Re: [PATCH] [RFC] xfrm: work around a clang-19 fortifiy-string false-positive

[Arnd Bergmann]

On Tue, Apr 9, 2024, at 18:15, Nathan Chancellor wrote:
> On Mon, Apr 08, 2024 at 09:06:21AM +0200, Arnd Bergmann wrote:
>> >
>> > The shorter fix (in the issue) is to explicitly range-check before
>> > the loop:
>> >
>> >        if (xp->xfrm_nr > XFRM_MAX_DEPTH)
>> >                return -ENOBUFS;
>> 
>> I ran into this issue again and I see that Nathan's fix has
>> made it into mainline and backports, but it's apparently
>> not sufficient.
>> 
>> I don't see the warning with my patch from this thread, but
>> there may still be a better fix.
>
> Is it the exact same warning? clang-19 or older?
> What > architecture/configuration? If my change is not sufficient then maybe
> there are two separate issues here? I have not seen this warning appear
> in our CI since my change was applied.

I only see it with clang-19. I've never seen it with arm32 and
currently only see it with arm64, though I had seen it with x86-64
as well in February before your patch.

The warning is the same as before aside from the line number,
which which is now include/linux/fortify-string.h:462:4
where it was line 420, but I think that is just a context
change.

I have a number of configs that reproduce this bug, see
https://pastebin.com/tMgfD7cu for an example with current
linux-next.

     Arnd