[OLPC-devel] Re: New buildrom is ready

[Jordan Crouse]

On 14/08/06 11:16 +0100, David Woodhouse wrote:
> On Sun, 2006-08-13 at 19:32 -0600, Ronald G Minnich wrote:
> > Or if we should just do this in linuxbios, and use the linuxbios i2c code? 
> 
> That makes sense -- it's only a single register write to enable the
> DCON.
> 
> However, we also want the framebuffer device in the Linux kernel to know
> that it's there, so that it can pick the special 1200x900 at 50 mode
> instead of a normal CRT mode. Can we just indicate its existence in the
> OpenFirmware device tree? Oh, wait.... :)

I2C + NAND in the kernel will add ~60k, FYI.

But I just had a thought - Why not probe the DCON in LinuxBIOS (using the
existing smbus code), and set a byte somewhere in the PCI config space if
it exists?  That way, we could communicate the existance of the DCON to the
kernel without going out of our way to do anything stupid with virtual 
registers.

Jordan

-- 
Jordan Crouse
Senior Linux Engineer
Advanced Micro Devices, Inc.
<www.amd.com/embeddedprocessors>