Re: Linux 2.4.35 tree abandoned/delayed?

Re: Linux 2.4.35 tree abandoned/delayed?

[Tony Borras]

On Tue, 17 Jul 2007 22:50:10 +0200
Willy Tarreau <w@1wt.eu> wrote:

> Hi Tony,
> 
> [ first, please always keep people in CC on LKML since it's
> very easy to
>   miss a thread ]
> 
> On Mon, Jul 16, 2007 at 11:23:57PM -0800, Tony Borras wrote:
> > On Tue, 17 Jul 2007 07:20:06 +0200
> > Willy Tarreau <w@1wt.eu> wrote:
> > 
> > > could you please report what the problem was, in case it's
> > > not fixed yet ?
> > 
> > Ok, reproduced error at ./net/ipv4/fib_semantics.c
> >   line 436 changed label:
> >                    errout -> err_inval
> 
> I don't know where you downloaded 2.4.35-pre5 from, but I've
> checked the sources and could not find any reference to
> "errout" here in any version!
> 
> Or what is possible is that your 2.4.34 kernel got corrupted
> on your disk after applying any patch. Sometimes, some shell
> scripts providede with some

Negatory on that, Willy!

Here is what I have verified:
  At kernel.org:/pub/linux/v2.4

file: patch-2.4.34.3.bz2 has the corruption at line 213:
  goto errout

That is where that came from, when I patched my 2.4.34 to arrive
at 2.4.34.5, before applying the 2.4.35.rc5 patch!

Hope that helps, maybe you can fix, because I cant post to
kernel.org v2.4 tree, that I know of.

TonyB


There are 10 types of people in this world, those that read
binary and those who don't!

--
  __      __  _     I N C.               http://www.sysdev.org
/ __|\\// __||  \  __   __          /         tonyb@sysdev.org
\__ \ \/\__ \||)|/ O_)\/ /        \/  System Tools / Utilities
|___/ || ___/|_ /\___|\_/        WIntel / Linux Device Drivers

Re: Linux 2.4.35 tree abandoned/delayed?

[Jesper Juhl]

On 19/07/07, Tony Borras <tonyb@sysdev.org> wrote:
> On Tue, 17 Jul 2007 22:50:10 +0200
> Willy Tarreau <w@1wt.eu> wrote:
>
> > Hi Tony,
> >
> > [ first, please always keep people in CC on LKML since it's
> > very easy to
> >   miss a thread ]
> >
> > On Mon, Jul 16, 2007 at 11:23:57PM -0800, Tony Borras wrote:
> > > On Tue, 17 Jul 2007 07:20:06 +0200
> > > Willy Tarreau <w@1wt.eu> wrote:
> > >
> > > > could you please report what the problem was, in case it's
> > > > not fixed yet ?
> > >
> > > Ok, reproduced error at ./net/ipv4/fib_semantics.c
> > >   line 436 changed label:
> > >                    errout -> err_inval
> >
> > I don't know where you downloaded 2.4.35-pre5 from, but I've
> > checked the sources and could not find any reference to
> > "errout" here in any version!
> >
> > Or what is possible is that your 2.4.34 kernel got corrupted
> > on your disk after applying any patch. Sometimes, some shell
> > scripts providede with some
>
> Negatory on that, Willy!
>
> Here is what I have verified:
>   At kernel.org:/pub/linux/v2.4
>
> file: patch-2.4.34.3.bz2 has the corruption at line 213:
>   goto errout
>
> That is where that came from, when I patched my 2.4.34 to arrive
> at 2.4.34.5, before applying the 2.4.35.rc5 patch!
>
That sounds wrong. To the best of my knowledge, the 2.4.35-rc5 patch
should be applied on top of 2.4.34, not on top of 2.4.34.5 ...

-- 
Jesper Juhl <jesper.juhl@gmail.com>
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

Re: Linux 2.4.35 tree abandoned/delayed?

[Tony Borras]

On Thu, 19 Jul 2007 02:10:58 +0200
"Jesper Juhl" <jesper.juhl@gmail.com> wrote:

> > file: patch-2.4.34.3.bz2 has the corruption at line 213:
> >   goto errout
> >
> > That is where that came from, when I patched my 2.4.34 to
> > arrive at 2.4.34.5, before applying the 2.4.35.rc5 patch!
> >
> That sounds wrong. To the best of my knowledge, the 2.4.35-rc5
> patch should be applied on top of 2.4.34, not on top of
> 2.4.34.5 ...
> 
> -- 
> Jesper Juhl <jesper.juhl@gmail.com>
> 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

You are absolutely right!  Since I was keeping a 2.4.34.current
and 2.4.35.pre5 is just a pre, I stupidly slapped it on top at
my current build tree!

too much Summer partying!

Thanks a lot!
TonyB

There are 10 types of people in this world, those that read
binary and those who don't!

--
  __      __  _     I N C.               http://www.sysdev.org
/ __|\\// __||  \  __   __          /         tonyb@sysdev.org
\__ \ \/\__ \||)|/ O_)\/ /        \/  System Tools / Utilities
|___/ || ___/|_ /\___|\_/        WIntel / Linux Device Drivers

Re: Linux 2.4.35 tree abandoned/delayed?

[Willy Tarreau]

On Wed, Jul 18, 2007 at 04:03:09PM -0800, Tony Borras wrote:
> On Tue, 17 Jul 2007 22:50:10 +0200
> Willy Tarreau <w@1wt.eu> wrote:
> 
> > Hi Tony,
> > 
> > [ first, please always keep people in CC on LKML since it's
> > very easy to
> >   miss a thread ]
> > 
> > On Mon, Jul 16, 2007 at 11:23:57PM -0800, Tony Borras wrote:
> > > On Tue, 17 Jul 2007 07:20:06 +0200
> > > Willy Tarreau <w@1wt.eu> wrote:
> > > 
> > > > could you please report what the problem was, in case it's
> > > > not fixed yet ?
> > > 
> > > Ok, reproduced error at ./net/ipv4/fib_semantics.c
> > >   line 436 changed label:
> > >                    errout -> err_inval
> > 
> > I don't know where you downloaded 2.4.35-pre5 from, but I've
> > checked the sources and could not find any reference to
> > "errout" here in any version!
> > 
> > Or what is possible is that your 2.4.34 kernel got corrupted
> > on your disk after applying any patch. Sometimes, some shell
> > scripts providede with some
> 
> Negatory on that, Willy!
> 
> Here is what I have verified:
>   At kernel.org:/pub/linux/v2.4
> 
> file: patch-2.4.34.3.bz2 has the corruption at line 213:
>   goto errout
> 
> That is where that came from, when I patched my 2.4.34 to arrive
> at 2.4.34.5,

So your 2.4.34.5 tree is wrong, you forgot to apply 2.4.34.4 to it,
which only fixes this goto. You should rebuild your 2.4.34.X tree
from scratch.

> before applying the 2.4.35.rc5 patch!

No, once again, you should not apply any 2.4.35-preX to 2.4.34.X.
You should apply it to plain 2.4.34.

If you don't have it anymore, I suggest that you revert the 2.4.34.X
patches with "patch -R".

Also, I think you got a lot of rejects when applying 2.4.35 on top of
2.4.34.X since most changes are the sames.

Regards,
Willy

Re: Linux 2.4.35 tree abandoned/delayed?

[Tony Borras]

On Tue, 17 Jul 2007 07:20:06 +0200
Willy Tarreau <w@1wt.eu> wrote:

> Hi Tony,
> 
> On Mon, Jul 16, 2007 at 04:30:47PM -0800, Tony Borras wrote:
> > I did get the current 'pre' set of 2.4.35 (circa early
> > June) patches and applied to 2.4.34.5.  After fixing a goto
> > mislable in one of the patches (dont remember if it was one
> > of the patch-2.4.34.? or patch-2.4.35.pre5), managed to
> > build a vmlinuz-2.4.35 successfully.
> 
> could you please report what the problem was, in case it's not
> fixed yet ?
> 
> > Just wondered whether to expect an abandonded tree, in which
> > case I will fly with this 2.4.35 build or should wait for
> > its eventual release.
>  
> No it's not abandonned at all !!!

My process was to start with 2.4.34 sources (that compiled
fine), applied the 2.4.34.[1-5] patches and then the 2.4.35.rc5.

I didnt patch the patches. I changed a jmp reference to a
non-existent label to the correct label, in the one file that
complained at build time.

So have to repeat the full process, again.  Will let you know...

TonyB


There are 10 types of people in this world, those that read
binary and those who don't!

--
  __      __  _     I N C.               http://www.sysdev.org
/ __|\\// __||  \  __   __          /         tonyb@sysdev.org
\__ \ \/\__ \||)|/ O_)\/ /        \/  System Tools / Utilities
|___/ || ___/|_ /\___|\_/        WIntel / Linux Device Drivers

Re: Linux 2.4.35 tree abandoned/delayed?

[Tony Borras]

On Tue, 17 Jul 2007 07:20:06 +0200
Willy Tarreau <w@1wt.eu> wrote:

> could you please report what the problem was, in case it's not
> fixed yet ?

Ok, reproduced error at ./net/ipv4/fib_semantics.c
  line 436 changed label:
                   errout -> err_inval

then it builds fine!

started with linux-2.4.34
then patched
     patch-2.4.34.1
     patch-2.4.34.2
     patch-2.4.34.3
     patch-2.4.34.4
     patch-2.4.34.5
     patch-2.4.35.pre5
I used a make oldconfig on a .config that builds in 2.4.34

Hope that helps
TonyB


There are 10 types of people in this world, those that read
binary and those who don't!

--
  __      __  _     I N C.               http://www.sysdev.org
/ __|\\// __||  \  __   __          /         tonyb@sysdev.org
\__ \ \/\__ \||)|/ O_)\/ /        \/  System Tools / Utilities
|___/ || ___/|_ /\___|\_/        WIntel / Linux Device Drivers

Re: Linux 2.4.35 tree abandoned/delayed?

[Willy Tarreau]

Hi Tony,

[ first, please always keep people in CC on LKML since it's very easy to
  miss a thread ]

On Mon, Jul 16, 2007 at 11:23:57PM -0800, Tony Borras wrote:
> On Tue, 17 Jul 2007 07:20:06 +0200
> Willy Tarreau <w@1wt.eu> wrote:
> 
> > could you please report what the problem was, in case it's not
> > fixed yet ?
> 
> Ok, reproduced error at ./net/ipv4/fib_semantics.c
>   line 436 changed label:
>                    errout -> err_inval

I don't know where you downloaded 2.4.35-pre5 from, but I've checked the
sources and could not find any reference to "errout" here in any version!

Or what is possible is that your 2.4.34 kernel got corrupted on your disk
after applying any patch. Sometimes, some shell scripts providede with some
drivers directly tamper with the sources. You should check that your tree
is valid by downloading another one and comparing them.

It had already happened to me in the past and I got 3 or 4 private kernels
with a fix that I did not know I had applied. Fortunately, this does not
happen anymore with git ;-)

> then it builds fine!
> 
> started with linux-2.4.34
> then patched
>      patch-2.4.34.1
>      patch-2.4.34.2
>      patch-2.4.34.3
>      patch-2.4.34.4
>      patch-2.4.34.5
>      patch-2.4.35.pre5

Huh !!! You should not apply all 2.4.34 patches before 2.4.35. You should
see the development process like this :

---> 2.4.34 ----> 2.4.35-pre1 ----> 2.4.35-pre2 ---> ... ---> 2.4.35
       \
        `--> 2.4.34.1 --> 2.4.34.2 --> 2.4.34.3 ...

So 2.4.35-preX and 2.4.34.X are two different branches forked from 2.4.34.
All 2.4.35-preX patches are made for 2.4.34, and all 2.4.34.X patches are
made for 2.4.34 too.

You should see 2.4.34 as the real official kernel, and other ones as
convenience versions released in order to help people keep up to date
without having to use possibly buggy pre-releases, even if they are
quite close.

> I used a make oldconfig on a .config that builds in 2.4.34
> 
> Hope that helps
> TonyB

Thanks anyway for your feedback !

Regards,
Willy

Linux 2.4.35 tree abandoned/delayed?

[Tony Borras]

I did get the current 'pre' set of 2.4.35 (circa early
June) patches and applied to 2.4.34.5.  After fixing a goto
mislable in one of the patches (dont remember if it was one of
the patch-2.4.34.? or patch-2.4.35.pre5), managed to build a
vmlinuz-2.4.35 successfully.

Just wondered whether to expect an abandonded tree, in which
case I will fly with this 2.4.35 build or should wait for its
eventual release.
 
Thanks Greg!
 
Tony Borras


There are 10 types of people in this world, those that read
binary and those who don't!

--
  __      __  _     I N C.               http://www.sysdev.org
/ __|\\// __||  \  __   __          /         tonyb@sysdev.org
\__ \ \/\__ \||)|/ O_)\/ /        \/  System Tools / Utilities
|___/ || ___/|_ /\___|\_/        WIntel / Linux Device Drivers

Re: Linux 2.4.35 tree abandoned/delayed?

[Willy Tarreau]

Hi Tony,

On Mon, Jul 16, 2007 at 04:30:47PM -0800, Tony Borras wrote:
> I did get the current 'pre' set of 2.4.35 (circa early
> June) patches and applied to 2.4.34.5.  After fixing a goto
> mislable in one of the patches (dont remember if it was one of
> the patch-2.4.34.? or patch-2.4.35.pre5), managed to build a
> vmlinuz-2.4.35 successfully.

could you please report what the problem was, in case it's not fixed
yet ?

> Just wondered whether to expect an abandonded tree, in which
> case I will fly with this 2.4.35 build or should wait for its
> eventual release.
 
No it's not abandonned at all !!!

The difficulty with 2.4 is to get user feedback on patches. While in 2.6,
there are hundreds or thousands of testers for every release, in 2.4,
I have to wait longer after every release in order to start collecting
problem reports, or confirmation of fixes. People using it in production
generally cannot reboot a machine in the evening just to try a patch.
Also, subsystem maintainers have less time to spend on it and are
themselves slowed down by the slow feedback process.

That's merely why I maintain a parallel 2.4.34.X stable branch. I thought
I could release one version every 6 months, but I think that this delay
will probably slightly increase after 2.4.35, we'll see.

I still have a few patches to merge, and I expect to be able to work on
them next week-end, then release -rc1, which I expect to be the only -rc
before -final (which will probably not be true as in every -rc).

Thanks,
Willy