|
From: <sv...@va...> - 2005-05-05 12:04:23
|
Author: sewardj
Date: 2005-05-05 13:04:14 +0100 (Thu, 05 May 2005)
New Revision: 1160
Modified:
trunk/priv/guest-x86/toIR.c
Log:
Update comment.
Modified: trunk/priv/guest-x86/toIR.c
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- trunk/priv/guest-x86/toIR.c 2005-05-03 12:20:15 UTC (rev 1159)
+++ trunk/priv/guest-x86/toIR.c 2005-05-05 12:04:14 UTC (rev 1160)
@@ -34,8 +34,12 @@
*/
=20
/* TODO:
- SBB reg with itself (copy from amd64 front end)
=20
+ Check the following. It afflicts the amd64 front end, not sure if
+ applicable here: MOVQ (sse) is wrong wrt is the upper half zeroed
+ or not? It always should be if dst is a reg; not quite the same as
+ MOVSD.
+
check flag settings for cmpxchg
FUCOMI(P): what happens to A and S flags? Currently are forced
to zero.
@@ -77,8 +81,7 @@
The delta values are 32-bit ints, not 64-bit ints. That means
this module may not work right if run on a 64-bit host. That should
be fixed properly, really -- if anyone ever wants to use Vex to
- translate x86 code for execution on a 64-bit host.
-*/
+ translate x86 code for execution on a 64-bit host. */
=20
/* Performance holes:
=20
|