[patch] PCI class test for VGA

[Ronald Hoogenboom]

Hi,

In pci_device.c, the class for VGA was not tested properly, leading to
no VGA output from coreboot, even after the boot-rom was executed
properly (CONFIG_PCI_ROM_RUN) or no boot-rom execution with
CONFIG_VGA_ROM_RUN at all. According to the header file device.h, the
class field of struct device is '3 bytes: (base,sub,prog-if)'.

Signed-off-by: Ronald Hoogenboom

[ron minnich]

Wow. Did this code ever work? and, if so, why ...

Can someone else test this?

Thanks Ronald

ron

[Torsten Duwe]

On Thursday 28 February 2008, ron minnich wrote:
> Wow. Did this code ever work? and, if so, why ...

Remember? VGA console never worked without PCI_ROM_RUN until Luc's proposal. 
Consequently, the first test got never even compiled.

If the second test fails only coreboot's own vga_tx_byte will not work. Direct 
VGA access especially from the payload is unharmed.
> Can someone else test this?

Ronald's patch is perfectly sane, so I committed it modulo the braces. My VGA 
still does not work, but if it breaks someone's console (which I strongly 
doubt) I'll back it out again.

Torsten

[ron minnich]

On Thu, Feb 28, 2008 at 3:18 PM, Torsten Duwe  wrote:
> On Thursday 28 February 2008, ron minnich wrote:
> > Wow. Did this code ever work? and, if so, why ...
>
> Remember? VGA console never worked without PCI_ROM_RUN until Luc's proposal.
>  Consequently, the first test got never even compiled.

I drink to forget

ron

[Torsten Duwe]

On Wednesday 27 February 2008, Ronald Hoogenboom wrote:
> In pci_device.c, the class for VGA was not tested properly, leading to
> no VGA output from coreboot, even after the boot-rom was executed
> properly (CONFIG_PCI_ROM_RUN) or no boot-rom execution with
> CONFIG_VGA_ROM_RUN at all. According to the header file device.h, the
> class field of struct device is '3 bytes: (base,sub,prog-if)'.
>
> Signed-off-by: Ronald Hoogenboom

Ack. r3119.