#3 calltree/callgrind can cause valgrind to segfault

[Anonymous]

I'm not sure if this is a bug in valgrind of calltree,
but I have an executable that I want to profile, but in
the middle of the profile I get the message:

--3561-- INTERNAL ERROR: Valgrind received a signal 11
(SIGSEGV) - exiting
--3561-- si_code=1 Fault EIP: 0xB806AAC3; Faulting
address: 0xBFB37000

valgrind: the `impossible' happened:
Killed by fatal signal
Basic block ctr is approximately 143600000

valgrind: vg_mylibc.c:1681
(vgPlain_get_memory_from_mmap): Assertion `p >= (void
*)vgPlain_valgrind_mmap_e
nd && p < (void *)vgPlain_valgrind_end' failed.

What I do know:

1. Both calltree nor callgrind cause this problem when
I profile this executable
2. Cachegrind works fine with this executable
3. The segfault doesn't occur until I start making
corba calls (over tcp/ip) into the application. There
may be some system call my application is making that
is causing the problem.
4. The application is single-threaded

I'm sure this isn't enough information to track this
problem down, but I don't know what information I need
to provide.

[Paul Brannan]

Logged In: YES
user_id=3554

Oops, I should have logged in before entering this bug.

-- Paul Brannan (cout)
pbrannan at atdesk.com

[Josef Weidendorfer]

Logged In: YES
user_id=621915

This looks like an out-of-memory condition from Valgrind because of
memory requirements of Calltree.
Valgrind 2.1.1 is quite memory sensitive.
Does it work with Valgrind 2.0.0 or 2.1.2?