* [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling
@ 2026-09-29 12:02 Linus Walleij
2026-09-29 12:02 ` [PATCH 1/3] dt-bindings: media: mali-c55: Allow wakeup source Linus Walleij
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Linus Walleij @ 2026-09-29 12:02 UTC (permalink / raw)
To: Li Youhong, Daniel Scally, Jacopo Mondi, Mauro Carvalho Chehab,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Hans Verkuil,
Nayden Kanchev
Cc: linux-media, devicetree, linux-kernel, Linus Walleij, stable
This is essentially my reply with the "proper fix" for:
https://lore.kernel.org/linux-media/20260903025526.2657105-1-dayou5941@163.com/
Keep the Mali-C55 interrupt requested for the driver's lifetime instead
of changing interrupt ownership along with the runtime power state.
Synchronize and disable the IRQ before power-off, and enable it after
power-on. Hold a PM reference until probe installs the IRQ, and stop
runtime PM and drain the IRQ before removing the media entities.
Allow platforms to mark the ISP as a wakeup source. Hold a runtime PM
reference while IRQ wake is armed, resuming an idle ISP if needed and
preserving an active ISP's configuration. Balance the reference on resume
and failure. Hibernation still uses the power-down path.
Wake events require a configured stream and an input that remains active
during system sleep. The internal TPG shares the ISP's power state;
external sensors and receivers need their own wake support. An idle
pipeline is not started by this series. Hardware wake remains untested.
The first patch extends the binding with the standard wakeup-source
property. The second patch updates the driver IRQ and PM handling. The
third patch retains ISP power while IRQ wake is armed.
Tested on the Renesas RZ/V2H board with the other in-flight patches
(DTS changes and the TPG enablement) doing full suspend/resume cycles
with wakeup from RTC and NMI and capturing images before and
after the test.
Assisted-by: LLM
Signed-off-by: Linus Walleij <linusw@kernel.org>
---
Linus Walleij (3):
dt-bindings: media: mali-c55: Allow wakeup source
media: mali-c55: Keep IRQ requested during suspend
media: mali-c55: Keep ISP powered while IRQ wake is armed
.../devicetree/bindings/media/arm,mali-c55.yaml | 2 +
.../media/platform/arm/mali-c55/mali-c55-core.c | 96 +++++++++++++++++-----
2 files changed, 78 insertions(+), 20 deletions(-)
---
base-commit: cee9395acd8043be0644b25c34bfa86623f2b935
change-id: 20260918-mali-c55-irq-supend-resume-30b507a20011
Best regards,
--
Linus Walleij <linusw@kernel.org>
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 1/3] dt-bindings: media: mali-c55: Allow wakeup source
2026-09-29 12:02 [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Linus Walleij
@ 2026-09-29 12:02 ` Linus Walleij
2026-09-29 12:02 ` [PATCH 2/3] media: mali-c55: Keep IRQ requested during suspend Linus Walleij
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Linus Walleij @ 2026-09-29 12:02 UTC (permalink / raw)
To: Li Youhong, Daniel Scally, Jacopo Mondi, Mauro Carvalho Chehab,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Hans Verkuil,
Nayden Kanchev
Cc: linux-media, devicetree, linux-kernel, Linus Walleij
The Mali-C55 interrupt can be wired as a wake source on platforms that
support it. Use the standard wakeup-source property.
