Re: intel 2700 register values

:-/ That is too bad. I was hoping we could at least speed things up a
bit. I don't think framebuffers use opengl, but when we get Xwindows
going we will definitely want it if we can, it'll speed things up a lot.

I think (hope) that we have all the info we need to write our own
graphics drivers for the chip. I don't know enough about opengl to find
out, though.

the reason it has to mess with the pci/agp stuff is because that's the
bus it uses to talk to the chip on a normal PC... also, it probably
makes no assumptions about whether or not the chip is initialized.

I do have framebuffer working, using the "vsfb.c" framebuffer (Very
Simple Framebuffer), which just maps a framebuffer to video ram, and
does no initializing or anything.

We're working on getting one that can initialize as well, but we won't
really need it until after we can get rid of wince from the device,
which will take awhile.

Richard

>http://www.powervr.com/Downloads/Drivers/2-01-21-0007/powervr-2.01.21-7.tgz
>http://www.pvrdev.com/Pub/MBX/
>
>http://www.khronos.org/opengles/
>http://www.khronos.org/opengles/documentation/gles-1.0c.tgz
>
>ramblings... words spewing out my ears:
>
>I think you're right about all the draw backs about the pci/agp bus, and
>it is definitely not a frame-buffer driver, and uses X's DRM.
>Why is it messing with pci/agp in X at all.... memory map stuff?
>Hope I am wrong about that.
>
>Worse their License is closed, dreadful, and talks about "No Reverse
>Engineering" when it is source code ;-) . This could reduce the code
>to being only useful for looking at. Could looking at the code
>poison a new driver's GPL status?
>
>So what is the association between the 2700G5 and MBX-whatever?
>The Intel 2700G5 incorporates the PowerVR MBX graphic core, and
>PowerVR video accel (from the Internet).
>So what do we do? Most of the APIs are windowblown.
>Maybe a OpenGL frame-buffer driver? (I see there are some, but
>possibly proprietary : Nvidia )
>What about VGA?
>LinuxBios (crossed with HaRET) for initializing the thing?
>Why didn't using a VGA console (vgacon) on the X50v work (I tried...
>but I at that point I had even less of a clue). Was it expecting PCI bus etc?
>I guess the VGA resources were not mapped in, something a bios could do.
>Thats enough for console until X (or something graphic) can start.
>
>I will retry my old experiments with a new cable (travel USB cable with
>Serial added, and if I get ambitious, VGA out just in case things don't
>go to the right place). Another thanks to Fisherss for the pretty
>cable diagram.
>
>As a last resort, get X up as quickly as possible, and soft boot the
>display ( I have a couple of old laptops like that).
>
> J.Currey
>
Received on Wed Jun 29 2005 - 23:51:36 EDT