Couple of bugs

Steve Grubb sgrubb at redhat.com
Sun Feb 22 19:21:21 CET 2009


Hi,

I am using libev as part of the Linux audit daemon. I have done some testing 
with 3.53 recently and run across a couple of "issues".

I created a simple "hello world" program.

#include <stdio.h>
#include <sys/types.h>
#include <ev.h>

volatile int stop = 0;
static void term_handler(struct ev_loop *loop, struct ev_signal *sig,
                        int revents)
{
	ev_unloop (ev_default_loop (EVFLAG_AUTO), EVUNLOOP_ALL);
	stop = 1;
}

int main(void)
{
	struct ev_loop *loop;
	struct ev_signal sigterm_watcher;

	loop = ev_default_loop (EVFLAG_NOENV);
        ev_signal_init (&sigterm_watcher, term_handler, SIGTERM);
        ev_signal_start (loop, &sigterm_watcher);

        while (!stop)
                ev_loop (loop, 0);

	ev_loop_destroy(loop);
	return 0;
}


When compiling, I get this:

In file included from test.c:3:
/home/sgrubb/working/BUILD/audit-1.8/src/libev/ev.h:123:5: 
warning: "__STDC_VERSION__" is not defined

It would seem that line 123 of ev.h should be changed to:

#if defined( __STDC_VERSION__) &&  __STDC_VERSION__ >= 199901L || __GNUC__ >= 
3

iow test to see if its defined before using it. But anyways, the real problem 
shows up when you use valgrind against the test program. This is what 
valgrind reports:

==10873== Memcheck, a memory error detector.
==10873== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al.
==10873== Using LibVEX rev 1804, a library for dynamic binary translation.
==10873== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP.
==10873== Using valgrind-3.3.0, a dynamic binary instrumentation framework.
==10873== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al.
==10873== For more details, rerun with: -v
==10873== 
==10873== Invalid free() / delete / delete[]
==10873==    at 0x4A0609F: free (vg_replace_malloc.c:323)
==10873==    by 0x40198C: ev_realloc_emul (ev.c:452)
==10873==    by 0x40142C: main (test.c:27)
==10873==  Address 0x6095c0 is not stack'd, malloc'd or (recently) free'd
==10873== 
==10873== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 4 from 1)
==10873== malloc/free: in use at exit: 512 bytes in 1 blocks.
==10873== malloc/free: 6 allocs, 6 frees, 1,480 bytes allocated.
==10873== For counts of detected errors, rerun with: -v
==10873== searching for pointers to 1 not-freed blocks.
==10873== checked 69,656 bytes.
==10873== 
==10873== 
==10873== 512 bytes in 1 blocks are still reachable in loss record 1 of 1
==10873==    at 0x4A0739E: malloc (vg_replace_malloc.c:207)
==10873==    by 0x4A07527: realloc (vg_replace_malloc.c:429)
==10873==    by 0x4033FF: array_realloc (ev.c:467)
==10873==    by 0x405954: ev_signal_start (ev.c:2386)
==10873==    by 0x405A20: ev_default_loop_init (ev.c:1679)
==10873==    by 0x4013A4: ev_default_loop (ev.h:447)
==10873==    by 0x4013CD: main (test.c:20)
==10873== 
==10873== LEAK SUMMARY:
==10873==    definitely lost: 0 bytes in 0 blocks.
==10873==      possibly lost: 0 bytes in 0 blocks.
==10873==    still reachable: 512 bytes in 1 blocks.
==10873==         suppressed: 0 bytes in 0 blocks.

We are leaking memory and have an invalid free. Any ideas?

Thanks,
-Steve



More information about the libev mailing list