Source-Changes-HG archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
[src/trunk]: src/sys/arch/powerpc/include Make sure to clear PSL_SE before go...
details: https://anonhg.NetBSD.org/src/rev/adf2ed197d63
branches: trunk
changeset: 327521:adf2ed197d63
user: matt <matt%NetBSD.org@localhost>
date: Sun Mar 09 23:09:32 2014 +0000
description:
Make sure to clear PSL_SE before going to usermode on BOOKE
Revert back to testing PSL_SE in the trapframe and not mdlwp md_flags
diffstat:
sys/arch/powerpc/include/userret.h | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
diffs (25 lines):
diff -r 71975e5ffffb -r adf2ed197d63 sys/arch/powerpc/include/userret.h
--- a/sys/arch/powerpc/include/userret.h Sun Mar 09 23:01:11 2014 +0000
+++ b/sys/arch/powerpc/include/userret.h Sun Mar 09 23:09:32 2014 +0000
@@ -1,4 +1,4 @@
-/* $NetBSD: userret.h,v 1.25 2014/03/09 22:31:25 matt Exp $ */
+/* $NetBSD: userret.h,v 1.26 2014/03/09 23:09:32 matt Exp $ */
/*
* Copyright (C) 1995, 1996 Wolfgang Solfrank.
@@ -75,11 +75,11 @@
#ifdef PPC_BOOKE
/*
* BookE doesn't have PSL_SE but it does have a debug instruction
- * completion exception but it needs PSL_DE to fire. Since we don't
- * want it to happen in the kernel, we must disable PSL_DE and let
- * it get restored by rfi/rfci.
+ * completion exception but it needs PSL_DE to fire. Instead we
+ * use IAC1/IAC2 to match the next PC.
*/
- if (__predict_false(l->l_md.md_flags & PSL_SE)) {
+ if (__predict_false(tf->tf_ssr1 & PSL_SE)) {
+ tf->tf_ssr1 &= ~PSL_SE;
extern void booke_sstep(struct trapframe *); /* ugly */
booke_sstep(tf);
}
Home |
Main Index |
Thread Index |
Old Index