Skip to contentSkip to conversation

precompiled 2.2.10 kernel available

PATCH New Paul Mackerras 8 participants 12 messages Jun 18, 1999 – Jun 25, 1999

lore
Share & tools
lore
#0 lore raw

I have put a precompiled 2.2.10 powermac kernel on dev.linuxppc.org and on samba.anu.edu.au (accessible only from within Australia). This kernel works on the iMac and should work on the blue&white G3 powermac, although I haven't been able to test it directly myself.

The source for this kernel is available as a .tar.gz file on the two ftp servers or by anonymous rsync from samba.anu.edu.au::linux-pmac-stable. (My current 2.3.6 tree is available by anon rsync at samba.anu.edu.au::linux-pmac-devel.)

*NOTE:* This kernel includes Linus' USB stack, which (in my version) translates the USB keycodes to ADB keycodes. If you have an iMac, or a B&W G3 with a USB keyboard, comment out (or change) the KEYTABLE definition in /etc/sysconfig/keyboard so that your init scripts don't load a USB keymap.

The relevant files are in ftp://dev.linuxppc.org/users/paulus or ftp://samba.anu.edu.au/pub/linux-pmac/. They are:

In kernel-binary/v2.2.10:
vmlinux kernel binary (ELF)
vmlinux.gz gzipped version of vmlinux
System.map symbol table for vmlinux
modules-2.2.10.tar.gz kernel modules, unpack in /
dot.config .config file used to build vmlinux

kernel-source/v2.2.10/linux-pmac-2.2.10.tar.gz
kernel source tree, tarred & gzipped

RPMS/pcmcia-cs-3.0.12-2210.ppc.rpm
PCMCIA kernel modules & user programs (compiled against glibc-1.99, i.e. Linux/PPC R4)

SRPMS/pcmcia-cs-3.0.12-2210.src.rpm
Source for PCMCIA support

Enjoy!

Paul.

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#1 lore raw

did the protocol of /dev/usbmouse have to change? :( now it appears as a ps2 mouse not a busmouse. x can't understand it either way but i'm able to use my usb mouse in x by having gpm use it as a ps2 mouse and export it, as in

gpm -t ps2 -R

and then have X use gpm

Section "Pointer"
Protocol "mousesystems"
Device "/dev/gpmdata"
EndSection

is there an easier way?

Quoted 1 line

*NOTE:* This kernel includes Linus' USB stack, which (in my version)

Brad
[email protected] | http://www.pht.com/~brad/

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#2 lore raw
Quoted 4 lines · Brad Midgley

did the protocol of /dev/usbmouse have to change? :( now it appears as a ps2 mouse not a busmouse. x can't understand it either way but i'm able to use my usb mouse in x by having gpm use it as a ps2 mouse and export it, as in

Linus' USB stack differs from Inkays in a few ways. This is one. It seems Busmouse is too limited or something along those lines, so a
modified PS/2 is "better". The goal eventually is for /dev/mouse to be a
real device, where all mouse data ends up (ala windows and macos w/ multi mice) and /dev/mouseX for mouseX. At least thats how it was at one point.

---
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#3 lore raw
In reply to Brad Midgley
Quoted 1 line · Brad Midgley

did the protocol of /dev/usbmouse have to change? :( now it appears as a

It was Linus' decision. Argue with him. :-)

Quoted 1 line

ps2 mouse not a busmouse. x can't understand it either way but i'm able to

Use protocol "IMPS/2" in your XF86Config.

Paul.

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#4 lore raw

[workaround] console-switch screws colors

hi,

i've noticed (at least) with a couple of machines that use ati graphics that switching away from and back to X screws up the colors (using xfb). fortunately, i left it that way long enough for the screensaver to kick in and realize it was fixed by the screensaver!

so if you switch back to x and the colors are screwed, invoke the screensaver:

xset s activate

hopefully this is one of those things that will be fixed soon anyway.

Brad
[email protected] | http://www.pht.com/~brad/

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#5 lore raw

Re: [workaround] console-switch screws colors

Hey, I've got that problem. At least there's finally a (sort of) workaround...

Quoted 11 lines · Brad Midgley

hi,

i've noticed (at least) with a couple of machines that use ati graphics that switching away from and back to X screws up the colors (using xfb). fortunately, i left it that way long enough for the screensaver to kick in and realize it was fixed by the screensaver!

so if you switch back to x and the colors are screwed, invoke the screensaver:

Quoted 10 lines

xset s activate

hopefully this is one of those things that will be fixed soon anyway.

Brad
[email protected] | http://www.pht.com/~brad/

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#6 lore raw

PATCH 1/1 : Re: [workaround] console-switch screws colors

In reply to Brad Midgley

Brad Midgley (Tue, Jun 22, 1999 at 11:11:51AM -0600):

Quoted 11 lines

i've noticed (at least) with a couple of machines that use ati graphics that switching away from and back to X screws up the colors (using xfb). fortunately, i left it that way long enough for the screensaver to kick in and realize it was fixed by the screensaver!

so if you switch back to x and the colors are screwed, invoke the screensaver:

xset s activate

hopefully this is one of those things that will be fixed soon anyway.

This may help. I don't know if Tom built new rpms yet.

home/atong/fbdev.c renamed from fbdev.c +5 -0
@@ -1261,6 +1261,11 @@
 	}
 
 	if (!xf86Resetting) {
+#ifdef DIRECTCOLORHACK
+	    if (fb_fix.visual == FB_VISUAL_DIRECTCOLOR)
+		fbdevSetLinearColormap(&fb_var);
+	    else
+#endif
 	    /* Update the colormap */
 	    fbdevRestoreColors(pScreen);
 	}

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#7 lore raw
In reply to Paul Mackerras

if the keyboard and mouse are going to be this way for good, i will patch xconfigurator so it does the right thing.

is there a way to get the middle- and right- mouse buttons via the imac keyboard or is that code missing from the linus-usb?

Quoted 4 lines
Quoted 1 line

did the protocol of /dev/usbmouse have to change? :( now it appears as a

It was Linus' decision. Argue with him. :-)
Use protocol "IMPS/2" in your XF86Config.
Paul.

Brad
[email protected] | http://www.pht.com/~brad/

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#8 lore raw

We can use xkb to handle mouse clicks, the mousekeys option already handles nouse movement/clicks, it might be a little hard to get used to but it will do as a temporary solution. Now that i am thinking of it we can write something similar (that will emulate only mouse buttons) and stop worrying about kernel support.

Kostas Gewrgiou.

Quoted 7 lines · Brad Midgley

if the keyboard and mouse are going to be this way for good, i will patch xconfigurator so it does the right thing.

is there a way to get the middle- and right- mouse buttons via the imac keyboard or is that code missing from the linus-usb?

Quoted 11 lines
Quoted 4 lines
Quoted 1 line

did the protocol of /dev/usbmouse have to change? :( now it appears as a

It was Linus' decision. Argue with him. :-)
Use protocol "IMPS/2" in your XF86Config.
Paul.

Brad
[email protected] | http://www.pht.com/~brad/

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#9 lore raw
In reply to Brad Midgley
Quoted 2 lines · Brad Midgley

if the keyboard and mouse are going to be this way for good, i will patch xconfigurator so it does the right thing.

They are, so that would be good.

Quoted 2 lines

is there a way to get the middle- and right- mouse buttons via the imac keyboard or is that code missing from the linus-usb?

Not at the moment, but we could add it with a #ifdef CONFIG_MAC_KEYBOARD.

Paul.

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#10 lore raw

Machine Check in Kernel Mode with Paul's 2.2.10 on B+W G3

In reply to Paul Mackerras

Hi,

Well I just received my B+W G3 yesterday and Paul's 2.2.10 kernel annoucement seemed like fate so I tried to boot with it to no avail:

Here are some of the interesting things the error report had to say:

System: Brand new B+W G3 with 12 gig ATI drive and Adatec AIC-7850 scsi host adapter found at PCI 2/0 with a 3 gig scsi Quantum drive with R5/YDL on it (which will boot fine on my PowerTower 180e).

Partion check:
sda: sda1 sda2 sda3 sda4 sda5 sda6 sda7 sda8 (sda5 is the root partition)
sdb: scsidisk I/O error: dev 08:10 sector 0
unable to read partition table (this is my zip drive without a disk in it)
hda: hda timeout waiting for DMA

Machine check in kernel mode
Caused by (from msr): regs c01a4c18 Unknown values in msr

Here are some register values:
NIP: C00BDE58 LR: C00BDE14 MSR: 00009030 Trap: 0200
Task= c01a2f58[0] 'swapper' mm->pgd
Last syscall: 112

Backtrace: C00BDE14 COO13C04 C001C850 C0003B68 C001F460 C00069F4 ...
Quoted 1 line

From looking at the booting log messages, everything seems fine up to the

very end (it finds the scsi drive, finds the partitions, finds the USB hub, USB mouse, etc). There does not seem to be anything unusual until the DMA timeout message waiting for hda.

Any ideas here? I used Paul's System.map to try and track down the offending routines but none of the backtrace addresses line up exactly with any of the map entry points (am I missing an offset here?)

Any ideas here would be greatly appreciated. My entire JDK build tree for glibc 2.1 is on the hard drive I am trying to boot from but my wife now has my old machine (PT 180e) and I can't go begging to borrow that machine back, so please any help in getting Linux to boot with this machine would be greatly appreciated.

Thanks,

Kevin

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]

#11 lore raw

Re: Machine Check in Kernel Mode with Paul's 2.2.10 on B+W G3

Quoted 26 lines · Kevin B. Hendricks

Well I just received my B+W G3 yesterday and Paul's 2.2.10 kernel annoucement seemed like fate so I tried to boot with it to no avail:

Here are some of the interesting things the error report had to say:

System: Brand new B+W G3 with 12 gig ATI drive and Adatec AIC-7850 scsi host adapter found at PCI 2/0 with a 3 gig scsi Quantum drive with R5/YDL on it (which will boot fine on my PowerTower 180e).

Partion check:
sda: sda1 sda2 sda3 sda4 sda5 sda6 sda7 sda8 (sda5 is the root partition)
sdb: scsidisk I/O error: dev 08:10 sector 0
unable to read partition table (this is my zip drive without a disk in it)
hda: hda timeout waiting for DMA

Machine check in kernel mode
Caused by (from msr): regs c01a4c18 Unknown values in msr

Here are some register values:

   NIP: C00BDE58  LR: C00BDE14  MSR: 00009030  Trap: 0200

Task= c01a2f58[0] 'swapper' mm->pgd
Last syscall: 112

Backtrace: C00BDE14 COO13C04 C001C850 C0003B68 C001F460 C00069F4 ...
Quoted 8 lines
Quoted 1 line

From looking at the booting log messages, everything seems fine up to the

very end (it finds the scsi drive, finds the partitions, finds the USB hub, USB mouse, etc). There does not seem to be anything unusual until the DMA timeout message waiting for hda.

Any ideas here? I used Paul's System.map to try and track down the offending routines but none of the backtrace addresses line up exactly with any of the map entry points (am I missing an offset here?)

It's normal that the addresses don't line up exactly: the addresses in System.map are the labels associated to routines, while chances are high your kernel crashes inside a routine. So you have to use the closest address that's smaller or equal than the addresses indicated by NIP and the backtrace.

Greetings,

Geert

--
Geert Uytterhoeven [email protected]
Wavelets, Linux/{m68k~Amiga,PPC~CHRP} http://www.cs.kuleuven.ac.be/~geert/
Department of Computer Science -- Katholieke Universiteit Leuven -- Belgium

[[ This message was sent via the linuxppc-dev mailing list. Replies are ]] [[ not forced back to the list, so be sure to Cc linuxppc-dev if your ]] [[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting. ]]