[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <26a86c55-7c35-484a-2e6f-9d5c36b078f4@nvidia.com>
Date: Tue, 15 Jun 2021 08:58:26 +0530
From: Vidya Sagar <vidyas@...dia.com>
To: Thierry Reding <treding@...dia.com>,
"bhelgaas@...gle.com" <bhelgaas@...gle.com>,
"lorenzo.pieralisi@....com" <lorenzo.pieralisi@....com>
CC: Gustavo Pimentel <Gustavo.Pimentel@...opsys.com>,
"robh+dt@...nel.org" <robh+dt@...nel.org>,
"amurray@...goodpenguin.co.uk" <amurray@...goodpenguin.co.uk>,
"jingoohan1@...il.com" <jingoohan1@...il.com>,
Joao Pinto <Joao.Pinto@...opsys.com>,
Jonathan Hunter <jonathanh@...dia.com>,
Krishna Thota <kthota@...dia.com>,
"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: Query regarding the use of pcie-designware-plat.c file
Bjorn / Lorenzo,
Could you please comment on the options proposed by Thierry?
- Vidya Sagar
On 6/9/2021 3:19 PM, Thierry Reding wrote:
> On Wed, Jun 09, 2021 at 10:56:48AM +0530, Vidya Sagar wrote:
>>
>>
>> On 6/9/2021 2:47 AM, Gustavo Pimentel wrote:
>>> External email: Use caution opening links or attachments
>>>
>>>
>>> Hi Vidya,
>>>
>>> The pcie-designware-plat.c is the driver for the Synopsys PCIe RC IP
>>> prototype.
>> Thanks for the info Gustavo.
>> But, I don't see any DT file having only "snps,dw-pcie" compatibility
>> string. All the DT files that have "snps,dw-pci" compatibility string also
>> have their platform specific compatibility string and their respective host
>> controller drivers. Also, it is the platform specific compatibility string
>> that is used for binding purpose with their respective drivers and not the
>> "snps,dw-pcie". So, wondering when will pcie-designware-plat.c be used as
>> there is not DT file which has only "snps,dw-pcie" as the compatibility
>> string.
>
> Sounds to me like we have two options:
>
> 1. If there's indeed real hardware that's identified by the existing
> "snps,dw-pcie" compatible string, then it's wrong for other devices
> to list that in their compatible string because they are likely not
> compatible with that (i.e. they might be from a register point of
> view, but at least from an integration point of view they usually
> differ).
>
> 2. If "snps,dw-pcie" is meant to describe the fact that these are all
> based off the same IP but may be differently integrated, then there
> should be no driver matching on that compatible string.
>
> Option 2 is not very robust because somebody could easily add a matching
> driver at some point in the future. Also, if we don't match on a
> compatible string there's not a lot of use in listing it in DT in the
> first place.
>
> So I think option 1 would be preferred.
>
> Thierry
>
Powered by blists - more mailing lists