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

Pull crypto updates from Herbert Xu:
 "API:
   - Add speed testing on 1420-byte blocks for networking

  Algorithms:
   - Improve performance of chacha on ARM for network packets
   - Improve performance of aegis128 on ARM for network packets

  Drivers:
   - Add support for Keem Bay OCS AES/SM4
   - Add support for QAT 4xxx devices
   - Enable crypto-engine retry mechanism in caam
   - Enable support for crypto engine on sdm845 in qce
   - Add HiSilicon PRNG driver support"

* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (161 commits)
  crypto: qat - add capability detection logic in qat_4xxx
  crypto: qat - add AES-XTS support for QAT GEN4 devices
  crypto: qat - add AES-CTR support for QAT GEN4 devices
  crypto: atmel-i2c - select CONFIG_BITREVERSE
  crypto: hisilicon/trng - replace atomic_add_return()
  crypto: keembay - Add support for Keem Bay OCS AES/SM4
  dt-bindings: Add Keem Bay OCS AES bindings
  crypto: aegis128 - avoid spurious references crypto_aegis128_update_simd
  crypto: seed - remove trailing semicolon in macro definition
  crypto: x86/poly1305 - Use TEST %reg,%reg instead of CMP $0,%reg
  crypto: x86/sha512 - Use TEST %reg,%reg instead of CMP $0,%reg
  crypto: aesni - Use TEST %reg,%reg instead of CMP $0,%reg
  crypto: cpt - Fix sparse warnings in cptpf
  hwrng: ks-sa - Add dependency on IOMEM and OF
  crypto: lib/blake2s - Move selftest prototype into header file
  crypto: arm/aes-ce - work around Cortex-A57/A72 silion errata
  crypto: ecdh - avoid unaligned accesses in ecdh_set_secret()
  crypto: ccree - rework cache parameters handling
  crypto: cavium - Use dma_set_mask_and_coherent to simplify code
  crypto: marvell/octeontx - Use dma_set_mask_and_coherent to simplify code
  ...

144 files changed, 7416 insertions, 1117 deletions

diff --git a/drivers/crypto/Kconfig b/drivers/crypto/Kconfig
index 37da0c070a88..bbd51703e738 100644
--- a/drivers/crypto/Kconfig
+++ b/drivers/crypto/Kconfig
@@ -548,6 +548,7 @@ config CRYPTO_DEV_ATMEL_SHA
 
 config CRYPTO_DEV_ATMEL_I2C
 	tristate
+	select BITREVERSE
 
 config CRYPTO_DEV_ATMEL_ECC
 	tristate "Support for Microchip / Atmel ECC hw accelerator"
@@ -648,7 +649,7 @@ choice
 	default CRYPTO_DEV_QCE_ENABLE_ALL
 	depends on CRYPTO_DEV_QCE
 	help
-	  This option allows to choose whether to build support for all algorihtms
+	  This option allows to choose whether to build support for all algorithms
 	  (default), hashes-only, or skciphers-only.
 
 	  The QCE engine does not appear to scale as well as the CPU to handle
@@ -900,4 +901,6 @@ config CRYPTO_DEV_SA2UL
 	  used for crypto offload.  Select this if you want to use hardware
 	  acceleration for cryptographic algorithms on these devices.
 
+source "drivers/crypto/keembay/Kconfig"
+
 endif # CRYPTO_HW
diff --git a/drivers/crypto/Makefile b/drivers/crypto/Makefile
index 53fc115cf459..fff9a70348e1 100644
--- a/drivers/crypto/Makefile
+++ b/drivers/crypto/Makefile
@@ -51,3 +51,4 @@ obj-$(CONFIG_CRYPTO_DEV_ARTPEC6) += axis/
 obj-$(CONFIG_CRYPTO_DEV_ZYNQMP_AES) += xilinx/
 obj-y += hisilicon/
 obj-$(CONFIG_CRYPTO_DEV_AMLOGIC_GXL) += amlogic/
+obj-y += keembay/
diff --git a/drivers/crypto/allwinner/Kconfig b/drivers/crypto/allwinner/Kconfig
index 0cdfe0e8cc66..180c8a9db819 100644
--- a/drivers/crypto/allwinner/Kconfig
+++ b/drivers/crypto/allwinner/Kconfig
@@ -43,7 +43,7 @@ config CRYPTO_DEV_SUN8I_CE
 	depends on CRYPTO_DEV_ALLWINNER
 	depends on PM
 	help
-	  Select y here to have support for the crypto Engine availlable on
+	  Select y here to have support for the crypto Engine available on
 	  Allwinner SoC H2+, H3, H5, H6, R40 and A64.
 	  The Crypto Engine handle AES/3DES ciphers in ECB/CBC mode.
 
diff --git a/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h b/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h
index 163962f9e284..5c291e4a6857 100644
--- a/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h
+++ b/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h
@@ -25,7 +25,7 @@
 #include <linux/pm_runtime.h>
 #include <crypto/md5.h>
 #include <crypto/skcipher.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
 #include <crypto/hash.h>
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
diff --git a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-hash.c b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-hash.c
index a94bf28f858a..2f09a37306e2 100644
--- a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-hash.c
+++ b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-hash.c
@@ -13,7 +13,8 @@
 #include <linux/pm_runtime.h>
 #include <linux/scatterlist.h>
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/md5.h>
 #include "sun8i-ce.h"
 
@@ -262,13 +263,13 @@ int sun8i_ce_hash_run(struct crypto_engine *engine, void *breq)
 	u32 common;
 	u64 byte_count;
 	__le32 *bf;
-	void *buf;
+	void *buf = NULL;
 	int j, i, todo;
 	int nbw = 0;
 	u64 fill, min_fill;
 	__be64 *bebits;
 	__le64 *lebits;
-	void *result;
+	void *result = NULL;
 	u64 bs;
 	int digestsize;
 	dma_addr_t addr_res, addr_pad;
@@ -285,13 +286,17 @@ int sun8i_ce_hash_run(struct crypto_engine *engine, void *breq)
 
 	/* the padding could be up to two block. */
 	buf = kzalloc(bs * 2, GFP_KERNEL | GFP_DMA);
-	if (!buf)
-		return -ENOMEM;
+	if (!buf) {
+		err = -ENOMEM;
+		goto theend;
+	}
 	bf = (__le32 *)buf;
 
 	result = kzalloc(digestsize, GFP_KERNEL | GFP_DMA);
-	if (!result)
-		return -ENOMEM;
+	if (!result) {
+		err = -ENOMEM;
+		goto theend;
+	}
 
 	flow = rctx->flow;
 	chan = &ce->chanlist[flow];
@@ -403,11 +408,11 @@ int sun8i_ce_hash_run(struct crypto_engine *engine, void *breq)
 	dma_unmap_sg(ce->dev, areq->src, nr_sgs, DMA_TO_DEVICE);
 	dma_unmap_single(ce->dev, addr_res, digestsize, DMA_FROM_DEVICE);
 
-	kfree(buf);
 
 	memcpy(areq->result, result, algt->alg.hash.halg.digestsize);
-	kfree(result);
 theend:
+	kfree(buf);
+	kfree(result);
 	crypto_finalize_hash_request(engine, breq, err);
 	return 0;
 }
diff --git a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h
index 558027516aed..cec781d5063c 100644
--- a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h
+++ b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h
@@ -16,7 +16,8 @@
 #include <crypto/internal/hash.h>
 #include <crypto/md5.h>
 #include <crypto/rng.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 /* CE Registers */
 #define CE_TDQ	0x00
diff --git a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-hash.c b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-hash.c
index b6ab2054f217..11cbcbc83a7b 100644
--- a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-hash.c
+++ b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-hash.c
@@ -13,7 +13,8 @@
 #include <linux/pm_runtime.h>
 #include <linux/scatterlist.h>
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/md5.h>
 #include "sun8i-ss.h"
 
diff --git a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h
index 1a66457f4a20..28188685b910 100644
--- a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h
+++ b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h
@@ -15,7 +15,8 @@
 #include <linux/crypto.h>
 #include <crypto/internal/hash.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #define SS_START	1
 
diff --git a/drivers/crypto/amcc/crypto4xx_alg.c b/drivers/crypto/amcc/crypto4xx_alg.c
index 7729a637fb02..a3fa849b139a 100644
--- a/drivers/crypto/amcc/crypto4xx_alg.c
+++ b/drivers/crypto/amcc/crypto4xx_alg.c
@@ -20,7 +20,7 @@
 #include <crypto/aead.h>
 #include <crypto/aes.h>
 #include <crypto/gcm.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
 #include <crypto/ctr.h>
 #include <crypto/skcipher.h>
 #include "crypto4xx_reg_def.h"
diff --git a/drivers/crypto/amcc/crypto4xx_core.c b/drivers/crypto/amcc/crypto4xx_core.c
index 981de43ea5e2..8d1b918a0533 100644
--- a/drivers/crypto/amcc/crypto4xx_core.c
+++ b/drivers/crypto/amcc/crypto4xx_core.c
@@ -30,7 +30,7 @@
 #include <crypto/aes.h>
 #include <crypto/ctr.h>
 #include <crypto/gcm.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
 #include <crypto/rng.h>
 #include <crypto/scatterwalk.h>
 #include <crypto/skcipher.h>
@@ -917,7 +917,7 @@ int crypto4xx_build_pd(struct crypto_async_request *req,
 	}
 
 	pd->pd_ctl.w = PD_CTL_HOST_READY |
-		((crypto_tfm_alg_type(req->tfm) == CRYPTO_ALG_TYPE_AHASH) |
+		((crypto_tfm_alg_type(req->tfm) == CRYPTO_ALG_TYPE_AHASH) ||
 		 (crypto_tfm_alg_type(req->tfm) == CRYPTO_ALG_TYPE_AEAD) ?
 			PD_CTL_HASH_FINAL : 0);
 	pd->pd_ctl_len.w = 0x00400000 | (assoclen + datalen);
diff --git a/drivers/crypto/atmel-authenc.h b/drivers/crypto/atmel-authenc.h
index c6530a1c8c20..45171e89a7d2 100644
--- a/drivers/crypto/atmel-authenc.h
+++ b/drivers/crypto/atmel-authenc.h
@@ -16,7 +16,8 @@
 
 #include <crypto/authenc.h>
 #include <crypto/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include "atmel-sha-regs.h"
 
 struct atmel_aes_dev;
diff --git a/drivers/crypto/atmel-sha.c b/drivers/crypto/atmel-sha.c
index 75ccf41a7cb9..352d80cb5ae9 100644
--- a/drivers/crypto/atmel-sha.c
+++ b/drivers/crypto/atmel-sha.c
@@ -33,7 +33,8 @@
 #include <linux/crypto.h>
 #include <crypto/scatterwalk.h>
 #include <crypto/algapi.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/hash.h>
 #include <crypto/internal/hash.h>
 #include "atmel-sha-regs.h"
@@ -459,7 +460,6 @@ static int atmel_sha_init(struct ahash_request *req)
 		break;
 	default:
 		return -EINVAL;
-		break;
 	}
 
 	ctx->bufcnt = 0;
diff --git a/drivers/crypto/axis/artpec6_crypto.c b/drivers/crypto/axis/artpec6_crypto.c
index 809c3033ca74..9ad188cffd0d 100644
--- a/drivers/crypto/axis/artpec6_crypto.c
+++ b/drivers/crypto/axis/artpec6_crypto.c
@@ -28,7 +28,8 @@
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/scatterwalk.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/xts.h>
 
 /* Max length of a line in all cache levels for Artpec SoCs. */
diff --git a/drivers/crypto/bcm/cipher.c b/drivers/crypto/bcm/cipher.c
index 50d169e61b41..30390a7324b2 100644
--- a/drivers/crypto/bcm/cipher.c
+++ b/drivers/crypto/bcm/cipher.c
@@ -26,11 +26,12 @@
 #include <crypto/aes.h>
 #include <crypto/internal/des.h>
 #include <crypto/hmac.h>
-#include <crypto/sha.h>
 #include <crypto/md5.h>
 #include <crypto/authenc.h>
 #include <crypto/skcipher.h>
 #include <crypto/hash.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/sha3.h>
 
 #include "util.h"
diff --git a/drivers/crypto/bcm/cipher.h b/drivers/crypto/bcm/cipher.h
index 035c8389cb3d..0ad5892b445d 100644
--- a/drivers/crypto/bcm/cipher.h
+++ b/drivers/crypto/bcm/cipher.h
@@ -16,7 +16,8 @@
 #include <crypto/aead.h>
 #include <crypto/arc4.h>
 #include <crypto/gcm.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/sha3.h>
 
 #include "spu.h"
diff --git a/drivers/crypto/bcm/spu.h b/drivers/crypto/bcm/spu.h
index dd132389bcaa..1c386a2d5506 100644
--- a/drivers/crypto/bcm/spu.h
+++ b/drivers/crypto/bcm/spu.h
@@ -17,7 +17,8 @@
 
 #include <linux/types.h>
 #include <linux/scatterlist.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 enum spu_cipher_alg {
 	CIPHER_ALG_NONE = 0x0,
diff --git a/drivers/crypto/caam/caamalg.c b/drivers/crypto/caam/caamalg.c
index cf5bd7666dfc..8697ae53b063 100644
--- a/drivers/crypto/caam/caamalg.c
+++ b/drivers/crypto/caam/caamalg.c
@@ -3404,8 +3404,8 @@ static int caam_cra_init(struct crypto_skcipher *tfm)
 		fallback = crypto_alloc_skcipher(tfm_name, 0,
 						 CRYPTO_ALG_NEED_FALLBACK);
 		if (IS_ERR(fallback)) {
-			dev_err(ctx->jrdev, "Failed to allocate %s fallback: %ld\n",
-				tfm_name, PTR_ERR(fallback));
+			pr_err("Failed to allocate %s fallback: %ld\n",
+			       tfm_name, PTR_ERR(fallback));
 			return PTR_ERR(fallback);
 		}
 
diff --git a/drivers/crypto/caam/caamalg_qi.c b/drivers/crypto/caam/caamalg_qi.c
index 66f60d78bdc8..189a7438b29c 100644
--- a/drivers/crypto/caam/caamalg_qi.c
+++ b/drivers/crypto/caam/caamalg_qi.c
@@ -852,7 +852,7 @@ static struct caam_drv_ctx *get_drv_ctx(struct caam_ctx *ctx,
 
 			cpu = smp_processor_id();
 			drv_ctx = caam_drv_ctx_init(ctx->qidev, &cpu, desc);
-			if (!IS_ERR_OR_NULL(drv_ctx))
+			if (!IS_ERR(drv_ctx))
 				drv_ctx->op_type = type;
 
 			ctx->drv_ctx[type] = drv_ctx;
@@ -955,7 +955,7 @@ static struct aead_edesc *aead_edesc_alloc(struct aead_request *req,
 	struct caam_drv_ctx *drv_ctx;
 
 	drv_ctx = get_drv_ctx(ctx, encrypt ? ENCRYPT : DECRYPT);
-	if (IS_ERR_OR_NULL(drv_ctx))
+	if (IS_ERR(drv_ctx))
 		return (struct aead_edesc *)drv_ctx;
 
 	/* allocate space for base edesc and hw desc commands, link tables */
@@ -1165,7 +1165,7 @@ static inline int aead_crypt(struct aead_request *req, bool encrypt)
 
 	/* allocate extended descriptor */
 	edesc = aead_edesc_alloc(req, encrypt);
-	if (IS_ERR_OR_NULL(edesc))
+	if (IS_ERR(edesc))
 		return PTR_ERR(edesc);
 
 	/* Create and submit job descriptor */
@@ -1259,7 +1259,7 @@ static struct skcipher_edesc *skcipher_edesc_alloc(struct skcipher_request *req,
 	struct caam_drv_ctx *drv_ctx;
 
 	drv_ctx = get_drv_ctx(ctx, encrypt ? ENCRYPT : DECRYPT);
-	if (IS_ERR_OR_NULL(drv_ctx))
+	if (IS_ERR(drv_ctx))
 		return (struct skcipher_edesc *)drv_ctx;
 
 	src_nents = sg_nents_for_len(req->src, req->cryptlen);
@@ -2502,8 +2502,8 @@ static int caam_cra_init(struct crypto_skcipher *tfm)
 		fallback = crypto_alloc_skcipher(tfm_name, 0,
 						 CRYPTO_ALG_NEED_FALLBACK);
 		if (IS_ERR(fallback)) {
-			dev_err(ctx->jrdev, "Failed to allocate %s fallback: %ld\n",
-				tfm_name, PTR_ERR(fallback));
+			pr_err("Failed to allocate %s fallback: %ld\n",
+			       tfm_name, PTR_ERR(fallback));
 			return PTR_ERR(fallback);
 		}
 
diff --git a/drivers/crypto/caam/caamalg_qi2.c b/drivers/crypto/caam/caamalg_qi2.c
index 98c1ff1744bb..a780e627838a 100644
--- a/drivers/crypto/caam/caamalg_qi2.c
+++ b/drivers/crypto/caam/caamalg_qi2.c
@@ -1611,7 +1611,8 @@ static int caam_cra_init_skcipher(struct crypto_skcipher *tfm)
 		fallback = crypto_alloc_skcipher(tfm_name, 0,
 						 CRYPTO_ALG_NEED_FALLBACK);
 		if (IS_ERR(fallback)) {
-			dev_err(ctx->dev, "Failed to allocate %s fallback: %ld\n",
+			dev_err(caam_alg->caam.dev,
+				"Failed to allocate %s fallback: %ld\n",
 				tfm_name, PTR_ERR(fallback));
 			return PTR_ERR(fallback);
 		}
diff --git a/drivers/crypto/caam/compat.h b/drivers/crypto/caam/compat.h
index c3c22a8de4c0..c4f79764172b 100644
--- a/drivers/crypto/caam/compat.h
+++ b/drivers/crypto/caam/compat.h
@@ -34,7 +34,8 @@
 #include <crypto/ctr.h>
 #include <crypto/internal/des.h>
 #include <crypto/gcm.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/md5.h>
 #include <crypto/chacha.h>
 #include <crypto/poly1305.h>
diff --git a/drivers/crypto/caam/intern.h b/drivers/crypto/caam/intern.h
index 9112279a4de0..7d45b21bd55a 100644
--- a/drivers/crypto/caam/intern.h
+++ b/drivers/crypto/caam/intern.h
@@ -16,6 +16,14 @@
 /* Currently comes from Kconfig param as a ^2 (driver-required) */
 #define JOBR_DEPTH (1 << CONFIG_CRYPTO_DEV_FSL_CAAM_RINGSIZE)
 
+/*
+ * Maximum size for crypto-engine software queue based on Job Ring
+ * size (JOBR_DEPTH) and a THRESHOLD (reserved for the non-crypto-API
+ * requests that are not passed through crypto-engine)
+ */
+#define THRESHOLD 15
+#define CRYPTO_ENGINE_MAX_QLEN (JOBR_DEPTH - THRESHOLD)
+
 /* Kconfig params for interrupt coalescing if selected (else zero) */
 #ifdef CONFIG_CRYPTO_DEV_FSL_CAAM_INTC
 #define JOBR_INTC JRCFG_ICEN
diff --git a/drivers/crypto/caam/jr.c b/drivers/crypto/caam/jr.c
index 6f669966ba2c..7f2b1101f567 100644
--- a/drivers/crypto/caam/jr.c
+++ b/drivers/crypto/caam/jr.c
@@ -550,7 +550,9 @@ static int caam_jr_probe(struct platform_device *pdev)
 	}
 
 	/* Initialize crypto engine */
-	jrpriv->engine = crypto_engine_alloc_init(jrdev, false);
+	jrpriv->engine = crypto_engine_alloc_init_and_set(jrdev, true, NULL,
+							  false,
+							  CRYPTO_ENGINE_MAX_QLEN);
 	if (!jrpriv->engine) {
 		dev_err(jrdev, "Could not init crypto-engine\n");
 		return -ENOMEM;
diff --git a/drivers/crypto/cavium/cpt/cptpf_main.c b/drivers/crypto/cavium/cpt/cptpf_main.c
index 781949027451..711b1acdd4e0 100644
--- a/drivers/crypto/cavium/cpt/cptpf_main.c
+++ b/drivers/crypto/cavium/cpt/cptpf_main.c
@@ -244,7 +244,7 @@ cpt_init_fail:
 
 struct ucode_header {
 	u8 version[CPT_UCODE_VERSION_SZ];
-	u32 code_length;
+	__be32 code_length;
 	u32 data_length;
 	u64 sram_address;
 };
@@ -288,10 +288,10 @@ static int cpt_ucode_load_fw(struct cpt_device *cpt, const u8 *fw, bool is_ae)
 
 	/* Byte swap 64-bit */
 	for (j = 0; j < (mcode->code_size / 8); j++)
-		((u64 *)mcode->code)[j] = cpu_to_be64(((u64 *)mcode->code)[j]);
+		((__be64 *)mcode->code)[j] = cpu_to_be64(((u64 *)mcode->code)[j]);
 	/*  MC needs 16-bit swap */
 	for (j = 0; j < (mcode->code_size / 2); j++)
-		((u16 *)mcode->code)[j] = cpu_to_be16(((u16 *)mcode->code)[j]);
+		((__be16 *)mcode->code)[j] = cpu_to_be16(((u16 *)mcode->code)[j]);
 
 	dev_dbg(dev, "mcode->code_size = %u\n", mcode->code_size);
 	dev_dbg(dev, "mcode->is_ae = %u\n", mcode->is_ae);
@@ -569,15 +569,9 @@ static int cpt_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto cpt_err_disable_device;
 	}
 
-	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(48));
+	err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
 	if (err) {
-		dev_err(dev, "Unable to get usable DMA configuration\n");
-		goto cpt_err_release_regions;
-	}
-
-	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(48));
-	if (err) {
-		dev_err(dev, "Unable to get 48-bit DMA for consistent allocations\n");
+		dev_err(dev, "Unable to get usable 48-bit DMA configuration\n");
 		goto cpt_err_release_regions;
 	}
 
diff --git a/drivers/crypto/cavium/cpt/cptvf_main.c b/drivers/crypto/cavium/cpt/cptvf_main.c
index a15245992cf9..f016448e43bb 100644
--- a/drivers/crypto/cavium/cpt/cptvf_main.c
+++ b/drivers/crypto/cavium/cpt/cptvf_main.c
@@ -687,15 +687,9 @@ static int cptvf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	}
 	/* Mark as VF driver */
 	cptvf->flags |= CPT_FLAG_VF_DRIVER;
-	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(48));
+	err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
 	if (err) {
-		dev_err(dev, "Unable to get usable DMA configuration\n");
-		goto cptvf_err_release_regions;
-	}
-
-	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(48));
-	if (err) {
-		dev_err(dev, "Unable to get 48-bit DMA for consistent allocations\n");
+		dev_err(dev, "Unable to get usable 48-bit DMA configuration\n");
 		goto cptvf_err_release_regions;
 	}
 
diff --git a/drivers/crypto/cavium/nitrox/nitrox_aead.c b/drivers/crypto/cavium/nitrox/nitrox_aead.c
index 1be2571363fe..c93c4e41d267 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_aead.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_aead.c
@@ -7,7 +7,6 @@
 #include <crypto/aead.h>
 #include <crypto/authenc.h>
 #include <crypto/des.h>
-#include <crypto/sha.h>
 #include <crypto/internal/aead.h>
 #include <crypto/scatterwalk.h>
 #include <crypto/gcm.h>
@@ -45,9 +44,9 @@ static int nitrox_aes_gcm_setkey(struct crypto_aead *aead, const u8 *key,
 
 	/* fill crypto context */
 	fctx = nctx->u.fctx;
-	flags.f = be64_to_cpu(fctx->flags.f);
+	flags.fu = be64_to_cpu(fctx->flags.f);
 	flags.w0.aes_keylen = aes_keylen;
-	fctx->flags.f = cpu_to_be64(flags.f);
+	fctx->flags.f = cpu_to_be64(flags.fu);
 
 	/* copy enc key to context */
 	memset(&fctx->crypto, 0, sizeof(fctx->crypto));
@@ -63,9 +62,9 @@ static int nitrox_aead_setauthsize(struct crypto_aead *aead,
 	struct flexi_crypto_context *fctx = nctx->u.fctx;
 	union fc_ctx_flags flags;
 
-	flags.f = be64_to_cpu(fctx->flags.f);
+	flags.fu = be64_to_cpu(fctx->flags.f);
 	flags.w0.mac_len = authsize;
-	fctx->flags.f = cpu_to_be64(flags.f);
+	fctx->flags.f = cpu_to_be64(flags.fu);
 
 	aead->authsize = authsize;
 
@@ -319,7 +318,7 @@ static int nitrox_gcm_common_init(struct crypto_aead *aead)
 	flags->w0.iv_source = IV_FROM_DPTR;
 	/* ask microcode to calculate ipad/opad */
 	flags->w0.auth_input_type = 1;
-	flags->f = be64_to_cpu(flags->f);
+	flags->f = cpu_to_be64(flags->fu);
 
 	return 0;
 }
diff --git a/drivers/crypto/cavium/nitrox/nitrox_debugfs.c b/drivers/crypto/cavium/nitrox/nitrox_debugfs.c
index 16f7d0bd1303..741572a01995 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_debugfs.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_debugfs.c
@@ -3,6 +3,7 @@
 #include <linux/debugfs.h>
 
 #include "nitrox_csr.h"
+#include "nitrox_debugfs.h"
 #include "nitrox_dev.h"
 
 static int firmware_show(struct seq_file *s, void *v)
diff --git a/drivers/crypto/cavium/nitrox/nitrox_hal.c b/drivers/crypto/cavium/nitrox/nitrox_hal.c
index 34a2f4f30a7e..13b137410b75 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_hal.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_hal.c
@@ -3,6 +3,7 @@
 
 #include "nitrox_dev.h"
 #include "nitrox_csr.h"
+#include "nitrox_hal.h"
 
 #define PLL_REF_CLK 50
 #define MAX_CSR_RETRIES 10
diff --git a/drivers/crypto/cavium/nitrox/nitrox_isr.c b/drivers/crypto/cavium/nitrox/nitrox_isr.c
index 3dec570a190a..99b053094f5a 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_isr.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_isr.c
@@ -7,6 +7,7 @@
 #include "nitrox_csr.h"
 #include "nitrox_common.h"
 #include "nitrox_hal.h"
+#include "nitrox_isr.h"
 #include "nitrox_mbx.h"
 
 /**
diff --git a/drivers/crypto/cavium/nitrox/nitrox_isr.h b/drivers/crypto/cavium/nitrox/nitrox_isr.h
index 1062c9336c1f..2bb123cd2f1c 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_isr.h
+++ b/drivers/crypto/cavium/nitrox/nitrox_isr.h
@@ -9,4 +9,13 @@ void nitrox_unregister_interrupts(struct nitrox_device *ndev);
 int nitrox_sriov_register_interupts(struct nitrox_device *ndev);
 void nitrox_sriov_unregister_interrupts(struct nitrox_device *ndev);
 
+#ifdef CONFIG_PCI_IOV
+int nitrox_sriov_configure(struct pci_dev *pdev, int num_vfs);
+#else
+static inline int nitrox_sriov_configure(struct pci_dev *pdev, int num_vfs)
+{
+	return 0;
+}
+#endif
+
 #endif /* __NITROX_ISR_H */
diff --git a/drivers/crypto/cavium/nitrox/nitrox_main.c b/drivers/crypto/cavium/nitrox/nitrox_main.c
index 9d14be97e381..facc8e6bc580 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_main.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_main.c
@@ -49,15 +49,6 @@ static unsigned int qlen = DEFAULT_CMD_QLEN;
 module_param(qlen, uint, 0644);
 MODULE_PARM_DESC(qlen, "Command queue length - default 2048");
 
-#ifdef CONFIG_PCI_IOV
-int nitrox_sriov_configure(struct pci_dev *pdev, int num_vfs);
-#else
-int nitrox_sriov_configure(struct pci_dev *pdev, int num_vfs)
-{
-	return 0;
-}
-#endif
-
 /**
  * struct ucode - Firmware Header
  * @id: microcode ID
@@ -555,10 +546,8 @@ static void nitrox_remove(struct pci_dev *pdev)
 
 	nitrox_remove_from_devlist(ndev);
 
-#ifdef CONFIG_PCI_IOV
 	/* disable SR-IOV */
 	nitrox_sriov_configure(pdev, 0);
-#endif
 	nitrox_crypto_unregister();
 	nitrox_debugfs_exit(ndev);
 	nitrox_pf_sw_cleanup(ndev);
@@ -584,9 +573,7 @@ static struct pci_driver nitrox_driver = {
 	.probe = nitrox_probe,
 	.remove	= nitrox_remove,
 	.shutdown = nitrox_shutdown,
-#ifdef CONFIG_PCI_IOV
 	.sriov_configure = nitrox_sriov_configure,
-#endif
 };
 
 module_pci_driver(nitrox_driver);
diff --git a/drivers/crypto/cavium/nitrox/nitrox_mbx.c b/drivers/crypto/cavium/nitrox/nitrox_mbx.c
index b51b0449b478..c1af9d4fca6e 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_mbx.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_mbx.c
@@ -4,6 +4,7 @@
 #include "nitrox_csr.h"
 #include "nitrox_hal.h"
 #include "nitrox_dev.h"
+#include "nitrox_mbx.h"
 
 #define RING_TO_VFNO(_x, _y)	((_x) / (_y))
 
@@ -112,7 +113,7 @@ static void pf2vf_resp_handler(struct work_struct *work)
 	case MBX_MSG_TYPE_ACK:
 	case MBX_MSG_TYPE_NACK:
 		break;
-	};
+	}
 
 	kfree(pf2vf_resp);
 }
diff --git a/drivers/crypto/cavium/nitrox/nitrox_req.h b/drivers/crypto/cavium/nitrox/nitrox_req.h
index 12282c1b14f5..ed174883c8e3 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_req.h
+++ b/drivers/crypto/cavium/nitrox/nitrox_req.h
@@ -149,6 +149,7 @@ struct auth_keys {
 
 union fc_ctx_flags {
 	__be64 f;
+	u64 fu;
 	struct {
 #if defined(__BIG_ENDIAN_BITFIELD)
 		u64 cipher_type	: 4;
@@ -280,6 +281,7 @@ struct nitrox_rfc4106_rctx {
  *   - packet payload bytes
  */
 union pkt_instr_hdr {
+	__be64 bev;
 	u64 value;
 	struct {
 #if defined(__BIG_ENDIAN_BITFIELD)
@@ -324,6 +326,7 @@ union pkt_instr_hdr {
  * @ctxp: Context pointer. CTXP<63,2:0> must be zero in all cases.
  */
 union pkt_hdr {
+	__be64 bev[2];
 	u64 value[2];
 	struct {
 #if defined(__BIG_ENDIAN_BITFIELD)
@@ -370,6 +373,7 @@ union pkt_hdr {
  *        sglist components at [RPTR] on the remote host.
  */
 union slc_store_info {
+	__be64 bev[2];
 	u64 value[2];
 	struct {
 #if defined(__BIG_ENDIAN_BITFIELD)
diff --git a/drivers/crypto/cavium/nitrox/nitrox_reqmgr.c b/drivers/crypto/cavium/nitrox/nitrox_reqmgr.c
index 5826c2c98a50..53ef06792133 100644
--- a/drivers/crypto/cavium/nitrox/nitrox_reqmgr.c
+++ b/drivers/crypto/cavium/nitrox/nitrox_reqmgr.c
@@ -3,6 +3,7 @@
 #include <linux/workqueue.h>
 #include <crypto/internal/skcipher.h>
 
+#include "nitrox_common.h"
 #include "nitrox_dev.h"
 #include "nitrox_req.h"
 #include "nitrox_csr.h"
@@ -448,7 +449,7 @@ int nitrox_process_se_request(struct nitrox_device *ndev,
 	sr->instr.ih.s.ssz = sr->out.sgmap_cnt;
 	sr->instr.ih.s.fsz = FDATA_SIZE + sizeof(struct gphdr);
 	sr->instr.ih.s.tlen = sr->instr.ih.s.fsz + sr->in.total_bytes;
-	sr->instr.ih.value = cpu_to_be64(sr->instr.ih.value);
+	sr->instr.ih.bev = cpu_to_be64(sr->instr.ih.value);
 
 	/* word 2 */
 	sr->instr.irh.value[0] = 0;
@@ -460,7 +461,7 @@ int nitrox_process_se_request(struct nitrox_device *ndev,
 	sr->instr.irh.s.ctxc = req->ctrl.s.ctxc;
 	sr->instr.irh.s.arg = req->ctrl.s.arg;
 	sr->instr.irh.s.opcode = req->opcode;
-	sr->instr.irh.value[0] = cpu_to_be64(sr->instr.irh.value[0]);
+	sr->instr.irh.bev[0] = cpu_to_be64(sr->instr.irh.value[0]);
 
 	/* word 3 */
 	sr->instr.irh.s.ctxp = cpu_to_be64(ctx_handle);
@@ -468,7 +469,7 @@ int nitrox_process_se_request(struct nitrox_device *ndev,
 	/* word 4 */
 	sr->instr.slc.value[0] = 0;
 	sr->instr.slc.s.ssz = sr->out.sgmap_cnt;
-	sr->instr.slc.value[0] = cpu_to_be64(sr->instr.slc.value[0]);
+	sr->instr.slc.bev[0] = cpu_to_be64(sr->instr.slc.value[0]);
 
 	/* word 5 */
 	sr->instr.slc.s.rptr = cpu_to_be64(sr->out.sgcomp_dma);
diff --git a/drivers/crypto/cavium/zip/zip_main.c b/drivers/crypto/cavium/zip/zip_main.c
index d35216e2f6cd..812b4ac9afd6 100644
--- a/drivers/crypto/cavium/zip/zip_main.c
+++ b/drivers/crypto/cavium/zip/zip_main.c
@@ -263,15 +263,9 @@ static int zip_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto err_disable_device;
 	}
 
-	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(48));
+	err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
 	if (err) {
-		dev_err(dev, "Unable to get usable DMA configuration\n");
-		goto err_release_regions;
-	}
-
-	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(48));
-	if (err) {
-		dev_err(dev, "Unable to get 48-bit DMA for allocations\n");
+		dev_err(dev, "Unable to get usable 48-bit DMA configuration\n");
 		goto err_release_regions;
 	}
 
diff --git a/drivers/crypto/ccp/ccp-crypto-sha.c b/drivers/crypto/ccp/ccp-crypto-sha.c
index 8fbfdb9e8cd3..74fa5360e722 100644
--- a/drivers/crypto/ccp/ccp-crypto-sha.c
+++ b/drivers/crypto/ccp/ccp-crypto-sha.c
@@ -17,7 +17,8 @@
 #include <crypto/hash.h>
 #include <crypto/hmac.h>
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/scatterwalk.h>
 #include <linux/string.h>
 
diff --git a/drivers/crypto/ccp/ccp-crypto.h b/drivers/crypto/ccp/ccp-crypto.h
index aed3d2192d01..e42450d07168 100644
--- a/drivers/crypto/ccp/ccp-crypto.h
+++ b/drivers/crypto/ccp/ccp-crypto.h
@@ -19,7 +19,8 @@
 #include <crypto/aead.h>
 #include <crypto/ctr.h>
 #include <crypto/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/akcipher.h>
 #include <crypto/skcipher.h>
 #include <crypto/internal/rsa.h>
diff --git a/drivers/crypto/ccree/cc_cipher.c b/drivers/crypto/ccree/cc_cipher.c
index dafa6577a845..cdfee501fbd9 100644
--- a/drivers/crypto/ccree/cc_cipher.c
+++ b/drivers/crypto/ccree/cc_cipher.c
@@ -97,6 +97,7 @@ static int validate_keys_sizes(struct cc_cipher_ctx *ctx_p, u32 size)
 	case S_DIN_to_SM4:
 		if (size == SM4_KEY_SIZE)
 			return 0;
+		break;
 	default:
 		break;
 	}
@@ -139,9 +140,11 @@ static int validate_data_size(struct cc_cipher_ctx *ctx_p,
 		case DRV_CIPHER_CBC:
 			if (IS_ALIGNED(size, SM4_BLOCK_SIZE))
 				return 0;
+			break;
 		default:
 			break;
 		}
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/crypto/ccree/cc_driver.c b/drivers/crypto/ccree/cc_driver.c
index 6f519d3e896c..d0e59e942568 100644
--- a/drivers/crypto/ccree/cc_driver.c
+++ b/drivers/crypto/ccree/cc_driver.c
@@ -100,6 +100,57 @@ static const struct of_device_id arm_ccree_dev_of_match[] = {
 };
 MODULE_DEVICE_TABLE(of, arm_ccree_dev_of_match);
 
+static void init_cc_cache_params(struct cc_drvdata *drvdata)
+{
+	struct device *dev = drvdata_to_dev(drvdata);
+	u32 cache_params, ace_const, val, mask;
+
+	/* compute CC_AXIM_CACHE_PARAMS */
+	cache_params = cc_ioread(drvdata, CC_REG(AXIM_CACHE_PARAMS));
+	dev_dbg(dev, "Cache params previous: 0x%08X\n", cache_params);
+
+	/* non cached or write-back, write allocate */
+	val = drvdata->coherent ? 0xb : 0x2;
+
+	mask = CC_GENMASK(CC_AXIM_CACHE_PARAMS_AWCACHE);
+	cache_params &= ~mask;
+	cache_params |= FIELD_PREP(mask, val);
+
+	mask = CC_GENMASK(CC_AXIM_CACHE_PARAMS_AWCACHE_LAST);
+	cache_params &= ~mask;
+	cache_params |= FIELD_PREP(mask, val);
+
+	mask = CC_GENMASK(CC_AXIM_CACHE_PARAMS_ARCACHE);
+	cache_params &= ~mask;
+	cache_params |= FIELD_PREP(mask, val);
+
+	drvdata->cache_params = cache_params;
+
+	dev_dbg(dev, "Cache params current: 0x%08X\n", cache_params);
+
+	if (drvdata->hw_rev <= CC_HW_REV_710)
+		return;
+
+	/* compute CC_AXIM_ACE_CONST */
+	ace_const = cc_ioread(drvdata, CC_REG(AXIM_ACE_CONST));
+	dev_dbg(dev, "ACE-const previous: 0x%08X\n", ace_const);
+
+	/* system or outer-sharable */
+	val = drvdata->coherent ? 0x2 : 0x3;
+
+	mask = CC_GENMASK(CC_AXIM_ACE_CONST_ARDOMAIN);
+	ace_const &= ~mask;
+	ace_const |= FIELD_PREP(mask, val);
+
+	mask = CC_GENMASK(CC_AXIM_ACE_CONST_AWDOMAIN);
+	ace_const &= ~mask;
+	ace_const |= FIELD_PREP(mask, val);
+
+	dev_dbg(dev, "ACE-const current: 0x%08X\n", ace_const);
+
+	drvdata->ace_const = ace_const;
+}
+
 static u32 cc_read_idr(struct cc_drvdata *drvdata, const u32 *idr_offsets)
 {
 	int i;
@@ -218,9 +269,9 @@ bool cc_wait_for_reset_completion(struct cc_drvdata *drvdata)
 	return false;
 }
 
-int init_cc_regs(struct cc_drvdata *drvdata, bool is_probe)
+int init_cc_regs(struct cc_drvdata *drvdata)
 {
-	unsigned int val, cache_params;
+	unsigned int val;
 	struct device *dev = drvdata_to_dev(drvdata);
 
 	/* Unmask all AXI interrupt sources AXI_CFG1 register   */
@@ -245,19 +296,9 @@ int init_cc_regs(struct cc_drvdata *drvdata, bool is_probe)
 
 	cc_iowrite(drvdata, CC_REG(HOST_IMR), ~val);
 
-	cache_params = (drvdata->coherent ? CC_COHERENT_CACHE_PARAMS : 0x0);
-
-	val = cc_ioread(drvdata, CC_REG(AXIM_CACHE_PARAMS));
-
-	if (is_probe)
-		dev_dbg(dev, "Cache params previous: 0x%08X\n", val);
-
-	cc_iowrite(drvdata, CC_REG(AXIM_CACHE_PARAMS), cache_params);
-	val = cc_ioread(drvdata, CC_REG(AXIM_CACHE_PARAMS));
-
-	if (is_probe)
-		dev_dbg(dev, "Cache params current: 0x%08X (expect: 0x%08X)\n",
-			val, cache_params);
+	cc_iowrite(drvdata, CC_REG(AXIM_CACHE_PARAMS), drvdata->cache_params);
+	if (drvdata->hw_rev >= CC_HW_REV_712)
+		cc_iowrite(drvdata, CC_REG(AXIM_ACE_CONST), drvdata->ace_const);
 
 	return 0;
 }
@@ -445,7 +486,9 @@ static int init_cc_resources(struct platform_device *plat_dev)
 	}
 	dev_dbg(dev, "Registered to IRQ: %d\n", irq);
 
-	rc = init_cc_regs(new_drvdata, true);
+	init_cc_cache_params(new_drvdata);
+
+	rc = init_cc_regs(new_drvdata);
 	if (rc) {
 		dev_err(dev, "init_cc_regs failed\n");
 		goto post_pm_err;
diff --git a/drivers/crypto/ccree/cc_driver.h b/drivers/crypto/ccree/cc_driver.h
index af77b2020350..5f1d4602eb8f 100644
--- a/drivers/crypto/ccree/cc_driver.h
+++ b/drivers/crypto/ccree/cc_driver.h
@@ -17,7 +17,8 @@
 #include <crypto/algapi.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/aes.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/aead.h>
 #include <crypto/authenc.h>
 #include <crypto/hash.h>
@@ -49,8 +50,6 @@ enum cc_std_body {
 	CC_STD_ALL = 0x3
 };
 
-#define CC_COHERENT_CACHE_PARAMS 0xEEE
-
 #define CC_PINS_FULL	0x0
 #define CC_PINS_SLIM	0x9F
 
@@ -155,6 +154,8 @@ struct cc_drvdata {
 	int std_bodies;
 	bool sec_disabled;
 	u32 comp_mask;
+	u32 cache_params;
+	u32 ace_const;
 };
 
 struct cc_crypto_alg {
@@ -205,7 +206,7 @@ static inline void dump_byte_array(const char *name, const u8 *the_array,
 }
 
 bool cc_wait_for_reset_completion(struct cc_drvdata *drvdata);
-int init_cc_regs(struct cc_drvdata *drvdata, bool is_probe);
+int init_cc_regs(struct cc_drvdata *drvdata);
 void fini_cc_regs(struct cc_drvdata *drvdata);
 unsigned int cc_get_default_hash_len(struct cc_drvdata *drvdata);
 
diff --git a/drivers/crypto/ccree/cc_pm.c b/drivers/crypto/ccree/cc_pm.c
index 3c65bf070c90..d5421b0c6831 100644
--- a/drivers/crypto/ccree/cc_pm.c
+++ b/drivers/crypto/ccree/cc_pm.c
@@ -45,7 +45,7 @@ static int cc_pm_resume(struct device *dev)
 	}
 
 	cc_iowrite(drvdata, CC_REG(HOST_POWER_DOWN_EN), POWER_DOWN_DISABLE);
-	rc = init_cc_regs(drvdata, false);
+	rc = init_cc_regs(drvdata);
 	if (rc) {
 		dev_err(dev, "init_cc_regs (%x)\n", rc);
 		return rc;
diff --git a/drivers/crypto/chelsio/chcr_algo.c b/drivers/crypto/chelsio/chcr_algo.c
index 13b908ea4873..f5a336634daa 100644
--- a/drivers/crypto/chelsio/chcr_algo.c
+++ b/drivers/crypto/chelsio/chcr_algo.c
@@ -53,7 +53,8 @@
 #include <crypto/algapi.h>
 #include <crypto/hash.h>
 #include <crypto/gcm.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/authenc.h>
 #include <crypto/ctr.h>
 #include <crypto/gf128mul.h>
diff --git a/drivers/crypto/hisilicon/Kconfig b/drivers/crypto/hisilicon/Kconfig
index 9c3b3ca815e6..843192666dc3 100644
--- a/drivers/crypto/hisilicon/Kconfig
+++ b/drivers/crypto/hisilicon/Kconfig
@@ -71,3 +71,11 @@ config CRYPTO_DEV_HISI_HPRE
 	help
 	  Support for HiSilicon HPRE(High Performance RSA Engine)
 	  accelerator, which can accelerate RSA and DH algorithms.
+
+config CRYPTO_DEV_HISI_TRNG
+	tristate "Support for HISI TRNG Driver"
+	depends on ARM64 && ACPI
+	select HW_RANDOM
+	select CRYPTO_RNG
+	help
+	  Support for HiSilicon TRNG Driver.
diff --git a/drivers/crypto/hisilicon/Makefile b/drivers/crypto/hisilicon/Makefile
index 7f5f74c72baa..1e89269a2e4b 100644
--- a/drivers/crypto/hisilicon/Makefile
+++ b/drivers/crypto/hisilicon/Makefile
@@ -5,3 +5,4 @@ obj-$(CONFIG_CRYPTO_DEV_HISI_SEC2) += sec2/
 obj-$(CONFIG_CRYPTO_DEV_HISI_QM) += hisi_qm.o
 hisi_qm-objs = qm.o sgl.o
 obj-$(CONFIG_CRYPTO_DEV_HISI_ZIP) += zip/
+obj-$(CONFIG_CRYPTO_DEV_HISI_TRNG) += trng/
diff --git a/drivers/crypto/hisilicon/hpre/hpre_main.c b/drivers/crypto/hisilicon/hpre/hpre_main.c
index a33394d91bbf..e5c991913f09 100644
--- a/drivers/crypto/hisilicon/hpre/hpre_main.c
+++ b/drivers/crypto/hisilicon/hpre/hpre_main.c
@@ -705,9 +705,7 @@ static int hpre_debugfs_init(struct hisi_qm *qm)
 
 	qm->debug.sqe_mask_offset = HPRE_SQE_MASK_OFFSET;
 	qm->debug.sqe_mask_len = HPRE_SQE_MASK_LEN;
-	ret = hisi_qm_debug_init(qm);
-	if (ret)
-		goto failed_to_create;
+	hisi_qm_debug_init(qm);
 
 	if (qm->pdev->device == HPRE_PCI_DEVICE_ID) {
 		ret = hpre_ctrl_debug_init(qm);
diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c
index 530f23116d7c..f21ccae0e8ea 100644
--- a/drivers/crypto/hisilicon/qm.c
+++ b/drivers/crypto/hisilicon/qm.c
@@ -473,7 +473,7 @@ static int qm_wait_mb_ready(struct hisi_qm *qm)
 
 	return readl_relaxed_poll_timeout(qm->io_base + QM_MB_CMD_SEND_BASE,
 					  val, !((val >> QM_MB_BUSY_SHIFT) &
-					  0x1), 10, 1000);
+					  0x1), POLL_PERIOD, POLL_TIMEOUT);
 }
 
 /* 128 bit should be written to hardware at one time to trigger a mailbox */
@@ -583,7 +583,8 @@ static int qm_dev_mem_reset(struct hisi_qm *qm)
 
 	writel(0x1, qm->io_base + QM_MEM_START_INIT);
 	return readl_relaxed_poll_timeout(qm->io_base + QM_MEM_INIT_DONE, val,
-					  val & BIT(0), 10, 1000);
+					  val & BIT(0), POLL_PERIOD,
+					  POLL_TIMEOUT);
 }
 
 static u32 qm_get_irq_num_v1(struct hisi_qm *qm)
@@ -804,7 +805,8 @@ static int qm_set_vft_common(struct hisi_qm *qm, enum vft_type type,
 	int ret;
 
 	ret = readl_relaxed_poll_timeout(qm->io_base + QM_VFT_CFG_RDY, val,
-					 val & BIT(0), 10, 1000);
+					 val & BIT(0), POLL_PERIOD,
+					 POLL_TIMEOUT);
 	if (ret)
 		return ret;
 
@@ -818,7 +820,8 @@ static int qm_set_vft_common(struct hisi_qm *qm, enum vft_type type,
 	writel(0x1, qm->io_base + QM_VFT_CFG_OP_ENABLE);
 
 	return readl_relaxed_poll_timeout(qm->io_base + QM_VFT_CFG_RDY, val,
-					  val & BIT(0), 10, 1000);
+					  val & BIT(0), POLL_PERIOD,
+					  POLL_TIMEOUT);
 }
 
 /* The config should be conducted after qm_dev_mem_reset() */
@@ -929,7 +932,8 @@ static ssize_t qm_debug_read(struct file *filp, char __user *buf,
 		return -EINVAL;
 	}
 	mutex_unlock(&file->lock);
-	ret = sprintf(tbuf, "%u\n", val);
+
+	ret = scnprintf(tbuf, QM_DBG_TMP_BUF_LEN, "%u\n", val);
 	return simple_read_from_buffer(buf, count, pos, tbuf, ret);
 }
 
@@ -1517,7 +1521,7 @@ static const struct file_operations qm_cmd_fops = {
 	.write = qm_cmd_write,
 };
 
-static int qm_create_debugfs_file(struct hisi_qm *qm, enum qm_debug_file index)
+static void qm_create_debugfs_file(struct hisi_qm *qm, enum qm_debug_file index)
 {
 	struct dentry *qm_d = qm->debug.qm_d;
 	struct debugfs_file *file = qm->debug.files + index;
@@ -1528,8 +1532,6 @@ static int qm_create_debugfs_file(struct hisi_qm *qm, enum qm_debug_file index)
 	file->index = index;
 	mutex_init(&file->lock);
 	file->debug = &qm->debug;
-
-	return 0;
 }
 
 static void qm_hw_error_init_v1(struct hisi_qm *qm, u32 ce, u32 nfe, u32 fe)
@@ -1733,19 +1735,15 @@ void hisi_qm_release_qp(struct hisi_qp *qp)
 }
 EXPORT_SYMBOL_GPL(hisi_qm_release_qp);
 
-static int qm_qp_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
+static int qm_sq_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
 {
 	struct hisi_qm *qm = qp->qm;
 	struct device *dev = &qm->pdev->dev;
 	enum qm_hw_ver ver = qm->ver;
 	struct qm_sqc *sqc;
-	struct qm_cqc *cqc;
 	dma_addr_t sqc_dma;
-	dma_addr_t cqc_dma;
 	int ret;
 
-	qm_init_qp_status(qp);
-
 	sqc = kzalloc(sizeof(struct qm_sqc), GFP_KERNEL);
 	if (!sqc)
 		return -ENOMEM;
@@ -1770,8 +1768,18 @@ static int qm_qp_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
 	ret = qm_mb(qm, QM_MB_CMD_SQC, sqc_dma, qp_id, 0);
 	dma_unmap_single(dev, sqc_dma, sizeof(struct qm_sqc), DMA_TO_DEVICE);
 	kfree(sqc);
-	if (ret)
-		return ret;
+
+	return ret;
+}
+
+static int qm_cq_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
+{
+	struct hisi_qm *qm = qp->qm;
+	struct device *dev = &qm->pdev->dev;
+	enum qm_hw_ver ver = qm->ver;
+	struct qm_cqc *cqc;
+	dma_addr_t cqc_dma;
+	int ret;
 
 	cqc = kzalloc(sizeof(struct qm_cqc), GFP_KERNEL);
 	if (!cqc)
@@ -1785,11 +1793,12 @@ static int qm_qp_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
 
 	INIT_QC_COMMON(cqc, qp->cqe_dma, pasid);
 	if (ver == QM_HW_V1) {
-		cqc->dw3 = cpu_to_le32(QM_MK_CQC_DW3_V1(0, 0, 0, 4));
+		cqc->dw3 = cpu_to_le32(QM_MK_CQC_DW3_V1(0, 0, 0,
+							QM_QC_CQE_SIZE));
 		cqc->w8 = cpu_to_le16(QM_Q_DEPTH - 1);
 	} else {
-		cqc->dw3 = cpu_to_le32(QM_MK_CQC_DW3_V2(4));
-		cqc->w8 = 0;
+		cqc->dw3 = cpu_to_le32(QM_MK_CQC_DW3_V2(QM_QC_CQE_SIZE));
+		cqc->w8 = 0; /* rand_qc */
 	}
 	cqc->dw6 = cpu_to_le32(1 << QM_CQ_PHASE_SHIFT | 1 << QM_CQ_FLAG_SHIFT);
 
@@ -1800,6 +1809,19 @@ static int qm_qp_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
 	return ret;
 }
 
+static int qm_qp_ctx_cfg(struct hisi_qp *qp, int qp_id, u32 pasid)
+{
+	int ret;
+
+	qm_init_qp_status(qp);
+
+	ret = qm_sq_ctx_cfg(qp, qp_id, pasid);
+	if (ret)
+		return ret;
+
+	return qm_cq_ctx_cfg(qp, qp_id, pasid);
+}
+
 static int qm_start_qp_nolock(struct hisi_qp *qp, unsigned long arg)
 {
 	struct hisi_qm *qm = qp->qm;
@@ -1843,6 +1865,9 @@ int hisi_qm_start_qp(struct hisi_qp *qp, unsigned long arg)
 EXPORT_SYMBOL_GPL(hisi_qm_start_qp);
 
 /**
+ * qm_drain_qp() - Drain a qp.
+ * @qp: The qp we want to drain.
+ *
  * Determine whether the queue is cleared by judging the tail pointers of
  * sq and cq.
  */
@@ -2008,7 +2033,8 @@ static void hisi_qm_cache_wb(struct hisi_qm *qm)
 
 	writel(0x1, qm->io_base + QM_CACHE_WB_START);
 	if (readl_relaxed_poll_timeout(qm->io_base + QM_CACHE_WB_DONE,
-					    val, val & BIT(0), 10, 1000))
+				       val, val & BIT(0), POLL_PERIOD,
+				       POLL_TIMEOUT))
 		dev_err(&qm->pdev->dev, "QM writeback sqc cache fail!\n");
 }
 
@@ -2112,7 +2138,7 @@ static void hisi_qm_uacce_stop_queue(struct uacce_queue *q)
 	hisi_qm_stop_qp(q->priv);
 }
 
-static int qm_set_sqctype(struct uacce_queue *q, u16 type)
+static void qm_set_sqctype(struct uacce_queue *q, u16 type)
 {
 	struct hisi_qm *qm = q->uacce->priv;
 	struct hisi_qp *qp = q->priv;
@@ -2120,8 +2146,6 @@ static int qm_set_sqctype(struct uacce_queue *q, u16 type)
 	down_write(&qm->qps_lock);
 	qp->alg_type = type;
 	up_write(&qm->qps_lock);
-
-	return 0;
 }
 
 static long hisi_qm_uacce_ioctl(struct uacce_queue *q, unsigned int cmd,
@@ -2418,6 +2442,16 @@ static void hisi_qm_pre_init(struct hisi_qm *qm)
 	qm->is_frozen = false;
 }
 
+static void hisi_qm_pci_uninit(struct hisi_qm *qm)
+{
+	struct pci_dev *pdev = qm->pdev;
+
+	pci_free_irq_vectors(pdev);
+	iounmap(qm->io_base);
+	pci_release_mem_regions(pdev);
+	pci_disable_device(pdev);
+}
+
 /**
  * hisi_qm_uninit() - Uninitialize qm.
  * @qm: The qm needed uninit.
@@ -2436,9 +2470,6 @@ void hisi_qm_uninit(struct hisi_qm *qm)
 		return;
 	}
 
-	uacce_remove(qm->uacce);
-	qm->uacce = NULL;
-
 	hisi_qp_memory_uninit(qm, qm->qp_num);
 	idr_destroy(&qm->qp_idr);
 
@@ -2450,10 +2481,9 @@ void hisi_qm_uninit(struct hisi_qm *qm)
 	}
 
 	qm_irq_unregister(qm);
-	pci_free_irq_vectors(pdev);
-	iounmap(qm->io_base);
-	pci_release_mem_regions(pdev);
-	pci_disable_device(pdev);
+	hisi_qm_pci_uninit(qm);
+	uacce_remove(qm->uacce);
+	qm->uacce = NULL;
 
 	up_write(&qm->qps_lock);
 }
@@ -2486,6 +2516,12 @@ int hisi_qm_get_vft(struct hisi_qm *qm, u32 *base, u32 *number)
 EXPORT_SYMBOL_GPL(hisi_qm_get_vft);
 
 /**
+ * hisi_qm_set_vft() - Set vft to a qm.
+ * @qm: The qm we want to set its vft.
+ * @fun_num: The function number.
+ * @base: The base number of queue in vft.
+ * @number: The number of queues in vft.
+ *
  * This function is alway called in PF driver, it is used to assign queues
  * among PF and VFs.
  *
@@ -2519,14 +2555,10 @@ static int qm_eq_ctx_cfg(struct hisi_qm *qm)
 {
 	struct device *dev = &qm->pdev->dev;
 	struct qm_eqc *eqc;
-	struct qm_aeqc *aeqc;
 	dma_addr_t eqc_dma;
-	dma_addr_t aeqc_dma;
 	int ret;
 
-	qm_init_eq_aeq_status(qm);
-
-	eqc = kzalloc(sizeof(struct qm_eqc), GFP_KERNEL);
+	eqc = kzalloc(sizeof(struct qm_eqc), GFP_KERNEL); //todo
 	if (!eqc)
 		return -ENOMEM;
 	eqc_dma = dma_map_single(dev, eqc, sizeof(struct qm_eqc),
@@ -2541,11 +2573,20 @@ static int qm_eq_ctx_cfg(struct hisi_qm *qm)
 	if (qm->ver == QM_HW_V1)
 		eqc->dw3 = cpu_to_le32(QM_EQE_AEQE_SIZE);
 	eqc->dw6 = cpu_to_le32((QM_EQ_DEPTH - 1) | (1 << QM_EQC_PHASE_SHIFT));
+
 	ret = qm_mb(qm, QM_MB_CMD_EQC, eqc_dma, 0, 0);
 	dma_unmap_single(dev, eqc_dma, sizeof(struct qm_eqc), DMA_TO_DEVICE);
 	kfree(eqc);
-	if (ret)
-		return ret;
+
+	return ret;
+}
+
+static int qm_aeq_ctx_cfg(struct hisi_qm *qm)
+{
+	struct device *dev = &qm->pdev->dev;
+	struct qm_aeqc *aeqc;
+	dma_addr_t aeqc_dma;
+	int ret;
 
 	aeqc = kzalloc(sizeof(struct qm_aeqc), GFP_KERNEL);
 	if (!aeqc)
@@ -2568,6 +2609,22 @@ static int qm_eq_ctx_cfg(struct hisi_qm *qm)
 	return ret;
 }
 
+static int qm_eq_aeq_ctx_cfg(struct hisi_qm *qm)
+{
+	struct device *dev = &qm->pdev->dev;
+	int ret;
+
+	qm_init_eq_aeq_status(qm);
+
+	ret = qm_eq_ctx_cfg(qm);
+	if (ret) {
+		dev_err(dev, "Set eqc failed!\n");
+		return ret;
+	}
+
+	return qm_aeq_ctx_cfg(qm);
+}
+
 static int __hisi_qm_start(struct hisi_qm *qm)
 {
 	int ret;
@@ -2584,7 +2641,7 @@ static int __hisi_qm_start(struct hisi_qm *qm)
 			return ret;
 	}
 
-	ret = qm_eq_ctx_cfg(qm);
+	ret = qm_eq_aeq_ctx_cfg(qm);
 	if (ret)
 		return ret;
 
@@ -2690,7 +2747,11 @@ static int qm_stop_started_qp(struct hisi_qm *qm)
 	return 0;
 }
 
+
 /**
+ * qm_clear_queues() - Clear all queues memory in a qm.
+ * @qm: The qm in which the queues will be cleared.
+ *
  * This function clears all queues memory in a qm. Reset of accelerator can
  * use this to clear queues.
  */
@@ -2806,12 +2867,12 @@ DEFINE_DEBUGFS_ATTRIBUTE(qm_atomic64_ops, qm_debugfs_atomic64_get,
  *
  * Create qm related debugfs files.
  */
-int hisi_qm_debug_init(struct hisi_qm *qm)
+void hisi_qm_debug_init(struct hisi_qm *qm)
 {
 	struct qm_dfx *dfx = &qm->debug.dfx;
 	struct dentry *qm_d;
 	void *data;
-	int i, ret;
+	int i;
 
 	qm_d = debugfs_create_dir("qm", qm->debug.debug_root);
 	qm->debug.qm_d = qm_d;
@@ -2819,10 +2880,7 @@ int hisi_qm_debug_init(struct hisi_qm *qm)
 	/* only show this in PF */
 	if (qm->fun_type == QM_HW_PF)
 		for (i = CURRENT_Q; i < DEBUG_FILE_NUM; i++)
-			if (qm_create_debugfs_file(qm, i)) {
-				ret = -ENOENT;
-				goto failed_to_create;
-			}
+			qm_create_debugfs_file(qm, i);
 
 	debugfs_create_file("regs", 0444, qm->debug.qm_d, qm, &qm_regs_fops);
 
@@ -2838,12 +2896,6 @@ int hisi_qm_debug_init(struct hisi_qm *qm)
 			data,
 			&qm_atomic64_ops);
 	}
-
-	return 0;
-
-failed_to_create:
-	debugfs_remove_recursive(qm_d);
-	return ret;
 }
 EXPORT_SYMBOL_GPL(hisi_qm_debug_init);
 
@@ -3273,7 +3325,7 @@ pci_ers_result_t hisi_qm_dev_err_detected(struct pci_dev *pdev,
 }
 EXPORT_SYMBOL_GPL(hisi_qm_dev_err_detected);
 
-static int qm_get_hw_error_status(struct hisi_qm *qm)
+static u32 qm_get_hw_error_status(struct hisi_qm *qm)
 {
 	return readl(qm->io_base + QM_ABNORMAL_INT_STATUS);
 }
@@ -3572,7 +3624,7 @@ restart_fail:
 	return ret;
 }
 
-static int qm_get_dev_err_status(struct hisi_qm *qm)
+static u32 qm_get_dev_err_status(struct hisi_qm *qm)
 {
 	return qm->err_ini->get_dev_hw_err_status(qm);
 }
@@ -3992,34 +4044,22 @@ void hisi_qm_alg_unregister(struct hisi_qm *qm, struct hisi_qm_list *qm_list)
 }
 EXPORT_SYMBOL_GPL(hisi_qm_alg_unregister);
 
-/**
- * hisi_qm_init() - Initialize configures about qm.
- * @qm: The qm needing init.
- *
- * This function init qm, then we can call hisi_qm_start to put qm into work.
- */
-int hisi_qm_init(struct hisi_qm *qm)
+static int hisi_qm_pci_init(struct hisi_qm *qm)
 {
 	struct pci_dev *pdev = qm->pdev;
 	struct device *dev = &pdev->dev;
 	unsigned int num_vec;
 	int ret;
 
-	hisi_qm_pre_init(qm);
-
-	ret = qm_alloc_uacce(qm);
-	if (ret < 0)
-		dev_warn(&pdev->dev, "fail to alloc uacce (%d)\n", ret);
-
 	ret = pci_enable_device_mem(pdev);
 	if (ret < 0) {
-		dev_err(&pdev->dev, "Failed to enable device mem!\n");
-		goto err_remove_uacce;
+		dev_err(dev, "Failed to enable device mem!\n");
+		return ret;
 	}
 
 	ret = pci_request_mem_regions(pdev, qm->dev_name);
 	if (ret < 0) {
-		dev_err(&pdev->dev, "Failed to request mem regions!\n");
+		dev_err(dev, "Failed to request mem regions!\n");
 		goto err_disable_pcidev;
 	}
 
@@ -4047,9 +4087,42 @@ int hisi_qm_init(struct hisi_qm *qm)
 		goto err_iounmap;
 	}
 
+	return 0;
+
+err_iounmap:
+	iounmap(qm->io_base);
+err_release_mem_regions:
+	pci_release_mem_regions(pdev);
+err_disable_pcidev:
+	pci_disable_device(pdev);
+	return ret;
+}
+
+/**
+ * hisi_qm_init() - Initialize configures about qm.
+ * @qm: The qm needing init.
+ *
+ * This function init qm, then we can call hisi_qm_start to put qm into work.
+ */
+int hisi_qm_init(struct hisi_qm *qm)
+{
+	struct pci_dev *pdev = qm->pdev;
+	struct device *dev = &pdev->dev;
+	int ret;
+
+	hisi_qm_pre_init(qm);
+
+	ret = qm_alloc_uacce(qm);
+	if (ret < 0)
+		dev_warn(dev, "fail to alloc uacce (%d)\n", ret);
+
+	ret = hisi_qm_pci_init(qm);
+	if (ret)
+		goto err_remove_uacce;
+
 	ret = qm_irq_register(qm);
 	if (ret)
-		goto err_free_irq_vectors;
+		goto err_pci_uninit;
 
 	if (qm->fun_type == QM_HW_VF && qm->ver != QM_HW_V1) {
 		/* v2 starts to support get vft by mailbox */
@@ -4072,14 +4145,8 @@ int hisi_qm_init(struct hisi_qm *qm)
 
 err_irq_unregister:
 	qm_irq_unregister(qm);
-err_free_irq_vectors:
-	pci_free_irq_vectors(pdev);
-err_iounmap:
-	iounmap(qm->io_base);
-err_release_mem_regions:
-	pci_release_mem_regions(pdev);
-err_disable_pcidev:
-	pci_disable_device(pdev);
+err_pci_uninit:
+	hisi_qm_pci_uninit(qm);
 err_remove_uacce:
 	uacce_remove(qm->uacce);
 	qm->uacce = NULL;
@@ -4087,7 +4154,6 @@ err_remove_uacce:
 }
 EXPORT_SYMBOL_GPL(hisi_qm_init);
 
-
 MODULE_LICENSE("GPL v2");
 MODULE_AUTHOR("Zhou Wang <wangzhou1@hisilicon.com>");
 MODULE_DESCRIPTION("HiSilicon Accelerator queue manager driver");
diff --git a/drivers/crypto/hisilicon/qm.h b/drivers/crypto/hisilicon/qm.h
index 0420f4ce7197..8624d1288afe 100644
--- a/drivers/crypto/hisilicon/qm.h
+++ b/drivers/crypto/hisilicon/qm.h
@@ -350,7 +350,7 @@ void hisi_qm_release_qp(struct hisi_qp *qp);
 int hisi_qp_send(struct hisi_qp *qp, const void *msg);
 int hisi_qm_get_free_qp_num(struct hisi_qm *qm);
 int hisi_qm_get_vft(struct hisi_qm *qm, u32 *base, u32 *number);
-int hisi_qm_debug_init(struct hisi_qm *qm);
+void hisi_qm_debug_init(struct hisi_qm *qm);
 enum qm_hw_ver hisi_qm_get_hw_version(struct pci_dev *pdev);
 void hisi_qm_debug_regs_clear(struct hisi_qm *qm);
 int hisi_qm_sriov_enable(struct pci_dev *pdev, int max_vfs);
diff --git a/drivers/crypto/hisilicon/sec2/sec.h b/drivers/crypto/hisilicon/sec2/sec.h
index 037762b531e2..08491912afd5 100644
--- a/drivers/crypto/hisilicon/sec2/sec.h
+++ b/drivers/crypto/hisilicon/sec2/sec.h
@@ -109,7 +109,6 @@ struct sec_qp_ctx {
 	struct list_head backlog;
 	struct hisi_acc_sgl_pool *c_in_pool;
 	struct hisi_acc_sgl_pool *c_out_pool;
-	atomic_t pending_reqs;
 };
 
 enum sec_alg_type {
@@ -180,7 +179,6 @@ struct sec_dev {
 	struct sec_debug debug;
 	u32 ctx_q_num;
 	bool iommu_used;
-	unsigned long status;
 };
 
 void sec_destroy_qps(struct hisi_qp **qps, int qp_num);
diff --git a/drivers/crypto/hisilicon/sec2/sec_crypto.c b/drivers/crypto/hisilicon/sec2/sec_crypto.c
index bb493423668c..2eaa516b3231 100644
--- a/drivers/crypto/hisilicon/sec2/sec_crypto.c
+++ b/drivers/crypto/hisilicon/sec2/sec_crypto.c
@@ -7,7 +7,8 @@
 #include <crypto/des.h>
 #include <crypto/hash.h>
 #include <crypto/internal/aead.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/skcipher.h>
 #include <crypto/xts.h>
 #include <linux/crypto.h>
@@ -101,6 +102,7 @@ static int sec_alloc_req_id(struct sec_req *req, struct sec_qp_ctx *qp_ctx)
 
 	req->qp_ctx = qp_ctx;
 	qp_ctx->req_list[req_id] = req;
+
 	return req_id;
 }
 
@@ -317,6 +319,7 @@ static int sec_alloc_pbuf_resource(struct device *dev, struct sec_alg_res *res)
 				j * SEC_PBUF_PKG + pbuf_page_offset;
 		}
 	}
+
 	return 0;
 }
 
@@ -345,12 +348,12 @@ static int sec_alg_resource_alloc(struct sec_ctx *ctx,
 	}
 
 	return 0;
+
 alloc_pbuf_fail:
 	if (ctx->alg_type == SEC_AEAD)
 		sec_free_mac_resource(dev, qp_ctx->res);
 alloc_fail:
 	sec_free_civ_resource(dev, res);
-
 	return ret;
 }
 
@@ -419,7 +422,6 @@ err_free_c_in_pool:
 	hisi_acc_free_sgl_pool(dev, qp_ctx->c_in_pool);
 err_destroy_idr:
 	idr_destroy(&qp_ctx->req_idr);
-
 	return ret;
 }
 
@@ -557,9 +559,9 @@ static int sec_skcipher_init(struct crypto_skcipher *tfm)
 		goto err_cipher_init;
 
 	return 0;
+
 err_cipher_init:
 	sec_ctx_base_uninit(ctx);
-
 	return ret;
 }
 
@@ -740,7 +742,6 @@ static void sec_cipher_pbuf_unmap(struct sec_ctx *ctx, struct sec_req *req,
 
 	if (unlikely(pbuf_length != copy_size))
 		dev_err(dev, "copy pbuf data to dst error!\n");
-
 }
 
 static int sec_cipher_map(struct sec_ctx *ctx, struct sec_req *req,
@@ -857,7 +858,7 @@ static int sec_aead_auth_set_key(struct sec_auth_ctx *ctx,
 				 struct crypto_authenc_keys *keys)
 {
 	struct crypto_shash *hash_tfm = ctx->hash_tfm;
-	int blocksize, ret;
+	int blocksize, digestsize, ret;
 
 	if (!keys->authkeylen) {
 		pr_err("hisi_sec2: aead auth key error!\n");
@@ -865,6 +866,7 @@ static int sec_aead_auth_set_key(struct sec_auth_ctx *ctx,
 	}
 
 	blocksize = crypto_shash_blocksize(hash_tfm);
+	digestsize = crypto_shash_digestsize(hash_tfm);
 	if (keys->authkeylen > blocksize) {
 		ret = crypto_shash_tfm_digest(hash_tfm, keys->authkey,
 					      keys->authkeylen, ctx->a_key);
@@ -872,7 +874,7 @@ static int sec_aead_auth_set_key(struct sec_auth_ctx *ctx,
 			pr_err("hisi_sec2: aead auth digest error!\n");
 			return -EINVAL;
 		}
-		ctx->a_key_len = blocksize;
+		ctx->a_key_len = digestsize;
 	} else {
 		memcpy(ctx->a_key, keys->authkey, keys->authkeylen);
 		ctx->a_key_len = keys->authkeylen;
@@ -913,9 +915,9 @@ static int sec_aead_setkey(struct crypto_aead *tfm, const u8 *key,
 	}
 
 	return 0;
+
 bad_key:
 	memzero_explicit(&keys, sizeof(struct crypto_authenc_keys));
-
 	return -EINVAL;
 }
 
@@ -966,7 +968,6 @@ static int sec_request_transfer(struct sec_ctx *ctx, struct sec_req *req)
 
 unmap_req_buf:
 	ctx->req_op->buf_unmap(ctx, req);
-
 	return ret;
 }
 
@@ -1107,7 +1108,6 @@ static void sec_skcipher_callback(struct sec_ctx *ctx, struct sec_req *req,
 		atomic64_inc(&ctx->sec->debug.dfx.recv_busy_cnt);
 	}
 
-
 	sk_req->base.complete(&sk_req->base, err);
 }
 
@@ -1279,7 +1279,6 @@ err_send_req:
 	sec_request_untransfer(ctx, req);
 err_uninit_req:
 	sec_request_uninit(ctx, req);
-
 	return ret;
 }
 
@@ -1349,7 +1348,6 @@ err_cipher_init:
 	sec_auth_uninit(ctx);
 err_auth_init:
 	sec_ctx_base_uninit(ctx);
-
 	return ret;
 }
 
@@ -1437,8 +1435,8 @@ static int sec_skcipher_param_check(struct sec_ctx *ctx, struct sec_req *sreq)
 		}
 		return 0;
 	}
-
 	dev_err(dev, "skcipher algorithm error!\n");
+
 	return -EINVAL;
 }
 
@@ -1554,7 +1552,6 @@ static int sec_aead_param_check(struct sec_ctx *ctx, struct sec_req *sreq)
 	if (unlikely(c_alg != SEC_CALG_AES)) {
 		dev_err(SEC_CTX_DEV(ctx), "aead crypto alg error!\n");
 		return -EINVAL;
-
 	}
 	if (sreq->c_req.encrypt)
 		sreq->c_req.c_len = req->cryptlen;
diff --git a/drivers/crypto/hisilicon/sec2/sec_main.c b/drivers/crypto/hisilicon/sec2/sec_main.c
index 548896394c4b..b35c1c2271a3 100644
--- a/drivers/crypto/hisilicon/sec2/sec_main.c
+++ b/drivers/crypto/hisilicon/sec2/sec_main.c
@@ -652,20 +652,16 @@ static int sec_debugfs_init(struct hisi_qm *qm)
 						  sec_debugfs_root);
 	qm->debug.sqe_mask_offset = SEC_SQE_MASK_OFFSET;
 	qm->debug.sqe_mask_len = SEC_SQE_MASK_LEN;
-	ret = hisi_qm_debug_init(qm);
-	if (ret)
-		goto failed_to_create;
+	hisi_qm_debug_init(qm);
 
 	ret = sec_debug_init(qm);
 	if (ret)
 		goto failed_to_create;
 
-
 	return 0;
 
 failed_to_create:
 	debugfs_remove_recursive(sec_debugfs_root);
-
 	return ret;
 }
 
@@ -683,13 +679,13 @@ static void sec_log_hw_error(struct hisi_qm *qm, u32 err_sts)
 	while (errs->msg) {
 		if (errs->int_msk & err_sts) {
 			dev_err(dev, "%s [error status=0x%x] found\n",
-				errs->msg, errs->int_msk);
+					errs->msg, errs->int_msk);
 
 			if (SEC_CORE_INT_STATUS_M_ECC & errs->int_msk) {
 				err_val = readl(qm->io_base +
 						SEC_CORE_SRAM_ECC_ERR_INFO);
 				dev_err(dev, "multi ecc sram num=0x%x\n",
-					SEC_ECC_NUM(err_val));
+						SEC_ECC_NUM(err_val));
 			}
 		}
 		errs++;
@@ -724,13 +720,13 @@ static const struct hisi_qm_err_ini sec_err_ini = {
 	.log_dev_hw_err		= sec_log_hw_error,
 	.open_axi_master_ooo	= sec_open_axi_master_ooo,
 	.err_info		= {
-		.ce			= QM_BASE_CE,
-		.nfe			= QM_BASE_NFE | QM_ACC_DO_TASK_TIMEOUT |
-					  QM_ACC_WB_NOT_READY_TIMEOUT,
-		.fe			= 0,
-		.ecc_2bits_mask		= SEC_CORE_INT_STATUS_M_ECC,
-		.msi_wr_port		= BIT(0),
-		.acpi_rst		= "SRST",
+		.ce		= QM_BASE_CE,
+		.nfe		= QM_BASE_NFE | QM_ACC_DO_TASK_TIMEOUT |
+				  QM_ACC_WB_NOT_READY_TIMEOUT,
+		.fe		= 0,
+		.ecc_2bits_mask	= SEC_CORE_INT_STATUS_M_ECC,
+		.msi_wr_port	= BIT(0),
+		.acpi_rst	= "SRST",
 	}
 };
 
@@ -899,17 +895,13 @@ static int sec_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 err_alg_unregister:
 	hisi_qm_alg_unregister(qm, &sec_devices);
-
 err_qm_stop:
 	sec_debugfs_exit(qm);
 	hisi_qm_stop(qm, QM_NORMAL);
-
 err_probe_uninit:
 	sec_probe_uninit(qm);
-
 err_qm_uninit:
 	sec_qm_uninit(qm);
-
 	return ret;
 }
 
@@ -936,9 +928,9 @@ static void sec_remove(struct pci_dev *pdev)
 
 static const struct pci_error_handlers sec_err_handler = {
 	.error_detected = hisi_qm_dev_err_detected,
-	.slot_reset =  hisi_qm_dev_slot_reset,
-	.reset_prepare		= hisi_qm_reset_prepare,
-	.reset_done		= hisi_qm_reset_done,
+	.slot_reset	= hisi_qm_dev_slot_reset,
+	.reset_prepare	= hisi_qm_reset_prepare,
+	.reset_done	= hisi_qm_reset_done,
 };
 
 static struct pci_driver sec_pci_driver = {
diff --git a/drivers/crypto/hisilicon/sgl.c b/drivers/crypto/hisilicon/sgl.c
index 725a739800b0..3bff6394acaf 100644
--- a/drivers/crypto/hisilicon/sgl.c
+++ b/drivers/crypto/hisilicon/sgl.c
@@ -246,8 +246,6 @@ EXPORT_SYMBOL_GPL(hisi_acc_sg_buf_map_to_hw_sgl);
  * @dev: The device which hw sgl belongs to.
  * @sgl: Related scatterlist.
  * @hw_sgl: Virtual address of hw sgl.
- * @hw_sgl_dma: DMA address of hw sgl.
- * @pool: Pool which hw sgl is allocated in.
  *
  * This function unmaps allocated hw sgl.
  */
diff --git a/drivers/crypto/hisilicon/trng/Makefile b/drivers/crypto/hisilicon/trng/Makefile
new file mode 100644
index 000000000000..d909079f351c
--- /dev/null
+++ b/drivers/crypto/hisilicon/trng/Makefile
@@ -0,0 +1,2 @@
+obj-$(CONFIG_CRYPTO_DEV_HISI_TRNG) += hisi-trng-v2.o
+hisi-trng-v2-objs = trng.o
diff --git a/drivers/crypto/hisilicon/trng/trng.c b/drivers/crypto/hisilicon/trng/trng.c
new file mode 100644
index 000000000000..29712685498a
--- /dev/null
+++ b/drivers/crypto/hisilicon/trng/trng.c
@@ -0,0 +1,334 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2019 HiSilicon Limited. */
+
+#include <linux/acpi.h>
+#include <linux/crypto.h>
+#include <linux/err.h>
+#include <linux/hw_random.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/platform_device.h>
+#include <linux/random.h>
+#include <crypto/internal/rng.h>
+
+#define HISI_TRNG_REG		0x00F0
+#define HISI_TRNG_BYTES		4
+#define HISI_TRNG_QUALITY	512
+#define SLEEP_US		10
+#define TIMEOUT_US		10000
+#define SW_DRBG_NUM_SHIFT	2
+#define SW_DRBG_KEY_BASE	0x082C
+#define SW_DRBG_SEED(n)         (SW_DRBG_KEY_BASE - ((n) << SW_DRBG_NUM_SHIFT))
+#define SW_DRBG_SEED_REGS_NUM	12
+#define SW_DRBG_SEED_SIZE	48
+#define SW_DRBG_BLOCKS		0x0830
+#define SW_DRBG_INIT		0x0834
+#define SW_DRBG_GEN		0x083c
+#define SW_DRBG_STATUS		0x0840
+#define SW_DRBG_BLOCKS_NUM	4095
+#define SW_DRBG_DATA_BASE	0x0850
+#define SW_DRBG_DATA_NUM	4
+#define SW_DRBG_DATA(n)		(SW_DRBG_DATA_BASE - ((n) << SW_DRBG_NUM_SHIFT))
+#define SW_DRBG_BYTES		16
+#define SW_DRBG_ENABLE_SHIFT	12
+#define SEED_SHIFT_24		24
+#define SEED_SHIFT_16		16
+#define SEED_SHIFT_8		8
+
+struct hisi_trng_list {
+	struct mutex lock;
+	struct list_head list;
+	bool is_init;
+};
+
+struct hisi_trng {
+	void __iomem *base;
+	struct hisi_trng_list *trng_list;
+	struct list_head list;
+	struct hwrng rng;
+	bool is_used;
+	struct mutex mutex;
+};
+
+struct hisi_trng_ctx {
+	struct hisi_trng *trng;
+};
+
+static atomic_t trng_active_devs;
+static struct hisi_trng_list trng_devices;
+
+static void hisi_trng_set_seed(struct hisi_trng *trng, const u8 *seed)
+{
+	u32 val, seed_reg, i;
+
+	for (i = 0; i < SW_DRBG_SEED_SIZE;
+	     i += SW_DRBG_SEED_SIZE / SW_DRBG_SEED_REGS_NUM) {
+		val = seed[i] << SEED_SHIFT_24;
+		val |= seed[i + 1UL] << SEED_SHIFT_16;
+		val |= seed[i + 2UL] << SEED_SHIFT_8;
+		val |= seed[i + 3UL];
+
+		seed_reg = (i >> SW_DRBG_NUM_SHIFT) % SW_DRBG_SEED_REGS_NUM;
+		writel(val, trng->base + SW_DRBG_SEED(seed_reg));
+	}
+}
+
+static int hisi_trng_seed(struct crypto_rng *tfm, const u8 *seed,
+			  unsigned int slen)
+{
+	struct hisi_trng_ctx *ctx = crypto_rng_ctx(tfm);
+	struct hisi_trng *trng = ctx->trng;
+	u32 val = 0;
+	int ret = 0;
+
+	if (slen < SW_DRBG_SEED_SIZE) {
+		pr_err("slen(%u) is not matched with trng(%d)\n", slen,
+			SW_DRBG_SEED_SIZE);
+		return -EINVAL;
+	}
+
+	writel(0x0, trng->base + SW_DRBG_BLOCKS);
+	hisi_trng_set_seed(trng, seed);
+
+	writel(SW_DRBG_BLOCKS_NUM | (0x1 << SW_DRBG_ENABLE_SHIFT),
+	       trng->base + SW_DRBG_BLOCKS);
+	writel(0x1, trng->base + SW_DRBG_INIT);
+
+	ret = readl_relaxed_poll_timeout(trng->base + SW_DRBG_STATUS,
+					val, val & BIT(0), SLEEP_US, TIMEOUT_US);
+	if (ret)
+		pr_err("fail to init trng(%d)\n", ret);
+
+	return ret;
+}
+
+static int hisi_trng_generate(struct crypto_rng *tfm, const u8 *src,
+			      unsigned int slen, u8 *dstn, unsigned int dlen)
+{
+	struct hisi_trng_ctx *ctx = crypto_rng_ctx(tfm);
+	struct hisi_trng *trng = ctx->trng;
+	u32 data[SW_DRBG_DATA_NUM];
+	u32 currsize = 0;
+	u32 val = 0;
+	int ret;
+	u32 i;
+
+	if (dlen > SW_DRBG_BLOCKS_NUM * SW_DRBG_BYTES || dlen == 0) {
+		pr_err("dlen(%d) exceeds limit(%d)!\n", dlen,
+			SW_DRBG_BLOCKS_NUM * SW_DRBG_BYTES);
+		return -EINVAL;
+	}
+
+	do {
+		ret = readl_relaxed_poll_timeout(trng->base + SW_DRBG_STATUS,
+		     val, val & BIT(1), SLEEP_US, TIMEOUT_US);
+		if (ret) {
+			pr_err("fail to generate random number(%d)!\n", ret);
+			break;
+		}
+
+		for (i = 0; i < SW_DRBG_DATA_NUM; i++)
+			data[i] = readl(trng->base + SW_DRBG_DATA(i));
+
+		if (dlen - currsize >= SW_DRBG_BYTES) {
+			memcpy(dstn + currsize, data, SW_DRBG_BYTES);
+			currsize += SW_DRBG_BYTES;
+		} else {
+			memcpy(dstn + currsize, data, dlen - currsize);
+			currsize = dlen;
+		}
+
+		writel(0x1, trng->base + SW_DRBG_GEN);
+	} while (currsize < dlen);
+
+	return ret;
+}
+
+static int hisi_trng_init(struct crypto_tfm *tfm)
+{
+	struct hisi_trng_ctx *ctx = crypto_tfm_ctx(tfm);
+	struct hisi_trng *trng;
+	int ret = -EBUSY;
+
+	mutex_lock(&trng_devices.lock);
+	list_for_each_entry(trng, &trng_devices.list, list) {
+		if (!trng->is_used) {
+			trng->is_used = true;
+			ctx->trng = trng;
+			ret = 0;
+			break;
+		}
+	}
+	mutex_unlock(&trng_devices.lock);
+
+	return ret;
+}
+
+static void hisi_trng_exit(struct crypto_tfm *tfm)
+{
+	struct hisi_trng_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	mutex_lock(&trng_devices.lock);
+	ctx->trng->is_used = false;
+	mutex_unlock(&trng_devices.lock);
+}
+
+static int hisi_trng_read(struct hwrng *rng, void *buf, size_t max, bool wait)
+{
+	struct hisi_trng *trng;
+	int currsize = 0;
+	u32 val = 0;
+	u32 ret;
+
+	trng = container_of(rng, struct hisi_trng, rng);
+
+	do {
+		ret = readl_poll_timeout(trng->base + HISI_TRNG_REG, val,
+					 val, SLEEP_US, TIMEOUT_US);
+		if (ret)
+			return currsize;
+
+		if (max - currsize >= HISI_TRNG_BYTES) {
+			memcpy(buf + currsize, &val, HISI_TRNG_BYTES);
+			currsize += HISI_TRNG_BYTES;
+			if (currsize == max)
+				return currsize;
+			continue;
+		}
+
+		/* copy remaining bytes */
+		memcpy(buf + currsize, &val, max - currsize);
+		currsize = max;
+	} while (currsize < max);
+
+	return currsize;
+}
+
+static struct rng_alg hisi_trng_alg = {
+	.generate = hisi_trng_generate,
+	.seed =	hisi_trng_seed,
+	.seedsize = SW_DRBG_SEED_SIZE,
+	.base = {
+		.cra_name = "stdrng",
+		.cra_driver_name = "hisi_stdrng",
+		.cra_priority = 300,
+		.cra_ctxsize = sizeof(struct hisi_trng_ctx),
+		.cra_module = THIS_MODULE,
+		.cra_init = hisi_trng_init,
+		.cra_exit = hisi_trng_exit,
+	},
+};
+
+static void hisi_trng_add_to_list(struct hisi_trng *trng)
+{
+	mutex_lock(&trng_devices.lock);
+	list_add_tail(&trng->list, &trng_devices.list);
+	mutex_unlock(&trng_devices.lock);
+}
+
+static int hisi_trng_del_from_list(struct hisi_trng *trng)
+{
+	int ret = -EBUSY;
+
+	mutex_lock(&trng_devices.lock);
+	if (!trng->is_used) {
+		list_del(&trng->list);
+		ret = 0;
+	}
+	mutex_unlock(&trng_devices.lock);
+
+	return ret;
+}
+
+static int hisi_trng_probe(struct platform_device *pdev)
+{
+	struct hisi_trng *trng;
+	int ret;
+
+	trng = devm_kzalloc(&pdev->dev, sizeof(*trng), GFP_KERNEL);
+	if (!trng)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, trng);
+
+	trng->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(trng->base))
+		return PTR_ERR(trng->base);
+
+	trng->is_used = false;
+	if (!trng_devices.is_init) {
+		INIT_LIST_HEAD(&trng_devices.list);
+		mutex_init(&trng_devices.lock);
+		trng_devices.is_init = true;
+	}
+
+	hisi_trng_add_to_list(trng);
+	if (atomic_inc_return(&trng_active_devs) == 1) {
+		ret = crypto_register_rng(&hisi_trng_alg);
+		if (ret) {
+			dev_err(&pdev->dev,
+				"failed to register crypto(%d)\n", ret);
+			atomic_dec_return(&trng_active_devs);
+			goto err_remove_from_list;
+		}
+	}
+
+	trng->rng.name = pdev->name;
+	trng->rng.read = hisi_trng_read;
+	trng->rng.quality = HISI_TRNG_QUALITY;
+	ret = devm_hwrng_register(&pdev->dev, &trng->rng);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to register hwrng: %d!\n", ret);
+		goto err_crypto_unregister;
+	}
+
+	return ret;
+
+err_crypto_unregister:
+	if (atomic_dec_return(&trng_active_devs) == 0)
+		crypto_unregister_rng(&hisi_trng_alg);
+
+err_remove_from_list:
+	hisi_trng_del_from_list(trng);
+	return ret;
+}
+
+static int hisi_trng_remove(struct platform_device *pdev)
+{
+	struct hisi_trng *trng = platform_get_drvdata(pdev);
+
+	/* Wait until the task is finished */
+	while (hisi_trng_del_from_list(trng))
+		;
+
+	if (atomic_dec_return(&trng_active_devs) == 0)
+		crypto_unregister_rng(&hisi_trng_alg);
+
+	return 0;
+}
+
+static const struct acpi_device_id hisi_trng_acpi_match[] = {
+	{ "HISI02B3", 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(acpi, hisi_trng_acpi_match);
+
+static struct platform_driver hisi_trng_driver = {
+	.probe		= hisi_trng_probe,
+	.remove         = hisi_trng_remove,
+	.driver		= {
+		.name	= "hisi-trng-v2",
+		.acpi_match_table = ACPI_PTR(hisi_trng_acpi_match),
+	},
+};
+
+module_platform_driver(hisi_trng_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Weili Qian <qianweili@huawei.com>");
+MODULE_AUTHOR("Zaibo Xu <xuzaibo@huawei.com>");
+MODULE_DESCRIPTION("HiSilicon true random number generator V2 driver");
diff --git a/drivers/crypto/hisilicon/zip/zip_main.c b/drivers/crypto/hisilicon/zip/zip_main.c
index 4bd2c811abba..4fb5a32bf830 100644
--- a/drivers/crypto/hisilicon/zip/zip_main.c
+++ b/drivers/crypto/hisilicon/zip/zip_main.c
@@ -590,9 +590,7 @@ static int hisi_zip_debugfs_init(struct hisi_qm *qm)
 	qm->debug.sqe_mask_offset = HZIP_SQE_MASK_OFFSET;
 	qm->debug.sqe_mask_len = HZIP_SQE_MASK_LEN;
 	qm->debug.debug_root = dev_d;
-	ret = hisi_qm_debug_init(qm);
-	if (ret)
-		goto failed_to_create;
+	hisi_qm_debug_init(qm);
 
 	if (qm->fun_type == QM_HW_PF) {
 		ret = hisi_zip_ctrl_debug_init(qm);
@@ -749,6 +747,8 @@ static int hisi_zip_pf_probe_init(struct hisi_zip *hisi_zip)
 
 static int hisi_zip_qm_init(struct hisi_qm *qm, struct pci_dev *pdev)
 {
+	int ret;
+
 	qm->pdev = pdev;
 	qm->ver = pdev->revision;
 	qm->algs = "zlib\ngzip";
@@ -774,7 +774,25 @@ static int hisi_zip_qm_init(struct hisi_qm *qm, struct pci_dev *pdev)
 		qm->qp_num = HZIP_QUEUE_NUM_V1 - HZIP_PF_DEF_Q_NUM;
 	}
 
-	return hisi_qm_init(qm);
+	qm->wq = alloc_workqueue("%s", WQ_HIGHPRI | WQ_MEM_RECLAIM |
+				 WQ_UNBOUND, num_online_cpus(),
+				 pci_name(qm->pdev));
+	if (!qm->wq) {
+		pci_err(qm->pdev, "fail to alloc workqueue\n");
+		return -ENOMEM;
+	}
+
+	ret = hisi_qm_init(qm);
+	if (ret)
+		destroy_workqueue(qm->wq);
+
+	return ret;
+}
+
+static void hisi_zip_qm_uninit(struct hisi_qm *qm)
+{
+	hisi_qm_uninit(qm);
+	destroy_workqueue(qm->wq);
 }
 
 static int hisi_zip_probe_init(struct hisi_zip *hisi_zip)
@@ -856,7 +874,7 @@ err_dev_err_uninit:
 	hisi_qm_dev_err_uninit(qm);
 
 err_qm_uninit:
-	hisi_qm_uninit(qm);
+	hisi_zip_qm_uninit(qm);
 
 	return ret;
 }
@@ -874,7 +892,7 @@ static void hisi_zip_remove(struct pci_dev *pdev)
 	hisi_zip_debugfs_exit(qm);
 	hisi_qm_stop(qm, QM_NORMAL);
 	hisi_qm_dev_err_uninit(qm);
-	hisi_qm_uninit(qm);
+	hisi_zip_qm_uninit(qm);
 }
 
 static const struct pci_error_handlers hisi_zip_err_handler = {
diff --git a/drivers/crypto/img-hash.c b/drivers/crypto/img-hash.c
index 91f555ccbb31..e813115d5432 100644
--- a/drivers/crypto/img-hash.c
+++ b/drivers/crypto/img-hash.c
@@ -19,7 +19,8 @@
 
 #include <crypto/internal/hash.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #define CR_RESET			0
 #define CR_RESET_SET			1
diff --git a/drivers/crypto/inside-secure/safexcel.c b/drivers/crypto/inside-secure/safexcel.c
index eb2418450f12..2e1562108a85 100644
--- a/drivers/crypto/inside-secure/safexcel.c
+++ b/drivers/crypto/inside-secure/safexcel.c
@@ -1639,7 +1639,7 @@ static int safexcel_probe_generic(void *pdev,
 
