Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6

Pull crypto updates from Herbert Xu:
 "API:
   - Remove VLA usage
   - Add cryptostat user-space interface
   - Add notifier for new crypto algorithms

  Algorithms:
   - Add OFB mode
   - Remove speck

  Drivers:
   - Remove x86/sha*-mb as they are buggy
   - Remove pcbc(aes) from x86/aesni
   - Improve performance of arm/ghash-ce by up to 85%
   - Implement CTS-CBC in arm64/aes-blk, faster by up to 50%
   - Remove PMULL based arm64/crc32 driver
   - Use PMULL in arm64/crct10dif
   - Add aes-ctr support in s5p-sss
   - Add caam/qi2 driver

  Others:
   - Pick better transform if one becomes available in crc-t10dif"

* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (124 commits)
  crypto: chelsio - Update ntx queue received from cxgb4
  crypto: ccree - avoid implicit enum conversion
  crypto: caam - add SPDX license identifier to all files
  crypto: caam/qi - simplify CGR allocation, freeing
  crypto: mxs-dcp - make symbols 'sha1_null_hash' and 'sha256_null_hash' static
  crypto: arm64/aes-blk - ensure XTS mask is always loaded
  crypto: testmgr - fix sizeof() on COMP_BUF_SIZE
  crypto: chtls - remove set but not used variable 'csk'
  crypto: axis - fix platform_no_drv_owner.cocci warnings
  crypto: x86/aes-ni - fix build error following fpu template removal
  crypto: arm64/aes - fix handling sub-block CTS-CBC inputs
  crypto: caam/qi2 - avoid double export
  crypto: mxs-dcp - Fix AES issues
  crypto: mxs-dcp - Fix SHA null hashes and output length
  crypto: mxs-dcp - Implement sha import/export
  crypto: aegis/generic - fix for big endian systems
  crypto: morus/generic - fix for big endian systems
  crypto: lrw - fix rebase error after out of bounds fix
  crypto: cavium/nitrox - use pci_alloc_irq_vectors() while enabling MSI-X.
  crypto: cavium/nitrox - NITROX command queue changes.
  ...

lib/chacha20.c

@@ -16,9 +16,9 @@
 #include <asm/unaligned.h>
 #include <crypto/chacha20.h>
 
-void chacha20_block(u32 *state, u32 *stream)
+void chacha20_block(u32 *state, u8 *stream)
 {
-	u32 x[16], *out = stream;
+	u32 x[16];
 	int i;
 
 	for (i = 0; i < ARRAY_SIZE(x); i++)
@@ -67,7 +67,7 @@ void chacha20_block(u32 *state, u32 *stream)
 	}
 
 	for (i = 0; i < ARRAY_SIZE(x); i++)
-		out[i] = cpu_to_le32(x[i] + state[i]);
+		put_unaligned_le32(x[i] + state[i], &stream[i * sizeof(u32)]);
 
 	state[12]++;
 }

lib/crc-t10dif.c

@@ -14,10 +14,47 @@
 #include <linux/err.h>
 #include <linux/init.h>
 #include <crypto/hash.h>
+#include <crypto/algapi.h>
 #include <linux/static_key.h>
+#include <linux/notifier.h>
 
-static struct crypto_shash *crct10dif_tfm;
+static struct crypto_shash __rcu *crct10dif_tfm;
 static struct static_key crct10dif_fallback __read_mostly;
+static DEFINE_MUTEX(crc_t10dif_mutex);
+
+static int crc_t10dif_rehash(struct notifier_block *self, unsigned long val, void *data)
+{
+	struct crypto_alg *alg = data;
+	struct crypto_shash *new, *old;
+
+	if (val != CRYPTO_MSG_ALG_LOADED ||
+	    static_key_false(&crct10dif_fallback) ||
+	    strncmp(alg->cra_name, CRC_T10DIF_STRING, strlen(CRC_T10DIF_STRING)))
+		return 0;
+
+	mutex_lock(&crc_t10dif_mutex);
+	old = rcu_dereference_protected(crct10dif_tfm,
+					lockdep_is_held(&crc_t10dif_mutex));
+	if (!old) {
+		mutex_unlock(&crc_t10dif_mutex);
+		return 0;
+	}
+	new = crypto_alloc_shash("crct10dif", 0, 0);
+	if (IS_ERR(new)) {
+		mutex_unlock(&crc_t10dif_mutex);
+		return 0;
+	}
+	rcu_assign_pointer(crct10dif_tfm, new);
+	mutex_unlock(&crc_t10dif_mutex);
+
+	synchronize_rcu();
+	crypto_free_shash(old);
+	return 0;
+}
+
+static struct notifier_block crc_t10dif_nb = {
+	.notifier_call = crc_t10dif_rehash,
+};
 
 __u16 crc_t10dif_update(__u16 crc, const unsigned char *buffer, size_t len)
 {
@@ -30,11 +67,14 @@ __u16 crc_t10dif_update(__u16 crc, const unsigned char *buffer, size_t len)
 	if (static_key_false(&crct10dif_fallback))
 		return crc_t10dif_generic(crc, buffer, len);
 
-	desc.shash.tfm = crct10dif_tfm;
+	rcu_read_lock();
+	desc.shash.tfm = rcu_dereference(crct10dif_tfm);
 	desc.shash.flags = 0;
 	*(__u16 *)desc.ctx = crc;
 
 	err = crypto_shash_update(&desc.shash, buffer, len);
+	rcu_read_unlock();
+
 	BUG_ON(err);
 
 	return *(__u16 *)desc.ctx;
@@ -49,6 +89,7 @@ EXPORT_SYMBOL(crc_t10dif);
 
 static int __init crc_t10dif_mod_init(void)
 {
+	crypto_register_notifier(&crc_t10dif_nb);
 	crct10dif_tfm = crypto_alloc_shash("crct10dif", 0, 0);
 	if (IS_ERR(crct10dif_tfm)) {
 		static_key_slow_inc(&crct10dif_fallback);
@@ -59,12 +100,24 @@ static int __init crc_t10dif_mod_init(void)
 
 static void __exit crc_t10dif_mod_fini(void)
 {
+	crypto_unregister_notifier(&crc_t10dif_nb);
 	crypto_free_shash(crct10dif_tfm);
 }
 
 module_init(crc_t10dif_mod_init);
 module_exit(crc_t10dif_mod_fini);
 
+static int crc_t10dif_transform_show(char *buffer, const struct kernel_param *kp)
+{
+	if (static_key_false(&crct10dif_fallback))
+		return sprintf(buffer, "fallback\n");
+
+	return sprintf(buffer, "%s\n",
+		crypto_tfm_alg_driver_name(crypto_shash_tfm(crct10dif_tfm)));
+}
+
+module_param_call(transform, NULL, crc_t10dif_transform_show, NULL, 0644);
+
 MODULE_DESCRIPTION("T10 DIF CRC calculation");
 MODULE_LICENSE("GPL");
 MODULE_SOFTDEP("pre: crct10dif");