Hi all,
A stupid question - buggered if I can find a kernel patch from 2.6.14.5 to
2.6.15?
Is there one?
Thanks,
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
On 1/4/06, Nick Warne <[email protected]> wrote:
> Hi all,
>
> A stupid question - buggered if I can find a kernel patch from 2.6.14.5 to
> 2.6.15?
>
> Is there one?
>
No.
What you do is you first revert the 2.6.14.5 patch so you are left
with a 2.6.14 kernel, then you apply the 2.6.15 patch.
For more info, please read Documentation/applying-patches.txt
(http://sosdg.org/~coywolf/lxr/source/Documentation/applying-patches.txt)
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wed, Jan 04, 2006 at 05:10:37PM +0000, Nick Warne wrote:
> A stupid question - buggered if I can find a kernel patch from 2.6.14.5 to
> 2.6.15?
You can try backing out the 2.6.14 -> 2.6.14.5 patch to generate a
"pristine" 2.6.14 to which you can apply 2.6.14 -> 2.6.15.
This mess is one reason while I usually keep old kernel tarballs
around for a while.
Greetings
Marc
--
-----------------------------------------------------------------------------
Marc Haber | "I don't trust Computers. They | Mailadresse im Header
Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834
Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
On Wednesday 04 January 2006 17:18, Jesper Juhl wrote:
> > Is there one?
>
> No.
>
> What you do is you first revert the 2.6.14.5 patch so you are left
> with a 2.6.14 kernel, then you apply the 2.6.15 patch.
> For more info, please read Documentation/applying-patches.txt
> (http://sosdg.org/~coywolf/lxr/source/Documentation/applying-patches.txt)
I thought about doing it that way, but convinced myself it was too
complicated.
I see it is the right way (whatever that is).
I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
I suppose I have to backtrack and revert all those patches in order?
Thanks.
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
In article <[email protected]> (at Wed, 4 Jan 2006 17:10:37 +0000), Nick Warne <[email protected]> says:
> A stupid question - buggered if I can find a kernel patch from 2.6.14.5 to
> 2.6.15?
>
> Is there one?
$ interdiff -p1 -z patch-2.6.14.5.bz2 patch-2.6.15.bz2 > patch-2.6.15_2.6.14.5
--yoshfuji
On Wed, 2006-01-04 at 18:19 +0100, Marc Haber wrote:
> On Wed, Jan 04, 2006 at 05:10:37PM +0000, Nick Warne wrote:
> > A stupid question - buggered if I can find a kernel patch from 2.6.14.5 to
> > 2.6.15?
>
> You can try backing out the 2.6.14 -> 2.6.14.5 patch to generate a
> "pristine" 2.6.14 to which you can apply 2.6.14 -> 2.6.15.
>
> This mess is one reason while I usually keep old kernel tarballs
> around for a while.
This mess is why I use ketchup:
http://www.selenic.com/ketchup/
-- Steve
On Wed, 4 Jan 2006, Nick Warne wrote:
> On Wednesday 04 January 2006 17:18, Jesper Juhl wrote:
>
> > > Is there one?
> >
> > No.
> >
> > What you do is you first revert the 2.6.14.5 patch so you are left
> > with a 2.6.14 kernel, then you apply the 2.6.15 patch.
> > For more info, please read Documentation/applying-patches.txt
> > (http://sosdg.org/~coywolf/lxr/source/Documentation/applying-patches.txt)
>
> I thought about doing it that way, but convinced myself it was too
> complicated.
>
> I see it is the right way (whatever that is).
>
> I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
and how did you do that?
Noone supplies such incremental patches AFAIK.
> I suppose I have to backtrack and revert all those patches in order?
No, you can revert 2.6.14.5 directly to 2.6.14.
--
~Randy
On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
>
> and how did you do that?
> Noone supplies such incremental patches AFAIK.
Yes, I got from kernel.org - I am _not_ that clever to devise my own
incremental patches, otherwise I wouldn't be asking stupid questions...
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
On 1/4/06, Nick Warne <[email protected]> wrote:
> On Wednesday 04 January 2006 17:18, Jesper Juhl wrote:
>
> > > Is there one?
> >
> > No.
> >
> > What you do is you first revert the 2.6.14.5 patch so you are left
> > with a 2.6.14 kernel, then you apply the 2.6.15 patch.
> > For more info, please read Documentation/applying-patches.txt
> > (http://sosdg.org/~coywolf/lxr/source/Documentation/applying-patches.txt)
>
> I thought about doing it that way, but convinced myself it was too
> complicated.
>
> I see it is the right way (whatever that is).
>
> I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
>
If you did that you did it wrong. The -stable patches are *not*
incremental, they all apply to the base 2.6.x kernel.
What you should have done is :
2.6.14 -> 2.6.14.1
then before applying the 2.6.14.2 patch you should have reverted the
2.6.14.1 patch
2.6.14.1 -> 2.6.14
Then you go from 2.6.14 directly to 2.6.14.2
2.6.14 -> 2.6.14.2
etc...
2.6.14.2 -> 2.6.14
2.6.14 -> 2.6.14.3
2.6.14.3 -> 2.6.14
2.6.14 -> 2.6.14.4
2.6.14.4 -> 2.6.14
2.6.14 -> 2.6.14.5
If what you say you did above actually worked that's pure luck.
> I suppose I have to backtrack and revert all those patches in order?
>
No, just revert the 2.6.14.5 patch and you'll be left with a plain
2.6.14 to which you can then apply the 2.6.15 patch.
It's only ever "revert one, apply one".
I cover this in the "The 2.6.x.y kernels" section in
Documentation/applying-patches.txt , was that section not clear? If
not, then feel free to offer suggestions on how I can improve the
wording to make it more clear.
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wednesday 04 January 2006 17:40, Jesper Juhl wrote:
> > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
>
> If you did that you did it wrong. The -stable patches are *not*
> incremental, they all apply to the base 2.6.x kernel.
nick@linuxamd:kernel$ ls -lsa | grep patch
24 -rw-r--r-- 1 nick users 20572 2005-11-11 06:07 patch-2.6.14.2
48 -rw-r--r-- 1 nick users 46260 2005-11-24 22:15 patch-2.6.14.3
24 -rw-r--r-- 1 nick users 22725 2005-12-15 00:27 patch-2.6.14.3-4
20 -rw-r--r-- 1 nick users 18651 2005-12-27 00:29 patch-2.6.14.4-5
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
On 1/4/06, Nick Warne <[email protected]> wrote:
> On Wednesday 04 January 2006 17:40, Jesper Juhl wrote:
>
> > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> >
> > If you did that you did it wrong. The -stable patches are *not*
> > incremental, they all apply to the base 2.6.x kernel.
>
> nick@linuxamd:kernel$ ls -lsa | grep patch
>
> 24 -rw-r--r-- 1 nick users 20572 2005-11-11 06:07 patch-2.6.14.2
> 48 -rw-r--r-- 1 nick users 46260 2005-11-24 22:15 patch-2.6.14.3
> 24 -rw-r--r-- 1 nick users 22725 2005-12-15 00:27 patch-2.6.14.3-4
> 20 -rw-r--r-- 1 nick users 18651 2005-12-27 00:29 patch-2.6.14.4-5
>
$ ncftp ftp.kernel.org
NcFTP 3.1.9 (Mar 24, 2005) by Mike Gleason (http://www.NcFTP.com/contact/).
Connecting to 204.152.191.37...
Welcome to ftp.kernel.org.
Logging in...
<-- snip -->
ncftp / > cd /pub/linux/kernel/v2.6/
Directory successfully changed.
ncftp /pub/linux/kernel/v2.6 > ls -l patch-2.6.14.?.gz
-rw-r--r-- 1 536 536 2841 Nov 9 01:01 patch-2.6.14.1.gz
-rw-r--r-- 1 536 536 6566 Nov 11 06:07 patch-2.6.14.2.gz
-rw-rw-r-- 1 536 536 13849 Nov 24 22:15 patch-2.6.14.3.gz
-rw-r--r-- 1 536 536 21012 Dec 15 00:27 patch-2.6.14.4.gz
-rw-r--r-- 1 536 536 25943 Dec 27 00:29 patch-2.6.14.5.gz
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wed, 4 Jan 2006, Jesper Juhl wrote:
> On 1/4/06, Nick Warne <[email protected]> wrote:
> > On Wednesday 04 January 2006 17:40, Jesper Juhl wrote:
> >
> > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > >
> > > If you did that you did it wrong. The -stable patches are *not*
> > > incremental, they all apply to the base 2.6.x kernel.
> >
> > nick@linuxamd:kernel$ ls -lsa | grep patch
> >
> > 24 -rw-r--r-- 1 nick users 20572 2005-11-11 06:07 patch-2.6.14.2
> > 48 -rw-r--r-- 1 nick users 46260 2005-11-24 22:15 patch-2.6.14.3
> > 24 -rw-r--r-- 1 nick users 22725 2005-12-15 00:27 patch-2.6.14.3-4
> > 20 -rw-r--r-- 1 nick users 18651 2005-12-27 00:29 patch-2.6.14.4-5
> >
>
> $ ncftp ftp.kernel.org
> NcFTP 3.1.9 (Mar 24, 2005) by Mike Gleason (http://www.NcFTP.com/contact/).
> Connecting to 204.152.191.37...
> Welcome to ftp.kernel.org.
> Logging in...
> <-- snip -->
> ncftp / > cd /pub/linux/kernel/v2.6/
> Directory successfully changed.
> ncftp /pub/linux/kernel/v2.6 > ls -l patch-2.6.14.?.gz
> -rw-r--r-- 1 536 536 2841 Nov 9 01:01 patch-2.6.14.1.gz
> -rw-r--r-- 1 536 536 6566 Nov 11 06:07 patch-2.6.14.2.gz
> -rw-rw-r-- 1 536 536 13849 Nov 24 22:15 patch-2.6.14.3.gz
> -rw-r--r-- 1 536 536 21012 Dec 15 00:27 patch-2.6.14.4.gz
> -rw-r--r-- 1 536 536 25943 Dec 27 00:29 patch-2.6.14.5.gz
but the incremental patches do appear to be in
http://www.kernel.org/pub/linux/kernel/v2.6/incr/
who generates these? are they automated?
--
~Randy
On 1/4/06, Randy.Dunlap <[email protected]> wrote:
> On Wed, 4 Jan 2006, Jesper Juhl wrote:
>
> > On 1/4/06, Nick Warne <[email protected]> wrote:
> > > On Wednesday 04 January 2006 17:40, Jesper Juhl wrote:
> > >
> > > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > > >
> > > > If you did that you did it wrong. The -stable patches are *not*
> > > > incremental, they all apply to the base 2.6.x kernel.
> > >
> > > nick@linuxamd:kernel$ ls -lsa | grep patch
> > >
> > > 24 -rw-r--r-- 1 nick users 20572 2005-11-11 06:07 patch-2.6.14.2
> > > 48 -rw-r--r-- 1 nick users 46260 2005-11-24 22:15 patch-2.6.14.3
> > > 24 -rw-r--r-- 1 nick users 22725 2005-12-15 00:27 patch-2.6.14.3-4
> > > 20 -rw-r--r-- 1 nick users 18651 2005-12-27 00:29 patch-2.6.14.4-5
> > >
> >
> > $ ncftp ftp.kernel.org
> > NcFTP 3.1.9 (Mar 24, 2005) by Mike Gleason (http://www.NcFTP.com/contact/).
> > Connecting to 204.152.191.37...
> > Welcome to ftp.kernel.org.
> > Logging in...
> > <-- snip -->
> > ncftp / > cd /pub/linux/kernel/v2.6/
> > Directory successfully changed.
> > ncftp /pub/linux/kernel/v2.6 > ls -l patch-2.6.14.?.gz
> > -rw-r--r-- 1 536 536 2841 Nov 9 01:01 patch-2.6.14.1.gz
> > -rw-r--r-- 1 536 536 6566 Nov 11 06:07 patch-2.6.14.2.gz
> > -rw-rw-r-- 1 536 536 13849 Nov 24 22:15 patch-2.6.14.3.gz
> > -rw-r--r-- 1 536 536 21012 Dec 15 00:27 patch-2.6.14.4.gz
> > -rw-r--r-- 1 536 536 25943 Dec 27 00:29 patch-2.6.14.5.gz
>
> but the incremental patches do appear to be in
> http://www.kernel.org/pub/linux/kernel/v2.6/incr/
>
> who generates these? are they automated?
>
Hmm, yes, you are right. I was not aware of those. When did those
start to apear?
Guess I need to update applying-patches.txt if those are automated...
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wednesday 04 January 2006 17:51, Randy.Dunlap wrote:
>
> but the incremental patches do appear to be in
> http://www.kernel.org/pub/linux/kernel/v2.6/incr/
>
> who generates these? are they automated?
OMG - am I the only person in the world to be H4><0R3D from kernel.org...
Nick :-D
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
Jesper Juhl wrote:
>
>>but the incremental patches do appear to be in
>> http://www.kernel.org/pub/linux/kernel/v2.6/incr/
..
>
> Hmm, yes, you are right. I was not aware of those. When did those
> start to apear?
> Guess I need to update applying-patches.txt if those are automated...
That's how Greg posts them to LKML also -- as incremental patches.
-ml
On 1/4/06, Mark Lord <[email protected]> wrote:
> Jesper Juhl wrote:
> >
> >>but the incremental patches do appear to be in
> >> http://www.kernel.org/pub/linux/kernel/v2.6/incr/
> ..
> >
> > Hmm, yes, you are right. I was not aware of those. When did those
> > start to apear?
> > Guess I need to update applying-patches.txt if those are automated...
>
> That's how Greg posts them to LKML also -- as incremental patches.
>
Yes, I know that's what he posts them on LKML, I just never knew that
they got archived on kernel.org in incr. form as well. Now I know :-)
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wed, 04 Jan 2006, Nick Warne wrote:
> Hi all,
>
> A stupid question - buggered if I can find a kernel patch from 2.6.14.5 to
> 2.6.15?
>
> Is there one?
Use ketchup. <http://www.selenic.com/ketchup/> - just typing "ketchup
2.6" in your /usr/src/linux directory should download the required patch
files, verify their signatures and apply them. No hassles.
--
Matthias Andree
On Wed, 2006-01-04 at 17:56 +0000, Nick Warne wrote:
> On Wednesday 04 January 2006 17:51, Randy.Dunlap wrote:
>
> >
> > but the incremental patches do appear to be in
> > http://www.kernel.org/pub/linux/kernel/v2.6/incr/
> >
> > who generates these? are they automated?
>
> OMG - am I the only person in the world to be H4><0R3D from kernel.org...
I still find ketchup the easiest solution:
Here's a cut and paste of what I did. The commands that I did was as
follows:
1) mkdir tmp
2) cd tmp
3) ketchup -r -G -f 2.6.14
# comment
# -r is rename the directory
# -G is to not verify against PGP keys
# -f is to force downloading a tar ball instead of using linux-2.6.13.tar.bz2
# and patch against it.
4) cd .
5) head Makefile
6) ketchup -r -G 2.6.14.1
# comment
# here I removed the -f since I now want to patch
7) cd .
8) head Makefile
9) ketchup -r -G 2.6.14.5
10) cd .
11) head Makefile
12) ketchup -r -G 2.6.15
13) cd .
14) head Makefile
It's as easy as that. The above commands put me through 4 different
versions of Linux. And the "cd ." and "head Makefile" was only done to
show that ketchup worked.
---
rostedt@gandalf:~$ mkdir tmp
rostedt@gandalf:~$ cd tmp
rostedt@gandalf:~/tmp$ ketchup -r -G -f 2.6.14
None -> 2.6.14
Downloading linux-2.6.14.tar.bz2
--13:12:54-- http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.14.tar.bz2
=> `/home/rostedt/.ketchup/linux-2.6.14.tar.bz2.partial'
Resolving http://www.kernel.org... 204.152.191.5, 204.152.191.37
Connecting to http://www.kernel.org|204.152.191.5|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 39,172,170 (37M) [application/x-bzip2]
100%[====================================>] 39,172,170 523.51K/s ETA 00:00
13:14:09 (511.71 KB/s) - `/home/rostedt/.ketchup/linux-2.6.14.tar.bz2.partial' saved [39172170/39172170]
Unpacking linux-2.6.14.tar.bz2
Current directory renamed to /home/rostedt/linux-2.6.14
rostedt@gandalf:~/tmp$ cd .
rostedt@gandalf:~/linux-2.6.14$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 14
EXTRAVERSION =
NAME=Affluent Albatross
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.14$ ketchup -r -G 2.6.14.1
2.6.14 -> 2.6.14.1
Downloading patch-2.6.14.1.bz2
--13:16:38-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.1.bz2
=> `/home/rostedt/.ketchup/patch-2.6.14.1.bz2.partial'
Resolving http://www.kernel.org... 204.152.191.5, 204.152.191.37
Connecting to http://www.kernel.org|204.152.191.5|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 2,808 (2.7K) [application/x-bzip2]
100%[====================================>] 2,808 --.--K/s
13:16:39 (29.14 KB/s) - `/home/rostedt/.ketchup/patch-2.6.14.1.bz2.partial' saved [2808/2808]
Applying patch-2.6.14.1.bz2
Current directory renamed to /home/rostedt/linux-2.6.14.1
rostedt@gandalf:~/linux-2.6.14$ cd .
