[PATCH 1/2] RISC-V: Update l2 cache DT documentation to add support for SiFive FU740

[Rob Herring]

On Mon, Nov 23, 2020 at 3:32 AM Yash Shah <yash.shah at openfive.com> wrote:
>
> > -----Original Message-----
> > From: Rob Herring <robh at kernel.org>
> > Sent: 21 November 2020 18:25
> > To: Yash Shah <yash.shah at openfive.com>
> > Cc: Paul Walmsley ( Sifive) <paul.walmsley at sifive.com>;
> > palmer at dabbelt.com; aou at eecs.berkeley.edu;
> > Jonathan.Cameron at huawei.com; wsa at kernel.org; sam at ravnborg.org;
> > Sagar Kadam <sagar.kadam at openfive.com>; anup at brainfault.org;
> > bp at suse.de; devicetree at vger.kernel.org; linux-riscv at lists.infradead.org;
> > linux-kernel at vger.kernel.org; Sachin Ghadi <sachin.ghadi at openfive.com>
> > Subject: Re: [PATCH 1/2] RISC-V: Update l2 cache DT documentation to add
> > support for SiFive FU740
> >
> > [External Email] Do not click links or attachments unless you recognize the
> > sender and know the content is safe
> >
> > On Thu, Nov 12, 2020 at 02:41:13PM +0530, Yash Shah wrote:
> > > The L2 cache controller in SiFive FU740 has 4 ECC interrupt sources as
> > > compared to 3 in FU540. Update the DT documentation accordingly with
> > > "compatible" and "interrupt" property changes.
> > >
> > > Signed-off-by: Yash Shah <yash.shah at sifive.com>
> > > ---
> > >  .../devicetree/bindings/riscv/sifive-l2-cache.yaml | 33
> > > +++++++++++++++++-----
> > >  1 file changed, 26 insertions(+), 7 deletions(-)
> > >
> > > diff --git
> > > a/Documentation/devicetree/bindings/riscv/sifive-l2-cache.yaml
> > > b/Documentation/devicetree/bindings/riscv/sifive-l2-cache.yaml
> > > index efc0198..4873d5c 100644
> > > --- a/Documentation/devicetree/bindings/riscv/sifive-l2-cache.yaml
> > > +++ b/Documentation/devicetree/bindings/riscv/sifive-l2-cache.yaml
>
> <...>
>
> > > @@ -51,12 +54,6 @@ properties:
> > >
> > >    cache-unified: true
> > >
> > > -  interrupts:
> > > -    description: |
> > > -      Must contain entries for DirError, DataError and DataFail signals.
> > > -    minItems: 3
> > > -    maxItems: 3
> >
> > Keep this here and just change maxItems to 4. Really, what each interrupt is
> > should be listed out as an 'items' entry.
> >
>
> Sure will send a v2 with the above modifications.
>
> <...>
>
> >
> > > +
> > > +else:
> > > +  properties:
> > > +    interrupts:
> > > +      description: |
> > > +        Must contain entries for DirError, DirFail, DataError, DataFail signals.
> >
> > DirFail should be last so you keep the same indices.
>
> Actually, the interrupts have been numbered like that in FU740 SoCs and the driver expects the interrupts to be in this order.
> I will keep the same order for v2 as well. Let me know if you still disagree.

Numbered within the cache block or the interrupt controller? If the
former, then fine. The latter would be outside the scope of the
binding. Another SoC could hook up interrupts differently.

It's going to be easier for the driver to deal with 1 new irq index
rather than 2 whole sets of different indices, but if you want to do
it the hard way...

Rob