[SPARC64]: Fix quad-float multiply emulation.
authorDavid S. Miller <davem@sunset.davemloft.net>
Thu, 27 Jul 2006 23:49:21 +0000 (16:49 -0700)
committerDavid S. Miller <davem@sunset.davemloft.net>
Thu, 27 Jul 2006 23:49:21 +0000 (16:49 -0700)
Something is wrong with the 3-multiply (vs. 4-multiply) optimized
version of _FP_MUL_MEAT_2_*(), so just use the slower version
which actually computes correct values.

Noticed by Rene Rebe

Signed-off-by: David S. Miller <davem@davemloft.net>
include/asm-sparc64/sfp-machine.h

index 5015bb8d6c3250a03720a50faefe485de13fc9ae..89d42431efb58e0cba76b903623c678cdfaf1e67 100644 (file)
@@ -34,7 +34,7 @@
 #define _FP_MUL_MEAT_D(R,X,Y)                                  \
   _FP_MUL_MEAT_1_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
 #define _FP_MUL_MEAT_Q(R,X,Y)                                  \
-  _FP_MUL_MEAT_2_wide_3mul(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
+  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
 
 #define _FP_DIV_MEAT_S(R,X,Y)  _FP_DIV_MEAT_1_imm(S,R,X,Y,_FP_DIV_HELP_imm)
 #define _FP_DIV_MEAT_D(R,X,Y)  _FP_DIV_MEAT_1_udiv_norm(D,R,X,Y)