All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Christian König" <deathsimple-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
Subject: [PATCH 6/7] drm/amdgpu: increase fragmentation size for Vega10
Date: Tue, 23 May 2017 18:52:53 +0200	[thread overview]
Message-ID: <1495558374-3513-7-git-send-email-deathsimple@vodafone.de> (raw)
In-Reply-To: <1495558374-3513-1-git-send-email-deathsimple-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>

From: Christian König <christian.koenig@amd.com>

The fragment bits work differently for Vega10 compared to previous generations.

Increase the fragment size to 2MB for now to better handle that.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 5 +++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c  | 5 +++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h  | 4 +++-
 3 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index f6edf68..83c4d0c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -548,8 +548,9 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 		dev_info.virtual_address_offset = AMDGPU_VA_RESERVED_SIZE;
 		dev_info.virtual_address_max = (uint64_t)adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE;
 		dev_info.virtual_address_alignment = max((int)PAGE_SIZE, AMDGPU_GPU_PAGE_SIZE);
-		dev_info.pte_fragment_size = (1 << AMDGPU_LOG2_PAGES_PER_FRAG) *
-					     AMDGPU_GPU_PAGE_SIZE;
+		dev_info.pte_fragment_size =
+			(1 << AMDGPU_LOG2_PAGES_PER_FRAG(adev)) *
+			AMDGPU_GPU_PAGE_SIZE;
 		dev_info.gart_page_size = AMDGPU_GPU_PAGE_SIZE;
 
 		dev_info.cu_active_number = adev->gfx.cu_info.number;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 062dadc..f7afdfa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -1262,8 +1262,9 @@ static int amdgpu_vm_frag_ptes(struct amdgpu_pte_update_params	*params,
 	 */
 
 	/* SI and newer are optimized for 64KB */
-	uint64_t frag_flags = AMDGPU_PTE_FRAG(AMDGPU_LOG2_PAGES_PER_FRAG);
-	uint64_t frag_align = 1 << AMDGPU_LOG2_PAGES_PER_FRAG;
+	unsigned pages_per_frag = AMDGPU_LOG2_PAGES_PER_FRAG(params->adev);
+	uint64_t frag_flags = AMDGPU_PTE_FRAG(pages_per_frag);
+	uint64_t frag_align = 1 << pages_per_frag;
 
 	uint64_t frag_start = ALIGN(start, frag_align);
 	uint64_t frag_end = end & ~(frag_align - 1);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 8309bc7..0f83fe3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -51,7 +51,9 @@ struct amdgpu_bo_list_entry;
 #define AMDGPU_VM_PTB_ALIGN_SIZE   32768
 
 /* LOG2 number of continuous pages for the fragment field */
-#define AMDGPU_LOG2_PAGES_PER_FRAG 4
+#define AMDGPU_LOG2_PAGES_PER_FRAG(adev) \
+	((adev)->asic_type < CHIP_VEGA10 ? 4 : \
+	 (adev)->vm_manager.block_size)
 
 #define AMDGPU_PTE_VALID	(1ULL << 0)
 #define AMDGPU_PTE_SYSTEM	(1ULL << 1)
-- 
2.7.4

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

  parent reply	other threads:[~2017-05-23 16:52 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-23 16:52 Huge page support for Vega10 Christian König
     [not found] ` <1495558374-3513-1-git-send-email-deathsimple-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-05-23 16:52   ` [PATCH 1/7] drm/amdgpu: cleanup adjust_mc_addr handling v3 Christian König
     [not found]     ` <1495558374-3513-2-git-send-email-deathsimple-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-05-23 19:03       ` Felix Kuehling
     [not found]         ` <b5b14e2f-c587-a83b-25ff-22f7197c3da9-5C7GfCeVMHo@public.gmane.org>
2017-05-24 11:22           ` Christian König
2017-05-23 16:52   ` [PATCH 2/7] drm/amdgpu: add some extra VM error handling Christian König
2017-05-23 16:52   ` [PATCH 3/7] drm/amdgpu: Return EINVAL if no PT BO Christian König
2017-05-23 16:52   ` [PATCH 4/7] drm/amdgpu: cache the complete pde Christian König
2017-05-23 16:52   ` [PATCH 5/7] drm/amdgpu: stop joining VM PTE updates Christian König
2017-05-23 16:52   ` Christian König [this message]
2017-05-23 16:52   ` [PATCH 7/7] drm/amdgpu: enable huge page handling in the VM v2 Christian König
     [not found]     ` <1495558374-3513-8-git-send-email-deathsimple-ANTagKRnAhcb1SvskN2V4Q@public.gmane.org>
2017-05-23 19:29       ` Felix Kuehling
     [not found]         ` <bed31c06-f9de-b31e-9972-b4c719f8982e-5C7GfCeVMHo@public.gmane.org>
2017-05-30 11:51           ` SPAM //Re: " Christian König

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1495558374-3513-7-git-send-email-deathsimple@vodafone.de \
    --to=deathsimple-antagkrnahcb1svskn2v4q@public.gmane.org \
    --cc=amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.