[RESEND PATCH V4] staging: vchiq_arm: Add compatibility wrappers for ioctls

Stefan Wahren stefan.wahren at i2se.com
Mon Mar 6 08:11:34 PST 2017


Am 06.03.2017 um 16:57 schrieb Michael Zoran:
> On Mon, 2017-03-06 at 14:01 +0100, Greg KH wrote:
>> On Sat, Mar 04, 2017 at 06:57:50AM -0800, Michael Zoran wrote:
>>> On Sat, 2017-03-04 at 14:58 +0300, Dan Carpenter wrote:
>>>> On Wed, Mar 01, 2017 at 07:41:46PM -0800, Michael Zoran wrote:
>>>>> Changes:
>>>>> 	V1 - Complete rewrite of the ioctl code.
>>>>> 	V2 - Rewrite of only ioctls that change
>>>>>              between 32 bit and 64 bit.
>>>>>         V3 - Minor changes.
>>>>> 	V4(This Version) - Abandon cleaning up the
>>>>> 	     exising code and completely write the
>>>>> 	     wrappers on top of the native ioctls.
>>>>> 	     No existing lines are changed.
>>>> I think I objected to earlier versions but this approach looks
>>>> nice.
>>>>
>>>> regards,
>>>> dan carpenter
>>>>
>>> Cool, does this mean I get an ACK?
>>>
>>> It really would be good to get this in since it's one of the last
>>> major
>>>   missing pieces to get Raspbian to work with a 64 bit kernel.
>> I'll be glad to take this if you respin it with the fixes that Eric
>> asked for :)
>>
>> thanks,
>>
>> greg k-h
> Hi Greg, 
>
> I have no problem doing a v5, just that Eric's fixes for await
> completion will not work at all.
>
> Only handling the case of 1 passed in would completely break vchiq_test
> which is the main test application and would break all of 32 bit user
> mode which would defeat the point of the wrappers.

If we couldn't find a solution, it would be nice to have an explaining
comment at this place.

>
> If Eric or anybody else has another way to fix await completion that
> will actually work, I'm more then willing to do it that way.

Regarding to my comments it's okay to keep the memdump and introduce the
Kconfig parameter for all dumps later.

Stefan



More information about the linux-rpi-kernel mailing list