PM / devfreq: Add dynamic scaling for imx8m ddr controller
Subsystems (6)
v4 cover PM / devfreq: Add dynamic scaling for imx8m ddr controller
This adds support for dynamic scaling of the DDR Controller (ddrc) present in imx8m series. Actual frequency switching is implemented inside TF-A, this driver wraps the SMC calls and synchronizes the clk tree.
DRAM frequency switching requires clock manipulation but during this operation DRAM itself is briefly inaccessible so this operation is performed a SMC call to by TF-A which runs from a SRAM area. Upon returning to linux the clock tree is updated to correspond to hardware configuration.
This is handled via CLK_GET_RATE_NO_CACHE for dividers but muxes are handled
manually: the driver will prepare/enable the new parents ahead of switching (so that the expected roots are enabled) and afterwards it will call clk_set_parent to ensure the parents in clock framework are up-to-date.
This series is atomically useful and roughly similar to devfreq drivers for tegra and rockchip.
Running at lower dram rates saves power but can affect the functionality of other blocks in the chip (display, vpu etc). Support for in-kernel constraints will some separately.
Angus/Martin: You previously attempted to test on purism boards, this updated version should work without hacks and has no dependencies.
Changes since v3:
* Rename to imx8m-ddrc. Similar blocks are present on imx7d and imx8qxp/imx8qm but soc integration is different.
* Move dt bindings to /memory-controllers/fsl/
* Fix dt validation issues
* Fix imx8mm.dtsi ddrc referencing ddrc_opp_table which is only defined in evk
* Move opps to child of ddrc device node
* Only add imx_ddrc_get_dev_status in perf patch.
* Adjust print messages
Link to v3: https://patchwork.kernel.org/cover/11221935/
Leonard Crestez (6):
clk: imx8m: Set CLK_GET_RATE_NOCACHE on dram clocks
clk: imx: Mark dram pll on 8mm and 8mn with CLK_GET_RATE_NOCACHE
dt-bindings: memory: Add bindings for imx8m ddr controller PM / devfreq: Add dynamic scaling for imx8m ddr controller PM / devfreq: imx8m-ddrc: Measure bandwidth with perf
arm64: dts: imx8m: Add ddr controller nodes
.../memory-controllers/fsl/imx8m-ddrc.yaml | 61 ++ arch/arm64/boot/dts/freescale/imx8mm-evk.dts | 18 + arch/arm64/boot/dts/freescale/imx8mm.dtsi | 13 +- .../boot/dts/freescale/imx8mn-ddr4-evk.dts | 18 + arch/arm64/boot/dts/freescale/imx8mn.dtsi | 13 +- arch/arm64/boot/dts/freescale/imx8mq-evk.dts | 24 + arch/arm64/boot/dts/freescale/imx8mq.dtsi | 13 +- drivers/clk/imx/clk-imx8mm.c | 13 +- drivers/clk/imx/clk-imx8mn.c | 14 +- drivers/clk/imx/clk-imx8mq.c | 15 +- drivers/clk/imx/clk-pll14xx.c | 7 + drivers/clk/imx/clk.h | 1 + drivers/devfreq/Kconfig | 10 + drivers/devfreq/Makefile | 1 + drivers/devfreq/imx8m-ddrc.c | 569 ++++++++++++++++++ 15 files changed, 777 insertions(+), 13 deletions(-) create mode 100644 Documentation/devicetree/bindings/memory-controllers/fsl/imx8m-ddrc.yaml create mode 100644 drivers/devfreq/imx8m-ddrc.c
Signature · 1 line
2.17.1
PATCH v4 2/6 clk: imx: Mark dram pll on 8mm and 8mn with CLK_GET_RATE_NOCACHE
DRAM frequency switches are executed in firmware and can change the configuration of the DRAM PLL outside linux. Mark these CLKs with CLK_GET_RATE_NOCACHE so we always read back the PLL config registers and recalculate rates.
In current DRAM frequency tables on 8mm/8mn only the maximum frequency uses the PLL so it's always configured in the same way. However reading back the PLL configuration is the correct behavior and allows additional setpoints in the future.
--- drivers/clk/imx/clk-imx8mm.c | 2 +- drivers/clk/imx/clk-imx8mn.c | 2 +- drivers/clk/imx/clk-pll14xx.c | 7 +++++++ drivers/clk/imx/clk.h | 1 + 4 files changed, 10 insertions(+), 2 deletions(-)
drivers/clk/imx/clk-imx8mm.c +1 -1
@@ -326,11 +326,11 @@ static int imx8mm_clocks_probe(struct platform_device *pdev) clks[IMX8MM_SYS_PLL3_REF_SEL ] =imx_clk_mux ("sys_pll3_ref_sel" , base +0x114 ,0 ,2 , pll_ref_sels,ARRAY_SIZE (pll_ref_sels)); clks[IMX8MM_AUDIO_PLL1 ] =imx_clk_pll14xx ("audio_pll1" ,"audio_pll1_ref_sel" , base, &imx_1443x_pll); clks[IMX8MM_AUDIO_PLL2 ] =imx_clk_pll14xx ("audio_pll2" ,"audio_pll2_ref_sel" , base +0x14 , &imx_1443x_pll); clks[IMX8MM_VIDEO_PLL1 ] =imx_clk_pll14xx ("video_pll1" ,"video_pll1_ref_sel" , base +0x28 , &imx_1443x_pll); - clks[IMX8MM_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_pll); + clks[IMX8MM_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_dram_pll); clks[IMX8MM_GPU_PLL ] =imx_clk_pll14xx ("gpu_pll" ,"gpu_pll_ref_sel" , base +0x64 , &imx_1416x_pll); clks[IMX8MM_VPU_PLL ] =imx_clk_pll14xx ("vpu_pll" ,"vpu_pll_ref_sel" , base +0x74 , &imx_1416x_pll); clks[IMX8MM_ARM_PLL ] =imx_clk_pll14xx ("arm_pll" ,"arm_pll_ref_sel" , base +0x84 , &imx_1416x_pll); clks[IMX8MM_SYS_PLL1 ] =imx_clk_fixed ("sys_pll1" ,800000000 ); clks[IMX8MM_SYS_PLL2 ] =imx_clk_fixed ("sys_pll2" ,1000000000 );
drivers/clk/imx/clk-imx8mn.c +1 -1
@@ -323,11 +323,11 @@ static int imx8mn_clocks_probe(struct platform_device *pdev) clks[IMX8MN_SYS_PLL3_REF_SEL ] =imx_clk_mux ("sys_pll3_ref_sel" , base +0x114 ,0 ,2 , pll_ref_sels,ARRAY_SIZE (pll_ref_sels)); clks[IMX8MN_AUDIO_PLL1 ] =imx_clk_pll14xx ("audio_pll1" ,"audio_pll1_ref_sel" , base, &imx_1443x_pll); clks[IMX8MN_AUDIO_PLL2 ] =imx_clk_pll14xx ("audio_pll2" ,"audio_pll2_ref_sel" , base +0x14 , &imx_1443x_pll); clks[IMX8MN_VIDEO_PLL1 ] =imx_clk_pll14xx ("video_pll1" ,"video_pll1_ref_sel" , base +0x28 , &imx_1443x_pll); - clks[IMX8MN_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_pll); + clks[IMX8MN_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_dram_pll); clks[IMX8MN_GPU_PLL ] =imx_clk_pll14xx ("gpu_pll" ,"gpu_pll_ref_sel" , base +0x64 , &imx_1416x_pll); clks[IMX8MN_VPU_PLL ] =imx_clk_pll14xx ("vpu_pll" ,"vpu_pll_ref_sel" , base +0x74 , &imx_1416x_pll); clks[IMX8MN_ARM_PLL ] =imx_clk_pll14xx ("arm_pll" ,"arm_pll_ref_sel" , base +0x84 , &imx_1416x_pll); clks[IMX8MN_SYS_PLL1 ] =imx_clk_fixed ("sys_pll1" ,800000000 ); clks[IMX8MN_SYS_PLL2 ] =imx_clk_fixed ("sys_pll2" ,1000000000 );
drivers/clk/imx/clk-pll14xx.c +7 -0
@@ -65,10 +65,17 @@ struct imx_pll14xx_clk imx_1443x_pll = { .type = PLL_1443X, .rate_table = imx_pll1443x_tbl, .rate_count = ARRAY_SIZE(imx_pll1443x_tbl), }; +struct imx_pll14xx_clk imx_1443x_dram_pll = { + .type =PLL_1443X , + .rate_table = imx_pll1443x_tbl, + .rate_count =ARRAY_SIZE (imx_pll1443x_tbl), + .flags =CLK_GET_RATE_NOCACHE , +}; +struct imx_pll14xx_clk imx_1416x_pll = { .type =PLL_1416X , .rate_table = imx_pll1416x_tbl, .rate_count =ARRAY_SIZE (imx_pll1416x_tbl), };
drivers/clk/imx/clk.h +1 -0
@@ -50,10 +50,11 @@ struct imx_pll14xx_clk {int flags; };extern struct imx_pll14xx_clk imx_1416x_pll;extern struct imx_pll14xx_clk imx_1443x_pll; +extern struct imx_pll14xx_clk imx_1443x_dram_pll;#define imx_clk_cpu (name, parent_name, div, mux, pll, step) \ imx_clk_hw_cpu(name, parent_name, div, mux, pll, step)->clk #define clk_register_gate2(dev, name, parent_name, flags, reg, bit_idx, \
Signature · 1 line
2.17.1
PATCH v4 1/6 clk: imx8m: Set CLK_GET_RATE_NOCACHE on dram clocks
In reply to Leonard CrestezThese clocks are only modified as part of DRAM frequency switches during which DRAM itself is briefly inaccessible. The switch is performed with a SMC call to by TF-A which runs from a SRAM area; upon returning to linux several clocks bits are modified and we need to update them.
For rate bits an easy solution is to just mark with CLK_GET_RATE_NOCACHE so that new rates are always read back from registers.
--- drivers/clk/imx/clk-imx8mm.c | 11 +++++++++-- drivers/clk/imx/clk-imx8mn.c | 12 ++++++++++-- drivers/clk/imx/clk-imx8mq.c | 15 +++++++++++---- 3 files changed, 30 insertions(+), 8 deletions(-)
drivers/clk/imx/clk-imx8mm.c +9 -2
@@ -440,13 +440,20 @@ static int imx8mm_clocks_probe(struct platform_device *pdev)/* IPG */ clks[IMX8MM_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MM_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware +*/ + clks[IMX8MM_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mm_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MM_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mm_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); +/* IP */ - clks[IMX8MM_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mm_dram_alt_sels, base +0xa000 ); - clks[IMX8MM_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mm_dram_apb_sels, base +0xa080 ); clks[IMX8MM_CLK_VPU_G1 ] =imx8m_clk_composite ("vpu_g1" , imx8mm_vpu_g1_sels, base +0xa100 ); clks[IMX8MM_CLK_VPU_G2 ] =imx8m_clk_composite ("vpu_g2" , imx8mm_vpu_g2_sels, base +0xa180 ); clks[IMX8MM_CLK_DISP_DTRC ] =imx8m_clk_composite ("disp_dtrc" , imx8mm_disp_dtrc_sels, base +0xa200 ); clks[IMX8MM_CLK_DISP_DC8000 ] =imx8m_clk_composite ("disp_dc8000" , imx8mm_disp_dc8000_sels, base +0xa280 ); clks[IMX8MM_CLK_PCIE1_CTRL ] =imx8m_clk_composite ("pcie1_ctrl" , imx8mm_pcie1_ctrl_sels, base +0xa300 );
drivers/clk/imx/clk-imx8mn.c +10 -2
@@ -428,12 +428,20 @@ static int imx8mn_clocks_probe(struct platform_device *pdev) clks[IMX8MN_CLK_AHB ] =imx8m_clk_composite_critical ("ahb" , imx8mn_ahb_sels, base +0x9000 ); clks[IMX8MN_CLK_AUDIO_AHB ] =imx8m_clk_composite ("audio_ahb" , imx8mn_audio_ahb_sels, base +0x9100 ); clks[IMX8MN_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MN_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); clks[IMX8MN_CLK_DRAM_CORE ] =imx_clk_mux2_flags ("dram_core_clk" , base +0x9800 ,24 ,1 , imx8mn_dram_core_sels,ARRAY_SIZE (imx8mn_dram_core_sels),CLK_IS_CRITICAL ); - clks[IMX8MN_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 ); - clks[IMX8MN_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 ); + +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware +*/ + clks[IMX8MN_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MN_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); + clks[IMX8MN_CLK_DISP_PIXEL ] =imx8m_clk_composite ("disp_pixel" , imx8mn_disp_pixel_sels, base +0xa500 ); clks[IMX8MN_CLK_SAI2 ] =imx8m_clk_composite ("sai2" , imx8mn_sai2_sels, base +0xa600 ); clks[IMX8MN_CLK_SAI3 ] =imx8m_clk_composite ("sai3" , imx8mn_sai3_sels, base +0xa680 ); clks[IMX8MN_CLK_SAI5 ] =imx8m_clk_composite ("sai5" , imx8mn_sai5_sels, base +0xa780 ); clks[IMX8MN_CLK_SAI6 ] =imx8m_clk_composite ("sai6" , imx8mn_sai6_sels, base +0xa800 );
drivers/clk/imx/clk-imx8mq.c +11 -4
@@ -341,11 +341,12 @@ static int imx8mq_clocks_probe(struct platform_device *pdev) clks[IMX8MQ_VIDEO_PLL1_OUT ] =imx_clk_gate ("video_pll1_out" ,"video_pll1_bypass" , base +0x10 ,21 ); clks[IMX8MQ_SYS1_PLL_OUT ] =imx_clk_fixed ("sys1_pll_out" ,800000000 ); clks[IMX8MQ_SYS2_PLL_OUT ] =imx_clk_fixed ("sys2_pll_out" ,1000000000 ); clks[IMX8MQ_SYS3_PLL_OUT ] =imx_clk_sccg_pll ("sys3_pll_out" , sys3_pll_out_sels,ARRAY_SIZE (sys3_pll_out_sels),0 ,0 ,0 , base +0x48 ,CLK_IS_CRITICAL ); - clks[IMX8MQ_DRAM_PLL_OUT ] =imx_clk_sccg_pll ("dram_pll_out" , dram_pll_out_sels,ARRAY_SIZE (dram_pll_out_sels),0 ,0 ,0 , base +0x60 ,CLK_IS_CRITICAL ); + clks[IMX8MQ_DRAM_PLL_OUT ] =imx_clk_sccg_pll ("dram_pll_out" , dram_pll_out_sels,ARRAY_SIZE (dram_pll_out_sels),0 ,0 ,0 , base +0x60 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); clks[IMX8MQ_VIDEO2_PLL_OUT ] =imx_clk_sccg_pll ("video2_pll_out" , video2_pll_out_sels,ARRAY_SIZE (video2_pll_out_sels),0 ,0 ,0 , base +0x54 ,0 );/* SYS PLL1 fixed output */ clks[IMX8MQ_SYS1_PLL_40M_CG ] =imx_clk_gate ("sys1_pll_40m_cg" ,"sys1_pll_out" , base +0x30 ,9 ); clks[IMX8MQ_SYS1_PLL_80M_CG ] =imx_clk_gate ("sys1_pll_80m_cg" ,"sys1_pll_out" , base +0x30 ,11 );
@@ -433,15 +434,21 @@ static int imx8mq_clocks_probe(struct platform_device *pdev)/* IPG */ clks[IMX8MQ_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MQ_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); -/* IP */ +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware +*/ clks[IMX8MQ_CLK_DRAM_CORE ] =imx_clk_mux2_flags ("dram_core_clk" , base +0x9800 ,24 ,1 , imx8mq_dram_core_sels,ARRAY_SIZE (imx8mq_dram_core_sels),CLK_IS_CRITICAL ); + clks[IMX8MQ_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mq_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MQ_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mq_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); - clks[IMX8MQ_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mq_dram_alt_sels, base +0xa000 ); - clks[IMX8MQ_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mq_dram_apb_sels, base +0xa080 ); +/* IP */ clks[IMX8MQ_CLK_VPU_G1 ] =imx8m_clk_composite ("vpu_g1" , imx8mq_vpu_g1_sels, base +0xa100 ); clks[IMX8MQ_CLK_VPU_G2 ] =imx8m_clk_composite ("vpu_g2" , imx8mq_vpu_g2_sels, base +0xa180 ); clks[IMX8MQ_CLK_DISP_DTRC ] =imx8m_clk_composite ("disp_dtrc" , imx8mq_disp_dtrc_sels, base +0xa200 ); clks[IMX8MQ_CLK_DISP_DC8000 ] =imx8m_clk_composite ("disp_dc8000" , imx8mq_disp_dc8000_sels, base +0xa280 ); clks[IMX8MQ_CLK_PCIE1_CTRL ] =imx8m_clk_composite ("pcie1_ctrl" , imx8mq_pcie1_ctrl_sels, base +0xa300 );
Signature · 1 line
2.17.1
PATCH v4 3/6 dt-bindings: memory: Add bindings for imx8m ddr controller
In reply to Leonard CrestezAdd devicetree bindings for the i.MX DDR Controller on imx8m series chips. It supports dynamic frequency switching between multiple data rates and this is exposed to Linux via the devfreq subsystem.
--- .../memory-controllers/fsl/imx8m-ddrc.yaml | 61 +++++++++++++++++++ 1 file changed, 61 insertions(+) create mode 100644 Documentation/devicetree/bindings/memory-controllers/fsl/imx8m-ddrc.yaml
Documentation/devicetree/bindings/memory-controllers/fsl/imx8m-ddrc.yaml new file +61 -0
@@ -0,0 +1,61 @@ +# SPDX-License-Identifier: GPL-2.0 +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/memory-controllers/fsl/imx8m-ddrc.yaml# +$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: i.MX8M DDR Controller + +maintainers: + - Leonard Crestez <[email protected]> + +properties: + compatible: + items: + - enum: + - fsl,imx8mn-ddrc + - fsl,imx8mm-ddrc + - fsl,imx8mq-ddrc + - const: fsl,imx8m-ddrc + + reg: + maxItems: 1 + + clocks: + maxItems: 4 + + clock-names: + items: + - const: core + - const: pll + - const: alt + - const: apb + + operating-points-v2: true + opp-table: true + + devfreq-events: + description: Phandle of PMU node + $ref: "/schemas/types.yaml#/definitions/phandle" + +required: + - reg + - compatible + - clocks + - clock-names + +additionalProperties: false + +examples: + - | + #include <dt-bindings/clock/imx8mm-clock.h> + ddrc: dram-controller@3d400000 { + compatible = "fsl,imx8mm-ddrc", "fsl,imx8m-ddrc"; + reg = <0x3d400000 0x400000>; + clock-names = "core", "pll", "alt", "apb"; + clocks = <&clk IMX8MM_CLK_DRAM_CORE>, + <&clk IMX8MM_DRAM_PLL>, + <&clk IMX8MM_CLK_DRAM_ALT>, + <&clk IMX8MM_CLK_DRAM_APB>; + operating-points-v2 = <&ddrc_opp_table>; + };
Signature · 1 line
2.17.1
PATCH v4 4/6 PM / devfreq: Add dynamic scaling for imx8m ddr controller
In reply to Leonard CrestezAdd driver for dynamic scaling the DDR Controller on imx8m chips. Actual frequency switching is implemented inside TF-A, this driver wraps the SMC calls and synchronizes the clk tree.
The DRAM clocks on imx8m have the following structure (abridged):
+----------+ |\ +------+
| dram_pll |-------|M| dram_core | |
+----------+ |U|---------->| D |
/--|X| | D |
dram_alt_root | |/ | R |
| | C |
+---------+ | |
|FIX DIV/4| | |
+---------+ | |
composite: | | |
+----------+ | | |
| dram_alt |----/ | |
+----------+ | |
| dram_apb |-------------------->| |
+----------+ +------+The dram_pll is used for higher rates and dram_alt is used for lower rates. The dram_alt and dram_apb clocks are "imx composite" and their parent can also be modified.
This driver will prepare/enable the new parents ahead of switching (so that the expected roots are enabled) and afterwards it will call clk_set_parent to ensure the parents in clock framework are up-to-date.
The driver relies on dram_pll dram_alt and dram_apb being marked with CLK_GET_RATE_NOCACHE for rate updates.
