clk: rockchip: clk-rk3588: Fix 32k clock name for pmu_24m_32k_100m_src_p
Subsystems (2)
PATCH v2 1/1 clk: rockchip: clk-rk3588: Fix 32k clock name for pmu_24m_32k_100m_src_p
The 32kHz input clock is named "xin32k" in the driver, so the name "32k" appears to be a typo in this case. Lets fix this.
--- drivers/clk/rockchip/clk-rk3588.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/clk/rockchip/clk-rk3588.c +1 -1
@@ -526,7 +526,7 @@ PNAME(pmu_200m_100m_p) = { "clk_pmu1_200m_src", "clk_pmu1_100m_src" };PNAME (pmu_300m_24m_p) = {"clk_300m_src" ,"xin24m" };PNAME (pmu_400m_24m_p) = {"clk_400m_src" ,"xin24m" };PNAME (pmu_100m_50m_24m_src_p) = {"clk_pmu1_100m_src" ,"clk_pmu1_50m_src" ,"xin24m" }; -PNAME (pmu_24m_32k_100m_src_p) = {"xin24m" ,"32k" ,"clk_pmu1_100m_src" }; +PNAME (pmu_24m_32k_100m_src_p) = {"xin24m" ,"xin32k" ,"clk_pmu1_100m_src" };PNAME (hclk_pmu1_root_p) = {"clk_pmu1_200m_src" ,"clk_pmu1_100m_src" ,"clk_pmu1_50m_src" ,"xin24m" };PNAME (hclk_pmu_cm0_root_p) = {"clk_pmu1_400m_src" ,"clk_pmu1_200m_src" ,"clk_pmu1_100m_src" ,"xin24m" };PNAME (mclk_pdm0_p) = {"clk_pmu1_300m_src" ,"clk_pmu1_200m_src" };
Signature · 1 line
2.38.2
Hello Alexander,
Quoted 4 lines · Alexander Shiyan
The 32kHz input clock is named "xin32k" in the driver, so the name "32k" appears to be a typo in this case. Lets fix this.
Signed-off-by Alexander Shiyan <[email protected]>
Makes sense to me, and it seems to be a typo inherited from the downstream code, [1] which the base RK3588 dtsi confirms, [2] as well as the RK3588 Hardware Design Guide, version 1.0.
Thus, please include:
I'd also suggest that this patch receives a Fixes tag, and gets submitted for inclusion into stable kernels. Thus:
... but you should actually submit the v2 with these tags, if you choose to agree with this suggestion.
Furthermore, it seems that the board dts for Radxa ROCK 5B needs some related fixes, because it deviates from the RK3588 EVB design, but I still need to dig deeper into that and actually do some testing to confirm it. Perhaps that will apply to other similar RK3588-based boards as well.
[1] https://raw.githubusercontent.com/rockchip-linux/kernel/develop-5.10/drivers/clk/rockchip/clk-rk3588.c
[2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tree/arch/arm64/boot/dts/rockchip/rk3588-base.dtsi#n423
Quoted 24 lines
--- drivers/clk/rockchip/clk-rk3588.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/clk/rockchip/clk-rk3588.c +1
-1@@ -526,7 +526,7 @@ PNAME(pmu_200m_100m_p) = { "clk_pmu1_200m_src", "clk_pmu1_100m_src" };PNAME (pmu_300m_24m_p) = {"clk_300m_src" ,"xin24m" };PNAME (pmu_400m_24m_p) = {"clk_400m_src" ,"xin24m" };PNAME (pmu_100m_50m_24m_src_p) = {"clk_pmu1_100m_src" ,"clk_pmu1_50m_src" ,"xin24m" }; -PNAME (pmu_24m_32k_100m_src_p) = {"xin24m" ,"32k" , "clk_pmu1_100m_src" }; +PNAME (pmu_24m_32k_100m_src_p) = {"xin24m" ,"xin32k" ,"
clk_pmu1_100m_src" };
PNAME(hclk_pmu1_root_p)= { "clk_pmu1_200m_src",
"clk_pmu1_100m_src", "clk_pmu1_50m_src", "xin24m" };
PNAME(hclk_pmu_cm0_root_p)= { "clk_pmu1_400m_src",
"clk_pmu1_200m_src", "clk_pmu1_100m_src", "xin24m" };
PNAME(mclk_pdm0_p)= { "clk_pmu1_300m_src", "clk_pmu1_200m_src" };
Just a brief reminder about this patch...
Quoted 58 lines · Dragan Simic
Hello Alexander,
Quoted 4 lines · Alexander Shiyan
The 32kHz input clock is named "xin32k" in the driver, so the name "32k" appears to be a typo in this case. Lets fix this.
Signed-off-by Alexander Shiyan <[email protected]>Makes sense to me, and it seems to be a typo inherited from the downstream code, [1] which the base RK3588 dtsi confirms, [2] as well as the RK3588 Hardware Design Guide, version 1.0.
Thus, please include:
Reviewed-by Dragan Simic <[email protected]>I'd also suggest that this patch receives a Fixes tag, and gets submitted for inclusion into stable kernels. Thus:
Fixes f1c506d152ff ("clk: rockchip: add clock controller for the RK3588")... but you should actually submit the v2 with these tags, if you choose to agree with this suggestion.
Furthermore, it seems that the board dts for Radxa ROCK 5B needs some related fixes, because it deviates from the RK3588 EVB design, but I still need to dig deeper into that and actually do some testing to confirm it. Perhaps that will apply to other similar RK3588-based boards as well.
[1] https://raw.githubusercontent.com/rockchip-linux/kernel/develop-5.10/drivers/clk/rockchip/clk-rk3588.c
[2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tree/arch/arm64/boot/dts/rockchip/rk3588-base.dtsi#n423Quoted 24 lines
--- drivers/clk/rockchip/clk-rk3588.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/clk/rockchip/clk-rk3588.c +1
-1@@ -526,7 +526,7 @@ PNAME(pmu_200m_100m_p) = { "clk_pmu1_200m_src", "clk_pmu1_100m_src" };PNAME (pmu_300m_24m_p) = {"clk_300m_src" ,"xin24m" };PNAME (pmu_400m_24m_p) = {"clk_400m_src" ,"xin24m" };PNAME (pmu_100m_50m_24m_src_p) = {"clk_pmu1_100m_src" ,"clk_pmu1_50m_src" ,"xin24m" }; -PNAME (pmu_24m_32k_100m_src_p) = {"xin24m" ,"32k" , "clk_pmu1_100m_src" }; +PNAME (pmu_24m_32k_100m_src_p) = {"xin24m" ,"xin32k" ,"
clk_pmu1_100m_src" };
PNAME(hclk_pmu1_root_p)= { "clk_pmu1_200m_src",
"clk_pmu1_100m_src", "clk_pmu1_50m_src", "xin24m" };
PNAME(hclk_pmu_cm0_root_p)= { "clk_pmu1_400m_src",
"clk_pmu1_200m_src", "clk_pmu1_100m_src", "xin24m" };
PNAME(mclk_pdm0_p)= { "clk_pmu1_300m_src", "clk_pmu1_200m_src" };