Re: net: GPF in eth_header

[Eric Dumazet]

On Sat, 2016-11-26 at 20:07 +0100, Andrey Konovalov wrote:
> On Sat, Nov 26, 2016 at 7:28 PM, 'Eric Dumazet' via syzkaller
> <syzkaller@xxxxxxxxxxxxxxxx> wrote:
> > On Sat, Nov 26, 2016 at 9:30 AM, Dmitry Vyukov <dvyukov@xxxxxxxxxx> wrote:
> >> Hello,
> >>
> >> The following program triggers GPF in eth_header:
> >>
> >> https://gist.githubusercontent.com/dvyukov/613cadf05543b55a419f237e419cd495/raw/5471231523d1a07c3de55f11f87472c2816ee06c/gistfile1.txt
> >>
> >> On commit 16ae16c6e5616c084168740990fc508bda6655d4 (Nov 24).
> >>
> >> BUG: unable to handle kernel paging request at ffffed002d14d74a
> >> IP: [<ffffffff86be3295>] eth_header+0x75/0x260 net/ethernet/eth.c:88
> >> PGD 7fff6067 [   50.787819] PUD 7fff5067
> >> PMD 0 [   50.787819]
> >> Oops: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN
> >> Modules linked in:
> >> CPU: 2 PID: 6712 Comm: a.out Not tainted 4.9.0-rc6+ #55
> >> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
> >> task: ffff88003a1841c0 task.stack: ffff880034d08000
> >> RIP: 0010:[<ffffffff86be3295>]  [<ffffffff86be3295>]
> >> eth_header+0x75/0x260 net/ethernet/eth.c:88
> >> RSP: 0018:ffff880034d0eb68  EFLAGS: 00010a03
> >> RAX: 1ffff1002d14d74a RBX: ffff880168a6ba4a RCX: ffff88006a9c7858
> >> RDX: 000000000000dd86 RSI: dffffc0000000000 RDI: ffff880168a6ba56
> >> RBP: ffff880034d0eb98 R08: 0000000000000000 R09: 0000000000000031
> >> R10: 0000000000000002 R11: 0000000000000000 R12: 0000000000000000
> >> R13: ffff88006c208d80 R14: 00000000000086dd R15: ffff88006a9c7858
> >> FS:  0000000001a02940(0000) GS:ffff88006d000000(0000) knlGS:0000000000000000
> >> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >> CR2: ffffed002d14d74a CR3: 0000000037373000 CR4: 00000000000006e0
> >> Stack:
> >>  000000316881ab40 ffff88006a9c76c0 ffff88006881ab40 ffff88006a9c77f8
> >>  0000000000000000 dffffc0000000000 ffff880034d0ee98 ffffffff86b31af9
> >>  ffffffff8719605c ffff880034d0f0f8 ffffffff000086dd ffffffff86be3220
> >> Call Trace:
> >>  [<     inline     >] dev_hard_header ./include/linux/netdevice.h:2762
> >>  [<ffffffff86b31af9>] neigh_resolve_output+0x659/0xb20 net/core/neighbour.c:1302
> >>  [<     inline     >] dst_neigh_output ./include/net/dst.h:464
> >>  [<ffffffff8719605c>] ip6_finish_output2+0xb3c/0x2500 net/ipv6/ip6_output.c:121
> >>  [<ffffffff871a0b0b>] ip6_finish_output+0x2eb/0x760 net/ipv6/ip6_output.c:139
> >>  [<     inline     >] NF_HOOK_COND ./include/linux/netfilter.h:246
> >>  [<ffffffff871a1157>] ip6_output+0x1d7/0x9a0 net/ipv6/ip6_output.c:153
> >>  [<     inline     >] dst_output ./include/net/dst.h:501
> >>  [<ffffffff873312ea>] ip6_local_out+0x9a/0x180 net/ipv6/output_core.c:170
> >>  [<ffffffff871a3886>] ip6_send_skb+0xa6/0x340 net/ipv6/ip6_output.c:1712
