[src/trunk]: src/sys/arch/arm/arm32 If the vector_page is not ARM_VECTORS_{LO...

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

details:   https://anonhg.NetBSD.org/src/rev/1ce83d89d1fe
branches:  trunk
changeset: 787352:1ce83d89d1fe
user:      matt <matt%NetBSD.org@localhost>
date:      Wed Jun 12 07:13:18 2013 +0000

description:
If the vector_page is not ARM_VECTORS_{LOW,HIGH}, assume it's in kernel
text and don't do anything special to map it.

diffstat:

 sys/arch/arm/arm32/pmap.c |  19 +++++++++++++++----
 1 files changed, 15 insertions(+), 4 deletions(-)

diffs (47 lines):

diff -r ba418e570ce9 -r 1ce83d89d1fe sys/arch/arm/arm32/pmap.c
--- a/sys/arch/arm/arm32/pmap.c Wed Jun 12 07:12:10 2013 +0000
+++ b/sys/arch/arm/arm32/pmap.c Wed Jun 12 07:13:18 2013 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: pmap.c,v 1.255 2013/05/11 10:15:43 skrll Exp $ */
+/*     $NetBSD: pmap.c,v 1.256 2013/06/12 07:13:18 matt Exp $  */
 
 /*
  * Copyright 2003 Wasabi Systems, Inc.
@@ -162,8 +162,8 @@
  *       the active domain on that cpu). I guess there are lots more tlb
  *       shootdown issues too...
  *
- *     o If the vector_page is at 0x00000000 instead of 0xffff0000, then
- *       MP systems will lose big-time because of the MMU domain hack.
+ *     o If the vector_page is at 0x00000000 instead of in kernel VA space,
+ *       then MP systems will lose big-time because of the MMU domain hack.
  *       The only way this can be solved (apart from moving the vector
  *       page to 0xffff0000) is to reserve the first 1MB of user address
  *       space for kernel use only. This would require re-linking all
@@ -212,7 +212,7 @@
 #include <arm/cpuconf.h>
 #include <arm/arm32/katelib.h>
 
-__KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.255 2013/05/11 10:15:43 skrll Exp $");
+__KERNEL_RCSID(0, "$NetBSD: pmap.c,v 1.256 2013/06/12 07:13:18 matt Exp $");
 
 #ifdef PMAP_DEBUG
 
@@ -5135,6 +5135,17 @@
        struct l2_bucket *l2b;
        pt_entry_t *ptep;
 
+#if defined(CPU_ARMV7) || defined(CPU_ARM11)
+       /*
+        * If we are using VBAR to use the vectors in the kernel, then it's
+        * already mapped in the kernel text so no need to anything here.
+        */
+       if (vector_page != ARM_VECTORS_LOW && vector_page != ARM_VECTORS_HIGH) {
+               KASSERT((armreg_pfr1_read() & ARM_PFR1_SEC_MASK) != 0);
+               return;
+       }
+#endif
+
        l2b = pmap_get_l2_bucket(pmap_kernel(), vector_page);
        KDASSERT(l2b != NULL);