On Tue, Apr 23, 2019 at 4:00 PM Nick Desaulniers <ndesaulni...@google.com> wrote: > > On Tue, Apr 23, 2019 at 3:57 PM Ard Biesheuvel > <ard.biesheu...@linaro.org> wrote: > > > > On Wed, 24 Apr 2019 at 00:55, Kees Cook <keesc...@chromium.org> wrote: > > > > > > On Mon, Apr 15, 2019 at 7:38 AM Kees Cook <keesc...@chromium.org> wrote: > > > > +#define __msr_s(r, v) \ > > > > + DEFINE_MSR_S \ > > > > +" msr_s " __stringify(r) ", %x0\n" \ > > > > + UNDEFINE_MSR_S : : "rZ" (v) > > > > > > BTW ... is "%x0" a typo here? Shouldn't this just be "%0"? > > > > > > > No, iirc, that is required to force clang to choose a x# encoding > > (which is the only one supported by msr/mrs instructions). Otherwise, > > it will infer from the operand size which encoding to use, and attempt > > to use w# for 32-bit system registers. > > Yes: > https://github.com/llvm/llvm-project/blob/80b578c7325352187d0aff7bd2bdb5a76b51cf37/llvm/lib/Target/AArch64/AArch64AsmPrinter.cpp#L466
Cool, yeah, I see now. Thanks! -- Kees Cook
