Message-ID: <E9BA11C2-0F15-4FFF-8E29-74640E82D046@vmware.com>
Date:   Mon, 4 Jun 2018 16:02:39 +0000
From:   "Jorgen S. Hansen" <jhansen@...are.com>
To:     Stefan Hajnoczi <stefanha@...hat.com>
CC:     Hangbin Liu <liuhangbin@...il.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "David S. Miller" <davem@...emloft.net>
Subject: Re: [PATCH net] VSOCK: check sk state before receive


> On May 30, 2018, at 11:17 AM, Stefan Hajnoczi <stefanha@...hat.com> wrote:
> 
> On Sun, May 27, 2018 at 11:29:45PM +0800, Hangbin Liu wrote:
>> Hmm...Although I won't reproduce this bug with my reproducer after
>> apply my patch. I could still get a similiar issue with syzkaller sock vnet test.
>> 
>> It looks this patch is not complete. Here is the KASAN call trace with my patch.
>> I can also reproduce it without my patch.
> 
> Seems like a race between vmci_datagram_destroy_handle() and the
> delayed callback, vmci_transport_recv_dgram_cb().
> 
> I don't know the VMCI transport well so I'll leave this to Jorgen.

Yes, it looks like we are calling the delayed callback after we return from vmci_datagram_destroy_handle(). I’ll take a closer look at the VMCI side here - the refcounting of VMCI datagram endpoints should guard against this, since the delayed callback does a get on the datagram resource, so this could a VMCI driver issue, and not a problem in the VMCI transport for AF_VSOCK.

> 
>> ==================================================================
>> BUG: KASAN: use-after-free in vmci_transport_allow_dgram.part.7+0x155/0x1a0 [vmw_vsock_vmci_transport]
>> Read of size 4 at addr ffff880026a3a914 by task kworker/0:2/96
>> 
>> CPU: 0 PID: 96 Comm: kworker/0:2 Not tainted 4.17.0-rc6.vsock+ #28
>> Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2011
>> Workqueue: events dg_delayed_dispatch [vmw_vmci]
>> Call Trace:
>> __dump_stack lib/dump_stack.c:77 [inline]
>> dump_stack+0xdd/0x18e lib/dump_stack.c:113
>> print_address_description+0x7a/0x3e0 mm/kasan/report.c:256
>> kasan_report_error mm/kasan/report.c:354 [inline]
>> kasan_report+0x1dd/0x460 mm/kasan/report.c:412
>> vmci_transport_allow_dgram.part.7+0x155/0x1a0 [vmw_vsock_vmci_transport]
>> vmci_transport_recv_dgram_cb+0x5d/0x200 [vmw_vsock_vmci_transport]
>> dg_delayed_dispatch+0x99/0x1b0 [vmw_vmci]
>> process_one_work+0xa4e/0x1720 kernel/workqueue.c:2145
>> worker_thread+0x1df/0x1400 kernel/workqueue.c:2279
>> kthread+0x343/0x4b0 kernel/kthread.c:240
>> ret_from_fork+0x35/0x40 arch/x86/entry/entry_64.S:412
>> 
>> Allocated by task 2684:
>> set_track mm/kasan/kasan.c:460 [inline]
>> kasan_kmalloc+0xa0/0xd0 mm/kasan/kasan.c:553
>> slab_post_alloc_hook mm/slab.h:444 [inline]
>> slab_alloc_node mm/slub.c:2741 [inline]
>> slab_alloc mm/slub.c:2749 [inline]
>> kmem_cache_alloc+0x105/0x330 mm/slub.c:2754
>> sk_prot_alloc+0x6a/0x2c0 net/core/sock.c:1468
>> sk_alloc+0xc9/0xbb0 net/core/sock.c:1528
>> __vsock_create+0xc8/0x9b0 [vsock]
>> vsock_create+0xfd/0x1a0 [vsock]
>> __sock_create+0x310/0x690 net/socket.c:1285
>> sock_create net/socket.c:1325 [inline]
>> __sys_socket+0x101/0x240 net/socket.c:1355
>> __do_sys_socket net/socket.c:1364 [inline]
>> __se_sys_socket net/socket.c:1362 [inline]
>> __x64_sys_socket+0x7d/0xd0 net/socket.c:1362
>> do_syscall_64+0x175/0x630 arch/x86/entry/common.c:287
>> entry_SYSCALL_64_after_hwframe+0x44/0xa9
>> 
>> Freed by task 2684:
>> set_track mm/kasan/kasan.c:460 [inline]
>> __kasan_slab_free+0x130/0x180 mm/kasan/kasan.c:521
>> slab_free_hook mm/slub.c:1388 [inline]
>> slab_free_freelist_hook mm/slub.c:1415 [inline]
>> slab_free mm/slub.c:2988 [inline]
>> kmem_cache_free+0xce/0x410 mm/slub.c:3004
>> sk_prot_free net/core/sock.c:1509 [inline]
>> __sk_destruct+0x629/0x940 net/core/sock.c:1593
>> sk_destruct+0x4e/0x90 net/core/sock.c:1601
>> __sk_free+0xd3/0x320 net/core/sock.c:1612
>> sk_free+0x2a/0x30 net/core/sock.c:1623
>> __vsock_release+0x431/0x610 [vsock]
>> vsock_release+0x3c/0xc0 [vsock]
>> sock_release+0x91/0x200 net/socket.c:594
>> sock_close+0x17/0x20 net/socket.c:1149
>> __fput+0x368/0xa20 fs/file_table.c:209
>> task_work_run+0x1c5/0x2a0 kernel/task_work.c:113
>> exit_task_work include/linux/task_work.h:22 [inline]
>> do_exit+0x1876/0x26c0 kernel/exit.c:865
>> do_group_exit+0x159/0x3e0 kernel/exit.c:968
>> get_signal+0x65a/0x1780 kernel/signal.c:2482
>> do_signal+0xa4/0x1fe0 arch/x86/kernel/signal.c:810
>> exit_to_usermode_loop+0x1b8/0x260 arch/x86/entry/common.c:162
>> prepare_exit_to_usermode arch/x86/entry/common.c:196 [inline]
>> syscall_return_slowpath arch/x86/entry/common.c:265 [inline]
>> do_syscall_64+0x505/0x630 arch/x86/entry/common.c:290
>> entry_SYSCALL_64_after_hwframe+0x44/0xa9
>> 
>> The buggy address belongs to the object at ffff880026a3a600
>> which belongs to the cache AF_VSOCK of size 1056
>> The buggy address is located 788 bytes inside of
>> 1056-byte region [ffff880026a3a600, ffff880026a3aa20)
>> The buggy address belongs to the page:
>> page:ffffea00009a8e00 count:1 mapcount:0 mapping:0000000000000000 index:0x0 compound_mapcount: 0
>> flags: 0xfffffc0008100(slab|head)
>> raw: 000fffffc0008100 0000000000000000 0000000000000000 00000001000d000d
>> raw: dead000000000100 dead000000000200 ffff880034471a40 0000000000000000
>> page dumped because: kasan: bad access detected
>> 
>> Memory state around the buggy address:
>> ffff880026a3a800: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>> ffff880026a3a880: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>>> ffff880026a3a900: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>>                         ^
>> ffff880026a3a980: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>> ffff880026a3aa00: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
>> ==================================================================

Powered by blists - more mailing lists