Josh Bonczkowski [EMAIL PROTECTED] writes:
I downloaded the RC1 ISOs over the weekend and attempted an install
last night. I'm very glad I started to read the old messages of
this list yesterday through the archives as it helped me get past
a few problems.
First, the supplied kernel images allowed me to boot mostly properly.
Because this is an older SCSI machine, I had to load two modules by hand.
One was for the SCSI controlled itself (I don't remember the name). The
other was the MESH module. Once both of these modules were loaded
the SCSI CD was found and the installation would begin.
Same here, although I only needed to load the MESH module. I've got a
PowerTower Pro 200 (604e-based clone with the same guts as the Apple
9600, IIRC).
I used the RC1 CD's, the supplied 2.4 kernel and ramdisk image all.gz.
Now we get to my problems. The display during the install is awful
in one of two ways depending on if I check the 'Force video' option on
BootX. I either get a nice looking display but the ADB mouse responds
extremely slowly or I get a squished display with odd colors, the images
wrap strangely on the screen, but the mouse functions fine. When the
mouse is unresponsive, the keyboard also responds poorly (very, very
slow).
I saw this too.
I'm not sure, but I think that X drawing performance is also horrible.
It takes a couple/few minutes for the Mandrake install screen to show
up with the language selection step. I'm installing off of a Yamaha
8x4x24 CDRW, but it doesn't look like that could be the bottleneck.
FWIW, I have an IMS TwinTurbo 4M video card. I've always had issues
with the video driver. I usually run in 8-bit 1280x1024 on my Sony
20 Multiscan with the following kernel args:
video=imsttb:vmode:20,cmode:8 -mach64
I've run both with and without the No video driver checked, and with
and without the above kernel args.
I get the same behavior with the on board (Chaos) controller as well as
with a Matrox Millenium II PCI card.
Well, I'm glad it isn't just *my* video card, since there aren't that
many of them around it seems, and it doesn't seem likely that I'll get
the fixes I need if the problem only occurs with them.
I also noticed several messages on the first console. The same message
kept scrolling by which basically said the default locale was set
incorrectly in a gtk module (perl I think?) and it could not find
the correct information. I doubt that has anything to do with the
problems I have but it could be something to clean up later.
I get this too. Here's what it says, many times over on VT #1:
...Gdk WARNING **: locale not supported by C library at
/usr/bin/perl-install/my_gtk.pm line 279
...Gdk WARNING **: locale not supported by Xlib, locale set to C at
/usr/bin/perl-install/my_gtk.pm line 279
...Gdk WARNING **: can not set locale modifiers at /usr/bin/perl-install/my_gtk.pm
line 279
Also, on VT #3, I see the following messages:
* warning rm of /usr/share/locale failed: No such file or directory
* getFile VERSION:
* Trying with server Xpmac
* starting step 'selectLanguage'
* getFile Mandrake/mdkinst/usr/bin/perl-install/po/en_US.po:
* errorOpeningFile Mandrake/mdkinst/usr/bin/perl-install/po/en_US.po
* getFile Mandrake/mdkinst/usr/bin/perl-install/po/en.po:
* errorOpeningFile Mandrake/mdkinst/usr/bin/perl-install/po/en.po
Sure enough, those two files are missing on the RC1 install CD.
And finally, on VT #4, I noticed the following kernel oops:
[...]
6 Partition check:
6 sda: [mac] sda1 sda3 sda4 sda5 sda6 sda7 sda8 sda9 sda10
Oops: Exception in kernel mode, sig: 4
4 NIP: C01C3980 XER: 2100 LR: C027E7A4 SP: C8C53C60 REGS: C8C53BB0 TRAP: 0700
4 MSR: 00089032 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11
4 TASK = C8C52000 [20] 'insmod_' Last syscall: 128
[...]
I doubt this is related to the horribly slow X performance, but it
looks wrong, particularly since it never seems to get to checking
partitions the other disks (I've got 2 other hard drives at sdb and
sdc).
Admittedly, I initialized the Mac partition map for this physical
drive using the recipe I saw somewhere (dd the 1st 16 blocks from
the raw device /dev/sda and use emacs/hexl to hex-edit the info
into place). pdisk seemed to think the partition map was ok enough
to work on, so I assumed everything was fine. Was that too hasty a
conclusion?
On the other hand, aren't I suppose to get a shell on VT #2?
I don't see one. Perhaps this is the reason?
-
Smooth Jazz is no jazz at all...
Dr. Robert M. Fuhrer [EMAIL PROTECTED]
IBM T.J. Watson Research Center
-