Skip to contentSkip to conversation

spi: spi-fsl-dspi: Fix lockup if device is removed during SPI transfer

PATCH In review Krzysztof Kozlowski 3 participants 8 messages Jun 22, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v4 1/4 spi: spi-fsl-dspi: Fix lockup if device is removed during SPI transfer

During device removal, the driver should unregister the SPI controller and stop the hardware. Otherwise the dspi_transfer_one_message() could wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device removal reverse-matches the probe function, where SPI controller is registered at the end.

Fixes 05209f457069 ("spi: fsl-dspi: add missing clk_disable_unprepare() in dspi_remove()")

---

Changes since v3:
1. New patch.

---
 drivers/spi/spi-fsl-dspi.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/spi/spi-fsl-dspi.c +10 -1
index 58190c94561f..ec0fd0d366eb 100644
@@ -1434,9 +1434,18 @@ static int dspi_remove(struct platform_device *pdev)
 	struct fsl_dspi *dspi = spi_controller_get_devdata(ctlr);
 
 	/* Disconnect from the SPI framework */
+	spi_unregister_controller(dspi->ctlr);
+
+	/* Disable RX and TX */
+	regmap_update_bits(dspi->regmap, SPI_MCR,
+			   SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF,
+			   SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF);
+
+	/* Stop Running */
+	regmap_update_bits(dspi->regmap, SPI_MCR, SPI_MCR_HALT, SPI_MCR_HALT);
+
 	dspi_release_dma(dspi);
 	clk_disable_unprepare(dspi->clk);
-	spi_unregister_controller(dspi->ctlr);
 
 	return 0;
 }
Signature · 1 line
2.17.1
#3 lore raw

PATCH v4 3/4 spi: spi-fsl-dspi: Fix external abort on interrupt in resume or exit paths

If shared interrupt comes late, during probe error path or device remove (could be triggered with CONFIG_DEBUG_SHIRQ), the interrupt handler dspi_interrupt() will access registers with the clock being disabled. This leads to external abort on non-linefetch on Toradex Colibri VF50 module (with Vybrid VF5xx):

    $ echo 4002d000.spi > /sys/devices/platform/soc/40000000.bus/4002d000.spi/driver/unbind
    Unhandled fault: external abort on non-linefetch (0x1008) at 0x8887f02c
    Internal error: : 1008 [#1] ARM
    Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
    Backtrace:
      (regmap_mmio_read32le)
      (regmap_mmio_read)
      (_regmap_bus_reg_read)
      (_regmap_read)
      (regmap_read)
      (dspi_interrupt)
      (free_irq)
      (devm_irq_release)
      (release_nodes)
      (devres_release_all)
      (device_release_driver_internal)

The resource-managed framework should not be used for shared interrupt handling, because the interrupt handler might be called after releasing other resources and disabling clocks.

Similar bug could happen during suspend - the shared interrupt handler could be invoked after suspending the device. Each device sharing this interrupt line should disable the IRQ during suspend so handler will be invoked only in following cases:
1. None suspended,
2. All devices resumed.

Fixes 349ad66c0ab0 ("spi:Add Freescale DSPI driver for Vybrid VF610 platform")

---

Changes since v3:
1. Use simpler if (dspi->irq).

Changes since v2:
1. Go back to v1 and use non-devm interface,
2. Fix also suspend/resume paths.

Changes since v1:
1. Disable the IRQ instead of using non-devm interface.

---
 drivers/spi/spi-fsl-dspi.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)
drivers/spi/spi-fsl-dspi.c +13 -4
index ec7919d9c0d9..e0b30e4b1b69 100644
@@ -1109,6 +1109,8 @@ static int dspi_suspend(struct device *dev)
 	struct spi_controller *ctlr = dev_get_drvdata(dev);
 	struct fsl_dspi *dspi = spi_controller_get_devdata(ctlr);
 
+	if (dspi->irq)
+		disable_irq(dspi->irq);
 	spi_controller_suspend(ctlr);
 	clk_disable_unprepare(dspi->clk);
 
@@ -1129,6 +1131,8 @@ static int dspi_resume(struct device *dev)
 	if (ret)
 		return ret;
 	spi_controller_resume(ctlr);
+	if (dspi->irq)
+		enable_irq(dspi->irq);
 
 	return 0;
 }
@@ -1385,8 +1389,8 @@ static int dspi_probe(struct platform_device *pdev)
 		goto poll_mode;
 	}
 
-	ret = devm_request_irq(&pdev->dev, dspi->irq, dspi_interrupt,
-			       IRQF_SHARED, pdev->name, dspi);
+	ret = request_threaded_irq(dspi->irq, dspi_interrupt, NULL,
+				   IRQF_SHARED, pdev->name, dspi);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "Unable to attach DSPI interrupt\n");
 		goto out_clk_put;
@@ -1400,7 +1404,7 @@ static int dspi_probe(struct platform_device *pdev)
 		ret = dspi_request_dma(dspi, res->start);
 		if (ret < 0) {
 			dev_err(&pdev->dev, "can't get dma channels\n");
-			goto out_clk_put;
+			goto out_free_irq;
 		}
 	}
 
@@ -1415,11 +1419,14 @@ static int dspi_probe(struct platform_device *pdev)
 	ret = spi_register_controller(ctlr);
 	if (ret != 0) {
 		dev_err(&pdev->dev, "Problem registering DSPI ctlr\n");
-		goto out_clk_put;
+		goto out_free_irq;
 	}
 
 	return ret;
 
+out_free_irq:
+	if (dspi->irq)
+		free_irq(dspi->irq, dspi);
 out_clk_put:
 	clk_disable_unprepare(dspi->clk);
 out_ctlr_put:
@@ -1445,6 +1452,8 @@ static int dspi_remove(struct platform_device *pdev)
 	regmap_update_bits(dspi->regmap, SPI_MCR, SPI_MCR_HALT, SPI_MCR_HALT);
 
 	dspi_release_dma(dspi);
+	if (dspi->irq)
+		free_irq(dspi->irq, dspi);
 	clk_disable_unprepare(dspi->clk);
 
 	return 0;
Signature · 1 line
2.17.1
#5 lore raw

PATCH v4 4/4 spi: spi-fsl-dspi: Initialize completion before possible interrupt

In reply to Krzysztof Kozlowski

The interrupt handler calls completion and is IRQ requested before the completion is initialized. Logically it should be the other way.

---

Changes since v2:
1. None

Changes since v1:
1. Rework the commit msg.

---
 drivers/spi/spi-fsl-dspi.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/spi/spi-fsl-dspi.c +2 -2
index e0b30e4b1b69..91c6affe139c 100644
@@ -1389,6 +1389,8 @@ static int dspi_probe(struct platform_device *pdev)
 		goto poll_mode;
 	}
 
+	init_completion(&dspi->xfer_done);
+
 	ret = request_threaded_irq(dspi->irq, dspi_interrupt, NULL,
 				   IRQF_SHARED, pdev->name, dspi);
 	if (ret < 0) {
@@ -1396,8 +1398,6 @@ static int dspi_probe(struct platform_device *pdev)
 		goto out_clk_put;
 	}
 
-	init_completion(&dspi->xfer_done);
-
 poll_mode:
 
 	if (dspi->devtype_data->trans_mode == DSPI_DMA_MODE) {
Signature · 1 line
2.17.1
#1 lore raw

PATCH v4 2/4 spi: spi-fsl-dspi: Fix lockup if device is shutdown during SPI transfer

In reply to Krzysztof Kozlowski

During shutdown, the driver should unregister the SPI controller and stop the hardware. Otherwise the dspi_transfer_one_message() could wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device shutdown reverse-matches the probe function, where SPI controller is registered at the end.

Fixes dc234825997e ("spi: spi-fsl-dspi: Adding shutdown hook")

---

kexec() not tested, only system shutdown.

Changes since v3:
1. New patch.

---
 drivers/spi/spi-fsl-dspi.c | 15 +--------------
 1 file changed, 1 insertion(+), 14 deletions(-)
drivers/spi/spi-fsl-dspi.c +1 -14
index ec0fd0d366eb..ec7919d9c0d9 100644
@@ -1452,20 +1452,7 @@ static int dspi_remove(struct platform_device *pdev)
 
 static void dspi_shutdown(struct platform_device *pdev)
 {
-	struct spi_controller *ctlr = platform_get_drvdata(pdev);
-	struct fsl_dspi *dspi = spi_controller_get_devdata(ctlr);
-
-	/* Disable RX and TX */
-	regmap_update_bits(dspi->regmap, SPI_MCR,
-			   SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF,
-			   SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF);
-
-	/* Stop Running */
-	regmap_update_bits(dspi->regmap, SPI_MCR, SPI_MCR_HALT, SPI_MCR_HALT);
-
-	dspi_release_dma(dspi);
-	clk_disable_unprepare(dspi->clk);
-	spi_unregister_controller(dspi->ctlr);
+	dspi_remove(pdev);
 }
 
 static struct platform_driver fsl_dspi_driver = {
Signature · 1 line
2.17.1
maintainer Reviewed-by Tested-by on v4 3/4
#4 lore raw

