| Message ID | 20161215182754.33921-1-icenowy@aosc.xyz (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, Dec 16, 2016 at 02:27:54AM +0800, Icenowy Zheng wrote: > An operating point table is needed for the cpu frequency adjusting to > work. > > The operating point table is converted from the common value in > extracted script.fex from many A33 board/tablets. > > Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> > --- > Changes since v1: > - Fix format problem (blank lines). > - Removed the 1.344GHz operating point, as it's overvoltage and overclocked. > > This patch depends on the following patchset: > > http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/473962.html > > It's the v2 of the [PATCH 4/6] in this patchset. > > I think this operating point table may also apply to A23, as there's no > difference except the points over 1.2GHz between A23 and A33's stock dvfs table. > > But as A23 CCU may not have the necessary fixes, I won't add the table to A23 > now. > > Chen-Yu, could you test the CCU fixes I described in the patchset above on A23, > then test this operating points table? > > If it's necessary, you can send out the CCU fixes and add one more patch that > moves this opp-v2 table to sun8i-a23-a33.dtsi . > > arch/arm/boot/dts/sun8i-a33.dtsi | 35 +++++++++++++++++++++++++++++++++++ > 1 file changed, 35 insertions(+) > > diff --git a/arch/arm/boot/dts/sun8i-a33.dtsi b/arch/arm/boot/dts/sun8i-a33.dtsi > index 504996cbee29..0f5b2af72981 100644 > --- a/arch/arm/boot/dts/sun8i-a33.dtsi > +++ b/arch/arm/boot/dts/sun8i-a33.dtsi > @@ -46,7 +46,42 @@ > #include <dt-bindings/dma/sun4i-a10.h> > > / { > + cpu0_opp_table: opp_table0 { > + compatible = "operating-points-v2"; > + opp-shared; > + > + opp@648000000 { > + opp-hz = /bits/ 64 <648000000>; > + opp-microvolt = <1040000>; > + clock-latency-ns = <244144>; /* 8 32k periods */ > + }; > + > + opp@816000000 { > + opp-hz = /bits/ 64 <816000000>; > + opp-microvolt = <1100000>; > + clock-latency-ns = <244144>; /* 8 32k periods */ > + }; > + > + opp@1008000000 { > + opp-hz = /bits/ 64 <1008000000>; > + opp-microvolt = <1200000>; > + clock-latency-ns = <244144>; /* 8 32k periods */ > + }; > + > + opp@1200000000 { > + opp-hz = /bits/ 64 <1200000000>; > + opp-microvolt = <1320000>; > + clock-latency-ns = <244144>; /* 8 32k periods */ > + }; > + }; > + > cpus { > + cpu0: cpu@0 { There's no need to duplicate the label here. I removed it and applied. Thanks! Maxime
On Mon, Dec 19, 2016 at 4:46 PM, Maxime Ripard <maxime.ripard@free-electrons.com> wrote: > On Fri, Dec 16, 2016 at 02:27:54AM +0800, Icenowy Zheng wrote: >> An operating point table is needed for the cpu frequency adjusting to >> work. >> >> The operating point table is converted from the common value in >> extracted script.fex from many A33 board/tablets. >> >> Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> >> --- >> Changes since v1: >> - Fix format problem (blank lines). >> - Removed the 1.344GHz operating point, as it's overvoltage and overclocked. >> >> This patch depends on the following patchset: >> >> http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/473962.html >> >> It's the v2 of the [PATCH 4/6] in this patchset. >> >> I think this operating point table may also apply to A23, as there's no >> difference except the points over 1.2GHz between A23 and A33's stock dvfs table. >> >> But as A23 CCU may not have the necessary fixes, I won't add the table to A23 >> now. >> >> Chen-Yu, could you test the CCU fixes I described in the patchset above on A23, >> then test this operating points table? >> >> If it's necessary, you can send out the CCU fixes and add one more patch that >> moves this opp-v2 table to sun8i-a23-a33.dtsi . >> >> arch/arm/boot/dts/sun8i-a33.dtsi | 35 +++++++++++++++++++++++++++++++++++ >> 1 file changed, 35 insertions(+) >> >> diff --git a/arch/arm/boot/dts/sun8i-a33.dtsi b/arch/arm/boot/dts/sun8i-a33.dtsi >> index 504996cbee29..0f5b2af72981 100644 >> --- a/arch/arm/boot/dts/sun8i-a33.dtsi >> +++ b/arch/arm/boot/dts/sun8i-a33.dtsi >> @@ -46,7 +46,42 @@ >> #include <dt-bindings/dma/sun4i-a10.h> >> >> / { >> + cpu0_opp_table: opp_table0 { >> + compatible = "operating-points-v2"; >> + opp-shared; >> + >> + opp@648000000 { >> + opp-hz = /bits/ 64 <648000000>; >> + opp-microvolt = <1040000>; >> + clock-latency-ns = <244144>; /* 8 32k periods */ >> + }; >> + >> + opp@816000000 { >> + opp-hz = /bits/ 64 <816000000>; >> + opp-microvolt = <1100000>; >> + clock-latency-ns = <244144>; /* 8 32k periods */ >> + }; >> + >> + opp@1008000000 { >> + opp-hz = /bits/ 64 <1008000000>; >> + opp-microvolt = <1200000>; >> + clock-latency-ns = <244144>; /* 8 32k periods */ >> + }; >> + >> + opp@1200000000 { >> + opp-hz = /bits/ 64 <1200000000>; >> + opp-microvolt = <1320000>; >> + clock-latency-ns = <244144>; /* 8 32k periods */ >> + }; >> + }; >> + >> cpus { >> + cpu0: cpu@0 { > > There's no need to duplicate the label here. I removed it and applied. I think using the label to directly reference cpu0 would be better, instead of duplicating the cpu@0 block. ChenYu
19.12.2016, 16:54, "Chen-Yu Tsai" <wens@csie.org>: > On Mon, Dec 19, 2016 at 4:46 PM, Maxime Ripard > <maxime.ripard@free-electrons.com> wrote: >> On Fri, Dec 16, 2016 at 02:27:54AM +0800, Icenowy Zheng wrote: >>> An operating point table is needed for the cpu frequency adjusting to >>> work. >>> >>> The operating point table is converted from the common value in >>> extracted script.fex from many A33 board/tablets. >>> >>> Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> >>> --- >>> Changes since v1: >>> - Fix format problem (blank lines). >>> - Removed the 1.344GHz operating point, as it's overvoltage and overclocked. >>> >>> This patch depends on the following patchset: >>> >>> http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/473962.html >>> >>> It's the v2 of the [PATCH 4/6] in this patchset. >>> >>> I think this operating point table may also apply to A23, as there's no >>> difference except the points over 1.2GHz between A23 and A33's stock dvfs table. >>> >>> But as A23 CCU may not have the necessary fixes, I won't add the table to A23 >>> now. >>> >>> Chen-Yu, could you test the CCU fixes I described in the patchset above on A23, >>> then test this operating points table? >>> >>> If it's necessary, you can send out the CCU fixes and add one more patch that >>> moves this opp-v2 table to sun8i-a23-a33.dtsi . >>> >>> arch/arm/boot/dts/sun8i-a33.dtsi | 35 +++++++++++++++++++++++++++++++++++ >>> 1 file changed, 35 insertions(+) >>> >>> diff --git a/arch/arm/boot/dts/sun8i-a33.dtsi b/arch/arm/boot/dts/sun8i-a33.dtsi >>> index 504996cbee29..0f5b2af72981 100644 >>> --- a/arch/arm/boot/dts/sun8i-a33.dtsi >>> +++ b/arch/arm/boot/dts/sun8i-a33.dtsi >>> @@ -46,7 +46,42 @@ >>> #include <dt-bindings/dma/sun4i-a10.h> >>> >>> / { >>> + cpu0_opp_table: opp_table0 { >>> + compatible = "operating-points-v2"; >>> + opp-shared; >>> + >>> + opp@648000000 { >>> + opp-hz = /bits/ 64 <648000000>; >>> + opp-microvolt = <1040000>; >>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>> + }; >>> + >>> + opp@816000000 { >>> + opp-hz = /bits/ 64 <816000000>; >>> + opp-microvolt = <1100000>; >>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>> + }; >>> + >>> + opp@1008000000 { >>> + opp-hz = /bits/ 64 <1008000000>; >>> + opp-microvolt = <1200000>; >>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>> + }; >>> + >>> + opp@1200000000 { >>> + opp-hz = /bits/ 64 <1200000000>; >>> + opp-microvolt = <1320000>; >>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>> + }; >>> + }; >>> + >>> cpus { >>> + cpu0: cpu@0 { >> >> There's no need to duplicate the label here. I removed it and applied. > > I think using the label to directly reference cpu0 would be better, > instead of duplicating the cpu@0 block. After proper testing of A23 ccu, the operating points can also apply to A23. (According to A23 devices' fex) > > ChenYu
Hi, On 19/12/2016 15:06, Icenowy Zheng wrote: > > After proper testing of A23 ccu, the operating points can also apply to A23. > (According to A23 devices' fex) > I disagree. I've looked into using the same operating points for A23 and A33 for CPUfreq and I came to the conclusion that not all A23-based boards share the same operating points. You can find the fex files for the different boards here: https://github.com/linux-sunxi/sunxi-boards/tree/master/sys_config/a23 After gathering all supported frequencies/voltages for each A23-based boards, I found that only the following frequencies and voltages are supported by all A23-based boards: - 1008 MHz, 1220 mV - 816 MHz, 1120 mV - 600 MHz, 1040 mV Quentin
On 19/12/2016 15:06, Icenowy Zheng wrote: > > > 19.12.2016, 16:54, "Chen-Yu Tsai" <wens@csie.org>: >> On Mon, Dec 19, 2016 at 4:46 PM, Maxime Ripard >> <maxime.ripard@free-electrons.com> wrote: >>> On Fri, Dec 16, 2016 at 02:27:54AM +0800, Icenowy Zheng wrote: >>>> An operating point table is needed for the cpu frequency adjusting to >>>> work. >>>> >>>> The operating point table is converted from the common value in >>>> extracted script.fex from many A33 board/tablets. >>>> >>>> Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> >>>> --- >>>> Changes since v1: >>>> - Fix format problem (blank lines). >>>> - Removed the 1.344GHz operating point, as it's overvoltage and overclocked. >>>> >>>> This patch depends on the following patchset: >>>> >>>> http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/473962.html >>>> >>>> It's the v2 of the [PATCH 4/6] in this patchset. >>>> >>>> I think this operating point table may also apply to A23, as there's no >>>> difference except the points over 1.2GHz between A23 and A33's stock dvfs table. >>>> >>>> But as A23 CCU may not have the necessary fixes, I won't add the table to A23 >>>> now. >>>> >>>> Chen-Yu, could you test the CCU fixes I described in the patchset above on A23, >>>> then test this operating points table? >>>> >>>> If it's necessary, you can send out the CCU fixes and add one more patch that >>>> moves this opp-v2 table to sun8i-a23-a33.dtsi . >>>> >>>> arch/arm/boot/dts/sun8i-a33.dtsi | 35 +++++++++++++++++++++++++++++++++++ >>>> 1 file changed, 35 insertions(+) >>>> >>>> diff --git a/arch/arm/boot/dts/sun8i-a33.dtsi b/arch/arm/boot/dts/sun8i-a33.dtsi >>>> index 504996cbee29..0f5b2af72981 100644 >>>> --- a/arch/arm/boot/dts/sun8i-a33.dtsi >>>> +++ b/arch/arm/boot/dts/sun8i-a33.dtsi >>>> @@ -46,7 +46,42 @@ >>>> #include <dt-bindings/dma/sun4i-a10.h> >>>> >>>> / { >>>> + cpu0_opp_table: opp_table0 { >>>> + compatible = "operating-points-v2"; >>>> + opp-shared; >>>> + >>>> + opp@648000000 { >>>> + opp-hz = /bits/ 64 <648000000>; >>>> + opp-microvolt = <1040000>; >>>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>>> + }; >>>> + >>>> + opp@816000000 { >>>> + opp-hz = /bits/ 64 <816000000>; >>>> + opp-microvolt = <1100000>; >>>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>>> + }; >>>> + >>>> + opp@1008000000 { >>>> + opp-hz = /bits/ 64 <1008000000>; >>>> + opp-microvolt = <1200000>; >>>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>>> + }; >>>> + >>>> + opp@1200000000 { >>>> + opp-hz = /bits/ 64 <1200000000>; >>>> + opp-microvolt = <1320000>; >>>> + clock-latency-ns = <244144>; /* 8 32k periods */ >>>> + }; >>>> + }; >>>> + Also, there are a lot more operating points for the A33, see: https://github.com/QSchulz/linux/blob/v4.9-rc4_adc_a31_v7/cpufreq_a33/arch/arm/boot/dts/sun8i-a33.dtsi#L323-L340 They are present in the Allwinner Linux source code and in the fex of all A33-based boards. Is there a reason for not adding all opp? Quentin
diff --git a/arch/arm/boot/dts/sun8i-a33.dtsi b/arch/arm/boot/dts/sun8i-a33.dtsi index 504996cbee29..0f5b2af72981 100644 --- a/arch/arm/boot/dts/sun8i-a33.dtsi +++ b/arch/arm/boot/dts/sun8i-a33.dtsi @@ -46,7 +46,42 @@ #include <dt-bindings/dma/sun4i-a10.h> / { + cpu0_opp_table: opp_table0 { + compatible = "operating-points-v2"; + opp-shared; + + opp@648000000 { + opp-hz = /bits/ 64 <648000000>; + opp-microvolt = <1040000>; + clock-latency-ns = <244144>; /* 8 32k periods */ + }; + + opp@816000000 { + opp-hz = /bits/ 64 <816000000>; + opp-microvolt = <1100000>; + clock-latency-ns = <244144>; /* 8 32k periods */ + }; + + opp@1008000000 { + opp-hz = /bits/ 64 <1008000000>; + opp-microvolt = <1200000>; + clock-latency-ns = <244144>; /* 8 32k periods */ + }; + + opp@1200000000 { + opp-hz = /bits/ 64 <1200000000>; + opp-microvolt = <1320000>; + clock-latency-ns = <244144>; /* 8 32k periods */ + }; + }; + cpus { + cpu0: cpu@0 { + clocks = <&ccu CLK_CPUX>; + clock-names = "cpu"; + operating-points-v2 = <&cpu0_opp_table>; + }; + cpu@2 { compatible = "arm,cortex-a7"; device_type = "cpu";
An operating point table is needed for the cpu frequency adjusting to work. The operating point table is converted from the common value in extracted script.fex from many A33 board/tablets. Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz> --- Changes since v1: - Fix format problem (blank lines). - Removed the 1.344GHz operating point, as it's overvoltage and overclocked. This patch depends on the following patchset: http://lists.infradead.org/pipermail/linux-arm-kernel/2016-December/473962.html It's the v2 of the [PATCH 4/6] in this patchset. I think this operating point table may also apply to A23, as there's no difference except the points over 1.2GHz between A23 and A33's stock dvfs table. But as A23 CCU may not have the necessary fixes, I won't add the table to A23 now. Chen-Yu, could you test the CCU fixes I described in the patchset above on A23, then test this operating points table? If it's necessary, you can send out the CCU fixes and add one more patch that moves this opp-v2 table to sun8i-a23-a33.dtsi . arch/arm/boot/dts/sun8i-a33.dtsi | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+)