ACK: Xenial SRU - CAPI: CGZIP AFU contexts do not receive interrupts after heavy afu open/close

[Stefan Bader]

On 08.06.2016 18:01, Tim Gardner wrote:
> On 06/08/2016 09:17 AM, Tim Gardner wrote:
>> On 06/08/2016 09:03 AM, Stefan Bader wrote:
>>> On 08.06.2016 16:52, Tim Gardner wrote:
>>>> On 06/08/2016 08:35 AM, Stefan Bader wrote:
>>>>> On 03.06.2016 21:19, Tim Gardner wrote:
>>>>>> The fix commit 2bc79ffcbb817873cc43d63118008ab75181b73d
>>>>>> ('cxl: Poll for outstanding IRQs when detaching a context')
>>>>>> requires a bunch of scaffolding. Since we already have a mix
>>>>>> of interdependent cxl and cxlflash patches I decided to just
>>>>>> all cxl/cxlflash patches up to and including the fix commit. 
>>>>>> Regression testing is straightforward since this driver only 
>>>>>> exists on one platform and architecture.
>>>>>>
>>>>>> See attached pull request.
>>>>
>>>>> Dunno, using the pull request yields more that just cxl changes
>>>>> for me (as does the diffstat of the pull request). I might be
>>>>> doing something wrong...
>>>>
>>>>
>>>> Details ?
>>
>>> drivers/scsi/mesh.c                        |    2 +- 
>>> drivers/usb/core/hcd-pci.c                 |    1 - 
>>> drivers/video/fbdev/aty/aty128fb.c         |    1 - 
>>> drivers/video/fbdev/aty/radeon_base.c      |    1 - 
>>> drivers/video/fbdev/imsttfb.c              |    1 - 
>>> drivers/video/fbdev/matrox/matroxfb_base.h |    1 - 
>>> drivers/video/fbdev/offb.c                 |    4 - 
>>> include/misc/cxl.h                         |    8 +- 
>>> include/uapi/misc/cxl.h                    |   24 + 
>>> sound/ppc/pmac.c                           |    1 -
>>>>
>>>>
>>
>>
>>
>> This isn't really a merge request. The patches up for review are from
>>  41b48dd5672ebaf077cc122af6b921ae81945753..7afdb0a5ea5a068eb9c91b110d276
>> 115d049761b as indicated in the pull request. The master or master-next
>> branch that you have has clearly moved since this pull request was origi
>> nally emitted.
>>
>> rtg
>>
> 
> OK, I think see where the problem lies. I've refreshed that branch after
> having dropped one of the scaffold patches. It turns out it was not
> required, and touched a bunch of drivers outside of cxl.

Ah ok, that looks much better. There is only one change outside the cxl code
left and that adds a return value which is introduced by the driver changes.
Hopefully this can be verified in testing in the correct environment because the
number of changes are too big for a reasonable code review.

-Stefan
> 
> rtg
> 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20160608/ecb7e659/attachment.sig>