mmc: pxamci: Simplify pxamci_probe() error handling using devm APIs

This patch refactors pxamci_probe() to use devm-managed resource allocation (e.g. devm_dma_request_chan()) and dev_err_probe() for improved readability and automatic cleanup on probe failure.

This eliminates redundant NULL assignments and manual release logic.

This issue was originally reported by Smatch:
drivers/mmc/host/pxamci.c:709 pxamci_probe() warn: passing zero to 'PTR_ERR'

The warning occurred because a pointer was set to NULL before using PTR_ERR(), leading to PTR_ERR(0) and an incorrect 0 return value. This refactor eliminates that condition while improving overall error handling robustness.

Reported-by kernel test robot <[email protected]>

Reported-by Dan Carpenter <[email protected]>

Closes https://lore.kernel.org/r/[email protected]/ open here

Fixes 58c40f3faf742c ("mmc: pxamci: Use devm_mmc_alloc_host() helper")

Suggested-by Uwe Kleine-König <[email protected]>

Signed-off-by Rakuram Eswaran <[email protected]>

---

Changes since v1:
Following Uwe Kleine-König’s suggestion:
- Replaced dma_request_chan() with devm_dma_request_chan() to make DMA channel allocation devm-managed and avoid manual release paths.
- Used dev_err_probe() for improved error reporting and consistent probe failure handling.
- Removed redundant NULL assignments and obsolete goto-based cleanup logic.
- Updated commit message to better describe the intent of the change.

Testing note:
I do not have access to appropriate hardware for runtime testing. Any help verifying on actual hardware would be appreciated.

Build and Analysis:
This patch was compiled against the configuration file reported by 0day CI in the above link (config: s390-randconfig-r071-20251004) using s390x-linux-gnu-gcc (Ubuntu 14.2.0-19ubuntu2) 14.2.0.

Static analysis was performed with Smatch to ensure the reported warning no longer reproduces after applying this fix.

Command used for verification:

   ARCH=s390 CROSS_COMPILE=s390x-linux-gnu- \
   ~/project/smatch/smatch_scripts/kchecker ./drivers/mmc/host/pxamci.c

  drivers/mmc/host/pxamci.c | 57 +++++++++++++++------------------------
  1 file changed, 21 insertions(+), 36 deletions(-)

drivers/mmc/host/pxamci.c +21 -29

index 26d03352af63..d03388f64934 100644
@@ -652,11 +652,14 @@ static int pxamci_probe(struct platform_device *pdev)
  	host->clkrt = CLKRT_OFF;
 
  	host->clk = devm_clk_get(dev, NULL);
-	if (IS_ERR(host->clk)) {
-		host->clk = NULL;
-		return PTR_ERR(host->clk);
-	}
+	if (IS_ERR(host->clk))
+		return dev_err_probe(dev, PTR_ERR(host->clk),
+				     "Failed to acquire clock\n");
 
+	/*
+	 * Note that the return value of clk_get_rate() is only valid
+	 * if the clock is enabled.
+	 */
  	host->clkrate = clk_get_rate(host->clk);
 
  	/*

@@ -703,46 +706,36 @@ static int pxamci_probe(struct platform_device *pdev)
 
  	platform_set_drvdata(pdev, mmc);
 
-	host->dma_chan_rx = dma_request_chan(dev, "rx");
-	if (IS_ERR(host->dma_chan_rx)) {
-		host->dma_chan_rx = NULL;
+	host->dma_chan_rx = devm_dma_request_chan(dev, "rx");
+	if (IS_ERR(host->dma_chan_rx))
  		return dev_err_probe(dev, PTR_ERR(host->dma_chan_rx),
  				     "unable to request rx dma channel\n");
-	}
 
-	host->dma_chan_tx = dma_request_chan(dev, "tx");
-	if (IS_ERR(host->dma_chan_tx)) {
-		dev_err(dev, "unable to request tx dma channel\n");
-		ret = PTR_ERR(host->dma_chan_tx);
-		host->dma_chan_tx = NULL;
-		goto out;
-	}
+	host->dma_chan_tx = devm_dma_request_chan(dev, "tx");
+	if (IS_ERR(host->dma_chan_tx))
+		return dev_err_probe(dev, PTR_ERR(host->dma_chan_tx),
+				     "unable to request tx dma channel\n");
 
  	if (host->pdata) {
  		host->detect_delay_ms = host->pdata->detect_delay_ms;
 
  		host->power = devm_gpiod_get_optional(dev, "power", GPIOD_OUT_LOW);
-		if (IS_ERR(host->power)) {
-			ret = PTR_ERR(host->power);
-			dev_err(dev, "Failed requesting gpio_power\n");
-			goto out;
-		}
+		if (IS_ERR(host->power))
+			return dev_err_probe(dev, PTR_ERR(host->power),
+						"Failed requesting gpio_power\n");
 
  		/* FIXME: should we pass detection delay to debounce? */
  		ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0);
-		if (ret && ret != -ENOENT) {
-			dev_err(dev, "Failed requesting gpio_cd\n");
-			goto out;
-		}
+		if (ret && ret != -ENOENT)
+			return dev_err_probe(dev, ret, "Failed requesting gpio_cd\n");
 
  		if (!host->pdata->gpio_card_ro_invert)
  			mmc->caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
 
  		ret = mmc_gpiod_request_ro(mmc, "wp", 0, 0);
-		if (ret && ret != -ENOENT) {
-			dev_err(dev, "Failed requesting gpio_ro\n");
-			goto out;
-		}
+		if (ret && ret != -ENOENT)
+			return dev_err_probe(dev, ret, "Failed requesting gpio_ro\n");
+
  		if (!ret)
  			host->use_ro_gpio = true;
 

@@ -759,16 +752,8 @@ static int pxamci_probe(struct platform_device *pdev)
  	if (ret) {
  		if (host->pdata && host->pdata->exit)
  			host->pdata->exit(dev, mmc);
-		goto out;
  	}
 

This patch refactors pxamci_probe() to use devm-managed resource allocation (e.g. devm_dma_request_chan()) and dev_err_probe() for improved readability and automatic cleanup on probe failure.

This eliminates redundant NULL assignments and manual release logic.

This issue was originally reported by Smatch:
drivers/mmc/host/pxamci.c:709 pxamci_probe() warn: passing zero to 'PTR_ERR'

The warning occurred because a pointer was set to NULL before using PTR_ERR(), leading to PTR_ERR(0) and an incorrect 0 return value. This refactor eliminates that condition while improving overall error handling robustness.

Reported-by kernel test robot <[email protected]>

Reported-by Dan Carpenter <[email protected]>

Closes https://lore.kernel.org/r/[email protected]/ open here

Fixes 58c40f3faf742c ("mmc: pxamci: Use devm_mmc_alloc_host() helper")

Suggested-by Uwe Kleine-König <[email protected]>

Signed-off-by Rakuram Eswaran <[email protected]>

---

Changes since v1:
Following Uwe Kleine-König’s suggestion:
- Replaced dma_request_chan() with devm_dma_request_chan() to make DMA channel allocation devm-managed and avoid manual release paths.
- Used dev_err_probe() for improved error reporting and consistent probe failure handling.
- Removed redundant NULL assignments and obsolete goto-based cleanup logic.
- Updated commit message to better describe the intent of the change.

Testing note:
I do not have access to appropriate hardware for runtime testing. Any help verifying on actual hardware would be appreciated.

Build and Analysis:
This patch was compiled against the configuration file reported by 0day CI in the above link (config: s390-randconfig-r071-20251004) using s390x-linux-gnu-gcc (Ubuntu 14.2.0-19ubuntu2) 14.2.0.

Static analysis was performed with Smatch to ensure the reported warning no longer reproduces after applying this fix.

Command used for verification:
ARCH=s390 CROSS_COMPILE=s390x-linux-gnu- \
~/project/smatch/smatch_scripts/kchecker ./drivers/mmc/host/pxamci.c

 drivers/mmc/host/pxamci.c | 57 +++++++++++++++------------------------
 1 file changed, 21 insertions(+), 36 deletions(-)

drivers/mmc/host/pxamci.c +0 -2

index 26d03352af63..d03388f64934 100644
@@ -652,11 +652,14 @@ static int pxamci_probe(struct platform_device *pdev)
 	host->clkrt = CLKRT_OFF;
 
 	host->clk = devm_clk_get(dev, NULL);
-	if (IS_ERR(host->clk)) {
-		host->clk = NULL;

drivers/mmc/host/pxamci.c +0 -0

 	host->clkrate = clk_get_rate(host->clk);
 
 	/*
[...]

@@ -759,16 +752,8 @@ static int pxamci_probe(struct platform_device *pdev)
 	if (ret) {
 		if (host->pdata && host->pdata->exit)
 			host->pdata->exit(dev, mmc);

Quoted 66 lines · Rakuram Eswaran

This patch refactors pxamci_probe() to use devm-managed resource allocation (e.g. devm_dma_request_chan()) and dev_err_probe() for improved readability and automatic cleanup on probe failure.

This eliminates redundant NULL assignments and manual release logic.

This issue was originally reported by Smatch:
drivers/mmc/host/pxamci.c:709 pxamci_probe() warn: passing zero to 'PTR_ERR'

The warning occurred because a pointer was set to NULL before using PTR_ERR(), leading to PTR_ERR(0) and an incorrect 0 return value. This refactor eliminates that condition while improving overall error handling robustness.

Reported-by kernel test robot <[email protected]>

Reported-by Dan Carpenter <[email protected]>

Closes https://lore.kernel.org/r/[email protected]/ open here

Fixes 58c40f3faf742c ("mmc: pxamci: Use devm_mmc_alloc_host() helper")

Suggested-by Uwe Kleine-König <[email protected]>

Signed-off-by Rakuram Eswaran <[email protected]>

---

Changes since v1:
Following Uwe Kleine-König’s suggestion:
- Replaced dma_request_chan() with devm_dma_request_chan() to make DMA channel allocation devm-managed and avoid manual release paths.
- Used dev_err_probe() for improved error reporting and consistent probe failure handling.
- Removed redundant NULL assignments and obsolete goto-based cleanup logic.
- Updated commit message to better describe the intent of the change.

Testing note:
I do not have access to appropriate hardware for runtime testing. Any help verifying on actual hardware would be appreciated.

Build and Analysis:
This patch was compiled against the configuration file reported by 0day CI in the above link (config: s390-randconfig-r071-20251004) using s390x-linux-gnu-gcc (Ubuntu 14.2.0-19ubuntu2) 14.2.0.

Static analysis was performed with Smatch to ensure the reported warning no longer reproduces after applying this fix.

Command used for verification:
  ARCH=s390 CROSS_COMPILE=s390x-linux-gnu- \
  ~/project/smatch/smatch_scripts/kchecker ./drivers/mmc/host/pxamci.c

 drivers/mmc/host/pxamci.c | 57 +++++++++++++++------------------------

 1 file changed, 21 insertions(+), 36 deletions(-)

drivers/mmc/host/pxamci.c +3 -4

index 26d03352af63..d03388f64934 100644
@@ -652,11 +652,14 @@ static int pxamci_probe(struct platform_device *pdev)
       host->clkrt = CLKRT_OFF;
 
       host->clk = devm_clk_get(dev, NULL);
-     if (IS_ERR(host->clk)) {
-             host->clk = NULL;
-             return PTR_ERR(host->clk);
-     }
+     if (IS_ERR(host->clk))
+             return dev_err_probe(dev, PTR_ERR(host->clk),
+                                  "Failed to acquire clock\n");
 

drivers/mmc/host/pxamci.c static int pxamci_probe(struct platform_device *pdev) · line 662

Quoted 4 lines

drivers/mmc/host/pxamci.c +4 -0

+     /*
+      * Note that the return value of clk_get_rate() is only valid
+      * if the clock is enabled.
+      */

The intention of this comment in my WIP suggestion was to point out another thing to fix as the precondition for calling clk_get_rate() isn't asserted. If you don't want to address this (which is fine), drop the comment (or improve my wording to make it more obvious that there is something to fix).

drivers/mmc/host/pxamci.c static int pxamci_probe(struct platform_device *pdev) · line 771

Quoted 5 lines

drivers/mmc/host/pxamci.c +0 -5

-out:
-     if (host->dma_chan_rx)
-             dma_release_channel(host->dma_chan_rx);
-     if (host->dma_chan_tx)
-             dma_release_channel(host->dma_chan_tx);

I was lazy in my prototype patch and didn't drop the calls to dma_release_channel() in pxamci_remove(). For a proper patch this is required though.

To continue the quest: Now that I looked at pxamci_remove(): mmc is always non-NULL, so the respective check can be dropped.

Quoted 1 line · Uwe Kleine-König

On Wed, Oct 15, 2025 at 12:16:57AM +0530, Rakuram Eswaran wrote:

Sorry for the delayed reply as I was in vacation.

Ah, got it — I’ll drop the clk_get_rate() comment since it was only a reminder from your WIP suggestion.

Just to confirm, are you referring to adding a call to clk_prepare_enable() before clk_get_rate()? I can move the clk_get_rate() call after clk_prepare_enable(), or drop the comment entirely.

If my understanding is correct, I’ll keep v3 focused on the current set of fixes and handle the clk_get_rate() precondition (by moving it after clk_prepare_enable()) in a follow-up patch. That should keep the scope of each change clean and review-friendly.

Quoted 10 lines

drivers/mmc/host/pxamci.c static int pxamci_probe(struct platform_device *pdev) · line 771

Quoted 5 lines

drivers/mmc/host/pxamci.c +0 -5

-out:
-     if (host->dma_chan_rx)
-             dma_release_channel(host->dma_chan_rx);
-     if (host->dma_chan_tx)
-             dma_release_channel(host->dma_chan_tx);

I was lazy in my prototype patch and didn't drop the calls to dma_release_channel() in pxamci_remove(). For a proper patch this is required though.

Quoted 3 lines

To continue the quest: Now that I looked at pxamci_remove(): mmc is always non-NULL, so the respective check can be dropped.

Understood. Since pxamci_remove() is only called after successful allocation and initialization in probe(), mmc will always be a valid pointer. I’ll drop the if (mmc) check in v3 as it can never be NULL in normal operation, and remove the dma_release_channel() calls as well.

I’ve prepared a preview of the v3 patch incorporating your previous comments. Before sending it out formally, I wanted to share it with you to confirm that the updates look good — especially the cleanup changes in pxamci_remove() and the dropped clk_get_rate() comment.

static void pxamci_remove(struct platform_device *pdev)

{
	struct mmc_host *mmc = platform_get_drvdata(pdev);
	struct pxamci_host *host = mmc_priv(mmc);

	mmc_remove_host(mmc);

	if (host->pdata && host->pdata->exit)
		host->pdata->exit(&pdev->dev, mmc);

	pxamci_stop_clock(host);
	writel(TXFIFO_WR_REQ|RXFIFO_RD_REQ|CLK_IS_OFF|STOP_CMD|
			END_CMD_RES|PRG_DONE|DATA_TRAN_DONE,
			host->base + MMC_I_MASK);

	dmaengine_terminate_all(host->dma_chan_rx);
	dmaengine_terminate_all(host->dma_chan_tx);
}

Hello Rakuram,

Quoted 3 lines · Rakuram Eswaran

Quoted 1 line · Uwe Kleine-König

On Wed, Oct 15, 2025 at 12:16:57AM +0530, Rakuram Eswaran wrote:

Sorry for the delayed reply as I was in vacation.

