[PATCH] ASYNC_TX: async_xor mapping fix

[Yuri Tikhonov]

 The destination address may be present in the source list, so we should
map the addresses from the source list first.
 Otherwise, if page corresponding to destination is not marked as write-
through (with regards to CPU cache), then mapping it with DMA_FROM_DEVICE
may lead to data loss, and finally to an incorrect result of calculations.

Signed-off-by: Yuri Tikhonov <yur at emcraft.com>
---
 crypto/async_tx/async_xor.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/crypto/async_tx/async_xor.c b/crypto/async_tx/async_xor.c
index 91dbf07..291f517 100644
--- a/crypto/async_tx/async_xor.c
+++ b/crypto/async_tx/async_xor.c
@@ -53,10 +53,10 @@ do_async_xor(struct dma_chan *chan, struct page *dest, struct page **src_list,
 	int xor_src_cnt;
 	dma_addr_t dma_dest;
 
-	dma_dest = dma_map_page(dma->dev, dest, offset, len, DMA_FROM_DEVICE);
 	for (i = 0; i < src_cnt; i++)
 		dma_src[i] = dma_map_page(dma->dev, src_list[i], offset,
 					  len, DMA_TO_DEVICE);
+	dma_dest = dma_map_page(dma->dev, dest, offset, len, DMA_FROM_DEVICE);
 
 	while (src_cnt) {
 		async_flags = flags;
-- 
1.5.6.1

-- 
Yuri Tikhonov, Senior Software Engineer
Emcraft Systems, www.emcraft.com