[uml-devel] uml kbuild cleanup (cross-build) - some fixes and comments

[Blaisorblade <blaisorblade@yahoo.it>]

First, Al, thanks a lot for succeeding in solving this... I had tried a lot of 
time ago to make sure that O= worked... (I think that -j already worked, 
unless somebody re-broke it after I fixed that). Plus you fix a lot of other 
stuff.

There are (at least) two bugs to fix about this patch, plus some comments:

http://user-mode-linux.sourceforge.net/work/current/2.6/2.6.11-rc3-mm1/patches/cross-build

1) typo here: "objtre" should be "objtree"

Index: linux-2.6.10/arch/um/kernel/Makefile
===================================================================

@@ -50,7 +50,7 @@
 quiet_cmd_quote2 = QUOTE   $@
       cmd_quote2 = sed -e '/CONFIG/{'          \
   -e 's/"CONFIG"\;/""/'        \
-  -e 'r $(obj)/config.tmp'     \
+  -e 'r $(objtre)/config.tmp'  \

2) More serious: the value for SETOPTIONS should be the 2.4 one, not the 2.6 
one which is incompatible with 2.4 host kernel (it's an issue for 2.6 headers 
in general).

In fact, a built UML binary must work on top of a 2.4 host kernel, too.

This IMHO would be a general issue about ABI compatibility, except that only 
userspace headers must define the correct value (the old, 2.4-compatible 
one)...

#define PTRACE_OLDSETOPTIONS         21

I just noticed we forgot to merge the patch for this in the current tree, 
which is:

+#ifndef PTRACE_OLDSETOPTIONS
+#define PTRACE_OLDSETOPTIONS PTRACE_SETOPTIONS
+#endif
and then use OLDSETOPTIONS instead of SETOPTIONS.

I'm sending it now for 2.6.11.

Then, some more comments:

3) I don't understand how vmlinux.lds.S is created... it's a symlink, but it's 
created nowhere after the patch - maybe I overlook something, maybe you 
didn't do "make clean" and retest.

4) Since you went with gen-asm-offsets, I think it would be nice (but also 
long, so if I find time I'll work on it) to use directly the generated 
headers, instead of compiling a program which mangles their content and 
reprints the same info in different form.

5) About the USER_OBJS: it was rejected in the current form for going in 
scripts/Makefile long time ago... the proposed alternatives where:

a) rebuild the thing from scratch, without using the obj-[ymn] code (a lot of 
code); or
b) do a) in a way which could also work for klibc (problem: klibc is not 
linked inside the kernel image)

when I have time, I'll try to get people to simply accept a way to specificate 
CFLAGS for some files which totally replace the normal ones.
-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729
http://www.user-mode-linux.org/~blaisorblade




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel