MR24 wifi broken by recent commit

[Christian Lamparter]

Almost lost this in the spam-folder :( .

On 2020-09-03 23:22, Russell Senior wrote:
>>>>>> Christian Lamparter <chunkeey at gmail.com> writes:
> 
>> On 2020-09-03 10:44, Russell Senior wrote:
>>> commit 9153955095f01a7ac5f2659a671f0229cbad3507 Author: Christian
>>> Lamparter <chunkeey at gmail.com <mailto:chunkeey at gmail.com>> Date:
>>> Wed Aug 12 18:26:43 2020 +0200
>>>
>>>      apm821xx: MR24: enumerate PCIe in device-tree
>>>
>>>      This patch adds the pcie-switch and bridge configuration for
>>> the Meraki MR24.
>>>
>>>      Signed-off-by: Christian Lamparter <chunkeey at gmail.com
>>> <mailto:chunkeey at gmail.com>>
>>>
>>> The symptom is client devices can't see the beacons. Wifi ifaces
>>> appear, can scan and hear other networks, but clients can't see the
>>> MR24's SSIDs. Reverting the commit above and it works normally again.
> 
>> Thanks for the report. CONFIG_PCI_DEBUG
> 
> 
> 
>> "ath9k 0000:44:00.0: runtime IRQ mapping not provided by arch"
> 
> 
> 
>> since DT overwrites the usual PCI(e) enumeration and MSI(x) is
> 
>> busted there's no choice but to specify legacy INTABCDs.
> 
> 
>> Can you please try the attached patch?
> 
> Your patch fixes it.
> 
> I am a little curious, if it works without your commit, what is the
> motivation for the commit?
> 

APM82181 needs to move forward. I'm trying to upstream these devices.
Since nobody is commenting on linux-ppcdev, I have to spin the wheels
and preemptively spif-things up from version to version.

In the end, I'm hoping to avoid smart questions about:
"hey, why does this dual-band accesspoint device of yours not define a
single wifi-device in the DT? Like is this even working or what?"

You could tell the linux-ppcdev about your experience with the MR24
by replying to the latest v3 series:
<https://lists.ozlabs.org/pipermail/linuxppc-dev/2020-September/217642.html>

Cheers,
Christian