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
Subject: [PATCH v2] drm/xe: Simplify locking in new_vma
Date: Mon, 17 Jun 2024 17:38:59 -0700	[thread overview]
Message-ID: <20240618003859.3239239-1-matthew.brost@intel.com> (raw)

Rather than acquiring and dropping the VM / BO dma-resv around
xe_vma_create and do the same thing upon adding preempt fences or an
error, hold these locks through the entire new_vma() function.

v2:
 - Rebase (CI)

Cc: Fei Yang <fei.yang@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_vm.c | 41 +++++++++++++++++---------------------
 1 file changed, 18 insertions(+), 23 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 61d4d95a5377..5b166fa03684 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -180,16 +180,14 @@ static int add_preempt_fences(struct xe_vm *vm, struct xe_bo *bo)
 	struct xe_exec_queue *q;
 	int err;
 
+	xe_bo_assert_held(bo);
+
 	if (!vm->preempt.num_exec_queues)
 		return 0;
 
-	err = xe_bo_lock(bo, true);
-	if (err)
-		return err;
-
 	err = dma_resv_reserve_fences(bo->ttm.base.resv, vm->preempt.num_exec_queues);
 	if (err)
-		goto out_unlock;
+		return err;
 
 	list_for_each_entry(q, &vm->preempt.exec_queues, lr.link)
 		if (q->lr.pfence) {
@@ -198,9 +196,7 @@ static int add_preempt_fences(struct xe_vm *vm, struct xe_bo *bo)
 					   DMA_RESV_USAGE_BOOKKEEP);
 		}
 
-out_unlock:
-	xe_bo_unlock(bo);
-	return err;
+	return 0;
 }
 
 static void resume_and_reinstall_preempt_fences(struct xe_vm *vm,
@@ -2140,7 +2136,7 @@ static struct xe_vma *new_vma(struct xe_vm *vm, struct drm_gpuva_op_map *op,
 	struct xe_bo *bo = op->gem.obj ? gem_to_xe_bo(op->gem.obj) : NULL;
 	struct drm_exec exec;
 	struct xe_vma *vma;
-	int err;
+	int err = 0;
 
 	lockdep_assert_held_write(&vm->lock);
 
@@ -2165,23 +2161,22 @@ static struct xe_vma *new_vma(struct xe_vm *vm, struct drm_gpuva_op_map *op,
 	vma = xe_vma_create(vm, bo, op->gem.offset,
 			    op->va.addr, op->va.addr +
 			    op->va.range - 1, pat_index, flags);
-	if (bo)
-		drm_exec_fini(&exec);
+	if (IS_ERR(vma))
+		goto err_unlock;
 
-	if (xe_vma_is_userptr(vma)) {
+	if (xe_vma_is_userptr(vma))
 		err = xe_vma_userptr_pin_pages(to_userptr_vma(vma));
-		if (err) {
-			prep_vma_destroy(vm, vma, false);
-			xe_vma_destroy_unlocked(vma);
-			return ERR_PTR(err);
-		}
-	} else if (!xe_vma_has_no_bo(vma) && !bo->vm) {
+	else if (!xe_vma_has_no_bo(vma) && !bo->vm)
 		err = add_preempt_fences(vm, bo);
-		if (err) {
-			prep_vma_destroy(vm, vma, false);
-			xe_vma_destroy_unlocked(vma);
-			return ERR_PTR(err);
-		}
+
+err_unlock:
+	if (bo)
+		drm_exec_fini(&exec);
+
+	if (err) {
+		prep_vma_destroy(vm, vma, false);
+		xe_vma_destroy_unlocked(vma);
+		vma = ERR_PTR(err);
 	}
 
 	return vma;
-- 
2.34.1


             reply	other threads:[~2024-06-18  0:38 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-06-18  0:38 Matthew Brost [this message]
2024-06-18  0:43 ` ✓ CI.Patch_applied: success for drm/xe: Simplify locking in new_vma (rev2) Patchwork
2024-06-18  0:43 ` ✓ CI.checkpatch: " Patchwork
2024-06-18  0:44 ` ✓ CI.KUnit: " Patchwork
2024-06-18  0:56 ` ✓ CI.Build: " Patchwork
2024-06-18  0:58 ` ✗ CI.Hooks: failure " Patchwork
2024-06-18  0:59 ` ✓ CI.checksparse: success " Patchwork
2024-06-18  1:21 ` ✓ CI.BAT: " Patchwork
2024-06-18 16:46 ` ✗ CI.FULL: failure " Patchwork
2024-06-18 17:59 ` [PATCH v2] drm/xe: Simplify locking in new_vma Randhawa, Jagmeet

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=20240618003859.3239239-1-matthew.brost@intel.com \
    --to=matthew.brost@intel.com \
    --cc=intel-xe@lists.freedesktop.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