Linux kernel and device drivers for NXP i.MX platforms
 help / color / mirror / Atom feed
* [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() for !CONFIG_PM
@ 2026-11-16 15:25 Frank.Li
  2026-11-16 15:25 ` [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() " Frank.Li
  2026-11-17 15:25 ` [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() " sashiko-bot
  0 siblings, 2 replies; 7+ messages in thread
From: Frank.Li @ 2026-11-16 15:25 UTC (permalink / raw)
  Cc: Frank Li, Alex Elder, Rafael J. Wysocki, Greg Kroah-Hartman,
	Andrzej Kacprowski, Karol Wachowski, Oded Gabbay, Jani Nikula,
	Joonas Lahtinen, Rodrigo Vivi, Tvrtko Ursulin, Matthew Brost,
	Thomas Hellstrom, David Airlie, Simona Vetter, Alex Elder,
	Will Deacon, Robin Murphy, Joerg Roedel, Rob Clark, Sakari Ailus,
	Mauro Carvalho Chehab, Guenter Roeck, Guan-Yu Lin, Mathias Nyman,
	linux-pm, linux-usb, dri-devel, intel-gfx, intel-xe, linux-hwmon,
	linux-arm-kernel, iommu, linux-arm-msm, linux-media,
	linux-staging, netdev, imx

From: Frank Li <Frank.Li@nxp.com>

When CONFIG_PM is not enabled, should be treated as if runtime PM were
enabled for the device and the runtime status is active.

It's a subtle distinction but the previous behavior returned -EINVAL, which
means "runtime PM is disabled for the device"

__pm_runtime_resume() returns 1 (called by pm_runtime_get()).
pm_runtime_get_if_active() is the try-and-get version of pm_runtime_get().
stub pm_runtime_active() also returns true. So pm_runtime_get_if_active()
should return the same value as the stub pm_runtime_get(), which is 1.

Analysis of all callers and their behavior with old (-EINVAL) vs new (1):

Pattern A: if (!pm_runtime_get_if_active(dev)) -- no impact
  Both -EINVAL and 1 are non-zero, so !ret is false in both cases.
  Callers: drivers/accel/ethosu/ethosu_perfmon.c,
           drivers/gpu/drm/msm/adreno/a6xx_gmu.c,
           drivers/gpu/drm/msm/msm_perfcntr.c,
           drivers/gpu/drm/v3d/v3d_mmu.c,
           drivers/gpu/drm/v3d/v3d_perfmon.c (x4),
           drivers/media/i2c/gc08a3.c gc05a2.c og0ve1b.c ov6211.c,
           drivers/media/i2c/ccs/ccs-core.c s5kjn1.c lm3560.c (x2),
           drivers/media/i2c/s5k3m5.c thp7312.c imx283.c ov64a40.c,
           drivers/media/pci/intel/ipu6/ipu6-buttress.c,
           sound/hda/core/device.c

Pattern B: if (!ret) after get -- no impact
  Both -EINVAL and 1 are non-zero, !ret is false in both cases.
  drivers/gpu/drm/amd/pm/amdgpu_pm.c (comment says "allow access"
    in kernels without CONFIG_PM), drivers/pci/pci.c

Pattern C: ret > 0 / ret <= 0 / ret != 1 checks -- FIXED
  drivers/usb/core/offload.c: checks (!= 1), old -EINVAL caused
    -EBUSY even with !CONFIG_PM. Fixed: now returns 1, proceeds.
  drivers/accel/ivpu/ivpu_job.c: drm_WARN_ON(<= 0), old -EINVAL
    spuriously triggered WARN. Fixed: 1 > 0, no warn.
  drivers/gpu/drm/i915/intel_runtime_pm.c: checks (<= 0), old
    -EINVAL triggered error path. Fixed: 1 > 0, correct path.
  drivers/accel/ivpu/ivpu_sysfs.c: (> 0), old skipped freq read.
    Fixed: proceeds correctly without CONFIG_PM.
  drivers/gpu/drm/xe/xe_pm.c: wraps (> 0), all xe callers
    (xe_bo.c, xe_shrinker.c, xe_vsec.c, xe_ggtt.c, xe_guc_ct.c,
    xe_pxp.c) now correctly treat !CONFIG_PM as active.
  drivers/net/ipa/ipa_smp2p.c: power_on = (> 0), now correctly true.
  drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c: (> 0), now correctly
    executes stall register update.
  drivers/staging/media/ipu7/ipu7-buttress.c: active (<= 0) early
    return now skipped, correctly proceeds.

Pattern D: explicit -EINVAL check
  drivers/hwmon/prom21-xhci.c: checks (> 0) then (== -EINVAL).
    Old: took -EINVAL fallback. New: takes (> 0) fast path.
    Both succeed when CONFIG_PM is disabled; functionally equivalent.
  drivers/hwmon/eic7700-pvt.c: call is guarded by IS_ENABLED(CONFIG_PM),
    stub is never reached. No impact.

Fixes: bd3d245b0fef5 ("usb: core: use dedicated spinlock for offload state")
Fixes: 7806bad76ac39 ("accel/ivpu: Prevent runtime suspend during context abort work")
Fixes: 9d58aa46291d4 ("drm/i915: Fix the GT fence revocation runtime PM logic")
Fixes: 1524c28b99527 ("accel/ivpu: Show NPU frequency in sysfs")
Fixes: 30c399529f4c6 ("drm/xe: Document Xe PM component")
Fixes: 0d08026ac6099 ("net: ipa: kill ipa_clock_get_additional()")
Fixes: 70892277ca2db ("iommu/arm-smmu-qcom: Make set_stall work when the device is on")
Fixes: 9ad65684b9285 ("media: staging/ipu7: Ignore interrupts when device is suspended")
Acked-by: Alex Elder <elder@riscstar.com>
Signed-off-by: Frank Li <Frank.Li@nxp.com>
---
Change in v2
- add fixes tag for patten C.
- add Alex Elder ack tags
- change subject to fix stub pm_runtime_get_if_active()

Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Andrzej Kacprowski <andrzej.kacprowski@linux.intel.com>
Cc: Karol Wachowski <karol.wachowski@linux.intel.com>
Cc: Oded Gabbay <ogabbay@kernel.org>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Tvrtko Ursulin <tursulin@ursulin.net>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: "Thomas Hellstrom" <thomas.hellstrom@linux.intel.com>
Cc: David Airlie <airlied@gmail.com>
Cc: Simona Vetter <simona@ffwll.ch>
Cc: Alex Elder <elder@kernel.org>
Cc: Will Deacon <will@kernel.org>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: "Joerg Roedel" <joro@8bytes.org>
Cc: Rob Clark <robin.clark@oss.qualcomm.com>
Cc: Sakari Ailus <sakari.ailus@linux.intel.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Guan-Yu Lin <guanyulin@google.com>
Cc: Mathias Nyman <mathias.nyman@linux.intel.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-usb@vger.kernel.org
Cc: dri-devel@lists.freedesktop.org
Cc: intel-gfx@lists.freedesktop.org
Cc: intel-xe@lists.freedesktop.org
Cc: linux-hwmon@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: iommu@lists.linux.dev
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-media@vger.kernel.org
Cc: linux-staging@lists.linux.dev
Cc: netdev@vger.kernel.org
Cc: imx@lists.linux.dev
---
 include/linux/pm_runtime.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
index 64921b10ac74f..de9e7518ecef2 100644
--- a/include/linux/pm_runtime.h
+++ b/include/linux/pm_runtime.h
@@ -280,7 +280,7 @@ static inline int pm_runtime_get_if_in_use(struct device *dev)
 }
 static inline int pm_runtime_get_if_active(struct device *dev)
 {
-	return -EINVAL;
+	return 1;
 }
 static inline int __pm_runtime_set_status(struct device *dev,
 					    unsigned int status) { return 0; }
--
2.43.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() for !CONFIG_PM
  2026-11-16 15:25 [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() for !CONFIG_PM Frank.Li
@ 2026-11-16 15:25 ` Frank.Li
  2026-11-16 15:31   ` Laurent Pinchart
  2026-11-17 15:25   ` sashiko-bot
  2026-11-17 15:25 ` [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() " sashiko-bot
  1 sibling, 2 replies; 7+ messages in thread
From: Frank.Li @ 2026-11-16 15:25 UTC (permalink / raw)
  Cc: Frank Li, Rafael J. Wysocki, linux-pm, Jani Nikula,
	Joonas Lahtinen, Rodrigo Vivi, Tvrtko Ursulin, Maarten Lankhorst,
	Maxime Ripard, Thomas Hellstrom, Simona Vetter, David Airlie,
	Matthew Brost, Thomas Zimmermann, Patrik Jakobsson, Alessio Belle,
	Luigi Santivetti, Liam Girdwood, Mark Brown, Jaroslav Kysela,
	Takashi Iwai, Bard Liao, Pierre-Louis Bossart, Peter Ujfalusi,
	Kai Vehmanen, Robin Murphy, Yong Wu, AngeloGioacchino Del Regno,
	Matthias Brugger, Heiko Stuebner, Mauro Carvalho Chehab,
	Dafna Hirschfeld, Laurent Pinchart, Michael Tretter,
	Daniel Baluta, Will Deacon, Vijendar Mukunda, dri-devel,
	imagination, intel-gfx, intel-xe, iommu, linux-arm-kernel,
	linux-mediatek, linux-media, linux-rockchip, linux-sound,
	sound-open-firmware, imx

From: Frank Li <Frank.Li@nxp.com>

The stub for pm_runtime_get_if_in_use() under !CONFIG_PM currently
returns -EINVAL.  The real implementation returns:
  1  - device is active and its usage counter is non-zero (bumped)
  0  - device is not in use or not active (counter not bumped)
 -EINVAL - runtime PM disabled for this device

When CONFIG_PM is not set every device is permanently "active" and
always considered "in use", so the stub must return 1, mirroring
__pm_runtime_resume() which already returns 1 in the same build.

Returning -EINVAL breaks callers that test the return value with
comparisons other than "!= 0 / == 0":

drivers/gpu/drm/i915/intel_runtime_pm.c:
  if (intel_runtime_pm_get_if_in_use(rpm))
  which wraps to: ret = pm_runtime_get_if_in_use(); return ret > 0;
  With -EINVAL the wrapper returns false, causing an incorrect error
  path in the GT fence revocation code.

drivers/gpu/drm/xe/xe_pm.c:
  if (pm_runtime_get_if_in_use(xe->drm.dev) > 0)
  With -EINVAL the condition is false; xe_pm_runtime_get_if_in_use()
  returns false, so callers believe the device is not usable.

drivers/gpu/drm/gma500/power.c:
  return pm_runtime_get_if_in_use(dev) == 1;
  -EINVAL != 1, so the helper always returns false when !CONFIG_PM.

drivers/gpu/drm/imagination/pvr_power.c:
  if (pm_runtime_get_if_in_use(...) <= 0)
  -EINVAL <= 0 is true, so the driver always takes the "requeue" error
  path instead of proceeding with the active device.

sound/soc/sof/debug.c:
  if (pm_runtime_get_if_in_use(sdev->dev) == 1)
  -EINVAL != 1, so d3_prevented is never set, corrupting exception
  debug accounting.

drivers/iommu/mtk_iommu.c:
  if (pm_runtime_get_if_in_use(dom->plat_dev->dev) <= 0)
  -EINVAL <= 0 skips the TLB flush, leaving stale IOMMU mappings.

drivers/iommu/rockchip-iommu.c:
  Three call sites use WARN_ON_ONCE(err < 0) or similar patterns.
  -EINVAL < 0 triggers a spurious WARNING on every call.

drivers/media/i2c/isl7998x.c:
  if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
  -EINVAL <= 0 causes the driver to always skip register programming
  when !CONFIG_PM.

drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c:
  if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
  -EINVAL <= 0 prevents any debug statistics from being collected.

Fixes: 9d58aa46291d ("drm/i915: Fix the GT fence revocation runtime PM logic")
Fixes: cbb6a7413b17 ("drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers")
Fixes: 672c473576ca ("drm/gma500: Rewrite power management code")
Fixes: 727538a4bbff ("drm/imagination: Implement power management")
Fixes: b30289e7fa92 ("ASoC: SOF: Fix runtime pm usage counter balance after fw exception")
Fixes: c0b57581b73b ("iommu/mediatek: Add power-domain operation")
Fixes: 0f181d3cf7d9 ("iommu/rockchip: Add runtime PM support")
Fixes: 51ef2be546e2 ("media: i2c: isl7998x: Add driver for Intersil ISL7998x")
Fixes: cbb819c363c5 ("media: rkisp1: debug: Collect input status by sampling ISP_FLAGS_SHD")
Signed-off-by: Frank Li <Frank.Li@nxp.com>
---
Change in v2
- new patches.

Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: linux-pm@vger.kernel.org
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Tvrtko Ursulin <tursulin@ursulin.net>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: "Thomas Hellstrom" <thomas.hellstrom@linux.intel.com>
Cc: Simona Vetter <simona@ffwll.ch>
Cc: David Airlie <airlied@gmail.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Jaroslav Kysela <perex@perex.cz>
Cc: Takashi Iwai <tiwai@suse.com>
Cc: Bard Liao <yung-chuan.liao@linux.intel.com>
Cc: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
Cc: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Yong Wu <yong.wu@mediatek.com>
Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Cc: Matthias Brugger <matthias.bgg@gmail.com>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Dafna Hirschfeld <dafna@fastmail.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Michael Tretter <m.tretter@pengutronix.de>
Cc: Daniel Baluta <daniel.baluta@nxp.com>
Cc: Will Deacon <will@kernel.org>
Cc: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Cc: dri-devel@lists.freedesktop.org
Cc: imagination@lists.freedesktop.org
Cc: intel-gfx@lists.freedesktop.org
Cc: intel-xe@lists.freedesktop.org
Cc: iommu@lists.linux.dev
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-mediatek@lists.infradead.org
Cc: linux-media@vger.kernel.org
Cc: linux-rockchip@lists.infradead.org
Cc: linux-sound@vger.kernel.org
Cc: sound-open-firmware@alsa-project.org
Cc: imx@lists.linux.dev
---
 include/linux/pm_runtime.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
index de9e7518ecef2..1d983fe55a16a 100644
--- a/include/linux/pm_runtime.h
+++ b/include/linux/pm_runtime.h
@@ -276,7 +276,7 @@ static inline int pm_schedule_suspend(struct device *dev, unsigned int delay)
 }
 static inline int pm_runtime_get_if_in_use(struct device *dev)
 {
-	return -EINVAL;
+	return 1;
 }
 static inline int pm_runtime_get_if_active(struct device *dev)
 {
-- 
2.43.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() for !CONFIG_PM
  2026-11-16 15:25 ` [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() " Frank.Li
@ 2026-11-16 15:31   ` Laurent Pinchart
  2026-11-16 15:43     ` Frank Li
  2026-11-17 15:25   ` sashiko-bot
  1 sibling, 1 reply; 7+ messages in thread
From: Laurent Pinchart @ 2026-11-16 15:31 UTC (permalink / raw)
  To: Frank.Li
  Cc: Frank Li, Rafael J. Wysocki, linux-pm, Jani Nikula,
	Joonas Lahtinen, Rodrigo Vivi, Tvrtko Ursulin, Maarten Lankhorst,
	Maxime Ripard, Thomas Hellstrom, Simona Vetter, David Airlie,
	Matthew Brost, Thomas Zimmermann, Patrik Jakobsson, Alessio Belle,
	Luigi Santivetti, Liam Girdwood, Mark Brown, Jaroslav Kysela,
	Takashi Iwai, Bard Liao, Pierre-Louis Bossart, Peter Ujfalusi,
	Kai Vehmanen, Robin Murphy, Yong Wu, AngeloGioacchino Del Regno,
	Matthias Brugger, Heiko Stuebner, Mauro Carvalho Chehab,
	Dafna Hirschfeld, Michael Tretter, Daniel Baluta, Will Deacon,
	Vijendar Mukunda, dri-devel, imagination, intel-gfx, intel-xe,
	iommu, linux-arm-kernel, linux-mediatek, linux-media,
	linux-rockchip, linux-sound, sound-open-firmware, imx

Hi Frank,

On Mon, Nov 16, 2026 at 10:25:11AM -0500, Frank.Li@oss.nxp.com wrote:
> From: Frank Li <Frank.Li@nxp.com>
> 
> The stub for pm_runtime_get_if_in_use() under !CONFIG_PM currently
> returns -EINVAL.  The real implementation returns:
>   1  - device is active and its usage counter is non-zero (bumped)
>   0  - device is not in use or not active (counter not bumped)
>  -EINVAL - runtime PM disabled for this device
> 
> When CONFIG_PM is not set every device is permanently "active" and
> always considered "in use", so the stub must return 1, mirroring
> __pm_runtime_resume() which already returns 1 in the same build.
> 
> Returning -EINVAL breaks callers that test the return value with
> comparisons other than "!= 0 / == 0":
> 
> drivers/gpu/drm/i915/intel_runtime_pm.c:
>   if (intel_runtime_pm_get_if_in_use(rpm))
>   which wraps to: ret = pm_runtime_get_if_in_use(); return ret > 0;
>   With -EINVAL the wrapper returns false, causing an incorrect error
>   path in the GT fence revocation code.
> 
> drivers/gpu/drm/xe/xe_pm.c:
>   if (pm_runtime_get_if_in_use(xe->drm.dev) > 0)
>   With -EINVAL the condition is false; xe_pm_runtime_get_if_in_use()
>   returns false, so callers believe the device is not usable.
> 
> drivers/gpu/drm/gma500/power.c:
>   return pm_runtime_get_if_in_use(dev) == 1;
>   -EINVAL != 1, so the helper always returns false when !CONFIG_PM.
> 
> drivers/gpu/drm/imagination/pvr_power.c:
>   if (pm_runtime_get_if_in_use(...) <= 0)
>   -EINVAL <= 0 is true, so the driver always takes the "requeue" error
>   path instead of proceeding with the active device.
> 
> sound/soc/sof/debug.c:
>   if (pm_runtime_get_if_in_use(sdev->dev) == 1)
>   -EINVAL != 1, so d3_prevented is never set, corrupting exception
>   debug accounting.
> 
> drivers/iommu/mtk_iommu.c:
>   if (pm_runtime_get_if_in_use(dom->plat_dev->dev) <= 0)
>   -EINVAL <= 0 skips the TLB flush, leaving stale IOMMU mappings.
> 
> drivers/iommu/rockchip-iommu.c:
>   Three call sites use WARN_ON_ONCE(err < 0) or similar patterns.
>   -EINVAL < 0 triggers a spurious WARNING on every call.
> 
> drivers/media/i2c/isl7998x.c:
>   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
>   -EINVAL <= 0 causes the driver to always skip register programming
>   when !CONFIG_PM.
> 
> drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c:
>   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
>   -EINVAL <= 0 prevents any debug statistics from being collected.
> 
> Fixes: 9d58aa46291d ("drm/i915: Fix the GT fence revocation runtime PM logic")
> Fixes: cbb6a7413b17 ("drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers")
> Fixes: 672c473576ca ("drm/gma500: Rewrite power management code")
> Fixes: 727538a4bbff ("drm/imagination: Implement power management")
> Fixes: b30289e7fa92 ("ASoC: SOF: Fix runtime pm usage counter balance after fw exception")
> Fixes: c0b57581b73b ("iommu/mediatek: Add power-domain operation")
> Fixes: 0f181d3cf7d9 ("iommu/rockchip: Add runtime PM support")
> Fixes: 51ef2be546e2 ("media: i2c: isl7998x: Add driver for Intersil ISL7998x")
> Fixes: cbb819c363c5 ("media: rkisp1: debug: Collect input status by sampling ISP_FLAGS_SHD")

I don't think this one is right. The rkisp1 driver only runs on ARCH_MXC
and ARCH_ROCKCHIP, which select CONFIG_PM. The same may be true for some
of the other fixes tags.

> Signed-off-by: Frank Li <Frank.Li@nxp.com>
> ---
> Change in v2
> - new patches.
> 
> Cc: "Rafael J. Wysocki" <rafael@kernel.org>
> Cc: linux-pm@vger.kernel.org
> Cc: Jani Nikula <jani.nikula@linux.intel.com>
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> Cc: Tvrtko Ursulin <tursulin@ursulin.net>
> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> Cc: Maxime Ripard <mripard@kernel.org>
> Cc: "Thomas Hellstrom" <thomas.hellstrom@linux.intel.com>
> Cc: Simona Vetter <simona@ffwll.ch>
> Cc: David Airlie <airlied@gmail.com>
> Cc: Matthew Brost <matthew.brost@intel.com>
> Cc: Thomas Zimmermann <tzimmermann@suse.de>
> Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
> Cc: Alessio Belle <alessio.belle@imgtec.com>
> Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
> Cc: Liam Girdwood <lgirdwood@gmail.com>
> Cc: Mark Brown <broonie@kernel.org>
> Cc: Jaroslav Kysela <perex@perex.cz>
> Cc: Takashi Iwai <tiwai@suse.com>
> Cc: Bard Liao <yung-chuan.liao@linux.intel.com>
> Cc: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
> Cc: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
> Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
> Cc: Robin Murphy <robin.murphy@arm.com>
> Cc: Yong Wu <yong.wu@mediatek.com>
> Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> Cc: Matthias Brugger <matthias.bgg@gmail.com>
> Cc: Heiko Stuebner <heiko@sntech.de>
> Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
> Cc: Dafna Hirschfeld <dafna@fastmail.com>
> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> Cc: Michael Tretter <m.tretter@pengutronix.de>
> Cc: Daniel Baluta <daniel.baluta@nxp.com>
> Cc: Will Deacon <will@kernel.org>
> Cc: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
> Cc: dri-devel@lists.freedesktop.org
> Cc: imagination@lists.freedesktop.org
> Cc: intel-gfx@lists.freedesktop.org
> Cc: intel-xe@lists.freedesktop.org
> Cc: iommu@lists.linux.dev
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-mediatek@lists.infradead.org
> Cc: linux-media@vger.kernel.org
> Cc: linux-rockchip@lists.infradead.org
> Cc: linux-sound@vger.kernel.org
> Cc: sound-open-firmware@alsa-project.org
> Cc: imx@lists.linux.dev
> ---
>  include/linux/pm_runtime.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
> index de9e7518ecef2..1d983fe55a16a 100644
> --- a/include/linux/pm_runtime.h
> +++ b/include/linux/pm_runtime.h
> @@ -276,7 +276,7 @@ static inline int pm_schedule_suspend(struct device *dev, unsigned int delay)
>  }
>  static inline int pm_runtime_get_if_in_use(struct device *dev)
>  {
> -	return -EINVAL;
> +	return 1;
>  }
>  static inline int pm_runtime_get_if_active(struct device *dev)
>  {

-- 
Regards,

Laurent Pinchart

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() for !CONFIG_PM
  2026-11-16 15:31   ` Laurent Pinchart
@ 2026-11-16 15:43     ` Frank Li
  2026-11-16 15:52       ` Laurent Pinchart
  0 siblings, 1 reply; 7+ messages in thread
From: Frank Li @ 2026-11-16 15:43 UTC (permalink / raw)
  To: Laurent Pinchart
  Cc: Frank Li, Rafael J. Wysocki, linux-pm, Jani Nikula,
	Joonas Lahtinen, Rodrigo Vivi, Tvrtko Ursulin, Maarten Lankhorst,
	Maxime Ripard, Thomas Hellstrom, Simona Vetter, David Airlie,
	Matthew Brost, Thomas Zimmermann, Patrik Jakobsson, Alessio Belle,
	Luigi Santivetti, Liam Girdwood, Mark Brown, Jaroslav Kysela,
	Takashi Iwai, Bard Liao, Pierre-Louis Bossart, Peter Ujfalusi,
	Kai Vehmanen, Robin Murphy, Yong Wu, AngeloGioacchino Del Regno,
	Matthias Brugger, Heiko Stuebner, Mauro Carvalho Chehab,
	Dafna Hirschfeld, Michael Tretter, Daniel Baluta, Will Deacon,
	Vijendar Mukunda, dri-devel, imagination, intel-gfx, intel-xe,
	iommu, linux-arm-kernel, linux-mediatek, linux-media,
	linux-rockchip, linux-sound, sound-open-firmware, imx

On Mon, Nov 16, 2026 at 05:31:39PM +0200, Laurent Pinchart wrote:
> Hi Frank,
>
> On Mon, Nov 16, 2026 at 10:25:11AM -0500, Frank.Li@oss.nxp.com wrote:
> > From: Frank Li <Frank.Li@nxp.com>
> >
> > The stub for pm_runtime_get_if_in_use() under !CONFIG_PM currently
> > returns -EINVAL.  The real implementation returns:
> >   1  - device is active and its usage counter is non-zero (bumped)
> >   0  - device is not in use or not active (counter not bumped)
> >  -EINVAL - runtime PM disabled for this device
> >
> > When CONFIG_PM is not set every device is permanently "active" and
> > always considered "in use", so the stub must return 1, mirroring
> > __pm_runtime_resume() which already returns 1 in the same build.
> >
> > Returning -EINVAL breaks callers that test the return value with
> > comparisons other than "!= 0 / == 0":
> >
> > drivers/gpu/drm/i915/intel_runtime_pm.c:
> >   if (intel_runtime_pm_get_if_in_use(rpm))
> >   which wraps to: ret = pm_runtime_get_if_in_use(); return ret > 0;
> >   With -EINVAL the wrapper returns false, causing an incorrect error
> >   path in the GT fence revocation code.
> >
> > drivers/gpu/drm/xe/xe_pm.c:
> >   if (pm_runtime_get_if_in_use(xe->drm.dev) > 0)
> >   With -EINVAL the condition is false; xe_pm_runtime_get_if_in_use()
> >   returns false, so callers believe the device is not usable.
> >
> > drivers/gpu/drm/gma500/power.c:
> >   return pm_runtime_get_if_in_use(dev) == 1;
> >   -EINVAL != 1, so the helper always returns false when !CONFIG_PM.
> >
> > drivers/gpu/drm/imagination/pvr_power.c:
> >   if (pm_runtime_get_if_in_use(...) <= 0)
> >   -EINVAL <= 0 is true, so the driver always takes the "requeue" error
> >   path instead of proceeding with the active device.
> >
> > sound/soc/sof/debug.c:
> >   if (pm_runtime_get_if_in_use(sdev->dev) == 1)
> >   -EINVAL != 1, so d3_prevented is never set, corrupting exception
> >   debug accounting.
> >
> > drivers/iommu/mtk_iommu.c:
> >   if (pm_runtime_get_if_in_use(dom->plat_dev->dev) <= 0)
> >   -EINVAL <= 0 skips the TLB flush, leaving stale IOMMU mappings.
> >
> > drivers/iommu/rockchip-iommu.c:
> >   Three call sites use WARN_ON_ONCE(err < 0) or similar patterns.
> >   -EINVAL < 0 triggers a spurious WARNING on every call.
> >
> > drivers/media/i2c/isl7998x.c:
> >   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
> >   -EINVAL <= 0 causes the driver to always skip register programming
> >   when !CONFIG_PM.
> >
> > drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c:
> >   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
> >   -EINVAL <= 0 prevents any debug statistics from being collected.
> >
> > Fixes: 9d58aa46291d ("drm/i915: Fix the GT fence revocation runtime PM logic")
> > Fixes: cbb6a7413b17 ("drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers")
> > Fixes: 672c473576ca ("drm/gma500: Rewrite power management code")
> > Fixes: 727538a4bbff ("drm/imagination: Implement power management")
> > Fixes: b30289e7fa92 ("ASoC: SOF: Fix runtime pm usage counter balance after fw exception")
> > Fixes: c0b57581b73b ("iommu/mediatek: Add power-domain operation")
> > Fixes: 0f181d3cf7d9 ("iommu/rockchip: Add runtime PM support")
> > Fixes: 51ef2be546e2 ("media: i2c: isl7998x: Add driver for Intersil ISL7998x")
> > Fixes: cbb819c363c5 ("media: rkisp1: debug: Collect input status by sampling ISP_FLAGS_SHD")
>
> I don't think this one is right. The rkisp1 driver only runs on ARCH_MXC
> and ARCH_ROCKCHIP, which select CONFIG_PM. The same may be true for some
> of the other fixes tags.

But code itself should be wrong even though these platform force enable
CONFIG_PM.

Frank

>
> > Signed-off-by: Frank Li <Frank.Li@nxp.com>
> > ---
> > Change in v2
> > - new patches.
> >
> > Cc: "Rafael J. Wysocki" <rafael@kernel.org>
> > Cc: linux-pm@vger.kernel.org
> > Cc: Jani Nikula <jani.nikula@linux.intel.com>
> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> > Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > Cc: Tvrtko Ursulin <tursulin@ursulin.net>
> > Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> > Cc: Maxime Ripard <mripard@kernel.org>
> > Cc: "Thomas Hellstrom" <thomas.hellstrom@linux.intel.com>
> > Cc: Simona Vetter <simona@ffwll.ch>
> > Cc: David Airlie <airlied@gmail.com>
> > Cc: Matthew Brost <matthew.brost@intel.com>
> > Cc: Thomas Zimmermann <tzimmermann@suse.de>
> > Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
> > Cc: Alessio Belle <alessio.belle@imgtec.com>
> > Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
> > Cc: Liam Girdwood <lgirdwood@gmail.com>
> > Cc: Mark Brown <broonie@kernel.org>
> > Cc: Jaroslav Kysela <perex@perex.cz>
> > Cc: Takashi Iwai <tiwai@suse.com>
> > Cc: Bard Liao <yung-chuan.liao@linux.intel.com>
> > Cc: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
> > Cc: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
> > Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
> > Cc: Robin Murphy <robin.murphy@arm.com>
> > Cc: Yong Wu <yong.wu@mediatek.com>
> > Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> > Cc: Matthias Brugger <matthias.bgg@gmail.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
> > Cc: Dafna Hirschfeld <dafna@fastmail.com>
> > Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > Cc: Michael Tretter <m.tretter@pengutronix.de>
> > Cc: Daniel Baluta <daniel.baluta@nxp.com>
> > Cc: Will Deacon <will@kernel.org>
> > Cc: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
> > Cc: dri-devel@lists.freedesktop.org
> > Cc: imagination@lists.freedesktop.org
> > Cc: intel-gfx@lists.freedesktop.org
> > Cc: intel-xe@lists.freedesktop.org
> > Cc: iommu@lists.linux.dev
> > Cc: linux-arm-kernel@lists.infradead.org
> > Cc: linux-mediatek@lists.infradead.org
> > Cc: linux-media@vger.kernel.org
> > Cc: linux-rockchip@lists.infradead.org
> > Cc: linux-sound@vger.kernel.org
> > Cc: sound-open-firmware@alsa-project.org
> > Cc: imx@lists.linux.dev
> > ---
> >  include/linux/pm_runtime.h | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
> > index de9e7518ecef2..1d983fe55a16a 100644
> > --- a/include/linux/pm_runtime.h
> > +++ b/include/linux/pm_runtime.h
> > @@ -276,7 +276,7 @@ static inline int pm_schedule_suspend(struct device *dev, unsigned int delay)
> >  }
> >  static inline int pm_runtime_get_if_in_use(struct device *dev)
> >  {
> > -	return -EINVAL;
> > +	return 1;
> >  }
> >  static inline int pm_runtime_get_if_active(struct device *dev)
> >  {
>
> --
> Regards,
>
> Laurent Pinchart

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() for !CONFIG_PM
  2026-11-16 15:43     ` Frank Li
@ 2026-11-16 15:52       ` Laurent Pinchart
  0 siblings, 0 replies; 7+ messages in thread
From: Laurent Pinchart @ 2026-11-16 15:52 UTC (permalink / raw)
  To: Frank Li
  Cc: Frank Li, Rafael J. Wysocki, linux-pm, Jani Nikula,
	Joonas Lahtinen, Rodrigo Vivi, Tvrtko Ursulin, Maarten Lankhorst,
	Maxime Ripard, Thomas Hellstrom, Simona Vetter, David Airlie,
	Matthew Brost, Thomas Zimmermann, Patrik Jakobsson, Alessio Belle,
	Luigi Santivetti, Liam Girdwood, Mark Brown, Jaroslav Kysela,
	Takashi Iwai, Bard Liao, Pierre-Louis Bossart, Peter Ujfalusi,
	Kai Vehmanen, Robin Murphy, Yong Wu, AngeloGioacchino Del Regno,
	Matthias Brugger, Heiko Stuebner, Mauro Carvalho Chehab,
	Dafna Hirschfeld, Michael Tretter, Daniel Baluta, Will Deacon,
	Vijendar Mukunda, dri-devel, imagination, intel-gfx, intel-xe,
	iommu, linux-arm-kernel, linux-mediatek, linux-media,
	linux-rockchip, linux-sound, sound-open-firmware, imx

On Mon, Nov 16, 2026 at 09:43:13AM -0600, Frank Li wrote:
> On Mon, Nov 16, 2026 at 05:31:39PM +0200, Laurent Pinchart wrote:
> > On Mon, Nov 16, 2026 at 10:25:11AM -0500, Frank.Li@oss.nxp.com wrote:
> > > From: Frank Li <Frank.Li@nxp.com>
> > >
> > > The stub for pm_runtime_get_if_in_use() under !CONFIG_PM currently
> > > returns -EINVAL.  The real implementation returns:
> > >   1  - device is active and its usage counter is non-zero (bumped)
> > >   0  - device is not in use or not active (counter not bumped)
> > >  -EINVAL - runtime PM disabled for this device
> > >
> > > When CONFIG_PM is not set every device is permanently "active" and
> > > always considered "in use", so the stub must return 1, mirroring
> > > __pm_runtime_resume() which already returns 1 in the same build.
> > >
> > > Returning -EINVAL breaks callers that test the return value with
> > > comparisons other than "!= 0 / == 0":
> > >
> > > drivers/gpu/drm/i915/intel_runtime_pm.c:
> > >   if (intel_runtime_pm_get_if_in_use(rpm))
> > >   which wraps to: ret = pm_runtime_get_if_in_use(); return ret > 0;
> > >   With -EINVAL the wrapper returns false, causing an incorrect error
> > >   path in the GT fence revocation code.
> > >
> > > drivers/gpu/drm/xe/xe_pm.c:
> > >   if (pm_runtime_get_if_in_use(xe->drm.dev) > 0)
> > >   With -EINVAL the condition is false; xe_pm_runtime_get_if_in_use()
> > >   returns false, so callers believe the device is not usable.
> > >
> > > drivers/gpu/drm/gma500/power.c:
> > >   return pm_runtime_get_if_in_use(dev) == 1;
> > >   -EINVAL != 1, so the helper always returns false when !CONFIG_PM.
> > >
> > > drivers/gpu/drm/imagination/pvr_power.c:
> > >   if (pm_runtime_get_if_in_use(...) <= 0)
> > >   -EINVAL <= 0 is true, so the driver always takes the "requeue" error
> > >   path instead of proceeding with the active device.
> > >
> > > sound/soc/sof/debug.c:
> > >   if (pm_runtime_get_if_in_use(sdev->dev) == 1)
> > >   -EINVAL != 1, so d3_prevented is never set, corrupting exception
> > >   debug accounting.
> > >
> > > drivers/iommu/mtk_iommu.c:
> > >   if (pm_runtime_get_if_in_use(dom->plat_dev->dev) <= 0)
> > >   -EINVAL <= 0 skips the TLB flush, leaving stale IOMMU mappings.
> > >
> > > drivers/iommu/rockchip-iommu.c:
> > >   Three call sites use WARN_ON_ONCE(err < 0) or similar patterns.
> > >   -EINVAL < 0 triggers a spurious WARNING on every call.
> > >
> > > drivers/media/i2c/isl7998x.c:
> > >   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
> > >   -EINVAL <= 0 causes the driver to always skip register programming
> > >   when !CONFIG_PM.
> > >
> > > drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c:
> > >   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
> > >   -EINVAL <= 0 prevents any debug statistics from being collected.
> > >
> > > Fixes: 9d58aa46291d ("drm/i915: Fix the GT fence revocation runtime PM logic")
> > > Fixes: cbb6a7413b17 ("drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers")
> > > Fixes: 672c473576ca ("drm/gma500: Rewrite power management code")
> > > Fixes: 727538a4bbff ("drm/imagination: Implement power management")
> > > Fixes: b30289e7fa92 ("ASoC: SOF: Fix runtime pm usage counter balance after fw exception")
> > > Fixes: c0b57581b73b ("iommu/mediatek: Add power-domain operation")
> > > Fixes: 0f181d3cf7d9 ("iommu/rockchip: Add runtime PM support")
> > > Fixes: 51ef2be546e2 ("media: i2c: isl7998x: Add driver for Intersil ISL7998x")
> > > Fixes: cbb819c363c5 ("media: rkisp1: debug: Collect input status by sampling ISP_FLAGS_SHD")
> >
> > I don't think this one is right. The rkisp1 driver only runs on ARCH_MXC
> > and ARCH_ROCKCHIP, which select CONFIG_PM. The same may be true for some
> > of the other fixes tags.
> 
> But code itself should be wrong even though these platform force enable
> CONFIG_PM.

How does it make the code wrong in the rkisp1 driver ? There's nothing
to fix there, because there's no bug.

I'm not saying this patch is wrong (I'll let the runtime PM folks
provide expert feedback), but I think some of the Fixes: tags are not
correct.

> > > Signed-off-by: Frank Li <Frank.Li@nxp.com>
> > > ---
> > > Change in v2
> > > - new patches.
> > >
> > > Cc: "Rafael J. Wysocki" <rafael@kernel.org>
> > > Cc: linux-pm@vger.kernel.org
> > > Cc: Jani Nikula <jani.nikula@linux.intel.com>
> > > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> > > Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > > Cc: Tvrtko Ursulin <tursulin@ursulin.net>
> > > Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> > > Cc: Maxime Ripard <mripard@kernel.org>
> > > Cc: "Thomas Hellstrom" <thomas.hellstrom@linux.intel.com>
> > > Cc: Simona Vetter <simona@ffwll.ch>
> > > Cc: David Airlie <airlied@gmail.com>
> > > Cc: Matthew Brost <matthew.brost@intel.com>
> > > Cc: Thomas Zimmermann <tzimmermann@suse.de>
> > > Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
> > > Cc: Alessio Belle <alessio.belle@imgtec.com>
> > > Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
> > > Cc: Liam Girdwood <lgirdwood@gmail.com>
> > > Cc: Mark Brown <broonie@kernel.org>
> > > Cc: Jaroslav Kysela <perex@perex.cz>
> > > Cc: Takashi Iwai <tiwai@suse.com>
> > > Cc: Bard Liao <yung-chuan.liao@linux.intel.com>
> > > Cc: Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>
> > > Cc: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
> > > Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
> > > Cc: Robin Murphy <robin.murphy@arm.com>
> > > Cc: Yong Wu <yong.wu@mediatek.com>
> > > Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> > > Cc: Matthias Brugger <matthias.bgg@gmail.com>
> > > Cc: Heiko Stuebner <heiko@sntech.de>
> > > Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
> > > Cc: Dafna Hirschfeld <dafna@fastmail.com>
> > > Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > > Cc: Michael Tretter <m.tretter@pengutronix.de>
> > > Cc: Daniel Baluta <daniel.baluta@nxp.com>
> > > Cc: Will Deacon <will@kernel.org>
> > > Cc: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
> > > Cc: dri-devel@lists.freedesktop.org
> > > Cc: imagination@lists.freedesktop.org
> > > Cc: intel-gfx@lists.freedesktop.org
> > > Cc: intel-xe@lists.freedesktop.org
> > > Cc: iommu@lists.linux.dev
> > > Cc: linux-arm-kernel@lists.infradead.org
> > > Cc: linux-mediatek@lists.infradead.org
> > > Cc: linux-media@vger.kernel.org
> > > Cc: linux-rockchip@lists.infradead.org
> > > Cc: linux-sound@vger.kernel.org
> > > Cc: sound-open-firmware@alsa-project.org
> > > Cc: imx@lists.linux.dev
> > > ---
> > >  include/linux/pm_runtime.h | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h
> > > index de9e7518ecef2..1d983fe55a16a 100644
> > > --- a/include/linux/pm_runtime.h
> > > +++ b/include/linux/pm_runtime.h
> > > @@ -276,7 +276,7 @@ static inline int pm_schedule_suspend(struct device *dev, unsigned int delay)
> > >  }
> > >  static inline int pm_runtime_get_if_in_use(struct device *dev)
> > >  {
> > > -	return -EINVAL;
> > > +	return 1;
> > >  }
> > >  static inline int pm_runtime_get_if_active(struct device *dev)
> > >  {

-- 
Regards,

Laurent Pinchart

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() for !CONFIG_PM
  2026-11-16 15:25 [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() for !CONFIG_PM Frank.Li
  2026-11-16 15:25 ` [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() " Frank.Li
@ 2026-11-17 15:25 ` sashiko-bot
  1 sibling, 0 replies; 7+ messages in thread
From: sashiko-bot @ 2026-11-17 15:25 UTC (permalink / raw)
  To: Frank.Li
  Cc: linux-usb, dri-devel, intel-gfx, intel-xe, linux-hwmon, imx,
	Frank.Li, media-ci

