NetCube-Systems-Austria/kumquat-buildroot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
9d8efeafaa
kumquat-buildroot/toolchain/kernel-headers/kernel-headers.mk

159 lines
5.5 KiB
Makefile
Raw Normal View History

Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
#############################################################
#
# Setup the kernel headers. I include a generic package of
# kernel headers here, so you shouldn't need to include your
# own. Be aware these kernel headers _will_ get blown away
# by a 'make clean' so don't put anything sacred in here...
#
#############################################################
Fix some silly bugs: a variable for gcc was used before it was defined, remove '.svn' not 'CVS' from the install, add 2.4.29 kernel headers as the default.
2005-02-11 21:52:06 +01:00
DEFAULT_KERNEL_HEADERS:=$(strip $(subst ",, $(BR2_DEFAULT_KERNEL_HEADERS)))
#"
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
ifneq ($(filter $(TARGETS),kernel-headers),)
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
DEFAULT_KERNEL_HEADERS:=$(strip $(DEFAULT_KERNEL_HEADERS))
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
LINUX_HEADERS_SITE:=127.0.0.1
LINUX_HEADERS_SOURCE:=unspecified-kernel-headers
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
ifeq ("$(strip $(DEFAULT_KERNEL_HEADERS))","2.4.25")
VERSION:=2
PATCHLEVEL:=4
SUBLEVEL:=25
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
LINUX_HEADERS_SITE:=http://www.uclibc.org/downloads/toolchain
LINUX_HEADERS_SOURCE:=linux-libc-headers-2.4.25.tar.bz2
LINUX_HEADERS_UNPACK_DIR:=$(TOOL_BUILD_DIR)/linux-libc-headers-2.4.25
Default to 2.4.27 kernel headers
2004-09-22 19:07:07 +02:00
endif
ifeq ("$(strip $(DEFAULT_KERNEL_HEADERS))","2.4.27")
VERSION:=2
PATCHLEVEL:=4
Fix some silly bugs: a variable for gcc was used before it was defined, remove '.svn' not 'CVS' from the install, add 2.4.29 kernel headers as the default.
2005-02-11 21:52:06 +01:00
SUBLEVEL:=27
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
LINUX_HEADERS_SITE:=http://www.uclibc.org/downloads/toolchain
LINUX_HEADERS_SOURCE:=linux-libc-headers-2.4.27.tar.bz2
LINUX_HEADERS_UNPACK_DIR:=$(TOOL_BUILD_DIR)/linux-libc-headers-2.4.27
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
endif
Revert to using 2.4.x kernel headers by default, due to public outcry....
2004-05-07 08:11:55 +02:00
Fix some silly bugs: a variable for gcc was used before it was defined, remove '.svn' not 'CVS' from the install, add 2.4.29 kernel headers as the default.
2005-02-11 21:52:06 +01:00
ifeq ("$(strip $(DEFAULT_KERNEL_HEADERS))","2.4.29")
VERSION:=2
PATCHLEVEL:=4
SUBLEVEL:=29
LINUX_HEADERS_SITE:=http://www.uclibc.org/downloads/toolchain
LINUX_HEADERS_SOURCE:=linux-libc-headers-2.4.29.tar.bz2
LINUX_HEADERS_UNPACK_DIR:=$(TOOL_BUILD_DIR)/linux-libc-headers-2.4.29
endif
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
ifeq ("$(strip $(DEFAULT_KERNEL_HEADERS))","2.6.9")
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
VERSION:=2
PATCHLEVEL:=6
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
SUBLEVEL:=9
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
LINUX_HEADERS_SITE:=http://ep09.pld-linux.org/~mmazur/linux-libc-headers/
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
LINUX_HEADERS_SOURCE:=linux-libc-headers-2.6.9.1.tar.bz2
LINUX_HEADERS_UNPACK_DIR:=$(TOOL_BUILD_DIR)/linux-libc-headers-2.6.9.1
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
endif
Revert to using 2.4.x kernel headers by default, due to public outcry....
2004-05-07 08:11:55 +02:00
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
ifeq ("$(strip $(DEFAULT_KERNEL_HEADERS))","2.6.10")
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
VERSION:=2
PATCHLEVEL:=6
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
SUBLEVEL:=10
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
LINUX_HEADERS_SITE:=http://ep09.pld-linux.org/~mmazur/linux-libc-headers/
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
LINUX_HEADERS_SOURCE:=linux-libc-headers-2.6.10.0.tar.bz2
LINUX_HEADERS_UNPACK_DIR:=$(TOOL_BUILD_DIR)/linux-libc-headers-2.6.10.0
Update to the latest linux-libc-headers
2004-10-25 23:23:18 +02:00
endif
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
ifeq ("$(strip $(DEFAULT_KERNEL_HEADERS))","2.6.11")
Update to the latest linux-libc-headers
2004-10-25 23:23:18 +02:00
VERSION:=2
PATCHLEVEL:=6
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
SUBLEVEL:=11
Update to the latest linux-libc-headers
2004-10-25 23:23:18 +02:00
LINUX_HEADERS_SITE:=http://ep09.pld-linux.org/~mmazur/linux-libc-headers/
Bump available sanitized kernel headers
2005-03-18 11:23:50 +01:00
LINUX_HEADERS_SOURCE:=linux-libc-headers-2.6.11.0.tar.bz2
LINUX_HEADERS_UNPACK_DIR:=$(TOOL_BUILD_DIR)/linux-libc-headers-2.6.11.0
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
endif
Revert to using 2.4.x kernel headers by default, due to public outcry....
2004-05-07 08:11:55 +02:00
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
LINUX_VERSION:=$(VERSION).$(PATCHLEVEL).$(SUBLEVEL)
Revert to using 2.4.x kernel headers by default, due to public outcry....
2004-05-07 08:11:55 +02:00
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
LINUX_HEADERS_DIR:=$(TOOL_BUILD_DIR)/linux
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
$(DL_DIR)/$(LINUX_HEADERS_SOURCE):
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
mkdir -p $(DL_DIR)
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
$(WGET) -P $(DL_DIR) $(LINUX_HEADERS_SITE)/$(LINUX_HEADERS_SOURCE)
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
$(LINUX_HEADERS_DIR)/.unpacked: $(DL_DIR)/$(LINUX_HEADERS_SOURCE)
If you update the kernel headers tarball and run make against an already built buildroot, the header files will be re-extracted, but they will not be put in exactly the right place if the name of the extracted directory includes the kernel version numbers. For example, linux-libc-headers-2.4.25.tar.bz2 will be extracted to $(TOOLCHAIN_BUILD_DIR)/linux-libc-headers-2.4.25, and then the mv command will move the newly extracted headers into the already existing $(TOOLCHAIN_BUILD_DIR)/linux directory, instead of replacing that directory as desired. The fix is to first remove the $(TOOLCHAIN_BUILD_DIR)/linux directory and then do the mv command. http://bugs.busybox.net/view.php?id=296
2005-06-24 01:25:25 +02:00
rm -rf $(LINUX_HEADERS_DIR)
Erik moved the kernel headers to a different directory, but didn't create the directory first.
2003-11-19 20:29:00 +01:00
mkdir -p $(TOOL_BUILD_DIR)
-xvf -> $(TAR_OPTIONS)
2004-10-11 16:01:07 +02:00
bzcat $(DL_DIR)/$(LINUX_HEADERS_SOURCE) | tar -C $(TOOL_BUILD_DIR) $(TAR_OPTIONS) -
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
ifneq ($(LINUX_HEADERS_UNPACK_DIR),$(LINUX_HEADERS_DIR))
mv $(LINUX_HEADERS_UNPACK_DIR) $(LINUX_HEADERS_DIR)
Revert to using 2.4.x kernel headers by default, due to public outcry....
2004-05-07 08:11:55 +02:00
endif
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
touch $(LINUX_HEADERS_DIR)/.unpacked
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
$(LINUX_HEADERS_DIR)/.patched: $(LINUX_HEADERS_DIR)/.unpacked
be a bit more consistant with naming, always escape wildcards
2005-06-24 10:36:13 +02:00
toolchain/patch-kernel.sh $(LINUX_HEADERS_DIR) toolchain/kernel-headers linux-libc-headers-$(LINUX_VERSION)\*.patch
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
touch $(LINUX_HEADERS_DIR)/.patched
Update to what I'm currently using. Lots of changes... 1) Allow selection of binutils/gcc/gdb/kernel headers to build, although some of the older tool patches probably need updating. 2) Rework gdb build so that remote debugging now works with gdbserver. 3) Misc. other package updates.3) Misc. other package updates.3) Misc. other package updates.
2004-09-03 02:49:43 +02:00
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
$(LINUX_HEADERS_DIR)/.configured: $(LINUX_HEADERS_DIR)/.patched
rm -f $(LINUX_HEADERS_DIR)/include/asm
@if [ ! -f $(LINUX_HEADERS_DIR)/Makefile ] ; then \
Revert to using 2.4.x kernel headers by default, due to public outcry....
2004-05-07 08:11:55 +02:00
echo -e "VERSION = $(VERSION)\nPATCHLEVEL = $(PATCHLEVEL)\n" > \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
$(LINUX_HEADERS_DIR)/Makefile; \
Fixup a stupid bug
2004-06-24 23:44:15 +02:00
echo -e "SUBLEVEL = $(SUBLEVEL)\nEXTRAVERSION =\n" >> \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
$(LINUX_HEADERS_DIR)/Makefile; \
Default to using the 2.6.x based linux-libc-headers -- yet another thing we don't have to maintain. These seem to be working nicely and everything now compiles as expected.
2004-04-21 22:03:39 +02:00
echo -e "KERNELRELEASE=\$$(VERSION).\$$(PATCHLEVEL).\$$(SUBLEVEL)\$$(EXTRAVERSION)" >> \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
$(LINUX_HEADERS_DIR)/Makefile; \
Default to using the 2.6.x based linux-libc-headers -- yet another thing we don't have to maintain. These seem to be working nicely and everything now compiles as expected.
2004-04-21 22:03:39 +02:00
fi;
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
@if [ "$(ARCH)" = "powerpc" ];then \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-ppc$(NOMMU) asm;) \
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
elif [ "$(ARCH)" = "mips" ];then \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-mips$(NOMMU) asm;) \
Use ARCH... not TARGET_ARCH in the mipsel test.
2003-11-25 00:23:16 +01:00
elif [ "$(ARCH)" = "mipsel" ];then \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-mips$(NOMMU) asm;) \
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
elif [ "$(ARCH)" = "arm" ];then \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-arm$(NOMMU) asm; \
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
cd asm; \
if [ ! -L proc ] ; then \
ln -fs proc-armv proc; \
ln -fs arch-ebsa285 arch; fi); \
For now, mimic arm behavior for armeb since that's what uClibc's fix-includes script winds up doing. Fix this later...
2004-09-03 03:24:18 +02:00
elif [ "$(ARCH)" = "armeb" ];then \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-arm$(NOMMU) asm; \
For now, mimic arm behavior for armeb since that's what uClibc's fix-includes script winds up doing. Fix this later...
2004-09-03 03:24:18 +02:00
cd asm; \
if [ ! -L proc ] ; then \
ln -fs proc-armv proc; \
ln -fs arch-ebsa285 arch; fi); \
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
elif [ "$(ARCH)" = "cris" ];then \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-cris asm;) \
Patch from Volmat Alain adding SH3/SH3EB SH4EB support
2004-12-11 10:56:58 +01:00
elif [ "$(ARCH)" = "sh3" ];then \
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-sh asm; \
cd asm; \
ln -s cpu-sh3 cpu) \
elif [ "$(ARCH)" = "sh3eb" ];then \
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-sh asm; \
cd asm; \
ln -s cpu-sh3 cpu) \
make sure sh4 headers get setup properly
2004-10-08 17:28:30 +02:00
elif [ "$(ARCH)" = "sh4" ];then \
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-sh asm; \
cd asm; \
ln -s cpu-sh4 cpu) \
Patch from Volmat Alain adding SH3/SH3EB SH4EB support
2004-12-11 10:56:58 +01:00
elif [ "$(ARCH)" = "sh4eb" ];then \
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-sh asm; \
cd asm; \
ln -s cpu-sh4 cpu) \
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
else \
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
(cd $(LINUX_HEADERS_DIR)/include; ln -fs asm-$(ARCH)$(NOMMU) asm;) \
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
fi
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
touch $(LINUX_HEADERS_DIR)/include/linux/autoconf.h;
touch $(LINUX_HEADERS_DIR)/.configured
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
$(LINUX_KERNEL): $(LINUX_HEADERS_DIR)/.configured
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
kernel-headers: $(LINUX_HEADERS_DIR)/.configured
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Avoid naming conflicts with 'linux' target
2004-10-05 09:16:53 +02:00
kernel-headers-source: $(DL_DIR)/$(LINUX_HEADERS_SOURCE)
Thomas Cameron writes: this patch fixes the source file downloads when executing a "make source", and a few cut-and-paste (tm) errors in the *.mk files. Again, this is a patch against the current CVS sources, and includes my previous patch.
2003-11-12 10:31:12 +01:00
kernel-headers-clean: clean
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
rm -f $(LINUX_KERNEL)
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
rm -rf $(LINUX_HEADERS_DIR)
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
Thomas Cameron writes: this patch fixes the source file downloads when executing a "make source", and a few cut-and-paste (tm) errors in the *.mk files. Again, this is a patch against the current CVS sources, and includes my previous patch.
2003-11-12 10:31:12 +01:00
kernel-headers-dirclean:
Always use kernel-headers.
2004-10-05 09:10:07 +02:00
rm -rf $(LINUX_HEADERS_DIR)
Rework toolchain build to make it much more modular and remove the redundant redundancy.
2003-09-14 13:38:35 +02:00
endif
Reference in New Issue Copy Permalink