crypto: sun4i-ss - reduce stack usage

After the latest addition, the stack usage of sun4i_ss_cipher_poll
grew beyond the warning limit when KASAN is enabled:

drivers/crypto/sunxi-ss/sun4i-ss-cipher.c:118:12: error: stack frame size of 1152 bytes in function 'sun4i_ss_cipher_poll' [-Werror,-Wframe-larger-than=]
static int sun4i_ss_cipher_poll(struct skcipher_request *areq)

Reduce it in three ways:

- split out the new code into a separate function so its stack
  usage can overlap that of the sun4i_ss_opti_poll() code path
- mark both special cases as noinline_for_stack, which should
  ideally result in a tail call that frees the rest of the
  stack
- move the buf and obuf variables into the code blocks in
  which they are used.

The three separate functions now use 144, 640 and 304 bytes of kernel
stack, respectively.

Fixes: 0ae1f46c ("crypto: sun4i-ss - fallback when length is not multiple of blocksize")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Corentin LABBE <clabbe.montjoie@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

drivers/crypto/sunxi-ss/sun4i-ss-cipher.c

@@ -16,7 +16,7 @@
  */
 #include "sun4i-ss.h"
 
-static int sun4i_ss_opti_poll(struct skcipher_request *areq)
+static int noinline_for_stack sun4i_ss_opti_poll(struct skcipher_request *areq)
 {
 	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(areq);
 	struct sun4i_tfm_ctx *op = crypto_skcipher_ctx(tfm);
@@ -118,6 +118,29 @@ static int sun4i_ss_opti_poll(struct skcipher_request *areq)
 	return err;
 }
 
+
+static int noinline_for_stack sun4i_ss_cipher_poll_fallback(struct skcipher_request *areq)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(areq);
+	struct sun4i_tfm_ctx *op = crypto_skcipher_ctx(tfm);
+	struct sun4i_cipher_req_ctx *ctx = skcipher_request_ctx(areq);
+	SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, op->fallback_tfm);
+	int err;
+
+	skcipher_request_set_sync_tfm(subreq, op->fallback_tfm);
+	skcipher_request_set_callback(subreq, areq->base.flags, NULL,
+				      NULL);
+	skcipher_request_set_crypt(subreq, areq->src, areq->dst,
+				   areq->cryptlen, areq->iv);
+	if (ctx->mode & SS_DECRYPTION)
+		err = crypto_skcipher_decrypt(subreq);
+	else
+		err = crypto_skcipher_encrypt(subreq);
+	skcipher_request_zero(subreq);
+
+	return err;
+}
+
 /* Generic function that support SG with size not multiple of 4 */
 static int sun4i_ss_cipher_poll(struct skcipher_request *areq)
 {
@@ -144,8 +167,6 @@ static int sun4i_ss_cipher_poll(struct skcipher_request *areq)
 	unsigned int todo;
 	struct sg_mapping_iter mi, mo;
 	unsigned int oi, oo;	/* offset for in and out */
-	char buf[4 * SS_RX_MAX];/* buffer for linearize SG src */
-	char bufo[4 * SS_TX_MAX]; /* buffer for linearize SG dst */
 	unsigned int ob = 0;	/* offset in buf */
 	unsigned int obo = 0;	/* offset in bufo*/
 	unsigned int obl = 0;	/* length of data in bufo */
@@ -182,20 +203,8 @@ static int sun4i_ss_cipher_poll(struct skcipher_request *areq)
 	if (no_chunk == 1 && !need_fallback)
 		return sun4i_ss_opti_poll(areq);
 
-	if (need_fallback) {
-		SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, op->fallback_tfm);
-		skcipher_request_set_sync_tfm(subreq, op->fallback_tfm);
-		skcipher_request_set_callback(subreq, areq->base.flags, NULL,
-					      NULL);
-		skcipher_request_set_crypt(subreq, areq->src, areq->dst,
-					   areq->cryptlen, areq->iv);
-		if (ctx->mode & SS_DECRYPTION)
-			err = crypto_skcipher_decrypt(subreq);
-		else
-			err = crypto_skcipher_encrypt(subreq);
-		skcipher_request_zero(subreq);
-		return err;
-	}
+	if (need_fallback)
+		return sun4i_ss_cipher_poll_fallback(areq);
 
 	spin_lock_irqsave(&ss->slock, flags);
 
@@ -228,6 +237,8 @@ static int sun4i_ss_cipher_poll(struct skcipher_request *areq)
 
 	while (oleft) {
 		if (ileft) {
+			char buf[4 * SS_RX_MAX];/* buffer for linearize SG src */
+
 			/*
 			 * todo is the number of consecutive 4byte word that we
 			 * can read from current SG
@@ -285,6 +296,8 @@ static int sun4i_ss_cipher_poll(struct skcipher_request *areq)
 				oo = 0;
 			}
 		} else {
+			char bufo[4 * SS_TX_MAX]; /* buffer for linearize SG dst */
+
 			/*
 			 * read obl bytes in bufo, we read at maximum for
 			 * emptying the device