Re: 2.6.39 Block layer regression was [Bug] Boot hangs with 2.6.39-rc[123]]

[Jens Axboe <jaxboe@fusionio.com>]

On 2011-04-15 06:22, Christoph Hellwig wrote:
> On Thu, Apr 14, 2011 at 08:25:33PM -0700, Linus Torvalds wrote:
>> What's the thinking there? It looks very confused to me.
> 
> It is.  I sent a patch a couple of days ago to fix it.

Yeah thanks for that, I agree it looks a bit confusing as-is. I'll queue
it up.

>> Now, clearly RAID seems to be involved in the problem? The main thing
>> with that would be that the execution of the requests would tend to
>> generate new requests, that go back on the plug queue. Yes? And the
>> loop in flush_plug_list() means that they all should get flushed out,
>> I assume. But something clearly isn't working, and it does seem to be
>> about the RAID kind of setup. So either they didn't get put on the
>> plug queue, or the task got a new plug (which _wasn't_ flushed).
>>
>> Because we're clearly waiting for some request that hasn't completed.
>> Where in the plug queues would it be hiding?
> 
> There's a thread where Neil explains what the problem with MD is - it
> needs a callback on unplug time to generate e.g. the write intent bitmap
> or as large as possible writes for RAID5.  Jens and Neil have been
> looking into it.

I think we are done, Neil just needs to rebase around the current
for-linus and then we should expedite things in.

-- 
Jens Axboe