Re: [Qemu-devel] [PATCH v3] inet_listen_opts: add error checking

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1829 bytes --]

On 12/13/2013 03:08 AM, Gerd Hoffmann wrote:
> Don't use atoi() function which doesn't detect errors, switch to
> strtol and error out on failures.  Also add a range check while
> being at it.
> 
> [ v3: oops, v2 didn't build ]
> [ v2: use parse_uint_full instead of strtol ]

Patch changelog belongs...

> 
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> ---

...here, so that 'git am' will strip it (it's useful for reviewers on
list, but not in 'git log', where you no longer have access to v1 or v2).

>  util/qemu-sockets.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/util/qemu-sockets.c b/util/qemu-sockets.c
> index 6b97dc1..c3560c1 100644
> --- a/util/qemu-sockets.c
> +++ b/util/qemu-sockets.c
> @@ -133,8 +133,18 @@ int inet_listen_opts(QemuOpts *opts, int port_offset, Error **errp)
>          ai.ai_family = PF_INET6;
>  
>      /* lookup */
> -    if (port_offset)
> -        snprintf(port, sizeof(port), "%d", atoi(port) + port_offset);
> +    if (port_offset) {
> +        int baseport;
> +        if (parse_uint_full(port, &baseport, 10) < 0) {

parse_uint_full takes an 'unsigned long long *', but you are passing an
'int *'.  I'm surprised it compiled for you.  It causes a buffer
overflow if the pointer is assigned to, and gives different results
depending on platform endianness.

> +            error_setg(errp, "can't convert to a number: %s", port);
> +            return -1;
> +        }
> +        if (baseport < 0 || baseport + port_offset > 65535) {
> +            error_setg(errp, "port %s out of range", port);

But errno is not set to a sane value at this point, so error_setg() is
wrong.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 621 bytes --]