stating: ccree: revert "staging: ccree: fix leak of import() after init()"
authorGilad Ben-Yossef <gilad@benyossef.com>
Sun, 7 Jan 2018 12:14:22 +0000 (12:14 +0000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Mon, 8 Jan 2018 15:27:23 +0000 (16:27 +0100)
This reverts commit c5f39d07860c ("staging: ccree: fix leak of import()
after init()") and commit aece09024414 ("staging: ccree: Uninitialized
return in ssi_ahash_import()").

This is the wrong solution and ends up relying on uninitialized memory,
although it was not obvious to me at the time.

Cc: stable@vger.kernel.org
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/ccree/ssi_hash.c

index ff05ac8..ee7370c 100644 (file)
@@ -1673,7 +1673,7 @@ static int cc_hash_import(struct ahash_request *req, const void *in)
        struct device *dev = drvdata_to_dev(ctx->drvdata);
        struct ahash_req_ctx *state = ahash_request_ctx(req);
        u32 tmp;
-       int rc = 0;
+       int rc;
 
        memcpy(&tmp, in, sizeof(u32));
        if (tmp != CC_EXPORT_MAGIC) {
@@ -1682,12 +1682,9 @@ static int cc_hash_import(struct ahash_request *req, const void *in)
        }
        in += sizeof(u32);
 
-       /* call init() to allocate bufs if the user hasn't */
-       if (!state->digest_buff) {
-               rc = cc_hash_init(req);
-               if (rc)
-                       goto out;
-       }
+       rc = cc_hash_init(req);
+       if (rc)
+               goto out;
 
        dma_sync_single_for_cpu(dev, state->digest_buff_dma_addr,
                                ctx->inter_digestsize, DMA_BIDIRECTIONAL);