Herbert Xu wrote: > On Wed, Nov 19, 2008 at 02:03:27PM +0000, Evgeniy Polyakov wrote: >> From: Patrick McHardy >> >> The descriptors need to be invalidated after processing for ring >> cleanup to work properly and to avoid using an old destination >> descriptor when the src and cmd descriptors are already set up >> and the dst descriptor isn't. >> >> Signed-off-by: Patrick McHardy >> Signed-off-by: Evgeniy Polyakov >> --- >> drivers/crypto/hifn_795x.c | 8 ++++---- >> 1 files changed, 4 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/crypto/hifn_795x.c b/drivers/crypto/hifn_795x.c >> index 320d08d..4d22b21 100644 >> --- a/drivers/crypto/hifn_795x.c >> +++ b/drivers/crypto/hifn_795x.c >> @@ -1296,7 +1296,7 @@ static int hifn_setup_src_desc(struct hifn_device *dev, struct page *page, >> >> dma->srcr[idx].p = __cpu_to_le32(addr); >> dma->srcr[idx].l = __cpu_to_le32(size | HIFN_D_VALID | >> - HIFN_D_MASKDONEIRQ | HIFN_D_NOINVALID | HIFN_D_LAST); >> + HIFN_D_MASKDONEIRQ | HIFN_D_LAST); > > Something is not right here. This patch is an exact reversal > of the previous patch in this series. Could you please take a > look? It seems the previous patch got mixed up, this is what I have in my old tree (just for reference, I don't know whether Evgeniy made any changes):