Re: [PATCH 2/2] wifi: ath11k: use unique QRTR instance ID

From: Kalle Valo <kvalo@kernel.org>
To: Robert Marko <robert.marko@sartura.hr>
Cc: Manivannan Sadhasivam <mani@kernel.org>,
	Robert Marko <robimarko@gmail.com>,
	davem@davemloft.net, edumazet@google.com, kuba@kernel.org,
	pabeni@redhat.com, gregkh@linuxfoundation.org, elder@linaro.org,
	hemantk@codeaurora.org, quic_jhugo@quicinc.com,
	quic_qianyu@quicinc.com, bbhatt@codeaurora.org,
	mhi@lists.linux.dev, linux-arm-msm@vger.kernel.org,
	ath11k@lists.infradead.org, linux-wireless@vger.kernel.org,
	netdev@vger.kernel.org, ansuelsmth@gmail.com
Subject: Re: [PATCH 2/2] wifi: ath11k: use unique QRTR instance ID
Date: Thu, 22 Dec 2022 15:57:32 +0200	[thread overview]
Message-ID: <87k02jzgkz.fsf@kernel.org> (raw)
In-Reply-To: <CA+HBbNFCFtJwzN=6SCsWnDmAjPkmxE4guH1RrLc+-HByLcVVXA@mail.gmail.com> (Robert Marko's message of "Wed, 14 Dec 2022 13:02:42 +0100")

Robert Marko <robert.marko@sartura.hr> writes:

> On Tue, Nov 22, 2022 at 12:26 PM Kalle Valo <kvalo@kernel.org> wrote:
>
>>
>> Kalle Valo <kvalo@kernel.org> writes:
>>
>> > Manivannan Sadhasivam <mani@kernel.org> writes:
>> >
>> >> On Sat, Nov 05, 2022 at 08:49:43PM +0100, Robert Marko wrote:
>> >>> Currently, trying to use AHB + PCI/MHI cards or multiple PCI/MHI cards
>> >>> will cause a clash in the QRTR instance node ID and prevent the driver
>> >>> from talking via QMI to the card and thus initializing it with:
>> >>> [    9.836329] ath11k c000000.wifi: host capability request failed: 1 90
>> >>> [    9.842047] ath11k c000000.wifi: failed to send qmi host cap: -22
>> >>>
>> >>
>> >> There is still an outstanding issue where you cannot connect two WLAN modules
>> >> with same node id.
>> >>
>> >>> So, in order to allow for this combination of cards, especially AHB + PCI
>> >>> cards like IPQ8074 + QCN9074 (Used by me and tested on) set the desired
>> >>> QRTR instance ID offset by calculating a unique one based on PCI domain
>> >>> and bus ID-s and writing it to bits 7-0 of BHI_ERRDBG2 MHI register by
>> >>> using the SBL state callback that is added as part of the series.
>> >>> We also have to make sure that new QRTR offset is added on top of the
>> >>> default QRTR instance ID-s that are currently used in the driver.
>> >>>
>> >>
>> >> Register BHI_ERRDBG2 is listed as Read only from Host as per the BHI spec.
>> >> So I'm not sure if this solution is going to work on all ath11k supported
>> >> chipsets.
>> >>
>> >> Kalle, can you confirm?
>> >
>> > I can't look at this in detail right now, but hopefully in few days.
>> > I'll get back to you.
>>
>> The solution we have been thinking internally would not use
>> MHI_CB_EE_SBL_MODE at all, it's not clear for me yet why the mode was
>> not needed in our solution. Maybe there are firmware modifications? I
>> think it's best that we submit our proposal as well, then we can then
>> compare implementations and see what is the best course of action.
>
> Kalle, any ETA when you will post your idea? I am constantly hitting
> this crazy limitation and my idea does not work on cards like QCA6390
> so it's not a viable workaround at all.

Really sorry, I just didn't manage to get this finalised due to other
stuff and now I'm leaving for a two week vacation :(

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches