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 57013C7EE2D for ; Mon, 22 May 2023 17:22:35 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 1F7AA10E1E9; Mon, 22 May 2023 17:22:35 +0000 (UTC) Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by gabe.freedesktop.org (Postfix) with ESMTPS id 349F310E19B for ; Mon, 22 May 2023 17:22:30 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1684776150; x=1716312150; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=1iYpYrCzHhHQtgQdipiQ7XpjKN4O2kruky9iAaNqsfY=; b=AihbrGF4450+DZA4UuVP/pdrKh9FBpFPEUGAadxA0VD3pUY/m4zRsnte vnvGVn/P/dD8burrybowQU+xbrboo8J9hhCNHrdeS8HEOakp9cYOjzadQ Prk4Wg6ninQb2XosFihy+l3Zflc9lkdWdaj4ab3JN5g+IVBuuniegtXz6 3/6PT33A+eOdtj+fvRone8TQ32cVCQx1yQCsir0QhMmKLWm0BlkNtUfUM qhv39xO0KrFSOS2tDBJ3by691jfp5sCB7NJW5i/qSgGdB3v/ElmeyEh3t /pizVW3qqENf/FLvtcfLNZfwFLNYLetl4ZN+Lh6EiNetsiZv28GU8iLjf w==; X-IronPort-AV: E=McAfee;i="6600,9927,10718"; a="342439578" X-IronPort-AV: E=Sophos;i="6.00,184,1681196400"; d="scan'208";a="342439578" Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by orsmga101.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 May 2023 10:22:30 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6600,9927,10718"; a="793409255" X-IronPort-AV: E=Sophos;i="6.00,184,1681196400"; d="scan'208";a="793409255" Received: from amigon-mobl.ger.corp.intel.com (HELO mwauld-desk1.intel.com) ([10.249.128.230]) by fmsmga003-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 22 May 2023 10:22:28 -0700 From: Matthew Auld To: intel-xe@lists.freedesktop.org Date: Mon, 22 May 2023 18:22:07 +0100 Message-Id: <20230522172211.90678-3-matthew.auld@intel.com> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20230522172211.90678-1-matthew.auld@intel.com> References: <20230522172211.90678-1-matthew.auld@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Subject: [Intel-xe] [PATCH v8 3/7] drm/xe: keep pulling mem_access_get further back 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: , Cc: Rodrigo Vivi Errors-To: intel-xe-bounces@lists.freedesktop.org Sender: "Intel-xe" Lockdep is unhappy about ggtt->lock -> mem_access.lock, where it seems to think this can somehow get inverted. The ggtt->lock looks like a potentially sensitive driver lock, so likely a sensible move to never call the runtime_pm routines while holding it. Actually it looks like d3cold wants to grab this, so perhaps this can indeed deadlock. v2: - Don't forget about xe_gt_tlb_invalidation_vma(), which now needs explicit access_get. Signed-off-by: Matthew Auld Cc: Thomas Hellström Cc: Rodrigo Vivi Reviewed-by: Rodrigo Vivi --- drivers/gpu/drm/xe/display/xe_fb_pin.c | 12 +++++++----- drivers/gpu/drm/xe/xe_ggtt.c | 6 ++++++ drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c | 10 ++++++---- 3 files changed, 19 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/xe/display/xe_fb_pin.c b/drivers/gpu/drm/xe/display/xe_fb_pin.c index c7c4df18e439..d140eee6473d 100644 --- a/drivers/gpu/drm/xe/display/xe_fb_pin.c +++ b/drivers/gpu/drm/xe/display/xe_fb_pin.c @@ -130,9 +130,10 @@ static int __xe_pin_fb_vma_ggtt(struct intel_framebuffer *fb, /* TODO: Consider sharing framebuffer mapping? * embed i915_vma inside intel_framebuffer */ + xe_device_mem_access_get(ggtt->gt->xe); ret = mutex_lock_interruptible(&ggtt->lock); if (ret) - return ret; + goto out; align = XE_PAGE_SIZE; if (xe_bo_is_vram(bo) && ggtt->flags & XE_GGTT_FLAGS_64K) @@ -146,7 +147,7 @@ static int __xe_pin_fb_vma_ggtt(struct intel_framebuffer *fb, ret = xe_ggtt_insert_special_node_locked(ggtt, &vma->node, size, align, 0); if (ret) - goto out; + goto out_unlock; for (x = 0; x < size; x += XE_PAGE_SIZE) xe_ggtt_set_pte(ggtt, vma->node.start + x, xe_ggtt_pte_encode(bo, x)); @@ -160,7 +161,7 @@ static int __xe_pin_fb_vma_ggtt(struct intel_framebuffer *fb, ret = xe_ggtt_insert_special_node_locked(ggtt, &vma->node, size, align, 0); if (ret) - goto out; + goto out_unlock; ggtt_ofs = vma->node.start; @@ -174,9 +175,10 @@ static int __xe_pin_fb_vma_ggtt(struct intel_framebuffer *fb, } xe_ggtt_invalidate(to_gt(xe)); - -out: +out_unlock: mutex_unlock(&ggtt->lock); +out: + xe_device_mem_access_put(ggtt->gt->xe); return ret; } diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c index 546240261e0a..f986e8218820 100644 --- a/drivers/gpu/drm/xe/xe_ggtt.c +++ b/drivers/gpu/drm/xe/xe_ggtt.c @@ -142,12 +142,14 @@ static void xe_ggtt_initial_clear(struct xe_ggtt *ggtt) u64 start, end; /* Display may have allocated inside ggtt, so be careful with clearing here */ + xe_device_mem_access_get(ggtt->gt->xe); mutex_lock(&ggtt->lock); drm_mm_for_each_hole(hole, &ggtt->mm, start, end) xe_ggtt_clear(ggtt, start, end - start); xe_ggtt_invalidate(ggtt->gt); mutex_unlock(&ggtt->lock); + xe_device_mem_access_put(ggtt->gt->xe); } int xe_ggtt_init(struct xe_gt *gt, struct xe_ggtt *ggtt) @@ -288,12 +290,14 @@ static int __xe_ggtt_insert_bo_at(struct xe_ggtt *ggtt, struct xe_bo *bo, if (err) return err; + xe_device_mem_access_get(ggtt->gt->xe); mutex_lock(&ggtt->lock); err = drm_mm_insert_node_in_range(&ggtt->mm, &bo->ggtt_node, bo->size, alignment, 0, start, end, 0); if (!err) xe_ggtt_map_bo(ggtt, bo); mutex_unlock(&ggtt->lock); + xe_device_mem_access_put(ggtt->gt->xe); return err; } @@ -311,6 +315,7 @@ int xe_ggtt_insert_bo(struct xe_ggtt *ggtt, struct xe_bo *bo) void xe_ggtt_remove_node(struct xe_ggtt *ggtt, struct drm_mm_node *node) { + xe_device_mem_access_get(ggtt->gt->xe); mutex_lock(&ggtt->lock); xe_ggtt_clear(ggtt, node->start, node->size); @@ -320,6 +325,7 @@ void xe_ggtt_remove_node(struct xe_ggtt *ggtt, struct drm_mm_node *node) xe_ggtt_invalidate(ggtt->gt); mutex_unlock(&ggtt->lock); + xe_device_mem_access_put(ggtt->gt->xe); } void xe_ggtt_remove_bo(struct xe_ggtt *ggtt, struct xe_bo *bo) diff --git a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c index 20f8f0aae6b4..44e442bf306c 100644 --- a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c +++ b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c @@ -114,7 +114,6 @@ static int send_tlb_invalidation(struct xe_guc *guc, * need to be updated. */ - xe_device_mem_access_get(gt->xe); mutex_lock(&guc->ct.lock); seqno = gt->tlb_invalidation.seqno; if (fence) { @@ -143,7 +142,6 @@ static int send_tlb_invalidation(struct xe_guc *guc, if (ret < 0 && fence) invalidation_fence_signal(fence); mutex_unlock(&guc->ct.lock); - xe_device_mem_access_put(gt->xe); return ret; } @@ -196,7 +194,7 @@ int xe_gt_tlb_invalidation_vma(struct xe_gt *gt, struct xe_device *xe = gt_to_xe(gt); #define MAX_TLB_INVALIDATION_LEN 7 u32 action[MAX_TLB_INVALIDATION_LEN]; - int len = 0; + int len = 0, ret; XE_BUG_ON(!vma); @@ -250,7 +248,11 @@ int xe_gt_tlb_invalidation_vma(struct xe_gt *gt, XE_BUG_ON(len > MAX_TLB_INVALIDATION_LEN); - return send_tlb_invalidation(>->uc.guc, fence, action, len); + xe_device_mem_access_get(gt->xe); + ret = send_tlb_invalidation(>->uc.guc, fence, action, len); + xe_device_mem_access_put(gt->xe); + + return ret; } static bool tlb_invalidation_seqno_past(struct xe_gt *gt, int seqno) -- 2.40.1