Bug 143350

Summary: ptrace horked in 2.6.9-1.1047_FC4smp
Product: [Fedora] Fedora Reporter: Ulrich Drepper <drepper>
Component: kernelAssignee: Roland McGrath <roland>
Status: CLOSED RAWHIDE QA Contact: Brian Brock <bbrock>
Severity: medium Docs Contact:
Priority: medium    
Version: rawhideCC: davej, mgb, wtogami
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2005-01-04 21:00:01 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 Ulrich Drepper 2004-12-19 21:16:16 UTC
Description of problem:
ptrace() seems to be completely broken in 2.6.9-1.1047_FC4smp.

$ strace /bin/echo
execve("/bin/echo", ["/bin/echo"], [/* 39 vars */]) = 0

... and then it's stuck.  The process has to be aborted.  gdb shows the same
behavior.

Version-Release number of selected component (if applicable):
2.6.9-1.1047_FC4smp

How reproducible:
I haven't rebooted and don't know whether this has been the case ever since the
last boot, but it surely is 100% reproducible now.

Steps to Reproduce:
1.boot said kernel
2.strace /bin/echo
3.
  
Actual results:
Nothing except the fake execve message

Expected results:
trace

Additional info:
strace-4.5.8-1.x86_64 is used, but this shouldn't matter since gdb is also
affected and both programs worked before.

Comment 1 Ulrich Drepper 2004-12-20 02:37:47 UTC
This happens on x86 as well.  So not x86-64 specific.

Comment 2 Roland McGrath 2004-12-20 22:34:07 UTC
This is ok in the upstream kernel and in the RHEL4 kernel.
So it's just some sort of merging snafu or whatnot in the FC4 kernel.
I didn't see anything obvious.  I'm not going to worry about this since I figure
bleeding edge FC should be rebasing to 2.6.10(rc*) soon enough anyway.

Comment 3 Michal Jaegermann 2004-12-23 23:48:03 UTC
This got corrected a bit but in 2.6.9-1.1049_FC4 but not that much.
'strace echo' just prints, at least on x86_64,

execve("/bin/echo", ["echo"], [/* 24 vars */]) = 0

and returns.  I would expect to see some 38 lines of output.

Comment 4 Dave Jones 2005-01-04 17:47:51 UTC
theres a 2.6.10 based update in rawhide now, can you test that please ?

Comment 5 Ulrich Drepper 2005-01-04 21:00:01 UTC
Seems to work again in 2.6.10-based kernels.  I've only tested x86 so far, but
since the symptoms were identical this should apply to other archs as well.