linux-kernel - Re: WARNING in block layer triggered in 3.17-rc3

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]

Date:	Mon, 8 Sep 2014 10:05:54 -0500
From:	Shirish Pargaonkar <shirishpargaonkar@...il.com>
To:	Alan Stern <stern@...land.harvard.edu>
Cc:	James Bottomley <James.Bottomley@...senpartnership.com>,
	Jens Axboe <axboe@...nel.dk>,
	SCSI development list <linux-scsi@...r.kernel.org>,
	Kernel development list <linux-kernel@...r.kernel.org>
Subject: Re: WARNING in block layer triggered in 3.17-rc3

So should a request queue be in bypass mode when the device is being detached
and queue is being unregistereed because requests can get queued up?


On Mon, Sep 8, 2014 at 9:51 AM, Alan Stern <stern@...land.harvard.edu> wrote:
> On Sun, 7 Sep 2014, Shirish Pargaonkar wrote:
>
>> I think the problem is, when a gendisk is detached, its request queue
>> is not put in bypass mode
>> cause when it is re-attached, code tries to put it out of bypass mode,
>> hence the warning.
>>
>> So either of these should work, I have not tested it, just coded it up.
>
> I'm pretty sure that both of your solutions are wrong.
>
> Jens and James, it appears the problem is in blk_register_queue().  The
> code does this:
>
>         /*
>          * Initialization must be complete by now.  Finish the initial
>          * bypass from queue allocation.
>          */
>         queue_flag_set_unlocked(QUEUE_FLAG_INIT_DONE, q);
>         blk_queue_bypass_end(q);
>
> This doesn't work well if the queue is unregistered later and then
> registered again -- which is what happens when the sd driver is unbound
> from a device and then bound again.  It looks like the code should be:
>
>         if (!queue_flag_test_and_set(QUEUE_FLAG_INIT_DONE, q))
>                 blk_queue_bypass_end(q);
>
> Do you agree?  If so, I'll send in patch.
>
> Alan Stern
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/