ptrace() fix to make strace work better again

From: David Mosberger <davidm_at_napali.hpl.hp.com>
Date: 2003-12-31 20:34:15
Some of you have noticed that strace -f was often causing the target
process to crash.  The problem seems to be due to NPTL-enabled glibc
stressing ptrace() in different ways than LinuxThreads and as a
result, linking against NPTL triggered the strace problems.  The root
cause of the problem has been there for a long time (both in 2.4 and
2.6 kernels) it's just that we didn't see it before.

I have now a patch which fixes the strace -f problems without causing
any regressions in the gdb test-suite.  See this change-set:

  http://lia64.bkbits.net:8080/to-linus-2.5/cset@1.1530

The changes are rather big, so I'd welcome thorough testing of it.  In
particular, it would be good to stress-test it more on largish
SMP-machines.  Also, if you have any apps that do strange things with
ptrace, or are using debuggers or strace a lot, please give it a good
whirl and let me know if you see any problems.

As for strace, with the patch I can now run "strace -f mozilla" without
any apparent problem.

Regards,

	--david
-
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Received on Wed Dec 31 04:34:29 2003

This archive was generated by hypermail 2.1.8 : 2005-08-02 09:20:21 EST