--- drivers/devfreq/Kconfig | 10 + drivers/devfreq/Makefile | 1 + drivers/devfreq/imx8m-ddrc.c | 416 +++++++++++++++++++++++++++++++++++ 3 files changed, 427 insertions(+) create mode 100644 drivers/devfreq/imx8m-ddrc.c
drivers/devfreq/Kconfig +10 -0
@@ -90,10 +90,20 @@ config ARM_EXYNOS_BUS_DEVFREQ Each memory bus group could contain many memoby bus block. It reads PPMU counters of memory controllers by using DEVFREQ-event device and adjusts the operating frequencies and voltages with OPP support. This does not yet operate with optimal voltages. +config ARM_IMX8M_DDRC_DEVFREQ + tristate "i.MX8M DDRC DEVFREQ Driver" + depends on ARCH_MXC || COMPILE_TEST + select DEVFREQ_GOV_SIMPLE_ONDEMAND + select DEVFREQ_GOV_USERSPACE + select PM_OPP + help + This adds the DEVFREQ driver for the i.MX8M DDR Controller. It allows + adjusting DRAM frequency. + config ARM_TEGRA_DEVFREQ tristate "NVIDIA Tegra30/114/124/210 DEVFREQ Driver" depends on ARCH_TEGRA_3x_SOC || ARCH_TEGRA_114_SOC || \ ARCH_TEGRA_132_SOC || ARCH_TEGRA_124_SOC || \ ARCH_TEGRA_210_SOC || \
drivers/devfreq/Makefile +1 -0
@@ -7,10 +7,11 @@ obj-$(CONFIG_DEVFREQ_GOV_POWERSAVE) += governor_powersave.o obj-$(CONFIG_DEVFREQ_GOV_USERSPACE) += governor_userspace.o obj-$(CONFIG_DEVFREQ_GOV_PASSIVE) += governor_passive.o # DEVFREQ Drivers obj-$(CONFIG_ARM_EXYNOS_BUS_DEVFREQ) += exynos-bus.o +obj-$(CONFIG_ARM_IMX8M_DDRC_DEVFREQ) += imx8m-ddrc.o obj-$(CONFIG_ARM_RK3399_DMC_DEVFREQ) += rk3399_dmc.o obj-$(CONFIG_ARM_TEGRA_DEVFREQ) += tegra30-devfreq.o obj-$(CONFIG_ARM_TEGRA20_DEVFREQ) += tegra20-devfreq.o # DEVFREQ Event Drivers
drivers/devfreq/imx8m-ddrc.c new file +416 -0
416 lines not on this page · raw message
Signature · 1 line
2.17.1
PATCH v4 5/6 PM / devfreq: imx8m-ddrc: Measure bandwidth with perf
In reply to Leonard CrestezThe imx8m ddrc has a performance monitoring block attached which can be used to measure bandwidth usage and automatically adjust frequency.
There is already a perf driver for that block so instead of implementing a devfreq events driver use the in-kernel perf API to implement get_dev_status directly.
--- drivers/devfreq/imx8m-ddrc.c | 153 +++++++++++++++++++++++++++++++++++ 1 file changed, 153 insertions(+)
drivers/devfreq/imx8m-ddrc.c +153 -0
@@ -11,10 +11,13 @@#include <linux/pm_opp.h> #include <linux/clk.h> #include <linux/clk-provider.h> #include <linux/arm-smccc.h> +#include <asm/perf_event.h> +#include <linux/perf_event.h> +#define IMX_SIP_DDR_DVFS 0xc2000004/* Values starting from 0 switch to specific frequency */ #define IMX_SIP_DDR_FREQ_SET_HIGH 0x00
@@ -78,10 +81,22 @@ struct imx8m_ddrc {struct clk *dram_alt;struct clk *dram_apb;int freq_count;struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; + +/* For measuring load with perf events: */ +struct platform_device *pmu_pdev; +struct pmu *pmu; + +struct perf_event_attr rd_event_attr; +struct perf_event_attr wr_event_attr; +struct perf_event *rd_event; +struct perf_event *wr_event; + +u64 last_rd_val, last_rd_ena, last_rd_run; +u64 last_wr_val, last_wr_ena, last_wr_run; };static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv,unsigned long rate) {
@@ -245,10 +260,131 @@ static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) *freq =clk_get_rate (priv->dram_core);return 0 ; } +static int imx8m_ddrc_get_dev_status (struct device *dev, +struct devfreq_dev_status *stat) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + stat->current_frequency =clk_get_rate (priv->dram_core); + +if (priv->rd_event && priv->wr_event) { +u64 rd_delta, rd_val, rd_ena, rd_run; +u64 wr_delta, wr_val, wr_ena, wr_run; + + rd_val =perf_event_read_value (priv->rd_event, + &rd_ena, &rd_run); + wr_val =perf_event_read_value (priv->wr_event, + &wr_ena, &wr_run); + + rd_delta = (rd_val - priv->last_rd_val) * + (rd_ena - priv->last_rd_ena); +do_div (rd_delta, rd_run - priv->last_rd_run); + priv->last_rd_val = rd_val; + priv->last_rd_ena = rd_ena; + priv->last_rd_run = rd_run; + + wr_delta = (wr_val - priv->last_wr_val) * + (wr_ena - priv->last_wr_ena); +do_div (wr_delta, wr_run - priv->last_wr_run); + priv->last_wr_val = wr_val; + priv->last_wr_ena = wr_ena; + priv->last_wr_run = wr_run; + +/* magic numbers, possibly wrong */ + stat->busy_time =4 * (rd_delta + wr_delta); + stat->total_time = stat->current_frequency; + }else { + stat->busy_time =0 ; + stat->total_time =0 ; + } + +return 0 ; +} + +static int imx8m_ddrc_perf_disable (struct imx8m_ddrc *priv) +{ +/* release and set to NULL */ +if (!IS_ERR_OR_NULL (priv->rd_event)) +perf_event_release_kernel (priv->rd_event); +if (!IS_ERR_OR_NULL (priv->wr_event)) +perf_event_release_kernel (priv->wr_event); + priv->rd_event =NULL ; + priv->wr_event =NULL ; + +return 0 ; +} + +static int imx8m_ddrc_perf_enable (struct imx8m_ddrc *priv) +{ +int ret; + + priv->rd_event_attr.size =sizeof (priv->rd_event_attr); + priv->rd_event_attr.type = priv->pmu->type; + priv->rd_event_attr.config =0x2a ; + + priv->rd_event =perf_event_create_kernel_counter ( + &priv->rd_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->rd_event)) { + ret =PTR_ERR (priv->rd_event); + goto err; + } + + priv->wr_event_attr.size =sizeof (priv->wr_event_attr); + priv->wr_event_attr.type = priv->pmu->type; + priv->wr_event_attr.config =0x2b ; + + priv->wr_event =perf_event_create_kernel_counter ( + &priv->wr_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->wr_event)) { + ret =PTR_ERR (priv->wr_event); + goto err; + } + +return 0 ; + +err: +imx8m_ddrc_perf_disable (priv); +return ret; +} + +static int imx8m_ddrc_init_events (struct device *dev, +struct device_node *events_node) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct device_driver *driver; + +/* +* We need pmu->type for perf_event_attr but there is no API for +* mapping device_node to pmu. Fetch private data for imx-ddr-pmu and +* cast that to a struct pmu instead. +*/ + priv->pmu_pdev =of_find_device_by_node (events_node); +if (!priv->pmu_pdev) +return -EPROBE_DEFER ; + driver = priv->pmu_pdev->dev.driver; +if (!driver) +return -EPROBE_DEFER ; +if (strcmp (driver->name,"imx-ddr-pmu" )) { +dev_warn (dev,"devfreq-events node %pOF has unexpected driver %s\n" , + events_node, driver->name); +return -ENODEV ; + } + + priv->pmu =platform_get_drvdata (priv->pmu_pdev); +if (!priv->pmu) { +dev_err (dev,"devfreq-events device missing private data\n" ); +return -EINVAL ; + } + +dev_dbg (dev,"events from pmu %s\n" , priv->pmu->name); + +return imx8m_ddrc_perf_enable (priv); +} +static int imx8m_ddrc_init_freq_info (struct device *dev) {struct imx8m_ddrc *priv =dev_get_drvdata (dev);struct arm_smccc_res res;int index;
@@ -328,17 +464,23 @@ static int imx8m_ddrc_check_opps(struct device *dev)return 0 ; }static void imx8m_ddrc_exit (struct device *dev) { +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev); +dev_pm_opp_of_remove_table (dev); }static int imx8m_ddrc_probe (struct platform_device *pdev) {struct device *dev = &pdev->dev;struct imx8m_ddrc *priv; +struct device_node *events_node;const char *gov =DEVFREQ_GOV_USERSPACE ;int ret; priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL );if (!priv)
@@ -350,10 +492,19 @@ static int imx8m_ddrc_probe(struct platform_device *pdev)if (ret) {dev_err (dev,"failed to init firmware freq info: %d\n" , ret);return ret; } + events_node =of_parse_phandle (dev->of_node,"devfreq-events" ,0 ); +if (events_node) { + ret =imx8m_ddrc_init_events (dev, events_node); +of_node_put (events_node); +if (ret) + goto err; + gov =DEVFREQ_GOV_SIMPLE_ONDEMAND ; + } + priv->dram_core =devm_clk_get (dev,"core" ); priv->dram_pll =devm_clk_get (dev,"pll" ); priv->dram_alt =devm_clk_get (dev,"alt" ); priv->dram_apb =devm_clk_get (dev,"apb" ); if (IS_ERR(priv->dram_core) ||
@@ -390,10 +541,12 @@ static int imx8m_ddrc_probe(struct platform_device *pdev) }return 0 ; err: +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev);dev_pm_opp_of_remove_table (dev);return ret; }static const struct of_device_id imx8m_ddrc_of_match[] = {
Signature · 1 line
2.17.1
PATCH v4 6/6 arm64: dts: imx8m: Add ddr controller nodes
In reply to Leonard CrestezThis is used by the imx-ddrc devfreq driver to implement dynamic frequency scaling of DRAM.
Add a devfreq-event link to the dram PMU in order to support on-demand scaling of ddrc based on measured dram bandwidth usage.
Support for proactive scaling via interconnect will come later. The high-performance bus masters which need that (display, vpu, gpu) are not yet enabled in upstream anyway.
--- arch/arm64/boot/dts/freescale/imx8mm-evk.dts | 18 ++++++++++++++ arch/arm64/boot/dts/freescale/imx8mm.dtsi | 13 +++++++++- .../boot/dts/freescale/imx8mn-ddr4-evk.dts | 18 ++++++++++++++ arch/arm64/boot/dts/freescale/imx8mn.dtsi | 13 +++++++++- arch/arm64/boot/dts/freescale/imx8mq-evk.dts | 24 +++++++++++++++++++ arch/arm64/boot/dts/freescale/imx8mq.dtsi | 13 +++++++++- 6 files changed, 96 insertions(+), 3 deletions(-)
arch/arm64/boot/dts/freescale/imx8mm-evk.dts +18 -0
@@ -75,10 +75,28 @@ &A53_0 { cpu-supply = <&buck2_reg>; }; +&ddrc { + operating-points-v2 = <&ddrc_opp_table>; + + ddrc_opp_table: opp-table { + compatible = "operating-points-v2"; + + opp-25M { + opp-hz = /bits/ 64 <25000000>; + }; + opp-100M { + opp-hz = /bits/ 64 <100000000>; + }; + opp-750M { + opp-hz = /bits/ 64 <750000000>; + }; + }; +}; + &fec1 { pinctrl-names = "default"; pinctrl-0 = <&pinctrl_fec1>; phy-mode = "rgmii-id"; phy-handle = <ðphy0>;
arch/arm64/boot/dts/freescale/imx8mm.dtsi +12 -1
@@ -856,11 +856,22 @@ #interrupt-cells = <3>; interrupt-controller; interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>; }; - ddr-pmu@3d800000 { + ddrc: dram-controller@3d400000 { + compatible = "fsl,imx8mm-ddrc", "fsl,imx8m-ddrc"; + reg = <0x3d400000 0x400000>; + clock-names = "core", "pll", "alt", "apb"; + clocks = <&clk IMX8MM_CLK_DRAM_CORE>, + <&clk IMX8MM_DRAM_PLL>, + <&clk IMX8MM_CLK_DRAM_ALT>, + <&clk IMX8MM_CLK_DRAM_APB>; + devfreq-events = <&ddr_pmu>; + }; + + ddr_pmu: ddr-pmu@3d800000 { compatible = "fsl,imx8mm-ddr-pmu", "fsl,imx8m-ddr-pmu"; reg = <0x3d800000 0x400000>; interrupt-parent = <&gic>; interrupts = <GIC_SPI 98 IRQ_TYPE_LEVEL_HIGH>; };
arch/arm64/boot/dts/freescale/imx8mn-ddr4-evk.dts +18 -0
@@ -15,10 +15,28 @@ &A53_0 { cpu-supply = <&buck2_reg>; }; +&ddrc { + operating-points-v2 = <&ddrc_opp_table>; + + ddrc_opp_table: opp-table { + compatible = "operating-points-v2"; + + opp-25M { + opp-hz = /bits/ 64 <25000000>; + }; + opp-100M { + opp-hz = /bits/ 64 <100000000>; + }; + opp-600M { + opp-hz = /bits/ 64 <600000000>; + }; + }; +}; + &i2c1 { pmic@4b { compatible = "rohm,bd71847"; reg = <0x4b>; pinctrl-0 = <&pinctrl_pmic>;
arch/arm64/boot/dts/freescale/imx8mn.dtsi +12 -1
@@ -757,11 +757,22 @@ #interrupt-cells = <3>; interrupt-controller; interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>; }; - ddr-pmu@3d800000 { + ddrc: dram-controller@3d400000 { + compatible = "fsl,imx8mn-ddrc", "fsl,imx8m-ddrc"; + reg = <0x3d400000 0x400000>; + clock-names = "core", "pll", "alt", "apb"; + clocks = <&clk IMX8MN_CLK_DRAM_CORE>, + <&clk IMX8MN_DRAM_PLL>, + <&clk IMX8MN_CLK_DRAM_ALT>, + <&clk IMX8MN_CLK_DRAM_APB>; + devfreq-events = <&ddr_pmu>; + }; + + ddr_pmu: ddr-pmu@3d800000 { compatible = "fsl,imx8mn-ddr-pmu", "fsl,imx8m-ddr-pmu"; reg = <0x3d800000 0x400000>; interrupts = <GIC_SPI 98 IRQ_TYPE_LEVEL_HIGH>; }; };
arch/arm64/boot/dts/freescale/imx8mq-evk.dts +24 -0
@@ -103,10 +103,34 @@ &A53_3 { cpu-supply = <&buck2_reg>; }; +&ddrc { + operating-points-v2 = <&ddrc_opp_table>; + + ddrc_opp_table: opp-table { + compatible = "operating-points-v2"; + + opp-25M { + opp-hz = /bits/ 64 <25000000>; + }; + opp-100M { + opp-hz = /bits/ 64 <100000000>; + }; + /* + * On imx8mq B0 PLL can't be bypassed so low bus is 166M + */ + opp-166M { + opp-hz = /bits/ 64 <166935483>; + }; + opp-800M { + opp-hz = /bits/ 64 <800000000>; + }; + }; +}; + &fec1 { pinctrl-names = "default"; pinctrl-0 = <&pinctrl_fec1>; phy-mode = "rgmii-id"; phy-handle = <ðphy0>;
arch/arm64/boot/dts/freescale/imx8mq.dtsi +12 -1
@@ -1111,11 +1111,22 @@ interrupt-controller; interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>; interrupt-parent = <&gic>; }; - ddr-pmu@3d800000 { + ddrc: dram-controller@3d400000 { + compatible = "fsl,imx8mq-ddrc", "fsl,imx8m-ddrc"; + reg = <0x3d400000 0x400000>; + clock-names = "core", "pll", "alt", "apb"; + clocks = <&clk IMX8MQ_CLK_DRAM_CORE>, + <&clk IMX8MQ_DRAM_PLL_OUT>, + <&clk IMX8MQ_CLK_DRAM_ALT>, + <&clk IMX8MQ_CLK_DRAM_APB>; + devfreq-events = <&ddr_pmu>; + }; + + ddr_pmu: ddr-pmu@3d800000 { compatible = "fsl,imx8mq-ddr-pmu", "fsl,imx8m-ddr-pmu"; reg = <0x3d800000 0x400000>; interrupt-parent = <&gic>; interrupts = <GIC_SPI 98 IRQ_TYPE_LEVEL_HIGH>; };
Signature · 1 line
2.17.1
Chanwoo Choi
maintainerHi Leonard,
Quoted 47 lines · Leonard Crestez
Add driver for dynamic scaling the DDR Controller on imx8m chips. Actual frequency switching is implemented inside TF-A, this driver wraps the SMC calls and synchronizes the clk tree.
The DRAM clocks on imx8m have the following structure (abridged):
+----------+ |\ +------+ | dram_pll |-------|M| dram_core | | +----------+ |U|---------->| D | /--|X| | D | dram_alt_root | |/ | R | | | C | +---------+ | | |FIX DIV/4| | | +---------+ | | composite: | | | +----------+ | | | | dram_alt |----/ | | +----------+ | | | dram_apb |-------------------->| | +----------+ +------+The
dram_pllis used for higher rates anddram_altis used for lower rates. Thedram_altanddram_apbclocks are "imx composite" and their parent can also be modified.This driver will prepare/enable the new parents ahead of switching (so that the expected roots are enabled) and afterwards it will call
clk_set_parentto ensure the parents in clock framework are up-to-date.The driver relies on
dram_plldram_altanddram_apbbeing marked withCLK_GET_RATE_NOCACHEfor rate updates.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/Kconfig | 10 + drivers/devfreq/Makefile | 1 + drivers/devfreq/imx8m-ddrc.c | 416 +++++++++++++++++++++++++++++++++++ 3 files changed, 427 insertions(+) create mode 100644 drivers/devfreq/imx8m-ddrc.cdrivers/devfreq/Kconfig +0
-0@@ -90,10 +90,20 @@ config ARM_EXYNOS_BUS_DEVFREQ Each memory bus group could contain many memoby bus block. It reads
drivers/devfreq/Kconfig config ARM_EXYNOS_BUS_DEVFREQ · line 100
Quoted 10 lines
drivers/devfreq/Kconfig +6
-0PPMU counters of memory controllers by using DEVFREQ-event device and adjusts the operating frequencies and voltages with OPP support. This does not yet operate with optimal voltages. +config ARM_IMX8M_DDRC_DEVFREQ + tristate "i.MX8M DDRC DEVFREQ Driver" + depends on ARCH_MXC || COMPILE_TEST + select DEVFREQ_GOV_SIMPLE_ONDEMAND + select DEVFREQ_GOV_USERSPACE + select PM_OPP
It doesn't need to add. CONFIG_DEVFREQ add 'select PM_OPP' because 'PM_OPP' is mandatory for devfreq.
Quoted 134 lines
drivers/devfreq/Kconfig +4
-0+ help + This adds the DEVFREQ driver for the i.MX8M DDR Controller. It allows + adjusting DRAM frequency. + config ARM_TEGRA_DEVFREQ tristate "NVIDIA Tegra30/114/124/210 DEVFREQ Driver" depends on ARCH_TEGRA_3x_SOC || ARCH_TEGRA_114_SOC || \ ARCH_TEGRA_132_SOC || ARCH_TEGRA_124_SOC || \ ARCH_TEGRA_210_SOC || \drivers/devfreq/Makefile +1
-0@@ -7,10 +7,11 @@ obj-$(CONFIG_DEVFREQ_GOV_POWERSAVE) += governor_powersave.o obj-$(CONFIG_DEVFREQ_GOV_USERSPACE) += governor_userspace.o obj-$(CONFIG_DEVFREQ_GOV_PASSIVE) += governor_passive.o # DEVFREQ Drivers obj-$(CONFIG_ARM_EXYNOS_BUS_DEVFREQ) += exynos-bus.o +obj-$(CONFIG_ARM_IMX8M_DDRC_DEVFREQ) += imx8m-ddrc.o obj-$(CONFIG_ARM_RK3399_DMC_DEVFREQ) += rk3399_dmc.o obj-$(CONFIG_ARM_TEGRA_DEVFREQ) += tegra30-devfreq.o obj-$(CONFIG_ARM_TEGRA20_DEVFREQ) += tegra20-devfreq.o # DEVFREQ Event Driversdrivers/devfreq/imx8m-ddrc.c new file +103
-0@@ -0,0 +1,416 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright 2019 NXP +*/ + +#include <linux/module.h> +#include <linux/device.h> +#include <linux/of_device.h> +#include <linux/platform_device.h> +#include <linux/devfreq.h> +#include <linux/pm_opp.h> +#include <linux/clk.h> +#include <linux/clk-provider.h> +#include <linux/arm-smccc.h> + +#define IMX_SIP_DDR_DVFS 0xc2000004 + +/* Values starting from 0 switch to specific frequency */ +#define IMX_SIP_DDR_FREQ_SET_HIGH 0x00 + +/* Deprecated after moving IRQ handling to ATF */ +#define IMX_SIP_DDR_DVFS_WAIT_CHANGE 0x0F + +/* Query available frequencies. */ +#define IMX_SIP_DDR_DVFS_GET_FREQ_COUNT 0x10 +#define IMX_SIP_DDR_DVFS_GET_FREQ_INFO 0x11 + +/* +* This should be in a 1:1 mapping with devicetree OPPs but +* firmware provides additional info. +*/ +struct imx8m_ddrc_freq { +unsigned long rate; +unsigned long smcarg; +int dram_core_parent_index; +int dram_alt_parent_index; +int dram_apb_parent_index; +}; + +/* Hardware limitation */ +#define IMX8M_DDRC_MAX_FREQ_COUNT 4 + +/* +* i.MX8M DRAM Controller clocks have the following structure (abridged): +* +* +----------+ |\ +------+ +* | dram_pll |-------|M| dram_core | | +* +----------+ |U|---------->| D | +* /--|X| | D | +* dram_alt_root | |/ | R | +* | | C | +* +---------+ | | +* |FIX DIV/4| | | +* +---------+ | | +* composite: | | | +* +----------+ | | | +* | dram_alt |----/ | | +* +----------+ | | +* | dram_apb |-------------------->| | +* +----------+ +------+ +* +* The dram_pll is used for higher rates and dram_alt is used for lower rates. +* +* Frequency switching is implemented in TF-A (via SMC call) and can change the +* configuration of the clocks, including mux parents. The dram_alt and +* dram_apb clocks are "imx composite" and their parent can change too. +* +* We need to prepare/enable the new mux parents head of switching and update +* their information afterwards. +*/ +struct imx8m_ddrc { +struct devfreq_dev_profile profile; +struct devfreq *devfreq; + +/* For frequency switching: */ +struct clk *dram_core; +struct clk *dram_pll; +struct clk *dram_alt; +struct clk *dram_apb; + +int freq_count; +struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; +}; + +static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv, +unsigned long rate) +{ +struct imx8m_ddrc_freq *freq; +int i; + +/* +* Firmware reports values in MT/s, so we round-down from Hz +* Rounding is extra generous to ensure a match. +*/ + rate =DIV_ROUND_CLOSEST (rate,250000 ); +for (i =0 ; i < priv->freq_count; ++i) { + freq = &priv->freq_table[i]; +if (freq->rate == rate || + freq->rate +1 == rate || + freq->rate -1 == rate) +return freq; + } +
drivers/devfreq/imx8m-ddrc.c line 113
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+return NULL ; +} + +static void imx8m_ddrc_smc_set_freq (int target_freq) +{ +struct arm_smccc_res res; +u32 online_cpus =0 ; +int cpu; + +local_irq_disable ();
local_irq_disable is more proper than local_irq_save()?
drivers/devfreq/imx8m-ddrc.c line 122
Quoted 9 lines
drivers/devfreq/imx8m-ddrc.c new file +9
-0+ +for_each_online_cpu (cpu) + online_cpus |= (1 << (cpu *8 )); + +/* change the ddr freqency */ +arm_smccc_smc (IMX_SIP_DDR_DVFS , target_freq, online_cpus, +0 ,0 ,0 ,0 ,0 , &res); + +local_irq_enable ();
ditto. local_irq_restore() instead of local_irq_enable()
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+} + +struct clk *clk_get_parent_by_index (struct clk *clk,int index) +{ +struct clk_hw *hw; + + hw =clk_hw_get_parent_by_index (__clk_get_hw (clk), index); + +return hw ? hw->clk :NULL ; +}
drivers/devfreq/imx8m-ddrc.c line 142
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+ +static int imx8m_ddrc_set_freq (struct device *dev,struct imx8m_ddrc_freq *freq) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct clk *new_dram_core_parent; +struct clk *new_dram_alt_parent; +struct clk *new_dram_apb_parent; +int ret; + + new_dram_core_parent = clk_get_parent_by_index(
You can use 'clk_hw_get_parent_by_index' directly. And, you need to check whether the return value is NULL or not.
drivers/devfreq/imx8m-ddrc.c line 144
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_core, freq->dram_core_parent_index - 1); + new_dram_alt_parent = clk_get_parent_by_index(
ditto.
drivers/devfreq/imx8m-ddrc.c line 146
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_alt, freq->dram_alt_parent_index - 1); + new_dram_apb_parent = clk_get_parent_by_index(
ditto.
drivers/devfreq/imx8m-ddrc.c line 158
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c new file +12
-0+ priv->dram_apb, freq->dram_apb_parent_index - 1); + +/* increase reference counts and ensure clks are ON before switch */ + ret =clk_prepare_enable (new_dram_core_parent); +if (ret) { +dev_err (dev,"failed enable new dram_core parent: %d\n" , ret); + goto out; + } + ret =clk_prepare_enable (new_dram_alt_parent); +if (ret) { +dev_err (dev,"failed enable new dram_alt parent: %d\n" , ret); + goto out_dis_core;
I think that 'dis' is not general expression for 'disable'. Just, I think that 'out_disable_core' is better than 'out_dis_core'.
drivers/devfreq/imx8m-ddrc.c line 163
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } + ret =clk_prepare_enable (new_dram_apb_parent); +if (ret) { +dev_err (dev,"failed enable new dram_apb parent: %d\n" , ret); + goto out_dis_alt;
ditto.
drivers/devfreq/imx8m-ddrc.c line 171
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +8
-0+ } + +imx8m_ddrc_smc_set_freq (freq->smcarg); + +/* update parents in clk tree after switch. */ + ret =clk_set_parent (priv->dram_core, new_dram_core_parent); +if (ret) +dev_err (dev,"failed set dram_core parent: %d\n" , ret);
if you don't return directly, you better to use 'dev_warn' instead of 'dev_err'.
drivers/devfreq/imx8m-ddrc.c line 175
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+if (new_dram_alt_parent) { + ret =clk_set_parent (priv->dram_alt, new_dram_alt_parent); +if (ret) +dev_err (dev,"failed set dram_alt parent: %d\n" , ret);
ditto. Use dev_warn.
drivers/devfreq/imx8m-ddrc.c line 180
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } +if (new_dram_apb_parent) { + ret =clk_set_parent (priv->dram_apb, new_dram_apb_parent); +if (ret) +dev_err (dev,"failed set dram_apb parent: %d\n" , ret);
ditto. Use dev_warn.
drivers/devfreq/imx8m-ddrc.c line 186
Quoted 6 lines
drivers/devfreq/imx8m-ddrc.c new file +6
-0+ } + +/* +* Explicitly refresh dram PLL rate. +* +* Even if it's marked with CLK_GET_RATE_NOCACHE the rate will not be
nitpick:
What is more proper description either or 'Even if' or 'if' ?
Quoted 31 lines
drivers/devfreq/imx8m-ddrc.c new file +31
-0+* automatically refreshed when clk_get_rate is called on children. +*/ +clk_get_rate (priv->dram_pll); + +/* +* clk_set_parent transfer the reference count from old parent. +* now we drop extra reference counts used during the switch +*/ +clk_disable_unprepare (new_dram_apb_parent); +out_dis_alt: +clk_disable_unprepare (new_dram_alt_parent); +out_dis_core: +clk_disable_unprepare (new_dram_core_parent); +out: +return ret; +} + +static int imx8m_ddrc_target (struct device *dev,unsigned long *freq,u32 flags) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *new_opp; +unsigned long old_freq, new_freq; +int ret; + + new_opp =devfreq_recommended_opp (dev, freq, flags); +if (IS_ERR (new_opp)) { + ret =PTR_ERR (new_opp); +dev_err (dev,"failed to get recommended opp: %d\n" , ret); +return ret; + }
drivers/devfreq/imx8m-ddrc.c line 227
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+dev_pm_opp_put (new_opp); + + old_freq =clk_get_rate (priv->dram_core); +if (*freq == old_freq) +return 0 ; + + freq_info =imx8m_ddrc_find_freq (priv, *freq); +if (!freq_info) +return -EINVAL ; + ret =imx8m_ddrc_set_freq (dev, freq_info);
Need to check whether 'ret' is valid or not.
drivers/devfreq/imx8m-ddrc.c line 231
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+ +/* Also read back the clk rate to verify switch was correct */ + new_freq =clk_get_rate (priv->dram_core); +if (ret || *freq != new_freq)
You should check 'ret' right after imx8m_ddrc_set_freq() instead of this position.
Quoted 17 lines
drivers/devfreq/imx8m-ddrc.c new file +17
-0+dev_err (dev,"ddrc failed freq set to %lu from %lu, now at %lu\n" , + old_freq, *freq, new_freq); + else + dev_dbg(dev, "ddrc freq set to %lu (was %lu)\n", + *freq, old_freq); + +return ret; +} + +static int imx8m_ddrc_get_cur_freq (struct device *dev,unsigned long *freq) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + *freq =clk_get_rate (priv->dram_core); + +return 0 ; +}
drivers/devfreq/imx8m-ddrc.c line 258
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+ +static int imx8m_ddrc_init_freq_info (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct arm_smccc_res res; +int index; + +/* +* An error here means DDR DVFS API not supported by firmware +*/
Don't need to add multiple line comments. It is possible to be change as following:
/* An error here means DDR DVFS API not supported by firmware */
Quoted 30 lines
drivers/devfreq/imx8m-ddrc.c new file +30
-0+arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_COUNT , +0 ,0 ,0 ,0 ,0 ,0 , &res); + priv->freq_count = res.a0; +if (priv->freq_count <=0 || + priv->freq_count >IMX8M_DDRC_MAX_FREQ_COUNT ) +return -ENODEV ; + +for (index =0 ; index < priv->freq_count; ++index) { +struct imx8m_ddrc_freq *freq = &priv->freq_table[index]; + +arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_INFO , + index,0 ,0 ,0 ,0 ,0 , &res); +/* Result should be strictly positive */ +if ((long )res.a0 <=0 ) +return -ENODEV ; + + freq->rate = res.a0; + freq->smcarg = index; + freq->dram_core_parent_index = res.a1; + freq->dram_alt_parent_index = res.a2; + freq->dram_apb_parent_index = res.a3; + +/* dram_core has 2 options: dram_pll or dram_alt_root */ +if (freq->dram_core_parent_index !=1 && + freq->dram_core_parent_index !=2 ) +return -ENODEV ; +/* dram_apb and dram_alt have exactly 8 possible parents */ +if (freq->dram_alt_parent_index >8 || + freq->dram_apb_parent_index >8 ) +return -ENODEV ;
drivers/devfreq/imx8m-ddrc.c line 298
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+/* dram_core from alt requires explicit dram_alt parent */ +if (freq->dram_core_parent_index ==2 && + freq->dram_alt_parent_index ==0 ) +return -ENODEV ; + } + +return 0 ; +} + +/* imx8m_ddrc_check_opps() - disable OPPs not supported by firmware */
nitpick:
On this driver, there are no some description for internal function. In order to keep the coding style, you better to move this comment into function before 'freq = ULONG_MAX'.
Quoted 119 lines
drivers/devfreq/imx8m-ddrc.c new file +118
-0+static int imx8m_ddrc_check_opps (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *opp; +unsigned long freq; + + freq =ULONG_MAX ; +while (true) { + opp =dev_pm_opp_find_freq_floor (dev, &freq); +if (opp ==ERR_PTR (-ERANGE )) +break ; +if (IS_ERR (opp)) { +dev_err (dev,"Failed enumerating OPPs: %ld\n" , +PTR_ERR (opp)); +return PTR_ERR (opp); + } +dev_pm_opp_put (opp); + + freq_info =imx8m_ddrc_find_freq (priv, freq); +if (!freq_info) { +dev_info (dev,"Disable unsupported OPP %luHz %luMT/s\n" , + freq,DIV_ROUND_CLOSEST (freq,250000 )); +dev_pm_opp_disable (dev, freq); + } + + freq--; + } + +return 0 ; +} + +static void imx8m_ddrc_exit (struct device *dev) +{ +dev_pm_opp_of_remove_table (dev); +} + +static int imx8m_ddrc_probe (struct platform_device *pdev) +{ +struct device *dev = &pdev->dev; +struct imx8m_ddrc *priv; +const char *gov =DEVFREQ_GOV_USERSPACE ; +int ret; + + priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL ); +if (!priv) +return -ENOMEM ; + +platform_set_drvdata (pdev, priv); + + ret =imx8m_ddrc_init_freq_info (dev); +if (ret) { +dev_err (dev,"failed to init firmware freq info: %d\n" , ret); +return ret; + } + + priv->dram_core =devm_clk_get (dev,"core" ); + priv->dram_pll =devm_clk_get (dev,"pll" ); + priv->dram_alt =devm_clk_get (dev,"alt" ); + priv->dram_apb =devm_clk_get (dev,"apb" ); +if (IS_ERR (priv->dram_core) || +IS_ERR (priv->dram_pll) || +IS_ERR (priv->dram_alt) || +IS_ERR (priv->dram_apb)) { + ret =PTR_ERR (priv->devfreq); +dev_err (dev,"failed to fetch clocks: %d\n" , ret); +return ret; + } + + ret =dev_pm_opp_of_add_table (dev); +if (ret <0 ) { +dev_err (dev,"failed to get OPP table\n" ); +return ret; + } + + ret =imx8m_ddrc_check_opps (dev); +if (ret <0 ) + goto err; + + priv->profile.polling_ms =1000 ; + priv->profile.target = imx8m_ddrc_target; + priv->profile.exit = imx8m_ddrc_exit; + priv->profile.get_cur_freq = imx8m_ddrc_get_cur_freq; + priv->profile.initial_freq =clk_get_rate (priv->dram_core); + + priv->devfreq =devm_devfreq_add_device (dev, &priv->profile, + gov,NULL ); +if (IS_ERR (priv->devfreq)) { + ret =PTR_ERR (priv->devfreq); +dev_err (dev,"failed to add devfreq device: %d\n" , ret); + goto err; + } + +return 0 ; + +err: +dev_pm_opp_of_remove_table (dev); +return ret; +} + +static const struct of_device_id imx8m_ddrc_of_match[] = { + { .compatible ="fsl,imx8m-ddrc" , }, + {/* sentinel */ }, +}; +MODULE_DEVICE_TABLE (of, imx8m_ddrc_of_match); + +static struct platform_driver imx8m_ddrc_platdrv = { + .probe = imx8m_ddrc_probe, + .driver = { + .name ="imx8m-ddrc-devfreq" , + .of_match_table =of_match_ptr (imx8m_ddrc_of_match), + }, +}; +module_platform_driver (imx8m_ddrc_platdrv); + +MODULE_DESCRIPTION ("i.MX8M DDR Controller frequency driver" ); +MODULE_AUTHOR ("Leonard Crestez <[email protected]>" ); +MODULE_LICENSE ("GPL v2" );
Signature · 3 lines
Best Regards, Chanwoo Choi Samsung Electronics
Chanwoo Choi
maintainerRe: [PATCH v4 5/6] PM / devfreq: imx8m-ddrc: Measure bandwidth with perf
In reply to Leonard CrestezHi Leonard,
Quoted 104 lines · Leonard Crestez
The imx8m ddrc has a performance monitoring block attached which can be used to measure bandwidth usage and automatically adjust frequency.
There is already a perf driver for that block so instead of implementing a devfreq events driver use the in-kernel perf API to implement
get_dev_statusdirectly.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/imx8m-ddrc.c | 153 +++++++++++++++++++++++++++++++++++ 1 file changed, 153 insertions(+)drivers/devfreq/imx8m-ddrc.c +60
-0@@ -11,10 +11,13 @@#include <linux/pm_opp.h> #include <linux/clk.h> #include <linux/clk-provider.h> #include <linux/arm-smccc.h> +#include <asm/perf_event.h> +#include <linux/perf_event.h> +#define IMX_SIP_DDR_DVFS 0xc2000004/* Values starting from 0 switch to specific frequency */ #define IMX_SIP_DDR_FREQ_SET_HIGH 0x00@@ -78,10 +81,22 @@ struct imx8m_ddrc {struct clk *dram_alt;struct clk *dram_apb;int freq_count;struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; + +/* For measuring load with perf events: */ +struct platform_device *pmu_pdev; +struct pmu *pmu; + +struct perf_event_attr rd_event_attr; +struct perf_event_attr wr_event_attr; +struct perf_event *rd_event; +struct perf_event *wr_event; + +u64 last_rd_val, last_rd_ena, last_rd_run; +u64 last_wr_val, last_wr_ena, last_wr_run; };static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv,unsigned long rate) {@@ -245,10 +260,131 @@ static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) *freq =clk_get_rate (priv->dram_core);return 0 ; } +static int imx8m_ddrc_get_dev_status (struct device *dev, +struct devfreq_dev_status *stat) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + stat->current_frequency =clk_get_rate (priv->dram_core); + +if (priv->rd_event && priv->wr_event) { +u64 rd_delta, rd_val, rd_ena, rd_run; +u64 wr_delta, wr_val, wr_ena, wr_run; + + rd_val =perf_event_read_value (priv->rd_event, + &rd_ena, &rd_run); + wr_val =perf_event_read_value (priv->wr_event, + &wr_ena, &wr_run); + + rd_delta = (rd_val - priv->last_rd_val) * + (rd_ena - priv->last_rd_ena); +do_div (rd_delta, rd_run - priv->last_rd_run); + priv->last_rd_val = rd_val; + priv->last_rd_ena = rd_ena; + priv->last_rd_run = rd_run; + + wr_delta = (wr_val - priv->last_wr_val) * + (wr_ena - priv->last_wr_ena); +do_div (wr_delta, wr_run - priv->last_wr_run); + priv->last_wr_val = wr_val; + priv->last_wr_ena = wr_ena; + priv->last_wr_run = wr_run; + +/* magic numbers, possibly wrong */ + stat->busy_time =4 * (rd_delta + wr_delta); + stat->total_time = stat->current_frequency; + }else { + stat->busy_time =0 ; + stat->total_time =0 ; + } + +return 0 ; +} + +static int imx8m_ddrc_perf_disable (struct imx8m_ddrc *priv) +{ +/* release and set to NULL */ +if (!IS_ERR_OR_NULL (priv->rd_event))
drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 319
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+perf_event_release_kernel (priv->rd_event); +if (!IS_ERR_OR_NULL (priv->wr_event)) +perf_event_release_kernel (priv->wr_event); + priv->rd_event =NULL ; + priv->wr_event =NULL ; + +return 0 ; +} + +static int imx8m_ddrc_perf_enable (struct imx8m_ddrc *priv)
Actually, this function have to call the just function for enabling the bandwidth monitoring instead of writing the detailed something. It looks like that you move the part of the different device driver into here.
Quoted 24 lines
drivers/devfreq/imx8m-ddrc.c +24
-0+{ +int ret; + + priv->rd_event_attr.size =sizeof (priv->rd_event_attr); + priv->rd_event_attr.type = priv->pmu->type; + priv->rd_event_attr.config =0x2a ; + + priv->rd_event =perf_event_create_kernel_counter ( + &priv->rd_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->rd_event)) { + ret =PTR_ERR (priv->rd_event); + goto err; + } + + priv->wr_event_attr.size =sizeof (priv->wr_event_attr); + priv->wr_event_attr.type = priv->pmu->type; + priv->wr_event_attr.config =0x2b ; + + priv->wr_event =perf_event_create_kernel_counter ( + &priv->wr_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->wr_event)) { + ret =PTR_ERR (priv->wr_event); + goto err; + }
drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 353
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+ +return 0 ; + +err: +imx8m_ddrc_perf_disable (priv); +return ret; +} + +static int imx8m_ddrc_init_events (struct device *dev, +struct device_node *events_node)
ditto.
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c +12
-0+{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct device_driver *driver; + +/* +* We need pmu->type for perf_event_attr but there is no API for +* mapping device_node to pmu. Fetch private data for imx-ddr-pmu and +* cast that to a struct pmu instead. +*/ + priv->pmu_pdev =of_find_device_by_node (events_node); +if (!priv->pmu_pdev) +return -EPROBE_DEFER ;
drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 375
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+ driver = priv->pmu_pdev->dev.driver; +if (!driver) +return -EPROBE_DEFER ; +if (strcmp (driver->name,"imx-ddr-pmu" )) { +dev_warn (dev,"devfreq-events node %pOF has unexpected driver %s\n" , + events_node, driver->name); +return -ENODEV ; + } + + priv->pmu =platform_get_drvdata (priv->pmu_pdev);
It seems that you access the different device driver without any standard interface from some linux kernel subsystem.
For the communication or control between different device drivers, we have to use the standard interface instead of direct access or call. I think that it make it too tightly coupled driver between them.
So, I developed the devfreq-event subsystem for this reason in order to remove the non-standard direct access to other device driver.
Unfortunately, I can't agree this approach. I don't prefer to use the direct access of other device driver(imx-ddr-pmu). You need to use standard interface provided from subsystem. or You need to make the new device driver with devfreq-event subsystem.
Quoted 73 lines
drivers/devfreq/imx8m-ddrc.c +27
-0+if (!priv->pmu) { +dev_err (dev,"devfreq-events device missing private data\n" ); +return -EINVAL ; + } + +dev_dbg (dev,"events from pmu %s\n" , priv->pmu->name); + +return imx8m_ddrc_perf_enable (priv); +} +static int imx8m_ddrc_init_freq_info (struct device *dev) {struct imx8m_ddrc *priv =dev_get_drvdata (dev);struct arm_smccc_res res;int index;@@ -328,17 +464,23 @@ static int imx8m_ddrc_check_opps(struct device *dev)return 0 ; }static void imx8m_ddrc_exit (struct device *dev) { +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev); +dev_pm_opp_of_remove_table (dev); }static int imx8m_ddrc_probe (struct platform_device *pdev) {struct device *dev = &pdev->dev;struct imx8m_ddrc *priv; +struct device_node *events_node;const char *gov =DEVFREQ_GOV_USERSPACE ;int ret; priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL );if (!priv)@@ -350,10 +492,19 @@ static int imx8m_ddrc_probe(struct platform_device *pdev)if (ret) {dev_err (dev,"failed to init firmware freq info: %d\n" , ret);return ret; } + events_node =of_parse_phandle (dev->of_node,"devfreq-events" ,0 ); +if (events_node) { + ret =imx8m_ddrc_init_events (dev, events_node); +of_node_put (events_node); +if (ret) + goto err; + gov =DEVFREQ_GOV_SIMPLE_ONDEMAND ; + } + priv->dram_core =devm_clk_get (dev,"core" ); priv->dram_pll =devm_clk_get (dev,"pll" ); priv->dram_alt =devm_clk_get (dev,"alt" ); priv->dram_apb =devm_clk_get (dev,"apb" ); if (IS_ERR(priv->dram_core) ||@@ -390,10 +541,12 @@ static int imx8m_ddrc_probe(struct platform_device *pdev) }return 0 ; err: +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev);dev_pm_opp_of_remove_table (dev);return ret; }static const struct of_device_id imx8m_ddrc_of_match[] = {
Signature · 3 lines
Best Regards, Chanwoo Choi Samsung Electronics
Quoted 54 lines · Chanwoo Choi
Hi Leonard,
Quoted 51 lines · Leonard Crestez
Add driver for dynamic scaling the DDR Controller on imx8m chips. Actual frequency switching is implemented inside TF-A, this driver wraps the SMC calls and synchronizes the clk tree.
The DRAM clocks on imx8m have the following structure (abridged):
+----------+ |\ +------+ | dram_pll |-------|M| dram_core | | +----------+ |U|---------->| D | /--|X| | D | dram_alt_root | |/ | R | | | C | +---------+ | | |FIX DIV/4| | | +---------+ | | composite: | | | +----------+ | | | | dram_alt |----/ | | +----------+ | | | dram_apb |-------------------->| | +----------+ +------+The
dram_pllis used for higher rates anddram_altis used for lower rates. Thedram_altanddram_apbclocks are "imx composite" and their parent can also be modified.This driver will prepare/enable the new parents ahead of switching (so that the expected roots are enabled) and afterwards it will call
clk_set_parentto ensure the parents in clock framework are up-to-date.The driver relies on
dram_plldram_altanddram_apbbeing marked withCLK_GET_RATE_NOCACHEfor rate updates.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/Kconfig | 10 + drivers/devfreq/Makefile | 1 + drivers/devfreq/imx8m-ddrc.c | 416 +++++++++++++++++++++++++++++++++++ 3 files changed, 427 insertions(+)create mode 100644
drivers/devfreq/imx8m-ddrc.cdrivers/devfreq/Kconfig +0
-0@@ -90,10 +90,20 @@ config ARM_EXYNOS_BUS_DEVFREQ Each memory bus group could contain many memoby bus block. It reads PPMU counters of memory controllers by using DEVFREQ-event device and adjusts the operating frequencies and voltages with OPP support. This does not yet operate with optimal voltages.
Quoted 9 lines
drivers/devfreq/Kconfig config ARM_EXYNOS_BUS_DEVFREQ · line 100
Quoted 6 lines
drivers/devfreq/Kconfig +6
-0+config ARM_IMX8M_DDRC_DEVFREQ + tristate "i.MX8M DDRC DEVFREQ Driver" + depends on ARCH_MXC || COMPILE_TEST + select DEVFREQ_GOV_SIMPLE_ONDEMAND + select DEVFREQ_GOV_USERSPACE + select PM_OPPIt doesn't need to add.
CONFIG_DEVFREQadd 'selectPM_OPP' because 'PM_OPP' is mandatory for devfreq.
OK.
All other drivers select PM_OPP, this was just copied
Quoted 138 lines
Quoted 138 lines
drivers/devfreq/Kconfig +4
-0+ help + This adds the DEVFREQ driver for the i.MX8M DDR Controller. It allows + adjusting DRAM frequency. + config ARM_TEGRA_DEVFREQ tristate "NVIDIA Tegra30/114/124/210 DEVFREQ Driver" depends on ARCH_TEGRA_3x_SOC || ARCH_TEGRA_114_SOC || \ ARCH_TEGRA_132_SOC || ARCH_TEGRA_124_SOC || \ ARCH_TEGRA_210_SOC || \drivers/devfreq/Makefile +1
-0@@ -7,10 +7,11 @@ obj-$(CONFIG_DEVFREQ_GOV_POWERSAVE) += governor_powersave.o obj-$(CONFIG_DEVFREQ_GOV_USERSPACE) += governor_userspace.o obj-$(CONFIG_DEVFREQ_GOV_PASSIVE) += governor_passive.o # DEVFREQ Drivers obj-$(CONFIG_ARM_EXYNOS_BUS_DEVFREQ) += exynos-bus.o +obj-$(CONFIG_ARM_IMX8M_DDRC_DEVFREQ) += imx8m-ddrc.o obj-$(CONFIG_ARM_RK3399_DMC_DEVFREQ) += rk3399_dmc.o obj-$(CONFIG_ARM_TEGRA_DEVFREQ) += tegra30-devfreq.o obj-$(CONFIG_ARM_TEGRA20_DEVFREQ) += tegra20-devfreq.o # DEVFREQ Event Driversdrivers/devfreq/imx8m-ddrc.c new file +107
-0@@ -0,0 +1,416 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright 2019 NXP +*/ + +#include <linux/module.h> +#include <linux/device.h> +#include <linux/of_device.h> +#include <linux/platform_device.h> +#include <linux/devfreq.h> +#include <linux/pm_opp.h> +#include <linux/clk.h> +#include <linux/clk-provider.h> +#include <linux/arm-smccc.h> + +#define IMX_SIP_DDR_DVFS 0xc2000004 + +/* Values starting from 0 switch to specific frequency */ +#define IMX_SIP_DDR_FREQ_SET_HIGH 0x00 + +/* Deprecated after moving IRQ handling to ATF */ +#define IMX_SIP_DDR_DVFS_WAIT_CHANGE 0x0F + +/* Query available frequencies. */ +#define IMX_SIP_DDR_DVFS_GET_FREQ_COUNT 0x10 +#define IMX_SIP_DDR_DVFS_GET_FREQ_INFO 0x11 + +/* +* This should be in a 1:1 mapping with devicetree OPPs but +* firmware provides additional info. +*/ +struct imx8m_ddrc_freq { +unsigned long rate; +unsigned long smcarg; +int dram_core_parent_index; +int dram_alt_parent_index; +int dram_apb_parent_index; +}; + +/* Hardware limitation */ +#define IMX8M_DDRC_MAX_FREQ_COUNT 4 + +/* +* i.MX8M DRAM Controller clocks have the following structure (abridged): +* +* +----------+ |\ +------+ +* | dram_pll |-------|M| dram_core | | +* +----------+ |U|---------->| D | +* /--|X| | D | +* dram_alt_root | |/ | R | +* | | C | +* +---------+ | | +* |FIX DIV/4| | | +* +---------+ | | +* composite: | | | +* +----------+ | | | +* | dram_alt |----/ | | +* +----------+ | | +* | dram_apb |-------------------->| | +* +----------+ +------+ +* +* The dram_pll is used for higher rates and dram_alt is used for lower rates. +* +* Frequency switching is implemented in TF-A (via SMC call) and can change the +* configuration of the clocks, including mux parents. The dram_alt and +* dram_apb clocks are "imx composite" and their parent can change too. +* +* We need to prepare/enable the new mux parents head of switching and update +* their information afterwards. +*/ +struct imx8m_ddrc { +struct devfreq_dev_profile profile; +struct devfreq *devfreq; + +/* For frequency switching: */ +struct clk *dram_core; +struct clk *dram_pll; +struct clk *dram_alt; +struct clk *dram_apb; + +int freq_count; +struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; +}; + +static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv, +unsigned long rate) +{ +struct imx8m_ddrc_freq *freq; +int i; + +/* +* Firmware reports values in MT/s, so we round-down from Hz +* Rounding is extra generous to ensure a match. +*/ + rate =DIV_ROUND_CLOSEST (rate,250000 ); +for (i =0 ; i < priv->freq_count; ++i) { + freq = &priv->freq_table[i]; +if (freq->rate == rate || + freq->rate +1 == rate || + freq->rate -1 == rate) +return freq; + } + +return NULL ; +} + +static void imx8m_ddrc_smc_set_freq (int target_freq)
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 113
Quoted 6 lines
drivers/devfreq/imx8m-ddrc.c new file +6
-0+{ +struct arm_smccc_res res; +u32 online_cpus =0 ; +int cpu; + +local_irq_disable ();
local_irq_disableis more proper thanlocal_irq_save()?
There's no need to use save/restore because we know that irqs are enabled when entering the function. This is only called from devfreq set_target which runs in process context.
Quoted 26 lines
drivers/devfreq/imx8m-ddrc.c line 122
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +9
-0+ +for_each_online_cpu (cpu) + online_cpus |= (1 << (cpu *8 )); + +/* change the ddr freqency */ +arm_smccc_smc (IMX_SIP_DDR_DVFS , target_freq, online_cpus, +0 ,0 ,0 ,0 ,0 , &res); + +local_irq_enable ();ditto.
local_irq_restore()instead oflocal_irq_enable()Quoted 13 lines
drivers/devfreq/imx8m-ddrc.c new file +13
-0+} + +struct clk *clk_get_parent_by_index (struct clk *clk,int index) +{ +struct clk_hw *hw; + + hw =clk_hw_get_parent_by_index (__clk_get_hw (clk), index); + +return hw ? hw->clk :NULL ; +} + +static int imx8m_ddrc_set_freq (struct device *dev,struct imx8m_ddrc_freq *freq) +{
Quoted 8 lines
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +7
-0+struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct clk *new_dram_core_parent; +struct clk *new_dram_alt_parent; +struct clk *new_dram_apb_parent; +int ret; + + new_dram_core_parent = clk_get_parent_by_index(You can use '
clk_hw_get_parent_by_index' directly.
The helper above avoids duplicating ?:
Quoted 1 line
And, you need to check whether the return value is NULL or not.
Yes for dram_core_parent but others are deliberately allowed to be NULL. For some setpoints some parents might be indifferent or require no changes. For example when dram_core has pll as parent the expectation is that dram_all will be OFF so parent is irrelevant.
Driver relies on the fact that clock API ignores prepare/enable for NULL clocks. I can add a comment explaining that.
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c line 144
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_core, freq->dram_core_parent_index - 1); + new_dram_alt_parent = clk_get_parent_by_index(ditto.
drivers/devfreq/imx8m-ddrc.c line 146
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_alt, freq->dram_alt_parent_index - 1); + new_dram_apb_parent = clk_get_parent_by_index(ditto.
Quoted 15 lines
drivers/devfreq/imx8m-ddrc.c line 158
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c new file +12
-0+ priv->dram_apb, freq->dram_apb_parent_index - 1); + +/* increase reference counts and ensure clks are ON before switch */ + ret =clk_prepare_enable (new_dram_core_parent); +if (ret) { +dev_err (dev,"failed enable new dram_core parent: %d\n" , ret); + goto out; + } + ret =clk_prepare_enable (new_dram_alt_parent); +if (ret) { +dev_err (dev,"failed enable new dram_alt parent: %d\n" , ret); + goto out_dis_core;I think that 'dis' is not general expression for 'disable'. Just, I think that '
out_disable_core' is better than 'out_dis_core'.
OK
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 163
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } + ret =clk_prepare_enable (new_dram_apb_parent); +if (ret) { +dev_err (dev,"failed enable new dram_apb parent: %d\n" , ret); + goto out_dis_alt;ditto.
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c line 171
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +8
-0+ } + +imx8m_ddrc_smc_set_freq (freq->smcarg); + +/* update parents in clk tree after switch. */ + ret =clk_set_parent (priv->dram_core, new_dram_core_parent); +if (ret) +dev_err (dev,"failed set dram_core parent: %d\n" , ret);if you don't return directly, you better to use '
dev_warn' instead of 'dev_err'.
OK.
Quoted 15 lines
drivers/devfreq/imx8m-ddrc.c line 175
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+if (new_dram_alt_parent) { + ret =clk_set_parent (priv->dram_alt, new_dram_alt_parent); +if (ret) +dev_err (dev,"failed set dram_alt parent: %d\n" , ret);ditto. Use
dev_warn.drivers/devfreq/imx8m-ddrc.c line 180
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } +if (new_dram_apb_parent) { + ret =clk_set_parent (priv->dram_apb, new_dram_apb_parent); +if (ret) +dev_err (dev,"failed set dram_apb parent: %d\n" , ret);ditto. Use
dev_warn.
Quoted 9 lines
drivers/devfreq/imx8m-ddrc.c line 186
Quoted 6 lines
drivers/devfreq/imx8m-ddrc.c new file +6
-0+ } + +/* +* Explicitly refresh dram PLL rate. +* +* Even if it's marked with CLK_GET_RATE_NOCACHE the rate will not be nitpick:
What is more proper description either or 'Even if' or 'if' ?
Rate updates work for dram_alt/apb but not for dram_pll because additional clocks might be present between the PLL and dram_core mux. This happens *even if* pll is marked with CLK_GET_RATE_NOCACHE.
Quoted 44 lines
Quoted 31 lines
drivers/devfreq/imx8m-ddrc.c new file +31
-0+* automatically refreshed when clk_get_rate is called on children. +*/ +clk_get_rate (priv->dram_pll); + +/* +* clk_set_parent transfer the reference count from old parent. +* now we drop extra reference counts used during the switch +*/ +clk_disable_unprepare (new_dram_apb_parent); +out_dis_alt: +clk_disable_unprepare (new_dram_alt_parent); +out_dis_core: +clk_disable_unprepare (new_dram_core_parent); +out: +return ret; +} + +static int imx8m_ddrc_target (struct device *dev,unsigned long *freq,u32 flags) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *new_opp; +unsigned long old_freq, new_freq; +int ret; + + new_opp =devfreq_recommended_opp (dev, freq, flags); +if (IS_ERR (new_opp)) { + ret =PTR_ERR (new_opp); +dev_err (dev,"failed to get recommended opp: %d\n" , ret); +return ret; + }drivers/devfreq/imx8m-ddrc.c line 227
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+dev_pm_opp_put (new_opp); + + old_freq =clk_get_rate (priv->dram_core); +if (*freq == old_freq) +return 0 ; + + freq_info =imx8m_ddrc_find_freq (priv, *freq); +if (!freq_info) +return -EINVAL ; + ret =imx8m_ddrc_set_freq (dev, freq_info);Need to check whether 'ret' is valid or not.
Quoted 7 lines
drivers/devfreq/imx8m-ddrc.c line 231
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+ +/* Also read back the clk rate to verify switch was correct */ + new_freq =clk_get_rate (priv->dram_core); +if (ret || *freq != new_freq)You should check 'ret' right after
imx8m_ddrc_set_freq()instead of this position.
OK, I can add two error paths.
Quoted 30 lines
Quoted 17 lines
drivers/devfreq/imx8m-ddrc.c new file +17
-0+dev_err (dev,"ddrc failed freq set to %lu from %lu, now at %lu\n" , + old_freq, *freq, new_freq); + else + dev_dbg(dev, "ddrc freq set to %lu (was %lu)\n", + *freq, old_freq); + +return ret; +} + +static int imx8m_ddrc_get_cur_freq (struct device *dev,unsigned long *freq) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + *freq =clk_get_rate (priv->dram_core); + +return 0 ; +}drivers/devfreq/imx8m-ddrc.c line 258
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+ +static int imx8m_ddrc_init_freq_info (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct arm_smccc_res res; +int index; + +/* +* An error here means DDR DVFS API not supported by firmware +*/ Don't need to add multiple line comments. It is possible to be change as following:
Quoted 1 line
/* An error here means DDR DVFS API not supported by firmware */
OK.
Quoted 37 lines
Quoted 37 lines
drivers/devfreq/imx8m-ddrc.c new file +37
-0+arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_COUNT , +0 ,0 ,0 ,0 ,0 ,0 , &res); + priv->freq_count = res.a0; +if (priv->freq_count <=0 || + priv->freq_count >IMX8M_DDRC_MAX_FREQ_COUNT ) +return -ENODEV ; + +for (index =0 ; index < priv->freq_count; ++index) { +struct imx8m_ddrc_freq *freq = &priv->freq_table[index]; + +arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_INFO , + index,0 ,0 ,0 ,0 ,0 , &res); +/* Result should be strictly positive */ +if ((long )res.a0 <=0 ) +return -ENODEV ; + + freq->rate = res.a0; + freq->smcarg = index; + freq->dram_core_parent_index = res.a1; + freq->dram_alt_parent_index = res.a2; + freq->dram_apb_parent_index = res.a3; + +/* dram_core has 2 options: dram_pll or dram_alt_root */ +if (freq->dram_core_parent_index !=1 && + freq->dram_core_parent_index !=2 ) +return -ENODEV ; +/* dram_apb and dram_alt have exactly 8 possible parents */ +if (freq->dram_alt_parent_index >8 || + freq->dram_apb_parent_index >8 ) +return -ENODEV ; +/* dram_core from alt requires explicit dram_alt parent */ +if (freq->dram_core_parent_index ==2 && + freq->dram_alt_parent_index ==0 ) +return -ENODEV ; + } + +return 0 ;
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 298
Quoted 3 lines
drivers/devfreq/imx8m-ddrc.c new file +3
-0+} + +/* imx8m_ddrc_check_opps() - disable OPPs not supported by firmware */ nitpick:
On this driver, there are no some description for internal function. In order to keep the coding style, you better to move this comment into function before 'freq =ULONG_MAX'.
Is it a problem if only some internal functions have documentation? I can add docs to more functions.
BTW: how can I generate devfreq documentation from these comments?
Quoted 110 lines
Quoted 110 lines
drivers/devfreq/imx8m-ddrc.c new file +110
-0+static int imx8m_ddrc_check_opps (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *opp; +unsigned long freq; + + freq =ULONG_MAX ; +while (true) { + opp =dev_pm_opp_find_freq_floor (dev, &freq); +if (opp ==ERR_PTR (-ERANGE )) +break ; +if (IS_ERR (opp)) { +dev_err (dev,"Failed enumerating OPPs: %ld\n" , +PTR_ERR (opp)); +return PTR_ERR (opp); + } +dev_pm_opp_put (opp); + + freq_info =imx8m_ddrc_find_freq (priv, freq); +if (!freq_info) { +dev_info (dev,"Disable unsupported OPP %luHz %luMT/s\n" , + freq,DIV_ROUND_CLOSEST (freq,250000 )); +dev_pm_opp_disable (dev, freq); + } + + freq--; + } + +return 0 ; +} + +static void imx8m_ddrc_exit (struct device *dev) +{ +dev_pm_opp_of_remove_table (dev); +} + +static int imx8m_ddrc_probe (struct platform_device *pdev) +{ +struct device *dev = &pdev->dev; +struct imx8m_ddrc *priv; +const char *gov =DEVFREQ_GOV_USERSPACE ; +int ret; + + priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL ); +if (!priv) +return -ENOMEM ; + +platform_set_drvdata (pdev, priv); + + ret =imx8m_ddrc_init_freq_info (dev); +if (ret) { +dev_err (dev,"failed to init firmware freq info: %d\n" , ret); +return ret; + } + + priv->dram_core =devm_clk_get (dev,"core" ); + priv->dram_pll =devm_clk_get (dev,"pll" ); + priv->dram_alt =devm_clk_get (dev,"alt" ); + priv->dram_apb =devm_clk_get (dev,"apb" ); +if (IS_ERR (priv->dram_core) || +IS_ERR (priv->dram_pll) || +IS_ERR (priv->dram_alt) || +IS_ERR (priv->dram_apb)) { + ret =PTR_ERR (priv->devfreq); +dev_err (dev,"failed to fetch clocks: %d\n" , ret); +return ret; + } + + ret =dev_pm_opp_of_add_table (dev); +if (ret <0 ) { +dev_err (dev,"failed to get OPP table\n" ); +return ret; + } + + ret =imx8m_ddrc_check_opps (dev); +if (ret <0 ) + goto err; + + priv->profile.polling_ms =1000 ; + priv->profile.target = imx8m_ddrc_target; + priv->profile.exit = imx8m_ddrc_exit; + priv->profile.get_cur_freq = imx8m_ddrc_get_cur_freq; + priv->profile.initial_freq =clk_get_rate (priv->dram_core); + + priv->devfreq =devm_devfreq_add_device (dev, &priv->profile, + gov,NULL ); +if (IS_ERR (priv->devfreq)) { + ret =PTR_ERR (priv->devfreq); +dev_err (dev,"failed to add devfreq device: %d\n" , ret); + goto err; + } + +return 0 ; + +err: +dev_pm_opp_of_remove_table (dev); +return ret; +} + +static const struct of_device_id imx8m_ddrc_of_match[] = { + { .compatible ="fsl,imx8m-ddrc" , }, + {/* sentinel */ }, +}; +MODULE_DEVICE_TABLE (of, imx8m_ddrc_of_match); + +static struct platform_driver imx8m_ddrc_platdrv = { + .probe = imx8m_ddrc_probe, + .driver = { + .name ="imx8m-ddrc-devfreq" ,
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 416
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +8
-0+ .of_match_table = of_match_ptr(imx8m_ddrc_of_match), + }, +}; +module_platform_driver (imx8m_ddrc_platdrv); + +MODULE_DESCRIPTION ("i.MX8M DDR Controller frequency driver" ); +MODULE_AUTHOR ("Leonard Crestez <[email protected]>" ); +MODULE_LICENSE ("GPL v2" );
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Chanwoo Choi
maintainerQuoted 216 lines · Leonard Crestez
Quoted 54 lines · Chanwoo Choi
Hi Leonard,
Quoted 51 lines · Leonard Crestez
Add driver for dynamic scaling the DDR Controller on imx8m chips. Actual frequency switching is implemented inside TF-A, this driver wraps the SMC calls and synchronizes the clk tree.
