re:Build regressions/improvements in v3.0-rc3

re:Build regressions/improvements in v3.0-rc3

[Wanlong Gao]

<snip>
I see that some mips fixes in ralf's tree but not been pulled. almost
for about one month.

Thanks

Re: Build regressions/improvements in v3.0-rc3

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 436 bytes --]

On Wed, 15 Jun 2011 19:44:41 +0800 Wanlong Gao <wanlong.gao@gmail.com> wrote:
>
> <snip>
> I see that some mips fixes in ralf's tree but not been pulled. almost
> for about one month.

Well, the mips tree in linux-next is
git://git.linux-mips.org/pub/scm/upstream-sfr.git#mips-for-linux-next and
I fetch it every day.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: Build regressions/improvements in v3.0-rc3

[Ralf Baechle]

On Wed, Jun 15, 2011 at 10:29:32PM +1000, Stephen Rothwell wrote:

> On Wed, 15 Jun 2011 19:44:41 +0800 Wanlong Gao <wanlong.gao@gmail.com> wrote:
> >
> > <snip>
> > I see that some mips fixes in ralf's tree but not been pulled. almost
> > for about one month.
> 
> Well, the mips tree in linux-next is
> git://git.linux-mips.org/pub/scm/upstream-sfr.git#mips-for-linux-next and
> I fetch it every day.

I only have the new bits in linux-next but not the pending for the current
work-in-progress kernel, that is 3.0.  Those are queued in

http://git.linux-mips.org/?p=upstream-linus.git;a=summary
git://git.linux-mips.org/pub/scm/upstream-linus.git

and I've sent two pull requests to Linus so far which went without
reaction ...

  Ralf

Re: Build regressions/improvements in v3.0-rc3

[Wanlong Gao]

Maybe linus had missed them, sadly.

Ralf Baechle <ralf@linux-mips.org> wrote:

>On Wed, Jun 15, 2011 at 10:29:32PM +1000, Stephen Rothwell wrote:
>
>> On Wed, 15 Jun 2011 19:44:41 +0800 Wanlong Gao <wanlong.gao@gmail.com> wrote:
>> >
>> > <snip>
>> > I see that some mips fixes in ralf's tree but not been pulled. almost
>> > for about one month.
>> 
>> Well, the mips tree in linux-next is
>> git://git.linux-mips.org/pub/scm/upstream-sfr.git#mips-for-linux-next and
>> I fetch it every day.
>
>I only have the new bits in linux-next but not the pending for the current
>work-in-progress kernel, that is 3.0.  Those are queued in
>
>http://git.linux-mips.org/?p=upstream-linus.git;a=summary
>git://git.linux-mips.org/pub/scm/upstream-linus.git
>
>and I've sent two pull requests to Linus so far which went without
>reaction ...
>
>  Ralf

Re: Build regressions/improvements in v3.0-rc3

[Geert Uytterhoeven]

On Wed, Jun 15, 2011 at 11:11, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> Below is the list of build error/warning regressions/improvements in
> v3.0-rc3[1] compared to v2.6.39[2].
>
> Summarized:
>  - build errors: +21/-66
>  - build warnings: +202/-421
>
> JFYI, when comparing v3.0-rc3 to v3.0-rc2[3], the summaries are:
>  - build errors: +20/-45
>  - build warnings: +92/-101

While looking what kind of fix got applied for one of the
improvements, I noticed it
didn't get fixed at all. But m68k/allmodconfig is broken due to an
unrelated change, after
which the build stops, so it never got to compile the file I was looking into.

Would it be possible to build with `make -k', so it continues in case
of an error, and
tries to build as much as possible?
Yes, I know this may increase build time considerably, but in an ideal
world, there
are no compile errors and everything is built anyway ;-)

Thanks!

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: Build regressions/improvements in v3.0-rc3

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 625 bytes --]

Hi Geert,

On Wed, 15 Jun 2011 11:19:48 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Would it be possible to build with `make -k', so it continues in case
> of an error, and
> tries to build as much as possible?
> Yes, I know this may increase build time considerably, but in an ideal
> world, there
> are no compile errors and everything is built anyway ;-)

Done (I think - it required very simple changes to a Python script and I
only do "monkey see, monkey do" Python programming :-))

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: Build regressions/improvements in v3.0-rc3

[Geert Uytterhoeven]

Hi Stephen,

On Sat, Jun 18, 2011 at 10:30, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> On Wed, 15 Jun 2011 11:19:48 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>> Would it be possible to build with `make -k', so it continues in case
>> of an error, and
>> tries to build as much as possible?
>> Yes, I know this may increase build time considerably, but in an ideal
>> world, there
>> are no compile errors and everything is built anyway ;-)
>
> Done (I think - it required very simple changes to a Python script and I

Thanks!

> only do "monkey see, monkey do" Python programming :-))

Good to know I'm not the only one ;-)

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: Build regressions/improvements in v3.0-rc3

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1004 bytes --]

Hi Geert,

On Sat, 18 Jun 2011 18:30:29 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> On Wed, 15 Jun 2011 11:19:48 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> >
> > Would it be possible to build with `make -k', so it continues in case
> > of an error, and
> > tries to build as much as possible?
> > Yes, I know this may increase build time considerably, but in an ideal
> > world, there
> > are no compile errors and everything is built anyway ;-)
> 
> Done (I think - it required very simple changes to a Python script and I
> only do "monkey see, monkey do" Python programming :-))

I also made the mistake of fixing things so that the automatic "git
bisect for build failures" would work.  That was taking much too much
time (over 4 hours for one build), so I have turned it off.  Hopefully,
the builds will now run in a more reasonable time frame.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: Build regressions/improvements in v3.0-rc3

[Geert Uytterhoeven]

Hi Stephen,

On Thu, Jun 23, 2011 at 13:27, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> On Sat, 18 Jun 2011 18:30:29 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>> On Wed, 15 Jun 2011 11:19:48 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>> > Would it be possible to build with `make -k', so it continues in case
>> > of an error, and
>> > tries to build as much as possible?
>> > Yes, I know this may increase build time considerably, but in an ideal
>> > world, there
>> > are no compile errors and everything is built anyway ;-)
>>
>> Done (I think - it required very simple changes to a Python script and I
>> only do "monkey see, monkey do" Python programming :-))
>
> I also made the mistake of fixing things so that the automatic "git
> bisect for build failures" would work.  That was taking much too much
> time (over 4 hours for one build), so I have turned it off.  Hopefully,
> the builds will now run in a more reasonable time frame.

Yeah, I noticed it. There are still no results for 3.0-rc4...

Probably you should only bisect for configs that where "green" recently.
And hope people fix reported regressions the-next-day-as-we-speak :-)

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: Build regressions/improvements in v3.0-rc3

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 842 bytes --]

Hi Geert,

On Fri, 24 Jun 2011 08:08:06 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> > I also made the mistake of fixing things so that the automatic "git
> > bisect for build failures" would work.  That was taking much too much
> > time (over 4 hours for one build), so I have turned it off.  Hopefully,
> > the builds will now run in a more reasonable time frame.
> 
> Yeah, I noticed it. There are still no results for 3.0-rc4...

3.0-rc4 should be complete in a few hours.  Hopefully rc5 will be faster.

> Probably you should only bisect for configs that where "green" recently.
> And hope people fix reported regressions the-next-day-as-we-speak :-)

Well, I will leave that for the author :-)

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: Build regressions/improvements in v3.0-rc3

[Geert Uytterhoeven]

Hi Stephen,

On Fri, Jun 24, 2011 at 08:10, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> 3.0-rc4 should be complete in a few hours.  Hopefully rc5 will be faster.

Completed at 101% 108/107! :-)
But there are only 107 logs on the page.

Anyway, the "-k" seems to work: much more problems in the logs. Thanks!

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds