Troubles with Potato/Woody on iBook 2001

[Laurent de Segur]

Hi all,

This message describes some problems I've ran into while installing a Debian
distr (2.2r3 + update stable) on a new iBook 2001 (P29 aka Marble).

First it seems that the kernel that ships on the CD is a pre 2.2.18 and
won't boot on this machine. So I got a 2.2.18 from Suse (god forbids!) just
to use with yaboot during installation.
 
After I was done with the installation, I downloaded the 2.2.19. More
exactly, I updated the stable (potato) on top of a freshly installed 2.2r3
and then moved the vmlinux kernel from / to my bootstrap partition where
yaboot is located, then changed the yaboot.conf file to point to the new
kernel.

The result when I reboot is a white screen with Do-QUIESCE finishedbooting
hanging (it seems to continue booting, I just can't see it.)

Rebooting with the original 2.2.18 works fine. On the iBook 2001, I have to
set "video=ofonly" to get video at boot time. Is this related? It seems that
the OS is still loading but I can't see it.


Next, I updated to the testing release (woody) on the linux partition and
leaving 2.2.18 in the bootstrap partition to start with yaboot, the keyboard
is mapped totally wrong just after loading the USB drivers from the linux
partition it seems (I type when booting and keys are ok until usb is
detected and loaded.) Note: the internal keyboard on the iBook is USB.

After installing woody, I noticed that the /lib/modules dir still doesn't
contain a 2.4 subdir in there, only a 2.2.19 (part of the potato update.) So
no drivers from 2.4 can be loaded. Is this intented to be that way on a PPC
machine? It would seem that doing a woody update on a potato update would
get me the modules for 2.4 in my linux partition, right?

Coming back to the white screen problem, if I move any kernel higher than
2.2.18 in my bootstrap partition, I get the white screen and DO-QUIESCE
frozen display even with woody installed in the linux partition (and I still
get the broken keyboard problem.)


Thanks for your help in advance,


Laurent

Problem with keyboard on iBook 2001 (marble)

[Laurent de Segur]

Hi all,

First, sorry if I use the wrong list. I am new around here.

I just noticed that running debian 2.2r3 (2.2.18) I get a message in this
message in the console when hitting the caps lock key:

"hid.c: ctrl urb status -32 received"

This happens with the internal keyboard and an external USB keyboard as
well.


Laurent

Kernel Drivers in PPC build: where are they?

[Laurent de Segur]

Hi,

Sorry if this questions looks strange, even stupid. I tried digging the
FAQ/Installation to get some answers for that, but it would seem that
PowerPC specific materials don't come out easily from the docs, at least
that is how I feel coming from the x86 side. I still like the machine better
though, so it's worse the extra efforts ;-)

Is the loading mechanism for kernel modules supported on the PPC platform?
Can I compile my kernel without a specific module then compile the driver
and store it in /lib/modules// or am I better off compiling
as a big monolithic kernel? My problem is understanding where the kernel
image stored on the bootstrap partition (vmlinux) kicks in the root
partition and can see the drivers if it can see them at all on it.

I was looking at the PowerPC installation disk for debian. I noticed that
there is a file in there called drivers.tar.gz. What am I supposed to do
with this file? Should I leave it in the bootstrap dir uncompressed next to
yaboot and vmlinux files? Decompress it and move it in the /lib/modules/ dir
in my root partition? Puzzled.

Thanks for your help in advance.


Laurent

Re: Kernel Drivers in PPC build: where are they?

[Laurent de Segur]

on 9/3/01 5:48 PM, Ethan Benson at [EMAIL PROTECTED] wrote:

> 
> the kernel is not stored on teh bootstrap partition, its stored on
> your ext2 / partition or your ext2 /boot partition.  never on teh
> bootstrap partition.
> 


Are you sure? That's big news to me! I need to readjust here (and I need a
drink too :-) 

If what you say is correct, then I have always misunderstood this and
probably got confused since the days of BootX where you needed to store your
kernel in the hfs partition (where bootx was residing.) Now that you are
saying that, I kinda understand why nowhere it was mentioned to copy the
kernel image to the bootstrap partition. Indulge me, but please confirm
this. Thanks,

Laurent

Re: Troubles with Potato/Woody on iBook 2001

[Laurent de Segur]

on 9/2/01 11:34 PM, Arrigo Benedetti at [EMAIL PROTECTED] wrote:

>> Rebooting with the original 2.2.18 works fine. On the iBook 2001, I have to
>> set "video=ofonly" to get video at boot time. Is this related? It seems that
>> the OS is still loading but I can't see it.
>> 
> 
> Yes, I got the Do-QUIESCE message until I started giving the "video=ofonly"
> argument to the kernel. You can put append "video=ofonly" in yaboot.conf as
> you would do with lilo.conf on x86 linux.
> 


Hi again,

The problem is still here. Reading debian docs for half a day didn't solve
anything :-( - I learned a lot more about various policies though...

Here is what I do:
- I built the kernel the debian way (make-kpkg yadi yada...)
- The kernel image I get in /boot after the kernel successfully built and
installed is vmlinux-2.2.19. This image, I copy to my bootstrap partition
(hfs) where yaboot resides.
- I reboot with the yaboot.conf pointing to the new image.
- I DO use 'video=atyfb video=ofonly' in the append string.

When the iBook restarts, I still get the Do-QUIESCE message. Linux boots
(supposedly as I hear the disk) but the OF screen is blocking the view
kinda.

Switching to linux-2.2.18 which I got from somewhere else works. It's just
not fair!

What's wrong with my procedure? Did I forget to tell set a flag when
building the kernel, like don't call quiesce or something (shouldn't it be
yaboot who calls quiesce anyway)? I am running out of ideas... Can someone
help plase?

Thanks a lot,

Laurent 

Re: X 4.1.0 Hints for french PowerBook users...

[Laurent de Segur]

Hi,

I support the idea of using command (apple) key instead of the alt key on
the mac keyboard (or if you use a mac keyboard on a pc.) One more reason is
that 'alt' is primarely labeled 'option' in the mac parlance and that should
be its major task: generating special chars. I wonder how many apps that
would break though (the ones that use option-command-something would
definitely suffer the change.)

Talking about keyboard mapping, I used a couple of powerbooks in the past
under macos 9/X, and as you probably know these laptops have a fn key.

On the MacOS, Apple uses the Fn key + F1-F12 to generate a function key
event. This, I think, is pretty stupid: how many times do you change the
screen luminosity vs. doing cut/paste in an app just to mention one example.

Still on the powerbook running macos, you can get hit enter by using
fn+return and enter has its own key on the right of command-right (note:
enter is different than return on the mac.) I wish they would have been put
an option/alt key instead of the single enter key. On MacOS, I was using a
utility that remapped the keyboard just like I wanted it to be (I guess I
was not the only one annoyed by this.)

How is it done under Linux/XF86? Is there any 'easy' way to change the
mapping of your keyboard without recompiling the kernel or whatever?

Thanks for your info,

Laurent

on 9/3/01 2:08 PM, Florent Pillet at [EMAIL PROTECTED] wrote:

> on 3/09/01 19:28, Tuomas Kuosmanen at [EMAIL PROTECTED] wrote:
> 
>> This has the effect of making []{} etc to work. However, it disables ALT
>> for me (though I tried it on the finnish/swedish keymap) - thus you
>> cannot Ctrl-Alt-Backspace to kill the X server, nor can you use Alt for
>> anything (alt-drag to move windows for example on many window managers)
>> 
>> Tuomas
> 
> Interesting.
> 
> I'm not an expert with this keyboard configuration yet. Since xkb setting
> files seem to be under-documented, I'm trying to make my way through them.
> I'm trying to find a fix for this. The default X mappings for Mac keyboards
> are not perfect since I think ALT is misused (the Command key should be used
> instead). Therefore, I wonder if it would not be more appropriate to use the
> Command key as "Alt_L" and leave the Alt key for what it is really used
> under MacOS, that is access additional characters on the various keys.

Key mapping issues

[Laurent de Segur]

Hi,

To follow up on the key mapping messages from earlier, I discovered that if
I want to switch to any virtual console, the left command key won't always
work on my iBook. I have to use the right command key instead.

On the iBook (and I believe pb, not sure), I also have to press the Fn key
to get any function key to work (ala Apple, but not the best choice IMHO.)

So for example, to get the virtual console 1, I'll have to press:
Fn-ctrl-RightCommand-F1 (No, I don't have to use my feet to help, but
almost.)

Is there a reason why the Left command key won't work to do the job?

Also, I'd really like not to have to press this Fn key on a *book just for
the F1-F12 keys (reverse with the audio/video controls which are used less
frequently.) Of course, the rest of the function keys could stay the same
(keypad and arrows.)

Has anyone any suggestions regarding this functionality?


Laurent

Re: debian-powerpc: yaboot & ybin bug ???

[Laurent de Segur]

on 9/3/01 8:05 PM, Ethan Benson at [EMAIL PROTECTED] wrote:

>> I could not reboot to my old system easily.  I had to boot off the CD
>> which was a pain because I had to search the web to find the right OF
>> commands and right kernel parameters to pass to the kernel (ie.
>> video=ofonly).
> 
> not true
> 
> at the yaboot boot: prompt:
> 
> hd:3,/boot/vmlinux root=/dev/hda3 ro



Actually, I noticed that, since with my particular config, I need to add
video=ofonly at the end of the param line (when I want to recover in the
same way as Brendan), the penguin logo will stick to the top of the screen
and never disappear until I reboot.

The cute mascot is eating about 4 lines of text on top of the display only
if I enter the parameter line like you show above (+ video=ofonly appened to
it). This won't happen when I do a standard selection. Is that a bug?

Laurent

A couple of more issues...

[Laurent de Segur]

Hi,

Okay, I got the new kernel 2.4.8-powerpc image installed on top of testing
(woody). I start to like this apt-get dpkg stuff. Coming from rpm, it was
somewhat hard to get used to. But I can see why people who have been using
it for a while, say good things about it.

About kernel 2.4.8 installed running on iBook 2001 (marble aka p29):

- There is some garbage just after kernel starts booting. By garbage I mean
some green semi-graphic rectangular shape in the middle of the display.
Reminds me the TRS-80 (these were the good old days...)

- I get a couple of pcmcia error messages at the end of the kernel startup:
i82365.o and ds.o complain about no such device and operation not permitted
respectively. No big deal but I didn't get that previously. IMHO, if I do an
install of a powerpc kernel, modules should be able to cope with any ppc
platforms without throwing error messages.

- The penguin logo is still sitting at the top left corner all the time,
blocking about 4 lines of screen. I can't get rid of it. I tried everything
in yaboot.conf (video=ofonly, video=atyfb, video=aty128fb in various
combinations). Looks like an add banner in my console. I wonder what is
causing it to happen. At one point I didn't have that, that was with 2.2.19
I think (there have been so many configs I tried... I kinda forgot.)

- Finally X windows won't start up. Period. It has nothing to do with kernel
2.4.8. It has been like that even since I did the update to woody. If I
enter the command startx, it tells me that it can't stat /etc/X11/X. All of
that after a simple testing upgrade on top of stable (which I need because
XF86 3.x won't work with aty128 on iBook.) What's weird is that I had to go
and edit the inittab to bump the run level from 2->5. Installing X, I
thought that it would do that for me. I thought wrong.

Voila. I really hope that I can get some if not all of these problems
resolved soon so I can one day enjoy running debian. Thanks for letting me
know if you have any suggestions. I certainly greatly appreciate your help.


Laurent

Re: A couple of more issues...

[Laurent de Segur]

on 9/4/01 8:49 AM, Colin Walters at [EMAIL PROTECTED] wrote:

>> - The penguin logo is still sitting at the top left corner all the
>> time, blocking about 4 lines of screen. I can't get rid of it. I
>> tried everything in yaboot.conf (video=ofonly, video=atyfb,
>> video=aty128fb in various combinations). Looks like an add banner in
>> my console. I wonder what is causing it to happen. At one point I
>> didn't have that, that was with 2.2.19 I think (there have been so
>> many configs I tried... I kinda forgot.)
> 
> Ethan suggested "setterm -reset" in another thread.  Does that work?
> I've never seen this on my tibook though.


Yes, setterm -reset works. Thanks for the help. One more thing to type after
I login each time.

Laurent

Re: Troubles with Potato/Woody on iBook 2001

[Laurent de Segur]

on 9/4/01 1:28 AM, Bastien Nocera at [EMAIL PROTECTED] wrote:

> When the iBook restarts, I still get the Do-QUIESCE message. Linux boots
>> (supposedly as I hear the disk) but the OF screen is blocking the view
>> kinda.
>> 
>> Switching to linux-2.2.18 which I got from somewhere else works. It's just
>> not fair!
>> 
>> What's wrong with my procedure? Did I forget to tell set a flag when
>> building the kernel, like don't call quiesce or something (shouldn't
> 
> It's probably working, you just have the wrong video driver. The iBook
> 2001 has an ATI Rage 128 which needs the aty128 driver, not the aty
> driver which is for Mach 64 cards.
> 
> Personally, I would install a 2.4 kernel (I have one I compiled for my
> iBook on my website). For laptop users, IMHO, 2.2 is useless.
> 
> Cheers

Hi Bastien,

Do you have a link to your precompiled kernel. I'd like to give it a try
(penguinppc is still down AFAIK).

Also, what do you pass in your append string. Right now I need to put
'video=ofonly'. Do I need to change that with the new kernel?

Thanks,


Laurent

Re: debian-powerpc: yaboot & ybin bug ???

[Laurent de Segur]

on 9/3/01 11:57 PM, Matt Brubeck at [EMAIL PROTECTED] wrote:

>> Actually, I noticed that, since with my particular config, I need to
>> add video=ofonly at the end of the param line (when I want to recover
>> in the same way as Brendan), the penguin logo will stick to the top of
>> the screen and never disappear until I reboot.
> 
> Does it stay there even if you switch to another virtual terminal
> (alt+Fx) or run an X server?
> 
> 

Matt,

I don't know. Since I installed kernel 2.4.8-powerpc from Debian, I can't
switch to another virtual terminal anymore :-(

I have tried any combination of keys even left and right alt but nothing
will bring another virtual terminal.

Any idea?

Laurent

Re: A couple of more issues...

[Laurent de Segur]

on 9/4/01 8:49 AM, Colin Walters at [EMAIL PROTECTED] wrote:

> 
>> - Finally X windows won't start up. Period. It has nothing to do
>> with kernel 2.4.8. It has been like that even since I did the update
>> to woody. If I enter the command startx, it tells me that it can't
>> stat /etc/X11/X.
> 
> Hm, strange.  Do you have the `xserver-xfree86' package installed?


No, I didn't. Thanks for pointing that problem to me. How come was I able to
install xfree86 with out being warn about not selecting xserver? Does that
mean that I need to select the individual xfree86 packages manually? This
would be really hard to remember all the names of various packages if it was
the case. I probably misunderstood something.


When I do apt-get install xserver-xfree86, I get a nasty string during
installation that shows in the dialog next to ok, stating that 'Failed to
open configuration file /etc/exim/exim.conf'. I think that I read this in
another message. Not sure what is causing it.

I got my XF86Config-4 file from someone who's got the same iBook 2001. When
I do a startx, the error message I get is: no screens found. The log stated
that Screen(s) found but none have a usable configuration (!) I checked that
my screen was set correctly for this machine and it seems like it. FYI, I
run the xfree86 from Woody (4.0.3). Has anyone with a similar machine ran
into that problem?

Do you have any more suggestions? Thanks for your help in advance.


Laurent

Re: A couple of more issues...

[Laurent de Segur]

Yeap, that works. Thanks Matt!

Switching virtual terminals is still funky (now, I need to press control on
top of fn/right-command/F keys...) Not perfect, but that works.

Oh, and sometimes switching terminals with these keys invokes a soft reboot,
so there is obviously something broken in the usb keyboard driver at least
for this model.


Laurent


on 9/4/01 10:44 AM, Matt Brubeck at [EMAIL PROTECTED] wrote:

> On Tue, 4 Sep 2001, Laurent de Segur wrote:
> 
>> Yes, setterm -reset works. Thanks for the help. One more thing to type
>> after I login each time.
> 
> To use Ethan's suggestion from earlier in this thread, just
> 'apt-get install mingetty' and then edit /etc/inittab to replace
> "/sbin/getty 38400" with "/sbin/mingetty".
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: A couple of more issues...

[Laurent de Segur]

on 9/4/01 12:24 PM, Michel Dänzer at [EMAIL PROTECTED] wrote:

> Laurent de Segur wrote:
>> 
>> I got my XF86Config-4 file from someone who's got the same iBook 2001. When
>> I do a startx, the error message I get is: no screens found. The log stated
>> that Screen(s) found but none have a usable configuration (!) I checked that
>> my screen was set correctly for this machine and it seems like it. FYI, I
>> run the xfree86 from Woody (4.0.3). Has anyone with a similar machine ran
>> into that problem?
> 
> Yes, almost everyone for whom it doesn't work. ;)
> 
> That's a generic error which can be caused by a myriad of problems. Looking
> carefully at /var/log/XFree86.0.log might reveal the cause. One common problem
> is that you have to provide a bus ID.
> 


Actually, I did. Here is my device setup:

Section "Device"
Identifier  "Card 1"
Driver  "r128"
VideoRam8192
BusID"PCI:0:16:0"
Option"UseFBDev" "true"
# Insert Clocks lines here if appropriate
EndSection


lspci will show that

00:10.0 VGA compatible controller: ATI Mobility M3 AGP 2x (rev 02)


So this should work but it doesn't.

My next step if I can't find the problem with 4.0.3 is to go to 4.1 (I
believe that sid includes xfree86 4.1.x) but I'd rather have it working now
the way it is.

If someone has any suggestion, I really appreciate.
Thanks,

Laurent

L

Problem with XFree and Internal LCD screen

[Laurent de Segur]

Hi,

In my tribulations to get xfree86 to show some graphics, I got scared to
death thinking that I had zapped the LCD monitor.

You will find below my XF86Config-4 that I copied. The machine is an iBook
2001 (dual-usb)

Under xfree86 4.0.1 (woody), I reported earlier that I was getting a message
stating that the monitor couldn't be found. I started experimenting with the
XF86Config  script and simply removed the 'Option "UseFBDev" "true"' from
the Device section. Then I entered startx and I didn't get this message
anymore. X started! So this particular option was causing the error message
I got under 4.0.1 all along.

But even if X started, things didn't work really, my LCD started to flicker
and the screen lighted up with weird patterns like if sand was flooding the
display from the bottom. I couldn't take it, having read horror stories
about monitors getting zapped with wrong params. So I hit the 3 finger
salute and rebooted.

Just after the reboot, the screen was still flickering even when booting
MacOS 9! I went under the monitor cp and changed a couple of time the
resolution, things got okay again after a while. Wow! I came close to call
1-800-MY-APPLE ;-)

Right now, I don't know how I can reboot in terminal mode only because I
know that if I restart with Linux, X will kick on again and do the same. I
don't want to take any risk.

Could someone please let me know the safest way to restart debian and log in
terminal mode without any init scripts trying to start XFree86?

I really appreciate your help.


Laurent


--




Section "Device"
Identifier  "Card 1"
Driver  "r128"
VideoRam8192
BusID"PCI:0:16:0"
Option"UseFBDev" "true"# I removed that line
EndSection


Section "Screen"
Identifier  "Screen 1"
Device  "Card 1"
Monitor "Monitor 1"
DefaultDepth 24

Subsection "Display"
Depth   8
Modes   "1024x768"
ViewPort0 0
EndSubsection
Subsection "Display"
Depth15
Modes"1024x768"
ViewPort0 0
EndSubsection
Subsection "Display"
Depth   16
Modes   "1024x768"
ViewPort0 0
EndSubsection
Subsection "Display"
Depth   24
Modes   "1024x768"
ViewPort0 0
EndSubsection
EndSection

Section "Monitor"
Identifier  "Monitor 1"
HorizSync   48.359
VertRefresh 60
EndSection

Re: A couple of more issues...

[Laurent de Segur]

What we could do is reference the machine by model and the year it was born.
Apple is less likely to release two machines of the same kind in the same
year (As we could witness during the last MW...)

We could also use the Apple tag number for the machine (SKU MM/)
although it's not really easy to remember.

For outside developers what would make the most sense would be to designate
the machine the same way as they are described internally during the
development phase (iBook 2001 = P29) or by their codename (iBook 2001 =
Marble).

I like the codename designation, and I believe there are a few mac sites
with a chart showing equivalence between models and codenames.

But from what I read on this list previously, some people seem to favor the
marketing designation of the machine, which as you pointed in your email,
can get really confusing.

There were a time when Apple used numbers for its models which made it
really simple to frame a model, that was when they had more than 20
platforms in their catalog, Performa anyone? SJ's return brought many
changes.


Laurent


on 9/4/01 4:06 PM, Ethan Benson at [EMAIL PROTECTED] wrote:

> On Tue, Sep 04, 2001 at 05:02:17PM +0100, Bastien Nocera wrote:
>> 
>> I can understand that saying "the new iBook" doesn't help very much
>> especially when a new one comes along. The Apple Tech Info Library uses
>> "iBook Dual-USB (2001)" as the full name. And saying "icebook" doesn't
>> help much, it's just like when somebody says they have a graphite g4, a
>> indigo imac or a tangerine ibook: it's useless (at least after a second
>> version with the same look pops up).
> 
> blame apple's marketing deptartment for naming all the computers
> indistincivly.
> 
> there are now 3 variations of `PowerMac G4' one is a BlueG3 minus
> firmware trojan, plus gray case, and plus G4 processor, the other is
> the AGP model, and now we have another AGP model that has the nickname
> quicksilver  
> 
> imagine the nightmare it must be to be an Apple tech support agent,
> typical MacOS users are unlikly to know about such names as `iBook
> Dual-USB' much less that they HAVE the Dual-USB version.
> 
> say what you will about the old numbering system at least you know
> what the differences are between a 7200 and a 8500. of course apple
> could have simply used a version number system
> 
> iBook Version 1
> iBook Version 2
> PowerMac G4 Version 1
> PowerMac G4 Version 2
> PowerMac G4 Version 3
> 
> but that would blow apples current `easy to use' naming system
> wouldn't it? 
> 
> -- 
> Ethan Benson
> http://www.alaska.net/~erbenson/
> 

Thanks

[Laurent de Segur]

Hi all,

Thanks to everyone on the list for helping me out, I got Debian running on
my iBook ;-)

I went from having almost fried the machine to now running gnome/sawfish and
using mozilla. What an strange and intense feeling...

I can't play mp3s yet but I am working on it.

Debian-PowerPC on the iBooks simply rocks!

Cheers,


Laurent

Re: OT: Darwin

[Laurent de Segur]

May I suggest that for Darwin discussions, you go to the proper place:

http://www.opensource.apple.com/

and subscribe to any of the numerous lists offered on the Apple web site.

Not only will you find a lot more valuable feedback regarding installing and
running darwin on apple lists, you won't even have to post a message
starting by an off-topic warning.

Cheers,

Laurent

on 9/7/01 5:27 AM, Berg, Björn at [EMAIL PROTECTED] wrote:

> Hello,
> 
> I know, it's a little bit off topic, but has someone used Darwin here on his
> Apple Macintosh? What do you think about it. Is it an alternative to Linux?
> 
> Greetz :-))
> 
> Björn
> 
> --
> Björn Berg
> http://www.bbinternet.de
> 
> 
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Comment regarding kernel version

[Laurent de Segur]

Hi,

I noticed that after installing either 2.2, either 2.4 pre-built kernel
binary package, I end up getting a kernel version with an appended string
(either -pmac, either -powerpc.) If I write a driver for this installed
kernel, should I change the version in the header version.h by appending the
string manually? If I don't and version dependency is use (automatically),
the module won't load on kernels that folks build themselves (I know, I
tried.)

What should I do if I still want to have version dependency turned on?
Any reason why this string is present in the version string (it can't be
even guaranteed that it will be the same in future versions.) Having the cpu
inside the version looks redundant to me since I get the package from the
platform directory already.

Thanks for your help,

Laurent

Re: Comment regarding kernel version

[Laurent de Segur]

Josh,

I still don't understand why the kernel package installed from a pre-built
image package returns either a -ppc on 2.2 either a -powerpc on 2.4 added to
the OS version since the platform is already returned using the uname -m
command.

Laurent

> Laurent de Segur <[EMAIL PROTECTED]> writes:
> 
>> I noticed that after installing either 2.2, either 2.4 pre-built
>> kernel binary package, I end up getting a kernel version with an
>> appended string (either -pmac, either -powerpc.) If I write a driver
>> for this installed kernel, should I change the version in the header
>> version.h by appending the string manually? If I don't and version
>> dependency is use (automatically), the module won't load on kernels
>> that folks build themselves (I know, I tried.)
> 
> No you shouldn't change the version string.  Build the module against
> the kernel source that you intend to use it with.
> -I/usr/src/linuxppc_2_4 (or whatever) when you're building should do
> the trick.
> 
>> What should I do if I still want to have version dependency turned
>> on?  Any reason why this string is present in the version string (it
>> can't be even guaranteed that it will be the same in future
>> versions.) Having the cpu inside the version looks redundant to me
>> since I get the package from the platform directory already.
> 
> Well, to put it simply, kernel modules are dependant on the kernel
> version that they were built for.  So, you need to distribute source
> to your module if you want other people to use it.  (or, distribute
> modules with a kernel to go along with them...)
> 
> ttyl,
> 
> -- 
> Josh Huber   | [EMAIL PROTECTED] |

Compile err with kernel-2.4.9 : kbd_rate undeclared when building on ppc.

[Laurent de Segur]

Hi,

I tried compiling the latest kernel (2.4.9) in order to use linux keycodes
(which seems to be required as I updated XF86 from 4.0 to 4.1.)

During the kernel build, the compiler stopped with an error:

drivers/char/vt.c:507: 'kdb_rate' undeclared.

Examining the file vt.c, the segment of code where the compile err happens
is apparently used for m68k/i386 according to the comment above the case
KDKBDREP: line. 

Note that this code doesn't use any conditional defines for this arch (is
that ok?) I did a grep to find out where kbd_rate is defined, and found that
it is not defined for the ppc arch, but only for m68k, i386, arm, and alpha
(in keyboard.h: #define kbd_rate pckbdrate.)

In order to build the kernel, I have been using the standard menuconfig to
set the build flags, so I don't think my .config is bogus: no flag has been
set by editing the .config manually, just by setting menu options.

Did I miss something? I was able to compile 2.4.8 on the same system
recently (but I didn't have to change the keyboard selection; this is the
only difference.)

Thanks for your help in advance,


Laurent 

Re: Compile err with kernel-2.4.9 : kbd_rate undeclared when building on ppc.

[Laurent de Segur]

on 9/11/01 12:20 AM, Ethan Benson at [EMAIL PROTECTED] wrote:

> On Mon, Sep 10, 2001 at 11:04:19PM -0700, Laurent de Segur wrote:
>> Hi,
>> 
>> I tried compiling the latest kernel (2.4.9) in order to use linux keycodes
>> (which seems to be required as I updated XF86 from 4.0 to 4.1.)
> 
> 2.4.9 is broken, use something else.  2.4.8 or 2.4.10pre[456] (maybe)


Anf apparently 2.4.8 is broken too. When building vmlinux, I get this
compile err:

In function pci_init_s182c105: undefined reference to ide_special_settings

I'd really like to know from which sources was built the image named:
2.4.8-powermac, the one I was running without any problems before upgrading
XFree and having the keymapping issue. All I would need to do is turn off
ADB event codes in the config and recompiled.

I tried getting a 2.4.10pre as you mentioned (setting my apt-get to
unstable, sid) but there seems to be some trouble when doing an apt-get
install kernel-image. Under testing, I get a choice of versions, with
sid/testing, I get a reported error that it can find the location.

Right now, I've got sources for 2.4.8 and 2.4.9 both unable to compile,
XFree86 with a broken keymapping since I updated from 4.0 to 4.1. Not good.

Thanks for your help,

Laurent

RE: XFree86 update question

[Laurent de Segur]

...and since most people don't read them anyway, expect great chaos on this
list when the sucker is released.

L
-Original Message-
From: Ethan Benson [mailto:[EMAIL PROTECTED]
Sent: Wednesday, September 12, 2001 7:24 AM
To: debian-powerpc@lists.debian.org
Subject: Re: XFree86 update question

when woody is released this WILL be in the release notes, if you don't
read them you get what you deserve.

--
Ethan Benson
http://www.alaska.net/~erbenson/

Re: XFree86 update question

[Laurent de Segur]

Hi,

I did my home work, and still get bitten by this keymapping problem.

I added append="keyboard_sends_linux_keycodes=1" in my yaboot.conf, and now
I get the XFree86 to play nice and give me the right keymapping.  The only
trouble is that the keymapping for the console before X starts is now wrong.

I know this since I have netenv installed, and it won't work correctly.
Actually, while booting if I hit keys, I can see that they don't map
correctly as they output to the console.

Same story with virtual consoles after X started. Keymapping is all screwed
up inside them too (so no way to find out what keys under this mapping I
should enter to go back to X.)

It seems that the append="keyboard_sends_linux_keycodes=1" option has fixed
the problem with XF86, it's just that now that the console has a broken
keymapping. Is it one or the other?


I have a kernel-image-2.4.8-powerpc needed for the iBook 2001 because of
aty128fb support (as mentioned earlier with my do-quiesce screen mail)
My other email mentioned a problem both compiling 2.4.8 and 2.4.9 from
kernel-sources so I am kinda stock with the binary image only.

Is the keyboard USB on the new iBook 2001? It looks like it, since the
keymapping isn't broken at boot time until USB gets loaded and replaces OF
driver.

If someone can give me a pointer, I would sincerely appreciate.


Laurent

on 9/12/01 4:05 PM, Ethan Benson at [EMAIL PROTECTED] wrote:

> On Wed, Sep 12, 2001 at 08:31:20PM +0200, Michael Schmitz wrote:
>> On Wed, 12 Sep 2001, Laurent de Segur wrote:
>> 
>>> ...and since most people don't read them anyway, expect great chaos on this
>>> list when the sucker is released.
>> 
>> That was clear from the beginning. It was predictable the moment the
>> Powers That Be decided to force everyone to switch over to Linux keycodes
>> now, not wait until release. Nothing that creative use of procmail
>> couldn't fix :-)
> 
> look at it this way, the input layer we switched to AFTER release, we
> broke debian stable.  this time were doing it right; in woody before
> its considered stable.
> 
> and people who refuse to rtfm can be ignored and blackholed with
> procmail yes, and they deserve it.
> 
>> Make no mistake: the new input layer is the cleaner of both options, and
>> having a common set of keytables for both ADB and USB keyboards also makes
>> things easier in the long run. Plus we better hash this out now and come
>> up with a few solutions for the transition. I just resent breaking
>> backwards compatibility, that's all.
> 
> there was no way to do this transition magically, the current method
> is as good as it can be, the keymap is corrected and the kernel sysctl
> switched, the user EXPLICITY WARNED that they MUST take action to fix
> thier broken kernel, the action is trivial:
> 
> 1) apt-get install kernel-image-2.2.19-pmac
> 2) add append="keyboard_sends_linux_keycodes=1" to /etc/yaboot.conf
> 
> finally major releases of debian usually do have some sort of breakage
> that requires user intervention, or specific action to deal with, just
> look at the sparc upgrade from slink to potato.  you MUST read the
> release notes, if you don't you deserve what you get: a broken upgrade
> and being killfiled on all the mailing lists.

Re: XFree86 update question

[Laurent de Segur]

on 9/12/01 8:23 PM, Steven Hanley at [EMAIL PROTECTED] wrote:

> I find choosing apple usb means my command key is alt in console, but alt is
> alt in X as it should be, so I chooise standard and console and X are the


That worked. Great! One less problem ;-)
Thanks a lot,


Laurent

Re: HFS Plus on Linux ?

[Laurent de Segur]

My experience while using hfs volumes under Linux has been mediocre at best.
Mounting ro is ok, but writing files to it is asking for trouble.

If you really have to write to hfs partition, then you should make some hfs
shuttle partition and keep it for that task only. I do a reformat on this
partition after switching to MacOS because I've lost files on the corrupted
volume after using Linux writing to it in the past. Having now two machines,
I get away from having to do this by using ftp.

One thing you can do to find out about the quality of hfs support on Linux
is trying to run Norton Utilities and examine the partition you store files
using Linux to. The report will usually show a list of problems, some
minors, some majors. Not a good thing(tm)

Laurent 

on 9/13/01 9:45 AM, Mike Fedyk at [EMAIL PROTECTED] wrote:

>>> Ethan, do you have any specific information on when/how HFS was corrupted?
>>> What kernel versions?
>>> 
>>> I've heard about this problem for a while, but I haven't seen anything that
>>> talked about what specific operations were performed to create the
>> corruption.
>> 
>> I can't tell for corruption (well, I did _once_ have a file written from
>> linux appearing corrupted on macos), but I know HFS has some nasty locking
>> bugs. At least on SMP kernels, it can easily lockup the box.
>> 
>> One problem with writing filesystems for linux is that the locking rules
>> and the underlying cache semantics keep changing.
>> 
> 
> Then why are so many FSes supported in Linux? ;)
> 
> Yes, I know that once it has been included in the mail kernel that when
> those underlying APIs are changed that the FSes that use them are usually
> changed with them...  Though, that isn't always the case.
> 
> Mike
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Java2 for Debian PPC

[Laurent de Segur]

Hi,

I stfw'ed without much success looking for the .deb packages of the
blackdown port of jdk 1.3.1 on linux-powerpc.

Although I found on the blackdown mirrors all the debian packages for the
i386 port, the ppc dir seems to be missing. I was wondering if anyone on
this list knows where I could find the jdk debian packages for ppc?
Logically, these would be located in {potato,woody}/non-free/binary-ppc,
right? 

Also, as anyone installed the 1.3 port of the jdk and had success running
it? Any input is greatly appreciated.

Thanks for your help,

Laurent

Re: HFS+ (MacOS) in contrast to EXT2 (Linux-i386)

[Laurent de Segur]

on 9/13/01 2:02 PM, Otto Wyss at [EMAIL PROTECTED] wrote:

> I leave it up to you to draw any conclusion.


I am not sure if we can draw any conclusions from individual and isolated
cases. I found ext2 extremely reliable so far. Most boo boo I had happened
with HFS in the past (but then I was using it extensively and Mac OS crashes
more than Linux too), so YMMV.

BeFS which I was running until I switched permanently to Debian recently was
#1 on my own rating scale in terms of reliability and transparency (never
had to run a utility to check the partition for the past 3 years.)

Talking of which... Is anyone using one of the journaling fs (Reiser, XFS,
JFS) on PowerPC daily and having some comments they would like to report? I
was planning on switching to ReiserFS or XFS for my home dir. Which one, if
any, would be the most stable?


Laurent

Re: Java2 for Debian PPC

[Laurent de Segur]

on 9/13/01 3:33 PM, Michel Dänzer at [EMAIL PROTECTED] wrote:

> Someone should probably ask those building the i386 .debs if the same could be
> done for powerpc.


Thanks for your pointer.

I will send them an email regarding packaging. I also found out that the
port on ppc is lagging  somewhat behind (1.3.1 not available for ppc yet.)
And unfortunately, HotSpot is only supported on i386 at this time, because
it's seems to be done by Sun and its licensees behind closed doors.

Looks like the nice folks at blackdown could use some help on the ppc side
though.


Laurent

Re: Java2 for Debian PPC

[Laurent de Segur]

on 9/13/01 3:33 PM, Michel Dänzer at [EMAIL PROTECTED] wrote:

> Laurent de Segur wrote:
> 
>> I stfw'ed without much success looking for the .deb packages of the
>> blackdown port of jdk 1.3.1 on linux-powerpc.
> 
> Because there aren't any. Download binary tarballs from
> http://penguinppc.org/usr/java/ . OpenMotif is available in the libmotif
> package.
> 
> Someone should probably ask those building the i386 .debs if the same could be
> done for powerpc.
> 


FYI, when I visited the link you sent me, I saw the photo of a really nice
person sitting on bed, using the same computer I do. I'll have to wait for
penguinppc to get back online. I surely appreciate the way they return a 404
Error though ;-)

Laurent

Re: Java2 for Debian PPC

[Laurent de Segur]

on 9/13/01 4:57 PM, Ethan Benson at [EMAIL PROTECTED] wrote:

> penguinppc has been online for about a week now


in terms of HTML pages, they are fine. In terms of file downloads, dead in
the water (you still get the cutie poster though...) Must be hard coded
broken links all over the place.

L

Non-free packages play hard to (apt-)get

[Laurent de Segur]

Hi,

Must be a classic, although I stfw and can't find any explanation for this.

Every time I have been wanting to install a non-free package (unzip, zip,
etc...) I had to revert back my /etc/apt/sources.list to stable vs. testing
to be able to access these packages. Is it intended to be that way? I would
seem that I should be able to access the non-free from any version, but that
is not the case apparently. No big deal, just curious.

L

uname -p

[Laurent de Segur]

Hi,

Entering 'uname -p' should return 'ppc' but returns 'unknow' running
2.4.8-powerpc with debian/woody.

The uname --version returns >uname (GNU sh-utils) 2.0.11

Any clues on what could be wrong?


Laurent

Re: XFree86 update question

[Laurent de Segur]

on 9/13/01 6:07 PM, Colin Walters at [EMAIL PROTECTED] wrote:

> It is an unfortunate fact of life that people simply don't read
> documentation (there have been studies done), and debconf


I am so tempted to add to this :  "Did they have studies about who read the
studies?" 

Sorry, couldn't resist ;-)


Laurent

Re: uname -p

[Laurent de Segur]

Michel,

Beside Ethan's answer and opinion given to your question (which I support
totally since I have lots of legacy makefile from SunOS myself), I'll just
describe what 'info uname' under Linux will return:

'-m'
'--machine'
Print the machine (hardware) type.

which should probably be something like 'powermac', 'chrp', or 'prep' in the
case of a powerpc machine.

'-p'
'--processor'
Print the machine's processor type.

which should in theory be 'ppc' for any PowerPC architecture, and x86 on any
Intel {3,4,5,6}86 (I could imagine having the processor type given more
precisely on intel, since it's easier to mask the first digit to find the
ISA with a script in that particular case: *86 or ?86. That is just
convenient.)


I wrote in a previous email to express my surprise about getting a
'-powerpc' extension at the end of the version returned by the command
'uname -r' with a pre-compiled kernel image, which I think should not be
done. I understand now why such a decision was made, as it can't be
determine at runtime by entering a simple 'uname -p'.

IMHO, uname is not correctly implemented under Linux (version-wise for ports
and machine/processor-wise in general.) According to your comment, it seems
to go beyond the PowerPC architecture unfortunately for the latter, as you
mentioned that you get 'unknown' running on x86.

Being new around here, I wonder if this shouldn't be filed as a bug?
Thanks for your feedback,


Laurent


on 9/14/01 4:39 AM, Ethan Benson at [EMAIL PROTECTED] wrote:

> On Fri, Sep 14, 2001 at 01:15:15PM +0200, Michel Dänzer wrote:
>> Laurent de Segur wrote:
>> 
>>> Entering 'uname -p' should return 'ppc' but returns 'unknow' running
>>> 2.4.8-powerpc with debian/woody.
>> 
>> What makes you think it should return ppc (uname -m does) ?
>> 
>> I just tried it on a sparc64 and an i686 system and both returned unknown as
>> well.
> 
> well lets see what SunOS does/says on the matter...
> 
> [EMAIL PROTECTED] ~$ uname -m
> sun4m
> [EMAIL PROTECTED] ~$ uname -p
> sparc
> [EMAIL PROTECTED] ~$ uname -a
> SunOS terminator 5.8 Generic_108528-09 sun4m sparc SUNW,SPARCstation-20
> [EMAIL PROTECTED] ~$
> 
> from the SunOS uname man page (which is much more informative then GNU's)
> 
> 
> -mPrints the machine hardware name (class). Use of  this
> option is discouraged; use uname -p instead. See NOTES
> section below.
> 
> [...]
> 
> -pPrints the current host's ISA or processor type.
> 
> [...]
> 
> NOTES
> Independent software vendors (ISVs) and others who  need  to
> determine  detailed characteristics of the platform on which
> their software is either being installed or executed  should
> use the uname command.
> 
> To determine the operating system name  and  release  level,
> use  uname  -sr.  To  determine  only  the  operating system
> release level, use uname -r. Notice  that  operating  system
> release  levels are not guaranteed to be in x.y format (such
> as 5.3, 5.4, 5.5, and so forth); future releases could be in
> the  x.y.z  format  (such  as  5.3.1,  5.3.2,  5.4.1, and so
> forth).
> 
> In SunOS 4.x releases, the arch(1) command was often used to
> obtain  information  similar  to  that obtained by using the
> uname command. The arch(1) command output "sun4"  was  often
> incorrectly  interpreted to signify a SunOS SPARC system. If
> hardware platform information is desired, use uname -sp.
> 
> The arch -k and uname -m commands return equivalent  values;
> however,  the use of either of these commands by third party
> programs is discouraged, as is the use of the  arch  command
> in  general.  To  determine  the  machine's  Instruction Set
> Architecture (ISA or processor type), use uname with the  -p
> option.
> 
> 
> my interpretation of this is that linux is wrong, -m should be giving
> more information about the machine itself, not just the processor, for
> the processor you should be using uname -p (which on linux is
> useless).
> 
> otoh SunOS is certainly not any great role model on how to do things
> correctly IMO, but its a more informative reference to look at then
> the GNU uname man page. 

Re: uname blues

[Laurent de Segur]

> From: Matt Brubeck <[EMAIL PROTECTED]>
>
> The issue of 'uname -pm' is separate, and I agree with you there.

Matt-

You are right. These are two separate issues.


- Regarding kernel -powermac appened in pre-compiled kernel images:

OK. I think I got it. I point my includes to the header files I downloaded
separately (kernel-headers) and then I get the correct version (the
version.h is correct.) BUT: if I compile a driver in my project using these
non-custom headers (with -I/usr/src/kernel-headers...) and then insmod the
built driver against this pre-compiled kernel I am using, I get an error
with version clashing still (which I made a warning by forcing the insmod to
ignore version for now.) This is still far from ideal. So unless I recompile
the kernel (which I can't do as it's buggy and doesn't even compile), I have
to either modify the version.h file, either ignore version when doing
insmod. Nothing fancy, I just want to follow what's in the Device Drivers
book from O'Reilly. And this doesn't work when I am on testing.


- Regarding the uname -p returning unknown:

What can I do to help resolve this issue? Right now, the kludge I have
implemented to get my makefiles to function on linux without changing
anything, let alone being able to use them cross-platform, is ugly at best
(I overload the bogus uname command on linux with a script that returns what
uname -m returns for uname -p. The script lives in a path that shows earlier
than /bin. Yuk!) I do that because I don't think it's right to change all my
makefiles. I could have tested uname -s then exec uname -m in case of linux,
or a uname -p if something else. Still Yuk! uname -p is far from being
unknown: it's the processor of the machine.

I have decided to report this as a bug to [EMAIL PROTECTED] Is that the
right thing to do? Your input is, as usual, greatly appreciated.


Laurent

Re: HFS+ (MacOS) in contrast to EXT2 (Linux-i386)

[Laurent de Segur]

What you are saying about ReiserFS is really unfortunate. When I read the
archive, there is no question that XFS is a lot more reliable than ReiserFS.
On the other hand, It seems that the Linux community is promoting heavily
ReiserFS vs XFS/JFS, and that seems totally contradictory with the feedback
I got so far. 

It's too bad that the XFS module (at least on ppc) is not part of the
pre-compiled kernel image (ReiserFS is), and that the XFS kernel patches lag
a few kernel revs behind, making it mostly obsolete by the time you have to
install it on the latest 2.4 kernel (don't know about 2.2.x but can't afford
this alternative due to platform support.)


Laurent

> From: Sven <[EMAIL PROTECTED]>
> Date: Tue, 18 Sep 2001 13:32:37 +0200
> To: Michel Dänzer <[EMAIL PROTECTED]>
> Cc: Laurent de Segur <[EMAIL PROTECTED]>, debian-powerpc@lists.debian.org
> Subject: Re: HFS+ (MacOS) in contrast to EXT2 (Linux-i386)
> 
> On Fri, Sep 14, 2001 at 12:27:55AM +0200, Michel Dänzer wrote:
>> Laurent de Segur wrote:
>> 
>>> Talking of which... Is anyone using one of the journaling fs (Reiser, XFS,
>>> JFS) on PowerPC daily and having some comments they would like to report? I
>>> was planning on switching to ReiserFS or XFS for my home dir. Which one, if
>>> any, would be the most stable?
>> 
>> I posted good reasons here not to use ReiserFS. I used it for a few months
>> but
>> switched to XFS a few weeks back and I'm more than happy I did.
> 
> BTW, even on i386, reiserfs has some serious problems in some cases, and its
> use is still discouraged. my (i386 smp) box crashed various times because of a
> ide dma error while running apt-get upgrade, and reiserfs completely corrupted
> the /var/lib/dpkg/status file, and a few others as well, which resulted in a
> rather hhopeless system, with regards to dpkg (it mostly worked well but the
> dpkg database was hoosed).
> 

openmotif not avail for ppc?

[Laurent de Segur]

Hi everyone,
 
It seems that the openmotif debian package is 
available only for x86 architecture, but doesn't seem to be present in the ppc 
side (on main/testing/unstable.) I went to visit the openmotif.org web site, and 
there, you can only get an RPM (yuk!) for the ppc platform (and debian packages 
for Ismell Inside platforms). Can someone points me to the right direction? I'd 
really like to install openmotif the Debian way.
 
Thanks,
 
Laurent

Re: openmotif not avail for ppc?

[Laurent de Segur]

Josh,

Things are not as simple for me as they are for you apparently.

Entering 'fakeroot apt-get -b source openmotif' breaks just after building
the docs with the compile error message:

make[3]: Entering directory
`/home/laurent/openmotif-2.1.30/work/motif/config/imake'
gcc DefaultGcc2PpcOpt -fsigned-char  -I../../include
-I../../imports/x11/include/X11  -I../.. -I../../exports/include
-I../../imports/x11/include  -Dlinux -D__powerpc__ -D_POSIX_C_SOURCE=199309L
-D_POSIX_SOURCE -D_XOPEN_SOURCE -D_BSD_SOURCE -D_SVID_SOURCE  -D_GNU_SOURCE
-DNO_MESSAGE_CATALOG  -DFUNCPROTO=15 -DNARROWPROTO
-DCPP_PROGRAM="\"/lib/cpp\""   -c -o imake.o imake.c
gcc: cannot specify -o with -c or -S and multiple compilations
make[3]: *** [imake.o] Error 1
make[3]: Leaving directory
`/home/laurent/openmotif-2.1.30/work/motif/config/imake'
make[2]: *** [imake.bootstrap] Error 2
make[2]: Leaving directory `/home/laurent/openmotif-2.1.30/work/motif'
make[1]: *** [World] Error 2
make[1]: Leaving directory `/home/laurent/openmotif-2.1.30/work/motif'
make: *** [build-stamp] Error 2
Build command 'cd openmotif-2.1.30 && dpkg-buildpackage -b -uc' failed.
E: Child process failed

I am running a vanilla testing system (did an apt-get update just before)
with gcc 2.95.4. How come your build succeeded? Curious,


Laurent



> From: Josh Huber <[EMAIL PROTECTED]>
> Organization: Debian GNU/Linux
> Date: Wed, 19 Sep 2001 17:04:45 -0400
> To: "Laurent de Segur" <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: openmotif not avail for ppc?
> 
> Hi.  Next time, could you not send HTML to the lists?
> 
> "Laurent de Segur" <[EMAIL PROTECTED]> writes:
> 
>> It seems that the openmotif debian package is available only for x86
>> architecture, but doesn't seem to be present in the ppc side (on
>> main/testing/unstable.) I went to visit the openmotif.org web site,
>> and there, you can only get an RPM (yuk!) for the ppc platform (and
>> debian packages for Ismell Inside platforms). Can someone points me
>> to the right direction? I'd really like to install openmotif the
>> Debian way.
> 
> fakeroot apt-get -b source openmotif spit out 5 debs for me, so i'm
> not sure why it's not being built.
> 
> ah, it's non-free, that's why.  the autobuilder doesn't build
> non-free.
> 
> take a look at:
> 
> http://people.debian.org/~huber/openmotif/
> 
> ttyl,
> 
> -- 
> Josh Huber   | [EMAIL PROTECTED] |
> 

Problem resizing windows horizontally

[Laurent de Segur]

Hi,

I have recently been experiencing a very annoying problem (since I updated
XFree to 4.1 in testing) with X running Sawfish default theme. In this
window style (and a couple of others), there is no right border on the
window frame, and resizing is done, I believe by selecting the window at the
bottom even when you want to resize the width. I am using a MS optical USB
wheel mouse (the white one, not the silver one, awesome mouse btw.)

When I try to select the bottom frame and resize the height by dragging the
mouse, no problemo. When I drag to the right to resize the width, no way. If
I select another style with a right border on the window, it will work.

If I select the 2nd and 3rd button on the mouse together, sometimes, it does
the trick, and will resize horizontally, sometimes not. It could be many
things (events, driver, xconfig.) and as I didn't touched any of those for a
while, I am just puzzled about this happening.

Has anyone else, seen the problem? I'd appreciate any help you can think
about, as I am running of ideas on what to try to fix this.

Thanks in advance for your help,


Laurent 

Re: [ANN] yaboot maintainer change

[Laurent de Segur]

> From: Benjamin Herrenschmidt <[EMAIL PROTECTED]>
> 
> however now get their versions from Ethan, I'll let him explain that all.

Arg! Let's get ready for this and turn on the "Vulgar Slang" filter in our
emailer.

Sorry, couldn't resist ;-)


Laurent

apt-get barfs with a SIGSEGV

[Laurent de Segur]

Hi,

Running Woody latest on ppc. I never had trouble running apt-get but today I
got apt-get to exit with with a SIGSEGV twice in a row. Fortunately, the
second time I was running the bogger with gdb and I caught this:

This GDB was configured as "powerpc-unknown-linux-gnu".
(gdb) file apt-get
Reading symbols from apt-get...(no debugging symbols found)...done.
(gdb) run -b source openmotif
Starting program: /usr/bin/apt-get -b source openmotif
(no debugging symbols found)...(no debugging symbols found)...
(no debugging symbols found)...(no debugging symbols found)...
Reading Package Lists... Done
Building Dependency Tree... Done
(no debugging symbols found)...
Program received signal SIGSEGV, Segmentation fault.
0xff7dc64 in pkgSrcRecords::~pkgSrcRecords ()
  from /usr/lib/libapt-pkg-libc6.2-3-2.so.3.1
(gdb) bt
#0  0xff7dc64 in pkgSrcRecords::~pkgSrcRecords ()
  from /usr/lib/libapt-pkg-libc6.2-3-2.so.3.1
#1  0x1001519c in _init ()
#2  0xff4bebc in CommandLine::DispatchArg ()
  from /usr/lib/libapt-pkg-libc6.2-3-2.so.3.1
#3  0x1001a248 in _init ()
#4  0xfc9a820 in __libc_start_main () from /lib/libc.so.6
#5  0x0 in ?? ()

Is this enough information to track the problem? I still have my gdb session
opened and can dump more if needed.

My apt-get is a vanilla testing release as shown below:

apt 0.5.3 for linux powerpc compiled on Mar  8 2001 20:47:22
Supported Modules:
*Ver: Standard .deb
*Pkg:  Debian dpkg interface (Priority 30)
S.L: 'deb' Standard Debian binary tree
S.L: 'deb-src' Standard Debian source tree
Idx: Debian Source Index
Idx: Debian Package Index
Idx: Debian dpkg status file

Has anyone seen that? Is this a known bug? (I'll have to admit: I am not
very familiar with the debian bug tracking system.)

Thanks for your help,


Laurent

Re: apt-get barfs with a SIGSEGV

[Laurent de Segur]

Hi,

Replying to my own post to provide more info. This problem is reproducible
when I edit my /etc/apt/sources.files and switch from testing to unstable
and vice-and-versa without doing an apt-get update and just directly doing
an apt-get source of some package. So doing the required apt-get update
after editing the file  won't show the problem (that's the workaround.)

Now, I just wished that there would be a friendlier way to tell about this
than throwing a SIGSEGV exception, because obviously there is a bogus
dereferenced pointer in one of the destructors in the apt-get lib.


Laurent

> From: Laurent de Segur <[EMAIL PROTECTED]>
> Date: Thu, 20 Sep 2001 10:42:42 -0700
> To: 
> Subject: apt-get barfs with a SIGSEGV
> Resent-From: debian-powerpc@lists.debian.org
> 
> Hi,
> 
> Running Woody latest on ppc. I never had trouble running apt-get but today I
> got apt-get to exit with with a SIGSEGV twice in a row. Fortunately, the
> second time I was running the bogger with gdb and I caught this:
> 
> This GDB was configured as "powerpc-unknown-linux-gnu".
> (gdb) file apt-get
> Reading symbols from apt-get...(no debugging symbols found)...done.
> (gdb) run -b source openmotif
> Starting program: /usr/bin/apt-get -b source openmotif
> (no debugging symbols found)...(no debugging symbols found)...
> (no debugging symbols found)...(no debugging symbols found)...
> Reading Package Lists... Done
> Building Dependency Tree... Done
> (no debugging symbols found)...
> Program received signal SIGSEGV, Segmentation fault.
> 0xff7dc64 in pkgSrcRecords::~pkgSrcRecords ()
> from /usr/lib/libapt-pkg-libc6.2-3-2.so.3.1
> (gdb) bt
> #0  0xff7dc64 in pkgSrcRecords::~pkgSrcRecords ()
> from /usr/lib/libapt-pkg-libc6.2-3-2.so.3.1
> #1  0x1001519c in _init ()
> #2  0xff4bebc in CommandLine::DispatchArg ()
> from /usr/lib/libapt-pkg-libc6.2-3-2.so.3.1
> #3  0x1001a248 in _init ()
> #4  0xfc9a820 in __libc_start_main () from /lib/libc.so.6
> #5  0x0 in ?? ()
> 
> Is this enough information to track the problem? I still have my gdb session
> opened and can dump more if needed.
> 
> My apt-get is a vanilla testing release as shown below:
> 
> apt 0.5.3 for linux powerpc compiled on Mar  8 2001 20:47:22
> Supported Modules:
> *Ver: Standard .deb
> *Pkg:  Debian dpkg interface (Priority 30)
> S.L: 'deb' Standard Debian binary tree
> S.L: 'deb-src' Standard Debian source tree
> Idx: Debian Source Index
> Idx: Debian Package Index
> Idx: Debian dpkg status file
> 
> Has anyone seen that? Is this a known bug? (I'll have to admit: I am not
> very familiar with the debian bug tracking system.)
> 
> Thanks for your help,
> 
> 
> Laurent
> 
> 
> 
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: Problem resizing windows horizontally

[Laurent de Segur]

Michel,

Was I wrong assuming that under 4.1.x, the lines in the "Module" section:

SubSection "extmod"
 Option "omit xfree86-dga"
EndSubSection

automagically load the "extmod", and therefore, I don't need to add a Load
"extmod" for this module? At least, that is why the comment above indicates.

At any rate, even adding the line with Load "extmod" just above the
SubSection doesn't correct the problem.

Did I precisely stated that this problem only shows up in specific themes
(no right window frame handle) in sawfish exclusively? windowmaker works
without any problem.


Laurent

> From: Michel Dänzer <[EMAIL PROTECTED]>
> Organization: Debian, XFree86
> Date: Thu, 20 Sep 2001 20:38:12 +0200
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: debian-powerpc@lists.debian.org
> Subject: Re: Problem resizing windows horizontally
> 
> Laurent de Segur wrote:
> 
>> I have recently been experiencing a very annoying problem (since I updated
>> XFree to 4.1 in testing) with X running Sawfish default theme. In this
>> window style (and a couple of others), there is no right border on the
>> window frame, and resizing is done, I believe by selecting the window at the
>> bottom even when you want to resize the width. I am using a MS optical USB
>> wheel mouse (the white one, not the silver one, awesome mouse btw.)
>> 
>> When I try to select the bottom frame and resize the height by dragging the
>> mouse, no problemo. When I drag to the right to resize the width, no way. If
>> I select another style with a right border on the window, it will work.
> 
> The only wild guess I can come up with is that you are missing the SHAPE
> extension. Make sure your XF86Config Section "Module" contains the line
> 
> Load"extmod"
> 
> 
> -- 
> Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
> XFree86 and DRI project member   /  CS student, Free Software enthusiast
> 

Re: openmotif not avail for ppc?

[Laurent de Segur]

And still the same problem reported originally in my first email (copied
again below) exists in unstable libmotif (2.1.30-5) that I just grabbed.
Won't compile on my woody ppc box. So far, no good.

Laurent

make[3]: Entering directory
`/home/laurent/openmotif-2.1.30/work/motif/config/imake'
gcc DefaultGcc2PpcOpt -fsigned-char  -I../../include
-I../../imports/x11/include/X11  -I../.. -I../../exports/include
-I../../imports/x11/include  -Dlinux -D__powerpc__ -D_POSIX_C_SOURCE=199309L
-D_POSIX_SOURCE -D_XOPEN_SOURCE -D_BSD_SOURCE -D_SVID_SOURCE  -D_GNU_SOURCE
-DNO_MESSAGE_CATALOG  -DFUNCPROTO=15 -DNARROWPROTO
-DCPP_PROGRAM="\"/lib/cpp\""   -c -o imake.o imake.c
gcc: cannot specify -o with -c or -S and multiple compilations
make[3]: *** [imake.o] Error 1
make[3]: Leaving directory
`/home/laurent/openmotif-2.1.30/work/motif/config/imake'
make[2]: *** [imake.bootstrap] Error 2
make[2]: Leaving directory `/home/laurent/openmotif-2.1.30/work/motif'
make[1]: *** [World] Error 2
make[1]: Leaving directory `/home/laurent/openmotif-2.1.30/work/motif'
make: *** [build-stamp] Error 2
Build command 'cd openmotif-2.1.30 && dpkg-buildpackage -b -uc' failed.
E: Child process failed




> From: Gerd Knorr <[EMAIL PROTECTED]>
> Organization: SuSE Labs, Außenstelle Berlin
> Newsgroups: lists.debian.powerpc
> Date: 20 Sep 2001 18:54:55 GMT
> To: debian-powerpc@lists.debian.org
> Subject: Re: openmotif not avail for ppc?
> Resent-From: debian-powerpc@lists.debian.org
> 
 libmotif |   2.1.30-5 |  unstable | alpha, i386, powerpc
 
>>> so it's just not in testing yet.
>> 
>> libmotif |   2.1.30-2 |   testing | alpha, i386, powerpc
>> 
>> Only packaging changes since then AFAICS...
> 
> Yup, a few compile bugs for !i386 archs and nasty dependences on the
> xfree86 version have been fixed since (see changelog).  It might be that
> 2.1.30-5 builds cleanly but 2.1.30-2 doesn't ...
> 
> Gerd
> 
> -- 
> Gerd Knorr <[EMAIL PROTECTED]>
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: apt-get barfs with a SIGSEGV

[Laurent de Segur]

Colin,

Thanks for the pointer, I just sent a detailed email to this address. I
added ppc dump, and the location in both disassembled code and source code
where I get the exception, with other additional info. Clearing the cache
didn't solve the issue, but since I provide a way for reproducing this bug
100%, it should make things easier.

Thanks again,


Laurent


> From: Colin Walters <[EMAIL PROTECTED]>
> Date: Thu, 20 Sep 2001 15:15:02 -0400
> To: debian-powerpc@lists.debian.org
> Subject: Re: apt-get barfs with a SIGSEGV
> Resent-From: debian-powerpc@lists.debian.org
> 
> Laurent de Segur <[EMAIL PROTECTED]> writes:
> 
>> Running Woody latest on ppc. I never had trouble running apt-get but
>> today I got apt-get to exit with with a SIGSEGV twice in a
>> row. Fortunately, the second time I was running the bogger with gdb
>> and I caught this:
> 
> It looks like there is a similar bug, #112361.  It doesn't look like
> it's dying in the same place, but it might be related.
> 
> I would send your debugging information to [EMAIL PROTECTED]
> Some people said that removing the cache files fixed the problem for
> them; I'd suggest trying that, and mention in the bug report whether
> or not that fixes the problem for you, too.
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: Stuck in 800x600

[Laurent de Segur]

May I suggest that you reset the nvram as OS X may have left it in mambo to
see if this fixes the problem. At boot time, press Option-Command-N-V until
you hear the chime and the machine reboots, then let go. The machine will
reboot to the first OS it finds (unless you press option) as you lost your
OF patches and env-vars, in case that matters... Re-enable them accordingly.


Laurent

> From: James Tyson <[EMAIL PROTECTED]>
> Date: Fri, 21 Sep 2001 11:03:34 +1200 (NZST)
> To: debian-powerpc@lists.debian.org
> Subject: Stuck in 800x600
> Resent-From: debian-powerpc@lists.debian.org
> 
> 
> Hi all.
> 
> I have an iBook2 which I have been running Linux on quite happily for a
> while now, however, the other day I was booting into OSX and it crashed on
> boot up, and since then the framebuffer has been stuck in 800x600.
> In OS9 or OSX it's working fine, and seems to be able to change video
> modes, however Linux is stuck in 800x600, but *thinks* it's in 1024x768.
> The output of fbset says "1024x768-60" and text scrolls past the bottom of
> the screen. Also, X seems to work, except that I can only see the top left
> 800x600 pixels of the screen (however X thinks it's in 1024x768).
> 
> Anyone know how to fix this?
> 
> Cheers.
> 
> James Tyson ---
> Samizdat New Media Solutions
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: Stuck in 800x600

[Laurent de Segur]

Er, make it Option-Command-P-R. It's getting late ;-)

Laurent

> From: Laurent de Segur <[EMAIL PROTECTED]>
> Date: Thu, 20 Sep 2001 18:26:49 -0700
> To: James Tyson <[EMAIL PROTECTED]>, 
> Subject: Re: Stuck in 800x600
> Resent-From: debian-powerpc@lists.debian.org
> 
> May I suggest that you reset the nvram as OS X may have left it in mambo to
> see if this fixes the problem. At boot time, press Option-Command-N-V until
> you hear the chime and the machine reboots, then let go. The machine will
> reboot to the first OS it finds (unless you press option) as you lost your
> OF patches and env-vars, in case that matters... Re-enable them accordingly.
> 
> 
> Laurent
> 
>> From: James Tyson <[EMAIL PROTECTED]>
>> Date: Fri, 21 Sep 2001 11:03:34 +1200 (NZST)
>> To: debian-powerpc@lists.debian.org
>> Subject: Stuck in 800x600
>> Resent-From: debian-powerpc@lists.debian.org
>> 
>> 
>> Hi all.
>> 
>> I have an iBook2 which I have been running Linux on quite happily for a
>> while now, however, the other day I was booting into OSX and it crashed on
>> boot up, and since then the framebuffer has been stuck in 800x600.
>> In OS9 or OSX it's working fine, and seems to be able to change video
>> modes, however Linux is stuck in 800x600, but *thinks* it's in 1024x768.
>> The output of fbset says "1024x768-60" and text scrolls past the bottom of
>> the screen. Also, X seems to work, except that I can only see the top left
>> 800x600 pixels of the screen (however X thinks it's in 1024x768).
>> 
>> Anyone know how to fix this?
>> 
>> Cheers.
>> 
>> James Tyson ---
>> Samizdat New Media Solutions
>> 
>> 
>> -- 
>> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
>> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>> 
>> 
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: How to Change resolution?

[Laurent de Segur]

Russel,

Why vi? Can't you just use emacs? ;-)

su -
cp /etc/X11/XF86Config-4 /etc/X11/XF86Config-4.OLD (you never know)
open /etc/X11/XF86Config-4 with your fav editor.
look for Section "Screen", and for each SubSection "Display",
change the Modes lines with your monitor value (i.e. "800x600")

I used Ctrl-Alt- Keypad +/- under KDE on a PeeSee to switch resolutions on
the fly. I don't know if this will work for you. YMMV.


Cheers,

Laurent

> From: Russell Hires <[EMAIL PROTECTED]>
> Date: Thu, 20 Sep 2001 22:25:46 -0400
> To: debian-powerpc@lists.debian.org
> Subject: Re: How to Change resolution?
> Resent-From: debian-powerpc@lists.debian.org
> 
> Well, silly me. I should have also asked this question as well: what
> exactly do I need to change in /etc/X11/XF86Config-4 to get the
> resolution to change? Also, is there a utility I could use to do that,
> not that I mind typing in vi to do the dirty work...
> 
> Thanks again!
> 
> Russell

Re: Debian Linux on PowerBook 1400cs -- eeep?

[Laurent de Segur]

If I were you, I would get more RAM for this machine (check auction sites)
and run MkLinux DR3 or RC1. I think that it is one of the only Linux ports
out there that really supports the PB 1400 old architecture.

Laurent



> From: Jørn Holm <[EMAIL PROTECTED]>
> Date: Fri, 21 Sep 2001 07:12:36 +0200
> To: debian-powerpc@lists.debian.org
> Subject: Debian Linux on PowerBook 1400cs -- eeep?
> Resent-From: debian-powerpc@lists.debian.org
> 
> Hi,
> I downloaded some of the 1.44-images for PowerPC debian, among others the
> boot-floppy-hfs.img-file...
> 
> I wrote it to a floppy using rawrite on a pc, and tested it on the
> powerbook... unfortunately, it appears to hang after two-three minutes of
> displaying that cute little logo with a mac and a penguin while reading a
> lot from the floppy drive...
> 
> i've only got 16mb ram, it appears to be the problem i've got when trying
> to install mklinux, but i didn't thing debian needed that much (i've
> installed debian on pc's with only 16mb, i believe?)
> 
> any help appreciated,
> 
> Jørn Holm
> [EMAIL PROTECTED]
> 
> 
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: Problem booting stable cd 1 on ibook2

[Laurent de Segur]

Henrik,

STFW.

Go to this excellent page put together by someone who's goal was to help
people who are trying to achieve what you want to do:

http://www.xiph.org/~jack/ibook/instnotes.html

Note that I got this link returned as the top choice by just entering
"ibook2 debian linux" using google.com. How easier do you need it to be?

In short debian ppc 2.2r3 doesn't boot from cd on the new iBook (dual USB).
All the gory details can be read on the web page above or other similar
links.

Laurent

- Original Message -
From: "Henrik Edlund" <[EMAIL PROTECTED]>
To: 
Sent: Friday, September 21, 2001 7:15 AM
Subject: Problem booting stable cd 1 on ibook2


> I am booting through OF with: boot cd:,\\:tbxi
>
> But it halts/freezes after printing "DO-QUIESCE finishedbooting...".
> Before that it says:
>
> ...ok
> copying OF device tree...done
> calling quiesce ...
>
> I then tried with "debian video=offb" at the boot: prompt but same result.
>
> Can anyone help me?
>
> --
> Henrik Edlund <[EMAIL PROTECTED]>
> http://www.edlund.org/
>
> "You're young, you're drunk, you're in bed, you have knives; shit
> happens." -- Angelina Jolie
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: imac kernel compilation

[Laurent de Segur]

You should recompile with CONFIG_MAC_ADBKEYCODES=n

If you run 2.2.19-pmac kernel, just append="keyboard_sends_linux_keycodes=1"
to /etc/yaboot.conf

Cheers,

Laurent

PS: check the mailing list archive, as this subject is a classic for the
past 2-3 weeks.


