From patchwork Thu Mar 18 17:04:14 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tvrtko Ursulin X-Patchwork-Id: 12148681 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,HK_RANDOM_FROM,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 82839C433E9 for ; Thu, 18 Mar 2021 17:04:36 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 471EC64E05 for ; Thu, 18 Mar 2021 17:04:36 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 471EC64E05 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 62D886E92C; Thu, 18 Mar 2021 17:04:30 +0000 (UTC) Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id 05E1B6E92B; Thu, 18 Mar 2021 17:04:27 +0000 (UTC) IronPort-SDR: o6XnlrfOahb7ObB50Quv3neW0BX+gQ+8s/jsePHt6Ua5MgQB+IrSrhhfPDMpvvtVl/JrdjIlCg pX7+C3bBDZBw== X-IronPort-AV: E=McAfee;i="6000,8403,9927"; a="209733836" X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="209733836" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:27 -0700 IronPort-SDR: tNRowEcFsagdAjjexagMQ2sA+HQr4WRD5/gnUAkX+PIO1ZD/MPO+ZKcW2ZvSxwvkxyA7oY3GWf XhP9XGABeNDg== X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="413164893" Received: from syeghiay-mobl.ger.corp.intel.com (HELO localhost.localdomain) ([10.252.51.182]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:25 -0700 From: Tvrtko Ursulin To: Intel-gfx@lists.freedesktop.org Subject: [PATCH 1/6] drm/i915: Individual request cancellation Date: Thu, 18 Mar 2021 17:04:14 +0000 Message-Id: <20210318170419.2107512-2-tvrtko.ursulin@linux.intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> References: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Tvrtko Ursulin , dri-devel@lists.freedesktop.org, Chris Wilson Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Chris Wilson Currently, we cancel outstanding requests within a context when the context is closed. We may also want to cancel individual requests using the same graceful preemption mechanism. v2 (Tvrtko): * Cancel waiters carefully considering no timeline lock and RCU. * Fixed selftests. v3 (Tvrtko): * Remove error propagation to waiters for now. Signed-off-by: Chris Wilson Signed-off-by: Tvrtko Ursulin Reviewed-by: Matthew Auld --- .../gpu/drm/i915/gt/intel_engine_heartbeat.c | 1 + .../drm/i915/gt/intel_execlists_submission.c | 9 +- drivers/gpu/drm/i915/i915_request.c | 52 ++++- drivers/gpu/drm/i915/i915_request.h | 4 +- drivers/gpu/drm/i915/selftests/i915_request.c | 201 ++++++++++++++++++ 5 files changed, 261 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c index 0b062fad1837..e2fb3ae2aaf3 100644 --- a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c +++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c @@ -314,6 +314,7 @@ int intel_engine_pulse(struct intel_engine_cs *engine) mutex_unlock(&ce->timeline->mutex); } + intel_engine_flush_scheduler(engine); intel_engine_pm_put(engine); return err; } diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c index 85ff5fe861b4..4c2acb5a6c0a 100644 --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c @@ -421,6 +421,11 @@ static void reset_active(struct i915_request *rq, ce->lrc.lrca = lrc_update_regs(ce, engine, head); } +static bool bad_request(const struct i915_request *rq) +{ + return rq->fence.error && i915_request_started(rq); +} + static struct intel_engine_cs * __execlists_schedule_in(struct i915_request *rq) { @@ -433,7 +438,7 @@ __execlists_schedule_in(struct i915_request *rq) !intel_engine_has_heartbeat(engine))) intel_context_set_banned(ce); - if (unlikely(intel_context_is_banned(ce))) + if (unlikely(intel_context_is_banned(ce) || bad_request(rq))) reset_active(rq, engine); if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) @@ -1112,7 +1117,7 @@ static unsigned long active_preempt_timeout(struct intel_engine_cs *engine, return 0; /* Force a fast reset for terminated contexts (ignoring sysfs!) */ - if (unlikely(intel_context_is_banned(rq->context))) + if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq))) return 1; return READ_ONCE(engine->props.preempt_timeout_ms); diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index e7b4c4bc41a6..b4511ac05e9a 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -33,7 +33,10 @@ #include "gem/i915_gem_context.h" #include "gt/intel_breadcrumbs.h" #include "gt/intel_context.h" +#include "gt/intel_engine.h" +#include "gt/intel_engine_heartbeat.h" #include "gt/intel_gpu_commands.h" +#include "gt/intel_reset.h" #include "gt/intel_ring.h" #include "gt/intel_rps.h" @@ -429,20 +432,22 @@ void __i915_request_skip(struct i915_request *rq) rq->infix = rq->postfix; } -void i915_request_set_error_once(struct i915_request *rq, int error) +bool i915_request_set_error_once(struct i915_request *rq, int error) { int old; GEM_BUG_ON(!IS_ERR_VALUE((long)error)); if (i915_request_signaled(rq)) - return; + return false; old = READ_ONCE(rq->fence.error); do { if (fatal_error(old)) - return; + return false; } while (!try_cmpxchg(&rq->fence.error, &old, error)); + + return true; } struct i915_request *i915_request_mark_eio(struct i915_request *rq) @@ -609,6 +614,47 @@ void i915_request_unsubmit(struct i915_request *request) spin_unlock_irqrestore(&se->lock, flags); } +static struct intel_engine_cs *active_engine(struct i915_request *rq) +{ + struct intel_engine_cs *engine, *locked; + + locked = READ_ONCE(rq->engine); + spin_lock_irq(&locked->sched.lock); + while (unlikely(locked != (engine = READ_ONCE(rq->engine)))) { + spin_unlock(&locked->sched.lock); + locked = engine; + spin_lock(&locked->sched.lock); + } + + engine = NULL; + if (i915_request_is_active(rq) && !__i915_request_is_complete(rq)) + engine = locked; + + spin_unlock_irq(&locked->sched.lock); + + return engine; +} + +static void __cancel_request(struct i915_request *rq) +{ + struct intel_engine_cs *engine = active_engine(rq); + + if (engine && intel_engine_pulse(engine)) + intel_gt_handle_error(engine->gt, engine->mask, 0, + "request cancellation by %s", + current->comm); +} + +void i915_request_cancel(struct i915_request *rq, int error) +{ + if (!i915_request_set_error_once(rq, error)) + return; + + set_bit(I915_FENCE_FLAG_SENTINEL, &rq->fence.flags); + + __cancel_request(rq); +} + static int __i915_sw_fence_call submit_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state) { diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h index dd10a6db3d21..64869a313b3e 100644 --- a/drivers/gpu/drm/i915/i915_request.h +++ b/drivers/gpu/drm/i915/i915_request.h @@ -312,7 +312,7 @@ struct i915_request * __must_check i915_request_create(struct intel_context *ce); void __i915_request_skip(struct i915_request *rq); -void i915_request_set_error_once(struct i915_request *rq, int error); +bool i915_request_set_error_once(struct i915_request *rq, int error); struct i915_request *i915_request_mark_eio(struct i915_request *rq); struct i915_request *__i915_request_commit(struct i915_request *request); @@ -368,6 +368,8 @@ void i915_request_submit(struct i915_request *request); void __i915_request_unsubmit(struct i915_request *request); void i915_request_unsubmit(struct i915_request *request); +void i915_request_cancel(struct i915_request *rq, int error); + long i915_request_wait(struct i915_request *rq, unsigned int flags, long timeout) diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c index 8035ea7565ed..72f85d9d5e8b 100644 --- a/drivers/gpu/drm/i915/selftests/i915_request.c +++ b/drivers/gpu/drm/i915/selftests/i915_request.c @@ -608,6 +608,206 @@ static int live_nop_request(void *arg) return err; } +static int __cancel_inactive(struct intel_engine_cs *engine) +{ + struct intel_context *ce; + struct igt_spinner spin; + struct i915_request *rq; + int err = 0; + + if (igt_spinner_init(&spin, engine->gt)) + return -ENOMEM; + + ce = intel_context_create(engine); + if (IS_ERR(ce)) { + err = PTR_ERR(ce); + goto out_spin; + } + + rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto out_ce; + } + + pr_debug("%s: Cancelling inactive request\n", engine->name); + i915_request_cancel(rq, -EINTR); + i915_request_get(rq); + i915_request_add(rq); + + if (i915_request_wait(rq, 0, HZ / 5) < 0) { + struct drm_printer p = drm_info_printer(engine->i915->drm.dev); + + pr_err("%s: Failed to cancel inactive request\n", engine->name); + intel_engine_dump(engine, &p, "%s\n", engine->name); + err = -ETIME; + goto out_rq; + } + + if (rq->fence.error != -EINTR) { + pr_err("%s: fence not cancelled (%u)\n", + engine->name, rq->fence.error); + err = -EINVAL; + } + +out_rq: + i915_request_put(rq); +out_ce: + intel_context_put(ce); +out_spin: + igt_spinner_fini(&spin); + if (err) + pr_err("%s: %s error %d\n", __func__, engine->name, err); + return err; +} + +static int __cancel_active(struct intel_engine_cs *engine) +{ + struct intel_context *ce; + struct igt_spinner spin; + struct i915_request *rq; + int err = 0; + + if (igt_spinner_init(&spin, engine->gt)) + return -ENOMEM; + + ce = intel_context_create(engine); + if (IS_ERR(ce)) { + err = PTR_ERR(ce); + goto out_spin; + } + + rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto out_ce; + } + + pr_debug("%s: Cancelling active request\n", engine->name); + i915_request_get(rq); + i915_request_add(rq); + if (!igt_wait_for_spinner(&spin, rq)) { + struct drm_printer p = drm_info_printer(engine->i915->drm.dev); + + pr_err("Failed to start spinner on %s\n", engine->name); + intel_engine_dump(engine, &p, "%s\n", engine->name); + err = -ETIME; + goto out_rq; + } + i915_request_cancel(rq, -EINTR); + + if (i915_request_wait(rq, 0, HZ / 5) < 0) { + struct drm_printer p = drm_info_printer(engine->i915->drm.dev); + + pr_err("%s: Failed to cancel active request\n", engine->name); + intel_engine_dump(engine, &p, "%s\n", engine->name); + err = -ETIME; + goto out_rq; + } + + if (rq->fence.error != -EINTR) { + pr_err("%s: fence not cancelled (%u)\n", + engine->name, rq->fence.error); + err = -EINVAL; + } + +out_rq: + i915_request_put(rq); +out_ce: + intel_context_put(ce); +out_spin: + igt_spinner_fini(&spin); + if (err) + pr_err("%s: %s error %d\n", __func__, engine->name, err); + return err; +} + +static int __cancel_completed(struct intel_engine_cs *engine) +{ + struct intel_context *ce; + struct igt_spinner spin; + struct i915_request *rq; + int err = 0; + + if (igt_spinner_init(&spin, engine->gt)) + return -ENOMEM; + + ce = intel_context_create(engine); + if (IS_ERR(ce)) { + err = PTR_ERR(ce); + goto out_spin; + } + + rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto out_ce; + } + igt_spinner_end(&spin); + i915_request_get(rq); + i915_request_add(rq); + + if (i915_request_wait(rq, 0, HZ / 5) < 0) { + err = -ETIME; + goto out_rq; + } + + pr_debug("%s: Cancelling completed request\n", engine->name); + i915_request_cancel(rq, -EINTR); + if (rq->fence.error) { + pr_err("%s: fence not cancelled (%u)\n", + engine->name, rq->fence.error); + err = -EINVAL; + } + +out_rq: + i915_request_put(rq); +out_ce: + intel_context_put(ce); +out_spin: + igt_spinner_fini(&spin); + if (err) + pr_err("%s: %s error %d\n", __func__, engine->name, err); + return err; +} + +static int live_cancel_request(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine; + + /* + * Check cancellation of requests. We expect to be able to immediately + * cancel active requests, even if they are currently on the GPU. + */ + + for_each_uabi_engine(engine, i915) { + struct igt_live_test t; + int err, err2; + + if (!intel_engine_has_preemption(engine)) + continue; + + err = igt_live_test_begin(&t, i915, __func__, engine->name); + if (err) + return err; + + err = __cancel_inactive(engine); + if (err == 0) + err = __cancel_active(engine); + if (err == 0) + err = __cancel_completed(engine); + + err2 = igt_live_test_end(&t); + if (err) + return err; + if (err2) + return err2; + } + + return 0; +} + static struct i915_vma *empty_batch(struct drm_i915_private *i915) { struct drm_i915_gem_object *obj; @@ -1485,6 +1685,7 @@ int i915_request_live_selftests(struct drm_i915_private *i915) SUBTEST(live_sequential_engines), SUBTEST(live_parallel_engines), SUBTEST(live_empty_request), + SUBTEST(live_cancel_request), SUBTEST(live_breadcrumbs_smoketest), }; From patchwork Thu Mar 18 17:04:15 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tvrtko Ursulin X-Patchwork-Id: 12148683 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,HK_RANDOM_FROM,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id D9AFFC433E0 for ; Thu, 18 Mar 2021 17:04:38 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id A884464EF6 for ; Thu, 18 Mar 2021 17:04:38 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org A884464EF6 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id EE0CA6E937; Thu, 18 Mar 2021 17:04:30 +0000 (UTC) Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id 6A8426E934; Thu, 18 Mar 2021 17:04:29 +0000 (UTC) IronPort-SDR: Ko1UXn4Xaa9K9wmcBQuCmjGUIG9WRwPNVDy85w2fwtCXP7wHF0hx2H7Wuc3gqvEoCG9fdTn1zE 7TP48UwFTeqg== X-IronPort-AV: E=McAfee;i="6000,8403,9927"; a="209733842" X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="209733842" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:29 -0700 IronPort-SDR: UqeLmsDra9bJCW/T0qrztBc+lLR64HmAOmN6MzErnc5xFpmhtAZ+ZvAn/QHUe32kuz4SedouBz HCpXpRomNTzQ== X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="413164912" Received: from syeghiay-mobl.ger.corp.intel.com (HELO localhost.localdomain) ([10.252.51.182]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:27 -0700 From: Tvrtko Ursulin To: Intel-gfx@lists.freedesktop.org Subject: [PATCH 2/6] drm/i915: Restrict sentinel requests further Date: Thu, 18 Mar 2021 17:04:15 +0000 Message-Id: <20210318170419.2107512-3-tvrtko.ursulin@linux.intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> References: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: dri-devel@lists.freedesktop.org, Tvrtko Ursulin Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Tvrtko Ursulin Disallow sentinel requests follow previous sentinels to make request cancellation work better when faced with a chain of requests which have all been marked as in error. Signed-off-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/gt/intel_execlists_submission.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c index 4c2acb5a6c0a..4b870eca9693 100644 --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c @@ -896,7 +896,7 @@ static bool can_merge_rq(const struct i915_request *prev, if (__i915_request_is_complete(next)) return true; - if (unlikely((i915_request_flags(prev) ^ i915_request_flags(next)) & + if (unlikely((i915_request_flags(prev) | i915_request_flags(next)) & (BIT(I915_FENCE_FLAG_NOPREEMPT) | BIT(I915_FENCE_FLAG_SENTINEL)))) return false; From patchwork Thu Mar 18 17:04:16 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tvrtko Ursulin X-Patchwork-Id: 12148685 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,HK_RANDOM_FROM,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E78B5C43332 for ; Thu, 18 Mar 2021 17:04:41 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id B01A764E05 for ; Thu, 18 Mar 2021 17:04:41 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org B01A764E05 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 2EBEF6E934; Thu, 18 Mar 2021 17:04:34 +0000 (UTC) Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id F07FC6E935; Thu, 18 Mar 2021 17:04:31 +0000 (UTC) IronPort-SDR: IgOFYAT8AGDadI1/Rm5ycFYMiwp7KejMg9hQQcc8aU0VXhwwpQDS+tN4FEjfhvc+YqArbhMRDC tvVt4/hAt8nw== X-IronPort-AV: E=McAfee;i="6000,8403,9927"; a="209733847" X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="209733847" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:31 -0700 IronPort-SDR: MtdxoRNfcJPry6Lqi5cQi/Dp+FQb8RLVDPS/L0Yb+zPdryJrbug8jW/hq+EU/SZrOR0ejrimDN eVJVyXZ98FEQ== X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="413164945" Received: from syeghiay-mobl.ger.corp.intel.com (HELO localhost.localdomain) ([10.252.51.182]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:29 -0700 From: Tvrtko Ursulin To: Intel-gfx@lists.freedesktop.org Subject: [PATCH 3/6] drm/i915: Handle async cancellation in sentinel assert Date: Thu, 18 Mar 2021 17:04:16 +0000 Message-Id: <20210318170419.2107512-4-tvrtko.ursulin@linux.intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> References: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: dri-devel@lists.freedesktop.org, Tvrtko Ursulin Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Tvrtko Ursulin With the watchdog cancelling requests asynchronously to preempt-to-busy we need to relax one assert making it apply only to requests not in error. v2: * Check against the correct request! Signed-off-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/gt/intel_execlists_submission.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c index 4b870eca9693..bf557290173a 100644 --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c @@ -815,6 +815,13 @@ assert_pending_valid(const struct intel_engine_execlists *execlists, spin_unlock_irqrestore(&rq->lock, flags); if (!ok) return false; + + /* + * Due async nature of preempt-to-busy and request cancellation + * we need to skip further asserts for cancelled requests. + */ + if (READ_ONCE(rq->fence.error)) + break; } return ce; From patchwork Thu Mar 18 17:04:17 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tvrtko Ursulin X-Patchwork-Id: 12148687 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,HK_RANDOM_FROM,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 82A30C433E0 for ; Thu, 18 Mar 2021 17:04:43 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 4A0C564E05 for ; Thu, 18 Mar 2021 17:04:43 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 4A0C564E05 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id F3EF86E939; Thu, 18 Mar 2021 17:04:34 +0000 (UTC) Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id 7A99D6E938; Thu, 18 Mar 2021 17:04:33 +0000 (UTC) IronPort-SDR: 9H06xzeKD2MOg24R7X2HJP7c0vgrAq8aXt1drs41pAIIjwZ8kWkPCTdq/z1snZAxvxEDFHIIgJ Xg420RilBLzQ== X-IronPort-AV: E=McAfee;i="6000,8403,9927"; a="209733856" X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="209733856" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:32 -0700 IronPort-SDR: cHEAIbJmlGyoa8A7nWUwS5AZYSHAeDgvawMGCsSl5KUDtjL+h0/VNfVUUTrA0D4LTTzRHISvKS lebD6ifJ8URA== X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="413164967" Received: from syeghiay-mobl.ger.corp.intel.com (HELO localhost.localdomain) ([10.252.51.182]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:30 -0700 From: Tvrtko Ursulin To: Intel-gfx@lists.freedesktop.org Subject: [PATCH 4/6] drm/i915: Request watchdog infrastructure Date: Thu, 18 Mar 2021 17:04:17 +0000 Message-Id: <20210318170419.2107512-5-tvrtko.ursulin@linux.intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> References: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Daniel Vetter , dri-devel@lists.freedesktop.org, Tvrtko Ursulin Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Tvrtko Ursulin Prepares the plumbing for setting request/fence expiration time. All code is put in place but is never activeted due yet missing ability to actually configure the timer. Outline of the basic operation: A timer is started when request is ready for execution. If the request completes (retires) before the timer fires, timer is cancelled and nothing further happens. If the timer fires request is added to a lockless list and worker queued. Purpose of this is twofold: a) It allows request cancellation from a more friendly context and b) coalesces multiple expirations into a single event of consuming the list. Worker locklessly consumes the list of expired requests and cancels them all using previous added i915_request_cancel(). Associated timeout value is stored in rq->context.watchdog.timeout_us. v2: * Log expiration. Signed-off-by: Tvrtko Ursulin Cc: Daniel Vetter --- drivers/gpu/drm/i915/gt/intel_context_types.h | 4 ++ .../drm/i915/gt/intel_execlists_submission.h | 2 + drivers/gpu/drm/i915/gt/intel_gt.c | 3 + drivers/gpu/drm/i915/gt/intel_gt.h | 2 + drivers/gpu/drm/i915/gt/intel_gt_requests.c | 26 +++++++++ drivers/gpu/drm/i915/gt/intel_gt_types.h | 7 +++ drivers/gpu/drm/i915/i915_request.c | 56 +++++++++++++++++++ drivers/gpu/drm/i915/i915_request.h | 8 +++ 8 files changed, 108 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h index 0ea18c9e2aca..65a5730a4f5b 100644 --- a/drivers/gpu/drm/i915/gt/intel_context_types.h +++ b/drivers/gpu/drm/i915/gt/intel_context_types.h @@ -99,6 +99,10 @@ struct intel_context { #define CONTEXT_FORCE_SINGLE_SUBMISSION 7 #define CONTEXT_NOPREEMPT 8 + struct { + u64 timeout_us; + } watchdog; + u32 *lrc_reg_state; union { struct { diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.h b/drivers/gpu/drm/i915/gt/intel_execlists_submission.h index f7bd3fccfee8..4ca9b475e252 100644 --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.h +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.h @@ -6,6 +6,7 @@ #ifndef __INTEL_EXECLISTS_SUBMISSION_H__ #define __INTEL_EXECLISTS_SUBMISSION_H__ +#include #include struct drm_printer; @@ -13,6 +14,7 @@ struct drm_printer; struct i915_request; struct intel_context; struct intel_engine_cs; +struct intel_gt; enum { INTEL_CONTEXT_SCHEDULE_IN = 0, diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c index ca76f93bc03d..8d77dcbad059 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt.c +++ b/drivers/gpu/drm/i915/gt/intel_gt.c @@ -31,6 +31,9 @@ void intel_gt_init_early(struct intel_gt *gt, struct drm_i915_private *i915) INIT_LIST_HEAD(>->closed_vma); spin_lock_init(>->closed_lock); + init_llist_head(>->watchdog.list); + INIT_WORK(>->watchdog.work, intel_gt_watchdog_work); + intel_gt_init_buffer_pool(gt); intel_gt_init_reset(gt); intel_gt_init_requests(gt); diff --git a/drivers/gpu/drm/i915/gt/intel_gt.h b/drivers/gpu/drm/i915/gt/intel_gt.h index a17bd8b3195f..7ec395cace69 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt.h +++ b/drivers/gpu/drm/i915/gt/intel_gt.h @@ -78,4 +78,6 @@ static inline bool intel_gt_is_wedged(const struct intel_gt *gt) void intel_gt_info_print(const struct intel_gt_info *info, struct drm_printer *p); +void intel_gt_watchdog_work(struct work_struct *work); + #endif /* __INTEL_GT_H__ */ diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.c b/drivers/gpu/drm/i915/gt/intel_gt_requests.c index 36ec97f79174..a7f7bd662fb7 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_requests.c +++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.c @@ -8,6 +8,7 @@ #include "i915_drv.h" /* for_each_engine() */ #include "i915_request.h" #include "intel_engine_heartbeat.h" +#include "intel_execlists_submission.h" #include "intel_gt.h" #include "intel_gt_pm.h" #include "intel_gt_requests.h" @@ -242,4 +243,29 @@ void intel_gt_fini_requests(struct intel_gt *gt) { /* Wait until the work is marked as finished before unloading! */ cancel_delayed_work_sync(>->requests.retire_work); + + flush_work(>->watchdog.work); +} + +void intel_gt_watchdog_work(struct work_struct *work) +{ + struct intel_gt *gt = + container_of(work, typeof(*gt), watchdog.work); + struct i915_request *rq, *rn; + struct llist_node *first; + + first = llist_del_all(>->watchdog.list); + if (!first) + return; + + llist_for_each_entry_safe(rq, rn, first, watchdog.link) { + if (!i915_request_completed(rq)) { + drm_notice(>->i915->drm, + "Fence expiration time out %llx:%llu!\n", + rq->fence.context, + rq->fence.seqno); + i915_request_cancel(rq, -EINTR); + } + i915_request_put(rq); + } } diff --git a/drivers/gpu/drm/i915/gt/intel_gt_types.h b/drivers/gpu/drm/i915/gt/intel_gt_types.h index 626af37c7790..d70ebcc6f19f 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_types.h +++ b/drivers/gpu/drm/i915/gt/intel_gt_types.h @@ -8,10 +8,12 @@ #include #include +#include #include #include #include #include +#include #include "uc/intel_uc.h" @@ -62,6 +64,11 @@ struct intel_gt { struct delayed_work retire_work; } requests; + struct { + struct llist_head list; + struct work_struct work; + } watchdog; + struct intel_wakeref wakeref; atomic_t user_wakeref; diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index b4511ac05e9a..9dd5e588b0a4 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -277,6 +277,57 @@ static void remove_from_engine(struct i915_request *rq) __notify_execute_cb_imm(rq); } +static void __rq_init_watchdog(struct i915_request *rq) +{ + rq->watchdog.timer.function = NULL; +} + +static enum hrtimer_restart __rq_watchdog_expired(struct hrtimer *hrtimer) +{ + struct i915_request *rq = + container_of(hrtimer, struct i915_request, watchdog.timer); + struct intel_gt *gt = rq->engine->gt; + + if (!i915_request_completed(rq)) { + if (llist_add(&rq->watchdog.link, >->watchdog.list)) + schedule_work(>->watchdog.work); + } else { + i915_request_put(rq); + } + + return HRTIMER_NORESTART; +} + +static void __rq_arm_watchdog(struct i915_request *rq) +{ + struct i915_request_watchdog *wdg = &rq->watchdog; + struct intel_context *ce = rq->context; + + if (!ce->watchdog.timeout_us) + return; + + hrtimer_init(&wdg->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + wdg->timer.function = __rq_watchdog_expired; + hrtimer_start_range_ns(&wdg->timer, + ns_to_ktime(ce->watchdog.timeout_us * + NSEC_PER_USEC), + /* + * FIXME check if it gives the "not sooner" + * guarantee or slack is both ways + */ + NSEC_PER_MSEC, + HRTIMER_MODE_REL); + i915_request_get(rq); +} + +static void __rq_cancel_watchdog(struct i915_request *rq) +{ + struct i915_request_watchdog *wdg = &rq->watchdog; + + if (wdg->timer.function && hrtimer_try_to_cancel(&wdg->timer) > 0) + i915_request_put(rq); +} + bool i915_request_retire(struct i915_request *rq) { if (!__i915_request_is_complete(rq)) @@ -288,6 +339,8 @@ bool i915_request_retire(struct i915_request *rq) trace_i915_request_retire(rq); i915_request_mark_complete(rq); + __rq_cancel_watchdog(rq); + /* * We know the GPU must have read the request to have * sent us the seqno + interrupt, so use the position @@ -667,6 +720,8 @@ submit_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state) if (unlikely(fence->error)) i915_request_set_error_once(request, fence->error); + else + __rq_arm_watchdog(request); /* * We need to serialize use of the submit_request() callback @@ -854,6 +909,7 @@ __i915_request_create(struct intel_context *ce, gfp_t gfp) /* No zalloc, everything must be cleared after use */ rq->batch = NULL; + __rq_init_watchdog(rq); GEM_BUG_ON(rq->capture_list); GEM_BUG_ON(!llist_empty(&rq->execute_cb)); diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h index 64869a313b3e..294f16e2163d 100644 --- a/drivers/gpu/drm/i915/i915_request.h +++ b/drivers/gpu/drm/i915/i915_request.h @@ -26,7 +26,9 @@ #define I915_REQUEST_H #include +#include #include +#include #include #include "gem/i915_gem_context_types.h" @@ -289,6 +291,12 @@ struct i915_request { /** timeline->request entry for this request */ struct list_head link; + /** Watchdog support fields. */ + struct i915_request_watchdog { + struct llist_node link; + struct hrtimer timer; + } watchdog; + I915_SELFTEST_DECLARE(struct { struct list_head link; unsigned long delay; From patchwork Thu Mar 18 17:04:18 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tvrtko Ursulin X-Patchwork-Id: 12148689 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,HK_RANDOM_FROM,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id A816BC433E6 for ; Thu, 18 Mar 2021 17:04:46 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 71A0564E38 for ; Thu, 18 Mar 2021 17:04:46 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 71A0564E38 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 754E36E93E; Thu, 18 Mar 2021 17:04:39 +0000 (UTC) Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id 400196E93A; Thu, 18 Mar 2021 17:04:35 +0000 (UTC) IronPort-SDR: QejXm9u9fB/b7Pc8Fo6c+Q/G2IyiSnwgXeHsVGJ+rIA0SS67BxYjdnZd12Sh/XY8vl8/lqMq86 y5ckEvMiPhjA== X-IronPort-AV: E=McAfee;i="6000,8403,9927"; a="209733871" X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="209733871" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:34 -0700 IronPort-SDR: GsABnf5FETu8uldvEh6wrY0PWcy/qu2KJtf18A8u+WdSO9jeQM9UlzKDVSQC2dX+JrLOSUzdPS QYB/syDlTm2g== X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="413164990" Received: from syeghiay-mobl.ger.corp.intel.com (HELO localhost.localdomain) ([10.252.51.182]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:33 -0700 From: Tvrtko Ursulin To: Intel-gfx@lists.freedesktop.org Subject: [PATCH 5/6] drm/i915: Fail too long user submissions by default Date: Thu, 18 Mar 2021 17:04:18 +0000 Message-Id: <20210318170419.2107512-6-tvrtko.ursulin@linux.intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> References: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Daniel Vetter , dri-devel@lists.freedesktop.org, Tvrtko Ursulin Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Tvrtko Ursulin A new Kconfig option CONFIG_DRM_I915_REQUEST_TIMEOUT is added, defaulting to 20s, and this timeout is applied to all users contexts using the previously added watchdog facility. Result of this is that any user submission will simply fail after this timeout, either causing a reset (for non-preemptable), or incomplete results. This can have an effect that workloads which used to work fine will suddenly start failing. Even workloads comprised of short batches but in long dependency chains can be terminated. And becuase of lack of agreement on usefulness and safety of fence error propagation this partial execution can be invisible to userspace even if it is "listening" to returned fence status. Another interaction is with hangcheck where care needs to be taken timeout is not set lower or close to three times the heartbeat interval. Otherwise a hang in any application can cause complete termination of all submissions from unrelated clients. Any users modifying the per engine heartbeat intervals therefore need to be aware of this potential denial of service to avoid inadvertently enabling it. Given all this I am personally not convinced the scheme is a good idea. Intuitively it feels object importers would be better positioned to enforce the time they are willing to wait for something to complete. v2: * Improved commit message and Kconfig text. * Pull in some helper code from patch which got dropped. v3: * Bump timeout to 20s to see if it helps Tigerlake. Signed-off-by: Tvrtko Ursulin Cc: Daniel Vetter Acked-by: Matthew Auld --- drivers/gpu/drm/i915/Kconfig.profile | 14 +++++++ drivers/gpu/drm/i915/gem/i915_gem_context.c | 38 +++++++++++++++++++ .../gpu/drm/i915/gem/i915_gem_context_types.h | 4 ++ drivers/gpu/drm/i915/gt/intel_context_param.h | 11 +++++- 4 files changed, 66 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/Kconfig.profile b/drivers/gpu/drm/i915/Kconfig.profile index 35bbe2b80596..39328567c200 100644 --- a/drivers/gpu/drm/i915/Kconfig.profile +++ b/drivers/gpu/drm/i915/Kconfig.profile @@ -1,3 +1,17 @@ +config DRM_I915_REQUEST_TIMEOUT + int "Default timeout for requests (ms)" + default 20000 # milliseconds + help + Configures the default timeout after which any user submissions will + be forcefully terminated. + + Beware setting this value lower, or close to heartbeat interval + rounded to whole seconds times three, in order to avoid allowing + misbehaving applications causing total rendering failure in unrelated + clients. + + May be 0 to disable the timeout. + config DRM_I915_FENCE_TIMEOUT int "Timeout for unsignaled foreign fences (ms, jiffy granularity)" default 10000 # milliseconds diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c index ca37d93ef5e7..be71be21800b 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c @@ -233,6 +233,8 @@ static void intel_context_set_gem(struct intel_context *ce, if (ctx->sched.priority >= I915_PRIORITY_NORMAL && intel_engine_has_timeslices(ce->engine)) __set_bit(CONTEXT_USE_SEMAPHORES, &ce->flags); + + intel_context_set_watchdog_us(ce, ctx->watchdog.timeout_us); } static void __free_engines(struct i915_gem_engines *e, unsigned int count) @@ -852,6 +854,40 @@ static void __assign_timeline(struct i915_gem_context *ctx, context_apply_all(ctx, __apply_timeline, timeline); } +static int __apply_watchdog(struct intel_context *ce, void *timeout_us) +{ + return intel_context_set_watchdog_us(ce, (uintptr_t)timeout_us); +} + +static int +__set_watchdog(struct i915_gem_context *ctx, unsigned long timeout_us) +{ + int ret; + + ret = context_apply_all(ctx, __apply_watchdog, + (void *)(uintptr_t)timeout_us); + if (!ret) + ctx->watchdog.timeout_us = timeout_us; + + return ret; +} + +static void __set_default_fence_expiry(struct i915_gem_context *ctx) +{ + struct drm_i915_private *i915 = ctx->i915; + int ret; + + if (!IS_ACTIVE(CONFIG_DRM_I915_REQUEST_TIMEOUT)) + return; + + /* Default expiry for user fences. */ + ret = __set_watchdog(ctx, CONFIG_DRM_I915_REQUEST_TIMEOUT * 1000); + if (ret) + drm_notice(&i915->drm, + "Failed to configure default fence expiry! (%d)", + ret); +} + static struct i915_gem_context * i915_gem_create_context(struct drm_i915_private *i915, unsigned int flags) { @@ -896,6 +932,8 @@ i915_gem_create_context(struct drm_i915_private *i915, unsigned int flags) intel_timeline_put(timeline); } + __set_default_fence_expiry(ctx); + trace_i915_context_create(ctx); return ctx; diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h index d5bc75508048..f17da7e26c43 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h @@ -150,6 +150,10 @@ struct i915_gem_context { */ atomic_t active_count; + struct { + u64 timeout_us; + } watchdog; + /** * @hang_timestamp: The last time(s) this context caused a GPU hang */ diff --git a/drivers/gpu/drm/i915/gt/intel_context_param.h b/drivers/gpu/drm/i915/gt/intel_context_param.h index f053d8633fe2..3ecacc675f41 100644 --- a/drivers/gpu/drm/i915/gt/intel_context_param.h +++ b/drivers/gpu/drm/i915/gt/intel_context_param.h @@ -6,9 +6,18 @@ #ifndef INTEL_CONTEXT_PARAM_H #define INTEL_CONTEXT_PARAM_H -struct intel_context; +#include + +#include "intel_context.h" int intel_context_set_ring_size(struct intel_context *ce, long sz); long intel_context_get_ring_size(struct intel_context *ce); +static inline int +intel_context_set_watchdog_us(struct intel_context *ce, u64 timeout_us) +{ + ce->watchdog.timeout_us = timeout_us; + return 0; +} + #endif /* INTEL_CONTEXT_PARAM_H */ From patchwork Thu Mar 18 17:04:19 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tvrtko Ursulin X-Patchwork-Id: 12148691 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,HK_RANDOM_FROM,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 388B5C433E0 for ; Thu, 18 Mar 2021 17:04:48 +0000 (UTC) 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 mail.kernel.org (Postfix) with ESMTPS id 0697864E38 for ; Thu, 18 Mar 2021 17:04:48 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0697864E38 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 765C56E93F; Thu, 18 Mar 2021 17:04:40 +0000 (UTC) Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by gabe.freedesktop.org (Postfix) with ESMTPS id 0644D6E93C; Thu, 18 Mar 2021 17:04:38 +0000 (UTC) IronPort-SDR: lH7OZbNLLmHcKNEdoQVI1Kpoeh7CkLcMEK4LuQao4dbHgpH2f2nkfjo0wLSXczNacgGMdbsWQG EVhr7FPm0mjw== X-IronPort-AV: E=McAfee;i="6000,8403,9927"; a="209733885" X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="209733885" Received: from orsmga008.jf.intel.com ([10.7.209.65]) by fmsmga101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:38 -0700 IronPort-SDR: EGCRi+aOWHl4dQwDkRxtSzuEVQ3B2egVS8rZiUaoWrPAl/7RwIN6iJBEOtExeAcFqzBc6yl2XX t5fJ4gfPyuGw== X-IronPort-AV: E=Sophos;i="5.81,259,1610438400"; d="scan'208";a="413165005" Received: from syeghiay-mobl.ger.corp.intel.com (HELO localhost.localdomain) ([10.252.51.182]) by orsmga008-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 18 Mar 2021 10:04:34 -0700 From: Tvrtko Ursulin To: Intel-gfx@lists.freedesktop.org Subject: [PATCH 6/6] drm/i915: Allow configuring default request expiry via modparam Date: Thu, 18 Mar 2021 17:04:19 +0000 Message-Id: <20210318170419.2107512-7-tvrtko.ursulin@linux.intel.com> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> References: <20210318170419.2107512-1-tvrtko.ursulin@linux.intel.com> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Daniel Vetter , dri-devel@lists.freedesktop.org, Tvrtko Ursulin Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Tvrtko Ursulin Module parameter is added (request_timeout_ms) to allow configuring the default request/fence expiry. Default value is inherited from CONFIG_DRM_I915_REQUEST_TIMEOUT. Signed-off-by: Tvrtko Ursulin Cc: Daniel Vetter Acked-by: Daniel Vetter --- drivers/gpu/drm/i915/gem/i915_gem_context.c | 5 +++-- drivers/gpu/drm/i915/i915_params.c | 5 +++++ drivers/gpu/drm/i915/i915_params.h | 1 + 3 files changed, 9 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c index be71be21800b..cf5b706a39b8 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c @@ -877,11 +877,12 @@ static void __set_default_fence_expiry(struct i915_gem_context *ctx) struct drm_i915_private *i915 = ctx->i915; int ret; - if (!IS_ACTIVE(CONFIG_DRM_I915_REQUEST_TIMEOUT)) + if (!IS_ACTIVE(CONFIG_DRM_I915_REQUEST_TIMEOUT) || + !i915->params.request_timeout_ms) return; /* Default expiry for user fences. */ - ret = __set_watchdog(ctx, CONFIG_DRM_I915_REQUEST_TIMEOUT * 1000); + ret = __set_watchdog(ctx, i915->params.request_timeout_ms * 1000); if (ret) drm_notice(&i915->drm, "Failed to configure default fence expiry! (%d)", diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c index 6939634e56ed..0320878d96b0 100644 --- a/drivers/gpu/drm/i915/i915_params.c +++ b/drivers/gpu/drm/i915/i915_params.c @@ -197,6 +197,11 @@ i915_param_named_unsafe(fake_lmem_start, ulong, 0400, "Fake LMEM start offset (default: 0)"); #endif +#if CONFIG_DRM_I915_REQUEST_TIMEOUT +i915_param_named_unsafe(request_timeout_ms, uint, 0600, + "Default request/fence/batch buffer expiration timeout."); +#endif + static __always_inline void _print_param(struct drm_printer *p, const char *name, const char *type, diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h index 48f47e44e848..34ebb0662547 100644 --- a/drivers/gpu/drm/i915/i915_params.h +++ b/drivers/gpu/drm/i915/i915_params.h @@ -72,6 +72,7 @@ struct drm_printer; param(int, enable_dpcd_backlight, -1, 0600) \ param(char *, force_probe, CONFIG_DRM_I915_FORCE_PROBE, 0400) \ param(unsigned long, fake_lmem_start, 0, 0400) \ + param(unsigned int, request_timeout_ms, CONFIG_DRM_I915_REQUEST_TIMEOUT, 0600) \ /* leave bools at the end to not create holes */ \ param(bool, enable_hangcheck, true, 0600) \ param(bool, load_detect_test, false, 0600) \