* [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains @ 2017-10-28 16:40 Neil Armstrong 2017-10-28 16:40 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong 2017-10-28 16:40 ` [RESEND PATCH v2 2/2] dt-bindings: power: add amlogic meson power domain bindings Neil Armstrong 0 siblings, 2 replies; 8+ messages in thread From: Neil Armstrong @ 2017-10-28 16:40 UTC (permalink / raw) To: linux-arm-kernel [Resent with *correct* linux-pm in CC] On the Amlogic Gx SoCs (GXBB, GXL & GXM), the VPU power domain is initialized by the vendor U-Boot code, but running mainline U-boot has been possible on these SoCs. But lacking such init made the system lock at kernel boot. This patchset adds the Video Processing Unit power domain driver to enable the same power-on and power-down sequences and was designed to allow booting with the power domain already initialized or not. This driver has been tested on : - Odroid-C2 (GXBB) with Vendor and Mainline U-Boot - P212 (GXL) with Vendor and Mainline U-Boot - Khadas Vim (GXL) with Vendor U-Boot - Khadas Vim2 with (GXM) Vendor U-Boot Changes since v1: - fixed bindings compatible string Neil Armstrong (2): soc: amlogic: add Meson GX VPU Domains driver dt-bindings: power: add amlogic meson power domain bindings .../bindings/power/amlogic,meson-gx-pwrc.txt | 61 ++++++ drivers/soc/amlogic/Kconfig | 10 + drivers/soc/amlogic/Makefile | 1 + drivers/soc/amlogic/meson-gx-pwrc-vpu.c | 234 +++++++++++++++++++++ 4 files changed, 306 insertions(+) create mode 100644 Documentation/devicetree/bindings/power/amlogic,meson-gx-pwrc.txt create mode 100644 drivers/soc/amlogic/meson-gx-pwrc-vpu.c -- 2.7.4 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver 2017-10-28 16:40 [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains Neil Armstrong @ 2017-10-28 16:40 ` Neil Armstrong 2017-10-29 14:59 ` Kevin Hilman 2017-10-28 16:40 ` [RESEND PATCH v2 2/2] dt-bindings: power: add amlogic meson power domain bindings Neil Armstrong 1 sibling, 1 reply; 8+ messages in thread From: Neil Armstrong @ 2017-10-28 16:40 UTC (permalink / raw) To: linux-arm-kernel The Video Processing Unit needs a specific Power Domain powering scheme this driver handles this as a PM Power Domain driver. Signed-off-by: Neil Armstrong <narmstrong@baylibre.com> --- drivers/soc/amlogic/Kconfig | 10 ++ drivers/soc/amlogic/Makefile | 1 + drivers/soc/amlogic/meson-gx-pwrc-vpu.c | 234 ++++++++++++++++++++++++++++++++ 3 files changed, 245 insertions(+) create mode 100644 drivers/soc/amlogic/meson-gx-pwrc-vpu.c diff --git a/drivers/soc/amlogic/Kconfig b/drivers/soc/amlogic/Kconfig index ef0b8f6..1621bb2 100644 --- a/drivers/soc/amlogic/Kconfig +++ b/drivers/soc/amlogic/Kconfig @@ -9,6 +9,16 @@ config MESON_GX_SOCINFO Say yes to support decoding of Amlogic Meson GX SoC family information about the type, package and version. +config MESON_GX_PM_DOMAINS + bool "Amlogic Meson GX Power Domains driver" + depends on ARCH_MESON || COMPILE_TEST + default ARCH_MESON + select PM_GENERIC_DOMAINS + select PM_GENERIC_DOMAINS_OF + help + Say yes to expose Amlogic Meson GX Power Domains as + Generic Power Domains. + config MESON_MX_SOCINFO bool "Amlogic Meson MX SoC Information driver" depends on ARCH_MESON || COMPILE_TEST diff --git a/drivers/soc/amlogic/Makefile b/drivers/soc/amlogic/Makefile index 1f5df50..8fa3218 100644 --- a/drivers/soc/amlogic/Makefile +++ b/drivers/soc/amlogic/Makefile @@ -1,2 +1,3 @@ obj-$(CONFIG_MESON_GX_SOCINFO) += meson-gx-socinfo.o +obj-$(CONFIG_MESON_GX_PM_DOMAINS) += meson-gx-pwrc-vpu.o obj-$(CONFIG_MESON_MX_SOCINFO) += meson-mx-socinfo.o diff --git a/drivers/soc/amlogic/meson-gx-pwrc-vpu.c b/drivers/soc/amlogic/meson-gx-pwrc-vpu.c new file mode 100644 index 0000000..bf5190b --- /dev/null +++ b/drivers/soc/amlogic/meson-gx-pwrc-vpu.c @@ -0,0 +1,234 @@ +/* + * Copyright (c) 2017 BayLibre, SAS + * Author: Neil Armstrong <narmstrong@baylibre.com> + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +#include <linux/of_address.h> +#include <linux/platform_device.h> +#include <linux/pm_domain.h> +#include <linux/bitfield.h> +#include <linux/regmap.h> +#include <linux/mfd/syscon.h> +#include <linux/reset.h> +#include <linux/clk.h> + +/* AO Offsets */ + +#define AO_RTI_GEN_PWR_SLEEP0 (0x3a << 2) + +#define GEN_PWR_VPU_HDMI BIT(8) +#define GEN_PWR_VPU_HDMI_ISO BIT(9) + +/* HHI Offsets */ + +#define HHI_MEM_PD_REG0 (0x40 << 2) +#define HHI_VPU_MEM_PD_REG0 (0x41 << 2) +#define HHI_VPU_MEM_PD_REG1 (0x42 << 2) + +struct meson_gx_pwrc_vpu { + struct generic_pm_domain genpd; + struct regmap *regmap_ao; + struct regmap *regmap_hhi; + struct reset_control *rstc; + struct clk *vpu_clk; + struct clk *vapb_clk; + bool powered; +}; + +static inline +struct meson_gx_pwrc_vpu *genpd_to_pd(struct generic_pm_domain *d) +{ + return container_of(d, struct meson_gx_pwrc_vpu, genpd); +} + +static int meson_gx_pwrc_vpu_power_off(struct generic_pm_domain *genpd) +{ + struct meson_gx_pwrc_vpu *pd = genpd_to_pd(genpd); + int i; + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI_ISO, GEN_PWR_VPU_HDMI_ISO); + udelay(20); + + /* Power Down Memories */ + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG0, + 0x2 << i, 0x3 << i); + udelay(5); + } + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG1, + 0x2 << i, 0x3 << i); + udelay(5); + } + for (i = 8; i < 16; i++) { + regmap_update_bits(pd->regmap_hhi, HHI_MEM_PD_REG0, + BIT(i), BIT(i)); + udelay(5); + } + udelay(20); + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI, GEN_PWR_VPU_HDMI); + + msleep(20); + + clk_disable_unprepare(pd->vpu_clk); + clk_disable_unprepare(pd->vapb_clk); + + pd->powered = false; + + return 0; +} + +static int meson_gx_pwrc_vpu_setup_clk(struct meson_gx_pwrc_vpu *pd) +{ + int ret; + + ret = clk_prepare_enable(pd->vpu_clk); + if (ret) + return ret; + + return clk_prepare_enable(pd->vapb_clk); +} + +static int meson_gx_pwrc_vpu_power_on(struct generic_pm_domain *genpd) +{ + struct meson_gx_pwrc_vpu *pd = genpd_to_pd(genpd); + int ret; + int i; + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI, 0); + udelay(20); + + /* Power Up Memories */ + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG0, + 0x2 << i, 0); + udelay(5); + } + + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG1, + 0x2 << i, 0); + udelay(5); + } + + for (i = 8; i < 16; i++) { + regmap_update_bits(pd->regmap_hhi, HHI_MEM_PD_REG0, + BIT(i), 0); + udelay(5); + } + udelay(20); + + ret = reset_control_assert(pd->rstc); + if (ret) + return ret; + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI_ISO, 0); + + ret = reset_control_deassert(pd->rstc); + if (ret) + return ret; + + ret = meson_gx_pwrc_vpu_setup_clk(pd); + if (ret) + return ret; + + pd->powered = true; + + return 0; +} + +static bool meson_gx_pwrc_vpu_get_power(struct meson_gx_pwrc_vpu *pd) +{ + u32 reg; + + regmap_read(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, ®); + + return (reg & GEN_PWR_VPU_HDMI); +} + +static struct meson_gx_pwrc_vpu vpu_hdmi_pd = { + .genpd = { + .name = "vpu_hdmi", + .power_off = meson_gx_pwrc_vpu_power_off, + .power_on = meson_gx_pwrc_vpu_power_on, + }, +}; + +static int meson_gx_pwrc_vpu_probe(struct platform_device *pdev) +{ + struct regmap *regmap_ao, *regmap_hhi; + struct reset_control *rstc; + struct clk *vpu_clk; + struct clk *vapb_clk; + + regmap_ao = syscon_node_to_regmap(of_get_parent(pdev->dev.of_node)); + if (IS_ERR(regmap_ao)) { + dev_err(&pdev->dev, "failed to get regmap\n"); + return PTR_ERR(regmap_ao); + } + + regmap_hhi = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "amlogic,hhi-sysctrl"); + if (IS_ERR(regmap_hhi)) { + dev_err(&pdev->dev, "failed to get HHI regmap\n"); + return PTR_ERR(regmap_hhi); + } + + rstc = devm_reset_control_array_get(&pdev->dev, false, false); + if (IS_ERR(rstc)) { + dev_err(&pdev->dev, "failed to get reset lines\n"); + return PTR_ERR(rstc); + } + + vpu_clk = devm_clk_get(&pdev->dev, "vpu"); + if (IS_ERR(vpu_clk)) { + dev_err(&pdev->dev, "vpu clock request failed\n"); + return PTR_ERR(vpu_clk); + } + + vapb_clk = devm_clk_get(&pdev->dev, "vapb"); + if (IS_ERR(vapb_clk)) { + dev_err(&pdev->dev, "vapb clock request failed\n"); + return PTR_ERR(vapb_clk); + } + + vpu_hdmi_pd.regmap_ao = regmap_ao; + vpu_hdmi_pd.regmap_hhi = regmap_hhi; + vpu_hdmi_pd.rstc = rstc; + vpu_hdmi_pd.vpu_clk = vpu_clk; + vpu_hdmi_pd.vapb_clk = vapb_clk; + + pm_genpd_init(&vpu_hdmi_pd.genpd, &simple_qos_governor, + meson_gx_pwrc_vpu_get_power(&vpu_hdmi_pd)); + + return of_genpd_add_provider_simple(pdev->dev.of_node, + &vpu_hdmi_pd.genpd); +} + +static void meson_gx_pwrc_vpu_shutdown(struct platform_device *pdev) +{ + if (vpu_hdmi_pd.powered) + meson_gx_pwrc_vpu_power_off(&vpu_hdmi_pd.genpd); +} + +static const struct of_device_id meson_gx_pwrc_vpu_match_table[] = { + { .compatible = "amlogic,meson-gx-pwrc-vpu" }, + { /* sentinel */ } +}; + +static struct platform_driver meson_gx_pwrc_vpu_driver = { + .probe = meson_gx_pwrc_vpu_probe, + .shutdown = meson_gx_pwrc_vpu_shutdown, + .driver = { + .name = "meson_gx_pwrc_vpu", + .of_match_table = meson_gx_pwrc_vpu_match_table, + }, +}; +builtin_platform_driver(meson_gx_pwrc_vpu_driver); -- 2.7.4 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver 2017-10-28 16:40 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong @ 2017-10-29 14:59 ` Kevin Hilman 0 siblings, 0 replies; 8+ messages in thread From: Kevin Hilman @ 2017-10-29 14:59 UTC (permalink / raw) To: linux-arm-kernel Hi Neil, Neil Armstrong <narmstrong@baylibre.com> writes: > The Video Processing Unit needs a specific Power Domain powering scheme > this driver handles this as a PM Power Domain driver. > > Signed-off-by: Neil Armstrong <narmstrong@baylibre.com> [...] > +static inline > +struct meson_gx_pwrc_vpu *genpd_to_pd(struct generic_pm_domain *d) > +{ > + return container_of(d, struct meson_gx_pwrc_vpu, genpd); > +} > + > +static int meson_gx_pwrc_vpu_power_off(struct generic_pm_domain *genpd) > +{ > + struct meson_gx_pwrc_vpu *pd = genpd_to_pd(genpd); > + int i; > + > + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, > + GEN_PWR_VPU_HDMI_ISO, GEN_PWR_VPU_HDMI_ISO); > + udelay(20); Some minor nits/questions. I know the vendor code does it this way, but... > + /* Power Down Memories */ > + for (i = 0; i < 32; i += 2) { > + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG0, > + 0x2 << i, 0x3 << i); > + udelay(5); > + } several of these bits are marked "reserved" in the datasheet. Have you tried without writing to the reserved bits? > + for (i = 0; i < 32; i += 2) { > + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG1, > + 0x2 << i, 0x3 << i); > + udelay(5); > + } ditto > + for (i = 8; i < 16; i++) { > + regmap_update_bits(pd->regmap_hhi, HHI_MEM_PD_REG0, > + BIT(i), BIT(i)); > + udelay(5); > + } Do these really have to be written one bit at a time? I'm actually OK to merge things like this, since it also captures the sequences used by the vendor kernel, but I'm just curious if you did any of the other experiments. If you can try those experiments, we can take them as follow-on patches. Kevin ^ permalink raw reply [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 2/2] dt-bindings: power: add amlogic meson power domain bindings 2017-10-28 16:40 [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains Neil Armstrong 2017-10-28 16:40 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong @ 2017-10-28 16:40 ` Neil Armstrong 1 sibling, 0 replies; 8+ messages in thread From: Neil Armstrong @ 2017-10-28 16:40 UTC (permalink / raw) To: linux-arm-kernel Acked-by: Rob Herring <robh@kernel.org> Signed-off-by: Neil Armstrong <narmstrong@baylibre.com> --- .../bindings/power/amlogic,meson-gx-pwrc.txt | 61 ++++++++++++++++++++++ 1 file changed, 61 insertions(+) create mode 100644 Documentation/devicetree/bindings/power/amlogic,meson-gx-pwrc.txt diff --git a/Documentation/devicetree/bindings/power/amlogic,meson-gx-pwrc.txt b/Documentation/devicetree/bindings/power/amlogic,meson-gx-pwrc.txt new file mode 100644 index 0000000..95ec49a --- /dev/null +++ b/Documentation/devicetree/bindings/power/amlogic,meson-gx-pwrc.txt @@ -0,0 +1,61 @@ +Amlogic Meson Power Controller +============================== + +The Amlogic Meson SoCs embeds an internal Power domain controller. + +VPU Power Domain +---------------- + +The Video Processing Unit power domain is controlled by this power controller, +but the domain requires some external resources to meet the correct power +sequences. +The bindings must respect the power domain bindings as described in the file +power_domain.txt + +Device Tree Bindings: +--------------------- + +Required properties: +- compatible: should be "amlogic,meson-gx-pwrc-vpu" for the Meson GX SoCs +- #power-domain-cells: should be 0 +- amlogic,hhi-sysctrl: phandle to the HHI sysctrl node +- resets: phandles to the reset lines needed for this power demain sequence + as described in ../reset/reset.txt +- clocks: from common clock binding: handle to VPU and VAPB clocks +- clock-names: from common clock binding: must contain "vpu", "vapb" + corresponding to entry in the clocks property. + +Parent node should have the following properties : +- compatible: "amlogic,meson-gx-ao-sysctrl", "syscon", "simple-mfd" +- reg: base address and size of the AO system control register space. + +Example: +------- + +ao_sysctrl: sys-ctrl at 0 { + compatible = "amlogic,meson-gx-ao-sysctrl", "syscon", "simple-mfd"; + reg = <0x0 0x0 0x0 0x100>; + + pwrc_vpu: power-controller-vpu { + compatible = "amlogic,meson-gx-pwrc-vpu"; + #power-domain-cells = <0>; + amlogic,hhi-sysctrl = <&sysctrl>; + resets = <&reset RESET_VIU>, + <&reset RESET_VENC>, + <&reset RESET_VCBUS>, + <&reset RESET_BT656>, + <&reset RESET_DVIN_RESET>, + <&reset RESET_RDMA>, + <&reset RESET_VENCI>, + <&reset RESET_VENCP>, + <&reset RESET_VDAC>, + <&reset RESET_VDI6>, + <&reset RESET_VENCL>, + <&reset RESET_VID_LOCK>; + clocks = <&clkc CLKID_VPU>, + <&clkc CLKID_VAPB>; + clock-names = "vpu", "vapb"; + }; +}; + + -- 2.7.4 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains @ 2017-10-28 16:36 Neil Armstrong 2017-10-28 16:36 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong 0 siblings, 1 reply; 8+ messages in thread From: Neil Armstrong @ 2017-10-28 16:36 UTC (permalink / raw) To: linux-arm-kernel [Resent with linux-pm in CC] On the Amlogic Gx SoCs (GXBB, GXL & GXM), the VPU power domain is initialized by the vendor U-Boot code, but running mainline U-boot has been possible on these SoCs. But lacking such init made the system lock at kernel boot. This patchset adds the Video Processing Unit power domain driver to enable the same power-on and power-down sequences and was designed to allow booting with the power domain already initialized or not. This driver has been tested on : - Odroid-C2 (GXBB) with Vendor and Mainline U-Boot - P212 (GXL) with Vendor and Mainline U-Boot - Khadas Vim (GXL) with Vendor U-Boot - Khadas Vim2 with (GXM) Vendor U-Boot Changes since v1: - fixed bindings compatible string Neil Armstrong (2): soc: amlogic: add Meson GX VPU Domains driver dt-bindings: power: add amlogic meson power domain bindings .../bindings/power/amlogic,meson-gx-pwrc.txt | 61 ++++++ drivers/soc/amlogic/Kconfig | 10 + drivers/soc/amlogic/Makefile | 1 + drivers/soc/amlogic/meson-gx-pwrc-vpu.c | 234 +++++++++++++++++++++ 4 files changed, 306 insertions(+) create mode 100644 Documentation/devicetree/bindings/power/amlogic,meson-gx-pwrc.txt create mode 100644 drivers/soc/amlogic/meson-gx-pwrc-vpu.c -- 2.7.4 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver 2017-10-28 16:36 [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains Neil Armstrong @ 2017-10-28 16:36 ` Neil Armstrong 2017-10-30 15:40 ` Ulf Hansson 0 siblings, 1 reply; 8+ messages in thread From: Neil Armstrong @ 2017-10-28 16:36 UTC (permalink / raw) To: linux-arm-kernel The Video Processing Unit needs a specific Power Domain powering scheme this driver handles this as a PM Power Domain driver. Signed-off-by: Neil Armstrong <narmstrong@baylibre.com> --- drivers/soc/amlogic/Kconfig | 10 ++ drivers/soc/amlogic/Makefile | 1 + drivers/soc/amlogic/meson-gx-pwrc-vpu.c | 234 ++++++++++++++++++++++++++++++++ 3 files changed, 245 insertions(+) create mode 100644 drivers/soc/amlogic/meson-gx-pwrc-vpu.c diff --git a/drivers/soc/amlogic/Kconfig b/drivers/soc/amlogic/Kconfig index ef0b8f6..1621bb2 100644 --- a/drivers/soc/amlogic/Kconfig +++ b/drivers/soc/amlogic/Kconfig @@ -9,6 +9,16 @@ config MESON_GX_SOCINFO Say yes to support decoding of Amlogic Meson GX SoC family information about the type, package and version. +config MESON_GX_PM_DOMAINS + bool "Amlogic Meson GX Power Domains driver" + depends on ARCH_MESON || COMPILE_TEST + default ARCH_MESON + select PM_GENERIC_DOMAINS + select PM_GENERIC_DOMAINS_OF + help + Say yes to expose Amlogic Meson GX Power Domains as + Generic Power Domains. + config MESON_MX_SOCINFO bool "Amlogic Meson MX SoC Information driver" depends on ARCH_MESON || COMPILE_TEST diff --git a/drivers/soc/amlogic/Makefile b/drivers/soc/amlogic/Makefile index 1f5df50..8fa3218 100644 --- a/drivers/soc/amlogic/Makefile +++ b/drivers/soc/amlogic/Makefile @@ -1,2 +1,3 @@ obj-$(CONFIG_MESON_GX_SOCINFO) += meson-gx-socinfo.o +obj-$(CONFIG_MESON_GX_PM_DOMAINS) += meson-gx-pwrc-vpu.o obj-$(CONFIG_MESON_MX_SOCINFO) += meson-mx-socinfo.o diff --git a/drivers/soc/amlogic/meson-gx-pwrc-vpu.c b/drivers/soc/amlogic/meson-gx-pwrc-vpu.c new file mode 100644 index 0000000..bf5190b --- /dev/null +++ b/drivers/soc/amlogic/meson-gx-pwrc-vpu.c @@ -0,0 +1,234 @@ +/* + * Copyright (c) 2017 BayLibre, SAS + * Author: Neil Armstrong <narmstrong@baylibre.com> + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +#include <linux/of_address.h> +#include <linux/platform_device.h> +#include <linux/pm_domain.h> +#include <linux/bitfield.h> +#include <linux/regmap.h> +#include <linux/mfd/syscon.h> +#include <linux/reset.h> +#include <linux/clk.h> + +/* AO Offsets */ + +#define AO_RTI_GEN_PWR_SLEEP0 (0x3a << 2) + +#define GEN_PWR_VPU_HDMI BIT(8) +#define GEN_PWR_VPU_HDMI_ISO BIT(9) + +/* HHI Offsets */ + +#define HHI_MEM_PD_REG0 (0x40 << 2) +#define HHI_VPU_MEM_PD_REG0 (0x41 << 2) +#define HHI_VPU_MEM_PD_REG1 (0x42 << 2) + +struct meson_gx_pwrc_vpu { + struct generic_pm_domain genpd; + struct regmap *regmap_ao; + struct regmap *regmap_hhi; + struct reset_control *rstc; + struct clk *vpu_clk; + struct clk *vapb_clk; + bool powered; +}; + +static inline +struct meson_gx_pwrc_vpu *genpd_to_pd(struct generic_pm_domain *d) +{ + return container_of(d, struct meson_gx_pwrc_vpu, genpd); +} + +static int meson_gx_pwrc_vpu_power_off(struct generic_pm_domain *genpd) +{ + struct meson_gx_pwrc_vpu *pd = genpd_to_pd(genpd); + int i; + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI_ISO, GEN_PWR_VPU_HDMI_ISO); + udelay(20); + + /* Power Down Memories */ + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG0, + 0x2 << i, 0x3 << i); + udelay(5); + } + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG1, + 0x2 << i, 0x3 << i); + udelay(5); + } + for (i = 8; i < 16; i++) { + regmap_update_bits(pd->regmap_hhi, HHI_MEM_PD_REG0, + BIT(i), BIT(i)); + udelay(5); + } + udelay(20); + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI, GEN_PWR_VPU_HDMI); + + msleep(20); + + clk_disable_unprepare(pd->vpu_clk); + clk_disable_unprepare(pd->vapb_clk); + + pd->powered = false; + + return 0; +} + +static int meson_gx_pwrc_vpu_setup_clk(struct meson_gx_pwrc_vpu *pd) +{ + int ret; + + ret = clk_prepare_enable(pd->vpu_clk); + if (ret) + return ret; + + return clk_prepare_enable(pd->vapb_clk); +} + +static int meson_gx_pwrc_vpu_power_on(struct generic_pm_domain *genpd) +{ + struct meson_gx_pwrc_vpu *pd = genpd_to_pd(genpd); + int ret; + int i; + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI, 0); + udelay(20); + + /* Power