Debian Bug report logs - #394970
finish-install: [powerpc64] Add support for IBM serial consoles (hvc and hvsi)

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Sven Luther <luther@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: finish-install: [powerpc64] Add support for IBM serial consoles (hvc and hvsi)

Date: Mon, 23 Oct 2006 22:19:33 +0200

[Message part 1 (text/plain, inline)]

Package: finish-install
Version: 2.4
Severity: normal
Tags: patch


Add support for hvsi and hvc IBM POWER serial consoles.

This patch is currently untested, but i don't want to lose it again, so i
submit it. I expect to test it nextly, but i can only do so for hvc as i have
no longer access to a non-virtualized pserver, and can only do so within the
constraints of my free time. Others are welcome to test it by themselves.

Also, i don't believe this is the most efficient way to do this, and that the
ttyS|hvc|hvsi could be factorised in the code, but my shell skill is not
enough to do it better without risking breakage, so i will only provide this
functional patch.

Friendly,

Sven Luther

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: powerpc (ppc)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-powerpc
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8)

[finish-install-hv.diff (text/plain, attachment)]

Message #10 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Frans Pop <elendil@planet.nl>

To: Sven Luther <luther@debian.org>, 394970@bugs.debian.org

Subject: Re: Bug#394970: finish-install: [powerpc64] Add support for IBM serial consoles (hvc and hvsi)

Date: Thu, 23 Nov 2006 16:24:34 +0100

[Message part 1 (text/plain, inline)]

On Monday 23 October 2006 22:19, Sven Luther wrote:
> This patch is currently untested, but i don't want to lose it again, so
> i submit it. I expect to test it nextly, but i can only do so for hvc
> as i have no longer access to a non-virtualized pserver, and can only
> do so within the constraints of my free time.

The patch had an error in the sed on /etc/inittab. Please test the 
attached patch, which also avoids the code duplication.

[finish-install_sercon.patch (text/x-diff, attachment)]

[Message part 3 (application/pgp-signature, inline)]

Message #15 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Sven Luther <sven.luther@wanadoo.fr>

To: Frans Pop <elendil@planet.nl>

Cc: Sven Luther <luther@debian.org>, 394970@bugs.debian.org

Subject: Re: Bug#394970: finish-install: [powerpc64] Add support for IBM serial consoles (hvc and hvsi)

Date: Thu, 23 Nov 2006 17:06:13 +0100

On Thu, Nov 23, 2006 at 04:24:34PM +0100, Frans Pop wrote:
> On Monday 23 October 2006 22:19, Sven Luther wrote:
> > This patch is currently untested, but i don't want to lose it again, so
> > i submit it. I expect to test it nextly, but i can only do so for hvc
> > as i have no longer access to a non-virtualized pserver, and can only
> > do so within the constraints of my free time.
> 
> The patch had an error in the sed on /etc/inittab. Please test the 
> attached patch, which also avoids the code duplication.

Ok, cool, will do this this WE, since i have no access to the box.

> Index: finish-install.d/90console
> ===================================================================
> --- finish-install.d/90console	(revision 42772)
> +++ finish-install.d/90console	(working copy)
> @@ -16,16 +16,18 @@
>  console=${console#/dev/}
>  
>  case "$console" in
> -ttyS*)
> +ttyS*|hvc*|hvsi*)

BTW, the new genesi efika board will use /dev/ttyPSC*, and i guess there are a
bunch of other serial ports out there.

Would it not be better to use the variable in rootskel to detect that the
terminal is a serial terminal, namely those who delcare themselves to be :

  TERM_TYPE=serial in rootskel/src/lib/debian-installer/detect-console-linux

The code there is : 

  /dev/console|/dev/tts/*|/dev/tty[A-Z]*|/dev/hvc*|/dev/hvsi*)

Furthermore, i have personally found out that it is interesting to enable
a serial console in /etc/inittab even in non-serial installs, Maybe a debconf
question of low priority would allow this to be preseeded or set in expert
installs ? Maybe too late for etch now, but still something to keep in mind.

Friendly,

Sven Luther

Message #22 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Holger Levsen <debian@layer-acht.org>

To: debian-powerpc@lists.debian.org

Cc: 394970@bugs.debian.org

Subject: Re: which serious bugs exactly? (Re: if you care about debian on powerpc, please react ...)

Date: Thu, 30 Nov 2006 12:17:07 +0100

[Message part 1 (text/plain, inline)]

Hi,

On Sunday 26 November 2006 15:09, Sven Luther wrote:
> > > currently shiping powerpc hardware are :
> > >   IBM pseries : not really supported, patch sitting without comment
> > > since months, early work lost because of svn commit conflicts.
> > raise the bug severity? which bug#, btw?
>
>   Outstanding bugs -- Normal bugs; Patch Available (1 bug)
>   #394970: finish-install: [powerpc64] Add support for IBM serial consoles
>   (hvc and hvsi)
>   Package: finish-install (finish-install 2.4); Reported by: Sven Luther
>   <luther@debian.org>; Tags: patch; 33 days old

Sven, the last info on this bug is from you, saying that you would test the 
patch Frans corrected the coming (and now past) weekend. You didn't post an 
update how the test went, so I'm not surprised the patch isn't commited.

How did the test went? Is the patch now fine? Could you please add this 
information to the bugreport?! Thanks.

As you mention this bug in this thread under the "powerpc should be removed 
because arch support is poor" label (which I still think isn't true), I 
wonder if the severity "normal" is correct.


regards,
	Holger

[Message part 2 (application/pgp-signature, inline)]

Message #27 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Frans Pop <elendil@planet.nl>

To: 394970@bugs.debian.org

Subject: Re: Bug#394970: which serious bugs exactly? (Re: if you care about debian on powerpc, please react ...)

Date: Thu, 30 Nov 2006 13:19:36 +0100

On Thursday 30 November 2006 12:17, Holger Levsen wrote:
> Sven, the last info on this bug is from you, saying that you would test
> the patch Frans corrected the coming (and now past) weekend. You didn't
> post an update how the test went, so I'm not surprised the patch isn't
> commited.

The patch is committed now and pending upload. A test on a machine using 
the newly supported devices would still be nice though.

Message #32 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Sven Luther <sven.luther@wanadoo.fr>

To: Holger Levsen <debian@layer-acht.org>

Cc: debian-powerpc@lists.debian.org, 394970@bugs.debian.org

Subject: Re: which serious bugs exactly? (Re: if you care about debian on powerpc, please react ...)

Date: Thu, 30 Nov 2006 15:30:05 +0100

On Thu, Nov 30, 2006 at 12:17:07PM +0100, Holger Levsen wrote:
> Hi,
> 
> On Sunday 26 November 2006 15:09, Sven Luther wrote:
> > > > currently shiping powerpc hardware are :
> > > >   IBM pseries : not really supported, patch sitting without comment
> > > > since months, early work lost because of svn commit conflicts.
> > > raise the bug severity? which bug#, btw?
> >
> >   Outstanding bugs -- Normal bugs; Patch Available (1 bug)
> >   #394970: finish-install: [powerpc64] Add support for IBM serial consoles
> >   (hvc and hvsi)
> >   Package: finish-install (finish-install 2.4); Reported by: Sven Luther
> >   <luther@debian.org>; Tags: patch; 33 days old
> 
> Sven, the last info on this bug is from you, saying that you would test the 
> patch Frans corrected the coming (and now past) weekend. You didn't post an 
> update how the test went, so I'm not surprised the patch isn't commited.

I didn't, because

  1) the box is not connected now, and i had the XServe G5 in the rack, which
  was being prepared to be put in production, and due to bugs like :

     #397973 [powerpci/mac] partman-md appears to not write back the raid flag to partitions.

  This took longer, and i was able to put it in the datacenter only yesterday,
  freeing the rack.

  2) The p505 i have on lend from IBM now has the VIOS server installed, and i
  need to read the 150+ pages of documentation to become familiar with the
  LPAR setup code, before doing the test.

  3) Last sunday, we did my 2 year son's birthday party, which didn't leave me
  time to do the testing, especially given the issues with 1), and other work
  related issues, which had more priority.

So, no, i didn't have time to test it, the day has only so much hours, and you
would understand, that family live and RL work takes priority, as so many have
been telling me recently.

Now, there is no excuse to not commit the patch, especially the one modified
by Frans, to the svn, and indeed i expect it to be commited already, just
waiting for my test to upload the package.

Even then, what really needs testing, is the check for the patch to not break
finish-install on other hardware, to make sure there are no regressions, and
this would be better served by the patch going in ASAP, rather than waiting
longer. This would also allow to setup a call for user to test, rather than
relying on only me.

> How did the test went? Is the patch now fine? Could you please add this 
> information to the bugreport?! Thanks.
> 
> As you mention this bug in this thread under the "powerpc should be removed 
> because arch support is poor" label (which I still think isn't true), I 
> wonder if the severity "normal" is correct.

Well, if i upgrade severity, i am sure to get an immediate backlash, but you
are free to do so if you like.

In general, i would seriously welcome more people working on the powerpc port,
and not people like you leaving all the pressure on me, or like frans said
once : "we won't bother fixing powerpc issues, since we know svenl will
eventually fix them". This doesn't scale well to me having less time due to RL
issues, as you might guess. And may have been one of the cause of the
escalation of those problems back in spring, there is such a thing like
over-stressing your human ressources, any good leader should know that :)

Friendly,

Sven Luther

Message #37 received at 394970-close@bugs.debian.org (full text, mbox, reply):

From: Frans Pop <fjp@debian.org>

To: 394970-close@bugs.debian.org

Subject: Bug#394970: fixed in finish-install 2.6

Date: Thu, 30 Nov 2006 14:02:04 +0000

Source: finish-install
Source-Version: 2.6

We believe that the bug you reported is fixed in the latest version of
finish-install, which is due to be installed in the Debian FTP archive:

finish-install_2.6.dsc
  to pool/main/f/finish-install/finish-install_2.6.dsc
finish-install_2.6.tar.gz
  to pool/main/f/finish-install/finish-install_2.6.tar.gz
finish-install_2.6_all.udeb
  to pool/main/f/finish-install/finish-install_2.6_all.udeb



A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 394970@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Frans Pop <fjp@debian.org> (supplier of updated finish-install package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Format: 1.7
Date: Thu, 30 Nov 2006 15:32:29 +0100
Source: finish-install
Binary: finish-install
Architecture: source all
Version: 2.6
Distribution: unstable
Urgency: low
Maintainer: Debian Install System Team <debian-boot@lists.debian.org>
Changed-By: Frans Pop <fjp@debian.org>
Description: 
 finish-install - Finish the installation and reboot (udeb)
Closes: 394970
Changes: 
 finish-install (2.6) unstable; urgency=low
 .
   * 90console: [powerpc64] Added support for hvc* and hvsi* serial consoles.
     Note: untested on powerpc. Closes: #394970.
   * 90console: add a blank line before pringing new lines to securetty.
 .
   [ Updated translations ]
   * Esperanto (eo.po) by Serge Leblanc
   * Georgian (ka.po) by Aiet Kolkhi
   * Kurdish (ku.po) by rizoye-xerzi
   * Panjabi (pa.po) by A S Alam
Files: 
 4b82c24c2f50cd83ca13519d433720eb 766 debian-installer required finish-install_2.6.dsc
 8ce608f547722bef73217bb65627eccf 40924 debian-installer required finish-install_2.6.tar.gz
 3bc1d99b5aee915977f3a9bbf5bd13d6 17968 debian-installer required finish-install_2.6_all.udeb
Package-Type: udeb

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFFbu3igm/Kwh6ICoQRAiFOAKCv9G/Lzqv0JdRpOxv3wI4eOiZG8ACfRUzp
1npPEOWIlGwworEUPChjgug=
=dCSb
-----END PGP SIGNATURE-----

Message #42 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Fabio Tranchitella <kobold@kobold.it>

To: Frans Pop <elendil@planet.nl>, debian@layer-acht.org

Cc: 394970@bugs.debian.org, Sven Luther <sven.luther@wanadoo.fr>

Subject: Please, add credits for patch while closing #394970

Date: Thu, 30 Nov 2006 18:02:01 +0100

[Message part 1 (text/plain, inline)]

Hi Frans and Holger,
  do you mind to retroactively modify the changelog for the upload 2.6 of
finish-install to credit Sven Luther about the patch he submitted? I do not
want to be pedantic, and I know that you modified and fixed his patch, but I
think it would be correct considering that in the same changelog entry you
already cite the names of the translators who submitted the updates.

Thanks, and have a nice day.

"""
 finish-install (2.6) unstable; urgency=low
 .
   * 90console: [powerpc64] Added support for hvc* and hvsi* serial consoles.
     Note: untested on powerpc. Closes: #394970.
   * 90console: add a blank line before pringing new lines to securetty.
 .
   [ Updated translations ]
   * Esperanto (eo.po) by Serge Leblanc
   * Georgian (ka.po) by Aiet Kolkhi
   * Kurdish (ku.po) by rizoye-xerzi
   * Panjabi (pa.po) by A S Alam
"""

-- 
Fabio Tranchitella                         http://www.kobold.it
Free Software Developer and Consultant     http://www.tranchitella.it
_____________________________________________________________________
1024D/7F961564, fpr 5465 6E69 E559 6466 BF3D 9F01 2BF8 EE2B 7F96 1564

[signature.asc (application/pgp-signature, inline)]

Message #47 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Frans Pop <elendil@planet.nl>

To: Fabio Tranchitella <kobold@kobold.it>

Cc: 394970@bugs.debian.org

Subject: Re: Please, add credits for patch while closing #394970

Date: Thu, 30 Nov 2006 18:33:07 +0100

[Message part 1 (text/plain, inline)]

On Thursday 30 November 2006 18:02, Fabio Tranchitella wrote:
>   do you mind to retroactively modify the changelog for the upload 2.6
> of finish-install to credit Sven Luther about the patch he submitted? I
> do not want to be pedantic, and I know that you modified and fixed his
> patch, but I think it would be correct considering that in the same
> changelog entry you already cite the names of the translators who
> submitted the updates.

No sorry.
The patch that was applied was not Sven's patch, but a patch I basically 
created myself from scratch. Sven's patch was broken (/etc/inittab would 
not have had the correct line) and had unnecessary code duplication (as 
he indicated himself).

So, IMO there is no need to credit him for the patch and we normally don't 
credit people just for submitting a bug report.

The citing of translators is automated and is a completely different issue 
and they actually do write and commit the changes they are credited for.

Sven has always been correctly credited for any patches he's submitted or 
work he's contributed.

Cheers,
FJP

[Message part 2 (application/pgp-signature, inline)]

Message #52 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Fabio Tranchitella <kobold@kobold.it>

To: Frans Pop <elendil@planet.nl>, Sven Luther <sven.luther@wanadoo.fr>

Cc: 394970@bugs.debian.org

Subject: Re: Please, add credits for patch while closing #394970

Date: Thu, 30 Nov 2006 18:40:16 +0100

[Message part 1 (text/plain, inline)]

Hi Frans,

* 2006-11-30 18:33, Frans Pop wrote:
> On Thursday 30 November 2006 18:02, Fabio Tranchitella wrote:
> >   do you mind to retroactively modify the changelog for the upload 2.6
> > of finish-install to credit Sven Luther about the patch he submitted? I
> > do not want to be pedantic, and I know that you modified and fixed his
> > patch, but I think it would be correct considering that in the same
> > changelog entry you already cite the names of the translators who
> > submitted the updates.
> 
> No sorry.
> The patch that was applied was not Sven's patch, but a patch I basically 
> created myself from scratch. Sven's patch was broken (/etc/inittab would 
> not have had the correct line) and had unnecessary code duplication (as 
> he indicated himself).

Fine, thanks for your explanation: I now understand your reasons, and I
accept it.

> credit people just for submitting a bug report.

That's normal.

> Sven has always been correctly credited for any patches he's submitted or 
> work he's contributed.

Thanks for confirming this, too.

Have a nice day,

-- 
Fabio Tranchitella                         http://www.kobold.it
Free Software Developer and Consultant     http://www.tranchitella.it
_____________________________________________________________________
1024D/7F961564, fpr 5465 6E69 E559 6466 BF3D 9F01 2BF8 EE2B 7F96 1564

[signature.asc (application/pgp-signature, inline)]

Message #57 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Sven Luther <sven.luther@wanadoo.fr>

To: Fabio Tranchitella <kobold@kobold.it>

Cc: Frans Pop <elendil@planet.nl>, Sven Luther <sven.luther@wanadoo.fr>, 394970@bugs.debian.org

Subject: Re: Please, add credits for patch while closing #394970

Date: Fri, 1 Dec 2006 07:55:16 +0100

On Thu, Nov 30, 2006 at 06:40:16PM +0100, Fabio Tranchitella wrote:
> Hi Frans,
> 
> * 2006-11-30 18:33, Frans Pop wrote:
> > On Thursday 30 November 2006 18:02, Fabio Tranchitella wrote:
> > >   do you mind to retroactively modify the changelog for the upload 2.6
> > > of finish-install to credit Sven Luther about the patch he submitted? I
> > > do not want to be pedantic, and I know that you modified and fixed his
> > > patch, but I think it would be correct considering that in the same
> > > changelog entry you already cite the names of the translators who
> > > submitted the updates.
> > 
> > No sorry.
> > The patch that was applied was not Sven's patch, but a patch I basically 
> > created myself from scratch. Sven's patch was broken (/etc/inittab would 
> > not have had the correct line) and had unnecessary code duplication (as 
> > he indicated himself).
> 
> Fine, thanks for your explanation: I now understand your reasons, and I
> accept it.

I don't though.

> > credit people just for submitting a bug report.
> 
> That's normal.

When people sen me bug reports and stuff, or even just push me to solve a bug
they can reproduce, i always say :

  - Special thanks to <xxx> for helping solve this bug.

But then, this is part of Frans continuous trend to discredit my technical
work, to lessen my contribution, and to outcast me in general.

> > Sven has always been correctly credited for any patches he's submitted or 
> > work he's contributed.
> 
> Thanks for confirming this, too.

This is not the full truth though, but the devil is in the detail.

Friendly,

Sven Luther

Message #62 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Rolf Brudeseth <rolfb@us.ibm.com>

To: 394970@bugs.debian.org, 420820@bugs.debian.org

Cc: Frans Pop <elendil@planet.nl>, Sven Luther <luther@debian.org>

Subject: Bug#394970/420820: no console set IBM p5 server

Date: Mon, 14 May 2007 12:28:20 -0500

[Message part 1 (text/plain, inline)]

Is it understood how one determines which device file (hvsi0, hvsi1 or
hvc0) Open Firmware associates with the console as it pertains to IBM
System p servers?

This can be determined by looking at properties under /proc/device-tree.

Please let me know if this is needed and I will find the documentation. I
know how it is accomplished, I just need to make sure that I point to
documentation already released by IBM.

Rolf Brudeseth
rolfb@us.ibm.com

[Message part 2 (text/html, inline)]

Message #67 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Frans Pop <elendil@planet.nl>

To: Rolf Brudeseth <rolfb@us.ibm.com>, 394970@bugs.debian.org, 420820@bugs.debian.org

Subject: Re: Bug#394970: /420820: no console set IBM p5 server

Date: Mon, 14 May 2007 19:47:43 +0200

On Monday 14 May 2007 19:28, Rolf Brudeseth wrote:
> Is it understood how one determines which device file (hvsi0, hvsi1 or
> hvc0) Open Firmware associates with the console as it pertains to IBM
> System p servers?

The question is if it is necessary to know that. For that I'd first like 
to see that script output I requested.

> This can be determined by looking at properties under
> /proc/device-tree.
>
> Please let me know if this is needed and I will find the documentation.
> I know how it is accomplished, I just need to make sure that I point to
> documentation already released by IBM.

Thanks for the offer. To be honest, if possible I'd prefer practical 
examples (e.g. content of (pseudo) files, examples of symlinks) over 
technical and theoretical documentation.
Whether or not we need to look in proc, I doubt fixing this issue will be 
very hard. The main thing is to have someone who can provide the needed 
info and test changes.

Message #72 received at 394970@bugs.debian.org (full text, mbox, reply):

From: Sven Luther <luther@debian.org>

To: Rolf Brudeseth <rolfb@us.ibm.com>, 394970@bugs.debian.org

Cc: 420820@bugs.debian.org, Frans Pop <elendil@planet.nl>, Sven Luther <luther@debian.org>

Subject: Re: Bug#394970: /420820: no console set IBM p5 server

Date: Tue, 15 May 2007 09:47:46 +0200

On Mon, May 14, 2007 at 12:28:20PM -0500, Rolf Brudeseth wrote:
> 
> Is it understood how one determines which device file (hvsi0, hvsi1 or
> hvc0) Open Firmware associates with the console as it pertains to IBM
> System p servers?
> 
> This can be determined by looking at properties under /proc/device-tree.
> 
> Please let me know if this is needed and I will find the documentation. I
> know how it is accomplished, I just need to make sure that I point to
> documentation already released by IBM.

ofpath or ofpathname (from yaboot or the ibm-power-utils or whatever
that package is named) are the one doing the OF path from linux path
mapping. Not sure it supports the other way around, but that would be
the right way to handle this.

Friendly,

Sven Luther

Send a report that this bug log contains spam.