Source-Changes-HG archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

[src/trunk]: src/sys/arch/i386/i386 The reason we are not using INTRENTRY her...



details:   https://anonhg.NetBSD.org/src/rev/1ceb3b89b308
branches:  trunk
changeset: 819020:1ceb3b89b308
user:      maxv <maxv%NetBSD.org@localhost>
date:      Sun Nov 13 12:38:14 2016 +0000

description:
The reason we are not using INTRENTRY here is because this interrupt goes
through a task gate that points to a TSS entry in the GDT, and therefore
the GPRs are saved in the TSS by the hardware itself. Explain it, otherwise
it easily looks buggy.

diffstat:

 sys/arch/i386/i386/vector.S |  8 ++++++--
 1 files changed, 6 insertions(+), 2 deletions(-)

diffs (29 lines):

diff -r a082c4490ad6 -r 1ceb3b89b308 sys/arch/i386/i386/vector.S
--- a/sys/arch/i386/i386/vector.S       Sat Nov 12 22:38:55 2016 +0000
+++ b/sys/arch/i386/i386/vector.S       Sun Nov 13 12:38:14 2016 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: vector.S,v 1.66 2016/10/16 10:38:49 maxv Exp $ */
+/*     $NetBSD: vector.S,v 1.67 2016/11/13 12:38:14 maxv Exp $ */
 
 /*
  * Copyright 2002 (c) Wasabi Systems, Inc.
@@ -65,7 +65,7 @@
  */
 
 #include <machine/asm.h>
-__KERNEL_RCSID(0, "$NetBSD: vector.S,v 1.66 2016/10/16 10:38:49 maxv Exp $");
+__KERNEL_RCSID(0, "$NetBSD: vector.S,v 1.67 2016/11/13 12:38:14 maxv Exp $");
 
 #include "opt_ddb.h"
 #include "opt_multiprocessor.h"
@@ -194,6 +194,10 @@
 IDTVEC_END(intr_lapic_tlb)
 
 #if defined(DDB)
+/*
+ * No need to use INTRENTRY, since we were brought here through a task-gate
+ * which triggered a hardware context switch and saved the GPRs in the TSS.
+ */
 IDTVEC(intrddbipi)
 1:
        str     %ax



Home | Main Index | Thread Index | Old Index