linux-kernel - Re: INFO: rcu detected stall in corrupted (3)

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: <daec0343-b905-f483-602b-9c0b9e9b1379@i-love.sakura.ne.jp>
Date:   Sat, 30 Mar 2019 19:40:11 +0900
From:   Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     syzbot <syzbot+65cecdd27b726c261799@...kaller.appspotmail.com>,
        syzkaller-bugs@...glegroups.com, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Andy Lutomirski <luto@...nel.org>, mingo@...hat.com,
        x86@...nel.org, Peter Zijlstra <peterz@...radead.org>,
        Juri Lelli <juri.lelli@...hat.com>
Subject: Re: INFO: rcu detected stall in corrupted (3)

On 2019/03/30 16:46, Thomas Gleixner wrote:
> On Sat, 30 Mar 2019, Tetsuo Handa wrote:
>> This reproducer does sched_setattr(SCHED_DEADLINE) with bogus value, as with
>> a reproducer for "INFO: rcu detected stall in sys_sendfile64" did.
>>
>> sched_setattr(0, {size=0, sched_policy=0x6 /* SCHED_DEADLINE */,
>> sched_flags=0, sched_nice=0, sched_priority=0, sched_runtime=65535,
>> sched_deadline=4611686018427453437, sched_period=0}, 0) = 0
>>
>> #syz invalid
> 
> Marking this invalid is not really the right thing to do. Bogus deadline
> parameters should not cause RCU stalls. They either need to be rejected or
> handled gracefully.

But how can the scheduler be aware of various watchdogs' thresholds?

The scheduler behaves differently based on watchdog's remaining grace periods?
That sounds quite strange. If administrator tunes watchdog thresholds in a way
schedulers can't survive (or vice versa), it must be an administrator's fault.

Since this stall might occur with any combination, not closing this kind of
report will result in flood of stall reports...