I am attempting to install Gentoo for PPC64 (using install-ppc64-minimal-2007.0-r1). I have tried all yaboot options that are displayed when pressing tab at the prompt, with the exception of the setting for a serial console, which I am not using. In every instance, I am able to get as far as the kernel loading before I am brought to a grey OpenFirmware screen with the following output: done instantiating rtas at 0x000000000076a40000 ...done 0000000000000000 : boot cpu 0000000000000000 0000000000000002 : starting cpu hw idx 0000000000000002 ...done copying OF device tree ... Building dt strings ... Building dt structure ... Device tree strings 0x0000000002b70000 -> 0x0000000002b7135b Device tree struct 0x0000000002b70000 -> 0x0000000002b860000 Calling quiesce ... returning from prom_init At which point the machine hangs for a very long time, and then reboots. This occurs each time I attempt to install, without fail. Again, I have tried all kernel image options given to me by yaboot, and have even included the "noprobe" and "video=ofonly" options passed to each kernel, with no avail. Reproducible: Always Steps to Reproduce: 1. Boot from CD on IBM Intellistation Power 285 2. Attempt loading any kernel image from cd install-ppc64-minimal-2007.0-r1 3. Crash occurs every time as described. Actual Results: I am brought to a grey OpenFirmware screen with the following output: done instantiating rtas at 0x000000000076a40000 ...done 0000000000000000 : boot cpu 0000000000000000 0000000000000002 : starting cpu hw idx 0000000000000002 ...done copying OF device tree ... Building dt strings ... Building dt structure ... Device tree strings 0x0000000002b70000 -> 0x0000000002b7135b Device tree struct 0x0000000002b70000 -> 0x0000000002b860000 Calling quiesce ... returning from prom_init Expected Results: Yaboot should have loaded the kernel image and installation should have began.
this looks like kernel bug #8372 [1]. what kind of graphics card do you have? could you please connect to the machine using a serial console? [1] http://bugzilla.kernel.org/show_bug.cgi?id=8372
I believe the graphics adapter is the Matrox G400 (or G450). I will attempt to connect via serial and install.
Did you also try the various framebuffer options? Like nofb?
Not yet, no. The only video option I had tried was video=ofonly. I have booted using a serial connection and the serial terminal I was using stopped updating the display after "starting local". I'm attempting to boot using my laptop and minicom as my terminal for this machine to see if that makes a difference. In the event that this is not successful, I will attempt other framebuffer options.
If the last thing you saw is starting local, it's likely a console issue. Does that machine use any special console devices and are you passing the correct console related variables at startup?
The machine uses a standard serial port for the console. The settings for the machine (and for minicom) are 19200 8/n/1. The last message which gets displayed is "Starting local" ... [ok] If I power down the machine the console updates and says that it is unmounting file systems. This is the only updated output I receive on the console after "Starting local"
So are you passing " console=ttyS0,19200" to the kernel image you need to boot? Or perhaps it's ttyS1 ?
ping, any luck?
I decided today to attempt loading the ibmpower-hvsi0 kernel image. I tried several times yesterday with no success, but today it worked and I am extracting the portage snapshot currently. I'm not sure what caused this kernel image to work today, but not on previous attempts. I didn't pass any flags to the kernel at boot, but it was successful. The original issue seems to be with the video card for this machine (Matrox G400 or G450). If the kernel attempts to detect a display, I get the aforementioned crash. Is there a possible work around that would enable display for this unit, or will I have to change video adapters?
can you log onto IRC? #gentoo-powerpc? Look for me and I can explain to you exactly what I think is going on; then have you confirm for me.
Absolutely. I'm there now. alockhart
This seems to be a kernel issue. Please report this bug at http://bugzilla.kernel.org