AMD-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Sasha Levin <sashal@kernel.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: Sasha Levin <sashal@kernel.org>,
	Philip Yang <Philip.Yang@amd.com>,
	Felix Kuehling <Felix.Kuehling@amd.com>,
	Xinhui.Pan@amd.com, amd-gfx@lists.freedesktop.org,
	dri-devel@lists.freedesktop.org, daniel@ffwll.ch,
	Alex Deucher <alexander.deucher@amd.com>,
	airlied@gmail.com, christian.koenig@amd.com
Subject: [PATCH AUTOSEL 6.2 12/60] drm/amdkfd: Page aligned memory reserve size
Date: Sun, 26 Feb 2023 20:59:57 -0500	[thread overview]
Message-ID: <20230227020045.1045105-12-sashal@kernel.org> (raw)
In-Reply-To: <20230227020045.1045105-1-sashal@kernel.org>

From: Philip Yang <Philip.Yang@amd.com>

[ Upstream commit 0c2dece8fb541ab07b68c3312a1065fa9c927a81 ]

Use page aligned size to reserve memory usage because page aligned TTM
BO size is used to unreserve memory usage, otherwise no page aligned
size causes memory usage accounting unbalanced.

Change vram_used definition type to int64_t to be able to trigger
WARN_ONCE(adev && adev->kfd.vram_used < 0, "..."), to help debug the
accounting issue with warning and backtrace.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h       |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 12 +++++++-----
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c         |  9 +++++++--
 3 files changed, 15 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 0040deaf8a83a..90a5254ec1387 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -97,7 +97,7 @@ struct amdgpu_amdkfd_fence {
 
 struct amdgpu_kfd_dev {
 	struct kfd_dev *dev;
-	uint64_t vram_used;
+	int64_t vram_used;
 	uint64_t vram_used_aligned;
 	bool init_complete;
 	struct work_struct reset_work;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 3b5c53712d319..05b884fe0a927 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1612,6 +1612,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	struct amdgpu_bo *bo;
 	struct drm_gem_object *gobj = NULL;
 	u32 domain, alloc_domain;
+	uint64_t aligned_size;
 	u64 alloc_flags;
 	int ret;
 
@@ -1667,22 +1668,23 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	 * the memory.
 	 */
 	if ((*mem)->aql_queue)
-		size = size >> 1;
+		size >>= 1;
+	aligned_size = PAGE_ALIGN(size);
 
 	(*mem)->alloc_flags = flags;
 
 	amdgpu_sync_create(&(*mem)->sync);
 
-	ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, flags);
+	ret = amdgpu_amdkfd_reserve_mem_limit(adev, aligned_size, flags);
 	if (ret) {
 		pr_debug("Insufficient memory\n");
 		goto err_reserve_limit;
 	}
 
 	pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s\n",
-			va, size, domain_string(alloc_domain));
+			va, (*mem)->aql_queue ? size << 1 : size, domain_string(alloc_domain));
 
-	ret = amdgpu_gem_object_create(adev, size, 1, alloc_domain, alloc_flags,
+	ret = amdgpu_gem_object_create(adev, aligned_size, 1, alloc_domain, alloc_flags,
 				       bo_type, NULL, &gobj);
 	if (ret) {
 		pr_debug("Failed to create BO on domain %s. ret %d\n",
@@ -1739,7 +1741,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	/* Don't unreserve system mem limit twice */
 	goto err_reserve_limit;
 err_bo_create:
-	amdgpu_amdkfd_unreserve_mem_limit(adev, size, flags);
+	amdgpu_amdkfd_unreserve_mem_limit(adev, aligned_size, flags);
 err_reserve_limit:
 	mutex_destroy(&(*mem)->lock);
 	if (gobj)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 6d291aa6386bd..f79b8e964140e 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -1127,8 +1127,13 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 	}
 
 	/* Update the VRAM usage count */
-	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)
-		WRITE_ONCE(pdd->vram_usage, pdd->vram_usage + args->size);
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		uint64_t size = args->size;
+
+		if (flags & KFD_IOC_ALLOC_MEM_FLAGS_AQL_QUEUE_MEM)
+			size >>= 1;
+		WRITE_ONCE(pdd->vram_usage, pdd->vram_usage + PAGE_ALIGN(size));
+	}
 
 	mutex_unlock(&p->mutex);
 
-- 
2.39.0


  parent reply	other threads:[~2023-02-27  2:01 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20230227020045.1045105-1-sashal@kernel.org>
2023-02-27  1:59 ` [PATCH AUTOSEL 6.2 03/60] drm/amd/display: Reduce expected sdp bandwidth for dcn321 Sasha Levin
2023-02-27  1:59 ` [PATCH AUTOSEL 6.2 04/60] drm/amd/display: Revert Reduce delay when sink device not able to ACK 00340h write Sasha Levin
2023-02-27  1:59 ` [PATCH AUTOSEL 6.2 05/60] drm/amd/display: Fix potential null-deref in dm_resume Sasha Levin
2023-02-27  1:59 ` [PATCH AUTOSEL 6.2 09/60] drm/amd/display: Defer DIG FIFO disable after VID stream enable Sasha Levin
2023-02-27  1:59 ` [PATCH AUTOSEL 6.2 10/60] drm/radeon: free iio for atombios when driver shutdown Sasha Levin
2023-02-27  1:59 ` [PATCH AUTOSEL 6.2 11/60] drm/amd: Avoid BUG() for case of SRIOV missing IP version Sasha Levin
2023-02-27  1:59 ` Sasha Levin [this message]
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 15/60] drm/amd: Avoid ASSERT for some message failures Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 16/60] drm: amd: display: Fix memory leakage Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 17/60] drm/amd/display: fix mapping to non-allocated address Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 28/60] drm/amd/display: Set hvm_enabled flag for S/G mode Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 38/60] Revert "drm/amdgpu: TA unload messages are not actually sent to psp when amdgpu is uninstalled" Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 39/60] drm/amd/display: fix FCLK pstate change underflow Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 45/60] drm/amd/display: Do not set DRR on pipe commit Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 46/60] drm/amd/display: Do not commit pipe when updating DRR Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 50/60] drm/amd/display: Move DCN314 DOMAIN power control to DMCUB Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 51/60] drm/amd/display: Enable P-state validation checks for DCN314 Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 53/60] drm/amd/display: Disable HUBP/DPP PG on DCN314 for now Sasha Levin
2023-02-27  2:00 ` [PATCH AUTOSEL 6.2 54/60] drm/amd/display: disable SubVP + DRR to prevent underflow Sasha Levin

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=20230227020045.1045105-12-sashal@kernel.org \
    --to=sashal@kernel.org \
    --cc=Felix.Kuehling@amd.com \
    --cc=Philip.Yang@amd.com \
    --cc=Xinhui.Pan@amd.com \
    --cc=airlied@gmail.com \
    --cc=alexander.deucher@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=christian.koenig@amd.com \
    --cc=daniel@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox