diff options
| author | Mark Kettenis <kettenis@cvs.openbsd.org> | 2010-05-02 22:26:59 +0000 |
|---|---|---|
| committer | Mark Kettenis <kettenis@cvs.openbsd.org> | 2010-05-02 22:26:59 +0000 |
| commit | 463582c1d2af352c0fc668f518d76a4291997276 (patch) | |
| tree | f1a5d8e710bccd1c64920bd700a3b66f183410c5 /sys/arch/i386 | |
| parent | 7494a552c9857ea80aecb9e1bb2b22e09a59c2af (diff) | |
The Cyrix "coma bug" workaround code has a really convoluted way to write 0
into a magic register. Simplify the code by making this explicit, but keep
the dummy read just in case this has a magic side-effect. And yes, as far as
I can tell, writing 0 is really what was intended here. Makes gcc4 happy.
No binary change with gcc3.
ok jsg@
Diffstat (limited to 'sys/arch/i386')
| -rw-r--r-- | sys/arch/i386/i386/machdep.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/sys/arch/i386/i386/machdep.c b/sys/arch/i386/i386/machdep.c index 1b8e2d8959d..dc2f970d364 100644 --- a/sys/arch/i386/i386/machdep.c +++ b/sys/arch/i386/i386/machdep.c @@ -1,4 +1,4 @@ -/* $OpenBSD: machdep.c,v 1.469 2010/03/21 23:00:57 jsg Exp $ */ +/* $OpenBSD: machdep.c,v 1.470 2010/05/02 22:26:58 kettenis Exp $ */ /* $NetBSD: machdep.c,v 1.214 1996/11/10 03:16:17 thorpej Exp $ */ /*- @@ -1239,7 +1239,7 @@ cyrix6x86_cpu_setup(struct cpu_info *ci) /* cyrix's workaround for the "coma bug" */ cyrix_write_reg(0x31, cyrix_read_reg(0x31) | 0xf8); cyrix_write_reg(0x32, cyrix_read_reg(0x32) | 0x7f); - cyrix_write_reg(0x33, cyrix_read_reg(0x33) & ~0xff); + cyrix_read_reg(0x33); cyrix_write_reg(0x33, 0); cyrix_write_reg(0x3c, cyrix_read_reg(0x3c) | 0x87); /* disable access to ccr4/ccr5 */ cyrix_write_reg(0xC3, cyrix_read_reg(0xC3) & ~0x10); |