[LEDE-DEV] MT7620A uart/uartlite no console or shell i/o but dmesg/bootearly0 ok [Edimax BR-6478AC V2]

[Rohan Murch]

Hi all,

I have been scratching away at an Edimax "BR-6478AC V2" router for a
few weeks now and have mostly successfully ported OpenWrt DD to this
model. I am having some major issues with the uart and uartlite serial
ports on the chip.

There is most definitely two serial ports, and the factory firmware
showed "console=ttyS0,57600n8 console=ttyS1". I have found what is
ttyS1 (uartlite 0xc00) and the front of the board and can get a
working shell on factory firmware with this port. I can see what looks
like a 2nd uart port, a row of 4x 2.54mm spaced pins, but they share 4
pins of the middle ethernet port (LAN3).

Has anyone heard of a MT7620A uart at 0x500 (ttyS0) sharing 4 pins of
the ethernet switch?

Has anyone dealt with the issue of getty/agetty/mingetty quietly doing
nothing when pointed to an active console, but dying with stderr
output in boot up when deliberately misconfigured?

I can get a full complete boot, past preinit, init and procd finished,
but no shell login. I can successfully configure either just uartlite
0xc00 as ttyS0 only, or uartlite 0xc00 as ttyS1 and uart(full) 0xc00
as ttyS0. They both seem to be working ok according to the kernel boot
line "console=ttyS0,57600" or "console=ttyS1,57600" (ie: setting 0x500
as console causes 0xc00 to stop outputting after "disabled bootearly0"
or "early0". So it appears as though the kernel thinks it has
correctly grabbed a console to use, but there is nothing outputted,
and no terminal input is accepted.

Link to OpenWrt forum thread on this topic:
https://forum.openwrt.org/viewtopic.php?id=67192

Any help would be most greatly appreciated!

Kindest regards, Roh.