include/linux/compile.h is a generated file, only init/Makefile knows 
about it - including it outside of init/* will cause trouble on
parallel builds.

Also, when compile.h already exists when 'make dep' is run, that'll pick
up a dependency on $(TOPDIR)/include/linux/compile.h. So init/Makefile
needs to tell make that this is actually the same file as 
../include/linux.compile.h

into home.transmeta.com:/home/torvalds/v2.5/linux

So I missed at least one file which still relied on implicitly exporting
symbols. Now fixed.

The following allows the ramdisk to work on 2.5.18; maybe we need a
comment in do_open() ?

If d_splice_alias() doesn't find the alias dentry, it returns NULL.
Then, msdos_lookup() dereference the NULL, and Oopses. Fixed here.

The vfat_lookup() part is cleanup only.

FORCE is the de-facto standard name for a prequisite to force
recompilation, so instead of using a mix of 'dummy','FORCE' and
'FORCE_RECOMPILE' use 'FORCE' everywhere.

Also, move figuring out the path relative to the top level dir
into Rules.make, instead of calling an external script. 

It's possible to say "make <subdir>", to descend into that subdir
and recursively build things there. This patch provides this
facility generally without the arch Makefiles needing to duplicate
it for arch/$(ARCH)/somedir.

We now have the information which objects are being built
modular / built-in in Rules.make, so use this information instead
of passing flags to the sub makes.

into tiger.hpl.hp.com:/data1/bk/lia64/to-linus-2.5

ia64_switch_mode().

Didn't get the net/ conversion quite right, and apparently overlooked
one subdir in sound/. Now fixed.

Use the new

	obj-$(CONFIG_FOO) += foo/

syntax to descend into a subdir and link in the result, if
CONFIG_FOO=y.

Build the headers only when we need them during the build, not always
at 'make dep' time.

We use O_TARGET everywhere else, so let's just convert the remaining
users (They get the implicit built-in.o O_TARGET now)

(plus a bit of rearranging in the ISDN Makefiles) 

If a Makefile defines neither O_TARGET nor L_TARGET, let's assume a
default of 'built-in.o'. The goal of this is, of course, to eventually
get rid of O_TARGET completely.

 - move the contents of bdev_get_queue() into do_open() and
   check_partitions() and cache the result in new field of struct
   block_device

 - clean it at the same places that reset ->bd_op, etc.

 - make bdev_get_queue() return cached pointer.

 - kill the last caller of get_hardsect_size() (switching it to
   bdev_hardsect_size()).

 - kill blk_get_queue() and switching to bdev_get_queue()

into home.transmeta.com:/home/torvalds/v2.5/linux

  - make sure return codes are negative (i.e. return -ENOMEM and not return ENOMEM et al)

into home.transmeta.com:/home/torvalds/v2.5/linux

into home.transmeta.com:/home/torvalds/v2.5/linux

Fixes a warning in ieee1394_types.h and an error in include/linux/usb.h

into tp1.ruhr-uni-bochum.de:/home/kai/kernel/v2.5/linux-2.5.misc

This kills unneccessary include from ide-disk.c, kills #ifdef from
reiserfs/journal.c, makes suspend_device local as it should be,
abstains from suspending devices two times in a row (typo), and makes
sure we do not run_task_queue() while we hold spinlock.

I asked Gabor if he'd like me to maintain swsusp, and he liked that
idea [<quote>Would you please take over maintaining? I offered this in
the list a while ago anyway.</quote>].

These are cleanups. They do not change any code, except killing
printk's.