This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Bug #41321

Description:
The file gcc/ada/init.c includes the assembly 
instruction "finit", which initializez the floating point 
unit, for 32-bit builds of NetBSD.  On 64-bit builds, this 
instruciton is omitted and the FPU is uninitialized.  This 
causes several of the acat chapter c4 tests to fail 
including C460007.

Testcases:
C460007

Changelong:
Initialize the FPU for x86_64/amd64 builds.

Bootstrapping and testing:
x86_64-unknown-netbsd5.99.16

Patch is attached to this e-mail and the bugzilla entry.

--- gcc-4.4.1/gcc/ada/init.c	2009-09-09 16:26:14.000000000 +0000
+++ gcc-4.4.1-orig/gcc/ada/init.c	2009-04-09 23:23:07.000000000 +0000
@@ -2116,7 +2116,7 @@
 void
 __gnat_init_float (void)
 {
-#if defined (__i386__) || defined (i386) || defined (__amd64__)
+#if defined (__i386__) || defined (i386)
 
   /* This is used to properly initialize the FPU on an x86 for each
      process thread.  */

Attachment: signature.asc
Description: This is a digitally signed message part.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]