SGI IP28 Kernels... anyone had any luck lately?

Hi all,
I'm not sure what's causing this error, could very well be PEBKAC, but
anyways.

I've been striking issues getting kernels for my IP28 to compile. So
far, I've tried both the 2.6.14 and 2.6.14-rc2 tags, the IP28 patches[1] apply successfully, but the subsequent compile fails with these
messages: http://pastebin.com/455158

Incidentally, the tarball you download containing the patches,
apparently has a README and .config file included. Well, there's symlinks to the files, but it seems the actual files themselves got missed in the archive. I used the /proc/config.gz from my currently running kernel (2.6.12-rc2, also works with 2.6.13.4).

I was hoping to try out Impact support, in console and X, as well as
HAL2 support (which was b0rked last time I tried it).

Has anyone had any luck, and if so, any ideas what I'm doing wrong?
Regards,

Stuart Longland (aka Redhatter)              .'''.
Gentoo Linux/MIPS Cobalt and Docs Developer  '.'` :
. . . . . . . . . . . . . . . . . . . . . .   .'.'
http://dev.gentoo.org/~redhatter             :.'

1. http://home.alphastar.de/fuerst/download.html

Hi all,

sorry, obviously forgot to "tar" with the "-h" option. The kernel patch-set is now repackaged with README and .config (same location). I wonder, why no one noticed their missing since Oct 17...
For exactness' sake: the patches are based on linux-2.6.14-rc2-mipscvs-20050925 Maybe .config will enable compiling, the error-messages seem to point to a misconfiguration, since the compiler didn't touch any of the patched files yet.

There's still a problem with the Xserver: often, when starting up the Xserver first after a cold boot, it likes to hang (in a loop, waiting for "dmabusy" to settle down, either in the kernel-driver or the Xserver itself, when re-mmapping the dma-buffer). Usually, after a reset the Xserver works okay. I couldn't find a solution for this yet, but otherwise (;-) i easily (can) use the machine for regular work (no more hangs after the Xserver started up once).

kind regards

pf

Hi,

Heh, I wondered that myself... Anyways, it's fixed now. :-)

I did try using the linux-2.6.14-rc2 tag out of git (repository as of 20051204; I didn't bother going to bleeding edge), but found I got the same errors.

The problems disappeared, however, when I tried 2.6.14-rc1... this seems to work just fine.

The only glitch I've hit so far, is that booting via the local framebuffer (okay, I know it isn't a true "framebuffer") causes a fatal oops. However, if I specify console=ttyS0,9600, I see a flashing cursor on the framebuffer, and the bootup proceeds to completion on the serial console. Which is good news... it wasn't long ago when I recall merely sneezing in front of the machine was enough to send it spiralling into a fatal oops.

As I say... this Linux port has really bounded ahead in the time I've had this IP28 of mine... considering the hardware issues that have to be worked around. :-)

X is my next step... for the moment though, I'm happy to be able to use a real monitor, and not a serial console. :-)

Stuart Longland (aka Redhatter)              .'''.
Gentoo Linux/MIPS Cobalt and Docs Developer  '.'` :
. . . . . . . . . . . . . . . . . . . . . .   .'.'
http://dev.gentoo.org/~redhatter             :.'