nope, none of them.
But the ath11k adapter using a mpci->m.2 chip does.
So, while it looks weird, do I need to consider all 3 ath10k mpci cards to be broken?
No, I don’t think the cards are broken. It can be, however, that the PCIe host weirdly rejects them because they only use one out of two available lanes. It shouldn’t do that obviously and I will also try a single lane mPCIe card later today.
I also tested mt7615 in both pci slots (of course sw4 off) and they worked,so nothing specific to the bpi wifi board. But atheros cards had some issues in past with bpi boards (because of missing capacitors or resistors).
Does this impact all cards? Or only 988x based chips?
I do not know,but r64 had some issues with some cards. Afaik there are no reports for r3/r4 yet.
and I assume a workaround isn’t possible if it’s due to missing capacitors …
do I need to add these cards to the wiki as “cards that are not bpi r4 compatible”? Or could there be some options so it can be brought alive after all (eg due to kernel using some pci quirks)?
I do not know if it is same issue on r3/r4 as i have not tried these cards yet. And solution for r64 was adding these capacitors/resistors
can I, based on a card’s datasheet, decide whether it will work or not? a sparklan 269ax ath11 card (ising mpcie->m.2 adapter) got detected, so is there some technical documentation about when it works and when it does not? if it’s the same issue compared to bpi r64, I may simply try to buy another mpcie card based on that documentation
i also had such a problem with 3 cards, 2 atheros (ath10k, ath9k) and 1 intel. none worked (was not recognized). now i bought a card that frank-w recommended me (Mediatek MTK MT7615 Minipci-express 802.11AC DBDC WIFI 5 Linux OpenWRT DD-WRT | eBay) and it works like a charm. its crazy… maybe you have the same problem?
btw. in my clearfog pro, all of those cards worked like a charm.
So it looks like mediatek cards work, and some other cards might, but we don’t know. Let’s hope it can be documented one day
Hi jpsollie,
Do you have links to these cards? I’ll try to buy and test it
I bought the ath10k cards at 524wifi.eu, so if you live in EU, simply pick one. I tested a 44, 33 and 2*2 mini pcie version, none worked
i could also ship you one or two of mine that do not work (if you live in EU). I have a ath9k/ath10k and an intel card that do not work in r4 but in a solidrun clearfog and in my desktop
Thank you for your kindness, I live in China, I will figure it out myself, THX again.
Thank you for your kindness, I live in China, I will figure it out myself, THX again.
I’m having a similar problem with a 5g card in the b-key slow (see GitHub issue here) and I’m seeing very similar issues in the kernel log.
[ 0.043952] mtk-pcie-gen3 11280000.pcie: host bridge /soc/pcie@11280000 ranges:
[ 0.043971] mtk-pcie-gen3 11280000.pcie: Parsing ranges property...
[ 0.043983] mtk-pcie-gen3 11280000.pcie: IO 0x0020000000..0x00201fffff -> 0x0020000000
[ 0.043995] mtk-pcie-gen3 11280000.pcie: MEM 0x0020200000..0x0027ffffff -> 0x0020200000
[ 0.044096] mtk-pcie-gen3 11290000.pcie: host bridge /soc/pcie@11290000 ranges:
[ 0.044107] mtk-pcie-gen3 11290000.pcie: Parsing ranges property...
[ 0.044116] mtk-pcie-gen3 11290000.pcie: IO 0x0028000000..0x00281fffff -> 0x0028000000
[ 0.044126] mtk-pcie-gen3 11290000.pcie: MEM 0x0028200000..0x002fffffff -> 0x0028200000
[ 0.044144] /soc/pcie@11290000: Failed to get clk index: 0 ret: -517
[ 0.044152] mtk-pcie-gen3 11290000.pcie: failed to get clocks
[ 0.044212] mtk-pcie-gen3 11300000.pcie: host bridge /soc/pcie@11300000 ranges:
[ 0.044221] mtk-pcie-gen3 11300000.pcie: Parsing ranges property...
[ 0.044230] mtk-pcie-gen3 11300000.pcie: IO 0x0030000000..0x00301fffff -> 0x0030000000
[ 0.044239] mtk-pcie-gen3 11300000.pcie: MEM 0x0030200000..0x0037ffffff -> 0x0030200000
[ 0.044254] /soc/pcie@11300000: Failed to get clk index: 0 ret: -517
[ 0.044262] mtk-pcie-gen3 11300000.pcie: failed to get clocks
[ 0.044332] mtk-pcie-gen3 11310000.pcie: host bridge /soc/pcie@11310000 ranges:
[ 0.044342] mtk-pcie-gen3 11310000.pcie: Parsing ranges property...
[ 0.044350] mtk-pcie-gen3 11310000.pcie: IO 0x0038000000..0x00381fffff -> 0x0038000000
[ 0.044359] mtk-pcie-gen3 11310000.pcie: MEM 0x0038200000..0x003fffffff -> 0x0038200000
[ 0.044374] /soc/pcie@11310000: Failed to get clk index: 0 ret: -517
[ 0.044381] mtk-pcie-gen3 11310000.pcie: failed to get clocks
...
[ 2.646088] mtk-pcie-gen3: probe of 11280000.pcie failed with error -110
[ 2.653130] mtk-pcie-gen3 11290000.pcie: host bridge /soc/pcie@11290000 ranges:
[ 2.660441] mtk-pcie-gen3 11290000.pcie: Parsing ranges property...
[ 2.666703] mtk-pcie-gen3 11290000.pcie: IO 0x0028000000..0x00281fffff -> 0x0028000000
[ 2.675138] mtk-pcie-gen3 11290000.pcie: MEM 0x0028200000..0x002fffffff -> 0x0028200000
[ 3.017289] mtk-pcie-gen3 11290000.pcie: PCIe link down, current LTSSM state: detect.quiet (0x1)
I think the first bridge is the mPCIe slots (where I have a QCNFA765 working fine - this is having no issues with the ath11k driver). The second and 3rd pci bridges seem like they are the B and M key slots.
What’s interesting is that I am able to get the 5g modem in the B slot to be recognized when I boot with the Bpi-provided image, but it’s not showing up in nightlies.
It seems like something to do with the device tree.
Any solution for QCA9880?
Thanks to the latest patches, I could make my qca9882 work, but not qca9880. as the qca9880 is pcie 1.1 only and a bit more power hungry, I guess you’re simply out of luck :s