Hi.
On Tue, Sep 14, 2004 at 06:39:10PM -0700, Joerg Sonnenberger wrote:
> joerg 2004/09/14 18:39:10 PDT
>
> DragonFly src repository
>
> Modified files:
> sys/dev/acpica acpi.c acpivar.h
> Log:
> Convert timeout ==> callout_*.
>
> Note that ACPI doesn't support detach / unloading, therefore no
> callout_stop is used.
>
> Revision Changes Path
> 1.7 +3 -1 src/sys/dev/acpica/acpi.c
> 1.3 +1 -0 src/sys/dev/acpica/acpivar.h
>
>
> http://www.dragonflybsd.org/cvsweb/src/sys/dev/acpica/acpi.c.diff?r1=1.6&r2=1.7&f=u
> http://www.dragonflybsd.org/cvsweb/src/sys/dev/acpica/acpivar.h.diff?r1=1.2&r2=1.3&f=u
After this commit, the acpi.ko(acpica5) panicked in acpi_machdep_init().
It turned out that there's been two places where MD-part of acpica5
code still using header files from older acpica code. This resulted in
MD code seeing the wrong place in the struct acpi_softc. I thought I
fixed this in the patches, but it seems I didn't.
Index: i386/acpica5/acpi_machdep.c
===================================================================
RCS file: /home/source/dragonfly/cvs/src/sys/i386/acpica5/acpi_machdep.c,v
retrieving revision 1.5
diff -u -r1.5 acpi_machdep.c
--- i386/acpica5/acpi_machdep.c 5 Jul 2004 00:07:35 -0000 1.5
+++ i386/acpica5/acpi_machdep.c 15 Sep 2004 15:51:11 -0000
@@ -36,8 +36,8 @@
#include <sys/uio.h>
#include "acpi.h"
-#include <dev/acpica/acpivar.h>
-#include <dev/acpica/acpiio.h>
+#include "acpivar.h"
+#include "acpiio.h"
static device_t acpi_dev;
Index: i386/acpica5/acpi_wakeup.c
===================================================================
RCS file: /home/source/dragonfly/cvs/src/sys/i386/acpica5/acpi_wakeup.c,v
retrieving revision 1.4
diff -u -r1.4 acpi_wakeup.c
--- i386/acpica5/acpi_wakeup.c 5 Jul 2004 00:14:01 -0000 1.4
+++ i386/acpica5/acpi_wakeup.c 15 Sep 2004 15:51:23 -0000
@@ -47,7 +47,7 @@
#include <machine/segments.h>
#include "acpi.h"
-#include <dev/acpica/acpivar.h>
+#include "acpivar.h"
#include "acpi_wakecode.h"