On Sat, Jun 28, 2025 at 3:04 PM Onur work@onurozkan.dev wrote:
Yeah, I meant that it't taking more effort than it should, like digging through historical changes in the relevant parts of the source code, trying to figuring out whether it was just a false positive or if there was a specific reason behind it, etc.
Yeah, that is a big part of kernel development, especially on the maintenance side :)
I definitely agree that a good comment in the source code is better than going through Git history, and the kernel sometimes has had some things documented in the Git log that should have been in the source code instead. It happens.
However, in some cases like this one it is not clear it would help. For instance, here the lint reason message could have been something that made sense back then when the lint was enabled, and yet we would still have had to notice the lint got disabled later on, so we would end up still going into the Git log.
`expect` is great to mitigate some of that -- sadly we cannot use it as much as we would like due to sometimes being conditional to an arch or the kernel config or the Rust version. (And your first patch may have some cases that perhaps we cannot convert due to that -- I didn't check)
Cheers, Miguel