sequencer module problems
First of all, my make is 3.77 and lacks support for the $(if CONDITION,THEN-PART[,ELSE-PART])
function. So I tried to replace it by some sort of equivalent selected by existing conditionals.
As make does not support the syntax below, it ignore it.
in alsa-driver/alsa-kernel/core/seq:
#
# this function returns: # "m" - CONFIG_SND_SEQUENCER is m # <empty string> - CONFIG_SND_SEQUENCER is undefined # otherwise parameter #1 value #
sequencer = $(if $(subst y,,$(CONFIG_SND_SEQUENCER)),$(if $(1),m),$(if $(CONFIG_SND_SEQUENCER),$(1)))
After reading some docs, I understand the following:CONFIG_SND_SEQUENCER=m (actually set to anything but "y") -> if $(1) is defined -> sequencer=mCONFIG_SND_SEQUENCER=y -> sequencer=$(1), since the subsequent test will always see it defined.CONFIG_SND_SEQUENCER undefined -> sequencer =
I checked the configure script and CONFIG_SND_SEQUENCER is set to "m" if --with-sequencer=yes or is left undefined otherwise. So any other value for CONFIG_SND_SEQUENCER is not possible.
I checked that Makefile again and I see that all calls to the function sequencer add $(RAWMIDI_OBJS), only some add $(OPL3_OBJS) and some add individual modules.
Then I believe the following could be done:
-obj-$(CONFIG_SND_SEQUENCER) += snd-seq.o snd-seq-device.o +obj-$(CONFIG_SND_SEQUENCER) += snd-seq.o snd-seq-device.o $(RAWMIDI_OBJS)
Then,
# Toplevel Module Dependency
obj-$(CONFIG_SND_VIRMIDI) += snd-seq-virmidi.o snd-seq-midi-event.o ifeq ($(CONFIG_SND_SEQUENCER),m) obj-$($(CONFIG_SND_SERIAL_U16550)) += obj-$($(CONFIG_SND_MTPAV)) +=
obj-$($(CONFIG_SND_MPU401)) +=
obj-$($(CONFIG_SND_ALS100)) += $(OPL3_OBJS) obj-$($(CONFIG_SND_AZT2320)) += $(OPL3_OBJS) obj-$($(CONFIG_SND_DT019X)) += $(OPL3_OBJS)
[snip]
obj-$($(CONFIG_SND_CS46XX)) += obj-$($(CONFIG_SND_EMU10K1)) += snd-seq-midi-emul.o snd-seq-virmidi.o obj-$($(CONFIG_SND_TRIDENT)) += snd-seq-midi-emul.o snd-seq-instr.o obj-$($(CONFIG_SND_YMFPCI)) += $(OPL3_OBJS) obj-$($(CONFIG_SND_USB_AUDIO)) += obj-$($(CONFIG_SND_HDSP)) +=
endif
obj-m := $(sort $(obj-m))
Then, the sequencer function above is no longer needed.
Signature · 11 lines
ing. Andrei Boros mailto:[email protected] / +40-21-303-1870 Centrul pt. Tehnologia Informatiei Societatea Romana de Radiodifuziune ------------------------------------------------------- This SF.net email is sponsored by: Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for? http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
At Wed, 19 Mar 2003 12:45:16 +0200,
Quoted 62 lines · Andrei Boros
First of all, my make is 3.77 and lacks support for the $(if CONDITION,THEN-PART[,ELSE-PART])
function. So I tried to replace it by some sort of equivalent selected by existing conditionals.As make does not support the syntax below, it ignore it.
in alsa-driver/alsa-kernel/core/seq:
## this function returns: # "m" - CONFIG_SND_SEQUENCER is m # <empty string> - CONFIG_SND_SEQUENCER is undefined # otherwise parameter #1 value #sequencer = $(if $(subst y,,$(CONFIG_SND_SEQUENCER)),$(if $(1),m),$(if $(CONFIG_SND_SEQUENCER),$(1)))
After reading some docs, I understand the following:
CONFIG_SND_SEQUENCER=m(actually set to anything but "y") -> if $(1) is defined -> sequencer=mCONFIG_SND_SEQUENCER=y-> sequencer=$(1), since the subsequent test will always see it defined.CONFIG_SND_SEQUENCERundefined -> sequencer =I checked the configure script and
CONFIG_SND_SEQUENCERis set to "m" if --with-sequencer=yes or is left undefined otherwise. So any other value forCONFIG_SND_SEQUENCERis not possible.I checked that
Makefileagain and I see that all calls to the function sequencer add $(RAWMIDI_OBJS), only some add $(OPL3_OBJS) and some add individual modules.Then I believe the following could be done:
(quoted diff) +1
-1-obj-$(CONFIG_SND_SEQUENCER) += snd-seq.o snd-seq-device.o +obj-$(CONFIG_SND_SEQUENCER) += snd-seq.o snd-seq-device.o$(RAWMIDI_OBJS)
Then,
# Toplevel Module Dependency
obj-$(CONFIG_SND_VIRMIDI) += snd-seq-virmidi.o snd-seq-midi-event.o ifeq ($(CONFIG_SND_SEQUENCER),m) obj-$($(CONFIG_SND_SERIAL_U16550)) += obj-$($(CONFIG_SND_MTPAV)) +=
obj-$($(CONFIG_SND_MPU401)) +=
obj-$($(CONFIG_SND_ALS100)) += $(OPL3_OBJS) obj-$($(CONFIG_SND_AZT2320)) += $(OPL3_OBJS) obj-$($(CONFIG_SND_DT019X)) += $(OPL3_OBJS)
[snip]
obj-$($(CONFIG_SND_CS46XX)) += obj-$($(CONFIG_SND_EMU10K1)) += snd-seq-midi-emul.o snd-seq-virmidi.o obj-$($(CONFIG_SND_TRIDENT)) += snd-seq-midi-emul.o snd-seq-instr.o obj-$($(CONFIG_SND_YMFPCI)) += $(OPL3_OBJS) obj-$($(CONFIG_SND_USB_AUDIO)) += obj-$($(CONFIG_SND_HDSP)) +=
endif
Quoted 3 lines
obj-m := $(sort $(obj-m))
Then, the sequencer function above is no longer needed.
well, some points noted.
1. CONFIG_SND_SEQUENCER is set to m only in the alsa-driver configure because alsa-driver package provides only the modules. you change above doesn't handle the case for built-in drivers.
2. rawmidi is not used in every system. this should belong to the card dependent defines.
3. the use of sequencer macro is intended for removal of if condition in the definitions of modules. so, going back to the style shown above is not acceptable for 2.5 kernel tree.
(2.5 kernel tree needs anyway the newer verision of make which can
handle the macro correctly.)
a workaround would be to rewrite the Makefile in alsa-driver/core/seq without reading the alsa-kernel Makefile.
but from the maintenace perspective, it's also not nice...
ciao,
Takashi
-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
At Wed, 19 Mar 2003 14:33:44 +0200,
Quoted 7 lines · Andrei Boros
Quoted 2 lines · Takashi Iwai
well, some points noted.
I am still searching for an acceptable way to have this compatible with make-3.77 without breaking anything.
Quoted 6 lines
Quoted 3 lines
1.
CONFIG_SND_SEQUENCERis set to m only in the alsa-driver configure because alsa-driver package provides only the modules. you change above doesn't handle the case for built-in drivers.Then the above may go
ifdef $(CONFIG_SND_SEQUENCER) instead of ifeq (...)
unfortunately, it's still not enough.
if sequencer is m, then the related modules have to be m, too, even if the card driver is y.
Quoted 6 lines
Quoted 5 lines
3. the use of sequencer macro is intended for removal of if condition in the definitions of modules. so, going back to the style shown above is not acceptable for 2.5 kernel tree.
(2.5 kernel tree needs anyway the newer verision of make which can
handle the macro correctly.)
Quoted 7 lines
Quoted 4 lines
a workaround would be to rewrite the
Makefilein alsa-driver/core/seq without reading the alsa-kernelMakefile.
but from the maintenace perspective, it's also not nice...Is there any way to do it and still be compatible?
I mean since the alsa package as a separate bundle is still available for 2.2 kernel, right?
yes, but it's rather a problem of make, not the kernel code.
Takashi
-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
Typo in configure.in : dummy sequencer
In reply to Takashi Iwaimake: Nothing to be done for `all-deps'.
autoconf: Undefined macros:
configure.in:797: AC_DEFINE_CONFIG_SND_SEQ_DUMMY)
Should read : AC_DEFINE(CONFIG_SND_SEQ_DUMMY)
Signature · 11 lines
ing. Andrei Boros mailto:[email protected] / +40-21-303-1870 Centrul pt. Tehnologia Informatiei Societatea Romana de Radiodifuziune ------------------------------------------------------- This SF.net email is sponsored by: Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for? http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
alsa-driver, make 3.79, cvscompile
In reply to Takashi IwaiI got make 3.79 installed, hopefully the sequencer module problems will be solved, but I get a tiny error that didn't appear before when runing cvscompile :
I updated the cvs tree, did a make clean, deleted Makefile, then ran ./cvscompile.
This is the first line it outputs:
make: *** No rule to make target `all-deps'. Stop.
Then the configure script starts and runs fine apparently.
Signature · 11 lines
ing. Andrei Boros mailto:[email protected] / +40-21-303-1870 Centrul pt. Tehnologia Informatiei Societatea Romana de Radiodifuziune ------------------------------------------------------- This SF.net email is sponsored by: Tablet PC. Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for? http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
Re: Typo in configure.in : dummy sequencer
In reply to Andrei BorosQuoted 5 lines · Andrei Boros
make: Nothing to be done for `all-deps'.
autoconf: Undefined macros:
configure.in:797:AC_DEFINE_CONFIG_SND_SEQ_DUMMY)Should read :
AC_DEFINE(CONFIG_SND_SEQ_DUMMY)
I have this line in my tree:
AC_DEFINE(CONFIG_SND_SEQ_DUMMY_MODULE)
Jaroslav
-----
Jaroslav Kysela <[email protected]>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs
-------------------------------------------------------
This SF.net email is sponsored by: Tablet PC.
Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
Re: dummy sequencer module problems
In reply to Takashi Iwai<<<<<<< Makefile.conf.inCONFIG_SND_SEQ_DUMMY = @CONFIG_SND_SEQ_DUMMY@
=======CONFIG_SND_SEQ_DUMMY = @CONFIG_SND_SEQUENCER@
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
1.52
I got again this inconsistency in Makefile.conf.in when updateing the cvs tree.
I am wandering, this looks to me that snd-seq-dummy.o will be built always if sequencer is built, regardless of the fact that the dummy soundcard driver is built or not. I checked the Makefile in alsa-kernel/core/seq.
Signature · 11 lines
ing. Andrei Boros mailto:[email protected] / +40-21-303-1870 Centrul pt. Tehnologia Informatiei Societatea Romana de Radiodifuziune ------------------------------------------------------- This SF.net email is sponsored by: Tablet PC. Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for? http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
Re: dummy sequencer module problems
At Thu, 20 Mar 2003 10:28:57 +0200,
Quoted 7 lines · Andrei Boros
<<<<<<< Makefile.conf.in
CONFIG_SND_SEQ_DUMMY= @CONFIG_SND_SEQ_DUMMY@
=======CONFIG_SND_SEQ_DUMMY= @CONFIG_SND_SEQUENCER@Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
1.52
Quoted 7 lines
I got again this inconsistency in Makefile.conf.in when updateing the cvs tree.
I am wandering, this looks to me that snd-seq-dummy.o will be built always if sequencer is built, regardless of the fact that the dummy soundcard driver is built or not. I checked the
Makefilein alsa-kernel/core/seq.
yes, we don't provide the configure option for this yet, so i just wrote like the above. (i.e. seq-dummy is provided always if sequencer is enabled).
note that snd-seq-dummy is completely different thing from snd-dummy. please don't make them together.
Takashi
-------------------------------------------------------
This SF.net email is sponsored by: Tablet PC.
Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
Re: alsa-driver, make 3.79, cvscompile
In reply to Andrei BorosAt Thu, 20 Mar 2003 10:20:53 +0200,
Quoted 7 lines · Andrei Boros
I got make 3.79 installed, hopefully the sequencer module problems will be solved, but I get a tiny error that didn't appear before when runing cvscompile :
I updated the cvs tree, did a make clean, deleted
Makefile, then ran
^^^^^^^^^^^^^^^^
?? you'll need it for cvscompile, too. Makefile itself won't generated by configure (but Makefile.conf will be).
Takashi
-------------------------------------------------------
This SF.net email is sponsored by: Tablet PC.
Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
Re: alsa-driver, make 3.79, cvscompile
In reply to a message not in the archiveAt Thu, 20 Mar 2003 12:49:26 +0200,
Quoted 7 lines · Andrei Boros
Quoted 5 lines · Takashi Iwai
Quoted 5 lines
I got make 3.79 installed, hopefully the sequencer module problems will be solved, but I get a tiny error that didn't appear before when runing cvscompile :
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
I updated the cvs tree, did a make clean, deleted
Makefile, then ran^^^^^^^^^^^^^^^^?? you'll need it for cvscompile, too.
Makefileitself won't generated by configure (but Makefile.conf will be).I corrected that.
It still wants to make all-deps before anything else, when I run cvscompile. Dunno it this is intended.
this is intentional behavior.
"make all-deps" will create the dependent files for configure script such as toplevel.config.in.
Takashi
-------------------------------------------------------
This SF.net email is sponsored by: Tablet PC.
Does your code think in ink? You could win a Tablet PC. Get a free Tablet PC hat just for playing. What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en