From patchwork Thu Nov 29 10:05:19 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Sharat Masetty X-Patchwork-Id: 10704279 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 74EA2109C for ; Thu, 29 Nov 2018 10:05:32 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 65FBB2E081 for ; Thu, 29 Nov 2018 10:05:32 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 5A6442E09B; Thu, 29 Nov 2018 10:05:32 +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=-7.7 required=2.0 tests=BAYES_00,DKIM_INVALID, DKIM_SIGNED,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 4BE7B2E081 for ; Thu, 29 Nov 2018 10:05:31 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726879AbeK2VKS (ORCPT ); Thu, 29 Nov 2018 16:10:18 -0500 Received: from smtp.codeaurora.org ([198.145.29.96]:45410 "EHLO smtp.codeaurora.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726740AbeK2VKS (ORCPT ); Thu, 29 Nov 2018 16:10:18 -0500 Received: by smtp.codeaurora.org (Postfix, from userid 1000) id 0B26260B7A; Thu, 29 Nov 2018 10:05:30 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=codeaurora.org; s=default; t=1543485930; bh=sHAJ9dA8QuV+S1Fh8N2VFAmtzlDXh7GsS0dH3RkMbVw=; h=From:To:Cc:Subject:Date:From; b=fT1CJ5KWqwjwMVY3B8pDlZBzZfYyxakWE0NlApup8OJu3ihEsQNB8cz9ycrIvA0vA pXIF0RWr7tcuYn5kGnUyu3gNflzMtUYC1DzNuwZm5//08U5CnaVEGRS4yCGNDpup+g FOPgZGQqdsErcAG1hYQIFjXqnyOPNZ1Ft2WFNKeA= Received: from smasetty-linux.qualcomm.com (blr-c-bdr-fw-01_globalnat_allzones-outside.qualcomm.com [103.229.19.19]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: smasetty@smtp.codeaurora.org) by smtp.codeaurora.org (Postfix) with ESMTPSA id A007D601D1; Thu, 29 Nov 2018 10:05:27 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=codeaurora.org; s=default; t=1543485929; bh=sHAJ9dA8QuV+S1Fh8N2VFAmtzlDXh7GsS0dH3RkMbVw=; h=From:To:Cc:Subject:Date:From; b=QQWDH8zM2A0qiJNdcPJux+pQH/NhsnPCreDdJUBztJHuxiRwszeqx6Pi4v4umLzwD zXF+/vvWHSJR++Zd54XJpXoVL/WHhVSVr12eid6OVLQnbqAQDmCr/VcvQfiRvXaKe4 EAcBUb7Ju22/1vSKXU9ALYPG9rTCGlaIAaiEf+RE= DMARC-Filter: OpenDMARC Filter v1.3.2 smtp.codeaurora.org A007D601D1 Authentication-Results: pdx-caf-mail.web.codeaurora.org; dmarc=none (p=none dis=none) header.from=codeaurora.org Authentication-Results: pdx-caf-mail.web.codeaurora.org; spf=none smtp.mailfrom=smasetty@codeaurora.org From: Sharat Masetty To: freedreno@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org, linux-arm-msm@vger.kernel.org, christian.koenig@amd.com, Sharat Masetty Subject: [PATCH 1/2] drm/scheduler: Set sched->thread to NULL on failure Date: Thu, 29 Nov 2018 15:35:19 +0530 Message-Id: <1543485920-28037-1-git-send-email-smasetty@codeaurora.org> X-Mailer: git-send-email 1.9.1 Sender: linux-arm-msm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-arm-msm@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP In cases where the scheduler instance is used as a base object of another driver object, it's not clear if the driver can call scheduler cleanup on the fail path. So, Set the sched->thread to NULL, so that the driver can safely call drm_sched_fini() during cleanup. Signed-off-by: Sharat Masetty Reviewed-by: Christian König --- drivers/gpu/drm/scheduler/sched_main.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 44fe587..c993d10 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -594,7 +594,7 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, long timeout, const char *name) { - int i; + int i, ret; sched->ops = ops; sched->hw_submission_limit = hw_submission; sched->name = name; @@ -615,8 +615,10 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, /* Each scheduler will run on a seperate kernel thread */ sched->thread = kthread_run(drm_sched_main, sched, sched->name); if (IS_ERR(sched->thread)) { + ret = PTR_ERR(sched->thread); + sched->thread = NULL; DRM_ERROR("Failed to create scheduler for %s.\n", name); - return PTR_ERR(sched->thread); + return ret; } return 0; From patchwork Thu Nov 29 10:05:20 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sharat Masetty X-Patchwork-Id: 10704285 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 90D37109C for ; Thu, 29 Nov 2018 10:05:35 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 821E02E081 for ; Thu, 29 Nov 2018 10:05:35 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 760DE2E08A; Thu, 29 Nov 2018 10:05:35 +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=-7.7 required=2.0 tests=BAYES_00,DKIM_INVALID, DKIM_SIGNED,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 121EB2E081 for ; Thu, 29 Nov 2018 10:05:34 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726926AbeK2VKV (ORCPT ); Thu, 29 Nov 2018 16:10:21 -0500 Received: from smtp.codeaurora.org ([198.145.29.96]:45568 "EHLO smtp.codeaurora.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726740AbeK2VKV (ORCPT ); Thu, 29 Nov 2018 16:10:21 -0500 Received: by smtp.codeaurora.org (Postfix, from userid 1000) id 6593F6060A; Thu, 29 Nov 2018 10:05:32 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=codeaurora.org; s=default; t=1543485932; bh=Cb7Xud+Cp87cXOPspFj3lRJiLnWtSuQIQFirkG54/vc=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=SS5Gai4Kqqdeh9cOBt64wY8IAHOlT0jn0gYOe1rBjWIMachxbf9zrnsJX64jZSQwk L/ak7cZTrPj+1Xl1R1dyHw6TYkMSzFcAeDKmQXS6r7S3Wr6OvFyC1lNHnL5z5Z/1wO aZ6ZkfBlkS2p9PqAy+fXWZO6kbv3ol+Z45WEvc/s= Received: from smasetty-linux.qualcomm.com (blr-c-bdr-fw-01_globalnat_allzones-outside.qualcomm.com [103.229.19.19]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: smasetty@smtp.codeaurora.org) by smtp.codeaurora.org (Postfix) with ESMTPSA id F40646060A; Thu, 29 Nov 2018 10:05:29 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=codeaurora.org; s=default; t=1543485931; bh=Cb7Xud+Cp87cXOPspFj3lRJiLnWtSuQIQFirkG54/vc=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=iu64q2v/Xu0U4H/9ITrs+fAwIzC33HA2nExeDjbfsJf+Mle1xXZVMeAuVU86tHbj2 /r1uzZFh0rOLVvFWiJGBz2568VdYy7q4RDGuYJBQv1WpSma9EnsOufHAP22apC1eqs q1Y8TU04nDI6JXCbHL+SHImDHDg3Qo+01iO8UGWQ= DMARC-Filter: OpenDMARC Filter v1.3.2 smtp.codeaurora.org F40646060A Authentication-Results: pdx-caf-mail.web.codeaurora.org; dmarc=none (p=none dis=none) header.from=codeaurora.org Authentication-Results: pdx-caf-mail.web.codeaurora.org; spf=none smtp.mailfrom=smasetty@codeaurora.org From: Sharat Masetty To: freedreno@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org, linux-arm-msm@vger.kernel.org, christian.koenig@amd.com, Sharat Masetty Subject: [PATCH 2/2] drm/scheduler: Add drm_sched_suspend/resume_timeout() Date: Thu, 29 Nov 2018 15:35:20 +0530 Message-Id: <1543485920-28037-2-git-send-email-smasetty@codeaurora.org> X-Mailer: git-send-email 1.9.1 In-Reply-To: <1543485920-28037-1-git-send-email-smasetty@codeaurora.org> References: <1543485920-28037-1-git-send-email-smasetty@codeaurora.org> Sender: linux-arm-msm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-arm-msm@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP This patch adds two new functions to help client drivers suspend and resume the scheduler job timeout. This can be useful in cases where the hardware has preemption support enabled. Using this, it is possible to have the timeout active only for the ring which is active on the ringbuffer. This patch also makes the job_list_lock IRQ safe. Suggested-by: Christian Koenig Signed-off-by: Sharat Masetty --- drivers/gpu/drm/etnaviv/etnaviv_dump.c | 9 ++-- drivers/gpu/drm/scheduler/sched_main.c | 91 ++++++++++++++++++++++++++++------ include/drm/gpu_scheduler.h | 4 ++ 3 files changed, 86 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/etnaviv/etnaviv_dump.c b/drivers/gpu/drm/etnaviv/etnaviv_dump.c index 9146e30..fd6bad2 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_dump.