Need help with RC #396817

[John Goerzen]

Hi,

#396817 was reported back in November.  Ian Lynagh, maintainer of GHC,
and I both believe that the build was proceeding normally and that on
the platform in question, it is not unreasonable to expect it to
take quite a bit of time to compile that file.

I have tried, in vain, to figure out someone to tweak the powerpc buildd
to have a longer timeout.

I was told to contact Daniel Jacobowitz, who told me to ask Ryan Murray.
I wrote to Ryan on Dec. 6, and have not yet heard back from him.

This seems like a very simple thing, and I'm frustrated that this
package is being kept out of etch.

Could someone:

1) Fix the autobuilder so this doesn't keep happening?

2) At least build and upload this package?

Thanks,

-- John


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

Re: Need help with RC #396817

[John Goerzen]

On Mon, Dec 18, 2006 at 10:55:48AM -0600, Carlo Segre wrote:
> On Mon, 18 Dec 2006, John Goerzen wrote:
> 
> >Hi,
> >
> >#396817 was reported back in November.  Ian Lynagh, maintainer of GHC,
> >and I both believe that the build was proceeding normally and that on
> >the platform in question, it is not unreasonable to expect it to
> >take quite a bit of time to compile that file.
> >
> 
> OK, it builds.  WOuld you like me just to upload?

Yes, please, and thanks for your help!

-- John


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

Re: A couple of questions

[John Goerzen]

On Tue, Jun 25, 2002 at 02:29:27PM -0500, Patrick Klee wrote:
> I am fully satisfied with my Debian iMac; However, I have a couple of 
> questions.  I saw the post on DebianPlanet for XF86 debs, but I could seem to 
> download any for powerpc, that was two weeks ago.

XF86 is included with Debian PowerPC; apt-get install xserver-xfree86

> Also, I heard Sid has 1+ packages.  That probably includes PPC, but I 
> picked everything, even non-free and I get something like 9500+ packages.

Some packages are architecture-specific (ie, LILO) and are not portable or
available on other architectures.  You are getting a full set of available
packages.

-- John


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

Re: Sound control on TiBook G4

[John Goerzen]

On my Powerbook, I have found that the "master" control has no effect, but
rather that separate "speaker" and "headphone" controls work.  Whether or
not that will be the same for you, I don't know.  This is with kmix.

-- John


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

Re: date is 1934 on reboot

[John Goerzen]

On Mon, Oct 21, 2002 at 07:24:51AM -0700, Chris Tillman wrote:
> Yes, usually the date/time not being kept is the first indication.
> It's fairly easy to replace, and although not a household battery,
> still fairly common.

I believe there are firmware and/or OS bugs that cause the date to show up
incorrectly.  I have seen this numerous times on my Powerbook, even when
brand new, especially when booting one OS and then another.

-- John

KDE progress on Alpha & PowerPC

[John Goerzen]

Hi,

A nasty bug[1] in the qt headers has been keeping KDE 3 out of all arches
other than i386.  This is bad, because the Arch: all parts of KDE 3 are
already there, so KDE apps were even more broken than on i386.

The Qt libs have been fixed (mostly; see the discussion at the end of [1]),
so I have uploaded a binary-only NMU of kdelibs to powerpc and alpha which
fix the KDE build problems.

But: a lot of KDE packages are not auto-buildable because their configure
scripts are not marked +x.  I've been manually building them as I can.

On Alpha, I've been seeing some weird relocation errors [2][3].  I haven't
tried those packages on powerpc yet (about to).

[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=180816
[2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=181624

Re: KDE progress on Alpha & PowerPC

[John Goerzen]

Strangely, this compiled fine for me on powerpc.  Has the X bug been
reported somewhere already?

On Wed, Feb 19, 2003 at 05:57:27PM +0100, Michel Dänzer wrote:
> On Mit, 2003-02-19 at 16:54, John Goerzen wrote:
> > 
> > On Alpha, I've been seeing some weird relocation errors [2][3].  I haven't
> > tried those packages on powerpc yet (about to).
> 
> [...]
> 
> > [2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=181624
> 
> Falk is correct; this will probably fail similarly on powerpc (maybe
> only at runtime though).
> 
> It should either use libXxf86dga_pic.a from xlibs-pic or wait for
> XFree86 4.3.0 packages.
> 
> 
> -- 
> Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
> XFree86 and DRI project member   /  CS student, Free Software enthusiast

Infrared broken on tibook

[John Goerzen]

Hi,

I have a 400MHz tibook and I'm trying to get infrared going.