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Low] PM: runtime: inaccurate Fixes trailers and caller analysis in
  commit message
--

commit ef3656b915b0fdf6f58ce6d87b9638c3f1514d1a
Author: Frank Li <Frank.Li@nxp.com>

> PM: runtime: fix stub pm_runtime_get_if_active() for !CONFIG_PM
>
> When CONFIG_PM is not enabled, should be treated as if runtime PM were
> enabled for the device and the runtime status is active.
>
> It's a subtle distinction but the previous behavior returned -EINVAL, which
> means "runtime PM is disabled for the device"
>
> __pm_runtime_resume() returns 1 (called by pm_runtime_get()).
> pm_runtime_get_if_active() is the try-and-get version of pm_runtime_get().
> stub pm_runtime_active() also returns true. So pm_runtime_get_if_active()
> should return the same value as the stub pm_runtime_get(), which is 1.
>
> Analysis of all callers and their behavior with old (-EINVAL) vs new (1):
>
> Pattern A: if (!pm_runtime_get_if_active(dev)) -- no impact
>   Both -EINVAL and 1 are non-zero, so !ret is false in both cases.
>   Callers: drivers/accel/ethosu/ethosu_perfmon.c,
>            drivers/gpu/drm/msm/adreno/a6xx_gmu.c,
>            drivers/gpu/drm/msm/msm_perfcntr.c,
>            drivers/gpu/drm/v3d/v3d_mmu.c,
>            drivers/gpu/drm/v3d/v3d_perfmon.c (x4),
>            drivers/media/i2c/gc08a3.c gc05a2.c og0ve1b.c ov6211.c,
>            drivers/media/i2c/ccs/ccs-core.c s5kjn1.c lm3560.c (x2),
>            drivers/media/i2c/s5k3m5.c thp7312.c imx283.c ov64a40.c,
>            drivers/media/pci/intel/ipu6/ipu6-buttress.c,
>            sound/hda/core/device.c
>
> Pattern B: if (!ret) after get -- no impact
>   Both -EINVAL and 1 are non-zero, !ret is false in both cases.
>   drivers/gpu/drm/amd/pm/amdgpu_pm.c (comment says "allow access"
>     in kernels without CONFIG_PM), drivers/pci/pci.c
>
> Pattern C: ret > 0 / ret <= 0 / ret != 1 checks -- FIXED
>   drivers/usb/core/offload.c: checks (!= 1), old -EINVAL caused
>     -EBUSY even with !CONFIG_PM. Fixed: now returns 1, proceeds.
>   drivers/accel/ivpu/ivpu_job.c: drm_WARN_ON(<= 0), old -EINVAL
>     spuriously triggered WARN. Fixed: 1 > 0, no warn.
>   drivers/gpu/drm/i915/intel_runtime_pm.c: checks (<= 0), old
>     -EINVAL triggered error path. Fixed: 1 > 0, correct path.
>   drivers/accel/ivpu/ivpu_sysfs.c: (> 0), old skipped freq read.
>     Fixed: proceeds correctly without CONFIG_PM.
>   drivers/gpu/drm/xe/xe_pm.c: wraps (> 0), all xe callers
>     (xe_bo.c, xe_shrinker.c, xe_vsec.c, xe_ggtt.c, xe_guc_ct.c,
>     xe_pxp.c) now correctly treat !CONFIG_PM as active.
>   drivers/net/ipa/ipa_smp2p.c: power_on = (> 0), now correctly true.
>   drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c: (> 0), now correctly
>     executes stall register update.
>   drivers/staging/media/ipu7/ipu7-buttress.c: active (<= 0) early
>     return now skipped, correctly proceeds.
>
> Pattern D: explicit -EINVAL check
>   drivers/hwmon/prom21-xhci.c: checks (> 0) then (== -EINVAL).
>     Old: took -EINVAL fallback. New: takes (> 0) fast path.
>     Both succeed when CONFIG_PM is disabled; functionally equivalent.
>   drivers/hwmon/eic7700-pvt.c: call is guarded by IS_ENABLED(CONFIG_PM),
>     stub is never reached. No impact.
>
> Fixes: bd3d245b0fef5 ("usb: core: use dedicated spinlock for offload state")
> Fixes: 7806bad76ac39 ("accel/ivpu: Prevent runtime suspend during context abort work")
> Fixes: 9d58aa46291d4 ("drm/i915: Fix the GT fence revocation runtime PM logic")
> Fixes: 1524c28b99527 ("accel/ivpu: Show NPU frequency in sysfs")
> Fixes: 30c399529f4c6 ("drm/xe: Document Xe PM component")
> Fixes: 0d08026ac6099 ("net: ipa: kill ipa_clock_get_additional()")
> Fixes: 70892277ca2db ("iommu/arm-smmu-qcom: Make set_stall work when the device is on")
> Fixes: 9ad65684b9285 ("media: staging/ipu7: Ignore interrupts when device is suspended")
> Acked-by: Alex Elder <elder@riscstar.com>
> Signed-off-by: Frank Li <Frank.Li@nxp.com>

[Severity: Low]
Are some of the Fixes: tags and caller notes in the commit message
pointing to the wrong commits or config guards?

The !CONFIG_PM stub for pm_runtime_get_if_active() was introduced
returning -EINVAL in c111566bea7c ("PM: runtime: Add
pm_runtime_get_if_active()") and updated in c0ef3df8dbae ("PM: runtime:
Simplify pm_runtime_get_if_active() usage"), neither of which is listed in
Fixes:.

For the caller-specific tags and descriptions:

- 30c399529f4c ("drm/xe: Document Xe PM component") only touched kernel-doc
  comments; the > 0 check in xe_pm_runtime_get_if_active() came from
  46edb0a3eb16 ("drm/xe: Fix xe_pm_runtime_get_if_active return").

- In __intel_runtime_pm_get_if_active() in
  drivers/gpu/drm/i915/intel_runtime_pm.c, the pm_runtime_get_if_active()
  call is guarded by if (IS_ENABLED(CONFIG_PM)), so the !CONFIG_PM stub is
  never reached there and 9d58aa46291d ("drm/i915: Fix the GT fence
  revocation runtime PM logic") was not affected.  Similarly,
  snd_hdac_keep_power_up() in sound/hda/core/device.c is inside #ifdef
  CONFIG_PM and also checks if (ret < 0).

- 0d08026ac609 ("net: ipa: kill ipa_clock_get_additional()") only inlined
  ipa_clock_get_additional() into ipa_smp2p_notify(); the
  pm_runtime_get_if_active() > 0 check was introduced in a71aeff3dd0a
  ("net: ipa: kill IPA clock reference count").

