(no subject)

[Keith Owens]

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

Content-Type: text/plain; charset=us-ascii

Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
release 2.4.

kbuild-2.5-core-13-1.
  Changes from core-9.

    Update documentation for asm-offsets.h.

    Remove the requirement that arch/$(ARCH)/Makefile.defs.*config had
    to be in the base tree.  Now a new architecture can be a completely
    separate drop in tree.

    Really force CONFIG_MODVERSIONS=n.  Modversions will not be
    supported until after kbuild 2.5 is in the main kernel.

    Change phase 4 message, it does a lot of integrity checks as well
    as generating the global makefile.  Will that shut up people who
    think that we don't need integrity checks?  Probably not :(.

    Correct bug in symlink message.

kbuild-2.5-common-2.5.13-1.
  Changes from common-2.5.12-1.

    Upgrade to kernel 2.5.13.
   
kbuild-2.5-i386-2.5.13-1.
  Changes from i386-2.5.12-1.

    Upgrade to kernel 2.5.13.

    Correct CC/CC_real in arch/$(ARCH)/Makefile.defs.*config.


You should be able to use kbuild-2.5-sparc64-2.5.12-1 with this
release.  ia64 is probably out of date by now and is not recommended
for use with 2.5.13.  I will do the kbuild 2.5 update for ia64 after we
get a new ia64 kernel patch.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: Exmh version 2.1.1 10/15/1999

iD8DBQE80pxci4UHNye0ZOoRAvYBAJ9eaE5JxM3vH3pmzI8ho6q3VsrRlACbBspY
u7SNlyBjmhgTD4YnSxw+Jas=
=Ek9t
-----END PGP SIGNATURE-----

Re: your mail

[tomas szepe]

> Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
> http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
> release 2.4.
>
> kbuild-2.5-core-13-1.

I believe you meant 's/13/10/'.

> kbuild-2.5-common-2.5.13-1.
> kbuild-2.5-i386-2.5.13-1.

hmmm.. doesn't look so good.

kala@nibbler:~$ tar xzf /usr/src/linux-2.5.13.tgz 
kala@nibbler:~$ cd linux-2.5.13 
kala@nibbler:~/linux-2.5.13$ zcat /usr/src/kbuild-2.5-core-10.gz /usr/src/kbuild-2.5-common-2.5.13-1.gz /usr/src/kbuild-2.5-i386-2.5.13-1.gz |patch -sp1
kala@nibbler:~/linux-2.5.13$ cp /lib/modules/2.5.13/.config .
kala@nibbler:~/linux-2.5.13$ make -f Makefile-2.5 oldconfig
Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
/home/kala/linux-2.5.13/scripts/Makefile-2.5:473: /no_such_file-arch/i386/Makefile.defs.config: No such file or directory
Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
/home/kala/linux-2.5.13/scripts/Makefile-2.5:473: /no_such_file-arch/i386/Makefile.defs.config: No such file or directory
Using ARCH='i386' AS='as' LD='ld' CC='/usr/bin/gcc' CPP='/usr/bin/gcc -E' AR='ar' HOSTAS='as' HOSTLD='gcc' HOSTCC='gcc' HOSTAR='ar'
Generating global Makefile
  phase 1 (find all inputs)
...

kala@nibbler:~/linux-2.5.13$ make -f Makefile-2.5 installable
Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
spec value %p not found
/home/kala/linux-2.5.13/scripts/Makefile-2.5:473: /no_such_file-arch/i386/Makefile.defs.config: No such file or directory
Using ARCH='i386' AS='as' LD='ld' CC='/usr/bin/gcc' CPP='/usr/bin/gcc -E' AR='ar' HOSTAS='as' HOSTLD='gcc' HOSTCC='gcc' HOSTAR='ar'
Generating global Makefile
  phase 1 (find all inputs)
  phase 2 (convert all Makefile.in files)
  phase 3 (evaluate selections)
  phase 4 (integrity checks, write global makefile)
pp_makefile4: arch/i386/lib/lib.a is selected but is not part of vmlinux, missing link_subdirs?
make: *** [phase4] Error 1


T.

Re: your mail

[tomas szepe]

Building as follows works, though.

$ cd /usr/src && tar xzf linux-2.5.13.tar.gz
$ cd ~ && mkdir build && cd build
$ export KBUILD_OBJTREE=$PWD
$ export KBUILD_SRCTREE_000=/usr/src/linux-2.5.13
$ alias M="make -f $KBUILD_SRCTREE_000/Makefile-2.5"
$ cp /lib/modules/2.5.13/.config .
$ M oldconfig
$ M installable

T.


> > Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
> > http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
> > release 2.4.
> >
> > kbuild-2.5-core-13-1.
> 
> I believe you meant 's/13/10/'.
> 
> > kbuild-2.5-common-2.5.13-1.
> > kbuild-2.5-i386-2.5.13-1.
> 
> hmmm.. doesn't look so good.
> 
> kala@nibbler:~$ tar xzf /usr/src/linux-2.5.13.tgz 
> kala@nibbler:~$ cd linux-2.5.13 
> kala@nibbler:~/linux-2.5.13$ zcat /usr/src/kbuild-2.5-core-10.gz /usr/src/kbuild-2.5-common-2.5.13-1.gz /usr/src/kbuild-2.5-i386-2.5.13-1.gz |patch -sp1
> kala@nibbler:~/linux-2.5.13$ cp /lib/modules/2.5.13/.config .
> kala@nibbler:~/linux-2.5.13$ make -f Makefile-2.5 oldconfig
> Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
> /home/kala/linux-2.5.13/scripts/Makefile-2.5:473: /no_such_file-arch/i386/Makefile.defs.config: No such file or directory
> Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
> /home/kala/linux-2.5.13/scripts/Makefile-2.5:473: /no_such_file-arch/i386/Makefile.defs.config: No such file or directory
> Using ARCH='i386' AS='as' LD='ld' CC='/usr/bin/gcc' CPP='/usr/bin/gcc -E' AR='ar' HOSTAS='as' HOSTLD='gcc' HOSTCC='gcc' HOSTAR='ar'
> Generating global Makefile
>   phase 1 (find all inputs)
> ...
> 
> kala@nibbler:~/linux-2.5.13$ make -f Makefile-2.5 installable
> Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
> spec value %p not found
> /home/kala/linux-2.5.13/scripts/Makefile-2.5:473: /no_such_file-arch/i386/Makefile.defs.config: No such file or directory
> Using ARCH='i386' AS='as' LD='ld' CC='/usr/bin/gcc' CPP='/usr/bin/gcc -E' AR='ar' HOSTAS='as' HOSTLD='gcc' HOSTCC='gcc' HOSTAR='ar'
> Generating global Makefile
>   phase 1 (find all inputs)
>   phase 2 (convert all Makefile.in files)
>   phase 3 (evaluate selections)
>   phase 4 (integrity checks, write global makefile)
> pp_makefile4: arch/i386/lib/lib.a is selected but is not part of vmlinux, missing link_subdirs?
> make: *** [phase4] Error 1
> 
> 
> T.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: your mail

[Keith Owens]

On Fri, 3 May 2002 16:37:38 +0200, 
tomas szepe <kala@pinerecords.com> wrote:
>kala@nibbler:~$ tar xzf /usr/src/linux-2.5.13.tgz 
>kala@nibbler:~$ cd linux-2.5.13 
>kala@nibbler:~/linux-2.5.13$ zcat /usr/src/kbuild-2.5-core-10.gz /usr/src/kbuild-2.5-common-2.5.13-1.gz /usr/src/kbuild-2.5-i386-2.5.13-1.gz |patch -sp1
>kala@nibbler:~/linux-2.5.13$ cp /lib/modules/2.5.13/.config .
>kala@nibbler:~/linux-2.5.13$ make -f Makefile-2.5 oldconfig
>Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory

The trailing '/' is omitted in one case.  Workaround for common source and object

export KBUILD_SRCTREE_000=`pwd`/
make -f Makefile-2.5 oldconfig

Re: your mail

[tomas szepe]

> On Fri, 3 May 2002 16:37:38 +0200, 
> tomas szepe <kala@pinerecords.com> wrote:
>
> >kala@nibbler:~$ tar xzf /usr/src/linux-2.5.13.tgz 
> >kala@nibbler:~$ cd linux-2.5.13 
> >kala@nibbler:~/linux-2.5.13$ zcat /usr/src/kbuild-2.5-core-10.gz /usr/src/kbuild-2.5-common-2.5.13-1.gz /usr/src/kbuild-2.5-i386-2.5.13-1.gz |patch -sp1
> >kala@nibbler:~/linux-2.5.13$ cp /lib/modules/2.5.13/.config .
> >kala@nibbler:~/linux-2.5.13$ make -f Makefile-2.5 oldconfig
> >Makefile-2.5:251: /no_such_file-arch/i386/Makefile.defs.noconfig: No such file or directory
> 
> The trailing '/' is omitted in one case.  Workaround for common source and object
> 
> export KBUILD_SRCTREE_000=`pwd`/
> make -f Makefile-2.5 oldconfig

Another problem/question:

$ cd build
$ export KBUILD_OBJTREE=$PWD
$ export KBUILD_SRCTREE_000=/usr/src/linux-2.5.13
$ alias M="make -f $KBUILD_SRCTREE_000/Makefile-2.5"
$ cp /lib/modules/2.5.13/.config .
$ M oldconfig
...

$ M installable
...

[so far so good]

$ make -f Makefile-2.5 menuconfig
[enable RAMDISK support, tweak ramdisk size, enable initrd]
...

Now, issuing "M installable" will result in nearly all files getting rebuilt.
The same happens when switching ramdisk off again. How's that?

I tried enabling/disabling many other config options and doing rebuilds but
couldn't find anything as damaging buildtime-wise as the ramdisk stuff.

Hopefully I'm causing no headaches,
T.

Re: kbuild 2.5 release 2.4

[Keith Owens]

On Fri, 3 May 2002 17:45:54 +0200, 
tomas szepe <kala@pinerecords.com> wrote:
>$ make -f Makefile-2.5 menuconfig
>[enable RAMDISK support, tweak ramdisk size, enable initrd]
>...
>
>Now, issuing "M installable" will result in nearly all files getting rebuilt.
>The same happens when switching ramdisk off again. How's that?
>
>I tried enabling/disabling many other config options and doing rebuilds but
>couldn't find anything as damaging buildtime-wise as the ramdisk stuff.

CONFIG_BLK_DEV_INITRD is tested in include/linux/fs.h.  Because of the
messy kernel include files, a config change to fs.h affects hundreds of
other files and forces lots of rebuilds.  This is a separate problem
from kbuild, other people are looking at cleaning up the include files.

Re: kbuild 2.5 release 2.4

[tomas szepe]

> >$ make -f Makefile-2.5 menuconfig
> >[enable RAMDISK support, tweak ramdisk size, enable initrd]
> >...
> >
> >Now, issuing "M installable" will result in nearly all files getting rebuilt.
> >The same happens when switching ramdisk off again. How's that?
> >
> >I tried enabling/disabling many other config options and doing rebuilds but
> >couldn't find anything as damaging buildtime-wise as the ramdisk stuff.
> 
> CONFIG_BLK_DEV_INITRD is tested in include/linux/fs.h.  Because of the
> messy kernel include files, a config change to fs.h affects hundreds of
> other files and forces lots of rebuilds.  This is a separate problem
> from kbuild, other people are looking at cleaning up the include files.


Hold on, there's more :)

/1/

$ cd ~ && rm -rf build && mkdir build && cd build
$ export KBUILD_SRCTREE_000=/usr/src/linux-2.5.13
$ export KBUILD_OBJTREE=$PWD
$ alias M="make -f $KBUILD_SRCTREE_000/Makefile-2.5"
$ touch foo bar baz
$ M mrproper
spec value %p not found
$ ls foo bar baz
/bin/ls: foo: No such file or directory
/bin/ls: bar: No such file or directory
/bin/ls: baz: No such file or directory
$ # oh my! wherez mah precious files gone!

--> *BUUUURN*

/2/

$ cd ~ && rm -rf build && mkdir build && cd build
$ cp /lib/modules/2.5.13/.config .
$ M oldconfig installable
...

--> kbuild repeats launching "oldconfig" till I hit C-c,
never proceeds to "installable".

Of course, issuing "M oldconfig && M installable" works.


T.

Re: kbuild 2.5 release 2.4

[tomas szepe]

> $ cd ~ && rm -rf build && mkdir build && cd build
> $ export KBUILD_SRCTREE_000=/usr/src/linux-2.5.13
> $ export KBUILD_OBJTREE=$PWD
> $ alias M="make -f $KBUILD_SRCTREE_000/Makefile-2.5"
> $ touch foo bar baz
> $ M mrproper
> spec value %p not found
> $ ls foo bar baz
> /bin/ls: foo: No such file or directory
> /bin/ls: bar: No such file or directory
> /bin/ls: baz: No such file or directory
> $ # oh my! wherez mah precious files gone!


Duh... I failed to demonstrate the real problem: kbuild-2.5's mrproper
also erases *non-zero-size* files in the object tree:

$ pwd
/home/kala/build
$ set| grep KBUILD
KBUILD_OBJTREE=/home/kala/build
KBUILD_SRCTREE_000=/usr/src/linux-2.5.13
$ alias M='make -f /usr/src/linux-2.5.13/Makefile-2.5'
$ echo foo >bar
$ ls -l bar
-rw-r--r--    1 kala     users           4 May  4 12:56 bar
$ M mrproper
spec value %p not found
$ ls -la
total 3
drwxr-xr-x    2 kala     users          48 May  4 12:56 ./
drwx--x--x   26 kala     users        1664 May  3 19:58 ../


T.

Re: kbuild 2.5 release 2.4

[Keith Owens]

On Sat, 4 May 2002 12:58:57 +0200, 
tomas szepe <kala@pinerecords.com> wrote:
>Duh... I failed to demonstrate the real problem: kbuild-2.5's mrproper
>also erases *non-zero-size* files in the object tree:

With a separate object tree, everything in the object directory is
generated by kbuild.  make mrproper deletes everything.

Re: kbuild 2.5 release 2.4

[Zwane Mwaikambo]

On Sat, 4 May 2002, Keith Owens wrote:

> On Sat, 4 May 2002 12:58:57 +0200, 
> tomas szepe <kala@pinerecords.com> wrote:
> >Duh... I failed to demonstrate the real problem: kbuild-2.5's mrproper
> >also erases *non-zero-size* files in the object tree:
> 
> With a separate object tree, everything in the object directory is
> generated by kbuild.  make mrproper deletes everything.

Hi Keith,
	Just a minor nit, could you change the rm parameter to have -r 
too? I've found that ^C halfway through a make can leave some old 
directories (.tmp_include or somesuch).

-- 
http://function.linuxpower.ca
		

Re: kbuild 2.5 release 2.4

[tomas szepe]

> >Duh... I failed to demonstrate the real problem: kbuild-2.5's mrproper
> >also erases *non-zero-size* files in the object tree:
> With a separate object tree, everything in the object directory is
> generated by kbuild.  make mrproper deletes everything.

Suppose I have scripts to automate kernel building in that dir?

Or even worse, suppose $KBUILD_OBJTREE gets reset to some weird
value by accident and my whole home dir goes bye bye upon make
mrproper?

Suppose many other scenarios, this "feature" can't stand -- it's inconceivable
that you'd just do "rm -rf $KBUILD_OBJTREE". And I can't see why you would even
try, given you "know" the names of the files generated anyway.

T.

Re: kbuild 2.5 release 2.4

[Zwane Mwaikambo]

On Sat, 4 May 2002, Keith Owens wrote:

> You will have to provide more details of the problem that you think you
> are seeing.  The .tmp_include files are part of the infrastructure that
> lets me separate source and object trees, they are meant to be there.
> make -f Makefile-2.5 mrproper deletes .tmp_include.