> From: Joe Paxton <[EMAIL PROTECTED]>
> Date: Sat, 22 Sep 2001 17:43:28 -0500
> To: debian-powerpc@lists.debian.org
> Subject: imac kernel compilation
> Resent-From: debian-powerpc@lists.debian.org
> 
> hello list,
> 
> i'm attempting to re-compile the kernel on my 2nd gen imac (lime,
> 266mHz)...but having little luck. i am using kernel-package from woody
> (though most of the software i'm running is from potato)...the major
> problem is that when i reboot after installing kernel, keyboard seems to
> be remapped in strange way. several sources have pointed me to new input
> layer...though i've had difficulty to this point comprehending exactly
> what it is that i need to do differently.
> 
> currently running kernel-2.2.18pre21...trying to update to
> kernel-2.2.19-2 (want to get a webcam, specifically logitech quickcam
> express - not even sure if that will work...but it's worth a try).
> 
> i would appreciate any recommendations, my .config for the kernel that
> i'm trying to compile is attached.
> 
> thanks,
> joe
> 
> #
> # Automatically generated by make menuconfig: don't edit
> #
> 
> #
> # Platform support
> #
> CONFIG_PPC=y
> CONFIG_6xx=y
> # CONFIG_8xx is not set
> CONFIG_PMAC=y
> # CONFIG_PREP is not set
> # CONFIG_CHRP is not set
> # CONFIG_ALL_PPC is not set
> # CONFIG_APUS is not set
> # CONFIG_GEMINI is not set
> # CONFIG_MBX is not set
> # CONFIG_SMP is not set
> # CONFIG_ALTIVEC is not set
> CONFIG_MACH_SPECIFIC=y
> CONFIG_POWERMAC=y
> 
> #
> # General setup
> #
> CONFIG_EXPERIMENTAL=y
> CONFIG_MODULES=y
> # CONFIG_MODVERSIONS is not set
> CONFIG_KMOD=y
> CONFIG_PCI=y
> # CONFIG_PCI_QUIRKS is not set
> CONFIG_PCI_OLD_PROC=y
> CONFIG_NET=y
> CONFIG_SYSCTL=y
> CONFIG_SYSVIPC=y
> # CONFIG_BSD_PROCESS_ACCT is not set
> CONFIG_BINFMT_ELF=y
> CONFIG_KERNEL_ELF=y
> CONFIG_BINFMT_MISC=y
> # CONFIG_BINFMT_JAVA is not set
> # CONFIG_PARPORT is not set
> # CONFIG_VGA_CONSOLE is not set
> CONFIG_FB=y
> # CONFIG_PMAC_PBOOK is not set
> CONFIG_PROC_DEVICETREE=y
> # CONFIG_TOTALMP is not set
> # CONFIG_BOOTX_TEXT is not set
> # CONFIG_MOTOROLA_HOTSWAP is not set
> 
> #
> # Plug and Play support
> #
> # CONFIG_PNP is not set
> 
> #
> # Block devices
> #
> # CONFIG_BLK_DEV_FD is not set
> CONFIG_BLK_DEV_IDE=y
> # CONFIG_BLK_DEV_HD_IDE is not set
> CONFIG_BLK_DEV_IDEDISK=y
> CONFIG_BLK_DEV_IDECD=y
> # CONFIG_BLK_DEV_IDETAPE is not set
> CONFIG_BLK_DEV_IDEFLOPPY=y
> CONFIG_BLK_DEV_IDESCSI=y
> # CONFIG_BLK_DEV_CMD640 is not set
> # CONFIG_BLK_DEV_RZ1000 is not set
> CONFIG_BLK_DEV_IDEPCI=y
> CONFIG_BLK_DEV_IDEDMA=y
> # CONFIG_BLK_DEV_OFFBOARD is not set
> CONFIG_IDEDMA_AUTO=y
> # CONFIG_BLK_DEV_OPTI621 is not set
> # CONFIG_BLK_DEV_ALI15X3 is not set
> # CONFIG_BLK_DEV_TRM290 is not set
> # CONFIG_BLK_DEV_NS87415 is not set
> # CONFIG_BLK_DEV_VIA82C586 is not set
> CONFIG_BLK_DEV_CMD646=y
> # CONFIG_BLK_DEV_CS5530 is not set
> # CONFIG_BLK_DEV_SL82C105 is not set
> CONFIG_BLK_DEV_IDE_PMAC=y
> CONFIG_BLK_DEV_IDEDMA_PMAC=y
> CONFIG_BLK_DEV_IDEDMA=y
> CONFIG_PMAC_IDEDMA_AUTO=y
> # CONFIG_IDE_CHIPSETS is not set
> CONFIG_BLK_DEV_LOOP=y
> # CONFIG_BLK_DEV_NBD is not set
> # CONFIG_BLK_DEV_MD is not set
> CONFIG_BLK_DEV_RAM=y
> CONFIG_BLK_DEV_RAM_SIZE=4096
> CONFIG_BLK_DEV_INITRD=y
> # CONFIG_BLK_DEV_XD is not set
> # CONFIG_BLK_DEV_DAC960 is not set
> CONFIG_PARIDE_PARPORT=y
> # CONFIG_PARIDE is not set
> # CONFIG_BLK_CPQ_DA is not set
> # CONFIG_BLK_CPQ_CISS_DA is not set
> # CONFIG_BLK_DEV_HD is not set
> 
> #
> # Networking options
> #
> CONFIG_PACKET=y
> CONFIG_NETLINK=y
> # CONFIG_RTNETLINK is not set
> # CONFIG_NETLINK_DEV is not set
> # CONFIG_FIREWALL is not set
> # CONFIG_FILTER is not set
> CONFIG_UNIX=y
> CONFIG_INET=y
> CONFIG_IP_MULTICAST=y
> # CONFIG_IP_ADVANCED_ROUTER is not set
> # CONFIG_IP_PNP is not set
> # CONFIG_IP_ROUTER is not set
> # CONFIG_NET_IPIP is not set
> # CONFIG_NET_IPGRE is not set
> # CONFIG_IP_MROUTE is not set
> CONFIG_IP_ALIAS=y
> # CONFIG_SYN_COOKIES is not set
> CONFIG_INET_RARP=y
> CONFIG_SKB_LARGE=y
> # CONFIG_IPV6 is not set
> # CONFIG_IPX is not set
> # CONFIG_ATALK is not set
> # CONFIG_X25 is not set
> # CONFIG_LAPB is not set
> # CONFIG_BRIDGE is not set
> # CONFIG_NET_DIVERT is not set
> # CONFIG_LLC is not set
> # CONFIG_ECONET is not set
> # CONFIG_WAN_ROUTER is not set
> # CONFIG_NET_FASTROUTE is not set
> # CONFIG_NET_HW_FLOWCONTROL is not set
> # CONFIG_CPU_IS_SLOW is not set
> 
> #
> # QoS and/or fair queueing
> #
> # CONFIG_NET_SCHED is not set
> 
> #
> # SCSI support
> #
> # CONFIG_SCSI is not set
> 
> #
> # Network device support
> #
> CONFIG_NETDEVICES=y
> 
> #
> # ARCnet devices
> #
> # CONFIG_ARCNET is not set
> # CONFIG_DUMMY is not set
> # CONFIG_BONDING is not set
> # CONFIG_EQUALIZER is not set
>

Problem compiling code with va_list assignment

[Laurent de Segur]

Hi,

I've been trying to understand the cause of the problem I ran into and
searched a great deal of time this weekend for a solution; I can't figure it
out yet.

Basically, I have a c++ class with a field variable declared as

Class Foo {
  protected:
va_list ap;

...

  public:
setVaArgs(va_list rap) {
  ap = rap;
}
}

I've got a setter method in this class to pass a va_list and store it into a
field var.

When I compile this code, I get the following error: incompatible types in
assignment of '__va_list_tag *' to '__va_list_tag[1]' at the line of the
assignment.

This doesn't seem to make sense, and I would appreciate if someone could
help me and point me in the right direction. Is there a special way for
handling va_list on the powerpc? I am using gcc 2.95.4.

Thanks in advance,

Laurent

Re: Problem compiling code with va_list assignment

[Laurent de Segur]

Thanks for the pointer. I was able to use a __va_copy(dst,src) instead of
the more direct approach dst = src.

However, I would like to comment on this problem. I am not certain that this
is related to the topics on this list, but this problem living on Linux PPC,
and my version of Linux installed being only Debian, I decided to post here.
Feel free to send me a pointer if you know of any.

The compiler error came to a surprise to me, because I used to compile the
code on Darwin PPC and never got this error reported. I went to look closer
at the gcc includes on Darwin and found that there is an extra section in
the va-ppc.h that has been added for AIX and APPLE OSs, perhaps to makes
things a little bit more compatible with other *nix flavors (i.e. va_list
ap, rap; ap = rap won't complain.) It seems that Apple in their version of
gcc given with Darwin, has added the patch for va-ppc.h (to be really
accurate, it seems more like they added an || defined (__APPLE__) after the
defined(_AIX) in this file, ending up to the same result.)

My question is: shouldn't the GCC compiler that comes with Linux (Debian)
PPC integrate this changes too? Running under APPLE, AIX, or LINUX on PPC,
shouldn't we all be using the same headers for one architecture? The ABI is
pervasive among all OSs running on the PowerPC (as far as using the GNU
tools anyway.)

IANAL, but it would seem that GCC is covered under LGPL in the first place,
and that should take care of licensing issues obviously associated with
Apple Darwin or AIX (I have my suspicion that the code started from the GNU
community originally anyway.)

Thanks for letting me know,


Laurent


> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Mon, 24 Sep 2001 12:04:55 +0200 (CEST)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Problem compiling code with va_list assignment
> Resent-From: debian-powerpc@lists.debian.org
> 
>> Basically, I have a c++ class with a field variable declared as
>> 
>> Class Foo {
>> protected:
>> va_list ap;
>> 
>> ...
>> 
>> public:
>> setVaArgs(va_list rap) {
>> ap = rap;
>> }
>> }
> 
> Please take a long look at va-ppc.h in the /usr/lib/gcc-lib/.../include/
> directory. That, together with a quick search of the list archives, should
> explain it.
> Hint: pay attention to the data type for va_list, and compare this with
> other archs. Also compare the defs for __va_copy.
> 
> Michael
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: 3 button emulations under X

[Laurent de Segur]

Hi David,

This option is available under the new input layer (2.2.18 and up on the mac
I believe.)

All I needed to do to get things to work was :

>echo "1" >/proc/sys/dev/mac_hid/mouse_button_emulation
>echo "97" >/proc/sys/dev/mac_hid/mouse_button2_keycode
>echo "100" >/proc/sys/dev/mac_hid/mouse_button3_keycode

and try it out. 

Note: keycode values may vary depending on the keyboard you are using. I own
an external extended keyboard. Here the button 2 corresponds to right option
key and button 3 corresponds to right control key.

To make things more permanent after I was happy with this configuration, I
added 3 lines to my /etc/sysctl.conf file :


dev.mac_hid.mouse_button_emulation=1
dev.mac_hid.mouse_button2_keycode=97
dev.mac_hid.mouse_button3_keycode=100

then :

>sysctl -p /etc/sysctl.conf

to make the changes effective.



One last word: do yourself a favor and buy a 3-button mouse. The real thing
is way easier to handle.

Cheers,


Laurent 

> From: David Howe <[EMAIL PROTECTED]>
> Date: Fri, 28 Sep 2001 21:45:37 +1000
> To: Linux PPC 
> Subject: 3 button emulations under X
> Resent-From: debian-powerpc@lists.debian.org
> 
> Hi Folks
> 
> I read somewhere that I can map my keyboard buttons to emulate buttons 2
> & 3 for my adb mouse.
> 
> Can someone help me out?
> 
> thanks in advance
> 
> david
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Kernel Panic exiting debug monitor under 2.4.10-ben0

[Laurent de Segur]

Just got a kernel panic by entered the debug console at login window with
cmd-power keys, after hitting x to exit. Got a sig 11 Kernel access
of bad area as a response. This is running 2.4.10-ben0 resynced as for
tonight. Beginning of BT available for the courageous ones. Ask me if
interested


Laurent

Packages kept back on woody

[Laurent de Segur]

Hi,

I was doing my usual weekend apt-get update/apt-get upgrade this morning on
testing and the upgrade stated that 3 packages (autoconf, lib-xt-java, and
mailtools) have been kept back.

As I haven't done absolutely nothing to these packages or their respective
installed files, I wonder why this is happening.

Thanks for your help,

Laurent

Re: Packages kept back on woody

[Laurent de Segur]

Michel,

That was my point which I hadn't make clear in the previous message: Doing
an apt-get install of the three individual packages goes without any
problems. Only when I did the apt-get upgrade, these packages would hold
back. All these three packages needed to install one new package but I can't
imagine that it is causing the problem (as other packages upgraded had some
associated packages in the past.) After doing an apt-get install of the
three separated kept back packages, I did a new apt-get upgrade and there is
no more problem. Everything shows as up-to-date. Why would I need to do
separate installs in the first place? I kept a log of the entire session
that shows the problem if you are interested to look at it.
Thanks,

Laurent

> From: Michel Dänzer <[EMAIL PROTECTED]>
> Organization: Debian, XFree86
> Date: Sat, 29 Sep 2001 19:54:31 +0200
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: debian-powerpc@lists.debian.org
> Subject: Re: Packages kept back on woody
> 
> Laurent de Segur wrote:
> 
>> I was doing my usual weekend apt-get update/apt-get upgrade this morning on
>> testing and the upgrade stated that 3 packages (autoconf, lib-xt-java, and
>> mailtools) have been kept back.
>> 
>> As I haven't done absolutely nothing to these packages or their respective
>> installed files, I wonder why this is happening.
> 
> apt-get install autoconf lib-xt-java mailtools
> 
> will show you the reasons.
> 
> 
> -- 
> Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
> XFree86 and DRI project member   /  CS student, Free Software enthusiast
> 

Back in time without even a warning

[Laurent de Segur]

Hi,

I just experienced my first battery replacement with my iBook 2002 running
Debian. While on the road and before running out juice, I swapped batteries.
Unfortunately, I discovered that there is no backup battery on the iBook
2001. So the clock got reset (I may have taken 30 secs to unpack the new
one.) 

What's curious is that when I rebooted, I didn't even get a message that the
clock was set earlier than when the system was shut down. Is this an option
that needs to be set and is off by default? It is not really hard to write
an init script that would do just that, but because of the serious
implications there is booting the machine with a reset clock, it would be
nice if this was a feature set by default system-wide. Right now, I need to
boot in single mode just to make sure the date is set correctly when ever I
swap batteries.


Laurent

Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks

[Laurent de Segur]

I just got a new iBook 2001 600MHz+COMBO-DVD/CDRW+256MB. It plays DVDs both
under macos9 and running debian using the vlc package without any problem at
all. The previous one had many troubles even playing simple MP3 due to 66MHz
bus speed mostly I believe. The increase in cpu speed is not phenomenal but
it can only helps.

Laurent

> From: Branden Robinson <[EMAIL PROTECTED]>
> Date: Tue, 23 Oct 2001 08:19:06 -0500
> To: debian-powerpc@lists.debian.org
> Subject: Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks
> Resent-From: debian-powerpc@lists.debian.org
> 
> On Tue, Oct 23, 2001 at 02:25:23PM +0200, David N. Welton wrote:
>> Branden Robinson <[EMAIL PROTECTED]> writes:
>> 
>>> 2) there is a combo DVD-ROM/CD-RW option available for it!
>>> (It completely blows my mind that this isn't an option on the TiBook.)
>> 
>> This is definitely something that pisses me off about my tibook.  I
>> could care less about DVD's, and would much rather have a CDRW.
> 
> Er, the new 667 MHz TiBook *is* available with a built-in CD-RW.
> 
> Just not the DVD-ROM+CD-RW combo that is an option on the iBook.
> 
> 
> 
> -- 
> G. Branden Robinson|
> Debian GNU/Linux   |   "Bother," said Pooh, as he was
> [EMAIL PROTECTED] |   assimilated by the Borg.
> http://people.debian.org/~branden/ |
> 

Trouble copying a large number of files to last ext2 partition

[Laurent de Segur]

Hi,

I run into this bizarre problem trying to copy a large amount of files to
the last partition on my disk. The disk is 10GB and the last partition
occupies the last 1GB of the disk (minus a few extra spare sectors.)

I did an fschk on the partition and no problem is reported.

I've got a compressed zip file with about 3 files I want to uncompress.
I know that the uncompressed size will end up filling the 1GB partition to
about 90%. 

If I unzip the file located on the same partition I get an error message
(can't create the file on device) for a few dozens files then the copy will
stop. At that point the disk is full at about 87%.

When I unzip the files on a previous partition, I don't have any problem.
When I then copy the files I just uncompressed from the previous partition
to the last partition, I get a message for the last few hundred files
stating that there is no space left on device. The space is 84% used and
there is about 161MB free.

If I persist doing this, I will eventually be able to unpack. The workaround
seems to be to cp -r from previous partition to the last which will not
succeed, then rm -rf the dir just created then unzipping to the last
partition directly. At that point, the drive is full at 90% which seems to
be correct.


My question is : is there a threshold on the last partition that would cause
the fs to report errors even if the partition is not completely full? Is
there a limit that the last partition should occupy?

If someone could give me a hint on how to track this problem, so I can
gather more info about this problem, that would be great.

Thanks for your help in advance,


Laurent

Re: Trouble copying a large number of files to last ext2 partition

[Laurent de Segur]

I use the partition for a bunch of cvs trees so this is a high probability.
Thanks for pointing this out. Is there a way to increase inode numbers on a
partitions (even without reformatting)? Is ext2 giving me less inodes than
ext3 on a partition? Should I switch to xfs or reiserfs for this partition
if I plan on storing lots of small files? Puzzled.

Thanks in advance for your help,

Laurent
- Original Message -
From: "Gordon Paynter" <[EMAIL PROTECTED]>
To: 
Cc: "Laurent de Segur" <[EMAIL PROTECTED]>
Sent: Thursday, October 25, 2001 9:59 AM
Subject: Re: Trouble copying a large number of files to last ext2 partition


> On Wednesday 24 October 2001 20:08, Laurent de Segur wrote:
> > I've got a compressed zip file with about 3 files I want to
> > uncompress. I know that the uncompressed size will end up filling the
> > 1GB partition to about 90%.
>
> A suggestion: maybe your drive has run out of inodes?
> (To find out: attempt the decompress, then run "df -i" on the drive.)
>
> Gordon
>

Re: Trouble copying a large number of files to last ext2 partition

[Laurent de Segur]

Andrew,

Thanks for your nice feedback. This is extremelly helpful.

Is it possible to change anything on the current partition without
reformatting it or am I just doomed to tarball all the content, move it
somewhere else (I can always zap my macos 9 partition I haven't touched for
months...) and then mkfs, then restore files? It would be nice if I can
change reserved space and goof with inode stuff without doing that though.

Thanks for your help,

Laurent
- Original Message -
From: "Andrew Sharp" <[EMAIL PROTECTED]>
To: 
Cc: 
Sent: Thursday, October 25, 2001 9:59 AM
Subject: Re: Trouble copying a large number of files to last ext2 partition


> Laurent de Segur wrote:
> >
> > Hi,
> >
> > I run into this bizarre problem trying to copy a large amount of files
to
> > the last partition on my disk. The disk is 10GB and the last partition
> > occupies the last 1GB of the disk (minus a few extra spare sectors.)
> >
> > I did an fschk on the partition and no problem is reported.
> >
> > I've got a compressed zip file with about 3 files I want to
uncompress.
> > I know that the uncompressed size will end up filling the 1GB partition
to
> > about 90%.
> >
> > If I unzip the file located on the same partition I get an error message
> > (can't create the file on device) for a few dozens files then the copy
will
> > stop. At that point the disk is full at about 87%.
>
> You can make the filesystem with a set percentage set aside for root
> so that the filesystem doesn't get all screwed up when it gets
> full-ish.  Actually getting completely full is a bad thing.  So set
> the percentage to 1 thusly:
>
> mkfs -m 1 /dev/
>
> Remember that other things consume space on the slice: inodes,
> directories, directory entries, block indirects, and so on.  So just
> 'cuz the slice is 1GB and the uncompressed size of the files is 90%
> of that is no guarantee that they will fit.  You could also mess
> with the number of bytes per inode to get the number of inodes to be
> very close to what you want, thus saving space by not having a large
> number of inodes unused.
>
> a
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: Trouble copying a large number of files to last ext2 partition

[Laurent de Segur]

Ok, currently tarballing and reading the man pages as we speak :-P

Thanks a lot for the tremendous help,

Laurent
- Original Message -
From: "Andrew Sharp" <[EMAIL PROTECTED]>
To: 
Sent: Thursday, October 25, 2001 1:09 PM
Subject: Re: Trouble copying a large number of files to last ext2 partition


>
> check out the tune2fs program, but you can't change the number of
> inodes with this program.  the reserved space can be, though.
>
> tar the files up and compress them "cd tar Icf /tmp/tarball.bz ."
>
> you can just put the tar file on the hfs partition, but you might be
> better off just sticking it somewhere safer while you re-mkfs the
> partition.  use the -v flag of mkfs to check how many inodes it is
> creating, etc.  you only need to store the tar file until you finish
> this step and then untar in onto the new file system.  you don't
> have to untar that file until then.  it all only takes a couple of
> minutes.
>
> read the man page for tune2fs (e2fsprogs package).
>
> a
>
>
> Laurent de Segur wrote:
> >
> > Andrew,
> >
> > Thanks for your nice feedback. This is extremelly helpful.
> >
> > Is it possible to change anything on the current partition without
> > reformatting it or am I just doomed to tarball all the content, move it
> > somewhere else (I can always zap my macos 9 partition I haven't touched
for
> > months...) and then mkfs, then restore files? It would be nice if I can
> > change reserved space and goof with inode stuff without doing that
though.
> >
> > Thanks for your help,
> >
> > Laurent
> > - Original Message -
> > From: "Andrew Sharp" <[EMAIL PROTECTED]>
> > To: 
> > Cc: 
> > Sent: Thursday, October 25, 2001 9:59 AM
> > Subject: Re: Trouble copying a large number of files to last ext2
partition
> >
> > > Laurent de Segur wrote:
> > > >
> > > > Hi,
> > > >
> > > > I run into this bizarre problem trying to copy a large amount of
files
> > to
> > > > the last partition on my disk. The disk is 10GB and the last
partition
> > > > occupies the last 1GB of the disk (minus a few extra spare sectors.)
> > > >
> > > > I did an fschk on the partition and no problem is reported.
> > > >
> > > > I've got a compressed zip file with about 3 files I want to
> > uncompress.
> > > > I know that the uncompressed size will end up filling the 1GB
partition
> > to
> > > > about 90%.
> > > >
> > > > If I unzip the file located on the same partition I get an error
message
> > > > (can't create the file on device) for a few dozens files then the
copy
> > will
> > > > stop. At that point the disk is full at about 87%.
> > >
> > > You can make the filesystem with a set percentage set aside for root
> > > so that the filesystem doesn't get all screwed up when it gets
> > > full-ish.  Actually getting completely full is a bad thing.  So set
> > > the percentage to 1 thusly:
> > >
> > > mkfs -m 1 /dev/
> > >
> > > Remember that other things consume space on the slice: inodes,
> > > directories, directory entries, block indirects, and so on.  So just
> > > 'cuz the slice is 1GB and the uncompressed size of the files is 90%
> > > of that is no guarantee that they will fit.  You could also mess
> > > with the number of bytes per inode to get the number of inodes to be
> > > very close to what you want, thus saving space by not having a large
> > > number of inodes unused.
> > >
> > > a
> > >
> > >
> > > --
> > > To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> > > with a subject of "unsubscribe". Trouble? Contact
> > [EMAIL PROTECTED]
> > >
> > >
> >
> > --
> > To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> > with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks

[Laurent de Segur]

You should show the example by including the decss code that now ships
standard with vlc into the debian package.

Right now, to be able to play commercial DVDs I need to get the decss module
from the vlc main site and install it on top of the stripped down version
that's available on debian servers.


LdS


> From: Branden Robinson <[EMAIL PROTECTED]>
> Date: Thu, 25 Oct 2001 23:30:01 -0500
> To: debian-powerpc@lists.debian.org
> Subject: Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks
> Resent-From: debian-powerpc@lists.debian.org
> 
> Fuck those Nazi bastards at Macrovision!  Stand up for your rights to
> view multimedia products you own on equipment you own!

Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks

[Laurent de Segur]

Andrew, 

ATI writes all their drivers for Windoz and the Mac internally. I guess
that's why it took so long for MacOS X to be able to play DVDs.
Cheers,

Laurent

> From: Andrew Sharp <[EMAIL PROTECTED]>
> Date: Fri, 26 Oct 2001 01:40:43 -0700
> To: debian-powerpc@lists.debian.org
> Subject: Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks
> Resent-From: debian-powerpc@lists.debian.org
> 
> Unfortunately, those aren't the only bastards skulking about.  Let's
> not get started about the MPAA or the RIAA.  But the documentation
> for the parts of the hardware from ATI had to be available to
> someone, because even Apple was able to write software for it.  OK,
> stop laughing, I was just kidding.  Actually, it is more likely that
> a library was provided by the company that designed the cell that
> ATI used, and ATI is contractually not able to release any
> documentation, and may not even have any documentation for
> programmers.  Another possibility is that a company wrote the
> software for ATI, and got it into the contract that they not release
> the documentation for how to program that part of the hardware.
> When you remember that the sellers of DVD playing software AND the
> manufacturers of DVD drives both pay a per unit fee to the MPAA, it
> is not hard to imagine the repressive contractual arrangements that
> may be in ATI's way.
> 
> a
> 
> 
> Laurent de Segur wrote:
>> 
>> You should show the example by including the decss code that now ships
>> standard with vlc into the debian package.
>> 
>> Right now, to be able to play commercial DVDs I need to get the decss module
>> from the vlc main site and install it on top of the stripped down version
>> that's available on debian servers.
>> 
>> LdS
>> 
>>> From: Branden Robinson <[EMAIL PROTECTED]>
>>> Date: Thu, 25 Oct 2001 23:30:01 -0500
>>> To: debian-powerpc@lists.debian.org
>>> Subject: Re: G3's, G4's, Altivec, DVD's, iBooks, and TiBooks
>>> Resent-From: debian-powerpc@lists.debian.org
>>> 
>>> Fuck those Nazi bastards at Macrovision!  Stand up for your rights to
>>> view multimedia products you own on equipment you own!
>> 
>> --
>> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
>> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

util-linux-locales and powerpc-utils installation conflict under woody

[Laurent de Segur]

Running an apt-get dist-upgrade as for tonight on woody reports the
following error:

.../...

Preparing to replace util-linux 2.11h-1 (using
.../util-linux_2.11l-3_powerpc.deb) ...
Unpacking replacement util-linux ...
dpkg: error processing
/var/cache/apt/archives/util-linux_2.11l-3_powerpc.deb (--unpack):
 trying to overwrite `/usr/share/man/man8/clock.8.gz', which is also in
package powerpc-utils
Errors were encountered while processing:
 /var/cache/apt/archives/util-linux_2.11l-3_powerpc.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)


So far the only I found to come around this is to remove the powerpc-utils.
This package won't install if the util-linux is already installed and vice
and versa. apt-get install -f will result in the same error.


Laurent

Re: util-linux upgrade problem

[Laurent de Segur]

See my other post, I wrote just a few mins while you were writing yours. I
removed the conflicting packages (util-linux-locales and powerpc-utils) then
ran apt-get dist-upgrade to install the rest then did apt-get install
util-linux-locales. I can live without powerpc-utils for a while.

Laurent

> From: Russell Hires <[EMAIL PROTECTED]>
> Reply-To: [EMAIL PROTECTED]
> Date: Wed, 31 Oct 2001 22:35:10 -0500
> To: debian-powerpc@lists.debian.org
> Subject: util-linux upgrade problem
> Resent-From: debian-powerpc@lists.debian.org
> 
> Hello everyone, 
> 
> I'm running woody on my PMac G3/266 and...
> I did my normal apt-get update && apt-get dist-upgrade and this is what I get:
> Preconfiguring packages ...
> (Reading database ... 71396 files and directories currently installed.)
> Preparing to replace util-linux 2.11h-1 (using
> .../util-linux_2.11l-3_powerpc.deb) ...
> Unpacking replacement util-linux ...
> dpkg: error processing /var/cache/apt/archives/util-linux_2.11l-3_powerpc.deb
> (--unpack):
> trying to overwrite `/usr/share/man/man8/clock.8.gz', which is also in
> package powerpc-utils
> Errors were encountered while processing:
> /var/cache/apt/archives/util-linux_2.11l-3_powerpc.deb
> E: Sub-process /usr/bin/dpkg returned an error code (1)
> 
> 
> What do I do to fix this?
> 
> Thanks!
> 
> Russell
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: Trackpad on the new TiBooks

[Laurent de Segur]

on 7/7/02 7:14 PM, Danial Pearce at [EMAIL PROTECTED] wrote:

> Gday people,
> 
> Stupid question, and not really crucial.
> 
> How do I get gnome to stop the mouse click emulation for the track pad?
> I mean how if you touch the mouse track pad, it emulates a click. In Mac
> OS X I can turn this on and off, I don't know how to turn it off in
> gnome.
> 
> cheers
> Danial.


trackpad notap

Type trackpad by itself to see the options available.


LdS


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Supported models

[Laurent de Segur]

If the real goal is finding out if the machine you bought can install and
run Debian, you should make that task really simple.

Something is the line of using a list of machines given by what they show
under the Apple System Profiler (running on the OS pre-installed on the
machine) associated with a status value (OK/NOT SUPPORTED/IN PROGRESS) would
be really neat.

For example, my desk machine shows:

Machine ID:   406  
Model Number: Power Mac G4 Cube

These two values are just enough to describe any particular model that Apple
manufactured as their software relies on them, and they are easily reachable
by any Mac user (ASP is pre-installed on both X and 9).

With these two values given in your list and a status associated with them,
the task of finding out if Debian can be installed becomes trivial. So far,
all the proposed schemes I have seen in this thread fail short to give a
unique name to multiple revisions of various models.

For those of you purists who don't even want to boot once under MacOS 9/X to
run ASP and look at these two values to find out if your machine is
supported, you probably don't even need the list anyway.


LdS


> From: Matt Brubeck <[EMAIL PROTECTED]>
> Date: Fri, 2 Nov 2001 21:01:14 -0800 (PST)
> To: Chris Tillman <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Supported models
> Resent-From: debian-powerpc@lists.debian.org
> 
> "Slot-loading" alone is too ambiguous, since the next several generations
> of iMacs will probably fit that description.  One can imagine a user
> purchasing a new iMac, seeing "iMac (slot loading)" on a compatible list,
> and thiking that his model will work when in fact it doesn't.  You and I
> know that it is Apple jargon for a particular model, but most users won't.
> 
> If we want to be really explicit we could expand the iMac models as
> follows.  This listing should make it easy for owners to pin down their
> exact models.  On the other hand it is a bit overwhelming, and perhaps a
> less-exhaustive compromise is better.
> 
>> iMac Bondi Blue, 5 Flavors, Slot Loadingpowermac-NewWorld
>> iMac Summer 2000, Early 2001powermac-NewWorld
> 
> iMac 233   "rev A/B" (bondi blue)
> iMac 266/333   "rev C/D" (five flavors)
> 
> iMac 350   "Slot-Loading" (blueberry)
> iMac DV 400"Slot-Loading" (five flavors)
> iMac DV SE 400 "Slot-Loading" (graphite)
> 
> iMac 350   "Summer 2000" (indigo)
> iMac DV 400"Summer 2000" (indigo/ruby)
> iMac DV+ 450   "Summer 2000" (indigo/ruby/sage)
> iMac DV SE 500 "Summer 2000" (graphite/snow)
> 
> iMac 400"Early 2001" (indigo)
> iMac 500"Early 2001" (flower power/blue dalmation)
> iMac SE 600 "Early 2001" (flower/dalmation/graphite)
> 
> iMac 500"Summer 2001" (indigo)
> iMac 600"Summer 2001" (graphite)
> iMac SE 700 "Summer 2001" (graphite/snow)
> 
> Similar details might be useful for the PowerBooks and iBooks.  Note that
> at least including a list of processor speeds greatly reduces the chance
> of unforseen ambiguity.
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: headless G4 Cube with serial console

[Laurent de Segur]

Eric,

Thanks for the info. A small, quite, and cute server for the living room is
how I imagined Apple would market this thing when I heard about it thru the
rumors sites. Good move! I'm heading to eBay as we speak...


LdS

> From: "Eric C. Cooper" <[EMAIL PROTECTED]>
> Date: Sat, 3 Nov 2001 12:17:17 -0500
> To: debian-powerpc@lists.debian.org
> Subject: headless G4 Cube with serial console
> Resent-From: debian-powerpc@lists.debian.org
> 
> It's a great server -- small, quiet, and very stylish.
> 
> -- 
> Eric C. Cooper  e c c @ c m u . e d u

missing package on http debian site

[Laurent de Segur]

A fresh and failing dist-upgrade as this morning (Sun 11/4 11am) reveals
that the package: 

http://http.us.debian.org/debian/pool/main/g/gqview/gqview_0.12.0-1_powerpc.
deb 

is missing. Please advise.

TIA,

LdS

Re: missing package on http debian site

[Laurent de Segur]

Thanks Michel. It came out after a dozen times. Note to myself: I need to be
more patient next time...

Another heads up: 

During the woody dist-upgrade, the Defoma package 0.4.12 config script
spills the same error (Use of unitialized value in string eq at
/usr/share/perl5/Debian/Defoma/IdCache.pm line 303) thousands of times.
Doing the reconfigure works fine though.

LdS

> From: Michel Dänzer <[EMAIL PROTECTED]>
> Date: 04 Nov 2001 20:04:14 +0100
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: debian-powerpc@lists.debian.org
> Subject: Re: missing package on http debian site
> 
> On Sun, 2001-11-04 at 20:01, Laurent de Segur wrote:
>> A fresh and failing dist-upgrade as this morning (Sun 11/4 11am) reveals
>> that the package:
>> 
>> http://http.us.debian.org/debian/pool/main/g/gqview/gqview_0.12.0-1_powerpc.
>> deb 
>> 
>> is missing. Please advise.
> 
> Try again, http.us.debian.org is several machines which tend to be out
> of sync. If you get tired of trying again, use a better mirror. :)
> 
> 
> -- 
> Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
> XFree86 and DRI project member   /  CS student, Free Software enthusiast
> 

powerpc-utils: still in limbo

[Laurent de Segur]

As for last night, trying to install testing or unstable powerpc-utils
package on top of a woody freshly updated system still shows the same
conflict error reported for about 10 days. Ethan's workaround of using sid
(unstable) doesn't work and causes the same problem (conflict happens with
'/usr/share/man/man8/clock.8.gz'.)

Because of this error, doing a dselect upgrade to woody on top of a potato
install is just a plain nightmare.


LdS


 

XFree 86 color palette screwed up.

[Laurent de Segur]

The machine is an iBook 2001 latest rev (600/100). Thanks to Apple marketing
for making this confusing. The video card in the machine is identical to the
previous model: a R128 M2 rev 2.

This problem came after a fresh install of Woody. XFree hasn't run on the
system yet. I am in ofonly mode installing all the bits and pieces. I set up
for the first time the XF86Config file to use 16 bit depth, and restart the
machine (I know the XF86 config file works because it's bitwise identical to
the one working on another iBook.)

At restart, I change my yaboot config to use the boot settings with the
video r128 card. When XFree86 comes up, I get a screen in inverted colors
(not sure if inverted, but really mixed up anyway, like lots of greens and
blue.) It looks exactly like the hw palette on the card hasn't been
initialized except that I don't know if there is a palette in true color.

I reboot and it does the same. Now, if I change XF86Config to use 24 bits
depth by default, and restart the X server (not the machine), then no
problem, colors are ok. I change to 16 bit depth again in XF86Config and
then restart using the r128 video mode and then it's working now. No more
inverted colors. 

To recap, It only happens when I boot first in ofonly video mode then
restart the system using the 16 bit depth default mode. Kinda easy to
reproduce.


Has anyone seen that problem yet?

LdS

Re: XFree 86 color palette screwed up.

[Laurent de Segur]

I'll try that tonight. I've never seen a XF86Config file with 15 used in the
bit depth though...

LdS


- Original Message -
From: "Michael Schmitz" <[EMAIL PROTECTED]>
To: "Laurent de Segur" <[EMAIL PROTECTED]>
Cc: 
Sent: Thursday, November 08, 2001 11:25 AM
Subject: Re: XFree 86 color palette screwed up.


> > This problem came after a fresh install of Woody. XFree hasn't run on
the
> > system yet. I am in ofonly mode installing all the bits and pieces. I
set up
> > for the first time the XF86Config file to use 16 bit depth, and restart
the
> > machine (I know the XF86 config file works because it's bitwise
identical to
> > the one working on another iBook.)
>
> How about 15 bit depth (16 bits per pixel framebuffer data format)?
>
> Michael
>
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: powerpc-utils: still in limbo

[Laurent de Segur]

Ron-

That's tempting, and I may do that if the problem still occur in a couple of
days. I just wanted to let people on these list who pointed me to sid for a
solution that it still didn't simply solved this conflict problem.
Thanks for your reply,

LdS
- Original Message -
From: "Ron Golan" <[EMAIL PROTECTED]>
To: 
Sent: Thursday, November 08, 2001 1:28 PM
Subject: Re: powerpc-utils: still in limbo


> On Thu, Nov 08, 2001 at 09:58:27AM -0800, Laurent de Segur wrote:
> > As for last night, trying to install testing or unstable powerpc-utils
> > package on top of a woody freshly updated system still shows the same
> > conflict error reported for about 10 days. Ethan's workaround of using
sid
> > (unstable) doesn't work and causes the same problem (conflict happens
with
> > '/usr/share/man/man8/clock.8.gz'.)
> >
>
> I can't say its the best solution but what I did was:
>
> cd /var/cache/apt/archives
> dpkg -i --force-overwrite ./powerpc-utils_1.1.3-5_powerpc.deb
>
>
> --
> Ron Golan
> [EMAIL PROTECTED]
> GPG key http://people.we.mediaone.net/rgolan/gpg.asc
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: XFree 86 color palette screwed up.

[Laurent de Segur]

Comments below...

- Original Message -
From: "Colin Walters" <[EMAIL PROTECTED]>
To: 
Sent: Thursday, November 08, 2001 11:08 AM
Subject: Re: XFree 86 color palette screwed up.


> Laurent de Segur <[EMAIL PROTECTED]> writes:
>
> > This problem came after a fresh install of Woody. XFree hasn't run
> > on the system yet. I am in ofonly mode installing all the bits and
> > pieces. I set up for the first time the XF86Config file to use 16
> > bit depth, and restart the machine (I know the XF86 config file
> > works because it's bitwise identical to the one working on another
> > iBook.)
>
> I don't think XFree is going to be happy when you've passed
> video=ofonly.

I know that... However, when I do a fresh install, I am in console mode and
Ineed to start somewhere anyway, right? So ofonly is the mode I used to do
the installation in console mode AT FIRST. The problem appears when I
restart in r128 mode after having installed all the packages as I stated in
my original message.

(FYI: when trying a start x in ofonly mode, it says that the screen can't be
found, which kinda makes senses. But you probably know that already...)

>
> > At restart, I change my yaboot config to use the boot settings with
> > the video r128 card. When XFree86 comes up, I get a screen in
> > inverted colors (not sure if inverted, but really mixed up anyway,
> > like lots of greens and blue.) It looks exactly like the hw palette
> > on the card hasn't been initialized except that I don't know if
> > there is a palette in true color.
>
> What happens if you boot with no video= argument at all?  The default
> on a recent kernel (this means 2.4.x at least) should be fine I
> believe.  You are running 2.4.x, right?

I am running 2.4.12 (not much choice on new HW, although I ran 2.2.19 for a
while without trouble but no extra stuff :-( , and whatever is in debian
testing as far as XFree86 (4.1.09?).

Tonight, I'll try passing nothing in the args and see what happens as well
as using 15 vs. 16 as Michel suggested in a previous reply..

Again, the procedure to reproduce: all I need to do is edit my XF86Config-4
to run in 16 bit depth, switch my run level to 2 (to avoid x to start),
restart in ofonly mode (that's ok because x won't start), then when I am at
the console login, log in, and restart the system with r128 fb video mode
passed as kernel args, log in again, and finally do a startx (or make sure
that I changed back my run level back to 4 to run xfree automagically.) This
will show the inverted palette(or whatever is the real bug) problem all the
time. And this happens only with 16. Pretty basic, uh?

LdS
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: XFree 86 color palette screwed up.

[Laurent de Segur]

Michel,

You are right. I compiled benh's kernel yesterday and I was unsuccessful
trying to reproduce the problem I reported earlier with the new kernel. Is
there an ETA for getting the patch into the woody train?

LdS

> From: Michel Dänzer <[EMAIL PROTECTED]>
> Date: 09 Nov 2001 17:33:04 +0100
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: debian-powerpc@lists.debian.org
> Subject: Re: XFree 86 color palette screwed up.
> Resent-From: debian-powerpc@lists.debian.org
> 
> On Thu, 2001-11-08 at 19:16, Laurent de Segur wrote:
> 
>> To recap, It only happens when I boot first in ofonly video mode then
>> restart the system using the 16 bit depth default mode. Kinda easy to
>> reproduce.
> 
> aty128fb used to be broken in depth 16, the fix is only in benh's tree
> yet AFAIK. If you're using that, maybe it's still broken for the rather
> complicated procedure you're using.
> 
> 
> -- 
> Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
> XFree86 and DRI project member   /  CS student, Free Software enthusiast
> 
> 
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Privilege Question

[Laurent de Segur]

Hi,

I was wondering what type of privileges I need to give to /dev/hdb (cdrom
device on my machine) to be able to access and eject the media.

The current mode has been set at install time by default as: 'brw-rw'
and the file is owned by 'root disk' causing a program with open("/dev/hdb",
...) or the command 'eject /cdrom' to fail executing.

Running the code or command as root will work of course, but I don't really
like to do that. I was wondering what other debian ppc users have in their
settings. Thanks for your help,

LdS

Problem cat'ing in terminal: buffer overflow?

[Laurent de Segur]

Hi,

I recently noticed that cat'ing a bin file in the terminal will render this
terminal unusable. Here is a short reproducible test case :

$cat /sbin/mkfs.ext2

Garbage is populating the output (which is ok in case of non-ascii chars),
but when it's done, the label before the prompt sign and everything typed is
just garbage too. I run 2.4.15 but I've seen this on prior 2.4.x builts.


LdS

ext3 anyone?

[Laurent de Segur]

I noticed a message on linuxtoday.com regarding ext3 being default in the
new 2.4.15-pre2 release. I built ben's tree which is, it seems, in sync with
this release. I was curious to know if some people on this list have been
switching their partitions to ext3 on ppc and how things are going so far. I
remember asking this about reiserfs and found the answers pretty valuable to
me and my precious data. Is it as easy as advertised to switch from ext2 to
ext3 and are all the tools available on ppc to do the migration?
Thanks to anyone for the pioneering adoption.

Cheers,

LdS

Re: Problem cat'ing in terminal: buffer overflow?

[Laurent de Segur]

No, no, reset works. What I don't understand is the corruption happening
before the prompt sign and the typed chars getting screwed up after the dump
is done. It seems that outputting non-ascii chars just overflow some
buffers. Do you think it's ok? Curious,

LdS


> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Sat, 10 Nov 2001 20:38:45 +0100 (CET)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Problem cat'ing in terminal: buffer overflow?
> 
>> I recently noticed that cat'ing a bin file in the terminal will render
> 
> You mean you cannot type 'reset' and get the shell back anymore? That
> would be bad. Otherwise, what did you expect?
> 
> Michael
> 

Re: Problem cat'ing in terminal: buffer overflow?

[Laurent de Segur]

Michael,

Thanks for the informative reply. I wasn't flaming, just interested about
the reason behind this behavior. My bad if I didn't figure it out. All you
had to do is spell VT100 really.
Thanks again,

LdS

> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Sat, 10 Nov 2001 20:56:17 +0100 (CET)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: Michael Schmitz <[EMAIL PROTECTED]>,
> debian-powerpc@lists.debian.org
> Subject: Re: Problem cat'ing in terminal: buffer overflow?
> Resent-From: debian-powerpc@lists.debian.org
> 
>> No, no, reset works. What I don't understand is the corruption happening
>> before the prompt sign and the typed chars getting screwed up after the dump
>> is done. It seems that outputting non-ascii chars just overflow some
>> buffers. Do you think it's ok? Curious,
> 
> I think that's what you rightly deserve for pumping random garbage to the
> screen. How the hell do you think scrolling regions, cursor position
> movement and character set switching works? Am I one of the very few that
> remember what 'vt100' actually means, or even used such a beast? Jeez...
> Read up on terminals, terminal emulation and related topics before
> someone comes throwing the Orange Wall (tm) at you, please.
> 
> Your random non-ASCII bytes instructed the terminal to have the cursor
> jump around at random, switch to an alternate character set, output some
> characters, jump around some more, scribble more garbage on the screen,
> etc. etc. ... Just what you asked for, really.
> 
> There's no buffer overflow or any other problem to be seen here. Nothing
> really happened. Move along.
> 
> Michael
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

Re: Problem cat'ing in terminal: buffer overflow?

[Laurent de Segur]

What's your point?

> From: Ethan Benson <[EMAIL PROTECTED]>
> Date: Sat, 10 Nov 2001 19:18:30 -0900
> To: debian-powerpc@lists.debian.org
> Subject: Re: Problem cat'ing in terminal: buffer overflow?
> Resent-From: debian-powerpc@lists.debian.org
> 
> On Sat, Nov 10, 2001 at 11:33:08AM -0800, Laurent de Segur wrote:
>> Hi,
>> 
>> I recently noticed that cat'ing a bin file in the terminal will render this
>> terminal unusable. Here is a short reproducible test case :
>> 
>> $cat /sbin/mkfs.ext2
>> 
>> Garbage is populating the output (which is ok in case of non-ascii chars),
>> but when it's done, the label before the prompt sign and everything typed is
>> just garbage too. I run 2.4.15 but I've seen this on prior 2.4.x builts.
> 
> you have obviously never used a unix system before.
> 
> -- 
> Ethan Benson
> http://www.alaska.net/~erbenson/
> 

Re: Problem cat'ing in terminal: buffer overflow?

[Laurent de Segur]

Fuck off Ethan. OK?

> From: Ethan Benson <[EMAIL PROTECTED]>
> Date: Sat, 10 Nov 2001 21:19:40 -0900
> To: debian-powerpc@lists.debian.org
> Subject: Re: Problem cat'ing in terminal: buffer overflow?
> Resent-From: debian-powerpc@lists.debian.org
> Resent-Date: Sat, 10 Nov 2001 22:19:55 -0800 (PST)
> 
> 
> what part of:
> 
> Mail-Followup-To: Debian-powerpc 
> Mail-Copies-To: nobody
> X-No-CC: I subscribe to this list; do not CC me on replies.
> 
> *don't* you understand?
> 
> On Sat, Nov 10, 2001 at 10:07:11PM -0800, Laurent de Segur wrote:
>> What's your point?
> 
> that you don't know what your talking about.
> 
>>> From: Ethan Benson <[EMAIL PROTECTED]>
>>> Date: Sat, 10 Nov 2001 19:18:30 -0900
>>> To: debian-powerpc@lists.debian.org
>>> Subject: Re: Problem cat'ing in terminal: buffer overflow?
>>> Resent-From: debian-powerpc@lists.debian.org
>>> 
>>> On Sat, Nov 10, 2001 at 11:33:08AM -0800, Laurent de Segur wrote:
>>>> Hi,
>>>> 
>>>> I recently noticed that cat'ing a bin file in the terminal will render this
>>>> terminal unusable. Here is a short reproducible test case :
>>>> 
>>>> $cat /sbin/mkfs.ext2
>>>> 
>>>> Garbage is populating the output (which is ok in case of non-ascii chars),
>>>> but when it's done, the label before the prompt sign and everything typed
>>>> is
>>>> just garbage too. I run 2.4.15 but I've seen this on prior 2.4.x builts.
>>> 
>>> you have obviously never used a unix system before.
>>> 
>>> -- 
>>> Ethan Benson
>>> http://www.alaska.net/~erbenson/
>>> 
>> 
> 
> -- 
> Ethan Benson
> http://www.alaska.net/~erbenson/
> 

Kernel Panic with mac-fdisk: It's reproducible...

[Laurent de Segur]

Hi,

This kernel panic happens on fast hardware just after the second sync()
following the ioctl() as Ethan wrote previously. I can reproduce the kernel
panic on two of my QuickSilver G4 dual 800 most of the time with vanilla
mac-fdisk. 

I wrote a little test case that reproduces this kernel panic most of the
time. the code doesn't need to write. Just read the partition map, sync +
reread it again = crash :-( Very basic.

On my slower machines (iBook, TiBook and my G4 Cube), it happens either
sometimes or all the time by lowering the sleep time before re-reading the
partition map. So as machine speed increase, it's doomed to happen more
often.

The backtrace for your reading pleasure:

vector:0 at pc=c0039d14, lr = c0039d14
msr = 9032, sp cbdd1db0 [cbdd1cf8]
current cbdd, pid=1222, comm = readmap

This is under the latest kernel from benh's tree (2.4.15-pre2 as for
tonight.)

Feel free to ask me more info if you need additional details.


LdS


> From: Ethan Benson <[EMAIL PROTECTED]>
> Date: Sun, 11 Nov 2001 17:12:56 -0900
> To: debian-powerpc@lists.debian.org
> Subject: Re: installation success using 3.0.16 on iBook Dual USB / HOWTO
> Resent-From: debian-powerpc@lists.debian.org
> 
> On Sun, Nov 11, 2001 at 09:37:26AM -0700, Tom Rini wrote:
>> 
>> So what happens exactly?  Can you reproduce this on something and decode
>> the oops?  Looking in 2.4.15-pre2 I don't see the mac partition table
>> stuffs not doing something the msdos ones do..
> 
> enter the `w' command to commit a new partition table
> 
> mac-fdisk writes out the new one, synced the disk
> 
> mac-fdisk calls the reread ioctl
> 
> kernel panics (sometimes)
> 
> i cannot reproduce this sorry, the only time ive seen it personally
> was when partitioning a silver G4 i was setting up for someone, and i
> obviously cannot and will not destroy the installed debian system
> there in an attempt to reproduce this.
> 
> one other note that just occured to me, i could only reproduce the
> problem on that G4 *once* and that one time was when i obliterated the
> crufty partition table containing all the macos crap, all subsequent
> partition tables were free of any driver cruft, only pure linux
> partitions and an Apple_Bootstrap.  when setting it up i intentionally
> tried writing a few new tables just to see, could never reproduce it
> again. 
> 
> perhaps all that cruft has something to do with it? though i don't see how...
> 
> -- 
> Ethan Benson
> http://www.alaska.net/~erbenson/
> 

Re: Kernel Panic with mac-fdisk: It's reproducible...

[Laurent de Segur]

> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Mon, 12 Nov 2001 10:54:36 +0100 (CET)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Kernel Panic with mac-fdisk: It's reproducible...
> 
>> I wrote a little test case that reproduces this kernel panic most of the
>> time. the code doesn't need to write. Just read the partition map, sync +
>> reread it again = crash :-( Very basic.
> 

> Now does this happen on other architectures as well (with Mac partitions),
> or other partition table formats?
>
I'll try to check on some x86 early this week. I don't have an ETA for that.
I'll do my best. My server farm and *tops are mostly Macs (running Debian.)

>> On my slower machines (iBook, TiBook and my G4 Cube), it happens either
>> sometimes or all the time by lowering the sleep time before re-reading the
>> partition map. So as machine speed increase, it's doomed to happen more
>> often.
> 
> What time to usleep() seems enough to prevent this from happening? I'll
> add a short sleep in mac-fdisk as a workaround while this is being fixed
> on the kernel side.
>
The sleeps are already in write_partition_map() (partition.c if memory
serves well), one of 2, the other of 4. I doubled these values prior to
recompiling pdisk last night and this workaround seemed ok for the faster
cpu machines. This is strange as I thought that the sleep time was
independent of cpu speed, and the last task should finish earlier on faster
machines. Weird.

Of course, I can not guarantee that it won't crash anymore. It just doesn't
easily anymore. BTW, you may want to lower these values to reproduce easily
on your system.
 
>> vector:0 at pc=c0039d14, lr = c0039d14
>> msr = 9032, sp cbdd1db0 [cbdd1cf8]
>> current cbdd, pid=1222, comm = readmap
> 
> That's mostly chinese for me without being fed through some ksymoops like
> tool (or your System.map; what's near c0039d14 in System.map?)
>
Michael, that's mostly Chinese for me even with the tool ;-)

Here are the two address in my System.map braketing the lr:

c0039c38 T invalidate_bdev
c0039db8 T __invalidate_buffers

Let me know if you need anything.
Thanks a lot,

LdS
 


> Michael
> 
> 

Re: Kernel Panic with mac-fdisk: It's reproducible...

[Laurent de Segur]

Michael,

Oki doki. I'll try to do all of that as soon as I get back tonight.

Heigh-Ho, Heigh-Ho, It's work from home I go.

LdS

> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Mon, 12 Nov 2001 15:27:29 +0100 (CET)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Kernel Panic with mac-fdisk: It's reproducible...
> 
>>> Now does this happen on other architectures as well (with Mac partitions),
>>> or other partition table formats?
>>> 
>> I'll try to check on some x86 early this week. I don't have an ETA for that.
>> I'll do my best. My server farm and *tops are mostly Macs (running Debian.)
> 
> Thanks; I'll wait for that.
> 
>>> What time to usleep() seems enough to prevent this from happening? I'll
>>> add a short sleep in mac-fdisk as a workaround while this is being fixed
>>> on the kernel side.
>>> 
>> The sleeps are already in write_partition_map() (partition.c if memory
>> serves well), one of 2, the other of 4. I doubled these values prior to
>> recompiling pdisk last night and this workaround seemed ok for the faster
>> cpu machines. This is strange as I thought that the sleep time was
>> independent of cpu speed, and the last task should finish earlier on faster
>> machines. Weird.
> 
> The sleep should be CPU speed independent indeed. But it shouldn't ever be
> necessary. Neither should the first sync (BLKRRPART invalidates the device
> thereby flushing all data to disk first). I really wonder what these
> sleep()s are meant for.
> 
> Please try another thing: move the close_device() after the second
> sync/sleep pair. Or double each sync(). If we go the voodo path, we may as
> well go all the way ;-)
> 
>> Of course, I can not guarantee that it won't crash anymore. It just doesn't
>> easily anymore. BTW, you may want to lower these values to reproduce easily
>> on your system.
> 
> I'll take them out for starters.
> 
>>>> current cbdd, pid=1222, comm = readmap
>>> 
>>> That's mostly chinese for me without being fed through some ksymoops like
>>> tool (or your System.map; what's near c0039d14 in System.map?)
>>> 
>> Michael, that's mostly Chinese for me even with the tool ;-)
>> 
>> Here are the two address in my System.map braketing the lr:
>> 
>> c0039c38 T invalidate_bdev
> 
> That's a good clue. Now I need you to "objdump -d
> --start-address=0xc0039c38 vmlinux " your kernel image and post the
> section up to and a few lines beyond the PC value.
> 
> Michael
> 
> 

Re: Kernel Panic with mac-fdisk: More Info

[Laurent de Segur]

e4 bnec0039c58 
c0039d78:37 bd ff ff addic.r29,r29,-1
c0039d7c:7f 9f e3 78 mrr31,r28
c0039d80:41 81 ff 08 bgtc0039c88 
c0039d84:3a f7 00 01 addir23,r23,1
c0039d88:2c 17 00 02 cmpwir23,2
c0039d8c:3b 7b 00 04 addir27,r27,4
c0039d90:40 81 fe dc blec0039c6c 
c0039d94:2c 1a 00 00 cmpwir26,0
c0039d98:40 82 fe c0 bnec0039c58 
c0039d9c:80 76 00 0c lwzr3,12(r22)
c0039da0:4b fe dc 69 blc0027a08 
c0039da4:80 01 00 44 lwzr0,68(r1)
c0039da8:7c 08 03 a6 mtlrr0
c0039dac:ba 81 00 10 lmwr20,16(r1)
c0039db0:38 21 00 40 addir1,r1,64
c0039db4:4e 80 00 20 blr

c0039db8 <__invalidate_buffers>:





> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Mon, 12 Nov 2001 15:27:29 +0100 (CET)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Kernel Panic with mac-fdisk: It's reproducible...
> 
>>> Now does this happen on other architectures as well (with Mac partitions),
>>> or other partition table formats?
>>> 
>> I'll try to check on some x86 early this week. I don't have an ETA for that.
>> I'll do my best. My server farm and *tops are mostly Macs (running Debian.)
> 
> Thanks; I'll wait for that.
> 
>>> What time to usleep() seems enough to prevent this from happening? I'll
>>> add a short sleep in mac-fdisk as a workaround while this is being fixed
>>> on the kernel side.
>>> 
>> The sleeps are already in write_partition_map() (partition.c if memory
>> serves well), one of 2, the other of 4. I doubled these values prior to
>> recompiling pdisk last night and this workaround seemed ok for the faster
>> cpu machines. This is strange as I thought that the sleep time was
>> independent of cpu speed, and the last task should finish earlier on faster
>> machines. Weird.
> 
> The sleep should be CPU speed independent indeed. But it shouldn't ever be
> necessary. Neither should the first sync (BLKRRPART invalidates the device
> thereby flushing all data to disk first). I really wonder what these
> sleep()s are meant for.
> 
> Please try another thing: move the close_device() after the second
> sync/sleep pair. Or double each sync(). If we go the voodo path, we may as
> well go all the way ;-)
> 
>> Of course, I can not guarantee that it won't crash anymore. It just doesn't
>> easily anymore. BTW, you may want to lower these values to reproduce easily
>> on your system.
> 
> I'll take them out for starters.
> 
>>>> current cbdd, pid=1222, comm = readmap
>>> 
>>> That's mostly chinese for me without being fed through some ksymoops like
>>> tool (or your System.map; what's near c0039d14 in System.map?)
>>> 
>> Michael, that's mostly Chinese for me even with the tool ;-)
>> 
>> Here are the two address in my System.map braketing the lr:
>> 
>> c0039c38 T invalidate_bdev
> 
> That's a good clue. Now I need you to "objdump -d
> --start-address=0xc0039c38 vmlinux " your kernel image and post the
> section up to and a few lines beyond the PC value.
> 
> Michael
> 
> 

Re: Font advice needed

[Laurent de Segur]

Rory,

Here is the link to help you:

http://www.macdisk.com/fontsen.php3

The tool is called TT Converter and let's you convert a mac resource file to
a true type file that can be used in Linux.

I added a line to my XFree86 config so my apps to find the mac fonts I added
in a sub-dir in my ~ dir. It's easier to re-install this way without doing
the conversion again.


LdS


> From: "Albert D. Cahalan" <[EMAIL PROTECTED]>
> Date: Tue, 13 Nov 2001 01:16:28 -0500 (EST)
> To: [EMAIL PROTECTED] (Rory Campbell-Lange)
> Cc: debian-powerpc@lists.debian.org
> Subject: Re: Font advice needed
> Resent-From: debian-powerpc@lists.debian.org
> Resent-Date: Mon, 12 Nov 2001 22:16:52 -0800 (PST)
> 
> Sure, 9pt Geneva (writing) or 9pt Monaco (coding). You have MacOS,
> so you have these fonts. There are tools to create the index files
> needed by XFree86 and other font users. Then add the directory to
> your XFree86 config file.
> 
> It's been a while since I did this, so don't ask for specifics.

Re: Quake and Quake 2 on PowerPC (Might be of topic)

[Laurent de Segur]

> From: "Patrick R. Klee" <[EMAIL PROTECTED]>
> Date: 13 Nov 2001 00:12:15 -0600
> To: debian-powerpc@lists.debian.org
> Subject: Quake and Quake 2 on PowerPC (Might be of topic)
> Resent-From: debian-powerpc@lists.debian.org
> 
> Anywho.  You guys are great.  After I get Quake working the laststep
> will be mastering compiling and installing a kernel, and then I will be
> a Linux guru :)


Hey! Aren't you missing a few steps to get to the top of the ladder?

I hope that I didn't spoil your euphoria :-)


LdS

Re: Kernel Panic with mac-fdisk: More Info

[Laurent de Segur]

Michael,

Forget about my non-sense with the x86 crap. R3 and R4 are pointing to valid
strings. I forgot about offsetting values by one when reading addis/addi
opcodes as I was getting tired. I just needed a good night sleep.

I've increased substantially the number of sync() (a dozen with small loops
in betweens.) Same crash. So, my guess is: sync() won't solve it.
Independently from this, it came to a point last night where I couldn't
reproduce this. I had to reinstall macos and then rerun mac-fdisk to get to
the bug again. Now I get it again. Painful experience. This makes me believe
that it must be some timing issue.

Now that I know what the code looks like (thanks to you), I will give you
the extra info as soon as I run it again (this evening.) Sorry about the
delay.

One thing I would like to know: can I drop into the 'mon' and set a break
point in the kernel code, then restart and trace when I hit the break point?
Does kernel trace works on PPC? I'll stfw for these info, but if you can
give me a quick pointer, I'll read and study that.

Thanks again for your help,

LdS

> From: Michael Schmitz <[EMAIL PROTECTED]>
> Date: Tue, 13 Nov 2001 12:20:07 +0100 (CET)
> To: Laurent de Segur <[EMAIL PROTECTED]>
> Cc: 
> Subject: Re: Kernel Panic with mac-fdisk: More Info
> Resent-From: debian-powerpc@lists.debian.org
> Resent-Date: Tue, 13 Nov 2001 03:20:27 -0800 (PST)
> 
>> The crash always happens with lr pointing to c0039d14 so apparently the test
>> at c0039cf0 fails and the kernel is trying to tell me something (printk)
>> except that I don't see it (no flush?). Apparently the message is still for
>> x86 (ead is a x86 right?).
> 
> Maybe you don't see it because BUG() maps to calling both printk and
> xmon.. That would mean we hit a buffer head that has just been unmapped
> from under us. Turn that BUG() into a call to printk and panic instead...
> 
> My guess is it crashes on the next line (buffer_dirty(bh))... what's in
> r31?

Re: debian-powerpc: how do I run a java app

[Laurent de Segur]

Isn't there some debian list where you can express yourself in your mother
tongue? I believe that the two lists you cc'ed your mail to are for English
usage only. Annoying.


LdS


> From: Viento <[EMAIL PROTECTED]>
> Date: Thu, 15 Nov 2001 09:22:15 -0600 (CST)
> To: Stephen Zander <[EMAIL PROTECTED]>
> Cc: <[EMAIL PROTECTED]>, debian-powerpc
> , , Javier
> Baliosian <[EMAIL PROTECTED]>
> Subject: Re: debian-powerpc: how do I run a java app
> Resent-From: debian-powerpc@lists.debian.org
> 
> Hola Javier... buen dia.. te importaria por favor decirme de donde podeis
> bajar el ISO de SuSe para una rs/6000 43p/140  me urge...o como lo
> puedo  conseguir... aqui en la lista he preguntado y no me han dado
> respuesta alguna..
> 
> 
> 
> -- 
> Saludos
> ---
> 
>     .--. Lic. Domingo Varela Yahuitl.
>    |o_o |Licenciado en Ciencias de la Computación
>    |:_/ |
>   //   \ \   Jefe del Depto de Administración de Sistemas Operativos -
> SICOM
>  (| | )  (Sistema de Información y Comunicación del Estado de Puebla)
> /'\_   _/`\  Tel(s): 01 22 30 00 57/58 Ext(s). 281,222
> \___)=(___/  Lada sin Costo:  01 800 224 30 00 ext 281
> TelCell: 044 23 44 49 63
> ICQ: 113476135
> 
> 
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 
> 

get_hardsect_size not exported

[Laurent de Segur]

Hi,

>From what I found when looking things up, starting from 2.4.4,
get_hardblocksize(dev) has been removed from the kernel. and seems to have
been replaced by a new call named get_hardsect_size(dev).

I compiled a module on a kernel prior to 2.4.4 that used to call the
get_hardblocksize(dev) and after upgrading to 2.4.15-pre2, I had to replace
it with the new call to successfully compile. Easy enough. The problem is
that the new symbol is not exported by the kernel anymore and the module
will fail to load (insmod -v says that get_hardsect_size can't be found.)
Now I wonder what is the best way to solve that.

Thanks in advance for anyone who can help me figuring this.


LdS

mingetty respawing too fast

[Laurent de Segur]

Hi,

On 2.4.12 and up, I get a lot of "respawing too fast" for id 2-6, "disabling
for 5 mins" messages. I looked in my /etc/inittab and the process for these
ids was  assigned to mingetty. Switching to getty for those solved the
problem (I don't see these spurious messages on the console), but I was
wondering if I need to give some peculiar option(s) on the inittab lines to
avoid this error and to be able to continue using mingetty, which I really
like for getting rid of the obnoxious penguin logo after startup.

Thanks for your help in advance,


LdS

Re: Any PPC Java JIT compiler project around ?

[Laurent de Segur]

A port of the HS JVM part of Sun JDK 1.4 is under way. Wanna help?

LdS
- Original Message -
From: "Jason E. Stewart" <[EMAIL PROTECTED]>
To: "David N. Welton" <[EMAIL PROTECTED]>
Cc: 
Sent: Wednesday, November 21, 2001 10:55 AM
Subject: Re: Any PPC Java JIT compiler project around ?


> "David N. Welton" <[EMAIL PROTECTED]> writes:
>
> > Franck Routier <[EMAIL PROTECTED]> writes:
> >
> > > I think this is a real bottleneck in using GNU/Linux PPC (with Java,
> > > of course !)
> >
> > Well, yes, that's what comes of becoming attached to proprietary
> > systems.  You find that they are not portable, and that you are tied
> > down to a whole series of things.
>
> That's not the issue. Without a fast JVM or JIT, java on linux is *dog
> slow*...
>
> > > I know gcj might be a partial solution, but didn't figured out how
> > > to use it.
> >
> > class HelloWorld {
> > public static void main(String[] args) {
> > System.out.println("Hello World!");
> > }
> > }
> >
> > @ashland [~/tmp] $ gcj-3.0 --main=HelloWorld HelloWorld.java
> > @ashland [~/tmp] $ ./a.out
> > Hello World!
> >
> > It's pretty easy, and it's free software!
>
> Doesn't help if you've got someone else's .class files. A fast JVM/JIT
> would help.
>
> jas.
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>
>

Re: Any PPC Java JIT compiler project around ?

[Laurent de Segur]

- Original Message -
From: "Jason E. Stewart" <[EMAIL PROTECTED]>
To: "David N. Welton" <[EMAIL PROTECTED]>
Cc: 
Sent: Wednesday, November 21, 2001 10:55 AM
Subject: Re: Any PPC Java JIT compiler project around ?


> "David N. Welton" <[EMAIL PROTECTED]> writes:
>
> > Franck Routier <[EMAIL PROTECTED]> writes:
> >
> Doesn't help if you've got someone else's .class files. A fast JVM/JIT
> would help.
>
> jas.
>

Wrong! gcj allows you to compile either source files (.java) either bytecode
files (.class). And it works great too. Check their web site for more info.

LdS

Nautilus unstable install fails.Bug?

[Laurent de Segur]

Hi,

I just tried the installation of Nautilus from sid running woody. There is a
dependency of libeel0 on libeel-data and vice-and-versa that causes the
package installer to barf an error message suggesting that I file a bug. I
tried force-installing the 2 individual libs but this shows the same
problem.

I was just curious to know if someone had succeeded in installing Nautilus
on woody lately. I found some reports back in March stating that what I am
trying to do is not working. Maybe things have changed since then.

Thanks for your help and feedback in advance,


LdS