Re: [PATCH V3 1/5] RDMA/core: Transport-independent access flags

[Sagi Grimberg <sagig@dev.mellanox.co.il>]

On 7/13/2015 7:50 PM, Jason Gunthorpe wrote:
> On Sun, Jul 12, 2015 at 10:49:08AM +0300, Sagi Grimberg wrote:
>> On 7/10/2015 10:34 PM, Christoph Hellwig wrote:
>>> On Thu, Jul 09, 2015 at 09:52:59AM -0400, Chuck Lever wrote:
>>>> There is one remaining kernel user of ib_reg_phys_mr() in 4.2: Lustre.
>>>
>>> It's in the staging tree, which proper in-tree code doesn't have to
>>> cater for.  So as soon as sunrpc is done using the interface we can and
>>> should kill it off.
>>>
>>
>> I think we should probably ask the Lustre folks if they have a real use
>> case for it before we remove it completely.
>
> I had an interesting conversation with some Lustre devs where they
> were very concerned that FMR was going to be removed and they didn't really
> seem to even know that FRMR existed.
>
> I'm sure their PHYS_MR usage is crufty old code to support old
> adaptors.
>
> And again, here is a great example of how the API is not helping the
> ULPs do what they want to do. Lustre doesn't care one bit about this
> stuff, they just want to send messages..

All protocols cares about transferring data and sending messages, so
it's not a good enough reason for a poor registration method choice.
This just emphasizes why we need to converge to a single method.

Sagi.