The DRAM clocks on imx8m have the following structure (abridged):
+----------+ |\ +------+ | dram_pll |-------|M| dram_core | | +----------+ |U|---------->| D | /--|X| | D | dram_alt_root | |/ | R | | | C | +---------+ | | |FIX DIV/4| | | +---------+ | | composite: | | | +----------+ | | | | dram_alt |----/ | | +----------+ | | | dram_apb |-------------------->| | +----------+ +------+The
dram_pllis used for higher rates anddram_altis used for lower rates. Thedram_altanddram_apbclocks are "imx composite" and their parent can also be modified.This driver will prepare/enable the new parents ahead of switching (so that the expected roots are enabled) and afterwards it will call
clk_set_parentto ensure the parents in clock framework are up-to-date.The driver relies on
dram_plldram_altanddram_apbbeing marked withCLK_GET_RATE_NOCACHEfor rate updates.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/Kconfig | 10 + drivers/devfreq/Makefile | 1 + drivers/devfreq/imx8m-ddrc.c | 416 +++++++++++++++++++++++++++++++++++ 3 files changed, 427 insertions(+)create mode 100644
drivers/devfreq/imx8m-ddrc.cdrivers/devfreq/Kconfig +0
-0@@ -90,10 +90,20 @@ config ARM_EXYNOS_BUS_DEVFREQ Each memory bus group could contain many memoby bus block. It reads PPMU counters of memory controllers by using DEVFREQ-event device and adjusts the operating frequencies and voltages with OPP support. This does not yet operate with optimal voltages.Quoted 9 lines
drivers/devfreq/Kconfig config ARM_EXYNOS_BUS_DEVFREQ · line 100
Quoted 6 lines
drivers/devfreq/Kconfig +6
-0+config ARM_IMX8M_DDRC_DEVFREQ + tristate "i.MX8M DDRC DEVFREQ Driver" + depends on ARCH_MXC || COMPILE_TEST + select DEVFREQ_GOV_SIMPLE_ONDEMAND + select DEVFREQ_GOV_USERSPACE + select PM_OPPIt doesn't need to add.
CONFIG_DEVFREQadd 'selectPM_OPP' because 'PM_OPP' is mandatory for devfreq.OK.
All other drivers select
PM_OPP, this was just copiedQuoted 138 lines
Quoted 138 lines
drivers/devfreq/Kconfig +4
-0+ help + This adds the DEVFREQ driver for the i.MX8M DDR Controller. It allows + adjusting DRAM frequency. + config ARM_TEGRA_DEVFREQ tristate "NVIDIA Tegra30/114/124/210 DEVFREQ Driver" depends on ARCH_TEGRA_3x_SOC || ARCH_TEGRA_114_SOC || \ ARCH_TEGRA_132_SOC || ARCH_TEGRA_124_SOC || \ ARCH_TEGRA_210_SOC || \drivers/devfreq/Makefile +1
-0@@ -7,10 +7,11 @@ obj-$(CONFIG_DEVFREQ_GOV_POWERSAVE) += governor_powersave.o obj-$(CONFIG_DEVFREQ_GOV_USERSPACE) += governor_userspace.o obj-$(CONFIG_DEVFREQ_GOV_PASSIVE) += governor_passive.o # DEVFREQ Drivers obj-$(CONFIG_ARM_EXYNOS_BUS_DEVFREQ) += exynos-bus.o +obj-$(CONFIG_ARM_IMX8M_DDRC_DEVFREQ) += imx8m-ddrc.o obj-$(CONFIG_ARM_RK3399_DMC_DEVFREQ) += rk3399_dmc.o obj-$(CONFIG_ARM_TEGRA_DEVFREQ) += tegra30-devfreq.o obj-$(CONFIG_ARM_TEGRA20_DEVFREQ) += tegra20-devfreq.o # DEVFREQ Event Driversdrivers/devfreq/imx8m-ddrc.c new file +107
-0@@ -0,0 +1,416 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright 2019 NXP +*/ + +#include <linux/module.h> +#include <linux/device.h> +#include <linux/of_device.h> +#include <linux/platform_device.h> +#include <linux/devfreq.h> +#include <linux/pm_opp.h> +#include <linux/clk.h> +#include <linux/clk-provider.h> +#include <linux/arm-smccc.h> + +#define IMX_SIP_DDR_DVFS 0xc2000004 + +/* Values starting from 0 switch to specific frequency */ +#define IMX_SIP_DDR_FREQ_SET_HIGH 0x00 + +/* Deprecated after moving IRQ handling to ATF */ +#define IMX_SIP_DDR_DVFS_WAIT_CHANGE 0x0F + +/* Query available frequencies. */ +#define IMX_SIP_DDR_DVFS_GET_FREQ_COUNT 0x10 +#define IMX_SIP_DDR_DVFS_GET_FREQ_INFO 0x11 + +/* +* This should be in a 1:1 mapping with devicetree OPPs but +* firmware provides additional info. +*/ +struct imx8m_ddrc_freq { +unsigned long rate; +unsigned long smcarg; +int dram_core_parent_index; +int dram_alt_parent_index; +int dram_apb_parent_index; +}; + +/* Hardware limitation */ +#define IMX8M_DDRC_MAX_FREQ_COUNT 4 + +/* +* i.MX8M DRAM Controller clocks have the following structure (abridged): +* +* +----------+ |\ +------+ +* | dram_pll |-------|M| dram_core | | +* +----------+ |U|---------->| D | +* /--|X| | D | +* dram_alt_root | |/ | R | +* | | C | +* +---------+ | | +* |FIX DIV/4| | | +* +---------+ | | +* composite: | | | +* +----------+ | | | +* | dram_alt |----/ | | +* +----------+ | | +* | dram_apb |-------------------->| | +* +----------+ +------+ +* +* The dram_pll is used for higher rates and dram_alt is used for lower rates. +* +* Frequency switching is implemented in TF-A (via SMC call) and can change the +* configuration of the clocks, including mux parents. The dram_alt and +* dram_apb clocks are "imx composite" and their parent can change too. +* +* We need to prepare/enable the new mux parents head of switching and update +* their information afterwards. +*/ +struct imx8m_ddrc { +struct devfreq_dev_profile profile; +struct devfreq *devfreq; + +/* For frequency switching: */ +struct clk *dram_core; +struct clk *dram_pll; +struct clk *dram_alt; +struct clk *dram_apb; + +int freq_count; +struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; +}; + +static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv, +unsigned long rate) +{ +struct imx8m_ddrc_freq *freq; +int i; + +/* +* Firmware reports values in MT/s, so we round-down from Hz +* Rounding is extra generous to ensure a match. +*/ + rate =DIV_ROUND_CLOSEST (rate,250000 ); +for (i =0 ; i < priv->freq_count; ++i) { + freq = &priv->freq_table[i]; +if (freq->rate == rate || + freq->rate +1 == rate || + freq->rate -1 == rate) +return freq; + } + +return NULL ; +} + +static void imx8m_ddrc_smc_set_freq (int target_freq)Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 113
Quoted 6 lines
drivers/devfreq/imx8m-ddrc.c new file +6
-0+{ +struct arm_smccc_res res; +u32 online_cpus =0 ; +int cpu; + +local_irq_disable ();
local_irq_disableis more proper thanlocal_irq_save()?
Quoted 3 lines
There's no need to use save/restore because we know that irqs are enabled when entering the function. This is only called from devfreq
set_targetwhich runs in process context.
local_irq might affect the whole architecture and other running device driver which requires the some CPU irq at the same time. Don't need to save the irqflag?
Quoted 36 lines
Quoted 27 lines
drivers/devfreq/imx8m-ddrc.c line 122
Quoted 11 lines
drivers/devfreq/imx8m-ddrc.c new file +9
-0+ +for_each_online_cpu (cpu) + online_cpus |= (1 << (cpu *8 )); + +/* change the ddr freqency */ +arm_smccc_smc (IMX_SIP_DDR_DVFS , target_freq, online_cpus, +0 ,0 ,0 ,0 ,0 , &res); + +local_irq_enable ();ditto.
local_irq_restore()instead oflocal_irq_enable()Quoted 13 lines
drivers/devfreq/imx8m-ddrc.c new file +13
-0+} + +struct clk *clk_get_parent_by_index (struct clk *clk,int index) +{ +struct clk_hw *hw; + + hw =clk_hw_get_parent_by_index (__clk_get_hw (clk), index); + +return hw ? hw->clk :NULL ; +} + +static int imx8m_ddrc_set_freq (struct device *dev,struct imx8m_ddrc_freq *freq) +{Quoted 8 lines
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +7
-0+struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct clk *new_dram_core_parent; +struct clk *new_dram_alt_parent; +struct clk *new_dram_apb_parent; +int ret; + + new_dram_core_parent = clk_get_parent_by_index(You can use '
clk_hw_get_parent_by_index' directly.
Quoted 1 line
The helper above avoids duplicating ?:
I think that it is not duplicate. You use 'clk_hw_get_parent_by_index' function and check the return value. it is essential behavior.
Quoted 6 lines
Quoted 2 lines
And, you need to check whether the return value is NULL or not.
Yes for
dram_core_parentbut others are deliberately allowed to be NULL. For some setpoints some parents might be indifferent or require no changes. For example whendram_corehas pll as parent the expectation is thatdram_allwill be OFF so parent is irrelevant.
If all case is not same as you commented, you just add the proper dev_warn or dev_info message according to the h/w characteristic. Because user cannot know the detailed reason why don't check the return value.
Quoted 78 lines
Driver relies on the fact that clock API ignores prepare/enable for NULL clocks. I can add a comment explaining that.
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c line 144
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_core, freq->dram_core_parent_index - 1); + new_dram_alt_parent = clk_get_parent_by_index(ditto.
drivers/devfreq/imx8m-ddrc.c line 146
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_alt, freq->dram_alt_parent_index - 1); + new_dram_apb_parent = clk_get_parent_by_index(ditto.
Quoted 15 lines
drivers/devfreq/imx8m-ddrc.c line 158
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c new file +12
-0+ priv->dram_apb, freq->dram_apb_parent_index - 1); + +/* increase reference counts and ensure clks are ON before switch */ + ret =clk_prepare_enable (new_dram_core_parent); +if (ret) { +dev_err (dev,"failed enable new dram_core parent: %d\n" , ret); + goto out; + } + ret =clk_prepare_enable (new_dram_alt_parent); +if (ret) { +dev_err (dev,"failed enable new dram_alt parent: %d\n" , ret); + goto out_dis_core;I think that 'dis' is not general expression for 'disable'. Just, I think that '
out_disable_core' is better than 'out_dis_core'.OK
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 163
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } + ret =clk_prepare_enable (new_dram_apb_parent); +if (ret) { +dev_err (dev,"failed enable new dram_apb parent: %d\n" , ret); + goto out_dis_alt;ditto.
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c line 171
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +8
-0+ } + +imx8m_ddrc_smc_set_freq (freq->smcarg); + +/* update parents in clk tree after switch. */ + ret =clk_set_parent (priv->dram_core, new_dram_core_parent); +if (ret) +dev_err (dev,"failed set dram_core parent: %d\n" , ret);if you don't return directly, you better to use '
dev_warn' instead of 'dev_err'.OK.
Quoted 15 lines
drivers/devfreq/imx8m-ddrc.c line 175
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+if (new_dram_alt_parent) { + ret =clk_set_parent (priv->dram_alt, new_dram_alt_parent); +if (ret) +dev_err (dev,"failed set dram_alt parent: %d\n" , ret);ditto. Use
dev_warn.drivers/devfreq/imx8m-ddrc.c line 180
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } +if (new_dram_apb_parent) { + ret =clk_set_parent (priv->dram_apb, new_dram_apb_parent); +if (ret) +dev_err (dev,"failed set dram_apb parent: %d\n" , ret);ditto. Use
dev_warn.Quoted 9 lines
drivers/devfreq/imx8m-ddrc.c line 186
Quoted 6 lines
drivers/devfreq/imx8m-ddrc.c new file +6
-0+ } + +/* +* Explicitly refresh dram PLL rate. +* +* Even if it's marked with CLK_GET_RATE_NOCACHE the rate will not be nitpick:
What is more proper description either or 'Even if' or 'if' ?
Quoted 3 lines
Rate updates work for dram_alt/apb but not for
dram_pllbecause additional clocks might be present between the PLL anddram_coremux. This happens *even if* pll is marked withCLK_GET_RATE_NOCACHE.
ok.
Quoted 134 lines
Quoted 45 lines
Quoted 32 lines
drivers/devfreq/imx8m-ddrc.c new file +31
-0+* automatically refreshed when clk_get_rate is called on children. +*/ +clk_get_rate (priv->dram_pll); + +/* +* clk_set_parent transfer the reference count from old parent. +* now we drop extra reference counts used during the switch +*/ +clk_disable_unprepare (new_dram_apb_parent); +out_dis_alt: +clk_disable_unprepare (new_dram_alt_parent); +out_dis_core: +clk_disable_unprepare (new_dram_core_parent); +out: +return ret; +} + +static int imx8m_ddrc_target (struct device *dev,unsigned long *freq,u32 flags) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *new_opp; +unsigned long old_freq, new_freq; +int ret; + + new_opp =devfreq_recommended_opp (dev, freq, flags); +if (IS_ERR (new_opp)) { + ret =PTR_ERR (new_opp); +dev_err (dev,"failed to get recommended opp: %d\n" , ret); +return ret; + }drivers/devfreq/imx8m-ddrc.c line 227
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+dev_pm_opp_put (new_opp); + + old_freq =clk_get_rate (priv->dram_core); +if (*freq == old_freq) +return 0 ; + + freq_info =imx8m_ddrc_find_freq (priv, *freq); +if (!freq_info) +return -EINVAL ; + ret =imx8m_ddrc_set_freq (dev, freq_info);Need to check whether 'ret' is valid or not.
Quoted 7 lines
drivers/devfreq/imx8m-ddrc.c line 231
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+ +/* Also read back the clk rate to verify switch was correct */ + new_freq =clk_get_rate (priv->dram_core); +if (ret || *freq != new_freq)You should check 'ret' right after
imx8m_ddrc_set_freq()instead of this position.OK, I can add two error paths.
Quoted 30 lines
Quoted 17 lines
drivers/devfreq/imx8m-ddrc.c new file +17
-0+dev_err (dev,"ddrc failed freq set to %lu from %lu, now at %lu\n" , + old_freq, *freq, new_freq); + else + dev_dbg(dev, "ddrc freq set to %lu (was %lu)\n", + *freq, old_freq); + +return ret; +} + +static int imx8m_ddrc_get_cur_freq (struct device *dev,unsigned long *freq) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + *freq =clk_get_rate (priv->dram_core); + +return 0 ; +}drivers/devfreq/imx8m-ddrc.c line 258
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+ +static int imx8m_ddrc_init_freq_info (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct arm_smccc_res res; +int index; + +/* +* An error here means DDR DVFS API not supported by firmware +*/ Don't need to add multiple line comments. It is possible to be change as following:
Quoted 1 line
/* An error here means DDR DVFS API not supported by firmware */OK.
Quoted 37 lines
Quoted 37 lines
drivers/devfreq/imx8m-ddrc.c new file +37
-0+arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_COUNT , +0 ,0 ,0 ,0 ,0 ,0 , &res); + priv->freq_count = res.a0; +if (priv->freq_count <=0 || + priv->freq_count >IMX8M_DDRC_MAX_FREQ_COUNT ) +return -ENODEV ; + +for (index =0 ; index < priv->freq_count; ++index) { +struct imx8m_ddrc_freq *freq = &priv->freq_table[index]; + +arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_INFO , + index,0 ,0 ,0 ,0 ,0 , &res); +/* Result should be strictly positive */ +if ((long )res.a0 <=0 ) +return -ENODEV ; + + freq->rate = res.a0; + freq->smcarg = index; + freq->dram_core_parent_index = res.a1; + freq->dram_alt_parent_index = res.a2; + freq->dram_apb_parent_index = res.a3; + +/* dram_core has 2 options: dram_pll or dram_alt_root */ +if (freq->dram_core_parent_index !=1 && + freq->dram_core_parent_index !=2 ) +return -ENODEV ; +/* dram_apb and dram_alt have exactly 8 possible parents */ +if (freq->dram_alt_parent_index >8 || + freq->dram_apb_parent_index >8 ) +return -ENODEV ; +/* dram_core from alt requires explicit dram_alt parent */ +if (freq->dram_core_parent_index ==2 && + freq->dram_alt_parent_index ==0 ) +return -ENODEV ; + } + +return 0 ;Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 298
Quoted 3 lines
drivers/devfreq/imx8m-ddrc.c new file +3
-0+} + +/* imx8m_ddrc_check_opps() - disable OPPs not supported by firmware */ nitpick:
On this driver, there are no some description for internal function. In order to keep the coding style, you better to move this comment into function before 'freq =ULONG_MAX'.
Quoted 2 lines
Is it a problem if only some internal functions have documentation? I can add docs to more functions.
It is not problem. It is just preferable way to add the comment. Usually, if the function is public and exported, I added the detailed function comment as following example of devfreq_add_device().
/**
656 * devfreq_add_device() - Add devfreq feature to the device 657 * @dev: the device to add devfreq feature. 658 * @profile: device-specific profile to run devfreq. 659 * @governor_name: name of the policy to choose frequency. 660 * @data: private data for the governor. The devfreq framework does not 661 * touch this value. 662 */ 663 struct devfreq *devfreq_add_device(struct device *dev, 664 struct devfreq_dev_profile *profile, 665 const char *governor_name, 666 void *data)
Otherwise, if function is used in the only local with static keyword, usually I don't add the above detailed description. Instead, add the detailed comment into the function.
It is not the standard way. Just preferable way for the function description if possible.
Quoted 124 lines
BTW: how can I generate devfreq documentation from these comments?
Quoted 118 lines
Quoted 118 lines
drivers/devfreq/imx8m-ddrc.c new file +118
-0+static int imx8m_ddrc_check_opps (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *opp; +unsigned long freq; + + freq =ULONG_MAX ; +while (true) { + opp =dev_pm_opp_find_freq_floor (dev, &freq); +if (opp ==ERR_PTR (-ERANGE )) +break ; +if (IS_ERR (opp)) { +dev_err (dev,"Failed enumerating OPPs: %ld\n" , +PTR_ERR (opp)); +return PTR_ERR (opp); + } +dev_pm_opp_put (opp); + + freq_info =imx8m_ddrc_find_freq (priv, freq); +if (!freq_info) { +dev_info (dev,"Disable unsupported OPP %luHz %luMT/s\n" , + freq,DIV_ROUND_CLOSEST (freq,250000 )); +dev_pm_opp_disable (dev, freq); + } + + freq--; + } + +return 0 ; +} + +static void imx8m_ddrc_exit (struct device *dev) +{ +dev_pm_opp_of_remove_table (dev); +} + +static int imx8m_ddrc_probe (struct platform_device *pdev) +{ +struct device *dev = &pdev->dev; +struct imx8m_ddrc *priv; +const char *gov =DEVFREQ_GOV_USERSPACE ; +int ret; + + priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL ); +if (!priv) +return -ENOMEM ; + +platform_set_drvdata (pdev, priv); + + ret =imx8m_ddrc_init_freq_info (dev); +if (ret) { +dev_err (dev,"failed to init firmware freq info: %d\n" , ret); +return ret; + } + + priv->dram_core =devm_clk_get (dev,"core" ); + priv->dram_pll =devm_clk_get (dev,"pll" ); + priv->dram_alt =devm_clk_get (dev,"alt" ); + priv->dram_apb =devm_clk_get (dev,"apb" ); +if (IS_ERR (priv->dram_core) || +IS_ERR (priv->dram_pll) || +IS_ERR (priv->dram_alt) || +IS_ERR (priv->dram_apb)) { + ret =PTR_ERR (priv->devfreq); +dev_err (dev,"failed to fetch clocks: %d\n" , ret); +return ret; + } + + ret =dev_pm_opp_of_add_table (dev); +if (ret <0 ) { +dev_err (dev,"failed to get OPP table\n" ); +return ret; + } + + ret =imx8m_ddrc_check_opps (dev); +if (ret <0 ) + goto err; + + priv->profile.polling_ms =1000 ; + priv->profile.target = imx8m_ddrc_target; + priv->profile.exit = imx8m_ddrc_exit; + priv->profile.get_cur_freq = imx8m_ddrc_get_cur_freq; + priv->profile.initial_freq =clk_get_rate (priv->dram_core); + + priv->devfreq =devm_devfreq_add_device (dev, &priv->profile, + gov,NULL ); +if (IS_ERR (priv->devfreq)) { + ret =PTR_ERR (priv->devfreq); +dev_err (dev,"failed to add devfreq device: %d\n" , ret); + goto err; + } + +return 0 ; + +err: +dev_pm_opp_of_remove_table (dev); +return ret; +} + +static const struct of_device_id imx8m_ddrc_of_match[] = { + { .compatible ="fsl,imx8m-ddrc" , }, + {/* sentinel */ }, +}; +MODULE_DEVICE_TABLE (of, imx8m_ddrc_of_match); + +static struct platform_driver imx8m_ddrc_platdrv = { + .probe = imx8m_ddrc_probe, + .driver = { + .name ="imx8m-ddrc-devfreq" , + .of_match_table =of_match_ptr (imx8m_ddrc_of_match), + }, +}; +module_platform_driver (imx8m_ddrc_platdrv); + +MODULE_DESCRIPTION ("i.MX8M DDR Controller frequency driver" ); +MODULE_AUTHOR ("Leonard Crestez <[email protected]>" ); +MODULE_LICENSE ("GPL v2" );
Signature · 3 lines
Best Regards, Chanwoo Choi Samsung Electronics
Re: [PATCH v4 1/6] clk: imx8m: Set CLK_GET_RATE_NOCACHE on dram clocks
In reply to Leonard CrestezQuoted 54 lines · Leonard Crestez
These clocks are only modified as part of DRAM frequency switches during which DRAM itself is briefly inaccessible. The switch is performed with a SMC call to by TF-A which runs from a SRAM area; upon returning to linux several clocks bits are modified and we need to update them.
