[v4] Enahance SVS's robustness

[PATCH v4 03/14] soc: mediatek: mtk-svs: enable the IRQ later

Posted by Roger Lu 1 year, 10 months ago

From: Ricardo Ribalda <ribalda@chromium.org>

If the system does not come from reset (like when is booted via
kexec(), the peripheral might trigger an IRQ before the data structures
are initialized.

Fixes:

[    0.227710] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000f08
[    0.227913] Call trace:
[    0.227918]  svs_isr+0x8c/0x538

Fixes: 681a02e95000 ("soc: mediatek: SVS: introduce MTK SVS engine")
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
Signed-off-by: Roger Lu <roger.lu@mediatek.com>
---
 drivers/soc/mediatek/mtk-svs.c | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
index a803b92afc3d..64d4d03ab71c 100644
--- a/drivers/soc/mediatek/mtk-svs.c
+++ b/drivers/soc/mediatek/mtk-svs.c
@@ -2403,20 +2403,6 @@ static int svs_probe(struct platform_device *pdev)
 		goto svs_probe_free_resource;
 	}
 
-	svsp_irq = platform_get_irq(pdev, 0);
-	if (svsp_irq < 0) {
-		ret = svsp_irq;
-		goto svs_probe_free_resource;
-	}
-
-	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
-					IRQF_ONESHOT, svsp->name, svsp);
-	if (ret) {
-		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
-			svsp_irq, ret);
-		goto svs_probe_free_resource;
-	}
-
 	svsp->main_clk = devm_clk_get(svsp->dev, "main");
 	if (IS_ERR(svsp->main_clk)) {
 		dev_err(svsp->dev, "failed to get clock: %ld\n",
@@ -2438,6 +2424,20 @@ static int svs_probe(struct platform_device *pdev)
 		goto svs_probe_clk_disable;
 	}
 
+	svsp_irq = platform_get_irq(pdev, 0);
+	if (svsp_irq < 0) {
+		ret = svsp_irq;
+		goto svs_probe_iounmap;
+	}
+
+	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
+					IRQF_ONESHOT, svsp->name, svsp);
+	if (ret) {
+		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
+			svsp_irq, ret);
+		goto svs_probe_iounmap;
+	}
+
 	ret = svs_start(svsp);
 	if (ret) {
 		dev_err(svsp->dev, "svs start fail: %d\n", ret);
-- 
2.18.0

Re: [PATCH v4 03/14] soc: mediatek: mtk-svs: enable the IRQ later

Posted by Matthias Brugger 1 year, 9 months ago


On 11/01/2023 08:45, Roger Lu wrote:
> From: Ricardo Ribalda <ribalda@chromium.org>
> 
> If the system does not come from reset (like when is booted via
> kexec(), the peripheral might trigger an IRQ before the data structures
> are initialized.
> 
> Fixes:
> 
> [    0.227710] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000f08
> [    0.227913] Call trace:
> [    0.227918]  svs_isr+0x8c/0x538
> 

This patch is different from
https://lore.kernel.org/r/20221127-mtk-svs-v2-0-145b07663ea8@chromium.org

If you think the patch from Ricardo is wrong, then you should have mentioned 
that in the review.

Regards,
Matthias

> Fixes: 681a02e95000 ("soc: mediatek: SVS: introduce MTK SVS engine")
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> Signed-off-by: Roger Lu <roger.lu@mediatek.com>
> ---
>   drivers/soc/mediatek/mtk-svs.c | 28 ++++++++++++++--------------
>   1 file changed, 14 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
> index a803b92afc3d..64d4d03ab71c 100644
> --- a/drivers/soc/mediatek/mtk-svs.c
> +++ b/drivers/soc/mediatek/mtk-svs.c
> @@ -2403,20 +2403,6 @@ static int svs_probe(struct platform_device *pdev)
>   		goto svs_probe_free_resource;
>   	}
>   
> -	svsp_irq = platform_get_irq(pdev, 0);
> -	if (svsp_irq < 0) {
> -		ret = svsp_irq;
> -		goto svs_probe_free_resource;
> -	}
> -
> -	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
> -					IRQF_ONESHOT, svsp->name, svsp);
> -	if (ret) {
> -		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
> -			svsp_irq, ret);
> -		goto svs_probe_free_resource;
> -	}
> -
>   	svsp->main_clk = devm_clk_get(svsp->dev, "main");
>   	if (IS_ERR(svsp->main_clk)) {
>   		dev_err(svsp->dev, "failed to get clock: %ld\n",
> @@ -2438,6 +2424,20 @@ static int svs_probe(struct platform_device *pdev)
>   		goto svs_probe_clk_disable;
>   	}
>   
> +	svsp_irq = platform_get_irq(pdev, 0);
> +	if (svsp_irq < 0) {
> +		ret = svsp_irq;
> +		goto svs_probe_iounmap;
> +	}
> +
> +	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
> +					IRQF_ONESHOT, svsp->name, svsp);
> +	if (ret) {
> +		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
> +			svsp_irq, ret);
> +		goto svs_probe_iounmap;
> +	}
> +
>   	ret = svs_start(svsp);
>   	if (ret) {
>   		dev_err(svsp->dev, "svs start fail: %d\n", ret);

Re: [PATCH v4 03/14] soc: mediatek: mtk-svs: enable the IRQ later

Posted by Roger Lu (陸瑞傑) 1 year, 9 months ago

Hi Matthias Sir,

On Tue, 2023-01-31 at 13:59 +0100, Matthias Brugger wrote:
> 
> On 11/01/2023 08:45, Roger Lu wrote:
> > From: Ricardo Ribalda <ribalda@chromium.org>
> > 
> > If the system does not come from reset (like when is booted via
> > kexec(), the peripheral might trigger an IRQ before the data structures
> > are initialized.
> > 
> > Fixes:
> > 
> > [    0.227710] Unable to handle kernel NULL pointer dereference at virtual
> > address 0000000000000f08
> > [    0.227913] Call trace:
> > [    0.227918]  svs_isr+0x8c/0x538
> > 
> 
> This patch is different from
> https://lore.kernel.org/r/20221127-mtk-svs-v2-0-145b07663ea8@chromium.org
> 
> If you think the patch from Ricardo is wrong, then you should have mentioned 
> that in the review.

My bad. Ricardo's patch is correct. Moving platform_get_irq() brings no
functional change and is for helping the svs probe readability as my preference.

> 
> Regards,
> Matthias
> 
> > Fixes: 681a02e95000 ("soc: mediatek: SVS: introduce MTK SVS engine")
> > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > Signed-off-by: Roger Lu <roger.lu@mediatek.com>
> > ---
> >   drivers/soc/mediatek/mtk-svs.c | 28 ++++++++++++++--------------
> >   1 file changed, 14 insertions(+), 14 deletions(-)
> > 
> > diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
> > index a803b92afc3d..64d4d03ab71c 100644
> > --- a/drivers/soc/mediatek/mtk-svs.c
> > +++ b/drivers/soc/mediatek/mtk-svs.c
> > @@ -2403,20 +2403,6 @@ static int svs_probe(struct platform_device *pdev)
> >   		goto svs_probe_free_resource;
> >   	}
> >   
> > -	svsp_irq = platform_get_irq(pdev, 0);
> > -	if (svsp_irq < 0) {
> > -		ret = svsp_irq;
> > -		goto svs_probe_free_resource;
> > -	}
> > -
> > -	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
> > -					IRQF_ONESHOT, svsp->name, svsp);
> > -	if (ret) {
> > -		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
> > -			svsp_irq, ret);
> > -		goto svs_probe_free_resource;
> > -	}
> > -
> >   	svsp->main_clk = devm_clk_get(svsp->dev, "main");
> >   	if (IS_ERR(svsp->main_clk)) {
> >   		dev_err(svsp->dev, "failed to get clock: %ld\n",
> > @@ -2438,6 +2424,20 @@ static int svs_probe(struct platform_device *pdev)
> >   		goto svs_probe_clk_disable;
> >   	}
> >   
> > +	svsp_irq = platform_get_irq(pdev, 0);
> > +	if (svsp_irq < 0) {
> > +		ret = svsp_irq;
> > +		goto svs_probe_iounmap;
> > +	}
> > +
> > +	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
> > +					IRQF_ONESHOT, svsp->name, svsp);
> > +	if (ret) {
> > +		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
> > +			svsp_irq, ret);
> > +		goto svs_probe_iounmap;
> > +	}
> > +
> >   	ret = svs_start(svsp);
> >   	if (ret) {
> >   		dev_err(svsp->dev, "svs start fail: %d\n", ret);

