Message-ID: <CACT4Y+YvFRCz7kbpDoq9dVghWQv6xD5a3_KZ0GXwxTxFx6i5xw@mail.gmail.com>
Date:   Sun, 5 Mar 2017 13:15:53 +0100
From:   Dmitry Vyukov <dvyukov@...gle.com>
To:     Matthew Wilcox <mawilcox@...rosoft.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        "ross.zwisler@...ux.intel.com" <ross.zwisler@...ux.intel.com>,
        Johannes Weiner <hannes@...xchg.org>, Jan Kara <jack@...e.cz>,
        LKML <linux-kernel@...r.kernel.org>, Tejun Heo <tj@...nel.org>,
        syzkaller <syzkaller@...glegroups.com>
Subject: Re: idr: use-after-free write in ida_get_new_above

On Fri, Mar 3, 2017 at 6:06 PM, Matthew Wilcox <mawilcox@...rosoft.com> wrote:
> Thanks!  I don't think it's syzcaller per se that's provoking this; it's a matter of an interrupt coming in at the right time and preempting a task that happens to be allocating from an IDA with another task that's also allocating from an IDA.  So any workload which makes heavy use of the kernel could provoke it.  I imagine syzcaller is exercising a lot of parts of the kernel which use IDAs.
>
> Funny that it didn't show up before now ... it's been in linux-next in this form since the start of February.
>
> I've created a test case which will reproduce this, and after applying this patch it no longer reproduces.  So I'm confident in your second patch :-)


Mailed the patch:
https://groups.google.com/d/msg/syzkaller/dpZ6ou1WOiI/7zfgSe1QEAAJ

Well, it's not syzkaller alone. Also KASAN + stressing kernel +
continuously monitoring oopses.
It's not too surprising to me that it wasn't caught before. If one
doesn't use KASAN (do you?), he/she would just track random memory
corruptions for months.


>> -----Original Message-----
>> From: Dmitry Vyukov [mailto:dvyukov@...gle.com]
>> Sent: Friday, March 3, 2017 5:06 AM
>> To: Matthew Wilcox <mawilcox@...rosoft.com>; Andrew Morton
>> <akpm@...ux-foundation.org>; ross.zwisler@...ux.intel.com; Johannes Weiner
>> <hannes@...xchg.org>; Jan Kara <jack@...e.cz>; LKML <linux-
>> kernel@...r.kernel.org>; Tejun Heo <tj@...nel.org>
>> Cc: syzkaller <syzkaller@...glegroups.com>
>> Subject: Re: idr: use-after-free write in ida_get_new_above
>>
>> On Fri, Mar 3, 2017 at 10:51 AM, Dmitry Vyukov <dvyukov@...gle.com>
>> wrote:
>> > Hello,
>> >
>> > I am hitting the following use-after-free reports while running
>> > syzkaller fuzzer on commit c82be9d2244aacea9851c86f4fb74694c99cd874:
>> >
>> >
>> ==================================================================
>> > BUG: KASAN: use-after-free in ida_get_new_above+0x564/0x9b0
>> > lib/idr.c:295 at addr ffff880160ef3600
>> > Write of size 128 by task syz-executor1/1757
>> > CPU: 1 PID: 1757 Comm: syz-executor1 Not tainted 4.10.0+ #98
>> > Hardware name: Google Google Compute Engine/Google Compute Engine,
>> > BIOS Google 01/01/2011
>> > Call Trace:
>> >  __dump_stack lib/dump_stack.c:15 [inline]
>> >  dump_stack+0x2ee/0x3ef lib/dump_stack.c:51
>> >  kasan_object_err+0x1c/0x70 mm/kasan/report.c:162
>> >  print_address_description mm/kasan/report.c:200 [inline]
>> >  kasan_report_error mm/kasan/report.c:289 [inline]
>> >  kasan_report.part.2+0x1e5/0x4b0 mm/kasan/report.c:311
>> >  kasan_report+0x21/0x30 mm/kasan/report.c:298
>> >  check_memory_region_inline mm/kasan/kasan.c:315 [inline]
>> >  check_memory_region+0x139/0x190 mm/kasan/kasan.c:322
>> >  memset+0x23/0x40 mm/kasan/kasan.c:340
>> >  ida_get_new_above+0x564/0x9b0 lib/idr.c:295
>> >  ida_get_new include/linux/idr.h:209 [inline]
>> >  proc_alloc_inum+0x97/0x1f0 fs/proc/generic.c:201
>> >  proc_register+0x25/0x2f0 fs/proc/generic.c:338
>> >  proc_create_data+0x100/0x190 fs/proc/generic.c:509
>> >  proc_create include/linux/proc_fs.h:35 [inline]
>> >  ip_rt_do_proc_init+0x53/0x110 net/ipv4/route.c:382
>> >  ops_init+0x10a/0x530 net/core/net_namespace.c:115
>> >  setup_net+0x2ed/0x690 net/core/net_namespace.c:291
>> >  copy_net_ns+0x26c/0x530 net/core/net_namespace.c:396
>> >  create_new_namespaces+0x409/0x860 kernel/nsproxy.c:106
>> >  unshare_nsproxy_namespaces+0xae/0x1e0 kernel/nsproxy.c:205
>> >  SYSC_unshare kernel/fork.c:2281 [inline]
>> >  SyS_unshare+0x64e/0xfc0 kernel/fork.c:2231
>> >  entry_SYSCALL_64_fastpath+0x1f/0xc2
>> > RIP: 0033:0x44fb79
>> > RSP: 002b:00007f1b66930b58 EFLAGS: 00000212 ORIG_RAX:
>> 0000000000000110
>> > RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 000000000044fb79
>> > RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000062020200
>> > RBP: 0000000062020200 R08: 0000000000000000 R09: 0000000000000000
>> > R10: 0000000000000000 R11: 0000000000000212 R12: 00000000007082a0
>> > R13: 0000000000000004 R14: 0000000000000400 R15: 0000000000000000
>> > Object at ffff880160ef3600, in cache kmalloc-128 size: 128
>> > Allocated:
>> > PID = 1730
>> >  save_stack_trace+0x16/0x20 arch/x86/kernel/stacktrace.c:57
>> >  save_stack+0x43/0xd0 mm/kasan/kasan.c:502
>> >  set_track mm/kasan/kasan.c:514 [inline]
>> >  kasan_kmalloc+0xaa/0xd0 mm/kasan/kasan.c:605
>> >  kmem_cache_alloc_trace+0x10b/0x670 mm/slab.c:3634
>> >  kmalloc include/linux/slab.h:490 [inline]
>> >  ida_pre_get+0xcd/0xe0 lib/radix-tree.c:2129
>> >  proc_alloc_inum+0xc5/0x1f0 fs/proc/generic.c:197
>> >  proc_register+0x25/0x2f0 fs/proc/generic.c:338
>> >  proc_create_data+0x100/0x190 fs/proc/generic.c:509
>> >  proc_create include/linux/proc_fs.h:35 [inline]
>> >  packet_net_init+0x8d/0xb0 net/packet/af_packet.c:4488
>> >  ops_init+0x10a/0x530 net/core/net_namespace.c:115
>> >  setup_net+0x2ed/0x690 net/core/net_namespace.c:291
>> >  copy_net_ns+0x26c/0x530 net/core/net_namespace.c:396
>> >  create_new_namespaces+0x409/0x860 kernel/nsproxy.c:106
>> >  unshare_nsproxy_namespaces+0xae/0x1e0 kernel/nsproxy.c:205
>> >  SYSC_unshare kernel/fork.c:2281 [inline]
>> >  SyS_unshare+0x64e/0xfc0 kernel/fork.c:2231
>> >  entry_SYSCALL_64_fastpath+0x1f/0xc2
>> > Freed:
>> > PID = 1648
>> >  save_stack_trace+0x16/0x20 arch/x86/kernel/stacktrace.c:57
>> >  save_stack+0x43/0xd0 mm/kasan/kasan.c:502
>> >  set_track mm/kasan/kasan.c:514 [inline]
>> >  kasan_slab_free+0x6f/0xb0 mm/kasan/kasan.c:578
>> >  __cache_free mm/slab.c:3510 [inline]
>> >  kfree+0xd3/0x250 mm/slab.c:3827
>> >  ida_pre_get+0x8a/0xe0 lib/radix-tree.c:2133
>> >  ida_simple_get+0x168/0x320 lib/idr.c:443
>> >  __kernfs_new_node+0xfe/0x3c0 fs/kernfs/dir.c:633
>> >  kernfs_new_node+0x80/0xe0 fs/kernfs/dir.c:661
>> >  kernfs_create_dir_ns+0x41/0x140 fs/kernfs/dir.c:933
>> >  kernfs_create_dir include/linux/kernfs.h:477 [inline]
>> >  internal_create_group+0xf7/0x8f0 fs/sysfs/group.c:124
>> >  sysfs_create_group fs/sysfs/group.c:156 [inline]
>> >  sysfs_create_groups+0x97/0x130 fs/sysfs/group.c:183
>> >  device_add_groups drivers/base/core.c:1031 [inline]
>> >  device_add_attrs drivers/base/core.c:1058 [inline]
>> >  device_add+0xb2f/0x1650 drivers/base/core.c:1679
>> >  netdev_register_kobject+0x183/0x360 net/core/net-sysfs.c:1603
>> >  register_netdevice+0xa27/0xed0 net/core/dev.c:7228
>> >  tun_set_iff drivers/net/tun.c:1843 [inline]
>> >  __tun_chr_ioctl+0x1bd0/0x3b70 drivers/net/tun.c:2046
>> >  tun_chr_ioctl+0x2a/0x40 drivers/net/tun.c:2291
>> >  vfs_ioctl fs/ioctl.c:43 [inline]
>> >  do_vfs_ioctl+0x1bf/0x1790 fs/ioctl.c:683
>> >  SYSC_ioctl fs/ioctl.c:698 [inline]
>> >  SyS_ioctl+0x8f/0xc0 fs/ioctl.c:689
>> >  entry_SYSCALL_64_fastpath+0x1f/0xc2
>> > Memory state around the buggy address:
>> >  ffff880160ef3500: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
>> >  ffff880160ef3580: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
>> >>ffff880160ef3600: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>> >                    ^
>> >  ffff880160ef3680: fc fc fc fc fc fc fc fc 00 00 00 00 00 00 00 00
>> >  ffff880160ef3700: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
>> >
>> ==================================================================
>> >
>> >
>> > I think we need:
>> >
>> > diff --git a/lib/radix-tree.c b/lib/radix-tree.c
>> > index 5ed506d648c4..d435916f6527 100644
>> > --- a/lib/radix-tree.c
>> > +++ b/lib/radix-tree.c
>> > @@ -2129,8 +2129,8 @@ int ida_pre_get(struct ida *ida, gfp_t gfp)
>> >                 struct ida_bitmap *bitmap = kmalloc(sizeof(*bitmap), gfp);
>> >                 if (!bitmap)
>> >                         return 0;
>> > -               bitmap = this_cpu_cmpxchg(ida_bitmap, NULL, bitmap);
>> > -               kfree(bitmap);
>> > +               if (!this_cpu_cmpxchg(ida_bitmap, NULL, bitmap))
>> > +                       kfree(bitmap);
>>
>> I meant:
>>
>> +               if (this_cpu_cmpxchg(ida_bitmap, NULL, bitmap))
>> +                       kfree(bitmap);
>>
>>
>> >         }
>> >
>> >
>> > Otherwise we both free the still installed old pointer and leak the new one.

Powered by blists - more mailing lists