[PATCH v2 17/22] drm/i915/gvt: Add 64K huge gtt support

[changbin.du at intel.com]

From: Changbin Du <changbin.du at intel.com>

Finally, this add the first huge gtt support for GVTg - 64K pages. Since
64K page and 4K page cannot be mixed on the same page table, so we always
split a 64K entry into small 4K page. And when unshadow guest 64K entry,
we need ensure all the shadowed entries in shadow page table also get
cleared.

Signed-off-by: Changbin Du <changbin.du at intel.com>
---
 drivers/gpu/drm/i915/gvt/gtt.c | 90 ++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 87 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c
index bcdf70a..2064b31 100644
--- a/drivers/gpu/drm/i915/gvt/gtt.c
+++ b/drivers/gpu/drm/i915/gvt/gtt.c
@@ -948,6 +948,9 @@ static int ppgtt_invalidate_spt(struct intel_vgpu_ppgtt_spt *spt)
 			gvt_vdbg_mm("invalidate 4K entry\n");
 			continue;
 		case GTT_TYPE_PPGTT_PTE_64K_ENTRY:
+			/* We don't setup 64K shadow entry so far. */
+			WARN(1, "suspicious 64K gtt entry\n");
+			continue;
 		case GTT_TYPE_PPGTT_PTE_2M_ENTRY:
 		case GTT_TYPE_PPGTT_PTE_1G_ENTRY:
 			WARN(1, "GVT doesn't support 64K/2M/1GB page\n");
@@ -1043,9 +1046,65 @@ static inline void ppgtt_generate_shadow_entry(struct intel_gvt_gtt_entry *se,
 	se->type = ge->type;
 	se->val64 = ge->val64;
 
+	/* Because we always split 64KB pages, so clear IPS in shadow PDE. */
+	if (se->type == GTT_TYPE_PPGTT_PDE_ENTRY)
+		ops->clear_ips(se);
+
 	ops->set_pfn(se, s->shadow_page.mfn);
 }
 
+/**
+ * Return 1 if huge gtt shadowing is possilbe, 0 if miscondition,
+ * negtive if found err.
+ */
+static int is_huge_gtt_possible(struct intel_vgpu *vgpu,
+	struct intel_gvt_gtt_entry *entry)
+{
+	switch (entry->type) {
+	case GTT_TYPE_PPGTT_PTE_64K_ENTRY:
+		/**
+		 * 64K page is special, the size if controlled by
+		 * uper PDE. To be simple, we always split it.
+		 */
+		return 0;
+	case GTT_TYPE_PPGTT_PTE_2M_ENTRY:
+		/* Will add later. */
+	default:
+		return 0;
+	}
+
+	return 0;
+}
+
+static int split_64KB_gtt_entry(struct intel_vgpu *vgpu,
+	struct intel_vgpu_ppgtt_spt *spt, unsigned long index,
+	struct intel_gvt_gtt_entry *se)
+{
+	struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_gvt_gtt_entry entry = *se;
+	unsigned long start_gfn, mfn;
+	int i;
+
+	gvt_vdbg_mm("Split 64K gtt entry, index %lu\n", index);
+
+	GEM_BUG_ON(index % GTT_64K_PTE_STRIDE);
+
+	start_gfn = ops->get_pfn(se);
+
+	entry.type = GTT_TYPE_PPGTT_PTE_4K_ENTRY;
+	ops->set_64k_splited(&entry);
+
+	for (i = 0; i < GTT_64K_PTE_STRIDE; i++) {
+		mfn = intel_gvt_hypervisor_gfn_to_mfn(vgpu, start_gfn + i);
+		if (mfn == INTEL_GVT_INVALID_ADDR)
+			return -ENXIO;
+
+		ops->set_pfn(&entry, mfn);
+		ppgtt_set_shadow_entry(spt, &entry, index + i);
+	}
+	return 0;
+}
+
 static int ppgtt_populate_shadow_entry(struct intel_vgpu *vgpu,
 	struct intel_vgpu_ppgtt_spt *spt, unsigned long index,
 	struct intel_gvt_gtt_entry *ge)
@@ -1053,6 +1112,7 @@ static int ppgtt_populate_shadow_entry(struct intel_vgpu *vgpu,
 	struct intel_gvt_gtt_pte_ops *pte_ops = vgpu->gvt->gtt.pte_ops;
 	struct intel_gvt_gtt_entry se = *ge;
 	unsigned long gfn, mfn;
+	int ret;
 
 	if (!pte_ops->test_present(ge))
 		return 0;
@@ -1064,6 +1124,13 @@ static int ppgtt_populate_shadow_entry(struct intel_vgpu *vgpu,
 		gvt_vdbg_mm("shadow 4K gtt entry\n");
 		break;
 	case GTT_TYPE_PPGTT_PTE_64K_ENTRY:
+		gvt_vdbg_mm("shadow 64K gtt entry\n");
+		ret = is_huge_gtt_possible(vgpu, ge);
+		if (ret == 0)
+			return split_64KB_gtt_entry(vgpu, spt, index, &se);
+		else if (ret < 0)
+			return -EINVAL;
+		break;
 	case GTT_TYPE_PPGTT_PTE_2M_ENTRY:
 	case GTT_TYPE_PPGTT_PTE_1G_ENTRY:
 		gvt_vgpu_err("GVT doesn't support 64K/2M/1GB entry\n");
@@ -1147,6 +1214,10 @@ static int ppgtt_handle_guest_entry_removal(struct intel_vgpu_ppgtt_spt *spt,
 		ret = ppgtt_invalidate_spt(s);
 		if (ret)
 			goto fail;
+	} else {
+		/* We don't setup 64K shadow entry so far. */
+		WARN(se->type == GTT_TYPE_PPGTT_PTE_64K_ENTRY,
+		     "suspicious 64K entry\n");
 	}
 	return 0;
 fail:
@@ -1369,7 +1440,7 @@ static int ppgtt_handle_guest_write_page_table(
 	struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
 	struct intel_gvt_gtt_entry old_se;
 	int new_present;
-	int ret;
+	int i, ret;
 
 	new_present = ops->test_present(we);
 
@@ -1391,8 +1462,21 @@ static int ppgtt_handle_guest_write_page_table(
 		goto fail;
 
 	if (!new_present) {
-		ops->set_pfn(&old_se, vgpu->gtt.scratch_pt[type].page_mfn);
-		ppgtt_set_shadow_entry(spt, &old_se, index);
+		/* For 64KB splited entries, we need clear them all. */
+		if (ops->test_64k_splited(&old_se) &&
+		    !(index % GTT_64K_PTE_STRIDE)) {
+			gvt_vdbg_mm("remove splited 64K shadow entries\n");
+			for (i = 0; i < GTT_64K_PTE_STRIDE; i++) {
+				ops->clear_64k_splited(&old_se);
+				ops->set_pfn(&old_se,
+					vgpu->gtt.scratch_pt[type].page_mfn);
+				ppgtt_set_shadow_entry(spt, &old_se, index + i);
+			}
+		} else {
+			ops->set_pfn(&old_se,
+				     vgpu->gtt.scratch_pt[type].page_mfn);
+			ppgtt_set_shadow_entry(spt, &old_se, index);
+		}
 	}
 
 	return 0;
-- 
2.7.4