aboutsummaryrefslogtreecommitdiffstats
path: root/meta-amd-bsp/recipes-kernel/linux/linux-yocto-4.14.71/3622-drm-amdgpu-Map-all-visible-VRAM-at-startup.patch
blob: 13818287ccb0305baead3307d8109a8edad011e2 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
From 06a8fa3e504a01c86accc9f035f2461e01697f28 Mon Sep 17 00:00:00 2001
From: Amber Lin <Amber.Lin@amd.com>
Date: Tue, 27 Feb 2018 10:01:59 -0500
Subject: [PATCH 3622/4131] drm/amdgpu: Map all visible VRAM at startup
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When using CPU to update page table, we need to kmap all the PDs/PTs after
they are allocated and that requires a TLB shot down on each CPU, which is
quite heavy.

Instead, we map the whole visible VRAM to a kernel address at once. Pages
can be obtained from the offset.

v2: move the mapping base from gmc to amdgpu_mman structure, and the
    implementation in amdgpu_ttm_* functions

Change-Id: I56574bd544dae273da50e8b5dd6894cd5d9454bd
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Kalyan Alle <kalyan.alle@amd.com>

Conflicts:
      drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 17 +++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h |  1 +
 2 files changed, 18 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 36a3d7e..1638a09 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -677,6 +677,7 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_
 	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct ttm_mem_reg backup;
+	struct drm_mm_node *mm_node = mem->mm_node;
 
 	backup = *mem;
 	mem->bus.addr = NULL;
@@ -699,6 +700,15 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_
 		/* check if it's visible */
 		if ((mem->bus.offset + mem->bus.size) > adev->gmc.visible_vram_size)
 			return -EINVAL;
+		/* Only physically contiguous buffers apply. In a contiguous
+		 * buffer, size of the first mm_node would match the number of
+		 * pages in ttm_mem_reg.
+		 */
+		if (adev->mman.aper_base_kaddr &&
+		    (mm_node->size == mem->num_pages))
+			mem->bus.addr = (u8 *)adev->mman.aper_base_kaddr +
+					mem->bus.offset;
+
 		mem->bus.base = adev->gmc.aper_base;
 		mem->bus.is_iomem = true;
 		break;
@@ -1638,6 +1648,10 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 
 	/* Change the size here instead of the init above so only lpfn is affected */
 	amdgpu_ttm_set_active_vram_size(adev, adev->gmc.visible_vram_size);
+#ifdef CONFIG_64BIT
+	adev->mman.aper_base_kaddr = ioremap_wc(adev->gmc.aper_base,
+						adev->gmc.visible_vram_size);
+#endif
 
 	/*
 	 *The reserved vram for firmware must be pinned to the specified
@@ -1738,6 +1752,9 @@ void amdgpu_ttm_fini(struct amdgpu_device *adev)
 
 	amdgpu_bo_free_kernel(&adev->stolen_vga_memory, NULL, NULL);
 	amdgpu_ttm_fw_reserve_vram_fini(adev);
+	if (adev->mman.aper_base_kaddr)
+		iounmap(adev->mman.aper_base_kaddr);
+	adev->mman.aper_base_kaddr = NULL;
 
 	amdgpu_ssg_fini(adev);
 	amdgpu_direct_gma_fini(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
index 92aba8b..eaf0aa8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -48,6 +48,7 @@ struct amdgpu_mman {
 	struct ttm_bo_device		bdev;
 	bool				mem_global_referenced;
 	bool				initialized;
+	void __iomem			*aper_base_kaddr;
 
 #if defined(CONFIG_DEBUG_FS)
 	struct dentry			*debugfs_entries[8];
-- 
2.7.4