netdev - Re: [syzbot] [bpf?] [net?] general protection fault in dev_map

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] [day] [month] [year] [list]

Message-ID: <20240902080232.wnhtxiWK@linutronix.de>
Date: Mon, 2 Sep 2024 10:02:32 +0200
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: syzbot <syzbot+cca39e6e84a367a7e6f6@...kaller.appspotmail.com>
Cc: alexei.starovoitov@...il.com, andrii@...nel.org, ast@...nel.org,
	bpf@...r.kernel.org, daniel@...earbox.net, davem@...emloft.net,
	eddyz87@...il.com, haoluo@...gle.com, hawk@...nel.org,
	john.fastabend@...il.com, jolsa@...nel.org, kpsingh@...nel.org,
	kuba@...nel.org, linux-kernel@...r.kernel.org, martin.lau@...ux.dev,
	michal.switala@...ogain.com, netdev@...r.kernel.org,
	revest@...gle.com, sdf@...ichev.me, sdf@...gle.com, song@...nel.org,
	syzkaller-bugs@...glegroups.com, toke@...hat.com,
	yonghong.song@...ux.dev
Subject: Re: [syzbot] [bpf?] [net?] general protection fault in
 dev_map_enqueue (2)

On 2024-08-31 13:55:02 [-0700], syzbot wrote:
> syzbot suspects this issue was fixed by commit:
> 
> commit 401cb7dae8130fd34eb84648e02ab4c506df7d5e
> Author: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
> Date:   Thu Jun 20 13:22:04 2024 +0000
> 
>     net: Reference bpf_redirect_info via task_struct on PREEMPT_RT.
> 
> bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=12597c63980000
> start commit:   36534d3c5453 tcp: use signed arithmetic in tcp_rtx_probe0_..
> git tree:       bpf
> kernel config:  https://syzkaller.appspot.com/x/.config?x=333ebe38d43c42e2
> dashboard link: https://syzkaller.appspot.com/bug?extid=cca39e6e84a367a7e6f6
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=13390aea980000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=10948741980000

This looks like ri->tgt_value is a NULL pointer (dst in
dev_map_enqueue()). The commit referenced by syz should not have fixed
that.
It is possible that there were leftovers in bpf_redirect_info (from a
previous invocation) which were memset(,0,) during the switch from
per-CPU to stack usage and now it does not trigger anymore. 

Sebastian