There are two cases where this was wrong, albeit in a benign way (the
compiler - according to my checking - didn't leverage the wrongness
for any optimizations affecting overall outcome).
Signed-off-by: Jan Beulich
--- a/xen/arch/x86/x86_emulate/x86_emulate.c
+++ b/xen/arch/x86/x86_emulate/x86_emulate.c
@@ -3683,7 +3683,7 @@ x86_emulate(
if ( (rc = ops->read(src.mem.seg, src.mem.off,
, src.bytes, ctxt)) != 0 )
goto done;
-emulate_fpu_insn_memdst("fldt", src.val);
+emulate_fpu_insn_memsrc("fldt", src.val);
break;
case 7: /* fstp m80fp */
ea.bytes = 10;
@@ -3780,7 +3780,7 @@ x86_emulate(
ea.bytes = 8;
dst = ea;
dst.type = OP_MEM;
-emulate_fpu_insn_memsrc("fstl", dst.val);
+emulate_fpu_insn_memdst("fstl", dst.val);
break;
case 3: /* fstp m64fp */
ea.bytes = 8;
x86emul: correct direction of FPU insn emulations
There are two cases where this was wrong, albeit in a benign way (the
compiler - according to my checking - didn't leverage the wrongness
for any optimizations affecting overall outcome).
Signed-off-by: Jan Beulich
--- a/xen/arch/x86/x86_emulate/x86_emulate.c
+++ b/xen/arch/x86/x86_emulate/x86_emulate.c
@@ -3683,7 +3683,7 @@ x86_emulate(
if ( (rc = ops->read(src.mem.seg, src.mem.off,
, src.bytes, ctxt)) != 0 )
goto done;
-emulate_fpu_insn_memdst("fldt", src.val);
+emulate_fpu_insn_memsrc("fldt", src.val);
break;
case 7: /* fstp m80fp */
ea.bytes = 10;
@@ -3780,7 +3780,7 @@ x86_emulate(
ea.bytes = 8;
dst = ea;
dst.type = OP_MEM;
-emulate_fpu_insn_memsrc("fstl", dst.val);
+emulate_fpu_insn_memdst("fstl", dst.val);
break;
case 3: /* fstp m64fp */
ea.bytes = 8;
___
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel