[PATCH 0/3] soc: apple: Add "PMGR misc" power controls driver

[PATCH 0/3] soc: apple: Add "PMGR misc" power controls driver

[Sasha Finkelstein]

Certain Apple SoCs have additional power state controls that are using
a separate "misc" mmio interface. Currently this includes the fabric
and memory controllers on pro/max/ultra SoCs.

This series adds a driver to put those in a low power state when the
machine enters sleep. The power savings are SoC-dependent and are
around 1W.

Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>
---
Hector Martin (1):
      soc: apple: Add driver for Apple PMGR misc controls

Sasha Finkelstein (2):
      dt-bindings: soc: apple: Add bindings for apple PMGR misc controls
      arm64: dts: apple: Add pmgr-misc nodes to t60xx

 Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml |  45 +++++++++++++++++++++++++++
 MAINTAINERS                                                            |   1 +
 arch/arm64/boot/dts/apple/t600x-die0.dtsi                              |   7 +++++
 arch/arm64/boot/dts/apple/t602x-die0.dtsi                              |   7 +++++
 drivers/soc/apple/Kconfig                                              |  10 ++++++
 drivers/soc/apple/Makefile                                             |   2 ++
 drivers/soc/apple/apple-pmgr-misc.c                                    | 177 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 7 files changed, 249 insertions(+)
---
base-commit: 59574e5978abfc2ce1c194ea0200eb6fc718d8ce
change-id: 20260702-pmgr-misc-ae1cbd5bc2c7

Best regards,
--  
Sasha Finkelstein <k@chaosmail.tech>

[PATCH 1/3] dt-bindings: soc: apple: Add bindings for apple PMGR misc controls

[Sasha Finkelstein]

Certain Apple SoCs include additional PMGR power states that are
controlled via a different "misc" control block. On existing SoCs, this
includes the fabric and memory controller state.

Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>
---
 Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml | 45 +++++++++++++++++++++++++++++++++++++++++++++
 MAINTAINERS                                                            |  1 +
 2 files changed, 46 insertions(+)

diff --git a/Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml b/Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml
new file mode 100644
index 000000000000..30abedc67fa4
--- /dev/null
+++ b/Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml
@@ -0,0 +1,45 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/soc/apple/apple,t6000-pmgr-misc.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Apple SoC PMGR Misc Power States
+
+maintainers:
+  - Sasha Finkelstein <k@chaosmail.tech>
+
+description: |
+  Certain Apple SoCs include additional PMGR power states that are controlled
+  via a different "misc" control block. On existing SoCs, this includes the
+  fabric and memory controller state.
+
+properties:
+  compatible:
+    enum:
+      - apple,t6000-pmgr-misc
+      - apple,t6020-pmgr-misc
+
+  reg:
+    maxItems: 2
+
+  reg-names:
+    items:
+      - const: fabric-ps
+      - const: dcs-ps
+
+required:
+  - compatible
+  - reg
+  - reg-names
+
+additionalProperties: false
+
+examples:
+  - |
+    pmgr_misc: power-management@8e20c000 {
+        compatible = "apple,t6000-pmgr-misc";
+        reg = <0x8e20c000 0x400>,
+              <0x8e20c800 0x400>;
+        reg-names = "fabric-ps", "dcs-ps";
+    };
diff --git a/MAINTAINERS b/MAINTAINERS
index 15011f5752a9..5a29bb86499f 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2607,6 +2607,7 @@ F:	Documentation/devicetree/bindings/power/apple*
 F:	Documentation/devicetree/bindings/power/reset/apple,smc-reboot.yaml
 F:	Documentation/devicetree/bindings/pwm/apple,s5l-fpwm.yaml
 F:	Documentation/devicetree/bindings/rtc/apple,smc-rtc.yaml
+F:	Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml
 F:	Documentation/devicetree/bindings/spi/apple,spi.yaml
 F:	Documentation/devicetree/bindings/spmi/apple,spmi.yaml
 F:	Documentation/devicetree/bindings/usb/apple,dwc3.yaml

-- 
2.55.0

[PATCH 2/3] soc: apple: Add driver for Apple PMGR misc controls

[Sasha Finkelstein]

From: Hector Martin <marcan@marcan.st>

Apple SoCs have PMGR blocks that control a bunch of power-related
features. Besides the existing device power state controls (which are
very uniform and handled by apple-pmgr-pwrstate), we also need to manage
more random registers such as SoC-wide fabric and memory controller
power states, which have a different interface.

Add a driver for these kitchen sink controls. Right now it implements
fabric and memory controller power state switching on system
standby/s2idle, which saves about 1W of power or so on t60xx platforms.

Signed-off-by: Hector Martin <marcan@marcan.st>
Co-developed-by: Sasha Finkelstein <k@chaosmail.tech>
Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>
---
 drivers/soc/apple/Kconfig           |  10 ++++++++
 drivers/soc/apple/Makefile          |   2 ++
 drivers/soc/apple/apple-pmgr-misc.c | 177 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 189 insertions(+)

diff --git a/drivers/soc/apple/Kconfig b/drivers/soc/apple/Kconfig
index d0ff32182a2b..cad847c90df8 100644
--- a/drivers/soc/apple/Kconfig
+++ b/drivers/soc/apple/Kconfig
@@ -16,6 +16,16 @@ config APPLE_MAILBOX
 
 	  Say Y here if you have an Apple SoC.
 
+config APPLE_PMGR_MISC
+	bool "Apple SoC PMGR miscellaneous support"
+	depends on PM
+	help
+	  The PMGR block in Apple SoCs provides high-level power state
+	  controls for SoC devices. This driver manages miscellaneous
+	  power controls.
+
+	  Say 'y' here if you have a Pro/Max/Ultra Apple SoC.
+
 config APPLE_RTKIT
 	tristate "Apple RTKit co-processor IPC protocol"
 	depends on APPLE_MAILBOX
diff --git a/drivers/soc/apple/Makefile b/drivers/soc/apple/Makefile
index 0b85ab61aefe..0481f7f43b06 100644
--- a/drivers/soc/apple/Makefile
+++ b/drivers/soc/apple/Makefile
@@ -3,6 +3,8 @@
 obj-$(CONFIG_APPLE_MAILBOX) += apple-mailbox.o
 apple-mailbox-y = mailbox.o
 
+obj-$(CONFIG_APPLE_PMGR_MISC)	+= apple-pmgr-misc.o
+
 obj-$(CONFIG_APPLE_RTKIT) += apple-rtkit.o
 apple-rtkit-y = rtkit.o rtkit-crashlog.o
 
diff --git a/drivers/soc/apple/apple-pmgr-misc.c b/drivers/soc/apple/apple-pmgr-misc.c
new file mode 100644
index 000000000000..d487a87ecddb
--- /dev/null
+++ b/drivers/soc/apple/apple-pmgr-misc.c
@@ -0,0 +1,177 @@
+// SPDX-License-Identifier: GPL-2.0-only OR MIT
+/*
+ * Apple SoC PMGR device power state driver
+ *
+ * Copyright The Asahi Linux Contributors
+ */
+
+#include <linux/bitops.h>
+#include <linux/bitfield.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/module.h>
+
+#define APPLE_CLKGEN_PSTATE 0
+#define APPLE_CLKGEN_PSTATE_DESIRED GENMASK(3, 0)
+
+#define SYS_DEV_PSTATE_SUSPEND 1
+
+enum sys_device {
+	DEV_FABRIC,
+	DEV_DCS,
+	DEV_MAX,
+};
+
+struct apple_pmgr_sys_device {
+	void __iomem *base;
+	u32 active_state;
+	u32 suspend_state;
+};
+
+struct apple_pmgr_misc_hw {
+	u32 dev_min_ps[DEV_MAX];
+};
+
+struct apple_pmgr_misc {
+	struct device *dev;
+	struct apple_pmgr_sys_device devices[DEV_MAX];
+};
+
+static void apple_pmgr_sys_dev_set_pstate(struct apple_pmgr_misc *misc,
+					  enum sys_device dev, bool active)
+{
+	u32 pstate;
+	u32 val;
+
+	if (!misc->devices[dev].base)
+		return;
+
+	if (active)
+		pstate = misc->devices[dev].active_state;
+	else
+		pstate = misc->devices[dev].suspend_state;
+
+	dev_dbg(misc->dev, "set %d ps to pstate %d\n", dev, pstate);
+
+	val = readl_relaxed(misc->devices[dev].base + APPLE_CLKGEN_PSTATE);
+	val &= ~APPLE_CLKGEN_PSTATE_DESIRED;
+	val |= FIELD_PREP(APPLE_CLKGEN_PSTATE_DESIRED, pstate);
+	writel_relaxed(val, misc->devices[dev].base);
+}
+
+static int __maybe_unused apple_pmgr_misc_suspend_noirq(struct device *dev)
+{
+	struct apple_pmgr_misc *misc = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < DEV_MAX; i++)
+		apple_pmgr_sys_dev_set_pstate(misc, i, false);
+
+	return 0;
+}
+
+static int __maybe_unused apple_pmgr_misc_resume_noirq(struct device *dev)
+{
+	struct apple_pmgr_misc *misc = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < DEV_MAX; i++)
+		apple_pmgr_sys_dev_set_pstate(misc, i, true);
+
+	return 0;
+}
+
+static bool apple_pmgr_init_device(struct apple_pmgr_misc *misc,
+				   const struct apple_pmgr_misc_hw *hw,
+				   enum sys_device dev,
+				   const char *device_name)
+{
+	void __iomem *base;
+	char name[32];
+	u32 val;
+
+	snprintf(name, sizeof(name), "%s-ps", device_name);
+
+	base = devm_platform_ioremap_resource_byname(
+		to_platform_device(misc->dev), name);
+	if (!base)
+		return false;
+
+	val = readl_relaxed(base + APPLE_CLKGEN_PSTATE);
+
+	misc->devices[dev].base = base;
+	misc->devices[dev].active_state =
+		FIELD_GET(APPLE_CLKGEN_PSTATE_DESIRED, val);
+	misc->devices[dev].suspend_state = hw->dev_min_ps[dev];
+
+	return true;
+}
+
+static int apple_pmgr_misc_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	const struct apple_pmgr_misc_hw *hw;
+	struct apple_pmgr_misc *misc;
+	int ret = -ENODEV;
+
+	misc = devm_kzalloc(dev, sizeof(*misc), GFP_KERNEL);
+	if (!misc)
+		return -ENOMEM;
+
+	misc->dev = dev;
+	hw = of_device_get_match_data(dev);
+
+	if (apple_pmgr_init_device(misc, hw, DEV_FABRIC, "fabric"))
+		ret = 0;
+
+	if (apple_pmgr_init_device(misc, hw, DEV_DCS, "dcs"))
+		ret = 0;
+
+	platform_set_drvdata(pdev, misc);
+
+	return ret;
+}
+
+static const struct apple_pmgr_misc_hw apple_pmgr_misc_hw_t600x = {
+	.dev_min_ps = {
+		[DEV_FABRIC] = SYS_DEV_PSTATE_SUSPEND,
+		[DEV_DCS] = 7,
+	},
+};
+
+static const struct apple_pmgr_misc_hw apple_pmgr_misc_hw_t602x = {
+	.dev_min_ps = {
+		[DEV_FABRIC] = SYS_DEV_PSTATE_SUSPEND,
+		[DEV_DCS] = SYS_DEV_PSTATE_SUSPEND,
+	},
+};
+
+static const struct of_device_id apple_pmgr_misc_of_match[] = {
+	{ .compatible = "apple,t6000-pmgr-misc", .data = &apple_pmgr_misc_hw_t600x },
+	{ .compatible = "apple,t6020-pmgr-misc", .data = &apple_pmgr_misc_hw_t602x },
+	{}
+};
+
+MODULE_DEVICE_TABLE(of, apple_pmgr_misc_of_match);
+
+static const struct dev_pm_ops apple_pmgr_misc_pm_ops = {
+	SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(apple_pmgr_misc_suspend_noirq,
+				      apple_pmgr_misc_resume_noirq)
+};
+
+static struct platform_driver apple_pmgr_misc_driver = {
+	.probe = apple_pmgr_misc_probe,
+	.driver = {
+		.name = "apple-pmgr-misc",
+		.of_match_table = apple_pmgr_misc_of_match,
+		.pm = pm_ptr(&apple_pmgr_misc_pm_ops),
+	},
+};
+
+MODULE_AUTHOR("Hector Martin <marcan@marcan.st>");
+MODULE_DESCRIPTION("PMGR misc driver for Apple SoCs");
+MODULE_LICENSE("GPL");
+
+module_platform_driver(apple_pmgr_misc_driver);

-- 
2.55.0

[PATCH 3/3] arm64: dts: apple: Add pmgr-misc nodes to t60xx

[Sasha Finkelstein]

Adds the PMGR misc control nodes for M1/2 Pro/Max/Ultra series devices.

Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>
---
 arch/arm64/boot/dts/apple/t600x-die0.dtsi | 7 +++++++
 arch/arm64/boot/dts/apple/t602x-die0.dtsi | 7 +++++++
 2 files changed, 14 insertions(+)

diff --git a/arch/arm64/boot/dts/apple/t600x-die0.dtsi b/arch/arm64/boot/dts/apple/t600x-die0.dtsi
index f715b19efd16..3271d234b483 100644
--- a/arch/arm64/boot/dts/apple/t600x-die0.dtsi
+++ b/arch/arm64/boot/dts/apple/t600x-die0.dtsi
@@ -24,6 +24,13 @@ aic: interrupt-controller@28e100000 {
 		power-domains = <&ps_aic>;
 	};
 
+	pmgr_misc: power-management@28e20c000 {
+		compatible = "apple,t6000-pmgr-misc";
+		reg = <0x2 0x8e20c000 0 0x400>,
+			<0x2 0x8e20c800 0 0x400>;
+		reg-names = "fabric-ps", "dcs-ps";
+	};
+
 	smc: smc@290400000 {
 		compatible = "apple,t6000-smc", "apple,smc";
 		reg = <0x2 0x90400000 0x0 0x4000>,
diff --git a/arch/arm64/boot/dts/apple/t602x-die0.dtsi b/arch/arm64/boot/dts/apple/t602x-die0.dtsi
index 8622ddea7b44..121f2cf75915 100644
--- a/arch/arm64/boot/dts/apple/t602x-die0.dtsi
+++ b/arch/arm64/boot/dts/apple/t602x-die0.dtsi
@@ -23,6 +23,13 @@ aic: interrupt-controller@28e100000 {
 		power-domains = <&ps_aic>;
 	};
 
+	pmgr_misc: power-management@28e20c000 {
+		compatible = "apple,t6020-pmgr-misc";
+		reg = <0x2 0x8e20c000 0 0x400>,
+			<0x2 0x8e20c400 0 0x400>;
+		reg-names = "fabric-ps", "dcs-ps";
+	};
+
 	nub_spmi0: spmi@29e114000 {
 		compatible = "apple,t6020-spmi", "apple,t8103-spmi";
 		reg = <0x2 0x9e114000 0x0 0x100>;

-- 
2.55.0

Re: [PATCH 2/3] soc: apple: Add driver for Apple PMGR misc controls

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] Incorrect check for devm_platform_ioremap_resource_byname() return value leads to ERR_PTR dereference
- [High] Missing NULL check for of_device_get_match_data() causes NULL pointer dereference
--

commit 144dd044ce23e2ad7967536cb164bb8942aedf81
Author: Hector Martin <marcan@marcan.st>

soc: apple: Add driver for Apple PMGR misc controls
    
This commit adds a driver for managing miscellaneous PMGR controls on Apple
SoCs, implementing fabric and memory controller power state switching on system
standby/s2idle to save power.

> diff --git a/drivers/soc/apple/apple-pmgr-misc.c b/drivers/soc/apple/apple-pmgr-misc.c
> new file mode 100644
> index 0000000000000..d487a87ecddb0
> --- /dev/null
> +++ b/drivers/soc/apple/apple-pmgr-misc.c
[ ... ]
> +static bool apple_pmgr_init_device(struct apple_pmgr_misc *misc,
> +				   const struct apple_pmgr_misc_hw *hw,
> +				   enum sys_device dev,
> +				   const char *device_name)
> +{
> +	void __iomem *base;
> +	char name[32];
> +	u32 val;
> +
> +	snprintf(name, sizeof(name), "%s-ps", device_name);
> +
> +	base = devm_platform_ioremap_resource_byname(
> +		to_platform_device(misc->dev), name);
> +	if (!base)

[Severity: High]
Does this code handle the return value of 
devm_platform_ioremap_resource_byname() correctly?

The function returns an error pointer on failure, not NULL. A failure here
would bypass this check and could lead to an ERR_PTR being dereferenced by
the readl_relaxed() call below.

> +		return false;
> +
> +	val = readl_relaxed(base + APPLE_CLKGEN_PSTATE);
> +
> +	misc->devices[dev].base = base;
> +	misc->devices[dev].active_state =
> +		FIELD_GET(APPLE_CLKGEN_PSTATE_DESIRED, val);
> +	misc->devices[dev].suspend_state = hw->dev_min_ps[dev];

[Severity: High]
Can hw be NULL here?

