diff --git a/drivers/crypto/hisilicon/Kconfig b/drivers/crypto/hisilicon/Kconfig
index 257c64809b48d245ff8fa86a09afd2ddc5357408..5eee12f5f26c46516949b9fc0a1c4ccb8fbafb62 100644
--- a/drivers/crypto/hisilicon/Kconfig
+++ b/drivers/crypto/hisilicon/Kconfig
@@ -57,6 +57,7 @@ config CRYPTO_DEV_HISI_ZIP
 	depends on UACCE || UACCE=n
 	depends on ACPI
 	select CRYPTO_DEV_HISI_QM
+	select CRYPTO_DEFLATE
 	help
 	  Support for HiSilicon ZIP Driver
 
diff --git a/drivers/crypto/hisilicon/hpre/hpre.h b/drivers/crypto/hisilicon/hpre/hpre.h
index 0f3ddbadbcf9913f6b59803b1bd04708023e555f..021dbd9a1d48fb0816285a817177b72812e313fe 100644
--- a/drivers/crypto/hisilicon/hpre/hpre.h
+++ b/drivers/crypto/hisilicon/hpre/hpre.h
@@ -94,9 +94,8 @@ struct hpre_sqe {
 	__le64 key;
 	__le64 in;
 	__le64 out;
-	__le16 tag;
-	__le16 resv2;
-#define _HPRE_SQE_ALIGN_EXT	7
+	__le64 tag;
+#define _HPRE_SQE_ALIGN_EXT	6
 	__le32 rsvd1[_HPRE_SQE_ALIGN_EXT];
 };
 
diff --git a/drivers/crypto/hisilicon/hpre/hpre_crypto.c b/drivers/crypto/hisilicon/hpre/hpre_crypto.c
index fbd2b3510752527bb7423bad7ae4122184d4fbe1..8fd809ba041138bd13cf88c6bd6a13cf1707b453 100644
--- a/drivers/crypto/hisilicon/hpre/hpre_crypto.c
+++ b/drivers/crypto/hisilicon/hpre/hpre_crypto.c
@@ -116,18 +116,14 @@ struct hpre_curve25519_ctx {
 	/* gx coordinate */
 	unsigned char *g;
 	dma_addr_t dma_g;
-	struct crypto_kpp *soft_tfm;
 };
 
 struct hpre_ctx {
 	struct hisi_qp *qp;
 	struct device *dev;
-	struct hpre_asym_request **req_list;
 	struct hpre *hpre;
-	spinlock_t req_lock;
 	unsigned int key_sz;
 	bool crt_g2_mode;
-	struct idr req_idr;
 	union {
 		struct hpre_rsa_ctx rsa;
 		struct hpre_dh_ctx dh;
@@ -153,7 +149,6 @@ struct hpre_asym_request {
 		struct kpp_request *curve25519;
 	} areq;
 	int err;
-	int req_id;
 	hpre_cb cb;
 	struct timespec64 req_time;
 };
@@ -168,79 +163,13 @@ static inline unsigned int hpre_align_pd(void)
 	return (hpre_align_sz() - 1) & ~(crypto_tfm_ctx_alignment() - 1);
 }
 
-static int hpre_alloc_req_id(struct hpre_ctx *ctx)
+static void hpre_dfx_add_req_time(struct hpre_asym_request *hpre_req)
 {
-	unsigned long flags;
-	int id;
-
-	spin_lock_irqsave(&ctx->req_lock, flags);
-	id = idr_alloc(&ctx->req_idr, NULL, 0, ctx->qp->sq_depth, GFP_ATOMIC);
-	spin_unlock_irqrestore(&ctx->req_lock, flags);
-
-	return id;
-}
-
-static void hpre_free_req_id(struct hpre_ctx *ctx, int req_id)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&ctx->req_lock, flags);
-	idr_remove(&ctx->req_idr, req_id);
-	spin_unlock_irqrestore(&ctx->req_lock, flags);
-}
-
-static int hpre_add_req_to_ctx(struct hpre_asym_request *hpre_req)
-{
-	struct hpre_ctx *ctx;
-	struct hpre_dfx *dfx;
-	int id;
-
-	ctx = hpre_req->ctx;
-	id = hpre_alloc_req_id(ctx);
-	if (unlikely(id < 0))
-		return -EINVAL;
-
-	ctx->req_list[id] = hpre_req;
-	hpre_req->req_id = id;
+	struct hpre_ctx *ctx = hpre_req->ctx;
+	struct hpre_dfx *dfx = ctx->hpre->debug.dfx;
 
