From patchwork Wed May 4 16:45:07 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Greg KH X-Patchwork-Id: 569670 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id A473DC433F5 for ; Wed, 4 May 2022 17:13:42 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1350628AbiEDRRO (ORCPT ); Wed, 4 May 2022 13:17:14 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:50158 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1354608AbiEDRHU (ORCPT ); Wed, 4 May 2022 13:07:20 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B6373515AE; Wed, 4 May 2022 09:54:25 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 6BCC6617DE; Wed, 4 May 2022 16:54:25 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id BA418C385A5; Wed, 4 May 2022 16:54:24 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1651683264; bh=FB6S71sPeFP0H53YPfgswpGWG6zlz1yjWXv/ilUPvWE=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=vuL7RtDrmsD/pszgwocosQWmtEhoqK8f7y6i8WSrCQDdKTgF3xCE8uz8tSc4p4w8s umq/YNORJfXr7AzLelS6pDWHA2ZcIzVDxc6xvoQpC31l70hMSymou58YLre8cCNQHO bLDv8aC+tCH9Pb3BqZMLy03+25vBTfxX5/omOftE= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, David Yat Sin , Felix Kuehling , Alex Deucher , Sasha Levin Subject: [PATCH 5.15 114/177] drm/amdkfd: Fix GWS queue count Date: Wed, 4 May 2022 18:45:07 +0200 Message-Id: <20220504153103.397174871@linuxfoundation.org> X-Mailer: git-send-email 2.36.0 In-Reply-To: <20220504153053.873100034@linuxfoundation.org> References: <20220504153053.873100034@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: stable@vger.kernel.org From: David Yat Sin [ Upstream commit 7c6b6e18c890f30965b0589b0a57645e1dbccfde ] dqm->gws_queue_count and pdd->qpd.mapped_gws_queue need to be updated each time the queue gets evicted. Fixes: b8020b0304c8 ("drm/amdkfd: Enable over-subscription with >1 GWS queue") Signed-off-by: David Yat Sin Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher Signed-off-by: Sasha Levin --- .../drm/amd/amdkfd/kfd_device_queue_manager.c | 83 +++++++++---------- 1 file changed, 37 insertions(+), 46 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index 4f2e0cc8a51a..442857f3bde7 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -138,19 +138,33 @@ void program_sh_mem_settings(struct device_queue_manager *dqm, } static void increment_queue_count(struct device_queue_manager *dqm, - enum kfd_queue_type type) + struct qcm_process_device *qpd, + struct queue *q) { dqm->active_queue_count++; - if (type == KFD_QUEUE_TYPE_COMPUTE || type == KFD_QUEUE_TYPE_DIQ) + if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE || + q->properties.type == KFD_QUEUE_TYPE_DIQ) dqm->active_cp_queue_count++; + + if (q->properties.is_gws) { + dqm->gws_queue_count++; + qpd->mapped_gws_queue = true; + } } static void decrement_queue_count(struct device_queue_manager *dqm, - enum kfd_queue_type type) + struct qcm_process_device *qpd, + struct queue *q) { dqm->active_queue_count--; - if (type == KFD_QUEUE_TYPE_COMPUTE || type == KFD_QUEUE_TYPE_DIQ) + if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE || + q->properties.type == KFD_QUEUE_TYPE_DIQ) dqm->active_cp_queue_count--; + + if (q->properties.is_gws) { + dqm->gws_queue_count--; + qpd->mapped_gws_queue = false; + } } static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q) @@ -390,7 +404,7 @@ static int create_queue_nocpsch(struct device_queue_manager *dqm, list_add(&q->list, &qpd->queues_list); qpd->queue_count++; if (q->properties.is_active) - increment_queue_count(dqm, q->properties.type); + increment_queue_count(dqm, qpd, q); /* * Unconditionally increment this counter, regardless of the queue's @@ -515,13 +529,8 @@ static int destroy_queue_nocpsch_locked(struct device_queue_manager *dqm, deallocate_vmid(dqm, qpd, q); } qpd->queue_count--; - if (q->properties.is_active) { - decrement_queue_count(dqm, q->properties.type); - if (q->properties.is_gws) { - dqm->gws_queue_count--; - qpd->mapped_gws_queue = false; - } - } + if (q->properties.is_active) + decrement_queue_count(dqm, qpd, q); return retval; } @@ -613,12 +622,11 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q) * dqm->active_queue_count to determine whether a new runlist must be * uploaded. */ - if (q->properties.is_active && !prev_active) - increment_queue_count(dqm, q->properties.type); - else if (!q->properties.is_active && prev_active) - decrement_queue_count(dqm, q->properties.type); - - if (q->gws && !q->properties.is_gws) { + if (q->properties.is_active && !prev_active) { + increment_queue_count(dqm, &pdd->qpd, q); + } else if (!q->properties.is_active && prev_active) { + decrement_queue_count(dqm, &pdd->qpd, q); + } else if (q->gws && !q->properties.is_gws) { if (q->properties.is_active) { dqm->gws_queue_count++; pdd->qpd.mapped_gws_queue = true; @@ -680,11 +688,7 @@ static int evict_process_queues_nocpsch(struct device_queue_manager *dqm, mqd_mgr = dqm->mqd_mgrs[get_mqd_type_from_queue_type( q->properties.type)]; q->properties.is_active = false; - decrement_queue_count(dqm, q->properties.type); - if (q->properties.is_gws) { - dqm->gws_queue_count--; - qpd->mapped_gws_queue = false; - } + decrement_queue_count(dqm, qpd, q); if (WARN_ONCE(!dqm->sched_running, "Evict when stopped\n")) continue; @@ -730,7 +734,7 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, continue; q->properties.is_active = false; - decrement_queue_count(dqm, q->properties.type); + decrement_queue_count(dqm, qpd, q); } pdd->last_evict_timestamp = get_jiffies_64(); retval = execute_queues_cpsch(dqm, @@ -801,11 +805,7 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm, mqd_mgr = dqm->mqd_mgrs[get_mqd_type_from_queue_type( q->properties.type)]; q->properties.is_active = true; - increment_queue_count(dqm, q->properties.type); - if (q->properties.is_gws) { - dqm->gws_queue_count++; - qpd->mapped_gws_queue = true; - } + increment_queue_count(dqm, qpd, q); if (WARN_ONCE(!dqm->sched_running, "Restore when stopped\n")) continue; @@ -863,7 +863,7 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm, continue; q->properties.is_active = true; - increment_queue_count(dqm, q->properties.type); + increment_queue_count(dqm, &pdd->qpd, q); } retval = execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); @@ -1265,7 +1265,7 @@ static int create_kernel_queue_cpsch(struct device_queue_manager *dqm, dqm->total_queue_count); list_add(&kq->list, &qpd->priv_queue_list); - increment_queue_count(dqm, kq->queue->properties.type); + increment_queue_count(dqm, qpd, kq->queue); qpd->is_debug = true; execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); dqm_unlock(dqm); @@ -1279,7 +1279,7 @@ static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm, { dqm_lock(dqm); list_del(&kq->list); - decrement_queue_count(dqm, kq->queue->properties.type); + decrement_queue_count(dqm, qpd, kq->queue); qpd->is_debug = false; execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0); /* @@ -1346,7 +1346,7 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q, qpd->queue_count++; if (q->properties.is_active) { - increment_queue_count(dqm, q->properties.type); + increment_queue_count(dqm, qpd, q); execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); @@ -1548,15 +1548,11 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm, list_del(&q->list); qpd->queue_count--; if (q->properties.is_active) { - decrement_queue_count(dqm, q->properties.type); + decrement_queue_count(dqm, qpd, q); retval = execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0); if (retval == -ETIME) qpd->reset_wavefronts = true; - if (q->properties.is_gws) { - dqm->gws_queue_count--; - qpd->mapped_gws_queue = false; - } } /* @@ -1747,7 +1743,7 @@ static int process_termination_cpsch(struct device_queue_manager *dqm, /* Clean all kernel queues */ list_for_each_entry_safe(kq, kq_next, &qpd->priv_queue_list, list) { list_del(&kq->list); - decrement_queue_count(dqm, kq->queue->properties.type); + decrement_queue_count(dqm, qpd, kq->queue); qpd->is_debug = false; dqm->total_queue_count--; filter = KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES; @@ -1760,13 +1756,8 @@ static int process_termination_cpsch(struct device_queue_manager *dqm, else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) deallocate_sdma_queue(dqm, q); - if (q->properties.is_active) { - decrement_queue_count(dqm, q->properties.type); - if (q->properties.is_gws) { - dqm->gws_queue_count--; - qpd->mapped_gws_queue = false; - } - } + if (q->properties.is_active) + decrement_queue_count(dqm, qpd, q); dqm->total_queue_count--; }