Message-Id: <681471515012954@web19j.yandex.ru>
Date:   Wed, 03 Jan 2018 23:55:54 +0300
From:   Ozgur <ozgur@...sey.org>
To:     Cong Wang <xiyou.wangcong@...il.com>,
        syzbot <syzbot+6dc95bddc6976b800b0b@...kaller.appspotmail.com>
Cc:     David Miller <davem@...emloft.net>,
        Alexey Kuznetsov <kuznet@....inr.ac.ru>,
        LKML <linux-kernel@...r.kernel.org>,
        Linux Kernel Network Developers <netdev@...r.kernel.org>,
        "syzkaller-bugs@...glegroups.com" <syzkaller-bugs@...glegroups.com>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        "avejwatson@...com" <avejwatson@...com>,
        "ilyal@...lanox.com" <ilyal@...lanox.com>,
        "aviadye@...lanox.com" <aviadye@...lanox.com>
Subject: Re: KASAN: slab-out-of-bounds Write in tcp_v6_syn_recv_sock



03.01.2018, 21:57, "Cong Wang" <xiyou.wangcong@...il.com>:
> On Tue, Jan 2, 2018 at 3:58 PM, syzbot
> <syzbot+6dc95bddc6976b800b0b@...kaller.appspotmail.com> wrote:
>>  Hello,
>>
>>  syzkaller hit the following crash on
>>  61233580f1f33c50e159c50e24d80ffd2ba2e06b
>>  git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/master
>>  compiler: gcc (GCC) 7.1.1 20170620
>>  .config is attached
>>  Raw console output is attached.
>>  C reproducer is attached
>>  syzkaller reproducer is attached. See https://goo.gl/kgGztJ
>>  for information about syzkaller reproducers
>>
>>  IMPORTANT: if you fix the bug, please add the following tag to the commit:
>>  Reported-by: syzbot+6dc95bddc6976b800b0b@...kaller.appspotmail.com
>>  It will help syzbot understand when the bug is fixed. See footer for
>>  details.
>>  If you forward the report, please keep this part and the footer.
>>
>>  TCP: request_sock_TCPv6: Possible SYN flooding on port 20002. Sending
>>  cookies. Check SNMP counters.
>>  ==================================================================
>>  BUG: KASAN: slab-out-of-bounds in memcpy include/linux/string.h:344 [inline]
>>  BUG: KASAN: slab-out-of-bounds in tcp_v6_syn_recv_sock+0x628/0x23a0
>>  net/ipv6/tcp_ipv6.c:1144
>>  Write of size 160 at addr ffff8801cbdd7460 by task syzkaller545407/3196
>>
>>  CPU: 1 PID: 3196 Comm: syzkaller545407 Not tainted 4.15.0-rc5+ #241
>>  Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
>>  Google 01/01/2011
>>  Call Trace:
>>   <IRQ>
>>   __dump_stack lib/dump_stack.c:17 [inline]
>>   dump_stack+0x194/0x257 lib/dump_stack.c:53
>>   print_address_description+0x73/0x250 mm/kasan/report.c:252
>>   kasan_report_error mm/kasan/report.c:351 [inline]
>>   kasan_report+0x25b/0x340 mm/kasan/report.c:409
>>   check_memory_region_inline mm/kasan/kasan.c:260 [inline]
>>   check_memory_region+0x137/0x190 mm/kasan/kasan.c:267
>>   memcpy+0x37/0x50 mm/kasan/kasan.c:303
>>   memcpy include/linux/string.h:344 [inline]
>>   tcp_v6_syn_recv_sock+0x628/0x23a0 net/ipv6/tcp_ipv6.c:1144
>
> tls_init() changes sk->sk_prot from IPv6 to IPv4, which leads
> to this bug. I guess IPv6 is not supported for TLS? If so, need
> a check on proto in tls_init()...

Hello,

I think IPv6 supports with TLS.
There was a previously posted commit by Mellanox:

https://patchwork.ozlabs.org/patch/801530/

Ozgur

>>   tcp_get_cookie_sock+0x102/0x540 net/ipv4/syncookies.c:213
>>   cookie_v6_check+0x177d/0x2160 net/ipv6/syncookies.c:255
>>   tcp_v6_cookie_check net/ipv6/tcp_ipv6.c:1008 [inline]
>>   tcp_v6_do_rcv+0xe4d/0x11c0 net/ipv6/tcp_ipv6.c:1316
>>   tcp_v6_rcv+0x22ee/0x2b40 net/ipv6/tcp_ipv6.c:1510
>>   ip6_input_finish+0x36f/0x1700 net/ipv6/ip6_input.c:284
>>   NF_HOOK include/linux/netfilter.h:250 [inline]
>>   ip6_input+0xe9/0x560 net/ipv6/ip6_input.c:327
>>   dst_input include/net/dst.h:466 [inline]
>>   ip6_rcv_finish+0x1a9/0x7a0 net/ipv6/ip6_input.c:71
>>   NF_HOOK include/linux/netfilter.h:250 [inline]
>>   ipv6_rcv+0xf1f/0x1f80 net/ipv6/ip6_input.c:208
>>   __netif_receive_skb_core+0x1a3e/0x3450 net/core/dev.c:4461
>>   __netif_receive_skb+0x2c/0x1b0 net/core/dev.c:4526
>>   process_backlog+0x203/0x740 net/core/dev.c:5205
>>   napi_poll net/core/dev.c:5603 [inline]
>>   net_rx_action+0x792/0x1910 net/core/dev.c:5669
>>   __do_softirq+0x2d7/0xb85 kernel/softirq.c:285
>>   do_softirq_own_stack+0x2a/0x40 arch/x86/entry/entry_64.S:1115
>>   </IRQ>
>>   do_softirq.part.21+0x14d/0x190 kernel/softirq.c:329
>>   do_softirq kernel/softirq.c:177 [inline]
>>   __local_bh_enable_ip+0x1ee/0x230 kernel/softirq.c:182
>>   local_bh_enable include/linux/bottom_half.h:32 [inline]
>>   rcu_read_unlock_bh include/linux/rcupdate.h:727 [inline]
>>   ip6_finish_output2+0xba6/0x2390 net/ipv6/ip6_output.c:121
>>   ip6_finish_output+0x2f9/0x920 net/ipv6/ip6_output.c:146
>>   NF_HOOK_COND include/linux/netfilter.h:239 [inline]
>>   ip6_output+0x1eb/0x840 net/ipv6/ip6_output.c:163
>>   dst_output include/net/dst.h:460 [inline]
>>   NF_HOOK include/linux/netfilter.h:250 [inline]
>>   ip6_xmit+0xd75/0x2080 net/ipv6/ip6_output.c:269
>>   inet6_csk_xmit+0x2fc/0x580 net/ipv6/inet6_connection_sock.c:139
>>   tcp_transmit_skb+0x1b12/0x38b0 net/ipv4/tcp_output.c:1176
>>   tcp_write_xmit+0x680/0x5190 net/ipv4/tcp_output.c:2367
>>   __tcp_push_pending_frames+0xa0/0x250 net/ipv4/tcp_output.c:2543
>>   tcp_send_fin+0x1b0/0xd20 net/ipv4/tcp_output.c:3087
>>   tcp_close+0xbe0/0xfc0 net/ipv4/tcp.c:2234
>>   inet_release+0xed/0x1c0 net/ipv4/af_inet.c:426
>>   inet6_release+0x50/0x70 net/ipv6/af_inet6.c:432
>>   sock_release+0x8d/0x1e0 net/socket.c:600
>>   sock_close+0x16/0x20 net/socket.c:1129
>>   __fput+0x327/0x7e0 fs/file_table.c:210
>>   ____fput+0x15/0x20 fs/file_table.c:244
>>   task_work_run+0x199/0x270 kernel/task_work.c:113
>>   exit_task_work include/linux/task_work.h:22 [inline]
>>   do_exit+0x9bb/0x1ad0 kernel/exit.c:865
>>   do_group_exit+0x149/0x400 kernel/exit.c:968
>>   get_signal+0x73f/0x16c0 kernel/signal.c:2335
>>   do_signal+0x94/0x1ee0 arch/x86/kernel/signal.c:809
>>   exit_to_usermode_loop+0x214/0x310 arch/x86/entry/common.c:158
>>   prepare_exit_to_usermode arch/x86/entry/common.c:195 [inline]
>>   syscall_return_slowpath+0x490/0x550 arch/x86/entry/common.c:264
>>   entry_SYSCALL_64_fastpath+0x94/0x96
>>  RIP: 0033:0x4456e9
>>  RSP: 002b:00007fb4de631da8 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
>>  RAX: fffffffffffffe00 RBX: 00000000006dac3c RCX: 00000000004456e9
>>  RDX: 0000000000000000 RSI: 0000000000000000 RDI: 00000000006dac3c
>>  RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
>>  R10: 0000000000000000 R11: 0000000000000246 R12: 00000000006dac38
>>  R13: 0100000000000000 R14: 00007fb4de6329c0 R15: 0000000000000009
>>
>>  Allocated by task 3196:
>>   save_stack+0x43/0xd0 mm/kasan/kasan.c:447
>>   set_track mm/kasan/kasan.c:459 [inline]
>>   kasan_kmalloc+0xad/0xe0 mm/kasan/kasan.c:551
>>   kasan_slab_alloc+0x12/0x20 mm/kasan/kasan.c:489
>>   kmem_cache_alloc+0x12e/0x760 mm/slab.c:3544
>>   sk_prot_alloc+0x65/0x2a0 net/core/sock.c:1463
>>   sk_clone_lock+0x152/0x1570 net/core/sock.c:1649
>>   inet_csk_clone_lock+0x92/0x4f0 net/ipv4/inet_connection_sock.c:781
>>   tcp_create_openreq_child+0x9b/0x1b70 net/ipv4/tcp_minisocks.c:449
>>   tcp_v6_syn_recv_sock+0x22d/0x23a0 net/ipv6/tcp_ipv6.c:1123
>>   tcp_get_cookie_sock+0x102/0x540 net/ipv4/syncookies.c:213
>>   cookie_v6_check+0x177d/0x2160 net/ipv6/syncookies.c:255
>>   tcp_v6_cookie_check net/ipv6/tcp_ipv6.c:1008 [inline]
>>   tcp_v6_do_rcv+0xe4d/0x11c0 net/ipv6/tcp_ipv6.c:1316
>>   tcp_v6_rcv+0x22ee/0x2b40 net/ipv6/tcp_ipv6.c:1510
>>   ip6_input_finish+0x36f/0x1700 net/ipv6/ip6_input.c:284
>>   NF_HOOK include/linux/netfilter.h:250 [inline]
>>   ip6_input+0xe9/0x560 net/ipv6/ip6_input.c:327
>>   dst_input include/net/dst.h:466 [inline]
>>   ip6_rcv_finish+0x1a9/0x7a0 net/ipv6/ip6_input.c:71
>>   NF_HOOK include/linux/netfilter.h:250 [inline]
>>   ipv6_rcv+0xf1f/0x1f80 net/ipv6/ip6_input.c:208
>>   __netif_receive_skb_core+0x1a3e/0x3450 net/core/dev.c:4461
>>   __netif_receive_skb+0x2c/0x1b0 net/core/dev.c:4526
>>   process_backlog+0x203/0x740 net/core/dev.c:5205
>>   napi_poll net/core/dev.c:5603 [inline]
>>   net_rx_action+0x792/0x1910 net/core/dev.c:5669
>>   __do_softirq+0x2d7/0xb85 kernel/softirq.c:285
>>
>>  Freed by task 0:
>>  (stack is not available)
>>
>>  The buggy address belongs to the object at ffff8801cbdd6a80
>>   which belongs to the cache TCP of size 2528
>>  The buggy address is located 0 bytes to the right of
>>   2528-byte region [ffff8801cbdd6a80, ffff8801cbdd7460)
>>  The buggy address belongs to the page:
>>  page:000000006145927c count:1 mapcount:0 mapping:00000000d41dd7c1
>>  index:0xffff8801cbdd7ffd compound_mapcount: 0
>>  flags: 0x2fffc0000008100(slab|head)
>>  raw: 02fffc0000008100 ffff8801cbdd6000 ffff8801cbdd7ffd 0000000100000003
>>  raw: ffffea00074ef120 ffff8801d82b7248 ffff8801d798b640 0000000000000000
>>  page dumped because: kasan: bad access detected
>>
>>  Memory state around the buggy address:
>>   ffff8801cbdd7300: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   ffff8801cbdd7380: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>>  ffff8801cbdd7400: 00 00 00 00 00 00 00 00 00 00 00 00 fc fc fc fc
>>
>>                                                         ^
>>   ffff8801cbdd7480: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>>   ffff8801cbdd7500: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>  ==================================================================
>>
>>  ---
>>  This bug is generated by a dumb bot. It may contain errors.
>>  See https://goo.gl/tpsmEJ for details.
>>  Direct all questions to syzkaller@...glegroups.com.
>>
>>  syzbot will keep track of this bug report.
>>  If you forgot to add the Reported-by tag, once the fix for this bug is
>>  merged
>>  into any tree, please reply to this email with:
>>  #syz fix: exact-commit-title
>>  If you want to test a patch for this bug, please reply with:
>>  #syz test: git://repo/address.git branch
>>  and provide the patch inline or as an attachment.
>>  To mark this as a duplicate of another syzbot report, please reply with:
>>  #syz dup: exact-subject-of-another-report
>>  If it's a one-off invalid bug report, please reply with:
>>  #syz invalid
>>  Note: if the crash happens again, it will cause creation of a new bug
>>  report.
>>  Note: all commands must start from beginning of the line in the email body.

Powered by blists - more mailing lists