during probe")

---
 sound/soc/mediatek/mt8192/mt8192-afe-pcm.c | 65 ++++++++--------------
 1 file changed, 22 insertions(+), 43 deletions(-)
sound/soc/mediatek/mt8192/mt8192-afe-pcm.c +22 -43
index 7ed36a760c87a2..75ceecad408078 100644
@@ -2209,17 +2209,12 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 
 	/* reset controller to reset audio regs before regmap cache */
 	rstc = devm_reset_control_get_exclusive(dev, "audiosys");
-	if (IS_ERR(rstc)) {
-		ret = PTR_ERR(rstc);
-		dev_err(dev, "could not get audiosys reset:%d\n", ret);
-		return ret;
-	}
+	if (IS_ERR(rstc))
+		return dev_err_probe(dev, PTR_ERR(rstc), "could not get audiosys reset\n");
 
 	ret = reset_control_reset(rstc);
-	if (ret) {
-		dev_err(dev, "failed to trigger audio reset:%d\n", ret);
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(dev, ret, "failed to trigger audio reset\n");
 
 	ret = devm_pm_runtime_enable(&pdev->dev);
 	if (ret)
@@ -2227,25 +2222,21 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 
 	/* regmap init */
 	afe->regmap = syscon_node_to_regmap(dev->parent->of_node);
-	if (IS_ERR(afe->regmap)) {
-		dev_err(dev, "could not get regmap from parent\n");
-		return PTR_ERR(afe->regmap);
-	}
+	if (IS_ERR(afe->regmap))
+		return dev_err_probe(dev, PTR_ERR(afe->regmap),
+				     "could not get regmap from parent");
+
 	ret = regmap_attach_dev(dev, afe->regmap, &mt8192_afe_regmap_config);
-	if (ret) {
-		dev_warn(dev, "regmap_attach_dev fail, ret %d\n", ret);
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(dev, ret, "regmap_attach_dev fail\n");
 
 	/* enable clock for regcache get default value from hw */
 	afe_priv->pm_runtime_bypass_reg_ctl = true;
 	pm_runtime_get_sync(&pdev->dev);
 
 	ret = regmap_reinit_cache(afe->regmap, &mt8192_afe_regmap_config);
-	if (ret) {
-		dev_err(dev, "regmap_reinit_cache fail, ret %d\n", ret);
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(dev, ret, "regmap_reinit_cache fail\n");
 
 	pm_runtime_put_sync(&pdev->dev);
 	afe_priv->pm_runtime_bypass_reg_ctl = false;
@@ -2285,30 +2276,22 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 
 	ret = devm_request_irq(dev, irq_id, mt8192_afe_irq_handler,
 			       IRQF_TRIGGER_NONE, "asys-isr", (void *)afe);
-	if (ret) {
-		dev_err(dev, "could not request_irq for Afe_ISR_Handle\n");
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(dev, ret, "could not request_irq for Afe_ISR_Handle\n");
 
 	/* init sub_dais */
 	INIT_LIST_HEAD(&afe->sub_dais);
 
 	for (i = 0; i < ARRAY_SIZE(dai_register_cbs); i++) {
 		ret = dai_register_cbs[i](afe);
-		if (ret) {
-			dev_warn(afe->dev, "dai register i %d fail, ret %d\n",
-				 i, ret);
-			return ret;
-		}
+		if (ret)
+			return dev_err_probe(afe->dev, ret, "dai %d register fail", i);
 	}
 
 	/* init dai_driver and component_driver */
 	ret = mtk_afe_combine_sub_dai(afe);
-	if (ret) {
-		dev_warn(afe->dev, "mtk_afe_combine_sub_dai fail, ret %d\n",
-			 ret);
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(afe->dev, ret, "mtk_afe_combine_sub_dai fail\n");
 
 	/* others */
 	afe->mtk_afe_hardware = &mt8192_afe_hardware;
@@ -2324,19 +2307,15 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	/* register platform */
 	ret = devm_snd_soc_register_component(&pdev->dev,
 					      &mt8192_afe_component, NULL, 0);
-	if (ret) {
-		dev_warn(dev, "err_platform\n");
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(dev, ret, "Couldn't register AFE component\n");
 
 	ret = devm_snd_soc_register_component(&pdev->dev,
 					      &mt8192_afe_pcm_component,
 					      afe->dai_drivers,
 					      afe->num_dai_drivers);
-	if (ret) {
-		dev_warn(dev, "err_dai_component\n");
-		return ret;
-	}
+	if (ret)
+		return dev_err_probe(dev, ret, "Couldn't register AFE-PCM component\n");
 
 	return 0;
 }
Signature · 1 line
2.53.0
maintainer
#3 lore raw

PATCH 3/5 ASoC: mediatek: Use common mtk_afe_pcm_platform with common probe cb

In reply to Sasha Levin

From: AngeloGioacchino Del Regno <[email protected]>

[ Upstream commit 1d2a6b3b5189b6c3e51f29e050aa9b2337b774e8 ]

Since the mtk-afe-platform-driver generic mtk_afe_pcm_platform now has a common .probe() callback, there is no reason to keep duplicating this function over and over in the SoC specific AFE-PCM drivers: switch over to register with the common bits instead.

Note that MT8186 was left out of this because it is registering some extra sinegen controls in the AFE-PCM probe callback and needs extra cleanups to be able to use the common bits.

Stable-dep-of: e24d5dde56a5 ("ASoC: mediatek: mt8192: Check runtime resume during probe")

---
 sound/soc/mediatek/mt8192/mt8192-afe-pcm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
sound/soc/mediatek/mt8192/mt8192-afe-pcm.c +2 -2
index 75ceecad408078..7562bea5b34c05 100644
@@ -2232,13 +2232,13 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 
 	/* enable clock for regcache get default value from hw */
 	afe_priv->pm_runtime_bypass_reg_ctl = true;
-	pm_runtime_get_sync(&pdev->dev);
+	pm_runtime_get_sync(dev);
 
 	ret = regmap_reinit_cache(afe->regmap, &mt8192_afe_regmap_config);
 	if (ret)
 		return dev_err_probe(dev, ret, "regmap_reinit_cache fail\n");
 
-	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_put_sync(dev);
 	afe_priv->pm_runtime_bypass_reg_ctl = false;
 
 	regcache_cache_only(afe->regmap, true);
Signature · 1 line
2.53.0
maintainer
#4 lore raw

PATCH 4/5 ASoC: mediatek: mt8192-afe-pcm: Simplify probe() with local dev variable

In reply to Sasha Levin

From: Tang Bin <[email protected]>

[ Upstream commit 01981565c764c554cc96e2d30a71c42975171416 ]

Simplify the function mt8192_afe_pcm_dev_probe() by using local 'dev' instead of '&pdev->dev'.

Stable-dep-of: e24d5dde56a5 ("ASoC: mediatek: mt8192: Check runtime resume during probe")

---
 sound/soc/mediatek/mt8192/mt8192-afe-pcm.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)
sound/soc/mediatek/mt8192/mt8192-afe-pcm.c +8 -9
index 7562bea5b34c05..0acf4fc8367d61 100644
@@ -2178,27 +2178,26 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 {
 	struct mtk_base_afe *afe;
 	struct mt8192_afe_private *afe_priv;
-	struct device *dev;
+	struct device *dev = &pdev->dev;
 	struct reset_control *rstc;
 	int i, ret, irq_id;
 
-	ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(34));
+	ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(34));
 	if (ret)
 		return ret;
 
-	afe = devm_kzalloc(&pdev->dev, sizeof(*afe), GFP_KERNEL);
+	afe = devm_kzalloc(dev, sizeof(*afe), GFP_KERNEL);
 	if (!afe)
 		return -ENOMEM;
 	platform_set_drvdata(pdev, afe);
 
-	afe->platform_priv = devm_kzalloc(&pdev->dev, sizeof(*afe_priv),
+	afe->platform_priv = devm_kzalloc(dev, sizeof(*afe_priv),
 					  GFP_KERNEL);
 	if (!afe->platform_priv)
 		return -ENOMEM;
 	afe_priv = afe->platform_priv;
 
-	afe->dev = &pdev->dev;
-	dev = afe->dev;
+	afe->dev = dev;
 
 	/* init audio related clock */
 	ret = mt8192_init_clock(afe);
@@ -2216,7 +2215,7 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	if (ret)
 		return dev_err_probe(dev, ret, "failed to trigger audio reset\n");
 
-	ret = devm_pm_runtime_enable(&pdev->dev);
+	ret = devm_pm_runtime_enable(dev);
 	if (ret)
 		return ret;
 
@@ -2305,12 +2304,12 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 	afe->runtime_suspend = mt8192_afe_runtime_suspend;
 
 	/* register platform */
-	ret = devm_snd_soc_register_component(&pdev->dev,
+	ret = devm_snd_soc_register_component(dev,
 					      &mt8192_afe_component, NULL, 0);
 	if (ret)
 		return dev_err_probe(dev, ret, "Couldn't register AFE component\n");
 
-	ret = devm_snd_soc_register_component(&pdev->dev,
+	ret = devm_snd_soc_register_component(dev,
 					      &mt8192_afe_pcm_component,
 					      afe->dai_drivers,
 					      afe->num_dai_drivers);
Signature · 1 line
2.53.0
maintainer
#5 lore raw

PATCH 5/5 ASoC: mediatek: mt8192: Check runtime resume during probe

In reply to Sasha Levin

From: Cássio Gabriel <[email protected]>

[ Upstream commit e24d5dde56a50946020b134fa8448869093db76a ]

The MT8192 AFE probe enables runtime PM temporarily while reinitializing the regmap cache from hardware, but it uses pm_runtime_get_sync() without checking the return value. If runtime resume fails, probe keeps going without the device necessarily being accessible, and pm_runtime_get_sync() may leave the PM usage count incremented.

The regmap_reinit_cache() failure path also returns before dropping the temporary PM reference and before clearing pm_runtime_bypass_reg_ctl.

Use pm_runtime_resume_and_get() so resume failures do not leak a usage count, and clear the temporary bypass flag after dropping the probe PM reference on all regmap_reinit_cache() outcomes.

Fixes 125ab5d588b0 ("ASoC: mediatek: mt8192: add platform driver")
---
 sound/soc/mediatek/mt8192/mt8192-afe-pcm.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)
sound/soc/mediatek/mt8192/mt8192-afe-pcm.c +8 -4
index 0acf4fc8367d61..765889f3fae0dc 100644
@@ -2231,15 +2231,19 @@ static int mt8192_afe_pcm_dev_probe(struct platform_device *pdev)
 
 	/* enable clock for regcache get default value from hw */
 	afe_priv->pm_runtime_bypass_reg_ctl = true;
-	pm_runtime_get_sync(dev);
+	ret = pm_runtime_resume_and_get(dev);
+	if (ret) {
+		afe_priv->pm_runtime_bypass_reg_ctl = false;
+		return dev_err_probe(dev, ret, "failed to resume device\n");
+	}
 
 	ret = regmap_reinit_cache(afe->regmap, &mt8192_afe_regmap_config);
-	if (ret)
-		return dev_err_probe(dev, ret, "regmap_reinit_cache fail\n");
-
 	pm_runtime_put_sync(dev);
 	afe_priv->pm_runtime_bypass_reg_ctl = false;
 
+	if (ret)
+		return dev_err_probe(dev, ret, "regmap_reinit_cache fail\n");
+
 	regcache_cache_only(afe->regmap, true);
 	regcache_mark_dirty(afe->regmap);
 
Signature · 1 line
2.53.0