CRIS/CRISv32 port, newlib part (first part)

[Hans-Peter Nilsson]

> Date: Wed, 26 Jan 2005 15:48:43 -0500
> From: Jeff Johnston <jjohnstn@redhat.com>

> > newlib:
> > 2005-01-25  Hans-Peter Nilsson  <hp@axis.com>
> > 
> > 	* configure.host: Add support for cris-*-* and crisv32-*-*.
> > 	* newlib/libc/include/machine/ieeefp.h: Ditto.
> > 	* newlib/libc/include/machine/setjmp.h: Ditto.
> > 	* newlib/libc/machine/cris: New directory.
> > 	* newlib/libc/sys/cris: New directory.
> > 
> 
> You have neglected to list all the files that you added.

Not listing all files but only the directory names, was not done
out of neglection.  This is apparently a per-project policy.
For the projects I've been involved in lately, "New directory"
has been a sufficient ChangeLog entry.  Note taken.

>  I have some issues.

Me too! :-)

> If you need to override sys header files, please note that I had recently added 
> the capability to put sys header files in the machine directory (under a sys 
> subdirectory) to avoid your situation.  Please use that instead and remove the 
> libc/sys/cris directory altogether as well as update your ChangeLog.

But... the way I read the stmp-targ-include rule in
newlib/Makefile.am, this doesn't let me override files in
include/ as sys/cris/include does; only include/machine/, so
this left out the include/pthread.h wrapper file.  A patch to
allow that follows.

When doing that, I noticed that sys/<sysname>/machine/*.h files
override the ones in machine/<cpuname>/*.h and that you can't
override sys/<sysname>/sys/*.h files from
machine/<cpuname>/sys/*.h.  Isn't that a bit backwards?
Shouldn't files in machine trump files in sys?  Consider for
example a future updated machine-independent libc/sys/linux;
machines would presumably need specific include files.  Hm, ok,
then it isn't good enough to put them in
machine/<cpudir>/sys/*.h, because they'd need to be both machine
and sys-specific and the machine subdir is only
machine-specific.  And the sys/linux/include/*.h files would
likely be augmented to include files from machine/.  So, is it
that the intersection of include/sys/*.h names from machine
respectively sys is supposed to be empty?  Anyway, I kept the
sys-trumps-machine order:

2005-01-27  Hans-Peter Nilsson  <hp@axis.com>

	* Makefile.am (stmp-targ-include): Support include header files
	from machine directories.
	* Makefile.in: Regenerate.

Index: Makefile.am
===================================================================
RCS file: /cvs/src/src/newlib/Makefile.am,v
retrieving revision 1.32
diff -c -p -r1.32 Makefile.am
*** Makefile.am	6 Jan 2005 23:31:55 -0000	1.32
--- Makefile.am	27 Jan 2005 03:09:49 -0000
*************** stmp-targ-include: config.status
*** 201,206 ****
--- 201,211 ----
  	      cp $$i targ-include/sys/`basename $$i`; \
  	    else true; fi ; \
  	  done
+ 	-for i in $(srcdir)/libc/machine/$(machine_dir)/include/*.h; do \
+ 	    if [ -f $$i ]; then \
+ 	      cp $$i targ-include/`basename $$i`; \
+ 	    else true; fi ; \
+ 	  done
  	-for i in $(srcdir)/libc/sys/$(sys_dir)/include/*.h; do \
  	    if [ -f $$i ]; then \
  	      cp $$i targ-include/`basename $$i`; \

Do you want me to send the files again or are you ok with me
committing them to the machine/ location with the following
changelog entry and the sys_dir=cris hunk in the previous
configure.host patch removed (after the machine/<cpu>/include
support is resolved)?

2005-01-27  Hans-Peter Nilsson  <hp@axis.com>

	* configure.host: Add support for cris-*-* and crisv32-*-*.
	* libc/include/machine/ieeefp.h: Ditto.
	* libc/include/machine/setjmp.h: Ditto.
	* libc/machine/cris/configure.in, libc/machine/cris/Makefile.am,
	libc/machine/cris/libcdtor.c, libc/machine/cris/setjmp.c,
	libc/machine/cris/memmove.c, libc/machine/cris/memcpy.c,
	libc/machine/cris/memset.c, libc/machine/cris/include/pthread.h,
	libc/machine/cris/sys/signal.h, libc/machine/cris/sys/fcntl.h,
	libc/machine/cris/sys/errno.h, libc/machine/cris/aclocal.m4,
	libc/machine/cris/configure, libc/machine/cris/Makefile.in: New
	files.

By the way, there was a bug in the submitted
libc/machine/cris/memmove.c; more about that next.

brgds, H-P