Assisted-by: LLM
Signed-off-by: Linus Walleij <linusw@kernel.org>
---
Documentation/devicetree/bindings/media/arm,mali-c55.yaml | 2 ++
1 file changed, 2 insertions(+)
diff --git a/Documentation/devicetree/bindings/media/arm,mali-c55.yaml b/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
index fc4fcd19922a..acd57799ccd0 100644
--- a/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
+++ b/Documentation/devicetree/bindings/media/arm,mali-c55.yaml
@@ -20,6 +20,8 @@ properties:
interrupts:
maxItems: 1
+ wakeup-source: true
+
clocks:
items:
- description: ISP Video Clock
--
2.55.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 2/3] media: mali-c55: Keep IRQ requested during suspend
2026-09-29 12:02 [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Linus Walleij
2026-09-29 12:02 ` [PATCH 1/3] dt-bindings: media: mali-c55: Allow wakeup source Linus Walleij
@ 2026-09-29 12:02 ` Linus Walleij
2026-09-29 12:02 ` [PATCH 3/3] media: mali-c55: Keep ISP powered while IRQ wake is armed Linus Walleij
2026-10-06 11:43 ` [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Lad, Prabhakar
3 siblings, 0 replies; 5+ messages in thread
From: Linus Walleij @ 2026-09-29 12:02 UTC (permalink / raw)
To: Li Youhong, Daniel Scally, Jacopo Mondi, Mauro Carvalho Chehab,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Hans Verkuil,
Nayden Kanchev
Cc: linux-media, devicetree, linux-kernel, Linus Walleij, stable
The interrupt is currently freed on every runtime suspend and requested
again on runtime resume. Apart from tying interrupt ownership to the
power state rather than to the driver lifetime, this leaves remove to
guess whether an action is installed for the IRQ.
Request the interrupt once during probe and free it during remove.
Disable and synchronize the IRQ before powering the ISP off, and enable
it only after a successful power-on.
Hold a runtime PM reference during probe until the IRQ is installed,
and stop runtime PM and drain the IRQ before unregistering the media
entities during remove.
If firmware marks the ISP as a wakeup source, initialize device wakeup
and enable IRQ wake during system suspend. Disable it again before
resuming the device.
This configures the interrupt controller wake path.
Fixes: d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")
Cc: stable@vger.kernel.org
Assisted-by: LLM
Signed-off-by: Linus Walleij <linusw@kernel.org>
---
.../media/platform/arm/mali-c55/mali-c55-core.c | 84 ++++++++++++++++------
1 file changed, 64 insertions(+), 20 deletions(-)
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-core.c b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
index f28e9f4354ac..07267b79801b 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-core.c
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
@@ -17,6 +17,8 @@
#include <linux/of_reserved_mem.h>
#include <linux/platform_device.h>
#include <linux/pm_runtime.h>
+#include <linux/pm_wakeup.h>
+#include <linux/property.h>
#include <linux/reset.h>
#include <linux/slab.h>
#include <linux/string.h>
@@ -675,8 +677,7 @@ static int __maybe_unused mali_c55_runtime_suspend(struct device *dev)
{
struct mali_c55 *mali_c55 = dev_get_drvdata(dev);
- if (irq_has_action(mali_c55->irqnum))
- free_irq(mali_c55->irqnum, dev);
+ disable_irq(mali_c55->irqnum);
__mali_c55_power_off(mali_c55);
return 0;
@@ -745,25 +746,41 @@ static int __maybe_unused mali_c55_runtime_resume(struct device *dev)
if (ret)
return ret;
- /*
- * The driver needs to transfer large amounts of register settings to
- * the ISP each frame, using either a DMA transfer or memcpy. We use a
- * threaded IRQ to avoid disabling interrupts the entire time that's
- * happening.
- */
- ret = request_threaded_irq(mali_c55->irqnum, NULL, mali_c55_isr,
- IRQF_ONESHOT, dev_driver_string(dev), dev);
- if (ret) {
- __mali_c55_power_off(mali_c55);
- dev_err(dev, "failed to request irq\n");
+ enable_irq(mali_c55->irqnum);
+
+ return 0;
+}
+
+static int __maybe_unused mali_c55_suspend(struct device *dev)
+{
+ struct mali_c55 *mali_c55 = dev_get_drvdata(dev);
+ int ret;
+
+ if (device_may_wakeup(dev)) {
+ ret = enable_irq_wake(mali_c55->irqnum);
+ if (ret)
+ return ret;
}
+ ret = pm_runtime_force_suspend(dev);
+ if (ret && device_may_wakeup(dev))
+ disable_irq_wake(mali_c55->irqnum);
+
return ret;
}
+static int __maybe_unused mali_c55_resume(struct device *dev)
+{
+ struct mali_c55 *mali_c55 = dev_get_drvdata(dev);
+
+ if (device_may_wakeup(dev))
+ disable_irq_wake(mali_c55->irqnum);
+
+ return pm_runtime_force_resume(dev);
+}
+
static const struct dev_pm_ops mali_c55_pm_ops = {
- SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
- pm_runtime_force_resume)
+ SET_SYSTEM_SLEEP_PM_OPS(mali_c55_suspend, mali_c55_resume)
SET_RUNTIME_PM_OPS(mali_c55_runtime_suspend, mali_c55_runtime_resume,
NULL)
};
@@ -825,27 +842,53 @@ static int mali_c55_probe(struct platform_device *pdev)
pm_runtime_set_autosuspend_delay(&pdev->dev, 2000);
pm_runtime_use_autosuspend(&pdev->dev);
pm_runtime_set_active(&pdev->dev);
+ pm_runtime_get_noresume(dev);
pm_runtime_enable(&pdev->dev);
ret = mali_c55_media_frameworks_init(mali_c55);
if (ret)
goto err_pm_runtime_disable;
- pm_runtime_idle(&pdev->dev);
-
mali_c55->irqnum = platform_get_irq(pdev, 0);
if (mali_c55->irqnum < 0) {
ret = mali_c55->irqnum;
goto err_deinit_media_frameworks;
}
+ /*
+ * The driver needs to transfer large amounts of register settings to
+ * the ISP each frame, using either a DMA transfer or memcpy. We use a
+ * threaded IRQ to avoid disabling interrupts the entire time that's
+ * happening.
+ */
+ ret = request_threaded_irq(mali_c55->irqnum, NULL, mali_c55_isr,
+ IRQF_ONESHOT, dev_driver_string(dev), dev);
+ if (ret) {
+ dev_err(dev, "failed to request irq\n");
+ goto err_deinit_media_frameworks;
+ }
+
+ if (device_property_read_bool(dev, "wakeup-source")) {
+ ret = devm_device_init_wakeup(dev);
+ if (ret) {
+ ret = dev_err_probe(dev, ret,
+ "failed to initialize wakeup\n");
+ goto err_free_irq;
+ }
+ }
+
+ pm_runtime_put_autosuspend(dev);
+
return 0;
+err_free_irq:
+ free_irq(mali_c55->irqnum, dev);
err_deinit_media_frameworks:
mali_c55_media_frameworks_deinit(mali_c55);
err_pm_runtime_disable:
- pm_runtime_set_suspended(&pdev->dev);
pm_runtime_disable(&pdev->dev);
+ pm_runtime_put_noidle(dev);
+ pm_runtime_set_suspended(&pdev->dev);
kfree(mali_c55->context.registers);
err_power_off:
__mali_c55_power_off(mali_c55);
@@ -859,12 +902,13 @@ static void mali_c55_remove(struct platform_device *pdev)
{
struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
+ pm_runtime_disable(&pdev->dev);
+ free_irq(mali_c55->irqnum, &pdev->dev);
mali_c55_media_frameworks_deinit(mali_c55);
- if (!pm_runtime_suspended(&pdev->dev)) {
+ if (!pm_runtime_status_suspended(&pdev->dev)) {
__mali_c55_power_off(mali_c55);
pm_runtime_set_suspended(&pdev->dev);
}
- pm_runtime_disable(&pdev->dev);
kfree(mali_c55->context.registers);
of_reserved_mem_device_release(&pdev->dev);
}
--
2.55.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 3/3] media: mali-c55: Keep ISP powered while IRQ wake is armed
2026-09-29 12:02 [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Linus Walleij
2026-09-29 12:02 ` [PATCH 1/3] dt-bindings: media: mali-c55: Allow wakeup source Linus Walleij
2026-09-29 12:02 ` [PATCH 2/3] media: mali-c55: Keep IRQ requested during suspend Linus Walleij
@ 2026-09-29 12:02 ` Linus Walleij
2026-10-06 11:43 ` [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Lad, Prabhakar
3 siblings, 0 replies; 5+ messages in thread
From: Linus Walleij @ 2026-09-29 12:02 UTC (permalink / raw)
To: Li Youhong, Daniel Scally, Jacopo Mondi, Mauro Carvalho Chehab,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Hans Verkuil,
Nayden Kanchev
Cc: linux-media, devicetree, linux-kernel, Linus Walleij
System suspend enables IRQ wake and then calls pm_runtime_force_suspend().
For an active ISP this asserts its resets and disables its clocks, so it
can no longer generate frame interrupts. An already runtime-suspended ISP
is left powered off as well.
Take a runtime PM reference before arming IRQ wake and retain it until
resume. This powers up an idle ISP and keeps an active ISP's configuration
intact. Release the reference if arming wake fails, and use the PM wake
policy in both suspend and resume to balance the reference. Continue to
use forced runtime suspend and resume when wake is disabled.
Use the forced power-down path for hibernation freeze and power-off even
when wake is enabled. Capture must not remain active while a hibernation
image is created.
Wake events require a configured stream and an input source that
remains operational during system sleep.
The internal TPG shares the ISP's power state. External sensor and
receiver drivers will need their own system-sleep wake support.
Assisted-by: LLM
Signed-off-by: Linus Walleij <linusw@kernel.org>
---
.../media/platform/arm/mali-c55/mali-c55-core.c | 38 ++++++++++++++--------
1 file changed, 25 insertions(+), 13 deletions(-)
diff --git a/drivers/media/platform/arm/mali-c55/mali-c55-core.c b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
index 07267b79801b..2629bb3aa628 100644
--- a/drivers/media/platform/arm/mali-c55/mali-c55-core.c
+++ b/drivers/media/platform/arm/mali-c55/mali-c55-core.c
@@ -756,31 +756,43 @@ static int __maybe_unused mali_c55_suspend(struct device *dev)
struct mali_c55 *mali_c55 = dev_get_drvdata(dev);
int ret;
- if (device_may_wakeup(dev)) {
- ret = enable_irq_wake(mali_c55->irqnum);
- if (ret)
- return ret;
- }
+ if (!device_may_wakeup(dev))
+ return pm_runtime_force_suspend(dev);
- ret = pm_runtime_force_suspend(dev);
- if (ret && device_may_wakeup(dev))
- disable_irq_wake(mali_c55->irqnum);
+ /* Keep the ISP clocked and out of reset while it can wake the system. */
+ ret = pm_runtime_resume_and_get(dev);
+ if (ret)
+ return ret;
- return ret;
+ ret = enable_irq_wake(mali_c55->irqnum);
+ if (ret) {
+ pm_runtime_put_autosuspend(dev);
+ return ret;
+ }
+
+ return 0;
}
static int __maybe_unused mali_c55_resume(struct device *dev)
{
struct mali_c55 *mali_c55 = dev_get_drvdata(dev);
- if (device_may_wakeup(dev))
- disable_irq_wake(mali_c55->irqnum);
+ if (!device_may_wakeup(dev))
+ return pm_runtime_force_resume(dev);
- return pm_runtime_force_resume(dev);
+ disable_irq_wake(mali_c55->irqnum);
+ pm_runtime_put_autosuspend(dev);
+
+ return 0;
}
static const struct dev_pm_ops mali_c55_pm_ops = {
- SET_SYSTEM_SLEEP_PM_OPS(mali_c55_suspend, mali_c55_resume)
+ .suspend = pm_sleep_ptr(mali_c55_suspend),
+ .resume = pm_sleep_ptr(mali_c55_resume),
+ .freeze = pm_sleep_ptr(pm_runtime_force_suspend),
+ .thaw = pm_sleep_ptr(pm_runtime_force_resume),
+ .poweroff = pm_sleep_ptr(pm_runtime_force_suspend),
+ .restore = pm_sleep_ptr(pm_runtime_force_resume),
SET_RUNTIME_PM_OPS(mali_c55_runtime_suspend, mali_c55_runtime_resume,
NULL)
};
--
2.55.0
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling
2026-09-29 12:02 [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Linus Walleij
` (2 preceding siblings ...)
2026-09-29 12:02 ` [PATCH 3/3] media: mali-c55: Keep ISP powered while IRQ wake is armed Linus Walleij
@ 2026-10-06 11:43 ` Lad, Prabhakar
3 siblings, 0 replies; 5+ messages in thread
From: Lad, Prabhakar @ 2026-10-06 11:43 UTC (permalink / raw)
To: Linus Walleij
Cc: Li Youhong, Daniel Scally, Jacopo Mondi, Mauro Carvalho Chehab,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Hans Verkuil,
Nayden Kanchev, linux-media, devicetree, linux-kernel, stable
Hi Linus,
On Tue, Sep 29, 2026 at 1:08 PM Linus Walleij <linusw@kernel.org> wrote:
>
> This is essentially my reply with the "proper fix" for:
> https://lore.kernel.org/linux-media/20260903025526.2657105-1-dayou5941@163.com/
>
> Keep the Mali-C55 interrupt requested for the driver's lifetime instead
> of changing interrupt ownership along with the runtime power state.
> Synchronize and disable the IRQ before power-off, and enable it after
> power-on. Hold a PM reference until probe installs the IRQ, and stop
> runtime PM and drain the IRQ before removing the media entities.
>
> Allow platforms to mark the ISP as a wakeup source. Hold a runtime PM
> reference while IRQ wake is armed, resuming an idle ISP if needed and
> preserving an active ISP's configuration. Balance the reference on resume
> and failure. Hibernation still uses the power-down path.
>
> Wake events require a configured stream and an input that remains active
> during system sleep. The internal TPG shares the ISP's power state;
> external sensors and receivers need their own wake support. An idle
> pipeline is not started by this series. Hardware wake remains untested.
>
> The first patch extends the binding with the standard wakeup-source
> property. The second patch updates the driver IRQ and PM handling. The
> third patch retains ISP power while IRQ wake is armed.
>
> Tested on the Renesas RZ/V2H board with the other in-flight patches
> (DTS changes and the TPG enablement) doing full suspend/resume cycles
> with wakeup from RTC and NMI and capturing images before and
> after the test.
>
Can you please share the dependency patches or branch, as these don't
apply cleanly. I tried the latest 'next' and 'media/next' branches
with no luck.
Cheers,
Prabhakar
> Assisted-by: LLM
> Signed-off-by: Linus Walleij <linusw@kernel.org>
> ---
> Linus Walleij (3):
> dt-bindings: media: mali-c55: Allow wakeup source
> media: mali-c55: Keep IRQ requested during suspend
> media: mali-c55: Keep ISP powered while IRQ wake is armed
>
> .../devicetree/bindings/media/arm,mali-c55.yaml | 2 +
> .../media/platform/arm/mali-c55/mali-c55-core.c | 96 +++++++++++++++++-----
> 2 files changed, 78 insertions(+), 20 deletions(-)
> ---
> base-commit: cee9395acd8043be0644b25c34bfa86623f2b935
> change-id: 20260918-mali-c55-irq-supend-resume-30b507a20011
>
> Best regards,
> --
> Linus Walleij <linusw@kernel.org>
>
>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2026-10-06 11:43 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-09-29 12:02 [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Linus Walleij
2026-09-29 12:02 ` [PATCH 1/3] dt-bindings: media: mali-c55: Allow wakeup source Linus Walleij
2026-09-29 12:02 ` [PATCH 2/3] media: mali-c55: Keep IRQ requested during suspend Linus Walleij
2026-09-29 12:02 ` [PATCH 3/3] media: mali-c55: Keep ISP powered while IRQ wake is armed Linus Walleij
2026-10-06 11:43 ` [PATCH 0/3] media: mali-c55: Fix IRQ lifetime and wake handling Lad, Prabhakar
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®