Add EMAX targets

[pkendall64]

Adds EMAX 2.4GHz and 900MHz targets:

• 2.4GHz OLED TX Module
• 2.4GHz RX (based on DIY/Generic target)
• 2.4GHz RX with PA (based on DIY/Generic target)
• 900MHz OLED TX Module
• 900MHz RX (based on DIY/Generic target)

[CapnBry]

I'm trying to test this on the EMAX OLED TX and it keeps locking up when I change the power level. EDIT: Actually it keeps locking up changing virtually any item in the Lua...

The output power doesn't go below 35mW either, so shouldn't this be one that starts at 25mW?

[pkendall64]

I'm trying to test this on the EMAX OLED TX and it keeps locking up when I change the power level. EDIT: Actually it keeps locking up changing virtually any item in the Lua...

The output power doesn't go below 35mW either, so shouldn't this be one that starts at 25mW?

I've remeasured mine now that I have a IRC power meter and it goes down to 11-12 dBm, so about 15mW on the 10mW setting.

I have seen a lockup but it's certainly not all the time, but I'll look into that.

[pkendall64]

@CapnBry I've tracked the lockup to a bug in the platform I2C code. It's the reason the Axis module was also locking up with the thermal device. I do not think it would wise to upgrade the platform at this late stage in the 3.0 release cycle, but we could upgrade early in 3.1.
One thing we can do is to move which devices are handled by core 0 and core 1, I did this as a test and if the radio, crsf and lua devices are on core 1 (the loop core) and all other ancillary devices are on core 0 then if the I2C does lock up at least the radio and LUA are still working.

[CapnBry]

One thing we can do is to move which devices are handled by core 0 and core 1, I did this as a test and if the radio, crsf and lua devices are on core 1 (the loop core) and all other ancillary devices are on core 0 then if the I2C does lock up at least the radio and LUA are still working.

I guess you didn't make this change? I keep getting lockups changing config through the Lua, and my power values are still all over the place even after Sunjun's calibration. I did find that my meter was on 5650MHz in the first test (/shame), but now on 2400 peak with a 10dBm attenuator.
10mW=23, 25=36, 50=66, 100=150, 250=320, 500=600, 1W=700

[wvarty]

Basic functional testing looks ok to me. Approving this after discussing the lockup issues raised above in comments, and agreeing that these are not specific to Emax targets, so should not hold up the merge.