Open CodeFetch opened 5 years ago
First, maybe see if this is only an issue with wave-1 ct firmware?
And second, try to create the simplest test case that uses little or no mesh logic. That would give me a better chance of reproducing and fixing the problem locally.
Try disabling all encryption to see if that helps, see if any packets hit the air using a sniffer, etc.
For what it is worth, I expect that IBSS should coexist with station devices on all of the ct firmware.
Did you have any luck testing this?
Unfortunately I didn't have time to test it, yet. QCA9886 is a Wave 1 chip, isn't it? I've read on the OpenWrt mailing list, that you've just released a new version. I'm going to test it with that one. Yesterday someone gave me three UniFI AC Mesh. Their chip seems to be a QCA9888 clone, which is a Wave 2 chip. I'm going to test them, too.
BTW I have some stupid questions. Are you planning to support 802.11ax chips like IPQ8074? And another one is: As you've signed a NDA, does Qualcomm offer you to get insight into the topical code of the "official" ath10k firmware? I ask because Gluon uses your driver only for IBSS as the "official" firmware seems to be more stable for 802.11s.
I do plan to support 11ax chips. I'd be interested in debugging 11s issues in my wave-2 firmware. I am not sure if the 'gluon' project had issues with wave-1 or wave-2 firmware. I do not have plans to support .11s in wave-1 firmware.
According to https://github.com/freifunk-gluon/gluon/issues/1584 two VIFs one being an IBSS and the other an AP don't work in parallel. I'm going to make some tests myself and try to provide logs. Maybe this is related to #68. This is a rather critical issue for mesh networks. Please tell me if you have some ideas what I can do to help you resolve it.