UPDATE - 4th of December 2020:
This puts back the 80mhz frequency on the SPI and fixed by using the mtd fast-read feature to stop the race condition.
Also some Wifi fixes - hopefully fixing the hardware reset issues and enabling a few more features
- ACS floor scanning
- P2P client and GO
This will probably be my last update as from me, as everything is working so far and will just be keeping my code up to date when new security updates come along.
When I get time, will try to push the patches upstream in the openwrt V19.07 branch.
Enjoy!
Original message:
Hi peeps,
So rather than looking through a thousand page topic, i've started this thread to get some feedback from owners or devs on my recent updates, fixes to this device.
Im not trying to taking any credit from the original thread - Xiaomi Mi Router 4A Gigabit Edition (R4AG/R4A Gigabit) -- fully supported and flashable with OpenWRTInvasion
Just getting to the bottom of the recent boot failures people and myself had experienced with this device.
My gitlab repo for the source code if you wish to fork or build:
https://gitlab.com/db260179/xiaomi-m4a.git
- Issue i had faced where - Squashfs data corruption
My original suggestion was the block size was the issue. This was incorrect.
After many days of testing and low level hardware probing, it appears that some of these boards might have a manufactuer issue with the spi chip clock speed - It is rated at 80mhz capable speed, but it seems that running at that speed causes race conditions on bootup.
I had to lower it all the way to 33mhz to create a stable no corrupted running mode.
Since this change i no longer get the squashfs corruption issue.
As there where no official patch for this device, only based on the work of the similiar device 3Gv2, the two settings don't quite fit for this device.
Feedback would be gratefully appreciated!