Jeremy Fitzhardinge wrote: > Zachary Amsden wrote: > >> Jeremy Fitzhardinge wrote: >> >>> Fix a few clobbers to include the return register. The clobbers set >>> is the set of all registers modified (or may be modified) by the code >>> snippet, regardless of whether it was deliberate or accidental. >>> >>> Signed-off-by: Jeremy Fitzhardinge <jeremy at xensource.com> >>> Cc: Rusty Russell <rusty at rustcorp.com.au> >>> Cc: Zachary Amsden <zach at vmware.com> >>> >>> --- >>> include/asm-i386/paravirt.h | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> =================================================================== >>> --- a/include/asm-i386/paravirt.h >>> +++ b/include/asm-i386/paravirt.h >>> @@ -556,7 +556,7 @@ static inline unsigned long __raw_local_ >>> "popl %%edx; popl %%ecx") >>> : "=a"(f): "m"(paravirt_ops.save_fl), >>> paravirt_type(PARAVIRT_PATCH(save_fl)), >>> - paravirt_clobber(CLBR_NONE) >>> + paravirt_clobber(CLBR_EAX) >>> : "memory", "cc"); >>> return f; >>> >>> >> Has this been tested on older gcc's? I seem to recall them barfing >> over things like this. >> > > Things like what? Do you mean the %[foo] asm parameter syntax? I think > those versions are no longer supported - Arjan posted a patch a few days > ago to convert a pile of asms to this form. Or do you mean something else? > I meant having an output in the clobber list, I didn't know we were dropping support for older versions already. Zach
