Question

I’m at a loss. Here is what happens in gdb:

571             void *aux = realloc(cs->body, new_size);
(gdb) p cs->body
$6 = 0x627b20 "POST /upload HTTP/1.1\r\nUser-Agent: curl/7.22.0 (x86_64-unknown-linux-gnu) libcurl/7.22.0 OpenSSL/1.0.0e zlib/1.2.5 c-ares/1.7.5 libidn/1.22 libssh2/1.2.9\r\nHost: localhost:2001\r\nAccept: */*\r\nContent-Le"...
(gdb) n
*** glibc detected *** /home/pc/bin/main: realloc(): invalid pointer: 0x0000000000627b20 ***

…far down after the backtrace and memory map (which I don’t know how to interpret):

Program received signal SIGABRT, Aborted.
0x00007ffff7660d95 in raise () from /lib64/libc.so.6

How is this possible? I printed cs->body right before the realloc call, and it looks fine right? How can it be an invalid pointer, but still contain characters just fine?
How can I debug this segfault?

Answer 1

Dereferencing invalid pointers invokes undefined behaviour, which means that anything could happen, including the behaviour that you’re observing. If I had to guess, I’d say that either:

• your pointer really is invalid (it wasn’t obtained from malloc, or you’ve already freed it, or you’ve modified it accidentally.
• you’ve overflowed a buffer somewhere, and trashed malloc‘s behind-the-scenes metadata.

I would suggest running your application in something like Valgrind, which is designed to help track down memory issues.