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_dump.c @@ -118,6 +118,7 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) unsigned int n_obj, n_bomap_pages; size_t file_size, mmu_size; __le64 *bomap, *bomap_start; + unsigned long flags; /* Only catch the first event, or when manually re-armed */ if (!etnaviv_dump_core) @@ -134,13 +135,13 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) mmu_size + gpu->buffer.size; /* Add in the active command buffers */ - spin_lock(&gpu->sched.job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry(s_job, &gpu->sched.ring_mirror_list, node) { submit = to_etnaviv_submit(s_job); file_size += submit->cmdbuf.size; n_obj++; } - spin_unlock(&gpu->sched.job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); /* Add in the active buffer objects */ list_for_each_entry(vram, &gpu->mmu->mappings, mmu_node) { @@ -182,14 +183,14 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) gpu->buffer.size, etnaviv_cmdbuf_get_va(&gpu->buffer)); - spin_lock(&gpu->sched.job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry(s_job, &gpu->sched.ring_mirror_list, node) { submit = to_etnaviv_submit(s_job); etnaviv_core_dump_mem(&iter, ETDUMP_BUF_CMD, submit->cmdbuf.vaddr, submit->cmdbuf.size, etnaviv_cmdbuf_get_va(&submit->cmdbuf)); } - spin_unlock(&gpu->sched.job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); /* Reserve space for the bomap */ if (n_bomap_pages) { diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index c993d10..ca09b4e 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -196,10 +196,68 @@ static void drm_sched_start_timeout(struct drm_gpu_scheduler *sched) schedule_delayed_work(&sched->work_tdr, sched->timeout); } +/** + * drm_sched_suspend_timeout - Suspend scheduler job timeout + * + * @sched: scheduler instance for which to suspend the timeout + * + * Suspend the delayed work timeout for the scheduler. This is done by + * modifying the delayed work timeout to an arbitrary large value, + * MAX_SCHEDULE_TIMEOUT in this case. Note that this function can be + * called from an IRQ context. + * + * Returns the timeout remaining + * + */ +unsigned long drm_sched_suspend_timeout(struct drm_gpu_scheduler *sched) +{ + unsigned long sched_timeout, now = jiffies; + + sched_timeout = sched->work_tdr.timer.expires; + + /* + * Modify the timeout to an arbitrarily large value. This also prevents + * the timeout to be restarted when new submissions arrive + */ + if (mod_delayed_work(system_wq, &sched->work_tdr, MAX_SCHEDULE_TIMEOUT) + && time_after(sched_timeout, now)) + return sched_timeout - now; + else + return sched->timeout; +} +EXPORT_SYMBOL(drm_sched_suspend_timeout); + +/** + * drm_sched_resume_timeout - Resume scheduler job timeout + * + * @sched: scheduler instance for which to resume the timeout + * @remaining: remaining timeout + * + * Resume the delayed work timeout for the scheduler. Note that + * this function can be called from an IRQ context. + */ +void drm_sched_resume_timeout(struct drm_gpu_scheduler *sched, + unsigned long remaining) +{ + unsigned long flags; + + spin_lock_irqsave(&sched->job_list_lock, flags); + + if (list_empty(&sched->ring_mirror_list)) + cancel_delayed_work(&sched->work_tdr); + else + mod_delayed_work(system_wq, &sched->work_tdr, remaining); + + spin_unlock_irqrestore(&sched->job_list_lock, flags); +} +EXPORT_SYMBOL(drm_sched_resume_timeout); + /* job_finish is called after hw fence signaled */ static void drm_sched_job_finish(struct work_struct *work) { + unsigned long flags; + struct drm_sched_job *s_job = container_of(work, struct drm_sched_job, finish_work); struct drm_gpu_scheduler *sched = s_job->sched; @@ -213,12 +271,12 @@ static void drm_sched_job_finish(struct work_struct *work) */ cancel_delayed_work_sync(&sched->work_tdr); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); /* remove job from ring_mirror_list */ list_del(&s_job->node); /* queue TDR for next job */ drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); dma_fence_put(&s_job->s_fence->finished); sched->ops->free_job(s_job); @@ -234,15 +292,17 @@ static void drm_sched_job_finish_cb(struct dma_fence *f, static void drm_sched_job_begin(struct drm_sched_job *s_job) { + unsigned long flags; + struct drm_gpu_scheduler *sched = s_job->sched; dma_fence_add_callback(&s_job->s_fence->finished, &s_job->finish_cb, drm_sched_job_finish_cb); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_add_tail(&s_job->node, &sched->ring_mirror_list); drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); } static void drm_sched_job_timedout(struct work_struct *work) @@ -250,10 +310,11 @@ static void drm_sched_job_timedout(struct work_struct *work) struct drm_gpu_scheduler *sched; struct drm_sched_job *job; int r; + unsigned long flags; sched = container_of(work, struct drm_gpu_scheduler, work_tdr.work); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry_reverse(job, &sched->ring_mirror_list, node) { struct drm_sched_fence *fence = job->s_fence; @@ -263,12 +324,12 @@ static void drm_sched_job_timedout(struct work_struct *work) job = list_first_entry_or_null(&sched->ring_mirror_list, struct drm_sched_job, node); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); if (job) sched->ops->timedout_job(job); - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry(job, &sched->ring_mirror_list, node) { struct drm_sched_fence *fence = job->s_fence; @@ -283,7 +344,7 @@ static void drm_sched_job_timedout(struct work_struct *work) already_signaled: ; } - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); } /** @@ -298,8 +359,9 @@ void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, struct drm_sched_jo struct drm_sched_job *s_job; struct drm_sched_entity *entity, *tmp; int i; + unsigned long flags; - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry_reverse(s_job, &sched->ring_mirror_list, node) { if (s_job->s_fence->parent && dma_fence_remove_callback(s_job->s_fence->parent, @@ -309,7 +371,7 @@ void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, struct drm_sched_jo atomic_dec(&sched->hw_rq_count); } } - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); if (bad && bad->s_priority != DRM_SCHED_PRIORITY_KERNEL) { atomic_inc(&bad->karma); @@ -348,8 +410,9 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) struct drm_sched_job *s_job, *tmp; bool found_guilty = false; int r; + unsigned long flags; - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) { struct drm_sched_fence *s_fence = s_job->s_fence; struct dma_fence *fence; @@ -363,7 +426,7 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) if (found_guilty && s_job->s_fence->scheduled.context == guilty_context) dma_fence_set_error(&s_fence->finished, -ECANCELED); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); fence = sched->ops->run_job(s_job); atomic_inc(&sched->hw_rq_count); @@ -380,10 +443,10 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) } else { drm_sched_process_job(NULL, &s_fence->cb); } - spin_lock(&sched->job_list_lock); + spin_lock_irqsave(&sched->job_list_lock, flags); } drm_sched_start_timeout(sched); - spin_unlock(&sched->job_list_lock); + spin_unlock_irqrestore(&sched->job_list_lock, flags); } EXPORT_SYMBOL(drm_sched_job_recovery); diff --git a/include/drm/gpu_scheduler.h b/include/drm/gpu_scheduler.h index d87b268..c9657a8 100644 --- a/include/drm/gpu_scheduler.h +++ b/include/drm/gpu_scheduler.h @@ -326,4 +326,8 @@ struct drm_sched_fence *drm_sched_fence_create( void drm_sched_fence_scheduled(struct drm_sched_fence *fence); void drm_sched_fence_finished(struct drm_sched_fence *fence); +unsigned long drm_sched_suspend_timeout(struct drm_gpu_scheduler *sched); +void drm_sched_resume_timeout(struct drm_gpu_scheduler *sched, + unsigned long remaining); + #endif