[PATCH] drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace

[PATCH] drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace

[Varun Gupta]

Reference-drop trace in drm_pagemap_dev_unhold_work() was changed
from drm_dbg() to drm_dbg_ratelimited() to reduce serial console
bottlenecks.
However, when the ratelimit kicks in, the kernel's ratelimiting subsystem
logs a "callbacks suppressed" notification at the KERN_WARNING level.
Revert the trace macro back to drm_dbg()

Fixes: b66b036c959f ("drm/pagemap: Prevent CPU stalls during unbounded pagemap teardown")
Signed-off-by: Varun Gupta <varun.gupta@intel.com>
---
 drivers/gpu/drm/drm_pagemap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/drm_pagemap.c b/drivers/gpu/drm/drm_pagemap.c
index 0867bfbde5c5..ba97daedcd49 100644
--- a/drivers/gpu/drm/drm_pagemap.c
+++ b/drivers/gpu/drm/drm_pagemap.c
@@ -1009,7 +1009,7 @@ static void drm_pagemap_dev_unhold_work(struct work_struct *work)
 		struct drm_device *drm = dev_hold->drm;
 		struct module *module = drm->driver->fops->owner;
 
-		drm_dbg_ratelimited(drm, "Releasing reference on provider device and module.\n");
+		drm_dbg(drm, "Releasing reference on provider device and module.\n");
 		drm_dev_put(drm);
 		module_put(module);
 		kfree(dev_hold);
-- 
2.43.0

Re: [PATCH] drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] drm/pagemap: CPU stall re-introduced by unbounded drm_dbg() loop
--

commit ea1ff5e58f9772cc1eaec3db21133d24e24bd4f0
Author: Varun Gupta <varun.gupta@intel.com>

drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace

This commit reverts the use of drm_dbg_ratelimited back to drm_dbg in
drm_pagemap_dev_unhold_work to avoid "callbacks suppressed" notifications
from the kernel's ratelimiting subsystem.

> diff --git a/drivers/gpu/drm/drm_pagemap.c b/drivers/gpu/drm/drm_pagemap.c
> index c66c58f85c345..743988036a5f8 100644
> --- a/drivers/gpu/drm/drm_pagemap.c
> +++ b/drivers/gpu/drm/drm_pagemap.c
> @@ -1102,7 +1102,7 @@ static void drm_pagemap_dev_unhold_work(struct work_struct *work)
>  		struct drm_device *drm = dev_hold->drm;
>  		struct module *module = drm->driver->fops->owner;
>  
> -		drm_dbg_ratelimited(drm, "Releasing reference on provider device and module.\n");
> +		drm_dbg(drm, "Releasing reference on provider device and module.\n");

[Severity: Medium]
Does this code re-introduce a serial console bottleneck and potential CPU stall
during teardown?

When tearing down an unbounded number of DRM pagemaps,
drm_pagemap_dev_unhold_work() iterates over the entire
drm_pagemap_unhold_list. If DRM debugging is enabled, drm_dbg() is now
called for every single entry unconditionally.

If the goal is to prevent the "callbacks suppressed" warning from the
ratelimiting subsystem, would it be better to drop the repetitive log
entirely, or perhaps log once outside the loop, rather than reverting
the CPU stall fix?

>  		drm_dev_put(drm);
>  		module_put(module);
>  		kfree(dev_hold);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260921062648.1166903-2-varun.gupta@intel.com?part=1

Re: [PATCH] drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace

[Gupta, Varun]

On 21-Sep-26 12:03 PM, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] drm/pagemap: CPU stall re-introduced by unbounded drm_dbg() loop
> --
>
> commit ea1ff5e58f9772cc1eaec3db21133d24e24bd4f0
> Author: Varun Gupta <varun.gupta@intel.com>
>
> drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace
>
> This commit reverts the use of drm_dbg_ratelimited back to drm_dbg in
> drm_pagemap_dev_unhold_work to avoid "callbacks suppressed" notifications
> from the kernel's ratelimiting subsystem.
>
>> diff --git a/drivers/gpu/drm/drm_pagemap.c b/drivers/gpu/drm/drm_pagemap.c
>> index c66c58f85c345..743988036a5f8 100644
>> --- a/drivers/gpu/drm/drm_pagemap.c
>> +++ b/drivers/gpu/drm/drm_pagemap.c
>> @@ -1102,7 +1102,7 @@ static void drm_pagemap_dev_unhold_work(struct work_struct *work)
>>   		struct drm_device *drm = dev_hold->drm;
>>   		struct module *module = drm->driver->fops->owner;
>>   
>> -		drm_dbg_ratelimited(drm, "Releasing reference on provider device and module.\n");
>> +		drm_dbg(drm, "Releasing reference on provider device and module.\n");
> [Severity: Medium]
> Does this code re-introduce a serial console bottleneck and potential CPU stall
> during teardown?
This does not introduce a stall, it is handled by the WQ_UNBOUND change. 
This log was ratelimited to
just prevent log flooding but it in turn introduces a suppressed warning.
> When tearing down an unbounded number of DRM pagemaps,
> drm_pagemap_dev_unhold_work() iterates over the entire
> drm_pagemap_unhold_list. If DRM debugging is enabled, drm_dbg() is now
> called for every single entry unconditionally.
>
> If the goal is to prevent the "callbacks suppressed" warning from the
> ratelimiting subsystem, would it be better to drop the repetitive log
> entirely, or perhaps log once outside the loop, rather than reverting
> the CPU stall fix?
>
>>   		drm_dev_put(drm);
>>   		module_put(module);
>>   		kfree(dev_hold);

Re: [PATCH] drm/pagemap: Drop ratelimiting from dev_unhold_work debug trace

[Thomas Hellström]

On Mon, 2026-09-21 at 11:56 +0530, Varun Gupta wrote:
> Reference-drop trace in drm_pagemap_dev_unhold_work() was changed
> from drm_dbg() to drm_dbg_ratelimited() to reduce serial console
> bottlenecks.
> However, when the ratelimit kicks in, the kernel's ratelimiting
> subsystem
> logs a "callbacks suppressed" notification at the KERN_WARNING level.
> Revert the trace macro back to drm_dbg()
> 
> Fixes: b66b036c959f ("drm/pagemap: Prevent CPU stalls during
> unbounded pagemap teardown")
> Signed-off-by: Varun Gupta <varun.gupta@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>

> ---
>  drivers/gpu/drm/drm_pagemap.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/drm_pagemap.c
> b/drivers/gpu/drm/drm_pagemap.c
> index 0867bfbde5c5..ba97daedcd49 100644
> --- a/drivers/gpu/drm/drm_pagemap.c
> +++ b/drivers/gpu/drm/drm_pagemap.c
> @@ -1009,7 +1009,7 @@ static void drm_pagemap_dev_unhold_work(struct
> work_struct *work)
>  		struct drm_device *drm = dev_hold->drm;
>  		struct module *module = drm->driver->fops->owner;
>  
> -		drm_dbg_ratelimited(drm, "Releasing reference on
> provider device and module.\n");
> +		drm_dbg(drm, "Releasing reference on provider device
> and module.\n");
>  		drm_dev_put(drm);
>  		module_put(module);
>  		kfree(dev_hold);