From mboxrd@z Thu Jan 1 00:00:00 1970 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 3FE55C52D71 for ; Thu, 8 Aug 2024 17:41:54 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 0DF0410E7B8; Thu, 8 Aug 2024 17:41:54 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=intel.com header.i=@intel.com header.b="KZC4peho"; dkim-atps=neutral Received: from mgamail.intel.com (mgamail.intel.com [198.175.65.21]) by gabe.freedesktop.org (Postfix) with ESMTPS id EE53710E7BF for ; Thu, 8 Aug 2024 17:41:49 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1723138910; x=1754674910; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=luHKOv1hPc2ayC6x6RNS8lunkizmXySbVRCyTw6lSzs=; b=KZC4peho9PEWKHaY+0Vd6kwiuOr+c1yyiT7FHsHX5rRLAxdOiVhaTIwu mZS90x27ww7WV4PISn83do3I65vCqIuPXCOKHRZ7HR5/7Q3knDe2UPUi7 MZwRHgPEs5K4bx4Nyy4JjR8D+BX4mmup/wDorwXn89bQmpHBlTq/OPePT iK0oBB2UlhvGIt7DcwpRu2s/2zRdP4ztK5PJoPg5iaj4ujylWYFkrx2Ib /cdmTeWbz6YyCeskjV/VKDNMLvTKQZ/4exdRaoisS72LBQlgHSiOH1SRu cAmKvk6IFeme7zZZ79SYcPZUUCO22qvRKsoM9wjL2cVRNrrQEqaNjpiP+ Q==; X-CSE-ConnectionGUID: q0pUZgcxRWOyzabvfkcQaw== X-CSE-MsgGUID: g9Lo/34dQqeFAyDqgEQh8Q== X-IronPort-AV: E=McAfee;i="6700,10204,11158"; a="21256063" X-IronPort-AV: E=Sophos;i="6.09,273,1716274800"; d="scan'208";a="21256063" Received: from orviesa007.jf.intel.com ([10.64.159.147]) by orvoesa113.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 08 Aug 2024 10:41:49 -0700 X-CSE-ConnectionGUID: Qb5j8Sg0Rwu+1igrT/L13g== X-CSE-MsgGUID: NpSMMRtKR/mr5jdKyx/7+g== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.09,273,1716274800"; d="scan'208";a="57850454" Received: from orsosgc001.jf.intel.com ([10.165.21.138]) by orviesa007-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 08 Aug 2024 10:41:49 -0700 From: Ashutosh Dixit To: intel-xe@lists.freedesktop.org Cc: Umesh Nerlige Ramappa , Jose Souza , Lionel Landwerlin Subject: [PATCH 5/8] drm/xe/oa: Signal output fences Date: Thu, 8 Aug 2024 10:41:36 -0700 Message-ID: <20240808174139.4027534-6-ashutosh.dixit@intel.com> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20240808174139.4027534-1-ashutosh.dixit@intel.com> References: <20240808174139.4027534-1-ashutosh.dixit@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 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: , Errors-To: intel-xe-bounces@lists.freedesktop.org Sender: "Intel-xe" Complete 'struct xe_oa_fence' to include the dma_fence used to signal output fences in the xe_sync array. The fences are signalled asynchronously. When there are no output fences to signal, the OA configuration wait is synchronously re-introduced into the ioctl. Signed-off-by: Ashutosh Dixit --- drivers/gpu/drm/xe/xe_oa.c | 46 +++++++++++++++++++++++++++++++++++--- 1 file changed, 43 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index 416e031ac454b..bc421cd0af6ba 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -96,6 +96,10 @@ struct xe_oa_config_bo { }; struct xe_oa_fence { + /* @base: dma fence base */ + struct dma_fence base; + /* @lock: lock for the fence */ + spinlock_t lock; /* @xe: pointer to xe device */ struct xe_device *xe; /* @work: work to signal that OA configuration is applied */ @@ -953,9 +957,26 @@ static void xe_oa_fence_work_fn(struct work_struct *w) /* Additional empirical delay needed for NOA programming after registers are written */ usleep_range(us, 2 * us); - kfree(ofence); + /* Now signal fence to indicate new OA configuration is active */ + dma_fence_signal(&ofence->base); + dma_fence_put(&ofence->base); } +static const char *xe_oa_get_driver_name(struct dma_fence *fence) +{ + return "xe_oa"; +} + +static const char *xe_oa_get_timeline_name(struct dma_fence *fence) +{ + return "unbound"; +} + +static const struct dma_fence_ops xe_oa_fence_ops = { + .get_driver_name = xe_oa_get_driver_name, + .get_timeline_name = xe_oa_get_timeline_name, +}; + static struct xe_oa_fence *xe_oa_fence_init(struct xe_device *xe, struct dma_fence *config_fence) { struct xe_oa_fence *ofence; @@ -967,6 +988,8 @@ static struct xe_oa_fence *xe_oa_fence_init(struct xe_device *xe, struct dma_fen ofence->xe = xe; INIT_WORK(&ofence->work, xe_oa_fence_work_fn); ofence->config_fence = config_fence; + spin_lock_init(&ofence->lock); + dma_fence_init(&ofence->base, &xe_oa_fence_ops, &ofence->lock, 0, 0); return ofence; } @@ -975,8 +998,8 @@ static int xe_oa_emit_oa_config(struct xe_oa_stream *stream, struct xe_oa_config { struct xe_oa_config_bo *oa_bo; struct xe_oa_fence *ofence; + int i, err, num_signal = 0; struct dma_fence *fence; - int err; /* Emit OA configuration batch */ oa_bo = xe_oa_alloc_config_buffer(stream, config); @@ -989,13 +1012,30 @@ static int xe_oa_emit_oa_config(struct xe_oa_stream *stream, struct xe_oa_config if (err) goto exit; + /* Initialize and set fence to signal */ ofence = xe_oa_fence_init(stream->oa->xe, fence); if (IS_ERR(ofence)) { err = PTR_ERR(ofence); goto put_fence; } - xe_oa_fence_work_fn(&ofence->work); + for (i = 0; i < stream->num_syncs; i++) + xe_sync_entry_signal(&stream->syncs[i], &ofence->base); + + /* Schedule work to signal the fence */ + queue_work(system_unbound_wq, &ofence->work); + + /* If nothing needs to be signaled we wait synchronously */ + for (i = 0; i < stream->num_syncs; i++) + if (stream->syncs[i].flags & DRM_XE_SYNC_FLAG_SIGNAL) + num_signal++; + if (!num_signal) + flush_work(&ofence->work); + + /* Done with syncs */ + for (i = 0; i < stream->num_syncs; i++) + xe_sync_entry_cleanup(&stream->syncs[i]); + kfree(stream->syncs); return 0; put_fence: -- 2.41.0