[PATCH v2] linux: Enforce zero fill of siginfo_t

[enh]

On Tue, Jan 27, 2026 at 11:01 AM Yury Khrustalev
<yury.khrustalev@arm.com> wrote:
>
> On Tue, Jan 27, 2026 at 04:38:56PM +0100, Andreas Schwab wrote:
> > On Jan 27 2026, Yury Khrustalev wrote:
> >
> > > I agree that memset is probably better here. It's unknown where the
> > > struct definition might come from. If a kernel header is used somehow,
> > > pad is not first in there:
> >
> > The layout of the union doesn't matter.
> >
>
> Yes, mea culpa.
>
> A compiler would probably call memset anyway (given the size of
> 128 bytes). Using `{0}` does look neat, but then this test is
> susceptible to compiler bugs.

all code is susceptible to compiler bugs :-)

> Using memset explicitly should be
> a good option for a test?
>
> Thanks,
> Yury
>