redonkable
/
remarkable-linux
1
0
Fork 0
Code Releases Activity

MLK-19365: crypto: gcm: Cache aligned auth_data 

Generic GCM is likely to end up using a hardware accelerator to do
part of the job. Allocating hash, iv and result in a contiguous memory
area increases the risk of dma mapping multiple ranges on the same
cacheline. Also having dma and cpu written data on the same cacheline
will cause coherence issues.

Signed-off-by: Franck LENORMAND <franck.lenormand@nxp.com>
pull/10/head
Franck LENORMAND 2018-09-14 10:31:57 +02:00 committed by Jason Liu
parent f389c802aa
commit d08385adac
1 changed files with 12 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
crypto/gcm.c
View File

@ -66,7 +66,18 @@ struct crypto_gcm_ghash_ctx {
struct crypto_gcm_req_priv_ctx {
u8 iv[16];
u8 auth_tag[16];
/*
* We need to force auth_tag to be on its own cacheline.
*
* We put it on its cacheline with the macro ____cacheline_aligned.
* The next fields must be on another cacheline so we add a dummy field
* which is located on another cacheline to enforce that.
*/
u8 auth_tag[16] ____cacheline_aligned;
u8 dummy_align_auth_tag ____cacheline_aligned;
u8 iauth_tag[16];
struct scatterlist src[3];
struct scatterlist dst[3];