diff options
author | Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru> | 2005-11-20 13:41:59 -0800 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2005-11-20 13:41:59 -0800 |
commit | 4909724b5dee8fb7c52bbe90afa40c65b17be9eb (patch) | |
tree | 7611a0f01ff0b0a1389928683b4223c160123d2a /arch | |
parent | fb0d366b0803571f06a5b838f02c6706fc287995 (diff) |
[COMPAT] net: SIOCGIFCONF data corruption
From: Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
From http://bugzilla.kernel.org/show_bug.cgi?id=4746
There is user data corruption when using ioctl(SIOCGIFCONF) in 32-bit
application running amd64 kernel. I do not think that this problem is
exploitable, but any data corruption may lead to security problems.
Following code demonstrates the problem
#include <stdint.h>
#include <stdio.h>
#include <sys/time.h>
#include <sys/socket.h>
#include <net/if.h>
#include <sys/ioctl.h>
char buf[256];
main()
{
int s = socket(AF_INET, SOCK_DGRAM, 0);
struct ifconf req;
int i;
req.ifc_buf = buf;
req.ifc_len = 41;
printf("Result %d\n", ioctl(s, SIOCGIFCONF, &req));
printf("Len %d\n", req.ifc_len);
for (i = 41; i < 256; i++)
if (buf[i] != 0)
printf("Byte %d is corrupted\n", i);
}
Steps to reproduce:
Compile the code above into 32-bit elf and run it. You'll get
Result 0
Len 32
Byte 48 is corrupted
Byte 52 is corrupted
Byte 53 is corrupted
Byte 54 is corrupted
Byte 55 is corrupted
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'arch')
0 files changed, 0 insertions, 0 deletions