Re: ppp/pppoe, still panic 4.15.3 in ppp_push

On 2018-02-22 20:30, Guillaume Nault wrote:

On Wed, Feb 21, 2018 at 12:04:30PM -0800, Cong Wang wrote:

quoted

On Thu, Feb 15, 2018 at 11:31 AM, Guillaume Nault [off-list ref] wrote:

quoted

On Thu, Feb 15, 2018 at 06:01:16PM +0200, Denys Fedoryshchenko wrote:

quoted

On 2018-02-15 17:55, Guillaume Nault wrote:

quoted

On Thu, Feb 15, 2018 at 12:19:52PM +0200, Denys Fedoryshchenko wrote:

quoted

Here we go:

 <srv> [24558.921549]
==================================================================
 <srv> [24558.922167] BUG: KASAN: use-after-free in
ppp_ioctl+0xa6a/0x1522
[ppp_generic]
 <srv> [24558.922776] Write of size 8 at addr ffff8803d35bf3f8 by task
accel-pppd/12622
 <srv> [24558.923113]
 <srv> [24558.923451] CPU: 0 PID: 12622 Comm: accel-pppd Tainted: G
W
4.15.3-build-0134 #1
 <srv> [24558.924058] Hardware name: HP ProLiant DL320e Gen8 v2,
BIOS P80
04/02/2015
 <srv> [24558.924406] Call Trace:
 <srv> [24558.924753]  dump_stack+0x46/0x59
 <srv> [24558.925103]  print_address_description+0x6b/0x23b
 <srv> [24558.925451]  ? ppp_ioctl+0xa6a/0x1522 [ppp_generic]
 <srv> [24558.925797]  kasan_report+0x21b/0x241
 <srv> [24558.926136]  ppp_ioctl+0xa6a/0x1522 [ppp_generic]
 <srv> [24558.926479]  ? ppp_nl_newlink+0x1da/0x1da [ppp_generic]
 <srv> [24558.926829]  ? sock_sendmsg+0x89/0x99
 <srv> [24558.927176]  ? __vfs_write+0xd9/0x4ad
 <srv> [24558.927523]  ? kernel_read+0xed/0xed
 <srv> [24558.927872]  ? SyS_getpeername+0x18c/0x18c
 <srv> [24558.928213]  ? bit_waitqueue+0x2a/0x2a
 <srv> [24558.928561]  ? wake_atomic_t_function+0x115/0x115
 <srv> [24558.928898]  vfs_ioctl+0x6e/0x81
 <srv> [24558.929228]  do_vfs_ioctl+0xa00/0xb10
 <srv> [24558.929571]  ? sigprocmask+0x1a6/0x1d0
 <srv> [24558.929907]  ? sigsuspend+0x13e/0x13e
 <srv> [24558.930239]  ? ioctl_preallocate+0x14e/0x14e
 <srv> [24558.930568]  ? SyS_rt_sigprocmask+0xf1/0x142
 <srv> [24558.930904]  ? sigprocmask+0x1d0/0x1d0
 <srv> [24558.931252]  SyS_ioctl+0x39/0x55
 <srv> [24558.931595]  ? do_vfs_ioctl+0xb10/0xb10
 <srv> [24558.931942]  do_syscall_64+0x1b1/0x31f
 <srv> [24558.932288]  entry_SYSCALL_64_after_hwframe+0x21/0x86
 <srv> [24558.932627] RIP: 0033:0x7f302849d8a7
 <srv> [24558.932965] RSP: 002b:00007f3029a52af8 EFLAGS: 00000206
ORIG_RAX:
0000000000000010
 <srv> [24558.933578] RAX: ffffffffffffffda RBX: 00007f3027d861e3 RCX:
00007f302849d8a7
 <srv> [24558.933927] RDX: 00007f3023f49468 RSI: 000000004004743a RDI:
0000000000003a67
 <srv> [24558.934266] RBP: 00007f3029a52b20 R08: 0000000000000000 R09:
000055c8308d8e40
 <srv> [24558.934607] R10: 0000000000000008 R11: 0000000000000206 R12:
00007f3023f49358
 <srv> [24558.934947] R13: 00007ffe86e5723f R14: 0000000000000000 R15:
00007f3029a53700
 <srv> [24558.935288]
 <srv> [24558.935626] Allocated by task 12622:
 <srv> [24558.935972]  ppp_register_net_channel+0x5f/0x5c6
[ppp_generic]
 <srv> [24558.936306]  pppoe_connect+0xab7/0xc71 [pppoe]
 <srv> [24558.936640]  SyS_connect+0x14b/0x1b7
 <srv> [24558.936975]  do_syscall_64+0x1b1/0x31f
 <srv> [24558.937319]  entry_SYSCALL_64_after_hwframe+0x21/0x86
 <srv> [24558.937655]
 <srv> [24558.937993] Freed by task 12622:
 <srv> [24558.938321]  kfree+0xb0/0x11d
 <srv> [24558.938658]  ppp_release+0x111/0x120 [ppp_generic]
 <srv> [24558.938994]  __fput+0x2ba/0x51a
 <srv> [24558.939332]  task_work_run+0x11c/0x13d
 <srv> [24558.939676]  exit_to_usermode_loop+0x7c/0xaf
 <srv> [24558.940022]  do_syscall_64+0x2ea/0x31f
 <srv> [24558.940368]  entry_SYSCALL_64_after_hwframe+0x21/0x86
 <srv> [24558.947099]

Your first guess was right. It looks like we have an issue with
reference counting on the channels. Can you send me your ppp_generic.o?

http://nuclearcat.com/ppp_generic.o
Compiled with gcc version 6.4.0 (Gentoo 6.4.0-r1 p1.3)

From what I can see, ppp_release() and ioctl(PPPIOCCONNECT) are called
concurrently on the same ppp_file. Even if this ppp_file was pointed at
by two different file descriptors, I can't see how this could defeat
the reference counting mechanism. I'm going to think more about it.

For me it looks like pch->clist is not removed from the list ppp->channels
when destroyed via ppp_release(). But I don't want to pretend I understand
ppp logic.

I've thought about that too, but couldn't find a scenario that could
trigger the bug.

To get ->private_data pointing to a struct channel pointer, a file needs
to ioctl(PPPIOCATTCHAN) first. For this call to succeed, the channel
must have been registered with ppp_register_net_channel(). Both
operations take a reference on the channel, which means that, before
adding pch->clist to a ppp->channels list (with ppp_connect_channel()),
the channel is already held by a /dev/ppp file and by the code that
registered the channel in the first place.

Therefore, closing the /dev/ppp file can't be enough to make
ppp_release() free the channel. We need to unregister the channel for
the refcount to drop to 0. And ppp_unregister_channel(), removes
pch->clist from ppp->channels before decrementing the reference
counter.

I'm simplifying a bit, and there could always be a race somewhere. But
I couldn't find anything so far.

I'm using accel-ppp that has unit-cache option, i guess for "reusing" ppp interfaces (because creating a lot of interfaces on BRAS with 8k users quite expensive).
Maybe it is somehow related and can be that scenario causing this bug?