[PATCH v2 0/5] ACPI: Replace acpi_get_first_physical_node() with acpi_bus_get_primary_device()

[PATCH v2 0/5] ACPI: Replace acpi_get_first_physical_node() with acpi_bus_get_primary_device()

[Rafael J. Wysocki]

Hi All,

This is a partial resend of

https://lore.kernel.org/linux-acpi/7995179.EvYhyI6sBW@rafael.j.wysocki/

that only includes patches that have not been ACKed and the last one that
depends on all of the others.  It is based on linux-next.

This series replaces acpi_get_first_physical_node(), which is inherently racy
and therefore generally unsafe to use, with acpi_bus_get_primary_device(), that
doesn't have this problem, everywhere in the code.

I will appreciate ACKs from maintainers of the code updated by the patches, so
I can pick them up and get rid of the problematic interface in one go.

Thanks!

[PATCH v2 1/5] efi/dev-path-parser: Switch to use acpi_bus_get_primary_device()

[Rafael J. Wysocki]

From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>

Replace acpi_get_first_physical_node() that is slated for removal
with acpi_bus_get_primary_device() that takes a reference to the
device it is about to return.

This addresses a potential use-after-free that may occur if the
device returned by acpi_get_first_physical_node() is removed right
after dropping its ACPI companion's physical_node_lock in that
function.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

This patch depends on new material in linux-next.

If you maintain the code updated by it, please consider ACKing it, so I can
pick it up and remove the problematic API replaced by it during the 7.4
merge window.

In the absence of feedback, it will be resent again when 7.4-rc1 is out.

Thanks!

v1 -> v2: No changes

---
 drivers/firmware/efi/dev-path-parser.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/firmware/efi/dev-path-parser.c b/drivers/firmware/efi/dev-path-parser.c
index 13ea141c0def..e0f864cbaa8c 100644
--- a/drivers/firmware/efi/dev-path-parser.c
+++ b/drivers/firmware/efi/dev-path-parser.c
@@ -37,9 +37,9 @@ static long __init parse_acpi_path(const struct efi_dev_path *node,
 	if (!adev)
 		return -ENODEV;
 
-	phys_dev = acpi_get_first_physical_node(adev);
+	phys_dev = acpi_bus_get_primary_device(adev);
 	if (phys_dev) {
-		*child = get_device(phys_dev);
+		*child = phys_dev;
 		acpi_dev_put(adev);
 	} else
 		*child = &adev->dev;
-- 
2.51.0

[PATCH v2 3/5] ALSA: hda: cs35l41: Switch to use acpi_bus_get_primary_device()

[Rafael J. Wysocki]

From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>

Replace acpi_get_first_physical_node() that is slated for removal
with acpi_bus_get_primary_device() that takes a reference to the
device it is about to return.

This addresses a potential use-after-free that may occur if the
device returned by acpi_get_first_physical_node() is removed right
after dropping its ACPI companion's physical_node_lock in that
function.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

This patch depends on new material in linux-next.

If you maintain the code updated by it, please consider ACKing it, so I can
pick it up and remove the problematic API replaced by it during the 7.4
merge window.

In the absence of feedback, it will be resent again when 7.4-rc1 is out.

Thanks!

v1 -> v2: No changes

---
 sound/hda/codecs/side-codecs/cs35l41_hda.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/sound/hda/codecs/side-codecs/cs35l41_hda.c b/sound/hda/codecs/side-codecs/cs35l41_hda.c
index 62296be45c55..32abb71dd401 100644
--- a/sound/hda/codecs/side-codecs/cs35l41_hda.c
+++ b/sound/hda/codecs/side-codecs/cs35l41_hda.c
@@ -1926,8 +1926,7 @@ static int cs35l41_hda_read_acpi(struct cs35l41_hda *cs35l41, const char *hid, i
 	}
 
 	cs35l41->dacpi = adev;
-	struct device *physdev __free(put_device) =
-		get_device(acpi_get_first_physical_node(adev));
+	struct device *physdev __free(put_device) = acpi_bus_get_primary_device(adev);
 	if (!physdev) {
 		acpi_dev_put(adev);
 		return -ENODEV;
-- 
2.51.0

Re: [PATCH v2 3/5] ALSA: hda: cs35l41: Switch to use acpi_bus_get_primary_device()

[Takashi Iwai]

On Wed, 30 Sep 2026 20:42:15 +0200,
Rafael J. Wysocki wrote:
> 
> From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>
> 
> Replace acpi_get_first_physical_node() that is slated for removal
> with acpi_bus_get_primary_device() that takes a reference to the
> device it is about to return.
> 
> This addresses a potential use-after-free that may occur if the
> device returned by acpi_get_first_physical_node() is removed right
> after dropping its ACPI companion's physical_node_lock in that
> function.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
> 
> This patch depends on new material in linux-next.
> 
> If you maintain the code updated by it, please consider ACKing it, so I can
> pick it up and remove the problematic API replaced by it during the 7.4
> merge window.
> 
> In the absence of feedback, it will be resent again when 7.4-rc1 is out.
> 
> Thanks!

Acked-by: Takashi Iwai <tiwai@suse.de>


thanks,

Takashi

[PATCH v2 4/5] ALSA: hda: tas2781: Switch to use acpi_bus_get_primary_device()

[Rafael J. Wysocki]

From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>

Replace acpi_get_first_physical_node() that is slated for removal
with acpi_bus_get_primary_device() that takes a reference to the
device it is about to return.

This addresses a potential use-after-free that may occur if the
device returned by acpi_get_first_physical_node() is removed right
after dropping its ACPI companion's physical_node_lock in that
function.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

This patch depends on new material in linux-next.

If you maintain the code updated by it, please consider ACKing it, so I can
pick it up and remove the problematic API replaced by it during the 7.4
merge window.

In the absence of feedback, it will be resent again when 7.4-rc1 is out.

Thanks!

v1 -> v2: No changes

---
 sound/hda/codecs/side-codecs/tas2781_hda_i2c.c | 3 +--
 sound/hda/codecs/side-codecs/tas2781_hda_spi.c | 3 +--
 2 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c b/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c
index 5e5f46c9fb6b..161ca965c63e 100644
--- a/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c
+++ b/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c
@@ -100,8 +100,7 @@ static int tas2781_read_acpi(struct tasdevice_priv *p, const char *hid)
 		return -ENODEV;
 	}
 
-	struct device *physdev __free(put_device) =
-		get_device(acpi_get_first_physical_node(adev));
+	struct device *physdev __free(put_device) = acpi_bus_get_primary_device(adev);
 	ret = acpi_dev_get_resources(adev, &resources, tas2781_get_i2c_res, p);
 	if (ret < 0) {
 		dev_err(p->dev, "Failed to get ACPI resource.\n");
diff --git a/sound/hda/codecs/side-codecs/tas2781_hda_spi.c b/sound/hda/codecs/side-codecs/tas2781_hda_spi.c
index da71b57a26ee..f3ba7132d2b8 100644
--- a/sound/hda/codecs/side-codecs/tas2781_hda_spi.c
+++ b/sound/hda/codecs/side-codecs/tas2781_hda_spi.c
@@ -341,8 +341,7 @@ static int tas2781_read_acpi(struct tas2781_hda *tas_hda,
 
 	strscpy(p->dev_name, hid, sizeof(p->dev_name));
 
-	struct device *physdev __free(put_device) =
-		get_device(acpi_get_first_physical_node(adev));
+	struct device *physdev __free(put_device) = acpi_bus_get_primary_device(adev);
 	acpi_dev_put(adev);
 	if (!physdev)
 		return -ENODEV;
-- 
2.51.0

Re: [PATCH v2 4/5] ALSA: hda: tas2781: Switch to use acpi_bus_get_primary_device()

[Takashi Iwai]

