Re: [PATCH 10/13] dm: add core functions for request-based dm

[Kiyoshi Ueda <k-ueda@ct.jp.nec.com>]

Hi Nikanth,

Nikanth Karthikesan wrote:
> Hi Kiyoshi
> 
>>>> On 10/28/2008 at 09:30 PM, Kiyoshi Ueda <k-ueda@ct.jp.nec.com> wrote: 
>> Hi Nikanth,
>>
>> On Fri, 24 Oct 2008 13:14:50 +0530, "Nikanth K" wrote:
>>> On Fri, Sep 12, 2008 at 8:16 PM, Kiyoshi Ueda wrote:
>>> <snip>
>>>
>>>> +static int dm_make_request(struct request_queue *q, struct bio *bio)
>>>> +{
>>>> +       struct mapped_device *md = (struct mapped_device *)q->queuedata;
>>>> +
>>>> +       if (unlikely(bio_barrier(bio))) {
>>>> +               bio_endio(bio, -EOPNOTSUPP);
>>>> +               return 0;
>>>> +       }
>>>> +
>>> <snip>
>>>
>>> Why not add barrier support in the beginning itself, so that targets
>>> can be developed with barriers in mind? At least can we make the target
>>> to return error, instead of the core?
>> Currently, there is no barrier support in dm, not only request-based.
>> Barrier support is a different feature in the next step, I think.
> 
> But there are some works in that direction to add support for barriers in dm.
> That is why I think building request-based dm with barriers from the 
> ground up might be a good idea.

I agree, if I or other people have a time to implement barrier support
for request-based dm.
But I think the some works you mentioned above are:
    - Andi Kleen: barrier support for linear (single device)
    - Milan Broz: full barrier support in dm core (no target patch)
so there is no barrier support work for dm-multipath yet.

Current request-based target is only dm-multipath, so we won't have
any feature regression even if request-based dm-multipath gets in.
And I don't have much time to implement barrier support for
request-based dm-multipath now, so I'd like to consider it as
the next step.

Thanks,
Kiyoshi Ueda