| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <7cdd692b-6325-4969-919f-7606c83391c9@huawei.com>
Date: Thu, 13 Mar 2025 14:43:36 +0800
From: Wang Liang <wangliang74@...wei.com>
To: Leon Romanovsky <leon@...nel.org>
CC: <jgg@...pe.ca>, <cmeiohas@...dia.com>, <michaelgur@...dia.com>,
<huangjunxian6@...ilicon.com>, <liyuyu6@...wei.com>, <markzhang@...dia.com>,
<linux@...blig.org>, <jbi.octave@...il.com>, <dsahern@...nel.org>,
<yuehaibing@...wei.com>, <zhangchangzhong@...wei.com>,
<linux-rdma@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net] infiniband: fix use-after-free when rename device
name
在 2025/3/12 21:49, Leon Romanovsky 写道:
> On Tue, Mar 11, 2025 at 10:55:42AM +0800, Wang Liang wrote:
>> 在 2025/3/10 18:14, Leon Romanovsky 写道:
>>> On Mon, Mar 10, 2025 at 02:45:16PM +0800, Wang Liang wrote:
>>>> Syzbot reported a slab-use-after-free with the following call trace:
>>>>
>>>> ==================================================================
>>>> BUG: KASAN: slab-use-after-free in nla_put+0xd3/0x150 lib/nlattr.c:1099
>>>> Read of size 5 at addr ffff888140ea1c60 by task syz.0.988/10025
>>>>
>>>> CPU: 0 UID: 0 PID: 10025 Comm: syz.0.988 Not tainted 6.14.0-rc4-syzkaller-00859-gf77f12010f67 #0
>>>> Hardware name: Google Compute Engine, BIOS Google 02/12/2025
>>>> Call Trace:
>>>> <TASK>
>>>> __dump_stack lib/dump_stack.c:94 [inline]
>>>> dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120
>>>> print_address_description mm/kasan/report.c:408 [inline]
>>>> print_report+0x16e/0x5b0 mm/kasan/report.c:521
>>>> kasan_report+0x143/0x180 mm/kasan/report.c:634
>>>> kasan_check_range+0x282/0x290 mm/kasan/generic.c:189
>>>> __asan_memcpy+0x29/0x70 mm/kasan/shadow.c:105
>>>> nla_put+0xd3/0x150 lib/nlattr.c:1099
>>>> nla_put_string include/net/netlink.h:1621 [inline]
>>>> fill_nldev_handle+0x16e/0x200 drivers/infiniband/core/nldev.c:265
>>>> rdma_nl_notify_event+0x561/0xef0 drivers/infiniband/core/nldev.c:2857
>>>> ib_device_notify_register+0x22/0x230 drivers/infiniband/core/device.c:1344
>>>> ib_register_device+0x1292/0x1460 drivers/infiniband/core/device.c:1460
>>>> rxe_register_device+0x233/0x350 drivers/infiniband/sw/rxe/rxe_verbs.c:1540
>>>> rxe_net_add+0x74/0xf0 drivers/infiniband/sw/rxe/rxe_net.c:550
>>>> rxe_newlink+0xde/0x1a0 drivers/infiniband/sw/rxe/rxe.c:212
>>>> nldev_newlink+0x5ea/0x680 drivers/infiniband/core/nldev.c:1795
>>>> rdma_nl_rcv_skb drivers/infiniband/core/netlink.c:239 [inline]
>>>> rdma_nl_rcv+0x6dd/0x9e0 drivers/infiniband/core/netlink.c:259
>>>> netlink_unicast_kernel net/netlink/af_netlink.c:1313 [inline]
>>>> netlink_unicast+0x7f6/0x990 net/netlink/af_netlink.c:1339
>>>> netlink_sendmsg+0x8de/0xcb0 net/netlink/af_netlink.c:1883
>>>> sock_sendmsg_nosec net/socket.c:709 [inline]
>>>> __sock_sendmsg+0x221/0x270 net/socket.c:724
>>>> ____sys_sendmsg+0x53a/0x860 net/socket.c:2564
>>>> ___sys_sendmsg net/socket.c:2618 [inline]
>>>> __sys_sendmsg+0x269/0x350 net/socket.c:2650
>>>> do_syscall_x64 arch/x86/entry/common.c:52 [inline]
>>>> do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83
>>>> entry_SYSCALL_64_after_hwframe+0x77/0x7f
>>>> RIP: 0033:0x7f42d1b8d169
>>>> Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 ...
>>>> RSP: 002b:00007f42d2960038 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
>>>> RAX: ffffffffffffffda RBX: 00007f42d1da6320 RCX: 00007f42d1b8d169
>>>> RDX: 0000000000000000 RSI: 00004000000002c0 RDI: 000000000000000c
>>>> RBP: 00007f42d1c0e2a0 R08: 0000000000000000 R09: 0000000000000000
>>>> R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
>>>> R13: 0000000000000000 R14: 00007f42d1da6320 R15: 00007ffe399344a8
>>>> </TASK>
>>>>
>>>> Allocated by task 10025:
>>>> kasan_save_stack mm/kasan/common.c:47 [inline]
>>>> kasan_save_track+0x3f/0x80 mm/kasan/common.c:68
>>>> poison_kmalloc_redzone mm/kasan/common.c:377 [inline]
>>>> __kasan_kmalloc+0x98/0xb0 mm/kasan/common.c:394
>>>> kasan_kmalloc include/linux/kasan.h:260 [inline]
>>>> __do_kmalloc_node mm/slub.c:4294 [inline]
>>>> __kmalloc_node_track_caller_noprof+0x28b/0x4c0 mm/slub.c:4313
>>>> __kmemdup_nul mm/util.c:61 [inline]
>>>> kstrdup+0x42/0x100 mm/util.c:81
>>>> kobject_set_name_vargs+0x61/0x120 lib/kobject.c:274
>>>> dev_set_name+0xd5/0x120 drivers/base/core.c:3468
>>>> assign_name drivers/infiniband/core/device.c:1202 [inline]
>>>> ib_register_device+0x178/0x1460 drivers/infiniband/core/device.c:1384
>>>> rxe_register_device+0x233/0x350 drivers/infiniband/sw/rxe/rxe_verbs.c:1540
>>>> rxe_net_add+0x74/0xf0 drivers/infiniband/sw/rxe/rxe_net.c:550
>>>> rxe_newlink+0xde/0x1a0 drivers/infiniband/sw/rxe/rxe.c:212
>>>> nldev_newlink+0x5ea/0x680 drivers/infiniband/core/nldev.c:1795
>>>> rdma_nl_rcv_skb drivers/infiniband/core/netlink.c:239 [inline]
>>>> rdma_nl_rcv+0x6dd/0x9e0 drivers/infiniband/core/netlink.c:259
>>>> netlink_unicast_kernel net/netlink/af_netlink.c:1313 [inline]
>>>> netlink_unicast+0x7f6/0x990 net/netlink/af_netlink.c:1339
>>>> netlink_sendmsg+0x8de/0xcb0 net/netlink/af_netlink.c:1883
>>>> sock_sendmsg_nosec net/socket.c:709 [inline]
>>>> __sock_sendmsg+0x221/0x270 net/socket.c:724
>>>> ____sys_sendmsg+0x53a/0x860 net/socket.c:2564
>>>> ___sys_sendmsg net/socket.c:2618 [inline]
>>>> __sys_sendmsg+0x269/0x350 net/socket.c:2650
>>>> do_syscall_x64 arch/x86/entry/common.c:52 [inline]
>>>> do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83
>>>> entry_SYSCALL_64_after_hwframe+0x77/0x7f
>>>>
>>>> Freed by task 10035:
>>>> kasan_save_stack mm/kasan/common.c:47 [inline]
>>>> kasan_save_track+0x3f/0x80 mm/kasan/common.c:68
>>>> kasan_save_free_info+0x40/0x50 mm/kasan/generic.c:576
>>>> poison_slab_object mm/kasan/common.c:247 [inline]
>>>> __kasan_slab_free+0x59/0x70 mm/kasan/common.c:264
>>>> kasan_slab_free include/linux/kasan.h:233 [inline]
>>>> slab_free_hook mm/slub.c:2353 [inline]
>>>> slab_free mm/slub.c:4609 [inline]
>>>> kfree+0x196/0x430 mm/slub.c:4757
>>>> kobject_rename+0x38f/0x410 lib/kobject.c:524
>>>> device_rename+0x16a/0x200 drivers/base/core.c:4525
>>>> ib_device_rename+0x270/0x710 drivers/infiniband/core/device.c:402
>>>> nldev_set_doit+0x30e/0x4c0 drivers/infiniband/core/nldev.c:1146
>>>> rdma_nl_rcv_skb drivers/infiniband/core/netlink.c:239 [inline]
>>>> rdma_nl_rcv+0x6dd/0x9e0 drivers/infiniband/core/netlink.c:259
>>>> netlink_unicast_kernel net/netlink/af_netlink.c:1313 [inline]
>>>> netlink_unicast+0x7f6/0x990 net/netlink/af_netlink.c:1339
>>>> netlink_sendmsg+0x8de/0xcb0 net/netlink/af_netlink.c:1883
>>>> sock_sendmsg_nosec net/socket.c:709 [inline]
>>>> __sock_sendmsg+0x221/0x270 net/socket.c:724
>>>> ____sys_sendmsg+0x53a/0x860 net/socket.c:2564
>>>> ___sys_sendmsg net/socket.c:2618 [inline]
>>>> __sys_sendmsg+0x269/0x350 net/socket.c:2650
>>>> do_syscall_x64 arch/x86/entry/common.c:52 [inline]
>>>> do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83
>>>> entry_SYSCALL_64_after_hwframe+0x77/0x7f
>>>>
>>>> This is because if rename device happens, the old name is freed in
>>>> ib_device_rename() with lock, but fill_nldev_handle() may visit the dev
>>>> name locklessly triggered by rxe_newlink().
>>>>
>>>> Fix this by add lock around rdma_nl_notify_event() in
>>>> ib_device_notify_register().
>>>>
>>>> Reported-by: syzbot+f60349ba1f9f08df349f@...kaller.appspotmail.com
>>>> Closes: https://syzkaller.appspot.com/bug?extid=25bc6f0ed2b88b9eb9b8
>>>> Fixes: 9cbed5aab5ae ("RDMA/nldev: Add support for RDMA monitoring")
>>>> Signed-off-by: Wang Liang <wangliang74@...wei.com>
>>>> ---
>>>> drivers/infiniband/core/device.c | 2 ++
>>>> 1 file changed, 2 insertions(+)
>>>>
>>>> diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
>>>> index 0ded91f056f3..4536621ada0d 100644
>>>> --- a/drivers/infiniband/core/device.c
>>>> +++ b/drivers/infiniband/core/device.c
>>>> @@ -1341,7 +1341,9 @@ static void ib_device_notify_register(struct ib_device *device)
>>>> u32 port;
>>>> int ret;
>>>> + down_write(&devices_rwsem);
>>> The analysis looks correct to me, however this should be down_read(&devices_rwsem)
>>> together with comment about possible race with RDMA netlink, which can change
>>> internals of struct ib_device.
>>>
>>> I wonder if this read semaphore should be hold for whole
>>> ib_device_notify_register() function and not only for RDMA_REGISTER_EVENT event.
>>>
>>> Thanks
>> Yes, you are right! The RDMA_NETDEV_ATTACH_EVENT event in function
>> ib_device_notify_register() can also visit the dev name locklessly:
>>
>> ......
>>
>> So only hold devices_rwsem for whole ib_device_notify_register() may be not
>> enough.
>>
>> How about add down_read(&devices_rwsem) around
>> fill_dev_info()/fill_port_info()/
>> fill_res_info()/ib_device_notify_register().
> I don't think so, as all flows in nldev.c except rdma_nl_notify_event() are
> single threaded and call to ib_device_rename() will make sure that no
> other RDMA_NLDEV_CMD_* commands are executed.
>
> In your case call to rdma_nl_notify_event() can happen in parallel to
> device_rename only.
>
> Thanks
Thanks for all your comments.
I will send a new patch later, and hold the devices_rwsem for whole
ib_device_notify_register().
Please check it. Thanks.
Powered by blists - more mailing lists