linux-kernel - Re: [PATCH RT] signal: Prevent double-free of user struct

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <f8152ef5-f71d-b8b6-fe9e-6a8678930775@monom.org>
Date:   Wed, 15 Apr 2020 21:05:14 +0200
From:   Daniel Wagner <wagi@...om.org>
To:     Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Matt Fleming <matt@...eblueprint.co.uk>
Cc:     linux-rt@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH RT] signal: Prevent double-free of user struct

Hi Sebastian,

On 15.04.20 18:46, Sebastian Andrzej Siewior wrote:
> On 2020-04-07 10:54:13 [+0100], Matt Fleming wrote:
>> The way user struct reference counting works changed significantly with,
>>
>>    fda31c50292a ("signal: avoid double atomic counter increments for user accounting")
>>
>> Now user structs are only freed once the last pending signal is
>> dequeued. Make sigqueue_free_current() follow this new convention to
>> avoid freeing the user struct multiple times and triggering this
>> warning:
>>
>>   refcount_t: underflow; use-after-free.
>>   WARNING: CPU: 0 PID: 6794 at lib/refcount.c:288 refcount_dec_not_one+0x45/0x50
>>   Call Trace:
>>    refcount_dec_and_lock_irqsave+0x16/0x60
>>    free_uid+0x31/0xa0
>>    ? schedule_hrtimeout_range_clock+0x104/0x110
>>    __dequeue_signal+0x17c/0x190
>>    dequeue_signal+0x5a/0x1b0
>>    do_sigtimedwait+0x208/0x250
>>    __x64_sys_rt_sigtimedwait+0x6f/0xd0
>>    do_syscall_64+0x72/0x200
>>    entry_SYSCALL_64_after_hwframe+0x49/0xbe
> 
> While all this sounds reasonable, may I ask what did you do to trigger
> this?

This can be triggered by running sigwaittest.

  # sigwaittest -t -a -p 98

a few seconds should be enough to get the splat.

> This is v5.6 only, correct?

I've seen this also with a frankstein version of 5.2-rt...v5-4-rt :)

Thanks,
Daniel