Re: [RFC PATCH 02/12] drm/dep: Add DRM dependency queue layer

[Matthew Brost <matthew.brost@intel.com>]

On Tue, Mar 17, 2026 at 08:41:24PM +0100, Miguel Ojeda wrote:
> On Tue, Mar 17, 2026 at 9:27 AM Matthew Brost <matthew.brost@intel.com> wrote:
> >
> > I hate cut off in thteads.
> >
> > I get it — you’re a Rust zealot.
> 
> Cut off? Zealot?
> 

I appologize here I shouldn't type when I get annoyed. This the 2nd
comment that pointing out difference between C and Rust which really
wasn't direction I have hoping this thread would take.

> Look, I got the email in my inbox, so I skimmed it to understand why I
> got it and why the Rust list was Cc'd. I happened to notice your
> (quite surprising) claims about Rust, so I decided to reply to a
> couple of those, since I proposed Rust for the kernel.
> 

Again my mistake.

> How is that a cut off and how does that make a maintainer a zealot?
> 
> Anyway, my understanding is that we agreed that the cleanup attribute
> in C doesn't enforce much of anything. We also agreed that it is
> important to think about ownership and lifetimes and to enforce the
> rules and to be disciplined. All good so far.
> 
> Now, what I said is simply that Rust fundamentally improves the
> situation -- C "RAII" not doing so is not comparable. For instance,
> that statically enforcing things is a meaningful improvement over
> runtime approaches (which generally require to trigger an issue, and
> which in some cases are not suitable for production settings).
> 

I agree the static checking in Rust is a very nice feature.

> Really, I just said Rust would help with things you already stated you
> care about. And nobody claims "Rust solves everything" as you stated.
> So I don't see zealots here, and insulting others doesn't help your
> argument.

I know, appologize.

Matt

> 
> Cheers,
> Miguel