Corsair RGB memory control or lack thereof

[goatcheese]

I've got Corsair RGB Vengeance Memory & out of curiosity I went searching to see if there was any Linux support for it. Short answer: no. But here's what I discovered on my way to that conclusion. FWIW motherboard is an Asus, not sure how much that matters.

The majority of Corsair projects are controlled via USB i.e. keyboards & AIO's. For those there exists projects like
- ckb / ckb-next https://github.com/ckb-next/ckb-next
- OpenCorsairLink https://github.com/audiohacked/OpenCorsairLink (works well on my H150i)

Lower level, a number of libraries & language bindings exist
- Official Corsair Windows only SDK http://forum.corsair.com/v3/forumdisplay.php?f=271
- Asus Aura SDK https://www.asus.com/campaign/aura/us/SDK.html also windows only but has support for controlling memory.
- RGB.net https://github.com/DarthAffe/RGB.NET

Decompiling the SDK dll or wrapping it would probably work, beyond my skill tho

Hows are the LEDs controlled?
Not much technical information on this, several reviews (e.g. https://www.bit-tech.net/reviews/tech/c ... -review/2/) mention the following
"Its LINK software communicates with the MCU on the DIMMs over the SM bus"

Some troubleshooting information here http://forum.corsair.com/v3/showthread.php?t=168055 mentions enable "SPD write" in your BIOS which also hints at how it works. No such setting on my Asus but it works in Windows without issue regardless.

This thread shows some success for gskill memory https://gitlab.com/CalcProgrammer1/Keyb ... /issues/85 via I2C

On my system i2cdetect on the SMBus comes up with three addresses which are empty so that's a no go.

Code: Select all

# i2cdetect -l
i2c-3	i2c       	NVIDIA i2c adapter 6 at 17:00.0 	I2C adapter
i2c-10	i2c       	NVIDIA i2c adapter 7 at 65:00.0 	I2C adapter
i2c-1	i2c       	NVIDIA i2c adapter 1 at 17:00.0 	I2C adapter
i2c-8	i2c       	NVIDIA i2c adapter 2 at 65:00.0 	I2C adapter
i2c-6	i2c       	NVIDIA i2c adapter 9 at 17:00.0 	I2C adapter
i2c-4	i2c       	NVIDIA i2c adapter 7 at 17:00.0 	I2C adapter
i2c-11	i2c       	NVIDIA i2c adapter 8 at 65:00.0 	I2C adapter
i2c-2	i2c       	NVIDIA i2c adapter 2 at 17:00.0 	I2C adapter
i2c-0	smbus     	SMBus I801 adapter at 5000      	SMBus adapter
i2c-9	i2c       	NVIDIA i2c adapter 6 at 65:00.0 	I2C adapter
i2c-7	i2c       	NVIDIA i2c adapter 1 at 65:00.0 	I2C adapter
i2c-5	i2c       	NVIDIA i2c adapter 8 at 17:00.0 	I2C adapter
i2c-12	i2c       	NVIDIA i2c adapter 9 at 65:00.0 	I2C adapter

# i2cdetect -y 0      
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:          -- -- -- -- -- 08 -- -- -- -- -- -- -- 
10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
40: -- -- -- -- 44 -- -- -- -- -- -- -- -- -- -- -- 
50: -- -- -- -- -- -- 56 -- -- -- -- -- -- -- -- -- 
60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
70: -- -- -- -- -- -- -- --  

No idea what MCU is on the RAM or if its a separate bus or what have you. Guessing it'd need a custom driver to be exposed as I2C anyway.

So yeah, its a fail for now but eventually someone smarter than me should be able to sort it out.

[Ralphred]

Long and short, there is good news and bad news.
Good news, explicit drivers for decorative rgb devices have been added and are to be included in the 4.20 kernel.
Bad news, asus keeps aurasync protocols close to it's chest and won't be supported at this stage.
Good news, the aura sync windows app does have some levels of persistence that hold through reboot, though things like "temperature" and "music" are likely to result in static colours.

As soon as 4.20 drops in the portage tree, I will be tinkering hard, anything that doesn't break things will get posted.

Back in the day I wrote cron jobs that hooked into the kernels cdrom I/O and stopped the draws open during "snacktime" to prevent the insertion of jam covered toast by the ~2 year old "snacker". If I can't make a light flash the right colour then I am just a shadow of my former self , the jury is out pending the release of gentoo-sources-4.20.*.

[goatcheese]

Cool, I'll keep an eye on the 4.20 / 5.0 release cycle.