On Dienstag, 28. August 2018 09:38:39 CEST John Crispin wrote: > > +gl-x1200) > > + ucidef_set_led_wlan "wlan2g" "WLAN2G" "$board:green:wlan2g" > > "phy1tpt" > > + ucidef_set_led_wlan "wlan5g" "WLAN5G" "$board:green:wlan5g" > > "phy0tpt" > > + ;; > > this section aswell as several other places in this patch is not > alphabetically ordered. please fix all occurrences
And this section could also be merged together with the gl-ar750s section.
> > + gl-x1200)
> > + ath10kcal_extract "art" 20480 12064
> > + ln -sf /lib/firmware/ath10k/pre-cal-pci-0000\:00\:00.0.bin \
> > + /lib/firmware/ath10k/QCA9888/hw2.0/board.bin
> > + ath10kcal_patch_mac $(macaddr_add $(cat
> > /sys/class/net/eth0/address) +1)
> > + ;;
[...]
> > + DEVICE_PACKAGES := kmod-ath10k ath10k-firmware-qca9887 kmod-usb-core \
> > + kmod-usb2 kmod-usb-storage
Is the 5GHz now a QCA9888 or a QCA9887? And why are you linking the pre-cal
data to board.bin? QCA9888 has its BDF stored in the board-2.bin [1] (selected
by
the bmi from the OTP and the variant string from DT/SMBIOS) and per device
pre-cal data is stored in /lib/firmware/ath10k/pre-cal-*.bin. The OTP is then
merging both and applying extra modifications on top of it. The result is then
automatically given to the actual QCA9888 firmware.
See for example the OpenMesh A62 for an device with an QCA9888. As you may
notice, it is using device tree - something which ar71xx is not using. But the
new target ath79 does.
Kind regards,
Sven
[1] https://github.com/openwrt/openwrt/pull/713#issuecomment-369511149
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