Re: [PATCH v4 03/14] soc: mediatek: mtk-svs: enable the IRQ later

Posted by Matthias Brugger 1 year, 9 months ago


On 01/02/2023 14:43, Roger Lu (陸瑞傑) wrote:
> Hi Matthias Sir,
> 
> On Tue, 2023-01-31 at 13:59 +0100, Matthias Brugger wrote:
>>
>> On 11/01/2023 08:45, Roger Lu wrote:
>>> From: Ricardo Ribalda <ribalda@chromium.org>
>>>
>>> If the system does not come from reset (like when is booted via
>>> kexec(), the peripheral might trigger an IRQ before the data structures
>>> are initialized.
>>>
>>> Fixes:
>>>
>>> [    0.227710] Unable to handle kernel NULL pointer dereference at virtual
>>> address 0000000000000f08
>>> [    0.227913] Call trace:
>>> [    0.227918]  svs_isr+0x8c/0x538
>>>
>>
>> This patch is different from
>> https://lore.kernel.org/r/20221127-mtk-svs-v2-0-145b07663ea8@chromium.org
>>
>> If you think the patch from Ricardo is wrong, then you should have mentioned
>> that in the review.
> 
> My bad. Ricardo's patch is correct. Moving platform_get_irq() brings no
> functional change and is for helping the svs probe readability as my preference.
> 

Thanks for the confirmation!

Regards,
Matthias

>>
>> Regards,
>> Matthias
>>
>>> Fixes: 681a02e95000 ("soc: mediatek: SVS: introduce MTK SVS engine")
>>> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
>>> Signed-off-by: Roger Lu <roger.lu@mediatek.com>
>>> ---
>>>    drivers/soc/mediatek/mtk-svs.c | 28 ++++++++++++++--------------
>>>    1 file changed, 14 insertions(+), 14 deletions(-)
>>>
>>> diff --git a/drivers/soc/mediatek/mtk-svs.c b/drivers/soc/mediatek/mtk-svs.c
>>> index a803b92afc3d..64d4d03ab71c 100644
>>> --- a/drivers/soc/mediatek/mtk-svs.c
>>> +++ b/drivers/soc/mediatek/mtk-svs.c
>>> @@ -2403,20 +2403,6 @@ static int svs_probe(struct platform_device *pdev)
>>>    		goto svs_probe_free_resource;
>>>    	}
>>>    
>>> -	svsp_irq = platform_get_irq(pdev, 0);
>>> -	if (svsp_irq < 0) {
>>> -		ret = svsp_irq;
>>> -		goto svs_probe_free_resource;
>>> -	}
>>> -
>>> -	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
>>> -					IRQF_ONESHOT, svsp->name, svsp);
>>> -	if (ret) {
>>> -		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
>>> -			svsp_irq, ret);
>>> -		goto svs_probe_free_resource;
>>> -	}
>>> -
>>>    	svsp->main_clk = devm_clk_get(svsp->dev, "main");
>>>    	if (IS_ERR(svsp->main_clk)) {
>>>    		dev_err(svsp->dev, "failed to get clock: %ld\n",
>>> @@ -2438,6 +2424,20 @@ static int svs_probe(struct platform_device *pdev)
>>>    		goto svs_probe_clk_disable;
>>>    	}
>>>    
>>> +	svsp_irq = platform_get_irq(pdev, 0);
>>> +	if (svsp_irq < 0) {
>>> +		ret = svsp_irq;
>>> +		goto svs_probe_iounmap;
>>> +	}
>>> +
>>> +	ret = devm_request_threaded_irq(svsp->dev, svsp_irq, NULL, svs_isr,
>>> +					IRQF_ONESHOT, svsp->name, svsp);
>>> +	if (ret) {
>>> +		dev_err(svsp->dev, "register irq(%d) failed: %d\n",
>>> +			svsp_irq, ret);
>>> +		goto svs_probe_iounmap;
>>> +	}
>>> +
>>>    	ret = svs_start(svsp);
>>>    	if (ret) {
>>>    		dev_err(svsp->dev, "svs start fail: %d\n", ret);