Re: [PATCH 1/8] bindings: PCI: designware: Example update

[Kishon Vijay Abraham I]

On Tuesday 03 April 2018 04:03 PM, Gustavo Pimentel wrote:
> Hi Kishon,
> 
> On 02/04/2018 06:23, Kishon Vijay Abraham I wrote:
>> Hi,
>>
>> On Wednesday 28 March 2018 05:08 PM, Gustavo Pimentel wrote:
>>> Changes the IP registers size to accommodate the ATU unroll space.
>>>
>>> Replaces "ctrlreg" reg-name by "dbi" to be coherent with similar drivers.
>>>
>>> Replaces the pcie base address example by a real pcie base address in use.
>>>
>>> Signed-off-by: Gustavo Pimentel <gustavo.pimentel@xxxxxxxxxxxx>
>>> ---
>>>  Documentation/devicetree/bindings/pci/designware-pcie.txt | 12 ++++++------
>>>  1 file changed, 6 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/pci/designware-pcie.txt b/Documentation/devicetree/bindings/pci/designware-pcie.txt
>>> index 1da7ade..6300762 100644
>>> --- a/Documentation/devicetree/bindings/pci/designware-pcie.txt
>>> +++ b/Documentation/devicetree/bindings/pci/designware-pcie.txt
>>> @@ -1,7 +1,8 @@
>>>  * Synopsys DesignWare PCIe interface
>>>  
>>>  Required properties:
>>> -- compatible: should contain "snps,dw-pcie" to identify the core.
>>> +- compatible:
>>> +	"snps,dw-pcie" for RC mode;
>>
>> I think irrespective of RC mode or EP mode, "snps,dw-pcie" can be used to
>> identify the pcie core?
> 
> I guess so. What you suggest? I was just folling the same guideline present
> here: https://lkml.org/lkml/2017/11/3/310

Okay, I think you should have
"snps,dw-pcie-rc", "snps,dw-pcie" for RC mode;

and in the later patch
"snps,dw-pcie-ep", "snps,dw-pcie" for EP mode;

> 
>>>  - reg: Should contain the configuration address space.
>>>  - reg-names: Must be "config" for the PCIe configuration space.
>>>      (The old way of getting the configuration address space from "ranges"
>>> @@ -41,11 +42,11 @@ EP mode:
>>>  
>>>  Example configuration:
>>>  
>>> -	pcie: pcie@dffff000 {
>>> +	pcie: pcie@dfc00000 {
>>>  		compatible = "snps,dw-pcie";
>>> -		reg = <0xdffff000 0x1000>, /* Controller registers */
>>> -		      <0xd0000000 0x2000>; /* PCI config space */
>>> -		reg-names = "ctrlreg", "config";
>>> +		reg = <0xdfc00000 0x302000>, /* IP registers */
>>
>> which version of synopsys IP is this. I think the ideal thing to do here is to
>> have a separate register space for iATU.
> 
> I also agree with that. The unroll iATU was introduced on Synopsys IP version
> 4.80 and the kernel patch was release on 2016-08-10
> https://patchwork.ozlabs.org/patch/657796/
> However a separate register space for iATU implies some extra code do handle it
> (and of course some tests) that don't fit into this patch series, in my point of
> view. Can this task enter in the backlog in order to be done in another patch
> series?

Yes sure. I think we should also make sure existing binding doesn't break.

Thanks
Kishon