From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lee Jones Subject: Re: [PATCH v4 2/2] dt: power: st: Provide bindings for ST's OPPs Date: Tue, 28 Jul 2015 08:34:39 +0100 Message-ID: <20150728073439.GK21114@x1> References: <1438010430-5802-1-git-send-email-lee.jones@linaro.org> <1438010430-5802-2-git-send-email-lee.jones@linaro.org> <20150728022936.GB1229@linux> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: <20150728022936.GB1229@linux> Sender: linux-pm-owner@vger.kernel.org To: Viresh Kumar Cc: rob.herring@linaro.org, nm@ti.com, arnd.bergmann@linaro.org, sboyd@codeaurora.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, kernel@stlinux.com, rjw@rjwysocki.net, linux-pm@vger.kernel.org, devicetree@vger.kernel.org, ajitpal.singh@st.com, sre@kernel.org, dbaryshkov@gmail.com List-Id: devicetree@vger.kernel.org On Tue, 28 Jul 2015, Viresh Kumar wrote: > Cc'ing few people (whom I cc'd last time as well :)). >=20 > On 27-07-15, 16:20, Lee Jones wrote: > > These OPPs are used in ST's CPUFreq implementation. > >=20 > > Signed-off-by: Lee Jones > > --- > >=20 > > Changelog: > > - None, new patch > >=20 > > Documentation/devicetree/bindings/power/opp-st.txt | 76 ++++++++++= ++++++++++++ > > 1 file changed, 76 insertions(+) > > create mode 100644 Documentation/devicetree/bindings/power/opp-st.= txt > >=20 > > diff --git a/Documentation/devicetree/bindings/power/opp-st.txt b/D= ocumentation/devicetree/bindings/power/opp-st.txt > > new file mode 100644 > > index 0000000..6eb2a91 > > --- /dev/null > > +++ b/Documentation/devicetree/bindings/power/opp-st.txt > > @@ -0,0 +1,76 @@ > > +STMicroelectronics OPP (Operating Performance Points) Bindings > > +-------------------------------------------------------------- > > + > > +Frequency Scaling only > > +---------------------- > > + > > +Located in CPU's node: > > + > > +- operating-points : [See: ./opp.txt] > > + > > +Example [safe] > > +-------------- > > + > > +cpus { > > + cpu@0 { > > + /* kHz uV */ > > + operating-points =3D <1500000 0 > > + 1200000 0 > > + 800000 0 > > + 500000 0>; > > + }; > > +}; > > + > > +Dynamic Voltage and Frequency Scaling (DVFS) > > +-------------------------------------------- > > + > > +Located in 'cpu0-opp-list' node [to be provided ONLY by the bootlo= ader]: > > + > > +- compatible : Should be "operating-points-v2-sti" > > +- opp{1..N} : Each 'oppX' subnode will contain the following pro= perties: >=20 > Or should we mention: > - opp{1..N} : Each 'oppX' subnode shall contain below properties, > over what ./opp.txt defines: >=20 > ? I disagree. For one, only 'opp-hz' is defined in ./opp.tx. Secondly it would be annoying to have to have to keep jumping between documents to obtain the whole picture. Finally, generic bindings are repeated in platform/device specific documentation all the time. Grep for 'clocks' or 'regulator-* or 'interrupts' or 'reg' or 'clock-frequency' (which IMHO I think you should have used instead of 'opp-hz', but that's by the by), or any number of other generic properties. > > + - opp-hz : CPU frequency [Hz] for this OPP [See: ./opp.txt] > > + - st,avs : List of available voltages [uV] indexed by process co= de > > + - st,cuts : Cut version this OPP is suitable for [0xFF means ALL= ] > > + - st,substrate : Substrate version this OPP is suitable for [0xF= =46 means ALL] > > +- st,syscfg : Phandle to Major number register > > + First cell: offset to major number > > +- st,syscfg-eng : Phandle to Minor number and Pcode registers > > + First cell: offset to process code > > + Second cell: offset to minor number > > + > > +WARNING: The opp{1..N} nodes will be provided by the bootloader. = Do not attempt to > > + artificially synthesise the opp{1..N} nodes or any of their desc= endants. > > + They are very platform specific and may damage the hardware if c= reated > > + incorrectly. > > + > > +Example [unsafe] > > +---------------- > > + > > +cpus { > > + cpu@0 { > > + operating-points-v2 =3D <&cpu0_opp_list>; > > + }; > > +}; > > + > > +/* ############################################################ */ > > +/* # WARNING: Do not attempt to copy/replicate this node, # */ > > +/* # it is only to be supplied by the bootloader !!! # */ > > +/* ############################################################ */ > > +cpu0-opp-list { > > + compatible =3D "operating-points-v2-sti"; > > + st,syscfg =3D <&syscfg [major_offset]>; > > + st,syscfg-eng =3D <&syscfg_eng [pcode_offset] [minor_offset]>; > > + > > + opp0 { > > + opp-hz =3D <1200000000>; > > + st,avs =3D <1110 1150 1100 1080 1040 1020 980 930>; > > + st,substrate =3D <0xff>; > > + st,cuts =3D <0xff>; > > + }; > > + opp1 { > > + opp-hz =3D <1500000000>; > > + st,avs =3D <1200 1200 1200 1200 1170 1140 1100 1070>; > > + st,substrate =3D <0xff>; > > + st,cuts =3D <0x2>; > > + }; > > +}; >=20 > I don't see more problems here, unless we can move some of this to th= e > generic bindings. >=20 > @Rob/Stephen: Please respond before it is late :) No one knows this stuff better than you. If you can't think of an already existing binding that could suit to portray our 'cuts' and 'substrate' information (with a similar way to support our "all cuts" and "all substrates" options, then there probably isn't one. ;) --=20 Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org =E2=94=82 Open source software for ARM SoCs =46ollow Linaro: Facebook | Twitter | Blog