drm/panfrost: Make sure MMU context lifetime is not bound to panfrost_priv

Jobs can be in-flight when the file descriptor is closed (either because
the process did not terminate properly, or because it didn't wait for
all GPU jobs to be finished), and apparently panfrost_job_close() does
not cancel already running jobs. Let's refcount the MMU context object
so it's lifetime is no longer bound to the FD lifetime and running jobs
can finish properly without generating spurious page faults.

Reported-by: Icecream95 <ixn@keemail.me>
Fixes: 7282f7645d06 ("drm/panfrost: Implement per FD address spaces")
Cc: <stable@vger.kernel.org>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210621133907.1683899-2-boris.brezillon@collabora.com

1 files changed, 108 insertions, 52 deletions

diff --git a/drivers/gpu/drm/panfrost/panfrost_mmu.c b/drivers/gpu/drm/panfrost/panfrost_mmu.c
index 0581186ebfb3..569509c2ba27 100644
--- a/drivers/gpu/drm/panfrost/panfrost_mmu.c
+++ b/drivers/gpu/drm/panfrost/panfrost_mmu.c
@@ -1,5 +1,8 @@
 // SPDX-License-Identifier:	GPL-2.0
 /* Copyright 2019 Linaro, Ltd, Rob Herring <robh@kernel.org> */
+
+#include <drm/panfrost_drm.h>
+
 #include <linux/atomic.h>
 #include <linux/bitfield.h>
 #include <linux/delay.h>
@@ -337,7 +340,7 @@ static void mmu_tlb_inv_context_s1(void *cookie)
 
 static void mmu_tlb_sync_context(void *cookie)
 {
-	//struct panfrost_device *pfdev = cookie;
+	//struct panfrost_mmu *mmu = cookie;
 	// TODO: Wait 1000 GPU cycles for HW_ISSUE_6367/T60X
 }
 
@@ -352,57 +355,10 @@ static const struct iommu_flush_ops mmu_tlb_ops = {
 	.tlb_flush_walk = mmu_tlb_flush_walk,
 };
 
-int panfrost_mmu_pgtable_alloc(struct panfrost_file_priv *priv)
-{
-	struct panfrost_mmu *mmu = &priv->mmu;
-	struct panfrost_device *pfdev = priv->pfdev;
-
-	INIT_LIST_HEAD(&mmu->list);
-	mmu->as = -1;
-
-	mmu->pgtbl_cfg = (struct io_pgtable_cfg) {
-		.pgsize_bitmap	= SZ_4K | SZ_2M,
-		.ias		= FIELD_GET(0xff, pfdev->features.mmu_features),
-		.oas		= FIELD_GET(0xff00, pfdev->features.mmu_features),
-		.coherent_walk	= pfdev->coherent,
-		.tlb		= &mmu_tlb_ops,
-		.iommu_dev	= pfdev->dev,
-	};
-
-	mmu->pgtbl_ops = alloc_io_pgtable_ops(ARM_MALI_LPAE, &mmu->pgtbl_cfg,
-					      priv);
-	if (!mmu->pgtbl_ops)
-		return -EINVAL;
-
-	return 0;
-}
-
-void panfrost_mmu_pgtable_free(struct panfrost_file_priv *priv)
-{
-	struct panfrost_device *pfdev = priv->pfdev;
-	struct panfrost_mmu *mmu = &priv->mmu;
-
-	spin_lock(&pfdev->as_lock);
-	if (mmu->as >= 0) {
-		pm_runtime_get_noresume(pfdev->dev);
-		if (pm_runtime_active(pfdev->dev))
-			panfrost_mmu_disable(pfdev, mmu->as);
-		pm_runtime_put_autosuspend(pfdev->dev);
-
-		clear_bit(mmu->as, &pfdev->as_alloc_mask);
-		clear_bit(mmu->as, &pfdev->as_in_use_mask);
-		list_del(&mmu->list);
-	}
-	spin_unlock(&pfdev->as_lock);
-
-	free_io_pgtable_ops(mmu->pgtbl_ops);
-}
-
 static struct panfrost_gem_mapping *
 addr_to_mapping(struct panfrost_device *pfdev, int as, u64 addr)
 {
 	struct panfrost_gem_mapping *mapping = NULL;
-	struct panfrost_file_priv *priv;
 	struct drm_mm_node *node;
 	u64 offset = addr >> PAGE_SHIFT;
 	struct panfrost_mmu *mmu;
@@ -415,11 +371,10 @@ addr_to_mapping(struct panfrost_device *pfdev, int as, u64 addr)
 	goto out;
 
 found_mmu:
-	priv = container_of(mmu, struct panfrost_file_priv, mmu);
 
-	spin_lock(&priv->mm_lock);
+	spin_lock(&mmu->mm_lock);
 
-	drm_mm_for_each_node(node, &priv->mm) {
+	drm_mm_for_each_node(node, &mmu->mm) {
 		if (offset >= node->start &&
 		    offset < (node->start + node->size)) {
 			mapping = drm_mm_node_to_panfrost_mapping(node);
@@ -429,7 +384,7 @@ found_mmu:
 		}
 	}
 
-	spin_unlock(&priv->mm_lock);
+	spin_unlock(&mmu->mm_lock);
 out:
 	spin_unlock(&pfdev->as_lock);
 	return mapping;
@@ -542,6 +497,107 @@ err_bo:
 	return ret;
 }
 
+static void panfrost_mmu_release_ctx(struct kref *kref)
+{
+	struct panfrost_mmu *mmu = container_of(kref, struct panfrost_mmu,
+						refcount);
+	struct panfrost_device *pfdev = mmu->pfdev;
+
+	spin_lock(&pfdev->as_lock);
+	if (mmu->as >= 0) {
+		pm_runtime_get_noresume(pfdev->dev);
+		if (pm_runtime_active(pfdev->dev))
+			panfrost_mmu_disable(pfdev, mmu->as);
+		pm_runtime_put_autosuspend(pfdev->dev);
+
+		clear_bit(mmu->as, &pfdev->as_alloc_mask);
+		clear_bit(mmu->as, &pfdev->as_in_use_mask);
+		list_del(&mmu->list);
+	}
+	spin_unlock(&pfdev->as_lock);
+
+	free_io_pgtable_ops(mmu->pgtbl_ops);
+	drm_mm_takedown(&mmu->mm);
+	kfree(mmu);
+}
+
+void panfrost_mmu_ctx_put(struct panfrost_mmu *mmu)
+{
+	kref_put(&mmu->refcount, panfrost_mmu_release_ctx);
+}
+
+struct panfrost_mmu *panfrost_mmu_ctx_get(struct panfrost_mmu *mmu)
+{
+	kref_get(&mmu->refcount);
+
+	return mmu;
+}
+
+#define PFN_4G		(SZ_4G >> PAGE_SHIFT)
+#define PFN_4G_MASK	(PFN_4G - 1)
+#define PFN_16M		(SZ_16M >> PAGE_SHIFT)
+
+static void panfrost_drm_mm_color_adjust(const struct drm_mm_node *node,
+					 unsigned long color,
+					 u64 *start, u64 *end)
+{
+	/* Executable buffers can't start or end on a 4GB boundary */
+	if (!(color & PANFROST_BO_NOEXEC)) {
+		u64 next_seg;
+
+		if ((*start & PFN_4G_MASK) == 0)
+			(*start)++;
+
+		if ((*end & PFN_4G_MASK) == 0)
+			(*end)--;
+
+		next_seg = ALIGN(*start, PFN_4G);
+		if (next_seg - *start <= PFN_16M)
+			*start = next_seg + 1;
+
+		*end = min(*end, ALIGN(*start, PFN_4G) - 1);
+	}
+}
+
+struct panfrost_mmu *panfrost_mmu_ctx_create(struct panfrost_device *pfdev)
+{
+	struct panfrost_mmu *mmu;
+
+	mmu = kzalloc(sizeof(*mmu), GFP_KERNEL);
+	if (!mmu)
+		return ERR_PTR(-ENOMEM);
+
+	mmu->pfdev = pfdev;
+	spin_lock_init(&mmu->mm_lock);
+
+	/* 4G enough for now. can be 48-bit */
+	drm_mm_init(&mmu->mm, SZ_32M >> PAGE_SHIFT, (SZ_4G - SZ_32M) >> PAGE_SHIFT);
+	mmu->mm.color_adjust = panfrost_drm_mm_color_adjust;
+
+	INIT_LIST_HEAD(&mmu->list);
+	mmu->as = -1;
+
+	mmu->pgtbl_cfg = (struct io_pgtable_cfg) {
+		.pgsize_bitmap	= SZ_4K | SZ_2M,
+		.ias		= FIELD_GET(0xff, pfdev->features.mmu_features),
+		.oas		= FIELD_GET(0xff00, pfdev->features.mmu_features),
+		.coherent_walk	= pfdev->coherent,
+		.tlb		= &mmu_tlb_ops,
+		.iommu_dev	= pfdev->dev,
+	};
+
+	mmu->pgtbl_ops = alloc_io_pgtable_ops(ARM_MALI_LPAE, &mmu->pgtbl_cfg,
+					      mmu);
+	if (!mmu->pgtbl_ops) {
+		kfree(mmu);
+		return ERR_PTR(-EINVAL);
+	}
+
+	kref_init(&mmu->refcount);
+
+	return mmu;
+}
+
 static const char *access_type_name(struct panfrost_device *pfdev,
 		u32 fault_status)
 {