Bug 150540

Summary:	valgrind segfaults under xen
Product:	[Fedora] Fedora	Reporter:	Robin Green <greenrd>
Component:	xen	Assignee:	Rik van Riel <riel>
Status:	CLOSED WONTFIX	QA Contact:
Severity:	high	Docs Contact:
Priority:	medium
Version:	3	CC:	sct
Target Milestone:	---
Target Release:	---
Hardware:	athlon
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2006-02-24 20:10:04 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Robin Green 2005-03-08 03:04:57 UTC

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.6) Gecko/20050302 Firefox/1.0.1 Fedora/1.0.1-1.3.2

Description of problem:
Running valgrind on a xen0-linux kernel produces consistent segfaults within fractions of a second of valgrind starting.

Version-Release number of selected component (if applicable):
xen-2-20050302

How reproducible:
Always

Steps to Reproduce:
1. Install xen-2-20050302 and kernel-xen0-2.6.11-1.1176_FC4
2. Reboot into xen
3. Run valgrind --tool=memcheck ls
  

Actual Results:  Valgrind shows a segfault in a random place in libc or /bin/ls. (Occassionally, ls completes successfully, but more often than not, it fails with a segfault.)

Expected Results:  No segfault

Additional info:

This bug does not occur running under 2.6.11-1.1176_FC4 (the equivalent non-xeno kernel).

Version of valgrind is valgrind-2.2.0-3
Version of coreutils is coreutils-5.2.1-31

I couldn't reproduce this bug in single-user mode with no services running. However, on the other hand, I could reproduce it consistently in runlevels 3 and 5, and whether xend was started made no difference.

Comment 1 Robin Green 2005-03-08 03:08:11 UTC

This bug also doesn't occur under a xen-testing kernel that I built myself.
(However, I built that kernel as a monolithic kernel without lots of things that
the runlevel 3 services were expecting, so it was not really a 100% comparable
test.)

Comment 2 Stephen Tweedie 2006-01-24 23:21:53 UTC

Does the problem persist in current FC5test2 or rawhide?  The Xen kernel and
hypervisor have been greatly rebased in that release.  Thanks!

Comment 3 Stephen Tweedie 2006-02-24 20:10:04 UTC

Closing due to lack of reporter feedback; please reopen if you can reproduce on
current kernels.