For rate bits an easy solution is to just mark with
CLK_GET_RATE_NOCACHEso that new rates are always read back from registers.Signed-off-by Leonard Crestez <[email protected]>--- drivers/clk/imx/clk-imx8mm.c | 11 +++++++++-- drivers/clk/imx/clk-imx8mn.c | 12 ++++++++++-- drivers/clk/imx/clk-imx8mq.c | 15 +++++++++++---- 3 files changed, 30 insertions(+), 8 deletions(-)drivers/clk/imx/clk-imx8mm.c +9
-2@@ -440,13 +440,20 @@ static int imx8mm_clocks_probe(struct platform_device *pdev)/* IPG */ clks[IMX8MM_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MM_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware +*/ + clks[IMX8MM_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mm_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MM_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mm_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); +/* IP */ - clks[IMX8MM_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mm_dram_alt_sels, base +0xa000 ); - clks[IMX8MM_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mm_dram_apb_sels, base +0xa080 ); clks[IMX8MM_CLK_VPU_G1 ] =imx8m_clk_composite ("vpu_g1" , imx8mm_vpu_g1_sels, base +0xa100 ); clks[IMX8MM_CLK_VPU_G2 ] =imx8m_clk_composite ("vpu_g2" , imx8mm_vpu_g2_sels, base +0xa180 ); clks[IMX8MM_CLK_DISP_DTRC ] =imx8m_clk_composite ("disp_dtrc" , imx8mm_disp_dtrc_sels, base +0xa200 ); clks[IMX8MM_CLK_DISP_DC8000 ] =imx8m_clk_composite ("disp_dc8000" , imx8mm_disp_dc8000_sels, base +0xa280 ); clks[IMX8MM_CLK_PCIE1_CTRL ] =imx8m_clk_composite ("pcie1_ctrl" , imx8mm_pcie1_ctrl_sels, base +0xa300 );drivers/clk/imx/clk-imx8mn.c +0
-1@@ -428,12 +428,20 @@ static int imx8mn_clocks_probe(struct platform_device *pdev) clks[IMX8MN_CLK_AHB ] =imx8m_clk_composite_critical ("ahb" , imx8mn_ahb_sels, base +0x9000 ); clks[IMX8MN_CLK_AUDIO_AHB ] =imx8m_clk_composite ("audio_ahb" , imx8mn_audio_ahb_sels, base +0x9100 ); clks[IMX8MN_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MN_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); clks[IMX8MN_CLK_DRAM_CORE ] =imx_clk_mux2_flags ("dram_core_clk" , base +0x9800 ,24 ,1 , imx8mn_dram_core_sels,ARRAY_SIZE (imx8mn_dram_core_sels),CLK_IS_CRITICAL ); - clks[IMX8MN_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 );
drivers/clk/imx/clk-imx8mn.c static int imx8mn_clocks_probe(struct platform_device *pdev) · line 441
Quoted 10 lines
drivers/clk/imx/clk-imx8mn.c +9
-1- clks[IMX8MN_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 ); + +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware +*/ + clks[IMX8MN_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MN_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE );
nitpick: I think it looks better if we stick to one line each clock. I know it's against the 80 chars rule, but at least is consistent.
Quoted 52 lines
drivers/clk/imx/clk-imx8mn.c +1
-0+ clks[IMX8MN_CLK_DISP_PIXEL ] =imx8m_clk_composite ("disp_pixel" , imx8mn_disp_pixel_sels, base +0xa500 ); clks[IMX8MN_CLK_SAI2 ] =imx8m_clk_composite ("sai2" , imx8mn_sai2_sels, base +0xa600 ); clks[IMX8MN_CLK_SAI3 ] =imx8m_clk_composite ("sai3" , imx8mn_sai3_sels, base +0xa680 ); clks[IMX8MN_CLK_SAI5 ] =imx8m_clk_composite ("sai5" , imx8mn_sai5_sels, base +0xa780 ); clks[IMX8MN_CLK_SAI6 ] =imx8m_clk_composite ("sai6" , imx8mn_sai6_sels, base +0xa800 );drivers/clk/imx/clk-imx8mq.c +11
-4@@ -341,11 +341,12 @@ static int imx8mq_clocks_probe(struct platform_device *pdev) clks[IMX8MQ_VIDEO_PLL1_OUT ] =imx_clk_gate ("video_pll1_out" ,"video_pll1_bypass" , base +0x10 ,21 ); clks[IMX8MQ_SYS1_PLL_OUT ] =imx_clk_fixed ("sys1_pll_out" ,800000000 ); clks[IMX8MQ_SYS2_PLL_OUT ] =imx_clk_fixed ("sys2_pll_out" ,1000000000 ); clks[IMX8MQ_SYS3_PLL_OUT ] =imx_clk_sccg_pll ("sys3_pll_out" , sys3_pll_out_sels,ARRAY_SIZE (sys3_pll_out_sels),0 ,0 ,0 , base +0x48 ,CLK_IS_CRITICAL ); - clks[IMX8MQ_DRAM_PLL_OUT ] =imx_clk_sccg_pll ("dram_pll_out" , dram_pll_out_sels,ARRAY_SIZE (dram_pll_out_sels),0 ,0 ,0 , base +0x60 ,CLK_IS_CRITICAL ); + clks[IMX8MQ_DRAM_PLL_OUT ] =imx_clk_sccg_pll ("dram_pll_out" , dram_pll_out_sels,ARRAY_SIZE (dram_pll_out_sels),0 ,0 ,0 , base +0x60 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); clks[IMX8MQ_VIDEO2_PLL_OUT ] =imx_clk_sccg_pll ("video2_pll_out" , video2_pll_out_sels,ARRAY_SIZE (video2_pll_out_sels),0 ,0 ,0 , base +0x54 ,0 );/* SYS PLL1 fixed output */ clks[IMX8MQ_SYS1_PLL_40M_CG ] =imx_clk_gate ("sys1_pll_40m_cg" ,"sys1_pll_out" , base +0x30 ,9 ); clks[IMX8MQ_SYS1_PLL_80M_CG ] =imx_clk_gate ("sys1_pll_80m_cg" ,"sys1_pll_out" , base +0x30 ,11 );@@ -433,15 +434,21 @@ static int imx8mq_clocks_probe(struct platform_device *pdev)/* IPG */ clks[IMX8MQ_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MQ_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); -/* IP */ +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware +*/ clks[IMX8MQ_CLK_DRAM_CORE ] =imx_clk_mux2_flags ("dram_core_clk" , base +0x9800 ,24 ,1 , imx8mq_dram_core_sels,ARRAY_SIZE (imx8mq_dram_core_sels),CLK_IS_CRITICAL ); + clks[IMX8MQ_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mq_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MQ_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mq_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE ); - clks[IMX8MQ_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mq_dram_alt_sels, base +0xa000 ); - clks[IMX8MQ_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mq_dram_apb_sels, base +0xa080 ); +/* IP */ clks[IMX8MQ_CLK_VPU_G1 ] =imx8m_clk_composite ("vpu_g1" , imx8mq_vpu_g1_sels, base +0xa100 ); clks[IMX8MQ_CLK_VPU_G2 ] =imx8m_clk_composite ("vpu_g2" , imx8mq_vpu_g2_sels, base +0xa180 ); clks[IMX8MQ_CLK_DISP_DTRC ] =imx8m_clk_composite ("disp_dtrc" , imx8mq_disp_dtrc_sels, base +0xa200 ); clks[IMX8MQ_CLK_DISP_DC8000 ] =imx8m_clk_composite ("disp_dc8000" , imx8mq_disp_dc8000_sels, base +0xa280 ); clks[IMX8MQ_CLK_PCIE1_CTRL ] =imx8m_clk_composite ("pcie1_ctrl" , imx8mq_pcie1_ctrl_sels, base +0xa300 );Signature · 1 line
2.17.1
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v4 2/6] clk: imx: Mark dram pll on 8mm and 8mn with CLK_GET_RATE_NOCACHE
In reply to Leonard CrestezQuoted 11 lines · Leonard Crestez
DRAM frequency switches are executed in firmware and can change the configuration of the DRAM PLL outside linux. Mark these CLKs with
CLK_GET_RATE_NOCACHEso we always read back the PLL config registers and recalculate rates.In current DRAM frequency tables on 8mm/8mn only the maximum frequency uses the PLL so it's always configured in the same way. However reading back the PLL configuration is the correct behavior and allows additional setpoints in the future.
Signed-off-by Leonard Crestez <[email protected]>
This one looks fine.
Quoted 82 lines
--- drivers/clk/imx/clk-imx8mm.c | 2 +- drivers/clk/imx/clk-imx8mn.c | 2 +- drivers/clk/imx/clk-pll14xx.c | 7 +++++++ drivers/clk/imx/clk.h | 1 + 4 files changed, 10 insertions(+), 2 deletions(-)drivers/clk/imx/clk-imx8mm.c +1
-1@@ -326,11 +326,11 @@ static int imx8mm_clocks_probe(struct platform_device *pdev) clks[IMX8MM_SYS_PLL3_REF_SEL ] =imx_clk_mux ("sys_pll3_ref_sel" , base +0x114 ,0 ,2 , pll_ref_sels,ARRAY_SIZE (pll_ref_sels)); clks[IMX8MM_AUDIO_PLL1 ] =imx_clk_pll14xx ("audio_pll1" ,"audio_pll1_ref_sel" , base, &imx_1443x_pll); clks[IMX8MM_AUDIO_PLL2 ] =imx_clk_pll14xx ("audio_pll2" ,"audio_pll2_ref_sel" , base +0x14 , &imx_1443x_pll); clks[IMX8MM_VIDEO_PLL1 ] =imx_clk_pll14xx ("video_pll1" ,"video_pll1_ref_sel" , base +0x28 , &imx_1443x_pll); - clks[IMX8MM_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_pll); + clks[IMX8MM_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_dram_pll); clks[IMX8MM_GPU_PLL ] =imx_clk_pll14xx ("gpu_pll" ,"gpu_pll_ref_sel" , base +0x64 , &imx_1416x_pll); clks[IMX8MM_VPU_PLL ] =imx_clk_pll14xx ("vpu_pll" ,"vpu_pll_ref_sel" , base +0x74 , &imx_1416x_pll); clks[IMX8MM_ARM_PLL ] =imx_clk_pll14xx ("arm_pll" ,"arm_pll_ref_sel" , base +0x84 , &imx_1416x_pll); clks[IMX8MM_SYS_PLL1 ] =imx_clk_fixed ("sys_pll1" ,800000000 ); clks[IMX8MM_SYS_PLL2 ] =imx_clk_fixed ("sys_pll2" ,1000000000 );drivers/clk/imx/clk-imx8mn.c +1
-1@@ -323,11 +323,11 @@ static int imx8mn_clocks_probe(struct platform_device *pdev) clks[IMX8MN_SYS_PLL3_REF_SEL ] =imx_clk_mux ("sys_pll3_ref_sel" , base +0x114 ,0 ,2 , pll_ref_sels,ARRAY_SIZE (pll_ref_sels)); clks[IMX8MN_AUDIO_PLL1 ] =imx_clk_pll14xx ("audio_pll1" ,"audio_pll1_ref_sel" , base, &imx_1443x_pll); clks[IMX8MN_AUDIO_PLL2 ] =imx_clk_pll14xx ("audio_pll2" ,"audio_pll2_ref_sel" , base +0x14 , &imx_1443x_pll); clks[IMX8MN_VIDEO_PLL1 ] =imx_clk_pll14xx ("video_pll1" ,"video_pll1_ref_sel" , base +0x28 , &imx_1443x_pll); - clks[IMX8MN_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_pll); + clks[IMX8MN_DRAM_PLL ] =imx_clk_pll14xx ("dram_pll" ,"dram_pll_ref_sel" , base +0x50 , &imx_1443x_dram_pll); clks[IMX8MN_GPU_PLL ] =imx_clk_pll14xx ("gpu_pll" ,"gpu_pll_ref_sel" , base +0x64 , &imx_1416x_pll); clks[IMX8MN_VPU_PLL ] =imx_clk_pll14xx ("vpu_pll" ,"vpu_pll_ref_sel" , base +0x74 , &imx_1416x_pll); clks[IMX8MN_ARM_PLL ] =imx_clk_pll14xx ("arm_pll" ,"arm_pll_ref_sel" , base +0x84 , &imx_1416x_pll); clks[IMX8MN_SYS_PLL1 ] =imx_clk_fixed ("sys_pll1" ,800000000 ); clks[IMX8MN_SYS_PLL2 ] =imx_clk_fixed ("sys_pll2" ,1000000000 );drivers/clk/imx/clk-pll14xx.c +7
-0@@ -65,10 +65,17 @@ struct imx_pll14xx_clk imx_1443x_pll = { .type = PLL_1443X, .rate_table = imx_pll1443x_tbl, .rate_count = ARRAY_SIZE(imx_pll1443x_tbl), }; +struct imx_pll14xx_clk imx_1443x_dram_pll = { + .type =PLL_1443X , + .rate_table = imx_pll1443x_tbl, + .rate_count =ARRAY_SIZE (imx_pll1443x_tbl), + .flags =CLK_GET_RATE_NOCACHE , +}; +struct imx_pll14xx_clk imx_1416x_pll = { .type =PLL_1416X , .rate_table = imx_pll1416x_tbl, .rate_count =ARRAY_SIZE (imx_pll1416x_tbl), };drivers/clk/imx/clk.h +1
-0@@ -50,10 +50,11 @@ struct imx_pll14xx_clk {int flags; };extern struct imx_pll14xx_clk imx_1416x_pll;extern struct imx_pll14xx_clk imx_1443x_pll; +extern struct imx_pll14xx_clk imx_1443x_dram_pll;#define imx_clk_cpu (name, parent_name, div, mux, pll, step) \ imx_clk_hw_cpu(name, parent_name, div, mux, pll, step)->clk #define clk_register_gate2(dev, name, parent_name, flags, reg, bit_idx, \Signature · 1 line
2.17.1
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v4 5/6] PM / devfreq: imx8m-ddrc: Measure bandwidth with perf
In reply to Chanwoo ChoiQuoted 113 lines · Chanwoo Choi
Hi Leonard,
Quoted 110 lines · Leonard Crestez
The imx8m ddrc has a performance monitoring block attached which can be used to measure bandwidth usage and automatically adjust frequency.
There is already a perf driver for that block so instead of implementing a devfreq events driver use the in-kernel perf API to implement
get_dev_statusdirectly.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/imx8m-ddrc.c | 153 +++++++++++++++++++++++++++++++++++ 1 file changed, 153 insertions(+)drivers/devfreq/imx8m-ddrc.c +66
-0@@ -11,10 +11,13 @@#include <linux/pm_opp.h> #include <linux/clk.h> #include <linux/clk-provider.h> #include <linux/arm-smccc.h> +#include <asm/perf_event.h> +#include <linux/perf_event.h> +#define IMX_SIP_DDR_DVFS 0xc2000004/* Values starting from 0 switch to specific frequency */ #define IMX_SIP_DDR_FREQ_SET_HIGH 0x00@@ -78,10 +81,22 @@ struct imx8m_ddrc {struct clk *dram_alt;struct clk *dram_apb;int freq_count;struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; + +/* For measuring load with perf events: */ +struct platform_device *pmu_pdev; +struct pmu *pmu; + +struct perf_event_attr rd_event_attr; +struct perf_event_attr wr_event_attr; +struct perf_event *rd_event; +struct perf_event *wr_event; + +u64 last_rd_val, last_rd_ena, last_rd_run; +u64 last_wr_val, last_wr_ena, last_wr_run; };static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv,unsigned long rate) {@@ -245,10 +260,131 @@ static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) *freq =clk_get_rate (priv->dram_core);return 0 ; } +static int imx8m_ddrc_get_dev_status (struct device *dev, +struct devfreq_dev_status *stat) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + stat->current_frequency =clk_get_rate (priv->dram_core); + +if (priv->rd_event && priv->wr_event) { +u64 rd_delta, rd_val, rd_ena, rd_run; +u64 wr_delta, wr_val, wr_ena, wr_run; + + rd_val =perf_event_read_value (priv->rd_event, + &rd_ena, &rd_run); + wr_val =perf_event_read_value (priv->wr_event, + &wr_ena, &wr_run); + + rd_delta = (rd_val - priv->last_rd_val) * + (rd_ena - priv->last_rd_ena); +do_div (rd_delta, rd_run - priv->last_rd_run); + priv->last_rd_val = rd_val; + priv->last_rd_ena = rd_ena; + priv->last_rd_run = rd_run; + + wr_delta = (wr_val - priv->last_wr_val) * + (wr_ena - priv->last_wr_ena); +do_div (wr_delta, wr_run - priv->last_wr_run); + priv->last_wr_val = wr_val; + priv->last_wr_ena = wr_ena; + priv->last_wr_run = wr_run; + +/* magic numbers, possibly wrong */ + stat->busy_time =4 * (rd_delta + wr_delta); + stat->total_time = stat->current_frequency; + }else { + stat->busy_time =0 ; + stat->total_time =0 ; + } + +return 0 ; +} + +static int imx8m_ddrc_perf_disable (struct imx8m_ddrc *priv) +{ +/* release and set to NULL */ +if (!IS_ERR_OR_NULL (priv->rd_event)) +perf_event_release_kernel (priv->rd_event); +if (!IS_ERR_OR_NULL (priv->wr_event)) +perf_event_release_kernel (priv->wr_event); + priv->rd_event =NULL ; + priv->wr_event =NULL ; +
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 319
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c +4
-0+return 0 ; +} + +static int imx8m_ddrc_perf_enable (struct imx8m_ddrc *priv)Actually, this function have to call the just function for enabling the bandwidth monitoring instead of writing the detailed something. It looks like that you move the part of the different device driver into here.
This is the code for enabling bandwith monitoring: it creates perf counters using in-kernel API. The perf api doesn't seem to expose anything else to enable/disable the counter after creation and honestly just create/destroy seems fine.
As far as I can tell bandwidth monitoring in devfreq is always enabled on probe anyway, no matter which governor is in use. It would be nice if devfreq drivers could receive a callback and dynamically acquire/release monitoring resources only when the ondemand governor is in used.
Until then this driver will permanently allocate 2 (out of 3) counters in ddr pmu hardware.
Quoted 67 lines
Quoted 24 lines
drivers/devfreq/imx8m-ddrc.c +24
-0+{ +int ret; + + priv->rd_event_attr.size =sizeof (priv->rd_event_attr); + priv->rd_event_attr.type = priv->pmu->type; + priv->rd_event_attr.config =0x2a ; + + priv->rd_event =perf_event_create_kernel_counter ( + &priv->rd_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->rd_event)) { + ret =PTR_ERR (priv->rd_event); + goto err; + } + + priv->wr_event_attr.size =sizeof (priv->wr_event_attr); + priv->wr_event_attr.type = priv->pmu->type; + priv->wr_event_attr.config =0x2b ; + + priv->wr_event =perf_event_create_kernel_counter ( + &priv->wr_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->wr_event)) { + ret =PTR_ERR (priv->wr_event); + goto err; + }drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 353
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+ +return 0 ; + +err: +imx8m_ddrc_perf_disable (priv); +return ret; +} + +static int imx8m_ddrc_init_events (struct device *dev, +struct device_node *events_node)ditto.
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c +12
-0+{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct device_driver *driver; + +/* +* We need pmu->type for perf_event_attr but there is no API for +* mapping device_node to pmu. Fetch private data for imx-ddr-pmu and +* cast that to a struct pmu instead. +*/ + priv->pmu_pdev =of_find_device_by_node (events_node); +if (!priv->pmu_pdev) +return -EPROBE_DEFER ;drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 375
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+ driver = priv->pmu_pdev->dev.driver; +if (!driver) +return -EPROBE_DEFER ; +if (strcmp (driver->name,"imx-ddr-pmu" )) { +dev_warn (dev,"devfreq-events node %pOF has unexpected driver %s\n" , + events_node, driver->name); +return -ENODEV ; + } + + priv->pmu =platform_get_drvdata (priv->pmu_pdev);It seems that you access the different device driver without any standard interface from some linux kernel subsystem.
For the communication or control between different device drivers, we have to use the standard interface instead of direct access or call. I think that it make it too tightly coupled driver between them.
Quoted 7 lines
So, I developed the devfreq-event subsystem for this reason in order to remove the non-standard direct access to other device driver.
Unfortunately, I can't agree this approach. I don't prefer to use the direct access of other device driver(imx-ddr-pmu). You need to use standard interface provided from subsystem. or You need to make the new device driver with devfreq-event subsystem.
This could be cleaned-up by adding a new API to "fetch struct pmu* by struct device_node*" as available for many other subsystems. The perf api is otherwise standard/generic and has a few other in-kernel users.
The perf driver for DDR PMU is already functional and useful for profiling and reusing it seem very worthwhile. If you're suggesting I implemented an unrelated "devfreq-event" driver then how would it get probed? There's only one PMU node in DT.
I wouldn't mind to delay the monitoring part into a second series.
Quoted 72 lines
Quoted 72 lines
drivers/devfreq/imx8m-ddrc.c +27
-0+if (!priv->pmu) { +dev_err (dev,"devfreq-events device missing private data\n" ); +return -EINVAL ; + } + +dev_dbg (dev,"events from pmu %s\n" , priv->pmu->name); + +return imx8m_ddrc_perf_enable (priv); +} +static int imx8m_ddrc_init_freq_info (struct device *dev) {struct imx8m_ddrc *priv =dev_get_drvdata (dev);struct arm_smccc_res res;int index;@@ -328,17 +464,23 @@ static int imx8m_ddrc_check_opps(struct device *dev)return 0 ; }static void imx8m_ddrc_exit (struct device *dev) { +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev); +dev_pm_opp_of_remove_table (dev); }static int imx8m_ddrc_probe (struct platform_device *pdev) {struct device *dev = &pdev->dev;struct imx8m_ddrc *priv; +struct device_node *events_node;const char *gov =DEVFREQ_GOV_USERSPACE ;int ret; priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL );if (!priv)@@ -350,10 +492,19 @@ static int imx8m_ddrc_probe(struct platform_device *pdev)if (ret) {dev_err (dev,"failed to init firmware freq info: %d\n" , ret);return ret; } + events_node =of_parse_phandle (dev->of_node,"devfreq-events" ,0 ); +if (events_node) { + ret =imx8m_ddrc_init_events (dev, events_node); +of_node_put (events_node); +if (ret) + goto err; + gov =DEVFREQ_GOV_SIMPLE_ONDEMAND ; + } + priv->dram_core =devm_clk_get (dev,"core" ); priv->dram_pll =devm_clk_get (dev,"pll" ); priv->dram_alt =devm_clk_get (dev,"alt" ); priv->dram_apb =devm_clk_get (dev,"apb" ); if (IS_ERR(priv->dram_core) ||@@ -390,10 +541,12 @@ static int imx8m_ddrc_probe(struct platform_device *pdev) }return 0 ; err: +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev);dev_pm_opp_of_remove_table (dev);return ret; }static const struct of_device_id imx8m_ddrc_of_match[] = {
Re: [PATCH v4 1/6] clk: imx8m: Set CLK_GET_RATE_NOCACHE on dram clocks
In reply to Abel VesaQuoted 31 lines · Abel Vesa
Quoted 30 lines · Leonard Crestez
These clocks are only modified as part of DRAM frequency switches during which DRAM itself is briefly inaccessible. The switch is performed with a SMC call to by TF-A which runs from a SRAM area; upon returning to linux several clocks bits are modified and we need to update them.
For rate bits an easy solution is to just mark with
CLK_GET_RATE_NOCACHEso that new rates are always read back from registers.Signed-off-by Leonard Crestez <[email protected]>--- drivers/clk/imx/clk-imx8mm.c | 11 +++++++++-- drivers/clk/imx/clk-imx8mn.c | 12 ++++++++++-- drivers/clk/imx/clk-imx8mq.c | 15 +++++++++++---- 3 files changed, 30 insertions(+), 8 deletions(-)drivers/clk/imx/clk-imx8mn.c +4
-2@@ -428,12 +428,20 @@ static int imx8mn_clocks_probe(struct platform_device *pdev) clks[IMX8MN_CLK_AHB ] =imx8m_clk_composite_critical ("ahb" , imx8mn_ahb_sels, base +0x9000 ); clks[IMX8MN_CLK_AUDIO_AHB ] =imx8m_clk_composite ("audio_ahb" , imx8mn_audio_ahb_sels, base +0x9100 ); clks[IMX8MN_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MN_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); clks[IMX8MN_CLK_DRAM_CORE ] =imx_clk_mux2_flags ("dram_core_clk" , base +0x9800 ,24 ,1 , imx8mn_dram_core_sels,ARRAY_SIZE (imx8mn_dram_core_sels),CLK_IS_CRITICAL ); - clks[IMX8MN_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 ); - clks[IMX8MN_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 ); + +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware
Quoted 8 lines
drivers/clk/imx/clk-imx8mn.c static int imx8mn_clocks_probe(struct platform_device *pdev) · line 441
Quoted 5 lines
drivers/clk/imx/clk-imx8mn.c +5
-0+*/ + clks[IMX8MN_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MN_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE );nitpick: I think it looks better if we stick to one line each clock. I know it's against the 80 chars rule, but at least is consistent.
Yes, there are longer lines in the imx8m* files anyway.
If I fix this (in all instances) can I also add a "reviewed-by"?
--
Regards,
Leonard
Quoted 217 lines · Chanwoo Choi
Quoted 216 lines · Leonard Crestez
Quoted 54 lines · Chanwoo Choi
Hi Leonard,
Quoted 51 lines · Leonard Crestez
Add driver for dynamic scaling the DDR Controller on imx8m chips. Actual frequency switching is implemented inside TF-A, this driver wraps the SMC calls and synchronizes the clk tree.
The DRAM clocks on imx8m have the following structure (abridged):
+----------+ |\ +------+ | dram_pll |-------|M| dram_core | | +----------+ |U|---------->| D | /--|X| | D | dram_alt_root | |/ | R | | | C | +---------+ | | |FIX DIV/4| | | +---------+ | | composite: | | | +----------+ | | | | dram_alt |----/ | | +----------+ | | | dram_apb |-------------------->| | +----------+ +------+The
dram_pllis used for higher rates anddram_altis used for lower rates. Thedram_altanddram_apbclocks are "imx composite" and their parent can also be modified.This driver will prepare/enable the new parents ahead of switching (so that the expected roots are enabled) and afterwards it will call
clk_set_parentto ensure the parents in clock framework are up-to-date.The driver relies on
dram_plldram_altanddram_apbbeing marked withCLK_GET_RATE_NOCACHEfor rate updates.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/Kconfig | 10 + drivers/devfreq/Makefile | 1 + drivers/devfreq/imx8m-ddrc.c | 416 +++++++++++++++++++++++++++++++++++ 3 files changed, 427 insertions(+)create mode 100644
drivers/devfreq/imx8m-ddrc.cdrivers/devfreq/Kconfig +0
-0@@ -90,10 +90,20 @@ config ARM_EXYNOS_BUS_DEVFREQ Each memory bus group could contain many memoby bus block. It reads PPMU counters of memory controllers by using DEVFREQ-event device and adjusts the operating frequencies and voltages with OPP support. This does not yet operate with optimal voltages.Quoted 9 lines
drivers/devfreq/Kconfig config ARM_EXYNOS_BUS_DEVFREQ · line 100
Quoted 6 lines
drivers/devfreq/Kconfig +6
-0+config ARM_IMX8M_DDRC_DEVFREQ + tristate "i.MX8M DDRC DEVFREQ Driver" + depends on ARCH_MXC || COMPILE_TEST + select DEVFREQ_GOV_SIMPLE_ONDEMAND + select DEVFREQ_GOV_USERSPACE + select PM_OPPIt doesn't need to add.
CONFIG_DEVFREQadd 'selectPM_OPP' because 'PM_OPP' is mandatory for devfreq.OK.
All other drivers select
PM_OPP, this was just copiedQuoted 146 lines
Quoted 134 lines
drivers/devfreq/Kconfig +4
-0+ help + This adds the DEVFREQ driver for the i.MX8M DDR Controller. It allows + adjusting DRAM frequency. + config ARM_TEGRA_DEVFREQ tristate "NVIDIA Tegra30/114/124/210 DEVFREQ Driver" depends on ARCH_TEGRA_3x_SOC || ARCH_TEGRA_114_SOC || \ ARCH_TEGRA_132_SOC || ARCH_TEGRA_124_SOC || \ ARCH_TEGRA_210_SOC || \drivers/devfreq/Makefile +1
-0@@ -7,10 +7,11 @@ obj-$(CONFIG_DEVFREQ_GOV_POWERSAVE) += governor_powersave.o obj-$(CONFIG_DEVFREQ_GOV_USERSPACE) += governor_userspace.o obj-$(CONFIG_DEVFREQ_GOV_PASSIVE) += governor_passive.o # DEVFREQ Drivers obj-$(CONFIG_ARM_EXYNOS_BUS_DEVFREQ) += exynos-bus.o +obj-$(CONFIG_ARM_IMX8M_DDRC_DEVFREQ) += imx8m-ddrc.o obj-$(CONFIG_ARM_RK3399_DMC_DEVFREQ) += rk3399_dmc.o obj-$(CONFIG_ARM_TEGRA_DEVFREQ) += tegra30-devfreq.o obj-$(CONFIG_ARM_TEGRA20_DEVFREQ) += tegra20-devfreq.o # DEVFREQ Event Driversdrivers/devfreq/imx8m-ddrc.c new file +103
-0@@ -0,0 +1,416 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright 2019 NXP +*/ + +#include <linux/module.h> +#include <linux/device.h> +#include <linux/of_device.h> +#include <linux/platform_device.h> +#include <linux/devfreq.h> +#include <linux/pm_opp.h> +#include <linux/clk.h> +#include <linux/clk-provider.h> +#include <linux/arm-smccc.h> + +#define IMX_SIP_DDR_DVFS 0xc2000004 + +/* Values starting from 0 switch to specific frequency */ +#define IMX_SIP_DDR_FREQ_SET_HIGH 0x00 + +/* Deprecated after moving IRQ handling to ATF */ +#define IMX_SIP_DDR_DVFS_WAIT_CHANGE 0x0F + +/* Query available frequencies. */ +#define IMX_SIP_DDR_DVFS_GET_FREQ_COUNT 0x10 +#define IMX_SIP_DDR_DVFS_GET_FREQ_INFO 0x11 + +/* +* This should be in a 1:1 mapping with devicetree OPPs but +* firmware provides additional info. +*/ +struct imx8m_ddrc_freq { +unsigned long rate; +unsigned long smcarg; +int dram_core_parent_index; +int dram_alt_parent_index; +int dram_apb_parent_index; +}; + +/* Hardware limitation */ +#define IMX8M_DDRC_MAX_FREQ_COUNT 4 + +/* +* i.MX8M DRAM Controller clocks have the following structure (abridged): +* +* +----------+ |\ +------+ +* | dram_pll |-------|M| dram_core | | +* +----------+ |U|---------->| D | +* /--|X| | D | +* dram_alt_root | |/ | R | +* | | C | +* +---------+ | | +* |FIX DIV/4| | | +* +---------+ | | +* composite: | | | +* +----------+ | | | +* | dram_alt |----/ | | +* +----------+ | | +* | dram_apb |-------------------->| | +* +----------+ +------+ +* +* The dram_pll is used for higher rates and dram_alt is used for lower rates. +* +* Frequency switching is implemented in TF-A (via SMC call) and can change the +* configuration of the clocks, including mux parents. The dram_alt and +* dram_apb clocks are "imx composite" and their parent can change too. +* +* We need to prepare/enable the new mux parents head of switching and update +* their information afterwards. +*/ +struct imx8m_ddrc { +struct devfreq_dev_profile profile; +struct devfreq *devfreq; + +/* For frequency switching: */ +struct clk *dram_core; +struct clk *dram_pll; +struct clk *dram_alt; +struct clk *dram_apb; + +int freq_count; +struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; +}; + +static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv, +unsigned long rate) +{ +struct imx8m_ddrc_freq *freq; +int i; + +/* +* Firmware reports values in MT/s, so we round-down from Hz +* Rounding is extra generous to ensure a match. +*/ + rate =DIV_ROUND_CLOSEST (rate,250000 ); +for (i =0 ; i < priv->freq_count; ++i) { + freq = &priv->freq_table[i]; +if (freq->rate == rate || + freq->rate +1 == rate || + freq->rate -1 == rate) +return freq; + } +drivers/devfreq/imx8m-ddrc.c line 113
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+return NULL ; +} + +static void imx8m_ddrc_smc_set_freq (int target_freq) +{ +struct arm_smccc_res res; +u32 online_cpus =0 ; +int cpu; + +local_irq_disable ();
local_irq_disableis more proper thanlocal_irq_save()?
Quoted 6 lines
Quoted 3 lines
There's no need to use save/restore because we know that irqs are enabled when entering the function. This is only called from devfreq
set_targetwhich runs in process context.
local_irqmight affect the whole architecture and other running device driver which requires the some CPU irq at the same time. Don't need to save the irqflag?
As far as I know IRQ disabling implicitly disables all forms of preemption, no other code can run until irqs are restored.
The difference between disable/enable and save/restore is that the latter also works when function is called with irqs already disabled but that is not required here. We know that devfreq set_target is called in process context with irqs enabled, otherwise it would be impossible to use the OPP framework (it uses mutexes internally) or touch regulators (they frequently sleep waiting on i2c transactions).
Quoted 33 lines
Quoted 33 lines
Quoted 32 lines
drivers/devfreq/imx8m-ddrc.c line 122
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +8
-0+for_each_online_cpu (cpu) + online_cpus |= (1 << (cpu *8 )); + +/* change the ddr freqency */ +arm_smccc_smc (IMX_SIP_DDR_DVFS , target_freq, online_cpus, +0 ,0 ,0 ,0 ,0 , &res); + +local_irq_enable ();ditto.
local_irq_restore()instead oflocal_irq_enable()Quoted 21 lines
drivers/devfreq/imx8m-ddrc.c new file +20
-0+} + +struct clk *clk_get_parent_by_index (struct clk *clk,int index) +{ +struct clk_hw *hw; + + hw =clk_hw_get_parent_by_index (__clk_get_hw (clk), index); + +return hw ? hw->clk :NULL ; +} + +static int imx8m_ddrc_set_freq (struct device *dev,struct imx8m_ddrc_freq *freq) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct clk *new_dram_core_parent; +struct clk *new_dram_alt_parent; +struct clk *new_dram_apb_parent; +int ret; + + new_dram_core_parent = clk_get_parent_by_index(You can use '
clk_hw_get_parent_by_index' directly.
Quoted 4 lines
Quoted 1 line
The helper above avoids duplicating ?:
I think that it is not duplicate. You use '
clk_hw_get_parent_by_index' function and check the return value. it is essential behavior.
The clk_hw_get_parent_by_index function returns a clk_hw and I'm using the inner hw->clk. Without clk_get_parent_by_index I'd have to write:
hw = clk_hw_get_parent_by_index
if (!hw || !hw->clk) {
dev_err(...);
return
}
clk = hw->clk;This seems pointlessly verbose.
Quoted 10 lines
Quoted 6 lines
Quoted 1 line
And, you need to check whether the return value is NULL or not.
Yes for
dram_core_parentbut others are deliberately allowed to be NULL. For some setpoints some parents might be indifferent or require no changes. For example whendram_corehas pll as parent the expectation is thatdram_allwill be OFF so parent is irrelevant.If all case is not same as you commented, you just add the proper
dev_warnordev_infomessage according to the h/w characteristic. Because user cannot know the detailed reason why don't check the return value.
Really it's only freq->dram_alt/apb_parent_index which can be zero, if that has a valid value then clk fetching shouldn't fail for any other reason. Current code could indeed hide some errors.
Quoted 238 lines
Quoted 77 lines
Driver relies on the fact that clock API ignores prepare/enable for NULL clocks. I can add a comment explaining that.
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c line 144
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_core, freq->dram_core_parent_index - 1); + new_dram_alt_parent = clk_get_parent_by_index(ditto.
drivers/devfreq/imx8m-ddrc.c line 146
Quoted 2 lines
drivers/devfreq/imx8m-ddrc.c new file +2
-0+ priv->dram_alt, freq->dram_alt_parent_index - 1); + new_dram_apb_parent = clk_get_parent_by_index(ditto.
Quoted 15 lines
drivers/devfreq/imx8m-ddrc.c line 158
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c new file +12
-0+ priv->dram_apb, freq->dram_apb_parent_index - 1); + +/* increase reference counts and ensure clks are ON before switch */ + ret =clk_prepare_enable (new_dram_core_parent); +if (ret) { +dev_err (dev,"failed enable new dram_core parent: %d\n" , ret); + goto out; + } + ret =clk_prepare_enable (new_dram_alt_parent); +if (ret) { +dev_err (dev,"failed enable new dram_alt parent: %d\n" , ret); + goto out_dis_core;I think that 'dis' is not general expression for 'disable'. Just, I think that '
out_disable_core' is better than 'out_dis_core'.OK
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 163
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } + ret =clk_prepare_enable (new_dram_apb_parent); +if (ret) { +dev_err (dev,"failed enable new dram_apb parent: %d\n" , ret); + goto out_dis_alt;ditto.
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c line 171
Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c new file +8
-0+ } + +imx8m_ddrc_smc_set_freq (freq->smcarg); + +/* update parents in clk tree after switch. */ + ret =clk_set_parent (priv->dram_core, new_dram_core_parent); +if (ret) +dev_err (dev,"failed set dram_core parent: %d\n" , ret);if you don't return directly, you better to use '
dev_warn' instead of 'dev_err'.OK.
Quoted 15 lines
drivers/devfreq/imx8m-ddrc.c line 175
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+if (new_dram_alt_parent) { + ret =clk_set_parent (priv->dram_alt, new_dram_alt_parent); +if (ret) +dev_err (dev,"failed set dram_alt parent: %d\n" , ret);ditto. Use
dev_warn.drivers/devfreq/imx8m-ddrc.c line 180
Quoted 5 lines
drivers/devfreq/imx8m-ddrc.c new file +5
-0+ } +if (new_dram_apb_parent) { + ret =clk_set_parent (priv->dram_apb, new_dram_apb_parent); +if (ret) +dev_err (dev,"failed set dram_apb parent: %d\n" , ret);ditto. Use
dev_warn.Quoted 9 lines
drivers/devfreq/imx8m-ddrc.c line 186
Quoted 6 lines
drivers/devfreq/imx8m-ddrc.c new file +6
-0+ } + +/* +* Explicitly refresh dram PLL rate. +* +* Even if it's marked with CLK_GET_RATE_NOCACHE the rate will not be nitpick:
What is more proper description either or 'Even if' or 'if' ?Quoted 3 lines
Rate updates work for dram_alt/apb but not for
dram_pllbecause additional clocks might be present between the PLL anddram_coremux. This happens *even if* pll is marked withCLK_GET_RATE_NOCACHE.ok.
Quoted 134 lines
Quoted 45 lines
Quoted 32 lines
drivers/devfreq/imx8m-ddrc.c new file +31
-0+* automatically refreshed when clk_get_rate is called on children. +*/ +clk_get_rate (priv->dram_pll); + +/* +* clk_set_parent transfer the reference count from old parent. +* now we drop extra reference counts used during the switch +*/ +clk_disable_unprepare (new_dram_apb_parent); +out_dis_alt: +clk_disable_unprepare (new_dram_alt_parent); +out_dis_core: +clk_disable_unprepare (new_dram_core_parent); +out: +return ret; +} + +static int imx8m_ddrc_target (struct device *dev,unsigned long *freq,u32 flags) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct imx8m_ddrc_freq *freq_info; +struct dev_pm_opp *new_opp; +unsigned long old_freq, new_freq; +int ret; + + new_opp =devfreq_recommended_opp (dev, freq, flags); +if (IS_ERR (new_opp)) { + ret =PTR_ERR (new_opp); +dev_err (dev,"failed to get recommended opp: %d\n" , ret); +return ret; + }drivers/devfreq/imx8m-ddrc.c line 227
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+dev_pm_opp_put (new_opp); + + old_freq =clk_get_rate (priv->dram_core); +if (*freq == old_freq) +return 0 ; + + freq_info =imx8m_ddrc_find_freq (priv, *freq); +if (!freq_info) +return -EINVAL ; + ret =imx8m_ddrc_set_freq (dev, freq_info);Need to check whether 'ret' is valid or not.
Quoted 7 lines
drivers/devfreq/imx8m-ddrc.c line 231
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c new file +4
-0+ +/* Also read back the clk rate to verify switch was correct */ + new_freq =clk_get_rate (priv->dram_core); +if (ret || *freq != new_freq)You should check 'ret' right after
imx8m_ddrc_set_freq()instead of this position.OK, I can add two error paths.
Quoted 30 lines
Quoted 17 lines
drivers/devfreq/imx8m-ddrc.c new file +17
-0+dev_err (dev,"ddrc failed freq set to %lu from %lu, now at %lu\n" , + old_freq, *freq, new_freq); + else + dev_dbg(dev, "ddrc freq set to %lu (was %lu)\n", + *freq, old_freq); + +return ret; +} + +static int imx8m_ddrc_get_cur_freq (struct device *dev,unsigned long *freq) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + *freq =clk_get_rate (priv->dram_core); + +return 0 ; +}drivers/devfreq/imx8m-ddrc.c line 258
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c new file +10
-0+ +static int imx8m_ddrc_init_freq_info (struct device *dev) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct arm_smccc_res res; +int index; + +/* +* An error here means DDR DVFS API not supported by firmware +*/ Don't need to add multiple line comments. It is possible to be change as following:
Quoted 1 line
/* An error here means DDR DVFS API not supported by firmware */OK.
