Skip to contentSkip to conversation

clk: imx8mp: Remove IPG_AUDIO_ROOT from imx8mp-clock.h

PATCH In review Hui Wang 2 participants 5 messages Nov 9, 2021 – Nov 23, 2021

Subsystems (4)
lore
Share & tools
lore

Hui Wang

author
#0 lore raw

PATCH 1/2 clk: imx8mp: Remove IPG_AUDIO_ROOT from imx8mp-clock.h

Since the commit b24e288d5063 ("clk: imx: Remove the audio ipg clock from imx8mp") removes the non-existing IPG_AUDIO_ROOT from the clk-imx8mp.c, and this definition is not used by anywhere, let us removed it in the imx8mp-clock.h as well.

---
 include/dt-bindings/clock/imx8mp-clock.h | 1 -
 1 file changed, 1 deletion(-)
include/dt-bindings/clock/imx8mp-clock.h +0 -1
index 43927a1b9e94..235c7a00d379 100644
@@ -117,7 +117,6 @@
 #define IMX8MP_CLK_AUDIO_AHB			108
 #define IMX8MP_CLK_MIPI_DSI_ESC_RX		109
 #define IMX8MP_CLK_IPG_ROOT			110
-#define IMX8MP_CLK_IPG_AUDIO_ROOT		111
 #define IMX8MP_CLK_DRAM_ALT			112
 #define IMX8MP_CLK_DRAM_APB			113
 #define IMX8MP_CLK_VPU_G1			114
Signature · 1 line
2.25.1
author
#1 lore raw

PATCH 2/2 clk: imx8mp: Fix the parent clk of the audio_root_clk

Recently we tried to enable the BSP on a platform based on imx8mp, we backported the audiomix related drivers to mainline kernel from
https://source.codeaurora.org/external/imx/linux-imx, when kernel boots to the audiomix powerdomain driver, the kernel will hang immediately. That is because we set the audio_root_clk to audiomix in the device tree, but the parent of the audio_root_clk is wrong in the clk-imx8mp.c.

And we could also refer to the section "5.1.4 System Clocks" of the IMX8MPRM.pdf, the parent clk of CCGR101 (Audiomix) is the AUDIO_AHB_CLK_ROOT.

---
 drivers/clk/imx/clk-imx8mp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/clk/imx/clk-imx8mp.c +1 -1
index 12837304545d..c990ad37882b 100644
@@ -700,7 +700,7 @@ static int imx8mp_clocks_probe(struct platform_device *pdev)
 	hws[IMX8MP_CLK_HDMI_ROOT] = imx_clk_hw_gate4("hdmi_root_clk", "hdmi_axi", ccm_base + 0x45f0, 0);
 	hws[IMX8MP_CLK_TSENSOR_ROOT] = imx_clk_hw_gate4("tsensor_root_clk", "ipg_root", ccm_base + 0x4620, 0);
 	hws[IMX8MP_CLK_VPU_ROOT] = imx_clk_hw_gate4("vpu_root_clk", "vpu_bus", ccm_base + 0x4630, 0);
-	hws[IMX8MP_CLK_AUDIO_ROOT] = imx_clk_hw_gate4("audio_root_clk", "ipg_root", ccm_base + 0x4650, 0);
+	hws[IMX8MP_CLK_AUDIO_ROOT] = imx_clk_hw_gate4("audio_root_clk", "audio_ahb", ccm_base + 0x4650, 0);
 
 	hws[IMX8MP_CLK_ARM] = imx_clk_hw_cpu("arm", "arm_a53_core",
 					     hws[IMX8MP_CLK_A53_CORE]->clk,
Signature · 1 line
2.25.1
Reviewed-by on v1 2/2
#2 lore raw

Re: [PATCH 2/2] clk: imx8mp: Fix the parent clk of the audio_root_clk

Quoted 8 lines · Hui Wang

Recently we tried to enable the BSP on a platform based on imx8mp, we backported the audiomix related drivers to mainline kernel from
https://source.codeaurora.org/external/imx/linux-imx, when kernel boots to the audiomix powerdomain driver, the kernel will hang immediately. That is because we set the audio_root_clk to audiomix in the device tree, but the parent of the audio_root_clk is wrong in the clk-imx8mp.c.

Quoted 4 lines

And we could also refer to the section "5.1.4 System Clocks" of the IMX8MPRM.pdf, the parent clk of CCGR101 (Audiomix) is the AUDIO_AHB_CLK_ROOT.

Thanks for the whole explanation, but I would not mention the audiomix and the downstream source of it in the commit message, since it will most probably never be upstreamed.

So lets just stick with the boot hang fixing explanation and what the RM mentions. I'll reword it myself, so you won't have to resend, and then I'll apply it to clk/imx.

Quoted 1 line
Quoted 20 lines
---
 drivers/clk/imx/clk-imx8mp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/clk/imx/clk-imx8mp.c +1 -1
index 12837304545d..c990ad37882b 100644
@@ -700,7 +700,7 @@ static int imx8mp_clocks_probe(struct platform_device *pdev)
 	hws[IMX8MP_CLK_HDMI_ROOT] = imx_clk_hw_gate4("hdmi_root_clk", "hdmi_axi", ccm_base + 0x45f0, 0);
 	hws[IMX8MP_CLK_TSENSOR_ROOT] = imx_clk_hw_gate4("tsensor_root_clk", "ipg_root", ccm_base + 0x4620, 0);
 	hws[IMX8MP_CLK_VPU_ROOT] = imx_clk_hw_gate4("vpu_root_clk", "vpu_bus", ccm_base + 0x4630, 0);
-	hws[IMX8MP_CLK_AUDIO_ROOT] = imx_clk_hw_gate4("audio_root_clk", "ipg_root", ccm_base + 0x4650, 0);
+	hws[IMX8MP_CLK_AUDIO_ROOT] = imx_clk_hw_gate4("audio_root_clk", "audio_ahb", ccm_base + 0x4650, 0);
 
 	hws[IMX8MP_CLK_ARM] = imx_clk_hw_cpu("arm", "arm_a53_core",
 					     hws[IMX8MP_CLK_A53_CORE]->clk,
Signature · 1 line
2.25.1
author
#3 lore raw

Re: [PATCH 2/2] clk: imx8mp: Fix the parent clk of the audio_root_clk

Quoted 13 lines · Abel Vesa
Quoted 8 lines · Hui Wang

Recently we tried to enable the BSP on a platform based on imx8mp, we backported the audiomix related drivers to mainline kernel from
https://source.codeaurora.org/external/imx/linux-imx, when kernel boots to the audiomix powerdomain driver, the kernel will hang immediately. That is because we set the audio_root_clk to audiomix in the device tree, but the parent of the audio_root_clk is wrong in the clk-imx8mp.c.

Quoted 4 lines

And we could also refer to the section "5.1.4 System Clocks" of the IMX8MPRM.pdf, the parent clk of CCGR101 (Audiomix) is the AUDIO_AHB_CLK_ROOT.

Quoted 7 lines

Thanks for the whole explanation, but I would not mention the audiomix and the downstream source of it in the commit message, since it will most probably never be upstreamed.

So lets just stick with the boot hang fixing explanation and what the RM mentions. I'll reword it myself, so you won't have to resend, and then I'll apply it to clk/imx.

OK, got it. Thanks.

Quoted 24 lines
Quoted 2 lines
Quoted 20 lines
---
  drivers/clk/imx/clk-imx8mp.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
drivers/clk/imx/clk-imx8mp.c +1 -1
index 12837304545d..c990ad37882b 100644
@@ -700,7 +700,7 @@ static int imx8mp_clocks_probe(struct platform_device *pdev)
  	hws[IMX8MP_CLK_HDMI_ROOT] = imx_clk_hw_gate4("hdmi_root_clk", "hdmi_axi", ccm_base + 0x45f0, 0);
  	hws[IMX8MP_CLK_TSENSOR_ROOT] = imx_clk_hw_gate4("tsensor_root_clk", "ipg_root", ccm_base + 0x4620, 0);
  	hws[IMX8MP_CLK_VPU_ROOT] = imx_clk_hw_gate4("vpu_root_clk", "vpu_bus", ccm_base + 0x4630, 0);
-	hws[IMX8MP_CLK_AUDIO_ROOT] = imx_clk_hw_gate4("audio_root_clk", "ipg_root", ccm_base + 0x4650, 0);
+	hws[IMX8MP_CLK_AUDIO_ROOT] = imx_clk_hw_gate4("audio_root_clk", "audio_ahb", ccm_base + 0x4650, 0);
 
  	hws[IMX8MP_CLK_ARM] = imx_clk_hw_cpu("arm", "arm_a53_core",
  					     hws[IMX8MP_CLK_A53_CORE]->clk,
Signature · 1 line
2.25.1
Reviewed-by on v1 1/2
#4 lore raw
In reply to Hui Wang
Quoted 6 lines · Hui Wang

Since the commit b24e288d5063 ("clk: imx: Remove the audio ipg clock from imx8mp") removes the non-existing IPG_AUDIO_ROOT from the clk-imx8mp.c, and this definition is not used by anywhere, let us removed it in the imx8mp-clock.h as well.

include/dt-bindings/clock/imx8mp-clock.h line 120

Quoted 13 lines
---
 include/dt-bindings/clock/imx8mp-clock.h | 1 -
 1 file changed, 1 deletion(-)
include/dt-bindings/clock/imx8mp-clock.h +0 -1
index 43927a1b9e94..235c7a00d379 100644
@@ -117,7 +117,6 @@
 #define IMX8MP_CLK_AUDIO_AHB			108
 #define IMX8MP_CLK_MIPI_DSI_ESC_RX		109
 #define IMX8MP_CLK_IPG_ROOT			110
-#define IMX8MP_CLK_IPG_AUDIO_ROOT		111

Seems this hasn't been used since 5.14, so I guess we're safe.

Quoted 6 lines
include/dt-bindings/clock/imx8mp-clock.h +0 -0
 #define IMX8MP_CLK_DRAM_ALT			112
 #define IMX8MP_CLK_DRAM_APB			113
 #define IMX8MP_CLK_VPU_G1			114
Signature · 1 line
2.25.1