Re: [PATCH] Warn about "unsigned value that used to be signed against zero"

[Vincent Mailhol <mailhol@kernel.org>]

On 24/09/2025 at 16:07, Chris Li wrote:
> On Mon, Sep 22, 2025 at 9:17 AM Linus Torvalds
> <torvalds@linux-foundation.org> wrote:
>>
>> On Mon, 22 Sept 2025 at 05:10, Chris Li <sparse@chrisli.org> wrote:
>>>
>>> On Sun, Sep 21, 2025 at 8:16 AM Linus Torvalds
>>> <torvalds@linux-foundation.org> wrote:
>>>>
>>>> That's why I had added that
>>>>
>>>>         info(def->pos, "signed value source");
>>>
>>> Do you want such a line in the final patch as well? Seems worth it.
>>
>> Oh, absolutely. Some of the cases it reported were really hard to
>> understand without it.
>>
>> If it ends up being noisy, there could possibly be some heuristic like
>> "if the source is very close to the use that triggers it, don't bother
>> talking about it", but that would be a later tweak.
> 
> Vincent, can you add the above info() line suggested by Linus as
> incremental or full patches?

I think that there is a small confusion here. That line is already in the patch
inside simplify_unsigned_zero_compare(). I think that Linus was just explaining
why he added it in the original patch.


Yours sincerely,
Vincent Mailhol