Skip to contentSkip to conversation

Docs for base maintainer expectations?

DISCUSSION Jakub Kicinski 3 participants 6 messages Jul 10, 2023 – Jul 11, 2023

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi,

do we have any docs describing what's expected from folks stepping up to maintain (small-ish) parts of the kernel like a driver or a protocol?

Experienced developers / maintainers differ like the beautiful snowflakes that we are, but outsiders have much less familiarity with the landscape, and frankly sometimes much less interest in participating once they code lands.

Which makes we wonder if a simple list of responsibilities would be useful as a baseline. I haven't spotted anything in Docs/process but perhaps someone has a local version for their subsystem?

#1 lore raw

Hey Jakub,

Quoted 8 lines · Jakub Kicinski

do we have any docs describing what's expected from folks stepping up to maintain (small-ish) parts of the kernel like a driver or a protocol?

Experienced developers / maintainers differ like the beautiful snowflakes that we are, but outsiders have much less familiarity with the landscape, and frankly sometimes much less interest in participating once they code lands.

Quoted 5 lines

Which makes we wonder if a simple list of responsibilities would be useful as a baseline.

I haven't spotted anything in Docs/process but perhaps someone has a local version for their subsystem?

Given I figure you did this on with a -rc1 based tree, which would mean that what I wrote probably does not fit the bill, but I tried to do something along these lines with
https://docs.kernel.org/process/maintainer-soc.html
for which my target audience was people picking up maintenance of DT/soc drivers, which I hope there'll be a few of in RISC-V land soon...

I suggested adding things to it, like putting the trees in linux-next etc, but review feedback suggested that was unsuited to a subsystem specific document.

Thanks,
Conor.

#2 lore raw
Quoted 22 lines · On Mon, 10 Jul 2023 20:06:56 +0100 Conor Dooley wrote:
Quoted 8 lines · Jakub Kicinski

do we have any docs describing what's expected from folks stepping up to maintain (small-ish) parts of the kernel like a driver or a protocol?

Experienced developers / maintainers differ like the beautiful snowflakes that we are, but outsiders have much less familiarity with the landscape, and frankly sometimes much less interest in participating once they code lands.

Quoted 5 lines

Which makes we wonder if a simple list of responsibilities would be useful as a baseline.

I haven't spotted anything in Docs/process but perhaps someone has a local version for their subsystem?

Given I figure you did this on with a -rc1 based tree, which would mean that what I wrote probably does not fit the bill, but I tried to do something along these lines with
https://docs.kernel.org/process/maintainer-soc.html
for which my target audience was people picking up maintenance of DT/soc drivers, which I hope there'll be a few of in RISC-V land soon...

Quoted 3 lines

I suggested adding things to it, like putting the trees in linux-next etc, but review feedback suggested that was unsuited to a subsystem specific document.

Thanks for the pointer! I haven't read it before because I assumed it'll describe workflow and suggestions for SoC sub-tree (which it does?) I was thinking about something way more basic, like "you are expected to reply to bug reports", "you are expected to investigate syzbot problems" etc. :(

IOW the SoC guide reads more like "how to get your code accepted" rather than "what are the externally-facing responsibilities of a maintainer".

#3 lore raw
Quoted 23 lines · Jakub Kicinski
Quoted 22 lines · On Mon, 10 Jul 2023 20:06:56 +0100 Conor Dooley wrote:
Quoted 8 lines · Jakub Kicinski

do we have any docs describing what's expected from folks stepping up to maintain (small-ish) parts of the kernel like a driver or a protocol?

Experienced developers / maintainers differ like the beautiful snowflakes that we are, but outsiders have much less familiarity with the landscape, and frankly sometimes much less interest in participating once they code lands.

Quoted 5 lines

Which makes we wonder if a simple list of responsibilities would be useful as a baseline.

I haven't spotted anything in Docs/process but perhaps someone has a local version for their subsystem?

Given I figure you did this on with a -rc1 based tree, which would mean that what I wrote probably does not fit the bill, but I tried to do something along these lines with
https://docs.kernel.org/process/maintainer-soc.html
for which my target audience was people picking up maintenance of DT/soc drivers, which I hope there'll be a few of in RISC-V land soon...

Quoted 7 lines
Quoted 3 lines

I suggested adding things to it, like putting the trees in linux-next etc, but review feedback suggested that was unsuited to a subsystem specific document.

Thanks for the pointer! I haven't read it before because I assumed it'll describe workflow and suggestions for SoC sub-tree (which it does?)

You did mention "local version for the subsystem", but maybe by local you meant on their own box, rather than local to the subsystem. I interpreted it as the latter, sorry about that!

Quoted 3 lines

I was thinking about something way more basic, like "you are expected to reply to bug reports", "you are expected to investigate syzbot problems" etc. :(

Yeah, along the lines of the generic stuff it was suggested that I should not add :)

Quoted 2 lines

IOW the SoC guide reads more like "how to get your code accepted" rather than "what are the externally-facing responsibilities of a maintainer".

I suppose it is a bit of a mix, since the document _is_ aimed at maintainers of platform support (there's details about sending PRs etc), but it does contain some conventions that "code" should follow, which they should in turn pass down to contributors.

It sounds almost as if you're looking for something to present to people when they add themselves as an M: entry against a driver, for when $company sends a patch w/ someone you've not really seen before?

#4 lore raw
Quoted 3 lines · On Mon, 10 Jul 2023 20:39:40 +0100 Conor Dooley wrote:

It sounds almost as if you're looking for something to present to people when they add themselves as an M: entry against a driver, for when $company sends a patch w/ someone you've not really seen before?

That is exactly the use I have in mind.

maintainer
#5 lore raw
In reply to Jakub Kicinski
Quoted 11 lines · Jakub Kicinski

do we have any docs describing what's expected from folks stepping up to maintain (small-ish) parts of the kernel like a driver or a protocol?

Experienced developers / maintainers differ like the beautiful snowflakes that we are, but outsiders have much less familiarity with the landscape, and frankly sometimes much less interest in participating once they code lands.

Which makes we wonder if a simple list of responsibilities would be useful as a baseline. I haven't spotted anything in Docs/process but perhaps someone has a local version for their subsystem?

I think that would be good to have. We often complain about a lack of maintainers, but we have nothing to point to telling people how they can step up and do that.

I've tried to fill in some pieces with the documents on merging and such, but we're missing the introduction.

Thanks,

jon