> Especially if it is -O0 only, I don't see why you think so. Just dg-skip-if
> it for -O1+ if you believe it is unreliable for some reason, but if you
> look at the parameter value after the prologue, not showing the right value
> at -O0 would be a serious bug everywhere. Having some GDB testcas
On Mon, Jun 17, 2013 at 09:46:30AM +0200, Eric Botcazou wrote:
> > This really should come with testcases (e.g. guality ones).
>
> guality testcases are barely maintainable, especially on non-x86 platforms,
> so
> I'd rather not enter this business. I'll discuss with Joel and see whether
> we
> This really should come with testcases (e.g. guality ones).
guality testcases are barely maintainable, especially on non-x86 platforms, so
I'd rather not enter this business. I'll discuss with Joel and see whether we
can coordinate with the GDB side.
--
Eric Botcazou
On Sun, Jun 16, 2013 at 12:51:25PM +0200, Eric Botcazou wrote:
> the subject is slightly misleading since it's actually about structures with
> integral modes which are passed by reference on some platforms, e.g. SPARC or
> PowerPC 32-bit. There are 3 issues:
>
> 1. At -O0, the parameters of t
Hi,
the subject is slightly misleading since it's actually about structures with
integral modes which are passed by reference on some platforms, e.g. SPARC or
PowerPC 32-bit. There are 3 issues:
1. At -O0, the parameters of this kind are not homed on the stack, which
means that backtraces ca
