Bug 126256

Summary:	(fnop) vex x86->IR: unhandled instruction bytes: 0xD9 0xD0 0x31 0xC0
Product:	[Developer tools] valgrind	Reporter:	Thomas Kühne <thomas-dloop>
Component:	vex	Assignee:	Julian Seward <jseward>
Status:	RESOLVED FIXED
Severity:	crash	CC:	mark, mark, vince
Priority:	NOR
Version First Reported In:	3.2 SVN
Target Milestone:	---
Platform:	Compiled Sources
OS:	Linux
Latest Commit:		Version Fixed/Implemented In:
Sentry Crash Report:
Bug Depends on:
Bug Blocks:	256630
Attachments:	Add support for fnop instruction on x86 and amd64

Description Thomas Kühne 2006-04-25 23:41:44 UTC

3:   e8 fc ff ff ff          call   4 <_Dmain+0x4>
   8:   d9 d0                   fnop   
   a:   5d                      pop    ebp

(http://dstress.kuehne.cn/run/a/asm_fnop_01.d)

Comment 1 Mark 2026-01-15 18:40:23 UTC

I have a similar case that I reported on the users list together with a proposed fix (see https://sourceforge.net/p/valgrind/mailman/message/59283753/). The attached patch has a test for reproducing the issue and a fix for x86 and amd64 (that should also address bug 253446).

Comment 2 Mark 2026-01-15 18:41:36 UTC

Created attachment 188561 [details]
Add support for fnop instruction on x86 and amd64

Comment 3 Mark Wielaard 2026-01-16 12:08:00 UTC

*** Bug 253446 has been marked as a duplicate of this bug. ***

Comment 4 Mark Wielaard 2026-01-16 12:11:26 UTC

(In reply to Mark from comment #2)
> Created attachment 188561 [details]
> Add support for fnop instruction on x86 and amd64

Thanks, this looks good. Apologies this was never "implemented".
At first I was afraid this needed some subtle floating point exception signaling.
Which technically it does, but that is currently a known limitation according to the TODO
at the top of the file
- x87 FP Limitations:
  * no FP exceptions, except for handling stack over/underflow

So this seems good to go, thanks for the testcases.

Comment 5 Mark Wielaard 2026-01-17 16:59:09 UTC

commit a615e4daebe2b7289beb66019493285cfc58c2bc
Author: Mark <mark@zbuffer.de>
Date:   Sat Jan 17 17:22:59 2026 +0100

    Add support for fnop instruction on x86 and amd64
    
    This implements FNOP in guest_amd64_toIR.c and guest_x86_toIR.c by
    simply skipping the instruction. Which is OK for now since one of the
    x87 FP Limitations is no FP exceptions, except for handling stack
    over/underflow.
    
    Add fnop to none/tests/{amd64,x86}/insn_fpu.{def,stdout.exp}.
    
    Also update old 3_1_BUGSTATUS.txt to correct bug number.
    
    https://bugs.kde.org/show_bug.cgi?id=126256