This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa mudflap]
- From: "Frank Ch. Eigler" <fche at redhat dot com>
- To: Eyal Lebedinsky <eyal at eyal dot emu dot id dot au>
- Cc: gcc list <gcc at gcc dot gnu dot org>
- Date: Fri, 3 Oct 2003 12:05:21 -0400
- Subject: Re: [tree-ssa mudflap]
- References: <3F7D672F.F88308D1@eyal.emu.id.au>
Hi -
On Fri, Oct 03, 2003 at 10:10:23PM +1000, Eyal Lebedinsky wrote:
> [...]
> My interpretation:
> At some point strerror() was called, which internally
> accessed pthread_rwlock_rdlock() (which I now trap in mf-hooks2.c)
> which eventually led to a malloc(12), which was trapped and
> registered.
> [...]
I believe that your hook function for pthread_rwlock_rdlock,
and other such functions, will have to set the reentrancy
state. This way, mallocs and frees run by the pthread runtime
will not be tracked by the libmudflap runtime.
> I also have an unexplained unregister violation in fork()
> which does not bother me too much at this point.
I suspect something similar is going on there.
> I use this libmudflap patch:
> http://users.bigpond.net.au/eyal/
Please do submit the patch formally, with a ChangeLog. There is
no need for you to maintain your separate patch, given that at the
moment, you are de facto the main developer of this part of
libmudflap. :-)
- FChE