[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1448854142.24696.109.camel@edumazet-glaptop2.roam.corp.google.com>
Date: Sun, 29 Nov 2015 19:29:02 -0800
From: Eric Dumazet <eric.dumazet@...il.com>
To: Dmitry Vyukov <dvyukov@...gle.com>
Cc: Vlad Yasevich <vyasevich@...il.com>,
"David S. Miller" <davem@...emloft.net>,
Alexey Kuznetsov <kuznet@....inr.ac.ru>,
James Morris <jmorris@...ei.org>,
Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
Patrick McHardy <kaber@...sh.net>,
netdev <netdev@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>,
Eric Dumazet <edumazet@...gle.com>,
William Dauchy <wdauchy@...il.com>,
Rainer Weikusat <rweikusat@...ileactivedefense.com>,
syzkaller <syzkaller@...glegroups.com>,
Kostya Serebryany <kcc@...gle.com>,
Alexander Potapenko <glider@...gle.com>,
Sasha Levin <sasha.levin@...cle.com>
Subject: Re: use-after-free in ip6_setup_cork
On Sat, 2015-11-28 at 09:23 -0800, Eric Dumazet wrote:
> On Sat, 2015-11-28 at 09:11 -0800, Eric Dumazet wrote:
> > On Sat, 2015-11-28 at 12:00 +0100, Dmitry Vyukov wrote:
> > > Hello,
> > >
> > > The following program triggers use-after-free in ip6_setup_cork:
> > >
> > > // autogenerated by syzkaller (http://github.com/google/syzkaller)
> > > #include <syscall.h>
> > > #include <string.h>
> > > #include <stdint.h>
> > > #include <pthread.h>
> > >
> > > int r1, r3, r4;
> > >
> > > void *thr0(void *arg)
> > > {
> > > *(uint64_t*)0x20000d90 = 0x20000fd3;
> > > *(uint64_t*)0x20000d98 = 0x2d;
> > > *(uint64_t*)0x20000da0 = 0x20000fa4;
> > > *(uint64_t*)0x20000da8 = 0x5c;
> > > *(uint64_t*)0x20000db0 = 0x20000fac;
> > > *(uint64_t*)0x20000db8 = 0x71;
> > > *(uint64_t*)0x20000dc0 = 0x20000fb6;
> > > *(uint64_t*)0x20000dc8 = 0xec;
> > > *(uint64_t*)0x20000dd0 = 0x20000fae;
> > > *(uint64_t*)0x20000dd8 = 0x70;
> > > syscall(SYS_vmsplice, r4, 0x20000d90ul, 0x5ul, 0x2ul, 0, 0);
> > > return 0;
> > > }
> > >
> > > void *thr1(void *arg)
> > > {
> > > memcpy((void*)0x200025e5, "\xbb\xef\x44\xd6\x33\x93", 6);
> > > syscall(SYS_setsockopt, r1, 0x29ul, 0x6ul, 0x200025e5ul, 0x6ul, 0);
> > > return 0;
> > > }
> > >
> > > void *thr2(void *arg)
> > > {
> > > syscall(SYS_splice, r3, 0x0ul, r1, 0x0ul, 0xaful, 0x3ul);
> > > return 0;
> > > }
> > >
> > > int main()
> > > {
> > > syscall(SYS_mmap, 0x20000000ul, 0x10000ul, 0x3ul, 0x32ul,
> > > 0xfffffffffffffffful, 0x0ul);
> > > r1 = syscall(SYS_socket, 0xaul, 0x80002ul, 0x0ul, 0, 0, 0);
> > > syscall(SYS_pipe2, 0x20001000ul, 0x800ul, 0, 0, 0, 0);
> > > r3 = *(uint32_t*)0x20001000;
> > > r4 = *(uint32_t*)0x20001004;
> > > memcpy((void*)0x20003000,
> > > "\x0a\x00\x33\xe2\x61\x44\xfe\x90\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\x5b\x11\x6d\x22",
> > > 28);
> > > long r6 = syscall(SYS_connect, r1, 0x20003000ul, 0x1cul, 0, 0, 0);
> > > pthread_t th[4];
> > > pthread_create(&th[0], 0, thr0, 0);
> > > pthread_create(&th[1], 0, thr1, 0);
> > > pthread_create(&th[2], 0, thr2, 0);
> > > pthread_create(&th[3], 0, thr1, 0);
> > > pthread_join(th[0], 0);
> > > pthread_join(th[1], 0);
> > > pthread_join(th[2], 0);
> > > pthread_join(th[3], 0);
> > > return 0;
> > > }
> > >
> > >
> > > ==================================================================
> > > BUG: KASAN: use-after-free in ip6_setup_cork+0xeb8/0x11a0 at addr
> > > ffff88006d36da08
> > > Read of size 4 by task executor/23958
> > > =============================================================================
> > > BUG kmalloc-64 (Not tainted): kasan: bad access detected
> > > -----------------------------------------------------------------------------
> > >
> > > Disabling lock debugging due to kernel taint
> > > INFO: Allocated in sock_kmalloc+0x7f/0xc0 age=9 cpu=3 pid=23946
> > > [< none >] ___slab_alloc+0x41c/0x460 mm/slub.c:2438
> > > [< none >] __slab_alloc+0x1b/0x30 mm/slub.c:2467
> > > [< inline >] slab_alloc_node mm/slub.c:2530
> > > [< inline >] slab_alloc mm/slub.c:2572
> > > [< none >] __kmalloc+0x156/0x1b0 mm/slub.c:3532
> > > [< inline >] kmalloc include/linux/slab.h:463
> > > [< none >] sock_kmalloc+0x7f/0xc0 net/core/sock.c:1774
> > > [< none >] do_ipv6_setsockopt.isra.8+0x779/0x2a60
> > > net/ipv6/ipv6_sockglue.c:483
> > > [< none >] ipv6_setsockopt+0x9b/0x140 net/ipv6/ipv6_sockglue.c:885
> > > [< none >] udpv6_setsockopt+0x45/0x80 net/ipv6/udp.c:1425
> > > [< none >] sock_common_setsockopt+0x95/0xd0 net/core/sock.c:2645
> > > [< inline >] SYSC_setsockopt net/socket.c:1757
> > > [< none >] SyS_setsockopt+0x158/0x240 net/socket.c:1736
> > > [< none >] entry_SYSCALL_64_fastpath+0x12/0x71
> > > arch/x86/entry/entry_64.S:185
> > >
> > > INFO: Freed in sock_kfree_s+0x29/0x70 age=11 cpu=2 pid=23957
> > > [< none >] __slab_free+0x1fb/0x300 mm/slub.c:2648
> > > [< inline >] slab_free mm/slub.c:2803
> > > [< none >] kfree+0x13b/0x160 mm/slub.c:3632
> > > [< inline >] __sock_kfree_s net/core/sock.c:1795
> > > [< none >] sock_kfree_s+0x29/0x70 net/core/sock.c:1801
> > > [< none >] do_ipv6_setsockopt.isra.8+0x815/0x2a60
> > > net/ipv6/ipv6_sockglue.c:506
> > > [< none >] ipv6_setsockopt+0x9b/0x140 net/ipv6/ipv6_sockglue.c:885
> > > [< none >] udpv6_setsockopt+0x45/0x80 net/ipv6/udp.c:1425
> > > [< none >] sock_common_setsockopt+0x95/0xd0 net/core/sock.c:2645
> > > [< inline >] SYSC_setsockopt net/socket.c:1757
> > > [< none >] SyS_setsockopt+0x158/0x240 net/socket.c:1736
> > > [< none >] entry_SYSCALL_64_fastpath+0x12/0x71
> > > arch/x86/entry/entry_64.S:185
> > >
> > > INFO: Slab 0xffffea0001b4db00 objects=20 used=8 fp=0xffff88006d36da08
> > > flags=0x500000000004080
> > > INFO: Object 0xffff88006d36da08 @offset=6664 fp=0xffff88006d36d3e8
> > > CPU: 3 PID: 23958 Comm: executor Tainted: G B 4.4.0-rc2+ #1
> > > Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
> > > 00000000ffffffff ffff88003c32f348 ffffffff81a2bdb0 ffff88003e807840
> > > ffff88006d36da08 ffff88006d36c000 ffff88003c32f378 ffffffff814541d4
> > > ffff88003e807840 ffffea0001b4db00 ffff88006d36da08 ffff88006d126000
> > >
> > > Call Trace:
> > > [<ffffffff8145b9ae>] __asan_report_load4_noabort+0x3e/0x40
> > > mm/kasan/report.c:279
> > > [<ffffffff829de1e8>] ip6_setup_cork+0xeb8/0x11a0 net/ipv6/ip6_output.c:1200
> > > [<ffffffff829e9e2a>] ip6_make_skb+0x19a/0x3b0 net/ipv6/ip6_output.c:1769
> > > [<ffffffff82a3c3e6>] udpv6_sendmsg+0x11c6/0x2120 net/ipv6/udp.c:1314
> > > [<ffffffff829038cc>] inet_sendmsg+0x23c/0x340 net/ipv4/af_inet.c:733
> > > [< inline >] sock_sendmsg_nosec net/socket.c:610
> > > [<ffffffff826adc2a>] sock_sendmsg+0xca/0x110 net/socket.c:620
> > > [<ffffffff826ae057>] kernel_sendmsg+0x47/0x60 net/socket.c:628
> > > [<ffffffff826b564a>] sock_no_sendpage+0xfa/0x130 net/core/sock.c:2270
> > > [<ffffffff826ac6d0>] kernel_sendpage+0x90/0xe0 net/socket.c:3278
> > > [<ffffffff826ac7c5>] sock_sendpage+0xa5/0xd0 net/socket.c:765
> > > [<ffffffff814feca4>] pipe_to_sendpage+0x264/0x320 fs/splice.c:720
> > > [< inline >] splice_from_pipe_feed fs/splice.c:772
> > > [<ffffffff81501435>] __splice_from_pipe+0x235/0x6d0 fs/splice.c:897
> > > [<ffffffff81504677>] splice_from_pipe+0xf7/0x140 fs/splice.c:932
> > > [<ffffffff81504700>] generic_splice_sendpage+0x40/0x50 fs/splice.c:1105
> > > [< inline >] do_splice_from fs/splice.c:1124
> > > [< inline >] do_splice fs/splice.c:1400
> > > [< inline >] SYSC_splice fs/splice.c:1703
> > > [<ffffffff815052f8>] SyS_splice+0x7c8/0x15c0 fs/splice.c:1686
> > > [<ffffffff82ddc0ae>] entry_SYSCALL_64_fastpath+0x12/0x71
> > > arch/x86/entry/entry_64.S:185
> > > ==================================================================
> > >
> > >
> > > On commit 78c4a49a69e910a162b05e4e8727b9bdbf948f13 (Nov 25).
> > > --
> >
> > Thanks for the report.
> >
> > Bug probably added in :
> >
> > commit 03485f2adcde0c2d4e9228b659be78e872486bbb
> > Author: Vlad Yasevich <vyasevich@...il.com>
> > Date: Sat Jan 31 10:40:17 2015 -0500
> >
> > udpv6: Add lockless sendmsg() support
> >
> > This commit adds the same functionaliy to IPv6 that
> > commit 903ab86d195cca295379699299c5fc10beba31c7
> > Author: Herbert Xu <herbert@...dor.apana.org.au>
> > Date: Tue Mar 1 02:36:48 2011 +0000
> >
> > udp: Add lockless transmit path
> >
> > added to IPv4.
> >
> > UDP transmit path can now run without a socket lock,
> > thus allowing multiple threads to send to a single socket
> > more efficiently.
> > This is only used when corking/MSG_MORE is not used.
> >
> > Signed-off-by: Vladislav Yasevich <vyasevic@...hat.com>
> > Signed-off-by: David S. Miller <davem@...emloft.net>
> >
> >
>
> Oh well, bug is older than that.
>
> inet6_sk(sk)->opt handling is completely racy, and has some xchg() calls
> to hide how ugly it is.
I have a patch that looks ready for submission.
While testing it, I also saw a problem in ipv6_update_options(),
calling sk_dst_reset() which directly conflicts with ip6_dst_store()
(ip6_dst_store() uses spin_lock(&sk->sk_dst_lock), which basically
protects councurrent ip6_dst_store() ... :( )
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists