Re: hwclock on a PowerMac 9500 and summertime
On Tue, Apr 01, 2003 at 08:25:39PM +0200, David Ulrich wrote: > > I put it on summertime and add a ntp server (ntp, ntp-simple and > ntpupdate). So the clock is always on time. > My debian is always on time now: > > dulrich# date > Tue Apr 1 20:14:04 CEST 2003 > > But I can fix my hardware clock on summertime: > > dulrich# hwclock --show > Tue Apr 1 19:15:32 2003 -0.984699 seconds > > I try: > hwclock --systohc, hwclock stop/start, hwclock force-reload > The hardware clock get the summertime: > > hwclock --systohc > daisy:/home/dulrich# hwclock --show > Tue Apr 1 20:15:21 2003 -0.454536 seconds > > but some minutes later: > dulrich# hwclock --show > Tue Apr 1 19:18:09 2003 -0.510788 seconds IMHO, the best practice is to keep hw clock coordinated to UTC(Greenwich), and use the right timezone..that way you never need to worry about summer time shift 8) -- Boris
Re: Install problem, no boot after all
On Sun, Apr 06, 2003 at 12:44:01PM +0200, Tamas Nagy wrote: > > It seems that the 'quik' bootloader doesn't work from Debian installer > with PowerMac 4400/200. The other boot possibility would be a 'boot > floppy' which after selection informs me that 'boot floppy' is not valid > option... You can (probably) boot from hfs formatted floppy. Just copy vmlinux.coff to floppy and when booting enter into OpenFirmware: 0> boot fd:vmlinux.coff kernel_arguments... Note that with recent 2.4 kernels there are some problems with COFF images.. > Is there any way to boot the installed system from MacOS or from a > custom made boot-floppy? I use quik with great success..although i must admit it didn't work out of the box after install. I have a PowerComputing box which is a bit different than yours i guess..but not much. My boot config: Mac partition type. /boot is on /dev/sda2(ext2 fs, 12MB). All files necessarry for boot are placed inside: System.map-2.2.22 System.map-2.4.21-pre7 boot first.b quik.conf tt System.map-2.4.19 blast lost+found second.b zorg /boot/quik.conf: # quik -f -v -r /boot -C /quik.conf # ^^ this is how i run quik..only need to run it once # after adding new kernel image to /boot you anly need to modify # quik.conf for the changes to take effect # timeout = 50 default = zorg partition = 2 append = "video=imsttfb:vmode:12 devfs=nomount" # 2.2.22 image = /tt root = /dev/sda3 label = tt read-only # 2.4.19 image = /blast root = /dev/sda3 label = blast read-only # 2.4.21-pre7 - testing swim3 image = /zorg root = /dev/sda3 label = zorg read-only NVRAM parameters: little-endian? false real-mode? false auto-boot? true diag-switch?false fcode-debug?false oem-banner? false oem-logo? false use-nvramrc?true real-base 0x7f0 real-size 0x10 virt-base 0x virt-size 0x10 load-base 0x4000 pci-probe-list 0x screen-#columns 0x64 screen-#rows0x28 selftest-#megs 0x0 boot-device scsi-int/[EMAIL PROTECTED]:0 boot-file diag-device fd:diags diag-file input-devicekbd output-device /[EMAIL PROTECTED]/IMS,tt128mb8 oem-banner oem-logo nvramrc : slowboot begin ." Waiting for boot device.." cr d# 1000 ms boot-device ['] $boot catch drop key? until ; boot-commandslowboot Note that your settings _will_ be different since u have different hardware. -- Boris
Re: RS6000 model 44p 170
On Tue, Apr 08, 2003 at 04:51:35PM +0100, Michael Cameron wrote: > Does anyone have a reference for Open Firmware commands? Try this: http://firmworks.com/QuickRef.html or http://www.netneurotic.de/mac/openfirmware.html > FAT: bogus logical sector size 15370 > FAT: bogus logical sector size 15370 > attempt to access beyond end of device > 08:03: rw=0, want=33, limit=12 > isofs_read_super: bread failed, dev=08:03, iso_blknum=16, block=32 > attempt to access beyond end of device > 08:03: rw=0, want=65, limit=12 > readsuper_block: bread failed (dev 08:03, block 32, size 1024) > readsuper_block: can't find a reiserfs filesystem on (dev 08:03, block 8, > size 1024) > attempt to access beyond end of device > 08:03: rw=0, want=36, limit=12 > attempt to access beyond end of device > 08:03: rw=0, want=36, limit=12 > Kernel Panic: VFS: unable to mount root fs on 08:03 > <0> Rebooting in 180 seconds Enter the right boot parameter for root device..or you are missing the driver for the root device. -- Boris
Re: nvsetenv
On Tue, Apr 08, 2003 at 05:14:36PM +0200, David Campillos wrote: > boot-device scsi/[EMAIL PROTECTED]:0 > boot-file /boot/vmlinux-2.2.19-pmac // I think it is the correct > name > diag-service fd:diags > diag-file > input-device kbd > output-device /[EMAIL PROTECTED]/ATY,[EMAIL PROTECTED] > oem-banner > oem-logo > nvramrc > boot-command boot > > The problem is that with this parameters the computer does not start. > The screen does not output any boot message, and a display says 'no > signal input or cable disconnected'. > > Where is the problem? Any idea? With a bit of bluff i would say the kernel doesn't init your mach64 board..have you tried some boot parameters like video=atyfb ? check the documentation for correct parameters as i never had that board myself.. Also..after cca. 180 sec after the OF loads the kernel, does the box resets ? Also try with serial console as system console, if u have some other box with serial port..might help to see where the fun stops. -- Boris
Re: PowerMac 9500 and sound
On Sat, Apr 26, 2003 at 07:18:11PM +0200, David Ulrich wrote: > > I'm on a PowerMac 9500 with 2.4.18. When I launch KDE I have a > crash of sound system. I have 9500 clone, but i don't use KDE. Is your sound working in console? what are the messages from dmasound_pmac module? -- Boris
Re: mac serial port settings in linux
On Mon, May 19, 2003 at 10:56:47AM -0700, dylan wrote: > > this might be a really dumb question... but: is there any easy way to set my > serial ports to a given speed, parity, ect - and have those changes stay? Take a look inside /etc/init.d/setserial and /var/lib/setserial/* -- Boris
Re: Installation problems on StarMax
On Mon, Jun 02, 2003 at 07:38:39AM -0700, Chris Tillman wrote: > > My first suspicion is that the kernel doesn't fit in the memory > space provided by quik. Do you still have MacOS on the machine? > If so, you can try it with BootX to see if it gives the same result. I have oldworld CHRP here booting 2.2/2.4/2.5 with quik just fine. Boot from OF is also possible with a little patch for coff loader (.data load address IIRC). No idea how much prep is different there.. -- Boris
Re: Quik and framebuffer problems on StarMax
On Thu, Jun 05, 2003 at 07:02:47PM -0700, Thanasis Kinias wrote: > > If I boot without any video= argument, _sometimes_ it will boot cleanly > with the atyfb driver running the ATI card and the imsttfb running the > TT128. If I explicitly use video=atyfb, however, I get a kernel oops as > soon as it loads the atyfb driver. I've tried different (and no) > vmode,cmode settings and it still oopses. I also have IMS TT 128 with 8MB ram..it may be similar to yours. With stock kernel i have problems initializing it(only works with video=ofonly..kind of). I patched the driver to force the detection of TI TVP3030 ramdac instead of IBM ramdac(seems like they use the same PCI ID and different ramdacs). I use append="video=imsttfb:vmode:17:cmode:8" with quik and it works well in console mode. In X i have some problems with wrapping, otherwise it works. See these pictures if you have the same problems (OpenFirmware/console/X): http://knjiznica-jesenice.org/~boris/boot/ No idea how to fix the wrap problem in X..still working on that. -- Boris
Re: Quik and framebuffer problems on StarMax
On Fri, Jun 06, 2003 at 03:54:01PM -0700, Thanasis Kinias wrote:
> scripsit Boris Bezlaj:
>
> > I also have IMS TT 128 with 8MB ram..it may be similar to yours. With
> > stock kernel i have problems initializing it(only works with
> > video=ofonly..kind of). I patched the driver to force the detection of
> > TI TVP3030 ramdac instead of IBM ramdac(seems like they use the same
> > PCI ID and different ramdacs).
>
> Can you share your patch? I noticed that when I got it to boot without
> video=ofonly, the TT128 was identified as `(IBM)', which struck me as
> odd.
It means IBM ramdac..the card is still IMS TT..weather with IBM or TVP
ramdac..if you have TVP try patch below, if not please send me as much info of
your car as possible.
..against 2.4.20(afaict); just a dirty patch but it works for me:
--- ORIG_imsttfb.c 2003-03-10 21:25:35.0 +0100
+++ imsttfb.c 2003-06-08 00:18:39.0 +0200
@@ -1934,7 +1934,7 @@
switch (pdev->device) {
case PCI_DEVICE_ID_IMS_TT128: /* IMS,tt128mbA */
- p->ramdac = IBM;
+ p->ramdac = TVP;
break;
case PCI_DEVICE_ID_IMS_TT3D: /* IMS,tt3d */
p->ramdac = TVP;
@@ -1942,6 +1942,8 @@
default:
printk(KERN_INFO "imsttfb: Device 0x%lx unknown, "
"contact maintainer.\n", pdev->device);
+ free_mem_region(addr, size);
+ kfree(p);
return -ENODEV;
}
> > I use append="video=imsttfb:vmode:17:cmode:8" with quik and it works
> > well in console mode. In X i have some problems with wrapping,
> > otherwise it works.
>
> That's with the patched kernel, right?
Yes, without it "detects" IBM ramdac according to PCI ID's, but i have
TVP 3030 (Texas Instruments) ramdac on the bord..looks like someone
didn't bother to change the PCI ID of the card..
>
> > See these pictures if you have the same problems
> > (OpenFirmware/console/X): http://knjiznica-jesenice.org/~boris/boot/
>
> Yep, that's it.
>
> > No idea how to fix the wrap problem in X..still working on that.
>
> Is the wrap problem using the imstt driver or a generic?
I only have IMS in here..
--
Boris
Re: Specific Documentation for PowerPC
On Wed, Jun 11, 2003 at 10:05:59AM +, Daniel Bolgheroni wrote: > In i386 platform, I used to configure XFree86 with "xf86config", but I didn't > found this script in powerpc platform. try /usr/X11R6/bin/xf86config it's probably not in the search path or something -- Boris
Re: Boot-floppies for PowerBook G3 Wallstreet
On Tue, Jun 17, 2003 at 09:03:48AM -0700, Henry House wrote: > > No error or messages. It appears that nothing was written to nvram. (See my > other message.) afaict, installer does not modify nvram settings. you have to do that yourself -- Boris
Re: some basic problems with ppc assembler
Try these links: http://www-106.ibm.com/developerworks/linux/library/l-ppc/ http://developer.apple.com/techpubs/macosx/DeveloperTools/Assembler/AssemblerTOC.html http://www.lightsoft.co.uk/Fantasm/Beginners/begin1.html http://www.mactech.com/articles/develop/issue_21/21balance.html http://penguinppc.org/~hollis/ -- Boris
Re: strange entry in /var/log/kern.log
On Wed, Jul 30, 2003 at 10:16:13AM -0800, dylan wrote: > [EMAIL PROTECTED]:/var/log$ cat kern.log.0 > Jul 23 14:44:40 cmn kernel: TCP: Treason uncloaked! Peer > 140.211.93.143:1198/80 shrinks window 4054999251:4055005711. Repaired. I get these all the time whenever someone prints through print server box. It tells the other end modifies some TCP parameters..in other words it's TCP layer is f*%&[EMAIL PROTECTED] -- Boris
Re: man -k not working
On Thu, Aug 07, 2003 at 09:30:21AM +0100, Tim Froggatt wrote: > displayed fine, but I can't seem to search through them for keywords. It > always prints "nothing appropriate". Try 'mandb'. > I've tried doing an "updatedb" but that hasn't helped either. Does anyone know updatedb is for 'locate' utility..see man page. AFAIK, it has nothing to do with man or apropos. -- Boris
Re: IMS TwinTurbo - wrong colors
On Fri, Aug 22, 2003 at 04:32:29PM +0200, fab wrote: > I am using Debian/Woody on a PowerMac 9600/350 with an IMS TwinTurbo > graphics adapter and have the following problem when > using XFree86 with the imstt- or fbdev-Driver : > > - the colors are only correct when in 8-Bit-Mode > - 16 bit depth yields in "psychedelic" color shifts > - when using 24 bit depth, red shows up as blue, and everything > that is meant to be blue appears as brown > > This problem occurs with kernel image 2.2.20 and with xserver version > 4.1.0.1 I have 9500/225 mac clone with IMS TT..working fine in 8 and 24 bit mode..not 16bit(same problem that you have). I got it working under fbdev and X almost fine. Using latest 2.4..also works in 2.6 but only under console..not X(working on that). Could you send me the chipset specs? I will get back in about 6 hours from now. -- Boris
Re: IMS TwinTurbo - wrong colors
On Fri, Aug 22, 2003 at 06:37:58PM +0200, fab wrote: > > here is what Open Firmware says: IMS,tt128mb8A > > an this is what lspci says: Disply Controller: Integrated Micro > Solutions Inc. IMS9129[Twin turbo 128] (rev01) lspci on my box: 01:0f.0 Class 0380: 10e0:9128 (rev 01) Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- Status: Cap- 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- SERR- second question: is it right that under the 2.2.20 kernel on a powermac > 9600/350 only the first 3 PCI-Slots are visible for Debian and that a > update to 2.4.. kernel will solve the problem ? > lspci - shows scans only the first 3 not all 6. lspci under 2.2.x does _not_ show all..under 2.4 it does(just tested..again :) btw, if you want you can look at some photos i made, showing the problems i still have. You can find them at: http://www.alpinizem.net/~boris/imsttfb/ I didn't take photos of 16bit mode..i belive the problem there is wrong color pallette or something similar. -- Boris
