Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <6e204026-d102-6d81-f58b-f4340e7d07ac@marcan.st>
Date:   Mon, 2 Jan 2023 23:01:10 +0900
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.9.1
Subject: Re: [PATCH v2 2/4] brcmfmac: Support 89459 pcie
Content-Language: en-US
To:     Arend van Spriel <arend.vanspriel@broadcom.com>,
        Aditya Garg <gargaditya08@live.com>
Cc:     Ian Lin <ian.lin@infineon.com>,
        "alep@cypress.com" <alep@cypress.com>,
        "brcm80211-dev-list@broadcom.com" <brcm80211-dev-list@broadcom.com>,
        "brcm80211-dev-list@cypress.com" <brcm80211-dev-list@cypress.com>,
        "franky.lin@broadcom.com" <franky.lin@broadcom.com>,
        "hante.meuleman@broadcom.com" <hante.meuleman@broadcom.com>,
        "kvalo@kernel.org" <kvalo@kernel.org>,
        "Double.Lo@infineon.com" <Double.Lo@infineon.com>,
        "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Asahi Linux <asahi@lists.linux.dev>
References: <20220922104140.11889-1-ian.lin@infineon.com>
 <20220922104140.11889-3-ian.lin@infineon.com>
 <3eb9cc12-db75-8042-4106-fa6521c684a7@marcan.st>
 <6D6035D7-894B-44F6-B35A-3EC9C57029B5@live.com>
 <91c8fc96-083d-74b7-3015-e76f1ab59cec@broadcom.com>
From:   Hector Martin <marcan@marcan.st>
In-Reply-To: <91c8fc96-083d-74b7-3015-e76f1ab59cec@broadcom.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 02/01/2023 22.58, Arend van Spriel wrote:
> On 1/2/2023 10:45 AM, Aditya Garg wrote:
>> Hi Hector
>>
>>>
>>> Is the CYW89459 just a rebrand of the BCM4355, or just a subset? If it
>>> is a rebrand, it's okay if we call our Apple firmware
>>> brcmfmac89459-pcie* (note that we use per-board firmware names, so it
>>> wouldn't conflict with a generic one). However, if CYW89459 only refers
>>> to specific variants, I think the firmware should be named after the
>>> overall bcm4355 family.
>>>
>>> I'm guessing you intend to ship firmware for this. Would that firmware
>>> work for all 4355 variants, or only the CYW one? If only the CYW one, is
>>> it possible to differentiate between them based on PCI revision ID? Note
>>> that our 4355 has revision ID 12, and Apple specifically calls it 4355C1
>>> (different chip revisions have different firmware builds, which is why I
>>> named our firmware brcmfmac4355c1-pcie). If the CYW variant uses other
>>> revision IDs that do not overlap, maybe we should have different
>>> firmware entries for them with different masks.
>>
>>
>> Can we make a separate table for the OTP Apple chips, something like here :-
>>
>> https://github.com/AdityaGarg8/linux/commit/fc41aac9283d2ba653a8b3191e8c0138c13d8ee1
> 
> I do not understand from this email thread why you would need separate 
> tables. Can you explain?
> 

I think he's proposing we special-case Apple chips into their own
firmware table just to avoid colliding with non-Apple firmware usage,
which is honestly kind of tempting as the safe option if nobody from the
Broadcom/Cypress side is willing to clarify what, exactly, is the
relationship between these chips and what their respective revision
numbers are so we can *correctly* represent them and avoid further
confusion and problems down the line.

You might be able to help with that ;)

- Hector