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 kanga.kvack.org (kanga.kvack.org [205.233.56.17]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id BD8BDC624C6 for ; Mon, 2 Mar 2026 16:33:25 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id 37BB46B0093; Mon, 2 Mar 2026 11:33:25 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 2FC496B0095; Mon, 2 Mar 2026 11:33:25 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 1BCE86B0096; Mon, 2 Mar 2026 11:33:25 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0016.hostedemail.com [216.40.44.16]) by kanga.kvack.org (Postfix) with ESMTP id 07EBF6B0093 for ; Mon, 2 Mar 2026 11:33:25 -0500 (EST) Received: from smtpin30.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay02.hostedemail.com (Postfix) with ESMTP id 89AD8139C51 for ; Mon, 2 Mar 2026 16:33:24 +0000 (UTC) X-FDA: 84501668328.30.8272F7A Received: from mgamail.intel.com (mgamail.intel.com [198.175.65.17]) by imf19.hostedemail.com (Postfix) with ESMTP id 2E1061A0015 for ; Mon, 2 Mar 2026 16:33:22 +0000 (UTC) Authentication-Results: imf19.hostedemail.com; dkim=pass header.d=intel.com header.s=Intel header.b=POUneRu2; spf=pass (imf19.hostedemail.com: domain of thomas.hellstrom@linux.intel.com designates 198.175.65.17 as permitted sender) smtp.mailfrom=thomas.hellstrom@linux.intel.com; dmarc=pass (policy=none) header.from=intel.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1772469202; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=RoJu2TZQJ6DUDvByKakWyLdyfBWLzd6QHY5e5dLWvAs=; b=hJtmJhtreZccZ5xTCLoGi+DW5cISEeGZGDcPVhtLM4kONCB7UrlFD3YoffyzZcU78F3thJ y9Iz1mMP0HMqh20bKp194OygtNV6YIaRzR2zUQiPoShebrTTTUHHCSkOSZGS25pk7P4Xrb Ay5SjrBt31JKA1NuE5GP9ydRD/zNLMI= ARC-Authentication-Results: i=1; imf19.hostedemail.com; dkim=pass header.d=intel.com header.s=Intel header.b=POUneRu2; spf=pass (imf19.hostedemail.com: domain of thomas.hellstrom@linux.intel.com designates 198.175.65.17 as permitted sender) smtp.mailfrom=thomas.hellstrom@linux.intel.com; dmarc=pass (policy=none) header.from=intel.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1772469202; a=rsa-sha256; cv=none; b=AIC8rP3mtyHf4giUjWCmlKw3uzf07cBbOcqgLdI0U9eNn28XLq9gPFc5fzPxUNikHLWPLm jm06NX8q6dVXg/DVFgcIZZDfO4N6O+41r0VEE8ZpCDFKA1S36fXxTXeW1r3v3xLxxof4ir 9UvM1/UUrPmNdN833NKQh46DKmXzNbI= DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1772469202; x=1804005202; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=gr/Rm/LmHwRqu0m9Cj42RM/GXccyftGf8yIDm0daZ4M=; b=POUneRu2jjEuFYt+S1RsLTv0hxC6rveLh9+NuqLI4S/ZUkcMMsepnTMd FLjSl02O7hVpLcGviioVoEENpCg5V6R15tdxf4i+v7eOXbrfBOu0KvhG9 qHck7G9Wq8Juv8IoHXLndaJmhFMy43NvnP4JM/XPjZGkzKi5is4SblRzq mYbOOhHMVTjPzITuKCT1y2Ij+0uD8Tz2jufpHJrQ5PK+E38ZAcH2w3ROA nOhht0bkcM+59OsmcNIJQ0Ta9qOFofQjx37AHZslY1aiYZu1ncphuoaJp DNGNAaO8JkGDo+WulttJyYtbx/OY/C5oWBXWx/a+eKAgHQLesENkbll84 Q==; X-CSE-ConnectionGUID: 2KobnuD9SuqeDTa78fNiBw== X-CSE-MsgGUID: /CBN0oonT8S5dOu2e8hWXQ== X-IronPort-AV: E=McAfee;i="6800,10657,11717"; a="73447871" X-IronPort-AV: E=Sophos;i="6.21,320,1763452800"; d="scan'208";a="73447871" Received: from orviesa001.jf.intel.com ([10.64.159.141]) by orvoesa109.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 02 Mar 2026 08:33:21 -0800 X-CSE-ConnectionGUID: TpC3DWhCS0uNy94/GNPzYg== X-CSE-MsgGUID: vuRGhG7cSu+atzC1e2EzwQ== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.21,320,1763452800"; d="scan'208";a="255564537" Received: from smoticic-mobl1.ger.corp.intel.com (HELO fedora) ([10.245.244.81]) by smtpauth.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 02 Mar 2026 08:33:19 -0800 From: =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= To: intel-xe@lists.freedesktop.org Cc: =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= , Matthew Brost , =?UTF-8?q?Christian=20K=C3=B6nig?= , dri-devel@lists.freedesktop.org, Jason Gunthorpe , Andrew Morton , Simona Vetter , Dave Airlie , Alistair Popple , linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: [PATCH v2 4/4] drm/xe/userptr: Defer Waiting for TLB invalidation to the second pass if possible Date: Mon, 2 Mar 2026 17:32:48 +0100 Message-ID: <20260302163248.105454-5-thomas.hellstrom@linux.intel.com> X-Mailer: git-send-email 2.53.0 In-Reply-To: <20260302163248.105454-1-thomas.hellstrom@linux.intel.com> References: <20260302163248.105454-1-thomas.hellstrom@linux.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Stat-Signature: m7npw3hm3anxzky8adwa3egrocsd6h8k X-Rspamd-Server: rspam09 X-Rspam-User: X-Rspamd-Queue-Id: 2E1061A0015 X-HE-Tag: 1772469202-258612 X-HE-Meta: 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 3YECLHXN Mav8tsu6pTygvGSO4mcWYzMBy6vCmRwfz0GohpZzGpSIERVxqF9Fh5L5B6yJcNeCZJ+tQUN66EpLgvbIWsox1MMtQHDgDGRluqiUO0iKM/ut4NRZuf4OgMgfReMyY1S7LfQMCX7iJR8JnHX9Yyvons6+JyJxgtdNtoB9KYqNKwAjeAefxwismzqLWZK+X2qTbSNWV6Eg5n84iINTuyFQ+18qdbP7Fb4tIshSqUf1wyjYDZZ8RYERx3c6X8nAWigACPGbDyUXMFk+TLN8= Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: Now that the two-pass notifier flow uses xe_vma_userptr_do_inval() for the fence-wait + TLB-invalidate work, extend it to support a further deferred TLB wait: - xe_vma_userptr_do_inval(): when the embedded finish handle is free, submit the TLB invalidation asynchronously (xe_vm_invalidate_vma_submit) and return &userptr->finish so the mmu_notifier core schedules a third pass. When the handle is occupied by a concurrent invalidation, fall back to the synchronous xe_vm_invalidate_vma() path. - xe_vma_userptr_complete_tlb_inval(): new helper called from invalidate_finish when tlb_inval_submitted is set. Waits for the previously submitted batch and unmaps the gpusvm pages. xe_vma_userptr_invalidate_finish() dispatches between the two helpers via tlb_inval_submitted, making the three possible flows explicit: pass1 (fences pending) -> invalidate_finish -> do_inval (sync TLB) pass1 (fences done) -> do_inval -> invalidate_finish -> complete_tlb_inval (deferred TLB) pass1 (finish occupied) -> do_inval (sync TLB, inline) In multi-GPU scenarios this allows TLB flushes to be submitted on all GPUs in one pass before any of them are waited on. Also adds xe_vm_invalidate_vma_submit() which submits the TLB range invalidation without blocking, populating a xe_tlb_inval_batch that the caller waits on separately. Assisted-by: GitHub Copilot:claude-sonnet-4.6 Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_userptr.c | 60 +++++++++++++++++++++++++++------ drivers/gpu/drm/xe/xe_userptr.h | 18 ++++++++++ drivers/gpu/drm/xe/xe_vm.c | 38 ++++++++++++++++----- drivers/gpu/drm/xe/xe_vm.h | 2 ++ 4 files changed, 99 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_userptr.c b/drivers/gpu/drm/xe/xe_userptr.c index 440b0a79d16f..a62b796afb93 100644 --- a/drivers/gpu/drm/xe/xe_userptr.c +++ b/drivers/gpu/drm/xe/xe_userptr.c @@ -8,6 +8,7 @@ #include +#include "xe_tlb_inval.h" #include "xe_trace_bo.h" /** @@ -73,8 +74,8 @@ int xe_vma_userptr_pin_pages(struct xe_userptr_vma *uvma) &ctx); } -static void xe_vma_userptr_do_inval(struct xe_vm *vm, struct xe_userptr_vma *uvma, - bool is_deferred) +static struct mmu_interval_notifier_finish * +xe_vma_userptr_do_inval(struct xe_vm *vm, struct xe_userptr_vma *uvma, bool is_deferred) { struct xe_userptr *userptr = &uvma->userptr; struct xe_vma *vma = &uvma->vma; @@ -84,12 +85,23 @@ static void xe_vma_userptr_do_inval(struct xe_vm *vm, struct xe_userptr_vma *uvm }; long err; - err = dma_resv_wait_timeout(xe_vm_resv(vm), - DMA_RESV_USAGE_BOOKKEEP, + err = dma_resv_wait_timeout(xe_vm_resv(vm), DMA_RESV_USAGE_BOOKKEEP, false, MAX_SCHEDULE_TIMEOUT); XE_WARN_ON(err <= 0); if (xe_vm_in_fault_mode(vm) && userptr->initial_bind) { + if (!userptr->finish_inuse) { + /* + * Defer the TLB wait to an extra pass so the caller + * can pipeline TLB flushes across GPUs before waiting + * on any of them. + */ + userptr->finish_inuse = true; + userptr->tlb_inval_submitted = true; + err = xe_vm_invalidate_vma_submit(vma, &userptr->inval_batch); + XE_WARN_ON(err); + return &userptr->finish; + } err = xe_vm_invalidate_vma(vma); XE_WARN_ON(err); } @@ -98,6 +110,24 @@ static void xe_vma_userptr_do_inval(struct xe_vm *vm, struct xe_userptr_vma *uvm userptr->finish_inuse = false; drm_gpusvm_unmap_pages(&vm->svm.gpusvm, &uvma->userptr.pages, xe_vma_size(vma) >> PAGE_SHIFT, &ctx); + return NULL; +} + +static void +xe_vma_userptr_complete_tlb_inval(struct xe_vm *vm, struct xe_userptr_vma *uvma) +{ + struct xe_userptr *userptr = &uvma->userptr; + struct xe_vma *vma = &uvma->vma; + struct drm_gpusvm_ctx ctx = { + .in_notifier = true, + .read_only = xe_vma_read_only(vma), + }; + + xe_tlb_inval_batch_wait(&userptr->inval_batch); + userptr->tlb_inval_submitted = false; + userptr->finish_inuse = false; + drm_gpusvm_unmap_pages(&vm->svm.gpusvm, &uvma->userptr.pages, + xe_vma_size(vma) >> PAGE_SHIFT, &ctx); } static struct mmu_interval_notifier_finish * @@ -141,13 +171,11 @@ xe_vma_userptr_invalidate_pass1(struct xe_vm *vm, struct xe_userptr_vma *uvma) * If it's already in use, or all fences are already signaled, * proceed directly to invalidation without deferring. */ - if (signaled || userptr->finish_inuse) { - xe_vma_userptr_do_inval(vm, uvma, false); - return NULL; - } + if (signaled || userptr->finish_inuse) + return xe_vma_userptr_do_inval(vm, uvma, false); + /* Defer: the notifier core will call invalidate_finish once done. */ userptr->finish_inuse = true; - return &userptr->finish; } @@ -193,7 +221,15 @@ static void xe_vma_userptr_invalidate_finish(struct mmu_interval_notifier_finish xe_vma_start(vma), xe_vma_size(vma)); down_write(&vm->svm.gpusvm.notifier_lock); - xe_vma_userptr_do_inval(vm, uvma, true); + /* + * If a TLB invalidation was previously submitted (deferred from the + * synchronous pass1 fallback), wait for it and unmap pages. + * Otherwise, fences have now completed: invalidate the TLB and unmap. + */ + if (uvma->userptr.tlb_inval_submitted) + xe_vma_userptr_complete_tlb_inval(vm, uvma); + else + xe_vma_userptr_do_inval(vm, uvma, true); up_write(&vm->svm.gpusvm.notifier_lock); trace_xe_vma_userptr_invalidate_complete(vma); } @@ -231,7 +267,9 @@ void xe_vma_userptr_force_invalidate(struct xe_userptr_vma *uvma) finish = xe_vma_userptr_invalidate_pass1(vm, uvma); if (finish) - xe_vma_userptr_do_inval(vm, uvma, true); + finish = xe_vma_userptr_do_inval(vm, uvma, true); + if (finish) + xe_vma_userptr_complete_tlb_inval(vm, uvma); } #endif diff --git a/drivers/gpu/drm/xe/xe_userptr.h b/drivers/gpu/drm/xe/xe_userptr.h index 4f42db61fd62..7477009651c2 100644 --- a/drivers/gpu/drm/xe/xe_userptr.h +++ b/drivers/gpu/drm/xe/xe_userptr.h @@ -14,6 +14,8 @@ #include +#include "xe_tlb_inval_types.h" + struct xe_vm; struct xe_vma; struct xe_userptr_vma; @@ -63,6 +65,15 @@ struct xe_userptr { * Protected by @vm::svm.gpusvm.notifier_lock. */ struct mmu_interval_notifier_finish finish; + + /** + * @inval_batch: TLB invalidation batch for deferred completion. + * Stores an in-flight TLB invalidation submitted during a two-pass + * notifier so the wait can be deferred to a subsequent pass, allowing + * multiple GPUs to be signalled before any of them are waited on. + * Protected by @vm::svm.gpusvm.notifier_lock. + */ + struct xe_tlb_inval_batch inval_batch; /** * @finish_inuse: Whether @finish is currently in use by an in-progress * two-pass invalidation. @@ -70,6 +81,13 @@ struct xe_userptr { */ bool finish_inuse; + /** + * @tlb_inval_submitted: Whether a TLB invalidation has been submitted + * via @inval_batch and is pending completion. When set, the next pass + * must call xe_tlb_inval_batch_wait() before reusing @inval_batch. + * Protected by @vm::svm.gpusvm.notifier_lock. + */ + bool tlb_inval_submitted; /** * @initial_bind: user pointer has been bound at least once. * write: vm->svm.gpusvm.notifier_lock in read mode and vm->resv held. diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 7f29d2b2972d..fdad9329dfb4 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -3967,20 +3967,23 @@ void xe_vm_unlock(struct xe_vm *vm) } /** - * xe_vm_invalidate_vma - invalidate GPU mappings for VMA without a lock + * xe_vm_invalidate_vma_submit - Submit a job to invalidate GPU mappings for + * VMA. * @vma: VMA to invalidate + * @batch: TLB invalidation batch to populate; caller must later call + * xe_tlb_inval_batch_wait() on it to wait for completion * * Walks a list of page tables leaves which it memset the entries owned by this - * VMA to zero, invalidates the TLBs, and block until TLBs invalidation is - * complete. + * VMA to zero, invalidates the TLBs, but doesn't block waiting for TLB flush + * to complete, but instead populates @batch which can be waited on using + * xe_tlb_inval_batch_wait(). * * Returns 0 for success, negative error code otherwise. */ -int xe_vm_invalidate_vma(struct xe_vma *vma) +int xe_vm_invalidate_vma_submit(struct xe_vma *vma, struct xe_tlb_inval_batch *batch) { struct xe_device *xe = xe_vma_vm(vma)->xe; struct xe_vm *vm = xe_vma_vm(vma); - struct xe_tlb_inval_batch _batch; struct xe_tile *tile; u8 tile_mask = 0; int ret = 0; @@ -4023,14 +4026,33 @@ int xe_vm_invalidate_vma(struct xe_vma *vma) ret = xe_tlb_inval_range_tilemask_submit(xe, xe_vma_vm(vma)->usm.asid, xe_vma_start(vma), xe_vma_end(vma), - tile_mask, &_batch); + tile_mask, batch); /* WRITE_ONCE pairs with READ_ONCE in xe_vm_has_valid_gpu_mapping() */ WRITE_ONCE(vma->tile_invalidated, vma->tile_mask); + return ret; +} + +/** + * xe_vm_invalidate_vma - invalidate GPU mappings for VMA without a lock + * @vma: VMA to invalidate + * + * Walks a list of page tables leaves which it memset the entries owned by this + * VMA to zero, invalidates the TLBs, and block until TLBs invalidation is + * complete. + * + * Returns 0 for success, negative error code otherwise. + */ +int xe_vm_invalidate_vma(struct xe_vma *vma) +{ + struct xe_tlb_inval_batch batch; + int ret; - if (!ret) - xe_tlb_inval_batch_wait(&_batch); + ret = xe_vm_invalidate_vma_submit(vma, &batch); + if (ret) + return ret; + xe_tlb_inval_batch_wait(&batch); return ret; } diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h index 62f4b6fec0bc..0bc7ed23eeae 100644 --- a/drivers/gpu/drm/xe/xe_vm.h +++ b/drivers/gpu/drm/xe/xe_vm.h @@ -242,6 +242,8 @@ struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm, int xe_vm_invalidate_vma(struct xe_vma *vma); +int xe_vm_invalidate_vma_submit(struct xe_vma *vma, struct xe_tlb_inval_batch *batch); + int xe_vm_validate_protected(struct xe_vm *vm); static inline void xe_vm_queue_rebind_worker(struct xe_vm *vm) -- 2.53.0