linux-kernel - Re: [PATCH 2/2] tracing: Fix wild-memory-access in register_synth

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

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <a456e1c64c9f1a6c5f60a3a864284199162db793.camel@kernel.org>
Date:   Sat, 12 Nov 2022 15:24:50 -0600
From:   Tom Zanussi <zanussi@...nel.org>
To:     Shang XiaoJing <shangxiaojing@...wei.com>, rostedt@...dmis.org,
        mhiramat@...nel.org, fengguang.wu@...el.com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] tracing: Fix wild-memory-access in
 register_synth_event()

Hi Shang,

On Tue, 2022-11-08 at 16:31 +0800, Shang XiaoJing wrote:
> In register_synth_event(), if set_synth_event_print_fmt() failed, then
> both trace_remove_event_call() and unregister_trace_event() will be
> called. If call->event.funcs is not NULL, then the trace_event_call will
> call __unregister_trace_event() twice. As the result, the second
> __unregister_trace_event() will causes the wild-memory-access.
> 
> register_synth_event
>     set_synth_event_print_fmt failed
>     trace_remove_event_call
>         event_remove
>             if call->event.funcs then
>             __unregister_trace_event (first call)
>     unregister_trace_event
>         __unregister_trace_event (second call)
> 
> Fix the bug by avoiding to call the second __unregister_trace_event() by
> checking if the first one is called.
> 
> general protection fault, probably for non-canonical address
>         0xfbd59c0000000024: 0000 [#1] SMP KASAN PTI
> KASAN: maybe wild-memory-access in range
> [0xdead000000000120-0xdead000000000127]
> CPU: 0 PID: 3807 Comm: modprobe Not tainted
> 6.1.0-rc1-00186-g76f33a7eedb4 #299
> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
> rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014
> RIP: 0010:unregister_trace_event+0x6e/0x280
> Code: 00 fc ff df 4c 89 ea 48 c1 ea 03 80 3c 02 00 0f 85 0e 02 00 00 48
> b8 00 00 00 00 00 fc ff df 4c 8b 63 08 4c 89 e2 48 c1 ea 03 <80> 3c 02
> 00 0f 85 e2 01 00 00 49 89 2c 24 48 85 ed 74 28 e8 7a 9b
> RSP: 0018:ffff88810413f370 EFLAGS: 00010a06
> RAX: dffffc0000000000 RBX: ffff888105d050b0 RCX: 0000000000000000
> RDX: 1bd5a00000000024 RSI: ffff888119e276e0 RDI: ffffffff835a8b20
> RBP: dead000000000100 R08: 0000000000000000 R09: fffffbfff0913481
> R10: ffffffff8489a407 R11: fffffbfff0913480 R12: dead000000000122
> R13: ffff888105d050b8 R14: 0000000000000000 R15: ffff888105d05028
> FS:  00007f7823e8d540(0000) GS:ffff888119e00000(0000)
> knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00007f7823e7ebec CR3: 000000010a058002 CR4: 0000000000330ef0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
> Call Trace:
>  <TASK>
>  __create_synth_event+0x1e37/0x1eb0
>  create_or_delete_synth_event+0x110/0x250
>  synth_event_run_command+0x2f/0x110
>  test_gen_synth_cmd+0x170/0x2eb [synth_event_gen_test]
>  synth_event_gen_test_init+0x76/0x9bc [synth_event_gen_test]
>  do_one_initcall+0xdb/0x480
>  do_init_module+0x1cf/0x680
>  load_module+0x6a50/0x70a0
>  __do_sys_finit_module+0x12f/0x1c0
>  do_syscall_64+0x3f/0x90
>  entry_SYSCALL_64_after_hwframe+0x63/0xcd
> 
> Fixes: 4b147936fa50 ("tracing: Add support for 'synthetic' events")
> Signed-off-by: Shang XiaoJing <shangxiaojing@...wei.com>
> Cc: stable@...r.kernel.org
> ---
>  kernel/trace/trace_events_synth.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/kernel/trace/trace_events_synth.c b/kernel/trace/trace_events_synth.c
> index e310052dc83c..a51280a153e3 100644
> --- a/kernel/trace/trace_events_synth.c
> +++ b/kernel/trace/trace_events_synth.c
> @@ -830,6 +830,8 @@ static int register_synth_event(struct synth_event *event)
>         ret = set_synth_event_print_fmt(call);
>         if (ret < 0) {
>                 trace_remove_event_call(call);
> +               if (call->event.funcs)
> +                       return ret;
>                 goto err;
>         }
>   out:

Good catch, thanks for finding this bug!

It looks like call->event.funcs will always be true here since it's set
to &synth_event_funcs above.

So it seems like you could just call trace_remove_event_call() and fall
through for this (ret < 0) case?  If so, it might be good to put a
comment there noting that trace_remove_event_call() will call
unregister_trace_event(), so it's ok to just return.

Thanks,

Tom