Up Memories */ + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG0, + 0x2 << i, 0); + udelay(5); + } + + for (i = 0; i < 32; i += 2) { + regmap_update_bits(pd->regmap_hhi, HHI_VPU_MEM_PD_REG1, + 0x2 << i, 0); + udelay(5); + } + + for (i = 8; i < 16; i++) { + regmap_update_bits(pd->regmap_hhi, HHI_MEM_PD_REG0, + BIT(i), 0); + udelay(5); + } + udelay(20); + + ret = reset_control_assert(pd->rstc); + if (ret) + return ret; + + regmap_update_bits(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, + GEN_PWR_VPU_HDMI_ISO, 0); + + ret = reset_control_deassert(pd->rstc); + if (ret) + return ret; + + ret = meson_gx_pwrc_vpu_setup_clk(pd); + if (ret) + return ret; + + pd->powered = true; + + return 0; +} + +static bool meson_gx_pwrc_vpu_get_power(struct meson_gx_pwrc_vpu *pd) +{ + u32 reg; + + regmap_read(pd->regmap_ao, AO_RTI_GEN_PWR_SLEEP0, ®); + + return (reg & GEN_PWR_VPU_HDMI); +} + +static struct meson_gx_pwrc_vpu vpu_hdmi_pd = { + .genpd = { + .name = "vpu_hdmi", + .power_off = meson_gx_pwrc_vpu_power_off, + .power_on = meson_gx_pwrc_vpu_power_on, + }, +}; + +static int meson_gx_pwrc_vpu_probe(struct platform_device *pdev) +{ + struct regmap *regmap_ao, *regmap_hhi; + struct reset_control *rstc; + struct clk *vpu_clk; + struct clk *vapb_clk; + + regmap_ao = syscon_node_to_regmap(of_get_parent(pdev->dev.of_node)); + if (IS_ERR(regmap_ao)) { + dev_err(&pdev->dev, "failed to get regmap\n"); + return PTR_ERR(regmap_ao); + } + + regmap_hhi = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "amlogic,hhi-sysctrl"); + if (IS_ERR(regmap_hhi)) { + dev_err(&pdev->dev, "failed to get HHI regmap\n"); + return PTR_ERR(regmap_hhi); + } + + rstc = devm_reset_control_array_get(&pdev->dev, false, false); + if (IS_ERR(rstc)) { + dev_err(&pdev->dev, "failed to get reset lines\n"); + return PTR_ERR(rstc); + } + + vpu_clk = devm_clk_get(&pdev->dev, "vpu"); + if (IS_ERR(vpu_clk)) { + dev_err(&pdev->dev, "vpu clock request failed\n"); + return PTR_ERR(vpu_clk); + } + + vapb_clk = devm_clk_get(&pdev->dev, "vapb"); + if (IS_ERR(vapb_clk)) { + dev_err(&pdev->dev, "vapb clock request failed\n"); + return PTR_ERR(vapb_clk); + } + + vpu_hdmi_pd.regmap_ao = regmap_ao; + vpu_hdmi_pd.regmap_hhi = regmap_hhi; + vpu_hdmi_pd.rstc = rstc; + vpu_hdmi_pd.vpu_clk = vpu_clk; + vpu_hdmi_pd.vapb_clk = vapb_clk; + + pm_genpd_init(&vpu_hdmi_pd.genpd, &simple_qos_governor, + meson_gx_pwrc_vpu_get_power(&vpu_hdmi_pd)); + + return of_genpd_add_provider_simple(pdev->dev.of_node, + &vpu_hdmi_pd.genpd); +} + +static void meson_gx_pwrc_vpu_shutdown(struct platform_device *pdev) +{ + if (vpu_hdmi_pd.powered) + meson_gx_pwrc_vpu_power_off(&vpu_hdmi_pd.genpd); +} + +static const struct of_device_id meson_gx_pwrc_vpu_match_table[] = { + { .compatible = "amlogic,meson-gx-pwrc-vpu" }, + { /* sentinel */ } +}; + +static struct platform_driver meson_gx_pwrc_vpu_driver = { + .probe = meson_gx_pwrc_vpu_probe, + .shutdown = meson_gx_pwrc_vpu_shutdown, + .driver = { + .name = "meson_gx_pwrc_vpu", + .of_match_table = meson_gx_pwrc_vpu_match_table, + }, +}; +builtin_platform_driver(meson_gx_pwrc_vpu_driver); -- 2.7.4 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver 2017-10-28 16:36 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong @ 2017-10-30 15:40 ` Ulf Hansson 2017-10-30 16:47 ` Neil Armstrong 2017-10-30 16:56 ` Kevin Hilman 0 siblings, 2 replies; 8+ messages in thread From: Ulf Hansson @ 2017-10-30 15:40 UTC (permalink / raw) To: linux-arm-kernel Hi Neil, [...] > + > + pm_genpd_init(&vpu_hdmi_pd.genpd, &simple_qos_governor, > + meson_gx_pwrc_vpu_get_power(&vpu_hdmi_pd)); > + This means that you may initialize the genpd in powered on state. For that scenario, I think you will hit a clock unprepare/disable in-balance issue when genpd calls your ->power_off() callback, simply because you didn't call clk_prepare_enable() for the clocks first. [...] Kind regards Uffe ^ permalink raw reply [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver 2017-10-30 15:40 ` Ulf Hansson @ 2017-10-30 16:47 ` Neil Armstrong 2017-10-30 16:56 ` Kevin Hilman 1 sibling, 0 replies; 8+ messages in thread From: Neil Armstrong @ 2017-10-30 16:47 UTC (permalink / raw) To: linux-arm-kernel On 30/10/2017 16:40, Ulf Hansson wrote: > Hi Neil, > > [...] > >> + >> + pm_genpd_init(&vpu_hdmi_pd.genpd, &simple_qos_governor, >> + meson_gx_pwrc_vpu_get_power(&vpu_hdmi_pd)); >> + > > This means that you may initialize the genpd in powered on state. > > For that scenario, I think you will hit a clock unprepare/disable > in-balance issue when genpd calls your ->power_off() callback, simply > because you didn't call clk_prepare_enable() for the clocks first. > > [...] > > Kind regards > Uffe > Hi, I added a "powered" attribute to trace if power-on was done by the driver, but I only protected the shutdown call.... I should indeed protect it in the power_off call instead. Neil ^ permalink raw reply [flat|nested] 8+ messages in thread
* [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver 2017-10-30 15:40 ` Ulf Hansson 2017-10-30 16:47 ` Neil Armstrong @ 2017-10-30 16:56 ` Kevin Hilman 1 sibling, 0 replies; 8+ messages in thread From: Kevin Hilman @ 2017-10-30 16:56 UTC (permalink / raw) To: linux-arm-kernel Ulf Hansson <ulf.hansson@linaro.org> writes: > Hi Neil, > > [...] > >> + >> + pm_genpd_init(&vpu_hdmi_pd.genpd, &simple_qos_governor, >> + meson_gx_pwrc_vpu_get_power(&vpu_hdmi_pd)); >> + > > This means that you may initialize the genpd in powered on state. > > For that scenario, I think you will hit a clock unprepare/disable > in-balance issue when genpd calls your ->power_off() callback, simply > because you didn't call clk_prepare_enable() for the clocks first. Yup, in my testing, I definitely hit this. Kevin ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2017-10-30 16:56 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2017-10-28 16:40 [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains Neil Armstrong 2017-10-28 16:40 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong 2017-10-29 14:59 ` Kevin Hilman 2017-10-28 16:40 ` [RESEND PATCH v2 2/2] dt-bindings: power: add amlogic meson power domain bindings Neil Armstrong -- strict thread matches above, loose matches on Subject: below -- 2017-10-28 16:36 [RESEND PATCH v2 0/2] soc: amlogic: add support for Meson GX VPU Domains Neil Armstrong 2017-10-28 16:36 ` [RESEND PATCH v2 1/2] soc: amlogic: add Meson GX VPU Domains driver Neil Armstrong 2017-10-30 15:40 ` Ulf Hansson 2017-10-30 16:47 ` Neil Armstrong 2017-10-30 16:56 ` Kevin Hilman
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).