[Ltrace-devel] fixes for tracing -p, -L

Ian Wienand ianw at gelato.unsw.edu.au
Mon Jan 22 07:59:46 CET 2007


On Fri, Jan 19, 2007 at 05:01:55PM +0100, Petr Machata wrote:
> anyway, trunk testsuite fails on mouth.

Is mouth a machine?  What architecture.

I am actually seeing a strange 'Operation not permitted' problem on my
386 Debian box with the clone test at the moment.  This happens with
or without your patch.

Executing on host: gcc ./ltrace.minor/trace-clone.c  -g  -lm   -o ./ltrace.minor/trace-clone    (timeout = 300)
__libc_start_main(0x80483fe, 1, 0xbfc01ac4, 0x80484c0, 0x8048470 <unfinished ...>
clone(0x80483f4, 0x80496e0, 512, 0, 0xb7ff1ff4Cannot attach to pid 10582: Operation not permitted
)                          = 10582
+++ exited (status 0) +++
FAIL: Operation not permitted, see testrun.log for details!
testcase ./ltrace.minor/trace-clone.exp completed in 0 seconds


> I'm attaching the patch that fixes these two problems for me.  I'd very 
> much appreciate if you gave it a test run on the distro of your choice, 
> as my experience is pretty much Fedora/RHEL-centric.

Modulo the error that already exists, this works on my Debian x86 box.

> Hmm, and I see my pie patch became completely wrecked over the 
> christmas.  Maybe indeed some change in kernel or toolchain...

I remember reading this message; maybe that has something to do with
it?

http://lkml.org/lkml/2007/1/6/124

-i



More information about the Ltrace-devel mailing list