> >>  [<ffffffff871a3bd8>] ip6_push_pending_frames+0xb8/0xe0
> >> net/ipv6/ip6_output.c:1732
> >>  [<     inline     >] rawv6_push_pending_frames net/ipv6/raw.c:607
> >>  [<ffffffff8722acfb>] rawv6_sendmsg+0x250b/0x2c20 net/ipv6/raw.c:920
> >>  [<ffffffff8701c4f5>] inet_sendmsg+0x385/0x590 net/ipv4/af_inet.c:734
> >>  [<     inline     >] sock_sendmsg_nosec net/socket.c:621
> >>  [<ffffffff86a6ea9f>] sock_sendmsg+0xcf/0x110 net/socket.c:631
> >>  [<ffffffff86a6ee0b>] sock_write_iter+0x32b/0x620 net/socket.c:829
> >>  [<ffffffff81a6f153>] do_iter_readv_writev+0x363/0x670 fs/read_write.c:695
> >>  [<ffffffff81a71ba1>] do_readv_writev+0x431/0x9b0 fs/read_write.c:872
> >>  [<ffffffff81a726dc>] vfs_writev+0x8c/0xc0 fs/read_write.c:911
> >>  [<ffffffff81a72825>] do_writev+0x115/0x2d0 fs/read_write.c:944
> >>  [<     inline     >] SYSC_writev fs/read_write.c:1017
> >>  [<ffffffff81a75fdc>] SyS_writev+0x2c/0x40 fs/read_write.c:1014
> >>  [<ffffffff8814cf85>] entry_SYSCALL_64_fastpath+0x23/0xc6
> >> arch/x86/entry/entry_64.S:209
> >> Code: 41 83 fe 04 0f 84 aa 00 00 00 e8 17 4e b0 fa 48 8d 7b 0c 48 be
> >> 00 00 00 00 00 fc ff df 44 89 f2 66 c1 c2 08 48 89 f8 48 c1 e8 03 <0f>
> >> b6 0c 30 48 8d 43 0d 49 89 c0 49 c1 e8 03 41 0f b6 34 30 49
> >> RIP  [<ffffffff86be3295>] eth_header+0x75/0x260 net/ethernet/eth.c:88
> >>  RSP <ffff880034d0eb68>
> >> CR2: ffffed002d14d74a
> >> ---[ end trace a73fedfdc11bd60c ]---
> >
> >
> > Hi Dmitry
> >
> > I could not reproduce the issue. Might need some specific configuration...
> >
> > loopback device has proper ethernet header (all 0)
> >
> > Fault happens in :
> >
> > 0f b6 0c 30             movzbl (%rax,%rsi,1),%ecx
> >
> > RAX=1ffff1002d14d74a  which is RDI>>3, and RSI=dffffc0000000000
> >
> > Could this be a KASAN problem ?
> 
> Hi Eric,
> 
> The crash happens when the kernel tries to access shadow for nonmapped memory.
> 
> The issue here is an integer overflow which happens in neigh_resolve_output().
> skb_network_offset(skb) can return negative number, but __skb_pull()
> accepts unsigned int as len.
> As a result, the least significat bit in higher 32 bits of skb->data
> gets set and we get an out-of-bounds with offset of 4 GB.
> 
> I've attached a short reproducer, but you either need KASAN or to add
> a BUG_ON to see the crash.
> In this reproducer skb_network_offset() becomes negative after merging
> two ipv6 fragments.
> 
> I actually see multiple places where skb_network_offset() is used as
> an argument to skb_pull().
> So I guess every place can potentially be buggy.
> 
> Thanks!

I can not reproduce the bug on my hosts.
Quite hard to debug for me.

skb_network_offset() can not be negative at this point, unless there is
a bug upper in the stack.

Hannes, do you have an idea of what could be wrong in IPv6 stack ?

Thanks.