netdev - Re: [PATCH net] macvlan: do not assume mac_header is set in macvlan

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: <20200108.125329.978670665805978687.davem@davemloft.net>
Date:   Wed, 08 Jan 2020 12:53:29 -0800 (PST)
From:   David Miller <davem@...emloft.net>
To:     edumazet@...gle.com
Cc:     netdev@...r.kernel.org, eric.dumazet@...il.com,
        syzkaller@...glegroups.com
Subject: Re: [PATCH net] macvlan: do not assume mac_header is set in
 macvlan_broadcast()

From: Eric Dumazet <edumazet@...gle.com>
Date: Mon,  6 Jan 2020 12:30:48 -0800

> Use of eth_hdr() in tx path is error prone.
> 
> Many drivers call skb_reset_mac_header() before using it,
> but others do not.
> 
> Commit 6d1ccff62780 ("net: reset mac header in dev_start_xmit()")
> attempted to fix this generically, but commit d346a3fae3ff
> ("packet: introduce PACKET_QDISC_BYPASS socket option") brought
> back the macvlan bug.
> 
> Lets add a new helper, so that tx paths no longer have
> to call skb_reset_mac_header() only to get a pointer
> to skb->data.
> 
> Hopefully we will be able to revert 6d1ccff62780
> ("net: reset mac header in dev_start_xmit()") and save few cycles
> in transmit fast path.
 ...
> Fixes: b863ceb7ddce ("[NET]: Add macvlan driver")
> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> Reported-by: syzbot <syzkaller@...glegroups.com>

Applied and queued up for -stable, thanks Eric.