netdev - Re: [Question] integer overflow in function __qdisc_calculate_pkt

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Mon, 5 Jun 2023 16:19:22 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: "luwei (O)" <luwei32@...wei.com>
Cc: Networking <netdev@...r.kernel.org>
Subject: Re: [Question] integer overflow in function
 __qdisc_calculate_pkt_len()

On Fri, 2 Jun 2023 10:50:44 +0800 luwei (O) wrote:
>      I found an integer overflow issue in function 
> __qdisc_calculate_pkt_len(), the root cause is overhead and cell_align 
> in stab is not checked.
> 
> For example, if overhead is set to -2147483559 and cell_align is set to 
> -32767 (tc tool limit it to 0 and -1, but other values can be set with 
> netlink api),
> 
> the integer overflow occurs:
> 
>   568 void __qdisc_calculate_pkt_len(struct sk_buff *skb,
>   569                                const struct qdisc_size_table *stab)
>   570 {
>   571         int pkt_len, slot;
>   572
>   573         pkt_len = skb->len + stab->szopts.overhead; (1)
>   574         if (unlikely(!stab->szopts.tsize))
>   575                 goto out;
>   576
>   577         slot = pkt_len + stab->szopts.cell_align;   (2)
>   578         if (unlikely(slot < 0))
>   579                 slot = 0;
> 
> if skb->len is 66, slot will be 66 + (-2147483559) + (-32767) = 
> 2147451036, and pkt_len will be 2147451040 finally.  I think the value 
> of overhead and cell_align
> 
> should be limited, but not sure to which values they should be limited, 
> can any one give me some suggestions?

on a quick look limiting the cell_align to S16_MIN at the netlink level
(NLA_POLICY_MIN()) seems reasonable, feel free to send a patch.