Re: [PATCH v2 0/3] ntb_hw_switchtec: Added support of >=4G memory windows

[Logan Gunthorpe]

On 2018-12-12 5:25 p.m., Jon Mason wrote:
> On Wed, Dec 12, 2018 at 7:01 PM Logan Gunthorpe <logang@xxxxxxxxxxxx> wrote:
>>
>>
>>
>> On 2018-12-12 4:57 p.m., Jon Mason wrote:
>>> On Wed, Dec 12, 2018 at 6:42 PM Logan Gunthorpe <logang@xxxxxxxxxxxx> wrote:
>>>>
>>>>
>>>>
>>>> On 2018-12-12 4:00 p.m., Jon Mason wrote:
>>>>> So, you based your patches on a series of patches not in the
>>>>> ntb/ntb-next branch?  Please don't do this.  I see nothing in these
>>>>> patches which requires that series, which makes this even more
>>>>> unnecessary.  Since these are fairly trivial, I'm taking them and
>>>>> pushing to the ntb-next branch to give these more time to be tested
>>>>> (due to not being tested on the proper branch).  I would really
>>>>> appreciate you testing the ntb-next branch as a sanity check.
>>>>
>>>> The NTB test tools don't work with switchtec hardware without that patch
>>>> set, so there's no way to test the changes without that branch.
>>>
>>> Then let's get those patches in.  IIRC, I asked you to split up the
>>> patch series to be bugfixes and features (or at least reorder the
>>> series so I can split it up that way in my branches).  Also, I think
>>> Serge had some comments that may/may not need to be addressed.  Could
>>> you please reorder and resend (and Serge can comment as needed on the
>>> resend)?
>>
>> I resent a while back and responded to all the feedback. Every patch in
>> that series fixes a bug. None of them add features.
> 
> Per https://lkml.org/lkml/2018/6/12/552, I asked the series be split
> up and the comments to be cleaned-up.  You repushed without addressing
> this (or Serge's comments), which caused me to ignore the series.
> Again, I'm happy to take it as a single series and split it up on my
> end, I just need the patches reordered to have the bugfixes I
> specified in the front to allow for this to be easily done.

And what was not clear about my response?

https://lkml.org/lkml/2018/6/12/577

That commit is absolutely *not* a feature request. Without it, none of
the other fixes will fix anything and are thus worse than useless.

I responded to Serge's comments and then responded *again* in the cover
letter of v2. What he was asking for was, and still is, physically
impossible.

Logan