Re: [PATCH 05/13] wifi: brcmfmac: Negotiate PCIe mailbox transport with firmware

[André Eikmeyer]

On Fri, 06 Nov 2026 18:36:52 +0100
Arend van Spriel <arend.vanspriel@xxxxxxxxxxxx> wrote:

> Sorry, you lost me there. What is the relation between the patch in
> this subject and D3 ack.

Hi Arend,

D3_INFORM and D3_ACK are mailbox messages, and 05 changes how they are
exchanged. Without 05 brcmfmac writes D3_INFORM to the TCM mailbox
and our cursed Apple 4377 firmware on the 9,1 doesn't read it.

That is why I mentioned the D3 ACK timeout handling patch. I
submitted that because I was sure the firmware wouldn't send it.
But actually the transport was broken.

https://lore.kernel.org/all/20261001085308.1688049-1-dev@deq.rocks/

which was superseded by Sebastian's patch here:

https://lore.kernel.org/linux-wireless/20260915-brcmfmac-bcm43752-suspend-fix-v2-1-1bd84914998b@xxxxxxxxxxxxx/

So probably for addressing Sebastian's issue this needs to stay
anyways. My memory served me wrong. I just read that his device is not
an Apple product, but his patch incidentally worked around the same
problem. Sorry if i was causing unneccessary noise. 

Andre