-	dfx = ctx->hpre->debug.dfx;
 	if (atomic64_read(&dfx[HPRE_OVERTIME_THRHLD].value))
 		ktime_get_ts64(&hpre_req->req_time);
-
-	return id;
-}
-
-static void hpre_rm_req_from_ctx(struct hpre_asym_request *hpre_req)
-{
-	struct hpre_ctx *ctx = hpre_req->ctx;
-	int id = hpre_req->req_id;
-
-	if (hpre_req->req_id >= 0) {
-		hpre_req->req_id = HPRE_INVLD_REQ_ID;
-		ctx->req_list[id] = NULL;
-		hpre_free_req_id(ctx, id);
-	}
-}
-
-static struct hisi_qp *hpre_get_qp_and_start(u8 type)
-{
-	struct hisi_qp *qp;
-	int ret;
-
-	qp = hpre_create_qp(type);
-	if (!qp) {
-		pr_info_ratelimited("Can not create hpre qp, alloc soft tfm!\n");
-		return ERR_PTR(-ENODEV);
-	}
-
-	ret = hisi_qm_start_qp(qp, 0);
-	if (ret < 0) {
-		hisi_qm_free_qps(&qp, 1);
-		pci_err(qp->qm->pdev, "Can not start qp!\n");
-		return ERR_PTR(-EINVAL);
-	}
-
-	return qp;
 }
 
 static int hpre_get_data_dma_addr(struct hpre_asym_request *hpre_req,
@@ -357,26 +286,19 @@ static void hpre_hw_data_clr_all(struct hpre_ctx *ctx,
 static int hpre_alg_res_post_hf(struct hpre_ctx *ctx, struct hpre_sqe *sqe,
 				void **kreq)
 {
-	struct hpre_asym_request *req;
 	unsigned int err, done, alg;
-	int id;
 
 #define HPRE_NO_HW_ERR		0
 #define HPRE_HW_TASK_DONE	3
 #define HREE_HW_ERR_MASK	GENMASK(10, 0)
 #define HREE_SQE_DONE_MASK	GENMASK(1, 0)
 #define HREE_ALG_TYPE_MASK	GENMASK(4, 0)
-	id = (int)le16_to_cpu(sqe->tag);
-	req = ctx->req_list[id];
-	hpre_rm_req_from_ctx(req);
-	*kreq = req;
+	*kreq = (void *)le64_to_cpu(sqe->tag);
 
 	err = (le32_to_cpu(sqe->dw0) >> HPRE_SQE_ALG_BITS) &
 		HREE_HW_ERR_MASK;
-
 	done = (le32_to_cpu(sqe->dw0) >> HPRE_SQE_DONE_SHIFT) &
 		HREE_SQE_DONE_MASK;
-
 	if (likely(err == HPRE_NO_HW_ERR && done == HPRE_HW_TASK_DONE))
 		return 0;
 
@@ -387,36 +309,10 @@ static int hpre_alg_res_post_hf(struct hpre_ctx *ctx, struct hpre_sqe *sqe,
 	return -EINVAL;
 }
 
-static int hpre_ctx_set(struct hpre_ctx *ctx, struct hisi_qp *qp, int qlen)
-{
-	struct hpre *hpre;
-
-	if (!ctx || !qp || qlen < 0)
-		return -EINVAL;
-
-	spin_lock_init(&ctx->req_lock);
-	ctx->qp = qp;
-	ctx->dev = &qp->qm->pdev->dev;
-
-	hpre = container_of(ctx->qp->qm, struct hpre, qm);
-	ctx->hpre = hpre;
-	ctx->req_list = kcalloc(qlen, sizeof(void *), GFP_KERNEL);
-	if (!ctx->req_list)
-		return -ENOMEM;
-	ctx->key_sz = 0;
-	ctx->crt_g2_mode = false;
-	idr_init(&ctx->req_idr);
-
-	return 0;
-}
-
 static void hpre_ctx_clear(struct hpre_ctx *ctx, bool is_clear_all)
 {
-	if (is_clear_all) {
-		idr_destroy(&ctx->req_idr);
-		kfree(ctx->req_list);
+	if (is_clear_all)
 		hisi_qm_free_qps(&ctx->qp, 1);
-	}
 
 	ctx->crt_g2_mode = false;
 	ctx->key_sz = 0;
@@ -484,51 +380,44 @@ static void hpre_rsa_cb(struct hpre_ctx *ctx, void *resp)
 
 static void hpre_alg_cb(struct hisi_qp *qp, void *resp)
 {
-	struct hpre_ctx *ctx = qp->qp_ctx;
-	struct hpre_dfx *dfx = ctx->hpre->debug.dfx;
+	struct hpre_asym_request *h_req;
 	struct hpre_sqe *sqe = resp;
-	struct hpre_asym_request *req = ctx->req_list[le16_to_cpu(sqe->tag)];
 
-	if (unlikely(!req)) {
-		atomic64_inc(&dfx[HPRE_INVALID_REQ_CNT].value);
+	h_req = (struct hpre_asym_request *)le64_to_cpu(sqe->tag);
+	if (unlikely(!h_req)) {
+		pr_err("Failed to get request, and qp_id is %u\n", qp->qp_id);
 		return;
 	}
 
-	req->cb(ctx, resp);
-}
-
-static void hpre_stop_qp_and_put(struct hisi_qp *qp)
-{
-	hisi_qm_stop_qp(qp);
-	hisi_qm_free_qps(&qp, 1);
+	h_req->cb(h_req->ctx, resp);
 }
 
 static int hpre_ctx_init(struct hpre_ctx *ctx, u8 type)
 {
 	struct hisi_qp *qp;
-	int ret;
+	struct hpre *hpre;
 
-	qp = hpre_get_qp_and_start(type);
-	if (IS_ERR(qp)) {
+	qp = hpre_create_qp(type);
+	if (!qp) {
 		ctx->qp = NULL;
 		return -ENODEV;
 	}
 
-	qp->qp_ctx = ctx;
 	qp->req_cb = hpre_alg_cb;
+	ctx->qp = qp;
+	ctx->dev = &qp->qm->pdev->dev;
+	hpre = container_of(ctx->qp->qm, struct hpre, qm);
+	ctx->hpre = hpre;
+	ctx->key_sz = 0;
+	ctx->crt_g2_mode = false;
 
-	ret = hpre_ctx_set(ctx, qp, qp->sq_depth);
-	if (ret)
-		hpre_stop_qp_and_put(qp);
-
-	return ret;
+	return 0;
 }
 
 static int hpre_msg_request_set(struct hpre_ctx *ctx, void *req, bool is_rsa)
 {
 	struct hpre_asym_request *h_req;
 	struct hpre_sqe *msg;
-	int req_id;
 	void *tmp;
 
 	if (is_rsa) {
@@ -568,11 +457,8 @@ static int hpre_msg_request_set(struct hpre_ctx *ctx, void *req, bool is_rsa)
 	msg->task_len1 = (ctx->key_sz >> HPRE_BITS_2_BYTES_SHIFT) - 1;
 	h_req->ctx = ctx;
 
-	req_id = hpre_add_req_to_ctx(h_req);
-	if (req_id < 0)
-		return -EBUSY;
-
-	msg->tag = cpu_to_le16((u16)req_id);
+	hpre_dfx_add_req_time(h_req);
+	msg->tag = cpu_to_le64((uintptr_t)h_req);
 
 	return 0;
 }
@@ -585,9 +471,7 @@ static int hpre_send(struct hpre_ctx *ctx, struct hpre_sqe *msg)
 
 	do {
 		atomic64_inc(&dfx[HPRE_SEND_CNT].value);
-		spin_lock_bh(&ctx->req_lock);
 		ret = hisi_qp_send(ctx->qp, msg);
-		spin_unlock_bh(&ctx->req_lock);
 		if (ret != -EBUSY)
 			break;
 		atomic64_inc(&dfx[HPRE_SEND_BUSY_CNT].value);
@@ -638,7 +522,6 @@ static int hpre_dh_compute_value(struct kpp_request *req)
 		return -EINPROGRESS;
 
 clear_all:
-	hpre_rm_req_from_ctx(hpre_req);
 	hpre_hw_data_clr_all(ctx, hpre_req, req->dst, req->src);
 
 	return ret;
@@ -747,9 +630,6 @@ static void hpre_dh_clear_ctx(struct hpre_ctx *ctx, bool is_clear_all)
 	if (!ctx->qp)
 		return;
 
-	if (is_clear_all)
-		hisi_qm_stop_qp(ctx->qp);
-
 	if (ctx->dh.g) {
 		dma_free_coherent(dev, sz, ctx->dh.g, ctx->dh.dma_g);
 		ctx->dh.g = NULL;
@@ -920,7 +800,6 @@ static int hpre_rsa_enc(struct akcipher_request *req)
 		return -EINPROGRESS;
 
 clear_all:
-	hpre_rm_req_from_ctx(hpre_req);
 	hpre_hw_data_clr_all(ctx, hpre_req, req->dst, req->src);
 
 	return ret;
@@ -974,7 +853,6 @@ static int hpre_rsa_dec(struct akcipher_request *req)
 		return -EINPROGRESS;
 
 clear_all:
-	hpre_rm_req_from_ctx(hpre_req);
 	hpre_hw_data_clr_all(ctx, hpre_req, req->dst, req->src);
 
 	return ret;
@@ -1131,9 +1009,6 @@ static void hpre_rsa_clear_ctx(struct hpre_ctx *ctx, bool is_clear_all)
 	if (!ctx->qp)
 		return;
 
-	if (is_clear_all)
-		hisi_qm_stop_qp(ctx->qp);
-
 	if (ctx->rsa.pubkey) {
 		dma_free_coherent(dev, ctx->key_sz << 1,
 				  ctx->rsa.pubkey, ctx->rsa.dma_pubkey);
@@ -1314,9 +1189,6 @@ static void hpre_ecc_clear_ctx(struct hpre_ctx *ctx, bool is_clear_all,
 	unsigned int sz = ctx->key_sz;
 	unsigned int shift = sz << 1;
 
