Re: [PATCH] staging: android: ion: Skip sync if not mapped

From: Dan Carpenter <dan.carpenter@oracle.com>
To: "Ørjan Eide" <orjan.eide@arm.com>
Cc: devel@driverdev.osuosl.org, "Laura Abbott" <labbott@redhat.com>,
	"Todd Kjos" <tkjos@android.com>,
	"Lecopzer Chen" <lecopzer.chen@mediatek.com>,
	"Arnd Bergmann" <arnd@arndb.de>,
	"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
	linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org,
	linaro-mm-sig@lists.linaro.org,
	"Arve Hjønnevåg" <arve@android.com>,
	anders.pedersen@arm.com, "Daniel Vetter" <daniel.vetter@ffwll.ch>,
	"Joel Fernandes" <joel@joelfernandes.org>,
	"Darren Hart (VMware)" <dvhart@infradead.org>,
	nd@arm.com, "Martijn Coenen" <maco@android.com>,
	"Christian Brauner" <christian@brauner.io>,
	linux-media@vger.kernel.org
Subject: Re: [PATCH] staging: android: ion: Skip sync if not mapped
Date: Thu, 16 Apr 2020 12:49:56 +0300	[thread overview]
Message-ID: <20200416094955.GM1163@kadam> (raw)
In-Reply-To: <20200414141849.55654-1-orjan.eide@arm.com>

On Tue, Apr 14, 2020 at 04:18:47PM +0200, Ørjan Eide wrote:
> @@ -238,6 +242,10 @@ static void ion_unmap_dma_buf(struct dma_buf_attachment *attachment,
>  			      struct sg_table *table,
>  			      enum dma_data_direction direction)
>  {
> +	struct ion_dma_buf_attachment *a = attachment->priv;
> +
> +	a->mapped = false;

Possibly a stupid question but here we're not holding a lock.  Is
concurrency an issue?

> +
>  	dma_unmap_sg(attachment->dev, table->sgl, table->nents, direction);
>  }
>  
> @@ -297,6 +305,8 @@ static int ion_dma_buf_begin_cpu_access(struct dma_buf *dmabuf,
>  
>  	mutex_lock(&buffer->lock);
>  	list_for_each_entry(a, &buffer->attachments, list) {
> +		if (!a->mapped)
> +			continue;
>  		dma_sync_sg_for_cpu(a->dev, a->table->sgl, a->table->nents,
>  				    direction);
>  	}

regards,
dan carpenter

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel