Re: KASAN: slab-out-of-bounds Write in tcp_v6_syn_recv_sock

[Ozgur]

03.01.2018, 21:57, "Cong Wang" <xiyou.wangcong@xxxxxxxxx>:
> On Tue, Jan 2, 2018 at 3:58 PM, syzbot
> <syzbot+6dc95bddc6976b800b0b@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>> ÂHello,
>>
>> Âsyzkaller hit the following crash on
>> Â61233580f1f33c50e159c50e24d80ffd2ba2e06b
>> Âgit://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/master
>> Âcompiler: gcc (GCC) 7.1.1 20170620
>> Â.config is attached
>> ÂRaw console output is attached.
>> ÂC reproducer is attached
>> Âsyzkaller reproducer is attached. See https://goo.gl/kgGztJ
>> Âfor information about syzkaller reproducers
>>
>> ÂIMPORTANT: if you fix the bug, please add the following tag to the commit:
>> ÂReported-by: syzbot+6dc95bddc6976b800b0b@xxxxxxxxxxxxxxxxxxxxxxxxx
>> ÂIt will help syzbot understand when the bug is fixed. See footer for
>> Âdetails.
>> ÂIf you forward the report, please keep this part and the footer.
>>
>> ÂTCP: request_sock_TCPv6: Possible SYN flooding on port 20002. Sending
>> Âcookies. Check SNMP counters.
>> Â==================================================================
>> ÂBUG: KASAN: slab-out-of-bounds in memcpy include/linux/string.h:344 [inline]
>> ÂBUG: KASAN: slab-out-of-bounds in tcp_v6_syn_recv_sock+0x628/0x23a0
>> Ânet/ipv6/tcp_ipv6.c:1144
>> ÂWrite of size 160 at addr ffff8801cbdd7460 by task syzkaller545407/3196
>>
>> ÂCPU: 1 PID: 3196 Comm: syzkaller545407 Not tainted 4.15.0-rc5+ #241
>> ÂHardware name: Google Google Compute Engine/Google Compute Engine, BIOS
>> ÂGoogle 01/01/2011
>> ÂCall Trace:
>> ÂÂ<IRQ>
>> ÂÂ__dump_stack lib/dump_stack.c:17 [inline]
>> ÂÂdump_stack+0x194/0x257 lib/dump_stack.c:53
>> ÂÂprint_address_description+0x73/0x250 mm/kasan/report.c:252
>> ÂÂkasan_report_error mm/kasan/report.c:351 [inline]
>> ÂÂkasan_report+0x25b/0x340 mm/kasan/report.c:409
>> ÂÂcheck_memory_region_inline mm/kasan/kasan.c:260 [inline]
>> ÂÂcheck_memory_region+0x137/0x190 mm/kasan/kasan.c:267
>> ÂÂmemcpy+0x37/0x50 mm/kasan/kasan.c:303
>> ÂÂmemcpy include/linux/string.h:344 [inline]
>> ÂÂtcp_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

>> ÂÂtcp_get_cookie_sock+0x102/0x540 net/ipv4/syncookies.c:213
>> ÂÂcookie_v6_check+0x177d/0x2160 net/ipv6/syncookies.c:255
>> ÂÂtcp_v6_cookie_check net/ipv6/tcp_ipv6.c:1008 [inline]
>> ÂÂtcp_v6_do_rcv+0xe4d/0x11c0 net/ipv6/tcp_ipv6.c:1316
>> ÂÂtcp_v6_rcv+0x22ee/0x2b40 net/ipv6/tcp_ipv6.c:1510
>> ÂÂip6_input_finish+0x36f/0x1700 net/ipv6/ip6_input.c:284
>> ÂÂNF_HOOK include/linux/netfilter.h:250 [inline]
>> ÂÂip6_input+0xe9/0x560 net/ipv6/ip6_input.c:327
>> ÂÂdst_input include/net/dst.h:466 [inline]
>> ÂÂip6_rcv_finish+0x1a9/0x7a0 net/ipv6/ip6_input.c:71
>> ÂÂNF_HOOK include/linux/netfilter.h:250 [inline]
>> ÂÂipv6_rcv+0xf1f/0x1f80 net/ipv6/ip6_input.c:208
>> ÂÂ__netif_receive_skb_core+0x1a3e/0x3450 net/core/dev.c:4461
>> ÂÂ__netif_receive_skb+0x2c/0x1b0 net/core/dev.c:4526
>> ÂÂprocess_backlog+0x203/0x740 net/core/dev.c:5205
>> ÂÂnapi_poll net/core/dev.c:5603 [inline]
>> ÂÂnet_rx_action+0x792/0x1910 net/core/dev.c:5669
