drm/radeon/kms: set gart pages to invalid on unbind and point to dummy page

this uses a new entrypoint to invalidate gart entries instead of using 0.
Changed to rather than pointing to 0 address point empty entry to dummy
page. This might help to avoid hard lockup if for some wrong
reasons GPU try to access unmapped GART entry.

I'm not 100% sure this is going to work, we probably need to allocate
a dummy page and point all the GTT entries at it similiar to what AGP does.
but we can test this first I suppose.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

10 files changed, 40 insertions, 3 deletions

diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index c2f9752..3368920 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -93,6 +93,7 @@ int evergreen_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
+	radeon_gart_restore(rdev);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
 				ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index 346ae3d..bc7d9e9 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -197,6 +197,7 @@ int r100_pci_gart_enable(struct radeon_device *rdev)
 {
 	uint32_t tmp;
 
+	radeon_gart_restore(rdev);
 	/* discard memory request outside of configured range */
 	tmp = RREG32(RADEON_AIC_CNTL) | RADEON_DIS_OUT_OF_PCI_GART_ACCESS;
 	WREG32(RADEON_AIC_CNTL, tmp);
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index 5897bd0..654aca1 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -117,6 +117,7 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
+	radeon_gart_restore(rdev);
 	/* discard memory request outside of configured range */
 	tmp = RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
 	WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp);
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index c9723dae..4facbab 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -416,6 +416,7 @@ int r600_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
+	radeon_gart_restore(rdev);
 
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 9eeca6f..993cdf2 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -1145,6 +1145,7 @@ static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v)
 /* AGP */
 extern void radeon_agp_disable(struct radeon_device *rdev);
 extern int radeon_gart_table_vram_pin(struct radeon_device *rdev);
+extern void radeon_gart_restore(struct radeon_device *rdev);
 extern int radeon_modeset_init(struct radeon_device *rdev);
 extern void radeon_modeset_fini(struct radeon_device *rdev);
 extern bool radeon_card_posted(struct radeon_device *rdev);
@@ -1269,7 +1270,6 @@ extern void r600_ring_init(struct radeon_device *rdev, unsigned ring_size);
 extern int r600_cp_resume(struct radeon_device *rdev);
 extern void r600_cp_fini(struct radeon_device *rdev);
 extern int r600_count_pipe_bits(uint32_t val);
-extern int r600_gart_clear_page(struct radeon_device *rdev, int i);
 extern int r600_mc_wait_for_idle(struct radeon_device *rdev);
 extern int r600_pcie_gart_init(struct radeon_device *rdev);
 extern void r600_pcie_gart_tlb_flush(struct radeon_device *rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index c224c1d..fb55faf 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -238,6 +238,8 @@ bool radeon_boot_test_post_card(struct radeon_device *rdev)
 
 int radeon_dummy_page_init(struct radeon_device *rdev)
 {
+	if (rdev->dummy_page.page)
+		return 0;
 	rdev->dummy_page.page = alloc_page(GFP_DMA32 | GFP_KERNEL | __GFP_ZERO);
 	if (rdev->dummy_page.page == NULL)
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index e73d56e..1770d3c 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -139,6 +139,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 	unsigned t;
 	unsigned p;
 	int i, j;
+	u64 page_base;
 
 	if (!rdev->gart.ready) {
 		WARN(1, "trying to unbind memory to unitialized GART !\n");
@@ -151,9 +152,11 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 			pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p],
 				       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
 			rdev->gart.pages[p] = NULL;
-			rdev->gart.pages_addr[p] = 0;
+			rdev->gart.pages_addr[p] = rdev->dummy_page.addr;
+			page_base = rdev->gart.pages_addr[p];
 			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-				radeon_gart_set_page(rdev, t, 0);
+				radeon_gart_set_page(rdev, t, page_base);
+				page_base += RADEON_GPU_PAGE_SIZE;
 			}
 		}
 	}
@@ -199,8 +202,26 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 	return 0;
 }
 
+void radeon_gart_restore(struct radeon_device *rdev)
+{
+	int i, j, t;
+	u64 page_base;
+
+	for (i = 0, t = 0; i < rdev->gart.num_cpu_pages; i++) {
+		page_base = rdev->gart.pages_addr[i];
+		for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
+			radeon_gart_set_page(rdev, t, page_base);
+			page_base += RADEON_GPU_PAGE_SIZE;
+		}
+	}
+	mb();
+	radeon_gart_tlb_flush(rdev);
+}
+
 int radeon_gart_init(struct radeon_device *rdev)
 {
+	int r, i;
+
 	if (rdev->gart.pages) {
 		return 0;
 	}
@@ -209,6 +230,9 @@ int radeon_gart_init(struct radeon_device *rdev)
 		DRM_ERROR("Page size is smaller than GPU page size!\n");
 		return -EINVAL;
 	}
+	r = radeon_dummy_page_init(rdev);
+	if (r)
+		return r;
 	/* Compute table size */
 	rdev->gart.num_cpu_pages = rdev->mc.gtt_size / PAGE_SIZE;
 	rdev->gart.num_gpu_pages = rdev->mc.gtt_size / RADEON_GPU_PAGE_SIZE;
@@ -227,6 +251,10 @@ int radeon_gart_init(struct radeon_device *rdev)
 		radeon_gart_fini(rdev);
 		return -ENOMEM;
 	}
+	/* set GART entry to point to the dummy page by default */
+	for (i = 0; i < rdev->gart.num_cpu_pages; i++) {
+		rdev->gart.pages_addr[i] = rdev->dummy_page.addr;
+	}
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index 287fceb..1e4582e 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -113,6 +113,7 @@ int rs400_gart_enable(struct radeon_device *rdev)
 	uint32_t size_reg;
 	uint32_t tmp;
 
+	radeon_gart_restore(rdev);
 	tmp = RREG32_MC(RS690_AIC_CTRL_SCRATCH);
 	tmp |= RS690_DIS_OUT_OF_PCI_GART_ACCESS;
 	WREG32_MC(RS690_AIC_CTRL_SCRATCH, tmp);
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index 979b000..28c8690 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -213,6 +213,7 @@ int rs600_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
+	radeon_gart_restore(rdev);
 	/* Enable bus master */
 	tmp = RREG32(R_00004C_BUS_CNTL) & C_00004C_BUS_MASTER_DIS;
 	WREG32(R_00004C_BUS_CNTL, tmp);
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index 3ecd138..6f1f4ab 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -56,6 +56,7 @@ int rv770_pcie_gart_enable(struct radeon_device *rdev)
 	r = radeon_gart_table_vram_pin(rdev);
 	if (r)
 		return r;
+	radeon_gart_restore(rdev);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING |
 				ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |