Re: [PATCH] vnc: avoid underflow when accessing user-provided address

[Philippe Mathieu-Daudé]

On 30/3/23 14:44, Paolo Bonzini wrote:
> If hostlen is zero, there is a possibility that addrstr[hostlen - 1]
> underflows and, if a closing bracked is there, hostlen - 2 is passed
> to g_strndup() on the next line.  If websocket==false then
> addrstr[0] would be a colon, but if websocket==true this could in
> principle happen.
>
> Fix it by checking hostlen.
>
> Reported by Coverity.

Also by Vladimir Sementsov-Ogievskiy few months ago:
https://lore.kernel.org/qemu-devel/20221206192334.65012-1-vsementsov@yandex-team.ru/

> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>   ui/vnc.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/ui/vnc.c b/ui/vnc.c
> index bbd8b6baaeca..9d8a24dd8a69 100644
> --- a/ui/vnc.c
> +++ b/ui/vnc.c
> @@ -3751,7 +3751,7 @@ static int vnc_display_get_address(const char *addrstr,
>   
>           addr->type = SOCKET_ADDRESS_TYPE_INET;
>           inet = &addr->u.inet;
> -        if (addrstr[0] == '[' && addrstr[hostlen - 1] == ']') {
> +        if (hostlen && addrstr[0] == '[' && addrstr[hostlen - 1] == ']') {
>               inet->host = g_strndup(addrstr + 1, hostlen - 2);
>           } else {
>               inet->host = g_strndup(addrstr, hostlen);