Skip to contentSkip to conversation

i.MX6 PU power domain support

PATCH Superseded Philipp Zabel 4 participants 13 messages Aug 26, 2014 – Sep 10, 2014

Subsystems (4)

v7 of 9 Superseded Latest is v9 →

lore
Share & tools
lore
#0 lore raw

v7 cover i.MX6 PU power domain support

Hi,

this is a rebased and updated version of the series I sent previously:

http://lists.infradead.org/pipermail/linux-arm-kernel/2014-March/238377.html

It is based on the "PM / Domains: Generic OF-based support" series that Ulf Hansson sent recently:

http://www.spinics.net/lists/arm-kernel/msg357003.html

Changes since v6:
- imx6qdl.dtsi entry now uses clock name macros instead of index numbers

regards
Philipp

Philipp Zabel (5):
Documentation: Add device tree bindings for Freescale i.MX GPC
ARM: imx6: gpc: Add PU power domain for GPU/VPU
ARM: dts: imx6qdl: Add power-domain information to gpc node
ARM: dts: imx6sl: Add power-domain information to gpc node
ARM: dts: imx6qdl: Allow disabling the PU regulator, add a enable ramp
delay

 .../devicetree/bindings/power/fsl,imx-gpc.txt      |  54 ++++++
 arch/arm/boot/dts/imx6qdl.dtsi                     |  11 +-
 arch/arm/boot/dts/imx6sl.dtsi                      |   4 +
 arch/arm/mach-imx/Kconfig                          |   1 +
 arch/arm/mach-imx/gpc.c                            | 203 +++++++++++++++++++++
 5 files changed, 272 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/power/fsl,imx-gpc.txt
Signature · 1 line
2.1.0.rc1
#1 lore raw

PATCH v7 1/5 Documentation: Add device tree bindings for Freescale i.MX GPC

The i.MX6 contains a power controller that controls power gating and sequencing for the SoC's power domains.

---
 .../devicetree/bindings/power/fsl,imx-gpc.txt      | 54 ++++++++++++++++++++++
 1 file changed, 54 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/power/fsl,imx-gpc.txt
Documentation/devicetree/bindings/power/fsl,imx-gpc.txt new file +54 -0
new file mode 100644
index 0000000..eaa8c93
@@ -0,0 +1,54 @@
+Freescale i.MX General Power Controller
+=======================================
+
+The i.MX6Q General Power Control (GPC) block contains DVFS load tracking
+counters and Power Gating Control (PGC) for the CPU and PU (GPU/VPU) power
+domains.
+
+Required properties:
+- compatible: Should be "fsl,imx6q-gpc" or "fsl,imx6sl-gpc"
+- reg: should be register base and length as documented in the
+  datasheet
+- interrupts: Should contain GPC interrupt request 1
+- pu-supply: Link to the LDO regulator powering the PU power domain
+- clocks: Clock phandles to devices in the PU power domain that need
+	  to be enabled during domain power-up for reset propagation.
+- #power-domain-cells: Should be 1, see below:
+
+The gpc node is a power-controller as documented by the generic power domain
+bindings in Documentation/devicetree/bindings/power/power_domain.txt.
+
+Example:
+
+	gpc: gpc at 020dc000 {
+		compatible = "fsl,imx6q-gpc";
+		reg = <0x020dc000 0x4000>;
+		interrupts = <0 89 0x04 0 90 0x04>;
+		pu-supply = <&reg_pu>;
+		clocks = <&clks 122>, <&clks 74>, <&clks 121>,
+			 <&clks 26>, <&clks 143>, <&clks 168>;
+		#power-domain-cells = <1>;
+	};
+
+
+Specifying power domain for IP modules
+======================================
+
+IP cores belonging to a power domain should contain a 'power-domain' property
+that is a phandle pointing to the gpc device node and a DOMAIN_INDEX specifying
+the power domain the device belongs to.
+
+Example of a device that is part of the PU power domain:
+
+	vpu: vpu at 02040000 {
+		reg = <0x02040000 0x3c000>;
+		/* ... */
+		power-domain = <&gpc 1>;
+		/* ... */
+	};
+
+The following DOMAIN_INDEX values are valid for i.MX6Q:
+ARM_DOMAIN     0
+PU_DOMAIN      1
+The following additional DOMAIN_INDEX value is valid for i.MX6SL:
+DISPLAY_DOMAIN 2
Signature · 1 line
2.1.0.rc1
#2 lore raw

PATCH v7 1/5 Documentation: Add device tree bindings for Freescale i.MX GPC

Hi Philipp,

Review

Quoted 4 lines · Philipp Zabel

+++ b/Documentation/devicetree/bindings/power/fsl,imx-gpc.txt

(quoted diff) +1 -0
@@ -0,0 +1,54 @@
 
+IP cores belonging to a power domain should contain a 'power-domain' property

power-domains

Documentation/devicetree/bindings/power/fsl,imx-gpc.txt line 46

Quoted 9 lines
Documentation/devicetree/bindings/power/fsl,imx-gpc.txt new file +9 -0
+that is a phandle pointing to the gpc device node and a DOMAIN_INDEX specifying
+the power domain the device belongs to.
+
+Example of a device that is part of the PU power domain:
+
+       vpu: vpu at 02040000 {
+               reg = <0x02040000 0x3c000>;
+               /* ... */
+               power-domain = <&gpc 1>;

power-domains

Review

Quoted 2 lines
(quoted diff) +2 -0
+               /* ... */
+       };

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#3 lore raw

PATCH v7 1/5 Documentation: Add device tree bindings for Freescale i.MX GPC

Hi Geert,

Am Freitag, den 05.09.2014, 13:47 +0200 schrieb Geert Uytterhoeven:

Quoted 10 lines

Hi Philipp,

Review

Quoted 4 lines · Philipp Zabel

+++ b/Documentation/devicetree/bindings/power/fsl,imx-gpc.txt

(quoted diff) +1 -0
@@ -0,0 +1,54 @@
 
+IP cores belonging to a power domain should contain a 'power-domain' property

power-domains

Quoted 11 lines

Documentation/devicetree/bindings/power/fsl,imx-gpc.txt line 46

Quoted 9 lines
Documentation/devicetree/bindings/power/fsl,imx-gpc.txt new file +9 -0
+that is a phandle pointing to the gpc device node and a DOMAIN_INDEX specifying
+the power domain the device belongs to.
+
+Example of a device that is part of the PU power domain:
+
+       vpu: vpu at 02040000 {
+               reg = <0x02040000 0x3c000>;
+               /* ... */
+               power-domain = <&gpc 1>;

power-domains

Thanks, I'll change those.

regards
Philipp

#4 lore raw

PATCH v7 2/5 ARM: imx6: gpc: Add PU power domain for GPU/VPU

In reply to Philipp Zabel

When generic pm domain support is enabled, the PGC can be used to completely gate power to the PU power domain containing GPU3D, GPU2D, and VPU cores.
This code triggers the PGC powerdown sequence to disable the GPU/VPU isolation cells and gate power and then disables the PU regulator. To reenable, the reverse powerup sequence is triggered after the PU regulator is enabled again.
The GPU and VPU devices in the PU power domain temporarily need to be clocked during powerup, so that the reset machinery can work.

---
 arch/arm/mach-imx/Kconfig |   1 +
 arch/arm/mach-imx/gpc.c   | 203 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 204 insertions(+)
arch/arm/mach-imx/Kconfig +1 -0
index 9de84a2..78d69cf 100644
@@ -50,6 +50,7 @@ config HAVE_IMX_ANATOP
 
 config HAVE_IMX_GPC
 	bool
+	select PM_GENERIC_DOMAINS if PM
 
 config HAVE_IMX_MMDC
 	bool
arch/arm/mach-imx/gpc.c +203 -0

225 lines not on this page · raw message

Signature · 1 line
2.1.0.rc1
#5 lore raw

PATCH v7 2/5 ARM: imx6: gpc: Add PU power domain for GPU/VPU

Quoted 75 lines · Philipp Zabel

When generic pm domain support is enabled, the PGC can be used to completely gate power to the PU power domain containing GPU3D, GPU2D, and VPU cores.
This code triggers the PGC powerdown sequence to disable the GPU/VPU isolation cells and gate power and then disables the PU regulator. To reenable, the reverse powerup sequence is triggered after the PU regulator is enabled again.
The GPU and VPU devices in the PU power domain temporarily need to be clocked during powerup, so that the reset machinery can work.