On Wed, 30 Sep 2026 20:42:36 +0200,
Rafael J. Wysocki wrote:
> 
> From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>
> 
> Replace acpi_get_first_physical_node() that is slated for removal
> with acpi_bus_get_primary_device() that takes a reference to the
> device it is about to return.
> 
> This addresses a potential use-after-free that may occur if the
> device returned by acpi_get_first_physical_node() is removed right
> after dropping its ACPI companion's physical_node_lock in that
> function.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
> 
> This patch depends on new material in linux-next.
> 
> If you maintain the code updated by it, please consider ACKing it, so I can
> pick it up and remove the problematic API replaced by it during the 7.4
> merge window.
> 
> In the absence of feedback, it will be resent again when 7.4-rc1 is out.
> 
> Thanks!

Acked-by: Takashi Iwai <tiwai@suse.de>


thanks,

Takashi

[PATCH v2 2/5] iommu/arm-smmu-v3: Switch to use acpi_bus_get_primary_device()

[Rafael J. Wysocki]

From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>

Replace acpi_get_first_physical_node() that is slated for removal
with acpi_bus_get_primary_device() that takes a reference to the
device it is about to return.

This addresses a potential use-after-free that may occur if the
device returned by acpi_get_first_physical_node() is removed right
after dropping its ACPI companion's physical_node_lock in that
function.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

This patch depends on new material in linux-next.

If you maintain the code updated by it, please consider ACKing it, so I can
pick it up and remove the problematic API replaced by it during the 7.4
merge window.

In the absence of feedback, it will be resent again when 7.4-rc1 is out.

Thanks!

v1 -> v2: No changes

---
 drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
index 5732f3ba0122..98f105798b89 100644
--- a/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
+++ b/drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c
@@ -5307,7 +5307,7 @@ static void acpi_smmu_dsdt_probe_tegra241_cmdqv(struct acpi_iort_node *node,
 	adev = acpi_dev_get_first_match_dev("NVDA200C", uid, -1);
 	if (adev) {
 		/* Tegra241 CMDQV driver is responsible for put_device() */
-		smmu->impl_dev = get_device(acpi_get_first_physical_node(adev));
+		smmu->impl_dev = acpi_bus_get_primary_device(adev);
 		smmu->options |= ARM_SMMU_OPT_TEGRA241_CMDQV;
 		dev_info(smmu->dev, "found companion CMDQV device: %s\n",
 			 dev_name(smmu->impl_dev));
-- 
2.51.0

Re: [PATCH v2 2/5] iommu/arm-smmu-v3: Switch to use acpi_bus_get_primary_device()

[Nicolin Chen]

On Wed, Sep 30, 2026 at 08:43:03PM +0200, Rafael J. Wysocki wrote:
> From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>
> 
> Replace acpi_get_first_physical_node() that is slated for removal
> with acpi_bus_get_primary_device() that takes a reference to the
> device it is about to return.
> 
> This addresses a potential use-after-free that may occur if the
> device returned by acpi_get_first_physical_node() is removed right
> after dropping its ACPI companion's physical_node_lock in that
> function.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
 
Reviewed-by: Nicolin Chen <nicolinc@nvidia.com>

[PATCH v2 5/5] ACPI: bus: Drop acpi_get_first_physical_node()

[Rafael J. Wysocki]

From: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>

Fold acpi_get_first_physical_node() into acpi_primary_dev_companion()
that is its only remaining caller and drop its signature from the
header file.

