[PATCH 11/19] kconfig: move .kernelrelease

[PATCH 11/19] kconfig: move .kernelrelease

[Roman Zippel]

This moves the .kernelrelease file into include/config directory.
Remove its generation from the config step, if the config step doesn't
leave a proper .config behind, it triggers a call to silentoldconfig.
Instead its generation can be done via proper dependencies.

Signed-off-by: Roman Zippel <zippel@linux-m68k.org>

---

 Makefile |   21 ++++++++++-----------
 1 file changed, 10 insertions(+), 11 deletions(-)

Index: linux-2.6-git/Makefile
===================================================================
--- linux-2.6-git.orig/Makefile
+++ linux-2.6-git/Makefile
@@ -309,8 +309,8 @@ CFLAGS 		:= -Wall -Wundef -Wstrict-proto
 	  	   -fno-strict-aliasing -fno-common
 AFLAGS		:= -D__ASSEMBLY__
 
-# Read KERNELRELEASE from .kernelrelease (if it exists)
-KERNELRELEASE = $(shell cat .kernelrelease 2> /dev/null)
+# Read KERNELRELEASE from include/config/kernel.release (if it exists)
+KERNELRELEASE = $(shell cat include/config/kernel.release 2> /dev/null)
 KERNELVERSION = $(VERSION).$(PATCHLEVEL).$(SUBLEVEL)$(EXTRAVERSION)
 
 export	VERSION PATCHLEVEL SUBLEVEL KERNELRELEASE KERNELVERSION \
@@ -408,7 +408,6 @@ export KBUILD_DEFCONFIG
 config %config: scripts_basic outputmakefile FORCE
 	$(Q)mkdir -p include/linux include/config
 	$(Q)$(MAKE) $(build)=scripts/kconfig $@
-	$(Q)$(MAKE) -C $(srctree) KBUILD_SRC= .kernelrelease
 
 else
 # ===========================================================================
@@ -716,7 +715,7 @@ $(vmlinux-dirs): prepare scripts
 	$(Q)$(MAKE) $(build)=$@
 
 # Build the kernel release string
-# The KERNELRELEASE is stored in a file named .kernelrelease
+# The KERNELRELEASE is stored in a file named include/config/kernel.release
 # to be used when executing for example make install or make modules_install
 #
 # Take the contents of any files called localversion* and the config
@@ -750,9 +749,9 @@ endif
 
 localver-full = $(localver)$(localver-auto)
 
-# Store (new) KERNELRELASE string in .kernelrelease
+# Store (new) KERNELRELASE string in include/config/kernel.release
 kernelrelease = $(KERNELVERSION)$(localver-full)
-.kernelrelease: FORCE
+include/config/kernel.release: include/config/auto.conf FORCE
 	$(Q)rm -f $@
 	$(Q)echo $(kernelrelease) > $@
 
@@ -773,7 +772,7 @@ PHONY += prepare-all
 # and if so do:
 # 1) Check that make has not been executed in the kernel src $(srctree)
 # 2) Create the include2 directory, used for the second asm symlink
-prepare3: .kernelrelease
+prepare3: include/config/kernel.release
 ifneq ($(KBUILD_SRC),)
 	@echo '  Using $(srctree) as source for kernel'
 	$(Q)if [ -f $(srctree)/.config -o -d $(srctree)/include/config ]; then \
@@ -836,7 +835,7 @@ define filechk_version.h
 	)
 endef
 
-include/linux/version.h: $(srctree)/Makefile .config .kernelrelease FORCE
+include/linux/version.h: $(srctree)/Makefile include/config/kernel.release FORCE
 	$(call filechk,version.h)
 
 # ---------------------------------------------------------------------------
@@ -932,7 +931,7 @@ CLEAN_FILES +=	vmlinux System.map \
 MRPROPER_DIRS  += include/config include2
 MRPROPER_FILES += .config .config.old include/asm .version .old_version \
                   include/linux/autoconf.h include/linux/version.h \
-		  .kernelrelease Module.symvers tags TAGS cscope*
+		  Module.symvers tags TAGS cscope*
 
 # clean - Delete most, but leave enough to build external modules
 #
@@ -1256,8 +1255,8 @@ checkstack:
 	$(PERL) $(src)/scripts/checkstack.pl $(ARCH)
 
 kernelrelease:
-	$(if $(wildcard .kernelrelease), $(Q)echo $(KERNELRELEASE), \
-	$(error kernelrelease not valid - run 'make *config' to update it))
+	$(if $(wildcard include/config/kernel.release), $(Q)echo $(KERNELRELEASE), \
+	$(error kernelrelease not valid - run 'make prepare' to update it))
 kernelversion:
 	@echo $(KERNELVERSION)
 

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Andrew Morton]

Roman Zippel <zippel@linux-m68k.org> wrote:
>
> This moves the .kernelrelease file into include/config directory.
>  Remove its generation from the config step, if the config step doesn't
>  leave a proper .config behind, it triggers a call to silentoldconfig.
>  Instead its generation can be done via proper dependencies.

Well that was a pita.  I was using that file in my kernel installation
script.

Your changelog says what the patch does, but gives no indication of why it
did it.

What do we get back for the breakage which this will cause?

Now I'm going to have to look for both .kernelrelease and
include/config/kernel.release and work out which one has the more recent
mtime.  grr.

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Andrew Morton]

Sam Ravnborg <sam@ravnborg.org> wrote:
>
> On Mon, Apr 10, 2006 at 01:57:27AM -0700, Andrew Morton wrote:
> > Roman Zippel <zippel@linux-m68k.org> wrote:
> > >
> > > This moves the .kernelrelease file into include/config directory.
> > >  Remove its generation from the config step, if the config step doesn't
> > >  leave a proper .config behind, it triggers a call to silentoldconfig.
> > >  Instead its generation can be done via proper dependencies.
> > 
> > Well that was a pita.  I was using that file in my kernel installation
> > script.
> > 
> > Your changelog says what the patch does, but gives no indication of why it
> > did it.
> > 
> > What do we get back for the breakage which this will cause?
> > 
> > Now I'm going to have to look for both .kernelrelease and
> > include/config/kernel.release and work out which one has the more recent
> > mtime.  grr.
> That you have for not using 'make kernelrelease' - he ;-)
> Did you not know, or did make kernelrelease not do what you expected?
> 

I stopped using `make kernelrelease' when it did something bad when used
from another machine across NFS.

<tries it>

hm, it takes nearly five seconds, but it wasn't that - something actually
broke.  But I forget what it was.  I'll put it back and will wait for it
to reoccur.

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Andrew Morton]

Andrew Morton <akpm@osdl.org> wrote:
>
> I stopped using `make kernelrelease' when it did something bad when used
>  from another machine across NFS.
> 
>  <tries it>
> 
>  hm, it takes nearly five seconds, but it wasn't that - something actually
>  broke.  But I forget what it was.  I'll put it back and will wait for it
>  to reoccur.
> 

Actually, I think it was a problem interacting with the weird things which
`git bisect' does with .kernelrelease.

bix:/usr/src/git26> git bisect start
bix:/usr/src/git26> git bisect bad v2.6.15
bix:/usr/src/git26> git bisect good v2.6.14
Bisecting:    2705 revisions left to test after this
bix:/usr/src/git26> cat .kernelrelease 
2.6.16-rc2-g6bd0e10e
bix:/usr/src/git26> make kernelrelease
2.6.14-gref: ref

That's cute, but it wasn't that..

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Sam Ravnborg]

On Mon, Apr 10, 2006 at 01:57:27AM -0700, Andrew Morton wrote:
> Roman Zippel <zippel@linux-m68k.org> wrote:
> >
> > This moves the .kernelrelease file into include/config directory.
> >  Remove its generation from the config step, if the config step doesn't
> >  leave a proper .config behind, it triggers a call to silentoldconfig.
> >  Instead its generation can be done via proper dependencies.
> 
> Well that was a pita.  I was using that file in my kernel installation
> script.
> 
> Your changelog says what the patch does, but gives no indication of why it
> did it.
> 
> What do we get back for the breakage which this will cause?
> 
> Now I'm going to have to look for both .kernelrelease and
> include/config/kernel.release and work out which one has the more recent
> mtime.  grr.
That you have for not using 'make kernelrelease' - he ;-)
Did you not know, or did make kernelrelease not do what you expected?

	Sam

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Roman Zippel]

Hi,

On Mon, 10 Apr 2006, Andrew Morton wrote:

> hm, it takes nearly five seconds, but it wasn't that - something actually
> broke.  But I forget what it was.  I'll put it back and will wait for it
> to reoccur.

The patch below should speed this up. You know that you have to update 
this file explicitly?

bye, Roman

---

 Makefile |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Index: linux-2.6-git/Makefile
===================================================================
--- linux-2.6-git.orig/Makefile
+++ linux-2.6-git/Makefile
@@ -366,7 +366,8 @@ outputmakefile:
 # of make so .config is not included in this case either (for *config).
 
 no-dot-config-targets := clean mrproper distclean \
-			 cscope TAGS tags help %docs check%
+			 cscope TAGS tags help %docs check% \
+			 kernelrelease kernelversion
 
 config-targets := 0
 mixed-targets  := 0
@@ -1251,7 +1252,7 @@ namespacecheck:
 endif #ifeq ($(config-targets),1)
 endif #ifeq ($(mixed-targets),1)
 
-PHONY += checkstack
+PHONY += checkstack kernelrelease kernelversion
 checkstack:
 	$(OBJDUMP) -d vmlinux $$(find . -name '*.ko') | \
 	$(PERL) $(src)/scripts/checkstack.pl $(ARCH)

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Andrew Morton]

Roman Zippel <zippel@linux-m68k.org> wrote:
>
> On Mon, 10 Apr 2006, Andrew Morton wrote:
> 
>  > hm, it takes nearly five seconds, but it wasn't that - something actually
>  > broke.  But I forget what it was.  I'll put it back and will wait for it
>  > to reoccur.
> 
>  The patch below should speed this up.

It went from 5 seconds down to 4 seconds.

> You know that you have to update 
>  this file explicitly?

That depends on what "explicitly" means.   `make oldconfig' updates it.

Re: [PATCH 11/19] kconfig: move .kernelrelease

[Roman Zippel]

Hi,

On Mon, 10 Apr 2006, Andrew Morton wrote:

> >  The patch below should speed this up.
> 
> It went from 5 seconds down to 4 seconds.

It shouldn't do much besides printing the file content.
An strace -ftt might help to find what it's wasting it time with.

> > You know that you have to update 
> >  this file explicitly?
> 
> That depends on what "explicitly" means.   `make oldconfig' updates it.

I had to change this, it's now "make prepare".
Putting it at the end of *config was an unfortunate choice, as it was 
updated independent of whether the config step left a valid config or not 
and with the fixed dependencies it even may have triggered a unexpected 
call to silentoldconfig (e.g. just by quitting xconfig/menuconfig).

bye, Roman