Bug#365010: Xserver G5 usb keyboard not loaded ...
Sven Luther
Boot method: netboot mini.iso.
Image version: http://people.debian.org/~luther/d-i/images/2006-04-26/powerpc64/netboot64/mini.iso
Date: jeu avr 27 11:54:50 CEST 2006
Machine: Apple XServer G5, RackMac3,1
Processor: Dual PPC970FX 2.3Ghz
Memory: 1GB
Partitions:
Output of lspci and lspci -n:
Base System Installation Checklist:
[O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it
Initial boot worked: [O]
Configure network HW: [O]
Config network: [O]
Detect CD: [ ]
Load installer modules: [O]
Detect hard drives: [O]
Partition hard drives: [O]
Create file systems: [O]
Mount partitions: [O]
Install base system: [O]
Install boot loader: [O]
Reboot: [O]
Comments/Problems:
1) daily build business card and netinst isos are failing to build since
april 1, which means they don't include the broadcom tg3 module at all,
and are thus not really usable for installs at this time.
2) ideally, we should include fan control modules (therm_pm72) in the
ramdisk, as the noise during install as the fan reach full speed quickly
becomes bothersome. Mmm, therm_pm72 doesn't seem to be included in the
2.6.15 kernels.
Friendly,
Sven Luther
--
To UNSUBSCRIBE, email to debian-bugs-...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Sven Luther
> Sven
>
> I tried this latest netboot64 image, but I get the same behavior as with
> the 2006-04-23 image, which I submitted a bug report for (bug 364637).
>
> The boot hangs at 'Trying to im_free nonexistent area (d0000800820d7000)'
>
> In the bug report, I included the whole console log.
> Should I submit another bug report, or update bug 364637?
> Toni
What hardware do you have ? I have an XServer dual G5 (RackMac3,1), and i just
did an installation without problem, using today's daily build.
Friendly,
Sven Luther
--
To UNSUBSCRIBE, email to debian-powe...@lists.debian.org
Andreas Barth
> On Thu, Apr 27, 2006 at 12:53:12PM +0200, Frans Pop wrote:
> > On Thursday 27 April 2006 12:14, Sven Luther wrote:
> > > 1) daily build business card and netinst isos are failing to build
> > > since april 1, which means they don't include the broadcom tg3 module
> > > at all, and are thus not really usable for installs at this time.
> >
> > builds. This results in debian-cd not being able to find a file needed
> > for mkisofs.
> >
> > I understand from comments on IRC that this is due to a kernel issue?
>
> Oh fun, i don't have any commit access to the d-i repo anymore, so i can't
> even fix this issue myself. This clearly shows the pettiness of the d-i team,
> i am disgusted. ...
>
> Dear fellow powerpc folk, this clearly means that the debian support for
> powerpc is dead or almost so, and i strongly recomend you to go find another
> distribution to run which cares a bit more about the powerpc architecture.
I'd rather prefer if you don't overexegerate.
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
Sven Luther
> On Thursday 27 April 2006 12:14, Sven Luther wrote:
> > since april 1, which means they don't include the broadcom tg3 module
> > at all, and are thus not really usable for installs at this time.
>
> builds. This results in debian-cd not being able to find a file needed
> for mkisofs.
>
> I understand from comments on IRC that this is due to a kernel issue?
Oh fun, i don't have any commit access to the d-i repo anymore, so i can't
even fix this issue myself. This clearly shows the pettiness of the d-i team,
i am disgusted. ...
Dear fellow powerpc folk, this clearly means that the debian support for
powerpc is dead or almost so, and i strongly recomend you to go find another
distribution to run which cares a bit more about the powerpc architecture.
Friendly,
Toni L. Harbaugh-Blackford [Contr]
> On Thu, Apr 27, 2006 at 07:18:07AM -0400, Toni L. Harbaugh-Blackford [Contr] wrote:
> > Sven
> >
> > I tried this latest netboot64 image, but I get the same behavior as with
> > the 2006-04-23 image, which I submitted a bug report for (bug 364637).
> >
> > The boot hangs at 'Trying to im_free nonexistent area (d0000800820d7000)'
> >
> > In the bug report, I included the whole console log.
> > Should I submit another bug report, or update bug 364637?
> > Toni
>
> What hardware do you have ? I have an XServer dual G5 (RackMac3,1), and i just
> did an installation without problem, using today's daily build.
Xserve G5 dual 2GHz (RackMac3,1) cluster node, headless
Does the 'Trying to im_free ...' message appear in your console log, and if
so what is before and after it?
Thanks,
Toni
>
> Friendly,
>
> Sven Luther
>
>
> --
> To UNSUBSCRIBE, email to debian-powe...@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
>
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Toni L. Harbaugh-Blackford [Contr]
On Thu, 27 Apr 2006, Andreas Barth wrote:
> * Sven Luther (sven....@wanadoo.fr) [060427 13:23]:
> > On Thu, Apr 27, 2006 at 12:53:12PM +0200, Frans Pop wrote:
> > > On Thursday 27 April 2006 12:14, Sven Luther wrote:
> > > > 1) daily build business card and netinst isos are failing to build
> > > > since april 1, which means they don't include the broadcom tg3 module
> > > > at all, and are thus not really usable for installs at this time.
> > >
> > > Reason is that prep and chrp d-i cdrom images are failing in daily d-i
> > > builds. This results in debian-cd not being able to find a file needed
> > > for mkisofs.
> > >
> > > I understand from comments on IRC that this is due to a kernel issue?
> >
> > Oh fun, i don't have any commit access to the d-i repo anymore, so i can't
> > even fix this issue myself. This clearly shows the pettiness of the d-i team,
> > i am disgusted. ...
> >
> > Dear fellow powerpc folk, this clearly means that the debian support for
> > powerpc is dead or almost so, and i strongly recomend you to go find another
> > distribution to run which cares a bit more about the powerpc architecture.
>
> I'd rather prefer if you don't overexegerate.
I'm not sure he's exagerating. There has been a grave silence about the
daily images not building, for anyone who doesn't access IRC anyway.
I would have expected to see some mention of it on the list so people who
don't stay on IRC (and people who are only interested in the ppc architecture)
would know what was going on.
Thanks,
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Sven Luther
> > Oh fun, i don't have any commit access to the d-i repo anymore, so i can't
> > even fix this issue myself. This clearly shows the pettiness of the d-i team,
> > i am disgusted. ...
> >
> > Dear fellow powerpc folk, this clearly means that the debian support for
> > powerpc is dead or almost so, and i strongly recomend you to go find another
> > distribution to run which cares a bit more about the powerpc architecture.
>
> I'd rather prefer if you don't overexegerate.
No i don't think i am overexagerating. The installer images are broken since
over a month, the d-i folk kicked me out after having joined Andres's
wolf-hunt, while i was in personal distress over my mother's sickness and
subsequent death, and now they are failing to take their responsabilities and
are screwing our powerpc users.
They are not worth to be debian developpers, if you would ask me, and in any
case, they are not true to the social contract.
Friendly,
Sven Luther
Andreas Barth
> > > powerpc is dead or almost so, and i strongly recomend you to go find another
> > > distribution to run which cares a bit more about the powerpc architecture.
> >
> > I'd rather prefer if you don't overexegerate.
>
> I'm not sure he's exagerating. There has been a grave silence about the
> daily images not building, for anyone who doesn't access IRC anyway.
I'm not too happy either to become aware of that issue in this way - but
well, I think his cited words "debian support for powerpc is dead or
almost so" and his recommendation people to change away from debian do
way more harm than the images not building the images for a month, and I
also don't really see any ground for these words.
Of course, best that could happen now would be if someone just takes up
the loose ends, and tries to fix the issues - and I hope someone just
does.
> I would have expected to see some mention of it on the list so people who
> don't stay on IRC (and people who are only interested in the ppc architecture)
> would know what was going on.
I never doubted that. :)
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
Sven Luther
> On Thu, 27 Apr 2006, Sven Luther wrote:
>
> > On Thu, Apr 27, 2006 at 07:18:07AM -0400, Toni L. Harbaugh-Blackford [Contr] wrote:
> > > Sven
> > >
> > > I tried this latest netboot64 image, but I get the same behavior as with
> > > the 2006-04-23 image, which I submitted a bug report for (bug 364637).
> > >
> > > The boot hangs at 'Trying to im_free nonexistent area (d0000800820d7000)'
> > >
> > > In the bug report, I included the whole console log.
> > > Should I submit another bug report, or update bug 364637?
> > > Toni
> >
> > What hardware do you have ? I have an XServer dual G5 (RackMac3,1), and i just
> > did an installation without problem, using today's daily build.
>
> Xserve G5 dual 2GHz (RackMac3,1) cluster node, headless
>
> Does the 'Trying to im_free ...' message appear in your console log, and if
> so what is before and after it?
Yep, tree times, and it doesn't seem to hurt.
I do have a pci graphic card in the box, so i suppose what you see is more
probably the broken serial driver i have read about elsewhere.
Mike Hommey
<sven....@wanadoo.fr> wrote:
> I was going to fix it, maybe it would have been fixed already, but no,
> the d-i team decided this otherwise.
You can't *commit*, but you still can send patches, can't you ?
Mike
Sven Luther
> * Toni L. Harbaugh-Blackford [Contr] (harb...@abcc.ncifcrf.gov) [060427 13:53]:
> > On Thu, 27 Apr 2006, Andreas Barth wrote:
> > > * Sven Luther (sven....@wanadoo.fr) [060427 13:23]:
> > > > Dear fellow powerpc folk, this clearly means that the debian support for
> > > > powerpc is dead or almost so, and i strongly recomend you to go find another
> > > > distribution to run which cares a bit more about the powerpc architecture.
> > >
> > > I'd rather prefer if you don't overexegerate.
> >
> > I'm not sure he's exagerating. There has been a grave silence about the
> > daily images not building, for anyone who doesn't access IRC anyway.
>
> I'm not too happy either to become aware of that issue in this way - but
> well, I think his cited words "debian support for powerpc is dead or
> almost so" and his recommendation people to change away from debian do
> way more harm than the images not building the images for a month, and I
> also don't really see any ground for these words.
Because they kicked out the powerpc porter ? Do you not think this is reason
enough to consider debian for powerpc dead, at least with the current set of
d-i team members ?
> Of course, best that could happen now would be if someone just takes up
> the loose ends, and tries to fix the issues - and I hope someone just
> does.
Yeah, problem is we where already there a month ago, and see what happened.
I was going to fix it, maybe it would have been fixed already, but no, the d-i
team decided this otherwise.
Friendly,
Sven Luther
Andreas Barth
> I don't think you understand. There was not a peep on the list about
> anything being up with the daily image until Sven explicitely stated it.
> I would have expected the maintainer or others privy to the situation
> to have said something earlier, along the lines of:
> [...]
Yes, that is what I assumed what happened, and that is what really
worries me. But I don't think it is as bad as Sven tries to make it look
like - it is currently at a level where it could be repaired, and where
we could still prevent that bad things remain.
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
--
To UNSUBSCRIBE, email to debian-powe...@lists.debian.org
Sven Luther
> * Toni L. Harbaugh-Blackford [Contr] (harb...@abcc.ncifcrf.gov) [060427 14:38]:
> > I don't think you understand. There was not a peep on the list about
> > anything being up with the daily image until Sven explicitely stated it.
> > I would have expected the maintainer or others privy to the situation
> > to have said something earlier, along the lines of:
> > [...]
>
> Yes, that is what I assumed what happened, and that is what really
> worries me. But I don't think it is as bad as Sven tries to make it look
> like - it is currently at a level where it could be repaired, and where
> we could still prevent that bad things remain.
Andreas, do you have an explanation of why d-i commit access was taken from
me, and why i find out only now as i was going to fix the issue ?
Does this together with the poor state of the d-i powerpc images not clearly
show that nobody is in charge anymore (or at least nobody who will notice or
have the time to fix it or even ask for help if he doesn't has the time) ?
And in light of those two facts, can you please tell me again with a straight
face that i am over-reacting ?
friendly,
Sven Luther
Sven Luther
> On Thu, 27 Apr 2006, Sven Luther wrote:
>
> > > >
> > > > What hardware do you have ? I have an XServer dual G5 (RackMac3,1), and i just
> > > > did an installation without problem, using today's daily build.
> > >
> > > Xserve G5 dual 2GHz (RackMac3,1) cluster node, headless
> > >
> > > Does the 'Trying to im_free ...' message appear in your console log, and if
> > > so what is before and after it?
> >
> > Yep, tree times, and it doesn't seem to hurt.
> >
> > I do have a pci graphic card in the box, so i suppose what you see is more
> > probably the broken serial driver i have read about elsewhere.
> >
>
>
> Do you have a reference the broken serial driver? Is it specific
> to 2.6.16-1?
I don't remember, now that i think of it, it is possible you who mentioned it.
Can you try booting a d-i image using the net-console trick,and access it with
a ssh connection ? Like they do on s390 ? I did this once a year or so back.
Sven Luther
> On Thu, Apr 27, 2006 at 02:19:34PM +0200, Sven Luther
> <sven....@wanadoo.fr> wrote:
> > I was going to fix it, maybe it would have been fixed already, but no,
> > the d-i team decided this otherwise.
>
> You can't *commit*, but you still can send patches, can't you ?
What good is a patch in the BTS, if there is nobody on the other side to apply
it ?
Still, there is a worse problem here than just svn commit access, and i think
it is my duty as DD to inform the powerpc community of how things stand, since
it is clear nobody else will talk about this.
Friendly,
Sven Luther
Toni L. Harbaugh-Blackford [Contr]
> > >
> > > What hardware do you have ? I have an XServer dual G5 (RackMac3,1), and i just
> > > did an installation without problem, using today's daily build.
> >
> > Xserve G5 dual 2GHz (RackMac3,1) cluster node, headless
> >
> > Does the 'Trying to im_free ...' message appear in your console log, and if
> > so what is before and after it?
>
> Yep, tree times, and it doesn't seem to hurt.
>
> I do have a pci graphic card in the box, so i suppose what you see is more
> probably the broken serial driver i have read about elsewhere.
>
That is probably it, since my box has no graphic card to get past the probes.
Do you have a reference the broken serial driver? Is it specific
to 2.6.16-1?
Thanks
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Toni L. Harbaugh-Blackford [Contr]
> > I would have expected to see some mention of it on the list so people who
> > don't stay on IRC (and people who are only interested in the ppc architecture)
> > would know what was going on.
>
> I never doubted that. :)
>
I don't think you understand. There was not a peep on the list about
anything being up with the daily image until Sven explicitely stated it.
I would have expected the maintainer or others privy to the situation
to have said something earlier, along the lines of:
"Hey folks, we know the daily image is busted and we are
working on it. We'll let you know when it's fixed so
don't bug us".
That nothing was said concerns me, because it implies that either
no one else noticed or no one thought anyone on the list cared, which
is bad in either case.
I don't mean to belabor the point, I just want to have some idea what
is going on from the ppc perspective, since sometimes it is difficult
to glean the basic status from the developer list or the other lists.
Thanks again,
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
--
To UNSUBSCRIBE, email to debian-powe...@lists.debian.org
Mike Hommey
> On Thu, Apr 27, 2006 at 02:27:35PM +0200, Mike Hommey wrote:
> > On Thu, Apr 27, 2006 at 02:19:34PM +0200, Sven Luther
> > <sven....@wanadoo.fr> wrote:
> > > I was going to fix it, maybe it would have been fixed already, but no,
> > > the d-i team decided this otherwise.
> >
> > You can't *commit*, but you still can send patches, can't you ?
>
> What good is a patch in the BTS, if there is nobody on the other side to apply
> it ?
http://lists.debian.org/debian-release/2006/04/msg00178.html
Mike
Toni L. Harbaugh-Blackford [Contr]
> >
> > Do you have a reference the broken serial driver? Is it specific
> > to 2.6.16-1?
>
> I don't remember, now that i think of it, it is possible you who mentioned it.
>
> Can you try booting a d-i image using the net-console trick,and access it with
> a ssh connection ? Like they do on s390 ? I did this once a year or so back.
>
Is this in the debian docs, or is there a reference somewhere to how
to do it? If not, can you just give me a quick overview so I can
figure out how to do it?
I know SuSE provides this via linuxrc, but I was not aware that debian
did something similar.
