Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Brost <matthew.brost@intel.com>
To: intel-xe@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Cc: himal.prasad.ghimiray@intel.com, apopple@nvidia.com,
	airlied@gmail.com, thomas.hellstrom@linux.intel.com,
	simona.vetter@ffwll.ch, felix.kuehling@amd.com, dakr@kernel.org
Subject: [PATCH v7 05/32] drm/xe/bo: Introduce xe_bo_put_async
Date: Wed,  5 Mar 2025 17:26:30 -0800	[thread overview]
Message-ID: <20250306012657.3505757-6-matthew.brost@intel.com> (raw)
In-Reply-To: <20250306012657.3505757-1-matthew.brost@intel.com>

From: Thomas Hellström <thomas.hellstrom@linux.intel.com>

Introduce xe_bo_put_async to put a bo where the context is such that
the bo destructor can't run due to lockdep problems or atomic context.

If the put is the final put, freeing will be done from a work item.

v5:
 - Kerenl doc for xe_bo_put_async (Thomas)
v7:
 - Fix kernel doc (CI)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Tested-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
---
 drivers/gpu/drm/xe/xe_bo.c           | 25 +++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_bo.h           | 19 +++++++++++++++++++
 drivers/gpu/drm/xe/xe_device.c       |  3 +++
 drivers/gpu/drm/xe/xe_device_types.h |  8 ++++++++
 4 files changed, 55 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index d6f44540820d..0305e4bdb18c 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -2828,6 +2828,31 @@ void xe_bo_put_commit(struct llist_head *deferred)
 		drm_gem_object_free(&bo->ttm.base.refcount);
 }
 
+static void xe_bo_dev_work_func(struct work_struct *work)
+{
+	struct xe_bo_dev *bo_dev = container_of(work, typeof(*bo_dev), async_free);
+
+	xe_bo_put_commit(&bo_dev->async_list);
+}
+
+/**
+ * xe_bo_dev_init() - Initialize BO dev to manage async BO freeing
+ * @bo_dev: The BO dev structure
+ */
+void xe_bo_dev_init(struct xe_bo_dev *bo_dev)
+{
+	INIT_WORK(&bo_dev->async_free, xe_bo_dev_work_func);
+}
+
+/**
+ * xe_bo_dev_fini() - Finalize BO dev managing async BO freeing
+ * @bo_dev: The BO dev structure
+ */
+void xe_bo_dev_fini(struct xe_bo_dev *bo_dev)
+{
+	flush_work(&bo_dev->async_free);
+}
+
 void xe_bo_put(struct xe_bo *bo)
 {
 	struct xe_tile *tile;
diff --git a/drivers/gpu/drm/xe/xe_bo.h b/drivers/gpu/drm/xe/xe_bo.h
index 2a7240c45511..b128becbe8cc 100644
--- a/drivers/gpu/drm/xe/xe_bo.h
+++ b/drivers/gpu/drm/xe/xe_bo.h
@@ -345,6 +345,25 @@ xe_bo_put_deferred(struct xe_bo *bo, struct llist_head *deferred)
 
 void xe_bo_put_commit(struct llist_head *deferred);
 
+/**
+ * xe_bo_put_async() - Put BO async
+ * @bo: The bo to put.
+ *
+ * Put BO async, the final put is deferred to a worker to exit an IRQ context.
+ */
+static inline void
+xe_bo_put_async(struct xe_bo *bo)
+{
+	struct xe_bo_dev *bo_device = &xe_bo_device(bo)->bo_device;
+
+	if (xe_bo_put_deferred(bo, &bo_device->async_list))
+		schedule_work(&bo_device->async_free);
+}
+
+void xe_bo_dev_init(struct xe_bo_dev *bo_device);
+
+void xe_bo_dev_fini(struct xe_bo_dev *bo_device);
+
 struct sg_table *xe_bo_sg(struct xe_bo *bo);
 
 /*
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 539da525aa14..5d79b439dd62 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -388,6 +388,8 @@ static void xe_device_destroy(struct drm_device *dev, void *dummy)
 {
 	struct xe_device *xe = to_xe_device(dev);
 
+	xe_bo_dev_fini(&xe->bo_device);
+
 	if (xe->preempt_fence_wq)
 		destroy_workqueue(xe->preempt_fence_wq);
 
@@ -428,6 +430,7 @@ struct xe_device *xe_device_create(struct pci_dev *pdev,
 	if (WARN_ON(err))
 		goto err;
 
+	xe_bo_dev_init(&xe->bo_device);
 	err = drmm_add_action_or_reset(&xe->drm, xe_device_destroy, NULL);
 	if (err)
 		goto err;
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 604a69c494e5..2f2a51fd9516 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -527,6 +527,14 @@ struct xe_device {
 		int mode;
 	} wedged;
 
+	/** @bo_device: Struct to control async free of BOs */
+	struct xe_bo_dev {
+		/** @bo_device.async_free: Free worker */
+		struct work_struct async_free;
+		/** @bo_device.async_list: List of BOs to be freed */
+		struct llist_head async_list;
+	} bo_device;
+
 	/** @pmu: performance monitoring unit */
 	struct xe_pmu pmu;
 
-- 
2.34.1


  parent reply	other threads:[~2025-03-06  1:26 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-03-06  1:26 [PATCH v7 00/32] Introduce GPU SVM and Xe SVM implementation Matthew Brost
2025-03-06  1:26 ` [PATCH v7 01/32] drm/xe: Retry BO allocation Matthew Brost
2025-03-06  1:26 ` [PATCH v7 02/32] mm/migrate: Add migrate_device_pfns Matthew Brost
2025-03-06  1:26 ` [PATCH v7 03/32] mm/migrate: Trylock device page in do_swap_page Matthew Brost
2025-03-06  1:26 ` [PATCH v7 04/32] drm/pagemap: Add DRM pagemap Matthew Brost
2025-03-06  1:26 ` Matthew Brost [this message]
2025-03-06  1:26 ` [PATCH v7 06/32] drm/gpusvm: Add support for GPU Shared Virtual Memory Matthew Brost
2025-03-06  1:26 ` [PATCH v7 07/32] drm/xe: Select DRM_GPUSVM Kconfig Matthew Brost
2025-03-06  1:26 ` [PATCH v7 08/32] drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR Matthew Brost
2025-03-06  1:26 ` [PATCH v7 09/32] drm/xe: Add SVM init / close / fini to faulting VMs Matthew Brost
2025-03-06  1:26 ` [PATCH v7 10/32] drm/xe: Add dma_addr res cursor Matthew Brost
2025-03-06  1:26 ` [PATCH v7 11/32] drm/xe: Nuke VM's mapping upon close Matthew Brost
2025-03-06  1:26 ` [PATCH v7 12/32] drm/xe: Add SVM range invalidation and page fault Matthew Brost
2025-03-06  1:26 ` [PATCH v7 13/32] drm/gpuvm: Add DRM_GPUVA_OP_DRIVER Matthew Brost
2025-03-06  1:26 ` [PATCH v7 14/32] drm/xe: Add (re)bind to SVM page fault handler Matthew Brost
2025-03-06  1:26 ` [PATCH v7 15/32] drm/xe: Add SVM garbage collector Matthew Brost
2025-03-06  1:26 ` [PATCH v7 16/32] drm/xe: Add unbind to " Matthew Brost
2025-03-06  1:26 ` [PATCH v7 17/32] drm/xe: Do not allow CPU address mirror VMA unbind if Matthew Brost
2025-03-06  1:26 ` [PATCH v7 18/32] drm/xe: Enable CPU address mirror uAPI Matthew Brost
2025-03-06  1:26 ` [PATCH v7 19/32] drm/xe/uapi: Add DRM_XE_QUERY_CONFIG_FLAG_HAS_CPU_ADDR_MIRROR Matthew Brost
2025-03-06  1:26 ` [PATCH v7 20/32] drm/xe: Add migrate layer functions for SVM support Matthew Brost
2025-03-06  1:26 ` [PATCH v7 21/32] drm/xe: Add SVM device memory mirroring Matthew Brost
2025-03-06  1:26 ` [PATCH v7 22/32] drm/xe: Add drm_gpusvm_devmem to xe_bo Matthew Brost
2025-03-06  1:26 ` [PATCH v7 23/32] drm/xe: Add drm_pagemap ops to SVM Matthew Brost
2025-03-06  1:26 ` [PATCH v7 24/32] drm/xe: Add GPUSVM device memory copy vfunc functions Matthew Brost
2025-03-06  1:26 ` [PATCH v7 25/32] drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc Matthew Brost
2025-03-06  1:26 ` [PATCH v7 26/32] drm/xe: Add Xe SVM devmem_release " Matthew Brost
2025-03-06  1:26 ` [PATCH v7 27/32] drm/xe: Add SVM VRAM migration Matthew Brost
2025-03-06  1:26 ` [PATCH v7 28/32] drm/xe: Basic SVM BO eviction Matthew Brost
2025-03-06  1:26 ` [PATCH v7 29/32] drm/xe: Add SVM debug Matthew Brost
2025-03-06  1:26 ` [PATCH v7 30/32] drm/xe: Add modparam for SVM notifier size Matthew Brost
2025-03-06  1:26 ` [PATCH v7 31/32] drm/xe: Add always_migrate_to_vram modparam Matthew Brost
2025-03-06  1:26 ` [PATCH v7 32/32] drm/doc: gpusvm: Add GPU SVM documentation Matthew Brost
2025-03-06  5:45   ` Alistair Popple
2025-03-06  6:08     ` Matthew Brost
2025-03-06  1:54 ` ✓ CI.Patch_applied: success for Introduce GPU SVM and Xe SVM implementation (rev7) Patchwork
2025-03-06  1:54 ` ✗ CI.checkpatch: warning " Patchwork
2025-03-06  1:56 ` ✓ CI.KUnit: success " Patchwork
2025-03-06  2:12 ` ✓ CI.Build: " Patchwork
2025-03-06  2:14 ` ✓ CI.Hooks: " Patchwork
2025-03-06  2:16 ` ✗ CI.checksparse: warning " Patchwork
2025-03-06  2:51 ` ✓ Xe.CI.BAT: success " Patchwork
2025-03-06  8:25 ` ✗ Xe.CI.Full: failure " Patchwork
2025-03-06  9:57   ` Matthew Brost
2025-03-06  9:22 ` ✓ CI.Patch_applied: success for Introduce GPU SVM and Xe SVM implementation (rev8) Patchwork
2025-03-06  9:23 ` ✗ CI.checkpatch: warning " Patchwork
2025-03-06  9:24 ` ✓ CI.KUnit: success " Patchwork
2025-03-06  9:41 ` ✓ CI.Build: " Patchwork
2025-03-06  9:43 ` ✓ CI.Hooks: " Patchwork
2025-03-06  9:44 ` ✗ CI.checksparse: warning " Patchwork
2025-03-06 10:17 ` ✗ Xe.CI.BAT: failure " Patchwork
2025-03-06 18:41 ` ✗ Xe.CI.Full: " Patchwork

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=20250306012657.3505757-6-matthew.brost@intel.com \
    --to=matthew.brost@intel.com \
    --cc=airlied@gmail.com \
    --cc=apopple@nvidia.com \
    --cc=dakr@kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=felix.kuehling@amd.com \
    --cc=himal.prasad.ghimiray@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=simona.vetter@ffwll.ch \
    --cc=thomas.hellstrom@linux.intel.com \
    /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