patches/gcc/3.3.3/pr15089-fix.patch
changeset 1 eeea35fbf182
     1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/patches/gcc/3.3.3/pr15089-fix.patch	Sat Feb 24 11:00:05 2007 +0000
     1.3 @@ -0,0 +1,25 @@
     1.4 +See http://gcc.gnu.org/PR15089
     1.5 +"In some cases the specified register for a local variable meant to be
     1.6 +used with inline assembly code is not respected.  This breaks code 
     1.7 +relying on this feature to implement non-standard calling convension."
     1.8 +
     1.9 +This fix is probably needed to properly implement linux syscalls in some architectures.
    1.10 +
    1.11 +First added to crosstool by Jamie Hicks, see
    1.12 +http://www.handhelds.org/hypermail/kernel-discuss/current/0066.html
    1.13 +
    1.14 +Index: gcc/loop.c
    1.15 +===================================================================
    1.16 +RCS file: /cvs/gcc/gcc/gcc/loop.c,v
    1.17 +retrieving revision 1.488.2.3
    1.18 +diff -u -r1.488.2.3 loop.c
    1.19 +--- gcc-old/gcc/loop.c	14 Feb 2004 14:46:03 -0000	1.488.2.3
    1.20 ++++ gcc-new/gcc/loop.c	28 Apr 2004 22:02:53 -0000
    1.21 +@@ -929,6 +929,7 @@
    1.22 + 			  || (! (GET_CODE (SET_SRC (set)) == REG
    1.23 + 				 && (REGNO (SET_SRC (set))
    1.24 + 				     < FIRST_PSEUDO_REGISTER))))
    1.25 ++		      && regno >= FIRST_PSEUDO_REGISTER 
    1.26 + 		      /* This test is not redundant; SET_SRC (set) might be
    1.27 + 			 a call-clobbered register and the life of REGNO
    1.28 + 			 might span a call.  */