I didn't hold my breath :-O

Quoted 26 lines

Ah, got it — I’ll drop the clk_get_rate() comment since it was only a reminder from your WIP suggestion.

Just to confirm, are you referring to adding a call to clk_prepare_enable() before clk_get_rate()? I can move the clk_get_rate() call after clk_prepare_enable(), or drop the comment entirely.

If my understanding is correct, I’ll keep v3 focused on the current set of fixes and handle the clk_get_rate() precondition (by moving it after clk_prepare_enable()) in a follow-up patch. That should keep the scope of each change clean and review-friendly.

Quoted 13 lines

Quoted 5 lines

-out:
-     if (host->dma_chan_rx)
-             dma_release_channel(host->dma_chan_rx);
-     if (host->dma_chan_tx)
-             dma_release_channel(host->dma_chan_tx);

I was lazy in my prototype patch and didn't drop the calls to dma_release_channel() in pxamci_remove(). For a proper patch this is required though.

To continue the quest: Now that I looked at pxamci_remove(): mmc is always non-NULL, so the respective check can be dropped.

Quoted 4 lines

Understood. Since pxamci_remove() is only called after successful allocation and initialization in probe(), mmc will always be a valid pointer. I’ll drop the if (mmc) check in v3 as it can never be NULL in normal operation, and remove the dma_release_channel() calls as well.

Yes, I suggest to make restructuring .remote a separate patch. (But removing dma_release_channel belongs into the patch that introduces devm to allocate the dma channels.)

Quoted 20 lines

I’ve prepared a preview of the v3 patch incorporating your previous comments. Before sending it out formally, I wanted to share it with you to confirm that the updates look good — especially the cleanup changes in pxamci_remove() and the dropped clk_get_rate() comment.

static void pxamci_remove(struct platform_device *pdev) {
      struct mmc_host *mmc = platform_get_drvdata(pdev); struct pxamci_host *host = mmc_priv(mmc);

      mmc_remove_host(mmc);

      if (host->pdata && host->pdata->exit)
              host->pdata->exit(&pdev->dev, mmc);

      pxamci_stop_clock(host); writel(TXFIFO_WR_REQ|RXFIFO_RD_REQ|CLK_IS_OFF|STOP_CMD| END_CMD_RES|PRG_DONE|DATA_TRAN_DONE, host->base + MMC_I_MASK);

Quoted 3 lines

      dmaengine_terminate_all(host->dma_chan_rx);
      dmaengine_terminate_all(host->dma_chan_tx);
}

Looks right.

Quoted 3 lines · Uwe Kleine-König

Yes, I suggest to make restructuring .remote a separate patch. (But removing dma_release_channel belongs into the patch that introduces devm to allocate the dma channels.)

I believe ".remote" is a typo and you're referring to the _remove() function. Removing if(mmc) condition check from pxamci_remove() can be handled in a separate cleanup patch, while removing redundant dma_release_channel() will be included in v3.

Is my above understanding correct?

Hello Rakuram,

Quoted 5 lines · Rakuram Eswaran

Quoted 3 lines · Uwe Kleine-König

Yes, I suggest to make restructuring .remote a separate patch. (But removing dma_release_channel belongs into the patch that introduces devm to allocate the dma channels.)

Quoted 6 lines

I believe ".remote" is a typo and you're referring to the _remove() function. Removing if(mmc) condition check from pxamci_remove() can be handled in a separate cleanup patch, while removing redundant dma_release_channel() will be included in v3.

Is my above understanding correct?

ack. remote vs. remove is one of my most-committed typos :-D