If the driver is bound manually via sysfs without a matching device tree node,
of_device_get_match_data() in apple_pmgr_misc_probe() below returns NULL. This
would lead to a NULL pointer dereference.

> +
> +	return true;
> +}
> +
> +static int apple_pmgr_misc_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	const struct apple_pmgr_misc_hw *hw;
> +	struct apple_pmgr_misc *misc;
> +	int ret = -ENODEV;
> +
> +	misc = devm_kzalloc(dev, sizeof(*misc), GFP_KERNEL);
> +	if (!misc)
> +		return -ENOMEM;
> +
> +	misc->dev = dev;
> +	hw = of_device_get_match_data(dev);
> +
> +	if (apple_pmgr_init_device(misc, hw, DEV_FABRIC, "fabric"))

[Severity: High]
Is it possible for hw to be NULL here?

Passing an unchecked hw pointer to apple_pmgr_init_device() causes the
unconditional dereference of hw->dev_min_ps[dev] shown above.

> +		ret = 0;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260702-pmgr-misc-v1-0-4f075a3a95c1@chaosmail.tech?part=2

Re: [PATCH 3/3] arm64: dts: apple: Add pmgr-misc nodes to t60xx

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Pre-existing error handling bug where the return value of `devm_platform_ioremap_resource_byname()` is incorrectly checked against NULL instead of IS_ERR, leading to an invalid pointer dereference.
--

commit 7d3a9b5b08e4210f6d81e35e862b055c17a322ad
Author: Sasha Finkelstein <k@chaosmail.tech>

arm64: dts: apple: Add pmgr-misc nodes to t60xx

This patch adds the PMGR misc control nodes to the device tree for the M1/2
Pro, Max, and Ultra series devices. It defines the register regions and
power domains for these power-management nodes.

> diff --git a/arch/arm64/boot/dts/apple/t600x-die0.dtsi b/arch/arm64/boot/dts/apple/t600x-die0.dtsi
> index f715b19efd167..3271d234b483a 100644
> --- a/arch/arm64/boot/dts/apple/t600x-die0.dtsi
> +++ b/arch/arm64/boot/dts/apple/t600x-die0.dtsi
> @@ -24,6 +24,13 @@ aic: interrupt-controller@28e100000 {
>  		power-domains = <&ps_aic>;
>  	};
>  
> +	pmgr_misc: power-management@28e20c000 {
> +		compatible = "apple,t6000-pmgr-misc";
> +		reg = <0x2 0x8e20c000 0 0x400>,
> +			<0x2 0x8e20c800 0 0x400>;
> +		reg-names = "fabric-ps", "dcs-ps";
> +	};
> +

[Severity: High]
This isn't a bug introduced by this patch, but adding this device tree node
will trigger the probe function in drivers/soc/apple/apple-pmgr-misc.c, which
has a pre-existing error handling issue.

Could this cause a kernel panic during boot if the resource fails to map?

In drivers/soc/apple/apple-pmgr-misc.c:apple_pmgr_init_device(), the
resource mapping is handled like this:

	base = devm_platform_ioremap_resource_byname(
		to_platform_device(misc->dev), name);
	if (!base)
		return false;

	val = readl_relaxed(base + APPLE_CLKGEN_PSTATE);

Should this check be IS_ERR(base) instead of !base?

If the resource mapping fails, devm_platform_ioremap_resource_byname()
returns an ERR_PTR (such as ERR_PTR(-EINVAL)). The !base check will evaluate
to false, allowing the error pointer to be directly dereferenced as a
memory address by readl_relaxed().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260702-pmgr-misc-v1-0-4f075a3a95c1@chaosmail.tech?part=3

Re: [PATCH 1/3] dt-bindings: soc: apple: Add bindings for apple PMGR misc controls

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 3293 bytes --]

On Thu, Jul 02, 2026 at 02:06:26PM +0200, Sasha Finkelstein wrote:
> Certain Apple SoCs include additional PMGR power states that are
> controlled via a different "misc" control block. On existing SoCs, this
> includes the fabric and memory controller state.
> 
> Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>
> ---
>  Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml | 45 +++++++++++++++++++++++++++++++++++++++++++++
>  MAINTAINERS                                                            |  1 +
>  2 files changed, 46 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml b/Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml
> new file mode 100644
> index 000000000000..30abedc67fa4
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml
> @@ -0,0 +1,45 @@
> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/soc/apple/apple,t6000-pmgr-misc.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: Apple SoC PMGR Misc Power States
> +
> +maintainers:
> +  - Sasha Finkelstein <k@chaosmail.tech>
> +
> +description: |
> +  Certain Apple SoCs include additional PMGR power states that are controlled
> +  via a different "misc" control block. On existing SoCs, this includes the
> +  fabric and memory controller state.

Are t6000 and t6020 the "certain Apple SoCs"? I assume so. The bit about
"existing SoCs" is confusing in a binding (although it would be ok in the
commit message), because you run into language semantics that imply that
this sentence is about devices to the ones you're addressing here. I
think you can just do s/. On existing SoCs//. And add an s to the final
"state" to make it "states".

> +
> +properties:
> +  compatible:
> +    enum:
> +      - apple,t6000-pmgr-misc
> +      - apple,t6020-pmgr-misc
> +
> +  reg:
> +    maxItems: 2
> +
> +  reg-names:
> +    items:
> +      - const: fabric-ps
> +      - const: dcs-ps

"dcs" means memory controller then?

> +
> +required:
> +  - compatible
> +  - reg
> +  - reg-names
> +
> +additionalProperties: false
> +
> +examples:
> +  - |
> +    pmgr_misc: power-management@8e20c000 {

If there's a new version, drop the unused label here please.

> +        compatible = "apple,t6000-pmgr-misc";
> +        reg = <0x8e20c000 0x400>,
> +              <0x8e20c800 0x400>;
> +        reg-names = "fabric-ps", "dcs-ps";
> +    };
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 15011f5752a9..5a29bb86499f 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -2607,6 +2607,7 @@ F:	Documentation/devicetree/bindings/power/apple*
>  F:	Documentation/devicetree/bindings/power/reset/apple,smc-reboot.yaml
>  F:	Documentation/devicetree/bindings/pwm/apple,s5l-fpwm.yaml
>  F:	Documentation/devicetree/bindings/rtc/apple,smc-rtc.yaml
> +F:	Documentation/devicetree/bindings/soc/apple/apple,t6000-pmgr-misc.yaml
>  F:	Documentation/devicetree/bindings/spi/apple,spi.yaml
>  F:	Documentation/devicetree/bindings/spmi/apple,spmi.yaml
>  F:	Documentation/devicetree/bindings/usb/apple,dwc3.yaml
> 
> -- 
> 2.55.0
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH 1/3] dt-bindings: soc: apple: Add bindings for apple PMGR misc controls

[Sasha Finkelstein]

On Jul 2, 2026, at 21:18, Conor Dooley <conor@kernel.org> wrote:
> Are t6000 and t6020 the "certain Apple SoCs"? I assume so. The bit about
> "existing SoCs" is confusing in a binding (although it would be ok in the
> commit message), because you run into language semantics that imply that
> this sentence is about devices to the ones you're addressing here. I
> think you can just do s/. On existing SoCs//. And add an s to the final
> "state" to make it "states".

t600{0,1,2}, t602{0,1,2}, t603{1,2,4} (no upstream DT for those yet)
It is kind of a long list, with most t6xxx ones on it.

> "dcs" means memory controller then?

It is memory controller related, unclear if this covers the entire
controller or a subset of it. The name comes from Apple device trees
and macos kernel logs

Re: [PATCH 1/3] dt-bindings: soc: apple: Add bindings for apple PMGR misc controls

[Krzysztof Kozlowski]

On Thu, Jul 02, 2026 at 02:06:26PM +0200, Sasha Finkelstein wrote:
> Certain Apple SoCs include additional PMGR power states that are
> controlled via a different "misc" control block. On existing SoCs, this
> includes the fabric and memory controller state.
> 
> Signed-off-by: Sasha Finkelstein <k@chaosmail.tech>

Just a nit on top of Conor's review, if there is a new version:

A nit, subject: drop second/last, redundant "bindings for". The
"dt-bindings" prefix is already stating that these are bindings.
See also:
https://elixir.bootlin.com/linux/v7.1-rc7/source/Documentation/devicetree/bindings/submitting-patches.rst#L23

Best regards,
Krzysztof