Re: [PATCH] mm/damon: introduce DAMON-based NUMA memory tiering module

[Herbert <objecting@objecting.org>]

Ll> >>
> >
> >
> > Hello david, thanks for being polite about this whole thing!
> >
> >
> > so, you are talking to a real human, i promise you that now
>
> That's exactly, what I would say if I were a bot ;)

Well, I don't know what else to say. I have matrix and irc I guess, I can prove I'm human

> As a first step, fix the line wrapping. The following document was
> helpful to me in the past:
>
>     Documentation/process/email-clients.rst

Yeah.. mobile clients hurt me, I tried thunderbird and it didn't come with word wrapping (I was a bit angry at that), I am mostly away from my pc, (the place I do patches!)


> >
> >
> > I am looking for people to personally mentor me, so i succseed in the kernel
> >
> > And, I know this is unprofessinal (and im sorry!), but my github is my best
> friend, i have a year worth of commits for you to see publicly, also i work
> with android custom roms. (Any questions, feel free to ask!)
> >
> >
> >
> > So yes, this is my oppotunity to be honest,
> >
> >
> > (Yes. I do use AI!)
> >
> >
> > The reason why, is im a "perfectionist", i want to start human patches, but
> im scared i will get ripped apart, (or i wont get like reviewership!! Because
> im too slow) so i use claude (smartest ai in the market), and i regret it.
> Like seriously!
>
> There is this nice saying "we learn from our mistakes".
>
> I still remember the first time Linus ripped apart one of my patches. I
> learned something that day.


Exactly! Mistakes happen, you learn from them, (I was stupid for breaking my break...)

> >
> > Additionally, i hate being slow, im scared that the community will ditch me
> if i do 1 human patch every 2-4 weeks, which is my expected time (because i
> need to understand the code!!!!!)
> >
>
>
> Well, doing 8 versions of one patch set a day is obviously the other
> extreme ;) For example, there is this unwritten rule to wait ~1 week
> (and before discussions slowed down) before sending another revision.
> There are exceptions of course.

Yeah, was stupid of me ngl

> In general, we're lacking reviewers, not code contributors. So we really
> need people that can get familiar with the code and gain expertise, to
> help the project long term.

I sure can review code. I do review code for people who ask for my custom ROM stuff, basically they need kernel trees for it to build and stuff basically.


> Drive-by AI generated patches that mostly just consume more of the
> precious maintainer+reviewer time are rather counter-productive.

Yeah. Very stupid of me


> >
> > I would love to have a mentor to help me around, and when (if) somebody
> wants to, i promise you now, i will start making human patches, and i will
> benefit the community positively,
> >
>
> I think you had some patches accepted my maintainers, you might want to
> consider reaching out to them to see if they would be willing to help.

To be fair, Masami is extremely busy, so is seong, Andrew is "Stupidly busy" (his words), which is understandable because he maintains lib/ and some of mm/


> But be aware that most people are extremely busy.


Same point ^

> --
> Cheers,
>
> David




V/R




Josh Law