Re: [PATCH v4 3/4] spi: spi-fsl-dspi: Fix external abort on interrupt in resume or exit paths

In reply to Krzysztof Kozlowski
Quoted 36 lines · Krzysztof Kozlowski

If shared interrupt comes late, during probe error path or device remove (could be triggered with CONFIG_DEBUG_SHIRQ), the interrupt handler dspi_interrupt() will access registers with the clock being disabled. This leads to external abort on non-linefetch on Toradex Colibri VF50 module (with Vybrid VF5xx):

    $ echo 4002d000.spi > /sys/devices/platform/soc/40000000.bus/4002d000.spi/driver/unbind
    Unhandled fault: external abort on non-linefetch (0x1008) at 0x8887f02c
    Internal error: : 1008 [#1] ARM
    Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
    Backtrace:
      (regmap_mmio_read32le)
      (regmap_mmio_read)
      (_regmap_bus_reg_read)
      (_regmap_read)
      (regmap_read)
      (dspi_interrupt)
      (free_irq)
      (devm_irq_release)
      (release_nodes)
      (devres_release_all)
      (device_release_driver_internal)

The resource-managed framework should not be used for shared interrupt handling, because the interrupt handler might be called after releasing other resources and disabling clocks.

Similar bug could happen during suspend - the shared interrupt handler could be invoked after suspending the device. Each device sharing this interrupt line should disable the IRQ during suspend so handler will be invoked only in following cases:
1. None suspended,
2. All devices resumed.

Quoted 5 lines
Fixes 349ad66c0ab0 ("spi:Add Freescale DSPI driver for Vybrid VF610 platform")

---

Quoted 84 lines

Changes since v3:
1. Use simpler if (dspi->irq).

Changes since v2:
1. Go back to v1 and use non-devm interface,
2. Fix also suspend/resume paths.

Changes since v1:
1. Disable the IRQ instead of using non-devm interface.

---
 drivers/spi/spi-fsl-dspi.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)
drivers/spi/spi-fsl-dspi.c +13 -5
index ec7919d9c0d9..e0b30e4b1b69 100644
@@ -1109,6 +1109,8 @@ static int dspi_suspend(struct device *dev)
        struct spi_controller *ctlr = dev_get_drvdata(dev);
        struct fsl_dspi *dspi = spi_controller_get_devdata(ctlr);
 
+       if (dspi->irq)
+               disable_irq(dspi->irq);
        spi_controller_suspend(ctlr);
        clk_disable_unprepare(dspi->clk);
 
@@ -1129,6 +1131,8 @@ static int dspi_resume(struct device *dev)
        if (ret)
                return ret;
        spi_controller_resume(ctlr);
+       if (dspi->irq)
+               enable_irq(dspi->irq);
 
        return 0;
 }
@@ -1385,8 +1389,8 @@ static int dspi_probe(struct platform_device *pdev)
                goto poll_mode;
        }
 
-       ret = devm_request_irq(&pdev->dev, dspi->irq, dspi_interrupt,
-                              IRQF_SHARED, pdev->name, dspi);
+       ret = request_threaded_irq(dspi->irq, dspi_interrupt, NULL,
+                                  IRQF_SHARED, pdev->name, dspi);
        if (ret < 0) {
                dev_err(&pdev->dev, "Unable to attach DSPI interrupt\n");
                goto out_clk_put;
@@ -1400,7 +1404,7 @@ static int dspi_probe(struct platform_device *pdev)
                ret = dspi_request_dma(dspi, res->start);
                if (ret < 0) {
                        dev_err(&pdev->dev, "can't get dma channels\n");
-                       goto out_clk_put;
+                       goto out_free_irq;
                }
        }
 
@@ -1415,11 +1419,14 @@ static int dspi_probe(struct platform_device *pdev)
        ret = spi_register_controller(ctlr);
        if (ret != 0) {
                dev_err(&pdev->dev, "Problem registering DSPI ctlr\n");
-               goto out_clk_put;
+               goto out_free_irq;
        }
 
        return ret;
 
+out_free_irq:
+       if (dspi->irq)
+               free_irq(dspi->irq, dspi);
 out_clk_put:
        clk_disable_unprepare(dspi->clk);
 out_ctlr_put:
@@ -1445,6 +1452,8 @@ static int dspi_remove(struct platform_device *pdev)
        regmap_update_bits(dspi->regmap, SPI_MCR, SPI_MCR_HALT, SPI_MCR_HALT);
 
        dspi_release_dma(dspi);
+       if (dspi->irq)
+               free_irq(dspi->irq, dspi);
        clk_disable_unprepare(dspi->clk);
 
        return 0;
--

2.17.1

maintainer Reviewed-by Tested-by on v4 2/4
#2 lore raw

Re: [PATCH v4 2/4] spi: spi-fsl-dspi: Fix lockup if device is shutdown during SPI transfer

In reply to Krzysztof Kozlowski
Quoted 9 lines · Krzysztof Kozlowski

During shutdown, the driver should unregister the SPI controller and stop the hardware. Otherwise the dspi_transfer_one_message() could wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device shutdown reverse-matches the probe function, where SPI controller is registered at the end.

Quoted 6 lines
Fixes dc234825997e ("spi: spi-fsl-dspi: Adding shutdown hook")

---

Quoted 38 lines

kexec() not tested, only system shutdown.

Changes since v3:
1. New patch.

---
 drivers/spi/spi-fsl-dspi.c | 15 +--------------
 1 file changed, 1 insertion(+), 14 deletions(-)
drivers/spi/spi-fsl-dspi.c +1 -15
index ec0fd0d366eb..ec7919d9c0d9 100644
@@ -1452,20 +1452,7 @@ static int dspi_remove(struct platform_device *pdev)
 
 static void dspi_shutdown(struct platform_device *pdev)
 {
-       struct spi_controller *ctlr = platform_get_drvdata(pdev);
-       struct fsl_dspi *dspi = spi_controller_get_devdata(ctlr);
-
-       /* Disable RX and TX */
-       regmap_update_bits(dspi->regmap, SPI_MCR,
-                          SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF,
-                          SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF);
-
-       /* Stop Running */
-       regmap_update_bits(dspi->regmap, SPI_MCR, SPI_MCR_HALT, SPI_MCR_HALT);
-
-       dspi_release_dma(dspi);
-       clk_disable_unprepare(dspi->clk);
-       spi_unregister_controller(dspi->ctlr);
+       dspi_remove(pdev);
 }
 
 static struct platform_driver fsl_dspi_driver = {
--

2.17.1

maintainer Reviewed-by Tested-by on v4 1/4
#6 lore raw
In reply to Krzysztof Kozlowski
Quoted 9 lines · Krzysztof Kozlowski

During device removal, the driver should unregister the SPI controller and stop the hardware. Otherwise the dspi_transfer_one_message() could wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device removal reverse-matches the probe function, where SPI controller is registered at the end.

Quoted 6 lines
Fixes 05209f457069 ("spi: fsl-dspi: add missing clk_disable_unprepare() in dspi_remove()")

---

Quoted 34 lines

Changes since v3:
1. New patch.

---
 drivers/spi/spi-fsl-dspi.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/spi/spi-fsl-dspi.c +10 -2
index 58190c94561f..ec0fd0d366eb 100644
@@ -1434,9 +1434,18 @@ static int dspi_remove(struct platform_device *pdev)
        struct fsl_dspi *dspi = spi_controller_get_devdata(ctlr);
 
        /* Disconnect from the SPI framework */
+       spi_unregister_controller(dspi->ctlr);
+
+       /* Disable RX and TX */
+       regmap_update_bits(dspi->regmap, SPI_MCR,
+                          SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF,
+                          SPI_MCR_DIS_TXF | SPI_MCR_DIS_RXF);
+
+       /* Stop Running */
+       regmap_update_bits(dspi->regmap, SPI_MCR, SPI_MCR_HALT, SPI_MCR_HALT);
+
        dspi_release_dma(dspi);
        clk_disable_unprepare(dspi->clk);
-       spi_unregister_controller(dspi->ctlr);
 
        return 0;
 }
--

2.17.1

maintainer
#7 lore raw
In reply to Krzysztof Kozlowski
Quoted 7 lines · Krzysztof Kozlowski

During device removal, the driver should unregister the SPI controller and stop the hardware. Otherwise the dspi_transfer_one_message() could wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device removal reverse-matches the probe function, where SPI controller is registered at the end.

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git for-next

Thanks!

[1/4] spi: spi-fsl-dspi: Fix lockup if device is removed during SPI transfer

      commit: 7684580d45bd3d84ed9b453a4cadf7a9a5605a3f
[2/4] spi: spi-fsl-dspi: Fix lockup if device is shutdown during SPI transfer
      commit: 3c525b69e8c1a9a6944e976603c7a1a713e728f9
[3/4] spi: spi-fsl-dspi: Fix external abort on interrupt in resume or exit paths
      commit: 3d87b613d6a3c6f0980e877ab0895785a2dde581
[4/4] spi: spi-fsl-dspi: Initialize completion before possible interrupt
      commit: f148915f91fccd8c3df1b0bff7d1c8458cad3be5

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark