Source-Changes-HG archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
[src/trunk]: src/sys/dev/sun wait for the blitter to go idle before trying to...
details: https://anonhg.NetBSD.org/src/rev/cd331e926794
branches: trunk
changeset: 780161:cd331e926794
user: macallan <macallan%NetBSD.org@localhost>
date: Fri Jul 13 19:48:45 2012 +0000
description:
wait for the blitter to go idle before trying to send another command
Apparently on some CG6 the GX_FULL bit is not a reliable indication wether the
engine is ready to accept another command ( LX onboard for example ) - might
be a hardware bug.
diffstat:
sys/dev/sun/cgsix.c | 14 +++++++++++---
1 files changed, 11 insertions(+), 3 deletions(-)
diffs (36 lines):
diff -r 557ddecb216c -r cd331e926794 sys/dev/sun/cgsix.c
--- a/sys/dev/sun/cgsix.c Fri Jul 13 19:32:51 2012 +0000
+++ b/sys/dev/sun/cgsix.c Fri Jul 13 19:48:45 2012 +0000
@@ -1,4 +1,4 @@
-/* $NetBSD: cgsix.c,v 1.57 2012/07/12 01:20:22 macallan Exp $ */
+/* $NetBSD: cgsix.c,v 1.58 2012/07/13 19:48:45 macallan Exp $ */
/*-
* Copyright (c) 1998 The NetBSD Foundation, Inc.
@@ -78,7 +78,7 @@
*/
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: cgsix.c,v 1.57 2012/07/12 01:20:22 macallan Exp $");
+__KERNEL_RCSID(0, "$NetBSD: cgsix.c,v 1.58 2012/07/13 19:48:45 macallan Exp $");
#include <sys/param.h>
#include <sys/systm.h>
@@ -303,8 +303,16 @@
/*EMPTY*/; \
} while (0)
+/*
+ * something is missing here
+ * Waiting for GX_FULL to clear should be enough to send another command
+ * but some CG6 ( LX onboard for example ) lock up if we do that while
+ * it works fine on others ( a 4MB TGX+ I've got here )
+ * So, until I figure out what's going on we wait for the blitter to go
+ * fully idle.
+ */
#define CG6_WAIT_READY(fbc) do { \
- while (((fbc)->fbc_s & GX_FULL) != 0) \
+ while (((fbc)->fbc_s & GX_INPROGRESS/*GX_FULL*/) != 0) \
/*EMPTY*/; \
} while (0)
Home |
Main Index |
Thread Index |
Old Index