On Thu, Dec 17, 2020 at 5:36 AM Jakub Jelinek via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> On Thu, Dec 17, 2020 at 08:25:12AM -0500, Nathan Sidwell wrote:
> > > Yeah. If it is meant as an optimization barrier, shouldn't it be just
> > > __asm__ volatile ("");
> > > or
> > > __asm__ volatile ("" : : : "memory");
> > > ? That said, shouldn't that be guarded on the compiler being GCC
> > > (or some other that supports inline asm)? Not all compilers have to
> > > support
> > > that.
> >
> > I can just take it out -- the reason it is there is so when you set a
> > breapoint on HCF (via a .gdbinit), gdb will land you at one of those two
> > lines, which is a great way of telling you where you are.
>
> So perhaps use something more portable, like:
> volatile int v = 0; v++;
> on which GDB should also put the breakpoint.
> The problem with nops is that not all targets have a simple nop instruction,
> as can be seen in gcc/testsuite/gcc.dg/nop.h:
> #if defined (__ia64__) || defined (__s390__) || defined (__s390x__)
> #define NOP "nop 0"
> #elif defined (__MMIX__)
> #define NOP "swym 0"
> #elif defined (__or1k__)
> #define NOP "l.nop"
> #else
> #define NOP "nop"
> #endif
> But another one is that non-GCC system compilers might not handle the inline
> asm
> at all.
Binutils 2.36 supports ".nop".
--
H.J.