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261116152513.2676573-1-Frank.Li@oss.nxp.com?part=1

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() for !CONFIG_PM
  2026-11-16 15:25 ` [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() " Frank.Li
  2026-11-16 15:31   ` Laurent Pinchart
@ 2026-11-17 15:25   ` sashiko-bot
  1 sibling, 0 replies; 7+ messages in thread
From: sashiko-bot @ 2026-11-17 15:25 UTC (permalink / raw)
  To: Frank.Li
  Cc: dri-devel, intel-gfx, intel-xe, linux-sound, imx, Frank.Li,
	media-ci

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Low] PM: runtime: inaccurate Fixes trailers and caller notes in
  commit message
--

commit 48057d65d03972fb7f912d84a7a35bf1b15bba82
Author: Frank Li <Frank.Li@nxp.com>

> PM: runtime: fix stub pm_runtime_get_if_in_use() for !CONFIG_PM
>
> The stub for pm_runtime_get_if_in_use() under !CONFIG_PM currently
> returns -EINVAL.  The real implementation returns:
>   1  - device is active and its usage counter is non-zero (bumped)
>   0  - device is not in use or not active (counter not bumped)
>  -EINVAL - runtime PM disabled for this device
>
> When CONFIG_PM is not set every device is permanently "active" and
> always considered "in use", so the stub must return 1, mirroring
> __pm_runtime_resume() which already returns 1 in the same build.
>
> Returning -EINVAL breaks callers that test the return value with
> comparisons other than "!= 0 / == 0":
>
> drivers/gpu/drm/i915/intel_runtime_pm.c:
>   if (intel_runtime_pm_get_if_in_use(rpm))
>   which wraps to: ret = pm_runtime_get_if_in_use(); return ret > 0;
>   With -EINVAL the wrapper returns false, causing an incorrect error
>   path in the GT fence revocation code.
>
> drivers/gpu/drm/xe/xe_pm.c:
>   if (pm_runtime_get_if_in_use(xe->drm.dev) > 0)
>   With -EINVAL the condition is false; xe_pm_runtime_get_if_in_use()
>   returns false, so callers believe the device is not usable.
>
> drivers/gpu/drm/gma500/power.c:
>   return pm_runtime_get_if_in_use(dev) == 1;
>   -EINVAL != 1, so the helper always returns false when !CONFIG_PM.
>
> drivers/gpu/drm/imagination/pvr_power.c:
>   if (pm_runtime_get_if_in_use(...) <= 0)
>   -EINVAL <= 0 is true, so the driver always takes the "requeue" error
>   path instead of proceeding with the active device.
>
> sound/soc/sof/debug.c:
>   if (pm_runtime_get_if_in_use(sdev->dev) == 1)
>   -EINVAL != 1, so d3_prevented is never set, corrupting exception
>   debug accounting.
>
> drivers/iommu/mtk_iommu.c:
>   if (pm_runtime_get_if_in_use(dom->plat_dev->dev) <= 0)
>   -EINVAL <= 0 skips the TLB flush, leaving stale IOMMU mappings.
>
> drivers/iommu/rockchip-iommu.c:
>   Three call sites use WARN_ON_ONCE(err < 0) or similar patterns.
>   -EINVAL < 0 triggers a spurious WARNING on every call.
>
> drivers/media/i2c/isl7998x.c:
>   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
>   -EINVAL <= 0 causes the driver to always skip register programming
>   when !CONFIG_PM.
>
> drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c:
>   if (pm_runtime_get_if_in_use(dev) <= 0)  (two call sites)
>   -EINVAL <= 0 prevents any debug statistics from being collected.
>
> Fixes: 9d58aa46291d ("drm/i915: Fix the GT fence revocation runtime PM logic")
> Fixes: cbb6a7413b17 ("drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers")
> Fixes: 672c473576ca ("drm/gma500: Rewrite power management code")
> Fixes: 727538a4bbff ("drm/imagination: Implement power management")
> Fixes: b30289e7fa92 ("ASoC: SOF: Fix runtime pm usage counter balance after fw exception")
> Fixes: c0b57581b73b ("iommu/mediatek: Add power-domain operation")
> Fixes: 0f181d3cf7d9 ("iommu/rockchip: Add runtime PM support")
> Fixes: 51ef2be546e2 ("media: i2c: isl7998x: Add driver for Intersil ISL7998x")
> Fixes: cbb819c363c5 ("media: rkisp1: debug: Collect input status by sampling ISP_FLAGS_SHD")
> Signed-off-by: Frank Li <Frank.Li@nxp.com>

[Severity: Low]
Are the Fixes tags and caller descriptions in the commit message accurate?

The !CONFIG_PM stub for pm_runtime_get_if_in_use() returning -EINVAL was
introduced in a436b6a19f57 ("PM / runtime: Add new helper for conditional
usage count incrementation"), which is missing from the Fixes tags.

For i915, __intel_runtime_pm_get_if_active() guards the call to
pm_runtime_get_if_in_use() with if (IS_ENABLED(CONFIG_PM)), so the
!CONFIG_PM stub is never called there, and 9d58aa46291d ("drm/i915: Fix the
GT fence revocation runtime PM logic") switched i915_vma_revoke_fence() to
with_intel_runtime_pm_if_active, which calls pm_runtime_get_if_active().

For imagination, DRM_POWERVR in drivers/gpu/drm/imagination/Kconfig has
"depends on PM" (added in 727538a4bbff ("drm/imagination: Implement power
management")), so pvr_power.c cannot be built with !CONFIG_PM.

Also, several of the remaining Fixes tags do not point to the commits that
introduced the affected checks:
- In drivers/gpu/drm/xe/xe_pm.c, the pm_runtime_get_if_in_use() > 0 check
  in xe_pm_runtime_get_if_in_use() was added by 3b85b7bcccf4 ("drm/xe: Fix
  display runtime_pm handling"), whereas cbb6a7413b17 ("drm/xe: Introduce
  xe_pm_runtime_get_noresume for inner callers") only added a caller.
- In drivers/iommu/rockchip-iommu.c, 0f181d3cf7d9 ("iommu/rockchip: Add
  runtime PM support") used boolean checks where -EINVAL evaluated to true;
  the WARN_ON_ONCE() checks on negative or non-positive returns were added
  by 3fc7c5c0cff3 ("iommu/rockchip: Handle errors returned from PM
  framework").
- In drivers/media/platform/rockchip/rkisp1/rkisp1-debug.c, the first
  pm_runtime_get_if_in_use() <= 0 check in rkisp1_debug_dump_regs() was
  added earlier by 2a6b586035ec ("media: rkisp1: debug: Add debugfs files
  to dump core and ISP registers").

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261116152513.2676573-1-Frank.Li@oss.nxp.com?part=2

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2026-11-17 15:25 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-16 15:25 [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() for !CONFIG_PM Frank.Li
2026-11-16 15:25 ` [PATCH v2 2/2] PM: runtime: fix stub pm_runtime_get_if_in_use() " Frank.Li
2026-11-16 15:31   ` Laurent Pinchart
2026-11-16 15:43     ` Frank Li
2026-11-16 15:52       ` Laurent Pinchart
2026-11-17 15:25   ` sashiko-bot
2026-11-17 15:25 ` [PATCH v2 1/2] PM: runtime: fix stub pm_runtime_get_if_active() " sashiko-bot

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox