lkml.org 
[lkml]   [2022]   [Nov]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 4/9] dt-bindings: Add RISC-V incoming MSI controller bindings
On Wed, Nov 16, 2022 at 04:04:45PM +0530, Anup Patel wrote:
> On Wed, Nov 16, 2022 at 2:30 PM Krzysztof Kozlowski
> <krzysztof.kozlowski@linaro.org> wrote:
> >
> > On 15/11/2022 23:34, Conor Dooley wrote:
> > > On Mon, Nov 14, 2022 at 05:59:00PM +0530, Anup Patel wrote:
> > >> On Sun, Nov 13, 2022 at 8:18 PM Conor Dooley <conor@kernel.org> wrote:
> > >
> > >>> Also, the file name says "riscv,imsic", the description says "IMSIC" but
> > >>> you've used "imsics" in the compatible. Is this a typo, or a plural?
> > >>
> > >> Yes, the file name should be consistent. I will update the file name.
> > >
> > > Is there a reason why the compatible is plural when all of the other
> > > mentions etc do not have an "s"? It really did look like a typo to me.
> > >
> > > It's the "incoming MSI controller", so I am unsure as to where the "s"
> > > actually even comes from. Why not just use "riscv,imsic"?
> >
> > Yep, should be rather consistent with all others, and IMSIC stands for
> > Integrated Circuit?
>
> This is intentionally plural because even though we have one
> IMSIC per-CPU, Linux (and various OSes) expect one DT node
> as MSI controller targeting all CPUs.

Even still, calling it "riscv,imsic" would seem fair to me given the
multiple regs make the distinct regions clear.

I think I must have missed the bit at the end of the description though:

+ The device tree of a RISC-V platform will have one IMSIC device tree node
+ for each privilege level (machine or supervisor) which collectively describe
+ IMSIC interrupt files at that privilege level across CPUs (or HARTs).

Perhaps, for eejits like me, that paragraph should become paragraph 3
instead of hiding it below the register layout etc?

Anyways, existing name seems fine to me then w/ the filename update &
increased prominence of the many-controllers-in-one statement.

Maybe the devicetree gods think differently!

> The plural compatible string "riscv,imsics" was chosen based
> on consensus on RISC-V AIA Task Group meetings.

btw, I see the following in the example:

+ reg = <0x28000000 0x2000>, /* Group0 IMSICs */
+ <0x29000000 0x2000>; /* Group1 IMSICs */

And in the property:
+ reg:
+ minItems: 1
+ maxItems: 128
+ description:
+ Base address of each IMSIC group.

It would appear that the comment there conflicts with the description of
the reg property itself & it's that lack of consistency me confused (:
Should the comments be in the singular form?

Thanks,
Conor.

\
 
 \ /
  Last update: 2022-11-16 14:31    [W:0.105 / U:0.124 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site