Re: [PATCH for 7.2?] target/i386: Remove compilation errors when -Werror=maybe-uninitialized

[Stefan Hajnoczi]

On Wed, 7 Dec 2022 at 14:11, Stefan Weil <sw@weilnetz.de> wrote:
>
> Am 07.12.22 um 19:22 schrieb Eric Auger:
> >
> > On 12/7/22 17:55, Philippe Mathieu-Daudé wrote:
> >> On 7/12/22 15:33, Eric Auger wrote:
> >>> On 12/7/22 15:09, Stefan Hajnoczi wrote:
> >>>> On Wed, 7 Dec 2022 at 08:31, Eric Auger <eric.auger@redhat.com> wrote:
> >>>>> On 12/7/22 14:24, Eric Auger wrote:
> >>>>>> Initialize r0-3 to avoid compilation errors when
> >>>>>> -Werror=maybe-uninitialized is used
> >>>>>>
> >>>>>> ../target/i386/ops_sse.h: In function ‘helper_vpermdq_ymm’:
> >>>>>> ../target/i386/ops_sse.h:2495:13: error: ‘r3’ may be used
> >>>>>> uninitialized in this function [-Werror=maybe-uninitialized]
> >>>>>>    2495 |     d->Q(3) = r3;
> >>>>>>         |     ~~~~~~~~^~~~
> >>>>>> ../target/i386/ops_sse.h:2494:13: error: ‘r2’ may be used
> >>>>>> uninitialized in this function [-Werror=maybe-uninitialized]
> >>>>>>    2494 |     d->Q(2) = r2;
> >>>>>>         |     ~~~~~~~~^~~~
> >>>>>> ../target/i386/ops_sse.h:2493:13: error: ‘r1’ may be used
> >>>>>> uninitialized in this function [-Werror=maybe-uninitialized]
> >>>>>>    2493 |     d->Q(1) = r1;
> >>>>>>         |     ~~~~~~~~^~~~
> >>>>>> ../target/i386/ops_sse.h:2492:13: error: ‘r0’ may be used
> >>>>>> uninitialized in this function [-Werror=maybe-uninitialized]
> >>>>>>    2492 |     d->Q(0) = r0;
> >>>>>>         |     ~~~~~~~~^~~~
> >>>>>>
> >>>>>> Signed-off-by: Eric Auger <eric.auger@redhat.com>
> >>>>>> Fixes: 790684776861 ("target/i386: reimplement 0x0f 0x3a, add AVX")
> >>>>>>
> >>>>>> ---
> >>>>>>
> >>>>>> Am I the only one getting this? Or anything wrong in my setup.
>
> Hi Eric,
>
> no, you are not the only one. I regularly build with higher warning
> levels, for example with -Weverything on macOS, and get a much longer
> list which includes the mentioned warnings (see below).
>
> The warnings for ops_sse.h are false positives, so I think no fix is
> needed for 7.2. The compiler is not clever enough to see that the switch
> statements handle all possible cases. It should be sufficient to replace
> `case 3` by `default` to help the compiler and fix the warning. Your fix
> might produce new compiler warnings because setting the variables to 0
> has no effect.

Eric: do you want to try the switch "default" approach for 8.0?

Thanks,
Stefan