| Message ID | 20200429143002.5050-2-t-kristo@ti.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ARM: omap4/5: crypto support fixes | expand |
* Tero Kristo <t-kristo@ti.com> [200429 14:31]: > The node name for these should be clk instead of clock. Otherwise the > clock driver won't be able to map the parent/child relationships > properly, and large number of clocks end up in orphaned state. > > Signed-off-by: Tero Kristo <t-kristo@ti.com> > --- > arch/arm/boot/dts/omap44xx-clocks.dtsi | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/arm/boot/dts/omap44xx-clocks.dtsi b/arch/arm/boot/dts/omap44xx-clocks.dtsi > index 532868591107..b82b1ca0e557 100644 > --- a/arch/arm/boot/dts/omap44xx-clocks.dtsi > +++ b/arch/arm/boot/dts/omap44xx-clocks.dtsi > @@ -1279,13 +1279,13 @@ > #size-cells = <1>; > ranges = <0 0x1400 0x200>; > > - l4_per_clkctrl: clock@20 { > + l4_per_clkctrl: clk@20 { > compatible = "ti,clkctrl-l4-per", "ti,clkctrl"; > reg = <0x20 0x144>; > #clock-cells = <2>; > }; > > - l4_secure_clkctrl: clock@1a0 { > + l4_secure_clkctrl: clk@1a0 { > compatible = "ti,clkctrl-l4-secure", "ti,clkctrl"; > reg = <0x1a0 0x3c>; > #clock-cells = <2>; Heh this is no longer needed since commit 6c3090520554 ("clk: ti: clkctrl: Fix hidden dependency to node name") that added support for using the compatible name :) Maybe you are using some older tree? Or else there's still something wrong somewhere. Regards, Tony
On 30/04/2020 01:07, Tony Lindgren wrote: > * Tero Kristo <t-kristo@ti.com> [200429 14:31]: >> The node name for these should be clk instead of clock. Otherwise the >> clock driver won't be able to map the parent/child relationships >> properly, and large number of clocks end up in orphaned state. >> >> Signed-off-by: Tero Kristo <t-kristo@ti.com> >> --- >> arch/arm/boot/dts/omap44xx-clocks.dtsi | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/arch/arm/boot/dts/omap44xx-clocks.dtsi b/arch/arm/boot/dts/omap44xx-clocks.dtsi >> index 532868591107..b82b1ca0e557 100644 >> --- a/arch/arm/boot/dts/omap44xx-clocks.dtsi >> +++ b/arch/arm/boot/dts/omap44xx-clocks.dtsi >> @@ -1279,13 +1279,13 @@ >> #size-cells = <1>; >> ranges = <0 0x1400 0x200>; >> >> - l4_per_clkctrl: clock@20 { >> + l4_per_clkctrl: clk@20 { >> compatible = "ti,clkctrl-l4-per", "ti,clkctrl"; >> reg = <0x20 0x144>; >> #clock-cells = <2>; >> }; >> >> - l4_secure_clkctrl: clock@1a0 { >> + l4_secure_clkctrl: clk@1a0 { >> compatible = "ti,clkctrl-l4-secure", "ti,clkctrl"; >> reg = <0x1a0 0x3c>; >> #clock-cells = <2>; > > Heh this is no longer needed since commit 6c3090520554 > ("clk: ti: clkctrl: Fix hidden dependency to node name") > that added support for using the compatible name :) > > Maybe you are using some older tree? Or else there's > still something wrong somewhere. I was using 5.7-rc1 as baseline so can't be that. Let me try to check this one again. -Tero -- Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
On 30/04/2020 07:55, Tero Kristo wrote: > On 30/04/2020 01:07, Tony Lindgren wrote: >> * Tero Kristo <t-kristo@ti.com> [200429 14:31]: >>> The node name for these should be clk instead of clock. Otherwise the >>> clock driver won't be able to map the parent/child relationships >>> properly, and large number of clocks end up in orphaned state. >>> >>> Signed-off-by: Tero Kristo <t-kristo@ti.com> >>> --- >>> arch/arm/boot/dts/omap44xx-clocks.dtsi | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> diff --git a/arch/arm/boot/dts/omap44xx-clocks.dtsi >>> b/arch/arm/boot/dts/omap44xx-clocks.dtsi >>> index 532868591107..b82b1ca0e557 100644 >>> --- a/arch/arm/boot/dts/omap44xx-clocks.dtsi >>> +++ b/arch/arm/boot/dts/omap44xx-clocks.dtsi >>> @@ -1279,13 +1279,13 @@ >>> #size-cells = <1>; >>> ranges = <0 0x1400 0x200>; >>> - l4_per_clkctrl: clock@20 { >>> + l4_per_clkctrl: clk@20 { >>> compatible = "ti,clkctrl-l4-per", "ti,clkctrl"; >>> reg = <0x20 0x144>; >>> #clock-cells = <2>; >>> }; >>> - l4_secure_clkctrl: clock@1a0 { >>> + l4_secure_clkctrl: clk@1a0 { >>> compatible = "ti,clkctrl-l4-secure", "ti,clkctrl"; >>> reg = <0x1a0 0x3c>; >>> #clock-cells = <2>; >> >> Heh this is no longer needed since commit 6c3090520554 >> ("clk: ti: clkctrl: Fix hidden dependency to node name") >> that added support for using the compatible name :) >> >> Maybe you are using some older tree? Or else there's >> still something wrong somewhere. > > I was using 5.7-rc1 as baseline so can't be that. > > Let me try to check this one again. Ok you can ignore this and patch #2 for omap5 for similar case. It seems like the patch 6c3090520554 actually forgot to fix the subclock names, and it causes issues in mixed clock node setup. Will post a fix against the clock driver shortly. -Tero -- Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
* Tero Kristo <t-kristo@ti.com> [200430 08:35]: > On 30/04/2020 07:55, Tero Kristo wrote: > > On 30/04/2020 01:07, Tony Lindgren wrote: > > > Heh this is no longer needed since commit 6c3090520554 > > > ("clk: ti: clkctrl: Fix hidden dependency to node name") > > > that added support for using the compatible name :) > > > > > > Maybe you are using some older tree? Or else there's > > > still something wrong somewhere. > > > > I was using 5.7-rc1 as baseline so can't be that. > > > > Let me try to check this one again. > > Ok you can ignore this and patch #2 for omap5 for similar case. It seems > like the patch 6c3090520554 actually forgot to fix the subclock names, and > it causes issues in mixed clock node setup. Will post a fix against the > clock driver shortly. OK good to hear. Thanks, Tony
* Tony Lindgren <tony@atomide.com> [200430 20:26]: > * Tero Kristo <t-kristo@ti.com> [200430 08:35]: > > On 30/04/2020 07:55, Tero Kristo wrote: > > > On 30/04/2020 01:07, Tony Lindgren wrote: > > > > Heh this is no longer needed since commit 6c3090520554 > > > > ("clk: ti: clkctrl: Fix hidden dependency to node name") > > > > that added support for using the compatible name :) > > > > > > > > Maybe you are using some older tree? Or else there's > > > > still something wrong somewhere. > > > > > > I was using 5.7-rc1 as baseline so can't be that. > > > > > > Let me try to check this one again. > > > > Ok you can ignore this and patch #2 for omap5 for similar case. It seems > > like the patch 6c3090520554 actually forgot to fix the subclock names, and > > it causes issues in mixed clock node setup. Will post a fix against the > > clock driver shortly. > > OK good to hear. Applying all but the first two patches into omap-for-v5.8/dt thanks. Tony
diff --git a/arch/arm/boot/dts/omap44xx-clocks.dtsi b/arch/arm/boot/dts/omap44xx-clocks.dtsi index 532868591107..b82b1ca0e557 100644 --- a/arch/arm/boot/dts/omap44xx-clocks.dtsi +++ b/arch/arm/boot/dts/omap44xx-clocks.dtsi @@ -1279,13 +1279,13 @@ #size-cells = <1>; ranges = <0 0x1400 0x200>; - l4_per_clkctrl: clock@20 { + l4_per_clkctrl: clk@20 { compatible = "ti,clkctrl-l4-per", "ti,clkctrl"; reg = <0x20 0x144>; #clock-cells = <2>; }; - l4_secure_clkctrl: clock@1a0 { + l4_secure_clkctrl: clk@1a0 { compatible = "ti,clkctrl-l4-secure", "ti,clkctrl"; reg = <0x1a0 0x3c>; #clock-cells = <2>;
The node name for these should be clk instead of clock. Otherwise the clock driver won't be able to map the parent/child relationships properly, and large number of clocks end up in orphaned state. Signed-off-by: Tero Kristo <t-kristo@ti.com> --- arch/arm/boot/dts/omap44xx-clocks.dtsi | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)