dmaengine: dw: don't perform DMA when dmaengine_submit is called

Accordingly to discussion [1] and followed up documentation the DMA controller
driver shouldn't start any DMA operations when dmaengine_submit() is called.

This patch fixes the workflow in dw_dmac driver to follow the documentation.

[1] http://www.spinics.net/lists/arm-kernel/msg125987.htmlSigned-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>

drivers/dma/TODO

@@ -7,7 +7,6 @@ TODO for slave dma
 	- imx-dma
 	- imx-sdma
 	- mxs-dma.c
-	- dw_dmac
 	- intel_mid_dma
 4. Check other subsystems for dma drivers and merge/move to dmaengine
 5. Remove dma_slave_config's dma direction.

drivers/dma/dw/core.c

@@ -684,17 +684,9 @@ static dma_cookie_t dwc_tx_submit(struct dma_async_tx_descriptor *tx)
 	 * possible, perhaps even appending to those already submitted
 	 * for DMA. But this is hard to do in a race-free manner.
 	 */
-	if (list_empty(&dwc->active_list)) {
-		dev_vdbg(chan2dev(tx->chan), "%s: started %u\n", __func__,
-				desc->txd.cookie);
-		list_add_tail(&desc->desc_node, &dwc->active_list);
-		dwc_dostart(dwc, dwc_first_active(dwc));
-	} else {
-		dev_vdbg(chan2dev(tx->chan), "%s: queued %u\n", __func__,
-				desc->txd.cookie);
 
+	dev_vdbg(chan2dev(tx->chan), "%s: queued %u\n", __func__, desc->txd.cookie);
 	list_add_tail(&desc->desc_node, &dwc->queue);
-	}
 
 	spin_unlock_irqrestore(&dwc->lock, flags);
 
@@ -1099,9 +1091,12 @@ dwc_tx_status(struct dma_chan *chan,
 static void dwc_issue_pending(struct dma_chan *chan)
 {
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
+	unsigned long		flags;
 
-	if (!list_empty(&dwc->queue))
-		dwc_scan_descriptors(to_dw_dma(chan->device), dwc);
+	spin_lock_irqsave(&dwc->lock, flags);
+	if (list_empty(&dwc->active_list))
+		dwc_dostart_first_queued(dwc);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 }
 
 static int dwc_alloc_chan_resources(struct dma_chan *chan)