https://bugs.kde.org/show_bug.cgi?id=374596
Bug ID: 374596
Summary: inconsistent RDTSCP support on x86_64
Product: valgrind
Version: 3.12.0
Platform: RedHat RPMs
OS: Linux
Status: UNCONFIRMED
Severity: normal
Priority: NOR
Component: vex
Assignee: jsew...@acm.org
Reporter: bugzi...@dcopp.net
Target Milestone: ---
Created attachment 103212
--> https://bugs.kde.org/attachment.cgi?id=103212&action=edit
g++ source for testing RDTSCP support
The attached test program attempts to determine support for, and then use,
RDTSCP instruction. On CPU that does not support RDTSCP is correctly reports
that the instruction is not supported and does not execute it. Otherwise it
does, without error.
Under valgrind, on a CPU that does not support RDTSCP the opcode is reported as
unsupported even though the program never executes it:
vex amd64->IR: unhandled instruction bytes: 0xF 0x1 0xF9 0xBE 0xD8 0x9 0x40 0x0
0xBF 0x80
Strangely (due to VEX simulating a different CPU stepping) if we comment out
the code that executes RDTCSP, the program under valgrind then reports the
instruction as being supported.
Expected behavior: under valgrind on a CPU that does not support RDTSCP the
program should not crash. valgrind (vex) should simulate the instruction
successfully since it advertises support for it.
--
You are receiving this mail because:
You are watching all bug changes.
