From patchwork Fri Oct 7 09:46:03 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chris Wilson X-Patchwork-Id: 9365841 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id E757F60487 for ; Fri, 7 Oct 2016 09:47:03 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id D77D329469 for ; Fri, 7 Oct 2016 09:47:03 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id CC5422946C; Fri, 7 Oct 2016 09:47:03 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-4.1 required=2.0 tests=BAYES_00,DKIM_SIGNED, RCVD_IN_DNSWL_MED,T_DKIM_INVALID autolearn=ham version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 42B2D29469 for ; Fri, 7 Oct 2016 09:47:03 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id A88C76EB3D; Fri, 7 Oct 2016 09:47:02 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from mail-wm0-x242.google.com (mail-wm0-x242.google.com [IPv6:2a00:1450:400c:c09::242]) by gabe.freedesktop.org (Postfix) with ESMTPS id B8C6E6EB38 for ; Fri, 7 Oct 2016 09:46:53 +0000 (UTC) Received: by mail-wm0-x242.google.com with SMTP id i130so2095789wmg.0 for ; Fri, 07 Oct 2016 02:46:53 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=sender:from:to:cc:subject:date:message-id:in-reply-to:references; bh=y+49irftMuMvsTqGlBFYjkLpX8ZueN9Gay7oKl/XVxg=; b=cFaL8gwwZFqtWbAyeAAql5uWZTPZYHKILxnDJNMzteLEXnlZ2twPbcY+j2zx+XEcNr GV5dR8gXPk6DaGmxPDYUEfLEj6Nx7HqhAnM2O4W3X/B2P6vXgHsDUEG6nGSfqGtjxu0f n8Ue5ieF7BuP3ZqMtl1kT+lyISzMj+HZUaE8Ahw5k60MqVKImXLelDJe98ykZvPc8pE/ LOhM0bCizWynSpgbdPqviN/usRTi1wcNNtPjwF3S9a/y8PGGPntFNngFXV6cRdf6WtAR DNMsBSOn/hSUEHdpKbLUg65BvFbooB98un5Aql2L7IV3Hmoz7zviV2rhTBj+lMH+04Ni 2J0g== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:sender:from:to:cc:subject:date:message-id :in-reply-to:references; bh=y+49irftMuMvsTqGlBFYjkLpX8ZueN9Gay7oKl/XVxg=; b=emp63J99OouCRac/zvk+spSeraWFAvaoK9P3qnXz/cF8y0dsXyfLcUIj+yotQ5Q3A+ 5UtMZuzri239mKIl2+dtK2eZgyCkos2lNvdsoyvaH6FPr7FIQk8IGKVMXkXb649QkbVZ oy0vI8jMHJfq51zsdOGSj7RkZcsyp0+S8jThCsy4/1bcEE3FTeFBOxlvs9SBpjXYTRL/ 45D/WI2Gs3rDCOm5nR/L+7/Y4sQGI0bdhWxZ4gr0F3Z07pXax9xTcYHVGoM1AeWKYzZ4 GiqTqx7IOYbikisIIIPRBJmvtQvFgx5B9kwcC65TyV+tia7Neg8za0BkmRXh7RYFqiFb zGbA== X-Gm-Message-State: AA6/9Rl6mKYwJWCM7J3DqjMEw+1e5WfbQ1M9Yu51u9WzwS56+7amb/HLpfYEjPjeSYXiwg== X-Received: by 10.28.95.87 with SMTP id t84mr17450330wmb.51.1475833612023; Fri, 07 Oct 2016 02:46:52 -0700 (PDT) Received: from haswell.alporthouse.com ([78.156.65.138]) by smtp.gmail.com with ESMTPSA id h3sm18877585wjp.45.2016.10.07.02.46.50 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 07 Oct 2016 02:46:50 -0700 (PDT) From: Chris Wilson To: intel-gfx@lists.freedesktop.org Date: Fri, 7 Oct 2016 10:46:03 +0100 Message-Id: <20161007094635.28319-11-chris@chris-wilson.co.uk> X-Mailer: git-send-email 2.9.3 In-Reply-To: <20161007094635.28319-1-chris@chris-wilson.co.uk> References: <20161007094635.28319-1-chris@chris-wilson.co.uk> Subject: [Intel-gfx] [PATCH 10/42] drm/i915: Defer active reference until required X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Virus-Scanned: ClamAV using ClamSMTP We only need the active reference to keep the object alive after the handle has been deleted (so as to prevent a synchronous gem_close). Why then pay the price of a kref on every execbuf when we can insert that final active ref just in time for the handle deletion? Signed-off-by: Chris Wilson Reviewed-by: Joonas Lahtinen --- drivers/gpu/drm/i915/i915_drv.h | 28 ++++++++++++++++++++++++++++ drivers/gpu/drm/i915/i915_gem.c | 22 +++++++++++++++++++++- drivers/gpu/drm/i915/i915_gem_batch_pool.c | 2 +- drivers/gpu/drm/i915/i915_gem_context.c | 2 +- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 -- drivers/gpu/drm/i915/i915_gem_gtt.c | 7 ++++++- drivers/gpu/drm/i915/i915_gem_render_state.c | 3 ++- drivers/gpu/drm/i915/intel_ringbuffer.c | 15 ++++++++++++--- 8 files changed, 71 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index ee25e265416f..fee5cc92e2f2 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -2232,6 +2232,12 @@ struct drm_i915_gem_object { ((READ_ONCE((bo)->flags) >> I915_BO_ACTIVE_SHIFT) & I915_BO_ACTIVE_MASK) /** + * Have we taken a reference for the object for incomplete GPU + * activity? + */ +#define I915_BO_ACTIVE_REF (I915_BO_ACTIVE_SHIFT + I915_NUM_ENGINES) + + /** * This is set if the object has been written to since last bound * to the GTT */ @@ -2399,6 +2405,28 @@ i915_gem_object_has_active_engine(const struct drm_i915_gem_object *obj, return obj->flags & BIT(engine + I915_BO_ACTIVE_SHIFT); } +static inline bool +i915_gem_object_has_active_reference(const struct drm_i915_gem_object *obj) +{ + return test_bit(I915_BO_ACTIVE_REF, &obj->flags); +} + +static inline void +i915_gem_object_set_active_reference(struct drm_i915_gem_object *obj) +{ + lockdep_assert_held(&obj->base.dev->struct_mutex); + __set_bit(I915_BO_ACTIVE_REF, &obj->flags); +} + +static inline void +i915_gem_object_clear_active_reference(struct drm_i915_gem_object *obj) +{ + lockdep_assert_held(&obj->base.dev->struct_mutex); + __clear_bit(I915_BO_ACTIVE_REF, &obj->flags); +} + +void __i915_gem_object_release_unless_active(struct drm_i915_gem_object *obj); + static inline unsigned int i915_gem_object_get_tiling(struct drm_i915_gem_object *obj) { diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 7fa5cb764739..b560263bf446 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2618,7 +2618,10 @@ i915_gem_object_retire__read(struct i915_gem_active *active, list_move_tail(&obj->global_list, &request->i915->mm.bound_list); - i915_gem_object_put(obj); + if (i915_gem_object_has_active_reference(obj)) { + i915_gem_object_clear_active_reference(obj); + i915_gem_object_put(obj); + } } static bool i915_context_is_banned(const struct i915_gem_context *ctx) @@ -2889,6 +2892,12 @@ void i915_gem_close_object(struct drm_gem_object *gem, struct drm_file *file) list_for_each_entry_safe(vma, vn, &obj->vma_list, obj_link) if (vma->vm->file == fpriv) i915_vma_close(vma); + + if (i915_gem_object_is_active(obj) && + !i915_gem_object_has_active_reference(obj)) { + i915_gem_object_set_active_reference(obj); + i915_gem_object_get(obj); + } mutex_unlock(&obj->base.dev->struct_mutex); } @@ -4365,6 +4374,17 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj) intel_runtime_pm_put(dev_priv); } +void __i915_gem_object_release_unless_active(struct drm_i915_gem_object *obj) +{ + lockdep_assert_held(&obj->base.dev->struct_mutex); + + GEM_BUG_ON(i915_gem_object_has_active_reference(obj)); + if (i915_gem_object_is_active(obj)) + i915_gem_object_set_active_reference(obj); + else + i915_gem_object_put(obj); +} + int i915_gem_suspend(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.c b/drivers/gpu/drm/i915/i915_gem_batch_pool.c index ed989596d9a3..cb25cad3318c 100644 --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.c +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.c @@ -73,7 +73,7 @@ void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool) list_for_each_entry_safe(obj, next, &pool->cache_list[n], batch_pool_link) - i915_gem_object_put(obj); + __i915_gem_object_release_unless_active(obj); INIT_LIST_HEAD(&pool->cache_list[n]); } diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index df10f4e95736..1d2ab73a8f43 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -155,7 +155,7 @@ void i915_gem_context_free(struct kref *ctx_ref) if (ce->ring) intel_ring_free(ce->ring); - i915_vma_put(ce->state); + __i915_gem_object_release_unless_active(ce->state->obj); } put_pid(ctx->pid); diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 72c7c1855e70..0deecd4e3b6c 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -1299,8 +1299,6 @@ void i915_vma_move_to_active(struct i915_vma *vma, * add the active reference first and queue for it to be dropped * *last*. */ - if (!i915_gem_object_is_active(obj)) - i915_gem_object_get(obj); i915_gem_object_set_active(obj, idx); i915_gem_active_set(&obj->last_read[idx], req); diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 2d846aa39ca5..1c95da8424cb 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -3712,11 +3712,16 @@ void __iomem *i915_vma_pin_iomap(struct i915_vma *vma) void i915_vma_unpin_and_release(struct i915_vma **p_vma) { struct i915_vma *vma; + struct drm_i915_gem_object *obj; vma = fetch_and_zero(p_vma); if (!vma) return; + obj = vma->obj; + i915_vma_unpin(vma); - i915_vma_put(vma); + i915_vma_close(vma); + + __i915_gem_object_release_unless_active(obj); } diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c index 95b7e9afd5f8..09cf4874c45f 100644 --- a/drivers/gpu/drm/i915/i915_gem_render_state.c +++ b/drivers/gpu/drm/i915/i915_gem_render_state.c @@ -224,7 +224,8 @@ int i915_gem_render_state_init(struct drm_i915_gem_request *req) i915_vma_move_to_active(so.vma, req, 0); err_unpin: i915_vma_unpin(so.vma); + i915_vma_close(so.vma); err_obj: - i915_gem_object_put(obj); + __i915_gem_object_release_unless_active(obj); return ret; } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index b60c6f09fbfd..f3dfb7ca625d 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1763,14 +1763,19 @@ static void cleanup_phys_status_page(struct intel_engine_cs *engine) static void cleanup_status_page(struct intel_engine_cs *engine) { struct i915_vma *vma; + struct drm_i915_gem_object *obj; vma = fetch_and_zero(&engine->status_page.vma); if (!vma) return; + obj = vma->obj; + i915_vma_unpin(vma); - i915_gem_object_unpin_map(vma->obj); - i915_vma_put(vma); + i915_vma_close(vma); + + i915_gem_object_unpin_map(obj); + __i915_gem_object_release_unless_active(obj); } static int init_status_page(struct intel_engine_cs *engine) @@ -1968,7 +1973,11 @@ intel_engine_create_ring(struct intel_engine_cs *engine, int size) void intel_ring_free(struct intel_ring *ring) { - i915_vma_put(ring->vma); + struct drm_i915_gem_object *obj = ring->vma->obj; + + i915_vma_close(ring->vma); + __i915_gem_object_release_unless_active(obj); + kfree(ring); }