Clevo (Malibal) p180HMx LED Color Project

[Blitz]

Thought I should make a thread for this. I'll try to keep it short.

The p180HMx has chassis LED's with customizable colors. However, information on how the end user can modify the color does not seem to exist. As a project I've decided to try to derive this information.

Known:
Probably foremost, that the hardware is there. Originally I thought, well maybe they have different color LED boards that they just install, but I've confirmed that the color parts actually have many different color LED's and wiring to them to control them.

Malibal sells as a service at time of purchase color customization. When I go into the AMI bios screen, I see that the name text for my EC is custom for my color.

I've compared the DMI tables of my bios to the stock clevo bios, and didn't notice anything. I was SO SURE it would be there too, like half the entries say "to be filled by OEM", but .. nothing

It's very possible I'm overlooking something obvious. I'm a professional developer, but usually don't get the opportunity to get this close to hardware or BIOS, so I guess you could call me an educated noob

What I'm asking is, for any thoughts as to where this setting may be, or additional things to consider. I'm sure some here have a lot more experience and might help me think outside my limited box on this.

For example, I recently read about NVRAM, which for all I know maybe this setting is stored there somewhere. But I have yet to locate a tool and do any comparisons, and apparently it's a little specific because I see 131072 bytes in the info page for AFUWINGUI, but some sources from older BIOS say there are just 256...

Or for example, when I do make a copy of my current bios with AFUWINGUI, it seems like some of the starting blocks aren't touched. I'm not sure why.

So there could be large holes I'm not aware of.

Any guidance would be greatly appreciated. Thanks to TheWiz for a few pointers earlier!

---

One other thing, if anyone can obtain an actual, saved BIOS from one of these machines (maybe you own it?) it might help a lot if I could get a copy.

[Blitz]

Ok. I got Somaddict's bios from another post and diffed every single module. 3 of them were different.
The first two, which don't have any names, Vol 0, Index 0 and Vol1 Index0. And one called "sbsmi". There are many differences in all 3 between my and his bios's so unfortunately I have to dig deeper.

First order of business is to figure out what each module does.

[Blitz]

Ok, here's something. I downloaded a utility called "RW Everything" from: http://rweverything.phpnet.us/download.html

Using the "Embedded controller" window, I'm able to toggle the "auto-dim" shortcut button on and off. This is at offset 0xDB, bit 4. 0xCB bit 7 controls the fan silencer button state.

[moral hazard]

I would compare 2 different versions of the EC with a hex editor (different colors), hexworkshop can do this automatically for you and it's free.

[Blitz]

I would compare 2 different versions of the EC with a hex editor (different colors), hexworkshop can do this automatically for you and it's free.

I think that's what I did. This is assuming AFUWin* pulls the EC off. These utilities all have an option to program the "main bios", program the EC bios, program the NVRAM, etc.... but the only read option is to "Save the current BIOS to a file" (/O option) which I'm assuming means reading EVERYTHING, but I've been suspicious. Can someone confirm?

Assuming I've done it right, as I said above, only 3 modules differ. But each one differs a lot. If the change is in there, I'll have to look at the actual code I think.

The only 2 other options I know of right now are to check the NVRAM or to try modifying the EC variables (which are accessible via an IO address) but try rebooting between changes. I have a feeling the color only gets set on a reboot, because when i messed up the colors with the "conductive finger" method, they stayed screwed up until I rebooted.

---

It would also be very helpful if anyone know what the 3 modules in the bios do:

Vol 0, Index 0
and Vol1 Index0
"sbsmi".

Since those are the only 3 that differed.

[Prema]

On these SB Clevos Afuwin pulls the same stuff, no matter what options we use, they all come out with an identical MD5.

The thing is, after reboots the dump differs a lot even without making any changes in windows, bios or stuff like XTU.

So the best way is to tell your reseller that you flashed the BIOS and EC from Clevos FTP and that now your logo & lights have changed and you want to revert it back to what you ordered.
Let them send you the (BIOS) EC & compare that to Clevos, that will bring you much closer to the bits they changed.
Maybe someone with an other color does the same thing and there you go!

[Blitz]

I'm resurrecting this project. I got sidetracked a bit trying to install MacOS Lion (which was successful).

In doing the Lion project, I had to rip the DSDT and fix a lot of compile errors in it (I think they use MS's compiler, but I was using Intel's which seems more strict).
To me it feels like this information, or perhaps some undocumented interface to it, may be in this big blob of code somewhere.

Back to the elusive LED color control. To recap, here's what I know:

• This is definitely possible.
  
• Diffing two BIOS from pretty much the same machine eliminated a lot but not quite enough to pin it down.
  
• Malibal has been extremely unhelpful on this. In fact, if I had a genuine BIOS problem that required the original files to fix, they wouldn't even help. Now that the machine is discontinued, they still refuse to release them. Bottom line, I am doing this without their help.
  
• TheWiz suggested it's somewhere in the ACPI table. The DSDT table is the "main" ACPI table, but there are two other tables that I haven't looked at yet (SSDT and SRAT).
  
• After diffing each module, the following 3 had differences:
  Vol 0, Index 0
  Vol1 Index0
  "sbsmi".
  I'm not sure how modules correlate to ACPI tables or AML code.
  
• AMI makes an EC chip, and in their brochure it mentions controlling custom LED's. I have to have another look at the internals to see if I can identify the EC chip, if it's even that easy. If it's AMI's, well I didn't really see any documention for it anywhere, though they mention there is full source code available somewhere (probably only for vendors).
  

And here's a list of the things that would be helpful:

• Any ripped or original bios from anyone with the same machine (any color)... In theory, with enough of these it might just stand out if only a few bits differ between colors, even if there are thousands of differences that are quasi-random elsewhere.
  
• If anyone knows how Alienware does theirs, that could be good confirmation for a normal way. I know they have software to control it, but at a lower level, is it somewhere in the DSDT table/AML code?
  
• How does the BIOS and the modules that make it up correspond to the DSDT table, is the DSDT table contained inside it or is it seperate?
  
• What are those 3 modules that differed?
  Vol 0, Index 0
  Vol1 Index0
  "sbsmi".
  
• If I rip my BIOS and then flash the (unchanged) rip, assuming no act of God or power surge or whatever, is it accurate to say that that is pretty safe?
  
• If I rip my BIOS and change a few things, and flash it, and my machine decides to not start up, is there a procedure where I can flash the original back?
  

[Blitz]

Ok next installment in the saga

The touch buttons module contains an IT8301E chip which is a GPIO, apparently just a way to send the button presses to the motherboard and possibly the LED states although those may be controlled in the EC.

The motherboard contains an ITE it8519E chip -- the embedded controller.

I'm looking for a datasheet for the 8519. I found one for a similar chip, an 8512.

For kicks I swapped 2 of the leads in one of the LED multi-color connectors and made my lid design blue. Lol. Those connectors won't take much abuse, though.