Re: [PATCH 4.14 16/68] net: ensure mac header is set in virtio_net_hdr_to_skb()

From: Balazs Nemeth
Date: Thu Apr 29 2021 - 05:02:17 EST


On Fri, 2021-04-16 at 11:32 +0200, Eric Dumazet wrote:
> On Fri, Apr 16, 2021 at 10:49 AM Balazs Nemeth <bnemeth@xxxxxxxxxx>
> wrote:
> >
> > On Thu, 2021-04-15 at 16:46 +0200, Greg Kroah-Hartman wrote:
> > > From: Eric Dumazet <edumazet@xxxxxxxxxx>
> > >
> > > commit 61431a5907fc36d0738e9a547c7e1556349a03e9 upstream.
> > >
> > > Commit 924a9bc362a5 ("net: check if protocol extracted by
> > > virtio_net_hdr_set_proto is correct")
> > > added a call to dev_parse_header_protocol() but mac_header is not
> > > yet
> > > set.
> > >
> > > This means that eth_hdr() reads complete garbage, and syzbot
> > > complained about it [1]
> > >
> > > This patch resets mac_header earlier, to get more coverage about
> > > this
> > > change.
> > >
> > > Audit of virtio_net_hdr_to_skb() callers shows that this change
> > > should be safe.
> > >
> > > [1]
> > >
> > > BUG: KASAN: use-after-free in eth_header_parse_protocol+0xdc/0xe0
> > > net/ethernet/eth.c:282
> > > Read of size 2 at addr ffff888017a6200b by task syz-
> > > executor313/8409
> > >
> > > CPU: 1 PID: 8409 Comm: syz-executor313 Not tainted 5.12.0-rc2-
> > > syzkaller #0
> > > Hardware name: Google Google Compute Engine/Google Compute Engine,
> > > BIOS Google 01/01/2011
> > > Call Trace:
> > >  __dump_stack lib/dump_stack.c:79 [inline]
> > >  dump_stack+0x141/0x1d7 lib/dump_stack.c:120
> > >  print_address_description.constprop.0.cold+0x5b/0x2f8
> > > mm/kasan/report.c:232
> > >  __kasan_report mm/kasan/report.c:399 [inline]
> > >  kasan_report.cold+0x7c/0xd8 mm/kasan/report.c:416
> > >  eth_header_parse_protocol+0xdc/0xe0 net/ethernet/eth.c:282
> > >  dev_parse_header_protocol include/linux/netdevice.h:3177 [inline]
> > >  virtio_net_hdr_to_skb.constprop.0+0x99d/0xcd0
> > > include/linux/virtio_net.h:83
> > >  packet_snd net/packet/af_packet.c:2994 [inline]
> > >  packet_sendmsg+0x2325/0x52b0 net/packet/af_packet.c:3031
> > >  sock_sendmsg_nosec net/socket.c:654 [inline]
> > >  sock_sendmsg+0xcf/0x120 net/socket.c:674
> > >  sock_no_sendpage+0xf3/0x130 net/core/sock.c:2860
> > >  kernel_sendpage.part.0+0x1ab/0x350 net/socket.c:3631
> > >  kernel_sendpage net/socket.c:3628 [inline]
> > >  sock_sendpage+0xe5/0x140 net/socket.c:947
> > >  pipe_to_sendpage+0x2ad/0x380 fs/splice.c:364
> > >  splice_from_pipe_feed fs/splice.c:418 [inline]
> > >  __splice_from_pipe+0x43e/0x8a0 fs/splice.c:562
> > >  splice_from_pipe fs/splice.c:597 [inline]
> > >  generic_splice_sendpage+0xd4/0x140 fs/splice.c:746
> > >  do_splice_from fs/splice.c:767 [inline]
> > >  do_splice+0xb7e/0x1940 fs/splice.c:1079
> > >  __do_splice+0x134/0x250 fs/splice.c:1144
> > >  __do_sys_splice fs/splice.c:1350 [inline]
> > >  __se_sys_splice fs/splice.c:1332 [inline]
> > >  __x64_sys_splice+0x198/0x250 fs/splice.c:1332
> > >  do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
> > >
> > > Fixes: 924a9bc362a5 ("net: check if protocol extracted by
> > > virtio_net_hdr_set_proto is correct")
> > > Signed-off-by: Eric Dumazet <edumazet@xxxxxxxxxx>
> > > Cc: Balazs Nemeth <bnemeth@xxxxxxxxxx>
> > > Cc: Willem de Bruijn <willemb@xxxxxxxxxx>
> > > Reported-by: syzbot <syzkaller@xxxxxxxxxxxxxxxx>
> > > Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
> > > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > > ---
> > >  include/linux/virtio_net.h |    2 ++
> > >  1 file changed, 2 insertions(+)
> > >
> > > --- a/include/linux/virtio_net.h
> > > +++ b/include/linux/virtio_net.h
> > > @@ -62,6 +62,8 @@ static inline int virtio_net_hdr_to_skb(
> > >                         return -EINVAL;
> > >         }
> > >
> > > +       skb_reset_mac_header(skb);
> > > +
> > >         if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
> > >                 u16 start = __virtio16_to_cpu(little_endian, hdr-
> > > > csum_start);
> > >                 u16 off = __virtio16_to_cpu(little_endian, hdr-
> > > > csum_offset);
> > >
> > >
> >
> > Hi,
> >
> > Since the call to dev_parse_header_protocol is only made for gso
> > packets where skb->protocol is not set, we could move
> > skb_reset_mac_header down closer to that call. Is there another
> > reason
> > to reset mac_header earlier (and affect handling of other packets as
> > well)? In any case, thanks for spotting this!
> >
>
> The answer to your question was in the changelog
>
> "This patch resets mac_header earlier, to get more coverage about this
> change."
>
> We want to detect if such a reset is going to hurt in general, not
> only for GSO packets.
>


Acked-by: Balazs Nemeth <bnemeth@xxxxxxxxxx>

I can't spot any issues caused by resetting the mac header early on.

Thanks!