Mesa (master): i965: Add #defines for Memory Object Control State fields on Gen7-7.5.

Kenneth Graunke kwg at kemper.freedesktop.org
Thu Jul 18 23:03:17 UTC 2013


Module: Mesa
Branch: master
Commit: 7791c9869b233e45a9089eae124dc9aa4f4e519a
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=7791c9869b233e45a9089eae124dc9aa4f4e519a

Author: Kenneth Graunke <kenneth at whitecape.org>
Date:   Tue Jul  9 20:47:54 2013 -0700

i965: Add #defines for Memory Object Control State fields on Gen7-7.5.

The L3 controls are identical on all platforms, but LLC differs:
- Ivybridge has a "cache in LLC" flag
- Baytrail has no LLC, but instead has a snoop bit:
  "data accesses in this page must be snooped in the CPU caches."
- Haswell has writeback/uncached flags for LLC and eLLC (eDRAM).

Signed-off-by: Kenneth Graunke <kenneth at whitecape.org>
Reviewed-by: Chad Versace <chad.versace at linux.intel.com>

---

 src/mesa/drivers/dri/i965/brw_defines.h |   26 ++++++++++++++++++++++++++
 1 files changed, 26 insertions(+), 0 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_defines.h b/src/mesa/drivers/dri/i965/brw_defines.h
index c6b8dab..d8b3b17 100644
--- a/src/mesa/drivers/dri/i965/brw_defines.h
+++ b/src/mesa/drivers/dri/i965/brw_defines.h
@@ -1730,6 +1730,32 @@ enum brw_wm_barycentric_interp_mode {
  */
 #define BRW_MAX_NUM_BUFFER_ENTRIES	(1 << 27)
 
+/* Memory Object Control State:
+ * Specifying zero for L3 means "uncached in L3", at least on Haswell
+ * and Baytrail, since there are no PTE flags for setting L3 cacheability.
+ * On Ivybridge, the PTEs do have a cache-in-L3 bit, so setting MOCS to 0
+ * may still respect that.
+ */
+#define GEN7_MOCS_L3                    1
+
+/* Ivybridge only: cache in LLC.
+ * Specifying zero here means to use the PTE values set by the kernel;
+ * non-zero overrides the PTE values.
+ */
+#define IVB_MOCS_LLC                    (1 << 1)
+
+/* Baytrail only: snoop in CPU cache */
+#define BYT_MOCS_SNOOP                  (1 << 1)
+
+/* Haswell only: LLC/eLLC controls (write-back or uncached).
+ * Specifying zero here means to use the PTE values set by the kernel,
+ * which is useful since it offers additional control (write-through
+ * cacheing and age).  Non-zero overrides the PTE values.
+ */
+#define HSW_MOCS_UC_LLC_UC_ELLC         (1 << 1)
+#define HSW_MOCS_WB_LLC_WB_ELLC         (2 << 1)
+#define HSW_MOCS_UC_LLC_WB_ELLC         (3 << 1)
+
 #include "intel_chipset.h"
 
 #endif




More information about the mesa-commit mailing list