On bootup, dmesg shows this when macserial.o is loaded:

macserial: i2c-modem detected, id: 1
PowerMac Z8530 serial driver version 2.0
tty00 at 0xd9867020 (irq = 22) is a Z8530 ESCC (internal modem)
tty01 at 0xd986e000 (irq = 23) is a Z8530 ESCC (IrDA)
i2c-core.o: i2c core module

So, I assume that /dev/ttyS1 is the SIR interface.  Strangely:

# cat /dev/ttyS1
cat: /dev/ttyS1: Input/output error

Obviously, irattach therefore fails.

I am running the latest 2.4.20 from benh.

-- John

Re: Of Mice and Modems

[John Goerzen]

The modem should be /dev/ttyS0 or /dev/ttyS1 like others.  It tripped
me up for a bit that you have to use macserial.o instead of serial.o
to drive it.  Incidentally, I can't make it work with efax yet.
Anybody have any tips?  (This is a tibook g4 box)

-- John

"mmissett" <[EMAIL PROTECTED]> writes:

> I finally have things up and running on the G4, but I'm stuck on a couple of 
> snags:  pppconfig can't figure out where the modem is and it's asking... me.  
> I've tried a few guesses, but with no evidence of success.  Also, startx is 
> failing, saying "cannot open mouse (no such file or directory)".  The mouse 
> in question is an Apple USB optical mouse.  My question is:  Is the  system 
> looking in all the wrong places for things or have I just given it more wrong 
> answers than it can tolerate?  Any advice, hints, etc. would be most 
> appreciated.
> Michael Missett  
> 
> 
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
> 

-- 
John Goerzen <[EMAIL PROTECTED]>   www.complete.org
Sr. Software Developer, Progeny Linux Systems, Inc.www.progenylinux.com
#include  <[EMAIL PROTECTED]>

Re: Of Mice and Modems

[John Goerzen]

Michael Schmitz <[EMAIL PROTECTED]> writes:

> Well, fax send craps out with a segfault, and efax won't dial unless I

Stopping the segfault can be done by removing ewin from the VERB=
setting in the conffile, oddly enough.

> beat it into submission with -iX3 (I'm behind a PBS with no dialtones).
> Other than that, it seems to talk to the modem OK. It just doesn's send
> successfully (only a few scan lines get sent before the remote machine
> chokes).

This last problem is the one I'm stuck on now.

Re: Of Mice and Modems

[John Goerzen]

Media 100 <[EMAIL PROTECTED]> writes:

> Thanks, everybody.  John, where does it get told macserial.o?  I apologize
> if that's an ignorant question, but then...

It's in benh's kernel available on
rsync.penguinppc.org::linux-2.4-benh

-- John

Re: Of Mice and Modems

[John Goerzen]

Michael Schmitz <[EMAIL PROTECTED]> writes:

> The conffile is duplicated in /usr/bin/fax so I had to remove it there. 

I didn't have that particular problem.

> Do you get the 'characters received while sending' message? I suspect flow
> control between modem and kernel driver is broken (even switching to
> hardware flow control doesn't cut it). 

No, I got "bad response from modem" if memory serves.  However I was
wondering if it wasn't a flow control issue too.  Since everything
seems to go fine until it starts sending.  Weird.

-- John

Sound problems

[John Goerzen]

Hi,

I've got a tibook G4 running sid and benh's latest kernels.  I've
loaded the dmasound_pmac module and it seems to be OK.  However I have
lots of troubles making it work.  For instance, when KDE starts, it
complains:

Sound server fatal error:
Error while initializing the sound driver:
Can't set playform format(_format = 17, asked driver to give 16, got 32)

Playing from apps like videolan often results in a couple of seconds
of very scratchy audio, followed by static, followed by silence.
Mixers seems to have no effect once the silence appears (everything's
muted).  I don't know what to try at this point.

dmasound_pmac appears to load OK:


dmasound_pmac: Awacs/Screamer Codec Mfct: 1 Rev 3
dmasound_pmac: found Keylargo rev 2 or later - H/W byte-swap disabled
PowerMac Screamer  DMA sound driver rev 016 installed
Core driver edition 01.06 : PowerMac Built-in Sound driver edition 00.06
Write will use4 fragments of   32768 bytes as default
Read  will use4 fragments of   32768 bytes as default

Any assistance appreciated!

Thanks,
John

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> I haven't tried VLC on anything, so I don't know about this. The
> developers of VideoLan, however, probably aren't making their client obey
> the driver's request for big-endian samples instead of little-endian
> samples.

It seems that very few programs do.  Perhaps it would be easier to
make the kernel driver accept little-endian samples, do the conversion
itself, and post those to the driver thanto go on a hunt for all the
poorly-written software out there?

> I didn't know that the PowerBook G4 also used KeyLargo. Does the Screamer
> chip in it support little-endian samples? Does anyone know?

It doesn't appear so; I hacked the kernel module so that it set the
supports little endian var to 1, but it behaved even worse then (no
sound plays at all, and arts gets all confused).

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> allowed. The plan is eventually to remove all the sample-conversion code,
> and only allow the sample formats that the hardware knows, expecting
> userspace to do the conversion work.

Ick.  Wave good-bye to all the non-esound/arts stuff out there, I
guess.

> > It doesn't appear so; I hacked the kernel module so that it set the
> > supports little endian var to 1, but it behaved even worse then (no
> > sound plays at all, and arts gets all confused).
> 
> That's good to know.

Well yes and no.  I changed your test for rev that was rev >= 2 to rev
== 2 (mine is a rev 3) which effectively said it could do the little
endian samples. However, rather than just playing with a bunch of
nasty-sounding static as I expected to occur, *nothing* happened other
than what looked like an entire lockup for awhile -- X entirely froze
for about 20 seconds when I tried to play something with arts.  No
idea why, but suffice it to say I'm back to the way it was :-)

-- John

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

An update -- xmms works fine on my tibook, with one exception.  It's
volume control has zero effect.  This might confirm my earlier
thoughts about that.  Unfortunately, arts doesn't work so my usual
mixer software won't work, xmix doesn't even excute on ppc, and aumix
seems ineffectual as well.  Hmmph.

-- John

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> Yay. I'm sure it's ALSO using an I2C-based mixer control arrangement,
> which is, I'm sure, different from Texas/Tumbler (iBook2) and DACA
> (iBook). Can you enable the i2c-keywest driver under the I2C configuration
> options, and let me know how many I2C buses it sees? (They'll be visible
> under /proc/bus.)

Loading the i2c-keywest module reported:

Keywest device found: /[EMAIL PROTECTED]/[EMAIL PROTECTED]
i2c-core.o: adapter keywest i2c registered as adapter 0.
Keywest device found: /[EMAIL PROTECTED]/[EMAIL PROTECTED]/[EMAIL PROTECTED]
i2c-core.o: adapter keywest i2c registered as adapter 1.

Then:

alexanderwohl:/proc/bus# ls
i2c  i2c-0  i2c-1  pccard  pci
alexanderwohl:/proc/bus# cat i2c
i2c-0   smbus   keywest i2c Non-I2C SMBus 
adapter
i2c-1   smbus   keywest i2c Non-I2C SMBus 
adapter
alexanderwohl:/proc/bus# cat i2c-0
alexanderwohl:/proc/bus# cat i2c-1

> 

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> On 11 Oct 2001, John Goerzen wrote:
> 
> > Keywest device found: /[EMAIL PROTECTED]/[EMAIL PROTECTED]
> > i2c-core.o: adapter keywest i2c registered as adapter 0.
> > Keywest device found: /[EMAIL PROTECTED]/[EMAIL PROTECTED]/[EMAIL PROTECTED]
> > i2c-core.o: adapter keywest i2c registered as adapter 1.
> 
> Yep. That's what I suspected. I suppose someone's gonna have to hack out a
> driver for the Screamer's I2C control device on the PBG4 now too.
> _Yikes_... as if DACA isn't enough of a pain (what with making the IDE bus
> lock if you have the wrong timings!)...

To whom do I write about that?  Also what is DACA stuff?  This is my
first PPC box and I don't know a lot about its internal architecture.

Thanks!

-- John

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> On 11 Oct 2001, Michel Dänzer wrote:
> 
> > Have you tried the main volume (there's an option about this in the xmms
> > OSS output plugin preferences) and speakers mixers (in case you're using
> > the internal speakers)?
> 
> Well, considering he said that no mixer app will even start up, that gives
> me the impression that mixer control isn't working on Screamer (or at
> least not with Screamer on the G4 PowerBook)...

I said that xmix wouldn't start up and that the kde mixer stuff is
useless because of arts not working.  aumix does start up but it
doesn't work for me.  However, that could just be due to my own
incompetance wrt not knowing what to set or how to drive aumix.

-- John

> 
> Derrik Pates  |   Sysadmin, Douglas School   |#linuxOS on EFnet
> [EMAIL PROTECTED] | District (dsdk12.net)|#linuxOS on OPN
> 

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Michel Dänzer <[EMAIL PROTECTED]> writes:

> On Thu, 2001-10-11 at 19:46, John Goerzen wrote:
> > 
> > An update -- xmms works fine on my tibook, with one exception.  It's
> > volume control has zero effect.  This might confirm my earlier
> > thoughts about that.  Unfortunately, arts doesn't work so my usual
> > mixer software won't work, xmix doesn't even excute on ppc, and aumix
> > seems ineffectual as well.  Hmmph.
> 
> Have you tried the main volume (there's an option about this in the xmms

No effect.

> OSS output plugin preferences) and speakers mixers (in case you're using
> the internal speakers)?

I am.  I'll try and figure out how to do that somewhere...


-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> Well, considering he said that no mixer app will even start up, that gives

I should also add that the reason xmix won't start is related to X or
shared libaries, not the audio system.

-- John

> me the impression that mixer control isn't working on Screamer (or at
> least not with Screamer on the G4 PowerBook)...
> 
> Derrik Pates  |   Sysadmin, Douglas School   |#linuxOS on EFnet
> [EMAIL PROTECTED] | District (dsdk12.net)    |#linuxOS on OPN
> 

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Sound problems

[John Goerzen]

Michel Dänzer <[EMAIL PROTECTED]> writes:

> > aumix does start up but it doesn't work for me.  However, that could
> > just be due to my own incompetance wrt not knowing what to set or how
> > to drive aumix.
> 
> Tried to help you there... 'Vol' is the headphone out, 'Spkr' is the
> speakers, they're independent. All other mixers seem useless.

Ahh, I fear I misunderstood.  Spkr does indeed work for me.  Thanks!

-- John

Re: Sound problems

[John Goerzen]

Derrik Pates <[EMAIL PROTECTED]> writes:

> KMix doesn't depend on aRts working. If it won't work, it's because it's
> having problems talking to the mixer device. aumix starting, otoh, may or
> may not be a good sign. Maybe install 'aumix-gtk' instead, so you have a
> GUI available, and see if you get any further that way?

Ahh, thanks again.  kmix also seems to work.  I hadn't even tried it
before thinking that "It's KDE, therefore it depends on arts; arts is
broken; therefore, kmix is broken."  Oops.

Re: Bug#116780: util-linux: hwclock shouldn't be run

[John Goerzen]

Daniel Jacobowitz <[EMAIL PROTECTED]> writes:

> John, what kernel/machine do you have hwclock causing a hang?  I'm not
> sure off the top of my head what the status is in 2.2, but 2.4 has a
> "proper" /dev/rtc driver.  We should be using hwclock.  Clock is an
> awful ADB-bit-bashing hack that needs to die.

I'm using BenH's 2.4.x kernels.  I observed this behavior in both
2.4.13pre and 2.4.10 release.  I had not installed hwclock when I was
using any previous version.

Running it causes this:   (from dmesg)

IN from bad port 71 at d983612c
Real Time Clock Driver v1.10e
IN from bad port 71 at d9835eac
IN from bad port 71 at d9835f08
IN from bad port 71 at d9836b64
IN from bad port 71 at d9836bc4
IN from bad port 71 at d9835eac
IN from bad port 71 at d9835f08

The clock program works flawlessly.

Re: Bug#116780: util-linux: hwclock shouldn't be run

[John Goerzen]

Ethan Benson <[EMAIL PROTECTED]> writes:

> you should not change it.  this is caused by misconfigured kernels, on
> powerpc the kernel *MUST* be compiled with CONFIG_PPC_RTC and NOT
> CONFIG_RTC, the latter must be disabled, completely.

And this is documented where?  Why is CONFIG_RTC even offered if it's
buggy?  Not only that, but I had both compiled as a module.  If the
kernel is loading the wrong one, again I maintain it's a bug in the
kernel and not with me.

> 
> > IOW I need a replacement for the following part of the postinst:
> 
> you need to change nothing, its PEBCAK.

Before you start flaming people, perhaps you ought to check to see
where people might find out this information while compiling the
kernel.

I see no reason that you should automatically assume that users will
magically divine that a certain commonly-used kernel option will
interfere with hwclock.  Suggesting that is ludicrous.

