[PATCH v6 0/7] PCI: get DMA configuration from parent device

[Murali Karicheri]

On 02/06/2015 12:53 PM, Bjorn Helgaas wrote:
> On Fri, Feb 6, 2015 at 9:28 AM, Murali Karicheri<m-karicheri2 at ti.com>  wrote:
>> On 02/06/2015 10:15 AM, Catalin Marinas wrote:
>>>
>>> On Thu, Feb 05, 2015 at 09:52:52PM +0000, Murali Karicheri wrote:
>>>>
>>>> This patch add an important capability to PCI driver on Keystone. I hope
>>>> to
>>>> have this merged to the upstream branch so that it is available for
>>>> v3.20.
>>>
>>> It's very late for 3.20 and the code hasn't been in linux-next at all
>>> (but it's not me who's merging this code), unless you can squeeze it in
>>> as a bug-fix.
>>
>> This is in fact a bug fix as PCI on Keystone is broken without this.
>
> Oh, sorry, I didn't realize that this was so urgent.  I guess I read
> "this adds an important capability" in the cover letter and concluded
> that it was new functionality.
Bjorn,

Thanks for responding.

Let me give you some context on this without which my explanation won't 
be complete. For using PCI driver on Keystone, I had submitted patches 
related to machine and DTS to the arm mailing list to enable the driver 
on Keystone. Subsequenty one of the patch from my series  was Nack-ed 
and I was asked to implememented in a different way and started this 
series. You can refer to the discussion of this at

http://www.gossamer-threads.com/lists/linux/kernel/2024591

The PCI driver enablement on Keystone is still a working in progress and 
I am trying to get it fully functional on the upstream. Another missing 
piece is the SerDes phy driver patch. We have started working on the 
other part (SerDes phy driver) already as the initial one was not 
accepted. So it is fine if we are too late for the v3.20 merge window to 
merge this series and this can be applied to the next branch for v3.21.

>
> Anyway, if it's broken, presumably PCI on Keystone *did* work at one
> point.  Can you identify the commit that broke it and requires these
> fixes, so we can figure out how far the fixes need to be backported?
>

I am trying to get this driver enabled on Keystone by adding the missing 
pieces as described above. So I guess we don't have to back port 
anything here.

> If I merge it, I would like to get into my for-linus branch and get a
> little time in -next before asking Linus to pull it.  The merge window
> is a little wrinkle there -- I don't like to add new things to the mix
> during the window.  But if it's an important fix we can still get it
> in before the final v3.20.

Please apply this to next branch for v3.21. It currently apply cleanly 
to v3.19-rc7. If you want me rebase to another branch, let me know I can 
apply and send you an updated patch.

Thanks

Murali

>
> Bjorn


-- 
Murali Karicheri
Linux Kernel, Texas Instruments