| Message ID | 20230323-topic-sm8450-upstream-dt-bindings-fixes-v1-6-3ead1e418fe4@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | arm64: qcom: sm8450: bindings check cleanup | expand |
On 23/03/2023 11:25, Neil Armstrong wrote: > Fixes the following DT bindings check error: > codec@33f0000: clocks: [[137, 57, 1], [137, 102, 1], [137, 103, 1], [137, 70, 1]] is too long > codec@33f0000: clock-names: 'oneOf' conditional failed, one must be fixed: > ['mclk', 'macro', 'dcodec', 'npl'] is too long > > The implementation was checked and this npl clock isn't used for the VA macro. > This does not look correct. DTS looks good, you miss some patches in your tree. Best regards, Krzysztof
On 23/03/2023 14:13, Neil Armstrong wrote: > On 23/03/2023 11:47, Krzysztof Kozlowski wrote: >> On 23/03/2023 11:25, Neil Armstrong wrote: >>> Fixes the following DT bindings check error: >>> codec@33f0000: clocks: [[137, 57, 1], [137, 102, 1], [137, 103, 1], [137, 70, 1]] is too long >>> codec@33f0000: clock-names: 'oneOf' conditional failed, one must be fixed: >>> ['mclk', 'macro', 'dcodec', 'npl'] is too long >>> >>> The implementation was checked and this npl clock isn't used for the VA macro. >>> >> >> This does not look correct. DTS looks good, you miss some patches in >> your tree. > > I'm based on today's linux-next, Which is unfortunately not enough. Several things were fixed/added/changed and are pending. I brought the topic of pending branch few times on IRC for that reason. > while the other lpass macros uses the npl clock, > the lpass vamacro bindings doesn't document the npl clock. > > And I found no fixes whatsover to add the npl clock to bindings. Really? lore finds it easily: https://lore.kernel.org/all/20221118071849.25506-2-srinivas.kandagatla@linaro.org/ Best regards, Krzysztof
On 24/03/2023 08:17, Krzysztof Kozlowski wrote: > On 23/03/2023 14:13, Neil Armstrong wrote: >> On 23/03/2023 11:47, Krzysztof Kozlowski wrote: >>> On 23/03/2023 11:25, Neil Armstrong wrote: >>>> Fixes the following DT bindings check error: >>>> codec@33f0000: clocks: [[137, 57, 1], [137, 102, 1], [137, 103, 1], [137, 70, 1]] is too long >>>> codec@33f0000: clock-names: 'oneOf' conditional failed, one must be fixed: >>>> ['mclk', 'macro', 'dcodec', 'npl'] is too long >>>> >>>> The implementation was checked and this npl clock isn't used for the VA macro. >>>> >>> >>> This does not look correct. DTS looks good, you miss some patches in >>> your tree. >> >> I'm based on today's linux-next, > > Which is unfortunately not enough. Several things were > fixed/added/changed and are pending. I brought the topic of pending > branch few times on IRC for that reason. > >> while the other lpass macros uses the npl clock, >> the lpass vamacro bindings doesn't document the npl clock. >> >> And I found no fixes whatsover to add the npl clock to bindings. > > Really? lore finds it easily: > > https://lore.kernel.org/all/20221118071849.25506-2-srinivas.kandagatla@linaro.org/ You're better at lore than me... thx anyway... Neil > > > Best regards, > Krzysztof >
diff --git a/arch/arm64/boot/dts/qcom/sm8450.dtsi b/arch/arm64/boot/dts/qcom/sm8450.dtsi index bcb51e612261..ef9bae2e6acc 100644 --- a/arch/arm64/boot/dts/qcom/sm8450.dtsi +++ b/arch/arm64/boot/dts/qcom/sm8450.dtsi @@ -2321,9 +2321,8 @@ vamacro: codec@33f0000 { reg = <0 0x033f0000 0 0x1000>; clocks = <&q6prmcc LPASS_CLK_ID_TX_CORE_MCLK LPASS_CLK_ATTRIBUTE_COUPLE_NO>, <&q6prmcc LPASS_HW_MACRO_VOTE LPASS_CLK_ATTRIBUTE_COUPLE_NO>, - <&q6prmcc LPASS_HW_DCODEC_VOTE LPASS_CLK_ATTRIBUTE_COUPLE_NO>, - <&q6prmcc LPASS_CLK_ID_RX_CORE_MCLK2_2X_MCLK LPASS_CLK_ATTRIBUTE_COUPLE_NO>; - clock-names = "mclk", "macro", "dcodec", "npl"; + <&q6prmcc LPASS_HW_DCODEC_VOTE LPASS_CLK_ATTRIBUTE_COUPLE_NO>; + clock-names = "mclk", "macro", "dcodec"; assigned-clocks = <&q6prmcc LPASS_CLK_ID_TX_CORE_MCLK LPASS_CLK_ATTRIBUTE_COUPLE_NO>; assigned-clock-rates = <19200000>;
Fixes the following DT bindings check error: codec@33f0000: clocks: [[137, 57, 1], [137, 102, 1], [137, 103, 1], [137, 70, 1]] is too long codec@33f0000: clock-names: 'oneOf' conditional failed, one must be fixed: ['mclk', 'macro', 'dcodec', 'npl'] is too long The implementation was checked and this npl clock isn't used for the VA macro. Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org> --- arch/arm64/boot/dts/qcom/sm8450.dtsi | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)