I didn't try reproducing it, i just rm -rf'd .tmp_include when it failed. 
I'll double check.

Thanks,
	Zwane Mwaikambo

-- 
http://function.linuxpower.ca
		

Re: kbuild 2.5 release 2.4

[Keith Owens]

On Sat, 4 May 2002 13:39:54 +0200, 
tomas szepe <kala@pinerecords.com> wrote:
>> >Duh... I failed to demonstrate the real problem: kbuild-2.5's mrproper
>> >also erases *non-zero-size* files in the object tree:
>> With a separate object tree, everything in the object directory is
>> generated by kbuild.  make mrproper deletes everything.
>
>Suppose I have scripts to automate kernel building in that dir?

Scripts go in the source tree, not the object tree.  Or better still,
completely outside the kernel source and object tree structure.

>Or even worse, suppose $KBUILD_OBJTREE gets reset to some weird
>value by accident and my whole home dir goes bye bye upon make
>mrproper?

<sarcasm>
Suppose cp gets aliased to rm -rf / by accident and my entire system
goes bye bye upon cp?
</sarcasm>

Re: kbuild 2.5 release 2.4

[Keith Owens]

On Sat, 4 May 2002 13:26:48 +0200 (SAST), 
Zwane Mwaikambo <zwane@linux.realnet.co.sz> wrote:
>	Just a minor nit, could you change the rm parameter to have -r 
>too? I've found that ^C halfway through a make can leave some old 
>directories (.tmp_include or somesuch).

You will have to provide more details of the problem that you think you
are seeing.  The .tmp_include files are part of the infrastructure that
lets me separate source and object trees, they are meant to be there.
make -f Makefile-2.5 mrproper deletes .tmp_include.

Re: kbuild 2.5 release 2.4

[tomas szepe]

> >Or even worse, suppose $KBUILD_OBJTREE gets reset to some weird
> >value by accident and my whole home dir goes bye bye upon make
> >mrproper?
> <sarcasm>
> Suppose cp gets aliased to rm -rf / by accident and my entire system
> goes bye bye upon cp?
> </sarcasm>

Hmm, I don't think this analogy will do -- working with aliases involving
fileutils as root is a way straight to hell, and hardly anyone ever walks
it. With kbuild-2.5, however, I have to set $KBUILD_OBJTREE every time
I want to build a kernel with objects out of the source dir -- and hey,
is there a single person on this list who's never made a typo on the
command line?

I don't know how to properly emphasize that this *is* asking for problems,
but still I'd be surprised if I were the only one scared by files not
connected to the build getting erased on make mrproper. Hello, anyone? :)

Would it be complicated to only kill the files the build knows it had
created?


-Tomas

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On Sat, 04 May 2002 00:19:10 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
>http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
>release 2.4.

kbuild-2.5-core-11
  Changes from core-10.

    Redo the passing of kbuild variables to $(shell) commands.

    Require make 3.79.1 for kbuild 2.5.

The make 'export' command does not export the variables to $(shell)
commands, which makes it messy when those commands need to access the
variables.  The workaround for this was a bit hackish and a core-10
change broke with common source and object.

core-11 uses a new method of passing the kbuild variables to $(shell)
commands.  It is cleaner, but it requires features that are not
available in make 3.77.  make 3.78 should work but since 3.79.1 has
been out for almost two years and is being used by any recent
distribution, the code says that make 3.79.1 is required for kbuild
2.5.

An update to 2.5/Documentation/Changes has been submitted.  Linus has
already done 2.5 changes that require make 3.79 features, the Changes
update is overdue.

Re: kbuild 2.5 release 2.4

[rddunlap]

On Sat, 4 May 2002, tomas szepe wrote:

| Hmm, I don't think this analogy will do -- working with aliases involving
| fileutils as root is a way straight to hell, and hardly anyone ever walks
| it. With kbuild-2.5, however, I have to set $KBUILD_OBJTREE every time
| I want to build a kernel with objects out of the source dir -- and hey,
| is there a single person on this list who's never made a typo on the
| command line?
|
| I don't know how to properly emphasize that this *is* asking for problems,
| but still I'd be surprised if I were the only one scared by files not
| connected to the build getting erased on make mrproper. Hello, anyone? :)

Too much policy here?

| Would it be complicated to only kill the files the build knows it had
| created?

That's what I would expect.

-- 
~Randy

Re: kbuild 2.5 release 2.4

[rddunlap]

On Sat, 4 May 2002, Zwane Mwaikambo wrote:

| On Sat, 4 May 2002, Keith Owens wrote:
|
| > You will have to provide more details of the problem that you think you
| > are seeing.  The .tmp_include files are part of the infrastructure that
| > lets me separate source and object trees, they are meant to be there.
| > make -f Makefile-2.5 mrproper deletes .tmp_include.
|
| I didn't try reproducing it, i just rm -rf'd .tmp_include when it failed.
| I'll double check.

Zwane-

I've seen this, but only when I did 'make clean' instead
of 'make -f Makefile-2.5 clean' (or was it mrproper?).
I.e., used the wrong Makefile.

-- 
~Randy

Re: kbuild 2.5 release 2.4

[Tomas Szepe]

> | Hmm, I don't think this analogy will do -- working with aliases involving
> | fileutils as root is a way straight to hell, and hardly anyone ever walks
> | it. With kbuild-2.5, however, I have to set $KBUILD_OBJTREE every time
> | I want to build a kernel with objects out of the source dir -- and hey,
> | is there a single person on this list who's never made a typo on the
> | command line?
> |
> | I don't know how to properly emphasize that this *is* asking for problems,
> | but still I'd be surprised if I were the only one scared by files not
> | connected to the build getting erased on make mrproper. Hello, anyone? :)
> 
> Too much policy here?
> 
> | Would it be complicated to only kill the files the build knows it had
> | created?
> 
> That's what I would expect.


Word. The purpose of all the 'clean' targets in Makefiles (distclean,
mrproper, etc.) everywhere has always been to get rid of the object/
build files, and the object/build files only.

Don't get me wrong, Keith, I believe kbuild 2.5's a terrific piece of
code, I'm just discussing what I've found myself to be in trouble with.


T.

Re: kbuild 2.5 release 2.4

[rddunlap]

On Sat, 4 May 2002, Tomas Szepe wrote:

| > | I don't know how to properly emphasize that this *is* asking for problems,
| > | but still I'd be surprised if I were the only one scared by files not
| > | connected to the build getting erased on make mrproper. Hello, anyone? :)
| >
| > Too much policy here?
| >
| > | Would it be complicated to only kill the files the build knows it had
| > | created?
| >
| > That's what I would expect.
|
|
| Word. The purpose of all the 'clean' targets in Makefiles (distclean,
| mrproper, etc.) everywhere has always been to get rid of the object/
| build files, and the object/build files only.
|
| Don't get me wrong, Keith, I believe kbuild 2.5's a terrific piece of
| code, I'm just discussing what I've found myself to be in trouble with.

Yes, I'd like to see kbuild 2.5 added to Linux 2.5 also,
but it does need some tweaking IMO.

-- 
~Randy

Re: kbuild 2.5 release 2.4

[Keith Owens]

On Sat, 4 May 2002 09:19:51 -0700 (PDT), 
<rddunlap@osdl.org> wrote:
>On Sat, 4 May 2002, tomas szepe wrote:
>
>| Hmm, I don't think this analogy will do -- working with aliases involving
>| fileutils as root is a way straight to hell, and hardly anyone ever walks
>| it. With kbuild-2.5, however, I have to set $KBUILD_OBJTREE every time
>| I want to build a kernel with objects out of the source dir -- and hey,
>| is there a single person on this list who's never made a typo on the
>| command line?
>|
>| I don't know how to properly emphasize that this *is* asking for problems,
>| but still I'd be surprised if I were the only one scared by files not
>| connected to the build getting erased on make mrproper. Hello, anyone? :)
>
>Too much policy here?
>
>| Would it be complicated to only kill the files the build knows it had
>| created?
>
>That's what I would expect.

OK, it is now officially classed as an over optimization.  I will
change clean/mrproper in core-12 to only delete known files.

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On Sat, 04 May 2002 00:19:10 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
>http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
>release 2.4.

kbuild-2.5-common-2.5.14-1 and kbuild-2.5-i386-2.5.14-1 are available.
Upgraded to kernel 2.5.14.

Re: [kbuild-devel] Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Thomas Duffy]

[-- Attachment #1: Type: text/plain, Size: 616 bytes --]

On Sun, 2002-05-05 at 23:27, Keith Owens wrote:
> kbuild-2.5-common-2.5.14-1 and kbuild-2.5-i386-2.5.14-1 are available.
> Upgraded to kernel 2.5.14.

here is the sparc64 kbuild against core-11 and 2.5.14

you must apply the linux-2.5.14-sparc64-hacks.patch first to get sparc64
to build on 2.5.14 (even using kbuild 2.4)

0) untar linux-2.5.14.tar.bz2
1) apply linux-2.5.14-sparc64-hacks.patch
2) apply kbuild-2.5-core-11
3) apply kbuild-2.5-common-2.5.14-1
4) apply kbuild-2.5-sparc64-2.5.14-1

you should be good to go...

this patch still needs work on the asm-offsets to get it work The Right
Way (tm)

-tduffy

[-- Attachment #2: linux-2.5.14-sparc64-hacks.patch --]
[-- Type: text/plain, Size: 2994 bytes --]

--- linux-2.5.14+kbuild-v2.1+common-1+core-11+i386-1/include/asm-sparc64/page.h	Sun May  5 20:38:01 2002
+++ linux-2.5.14+kbuild-v2.1+common-1+core-11+i386-1+sparc64-1/include/asm-sparc64/page.h	Mon May  6 11:02:04 2002
@@ -113,8 +113,17 @@
 
 #define __pa(x)			((unsigned long)(x) - PAGE_OFFSET)
 #define __va(x)			((void *)((unsigned long) (x) + PAGE_OFFSET))
+
+#if 0
 #define virt_to_page(kaddr)	(mem_map + ((__pa(kaddr)-phys_base) >> PAGE_SHIFT))
 #define VALID_PAGE(page)	((page - mem_map) < max_mapnr)
+#endif
+
+#define pfn_to_page(pfn)	(mem_map + (pfn))
+#define page_to_pfn(page)	((unsigned long)((page) - mem_map))
+#define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr-phys_base) >> PAGE_SHIFT)
+#define pfn_valid(pfn)		((pfn) < max_mapnr)
+#define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr-phys_base) >> PAGE_SHIFT)
 
 #define virt_to_phys __pa
 #define phys_to_virt __va
--- linux-2.5.14+kbuild-v2.1+common-1+core-11+i386-1/arch/sparc64/kernel/ioctl32.c	Sun May  5 20:37:59 2002
+++ linux-2.5.14+kbuild-v2.1+common-1+core-11+i386-1+sparc64-1/arch/sparc64/kernel/ioctl32.c	Mon May  6 11:17:15 2002
@@ -4553,8 +4553,8 @@
 COMPATIBLE_IOCTL(HCIDEVUP)
 COMPATIBLE_IOCTL(HCIDEVDOWN)
 COMPATIBLE_IOCTL(HCIDEVRESET)
-COMPATIBLE_IOCTL(HCIRESETSTAT)
-COMPATIBLE_IOCTL(HCIGETINFO)
+/* COMPATIBLE_IOCTL(HCIRESETSTAT) */
+/* COMPATIBLE_IOCTL(HCIGETINFO) */
 COMPATIBLE_IOCTL(HCIGETDEVLIST)
 COMPATIBLE_IOCTL(HCISETRAW)
 COMPATIBLE_IOCTL(HCISETSCAN)
--- linux-2.5.14+kbuild-v2.1+common-1+core-11+i386-1/include/asm-sparc64/pgtable.h	Sun May  5 20:37:53 2002
+++ linux-2.5.14+kbuild-v2.1+common-1+core-11+i386-1+sparc64-1/include/asm-sparc64/pgtable.h	Tue May  7 09:41:53 2002
@@ -194,13 +194,22 @@
 extern struct page *mem_map_zero;
 #define ZERO_PAGE(vaddr)	(mem_map_zero)
 
+#if 0
 /* Warning: These take pointers to page structs now... */
 #define mk_pte(page, pgprot)		\
 	__pte((((page - mem_map) << PAGE_SHIFT)+phys_base) | pgprot_val(pgprot) | _PAGE_SZBITS)
+#endif
+
+#define pfn_pte(pfn, pgprot)		\
+	__pte((((pfn) << PAGE_SHIFT)+phys_base) | pgprot_val(pgprot) | _PAGE_SZBITS)
+#define mk_pte(page, pgprot)	pfn_pte(page_to_pfn(page), (pgprot))
+
 #define page_pte_prot(page, prot)	mk_pte(page, prot)
 #define page_pte(page)			page_pte_prot(page, __pgprot(0))
 
+#if 0
 #define mk_pte_phys(physpage, pgprot)	(__pte((physpage) | pgprot_val(pgprot) | _PAGE_SZBITS))
+#endif
 
 extern inline pte_t pte_modify(pte_t orig_pte, pgprot_t new_prot)
 {
@@ -246,7 +255,12 @@
 /* Permanent address of a page. */
 #define __page_address(page)	page_address(page)
 
+#if 0
 #define pte_page(x) (mem_map+(((pte_val(x)&_PAGE_PADDR)-phys_base)>>PAGE_SHIFT))
+#endif
+
+#define pte_page(x) pfn_to_page(pte_pfn(x))
+#define pte_pfn(x) (((pte_val(x)&_PAGE_PADDR)-phys_base)>>PAGE_SHIFT)
 
 /* Be very careful when you change these three, they are delicate. */
 #define pte_mkyoung(pte)	(__pte(pte_val(pte) | _PAGE_ACCESSED | _PAGE_R))

[-- Attachment #3: kbuild-2.5-sparc64-2.5.14-1.bz2 --]
[-- Type: application/x-bzip, Size: 6641 bytes --]

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On Sat, 04 May 2002 23:01:17 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
>http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
>release 2.4.

Extra architecture support added to release 2.4:

  kbuild-2.5-ppc-2.5.14-1.bz2 from Paul Mackerras
  kbuild-2.5-sparc64-2.5.14-1.bz2 from Tom Duffy
  kbuild-2.5-sh-2.5.12-1.bz2 from Greg Banks

  Read the start of each patch for special instructions.

kbuild-2.5-core-12.bz2 added to release 2.4

  Changes from core-11.

    Verify that KBUILD_OBJTREE points to an objtree directory during
    make clean/mrproper.  Old object trees will get a warning,
      echo "must not be empty" > $KBUILD_OBJTREE/_objtree
    to mark old objtrees, only needed once.

    Change message when switching from normal mode to NO_MAKEFILE_GEN
    mode, document why the switch requires additional processing.

    Fix standardization of file names in commands.

    pp_ programs are compiled with -O2 -NDEBUG=1.

    Tune database layout to improve speed, especially for small builds.
    The new database format requires and will force a complete rebuild.

Re: [kbuild-devel] Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Thomas Duffy]

[-- Attachment #1: Type: text/plain, Size: 758 bytes --]

On Thu, 2002-05-09 at 04:45, Keith Owens wrote: 
> Extra architecture support added to release 2.4:
> 
>   kbuild-2.5-sparc64-2.5.14-1.bz2 from Tom Duffy

here is kbuild-2.5-sparc64-2.5.14-2.  you still need to apply the hacks
patch before using this from
http://prdownloads.sourceforge.net/kbuild/linux-2.5.14-sparc64-hacks.patch.bz2.  This won't be necessary once 2.5.15 comes out as Dave has already sent these fixes to Linus. 

  Changes from kbuild-2.5-sparc64-2.5.14-1 

    Build against core-12 

    Builds with kbuild 2.4 now as well

    asm-offsets.c now uses the new thread_info offsets 

    Had to use the CFLAG ugliness of ia64 to get asm-offsets.c to
    work properly as include/asm-sparc64/system.h uses thread_info 
    offsets.

-tduffy

[-- Attachment #2: kbuild-2.5-sparc64-2.5.14-2.bz2 --]
[-- Type: application/x-bzip, Size: 7354 bytes --]

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On 09 May 2002 18:46:18 -0700, 
Thomas Duffy <tduffy@directvinternet.com> wrote:
>here is kbuild-2.5-sparc64-2.5.14-2.  you still need to apply the hacks
>patch before using this from
>http://prdownloads.sourceforge.net/kbuild/linux-2.5.14-sparc64-hacks.patch.bz2.  This won't be necessary once 2.5.15 comes out as Dave has already sent these fixes to Linus. 
>
>  Changes from kbuild-2.5-sparc64-2.5.14-1 
>
>    Build against core-12 
>
>    Builds with kbuild 2.4 now as well
>
>    asm-offsets.c now uses the new thread_info offsets 
>
>    Had to use the CFLAG ugliness of ia64 to get asm-offsets.c to
>    work properly as include/asm-sparc64/system.h uses thread_info 
>    offsets.

Uploaded to http://sourceforge.net/project/showfiles.php?group_id=18813
under release 2.4.

Also uploaded are kbuild-2.5-common-2.5.15-1.bz2 and
kbuild-2.5-i386-2.5.15-1.bz2 for kernel 2.5.15 support.

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On Sat, 04 May 2002 00:19:10 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
>http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
>release 2.4.

kbuild-2.5-core-13
  Changes from core-12.

    Disable the ability to generate .[is] files on demand from the
    command line.  There are corner cases with autogenerated .[is] that
    are causing kbuild database problems so this new feature has been
    disabled for the moment.  If you try to generate a .[is] file,
    kbuild will give you instructions on how to do it by hand.

    Tweak standardize_name code to cope with aic7xxx shipping files and
    overwriting them with the same name.

kbuild-2.5-common-2.5.15-2
  Changes from common-2.5.15-1.

    Add -f to cp/mv.  Mainly to cope with aic7xxx shipping files and 
    overwriting them with the same name.

    Add missing select for aic7xxx.

kbuild-2.5-i386-2.5.15-2
  Changes from i386-2.5.15-1.

    arch/i386/pci/Makefile.in fixups.  Add acpi flags to acpi.o.  Only
    build pci objects when CONFIG_PCI=y.

Re: [kbuild-devel] Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Thomas Duffy]

[-- Attachment #1: Type: text/plain, Size: 1453 bytes --]

On Sat, 2002-05-11 at 02:32, Keith Owens wrote:
> On Sat, 04 May 2002 00:19:10 +1000, 
> Keith Owens <kaos@ocs.com.au> wrote:
> >Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
> >http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
> >release 2.4.

ok, kbuild-2.5-sparc64-2.5.15-1 is ready.

   Changes from 2.5.14-2

    Update to 2.5.15

    Change all #include <asm-offsets.h> to #include "asm-offsets.h"

    Fix up setup hack for recursive include in assembler offsets.

> kbuild-2.5-core-13
>   Changes from core-12.

<snip>

>     Tweak standardize_name code to cope with aic7xxx shipping files and
>     overwriting them with the same name.
> 
> kbuild-2.5-common-2.5.15-2
>   Changes from common-2.5.15-1.
> 
>     Add -f to cp/mv.  Mainly to cope with aic7xxx shipping files and 
>     overwriting them with the same name.
> 
>     Add missing select for aic7xxx.

Keith, now I am getting this error with latest kbuild core-13,
common-2.5.15-2 (that wasn't happening before):

  CC drivers/scsi/aic7xxx/aic7xxx_core.o
pp_makefile5: stat (drivers/scsi/aic7xxx/aic7xxx_seq.h) failed: No such file or directory
pp_makefile5: dependencies (pid 15566) returned 1
make[1]: *** [drivers/scsi/aic7xxx/aic7xxx_core.o] Error 1

$ ls $KBUILD_SRCTREE_000/drivers/scsi/aic7xxx/aic7xxx_seq.h
/darwin_build2/tduffy/linux-2.5.15+kbuild-v2.4+common-2+core-13+i386-2+sparc64-1/drivers/scsi/aic7xxx/aic7xxx_seq.h

-tduffy

[-- Attachment #2: kbuild-2.5-sparc64-2.5.15-1.bz2 --]
[-- Type: application/x-bzip, Size: 7307 bytes --]

Re: [kbuild-devel] Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On 13 May 2002 17:12:08 -0700, 
Thomas Duffy <tduffy@directvinternet.com> wrote:
>ok, kbuild-2.5-sparc64-2.5.15-1 is ready.

Does it still need sparc64 hacks or is the base 2.5.15 kernel clean for
sparc64?

>  CC drivers/scsi/aic7xxx/aic7xxx_core.o
>pp_makefile5: stat (drivers/scsi/aic7xxx/aic7xxx_seq.h) failed: No such file or directory
>pp_makefile5: dependencies (pid 15566) returned 1

Did not show up in my testing, it was hidden by out of order compiles
with -j4.  Will be in common-2.5.15-3.

diff -ur 2.5.15-kbuild-2.5/drivers/scsi/aic7xxx/Makefile.in 2.5.15-kbuild-2.5.new/drivers/scsi/aic7xxx/Makefile.in
--- 2.5.15-kbuild-2.5/drivers/scsi/aic7xxx/Makefile.in	Tue May 14 12:38:47 2002
+++ 2.5.15-kbuild-2.5.new/drivers/scsi/aic7xxx/Makefile.in	Tue May 14 12:37:20 2002
@@ -19,9 +19,9 @@
 
 extra_cflags_all($(src_includelist /drivers/scsi))
 
-# Only aic7xxx.c includes aic7xxx_seq.h.
+# Only aic7xxx_core.c includes aic7xxx_seq.h.
 
-$(objfile aic7xxx.o): $(objfile aic7xxx_seq.h)
+$(objfile aic7xxx_core.o): $(objfile aic7xxx_seq.h)
 
 # aic7xxx_reg.h is messy.  It is included by aic7xxx.h which is included by
 # aic7xxx_osm.h which is included by several .c files.  Play safe and make all

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On Sat, 04 May 2002 00:19:10 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
>http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
>release 2.4.

kbuild-2.5-core-14
  Changes from core-13.

    Correct typo in safety check for make clean/mrproper.

    Add sys/time.h to pp_makefile5, not all libraries include
    sys/time.h from time.h.

kbuild-2.5-common-2.5.15-3
  Changes from common-2.5.15-2.

    Revert cp -fa to cp for modules_install.  Modules install as root,
    not builder.

    Tweak aic7xxx dependencies.

    Makefile.in changes fr9om s390 group.

kbuild-2.5-sparc64-2.5.15-1 (new)
kbuild-2.5-s390-2.5.15-1 (new)
kbuild-2.5-s390x-2.5.15-1 (new)
kbuild-2.5-ppc-2.5.14-1 (new)

Greg Banks reports that SuperH is up to 2.5.13 and the existing
kbuild-2.5-sh-2.5.12-1 patch will work on 2.5.13.

Re: [kbuild-devel] Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Thomas Duffy]

[-- Attachment #1: Type: text/plain, Size: 1039 bytes --]

On Mon, 2002-05-13 at 22:44, Keith Owens wrote:
> On Sat, 04 May 2002 00:19:10 +1000, 
> Keith Owens <kaos@ocs.com.au> wrote:
> >Release 2.4 of kernel build for kernel 2.5 (kbuild 2.5) is available.
> >http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
> >release 2.4.

Just for shits and giggles, I updated kbuild 2.5 for 2.4.19-pre8.

(actually, I did this because sparc32 builds on that kernel and I am
going to start work to get that port underway and wanted a buildable
base to start with)

this patch is just the diff between 2.4.18 and 2.4.19-pre8

0) untar 2.4.18 and apply patch-2.4.19-pre8
1) apply kbuild-2.5-core-14
2) apply kbuild-2.5-common-2.4.18-6
3) apply kbuild-2.5-i386-2.4.18-2
4) then apply kbuild-2.5-common-2.4.19-pre8-update

I have only tested x86 as of this moment.  If anyone is dying for a
clean kbuild-2.5-common-2.4.19-pre8-1, let me know and I can turn those
wheels.  BTW, there are a few things that don't build on this kernel in
kbuild 2.4 as probably already reported on lkml.

-tduffy

[-- Attachment #2: kbuild-2.5-common-2.4.19-pre8-update.bz2 --]
[-- Type: application/x-bzip, Size: 10004 bytes --]

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On 14 May 2002 20:46:38 -0700, 
Thomas Duffy <tduffy@directvinternet.com> wrote:
>I have only tested x86 as of this moment.  If anyone is dying for a
>clean kbuild-2.5-common-2.4.19-pre8-1, let me know and I can turn those
>wheels.

I have uploaded common-2.4.18-7 and common-2.5.15-4 to sourceforge.
http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
release 2.4.  This cross syncs the 2.4.18 and 2.5.15 architecture
independent code, where the two kernels have similar requirements.  I
will take Tom Duffy's 2.4.19-pre8 patch, add it to common-2.4.18-7 and
issue common-2.4.19-pre8-1 tomorrow.

Re: Announce: Kernel Build for 2.5, Release 2.4 is available

[Keith Owens]

On Wed, 15 May 2002 23:20:38 +1000, 
Keith Owens <kaos@ocs.com.au> wrote:
>On 14 May 2002 20:46:38 -0700, 
>Thomas Duffy <tduffy@directvinternet.com> wrote:
>>I have only tested x86 as of this moment.  If anyone is dying for a
>>clean kbuild-2.5-common-2.4.19-pre8-1, let me know and I can turn those
>>wheels.
>
>I have uploaded common-2.4.18-7 and common-2.5.15-4 to sourceforge.
>http://sourceforge.net/projects/kbuild/, package kbuild-2.5, download
>release 2.4.  This cross syncs the 2.4.18 and 2.5.15 architecture
>independent code, where the two kernels have similar requirements.  I
>will take Tom Duffy's 2.4.19-pre8 patch, add it to common-2.4.18-7 and
>issue common-2.4.19-pre8-1 tomorrow.

kbuild-2.5-common-2.4.19-pre8-1.bz2, kbuild-2.5-i386-2.4.19-pre8-1.bz2
are available.  Use with core-14.

I built these from common-2.4.18-7, i386-2.4.18-2 plus some previous
work on 2.4.19-pre6 then updated to -pre8.  A cross check against Tom
Duffy's 19-pre8 patch found errors in my patches and Tom's, which have
been reconciled.