platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com --- Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar --- drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
- res = platform_get_resource(pdev, IORESOURCE_IRQ, 0); - if (!res) { - dev_err(&pdev->dev, "no slimbus IRQ resource\n"); - return -ENODEV; - } + ret = platform_get_irq(pdev, 0); + if (ret < 0) + return ret;
- ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt, + ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");
On Fri 24 Dec 08:13 PST 2021, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
Reviewed-by: Bjorn Andersson bjorn.andersson@linaro.org
Regards, Bjorn
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com
Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar
drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
- res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
- if (!res) {
dev_err(&pdev->dev, "no slimbus IRQ resource\n");return -ENODEV;- }
- ret = platform_get_irq(pdev, 0);
- if (ret < 0)
return ret;
- ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt,
- ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");
-- 2.17.1
On 12/24/21 10:13 AM, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com
Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar
drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
- res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
- if (!res) {
dev_err(&pdev->dev, "no slimbus IRQ resource\n");return -ENODEV;- }
- ret = platform_get_irq(pdev, 0);
- if (ret < 0)
return ret;
- ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt,
- ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");
Tested on Lenovo Yoga C630
Tested-By: Steev Klimaszewski steev@kali.org
Hi Srinivas,
On Fri, Dec 24, 2021 at 4:13 PM Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com
Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar
drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
This patch is not in -next yet. When do you plan to merge this patch?
Cheers, Prabhakar
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);if (!res) {dev_err(&pdev->dev, "no slimbus IRQ resource\n");return -ENODEV;}
ret = platform_get_irq(pdev, 0);if (ret < 0)return ret;
ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt,
ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");-- 2.17.1
On 24/12/2021 16:13, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue
Are you saying that we should not be using platform_get_resource(pdev, IORESOURCE_IRQ, ...) on drivers that support DT?
when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
Should this not be fixed in the DT core itself?
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
I would prefer this patch to be part of the series that removes IRQ resource handling from DT core.
--srini
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com
Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar
drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
- res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
- if (!res) {
dev_err(&pdev->dev, "no slimbus IRQ resource\n");return -ENODEV;- }
- ret = platform_get_irq(pdev, 0);
- if (ret < 0)
return ret;
- ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt,
- ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");
On Thu, Mar 10, 2022 at 10:16 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 24/12/2021 16:13, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue
Are you saying that we should not be using platform_get_resource(pdev, IORESOURCE_IRQ, ...) on drivers that support DT?
when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
Should this not be fixed in the DT core itself?
Yes the plan is to fix in the DT core itself (refer [0]).
[0] https://patchwork.kernel.org/project/linux-renesas-soc/patch/20211209001056....
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
I would prefer this patch to be part of the series that removes IRQ resource handling from DT core.
Since there are too many users (which are in different subsystems) getting this all in single series would be a pain. As a result it is split up into individual subsystems.
Cheers, Prabhakar
--srini
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com
Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar
drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);if (!res) {dev_err(&pdev->dev, "no slimbus IRQ resource\n");return -ENODEV;}
ret = platform_get_irq(pdev, 0);if (ret < 0)return ret;
ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt,
ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");
On 10/03/2022 10:23, Lad, Prabhakar wrote:
On Thu, Mar 10, 2022 at 10:16 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 24/12/2021 16:13, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue
Are you saying that we should not be using platform_get_resource(pdev, IORESOURCE_IRQ, ...) on drivers that support DT?
when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
Should this not be fixed in the DT core itself?
Yes the plan is to fix in the DT core itself (refer [0]).
[0] https://patchwork.kernel.org/project/linux-renesas-soc/patch/20211209001056....
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
I would prefer this patch to be part of the series that removes IRQ resource handling from DT core.
Since there are too many users (which are in different subsystems) getting this all in single series would be a pain. As a result it is split up into individual subsystems.
Am happy for this to be included in that series, TBH, this patch make more sense along with that series than by itself.
This would also give better insight of what exactly is changing in platform_get_resource() w.r.t handling IRQ resources.
--srini
Cheers, Prabhakar
--srini
Signed-off-by: Lad Prabhakar prabhakar.mahadev-lad.rj@bp.renesas.com
Hi,
Dropping usage of platform_get_resource() was agreed based on the discussion [0].
[0] https://patchwork.kernel.org/project/linux-renesas-soc/ patch/20211209001056.29774-1-prabhakar.mahadev-lad.rj@bp.renesas.com/
Cheers, Prabhakar
drivers/slimbus/qcom-ngd-ctrl.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/slimbus/qcom-ngd-ctrl.c b/drivers/slimbus/qcom-ngd-ctrl.c index 7040293c2ee8..0f29a08b4c09 100644 --- a/drivers/slimbus/qcom-ngd-ctrl.c +++ b/drivers/slimbus/qcom-ngd-ctrl.c @@ -1526,13 +1526,11 @@ static int qcom_slim_ngd_ctrl_probe(struct platform_device *pdev) if (IS_ERR(ctrl->base)) return PTR_ERR(ctrl->base);
res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);if (!res) {dev_err(&pdev->dev, "no slimbus IRQ resource\n");return -ENODEV;}
ret = platform_get_irq(pdev, 0);if (ret < 0)return ret;
ret = devm_request_irq(dev, res->start, qcom_slim_ngd_interrupt,
ret = devm_request_irq(dev, ret, qcom_slim_ngd_interrupt, IRQF_TRIGGER_HIGH, "slim-ngd", ctrl); if (ret) { dev_err(&pdev->dev, "request IRQ failed\n");
On Thu, Mar 10, 2022 at 4:42 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 10/03/2022 10:23, Lad, Prabhakar wrote:
On Thu, Mar 10, 2022 at 10:16 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 24/12/2021 16:13, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue
Are you saying that we should not be using platform_get_resource(pdev, IORESOURCE_IRQ, ...) on drivers that support DT?
We should be using platform_get_irq(). (period, on all platform drivers)
when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
Should this not be fixed in the DT core itself?
Yes the plan is to fix in the DT core itself (refer [0]).
[0] https://patchwork.kernel.org/project/linux-renesas-soc/patch/20211209001056....
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
I would prefer this patch to be part of the series that removes IRQ resource handling from DT core.
Since there are too many users (which are in different subsystems) getting this all in single series would be a pain. As a result it is split up into individual subsystems.
Am happy for this to be included in that series, TBH, this patch make more sense along with that series than by itself.
No it doesn't. This is no different than converting to devm_* variants or other cleanups to match current preferred styles.
Treewide cross subsystem clean-ups are a huge pain to merge. Why would you ask for that when it is clearly not necessary?
Rob
On 10/03/2022 14:14, Rob Herring wrote:
On Thu, Mar 10, 2022 at 4:42 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 10/03/2022 10:23, Lad, Prabhakar wrote:
On Thu, Mar 10, 2022 at 10:16 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 24/12/2021 16:13, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue
Are you saying that we should not be using platform_get_resource(pdev, IORESOURCE_IRQ, ...) on drivers that support DT?
We should be using platform_get_irq(). (period, on all platform drivers)
Thanks, I see why is it preferred.
Code as of now will not prevent drivers from calling platform_get_resource(..IORESOURCE_IRQ).
Are we planning to enforce this in any way?
when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
Should this not be fixed in the DT core itself?
Yes the plan is to fix in the DT core itself (refer [0]).
[0] https://patchwork.kernel.org/project/linux-renesas-soc/patch/20211209001056....
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
I would prefer this patch to be part of the series that removes IRQ resource handling from DT core.
Since there are too many users (which are in different subsystems) getting this all in single series would be a pain. As a result it is split up into individual subsystems.
Am happy for this to be included in that series, TBH, this patch make more sense along with that series than by itself.
No it doesn't. This is no different than converting to devm_* variants or other cleanups to match current preferred styles.
Treewide cross subsystem clean-ups are a huge pain to merge. Why would you ask for that when it is clearly not necessary?
Only reason for this ask was to understand how platform_get_resource() will change moving forward, if this is something that you are planning to include in your fix patches.
I can go ahead and apply the patch, if that helps.
--srini
Rob
On Thu, Mar 10, 2022 at 2:59 PM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 10/03/2022 14:14, Rob Herring wrote:
On Thu, Mar 10, 2022 at 4:42 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 10/03/2022 10:23, Lad, Prabhakar wrote:
On Thu, Mar 10, 2022 at 10:16 AM Srinivas Kandagatla srinivas.kandagatla@linaro.org wrote:
On 24/12/2021 16:13, Lad Prabhakar wrote:
platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on static allocation of IRQ resources in DT core code, this causes an issue
Are you saying that we should not be using platform_get_resource(pdev, IORESOURCE_IRQ, ...) on drivers that support DT?
We should be using platform_get_irq(). (period, on all platform drivers)
Thanks, I see why is it preferred.
Code as of now will not prevent drivers from calling platform_get_resource(..IORESOURCE_IRQ).
Are we planning to enforce this in any way?
when using hierarchical interrupt domains using "interrupts" property in the node as this bypasses the hierarchical setup and messes up the irq chaining.
Should this not be fixed in the DT core itself?
Yes the plan is to fix in the DT core itself (refer [0]).
[0] https://patchwork.kernel.org/project/linux-renesas-soc/patch/20211209001056....
In preparation for removal of static setup of IRQ resource from DT core code use platform_get_irq().
I would prefer this patch to be part of the series that removes IRQ resource handling from DT core.
Since there are too many users (which are in different subsystems) getting this all in single series would be a pain. As a result it is split up into individual subsystems.
Am happy for this to be included in that series, TBH, this patch make more sense along with that series than by itself.
No it doesn't. This is no different than converting to devm_* variants or other cleanups to match current preferred styles.
Treewide cross subsystem clean-ups are a huge pain to merge. Why would you ask for that when it is clearly not necessary?
Only reason for this ask was to understand how platform_get_resource() will change moving forward, if this is something that you are planning to include in your fix patches.
I can go ahead and apply the patch, if that helps.
Yes please, that would be helpful.
Cheers, Prabhakar
-
Bjorn Andersson -
Lad Prabhakar -
Lad, Prabhakar -
Rob Herring -
Srinivas Kandagatla -
Steev Klimaszewski