Thanks,
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Sven Luther
> On Thu, 27 Apr 2006, Sven Luther wrote:
>
> > >
> > > Do you have a reference the broken serial driver? Is it specific
> > > to 2.6.16-1?
> >
> > I don't remember, now that i think of it, it is possible you who mentioned it.
> >
> > Can you try booting a d-i image using the net-console trick,and access it with
> > a ssh connection ? Like they do on s390 ? I did this once a year or so back.
> >
>
> Is this in the debian docs, or is there a reference somewhere to how
> to do it? If not, can you just give me a quick overview so I can
> figure out how to do it?
the debian-installer documentation used to have some (incomplete) guidelines
to do this, not sure where they are now. The s390 installer guide sahould help
you since they have neither display nor serial console.
> I know SuSE provides this via linuxrc, but I was not aware that debian
> did something similar.
Well, it is a d-i feature, you use preseeding together with the ssh remote
console.
Friendly,
Sven Luther
Sven Luther
> On Thu, Apr 27, 2006 at 03:07:36PM +0200, Sven Luther <sven....@wanadoo.fr> wrote:
> > On Thu, Apr 27, 2006 at 02:27:35PM +0200, Mike Hommey wrote:
> > > On Thu, Apr 27, 2006 at 02:19:34PM +0200, Sven Luther
> > > <sven....@wanadoo.fr> wrote:
> > > > I was going to fix it, maybe it would have been fixed already, but no,
> > > > the d-i team decided this otherwise.
> > >
> > > You can't *commit*, but you still can send patches, can't you ?
> >
> > What good is a patch in the BTS, if there is nobody on the other side to apply
> > it ?
>
> http://lists.debian.org/debian-release/2006/04/msg00178.html
Why does this come now only after the brokeness was there for almost 4 weeks,
and activity only happened after i made a fuss over it ? This was already the
case with the yaird problem, and going a bit farther away, Ethan's rejection
of the yaboot/amiga-partition-table patch.
And notice, yaboot is currently orphaned, and nobody has shown up and took it
over, mol is currently in sad dissaray, and people go about and expulse and
kick the few powerpc porters we have remaining.
So, things clearly look grim for debian/powerpc.
Friendly,
Sven Luther
Toni L. Harbaugh-Blackford [Contr]
> >
> > Is this in the debian docs, or is there a reference somewhere to how
> > to do it? If not, can you just give me a quick overview so I can
> > figure out how to do it?
>
> the debian-installer documentation used to have some (incomplete) guidelines
> to do this, not sure where they are now. The s390 installer guide sahould help
> you since they have neither display nor serial console.
>
> > I know SuSE provides this via linuxrc, but I was not aware that debian
> > did something similar.
>
> Well, it is a d-i feature, you use preseeding together with the ssh remote
> console.
>
I don't see any mention of this in the s390 manual. The only description
that seems relevant indicates that the installation system has to successfully
boot before the network interface can be configured, which does not
happen in this case.
What is supposed to happen after 'Trying to im_free...' ?
Thanks,
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Andreas Barth
> Andreas, do you have an explanation of why d-i commit access was taken from
> me, and why i find out only now as i was going to fix the issue ?
http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
you stepped back, and
http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
that.
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
Sven Luther
> * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > Andreas, do you have an explanation of why d-i commit access was taken from
> > me, and why i find out only now as i was going to fix the issue ?
>
> http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> you stepped back, and
> http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> that.
I stepped back from d-i powerpc port maintainer, not as d-i contributor.
Also, i don't believe there is any justification for taking away svn commit
access except when there was a clear misuse of it being made.
So, basically, you are saying that it is ok to take commit right away because
of personal dislike situations like apparently happened here ?
Again, i ask you, what possible reason where there to take svn commit access
away from me ?
Friendly,
Sven Luther
Rene Engelhard
Sven Luther wrote:
> On Thu, Apr 27, 2006 at 04:52:09PM +0200, Andreas Barth wrote:
> > * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > > Andreas, do you have an explanation of why d-i commit access was taken from
> > > me, and why i find out only now as i was going to fix the issue ?
> >
> > http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> > you stepped back, and
> > http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> > that.
>
> I stepped back from d-i powerpc port maintainer, not as d-i contributor.
So there apparently was a misunderstanding.
> Also, i don't believe there is any justification for taking away svn commit
> access except when there was a clear misuse of it being made.
>
> So, basically, you are saying that it is ok to take commit right away because
> of personal dislike situations like apparently happened here ?
It is OK to remove svn access for a person which stepped down from d-i
activity (see above) imho, yes.
> Again, i ask you, what possible reason where there to take svn commit access
> away from me ?
See above.
Regards,
Rene
Sven Luther
> Hi,
>
> Sven Luther wrote:
> > On Thu, Apr 27, 2006 at 04:52:09PM +0200, Andreas Barth wrote:
> > > * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > > > Andreas, do you have an explanation of why d-i commit access was taken from
> > > > me, and why i find out only now as i was going to fix the issue ?
> > >
> > > http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> > > you stepped back, and
> > > http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> > > that.
> >
> > I stepped back from d-i powerpc port maintainer, not as d-i contributor.
>
> So there apparently was a misunderstanding.
No, there was a deliberate intention for revenge and hurting. I doubt any of
the other d-i contributors commit right where removed when they stopped
contributing, and there must be over a 100 such dormant contributors by now.
> > Also, i don't believe there is any justification for taking away svn commit
> > access except when there was a clear misuse of it being made.
> >
> > So, basically, you are saying that it is ok to take commit right away because
> > of personal dislike situations like apparently happened here ?
>
> It is OK to remove svn access for a person which stepped down from d-i
> activity (see above) imho, yes.
who was kicked out you mean ?
> > Again, i ask you, what possible reason where there to take svn commit access
> > away from me ?
>
> See above.
That is no reply. I don't believe such was done in any of the other previous
cases, at least not without months of waiting and at least a ping tentative.
So tell me, could this really be interpreted as anythind else than a tentative
to get ride of me ?
Friendly,
Sven Luther
Rene Engelhard
> On Thu, Apr 27, 2006 at 05:10:17PM +0200, Rene Engelhard wrote:
> > Hi,
> >
> > Sven Luther wrote:
> > > On Thu, Apr 27, 2006 at 04:52:09PM +0200, Andreas Barth wrote:
> > > > * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > > > > Andreas, do you have an explanation of why d-i commit access was taken from
> > > > > me, and why i find out only now as i was going to fix the issue ?
> > > >
> > > > http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> > > > you stepped back, and
> > > > http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> > > > that.
> > >
> > > I stepped back from d-i powerpc port maintainer, not as d-i contributor.
> >
> > So there apparently was a misunderstanding.
> who was kicked out you mean ?
[...]
Please decide.
Regards,
Rene
Sven Luther
> Sven Luther wrote:
> > On Thu, Apr 27, 2006 at 05:10:17PM +0200, Rene Engelhard wrote:
> > > Hi,
> > >
> > > Sven Luther wrote:
> > > > On Thu, Apr 27, 2006 at 04:52:09PM +0200, Andreas Barth wrote:
> > > > > * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > > > > > Andreas, do you have an explanation of why d-i commit access was taken from
> > > > > > me, and why i find out only now as i was going to fix the issue ?
> > > > >
> > > > > http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> > > > > you stepped back, and
> > > > > http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> > > > > that.
> > > >
> > > > I stepped back from d-i powerpc port maintainer, not as d-i contributor.
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > >
> > > So there apparently was a misunderstanding.
> [...]
> > who was kicked out you mean ?
> ^^^^^^^^^^
> [...]
>
> Please decide.
I was kicked out. If you reread the message, i told that i not interested in
continuing being the powerpc porter, just so franz could blame me for every
time, even if i tried to help as best i could as i did in the parent post to
those you quoted (which was when i was with my dying mother, and Frans
perfectly knew that). So i said i was not interested to continue this role
unless those d-i team member who believe debian would be better off without
me, and that i was no irreplacable, and who happily joined in in my expulsion
request, at least apologized for the hurt they did me during the expulsion
request.
Less than a few minutes after this came Franz mail officializing my expulsion
from the d-i team, and searching for contributors, contributors who failed to
show up since then i notice, and a few days after the powerpc d-i cds broke.
So, please tell me how this whole process can be seen as anything else but
kicking me out, or how it could be called a 'misunderstanding' as some did ?
Friendly,
Sven Luther
Sven Luther
> Just what are the rules for someone to have commit access?
None, it is the full decision of the project admin, and i believe what
happened here is that one such project admin did let some petty personal
considerations overstep his responsabilities.
Friendly,
Sven Luther
Toni L. Harbaugh-Blackford [Contr]
> > > >
> > > > I stepped back from d-i powerpc port maintainer, not as d-i contributor.
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > >
> > > So there apparently was a misunderstanding.
> [...]
> > who was kicked out you mean ?
> ^^^^^^^^^^
> [...]
>
> Please decide.
>
> Regards,
>
> Rene
So only maintainers have d-i commit access? This would seem to put
a lot of extra work on the maintainer.
Just what are the rules for someone to have commit access?
Regards,
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Olaf van der Spek
> On Thu, Apr 27, 2006 at 11:43:02AM -0400, Toni L. Harbaugh-Blackford [Contr] wrote:
> > Just what are the rules for someone to have commit access?
>
> None, it is the full decision of the project admin, and i believe what
> happened here is that one such project admin did let some petty personal
> considerations overstep his responsabilities.
Why don't you ask the admins instead of continuing based on beliefs?
Joey Hess
> anything being up with the daily image until Sven explicitely stated it.
All of us who read http://people.debian.org/~joeyh/d-i/build-logs.html
on a daily basis knew that the powerpc builds had not updated in a
while, but given that
a) They're marked as "currently on a laptop so not cronned, should be
moved soon"
b) The mips builds had been down longer until recently.
c) The mipsel builds had been down nearly as long. And that machine
seems to be dead. Argh.
d) The i386 floppy builds had been broken for longer until recently.
e) The amd64 builds have been down just as long and still are.
f) The arm CD builds have been broken for quite a while.
It wasn't very clear, to me at least, that anything was exceptionally
wrong with the powerpc builds.
> I would have expected the maintainer or others privy to the situation
> to have said something earlier, along the lines of:
>
> "Hey folks, we know the daily image is busted and we are
> working on it. We'll let you know when it's fixed so
> don't bug us".
We have an automated page to track the status of the builds so that we
don't have to waste everyone's time doing that on a continual basis.
--
see shy jo
Frans Pop
On Thursday 27 April 2006 17:47, Sven Luther wrote:
> None, it is the full decision of the project admin, and i believe what
> happened here is that one such project admin did let some petty
> personal considerations overstep his responsabilities.
Yep, and the "admin" feels completely justified by this ridiculous
escalation of a minor issue [0].
The commit access was revoked based on the fact Sven resigned but that
decision was influenced by the circumstances in which that happened. Not
only I, but several members of the d-i team, have long had issues working
with Sven and basically we decided enough was enough.
On Thursday 27 April 2006 17:43, Toni L. Harbaugh-Blackford [Contr] wrote:
> So only maintainers have d-i commit access? This would seem to put
> a lot of extra work on the maintainer.
>
> Just what are the rules for someone to have commit access?
No, anybody, both Debian Developers and non-Developers can get commit
access to the d-i SVN repository (actually there are 140 people with
access in total of whom about 65% are non-DDs, a lot of them
translators), and in general we are quite quick to grant access and very
slow to revoke it.
We are still [1] very interested in welcoming other powerpc people to help
with support for powerpc in d-i. We _do_ need porter help to be able to
continue support for an architecture, especially one as complex as
powerpc (with several port-specific subarches, bootloaders and
filesystems, non-free issues).
Without your (powerpc community, not Sven) help support for subarches like
apus, prep and chrp may loose their support in d-i. The issues that are
listed in [2] are still there (except for the missing tg3 driver) and
we'd dearly like to see those worked on.
Cheers,
Frans Pop
P.S. This is the only mail I'm going to waste on this issue.
P.P.S. Note that the daily builds that Sven is still running should not be
used. Instead, please use the builds lined from [3].
P.P.P.S. Just for the record: I did _not_ support the expulsion request
that was started a while back against Sven and have said so publically,
although I very much did and do understand the reasons behind it.
[0] Note: we are only talking about daily and weekly _development_ builds
here, not about Sarge CDs or even Etch Beta2 CDs. Installing powerpc is
still very much possible.
[1] http://lists.debian.org/debian-powerpc/2006/03/msg00490.html
[2] http://www.debian.org/devel/debian-installer/errata
[3] http://www.debian.org/devel/debian-installer/
Michael Banck
> On Thu, 27 Apr 2006, Rene Engelhard wrote:
>
> > > > >
> > > > > I stepped back from d-i powerpc port maintainer, not as d-i contributor.
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > >
> > > > So there apparently was a misunderstanding.
> > [...]
> > > who was kicked out you mean ?
> > ^^^^^^^^^^
> > [...]
> >
> > Please decide.
> >
> > Regards,
> >
> > Rene
>
> So only maintainers have d-i commit access?
No, people who are part of the d-i team have commit access.
When Sven stepped down as the powerpc maintainer, the d-i project admins
decided that commit access is no longer needed for him, assuming he did
not need commit access for the core parts, but could just send patches
members of the team would apply. This looks quite reasonable to me and
the project admins are totally fine to do this.
This is just an educated guess of course.
> Just what are the rules for someone to have commit access?
This question is off topic for debian-devel (and probably debian-powerpc
as well).
Michael
--
"I'm glad that we're compatible, mature and grown,
'Cause this is not something U can do alone"
-- Prince, "Incense and Candles"
Sven Luther
> (dropping the ridiculous CC list, AFAICT everybody is subscribed)
>
> On Thursday 27 April 2006 17:47, Sven Luther wrote:
> > None, it is the full decision of the project admin, and i believe what
> > happened here is that one such project admin did let some petty
> > personal considerations overstep his responsabilities.
>
> Yep, and the "admin" feels completely justified by this ridiculous
> escalation of a minor issue [0].
>
> The commit access was revoked based on the fact Sven resigned but that
> decision was influenced by the circumstances in which that happened. Not
> only I, but several members of the d-i team, have long had issues working
> with Sven and basically we decided enough was enough.
Then live up to your responsability and don't let the powerpc port break
again, because i promise you i will be watching and remembering the world of
your failure if that happens.
This is the price you pay for kicking people out, after doing your best for
hurting them in the first place.
Sven Luther
> Toni L. Harbaugh-Blackford [Contr] wrote:
> > I don't think you understand. There was not a peep on the list about
> > anything being up with the daily image until Sven explicitely stated it.
>
> All of us who read http://people.debian.org/~joeyh/d-i/build-logs.html
> on a daily basis knew that the powerpc builds had not updated in a
Ah, yeah, and when where you going to fix this ? And when where you going to
inform the powerpc users ?
> while, but given that
>
> a) They're marked as "currently on a laptop so not cronned, should be
> moved soon"
That is no excuse, since you kicked me out, and where very quick to throw
away the buildd i maintained, just to get ride of me, without having proper
replacement.
Notice also that both you and Colin Watson, where donated pegasos machines,
(and guess who arranged that), so the unavailability of a decent build machine
is no excuse. At worse you could have used my buildd until a decent solution
was found, but no, it was more important to get ride of sven, than to care for
the powerpc users.
> b) The mips builds had been down longer until recently.
> c) The mipsel builds had been down nearly as long. And that machine
> seems to be dead. Argh.
You didn't kick the mips/mipsel maintainers out of the d-i team though, so
your responsability is not engaged.
> d) The i386 floppy builds had been broken for longer until recently.
> e) The amd64 builds have been down just as long and still are.
> f) The arm CD builds have been broken for quite a while.
Same here.
> It wasn't very clear, to me at least, that anything was exceptionally
> wrong with the powerpc builds.
There is something exceptionally wrong, and this is the unprecedented fact
that you kicked the powerpc maintainer out of the project, with uther contempt
for all the work i did do in the past for this, promised you had found a
replacement, and that the port would be well taken care off, and failed in
this.
> > I would have expected the maintainer or others privy to the situation
> > to have said something earlier, along the lines of:
> >
> > "Hey folks, we know the daily image is busted and we are
> > working on it. We'll let you know when it's fixed so
> > don't bug us".
>
> We have an automated page to track the status of the builds so that we
> don't have to waste everyone's time doing that on a continual basis.
Yeah, but you are to chicken to take the next logical step, namely to have a
mailing list report scheme or something to make people aware of this who don't
necessarily have time to daily look at the build log web page. But then, i
suppose you prefer bugzilla over our BTS too.
This is just a bunch of excuses, you expulsed me, and then the build failed,
and it was gone almost 4 weeks without being fixed. You are fully responsible
for this failure, and should assume your responsability.
But then i know already that you are not going to assume it, and like always
just do on more round of sven-bashing, and hope everything will be fine.
Sven Luther
Sven Luther
> On Thu, 27 Apr 2006, Sven Luther wrote:
>
> > >
> > > Is this in the debian docs, or is there a reference somewhere to how
> > > to do it? If not, can you just give me a quick overview so I can
> > > figure out how to do it?
> >
> > the debian-installer documentation used to have some (incomplete) guidelines
> > to do this, not sure where they are now. The s390 installer guide sahould help
> > you since they have neither display nor serial console.
> >
> > > I know SuSE provides this via linuxrc, but I was not aware that debian
> > > did something similar.
> >
> > Well, it is a d-i feature, you use preseeding together with the ssh remote
> > console.
> >
>
> I don't see any mention of this in the s390 manual. The only description
> that seems relevant indicates that the installation system has to successfully
> boot before the network interface can be configured, which does not
> happen in this case.
Ask the d-i folk, they should know. look at the preseed stuff in the
documentation or something too, that is what i did.
>
> What is supposed to happen after 'Trying to im_free...' ?
Attaching the full log. The isa stuff is probably related to the graphic card.
Friendly,
Sven Luther
Andreas Barth
> On Thu, Apr 27, 2006 at 04:52:09PM +0200, Andreas Barth wrote:
> > * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > > Andreas, do you have an explanation of why d-i commit access was taken from
> > > me, and why i find out only now as i was going to fix the issue ?
> >
> > http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> > you stepped back, and
> > http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> > that.
>
> I stepped back from d-i powerpc port maintainer, not as d-i contributor.
I don't call myself authoritative what your other contributions to d-i
include. It might be a good idea to clarify your status within the d-i
team first, instead of telling all people that debian has stopped ppc
support (which is just wrong, and is definitly neither the intention of
the d-i team nor the release team).
> Also, i don't believe there is any justification for taking away svn commit
> access except when there was a clear misuse of it being made.
Actually, I think revoking svn commit access from people who stopped to
work on d-i (or any project, that is not d-i specific) is ok. E.g. we
also "revoke" upload privileges of people who don't do Debian work
anymore. If that was a misunderstanding - please try to clarify it in
private first (without the noise that this thread makes).
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
Andreas Barth
Hey, that would be violating the rules of a flame fest. :)
Cheers,
Andi
--
http://home.arcor.de/andreas-barth/
--
To UNSUBSCRIBE, email to debian-dev...@lists.debian.org
Brad Boyer
> > What is supposed to happen after 'Trying to im_free...' ?
>
> Attaching the full log. The isa stuff is probably related to the graphic card.
>
> Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled
> pmac_zilog: 0.6 (Benjamin Herrenschmidt <be...@kernel.crashing.org>)
> pmac_zilog: Error registering serial device, disabling pmac_zilog.
> pmac_zilog: Did another serial driver already claim the minors?
> Trying to im_free nonexistent area (d00008008216c000)
> Trying to im_free nonexistent area (d00008008216e000)
Here is the problem. The 8250/16550 driver used to have a check to make
it bail out on all PowerMac models. It used to print out a message to
that fact, too. The pmac_zilog driver is dying a miserable death because
it thinks it should be the only serial driver, but in this case it isn't.
Brad Boyer
fl...@allandria.com
Sven Luther
> * Sven Luther (sven....@wanadoo.fr) [060427 17:03]:
> > On Thu, Apr 27, 2006 at 04:52:09PM +0200, Andreas Barth wrote:
> > > * Sven Luther (sven....@wanadoo.fr) [060427 15:05]:
> > > > Andreas, do you have an explanation of why d-i commit access was taken from
> > > > me, and why i find out only now as i was going to fix the issue ?
> > >
> > > http://lists.debian.org/debian-boot/2006/03/msg01075.html sounds like
> > > you stepped back, and
> > > http://lists.debian.org/debian-powerpc/2006/03/msg00490.html confirms
> > > that.
> >
> > I stepped back from d-i powerpc port maintainer, not as d-i contributor.
>
> I don't call myself authoritative what your other contributions to d-i
The maintainer is held responsible (and frans and joeyh have not stepped down
from reminding me of this in the past) of the build failure, while a
contributor is free commit fixes, without necessarily being the one to blame
for every problem of the port.
> include. It might be a good idea to clarify your status within the d-i
> team first, instead of telling all people that debian has stopped ppc
> support (which is just wrong, and is definitly neither the intention of
> the d-i team nor the release team).
Well, it is what i see. There is nothing to clarify on this point. They
removed me to get ride of me, and despite Colin's help, they don't really have
someone to make the real work.
This is actually the real problem of the d-i team, they lack manpower to
properly care about d-i as it deserves, and this makes them irritable.
> > Also, i don't believe there is any justification for taking away svn commit
> > access except when there was a clear misuse of it being made.
>
> Actually, I think revoking svn commit access from people who stopped to
> work on d-i (or any project, that is not d-i specific) is ok. E.g. we
> also "revoke" upload privileges of people who don't do Debian work
> anymore. If that was a misunderstanding - please try to clarify it in
> private first (without the noise that this thread makes).
After having read the replies of both Joeyh and fjp, i don't believe it was a
misunderstanding, but a deliberate decision of the d-i team. If you read
something else out of it, i would be very interested in you interpretation of
it.
Friendly,
Sven Luther
Sven Luther
> On Thu, Apr 27, 2006 at 09:11:32PM +0200, Sven Luther wrote:
> > > What is supposed to happen after 'Trying to im_free...' ?
> >
> > Attaching the full log. The isa stuff is probably related to the graphic card.
> >
>
> > Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled
> > pmac_zilog: 0.6 (Benjamin Herrenschmidt <be...@kernel.crashing.org>)
> > pmac_zilog: Error registering serial device, disabling pmac_zilog.
> > pmac_zilog: Did another serial driver already claim the minors?
> > Trying to im_free nonexistent area (d00008008216c000)
> > Trying to im_free nonexistent area (d00008008216e000)
>
> Here is the problem. The 8250/16550 driver used to have a check to make
> it bail out on all PowerMac models. It used to print out a message to
> that fact, too. The pmac_zilog driver is dying a miserable death because
> it thinks it should be the only serial driver, but in this case it isn't.
Oh well, i know what this problem is then. We had a patch to stop the
8250/16550 driver from loading on arches it had nothing to do (since it used
to poke random pmac register and die), i guess this one was not ported over to
the 2.6.16 arch=powerpc tree.
Investigation is needed here.
Friendly,
Sven Luther
Steve Langasek
> The maintainer is held responsible (and frans and joeyh have not stepped down
> from reminding me of this in the past) of the build failure, while a
> contributor is free commit fixes, without necessarily being the one to blame
> for every problem of the port.
And whereas when the alpha daily builds are broken and require build env
updates to get them working again I simply fix them at my earliest
opportunity, you invariably used this as an excuse to accuse the rest of the
d-i team of misconduct or negligence. Your technical skills and committment
to powerpc are valuable traits, but the only thing you bring to the table
that's irreplaceable is your penchant for vitriol, and I'm quite sure
everyone involved would be happy to be rid of that.
It's unfortunate that even your resignation as d-i porter doesn't spare the
rest of the d-i team from having their time wasted by threads like this.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
vor...@debian.org http://www.debian.org/
Wolfgang Pfeiffer
> Well, it is what i see. There is nothing to clarify on this
> point. They removed me to get ride of me, and despite Colin's help,
> they don't really have someone to make the real work.
So Colin's help isn't real, or what?
Please don't get me wrong: I definitely appreciate the outstanding
entertainment value of this thread ... But please make sure *everyone*
can laugh ...
Regards
Wolfgang
--
Wolfgang Pfeiffer: /ICQ: 286585973/ + + + /AIM: crashinglinux/
http://profiles.yahoo.com/wolfgangpfeiffer
Key ID: E3037113
http://keyserver.mine.nu/pks/lookup?search=0xE3037113&fingerprint=on
Sven Luther
> On Thu, Apr 27, 2006 at 10:46:24PM +0200, Sven Luther wrote:
> > The maintainer is held responsible (and frans and joeyh have not stepped down
> > from reminding me of this in the past) of the build failure, while a
> > contributor is free commit fixes, without necessarily being the one to blame
> > for every problem of the port.
>
> And whereas when the alpha daily builds are broken and require build env
> updates to get them working again I simply fix them at my earliest
> opportunity, you invariably used this as an excuse to accuse the rest of the
I don't remember it such, i remember frans accusing me of negligence and
misconduct because i did give a (maybe a bit uninformed) advice to a powerpc
user.
> d-i team of misconduct or negligence. Your technical skills and committment
> to powerpc are valuable traits, but the only thing you bring to the table
> that's irreplaceable is your penchant for vitriol, and I'm quite sure
> everyone involved would be happy to be rid of that.
I am most assuredly not the only one in debian with this trait, who will you
get ride of next ?
> It's unfortunate that even your resignation as d-i porter doesn't spare the
> rest of the d-i team from having their time wasted by threads like this.
Oh, thanks. so you also believe that the removal of my d-i commit rights was
warranted.
Could you please explain this in the open, and not in this cabal like fasion ?
(22:36:45)< vorlon> fjp: can you speak to why svenl's commit access to d-i was
revoked? I vaguely remember a clean-up of unused d-i accounts, but I thought
that only covered accounts that had been unused for some time.
(22:37:07)< fjp> vorlon: I'd prefer /msg
So, why was i not informed of that fact ? And why is the powerpc port broken
since 4 weeks without nobody noticing ?
Friendly,
Sven Luther
--
To UNSUBSCRIBE, email to debian-dev...@lists.debian.org
Sven Luther
> On Thu, Apr 27, 2006 at 10:46:24PM +0200, Sven Luther wrote:
>
> > Well, it is what i see. There is nothing to clarify on this
> > point. They removed me to get ride of me, and despite Colin's help,
> > they don't really have someone to make the real work.
>
> So Colin's help isn't real, or what?
Sure it is, if it confirms itself that he has enough time to fullfill it. The
reality is that the powerpc d-i port was broken since *4 WEEKS*, and nobody
noticed, and nobody informed the powerpc users who where complaining and
wondering.
> Please don't get me wrong: I definitely appreciate the outstanding
> entertainment value of this thread ... But please make sure *everyone*
> can laugh ...
At my depends, right ?
And when i am gone, who will be the next scape-goat ?
Friendly,
Sven Luther
Steve Langasek
> On Thu, Apr 27, 2006 at 03:34:30PM -0700, Steve Langasek wrote:
> > On Thu, Apr 27, 2006 at 10:46:24PM +0200, Sven Luther wrote:
> > > The maintainer is held responsible (and frans and joeyh have not stepped down
> > > from reminding me of this in the past) of the build failure, while a
> > > contributor is free commit fixes, without necessarily being the one to blame
> > > for every problem of the port.
> > And whereas when the alpha daily builds are broken and require build env
> > updates to get them working again I simply fix them at my earliest
> > opportunity, you invariably used this as an excuse to accuse the rest of the
> I don't remember it such, i remember frans accusing me of negligence and
> misconduct because i did give a (maybe a bit uninformed) advice to a powerpc
> user.
You had quite strong words for Frans, accusing him (and other d-i folk) of
breaking your daily builds. They didn't do anything of the sort; *I* broke
the daily builds, because there was a libnewt soname bump and udebs needed
to be rebuilt against the new soname, which broke the daily builds until
libnewt0.52 was installed in the build env because this was before we had
support for udeb shlibs. That didn't stop you from accusing Frans of first
breaking the build and then picking on you.
> > It's unfortunate that even your resignation as d-i porter doesn't spare the
> > rest of the d-i team from having their time wasted by threads like this.
> Oh, thanks. so you also believe that the removal of my d-i commit rights was
> warranted.
As discussed on IRC, yes, I believe the d-i repo admins have the authority
to remove the d-i commit rights of committers who have resigned, or
committers that they believe are abusive, or committers who have idled out,
and probably the authority to remove commit rights for other reasons I'm not
thinking of right now.
> Could you please explain this in the open, and not in this cabal like
> fasion ?
> (22:36:45)< vorlon> fjp: can you speak to why svenl's commit access to d-i was
> revoked? I vaguely remember a clean-up of unused d-i accounts, but I thought
> that only covered accounts that had been unused for some time.
> (22:37:07)< fjp> vorlon: I'd prefer /msg
No, it's up to fjp to decide if he wants to say more than he already has;
though I think at the time I made that comment on IRC, he had already posted
a reply to this thread, which simply had not reached my inbox yet.
Sven Luther
> On Fri, Apr 28, 2006 at 12:40:48AM +0200, Sven Luther wrote:
> > On Thu, Apr 27, 2006 at 03:34:30PM -0700, Steve Langasek wrote:
> > > On Thu, Apr 27, 2006 at 10:46:24PM +0200, Sven Luther wrote:
> > > > The maintainer is held responsible (and frans and joeyh have not stepped down
> > > > from reminding me of this in the past) of the build failure, while a
> > > > contributor is free commit fixes, without necessarily being the one to blame
> > > > for every problem of the port.
>
> > > And whereas when the alpha daily builds are broken and require build env
> > > updates to get them working again I simply fix them at my earliest
> > > opportunity, you invariably used this as an excuse to accuse the rest of the
>
> > I don't remember it such, i remember frans accusing me of negligence and
> > misconduct because i did give a (maybe a bit uninformed) advice to a powerpc
> > user.
>
> You had quite strong words for Frans, accusing him (and other d-i folk) of
> breaking your daily builds. They didn't do anything of the sort; *I* broke
> the daily builds, because there was a libnewt soname bump and udebs needed
> to be rebuilt against the new soname, which broke the daily builds until
> libnewt0.52 was installed in the build env because this was before we had
> support for udeb shlibs. That didn't stop you from accusing Frans of first
> breaking the build and then picking on you.
I didn't accuse Frans, i did accuse the d-i folk in general. English has this
poor feature of not distinguishing between the polite you, the singular you
and the plural you, maybe we should switch to a more advanced language instead
:)
Now, putting things in context, some user complained about brokeness on
debian-powerpc. I know that appart from me, nobody is reading debian-powerpc
from the d-i team, so i told Shaymal that he should post on debian-boot
instead, or better file a bug report directly, since i was hardly available to
do real work, and to do the bridge between debian-powerpc and d-i at that
time. I was away in el salvador at the time, my mother had just passed a
severe respipratory crisis a few hours before, and i had gone to read
debian-powerpc in order to change my mind a bit, and being me, i could not
help trying to be helpful to users, even though i didn't really take the time
to investigate fully, and may have made a mistake, but given my situation, you
have to admit that this is understandable, no ?
As a result, i got an immediate response from Frans, not only telling me i was
wrong, and that the i failed miserably to keep the daily builds going, but
also adding that little bashing paragraph, the kind that frans has been giving
to me with various degrees of subtletly since over 8 month now or so, and
which yourself agreed yesterday was not correct.
Given this, two things happened. I wrote Frans a personal mail asking him for
comprehension, and kind of explaining my personal situation, which i don't
really feel he has acted upon, and second i was pretty much feedup that even
in the situation i found myself, there was nobody who would take care of
either fixing this issue, or at least inform the users that it was a known
problem, that i was currently unavailable for severe personal reasons, and it
would be fixed soon.
Seeing things in that light, and given of what Frans did know at that time, is
there still any doubt left that the removal of my commit access was nothing
more that an unfeeling attempt to get ride of me, and that the resignation
letter is nothing but an excuse ? I was also told a bit before this events (on
irc and i saddly don't kep logs) that some people didn't really want (after
the expulsion event) for me to make the effort to come back to debian, and
would be happy to be ride of me. I wasn't told who those people are, but given
these events, one can guess.
Now, the critic i have is of another kind, and one i have done repeteadly in
the past, and for which the d-i team had marked me as someone to bash at will.
The problem here is very speaking, As you say, something happened, so the
build broke. The build breaking is listed on joeyh's web page, but depending
on folk, browsing a web page daily is a poor substitute for email
notification, and i guess we all agree on this, or we would be using bugzilla
over our BTS :). So, it broke, some folk noticed this, and fixed their daily
builds (joeyh or whoever maintains the x86 daily build among them), but nobody
informed the other daily build maintainers, so each one would have to discover
the issue alone, investigate the problem and do the fix. This i believe is not
efficient, and i said so, and something which is mirrored in the way the
kernel .udebs are handled, and i have said so in the past, tried to start a
discussion to get more efficiency into this, and proposed some possible
solutions. At the same time, Joey was repeteadly blaming the lazy porters for
for the d-i state and stuff like that.
This exact issue, is why i believe that some of the d-i team have marked me as
someone to eliminate or whatever, this is the technical reason i spoke about,
and altough i may have not been the most clever in handling this, you can look
over the email archive and see it for yourself.
> > > It's unfortunate that even your resignation as d-i porter doesn't spare the
> > > rest of the d-i team from having their time wasted by threads like this.
>
> > Oh, thanks. so you also believe that the removal of my d-i commit rights was
> > warranted.
>
> As discussed on IRC, yes, I believe the d-i repo admins have the authority
> to remove the d-i commit rights of committers who have resigned, or
> committers that they believe are abusive, or committers who have idled out,
> and probably the authority to remove commit rights for other reasons I'm not
> thinking of right now.
The question is different though. Do you believe that an alioth project admin
has the right to use personal opinion in order to do actions which potentially
hurt the project, and given the above, as well as the personal mail i
forwarded you, do you really believe that this action is warranted in this
case.
Do you believe also that there are different categories of DDs ? Some with
more rights than others ? This is indeed a very serious question, as when i
joined, it was clear that every DD was equal.
> > Could you please explain this in the open, and not in this cabal like
> > fasion ?
>
> > (22:36:45)< vorlon> fjp: can you speak to why svenl's commit access to d-i was
> > revoked? I vaguely remember a clean-up of unused d-i accounts, but I thought
> > that only covered accounts that had been unused for some time.
> > (22:37:07)< fjp> vorlon: I'd prefer /msg
>
> No, it's up to fjp to decide if he wants to say more than he already has;
> though I think at the time I made that comment on IRC, he had already posted
> a reply to this thread, which simply had not reached my inbox yet.
Ok, fine, i will let the issue in the hand of the DPL and his delegate on this
matter. All i ask is for bystander to be honest with themselves given the
facts i exposed here and elsewhere. I have the feeling that this was not the
case upto now though. And some would have reacted differently if the persons
involved where other folk.
Daniel Stone
> Notice also that both you and Colin Watson, where donated pegasos machines,
> (and guess who arranged that), so the unavailability of a decent build machine
> is no excuse.
I can't speak for the other guys, but I have a Pegasos machine (sitting
under my desk at the moment, actually), sent to me by yourself. You
offered it to me when I told you that I had no powerpc machines, and
thus couldn't test X with PowerPC. I made it very, very clear to you
that I could not guarantee that the machine would ever get turned on,
let alone used productively. Repeatedly. You said that was fine.
However, you then got upset when Pegasos support lapsed, and ripped into
me for not doing enough to fix it, given that you sent me an ODW. So, I
can't help but think, maybe this is another case where people explicitly
told you that they couldn't ensure the machine was used productively,
but you still got upset when it wasn't?
Sven Luther
> On Thu, Apr 27, 2006 at 08:28:11PM +0200, Sven Luther wrote:
> > Notice also that both you and Colin Watson, where donated pegasos machines,
> > (and guess who arranged that), so the unavailability of a decent build machine
> > is no excuse.
>
> I can't speak for the other guys, but I have a Pegasos machine (sitting
> under my desk at the moment, actually), sent to me by yourself. You
> offered it to me when I told you that I had no powerpc machines, and
> thus couldn't test X with PowerPC. I made it very, very clear to you
> that I could not guarantee that the machine would ever get turned on,
> let alone used productively. Repeatedly. You said that was fine.
Indeed.
> However, you then got upset when Pegasos support lapsed, and ripped into
> me for not doing enough to fix it, given that you sent me an ODW. So, I
Notice that first, this was an ubuntu matter, and i got upset, not because you
didn't get the pegasos/X support fixed in a timely way, but because my patches
sent to the ubuntu BTS where coldly received, and even ignored until i pushed
some.
> can't help but think, maybe this is another case where people explicitly
> told you that they couldn't ensure the machine was used productively,
> but you still got upset when it wasn't?
Joey Hess received a machine in order to do daily d-i powerpc tests. This was
the exact wording, and it was an additional machine outside of the first
donation. I know since then that his machine broke, for whatever reason, some
time ago, but he never informed me of this (we would have replaced it), and i
only knew about this in Helsinki, and if he would have told me about this a
week before, i would have swapped it out. The 4 machines i had in Heslinki for
debconf admin purpose where all donated to to debian or skolelinux purpose,
except the one where i burned the power supply, and supposedly the debconf
assurance would have covered.
So, with all that said, do you still believe it is normal that a perfectly
running daily build was rejected in maybe a few minutes/hours after i sent
that email, while i had offered to continue running it until a proper
replacement was found, and some unstable solution has been used ever since,
which doesn't even include to this day the miboot support ?
Friendly,
Daniel Stone
> So, with all that said, do you still believe it is normal that a perfectly
> running daily build was rejected in maybe a few minutes/hours after i sent
> that email, while i had offered to continue running it until a proper
> replacement was found, and some unstable solution has been used ever since,
> which doesn't even include to this day the miboot support ?
If you're going to attempt to drag other people into your petty personal
tiffs, you might as well at least try to rope in people who are
sympathetic to your cause.
Sven Luther
Please tell me then, what interest is your post to the issue at hand ? Or does
this mean, that whatever i say, you would still be against me, because you are
'unsympathetic to my cause' ?
I also don't see you aknowledging the correction on the misleading information
you mentioned in your first post.
Toni L. Harbaugh-Blackford [Contr]
> On Thu, Apr 27, 2006 at 09:11:32PM +0200, Sven Luther wrote:
> > > What is supposed to happen after 'Trying to im_free...' ?
> >
> > Attaching the full log. The isa stuff is probably related to the graphic card.
> >
>
> > Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled
> > pmac_zilog: 0.6 (Benjamin Herrenschmidt <be...@kernel.crashing.org>)
> > pmac_zilog: Error registering serial device, disabling pmac_zilog.
> > pmac_zilog: Did another serial driver already claim the minors?
> > Trying to im_free nonexistent area (d00008008216c000)
> > Trying to im_free nonexistent area (d00008008216e000)
>
> Here is the problem. The 8250/16550 driver used to have a check to make
> it bail out on all PowerMac models. It used to print out a message to
> that fact, too. The pmac_zilog driver is dying a miserable death because
> it thinks it should be the only serial driver, but in this case it isn't.
>
> Brad Boyer
> fl...@allandria.com
>
I see. Comparing drivers/serial/8250.c between the debian kernel source
packages for testing (2.6.15-8) and unstable (2.6.16-12) shows a number
of missing ifdefs. But are these differences because changes which need
to be continued were just not applied to 2.6.16, or did someone feel
that these ifdefs were no longer needed and removed them?
$ diff -u ./2.6.15/2.6.15-8/usr/src/linux-source-2.6.15/drivers/serial/8250.c \
> ./2.6.16/2.6.16-12/usr/src/linux-source-2.6.16/drivers/serial/8250.c
--- ./2.6.15/2.6.15-8/usr/src/linux-source-2.6.15/drivers/serial/8250.c 2006-03-06 05:29:37.000000000 -0500
+++ ./2.6.16/2.6.16-12/usr/src/linux-source-2.6.16/drivers/serial/8250.c 2006-03-20 00:53:29.000000000 -0500
@@ -31,7 +31,6 @@
.
.
.
@@ -2291,12 +2271,6 @@
static int __init serial8250_console_init(void)
{
-#ifdef CONFIG_PPC_MULTIPLATFORM
- if(_machine == _MACH_Pmac) {
- printk("%s: nothing to do on PowerMac\n",__FUNCTION__);
- return -ENODEV;
- }
-#endif
serial8250_isa_init_ports();
register_console(&serial8250_console);
return 0;
@@ -2308,11 +2282,9 @@
.
.
.
@@ -2602,14 +2581,11 @@
{
int ret, i;
-#ifdef CONFIG_PPC_MULTIPLATFORM
- if(_machine == _MACH_Pmac) {
- printk("%s: nothing to do on PowerMac\n",__FUNCTION__);
- return -ENODEV;
- }
-#endif
+ if (nr_uarts > UART_NR)
+ nr_uarts = UART_NR;
+
printk(KERN_INFO "Serial: 8250/16550 driver $Revision: 1.90 $ "
- "%d ports, IRQ sharing %sabled\n", (int) UART_NR,
+ "%d ports, IRQ sharing %sabled\n", nr_uarts,
share_irqs ? "en" : "dis");
for (i = 0; i < NR_IRQS; i++)
@@ -2619,21 +2595,27 @@
.
.
.
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
--
To UNSUBSCRIBE, email to debian-bugs-...@lists.debian.org
Sven Luther
> On Thu, 27 Apr 2006, Brad Boyer wrote:
>
> > On Thu, Apr 27, 2006 at 09:11:32PM +0200, Sven Luther wrote:
> > > > What is supposed to happen after 'Trying to im_free...' ?
> > >
> > > Attaching the full log. The isa stuff is probably related to the graphic card.
> > >
> >
> > > Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled
> > > pmac_zilog: 0.6 (Benjamin Herrenschmidt <be...@kernel.crashing.org>)
> > > pmac_zilog: Error registering serial device, disabling pmac_zilog.
> > > pmac_zilog: Did another serial driver already claim the minors?
> > > Trying to im_free nonexistent area (d00008008216c000)
> > > Trying to im_free nonexistent area (d00008008216e000)
> >
> > Here is the problem. The 8250/16550 driver used to have a check to make
> > it bail out on all PowerMac models. It used to print out a message to
> > that fact, too. The pmac_zilog driver is dying a miserable death because
> > it thinks it should be the only serial driver, but in this case it isn't.
> >
> > Brad Boyer
> > fl...@allandria.com
> >
>
> I see. Comparing drivers/serial/8250.c between the debian kernel source
> packages for testing (2.6.15-8) and unstable (2.6.16-12) shows a number
> of missing ifdefs. But are these differences because changes which need
> to be continued were just not applied to 2.6.16, or did someone feel
> that these ifdefs were no longer needed and removed them?
Indeed, that was the patch. It is disabled, probablybecause it doesn't apply
anymore, and needs redoing.
Friendly,
Sven Luther
Toni L. Harbaugh-Blackford [Contr]
The conflict between pmac_zilog and the 8250/16550 driver still exists in the
latest daily images. Is there any other way around the issue besides patching
the source for 8250.c?
Is there a way to change the order in which the kernel attempts the drivers,
to have pmac_zilog go before 8250.c?
I can't really fix this myself, because I can't get a working debian installation
to do it on. Anyone have a workaround?
Thanks,
Toni
> To UNSUBSCRIBE, email to debian-powe...@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
>
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
--
To UNSUBSCRIBE, email to debian-powe...@lists.debian.org
Sven Luther
> Hi
>
> The conflict between pmac_zilog and the 8250/16550 driver still exists in the
> latest daily images. Is there any other way around the issue besides patching
> the source for 8250.c?
If 2.6.17 doesn't fix this, then i will have a look.
> Is there a way to change the order in which the kernel attempts the drivers,
> to have pmac_zilog go before 8250.c?
Since they are both builtin, i think it would be difficult.
> I can't really fix this myself, because I can't get a working debian installation
> to do it on. Anyone have a workaround?
Friendly,
Sven Luther
--
To UNSUBSCRIBE, email to debian-bugs-...@lists.debian.org
Mark Hymers
> If 2.6.17 doesn't fix this, then i will have a look.
It still doesn't work on 2.6.17 for me. I've tried this patch and it
works on the Xserve G5s at work but it needs review as I'm not a kernel
coder and I'm worried it might break the serial ports on other machines.
This was tested building against the debian kernel source 2.6.17-5.
Cheers,
Mark
--
Mark Hymers <mark at hymers dot org dot uk>
"I told you I was ill"
The epitaph of Spike Milligan (1918-2002)
Toni L. Harbaugh-Blackford [Contr]
pardon if this is a dumb question, but does anyone know what archs use
the 8250 driver that would be good choices for a test?
Thanks,
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
Sven Luther
> #! /bin/sh -e
> ##
> ## All lines beginning with `## DP:' are a description of the patch.
> ## DP: Description: Disables legacy serial driver on powermacs.
> ## DP: Patch author: Mark Hymers <ma...@hymers.org.uk>
> ## DP: Patch author: adapted from Sven's earlier version
> ## DP: Patch author: Sven Luther <lut...@debian.org>
> ## DP: Patch author: adapted from the SuSE kernel tree.
> ## DP: Upstream status: workaround hack waiting for a clean legacy device solution.
>
> diff -aurN a/drivers/serial/8250.c b/drivers/serial/8250.c
> --- a/drivers/serial/8250.c 2006-06-18 02:49:35.000000000 +0100
> +++ b/drivers/serial/8250.c 2006-08-01 00:31:40.000000000 +0100
> @@ -2307,9 +2312,14 @@
>
> static int __init serial8250_console_init(void)
> {
> +#ifdef CONFIG_PPC_PMAC
> + printk("%s: nothing to do on PowerMac\n",__FUNCTION__);
> + return -ENODEV;
> +#else
> serial8250_isa_init_ports();
> register_console(&serial8250_console);
> return 0;
> +#endif
I don't believe this will work, since we have hardware who has a pc-like 8250
serial interface, and is powerpc, and will run the same kernel that will also
work on powermacs. This is why a static built-time check like you do is not
valid, and why the previous patch used a dynamic checked subarch variable.
There must be some workaround for that one, but i also believe that people
like benh have been working on a better solution to this. I lost sight of the
current state of this though.
Mark Hymers
> I don't believe this will work, since we have hardware who has a pc-like 8250
> serial interface, and is powerpc, and will run the same kernel that will also
> work on powermacs. This is why a static built-time check like you do is not
> valid, and why the previous patch used a dynamic checked subarch variable.
> There must be some workaround for that one, but i also believe that people
> like benh have been working on a better solution to this. I lost sight of the
> current state of this though.
Oh, absolutely. I was mainly pointing out that people can use it if
they have Xserves which they really need the serial ports working on.
Of course, it occurs that if you're rebuilding the kernel you could just
disable the 8250 driver anyways :-)
Cheers,
Mark
--
Mark Hymers <mark at hymers dot org dot uk>
"Irish police are being handicapped in a search for a stolen van, because
they cannot issue a description. It's a special branch vehicle, and they
don't want the public to know what it looks like."
The Guardian
Toni L. Harbaugh-Blackford [Contr]
> On Thu, 03, Aug, 2006 at 11:50:19PM +0200, Sven Luther spoke thus..
> > I don't believe this will work, since we have hardware who has a pc-like 8250
> > serial interface, and is powerpc, and will run the same kernel that will also
> > work on powermacs. This is why a static built-time check like you do is not
> > valid, and why the previous patch used a dynamic checked subarch variable.
> > There must be some workaround for that one, but i also believe that people
> > like benh have been working on a better solution to this. I lost sight of the
> > current state of this though.
>
> Oh, absolutely. I was mainly pointing out that people can use it if
> they have Xserves which they really need the serial ports working on.
> Of course, it occurs that if you're rebuilding the kernel you could just
> disable the 8250 driver anyways :-)
Yes, but you have to have a compatible base system installed first, in order
to build the kernel. That's fine on x86, but getting a working linux
build on an Xserve without a an installer or a preinstalled compatible
system *elsewhere* (to debootstrap from) is very difficult.
Toni
-------------------------------------------------------------------
Toni Harbaugh-Blackford harb...@abcc.ncifcrf.gov
System Administrator
Advanced Biomedical Computing Center (ABCC)
National Cancer Institute
Contractor - SAIC/Frederick
--
To UNSUBSCRIBE, email to debian-ker...@lists.debian.org