-	if (is_clear_all)
-		hisi_qm_stop_qp(ctx->qp);
-
 	if (is_ecdh && ctx->ecdh.p) {
 		/* ecdh: p->a->k->b */
 		memzero_explicit(ctx->ecdh.p + shift, sz);
@@ -1459,7 +1331,7 @@ static int hpre_ecdh_set_param(struct hpre_ctx *ctx, struct ecdh *params)
 	return 0;
 }
 
-static bool hpre_key_is_zero(char *key, unsigned short key_sz)
+static bool hpre_key_is_zero(const char *key, unsigned short key_sz)
 {
 	int i;
 
@@ -1604,7 +1476,6 @@ static int hpre_ecdh_msg_request_set(struct hpre_ctx *ctx,
 {
 	struct hpre_asym_request *h_req;
 	struct hpre_sqe *msg;
-	int req_id;
 	void *tmp;
 
 	if (req->dst_len < ctx->key_sz << 1) {
@@ -1626,11 +1497,8 @@ static int hpre_ecdh_msg_request_set(struct hpre_ctx *ctx,
 	msg->task_len1 = (ctx->key_sz >> HPRE_BITS_2_BYTES_SHIFT) - 1;
 	h_req->ctx = ctx;
 
-	req_id = hpre_add_req_to_ctx(h_req);
-	if (req_id < 0)
-		return -EBUSY;
-
-	msg->tag = cpu_to_le16((u16)req_id);
+	hpre_dfx_add_req_time(h_req);
+	msg->tag = cpu_to_le64((uintptr_t)h_req);
 	return 0;
 }
 
@@ -1728,7 +1596,6 @@ static int hpre_ecdh_compute_value(struct kpp_request *req)
 		return -EINPROGRESS;
 
 clear_all:
-	hpre_rm_req_from_ctx(hpre_req);
 	hpre_ecdh_hw_data_clr_all(ctx, hpre_req, req->dst, req->src);
 	return ret;
 }
@@ -1904,9 +1771,6 @@ static int hpre_curve25519_set_secret(struct crypto_kpp *tfm, const void *buf,
 	struct device *dev = ctx->dev;
 	int ret = -EINVAL;
 
-	if (ctx->fallback)
-		return crypto_kpp_set_secret(ctx->curve25519.soft_tfm, buf, len);
-
 	if (len != CURVE25519_KEY_SIZE ||
 	    !crypto_memneq(buf, curve25519_null_point, CURVE25519_KEY_SIZE)) {
 		dev_err(dev, "key is null or key len is not 32bytes!\n");
@@ -1984,7 +1848,6 @@ static int hpre_curve25519_msg_request_set(struct hpre_ctx *ctx,
 {
 	struct hpre_asym_request *h_req;
 	struct hpre_sqe *msg;
-	int req_id;
 	void *tmp;
 
 	if (unlikely(req->dst_len < ctx->key_sz)) {
@@ -2006,11 +1869,8 @@ static int hpre_curve25519_msg_request_set(struct hpre_ctx *ctx,
 	msg->task_len1 = (ctx->key_sz >> HPRE_BITS_2_BYTES_SHIFT) - 1;
 	h_req->ctx = ctx;
 
-	req_id = hpre_add_req_to_ctx(h_req);
-	if (req_id < 0)
-		return -EBUSY;
-
-	msg->tag = cpu_to_le16((u16)req_id);
+	hpre_dfx_add_req_time(h_req);
+	msg->tag = cpu_to_le64((uintptr_t)h_req);
 	return 0;
 }
 
@@ -2147,88 +2007,30 @@ static int hpre_curve25519_compute_value(struct kpp_request *req)
 		return -EINPROGRESS;
 
 clear_all:
-	hpre_rm_req_from_ctx(hpre_req);
 	hpre_curve25519_hw_data_clr_all(ctx, hpre_req, req->dst, req->src);
 	return ret;
 }
 
-static int hpre_curve25519_generate_public_key(struct kpp_request *req)
-{
-	struct crypto_kpp *tfm = crypto_kpp_reqtfm(req);
-	struct hpre_ctx *ctx = kpp_tfm_ctx(tfm);
-	int ret;
-
-	if (ctx->fallback) {
-		kpp_request_set_tfm(req, ctx->curve25519.soft_tfm);
-		ret = crypto_kpp_generate_public_key(req);
-		kpp_request_set_tfm(req, tfm);
-		return ret;
-	}
-
-	return hpre_curve25519_compute_value(req);
-}
-
-static int hpre_curve25519_compute_shared_secret(struct kpp_request *req)
-{
-	struct crypto_kpp *tfm = crypto_kpp_reqtfm(req);
-	struct hpre_ctx *ctx = kpp_tfm_ctx(tfm);
-	int ret;
-
-	if (ctx->fallback) {
-		kpp_request_set_tfm(req, ctx->curve25519.soft_tfm);
-		ret = crypto_kpp_compute_shared_secret(req);
-		kpp_request_set_tfm(req, tfm);
-		return ret;
-	}
-
-	return hpre_curve25519_compute_value(req);
-}
-
 static unsigned int hpre_curve25519_max_size(struct crypto_kpp *tfm)
 {
 	struct hpre_ctx *ctx = kpp_tfm_ctx(tfm);
 