---
 arch/arm/mach-imx/Kconfig |   1 +
 arch/arm/mach-imx/gpc.c   | 203 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 204 insertions(+)
arch/arm/mach-imx/Kconfig +1 -0
index 9de84a2..78d69cf 100644
@@ -50,6 +50,7 @@ config HAVE_IMX_ANATOP
 
 config HAVE_IMX_GPC
        bool
+       select PM_GENERIC_DOMAINS if PM
 
 config HAVE_IMX_MMDC
        bool
arch/arm/mach-imx/gpc.c +22 -0
index 82ea74e..5eec828 100644
@@ -10,19 +10,41 @@
  * http://www.gnu.org/copyleft/gpl.html
  */
 
+#include <linux/clk.h>
+#include <linux/delay.h>
 #include <linux/io.h>
 #include <linux/irq.h>
 #include <linux/of.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
+#include <linux/platform_device.h>
+#include <linux/pm_domain.h>
+#include <linux/regulator/consumer.h>
 #include <linux/irqchip/arm-gic.h>
 #include "common.h"
+#include "hardware.h"
 
+#define GPC_CNTR               0x000
 #define GPC_IMR1               0x008
+#define GPC_PGC_GPU_PDN                0x260
+#define GPC_PGC_GPU_PUPSCR     0x264
+#define GPC_PGC_GPU_PDNSCR     0x268
 #define GPC_PGC_CPU_PDN                0x2a0
 
 #define IMR_NUM                        4
 
+#define GPU_VPU_PUP_REQ                BIT(1)
+#define GPU_VPU_PDN_REQ                BIT(0)
+
+#define GPC_CLK_MAX            6
+
+struct pu_domain {
+       struct generic_pm_domain base;
+       struct regulator *reg;
+       struct clk *clk[GPC_CLK_MAX];
+       int num_clks;
+};
+
 static void __iomem *gpc_base;
 static u32 gpc_wake_irqs[IMR_NUM];
 static u32 gpc_saved_imrs[IMR_NUM];

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 165

Quoted 5 lines
arch/arm/mach-imx/gpc.c +2 -0
@@ -139,3 +161,184 @@ void __init imx_gpc_init(void)
        gic_arch_extn.irq_unmask = imx_gpc_irq_unmask;
        gic_arch_extn.irq_set_wake = imx_gpc_irq_set_wake;
 }
+
+#ifdef CONFIG_PM

Hi Philipp,

Should this be CONFIG_PM_GENERIC_DOMAINS?

Quoted 110 lines
arch/arm/mach-imx/gpc.c +110 -0
+
+static int imx6q_pm_pu_power_off(struct generic_pm_domain *genpd)
+{
+       struct pu_domain *pu = container_of(genpd, struct pu_domain, base);
+       int iso, iso2sw;
+       u32 val;
+
+       /* Read ISO and ISO2SW power down delays */
+       val = readl_relaxed(gpc_base + GPC_PGC_GPU_PDNSCR);
+       iso = val & 0x3f;
+       iso2sw = (val >> 8) & 0x3f;
+
+       /* Gate off PU domain when GPU/VPU when powered down */
+       writel_relaxed(0x1, gpc_base + GPC_PGC_GPU_PDN);
+
+       /* Request GPC to power down GPU/VPU */
+       val = readl_relaxed(gpc_base + GPC_CNTR);
+       val |= GPU_VPU_PDN_REQ;
+       writel_relaxed(val, gpc_base + GPC_CNTR);
+
+       /* Wait ISO + ISO2SW IPG clock cycles */
+       ndelay((iso + iso2sw) * 1000 / 66);
+
+       regulator_disable(pu->reg);
+
+       return 0;
+}
+
+static int imx6q_pm_pu_power_on(struct generic_pm_domain *genpd)
+{
+       struct pu_domain *pu = container_of(genpd, struct pu_domain, base);
+       int i, ret, sw, sw2iso;
+       u32 val;
+
+       ret = regulator_enable(pu->reg);
+       if (ret) {
+               pr_err("%s: failed to enable regulator: %d\n", __func__, ret);
+               return ret;
+       }
+
+       /* Enable reset clocks for all devices in the PU domain */
+       for (i = 0; i < pu->num_clks; i++)
+               clk_prepare_enable(pu->clk[i]);
+
+       /* Gate off PU domain when GPU/VPU when powered down */
+       writel_relaxed(0x1, gpc_base + GPC_PGC_GPU_PDN);
+
+       /* Read ISO and ISO2SW power down delays */
+       val = readl_relaxed(gpc_base + GPC_PGC_GPU_PUPSCR);
+       sw = val & 0x3f;
+       sw2iso = (val >> 8) & 0x3f;
+
+       /* Request GPC to power up GPU/VPU */
+       val = readl_relaxed(gpc_base + GPC_CNTR);
+       val |= GPU_VPU_PUP_REQ;
+       writel_relaxed(val, gpc_base + GPC_CNTR);
+
+       /* Wait ISO + ISO2SW IPG clock cycles */
+       ndelay((sw + sw2iso) * 1000 / 66);
+
+       /* Disable reset clocks for all devices in the PU domain */
+       for (i = 0; i < pu->num_clks; i++)
+               clk_disable_unprepare(pu->clk[i]);
+
+       return 0;
+}
+
+static struct generic_pm_domain imx6q_arm_domain = {
+       .name = "ARM",
+};
+
+static struct pu_domain imx6q_pu_domain = {
+       .base = {
+               .name = "PU",
+               .power_off = imx6q_pm_pu_power_off,
+               .power_on = imx6q_pm_pu_power_on,
+               .power_off_latency_ns = 25000,
+               .power_on_latency_ns = 2000000,
+       },
+};
+
+static struct generic_pm_domain imx6sl_display_domain = {
+       .name = "DISPLAY",
+};
+
+static struct generic_pm_domain *imx_gpc_domains[] = {
+       &imx6q_arm_domain,
+       &imx6q_pu_domain.base,
+       &imx6sl_display_domain,
+};
+
+static struct genpd_onecell_data imx_gpc_onecell_data = {
+       .domains = imx_gpc_domains,
+       .domain_num = ARRAY_SIZE(imx_gpc_domains),
+};
+
+static int imx_gpc_genpd_init(struct device *dev, struct regulator *pu_reg)
+{
+       struct clk *clk;
+       bool is_off;
+       int i;
+
+       imx6q_pu_domain.base.of_node = dev->of_node;
+       imx6q_pu_domain.reg = pu_reg;
+
+       for (i = 0; ; i++) {
+               clk = of_clk_get(dev->of_node, i);
+               if (IS_ERR(clk))
+                       break;
+               if (i >= GPC_CLK_MAX) {

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 285

Quoted 10 lines
arch/arm/mach-imx/gpc.c +10 -0
+                       dev_err(dev, "more than %d clocks\n", GPC_CLK_MAX);
+                       goto clk_err;
+               }
+               imx6q_pu_domain.clk[i] = clk;
+       }
+       imx6q_pu_domain.num_clks = i;
+
+       is_off = IS_ENABLED(CONFIG_PM_RUNTIME);
+       if (is_off)
+               imx6q_pm_pu_power_off(&imx6q_pu_domain.base);

Could you elaborate why is_off depends on CONFIG_PM_RUNTIME? That seems strange to me. :-)

Additionally, I think it would be better to leave the domain "on" state. Wouldn't that even be required for some drivers to be able to succeed probing of these devices?

Quoted 49 lines
arch/arm/mach-imx/gpc.c +49 -0
+
+       pm_genpd_init(&imx6q_pu_domain.base, NULL, is_off);
+       return of_genpd_add_provider(dev->of_node, of_genpd_xlate_onecell,
+                                    &imx_gpc_onecell_data);
+
+clk_err:
+       while (i--)
+               clk_put(imx6q_pu_domain.clk[i]);
+       return -EINVAL;
+}
+
+#else
+static inline int imx_gpc_genpd_init(struct device *dev, struct regulator *reg)
+{
+       return 0;
+}
+#endif /* CONFIG_PM */
+
+static int imx_gpc_probe(struct platform_device *pdev)
+{
+       struct regulator *pu_reg;
+       int ret;
+
+       pu_reg = devm_regulator_get(&pdev->dev, "pu");
+       if (IS_ERR(pu_reg)) {
+               ret = PTR_ERR(pu_reg);
+               dev_err(&pdev->dev, "failed to get pu regulator: %d\n", ret);
+               return ret;
+       }
+
+       /* The regulator is initially enabled */
+       ret = regulator_enable(pu_reg);
+       if (ret < 0) {
+               dev_err(&pdev->dev, "failed to enable pu regulator: %d\n", ret);
+               return ret;
+       }
+       return imx_gpc_genpd_init(&pdev->dev, pu_reg);
+}
+
+static struct of_device_id imx_gpc_dt_ids[] = {
+       { .compatible = "fsl,imx6q-gpc" },
+       { .compatible = "fsl,imx6sl-gpc" },
+       { }
+};
+
+static struct platform_driver imx_gpc_driver = {
+       .driver = {
+               .name = "imx-gpc",
+               .owner = THIS_MODULE,

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 344

Quoted 10 lines
arch/arm/mach-imx/gpc.c +10 -0
+               .of_match_table = imx_gpc_dt_ids,
+       },
+       .probe = imx_gpc_probe,
+};
+
+static int __init imx_pgc_init(void)
+{
+       return platform_driver_register(&imx_gpc_driver);
+}
+subsys_initcall(imx_pgc_init);

Could initialization of the generic power domain at subsys_init level, mean that some other drivers may already be probed?

Are you sure none of those drivers handles devices within your power domain, thus causing attachment of the power domain to fail?

Kind regards
Uffe

Quoted 3 lines

--
2.1.0.rc1

#6 lore raw

PATCH v7 2/5 ARM: imx6: gpc: Add PU power domain for GPU/VPU

Hi Ulf,

thank you for the comments.

Am Dienstag, den 26.08.2014, 21:48 +0200 schrieb Ulf Hansson:
[...]

Quoted 10 lines

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 165

Quoted 6 lines
arch/arm/mach-imx/gpc.c +2 -0
@@ -139,3 +161,184 @@ void __init imx_gpc_init(void)
        gic_arch_extn.irq_unmask = imx_gpc_irq_unmask;
        gic_arch_extn.irq_set_wake = imx_gpc_irq_set_wake;
 }
+
+#ifdef CONFIG_PM

Hi Philipp,

Should this be CONFIG_PM_GENERIC_DOMAINS?

Yes, I'll change this.

[...]

Quoted 28 lines
Quoted 14 lines
arch/arm/mach-imx/gpc.c +14 -0
+static int imx_gpc_genpd_init(struct device *dev, struct regulator *pu_reg)
+{
+       struct clk *clk;
+       bool is_off;
+       int i;
+
+       imx6q_pu_domain.base.of_node = dev->of_node;
+       imx6q_pu_domain.reg = pu_reg;
+
+       for (i = 0; ; i++) {
+               clk = of_clk_get(dev->of_node, i);
+               if (IS_ERR(clk))
+                       break;
+               if (i >= GPC_CLK_MAX) {

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 285

Quoted 10 lines
arch/arm/mach-imx/gpc.c +10 -0
+                       dev_err(dev, "more than %d clocks\n", GPC_CLK_MAX);
+                       goto clk_err;
+               }
+               imx6q_pu_domain.clk[i] = clk;
+       }
+       imx6q_pu_domain.num_clks = i;
+
+       is_off = IS_ENABLED(CONFIG_PM_RUNTIME);
+       if (is_off)
+               imx6q_pm_pu_power_off(&imx6q_pu_domain.base);

Could you elaborate why is_off depends on CONFIG_PM_RUNTIME? That seems strange to me. :-)

Quoted 3 lines

Additionally, I think it would be better to leave the domain "on" state. Wouldn't that even be required for some drivers to be able to succeed probing of these devices?

The devices in the PU power domain are Vivante GPUs and the CODA Video Processing Unit. These are platform devices that don't need to be active to enumerate before being probed. The drivers support runtime PM, so if CONFIG_PM_RUNTIME is enabled, it is safe to disable the PU power domain on boot. The drivers will probe and request for the power domain to be enabled via runtime PM before accessing the hardware.

If CONFIG_PM_RUNTIME is disabled, the power domain needs to stay enabled at all times.

[...]

Quoted 18 lines
Quoted 4 lines
arch/arm/mach-imx/gpc.c +4 -0
+static struct platform_driver imx_gpc_driver = {
+       .driver = {
+               .name = "imx-gpc",
+               .owner = THIS_MODULE,

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 344

Quoted 10 lines
arch/arm/mach-imx/gpc.c +10 -0
+               .of_match_table = imx_gpc_dt_ids,
+       },
+       .probe = imx_gpc_probe,
+};
+
+static int __init imx_pgc_init(void)
+{
+       return platform_driver_register(&imx_gpc_driver);
+}
+subsys_initcall(imx_pgc_init);

Could initialization of the generic power domain at subsys_init level, mean that some other drivers may already be probed?

Quoted 2 lines

Are you sure none of those drivers handles devices within your power domain, thus causing attachment of the power domain to fail?

The GPU and VPU drivers are standard module_platform_drivers, those will be probed later at device_initcall times.

regards
Philipp

#7 lore raw

PATCH v7 2/5 ARM: imx6: gpc: Add PU power domain for GPU/VPU

[...]

Quoted 18 lines
Quoted 10 lines

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 285

Quoted 3 lines
arch/arm/mach-imx/gpc.c +3 -0
+       is_off = IS_ENABLED(CONFIG_PM_RUNTIME);
+       if (is_off)
+               imx6q_pm_pu_power_off(&imx6q_pu_domain.base);

Could you elaborate why is_off depends on CONFIG_PM_RUNTIME? That seems strange to me. :-)

Additionally, I think it would be better to leave the domain "on" state. Wouldn't that even be required for some drivers to be able to succeed probing of these devices?

The devices in the PU power domain are Vivante GPUs and the CODA Video Processing Unit. These are platform devices that don't need to be active to enumerate before being probed. The drivers support runtime PM, so if CONFIG_PM_RUNTIME is enabled, it is safe to disable the PU power domain on boot. The drivers will probe and request for the power domain to be enabled via runtime PM before accessing the hardware.

Quoted 2 lines

If CONFIG_PM_RUNTIME is disabled, the power domain needs to stay enabled at all times.

Hi Philipp,

Thanks for your reply, that certainly made it clear on what you would like to accomplish.

Before discussing further consequences in genpd of this approach (using IS_ENABLED(CONFIG_PM_RUNTIME)), I would be interested to see how the runtime PM support is implemented for any of the drivers you refer to. Do you mind pointing me to any of them?

Kind regards
Uffe

#8 lore raw

PATCH v7 2/5 ARM: imx6: gpc: Add PU power domain for GPU/VPU

Quoted 25 lines · Ulf Hansson

[...]

Quoted 18 lines
Quoted 10 lines

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 285

Quoted 3 lines
arch/arm/mach-imx/gpc.c +3 -0
+       is_off = IS_ENABLED(CONFIG_PM_RUNTIME);
+       if (is_off)
+               imx6q_pm_pu_power_off(&imx6q_pu_domain.base);

Could you elaborate why is_off depends on CONFIG_PM_RUNTIME? That seems strange to me. :-)

Additionally, I think it would be better to leave the domain "on" state. Wouldn't that even be required for some drivers to be able to succeed probing of these devices?

The devices in the PU power domain are Vivante GPUs and the CODA Video Processing Unit. These are platform devices that don't need to be active to enumerate before being probed. The drivers support runtime PM, so if CONFIG_PM_RUNTIME is enabled, it is safe to disable the PU power domain on boot. The drivers will probe and request for the power domain to be enabled via runtime PM before accessing the hardware.

Quoted 2 lines

If CONFIG_PM_RUNTIME is disabled, the power domain needs to stay enabled at all times.

Hi Philipp,

Quoted 7 lines

Thanks for your reply, that certainly made it clear on what you would like to accomplish.

Before discussing further consequences in genpd of this approach (using IS_ENABLED(CONFIG_PM_RUNTIME)), I would be interested to see how the runtime PM support is implemented for any of the drivers you refer to. Do you mind pointing me to any of them?

Of course, the CODA driver is at drivers/media/platform/coda.c in mainline kernels (moved to drivers/media/platform/coda-common.c in linux-next).

regards
Philipp

#9 lore raw

PATCH v7 2/5 ARM: imx6: gpc: Add PU power domain for GPU/VPU

Quoted 29 lines · Philipp Zabel
Quoted 28 lines · Ulf Hansson

[...]

Quoted 18 lines
Quoted 10 lines

arch/arm/mach-imx/gpc.c void __init imx_gpc_init(void) · line 285

Quoted 3 lines
arch/arm/mach-imx/gpc.c +3 -0
+       is_off = IS_ENABLED(CONFIG_PM_RUNTIME);
+       if (is_off)
+               imx6q_pm_pu_power_off(&imx6q_pu_domain.base);

Could you elaborate why is_off depends on CONFIG_PM_RUNTIME? That seems strange to me. :-)

Additionally, I think it would be better to leave the domain "on" state. Wouldn't that even be required for some drivers to be able to succeed probing of these devices?

The devices in the PU power domain are Vivante GPUs and the CODA Video Processing Unit. These are platform devices that don't need to be active to enumerate before being probed. The drivers support runtime PM, so if CONFIG_PM_RUNTIME is enabled, it is safe to disable the PU power domain on boot. The drivers will probe and request for the power domain to be enabled via runtime PM before accessing the hardware.

Quoted 2 lines

If CONFIG_PM_RUNTIME is disabled, the power domain needs to stay enabled at all times.

Hi Philipp,

Thanks for your reply, that certainly made it clear on what you would like to accomplish.

Quoted 7 lines
Quoted 4 lines

Before discussing further consequences in genpd of this approach (using IS_ENABLED(CONFIG_PM_RUNTIME)), I would be interested to see how the runtime PM support is implemented for any of the drivers you refer to. Do you mind pointing me to any of them?

Of course, the CODA driver is at drivers/media/platform/coda.c in mainline kernels (moved to drivers/media/platform/coda-common.c in linux-next).

Thanks! I went for a look and found some parts that concerns me. :-)

In principle, the coda platform driver will have a close dependency to its corresponding PM domain state (powered or not powered) while probing, which seems fragile.

In the case of CONFIG_PM_RUNTIME, the driver will not bring the coda device into active state (runtime PM resumed) during probing. As you also stated above, but this have consequences. Let me elaborate.

When the device gets attached to its PM domain, which is prior the coda driver is probed - the device specific flag in genpd, "need_restore", get assigned to a value depending on what status the PM domain currently has. The "need_restore" flag is used internally in genpd to understand when the device's runtime PM callbacks shall or shall not be invoked.

As long as the state of the PM domain is powered off, when it gets attached to the coda device - this works. Can you guarantee that's true for all SoCs and PM domains that uses the coda driver?

It won't work if the PM domain is powered on while attaching occurs. That's because the device's runtime PM state will not be in sync with the "need_restore" flag in genpd. In other words, genpd will prevent the coda driver's runtime PM callbacks from being invoked properly.

So this is a generic problem while using genpd and there are similar issues for the opposite situation. I am already working on a fix for it and will keep you posted. Let's see if that fix may effect this patch as well.

