[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20240118111751.1426-1-hdanton@sina.com>
Date: Thu, 18 Jan 2024 19:17:51 +0800
From: Hillf Danton <hdanton@...a.com>
To: syzbot <syzbot+bfde3bef047a81b8fde6@...kaller.appspotmail.com>
Cc: linux-kernel@...r.kernel.org,
syzkaller-bugs@...glegroups.com
Subject: Re: [syzbot] [net?] KASAN: use-after-free Read in __skb_flow_dissect (3)
On Mon, 01 Jan 2024 09:18:16 -0800
> syzbot found the following issue on:
>
> HEAD commit: f5837722ffec Merge tag 'mm-hotfixes-stable-2023-12-27-15-0..
> git tree: upstream
> C reproducer: https://syzkaller.appspot.com/x/repro.c?x=122dfc65e80000
#syz test https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
--- x/net/core/filter.c
+++ y/net/core/filter.c
@@ -2431,7 +2431,7 @@ enum {
BPF_CALL_3(bpf_clone_redirect, struct sk_buff *, skb, u32, ifindex, u64, flags)
{
struct net_device *dev;
- struct sk_buff *clone;
+ struct sk_buff *buf;
int ret;
if (unlikely(flags & (~(BPF_F_INGRESS) | BPF_F_REDIRECT_INTERNAL)))
@@ -2441,22 +2441,11 @@ BPF_CALL_3(bpf_clone_redirect, struct sk
if (unlikely(!dev))
return -EINVAL;
- clone = skb_clone(skb, GFP_ATOMIC);
- if (unlikely(!clone))
+ buf = skb_copy(skb, GFP_ATOMIC);
+ if (unlikely(!buf))
return -ENOMEM;
- /* For direct write, we need to keep the invariant that the skbs
- * we're dealing with need to be uncloned. Should uncloning fail
- * here, we need to free the just generated clone to unclone once
- * again.
- */
- ret = bpf_try_make_head_writable(skb);
- if (unlikely(ret)) {
- kfree_skb(clone);
- return -ENOMEM;
- }
-
- return __bpf_redirect(clone, dev, flags);
+ return __bpf_redirect(buf, dev, flags);
}
static const struct bpf_func_proto bpf_clone_redirect_proto = {
--
Powered by blists - more mailing lists