-	if (ctx->fallback)
-		return crypto_kpp_maxsize(ctx->curve25519.soft_tfm);
-
 	return ctx->key_sz;
 }
 
 static int hpre_curve25519_init_tfm(struct crypto_kpp *tfm)
 {
 	struct hpre_ctx *ctx = kpp_tfm_ctx(tfm);
-	const char *alg = kpp_alg_name(tfm);
-	int ret;
 
-	ret = hpre_ctx_init(ctx, HPRE_V3_ECC_ALG_TYPE);
-	if (!ret) {
-		kpp_set_reqsize(tfm, sizeof(struct hpre_asym_request) + hpre_align_pd());
-		return 0;
-	} else if (ret && ret != -ENODEV) {
-		return ret;
-	}
-
-	ctx->curve25519.soft_tfm = crypto_alloc_kpp(alg, 0, CRYPTO_ALG_NEED_FALLBACK);
-	if (IS_ERR(ctx->curve25519.soft_tfm)) {
-		pr_err("Failed to alloc curve25519 tfm!\n");
-		return PTR_ERR(ctx->curve25519.soft_tfm);
-	}
+	kpp_set_reqsize(tfm, sizeof(struct hpre_asym_request) + hpre_align_pd());
 
-	crypto_kpp_set_flags(ctx->curve25519.soft_tfm, crypto_kpp_get_flags(tfm));
-	ctx->fallback = true;
-
-	return 0;
+	return hpre_ctx_init(ctx, HPRE_V3_ECC_ALG_TYPE);
 }
 
 static void hpre_curve25519_exit_tfm(struct crypto_kpp *tfm)
 {
 	struct hpre_ctx *ctx = kpp_tfm_ctx(tfm);
 
-	if (ctx->fallback) {
-		crypto_free_kpp(ctx->curve25519.soft_tfm);
-		return;
-	}
-
 	hpre_ecc_clear_ctx(ctx, true, false);
 }
 
@@ -2320,8 +2122,8 @@ static struct kpp_alg ecdh_curves[] = {
 
 static struct kpp_alg curve25519_alg = {
 	.set_secret = hpre_curve25519_set_secret,
-	.generate_public_key = hpre_curve25519_generate_public_key,
-	.compute_shared_secret = hpre_curve25519_compute_shared_secret,
+	.generate_public_key = hpre_curve25519_compute_value,
+	.compute_shared_secret = hpre_curve25519_compute_value,
 	.max_size = hpre_curve25519_max_size,
 	.init = hpre_curve25519_init_tfm,
 	.exit = hpre_curve25519_exit_tfm,
@@ -2331,7 +2133,6 @@ static struct kpp_alg curve25519_alg = {
 		.cra_name = "curve25519",
 		.cra_driver_name = "hpre-curve25519",
 		.cra_module = THIS_MODULE,
-		.cra_flags = CRYPTO_ALG_NEED_FALLBACK,
 	},
 };
 
diff --git a/drivers/crypto/hisilicon/hpre/hpre_main.c b/drivers/crypto/hisilicon/hpre/hpre_main.c
index 1bd362d4704e8216f09fc36cfd7427e8220592c8..430031c442e143e3767a66467c7f2d4dbac9275f 100644
--- a/drivers/crypto/hisilicon/hpre/hpre_main.c
+++ b/drivers/crypto/hisilicon/hpre/hpre_main.c
@@ -464,7 +464,7 @@ struct hisi_qp *hpre_create_qp(u8 type)
 	 * type: 0 - RSA/DH. algorithm supported in V2,
 	 *       1 - ECC algorithm in V3.
 	 */
-	ret = hisi_qm_alloc_qps_node(&hpre_devices, 1, type, node, &qp);
+	ret = hisi_qm_alloc_qps_node(&hpre_devices, 1, &type, node, &qp);
 	if (!ret)
 		return qp;
 
diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c
index 4c9c2913f1e41e07236c7f60366892565500f3f1..8727e1cb6a866e8115610c6c0538e8d43a3ba762 100644
--- a/drivers/crypto/hisilicon/qm.c
+++ b/drivers/crypto/hisilicon/qm.c
@@ -2106,7 +2106,38 @@ static void hisi_qm_unset_hw_reset(struct hisi_qp *qp)
 	*addr = 0;
 }
 
-static struct hisi_qp *qm_create_qp_nolock(struct hisi_qm *qm, u8 alg_type)
+static struct hisi_qp *find_shareable_qp(struct hisi_qm *qm, u8 alg_type, bool is_in_kernel)
+{
+	struct device *dev = &qm->pdev->dev;
+	struct hisi_qp *share_qp = NULL;
+	struct hisi_qp *qp;
+	u32 ref_count = ~0;
+	int i;
+
+	if (!is_in_kernel)
+		goto queues_busy;
+
+	for (i = 0; i < qm->qp_num; i++) {
+		qp = &qm->qp_array[i];
+		if (qp->is_in_kernel && qp->alg_type == alg_type && qp->ref_count < ref_count) {
+			ref_count = qp->ref_count;
+			share_qp = qp;
+		}
+	}
+
+	if (share_qp) {
+		share_qp->ref_count++;
+		return share_qp;
+	}
+
+queues_busy:
+	dev_info_ratelimited(dev, "All %u queues of QM are busy and no shareable queue\n",
+			     qm->qp_num);
+	atomic64_inc(&qm->debug.dfx.create_qp_err_cnt);
+	return ERR_PTR(-EBUSY);
+}
+
+static struct hisi_qp *qm_create_qp_nolock(struct hisi_qm *qm, u8 alg_type, bool is_in_kernel)
 {
 	struct device *dev = &qm->pdev->dev;
 	struct hisi_qp *qp;
@@ -2117,12 +2148,9 @@ static struct hisi_qp *qm_create_qp_nolock(struct hisi_qm *qm, u8 alg_type)
 		return ERR_PTR(-EPERM);
 	}
 
-	if (qm->qp_in_used == qm->qp_num) {
-		dev_info_ratelimited(dev, "All %u queues of QM are busy!\n",
-				     qm->qp_num);
-		atomic64_inc(&qm->debug.dfx.create_qp_err_cnt);
-		return ERR_PTR(-EBUSY);
-	}
+	/* Try to find a shareable queue when all queues are busy */
+	if (qm->qp_in_used == qm->qp_num)
+		return find_shareable_qp(qm, alg_type, is_in_kernel);
 
 	qp_id = idr_alloc_cyclic(&qm->qp_idr, NULL, 0, qm->qp_num, GFP_ATOMIC);
 	if (qp_id < 0) {
@@ -2138,10 +2166,10 @@ static struct hisi_qp *qm_create_qp_nolock(struct hisi_qm *qm, u8 alg_type)
 
 	qp->event_cb = NULL;
 	qp->req_cb = NULL;
-	qp->qp_id = qp_id;
 	qp->alg_type = alg_type;
-	qp->is_in_kernel = true;
+	qp->is_in_kernel = is_in_kernel;
 	qm->qp_in_used++;
+	qp->ref_count = 1;
 
 	return qp;
 }
@@ -2163,7 +2191,7 @@ static struct hisi_qp *hisi_qm_create_qp(struct hisi_qm *qm, u8 alg_type)
 		return ERR_PTR(ret);
 
 	down_write(&qm->qps_lock);
-	qp = qm_create_qp_nolock(qm, alg_type);
+	qp = qm_create_qp_nolock(qm, alg_type, false);
 	up_write(&qm->qps_lock);
 
 	if (IS_ERR(qp))
@@ -2323,6 +2351,7 @@ static void qp_stop_fail_cb(struct hisi_qp *qp)
 	for (i = 0; i < qp_used; i++) {
 		pos = (i + cur_head) % sq_depth;
 		qp->req_cb(qp, qp->sqe + (u32)(qm->sqe_size * pos));
+		qm_cq_head_update(qp);
 		atomic_dec(&qp->qp_status.used);
 	}
 }
@@ -2472,25 +2501,33 @@ EXPORT_SYMBOL_GPL(hisi_qm_stop_qp);
 int hisi_qp_send(struct hisi_qp *qp, const void *msg)
 {
 	struct hisi_qp_status *qp_status = &qp->qp_status;
-	u16 sq_tail = qp_status->sq_tail;
-	u16 sq_tail_next = (sq_tail + 1) % qp->sq_depth;
-	void *sqe = qm_get_avail_sqe(qp);
+	u16 sq_tail, sq_tail_next;
+	void *sqe;
 
+	spin_lock_bh(&qp->qp_lock);
 	if (unlikely(atomic_read(&qp->qp_status.flags) == QP_STOP ||
 		     atomic_read(&qp->qm->status.flags) == QM_STOP ||
 		     qp->is_resetting)) {
+		spin_unlock_bh(&qp->qp_lock);
 		dev_info_ratelimited(&qp->qm->pdev->dev, "QP is stopped or resetting\n");
 		return -EAGAIN;
 	}
 
-	if (!sqe)
+	sqe = qm_get_avail_sqe(qp);
+	if (!sqe) {
+		spin_unlock_bh(&qp->qp_lock);
 		return -EBUSY;
+	}
 
