Re: kernel building procedure changed?

[Cyrill Gorcunov <gorcunov@gmail.com>]

[Sam Ravnborg - Mon, Feb 23, 2009 at 12:00:31AM +0100]
| On Sun, Feb 22, 2009 at 11:27:51PM +0100, Jiri Slaby wrote:
| > On 22.2.2009 23:16, Sam Ravnborg wrote:
| > >On Mon, Feb 23, 2009 at 01:07:48AM +0300, Cyrill Gorcunov wrote:
| > >>Hi all,
| > >>
| > >>I found a bit strange thing happens now --
| > >>plain
| > >>
| > >>	make arch/x86/kernel/head64.o
| > >>
| > >>doesn't work anymore. Instead of compiling
| > >>_just_ this file I've got the following:
| > >>
| > >>---
| > >>cyrill@lenovo linux-2.6.git $ make arch/x86/kernel/head64.o
| > >>   CHK     include/linux/version.h
| > >>   CHK     include/linux/utsrelease.h
| > >>   SYMLINK include/asm ->  include/asm-x86
| > >>   CALL    scripts/checksyscalls.sh
| > >>   CHK     include/linux/compile.h
| > >>   AS      arch/x86/kernel/head_64.o
| > >>   VDSOSYM arch/x86/vdso/vdso-syms.lds
| > >>   VDSOSYM arch/x86/vdso/vdso32-int80-syms.lds
| > >>   VDSOSYM arch/x86/vdso/vdso32-syscall-syms.lds
| > >>   VDSOSYM arch/x86/vdso/vdso32-sysenter-syms.lds
| > >>   VDSOSYM arch/x86/vdso/vdso32-syms.lds
| > >>   LD      arch/x86/vdso/built-in.o
| > >>   CC      fs/hpfs/dnode.o
| > >>make[2]: *** [fs/hpfs/dnode.o] Interrupt
| > >>make[1]: *** [fs/hpfs] Interrupt
| > >>make: *** [fs] Interrupt
| > >>---
| > >>
| > >>I've done make allyesconfig previously.
| > >>I don't understand what I'm doing wrong... :(
| > >
| > >I should work...
| > 
| > It is matter of days, new mmotm doesn't work either, older did.
| 
| Does it happen with -next?
| Does it happen with mainline?
| I need to know what tree to look at - and I do not have mmotm handy.
| 
| 	Sam
| 

Sam, it seems I never paid attention on Makefile
rules before. Which means -- I don't know if kernel
build procedure should behave like it do now.

We have the following rule (arch/x86/kernel/Makefile)

extra-y                := head_$(BITS).o head$(BITS).o head.o init_task.o
vmlinux.lds

Does it mean that any obj-y targets should be built before
extra-y? If "yes" -- all is fine and works well and I'm just
noising here (sorry for that).

To be precise -- I don't remember how it worked before :)
I mean it's possible that my previous workflow was:

1) build whole kernel by "make"
2) modify head_64.S
3) run "make arch/x86/kernel/head_64.o"

So I will try to repeat these 3 points and report.

	- Cyrill -