Re: E250 crashing hard with 2.4.16 and 2.4.17

[B.C.J.O]

On Thu, 28 Feb 2002, Tille, Andreas wrote:

 once I had hardware failures on my E250 server which I was able to
 reproduce also under Solaris.  This enabled me to bother the Sun
 support.  Now I am in trouble because the problem only occures under
 Linux :-(.

 After the crash I had to switch on the machine manually.  A can't to
 anything via serial line (seyon).  It does not take any signal (at least
 'Break' does not work and I did not found any other signal to realive
 the box).

Andreas,

I'm experiencing a similar problem with a similarly configured (logically
speaking) machine.

 I'm running a kernel compiled by myself with ext3fs.  It happens at least
 with 2.4.16 and 2.4.17.  Whatever problem /dev/sda4 has - the machine
 should not crash, shouldn't it?

 I tried the analogous script on /dev/sdd where I have a Solaris test
 partition.  At least I tried it with a partition at /dev/sdd, because I
 do not know which is the syntax under solaris to adress exactly the
 physical /dev/sda4.

 Any ideas to get the machine working again?
 What additional information is needed to track down the problem?

I have been running the 2.4.17 kernel (2.4.18 doesn't compile) with ext3
built right into the image on an ultra60 with two cpus and 512mb of ram.
the root filesystem is running on ext2, but I have /home mounted on a
large (36gb) ext3 partition. I have been experiencing random hard locks
that I have not been able to track down; This machine previously ran with
only ext2 filesystems (even on 2.4.17) for long periods, but as soon as I
mounted the ext3 filesystem, I started to see the locking behaviour. I
have been trying to move to a 2.4.18 or a 2.5.x development kernel from
vger, but I can't get a viable set of kernel packages built out of any of
the post 2.4.17 source. =/

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

woody/sparc64/kde

[B.C.J.O]

Currently my only woody machine is my u60, and it is exhibiting the
problems associated with the libpng/libqt snafu. I use gnome on my
desktop, but prefer konqueror to mozilla as a webbrowser; obviously
this has put a crimp in my workstyle. =) I've noticed people in
#debian claiming it has been fixed, but the fixes don't seem to have made
it into the woody pool for sparc. Any ETA on updates? My intel box is
tracking sid, and I seem to have skipped the problem entirely there. Any
help appreciated.

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

Re: woody/sparc64/kde

[B.C.J.O]

On Wed, 20 Feb 2002, Ben Collins wrote:

  Currently my only woody machine is my u60, and it is exhibiting the
  problems associated with the libpng/libqt snafu. I use gnome on my
  desktop, but prefer konqueror to mozilla as a webbrowser; obviously
  this has put a crimp in my workstyle. =) I've noticed people in
  #debian claiming it has been fixed, but the fixes don't seem to have made
  it into the woody pool for sparc. Any ETA on updates? My intel box is
  tracking sid, and I seem to have skipped the problem entirely there. Any
  help appreciated.

 woody takes time to update.

Aye.. I was just under the impression that woody had been updated for x86.

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

Re: Xsun has me beat

[B.C.J.O]

On Wed, 14 Nov 2001, Hakan Kuecuekyilmaz wrote:

  typed Xsun, but now I can't shut it down.  What's the key sequence to
  kill Xsun?  It's stuck at the Debian Welcome screen.
 
  Mario

 can't you switch your console with ctrl+alt F1, or kill the server with
 ctrl+alt+backspace...
 you will find following similar line in your startx script:

 defaultserver=/usr/X11R6/bin/X

 change this to your path to xsun.

This was a real pain when I was setting up the xconfig on my U60, because
the server would grab control of the keyboard, whether it had a working
map or not, which often left the keys dead dead dead. I just ssh'd to it
from another machine and killed off the xserver. If you don't have access
to it from another computer you may have to do resort to the power switch.
=/

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

Re: More Questions on Ultra 10: Solved

[B.C.J.O]

On Fri, 6 Jul 2001, Onno Benschop wrote:

 At 20:10 5/07/01 +0200, Karsten Heymann wrote:
 On Thu, Jul 05, 2001 at 01:04:51AM +0200 or thereabouts, Karsten Heymann 
 wrote:
 
 My problems are solved:
 
 Sound is working too without problems (and without a speaker; i never
 thought there would be a builtin speaker!).
 
 I have to disappoint you here. It's not a builtin speaker, it's the hard 
 drive head moving... ;-)

LOL! shades of 'disk musik' from the C=64 days. ;)

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

Re: X keyboard map?

[B.C.J.O]

On Fri, 11 May 2001, Jim Crumley wrote:

  Man, I've had this issue for a very long time. As far as I can tell, X has
  been completely broken on debian/sparc since xf4. I've emailed this list
  and debian-x about it, but haven't received any response, and I haven't
  made any headway on the problem myself. I can start X, the server log
  indicates that it recognizes my framebuffer, but the keyboard and mouse
  are completely ignored. Same story with xf86cfg. I'd be really very
  greatfull for any input on this problem. For reference, my own config is:
  
  Sun Ultra60,
  Creator3d ffb,
  type6 keyboard/mouse,
 
 Take a look at the list archives - this problem comes up fairly often.
 For the mouse try device /dev/sunmouse and protocol busmouse.
 For the keyboard here's what the relevant sections of my config looks like:
 
 Section InputDevice
 Identifier  Keyboard0
 Driver  keyboard
 Option  CoreKeyboard
 Option  XkbRules  sun
 Option  XkbModel  type4
 Option  XkbLayout us
 EndSection

Jim,

Thanks. This worked perfectly. I've searched through the archives
several times, but didn't come across the requisite info -- my bad. Any
idea why you have to describe a type6 keyboard as a typ4 to get it
working? =)

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

XF86 4.0.2-13, sun4u, keyboard, mouse, snafu.

[B.C.J.O]

The last time I tried to work through this problem was over a month ago,
but I ran out of time and had to put it away. I have an Ultra60 with a
creator3d framebuffer in it. This framebuffer appears to be much better
supported under woody/potato /w xf86 4.0.x, but when I move from potato to
either woody or sid the keymaps in the xserver go crazy, and the system
has no idea about the maus. The keyboard is a newish sun type6 /w
accompanying mouse. I would appreciate it if somebody would be good enough
to paste an operational InputDevice XF86Config section for a type6
keyboard and mouse.. even any pointers to the answer would be greatly
appreciated. This problem doesn't appear to be a transient bug,
unfortunately, as I've been fiddling away at the problem /w frequent
updates from the relevant sources for some time now. Any clues welcome.

Brian
[EMAIL PROTECTED]

You can't depend on your judgement when your imagination
is out of focus.   -- Mark Twain

Re: some strange kernel message...

[B.C.J.O]

On Mon, 19 Feb 2001, Thomas 'Mike' Michlmayr wrote:

 On Fri, Feb 16, 2001 at 17:08:36 +0100,
 Marco Gaiarin [EMAIL PROTECTED] wrote:
  After some days of uptime (4-5) if i do ps i got:
  
  {iommu_get_scsi_one_gflush} {___f_mmu_get_scsi_one}
  Warning: /boot/System.map-2.2.18 does not match kernel data.
 
 same problem here, with kernel-image-2.2.17-sun4dm-smp on an SS10.
 
 a quick google search did not yield a positive result.

Ensure that /boot/System.map is a link to System.map-`uname -r`. In theory
removing the System.map softlink will cause the appropriate map to load on
boot.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: Lost root password...

[B.C.J.O]

On Mon, 5 Feb 2001, Marko Dinic wrote:

 I had a SparcStation 2 (full system) for a while in my closet (dead NVRAM) 
 and I just got some parts for it so it's back up and running. Now I need to 
 reset the root password. Ok, that sounds dumb. So, I don't have the cdrom, I 
 tried booting of a debian floppy (potato rescue for Sparc) and mounting the 
 sda1 (my first SCSI disk, running Solaris 2.5). The problem is that I can 
 only mount it read-only. I tried everything from mount /dev/sda1 /sunhdd to 
 mount -t ufs /dev/sda1 /sunhdd -o w,ufstype=sun. All of the commands mount 
 it, but mount it read-only. I need write access so I can change /etc/shadow 
 and delete the password hash. 
 
 Does anyone have any suggestions? I can't seem to figure out what mount 
 wants. It's weirding me out.
 
 Regards,
 Marko
 
 PS: Can you include me in the reply for now, I'll be joining the mailing list 
 shortly. Thanks.

It's been awhile since I looked at the rescue floppy for potato on sparc,
but my guess is that the ufs driver on that floppy doesn't include write
support, which is an experimental option iirc. Do you have Sun
installation media? you could boot that to change the pass, but you would
have probably done that already if that were the case. =) If you have
access to another Debian box, you could use it to generate a custom .iso
image, or you could mount the hard-drive on another box /w ufs rw support.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: Ultra60, XF4.0.1pre2.RC3-1, woody

[B.C.J.O]

On Wed, 20 Dec 2000, Ben Collins wrote:

Section InputDevice
Identifier  Keyboard0
Driver  keyboard
Option  XkbRules  xfree86
Option  XkbModel  sun5
Option  XkbLayout us
EndSection
   
   Change sun5 to type5
  
  the punctuation keys are still all fubar. enter gives lower c etc.
 
 Ok, also change the XkbRules to sun

Ben, I have changed the XkbRules option to sun which completely flumox's
the keymap so nothing is as it should be. The current config clause shows:

Section InputDevice
Identifier  Keyboard0
Driver  keyboard
Option  XkbRules sun
Option  XkbModel type5
Option  XkbLayout us
EndSection

... and as I say nothing is as it should be, now. =) I'm getting this line
in the /var/log output again:

Couldn't load XKB keymap, falling back to pre-XKB keymap

... suggestions readily welcomed.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Ultra60, XF4.0.1pre2.RC3-1, woody

[B.C.J.O]

'morning

I'm pretty baffled by an issue I'm having /w a new Ultra running woody /w
XF4.0.1pre2.RC3-1 .. I've fidled with the input clause for the keyboard in
a seemingly endless series of variations, but I can't get the X server to
come up /w sane key mappings. The primary letter keys seem to function,
but the control and command keys are all twisted and mismatched. This is
the clause I'm using:

Section InputDevice
Identifier  Keyboard0
Driver  keyboard
Option  XkbRules  xfree86
Option  XkbModel  sun5
Option  XkbLayout us
EndSection

the keyboard is a sun type 5, recognized at boot:

su(mouse) at 0x1fff13062f8 (irq = 4,7ea) is a 16550A
Sun Mouse-Systems mouse driver version 1.00
su(kbd) at 0x1fff13083f8 (irq = 9,7e9) is a 16550A
Sun TYPE 5 keyboard detected without keyclick

On the console the key mappings are all fine. The machine is running a
custom kernel:

Linux wookiee 2.4.0-test12 #2 SMP Mon Dec 18 02:55:28 EST 2000 sparc64
unknown

which I installed to get the SMP support for the second CPU in the box. I
imagine I've done something completely boneheaded, but I'm not sure
what. =( This is the output from /var/log:

(==) ServerLayout XFree86 Configured
(**) |--Screen Screen0 (0)
(**) |   |--Monitor Monitor0
(**) |   |--Device Card0
(**) |--Input Device Mouse0
(**) |--Input Device Keyboard0
(WW) The directory /usr/X11R6/lib/X11/fonts/CID/ does not exist.

[clip]

(II) /dev/fb0: Detected FFB2+/vertical, Z-buffer, Double-buffered.
(II) /dev/fb0: BT498 (PAC2) ramdac detected
(II) /dev/fb0: Detected Creator/Creator3D
(0): [drm] created ffb driver at busid SBUS:/dev/fb0
(0): [drm] added 8192 byte SAREA at 0x1bd84000
(0): [drm] mapped SAREA 0x1bd84000 to 0x70028000
(0): [drm] framebuffer handle = 0x0200
(0): [drm] added 1 reserved context for kernel
(II) SUNFFB(0): [drm] FBC Register handle = 0x0060
(II) SUNFFB(0): [drm] DAC Register handle = 0x0040
(II) SUNFFB(0): [drm] SFB8R handle = 0x0400
(II) SUNFFB(0): [drm] SFB32 handle = 0x0500
(II) SUNFFB(0): [drm] SFB64 handle = 0x0600
(II) SUNFFB(0): [drm] Visual configs initialized
(II) /dev/fb0: DRM initialized
(==) SUNFFB(0): Backing store disabled
(==) SUNFFB(0): Silken mouse enabled
(II) /dev/fb0: Using acceleration
(II) /dev/fb0: DGA support initialized.
(0): X context handle = 0x0001
(0): [DRI] installation complete
(II) /dev/fb0: DRM finish setup complete

[clip]

(**) Option Protocol Microsoft
(**) Mouse0: Protocol: Microsoft
(**) Option CorePointer
(**) Mouse0: Core Pointer
(**) Option Device /dev/gpmdata
(==) Mouse0: Buttons: 3
(**) Option BaudRate 1200
(**) Mouse0: BaudRate: 1200
(II) Keyboard Keyboard0 handled by legacy driver
(II) XINPUT: Adding extended input device Mouse0 (type: MOUSE)
Couldn't load XKB keymap, falling back to pre-XKB keymap


this last line is obviously pointing at the culprit, but I'm unsure of how
to get the server to acknowledge the maps.. any help appreciated. 'cc me
off list as well pls. Cheers.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: Ultra60, XF4.0.1pre2.RC3-1, woody

[B.C.J.O]

On Wed, 20 Dec 2000, tony mancill wrote:

 Not that this is much help, but I'm having the same problem on my Ultra5
 after upgrading.  In fact, even at the console (i.e. X or no X), my
 punctuation keys are all whacked.

Well, I did find that the kbd package wasn't installed correctly when I
upgraded the box from potato. Once I manually shoved that in and copied
the appropriate keymap from /usr/share/keymaps/sun to
/etc/kbd/default.kmap.gz the console got healthy again. X is still way
snafu'd tho. =(

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: Ultra60, XF4.0.1pre2.RC3-1, woody

[B.C.J.O]

On Wed, 20 Dec 2000, Ben Collins wrote:

  I'm pretty baffled by an issue I'm having /w a new Ultra running woody /w
  XF4.0.1pre2.RC3-1 .. I've fidled with the input clause for the keyboard in
  a seemingly endless series of variations, but I can't get the X server to
  come up /w sane key mappings. The primary letter keys seem to function,
  but the control and command keys are all twisted and mismatched. This is
  the clause I'm using:
  
  Section InputDevice
  Identifier  Keyboard0
  Driver  keyboard
  Option  XkbRules  xfree86
  Option  XkbModel  sun5
  Option  XkbLayout us
  EndSection
 
 Change sun5 to type5

the punctuation keys are still all fubar. enter gives lower c etc.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: tftp boot image...

[B.C.J.O]

On Thu, 22 Jun 2000, Ben Collins wrote:

 On Thu, Jun 22, 2000 at 05:48:19PM -0400, B.C.J.O wrote:
  I'm trying to build a custom tftp boot image for a series of ultra5's with
  support for / on nfs and I've been having a hell of a time tracking down
  documentation on the process. I would be deeply gratefull if somebody
  would direct me towards some docs. The diskless boot howto on the linuxdoc
  site neglects sparc; I know the tftp boot images in the debian archives
  are catenated kernels /w ramdisks, and tilo is used to glue the components
  together, but I'm not clear on how to integrate these components.
 
 Small question, why do you want a tftp image if you are going to use NFS
 for / ?
 
 If that's all you want, compile a kernel, run elftoaout on it and use the
 resulting image for the netboot. You'll need to either pass special params
 on the PROM command, or use bootp so that it will mount the NFS.

Because I didn't know that this was possible. =) I looked at the tftp
images on the debian archive and jumped to some conclusions. These boxes
don't have disk, so I can't write the kernel out locally.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: what do I do with the System.map file after I build a new smp kernel

[B.C.J.O]

On Wed, 24 May 2000, John F. Davis wrote:

 Hello
 
 What am I supposed to do with the new /usr/src/linux/System.Map file 
 after I build a new smp kernel?
 
 I copied it to /boot but, I get this weird message everytime I
 run top.
 
 {__smp4m_processor_id} {___f___smp_processor_id}
 Warning: /boot/System.map-2.2.14 does not match kernel data.
 {__smp4m_processor_id} {___f___smp_processor_id}
 Warning: /boot/System.map does not match kernel data.

append a hyphen and the kernel version number for your new kernel after
System.map. 

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: Compiling a kernel on Sparc.

[B.C.J.O]

On Tue, 9 May 2000, Aaron Daniels wrote:

 Hi,
 I am trying to upgrade from slink to potato on my sparc, which is running the
 default kernel until now.
 libc6 says it needs kernel 2.2.7 to install so I have downloaded 2.2.7 and
 configured it then ran make dep clean modules modules_install bzImage as I
 normally would on my i386 box and all runs fine except for the bzImage. zImage
 doesnt work either, is there some special image I have to make for a sparc?

you're looking for make vmlinux.

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller

Re: Booting CD-ROM Installation

[B.C.J.O]

On Fri, 11 Feb 2000, Nornet Helpline wrote:

 I have been using Debian on x86 systems for about 3 years and have now
 come into possession of some Sun Sparc5 machines.
 These have no Floppy drive and internal hard disk I have an External
 CD-ROM which I can use to boot into the Solaris installation fine, but
 when I try to boot into the Debian CD(which boots fine on my x86 boxes)
 following the instructions on the Debian on Sparc installation
 site(http://www.debian.org/releases/slink/sparc/ch-install-methods.en.html#s-install-cd)
 it doesnt work.
 
 is a bootable sparc debian cd different to a x86 cd?

The way you phrased that explanation made it seem like you were trying to
use the x86 cd to boot the sparc machine... now, if this is the case, then
that is your problem -- you need the sparc specific dist cd. If that is
not the case, then I apologise for being pedantic. Assuming that the boot
prom knows where the bootable cdrom is, you may have a bad cd. Make sure
that the cdrom device is at scsi ID 6 if you have the defaults set on the
boot prom. 

Brian
[EMAIL PROTECTED]

How often I found where I should be going only by setting out for
somewhere else..
   -- R. Buckminster Fuller