VAX Ultrix bootstrap with gcc-2.96 20000519: genrecog failure

John David Anglin dave@hiauly1.hia.nrc.ca
Wed May 24 10:58:00 GMT 2000


The following bootstrap failure occurs in stage1:

stage1/xgcc -Bstage1/ -B/usr/local/vax-dec-ultrix4.3/bin/  -DIN_GCC    -W -Wall -Wtraditional -pedantic -Wno-long-long -Wwrite-strings -g -O2 -W -Wall  -DHAVE_CONFIG_H  -o genrecog \
 genrecog.o rtl.o bitmap.o ggc-none.o gensupport.o print-rtl.o errors.o ` case "obstack.o" in ?*) echo obstack.o ;; esac ` ` case "" in ?*) echo  ;; esac ` ` case "-lmalloc" in ?*) echo -lmalloc ;; esac `  ` case "" in ?*) echo  ;; esac ` ` case "" in ?*) echo  ;; esac `
./genrecog ../../gcc/config/vax/vax.md > tmp-recog.c
../../gcc/config/vax/vax.md:139: warning: source missing a mode?
../../gcc/config/vax/vax.md:139: warning: destination operand 0 allows non-lvalue
../../gcc/config/vax/vax.md:145: warning: source missing a mode?
../../gcc/config/vax/vax.md:145: warning: destination operand 0 allows non-lvalue
...

This is probably the RELOAD_FOR_OPERAND_ADDRESS bug that I reported previously
for the gcc-2.95 and 2.8.0 branches.  The following fix was applied to the
gcc-2.8.0 tree by Kenner:

Wed Feb 18 07:14:47 1998  Richard Kenner  <kenner@vlsi1.ultra.nyu.edu>

	* reload.c (find_reloads): Always convert address reload for
	non-reloaded operand to RELOAD_FOR_OPERAND_ADDRESS.

--- reload.c.orig	Wed Jul 14 18:29:48 1999
+++ reload.c	Fri Apr  7 15:48:47 2000
@@ -4042,11 +4042,8 @@
 		  = RELOAD_FOR_OPADDR_ADDR;
 	    }
 
-	  if (reload_when_needed[i] == RELOAD_FOR_INPADDR_ADDRESS
-	      || reload_when_needed[i] == RELOAD_FOR_OUTADDR_ADDRESS)
-	    reload_when_needed[i] = RELOAD_FOR_OPADDR_ADDR;
-	  else
-	    reload_when_needed[i] = RELOAD_FOR_OPERAND_ADDRESS;
+
+	  reload_when_needed[i] = RELOAD_FOR_OPERAND_ADDRESS;
 	}
 
       if ((reload_when_needed[i] == RELOAD_FOR_INPUT_ADDRESS

Dave
--
J. David Anglin                                  dave.anglin@nrc.ca
National Research Council of Canada              (613) 990-0752 (FAX: 952-6605)


More information about the Gcc-bugs mailing list