[Buildroot] ALSA for build root

[Buildroot] ALSA for build root

[Rod Boyce]

All,

I have been adding ALSA to the build root package system I am only able 
to test on a x86 platform.  Is anybody else interested in ALSA for 
build-root?  I have also been adding libogg and libvorbis would it be OK 
if I submitted a combined patch or would you prefer 2 separate patches?  
Is anybody else working on this who would like an advanced copy to test 
out and or compare against?

This would include the following packages:
alsa-lib-1.0.14rc4
alsa-utils-1.0.14rc4
libogg-1.1.3
libvorbis-1.1.2


Regards,
Rod Boyce

[Buildroot] ALSA for build root

[Daniel Laird]

We have added alsa for use on our projects so would like to see someone get
this into mainline buildroot

Cheers
Daniel Laird


Rod Boyce-2 wrote:
> 
> All,
> 
> I have been adding ALSA to the build root package system I am only able 
> to test on a x86 platform.  Is anybody else interested in ALSA for 
> build-root?  I have also been adding libogg and libvorbis would it be OK 
> if I submitted a combined patch or would you prefer 2 separate patches?  
> Is anybody else working on this who would like an advanced copy to test 
> out and or compare against?
> 
> This would include the following packages:
> alsa-lib-1.0.14rc4
> alsa-utils-1.0.14rc4
> libogg-1.1.3
> libvorbis-1.1.2
> 
> 
> Regards,
> Rod Boyce
> _______________________________________________
> buildroot mailing list
> buildroot at uclibc.org
> http://busybox.net/mailman/listinfo/buildroot
> 
> 

-- 
View this message in context: http://www.nabble.com/ALSA-for-build-root-tf3816584.html#a10805002
Sent from the BuildRoot mailing list archive at Nabble.com.

[Buildroot] ALSA for build root

[Peter at busybox.net]

>>>>> "RB" == Rod Boyce <buildroot@teamboyce.co.uk> writes:

Hi,

RB> All, I have been adding ALSA to the build root package system I am
RB> only able to test on a x86 platform.  Is anybody else interested
RB> in ALSA for build-root?

Yes, please. I can test on powerpc.

RB> I have also been adding libogg and libvorbis would it be OK if I
RB> submitted a combined patch or would you prefer 2 separate patches?
RB> Is anybody else working on this who would like an advanced copy to
RB> test out and or compare against?

Seperate patches please.

Thanks!

-- 
Bye, Peter Korsgaard

[Buildroot] ALSA for build root

[Stuart Wood]

I'm doing ARM with a custom audio section that still needs a driver, but
I'll try when I'm there.


-----Original Message-----
From: buildroot-bounces@uclibc.org [mailto:buildroot-bounces at uclibc.org] On
Behalf Of Peter at busybox.net
Sent: Friday, May 25, 2007 2:57 PM
To: Rod Boyce
Cc: buildroot at uclibc.org
Subject: Re: [Buildroot] ALSA for build root


>>>>> "RB" == Rod Boyce <buildroot@teamboyce.co.uk> writes:

Hi,

RB> All, I have been adding ALSA to the build root package system I am 
RB> only able to test on a x86 platform.  Is anybody else interested in 
RB> ALSA for build-root?

Yes, please. I can test on powerpc.

RB> I have also been adding libogg and libvorbis would it be OK if I 
RB> submitted a combined patch or would you prefer 2 separate patches? 
RB> Is anybody else working on this who would like an advanced copy to 
RB> test out and or compare against?

Seperate patches please.

Thanks!

-- 
Bye, Peter Korsgaard _______________________________________________
buildroot mailing list
buildroot at uclibc.org http://busybox.net/mailman/listinfo/buildroot

[Buildroot] Package selection question

[Tom]

Hello,

I just want to build the toolchain and the debugger on host: cygwin, 
target: arm.

So, even after I unchecked busybox and all packages, it still builds gmp 
(and, unsuccessfully, mpfr). Why does it need it please and why is this 
an option then? Or am I doing something wrong here?

Thanks
Tom

[Buildroot] Package selection question

[Yann E. MORIN]

Tom,
All,

On Friday 25 May 2007 21:23, Tom wrote:
> So, even after I unchecked busybox and all packages, it still builds gmp 
> (and, unsuccessfully, mpfr). Why does it need it please and why is this 
> an option then? Or am I doing something wrong here?

gmp and mpfr are needed to build gcc. They are to run on host, not on target.
Can you post the build logs for mpfr (and gmp if it also fails)?

Regards,
Yann E. MORIN.

PS. gcc-4.1.x and earlier require neither gmp nor mpfr.
    gcc-4.2.x require gmp and mpfr to build fortran (at least, may be others)
    gcc-4.3.x will require gmp and mpfr to build all frontends.
YEM.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |   ^                |
| --==< ?_? >==-- ?------------.-------:  X  AGAINST      |  /e\  There is no  |
| http://ymorin.is-a-geek.org/ | (*_*) | / \ HTML MAIL    |  """  conspiracy.  |
?------------------------------?-------?------------------?--------------------?

[Buildroot] Package selection question

[Bernhard Fischer]

On Fri, May 25, 2007 at 12:23:37PM -0700, Tom wrote:
>
>Hello,
>
>I just want to build the toolchain and the debugger on host: cygwin, 
>target: arm.
>
>So, even after I unchecked busybox and all packages, it still builds gmp 
>(and, unsuccessfully, mpfr). Why does it need it please and why is this 
>an option then? Or am I doing something wrong here?

gcc uses gmp/mpfr, you're doing nothing wrong. Please report build
failures of gmp/mpfr to the respective lists.

Before you report it to the mpfr folks, please paste the error here,
since i suspect that it could have to do something with the
binary/library name extensions, which (IIRC) we currently do not handle
at all (there's a bug somewhere about this, if memory serves me right).

[Buildroot] Package selection question

[Bernhard Fischer]

On Fri, May 25, 2007 at 12:23:37PM -0700, Tom wrote:
>
>Hello,

heck, and please don't hijack unrelated threads!

[Buildroot] Package selection question

[Bernhard Fischer]

On Fri, May 25, 2007 at 09:51:38PM +0200, Yann E. MORIN wrote:

>PS. gcc-4.1.x and earlier require neither gmp nor mpfr.

Is this true? IIRC the middle-end of 4.1 does use them, no?

>    gcc-4.2.x require gmp and mpfr to build fortran (at least, may be others)
>    gcc-4.3.x will require gmp and mpfr to build all frontends.

[Buildroot] Package selection question

[Yann E. MORIN]

Bernhard,
All,

On Friday 25 May 2007 22:38, Bernhard Fischer wrote:
> On Fri, May 25, 2007 at 09:51:38PM +0200, Yann E. MORIN wrote:
> >PS. gcc-4.1.x and earlier require neither gmp nor mpfr.
> Is this true? IIRC the middle-end of 4.1 does use them, no?

I don't have mpfr here and gcc-4.1.2 does build OK. I do have gmp, but gcc does
not linked against it.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |   ^                |
| --==< ?_? >==-- ?------------.-------:  X  AGAINST      |  /e\  There is no  |
| http://ymorin.is-a-geek.org/ | (*_*) | / \ HTML MAIL    |  """  conspiracy.  |
?------------------------------?-------?------------------?--------------------?

[Buildroot] Package selection question

[Tom]

Thanks.

I am working to isolate and find a workaround to my mpfr problem on 
cygwin. Once I do, I will post the results of all workarounds I had to 
do (or the error log should I not get it done). In a separate thread as 
requested.

Bernhard Fischer wrote:
> On Fri, May 25, 2007 at 12:23:37PM -0700, Tom wrote:
>> Hello,
>>
>> I just want to build the toolchain and the debugger on host: cygwin, 
>> target: arm.
>>
>> So, even after I unchecked busybox and all packages, it still builds gmp 
>> (and, unsuccessfully, mpfr). Why does it need it please and why is this 
>> an option then? Or am I doing something wrong here?
> 
> gcc uses gmp/mpfr, you're doing nothing wrong. Please report build
> failures of gmp/mpfr to the respective lists.
> 
> Before you report it to the mpfr folks, please paste the error here,
> since i suspect that it could have to do something with the
> binary/library name extensions, which (IIRC) we currently do not handle
> at all (there's a bug somewhere about this, if memory serves me right).
> 

[Buildroot] Package selection question

[Tom]

I have selected "Buildroot toolchain" and in the toolchain->GCC option I 
have selected 4.1.2.

(HOSTCC I have tried both 3.4.4 and 4.2.0)

Why does it download and try to build mpfr if it is not used?


Yann E. MORIN wrote:
> Bernhard,
> All,
> 
> On Friday 25 May 2007 22:38, Bernhard Fischer wrote:
>> On Fri, May 25, 2007 at 09:51:38PM +0200, Yann E. MORIN wrote:
>>> PS. gcc-4.1.x and earlier require neither gmp nor mpfr.
>> Is this true? IIRC the middle-end of 4.1 does use them, no?
> 
> I don't have mpfr here and gcc-4.1.2 does build OK. I do have gmp, but gcc does
> not linked against it.
> 
> Regards,
> Yann E. MORIN.
> 

[Buildroot] mpfr error on cygwin for arm

[Tom]

Tom wrote:
> because it does not seem to 
> recognize the "--rpath" options,

Never mind this part. Sorry, it has been a long day.

[Buildroot] Package selection question

[Bernhard Fischer]

On Fri, May 25, 2007 at 10:45:32PM +0200, Yann E. MORIN wrote:
>Bernhard,
>All,
>
>On Friday 25 May 2007 22:38, Bernhard Fischer wrote:
>> On Fri, May 25, 2007 at 09:51:38PM +0200, Yann E. MORIN wrote:
>> >PS. gcc-4.1.x and earlier require neither gmp nor mpfr.
>> Is this true? IIRC the middle-end of 4.1 does use them, no?
>
>I don't have mpfr here and gcc-4.1.2 does build OK. I do have gmp, but gcc does
>not linked against it.

Fair enough, in 4.1, only the fortran frontend needs it.

[Buildroot] mpfr error on cygwin for arm

[Bernhard Fischer]

On Fri, May 25, 2007 at 11:52:35PM -0700, Tom wrote:
>Hi Bernhard and all,
>
>sorry if this thread is a little out of sequence, had problems with text 
>attachment.
>
>Bernhard Fischer wrote:
>>Before you report it to the mpfr folks, please paste the error here,
>>since i suspect that it could have to do something with the
>>binary/library name extensions, which (IIRC) we currently do not handle
>>at all (there's a bug somewhere about this, if memory serves me right).
>>
>
>You were right about extensions, below is the error as produced by
>buildroot-20050516 built, I just added -verbose for ld.
>It can not resolve the depency to libgmp.so, because cygwin ld assumes 
>libs to end in .a, .lib or .dll, and will append them to .so libs, which 
>thus can never be found.
>Can I help to add support for this?

Yes, you can.

Depending on the target, set
ifneq $($(findstring linux,$(BR2_GNU_BUILD_SUFFIX)),)
EXEEXT:=
LIBEXT:=.a
SHREXT:=.so
endif
ifneq $($(findstring apple,$(BR2_GNU_BUILD_SUFFIX)),)
EXEEXT:=.pear
LIBEXT:=.dunno
SHREXT:=.dylib
endif
ifneq $($(findstring cygwin,$(BR2_GNU_BUILD_SUFFIX)),)
EXEEXT:=.exe
LIBEXT:=.dunno
SHREXT:=.dll
endif
ifneq $($(findstring mingw,$(BR2_GNU_BUILD_SUFFIX)),)
EXEEXT:=.exe
LIBEXT:=.dunno
SHREXT:=.dll
endif

in the toplevel Makefile, before the
all: world
target, and use them accordingly.


While you're at it, it would be awesome if you could deal with this too:
We want to be able to select which flavour of lib is built (shared or
static) with a single config-option.
So, instead of changing all ".so" blindly to $(SHREXT), change it to
$(LIBTGTEXT) which is set either to

ifeq $($(BR2_DEFAULT_LIB_TARGET),shared)
LIBTGTEXT=$(SHREXT)
else
LIBTGTEXT=$(LIBEXT)
endif

below the newly added block i mentioned above.

TIA and cheers,

[Buildroot] Package selection question

[Bernhard Fischer]

On Fri, May 25, 2007 at 01:59:06PM -0700, Tom wrote:
>Thanks.
>
>I am working to isolate and find a workaround to my mpfr problem on 

See this for a real fix instead:
http://buildroot.uclibc.org/lists/buildroot/2007-May/002801.html

>cygwin. Once I do, I will post the results of all workarounds I had to 
>do (or the error log should I not get it done). In a separate thread as 
>requested.
>
>Bernhard Fischer wrote:
>>On Fri, May 25, 2007 at 12:23:37PM -0700, Tom wrote:
>>>Hello,
>>>
>>>I just want to build the toolchain and the debugger on host: cygwin, 
>>>target: arm.
>>>
>>>So, even after I unchecked busybox and all packages, it still builds gmp 
>>>(and, unsuccessfully, mpfr). Why does it need it please and why is this 
>>>an option then? Or am I doing something wrong here?
>>
>>gcc uses gmp/mpfr, you're doing nothing wrong. Please report build
>>failures of gmp/mpfr to the respective lists.
>>
>>Before you report it to the mpfr folks, please paste the error here,
>>since i suspect that it could have to do something with the
>>binary/library name extensions, which (IIRC) we currently do not handle
>>at all (there's a bug somewhere about this, if memory serves me right).
>>
>

[Buildroot] Package selection question

[Yann E. MORIN]

Bernhard,
All,

On Saturday 26 May 2007 11:23, Bernhard Fischer wrote:
> On Fri, May 25, 2007 at 10:45:32PM +0200, Yann E. MORIN wrote:
> >I don't have mpfr here and gcc-4.1.2 does build OK. I do have gmp, but gcc does
> >not linked against it.
> Fair enough, in 4.1, only the fortran frontend needs it.

So gmp and mpfr are needed for Fortran from 4.1 onward. I thought only 4.2
would need them. I didn't stumble upon this as I don't do Fortran...

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |   ^                |
| --==< ?_? >==-- ?------------.-------:  X  AGAINST      |  /e\  There is no  |
| http://ymorin.is-a-geek.org/ | (*_*) | / \ HTML MAIL    |  """  conspiracy.  |
?------------------------------?-------?------------------?--------------------?

[Buildroot] ALSA for build root

[Rod Boyce]

Ulf Samuelsson wrote:
>  ________________ ?rende:	[Buildroot] ALSA for build root F?rfattare:	Rod Boyce <buildroot@teamboyce.co.uk> Datum:		2007 maj 25:e 14:50 
> All,
>
> I have been adding ALSA to the build root package system I am only able 
> to test on a x86 platform.  Is anybody else interested in ALSA for 
> build-root?  I have also been adding libogg and libvorbis would it be OK 
> if I submitted a combined patch or would you prefer 2 separate patches?  
> Is anybody else working on this who would like an advanced copy to test 
> out and or compare against?
>
> This would include the following packages:
> alsa-lib-1.0.14rc4
> alsa-utils-1.0.14rc4
> libogg-1.1.3
> libvorbis-1.1.2
>
>
> Regards,
> Rod Boyce
>
>
> I know that some guys are working on moving the 
> existing AVR32 ALSA stuff
> to buildroot.
> You might want to include
> the ALSA patches found
> at avr32linux.org
>
> Best Regards
> Ulf Samuelsson
>
>
>   
Ulf & All,

My preference would be to submit what I have so others can play and 
submit their patches after testing.
I am ready to e-mail my patch now without the AVR32 changes.

Regards,
Rod Boyce

[Buildroot] ALSA for build root

[Rod Boyce]

Rod Boyce wrote:
> All,
>
> I have been adding ALSA to the build root package system I am only able 
> to test on a x86 platform.  Is anybody else interested in ALSA for 
> build-root?  I have also been adding libogg and libvorbis would it be OK 
> if I submitted a combined patch or would you prefer 2 separate patches?  
> Is anybody else working on this who would like an advanced copy to test 
> out and or compare against?
>
> This would include the following packages:
> alsa-lib-1.0.14rc4
> alsa-utils-1.0.14rc4
> libogg-1.1.3
> libvorbis-1.1.2
>
>
> Regards,
> Rod Boyce
> _______________________________________________
> buildroot mailing list
> buildroot at uclibc.org
> http://busybox.net/mailman/listinfo/buildroot
>
>   
All,

As promised attached is a patch to add ALSA packages to build root I 
have only tested on i386 but it seems to work for me.  This patch is 
against the latest SVN repository No. 18694.  I have not tested the 
alsa-tools well and in-fact because I am not using X or GTK most will 
not compile I have disabled all the the ac3dec by default.  Also I am 
assuming the latest version of ALSA and a compatible kernel.

Regards,
Rod Boyce
-------------- next part --------------
A non-text attachment was scrubbed...
Name: alsa.patch
Type: text/x-patch
Size: 23409 bytes
Desc: not available
Url : http://busybox.net/lists/buildroot/attachments/20070526/0dee6bb5/attachment.bin 

[Buildroot] ALSA for build root

[Hans-Christian Egtvedt]

On Sat, 2007-05-26 at 15:45 +0100, Rod Boyce wrote:
> Rod Boyce wrote:

<cut intro about alsa patches>

> As promised attached is a patch to add ALSA packages to build root I 
> have only tested on i386 but it seems to work for me.  This patch is 
> against the latest SVN repository No. 18694.  I have not tested the 
> alsa-tools well and in-fact because I am not using X or GTK most will 
> not compile I have disabled all the the ac3dec by default.  Also I am 
> assuming the latest version of ALSA and a compatible kernel.

I am working on the exactly same package (alsa + alsa-util) for
buildroot, so far testing on an AVR32 platform.

> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
> --exclude=libvorbis ../trunk/package/alsa-lib/alsa-lib.mk
> package/alsa-lib/alsa-lib.mk
> --- ../trunk/package/alsa-lib/alsa-lib.mk       1970-01-01
> 01:00:00.000000000 +0100
> +++ package/alsa-lib/alsa-lib.mk        2007-05-26 08:17:42.849975524
> +0100
> @@ -0,0 +1,135 @@
> +#############################################################
> +#
> +# ALSA library
> +#
> +#############################################################
> +
> +ALSA-LIB_VERSION=1.0.14rc4
> +ALSA-LIB_SOURCE=alsa-lib-$(ALSA-LIB_VERSION).tar.bz2
> +ALSA-LIB_SITE=ftp://ftp.alsa-project.org/pub/lib/
> +ALSA-LIB_DIR=$(BUILD_DIR)/${shell basename
> $(ALSA-LIB_SOURCE) .tar.bz2}
> +ALSA-LIB_WORKDIR=$(BUILD_DIR)/alsa-lib-$(ALSA-LIB_VERSION)
> +ALSA-LIB_CAT:=$(BZCAT)
> +ALLNKSTRGS :=
> +
> +$(DL_DIR)/$(ALSA-LIB_SOURCE):
> +       $(WGET) -P $(DL_DIR) $(ALSA-LIB_SITE)/$(ALSA-LIB_SOURCE)
> +
> +$(ALSA-LIB_DIR)/.unpacked: $(DL_DIR)/$(ALSA-LIB_SOURCE)
> +       $(ALSA-LIB_CAT) $(DL_DIR)/$(ALSA-LIB_SOURCE) | tar -C
> $(BUILD_DIR) $(TAR_OPTIONS) -
> +       touch $(ALSA-LIB_DIR)/.unpacked
> +
> +$(ALSA-LIB_DIR)/.configured: $(ALSA-LIB_DIR)/.unpacked
> +       (cd $(ALSA-LIB_DIR); rm -rf config.cache; \
> +               $(TARGET_CONFIGURE_OPTS) \
> +               ./configure \
> +               --target=$(GNU_TARGET_NAME) \
> +               --host=$(GNU_TARGET_NAME) \
> +               --build=$(GNU_HOST_NAME) \
> +               --prefix=/usr \
> +               --exec-prefix=/usr \
> +               --sysconfdir=/etc \
> +               --libdir=/lib \

Why /lib? I prefer user libraries in /usr/lib.

AFAIK specifying --prefix=/usr should be more than enough.

> +               --datarootdir=$(STAGING_DIR)/usr/share \
> +               --datadir=/usr/share \
> +               --exec-prefix=/usr/bin \
> +               --includedir=$(STAGING_DIR)/include \
> +               --disable-docs \
> +               --with-debug=yes \

Make this optional at least, I do not need debug symbols when it is
working.

> +               $(DISABLE_NLS) \
> +       );
> +       touch $(ALSA-LIB_DIR)/.configured
> +
> +$(ALSA-LIB_WORKDIR)/src/libasound.la: $(ALSA-LIB_DIR)/.configured
> +       rm -f $@
> +       $(MAKE) CFLAGS=-g CC=$(TARGET_CC) -C $(ALSA-LIB_WORKDIR)
> +
> +$(STAGING_DIR)/lib/libasound.so.2.0.0:
> $(ALSA-LIB_WORKDIR)/src/libasound.la
> +       $(MAKE) prefix=$(STAGING_DIR) \
> +           exec_prefix=$(STAGING_DIR) \
> +           bindir=$(STAGING_DIR)/bin \
> +           sbindir=$(STAGING_DIR)/sbin \
> +           libexecdir=$(STAGING_DIR)/bin \
> +           datadir=$(STAGING_DIR)/share \
> +           sysconfdir=$(STAGING_DIR)/etc \
> +           sharedstatedir=$(STAGING_DIR)/com \
> +           localstatedir=$(STAGING_DIR)/var \
> +           libdir=$(STAGING_DIR)/lib \
> +           includedir=$(STAGING_DIR)/include \
> +           oldincludedir=$(STAGING_DIR)/include \
> +           infodir=$(STAGING_DIR)/info \
> +           mandir=$(STAGING_DIR)/man \

All these dir can easily be replaced with DESTDIR=$(STAGING_DIR).

> +           -C $(ALSA-LIB_DIR) install;
> +       @touch -c $@
> +
> +$(TARGET_DIR)/lib/libasound.so.2.0.0:
> $(STAGING_DIR)/lib/libasound.so.2.0.0
> +       @mkdir -p $(TARGET_DIR)/usr/share/alsa
> +       @mkdir -p $(TARGET_DIR)/usr/lib/alsa-lib/smixer
> +       cp -dpf $(STAGING_DIR)/lib/libasound.so* $(TARGET_DIR)/lib
> +       cp -rdpf $(STAGING_DIR)/share/alsa/*
> $(TARGET_DIR)/usr/share/alsa/
> +       cp -rdpf $(STAGING_DIR)/lib/alsa-lib/*
> $(TARGET_DIR)/usr/lib/alsa-lib/
> +       -$(STRIP) --strip-unneeded $(TARGET_DIR)/lib/libasound.so*
> +       -$(STRIP) --strip-unneeded
> $(TARGET_DIR)/usr/lib/alsa-lib/smixer/smixer-*.so
> +       @touch -c $(TARGET_DIR)/lib/libasound.so.2.0.0
> +
> +$(TARGET_DIR)/dev/dsp:
> +       @(cd $(dir $@) && ln -sf dsp0 $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/dsp

This is handled by mdev in busybox. And when using tmpfs on /dev this is
not even visible for the filesystem. If these symlinks are needed, I
think they should added to target/ instead.

> +$(TARGET_DIR)/dev/adsp:
> +       @(cd $(dir $@) && ln -sf adsp0 $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/adsp

Same as above.

> +$(TARGET_DIR)/dev/sequencer2:
> +       @(cd $(dir $@) && ln -sf music $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/sequencer2

Same as above.

> +$(TARGET_DIR)/dev/amidi:
> +       @(cd $(dir $@) && ln -sf amidi0 $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/amidi

Same as above.

> +$(TARGET_DIR)/dev/audio:
> +       @(cd $(dir $@) && ln -sf audio0 $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/audio

Same as above.

> +$(TARGET_DIR)/dev/midi:
> +       @(cd $(dir $@) && ln -sf midi00 $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/midi

Same as above.

> +$(TARGET_DIR)/dev/mixer:
> +       @(cd $(dir $@) && ln -sf mixer0 $(notdir $@))
> +ALLNKSTRGS += $(TARGET_DIR)/dev/mixer

Same as above.

> +$(TARGET_DIR)/usr/lib/libasound.la:
> $(STAGING_DIR)/lib/libasound.so.2.0.0
> +       @mkdir -p $(TARGET_DIR)/usr/include/alsa/sound
> +       @mkdir -p $(TARGET_DIR)/usr/include/sys
> +       cp -rdpf $(STAGING_DIR)/include/alsa $(TARGET_DIR)/usr/include
> +       cp -dpf $(STAGING_DIR)/include/sys/asoundlib.h
> $(TARGET_DIR)/usr/include/sys/
> +       cp -dpf $(STAGING_DIR)/lib/libasound.la $(TARGET_DIR)/usr/lib/
> +       @touch -c $(TARGET_DIR)/usr/lib/libasound.la
> +
> +alsa-lib-headers: $(TARGET_DIR)/usr/lib/libasound.la
> +
> +alsa-lib: linux26 $(TARGET_DIR)/lib/libasound.so.2.0.0 $(ALLNKSTRGS)
> +
> +alsa-lib-source: $(DL_DIR)/$(ALSA-LIB_SOURCE)
> +
> +alsa-lib-clean:
> +       @if [ -d $(ALSA-LIB_WORKDIR)/Makefile ] ; then \
> +               $(MAKE) -C $(ALSA-LIB_WORKDIR) clean ; \
> +       fi;
> +
> +alsa-lib-dirclean:
> +       rm -rf $(ALSA-LIB_DIR) $(ALSA-LIB_WORKDIR)
> +#############################################################
> +#
> +# Toplevel Makefile options
> +#
> +#############################################################
> +ifeq ($(strip $(BR2_PACKAGE_ALSA-LIB)),y)
> +TARGETS+=alsa-lib
> +endif
> +
> +ifeq ($(strip $(BR2_PACKAGE_ALSA-LIB_TARGET_HEADERS)),y)
> +TARGETS+=alsa-lib-headers
> +endif
> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
> --exclude=libvorbis ../trunk/package/alsa-lib/Config.in
> package/alsa-lib/Config.in
> --- ../trunk/package/alsa-lib/Config.in 1970-01-01 01:00:00.000000000
> +0100
> +++ package/alsa-lib/Config.in  2007-05-07 18:16:23.220365544 +0100
> @@ -0,0 +1,15 @@
> +config BR2_PACKAGE_ALSA-LIB
> +       bool "alsa-lib"
> +       default n
> +       help
> +         Advanced Linux Sound Architecture
> +
> +         http://www.alsa-project.org/
> +
> +config BR2_PACKAGE_ALSA-LIB_TARGET_HEADERS
> +       bool "alsa-lib headers in target"
> +       default n
> +       depends on BR2_PACKAGE_ALSA-LIB
> +       help
> +         Put headers files in the target.
> +

<cut alsa-tools Makefile and Config.in>

> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
> --exclude=libvorbis ../trunk/package/alsa-utils/alsa-utils.mk
> package/alsa-utils/alsa-utils.mk
> --- ../trunk/package/alsa-utils/alsa-utils.mk   1970-01-01
> 01:00:00.000000000 +0100
> +++ package/alsa-utils/alsa-utils.mk    2007-05-13 10:25:35.339532842
> +0100
> @@ -0,0 +1,114 @@
> +#############################################################
> +#
> +# ALSA Tools

Copy-paste (-:

> +#
> +#############################################################
> +ALSA-UTILS_VERSION=1.0.14rc4
> +ALSA-UTILS_SOURCE=alsa-utils-$(ALSA-UTILS_VERSION).tar.bz2
> +ALSA-UTILS_SITE=ftp://ftp.alsa-project.org/pub/utils/
> +ALSA-UTILS_DIR=$(BUILD_DIR)/${shell basename
> $(ALSA-UTILS_SOURCE) .tar.bz2}
> +ALSA-UTILS_WORKDIR=$(BUILD_DIR)/alsa-utils-$(ALSA-UTILS_VERSION)
> +ALSA-UTILS_CAT:=$(BZCAT)
> +
> +AU_TOOLS :=
> +AU_BIN := 
> +AU_SBIN := 
> +
> +AUTOOLS += aplay/aplay
> +AUTOOLS += aplay/arecord
> +AUTOOLS += alsaconf/alsaconf
> +AUTOOLS += alsactl/alsactl
> +AUTOOLS += alsamixer/alsamixer
> +AUTOOLS += amidi/amidi
> +AUTOOLS += amixer/amixer
> +AUTOOLS += seq/aplaymidi/arecordmidi
> +AUTOOLS += seq/aseqdump/aseqdump
> +AUTOOLS += seq/aseqnet/aseqnet
> +AUTOOLS += iecset/iecset
> +AUTOOLS += speaker-test/speaker-test
> +
> +AUBIN += aconnect
> +AUBIN += alsamixer
> +AUBIN += amidi
> +AUBIN += amixer
> +AUBIN += aplay
> +AUBIN += aplaymidi
> +AUBIN += arecord
> +AUBIN += arecordmidi
> +AUBIN += aseqdump
> +AUBIN += aseqnet
> +AUBIN += iecset
> +AUBIN += speaker-test
> +
> +AUSBIN += alsaconf
> +AUSBIN += alsactl
> +
> +
> +$(DL_DIR)/$(ALSA-UTILS_SOURCE):
> +       $(WGET) -P $(DL_DIR) $(ALSA-UTILS_SITE)/$(ALSA-UTILS_SOURCE)
> +
> +$(ALSA-UTILS_DIR)/.unpacked: $(DL_DIR)/$(ALSA-UTILS_SOURCE)
> +       $(ALSA-UTILS_CAT) $(DL_DIR)/$(ALSA-UTILS_SOURCE) | tar -C
> $(BUILD_DIR) $(TAR_OPTIONS) -
> +       touch $(ALSA-UTILS_DIR)/.unpacked
> +
> +$(ALSA-UTILS_DIR)/.configured: $(ALSA-UTILS_DIR)/.unpacked
> +       (cd $(ALSA-UTILS_DIR); rm -rf config.cache; \
> +               $(TARGET_CONFIGURE_OPTS) \
> +               ./configure \
> +               --target=$(GNU_TARGET_NAME) \
> +               --host=$(GNU_TARGET_NAME) \
> +               --build=$(GNU_HOST_NAME) \
> +               --disable-docs \
> +               --prefix=/usr \
> +               --sysconfdir=/etc \
> +        --disable-alsatest \
> +               $(DISABLE_NLS) \
> +               --exec-prefix=$(STAGING_DIR)/usr/bin \
> +               --libdir=$(STAGING_DIR)/lib \
> +               --includedir=$(STAGING_DIR)/include \
> +               --datadir=$(STAGING_DIR)/usr/share \
> +       );
> +       touch $(ALSA-UTILS_DIR)/.configured
> +
> +$(addprefix $(ALSA-UTILS_WORKDIR)/,$(AUTOOLS)):
> $(ALSA-UTILS_DIR)/.configured
> +       $(MAKE) CC=$(TARGET_CC) -C $(ALSA-UTILS_WORKDIR)
> +
> +
> +$(addprefix $(STAGING_DIR)/,$(addprefix bin/,$(AUBIN)) $(addprefix
> sbin/,$(AUSBIN))): $(addprefix $(ALSA-UTILS_WORKDIR)/,$(AUTOOLS))
> +       $(MAKE) prefix=$(STAGING_DIR) exec_prefix=$(STAGING_DIR) \
> +           bindir=$(STAGING_DIR)/bin sbindir=$(STAGING_DIR)/sbin \
> +           libexecdir=$(STAGING_DIR)/bin datadir=$(STAGING_DIR)/share
> \
> +           sysconfdir=$(STAGING_DIR)/etc sharedstatedir=
> $(STAGING_DIR)/com \
> +           localstatedir=$(STAGING_DIR)/var libdir=$(STAGING_DIR)/lib
> \
> +           includedir=$(STAGING_DIR)/include oldincludedir=
> $(STAGING_DIR)/include \
> +           infodir=$(STAGING_DIR)/info mandir=$(STAGING_DIR)/man \
> +           -C $(ALSA-UTILS_DIR) install;

Use DESTDIR=$(STAGING_DIR) instead.

> +       touch -c $@
> +
> +$(addprefix $(TARGET_DIR)/,$(addprefix usr/bin/,$(AUBIN)) $(addprefix
> usr/sbin/,$(AUSBIN))): $(addprefix $(STAGING_DIR)/,$(addprefix bin/,
> $(AUBIN)) $(addprefix sbin/,$(AUSBIN)))
> +       @mkdir -p $(TARGET_DIR)/usr/bin
> +       @mkdir -p $(TARGET_DIR)/usr/sbin
> +       cp -dpf $(filter $(STAGING_DIR)/bin/%, $^)
> $(TARGET_DIR)/usr/bin/
> +       cp -dpf $(filter $(STAGING_DIR)/sbin/%, $^)
> $(TARGET_DIR)/usr/sbin/
> +       -$(STRIP) --strip-unneeded $@ > /dev/null 2>&1
> +       touch -c $@
> +
> +
> +alsa-utils: uclibc alsa-lib $(addprefix $(TARGET_DIR)/,$(addprefix
> usr/bin/,$(AUBIN)) $(addprefix usr/sbin/,$(AUSBIN)))
> +
> +alsa-utils-clean:
> +       @if [ -d $(ALSA-UTILS_WORKDIR)/Makefile ] ; then \
> +               $(MAKE) -C $(ALSA-UTILS_WORKDIR) clean ; \
> +       fi;
> +
> +alsa-utils-dirclean:
> +       rm -rf $(ALSA-UTILS_DIR) $(ALSA-UTILS_WORKDIR)
> +
> +#############################################################
> +#
> +# Toplevel Makefile options
> +#
> +#############################################################
> +ifeq ($(strip $(BR2_PACKAGE_ALSA-UTILS)),y)
> +TARGETS+=alsa-utils
> +endif
> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
> --exclude=libvorbis ../trunk/package/alsa-utils/Config.in
> package/alsa-utils/Config.in
> --- ../trunk/package/alsa-utils/Config.in       1970-01-01
> 01:00:00.000000000 +0100
> +++ package/alsa-utils/Config.in        2007-05-07 22:13:02.134609691
> +0100
> @@ -0,0 +1,8 @@
> +config BR2_PACKAGE_ALSA-UTILS
> +       bool "alsa-utils"
> +       default n
> +       help
> +         Utilities from Advanced Linux Sound Architecture
> +
> +         http://www.alsa-project.org/
> +
> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
> --exclude=libvorbis ../trunk/package/Config.in package/Config.in
> --- ../trunk/package/Config.in  2007-05-25 17:08:32.583594165 +0100
> +++ package/Config.in   2007-05-25 15:22:03.475398547 +0100
> @@ -239,12 +239,15 @@
>           Support for audio libraries and applications
>  
>  if BR2_AUDIO_SUPPORT
> +source "package/alsa-lib/Config.in"
> +source "package/alsa-tools/Config.in"
> +source "package/alsa-utils/Config.in"
>  source "package/asterisk/Config.in"
>  source "package/libid3tag/Config.in"
>  source "package/libmad/Config.in"
>  source "package/libsndfile/Config.in"
>  source "package/madplay/Config.in"
>  source "package/mpg123/Config.in"
>  endif
>  
>  menuconfig BR2_GRAPHIC_SUPPORT

-- 
With kind regards,

Hans-Christian Egtvedt, siv.ing. (M.Sc.)
Applications Engineer - AVR32 System Solutions - Atmel Norway
-- 
Mvh
Hans-Christian Egtvedt

[Buildroot] ALSA for build root

[Bernhard Fischer]

On Tue, May 29, 2007 at 09:03:12AM +0200, Hans-Christian Egtvedt wrote:
>On Sat, 2007-05-26 at 15:45 +0100, Rod Boyce wrote:
>> Rod Boyce wrote:

Rod, care t osend an updated patch taking Hans-Christian's comments
(thanks for those!) into account?

TIA,

[Buildroot] ALSA for build root

[Bernhard Fischer]

On Tue, May 29, 2007 at 09:03:12AM +0200, Hans-Christian Egtvedt wrote:
>On Sat, 2007-05-26 at 15:45 +0100, Rod Boyce wrote:
>> Rod Boyce wrote:
>
><cut intro about alsa patches>
>
>> As promised attached is a patch to add ALSA packages to build root I 
>> have only tested on i386 but it seems to work for me.  This patch is 
>> against the latest SVN repository No. 18694.  I have not tested the 
>> alsa-tools well and in-fact because I am not using X or GTK most will 
>> not compile I have disabled all the the ac3dec by default.  Also I am 
>> assuming the latest version of ALSA and a compatible kernel.
>
>I am working on the exactly same package (alsa + alsa-util) for
>buildroot, so far testing on an AVR32 platform.
>
>> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
>> --exclude=libvorbis ../trunk/package/alsa-lib/alsa-lib.mk
>> package/alsa-lib/alsa-lib.mk
>> --- ../trunk/package/alsa-lib/alsa-lib.mk       1970-01-01
>> 01:00:00.000000000 +0100
>> +++ package/alsa-lib/alsa-lib.mk        2007-05-26 08:17:42.849975524
>> +0100
>> @@ -0,0 +1,135 @@
>> +#############################################################
>> +#
>> +# ALSA library
>> +#
>> +#############################################################
>> +
>> +ALSA-LIB_VERSION=1.0.14rc4
>> +ALSA-LIB_SOURCE=alsa-lib-$(ALSA-LIB_VERSION).tar.bz2
>> +ALSA-LIB_SITE=ftp://ftp.alsa-project.org/pub/lib/
>> +ALSA-LIB_DIR=$(BUILD_DIR)/${shell basename
>> $(ALSA-LIB_SOURCE) .tar.bz2}

And please don't forget to strip this basename nonsense off.

thanks,

[Buildroot] ALSA for build root

[Rod Boyce]

Bernhard Fischer wrote:
> On Tue, May 29, 2007 at 09:03:12AM +0200, Hans-Christian Egtvedt wrote:
>   
>> On Sat, 2007-05-26 at 15:45 +0100, Rod Boyce wrote:
>>     
>>> Rod Boyce wrote:
>>>       
>> <cut intro about alsa patches>
>>
>>     
>>> As promised attached is a patch to add ALSA packages to build root I 
>>> have only tested on i386 but it seems to work for me.  This patch is 
>>> against the latest SVN repository No. 18694.  I have not tested the 
>>> alsa-tools well and in-fact because I am not using X or GTK most will 
>>> not compile I have disabled all the the ac3dec by default.  Also I am 
>>> assuming the latest version of ALSA and a compatible kernel.
>>>       
>> I am working on the exactly same package (alsa + alsa-util) for
>> buildroot, so far testing on an AVR32 platform.
>>
>>     
>>> diff -Nu --recursive --exclude=.svn --exclude=config --exclude=libogg
>>> --exclude=libvorbis ../trunk/package/alsa-lib/alsa-lib.mk
>>> package/alsa-lib/alsa-lib.mk
>>> --- ../trunk/package/alsa-lib/alsa-lib.mk       1970-01-01
>>> 01:00:00.000000000 +0100
>>> +++ package/alsa-lib/alsa-lib.mk        2007-05-26 08:17:42.849975524
>>> +0100
>>> @@ -0,0 +1,135 @@
>>> +#############################################################
>>> +#
>>> +# ALSA library
>>> +#
>>> +#############################################################
>>> +
>>> +ALSA-LIB_VERSION=1.0.14rc4
>>> +ALSA-LIB_SOURCE=alsa-lib-$(ALSA-LIB_VERSION).tar.bz2
>>> +ALSA-LIB_SITE=ftp://ftp.alsa-project.org/pub/lib/
>>> +ALSA-LIB_DIR=$(BUILD_DIR)/${shell basename
>>> $(ALSA-LIB_SOURCE) .tar.bz2}
>>>       
>
> And please don't forget to strip this basename nonsense off.
>
> thanks,
>
>   

All,

As usually happens when sending this stuff out I did not tidy it as much 
as I would have liked.  I will be happy to make some of the changes.  
But firstly I started out using usr/lib for the libraries but noticed 
that some libraries had been flattened to just /lib so this is what I 
did for the ALSA stuff in the staging dir.

Also a comment about basename nonsense might make sense if you have been 
working with build-root for many months and know every part inside out.  
But I have been using build root at home in my spare time which consists 
of about the last 6-weekends so please explain what you mean by 
basename?  I am happy to comply with a more modern standard if I know 
what it is.  I think I might have used zlib as my starting point when I 
firstly started compiled to get the ALSA lib library integrated into 
build root.  This might explain some of the interesting coding in the patch.

Regards,
Rod Boyce

[Buildroot] mpfr error on cygwin for arm

[Tom]

Hi Bernhard, All

Enclosed is cygwin-1.patch for buildroot-20050524.tar.bz2.

- requires cygwin 1.5.24 with development package

- configure buildroot for arm, arm-generic, eabi, "pc-cygwin", 
uclibc-0.9.29, buildroot toolchain with gcc 4.1.2

- FIRST do one build attempt (will fail, just to get and extract uclibc)

- apply cygwin-1.patch with p1 (it is at very bottom of this email)

- THEN do another build attempt, which will build
   gmp, mpfr, newlib and arm-linux-uclibcgnueabi-gcc-4.1.2.exe
   before failing (*).

Thanks
Tom

PS: if someone could take care of the apple part, that would be great. 
Also, the one fix belongs to uclibc and not to buildroot...


(*) it fails when cross compiling uclibc, maybe my UCLIBC options? Hints 
appreciated, need this to test on target.

make MAKE="make -j1" -C /tw/buildroot/toolchain_build_arm/uClibc-0.9.29
\
                 PREFIX= \
                 DEVEL_PREFIX=/ \
                 RUNTIME_PREFIX=/ \
                 HOSTCC="gcc" \
                 all
arm-linux-uclibcgnueabi-gcc: n: No such file or directory
make[1]: Entering directory
`/tw/buildroot/toolchain_build_arm/uClibc-0.9.29'
arm-linux-uclibcgnueabi-gcc: n: No such file or directory
make[2]: `conf' is up to date.
arm-linux-uclibcgnueabi-gcc: n: No such file or directory
   CC ldso/ldso/ldso.oS
arm-linux-uclibcgnueabi-gcc: n: No such file or directory
make[1]: *** [ldso/ldso/ldso.oS] Error 1
make[1]: Leaving directory
`/tw/buildroot/toolchain_build_arm/uClibc-0.9.29'
make: *** [/tw/buildroot/toolchain_build_arm/uClibc-0.9.29/lib/libc.a]
Error 2





Bernhard Fischer wrote:
> On Fri, May 25, 2007 at 11:52:35PM -0700, Tom wrote:
>> Hi Bernhard and all,
>>
>> sorry if this thread is a little out of sequence, had problems with text 
>> attachment.
>>
>> Bernhard Fischer wrote:
>>> Before you report it to the mpfr folks, please paste the error here,
>>> since i suspect that it could have to do something with the
>>> binary/library name extensions, which (IIRC) we currently do not handle
>>> at all (there's a bug somewhere about this, if memory serves me right).
>>>
>> You were right about extensions, below is the error as produced by
>> buildroot-20050516 built, I just added -verbose for ld.
>> It can not resolve the depency to libgmp.so, because cygwin ld assumes 
>> libs to end in .a, .lib or .dll, and will append them to .so libs, which 
>> thus can never be found.
>> Can I help to add support for this?
> 
> Yes, you can.
> 
> Depending on the target, set
> ifneq $($(findstring linux,$(BR2_GNU_BUILD_SUFFIX)),)
> EXEEXT:=
> LIBEXT:=.a
> SHREXT:=.so
> endif
> ifneq $($(findstring apple,$(BR2_GNU_BUILD_SUFFIX)),)
> EXEEXT:=.pear
> LIBEXT:=.dunno
> SHREXT:=.dylib
> endif
> ifneq $($(findstring cygwin,$(BR2_GNU_BUILD_SUFFIX)),)
> EXEEXT:=.exe
> LIBEXT:=.dunno
> SHREXT:=.dll
> endif
> ifneq $($(findstring mingw,$(BR2_GNU_BUILD_SUFFIX)),)
> EXEEXT:=.exe
> LIBEXT:=.dunno
> SHREXT:=.dll
> endif
> 
> in the toplevel Makefile, before the
> all: world
> target, and use them accordingly.
> 
> 
> While you're at it, it would be awesome if you could deal with this too:
> We want to be able to select which flavour of lib is built (shared or
> static) with a single config-option.
> So, instead of changing all ".so" blindly to $(SHREXT), change it to
> $(LIBTGTEXT) which is set either to
> 
> ifeq $($(BR2_DEFAULT_LIB_TARGET),shared)
> LIBTGTEXT=$(SHREXT)
> else
> LIBTGTEXT=$(LIBEXT)
> endif
> 
> below the newly added block i mentioned above.
> 
> TIA and cheers,
> 


cygwin-1.patch:

--- buildroot.orig/Makefile	2007-05-29 11:55:38.025317600 -0700
+++ buildroot/Makefile	2007-05-29 19:54:26.130869700 -0700
@@ -66,7 +66,32 @@
  #
  #############################################################

+ifneq (,$(findstring linux,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=
+LIBEXT:=.a
+SHREXT:=.so
+endif
+ifneq (,$(findstring apple,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=.pear
+LIBEXT:=.dunno
+SHREXT:=.dylib
+endif
+ifneq (,$(findstring cygwin,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=.exe
+LIBEXT:=.lib
+SHREXT:=.dll
+endif
+ifneq (,$(findstring mingw,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=.exe
+LIBEXT:=.lib
+SHREXT:=.dll
+endif

+ifeq ($(BR2_PREFER_STATIC_LIB),y)
+LIBTGTEXT=$(LIBEXT)
+else
+LIBTGTEXT=$(SHREXT)
+endif

  all:   world

--- buildroot.orig/Config.in	2007-05-29 11:54:55.651131200 -0700
+++ buildroot/Config.in	2007-05-29 19:39:05.976662000 -0700
@@ -373,6 +373,19 @@
  	help
  	  This option hides outdated/obsolete versions of packages.

+config BR2_PREFER_STATIC_LIB
+	bool "prefer static libraries"
+	default n
+	help
+	  Where possible, use static libraries.
+	  This increases your code size a lot and should only be
+	  used with a good reason why not use the default, which
+	  is dynamic libraries.
+
+	  If unsure, say No.
+
+	  WARNING: This is highly experimental at the moment.
+
  endmenu

  source "toolchain/Config.in"
--- buildroot.orig/package/gmp/gmp.mk	2007-05-29 11:56:19.118278600 -0700
+++ buildroot/package/gmp/gmp.mk	2007-05-29 19:34:47.184755900 -0700
@@ -18,6 +18,18 @@
  GMP_BE:=no
  endif

+ifeq ($(BR2_PREFER_STATIC_LIB),y)
+GMP_LIB_FLAGS:=--enable-static --disable-shared
+else
+GMP_LIB_FLAGS:=--disable-static --enable-shared
+endif
+
+ifeq ($(EXEEXT),".exe")
+GMP_CPP_FLAGS:=-DDLL_EXPORT
+else
+GMP_CPP_FLAGS:=-DDLL_EXPORT
+endif
+
  $(DL_DIR)/$(GMP_SOURCE):
  	 $(WGET) -P $(DL_DIR) $(GMP_SITE)/$(GMP_SOURCE)

@@ -35,6 +47,7 @@
  		$(TARGET_CONFIGURE_OPTS) \
  		CFLAGS="$(TARGET_CFLAGS)" \
  		LDFLAGS="$(TARGET_LDFLAGS)" \
+		CPPFLAGS="$(GMP_CPP_FLAGS)" \
  		ac_cv_c_bigendian=$(GMP_BE) \
  		$(GMP_DIR)/configure \
  		--target=$(GNU_TARGET_NAME) \
@@ -52,7 +65,7 @@
  		--includedir=/include \
  		--mandir=/usr/man \
  		--infodir=/usr/info \
-		--enable-shared \
+		$(GMP_LIB_FLAGS) \
  		$(DISABLE_NLS) \
  	);
  	touch $@
@@ -105,12 +118,12 @@
  		CC_FOR_BUILD="$(HOSTCC)" \
  		CC="$(HOSTCC)" \
  		CFLAGS="$(HOST_CFLAGS)" \
+		CPPFLAGS="$(GMP_CPP_FLAGS)" \
  		$(GMP_DIR)/configure \
  		--prefix="$(GMP_HOST_DIR)" \
  		--build=$(GNU_HOST_NAME) \
  		--host=$(GNU_HOST_NAME) \
-		--enable-shared \
-		--enable-static \
+		$(GMP_LIB_FLAGS) \
  		$(DISABLE_NLS) \
  	);
  	touch $@
--- 
buildroot.orig/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile 
2007-04-17 04:38:21.000000000 -0700
+++ buildroot/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile 
2007-05-29 19:34:47.200380600 -0700
@@ -21,7 +21,7 @@
  	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) 
$(HOST_LOADLIBES) $< -o $@

  $(host-cmulti): %: $(host-cobjs) $(host-cshlib)
-	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) 
$(HOST_LOADLIBES) $($@-objs) -o $@
+	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) 
$(HOST_LOADLIBES) $($@-objs) -lintl -o $@

  $(host-cobjs): %.o: %.c
  	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) -c $< -o $@

[Buildroot] mpfr error on cygwin for arm

[Bernhard Fischer]

On Tue, May 29, 2007 at 08:07:48PM -0700, Tom wrote:
>Hi Bernhard, All
>
>Enclosed is cygwin-1.patch for buildroot-20050524.tar.bz2.

Thanks for doing this!

>- requires cygwin 1.5.24 with development package
>
>- configure buildroot for arm, arm-generic, eabi, "pc-cygwin", 
>uclibc-0.9.29, buildroot toolchain with gcc 4.1.2
>
>- FIRST do one build attempt (will fail, just to get and extract uclibc)
>
>- apply cygwin-1.patch with p1 (it is at very bottom of this email)
>
>- THEN do another build attempt, which will build
>  gmp, mpfr, newlib and arm-linux-uclibcgnueabi-gcc-4.1.2.exe
>  before failing (*).
>
>Thanks
>Tom
>
>PS: if someone could take care of the apple part, that would be great. 
>Also, the one fix belongs to uclibc and not to buildroot...
>
>
>(*) it fails when cross compiling uclibc, maybe my UCLIBC options? Hints 
>appreciated, need this to test on target.
>
>make MAKE="make -j1" -C /tw/buildroot/toolchain_build_arm/uClibc-0.9.29
>\
>                PREFIX= \
>                DEVEL_PREFIX=/ \
>                RUNTIME_PREFIX=/ \
>                HOSTCC="gcc" \
>                all
>arm-linux-uclibcgnueabi-gcc: n: No such file or directory
>make[1]: Entering directory
>`/tw/buildroot/toolchain_build_arm/uClibc-0.9.29'
>arm-linux-uclibcgnueabi-gcc: n: No such file or directory
>make[2]: `conf' is up to date.
>arm-linux-uclibcgnueabi-gcc: n: No such file or directory
>  CC ldso/ldso/ldso.oS
>arm-linux-uclibcgnueabi-gcc: n: No such file or directory
>make[1]: *** [ldso/ldso/ldso.oS] Error 1
>make[1]: Leaving directory
>`/tw/buildroot/toolchain_build_arm/uClibc-0.9.29'
>make: *** [/tw/buildroot/toolchain_build_arm/uClibc-0.9.29/lib/libc.a]
>Error 2

[snip]
>cygwin-1.patch:
>
>--- buildroot.orig/Makefile	2007-05-29 11:55:38.025317600 -0700
>+++ buildroot/Makefile	2007-05-29 19:54:26.130869700 -0700
[snip]
> source "toolchain/Config.in"
>--- buildroot.orig/package/gmp/gmp.mk	2007-05-29 11:56:19.118278600 -0700
>+++ buildroot/package/gmp/gmp.mk	2007-05-29 19:34:47.184755900 -0700
>@@ -18,6 +18,18 @@
> GMP_BE:=no
> endif
>
>+ifeq ($(BR2_PREFER_STATIC_LIB),y)
>+GMP_LIB_FLAGS:=--enable-static --disable-shared
>+else
>+GMP_LIB_FLAGS:=--disable-static --enable-shared
>+endif

Better put this into the toolchain/Makefile.in, like:
+ifeq ($(BR2_PREFER_STATIC_LIB),y)
+BR2_SHARED_LIBS:=--enable-static --disable-shared
+else
+BR2_SHARED_LIBS:=--disable-static --enable-shared
+endif
and use ./configure $(BR2_SHARED_LIBS) ...

>+
>+ifeq ($(EXEEXT),".exe")
>+GMP_CPP_FLAGS:=-DDLL_EXPORT
>+else
>+GMP_CPP_FLAGS:=-DDLL_EXPORT
>+endif

This can't be right.
[snip]

>buildroot.orig/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile 
>2007-04-17 04:38:21.000000000 -0700
>+++ buildroot/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile 
>2007-05-29 19:34:47.200380600 -0700
>@@ -21,7 +21,7 @@
> 	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) 
>$(HOST_LOADLIBES) $< -o $@
>
> $(host-cmulti): %: $(host-cobjs) $(host-cshlib)
>-	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) 
>$(HOST_LOADLIBES) $($@-objs) -o $@
>+	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) 
>$(HOST_LOADLIBES) $($@-objs) -lintl -o $@

Shouldn't this libintel rather go into HOST_LOADLIBES for windows?

[Buildroot] mpfr error on cygwin for arm

[Tom]

Hi Bernhard,


will do changes below.

Bernhard Fischer wrote:
 > Better put this into the toolchain/Makefile.in, like:
 > +ifeq ($(BR2_PREFER_STATIC_LIB),y)
 > +BR2_SHARED_LIBS:=--enable-static --disable-shared
 > +else
 > +BR2_SHARED_LIBS:=--disable-static --enable-shared
 > +endif
 > and use ./configure $(BR2_SHARED_LIBS) ...

will try this.

 >
 >> +
 >> +ifeq ($(EXEEXT),".exe")
 >> +GMP_CPP_FLAGS:=-DDLL_EXPORT
 >> +else
 >> +GMP_CPP_FLAGS:=-DDLL_EXPORT
 >> +endif
 >
 > This can't be right.

True. The non-.exe case should be empty, sorry.

 > [snip]
 >
 >> buildroot.orig/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile
 >> 2007-04-17 04:38:21.000000000 -0700
 >> +++ buildroot/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile
 >> 2007-05-29 19:34:47.200380600 -0700
 >> @@ -21,7 +21,7 @@
 >> 	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@)
 >> $(HOST_LOADLIBES) $< -o $@
 >>
 >> $(host-cmulti): %: $(host-cobjs) $(host-cshlib)
 >> -	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@)
 >> $(HOST_LOADLIBES) $($@-objs) -o $@
 >> +	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@)
 >> $(HOST_LOADLIBES) $($@-objs) -lintl -o $@
 >
 > Shouldn't this libintel rather go into HOST_LOADLIBES for windows?

It seems there is an issue with the sequence. (Just moving this -lintl 
before the $($@-objs) caused it to fail.)

[Buildroot] ALSA for build root Part 2

[Rod Boyce]

Bernhard Fischer wrote:
> On Tue, May 29, 2007 at 09:03:12AM +0200, Hans-Christian Egtvedt wrote:
>   
>> On Sat, 2007-05-26 at 15:45 +0100, Rod Boyce wrote:
>>     
>>> Rod Boyce wrote:
>>>       
>
> Rod, care t osend an updated patch taking Hans-Christian's comments
> (thanks for those!) into account?
>
> TIA,
>
>   
All,

Attached is an updated patch that does not offend the sensibilities of 
other ALSA people working on build root.  This is just alsa-lib and 
alsa-utils I will send out a patch for more alsa stuff latter.  This 
outputs to $(STAGING_DIR)/usr/lib which on my build happens to be a link 
back to $(STAGING_DIR)/lib.  I have also removed the device links that I 
was putting in the dev directory I had not removed this since I switched 
over to udev.  Just --prefix=/usr did not fix all paths all of the time 
although I think that I have the minimum custom options for configure 
and install targets for both utilities and library makefile fragments 
now.  I also figured out what the "basename nonsense" was and while I 
approve of the intent I used a more makefile way of doing the same thing.

Hopefully this is a more acceptable patch, if anybody has some more 
suggestions then perhaps they would like to contribute their own ALSA 
patch's so we can either pick the best or merge and get the best of both 
worlds and produce a better build-root system.

Regards,
Rod Boyce
-------------- next part --------------
A non-text attachment was scrubbed...
Name: alsa.patch
Type: text/x-patch
Size: 8564 bytes
Desc: not available
Url : http://busybox.net/lists/buildroot/attachments/20070530/bcc6f0ad/attachment.bin 

[Buildroot] mpfr error on cygwin for arm

[Tom]

Hi Bernhard,

Below is cygwin-2.patch, which takes your recommendations into account, 
proceedings for cygwin-1.patch still apply.

Another issue I ran into now is the following for building the host 
gdb.exe, let me know if this is worth a separate thread:

make[1]: Leaving directory
`/tw/buildroot/toolchain_build_arm/gdbhost-6.6'
strip /tw/buildroot/toolchain_build_arm/gdbhost-6.6/gdb/gdb
strip: /tw/buildroot/toolchain_build_arm/gdbhost-6.6/gdb/gdb: No such 
file or directory
make: *** [/tw/buildroot/toolchain_build_arm/gdbhost-6.6/gdb/gdb] Error 1

The problem is that gdbhost-6.6/gdb/gdb does not exist. 
gdbhost-6.6/gdb/gdb.exe does exist (and works fine) and should not be 
stripped. Where should the cygwin exclusion for this happen?

Thanks
Tom

cygwin-2.patch
--------------
--- buildroot.orig/Makefile	2007-05-24 00:15:29.000000000 -0700
+++ buildroot/Makefile	2007-05-29 19:54:26.130869700 -0700
@@ -66,7 +66,32 @@
  #
  #############################################################

+ifneq (,$(findstring linux,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=
+LIBEXT:=.a
+SHREXT:=.so
+endif
+ifneq (,$(findstring apple,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=.pear
+LIBEXT:=.dunno
+SHREXT:=.dylib
+endif
+ifneq (,$(findstring cygwin,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=.exe
+LIBEXT:=.lib
+SHREXT:=.dll
+endif
+ifneq (,$(findstring mingw,$(BR2_GNU_BUILD_SUFFIX)))
+EXEEXT:=.exe
+LIBEXT:=.lib
+SHREXT:=.dll
+endif

+ifeq ($(BR2_PREFER_STATIC_LIB),y)
+LIBTGTEXT=$(LIBEXT)
+else
+LIBTGTEXT=$(SHREXT)
+endif

  all:   world

--- buildroot.orig/Config.in	2007-05-24 00:15:29.000000000 -0700
+++ buildroot/Config.in	2007-05-29 19:39:05.976662000 -0700
@@ -373,6 +373,19 @@
  	help
  	  This option hides outdated/obsolete versions of packages.

+config BR2_PREFER_STATIC_LIB
+	bool "prefer static libraries"
+	default n
+	help
+	  Where possible, use static libraries.
+	  This increases your code size a lot and should only be
+	  used with a good reason why not use the default, which
+	  is dynamic libraries.
+
+	  If unsure, say No.
+
+	  WARNING: This is highly experimental at the moment.
+
  endmenu

  source "toolchain/Config.in"
--- buildroot.orig/toolchain/Makefile.in	2007-05-24
00:15:08.000000000 -0700
+++ buildroot/toolchain/Makefile.in	2007-05-30 19:30:26.858173300
-0700
@@ -10,6 +10,12 @@
  MULTILIB:=--disable-multilib
  endif

+ifeq ($(BR2_PREFER_STATIC_LIB),y)
+PREFERRED_LIB_FLAGS:=--enable-static --disable-shared
+else
+PREFERRED_LIB_FLAGS:=--disable-static --enable-shared
+endif
+

  # FIXME -- this is temporary
  OPTIMIZE_FOR_CPU=$(ARCH)
---
buildroot.orig/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile
2007-04-17 04:38:21.000000000 -0700
+++ buildroot/toolchain_build_arm/uClibc-0.9.29/extra/config/Makefile
2007-05-29 19:34:47.200380600 -0700
@@ -21,7 +21,7 @@
  	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@)
$(HOST_LOADLIBES) $< -o $@

  $(host-cmulti): %: $(host-cobjs) $(host-cshlib)
-	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@)
$(HOST_LOADLIBES) $($@-objs) -o $@
+	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@)
$(HOST_LOADLIBES) $($@-objs) -lintl -o $@

  $(host-cobjs): %.o: %.c
  	$(HOSTCC) $(HOST_EXTRACFLAGS) $(HOSTCFLAGS) $(HOSTCFLAGS_$@) -c
$< -o $@
--- buildroot.orig/package/gmp/gmp.mk	2007-05-24 00:15:24.000000000
-0700
+++ buildroot/package/gmp/gmp.mk	2007-05-30 20:56:55.313191500
-0700
@@ -18,6 +18,14 @@
  GMP_BE:=no
  endif

+# this is a workaround for a bug in GMP, please see
+# http://gmplib.org/list-archives/gmp-devel/2006-April/000618.html
+ifeq ($(EXEEXT),.exe)
+GMP_CPP_FLAGS:=-DDLL_EXPORT
+else
+GMP_CPP_FLAGS:=
+endif
+
  $(DL_DIR)/$(GMP_SOURCE):
  	 $(WGET) -P $(DL_DIR) $(GMP_SITE)/$(GMP_SOURCE)

@@ -35,6 +43,7 @@
  		$(TARGET_CONFIGURE_OPTS) \
  		CFLAGS="$(TARGET_CFLAGS)" \
  		LDFLAGS="$(TARGET_LDFLAGS)" \
+		CPPFLAGS="$(GMP_CPP_FLAGS)" \
  		ac_cv_c_bigendian=$(GMP_BE) \
  		$(GMP_DIR)/configure \
  		--target=$(GNU_TARGET_NAME) \
@@ -52,7 +61,7 @@
  		--includedir=/include \
  		--mandir=/usr/man \
  		--infodir=/usr/info \
-		--enable-shared \
+		$(PREFERRED_LIB_FLAGS) \
  		$(DISABLE_NLS) \
  	);
  	touch $@
@@ -105,12 +114,12 @@
  		CC_FOR_BUILD="$(HOSTCC)" \
  		CC="$(HOSTCC)" \
  		CFLAGS="$(HOST_CFLAGS)" \
+		CPPFLAGS="$(GMP_CPP_FLAGS)" \
  		$(GMP_DIR)/configure \
  		--prefix="$(GMP_HOST_DIR)" \
  		--build=$(GNU_HOST_NAME) \
  		--host=$(GNU_HOST_NAME) \
-		--enable-shared \
-		--enable-static \
+		$(PREFERRED_LIB_FLAGS) \
  		$(DISABLE_NLS) \
  	);
  	touch $@

[Buildroot] ALSA for build root Part 2

[Hans-Christian Egtvedt]

On Wed, 2007-05-30 at 23:29 +0100, Rod Boyce wrote:
> Bernhard Fischer wrote:
> > On Tue, May 29, 2007 at 09:03:12AM +0200, Hans-Christian Egtvedt wrote:

<cut>

> Attached is an updated patch that does not offend the sensibilities of 
> other ALSA people working on build root.  This is just alsa-lib and 
> alsa-utils I will send out a patch for more alsa stuff latter.  This 
> outputs to $(STAGING_DIR)/usr/lib which on my build happens to be a link 
> back to $(STAGING_DIR)/lib. 

Same here, AFAIK it is supposed to be that way.

> I have also removed the device links that I 
> was putting in the dev directory I had not removed this since I switched 
> over to udev.  Just --prefix=/usr did not fix all paths all of the time 
> although I think that I have the minimum custom options for configure 
> and install targets for both utilities and library makefile fragments 
> now.

Weird, what prefixes are not correct? I only have to give --prefix=/usr
and --sysconfdir=/etc .

> I also figured out what the "basename nonsense" was and while I 
> approve of the intent I used a more makefile way of doing the same thing.
> 
> Hopefully this is a more acceptable patch, if anybody has some more 
> suggestions then perhaps they would like to contribute their own ALSA 
> patch's so we can either pick the best or merge and get the best of both 
> worlds and produce a better build-root system.

Comments below.

> diff -uN --recursive ../trunk/package/alsa-lib/alsa-lib.mk
> package/alsa-lib/alsa-lib.mk
> --- ../trunk/package/alsa-lib/alsa-lib.mk       1970-01-01
> 01:00:00.000000000 +0100
> +++ package/alsa-lib/alsa-lib.mk        2007-05-30 22:58:14.604803783
> +0100
> @@ -0,0 +1,95 @@
> +#############################################################
> +#
> +# ALSA library
> +#
> +#############################################################
> +
> +ALSA-LIB_VERSION := 1.0.14rc4

Same as for alsa-tools, I prefer the ALSA_LIB_VERSION, all underscores.

> +ALSA-LIB_SOURCE := alsa-lib-$(ALSA-LIB_VERSION).tar.bz2
> +ALSA-LIB_SITE := ftp://ftp.alsa-project.org/pub/lib/
> +ALSA-LIB_DIR := $(BUILD_DIR)/$(patsubst %.tar.bz2,%,$(ALSA-LIB_SOURCE))

Again, I prefer the all written approach, to easier see what the
variable is.

> +ALSA-LIB_CAT := $(BZCAT)
> +ALSA-LIB_PATH := /usr/lib

usr/lib, skip leading /

> +ALSA-LIB_DEBUG := yes
> +ifneq ($(BR2_PACKAGE_ALSA-LIB_DEBUG_ENABLE),y)
> +ALSA-LIB_DEBUG := no
> +endif

Nice, but better to set them as "--with-debug" or "--without-debug" and
append them to the ./configure.

> +$(DL_DIR)/$(ALSA-LIB_SOURCE):
> +       $(WGET) -P $(DL_DIR) $(ALSA-LIB_SITE)/$(ALSA-LIB_SOURCE)
> +
> +$(ALSA-LIB_DIR)/.unpacked: $(DL_DIR)/$(ALSA-LIB_SOURCE)
> +       $(ALSA-LIB_CAT) $(DL_DIR)/$(ALSA-LIB_SOURCE) | tar -C
> $(BUILD_DIR) $(TAR_OPTIONS) -
> +       touch $@
> +
> +$(ALSA-LIB_DIR)/.configured: $(ALSA-LIB_DIR)/.unpacked
> +       (cd $(ALSA-LIB_DIR); rm -rf config.cache; \
> +               $(TARGET_CONFIGURE_OPTS) \
> +               ./configure \
> +               --target=$(GNU_TARGET_NAME) \
> +               --host=$(GNU_TARGET_NAME) \
> +               --build=$(GNU_HOST_NAME) \
> +               --prefix=/usr \
> +               --sysconfdir=/etc \
> +               --libdir=$(ALSA-LIB_PATH) \

This should be set with the prefix, not needed.

Add the ALSA-LIB_DEBUG variable.

> +               --disable-docs \
> +               --with-debug=$(ALSA-LIB_DEBUG) \
> +               $(DISABLE_NLS) \
> +       );
> +       touch $@
> +
> +$(ALSA-LIB_DIR)/src/libasound.la: $(ALSA-LIB_DIR)/.configured
> +       rm -f $@
> +       $(MAKE) CFLAGS=-g CC=$(TARGET_CC) -C $(ALSA-LIB_DIR)

Only append -g if you actually are making debug target, see comment on
debug option further up.

> +$(STAGING_DIR)$(ALSA-LIB_PATH)/libasound.so.2.0.0:
> $(ALSA-LIB_DIR)/src/libasound.la
> +       $(MAKE) prefix=$(STAGING_DIR) \
> +           exec_prefix=$(STAGING_DIR) \
> +           libdir=$(STAGING_DIR)$(ALSA-LIB_PATH) \
> +           -C $(ALSA-LIB_DIR) install;
> +       @touch -c $@
> +
> +$(TARGET_DIR)$(ALSA-LIB_PATH)/libasound.so.2.0.0:
> $(STAGING_DIR)$(ALSA-LIB_PATH)/libasound.so.2.0.0
> +       @mkdir -p $(TARGET_DIR)/usr/share/alsa
> +       @mkdir -p $(TARGET_DIR)$(ALSA-LIB_PATH)/alsa-lib/smixer

$(TARGET_DIR)/$(ALSA-LIB_PATH) (add a / in between).

> +       cp -dpf $(STAGING_DIR)/lib/libasound.so* $(TARGET_DIR)/lib

exchange /lib/ with /$(ALSA-LIB_PATH)/ on both sides.

> +       cp -rdpf $(STAGING_DIR)/share/alsa/* $(TARGET_DIR)/usr/share/alsa/
> +       cp -rdpf $(STAGING_DIR)$(ALSA-LIB_PATH)/alsa-lib/* $(TARGET_DIR)$(ALSA-LIB_PATH)/alsa-lib/

$(TARGET_DIR)/$(ALSA-LIB_PATH) (add a / in between).

> +       -$(STRIP) --strip-unneeded $(TARGET_DIR)/lib/libasound.so*

exchange /lib/ with /$(ALSA-LIB_PATH)/

> +       -$(STRIP) --strip-unneeded $(TARGET_DIR)$(ALSA-LIB_PATH)/alsa-lib/smixer/smixer-*.so

$(TARGET_DIR)/$(ALSA-LIB_PATH) (add a / in between).

> +       @touch -c $@
> +
> +
> +$(TARGET_DIR)$(ALSA-LIB_PATH)/libasound.la: $(STAGING_DIR)$(ALSA-LIB_PATH)/libasound.so.2.0.0
> +       @mkdir -p $(TARGET_DIR)/usr/include/alsa/sound
> +       @mkdir -p $(TARGET_DIR)/usr/include/sys
> +       cp -rdpf $(STAGING_DIR)/include/alsa $(TARGET_DIR)/usr/include
> +       cp -dpf $(STAGING_DIR)/include/sys/asoundlib.h $(TARGET_DIR)/usr/include/sys/
> +       cp -dpf $(STAGING_DIR)/lib/libasound.la $(TARGET_DIR)$(ALSA-LIB_PATH)

$(TARGET_DIR)/$(ALSA-LIB_PATH) (add a / in between).

> +       @touch -c $@
> +
> +alsa-lib-headers: $(TARGET_DIR)$(ALSA-LIB_PATH)/libasound.la

$(TARGET_DIR)/$(ALSA-LIB_PATH) (add a / in between).

> +
> +alsa-lib: linux26 uclibc $(TARGET_DIR)$(ALSA-LIB_PATH)/libasound.so.2.0.0
> +
> +alsa-lib-source: $(DL_DIR)/$(ALSA-LIB_SOURCE)
> +
> +alsa-lib-clean:
> +       @if [ -d $(ALSA-LIB_DIR)/Makefile ] ; then \
> +               $(MAKE) -C $(ALSA-LIB_DIR) clean ; \
> +       fi;
> +
> +alsa-lib-dirclean:
> +       rm -rf $(ALSA-LIB_DIR) $(ALSA-LIB_DIR)
> +#############################################################
> +#
> +# Toplevel Makefile options
> +#
> +#############################################################
> +ifeq ($(strip $(BR2_PACKAGE_ALSA-LIB)),y)
> +TARGETS+=alsa-lib
> +endif
> +
> +ifeq ($(strip $(BR2_PACKAGE_ALSA-LIB_TARGET_HEADERS)),y)
> +TARGETS+=alsa-lib-headers
> +endif
> diff -uN --recursive ../trunk/package/alsa-lib/Config.in package/alsa-lib/Config.in
> --- ../trunk/package/alsa-lib/Config.in 1970-01-01 01:00:00.000000000 +0100
> +++ package/alsa-lib/Config.in  2007-05-30 18:58:51.493711408 +0100
> @@ -0,0 +1,23 @@
> +config BR2_PACKAGE_ALSA-LIB
> +       bool "alsa-lib"
> +       default n
> +       help
> +         Advanced Linux Sound Architecture
> +
> +         http://www.alsa-project.org/
> +
> +config BR2_PACKAGE_ALSA-LIB_DEBUG_ENABLE
> +       bool "Build alsa-lib with debug"
> +       default y
> +       depends on BR2_PACKAGE_ALSA-LIB
> +       help
> +         Build libraries with debug turned on see alsa-lib/NOTES
> +
> +config BR2_PACKAGE_ALSA-LIB_TARGET_HEADERS
> +       bool "alsa-lib headers in target"
> +       default n
> +       depends on BR2_PACKAGE_ALSA-LIB
> +       help
> +         Put headers files in the target.

Other than that, alsa-lib looks good.

> diff -uN --recursive ../trunk/package/alsa-utils/alsa-utils.mk
> package/alsa-utils/alsa-utils.mk
> --- ../trunk/package/alsa-utils/alsa-utils.mk   1970-01-01
> 01:00:00.000000000 +0100
> +++ package/alsa-utils/alsa-utils.mk    2007-05-30 22:57:20.186203681
> +0100
> @@ -0,0 +1,105 @@
> +#############################################################
> +#
> +# ALSA Utils
> +#
> +#############################################################
> +ALSA-UTILS_VERSION := 1.0.14rc4

Same comment about underscores in variables.

> +ALSA-UTILS_SOURCE := alsa-utils-$(ALSA-UTILS_VERSION).tar.bz2
> +ALSA-UTILS_SITE := ftp://ftp.alsa-project.org/pub/utils/
> +ALSA-UTILS_DIR := $(BUILD_DIR)/$(patsubst %.tar.bz2,%,$(ALSA-UTILS_SOURCE))

Again, I prefer the all written approach, to easier see what the
variable is.

> +ALSA-UTILS_CAT := $(BZCAT)
> +
> +AUTOOLS :=
> +AUBIN := 
> +AUSBIN := 
> +
> +AUTOOLS += aplay/aplay
> +AUTOOLS += aplay/arecord
> +AUTOOLS += alsaconf/alsaconf
> +AUTOOLS += alsactl/alsactl
> +AUTOOLS += alsamixer/alsamixer
> +AUTOOLS += amidi/amidi
> +AUTOOLS += amixer/amixer
> +AUTOOLS += seq/aplaymidi/arecordmidi
> +AUTOOLS += seq/aseqdump/aseqdump
> +AUTOOLS += seq/aseqnet/aseqnet
> +AUTOOLS += seq/aconnect/aconnect
> +AUTOOLS += iecset/iecset
> +AUTOOLS += speaker-test/speaker-test
> +
> +AUBIN += aconnect
> +AUBIN += alsamixer
> +AUBIN += amidi
> +AUBIN += amixer
> +AUBIN += aplay
> +AUBIN += aplaymidi
> +AUBIN += arecord
> +AUBIN += arecordmidi
> +AUBIN += aseqdump
> +AUBIN += aseqnet
> +AUBIN += iecset
> +AUBIN += speaker-test
> +
> +AUSBIN += alsaconf
> +AUSBIN += alsactl

Make all this configurable in Kconfig, will save some space on targets
where you only need alsactl and a mixer.

Same approach as the tools:

LSA_UTILS_TARGETS_ :=
ALSA_UTILS_TARGETS_y :=

ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ALSACONF)	+= usr/sbin/alsaconf
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ALSACTL)	+= usr/sbin/alsactl
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ALSAMIXER)	+= usr/bin/alsamixer
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_AMIDI)	+= usr/bin/amidi
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_AMIXER)	+= usr/bin/amixer
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_APLAY)	+= usr/bin/aplay
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ARECORD)	+= usr/bin/arecord
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_IECSET)	+= usr/bin/iecset
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ACONNECT)	+= usr/bin/aconnect
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_APLAYMIDI)	+= usr/bin/aplaymidi
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ARECORDMIDI)	+= usr/bin/arecordmidi
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ASEQDUMP)	+= usr/bin/aseqdump
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_ASEQNET)	+= usr/bin/aseqnet
ALSA_UTILS_TARGETS_$(BR2_PACKAGE_ALSA_UTILS_SPEAKER_TEST)	+= usr/bin/speaker-test

> +$(DL_DIR)/$(ALSA-UTILS_SOURCE):
> +       $(WGET) -P $(DL_DIR) $(ALSA-UTILS_SITE)/$(ALSA-UTILS_SOURCE)
> +
> +$(ALSA-UTILS_DIR)/.unpacked: $(DL_DIR)/$(ALSA-UTILS_SOURCE)
> +       $(ALSA-UTILS_CAT) $(DL_DIR)/$(ALSA-UTILS_SOURCE) | tar -C $(BUILD_DIR) $(TAR_OPTIONS) -
> +       touch $(ALSA-UTILS_DIR)/.unpacked
> +
> +$(ALSA-UTILS_DIR)/.configured: $(ALSA-UTILS_DIR)/.unpacked
> +       (cd $(ALSA-UTILS_DIR); rm -rf config.cache; \
> +               $(TARGET_CONFIGURE_OPTS) \
> +               ./configure \
> +               --target=$(GNU_TARGET_NAME) \
> +               --host=$(GNU_TARGET_NAME) \
> +               --build=$(GNU_HOST_NAME) \
> +               --disable-docs \
> +               --prefix=/usr \
> +               --sysconfdir=/etc \
> +        --disable-alsatest \
> +               $(DISABLE_NLS) \
> +       );
> +       touch $@

@ out the touches, I may have overseen some of them.

> +$(ALSA-UTILS_DIR)/aplay/aplay: $(ALSA-UTILS_DIR)/.configured
> +       $(MAKE) CC=$(TARGET_CC) -C $(ALSA-UTILS_DIR)
> +
> +
> +$(STAGING_DIR)/usr/bin/aplay: $(ALSA-UTILS_DIR)/aplay/aplay
> +       $(MAKE) prefix=$(STAGING_DIR)/usr \
> +               exec_prefix=$(STAGING_DIR)/usr \
> +           -C $(ALSA-UTILS_DIR) install;

Use DESTDIR=$(STAGING_DIR) instead.

> +$(TARGET_DIR)/usr/bin/aplay: $(STAGING_DIR)/usr/bin/aplay
> +       @mkdir -p $(TARGET_DIR)/usr/bin
> +       @mkdir -p $(TARGET_DIR)/usr/sbin
> +       cp -dpf $(addprefix $(STAGING_DIR)/usr/bin/,$(AUBIN)) $(TARGET_DIR)/usr/bin/
> +       cp -dpf $(addprefix $(STAGING_DIR)/usr/sbin/,$(AUSBIN)) $(TARGET_DIR)/usr/sbin/
> +       for i in $(addprefix $(STAGING_DIR)/usr/bin/,$(AUBIN)) $(addprefix $(STAGING_DIR)/usr/sbin/,$(AUSBIN)); \
> +       do $(STRIP) --strip-unneeded $$i > /dev/null 2>&1; done

Replace install procedure with something like:
	@mkdir -p $(TARGET_DIR)/usr/bin
	@mkdir -p $(TARGET_DIR)/usr/sbin
	for file in $(ALSA_UTILS_TARGETS_y) ; do \
		cp -dpf $(STAGING_DIR)/$$file $(TARGET_DIR)/$$file; \
		$(STRIP) --strip-unneeded $(STAGING_DIR)/$$file; \
	done
	@if [ -x "$(TARGET_DIR)/usr/bin/speaker-test" ] ; then \
		mkdir -p $(TARGET_DIR)/usr/share/alsa/speaker-test; \
		mkdir -p $(TARGET_DIR)/usr/share/sounds/alsa; \
		cp -rdpf $(STAGING_DIR)/usr/share/alsa/speaker-test/* $(TARGET_DIR)/usr/share/alsa/speaker-test/; \
		cp -rdpf $(STAGING_DIR)/usr/share/sounds/alsa/* $(TARGET_DIR)/usr/share/sounds/alsa/; \
	fi;

> +
> +alsa-utils: uclibc alsa-lib $(TARGET_DIR)/usr/bin/aplay

IMHO alsactl is the "main" util.

> +
> +alsa-utils-clean:
> +       @if [ -d $(ALSA-UTILS_DIR)/Makefile ] ; then \
> +               $(MAKE) -C $(ALSA-UTILS_DIR) clean ; \
> +       fi;
> +
> +alsa-utils-dirclean:
> +       rm -rf $(ALSA-UTILS_DIR) $(ALSA-UTILS_DIR)
> +
> +#############################################################
> +#
> +# Toplevel Makefile options
> +#
> +#############################################################
> +ifeq ($(strip $(BR2_PACKAGE_ALSA-UTILS)),y)
> +TARGETS+=alsa-utils
> +endif
> diff -uN --recursive ../trunk/package/alsa-utils/Config.in package/alsa-utils/Config.in
> --- ../trunk/package/alsa-utils/Config.in       1970-01-01 01:00:00.000000000 +0100
> +++ package/alsa-utils/Config.in        2007-05-07 22:13:02.134609691 +0100
> @@ -0,0 +1,8 @@
> +config BR2_PACKAGE_ALSA-UTILS
> +       bool "alsa-utils"
> +       default n
> +       help
> +         Utilities from Advanced Linux Sound Architecture
> +
> +         http://www.alsa-project.org/

And add the configuration options:
menu "ALSA utils selection"
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ALSACONF
	bool "alsaconf"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ALSACTL
	bool "alsactl"
	default y
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ALSAMIXER
	bool "alsamixer"
	default y
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_AMIDI
	bool "amidi"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_AMIXER
	bool "amixer"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_APLAY
	bool "aplay"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ARECORD
	bool "arecord"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_IECSET
	bool "iecset"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ACONNECT
	bool "aconnect"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_APLAYMIDI
	bool "aplaymidi"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ARECORDMIDI
	bool "arecordmidi"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ASEQDUMP
	bool "aseqdump"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_ASEQNET
	bool "aseqnet"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

config BR2_PACKAGE_ALSA_UTILS_SPEAKER_TEST
	bool "speaker-test"
	default n
	depends on BR2_PACKAGE_ALSA_UTILS

endmenu

> diff -uN --recursive ../trunk/package/Config.in package/Config.in
> --- ../trunk/package/Config.in  2007-05-30 23:10:52.748747759 +0100
> +++ package/Config.in   2007-05-28 18:53:59.880317482 +0100
> @@ -239,12 +239,14 @@
>           Support for audio libraries and applications
>  
>  if BR2_AUDIO_SUPPORT
> +source "package/alsa-lib/Config.in"
> +source "package/alsa-utils/Config.in"
>  source "package/asterisk/Config.in"
>  source "package/libid3tag/Config.in"
>  source "package/libmad/Config.in"
>  source "package/libsndfile/Config.in"
>  source "package/madplay/Config.in"
>  source "package/mpg123/Config.in"
>  endif
>  
>  menuconfig BR2_GRAPHIC_SUPPORT

Again, looks nice, will be good to have full alsa support i Buildroot.

-- 
Mvh
Hans-Christian Egtvedt