Kind regards
Uffe

#10 lore raw

PATCH v7 3/5 ARM: dts: imx6qdl: Add power-domain information to gpc node

In reply to Philipp Zabel

The PGC that is part of GPC controls isolation and power sequencing of the power domains. The PU power domain will be handled by the generic pm domain framework. It needs a phandle to the PU regulator to turn off power when the domain is disabled, and a list of phandles to all clocks that must be enabled during powerup for reset propagation.

---
Changes since v5:
- Replaced clock indices with name macros

---
 arch/arm/boot/dts/imx6qdl.dtsi | 8 ++++++++
 1 file changed, 8 insertions(+)
arch/arm/boot/dts/imx6qdl.dtsi +8 -0
index c701af9..0b1384c 100644
@@ -670,6 +670,14 @@
 				reg = <0x020dc000 0x4000>;
 				interrupts = <0 89 IRQ_TYPE_LEVEL_HIGH>,
 					     <0 90 IRQ_TYPE_LEVEL_HIGH>;
+				pu-supply = <&reg_pu>;
+				clocks = <&clks IMX6QDL_CLK_GPU3D_CORE>,
+					 <&clks IMX6QDL_CLK_GPU3D_SHADER>,
+					 <&clks IMX6QDL_CLK_GPU2D_CORE>,
+					 <&clks IMX6QDL_CLK_GPU2D_AXI>,
+					 <&clks IMX6QDL_CLK_OPENVG_AXI>,
+					 <&clks IMX6QDL_CLK_VPU_AXI>;
+				#power-domain-cells = <1>;
 			};
 
 			gpr: iomuxc-gpr at 020e0000 {
Signature · 1 line
2.1.0.rc1
#11 lore raw

PATCH v7 4/5 ARM: dts: imx6sl: Add power-domain information to gpc node

In reply to Philipp Zabel

The PGC that is part of GPC controls isolation and power sequencing of the power domains. The PU power domain will be handled by the generic pm domain framework. It needs a phandle to the PU regulator to turn off power when the domain is disabled and a list of clocks to be enabled during powerup for reset propagation.

---
 arch/arm/boot/dts/imx6sl.dtsi | 4 ++++
 1 file changed, 4 insertions(+)
arch/arm/boot/dts/imx6sl.dtsi +4 -0
index c75800c..7ae4847 100644
@@ -581,6 +581,10 @@
 				compatible = "fsl,imx6sl-gpc", "fsl,imx6q-gpc";
 				reg = <0x020dc000 0x4000>;
 				interrupts = <0 89 IRQ_TYPE_LEVEL_HIGH>;
+				pu-supply = <&reg_pu>;
+				clocks = <&clks IMX6SL_CLK_GPU2D_OVG>,
+					 <&clks IMX6SL_CLK_GPU2D_PODF>;
+				#power-domain-cells = <1>;
 			};
 
 			gpr: iomuxc-gpr at 020e0000 {
Signature · 1 line
2.1.0.rc1
#12 lore raw

PATCH v7 5/5 ARM: dts: imx6qdl: Allow disabling the PU regulator, add a enable ramp delay

In reply to Philipp Zabel

The PU regulator is enabled during boot, but not necessarily always-on. It can be disabled by the generic pm domain framework when the PU power domain is shut down. The ramp delay of 150 us might be a bit conservative, the value is taken from the Freescale kernel.

---
 arch/arm/boot/dts/imx6qdl.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/arm/boot/dts/imx6qdl.dtsi +2 -1
index 0b1384c..d8d3d93 100644
@@ -579,7 +579,8 @@
 					regulator-name = "vddpu";
 					regulator-min-microvolt = <725000>;
 					regulator-max-microvolt = <1450000>;
-					regulator-always-on;
+					regulator-enable-ramp-delay = <150>;
+					regulator-boot-on;
 					anatop-reg-offset = <0x140>;
 					anatop-vol-bit-shift = <9>;
 					anatop-vol-bit-width = <5>;
Signature · 1 line
2.1.0.rc1