On Fri, Aug 30, 2019 at 9:25 PM Stephen Boyd <swboyd@xxxxxxxxxxxx> wrote: > > Quoting Amit Kucheria (2019-08-30 04:32:54) > > On Thu, Aug 29, 2019 at 10:04 PM Amit Kucheria <amit.kucheria@xxxxxxxxxx> wrote: > > > > > > On Thu, Aug 29, 2019 at 8:23 PM Stephen Boyd <swboyd@xxxxxxxxxxxx> wrote: > > > > > > > > Can we get a known quantity of interrupts for a particular compatible > > > > string instead? Let's be as specific as possible. The index matters too, > > > > so please list them in the order that is desired. > > > > > > I *think* we can predict what platforms have uplow and critical > > > interrupts based on IP version currently[1]. For newer interrupt > > > types, we might need more fine-grained platform compatibles. > > > > > > [1] Caveat: this is based only on the list of platforms I've currently > > > looked at, there might be something internally that breaks these > > > rules. > > > > What do you think if we changed the wording to something like the following, > > > > - interrupt-names: Must be one of the following depending on IP version: > > For compatibles qcom,msm8916-tsens, qcom,msm8974-tsens, > > qcom,qcs404-tsens, qcom,tsens-v1, use > > interrupt-names = "uplow"; > > For compatibles qcom,msm8996-tsens, qcom,msm8998-tsens, > > qcom,sdm845-tsens, qcom,tsens-v2, use > > interrupt-names = "uplow", "critical"; > > Ok. I would still prefer YAML/JSON schema for this binding so that it's > much more explicit about numbers and the order of interrupts, etc. OK, I'll look around for some examples.
