Messages in this thread | | | Date | Thu, 22 Sep 2022 16:46:15 +0300 | From | Laurent Pinchart <> | Subject | Re: [PATCH v2 2/4] media: dt-bindings: Document Renesas RZ/G2L CRU block |
| |
On Wed, Sep 21, 2022 at 08:58:26PM +0200, Krzysztof Kozlowski wrote: > On 21/09/2022 19:29, Laurent Pinchart wrote: > >>>>> + clock-names: > >>>>> + items: > >>>>> + - const: vclk > >>>>> + - const: pclk > >>>>> + - const: aclk > >>>> > >>>> Drop the "clk" suffixes. Remaining names could be made a bit more readable. > >>> > >>> These names come from the documentation, isn't it better to match the > >>> datasheet ? > >> > >> If datasheet calls it "vclk_really_clk_it_is_clk_clk", it's not the > >> reason to use it. :) > >> > >> The "clk" is redundant even if the hardware engineer thought different. > >> > >> The same for IRQs ("tx" not "txirq"), for dmas ("tx" not "txdma"). > > > > I'd argue that naming clocks "v", "p" and "a" would be less readable and > > more confusing. Is this a new rule ? > > Not really, but also it's only a style issue. > > Indeed "v" and "p" are not much better... but still "vclk" does not > bring any additional information over "v". It's redundant. > > You can also drop entire entry - unless it helps in particular > implementation.
There are multiple clocks, so I think names are better than indices. If you really insist we could name them "v", "p" and "a", but I think the clk suffix here improves readability. If those clocks were named "videoclk", "pixelclk" and "axiclk" I'd be fine dropping the suffix, but that's not the case here.
> https://lore.kernel.org/all/20220517175958.GA1321687-robh@kernel.org/ > https://lore.kernel.org/all/20210815133926.22860-1-biju.das.jz@bp.renesas.com/ > https://lore.kernel.org/all/YYFCaHI%2FDASUz+Vu@robh.at.kernel.org/ > https://lore.kernel.org/all/20220830182540.GA1797396-robh@kernel.org/
-- Regards,
Laurent Pinchart
| |