Fix a side effect of adding xcbc support, when the next_buffer is not
copied.
The issue occurs, when there is stored from previous state a blocksize
buffer and received, a less than blocksize, from user. In this case, the
nents for req->src is 0, and the next_buffer is not copied.
An example is:
{
        .tap    = { 17, 15, 8 },
        .psize  = 40,
        .np     = 3,
        .ksize  = 16,
}

Cc: <sta...@vger.kernel.org>
Fixes: 12b8567f6fa4 ("crypto: caam - add support for xcbc(aes)")
Signed-off-by: Iuliana Prodan <iuliana.pro...@nxp.com>
---
 drivers/crypto/caam/caamhash.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/crypto/caam/caamhash.c b/drivers/crypto/caam/caamhash.c
index d7483e4..cd03767 100644
--- a/drivers/crypto/caam/caamhash.c
+++ b/drivers/crypto/caam/caamhash.c
@@ -865,19 +865,18 @@ static int ahash_update_ctx(struct ahash_request *req)
                if (ret)
                        goto unmap_ctx;
 
-               if (mapped_nents) {
+               if (mapped_nents)
                        sg_to_sec4_sg_last(req->src, mapped_nents,
                                           edesc->sec4_sg + sec4_sg_src_index,
                                           0);
-                       if (*next_buflen)
-                               scatterwalk_map_and_copy(next_buf, req->src,
-                                                        to_hash - *buflen,
-                                                        *next_buflen, 0);
-               } else {
+               else
                        sg_to_sec4_set_last(edesc->sec4_sg + sec4_sg_src_index -
                                            1);
-               }
 
+               if (*next_buflen)
+                       scatterwalk_map_and_copy(next_buf, req->src,
+                                                to_hash - *buflen,
+                                                *next_buflen, 0);
                desc = edesc->hw_desc;
 
                edesc->sec4_sg_dma = dma_map_single(jrdev, edesc->sec4_sg,
-- 
2.1.0

Reply via email to