Possible fix for ALL non booting linux kernels

[Guest]

Could it have something to do with the linux kernel trying to initialize the programmable interrupt controller? I seem to recall that when I compared the screens of Yellow Dog linux booting on a real mac vs. pearpc that the APIC initialization is where it stopped at in PearPC. But I'm no programmer...

[dfoesch]

[quote="Anonymous":3380e1ee26]
Could it have something to do with the linux kernel trying to initialize the programmable interrupt controller? I seem to recall that when I compared the screens of Yellow Dog linux booting on a real mac vs. pearpc that the APIC initialization is where it stopped at in PearPC. But I'm no programmer...
[/quote:3380e1ee26]

Usually the problem has been decompressing the kernel overwrites the yaboot executable (the kernel being much larger at 2.6.*) Loading the yaboot executable into a higher address has pretty much fixed that problem.

[Guest]

And how do you do that?

[dfoesch]

[quote="Anonymous":04ed6ab8ad]
And how do you do that?
[/quote:04ed6ab8ad]

Open up src/io/prom/promboot.cc

Find the line that defines LOAD_ADDRESS, and increase it. Then recompile.

NOTE: Linux 2.6 is still not working though, as its framebuffer is not working, and it also crashes at some point during the cdrom initialization.

[Guru3]

So you're saying that there's something in the frambuffer code that kills it as well?

[pearpctweaker]

[quote="dfoesch":770c421440]
[quote="Anonymous":770c421440]And how do you do that?
[/quote:770c421440]

Open up src/io/prom/promboot.cc

Find the line that defines LOAD_ADDRESS, and increase it. Then recompile.

NOTE: Linux 2.6 is still not working though, as its framebuffer is not working, and it also crashes at some point during the cdrom initialization.
[/quote:770c421440]

Have you tried getting an already installed hard disk image for a linux system and see what happens? As in, one that's not loading off of CD as far as the OS sees....

[dfoesch]

[quote="Guru3":6715aa276f]
So you're saying that there's something in the frambuffer code that kills it as well?
[/quote:6715aa276f]

Not in our framebuffer code necessarily, and it doesn't kill the bootup. The CDROM error is what kills the boot.

For some reason the framebuffer doesn't work, and it's unclear if this is an error in the linux code (highly unlikely) or a lack of compatibility on our end (much more likely)

Basically, there's a lot broken besides just loading the Linux 2.6 kernel into memory.