Quoted 37 lines
Quoted 37 lines
drivers/devfreq/imx8m-ddrc.c new file +37
-0+arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_COUNT , +0 ,0 ,0 ,0 ,0 ,0 , &res); + priv->freq_count = res.a0; +if (priv->freq_count <=0 || + priv->freq_count >IMX8M_DDRC_MAX_FREQ_COUNT ) +return -ENODEV ; + +for (index =0 ; index < priv->freq_count; ++index) { +struct imx8m_ddrc_freq *freq = &priv->freq_table[index]; + +arm_smccc_smc (IMX_SIP_DDR_DVFS ,IMX_SIP_DDR_DVFS_GET_FREQ_INFO , + index,0 ,0 ,0 ,0 ,0 , &res); +/* Result should be strictly positive */ +if ((long )res.a0 <=0 ) +return -ENODEV ; + + freq->rate = res.a0; + freq->smcarg = index; + freq->dram_core_parent_index = res.a1; + freq->dram_alt_parent_index = res.a2; + freq->dram_apb_parent_index = res.a3; + +/* dram_core has 2 options: dram_pll or dram_alt_root */ +if (freq->dram_core_parent_index !=1 && + freq->dram_core_parent_index !=2 ) +return -ENODEV ; +/* dram_apb and dram_alt have exactly 8 possible parents */ +if (freq->dram_alt_parent_index >8 || + freq->dram_apb_parent_index >8 ) +return -ENODEV ; +/* dram_core from alt requires explicit dram_alt parent */ +if (freq->dram_core_parent_index ==2 && + freq->dram_alt_parent_index ==0 ) +return -ENODEV ; + } + +return 0 ;Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c line 298
Quoted 3 lines
drivers/devfreq/imx8m-ddrc.c new file +3
-0+} + +/* imx8m_ddrc_check_opps() - disable OPPs not supported by firmware */ nitpick:
On this driver, there are no some description for internal function. In order to keep the coding style, you better to move this comment into function before 'freq =ULONG_MAX'.Quoted 2 lines
Is it a problem if only some internal functions have documentation? I can add docs to more functions.
It is not problem. It is just preferable way to add the comment. Usually, if the function is public and exported, I added the detailed function comment as following example of
devfreq_add_device()./**
656 * devfreq_add_device() - Add devfreq feature to the device 657 * @dev: the device to add devfreq feature. 658 * @profile: device-specific profile to run devfreq. 659 * @governor_name: name of the policy to choose frequency. 660 * @data: private data for the governor. The devfreq framework does not 661 * touch this value. 662 */ 663 struct devfreq *devfreq_add_device(struct device *dev, 664 struct devfreq_dev_profile *profile, 665 const char *governor_name, 666 void *data)
Quoted 12 lines
Otherwise, if function is used in the only local with static keyword, usually I don't add the above detailed description. Instead, add the detailed comment into the function.
It is not the standard way. Just preferable way for the function description if possible.
Quoted 3 lines
BTW: how can I generate devfreq documentation from these comments?
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Abel Vesa
Reviewed-byRe: [PATCH v4 1/6] clk: imx8m: Set CLK_GET_RATE_NOCACHE on dram clocks
In reply to Leonard CrestezQuoted 41 lines · Leonard Crestez
Quoted 31 lines · Abel Vesa
Quoted 30 lines · Leonard Crestez
These clocks are only modified as part of DRAM frequency switches during which DRAM itself is briefly inaccessible. The switch is performed with a SMC call to by TF-A which runs from a SRAM area; upon returning to linux several clocks bits are modified and we need to update them.
For rate bits an easy solution is to just mark with
CLK_GET_RATE_NOCACHEso that new rates are always read back from registers.Signed-off-by Leonard Crestez <[email protected]>--- drivers/clk/imx/clk-imx8mm.c | 11 +++++++++-- drivers/clk/imx/clk-imx8mn.c | 12 ++++++++++-- drivers/clk/imx/clk-imx8mq.c | 15 +++++++++++---- 3 files changed, 30 insertions(+), 8 deletions(-)drivers/clk/imx/clk-imx8mn.c +4
-2@@ -428,12 +428,20 @@ static int imx8mn_clocks_probe(struct platform_device *pdev) clks[IMX8MN_CLK_AHB ] =imx8m_clk_composite_critical ("ahb" , imx8mn_ahb_sels, base +0x9000 ); clks[IMX8MN_CLK_AUDIO_AHB ] =imx8m_clk_composite ("audio_ahb" , imx8mn_audio_ahb_sels, base +0x9100 ); clks[IMX8MN_CLK_IPG_ROOT ] =imx_clk_divider2 ("ipg_root" ,"ahb" , base +0x9080 ,0 ,1 ); clks[IMX8MN_CLK_IPG_AUDIO_ROOT ] =imx_clk_divider2 ("ipg_audio_root" ,"audio_ahb" , base +0x9180 ,0 ,1 ); clks[IMX8MN_CLK_DRAM_CORE ] =imx_clk_mux2_flags ("dram_core_clk" , base +0x9800 ,24 ,1 , imx8mn_dram_core_sels,ARRAY_SIZE (imx8mn_dram_core_sels),CLK_IS_CRITICAL ); - clks[IMX8MN_CLK_DRAM_ALT ] =imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 ); - clks[IMX8MN_CLK_DRAM_APB ] =imx8m_clk_composite_critical ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 ); + +/* +* DRAM clocks are manipulated from TF-A outside clock framework. +* Mark with GET_RATE_NOCACHE to always read div value from hardware Quoted 8 lines
drivers/clk/imx/clk-imx8mn.c static int imx8mn_clocks_probe(struct platform_device *pdev) · line 441
Quoted 5 lines
drivers/clk/imx/clk-imx8mn.c +5
-0+*/ + clks[IMX8MN_CLK_DRAM_ALT ] =__imx8m_clk_composite ("dram_alt" , imx8mn_dram_alt_sels, base +0xa000 , +CLK_GET_RATE_NOCACHE ); + clks[IMX8MN_CLK_DRAM_APB ] =__imx8m_clk_composite ("dram_apb" , imx8mn_dram_apb_sels, base +0xa080 , +CLK_IS_CRITICAL |CLK_GET_RATE_NOCACHE );nitpick: I think it looks better if we stick to one line each clock. I know it's against the 80 chars rule, but at least is consistent.
Quoted 4 lines
Yes, there are longer lines in the imx8m* files anyway.
If I fix this (in all instances) can I also add a "reviewed-by"?
Sorry, I forgot to add the line.
For all the clock related changes:
Quoted 3 lines
--
Regards,
Leonard
Chanwoo Choi
maintainerRe: [PATCH v4 5/6] PM / devfreq: imx8m-ddrc: Measure bandwidth with perf
In reply to Leonard CrestezHi Leonard,
Quoted 128 lines · Leonard Crestez
Quoted 113 lines · Chanwoo Choi
Hi Leonard,
Quoted 110 lines · Leonard Crestez
The imx8m ddrc has a performance monitoring block attached which can be used to measure bandwidth usage and automatically adjust frequency.
There is already a perf driver for that block so instead of implementing a devfreq events driver use the in-kernel perf API to implement
get_dev_statusdirectly.Signed-off-by Leonard Crestez <[email protected]>--- drivers/devfreq/imx8m-ddrc.c | 153 +++++++++++++++++++++++++++++++++++ 1 file changed, 153 insertions(+)drivers/devfreq/imx8m-ddrc.c +66
-0@@ -11,10 +11,13 @@#include <linux/pm_opp.h> #include <linux/clk.h> #include <linux/clk-provider.h> #include <linux/arm-smccc.h> +#include <asm/perf_event.h> +#include <linux/perf_event.h> +#define IMX_SIP_DDR_DVFS 0xc2000004/* Values starting from 0 switch to specific frequency */ #define IMX_SIP_DDR_FREQ_SET_HIGH 0x00@@ -78,10 +81,22 @@ struct imx8m_ddrc {struct clk *dram_alt;struct clk *dram_apb;int freq_count;struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT ]; + +/* For measuring load with perf events: */ +struct platform_device *pmu_pdev; +struct pmu *pmu; + +struct perf_event_attr rd_event_attr; +struct perf_event_attr wr_event_attr; +struct perf_event *rd_event; +struct perf_event *wr_event; + +u64 last_rd_val, last_rd_ena, last_rd_run; +u64 last_wr_val, last_wr_ena, last_wr_run; };static struct imx8m_ddrc_freq *imx8m_ddrc_find_freq (struct imx8m_ddrc *priv,unsigned long rate) {@@ -245,10 +260,131 @@ static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) *freq =clk_get_rate (priv->dram_core);return 0 ; } +static int imx8m_ddrc_get_dev_status (struct device *dev, +struct devfreq_dev_status *stat) +{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + + stat->current_frequency =clk_get_rate (priv->dram_core); + +if (priv->rd_event && priv->wr_event) { +u64 rd_delta, rd_val, rd_ena, rd_run; +u64 wr_delta, wr_val, wr_ena, wr_run; + + rd_val =perf_event_read_value (priv->rd_event, + &rd_ena, &rd_run); + wr_val =perf_event_read_value (priv->wr_event, + &wr_ena, &wr_run); + + rd_delta = (rd_val - priv->last_rd_val) * + (rd_ena - priv->last_rd_ena); +do_div (rd_delta, rd_run - priv->last_rd_run); + priv->last_rd_val = rd_val; + priv->last_rd_ena = rd_ena; + priv->last_rd_run = rd_run; + + wr_delta = (wr_val - priv->last_wr_val) * + (wr_ena - priv->last_wr_ena); +do_div (wr_delta, wr_run - priv->last_wr_run); + priv->last_wr_val = wr_val; + priv->last_wr_ena = wr_ena; + priv->last_wr_run = wr_run; + +/* magic numbers, possibly wrong */ + stat->busy_time =4 * (rd_delta + wr_delta); + stat->total_time = stat->current_frequency; + }else { + stat->busy_time =0 ; + stat->total_time =0 ; + } + +return 0 ; +} + +static int imx8m_ddrc_perf_disable (struct imx8m_ddrc *priv) +{ +/* release and set to NULL */ +if (!IS_ERR_OR_NULL (priv->rd_event)) +perf_event_release_kernel (priv->rd_event); +if (!IS_ERR_OR_NULL (priv->wr_event)) +perf_event_release_kernel (priv->wr_event); + priv->rd_event =NULL ; + priv->wr_event =NULL ; +Quoted 8 lines
drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 319
Quoted 4 lines
drivers/devfreq/imx8m-ddrc.c +4
-0+return 0 ; +} + +static int imx8m_ddrc_perf_enable (struct imx8m_ddrc *priv)Actually, this function have to call the just function for enabling the bandwidth monitoring instead of writing the detailed something. It looks like that you move the part of the different device driver into here.
This is the code for enabling bandwith monitoring: it creates perf counters using in-kernel API. The perf api doesn't seem to expose anything else to enable/disable the counter after creation and honestly just create/destroy seems fine.
Quoted 4 lines
As far as I can tell bandwidth monitoring in devfreq is always enabled on probe anyway, no matter which governor is in use. It would be nice if devfreq drivers could receive a callback and dynamically acquire/release monitoring resources only when the ondemand governor is in used.
As you commented, it doesn't matter to keep the enabling state of perf. Just, I want to say that I think that it is not proper to access the different device driver directly without any standard subsystem interface.
Quoted 74 lines
Until then this driver will permanently allocate 2 (out of 3) counters in ddr pmu hardware.
Quoted 70 lines
Quoted 24 lines
drivers/devfreq/imx8m-ddrc.c +24
-0+{ +int ret; + + priv->rd_event_attr.size =sizeof (priv->rd_event_attr); + priv->rd_event_attr.type = priv->pmu->type; + priv->rd_event_attr.config =0x2a ; + + priv->rd_event =perf_event_create_kernel_counter ( + &priv->rd_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->rd_event)) { + ret =PTR_ERR (priv->rd_event); + goto err; + } + + priv->wr_event_attr.size =sizeof (priv->wr_event_attr); + priv->wr_event_attr.type = priv->pmu->type; + priv->wr_event_attr.config =0x2b ; + + priv->wr_event =perf_event_create_kernel_counter ( + &priv->wr_event_attr,0 ,NULL ,NULL ,NULL ); +if (IS_ERR (priv->wr_event)) { + ret =PTR_ERR (priv->wr_event); + goto err; + }drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 353
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+ +return 0 ; + +err: +imx8m_ddrc_perf_disable (priv); +return ret; +} + +static int imx8m_ddrc_init_events (struct device *dev, +struct device_node *events_node)ditto.
Quoted 12 lines
drivers/devfreq/imx8m-ddrc.c +12
-0+{ +struct imx8m_ddrc *priv =dev_get_drvdata (dev); +struct device_driver *driver; + +/* +* We need pmu->type for perf_event_attr but there is no API for +* mapping device_node to pmu. Fetch private data for imx-ddr-pmu and +* cast that to a struct pmu instead. +*/ + priv->pmu_pdev =of_find_device_by_node (events_node); +if (!priv->pmu_pdev) +return -EPROBE_DEFER ;drivers/devfreq/imx8m-ddrc.c static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq) · line 375
Quoted 10 lines
drivers/devfreq/imx8m-ddrc.c +10
-0+ driver = priv->pmu_pdev->dev.driver; +if (!driver) +return -EPROBE_DEFER ; +if (strcmp (driver->name,"imx-ddr-pmu" )) { +dev_warn (dev,"devfreq-events node %pOF has unexpected driver %s\n" , + events_node, driver->name); +return -ENODEV ; + } + + priv->pmu =platform_get_drvdata (priv->pmu_pdev);It seems that you access the different device driver without any standard interface from some linux kernel subsystem.
For the communication or control between different device drivers, we have to use the standard interface instead of direct access or call. I think that it make it too tightly coupled driver between them.
So, I developed the devfreq-event subsystem for this reason in order to remove the non-standard direct access to other device driver.
Quoted 7 lines
Quoted 4 lines
Unfortunately, I can't agree this approach. I don't prefer to use the direct access of other device driver(imx-ddr-pmu). You need to use standard interface provided from subsystem. or You need to make the new device driver with devfreq-event subsystem.
This could be cleaned-up by adding a new API to "fetch struct pmu* by struct device_node*" as available for many other subsystems.
OK. I'll expect the next version with keeping the rule without any direct access between two device drivers.
The perf
Quoted 7 lines
api is otherwise standard/generic and has a few other in-kernel users.
The perf driver for DDR PMU is already functional and useful for profiling and reusing it seem very worthwhile. If you're suggesting I implemented an unrelated "devfreq-event" driver then how would it get probed? There's only one PMU node in DT.
No, I don't force to develop the new devfreq-event instead of exiting perf driver. As I said already,I think that it is not proper to access the different device driver directly without any standard subsystem interface.
The following style access is not safe. Any device driver might change the platform data of the other device driver.
priv->pmu = platform_get_drvdata(priv->pmu_pdev);
If you support this patch with standard subsystem interface, I'm OK. Just, I don't want to access the data of different device driver.
Quoted 77 lines
I wouldn't mind to delay the monitoring part into a second series.
Quoted 72 lines
Quoted 72 lines
drivers/devfreq/imx8m-ddrc.c +27
-0+if (!priv->pmu) { +dev_err (dev,"devfreq-events device missing private data\n" ); +return -EINVAL ; + } + +dev_dbg (dev,"events from pmu %s\n" , priv->pmu->name); + +return imx8m_ddrc_perf_enable (priv); +} +static int imx8m_ddrc_init_freq_info (struct device *dev) {struct imx8m_ddrc *priv =dev_get_drvdata (dev);struct arm_smccc_res res;int index;@@ -328,17 +464,23 @@ static int imx8m_ddrc_check_opps(struct device *dev)return 0 ; }static void imx8m_ddrc_exit (struct device *dev) { +struct imx8m_ddrc *priv =dev_get_drvdata (dev); + +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev); +dev_pm_opp_of_remove_table (dev); }static int imx8m_ddrc_probe (struct platform_device *pdev) {struct device *dev = &pdev->dev;struct imx8m_ddrc *priv; +struct device_node *events_node;const char *gov =DEVFREQ_GOV_USERSPACE ;int ret; priv =devm_kzalloc (dev,sizeof (*priv),GFP_KERNEL );if (!priv)@@ -350,10 +492,19 @@ static int imx8m_ddrc_probe(struct platform_device *pdev)if (ret) {dev_err (dev,"failed to init firmware freq info: %d\n" , ret);return ret; } + events_node =of_parse_phandle (dev->of_node,"devfreq-events" ,0 ); +if (events_node) { + ret =imx8m_ddrc_init_events (dev, events_node); +of_node_put (events_node); +if (ret) + goto err; + gov =DEVFREQ_GOV_SIMPLE_ONDEMAND ; + } + priv->dram_core =devm_clk_get (dev,"core" ); priv->dram_pll =devm_clk_get (dev,"pll" ); priv->dram_alt =devm_clk_get (dev,"alt" ); priv->dram_apb =devm_clk_get (dev,"apb" ); if (IS_ERR(priv->dram_core) ||@@ -390,10 +541,12 @@ static int imx8m_ddrc_probe(struct platform_device *pdev) }return 0 ; err: +imx8m_ddrc_perf_disable (priv); +platform_device_put (priv->pmu_pdev);dev_pm_opp_of_remove_table (dev);return ret; }static const struct of_device_id imx8m_ddrc_of_match[] = {
Signature · 3 lines
Best Regards, Chanwoo Choi Samsung Electronics
hi Leondard,
before using this patchset I'd like to ask: Do you have plans to create an update and push this forward? It is useful.
thanks a lot,
martin
Quoted 4 lines · Martin Kepplinger
hi Leondard,
before using this patchset I'd like to ask: Do you have plans to create an update and push this forward? It is useful.
Hello.
I am no longer with NXP and don't have access to imx hardware right now.
However the series that you replied to is very old and was accepted many months ago. You shouldn't have to apply out-of-tree kernel patches.
--
Regards,
Leonard
hi Leonard,
Quoted 9 lines · Leonard Crestez
Quoted 4 lines · Martin Kepplinger
hi Leondard,
before using this patchset I'd like to ask: Do you have plans to create an update and push this forward? It is useful.
Hello.
I am no longer with NXP and don't have access to imx hardware right now.
I guess it'll get even harder to get the ATF part for devfreq implemented now :) Thanks for the update and all the best for your new stuff.
Quoted 4 lines
However the series that you replied to is very old and was accepted many months ago. You shouldn't have to apply out-of-tree kernel patches.
that particular series doesn't seem to be in mainline, see
https://elixir.bootlin.com/linux/latest/source/drivers/devfreq/imx8m-ddrc.c#L283
or do I miss something?
do you know who at nxp would be likely actively working on devfreq?
thanks,
martin
Quoted 22 lines · Martin Kepplinger
hi Leonard,
Quoted 9 lines · Leonard Crestez
Quoted 4 lines · Martin Kepplinger
hi Leondard,
before using this patchset I'd like to ask: Do you have plans to create an update and push this forward? It is useful.
Hello.
I am no longer with NXP and don't have access to imx hardware right now.
I guess it'll get even harder to get the ATF part for devfreq implemented now :) Thanks for the update and all the best for your new stuff.
Quoted 4 lines
However the series that you replied to is very old and was accepted many months ago. You shouldn't have to apply out-of-tree kernel patches.
Quoted 5 lines
that particular series doesn't seem to be in mainline, see
https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Felixir.bootlin.com%2Flinux%2Flatest%2Fsource%2Fdrivers%2Fdevfreq%2Fimx8m-ddrc.c%23L283&data=02%7C01%7Cabel.vesa%40nxp.com%7Cb00f437e756d4850238f08d818d51b59%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637286650857331523&sdata=S7%2BN3%2BiTFkUW5YnmVzl36wEBlr%2BkTatGoDDrvY9XfTk%3D&reserved=0
or do I miss something?do you know who at nxp would be likely actively working on devfreq?
Hi Martin,
I will be working on this in the following weeks.
Quoted 3 lines
thanks,
martin
Quoted 27 lines · Abel Vesa
Quoted 26 lines · Martin Kepplinger
hi Leonard,
Quoted 9 lines · Leonard Crestez
Quoted 4 lines · Martin Kepplinger
hi Leondard,
before using this patchset I'd like to ask: Do you have plans to create an update and push this forward? It is useful.
Hello.
I am no longer with NXP and don't have access to imx hardware right now.
I guess it'll get even harder to get the ATF part for devfreq implemented now :) Thanks for the update and all the best for your new stuff.
Quoted 4 lines
However the series that you replied to is very old and was accepted many months ago. You shouldn't have to apply out-of-tree kernel patches.
that particular series doesn't seem to be in mainline, see
https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Felixir.bootlin.com%2Flinux%2Flatest%2Fsource%2Fdrivers%2Fdevfreq%2Fimx8m-ddrc.c%23L283&data=02%7C01%7Cabel.vesa%40nxp.com%7Cb00f437e756d4850238f08d818d51b59%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637286650857331523&sdata=S7%2BN3%2BiTFkUW5YnmVzl36wEBlr%2BkTatGoDDrvY9XfTk%3D&reserved=0
or do I miss something?
Quoted 6 lines
Quoted 1 line
do you know who at nxp would be likely actively working on devfreq?
Hi Martin,
I will be working on this in the following weeks.
hi Abel,
that's good to hear. I'm basically always happy to test changes to devfreq (the ondemand governor or an atf implementation).
thanks,
martin
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel