From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 1354BE95A8E for ; Mon, 9 Oct 2023 13:14:27 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id D3ABF10E0DD; Mon, 9 Oct 2023 13:14:26 +0000 (UTC) Received: from mgamail.intel.com (mgamail.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id B705610E0DD for ; Mon, 9 Oct 2023 13:14:24 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1696857264; x=1728393264; h=from:to:subject:date:message-id:in-reply-to:references: mime-version:content-transfer-encoding; bh=FNt8RODRttcMW7z05Xjq+3hgtmqJnl5Wz8LJG1Jd1qA=; b=jYDsfNl/umOq+8U3vUdQkSPmBADy0n2ZsYfZAfveXtYVuFdq/IhueO3K B9XPDLEXIqzFgdgqxpWm17Wk3ExtKp/yIg2u0uA6TG0qvOhe0VmbsyVlU lBam5rbZA2PluaP99t2Yd5LUQYL+VAOu8cYLscFqiBGBDO+m6xddrBXdX zOp73N6DpxOaTgyTDjH79VkFNhCo4/ZwJ9KToshxCo7G9tdRF+syLokHW gNkLPBU1ug876NLLwjnlCgsWmFozdN7SbnkY58uQmpyHVWGWROZGL7QZv GwNy2JiXn0/n26GjY07KHmvR/+eAvS+k6h3+bj/t3g9hVA3dJxRSMbDib w==; X-IronPort-AV: E=McAfee;i="6600,9927,10858"; a="415128981" X-IronPort-AV: E=Sophos;i="6.03,210,1694761200"; d="scan'208";a="415128981" Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 09 Oct 2023 06:14:02 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6600,9927,10858"; a="843695753" X-IronPort-AV: E=Sophos;i="6.03,210,1694761200"; d="scan'208";a="843695753" Received: from sorenthe-mobl.ger.corp.intel.com (HELO fedora..) ([10.249.254.151]) by fmsmga003-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 09 Oct 2023 06:14:01 -0700 From: =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= To: intel-xe@lists.freedesktop.org Date: Mon, 9 Oct 2023 15:13:32 +0200 Message-ID: <20231009131341.18355-2-thomas.hellstrom@linux.intel.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20231009131341.18355-1-thomas.hellstrom@linux.intel.com> References: <20231009131341.18355-1-thomas.hellstrom@linux.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Subject: [Intel-xe] [CI 02/10] drm/xe: Adjust to "drm/gpuvm: add common dma-resv per struct drm_gpuvm" X-BeenThere: intel-xe@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel Xe graphics driver List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-xe-bounces@lists.freedesktop.org Sender: "Intel-xe" Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_bo.c | 10 ++--- drivers/gpu/drm/xe/xe_bo.h | 2 +- drivers/gpu/drm/xe/xe_exec.c | 4 +- drivers/gpu/drm/xe/xe_migrate.c | 4 +- drivers/gpu/drm/xe/xe_pt.c | 6 +-- drivers/gpu/drm/xe/xe_vm.c | 72 ++++++++++++++++---------------- drivers/gpu/drm/xe/xe_vm.h | 4 -- drivers/gpu/drm/xe/xe_vm_types.h | 12 +++--- 8 files changed, 57 insertions(+), 57 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index 61789c0e88fb..772ffcd94101 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -501,9 +501,9 @@ static int xe_bo_trigger_rebind(struct xe_device *xe, struct xe_bo *bo, * that we indeed have it locked, put the vma an the * vm's notifier.rebind_list instead and scoop later. */ - if (dma_resv_trylock(&vm->resv)) + if (dma_resv_trylock(xe_vm_resv(vm))) vm_resv_locked = true; - else if (ctx->resv != &vm->resv) { + else if (ctx->resv != xe_vm_resv(vm)) { spin_lock(&vm->notifier.list_lock); if (!(vma->gpuva.flags & XE_VMA_DESTROYED)) list_move_tail(&vma->notifier.rebind_link, @@ -520,7 +520,7 @@ static int xe_bo_trigger_rebind(struct xe_device *xe, struct xe_bo *bo, &vm->rebind_list); if (vm_resv_locked) - dma_resv_unlock(&vm->resv); + dma_resv_unlock(xe_vm_resv(vm)); } } @@ -1340,7 +1340,7 @@ xe_bo_create_locked_range(struct xe_device *xe, } } - bo = __xe_bo_create_locked(xe, bo, tile, vm ? &vm->resv : NULL, + bo = __xe_bo_create_locked(xe, bo, tile, vm ? xe_vm_resv(vm) : NULL, vm && !xe_vm_in_fault_mode(vm) && flags & XE_BO_CREATE_USER_BIT ? &vm->lru_bulk_move : NULL, size, @@ -1656,7 +1656,7 @@ int xe_bo_validate(struct xe_bo *bo, struct xe_vm *vm, bool allow_res_evict) xe_vm_assert_held(vm); ctx.allow_res_evict = allow_res_evict; - ctx.resv = &vm->resv; + ctx.resv = xe_vm_resv(vm); } return ttm_bo_validate(&bo->ttm, &bo->placement, &ctx); diff --git a/drivers/gpu/drm/xe/xe_bo.h b/drivers/gpu/drm/xe/xe_bo.h index 78da28f4ea13..fe000a390cd0 100644 --- a/drivers/gpu/drm/xe/xe_bo.h +++ b/drivers/gpu/drm/xe/xe_bo.h @@ -159,7 +159,7 @@ void xe_bo_unlock(struct xe_bo *bo); static inline void xe_bo_unlock_vm_held(struct xe_bo *bo) { if (bo) { - XE_WARN_ON(bo->vm && bo->ttm.base.resv != &bo->vm->resv); + XE_WARN_ON(bo->vm && bo->ttm.base.resv != xe_vm_resv(bo->vm)); if (bo->vm) xe_vm_assert_held(bo->vm); else diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index 7d0381d41b32..7cf4215b2b2e 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -324,7 +324,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) /* Wait behind munmap style rebinds */ if (!xe_vm_no_dma_fences(vm)) { err = drm_sched_job_add_resv_dependencies(&job->drm, - &vm->resv, + xe_vm_resv(vm), DMA_RESV_USAGE_KERNEL); if (err) goto err_put_job; @@ -352,7 +352,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) xe_sched_job_arm(job); if (!xe_vm_no_dma_fences(vm)) { /* Block userptr invalidations / BO eviction */ - dma_resv_add_fence(&vm->resv, + dma_resv_add_fence(xe_vm_resv(vm), &job->drm.s_fence->finished, DMA_RESV_USAGE_BOOKKEEP); diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index 3049595e288a..02d7c6344baa 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -1133,7 +1133,7 @@ xe_migrate_update_pgtables_cpu(struct xe_migrate *m, DMA_RESV_USAGE_KERNEL)) return ERR_PTR(-ETIME); - if (wait_vm && !dma_resv_test_signaled(&vm->resv, + if (wait_vm && !dma_resv_test_signaled(xe_vm_resv(vm), DMA_RESV_USAGE_BOOKKEEP)) return ERR_PTR(-ETIME); @@ -1348,7 +1348,7 @@ xe_migrate_update_pgtables(struct xe_migrate *m, * trigger preempts before moving forward */ if (first_munmap_rebind) { - err = job_add_deps(job, &vm->resv, + err = job_add_deps(job, xe_vm_resv(vm), DMA_RESV_USAGE_BOOKKEEP); if (err) goto err_job; diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c index f7553a2415ac..15f7c9d5b311 100644 --- a/drivers/gpu/drm/xe/xe_pt.c +++ b/drivers/gpu/drm/xe/xe_pt.c @@ -870,7 +870,7 @@ static void xe_pt_commit_locks_assert(struct xe_vma *vma) else if (!xe_vma_is_null(vma)) dma_resv_assert_held(xe_vma_bo(vma)->ttm.base.resv); - dma_resv_assert_held(&vm->resv); + xe_vm_assert_held(vm); } static void xe_pt_commit_bind(struct xe_vma *vma, @@ -1332,7 +1332,7 @@ __xe_pt_bind_vma(struct xe_tile *tile, struct xe_vma *vma, struct xe_exec_queue } /* add shared fence now for pagetable delayed destroy */ - dma_resv_add_fence(&vm->resv, fence, !rebind && + dma_resv_add_fence(xe_vm_resv(vm), fence, !rebind && last_munmap_rebind ? DMA_RESV_USAGE_KERNEL : DMA_RESV_USAGE_BOOKKEEP); @@ -1669,7 +1669,7 @@ __xe_pt_unbind_vma(struct xe_tile *tile, struct xe_vma *vma, struct xe_exec_queu fence = &ifence->base.base; /* add shared fence now for pagetable delayed destroy */ - dma_resv_add_fence(&vm->resv, fence, + dma_resv_add_fence(xe_vm_resv(vm), fence, DMA_RESV_USAGE_BOOKKEEP); /* This fence will be installed by caller when doing eviction */ diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index d25dfa402267..e46812c01154 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -37,6 +37,11 @@ #define TEST_VM_ASYNC_OPS_ERROR +static struct drm_gem_object *xe_vm_obj(struct xe_vm *vm) +{ + return vm->gpuvm.r_obj; +} + /** * xe_vma_userptr_check_repin() - Advisory check for repin needed * @vma: The userptr vma @@ -321,7 +326,7 @@ static void resume_and_reinstall_preempt_fences(struct xe_vm *vm) list_for_each_entry(q, &vm->preempt.exec_queues, compute.link) { q->ops->resume(q); - dma_resv_add_fence(&vm->resv, q->compute.pfence, + dma_resv_add_fence(xe_vm_resv(vm), q->compute.pfence, DMA_RESV_USAGE_BOOKKEEP); xe_vm_fence_all_extobjs(vm, q->compute.pfence, DMA_RESV_USAGE_BOOKKEEP); @@ -359,7 +364,7 @@ int xe_vm_add_compute_exec_queue(struct xe_vm *vm, struct xe_exec_queue *q) down_read(&vm->userptr.notifier_lock); - dma_resv_add_fence(&vm->resv, pfence, + dma_resv_add_fence(xe_vm_resv(vm), pfence, DMA_RESV_USAGE_BOOKKEEP); xe_vm_fence_all_extobjs(vm, pfence, DMA_RESV_USAGE_BOOKKEEP); @@ -424,8 +429,7 @@ int xe_vm_lock_dma_resv(struct xe_vm *vm, struct drm_exec *exec, lockdep_assert_held(&vm->lock); if (lock_vm) { - err = drm_exec_prepare_obj(exec, &xe_vm_ttm_bo(vm)->base, - num_shared); + err = drm_exec_prepare_obj(exec, xe_vm_obj(vm), num_shared); if (err) return err; } @@ -521,7 +525,7 @@ static int xe_preempt_work_begin(struct drm_exec *exec, struct xe_vm *vm, * 1 fence for each preempt fence plus a fence for each tile from a * possible rebind */ - err = drm_exec_prepare_obj(exec, &xe_vm_ttm_bo(vm)->base, + err = drm_exec_prepare_obj(exec, xe_vm_obj(vm), vm->preempt.num_exec_queues + vm->xe->info.tile_count); if (err) @@ -636,7 +640,7 @@ static void preempt_rebind_work_func(struct work_struct *w) } /* Wait on munmap style VM unbinds */ - wait = dma_resv_wait_timeout(&vm->resv, + wait = dma_resv_wait_timeout(xe_vm_resv(vm), DMA_RESV_USAGE_KERNEL, false, MAX_SCHEDULE_TIMEOUT); if (wait <= 0) { @@ -731,13 +735,13 @@ static bool vma_userptr_invalidate(struct mmu_interval_notifier *mni, * unbinds to complete, and those are attached as BOOKMARK fences * to the vm. */ - dma_resv_iter_begin(&cursor, &vm->resv, + dma_resv_iter_begin(&cursor, xe_vm_resv(vm), DMA_RESV_USAGE_BOOKKEEP); dma_resv_for_each_fence_unlocked(&cursor, fence) dma_fence_enable_sw_signaling(fence); dma_resv_iter_end(&cursor); - err = dma_resv_wait_timeout(&vm->resv, + err = dma_resv_wait_timeout(xe_vm_resv(vm), DMA_RESV_USAGE_BOOKKEEP, false, MAX_SCHEDULE_TIMEOUT); XE_WARN_ON(err <= 0); @@ -786,14 +790,14 @@ int xe_vm_userptr_pin(struct xe_vm *vm) } /* Take lock and move to rebind_list for rebinding. */ - err = dma_resv_lock_interruptible(&vm->resv, NULL); + err = dma_resv_lock_interruptible(xe_vm_resv(vm), NULL); if (err) goto out_err; list_for_each_entry_safe(vma, next, &tmp_evict, combined_links.userptr) list_move_tail(&vma->combined_links.rebind, &vm->rebind_list); - dma_resv_unlock(&vm->resv); + dma_resv_unlock(xe_vm_resv(vm)); return 0; @@ -1111,7 +1115,7 @@ int xe_vm_prepare_vma(struct drm_exec *exec, struct xe_vma *vma, int err; XE_WARN_ON(!vm); - err = drm_exec_prepare_obj(exec, &xe_vm_ttm_bo(vm)->base, num_shared); + err = drm_exec_prepare_obj(exec, xe_vm_obj(vm), num_shared); if (!err && bo && !bo->vm) err = drm_exec_prepare_obj(exec, &bo->ttm.base, num_shared); @@ -1322,6 +1326,7 @@ static void vm_destroy_work_func(struct work_struct *w); struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags) { + struct drm_gem_object *vm_resv_obj; struct xe_vm *vm; int err, i = 0, number_tiles = 0; struct xe_tile *tile; @@ -1333,7 +1338,6 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags) vm->xe = xe; kref_init(&vm->refcount); - dma_resv_init(&vm->resv); vm->size = 1ull << xe->info.va_bits; @@ -1370,12 +1374,21 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags) if (!(flags & XE_VM_FLAG_MIGRATION)) xe_device_mem_access_get(xe); - err = dma_resv_lock_interruptible(&vm->resv, NULL); + vm_resv_obj = drm_gpuvm_root_object_alloc(&xe->drm); + if (!vm_resv_obj) { + err = -ENOMEM; + goto err_no_resv; + } + + drm_gpuvm_init(&vm->gpuvm, vm_resv_obj, "Xe VM", 0, vm->size, 0, 0, + &gpuvm_ops); + + drm_gem_object_put(vm_resv_obj); + + err = dma_resv_lock_interruptible(xe_vm_resv(vm), NULL); if (err) goto err_put; - drm_gpuvm_init(&vm->gpuvm, "Xe VM", 0, vm->size, 0, 0, - &gpuvm_ops); if (IS_DGFX(xe) && xe->info.vram_flags & XE_VRAM_FLAGS_NEED64K) vm->flags |= XE_VM_FLAG_64K; @@ -1422,7 +1435,7 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags) xe_pt_populate_empty(tile, vm, vm->pt_root[id]); } - dma_resv_unlock(&vm->resv); + dma_resv_unlock(xe_vm_resv(vm)); /* Kernel migration VM shouldn't have a circular loop.. */ if (!(flags & XE_VM_FLAG_MIGRATION)) { @@ -1480,10 +1493,10 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags) if (vm->pt_root[id]) xe_pt_destroy(vm->pt_root[id], vm->flags, NULL); } - dma_resv_unlock(&vm->resv); - drm_gpuvm_destroy(&vm->gpuvm); + dma_resv_unlock(xe_vm_resv(vm)); err_put: - dma_resv_fini(&vm->resv); + drm_gpuvm_destroy(&vm->gpuvm); +err_no_resv: for_each_tile(tile, xe, id) xe_range_fence_tree_fini(&vm->rftree[id]); kfree(vm); @@ -1621,8 +1634,6 @@ void xe_vm_close_and_put(struct xe_vm *vm) xe_assert(xe, list_empty(&vm->extobj.list)); up_write(&vm->lock); - drm_gpuvm_destroy(&vm->gpuvm); - mutex_lock(&xe->usm.lock); if (vm->flags & XE_VM_FLAG_FAULT_MODE) xe->usm.num_vm_in_fault_mode--; @@ -1675,7 +1686,7 @@ static void vm_destroy_work_func(struct work_struct *w) trace_xe_vm_free(vm); dma_fence_put(vm->rebind_fence); - dma_resv_fini(&vm->resv); + drm_gpuvm_destroy(&vm->gpuvm); kfree(vm); } @@ -2273,15 +2284,6 @@ static int xe_vm_prefetch(struct xe_vm *vm, struct xe_vma *vma, #define VM_BIND_OP(op) (op & 0xffff) -struct ttm_buffer_object *xe_vm_ttm_bo(struct xe_vm *vm) -{ - int idx = vm->flags & XE_VM_FLAG_MIGRATION ? - XE_VM_FLAG_TILE_ID(vm->flags) : 0; - - /* Safe to use index 0 as all BO in the VM share a single dma-resv lock */ - return &vm->pt_root[idx]->bo->ttm; -} - static void vm_set_async_error(struct xe_vm *vm, int err) { lockdep_assert_held(&vm->lock); @@ -3569,9 +3571,9 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) int xe_vm_lock(struct xe_vm *vm, bool intr) { if (intr) - return dma_resv_lock_interruptible(&vm->resv, NULL); + return dma_resv_lock_interruptible(xe_vm_resv(vm), NULL); - return dma_resv_lock(&vm->resv, NULL); + return dma_resv_lock(xe_vm_resv(vm), NULL); } /** @@ -3582,7 +3584,7 @@ int xe_vm_lock(struct xe_vm *vm, bool intr) */ void xe_vm_unlock(struct xe_vm *vm) { - dma_resv_unlock(&vm->resv); + dma_resv_unlock(xe_vm_resv(vm)); } /** @@ -3614,7 +3616,7 @@ int xe_vm_invalidate_vma(struct xe_vma *vma) WARN_ON_ONCE(!mmu_interval_check_retry (&vma->userptr.notifier, vma->userptr.notifier_seq)); - WARN_ON_ONCE(!dma_resv_test_signaled(&xe_vma_vm(vma)->resv, + WARN_ON_ONCE(!dma_resv_test_signaled(xe_vm_resv(xe_vma_vm(vma)), DMA_RESV_USAGE_BOOKKEEP)); } else { diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h index 4b32ce9cf9b1..12150fc0a7fd 100644 --- a/drivers/gpu/drm/xe/xe_vm.h +++ b/drivers/gpu/drm/xe/xe_vm.h @@ -139,8 +139,6 @@ static inline bool xe_vma_is_userptr(struct xe_vma *vma) return xe_vma_has_no_bo(vma) && !xe_vma_is_null(vma); } -#define xe_vm_assert_held(vm) dma_resv_assert_held(&(vm)->resv) - u64 xe_vm_pdp4_descriptor(struct xe_vm *vm, struct xe_tile *tile); int xe_vm_create_ioctl(struct drm_device *dev, void *data, @@ -183,8 +181,6 @@ int xe_vm_async_fence_wait_start(struct dma_fence *fence); extern struct ttm_device_funcs xe_ttm_funcs; -struct ttm_buffer_object *xe_vm_ttm_bo(struct xe_vm *vm); - static inline void xe_vm_queue_rebind_worker(struct xe_vm *vm) { xe_assert(vm->xe, xe_vm_in_compute_mode(vm)); diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h index da5e6cb6f094..036d5c53a206 100644 --- a/drivers/gpu/drm/xe/xe_vm_types.h +++ b/drivers/gpu/drm/xe/xe_vm_types.h @@ -132,8 +132,6 @@ struct xe_vma { struct xe_device; -#define xe_vm_assert_held(vm) dma_resv_assert_held(&(vm)->resv) - struct xe_vm { /** @gpuvm: base GPUVM used to track VMAs */ struct drm_gpuvm gpuvm; @@ -145,9 +143,6 @@ struct xe_vm { /* exec queue used for (un)binding vma's */ struct xe_exec_queue *q[XE_MAX_TILES_PER_DEVICE]; - /** Protects @rebind_list and the page-table structures */ - struct dma_resv resv; - /** @lru_bulk_move: Bulk LRU move list for this VM's BOs */ struct ttm_lru_bulk_move lru_bulk_move; @@ -436,4 +431,11 @@ struct xe_vma_op { }; }; +static inline struct dma_resv *xe_vm_resv(struct xe_vm *vm) +{ + return drm_gpuvm_resv(&vm->gpuvm); +} + +#define xe_vm_assert_held(vm) dma_resv_assert_held(xe_vm_resv(vm)) + #endif -- 2.41.0