[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
Re: [dm-devel] [PATCH and question]: dm-table should check for valid queue]
- From: Stefan Bader <Stefan Bader de ibm com>
- To: agk redhat com
- Cc: dm-devel redhat com
- Subject: Re: [dm-devel] [PATCH and question]: dm-table should check for valid queue]
- Date: Tue, 15 Aug 2006 12:02:04 +0200
> On Thu, Aug 10, 2006 at 01:29:08PM +0200, Stefan Bader wrote:
>> But what effect should a missing queue have? Same
>> as device is congested or not or let the call fail somehow?
>
> I'm curious - under what circumstances is this missing:
> Is it intentional, or is it perhaps a race that needs fixing?
>
More or less intentional. If a device is (virtually) ripped from the
system the driver will throw everything away. Device-mapper still has a
reference to the bdev but the queue is cleaned and removed.
I am not sure whether the any congested call would ever be triggered in
that situation. But at least the unplug all is when user-space tries to
change the maps to remove or replace the missing device. And very likely
the flush all call would be used as well (probably after that).
Stefan
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]