Re: Bug#116780: util-linux: hwclock shouldn't be run

[John Goerzen]

Ethan Benson <[EMAIL PROTECTED]> writes:

> you configured it wrong then.
> 
> CONFIG_RTC=n
> CONFIG_PPC_RTC=y
> 
> that is the correct config.  yours is no doubt backwards.

No, mine has both compiled as a module.

> the clock program as Dan put it is a kludge that must die, and a
> couple users misconfiguring thier kernel is no reason to bring it
> back.

Why is the option in the kernel if it's useless?  Your constant
badgering me for doing what all information presented leads me to
believe is the correct path is becoming annoying.  Perhaps you should
implement a telepathic kernel configuration interface so that
everybody that reads the documentation knows the extra five paragraphs
that appear nowhere?

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Bug#116780: util-linux: hwclock shouldn't be run

[John Goerzen]

Adrian Bunk <[EMAIL PROTECTED]> writes:

> As far as I understand it clock from powerpc-utils is obsolete and you can
> use hwclock on all machines (including PReP machines?). Does that mean you
> will remove the clock program from powerpc-utils?

I'm not so sure this is true.  I've had terrible trouble with hwclock
setting the CMOS clock to either 1903 or 1933.  It seems to cause
significant problems in this area.  The clock program works fine.

To those that would flame me: yes, my kernel is configured as
recommended now.

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Re: Bug#116780: util-linux: hwclock shouldn't be run

[John Goerzen]

Tom Rini <[EMAIL PROTECTED]> writes:

> Is this a PReP box you're talking about?  What kernel are you running
> and did you set a date on the box from PPCBUG to start with?  hwclock is
> quite happy on mine, and PRePs don't have ADB anyhow so the evil clock
> hack program can't do anything.

It's a G4 PowerBook (tibook).  Kernel 2.4.13-pre3-ben0.  I don't know
what PPCBUG is.

-- John

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

-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

[Stefan Westerfeld ] Re: Not entirely fixed. - big-endian arts issues

[John Goerzen]

--- Begin Message ---
   Hi!

On Wed, Oct 31, 2001 at 08:50:48PM -0700, Ivan E. Moore II wrote:
> Response on the arts patch you supplied.
> 
> On Wed, Oct 31, 2001 at 09:04:51AM -0500, John Goerzen wrote:
> > 
> > There seems to have been progress, but it's not entirely there yet.
> > The arts daemon no longer complains about endianness, but still
> > doesn't work.  It sometimes complains of a "sound server fatal error:
> > CPU overload" when trying to play a sound or when starting up the
> > server.  It still doesn't actually play anything.  I did make sure to
> > try forcing it to 44100 Hz and 16-bit (formats supported by the
> > hardware), but this did not help at all.  I'm thinking there is still
> > an upstream bug.

Ok - thanks for the feedback, I am closing the bug now (and putting the fix
in the KDE_2_2_BRANCH). To the other issue, CPU overload:

This is probably not an aRts bug. Basically, the kernel driver can cause artsd
to fail in two ways:

 a) select for writing on audio FD wakes up too early (nothing to do)
 b) GETOSPACE ioctl implemented wrong (lies about the actual free space)

If you use strace, you will see that artsd is woken up again and again by
select() far too often. I know that the problem exists for a few kernel drivers
(like i810), but aRts isn't the place to fix it. I also know that other
programs that use the kernel drivers differently don't experience trouble
(i.e. xmms), but still its a kernel bug. If you need more information, let me
know. Please also consider reporting it to the author of the kernel driver.

   Cu... Stefan
-- 
  -* Stefan Westerfeld, [EMAIL PROTECTED] (PGP!), Hamburg/Germany
 KDE Developer, project infos at http://space.twc.de/~stefan/kde *- 
--- End Message ---


-- 
John Goerzen <[EMAIL PROTECTED]>GPG: 0x8A1D9A1Fwww.complete.org

Sound broken in KDE again

[John Goerzen]

Sigh...

I upgraded from 2.4.15-pre6-ben0 to 2.4.17-pre2-ben0 and now I get
this from KDE on startup:

Error while initializing the sound driver:
SNDCTL_DSP_SETFMT failed - Invalid argument

and sound in KDE is broken.  Ideas?

-- John

Airport "Tx error" problem

[John Goerzen]

Hello,

I have a problem with my Apple Airport card in a PowerBook G4 400MHz:

I will load the module, use iwconfig to set my key, and then try to 
start the normal DHCP process.  The kernel logs these errors:

eth1: Tx error, status 4 (FID=01E3)
eth1: Tx error, status 4 (FID=01E3)
eth1: Tx error, status 4 (FID=01CF)
etc ...

I'm using kernel 2.2.18-ben0 with version 0.09b of airport.c and 
orinoco.c.  I found a message at 
http://lists.samba.org/pipermail/wireless/2001-December/003534.html 
that seems to include a patch for a related problem.  Looking at the 
source, it doesn't include the comments in that patch but seems 
different enough that I can't tell quite whether or not it's been 
patched.  (I'm not at all an expert in these things)

I had previously been using a Wavelan Silver PCMCIA card with the 
same configuration, and it worked fine.  I have also tried an earlier 
ben0 kernel and got the same results.  The Airport card works fine in MacOS
X.

Thanks,
John Goerzen
[EMAIL PROTECTED]

Re: Airport "Tx error" problem

[John Goerzen]

Well, I have some additional information on this problem.

It appears that when I configure the interface statically, it all works 
fine.  If I use DHCP to configure it (with pump), then I have the problem 
below -- can't communicate to anywhere.  I have to rmmod airport and 
orinoco and then bring it up statically for it to work again.


I get the Tx errors whether it's configured statically or dynamically.  
Only when it is configured statically can I actually communicate.



On Friday, March 8, 2002, at 10:08  AM, John Goerzen wrote:


Hello,

I have a problem with my Apple Airport card in a PowerBook G4 400MHz:

I will load the module, use iwconfig to set my key, and then try to
start the normal DHCP process.  The kernel logs these errors:

eth1: Tx error, status 4 (FID=01E3)
eth1: Tx error, status 4 (FID=01E3)
eth1: Tx error, status 4 (FID=01CF)
etc ...

I'm using kernel 2.2.18-ben0 with version 0.09b of airport.c and
orinoco.c.  I found a message at
http://lists.samba.org/pipermail/wireless/2001-December/003534.html
that seems to include a patch for a related problem.  Looking at the
source, it doesn't include the comments in that patch but seems
different enough that I can't tell quite whether or not it's been
patched.  (I'm not at all an expert in these things)

I had previously been using a Wavelan Silver PCMCIA card with the
same configuration, and it worked fine.  I have also tried an earlier
ben0 kernel and got the same results.  The Airport card works fine in 
MacOS

X.

Thanks,
John Goerzen
[EMAIL PROTECTED]


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

Delete key

[John Goerzen]

Hi,

I have a titanium Powerbook G4.  This laptop does not have a key that
sends "delete".  It has a "delete" key, but Linux uses it as Backspace.
   Mozilla's mail/news reader wants to use a "delete" key for deleting a
message.  How can I fake a delete key?

-- John


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

Enigmail for Linux/PowerPC

[John Goerzen]

Hello,

I have built IPC and Enigmail 0.49.1 (though it calls itself 0.49.2) for
Linux/PowerPC systems for Mozilla 1.0RC1.  It's at
gopher://quux.org/1/devel/debian/mozilla or
http://quux.org:70/devel/debian/mozilla.

-- John


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

Re: Test-DFS for PowerPC

[John Goerzen]

On Mon, Jun 07, 2004 at 06:40:12AM +0200, Sven Luther wrote:
> > Bad, bad, bad. You are doing "boot cd:,\boot\yaboot" from OF, right?
> > (its not \install as on regular debian cds).
> 
> Yeah, directly copied from your other mail. I had trouble finding the
> back-/ at first, as it seems to be inexistant on french mac keyboards, ,
> but then i noticed that the OF is in qwerty layout anyway :)
> 
> > > BTW, where does the kernel on the iso comes from ? 
> > 
> > Selfmade since we can't build things like ide/cdrom... as modules
> > because the primary initrd does not contain any modules. The config is
> > on the image, it's debian 2.6.6 sources, config modelled after
> > pmac_defconfig and the stuff from John's i386 config added.
> 
> Why not add those modules to the initrd and have it loaded, like the
> mkinitrd generated initrd does ? Indeed it should even be possible to
> use mkinitrd for this purpose, and add an own startup script or
> something such. Not sure how mkinitrd would take to having ocaml
> programs (or scripts) though, but i have the feeling that it should make
> no difference.

What's the point, though?  We bloat the initrd size by putting libc, all
the module loading tools, etc. on it, or we bloat the kernel size by
compiling things into the kernel.

Also, mkinitrd seems to just load modules that are specified in a file
like /etc/modules.  It doesn't do any real autodetection, so I don't see
what it buys us at all.

