Skip to contentSkip to conversation

Stop calling devm_request_irq() with invalid IRQs in the USB drivers

PATCH In review Sergey Shtylyov 4 participants 13 messages Aug 8, 2021 – Aug 13, 2021

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

v2 cover Stop calling devm_request_irq() with invalid IRQs in the USB drivers

Here are 9 patches against the 'usb-linus' branch of GregKH's 'usb.git' repo. The affected drivers call platform_get_irq() but largely ignore its result -- they blithely pass the negative error codes to request_irq() (and its ilk) which expects *unsinged* IRQ #s. Stop doing that by checking what exactly platform_get_irq() returns.

[1/9] usb: dwc3: meson-g12a: add IRQ check [2/9] usb: dwc3: qcom: add IRQ check [3/61] usb: gadget: udc: at91: add IRQ check [4/9] usb: gadget: udc: s3c2410: add IRQ check [5/9] usb: host: ohci-tmio: add IRQ check [6/9] usb: misc: brcmstb-usb-pinmap: add IRQ check [7/9] usb: phy: fsl-usb: add IRQ check [8/9] usb: phy: tahvo: add IRQ check
[9/9] usb: phy: twl6030: add IRQ checks

#1 lore raw

PATCH v2 1/9 usb: dwc3: meson-g12a: add IRQ check

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_threaded_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_threaded_irq() with the invalid IRQ #s.

Fixes f90db10779ad ("usb: dwc3: meson-g12a: Add support for IRQ based OTG switching")
---
 drivers/usb/dwc3/dwc3-meson-g12a.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/usb/dwc3/dwc3-meson-g12a.c +2 -0
Index: usb/drivers/usb/dwc3/dwc3-meson-g12a.c
===================================================================
@@ -598,6 +598,8 @@ static int dwc3_meson_g12a_otg_init(stru
 				   USB_R5_ID_DIG_IRQ, 0);
 
 		irq = platform_get_irq(pdev, 0);
+		if (irq < 0)
+			return irq;
 		ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
 						dwc3_meson_g12a_irq_thread,
 						IRQF_ONESHOT, pdev->name, priv);
reviewer Reviewed-by on v2 1/9
#2 lore raw

Re: [PATCH 1/9] usb: dwc3: meson-g12a: add IRQ check

Hi Sergey,

Quoted 9 lines · Sergey Shtylyov

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_threaded_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_threaded_irq() with the invalid IRQ #s.

Fixes f90db10779ad ("usb: dwc3: meson-g12a: Add support for IRQ based OTG switching")

Thanks for spotting and fixing this issue!

Acked-by on v2 1/9
#3 lore raw

Re: [PATCH 1/9] usb: dwc3: meson-g12a: add IRQ check

In reply to Sergey Shtylyov
Quoted 8 lines · Sergey Shtylyov

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_threaded_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_threaded_irq() with the invalid IRQ #s.

Fixes f90db10779ad ("usb: dwc3: meson-g12a: Add support for IRQ based OTG switching")
Signature · 1 line
balbi
#4 lore raw

PATCH v2 2/9 usb: dwc3: qcom: add IRQ check

In reply to Sergey Shtylyov

In dwc3_qcom_acpi_register_core(), the driver neglects to check the result of platform_get_irq()'s call and blithely assigns the negative error codes to the allocated child device's IRQ resource and then passing this resource to platform_device_add_resources() and later causing dwc3_otg_get_irq() to fail anyway. Stop calling platform_device_add_resources() with the invalid IRQ #s, so that there's less complexity in the IRQ error checking.

Fixes 2bc02355f8ba ("usb: dwc3: qcom: Add support for booting with ACPI")
---
 drivers/usb/dwc3/dwc3-qcom.c |    4 ++++
 1 file changed, 4 insertions(+)
drivers/usb/dwc3/dwc3-qcom.c +4 -0
Index: usb/drivers/usb/dwc3/dwc3-qcom.c
===================================================================
@@ -614,6 +614,10 @@ static int dwc3_qcom_acpi_register_core(
 		qcom->acpi_pdata->dwc3_core_base_size;
 
 	irq = platform_get_irq(pdev_irq, 0);
+	if (irq < 0) {
+		ret = irq;
+		goto out;
+	}
 	child_res[1].flags = IORESOURCE_IRQ;
 	child_res[1].start = child_res[1].end = irq;
 
Acked-by on v2 2/9
#5 lore raw

Re: [PATCH 2/9] usb: dwc3: qcom: add IRQ check

Quoted 9 lines · Sergey Shtylyov

In dwc3_qcom_acpi_register_core(), the driver neglects to check the result of platform_get_irq()'s call and blithely assigns the negative error codes to the allocated child device's IRQ resource and then passing this resource to platform_device_add_resources() and later causing dwc3_otg_get_irq() to fail anyway. Stop calling platform_device_add_resources() with the invalid IRQ #s, so that there's less complexity in the IRQ error checking.

Fixes 2bc02355f8ba ("usb: dwc3: qcom: Add support for booting with ACPI")
Signature · 1 line
balbi
#6 lore raw

PATCH v2 3/9 usb: gadget: udc: at91: add IRQ check

In reply to Sergey Shtylyov

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_irq() with the invalid IRQ #s.

Fixes 8b2e76687b39 ("USB: AT91 UDC updates, mostly power management")
---
 drivers/usb/gadget/udc/at91_udc.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/usb/gadget/udc/at91_udc.c +3 -1
Index: usb/drivers/usb/gadget/udc/at91_udc.c
===================================================================
@@ -1876,7 +1876,9 @@ static int at91udc_probe(struct platform
 	clk_disable(udc->iclk);
 
 	/* request UDC and maybe VBUS irqs */
-	udc->udp_irq = platform_get_irq(pdev, 0);
+	udc->udp_irq = retval = platform_get_irq(pdev, 0);
+	if (retval < 0)
+		goto err_unprepare_iclk;
 	retval = devm_request_irq(dev, udc->udp_irq, at91_udc_irq, 0,
 				  driver_name, udc);
 	if (retval) {
#7 lore raw

Re: [PATCH 3/9] usb: gadget: udc: at91: add IRQ check

Quoted 17 lines · Sergey Shtylyov

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_irq() with the invalid IRQ #s.

Fixes 8b2e76687b39 ("USB: AT91 UDC updates, mostly power management")
---
 drivers/usb/gadget/udc/at91_udc.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

Index: usb/drivers/usb/gadget/udc/at91_udc.c
===================================================================

drivers/usb/gadget/udc/at91_udc.c static int at91udc_probe(struct platform · line 1881

Quoted 7 lines
drivers/usb/gadget/udc/at91_udc.c +3 -1
@@ -1876,7 +1876,9 @@ static int at91udc_probe(struct platform
 	clk_disable(udc->iclk);
 
 	/* request UDC and maybe VBUS irqs */
-	udc->udp_irq = platform_get_irq(pdev, 0);
+	udc->udp_irq = retval = platform_get_irq(pdev, 0);
+	if (retval < 0)
+		goto err_unprepare_iclk;

let's avoid multiple assignments in one line. How about:

	if (udc->udp_irq < 0) {
        	retval = udc->udp_irq;
                goto err_unprepare_iclk;
        }
Signature · 1 line
balbi
#8 lore raw

PATCH v2 4/9 usb: gadget: udc: s3c2410: add IRQ check

In reply to Sergey Shtylyov

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_irq() with the invalid IRQ #s.

Fixes 8b2e76687b39 ("USB: AT91 UDC updates, mostly power management")
---
 drivers/usb/gadget/udc/at91_udc.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/usb/gadget/udc/at91_udc.c +3 -1
Index: usb/drivers/usb/gadget/udc/at91_udc.c
===================================================================
@@ -1876,7 +1876,9 @@ static int at91udc_probe(struct platform
 	clk_disable(udc->iclk);
 
 	/* request UDC and maybe VBUS irqs */
-	udc->udp_irq = platform_get_irq(pdev, 0);
+	udc->udp_irq = retval = platform_get_irq(pdev, 0);
+	if (retval < 0)
+		goto err_unprepare_iclk;
 	retval = devm_request_irq(dev, udc->udp_irq, at91_udc_irq, 0,
 				  driver_name, udc);
 	if (retval) {
#9 lore raw

Re: [PATCH 4/9] usb: gadget: udc: s3c2410: add IRQ check

Oops, duplicate patch. Scratch the series, I'm going to restart posting tomorrow... :-)

Acked-by on v2 4/9
#10 lore raw

Re: [PATCH 4/9] usb: gadget: udc: s3c2410: add IRQ check

In reply to Sergey Shtylyov
Quoted 7 lines · Sergey Shtylyov

The driver neglects to check the result of platform_get_irq()'s call and blithely passes the negative error codes to devm_request_irq() (which takes *unsigned* IRQ #), causing it to fail with -EINVAL, overriding an original error code. Stop calling devm_request_irq() with the invalid IRQ #s.

Fixes 8b2e76687b39 ("USB: AT91 UDC updates, mostly power management")
Signature · 1 line
balbi
#11 lore raw
In reply to Sergey Shtylyov
Quoted 9 lines · Sergey Shtylyov

Here are 9 patches against the 'usb-linus' branch of GregKH's 'usb.git' repo. The affected drivers call platform_get_irq() but largely ignore its result -- they blithely pass the negative error codes to request_irq() (and its ilk) which expects *unsinged* IRQ #s. Stop doing that by checking what exactly platform_get_irq() returns.

[1/9] usb: dwc3: meson-g12a: add IRQ check
[2/9] usb: dwc3: qcom: add IRQ check
[3/61] usb: gadget: udc: at91: add IRQ check

61?

Quoted 6 lines

[4/9] usb: gadget: udc: s3c2410: add IRQ check [5/9] usb: host: ohci-tmio: add IRQ check [6/9] usb: misc: brcmstb-usb-pinmap: add IRQ check [7/9] usb: phy: fsl-usb: add IRQ check [8/9] usb: phy: tahvo: add IRQ check
[9/9] usb: phy: twl6030: add IRQ checks

I only see 4 patches in this series, and the 4th, did not apply to my tree :(

Can you please fix this up and resend it correctly?

thanks,

greg k-h

#12 lore raw
Quoted 21 lines · Greg Kroah-Hartman
Quoted 9 lines · Sergey Shtylyov

Here are 9 patches against the 'usb-linus' branch of GregKH's 'usb.git' repo. The affected drivers call platform_get_irq() but largely ignore its result -- they blithely pass the negative error codes to request_irq() (and its ilk) which expects *unsinged* IRQ #s. Stop doing that by checking what exactly platform_get_irq() returns.

[1/9] usb: dwc3: meson-g12a: add IRQ check
[2/9] usb: dwc3: qcom: add IRQ check
[3/61] usb: gadget: udc: at91: add IRQ check

61?

Quoted 6 lines

[4/9] usb: gadget: udc: s3c2410: add IRQ check [5/9] usb: host: ohci-tmio: add IRQ check [6/9] usb: misc: brcmstb-usb-pinmap: add IRQ check [7/9] usb: phy: fsl-usb: add IRQ check [8/9] usb: phy: tahvo: add IRQ check
[9/9] usb: phy: twl6030: add IRQ checks

Quoted 4 lines

I only see 4 patches in this series, and the 4th, did not apply to my tree :(

Can you please fix this up and resend it correctly?

Ah, see your v2 now...