netdev - Re: [PATCH net] net: qrtr: Fix an uninit variable access bug in qrtr_tx

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:   Sat, 8 Apr 2023 14:05:28 +0530
From:   Manivannan Sadhasivam <mani@...nel.org>
To:     Ziyang Xuan <william.xuanziyang@...wei.com>
Cc:     davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
        pabeni@...hat.com, andersson@...nel.org, luca@...tu.xyz,
        linux-arm-msm@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [PATCH net] net: qrtr: Fix an uninit variable access bug in
 qrtr_tx_resume()

On Mon, Apr 03, 2023 at 03:54:17PM +0800, Ziyang Xuan wrote:
> Syzbot reported a bug as following:
> 
> =====================================================
> BUG: KMSAN: uninit-value in qrtr_tx_resume+0x185/0x1f0 net/qrtr/af_qrtr.c:230
>  qrtr_tx_resume+0x185/0x1f0 net/qrtr/af_qrtr.c:230
>  qrtr_endpoint_post+0xf85/0x11b0 net/qrtr/af_qrtr.c:519
>  qrtr_tun_write_iter+0x270/0x400 net/qrtr/tun.c:108
>  call_write_iter include/linux/fs.h:2189 [inline]
>  aio_write+0x63a/0x950 fs/aio.c:1600
>  io_submit_one+0x1d1c/0x3bf0 fs/aio.c:2019
>  __do_sys_io_submit fs/aio.c:2078 [inline]
>  __se_sys_io_submit+0x293/0x770 fs/aio.c:2048
>  __x64_sys_io_submit+0x92/0xd0 fs/aio.c:2048
>  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>  do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
>  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> 
> Uninit was created at:
>  slab_post_alloc_hook mm/slab.h:766 [inline]
>  slab_alloc_node mm/slub.c:3452 [inline]
>  __kmem_cache_alloc_node+0x71f/0xce0 mm/slub.c:3491
>  __do_kmalloc_node mm/slab_common.c:967 [inline]
>  __kmalloc_node_track_caller+0x114/0x3b0 mm/slab_common.c:988
>  kmalloc_reserve net/core/skbuff.c:492 [inline]
>  __alloc_skb+0x3af/0x8f0 net/core/skbuff.c:565
>  __netdev_alloc_skb+0x120/0x7d0 net/core/skbuff.c:630
>  qrtr_endpoint_post+0xbd/0x11b0 net/qrtr/af_qrtr.c:446
>  qrtr_tun_write_iter+0x270/0x400 net/qrtr/tun.c:108
>  call_write_iter include/linux/fs.h:2189 [inline]
>  aio_write+0x63a/0x950 fs/aio.c:1600
>  io_submit_one+0x1d1c/0x3bf0 fs/aio.c:2019
>  __do_sys_io_submit fs/aio.c:2078 [inline]
>  __se_sys_io_submit+0x293/0x770 fs/aio.c:2048
>  __x64_sys_io_submit+0x92/0xd0 fs/aio.c:2048
>  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>  do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
>  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> 
> It is because that skb->len requires at least sizeof(struct qrtr_ctrl_pkt)
> in qrtr_tx_resume(). And skb->len equals to size in qrtr_endpoint_post().
> But size is less than sizeof(struct qrtr_ctrl_pkt) when qrtr_cb->type
> equals to QRTR_TYPE_RESUME_TX in qrtr_endpoint_post() under the syzbot
> scenario. This triggers the uninit variable access bug.
> 
> Add size check when qrtr_cb->type equals to QRTR_TYPE_RESUME_TX in
> qrtr_endpoint_post() to fix the bug.
> 
> Fixes: 5fdeb0d372ab ("net: qrtr: Implement outgoing flow control")
> Reported-by: syzbot+4436c9630a45820fda76@...kaller.appspotmail.com
> Link: https://syzkaller.appspot.com/bug?id=c14607f0963d27d5a3d5f4c8639b500909e43540
> Signed-off-by: Ziyang Xuan <william.xuanziyang@...wei.com>

Looks good to me. But I have a small suggestion below.

> ---
>  net/qrtr/af_qrtr.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/net/qrtr/af_qrtr.c b/net/qrtr/af_qrtr.c
> index 3a70255c8d02..631e81a8a368 100644
> --- a/net/qrtr/af_qrtr.c
> +++ b/net/qrtr/af_qrtr.c
> @@ -498,6 +498,10 @@ int qrtr_endpoint_post(struct qrtr_endpoint *ep, const void *data, size_t len)
>  	if (!size || len != ALIGN(size, 4) + hdrlen)
>  		goto err;
>  
> +	if (cb->type == QRTR_TYPE_RESUME_TX &&
> +	    size < sizeof(struct qrtr_ctrl_pkt))

There is already a check for QRTR_TYPE_NEW_SERVER below. So you can combine both:

	if ((cb->type == QRTR_TYPE_NEW_SERVER ||
	     cb->type == QRTR_TYPE_RESUME_TX)
	     && size < sizeof(struct qrtr_ctrl_pkt))
		goto err;

- Mani

> +		goto err;
> +
>  	if (cb->dst_port != QRTR_PORT_CTRL && cb->type != QRTR_TYPE_DATA &&
>  	    cb->type != QRTR_TYPE_RESUME_TX)
>  		goto err;
> -- 
> 2.25.1
> 

-- 
மணிவண்ணன் சதாசிவம்