Skip to contentSkip to conversation

drm/i915: Report enabled slices on Haswell GT3

PATCH Applied Rodrigo Vivi 3 participants 3 messages Aug 28, 2013 – Aug 29, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: Report enabled slices on Haswell GT3

Batchbuffers constructed by userspace can conditionalise their URB allocations through the use of the MI_SET_PREDICATE command. This command can read the MI_PREDICATE_RESULT_2 register to see how many slices are enabled on GT3, and by virtue of the result, scale their memory allocations to fit enabled memory.

Of course, this only works if the kernel sets the appropriate bit in the register first.

v2: Better commit subject and message by Chris Wilson.

---
 drivers/gpu/drm/i915/i915_drv.h | 2 ++
 drivers/gpu/drm/i915/i915_gem.c | 5 +++++
 drivers/gpu/drm/i915/i915_reg.h | 5 +++++
 3 files changed, 12 insertions(+)
drivers/gpu/drm/i915/i915_drv.h +2 -0
index 14299a4..9e25079 100644
@@ -1519,6 +1519,8 @@ struct drm_i915_file_private {
 				 ((dev)->pci_device & 0xFF00) == 0x0C00)
 #define IS_ULT(dev)		(IS_HASWELL(dev) && \
 				 ((dev)->pci_device & 0xFF00) == 0x0A00)
+#define IS_HSW_GT3(dev)		(IS_HASWELL(dev) && \
+				 ((dev)->pci_device & 0x00F0) == 0x0020)
 
 /*
  * The genX designation typically refers to the render engine, so render
drivers/gpu/drm/i915/i915_gem.c +5 -0
index f705314..d88c8cb 100644
@@ -4328,6 +4328,11 @@ i915_gem_init_hw(struct drm_device *dev)
 	if (dev_priv->ellc_size)
 		I915_WRITE(HSW_IDICR, I915_READ(HSW_IDICR) | IDIHASHMSK(0xf));
 
+	if (IS_HSW_GT3(dev))
+		I915_WRITE(MI_PREDICATE_RESULT_2, LOWER_SLICE_ENABLED);
+	else
+		I915_WRITE(MI_PREDICATE_RESULT_2, LOWER_SLICE_DISABLED);
+
 	if (HAS_PCH_NOP(dev)) {
 		u32 temp = I915_READ(GEN7_MSG_CTL);
 		temp &= ~(WAIT_FOR_PCH_FLR_ACK | WAIT_FOR_PCH_RESET_ACK);
drivers/gpu/drm/i915/i915_reg.h +5 -0
index 53d0e70..2ab0de6 100644
@@ -278,6 +278,11 @@
 #define  MI_SEMAPHORE_SYNC_VVE	    (1<<16) /* VECS wait for VCS  (VEVSYNC) */
 #define  MI_SEMAPHORE_SYNC_RVE	    (2<<16) /* VECS wait for RCS  (VERSYNC) */
 #define  MI_SEMAPHORE_SYNC_INVALID  (3<<16)
+
+#define MI_PREDICATE_RESULT_2	(0x2214)
+#define  LOWER_SLICE_ENABLED	(1<<0)
+#define  LOWER_SLICE_DISABLED	(0<<0)
+
 /*
  * 3D instructions used by the kernel
  */
Signature · 1 line
1.8.1.4
Reviewed-by on v1 1/1
#1 lore raw
Quoted 6 lines · Rodrigo Vivi

Batchbuffers constructed by userspace can conditionalise their URB allocations through the use of the MI_SET_PREDICATE command. This command can read the MI_PREDICATE_RESULT_2 register to see how many slices are enabled on GT3, and by virtue of the result, scale their memory allocations to fit enabled memory.

Quoted 8 lines

Of course, this only works if the kernel sets the appropriate bit in the register first.

v2: Better commit subject and message by Chris Wilson.

I would have written the I915_WRITE() differently but,

-Chris

Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
#2 lore raw
Quoted 10 lines · Chris Wilson
Quoted 9 lines · Rodrigo Vivi

Batchbuffers constructed by userspace can conditionalise their URB allocations through the use of the MI_SET_PREDICATE command. This command can read the MI_PREDICATE_RESULT_2 register to see how many slices are enabled on GT3, and by virtue of the result, scale their memory allocations to fit enabled memory.

Of course, this only works if the kernel sets the appropriate bit in the register first.

Quoted 8 lines
Quoted 5 lines

v2: Better commit subject and message by Chris Wilson.

I would have written the I915_WRITE() differently but,

Queued for -next, thanks for the patch.
-Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch