Full Disclosure mailing list archives

Re: Linux Kernel sctp_setsockopt() Integer Overflow


From: "Evgeny Demidov" <demidov () gleg net>
Date: Sun, 16 May 2004 03:24:08 +0400

Hello,

On Sat, 15 May 2004 23:07:00 +0200 (CEST)
Jirka Kosina <jikos () jikos cz> wrote:
On Sat, 15 May 2004, Michael Tokarev wrote:

But kmalloc(0) will return NULL, and the whole setsockopt
will finish with errno set to ENOMEM.

From 2.4 mm/slab.c:
void * kmalloc (size_t size, int flags)
{
        cache_sizes_t *csizep = cache_sizes;

        for (; csizep->cs_size; csizep++) {
                if (size > csizep->cs_size)
                        continue;
return __kmem_cache_alloc(flags & GFP_DMA ? csizep->cs_dmacachep : csizep->cs_cachep, flags);
        }
        return NULL;
}

How did you come from the above snippet of the code to the idea that
kmalloc(0) returns NULL?

It allocates the number of bytes equal to the closest larger value of cache_sizes->cs_size entries ... so on typical system this would be something like 32 or 64 bytes, depending on the page size (see include/linux/kmalloc_sizes.h) ... and of course returns pointer to this data, which is definitely not NULL.


Nice advisory, there are couple of bugs like this one ;-), but I always thought that negative 'optlen' values are filtered out in sys_setsockopt (net/socket.c):

<snip>
asmlinkage long sys_setsockopt(int fd, int level, int optname, char *optval, int optlen)
{
        int err;
        struct socket *sock;

[!]     if (optlen < 0)
                return -EINVAL;

        if ((sock = sockfd_lookup(fd, &err))!=NULL)
        {
                if (level == SOL_SOCKET)
                        err=sock_setsockopt(sock,level,optname,optval,optlen);
                else
err=sock->ops->setsockopt(sock, level, optname, optval, optlen);
                sockfd_put(sock);
        }
        return err;
}

</snip>

Best regards
-Evgeny Demidov

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.netsys.com/full-disclosure-charter.html


Current thread: