Ubuntu kernel development ()

Andrew Zajac | 11 Nov 2005 15:33

Badram kernel patch for Dapper?

Hello.

Would it be possible to include the badram kernel patch as part of a stock kernel for Dapper?
http://rick.vanrein.org/linux/badram/

http://rick.vanrein.org/linux/badram/download.html

I have tried the 2.6.12 patch and it seems to work fine and not cause any problems, although I have not tested it extensively.

It is problematic to swap out the ram in a machine to install the OS and then install the badram patched kernel to be able to run it with the bad ram stick(s). Ideally, this feature should be in the kernel on the cd from which the user would boot. They could run memtest, input the parameters once and then install.

I think this could allow a lot of users who own failing or recycled hardware to run Ubuntu. I think it could benefit some users who would run the planned "light" version of edubuntu which would be built on top of Xubuntu.

As well, it is a really big selling point which goes against the trend set by other operating systems which neccessitate a hardware upgrade to properly run the latest version.

Andrew.

<div><p>Hello.<br><br>

Would it be possible to include the <span name="st" class="st0">badram</span> kernel patch as part of a stock kernel for Dapper?<br><a href="http://rick.vanrein.org/linux/badram/" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://rick.vanrein.org/linux/<span name="st" class="st0">badram</span>/</a><br><br><a href="http://rick.vanrein.org/linux/badram/download.html" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://rick.vanrein.org/linux/<span name="st" class="st0">badram</span>/download.html
</a><br><br>

I have tried the 2.6.12 patch and it seems to work fine and not cause
any problems, although I have not tested it extensively.&nbsp; <br><br>

It is problematic to swap out the ram in a machine to install the OS
and then install the <span name="st" class="st0">badram</span> patched kernel to be able to run it with
the bad ram stick(s).&nbsp; Ideally, this feature should be in the
kernel on the cd from which the user would boot.&nbsp; They could run
memtest, input the parameters once and then install.<br><br>

I think this could allow a lot of users who own failing or recycled
hardware to run Ubuntu.&nbsp; I think it could benefit some users who
would run the planned "light" version of edubuntu which would be built
on top of Xubuntu.<br><br>

As well, it is a really big selling point which goes against the trend
set by other operating systems which neccessitate a hardware upgrade to
properly run the latest version. <br><span class="sg">
<br>
Andrew.<br></span></p></div>

headers

Ben Collins | 16 Nov 2005 02:38

Status, 2.6.15-2.2, and so on

In case you don't follow the buildd's like a stock market ticker,
2.6.15-2.2 was uploaded today. It built on x86 and amd64, and I had
succsessful reports from amd64 and x86 users.

THere were several reports (all within the first hour of availability):

- Usplash stopped working. I was able to peg this one on usplash :) Needed
  to tweak the ordering of the fbcon and related modules.

- ipw2200 firmware didn't load. -3.3 will have the newer firmware (already
  in git).

- Atmel USB driver did not exist. Got this one in, so exepct it in -3.3.

- FTBFS for powerpc. Fixed this. Needed to workaround a bug. Very minor.

So expect 2.6.15-3.3 by the end of tomorrow.

--

-- 
   Ben Collins <ben.collins <at> ubuntu.com>
   Developer
   Ubuntu Linux

headers

Fabio Massimo Di Nitto | 16 Nov 2005 07:52

Re: Status, 2.6.15-2.2, and so on

Ben Collins wrote:

> 
> - FTBFS for powerpc. Fixed this. Needed to workaround a bug. Very minor.

PPC users should be extra careful NOT to purge old kernels.

Upstream is in the middle of a big code cleanup/polish/reorganization phase and
it is known NOT to boot on all machines or NOT to work properly (specially in
32bit mode).

Fabio

--

-- 
I'm going to make him an offer he can't refuse.

headers

Ben Collins | 16 Nov 2005 13:37

Re: Status, 2.6.15-2.2, and so on

On Wed, Nov 16, 2005 at 07:52:02AM +0100, Fabio Massimo Di Nitto wrote:
> Ben Collins wrote:
> 
> > 
> > - FTBFS for powerpc. Fixed this. Needed to workaround a bug. Very minor.
> 
> PPC users should be extra careful NOT to purge old kernels.
> 
> Upstream is in the middle of a big code cleanup/polish/reorganization phase and
> it is known NOT to boot on all machines or NOT to work properly (specially in
> 32bit mode).

Well, one good note is that it works perfectly on my G4 (Grey case), with
airo card and ATI graphics. Been up for 36 hours now (my main desktop)
with no problems.

--

-- 
   Ben Collins <ben.collins <at> ubuntu.com>
   Developer
   Ubuntu Linux

headers

Dr. David Alan Gilbert | 16 Nov 2005 18:24

incorrect change of [Bug 8907] to old

Hi,
  My bug report 8907 has been incorrectly marked as old saying that
it hasn't been tested on Breezy.
You can see my comment #2 that was the result of my test on breezy.

Dave
--
 -----Open up your eyes, open up your mind, open up your code -------   
/ Dr. David Alan Gilbert    | Running GNU/Linux on Alpha,68K| Happy  \ 
\ gro.gilbert  <at>  treblig.org | MIPS,x86,ARM,SPARC,PPC & HPPA | In Hex /
 \ _________________________|_____ http://www.treblig.org   |_______/

headers

Andrew Zajac | 10 Nov 2005 02:51

Badram kernel patch for Dapper?

Hello.

Would it be possible to include the badram kernel patch as part of a stock
kernel for Dapper?

http://rick.vanrein.org/linux/badram/

http://rick.vanrein.org/linux/badram/download.html

I have tried the 2.6.12 patch and it seems to work fine and not cause any
problems, although I have not tested it extensively.

It is problematic to swap out the ram in a machine to install the OS and
then install the badram patched kernel to be able to run it with the bad ram
stick(s). Ideally, this feature should be in the kernel on the cd from which
the user would boot. They could run memtest, input the parameters once and
then install.

I think this could allow a lot of users who own failing or recycled hardware
to run Ubuntu. I think it could benefit some users who would run the planned
"light" version of edubuntu which would be built on top of Xubuntu.

As well, it is a really big selling point which goes against the trend set
by other operating systems which neccessitate a hardware upgrade to properly
run the latest version.

Andrew.

<div>Hello. <br><br>Would it be possible to include the badram kernel patch as part of a stock<br>kernel for Dapper?<br><br><a href="http://rick.vanrein.org/linux/badram/">http://rick.vanrein.org/linux/badram/</a><span>
<br><span><br><a href="http://rick.vanrein.org/linux/badram/download.html">http://rick.vanrein.org/linux/badram/download.html</a><br><br></span></span>I have tried the 2.6.12 patch and it seems to work fine and not cause any
<br>problems, although I have not tested it extensively.<br><br>It is problematic to swap out the ram in a machine to install the OS and<br>then install the badram patched kernel to be able to run it with the bad ram<br>stick(s). Ideally, this feature should be in the kernel on the cd from which
<br>the user would boot. They could run memtest, input the parameters once and<br>then install.<br><br>I think this could allow a lot of users who own failing or recycled hardware<br>to run Ubuntu. I think it could benefit some users who would run the planned
<br>"light" version of edubuntu which would be built on top of Xubuntu.<br><br>As well, it is a really big selling point which goes against the trend set<br>by other operating systems which neccessitate a hardware upgrade to properly
<br>run the latest version.<br><br>Andrew.</div>

headers

Dorneles Treméa | 17 Nov 2005 07:18

Re: Status, 2.6.15-2.2, and so on

Hello Ben,

> In case you don't follow the buildd's like a stock market ticker,
> 2.6.15-2.2 was uploaded today. It built on x86 and amd64, and I had
> succsessful reports from amd64 and x86 users.
> 
> THere were several reports (all within the first hour of availability):
> 
> [...]

2.6.12 had a CVS copy from drivers/acpi/asus_acpi.c while the current
2.6.15 comes with the standard 0.29 version.

What's the policy here?

--

-- 

Dorneles Treméa
X3ng Web Technology

headers

Fabio Massimo Di Nitto | 17 Nov 2005 13:55

Next kernel security release hell


Hi guys,
	I have basically completed the packages that will hit warty/hoary/breezy in
-security.

There are two problems I need help to address:

1) this huge update requires deep testing, because the changes are intrusive and
that means covering all 3 main arches on all 3 releases in as many flavours as
possible. I simply don't have hw available to cover everything here.

2) due to the nature of the changes, there is a kernel ABI bump in all 3
releases. AFAIK we never had this situation before and this drag in the problem
of uploading linux-restricted-modules and linux-meta.
We did never agree (or talk) if the rebuild of the latters should be done via
-security or -updates.
I personally would like to see them entering the same suite (-security) as the
kernel even if they do not contain security updates themself. Any objection?
Better ideas?

Colin: AFAIR warty did not build udeb from the kernel itself. I assume we will
not need to update d-i, but for a person that has -security in his sources.list
it will make one package unbuildable (the one that was doing deb -> udeb
conversion and i can't remember the name)
How should we address this problem IF we have to address it.

Fabio

--
I'm going to make him an offer he can't refuse.

headers

Colin Watson | 17 Nov 2005 14:10

Re: Next kernel security release hell

On Thu, Nov 17, 2005 at 01:55:14PM +0100, Fabio Massimo Di Nitto wrote:
> 	I have basically completed the packages that will hit warty/hoary/breezy in
> -security.
> 
> There are two problems I need help to address:
> 
> 1) this huge update requires deep testing, because the changes are intrusive and
> that means covering all 3 main arches on all 3 releases in as many flavours as
> possible. I simply don't have hw available to cover everything here.
> 
> 2) due to the nature of the changes, there is a kernel ABI bump in all 3
> releases. AFAIK we never had this situation before and this drag in the problem
> of uploading linux-restricted-modules and linux-meta.

We never had ABI bumps in all releases before, but we've already had two
ABI bumps in hoary.

> We did never agree (or talk) if the rebuild of the latters should be done via
> - -security or -updates.
> I personally would like to see them entering the same suite (-security) as the
> kernel even if they do not contain security updates themself.

I agree; this makes sense.

> Colin: AFAIR warty did not build udeb from the kernel itself. I assume we will
> not need to update d-i, but for a person that has -security in his sources.list
> it will make one package unbuildable (the one that was doing deb -> udeb
> conversion and i can't remember the name)
> How should we address this problem IF we have to address it.

That would be linux-kernel-di-{amd64,i386,powerpc}-2.6 in warty. We
could (and arguably should) certainly upload these to -security as well
for completeness' sake, although as you say it's unlikely that we'll
update debian-installer itself.

--

-- 
Colin Watson                                       [cjwatson <at> ubuntu.com]

headers

Adam Conrad | 17 Nov 2005 14:13

Re: Next kernel security release hell

Fabio Massimo Di Nitto wrote:

> 2) due to the nature of the changes, there is a kernel ABI bump in all 3
> releases. AFAIK we never had this situation before and this drag in
> the problem
> of uploading linux-restricted-modules and linux-meta.
> We did never agree (or talk) if the rebuild of the latters should be
> done via
> -security or -updates.
> I personally would like to see them entering the same suite
> (-security) as the
> kernel even if they do not contain security updates themself. Any
> objection?
> Better ideas?

We pretty much have to send all updates to the same suite, since some
people may not have -updates enabled, and we'd end up breaking their
setups.  This precedent has been set in the past (for instance, enigmail
updates to match new mozilla/thunderbird in -security).  If you
need/want my help getting lrm updated, ping me on IRC (tomorrow, I'm
about to go back to bed and nurse my cold some more)

... Adam

488	June 2013
901	May 2013
663	April 2013
833	March 2013
876	February 2013
972	January 2013
724	December 2012
739	November 2012
405	October 2012
368	September 2012
368	August 2012
360	July 2012
444	June 2012
373	May 2012
449	April 2012
545	March 2012
365	February 2012
346	January 2012
348	December 2011
428	November 2011
389	October 2011
305	September 2011
377	August 2011
468	July 2011
295	June 2011
227	May 2011
267	April 2011
477	March 2011
446	February 2011
351	January 2011
204	December 2010
323	November 2010
253	October 2010
423	September 2010
526	August 2010
436	July 2010
509	June 2010
434	May 2010
566	April 2010
598	March 2010
453	February 2010
328	January 2010
271	December 2009
152	November 2009
340	October 2009
249	September 2009
348	August 2009
468	July 2009
327	June 2009
207	May 2009
466	April 2009
536	March 2009
244	February 2009
258	January 2009
274	December 2008
253	November 2008
157	October 2008
191	September 2008
185	August 2008
126	July 2008
163	June 2008
128	May 2008
98	April 2008
81	March 2008
103	February 2008
99	January 2008
38	December 2007
42	November 2007
60	October 2007
88	September 2007
113	August 2007
52	July 2007
48	June 2007
75	May 2007
69	April 2007
55	March 2007
24	February 2007
21	January 2007
11	December 2006
38	November 2006
82	October 2006
30	September 2006
84	August 2006
25	July 2006
51	June 2006
34	May 2006
62	April 2006
100	March 2006
25	February 2006
20	January 2006
24	December 2005
20	November 2005
27	October 2005
17	September 2005
8	August 2005
5	July 2005
17	June 2005
33	May 2005
25	April 2005
3	March 2005

Mon	Tue	Wed	Thu	Fri	Sat	Sun

	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28	29	30