Skip to contentSkip to conversation

drm/i915: Enable guest i915 full ppgtt functionality

PATCH In review Tina Zhang 3 participants 4 messages Aug 9, 2017 – Aug 14, 2017 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Tina Zhang

author
#1 lore raw

PATCH v9 2/3 drm/i915: Enable guest i915 full ppgtt functionality

Enable the guest i915 full ppgtt functionality when host can provide this capability. vgt_caps is introduced to guest i915 driver to get the vgpu capabilities from the device model. VGT_CPAS_FULL_PPGTT is one of the capabilities type to let guest i915 dirver know that the guest i915 full ppgtt is supported by device model.

Notice that the minor version of pvinfo isn't bumped because of this vgt_caps introduction, due to older guest would be broken by simply increasing the pvinfo version. Although the pvinfo minor version doesn't increase, the compatibility won't be blocked. The compatibility is ensured by checking the value of caps field in pvinfo. Zero means no full ppgtt support and BIT(2) means this feature is provided.

Changes since v1:
- Use u32 instead of uint32_t (Joonas)
- Move VGT_CAPS_FULL_PPGTT introduction to this patch and use #define instead of enum (Joonas)
- Rewrite the vgpu full ppgtt capability checking logic. (Joonas)
- Some coding style refine. (Joonas)

Changes since v2:
- Divide the whole patch set into two separate patch series, with one patch in i915 side to check guest i915 full ppgtt capability and enable it when this capability is supported by the device model, and the other one in gvt side which fixs the blocking issue and enables the device model to provide the capability to guest. And this patch focuses on guest i915 side. (Joonas)
- Change the title from "introduce vgt_caps to pvinfo" to "Enable guest i915 full ppgtt functionality". (Tina)

Change since v3:
- Add some comments about pvinfo caps and version. (Joonas)

Change since v4:
- Tested by Tina Zhang.

Change since v5:
- Add limitation about supporting 32bit full ppgtt.

Change since v6:
- Change the fallback to 48bit full ppgtt if i915.ppgtt_enable=2. (Zhenyu)

Change in v9:
- Remove the fixme comment due to no plan for 32bit full ppgtt support. (Zhenyu)
- Reorder the patch-set to fix compiling issue with git-bisect. (Zhenyu)
- Add print log when forcing guest 48bit full ppgtt. (Zhenyu)

drivers/gpu/drm/i915/i915_drv.h +1 -0
index b051122..1b95069 100644
@@ -1879,6 +1879,7 @@ struct i915_workarounds {
 
 struct i915_virtual_gpu {
 	bool active;
+	u32 caps;
 };
 
 /* used in computing the new watermarks state */
drivers/gpu/drm/i915/i915_gem_gtt.c +9 -4
index 10aa776..090ceb7 100644
@@ -141,14 +141,19 @@ int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv,
 
 	has_aliasing_ppgtt = dev_priv->info.has_aliasing_ppgtt;
 	has_full_ppgtt = dev_priv->info.has_full_ppgtt;
-	has_full_48bit_ppgtt = dev_priv->info.has_full_48bit_ppgtt;
 
 	if (intel_vgpu_active(dev_priv)) {
-		/* emulation is too hard */
-		has_full_ppgtt = false;
-		has_full_48bit_ppgtt = false;
+		has_full_ppgtt = intel_vgpu_has_full_ppgtt(dev_priv);
+		/* GVT-g has no support for 32bit ppgtt */
+		if (enable_ppgtt == 2 && has_full_ppgtt) {
+			DRM_DEBUG_DRIVER("Force 48bit ppgtt for vGPU\n");
+			enable_ppgtt = 3;
+		}
 	}
 
+	has_full_48bit_ppgtt = has_full_ppgtt &&
+		dev_priv->info.has_full_48bit_ppgtt;
+
 	if (!has_aliasing_ppgtt)
 		return 0;
 
drivers/gpu/drm/i915/i915_pvinfo.h +7 -1
index 2cfe96d3..7e958d7 100644
@@ -49,12 +49,18 @@ enum vgt_g2v_type {
 	VGT_G2V_MAX,
 };
 
+/*
+ * VGT capabilities type
+ */
+#define VGT_CAPS_FULL_PPGTT_48BIT	BIT(2)
+
 struct vgt_if {
 	u64 magic;		/* VGT_MAGIC */
 	u16 version_major;
 	u16 version_minor;
 	u32 vgt_id;		/* ID of vGT instance */
-	u32 rsv1[12];		/* pad to offset 0x40 */
+	u32 vgt_caps;		/* VGT capabilities */
+	u32 rsv1[11];		/* pad to offset 0x40 */
 	/*
 	 *  Data structure to describe the balooning info of resources.
 	 *  Each VM can only have one portion of continuous area for now.
drivers/gpu/drm/i915/i915_vgpu.c +7 -0
index cf7a958..e85e27c 100644
@@ -75,10 +75,17 @@ void i915_check_vgpu(struct drm_i915_private *dev_priv)
 		return;
 	}
 
+	dev_priv->vgpu.caps = __raw_i915_read32(dev_priv, vgtif_reg(vgt_caps));
+
 	dev_priv->vgpu.active = true;
 	DRM_INFO("Virtual GPU for Intel GVT-g detected.\n");
 }
 
+bool intel_vgpu_has_full_ppgtt(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->vgpu.caps & VGT_CAPS_FULL_PPGTT_48BIT;
+}
+
 struct _balloon_info_ {
 	/*
 	 * There are up to 2 regions per mappable/unmappable graphic
drivers/gpu/drm/i915/i915_vgpu.h +3 -0
index 3c3b2d2..b4e04eb 100644
@@ -27,6 +27,9 @@
 #include "i915_pvinfo.h"
 
 void i915_check_vgpu(struct drm_i915_private *dev_priv);
+
+bool intel_vgpu_has_full_ppgtt(struct drm_i915_private *dev_priv);
+
 int intel_vgt_balloon(struct drm_i915_private *dev_priv);
 void intel_vgt_deballoon(struct drm_i915_private *dev_priv);
 
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#2 lore raw
Quoted 50 lines · Tina Zhang

Enable the guest i915 full ppgtt functionality when host can provide this capability. vgt_caps is introduced to guest i915 driver to get the vgpu capabilities from the device model. VGT_CPAS_FULL_PPGTT is one of the capabilities type to let guest i915 dirver know that the guest i915 full ppgtt is supported by device model.

Notice that the minor version of pvinfo isn't bumped because of this vgt_caps introduction, due to older guest would be broken by simply increasing the pvinfo version. Although the pvinfo minor version doesn't increase, the compatibility won't be blocked. The compatibility is ensured by checking the value of caps field in pvinfo. Zero means no full ppgtt support and BIT(2) means this feature is provided.

Changes since v1:
- Use u32 instead of uint32_t (Joonas)
- Move VGT_CAPS_FULL_PPGTT introduction to this patch and use #define instead of enum (Joonas)
- Rewrite the vgpu full ppgtt capability checking logic. (Joonas)
- Some coding style refine. (Joonas)

Changes since v2:
- Divide the whole patch set into two separate patch series, with one patch in i915 side to check guest i915 full ppgtt capability and enable it when this capability is supported by the device model, and the other one in gvt side which fixs the blocking issue and enables the device model to provide the capability to guest. And this patch focuses on guest i915 side. (Joonas)
- Change the title from "introduce vgt_caps to pvinfo" to "Enable guest i915 full ppgtt functionality". (Tina)

Change since v3:
- Add some comments about pvinfo caps and version. (Joonas)

Change since v4:
- Tested by Tina Zhang.

Change since v5:
- Add limitation about supporting 32bit full ppgtt.

Change since v6:
- Change the fallback to 48bit full ppgtt if i915.ppgtt_enable=2. (Zhenyu)

Change in v9:
- Remove the fixme comment due to no plan for 32bit full ppgtt support. (Zhenyu)
- Reorder the patch-set to fix compiling issue with git-bisect. (Zhenyu)
- Add print log when forcing guest 48bit full ppgtt. (Zhenyu)

<SNIP>

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 144

Quoted 5 lines
(quoted diff) +0 -0
@@ -141,14 +141,19 @@ int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv,


  has_aliasing_ppgtt = dev_priv->info.has_aliasing_ppgtt; has_full_ppgtt = dev_priv->info.has_full_ppgtt;

drivers/gpu/drm/i915/i915_gem_gtt.c +0 -1
-	has_full_48bit_ppgtt = dev_priv->info.has_full_48bit_ppgtt;

Leave above line.

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 149

Quoted 5 lines


  if (intel_vgpu_active(dev_priv)) {

drivers/gpu/drm/i915/i915_gem_gtt.c +0 -3
-		/* emulation is too hard */
-		has_full_ppgtt = false;
-		has_full_48bit_ppgtt = false;
		has_full_ppgtt = false;
		has_full_48bit_ppgtt =
			intel_vgpu_has_full_48bit_ppgtt(dev_priv);

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 150

Quoted 5 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +5 -0
+		has_full_ppgtt = intel_vgpu_has_full_ppgtt(dev_priv);
+		/* GVT-g has no support for 32bit ppgtt */
+		if (enable_ppgtt == 2 && has_full_ppgtt) {
+			DRM_DEBUG_DRIVER("Force 48bit ppgtt for vGPU\n");
+			enable_ppgtt = 3;

Lets not do an upgrade for the user if they explicitly requested lower level. I'd just leave the option to be and it'll get downgraded to aliasing ppgtt.

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 156

Quoted 6 lines
(quoted diff) +1 -0
+		}

  }

drivers/gpu/drm/i915/i915_gem_gtt.c +3 -0
+	has_full_48bit_ppgtt = has_full_ppgtt &&
+		dev_priv->info.has_full_48bit_ppgtt;
+

This can then be dropped too.

I'll send a patch to disconnect the has_full_ppgtt and has_full_48bit_ppgtt flags.

drivers/gpu/drm/i915/i915_pvinfo.h enum vgt_g2v_type { · line 55

Quoted 11 lines

  if (!has_aliasing_ppgtt)
  return 0;

(quoted diff) +0 -0
@@ -49,12 +49,18 @@ enum vgt_g2v_type {
 >  	VGT_G2V_MAX,
  };

drivers/gpu/drm/i915/i915_pvinfo.h +4 -0
+/*
+ * VGT capabilities type
+ */
+#define VGT_CAPS_FULL_PPGTT_48BIT	BIT(2)

VGT_CAPS_FULL_48BIT_PPGTT for consistency

Quoted 11 lines
(quoted diff) +0 -0
@@ -75,10 +75,17 @@ void i915_check_vgpu(struct drm_i915_private *dev_priv)
  		return;
  	}

(quoted diff) +2 -0
+	dev_priv->vgpu.caps = __raw_i915_read32(dev_priv, vgtif_reg(vgt_caps));
+

  dev_priv->vgpu.active = true;
  DRM_INFO("Virtual GPU for Intel GVT-g detected.\n"); }

+bool intel_vgpu_has_full_ppgtt(struct drm_i915_private *dev_priv)

This should really be has_full_48bit_ppgtt() too.

Regards, Joonas

Signature · 7 lines
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
Quoted 19 lines · Joonas Lahtinen

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 144

Quoted 6 lines
(quoted diff) +0 -0
@@ -141,14 +141,19 @@ int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv,


  has_aliasing_ppgtt = dev_priv->info.has_aliasing_ppgtt; has_full_ppgtt = dev_priv->info.has_full_ppgtt;

drivers/gpu/drm/i915/i915_gem_gtt.c +0 -1
-	has_full_48bit_ppgtt = dev_priv->info.has_full_48bit_ppgtt;

Leave above line.

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 149

Quoted 5 lines


  if (intel_vgpu_active(dev_priv)) {

drivers/gpu/drm/i915/i915_gem_gtt.c +0 -3
-		/* emulation is too hard */
-		has_full_ppgtt = false;
-		has_full_48bit_ppgtt = false;
		has_full_ppgtt = false;
		has_full_48bit_ppgtt =
			intel_vgpu_has_full_48bit_ppgtt(dev_priv);
Quoted 10 lines

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 150

Quoted 5 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +5 -0
+		has_full_ppgtt = intel_vgpu_has_full_ppgtt(dev_priv);
+		/* GVT-g has no support for 32bit ppgtt */
+		if (enable_ppgtt == 2 && has_full_ppgtt) {
+			DRM_DEBUG_DRIVER("Force 48bit ppgtt for vGPU\n");
+			enable_ppgtt = 3;

Lets not do an upgrade for the user if they explicitly requested lower level. I'd just leave the option to be and it'll get downgraded to aliasing ppgtt.

If user specified enable_ppgtt=2, it will get enable_ppgtt=2 instead of aliasing ppgtt, so we try to guard against that for current 48bit ppgtt support on vgpu. Well either upgrade to full 48bit or downgrade to aliasing.

Quoted 9 lines

drivers/gpu/drm/i915/i915_gem_gtt.c int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, · line 156

Quoted 6 lines
(quoted diff) +1 -0
+		}

  }

drivers/gpu/drm/i915/i915_gem_gtt.c +3 -0
+	has_full_48bit_ppgtt = has_full_ppgtt &&
+		dev_priv->info.has_full_48bit_ppgtt;
+

This can then be dropped too.

Quoted 3 lines

I'll send a patch to disconnect the has_full_ppgtt and has_full_48bit_ppgtt flags.

That'll be good. As we have validated this full ppgtt on vgpu for quite some time now and plan to push for 4.14. Are you able to align with that? So I might send first 4.14 pull without this series, then add them later for dinf.

Quoted 34 lines

drivers/gpu/drm/i915/i915_pvinfo.h enum vgt_g2v_type { · line 55

Quoted 12 lines

  if (!has_aliasing_ppgtt)
  return 0;

(quoted diff) +0 -0
@@ -49,12 +49,18 @@ enum vgt_g2v_type {
 >  	VGT_G2V_MAX,
  };

drivers/gpu/drm/i915/i915_pvinfo.h +4 -0
+/*
+ * VGT capabilities type
+ */
+#define VGT_CAPS_FULL_PPGTT_48BIT	BIT(2)

VGT_CAPS_FULL_48BIT_PPGTT for consistency

Quoted 11 lines
(quoted diff) +0 -0
@@ -75,10 +75,17 @@ void i915_check_vgpu(struct drm_i915_private *dev_priv)
  		return;
  	}

(quoted diff) +2 -0
+	dev_priv->vgpu.caps = __raw_i915_read32(dev_priv, vgtif_reg(vgt_caps));
+

  dev_priv->vgpu.active = true;
  DRM_INFO("Virtual GPU for Intel GVT-g detected.\n"); }

+bool intel_vgpu_has_full_ppgtt(struct drm_i915_private *dev_priv)

This should really be has_full_48bit_ppgtt() too.

Regards, Joonas

Signature · 3 lines
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Applied
#4 lore raw

Re: [PATCH v9 0/3] drm/i915/gvt: Enable guest i915 48bit full ppgtt support

In reply to a message not in the archive
Quoted 9 lines · Tina Zhang

This patch set is used to enable guest i915 48bit full ppgtt support.

Kechen Lu (1):
drm/i915/gvt: Add shadow context descriptor updating

Tina Zhang (2):
drm/i915: Enable guest i915 full ppgtt functionality drm/i915/gvt: Fix guest i915 full ppgtt blocking issue

Quoted 12 lines
 drivers/gpu/drm/i915/gvt/execlist.c  |  2 ++
 drivers/gpu/drm/i915/gvt/gtt.c       | 45 ++++++++++++++++++++++--------------
 drivers/gpu/drm/i915/gvt/gvt.h       |  1 +
 drivers/gpu/drm/i915/gvt/scheduler.c | 23 ++++++++++++++++++
 drivers/gpu/drm/i915/gvt/vgpu.c      |  1 +
 drivers/gpu/drm/i915/i915_drv.h      |  1 +
 drivers/gpu/drm/i915/i915_gem_gtt.c  | 13 +++++++----
 drivers/gpu/drm/i915/i915_pvinfo.h   |  8 ++++++-
 drivers/gpu/drm/i915/i915_vgpu.c     |  7 ++++++
 drivers/gpu/drm/i915/i915_vgpu.h     |  3 +++
 10 files changed, 82 insertions(+), 22 deletions(-)

As agreed before and need to apply this series in order, i915 change for this will be merged through gvt pull.

Applied this to gvt-next. Thanks!

Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx