GCC PR target/843

Philip Blundell pb@nexus.co.uk
Fri Apr 5 09:46:00 GMT 2002


Richard, do you think it would be OK to check in your patch for this
problem (reproduced below)?  We have had it in the Debian gcc-3.0
patchset for almost a year now and it doesn't seem to have caused any
untoward side-effects.

p.

Index: arm.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/config/arm/arm.c,v
retrieving revision 1.142
diff -p -r1.142 arm.c
*** arm.c	2001/02/09 02:00:44	1.142
--- arm.c	2001/04/22 14:40:49
*************** symbol_mentioned_p (x)
*** 3418,3423 ****
--- 3414,3424 ----
    if (GET_CODE (x) == SYMBOL_REF)
      return 1;
  
+   /* CONST_DOUBLE can contain a symbol ref for its in-memory
+      representation.  We are not interested about that case here.  */
+   if (GET_CODE (x) == CONST_DOUBLE)
+     return 0;
+ 
    fmt = GET_RTX_FORMAT (GET_CODE (x));
    
    for (i = GET_RTX_LENGTH (GET_CODE (x)) - 1; i >= 0; i--)






More information about the Gcc-bugs mailing list