Re: [PATCH v2 00/12] IB: Replace safe uses for ib_get_dma_mr with pd->local_dma_lkey

[Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 2387 bytes --]

On 08/14/2015 12:20 PM, Bart Van Assche wrote:
> On 08/14/2015 06:36 AM, Doug Ledford wrote:
>> On 08/11/2015 04:50 PM, Bart Van Assche wrote:
>>> On 08/01/2015 01:05 PM, Doug Ledford wrote:
>>>> On 07/31/2015 07:14 PM, Jason Gunthorpe wrote:
>>>>> On Fri, Jul 31, 2015 at 04:04:35PM -0700, Bart Van Assche wrote:
>>>>>
>>>>>> With only patches 1/12 and 8/12 applied my test passes.
>>>>>
>>>>> Thanks Bart!
>>>>>
>>>>> If you like, I can try and work on #9 (no promises :(), but I think
>>>>> I'll need some way to test it here.
>>>>>
>>>>> Do you by chance have a straightforward recipe to setup SRP and SRPT
>>>>> on two Linux's for this simple purpose?
>>>>>
>>>>> Doug, just drop #9 'IB/srp: Do not create an all physical insecure
>>>>> rkey by default' please, thanks.
>>>>
>>>> I've pulled in the series for 4.3.  I've included patch #9, but that's
>>>> on the understanding that if it can't be fixed before the initial pull
>>>> request, I'll drop it at that time.  I have the resources to test this,
>>>> so I can work on it, and that factors into my path forward here.
>>>
>>> Hi Doug,
>>>
>>> Do you feel comfortable enough to queue the nine patches that replace
>>> Jason's patch #9 for kernel v4.3 (see also
>>> http://thread.gmane.org/gmane.linux.drivers.rdma/28153/focus=28419) or
>>> would you rather prefer to take out patch #9 and queue these nine
>>> patches for kernel v4.4 ? Please advise.
>>
>> I was planning on queueing up the 9 replacement patches.
> 
> Hello Doug,
> 
> I hope that someone who has access to a DDN system has the time to test
> this patch series against a DDN array. The memory key that is associated
> with the indirect data buffer descriptor is never used by the upstream
> SRP target driver (ib_srpt) but is used by DDN arrays to transfer the
> indirect data buffer descriptor. DDN arrays support indirect data
> buffers that are larger than what fits in the SRP_CMD information unit
> but the upstream SRP target driver not.

It will be ready to go very early in the merge window, so they would
have plenty of time to test.  Even pre-merge window if they wanted to.

Ben, can you find out if there would be an ability to test an upstream
kernel against a DDN array where you are?

-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
              GPG KeyID: 0E572FDD



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]