>> ÂÂ__do_softirq+0x2d7/0xb85 kernel/softirq.c:285
>> ÂÂdo_softirq_own_stack+0x2a/0x40 arch/x86/entry/entry_64.S:1115
>> ÂÂ</IRQ>
>> ÂÂdo_softirq.part.21+0x14d/0x190 kernel/softirq.c:329
>> ÂÂdo_softirq kernel/softirq.c:177 [inline]
>> ÂÂ__local_bh_enable_ip+0x1ee/0x230 kernel/softirq.c:182
>> ÂÂlocal_bh_enable include/linux/bottom_half.h:32 [inline]
>> ÂÂrcu_read_unlock_bh include/linux/rcupdate.h:727 [inline]
>> ÂÂip6_finish_output2+0xba6/0x2390 net/ipv6/ip6_output.c:121
>> ÂÂip6_finish_output+0x2f9/0x920 net/ipv6/ip6_output.c:146
>> ÂÂNF_HOOK_COND include/linux/netfilter.h:239 [inline]
>> ÂÂip6_output+0x1eb/0x840 net/ipv6/ip6_output.c:163
>> ÂÂdst_output include/net/dst.h:460 [inline]
>> ÂÂNF_HOOK include/linux/netfilter.h:250 [inline]
>> ÂÂip6_xmit+0xd75/0x2080 net/ipv6/ip6_output.c:269
>> ÂÂinet6_csk_xmit+0x2fc/0x580 net/ipv6/inet6_connection_sock.c:139
>> ÂÂtcp_transmit_skb+0x1b12/0x38b0 net/ipv4/tcp_output.c:1176
>> ÂÂtcp_write_xmit+0x680/0x5190 net/ipv4/tcp_output.c:2367
>> ÂÂ__tcp_push_pending_frames+0xa0/0x250 net/ipv4/tcp_output.c:2543
>> ÂÂtcp_send_fin+0x1b0/0xd20 net/ipv4/tcp_output.c:3087
>> ÂÂtcp_close+0xbe0/0xfc0 net/ipv4/tcp.c:2234
>> ÂÂinet_release+0xed/0x1c0 net/ipv4/af_inet.c:426
>> ÂÂinet6_release+0x50/0x70 net/ipv6/af_inet6.c:432
>> ÂÂsock_release+0x8d/0x1e0 net/socket.c:600
>> ÂÂsock_close+0x16/0x20 net/socket.c:1129
>> ÂÂ__fput+0x327/0x7e0 fs/file_table.c:210
>> ÂÂ____fput+0x15/0x20 fs/file_table.c:244
>> ÂÂtask_work_run+0x199/0x270 kernel/task_work.c:113
>> ÂÂexit_task_work include/linux/task_work.h:22 [inline]
>> ÂÂdo_exit+0x9bb/0x1ad0 kernel/exit.c:865
>> ÂÂdo_group_exit+0x149/0x400 kernel/exit.c:968
>> ÂÂget_signal+0x73f/0x16c0 kernel/signal.c:2335
>> ÂÂdo_signal+0x94/0x1ee0 arch/x86/kernel/signal.c:809
>> ÂÂexit_to_usermode_loop+0x214/0x310 arch/x86/entry/common.c:158
>> ÂÂprepare_exit_to_usermode arch/x86/entry/common.c:195 [inline]
>> ÂÂsyscall_return_slowpath+0x490/0x550 arch/x86/entry/common.c:264
>> ÂÂentry_SYSCALL_64_fastpath+0x94/0x96
>> ÂRIP: 0033:0x4456e9
>> ÂRSP: 002b:00007fb4de631da8 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
>> ÂRAX: fffffffffffffe00 RBX: 00000000006dac3c RCX: 00000000004456e9
>> ÂRDX: 0000000000000000 RSI: 0000000000000000 RDI: 00000000006dac3c
>> ÂRBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
>> ÂR10: 0000000000000000 R11: 0000000000000246 R12: 00000000006dac38
>> ÂR13: 0100000000000000 R14: 00007fb4de6329c0 R15: 0000000000000009
>>
>> ÂAllocated by task 3196:
>> ÂÂsave_stack+0x43/0xd0 mm/kasan/kasan.c:447
>> ÂÂset_track mm/kasan/kasan.c:459 [inline]
>> ÂÂkasan_kmalloc+0xad/0xe0 mm/kasan/kasan.c:551
>> ÂÂkasan_slab_alloc+0x12/0x20 mm/kasan/kasan.c:489
>> ÂÂkmem_cache_alloc+0x12e/0x760 mm/slab.c:3544
>> ÂÂsk_prot_alloc+0x65/0x2a0 net/core/sock.c:1463
>> ÂÂsk_clone_lock+0x152/0x1570 net/core/sock.c:1649
>> ÂÂinet_csk_clone_lock+0x92/0x4f0 net/ipv4/inet_connection_sock.c:781
>> ÂÂtcp_create_openreq_child+0x9b/0x1b70 net/ipv4/tcp_minisocks.c:449
>> ÂÂtcp_v6_syn_recv_sock+0x22d/0x23a0 net/ipv6/tcp_ipv6.c:1123
>> ÂÂtcp_get_cookie_sock+0x102/0x540 net/ipv4/syncookies.c:213
>> ÂÂcookie_v6_check+0x177d/0x2160 net/ipv6/syncookies.c:255
>> ÂÂtcp_v6_cookie_check net/ipv6/tcp_ipv6.c:1008 [inline]
>> ÂÂtcp_v6_do_rcv+0xe4d/0x11c0 net/ipv6/tcp_ipv6.c:1316
>> ÂÂtcp_v6_rcv+0x22ee/0x2b40 net/ipv6/tcp_ipv6.c:1510
>> ÂÂip6_input_finish+0x36f/0x1700 net/ipv6/ip6_input.c:284
>> ÂÂNF_HOOK include/linux/netfilter.h:250 [inline]
>> ÂÂip6_input+0xe9/0x560 net/ipv6/ip6_input.c:327
>> ÂÂdst_input include/net/dst.h:466 [inline]
>> ÂÂip6_rcv_finish+0x1a9/0x7a0 net/ipv6/ip6_input.c:71
>> ÂÂNF_HOOK include/linux/netfilter.h:250 [inline]
>> ÂÂipv6_rcv+0xf1f/0x1f80 net/ipv6/ip6_input.c:208
>> ÂÂ__netif_receive_skb_core+0x1a3e/0x3450 net/core/dev.c:4461
>> ÂÂ__netif_receive_skb+0x2c/0x1b0 net/core/dev.c:4526
>> ÂÂprocess_backlog+0x203/0x740 net/core/dev.c:5205
>> ÂÂnapi_poll net/core/dev.c:5603 [inline]
>> ÂÂnet_rx_action+0x792/0x1910 net/core/dev.c:5669
>> ÂÂ__do_softirq+0x2d7/0xb85 kernel/softirq.c:285
>>
>> ÂFreed by task 0:
>> Â(stack is not available)
>>
>> ÂThe buggy address belongs to the object at ffff8801cbdd6a80
>> ÂÂwhich belongs to the cache TCP of size 2528
>> ÂThe buggy address is located 0 bytes to the right of
>> ÂÂ2528-byte region [ffff8801cbdd6a80, ffff8801cbdd7460)
>> ÂThe buggy address belongs to the page:
>> Âpage:000000006145927c count:1 mapcount:0 mapping:00000000d41dd7c1
>> Âindex:0xffff8801cbdd7ffd compound_mapcount: 0
>> Âflags: 0x2fffc0000008100(slab|head)
>> Âraw: 02fffc0000008100 ffff8801cbdd6000 ffff8801cbdd7ffd 0000000100000003
>> Âraw: ffffea00074ef120 ffff8801d82b7248 ffff8801d798b640 0000000000000000
>> Âpage dumped because: kasan: bad access detected
>>
>> ÂMemory state around the buggy address:
>> ÂÂffff8801cbdd7300: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>> ÂÂffff8801cbdd7380: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>> Âffff8801cbdd7400: 00 00 00 00 00 00 00 00 00 00 00 00 fc fc fc fc
>>
>> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ^
>> ÂÂffff8801cbdd7480: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>> ÂÂffff8801cbdd7500: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>> Â==================================================================
>>
>> Â---
>> ÂThis bug is generated by a dumb bot. It may contain errors.
>> ÂSee https://goo.gl/tpsmEJ for details.
>> ÂDirect all questions to syzkaller@xxxxxxxxxxxxxxxxx
>>
>> Âsyzbot will keep track of this bug report.
>> ÂIf you forgot to add the Reported-by tag, once the fix for this bug is
>> Âmerged
>> Âinto any tree, please reply to this email with:
>> Â#syz fix: exact-commit-title
>> ÂIf you want to test a patch for this bug, please reply with:
>> Â#syz test: git://repo/address.git branch
>> Âand provide the patch inline or as an attachment.
>> ÂTo mark this as a duplicate of another syzbot report, please reply with:
>> Â#syz dup: exact-subject-of-another-report
>> ÂIf it's a one-off invalid bug report, please reply with:
>> Â#syz invalid
>> ÂNote: if the crash happens again, it will cause creation of a new bug
>> Âreport.
>> ÂNote: all commands must start from beginning of the line in the email body.