Subject: CVS commit: src/gnu/dist/gcc4/gcc/config/m68k
To: None <source-changes@NetBSD.org>
From: matthew green <mrg@netbsd.org>
List: source-changes
Date: 05/12/2006 00:46:33
Module Name:	src
Committed By:	mrg
Date:		Fri May 12 00:46:33 UTC 2006

Modified Files:
	src/gnu/dist/gcc4/gcc/config/m68k: m68k.md

Log Message:
pull across from the GCC3 tree:
>revision 1.2
>date: 2003/10/29 21:25:03;  author: mycroft;  state: Exp;  lines: +40 -11
>As seen on tech-toolchain:
>
>So the problem is thus:
>
>* The {u,}mulsidi3 generate two parallel sets which modify the upper and lower
>halves of the target register.
>
>* life_analysis() does not track subregister modifications -- if you don't
>modify the whole register with a single set, it considers the register
>unused.
>
>The simple, if klugy, solution to this is to stick an explicit clobber in.  It
>seems to work.
>
>While doing this, I noticed that constant folding was not happening for
>32x32->64 multiplies.  This is because the parallel set generated by
>{u,}mulsidi3 cannot be folded at all.  To solve this, I first expand to a
>normal multiply, and then use a define_insn_and_split to convert it to the
>parallel set after CSE and constant folding.
>
>This patch has also been submitted to GCC bugzilla, but who knows if I'll get
>a reply to that.


To generate a diff of this commit:
cvs rdiff -r1.1.1.1 -r1.2 src/gnu/dist/gcc4/gcc/config/m68k/m68k.md

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.