On Sat, May 31, 2025 at 12:23:04PM +0200, Peter Zijlstra wrote:
On Fri, May 30, 2025 at 10:48:47AM -0700, Kees Cook wrote:
On Fri, May 30, 2025 at 04:01:40PM +0200, Peter Zijlstra wrote:
I'm not really concerned with performance here, but more with the size of the code emitted by WARN_ONCE(). There are a *ton* of WARN sites, while only one report_bug() and printk().
The really offensive thing is that this is for a feature most nobody will ever need :/
Well, it won't be enabled often -- this reminds me of ftrace: it needs to work, but it'll be off most of the time.
Well, ftrace is useful, but when would I *ever* care about this stuff? I can't operate kunit
Why not?
don't give a crap about kunit
That's your choice, of course, and it might not be useful to you anyway, but it's *really* nice and closed a major gap in testing in some other areas.
I'd still encourage you to try it, it might be worth your time.
and if I were to magically run it, I would be more than capable of ignoring WARNs.
Yeah, it's not just about ignoring WARNs, but mostly about knowing which ones you can ignore, and which ones you should fix.
We're getting at a point (on some subsystems I guess) where we actually have a decent testing suite we can ask contributors to run and have all tests passing.
We also want to ask them to fix whatever issue they might introduce :)
Thanks for your help on getting a cleaner solution!
Maxime