linux-kernel - Re: BLKSECDISCARD ioctl and hung tasks

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: <20200219013714.GB31488@ming.t460p>
Date:   Wed, 19 Feb 2020 09:37:14 +0800
From:   Ming Lei <ming.lei@...hat.com>
To:     Jesse Barnes <jsbarnes@...gle.com>
Cc:     Salman Qazi <sqazi@...gle.com>, Ming Lei <tom.leiming@...il.com>,
        Bart Van Assche <bvanassche@....org>,
        Jens Axboe <axboe@...nel.dk>, Christoph Hellwig <hch@....de>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-block <linux-block@...r.kernel.org>,
        Gwendal Grignou <gwendal@...gle.com>
Subject: Re: BLKSECDISCARD ioctl and hung tasks

On Tue, Feb 18, 2020 at 08:11:53AM -0800, Jesse Barnes wrote:
> On Fri, Feb 14, 2020 at 7:47 PM Ming Lei <ming.lei@...hat.com> wrote:
> > What are the 'other operations'? Are they block IOs?
> >
> > If yes, that is why I suggest to fix submit_bio_wait(), which should cover
> > most of sync bio submission.
> >
> > Anyway, the fix is simple & generic enough, I'd plan to post a formal
> > patch if no one figures out better doable approaches.
> 
> Yeah I think any block I/O operation that occurs after the
> BLKSECDISCARD is submitted will also potentially be affected by the
> hung task timeouts, and I think your patch will address that.  My only
> concern with it is that it might hide some other I/O "hangs" that are
> due to device misbehavior instead.  Yes driver and device timeouts
> should generally catch those, but with this in place we might miss a
> few bugs.

We may add a warning(device name, request, ...) when one IO isn't completed in
sysctl_hung_task_timeout_secs, so this device misbehavior still can be
caught.

> 
> Given the nature of these types of storage devices though, I think
> that's a minor issue and not worth blocking the patch on, given that
> it should prevent a lot of false positive hang reports as Salman
> demonstrated.

OK, I will post it soon.


Thanks,
Ming