-- John

Re: Test-DFS for PowerPC

[John Goerzen]

On Mon, Jun 07, 2004 at 05:06:27PM +0200, Sven Luther wrote:
> reimplementation using the discover database), as the debian-installer
> does.

Well, the purposes of DFS and of debian-installer are a little
different.

DFS explicitly does *not* do hardware autodetection on boot, though it
prints out a message on boot telling the user how to run discover or
hotplug.  The reason is that, as a recovery tool, you may not want to
attempt this discovery (and that could especially be the case if, for
instance, your network controller is fried and the kernel hangs trying
to bring it up).

Another goal is to make the kernel be one that somebody can slap onto
their hard disk in an emergency and be able to boot from their hard disk
again.  I actually initially was going to not use modules at all to make
it easier, but when I realized that some Ethernet module's probing code
hangs machines, I abandoned that course :-)  However, I still think that
mkinitrd or some such solution is too complex and error-prone for this
use.

The stock Debian kernels are unsuitable for a number of reasons.  One is
that they do not have enough filesystems, LVM/RAID, etc. compiled in, so
they are unsuitable for emergency use in a lot of situations (requiring
an initrd just to be able to grok the root fs in these cases).

I am not terribly troubled by a 5MB kernel size for this application,
actually.  I don't know a lot about PowerPC subarchs, but on Alpha we do
have subarchs but can compile a "generic" kernel that runs everywhere,
which is what I have done.  I'm not all that concerned, in general,
about running on machines old enough that they can't run a 5MB kernel
and bash at the same time.

DFS kernels also omit things that are not necessary for
installation/repair work.  For instance, sound and video4linux are
completely disabled.  The DFS kernels are here to give people a live,
working system to use to repair or install Debian and build a custom
kernel from there.

> Actually, i believe that the initrd of the default debian kernel should
> use discover too, and not the non-robust mess that is produced by
> mkinitrd right now. And then, having an ocaml reimplementation of

Yes, mkinitrd mostly sucks.  OTOH neither discover nor hotplug are
complete solutions.  Both miss certain hardware.

> If ytou don't do that, and which to support more than just the subarches
> that are most common, you will need to do agiant kernel like we did for
> 2.4.25 (the debian kernel is over 5Mo big i believe).

Can you clarify how the powerpc subarchs work so I can get a better
understanding of the problem?  Is it possible to compile a kernel that
runs on all subarchs?

Thanks,
John

Re: Test-DFS for PowerPC

[John Goerzen]

On Mon, Jun 07, 2004 at 05:49:02PM +0200, Robert Jordens wrote:
> > DFS kernels also omit things that are not necessary for
> > installation/repair work.  For instance, sound and video4linux are
> > completely disabled.  The DFS kernels are here to give people a live,
> > working system to use to repair or install Debian and build a custom
> > kernel from there.
> 
> I'd really like to be able to build DFS images "en masse" for each and
> every purpose. Just like all those crazy people that praise their "Yet
> another Knoppix remastered" as the new holy grail. It's just that their
> remastering process is so "manual" (and BTW their mental contribution is
> not that much). ;-]
> 
> So. I'd really like to see DFS not only becoming a "Yet another Rescue
> CD" but also "the first fully automatized live CD generation system".

Yes.  I agree!  In fact, that's why I started the project.  I guess it's
a little confusing to use DFS to refer both to the "standard" DFS ISOs
and to the build system.

dfsbuild is already very far along that path, with the sets of packages
(save for the base system) fully configurable, kernels configurable too,
etc.

-- John

Re: Test-DFS for PowerPC

[John Goerzen]

On Mon, Jun 07, 2004 at 10:58:02AM -0500, John Goerzen wrote:
> dfsbuild is already very far along that path, with the sets of packages
> (save for the base system) fully configurable, kernels configurable too,
> etc.

I should add that I view dfsbuild as a far more useful tool than the
custom remastering hack that Knoppix uses.  dfsbuild is scriptable and
does its entire job without any user intervention, and also doesn't
require an existing ISO to do so.  All it needs is a Debian mirror :-)

-- John

Re: Test-DFS for PowerPC

[John Goerzen]

On Mon, Jun 07, 2004 at 06:13:56PM +0200, Robert Jordens wrote:
> 
> [Mon, 07 Jun 2004] John Goerzen wrote:
> > I should add that I view dfsbuild as a far more useful tool than the
> > custom remastering hack that Knoppix uses.  dfsbuild is scriptable and
> > does its entire job without any user intervention, and also doesn't
> > require an existing ISO to do so.  All it needs is a Debian mirror :-)
> 
> Exactly. And because standard Debian is what we build on, I think we
> should be able to also use standard Debian kernels. 

OK, I get it now.  Thanks for your patience :-)

> If packages are well built and configure automatically, they are
> suitable for DFS. If kernels are well built and configure automatically
> with the help of discover and/or hotplug, why not let them be suitable
> for DFS also. Yes kernels for a Rescue CD are special, no need for
> hardware discovery. But for a DFS with Knoppix functions, The Debian
> standard Kernels are ideal.

OK, I had thought we were talking specifically about the kernels for the
standard images.  You have a good point that it would be useful for
dfsbuild users to be able to use standard kernels.  Obviously our
ramdisk generation code will need some work to make it happen.  I'll add
an entry to the TODO file :-)

-- John

Re: Test-DFS for PowerPC

[John Goerzen]

On Wed, Jun 09, 2004 at 01:07:01AM +0200, Robert Jordens wrote:
> > No, the dfsbuild packages modified to build the powerpc version, (the
> > source package of the -rj* versions) ?
> 
> The Source is under Arch:
> 
> dfsbuild--jordens--0.5
> 
> If you want packages, I'll do them but I'd rather prefer John to do the
> releases.

Robert and I sync up trees frequently.  If you're working on PowerPC
stuff, it makes sense for you to use Robert's tree.  (Robert, do please
drop me an e-mail if you notice I haven't synced with you for awhile and
there's stuff to go in.)

If you're working on non-PowerPC stuff, it may make sense for you to
work directly from my tree, but it's still OK if you work from Robert's
as long as he syncs with you too.

-- John

True state of support for 12" iBook G4

[John Goerzen]

Hello,

I am considering purchasing a new 12" iBook from Apple.  The intent is
to use this with Debian.  However, in researching the state of support
for this unit, I've found wildly contradictory information.  So I'm
wondering if someone could help clear this up for me.

1. Sleep mode.  According to [1], sleep doesn't work at all.  However,
   others are not reporting any trouble with it (basically saying
   everything except the Airport Extreme works fine).  Does sleep work
   on this unit?

2. Video.  Not only are there conflicting stories (ranging from
   "accelerated video will never be available on this unit" to
   "dri trunk already supports it fine"), but I've even seen a thread
   where people were debating whether the unit has NVidia or ATI
   graphics.  What's the story here?

3. The Airport Extreme.  Everybody agrees that it doesn't work now.
   Some people are saying there are projects to reverse engineer it.
   But I've never found any solid info on any of these projects.
   Is there really any hope here, or should I just buy a USB
   device and be done with it?  If I get a USB device, what do you
   recommend?

Thanks!

[1] http://geekounet.org/powerbook/ibookg4.html

Re: True state of support for 12" iBook G4

[John Goerzen]

On 2004-10-28, Jérôme Marant <[EMAIL PROTECTED]> wrote:
> Quoting John Goerzen <[EMAIL PROTECTED]>:
>
>> Hello,
>>
>> I am considering purchasing a new 12" iBook from Apple.  The intent is
>> to use this with Debian.  However, in researching the state of support
>> for this unit, I've found wildly contradictory information.  So I'm
>> wondering if someone could help clear this up for me.
>
> Buying Apple hardware is often a bad choice regarding linux
> support. I bought an iBook 2 years ago because the ratio quality/price
> was good, but these days, there are much nicer alternatives in
> the x86 world.

Perhaps you have a recommendation for me then.  I am looking for a
small, very rugged, and inexpensive laptop.  The iBook seems to fit the
bill nicely.  I can get it configured how I want for about $1200.  I
could get a Sony that does what I want for about $1000 more, but it
would probably be less rugged than the iBook.

>
> Cheers,
>
> --
> Jérôme Marant
>
>


-- 
John Goerzen
Author, Foundations of Python Network Programming
http://www.amazon.com/exec/obidos/tg/detail/-/1590593715

Re: True state of support for 12" iBook G4

[John Goerzen]

On 2004-10-28, Jérôme Marant <[EMAIL PROTECTED]> wrote:
> John Goerzen <[EMAIL PROTECTED]> writes:
>
> These days, people seem to recommend IBM X40 series, but I don't
> know if they are cheap enough for you.

Yep, nice series.  Unfortunately, configured the same as an ibook for
the things that matter to me, they're more than $1000 more expensive
than the iBook 12".

-- John