patches/glibc/2.6/160-use_short_for_fnstsw.patch
author "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
Sun Jul 17 18:01:28 2011 +0200 (2011-07-17)
changeset 2895 86ea5204e9ce
parent 607 aaf7ae5cb11a
permissions -rw-r--r--
libc/newlib: build newlib in the proper step

Currently, newlib is built in the start_file step, which is wrong, but was
needed when the baremetal integration was... well, 'unfinished'.

Now that we build the baremetal compiler from the final cc step, and a
proper core gcc in pass-1 and pass-2, we can move the newlib build to the
step do_libc, where it belongs.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
     1 Original patch from H.J. Lu, as reported there:
     2 http://www.nabble.com/PATCH:-Use-short-for-fnstsw-td14775171.html
     3 
     4 fnstsw stores 16bit into %ax. The upper 16bit of %eax is unchanged.
     5 The new assembler (binutils-2.18.50.0.4 and up) will disallow "fnstsw %eax".
     6 
     7 diff -dur glibc-2.6.1.orig/sysdeps/i386/fpu/ftestexcept.c glibc-2.6.1/sysdeps/i386/fpu/ftestexcept.c
     8 --- glibc-2.6.1.orig/sysdeps/i386/fpu/ftestexcept.c	2004-03-05 11:14:48.000000000 +0100
     9 +++ glibc-2.6.1/sysdeps/i386/fpu/ftestexcept.c	2008-06-24 10:56:14.000000000 +0200
    10 @@ -26,7 +26,7 @@
    11  int
    12  fetestexcept (int excepts)
    13  {
    14 -  int temp;
    15 +  short temp;
    16    int xtemp = 0;
    17  
    18    /* Get current exceptions.  */