[Intel-gfx] [PATCH 09/10 v2] drm/i915: flush system agent TLBs on SNB

Ben Widawsky ben at bwidawsk.net
Fri Oct 26 22:55:05 CEST 2012


This allows us to map the PTEs WC. I've not done thorough testing or
performance measurements with this patch, but it should be decent.

This is based on a patch from Jesse with the original commit message
> I've only lightly tested this so far, but the corruption seems to be
> gone if I write the GFX_FLSH_CNTL reg after binding an object.  This
> register should control the TLB for the system agent, which is what CPU
> mapped objects will go through.

It has been updated for the new AGP-less code by me, and included with
it is feedback from the original patch.

v2: Updated to reflect paranoia on pte updates/register posting reads.
This entails doing the flush (with posting_read) after the pte readback

Signed-off-by: Ben Widawsky <ben at bwidawsk.net>
Reviewed-by [v1]: Jesse Barnes <jbarnes at virtuousgeek.org>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 11 +++++++++--
 drivers/gpu/drm/i915/i915_reg.h     |  2 ++
 2 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index f832ab0..47a408a 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -402,6 +402,13 @@ out:
 	 * hardware should work, we must keep this posting read for paranoia.
 	 */
 	WARN_ON(readl(&gtt_entries[i-1]) != pte_encode(dev, addr, level));
+
+	/* This next bit makes the above posting read even more important. We
+	 * want to flush the TLBs only after we're certain all the PTE updates
+	 * have finished.
+	 */
+	I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN);
+	POSTING_READ(GFX_FLSH_CNTL_GEN6);
 }
 
 void i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj,
@@ -599,8 +606,8 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	if (ret)
 		return ret;
 
-	dev_priv->mm.gtt->gtt = ioremap(gtt_bus_addr,
-					dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
+	dev_priv->mm.gtt->gtt = ioremap_wc(gtt_bus_addr,
+					   dev_priv->mm.gtt->gtt_total_entries * sizeof(gtt_pte_t));
 
 	/* GMADR is the PCI aperture used by SW to access tiled GFX surfaces in a linear fashion. */
 	DRM_DEBUG_DRIVER("GMADR size = %dM\n", dev_priv->mm.gtt->gtt_mappable_entries >> 8);
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 1d54328..e8c578f 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -683,6 +683,8 @@
 #define   CM0_RC_OP_FLUSH_DISABLE (1<<0)
 #define BB_ADDR		0x02140 /* 8 bytes */
 #define GFX_FLSH_CNTL	0x02170 /* 915+ only */
+#define GFX_FLSH_CNTL_GEN6	0x101008
+#define   GFX_FLSH_CNTL_EN	(1<<0)
 #define ECOSKPD		0x021d0
 #define   ECO_GATING_CX_ONLY	(1<<3)
 #define   ECO_FLIP_DONE		(1<<0)
-- 
1.8.0




More information about the Intel-gfx mailing list