On 30.11.21 00:38, Bjorn Helgaas wrote:

On Mon, Nov 29, 2021 at 05:32:11PM -0600, Bjorn Helgaas wrote:

quoted

On Wed, Nov 03, 2021 at 09:58:27PM +0100, Jonas Dreßler wrote:

quoted

The firmware of the 88W8897 PCIe+USB card sends those events very
unreliably, sometimes bluetooth together with 2.4ghz-wifi is used and no
COEX event comes in, and sometimes bluetooth is disabled but the
coexistance mode doesn't get disabled.

s/sends those events/sends BTCOEX events/ so it reads well without the
subject.

s/coexistance/coexistence/

Is BTCOEX a standard Bluetooth thing?  Is there a spec reference that
could be useful here?  I've never seen those specs, so this is just
curiosity.  I did download the "Bluetooth Core Spec v5.3", which does
have a "Wireless Coexistence Signaling and Interfaces" chapter, but
"BTCOEX" doesn't appear in that doc.

quoted

This means we sometimes end up capping the rx/tx window size while
bluetooth is not enabled anymore, artifically limiting wifi speeds even
though bluetooth is not being used.

s/artifically/artificially/

quoted

Since we can't fix the firmware, let's just ignore those events on the
88W8897 device. From some Wireshark capture sessions it seems that the
Windows driver also doesn't change the rx/tx window sizes when bluetooth
gets enabled or disabled, so this is fairly consistent with the Windows
driver.

I hadn't read far enough to see that the patch was already applied,
sorry for the noise :)

No problem, in case you still want to know about BTCOEX:

 From what I've seen that's not something defined in any standards, but
it's usually the name of the (sometimes patented) tricks every manufacturer
has to make wifi and bt (which are both on the 2.4ghz band) behave well
together.

In almost every wifi driver you'll find functionality named
btcoex/coexist/coexistence. The way it usually works is that the card
sends an event to the kernel driver (in our case that event is called
BTCOEX), and then the driver decides which quirks to apply to make wifi
more interference-resistant (here's where the patents come in because
some of those quirks are quite tricky, see for example
https://patents.google.com/patent/US9226102B1/en).

Now with our Marvell card the firmware is buggy and sends those events
so unreliably (the card "forgets" to inform us that the BT connection
has ended) that we're better off ignoring them.