2018-03-07 12:50+0100, Paolo Bonzini: > On 06/03/2018 21:50, Radim Krčmář wrote: > > GCC 8 emitted MOVDQU when compiling vmx.elf, but we do not enable > > CR4.OSFXSR in that test, so the instruction throws #UD. > > This patch forbids all sse2 instructions, instead of enabling > > CR4.OSFXSR, as I think it's better to keep the environment minimal. > > > > Signed-off-by: Radim Krčmář <rkrcmar@xxxxxxxxxx> > > --- > > This probably hit a minor KVM bug: the instruction was "f3 0f 6f 41 28", > > but KVM reported that #UD happened on "f3 0f 6f 41". > > That's probably just because "no opcode" does not decode mod/rm/sib. Good point, moving the EmulateOnUD check down gets the correct result. I'll send a patch as doing more decoding before dying isn't harmful. > > --- > > x86/Makefile.x86_64 | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/x86/Makefile.x86_64 b/x86/Makefile.x86_64 > > index 623fc5b37726..6caa3a8863f1 100644 > > --- a/x86/Makefile.x86_64 > > +++ b/x86/Makefile.x86_64 > > @@ -1,7 +1,7 @@ > > cstart.o = $(TEST_DIR)/cstart64.o > > bits = 64 > > ldarch = elf64-x86-64 > > -COMMON_CFLAGS += -mno-red-zone > > +COMMON_CFLAGS += -mno-red-zone -mno-sse2 > > Why not -mno-sse? We have asm some code with the "x" constraint and it wouldn't compile without -msse. Should I rewrite that to be explicit? Thanks.
