126389 2006-04-28 03:29:49 +0000 (fxrstr) vex x86->IR: unhandled instruction bytes: 0xF 0xAE 0x4D 0xF0 2011-08-11 10:09:11 +0000 1 1 6 Developer tools valgrind vex 3.2 SVN Compiled Sources Linux RESOLVED FIXED NOR normal --- 1 thomas-dloop jseward m.b.lankhorst nilesh_asher 0 oldest_to_newest 434637 0 thomas-dloop 2006-04-28 03:29:49 +0000 89 55 f4 mov DWORD PTR [ebp-12],edx 0f ae 4d f0 fxrstor DWORD PTR [ebp-16] 31 c0 xor eax,eax (http://dstress.kuehne.cn/norun/a/asm_fxrstor_01.d) 579758 1 dank 2008-02-12 03:05:53 +0000 I can confirm this. As of three weeks ago, Wine now uses this instruction; see http://www.winehq.org/pipermail/wine-cvs/2008-January/039635.html and today's valgrind from svn still doesn't support it. To reproduce, just run the wine test suite under valgrind. To pull the interesting bit out of the log, do egrep 'asm|vex' log This shows a bunch of instances of vex x86->IR: unhandled instruction bytes: 0xF 0xAE 0xA 0xE9 701 __asm__ __volatile__( "fxrstor %0" : : "m" (*state) ); 579802 2 jseward 2008-02-12 11:12:34 +0000 > vex x86->IR: unhandled instruction bytes: 0xF 0xAE 0xA 0xE9 > 701 __asm__ __volatile__( "fxrstor %0" : : "m" (*state) ); Hmm, this got punted because it's rarely used and a pain to implement. I'll try to get to it in the next couple of weeks. 583405 3 dank 2008-02-28 15:35:23 +0000 This is more important than it sounds. The instruction is used in the exception handler in Wine. That means that any app that tends to use exceptions in normal operation, like oh say Adobe InDesign, can't be valground. Which is a shame. Can you at least confirm the bug? :-) 583409 4 jseward 2008-02-28 16:01:40 +0000 Yes. It's a real bug :-) 583412 5 tom 2008-02-28 16:24:13 +0000 *** Bug 158525 has been marked as a duplicate of this bug. *** 599930 6 jseward 2008-05-09 15:27:36 +0000 Fixed (vex r1836). 1152537 7 tom 2011-08-11 10:09:11 +0000 *** Bug 149838 has been marked as a duplicate of this bug. ***