[src/trunk]: src/sys/arch/arm/vfp Remove the call to vfp_fpscr_handler() from...

[bouyer <bouyer%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/651bc850cd68
branches:  trunk
changeset: 321861:651bc850cd68
user:      bouyer <bouyer%NetBSD.org@localhost>
date:      Sun Apr 08 09:19:27 2018 +0000

description:
Remove the call to vfp_fpscr_handler() from vfp_handler(). It actually never
avoids a full FPU switch, and costs a function call and a few tests.

Discussed on port-arm@ on october 2017:
http://mail-index.netbsd.org/port-arm/2017/10/16/msg004411.html

diffstat:

 sys/arch/arm/vfp/vfp_init.c |  9 +--------
 1 files changed, 1 insertions(+), 8 deletions(-)

diffs (23 lines):

diff -r f4121528fba4 -r 651bc850cd68 sys/arch/arm/vfp/vfp_init.c
--- a/sys/arch/arm/vfp/vfp_init.c       Sun Apr 08 08:57:37 2018 +0000
+++ b/sys/arch/arm/vfp/vfp_init.c       Sun Apr 08 09:19:27 2018 +0000
@@ -1,4 +1,4 @@
-/*      $NetBSD: vfp_init.c,v 1.56 2018/03/02 23:07:55 christos Exp $ */
+/*      $NetBSD: vfp_init.c,v 1.57 2018/04/08 09:19:27 bouyer Exp $ */
 
 /*
  * Copyright (c) 2008 ARM Ltd
@@ -409,13 +409,6 @@
                return 1;
        }
 
-       /*
-        * If we are just changing/fetching FPSCR, don't bother loading it
-        * just emulate the instruction.
-        */
-       if (!vfp_fpscr_handler(address, insn, frame, fault_code))
-               return 0;
-
        /* 
         * If we already own the FPU and it's enabled (and no exception), raise
         * SIGILL.  If there is an exception, drop through to raise a SIGFPE.