rostedt@gandalf:~/linux-2.6.14.1$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 14
EXTRAVERSION = .1
NAME=Affluent Albatross
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.14.1$ ketchup -r -G 2.6.14.5
2.6.14.1 -> 2.6.14.5
Applying patch-2.6.14.1.bz2 -R
Downloading patch-2.6.14.5.bz2
--13:16:50-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.5.bz2
=> `/home/rostedt/.ketchup/patch-2.6.14.5.bz2.partial'
Resolving http://www.kernel.org... 204.152.191.37, 204.152.191.5
Connecting to http://www.kernel.org|204.152.191.37|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 24,110 (24K) [application/x-bzip2]
100%[====================================>] 24,110 73.62K/s
13:16:51 (73.44 KB/s) - `/home/rostedt/.ketchup/patch-2.6.14.5.bz2.partial' saved [24110/24110]
Applying patch-2.6.14.5.bz2
Current directory renamed to /home/rostedt/linux-2.6.14.5
rostedt@gandalf:~/linux-2.6.14.1$ cd .
rostedt@gandalf:~/linux-2.6.14.5$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 14
EXTRAVERSION = .5
NAME=Affluent Albatross
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.14.5$ ketchup -r -G 2.6.15
2.6.14.5 -> 2.6.15
Applying patch-2.6.14.5.bz2 -R
Downloading patch-2.6.15.bz2
--13:17:03-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.15.bz2
=> `/home/rostedt/.ketchup/patch-2.6.15.bz2.partial'
Resolving http://www.kernel.org... 204.152.191.5, 204.152.191.37
Connecting to http://www.kernel.org|204.152.191.5|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6,254,721 (6.0M) [application/x-bzip2]
100%[====================================>] 6,254,721 377.90K/s ETA 00:00
13:17:18 (436.92 KB/s) - `/home/rostedt/.ketchup/patch-2.6.15.bz2.partial' saved [6254721/6254721]
Applying patch-2.6.15.bz2
Current directory renamed to /home/rostedt/linux-2.6.15
rostedt@gandalf:~/linux-2.6.14.5$ cd .
rostedt@gandalf:~/linux-2.6.15$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 15
EXTRAVERSION =
NAME=Sliding Snow Leopard
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.15$
---
-- Steve
On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> >
> > and how did you do that?
> > Noone supplies such incremental patches AFAIK.
>
> Yes, I got from kernel.org - I am _not_ that clever to devise my own
> incremental patches, otherwise I wouldn't be asking stupid questions...
Nick's right, both are provided automatically by kernel.org.
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wednesday 04 January 2006 18:34, Alistair John Strachan wrote:
> On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> > On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > >
> > > and how did you do that?
> > > Noone supplies such incremental patches AFAIK.
> >
> > Yes, I got from kernel.org - I am _not_ that clever to devise my own
> > incremental patches, otherwise I wouldn't be asking stupid questions...
>
> Nick's right, both are provided automatically by kernel.org.
Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then make
oldconfig etc.
I think there needs to be a way out of this that is easily discernible - it
does get confusing sometimes with all the patches flying around on a 'stable
release'.
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
Nick Warne wrote:
> Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then make
> oldconfig etc.
>
> I think there needs to be a way out of this that is easily discernible - it
> does get confusing sometimes with all the patches flying around on a 'stable
> release'.
>
> Nick
Whats wrong with ketchup?
--
[Name ] :: [Matan I. Peled ]
[Location ] :: [Israel ]
[Public Key] :: [0xD6F42CA5 ]
[Keyserver ] :: [keyserver.kjsl.com]
encrypted/signed plain text preferred
On Wednesday 04 January 2006 19:53, Nick Warne wrote:
> On Wednesday 04 January 2006 18:34, Alistair John Strachan wrote:
> > On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> > > On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > > >
> > > > and how did you do that?
> > > > Noone supplies such incremental patches AFAIK.
> > >
> > > Yes, I got from kernel.org - I am _not_ that clever to devise my own
> > > incremental patches, otherwise I wouldn't be asking stupid questions...
> >
> > Nick's right, both are provided automatically by kernel.org.
>
> Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then make
> oldconfig etc.
>
> I think there needs to be a way out of this that is easily discernible - it
> does get confusing sometimes with all the patches flying around on a
> 'stable release'.
It's documented in the kernel.
There's something in the kernel.org FAQ there about -rc kernels, but it might
be better to generalise this for stable releases. Added hpa to CC.
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wed, 4 Jan 2006 19:13:08 +0100, Jesper Juhl <[email protected]> wrote:
>On 1/4/06, Mark Lord <[email protected]> wrote:
>> Jesper Juhl wrote:
>> >
>> >>but the incremental patches do appear to be in
>> >> http://www.kernel.org/pub/linux/kernel/v2.6/incr/
>> ..
>> >
>> > Hmm, yes, you are right. I was not aware of those. When did those
>> > start to apear?
>> > Guess I need to update applying-patches.txt if those are automated...
>>
>> That's how Greg posts them to LKML also -- as incremental patches.
>>
>Yes, I know that's what he posts them on LKML, I just never knew that
>they got archived on kernel.org in incr. form as well. Now I know :-)
Easy to revert 2.6.14.5 to 2.6.14 then patch to 2.6.15 if you follow
the stable series. Still saves ~30MB downloading source.
Because I'm also compiling the -rc? and -mm? I keep 2.6.14 tree and use
hardlink trees, when 2.6.15 came out I patched the 2.6.14 to 2.6.15
and deleted the development trees -- works for me ;)
And no, I'll not automate this, it's enough dealing with finger trouble
let alone scripts doing things behind my back! I have vim trained to
break hardlink files on write, I don't get myself in trouble so much.
Grant.
On 1/4/06, Steven Rostedt <[email protected]> wrote:
> On Wed, 2006-01-04 at 17:56 +0000, Nick Warne wrote:
> > On Wednesday 04 January 2006 17:51, Randy.Dunlap wrote:
> >
> > >
> > > but the incremental patches do appear to be in
> > > http://www.kernel.org/pub/linux/kernel/v2.6/incr/
> > >
> > > who generates these? are they automated?
> >
> > OMG - am I the only person in the world to be H4><0R3D from kernel.org...
>
> I still find ketchup the easiest solution:
>
> Here's a cut and paste of what I did. The commands that I did was as
14 steps?
Surely a simple 5 step procedure is easier :
$ cd ~/linux-2.6.14.5 # change into the kernel source dir
$ patch -p1 -R < ../patch-2.6.14.5 # revert the 2.6.14.5 patch
$ patch -p1 < ../patch-2.6.15 # apply the new 2.6.15 patch
$ cd ..
$ mv linux-2.6.14.5 linux-2.6.15 # rename the kernel source dir
That's assuming you have already gunzip'ed the patch, but even if you
have not it's still just as easy :
$ cd ~/linux-2.6.14.5 # change into the kernel source dir
$ zcat ../patch-2.6.14.5.gz | patch -p1 -R # revert the 2.6.14.5 patch
$ zcat ../patch-2.6.15.gz | patch -p1 # apply the new 2.6.15 patch
$ cd ..
$ mv linux-2.6.14.5 linux-2.6.15 # rename the kernel source dir
No need to jump through 14 hoops.
And it's even nicely documented.
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On 1/4/06, Grant Coady <[email protected]> wrote:
> On Wed, 4 Jan 2006 19:13:08 +0100, Jesper Juhl <[email protected]> wrote:
>
> >On 1/4/06, Mark Lord <[email protected]> wrote:
> >> Jesper Juhl wrote:
> >> >
> >> >>but the incremental patches do appear to be in
> >> >> http://www.kernel.org/pub/linux/kernel/v2.6/incr/
> >> ..
> >> >
> >> > Hmm, yes, you are right. I was not aware of those. When did those
> >> > start to apear?
> >> > Guess I need to update applying-patches.txt if those are automated...
> >>
> >> That's how Greg posts them to LKML also -- as incremental patches.
> >>
> >Yes, I know that's what he posts them on LKML, I just never knew that
> >they got archived on kernel.org in incr. form as well. Now I know :-)
>
> Easy to revert 2.6.14.5 to 2.6.14 then patch to 2.6.15 if you follow
> the stable series. Still saves ~30MB downloading source.
>
Yup, no point in downloading everything all the time.
> Because I'm also compiling the -rc? and -mm? I keep 2.6.14 tree and use
> hardlink trees, when 2.6.15 came out I patched the 2.6.14 to 2.6.15
> and deleted the development trees -- works for me ;)
>
Very similar to what I do myself.
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wed, 2006-01-04 at 22:17 +0100, Jesper Juhl wrote:
>
> 14 steps?
Sorry I wasn't clearer. I was demonstrating how one can go from 2.6.14
to 2.6.14.1 to 2.6.14.5 then to 2.6.15. Not how to go from 2.6.14.5 to
2.6.15. Those were 14 steps of going from 2.6.14 to 2.6.14.1 to
2.6.14.5 to 2.6.15 with showing that it worked. If all i wanted to do
was to go from 2.6.14.5 to 2.6.15, it's one step: And it even downloads
and renames the directory for me :)
All patches are archived in ~/.ketchup so that it doesn't have to
download them each time. I actually went into that directory and
manually removed the patch so that it would force ketchup to download it
again, for this example.
===
rostedt@gandalf:~$ cd linux-2.6.14.5/
rostedt@gandalf:~/linux-2.6.14.5$ ketchup -r -G 2.6.15
2.6.14.5 -> 2.6.15
Applying patch-2.6.14.5.bz2 -R
Downloading patch-2.6.15.bz2
--16:26:05-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.15.bz2
=> `/home/rostedt/.ketchup/patch-2.6.15.bz2.partial'
Resolving http://www.kernel.org... 204.152.191.5, 204.152.191.37
Connecting to http://www.kernel.org|204.152.191.5|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6,254,721 (6.0M) [application/x-bzip2]
100%[====================================>] 6,254,721 475.16K/s ETA 00:00
16:26:21 (413.28 KB/s) - `/home/rostedt/.ketchup/patch-2.6.15.bz2.partial' saved [6254721/6254721]
Applying patch-2.6.15.bz2
Current directory renamed to /home/rostedt/linux-2.6.15
rostedt@gandalf:~/linux-2.6.14.5$ cd .
===
So removing all the crap that is spit out by ketchup, I have three
steps. Two to change directories, and one real ketchup step!
1) cd linux-2.6.14.5/
2) ketchup -r -G 2.6.15
3) cd .
Can it get any easier?
> Surely a simple 5 step procedure is easier :
>
> $ cd ~/linux-2.6.14.5 # change into the kernel source dir
> $ patch -p1 -R < ../patch-2.6.14.5 # revert the 2.6.14.5 patch
> $ patch -p1 < ../patch-2.6.15 # apply the new 2.6.15 patch
> $ cd ..
> $ mv linux-2.6.14.5 linux-2.6.15 # rename the kernel source dir
>
> That's assuming you have already gunzip'ed the patch, but even if you
> have not it's still just as easy :
>
> $ cd ~/linux-2.6.14.5 # change into the kernel source dir
> $ zcat ../patch-2.6.14.5.gz | patch -p1 -R # revert the 2.6.14.5 patch
> $ zcat ../patch-2.6.15.gz | patch -p1 # apply the new 2.6.15 patch
> $ cd ..
> $ mv linux-2.6.14.5 linux-2.6.15 # rename the kernel source dir
You never showed the step of downloading the patch. That was done for
me too. So if I would do it your way, and do what I did in those 14
steps...
====
rostedt@gandalf:~$ wget http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.14.tar.bz2
--16:31:54-- http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.14.tar.bz2
=> `linux-2.6.14.tar.bz2'
Resolving http://www.kernel.org... 204.152.191.5, 204.152.191.37
Connecting to http://www.kernel.org|204.152.191.5|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 39,172,170 (37M) [application/x-bzip2]
100%[====================================>] 39,172,170 423.28K/s ETA 00:00
16:33:19 (451.00 KB/s) - `linux-2.6.14.tar.bz2' saved [39172170/39172170]
rostedt@gandalf:~$ tar -xjf linux-2.6.14.tar.bz2
rostedt@gandalf:~$ cd linux-2.6.14
rostedt@gandalf:~/linux-2.6.14$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 14
EXTRAVERSION =
NAME=Affluent Albatross
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.14$ wget http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.1.bz2
--16:35:55-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.1.bz2
=> `patch-2.6.14.1.bz2'
Resolving http://www.kernel.org... 204.152.191.37, 204.152.191.5
Connecting to http://www.kernel.org|204.152.191.37|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 2,808 (2.7K) [application/x-bzip2]
100%[====================================>] 2,808 --.--K/s
16:35:56 (79.00 KB/s) - `patch-2.6.14.1.bz2' saved [2808/2808]
rostedt@gandalf:~/linux-2.6.14$ bzcat patch-2.6.14.1.bz2 | patch -p1 -s
rostedt@gandalf:~/linux-2.6.14$ mv ../linux-2.6.14 ../linux-2.6.14.1
rostedt@gandalf:~/linux-2.6.14$ cd .
rostedt@gandalf:~/linux-2.6.14.1$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 14
EXTRAVERSION = .1
NAME=Affluent Albatross
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.14.1$ bzcat patch-2.6.14.1.bz2 | patch -p1 -s -R
rostedt@gandalf:~/linux-2.6.14.1$ wget http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.5.bz2
--16:37:42-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.5.bz2
=> `patch-2.6.14.5.bz2'
Resolving http://www.kernel.org... 204.152.191.5, 204.152.191.37
Connecting to http://www.kernel.org|204.152.191.5|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 24,110 (24K) [application/x-bzip2]
100%[====================================>] 24,110 69.39K/s
16:37:43 (69.38 KB/s) - `patch-2.6.14.5.bz2' saved [24110/24110]
rostedt@gandalf:~/linux-2.6.14.1$ bzcat patch-2.6.14.5.bz2 | patch -p1 -s rostedt@gandalf:~/linux-2.6.14.1$ mv ../linux-2.6.14.1 ../linux-2.6.14.5
rostedt@gandalf:~/linux-2.6.14.1$ cd .
rostedt@gandalf:~/linux-2.6.14.5$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 14
EXTRAVERSION = .5
NAME=Affluent Albatross
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
rostedt@gandalf:~/linux-2.6.14.5$ bzcat patch-2.6.14.5.bz2 | patch -p1 -s -R
rostedt@gandalf:~/linux-2.6.14.5$ wget http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.15.bz2
--16:38:44-- http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.15.bz2
=> `patch-2.6.15.bz2'
Resolving http://www.kernel.org... 204.152.191.37, 204.152.191.5
Connecting to http://www.kernel.org|204.152.191.37|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6,254,721 (6.0M) [application/x-bzip2]
100%[====================================>] 6,254,721 386.49K/s ETA 00:00
16:39:02 (344.52 KB/s) - `patch-2.6.15.bz2' saved [6254721/6254721]
rostedt@gandalf:~/linux-2.6.14.5$ bzcat patch-2.6.15.bz2 | patch -p1 -s rostedt@gandalf:~/linux-2.6.14.5$ mv ../linux-2.6.14.5 ../linux-2.6.15
rostedt@gandalf:~/linux-2.6.14.5$ cd .
rostedt@gandalf:~/linux-2.6.15$ head Makefile
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 15
EXTRAVERSION =
NAME=Sliding Snow Leopard
# *DOCUMENTATION*
# To see a list of typical targets execute "make help"
# More info can be located in ./README
# Comments in this file are targeted only to the developer, do not
====
Here's the steps I did:
1) wget http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.14.tar.bz2
2) tar -xjf linux-2.6.14.tar.bz2
3) cd linux-2.6.14
4) head Makefile
5) wget http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.1.bz2
6) bzcat patch-2.6.14.1.bz2 | patch -p1 -s
7) mv ../linux-2.6.14 ../linux-2.6.14.1
8) cd .
9) head Makefile
10) bzcat patch-2.6.14.1.bz2 | patch -p1 -s -R
11) wget http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.14.5.bz2
12) bzcat patch-2.6.14.5.bz2 | patch -p1 -s
13) mv ../linux-2.6.14.1 ../linux-2.6.14.5
14) cd .
15) head Makefile
16) bzcat patch-2.6.14.5.bz2 | patch -p1 -s -R
17) wget http://www.kernel.org/pub/linux/kernel/v2.6/patch-2.6.15.bz2
18) bzcat patch-2.6.15.bz2 | patch -p1 -s
19) mv ../linux-2.6.14.5 ../linux-2.6.15
20) cd .
21) head Makefile
22 steps is the equivalent of ketchup's 14, not to mention you need to
type a hell of a lot more, and remember where to download the patches
from. Also, you need to manage the patches you download.
-- Steve
On Wed, Jan 04, 2006 at 06:34:23PM +0000, Alistair John Strachan wrote:
> On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> > On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > >
> > > and how did you do that?
> > > Noone supplies such incremental patches AFAIK.
> >
> > Yes, I got from kernel.org - I am _not_ that clever to devise my own
> > incremental patches, otherwise I wouldn't be asking stupid questions...
>
> Nick's right, both are provided automatically by kernel.org.
Not "automatically", the -stable team lovingly hand crafts them for
every release, just because we are a people-pleasing group.
Well ok, we have a script that does it for us as part of our release,
but someone had to write that script, and it isn't automatically
generated by kernel.org like some of the other interdiffs are...
thanks,
greg k-h
On Wed, Jan 04, 2006 at 08:10:36PM +0000, Alistair John Strachan wrote:
> On Wednesday 04 January 2006 19:53, Nick Warne wrote:
> > On Wednesday 04 January 2006 18:34, Alistair John Strachan wrote:
> > > On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> > > > On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > > > >
> > > > > and how did you do that?
> > > > > Noone supplies such incremental patches AFAIK.
> > > >
> > > > Yes, I got from kernel.org - I am _not_ that clever to devise my own
> > > > incremental patches, otherwise I wouldn't be asking stupid questions...
> > >
> > > Nick's right, both are provided automatically by kernel.org.
> >
> > Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then make
> > oldconfig etc.
> >
> > I think there needs to be a way out of this that is easily discernible - it
> > does get confusing sometimes with all the patches flying around on a
> > 'stable release'.
>
> It's documented in the kernel.
>
> There's something in the kernel.org FAQ there about -rc kernels, but it might
> be better to generalise this for stable releases. Added hpa to CC.
What do you mean, "generalize" this? Where else could we document it
better?
thanks,
greg k-h
On Wednesday 04 January 2006 22:01, Greg KH wrote:
> > > > Nick's right, both are provided automatically by kernel.org.
> > >
> > > Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then
> > > make oldconfig etc.
> > >
> > > I think there needs to be a way out of this that is easily discernible
> > > - it does get confusing sometimes with all the patches flying around on
> > > a 'stable release'.
> >
> > It's documented in the kernel.
> >
> > There's something in the kernel.org FAQ there about -rc kernels, but it
> > might be better to generalise this for stable releases. Added hpa to CC.
>
> What do you mean, "generalize" this? Where else could we document it
> better?
The issue I hit was we have a 'latest stable kernel release 2.6.14.5' and
under it a 'the latest stable kernel' (or words to that effect) on
kernel.org.
Then when 2.6.15 came out, that was it! No patch for the 'latest stable
kernel release 2.6.14.5'. It was GONE!
OK, I suppose we are all capable of getting back to where we are on rebuilding
to latest 'stable', but there _is_ a missing link for somebody that doesn't
know - and I think backtracking patches isn't really the way to go if the
'latest stable release' isn't catered for.
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
On Wed, Jan 04, 2006 at 10:10:47PM +0000, Nick Warne wrote:
> On Wednesday 04 January 2006 22:01, Greg KH wrote:
>
> > > > > Nick's right, both are provided automatically by kernel.org.
> > > >
> > > > Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then
> > > > make oldconfig etc.
> > > >
> > > > I think there needs to be a way out of this that is easily discernible
> > > > - it does get confusing sometimes with all the patches flying around on
> > > > a 'stable release'.
> > >
> > > It's documented in the kernel.
> > >
> > > There's something in the kernel.org FAQ there about -rc kernels, but it
> > > might be better to generalise this for stable releases. Added hpa to CC.
> >
> > What do you mean, "generalize" this? Where else could we document it
> > better?
>
> The issue I hit was we have a 'latest stable kernel release 2.6.14.5' and
> under it a 'the latest stable kernel' (or words to that effect) on
> kernel.org.
>
> Then when 2.6.15 came out, that was it! No patch for the 'latest stable
> kernel release 2.6.14.5'. It was GONE!
That's because 2.6.15 is the latest stable release.
> OK, I suppose we are all capable of getting back to where we are on rebuilding
> to latest 'stable', but there _is_ a missing link for somebody that doesn't
> know - and I think backtracking patches isn't really the way to go if the
> 'latest stable release' isn't catered for.
Um, it is, see my sentance above. And if you want to download older
stable releases, you can jump to the proper directory, how long do you
want us to put older stable releases on the main page for? :)
thanks,
greg k-h
On Wed, 4 Jan 2006, Nick Warne wrote:
> On Wednesday 04 January 2006 22:01, Greg KH wrote:
>
> > > > > Nick's right, both are provided automatically by kernel.org.
> > > >
> > > > Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then
> > > > make oldconfig etc.
> > > >
> > > > I think there needs to be a way out of this that is easily discernible
> > > > - it does get confusing sometimes with all the patches flying around on
> > > > a 'stable release'.
> > >
> > > It's documented in the kernel.
> > >
> > > There's something in the kernel.org FAQ there about -rc kernels, but it
> > > might be better to generalise this for stable releases. Added hpa to CC.
> >
> > What do you mean, "generalize" this? Where else could we document it
> > better?
>
> The issue I hit was we have a 'latest stable kernel release 2.6.14.5' and
> under it a 'the latest stable kernel' (or words to that effect) on
> kernel.org.
>
> Then when 2.6.15 came out, that was it! No patch for the 'latest stable
> kernel release 2.6.14.5'. It was GONE!
Yes, I brought this up a couple of weeks ago, but I was told
that I was wrong (in some such words).
I agree that it needs to be fixed.
> OK, I suppose we are all capable of getting back to where we are on rebuilding
> to latest 'stable', but there _is_ a missing link for somebody that doesn't
> know - and I think backtracking patches isn't really the way to go if the
> 'latest stable release' isn't catered for.
--
~Randy
On Wed, 2006-01-04 at 16:48 -0500, Steven Rostedt wrote:
> 21) head Makefile
>
>
> 22 steps is the equivalent of ketchup's 14, not to mention you need to
OK, after doing all that, I couldn't count anymore. I meant 21 steps,
not 22 :P
-- Steve
On Wednesday 04 January 2006 22:15, Greg KH wrote:
> > Then when 2.6.15 came out, that was it! No patch for the 'latest stable
> > kernel release 2.6.14.5'. It was GONE!
>
> That's because 2.6.15 is the latest stable release.
>
> > OK, I suppose we are all capable of getting back to where we are on
> > rebuilding to latest 'stable', but there _is_ a missing link for somebody
> > that doesn't know - and I think backtracking patches isn't really the way
> > to go if the 'latest stable release' isn't catered for.
>
> Um, it is, see my sentance above. And if you want to download older
> stable releases, you can jump to the proper directory, how long do you
> want us to put older stable releases on the main page for? :)
OK, I see what you mean, but 2.6.14 wasn't the latest 'release' - 2.6.14.5 was
(according to kernel.org). Yet there is no upgrade path for that build (or
any other .x releases)
It is a bit of a mess really.
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/
On 1/4/06, Nick Warne <[email protected]> wrote:
> On Wednesday 04 January 2006 22:15, Greg KH wrote:
>
> > > Then when 2.6.15 came out, that was it! No patch for the 'latest stable
> > > kernel release 2.6.14.5'. It was GONE!
> >
> > That's because 2.6.15 is the latest stable release.
> >
> > > OK, I suppose we are all capable of getting back to where we are on
> > > rebuilding to latest 'stable', but there _is_ a missing link for somebody
> > > that doesn't know - and I think backtracking patches isn't really the way
> > > to go if the 'latest stable release' isn't catered for.
> >
> > Um, it is, see my sentance above. And if you want to download older
> > stable releases, you can jump to the proper directory, how long do you
> > want us to put older stable releases on the main page for? :)
>
> OK, I see what you mean, but 2.6.14 wasn't the latest 'release' - 2.6.14.5 was
2.6.14 was indeed the latest mainline/Linus release, all the 2.6.14.x
kernels were -stable kernels released by the -stable team, and when
2.6.15 shows up noone knows how many further -stable kernels the team
will release for 2.6.14 (most likely 1 at most, but it's not set in
stone).
> (according to kernel.org). Yet there is no upgrade path for that build (or
> any other .x releases)
>
> It is a bit of a mess really.
>
but, a 2.6.14.6 kernel might come out *after* 2.6.15, then what?
There's a 2.6.15 patch on the kernel.org frontpage that's a 2.6.14 ->
2.6.15 delta, if people are using 2.6.14.x then I think it's fair to
assume they have that .x patch around somewhere (or know where to find
it) and can easily revert it to then apply the 2.6.15 patch.
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Wed, Jan 04, 2006 at 02:16:37PM -0800, Randy.Dunlap wrote:
> On Wed, 4 Jan 2006, Nick Warne wrote:
>
> > On Wednesday 04 January 2006 22:01, Greg KH wrote:
> >
> > > > > > Nick's right, both are provided automatically by kernel.org.
> > > > >
> > > > > Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then
> > > > > make oldconfig etc.
> > > > >
> > > > > I think there needs to be a way out of this that is easily discernible
> > > > > - it does get confusing sometimes with all the patches flying around on
> > > > > a 'stable release'.
> > > >
> > > > It's documented in the kernel.
> > > >
> > > > There's something in the kernel.org FAQ there about -rc kernels, but it
> > > > might be better to generalise this for stable releases. Added hpa to CC.
> > >
> > > What do you mean, "generalize" this? Where else could we document it
> > > better?
> >
> > The issue I hit was we have a 'latest stable kernel release 2.6.14.5' and
> > under it a 'the latest stable kernel' (or words to that effect) on
> > kernel.org.
> >
> > Then when 2.6.15 came out, that was it! No patch for the 'latest stable
> > kernel release 2.6.14.5'. It was GONE!
>
> Yes, I brought this up a couple of weeks ago, but I was told
> that I was wrong (in some such words).
> I agree that it needs to be fixed.
How would you suggest that it be fixed?
thanks,
greg k-h
On Wed, Jan 04, 2006 at 10:20:59PM +0000, Nick Warne wrote:
> On Wednesday 04 January 2006 22:15, Greg KH wrote:
>
> > > Then when 2.6.15 came out, that was it! No patch for the 'latest stable
> > > kernel release 2.6.14.5'. It was GONE!
> >
> > That's because 2.6.15 is the latest stable release.
> >
> > > OK, I suppose we are all capable of getting back to where we are on
> > > rebuilding to latest 'stable', but there _is_ a missing link for somebody
> > > that doesn't know - and I think backtracking patches isn't really the way
> > > to go if the 'latest stable release' isn't catered for.
> >
> > Um, it is, see my sentance above. And if you want to download older
> > stable releases, you can jump to the proper directory, how long do you
> > want us to put older stable releases on the main page for? :)
>
> OK, I see what you mean, but 2.6.14 wasn't the latest 'release' - 2.6.14.5 was
> (according to kernel.org). Yet there is no upgrade path for that build (or
> any other .x releases)
>
> It is a bit of a mess really.
Huh? I'm confused, what exactly would you like the kernel.org site to
look like?
thanks,
greg k-h
On Wednesday 04 January 2006 22:30, Jesper Juhl wrote:
> > (according to kernel.org). Yet there is no upgrade path for that build
> > (or any other .x releases)
> >
> > It is a bit of a mess really.
>
> but, a 2.6.14.6 kernel might come out *after* 2.6.15, then what?
Nightmares...
I get all the points. Let me say then it needs better distinction on what the
'latest kernel' is and on what the 'latest -stable' is.
Nick
--
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
On Wednesday 04 January 2006 22:01, Greg KH wrote:
> On Wed, Jan 04, 2006 at 08:10:36PM +0000, Alistair John Strachan wrote:
> > On Wednesday 04 January 2006 19:53, Nick Warne wrote:
> > > On Wednesday 04 January 2006 18:34, Alistair John Strachan wrote:
> > > > On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> > > > > On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > > > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > > > > >
> > > > > > and how did you do that?
> > > > > > Noone supplies such incremental patches AFAIK.
> > > > >
> > > > > Yes, I got from kernel.org - I am _not_ that clever to devise my
> > > > > own incremental patches, otherwise I wouldn't be asking stupid
> > > > > questions...
> > > >
> > > > Nick's right, both are provided automatically by kernel.org.
> > >
> > > Anyway, I started from scratch - 2.6.14, patched to 2.6.15 and then
> > > make oldconfig etc.
> > >
> > > I think there needs to be a way out of this that is easily discernible
> > > - it does get confusing sometimes with all the patches flying around on
> > > a 'stable release'.
> >
> > It's documented in the kernel.
> >
> > There's something in the kernel.org FAQ there about -rc kernels, but it
> > might be better to generalise this for stable releases. Added hpa to CC.
>
> What do you mean, "generalize" this? Where else could we document it
> better?
Re-read the thread. The confusion here is about "going back" to 2.6.14 before
patching 2.6.15. This has nothing to do with "rc kernels". We have this
documented explicitly in the kernel but not on the kernel.org FAQ.
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wed, 4 Jan 2006 22:20:59 +0000, Nick Warne <[email protected]> wrote:
>On Wednesday 04 January 2006 22:15, Greg KH wrote:
>
>> > Then when 2.6.15 came out, that was it! No patch for the 'latest stable
>> > kernel release 2.6.14.5'. It was GONE!
>>
>> That's because 2.6.15 is the latest stable release.
>>
>> > OK, I suppose we are all capable of getting back to where we are on
>> > rebuilding to latest 'stable', but there _is_ a missing link for somebody
>> > that doesn't know - and I think backtracking patches isn't really the way
>> > to go if the 'latest stable release' isn't catered for.
>>
>> Um, it is, see my sentance above. And if you want to download older
>> stable releases, you can jump to the proper directory, how long do you
>> want us to put older stable releases on the main page for? :)
>
>OK, I see what you mean, but 2.6.14 wasn't the latest 'release' - 2.6.14.5 was
>(according to kernel.org). Yet there is no upgrade path for that build (or
>any other .x releases)
>
>It is a bit of a mess really.
Nah, search the archives for 'sucker tree' -- was much worse before
we have the stable bugfix patchlets in between the main kernel
releases, there was never going to be a 2.6.X.Y to 2.6.X+1 patch as
mainstream kernel development may fix issues entirely differently.
Kernel development is an ongoing process, the stable trees are
temporary branches outside of (or beside) mainstream development,
that easier for you?
Grant.
On Wednesday 04 January 2006 22:01, Greg KH wrote:
> On Wed, Jan 04, 2006 at 06:34:23PM +0000, Alistair John Strachan wrote:
> > On Wednesday 04 January 2006 17:39, Nick Warne wrote:
> > > On Wednesday 04 January 2006 17:36, Randy.Dunlap wrote:
> > > > > I went from 2.6.14 -> 2.6.14.2 -> .2-.3 -> .3-.4 -> .4-.5
> > > >
> > > > and how did you do that?
> > > > Noone supplies such incremental patches AFAIK.
> > >
> > > Yes, I got from kernel.org - I am _not_ that clever to devise my own
> > > incremental patches, otherwise I wouldn't be asking stupid questions...
> >
> > Nick's right, both are provided automatically by kernel.org.
>
> Not "automatically", the -stable team lovingly hand crafts them for
> every release, just because we are a people-pleasing group.
>
> Well ok, we have a script that does it for us as part of our release,
> but someone had to write that script, and it isn't automatically
> generated by kernel.org like some of the other interdiffs are...
My mistake Greg ;-)
We love the stable team all the more for it.
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wednesday 04 January 2006 22:31, Greg KH wrote:
[snip]
> > >
> > > The issue I hit was we have a 'latest stable kernel release 2.6.14.5'
> > > and under it a 'the latest stable kernel' (or words to that effect) on
> > > kernel.org.
> > >
> > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > stable kernel release 2.6.14.5'. It was GONE!
> >
> > Yes, I brought this up a couple of weeks ago, but I was told
> > that I was wrong (in some such words).
> > I agree that it needs to be fixed.
>
> How would you suggest that it be fixed?
It's difficult, but perhaps providing a link to the latest "stable team"
release in addition to Linus's release would solve the problem.
At least then you can do what Nick wanted (assuming the kernel.org FAQ gets
fixed) and download the "patch" for 2.6.14.5, say, revert it, then apply
Linus's latest and greatest (one or more times as required).
Bloats the front page though. I think as long as something is documented
properly it doesn't really matter. Currently it isn't.
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wed, Jan 04, 2006 at 10:36:10PM +0000, Nick Warne wrote:
> On Wednesday 04 January 2006 22:30, Jesper Juhl wrote:
>
> > > (according to kernel.org). Yet there is no upgrade path for that build
> > > (or any other .x releases)
> > >
> > > It is a bit of a mess really.
> >
> > but, a 2.6.14.6 kernel might come out *after* 2.6.15, then what?
>
> Nightmares...
It's already happened, we did a 2.6.13.y release after 2.6.14 was out,
and I'm about to do a 2.6.14.y release in a few days too :)
> I get all the points. Let me say then it needs better distinction on
> what the 'latest kernel' is and on what the 'latest -stable' is.
Ok, exactly how would you like this to be shown?
Again, 2.6.15 is the latest stable kernel right now. Just because there
isn't a .1 release yet, doesn't mean it isn't.
thanks,
greg k-h
On Wed, Jan 04, 2006 at 10:49:11PM +0000, Alistair John Strachan wrote:
> Re-read the thread. The confusion here is about "going back" to 2.6.14 before
> patching 2.6.15. This has nothing to do with "rc kernels". We have this
> documented explicitly in the kernel but not on the kernel.org FAQ.
The kernel.org FAQ does not deal with Linux kernel specific things, only
kernel.org specific things. So documenting it in the kernel itself is
the proper place for it :)
thanks,
greg k-h
On Wed, Jan 04, 2006 at 10:58:24PM +0000, Alistair John Strachan wrote:
> On Wednesday 04 January 2006 22:31, Greg KH wrote:
> [snip]
> > > >
> > > > The issue I hit was we have a 'latest stable kernel release 2.6.14.5'
> > > > and under it a 'the latest stable kernel' (or words to that effect) on
> > > > kernel.org.
> > > >
> > > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > > stable kernel release 2.6.14.5'. It was GONE!
> > >
> > > Yes, I brought this up a couple of weeks ago, but I was told
> > > that I was wrong (in some such words).
> > > I agree that it needs to be fixed.
> >
> > How would you suggest that it be fixed?
>
> It's difficult, but perhaps providing a link to the latest "stable team"
> release in addition to Linus's release would solve the problem.
But what happens when we release a 2.6.14.y release and a 2.6.15.y
release at the same time (as people have requested this in previous
threads...)? What would show up where?
thanks,
greg k-h
On Wednesday 04 January 2006 23:12, Greg KH wrote:
> On Wed, Jan 04, 2006 at 10:49:11PM +0000, Alistair John Strachan wrote:
> > Re-read the thread. The confusion here is about "going back" to 2.6.14
> > before patching 2.6.15. This has nothing to do with "rc kernels". We have
> > this documented explicitly in the kernel but not on the kernel.org FAQ.
>
> The kernel.org FAQ does not deal with Linux kernel specific things, only
> kernel.org specific things. So documenting it in the kernel itself is
> the proper place for it :)
The note on -rc kernels refutes this claim. How difficult it is really to
modify the sentence?
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wednesday 04 January 2006 23:13, Greg KH wrote:
> On Wed, Jan 04, 2006 at 10:58:24PM +0000, Alistair John Strachan wrote:
> > On Wednesday 04 January 2006 22:31, Greg KH wrote:
> > [snip]
> >
> > > > > The issue I hit was we have a 'latest stable kernel release
> > > > > 2.6.14.5' and under it a 'the latest stable kernel' (or words to
> > > > > that effect) on kernel.org.
> > > > >
> > > > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > > > stable kernel release 2.6.14.5'. It was GONE!
> > > >
> > > > Yes, I brought this up a couple of weeks ago, but I was told
> > > > that I was wrong (in some such words).
> > > > I agree that it needs to be fixed.
> > >
> > > How would you suggest that it be fixed?
> >
> > It's difficult, but perhaps providing a link to the latest "stable team"
> > release in addition to Linus's release would solve the problem.
>
> But what happens when we release a 2.6.14.y release and a 2.6.15.y
> release at the same time (as people have requested this in previous
> threads...)? What would show up where?
You're right, it's complicated. In that case I'd still opt for showing
2.6.15.y, as the vast majority of people manually installing vanilla kernels
will either be on the latest-ish kernel, or have a clue about what they're
doing (who doesn't know the ftp URL off by heart now).
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On Wed, 4 Jan 2006, Alistair John Strachan wrote:
> On Wednesday 04 January 2006 23:13, Greg KH wrote:
> > On Wed, Jan 04, 2006 at 10:58:24PM +0000, Alistair John Strachan wrote:
> > > On Wednesday 04 January 2006 22:31, Greg KH wrote:
> > > [snip]
> > >
> > > > > > The issue I hit was we have a 'latest stable kernel release
> > > > > > 2.6.14.5' and under it a 'the latest stable kernel' (or words to
> > > > > > that effect) on kernel.org.
> > > > > >
> > > > > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > > > > stable kernel release 2.6.14.5'. It was GONE!
> > > > >
> > > > > Yes, I brought this up a couple of weeks ago, but I was told
> > > > > that I was wrong (in some such words).
> > > > > I agree that it needs to be fixed.
> > > >
> > > > How would you suggest that it be fixed?
> > >
> > > It's difficult, but perhaps providing a link to the latest "stable team"
> > > release in addition to Linus's release would solve the problem.
> >
> > But what happens when we release a 2.6.14.y release and a 2.6.15.y
> > release at the same time (as people have requested this in previous
> > threads...)? What would show up where?
>
> You're right, it's complicated. In that case I'd still opt for showing
> 2.6.15.y, as the vast majority of people manually installing vanilla kernels
> will either be on the latest-ish kernel, or have a clue about what they're
> doing (who doesn't know the ftp URL off by heart now).
I agree. I think that one previous -stable patch version should always
be listed there, even if we think that 2.6.N is stable. :)
--
~Randy
On Wed, Jan 04, 2006 at 03:31:01PM -0800, Randy.Dunlap wrote:
> On Wed, 4 Jan 2006, Alistair John Strachan wrote:
>
> > On Wednesday 04 January 2006 23:13, Greg KH wrote:
> > > On Wed, Jan 04, 2006 at 10:58:24PM +0000, Alistair John Strachan wrote:
> > > > On Wednesday 04 January 2006 22:31, Greg KH wrote:
> > > > [snip]
> > > >
> > > > > > > The issue I hit was we have a 'latest stable kernel release
> > > > > > > 2.6.14.5' and under it a 'the latest stable kernel' (or words to
> > > > > > > that effect) on kernel.org.
> > > > > > >
> > > > > > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > > > > > stable kernel release 2.6.14.5'. It was GONE!
> > > > > >
> > > > > > Yes, I brought this up a couple of weeks ago, but I was told
> > > > > > that I was wrong (in some such words).
> > > > > > I agree that it needs to be fixed.
> > > > >
> > > > > How would you suggest that it be fixed?
> > > >
> > > > It's difficult, but perhaps providing a link to the latest "stable team"
> > > > release in addition to Linus's release would solve the problem.
> > >
> > > But what happens when we release a 2.6.14.y release and a 2.6.15.y
> > > release at the same time (as people have requested this in previous
> > > threads...)? What would show up where?
> >
> > You're right, it's complicated. In that case I'd still opt for showing
> > 2.6.15.y, as the vast majority of people manually installing vanilla kernels
> > will either be on the latest-ish kernel, or have a clue about what they're
> > doing (who doesn't know the ftp URL off by heart now).
>
> I agree. I think that one previous -stable patch version should always
> be listed there, even if we think that 2.6.N is stable. :)
I don't at all. If we do that, people will assume that they need to
wait till 2.6.N.1 before trying that kernel (as it wouldn't be "stable"
otherwise.) So no one will test it, to really generate the bug reports
that we need to get to that .1 release.
Or should we just throw out a .1 release with the first simple patch
that comes along just to make the kernel.org page update properly? I
don't think so...
thanks,
greg k-h
On Wed, 4 Jan 2006, Greg KH wrote:
> On Wed, Jan 04, 2006 at 03:31:01PM -0800, Randy.Dunlap wrote:
> > On Wed, 4 Jan 2006, Alistair John Strachan wrote:
> >
> > > On Wednesday 04 January 2006 23:13, Greg KH wrote:
> > > > On Wed, Jan 04, 2006 at 10:58:24PM +0000, Alistair John Strachan wrote:
> > > > > On Wednesday 04 January 2006 22:31, Greg KH wrote:
> > > > > [snip]
> > > > >
> > > > > > > > The issue I hit was we have a 'latest stable kernel release
> > > > > > > > 2.6.14.5' and under it a 'the latest stable kernel' (or words to
> > > > > > > > that effect) on kernel.org.
> > > > > > > >
> > > > > > > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > > > > > > stable kernel release 2.6.14.5'. It was GONE!
> > > > > > >
> > > > > > > Yes, I brought this up a couple of weeks ago, but I was told
> > > > > > > that I was wrong (in some such words).
> > > > > > > I agree that it needs to be fixed.
> > > > > >
> > > > > > How would you suggest that it be fixed?
> > > > >
> > > > > It's difficult, but perhaps providing a link to the latest "stable team"
> > > > > release in addition to Linus's release would solve the problem.
> > > >
> > > > But what happens when we release a 2.6.14.y release and a 2.6.15.y
> > > > release at the same time (as people have requested this in previous
> > > > threads...)? What would show up where?
> > >
> > > You're right, it's complicated. In that case I'd still opt for showing
> > > 2.6.15.y, as the vast majority of people manually installing vanilla kernels
> > > will either be on the latest-ish kernel, or have a clue about what they're
> > > doing (who doesn't know the ftp URL off by heart now).
> >
> > I agree. I think that one previous -stable patch version should always
> > be listed there, even if we think that 2.6.N is stable. :)
>
> I don't at all. If we do that, people will assume that they need to
> wait till 2.6.N.1 before trying that kernel (as it wouldn't be "stable"
> otherwise.) So no one will test it, to really generate the bug reports
> that we need to get to that .1 release.
>
> Or should we just throw out a .1 release with the first simple patch
> that comes along just to make the kernel.org page update properly? I
> don't think so...
and the circle continues.
You are reading too much PR. 8:)
--
~Randy
On 1/5/06, Greg KH <[email protected]> wrote:
> On Wed, Jan 04, 2006 at 03:31:01PM -0800, Randy.Dunlap wrote:
> > On Wed, 4 Jan 2006, Alistair John Strachan wrote:
> >
> > > On Wednesday 04 January 2006 23:13, Greg KH wrote:
> > > > On Wed, Jan 04, 2006 at 10:58:24PM +0000, Alistair John Strachan wrote:
> > > > > On Wednesday 04 January 2006 22:31, Greg KH wrote:
> > > > > [snip]
> > > > >
> > > > > > > > The issue I hit was we have a 'latest stable kernel release
> > > > > > > > 2.6.14.5' and under it a 'the latest stable kernel' (or words to
> > > > > > > > that effect) on kernel.org.
> > > > > > > >
> > > > > > > > Then when 2.6.15 came out, that was it! No patch for the 'latest
> > > > > > > > stable kernel release 2.6.14.5'. It was GONE!
> > > > > > >
> > > > > > > Yes, I brought this up a couple of weeks ago, but I was told
> > > > > > > that I was wrong (in some such words).
> > > > > > > I agree that it needs to be fixed.
> > > > > >
> > > > > > How would you suggest that it be fixed?
> > > > >
> > > > > It's difficult, but perhaps providing a link to the latest "stable team"
> > > > > release in addition to Linus's release would solve the problem.
> > > >
> > > > But what happens when we release a 2.6.14.y release and a 2.6.15.y
> > > > release at the same time (as people have requested this in previous
> > > > threads...)? What would show up where?
> > >
> > > You're right, it's complicated. In that case I'd still opt for showing
> > > 2.6.15.y, as the vast majority of people manually installing vanilla kernels
> > > will either be on the latest-ish kernel, or have a clue about what they're
> > > doing (who doesn't know the ftp URL off by heart now).
> >
> > I agree. I think that one previous -stable patch version should always
> > be listed there, even if we think that 2.6.N is stable. :)
>
> I don't at all. If we do that, people will assume that they need to
> wait till 2.6.N.1 before trying that kernel (as it wouldn't be "stable"
> otherwise.) So no one will test it, to really generate the bug reports
> that we need to get to that .1 release.
>
> Or should we just throw out a .1 release with the first simple patch
> that comes along just to make the kernel.org page update properly? I
> don't think so...
>
How about simply stating a bit more clearly for the 2.6.15 release
that "This patch is based on the base 2.6.14 kernel tree" and then
also provide a link to documentation explaining how to patch between
kernel versions and place text next to that link along the lines of
"If you want to know how to patch from your current kernel version to
the latest stable, please read the document at <some url>" - feel free
to copy at will from Documentation/applying-patches.txt for that <some
url> bit...
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
Alistair John Strachan wrote:
>
> Re-read the thread. The confusion here is about "going back" to 2.6.14 before
> patching 2.6.15. This has nothing to do with "rc kernels". We have this
> documented explicitly in the kernel but not on the kernel.org FAQ.
>
If you can send me some suggested verbiage I'll put it in the FAQ. We
can also make a page that's directly linked from the "stable release",
kind of like we have info links for -mm patches etc.
-hpa
>> I agree. I think that one previous -stable patch version should always
>> be listed there, even if we think that 2.6.N is stable. :)
>
>I don't at all. If we do that, people will assume that they need to
>wait till 2.6.N.1 before trying that kernel (as it wouldn't be "stable"
>otherwise.) So no one will test it, to really generate the bug reports
>that we need to get to that .1 release.
>
>Or should we just throw out a .1 release with the first simple patch
>that comes along just to make the kernel.org page update properly? I
>don't think so...
>
Or call the "2.6.X" as "2.6.X.0", so they got at least a clue that this is
becoming .1
Jan Engelhardt
--
On Thursday 05 January 2006 00:08, H. Peter Anvin wrote:
> Alistair John Strachan wrote:
> > Re-read the thread. The confusion here is about "going back" to 2.6.14
> > before patching 2.6.15. This has nothing to do with "rc kernels". We have
> > this documented explicitly in the kernel but not on the kernel.org FAQ.
>
> If you can send me some suggested verbiage I'll put it in the FAQ. We
> can also make a page that's directly linked from the "stable release",
> kind of like we have info links for -mm patches etc.
I hope somebody else here can minimise my logic; I think the verbosity is
necessary to completely explain the "patch nightmare" to everybody concerned.
I'm getting warnings about "Reversed (or previously applied) patch detected!"
when I attempt to patch to an -rc, -mm or -stable kernel. How should this be
done?
*Terms*
A Linus mainline "release" kernel is given an x.y.z version number (e.g.
2.6.14). A -stable kernel is given an x.y.z.a version number (e.g.,
2.6.14.3).
*Patching*
New -stable or -rc patches are to be applied to Linus's mainline "release"
kernels, not -stable, -rc or -mm kernel versions.
New -mm patches are to be applied either to the "release" kernel, or the
applicable -rc kernel (whichever is newer), which are given an x.y.z-rcN
version number (e.g. 2.6.15-rc2).
- If you are on a -stable kernel, you need to revert the -stable patch before
applying either an -rc, a new -stable, or new Linus "release" patch.
- If you are on an -rc kernel, you need to revert the -rc patch if you want to
apply another -rc patch, a -stable patch, or a new Linus "release" patch.
- If you are on an -mm kernel, you need to revert the -mm patch and the -rc
patch (if applicable), before applying either an -rc, -stable or new Linus
"release" patch.
*Patch Process*
For example, to patch to 2.6.14-rc3-mm3, when you were previously on 2.6.13.4,
you would need to:
- Download the full Linux 2.6.13 "release" tarball, or reverse the 2.6.13.4
-stable patch.
tar jxvf linux-2.6.13.tar.bz2 OR
bzip2 -cd ../patch-2.6.13.4.bz2 | patch -Rp1 (whilst in linux directory).
- Apply the 2.6.14-rc3 patch.
bzip2 -cd ../patch-2.6.14-rc3.bz2 | patch -Np1 (whilst in linux directory).
- Apply the 2.6.14-rc3-mm3 patch.
bzip2 -cd ../2.6.14-rc3-mm3.bz2 | patch -Np1 (whilst in linux directory).
Most users will be confused by running a -stable kernel, and not being able to
patch to the latest Linus "release" kernel. If you are on 2.6.13.4, and you
want to go to 2.6.14, all you need to do is:
- Download the full Linux 2.6.14 "release" tarball. Done.
tar jxvf linux-2.6.14.tar.bz2
- OR, Reverse the 2.6.13.4 patch from your existing tree. This will return you
to Linus "release" 2.6.13.
bzip2 -cd ../patch-2.6.13.4.bz2 | patch -Rp1 (whilst in linux directory).
- Apply the 2.6.14 patch.
bzip2 -cd ../patch-2.6.14.bz2 | patch -Np1 (whilst in linux directory).
Alternatively, if the above is too much for you, check out Matt Mackall's
ketchup utility. This will automate the above process.
http://www.selenic.com/ketchup/
.. Okay, after writing that maybe Greg was correct. This will easily become
the largest FAQ entry, even if it is a complete description of the issue at
hand.
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
On 1/5/06, Alistair John Strachan <[email protected]> wrote:
> On Thursday 05 January 2006 00:08, H. Peter Anvin wrote:
> > Alistair John Strachan wrote:
> > > Re-read the thread. The confusion here is about "going back" to 2.6.14
> > > before patching 2.6.15. This has nothing to do with "rc kernels". We have
> > > this documented explicitly in the kernel but not on the kernel.org FAQ.
> >
> > If you can send me some suggested verbiage I'll put it in the FAQ. We
> > can also make a page that's directly linked from the "stable release",
> > kind of like we have info links for -mm patches etc.
>
> I hope somebody else here can minimise my logic; I think the verbosity is
> necessary to completely explain the "patch nightmare" to everybody concerned.
>
[snip]
Nice writeup, but why not simply put a copy of
Documentation/applying-patches.txt online and link to that?
It contains more or less the same stuff you just wrote.
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Thursday 05 January 2006 15:37, Jesper Juhl wrote:
> On 1/5/06, Alistair John Strachan <[email protected]> wrote:
> > On Thursday 05 January 2006 00:08, H. Peter Anvin wrote:
> > > Alistair John Strachan wrote:
> > > > Re-read the thread. The confusion here is about "going back" to
> > > > 2.6.14 before patching 2.6.15. This has nothing to do with "rc
> > > > kernels". We have this documented explicitly in the kernel but not on
> > > > the kernel.org FAQ.
> > >
> > > If you can send me some suggested verbiage I'll put it in the FAQ. We
> > > can also make a page that's directly linked from the "stable release",
> > > kind of like we have info links for -mm patches etc.
> >
> > I hope somebody else here can minimise my logic; I think the verbosity is
> > necessary to completely explain the "patch nightmare" to everybody
> > concerned.
>
> [snip]
>
> Nice writeup, but why not simply put a copy of
> Documentation/applying-patches.txt online and link to that?
> It contains more or less the same stuff you just wrote.
It's certainly one possibility, but this file is at least 4x more verbose than
my summary. I aimed to write something instructional, rather than provide a
complete explanation of the "patch problem".
If this issue is large enough to get its own page on kernel.org, then a more
complete description may indeed be justified. Comments?
--
Cheers,
Alistair.
'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.
Alistair John Strachan wrote:
>>
>>Nice writeup, but why not simply put a copy of
>>Documentation/applying-patches.txt online and link to that?
>>It contains more or less the same stuff you just wrote.
>
> It's certainly one possibility, but this file is at least 4x more verbose than
> my summary. I aimed to write something instructional, rather than provide a
> complete explanation of the "patch problem".
>
> If this issue is large enough to get its own page on kernel.org, then a more
> complete description may indeed be justified. Comments?
>
It'd be easy to make its own page, and then link it from the FAQ. Heck,
the easiest might be to link it to the gitweb checkout of
Documentation/applying-patches.txt, but having it be nicely HTML
formatted would probably be better.
-hpa
I don't know how much of a pain this would be, but maybe you could
arrange the links on the Kernel.org front page to visually cue people
that the 2.6.15 Linus release is a base for all of the patches. Say,
put the 2.6.15 in double-size text and put the other branches
underneath in smaller text and leaders, like a family tree.