Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH v3 0/6] PCI: J7200/J721E PCIe bindings
To:     Nishanth Menon <nm@ti.com>
CC:     Rob Herring <robh+dt@kernel.org>,
        <linux-arm-kernel@lists.infradead.org>,
        <devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>
References: <20210104122232.24071-1-kishon@ti.com>
 <20210104125910.qaf7vi3dx6vsne6t@backfield>
 <465097c1-2d38-ad45-cc32-d92c385114c5@ti.com>
 <20210104131642.m5rgpkhnrffq5nrx@circular>
 <d50fa957-7371-090a-754c-816760b43a27@ti.com>
 <20210104145137.wmtppcvjzu374yly@kahuna>
From:   Kishon Vijay Abraham I <kishon@ti.com>
Message-ID: <2c0846ff-3565-5d95-3408-b8b629a9fb48@ti.com>
Date:   Mon, 4 Jan 2021 21:01:06 +0530
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.10.0
MIME-Version: 1.0
In-Reply-To: <20210104145137.wmtppcvjzu374yly@kahuna>
Content-Type: text/plain; charset="utf-8"
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

Nishanth,

On 04/01/21 8:21 pm, Nishanth Menon wrote:
> On 18:52-20210104, Kishon Vijay Abraham I wrote:
>> Nishanth,
>>
>> On 04/01/21 6:46 pm, Nishanth Menon wrote:
>>> On 18:40-20210104, Kishon Vijay Abraham I wrote:
>>>> Nishanth,
>>>>
>>>> On 04/01/21 6:29 pm, Nishanth Menon wrote:
>>>>> On 17:52-20210104, Kishon Vijay Abraham I wrote:
>>>>>> Patch series adds DT nodes in order to get PCIe working in J7200.
>>>>>> Also includes couple of fixes for J721e.
>>>>>>
>>>>>> v1 of the patch series can be found @ [1]
>>>>>> v2 of the patch series can be found @ [2]
>>>>>>
>>>>>> Changes from v2:
>>>>>> 1) Moved serdes_refclk node out of interconnect node and also replaced
>>>>>>    "_" with "-"
>>>>>>
>>>>>> Changes from v1:
>>>>>> 1) Include only the device tree patches here (the binding patch is sent
>>>>>> separately)
>>>>>> 2) Include couple of patches that fixes J721E DTS.
>>>>>>
>>>>>> [1] -> http://lore.kernel.org/r/20201102101154.13598-1-kishon@ti.com
>>>>>> [2] -> http://lore.kernel.org/r/20201210130747.25436-1-kishon@ti.com
>>>>>>
>>>>>> Kishon Vijay Abraham I (6):
>>>>>>   arm64: dts: ti: k3-j721e-main: Fix supported max outbound regions
>>>>>>   arm64: dts: ti: k3-j721e-main: Remove "syscon" nodes added for
>>>>>>     pcieX_ctrl
>>>>>>   arm64: dts: ti: k3-j7200-main: Add SERDES and WIZ device tree node
>>>>>>   arm64: dts: ti: k3-j7200-main: Add PCIe device tree node
>>>>>>   arm64: dts: ti: k3-j7200-common-proc-board: Enable SERDES0
>>>>>>   arm64: dts: ti: k3-j7200-common-proc-board: Enable PCIe
>>>>>>
>>>>>>  .../dts/ti/k3-j7200-common-proc-board.dts     |  38 ++++++
>>>>>>  arch/arm64/boot/dts/ti/k3-j7200-main.dtsi     | 111 ++++++++++++++++++
>>>>>>  arch/arm64/boot/dts/ti/k3-j721e-main.dtsi     |  52 ++------
>>>>>>  3 files changed, 157 insertions(+), 44 deletions(-)
>>>>>
>>>>>
>>>>> A bit confused on the dependency here. is there something merged into
>>>>> next-20210104 that makes this series ready for pickup? is there a way
>>>>> I can get a immutable tag for driver fixups to pull so that my dts
>>>>> next is not broken for PCIe (I am assuming looking at the series that
>>>>> this is probably not a backward compatible series?)?
>>>>
>>>> There are no driver changes for the basic J7200 PCIe support and the DT
>>>> bindings are already merged [1].
>>>>
>>>> There are few errata fixes applicable for J721E which has to be removed
>>>> for J7200 but that depends on other patches to be merged [1] but that
>>>> doesn't impact j7200 functionality.
>>>>
>>>> [1] ->
>>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/pci/ti,j721e-pci-host.yaml#n19
>>>> [2] -> http://lore.kernel.org/r/20201230120515.2348-1-nadeem@cadence.com
>>>
>>> So, Dropping stuff like "cdns,max-outbound-regions" (change from 16 to
>>> 32) will work on older kernels? Could you do a quick sanity check on the
>>> couple of "fixes" patches in this thread is not breaking functionality
>>> introduced in the older stable kernels?
>>
>> Okay, the driver changes are done such that it works with both old DT
>> and new DT however the newer DT will not work with old kernel. So I
>> think I should drop the "Fixes" tag in the DT patches.
> 
> If there is a specific stable kernel version you might like to use, you
> could use that as well for those stable tags (see [1])

It's not a fix that might bother people (as it doesn't impact existing
functionality), so don't think it's a candidate for stable releases.

Thanks
Kishon
> 
> 
> [1] https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
>