* Realloc() problem with Efence
@ 1999-07-28 21:48 Alain Birtz
1999-07-28 20:03 ` Nicholas T Ingolia
1999-07-28 20:54 ` Jerry Quinn
0 siblings, 2 replies; 3+ messages in thread
From: Alain Birtz @ 1999-07-28 21:48 UTC (permalink / raw)
To: linuxppc-dev
What is wrong with this code ?
Code is executed correctly without Electric Fence
But linked with Efence library, dbg complaint:
Electric Fence 2.0.5 Copyright (C) 1987-1998 Bruce Perens.
...
...
Program received signal SIGSEGV, Segmentation fault.
when trying to write the first B, just after the first memory allocation
buffer
(at the instruction (*c_ptr++ = 'B'))
void tst ()
{
int i;
char *c_buf, *c_ptr, *temp_c_buf;
c_buf = (char *)malloc(5000);
if (c_buf != NULL)
{
c_ptr = c_buf;
for (i = 0; i < 5000; i++)
*c_ptr++ = 'A';
temp_c_buf = c_buf;
realloc(temp_c_buf, 5000 + 5000);
if (temp_c_buf != NULL)
{
for (i = 0; i < 5000; i++)
*c_ptr++ = 'B';
}
}
}
[[ This message was sent via the linuxppc-dev mailing list. Replies are ]]
[[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: Realloc() problem with Efence
1999-07-28 21:48 Realloc() problem with Efence Alain Birtz
@ 1999-07-28 20:03 ` Nicholas T Ingolia
1999-07-28 20:54 ` Jerry Quinn
1 sibling, 0 replies; 3+ messages in thread
From: Nicholas T Ingolia @ 1999-07-28 20:03 UTC (permalink / raw)
To: Alain Birtz; +Cc: linuxppc-dev
-----BEGIN PGP SIGNED MESSAGE-----
Hello...
realloc() may move the memory being realloated. To quote from the man page,
realloc() returns a pointer to the newly allocated memory,
which is suitably aligned for any kind of variable and may
be different from ptr...
It doesn't (and couldn't) mutate the pointer temp_c_buf directly.
Thus, you probably want to re-write your code to do
temp_c_buf = realloc(temp_c_buf, 5000 + 5000);
- --Nicholas Ingolia
ingolia@mit.edu
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0i for non-commercial use
Comment: Processed by Mailcrypt 3.5.3, an Emacs/PGP interface
Charset: noconv
iQCVAwUBN59h+gRHXKx72OUhAQEkNAP/bYZO+zzGI2bDBRuCYabEfaxo8/EAffTN
hliGzKLwhhEetuvrUNv+pam5qBXGb3OLWWAZdRmypEvPXfYP/w3wGlJlGKWxAbGm
WhBJkzwVWEkqKmqADiClP7ucJcMIYkMcLLVcCBnVEsjoPFDumoc2ncEe6beeRDIB
gma/mrLVmZs=
=tyhG
-----END PGP SIGNATURE-----
[[ This message was sent via the linuxppc-dev mailing list. Replies are ]]
[[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]
^ permalink raw reply [flat|nested] 3+ messages in thread* Realloc() problem with Efence
1999-07-28 21:48 Realloc() problem with Efence Alain Birtz
1999-07-28 20:03 ` Nicholas T Ingolia
@ 1999-07-28 20:54 ` Jerry Quinn
1 sibling, 0 replies; 3+ messages in thread
From: Jerry Quinn @ 1999-07-28 20:54 UTC (permalink / raw)
To: Alain Birtz; +Cc: linuxppc-dev
Alain Birtz <abz@videotron.ca> writes:
Alain> What is wrong with this code ?
Alain> Code is executed correctly without Electric Fence
Alain> But linked with Efence library, dbg complaint:
Alain>
Alain> Electric Fence 2.0.5 Copyright (C) 1987-1998 Bruce Perens.
Alain> ...
Alain> ...
Alain> Program received signal SIGSEGV, Segmentation fault.
Alain>
Alain> when trying to write the first B, just after the first memory allocation
Alain> buffer
Alain> (at the instruction (*c_ptr++ = 'B'))
Alain>
Alain>
Alain> void tst ()
Alain> {
Alain> int i;
Alain> char *c_buf, *c_ptr, *temp_c_buf;
Alain>
Alain> c_buf = (char *)malloc(5000);
Alain> if (c_buf != NULL)
Alain> {
Alain> c_ptr = c_buf;
Alain> for (i = 0; i < 5000; i++)
Alain> *c_ptr++ = 'A';
Alain> temp_c_buf = c_buf;
Alain> realloc(temp_c_buf, 5000 + 5000);
Alain> if (temp_c_buf != NULL)
Alain> {
Alain> for (i = 0; i < 5000; i++)
Alain> *c_ptr++ = 'B';
Alain> }
Alain> }
Alain> }
Alain>
realloc can (and did) move the block. Therefore, when you start writing 'B's,
they are going into invalid memory. The fact that it works without Electric
Fence is luck.
To do this, you need to set the pointer to the whole block to the output of
realloc and then you can start writing B's at the new location + 5000.
--
Jerry Quinn Tel: (514) 761-8737
jquinn@nortelnetworks.com Fax: (514) 761-8505
Speech Recognition Research
[[ This message was sent via the linuxppc-dev mailing list. Replies are ]]
[[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~1999-07-28 21:48 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
1999-07-28 21:48 Realloc() problem with Efence Alain Birtz
1999-07-28 20:03 ` Nicholas T Ingolia
1999-07-28 20:54 ` Jerry Quinn
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).