Re: [PATCH v5 2/2] dt-bindings: media: rockchip-vpu: Add rk3588 vdpu121 compatible string

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2502 bytes --]

On Thu, Apr 18, 2024 at 11:19:56AM +0200, Diederik de Haas wrote:
> On Saturday, 13 April 2024 17:57:09 CEST Jianfeng Liu wrote:
> > I'm sorry for my unkonwing about the kernel patching process. And I'm
> > sorry to let maintainers do extra work. Thank you for teaching me this.
> > I will do this right in future patches.
> > 
> > I did received a Acked-by tag from Conor in v4:
> > Acked-by: Conor Dooley <conor.dooley@microchip.com>
> > 
> > I note it here in case someone forgets this tag.
> 
> I think it's beneficial to send a v6 with the following changes:
> 1) Make this dt-bindings patch the first in the series
> 2) Make sure you've collected all the tags you've received to all the patches
> 3) Specify the base commit
> 
> ad 1) I don't know if it's a hard rule, but I've seen a consistent pattern 
> where the dt-binding changes come before those changes being applied to 
> DeviceTree files. It also makes sense as when the dt-binding change hasn't been 
> applied, then the DT file is technically invalid.

It is definitely preferred, since there is tooling that checks for
undocumented compatibles etc that would see spurious errors during
bisection, were that to be done. Generally I wouldn't suggest resending
for the order though if it were the only thing amiss.

> ad 2) You shouldn't make maintainers do extra work to get your patch(es) 
> merged; you want to make their work as easy as possible. Thus you do the 
> (extra) work and provide a new version of the patch(es).
> Sending multiple versions in a single day is generally not recommended as you 
> should give reviewers some time to do the review. But it should be fine now as 
> several days have past without new reviews.

I dunno, the best way to save our time is to not omit the tags in the
first place (or give a reason as to why you did) as we'll likely pull up
the previous version of the series to see if all comments were
addressed, if made by another maintainer (at least I do that and Krzysztof
must have here). In my workflow sending another ack takes much less time
than looking up previous versions and checking to see if things were
dealt with - probably that's in-part to me lacking automation for dfn:
lore search from mutt though..

> 
> ad 3) The `git format-patch` command has a `--base=<commit>` parameter with 
> which you can make explicit upon which commit the patch is based.
> That works a lot better/easier then a textual description.
> 
> HTH



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]