No intentional functional impact.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Sudeep Holla <sudeep.holla@kernel.org>
---

v1 -> v2:
   * Add R-by from Sudeep

---
 drivers/acpi/bus.c   | 20 +++++---------------
 include/linux/acpi.h |  7 -------
 2 files changed, 5 insertions(+), 22 deletions(-)

diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c
index 808c6746be14..cc6989d88645 100644
--- a/drivers/acpi/bus.c
+++ b/drivers/acpi/bus.c
@@ -809,24 +809,14 @@ struct device *acpi_bus_get_primary_device(struct acpi_device *adev)
 }
 EXPORT_SYMBOL_GPL(acpi_bus_get_primary_device);
 
-/**
- * acpi_get_first_physical_node - Find first physical node of an ACPI device
- * @adev:	ACPI device in question
- *
- * Return: First physical node of ACPI device @adev
- */
-struct device *acpi_get_first_physical_node(struct acpi_device *adev)
-{
-	guard(mutex)(&adev->physical_node_lock);
-
-	return primary_physical_device(adev);
-}
-EXPORT_SYMBOL_GPL(acpi_get_first_physical_node);
-
 static struct acpi_device *acpi_primary_dev_companion(struct acpi_device *adev,
 						      const struct device *dev)
 {
-	const struct device *phys_dev = acpi_get_first_physical_node(adev);
+	const struct device *phys_dev;
+
+	guard(mutex)(&adev->physical_node_lock);
+
+	phys_dev = primary_physical_device(adev);
 
 	return phys_dev && phys_dev == dev ? adev : NULL;
 }
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index ba5234bd93a2..6f69819f9c9a 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -105,8 +105,6 @@ static inline const char *acpi_dev_name(struct acpi_device *adev)
 	return dev_name(&adev->dev);
 }
 
-struct device *acpi_get_first_physical_node(struct acpi_device *adev);
-
 enum acpi_irq_model_id {
 	ACPI_IRQ_MODEL_PIC = 0,
 	ACPI_IRQ_MODEL_IOAPIC,
@@ -938,11 +936,6 @@ static inline const char *acpi_dev_name(struct acpi_device *adev)
 	return NULL;
 }
 
-static inline struct device *acpi_get_first_physical_node(struct acpi_device *adev)
-{
-	return NULL;
-}
-
 static inline void acpi_early_init(void) { }
 static inline void acpi_subsystem_init(void) { }
 
-- 
2.51.0

Re: [PATCH v2 0/5] ACPI: Replace acpi_get_first_physical_node() with acpi_bus_get_primary_device()

[Andy Shevchenko]

On Wed, Sep 30, 2026 at 08:30:47PM +0200, Rafael J. Wysocki wrote:
> Hi All,
> 
> This is a partial resend of
> 
> https://lore.kernel.org/linux-acpi/7995179.EvYhyI6sBW@rafael.j.wysocki/
> 
> that only includes patches that have not been ACKed and the last one that
> depends on all of the others.  It is based on linux-next.
> 
> This series replaces acpi_get_first_physical_node(), which is inherently racy
> and therefore generally unsafe to use, with acpi_bus_get_primary_device(), that
> doesn't have this problem, everywhere in the code.
> 
> I will appreciate ACKs from maintainers of the code updated by the patches, so
> I can pick them up and get rid of the problematic interface in one go.

The whole lot is straightforward,
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v2 0/5] ACPI: Replace acpi_get_first_physical_node() with acpi_bus_get_primary_device()

[Rafael J. Wysocki (Intel)]

On Thu, Oct 1, 2026 at 8:22 PM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
>
> On Wed, Sep 30, 2026 at 08:30:47PM +0200, Rafael J. Wysocki wrote:
> > Hi All,
> >
> > This is a partial resend of
> >
> > https://lore.kernel.org/linux-acpi/7995179.EvYhyI6sBW@rafael.j.wysocki/
> >
> > that only includes patches that have not been ACKed and the last one that
> > depends on all of the others.  It is based on linux-next.
> >
> > This series replaces acpi_get_first_physical_node(), which is inherently racy
> > and therefore generally unsafe to use, with acpi_bus_get_primary_device(), that
> > doesn't have this problem, everywhere in the code.
> >
> > I will appreciate ACKs from maintainers of the code updated by the patches, so
> > I can pick them up and get rid of the problematic interface in one go.
>
> The whole lot is straightforward,
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Yes, it is, thank you!