dmaengine: dw-edma: Fix scatter-gather address calculation

Fix the source and destination physical address calculation of a
peripheral device on scatter-gather implementation.

This issue manifested during tests using a 64 bits architecture system.
The abnormal behavior wasn't visible before due to all previous tests
were done using 32 bits architecture system, that masked his effect.

Fixes: e63d79d ("dmaengine: Add Synopsys eDMA IP core driver")
Cc: stable@vger.kernel.org
Signed-off-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Link: https://lore.kernel.org/r/8d3ab7e2ba96563fe3495b32f60077fffb85307d.1597327623.git.gustavo.pimentel@synopsys.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>

drivers/dma/dw-edma/dw-edma-core.c

@@ -405,35 +405,36 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 			if (xfer->cyclic) {
 				burst->dar = xfer->xfer.cyclic.paddr;
 			} else {
-				burst->dar = sg_dma_address(sg);
+				burst->dar = dst_addr;
 				/* Unlike the typical assumption by other
 				 * drivers/IPs the peripheral memory isn't
 				 * a FIFO memory, in this case, it's a
 				 * linear memory and that why the source
 				 * and destination addresses are increased
 				 * by the same portion (data length)
 				 */
-				src_addr += sg_dma_len(sg);
 			}
 		} else {
 			burst->dar = dst_addr;
 			if (xfer->cyclic) {
 				burst->sar = xfer->xfer.cyclic.paddr;
 			} else {
-				burst->sar = sg_dma_address(sg);
+				burst->sar = src_addr;
 				/* Unlike the typical assumption by other
 				 * drivers/IPs the peripheral memory isn't
 				 * a FIFO memory, in this case, it's a
 				 * linear memory and that why the source
 				 * and destination addresses are increased
 				 * by the same portion (data length)
 				 */
-				dst_addr += sg_dma_len(sg);
 			}
 		}
 
-		if (!xfer->cyclic)
+		if (!xfer->cyclic) {
+			src_addr += sg_dma_len(sg);
+			dst_addr += sg_dma_len(sg);
 			sg = sg_next(sg);
+		}
 	}
 
 	return vchan_tx_prep(&chan->vc, &desc->vd, xfer->flags);