| Message ID | 2d52388bd7d3cc546ac3ab5afeb47bfcb3012213.1446167359.git.viresh.kumar@linaro.org |
|---|---|
| State | Superseded, archived |
| Headers | show |
On 10/30, Viresh Kumar wrote: > +- opp-supported-hw: User defined array containing a hierarchy of hardware > + version numbers, supported by the OPP. For example: a platform with hierarchy > + of three levels of versions (A, B and C), this field should be like <X Y Z>, > + where X corresponds to Version hierarchy A, Y corresponds to version hierarchy > + B and Z corresponds to version hierarchy C. > + > + Each level of hierarchy is represented by a 32 bit value, and so there can be > + only 32 different supported version per hierarchy. i.e. 1 bit per version. A > + value of 0xFFFFFFFF will enable the OPP for all versions for that hierarchy > + level. And a value of 0x00000000 will disable the OPP completely, and so we > + never want that to happen. I suppose if you wanted to have 64 possible combinations of some attribute you would just extend it to two 32 bit numbers in sequence? I don't see the limitation here, and hopefully there isn't a limitation so that we can specify sufficiently large numbers with more bits if we need to.
On 10/30, Viresh Kumar wrote: > + opp_table { > + compatible = "operating-points-v2"; > + status = "okay"; > + opp-shared; > + > + opp00 { A side-note. I wonder if it would be better style to have the node name be: opp@600000000 { At least it seems that the assumption is we can store all the possible combinations of OPP values for a particular frequency in the same node. Following this style would make dt compilation fail if two nodes have the same frequency. Also, this makes it sound like opp-supported-hw is really just telling us if this is a supported frequency or not for the particular device we're running on. The current wording makes it sound like we could have two OPP nodes with the same frequency but different voltages inside them, which we're trying to discourage by compressing the tables into less nodes. I think in Lee's case we're only going to use the cuts parameter to figure out if the OPP is supported or not. On qcom platforms we will only use one parameter for this property as well, the speed bin. The slow/fast and version stuff will be handled by named opp properties. > + /* > + * Supports all substrate and process versions for 0xF > + * cuts, i.e. only first four cuts. > + */ > + opp-supported-hw = <0xF 0xFFFFFFFF 0xFFFFFFFF> > + opp-hz = /bits/ 64 <600000000>; > + opp-microvolt = <900000 915000 925000>;
On 30-10-15, 14:49, Stephen Boyd wrote: > I suppose if you wanted to have 64 possible combinations of some > attribute you would just extend it to two 32 bit numbers in > sequence? I don't see the limitation here, and hopefully there > isn't a limitation so that we can specify sufficiently large > numbers with more bits if we need to. Yeah, we discussed this earlier when Lee had the same query and I suggested the exact same thing to him then.
On 30-10-15, 15:18, Stephen Boyd wrote: > A side-note. I wonder if it would be better style to have the > node name be: > > opp@600000000 { I thought the @... had a special meaning and we might end up creating some device for the node then? Perhaps I am mistaken. But then, yeah it will make it more readable as you mentioned. > At least it seems that the assumption is we can store all the > possible combinations of OPP values for a particular frequency in > the same node. Following this style would make dt compilation > fail if two nodes have the same frequency. Right. > Also, this makes it sound like opp-supported-hw is really just > telling us if this is a supported frequency or not for the > particular device we're running on. That's right. > The current wording makes it Of the commit log ? Or the way the nodes are written? > sound like we could have two OPP nodes with the same frequency > but different voltages inside them, which we're trying to > discourage by compressing the tables into less nodes. No no, we can't have two nodes with same frequency.
On Fri, Oct 30, 2015 at 9:20 PM, Viresh Kumar <viresh.kumar@linaro.org> wrote: > On 30-10-15, 15:18, Stephen Boyd wrote: >> A side-note. I wonder if it would be better style to have the >> node name be: >> >> opp@600000000 { > > I thought the @... had a special meaning and we might end up creating > some device for the node then? Perhaps I am mistaken. There is no special meaning, just convention which is the unit-address should match the reg property address. I'm okay with an exception here. Rob -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 30-10-15, 14:49, Stephen Boyd wrote: > I suppose if you wanted to have 64 possible combinations of some > attribute you would just extend it to two 32 bit numbers in > sequence? I don't see the limitation here, and hopefully there > isn't a limitation so that we can specify sufficiently large > numbers with more bits if we need to. I hope you want to mark this patch with your reviewed-by tag?
On 02-11-15, 09:13, Rob Herring wrote: > There is no special meaning, just convention which is the unit-address > should match the reg property address. I'm okay with an exception > here. Thanks, I will update this separately.
On 10/31, Viresh Kumar wrote: > On 30-10-15, 15:18, Stephen Boyd wrote: > > > Also, this makes it sound like opp-supported-hw is really just > > telling us if this is a supported frequency or not for the > > particular device we're running on. > > That's right. > > > The current wording makes it > > Of the commit log ? Or the way the nodes are written? The way the documentation is written.
On 10/31, Viresh Kumar wrote: > On 30-10-15, 14:49, Stephen Boyd wrote: > > I suppose if you wanted to have 64 possible combinations of some > > attribute you would just extend it to two 32 bit numbers in > > sequence? I don't see the limitation here, and hopefully there > > isn't a limitation so that we can specify sufficiently large > > numbers with more bits if we need to. > > Yeah, we discussed this earlier when Lee had the same query and I > suggested the exact same thing to him then. > Ah I see that after looking at the previous thread. Perhaps we can add such information into the documentation so that people aren't misled into thinking they're limited to 32 bits?
diff --git a/Documentation/devicetree/bindings/opp/opp.txt b/Documentation/devicetree/bindings/opp/opp.txt index 0cb44dc21f97..96892057586a 100644 --- a/Documentation/devicetree/bindings/opp/opp.txt +++ b/Documentation/devicetree/bindings/opp/opp.txt @@ -123,6 +123,18 @@ properties. - opp-suspend: Marks the OPP to be used during device suspend. Only one OPP in the table should have this. +- opp-supported-hw: User defined array containing a hierarchy of hardware + version numbers, supported by the OPP. For example: a platform with hierarchy + of three levels of versions (A, B and C), this field should be like <X Y Z>, + where X corresponds to Version hierarchy A, Y corresponds to version hierarchy + B and Z corresponds to version hierarchy C. + + Each level of hierarchy is represented by a 32 bit value, and so there can be + only 32 different supported version per hierarchy. i.e. 1 bit per version. A + value of 0xFFFFFFFF will enable the OPP for all versions for that hierarchy + level. And a value of 0x00000000 will disable the OPP completely, and so we + never want that to happen. + - status: Marks the node enabled/disabled. Example 1: Single cluster Dual-core ARM cortex A9, switch DVFS states together. @@ -463,3 +475,48 @@ Example 5: Multiple OPP tables }; }; }; + +Example 6: opp-supported-hw +(example: three level hierarchy of versions: cuts, substrate and process) + +/ { + cpus { + cpu@0 { + compatible = "arm,cortex-a7"; + ... + + cpu-supply = <&cpu_supply> + operating-points-v2 = <&cpu0_opp_table_slow>; + }; + }; + + opp_table { + compatible = "operating-points-v2"; + status = "okay"; + opp-shared; + + opp00 { + /* + * Supports all substrate and process versions for 0xF + * cuts, i.e. only first four cuts. + */ + opp-supported-hw = <0xF 0xFFFFFFFF 0xFFFFFFFF> + opp-hz = /bits/ 64 <600000000>; + opp-microvolt = <900000 915000 925000>; + ... + }; + + opp01 { + /* + * Supports: + * - cuts: only one, 6th cut (represented by 6th bit). + * - substrate: supports 16 different substrate versions + * - process: supports 9 different process versions + */ + opp-supported-hw = <0x20 0xff0000ff 0x0000f4f0> + opp-hz = /bits/ 64 <800000000>; + opp-microvolt = <900000 915000 925000>; + ... + }; + }; +};
We may want to enable only a subset of OPPs, from the bigger list of OPPs, based on what version of the hardware we are running on. This would enable us to not duplicate OPP tables for every version of the hardware we support. To enable that, this patch defines a new property 'opp-supported-hw'. It can support any number of hierarchy levels of the versions the hardware follows. And based on the selected hardware versions, we can pick only the relevant OPPs at runtime. Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> --- Documentation/devicetree/bindings/opp/opp.txt | 57 +++++++++++++++++++++++++++ 1 file changed, 57 insertions(+)