 		priv->ring[i].rdr_req = devm_kcalloc(dev,
 			EIP197_DEFAULT_RING_SIZE,
-			sizeof(priv->ring[i].rdr_req),
+			sizeof(*priv->ring[i].rdr_req),
 			GFP_KERNEL);
 		if (!priv->ring[i].rdr_req)
 			return -ENOMEM;
diff --git a/drivers/crypto/inside-secure/safexcel.h b/drivers/crypto/inside-secure/safexcel.h
index 9045f2d7f4c6..ce1e611a163e 100644
--- a/drivers/crypto/inside-secure/safexcel.h
+++ b/drivers/crypto/inside-secure/safexcel.h
@@ -11,7 +11,8 @@
 #include <crypto/aead.h>
 #include <crypto/algapi.h>
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/sha3.h>
 #include <crypto/skcipher.h>
 #include <linux/types.h>
diff --git a/drivers/crypto/inside-secure/safexcel_cipher.c b/drivers/crypto/inside-secure/safexcel_cipher.c
index 9bcfb79a030f..d68ef16650d4 100644
--- a/drivers/crypto/inside-secure/safexcel_cipher.c
+++ b/drivers/crypto/inside-secure/safexcel_cipher.c
@@ -18,7 +18,8 @@
 #include <crypto/gcm.h>
 #include <crypto/ghash.h>
 #include <crypto/poly1305.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/sm3.h>
 #include <crypto/sm4.h>
 #include <crypto/xts.h>
diff --git a/drivers/crypto/inside-secure/safexcel_hash.c b/drivers/crypto/inside-secure/safexcel_hash.c
index 56d5ccb5cc00..50fb6d90a2e0 100644
--- a/drivers/crypto/inside-secure/safexcel_hash.c
+++ b/drivers/crypto/inside-secure/safexcel_hash.c
@@ -8,7 +8,8 @@
 #include <crypto/aes.h>
 #include <crypto/hmac.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/sha3.h>
 #include <crypto/skcipher.h>
 #include <crypto/sm3.h>
diff --git a/drivers/crypto/ixp4xx_crypto.c b/drivers/crypto/ixp4xx_crypto.c
index 276012e7c482..8b0f17fc09fb 100644
--- a/drivers/crypto/ixp4xx_crypto.c
+++ b/drivers/crypto/ixp4xx_crypto.c
@@ -20,7 +20,7 @@
 #include <crypto/internal/des.h>
 #include <crypto/aes.h>
 #include <crypto/hmac.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
 #include <crypto/algapi.h>
 #include <crypto/internal/aead.h>
 #include <crypto/internal/skcipher.h>
diff --git a/drivers/crypto/keembay/Kconfig b/drivers/crypto/keembay/Kconfig
new file mode 100644
index 000000000000..3c16797b25b9
--- /dev/null
+++ b/drivers/crypto/keembay/Kconfig
@@ -0,0 +1,39 @@
+config CRYPTO_DEV_KEEMBAY_OCS_AES_SM4
+	tristate "Support for Intel Keem Bay OCS AES/SM4 HW acceleration"
+	depends on OF || COMPILE_TEST
+	select CRYPTO_SKCIPHER
+	select CRYPTO_AEAD
+	select CRYPTO_ENGINE
+	help
+	  Support for Intel Keem Bay Offload and Crypto Subsystem (OCS) AES and
+	  SM4 cihper hardware acceleration for use with Crypto API.
+
+	  Provides HW acceleration for the following transformations:
+	  cbc(aes), ctr(aes), ccm(aes), gcm(aes), cbc(sm4), ctr(sm4), ccm(sm4)
+	  and gcm(sm4).
+
+	  Optionally, support for the following transformations can also be
+	  enabled: ecb(aes), cts(cbc(aes)), ecb(sm4) and cts(cbc(sm4)).
+
+config CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB
+	bool "Support for Intel Keem Bay OCS AES/SM4 ECB HW acceleration"
+	depends on CRYPTO_DEV_KEEMBAY_OCS_AES_SM4
+	help
+	  Support for Intel Keem Bay Offload and Crypto Subsystem (OCS)
+	  AES/SM4 ECB mode hardware acceleration for use with Crypto API.
+
+	  Provides OCS version of ecb(aes) and ecb(sm4)
+
+	  Intel does not recommend use of ECB mode with AES/SM4.
+
+config CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS
+	bool "Support for Intel Keem Bay OCS AES/SM4 CTS HW acceleration"
+	depends on CRYPTO_DEV_KEEMBAY_OCS_AES_SM4
+	help
+	  Support for Intel Keem Bay Offload and Crypto Subsystem (OCS)
+	  AES/SM4 CBC with CTS mode hardware acceleration for use with
+	  Crypto API.
+
+	  Provides OCS version of cts(cbc(aes)) and cts(cbc(sm4)).
+
+	  Intel does not recommend use of CTS mode with AES/SM4.
diff --git a/drivers/crypto/keembay/Makefile b/drivers/crypto/keembay/Makefile
new file mode 100644
index 000000000000..f21e2c4ab3b3
--- /dev/null
+++ b/drivers/crypto/keembay/Makefile
@@ -0,0 +1,5 @@
+#
+# Makefile for Intel Keem Bay OCS Crypto API Linux drivers
+#
+obj-$(CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4) += keembay-ocs-aes.o
+keembay-ocs-aes-objs := keembay-ocs-aes-core.o ocs-aes.o
diff --git a/drivers/crypto/keembay/keembay-ocs-aes-core.c b/drivers/crypto/keembay/keembay-ocs-aes-core.c
new file mode 100644
index 000000000000..b6b25d994af3
--- /dev/null
+++ b/drivers/crypto/keembay/keembay-ocs-aes-core.c
@@ -0,0 +1,1713 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Intel Keem Bay OCS AES Crypto Driver.
+ *
+ * Copyright (C) 2018-2020 Intel Corporation
+ */
+
+#include <linux/clk.h>
+#include <linux/completion.h>
+#include <linux/crypto.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/types.h>
+
+#include <crypto/aes.h>
+#include <crypto/engine.h>
+#include <crypto/gcm.h>
+#include <crypto/scatterwalk.h>
+
+#include <crypto/internal/aead.h>
+#include <crypto/internal/skcipher.h>
+
+#include "ocs-aes.h"
+
+#define KMB_OCS_PRIORITY	350
+#define DRV_NAME		"keembay-ocs-aes"
+
+#define OCS_AES_MIN_KEY_SIZE	16
+#define OCS_AES_MAX_KEY_SIZE	32
+#define OCS_AES_KEYSIZE_128	16
+#define OCS_AES_KEYSIZE_192	24
+#define OCS_AES_KEYSIZE_256	32
+#define OCS_SM4_KEY_SIZE	16
+
+/**
+ * struct ocs_aes_tctx - OCS AES Transform context
+ * @engine_ctx:		Engine context.
+ * @aes_dev:		The OCS AES device.
+ * @key:		AES/SM4 key.
+ * @key_len:		The length (in bytes) of @key.
+ * @cipher:		OCS cipher to use (either AES or SM4).
+ * @sw_cipher:		The cipher to use as fallback.
+ * @use_fallback:	Whether or not fallback cipher should be used.
+ */
+struct ocs_aes_tctx {
+	struct crypto_engine_ctx engine_ctx;
+	struct ocs_aes_dev *aes_dev;
+	u8 key[OCS_AES_KEYSIZE_256];
+	unsigned int key_len;
+	enum ocs_cipher cipher;
+	union {
+		struct crypto_sync_skcipher *sk;
+		struct crypto_aead *aead;
+	} sw_cipher;
+	bool use_fallback;
+};
+
+/**
+ * struct ocs_aes_rctx - OCS AES Request context.
+ * @instruction:	Instruction to be executed (encrypt / decrypt).
+ * @mode:		Mode to use (ECB, CBC, CTR, CCm, GCM, CTS)
+ * @src_nents:		Number of source SG entries.
+ * @dst_nents:		Number of destination SG entries.
+ * @src_dma_count:	The number of DMA-mapped entries of the source SG.
+ * @dst_dma_count:	The number of DMA-mapped entries of the destination SG.
+ * @in_place:		Whether or not this is an in place request, i.e.,
+ *			src_sg == dst_sg.
+ * @src_dll:		OCS DMA linked list for input data.
+ * @dst_dll:		OCS DMA linked list for output data.
+ * @last_ct_blk:	Buffer to hold last cipher text block (only used in CBC
+ *			mode).
+ * @cts_swap:		Whether or not CTS swap must be performed.
+ * @aad_src_dll:	OCS DMA linked list for input AAD data.
+ * @aad_dst_dll:	OCS DMA linked list for output AAD data.
+ * @in_tag:		Buffer to hold input encrypted tag (only used for
+ *			CCM/GCM decrypt).
+ * @out_tag:		Buffer to hold output encrypted / decrypted tag (only
+ *			used for GCM encrypt / decrypt).
+ */
+struct ocs_aes_rctx {
+	/* Fields common across all modes. */
+	enum ocs_instruction	instruction;
+	enum ocs_mode		mode;
+	int			src_nents;
+	int			dst_nents;
+	int			src_dma_count;
+	int			dst_dma_count;
+	bool			in_place;
+	struct ocs_dll_desc	src_dll;
+	struct ocs_dll_desc	dst_dll;
+
+	/* CBC specific */
+	u8			last_ct_blk[AES_BLOCK_SIZE];
+
+	/* CTS specific */
+	int			cts_swap;
+
+	/* CCM/GCM specific */
+	struct ocs_dll_desc	aad_src_dll;
+	struct ocs_dll_desc	aad_dst_dll;
+	u8			in_tag[AES_BLOCK_SIZE];
+
+	/* GCM specific */
+	u8			out_tag[AES_BLOCK_SIZE];
+};
+
+/* Driver data. */
+struct ocs_aes_drv {
+	struct list_head dev_list;
+	spinlock_t lock;	/* Protects dev_list. */
+};
+
+static struct ocs_aes_drv ocs_aes = {
+	.dev_list = LIST_HEAD_INIT(ocs_aes.dev_list),
+	.lock = __SPIN_LOCK_UNLOCKED(ocs_aes.lock),
+};
+
+static struct ocs_aes_dev *kmb_ocs_aes_find_dev(struct ocs_aes_tctx *tctx)
+{
+	struct ocs_aes_dev *aes_dev;
+
+	spin_lock(&ocs_aes.lock);
+
+	if (tctx->aes_dev) {
+		aes_dev = tctx->aes_dev;
+		goto exit;
+	}
+
+	/* Only a single OCS device available */
+	aes_dev = list_first_entry(&ocs_aes.dev_list, struct ocs_aes_dev, list);
+	tctx->aes_dev = aes_dev;
+
+exit:
+	spin_unlock(&ocs_aes.lock);
+
+	return aes_dev;
+}
+
+/*
+ * Ensure key is 128-bit or 256-bit for AES or 128-bit for SM4 and an actual
+ * key is being passed in.
+ *
+ * Return: 0 if key is valid, -EINVAL otherwise.
+ */
+static int check_key(const u8 *in_key, size_t key_len, enum ocs_cipher cipher)
+{
+	if (!in_key)
+		return -EINVAL;
+
+	/* For AES, only 128-byte or 256-byte keys are supported. */
+	if (cipher == OCS_AES && (key_len == OCS_AES_KEYSIZE_128 ||
+				  key_len == OCS_AES_KEYSIZE_256))
+		return 0;
+
+	/* For SM4, only 128-byte keys are supported. */
+	if (cipher == OCS_SM4 && key_len == OCS_AES_KEYSIZE_128)
+		return 0;
+
+	/* Everything else is unsupported. */
+	return -EINVAL;
+}
+
+/* Save key into transformation context. */
+static int save_key(struct ocs_aes_tctx *tctx, const u8 *in_key, size_t key_len,
+		    enum ocs_cipher cipher)
+{
+	int ret;
+
+	ret = check_key(in_key, key_len, cipher);
+	if (ret)
+		return ret;
+
+	memcpy(tctx->key, in_key, key_len);
+	tctx->key_len = key_len;
+	tctx->cipher = cipher;
+
+	return 0;
+}
+
+/* Set key for symmetric cypher. */
+static int kmb_ocs_sk_set_key(struct crypto_skcipher *tfm, const u8 *in_key,
+			      size_t key_len, enum ocs_cipher cipher)
+{
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+
+	/* Fallback is used for AES with 192-bit key. */
+	tctx->use_fallback = (cipher == OCS_AES &&
+			      key_len == OCS_AES_KEYSIZE_192);
+
+	if (!tctx->use_fallback)
+		return save_key(tctx, in_key, key_len, cipher);
+
+	crypto_sync_skcipher_clear_flags(tctx->sw_cipher.sk,
+					 CRYPTO_TFM_REQ_MASK);
+	crypto_sync_skcipher_set_flags(tctx->sw_cipher.sk,
+				       tfm->base.crt_flags &
+				       CRYPTO_TFM_REQ_MASK);
+
+	return crypto_sync_skcipher_setkey(tctx->sw_cipher.sk, in_key, key_len);
+}
+
+/* Set key for AEAD cipher. */
+static int kmb_ocs_aead_set_key(struct crypto_aead *tfm, const u8 *in_key,
+				size_t key_len, enum ocs_cipher cipher)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(tfm);
+
+	/* Fallback is used for AES with 192-bit key. */
+	tctx->use_fallback = (cipher == OCS_AES &&
+			      key_len == OCS_AES_KEYSIZE_192);
+
+	if (!tctx->use_fallback)
+		return save_key(tctx, in_key, key_len, cipher);
+
+	crypto_aead_clear_flags(tctx->sw_cipher.aead, CRYPTO_TFM_REQ_MASK);
+	crypto_aead_set_flags(tctx->sw_cipher.aead,
+			      crypto_aead_get_flags(tfm) & CRYPTO_TFM_REQ_MASK);
+
+	return crypto_aead_setkey(tctx->sw_cipher.aead, in_key, key_len);
+}
+
+/* Swap two AES blocks in SG lists. */
+static void sg_swap_blocks(struct scatterlist *sgl, unsigned int nents,
+			   off_t blk1_offset, off_t blk2_offset)
+{
+	u8 tmp_buf1[AES_BLOCK_SIZE], tmp_buf2[AES_BLOCK_SIZE];
+
+	/*
+	 * No easy way to copy within sg list, so copy both blocks to temporary
+	 * buffers first.
+	 */
+	sg_pcopy_to_buffer(sgl, nents, tmp_buf1, AES_BLOCK_SIZE, blk1_offset);
+	sg_pcopy_to_buffer(sgl, nents, tmp_buf2, AES_BLOCK_SIZE, blk2_offset);
+	sg_pcopy_from_buffer(sgl, nents, tmp_buf1, AES_BLOCK_SIZE, blk2_offset);
+	sg_pcopy_from_buffer(sgl, nents, tmp_buf2, AES_BLOCK_SIZE, blk1_offset);
+}
+
+/* Initialize request context to default values. */
+static void ocs_aes_init_rctx(struct ocs_aes_rctx *rctx)
+{
+	/* Zero everything. */
+	memset(rctx, 0, sizeof(*rctx));
+
+	/* Set initial value for DMA addresses. */
+	rctx->src_dll.dma_addr = DMA_MAPPING_ERROR;
+	rctx->dst_dll.dma_addr = DMA_MAPPING_ERROR;
+	rctx->aad_src_dll.dma_addr = DMA_MAPPING_ERROR;
+	rctx->aad_dst_dll.dma_addr = DMA_MAPPING_ERROR;
+}
+
+static int kmb_ocs_sk_validate_input(struct skcipher_request *req,
+				     enum ocs_mode mode)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	int iv_size = crypto_skcipher_ivsize(tfm);
+
+	switch (mode) {
+	case OCS_MODE_ECB:
+		/* Ensure input length is multiple of block size */
+		if (req->cryptlen % AES_BLOCK_SIZE != 0)
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_CBC:
+		/* Ensure input length is multiple of block size */
+		if (req->cryptlen % AES_BLOCK_SIZE != 0)
+			return -EINVAL;
+
+		/* Ensure IV is present and block size in length */
+		if (!req->iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+		/*
+		 * NOTE: Since req->cryptlen == 0 case was already handled in
+		 * kmb_ocs_sk_common(), the above two conditions also guarantee
+		 * that: cryptlen >= iv_size
+		 */
+		return 0;
+
+	case OCS_MODE_CTR:
+		/* Ensure IV is present and block size in length */
+		if (!req->iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+		return 0;
+
+	case OCS_MODE_CTS:
+		/* Ensure input length >= block size */
+		if (req->cryptlen < AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		/* Ensure IV is present and block size in length */
+		if (!req->iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+/*
+ * Called by encrypt() / decrypt() skcipher functions.
+ *
+ * Use fallback if needed, otherwise initialize context and enqueue request
+ * into engine.
+ */
+static int kmb_ocs_sk_common(struct skcipher_request *req,
+			     enum ocs_cipher cipher,
+			     enum ocs_instruction instruction,
+			     enum ocs_mode mode)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ocs_aes_rctx *rctx = skcipher_request_ctx(req);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	struct ocs_aes_dev *aes_dev;
+	int rc;
+
+	if (tctx->use_fallback) {
+		SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, tctx->sw_cipher.sk);
+
+		skcipher_request_set_sync_tfm(subreq, tctx->sw_cipher.sk);
+		skcipher_request_set_callback(subreq, req->base.flags, NULL,
+					      NULL);
+		skcipher_request_set_crypt(subreq, req->src, req->dst,
+					   req->cryptlen, req->iv);
+
+		if (instruction == OCS_ENCRYPT)
+			rc = crypto_skcipher_encrypt(subreq);
+		else
+			rc = crypto_skcipher_decrypt(subreq);
+
+		skcipher_request_zero(subreq);
+
+		return rc;
+	}
+
+	/*
+	 * If cryptlen == 0, no processing needed for ECB, CBC and CTR.
+	 *
+	 * For CTS continue: kmb_ocs_sk_validate_input() will return -EINVAL.
+	 */
+	if (!req->cryptlen && mode != OCS_MODE_CTS)
+		return 0;
+
+	rc = kmb_ocs_sk_validate_input(req, mode);
+	if (rc)
+		return rc;
+
+	aes_dev = kmb_ocs_aes_find_dev(tctx);
+	if (!aes_dev)
+		return -ENODEV;
+
+	if (cipher != tctx->cipher)
+		return -EINVAL;
+
+	ocs_aes_init_rctx(rctx);
+	rctx->instruction = instruction;
+	rctx->mode = mode;
+
+	return crypto_transfer_skcipher_request_to_engine(aes_dev->engine, req);
+}
+
+static void cleanup_ocs_dma_linked_list(struct device *dev,
+					struct ocs_dll_desc *dll)
+{
+	if (dll->vaddr)
+		dma_free_coherent(dev, dll->size, dll->vaddr, dll->dma_addr);
+	dll->vaddr = NULL;
+	dll->size = 0;
+	dll->dma_addr = DMA_MAPPING_ERROR;
+}
+
+static void kmb_ocs_sk_dma_cleanup(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ocs_aes_rctx *rctx = skcipher_request_ctx(req);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	struct device *dev = tctx->aes_dev->dev;
+
+	if (rctx->src_dma_count) {
+		dma_unmap_sg(dev, req->src, rctx->src_nents, DMA_TO_DEVICE);
+		rctx->src_dma_count = 0;
+	}
+
+	if (rctx->dst_dma_count) {
+		dma_unmap_sg(dev, req->dst, rctx->dst_nents, rctx->in_place ?
+							     DMA_BIDIRECTIONAL :
+							     DMA_FROM_DEVICE);
+		rctx->dst_dma_count = 0;
+	}
+
+	/* Clean up OCS DMA linked lists */
+	cleanup_ocs_dma_linked_list(dev, &rctx->src_dll);
+	cleanup_ocs_dma_linked_list(dev, &rctx->dst_dll);
+}
+
+static int kmb_ocs_sk_prepare_inplace(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ocs_aes_rctx *rctx = skcipher_request_ctx(req);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	int iv_size = crypto_skcipher_ivsize(tfm);
+	int rc;
+
+	/*
+	 * For CBC decrypt, save last block (iv) to last_ct_blk buffer.
+	 *
+	 * Note: if we are here, we already checked that cryptlen >= iv_size
+	 * and iv_size == AES_BLOCK_SIZE (i.e., the size of last_ct_blk); see
+	 * kmb_ocs_sk_validate_input().
+	 */
+	if (rctx->mode == OCS_MODE_CBC && rctx->instruction == OCS_DECRYPT)
+		scatterwalk_map_and_copy(rctx->last_ct_blk, req->src,
+					 req->cryptlen - iv_size, iv_size, 0);
+
+	/* For CTS decrypt, swap last two blocks, if needed. */
+	if (rctx->cts_swap && rctx->instruction == OCS_DECRYPT)
+		sg_swap_blocks(req->dst, rctx->dst_nents,
+			       req->cryptlen - AES_BLOCK_SIZE,
+			       req->cryptlen - (2 * AES_BLOCK_SIZE));
+
+	/* src and dst buffers are the same, use bidirectional DMA mapping. */
+	rctx->dst_dma_count = dma_map_sg(tctx->aes_dev->dev, req->dst,
+					 rctx->dst_nents, DMA_BIDIRECTIONAL);
+	if (rctx->dst_dma_count == 0) {
+		dev_err(tctx->aes_dev->dev, "Failed to map destination sg\n");
+		return -ENOMEM;
+	}
+
+	/* Create DST linked list */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->dst,
+					    rctx->dst_dma_count, &rctx->dst_dll,
+					    req->cryptlen, 0);
+	if (rc)
+		return rc;
+	/*
+	 * If descriptor creation was successful, set the src_dll.dma_addr to
+	 * the value of dst_dll.dma_addr, as we do in-place AES operation on
+	 * the src.
+	 */
+	rctx->src_dll.dma_addr = rctx->dst_dll.dma_addr;
+
+	return 0;
+}
+
+static int kmb_ocs_sk_prepare_notinplace(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ocs_aes_rctx *rctx = skcipher_request_ctx(req);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	int rc;
+
+	rctx->src_nents =  sg_nents_for_len(req->src, req->cryptlen);
+	if (rctx->src_nents < 0)
+		return -EBADMSG;
+
+	/* Map SRC SG. */
+	rctx->src_dma_count = dma_map_sg(tctx->aes_dev->dev, req->src,
+					 rctx->src_nents, DMA_TO_DEVICE);
+	if (rctx->src_dma_count == 0) {
+		dev_err(tctx->aes_dev->dev, "Failed to map source sg\n");
+		return -ENOMEM;
+	}
+
+	/* Create SRC linked list */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->src,
+					    rctx->src_dma_count, &rctx->src_dll,
+					    req->cryptlen, 0);
+	if (rc)
+		return rc;
+
+	/* Map DST SG. */
+	rctx->dst_dma_count = dma_map_sg(tctx->aes_dev->dev, req->dst,
+					 rctx->dst_nents, DMA_FROM_DEVICE);
+	if (rctx->dst_dma_count == 0) {
+		dev_err(tctx->aes_dev->dev, "Failed to map destination sg\n");
+		return -ENOMEM;
+	}
+
+	/* Create DST linked list */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->dst,
+					    rctx->dst_dma_count, &rctx->dst_dll,
+					    req->cryptlen, 0);
+	if (rc)
+		return rc;
+
+	/* If this is not a CTS decrypt operation with swapping, we are done. */
+	if (!(rctx->cts_swap && rctx->instruction == OCS_DECRYPT))
+		return 0;
+
+	/*
+	 * Otherwise, we have to copy src to dst (as we cannot modify src).
+	 * Use OCS AES bypass mode to copy src to dst via DMA.
+	 *
+	 * NOTE: for anything other than small data sizes this is rather
+	 * inefficient.
+	 */
+	rc = ocs_aes_bypass_op(tctx->aes_dev, rctx->dst_dll.dma_addr,
+			       rctx->src_dll.dma_addr, req->cryptlen);
+	if (rc)
+		return rc;
+
+	/*
+	 * Now dst == src, so clean up what we did so far and use in_place
+	 * logic.
+	 */
+	kmb_ocs_sk_dma_cleanup(req);
+	rctx->in_place = true;
+
+	return kmb_ocs_sk_prepare_inplace(req);
+}
+
+static int kmb_ocs_sk_run(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ocs_aes_rctx *rctx = skcipher_request_ctx(req);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	struct ocs_aes_dev *aes_dev = tctx->aes_dev;
+	int iv_size = crypto_skcipher_ivsize(tfm);
+	int rc;
+
+	rctx->dst_nents = sg_nents_for_len(req->dst, req->cryptlen);
+	if (rctx->dst_nents < 0)
+		return -EBADMSG;
+
+	/*
+	 * If 2 blocks or greater, and multiple of block size swap last two
+	 * blocks to be compatible with other crypto API CTS implementations:
+	 * OCS mode uses CBC-CS2, whereas other crypto API implementations use
+	 * CBC-CS3.
+	 * CBC-CS2 and CBC-CS3 defined by:
+	 * https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-38a-add.pdf
+	 */
+	rctx->cts_swap = (rctx->mode == OCS_MODE_CTS &&
+			  req->cryptlen > AES_BLOCK_SIZE &&
+			  req->cryptlen % AES_BLOCK_SIZE == 0);
+
+	rctx->in_place = (req->src == req->dst);
+
+	if (rctx->in_place)
+		rc = kmb_ocs_sk_prepare_inplace(req);
+	else
+		rc = kmb_ocs_sk_prepare_notinplace(req);
+
+	if (rc)
+		goto error;
+
+	rc = ocs_aes_op(aes_dev, rctx->mode, tctx->cipher, rctx->instruction,
+			rctx->dst_dll.dma_addr, rctx->src_dll.dma_addr,
+			req->cryptlen, req->iv, iv_size);
+	if (rc)
+		goto error;
+
+	/* Clean-up DMA before further processing output. */
+	kmb_ocs_sk_dma_cleanup(req);
+
+	/* For CTS Encrypt, swap last 2 blocks, if needed. */
+	if (rctx->cts_swap && rctx->instruction == OCS_ENCRYPT) {
+		sg_swap_blocks(req->dst, rctx->dst_nents,
+			       req->cryptlen - AES_BLOCK_SIZE,
+			       req->cryptlen - (2 * AES_BLOCK_SIZE));
+		return 0;
+	}
+
+	/* For CBC copy IV to req->IV. */
+	if (rctx->mode == OCS_MODE_CBC) {
+		/* CBC encrypt case. */
+		if (rctx->instruction == OCS_ENCRYPT) {
+			scatterwalk_map_and_copy(req->iv, req->dst,
+						 req->cryptlen - iv_size,
+						 iv_size, 0);
+			return 0;
+		}
+		/* CBC decrypt case. */
+		if (rctx->in_place)
+			memcpy(req->iv, rctx->last_ct_blk, iv_size);
+		else
+			scatterwalk_map_and_copy(req->iv, req->src,
+						 req->cryptlen - iv_size,
+						 iv_size, 0);
+		return 0;
+	}
+	/* For all other modes there's nothing to do. */
+
+	return 0;
+
+error:
+	kmb_ocs_sk_dma_cleanup(req);
+
+	return rc;
+}
+
+static int kmb_ocs_aead_validate_input(struct aead_request *req,
+				       enum ocs_instruction instruction,
+				       enum ocs_mode mode)
+{
+	struct crypto_aead *tfm = crypto_aead_reqtfm(req);
+	int tag_size = crypto_aead_authsize(tfm);
+	int iv_size = crypto_aead_ivsize(tfm);
+
+	/* For decrypt crytplen == len(PT) + len(tag). */
+	if (instruction == OCS_DECRYPT && req->cryptlen < tag_size)
+		return -EINVAL;
+
+	/* IV is mandatory. */
+	if (!req->iv)
+		return -EINVAL;
+
+	switch (mode) {
+	case OCS_MODE_GCM:
+		if (iv_size != GCM_AES_IV_SIZE)
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_CCM:
+		/* Ensure IV is present and block size in length */
+		if (iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		return 0;
+
+	default:
+		return -EINVAL;
+	}
+}
+
+/*
+ * Called by encrypt() / decrypt() aead functions.
+ *
+ * Use fallback if needed, otherwise initialize context and enqueue request
+ * into engine.
+ */
+static int kmb_ocs_aead_common(struct aead_request *req,
+			       enum ocs_cipher cipher,
+			       enum ocs_instruction instruction,
+			       enum ocs_mode mode)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(crypto_aead_reqtfm(req));
+	struct ocs_aes_rctx *rctx = aead_request_ctx(req);
+	struct ocs_aes_dev *dd;
+	int rc;
+
+	if (tctx->use_fallback) {
+		struct aead_request *subreq = aead_request_ctx(req);
+
+		aead_request_set_tfm(subreq, tctx->sw_cipher.aead);
+		aead_request_set_callback(subreq, req->base.flags,
+					  req->base.complete, req->base.data);
+		aead_request_set_crypt(subreq, req->src, req->dst,
+				       req->cryptlen, req->iv);
+		aead_request_set_ad(subreq, req->assoclen);
+		rc = crypto_aead_setauthsize(tctx->sw_cipher.aead,
+					     crypto_aead_authsize(crypto_aead_reqtfm(req)));
+		if (rc)
+			return rc;
+
+		return (instruction == OCS_ENCRYPT) ?
+		       crypto_aead_encrypt(subreq) :
+		       crypto_aead_decrypt(subreq);
+	}
+
+	rc = kmb_ocs_aead_validate_input(req, instruction, mode);
+	if (rc)
+		return rc;
+
+	dd = kmb_ocs_aes_find_dev(tctx);
+	if (!dd)
+		return -ENODEV;
+
+	if (cipher != tctx->cipher)
+		return -EINVAL;
+
+	ocs_aes_init_rctx(rctx);
+	rctx->instruction = instruction;
+	rctx->mode = mode;
+
+	return crypto_transfer_aead_request_to_engine(dd->engine, req);
+}
+
+static void kmb_ocs_aead_dma_cleanup(struct aead_request *req)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(crypto_aead_reqtfm(req));
+	struct ocs_aes_rctx *rctx = aead_request_ctx(req);
+	struct device *dev = tctx->aes_dev->dev;
+
+	if (rctx->src_dma_count) {
+		dma_unmap_sg(dev, req->src, rctx->src_nents, DMA_TO_DEVICE);
+		rctx->src_dma_count = 0;
+	}
+
+	if (rctx->dst_dma_count) {
+		dma_unmap_sg(dev, req->dst, rctx->dst_nents, rctx->in_place ?
+							     DMA_BIDIRECTIONAL :
+							     DMA_FROM_DEVICE);
+		rctx->dst_dma_count = 0;
+	}
+	/* Clean up OCS DMA linked lists */
+	cleanup_ocs_dma_linked_list(dev, &rctx->src_dll);
+	cleanup_ocs_dma_linked_list(dev, &rctx->dst_dll);
+	cleanup_ocs_dma_linked_list(dev, &rctx->aad_src_dll);
+	cleanup_ocs_dma_linked_list(dev, &rctx->aad_dst_dll);
+}
+
+/**
+ * kmb_ocs_aead_dma_prepare() - Do DMA mapping for AEAD processing.
+ * @req:		The AEAD request being processed.
+ * @src_dll_size:	Where to store the length of the data mapped into the
+ *			src_dll OCS DMA list.
+ *
+ * Do the following:
+ * - DMA map req->src and req->dst
+ * - Initialize the following OCS DMA linked lists: rctx->src_dll,
+ *   rctx->dst_dll, rctx->aad_src_dll and rxtc->aad_dst_dll.
+ *
+ * Return: 0 on success, negative error code otherwise.
+ */
+static int kmb_ocs_aead_dma_prepare(struct aead_request *req, u32 *src_dll_size)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(crypto_aead_reqtfm(req));
+	const int tag_size = crypto_aead_authsize(crypto_aead_reqtfm(req));
+	struct ocs_aes_rctx *rctx = aead_request_ctx(req);
+	u32 in_size;	/* The length of the data to be mapped by src_dll. */
+	u32 out_size;	/* The length of the data to be mapped by dst_dll. */
+	u32 dst_size;	/* The length of the data in dst_sg. */
+	int rc;
+
+	/* Get number of entries in input data SG list. */
+	rctx->src_nents = sg_nents_for_len(req->src,
+					   req->assoclen + req->cryptlen);
+	if (rctx->src_nents < 0)
+		return -EBADMSG;
+
+	if (rctx->instruction == OCS_DECRYPT) {
+		/*
+		 * For decrypt:
+		 * - src sg list is:		AAD|CT|tag
+		 * - dst sg list expects:	AAD|PT
+		 *
+		 * in_size == len(CT); out_size == len(PT)
+		 */
+
+		/* req->cryptlen includes both CT and tag. */
+		in_size = req->cryptlen - tag_size;
+
+		/* out_size = PT size == CT size */
+		out_size = in_size;
+
+		/* len(dst_sg) == len(AAD) + len(PT) */
+		dst_size = req->assoclen + out_size;
+
+		/*
+		 * Copy tag from source SG list to 'in_tag' buffer.
+		 *
+		 * Note: this needs to be done here, before DMA mapping src_sg.
+		 */
+		sg_pcopy_to_buffer(req->src, rctx->src_nents, rctx->in_tag,
+				   tag_size, req->assoclen + in_size);
+
+	} else { /* OCS_ENCRYPT */
+		/*
+		 * For encrypt:
+		 *	src sg list is:		AAD|PT
+		 *	dst sg list expects:	AAD|CT|tag
+		 */
+		/* in_size == len(PT) */
+		in_size = req->cryptlen;
+
+		/*
+		 * In CCM mode the OCS engine appends the tag to the ciphertext,
+		 * but in GCM mode the tag must be read from the tag registers
+		 * and appended manually below
+		 */
+		out_size = (rctx->mode == OCS_MODE_CCM) ? in_size + tag_size :
+							  in_size;
+		/* len(dst_sg) == len(AAD) + len(CT) + len(tag) */
+		dst_size = req->assoclen + in_size + tag_size;
+	}
+	*src_dll_size = in_size;
+
+	/* Get number of entries in output data SG list. */
+	rctx->dst_nents = sg_nents_for_len(req->dst, dst_size);
+	if (rctx->dst_nents < 0)
+		return -EBADMSG;
+
+	rctx->in_place = (req->src == req->dst) ? 1 : 0;
+
+	/* Map destination; use bidirectional mapping for in-place case. */
+	rctx->dst_dma_count = dma_map_sg(tctx->aes_dev->dev, req->dst,
+					 rctx->dst_nents,
+					 rctx->in_place ? DMA_BIDIRECTIONAL :
+							  DMA_FROM_DEVICE);
+	if (rctx->dst_dma_count == 0 && rctx->dst_nents != 0) {
+		dev_err(tctx->aes_dev->dev, "Failed to map destination sg\n");
+		return -ENOMEM;
+	}
+
+	/* Create AAD DST list: maps dst[0:AAD_SIZE-1]. */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->dst,
+					    rctx->dst_dma_count,
+					    &rctx->aad_dst_dll, req->assoclen,
+					    0);
+	if (rc)
+		return rc;
+
+	/* Create DST list: maps dst[AAD_SIZE:out_size] */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->dst,
+					    rctx->dst_dma_count, &rctx->dst_dll,
+					    out_size, req->assoclen);
+	if (rc)
+		return rc;
+
+	if (rctx->in_place) {
+		/* If this is not CCM encrypt, we are done. */
+		if (!(rctx->mode == OCS_MODE_CCM &&
+		      rctx->instruction == OCS_ENCRYPT)) {
+			/*
+			 * SRC and DST are the same, so re-use the same DMA
+			 * addresses (to avoid allocating new DMA lists
+			 * identical to the dst ones).
+			 */
+			rctx->src_dll.dma_addr = rctx->dst_dll.dma_addr;
+			rctx->aad_src_dll.dma_addr = rctx->aad_dst_dll.dma_addr;
+
+			return 0;
+		}
+		/*
+		 * For CCM encrypt the input and output linked lists contain
+		 * different amounts of data, so, we need to create different
+		 * SRC and AAD SRC lists, even for the in-place case.
+		 */
+		rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->dst,
+						    rctx->dst_dma_count,
+						    &rctx->aad_src_dll,
+						    req->assoclen, 0);
+		if (rc)
+			return rc;
+		rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->dst,
+						    rctx->dst_dma_count,
+						    &rctx->src_dll, in_size,
+						    req->assoclen);
+		if (rc)
+			return rc;
+
+		return 0;
+	}
+	/* Not in-place case. */
+
+	/* Map source SG. */
+	rctx->src_dma_count = dma_map_sg(tctx->aes_dev->dev, req->src,
+					 rctx->src_nents, DMA_TO_DEVICE);
+	if (rctx->src_dma_count == 0 && rctx->src_nents != 0) {
+		dev_err(tctx->aes_dev->dev, "Failed to map source sg\n");
+		return -ENOMEM;
+	}
+
+	/* Create AAD SRC list. */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->src,
+					    rctx->src_dma_count,
+					    &rctx->aad_src_dll,
+					    req->assoclen, 0);
+	if (rc)
+		return rc;
+
+	/* Create SRC list. */
+	rc = ocs_create_linked_list_from_sg(tctx->aes_dev, req->src,
+					    rctx->src_dma_count,
+					    &rctx->src_dll, in_size,
+					    req->assoclen);
+	if (rc)
+		return rc;
+
+	if (req->assoclen == 0)
+		return 0;
+
+	/* Copy AAD from src sg to dst sg using OCS DMA. */
+	rc = ocs_aes_bypass_op(tctx->aes_dev, rctx->aad_dst_dll.dma_addr,
+			       rctx->aad_src_dll.dma_addr, req->cryptlen);
+	if (rc)
+		dev_err(tctx->aes_dev->dev,
+			"Failed to copy source AAD to destination AAD\n");
+
+	return rc;
+}
+
+static int kmb_ocs_aead_run(struct aead_request *req)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(crypto_aead_reqtfm(req));
+	const int tag_size = crypto_aead_authsize(crypto_aead_reqtfm(req));
+	struct ocs_aes_rctx *rctx = aead_request_ctx(req);
+	u32 in_size;	/* The length of the data mapped by src_dll. */
+	int rc;
+
+	rc = kmb_ocs_aead_dma_prepare(req, &in_size);
+	if (rc)
+		goto exit;
+
+	/* For CCM, we just call the OCS processing and we are done. */
+	if (rctx->mode == OCS_MODE_CCM) {
+		rc = ocs_aes_ccm_op(tctx->aes_dev, tctx->cipher,
+				    rctx->instruction, rctx->dst_dll.dma_addr,
+				    rctx->src_dll.dma_addr, in_size,
+				    req->iv,
+				    rctx->aad_src_dll.dma_addr, req->assoclen,
+				    rctx->in_tag, tag_size);
+		goto exit;
+	}
+	/* GCM case; invoke OCS processing. */
+	rc = ocs_aes_gcm_op(tctx->aes_dev, tctx->cipher,
+			    rctx->instruction,
+			    rctx->dst_dll.dma_addr,
+			    rctx->src_dll.dma_addr, in_size,
+			    req->iv,
+			    rctx->aad_src_dll.dma_addr, req->assoclen,
+			    rctx->out_tag, tag_size);
+	if (rc)
+		goto exit;
+
+	/* For GCM decrypt, we have to compare in_tag with out_tag. */
+	if (rctx->instruction == OCS_DECRYPT) {
+		rc = memcmp(rctx->in_tag, rctx->out_tag, tag_size) ?
+		     -EBADMSG : 0;
+		goto exit;
+	}
+
+	/* For GCM encrypt, we must manually copy out_tag to DST sg. */
+
+	/* Clean-up must be called before the sg_pcopy_from_buffer() below. */
+	kmb_ocs_aead_dma_cleanup(req);
+
+	/* Copy tag to destination sg after AAD and CT. */
+	sg_pcopy_from_buffer(req->dst, rctx->dst_nents, rctx->out_tag,
+			     tag_size, req->assoclen + req->cryptlen);
+
+	/* Return directly as DMA cleanup already done. */
+	return 0;
+
+exit:
+	kmb_ocs_aead_dma_cleanup(req);
+
+	return rc;
+}
+
+static int kmb_ocs_aes_sk_do_one_request(struct crypto_engine *engine,
+					 void *areq)
+{
+	struct skcipher_request *req =
+			container_of(areq, struct skcipher_request, base);
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	int err;
+
+	if (!tctx->aes_dev) {
+		err = -ENODEV;
+		goto exit;
+	}
+
+	err = ocs_aes_set_key(tctx->aes_dev, tctx->key_len, tctx->key,
+			      tctx->cipher);
+	if (err)
+		goto exit;
+
+	err = kmb_ocs_sk_run(req);
+
+exit:
+	crypto_finalize_skcipher_request(engine, req, err);
+
+	return 0;
+}
+
+static int kmb_ocs_aes_aead_do_one_request(struct crypto_engine *engine,
+					   void *areq)
+{
+	struct aead_request *req = container_of(areq,
+						struct aead_request, base);
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(crypto_aead_reqtfm(req));
+	int err;
+
+	if (!tctx->aes_dev)
+		return -ENODEV;
+
+	err = ocs_aes_set_key(tctx->aes_dev, tctx->key_len, tctx->key,
+			      tctx->cipher);
+	if (err)
+		goto exit;
+
+	err = kmb_ocs_aead_run(req);
+
+exit:
+	crypto_finalize_aead_request(tctx->aes_dev->engine, req, err);
+
+	return 0;
+}
+
+static int kmb_ocs_aes_set_key(struct crypto_skcipher *tfm, const u8 *in_key,
+			       unsigned int key_len)
+{
+	return kmb_ocs_sk_set_key(tfm, in_key, key_len, OCS_AES);
+}
+
+static int kmb_ocs_aes_aead_set_key(struct crypto_aead *tfm, const u8 *in_key,
+				    unsigned int key_len)
+{
+	return kmb_ocs_aead_set_key(tfm, in_key, key_len, OCS_AES);
+}
+
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB
+static int kmb_ocs_aes_ecb_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_ENCRYPT, OCS_MODE_ECB);
+}
+
+static int kmb_ocs_aes_ecb_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_DECRYPT, OCS_MODE_ECB);
+}
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB */
+
+static int kmb_ocs_aes_cbc_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_ENCRYPT, OCS_MODE_CBC);
+}
+
+static int kmb_ocs_aes_cbc_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_DECRYPT, OCS_MODE_CBC);
+}
+
+static int kmb_ocs_aes_ctr_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_ENCRYPT, OCS_MODE_CTR);
+}
+
+static int kmb_ocs_aes_ctr_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_DECRYPT, OCS_MODE_CTR);
+}
+
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS
+static int kmb_ocs_aes_cts_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_ENCRYPT, OCS_MODE_CTS);
+}
+
+static int kmb_ocs_aes_cts_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_AES, OCS_DECRYPT, OCS_MODE_CTS);
+}
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS */
+
+static int kmb_ocs_aes_gcm_encrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_AES, OCS_ENCRYPT, OCS_MODE_GCM);
+}
+
+static int kmb_ocs_aes_gcm_decrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_AES, OCS_DECRYPT, OCS_MODE_GCM);
+}
+
+static int kmb_ocs_aes_ccm_encrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_AES, OCS_ENCRYPT, OCS_MODE_CCM);
+}
+
+static int kmb_ocs_aes_ccm_decrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_AES, OCS_DECRYPT, OCS_MODE_CCM);
+}
+
+static int kmb_ocs_sm4_set_key(struct crypto_skcipher *tfm, const u8 *in_key,
+			       unsigned int key_len)
+{
+	return kmb_ocs_sk_set_key(tfm, in_key, key_len, OCS_SM4);
+}
+
+static int kmb_ocs_sm4_aead_set_key(struct crypto_aead *tfm, const u8 *in_key,
+				    unsigned int key_len)
+{
+	return kmb_ocs_aead_set_key(tfm, in_key, key_len, OCS_SM4);
+}
+
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB
+static int kmb_ocs_sm4_ecb_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_ENCRYPT, OCS_MODE_ECB);
+}
+
+static int kmb_ocs_sm4_ecb_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_DECRYPT, OCS_MODE_ECB);
+}
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB */
+
+static int kmb_ocs_sm4_cbc_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_ENCRYPT, OCS_MODE_CBC);
+}
+
+static int kmb_ocs_sm4_cbc_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_DECRYPT, OCS_MODE_CBC);
+}
+
+static int kmb_ocs_sm4_ctr_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_ENCRYPT, OCS_MODE_CTR);
+}
+
+static int kmb_ocs_sm4_ctr_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_DECRYPT, OCS_MODE_CTR);
+}
+
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS
+static int kmb_ocs_sm4_cts_encrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_ENCRYPT, OCS_MODE_CTS);
+}
+
+static int kmb_ocs_sm4_cts_decrypt(struct skcipher_request *req)
+{
+	return kmb_ocs_sk_common(req, OCS_SM4, OCS_DECRYPT, OCS_MODE_CTS);
+}
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS */
+
+static int kmb_ocs_sm4_gcm_encrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_SM4, OCS_ENCRYPT, OCS_MODE_GCM);
+}
+
+static int kmb_ocs_sm4_gcm_decrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_SM4, OCS_DECRYPT, OCS_MODE_GCM);
+}
+
+static int kmb_ocs_sm4_ccm_encrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_SM4, OCS_ENCRYPT, OCS_MODE_CCM);
+}
+
+static int kmb_ocs_sm4_ccm_decrypt(struct aead_request *req)
+{
+	return kmb_ocs_aead_common(req, OCS_SM4, OCS_DECRYPT, OCS_MODE_CCM);
+}
+
+static inline int ocs_common_init(struct ocs_aes_tctx *tctx)
+{
+	tctx->engine_ctx.op.prepare_request = NULL;
+	tctx->engine_ctx.op.do_one_request = kmb_ocs_aes_sk_do_one_request;
+	tctx->engine_ctx.op.unprepare_request = NULL;
+
+	return 0;
+}
+
+static int ocs_aes_init_tfm(struct crypto_skcipher *tfm)
+{
+	const char *alg_name = crypto_tfm_alg_name(&tfm->base);
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+	struct crypto_sync_skcipher *blk;
+
+	/* set fallback cipher in case it will be needed */
+	blk = crypto_alloc_sync_skcipher(alg_name, 0, CRYPTO_ALG_NEED_FALLBACK);
+	if (IS_ERR(blk))
+		return PTR_ERR(blk);
+
+	tctx->sw_cipher.sk = blk;
+
+	crypto_skcipher_set_reqsize(tfm, sizeof(struct ocs_aes_rctx));
+
+	return ocs_common_init(tctx);
+}
+
+static int ocs_sm4_init_tfm(struct crypto_skcipher *tfm)
+{
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+
+	crypto_skcipher_set_reqsize(tfm, sizeof(struct ocs_aes_rctx));
+
+	return ocs_common_init(tctx);
+}
+
+static inline void clear_key(struct ocs_aes_tctx *tctx)
+{
+	memzero_explicit(tctx->key, OCS_AES_KEYSIZE_256);
+
+	/* Zero key registers if set */
+	if (tctx->aes_dev)
+		ocs_aes_set_key(tctx->aes_dev, OCS_AES_KEYSIZE_256,
+				tctx->key, OCS_AES);
+}
+
+static void ocs_exit_tfm(struct crypto_skcipher *tfm)
+{
+	struct ocs_aes_tctx *tctx = crypto_skcipher_ctx(tfm);
+
+	clear_key(tctx);
+
+	if (tctx->sw_cipher.sk) {
+		crypto_free_sync_skcipher(tctx->sw_cipher.sk);
+		tctx->sw_cipher.sk = NULL;
+	}
+}
+
+static inline int ocs_common_aead_init(struct ocs_aes_tctx *tctx)
+{
+	tctx->engine_ctx.op.prepare_request = NULL;
+	tctx->engine_ctx.op.do_one_request = kmb_ocs_aes_aead_do_one_request;
+	tctx->engine_ctx.op.unprepare_request = NULL;
+
+	return 0;
+}
+
+static int ocs_aes_aead_cra_init(struct crypto_aead *tfm)
+{
+	const char *alg_name = crypto_tfm_alg_name(&tfm->base);
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(tfm);
+	struct crypto_aead *blk;
+
+	/* Set fallback cipher in case it will be needed */
+	blk = crypto_alloc_aead(alg_name, 0, CRYPTO_ALG_NEED_FALLBACK);
+	if (IS_ERR(blk))
+		return PTR_ERR(blk);
+
+	tctx->sw_cipher.aead = blk;
+
+	crypto_aead_set_reqsize(tfm,
+				max(sizeof(struct ocs_aes_rctx),
+				    (sizeof(struct aead_request) +
+				     crypto_aead_reqsize(tctx->sw_cipher.aead))));
+
+	return ocs_common_aead_init(tctx);
+}
+
+static int kmb_ocs_aead_ccm_setauthsize(struct crypto_aead *tfm,
+					unsigned int authsize)
+{
+	switch (authsize) {
+	case 4:
+	case 6:
+	case 8:
+	case 10:
+	case 12:
+	case 14:
+	case 16:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int kmb_ocs_aead_gcm_setauthsize(struct crypto_aead *tfm,
+					unsigned int authsize)
+{
+	return crypto_gcm_check_authsize(authsize);
+}
+
+static int ocs_sm4_aead_cra_init(struct crypto_aead *tfm)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(tfm);
+
+	crypto_aead_set_reqsize(tfm, sizeof(struct ocs_aes_rctx));
+
+	return ocs_common_aead_init(tctx);
+}
+
+static void ocs_aead_cra_exit(struct crypto_aead *tfm)
+{
+	struct ocs_aes_tctx *tctx = crypto_aead_ctx(tfm);
+
+	clear_key(tctx);
+
+	if (tctx->sw_cipher.aead) {
+		crypto_free_aead(tctx->sw_cipher.aead);
+		tctx->sw_cipher.aead = NULL;
+	}
+}
+
+static struct skcipher_alg algs[] = {
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB
+	{
+		.base.cra_name = "ecb(aes)",
+		.base.cra_driver_name = "ecb-aes-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY |
+				  CRYPTO_ALG_NEED_FALLBACK,
+		.base.cra_blocksize = AES_BLOCK_SIZE,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_AES_MIN_KEY_SIZE,
+		.max_keysize = OCS_AES_MAX_KEY_SIZE,
+		.setkey = kmb_ocs_aes_set_key,
+		.encrypt = kmb_ocs_aes_ecb_encrypt,
+		.decrypt = kmb_ocs_aes_ecb_decrypt,
+		.init = ocs_aes_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB */
+	{
+		.base.cra_name = "cbc(aes)",
+		.base.cra_driver_name = "cbc-aes-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY |
+				  CRYPTO_ALG_NEED_FALLBACK,
+		.base.cra_blocksize = AES_BLOCK_SIZE,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_AES_MIN_KEY_SIZE,
+		.max_keysize = OCS_AES_MAX_KEY_SIZE,
+		.ivsize = AES_BLOCK_SIZE,
+		.setkey = kmb_ocs_aes_set_key,
+		.encrypt = kmb_ocs_aes_cbc_encrypt,
+		.decrypt = kmb_ocs_aes_cbc_decrypt,
+		.init = ocs_aes_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+	{
+		.base.cra_name = "ctr(aes)",
+		.base.cra_driver_name = "ctr-aes-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY |
+				  CRYPTO_ALG_NEED_FALLBACK,
+		.base.cra_blocksize = 1,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_AES_MIN_KEY_SIZE,
+		.max_keysize = OCS_AES_MAX_KEY_SIZE,
+		.ivsize = AES_BLOCK_SIZE,
+		.setkey = kmb_ocs_aes_set_key,
+		.encrypt = kmb_ocs_aes_ctr_encrypt,
+		.decrypt = kmb_ocs_aes_ctr_decrypt,
+		.init = ocs_aes_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS
+	{
+		.base.cra_name = "cts(cbc(aes))",
+		.base.cra_driver_name = "cts-aes-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY |
+				  CRYPTO_ALG_NEED_FALLBACK,
+		.base.cra_blocksize = AES_BLOCK_SIZE,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_AES_MIN_KEY_SIZE,
+		.max_keysize = OCS_AES_MAX_KEY_SIZE,
+		.ivsize = AES_BLOCK_SIZE,
+		.setkey = kmb_ocs_aes_set_key,
+		.encrypt = kmb_ocs_aes_cts_encrypt,
+		.decrypt = kmb_ocs_aes_cts_decrypt,
+		.init = ocs_aes_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS */
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB
+	{
+		.base.cra_name = "ecb(sm4)",
+		.base.cra_driver_name = "ecb-sm4-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY,
+		.base.cra_blocksize = AES_BLOCK_SIZE,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_SM4_KEY_SIZE,
+		.max_keysize = OCS_SM4_KEY_SIZE,
+		.setkey = kmb_ocs_sm4_set_key,
+		.encrypt = kmb_ocs_sm4_ecb_encrypt,
+		.decrypt = kmb_ocs_sm4_ecb_decrypt,
+		.init = ocs_sm4_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB */
+	{
+		.base.cra_name = "cbc(sm4)",
+		.base.cra_driver_name = "cbc-sm4-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY,
+		.base.cra_blocksize = AES_BLOCK_SIZE,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_SM4_KEY_SIZE,
+		.max_keysize = OCS_SM4_KEY_SIZE,
+		.ivsize = AES_BLOCK_SIZE,
+		.setkey = kmb_ocs_sm4_set_key,
+		.encrypt = kmb_ocs_sm4_cbc_encrypt,
+		.decrypt = kmb_ocs_sm4_cbc_decrypt,
+		.init = ocs_sm4_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+	{
+		.base.cra_name = "ctr(sm4)",
+		.base.cra_driver_name = "ctr-sm4-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY,
+		.base.cra_blocksize = 1,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_SM4_KEY_SIZE,
+		.max_keysize = OCS_SM4_KEY_SIZE,
+		.ivsize = AES_BLOCK_SIZE,
+		.setkey = kmb_ocs_sm4_set_key,
+		.encrypt = kmb_ocs_sm4_ctr_encrypt,
+		.decrypt = kmb_ocs_sm4_ctr_decrypt,
+		.init = ocs_sm4_init_tfm,
+		.exit = ocs_exit_tfm,
+	},
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS
+	{
+		.base.cra_name = "cts(cbc(sm4))",
+		.base.cra_driver_name = "cts-sm4-keembay-ocs",
+		.base.cra_priority = KMB_OCS_PRIORITY,
+		.base.cra_flags = CRYPTO_ALG_ASYNC |
+				  CRYPTO_ALG_KERN_DRIVER_ONLY,
+		.base.cra_blocksize = AES_BLOCK_SIZE,
+		.base.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+		.base.cra_module = THIS_MODULE,
+		.base.cra_alignmask = 0,
+
+		.min_keysize = OCS_SM4_KEY_SIZE,
+		.max_keysize = OCS_SM4_KEY_SIZE,
+		.ivsize = AES_BLOCK_SIZE,
+		.setkey = kmb_ocs_sm4_set_key,
+		.encrypt = kmb_ocs_sm4_cts_encrypt,
+		.decrypt = kmb_ocs_sm4_cts_decrypt,
+		.init = ocs_sm4_init_tfm,
+		.exit = ocs_exit_tfm,
+	}
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS */
+};
+
+static struct aead_alg algs_aead[] = {
+	{
+		.base = {
+			.cra_name = "gcm(aes)",
+			.cra_driver_name = "gcm-aes-keembay-ocs",
+			.cra_priority = KMB_OCS_PRIORITY,
+			.cra_flags = CRYPTO_ALG_ASYNC |
+				     CRYPTO_ALG_KERN_DRIVER_ONLY |
+				     CRYPTO_ALG_NEED_FALLBACK,
+			.cra_blocksize = 1,
+			.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+			.cra_alignmask = 0,
+			.cra_module = THIS_MODULE,
+		},
+		.init = ocs_aes_aead_cra_init,
+		.exit = ocs_aead_cra_exit,
+		.ivsize = GCM_AES_IV_SIZE,
+		.maxauthsize = AES_BLOCK_SIZE,
+		.setauthsize = kmb_ocs_aead_gcm_setauthsize,
+		.setkey = kmb_ocs_aes_aead_set_key,
+		.encrypt = kmb_ocs_aes_gcm_encrypt,
+		.decrypt = kmb_ocs_aes_gcm_decrypt,
+	},
+	{
+		.base = {
+			.cra_name = "ccm(aes)",
+			.cra_driver_name = "ccm-aes-keembay-ocs",
+			.cra_priority = KMB_OCS_PRIORITY,
+			.cra_flags = CRYPTO_ALG_ASYNC |
+				     CRYPTO_ALG_KERN_DRIVER_ONLY |
+				     CRYPTO_ALG_NEED_FALLBACK,
+			.cra_blocksize = 1,
+			.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+			.cra_alignmask = 0,
+			.cra_module = THIS_MODULE,
+		},
+		.init = ocs_aes_aead_cra_init,
+		.exit = ocs_aead_cra_exit,
+		.ivsize = AES_BLOCK_SIZE,
+		.maxauthsize = AES_BLOCK_SIZE,
+		.setauthsize = kmb_ocs_aead_ccm_setauthsize,
+		.setkey = kmb_ocs_aes_aead_set_key,
+		.encrypt = kmb_ocs_aes_ccm_encrypt,
+		.decrypt = kmb_ocs_aes_ccm_decrypt,
+	},
+	{
+		.base = {
+			.cra_name = "gcm(sm4)",
+			.cra_driver_name = "gcm-sm4-keembay-ocs",
+			.cra_priority = KMB_OCS_PRIORITY,
+			.cra_flags = CRYPTO_ALG_ASYNC |
+				     CRYPTO_ALG_KERN_DRIVER_ONLY,
+			.cra_blocksize = 1,
+			.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+			.cra_alignmask = 0,
+			.cra_module = THIS_MODULE,
+		},
+		.init = ocs_sm4_aead_cra_init,
+		.exit = ocs_aead_cra_exit,
+		.ivsize = GCM_AES_IV_SIZE,
+		.maxauthsize = AES_BLOCK_SIZE,
+		.setauthsize = kmb_ocs_aead_gcm_setauthsize,
+		.setkey = kmb_ocs_sm4_aead_set_key,
+		.encrypt = kmb_ocs_sm4_gcm_encrypt,
+		.decrypt = kmb_ocs_sm4_gcm_decrypt,
+	},
+	{
+		.base = {
+			.cra_name = "ccm(sm4)",
+			.cra_driver_name = "ccm-sm4-keembay-ocs",
+			.cra_priority = KMB_OCS_PRIORITY,
+			.cra_flags = CRYPTO_ALG_ASYNC |
+				     CRYPTO_ALG_KERN_DRIVER_ONLY,
+			.cra_blocksize = 1,
+			.cra_ctxsize = sizeof(struct ocs_aes_tctx),
+			.cra_alignmask = 0,
+			.cra_module = THIS_MODULE,
+		},
+		.init = ocs_sm4_aead_cra_init,
+		.exit = ocs_aead_cra_exit,
+		.ivsize = AES_BLOCK_SIZE,
+		.maxauthsize = AES_BLOCK_SIZE,
+		.setauthsize = kmb_ocs_aead_ccm_setauthsize,
+		.setkey = kmb_ocs_sm4_aead_set_key,
+		.encrypt = kmb_ocs_sm4_ccm_encrypt,
+		.decrypt = kmb_ocs_sm4_ccm_decrypt,
+	}
+};
+
+static void unregister_aes_algs(struct ocs_aes_dev *aes_dev)
+{
+	crypto_unregister_aeads(algs_aead, ARRAY_SIZE(algs_aead));
+	crypto_unregister_skciphers(algs, ARRAY_SIZE(algs));
+}
+
+static int register_aes_algs(struct ocs_aes_dev *aes_dev)
+{
+	int ret;
+
+	/*
+	 * If any algorithm fails to register, all preceding algorithms that
+	 * were successfully registered will be automatically unregistered.
+	 */
+	ret = crypto_register_aeads(algs_aead, ARRAY_SIZE(algs_aead));
+	if (ret)
+		return ret;
+
+	ret = crypto_register_skciphers(algs, ARRAY_SIZE(algs));
+	if (ret)
+		crypto_unregister_aeads(algs_aead, ARRAY_SIZE(algs));
+
+	return ret;
+}
+
+/* Device tree driver match. */
+static const struct of_device_id kmb_ocs_aes_of_match[] = {
+	{
+		.compatible = "intel,keembay-ocs-aes",
+	},
+	{}
+};
+
+static int kmb_ocs_aes_remove(struct platform_device *pdev)
+{
+	struct ocs_aes_dev *aes_dev;
+
+	aes_dev = platform_get_drvdata(pdev);
+	if (!aes_dev)
+		return -ENODEV;
+
+	unregister_aes_algs(aes_dev);
+
+	spin_lock(&ocs_aes.lock);
+	list_del(&aes_dev->list);
+	spin_unlock(&ocs_aes.lock);
+
+	crypto_engine_exit(aes_dev->engine);
+
+	return 0;
+}
+
+static int kmb_ocs_aes_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct ocs_aes_dev *aes_dev;
+	struct resource *aes_mem;
+	int rc;
+
+	aes_dev = devm_kzalloc(dev, sizeof(*aes_dev), GFP_KERNEL);
+	if (!aes_dev)
+		return -ENOMEM;
+
+	aes_dev->dev = dev;
+
+	platform_set_drvdata(pdev, aes_dev);
+
+	rc = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32));
+	if (rc) {
+		dev_err(dev, "Failed to set 32 bit dma mask %d\n", rc);
+		return rc;
+	}
+
+	/* Get base register address. */
+	aes_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!aes_mem) {
+		dev_err(dev, "Could not retrieve io mem resource\n");
+		return -ENODEV;
+	}
+
+	aes_dev->base_reg = devm_ioremap_resource(&pdev->dev, aes_mem);
+	if (IS_ERR(aes_dev->base_reg)) {
+		dev_err(dev, "Failed to get base address\n");
+		return PTR_ERR(aes_dev->base_reg);
+	}
+
+	/* Get and request IRQ */
+	aes_dev->irq = platform_get_irq(pdev, 0);
+	if (aes_dev->irq < 0)
+		return aes_dev->irq;
+
+	rc = devm_request_threaded_irq(dev, aes_dev->irq, ocs_aes_irq_handler,
+				       NULL, 0, "keembay-ocs-aes", aes_dev);
+	if (rc < 0) {
+		dev_err(dev, "Could not request IRQ\n");
+		return rc;
+	}
+
+	INIT_LIST_HEAD(&aes_dev->list);
+	spin_lock(&ocs_aes.lock);
+	list_add_tail(&aes_dev->list, &ocs_aes.dev_list);
+	spin_unlock(&ocs_aes.lock);
+
+	init_completion(&aes_dev->irq_completion);
+
+	/* Initialize crypto engine */
+	aes_dev->engine = crypto_engine_alloc_init(dev, true);
+	if (!aes_dev->engine)
+		goto list_del;
+
+	rc = crypto_engine_start(aes_dev->engine);
+	if (rc) {
+		dev_err(dev, "Could not start crypto engine\n");
+		goto cleanup;
+	}
+
+	rc = register_aes_algs(aes_dev);
+	if (rc) {
+		dev_err(dev,
+			"Could not register OCS algorithms with Crypto API\n");
+		goto cleanup;
+	}
+
+	return 0;
+
+cleanup:
+	crypto_engine_exit(aes_dev->engine);
+list_del:
+	spin_lock(&ocs_aes.lock);
+	list_del(&aes_dev->list);
+	spin_unlock(&ocs_aes.lock);
+
+	return rc;
+}
+
+/* The OCS driver is a platform device. */
+static struct platform_driver kmb_ocs_aes_driver = {
+	.probe = kmb_ocs_aes_probe,
+	.remove = kmb_ocs_aes_remove,
+	.driver = {
+			.name = DRV_NAME,
+			.of_match_table = kmb_ocs_aes_of_match,
+		},
+};
+
+module_platform_driver(kmb_ocs_aes_driver);
+
+MODULE_DESCRIPTION("Intel Keem Bay Offload and Crypto Subsystem (OCS) AES/SM4 Driver");
+MODULE_LICENSE("GPL");
+
+MODULE_ALIAS_CRYPTO("cbc-aes-keembay-ocs");
+MODULE_ALIAS_CRYPTO("ctr-aes-keembay-ocs");
+MODULE_ALIAS_CRYPTO("gcm-aes-keembay-ocs");
+MODULE_ALIAS_CRYPTO("ccm-aes-keembay-ocs");
+
+MODULE_ALIAS_CRYPTO("cbc-sm4-keembay-ocs");
+MODULE_ALIAS_CRYPTO("ctr-sm4-keembay-ocs");
+MODULE_ALIAS_CRYPTO("gcm-sm4-keembay-ocs");
+MODULE_ALIAS_CRYPTO("ccm-sm4-keembay-ocs");
+
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB
+MODULE_ALIAS_CRYPTO("ecb-aes-keembay-ocs");
+MODULE_ALIAS_CRYPTO("ecb-sm4-keembay-ocs");
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_ECB */
+
+#ifdef CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS
+MODULE_ALIAS_CRYPTO("cts-aes-keembay-ocs");
+MODULE_ALIAS_CRYPTO("cts-sm4-keembay-ocs");
+#endif /* CONFIG_CRYPTO_DEV_KEEMBAY_OCS_AES_SM4_CTS */
diff --git a/drivers/crypto/keembay/ocs-aes.c b/drivers/crypto/keembay/ocs-aes.c
new file mode 100644
index 000000000000..cc286adb1c4a
--- /dev/null
+++ b/drivers/crypto/keembay/ocs-aes.c
@@ -0,0 +1,1489 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Intel Keem Bay OCS AES Crypto Driver.
+ *
+ * Copyright (C) 2018-2020 Intel Corporation
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/swab.h>
+
+#include <asm/byteorder.h>
+#include <asm/errno.h>
+
+#include <crypto/aes.h>
+#include <crypto/gcm.h>
+
+#include "ocs-aes.h"
+
+#define AES_COMMAND_OFFSET			0x0000
+#define AES_KEY_0_OFFSET			0x0004
+#define AES_KEY_1_OFFSET			0x0008
+#define AES_KEY_2_OFFSET			0x000C
+#define AES_KEY_3_OFFSET			0x0010
+#define AES_KEY_4_OFFSET			0x0014
+#define AES_KEY_5_OFFSET			0x0018
+#define AES_KEY_6_OFFSET			0x001C
+#define AES_KEY_7_OFFSET			0x0020
+#define AES_IV_0_OFFSET				0x0024
+#define AES_IV_1_OFFSET				0x0028
+#define AES_IV_2_OFFSET				0x002C
+#define AES_IV_3_OFFSET				0x0030
+#define AES_ACTIVE_OFFSET			0x0034
+#define AES_STATUS_OFFSET			0x0038
+#define AES_KEY_SIZE_OFFSET			0x0044
+#define AES_IER_OFFSET				0x0048
+#define AES_ISR_OFFSET				0x005C
+#define AES_MULTIPURPOSE1_0_OFFSET		0x0200
+#define AES_MULTIPURPOSE1_1_OFFSET		0x0204
+#define AES_MULTIPURPOSE1_2_OFFSET		0x0208
+#define AES_MULTIPURPOSE1_3_OFFSET		0x020C
+#define AES_MULTIPURPOSE2_0_OFFSET		0x0220
+#define AES_MULTIPURPOSE2_1_OFFSET		0x0224
+#define AES_MULTIPURPOSE2_2_OFFSET		0x0228
+#define AES_MULTIPURPOSE2_3_OFFSET		0x022C
+#define AES_BYTE_ORDER_CFG_OFFSET		0x02C0
+#define AES_TLEN_OFFSET				0x0300
+#define AES_T_MAC_0_OFFSET			0x0304
+#define AES_T_MAC_1_OFFSET			0x0308
+#define AES_T_MAC_2_OFFSET			0x030C
+#define AES_T_MAC_3_OFFSET			0x0310
+#define AES_PLEN_OFFSET				0x0314
+#define AES_A_DMA_SRC_ADDR_OFFSET		0x0400
+#define AES_A_DMA_DST_ADDR_OFFSET		0x0404
+#define AES_A_DMA_SRC_SIZE_OFFSET		0x0408
+#define AES_A_DMA_DST_SIZE_OFFSET		0x040C
+#define AES_A_DMA_DMA_MODE_OFFSET		0x0410
+#define AES_A_DMA_NEXT_SRC_DESCR_OFFSET		0x0418
+#define AES_A_DMA_NEXT_DST_DESCR_OFFSET		0x041C
+#define AES_A_DMA_WHILE_ACTIVE_MODE_OFFSET	0x0420
+#define AES_A_DMA_LOG_OFFSET			0x0424
+#define AES_A_DMA_STATUS_OFFSET			0x0428
+#define AES_A_DMA_PERF_CNTR_OFFSET		0x042C
+#define AES_A_DMA_MSI_ISR_OFFSET		0x0480
+#define AES_A_DMA_MSI_IER_OFFSET		0x0484
+#define AES_A_DMA_MSI_MASK_OFFSET		0x0488
+#define AES_A_DMA_INBUFFER_WRITE_FIFO_OFFSET	0x0600
+#define AES_A_DMA_OUTBUFFER_READ_FIFO_OFFSET	0x0700
+
+/*
+ * AES_A_DMA_DMA_MODE register.
+ * Default: 0x00000000.
+ * bit[31]	ACTIVE
+ *		This bit activates the DMA. When the DMA finishes, it resets
+ *		this bit to zero.
+ * bit[30:26]	Unused by this driver.
+ * bit[25]	SRC_LINK_LIST_EN
+ *		Source link list enable bit. When the linked list is terminated
+ *		this bit is reset by the DMA.
+ * bit[24]	DST_LINK_LIST_EN
+ *		Destination link list enable bit. When the linked list is
+ *		terminated this bit is reset by the DMA.
+ * bit[23:0]	Unused by this driver.
+ */
+#define AES_A_DMA_DMA_MODE_ACTIVE		BIT(31)
+#define AES_A_DMA_DMA_MODE_SRC_LINK_LIST_EN	BIT(25)
+#define AES_A_DMA_DMA_MODE_DST_LINK_LIST_EN	BIT(24)
+
+/*
+ * AES_ACTIVE register
+ * default 0x00000000
+ * bit[31:10]	Reserved
+ * bit[9]	LAST_ADATA
+ * bit[8]	LAST_GCX
+ * bit[7:2]	Reserved
+ * bit[1]	TERMINATION
+ * bit[0]	TRIGGER
+ */
+#define AES_ACTIVE_LAST_ADATA			BIT(9)
+#define AES_ACTIVE_LAST_CCM_GCM			BIT(8)
+#define AES_ACTIVE_TERMINATION			BIT(1)
+#define AES_ACTIVE_TRIGGER			BIT(0)
+
+#define AES_DISABLE_INT				0x00000000
+#define AES_DMA_CPD_ERR_INT			BIT(8)
+#define AES_DMA_OUTBUF_RD_ERR_INT		BIT(7)
+#define AES_DMA_OUTBUF_WR_ERR_INT		BIT(6)
+#define AES_DMA_INBUF_RD_ERR_INT		BIT(5)
+#define AES_DMA_INBUF_WR_ERR_INT		BIT(4)
+#define AES_DMA_BAD_COMP_INT			BIT(3)
+#define AES_DMA_SAI_INT				BIT(2)
+#define AES_DMA_SRC_DONE_INT			BIT(0)
+#define AES_COMPLETE_INT			BIT(1)
+
+#define AES_DMA_MSI_MASK_CLEAR			BIT(0)
+
+#define AES_128_BIT_KEY				0x00000000
+#define AES_256_BIT_KEY				BIT(0)
+
+#define AES_DEACTIVATE_PERF_CNTR		0x00000000
+#define AES_ACTIVATE_PERF_CNTR			BIT(0)
+
+#define AES_MAX_TAG_SIZE_U32			4
+
+#define OCS_LL_DMA_FLAG_TERMINATE		BIT(31)
+
+/*
+ * There is an inconsistency in the documentation. This is documented as a
+ * 11-bit value, but it is actually 10-bits.
+ */
+#define AES_DMA_STATUS_INPUT_BUFFER_OCCUPANCY_MASK	0x3FF
+
+/*
+ * During CCM decrypt, the OCS block needs to finish processing the ciphertext
+ * before the tag is written. For 128-bit mode this required delay is 28 OCS
+ * clock cycles. For 256-bit mode it is 36 OCS clock cycles.
+ */
+#define CCM_DECRYPT_DELAY_TAG_CLK_COUNT		36UL
+
+/*
+ * During CCM decrypt there must be a delay of at least 42 OCS clock cycles
+ * between setting the TRIGGER bit in AES_ACTIVE and setting the LAST_CCM_GCM
+ * bit in the same register (as stated in the OCS databook)
+ */
+#define CCM_DECRYPT_DELAY_LAST_GCX_CLK_COUNT	42UL
+
+/* See RFC3610 section 2.2 */
+#define L_PRIME_MIN (1)
+#define L_PRIME_MAX (7)
+/*
+ * CCM IV format from RFC 3610 section 2.3
+ *
+ *   Octet Number   Contents
+ *   ------------   ---------
+ *   0              Flags
+ *   1 ... 15-L     Nonce N
+ *   16-L ... 15    Counter i
+ *
+ * Flags = L' = L - 1
+ */
+#define L_PRIME_IDX		0
+#define COUNTER_START(lprime)	(16 - ((lprime) + 1))
+#define COUNTER_LEN(lprime)	((lprime) + 1)
+
+enum aes_counter_mode {
+	AES_CTR_M_NO_INC = 0,
+	AES_CTR_M_32_INC = 1,
+	AES_CTR_M_64_INC = 2,
+	AES_CTR_M_128_INC = 3,
+};
+
+/**
+ * struct ocs_dma_linked_list - OCS DMA linked list entry.
+ * @src_addr:   Source address of the data.
+ * @src_len:    Length of data to be fetched.
+ * @next:	Next dma_list to fetch.
+ * @ll_flags:   Flags (Freeze @ terminate) for the DMA engine.
+ */
+struct ocs_dma_linked_list {
+	u32 src_addr;
+	u32 src_len;
+	u32 next;
+	u32 ll_flags;
+} __packed;
+
+/*
+ * Set endianness of inputs and outputs
+ * AES_BYTE_ORDER_CFG
+ * default 0x00000000
+ * bit [10] - KEY_HI_LO_SWAP
+ * bit [9] - KEY_HI_SWAP_DWORDS_IN_OCTWORD
+ * bit [8] - KEY_HI_SWAP_BYTES_IN_DWORD
+ * bit [7] - KEY_LO_SWAP_DWORDS_IN_OCTWORD
+ * bit [6] - KEY_LO_SWAP_BYTES_IN_DWORD
+ * bit [5] - IV_SWAP_DWORDS_IN_OCTWORD
+ * bit [4] - IV_SWAP_BYTES_IN_DWORD
+ * bit [3] - DOUT_SWAP_DWORDS_IN_OCTWORD
+ * bit [2] - DOUT_SWAP_BYTES_IN_DWORD
+ * bit [1] - DOUT_SWAP_DWORDS_IN_OCTWORD
+ * bit [0] - DOUT_SWAP_BYTES_IN_DWORD
+ */
+static inline void aes_a_set_endianness(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(0x7FF, aes_dev->base_reg + AES_BYTE_ORDER_CFG_OFFSET);
+}
+
+/* Trigger AES process start. */
+static inline void aes_a_op_trigger(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_ACTIVE_TRIGGER, aes_dev->base_reg + AES_ACTIVE_OFFSET);
+}
+
+/* Indicate last bulk of data. */
+static inline void aes_a_op_termination(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_ACTIVE_TERMINATION,
+		  aes_dev->base_reg + AES_ACTIVE_OFFSET);
+}
+
+/*
+ * Set LAST_CCM_GCM in AES_ACTIVE register and clear all other bits.
+ *
+ * Called when DMA is programmed to fetch the last batch of data.
+ * - For AES-CCM it is called for the last batch of Payload data and Ciphertext
+ *   data.
+ * - For AES-GCM, it is called for the last batch of Plaintext data and
+ *   Ciphertext data.
+ */
+static inline void aes_a_set_last_gcx(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_ACTIVE_LAST_CCM_GCM,
+		  aes_dev->base_reg + AES_ACTIVE_OFFSET);
+}
+
+/* Wait for LAST_CCM_GCM bit to be unset. */
+static inline void aes_a_wait_last_gcx(const struct ocs_aes_dev *aes_dev)
+{
+	u32 aes_active_reg;
+
+	do {
+		aes_active_reg = ioread32(aes_dev->base_reg +
+					  AES_ACTIVE_OFFSET);
+	} while (aes_active_reg & AES_ACTIVE_LAST_CCM_GCM);
+}
+
+/* Wait for 10 bits of input occupancy. */
+static void aes_a_dma_wait_input_buffer_occupancy(const struct ocs_aes_dev *aes_dev)
+{
+	u32 reg;
+
+	do {
+		reg = ioread32(aes_dev->base_reg + AES_A_DMA_STATUS_OFFSET);
+	} while (reg & AES_DMA_STATUS_INPUT_BUFFER_OCCUPANCY_MASK);
+}
+
+ /*
+  * Set LAST_CCM_GCM and LAST_ADATA bits in AES_ACTIVE register (and clear all
+  * other bits).
+  *
+  * Called when DMA is programmed to fetch the last batch of Associated Data
+  * (CCM case) or Additional Authenticated Data (GCM case).
+  */
+static inline void aes_a_set_last_gcx_and_adata(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_ACTIVE_LAST_ADATA | AES_ACTIVE_LAST_CCM_GCM,
+		  aes_dev->base_reg + AES_ACTIVE_OFFSET);
+}
+
+/* Set DMA src and dst transfer size to 0 */
+static inline void aes_a_dma_set_xfer_size_zero(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(0, aes_dev->base_reg + AES_A_DMA_SRC_SIZE_OFFSET);
+	iowrite32(0, aes_dev->base_reg + AES_A_DMA_DST_SIZE_OFFSET);
+}
+
+/* Activate DMA for zero-byte transfer case. */
+static inline void aes_a_dma_active(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_A_DMA_DMA_MODE_ACTIVE,
+		  aes_dev->base_reg + AES_A_DMA_DMA_MODE_OFFSET);
+}
+
+/* Activate DMA and enable src linked list */
+static inline void aes_a_dma_active_src_ll_en(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_A_DMA_DMA_MODE_ACTIVE |
+		  AES_A_DMA_DMA_MODE_SRC_LINK_LIST_EN,
+		  aes_dev->base_reg + AES_A_DMA_DMA_MODE_OFFSET);
+}
+
+/* Activate DMA and enable dst linked list */
+static inline void aes_a_dma_active_dst_ll_en(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_A_DMA_DMA_MODE_ACTIVE |
+		  AES_A_DMA_DMA_MODE_DST_LINK_LIST_EN,
+		  aes_dev->base_reg + AES_A_DMA_DMA_MODE_OFFSET);
+}
+
+/* Activate DMA and enable src and dst linked lists */
+static inline void aes_a_dma_active_src_dst_ll_en(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(AES_A_DMA_DMA_MODE_ACTIVE |
+		  AES_A_DMA_DMA_MODE_SRC_LINK_LIST_EN |
+		  AES_A_DMA_DMA_MODE_DST_LINK_LIST_EN,
+		  aes_dev->base_reg + AES_A_DMA_DMA_MODE_OFFSET);
+}
+
+/* Reset PERF_CNTR to 0 and activate it */
+static inline void aes_a_dma_reset_and_activate_perf_cntr(const struct ocs_aes_dev *aes_dev)
+{
+	iowrite32(0x00000000, aes_dev->base_reg + AES_A_DMA_PERF_CNTR_OFFSET);
+	iowrite32(AES_ACTIVATE_PERF_CNTR,
+		  aes_dev->base_reg + AES_A_DMA_WHILE_ACTIVE_MODE_OFFSET);
+}
+
+/* Wait until PERF_CNTR is > delay, then deactivate it */
+static inline void aes_a_dma_wait_and_deactivate_perf_cntr(const struct ocs_aes_dev *aes_dev,
+							   int delay)
+{
+	while (ioread32(aes_dev->base_reg + AES_A_DMA_PERF_CNTR_OFFSET) < delay)
+		;
+	iowrite32(AES_DEACTIVATE_PERF_CNTR,
+		  aes_dev->base_reg + AES_A_DMA_WHILE_ACTIVE_MODE_OFFSET);
+}
+
+/* Disable AES and DMA IRQ. */
+static void aes_irq_disable(struct ocs_aes_dev *aes_dev)
+{
+	u32 isr_val = 0;
+
+	/* Disable interrupts */
+	iowrite32(AES_DISABLE_INT,
+		  aes_dev->base_reg + AES_A_DMA_MSI_IER_OFFSET);
+	iowrite32(AES_DISABLE_INT, aes_dev->base_reg + AES_IER_OFFSET);
+
+	/* Clear any pending interrupt */
+	isr_val = ioread32(aes_dev->base_reg + AES_A_DMA_MSI_ISR_OFFSET);
+	if (isr_val)
+		iowrite32(isr_val,
+			  aes_dev->base_reg + AES_A_DMA_MSI_ISR_OFFSET);
+
+	isr_val = ioread32(aes_dev->base_reg + AES_A_DMA_MSI_MASK_OFFSET);
+	if (isr_val)
+		iowrite32(isr_val,
+			  aes_dev->base_reg + AES_A_DMA_MSI_MASK_OFFSET);
+
+	isr_val = ioread32(aes_dev->base_reg + AES_ISR_OFFSET);
+	if (isr_val)
+		iowrite32(isr_val, aes_dev->base_reg + AES_ISR_OFFSET);
+}
+
+/* Enable AES or DMA IRQ.  IRQ is disabled once fired. */
+static void aes_irq_enable(struct ocs_aes_dev *aes_dev, u8 irq)
+{
+	if (irq == AES_COMPLETE_INT) {
+		/* Ensure DMA error interrupts are enabled */
+		iowrite32(AES_DMA_CPD_ERR_INT |
+			  AES_DMA_OUTBUF_RD_ERR_INT |
+			  AES_DMA_OUTBUF_WR_ERR_INT |
+			  AES_DMA_INBUF_RD_ERR_INT |
+			  AES_DMA_INBUF_WR_ERR_INT |
+			  AES_DMA_BAD_COMP_INT |
+			  AES_DMA_SAI_INT,
+			  aes_dev->base_reg + AES_A_DMA_MSI_IER_OFFSET);
+		/*
+		 * AES_IER
+		 * default 0x00000000
+		 * bits [31:3] - reserved
+		 * bit [2] - EN_SKS_ERR
+		 * bit [1] - EN_AES_COMPLETE
+		 * bit [0] - reserved
+		 */
+		iowrite32(AES_COMPLETE_INT, aes_dev->base_reg + AES_IER_OFFSET);
+		return;
+	}
+	if (irq == AES_DMA_SRC_DONE_INT) {
+		/* Ensure AES interrupts are disabled */
+		iowrite32(AES_DISABLE_INT, aes_dev->base_reg + AES_IER_OFFSET);
+		/*
+		 * DMA_MSI_IER
+		 * default 0x00000000
+		 * bits [31:9] - reserved
+		 * bit [8] - CPD_ERR_INT_EN
+		 * bit [7] - OUTBUF_RD_ERR_INT_EN
+		 * bit [6] - OUTBUF_WR_ERR_INT_EN
+		 * bit [5] - INBUF_RD_ERR_INT_EN
+		 * bit [4] - INBUF_WR_ERR_INT_EN
+		 * bit [3] - BAD_COMP_INT_EN
+		 * bit [2] - SAI_INT_EN
+		 * bit [1] - DST_DONE_INT_EN
+		 * bit [0] - SRC_DONE_INT_EN
+		 */
+		iowrite32(AES_DMA_CPD_ERR_INT |
+			  AES_DMA_OUTBUF_RD_ERR_INT |
+			  AES_DMA_OUTBUF_WR_ERR_INT |
+			  AES_DMA_INBUF_RD_ERR_INT |
+			  AES_DMA_INBUF_WR_ERR_INT |
+			  AES_DMA_BAD_COMP_INT |
+			  AES_DMA_SAI_INT |
+			  AES_DMA_SRC_DONE_INT,
+			  aes_dev->base_reg + AES_A_DMA_MSI_IER_OFFSET);
+	}
+}
+
+/* Enable and wait for IRQ (either from OCS AES engine or DMA) */
+static int ocs_aes_irq_enable_and_wait(struct ocs_aes_dev *aes_dev, u8 irq)
+{
+	int rc;
+
+	reinit_completion(&aes_dev->irq_completion);
+	aes_irq_enable(aes_dev, irq);
+	rc = wait_for_completion_interruptible(&aes_dev->irq_completion);
+	if (rc)
+		return rc;
+
+	return aes_dev->dma_err_mask ? -EIO : 0;
+}
+
+/* Configure DMA to OCS, linked list mode */
+static inline void dma_to_ocs_aes_ll(struct ocs_aes_dev *aes_dev,
+				     dma_addr_t dma_list)
+{
+	iowrite32(0, aes_dev->base_reg + AES_A_DMA_SRC_SIZE_OFFSET);
+	iowrite32(dma_list,
+		  aes_dev->base_reg + AES_A_DMA_NEXT_SRC_DESCR_OFFSET);
+}
+
+/* Configure DMA from OCS, linked list mode */
+static inline void dma_from_ocs_aes_ll(struct ocs_aes_dev *aes_dev,
+				       dma_addr_t dma_list)
+{
+	iowrite32(0, aes_dev->base_reg + AES_A_DMA_DST_SIZE_OFFSET);
+	iowrite32(dma_list,
+		  aes_dev->base_reg + AES_A_DMA_NEXT_DST_DESCR_OFFSET);
+}
+
+irqreturn_t ocs_aes_irq_handler(int irq, void *dev_id)
+{
+	struct ocs_aes_dev *aes_dev = dev_id;
+	u32 aes_dma_isr;
+
+	/* Read DMA ISR status. */
+	aes_dma_isr = ioread32(aes_dev->base_reg + AES_A_DMA_MSI_ISR_OFFSET);
+
+	/* Disable and clear interrupts. */
+	aes_irq_disable(aes_dev);
+
+	/* Save DMA error status. */
+	aes_dev->dma_err_mask = aes_dma_isr &
+				(AES_DMA_CPD_ERR_INT |
+				 AES_DMA_OUTBUF_RD_ERR_INT |
+				 AES_DMA_OUTBUF_WR_ERR_INT |
+				 AES_DMA_INBUF_RD_ERR_INT |
+				 AES_DMA_INBUF_WR_ERR_INT |
+				 AES_DMA_BAD_COMP_INT |
+				 AES_DMA_SAI_INT);
+
+	/* Signal IRQ completion. */
+	complete(&aes_dev->irq_completion);
+
+	return IRQ_HANDLED;
+}
+
+/**
+ * ocs_aes_set_key() - Write key into OCS AES hardware.
+ * @aes_dev:	The OCS AES device to write the key to.
+ * @key_size:	The size of the key (in bytes).
+ * @key:	The key to write.
+ * @cipher:	The cipher the key is for.
+ *
+ * For AES @key_size must be either 16 or 32. For SM4 @key_size must be 16.
+ *
+ * Return:	0 on success, negative error code otherwise.
+ */
+int ocs_aes_set_key(struct ocs_aes_dev *aes_dev, u32 key_size, const u8 *key,
+		    enum ocs_cipher cipher)
+{
+	const u32 *key_u32;
+	u32 val;
+	int i;
+
+	/* OCS AES supports 128-bit and 256-bit keys only. */
+	if (cipher == OCS_AES && !(key_size == 32 || key_size == 16)) {
+		dev_err(aes_dev->dev,
+			"%d-bit keys not supported by AES cipher\n",
+			key_size * 8);
+		return -EINVAL;
+	}
+	/* OCS SM4 supports 128-bit keys only. */
+	if (cipher == OCS_SM4 && key_size != 16) {
+		dev_err(aes_dev->dev,
+			"%d-bit keys not supported for SM4 cipher\n",
+			key_size * 8);
+		return -EINVAL;
+	}
+
+	if (!key)
+		return -EINVAL;
+
+	key_u32 = (const u32 *)key;
+
+	/* Write key to AES_KEY[0-7] registers */
+	for (i = 0; i < (key_size / sizeof(u32)); i++) {
+		iowrite32(key_u32[i],
+			  aes_dev->base_reg + AES_KEY_0_OFFSET +
+			  (i * sizeof(u32)));
+	}
+	/*
+	 * Write key size
+	 * bits [31:1] - reserved
+	 * bit [0] - AES_KEY_SIZE
+	 *           0 - 128 bit key
+	 *           1 - 256 bit key
+	 */
+	val = (key_size == 16) ? AES_128_BIT_KEY : AES_256_BIT_KEY;
+	iowrite32(val, aes_dev->base_reg + AES_KEY_SIZE_OFFSET);
+
+	return 0;
+}
+
+/* Write AES_COMMAND */
+static inline void set_ocs_aes_command(struct ocs_aes_dev *aes_dev,
+				       enum ocs_cipher cipher,
+				       enum ocs_mode mode,
+				       enum ocs_instruction instruction)
+{
+	u32 val;
+
+	/* AES_COMMAND
+	 * default 0x000000CC
+	 * bit [14] - CIPHER_SELECT
+	 *            0 - AES
+	 *            1 - SM4
+	 * bits [11:8] - OCS_AES_MODE
+	 *               0000 - ECB
+	 *               0001 - CBC
+	 *               0010 - CTR
+	 *               0110 - CCM
+	 *               0111 - GCM
+	 *               1001 - CTS
+	 * bits [7:6] - AES_INSTRUCTION
+	 *              00 - ENCRYPT
+	 *              01 - DECRYPT
+	 *              10 - EXPAND
+	 *              11 - BYPASS
+	 * bits [3:2] - CTR_M_BITS
+	 *              00 - No increment
+	 *              01 - Least significant 32 bits are incremented
+	 *              10 - Least significant 64 bits are incremented
+	 *              11 - Full 128 bits are incremented
+	 */
+	val = (cipher << 14) | (mode << 8) | (instruction << 6) |
+	      (AES_CTR_M_128_INC << 2);
+	iowrite32(val, aes_dev->base_reg + AES_COMMAND_OFFSET);
+}
+
+static void ocs_aes_init(struct ocs_aes_dev *aes_dev,
+			 enum ocs_mode mode,
+			 enum ocs_cipher cipher,
+			 enum ocs_instruction instruction)
+{
+	/* Ensure interrupts are disabled and pending interrupts cleared. */
+	aes_irq_disable(aes_dev);
+
+	/* Set endianness recommended by data-sheet. */
+	aes_a_set_endianness(aes_dev);
+
+	/* Set AES_COMMAND register. */
+	set_ocs_aes_command(aes_dev, cipher, mode, instruction);
+}
+
+/*
+ * Write the byte length of the last AES/SM4 block of Payload data (without
+ * zero padding and without the length of the MAC) in register AES_PLEN.
+ */
+static inline void ocs_aes_write_last_data_blk_len(struct ocs_aes_dev *aes_dev,
+						   u32 size)
+{
+	u32 val;
+
+	if (size == 0) {
+		val = 0;
+		goto exit;
+	}
+
+	val = size % AES_BLOCK_SIZE;
+	if (val == 0)
+		val = AES_BLOCK_SIZE;
+
+exit:
+	iowrite32(val, aes_dev->base_reg + AES_PLEN_OFFSET);
+}
+
+/*
+ * Validate inputs according to mode.
+ * If OK return 0; else return -EINVAL.
+ */
+static int ocs_aes_validate_inputs(dma_addr_t src_dma_list, u32 src_size,
+				   const u8 *iv, u32 iv_size,
+				   dma_addr_t aad_dma_list, u32 aad_size,
+				   const u8 *tag, u32 tag_size,
+				   enum ocs_cipher cipher, enum ocs_mode mode,
+				   enum ocs_instruction instruction,
+				   dma_addr_t dst_dma_list)
+{
+	/* Ensure cipher, mode and instruction are valid. */
+	if (!(cipher == OCS_AES || cipher == OCS_SM4))
+		return -EINVAL;
+
+	if (mode != OCS_MODE_ECB && mode != OCS_MODE_CBC &&
+	    mode != OCS_MODE_CTR && mode != OCS_MODE_CCM &&
+	    mode != OCS_MODE_GCM && mode != OCS_MODE_CTS)
+		return -EINVAL;
+
+	if (instruction != OCS_ENCRYPT && instruction != OCS_DECRYPT &&
+	    instruction != OCS_EXPAND  && instruction != OCS_BYPASS)
+		return -EINVAL;
+
+	/*
+	 * When instruction is OCS_BYPASS, OCS simply copies data from source
+	 * to destination using DMA.
+	 *
+	 * AES mode is irrelevant, but both source and destination DMA
+	 * linked-list must be defined.
+	 */
+	if (instruction == OCS_BYPASS) {
+		if (src_dma_list == DMA_MAPPING_ERROR ||
+		    dst_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		return 0;
+	}
+
+	/*
+	 * For performance reasons switch based on mode to limit unnecessary
+	 * conditionals for each mode
+	 */
+	switch (mode) {
+	case OCS_MODE_ECB:
+		/* Ensure input length is multiple of block size */
+		if (src_size % AES_BLOCK_SIZE != 0)
+			return -EINVAL;
+
+		/* Ensure source and destination linked lists are created */
+		if (src_dma_list == DMA_MAPPING_ERROR ||
+		    dst_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_CBC:
+		/* Ensure input length is multiple of block size */
+		if (src_size % AES_BLOCK_SIZE != 0)
+			return -EINVAL;
+
+		/* Ensure source and destination linked lists are created */
+		if (src_dma_list == DMA_MAPPING_ERROR ||
+		    dst_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		/* Ensure IV is present and block size in length */
+		if (!iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_CTR:
+		/* Ensure input length of 1 byte or greater */
+		if (src_size == 0)
+			return -EINVAL;
+
+		/* Ensure source and destination linked lists are created */
+		if (src_dma_list == DMA_MAPPING_ERROR ||
+		    dst_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		/* Ensure IV is present and block size in length */
+		if (!iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_CTS:
+		/* Ensure input length >= block size */
+		if (src_size < AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		/* Ensure source and destination linked lists are created */
+		if (src_dma_list == DMA_MAPPING_ERROR ||
+		    dst_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		/* Ensure IV is present and block size in length */
+		if (!iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_GCM:
+		/* Ensure IV is present and GCM_AES_IV_SIZE in length */
+		if (!iv || iv_size != GCM_AES_IV_SIZE)
+			return -EINVAL;
+
+		/*
+		 * If input data present ensure source and destination linked
+		 * lists are created
+		 */
+		if (src_size && (src_dma_list == DMA_MAPPING_ERROR ||
+				 dst_dma_list == DMA_MAPPING_ERROR))
+			return -EINVAL;
+
+		/* If aad present ensure aad linked list is created */
+		if (aad_size && aad_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		/* Ensure tag destination is set */
+		if (!tag)
+			return -EINVAL;
+
+		/* Just ensure that tag_size doesn't cause overflows. */
+		if (tag_size > (AES_MAX_TAG_SIZE_U32 * sizeof(u32)))
+			return -EINVAL;
+
+		return 0;
+
+	case OCS_MODE_CCM:
+		/* Ensure IV is present and block size in length */
+		if (!iv || iv_size != AES_BLOCK_SIZE)
+			return -EINVAL;
+
+		/* 2 <= L <= 8, so 1 <= L' <= 7 */
+		if (iv[L_PRIME_IDX] < L_PRIME_MIN ||
+		    iv[L_PRIME_IDX] > L_PRIME_MAX)
+			return -EINVAL;
+
+		/* If aad present ensure aad linked list is created */
+		if (aad_size && aad_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		/* Just ensure that tag_size doesn't cause overflows. */
+		if (tag_size > (AES_MAX_TAG_SIZE_U32 * sizeof(u32)))
+			return -EINVAL;
+
+		if (instruction == OCS_DECRYPT) {
+			/*
+			 * If input data present ensure source and destination
+			 * linked lists are created
+			 */
+			if (src_size && (src_dma_list == DMA_MAPPING_ERROR ||
+					 dst_dma_list == DMA_MAPPING_ERROR))
+				return -EINVAL;
+
+			/* Ensure input tag is present */
+			if (!tag)
+				return -EINVAL;
+
+			return 0;
+		}
+
+		/* Instruction == OCS_ENCRYPT */
+
+		/*
+		 * Destination linked list always required (for tag even if no
+		 * input data)
+		 */
+		if (dst_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		/* If input data present ensure src linked list is created */
+		if (src_size && src_dma_list == DMA_MAPPING_ERROR)
+			return -EINVAL;
+
+		return 0;
+
+	default:
+		return -EINVAL;
+	}
+}
+
+/**
+ * ocs_aes_op() - Perform AES/SM4 operation.
+ * @aes_dev:		The OCS AES device to use.
+ * @mode:		The mode to use (ECB, CBC, CTR, or CTS).
+ * @cipher:		The cipher to use (AES or SM4).
+ * @instruction:	The instruction to perform (encrypt or decrypt).
+ * @dst_dma_list:	The OCS DMA list mapping output memory.
+ * @src_dma_list:	The OCS DMA list mapping input payload data.
+ * @src_size:		The amount of data mapped by @src_dma_list.
+ * @iv:			The IV vector.
+ * @iv_size:		The size (in bytes) of @iv.
+ *
+ * Return: 0 on success, negative error code otherwise.
+ */
+int ocs_aes_op(struct ocs_aes_dev *aes_dev,
+	       enum ocs_mode mode,
+	       enum ocs_cipher cipher,
+	       enum ocs_instruction instruction,
+	       dma_addr_t dst_dma_list,
+	       dma_addr_t src_dma_list,
+	       u32 src_size,
+	       u8 *iv,
+	       u32 iv_size)
+{
+	u32 *iv32;
+	int rc;
+
+	rc = ocs_aes_validate_inputs(src_dma_list, src_size, iv, iv_size, 0, 0,
+				     NULL, 0, cipher, mode, instruction,
+				     dst_dma_list);
+	if (rc)
+		return rc;
+	/*
+	 * ocs_aes_validate_inputs() is a generic check, now ensure mode is not
+	 * GCM or CCM.
+	 */
+	if (mode == OCS_MODE_GCM || mode == OCS_MODE_CCM)
+		return -EINVAL;
+
+	/* Cast IV to u32 array. */
+	iv32 = (u32 *)iv;
+
+	ocs_aes_init(aes_dev, mode, cipher, instruction);
+
+	if (mode == OCS_MODE_CTS) {
+		/* Write the byte length of the last data block to engine. */
+		ocs_aes_write_last_data_blk_len(aes_dev, src_size);
+	}
+
+	/* ECB is the only mode that doesn't use IV. */
+	if (mode != OCS_MODE_ECB) {
+		iowrite32(iv32[0], aes_dev->base_reg + AES_IV_0_OFFSET);
+		iowrite32(iv32[1], aes_dev->base_reg + AES_IV_1_OFFSET);
+		iowrite32(iv32[2], aes_dev->base_reg + AES_IV_2_OFFSET);
+		iowrite32(iv32[3], aes_dev->base_reg + AES_IV_3_OFFSET);
+	}
+
+	/* Set AES_ACTIVE.TRIGGER to start the operation. */
+	aes_a_op_trigger(aes_dev);
+
+	/* Configure and activate input / output DMA. */
+	dma_to_ocs_aes_ll(aes_dev, src_dma_list);
+	dma_from_ocs_aes_ll(aes_dev, dst_dma_list);
+	aes_a_dma_active_src_dst_ll_en(aes_dev);
+
+	if (mode == OCS_MODE_CTS) {
+		/*
+		 * For CTS mode, instruct engine to activate ciphertext
+		 * stealing if last block of data is incomplete.
+		 */
+		aes_a_set_last_gcx(aes_dev);
+	} else {
+		/* For all other modes, just write the 'termination' bit. */
+		aes_a_op_termination(aes_dev);
+	}
+
+	/* Wait for engine to complete processing. */
+	rc = ocs_aes_irq_enable_and_wait(aes_dev, AES_COMPLETE_INT);
+	if (rc)
+		return rc;
+
+	if (mode == OCS_MODE_CTR) {
+		/* Read back IV for streaming mode */
+		iv32[0] = ioread32(aes_dev->base_reg + AES_IV_0_OFFSET);
+		iv32[1] = ioread32(aes_dev->base_reg + AES_IV_1_OFFSET);
+		iv32[2] = ioread32(aes_dev->base_reg + AES_IV_2_OFFSET);
+		iv32[3] = ioread32(aes_dev->base_reg + AES_IV_3_OFFSET);
+	}
+
+	return 0;
+}
+
+/* Compute and write J0 to engine registers. */
+static void ocs_aes_gcm_write_j0(const struct ocs_aes_dev *aes_dev,
+				 const u8 *iv)
+{
+	const u32 *j0 = (u32 *)iv;
+
+	/*
+	 * IV must be 12 bytes; Other sizes not supported as Linux crypto API
+	 * does only expects/allows 12 byte IV for GCM
+	 */
+	iowrite32(0x00000001, aes_dev->base_reg + AES_IV_0_OFFSET);
+	iowrite32(__swab32(j0[2]), aes_dev->base_reg + AES_IV_1_OFFSET);
+	iowrite32(__swab32(j0[1]), aes_dev->base_reg + AES_IV_2_OFFSET);
+	iowrite32(__swab32(j0[0]), aes_dev->base_reg + AES_IV_3_OFFSET);
+}
+
+/* Read GCM tag from engine registers. */
+static inline void ocs_aes_gcm_read_tag(struct ocs_aes_dev *aes_dev,
+					u8 *tag, u32 tag_size)
+{
+	u32 tag_u32[AES_MAX_TAG_SIZE_U32];
+
+	/*
+	 * The Authentication Tag T is stored in Little Endian order in the
+	 * registers with the most significant bytes stored from AES_T_MAC[3]
+	 * downward.
+	 */
+	tag_u32[0] = __swab32(ioread32(aes_dev->base_reg + AES_T_MAC_3_OFFSET));
+	tag_u32[1] = __swab32(ioread32(aes_dev->base_reg + AES_T_MAC_2_OFFSET));
+	tag_u32[2] = __swab32(ioread32(aes_dev->base_reg + AES_T_MAC_1_OFFSET));
+	tag_u32[3] = __swab32(ioread32(aes_dev->base_reg + AES_T_MAC_0_OFFSET));
+
+	memcpy(tag, tag_u32, tag_size);
+}
+
+/**
+ * ocs_aes_gcm_op() - Perform GCM operation.
+ * @aes_dev:		The OCS AES device to use.
+ * @cipher:		The Cipher to use (AES or SM4).
+ * @instruction:	The instruction to perform (encrypt or decrypt).
+ * @dst_dma_list:	The OCS DMA list mapping output memory.
+ * @src_dma_list:	The OCS DMA list mapping input payload data.
+ * @src_size:		The amount of data mapped by @src_dma_list.
+ * @iv:			The input IV vector.
+ * @aad_dma_list:	The OCS DMA list mapping input AAD data.
+ * @aad_size:		The amount of data mapped by @aad_dma_list.
+ * @out_tag:		Where to store computed tag.
+ * @tag_size:		The size (in bytes) of @out_tag.
+ *
+ * Return: 0 on success, negative error code otherwise.
+ */
+int ocs_aes_gcm_op(struct ocs_aes_dev *aes_dev,
+		   enum ocs_cipher cipher,
+		   enum ocs_instruction instruction,
+		   dma_addr_t dst_dma_list,
+		   dma_addr_t src_dma_list,
+		   u32 src_size,
+		   const u8 *iv,
+		   dma_addr_t aad_dma_list,
+		   u32 aad_size,
+		   u8 *out_tag,
+		   u32 tag_size)
+{
+	u64 bit_len;
+	u32 val;
+	int rc;
+
+	rc = ocs_aes_validate_inputs(src_dma_list, src_size, iv,
+				     GCM_AES_IV_SIZE, aad_dma_list,
+				     aad_size, out_tag, tag_size, cipher,
+				     OCS_MODE_GCM, instruction,
+				     dst_dma_list);
+	if (rc)
+		return rc;
+
+	ocs_aes_init(aes_dev, OCS_MODE_GCM, cipher, instruction);
+
+	/* Compute and write J0 to OCS HW. */
+	ocs_aes_gcm_write_j0(aes_dev, iv);
+
+	/* Write out_tag byte length */
+	iowrite32(tag_size, aes_dev->base_reg + AES_TLEN_OFFSET);
+
+	/* Write the byte length of the last plaintext / ciphertext block. */
+	ocs_aes_write_last_data_blk_len(aes_dev, src_size);
+
+	/* Write ciphertext bit length */
+	bit_len = src_size * 8;
+	val = bit_len & 0xFFFFFFFF;
+	iowrite32(val, aes_dev->base_reg + AES_MULTIPURPOSE2_0_OFFSET);
+	val = bit_len >> 32;
+	iowrite32(val, aes_dev->base_reg + AES_MULTIPURPOSE2_1_OFFSET);
+
+	/* Write aad bit length */
+	bit_len = aad_size * 8;
+	val = bit_len & 0xFFFFFFFF;
+	iowrite32(val, aes_dev->base_reg + AES_MULTIPURPOSE2_2_OFFSET);
+	val = bit_len >> 32;
+	iowrite32(val, aes_dev->base_reg + AES_MULTIPURPOSE2_3_OFFSET);
+
+	/* Set AES_ACTIVE.TRIGGER to start the operation. */
+	aes_a_op_trigger(aes_dev);
+
+	/* Process AAD. */
+	if (aad_size) {
+		/* If aad present, configure DMA to feed it to the engine. */
+		dma_to_ocs_aes_ll(aes_dev, aad_dma_list);
+		aes_a_dma_active_src_ll_en(aes_dev);
+
+		/* Instructs engine to pad last block of aad, if needed. */
+		aes_a_set_last_gcx_and_adata(aes_dev);
+
+		/* Wait for DMA transfer to complete. */
+		rc = ocs_aes_irq_enable_and_wait(aes_dev, AES_DMA_SRC_DONE_INT);
+		if (rc)
+			return rc;
+	} else {
+		aes_a_set_last_gcx_and_adata(aes_dev);
+	}
+
+	/* Wait until adata (if present) has been processed. */
+	aes_a_wait_last_gcx(aes_dev);
+	aes_a_dma_wait_input_buffer_occupancy(aes_dev);
+
+	/* Now process payload. */
+	if (src_size) {
+		/* Configure and activate DMA for both input and output data. */
+		dma_to_ocs_aes_ll(aes_dev, src_dma_list);
+		dma_from_ocs_aes_ll(aes_dev, dst_dma_list);
+		aes_a_dma_active_src_dst_ll_en(aes_dev);
+	} else {
+		aes_a_dma_set_xfer_size_zero(aes_dev);
+		aes_a_dma_active(aes_dev);
+	}
+
+	/* Instruct AES/SMA4 engine payload processing is over. */
+	aes_a_set_last_gcx(aes_dev);
+
+	/* Wait for OCS AES engine to complete processing. */
+	rc = ocs_aes_irq_enable_and_wait(aes_dev, AES_COMPLETE_INT);
+	if (rc)
+		return rc;
+
+	ocs_aes_gcm_read_tag(aes_dev, out_tag, tag_size);
+
+	return 0;
+}
+
+/* Write encrypted tag to AES/SM4 engine. */
+static void ocs_aes_ccm_write_encrypted_tag(struct ocs_aes_dev *aes_dev,
+					    const u8 *in_tag, u32 tag_size)
+{
+	int i;
+
+	/* Ensure DMA input buffer is empty */
+	aes_a_dma_wait_input_buffer_occupancy(aes_dev);
+
+	/*
+	 * During CCM decrypt, the OCS block needs to finish processing the
+	 * ciphertext before the tag is written.  So delay needed after DMA has
+	 * completed writing the ciphertext
+	 */
+	aes_a_dma_reset_and_activate_perf_cntr(aes_dev);
+	aes_a_dma_wait_and_deactivate_perf_cntr(aes_dev,
+						CCM_DECRYPT_DELAY_TAG_CLK_COUNT);
+
+	/* Write encrypted tag to AES/SM4 engine. */
+	for (i = 0; i < tag_size; i++) {
+		iowrite8(in_tag[i], aes_dev->base_reg +
+				    AES_A_DMA_INBUFFER_WRITE_FIFO_OFFSET);
+	}
+}
+
+/*
+ * Write B0 CCM block to OCS AES HW.
+ *
+ * Note: B0 format is documented in NIST Special Publication 800-38C
+ * https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-38c.pdf
+ * (see Section A.2.1)
+ */
+static int ocs_aes_ccm_write_b0(const struct ocs_aes_dev *aes_dev,
+				const u8 *iv, u32 adata_size, u32 tag_size,
+				u32 cryptlen)
+{
+	u8 b0[16]; /* CCM B0 block is 16 bytes long. */
+	int i, q;
+
+	/* Initialize B0 to 0. */
+	memset(b0, 0, sizeof(b0));
+
+	/*
+	 * B0[0] is the 'Flags Octet' and has the following structure:
+	 *   bit 7: Reserved
+	 *   bit 6: Adata flag
+	 *   bit 5-3: t value encoded as (t-2)/2
+	 *   bit 2-0: q value encoded as q - 1
+	 */
+	/* If there is AAD data, set the Adata flag. */
+	if (adata_size)
+		b0[0] |= BIT(6);
+	/*
+	 * t denotes the octet length of T.
+	 * t can only be an element of { 4, 6, 8, 10, 12, 14, 16} and is
+	 * encoded as (t - 2) / 2
+	 */
+	b0[0] |= (((tag_size - 2) / 2) & 0x7)  << 3;
+	/*
+	 * q is the octet length of Q.
+	 * q can only be an element of {2, 3, 4, 5, 6, 7, 8} and is encoded as
+	 * q - 1 == iv[0]
+	 */
+	b0[0] |= iv[0] & 0x7;
+	/*
+	 * Copy the Nonce N from IV to B0; N is located in iv[1]..iv[15 - q]
+	 * and must be copied to b0[1]..b0[15-q].
+	 * q == iv[0] + 1
+	 */
+	q = iv[0] + 1;
+	for (i = 1; i <= 15 - q; i++)
+		b0[i] = iv[i];
+	/*
+	 * The rest of B0 must contain Q, i.e., the message length.
+	 * Q is encoded in q octets, in big-endian order, so to write it, we
+	 * start from the end of B0 and we move backward.
+	 */
+	i = sizeof(b0) - 1;
+	while (q) {
+		b0[i] = cryptlen & 0xff;
+		cryptlen >>= 8;
+		i--;
+		q--;
+	}
+	/*
+	 * If cryptlen is not zero at this point, it means that its original
+	 * value was too big.
+	 */
+	if (cryptlen)
+		return -EOVERFLOW;
+	/* Now write B0 to OCS AES input buffer. */
+	for (i = 0; i < sizeof(b0); i++)
+		iowrite8(b0[i], aes_dev->base_reg +
+				AES_A_DMA_INBUFFER_WRITE_FIFO_OFFSET);
+	return 0;
+}
+
+/*
+ * Write adata length to OCS AES HW.
+ *
+ * Note: adata len encoding is documented in NIST Special Publication 800-38C
+ * https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-38c.pdf
+ * (see Section A.2.2)
+ */
+static void ocs_aes_ccm_write_adata_len(const struct ocs_aes_dev *aes_dev,
+					u64 adata_len)
+{
+	u8 enc_a[10]; /* Maximum encoded size: 10 octets. */
+	int i, len;
+
+	/*
+	 * adata_len ('a') is encoded as follows:
+	 * If 0 < a < 2^16 - 2^8    ==> 'a' encoded as [a]16, i.e., two octets
+	 *				(big endian).
+	 * If 2^16 - 2^8 ≤ a < 2^32 ==> 'a' encoded as 0xff || 0xfe || [a]32,
+	 *				i.e., six octets (big endian).
+	 * If 2^32 ≤ a < 2^64       ==> 'a' encoded as 0xff || 0xff || [a]64,
+	 *				i.e., ten octets (big endian).
+	 */
+	if (adata_len < 65280) {
+		len = 2;
+		*(__be16 *)enc_a = cpu_to_be16(adata_len);
+	} else if (adata_len <= 0xFFFFFFFF) {
+		len = 6;
+		*(__be16 *)enc_a = cpu_to_be16(0xfffe);
+		*(__be32 *)&enc_a[2] = cpu_to_be32(adata_len);
+	} else { /* adata_len >= 2^32 */
+		len = 10;
+		*(__be16 *)enc_a = cpu_to_be16(0xffff);
+		*(__be64 *)&enc_a[2] = cpu_to_be64(adata_len);
+	}
+	for (i = 0; i < len; i++)
+		iowrite8(enc_a[i],
+			 aes_dev->base_reg +
+			 AES_A_DMA_INBUFFER_WRITE_FIFO_OFFSET);
+}
+
+static int ocs_aes_ccm_do_adata(struct ocs_aes_dev *aes_dev,
+				dma_addr_t adata_dma_list, u32 adata_size)
+{
+	int rc;
+
+	if (!adata_size) {
+		/* Since no aad the LAST_GCX bit can be set now */
+		aes_a_set_last_gcx_and_adata(aes_dev);
+		goto exit;
+	}
+
+	/* Adata case. */
+
+	/*
+	 * Form the encoding of the Associated data length and write it
+	 * to the AES/SM4 input buffer.
+	 */
+	ocs_aes_ccm_write_adata_len(aes_dev, adata_size);
+
+	/* Configure the AES/SM4 DMA to fetch the Associated Data */
+	dma_to_ocs_aes_ll(aes_dev, adata_dma_list);
+
+	/* Activate DMA to fetch Associated data. */
+	aes_a_dma_active_src_ll_en(aes_dev);
+
+	/* Set LAST_GCX and LAST_ADATA in AES ACTIVE register. */
+	aes_a_set_last_gcx_and_adata(aes_dev);
+
+	/* Wait for DMA transfer to complete. */
+	rc = ocs_aes_irq_enable_and_wait(aes_dev, AES_DMA_SRC_DONE_INT);
+	if (rc)
+		return rc;
+
+exit:
+	/* Wait until adata (if present) has been processed. */
+	aes_a_wait_last_gcx(aes_dev);
+	aes_a_dma_wait_input_buffer_occupancy(aes_dev);
+
+	return 0;
+}
+
+static int ocs_aes_ccm_encrypt_do_payload(struct ocs_aes_dev *aes_dev,
+					  dma_addr_t dst_dma_list,
+					  dma_addr_t src_dma_list,
+					  u32 src_size)
+{
+	if (src_size) {
+		/*
+		 * Configure and activate DMA for both input and output
+		 * data.
+		 */
+		dma_to_ocs_aes_ll(aes_dev, src_dma_list);
+		dma_from_ocs_aes_ll(aes_dev, dst_dma_list);
+		aes_a_dma_active_src_dst_ll_en(aes_dev);
+	} else {
+		/* Configure and activate DMA for output data only. */
+		dma_from_ocs_aes_ll(aes_dev, dst_dma_list);
+		aes_a_dma_active_dst_ll_en(aes_dev);
+	}
+
+	/*
+	 * Set the LAST GCX bit in AES_ACTIVE Register to instruct
+	 * AES/SM4 engine to pad the last block of data.
+	 */
+	aes_a_set_last_gcx(aes_dev);
+
+	/* We are done, wait for IRQ and return. */
+	return ocs_aes_irq_enable_and_wait(aes_dev, AES_COMPLETE_INT);
+}
+
+static int ocs_aes_ccm_decrypt_do_payload(struct ocs_aes_dev *aes_dev,
+					  dma_addr_t dst_dma_list,
+					  dma_addr_t src_dma_list,
+					  u32 src_size)
+{
+	if (!src_size) {
+		/* Let engine process 0-length input. */
+		aes_a_dma_set_xfer_size_zero(aes_dev);
+		aes_a_dma_active(aes_dev);
+		aes_a_set_last_gcx(aes_dev);
+
+		return 0;
+	}
+
+	/*
+	 * Configure and activate DMA for both input and output
+	 * data.
+	 */
+	dma_to_ocs_aes_ll(aes_dev, src_dma_list);
+	dma_from_ocs_aes_ll(aes_dev, dst_dma_list);
+	aes_a_dma_active_src_dst_ll_en(aes_dev);
+	/*
+	 * Set the LAST GCX bit in AES_ACTIVE Register; this allows the
+	 * AES/SM4 engine to differentiate between encrypted data and
+	 * encrypted MAC.
+	 */
+	aes_a_set_last_gcx(aes_dev);
+	 /*
+	  * Enable DMA DONE interrupt; once DMA transfer is over,
+	  * interrupt handler will process the MAC/tag.
+	  */
+	return ocs_aes_irq_enable_and_wait(aes_dev, AES_DMA_SRC_DONE_INT);
+}
+
+/*
+ * Compare Tag to Yr.
+ *
+ * Only used at the end of CCM decrypt. If tag == yr, message authentication
+ * has succeeded.
+ */
+static inline int ccm_compare_tag_to_yr(struct ocs_aes_dev *aes_dev,
+					u8 tag_size_bytes)
+{
+	u32 tag[AES_MAX_TAG_SIZE_U32];
+	u32 yr[AES_MAX_TAG_SIZE_U32];
+	u8 i;
+
+	/* Read Tag and Yr from AES registers. */
+	for (i = 0; i < AES_MAX_TAG_SIZE_U32; i++) {
+		tag[i] = ioread32(aes_dev->base_reg +
+				  AES_T_MAC_0_OFFSET + (i * sizeof(u32)));
+		yr[i] = ioread32(aes_dev->base_reg +
+				 AES_MULTIPURPOSE2_0_OFFSET +
+				 (i * sizeof(u32)));
+	}
+
+	return memcmp(tag, yr, tag_size_bytes) ? -EBADMSG : 0;
+}
+
+/**
+ * ocs_aes_ccm_op() - Perform CCM operation.
+ * @aes_dev:		The OCS AES device to use.
+ * @cipher:		The Cipher to use (AES or SM4).
+ * @instruction:	The instruction to perform (encrypt or decrypt).
+ * @dst_dma_list:	The OCS DMA list mapping output memory.
+ * @src_dma_list:	The OCS DMA list mapping input payload data.
+ * @src_size:		The amount of data mapped by @src_dma_list.
+ * @iv:			The input IV vector.
+ * @adata_dma_list:	The OCS DMA list mapping input A-data.
+ * @adata_size:		The amount of data mapped by @adata_dma_list.
+ * @in_tag:		Input tag.
+ * @tag_size:		The size (in bytes) of @in_tag.
+ *
+ * Note: for encrypt the tag is appended to the ciphertext (in the memory
+ *	 mapped by @dst_dma_list).
+ *
+ * Return: 0 on success, negative error code otherwise.
+ */
+int ocs_aes_ccm_op(struct ocs_aes_dev *aes_dev,
+		   enum ocs_cipher cipher,
+		   enum ocs_instruction instruction,
+		   dma_addr_t dst_dma_list,
+		   dma_addr_t src_dma_list,
+		   u32 src_size,
+		   u8 *iv,
+		   dma_addr_t adata_dma_list,
+		   u32 adata_size,
+		   u8 *in_tag,
+		   u32 tag_size)
+{
+	u32 *iv_32;
+	u8 lprime;
+	int rc;
+
+	rc = ocs_aes_validate_inputs(src_dma_list, src_size, iv,
+				     AES_BLOCK_SIZE, adata_dma_list, adata_size,
+				     in_tag, tag_size, cipher, OCS_MODE_CCM,
+				     instruction, dst_dma_list);
+	if (rc)
+		return rc;
+
+	ocs_aes_init(aes_dev, OCS_MODE_CCM, cipher, instruction);
+
+	/*
+	 * Note: rfc 3610 and NIST 800-38C require counter of zero to encrypt
+	 * auth tag so ensure this is the case
+	 */
+	lprime = iv[L_PRIME_IDX];
+	memset(&iv[COUNTER_START(lprime)], 0, COUNTER_LEN(lprime));
+
+	/*
+	 * Nonce is already converted to ctr0 before being passed into this
+	 * function as iv.
+	 */
+	iv_32 = (u32 *)iv;
+	iowrite32(__swab32(iv_32[0]),
+		  aes_dev->base_reg + AES_MULTIPURPOSE1_3_OFFSET);
+	iowrite32(__swab32(iv_32[1]),
+		  aes_dev->base_reg + AES_MULTIPURPOSE1_2_OFFSET);
+	iowrite32(__swab32(iv_32[2]),
+		  aes_dev->base_reg + AES_MULTIPURPOSE1_1_OFFSET);
+	iowrite32(__swab32(iv_32[3]),
+		  aes_dev->base_reg + AES_MULTIPURPOSE1_0_OFFSET);
+
+	/* Write MAC/tag length in register AES_TLEN */
+	iowrite32(tag_size, aes_dev->base_reg + AES_TLEN_OFFSET);
+	/*
+	 * Write the byte length of the last AES/SM4 block of Payload data
+	 * (without zero padding and without the length of the MAC) in register
+	 * AES_PLEN.
+	 */
+	ocs_aes_write_last_data_blk_len(aes_dev, src_size);
+
+	/* Set AES_ACTIVE.TRIGGER to start the operation. */
+	aes_a_op_trigger(aes_dev);
+
+	aes_a_dma_reset_and_activate_perf_cntr(aes_dev);
+
+	/* Form block B0 and write it to the AES/SM4 input buffer. */
+	rc = ocs_aes_ccm_write_b0(aes_dev, iv, adata_size, tag_size, src_size);
+	if (rc)
+		return rc;
+	/*
+	 * Ensure there has been at least CCM_DECRYPT_DELAY_LAST_GCX_CLK_COUNT
+	 * clock cycles since TRIGGER bit was set
+	 */
+	aes_a_dma_wait_and_deactivate_perf_cntr(aes_dev,
+						CCM_DECRYPT_DELAY_LAST_GCX_CLK_COUNT);
+
+	/* Process Adata. */
+	ocs_aes_ccm_do_adata(aes_dev, adata_dma_list, adata_size);
+
+	/* For Encrypt case we just process the payload and return. */
+	if (instruction == OCS_ENCRYPT) {
+		return ocs_aes_ccm_encrypt_do_payload(aes_dev, dst_dma_list,
+						      src_dma_list, src_size);
+	}
+	/* For Decypt we need to process the payload and then the tag. */
+	rc = ocs_aes_ccm_decrypt_do_payload(aes_dev, dst_dma_list,
+					    src_dma_list, src_size);
+	if (rc)
+		return rc;
+
+	/* Process MAC/tag directly: feed tag to engine and wait for IRQ. */
+	ocs_aes_ccm_write_encrypted_tag(aes_dev, in_tag, tag_size);
+	rc = ocs_aes_irq_enable_and_wait(aes_dev, AES_COMPLETE_INT);
+	if (rc)
+		return rc;
+
+	return ccm_compare_tag_to_yr(aes_dev, tag_size);
+}
+
+/**
+ * ocs_create_linked_list_from_sg() - Create OCS DMA linked list from SG list.
+ * @aes_dev:	  The OCS AES device the list will be created for.
+ * @sg:		  The SG list OCS DMA linked list will be created from. When
+ *		  passed to this function, @sg must have been already mapped
+ *		  with dma_map_sg().
+ * @sg_dma_count: The number of DMA-mapped entries in @sg. This must be the
+ *		  value returned by dma_map_sg() when @sg was mapped.
+ * @dll_desc:	  The OCS DMA dma_list to use to store information about the
+ *		  created linked list.
+ * @data_size:	  The size of the data (from the SG list) to be mapped into the
+ *		  OCS DMA linked list.
+ * @data_offset:  The offset (within the SG list) of the data to be mapped.
+ *
+ * Return:	0 on success, negative error code otherwise.
+ */
+int ocs_create_linked_list_from_sg(const struct ocs_aes_dev *aes_dev,
+				   struct scatterlist *sg,
+				   int sg_dma_count,
+				   struct ocs_dll_desc *dll_desc,
+				   size_t data_size, size_t data_offset)
+{
+	struct ocs_dma_linked_list *ll = NULL;
+	struct scatterlist *sg_tmp;
+	unsigned int tmp;
+	int dma_nents;
+	int i;
+
+	if (!dll_desc || !sg || !aes_dev)
+		return -EINVAL;
+
+	/* Default values for when no ddl_desc is created. */
+	dll_desc->vaddr = NULL;
+	dll_desc->dma_addr = DMA_MAPPING_ERROR;
+	dll_desc->size = 0;
+
+	if (data_size == 0)
+		return 0;
+
+	/* Loop over sg_list until we reach entry at specified offset. */
+	while (data_offset >= sg_dma_len(sg)) {
+		data_offset -= sg_dma_len(sg);
+		sg_dma_count--;
+		sg = sg_next(sg);
+		/* If we reach the end of the list, offset was invalid. */
+		if (!sg || sg_dma_count == 0)
+			return -EINVAL;
+	}
+
+	/* Compute number of DMA-mapped SG entries to add into OCS DMA list. */
+	dma_nents = 0;
+	tmp = 0;
+	sg_tmp = sg;
+	while (tmp < data_offset + data_size) {
+		/* If we reach the end of the list, data_size was invalid. */
+		if (!sg_tmp)
+			return -EINVAL;
+		tmp += sg_dma_len(sg_tmp);
+		dma_nents++;
+		sg_tmp = sg_next(sg_tmp);
+	}
+	if (dma_nents > sg_dma_count)
+		return -EINVAL;
+
+	/* Allocate the DMA list, one entry for each SG entry. */
+	dll_desc->size = sizeof(struct ocs_dma_linked_list) * dma_nents;
+	dll_desc->vaddr = dma_alloc_coherent(aes_dev->dev, dll_desc->size,
+					     &dll_desc->dma_addr, GFP_KERNEL);
+	if (!dll_desc->vaddr)
+		return -ENOMEM;
+
+	/* Populate DMA linked list entries. */
+	ll = dll_desc->vaddr;
+	for (i = 0; i < dma_nents; i++, sg = sg_next(sg)) {
+		ll[i].src_addr = sg_dma_address(sg) + data_offset;
+		ll[i].src_len = (sg_dma_len(sg) - data_offset) < data_size ?
+				(sg_dma_len(sg) - data_offset) : data_size;
+		data_offset = 0;
+		data_size -= ll[i].src_len;
+		/* Current element points to the DMA address of the next one. */
+		ll[i].next = dll_desc->dma_addr + (sizeof(*ll) * (i + 1));
+		ll[i].ll_flags = 0;
+	}
+	/* Terminate last element. */
+	ll[i - 1].next = 0;
+	ll[i - 1].ll_flags = OCS_LL_DMA_FLAG_TERMINATE;
+
+	return 0;
+}
diff --git a/drivers/crypto/keembay/ocs-aes.h b/drivers/crypto/keembay/ocs-aes.h
new file mode 100644
index 000000000000..c035fc48b7ed
--- /dev/null
+++ b/drivers/crypto/keembay/ocs-aes.h
@@ -0,0 +1,129 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Intel Keem Bay OCS AES Crypto Driver.
+ *
+ * Copyright (C) 2018-2020 Intel Corporation
+ */
+
+#ifndef _CRYPTO_OCS_AES_H
+#define _CRYPTO_OCS_AES_H
+
+#include <linux/dma-mapping.h>
+
+enum ocs_cipher {
+	OCS_AES = 0,
+	OCS_SM4 = 1,
+};
+
+enum ocs_mode {
+	OCS_MODE_ECB = 0,
+	OCS_MODE_CBC = 1,
+	OCS_MODE_CTR = 2,
+	OCS_MODE_CCM = 6,
+	OCS_MODE_GCM = 7,
+	OCS_MODE_CTS = 9,
+};
+
+enum ocs_instruction {
+	OCS_ENCRYPT = 0,
+	OCS_DECRYPT = 1,
+	OCS_EXPAND  = 2,
+	OCS_BYPASS  = 3,
+};
+
+/**
+ * struct ocs_aes_dev - AES device context.
+ * @list:			List head for insertion into device list hold
+ *				by driver.
+ * @dev:			OCS AES device.
+ * @irq:			IRQ number.
+ * @base_reg:			IO base address of OCS AES.
+ * @irq_copy_completion:	Completion to indicate IRQ has been triggered.
+ * @dma_err_mask:		Error reported by OCS DMA interrupts.
+ * @engine:			Crypto engine for the device.
+ */
+struct ocs_aes_dev {
+	struct list_head list;
+	struct device *dev;
+	int irq;
+	void __iomem *base_reg;
+	struct completion irq_completion;
+	u32 dma_err_mask;
+	struct crypto_engine *engine;
+};
+
+/**
+ * struct ocs_dll_desc - Descriptor of an OCS DMA Linked List.
+ * @vaddr:	Virtual address of the linked list head.
+ * @dma_addr:	DMA address of the linked list head.
+ * @size:	Size (in bytes) of the linked list.
+ */
+struct ocs_dll_desc {
+	void		*vaddr;
+	dma_addr_t	dma_addr;
+	size_t		size;
+};
+
+int ocs_aes_set_key(struct ocs_aes_dev *aes_dev, const u32 key_size,
+		    const u8 *key, const enum ocs_cipher cipher);
+
+int ocs_aes_op(struct ocs_aes_dev *aes_dev,
+	       enum ocs_mode mode,
+	       enum ocs_cipher cipher,
+	       enum ocs_instruction instruction,
+	       dma_addr_t dst_dma_list,
+	       dma_addr_t src_dma_list,
+	       u32 src_size,
+	       u8 *iv,
+	       u32 iv_size);
+
+/**
+ * ocs_aes_bypass_op() - Use OCS DMA to copy data.
+ * @aes_dev:            The OCS AES device to use.
+ * @dst_dma_list:	The OCS DMA list mapping the memory where input data
+ *			will be copied to.
+ * @src_dma_list:	The OCS DMA list mapping input data.
+ * @src_size:		The amount of data to copy.
+ */
+static inline int ocs_aes_bypass_op(struct ocs_aes_dev *aes_dev,
+				    dma_addr_t dst_dma_list,
+				    dma_addr_t src_dma_list, u32 src_size)
+{
+	return ocs_aes_op(aes_dev, OCS_MODE_ECB, OCS_AES, OCS_BYPASS,
+			  dst_dma_list, src_dma_list, src_size, NULL, 0);
+}
+
+int ocs_aes_gcm_op(struct ocs_aes_dev *aes_dev,
+		   enum ocs_cipher cipher,
+		   enum ocs_instruction instruction,
+		   dma_addr_t dst_dma_list,
+		   dma_addr_t src_dma_list,
+		   u32 src_size,
+		   const u8 *iv,
+		   dma_addr_t aad_dma_list,
+		   u32 aad_size,
+		   u8 *out_tag,
+		   u32 tag_size);
+
+int ocs_aes_ccm_op(struct ocs_aes_dev *aes_dev,
+		   enum ocs_cipher cipher,
+		   enum ocs_instruction instruction,
+		   dma_addr_t dst_dma_list,
+		   dma_addr_t src_dma_list,
+		   u32 src_size,
+		   u8 *iv,
+		   dma_addr_t adata_dma_list,
+		   u32 adata_size,
+		   u8 *in_tag,
+		   u32 tag_size);
+
+int ocs_create_linked_list_from_sg(const struct ocs_aes_dev *aes_dev,
+				   struct scatterlist *sg,
+				   int sg_dma_count,
+				   struct ocs_dll_desc *dll_desc,
+				   size_t data_size,
+				   size_t data_offset);
+
+irqreturn_t ocs_aes_irq_handler(int irq, void *dev_id);
+
+#endif
diff --git a/drivers/crypto/marvell/cesa/hash.c b/drivers/crypto/marvell/cesa/hash.c
index add7ea011c98..8cf9fd518d86 100644
--- a/drivers/crypto/marvell/cesa/hash.c
+++ b/drivers/crypto/marvell/cesa/hash.c
@@ -11,7 +11,8 @@
 
 #include <crypto/hmac.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <linux/device.h>
 #include <linux/dma-mapping.h>
 
diff --git a/drivers/crypto/marvell/octeontx/otx_cptpf_main.c b/drivers/crypto/marvell/octeontx/otx_cptpf_main.c
index 34bb3063eb70..14a42559f81d 100644
--- a/drivers/crypto/marvell/octeontx/otx_cptpf_main.c
+++ b/drivers/crypto/marvell/octeontx/otx_cptpf_main.c
@@ -212,15 +212,9 @@ static int otx_cpt_probe(struct pci_dev *pdev,
 		goto err_disable_device;
 	}
 
-	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(48));
+	err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
 	if (err) {
-		dev_err(dev, "Unable to get usable DMA configuration\n");
-		goto err_release_regions;
-	}
-
-	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(48));
-	if (err) {
-		dev_err(dev, "Unable to get 48-bit DMA for consistent allocations\n");
+		dev_err(dev, "Unable to get usable 48-bit DMA configuration\n");
 		goto err_release_regions;
 	}
 
diff --git a/drivers/crypto/marvell/octeontx/otx_cptvf_algs.c b/drivers/crypto/marvell/octeontx/otx_cptvf_algs.c
index 90bb31329d4b..ccbef01888d4 100644
--- a/drivers/crypto/marvell/octeontx/otx_cptvf_algs.c
+++ b/drivers/crypto/marvell/octeontx/otx_cptvf_algs.c
@@ -13,7 +13,8 @@
 #include <crypto/cryptd.h>
 #include <crypto/des.h>
 #include <crypto/internal/aead.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/xts.h>
 #include <crypto/scatterwalk.h>
 #include <linux/rtnetlink.h>
diff --git a/drivers/crypto/marvell/octeontx/otx_cptvf_main.c b/drivers/crypto/marvell/octeontx/otx_cptvf_main.c
index 228fe8e47e0e..c076d0b3ad5f 100644
--- a/drivers/crypto/marvell/octeontx/otx_cptvf_main.c
+++ b/drivers/crypto/marvell/octeontx/otx_cptvf_main.c
@@ -804,15 +804,9 @@ static int otx_cptvf_probe(struct pci_dev *pdev,
 		dev_err(dev, "PCI request regions failed 0x%x\n", err);
 		goto disable_device;
 	}
-	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(48));
+	err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
 	if (err) {
-		dev_err(dev, "Unable to get usable DMA configuration\n");
-		goto release_regions;
-	}
-
-	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(48));
-	if (err) {
-		dev_err(dev, "Unable to get 48-bit DMA for consistent allocations\n");
+		dev_err(dev, "Unable to get usable 48-bit DMA configuration\n");
 		goto release_regions;
 	}
 
diff --git a/drivers/crypto/mediatek/mtk-sha.c b/drivers/crypto/mediatek/mtk-sha.c
index 3d5d7d68b03b..f55aacdafbef 100644
--- a/drivers/crypto/mediatek/mtk-sha.c
+++ b/drivers/crypto/mediatek/mtk-sha.c
@@ -10,7 +10,8 @@
  */
 
 #include <crypto/hmac.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include "mtk-platform.h"
 
 #define SHA_ALIGN_MSK		(sizeof(u32) - 1)
diff --git a/drivers/crypto/mxs-dcp.c b/drivers/crypto/mxs-dcp.c
index 909a7eb748e3..d6a7784d2988 100644
--- a/drivers/crypto/mxs-dcp.c
+++ b/drivers/crypto/mxs-dcp.c
@@ -17,7 +17,8 @@
 #include <linux/clk.h>
 
 #include <crypto/aes.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/scatterwalk.h>
diff --git a/drivers/crypto/n2_core.c b/drivers/crypto/n2_core.c
index 3642bf83d809..3b0bf6fea491 100644
--- a/drivers/crypto/n2_core.c
+++ b/drivers/crypto/n2_core.c
@@ -15,7 +15,8 @@
 #include <linux/interrupt.h>
 #include <linux/crypto.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/aes.h>
 #include <crypto/internal/des.h>
 #include <linux/mutex.h>
diff --git a/drivers/crypto/nx/nx-sha256.c b/drivers/crypto/nx/nx-sha256.c
index 02fb53453195..90d9a37a57f6 100644
--- a/drivers/crypto/nx/nx-sha256.c
+++ b/drivers/crypto/nx/nx-sha256.c
@@ -8,7 +8,7 @@
  */
 
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha2.h>
 #include <linux/module.h>
 #include <asm/vio.h>
 #include <asm/byteorder.h>
diff --git a/drivers/crypto/nx/nx-sha512.c b/drivers/crypto/nx/nx-sha512.c
index 4c7a3e3eeebf..eb8627a0f317 100644
--- a/drivers/crypto/nx/nx-sha512.c
+++ b/drivers/crypto/nx/nx-sha512.c
@@ -8,7 +8,7 @@
  */
 
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha2.h>
 #include <linux/module.h>
 #include <asm/vio.h>
 
diff --git a/drivers/crypto/nx/nx.c b/drivers/crypto/nx/nx.c
index 40882d6d52c1..0d2dc5be7f19 100644
--- a/drivers/crypto/nx/nx.c
+++ b/drivers/crypto/nx/nx.c
@@ -10,7 +10,7 @@
 #include <crypto/internal/aead.h>
 #include <crypto/internal/hash.h>
 #include <crypto/aes.h>
-#include <crypto/sha.h>
+#include <crypto/sha2.h>
 #include <crypto/algapi.h>
 #include <crypto/scatterwalk.h>
 #include <linux/module.h>
diff --git a/drivers/crypto/omap-aes.c b/drivers/crypto/omap-aes.c
index 4fd14d90cc40..a45bdcf3026d 100644
--- a/drivers/crypto/omap-aes.c
+++ b/drivers/crypto/omap-aes.c
@@ -105,6 +105,7 @@ static int omap_aes_hw_init(struct omap_aes_dev *dd)
 
 	err = pm_runtime_get_sync(dd->dev);
 	if (err < 0) {
+		pm_runtime_put_noidle(dd->dev);
 		dev_err(dd->dev, "failed to get sync: %d\n", err);
 		return err;
 	}
@@ -1137,7 +1138,7 @@ static int omap_aes_probe(struct platform_device *pdev)
 	if (err < 0) {
 		dev_err(dev, "%s: failed to get_sync(%d)\n",
 			__func__, err);
-		goto err_res;
+		goto err_pm_disable;
 	}
 
 	omap_aes_dma_stop(dd);
@@ -1246,6 +1247,7 @@ err_engine:
 	omap_aes_dma_cleanup(dd);
 err_irq:
 	tasklet_kill(&dd->done_task);
+err_pm_disable:
 	pm_runtime_disable(dev);
 err_res:
 	dd = NULL;
diff --git a/drivers/crypto/omap-sham.c b/drivers/crypto/omap-sham.c
index a3b38d2c92e7..ae0d320d3c60 100644
--- a/drivers/crypto/omap-sham.c
+++ b/drivers/crypto/omap-sham.c
@@ -35,7 +35,8 @@
 #include <linux/crypto.h>
 #include <crypto/scatterwalk.h>
 #include <crypto/algapi.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/hash.h>
 #include <crypto/hmac.h>
 #include <crypto/internal/hash.h>
diff --git a/drivers/crypto/padlock-sha.c b/drivers/crypto/padlock-sha.c
index a697a4a3f2d0..6865c7f1fc1a 100644
--- a/drivers/crypto/padlock-sha.c
+++ b/drivers/crypto/padlock-sha.c
@@ -9,7 +9,8 @@
 
 #include <crypto/internal/hash.h>
 #include <crypto/padlock.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <linux/err.h>
 #include <linux/module.h>
 #include <linux/init.h>
diff --git a/drivers/crypto/picoxcell_crypto.c b/drivers/crypto/picoxcell_crypto.c
index fb34bf92861d..84f9c16d984c 100644
--- a/drivers/crypto/picoxcell_crypto.c
+++ b/drivers/crypto/picoxcell_crypto.c
@@ -8,7 +8,8 @@
 #include <crypto/authenc.h>
 #include <crypto/internal/des.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/internal/skcipher.h>
 #include <linux/clk.h>
 #include <linux/crypto.h>
diff --git a/drivers/crypto/qat/Kconfig b/drivers/crypto/qat/Kconfig
index 2006322345de..beb379b23dc3 100644
--- a/drivers/crypto/qat/Kconfig
+++ b/drivers/crypto/qat/Kconfig
@@ -46,6 +46,17 @@ config CRYPTO_DEV_QAT_C62X
 	  To compile this as a module, choose M here: the module
 	  will be called qat_c62x.
 
+config CRYPTO_DEV_QAT_4XXX
+	tristate "Support for Intel(R) QAT_4XXX"
+	depends on X86 && PCI
+	select CRYPTO_DEV_QAT
+	help
+	  Support for Intel(R) QuickAssist Technology QAT_4xxx
+	  for accelerating crypto and compression workloads.
+
+	  To compile this as a module, choose M here: the module
+	  will be called qat_4xxx.
+
 config CRYPTO_DEV_QAT_DH895xCCVF
 	tristate "Support for Intel(R) DH895xCC Virtual Function"
 	depends on X86 && PCI
diff --git a/drivers/crypto/qat/Makefile b/drivers/crypto/qat/Makefile
index 7dd15e751d02..258c8a626ce0 100644
--- a/drivers/crypto/qat/Makefile
+++ b/drivers/crypto/qat/Makefile
@@ -3,6 +3,7 @@ obj-$(CONFIG_CRYPTO_DEV_QAT) += qat_common/
 obj-$(CONFIG_CRYPTO_DEV_QAT_DH895xCC) += qat_dh895xcc/
 obj-$(CONFIG_CRYPTO_DEV_QAT_C3XXX) += qat_c3xxx/
 obj-$(CONFIG_CRYPTO_DEV_QAT_C62X) += qat_c62x/
+obj-$(CONFIG_CRYPTO_DEV_QAT_4XXX) += qat_4xxx/
 obj-$(CONFIG_CRYPTO_DEV_QAT_DH895xCCVF) += qat_dh895xccvf/
 obj-$(CONFIG_CRYPTO_DEV_QAT_C3XXXVF) += qat_c3xxxvf/
 obj-$(CONFIG_CRYPTO_DEV_QAT_C62XVF) += qat_c62xvf/
diff --git a/drivers/crypto/qat/qat_4xxx/Makefile b/drivers/crypto/qat/qat_4xxx/Makefile
new file mode 100644
index 000000000000..ff9c8b5897ea
--- /dev/null
+++ b/drivers/crypto/qat/qat_4xxx/Makefile
@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only)
+ccflags-y := -I $(srctree)/$(src)/../qat_common
+obj-$(CONFIG_CRYPTO_DEV_QAT_4XXX) += qat_4xxx.o
+qat_4xxx-objs := adf_drv.o adf_4xxx_hw_data.o
diff --git a/drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.c b/drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.c
new file mode 100644
index 000000000000..344bfae45bff
--- /dev/null
+++ b/drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.c
@@ -0,0 +1,242 @@
+// SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only)
+/* Copyright(c) 2020 Intel Corporation */
+#include <adf_accel_devices.h>
+#include <adf_common_drv.h>
+#include <adf_pf2vf_msg.h>
+#include <adf_gen4_hw_data.h>
+#include "adf_4xxx_hw_data.h"
+#include "icp_qat_hw.h"
+
+struct adf_fw_config {
+	u32 ae_mask;
+	char *obj_name;
+};
+
+static struct adf_fw_config adf_4xxx_fw_config[] = {
+	{0xF0, ADF_4XXX_SYM_OBJ},
+	{0xF, ADF_4XXX_ASYM_OBJ},
+	{0x100, ADF_4XXX_ADMIN_OBJ},
+};
+
+/* Worker thread to service arbiter mappings */
+static u32 thrd_to_arb_map[] = {
+	0x5555555, 0x5555555, 0x5555555, 0x5555555,
+	0xAAAAAAA, 0xAAAAAAA, 0xAAAAAAA, 0xAAAAAAA,
+	0x0
+};
+
+static struct adf_hw_device_class adf_4xxx_class = {
+	.name = ADF_4XXX_DEVICE_NAME,
+	.type = DEV_4XXX,
+	.instances = 0,
+};
+
+static u32 get_accel_mask(struct adf_hw_device_data *self)
+{
+	return ADF_4XXX_ACCELERATORS_MASK;
+}
+
+static u32 get_ae_mask(struct adf_hw_device_data *self)
+{
+	u32 me_disable = self->fuses;
+
+	return ~me_disable & ADF_4XXX_ACCELENGINES_MASK;
+}
+
+static u32 get_num_accels(struct adf_hw_device_data *self)
+{
+	return ADF_4XXX_MAX_ACCELERATORS;
+}
+
+static u32 get_num_aes(struct adf_hw_device_data *self)
+{
+	if (!self || !self->ae_mask)
+		return 0;
+
+	return hweight32(self->ae_mask);
+}
+
+static u32 get_misc_bar_id(struct adf_hw_device_data *self)
+{
+	return ADF_4XXX_PMISC_BAR;
+}
+
+static u32 get_etr_bar_id(struct adf_hw_device_data *self)
+{
+	return ADF_4XXX_ETR_BAR;
+}
+
+static u32 get_sram_bar_id(struct adf_hw_device_data *self)
+{
+	return ADF_4XXX_SRAM_BAR;
+}
+
+/*
+ * The vector routing table is used to select the MSI-X entry to use for each
+ * interrupt source.
+ * The first ADF_4XXX_ETR_MAX_BANKS entries correspond to ring interrupts.
+ * The final entry corresponds to VF2PF or error interrupts.
+ * This vector table could be used to configure one MSI-X entry to be shared
+ * between multiple interrupt sources.
+ *
+ * The default routing is set to have a one to one correspondence between the
+ * interrupt source and the MSI-X entry used.
+ */
+static void set_msix_default_rttable(struct adf_accel_dev *accel_dev)
+{
+	void __iomem *csr;
+	int i;
+
+	csr = (&GET_BARS(accel_dev)[ADF_4XXX_PMISC_BAR])->virt_addr;
+	for (i = 0; i <= ADF_4XXX_ETR_MAX_BANKS; i++)
+		ADF_CSR_WR(csr, ADF_4XXX_MSIX_RTTABLE_OFFSET(i), i);
+}
+
+static u32 get_accel_cap(struct adf_accel_dev *accel_dev)
+{
+	struct pci_dev *pdev = accel_dev->accel_pci_dev.pci_dev;
+	u32 fusectl1;
+	u32 capabilities = ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC |
+			   ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC |
+			   ICP_ACCEL_CAPABILITIES_AUTHENTICATION |
+			   ICP_ACCEL_CAPABILITIES_AES_V2;
+
+	/* Read accelerator capabilities mask */
+	pci_read_config_dword(pdev, ADF_4XXX_FUSECTL1_OFFSET, &fusectl1);
+
+	if (fusectl1 & ICP_ACCEL_4XXX_MASK_CIPHER_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC;
+	if (fusectl1 & ICP_ACCEL_4XXX_MASK_AUTH_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_AUTHENTICATION;
+	if (fusectl1 & ICP_ACCEL_4XXX_MASK_PKE_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC;
+
+	return capabilities;
+}
+
+static enum dev_sku_info get_sku(struct adf_hw_device_data *self)
+{
+	return DEV_SKU_1;
+}
+
+static void adf_get_arbiter_mapping(struct adf_accel_dev *accel_dev,
+				    u32 const **arb_map_config)
+{
+	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
+	unsigned long ae_mask = hw_device->ae_mask;
+	int i;
+
+	for_each_clear_bit(i, &ae_mask, ADF_4XXX_MAX_ACCELENGINES)
+		thrd_to_arb_map[i] = 0;
+
+	*arb_map_config = thrd_to_arb_map;
+}
+
+static void get_arb_info(struct arb_info *arb_info)
+{
+	arb_info->arb_cfg = ADF_4XXX_ARB_CONFIG;
+	arb_info->arb_offset = ADF_4XXX_ARB_OFFSET;
+	arb_info->wt2sam_offset = ADF_4XXX_ARB_WRK_2_SER_MAP_OFFSET;
+}
+
+static void get_admin_info(struct admin_info *admin_csrs_info)
+{
+	admin_csrs_info->mailbox_offset = ADF_4XXX_MAILBOX_BASE_OFFSET;
+	admin_csrs_info->admin_msg_ur = ADF_4XXX_ADMINMSGUR_OFFSET;
+	admin_csrs_info->admin_msg_lr = ADF_4XXX_ADMINMSGLR_OFFSET;
+}
+
+static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
+{
+	struct adf_bar *misc_bar = &GET_BARS(accel_dev)[ADF_4XXX_PMISC_BAR];
+	void __iomem *csr = misc_bar->virt_addr;
+
+	/* Enable all in errsou3 except VFLR notification on host */
+	ADF_CSR_WR(csr, ADF_4XXX_ERRMSK3, ADF_4XXX_VFLNOTIFY);
+}
+
+static void adf_enable_ints(struct adf_accel_dev *accel_dev)
+{
+	void __iomem *addr;
+
+	addr = (&GET_BARS(accel_dev)[ADF_4XXX_PMISC_BAR])->virt_addr;
+
+	/* Enable bundle interrupts */
+	ADF_CSR_WR(addr, ADF_4XXX_SMIAPF_RP_X0_MASK_OFFSET, 0);
+	ADF_CSR_WR(addr, ADF_4XXX_SMIAPF_RP_X1_MASK_OFFSET, 0);
+
+	/* Enable misc interrupts */
+	ADF_CSR_WR(addr, ADF_4XXX_SMIAPF_MASK_OFFSET, 0);
+}
+
+static int adf_pf_enable_vf2pf_comms(struct adf_accel_dev *accel_dev)
+{
+	return 0;
+}
+
+static u32 uof_get_num_objs(void)
+{
+	return ARRAY_SIZE(adf_4xxx_fw_config);
+}
+
+static char *uof_get_name(u32 obj_num)
+{
+	return adf_4xxx_fw_config[obj_num].obj_name;
+}
+
+static u32 uof_get_ae_mask(u32 obj_num)
+{
+	return adf_4xxx_fw_config[obj_num].ae_mask;
+}
+
+void adf_init_hw_data_4xxx(struct adf_hw_device_data *hw_data)
+{
+	hw_data->dev_class = &adf_4xxx_class;
+	hw_data->instance_id = adf_4xxx_class.instances++;
+	hw_data->num_banks = ADF_4XXX_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_4XXX_NUM_RINGS_PER_BANK;
+	hw_data->num_accel = ADF_4XXX_MAX_ACCELERATORS;
+	hw_data->num_engines = ADF_4XXX_MAX_ACCELENGINES;
+	hw_data->num_logical_accel = 1;
+	hw_data->tx_rx_gap = ADF_4XXX_RX_RINGS_OFFSET;
+	hw_data->tx_rings_mask = ADF_4XXX_TX_RINGS_MASK;
+	hw_data->alloc_irq = adf_isr_resource_alloc;
+	hw_data->free_irq = adf_isr_resource_free;
+	hw_data->enable_error_correction = adf_enable_error_correction;
+	hw_data->get_accel_mask = get_accel_mask;
+	hw_data->get_ae_mask = get_ae_mask;
+	hw_data->get_num_accels = get_num_accels;
+	hw_data->get_num_aes = get_num_aes;
+	hw_data->get_sram_bar_id = get_sram_bar_id;
+	hw_data->get_etr_bar_id = get_etr_bar_id;
+	hw_data->get_misc_bar_id = get_misc_bar_id;
+	hw_data->get_arb_info = get_arb_info;
+	hw_data->get_admin_info = get_admin_info;
+	hw_data->get_accel_cap = get_accel_cap;
+	hw_data->get_sku = get_sku;
+	hw_data->fw_name = ADF_4XXX_FW;
+	hw_data->fw_mmp_name = ADF_4XXX_MMP;
+	hw_data->init_admin_comms = adf_init_admin_comms;
+	hw_data->exit_admin_comms = adf_exit_admin_comms;
+	hw_data->disable_iov = adf_disable_sriov;
+	hw_data->send_admin_init = adf_send_admin_init;
+	hw_data->init_arb = adf_init_arb;
+	hw_data->exit_arb = adf_exit_arb;
+	hw_data->get_arb_mapping = adf_get_arbiter_mapping;
+	hw_data->enable_ints = adf_enable_ints;
+	hw_data->enable_vf2pf_comms = adf_pf_enable_vf2pf_comms;
+	hw_data->reset_device = adf_reset_flr;
+	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
+	hw_data->admin_ae_mask = ADF_4XXX_ADMIN_AE_MASK;
+	hw_data->uof_get_num_objs = uof_get_num_objs;
+	hw_data->uof_get_name = uof_get_name;
+	hw_data->uof_get_ae_mask = uof_get_ae_mask;
+	hw_data->set_msix_rttable = set_msix_default_rttable;
+
+	adf_gen4_init_hw_csr_ops(&hw_data->csr_ops);
+}
+
+void adf_clean_hw_data_4xxx(struct adf_hw_device_data *hw_data)
+{
+	hw_data->dev_class->instances--;
+}
diff --git a/drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.h b/drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.h
new file mode 100644
index 000000000000..4fe2a776293c
--- /dev/null
+++ b/drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.h
@@ -0,0 +1,86 @@
+/* SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only) */
+/* Copyright(c) 2014 - 2020 Intel Corporation */
+#ifndef ADF_4XXX_HW_DATA_H_
+#define ADF_4XXX_HW_DATA_H_
+
+#include <adf_accel_devices.h>
+
+/* PCIe configuration space */
+#define ADF_4XXX_SRAM_BAR		0
+#define ADF_4XXX_PMISC_BAR		1
+#define ADF_4XXX_ETR_BAR		2
+#define ADF_4XXX_RX_RINGS_OFFSET	1
+#define ADF_4XXX_TX_RINGS_MASK		0x1
+#define ADF_4XXX_MAX_ACCELERATORS	1
+#define ADF_4XXX_MAX_ACCELENGINES	9
+#define ADF_4XXX_BAR_MASK		(BIT(0) | BIT(2) | BIT(4))
+
+/* Physical function fuses */
+#define ADF_4XXX_FUSECTL0_OFFSET	(0x2C8)
+#define ADF_4XXX_FUSECTL1_OFFSET	(0x2CC)
+#define ADF_4XXX_FUSECTL2_OFFSET	(0x2D0)
+#define ADF_4XXX_FUSECTL3_OFFSET	(0x2D4)
+#define ADF_4XXX_FUSECTL4_OFFSET	(0x2D8)
+#define ADF_4XXX_FUSECTL5_OFFSET	(0x2DC)
+
+#define ADF_4XXX_ACCELERATORS_MASK	(0x1)
+#define ADF_4XXX_ACCELENGINES_MASK	(0x1FF)
+#define ADF_4XXX_ADMIN_AE_MASK		(0x100)
+
+#define ADF_4XXX_ETR_MAX_BANKS		64
+
+/* MSIX interrupt */
+#define ADF_4XXX_SMIAPF_RP_X0_MASK_OFFSET	(0x41A040)
+#define ADF_4XXX_SMIAPF_RP_X1_MASK_OFFSET	(0x41A044)
+#define ADF_4XXX_SMIAPF_MASK_OFFSET		(0x41A084)
+#define ADF_4XXX_MSIX_RTTABLE_OFFSET(i)		(0x409000 + ((i) * 0x04))
+
+/* Bank and ring configuration */
+#define ADF_4XXX_NUM_RINGS_PER_BANK	2
+
+/* Error source registers */
+#define ADF_4XXX_ERRSOU0	(0x41A200)
+#define ADF_4XXX_ERRSOU1	(0x41A204)
+#define ADF_4XXX_ERRSOU2	(0x41A208)
+#define ADF_4XXX_ERRSOU3	(0x41A20C)
+
+/* Error source mask registers */
+#define ADF_4XXX_ERRMSK0	(0x41A210)
+#define ADF_4XXX_ERRMSK1	(0x41A214)
+#define ADF_4XXX_ERRMSK2	(0x41A218)
+#define ADF_4XXX_ERRMSK3	(0x41A21C)
+
+#define ADF_4XXX_VFLNOTIFY	BIT(7)
+
+/* Arbiter configuration */
+#define ADF_4XXX_ARB_CONFIG			(BIT(31) | BIT(6) | BIT(0))
+#define ADF_4XXX_ARB_OFFSET			(0x0)
+#define ADF_4XXX_ARB_WRK_2_SER_MAP_OFFSET	(0x400)
+
+/* Admin Interface Reg Offset */
+#define ADF_4XXX_ADMINMSGUR_OFFSET	(0x500574)
+#define ADF_4XXX_ADMINMSGLR_OFFSET	(0x500578)
+#define ADF_4XXX_MAILBOX_BASE_OFFSET	(0x600970)
+
+/* Firmware Binaries */
+#define ADF_4XXX_FW		"qat_4xxx.bin"
+#define ADF_4XXX_MMP		"qat_4xxx_mmp.bin"
+#define ADF_4XXX_SYM_OBJ	"qat_4xxx_sym.bin"
+#define ADF_4XXX_ASYM_OBJ	"qat_4xxx_asym.bin"
+#define ADF_4XXX_ADMIN_OBJ	"qat_4xxx_admin.bin"
+
+/* qat_4xxx fuse bits are different from old GENs, redefine them */
+enum icp_qat_4xxx_slice_mask {
+	ICP_ACCEL_4XXX_MASK_CIPHER_SLICE = BIT(0),
+	ICP_ACCEL_4XXX_MASK_AUTH_SLICE = BIT(1),
+	ICP_ACCEL_4XXX_MASK_PKE_SLICE = BIT(2),
+	ICP_ACCEL_4XXX_MASK_COMPRESS_SLICE = BIT(3),
+	ICP_ACCEL_4XXX_MASK_UCS_SLICE = BIT(4),
+	ICP_ACCEL_4XXX_MASK_EIA3_SLICE = BIT(5),
+	ICP_ACCEL_4XXX_MASK_SMX_SLICE = BIT(6),
+};
+
+void adf_init_hw_data_4xxx(struct adf_hw_device_data *hw_data);
+void adf_clean_hw_data_4xxx(struct adf_hw_device_data *hw_data);
+
+#endif
diff --git a/drivers/crypto/qat/qat_4xxx/adf_drv.c b/drivers/crypto/qat/qat_4xxx/adf_drv.c
new file mode 100644
index 000000000000..a8805c815d16
--- /dev/null
+++ b/drivers/crypto/qat/qat_4xxx/adf_drv.c
@@ -0,0 +1,323 @@
+// SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only)
+/* Copyright(c) 2020 Intel Corporation */
+#include <linux/device.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+
+#include <adf_accel_devices.h>
+#include <adf_cfg.h>
+#include <adf_common_drv.h>
+
+#include "adf_4xxx_hw_data.h"
+#include "qat_crypto.h"
+#include "adf_transport_access_macros.h"
+
+static const struct pci_device_id adf_pci_tbl[] = {
+	{ PCI_VDEVICE(INTEL, ADF_4XXX_PCI_DEVICE_ID), },
+	{ }
+};
+MODULE_DEVICE_TABLE(pci, adf_pci_tbl);
+
+static void adf_cleanup_accel(struct adf_accel_dev *accel_dev)
+{
+	if (accel_dev->hw_device) {
+		adf_clean_hw_data_4xxx(accel_dev->hw_device);
+		accel_dev->hw_device = NULL;
+	}
+	adf_cfg_dev_remove(accel_dev);
+	debugfs_remove(accel_dev->debugfs_dir);
+	adf_devmgr_rm_dev(accel_dev, NULL);
+}
+
+static int adf_crypto_dev_config(struct adf_accel_dev *accel_dev)
+{
+	char key[ADF_CFG_MAX_KEY_LEN_IN_BYTES];
+	int banks = GET_MAX_BANKS(accel_dev);
+	int cpus = num_online_cpus();
+	unsigned long bank, val;
+	int instances;
+	int ret;
+	int i;
+
+	if (adf_hw_dev_has_crypto(accel_dev))
+		instances = min(cpus, banks / 2);
+	else
+		instances = 0;
+
+	ret = adf_cfg_section_add(accel_dev, ADF_KERNEL_SEC);
+	if (ret)
+		goto err;
+
+	ret = adf_cfg_section_add(accel_dev, "Accelerator0");
+	if (ret)
+		goto err;
+
+	for (i = 0; i < instances; i++) {
+		val = i;
+		bank = i * 2;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_BANK_NUM, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &bank, ADF_DEC);
+		if (ret)
+			goto err;
+
+		bank += 1;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_BANK_NUM, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &bank, ADF_DEC);
+		if (ret)
+			goto err;
+
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_ETRMGR_CORE_AFFINITY,
+			 i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_SIZE, i);
+		val = 128;
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		val = 512;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_SIZE, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		val = 0;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_TX, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		val = 0;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_TX, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		val = 1;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_RX, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		val = 1;
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_RX, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		val = ADF_COALESCING_DEF_TIME;
+		snprintf(key, sizeof(key), ADF_ETRMGR_COALESCE_TIMER_FORMAT, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, "Accelerator0",
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+	}
+
+	val = i;
+	ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC, ADF_NUM_CY,
+					  &val, ADF_DEC);
+	if (ret)
+		goto err;
+
+	set_bit(ADF_STATUS_CONFIGURED, &accel_dev->status);
+	return 0;
+err:
+	dev_err(&GET_DEV(accel_dev), "Failed to start QAT accel dev\n");
+	return ret;
+}
+
+static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	struct adf_accel_dev *accel_dev;
+	struct adf_accel_pci *accel_pci_dev;
+	struct adf_hw_device_data *hw_data;
+	char name[ADF_DEVICE_NAME_LENGTH];
+	unsigned int i, bar_nr;
+	unsigned long bar_mask;
+	struct adf_bar *bar;
+	int ret;
+
+	if (num_possible_nodes() > 1 && dev_to_node(&pdev->dev) < 0) {
+		/*
+		 * If the accelerator is connected to a node with no memory
+		 * there is no point in using the accelerator since the remote
+		 * memory transaction will be very slow.
+		 */
+		dev_err(&pdev->dev, "Invalid NUMA configuration.\n");
+		return -EINVAL;
+	}
+
+	accel_dev = devm_kzalloc(&pdev->dev, sizeof(*accel_dev), GFP_KERNEL);
+	if (!accel_dev)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&accel_dev->crypto_list);
+	accel_pci_dev = &accel_dev->accel_pci_dev;
+	accel_pci_dev->pci_dev = pdev;
+
+	/*
+	 * Add accel device to accel table
+	 * This should be called before adf_cleanup_accel is called
+	 */
+	if (adf_devmgr_add_dev(accel_dev, NULL)) {
+		dev_err(&pdev->dev, "Failed to add new accelerator device.\n");
+		return -EFAULT;
+	}
+
+	accel_dev->owner = THIS_MODULE;
+	/* Allocate and initialise device hardware meta-data structure */
+	hw_data = devm_kzalloc(&pdev->dev, sizeof(*hw_data), GFP_KERNEL);
+	if (!hw_data) {
+		ret = -ENOMEM;
+		goto out_err;
+	}
+
+	accel_dev->hw_device = hw_data;
+	adf_init_hw_data_4xxx(accel_dev->hw_device);
+
+	pci_read_config_byte(pdev, PCI_REVISION_ID, &accel_pci_dev->revid);
+	pci_read_config_dword(pdev, ADF_4XXX_FUSECTL4_OFFSET, &hw_data->fuses);
+
+	/* Get Accelerators and Accelerators Engines masks */
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
+	accel_pci_dev->sku = hw_data->get_sku(hw_data);
+	/* If the device has no acceleration engines then ignore it */
+	if (!hw_data->accel_mask || !hw_data->ae_mask ||
+	    (~hw_data->ae_mask & 0x01)) {
+		dev_err(&pdev->dev, "No acceleration units found.\n");
+		ret = -EFAULT;
+		goto out_err;
+	}
+
+	/* Create dev top level debugfs entry */
+	snprintf(name, sizeof(name), "%s%s_%s", ADF_DEVICE_NAME_PREFIX,
+		 hw_data->dev_class->name, pci_name(pdev));
+
+	accel_dev->debugfs_dir = debugfs_create_dir(name, NULL);
+
+	/* Create device configuration table */
+	ret = adf_cfg_dev_add(accel_dev);
+	if (ret)
+		goto out_err;
+
+	/* Enable PCI device */
+	ret = pcim_enable_device(pdev);
+	if (ret) {
+		dev_err(&pdev->dev, "Can't enable PCI device.\n");
+		goto out_err;
+	}
+
+	/* Set DMA identifier */
+	if (pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) {
+		if ((pci_set_dma_mask(pdev, DMA_BIT_MASK(32)))) {
+			dev_err(&pdev->dev, "No usable DMA configuration.\n");
+			ret = -EFAULT;
+			goto out_err;
+		} else {
+			pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
+		}
+	} else {
+		pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+	}
+
+	/* Get accelerator capabilities mask */
+	hw_data->accel_capabilities_mask = hw_data->get_accel_cap(accel_dev);
+
+	/* Find and map all the device's BARS */
+	bar_mask = pci_select_bars(pdev, IORESOURCE_MEM) & ADF_4XXX_BAR_MASK;
+
+	ret = pcim_iomap_regions_request_all(pdev, bar_mask, pci_name(pdev));
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to map pci regions.\n");
+		goto out_err;
+	}
+
+	i = 0;
+	for_each_set_bit(bar_nr, &bar_mask, PCI_STD_NUM_BARS) {
+		bar = &accel_pci_dev->pci_bars[i++];
+		bar->virt_addr = pcim_iomap_table(pdev)[bar_nr];
+	}
+
+	pci_set_master(pdev);
+
+	if (adf_enable_aer(accel_dev)) {
+		dev_err(&pdev->dev, "Failed to enable aer.\n");
+		ret = -EFAULT;
+		goto out_err;
+	}
+
+	if (pci_save_state(pdev)) {
+		dev_err(&pdev->dev, "Failed to save pci state.\n");
+		ret = -ENOMEM;
+		goto out_err_disable_aer;
+	}
+
+	ret = adf_crypto_dev_config(accel_dev);
+	if (ret)
+		goto out_err_disable_aer;
+
+	ret = adf_dev_init(accel_dev);
+	if (ret)
+		goto out_err_dev_shutdown;
+
+	ret = adf_dev_start(accel_dev);
+	if (ret)
+		goto out_err_dev_stop;
+
+	return ret;
+
+out_err_dev_stop:
+	adf_dev_stop(accel_dev);
+out_err_dev_shutdown:
+	adf_dev_shutdown(accel_dev);
+out_err_disable_aer:
+	adf_disable_aer(accel_dev);
+out_err:
+	adf_cleanup_accel(accel_dev);
+	return ret;
+}
+
+static void adf_remove(struct pci_dev *pdev)
+{
+	struct adf_accel_dev *accel_dev = adf_devmgr_pci_to_accel_dev(pdev);
+
+	if (!accel_dev) {
+		pr_err("QAT: Driver removal failed\n");
+		return;
+	}
+	adf_dev_stop(accel_dev);
+	adf_dev_shutdown(accel_dev);
+	adf_disable_aer(accel_dev);
+	adf_cleanup_accel(accel_dev);
+}
+
+static struct pci_driver adf_driver = {
+	.id_table = adf_pci_tbl,
+	.name = ADF_4XXX_DEVICE_NAME,
+	.probe = adf_probe,
+	.remove = adf_remove,
+	.sriov_configure = adf_sriov_configure,
+};
+
+module_pci_driver(adf_driver);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Intel");
+MODULE_FIRMWARE(ADF_4XXX_FW);
+MODULE_FIRMWARE(ADF_4XXX_MMP);
+MODULE_DESCRIPTION("Intel(R) QuickAssist Technology");
+MODULE_VERSION(ADF_DRV_VERSION);
+MODULE_SOFTDEP("pre: crypto-intel_qat");
diff --git a/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.c b/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.c
index aee494d3da52..eb45f1b1ae3e 100644
--- a/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.c
+++ b/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.c
@@ -3,7 +3,9 @@
 #include <adf_accel_devices.h>
 #include <adf_common_drv.h>
 #include <adf_pf2vf_msg.h>
+#include <adf_gen2_hw_data.h>
 #include "adf_c3xxx_hw_data.h"
+#include "icp_qat_hw.h"
 
 /* Worker thread to service arbiter mappings based on dev SKUs */
 static const u32 thrd_to_arb_map_6_me_sku[] = {
@@ -17,15 +19,33 @@ static struct adf_hw_device_class c3xxx_class = {
 	.instances = 0
 };
 
-static u32 get_accel_mask(u32 fuse)
+static u32 get_accel_mask(struct adf_hw_device_data *self)
 {
-	return (~fuse) >> ADF_C3XXX_ACCELERATORS_REG_OFFSET &
-		ADF_C3XXX_ACCELERATORS_MASK;
+	u32 straps = self->straps;
+	u32 fuses = self->fuses;
+	u32 accel;
+
+	accel = ~(fuses | straps) >> ADF_C3XXX_ACCELERATORS_REG_OFFSET;
+	accel &= ADF_C3XXX_ACCELERATORS_MASK;
+
+	return accel;
 }
 
-static u32 get_ae_mask(u32 fuse)
+static u32 get_ae_mask(struct adf_hw_device_data *self)
 {
-	return (~fuse) & ADF_C3XXX_ACCELENGINES_MASK;
+	u32 straps = self->straps;
+	u32 fuses = self->fuses;
+	unsigned long disabled;
+	u32 ae_disable;
+	int accel;
+
+	/* If an accel is disabled, then disable the corresponding two AEs */
+	disabled = ~get_accel_mask(self) & ADF_C3XXX_ACCELERATORS_MASK;
+	ae_disable = BIT(1) | BIT(0);
+	for_each_set_bit(accel, &disabled, ADF_C3XXX_MAX_ACCELERATORS)
+		straps |= ae_disable << (accel << 1);
+
+	return ~(fuses | straps) & ADF_C3XXX_ACCELENGINES_MASK;
 }
 
 static u32 get_num_accels(struct adf_hw_device_data *self)
@@ -109,11 +129,13 @@ static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
 {
 	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
 	struct adf_bar *misc_bar = &GET_BARS(accel_dev)[ADF_C3XXX_PMISC_BAR];
+	unsigned long accel_mask = hw_device->accel_mask;
+	unsigned long ae_mask = hw_device->ae_mask;
 	void __iomem *csr = misc_bar->virt_addr;
 	unsigned int val, i;
 
 	/* Enable Accel Engine error detection & correction */
-	for (i = 0; i < hw_device->get_num_aes(hw_device); i++) {
+	for_each_set_bit(i, &ae_mask, GET_MAX_ACCELENGINES(accel_dev)) {
 		val = ADF_CSR_RD(csr, ADF_C3XXX_AE_CTX_ENABLES(i));
 		val |= ADF_C3XXX_ENABLE_AE_ECC_ERR;
 		ADF_CSR_WR(csr, ADF_C3XXX_AE_CTX_ENABLES(i), val);
@@ -123,7 +145,7 @@ static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
 	}
 
 	/* Enable shared memory error detection & correction */
-	for (i = 0; i < hw_device->get_num_accels(hw_device); i++) {
+	for_each_set_bit(i, &accel_mask, ADF_C3XXX_MAX_ACCELERATORS) {
 		val = ADF_CSR_RD(csr, ADF_C3XXX_UERRSSMSH(i));
 		val |= ADF_C3XXX_ERRSSMSH_EN;
 		ADF_CSR_WR(csr, ADF_C3XXX_UERRSSMSH(i), val);
@@ -151,11 +173,19 @@ static int adf_pf_enable_vf2pf_comms(struct adf_accel_dev *accel_dev)
 	return 0;
 }
 
+static void configure_iov_threads(struct adf_accel_dev *accel_dev, bool enable)
+{
+	adf_gen2_cfg_iov_thds(accel_dev, enable,
+			      ADF_C3XXX_AE2FUNC_MAP_GRP_A_NUM_REGS,
+			      ADF_C3XXX_AE2FUNC_MAP_GRP_B_NUM_REGS);
+}
+
 void adf_init_hw_data_c3xxx(struct adf_hw_device_data *hw_data)
 {
 	hw_data->dev_class = &c3xxx_class;
 	hw_data->instance_id = c3xxx_class.instances++;
 	hw_data->num_banks = ADF_C3XXX_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_ETR_MAX_RINGS_PER_BANK;
 	hw_data->num_accel = ADF_C3XXX_MAX_ACCELERATORS;
 	hw_data->num_logical_accel = 1;
 	hw_data->num_engines = ADF_C3XXX_MAX_ACCELENGINES;
@@ -166,6 +196,7 @@ void adf_init_hw_data_c3xxx(struct adf_hw_device_data *hw_data)
 	hw_data->enable_error_correction = adf_enable_error_correction;
 	hw_data->get_accel_mask = get_accel_mask;
 	hw_data->get_ae_mask = get_ae_mask;
+	hw_data->get_accel_cap = adf_gen2_get_accel_cap;
 	hw_data->get_num_accels = get_num_accels;
 	hw_data->get_num_aes = get_num_aes;
 	hw_data->get_sram_bar_id = get_sram_bar_id;
@@ -173,11 +204,14 @@ void adf_init_hw_data_c3xxx(struct adf_hw_device_data *hw_data)
 	hw_data->get_misc_bar_id = get_misc_bar_id;
 	hw_data->get_pf2vf_offset = get_pf2vf_offset;
 	hw_data->get_vintmsk_offset = get_vintmsk_offset;
+	hw_data->get_admin_info = adf_gen2_get_admin_info;
+	hw_data->get_arb_info = adf_gen2_get_arb_info;
 	hw_data->get_sku = get_sku;
 	hw_data->fw_name = ADF_C3XXX_FW;
 	hw_data->fw_mmp_name = ADF_C3XXX_MMP;
 	hw_data->init_admin_comms = adf_init_admin_comms;
 	hw_data->exit_admin_comms = adf_exit_admin_comms;
+	hw_data->configure_iov_threads = configure_iov_threads;
 	hw_data->disable_iov = adf_disable_sriov;
 	hw_data->send_admin_init = adf_send_admin_init;
 	hw_data->init_arb = adf_init_arb;
@@ -187,6 +221,7 @@ void adf_init_hw_data_c3xxx(struct adf_hw_device_data *hw_data)
 	hw_data->enable_vf2pf_comms = adf_pf_enable_vf2pf_comms;
 	hw_data->reset_device = adf_reset_flr;
 	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
+	adf_gen2_init_hw_csr_ops(&hw_data->csr_ops);
 }
 
 void adf_clean_hw_data_c3xxx(struct adf_hw_device_data *hw_data)
diff --git a/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.h b/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.h
index 8b5dd2c94ebf..fece8e38025a 100644
--- a/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.h
+++ b/drivers/crypto/qat/qat_c3xxx/adf_c3xxx_hw_data.h
@@ -18,6 +18,7 @@
 #define ADF_C3XXX_SMIAPF1_MASK_OFFSET (0x3A000 + 0x30)
 #define ADF_C3XXX_SMIA0_MASK 0xFFFF
 #define ADF_C3XXX_SMIA1_MASK 0x1
+#define ADF_C3XXX_SOFTSTRAP_CSR_OFFSET 0x2EC
 /* Error detection and correction */
 #define ADF_C3XXX_AE_CTX_ENABLES(i) (i * 0x1000 + 0x20818)
 #define ADF_C3XXX_AE_MISC_CONTROL(i) (i * 0x1000 + 0x20960)
@@ -30,6 +31,10 @@
 #define ADF_C3XXX_PF2VF_OFFSET(i)	(0x3A000 + 0x280 + ((i) * 0x04))
 #define ADF_C3XXX_VINTMSK_OFFSET(i)	(0x3A000 + 0x200 + ((i) * 0x04))
 
+/* AE to function mapping */
+#define ADF_C3XXX_AE2FUNC_MAP_GRP_A_NUM_REGS 48
+#define ADF_C3XXX_AE2FUNC_MAP_GRP_B_NUM_REGS 6
+
 /* Firmware Binary */
 #define ADF_C3XXX_FW "qat_c3xxx.bin"
 #define ADF_C3XXX_MMP "qat_c3xxx_mmp.bin"
diff --git a/drivers/crypto/qat/qat_c3xxx/adf_drv.c b/drivers/crypto/qat/qat_c3xxx/adf_drv.c
index ed0e8e33fe4b..7fb3343ae8b0 100644
--- a/drivers/crypto/qat/qat_c3xxx/adf_drv.c
+++ b/drivers/crypto/qat/qat_c3xxx/adf_drv.c
@@ -126,10 +126,12 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	pci_read_config_byte(pdev, PCI_REVISION_ID, &accel_pci_dev->revid);
 	pci_read_config_dword(pdev, ADF_DEVICE_FUSECTL_OFFSET,
 			      &hw_data->fuses);
+	pci_read_config_dword(pdev, ADF_C3XXX_SOFTSTRAP_CSR_OFFSET,
+			      &hw_data->straps);
 
 	/* Get Accelerators and Accelerators Engines masks */
-	hw_data->accel_mask = hw_data->get_accel_mask(hw_data->fuses);
-	hw_data->ae_mask = hw_data->get_ae_mask(hw_data->fuses);
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
 	accel_pci_dev->sku = hw_data->get_sku(hw_data);
 	/* If the device has no acceleration engines then ignore it. */
 	if (!hw_data->accel_mask || !hw_data->ae_mask ||
@@ -175,9 +177,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto out_err_disable;
 	}
 
-	/* Read accelerator capabilities mask */
-	pci_read_config_dword(pdev, ADF_DEVICE_LEGFUSE_OFFSET,
-			      &hw_data->accel_capabilities_mask);
+	/* Get accelerator capabilities mask */
+	hw_data->accel_capabilities_mask = hw_data->get_accel_cap(accel_dev);
 
 	/* Find and map all the device's BARS */
 	i = 0;
diff --git a/drivers/crypto/qat/qat_c3xxxvf/adf_c3xxxvf_hw_data.c b/drivers/crypto/qat/qat_c3xxxvf/adf_c3xxxvf_hw_data.c
index d2fedbd7113c..15f6b9bdfb22 100644
--- a/drivers/crypto/qat/qat_c3xxxvf/adf_c3xxxvf_hw_data.c
+++ b/drivers/crypto/qat/qat_c3xxxvf/adf_c3xxxvf_hw_data.c
@@ -3,6 +3,7 @@
 #include <adf_accel_devices.h>
 #include <adf_pf2vf_msg.h>
 #include <adf_common_drv.h>
+#include <adf_gen2_hw_data.h>
 #include "adf_c3xxxvf_hw_data.h"
 
 static struct adf_hw_device_class c3xxxiov_class = {
@@ -11,12 +12,12 @@ static struct adf_hw_device_class c3xxxiov_class = {
 	.instances = 0
 };
 
-static u32 get_accel_mask(u32 fuse)
+static u32 get_accel_mask(struct adf_hw_device_data *self)
 {
 	return ADF_C3XXXIOV_ACCELERATORS_MASK;
 }
 
-static u32 get_ae_mask(u32 fuse)
+static u32 get_ae_mask(struct adf_hw_device_data *self)
 {
 	return ADF_C3XXXIOV_ACCELENGINES_MASK;
 }
@@ -69,6 +70,7 @@ void adf_init_hw_data_c3xxxiov(struct adf_hw_device_data *hw_data)
 {
 	hw_data->dev_class = &c3xxxiov_class;
 	hw_data->num_banks = ADF_C3XXXIOV_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_ETR_MAX_RINGS_PER_BANK;
 	hw_data->num_accel = ADF_C3XXXIOV_MAX_ACCELERATORS;
 	hw_data->num_logical_accel = 1;
 	hw_data->num_engines = ADF_C3XXXIOV_MAX_ACCELENGINES;
@@ -97,6 +99,7 @@ void adf_init_hw_data_c3xxxiov(struct adf_hw_device_data *hw_data)
 	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
 	hw_data->dev_class->instances++;
 	adf_devmgr_update_class_index(hw_data);
+	adf_gen2_init_hw_csr_ops(&hw_data->csr_ops);
 }
 
 void adf_clean_hw_data_c3xxxiov(struct adf_hw_device_data *hw_data)
diff --git a/drivers/crypto/qat/qat_c3xxxvf/adf_drv.c b/drivers/crypto/qat/qat_c3xxxvf/adf_drv.c
index 456979b136a2..1d1532e8fb6d 100644
--- a/drivers/crypto/qat/qat_c3xxxvf/adf_drv.c
+++ b/drivers/crypto/qat/qat_c3xxxvf/adf_drv.c
@@ -119,8 +119,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	adf_init_hw_data_c3xxxiov(accel_dev->hw_device);
 
 	/* Get Accelerators and Accelerators Engines masks */
-	hw_data->accel_mask = hw_data->get_accel_mask(hw_data->fuses);
-	hw_data->ae_mask = hw_data->get_ae_mask(hw_data->fuses);
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
 	accel_pci_dev->sku = hw_data->get_sku(hw_data);
 
 	/* Create dev top level debugfs entry */
diff --git a/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.c b/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.c
index 844ad5ed33fc..babdffbcb846 100644
--- a/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.c
+++ b/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.c
@@ -3,7 +3,9 @@
 #include <adf_accel_devices.h>
 #include <adf_common_drv.h>
 #include <adf_pf2vf_msg.h>
+#include <adf_gen2_hw_data.h>
 #include "adf_c62x_hw_data.h"
+#include "icp_qat_hw.h"
 
 /* Worker thread to service arbiter mappings based on dev SKUs */
 static const u32 thrd_to_arb_map_8_me_sku[] = {
@@ -22,15 +24,33 @@ static struct adf_hw_device_class c62x_class = {
 	.instances = 0
 };
 
-static u32 get_accel_mask(u32 fuse)
+static u32 get_accel_mask(struct adf_hw_device_data *self)
 {
-	return (~fuse) >> ADF_C62X_ACCELERATORS_REG_OFFSET &
-			  ADF_C62X_ACCELERATORS_MASK;
+	u32 straps = self->straps;
+	u32 fuses = self->fuses;
+	u32 accel;
+
+	accel = ~(fuses | straps) >> ADF_C62X_ACCELERATORS_REG_OFFSET;
+	accel &= ADF_C62X_ACCELERATORS_MASK;
+
+	return accel;
 }
 
-static u32 get_ae_mask(u32 fuse)
+static u32 get_ae_mask(struct adf_hw_device_data *self)
 {
-	return (~fuse) & ADF_C62X_ACCELENGINES_MASK;
+	u32 straps = self->straps;
+	u32 fuses = self->fuses;
+	unsigned long disabled;
+	u32 ae_disable;
+	int accel;
+
+	/* If an accel is disabled, then disable the corresponding two AEs */
+	disabled = ~get_accel_mask(self) & ADF_C62X_ACCELERATORS_MASK;
+	ae_disable = BIT(1) | BIT(0);
+	for_each_set_bit(accel, &disabled, ADF_C62X_MAX_ACCELERATORS)
+		straps |= ae_disable << (accel << 1);
+
+	return ~(fuses | straps) & ADF_C62X_ACCELENGINES_MASK;
 }
 
 static u32 get_num_accels(struct adf_hw_device_data *self)
@@ -119,11 +139,13 @@ static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
 {
 	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
 	struct adf_bar *misc_bar = &GET_BARS(accel_dev)[ADF_C62X_PMISC_BAR];
+	unsigned long accel_mask = hw_device->accel_mask;
+	unsigned long ae_mask = hw_device->ae_mask;
 	void __iomem *csr = misc_bar->virt_addr;
 	unsigned int val, i;
 
 	/* Enable Accel Engine error detection & correction */
-	for (i = 0; i < hw_device->get_num_aes(hw_device); i++) {
+	for_each_set_bit(i, &ae_mask, GET_MAX_ACCELENGINES(accel_dev)) {
 		val = ADF_CSR_RD(csr, ADF_C62X_AE_CTX_ENABLES(i));
 		val |= ADF_C62X_ENABLE_AE_ECC_ERR;
 		ADF_CSR_WR(csr, ADF_C62X_AE_CTX_ENABLES(i), val);
@@ -133,7 +155,7 @@ static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
 	}
 
 	/* Enable shared memory error detection & correction */
-	for (i = 0; i < hw_device->get_num_accels(hw_device); i++) {
+	for_each_set_bit(i, &accel_mask, ADF_C62X_MAX_ACCELERATORS) {
 		val = ADF_CSR_RD(csr, ADF_C62X_UERRSSMSH(i));
 		val |= ADF_C62X_ERRSSMSH_EN;
 		ADF_CSR_WR(csr, ADF_C62X_UERRSSMSH(i), val);
@@ -161,11 +183,19 @@ static int adf_pf_enable_vf2pf_comms(struct adf_accel_dev *accel_dev)
 	return 0;
 }
 
+static void configure_iov_threads(struct adf_accel_dev *accel_dev, bool enable)
+{
+	adf_gen2_cfg_iov_thds(accel_dev, enable,
+			      ADF_C62X_AE2FUNC_MAP_GRP_A_NUM_REGS,
+			      ADF_C62X_AE2FUNC_MAP_GRP_B_NUM_REGS);
+}
+
 void adf_init_hw_data_c62x(struct adf_hw_device_data *hw_data)
 {
 	hw_data->dev_class = &c62x_class;
 	hw_data->instance_id = c62x_class.instances++;
 	hw_data->num_banks = ADF_C62X_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_ETR_MAX_RINGS_PER_BANK;
 	hw_data->num_accel = ADF_C62X_MAX_ACCELERATORS;
 	hw_data->num_logical_accel = 1;
 	hw_data->num_engines = ADF_C62X_MAX_ACCELENGINES;
@@ -176,6 +206,7 @@ void adf_init_hw_data_c62x(struct adf_hw_device_data *hw_data)
 	hw_data->enable_error_correction = adf_enable_error_correction;
 	hw_data->get_accel_mask = get_accel_mask;
 	hw_data->get_ae_mask = get_ae_mask;
+	hw_data->get_accel_cap = adf_gen2_get_accel_cap;
 	hw_data->get_num_accels = get_num_accels;
 	hw_data->get_num_aes = get_num_aes;
 	hw_data->get_sram_bar_id = get_sram_bar_id;
@@ -183,11 +214,14 @@ void adf_init_hw_data_c62x(struct adf_hw_device_data *hw_data)
 	hw_data->get_misc_bar_id = get_misc_bar_id;
 	hw_data->get_pf2vf_offset = get_pf2vf_offset;
 	hw_data->get_vintmsk_offset = get_vintmsk_offset;
+	hw_data->get_admin_info = adf_gen2_get_admin_info;
+	hw_data->get_arb_info = adf_gen2_get_arb_info;
 	hw_data->get_sku = get_sku;
 	hw_data->fw_name = ADF_C62X_FW;
 	hw_data->fw_mmp_name = ADF_C62X_MMP;
 	hw_data->init_admin_comms = adf_init_admin_comms;
 	hw_data->exit_admin_comms = adf_exit_admin_comms;
+	hw_data->configure_iov_threads = configure_iov_threads;
 	hw_data->disable_iov = adf_disable_sriov;
 	hw_data->send_admin_init = adf_send_admin_init;
 	hw_data->init_arb = adf_init_arb;
@@ -197,6 +231,7 @@ void adf_init_hw_data_c62x(struct adf_hw_device_data *hw_data)
 	hw_data->enable_vf2pf_comms = adf_pf_enable_vf2pf_comms;
 	hw_data->reset_device = adf_reset_flr;
 	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
+	adf_gen2_init_hw_csr_ops(&hw_data->csr_ops);
 }
 
 void adf_clean_hw_data_c62x(struct adf_hw_device_data *hw_data)
diff --git a/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.h b/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.h
index 88504d2bf30d..53d3cb577f5b 100644
--- a/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.h
+++ b/drivers/crypto/qat/qat_c62x/adf_c62x_hw_data.h
@@ -19,6 +19,7 @@
 #define ADF_C62X_SMIAPF1_MASK_OFFSET (0x3A000 + 0x30)
 #define ADF_C62X_SMIA0_MASK 0xFFFF
 #define ADF_C62X_SMIA1_MASK 0x1
+#define ADF_C62X_SOFTSTRAP_CSR_OFFSET 0x2EC
 /* Error detection and correction */
 #define ADF_C62X_AE_CTX_ENABLES(i) (i * 0x1000 + 0x20818)
 #define ADF_C62X_AE_MISC_CONTROL(i) (i * 0x1000 + 0x20960)
@@ -31,6 +32,10 @@
 #define ADF_C62X_PF2VF_OFFSET(i)	(0x3A000 + 0x280 + ((i) * 0x04))
 #define ADF_C62X_VINTMSK_OFFSET(i)	(0x3A000 + 0x200 + ((i) * 0x04))
 
+/* AE to function mapping */
+#define ADF_C62X_AE2FUNC_MAP_GRP_A_NUM_REGS 80
+#define ADF_C62X_AE2FUNC_MAP_GRP_B_NUM_REGS 10
+
 /* Firmware Binary */
 #define ADF_C62X_FW "qat_c62x.bin"
 #define ADF_C62X_MMP "qat_c62x_mmp.bin"
diff --git a/drivers/crypto/qat/qat_c62x/adf_drv.c b/drivers/crypto/qat/qat_c62x/adf_drv.c
index d8e7c9c25590..1f5de442e1e6 100644
--- a/drivers/crypto/qat/qat_c62x/adf_drv.c
+++ b/drivers/crypto/qat/qat_c62x/adf_drv.c
@@ -126,10 +126,12 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	pci_read_config_byte(pdev, PCI_REVISION_ID, &accel_pci_dev->revid);
 	pci_read_config_dword(pdev, ADF_DEVICE_FUSECTL_OFFSET,
 			      &hw_data->fuses);
+	pci_read_config_dword(pdev, ADF_C62X_SOFTSTRAP_CSR_OFFSET,
+			      &hw_data->straps);
 
 	/* Get Accelerators and Accelerators Engines masks */
-	hw_data->accel_mask = hw_data->get_accel_mask(hw_data->fuses);
-	hw_data->ae_mask = hw_data->get_ae_mask(hw_data->fuses);
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
 	accel_pci_dev->sku = hw_data->get_sku(hw_data);
 	/* If the device has no acceleration engines then ignore it. */
 	if (!hw_data->accel_mask || !hw_data->ae_mask ||
@@ -175,9 +177,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto out_err_disable;
 	}
 
-	/* Read accelerator capabilities mask */
-	pci_read_config_dword(pdev, ADF_DEVICE_LEGFUSE_OFFSET,
-			      &hw_data->accel_capabilities_mask);
+	/* Get accelerator capabilities mask */
+	hw_data->accel_capabilities_mask = hw_data->get_accel_cap(accel_dev);
 
 	/* Find and map all the device's BARS */
 	i = (hw_data->fuses & ADF_DEVICE_FUSECTL_MASK) ? 1 : 0;
diff --git a/drivers/crypto/qat/qat_c62xvf/adf_c62xvf_hw_data.c b/drivers/crypto/qat/qat_c62xvf/adf_c62xvf_hw_data.c
index 29fd3f1091ab..d231583428c9 100644
--- a/drivers/crypto/qat/qat_c62xvf/adf_c62xvf_hw_data.c
+++ b/drivers/crypto/qat/qat_c62xvf/adf_c62xvf_hw_data.c
@@ -3,6 +3,7 @@
 #include <adf_accel_devices.h>
 #include <adf_pf2vf_msg.h>
 #include <adf_common_drv.h>
+#include <adf_gen2_hw_data.h>
 #include "adf_c62xvf_hw_data.h"
 
 static struct adf_hw_device_class c62xiov_class = {
@@ -11,12 +12,12 @@ static struct adf_hw_device_class c62xiov_class = {
 	.instances = 0
 };
 
-static u32 get_accel_mask(u32 fuse)
+static u32 get_accel_mask(struct adf_hw_device_data *self)
 {
 	return ADF_C62XIOV_ACCELERATORS_MASK;
 }
 
-static u32 get_ae_mask(u32 fuse)
+static u32 get_ae_mask(struct adf_hw_device_data *self)
 {
 	return ADF_C62XIOV_ACCELENGINES_MASK;
 }
@@ -69,6 +70,7 @@ void adf_init_hw_data_c62xiov(struct adf_hw_device_data *hw_data)
 {
 	hw_data->dev_class = &c62xiov_class;
 	hw_data->num_banks = ADF_C62XIOV_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_ETR_MAX_RINGS_PER_BANK;
 	hw_data->num_accel = ADF_C62XIOV_MAX_ACCELERATORS;
 	hw_data->num_logical_accel = 1;
 	hw_data->num_engines = ADF_C62XIOV_MAX_ACCELENGINES;
@@ -97,6 +99,7 @@ void adf_init_hw_data_c62xiov(struct adf_hw_device_data *hw_data)
 	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
 	hw_data->dev_class->instances++;
 	adf_devmgr_update_class_index(hw_data);
+	adf_gen2_init_hw_csr_ops(&hw_data->csr_ops);
 }
 
 void adf_clean_hw_data_c62xiov(struct adf_hw_device_data *hw_data)
diff --git a/drivers/crypto/qat/qat_c62xvf/adf_drv.c b/drivers/crypto/qat/qat_c62xvf/adf_drv.c
index b9810f79eb84..04742a6d91ca 100644
--- a/drivers/crypto/qat/qat_c62xvf/adf_drv.c
+++ b/drivers/crypto/qat/qat_c62xvf/adf_drv.c
@@ -119,8 +119,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	adf_init_hw_data_c62xiov(accel_dev->hw_device);
 
 	/* Get Accelerators and Accelerators Engines masks */
-	hw_data->accel_mask = hw_data->get_accel_mask(hw_data->fuses);
-	hw_data->ae_mask = hw_data->get_ae_mask(hw_data->fuses);
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
 	accel_pci_dev->sku = hw_data->get_sku(hw_data);
 
 	/* Create dev top level debugfs entry */
diff --git a/drivers/crypto/qat/qat_common/Makefile b/drivers/crypto/qat/qat_common/Makefile
index 47a8e3d8b81a..9c57abdf56b7 100644
--- a/drivers/crypto/qat/qat_common/Makefile
+++ b/drivers/crypto/qat/qat_common/Makefile
@@ -10,6 +10,8 @@ intel_qat-objs := adf_cfg.o \
 	adf_transport.o \
 	adf_admin.o \
 	adf_hw_arbiter.o \
+	adf_gen2_hw_data.o \
+	adf_gen4_hw_data.o \
 	qat_crypto.o \
 	qat_algs.o \
 	qat_asym_algs.o \
diff --git a/drivers/crypto/qat/qat_common/adf_accel_devices.h b/drivers/crypto/qat/qat_common/adf_accel_devices.h
index 06952ece53d9..c46a5805b294 100644
--- a/drivers/crypto/qat/qat_common/adf_accel_devices.h
+++ b/drivers/crypto/qat/qat_common/adf_accel_devices.h
@@ -15,6 +15,9 @@
 #define ADF_C62XVF_DEVICE_NAME "c6xxvf"
 #define ADF_C3XXX_DEVICE_NAME "c3xxx"
 #define ADF_C3XXXVF_DEVICE_NAME "c3xxxvf"
+#define ADF_4XXX_DEVICE_NAME "4xxx"
+#define ADF_4XXX_PCI_DEVICE_ID 0x4940
+#define ADF_4XXXIOV_PCI_DEVICE_ID 0x4941
 #define ADF_ERRSOU3 (0x3A000 + 0x0C)
 #define ADF_ERRSOU5 (0x3A000 + 0xD8)
 #define ADF_DEVICE_FUSECTL_OFFSET 0x40
@@ -97,6 +100,46 @@ struct adf_hw_device_class {
 	u32 instances;
 } __packed;
 
+struct arb_info {
+	u32 arb_cfg;
+	u32 arb_offset;
+	u32 wt2sam_offset;
+};
+
+struct admin_info {
+	u32 admin_msg_ur;
+	u32 admin_msg_lr;
+	u32 mailbox_offset;
+};
+
+struct adf_hw_csr_ops {
+	u64 (*build_csr_ring_base_addr)(dma_addr_t addr, u32 size);
+	u32 (*read_csr_ring_head)(void __iomem *csr_base_addr, u32 bank,
+				  u32 ring);
+	void (*write_csr_ring_head)(void __iomem *csr_base_addr, u32 bank,
+				    u32 ring, u32 value);
+	u32 (*read_csr_ring_tail)(void __iomem *csr_base_addr, u32 bank,
+				  u32 ring);
+	void (*write_csr_ring_tail)(void __iomem *csr_base_addr, u32 bank,
+				    u32 ring, u32 value);
+	u32 (*read_csr_e_stat)(void __iomem *csr_base_addr, u32 bank);
+	void (*write_csr_ring_config)(void __iomem *csr_base_addr, u32 bank,
+				      u32 ring, u32 value);
+	void (*write_csr_ring_base)(void __iomem *csr_base_addr, u32 bank,
+				    u32 ring, dma_addr_t addr);
+	void (*write_csr_int_flag)(void __iomem *csr_base_addr, u32 bank,
+				   u32 value);
+	void (*write_csr_int_srcsel)(void __iomem *csr_base_addr, u32 bank);
+	void (*write_csr_int_col_en)(void __iomem *csr_base_addr, u32 bank,
+				     u32 value);
+	void (*write_csr_int_col_ctl)(void __iomem *csr_base_addr, u32 bank,
+				      u32 value);
+	void (*write_csr_int_flag_and_col)(void __iomem *csr_base_addr,
+					   u32 bank, u32 value);
+	void (*write_csr_ring_srv_arb_en)(void __iomem *csr_base_addr, u32 bank,
+					  u32 value);
+};
+
 struct adf_cfg_device_data;
 struct adf_accel_dev;
 struct adf_etr_data;
@@ -104,8 +147,9 @@ struct adf_etr_ring_data;
 
 struct adf_hw_device_data {
 	struct adf_hw_device_class *dev_class;
-	u32 (*get_accel_mask)(u32 fuse);
-	u32 (*get_ae_mask)(u32 fuse);
+	u32 (*get_accel_mask)(struct adf_hw_device_data *self);
+	u32 (*get_ae_mask)(struct adf_hw_device_data *self);
+	u32 (*get_accel_cap)(struct adf_accel_dev *accel_dev);
 	u32 (*get_sram_bar_id)(struct adf_hw_device_data *self);
 	u32 (*get_misc_bar_id)(struct adf_hw_device_data *self);
 	u32 (*get_etr_bar_id)(struct adf_hw_device_data *self);
@@ -113,6 +157,8 @@ struct adf_hw_device_data {
 	u32 (*get_num_accels)(struct adf_hw_device_data *self);
 	u32 (*get_pf2vf_offset)(u32 i);
 	u32 (*get_vintmsk_offset)(u32 i);
+	void (*get_arb_info)(struct arb_info *arb_csrs_info);
+	void (*get_admin_info)(struct admin_info *admin_csrs_info);
 	enum dev_sku_info (*get_sku)(struct adf_hw_device_data *self);
 	int (*alloc_irq)(struct adf_accel_dev *accel_dev);
 	void (*free_irq)(struct adf_accel_dev *accel_dev);
@@ -125,19 +171,29 @@ struct adf_hw_device_data {
 	void (*get_arb_mapping)(struct adf_accel_dev *accel_dev,
 				const u32 **cfg);
 	void (*disable_iov)(struct adf_accel_dev *accel_dev);
+	void (*configure_iov_threads)(struct adf_accel_dev *accel_dev,
+				      bool enable);
 	void (*enable_ints)(struct adf_accel_dev *accel_dev);
 	int (*enable_vf2pf_comms)(struct adf_accel_dev *accel_dev);
 	void (*reset_device)(struct adf_accel_dev *accel_dev);
+	void (*set_msix_rttable)(struct adf_accel_dev *accel_dev);
+	char *(*uof_get_name)(u32 obj_num);
+	u32 (*uof_get_num_objs)(void);
+	u32 (*uof_get_ae_mask)(u32 obj_num);
+	struct adf_hw_csr_ops csr_ops;
 	const char *fw_name;
 	const char *fw_mmp_name;
 	u32 fuses;
+	u32 straps;
 	u32 accel_capabilities_mask;
 	u32 instance_id;
 	u16 accel_mask;
-	u16 ae_mask;
+	u32 ae_mask;
+	u32 admin_ae_mask;
 	u16 tx_rings_mask;
 	u8 tx_rx_gap;
 	u8 num_banks;
+	u8 num_rings_per_bank;
 	u8 num_accel;
 	u8 num_logical_accel;
 	u8 num_engines;
@@ -155,7 +211,10 @@ struct adf_hw_device_data {
 #define GET_BARS(accel_dev) ((accel_dev)->accel_pci_dev.pci_bars)
 #define GET_HW_DATA(accel_dev) (accel_dev->hw_device)
 #define GET_MAX_BANKS(accel_dev) (GET_HW_DATA(accel_dev)->num_banks)
+#define GET_NUM_RINGS_PER_BANK(accel_dev) \
+	GET_HW_DATA(accel_dev)->num_rings_per_bank
 #define GET_MAX_ACCELENGINES(accel_dev) (GET_HW_DATA(accel_dev)->num_engines)
+#define GET_CSR_OPS(accel_dev) (&(accel_dev)->hw_device->csr_ops)
 #define accel_to_pci_dev(accel_ptr) accel_ptr->accel_pci_dev.pci_dev
 
 struct adf_admin_comms;
diff --git a/drivers/crypto/qat/qat_common/adf_accel_engine.c b/drivers/crypto/qat/qat_common/adf_accel_engine.c
index c8ad85b882be..ca4eae8cdd0b 100644
--- a/drivers/crypto/qat/qat_common/adf_accel_engine.c
+++ b/drivers/crypto/qat/qat_common/adf_accel_engine.c
@@ -7,12 +7,55 @@
 #include "adf_common_drv.h"
 #include "icp_qat_uclo.h"
 
+static int adf_ae_fw_load_images(struct adf_accel_dev *accel_dev, void *fw_addr,
+				 u32 fw_size)
+{
+	struct adf_fw_loader_data *loader_data = accel_dev->fw_loader;
+	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
+	struct icp_qat_fw_loader_handle *loader;
+	char *obj_name;
+	u32 num_objs;
+	u32 ae_mask;
+	int i;
+
+	loader = loader_data->fw_loader;
+	num_objs = hw_device->uof_get_num_objs();
+
+	for (i = 0; i < num_objs; i++) {
+		obj_name = hw_device->uof_get_name(i);
+		ae_mask = hw_device->uof_get_ae_mask(i);
+
+		if (qat_uclo_set_cfg_ae_mask(loader, ae_mask)) {
+			dev_err(&GET_DEV(accel_dev),
+				"Invalid mask for UOF image\n");
+			goto out_err;
+		}
+		if (qat_uclo_map_obj(loader, fw_addr, fw_size, obj_name)) {
+			dev_err(&GET_DEV(accel_dev),
+				"Failed to map UOF firmware\n");
+			goto out_err;
+		}
+		if (qat_uclo_wr_all_uimage(loader)) {
+			dev_err(&GET_DEV(accel_dev),
+				"Failed to load UOF firmware\n");
+			goto out_err;
+		}
+		qat_uclo_del_obj(loader);
+	}
+
+	return 0;
+
+out_err:
+	adf_ae_fw_release(accel_dev);
+	return -EFAULT;
+}
+
 int adf_ae_fw_load(struct adf_accel_dev *accel_dev)
 {
 	struct adf_fw_loader_data *loader_data = accel_dev->fw_loader;
 	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
-	void *uof_addr, *mmp_addr;
-	u32 uof_size, mmp_size;
+	void *fw_addr, *mmp_addr;
+	u32 fw_size, mmp_size;
 
 	if (!hw_device->fw_name)
 		return 0;
@@ -30,15 +73,20 @@ int adf_ae_fw_load(struct adf_accel_dev *accel_dev)
 		goto out_err;
 	}
 
-	uof_size = loader_data->uof_fw->size;
-	uof_addr = (void *)loader_data->uof_fw->data;
+	fw_size = loader_data->uof_fw->size;
+	fw_addr = (void *)loader_data->uof_fw->data;
 	mmp_size = loader_data->mmp_fw->size;
 	mmp_addr = (void *)loader_data->mmp_fw->data;
+
 	if (qat_uclo_wr_mimage(loader_data->fw_loader, mmp_addr, mmp_size)) {
 		dev_err(&GET_DEV(accel_dev), "Failed to load MMP\n");
 		goto out_err;
 	}
-	if (qat_uclo_map_obj(loader_data->fw_loader, uof_addr, uof_size)) {
+
+	if (hw_device->uof_get_num_objs)
+		return adf_ae_fw_load_images(accel_dev, fw_addr, fw_size);
+
+	if (qat_uclo_map_obj(loader_data->fw_loader, fw_addr, fw_size, NULL)) {
 		dev_err(&GET_DEV(accel_dev), "Failed to map FW\n");
 		goto out_err;
 	}
@@ -61,7 +109,7 @@ void adf_ae_fw_release(struct adf_accel_dev *accel_dev)
 	if (!hw_device->fw_name)
 		return;
 
-	qat_uclo_del_uof_obj(loader_data->fw_loader);
+	qat_uclo_del_obj(loader_data->fw_loader);
 	qat_hal_deinit(loader_data->fw_loader);
 	release_firmware(loader_data->uof_fw);
 	release_firmware(loader_data->mmp_fw);
@@ -74,17 +122,12 @@ int adf_ae_start(struct adf_accel_dev *accel_dev)
 {
 	struct adf_fw_loader_data *loader_data = accel_dev->fw_loader;
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
-	u32 ae_ctr, ae, max_aes = GET_MAX_ACCELENGINES(accel_dev);
+	u32 ae_ctr;
 
 	if (!hw_data->fw_name)
 		return 0;
 
-	for (ae = 0, ae_ctr = 0; ae < max_aes; ae++) {
-		if (hw_data->ae_mask & (1 << ae)) {
-			qat_hal_start(loader_data->fw_loader, ae, 0xFF);
-			ae_ctr++;
-		}
-	}
+	ae_ctr = qat_hal_start(loader_data->fw_loader);
 	dev_info(&GET_DEV(accel_dev),
 		 "qat_dev%d started %d acceleration engines\n",
 		 accel_dev->accel_id, ae_ctr);
diff --git a/drivers/crypto/qat/qat_common/adf_admin.c b/drivers/crypto/qat/qat_common/adf_admin.c
index ec9b390276d6..43680e178242 100644
--- a/drivers/crypto/qat/qat_common/adf_admin.c
+++ b/drivers/crypto/qat/qat_common/adf_admin.c
@@ -10,11 +10,7 @@
 #include "adf_common_drv.h"
 #include "icp_qat_fw_init_admin.h"
 
-/* Admin Messages Registers */
-#define ADF_DH895XCC_ADMINMSGUR_OFFSET (0x3A000 + 0x574)
-#define ADF_DH895XCC_ADMINMSGLR_OFFSET (0x3A000 + 0x578)
-#define ADF_DH895XCC_MAILBOX_BASE_OFFSET 0x20970
-#define ADF_DH895XCC_MAILBOX_STRIDE 0x1000
+#define ADF_ADMIN_MAILBOX_STRIDE 0x1000
 #define ADF_ADMINMSG_LEN 32
 #define ADF_CONST_TABLE_SIZE 1024
 #define ADF_ADMIN_POLL_DELAY_US 20
@@ -70,28 +66,28 @@ static const u8 const_tab[1024] __aligned(1024) = {
 0xf8, 0x2b, 0xa5, 0x4f, 0xf5, 0x3a, 0x5f, 0x1d, 0x36, 0xf1, 0x51, 0x0e, 0x52,
 0x7f, 0xad, 0xe6, 0x82, 0xd1, 0x9b, 0x05, 0x68, 0x8c, 0x2b, 0x3e, 0x6c, 0x1f,
 0x1f, 0x83, 0xd9, 0xab, 0xfb, 0x41, 0xbd, 0x6b, 0x5b, 0xe0, 0xcd, 0x19, 0x13,
-0x7e, 0x21, 0x79, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x7e, 0x21, 0x79, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x80, 0x00, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x18,
 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x01, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x15, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x14, 0x02, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x14, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x15, 0x02,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x15, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x24, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x24, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x25,
+0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x12, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x43, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x45, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x45, 0x01, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x44, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x44, 0x01,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2B, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x2B, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x80, 0x00, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00,
 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
@@ -118,7 +114,7 @@ static int adf_put_admin_msg_sync(struct adf_accel_dev *accel_dev, u32 ae,
 	struct adf_admin_comms *admin = accel_dev->admin;
 	int offset = ae * ADF_ADMINMSG_LEN * 2;
 	void __iomem *mailbox = admin->mailbox_addr;
-	int mb_offset = ae * ADF_DH895XCC_MAILBOX_STRIDE;
+	int mb_offset = ae * ADF_ADMIN_MAILBOX_STRIDE;
 	struct icp_qat_fw_init_admin_req *request = in;
 
 	mutex_lock(&admin->lock);
@@ -167,7 +163,7 @@ static int adf_send_admin(struct adf_accel_dev *accel_dev,
 	return 0;
 }
 
-static int adf_init_me(struct adf_accel_dev *accel_dev)
+static int adf_init_ae(struct adf_accel_dev *accel_dev)
 {
 	struct icp_qat_fw_init_admin_req req;
 	struct icp_qat_fw_init_admin_resp resp;
@@ -176,7 +172,7 @@ static int adf_init_me(struct adf_accel_dev *accel_dev)
 
 	memset(&req, 0, sizeof(req));
 	memset(&resp, 0, sizeof(resp));
-	req.cmd_id = ICP_QAT_FW_INIT_ME;
+	req.cmd_id = ICP_QAT_FW_INIT_AE;
 
 	return adf_send_admin(accel_dev, &req, &resp, ae_mask);
 }
@@ -186,7 +182,7 @@ static int adf_set_fw_constants(struct adf_accel_dev *accel_dev)
 	struct icp_qat_fw_init_admin_req req;
 	struct icp_qat_fw_init_admin_resp resp;
 	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
-	u32 ae_mask = hw_device->ae_mask;
+	u32 ae_mask = hw_device->admin_ae_mask ?: hw_device->ae_mask;
 
 	memset(&req, 0, sizeof(req));
 	memset(&resp, 0, sizeof(resp));
@@ -210,11 +206,11 @@ int adf_send_admin_init(struct adf_accel_dev *accel_dev)
 {
 	int ret;
 
-	ret = adf_init_me(accel_dev);
+	ret = adf_set_fw_constants(accel_dev);
 	if (ret)
 		return ret;
 
-	return adf_set_fw_constants(accel_dev);
+	return adf_init_ae(accel_dev);
 }
 EXPORT_SYMBOL_GPL(adf_send_admin_init);
 
@@ -225,8 +221,9 @@ int adf_init_admin_comms(struct adf_accel_dev *accel_dev)
 	struct adf_bar *pmisc =
 		&GET_BARS(accel_dev)[hw_data->get_misc_bar_id(hw_data)];
 	void __iomem *csr = pmisc->virt_addr;
-	void __iomem *mailbox = (void __iomem *)((uintptr_t)csr +
-				 ADF_DH895XCC_MAILBOX_BASE_OFFSET);
+	struct admin_info admin_csrs_info;
+	u32 mailbox_offset, adminmsg_u, adminmsg_l;
+	void __iomem *mailbox;
 	u64 reg_val;
 
 	admin = kzalloc_node(sizeof(*accel_dev->admin), GFP_KERNEL,
@@ -254,9 +251,17 @@ int adf_init_admin_comms(struct adf_accel_dev *accel_dev)
 	}
 
 	memcpy(admin->virt_tbl_addr, const_tab, sizeof(const_tab));
+	hw_data->get_admin_info(&admin_csrs_info);
+
+	mailbox_offset = admin_csrs_info.mailbox_offset;
+	mailbox = csr + mailbox_offset;
+	adminmsg_u = admin_csrs_info.admin_msg_ur;
+	adminmsg_l = admin_csrs_info.admin_msg_lr;
+
 	reg_val = (u64)admin->phy_addr;
-	ADF_CSR_WR(csr, ADF_DH895XCC_ADMINMSGUR_OFFSET, reg_val >> 32);
-	ADF_CSR_WR(csr, ADF_DH895XCC_ADMINMSGLR_OFFSET, reg_val);
+	ADF_CSR_WR(csr, adminmsg_u, upper_32_bits(reg_val));
+	ADF_CSR_WR(csr, adminmsg_l, lower_32_bits(reg_val));
+
 	mutex_init(&admin->lock);
 	admin->mailbox_addr = mailbox;
 	accel_dev->admin = admin;
diff --git a/drivers/crypto/qat/qat_common/adf_cfg.c b/drivers/crypto/qat/qat_common/adf_cfg.c
index 22ae32838113..575b6f002303 100644
--- a/drivers/crypto/qat/qat_common/adf_cfg.c
+++ b/drivers/crypto/qat/qat_common/adf_cfg.c
@@ -196,7 +196,7 @@ static int adf_cfg_key_val_get(struct adf_accel_dev *accel_dev,
 		memcpy(val, keyval->val, ADF_CFG_MAX_VAL_LEN_IN_BYTES);
 		return 0;
 	}
-	return -1;
+	return -ENODATA;
 }
 
 /**
@@ -243,7 +243,7 @@ int adf_cfg_add_key_value_param(struct adf_accel_dev *accel_dev,
 	} else {
 		dev_err(&GET_DEV(accel_dev), "Unknown type given.\n");
 		kfree(key_val);
-		return -1;
+		return -EINVAL;
 	}
 	key_val->type = type;
 	down_write(&cfg->lock);
diff --git a/drivers/crypto/qat/qat_common/adf_cfg_common.h b/drivers/crypto/qat/qat_common/adf_cfg_common.h
index 1ef46ccfba47..4fabb70b1f18 100644
--- a/drivers/crypto/qat/qat_common/adf_cfg_common.h
+++ b/drivers/crypto/qat/qat_common/adf_cfg_common.h
@@ -32,7 +32,8 @@ enum adf_device_type {
 	DEV_C62X,
 	DEV_C62XVF,
 	DEV_C3XXX,
-	DEV_C3XXXVF
+	DEV_C3XXXVF,
+	DEV_4XXX,
 };
 
 struct adf_dev_status_info {
diff --git a/drivers/crypto/qat/qat_common/adf_cfg_strings.h b/drivers/crypto/qat/qat_common/adf_cfg_strings.h
index 314790f5b0af..09651e1f937a 100644
--- a/drivers/crypto/qat/qat_common/adf_cfg_strings.h
+++ b/drivers/crypto/qat/qat_common/adf_cfg_strings.h
@@ -18,7 +18,8 @@
 #define ADF_RING_DC_TX "RingTx"
 #define ADF_RING_DC_RX "RingRx"
 #define ADF_ETRMGR_BANK "Bank"
-#define ADF_RING_BANK_NUM "BankNumber"
+#define ADF_RING_SYM_BANK_NUM "BankSymNumber"
+#define ADF_RING_ASYM_BANK_NUM "BankAsymNumber"
 #define ADF_CY "Cy"
 #define ADF_DC "Dc"
 #define ADF_ETRMGR_COALESCING_ENABLED "InterruptCoalescingEnabled"
diff --git a/drivers/crypto/qat/qat_common/adf_common_drv.h b/drivers/crypto/qat/qat_common/adf_common_drv.h
index f22342f612c1..c61476553728 100644
--- a/drivers/crypto/qat/qat_common/adf_common_drv.h
+++ b/drivers/crypto/qat/qat_common/adf_common_drv.h
@@ -133,8 +133,7 @@ void adf_vf_isr_resource_free(struct adf_accel_dev *accel_dev);
 
 int qat_hal_init(struct adf_accel_dev *accel_dev);
 void qat_hal_deinit(struct icp_qat_fw_loader_handle *handle);
-void qat_hal_start(struct icp_qat_fw_loader_handle *handle, unsigned char ae,
-		   unsigned int ctx_mask);
+int qat_hal_start(struct icp_qat_fw_loader_handle *handle);
 void qat_hal_stop(struct icp_qat_fw_loader_handle *handle, unsigned char ae,
 		  unsigned int ctx_mask);
 void qat_hal_reset(struct icp_qat_fw_loader_handle *handle);
@@ -163,28 +162,32 @@ int qat_hal_batch_wr_lm(struct icp_qat_fw_loader_handle *handle,
 			unsigned char ae,
 			struct icp_qat_uof_batch_init *lm_init_header);
 int qat_hal_init_gpr(struct icp_qat_fw_loader_handle *handle,
-		     unsigned char ae, unsigned char ctx_mask,
+		     unsigned char ae, unsigned long ctx_mask,
 		     enum icp_qat_uof_regtype reg_type,
 		     unsigned short reg_num, unsigned int regdata);
 int qat_hal_init_wr_xfer(struct icp_qat_fw_loader_handle *handle,
-			 unsigned char ae, unsigned char ctx_mask,
+			 unsigned char ae, unsigned long ctx_mask,
 			 enum icp_qat_uof_regtype reg_type,
 			 unsigned short reg_num, unsigned int regdata);
 int qat_hal_init_rd_xfer(struct icp_qat_fw_loader_handle *handle,
-			 unsigned char ae, unsigned char ctx_mask,
+			 unsigned char ae, unsigned long ctx_mask,
 			 enum icp_qat_uof_regtype reg_type,
 			 unsigned short reg_num, unsigned int regdata);
 int qat_hal_init_nn(struct icp_qat_fw_loader_handle *handle,
-		    unsigned char ae, unsigned char ctx_mask,
+		    unsigned char ae, unsigned long ctx_mask,
 		    unsigned short reg_num, unsigned int regdata);
 int qat_hal_wr_lm(struct icp_qat_fw_loader_handle *handle,
 		  unsigned char ae, unsigned short lm_addr, unsigned int value);
+void qat_hal_set_ae_tindex_mode(struct icp_qat_fw_loader_handle *handle,
+				unsigned char ae, unsigned char mode);
 int qat_uclo_wr_all_uimage(struct icp_qat_fw_loader_handle *handle);
-void qat_uclo_del_uof_obj(struct icp_qat_fw_loader_handle *handle);
+void qat_uclo_del_obj(struct icp_qat_fw_loader_handle *handle);
 int qat_uclo_wr_mimage(struct icp_qat_fw_loader_handle *handle, void *addr_ptr,
 		       int mem_size);
 int qat_uclo_map_obj(struct icp_qat_fw_loader_handle *handle,
-		     void *addr_ptr, int mem_size);
+		     void *addr_ptr, u32 mem_size, char *obj_name);
+int qat_uclo_set_cfg_ae_mask(struct icp_qat_fw_loader_handle *handle,
+			     unsigned int cfg_ae_mask);
 #if defined(CONFIG_PCI_IOV)
 int adf_sriov_configure(struct pci_dev *pdev, int numvfs);
 void adf_disable_sriov(struct adf_accel_dev *accel_dev);
diff --git a/drivers/crypto/qat/qat_common/adf_dev_mgr.c b/drivers/crypto/qat/qat_common/adf_dev_mgr.c
index 92ec035576df..4c752eed10fe 100644
--- a/drivers/crypto/qat/qat_common/adf_dev_mgr.c
+++ b/drivers/crypto/qat/qat_common/adf_dev_mgr.c
@@ -151,8 +151,8 @@ int adf_devmgr_add_dev(struct adf_accel_dev *accel_dev,
 	mutex_lock(&table_lock);
 	atomic_set(&accel_dev->ref_count, 0);
 
-	/* PF on host or VF on guest */
-	if (!accel_dev->is_vf || (accel_dev->is_vf && !pf)) {
+	/* PF on host or VF on guest - optimized to remove redundant is_vf */
+	if (!accel_dev->is_vf || !pf) {
 		struct vf_id_map *map;
 
 		list_for_each(itr, &accel_table) {
@@ -248,7 +248,8 @@ void adf_devmgr_rm_dev(struct adf_accel_dev *accel_dev,
 		       struct adf_accel_dev *pf)
 {
 	mutex_lock(&table_lock);
-	if (!accel_dev->is_vf || (accel_dev->is_vf && !pf)) {
+	/* PF on host or VF on guest - optimized to remove redundant is_vf */
+	if (!accel_dev->is_vf || !pf) {
 		id_map[accel_dev->accel_id] = 0;
 		num_devices--;
 	} else if (accel_dev->is_vf && pf) {
@@ -285,9 +286,9 @@ struct adf_accel_dev *adf_devmgr_get_first(void)
 
 /**
  * adf_devmgr_pci_to_accel_dev() - Get accel_dev associated with the pci_dev.
- * @pci_dev:  Pointer to pci device.
+ * @pci_dev:  Pointer to PCI device.
  *
- * Function returns acceleration device associated with the given pci device.
+ * Function returns acceleration device associated with the given PCI device.
  * To be used by QAT device specific drivers.
  *
  * Return: pointer to accel_dev or NULL if not found.
diff --git a/drivers/crypto/qat/qat_common/adf_gen2_hw_data.c b/drivers/crypto/qat/qat_common/adf_gen2_hw_data.c
new file mode 100644
index 000000000000..1aa17303838d
--- /dev/null
+++ b/drivers/crypto/qat/qat_common/adf_gen2_hw_data.c
@@ -0,0 +1,181 @@
+// SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only)
+/* Copyright(c) 2020 Intel Corporation */
+#include "adf_gen2_hw_data.h"
+#include "icp_qat_hw.h"
+#include <linux/pci.h>
+
+void adf_gen2_cfg_iov_thds(struct adf_accel_dev *accel_dev, bool enable,
+			   int num_a_regs, int num_b_regs)
+{
+	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	void __iomem *pmisc_addr;
+	struct adf_bar *pmisc;
+	int pmisc_id, i;
+	u32 reg;
+
+	pmisc_id = hw_data->get_misc_bar_id(hw_data);
+	pmisc = &GET_BARS(accel_dev)[pmisc_id];
+	pmisc_addr = pmisc->virt_addr;
+
+	/* Set/Unset Valid bit in AE Thread to PCIe Function Mapping Group A */
+	for (i = 0; i < num_a_regs; i++) {
+		reg = READ_CSR_AE2FUNCTION_MAP_A(pmisc_addr, i);
+		if (enable)
+			reg |= AE2FUNCTION_MAP_VALID;
+		else
+			reg &= ~AE2FUNCTION_MAP_VALID;
+		WRITE_CSR_AE2FUNCTION_MAP_A(pmisc_addr, i, reg);
+	}
+
+	/* Set/Unset Valid bit in AE Thread to PCIe Function Mapping Group B */
+	for (i = 0; i < num_b_regs; i++) {
+		reg = READ_CSR_AE2FUNCTION_MAP_B(pmisc_addr, i);
+		if (enable)
+			reg |= AE2FUNCTION_MAP_VALID;
+		else
+			reg &= ~AE2FUNCTION_MAP_VALID;
+		WRITE_CSR_AE2FUNCTION_MAP_B(pmisc_addr, i, reg);
+	}
+}
+EXPORT_SYMBOL_GPL(adf_gen2_cfg_iov_thds);
+
+void adf_gen2_get_admin_info(struct admin_info *admin_csrs_info)
+{
+	admin_csrs_info->mailbox_offset = ADF_MAILBOX_BASE_OFFSET;
+	admin_csrs_info->admin_msg_ur = ADF_ADMINMSGUR_OFFSET;
+	admin_csrs_info->admin_msg_lr = ADF_ADMINMSGLR_OFFSET;
+}
+EXPORT_SYMBOL_GPL(adf_gen2_get_admin_info);
+
+void adf_gen2_get_arb_info(struct arb_info *arb_info)
+{
+	arb_info->arb_cfg = ADF_ARB_CONFIG;
+	arb_info->arb_offset = ADF_ARB_OFFSET;
+	arb_info->wt2sam_offset = ADF_ARB_WRK_2_SER_MAP_OFFSET;
+}
+EXPORT_SYMBOL_GPL(adf_gen2_get_arb_info);
+
+static u64 build_csr_ring_base_addr(dma_addr_t addr, u32 size)
+{
+	return BUILD_RING_BASE_ADDR(addr, size);
+}
+
+static u32 read_csr_ring_head(void __iomem *csr_base_addr, u32 bank, u32 ring)
+{
+	return READ_CSR_RING_HEAD(csr_base_addr, bank, ring);
+}
+
+static void write_csr_ring_head(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				u32 value)
+{
+	WRITE_CSR_RING_HEAD(csr_base_addr, bank, ring, value);
+}
+
+static u32 read_csr_ring_tail(void __iomem *csr_base_addr, u32 bank, u32 ring)
+{
+	return READ_CSR_RING_TAIL(csr_base_addr, bank, ring);
+}
+
+static void write_csr_ring_tail(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				u32 value)
+{
+	WRITE_CSR_RING_TAIL(csr_base_addr, bank, ring, value);
+}
+
+static u32 read_csr_e_stat(void __iomem *csr_base_addr, u32 bank)
+{
+	return READ_CSR_E_STAT(csr_base_addr, bank);
+}
+
+static void write_csr_ring_config(void __iomem *csr_base_addr, u32 bank,
+				  u32 ring, u32 value)
+{
+	WRITE_CSR_RING_CONFIG(csr_base_addr, bank, ring, value);
+}
+
+static void write_csr_ring_base(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				dma_addr_t addr)
+{
+	WRITE_CSR_RING_BASE(csr_base_addr, bank, ring, addr);
+}
+
+static void write_csr_int_flag(void __iomem *csr_base_addr, u32 bank, u32 value)
+{
+	WRITE_CSR_INT_FLAG(csr_base_addr, bank, value);
+}
+
+static void write_csr_int_srcsel(void __iomem *csr_base_addr, u32 bank)
+{
+	WRITE_CSR_INT_SRCSEL(csr_base_addr, bank);
+}
+
+static void write_csr_int_col_en(void __iomem *csr_base_addr, u32 bank,
+				 u32 value)
+{
+	WRITE_CSR_INT_COL_EN(csr_base_addr, bank, value);
+}
+
+static void write_csr_int_col_ctl(void __iomem *csr_base_addr, u32 bank,
+				  u32 value)
+{
+	WRITE_CSR_INT_COL_CTL(csr_base_addr, bank, value);
+}
+
+static void write_csr_int_flag_and_col(void __iomem *csr_base_addr, u32 bank,
+				       u32 value)
+{
+	WRITE_CSR_INT_FLAG_AND_COL(csr_base_addr, bank, value);
+}
+
+static void write_csr_ring_srv_arb_en(void __iomem *csr_base_addr, u32 bank,
+				      u32 value)
+{
+	WRITE_CSR_RING_SRV_ARB_EN(csr_base_addr, bank, value);
+}
+
+void adf_gen2_init_hw_csr_ops(struct adf_hw_csr_ops *csr_ops)
+{
+	csr_ops->build_csr_ring_base_addr = build_csr_ring_base_addr;
+	csr_ops->read_csr_ring_head = read_csr_ring_head;
+	csr_ops->write_csr_ring_head = write_csr_ring_head;
+	csr_ops->read_csr_ring_tail = read_csr_ring_tail;
+	csr_ops->write_csr_ring_tail = write_csr_ring_tail;
+	csr_ops->read_csr_e_stat = read_csr_e_stat;
+	csr_ops->write_csr_ring_config = write_csr_ring_config;
+	csr_ops->write_csr_ring_base = write_csr_ring_base;
+	csr_ops->write_csr_int_flag = write_csr_int_flag;
+	csr_ops->write_csr_int_srcsel = write_csr_int_srcsel;
+	csr_ops->write_csr_int_col_en = write_csr_int_col_en;
+	csr_ops->write_csr_int_col_ctl = write_csr_int_col_ctl;
+	csr_ops->write_csr_int_flag_and_col = write_csr_int_flag_and_col;
+	csr_ops->write_csr_ring_srv_arb_en = write_csr_ring_srv_arb_en;
+}
+EXPORT_SYMBOL_GPL(adf_gen2_init_hw_csr_ops);
+
+u32 adf_gen2_get_accel_cap(struct adf_accel_dev *accel_dev)
+{
+	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	struct pci_dev *pdev = accel_dev->accel_pci_dev.pci_dev;
+	u32 straps = hw_data->straps;
+	u32 fuses = hw_data->fuses;
+	u32 legfuses;
+	u32 capabilities = ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC |
+			   ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC |
+			   ICP_ACCEL_CAPABILITIES_AUTHENTICATION;
+
+	/* Read accelerator capabilities mask */
+	pci_read_config_dword(pdev, ADF_DEVICE_LEGFUSE_OFFSET, &legfuses);
+
+	if (legfuses & ICP_ACCEL_MASK_CIPHER_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC;
+	if (legfuses & ICP_ACCEL_MASK_PKE_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC;
+	if (legfuses & ICP_ACCEL_MASK_AUTH_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_AUTHENTICATION;
+
+	if ((straps | fuses) & ADF_POWERGATE_PKE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC;
+
+	return capabilities;
+}
+EXPORT_SYMBOL_GPL(adf_gen2_get_accel_cap);
diff --git a/drivers/crypto/qat/qat_common/adf_gen2_hw_data.h b/drivers/crypto/qat/qat_common/adf_gen2_hw_data.h
new file mode 100644
index 000000000000..3816e6500352
--- /dev/null
+++ b/drivers/crypto/qat/qat_common/adf_gen2_hw_data.h
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only) */
+/* Copyright(c) 2020 Intel Corporation */
+#ifndef ADF_GEN2_HW_DATA_H_
+#define ADF_GEN2_HW_DATA_H_
+
+#include "adf_accel_devices.h"
+
+/* Transport access */
+#define ADF_BANK_INT_SRC_SEL_MASK_0	0x4444444CUL
+#define ADF_BANK_INT_SRC_SEL_MASK_X	0x44444444UL
+#define ADF_RING_CSR_RING_CONFIG	0x000
+#define ADF_RING_CSR_RING_LBASE		0x040
+#define ADF_RING_CSR_RING_UBASE		0x080
+#define ADF_RING_CSR_RING_HEAD		0x0C0
+#define ADF_RING_CSR_RING_TAIL		0x100
+#define ADF_RING_CSR_E_STAT		0x14C
+#define ADF_RING_CSR_INT_FLAG		0x170
+#define ADF_RING_CSR_INT_SRCSEL		0x174
+#define ADF_RING_CSR_INT_SRCSEL_2	0x178
+#define ADF_RING_CSR_INT_COL_EN		0x17C
+#define ADF_RING_CSR_INT_COL_CTL	0x180
+#define ADF_RING_CSR_INT_FLAG_AND_COL	0x184
+#define ADF_RING_CSR_INT_COL_CTL_ENABLE	0x80000000
+#define ADF_RING_BUNDLE_SIZE		0x1000
+
+#define BUILD_RING_BASE_ADDR(addr, size) \
+	(((addr) >> 6) & (GENMASK_ULL(63, 0) << (size)))
+#define READ_CSR_RING_HEAD(csr_base_addr, bank, ring) \
+	ADF_CSR_RD(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_HEAD + ((ring) << 2))
+#define READ_CSR_RING_TAIL(csr_base_addr, bank, ring) \
+	ADF_CSR_RD(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_TAIL + ((ring) << 2))
+#define READ_CSR_E_STAT(csr_base_addr, bank) \
+	ADF_CSR_RD(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_E_STAT)
+#define WRITE_CSR_RING_CONFIG(csr_base_addr, bank, ring, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_CONFIG + ((ring) << 2), value)
+#define WRITE_CSR_RING_BASE(csr_base_addr, bank, ring, value) \
+do { \
+	u32 l_base = 0, u_base = 0; \
+	l_base = (u32)((value) & 0xFFFFFFFF); \
+	u_base = (u32)(((value) & 0xFFFFFFFF00000000ULL) >> 32); \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_LBASE + ((ring) << 2), l_base); \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_UBASE + ((ring) << 2), u_base); \
+} while (0)
+
+#define WRITE_CSR_RING_HEAD(csr_base_addr, bank, ring, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_HEAD + ((ring) << 2), value)
+#define WRITE_CSR_RING_TAIL(csr_base_addr, bank, ring, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_RING_TAIL + ((ring) << 2), value)
+#define WRITE_CSR_INT_FLAG(csr_base_addr, bank, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_INT_FLAG, value)
+#define WRITE_CSR_INT_SRCSEL(csr_base_addr, bank) \
+do { \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+	ADF_RING_CSR_INT_SRCSEL, ADF_BANK_INT_SRC_SEL_MASK_0); \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+	ADF_RING_CSR_INT_SRCSEL_2, ADF_BANK_INT_SRC_SEL_MASK_X); \
+} while (0)
+#define WRITE_CSR_INT_COL_EN(csr_base_addr, bank, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_INT_COL_EN, value)
+#define WRITE_CSR_INT_COL_CTL(csr_base_addr, bank, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_INT_COL_CTL, \
+		   ADF_RING_CSR_INT_COL_CTL_ENABLE | (value))
+#define WRITE_CSR_INT_FLAG_AND_COL(csr_base_addr, bank, value) \
+	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
+		   ADF_RING_CSR_INT_FLAG_AND_COL, value)
+
+/* AE to function map */
+#define AE2FUNCTION_MAP_A_OFFSET	(0x3A400 + 0x190)
+#define AE2FUNCTION_MAP_B_OFFSET	(0x3A400 + 0x310)
+#define AE2FUNCTION_MAP_REG_SIZE	4
+#define AE2FUNCTION_MAP_VALID		BIT(7)
+
+#define READ_CSR_AE2FUNCTION_MAP_A(pmisc_bar_addr, index) \
+	ADF_CSR_RD(pmisc_bar_addr, AE2FUNCTION_MAP_A_OFFSET + \
+		   AE2FUNCTION_MAP_REG_SIZE * (index))
+#define WRITE_CSR_AE2FUNCTION_MAP_A(pmisc_bar_addr, index, value) \
+	ADF_CSR_WR(pmisc_bar_addr, AE2FUNCTION_MAP_A_OFFSET + \
+		   AE2FUNCTION_MAP_REG_SIZE * (index), value)
+#define READ_CSR_AE2FUNCTION_MAP_B(pmisc_bar_addr, index) \
+	ADF_CSR_RD(pmisc_bar_addr, AE2FUNCTION_MAP_B_OFFSET + \
+		   AE2FUNCTION_MAP_REG_SIZE * (index))
+#define WRITE_CSR_AE2FUNCTION_MAP_B(pmisc_bar_addr, index, value) \
+	ADF_CSR_WR(pmisc_bar_addr, AE2FUNCTION_MAP_B_OFFSET + \
+		   AE2FUNCTION_MAP_REG_SIZE * (index), value)
+
+/* Admin Interface Offsets */
+#define ADF_ADMINMSGUR_OFFSET	(0x3A000 + 0x574)
+#define ADF_ADMINMSGLR_OFFSET	(0x3A000 + 0x578)
+#define ADF_MAILBOX_BASE_OFFSET	0x20970
+
+/* Arbiter configuration */
+#define ADF_ARB_OFFSET			0x30000
+#define ADF_ARB_WRK_2_SER_MAP_OFFSET	0x180
+#define ADF_ARB_CONFIG			(BIT(31) | BIT(6) | BIT(0))
+#define ADF_ARB_REG_SLOT		0x1000
+#define ADF_ARB_RINGSRVARBEN_OFFSET	0x19C
+
+#define WRITE_CSR_RING_SRV_ARB_EN(csr_addr, index, value) \
+	ADF_CSR_WR(csr_addr, ADF_ARB_RINGSRVARBEN_OFFSET + \
+	(ADF_ARB_REG_SLOT * (index)), value)
+
+/* Power gating */
+#define ADF_POWERGATE_PKE		BIT(24)
+
+void adf_gen2_cfg_iov_thds(struct adf_accel_dev *accel_dev, bool enable,
+			   int num_a_regs, int num_b_regs);
+void adf_gen2_init_hw_csr_ops(struct adf_hw_csr_ops *csr_ops);
+void adf_gen2_get_admin_info(struct admin_info *admin_csrs_info);
+void adf_gen2_get_arb_info(struct arb_info *arb_info);
+u32 adf_gen2_get_accel_cap(struct adf_accel_dev *accel_dev);
+
+#endif
diff --git a/drivers/crypto/qat/qat_common/adf_gen4_hw_data.c b/drivers/crypto/qat/qat_common/adf_gen4_hw_data.c
new file mode 100644
index 000000000000..b72ff58e0bc7
--- /dev/null
+++ b/drivers/crypto/qat/qat_common/adf_gen4_hw_data.c
@@ -0,0 +1,101 @@
+// SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only)
+/* Copyright(c) 2020 Intel Corporation */
+#include "adf_accel_devices.h"
+#include "adf_gen4_hw_data.h"
+
+static u64 build_csr_ring_base_addr(dma_addr_t addr, u32 size)
+{
+	return BUILD_RING_BASE_ADDR(addr, size);
+}
+
+static u32 read_csr_ring_head(void __iomem *csr_base_addr, u32 bank, u32 ring)
+{
+	return READ_CSR_RING_HEAD(csr_base_addr, bank, ring);
+}
+
+static void write_csr_ring_head(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				u32 value)
+{
+	WRITE_CSR_RING_HEAD(csr_base_addr, bank, ring, value);
+}
+
+static u32 read_csr_ring_tail(void __iomem *csr_base_addr, u32 bank, u32 ring)
+{
+	return READ_CSR_RING_TAIL(csr_base_addr, bank, ring);
+}
+
+static void write_csr_ring_tail(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				u32 value)
+{
+	WRITE_CSR_RING_TAIL(csr_base_addr, bank, ring, value);
+}
+
+static u32 read_csr_e_stat(void __iomem *csr_base_addr, u32 bank)
+{
+	return READ_CSR_E_STAT(csr_base_addr, bank);
+}
+
+static void write_csr_ring_config(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				  u32 value)
+{
+	WRITE_CSR_RING_CONFIG(csr_base_addr, bank, ring, value);
+}
+
+static void write_csr_ring_base(void __iomem *csr_base_addr, u32 bank, u32 ring,
+				dma_addr_t addr)
+{
+	WRITE_CSR_RING_BASE(csr_base_addr, bank, ring, addr);
+}
+
+static void write_csr_int_flag(void __iomem *csr_base_addr, u32 bank,
+			       u32 value)
+{
+	WRITE_CSR_INT_FLAG(csr_base_addr, bank, value);
+}
+
+static void write_csr_int_srcsel(void __iomem *csr_base_addr, u32 bank)
+{
+	WRITE_CSR_INT_SRCSEL(csr_base_addr, bank);
+}
+
+static void write_csr_int_col_en(void __iomem *csr_base_addr, u32 bank, u32 value)
+{
+	WRITE_CSR_INT_COL_EN(csr_base_addr, bank, value);
+}
+
+static void write_csr_int_col_ctl(void __iomem *csr_base_addr, u32 bank,
+				  u32 value)
+{
+	WRITE_CSR_INT_COL_CTL(csr_base_addr, bank, value);
+}
+
+static void write_csr_int_flag_and_col(void __iomem *csr_base_addr, u32 bank,
+				       u32 value)
+{
+	WRITE_CSR_INT_FLAG_AND_COL(csr_base_addr, bank, value);
+}
+
+static void write_csr_ring_srv_arb_en(void __iomem *csr_base_addr, u32 bank,
+				      u32 value)
+{
+	WRITE_CSR_RING_SRV_ARB_EN(csr_base_addr, bank, value);
+}
+
+void adf_gen4_init_hw_csr_ops(struct adf_hw_csr_ops *csr_ops)
+{
+	csr_ops->build_csr_ring_base_addr = build_csr_ring_base_addr;
+	csr_ops->read_csr_ring_head = read_csr_ring_head;
+	csr_ops->write_csr_ring_head = write_csr_ring_head;
+	csr_ops->read_csr_ring_tail = read_csr_ring_tail;
+	csr_ops->write_csr_ring_tail = write_csr_ring_tail;
+	csr_ops->read_csr_e_stat = read_csr_e_stat;
+	csr_ops->write_csr_ring_config = write_csr_ring_config;
+	csr_ops->write_csr_ring_base = write_csr_ring_base;
+	csr_ops->write_csr_int_flag = write_csr_int_flag;
+	csr_ops->write_csr_int_srcsel = write_csr_int_srcsel;
+	csr_ops->write_csr_int_col_en = write_csr_int_col_en;
+	csr_ops->write_csr_int_col_ctl = write_csr_int_col_ctl;
+	csr_ops->write_csr_int_flag_and_col = write_csr_int_flag_and_col;
+	csr_ops->write_csr_ring_srv_arb_en = write_csr_ring_srv_arb_en;
+}
+EXPORT_SYMBOL_GPL(adf_gen4_init_hw_csr_ops);
diff --git a/drivers/crypto/qat/qat_common/adf_gen4_hw_data.h b/drivers/crypto/qat/qat_common/adf_gen4_hw_data.h
new file mode 100644
index 000000000000..8ab62b2ac311
--- /dev/null
+++ b/drivers/crypto/qat/qat_common/adf_gen4_hw_data.h
@@ -0,0 +1,99 @@
+/* SPDX-License-Identifier: (BSD-3-Clause OR GPL-2.0-only) */
+/* Copyright(c) 2020 Intel Corporation */
+#ifndef ADF_GEN4_HW_CSR_DATA_H_
+#define ADF_GEN4_HW_CSR_DATA_H_
+
+#include "adf_accel_devices.h"
+
+/* Transport access */
+#define ADF_BANK_INT_SRC_SEL_MASK	0x44UL
+#define ADF_RING_CSR_RING_CONFIG	0x1000
+#define ADF_RING_CSR_RING_LBASE		0x1040
+#define ADF_RING_CSR_RING_UBASE		0x1080
+#define ADF_RING_CSR_RING_HEAD		0x0C0
+#define ADF_RING_CSR_RING_TAIL		0x100
+#define ADF_RING_CSR_E_STAT		0x14C
+#define ADF_RING_CSR_INT_FLAG		0x170
+#define ADF_RING_CSR_INT_SRCSEL		0x174
+#define ADF_RING_CSR_INT_COL_CTL	0x180
+#define ADF_RING_CSR_INT_FLAG_AND_COL	0x184
+#define ADF_RING_CSR_INT_COL_CTL_ENABLE	0x80000000
+#define ADF_RING_CSR_INT_COL_EN		0x17C
+#define ADF_RING_CSR_ADDR_OFFSET	0x100000
+#define ADF_RING_BUNDLE_SIZE		0x2000
+
+#define BUILD_RING_BASE_ADDR(addr, size) \
+	((((addr) >> 6) & (GENMASK_ULL(63, 0) << (size))) << 6)
+#define READ_CSR_RING_HEAD(csr_base_addr, bank, ring) \
+	ADF_CSR_RD((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_RING_HEAD + ((ring) << 2))
+#define READ_CSR_RING_TAIL(csr_base_addr, bank, ring) \
+	ADF_CSR_RD((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_RING_TAIL + ((ring) << 2))
+#define READ_CSR_E_STAT(csr_base_addr, bank) \
+	ADF_CSR_RD((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + ADF_RING_CSR_E_STAT)
+#define WRITE_CSR_RING_CONFIG(csr_base_addr, bank, ring, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_RING_CONFIG + ((ring) << 2), value)
+#define WRITE_CSR_RING_BASE(csr_base_addr, bank, ring, value)	\
+do { \
+	void __iomem *_csr_base_addr = csr_base_addr; \
+	u32 _bank = bank;						\
+	u32 _ring = ring;						\
+	dma_addr_t _value = value;					\
+	u32 l_base = 0, u_base = 0;					\
+	l_base = lower_32_bits(_value);					\
+	u_base = upper_32_bits(_value);					\
+	ADF_CSR_WR((_csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET,		\
+		   ADF_RING_BUNDLE_SIZE * (_bank) +			\
+		   ADF_RING_CSR_RING_LBASE + ((_ring) << 2), l_base);	\
+	ADF_CSR_WR((_csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET,		\
+		   ADF_RING_BUNDLE_SIZE * (_bank) +			\
+		   ADF_RING_CSR_RING_UBASE + ((_ring) << 2), u_base);	\
+} while (0)
+
+#define WRITE_CSR_RING_HEAD(csr_base_addr, bank, ring, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_RING_HEAD + ((ring) << 2), value)
+#define WRITE_CSR_RING_TAIL(csr_base_addr, bank, ring, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_RING_TAIL + ((ring) << 2), value)
+#define WRITE_CSR_INT_FLAG(csr_base_addr, bank, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_INT_FLAG, (value))
+#define WRITE_CSR_INT_SRCSEL(csr_base_addr, bank) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_INT_SRCSEL, ADF_BANK_INT_SRC_SEL_MASK)
+#define WRITE_CSR_INT_COL_EN(csr_base_addr, bank, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_INT_COL_EN, (value))
+#define WRITE_CSR_INT_COL_CTL(csr_base_addr, bank, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_INT_COL_CTL, \
+		   ADF_RING_CSR_INT_COL_CTL_ENABLE | (value))
+#define WRITE_CSR_INT_FLAG_AND_COL(csr_base_addr, bank, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_INT_FLAG_AND_COL, (value))
+
+/* Arbiter configuration */
+#define ADF_RING_CSR_RING_SRV_ARB_EN 0x19C
+
+#define WRITE_CSR_RING_SRV_ARB_EN(csr_base_addr, bank, value) \
+	ADF_CSR_WR((csr_base_addr) + ADF_RING_CSR_ADDR_OFFSET, \
+		   ADF_RING_BUNDLE_SIZE * (bank) + \
+		   ADF_RING_CSR_RING_SRV_ARB_EN, (value))
+
+void adf_gen4_init_hw_csr_ops(struct adf_hw_csr_ops *csr_ops);
+
+#endif
diff --git a/drivers/crypto/qat/qat_common/adf_hw_arbiter.c b/drivers/crypto/qat/qat_common/adf_hw_arbiter.c
index d4162783f970..9f5240d9488b 100644
--- a/drivers/crypto/qat/qat_common/adf_hw_arbiter.c
+++ b/drivers/crypto/qat/qat_common/adf_hw_arbiter.c
@@ -6,48 +6,33 @@
 
 #define ADF_ARB_NUM 4
 #define ADF_ARB_REG_SIZE 0x4
-#define ADF_ARB_WTR_SIZE 0x20
-#define ADF_ARB_OFFSET 0x30000
-#define ADF_ARB_REG_SLOT 0x1000
-#define ADF_ARB_WTR_OFFSET 0x010
-#define ADF_ARB_RO_EN_OFFSET 0x090
-#define ADF_ARB_WQCFG_OFFSET 0x100
-#define ADF_ARB_WRK_2_SER_MAP_OFFSET 0x180
-#define ADF_ARB_RINGSRVARBEN_OFFSET 0x19C
-
-#define WRITE_CSR_ARB_RINGSRVARBEN(csr_addr, index, value) \
-	ADF_CSR_WR(csr_addr, ADF_ARB_RINGSRVARBEN_OFFSET + \
-	(ADF_ARB_REG_SLOT * index), value)
-
-#define WRITE_CSR_ARB_SARCONFIG(csr_addr, index, value) \
-	ADF_CSR_WR(csr_addr, ADF_ARB_OFFSET + \
-	(ADF_ARB_REG_SIZE * index), value)
-
-#define WRITE_CSR_ARB_WRK_2_SER_MAP(csr_addr, index, value) \
-	ADF_CSR_WR(csr_addr, (ADF_ARB_OFFSET + \
-	ADF_ARB_WRK_2_SER_MAP_OFFSET) + \
-	(ADF_ARB_REG_SIZE * index), value)
-
-#define WRITE_CSR_ARB_WQCFG(csr_addr, index, value) \
-	ADF_CSR_WR(csr_addr, (ADF_ARB_OFFSET + \
-	ADF_ARB_WQCFG_OFFSET) + (ADF_ARB_REG_SIZE * index), value)
+
+#define WRITE_CSR_ARB_SARCONFIG(csr_addr, arb_offset, index, value) \
+	ADF_CSR_WR(csr_addr, (arb_offset) + \
+	(ADF_ARB_REG_SIZE * (index)), value)
+
+#define WRITE_CSR_ARB_WT2SAM(csr_addr, arb_offset, wt_offset, index, value) \
+	ADF_CSR_WR(csr_addr, ((arb_offset) + (wt_offset)) + \
+	(ADF_ARB_REG_SIZE * (index)), value)
 
 int adf_init_arb(struct adf_accel_dev *accel_dev)
 {
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
 	void __iomem *csr = accel_dev->transport->banks[0].csr_addr;
-	u32 arb_cfg = 0x1 << 31 | 0x4 << 4 | 0x1;
-	u32 arb, i;
+	u32 arb_off, wt_off, arb_cfg;
 	const u32 *thd_2_arb_cfg;
+	struct arb_info info;
+	int arb, i;
+
+	hw_data->get_arb_info(&info);
+	arb_cfg = info.arb_cfg;
+	arb_off = info.arb_offset;
+	wt_off = info.wt2sam_offset;
 
 	/* Service arb configured for 32 bytes responses and
 	 * ring flow control check enabled. */
 	for (arb = 0; arb < ADF_ARB_NUM; arb++)
-		WRITE_CSR_ARB_SARCONFIG(csr, arb, arb_cfg);
-
-	/* Setup worker queue registers */
-	for (i = 0; i < hw_data->num_engines; i++)
-		WRITE_CSR_ARB_WQCFG(csr, i, i);
+		WRITE_CSR_ARB_SARCONFIG(csr, arb_off, arb, arb_cfg);
 
 	/* Map worker threads to service arbiters */
 	hw_data->get_arb_mapping(accel_dev, &thd_2_arb_cfg);
@@ -56,7 +41,7 @@ int adf_init_arb(struct adf_accel_dev *accel_dev)
 		return -EFAULT;
 
 	for (i = 0; i < hw_data->num_engines; i++)
-		WRITE_CSR_ARB_WRK_2_SER_MAP(csr, i, *(thd_2_arb_cfg + i));
+		WRITE_CSR_ARB_WT2SAM(csr, arb_off, wt_off, i, thd_2_arb_cfg[i]);
 
 	return 0;
 }
@@ -64,36 +49,59 @@ EXPORT_SYMBOL_GPL(adf_init_arb);
 
 void adf_update_ring_arb(struct adf_etr_ring_data *ring)
 {
-	WRITE_CSR_ARB_RINGSRVARBEN(ring->bank->csr_addr,
-				   ring->bank->bank_number,
-				   ring->bank->ring_mask & 0xFF);
+	struct adf_accel_dev *accel_dev = ring->bank->accel_dev;
+	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(accel_dev);
+	u32 tx_ring_mask = hw_data->tx_rings_mask;
+	u32 shift = hw_data->tx_rx_gap;
+	u32 arben, arben_tx, arben_rx;
+	u32 rx_ring_mask;
+
+	/*
+	 * Enable arbitration on a ring only if the TX half of the ring mask
+	 * matches the RX part. This results in writes to CSR on both TX and
+	 * RX update - only one is necessary, but both are done for
+	 * simplicity.
+	 */
+	rx_ring_mask = tx_ring_mask << shift;
+	arben_tx = (ring->bank->ring_mask & tx_ring_mask) >> 0;
+	arben_rx = (ring->bank->ring_mask & rx_ring_mask) >> shift;
+	arben = arben_tx & arben_rx;
+
+	csr_ops->write_csr_ring_srv_arb_en(ring->bank->csr_addr,
+					   ring->bank->bank_number, arben);
 }
 
 void adf_exit_arb(struct adf_accel_dev *accel_dev)
 {
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(accel_dev);
+	u32 arb_off, wt_off;
+	struct arb_info info;
 	void __iomem *csr;
 	unsigned int i;
 
+	hw_data->get_arb_info(&info);
+	arb_off = info.arb_offset;
+	wt_off = info.wt2sam_offset;
+
 	if (!accel_dev->transport)
 		return;
 
 	csr = accel_dev->transport->banks[0].csr_addr;
 
+	hw_data->get_arb_info(&info);
+
 	/* Reset arbiter configuration */
 	for (i = 0; i < ADF_ARB_NUM; i++)
-		WRITE_CSR_ARB_SARCONFIG(csr, i, 0);
-
-	/* Shutdown work queue */
-	for (i = 0; i < hw_data->num_engines; i++)
-		WRITE_CSR_ARB_WQCFG(csr, i, 0);
+		WRITE_CSR_ARB_SARCONFIG(csr, arb_off, i, 0);
 
 	/* Unmap worker threads to service arbiters */
 	for (i = 0; i < hw_data->num_engines; i++)
-		WRITE_CSR_ARB_WRK_2_SER_MAP(csr, i, 0);
+		WRITE_CSR_ARB_WT2SAM(csr, arb_off, wt_off, i, 0);
 
 	/* Disable arbitration on all rings */
 	for (i = 0; i < GET_MAX_BANKS(accel_dev); i++)
-		WRITE_CSR_ARB_RINGSRVARBEN(csr, i, 0);
+		csr_ops->write_csr_ring_srv_arb_en(csr, i, 0);
 }
 EXPORT_SYMBOL_GPL(adf_exit_arb);
diff --git a/drivers/crypto/qat/qat_common/adf_isr.c b/drivers/crypto/qat/qat_common/adf_isr.c
index 36136f7db509..c45853463530 100644
--- a/drivers/crypto/qat/qat_common/adf_isr.c
+++ b/drivers/crypto/qat/qat_common/adf_isr.c
@@ -21,6 +21,9 @@ static int adf_enable_msix(struct adf_accel_dev *accel_dev)
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
 	u32 msix_num_entries = 1;
 
+	if (hw_data->set_msix_rttable)
+		hw_data->set_msix_rttable(accel_dev);
+
 	/* If SR-IOV is disabled, add entries for each bank */
 	if (!accel_dev->pf.vf_info) {
 		int i;
@@ -50,8 +53,10 @@ static void adf_disable_msix(struct adf_accel_pci *pci_dev_info)
 static irqreturn_t adf_msix_isr_bundle(int irq, void *bank_ptr)
 {
 	struct adf_etr_bank_data *bank = bank_ptr;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
 
-	WRITE_CSR_INT_FLAG_AND_COL(bank->csr_addr, bank->bank_number, 0);
+	csr_ops->write_csr_int_flag_and_col(bank->csr_addr, bank->bank_number,
+					    0);
 	tasklet_hi_schedule(&bank->resp_handler);
 	return IRQ_HANDLED;
 }
diff --git a/drivers/crypto/qat/qat_common/adf_sriov.c b/drivers/crypto/qat/qat_common/adf_sriov.c
index 963b2bea78f2..8c822c2861c2 100644
--- a/drivers/crypto/qat/qat_common/adf_sriov.c
+++ b/drivers/crypto/qat/qat_common/adf_sriov.c
@@ -10,31 +10,6 @@
 
 static struct workqueue_struct *pf2vf_resp_wq;
 
-#define ME2FUNCTION_MAP_A_OFFSET	(0x3A400 + 0x190)
-#define ME2FUNCTION_MAP_A_NUM_REGS	96
-
-#define ME2FUNCTION_MAP_B_OFFSET	(0x3A400 + 0x310)
-#define ME2FUNCTION_MAP_B_NUM_REGS	12
-
-#define ME2FUNCTION_MAP_REG_SIZE	4
-#define ME2FUNCTION_MAP_VALID		BIT(7)
-
-#define READ_CSR_ME2FUNCTION_MAP_A(pmisc_bar_addr, index)		\
-	ADF_CSR_RD(pmisc_bar_addr, ME2FUNCTION_MAP_A_OFFSET +		\
-		   ME2FUNCTION_MAP_REG_SIZE * index)
-
-#define WRITE_CSR_ME2FUNCTION_MAP_A(pmisc_bar_addr, index, value)	\
-	ADF_CSR_WR(pmisc_bar_addr, ME2FUNCTION_MAP_A_OFFSET +		\
-		   ME2FUNCTION_MAP_REG_SIZE * index, value)
-
-#define READ_CSR_ME2FUNCTION_MAP_B(pmisc_bar_addr, index)		\
-	ADF_CSR_RD(pmisc_bar_addr, ME2FUNCTION_MAP_B_OFFSET +		\
-		   ME2FUNCTION_MAP_REG_SIZE * index)
-
-#define WRITE_CSR_ME2FUNCTION_MAP_B(pmisc_bar_addr, index, value)	\
-	ADF_CSR_WR(pmisc_bar_addr, ME2FUNCTION_MAP_B_OFFSET +		\
-		   ME2FUNCTION_MAP_REG_SIZE * index, value)
-
 struct adf_pf2vf_resp {
 	struct work_struct pf2vf_resp_work;
 	struct adf_accel_vf_info *vf_info;
@@ -68,12 +43,8 @@ static int adf_enable_sriov(struct adf_accel_dev *accel_dev)
 	struct pci_dev *pdev = accel_to_pci_dev(accel_dev);
 	int totalvfs = pci_sriov_get_totalvfs(pdev);
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
-	struct adf_bar *pmisc =
-			&GET_BARS(accel_dev)[hw_data->get_misc_bar_id(hw_data)];
-	void __iomem *pmisc_addr = pmisc->virt_addr;
 	struct adf_accel_vf_info *vf_info;
 	int i;
-	u32 reg;
 
 	for (i = 0, vf_info = accel_dev->pf.vf_info; i < totalvfs;
 	     i++, vf_info++) {
@@ -90,22 +61,13 @@ static int adf_enable_sriov(struct adf_accel_dev *accel_dev)
 				     DEFAULT_RATELIMIT_BURST);
 	}
 
-	/* Set Valid bits in ME Thread to PCIe Function Mapping Group A */
-	for (i = 0; i < ME2FUNCTION_MAP_A_NUM_REGS; i++) {
-		reg = READ_CSR_ME2FUNCTION_MAP_A(pmisc_addr, i);
-		reg |= ME2FUNCTION_MAP_VALID;
-		WRITE_CSR_ME2FUNCTION_MAP_A(pmisc_addr, i, reg);
-	}
-
-	/* Set Valid bits in ME Thread to PCIe Function Mapping Group B */
-	for (i = 0; i < ME2FUNCTION_MAP_B_NUM_REGS; i++) {
-		reg = READ_CSR_ME2FUNCTION_MAP_B(pmisc_addr, i);
-		reg |= ME2FUNCTION_MAP_VALID;
-		WRITE_CSR_ME2FUNCTION_MAP_B(pmisc_addr, i, reg);
-	}
+	/* Set Valid bits in AE Thread to PCIe Function Mapping */
+	if (hw_data->configure_iov_threads)
+		hw_data->configure_iov_threads(accel_dev, true);
 
 	/* Enable VF to PF interrupts for all VFs */
-	adf_enable_vf2pf_interrupts(accel_dev, GENMASK_ULL(totalvfs - 1, 0));
+	if (hw_data->get_pf2vf_offset)
+		adf_enable_vf2pf_interrupts(accel_dev, BIT_ULL(totalvfs) - 1);
 
 	/*
 	 * Due to the hardware design, when SR-IOV and the ring arbiter
@@ -127,37 +89,25 @@ static int adf_enable_sriov(struct adf_accel_dev *accel_dev)
 void adf_disable_sriov(struct adf_accel_dev *accel_dev)
 {
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
-	struct adf_bar *pmisc =
-			&GET_BARS(accel_dev)[hw_data->get_misc_bar_id(hw_data)];
-	void __iomem *pmisc_addr = pmisc->virt_addr;
 	int totalvfs = pci_sriov_get_totalvfs(accel_to_pci_dev(accel_dev));
 	struct adf_accel_vf_info *vf;
-	u32 reg;
 	int i;
 
 	if (!accel_dev->pf.vf_info)
 		return;
 
-	adf_pf2vf_notify_restarting(accel_dev);
+	if (hw_data->get_pf2vf_offset)
+		adf_pf2vf_notify_restarting(accel_dev);
 
 	pci_disable_sriov(accel_to_pci_dev(accel_dev));
 
 	/* Disable VF to PF interrupts */
-	adf_disable_vf2pf_interrupts(accel_dev, 0xFFFFFFFF);
+	if (hw_data->get_pf2vf_offset)
+		adf_disable_vf2pf_interrupts(accel_dev, GENMASK(31, 0));
 
-	/* Clear Valid bits in ME Thread to PCIe Function Mapping Group A */
-	for (i = 0; i < ME2FUNCTION_MAP_A_NUM_REGS; i++) {
-		reg = READ_CSR_ME2FUNCTION_MAP_A(pmisc_addr, i);
-		reg &= ~ME2FUNCTION_MAP_VALID;
-		WRITE_CSR_ME2FUNCTION_MAP_A(pmisc_addr, i, reg);
-	}
-
-	/* Clear Valid bits in ME Thread to PCIe Function Mapping Group B */
-	for (i = 0; i < ME2FUNCTION_MAP_B_NUM_REGS; i++) {
-		reg = READ_CSR_ME2FUNCTION_MAP_B(pmisc_addr, i);
-		reg &= ~ME2FUNCTION_MAP_VALID;
-		WRITE_CSR_ME2FUNCTION_MAP_B(pmisc_addr, i, reg);
-	}
+	/* Clear Valid bits in AE Thread to PCIe Function Mapping */
+	if (hw_data->configure_iov_threads)
+		hw_data->configure_iov_threads(accel_dev, false);
 
 	for (i = 0, vf = accel_dev->pf.vf_info; i < totalvfs; i++, vf++) {
 		tasklet_disable(&vf->vf2pf_bh_tasklet);
@@ -172,13 +122,13 @@ EXPORT_SYMBOL_GPL(adf_disable_sriov);
 
 /**
  * adf_sriov_configure() - Enable SRIOV for the device
- * @pdev:  Pointer to pci device.
+ * @pdev:  Pointer to PCI device.
  * @numvfs: Number of virtual functions (VFs) to enable.
  *
  * Note that the @numvfs parameter is ignored and all VFs supported by the
  * device are enabled due to the design of the hardware.
  *
- * Function enables SRIOV for the pci device.
+ * Function enables SRIOV for the PCI device.
  *
  * Return: number of VFs enabled on success, error code otherwise.
  */
diff --git a/drivers/crypto/qat/qat_common/adf_transport.c b/drivers/crypto/qat/qat_common/adf_transport.c
index 2ad774017200..5a7030acdc33 100644
--- a/drivers/crypto/qat/qat_common/adf_transport.c
+++ b/drivers/crypto/qat/qat_common/adf_transport.c
@@ -54,24 +54,32 @@ static void adf_unreserve_ring(struct adf_etr_bank_data *bank, u32 ring)
 
 static void adf_enable_ring_irq(struct adf_etr_bank_data *bank, u32 ring)
 {
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
+
 	spin_lock_bh(&bank->lock);
 	bank->irq_mask |= (1 << ring);
 	spin_unlock_bh(&bank->lock);
-	WRITE_CSR_INT_COL_EN(bank->csr_addr, bank->bank_number, bank->irq_mask);
-	WRITE_CSR_INT_COL_CTL(bank->csr_addr, bank->bank_number,
-			      bank->irq_coalesc_timer);
+	csr_ops->write_csr_int_col_en(bank->csr_addr, bank->bank_number,
+				      bank->irq_mask);
+	csr_ops->write_csr_int_col_ctl(bank->csr_addr, bank->bank_number,
+				       bank->irq_coalesc_timer);
 }
 
 static void adf_disable_ring_irq(struct adf_etr_bank_data *bank, u32 ring)
 {
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
+
 	spin_lock_bh(&bank->lock);
 	bank->irq_mask &= ~(1 << ring);
 	spin_unlock_bh(&bank->lock);
-	WRITE_CSR_INT_COL_EN(bank->csr_addr, bank->bank_number, bank->irq_mask);
+	csr_ops->write_csr_int_col_en(bank->csr_addr, bank->bank_number,
+				      bank->irq_mask);
 }
 
 int adf_send_message(struct adf_etr_ring_data *ring, u32 *msg)
 {
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(ring->bank->accel_dev);
+
 	if (atomic_add_return(1, ring->inflights) >
 	    ADF_MAX_INFLIGHTS(ring->ring_size, ring->msg_size)) {
 		atomic_dec(ring->inflights);
@@ -84,14 +92,17 @@ int adf_send_message(struct adf_etr_ring_data *ring, u32 *msg)
 	ring->tail = adf_modulo(ring->tail +
 				ADF_MSG_SIZE_TO_BYTES(ring->msg_size),
 				ADF_RING_SIZE_MODULO(ring->ring_size));
-	WRITE_CSR_RING_TAIL(ring->bank->csr_addr, ring->bank->bank_number,
-			    ring->ring_number, ring->tail);
+	csr_ops->write_csr_ring_tail(ring->bank->csr_addr,
+				     ring->bank->bank_number, ring->ring_number,
+				     ring->tail);
 	spin_unlock_bh(&ring->lock);
+
 	return 0;
 }
 
 static int adf_handle_response(struct adf_etr_ring_data *ring)
 {
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(ring->bank->accel_dev);
 	u32 msg_counter = 0;
 	u32 *msg = (u32 *)((uintptr_t)ring->base_addr + ring->head);
 
@@ -105,30 +116,36 @@ static int adf_handle_response(struct adf_etr_ring_data *ring)
 		msg_counter++;
 		msg = (u32 *)((uintptr_t)ring->base_addr + ring->head);
 	}
-	if (msg_counter > 0)
-		WRITE_CSR_RING_HEAD(ring->bank->csr_addr,
-				    ring->bank->bank_number,
-				    ring->ring_number, ring->head);
+	if (msg_counter > 0) {
+		csr_ops->write_csr_ring_head(ring->bank->csr_addr,
+					     ring->bank->bank_number,
+					     ring->ring_number, ring->head);
+	}
 	return 0;
 }
 
 static void adf_configure_tx_ring(struct adf_etr_ring_data *ring)
 {
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(ring->bank->accel_dev);
 	u32 ring_config = BUILD_RING_CONFIG(ring->ring_size);
 
-	WRITE_CSR_RING_CONFIG(ring->bank->csr_addr, ring->bank->bank_number,
-			      ring->ring_number, ring_config);
+	csr_ops->write_csr_ring_config(ring->bank->csr_addr,
+				       ring->bank->bank_number,
+				       ring->ring_number, ring_config);
+
 }
 
 static void adf_configure_rx_ring(struct adf_etr_ring_data *ring)
 {
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(ring->bank->accel_dev);
 	u32 ring_config =
 			BUILD_RESP_RING_CONFIG(ring->ring_size,
 					       ADF_RING_NEAR_WATERMARK_512,
 					       ADF_RING_NEAR_WATERMARK_0);
 
-	WRITE_CSR_RING_CONFIG(ring->bank->csr_addr, ring->bank->bank_number,
-			      ring->ring_number, ring_config);
+	csr_ops->write_csr_ring_config(ring->bank->csr_addr,
+				       ring->bank->bank_number,
+				       ring->ring_number, ring_config);
 }
 
 static int adf_init_ring(struct adf_etr_ring_data *ring)
@@ -136,6 +153,7 @@ static int adf_init_ring(struct adf_etr_ring_data *ring)
 	struct adf_etr_bank_data *bank = ring->bank;
 	struct adf_accel_dev *accel_dev = bank->accel_dev;
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(accel_dev);
 	u64 ring_base;
 	u32 ring_size_bytes =
 			ADF_SIZE_TO_RING_SIZE_IN_BYTES(ring->ring_size);
@@ -162,9 +180,12 @@ static int adf_init_ring(struct adf_etr_ring_data *ring)
 	else
 		adf_configure_rx_ring(ring);
 
-	ring_base = BUILD_RING_BASE_ADDR(ring->dma_addr, ring->ring_size);
-	WRITE_CSR_RING_BASE(ring->bank->csr_addr, ring->bank->bank_number,
-			    ring->ring_number, ring_base);
+	ring_base = csr_ops->build_csr_ring_base_addr(ring->dma_addr,
+						      ring->ring_size);
+
+	csr_ops->write_csr_ring_base(ring->bank->csr_addr,
+				     ring->bank->bank_number, ring->ring_number,
+				     ring_base);
 	spin_lock_init(&ring->lock);
 	return 0;
 }
@@ -190,6 +211,7 @@ int adf_create_ring(struct adf_accel_dev *accel_dev, const char *section,
 		    struct adf_etr_ring_data **ring_ptr)
 {
 	struct adf_etr_data *transport_data = accel_dev->transport;
+	u8 num_rings_per_bank = GET_NUM_RINGS_PER_BANK(accel_dev);
 	struct adf_etr_bank_data *bank;
 	struct adf_etr_ring_data *ring;
 	char val[ADF_CFG_MAX_VAL_LEN_IN_BYTES];
@@ -219,7 +241,7 @@ int adf_create_ring(struct adf_accel_dev *accel_dev, const char *section,
 		dev_err(&GET_DEV(accel_dev), "Can't get ring number\n");
 		return -EFAULT;
 	}
-	if (ring_num >= ADF_ETR_MAX_RINGS_PER_BANK) {
+	if (ring_num >= num_rings_per_bank) {
 		dev_err(&GET_DEV(accel_dev), "Invalid ring number\n");
 		return -EFAULT;
 	}
@@ -268,15 +290,17 @@ err:
 void adf_remove_ring(struct adf_etr_ring_data *ring)
 {
 	struct adf_etr_bank_data *bank = ring->bank;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
 
 	/* Disable interrupts for the given ring */
 	adf_disable_ring_irq(bank, ring->ring_number);
 
 	/* Clear PCI config space */
-	WRITE_CSR_RING_CONFIG(bank->csr_addr, bank->bank_number,
-			      ring->ring_number, 0);
-	WRITE_CSR_RING_BASE(bank->csr_addr, bank->bank_number,
-			    ring->ring_number, 0);
+
+	csr_ops->write_csr_ring_config(bank->csr_addr, bank->bank_number,
+				       ring->ring_number, 0);
+	csr_ops->write_csr_ring_base(bank->csr_addr, bank->bank_number,
+				     ring->ring_number, 0);
 	adf_ring_debugfs_rm(ring);
 	adf_unreserve_ring(bank, ring->ring_number);
 	/* Disable HW arbitration for the given ring */
@@ -286,25 +310,30 @@ void adf_remove_ring(struct adf_etr_ring_data *ring)
 
 static void adf_ring_response_handler(struct adf_etr_bank_data *bank)
 {
-	u32 empty_rings, i;
+	struct adf_accel_dev *accel_dev = bank->accel_dev;
+	u8 num_rings_per_bank = GET_NUM_RINGS_PER_BANK(accel_dev);
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(accel_dev);
+	unsigned long empty_rings;
+	int i;
 
-	empty_rings = READ_CSR_E_STAT(bank->csr_addr, bank->bank_number);
+	empty_rings = csr_ops->read_csr_e_stat(bank->csr_addr,
+					       bank->bank_number);
 	empty_rings = ~empty_rings & bank->irq_mask;
 
-	for (i = 0; i < ADF_ETR_MAX_RINGS_PER_BANK; ++i) {
-		if (empty_rings & (1 << i))
-			adf_handle_response(&bank->rings[i]);
-	}
+	for_each_set_bit(i, &empty_rings, num_rings_per_bank)
+		adf_handle_response(&bank->rings[i]);
 }
 
 void adf_response_handler(uintptr_t bank_addr)
 {
 	struct adf_etr_bank_data *bank = (void *)bank_addr;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
 
 	/* Handle all the responses and reenable IRQs */
 	adf_ring_response_handler(bank);
-	WRITE_CSR_INT_FLAG_AND_COL(bank->csr_addr, bank->bank_number,
-				   bank->irq_mask);
+
+	csr_ops->write_csr_int_flag_and_col(bank->csr_addr, bank->bank_number,
+					    bank->irq_mask);
 }
 
 static inline int adf_get_cfg_int(struct adf_accel_dev *accel_dev,
@@ -343,9 +372,14 @@ static int adf_init_bank(struct adf_accel_dev *accel_dev,
 			 u32 bank_num, void __iomem *csr_addr)
 {
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	u8 num_rings_per_bank = hw_data->num_rings_per_bank;
+	struct adf_hw_csr_ops *csr_ops = &hw_data->csr_ops;
+	u32 irq_mask = BIT(num_rings_per_bank) - 1;
 	struct adf_etr_ring_data *ring;
 	struct adf_etr_ring_data *tx_ring;
 	u32 i, coalesc_enabled = 0;
+	unsigned long ring_mask;
+	int size;
 
 	memset(bank, 0, sizeof(*bank));
 	bank->bank_number = bank_num;
@@ -353,6 +387,13 @@ static int adf_init_bank(struct adf_accel_dev *accel_dev,
 	bank->accel_dev = accel_dev;
 	spin_lock_init(&bank->lock);
 
+	/* Allocate the rings in the bank */
+	size = num_rings_per_bank * sizeof(struct adf_etr_ring_data);
+	bank->rings = kzalloc_node(size, GFP_KERNEL,
+				   dev_to_node(&GET_DEV(accel_dev)));
+	if (!bank->rings)
+		return -ENOMEM;
+
 	/* Enable IRQ coalescing always. This will allow to use
 	 * the optimised flag and coalesc register.
 	 * If it is disabled in the config file just use min time value */
@@ -363,9 +404,10 @@ static int adf_init_bank(struct adf_accel_dev *accel_dev,
 	else
 		bank->irq_coalesc_timer = ADF_COALESCING_MIN_TIME;
 
-	for (i = 0; i < ADF_ETR_MAX_RINGS_PER_BANK; i++) {
-		WRITE_CSR_RING_CONFIG(csr_addr, bank_num, i, 0);
-		WRITE_CSR_RING_BASE(csr_addr, bank_num, i, 0);
+	for (i = 0; i < num_rings_per_bank; i++) {
+		csr_ops->write_csr_ring_config(csr_addr, bank_num, i, 0);
+		csr_ops->write_csr_ring_base(csr_addr, bank_num, i, 0);
+
 		ring = &bank->rings[i];
 		if (hw_data->tx_rings_mask & (1 << i)) {
 			ring->inflights =
@@ -390,15 +432,18 @@ static int adf_init_bank(struct adf_accel_dev *accel_dev,
 		goto err;
 	}
 
-	WRITE_CSR_INT_FLAG(csr_addr, bank_num, ADF_BANK_INT_FLAG_CLEAR_MASK);
-	WRITE_CSR_INT_SRCSEL(csr_addr, bank_num);
+	csr_ops->write_csr_int_flag(csr_addr, bank_num, irq_mask);
+	csr_ops->write_csr_int_srcsel(csr_addr, bank_num);
+
 	return 0;
 err:
-	for (i = 0; i < ADF_ETR_MAX_RINGS_PER_BANK; i++) {
+	ring_mask = hw_data->tx_rings_mask;
+	for_each_set_bit(i, &ring_mask, num_rings_per_bank) {
 		ring = &bank->rings[i];
-		if (hw_data->tx_rings_mask & (1 << i))
-			kfree(ring->inflights);
+		kfree(ring->inflights);
+		ring->inflights = NULL;
 	}
+	kfree(bank->rings);
 	return -ENOMEM;
 }
 
@@ -464,11 +509,12 @@ EXPORT_SYMBOL_GPL(adf_init_etr_data);
 
 static void cleanup_bank(struct adf_etr_bank_data *bank)
 {
+	struct adf_accel_dev *accel_dev = bank->accel_dev;
+	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	u8 num_rings_per_bank = hw_data->num_rings_per_bank;
 	u32 i;
 
-	for (i = 0; i < ADF_ETR_MAX_RINGS_PER_BANK; i++) {
-		struct adf_accel_dev *accel_dev = bank->accel_dev;
-		struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	for (i = 0; i < num_rings_per_bank; i++) {
 		struct adf_etr_ring_data *ring = &bank->rings[i];
 
 		if (bank->ring_mask & (1 << i))
@@ -477,6 +523,7 @@ static void cleanup_bank(struct adf_etr_bank_data *bank)
 		if (hw_data->tx_rings_mask & (1 << i))
 			kfree(ring->inflights);
 	}
+	kfree(bank->rings);
 	adf_bank_debugfs_rm(bank);
 	memset(bank, 0, sizeof(*bank));
 }
@@ -507,6 +554,7 @@ void adf_cleanup_etr_data(struct adf_accel_dev *accel_dev)
 	if (etr_data) {
 		adf_cleanup_etr_handles(accel_dev);
 		debugfs_remove(etr_data->debug);
+		kfree(etr_data->banks->rings);
 		kfree(etr_data->banks);
 		kfree(etr_data);
 		accel_dev->transport = NULL;
diff --git a/drivers/crypto/qat/qat_common/adf_transport_access_macros.h b/drivers/crypto/qat/qat_common/adf_transport_access_macros.h
index 950d1988556c..3b6b0267bbec 100644
--- a/drivers/crypto/qat/qat_common/adf_transport_access_macros.h
+++ b/drivers/crypto/qat/qat_common/adf_transport_access_macros.h
@@ -4,23 +4,6 @@
 #define ADF_TRANSPORT_ACCESS_MACROS_H
 
 #include "adf_accel_devices.h"
-#define ADF_BANK_INT_SRC_SEL_MASK_0 0x4444444CUL
-#define ADF_BANK_INT_SRC_SEL_MASK_X 0x44444444UL
-#define ADF_BANK_INT_FLAG_CLEAR_MASK 0xFFFF
-#define ADF_RING_CSR_RING_CONFIG 0x000
-#define ADF_RING_CSR_RING_LBASE 0x040
-#define ADF_RING_CSR_RING_UBASE 0x080
-#define ADF_RING_CSR_RING_HEAD 0x0C0
-#define ADF_RING_CSR_RING_TAIL 0x100
-#define ADF_RING_CSR_E_STAT 0x14C
-#define ADF_RING_CSR_INT_FLAG	0x170
-#define ADF_RING_CSR_INT_SRCSEL 0x174
-#define ADF_RING_CSR_INT_SRCSEL_2 0x178
-#define ADF_RING_CSR_INT_COL_EN 0x17C
-#define ADF_RING_CSR_INT_COL_CTL 0x180
-#define ADF_RING_CSR_INT_FLAG_AND_COL 0x184
-#define ADF_RING_CSR_INT_COL_CTL_ENABLE	0x80000000
-#define ADF_RING_BUNDLE_SIZE 0x1000
 #define ADF_RING_CONFIG_NEAR_FULL_WM 0x0A
 #define ADF_RING_CONFIG_NEAR_EMPTY_WM 0x05
 #define ADF_COALESCING_MIN_TIME 0x1FF
@@ -72,54 +55,4 @@
 	((watermark_nf << ADF_RING_CONFIG_NEAR_FULL_WM)	\
 	| (watermark_ne << ADF_RING_CONFIG_NEAR_EMPTY_WM) \
 	| size)
-#define BUILD_RING_BASE_ADDR(addr, size) \
-	((addr >> 6) & (0xFFFFFFFFFFFFFFFFULL << size))
-#define READ_CSR_RING_HEAD(csr_base_addr, bank, ring) \
-	ADF_CSR_RD(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-			ADF_RING_CSR_RING_HEAD + (ring << 2))
-#define READ_CSR_RING_TAIL(csr_base_addr, bank, ring) \
-	ADF_CSR_RD(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-			ADF_RING_CSR_RING_TAIL + (ring << 2))
-#define READ_CSR_E_STAT(csr_base_addr, bank) \
-	ADF_CSR_RD(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-			ADF_RING_CSR_E_STAT)
-#define WRITE_CSR_RING_CONFIG(csr_base_addr, bank, ring, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-		ADF_RING_CSR_RING_CONFIG + (ring << 2), value)
-#define WRITE_CSR_RING_BASE(csr_base_addr, bank, ring, value) \
-do { \
-	u32 l_base = 0, u_base = 0; \
-	l_base = (u32)(value & 0xFFFFFFFF); \
-	u_base = (u32)((value & 0xFFFFFFFF00000000ULL) >> 32); \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-		ADF_RING_CSR_RING_LBASE + (ring << 2), l_base);	\
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-		ADF_RING_CSR_RING_UBASE + (ring << 2), u_base);	\
-} while (0)
-#define WRITE_CSR_RING_HEAD(csr_base_addr, bank, ring, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-		ADF_RING_CSR_RING_HEAD + (ring << 2), value)
-#define WRITE_CSR_RING_TAIL(csr_base_addr, bank, ring, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-		ADF_RING_CSR_RING_TAIL + (ring << 2), value)
-#define WRITE_CSR_INT_FLAG(csr_base_addr, bank, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * (bank)) + \
-			ADF_RING_CSR_INT_FLAG, value)
-#define WRITE_CSR_INT_SRCSEL(csr_base_addr, bank) \
-do { \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-	ADF_RING_CSR_INT_SRCSEL, ADF_BANK_INT_SRC_SEL_MASK_0);	\
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-	ADF_RING_CSR_INT_SRCSEL_2, ADF_BANK_INT_SRC_SEL_MASK_X); \
-} while (0)
-#define WRITE_CSR_INT_COL_EN(csr_base_addr, bank, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-			ADF_RING_CSR_INT_COL_EN, value)
-#define WRITE_CSR_INT_COL_CTL(csr_base_addr, bank, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-			ADF_RING_CSR_INT_COL_CTL, \
-			ADF_RING_CSR_INT_COL_CTL_ENABLE | value)
-#define WRITE_CSR_INT_FLAG_AND_COL(csr_base_addr, bank, value) \
-	ADF_CSR_WR(csr_base_addr, (ADF_RING_BUNDLE_SIZE * bank) + \
-			ADF_RING_CSR_INT_FLAG_AND_COL, value)
 #endif
diff --git a/drivers/crypto/qat/qat_common/adf_transport_debug.c b/drivers/crypto/qat/qat_common/adf_transport_debug.c
index dac25ba47260..1205186ad51e 100644
--- a/drivers/crypto/qat/qat_common/adf_transport_debug.c
+++ b/drivers/crypto/qat/qat_common/adf_transport_debug.c
@@ -42,16 +42,17 @@ static int adf_ring_show(struct seq_file *sfile, void *v)
 {
 	struct adf_etr_ring_data *ring = sfile->private;
 	struct adf_etr_bank_data *bank = ring->bank;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
 	void __iomem *csr = ring->bank->csr_addr;
 
 	if (v == SEQ_START_TOKEN) {
 		int head, tail, empty;
 
-		head = READ_CSR_RING_HEAD(csr, bank->bank_number,
-					  ring->ring_number);
-		tail = READ_CSR_RING_TAIL(csr, bank->bank_number,
-					  ring->ring_number);
-		empty = READ_CSR_E_STAT(csr, bank->bank_number);
+		head = csr_ops->read_csr_ring_head(csr, bank->bank_number,
+						   ring->ring_number);
+		tail = csr_ops->read_csr_ring_tail(csr, bank->bank_number,
+						   ring->ring_number);
+		empty = csr_ops->read_csr_e_stat(csr, bank->bank_number);
 
 		seq_puts(sfile, "------- Ring configuration -------\n");
 		seq_printf(sfile, "ring name: %s\n",
@@ -117,11 +118,14 @@ void adf_ring_debugfs_rm(struct adf_etr_ring_data *ring)
 
 static void *adf_bank_start(struct seq_file *sfile, loff_t *pos)
 {
+	struct adf_etr_bank_data *bank = sfile->private;
+	u8 num_rings_per_bank = GET_NUM_RINGS_PER_BANK(bank->accel_dev);
+
 	mutex_lock(&bank_read_lock);
 	if (*pos == 0)
 		return SEQ_START_TOKEN;
 
-	if (*pos >= ADF_ETR_MAX_RINGS_PER_BANK)
+	if (*pos >= num_rings_per_bank)
 		return NULL;
 
 	return pos;
@@ -129,7 +133,10 @@ static void *adf_bank_start(struct seq_file *sfile, loff_t *pos)
 
 static void *adf_bank_next(struct seq_file *sfile, void *v, loff_t *pos)
 {
-	if (++(*pos) >= ADF_ETR_MAX_RINGS_PER_BANK)
+	struct adf_etr_bank_data *bank = sfile->private;
+	u8 num_rings_per_bank = GET_NUM_RINGS_PER_BANK(bank->accel_dev);
+
+	if (++(*pos) >= num_rings_per_bank)
 		return NULL;
 
 	return pos;
@@ -138,6 +145,7 @@ static void *adf_bank_next(struct seq_file *sfile, void *v, loff_t *pos)
 static int adf_bank_show(struct seq_file *sfile, void *v)
 {
 	struct adf_etr_bank_data *bank = sfile->private;
+	struct adf_hw_csr_ops *csr_ops = GET_CSR_OPS(bank->accel_dev);
 
 	if (v == SEQ_START_TOKEN) {
 		seq_printf(sfile, "------- Bank %d configuration -------\n",
@@ -151,11 +159,11 @@ static int adf_bank_show(struct seq_file *sfile, void *v)
 		if (!(bank->ring_mask & 1 << ring_id))
 			return 0;
 
-		head = READ_CSR_RING_HEAD(csr, bank->bank_number,
-					  ring->ring_number);
-		tail = READ_CSR_RING_TAIL(csr, bank->bank_number,
-					  ring->ring_number);
-		empty = READ_CSR_E_STAT(csr, bank->bank_number);
+		head = csr_ops->read_csr_ring_head(csr, bank->bank_number,
+						   ring->ring_number);
+		tail = csr_ops->read_csr_ring_tail(csr, bank->bank_number,
+						   ring->ring_number);
+		empty = csr_ops->read_csr_e_stat(csr, bank->bank_number);
 
 		seq_printf(sfile,
 			   "ring num %02d, head %04x, tail %04x, empty: %d\n",
diff --git a/drivers/crypto/qat/qat_common/adf_transport_internal.h b/drivers/crypto/qat/qat_common/adf_transport_internal.h
index c7faf4e2d302..501bcf0f1809 100644
--- a/drivers/crypto/qat/qat_common/adf_transport_internal.h
+++ b/drivers/crypto/qat/qat_common/adf_transport_internal.h
@@ -28,7 +28,7 @@ struct adf_etr_ring_data {
 };
 
 struct adf_etr_bank_data {
-	struct adf_etr_ring_data rings[ADF_ETR_MAX_RINGS_PER_BANK];
+	struct adf_etr_ring_data *rings;
 	struct tasklet_struct resp_handler;
 	void __iomem *csr_addr;
 	u32 irq_coalesc_timer;
diff --git a/drivers/crypto/qat/qat_common/adf_vf_isr.c b/drivers/crypto/qat/qat_common/adf_vf_isr.c
index c4a44dc6af3e..38d316a42ba6 100644
--- a/drivers/crypto/qat/qat_common/adf_vf_isr.c
+++ b/drivers/crypto/qat/qat_common/adf_vf_isr.c
@@ -156,6 +156,7 @@ static irqreturn_t adf_isr(int irq, void *privdata)
 {
 	struct adf_accel_dev *accel_dev = privdata;
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	struct adf_hw_csr_ops *csr_ops = &hw_data->csr_ops;
 	struct adf_bar *pmisc =
 			&GET_BARS(accel_dev)[hw_data->get_misc_bar_id(hw_data)];
 	void __iomem *pmisc_bar_addr = pmisc->virt_addr;
@@ -180,8 +181,8 @@ static irqreturn_t adf_isr(int irq, void *privdata)
 		struct adf_etr_bank_data *bank = &etr_data->banks[0];
 
 		/* Disable Flag and Coalesce Ring Interrupts */
-		WRITE_CSR_INT_FLAG_AND_COL(bank->csr_addr, bank->bank_number,
-					   0);
+		csr_ops->write_csr_int_flag_and_col(bank->csr_addr,
+						    bank->bank_number, 0);
 		tasklet_hi_schedule(&bank->resp_handler);
 		return IRQ_HANDLED;
 	}
diff --git a/drivers/crypto/qat/qat_common/icp_qat_fw_init_admin.h b/drivers/crypto/qat/qat_common/icp_qat_fw_init_admin.h
index d4d188cd7ed0..f05ad17fbdd6 100644
--- a/drivers/crypto/qat/qat_common/icp_qat_fw_init_admin.h
+++ b/drivers/crypto/qat/qat_common/icp_qat_fw_init_admin.h
@@ -6,7 +6,7 @@
 #include "icp_qat_fw.h"
 
 enum icp_qat_fw_init_admin_cmd_id {
-	ICP_QAT_FW_INIT_ME = 0,
+	ICP_QAT_FW_INIT_AE = 0,
 	ICP_QAT_FW_TRNG_ENABLE = 1,
 	ICP_QAT_FW_TRNG_DISABLE = 2,
 	ICP_QAT_FW_CONSTANTS_CFG = 3,
@@ -39,7 +39,7 @@ struct icp_qat_fw_init_admin_req {
 	};
 
 	__u32 resrvd4;
-};
+} __packed;
 
 struct icp_qat_fw_init_admin_resp {
 	__u8 flags;
@@ -92,7 +92,7 @@ struct icp_qat_fw_init_admin_resp {
 			__u64 resrvd8;
 		};
 	};
-};
+} __packed;
 
 #define ICP_QAT_FW_COMN_HEARTBEAT_OK 0
 #define ICP_QAT_FW_COMN_HEARTBEAT_BLOCKED 1
diff --git a/drivers/crypto/qat/qat_common/icp_qat_fw_la.h b/drivers/crypto/qat/qat_common/icp_qat_fw_la.h
index 6757ec09d81f..28fa17f14be4 100644
--- a/drivers/crypto/qat/qat_common/icp_qat_fw_la.h
+++ b/drivers/crypto/qat/qat_common/icp_qat_fw_la.h
@@ -33,6 +33,9 @@ struct icp_qat_fw_la_bulk_req {
 	struct icp_qat_fw_comn_req_cd_ctrl cd_ctrl;
 };
 
+#define ICP_QAT_FW_LA_USE_UCS_SLICE_TYPE 1
+#define QAT_LA_SLICE_TYPE_BITPOS 14
+#define QAT_LA_SLICE_TYPE_MASK 0x3
 #define ICP_QAT_FW_LA_GCM_IV_LEN_12_OCTETS 1
 #define ICP_QAT_FW_LA_GCM_IV_LEN_NOT_12_OCTETS 0
 #define QAT_FW_LA_ZUC_3G_PROTO_FLAG_BITPOS 12
@@ -179,6 +182,10 @@ struct icp_qat_fw_la_bulk_req {
 	QAT_FIELD_SET(flags, val, QAT_LA_PARTIAL_BITPOS, \
 	QAT_LA_PARTIAL_MASK)
 
+#define ICP_QAT_FW_LA_SLICE_TYPE_SET(flags, val) \
+	QAT_FIELD_SET(flags, val, QAT_LA_SLICE_TYPE_BITPOS, \
+	QAT_LA_SLICE_TYPE_MASK)
+
 struct icp_qat_fw_cipher_req_hdr_cd_pars {
 	union {
 		struct {
diff --git a/drivers/crypto/qat/qat_common/icp_qat_fw_loader_handle.h b/drivers/crypto/qat/qat_common/icp_qat_fw_loader_handle.h
index 3e8e291cd122..b8f3463be6ef 100644
--- a/drivers/crypto/qat/qat_common/icp_qat_fw_loader_handle.h
+++ b/drivers/crypto/qat/qat_common/icp_qat_fw_loader_handle.h
@@ -15,6 +15,7 @@ struct icp_qat_fw_loader_ae_data {
 struct icp_qat_fw_loader_hal_handle {
 	struct icp_qat_fw_loader_ae_data aes[ICP_QAT_UCLO_MAX_AE];
 	unsigned int ae_mask;
+	unsigned int admin_ae_mask;
 	unsigned int slice_mask;
 	unsigned int revision_id;
 	unsigned int ae_max_num;
@@ -22,12 +23,35 @@ struct icp_qat_fw_loader_hal_handle {
 	unsigned int max_ustore;
 };
 
+struct icp_qat_fw_loader_chip_info {
+	bool sram_visible;
+	bool nn;
+	bool lm2lm3;
+	u32 lm_size;
+	u32 icp_rst_csr;
+	u32 icp_rst_mask;
+	u32 glb_clk_enable_csr;
+	u32 misc_ctl_csr;
+	u32 wakeup_event_val;
+	bool fw_auth;
+	bool css_3k;
+	bool tgroup_share_ustore;
+	u32 fcu_ctl_csr;
+	u32 fcu_sts_csr;
+	u32 fcu_dram_addr_hi;
+	u32 fcu_dram_addr_lo;
+	u32 fcu_loaded_ae_csr;
+	u8 fcu_loaded_ae_pos;
+};
+
 struct icp_qat_fw_loader_handle {
 	struct icp_qat_fw_loader_hal_handle *hal_handle;
+	struct icp_qat_fw_loader_chip_info *chip_info;
 	struct pci_dev *pci_dev;
 	void *obj_handle;
 	void *sobj_handle;
-	bool fw_auth;
+	void *mobj_handle;
+	unsigned int cfg_ae_mask;
 	void __iomem *hal_sram_addr_v;
 	void __iomem *hal_cap_g_ctl_csr_addr_v;
 	void __iomem *hal_cap_ae_xfer_csr_addr_v;
diff --git a/drivers/crypto/qat/qat_common/icp_qat_hal.h b/drivers/crypto/qat/qat_common/icp_qat_hal.h
index c0e9fc0c93dd..20b2ee1fc65a 100644
--- a/drivers/crypto/qat/qat_common/icp_qat_hal.h
+++ b/drivers/crypto/qat/qat_common/icp_qat_hal.h
@@ -5,9 +5,17 @@
 #include "icp_qat_fw_loader_handle.h"
 
 enum hal_global_csr {
-	MISC_CONTROL = 0x04,
-	ICP_RESET = 0x0c,
-	ICP_GLOBAL_CLK_ENABLE = 0x50
+	MISC_CONTROL = 0xA04,
+	ICP_RESET = 0xA0c,
+	ICP_GLOBAL_CLK_ENABLE = 0xA50
+};
+
+enum {
+	MISC_CONTROL_C4XXX = 0xAA0,
+	ICP_RESET_CPP0 = 0x938,
+	ICP_RESET_CPP1 = 0x93c,
+	ICP_GLOBAL_CLK_ENABLE_CPP0 = 0x964,
+	ICP_GLOBAL_CLK_ENABLE_CPP1 = 0x968
 };
 
 enum hal_ae_csr {
@@ -26,8 +34,14 @@ enum hal_ae_csr {
 	CTX_WAKEUP_EVENTS_INDIRECT = 0x050,
 	LM_ADDR_0_INDIRECT = 0x060,
 	LM_ADDR_1_INDIRECT = 0x068,
+	LM_ADDR_2_INDIRECT = 0x0cc,
+	LM_ADDR_3_INDIRECT = 0x0d4,
 	INDIRECT_LM_ADDR_0_BYTE_INDEX = 0x0e0,
 	INDIRECT_LM_ADDR_1_BYTE_INDEX = 0x0e8,
+	INDIRECT_LM_ADDR_2_BYTE_INDEX = 0x10c,
+	INDIRECT_LM_ADDR_3_BYTE_INDEX = 0x114,
+	INDIRECT_T_INDEX = 0x0f8,
+	INDIRECT_T_INDEX_BYTE_INDEX = 0x0fc,
 	FUTURE_COUNT_SIGNAL_INDIRECT = 0x078,
 	TIMESTAMP_LOW = 0x0c0,
 	TIMESTAMP_HIGH = 0x0c4,
@@ -47,6 +61,15 @@ enum fcu_csr {
 	FCU_RAMBASE_ADDR_LO   = 0x8d8
 };
 
+enum fcu_csr_4xxx {
+	FCU_CONTROL_4XXX           = 0x1000,
+	FCU_STATUS_4XXX            = 0x1004,
+	FCU_ME_BROADCAST_MASK_TYPE = 0x1008,
+	FCU_AE_LOADED_4XXX         = 0x1010,
+	FCU_DRAM_ADDR_LO_4XXX      = 0x1014,
+	FCU_DRAM_ADDR_HI_4XXX      = 0x1018,
+};
+
 enum fcu_cmd {
 	FCU_CTRL_CMD_NOOP  = 0,
 	FCU_CTRL_CMD_AUTH  = 1,
@@ -62,12 +85,17 @@ enum fcu_sts {
 	FCU_STS_LOAD_FAIL = 4,
 	FCU_STS_BUSY      = 5
 };
+
+#define ALL_AE_MASK                 0xFFFFFFFF
 #define UA_ECS                      (0x1 << 31)
 #define ACS_ABO_BITPOS              31
 #define ACS_ACNO                    0x7
 #define CE_ENABLE_BITPOS            0x8
 #define CE_LMADDR_0_GLOBAL_BITPOS   16
 #define CE_LMADDR_1_GLOBAL_BITPOS   17
+#define CE_LMADDR_2_GLOBAL_BITPOS   22
+#define CE_LMADDR_3_GLOBAL_BITPOS   23
+#define CE_T_INDEX_GLOBAL_BITPOS    21
 #define CE_NN_MODE_BITPOS           20
 #define CE_REG_PAR_ERR_BITPOS       25
 #define CE_BREAKPOINT_BITPOS        27
@@ -78,7 +106,8 @@ enum fcu_sts {
 #define XCWE_VOLUNTARY              (0x1)
 #define LCS_STATUS          (0x1)
 #define MMC_SHARE_CS_BITPOS         2
-#define GLOBAL_CSR                0xA00
+#define WAKEUP_EVENT 0x10000
+#define FCU_CTRL_BROADCAST_POS   0x4
 #define FCU_CTRL_AE_POS     0x8
 #define FCU_AUTH_STS_MASK   0x7
 #define FCU_STS_DONE_POS    0x9
@@ -86,27 +115,29 @@ enum fcu_sts {
 #define FCU_LOADED_AE_POS   0x16
 #define FW_AUTH_WAIT_PERIOD 10
 #define FW_AUTH_MAX_RETRY   300
-
+#define ICP_QAT_AE_OFFSET 0x20000
+#define ICP_QAT_CAP_OFFSET (ICP_QAT_AE_OFFSET + 0x10000)
+#define LOCAL_TO_XFER_REG_OFFSET 0x800
+#define ICP_QAT_EP_OFFSET 0x3a000
+#define ICP_QAT_EP_OFFSET_4XXX   0x200000 /* HI MMIO CSRs */
+#define ICP_QAT_AE_OFFSET_4XXX   0x600000
+#define ICP_QAT_CAP_OFFSET_4XXX  0x640000
 #define SET_CAP_CSR(handle, csr, val) \
-	ADF_CSR_WR(handle->hal_cap_g_ctl_csr_addr_v, csr, val)
+	ADF_CSR_WR((handle)->hal_cap_g_ctl_csr_addr_v, csr, val)
 #define GET_CAP_CSR(handle, csr) \
-	ADF_CSR_RD(handle->hal_cap_g_ctl_csr_addr_v, csr)
-#define SET_GLB_CSR(handle, csr, val) SET_CAP_CSR(handle, csr + GLOBAL_CSR, val)
-#define GET_GLB_CSR(handle, csr) GET_CAP_CSR(handle, GLOBAL_CSR + csr)
+	ADF_CSR_RD((handle)->hal_cap_g_ctl_csr_addr_v, csr)
 #define AE_CSR(handle, ae) \
-	((char __iomem *)handle->hal_cap_ae_local_csr_addr_v + \
-	((ae & handle->hal_handle->ae_mask) << 12))
-#define AE_CSR_ADDR(handle, ae, csr) (AE_CSR(handle, ae) + (0x3ff & csr))
+	((char __iomem *)(handle)->hal_cap_ae_local_csr_addr_v + ((ae) << 12))
+#define AE_CSR_ADDR(handle, ae, csr) (AE_CSR(handle, ae) + (0x3ff & (csr)))
 #define SET_AE_CSR(handle, ae, csr, val) \
 	ADF_CSR_WR(AE_CSR_ADDR(handle, ae, csr), 0, val)
 #define GET_AE_CSR(handle, ae, csr) ADF_CSR_RD(AE_CSR_ADDR(handle, ae, csr), 0)
 #define AE_XFER(handle, ae) \
-	((char __iomem *)handle->hal_cap_ae_xfer_csr_addr_v + \
-	((ae & handle->hal_handle->ae_mask) << 12))
+	((char __iomem *)(handle)->hal_cap_ae_xfer_csr_addr_v + ((ae) << 12))
 #define AE_XFER_ADDR(handle, ae, reg) (AE_XFER(handle, ae) + \
-	((reg & 0xff) << 2))
+	(((reg) & 0xff) << 2))
 #define SET_AE_XFER(handle, ae, reg, val) \
 	ADF_CSR_WR(AE_XFER_ADDR(handle, ae, reg), 0, val)
 #define SRAM_WRITE(handle, addr, val) \
-	ADF_CSR_WR(handle->hal_sram_addr_v, addr, val)
+	ADF_CSR_WR((handle)->hal_sram_addr_v, addr, val)
 #endif
diff --git a/drivers/crypto/qat/qat_common/icp_qat_hw.h b/drivers/crypto/qat/qat_common/icp_qat_hw.h
index c4b6ef1506ab..e39e8a2d51a7 100644
--- a/drivers/crypto/qat/qat_common/icp_qat_hw.h
+++ b/drivers/crypto/qat/qat_common/icp_qat_hw.h
@@ -65,6 +65,36 @@ struct icp_qat_hw_auth_config {
 	__u32 reserved;
 };
 
+struct icp_qat_hw_ucs_cipher_config {
+	__u32 val;
+	__u32 reserved[3];
+};
+
+enum icp_qat_slice_mask {
+	ICP_ACCEL_MASK_CIPHER_SLICE = BIT(0),
+	ICP_ACCEL_MASK_AUTH_SLICE = BIT(1),
+	ICP_ACCEL_MASK_PKE_SLICE = BIT(2),
+	ICP_ACCEL_MASK_COMPRESS_SLICE = BIT(3),
+	ICP_ACCEL_MASK_LZS_SLICE = BIT(4),
+	ICP_ACCEL_MASK_EIA3_SLICE = BIT(5),
+	ICP_ACCEL_MASK_SHA3_SLICE = BIT(6),
+};
+
+enum icp_qat_capabilities_mask {
+	ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC = BIT(0),
+	ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC = BIT(1),
+	ICP_ACCEL_CAPABILITIES_CIPHER = BIT(2),
+	ICP_ACCEL_CAPABILITIES_AUTHENTICATION = BIT(3),
+	ICP_ACCEL_CAPABILITIES_RESERVED_1 = BIT(4),
+	ICP_ACCEL_CAPABILITIES_COMPRESSION = BIT(5),
+	ICP_ACCEL_CAPABILITIES_LZS_COMPRESSION = BIT(6),
+	ICP_ACCEL_CAPABILITIES_RAND = BIT(7),
+	ICP_ACCEL_CAPABILITIES_ZUC = BIT(8),
+	ICP_ACCEL_CAPABILITIES_SHA3 = BIT(9),
+	/* Bits 10-25 are currently reserved */
+	ICP_ACCEL_CAPABILITIES_AES_V2 = BIT(26)
+};
+
 #define QAT_AUTH_MODE_BITPOS 4
 #define QAT_AUTH_MODE_MASK 0xF
 #define QAT_AUTH_ALGO_BITPOS 0
@@ -255,7 +285,15 @@ struct icp_qat_hw_cipher_aes256_f8 {
 	__u8 key[ICP_QAT_HW_AES_256_F8_KEY_SZ];
 };
 
+struct icp_qat_hw_ucs_cipher_aes256_f8 {
+	struct icp_qat_hw_ucs_cipher_config cipher_config;
+	__u8 key[ICP_QAT_HW_AES_256_F8_KEY_SZ];
+};
+
 struct icp_qat_hw_cipher_algo_blk {
-	struct icp_qat_hw_cipher_aes256_f8 aes;
+	union {
+		struct icp_qat_hw_cipher_aes256_f8 aes;
+		struct icp_qat_hw_ucs_cipher_aes256_f8 ucs_aes;
+	};
 } __aligned(64);
 #endif
diff --git a/drivers/crypto/qat/qat_common/icp_qat_uclo.h b/drivers/crypto/qat/qat_common/icp_qat_uclo.h
index 8fe1ec344fa2..4b36869bf460 100644
--- a/drivers/crypto/qat/qat_common/icp_qat_uclo.h
+++ b/drivers/crypto/qat/qat_common/icp_qat_uclo.h
@@ -6,6 +6,7 @@
 #define ICP_QAT_AC_895XCC_DEV_TYPE 0x00400000
 #define ICP_QAT_AC_C62X_DEV_TYPE   0x01000000
 #define ICP_QAT_AC_C3XXX_DEV_TYPE  0x02000000
+#define ICP_QAT_AC_4XXX_A_DEV_TYPE 0x08000000
 #define ICP_QAT_UCLO_MAX_AE       12
 #define ICP_QAT_UCLO_MAX_CTX      8
 #define ICP_QAT_UCLO_MAX_UIMAGE   (ICP_QAT_UCLO_MAX_AE * ICP_QAT_UCLO_MAX_CTX)
@@ -13,6 +14,7 @@
 #define ICP_QAT_UCLO_MAX_XFER_REG 128
 #define ICP_QAT_UCLO_MAX_GPR_REG  128
 #define ICP_QAT_UCLO_MAX_LMEM_REG 1024
+#define ICP_QAT_UCLO_MAX_LMEM_REG_2X 1280
 #define ICP_QAT_UCLO_AE_ALL_CTX   0xff
 #define ICP_QAT_UOF_OBJID_LEN     8
 #define ICP_QAT_UOF_FID 0xc6c2
@@ -31,26 +33,59 @@
 #define ICP_QAT_SUOF_FID  0x53554f46
 #define ICP_QAT_SUOF_MAJVER 0x0
 #define ICP_QAT_SUOF_MINVER 0x1
+#define ICP_QAT_SUOF_OBJ_NAME_LEN 128
+#define ICP_QAT_MOF_OBJ_ID_LEN 8
+#define ICP_QAT_MOF_OBJ_CHUNKID_LEN 8
+#define ICP_QAT_MOF_FID 0x00666f6d
+#define ICP_QAT_MOF_MAJVER 0x0
+#define ICP_QAT_MOF_MINVER 0x1
+#define ICP_QAT_MOF_SYM_OBJS "SYM_OBJS"
+#define ICP_QAT_SUOF_OBJS "SUF_OBJS"
+#define ICP_QAT_SUOF_IMAG "SUF_IMAG"
 #define ICP_QAT_SIMG_AE_INIT_SEQ_LEN    (50 * sizeof(unsigned long long))
 #define ICP_QAT_SIMG_AE_INSTS_LEN       (0x4000 * sizeof(unsigned long long))
-#define ICP_QAT_CSS_FWSK_MODULUS_LEN    256
-#define ICP_QAT_CSS_FWSK_EXPONENT_LEN   4
-#define ICP_QAT_CSS_FWSK_PAD_LEN        252
-#define ICP_QAT_CSS_FWSK_PUB_LEN   (ICP_QAT_CSS_FWSK_MODULUS_LEN + \
-				    ICP_QAT_CSS_FWSK_EXPONENT_LEN + \
-				    ICP_QAT_CSS_FWSK_PAD_LEN)
-#define ICP_QAT_CSS_SIGNATURE_LEN   256
+
+#define DSS_FWSK_MODULUS_LEN    384 /* RSA3K */
+#define DSS_FWSK_EXPONENT_LEN   4
+#define DSS_FWSK_PADDING_LEN    380
+#define DSS_SIGNATURE_LEN       384 /* RSA3K */
+
+#define CSS_FWSK_MODULUS_LEN    256 /* RSA2K */
+#define CSS_FWSK_EXPONENT_LEN   4
+#define CSS_FWSK_PADDING_LEN    252
+#define CSS_SIGNATURE_LEN       256 /* RSA2K */
+
+#define ICP_QAT_CSS_FWSK_MODULUS_LEN(handle)	((handle)->chip_info->css_3k ? \
+						DSS_FWSK_MODULUS_LEN  : \
+						CSS_FWSK_MODULUS_LEN)
+
+#define ICP_QAT_CSS_FWSK_EXPONENT_LEN(handle)	((handle)->chip_info->css_3k ? \
+						DSS_FWSK_EXPONENT_LEN : \
+						CSS_FWSK_EXPONENT_LEN)
+
+#define ICP_QAT_CSS_FWSK_PAD_LEN(handle)	((handle)->chip_info->css_3k ? \
+						DSS_FWSK_PADDING_LEN : \
+						CSS_FWSK_PADDING_LEN)
+
+#define ICP_QAT_CSS_FWSK_PUB_LEN(handle)	(ICP_QAT_CSS_FWSK_MODULUS_LEN(handle) + \
+						ICP_QAT_CSS_FWSK_EXPONENT_LEN(handle) + \
+						ICP_QAT_CSS_FWSK_PAD_LEN(handle))
+
+#define ICP_QAT_CSS_SIGNATURE_LEN(handle)	((handle)->chip_info->css_3k ? \
+						DSS_SIGNATURE_LEN : \
+						CSS_SIGNATURE_LEN)
+
 #define ICP_QAT_CSS_AE_IMG_LEN     (sizeof(struct icp_qat_simg_ae_mode) + \
 				    ICP_QAT_SIMG_AE_INIT_SEQ_LEN +         \
 				    ICP_QAT_SIMG_AE_INSTS_LEN)
-#define ICP_QAT_CSS_AE_SIMG_LEN    (sizeof(struct icp_qat_css_hdr) + \
-				    ICP_QAT_CSS_FWSK_PUB_LEN + \
-				    ICP_QAT_CSS_SIGNATURE_LEN + \
-				    ICP_QAT_CSS_AE_IMG_LEN)
-#define ICP_QAT_AE_IMG_OFFSET	   (sizeof(struct icp_qat_css_hdr) + \
-				    ICP_QAT_CSS_FWSK_MODULUS_LEN + \
-				    ICP_QAT_CSS_FWSK_EXPONENT_LEN + \
-				    ICP_QAT_CSS_SIGNATURE_LEN)
+#define ICP_QAT_CSS_AE_SIMG_LEN(handle) (sizeof(struct icp_qat_css_hdr) + \
+					ICP_QAT_CSS_FWSK_PUB_LEN(handle) + \
+					ICP_QAT_CSS_SIGNATURE_LEN(handle) + \
+					ICP_QAT_CSS_AE_IMG_LEN)
+#define ICP_QAT_AE_IMG_OFFSET(handle) (sizeof(struct icp_qat_css_hdr) + \
+					ICP_QAT_CSS_FWSK_MODULUS_LEN(handle) + \
+					ICP_QAT_CSS_FWSK_EXPONENT_LEN(handle) + \
+					ICP_QAT_CSS_SIGNATURE_LEN(handle))
 #define ICP_QAT_CSS_MAX_IMAGE_LEN   0x40000
 
 #define ICP_QAT_CTX_MODE(ae_mode) ((ae_mode) & 0xf)
@@ -60,6 +95,9 @@
 
 #define ICP_QAT_LOC_MEM0_MODE(ae_mode) (((ae_mode) >> 0x8) & 0x1)
 #define ICP_QAT_LOC_MEM1_MODE(ae_mode) (((ae_mode) >> 0x9) & 0x1)
+#define ICP_QAT_LOC_MEM2_MODE(ae_mode) (((ae_mode) >> 0x6) & 0x1)
+#define ICP_QAT_LOC_MEM3_MODE(ae_mode) (((ae_mode) >> 0x7) & 0x1)
+#define ICP_QAT_LOC_TINDEX_MODE(ae_mode) (((ae_mode) >> 0xe) & 0x1)
 
 enum icp_qat_uof_mem_region {
 	ICP_QAT_UOF_SRAM_REGION = 0x0,
@@ -89,6 +127,8 @@ enum icp_qat_uof_regtype {
 	ICP_LMEM0	= 27,
 	ICP_LMEM1	= 28,
 	ICP_NEIGH_REL	= 31,
+	ICP_LMEM2	= 61,
+	ICP_LMEM3	= 62,
 };
 
 enum icp_qat_css_fwtype {
@@ -394,7 +434,7 @@ struct icp_qat_suof_handle {
 
 struct icp_qat_fw_auth_desc {
 	unsigned int   img_len;
-	unsigned int   reserved;
+	unsigned int   ae_mask;
 	unsigned int   css_hdr_high;
 	unsigned int   css_hdr_low;
 	unsigned int   img_high;
@@ -481,4 +521,64 @@ struct icp_qat_suof_objhdr {
 	unsigned int img_length;
 	unsigned int reserved;
 };
+
+struct icp_qat_mof_file_hdr {
+	unsigned int file_id;
+	unsigned int checksum;
+	char min_ver;
+	char maj_ver;
+	unsigned short reserved;
+	unsigned short max_chunks;
+	unsigned short num_chunks;
+};
+
+struct icp_qat_mof_chunkhdr {
+	char chunk_id[ICP_QAT_MOF_OBJ_ID_LEN];
+	u64 offset;
+	u64 size;
+};
+
+struct icp_qat_mof_str_table {
+	unsigned int tab_len;
+	unsigned int strings;
+};
+
+struct icp_qat_mof_obj_hdr {
+	unsigned short max_chunks;
+	unsigned short num_chunks;
+	unsigned int reserved;
+};
+
+struct icp_qat_mof_obj_chunkhdr {
+	char chunk_id[ICP_QAT_MOF_OBJ_CHUNKID_LEN];
+	u64 offset;
+	u64 size;
+	unsigned int name;
+	unsigned int reserved;
+};
+
+struct icp_qat_mof_objhdr {
+	char *obj_name;
+	char *obj_buf;
+	unsigned int obj_size;
+};
+
+struct icp_qat_mof_table {
+	unsigned int num_objs;
+	struct icp_qat_mof_objhdr *obj_hdr;
+};
+
+struct icp_qat_mof_handle {
+	unsigned int file_id;
+	unsigned int checksum;
+	char min_ver;
+	char maj_ver;
+	char *mof_buf;
+	u32 mof_size;
+	char *sym_str;
+	unsigned int sym_size;
+	char *uobjs_hdr;
+	char *sobjs_hdr;
+	struct icp_qat_mof_table obj_table;
+};
 #endif
diff --git a/drivers/crypto/qat/qat_common/qat_algs.c b/drivers/crypto/qat/qat_common/qat_algs.c
index d552dbcfe0a0..31c7a206a629 100644
--- a/drivers/crypto/qat/qat_common/qat_algs.c
+++ b/drivers/crypto/qat/qat_common/qat_algs.c
@@ -6,11 +6,13 @@
 #include <crypto/internal/aead.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/aes.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/hash.h>
 #include <crypto/hmac.h>
 #include <crypto/algapi.h>
 #include <crypto/authenc.h>
+#include <crypto/scatterwalk.h>
 #include <crypto/xts.h>
 #include <linux/dma-mapping.h>
 #include "adf_accel_devices.h"
@@ -31,6 +33,15 @@
 				       ICP_QAT_HW_CIPHER_KEY_CONVERT, \
 				       ICP_QAT_HW_CIPHER_DECRYPT)
 
+#define QAT_AES_HW_CONFIG_DEC_NO_CONV(alg, mode) \
+	ICP_QAT_HW_CIPHER_CONFIG_BUILD(mode, alg, \
+				       ICP_QAT_HW_CIPHER_NO_CONVERT, \
+				       ICP_QAT_HW_CIPHER_DECRYPT)
+
+#define HW_CAP_AES_V2(accel_dev) \
+	(GET_HW_DATA(accel_dev)->accel_capabilities_mask & \
+	 ICP_ACCEL_CAPABILITIES_AES_V2)
+
 static DEFINE_MUTEX(algs_lock);
 static unsigned int active_devs;
 
@@ -89,7 +100,9 @@ struct qat_alg_skcipher_ctx {
 	struct icp_qat_fw_la_bulk_req dec_fw_req;
 	struct qat_crypto_instance *inst;
 	struct crypto_skcipher *ftfm;
+	struct crypto_cipher *tweak;
 	bool fallback;
+	int mode;
 };
 
 static int qat_get_inter_state_size(enum icp_qat_hw_auth_algo qat_hash_alg)
@@ -103,7 +116,7 @@ static int qat_get_inter_state_size(enum icp_qat_hw_auth_algo qat_hash_alg)
 		return ICP_QAT_HW_SHA512_STATE1_SZ;
 	default:
 		return -EFAULT;
-	};
+	}
 	return -EFAULT;
 }
 
@@ -214,24 +227,7 @@ static int qat_alg_do_precomputes(struct icp_qat_hw_auth_algo_blk *hash,
 	return 0;
 }
 
-static void qat_alg_init_hdr_iv_updt(struct icp_qat_fw_comn_req_hdr *header)
-{
-	ICP_QAT_FW_LA_CIPH_IV_FLD_FLAG_SET(header->serv_specif_flags,
-					   ICP_QAT_FW_CIPH_IV_64BIT_PTR);
-	ICP_QAT_FW_LA_UPDATE_STATE_SET(header->serv_specif_flags,
-				       ICP_QAT_FW_LA_UPDATE_STATE);
-}
-
-static void qat_alg_init_hdr_no_iv_updt(struct icp_qat_fw_comn_req_hdr *header)
-{
-	ICP_QAT_FW_LA_CIPH_IV_FLD_FLAG_SET(header->serv_specif_flags,
-					   ICP_QAT_FW_CIPH_IV_16BYTE_DATA);
-	ICP_QAT_FW_LA_UPDATE_STATE_SET(header->serv_specif_flags,
-				       ICP_QAT_FW_LA_NO_UPDATE_STATE);
-}
-
-static void qat_alg_init_common_hdr(struct icp_qat_fw_comn_req_hdr *header,
-				    int aead)
+static void qat_alg_init_common_hdr(struct icp_qat_fw_comn_req_hdr *header)
 {
 	header->hdr_flags =
 		ICP_QAT_FW_COMN_HDR_FLAGS_BUILD(ICP_QAT_FW_COMN_REQ_FLAG_SET);
@@ -241,12 +237,12 @@ static void qat_alg_init_common_hdr(struct icp_qat_fw_comn_req_hdr *header,
 					    QAT_COMN_PTR_TYPE_SGL);
 	ICP_QAT_FW_LA_PARTIAL_SET(header->serv_specif_flags,
 				  ICP_QAT_FW_LA_PARTIAL_NONE);
-	if (aead)
-		qat_alg_init_hdr_no_iv_updt(header);
-	else
-		qat_alg_init_hdr_iv_updt(header);
+	ICP_QAT_FW_LA_CIPH_IV_FLD_FLAG_SET(header->serv_specif_flags,
+					   ICP_QAT_FW_CIPH_IV_16BYTE_DATA);
 	ICP_QAT_FW_LA_PROTO_SET(header->serv_specif_flags,
 				ICP_QAT_FW_LA_NO_PROTO);
+	ICP_QAT_FW_LA_UPDATE_STATE_SET(header->serv_specif_flags,
+				       ICP_QAT_FW_LA_NO_UPDATE_STATE);
 }
 
 static int qat_alg_aead_init_enc_session(struct crypto_aead *aead_tfm,
@@ -281,7 +277,7 @@ static int qat_alg_aead_init_enc_session(struct crypto_aead *aead_tfm,
 		return -EFAULT;
 
 	/* Request setup */
-	qat_alg_init_common_hdr(header, 1);
+	qat_alg_init_common_hdr(header);
 	header->service_cmd_id = ICP_QAT_FW_LA_CMD_CIPHER_HASH;
 	ICP_QAT_FW_LA_DIGEST_IN_BUFFER_SET(header->serv_specif_flags,
 					   ICP_QAT_FW_LA_DIGEST_IN_BUFFER);
@@ -368,7 +364,7 @@ static int qat_alg_aead_init_dec_session(struct crypto_aead *aead_tfm,
 		return -EFAULT;
 
 	/* Request setup */
-	qat_alg_init_common_hdr(header, 1);
+	qat_alg_init_common_hdr(header);
 	header->service_cmd_id = ICP_QAT_FW_LA_CMD_HASH_CIPHER;
 	ICP_QAT_FW_LA_DIGEST_IN_BUFFER_SET(header->serv_specif_flags,
 					   ICP_QAT_FW_LA_DIGEST_IN_BUFFER);
@@ -430,12 +426,32 @@ static void qat_alg_skcipher_init_com(struct qat_alg_skcipher_ctx *ctx,
 	struct icp_qat_fw_comn_req_hdr_cd_pars *cd_pars = &req->cd_pars;
 	struct icp_qat_fw_comn_req_hdr *header = &req->comn_hdr;
 	struct icp_qat_fw_cipher_cd_ctrl_hdr *cd_ctrl = (void *)&req->cd_ctrl;
+	bool aes_v2_capable = HW_CAP_AES_V2(ctx->inst->accel_dev);
+	int mode = ctx->mode;
 
-	memcpy(cd->aes.key, key, keylen);
-	qat_alg_init_common_hdr(header, 0);
+	qat_alg_init_common_hdr(header);
 	header->service_cmd_id = ICP_QAT_FW_LA_CMD_CIPHER;
 	cd_pars->u.s.content_desc_params_sz =
 				sizeof(struct icp_qat_hw_cipher_algo_blk) >> 3;
+
+	if (aes_v2_capable && mode == ICP_QAT_HW_CIPHER_XTS_MODE) {
+		ICP_QAT_FW_LA_SLICE_TYPE_SET(header->serv_specif_flags,
+					     ICP_QAT_FW_LA_USE_UCS_SLICE_TYPE);
+
+		/* Store both XTS keys in CD, only the first key is sent
+		 * to the HW, the second key is used for tweak calculation
+		 */
+		memcpy(cd->ucs_aes.key, key, keylen);
+		keylen = keylen / 2;
+	} else if (aes_v2_capable && mode == ICP_QAT_HW_CIPHER_CTR_MODE) {
+		ICP_QAT_FW_LA_SLICE_TYPE_SET(header->serv_specif_flags,
+					     ICP_QAT_FW_LA_USE_UCS_SLICE_TYPE);
+		keylen = round_up(keylen, 16);
+		memcpy(cd->ucs_aes.key, key, keylen);
+	} else {
+		memcpy(cd->aes.key, key, keylen);
+	}
+
 	/* Cipher CD config setup */
 	cd_ctrl->cipher_key_sz = keylen >> 3;
 	cd_ctrl->cipher_state_sz = AES_BLOCK_SIZE >> 3;
@@ -457,6 +473,28 @@ static void qat_alg_skcipher_init_enc(struct qat_alg_skcipher_ctx *ctx,
 	enc_cd->aes.cipher_config.val = QAT_AES_HW_CONFIG_ENC(alg, mode);
 }
 
+static void qat_alg_xts_reverse_key(const u8 *key_forward, unsigned int keylen,
+				    u8 *key_reverse)
+{
+	struct crypto_aes_ctx aes_expanded;
+	int nrounds;
+	u8 *key;
+
+	aes_expandkey(&aes_expanded, key_forward, keylen);
+	if (keylen == AES_KEYSIZE_128) {
+		nrounds = 10;
+		key = (u8 *)aes_expanded.key_enc + (AES_BLOCK_SIZE * nrounds);
+		memcpy(key_reverse, key, AES_BLOCK_SIZE);
+	} else {
+		/* AES_KEYSIZE_256 */
+		nrounds = 14;
+		key = (u8 *)aes_expanded.key_enc + (AES_BLOCK_SIZE * nrounds);
+		memcpy(key_reverse, key, AES_BLOCK_SIZE);
+		memcpy(key_reverse + AES_BLOCK_SIZE, key - AES_BLOCK_SIZE,
+		       AES_BLOCK_SIZE);
+	}
+}
+
 static void qat_alg_skcipher_init_dec(struct qat_alg_skcipher_ctx *ctx,
 				      int alg, const u8 *key,
 				      unsigned int keylen, int mode)
@@ -464,16 +502,26 @@ static void qat_alg_skcipher_init_dec(struct qat_alg_skcipher_ctx *ctx,
 	struct icp_qat_hw_cipher_algo_blk *dec_cd = ctx->dec_cd;
 	struct icp_qat_fw_la_bulk_req *req = &ctx->dec_fw_req;
 	struct icp_qat_fw_comn_req_hdr_cd_pars *cd_pars = &req->cd_pars;
+	bool aes_v2_capable = HW_CAP_AES_V2(ctx->inst->accel_dev);
 
 	qat_alg_skcipher_init_com(ctx, req, dec_cd, key, keylen);
 	cd_pars->u.s.content_desc_addr = ctx->dec_cd_paddr;
 
-	if (mode != ICP_QAT_HW_CIPHER_CTR_MODE)
+	if (aes_v2_capable && mode == ICP_QAT_HW_CIPHER_XTS_MODE) {
+		/* Key reversing not supported, set no convert */
+		dec_cd->aes.cipher_config.val =
+				QAT_AES_HW_CONFIG_DEC_NO_CONV(alg, mode);
+
+		/* In-place key reversal */
+		qat_alg_xts_reverse_key(dec_cd->ucs_aes.key, keylen / 2,
+					dec_cd->ucs_aes.key);
+	} else if (mode != ICP_QAT_HW_CIPHER_CTR_MODE) {
 		dec_cd->aes.cipher_config.val =
 					QAT_AES_HW_CONFIG_DEC(alg, mode);
-	else
+	} else {
 		dec_cd->aes.cipher_config.val =
 					QAT_AES_HW_CONFIG_ENC(alg, mode);
+	}
 }
 
 static int qat_alg_validate_key(int key_len, int *alg, int mode)
@@ -787,6 +835,61 @@ static void qat_aead_alg_callback(struct icp_qat_fw_la_resp *qat_resp,
 	areq->base.complete(&areq->base, res);
 }
 
+static void qat_alg_update_iv_ctr_mode(struct qat_crypto_request *qat_req)
+{
+	struct skcipher_request *sreq = qat_req->skcipher_req;
+	u64 iv_lo_prev;
+	u64 iv_lo;
+	u64 iv_hi;
+
+	memcpy(qat_req->iv, sreq->iv, AES_BLOCK_SIZE);
+
+	iv_lo = be64_to_cpu(qat_req->iv_lo);
+	iv_hi = be64_to_cpu(qat_req->iv_hi);
+
+	iv_lo_prev = iv_lo;
+	iv_lo += DIV_ROUND_UP(sreq->cryptlen, AES_BLOCK_SIZE);
+	if (iv_lo < iv_lo_prev)
+		iv_hi++;
+
+	qat_req->iv_lo = cpu_to_be64(iv_lo);
+	qat_req->iv_hi = cpu_to_be64(iv_hi);
+}
+
+static void qat_alg_update_iv_cbc_mode(struct qat_crypto_request *qat_req)
+{
+	struct skcipher_request *sreq = qat_req->skcipher_req;
+	int offset = sreq->cryptlen - AES_BLOCK_SIZE;
+	struct scatterlist *sgl;
+
+	if (qat_req->encryption)
+		sgl = sreq->dst;
+	else
+		sgl = sreq->src;
+
+	scatterwalk_map_and_copy(qat_req->iv, sgl, offset, AES_BLOCK_SIZE, 0);
+}
+
+static void qat_alg_update_iv(struct qat_crypto_request *qat_req)
+{
+	struct qat_alg_skcipher_ctx *ctx = qat_req->skcipher_ctx;
+	struct device *dev = &GET_DEV(ctx->inst->accel_dev);
+
+	switch (ctx->mode) {
+	case ICP_QAT_HW_CIPHER_CTR_MODE:
+		qat_alg_update_iv_ctr_mode(qat_req);
+		break;
+	case ICP_QAT_HW_CIPHER_CBC_MODE:
+		qat_alg_update_iv_cbc_mode(qat_req);
+		break;
+	case ICP_QAT_HW_CIPHER_XTS_MODE:
+		break;
+	default:
+		dev_warn(dev, "Unsupported IV update for cipher mode %d\n",
+			 ctx->mode);
+	}
+}
+
 static void qat_skcipher_alg_callback(struct icp_qat_fw_la_resp *qat_resp,
 				      struct qat_crypto_request *qat_req)
 {
@@ -794,16 +897,16 @@ static void qat_skcipher_alg_callback(struct icp_qat_fw_la_resp *qat_resp,
 	struct qat_crypto_instance *inst = ctx->inst;
 	struct skcipher_request *sreq = qat_req->skcipher_req;
 	u8 stat_filed = qat_resp->comn_resp.comn_status;
-	struct device *dev = &GET_DEV(ctx->inst->accel_dev);
 	int res = 0, qat_res = ICP_QAT_FW_COMN_RESP_CRYPTO_STAT_GET(stat_filed);
 
 	qat_alg_free_bufl(inst, qat_req);
 	if (unlikely(qat_res != ICP_QAT_FW_COMN_STATUS_FLAG_OK))
 		res = -EINVAL;
 
+	if (qat_req->encryption)
+		qat_alg_update_iv(qat_req);
+
 	memcpy(sreq->iv, qat_req->iv, AES_BLOCK_SIZE);
-	dma_free_coherent(dev, AES_BLOCK_SIZE, qat_req->iv,
-			  qat_req->iv_paddr);
 
 	sreq->base.complete(&sreq->base, res);
 }
@@ -981,6 +1084,8 @@ static int qat_alg_skcipher_setkey(struct crypto_skcipher *tfm,
 {
 	struct qat_alg_skcipher_ctx *ctx = crypto_skcipher_ctx(tfm);
 
+	ctx->mode = mode;
+
 	if (ctx->enc_cd)
 		return qat_alg_skcipher_rekey(ctx, key, keylen, mode);
 	else
@@ -1023,8 +1128,33 @@ static int qat_alg_skcipher_xts_setkey(struct crypto_skcipher *tfm,
 
 	ctx->fallback = false;
 
-	return qat_alg_skcipher_setkey(tfm, key, keylen,
-				       ICP_QAT_HW_CIPHER_XTS_MODE);
+	ret = qat_alg_skcipher_setkey(tfm, key, keylen,
+				      ICP_QAT_HW_CIPHER_XTS_MODE);
+	if (ret)
+		return ret;
+
+	if (HW_CAP_AES_V2(ctx->inst->accel_dev))
+		ret = crypto_cipher_setkey(ctx->tweak, key + (keylen / 2),
+					   keylen / 2);
+
+	return ret;
+}
+
+static void qat_alg_set_req_iv(struct qat_crypto_request *qat_req)
+{
+	struct icp_qat_fw_la_cipher_req_params *cipher_param;
+	struct qat_alg_skcipher_ctx *ctx = qat_req->skcipher_ctx;
+	bool aes_v2_capable = HW_CAP_AES_V2(ctx->inst->accel_dev);
+	u8 *iv = qat_req->skcipher_req->iv;
+
+	cipher_param = (void *)&qat_req->req.serv_specif_rqpars;
+
+	if (aes_v2_capable && ctx->mode == ICP_QAT_HW_CIPHER_XTS_MODE)
+		crypto_cipher_encrypt_one(ctx->tweak,
+					  (u8 *)cipher_param->u.cipher_IV_array,
+					  iv);
+	else
+		memcpy(cipher_param->u.cipher_IV_array, iv, AES_BLOCK_SIZE);
 }
 
 static int qat_alg_skcipher_encrypt(struct skcipher_request *req)
@@ -1035,23 +1165,14 @@ static int qat_alg_skcipher_encrypt(struct skcipher_request *req)
 	struct qat_crypto_request *qat_req = skcipher_request_ctx(req);
 	struct icp_qat_fw_la_cipher_req_params *cipher_param;
 	struct icp_qat_fw_la_bulk_req *msg;
-	struct device *dev = &GET_DEV(ctx->inst->accel_dev);
 	int ret, ctr = 0;
 
 	if (req->cryptlen == 0)
 		return 0;
 
-	qat_req->iv = dma_alloc_coherent(dev, AES_BLOCK_SIZE,
-					 &qat_req->iv_paddr, GFP_ATOMIC);
-	if (!qat_req->iv)
-		return -ENOMEM;
-
 	ret = qat_alg_sgl_to_bufl(ctx->inst, req->src, req->dst, qat_req);
-	if (unlikely(ret)) {
-		dma_free_coherent(dev, AES_BLOCK_SIZE, qat_req->iv,
-				  qat_req->iv_paddr);
+	if (unlikely(ret))
 		return ret;
-	}
 
 	msg = &qat_req->req;
 	*msg = ctx->enc_fw_req;
@@ -1061,19 +1182,19 @@ static int qat_alg_skcipher_encrypt(struct skcipher_request *req)
 	qat_req->req.comn_mid.opaque_data = (u64)(__force long)qat_req;
 	qat_req->req.comn_mid.src_data_addr = qat_req->buf.blp;
 	qat_req->req.comn_mid.dest_data_addr = qat_req->buf.bloutp;
+	qat_req->encryption = true;
 	cipher_param = (void *)&qat_req->req.serv_specif_rqpars;
 	cipher_param->cipher_length = req->cryptlen;
 	cipher_param->cipher_offset = 0;
-	cipher_param->u.s.cipher_IV_ptr = qat_req->iv_paddr;
-	memcpy(qat_req->iv, req->iv, AES_BLOCK_SIZE);
+
+	qat_alg_set_req_iv(qat_req);
+
 	do {
 		ret = adf_send_message(ctx->inst->sym_tx, (u32 *)msg);
 	} while (ret == -EAGAIN && ctr++ < 10);
 
 	if (ret == -EAGAIN) {
 		qat_alg_free_bufl(ctx->inst, qat_req);
-		dma_free_coherent(dev, AES_BLOCK_SIZE, qat_req->iv,
-				  qat_req->iv_paddr);
 		return -EBUSY;
 	}
 	return -EINPROGRESS;
@@ -1113,23 +1234,14 @@ static int qat_alg_skcipher_decrypt(struct skcipher_request *req)
 	struct qat_crypto_request *qat_req = skcipher_request_ctx(req);
 	struct icp_qat_fw_la_cipher_req_params *cipher_param;
 	struct icp_qat_fw_la_bulk_req *msg;
-	struct device *dev = &GET_DEV(ctx->inst->accel_dev);
 	int ret, ctr = 0;
 
 	if (req->cryptlen == 0)
 		return 0;
 
-	qat_req->iv = dma_alloc_coherent(dev, AES_BLOCK_SIZE,
-					 &qat_req->iv_paddr, GFP_ATOMIC);
-	if (!qat_req->iv)
-		return -ENOMEM;
-
 	ret = qat_alg_sgl_to_bufl(ctx->inst, req->src, req->dst, qat_req);
-	if (unlikely(ret)) {
-		dma_free_coherent(dev, AES_BLOCK_SIZE, qat_req->iv,
-				  qat_req->iv_paddr);
+	if (unlikely(ret))
 		return ret;
-	}
 
 	msg = &qat_req->req;
 	*msg = ctx->dec_fw_req;
@@ -1139,19 +1251,20 @@ static int qat_alg_skcipher_decrypt(struct skcipher_request *req)
 	qat_req->req.comn_mid.opaque_data = (u64)(__force long)qat_req;
 	qat_req->req.comn_mid.src_data_addr = qat_req->buf.blp;
 	qat_req->req.comn_mid.dest_data_addr = qat_req->buf.bloutp;
+	qat_req->encryption = false;
 	cipher_param = (void *)&qat_req->req.serv_specif_rqpars;
 	cipher_param->cipher_length = req->cryptlen;
 	cipher_param->cipher_offset = 0;
-	cipher_param->u.s.cipher_IV_ptr = qat_req->iv_paddr;
-	memcpy(qat_req->iv, req->iv, AES_BLOCK_SIZE);
+
+	qat_alg_set_req_iv(qat_req);
+	qat_alg_update_iv(qat_req);
+
 	do {
 		ret = adf_send_message(ctx->inst->sym_tx, (u32 *)msg);
 	} while (ret == -EAGAIN && ctr++ < 10);
 
 	if (ret == -EAGAIN) {
 		qat_alg_free_bufl(ctx->inst, qat_req);
-		dma_free_coherent(dev, AES_BLOCK_SIZE, qat_req->iv,
-				  qat_req->iv_paddr);
 		return -EBUSY;
 	}
 	return -EINPROGRESS;
@@ -1253,6 +1366,12 @@ static int qat_alg_skcipher_init_xts_tfm(struct crypto_skcipher *tfm)
 	if (IS_ERR(ctx->ftfm))
 		return PTR_ERR(ctx->ftfm);
 
+	ctx->tweak = crypto_alloc_cipher("aes", 0, 0);
+	if (IS_ERR(ctx->tweak)) {
+		crypto_free_skcipher(ctx->ftfm);
+		return PTR_ERR(ctx->tweak);
+	}
+
 	reqsize = max(sizeof(struct qat_crypto_request),
 		      sizeof(struct skcipher_request) +
 		      crypto_skcipher_reqsize(ctx->ftfm));
@@ -1295,6 +1414,9 @@ static void qat_alg_skcipher_exit_xts_tfm(struct crypto_skcipher *tfm)
 	if (ctx->ftfm)
 		crypto_free_skcipher(ctx->ftfm);
 
+	if (ctx->tweak)
+		crypto_free_cipher(ctx->tweak);
+
 	qat_alg_skcipher_exit_tfm(tfm);
 }
 
diff --git a/drivers/crypto/qat/qat_common/qat_asym_algs.c b/drivers/crypto/qat/qat_common/qat_asym_algs.c
index 846569ec9066..2c863d25327a 100644
--- a/drivers/crypto/qat/qat_common/qat_asym_algs.c
+++ b/drivers/crypto/qat/qat_common/qat_asym_algs.c
@@ -201,12 +201,7 @@ static unsigned long qat_dh_fn_id(unsigned int len, bool g2)
 		return g2 ? PKE_DH_G2_4096 : PKE_DH_4096;
 	default:
 		return 0;
-	};
-}
-
-static inline struct qat_dh_ctx *qat_dh_get_params(struct crypto_kpp *tfm)
-{
-	return kpp_tfm_ctx(tfm);
+	}
 }
 
 static int qat_dh_compute_value(struct kpp_request *req)
@@ -577,7 +572,7 @@ static unsigned long qat_rsa_enc_fn_id(unsigned int len)
 		return PKE_RSA_EP_4096;
 	default:
 		return 0;
-	};
+	}
 }
 
 #define PKE_RSA_DP1_512 0x1c161b3c
@@ -606,7 +601,7 @@ static unsigned long qat_rsa_dec_fn_id(unsigned int len)
 		return PKE_RSA_DP1_4096;
 	default:
 		return 0;
-	};
+	}
 }
 
 #define PKE_RSA_DP2_512 0x1c131b57
@@ -635,7 +630,7 @@ static unsigned long qat_rsa_dec_fn_id_crt(unsigned int len)
 		return PKE_RSA_DP2_4096;
 	default:
 		return 0;
-	};
+	}
 }
 
 static int qat_rsa_enc(struct akcipher_request *req)
diff --git a/drivers/crypto/qat/qat_common/qat_crypto.c b/drivers/crypto/qat/qat_common/qat_crypto.c
index ab621b7dbd20..ece6776fbd53 100644
--- a/drivers/crypto/qat/qat_common/qat_crypto.c
+++ b/drivers/crypto/qat/qat_common/qat_crypto.c
@@ -115,165 +115,217 @@ struct qat_crypto_instance *qat_crypto_get_instance_node(int node)
  */
 int qat_crypto_dev_config(struct adf_accel_dev *accel_dev)
 {
-	int cpus = num_online_cpus();
-	int banks = GET_MAX_BANKS(accel_dev);
-	int instances = min(cpus, banks);
 	char key[ADF_CFG_MAX_KEY_LEN_IN_BYTES];
-	int i;
+	int banks = GET_MAX_BANKS(accel_dev);
+	int cpus = num_online_cpus();
 	unsigned long val;
+	int instances;
+	int ret;
+	int i;
+
+	if (adf_hw_dev_has_crypto(accel_dev))
+		instances = min(cpus, banks);
+	else
+		instances = 0;
 
-	if (adf_cfg_section_add(accel_dev, ADF_KERNEL_SEC))
+	ret = adf_cfg_section_add(accel_dev, ADF_KERNEL_SEC);
+	if (ret)
 		goto err;
-	if (adf_cfg_section_add(accel_dev, "Accelerator0"))
+
+	ret = adf_cfg_section_add(accel_dev, "Accelerator0");
+	if (ret)
 		goto err;
+
 	for (i = 0; i < instances; i++) {
 		val = i;
-		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_BANK_NUM, i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_BANK_NUM, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
+			goto err;
+
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_BANK_NUM, i);
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_ETRMGR_CORE_AFFINITY,
 			 i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_SIZE, i);
 		val = 128;
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		val = 512;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_SIZE, i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		val = 0;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_TX, i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		val = 2;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_TX, i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		val = 8;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_RX, i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		val = 10;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_RX, i);
-		if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 
 		val = ADF_COALESCING_DEF_TIME;
 		snprintf(key, sizeof(key), ADF_ETRMGR_COALESCE_TIMER_FORMAT, i);
-		if (adf_cfg_add_key_value_param(accel_dev, "Accelerator0",
-						key, (void *)&val, ADF_DEC))
+		ret = adf_cfg_add_key_value_param(accel_dev, "Accelerator0",
+						  key, &val, ADF_DEC);
+		if (ret)
 			goto err;
 	}
 
 	val = i;
-	if (adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC,
-					ADF_NUM_CY, (void *)&val, ADF_DEC))
+	ret = adf_cfg_add_key_value_param(accel_dev, ADF_KERNEL_SEC, ADF_NUM_CY,
+					  &val, ADF_DEC);
+	if (ret)
 		goto err;
 
 	set_bit(ADF_STATUS_CONFIGURED, &accel_dev->status);
 	return 0;
 err:
 	dev_err(&GET_DEV(accel_dev), "Failed to start QAT accel dev\n");
-	return -EINVAL;
+	return ret;
 }
 EXPORT_SYMBOL_GPL(qat_crypto_dev_config);
 
 static int qat_crypto_create_instances(struct adf_accel_dev *accel_dev)
 {
-	int i;
-	unsigned long bank;
 	unsigned long num_inst, num_msg_sym, num_msg_asym;
-	int msg_size;
-	struct qat_crypto_instance *inst;
 	char key[ADF_CFG_MAX_KEY_LEN_IN_BYTES];
 	char val[ADF_CFG_MAX_VAL_LEN_IN_BYTES];
+	unsigned long sym_bank, asym_bank;
+	struct qat_crypto_instance *inst;
+	int msg_size;
+	int ret;
+	int i;
 
 	INIT_LIST_HEAD(&accel_dev->crypto_list);
-	if (adf_cfg_get_param_value(accel_dev, SEC, ADF_NUM_CY, val))
-		return -EFAULT;
+	ret = adf_cfg_get_param_value(accel_dev, SEC, ADF_NUM_CY, val);
+	if (ret)
+		return ret;
 
-	if (kstrtoul(val, 0, &num_inst))
-		return -EFAULT;
+	ret = kstrtoul(val, 0, &num_inst);
+	if (ret)
+		return ret;
 
 	for (i = 0; i < num_inst; i++) {
 		inst = kzalloc_node(sizeof(*inst), GFP_KERNEL,
 				    dev_to_node(&GET_DEV(accel_dev)));
-		if (!inst)
+		if (!inst) {
+			ret = -ENOMEM;
 			goto err;
+		}
 
 		list_add_tail(&inst->list, &accel_dev->crypto_list);
 		inst->id = i;
 		atomic_set(&inst->refctr, 0);
 		inst->accel_dev = accel_dev;
-		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_BANK_NUM, i);
-		if (adf_cfg_get_param_value(accel_dev, SEC, key, val))
+
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_BANK_NUM, i);
+		ret = adf_cfg_get_param_value(accel_dev, SEC, key, val);
+		if (ret)
+			goto err;
+
+		ret = kstrtoul(val, 10, &sym_bank);
+		if (ret)
 			goto err;
 
-		if (kstrtoul(val, 10, &bank))
+		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_BANK_NUM, i);
+		ret = adf_cfg_get_param_value(accel_dev, SEC, key, val);
+		if (ret)
 			goto err;
+
+		ret = kstrtoul(val, 10, &asym_bank);
+		if (ret)
+			goto err;
+
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_SIZE, i);
-		if (adf_cfg_get_param_value(accel_dev, SEC, key, val))
+		ret = adf_cfg_get_param_value(accel_dev, SEC, key, val);
+		if (ret)
 			goto err;
 
-		if (kstrtoul(val, 10, &num_msg_sym))
+		ret = kstrtoul(val, 10, &num_msg_sym);
+		if (ret)
 			goto err;
 
 		num_msg_sym = num_msg_sym >> 1;
 
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_SIZE, i);
-		if (adf_cfg_get_param_value(accel_dev, SEC, key, val))
+		ret = adf_cfg_get_param_value(accel_dev, SEC, key, val);
+		if (ret)
 			goto err;
 
-		if (kstrtoul(val, 10, &num_msg_asym))
+		ret = kstrtoul(val, 10, &num_msg_asym);
+		if (ret)
 			goto err;
 		num_msg_asym = num_msg_asym >> 1;
 
 		msg_size = ICP_QAT_FW_REQ_DEFAULT_SZ;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_TX, i);
-		if (adf_create_ring(accel_dev, SEC, bank, num_msg_sym,
-				    msg_size, key, NULL, 0, &inst->sym_tx))
+		ret = adf_create_ring(accel_dev, SEC, sym_bank, num_msg_sym,
+				      msg_size, key, NULL, 0, &inst->sym_tx);
+		if (ret)
 			goto err;
 
 		msg_size = msg_size >> 1;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_TX, i);
-		if (adf_create_ring(accel_dev, SEC, bank, num_msg_asym,
-				    msg_size, key, NULL, 0, &inst->pke_tx))
+		ret = adf_create_ring(accel_dev, SEC, asym_bank, num_msg_asym,
+				      msg_size, key, NULL, 0, &inst->pke_tx);
+		if (ret)
 			goto err;
 
 		msg_size = ICP_QAT_FW_RESP_DEFAULT_SZ;
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_SYM_RX, i);
-		if (adf_create_ring(accel_dev, SEC, bank, num_msg_sym,
-				    msg_size, key, qat_alg_callback, 0,
-				    &inst->sym_rx))
+		ret = adf_create_ring(accel_dev, SEC, sym_bank, num_msg_sym,
+				      msg_size, key, qat_alg_callback, 0,
+				      &inst->sym_rx);
+		if (ret)
 			goto err;
 
 		snprintf(key, sizeof(key), ADF_CY "%d" ADF_RING_ASYM_RX, i);
-		if (adf_create_ring(accel_dev, SEC, bank, num_msg_asym,
-				    msg_size, key, qat_alg_asym_callback, 0,
-				    &inst->pke_rx))
+		ret = adf_create_ring(accel_dev, SEC, asym_bank, num_msg_asym,
+				      msg_size, key, qat_alg_asym_callback, 0,
+				      &inst->pke_rx);
+		if (ret)
 			goto err;
 	}
 	return 0;
 err:
 	qat_crypto_free_instances(accel_dev);
-	return -ENOMEM;
+	return ret;
 }
 
 static int qat_crypto_init(struct adf_accel_dev *accel_dev)
diff --git a/drivers/crypto/qat/qat_common/qat_crypto.h b/drivers/crypto/qat/qat_common/qat_crypto.h
index 12682d1e9f5f..b6a4c95ae003 100644
--- a/drivers/crypto/qat/qat_common/qat_crypto.h
+++ b/drivers/crypto/qat/qat_common/qat_crypto.h
@@ -3,6 +3,7 @@
 #ifndef _QAT_CRYPTO_INSTANCE_H_
 #define _QAT_CRYPTO_INSTANCE_H_
 
+#include <crypto/aes.h>
 #include <linux/list.h>
 #include <linux/slab.h>
 #include "adf_accel_devices.h"
@@ -44,8 +45,29 @@ struct qat_crypto_request {
 	struct qat_crypto_request_buffs buf;
 	void (*cb)(struct icp_qat_fw_la_resp *resp,
 		   struct qat_crypto_request *req);
-	void *iv;
-	dma_addr_t iv_paddr;
+	union {
+		struct {
+			__be64 iv_hi;
+			__be64 iv_lo;
+		};
+		u8 iv[AES_BLOCK_SIZE];
+	};
+	bool encryption;
 };
 
+static inline bool adf_hw_dev_has_crypto(struct adf_accel_dev *accel_dev)
+{
+	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
+	u32 mask = ~hw_device->accel_capabilities_mask;
+
+	if (mask & ADF_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC)
+		return false;
+	if (mask & ADF_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC)
+		return false;
+	if (mask & ADF_ACCEL_CAPABILITIES_AUTHENTICATION)
+		return false;
+
+	return true;
+}
+
 #endif
diff --git a/drivers/crypto/qat/qat_common/qat_hal.c b/drivers/crypto/qat/qat_common/qat_hal.c
index 6b9d47682d04..bd3028126cbe 100644
--- a/drivers/crypto/qat/qat_common/qat_hal.c
+++ b/drivers/crypto/qat/qat_common/qat_hal.c
@@ -33,7 +33,7 @@
 		((((const_val) << 12) & 0x0FF00000ull) | \
 		(((const_val) <<  0) & 0x000000FFull))))
 
-#define AE(handle, ae) handle->hal_handle->aes[ae]
+#define AE(handle, ae) ((handle)->hal_handle->aes[ae])
 
 static const u64 inst_4b[] = {
 	0x0F0400C0000ull, 0x0F4400C0000ull, 0x0F040000300ull, 0x0F440000300ull,
@@ -150,15 +150,15 @@ static int qat_hal_wait_cycles(struct icp_qat_fw_loader_handle *handle,
 	return 0;
 }
 
-#define CLR_BIT(wrd, bit) (wrd & ~(1 << bit))
-#define SET_BIT(wrd, bit) (wrd | 1 << bit)
+#define CLR_BIT(wrd, bit) ((wrd) & ~(1 << (bit)))
+#define SET_BIT(wrd, bit) ((wrd) | 1 << (bit))
 
 int qat_hal_set_ae_ctx_mode(struct icp_qat_fw_loader_handle *handle,
 			    unsigned char ae, unsigned char mode)
 {
 	unsigned int csr, new_csr;
 
-	if ((mode != 4) && (mode != 8)) {
+	if (mode != 4 && mode != 8) {
 		pr_err("QAT: bad ctx mode=%d\n", mode);
 		return -EINVAL;
 	}
@@ -210,6 +210,16 @@ int qat_hal_set_ae_lm_mode(struct icp_qat_fw_loader_handle *handle,
 			SET_BIT(csr, CE_LMADDR_1_GLOBAL_BITPOS) :
 			CLR_BIT(csr, CE_LMADDR_1_GLOBAL_BITPOS);
 		break;
+	case ICP_LMEM2:
+		new_csr = (mode) ?
+			SET_BIT(csr, CE_LMADDR_2_GLOBAL_BITPOS) :
+			CLR_BIT(csr, CE_LMADDR_2_GLOBAL_BITPOS);
+		break;
+	case ICP_LMEM3:
+		new_csr = (mode) ?
+			SET_BIT(csr, CE_LMADDR_3_GLOBAL_BITPOS) :
+			CLR_BIT(csr, CE_LMADDR_3_GLOBAL_BITPOS);
+		break;
 	default:
 		pr_err("QAT: lmType = 0x%x\n", lm_type);
 		return -EINVAL;
@@ -220,6 +230,20 @@ int qat_hal_set_ae_lm_mode(struct icp_qat_fw_loader_handle *handle,
 	return 0;
 }
 
+void qat_hal_set_ae_tindex_mode(struct icp_qat_fw_loader_handle *handle,
+				unsigned char ae, unsigned char mode)
+{
+	unsigned int csr, new_csr;
+
+	csr = qat_hal_rd_ae_csr(handle, ae, CTX_ENABLES);
+	csr &= IGNORE_W1C_MASK;
+	new_csr = (mode) ?
+		  SET_BIT(csr, CE_T_INDEX_GLOBAL_BITPOS) :
+		  CLR_BIT(csr, CE_T_INDEX_GLOBAL_BITPOS);
+	if (new_csr != csr)
+		qat_hal_wr_ae_csr(handle, ae, CTX_ENABLES, new_csr);
+}
+
 static unsigned short qat_hal_get_reg_addr(unsigned int type,
 					   unsigned short reg_num)
 {
@@ -259,6 +283,12 @@ static unsigned short qat_hal_get_reg_addr(unsigned int type,
 	case ICP_LMEM1:
 		reg_addr = 0x220;
 		break;
+	case ICP_LMEM2:
+		reg_addr = 0x2c0;
+		break;
+	case ICP_LMEM3:
+		reg_addr = 0x2e0;
+		break;
 	case ICP_NO_DEST:
 		reg_addr = 0x300 | (reg_num & 0xff);
 		break;
@@ -271,12 +301,13 @@ static unsigned short qat_hal_get_reg_addr(unsigned int type,
 
 void qat_hal_reset(struct icp_qat_fw_loader_handle *handle)
 {
-	unsigned int ae_reset_csr;
+	unsigned int reset_mask = handle->chip_info->icp_rst_mask;
+	unsigned int reset_csr = handle->chip_info->icp_rst_csr;
+	unsigned int csr_val;
 
-	ae_reset_csr = GET_GLB_CSR(handle, ICP_RESET);
-	ae_reset_csr |= handle->hal_handle->ae_mask << RST_CSR_AE_LSB;
-	ae_reset_csr |= handle->hal_handle->slice_mask << RST_CSR_QAT_LSB;
-	SET_GLB_CSR(handle, ICP_RESET, ae_reset_csr);
+	csr_val = GET_CAP_CSR(handle, reset_csr);
+	csr_val |= reset_mask;
+	SET_CAP_CSR(handle, reset_csr, csr_val);
 }
 
 static void qat_hal_wr_indr_csr(struct icp_qat_fw_loader_handle *handle,
@@ -346,11 +377,12 @@ static void qat_hal_put_wakeup_event(struct icp_qat_fw_loader_handle *handle,
 
 static int qat_hal_check_ae_alive(struct icp_qat_fw_loader_handle *handle)
 {
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
 	unsigned int base_cnt, cur_cnt;
 	unsigned char ae;
 	int times = MAX_RETRY_TIMES;
 
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		base_cnt = qat_hal_rd_ae_csr(handle, ae, PROFILE_COUNT);
 		base_cnt &= 0xffff;
 
@@ -384,21 +416,23 @@ int qat_hal_check_ae_active(struct icp_qat_fw_loader_handle *handle,
 
 static void qat_hal_reset_timestamp(struct icp_qat_fw_loader_handle *handle)
 {
-	unsigned int misc_ctl;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned int misc_ctl_csr, misc_ctl;
 	unsigned char ae;
 
+	misc_ctl_csr = handle->chip_info->misc_ctl_csr;
 	/* stop the timestamp timers */
-	misc_ctl = GET_GLB_CSR(handle, MISC_CONTROL);
+	misc_ctl = GET_CAP_CSR(handle, misc_ctl_csr);
 	if (misc_ctl & MC_TIMESTAMP_ENABLE)
-		SET_GLB_CSR(handle, MISC_CONTROL, misc_ctl &
+		SET_CAP_CSR(handle, misc_ctl_csr, misc_ctl &
 			    (~MC_TIMESTAMP_ENABLE));
 
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		qat_hal_wr_ae_csr(handle, ae, TIMESTAMP_LOW, 0);
 		qat_hal_wr_ae_csr(handle, ae, TIMESTAMP_HIGH, 0);
 	}
 	/* start timestamp timers */
-	SET_GLB_CSR(handle, MISC_CONTROL, misc_ctl | MC_TIMESTAMP_ENABLE);
+	SET_CAP_CSR(handle, misc_ctl_csr, misc_ctl | MC_TIMESTAMP_ENABLE);
 }
 
 #define ESRAM_AUTO_TINIT	BIT(2)
@@ -428,7 +462,7 @@ static int qat_hal_init_esram(struct icp_qat_fw_loader_handle *handle)
 		qat_hal_wait_cycles(handle, 0, ESRAM_AUTO_INIT_USED_CYCLES, 0);
 		csr_val = ADF_CSR_RD(csr_addr, 0);
 	} while (!(csr_val & ESRAM_AUTO_TINIT_DONE) && times--);
-	if ((times < 0)) {
+	if (times < 0) {
 		pr_err("QAT: Fail to init eSram!\n");
 		return -EFAULT;
 	}
@@ -438,33 +472,33 @@ static int qat_hal_init_esram(struct icp_qat_fw_loader_handle *handle)
 #define SHRAM_INIT_CYCLES 2060
 int qat_hal_clr_reset(struct icp_qat_fw_loader_handle *handle)
 {
-	unsigned int ae_reset_csr;
-	unsigned char ae;
-	unsigned int clk_csr;
+	unsigned int clk_csr = handle->chip_info->glb_clk_enable_csr;
+	unsigned int reset_mask = handle->chip_info->icp_rst_mask;
+	unsigned int reset_csr = handle->chip_info->icp_rst_csr;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned char ae = 0;
 	unsigned int times = 100;
-	unsigned int csr;
+	unsigned int csr_val;
 
 	/* write to the reset csr */
-	ae_reset_csr = GET_GLB_CSR(handle, ICP_RESET);
-	ae_reset_csr &= ~(handle->hal_handle->ae_mask << RST_CSR_AE_LSB);
-	ae_reset_csr &= ~(handle->hal_handle->slice_mask << RST_CSR_QAT_LSB);
+	csr_val = GET_CAP_CSR(handle, reset_csr);
+	csr_val &= ~reset_mask;
 	do {
-		SET_GLB_CSR(handle, ICP_RESET, ae_reset_csr);
+		SET_CAP_CSR(handle, reset_csr, csr_val);
 		if (!(times--))
 			goto out_err;
-		csr = GET_GLB_CSR(handle, ICP_RESET);
-	} while ((handle->hal_handle->ae_mask |
-		 (handle->hal_handle->slice_mask << RST_CSR_QAT_LSB)) & csr);
+		csr_val = GET_CAP_CSR(handle, reset_csr);
+		csr_val &= reset_mask;
+	} while (csr_val);
 	/* enable clock */
-	clk_csr = GET_GLB_CSR(handle, ICP_GLOBAL_CLK_ENABLE);
-	clk_csr |= handle->hal_handle->ae_mask << 0;
-	clk_csr |= handle->hal_handle->slice_mask << 20;
-	SET_GLB_CSR(handle, ICP_GLOBAL_CLK_ENABLE, clk_csr);
+	csr_val = GET_CAP_CSR(handle, clk_csr);
+	csr_val |= reset_mask;
+	SET_CAP_CSR(handle, clk_csr, csr_val);
 	if (qat_hal_check_ae_alive(handle))
 		goto out_err;
 
 	/* Set undefined power-up/reset states to reasonable default values */
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		qat_hal_wr_ae_csr(handle, ae, CTX_ENABLES,
 				  INIT_CTX_ENABLE_VALUE);
 		qat_hal_wr_indr_csr(handle, ae, ICP_QAT_UCLO_AE_ALL_CTX,
@@ -570,10 +604,11 @@ static void qat_hal_enable_ctx(struct icp_qat_fw_loader_handle *handle,
 
 static void qat_hal_clear_xfer(struct icp_qat_fw_loader_handle *handle)
 {
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
 	unsigned char ae;
 	unsigned short reg;
 
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		for (reg = 0; reg < ICP_QAT_UCLO_MAX_GPR_REG; reg++) {
 			qat_hal_init_rd_xfer(handle, ae, 0, ICP_SR_RD_ABS,
 					     reg, 0);
@@ -585,6 +620,7 @@ static void qat_hal_clear_xfer(struct icp_qat_fw_loader_handle *handle)
 
 static int qat_hal_clear_gpr(struct icp_qat_fw_loader_handle *handle)
 {
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
 	unsigned char ae;
 	unsigned int ctx_mask = ICP_QAT_UCLO_AE_ALL_CTX;
 	int times = MAX_RETRY_TIMES;
@@ -592,13 +628,15 @@ static int qat_hal_clear_gpr(struct icp_qat_fw_loader_handle *handle)
 	unsigned int savctx = 0;
 	int ret = 0;
 
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		csr_val = qat_hal_rd_ae_csr(handle, ae, AE_MISC_CONTROL);
 		csr_val &= ~(1 << MMC_SHARE_CS_BITPOS);
 		qat_hal_wr_ae_csr(handle, ae, AE_MISC_CONTROL, csr_val);
 		csr_val = qat_hal_rd_ae_csr(handle, ae, CTX_ENABLES);
 		csr_val &= IGNORE_W1C_MASK;
-		csr_val |= CE_NN_MODE;
+		if (handle->chip_info->nn)
+			csr_val |= CE_NN_MODE;
+
 		qat_hal_wr_ae_csr(handle, ae, CTX_ENABLES, csr_val);
 		qat_hal_wr_uwords(handle, ae, 0, ARRAY_SIZE(inst),
 				  (u64 *)inst);
@@ -613,7 +651,7 @@ static int qat_hal_clear_gpr(struct icp_qat_fw_loader_handle *handle)
 		qat_hal_wr_ae_csr(handle, ae, CTX_SIG_EVENTS_ACTIVE, 0);
 		qat_hal_enable_ctx(handle, ae, ctx_mask);
 	}
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		/* wait for AE to finish */
 		do {
 			ret = qat_hal_wait_cycles(handle, ae, 20, 1);
@@ -641,57 +679,143 @@ static int qat_hal_clear_gpr(struct icp_qat_fw_loader_handle *handle)
 	return 0;
 }
 
-#define ICP_QAT_AE_OFFSET	0x20000
-#define ICP_QAT_CAP_OFFSET       (ICP_QAT_AE_OFFSET + 0x10000)
-#define LOCAL_TO_XFER_REG_OFFSET    0x800
-#define ICP_QAT_EP_OFFSET	0x3a000
-int qat_hal_init(struct adf_accel_dev *accel_dev)
+static int qat_hal_chip_init(struct icp_qat_fw_loader_handle *handle,
+			     struct adf_accel_dev *accel_dev)
 {
-	unsigned char ae;
-	unsigned int max_en_ae_id = 0;
-	struct icp_qat_fw_loader_handle *handle;
 	struct adf_accel_pci *pci_info = &accel_dev->accel_pci_dev;
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
 	struct adf_bar *misc_bar =
 			&pci_info->pci_bars[hw_data->get_misc_bar_id(hw_data)];
+	unsigned int max_en_ae_id = 0;
 	struct adf_bar *sram_bar;
+	unsigned int csr_val = 0;
+	unsigned long ae_mask;
+	unsigned char ae = 0;
+	int ret = 0;
 
-	handle = kzalloc(sizeof(*handle), GFP_KERNEL);
-	if (!handle)
-		return -ENOMEM;
-
-	handle->hal_cap_g_ctl_csr_addr_v =
-		(void __iomem *)((uintptr_t)misc_bar->virt_addr +
-				 ICP_QAT_CAP_OFFSET);
-	handle->hal_cap_ae_xfer_csr_addr_v =
-		(void __iomem *)((uintptr_t)misc_bar->virt_addr +
-				 ICP_QAT_AE_OFFSET);
-	handle->hal_ep_csr_addr_v =
-		(void __iomem *)((uintptr_t)misc_bar->virt_addr +
-				 ICP_QAT_EP_OFFSET);
-	handle->hal_cap_ae_local_csr_addr_v =
-		(void __iomem *)((uintptr_t)handle->hal_cap_ae_xfer_csr_addr_v +
-				 LOCAL_TO_XFER_REG_OFFSET);
 	handle->pci_dev = pci_info->pci_dev;
-	if (handle->pci_dev->device == PCI_DEVICE_ID_INTEL_QAT_DH895XCC) {
+	switch (handle->pci_dev->device) {
+	case ADF_4XXX_PCI_DEVICE_ID:
+		handle->chip_info->sram_visible = false;
+		handle->chip_info->nn = false;
+		handle->chip_info->lm2lm3 = true;
+		handle->chip_info->lm_size = ICP_QAT_UCLO_MAX_LMEM_REG_2X;
+		handle->chip_info->icp_rst_csr = ICP_RESET_CPP0;
+		handle->chip_info->icp_rst_mask = 0x100015;
+		handle->chip_info->glb_clk_enable_csr = ICP_GLOBAL_CLK_ENABLE_CPP0;
+		handle->chip_info->misc_ctl_csr = MISC_CONTROL_C4XXX;
+		handle->chip_info->wakeup_event_val = 0x80000000;
+		handle->chip_info->fw_auth = true;
+		handle->chip_info->css_3k = true;
+		handle->chip_info->tgroup_share_ustore = true;
+		handle->chip_info->fcu_ctl_csr = FCU_CONTROL_4XXX;
+		handle->chip_info->fcu_sts_csr = FCU_STATUS_4XXX;
+		handle->chip_info->fcu_dram_addr_hi = FCU_DRAM_ADDR_HI_4XXX;
+		handle->chip_info->fcu_dram_addr_lo = FCU_DRAM_ADDR_LO_4XXX;
+		handle->chip_info->fcu_loaded_ae_csr = FCU_AE_LOADED_4XXX;
+		handle->chip_info->fcu_loaded_ae_pos = 0;
+
+		handle->hal_cap_g_ctl_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_CAP_OFFSET_4XXX);
+		handle->hal_cap_ae_xfer_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_AE_OFFSET_4XXX);
+		handle->hal_ep_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_EP_OFFSET_4XXX);
+		handle->hal_cap_ae_local_csr_addr_v =
+			(void __iomem *)((uintptr_t)handle->hal_cap_ae_xfer_csr_addr_v
+							+ LOCAL_TO_XFER_REG_OFFSET);
+		break;
+	case PCI_DEVICE_ID_INTEL_QAT_C62X:
+	case PCI_DEVICE_ID_INTEL_QAT_C3XXX:
+		handle->chip_info->sram_visible = false;
+		handle->chip_info->nn = true;
+		handle->chip_info->lm2lm3 = false;
+		handle->chip_info->lm_size = ICP_QAT_UCLO_MAX_LMEM_REG;
+		handle->chip_info->icp_rst_csr = ICP_RESET;
+		handle->chip_info->icp_rst_mask = (hw_data->ae_mask << RST_CSR_AE_LSB) |
+						  (hw_data->accel_mask << RST_CSR_QAT_LSB);
+		handle->chip_info->glb_clk_enable_csr = ICP_GLOBAL_CLK_ENABLE;
+		handle->chip_info->misc_ctl_csr = MISC_CONTROL;
+		handle->chip_info->wakeup_event_val = WAKEUP_EVENT;
+		handle->chip_info->fw_auth = true;
+		handle->chip_info->css_3k = false;
+		handle->chip_info->tgroup_share_ustore = false;
+		handle->chip_info->fcu_ctl_csr = FCU_CONTROL;
+		handle->chip_info->fcu_sts_csr = FCU_STATUS;
+		handle->chip_info->fcu_dram_addr_hi = FCU_DRAM_ADDR_HI;
+		handle->chip_info->fcu_dram_addr_lo = FCU_DRAM_ADDR_LO;
+		handle->chip_info->fcu_loaded_ae_csr = FCU_STATUS;
+		handle->chip_info->fcu_loaded_ae_pos = FCU_LOADED_AE_POS;
+		handle->hal_cap_g_ctl_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_CAP_OFFSET);
+		handle->hal_cap_ae_xfer_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_AE_OFFSET);
+		handle->hal_ep_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_EP_OFFSET);
+		handle->hal_cap_ae_local_csr_addr_v =
+			(void __iomem *)((uintptr_t)handle->hal_cap_ae_xfer_csr_addr_v
+			+ LOCAL_TO_XFER_REG_OFFSET);
+		break;
+	case PCI_DEVICE_ID_INTEL_QAT_DH895XCC:
+		handle->chip_info->sram_visible = true;
+		handle->chip_info->nn = true;
+		handle->chip_info->lm2lm3 = false;
+		handle->chip_info->lm_size = ICP_QAT_UCLO_MAX_LMEM_REG;
+		handle->chip_info->icp_rst_csr = ICP_RESET;
+		handle->chip_info->icp_rst_mask = (hw_data->ae_mask << RST_CSR_AE_LSB) |
+						  (hw_data->accel_mask << RST_CSR_QAT_LSB);
+		handle->chip_info->glb_clk_enable_csr = ICP_GLOBAL_CLK_ENABLE;
+		handle->chip_info->misc_ctl_csr = MISC_CONTROL;
+		handle->chip_info->wakeup_event_val = WAKEUP_EVENT;
+		handle->chip_info->fw_auth = false;
+		handle->chip_info->css_3k = false;
+		handle->chip_info->tgroup_share_ustore = false;
+		handle->chip_info->fcu_ctl_csr = 0;
+		handle->chip_info->fcu_sts_csr = 0;
+		handle->chip_info->fcu_dram_addr_hi = 0;
+		handle->chip_info->fcu_dram_addr_lo = 0;
+		handle->chip_info->fcu_loaded_ae_csr = 0;
+		handle->chip_info->fcu_loaded_ae_pos = 0;
+		handle->hal_cap_g_ctl_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_CAP_OFFSET);
+		handle->hal_cap_ae_xfer_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_AE_OFFSET);
+		handle->hal_ep_csr_addr_v =
+			(void __iomem *)((uintptr_t)misc_bar->virt_addr +
+			ICP_QAT_EP_OFFSET);
+		handle->hal_cap_ae_local_csr_addr_v =
+			(void __iomem *)((uintptr_t)handle->hal_cap_ae_xfer_csr_addr_v
+			+ LOCAL_TO_XFER_REG_OFFSET);
+		break;
+	default:
+		ret = -EINVAL;
+		goto out_err;
+	}
+
+	if (handle->chip_info->sram_visible) {
 		sram_bar =
 			&pci_info->pci_bars[hw_data->get_sram_bar_id(hw_data)];
 		handle->hal_sram_addr_v = sram_bar->virt_addr;
 	}
-	handle->fw_auth = (handle->pci_dev->device ==
-			   PCI_DEVICE_ID_INTEL_QAT_DH895XCC) ? false : true;
-	handle->hal_handle = kzalloc(sizeof(*handle->hal_handle), GFP_KERNEL);
-	if (!handle->hal_handle)
-		goto out_hal_handle;
 	handle->hal_handle->revision_id = accel_dev->accel_pci_dev.revid;
 	handle->hal_handle->ae_mask = hw_data->ae_mask;
+	handle->hal_handle->admin_ae_mask = hw_data->admin_ae_mask;
 	handle->hal_handle->slice_mask = hw_data->accel_mask;
+	handle->cfg_ae_mask = ALL_AE_MASK;
 	/* create AE objects */
 	handle->hal_handle->upc_mask = 0x1ffff;
 	handle->hal_handle->max_ustore = 0x4000;
-	for (ae = 0; ae < ICP_QAT_UCLO_MAX_AE; ae++) {
-		if (!(hw_data->ae_mask & (1 << ae)))
-			continue;
+
+	ae_mask = handle->hal_handle->ae_mask;
+	for_each_set_bit(ae, &ae_mask, ICP_QAT_UCLO_MAX_AE) {
 		handle->hal_handle->aes[ae].free_addr = 0;
 		handle->hal_handle->aes[ae].free_size =
 		    handle->hal_handle->max_ustore;
@@ -702,70 +826,116 @@ int qat_hal_init(struct adf_accel_dev *accel_dev)
 		max_en_ae_id = ae;
 	}
 	handle->hal_handle->ae_max_num = max_en_ae_id + 1;
+
+	/* Set SIGNATURE_ENABLE[0] to 0x1 in order to enable ALU_OUT csr */
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
+		csr_val = qat_hal_rd_ae_csr(handle, ae, SIGNATURE_ENABLE);
+		csr_val |= 0x1;
+		qat_hal_wr_ae_csr(handle, ae, SIGNATURE_ENABLE, csr_val);
+	}
+out_err:
+	return ret;
+}
+
+int qat_hal_init(struct adf_accel_dev *accel_dev)
+{
+	struct icp_qat_fw_loader_handle *handle;
+	int ret = 0;
+
+	handle = kzalloc(sizeof(*handle), GFP_KERNEL);
+	if (!handle)
+		return -ENOMEM;
+
+	handle->hal_handle = kzalloc(sizeof(*handle->hal_handle), GFP_KERNEL);
+	if (!handle->hal_handle) {
+		ret = -ENOMEM;
+		goto out_hal_handle;
+	}
+
+	handle->chip_info = kzalloc(sizeof(*handle->chip_info), GFP_KERNEL);
+	if (!handle->chip_info) {
+		ret = -ENOMEM;
+		goto out_chip_info;
+	}
+
+	ret = qat_hal_chip_init(handle, accel_dev);
+	if (ret) {
+		dev_err(&GET_DEV(accel_dev), "qat_hal_chip_init error\n");
+		goto out_err;
+	}
+
 	/* take all AEs out of reset */
-	if (qat_hal_clr_reset(handle)) {
+	ret = qat_hal_clr_reset(handle);
+	if (ret) {
 		dev_err(&GET_DEV(accel_dev), "qat_hal_clr_reset error\n");
 		goto out_err;
 	}
+
 	qat_hal_clear_xfer(handle);
-	if (!handle->fw_auth) {
-		if (qat_hal_clear_gpr(handle))
+	if (!handle->chip_info->fw_auth) {
+		ret = qat_hal_clear_gpr(handle);
+		if (ret)
 			goto out_err;
 	}
 
-	/* Set SIGNATURE_ENABLE[0] to 0x1 in order to enable ALU_OUT csr */
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
-		unsigned int csr_val = 0;
-
-		csr_val = qat_hal_rd_ae_csr(handle, ae, SIGNATURE_ENABLE);
-		csr_val |= 0x1;
-		qat_hal_wr_ae_csr(handle, ae, SIGNATURE_ENABLE, csr_val);
-	}
 	accel_dev->fw_loader->fw_loader = handle;
 	return 0;
 
 out_err:
+	kfree(handle->chip_info);
+out_chip_info:
 	kfree(handle->hal_handle);
 out_hal_handle:
 	kfree(handle);
-	return -EFAULT;
+	return ret;
 }
 
 void qat_hal_deinit(struct icp_qat_fw_loader_handle *handle)
 {
 	if (!handle)
 		return;
+	kfree(handle->chip_info);
 	kfree(handle->hal_handle);
 	kfree(handle);
 }
 
-void qat_hal_start(struct icp_qat_fw_loader_handle *handle, unsigned char ae,
-		   unsigned int ctx_mask)
+int qat_hal_start(struct icp_qat_fw_loader_handle *handle)
 {
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	u32 wakeup_val = handle->chip_info->wakeup_event_val;
+	u32 fcu_ctl_csr, fcu_sts_csr;
+	unsigned int fcu_sts;
+	unsigned char ae;
+	u32 ae_ctr = 0;
 	int retry = 0;
-	unsigned int fcu_sts = 0;
 
-	if (handle->fw_auth) {
-		SET_CAP_CSR(handle, FCU_CONTROL, FCU_CTRL_CMD_START);
+	if (handle->chip_info->fw_auth) {
+		fcu_ctl_csr = handle->chip_info->fcu_ctl_csr;
+		fcu_sts_csr = handle->chip_info->fcu_sts_csr;
+		ae_ctr = hweight32(ae_mask);
+		SET_CAP_CSR(handle, fcu_ctl_csr, FCU_CTRL_CMD_START);
 		do {
 			msleep(FW_AUTH_WAIT_PERIOD);
-			fcu_sts = GET_CAP_CSR(handle, FCU_STATUS);
+			fcu_sts = GET_CAP_CSR(handle, fcu_sts_csr);
 			if (((fcu_sts >> FCU_STS_DONE_POS) & 0x1))
-				return;
+				return ae_ctr;
 		} while (retry++ < FW_AUTH_MAX_RETRY);
-		pr_err("QAT: start error (AE 0x%x FCU_STS = 0x%x)\n", ae,
-		       fcu_sts);
+		pr_err("QAT: start error (FCU_STS = 0x%x)\n", fcu_sts);
+		return 0;
 	} else {
-		qat_hal_put_wakeup_event(handle, ae, (~ctx_mask) &
-				 ICP_QAT_UCLO_AE_ALL_CTX, 0x10000);
-		qat_hal_enable_ctx(handle, ae, ctx_mask);
+		for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
+			qat_hal_put_wakeup_event(handle, ae, 0, wakeup_val);
+			qat_hal_enable_ctx(handle, ae, ICP_QAT_UCLO_AE_ALL_CTX);
+			ae_ctr++;
+		}
+		return ae_ctr;
 	}
 }
 
 void qat_hal_stop(struct icp_qat_fw_loader_handle *handle, unsigned char ae,
 		  unsigned int ctx_mask)
 {
-	if (!handle->fw_auth)
+	if (!handle->chip_info->fw_auth)
 		qat_hal_disable_ctx(handle, ae, ctx_mask);
 }
 
@@ -832,9 +1002,12 @@ static int qat_hal_exec_micro_inst(struct icp_qat_fw_loader_handle *handle,
 				   int code_off, unsigned int max_cycle,
 				   unsigned int *endpc)
 {
+	unsigned int ind_lm_addr_byte0 = 0, ind_lm_addr_byte1 = 0;
+	unsigned int ind_lm_addr_byte2 = 0, ind_lm_addr_byte3 = 0;
+	unsigned int ind_t_index = 0, ind_t_index_byte = 0;
+	unsigned int ind_lm_addr0 = 0, ind_lm_addr1 = 0;
+	unsigned int ind_lm_addr2 = 0, ind_lm_addr3 = 0;
 	u64 savuwords[MAX_EXEC_INST];
-	unsigned int ind_lm_addr0, ind_lm_addr1;
-	unsigned int ind_lm_addr_byte0, ind_lm_addr_byte1;
 	unsigned int ind_cnt_sig;
 	unsigned int ind_sig, act_sig;
 	unsigned int csr_val = 0, newcsr_val;
@@ -853,6 +1026,20 @@ static int qat_hal_exec_micro_inst(struct icp_qat_fw_loader_handle *handle,
 						INDIRECT_LM_ADDR_0_BYTE_INDEX);
 	ind_lm_addr_byte1 = qat_hal_rd_indr_csr(handle, ae, ctx,
 						INDIRECT_LM_ADDR_1_BYTE_INDEX);
+	if (handle->chip_info->lm2lm3) {
+		ind_lm_addr2 = qat_hal_rd_indr_csr(handle, ae, ctx,
+						   LM_ADDR_2_INDIRECT);
+		ind_lm_addr3 = qat_hal_rd_indr_csr(handle, ae, ctx,
+						   LM_ADDR_3_INDIRECT);
+		ind_lm_addr_byte2 = qat_hal_rd_indr_csr(handle, ae, ctx,
+							INDIRECT_LM_ADDR_2_BYTE_INDEX);
+		ind_lm_addr_byte3 = qat_hal_rd_indr_csr(handle, ae, ctx,
+							INDIRECT_LM_ADDR_3_BYTE_INDEX);
+		ind_t_index = qat_hal_rd_indr_csr(handle, ae, ctx,
+						  INDIRECT_T_INDEX);
+		ind_t_index_byte = qat_hal_rd_indr_csr(handle, ae, ctx,
+						       INDIRECT_T_INDEX_BYTE_INDEX);
+	}
 	if (inst_num <= MAX_EXEC_INST)
 		qat_hal_get_uwords(handle, ae, 0, inst_num, savuwords);
 	qat_hal_get_wakeup_event(handle, ae, ctx, &wakeup_events);
@@ -910,6 +1097,23 @@ static int qat_hal_exec_micro_inst(struct icp_qat_fw_loader_handle *handle,
 			    INDIRECT_LM_ADDR_0_BYTE_INDEX, ind_lm_addr_byte0);
 	qat_hal_wr_indr_csr(handle, ae, (1 << ctx),
 			    INDIRECT_LM_ADDR_1_BYTE_INDEX, ind_lm_addr_byte1);
+	if (handle->chip_info->lm2lm3) {
+		qat_hal_wr_indr_csr(handle, ae, BIT(ctx), LM_ADDR_2_INDIRECT,
+				    ind_lm_addr2);
+		qat_hal_wr_indr_csr(handle, ae, BIT(ctx), LM_ADDR_3_INDIRECT,
+				    ind_lm_addr3);
+		qat_hal_wr_indr_csr(handle, ae, BIT(ctx),
+				    INDIRECT_LM_ADDR_2_BYTE_INDEX,
+				    ind_lm_addr_byte2);
+		qat_hal_wr_indr_csr(handle, ae, BIT(ctx),
+				    INDIRECT_LM_ADDR_3_BYTE_INDEX,
+				    ind_lm_addr_byte3);
+		qat_hal_wr_indr_csr(handle, ae, BIT(ctx),
+				    INDIRECT_T_INDEX, ind_t_index);
+		qat_hal_wr_indr_csr(handle, ae, BIT(ctx),
+				    INDIRECT_T_INDEX_BYTE_INDEX,
+				    ind_t_index_byte);
+	}
 	qat_hal_wr_indr_csr(handle, ae, (1 << ctx),
 			    FUTURE_COUNT_SIGNAL_INDIRECT, ind_cnt_sig);
 	qat_hal_wr_indr_csr(handle, ae, (1 << ctx),
@@ -1125,7 +1329,7 @@ int qat_hal_batch_wr_lm(struct icp_qat_fw_loader_handle *handle,
 		plm_init = plm_init->next;
 	}
 	/* exec micro codes */
-	if (micro_inst_arry && (micro_inst_num > 0)) {
+	if (micro_inst_arry && micro_inst_num > 0) {
 		micro_inst_arry[micro_inst_num++] = 0x0E000010000ull;
 		stat = qat_hal_exec_micro_init_lm(handle, ae, 0, &first_exec,
 						  micro_inst_arry,
@@ -1146,7 +1350,7 @@ static int qat_hal_put_rel_rd_xfer(struct icp_qat_fw_loader_handle *handle,
 	unsigned short mask;
 	unsigned short dr_offset = 0x10;
 
-	status = ctx_enables = qat_hal_rd_ae_csr(handle, ae, CTX_ENABLES);
+	ctx_enables = qat_hal_rd_ae_csr(handle, ae, CTX_ENABLES);
 	if (CE_INUSE_CONTEXTS & ctx_enables) {
 		if (ctx & 0x1) {
 			pr_err("QAT: bad 4-ctx mode,ctx=0x%x\n", ctx);
@@ -1271,7 +1475,7 @@ static int qat_hal_convert_abs_to_rel(struct icp_qat_fw_loader_handle
 }
 
 int qat_hal_init_gpr(struct icp_qat_fw_loader_handle *handle,
-		     unsigned char ae, unsigned char ctx_mask,
+		     unsigned char ae, unsigned long ctx_mask,
 		     enum icp_qat_uof_regtype reg_type,
 		     unsigned short reg_num, unsigned int regdata)
 {
@@ -1291,7 +1495,7 @@ int qat_hal_init_gpr(struct icp_qat_fw_loader_handle *handle,
 		} else {
 			reg = reg_num;
 			type = reg_type;
-			if (!test_bit(ctx, (unsigned long *)&ctx_mask))
+			if (!test_bit(ctx, &ctx_mask))
 				continue;
 		}
 		stat = qat_hal_wr_rel_reg(handle, ae, ctx, type, reg, regdata);
@@ -1305,7 +1509,7 @@ int qat_hal_init_gpr(struct icp_qat_fw_loader_handle *handle,
 }
 
 int qat_hal_init_wr_xfer(struct icp_qat_fw_loader_handle *handle,
-			 unsigned char ae, unsigned char ctx_mask,
+			 unsigned char ae, unsigned long ctx_mask,
 			 enum icp_qat_uof_regtype reg_type,
 			 unsigned short reg_num, unsigned int regdata)
 {
@@ -1325,7 +1529,7 @@ int qat_hal_init_wr_xfer(struct icp_qat_fw_loader_handle *handle,
 		} else {
 			reg = reg_num;
 			type = reg_type;
-			if (!test_bit(ctx, (unsigned long *)&ctx_mask))
+			if (!test_bit(ctx, &ctx_mask))
 				continue;
 		}
 		stat = qat_hal_put_rel_wr_xfer(handle, ae, ctx, type, reg,
@@ -1340,7 +1544,7 @@ int qat_hal_init_wr_xfer(struct icp_qat_fw_loader_handle *handle,
 }
 
 int qat_hal_init_rd_xfer(struct icp_qat_fw_loader_handle *handle,
-			 unsigned char ae, unsigned char ctx_mask,
+			 unsigned char ae, unsigned long ctx_mask,
 			 enum icp_qat_uof_regtype reg_type,
 			 unsigned short reg_num, unsigned int regdata)
 {
@@ -1360,7 +1564,7 @@ int qat_hal_init_rd_xfer(struct icp_qat_fw_loader_handle *handle,
 		} else {
 			reg = reg_num;
 			type = reg_type;
-			if (!test_bit(ctx, (unsigned long *)&ctx_mask))
+			if (!test_bit(ctx, &ctx_mask))
 				continue;
 		}
 		stat = qat_hal_put_rel_rd_xfer(handle, ae, ctx, type, reg,
@@ -1375,17 +1579,22 @@ int qat_hal_init_rd_xfer(struct icp_qat_fw_loader_handle *handle,
 }
 
 int qat_hal_init_nn(struct icp_qat_fw_loader_handle *handle,
-		    unsigned char ae, unsigned char ctx_mask,
+		    unsigned char ae, unsigned long ctx_mask,
 		    unsigned short reg_num, unsigned int regdata)
 {
 	int stat = 0;
 	unsigned char ctx;
+	if (!handle->chip_info->nn) {
+		dev_err(&handle->pci_dev->dev, "QAT: No next neigh in 0x%x\n",
+			handle->pci_dev->device);
+		return -EINVAL;
+	}
 
 	if (ctx_mask == 0)
 		return -EINVAL;
 
 	for (ctx = 0; ctx < ICP_QAT_UCLO_MAX_CTX; ctx++) {
-		if (!test_bit(ctx, (unsigned long *)&ctx_mask))
+		if (!test_bit(ctx, &ctx_mask))
 			continue;
 		stat = qat_hal_put_rel_nn(handle, ae, ctx, reg_num, regdata);
 		if (stat) {
diff --git a/drivers/crypto/qat/qat_common/qat_uclo.c b/drivers/crypto/qat/qat_common/qat_uclo.c
index 5d1f28cd6680..1fb5fc852f6b 100644
--- a/drivers/crypto/qat/qat_common/qat_uclo.c
+++ b/drivers/crypto/qat/qat_common/qat_uclo.c
@@ -74,7 +74,7 @@ static int qat_uclo_free_ae_data(struct icp_qat_uclo_aedata *ae_data)
 static char *qat_uclo_get_string(struct icp_qat_uof_strtable *str_table,
 				 unsigned int str_offset)
 {
-	if ((!str_table->table_len) || (str_offset > str_table->table_len))
+	if (!str_table->table_len || str_offset > str_table->table_len)
 		return NULL;
 	return (char *)(((uintptr_t)(str_table->strings)) + str_offset);
 }
@@ -311,7 +311,7 @@ static int qat_uclo_init_lmem_seg(struct icp_qat_fw_loader_handle *handle,
 	unsigned int ae;
 
 	if (qat_uclo_fetch_initmem_ae(handle, init_mem,
-				      ICP_QAT_UCLO_MAX_LMEM_REG, &ae))
+				      handle->chip_info->lm_size, &ae))
 		return -EINVAL;
 	if (qat_uclo_create_batch_init_list(handle, init_mem, ae,
 					    &obj_handle->lm_init_tab[ae]))
@@ -324,6 +324,7 @@ static int qat_uclo_init_umem_seg(struct icp_qat_fw_loader_handle *handle,
 {
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
 	unsigned int ae, ustore_size, uaddr, i;
+	struct icp_qat_uclo_aedata *aed;
 
 	ustore_size = obj_handle->ustore_phy_size;
 	if (qat_uclo_fetch_initmem_ae(handle, init_mem, ustore_size, &ae))
@@ -333,11 +334,10 @@ static int qat_uclo_init_umem_seg(struct icp_qat_fw_loader_handle *handle,
 		return -EINVAL;
 	/* set the highest ustore address referenced */
 	uaddr = (init_mem->addr + init_mem->num_in_bytes) >> 0x2;
-	for (i = 0; i < obj_handle->ae_data[ae].slice_num; i++) {
-		if (obj_handle->ae_data[ae].ae_slices[i].
-		    encap_image->uwords_num < uaddr)
-			obj_handle->ae_data[ae].ae_slices[i].
-			encap_image->uwords_num = uaddr;
+	aed = &obj_handle->ae_data[ae];
+	for (i = 0; i < aed->slice_num; i++) {
+		if (aed->ae_slices[i].encap_image->uwords_num < uaddr)
+			aed->ae_slices[i].encap_image->uwords_num = uaddr;
 	}
 	return 0;
 }
@@ -373,6 +373,8 @@ static int qat_uclo_init_ustore(struct icp_qat_fw_loader_handle *handle,
 	unsigned int ustore_size;
 	unsigned int patt_pos;
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned long cfg_ae_mask = handle->cfg_ae_mask;
 	u64 *fill_data;
 
 	uof_image = image->img_ptr;
@@ -385,9 +387,13 @@ static int qat_uclo_init_ustore(struct icp_qat_fw_loader_handle *handle,
 		       sizeof(u64));
 	page = image->page;
 
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		if (!test_bit(ae, (unsigned long *)&uof_image->ae_assigned))
 			continue;
+
+		if (!test_bit(ae, &cfg_ae_mask))
+			continue;
+
 		ustore_size = obj_handle->ae_data[ae].eff_ustore_size;
 		patt_pos = page->beg_addr_p + page->micro_words_num;
 
@@ -406,6 +412,7 @@ static int qat_uclo_init_memory(struct icp_qat_fw_loader_handle *handle)
 	int i, ae;
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
 	struct icp_qat_uof_initmem *initmem = obj_handle->init_mem_tab.init_mem;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
 
 	for (i = 0; i < obj_handle->init_mem_tab.entry_num; i++) {
 		if (initmem->num_in_bytes) {
@@ -418,7 +425,8 @@ static int qat_uclo_init_memory(struct icp_qat_fw_loader_handle *handle)
 			(sizeof(struct icp_qat_uof_memvar_attr) *
 			initmem->val_attr_num));
 	}
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
+
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
 		if (qat_hal_batch_wr_lm(handle, ae,
 					obj_handle->lm_init_tab[ae])) {
 			pr_err("QAT: fail to batch init lmem for AE %d\n", ae);
@@ -536,7 +544,7 @@ qat_uclo_check_image_compat(struct icp_qat_uof_encap_obj *encap_uof_obj,
 			(encap_uof_obj->beg_uof +
 			code_page->neigh_reg_tab_offset);
 	if (neigh_reg_tab->entry_num) {
-		pr_err("QAT: UOF can't contain shared control store feature\n");
+		pr_err("QAT: UOF can't contain neighbor register table\n");
 		return -EINVAL;
 	}
 	if (image->numpages > 1) {
@@ -649,11 +657,13 @@ static int qat_uclo_map_ae(struct icp_qat_fw_loader_handle *handle, int max_ae)
 	int i, ae;
 	int mflag = 0;
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned long cfg_ae_mask = handle->cfg_ae_mask;
 
-	for (ae = 0; ae < max_ae; ae++) {
-		if (!test_bit(ae,
-			      (unsigned long *)&handle->hal_handle->ae_mask))
+	for_each_set_bit(ae, &ae_mask, max_ae) {
+		if (!test_bit(ae, &cfg_ae_mask))
 			continue;
+
 		for (i = 0; i < obj_handle->uimage_num; i++) {
 			if (!test_bit(ae, (unsigned long *)
 			&obj_handle->ae_uimage[i].img_ptr->ae_assigned))
@@ -718,6 +728,8 @@ qat_uclo_get_dev_type(struct icp_qat_fw_loader_handle *handle)
 		return ICP_QAT_AC_C62X_DEV_TYPE;
 	case PCI_DEVICE_ID_INTEL_QAT_C3XXX:
 		return ICP_QAT_AC_C3XXX_DEV_TYPE;
+	case ADF_4XXX_PCI_DEVICE_ID:
+		return ICP_QAT_AC_4XXX_A_DEV_TYPE;
 	default:
 		pr_err("QAT: unsupported device 0x%x\n",
 		       handle->pci_dev->device);
@@ -736,8 +748,8 @@ static int qat_uclo_check_uof_compat(struct icp_qat_uclo_objhandle *obj_handle)
 		return -EINVAL;
 	}
 	maj_ver = obj_handle->prod_rev & 0xff;
-	if ((obj_handle->encap_uof_obj.obj_hdr->max_cpu_ver < maj_ver) ||
-	    (obj_handle->encap_uof_obj.obj_hdr->min_cpu_ver > maj_ver)) {
+	if (obj_handle->encap_uof_obj.obj_hdr->max_cpu_ver < maj_ver ||
+	    obj_handle->encap_uof_obj.obj_hdr->min_cpu_ver > maj_ver) {
 		pr_err("QAT: UOF majVer 0x%x out of range\n", maj_ver);
 		return -EINVAL;
 	}
@@ -845,6 +857,8 @@ static int qat_uclo_init_reg_sym(struct icp_qat_fw_loader_handle *handle,
 static int qat_uclo_init_globals(struct icp_qat_fw_loader_handle *handle)
 {
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	struct icp_qat_uclo_aedata *aed;
 	unsigned int s, ae;
 
 	if (obj_handle->global_inited)
@@ -855,13 +869,13 @@ static int qat_uclo_init_globals(struct icp_qat_fw_loader_handle *handle)
 			return -EINVAL;
 		}
 	}
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
-		for (s = 0; s < obj_handle->ae_data[ae].slice_num; s++) {
-			if (!obj_handle->ae_data[ae].ae_slices[s].encap_image)
+
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
+		aed = &obj_handle->ae_data[ae];
+		for (s = 0; s < aed->slice_num; s++) {
+			if (!aed->ae_slices[s].encap_image)
 				continue;
-			if (qat_uclo_init_reg_sym(handle, ae,
-						  obj_handle->ae_data[ae].
-						  ae_slices[s].encap_image))
+			if (qat_uclo_init_reg_sym(handle, ae, aed->ae_slices[s].encap_image))
 				return -EINVAL;
 		}
 	}
@@ -869,46 +883,83 @@ static int qat_uclo_init_globals(struct icp_qat_fw_loader_handle *handle)
 	return 0;
 }
 
+static int qat_hal_set_modes(struct icp_qat_fw_loader_handle *handle,
+			     struct icp_qat_uclo_objhandle *obj_handle,
+			     unsigned char ae,
+			     struct icp_qat_uof_image *uof_image)
+{
+	unsigned char mode;
+	int ret;
+
+	mode = ICP_QAT_CTX_MODE(uof_image->ae_mode);
+	ret = qat_hal_set_ae_ctx_mode(handle, ae, mode);
+	if (ret) {
+		pr_err("QAT: qat_hal_set_ae_ctx_mode error\n");
+		return ret;
+	}
+	if (handle->chip_info->nn) {
+		mode = ICP_QAT_NN_MODE(uof_image->ae_mode);
+		ret = qat_hal_set_ae_nn_mode(handle, ae, mode);
+		if (ret) {
+			pr_err("QAT: qat_hal_set_ae_nn_mode error\n");
+			return ret;
+		}
+	}
+	mode = ICP_QAT_LOC_MEM0_MODE(uof_image->ae_mode);
+	ret = qat_hal_set_ae_lm_mode(handle, ae, ICP_LMEM0, mode);
+	if (ret) {
+		pr_err("QAT: qat_hal_set_ae_lm_mode LMEM0 error\n");
+		return ret;
+	}
+	mode = ICP_QAT_LOC_MEM1_MODE(uof_image->ae_mode);
+	ret = qat_hal_set_ae_lm_mode(handle, ae, ICP_LMEM1, mode);
+	if (ret) {
+		pr_err("QAT: qat_hal_set_ae_lm_mode LMEM1 error\n");
+		return ret;
+	}
+	if (handle->chip_info->lm2lm3) {
+		mode = ICP_QAT_LOC_MEM2_MODE(uof_image->ae_mode);
+		ret = qat_hal_set_ae_lm_mode(handle, ae, ICP_LMEM2, mode);
+		if (ret) {
+			pr_err("QAT: qat_hal_set_ae_lm_mode LMEM2 error\n");
+			return ret;
+		}
+		mode = ICP_QAT_LOC_MEM3_MODE(uof_image->ae_mode);
+		ret = qat_hal_set_ae_lm_mode(handle, ae, ICP_LMEM3, mode);
+		if (ret) {
+			pr_err("QAT: qat_hal_set_ae_lm_mode LMEM3 error\n");
+			return ret;
+		}
+		mode = ICP_QAT_LOC_TINDEX_MODE(uof_image->ae_mode);
+		qat_hal_set_ae_tindex_mode(handle, ae, mode);
+	}
+	return 0;
+}
+
 static int qat_uclo_set_ae_mode(struct icp_qat_fw_loader_handle *handle)
 {
-	unsigned char ae, nn_mode, s;
 	struct icp_qat_uof_image *uof_image;
 	struct icp_qat_uclo_aedata *ae_data;
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned long cfg_ae_mask = handle->cfg_ae_mask;
+	unsigned char ae, s;
+	int error;
 
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
-		if (!test_bit(ae,
-			      (unsigned long *)&handle->hal_handle->ae_mask))
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
+		if (!test_bit(ae, &cfg_ae_mask))
 			continue;
+
 		ae_data = &obj_handle->ae_data[ae];
 		for (s = 0; s < min_t(unsigned int, ae_data->slice_num,
 				      ICP_QAT_UCLO_MAX_CTX); s++) {
 			if (!obj_handle->ae_data[ae].ae_slices[s].encap_image)
 				continue;
 			uof_image = ae_data->ae_slices[s].encap_image->img_ptr;
-			if (qat_hal_set_ae_ctx_mode(handle, ae,
-						    (char)ICP_QAT_CTX_MODE
-						    (uof_image->ae_mode))) {
-				pr_err("QAT: qat_hal_set_ae_ctx_mode error\n");
-				return -EFAULT;
-			}
-			nn_mode = ICP_QAT_NN_MODE(uof_image->ae_mode);
-			if (qat_hal_set_ae_nn_mode(handle, ae, nn_mode)) {
-				pr_err("QAT: qat_hal_set_ae_nn_mode error\n");
-				return -EFAULT;
-			}
-			if (qat_hal_set_ae_lm_mode(handle, ae, ICP_LMEM0,
-						   (char)ICP_QAT_LOC_MEM0_MODE
-						   (uof_image->ae_mode))) {
-				pr_err("QAT: qat_hal_set_ae_lm_mode LMEM0 error\n");
-				return -EFAULT;
-			}
-			if (qat_hal_set_ae_lm_mode(handle, ae, ICP_LMEM1,
-						   (char)ICP_QAT_LOC_MEM1_MODE
-						   (uof_image->ae_mode))) {
-				pr_err("QAT: qat_hal_set_ae_lm_mode LMEM1 error\n");
-				return -EFAULT;
-			}
+			error = qat_hal_set_modes(handle, obj_handle, ae,
+						  uof_image);
+			if (error)
+				return error;
 		}
 	}
 	return 0;
@@ -1003,10 +1054,11 @@ static int qat_uclo_map_suof_file_hdr(struct icp_qat_fw_loader_handle *handle,
 	return 0;
 }
 
-static void qat_uclo_map_simg(struct icp_qat_suof_handle *suof_handle,
+static void qat_uclo_map_simg(struct icp_qat_fw_loader_handle *handle,
 			      struct icp_qat_suof_img_hdr *suof_img_hdr,
 			      struct icp_qat_suof_chunk_hdr *suof_chunk_hdr)
 {
+	struct icp_qat_suof_handle *suof_handle = handle->sobj_handle;
 	struct icp_qat_simg_ae_mode *ae_mode;
 	struct icp_qat_suof_objhdr *suof_objhdr;
 
@@ -1021,10 +1073,10 @@ static void qat_uclo_map_simg(struct icp_qat_suof_handle *suof_handle,
 	suof_img_hdr->css_key = (suof_img_hdr->css_header +
 				 sizeof(struct icp_qat_css_hdr));
 	suof_img_hdr->css_signature = suof_img_hdr->css_key +
-				      ICP_QAT_CSS_FWSK_MODULUS_LEN +
-				      ICP_QAT_CSS_FWSK_EXPONENT_LEN;
+				      ICP_QAT_CSS_FWSK_MODULUS_LEN(handle) +
+				      ICP_QAT_CSS_FWSK_EXPONENT_LEN(handle);
 	suof_img_hdr->css_simg = suof_img_hdr->css_signature +
-				 ICP_QAT_CSS_SIGNATURE_LEN;
+				 ICP_QAT_CSS_SIGNATURE_LEN(handle);
 
 	ae_mode = (struct icp_qat_simg_ae_mode *)(suof_img_hdr->css_simg);
 	suof_img_hdr->ae_mask = ae_mode->ae_mask;
@@ -1064,8 +1116,8 @@ static int qat_uclo_check_simg_compat(struct icp_qat_fw_loader_handle *handle,
 		return -EINVAL;
 	}
 	maj_ver = prod_rev & 0xff;
-	if ((maj_ver > img_ae_mode->devmax_ver) ||
-	    (maj_ver < img_ae_mode->devmin_ver)) {
+	if (maj_ver > img_ae_mode->devmax_ver ||
+	    maj_ver < img_ae_mode->devmin_ver) {
 		pr_err("QAT: incompatible device majver 0x%x\n", maj_ver);
 		return -EINVAL;
 	}
@@ -1108,7 +1160,7 @@ static int qat_uclo_map_suof(struct icp_qat_fw_loader_handle *handle,
 	unsigned int i = 0;
 	struct icp_qat_suof_img_hdr img_header;
 
-	if (!suof_ptr || (suof_size == 0)) {
+	if (!suof_ptr || suof_size == 0) {
 		pr_err("QAT: input parameter SUOF pointer/size is NULL\n");
 		return -EINVAL;
 	}
@@ -1130,20 +1182,24 @@ static int qat_uclo_map_suof(struct icp_qat_fw_loader_handle *handle,
 		if (!suof_img_hdr)
 			return -ENOMEM;
 		suof_handle->img_table.simg_hdr = suof_img_hdr;
-	}
 
-	for (i = 0; i < suof_handle->img_table.num_simgs; i++) {
-		qat_uclo_map_simg(handle->sobj_handle, &suof_img_hdr[i],
-				  &suof_chunk_hdr[1 + i]);
-		ret = qat_uclo_check_simg_compat(handle,
-						 &suof_img_hdr[i]);
-		if (ret)
-			return ret;
-		if ((suof_img_hdr[i].ae_mask & 0x1) != 0)
-			ae0_img = i;
+		for (i = 0; i < suof_handle->img_table.num_simgs; i++) {
+			qat_uclo_map_simg(handle, &suof_img_hdr[i],
+					  &suof_chunk_hdr[1 + i]);
+			ret = qat_uclo_check_simg_compat(handle,
+							 &suof_img_hdr[i]);
+			if (ret)
+				return ret;
+			suof_img_hdr[i].ae_mask &= handle->cfg_ae_mask;
+			if ((suof_img_hdr[i].ae_mask & 0x1) != 0)
+				ae0_img = i;
+		}
+
+		if (!handle->chip_info->tgroup_share_ustore) {
+			qat_uclo_tail_img(suof_img_hdr, ae0_img,
+					  suof_handle->img_table.num_simgs);
+		}
 	}
-	qat_uclo_tail_img(suof_img_hdr, ae0_img,
-			  suof_handle->img_table.num_simgs);
 	return 0;
 }
 
@@ -1153,18 +1209,26 @@ static int qat_uclo_map_suof(struct icp_qat_fw_loader_handle *handle,
 static int qat_uclo_auth_fw(struct icp_qat_fw_loader_handle *handle,
 			    struct icp_qat_fw_auth_desc *desc)
 {
-	unsigned int fcu_sts, retry = 0;
+	u32 fcu_sts, retry = 0;
+	u32 fcu_ctl_csr, fcu_sts_csr;
+	u32 fcu_dram_hi_csr, fcu_dram_lo_csr;
 	u64 bus_addr;
 
 	bus_addr = ADD_ADDR(desc->css_hdr_high, desc->css_hdr_low)
 			   - sizeof(struct icp_qat_auth_chunk);
-	SET_CAP_CSR(handle, FCU_DRAM_ADDR_HI, (bus_addr >> BITS_IN_DWORD));
-	SET_CAP_CSR(handle, FCU_DRAM_ADDR_LO, bus_addr);
-	SET_CAP_CSR(handle, FCU_CONTROL, FCU_CTRL_CMD_AUTH);
+
+	fcu_ctl_csr = handle->chip_info->fcu_ctl_csr;
+	fcu_sts_csr = handle->chip_info->fcu_sts_csr;
+	fcu_dram_hi_csr = handle->chip_info->fcu_dram_addr_hi;
+	fcu_dram_lo_csr = handle->chip_info->fcu_dram_addr_lo;
+
+	SET_CAP_CSR(handle, fcu_dram_hi_csr, (bus_addr >> BITS_IN_DWORD));
+	SET_CAP_CSR(handle, fcu_dram_lo_csr, bus_addr);
+	SET_CAP_CSR(handle, fcu_ctl_csr, FCU_CTRL_CMD_AUTH);
 
 	do {
 		msleep(FW_AUTH_WAIT_PERIOD);
-		fcu_sts = GET_CAP_CSR(handle, FCU_STATUS);
+		fcu_sts = GET_CAP_CSR(handle, fcu_sts_csr);
 		if ((fcu_sts & FCU_AUTH_STS_MASK) == FCU_STS_VERI_FAIL)
 			goto auth_fail;
 		if (((fcu_sts >> FCU_STS_AUTHFWLD_POS) & 0x1))
@@ -1177,6 +1241,83 @@ auth_fail:
 	return -EINVAL;
 }
 
+static bool qat_uclo_is_broadcast(struct icp_qat_fw_loader_handle *handle,
+				  int imgid)
+{
+	struct icp_qat_suof_handle *sobj_handle;
+
+	if (!handle->chip_info->tgroup_share_ustore)
+		return false;
+
+	sobj_handle = (struct icp_qat_suof_handle *)handle->sobj_handle;
+	if (handle->hal_handle->admin_ae_mask &
+	    sobj_handle->img_table.simg_hdr[imgid].ae_mask)
+		return false;
+
+	return true;
+}
+
+static int qat_uclo_broadcast_load_fw(struct icp_qat_fw_loader_handle *handle,
+				      struct icp_qat_fw_auth_desc *desc)
+{
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned long desc_ae_mask = desc->ae_mask;
+	u32 fcu_sts, ae_broadcast_mask = 0;
+	u32 fcu_loaded_csr, ae_loaded;
+	u32 fcu_sts_csr, fcu_ctl_csr;
+	unsigned int ae, retry = 0;
+
+	if (handle->chip_info->tgroup_share_ustore) {
+		fcu_ctl_csr = handle->chip_info->fcu_ctl_csr;
+		fcu_sts_csr = handle->chip_info->fcu_sts_csr;
+		fcu_loaded_csr = handle->chip_info->fcu_loaded_ae_csr;
+	} else {
+		pr_err("Chip 0x%x doesn't support broadcast load\n",
+		       handle->pci_dev->device);
+		return -EINVAL;
+	}
+
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
+		if (qat_hal_check_ae_active(handle, (unsigned char)ae)) {
+			pr_err("QAT: Broadcast load failed. AE is not enabled or active.\n");
+			return -EINVAL;
+		}
+
+		if (test_bit(ae, &desc_ae_mask))
+			ae_broadcast_mask |= 1 << ae;
+	}
+
+	if (ae_broadcast_mask) {
+		SET_CAP_CSR(handle, FCU_ME_BROADCAST_MASK_TYPE,
+			    ae_broadcast_mask);
+
+		SET_CAP_CSR(handle, fcu_ctl_csr, FCU_CTRL_CMD_LOAD);
+
+		do {
+			msleep(FW_AUTH_WAIT_PERIOD);
+			fcu_sts = GET_CAP_CSR(handle, fcu_sts_csr);
+			fcu_sts &= FCU_AUTH_STS_MASK;
+
+			if (fcu_sts == FCU_STS_LOAD_FAIL) {
+				pr_err("Broadcast load failed: 0x%x)\n", fcu_sts);
+				return -EINVAL;
+			} else if (fcu_sts == FCU_STS_LOAD_DONE) {
+				ae_loaded = GET_CAP_CSR(handle, fcu_loaded_csr);
+				ae_loaded >>= handle->chip_info->fcu_loaded_ae_pos;
+
+				if ((ae_loaded & ae_broadcast_mask) == ae_broadcast_mask)
+					break;
+			}
+		} while (retry++ < FW_AUTH_MAX_RETRY);
+
+		if (retry > FW_AUTH_MAX_RETRY) {
+			pr_err("QAT: broadcast load failed timeout %d\n", retry);
+			return -EINVAL;
+		}
+	}
+	return 0;
+}
+
 static int qat_uclo_simg_alloc(struct icp_qat_fw_loader_handle *handle,
 			       struct icp_firml_dram_desc *dram_desc,
 			       unsigned int size)
@@ -1197,11 +1338,15 @@ static int qat_uclo_simg_alloc(struct icp_qat_fw_loader_handle *handle,
 static void qat_uclo_simg_free(struct icp_qat_fw_loader_handle *handle,
 			       struct icp_firml_dram_desc *dram_desc)
 {
-	dma_free_coherent(&handle->pci_dev->dev,
-			  (size_t)(dram_desc->dram_size),
-			  (dram_desc->dram_base_addr_v),
-			  dram_desc->dram_bus_addr);
-	memset(dram_desc, 0, sizeof(*dram_desc));
+	if (handle && dram_desc && dram_desc->dram_base_addr_v) {
+		dma_free_coherent(&handle->pci_dev->dev,
+				  (size_t)(dram_desc->dram_size),
+				  dram_desc->dram_base_addr_v,
+				  dram_desc->dram_bus_addr);
+	}
+
+	if (dram_desc)
+		memset(dram_desc, 0, sizeof(*dram_desc));
 }
 
 static void qat_uclo_ummap_auth_fw(struct icp_qat_fw_loader_handle *handle,
@@ -1209,12 +1354,14 @@ static void qat_uclo_ummap_auth_fw(struct icp_qat_fw_loader_handle *handle,
 {
 	struct icp_firml_dram_desc dram_desc;
 
-	dram_desc.dram_base_addr_v = *desc;
-	dram_desc.dram_bus_addr = ((struct icp_qat_auth_chunk *)
-				   (*desc))->chunk_bus_addr;
-	dram_desc.dram_size = ((struct icp_qat_auth_chunk *)
-			       (*desc))->chunk_size;
-	qat_uclo_simg_free(handle, &dram_desc);
+	if (*desc) {
+		dram_desc.dram_base_addr_v = *desc;
+		dram_desc.dram_bus_addr = ((struct icp_qat_auth_chunk *)
+					   (*desc))->chunk_bus_addr;
+		dram_desc.dram_size = ((struct icp_qat_auth_chunk *)
+				       (*desc))->chunk_size;
+		qat_uclo_simg_free(handle, &dram_desc);
+	}
 }
 
 static int qat_uclo_map_auth_fw(struct icp_qat_fw_loader_handle *handle,
@@ -1226,15 +1373,16 @@ static int qat_uclo_map_auth_fw(struct icp_qat_fw_loader_handle *handle,
 	struct icp_qat_auth_chunk *auth_chunk;
 	u64 virt_addr,  bus_addr, virt_base;
 	unsigned int length, simg_offset = sizeof(*auth_chunk);
+	struct icp_qat_simg_ae_mode *simg_ae_mode;
 	struct icp_firml_dram_desc img_desc;
 
-	if (size > (ICP_QAT_AE_IMG_OFFSET + ICP_QAT_CSS_MAX_IMAGE_LEN)) {
+	if (size > (ICP_QAT_AE_IMG_OFFSET(handle) + ICP_QAT_CSS_MAX_IMAGE_LEN)) {
 		pr_err("QAT: error, input image size overflow %d\n", size);
 		return -EINVAL;
 	}
 	length = (css_hdr->fw_type == CSS_AE_FIRMWARE) ?
-		 ICP_QAT_CSS_AE_SIMG_LEN + simg_offset :
-		 size + ICP_QAT_CSS_FWSK_PAD_LEN + simg_offset;
+		 ICP_QAT_CSS_AE_SIMG_LEN(handle) + simg_offset :
+		 size + ICP_QAT_CSS_FWSK_PAD_LEN(handle) + simg_offset;
 	if (qat_uclo_simg_alloc(handle, &img_desc, length)) {
 		pr_err("QAT: error, allocate continuous dram fail\n");
 		return -ENOMEM;
@@ -1261,42 +1409,42 @@ static int qat_uclo_map_auth_fw(struct icp_qat_fw_loader_handle *handle,
 
 	memcpy((void *)(uintptr_t)virt_addr,
 	       (void *)(image + sizeof(*css_hdr)),
-	       ICP_QAT_CSS_FWSK_MODULUS_LEN);
+	       ICP_QAT_CSS_FWSK_MODULUS_LEN(handle));
 	/* padding */
-	memset((void *)(uintptr_t)(virt_addr + ICP_QAT_CSS_FWSK_MODULUS_LEN),
-	       0, ICP_QAT_CSS_FWSK_PAD_LEN);
+	memset((void *)(uintptr_t)(virt_addr + ICP_QAT_CSS_FWSK_MODULUS_LEN(handle)),
+	       0, ICP_QAT_CSS_FWSK_PAD_LEN(handle));
 
 	/* exponent */
-	memcpy((void *)(uintptr_t)(virt_addr + ICP_QAT_CSS_FWSK_MODULUS_LEN +
-	       ICP_QAT_CSS_FWSK_PAD_LEN),
+	memcpy((void *)(uintptr_t)(virt_addr + ICP_QAT_CSS_FWSK_MODULUS_LEN(handle) +
+	       ICP_QAT_CSS_FWSK_PAD_LEN(handle)),
 	       (void *)(image + sizeof(*css_hdr) +
-			ICP_QAT_CSS_FWSK_MODULUS_LEN),
+			ICP_QAT_CSS_FWSK_MODULUS_LEN(handle)),
 	       sizeof(unsigned int));
 
 	/* signature */
 	bus_addr = ADD_ADDR(auth_desc->fwsk_pub_high,
 			    auth_desc->fwsk_pub_low) +
-		   ICP_QAT_CSS_FWSK_PUB_LEN;
-	virt_addr = virt_addr + ICP_QAT_CSS_FWSK_PUB_LEN;
+		   ICP_QAT_CSS_FWSK_PUB_LEN(handle);
+	virt_addr = virt_addr + ICP_QAT_CSS_FWSK_PUB_LEN(handle);
 	auth_desc->signature_high = (unsigned int)(bus_addr >> BITS_IN_DWORD);
 	auth_desc->signature_low = (unsigned int)bus_addr;
 
 	memcpy((void *)(uintptr_t)virt_addr,
 	       (void *)(image + sizeof(*css_hdr) +
-	       ICP_QAT_CSS_FWSK_MODULUS_LEN +
-	       ICP_QAT_CSS_FWSK_EXPONENT_LEN),
-	       ICP_QAT_CSS_SIGNATURE_LEN);
+	       ICP_QAT_CSS_FWSK_MODULUS_LEN(handle) +
+	       ICP_QAT_CSS_FWSK_EXPONENT_LEN(handle)),
+	       ICP_QAT_CSS_SIGNATURE_LEN(handle));
 
 	bus_addr = ADD_ADDR(auth_desc->signature_high,
 			    auth_desc->signature_low) +
-		   ICP_QAT_CSS_SIGNATURE_LEN;
-	virt_addr += ICP_QAT_CSS_SIGNATURE_LEN;
+		   ICP_QAT_CSS_SIGNATURE_LEN(handle);
+	virt_addr += ICP_QAT_CSS_SIGNATURE_LEN(handle);
 
 	auth_desc->img_high = (unsigned int)(bus_addr >> BITS_IN_DWORD);
 	auth_desc->img_low = (unsigned int)bus_addr;
-	auth_desc->img_len = size - ICP_QAT_AE_IMG_OFFSET;
+	auth_desc->img_len = size - ICP_QAT_AE_IMG_OFFSET(handle);
 	memcpy((void *)(uintptr_t)virt_addr,
-	       (void *)(image + ICP_QAT_AE_IMG_OFFSET),
+	       (void *)(image + ICP_QAT_AE_IMG_OFFSET(handle)),
 	       auth_desc->img_len);
 	virt_addr = virt_base;
 	/* AE firmware */
@@ -1315,6 +1463,11 @@ static int qat_uclo_map_auth_fw(struct icp_qat_fw_loader_handle *handle,
 		auth_desc->img_ae_insts_high = (unsigned int)
 					     (bus_addr >> BITS_IN_DWORD);
 		auth_desc->img_ae_insts_low = (unsigned int)bus_addr;
+		virt_addr += sizeof(struct icp_qat_css_hdr);
+		virt_addr += ICP_QAT_CSS_FWSK_PUB_LEN(handle);
+		virt_addr += ICP_QAT_CSS_SIGNATURE_LEN(handle);
+		simg_ae_mode = (struct icp_qat_simg_ae_mode *)(uintptr_t)virt_addr;
+		auth_desc->ae_mask = simg_ae_mode->ae_mask & handle->cfg_ae_mask;
 	} else {
 		auth_desc->img_ae_insts_high = auth_desc->img_high;
 		auth_desc->img_ae_insts_low = auth_desc->img_low;
@@ -1326,35 +1479,40 @@ static int qat_uclo_map_auth_fw(struct icp_qat_fw_loader_handle *handle,
 static int qat_uclo_load_fw(struct icp_qat_fw_loader_handle *handle,
 			    struct icp_qat_fw_auth_desc *desc)
 {
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	u32 fcu_sts_csr, fcu_ctl_csr;
+	u32 loaded_aes, loaded_csr;
 	unsigned int i;
-	unsigned int fcu_sts;
-	struct icp_qat_simg_ae_mode *virt_addr;
-	unsigned int fcu_loaded_ae_pos = FCU_LOADED_AE_POS;
-
-	virt_addr = (void *)((uintptr_t)desc +
-		     sizeof(struct icp_qat_auth_chunk) +
-		     sizeof(struct icp_qat_css_hdr) +
-		     ICP_QAT_CSS_FWSK_PUB_LEN +
-		     ICP_QAT_CSS_SIGNATURE_LEN);
-	for (i = 0; i < handle->hal_handle->ae_max_num; i++) {
+	u32 fcu_sts;
+
+	fcu_ctl_csr = handle->chip_info->fcu_ctl_csr;
+	fcu_sts_csr = handle->chip_info->fcu_sts_csr;
+	loaded_csr = handle->chip_info->fcu_loaded_ae_csr;
+
+	for_each_set_bit(i, &ae_mask, handle->hal_handle->ae_max_num) {
 		int retry = 0;
 
-		if (!((virt_addr->ae_mask >> i) & 0x1))
+		if (!((desc->ae_mask >> i) & 0x1))
 			continue;
 		if (qat_hal_check_ae_active(handle, i)) {
 			pr_err("QAT: AE %d is active\n", i);
 			return -EINVAL;
 		}
-		SET_CAP_CSR(handle, FCU_CONTROL,
-			    (FCU_CTRL_CMD_LOAD | (i << FCU_CTRL_AE_POS)));
+		SET_CAP_CSR(handle, fcu_ctl_csr,
+			    (FCU_CTRL_CMD_LOAD |
+			    (1 << FCU_CTRL_BROADCAST_POS) |
+			    (i << FCU_CTRL_AE_POS)));
 
 		do {
 			msleep(FW_AUTH_WAIT_PERIOD);
-			fcu_sts = GET_CAP_CSR(handle, FCU_STATUS);
-			if (((fcu_sts & FCU_AUTH_STS_MASK) ==
-			    FCU_STS_LOAD_DONE) &&
-			    ((fcu_sts >> fcu_loaded_ae_pos) & (1 << i)))
-				break;
+			fcu_sts = GET_CAP_CSR(handle, fcu_sts_csr);
+			if ((fcu_sts & FCU_AUTH_STS_MASK) ==
+			    FCU_STS_LOAD_DONE) {
+				loaded_aes = GET_CAP_CSR(handle, loaded_csr);
+				loaded_aes >>= handle->chip_info->fcu_loaded_ae_pos;
+				if (loaded_aes & (1 << i))
+					break;
+			}
 		} while (retry++ < FW_AUTH_MAX_RETRY);
 		if (retry > FW_AUTH_MAX_RETRY) {
 			pr_err("QAT: firmware load failed timeout %x\n", retry);
@@ -1387,14 +1545,16 @@ int qat_uclo_wr_mimage(struct icp_qat_fw_loader_handle *handle,
 	struct icp_qat_fw_auth_desc *desc = NULL;
 	int status = 0;
 
-	if (handle->fw_auth) {
+	if (handle->chip_info->fw_auth) {
 		if (!qat_uclo_map_auth_fw(handle, addr_ptr, mem_size, &desc))
 			status = qat_uclo_auth_fw(handle, desc);
 		qat_uclo_ummap_auth_fw(handle, &desc);
 	} else {
-		if (handle->pci_dev->device == PCI_DEVICE_ID_INTEL_QAT_C3XXX) {
-			pr_err("QAT: C3XXX doesn't support unsigned MMP\n");
-			return -EINVAL;
+		if (!handle->chip_info->sram_visible) {
+			dev_dbg(&handle->pci_dev->dev,
+				"QAT MMP fw not loaded for device 0x%x",
+				handle->pci_dev->device);
+			return status;
 		}
 		qat_uclo_wr_sram_by_words(handle, 0, addr_ptr, mem_size);
 	}
@@ -1437,25 +1597,281 @@ out_objbuf_err:
 	return -ENOMEM;
 }
 
+static int qat_uclo_map_mof_file_hdr(struct icp_qat_fw_loader_handle *handle,
+				     struct icp_qat_mof_file_hdr *mof_ptr,
+				     u32 mof_size)
+{
+	struct icp_qat_mof_handle *mobj_handle = handle->mobj_handle;
+	unsigned int min_ver_offset;
+	unsigned int checksum;
+
+	mobj_handle->file_id = ICP_QAT_MOF_FID;
+	mobj_handle->mof_buf = (char *)mof_ptr;
+	mobj_handle->mof_size = mof_size;
+
+	min_ver_offset = mof_size - offsetof(struct icp_qat_mof_file_hdr,
+					     min_ver);
+	checksum = qat_uclo_calc_str_checksum(&mof_ptr->min_ver,
+					      min_ver_offset);
+	if (checksum != mof_ptr->checksum) {
+		pr_err("QAT: incorrect MOF checksum\n");
+		return -EINVAL;
+	}
+
+	mobj_handle->checksum = mof_ptr->checksum;
+	mobj_handle->min_ver = mof_ptr->min_ver;
+	mobj_handle->maj_ver = mof_ptr->maj_ver;
+	return 0;
+}
+
+static void qat_uclo_del_mof(struct icp_qat_fw_loader_handle *handle)
+{
+	struct icp_qat_mof_handle *mobj_handle = handle->mobj_handle;
+
+	kfree(mobj_handle->obj_table.obj_hdr);
+	mobj_handle->obj_table.obj_hdr = NULL;
+	kfree(handle->mobj_handle);
+	handle->mobj_handle = NULL;
+}
+
+static int qat_uclo_seek_obj_inside_mof(struct icp_qat_mof_handle *mobj_handle,
+					char *obj_name, char **obj_ptr,
+					unsigned int *obj_size)
+{
+	struct icp_qat_mof_objhdr *obj_hdr = mobj_handle->obj_table.obj_hdr;
+	unsigned int i;
+
+	for (i = 0; i < mobj_handle->obj_table.num_objs; i++) {
+		if (!strncmp(obj_hdr[i].obj_name, obj_name,
+			     ICP_QAT_SUOF_OBJ_NAME_LEN)) {
+			*obj_ptr  = obj_hdr[i].obj_buf;
+			*obj_size = obj_hdr[i].obj_size;
+			return 0;
+		}
+	}
+
+	pr_err("QAT: object %s is not found inside MOF\n", obj_name);
+	return -EINVAL;
+}
+
+static int qat_uclo_map_obj_from_mof(struct icp_qat_mof_handle *mobj_handle,
+				     struct icp_qat_mof_objhdr *mobj_hdr,
+				     struct icp_qat_mof_obj_chunkhdr *obj_chunkhdr)
+{
+	u8 *obj;
+
+	if (!strncmp(obj_chunkhdr->chunk_id, ICP_QAT_UOF_IMAG,
+		     ICP_QAT_MOF_OBJ_CHUNKID_LEN)) {
+		obj = mobj_handle->uobjs_hdr + obj_chunkhdr->offset;
+	} else if (!strncmp(obj_chunkhdr->chunk_id, ICP_QAT_SUOF_IMAG,
+			    ICP_QAT_MOF_OBJ_CHUNKID_LEN)) {
+		obj = mobj_handle->sobjs_hdr + obj_chunkhdr->offset;
+	} else {
+		pr_err("QAT: unsupported chunk id\n");
+		return -EINVAL;
+	}
+	mobj_hdr->obj_buf = obj;
+	mobj_hdr->obj_size = (unsigned int)obj_chunkhdr->size;
+	mobj_hdr->obj_name = obj_chunkhdr->name + mobj_handle->sym_str;
+	return 0;
+}
+
+static int qat_uclo_map_objs_from_mof(struct icp_qat_mof_handle *mobj_handle)
+{
+	struct icp_qat_mof_obj_chunkhdr *uobj_chunkhdr;
+	struct icp_qat_mof_obj_chunkhdr *sobj_chunkhdr;
+	struct icp_qat_mof_obj_hdr *uobj_hdr;
+	struct icp_qat_mof_obj_hdr *sobj_hdr;
+	struct icp_qat_mof_objhdr *mobj_hdr;
+	unsigned int uobj_chunk_num = 0;
+	unsigned int sobj_chunk_num = 0;
+	unsigned int *valid_chunk;
+	int ret, i;
+
+	uobj_hdr = (struct icp_qat_mof_obj_hdr *)mobj_handle->uobjs_hdr;
+	sobj_hdr = (struct icp_qat_mof_obj_hdr *)mobj_handle->sobjs_hdr;
+	if (uobj_hdr)
+		uobj_chunk_num = uobj_hdr->num_chunks;
+	if (sobj_hdr)
+		sobj_chunk_num = sobj_hdr->num_chunks;
+
+	mobj_hdr = kzalloc((uobj_chunk_num + sobj_chunk_num) *
+			   sizeof(*mobj_hdr), GFP_KERNEL);
+	if (!mobj_hdr)
+		return -ENOMEM;
+
+	mobj_handle->obj_table.obj_hdr = mobj_hdr;
+	valid_chunk = &mobj_handle->obj_table.num_objs;
+	uobj_chunkhdr = (struct icp_qat_mof_obj_chunkhdr *)
+			 ((uintptr_t)uobj_hdr + sizeof(*uobj_hdr));
+	sobj_chunkhdr = (struct icp_qat_mof_obj_chunkhdr *)
+			((uintptr_t)sobj_hdr + sizeof(*sobj_hdr));
+
+	/* map uof objects */
+	for (i = 0; i < uobj_chunk_num; i++) {
+		ret = qat_uclo_map_obj_from_mof(mobj_handle,
+						&mobj_hdr[*valid_chunk],
+						&uobj_chunkhdr[i]);
+		if (ret)
+			return ret;
+		(*valid_chunk)++;
+	}
+
+	/* map suof objects */
+	for (i = 0; i < sobj_chunk_num; i++) {
+		ret = qat_uclo_map_obj_from_mof(mobj_handle,
+						&mobj_hdr[*valid_chunk],
+						&sobj_chunkhdr[i]);
+		if (ret)
+			return ret;
+		(*valid_chunk)++;
+	}
+
+	if ((uobj_chunk_num + sobj_chunk_num) != *valid_chunk) {
+		pr_err("QAT: inconsistent UOF/SUOF chunk amount\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static void qat_uclo_map_mof_symobjs(struct icp_qat_mof_handle *mobj_handle,
+				     struct icp_qat_mof_chunkhdr *mof_chunkhdr)
+{
+	char **sym_str = (char **)&mobj_handle->sym_str;
+	unsigned int *sym_size = &mobj_handle->sym_size;
+	struct icp_qat_mof_str_table *str_table_obj;
+
+	*sym_size = *(unsigned int *)(uintptr_t)
+		    (mof_chunkhdr->offset + mobj_handle->mof_buf);
+	*sym_str = (char *)(uintptr_t)
+		   (mobj_handle->mof_buf + mof_chunkhdr->offset +
+		    sizeof(str_table_obj->tab_len));
+}
+
+static void qat_uclo_map_mof_chunk(struct icp_qat_mof_handle *mobj_handle,
+				   struct icp_qat_mof_chunkhdr *mof_chunkhdr)
+{
+	char *chunk_id = mof_chunkhdr->chunk_id;
+
+	if (!strncmp(chunk_id, ICP_QAT_MOF_SYM_OBJS, ICP_QAT_MOF_OBJ_ID_LEN))
+		qat_uclo_map_mof_symobjs(mobj_handle, mof_chunkhdr);
+	else if (!strncmp(chunk_id, ICP_QAT_UOF_OBJS, ICP_QAT_MOF_OBJ_ID_LEN))
+		mobj_handle->uobjs_hdr = mobj_handle->mof_buf +
+					 mof_chunkhdr->offset;
+	else if (!strncmp(chunk_id, ICP_QAT_SUOF_OBJS, ICP_QAT_MOF_OBJ_ID_LEN))
+		mobj_handle->sobjs_hdr = mobj_handle->mof_buf +
+					 mof_chunkhdr->offset;
+}
+
+static int qat_uclo_check_mof_format(struct icp_qat_mof_file_hdr *mof_hdr)
+{
+	int maj = mof_hdr->maj_ver & 0xff;
+	int min = mof_hdr->min_ver & 0xff;
+
+	if (mof_hdr->file_id != ICP_QAT_MOF_FID) {
+		pr_err("QAT: invalid header 0x%x\n", mof_hdr->file_id);
+		return -EINVAL;
+	}
+
+	if (mof_hdr->num_chunks <= 0x1) {
+		pr_err("QAT: MOF chunk amount is incorrect\n");
+		return -EINVAL;
+	}
+	if (maj != ICP_QAT_MOF_MAJVER || min != ICP_QAT_MOF_MINVER) {
+		pr_err("QAT: bad MOF version, major 0x%x, minor 0x%x\n",
+		       maj, min);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int qat_uclo_map_mof_obj(struct icp_qat_fw_loader_handle *handle,
+				struct icp_qat_mof_file_hdr *mof_ptr,
+				u32 mof_size, char *obj_name, char **obj_ptr,
+				unsigned int *obj_size)
+{
+	struct icp_qat_mof_chunkhdr *mof_chunkhdr;
+	unsigned int file_id = mof_ptr->file_id;
+	struct icp_qat_mof_handle *mobj_handle;
+	unsigned short chunks_num;
+	unsigned int i;
+	int ret;
+
+	if (file_id == ICP_QAT_UOF_FID || file_id == ICP_QAT_SUOF_FID) {
+		if (obj_ptr)
+			*obj_ptr = (char *)mof_ptr;
+		if (obj_size)
+			*obj_size = mof_size;
+		return 0;
+	}
+	if (qat_uclo_check_mof_format(mof_ptr))
+		return -EINVAL;
+
+	mobj_handle = kzalloc(sizeof(*mobj_handle), GFP_KERNEL);
+	if (!mobj_handle)
+		return -ENOMEM;
+
+	handle->mobj_handle = mobj_handle;
+	ret = qat_uclo_map_mof_file_hdr(handle, mof_ptr, mof_size);
+	if (ret)
+		return ret;
+
+	mof_chunkhdr = (void *)mof_ptr + sizeof(*mof_ptr);
+	chunks_num = mof_ptr->num_chunks;
+
+	/* Parse MOF file chunks */
+	for (i = 0; i < chunks_num; i++)
+		qat_uclo_map_mof_chunk(mobj_handle, &mof_chunkhdr[i]);
+
+	/* All sym_objs uobjs and sobjs should be available */
+	if (!mobj_handle->sym_str ||
+	    (!mobj_handle->uobjs_hdr && !mobj_handle->sobjs_hdr))
+		return -EINVAL;
+
+	ret = qat_uclo_map_objs_from_mof(mobj_handle);
+	if (ret)
+		return ret;
+
+	/* Seek specified uof object in MOF */
+	return qat_uclo_seek_obj_inside_mof(mobj_handle, obj_name,
+					    obj_ptr, obj_size);
+}
+
 int qat_uclo_map_obj(struct icp_qat_fw_loader_handle *handle,
-		     void *addr_ptr, int mem_size)
+		     void *addr_ptr, u32 mem_size, char *obj_name)
 {
+	char *obj_addr;
+	u32 obj_size;
+	int ret;
+
 	BUILD_BUG_ON(ICP_QAT_UCLO_MAX_AE >=
 		     (sizeof(handle->hal_handle->ae_mask) * 8));
 
 	if (!handle || !addr_ptr || mem_size < 24)
 		return -EINVAL;
 
-	return (handle->fw_auth) ?
-			qat_uclo_map_suof_obj(handle, addr_ptr, mem_size) :
-			qat_uclo_map_uof_obj(handle, addr_ptr, mem_size);
+	if (obj_name) {
+		ret = qat_uclo_map_mof_obj(handle, addr_ptr, mem_size, obj_name,
+					   &obj_addr, &obj_size);
+		if (ret)
+			return ret;
+	} else {
+		obj_addr = addr_ptr;
+		obj_size = mem_size;
+	}
+
+	return (handle->chip_info->fw_auth) ?
+			qat_uclo_map_suof_obj(handle, obj_addr, obj_size) :
+			qat_uclo_map_uof_obj(handle, obj_addr, obj_size);
 }
 
-void qat_uclo_del_uof_obj(struct icp_qat_fw_loader_handle *handle)
+void qat_uclo_del_obj(struct icp_qat_fw_loader_handle *handle)
 {
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
 	unsigned int a;
 
+	if (handle->mobj_handle)
+		qat_uclo_del_mof(handle);
 	if (handle->sobj_handle)
 		qat_uclo_del_suof(handle);
 	if (!obj_handle)
@@ -1479,23 +1895,24 @@ static void qat_uclo_fill_uwords(struct icp_qat_uclo_objhandle *obj_handle,
 				 u64 *uword, unsigned int addr_p,
 				 unsigned int raddr, u64 fill)
 {
+	unsigned int i, addr;
 	u64 uwrd = 0;
-	unsigned int i;
 
 	if (!encap_page) {
 		*uword = fill;
 		return;
 	}
+	addr = (encap_page->page_region) ? raddr : addr_p;
 	for (i = 0; i < encap_page->uwblock_num; i++) {
-		if (raddr >= encap_page->uwblock[i].start_addr &&
-		    raddr <= encap_page->uwblock[i].start_addr +
+		if (addr >= encap_page->uwblock[i].start_addr &&
+		    addr <= encap_page->uwblock[i].start_addr +
 		    encap_page->uwblock[i].words_num - 1) {
-			raddr -= encap_page->uwblock[i].start_addr;
-			raddr *= obj_handle->uword_in_bytes;
+			addr -= encap_page->uwblock[i].start_addr;
+			addr *= obj_handle->uword_in_bytes;
 			memcpy(&uwrd, (void *)(((uintptr_t)
-			       encap_page->uwblock[i].micro_words) + raddr),
+			       encap_page->uwblock[i].micro_words) + addr),
 			       obj_handle->uword_in_bytes);
-			uwrd = uwrd & 0xbffffffffffull;
+			uwrd = uwrd & GENMASK_ULL(43, 0);
 		}
 	}
 	*uword = uwrd;
@@ -1546,6 +1963,10 @@ static void qat_uclo_wr_uimage_page(struct icp_qat_fw_loader_handle *handle,
 				    struct icp_qat_uof_image *image)
 {
 	struct icp_qat_uclo_objhandle *obj_handle = handle->obj_handle;
+	unsigned long ae_mask = handle->hal_handle->ae_mask;
+	unsigned long cfg_ae_mask = handle->cfg_ae_mask;
+	unsigned long ae_assigned = image->ae_assigned;
+	struct icp_qat_uclo_aedata *aed;
 	unsigned int ctx_mask, s;
 	struct icp_qat_uclo_page *page;
 	unsigned char ae;
@@ -1557,25 +1978,30 @@ static void qat_uclo_wr_uimage_page(struct icp_qat_fw_loader_handle *handle,
 		ctx_mask = 0x55;
 	/* load the default page and set assigned CTX PC
 	 * to the entrypoint address */
-	for (ae = 0; ae < handle->hal_handle->ae_max_num; ae++) {
-		if (!test_bit(ae, (unsigned long *)&image->ae_assigned))
+	for_each_set_bit(ae, &ae_mask, handle->hal_handle->ae_max_num) {
+		if (!test_bit(ae, &cfg_ae_mask))
 			continue;
+
+		if (!test_bit(ae, &ae_assigned))
+			continue;
+
+		aed = &obj_handle->ae_data[ae];
 		/* find the slice to which this image is assigned */
-		for (s = 0; s < obj_handle->ae_data[ae].slice_num; s++) {
-			if (image->ctx_assigned & obj_handle->ae_data[ae].
-			    ae_slices[s].ctx_mask_assigned)
+		for (s = 0; s < aed->slice_num; s++) {
+			if (image->ctx_assigned &
+			    aed->ae_slices[s].ctx_mask_assigned)
 				break;
 		}
-		if (s >= obj_handle->ae_data[ae].slice_num)
+		if (s >= aed->slice_num)
 			continue;
-		page = obj_handle->ae_data[ae].ae_slices[s].page;
+		page = aed->ae_slices[s].page;
 		if (!page->encap_page->def_page)
 			continue;
 		qat_uclo_wr_uimage_raw_page(handle, page->encap_page, ae);
 
-		page = obj_handle->ae_data[ae].ae_slices[s].page;
+		page = aed->ae_slices[s].page;
 		for (ctx = 0; ctx < ICP_QAT_UCLO_MAX_CTX; ctx++)
-			obj_handle->ae_data[ae].ae_slices[s].cur_page[ctx] =
+			aed->ae_slices[s].cur_page[ctx] =
 					(ctx_mask & (1 << ctx)) ? page : NULL;
 		qat_hal_set_live_ctx(handle, (unsigned char)ae,
 				     image->ctx_assigned);
@@ -1595,13 +2021,18 @@ static int qat_uclo_wr_suof_img(struct icp_qat_fw_loader_handle *handle)
 		if (qat_uclo_map_auth_fw(handle,
 					 (char *)simg_hdr[i].simg_buf,
 					 (unsigned int)
-					 (simg_hdr[i].simg_len),
+					 simg_hdr[i].simg_len,
 					 &desc))
 			goto wr_err;
 		if (qat_uclo_auth_fw(handle, desc))
 			goto wr_err;
-		if (qat_uclo_load_fw(handle, desc))
-			goto wr_err;
+		if (qat_uclo_is_broadcast(handle, i)) {
+			if (qat_uclo_broadcast_load_fw(handle, desc))
+				goto wr_err;
+		} else {
+			if (qat_uclo_load_fw(handle, desc))
+				goto wr_err;
+		}
 		qat_uclo_ummap_auth_fw(handle, &desc);
 	}
 	return 0;
@@ -1630,6 +2061,16 @@ static int qat_uclo_wr_uof_img(struct icp_qat_fw_loader_handle *handle)
 
 int qat_uclo_wr_all_uimage(struct icp_qat_fw_loader_handle *handle)
 {
-	return (handle->fw_auth) ? qat_uclo_wr_suof_img(handle) :
+	return (handle->chip_info->fw_auth) ? qat_uclo_wr_suof_img(handle) :
 				   qat_uclo_wr_uof_img(handle);
 }
+
+int qat_uclo_set_cfg_ae_mask(struct icp_qat_fw_loader_handle *handle,
+			     unsigned int cfg_ae_mask)
+{
+	if (!cfg_ae_mask)
+		return -EINVAL;
+
+	handle->cfg_ae_mask = cfg_ae_mask;
+	return 0;
+}
diff --git a/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.c b/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.c
index b975c263446d..1e83d9397b11 100644
--- a/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.c
+++ b/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.c
@@ -3,7 +3,9 @@
 #include <adf_accel_devices.h>
 #include <adf_pf2vf_msg.h>
 #include <adf_common_drv.h>
+#include <adf_gen2_hw_data.h>
 #include "adf_dh895xcc_hw_data.h"
+#include "icp_qat_hw.h"
 
 /* Worker thread to service arbiter mappings based on dev SKUs */
 static const u32 thrd_to_arb_map_sku4[] = {
@@ -24,15 +26,19 @@ static struct adf_hw_device_class dh895xcc_class = {
 	.instances = 0
 };
 
-static u32 get_accel_mask(u32 fuse)
+static u32 get_accel_mask(struct adf_hw_device_data *self)
 {
-	return (~fuse) >> ADF_DH895XCC_ACCELERATORS_REG_OFFSET &
-			  ADF_DH895XCC_ACCELERATORS_MASK;
+	u32 fuses = self->fuses;
+
+	return ~fuses >> ADF_DH895XCC_ACCELERATORS_REG_OFFSET &
+			 ADF_DH895XCC_ACCELERATORS_MASK;
 }
 
-static u32 get_ae_mask(u32 fuse)
+static u32 get_ae_mask(struct adf_hw_device_data *self)
 {
-	return (~fuse) & ADF_DH895XCC_ACCELENGINES_MASK;
+	u32 fuses = self->fuses;
+
+	return ~fuses & ADF_DH895XCC_ACCELENGINES_MASK;
 }
 
 static u32 get_num_accels(struct adf_hw_device_data *self)
@@ -78,6 +84,29 @@ static u32 get_sram_bar_id(struct adf_hw_device_data *self)
 	return ADF_DH895XCC_SRAM_BAR;
 }
 
+static u32 get_accel_cap(struct adf_accel_dev *accel_dev)
+{
+	struct pci_dev *pdev = accel_dev->accel_pci_dev.pci_dev;
+	u32 capabilities;
+	u32 legfuses;
+
+	capabilities = ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC |
+		       ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC |
+		       ICP_ACCEL_CAPABILITIES_AUTHENTICATION;
+
+	/* Read accelerator capabilities mask */
+	pci_read_config_dword(pdev, ADF_DEVICE_LEGFUSE_OFFSET, &legfuses);
+
+	if (legfuses & ICP_ACCEL_MASK_CIPHER_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_SYMMETRIC;
+	if (legfuses & ICP_ACCEL_MASK_PKE_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_CRYPTO_ASYMMETRIC;
+	if (legfuses & ICP_ACCEL_MASK_AUTH_SLICE)
+		capabilities &= ~ICP_ACCEL_CAPABILITIES_AUTHENTICATION;
+
+	return capabilities;
+}
+
 static enum dev_sku_info get_sku(struct adf_hw_device_data *self)
 {
 	int sku = (self->fuses & ADF_DH895XCC_FUSECTL_SKU_MASK)
@@ -131,11 +160,13 @@ static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
 {
 	struct adf_hw_device_data *hw_device = accel_dev->hw_device;
 	struct adf_bar *misc_bar = &GET_BARS(accel_dev)[ADF_DH895XCC_PMISC_BAR];
+	unsigned long accel_mask = hw_device->accel_mask;
+	unsigned long ae_mask = hw_device->ae_mask;
 	void __iomem *csr = misc_bar->virt_addr;
 	unsigned int val, i;
 
 	/* Enable Accel Engine error detection & correction */
-	for (i = 0; i < hw_device->get_num_aes(hw_device); i++) {
+	for_each_set_bit(i, &ae_mask, GET_MAX_ACCELENGINES(accel_dev)) {
 		val = ADF_CSR_RD(csr, ADF_DH895XCC_AE_CTX_ENABLES(i));
 		val |= ADF_DH895XCC_ENABLE_AE_ECC_ERR;
 		ADF_CSR_WR(csr, ADF_DH895XCC_AE_CTX_ENABLES(i), val);
@@ -145,7 +176,7 @@ static void adf_enable_error_correction(struct adf_accel_dev *accel_dev)
 	}
 
 	/* Enable shared memory error detection & correction */
-	for (i = 0; i < hw_device->get_num_accels(hw_device); i++) {
+	for_each_set_bit(i, &accel_mask, ADF_DH895XCC_MAX_ACCELERATORS) {
 		val = ADF_CSR_RD(csr, ADF_DH895XCC_UERRSSMSH(i));
 		val |= ADF_DH895XCC_ERRSSMSH_EN;
 		ADF_CSR_WR(csr, ADF_DH895XCC_UERRSSMSH(i), val);
@@ -164,7 +195,7 @@ static void adf_enable_ints(struct adf_accel_dev *accel_dev)
 	/* Enable bundle and misc interrupts */
 	ADF_CSR_WR(addr, ADF_DH895XCC_SMIAPF0_MASK_OFFSET,
 		   accel_dev->pf.vf_info ? 0 :
-			GENMASK_ULL(GET_MAX_BANKS(accel_dev) - 1, 0));
+			BIT_ULL(GET_MAX_BANKS(accel_dev)) - 1);
 	ADF_CSR_WR(addr, ADF_DH895XCC_SMIAPF1_MASK_OFFSET,
 		   ADF_DH895XCC_SMIA1_MASK);
 }
@@ -174,11 +205,19 @@ static int adf_pf_enable_vf2pf_comms(struct adf_accel_dev *accel_dev)
 	return 0;
 }
 
+static void configure_iov_threads(struct adf_accel_dev *accel_dev, bool enable)
+{
+	adf_gen2_cfg_iov_thds(accel_dev, enable,
+			      ADF_DH895XCC_AE2FUNC_MAP_GRP_A_NUM_REGS,
+			      ADF_DH895XCC_AE2FUNC_MAP_GRP_B_NUM_REGS);
+}
+
 void adf_init_hw_data_dh895xcc(struct adf_hw_device_data *hw_data)
 {
 	hw_data->dev_class = &dh895xcc_class;
 	hw_data->instance_id = dh895xcc_class.instances++;
 	hw_data->num_banks = ADF_DH895XCC_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_ETR_MAX_RINGS_PER_BANK;
 	hw_data->num_accel = ADF_DH895XCC_MAX_ACCELERATORS;
 	hw_data->num_logical_accel = 1;
 	hw_data->num_engines = ADF_DH895XCC_MAX_ACCELENGINES;
@@ -189,18 +228,22 @@ void adf_init_hw_data_dh895xcc(struct adf_hw_device_data *hw_data)
 	hw_data->enable_error_correction = adf_enable_error_correction;
 	hw_data->get_accel_mask = get_accel_mask;
 	hw_data->get_ae_mask = get_ae_mask;
+	hw_data->get_accel_cap = get_accel_cap;
 	hw_data->get_num_accels = get_num_accels;
 	hw_data->get_num_aes = get_num_aes;
 	hw_data->get_etr_bar_id = get_etr_bar_id;
 	hw_data->get_misc_bar_id = get_misc_bar_id;
 	hw_data->get_pf2vf_offset = get_pf2vf_offset;
 	hw_data->get_vintmsk_offset = get_vintmsk_offset;
+	hw_data->get_admin_info = adf_gen2_get_admin_info;
+	hw_data->get_arb_info = adf_gen2_get_arb_info;
 	hw_data->get_sram_bar_id = get_sram_bar_id;
 	hw_data->get_sku = get_sku;
 	hw_data->fw_name = ADF_DH895XCC_FW;
 	hw_data->fw_mmp_name = ADF_DH895XCC_MMP;
 	hw_data->init_admin_comms = adf_init_admin_comms;
 	hw_data->exit_admin_comms = adf_exit_admin_comms;
+	hw_data->configure_iov_threads = configure_iov_threads;
 	hw_data->disable_iov = adf_disable_sriov;
 	hw_data->send_admin_init = adf_send_admin_init;
 	hw_data->init_arb = adf_init_arb;
@@ -210,6 +253,7 @@ void adf_init_hw_data_dh895xcc(struct adf_hw_device_data *hw_data)
 	hw_data->enable_vf2pf_comms = adf_pf_enable_vf2pf_comms;
 	hw_data->reset_device = adf_reset_sbr;
 	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
+	adf_gen2_init_hw_csr_ops(&hw_data->csr_ops);
 }
 
 void adf_clean_hw_data_dh895xcc(struct adf_hw_device_data *hw_data)
diff --git a/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.h b/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.h
index 082a04466dca..4d613923d155 100644
--- a/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.h
+++ b/drivers/crypto/qat/qat_dh895xcc/adf_dh895xcc_hw_data.h
@@ -36,6 +36,11 @@
 
 #define ADF_DH895XCC_PF2VF_OFFSET(i)	(0x3A000 + 0x280 + ((i) * 0x04))
 #define ADF_DH895XCC_VINTMSK_OFFSET(i)	(0x3A000 + 0x200 + ((i) * 0x04))
+
+/* AE to function mapping */
+#define ADF_DH895XCC_AE2FUNC_MAP_GRP_A_NUM_REGS 96
+#define ADF_DH895XCC_AE2FUNC_MAP_GRP_B_NUM_REGS 12
+
 /* FW names */
 #define ADF_DH895XCC_FW "qat_895xcc.bin"
 #define ADF_DH895XCC_MMP "qat_895xcc_mmp.bin"
diff --git a/drivers/crypto/qat/qat_dh895xcc/adf_drv.c b/drivers/crypto/qat/qat_dh895xcc/adf_drv.c
index ecb4f6f20e22..a9ec4357144c 100644
--- a/drivers/crypto/qat/qat_dh895xcc/adf_drv.c
+++ b/drivers/crypto/qat/qat_dh895xcc/adf_drv.c
@@ -128,8 +128,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 			      &hw_data->fuses);
 
 	/* Get Accelerators and Accelerators Engines masks */
-	hw_data->accel_mask = hw_data->get_accel_mask(hw_data->fuses);
-	hw_data->ae_mask = hw_data->get_ae_mask(hw_data->fuses);
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
 	accel_pci_dev->sku = hw_data->get_sku(hw_data);
 	/* If the device has no acceleration engines then ignore it. */
 	if (!hw_data->accel_mask || !hw_data->ae_mask ||
@@ -177,9 +177,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto out_err_disable;
 	}
 
-	/* Read accelerator capabilities mask */
-	pci_read_config_dword(pdev, ADF_DEVICE_LEGFUSE_OFFSET,
-			      &hw_data->accel_capabilities_mask);
+	/* Get accelerator capabilities mask */
+	hw_data->accel_capabilities_mask = hw_data->get_accel_cap(accel_dev);
 
 	/* Find and map all the device's BARS */
 	i = 0;
diff --git a/drivers/crypto/qat/qat_dh895xccvf/adf_dh895xccvf_hw_data.c b/drivers/crypto/qat/qat_dh895xccvf/adf_dh895xccvf_hw_data.c
index 5246f0524ca3..f14fb82ed6df 100644
--- a/drivers/crypto/qat/qat_dh895xccvf/adf_dh895xccvf_hw_data.c
+++ b/drivers/crypto/qat/qat_dh895xccvf/adf_dh895xccvf_hw_data.c
@@ -3,6 +3,7 @@
 #include <adf_accel_devices.h>
 #include <adf_pf2vf_msg.h>
 #include <adf_common_drv.h>
+#include <adf_gen2_hw_data.h>
 #include "adf_dh895xccvf_hw_data.h"
 
 static struct adf_hw_device_class dh895xcciov_class = {
@@ -11,12 +12,12 @@ static struct adf_hw_device_class dh895xcciov_class = {
 	.instances = 0
 };
 
-static u32 get_accel_mask(u32 fuse)
+static u32 get_accel_mask(struct adf_hw_device_data *self)
 {
 	return ADF_DH895XCCIOV_ACCELERATORS_MASK;
 }
 
-static u32 get_ae_mask(u32 fuse)
+static u32 get_ae_mask(struct adf_hw_device_data *self)
 {
 	return ADF_DH895XCCIOV_ACCELENGINES_MASK;
 }
@@ -69,6 +70,7 @@ void adf_init_hw_data_dh895xcciov(struct adf_hw_device_data *hw_data)
 {
 	hw_data->dev_class = &dh895xcciov_class;
 	hw_data->num_banks = ADF_DH895XCCIOV_ETR_MAX_BANKS;
+	hw_data->num_rings_per_bank = ADF_ETR_MAX_RINGS_PER_BANK;
 	hw_data->num_accel = ADF_DH895XCCIOV_MAX_ACCELERATORS;
 	hw_data->num_logical_accel = 1;
 	hw_data->num_engines = ADF_DH895XCCIOV_MAX_ACCELENGINES;
@@ -97,6 +99,7 @@ void adf_init_hw_data_dh895xcciov(struct adf_hw_device_data *hw_data)
 	hw_data->min_iov_compat_ver = ADF_PFVF_COMPATIBILITY_VERSION;
 	hw_data->dev_class->instances++;
 	adf_devmgr_update_class_index(hw_data);
+	adf_gen2_init_hw_csr_ops(&hw_data->csr_ops);
 }
 
 void adf_clean_hw_data_dh895xcciov(struct adf_hw_device_data *hw_data)
diff --git a/drivers/crypto/qat/qat_dh895xccvf/adf_drv.c b/drivers/crypto/qat/qat_dh895xccvf/adf_drv.c
index 404cf9df6922..c972554a755e 100644
--- a/drivers/crypto/qat/qat_dh895xccvf/adf_drv.c
+++ b/drivers/crypto/qat/qat_dh895xccvf/adf_drv.c
@@ -119,8 +119,8 @@ static int adf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	adf_init_hw_data_dh895xcciov(accel_dev->hw_device);
 
 	/* Get Accelerators and Accelerators Engines masks */
-	hw_data->accel_mask = hw_data->get_accel_mask(hw_data->fuses);
-	hw_data->ae_mask = hw_data->get_ae_mask(hw_data->fuses);
+	hw_data->accel_mask = hw_data->get_accel_mask(hw_data);
+	hw_data->ae_mask = hw_data->get_ae_mask(hw_data);
 	accel_pci_dev->sku = hw_data->get_sku(hw_data);
 
 	/* Create dev top level debugfs entry */
diff --git a/drivers/crypto/qce/common.c b/drivers/crypto/qce/common.c
index 5006e74c40cd..a73db2a5637f 100644
--- a/drivers/crypto/qce/common.c
+++ b/drivers/crypto/qce/common.c
@@ -7,7 +7,8 @@
 #include <linux/interrupt.h>
 #include <linux/types.h>
 #include <crypto/scatterwalk.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #include "cipher.h"
 #include "common.h"
diff --git a/drivers/crypto/qce/core.c b/drivers/crypto/qce/core.c
index ea616b7259ae..80b75085c265 100644
--- a/drivers/crypto/qce/core.c
+++ b/drivers/crypto/qce/core.c
@@ -13,7 +13,6 @@
 #include <linux/types.h>
 #include <crypto/algapi.h>
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
 
 #include "core.h"
 #include "cipher.h"
@@ -160,7 +159,21 @@ static int qce_check_version(struct qce_device *qce)
 		return -ENODEV;
 
 	qce->burst_size = QCE_BAM_BURST_SIZE;
-	qce->pipe_pair_id = 1;
+
+	/*
+	 * Rx and tx pipes are treated as a pair inside CE.
+	 * Pipe pair number depends on the actual BAM dma pipe
+	 * that is used for transfers. The BAM dma pipes are passed
+	 * from the device tree and used to derive the pipe pair
+	 * id in the CE driver as follows.
+	 * 	BAM dma pipes(rx, tx)		CE pipe pair id
+	 *		0,1				0
+	 *		2,3				1
+	 *		4,5				2
+	 *		6,7				3
+	 *		...
+	 */
+	qce->pipe_pair_id = qce->dma.rxchan->chan_id >> 1;
 
 	dev_dbg(qce->dev, "Crypto device found, version %d.%d.%d\n",
 		major, minor, step);
@@ -261,6 +274,7 @@ static int qce_crypto_remove(struct platform_device *pdev)
 
 static const struct of_device_id qce_crypto_of_match[] = {
 	{ .compatible = "qcom,crypto-v5.1", },
+	{ .compatible = "qcom,crypto-v5.4", },
 	{}
 };
 MODULE_DEVICE_TABLE(of, qce_crypto_of_match);
diff --git a/drivers/crypto/qce/sha.c b/drivers/crypto/qce/sha.c
index 87be96a0b0bb..61c418c12345 100644
--- a/drivers/crypto/qce/sha.c
+++ b/drivers/crypto/qce/sha.c
@@ -48,7 +48,7 @@ static void qce_ahash_done(void *data)
 	dma_unmap_sg(qce->dev, &rctx->result_sg, 1, DMA_FROM_DEVICE);
 
 	memcpy(rctx->digest, result->auth_iv, digestsize);
-	if (req->result)
+	if (req->result && rctx->last_blk)
 		memcpy(req->result, result->auth_iv, digestsize);
 
 	rctx->byte_count[0] = cpu_to_be32(result->auth_byte_count[0]);
diff --git a/drivers/crypto/qce/sha.h b/drivers/crypto/qce/sha.h
index d63526e3804d..a22695361f16 100644
--- a/drivers/crypto/qce/sha.h
+++ b/drivers/crypto/qce/sha.h
@@ -7,7 +7,8 @@
 #define _SHA_H_
 
 #include <crypto/scatterwalk.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #include "common.h"
 #include "core.h"
diff --git a/drivers/crypto/rockchip/rk3288_crypto.h b/drivers/crypto/rockchip/rk3288_crypto.h
index 3db595570c9c..97278c2574ff 100644
--- a/drivers/crypto/rockchip/rk3288_crypto.h
+++ b/drivers/crypto/rockchip/rk3288_crypto.h
@@ -12,7 +12,8 @@
 #include <crypto/internal/skcipher.h>
 
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #define _SBF(v, f)			((v) << (f))
 
diff --git a/drivers/crypto/s5p-sss.c b/drivers/crypto/s5p-sss.c
index 88a6c853ffd7..682c8a450a57 100644
--- a/drivers/crypto/s5p-sss.c
+++ b/drivers/crypto/s5p-sss.c
@@ -30,7 +30,8 @@
 
 #include <crypto/hash.h>
 #include <crypto/md5.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/internal/hash.h>
 
 #define _SBF(s, v)			((v) << (s))
diff --git a/drivers/crypto/sa2ul.c b/drivers/crypto/sa2ul.c
index eda93fab95fe..f300b0a5958a 100644
--- a/drivers/crypto/sa2ul.c
+++ b/drivers/crypto/sa2ul.c
@@ -9,8 +9,10 @@
  *		Tero Kristo
  */
 #include <linux/clk.h>
+#include <linux/dma-mapping.h>
 #include <linux/dmaengine.h>
 #include <linux/dmapool.h>
+#include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
@@ -23,7 +25,8 @@
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/scatterwalk.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #include "sa2ul.h"
 
@@ -361,42 +364,45 @@ static void sa_swiz_128(u8 *in, u16 len)
 }
 
 /* Prepare the ipad and opad from key as per SHA algorithm step 1*/
-static void prepare_kiopad(u8 *k_ipad, u8 *k_opad, const u8 *key, u16 key_sz)
+static void prepare_kipad(u8 *k_ipad, const u8 *key, u16 key_sz)
 {
 	int i;
 
-	for (i = 0; i < key_sz; i++) {
+	for (i = 0; i < key_sz; i++)
 		k_ipad[i] = key[i] ^ 0x36;
-		k_opad[i] = key[i] ^ 0x5c;
-	}
 
 	/* Instead of XOR with 0 */
-	for (; i < SHA1_BLOCK_SIZE; i++) {
+	for (; i < SHA1_BLOCK_SIZE; i++)
 		k_ipad[i] = 0x36;
+}
+
+static void prepare_kopad(u8 *k_opad, const u8 *key, u16 key_sz)
+{
+	int i;
+
+	for (i = 0; i < key_sz; i++)
+		k_opad[i] = key[i] ^ 0x5c;
+
+	/* Instead of XOR with 0 */
+	for (; i < SHA1_BLOCK_SIZE; i++)
 		k_opad[i] = 0x5c;
-	}
 }
 
-static void sa_export_shash(struct shash_desc *hash, int block_size,
+static void sa_export_shash(void *state, struct shash_desc *hash,
 			    int digest_size, __be32 *out)
 {
-	union {
-		struct sha1_state sha1;
-		struct sha256_state sha256;
-		struct sha512_state sha512;
-	} sha;
-	void *state;
+	struct sha1_state *sha1;
+	struct sha256_state *sha256;
 	u32 *result;
-	int i;
 
 	switch (digest_size) {
 	case SHA1_DIGEST_SIZE:
-		state = &sha.sha1;
-		result = sha.sha1.state;
+		sha1 = state;
+		result = sha1->state;
 		break;
 	case SHA256_DIGEST_SIZE:
-		state = &sha.sha256;
-		result = sha.sha256.state;
+		sha256 = state;
+		result = sha256->state;
 		break;
 	default:
 		dev_err(sa_k3_dev, "%s: bad digest_size=%d\n", __func__,
@@ -406,8 +412,7 @@ static void sa_export_shash(struct shash_desc *hash, int block_size,
 
 	crypto_shash_export(hash, state);
 
-	for (i = 0; i < digest_size >> 2; i++)
-		out[i] = cpu_to_be32(result[i]);
+	cpu_to_be32_array(out, result, digest_size / 4);
 }
 
 static void sa_prepare_iopads(struct algo_data *data, const u8 *key,
@@ -416,24 +421,28 @@ static void sa_prepare_iopads(struct algo_data *data, const u8 *key,
 	SHASH_DESC_ON_STACK(shash, data->ctx->shash);
 	int block_size = crypto_shash_blocksize(data->ctx->shash);
 	int digest_size = crypto_shash_digestsize(data->ctx->shash);
-	u8 k_ipad[SHA1_BLOCK_SIZE];
-	u8 k_opad[SHA1_BLOCK_SIZE];
+	union {
+		struct sha1_state sha1;
+		struct sha256_state sha256;
+		u8 k_pad[SHA1_BLOCK_SIZE];
+	} sha;
 
 	shash->tfm = data->ctx->shash;
 
-	prepare_kiopad(k_ipad, k_opad, key, key_sz);
-
-	memzero_explicit(ipad, block_size);
-	memzero_explicit(opad, block_size);
+	prepare_kipad(sha.k_pad, key, key_sz);
 
 	crypto_shash_init(shash);
-	crypto_shash_update(shash, k_ipad, block_size);
-	sa_export_shash(shash, block_size, digest_size, ipad);
+	crypto_shash_update(shash, sha.k_pad, block_size);
+	sa_export_shash(&sha, shash, digest_size, ipad);
+
+	prepare_kopad(sha.k_pad, key, key_sz);
 
 	crypto_shash_init(shash);
-	crypto_shash_update(shash, k_opad, block_size);
+	crypto_shash_update(shash, sha.k_pad, block_size);
 
-	sa_export_shash(shash, block_size, digest_size, opad);
+	sa_export_shash(&sha, shash, digest_size, opad);
+
+	memzero_explicit(&sha, sizeof(sha));
 }
 
 /* Derive the inverse key used in AES-CBC decryption operation */
@@ -506,7 +515,8 @@ static int sa_set_sc_enc(struct algo_data *ad, const u8 *key, u16 key_sz,
 static void sa_set_sc_auth(struct algo_data *ad, const u8 *key, u16 key_sz,
 			   u8 *sc_buf)
 {
-	__be32 ipad[64], opad[64];
+	__be32 *ipad = (void *)(sc_buf + 32);
+	__be32 *opad = (void *)(sc_buf + 64);
 
 	/* Set Authentication mode selector to hash processing */
 	sc_buf[0] = SA_HASH_PROCESSING;
@@ -515,14 +525,9 @@ static void sa_set_sc_auth(struct algo_data *ad, const u8 *key, u16 key_sz,
 	sc_buf[1] |= ad->auth_ctrl;
 
 	/* Copy the keys or ipad/opad */
-	if (ad->keyed_mac) {
+	if (ad->keyed_mac)
 		ad->prep_iopad(ad, key, key_sz, ipad, opad);
-
-		/* Copy ipad to AuthKey */
-		memcpy(&sc_buf[32], ipad, ad->hash_size);
-		/* Copy opad to Aux-1 */
-		memcpy(&sc_buf[64], opad, ad->hash_size);
-	} else {
+	else {
 		/* basic hash */
 		sc_buf[1] |= SA_BASIC_HASH;
 	}
@@ -819,7 +824,7 @@ static void sa_cipher_cra_exit(struct crypto_skcipher *tfm)
 	sa_free_ctx_info(&ctx->enc, data);
 	sa_free_ctx_info(&ctx->dec, data);
 
-	crypto_free_sync_skcipher(ctx->fallback.skcipher);
+	crypto_free_skcipher(ctx->fallback.skcipher);
 }
 
 static int sa_cipher_cra_init(struct crypto_skcipher *tfm)
@@ -827,6 +832,7 @@ static int sa_cipher_cra_init(struct crypto_skcipher *tfm)
 	struct sa_tfm_ctx *ctx = crypto_skcipher_ctx(tfm);
 	struct sa_crypto_data *data = dev_get_drvdata(sa_k3_dev);
 	const char *name = crypto_tfm_alg_name(&tfm->base);
+	struct crypto_skcipher *child;
 	int ret;
 
 	memzero_explicit(ctx, sizeof(*ctx));
@@ -841,14 +847,17 @@ static int sa_cipher_cra_init(struct crypto_skcipher *tfm)
 		return ret;
 	}
 
-	ctx->fallback.skcipher =
-		crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK);
+	child = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK);
 
-	if (IS_ERR(ctx->fallback.skcipher)) {
+	if (IS_ERR(child)) {
 		dev_err(sa_k3_dev, "Error allocating fallback algo %s\n", name);
-		return PTR_ERR(ctx->fallback.skcipher);
+		return PTR_ERR(child);
 	}
 
+	ctx->fallback.skcipher = child;
+	crypto_skcipher_set_reqsize(tfm, crypto_skcipher_reqsize(child) +
+					 sizeof(struct skcipher_request));
+
 	dev_dbg(sa_k3_dev, "%s(0x%p) sc-ids(0x%x(0x%pad), 0x%x(0x%pad))\n",
 		__func__, tfm, ctx->enc.sc_id, &ctx->enc.sc_phys,
 		ctx->dec.sc_id, &ctx->dec.sc_phys);
@@ -859,6 +868,7 @@ static int sa_cipher_setkey(struct crypto_skcipher *tfm, const u8 *key,
 			    unsigned int keylen, struct algo_data *ad)
 {
 	struct sa_tfm_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct crypto_skcipher *child = ctx->fallback.skcipher;
 	int cmdl_len;
 	struct sa_cmdl_cfg cfg;
 	int ret;
@@ -874,12 +884,10 @@ static int sa_cipher_setkey(struct crypto_skcipher *tfm, const u8 *key,
 	cfg.enc_eng_id = ad->enc_eng.eng_id;
 	cfg.iv_size = crypto_skcipher_ivsize(tfm);
 
-	crypto_sync_skcipher_clear_flags(ctx->fallback.skcipher,
+	crypto_skcipher_clear_flags(child, CRYPTO_TFM_REQ_MASK);
+	crypto_skcipher_set_flags(child, tfm->base.crt_flags &
 					 CRYPTO_TFM_REQ_MASK);
-	crypto_sync_skcipher_set_flags(ctx->fallback.skcipher,
-				       tfm->base.crt_flags &
-				       CRYPTO_TFM_REQ_MASK);
-	ret = crypto_sync_skcipher_setkey(ctx->fallback.skcipher, key, keylen);
+	ret = crypto_skcipher_setkey(child, key, keylen);
 	if (ret)
 		return ret;
 
@@ -1270,7 +1278,6 @@ static int sa_cipher_run(struct skcipher_request *req, u8 *iv, int enc)
 	    crypto_skcipher_ctx(crypto_skcipher_reqtfm(req));
 	struct crypto_alg *alg = req->base.tfm->__crt_alg;
 	struct sa_req sa_req = { 0 };
-	int ret;
 
 	if (!req->cryptlen)
 		return 0;
@@ -1282,20 +1289,18 @@ static int sa_cipher_run(struct skcipher_request *req, u8 *iv, int enc)
 	if (req->cryptlen > SA_MAX_DATA_SZ ||
 	    (req->cryptlen >= SA_UNSAFE_DATA_SZ_MIN &&
 	     req->cryptlen <= SA_UNSAFE_DATA_SZ_MAX)) {
-		SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback.skcipher);
+		struct skcipher_request *subreq = skcipher_request_ctx(req);
 
-		skcipher_request_set_sync_tfm(subreq, ctx->fallback.skcipher);
+		skcipher_request_set_tfm(subreq, ctx->fallback.skcipher);
 		skcipher_request_set_callback(subreq, req->base.flags,
-					      NULL, NULL);
+					      req->base.complete,
+					      req->base.data);
 		skcipher_request_set_crypt(subreq, req->src, req->dst,
 					   req->cryptlen, req->iv);
 		if (enc)
-			ret = crypto_skcipher_encrypt(subreq);
+			return crypto_skcipher_encrypt(subreq);
 		else
-			ret = crypto_skcipher_decrypt(subreq);
-
-		skcipher_request_zero(subreq);
-		return ret;
+			return crypto_skcipher_decrypt(subreq);
 	}
 
 	sa_req.size = req->cryptlen;
diff --git a/drivers/crypto/sa2ul.h b/drivers/crypto/sa2ul.h
index 7f7e3fe60d11..f597ddecde34 100644
--- a/drivers/crypto/sa2ul.h
+++ b/drivers/crypto/sa2ul.h
@@ -12,10 +12,9 @@
 #ifndef _K3_SA2UL_
 #define _K3_SA2UL_
 
-#include <linux/interrupt.h>
-#include <linux/skbuff.h>
-#include <linux/hw_random.h>
 #include <crypto/aes.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #define SA_ENGINE_ENABLE_CONTROL	0x1000
 
@@ -311,7 +310,7 @@ struct sa_tfm_ctx {
 	struct crypto_shash	*shash;
 	/* for fallback */
 	union {
-		struct crypto_sync_skcipher	*skcipher;
+		struct crypto_skcipher		*skcipher;
 		struct crypto_ahash		*ahash;
 		struct crypto_aead		*aead;
 	} fallback;
diff --git a/drivers/crypto/sahara.c b/drivers/crypto/sahara.c
index d60679c79822..8b5be29cb4dc 100644
--- a/drivers/crypto/sahara.c
+++ b/drivers/crypto/sahara.c
@@ -15,7 +15,8 @@
 #include <crypto/internal/hash.h>
 #include <crypto/internal/skcipher.h>
 #include <crypto/scatterwalk.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 
 #include <linux/clk.h>
 #include <linux/dma-mapping.h>
diff --git a/drivers/crypto/stm32/stm32-hash.c b/drivers/crypto/stm32/stm32-hash.c
index e3e25278a970..7ac0573ef663 100644
--- a/drivers/crypto/stm32/stm32-hash.c
+++ b/drivers/crypto/stm32/stm32-hash.c
@@ -25,7 +25,8 @@
 #include <crypto/hash.h>
 #include <crypto/md5.h>
 #include <crypto/scatterwalk.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/internal/hash.h>
 
 #define HASH_CR				0x00
diff --git a/drivers/crypto/talitos.c b/drivers/crypto/talitos.c
index 66773892f665..4fd85f31630a 100644
--- a/drivers/crypto/talitos.c
+++ b/drivers/crypto/talitos.c
@@ -31,7 +31,8 @@
 #include <crypto/algapi.h>
 #include <crypto/aes.h>
 #include <crypto/internal/des.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/md5.h>
 #include <crypto/internal/aead.h>
 #include <crypto/authenc.h>
@@ -460,7 +461,7 @@ DEF_TALITOS2_DONE(ch1_3, TALITOS2_ISR_CH_1_3_DONE)
 /*
  * locate current (offending) descriptor
  */
-static u32 current_desc_hdr(struct device *dev, int ch)
+static __be32 current_desc_hdr(struct device *dev, int ch)
 {
 	struct talitos_private *priv = dev_get_drvdata(dev);
 	int tail, iter;
@@ -478,7 +479,7 @@ static u32 current_desc_hdr(struct device *dev, int ch)
 
 	iter = tail;
 	while (priv->chan[ch].fifo[iter].dma_desc != cur_desc &&
-	       priv->chan[ch].fifo[iter].desc->next_desc != cur_desc) {
+	       priv->chan[ch].fifo[iter].desc->next_desc != cpu_to_be32(cur_desc)) {
 		iter = (iter + 1) & (priv->fifo_len - 1);
 		if (iter == tail) {
 			dev_err(dev, "couldn't locate current descriptor\n");
@@ -486,7 +487,7 @@ static u32 current_desc_hdr(struct device *dev, int ch)
 		}
 	}
 
-	if (priv->chan[ch].fifo[iter].desc->next_desc == cur_desc) {
+	if (priv->chan[ch].fifo[iter].desc->next_desc == cpu_to_be32(cur_desc)) {
 		struct talitos_edesc *edesc;
 
 		edesc = container_of(priv->chan[ch].fifo[iter].desc,
@@ -501,13 +502,13 @@ static u32 current_desc_hdr(struct device *dev, int ch)
 /*
  * user diagnostics; report root cause of error based on execution unit status
  */
-static void report_eu_error(struct device *dev, int ch, u32 desc_hdr)
+static void report_eu_error(struct device *dev, int ch, __be32 desc_hdr)
 {
 	struct talitos_private *priv = dev_get_drvdata(dev);
 	int i;
 
 	if (!desc_hdr)
-		desc_hdr = in_be32(priv->chan[ch].reg + TALITOS_DESCBUF);
+		desc_hdr = cpu_to_be32(in_be32(priv->chan[ch].reg + TALITOS_DESCBUF));
 
 	switch (desc_hdr & DESC_HDR_SEL0_MASK) {
 	case DESC_HDR_SEL0_AFEU:
diff --git a/drivers/crypto/ux500/hash/hash_core.c b/drivers/crypto/ux500/hash/hash_core.c
index 3d407eebb2ba..da284b0ea1b2 100644
--- a/drivers/crypto/ux500/hash/hash_core.c
+++ b/drivers/crypto/ux500/hash/hash_core.c
@@ -31,7 +31,8 @@
 #include <linux/bitops.h>
 
 #include <crypto/internal/hash.h>
-#include <crypto/sha.h>
+#include <crypto/sha1.h>
+#include <crypto/sha2.h>
 #include <crypto/scatterwalk.h>
 #include <crypto/algapi.h>