Blob Blame History Raw
From: Amber Lin <Amber.Lin@amd.com>
Date: Tue, 27 Feb 2018 10:01:59 -0500
Subject: drm/amdgpu: Map all visible VRAM at startup
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Git-commit: f8f4b9a679e52298b14358db22c37a135e7086ec
Patch-mainline: v4.17-rc1
References: FATE#326289 FATE#326079 FATE#326049 FATE#322398 FATE#326166

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

Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Petr Tesarik <ptesarik@suse.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c |   17 +++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h |    1 +
 2 files changed, 18 insertions(+)

--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -621,6 +621,7 @@ static int amdgpu_ttm_io_mem_reserve(str
 {
 	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
+	struct drm_mm_node *mm_node = mem->mm_node;
 
 	mem->bus.addr = NULL;
 	mem->bus.offset = 0;
@@ -640,6 +641,15 @@ static int amdgpu_ttm_io_mem_reserve(str
 		/* 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;
@@ -1403,6 +1413,10 @@ int amdgpu_ttm_init(struct amdgpu_device
 
 	/* 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
@@ -1495,6 +1509,9 @@ void amdgpu_ttm_fini(struct amdgpu_devic
 	amdgpu_ttm_debugfs_fini(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;
 
 	ttm_bo_clean_mm(&adev->mman.bdev, TTM_PL_VRAM);
 	ttm_bo_clean_mm(&adev->mman.bdev, TTM_PL_TT);
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h
@@ -44,6 +44,7 @@ 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];