From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S932561AbcLMJOo convert rfc822-to-8bit (ORCPT <rfc822;w@1wt.eu>);
        Tue, 13 Dec 2016 04:14:44 -0500
Received: from mx1.redhat.com ([209.132.183.28]:45916 "EHLO mx1.redhat.com"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1752456AbcLMJOk (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Tue, 13 Dec 2016 04:14:40 -0500
Message-ID: <1481611645.27088.56.camel@redhat.com>
Subject: Re: [RFC 1/5] drm/virtio: add virtio_gpu_alloc_fence()
From: Gerd Hoffmann <kraxel@redhat.com>
To: Gustavo Padovan <gustavo@padovan.org>
Cc: dri-devel@lists.freedesktop.org, robdclark@gmail.com,
        Gustavo Padovan <gustavo.padovan@collabora.co.uk>,
        David Airlie <airlied@linux.ie>,
        "open list:VIRTIO GPU DRIVER" 
        <virtualization@lists.linux-foundation.org>,
        open list <linux-kernel@vger.kernel.org>
Date: Tue, 13 Dec 2016 07:47:25 +0100
In-Reply-To: <1481575710-12535-1-git-send-email-gustavo@padovan.org>
References: <1481575710-12535-1-git-send-email-gustavo@padovan.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Mime-Version: 1.0
X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.29]); Tue, 13 Dec 2016 09:14:40 +0000 (UTC)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

  Hi,

> +struct virtio_gpu_fence *virtio_gpu_fence_alloc(struct virtio_gpu_device *vgdev)
> +{
> +	struct virtio_gpu_fence_driver *drv = &vgdev->fence_drv;
> +	struct virtio_gpu_fence *fence;
> +	unsigned long irq_flags;
> +
> +	fence = kmalloc(sizeof(struct virtio_gpu_fence), GFP_ATOMIC);
> +	if (!fence)
> +		return NULL;
> +

> +	spin_lock_irqsave(&drv->lock, irq_flags);
> +	fence->drv = drv;
> +	fence->seq = ++drv->sync_seq;
> +	dma_fence_init(&fence->f, &virtio_fence_ops, &drv->lock,
> +		       drv->context, fence->seq);
> +	spin_unlock_irqrestore(&drv->lock, irq_flags);

seq assignment ...

> +
> +	return fence;
> +}
> +
>  int virtio_gpu_fence_emit(struct virtio_gpu_device *vgdev,
>  			  struct virtio_gpu_ctrl_hdr *cmd_hdr,
> -			  struct virtio_gpu_fence **fence)
> +			  struct virtio_gpu_fence *fence)
>  {
>  	struct virtio_gpu_fence_driver *drv = &vgdev->fence_drv;
>  	unsigned long irq_flags;
>  
> -	*fence = kmalloc(sizeof(struct virtio_gpu_fence), GFP_ATOMIC);
> -	if ((*fence) == NULL)
> -		return -ENOMEM;
> -
>  	spin_lock_irqsave(&drv->lock, irq_flags);
> -	(*fence)->drv = drv;
> -	(*fence)->seq = ++drv->sync_seq;
> -	dma_fence_init(&(*fence)->f, &virtio_fence_ops, &drv->lock,
> -		       drv->context, (*fence)->seq);

... must stay here.  Otherwise requests can be submitted to the virt
queue with fence sequence numbers out of order.

cheers,
  Gerd