[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PULL V2 25/25] net/colo.c: fix segmentation fault when packet is no
From: |
Peter Maydell |
Subject: |
Re: [PULL V2 25/25] net/colo.c: fix segmentation fault when packet is not parsed correctly |
Date: |
Fri, 29 Jul 2022 14:58:01 +0100 |
On Wed, 20 Jul 2022 at 10:04, Jason Wang <jasowang@redhat.com> wrote:
>
> From: Zhang Chen <chen.zhang@intel.com>
>
> When COLO use only one vnet_hdr_support parameter between
> filter-redirector and filter-mirror(or colo-compare), COLO will crash
> with segmentation fault. Back track as follow:
>
> Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault.
> 0x0000555555cb200b in eth_get_l2_hdr_length (p=0x0)
> at /home/tao/project/COLO/colo-qemu/include/net/eth.h:296
> 296 uint16_t proto = be16_to_cpu(PKT_GET_ETH_HDR(p)->h_proto);
> (gdb) bt
> 0 0x0000555555cb200b in eth_get_l2_hdr_length (p=0x0)
> at /home/tao/project/COLO/colo-qemu/include/net/eth.h:296
> 1 0x0000555555cb22b4 in parse_packet_early (pkt=0x555556a44840) at
> net/colo.c:49
> 2 0x0000555555cb2b91 in is_tcp_packet (pkt=0x555556a44840) at
> net/filter-rewriter.c:63
>
> So wrong vnet_hdr_len will cause pkt->data become NULL. Add check to
> raise error and add trace-events to track vnet_hdr_len.
>
> Signed-off-by: Tao Xu <tao3.xu@intel.com>
> Signed-off-by: Zhang Chen <chen.zhang@intel.com>
> Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
> Signed-off-by: Jason Wang <jasowang@redhat.com>
Hi -- Coverity points out that there's a problem with this fix
(CID 1490786):
> @@ -46,7 +46,14 @@ int parse_packet_early(Packet *pkt)
> static const uint8_t vlan[] = {0x81, 0x00};
> uint8_t *data = pkt->data + pkt->vnet_hdr_len;
data here is set to pkt->data + pkt->vnet_hdr_len.
If pkt->data is NULL then this addition is C undefined behaviour,
and if pkt->data is not NULL but the integer added is such
that the pointer ends up not pointing within data, then that
is also C undefined behaviour...
> uint16_t l3_proto;
> - ssize_t l2hdr_len = eth_get_l2_hdr_length(data);
> + ssize_t l2hdr_len;
> +
> + if (data == NULL) {
...so the compiler is allowed to assume that data cannot be NULL
here, and this entire error checking clause is logically dead code.
If you're trying to check whether vnet_hdr_len is valid, you
need to do that as an integer arithmetic check before you start
using it for pointer arithmetic. And you probably want to be
checking against some kind of bound, not just for whether the
result is going to be NULL.
Overall this looks kinda sketchy and could probably use more
detailed code review. Where does the bogus vnet_hdr_len come from in
the first place? Is this data from the guest, or from the network
(ie uncontrolled)?
> + trace_colo_proxy_main_vnet_info("This packet is not parsed
> correctly, "
> + "pkt->vnet_hdr_len",
> pkt->vnet_hdr_len);
> + return 1;
> + }
thanks
-- PMM
- [PULL V2 12/25] vhost: Expose vhost_svq_add, (continued)
- [PULL V2 12/25] vhost: Expose vhost_svq_add, Jason Wang, 2022/07/20
- [PULL V2 14/25] vhost: Add svq avail_handler callback, Jason Wang, 2022/07/20
- [PULL V2 13/25] vhost: add vhost_svq_poll, Jason Wang, 2022/07/20
- [PULL V2 16/25] vhost-net-vdpa: add stubs for when no virtio-net device is present, Jason Wang, 2022/07/20
- [PULL V2 18/25] vdpa: Buffer CVQ support on shadow virtqueue, Jason Wang, 2022/07/20
- [PULL V2 20/25] vdpa: Add device migration blocker, Jason Wang, 2022/07/20
- [PULL V2 21/25] vdpa: Add x-svq to NetdevVhostVDPAOptions, Jason Wang, 2022/07/20
- [PULL V2 23/25] net/colo: Fix a "double free" crash to clear the conn_list, Jason Wang, 2022/07/20
- [PULL V2 24/25] net/colo.c: No need to track conn_list for filter-rewriter, Jason Wang, 2022/07/20
- [PULL V2 25/25] net/colo.c: fix segmentation fault when packet is not parsed correctly, Jason Wang, 2022/07/20
- Re: [PULL V2 25/25] net/colo.c: fix segmentation fault when packet is not parsed correctly,
Peter Maydell <=
- Re: [PULL V2 00/25] Net patches, Peter Maydell, 2022/07/20