+	sq_tail = qp_status->sq_tail;
+	sq_tail_next = (sq_tail + 1) % qp->sq_depth;
 	memcpy(sqe, msg, qp->qm->sqe_size);
+	qp->msg[sq_tail] = msg;
 
 	qm_db(qp->qm, qp->qp_id, QM_DOORBELL_CMD_SQ, sq_tail_next, 0);
 	atomic_inc(&qp->qp_status.used);
 	qp_status->sq_tail = sq_tail_next;
+	spin_unlock_bh(&qp->qp_lock);
 
 	return 0;
 }
@@ -2553,7 +2590,6 @@ static int hisi_qm_uacce_get_queue(struct uacce_device *uacce,
 	qp->uacce_q = q;
 	qp->event_cb = qm_qp_event_notifier;
 	qp->pasid = arg;
-	qp->is_in_kernel = false;
 
 	return 0;
 }
@@ -3067,12 +3103,13 @@ EXPORT_SYMBOL_GPL(hisi_qm_wait_task_finish);
 static void hisi_qp_memory_uninit(struct hisi_qm *qm, int num)
 {
 	struct device *dev = &qm->pdev->dev;
-	struct qm_dma *qdma;
+	struct hisi_qp *qp;
 	int i;
 
 	for (i = num - 1; i >= 0; i--) {
-		qdma = &qm->qp_array[i].qdma;
-		dma_free_coherent(dev, qdma->size, qdma->va, qdma->dma);
+		qp = &qm->qp_array[i];
+		dma_free_coherent(dev, qp->qdma.size, qp->qdma.va, qp->qdma.dma);
+		kfree(qp->msg);
 	}
 
 	kfree(qm->qp_array);
@@ -3086,10 +3123,16 @@ static int hisi_qp_memory_init(struct hisi_qm *qm, size_t dma_size, int id,
 	struct hisi_qp *qp;
 
 	qp = &qm->qp_array[id];
+	qp->msg = kmalloc_array(sq_depth, sizeof(void *), GFP_KERNEL);
+	if (!qp->msg)
+		return -ENOMEM;
+
 	qp->qdma.va = dma_alloc_coherent(dev, dma_size, &qp->qdma.dma,
 					 GFP_KERNEL);
-	if (!qp->qdma.va)
+	if (!qp->qdma.va) {
+		kfree(qp->msg);
 		return -ENOMEM;
+	}
 
 	qp->sqe = qp->qdma.va;
 	qp->sqe_dma = qp->qdma.dma;
@@ -3101,6 +3144,10 @@ static int hisi_qp_memory_init(struct hisi_qm *qm, size_t dma_size, int id,
 	qp->qm = qm;
 	qp->qp_id = id;
 
+	spin_lock_init(&qp->qp_lock);
+	spin_lock_init(&qp->backlog.lock);
+	INIT_LIST_HEAD(&qp->backlog.list);
+
 	return 0;
 }
 
@@ -3684,6 +3731,17 @@ static void qm_dev_err_uninit(struct hisi_qm *qm)
 	qm->err_ini->hw_err_disable(qm);
 }
 
+static void qm_release_qp_nolock(struct hisi_qp *qp)
+{
+	struct hisi_qm *qm = qp->qm;
+
+	if (--qp->ref_count)
+		return;
+
+	qm->qp_in_used--;
+	idr_remove(&qm->qp_idr, qp->qp_id);
+}
+
 /**
  * hisi_qm_free_qps() - free multiple queue pairs.
  * @qps: The queue pairs need to be freed.
@@ -3696,11 +3754,34 @@ void hisi_qm_free_qps(struct hisi_qp **qps, int qp_num)
 	if (!qps || qp_num <= 0)
 		return;
 
-	for (i = qp_num - 1; i >= 0; i--)
-		hisi_qm_release_qp(qps[i]);
+	down_write(&qps[0]->qm->qps_lock);
+
+	for (i = qp_num - 1; i >= 0; i--) {
+		if (qps[i]->ref_count == 1)
+			qm_stop_qp_nolock(qps[i]);
+
+		qm_release_qp_nolock(qps[i]);
+	}
+
+	up_write(&qps[0]->qm->qps_lock);
+	qm_pm_put_sync(qps[0]->qm);
 }
 EXPORT_SYMBOL_GPL(hisi_qm_free_qps);
 
+static void qm_insert_sorted(struct list_head *head, struct hisi_qm_resource *res)
+{
+	struct hisi_qm_resource *tmp;
+	struct list_head *n = head;
+
+	list_for_each_entry(tmp, head, list) {
+		if (res->distance < tmp->distance) {
+			n = &tmp->list;
+			break;
+		}
+	}
+	list_add_tail(&res->list, n);
+}
+
 static void free_list(struct list_head *head)
 {
 	struct hisi_qm_resource *res, *tmp;
@@ -3711,14 +3792,57 @@ static void free_list(struct list_head *head)
 	}
 }
 
+static int qm_get_and_start_qp(struct hisi_qm *qm, int qp_num, struct hisi_qp **qps, u8 *alg_type)
+{
+	int i, ret;
+
+	ret = qm_pm_get_sync(qm);
+	if (ret)
+		return ret;
+
+	down_write(&qm->qps_lock);
+	for (i = 0; i < qp_num; i++) {
+		qps[i] = qm_create_qp_nolock(qm, alg_type[i], true);
+		if (IS_ERR(qps[i])) {
+			ret = -ENODEV;
+			goto stop_and_free;
+		}
+
+		if (qps[i]->ref_count != 1)
+			continue;
+
+		ret = qm_start_qp_nolock(qps[i], 0);
+		if (ret) {
+			qm_release_qp_nolock(qps[i]);
+			goto stop_and_free;
+		}
+	}
+	up_write(&qm->qps_lock);
+
+	return 0;
+
+stop_and_free:
+	for (i--; i >= 0; i--) {
+		if (qps[i]->ref_count == 1)
+			qm_stop_qp_nolock(qps[i]);
+
+		qm_release_qp_nolock(qps[i]);
+	}
+	up_write(&qm->qps_lock);
+	qm_pm_put_sync(qm);
+
+	return ret;
+}
+
 static int hisi_qm_sort_devices(int node, struct list_head *head,
 				struct hisi_qm_list *qm_list)
 {
-	struct hisi_qm_resource *res, *tmp;
+	struct hisi_qm_resource *res;
 	struct hisi_qm *qm;
-	struct list_head *n;
 	struct device *dev;
 	int dev_node;
+	LIST_HEAD(non_full_list);
+	LIST_HEAD(full_list);
 
 	list_for_each_entry(qm, &qm_list->list, list) {
 		dev = &qm->pdev->dev;
@@ -3733,16 +3857,16 @@ static int hisi_qm_sort_devices(int node, struct list_head *head,
 
 		res->qm = qm;
 		res->distance = node_distance(dev_node, node);
-		n = head;
-		list_for_each_entry(tmp, head, list) {
-			if (res->distance < tmp->distance) {
-				n = &tmp->list;
-				break;
-			}
-		}
-		list_add_tail(&res->list, n);
+
+		if (qm->qp_in_used == qm->qp_num)
+			qm_insert_sorted(&full_list, res);
+		else
+			qm_insert_sorted(&non_full_list, res);
 	}
 
+	list_splice_tail(&non_full_list, head);
+	list_splice_tail(&full_list, head);
+
 	return 0;
 }
 
@@ -3759,12 +3883,11 @@ static int hisi_qm_sort_devices(int node, struct list_head *head,
  * not meet the requirements will return error.
  */
 int hisi_qm_alloc_qps_node(struct hisi_qm_list *qm_list, int qp_num,
-			   u8 alg_type, int node, struct hisi_qp **qps)
+			   u8 *alg_type, int node, struct hisi_qp **qps)
 {
 	struct hisi_qm_resource *tmp;
 	int ret = -ENODEV;
 	LIST_HEAD(head);
-	int i;
 
 	if (!qps || !qm_list || qp_num <= 0)
 		return -EINVAL;
@@ -3776,24 +3899,15 @@ int hisi_qm_alloc_qps_node(struct hisi_qm_list *qm_list, int qp_num,
 	}
 
 	list_for_each_entry(tmp, &head, list) {
-		for (i = 0; i < qp_num; i++) {
-			qps[i] = hisi_qm_create_qp(tmp->qm, alg_type);
-			if (IS_ERR(qps[i])) {
-				hisi_qm_free_qps(qps, i);
-				break;
-			}
-		}
-
-		if (i == qp_num) {
-			ret = 0;
+		ret = qm_get_and_start_qp(tmp->qm, qp_num, qps, alg_type);
+		if (!ret)
 			break;
-		}
 	}
 
 	mutex_unlock(&qm_list->lock);
 	if (ret)
-		pr_info_ratelimited("Too busy to create qps, node[%d], alg[%u], qp[%d]!\n",
-				     node, alg_type, qp_num);
+		pr_info("Failed to create qps, node[%d], qp[%d]!\n",
+			node, qp_num);
 
 err:
 	free_list(&head);
diff --git a/drivers/crypto/hisilicon/sec2/sec.h b/drivers/crypto/hisilicon/sec2/sec.h
index 81d0beda93b2bbbd3903bac7b257602a427de343..0710977861f32343f34b980db7f5ad7057c4360d 100644
--- a/drivers/crypto/hisilicon/sec2/sec.h
+++ b/drivers/crypto/hisilicon/sec2/sec.h
@@ -82,11 +82,6 @@ struct sec_aead_req {
 	__u8 out_mac_buf[SEC_MAX_MAC_LEN];
 };
 
-struct sec_instance_backlog {
-	struct list_head list;
-	spinlock_t lock;
-};
-
 /* SEC request of Crypto */
 struct sec_req {
 	union {
@@ -112,7 +107,6 @@ struct sec_req {
 	bool use_pbuf;
 
 	struct list_head list;
-	struct sec_instance_backlog *backlog;
 	struct sec_request_buf buf;
 };
 
@@ -172,7 +166,6 @@ struct sec_qp_ctx {
 	spinlock_t id_lock;
 	struct hisi_acc_sgl_pool *c_in_pool;
 	struct hisi_acc_sgl_pool *c_out_pool;
-	struct sec_instance_backlog backlog;
 	u16 send_head;
 };
 
diff --git a/drivers/crypto/hisilicon/sec2/sec_crypto.c b/drivers/crypto/hisilicon/sec2/sec_crypto.c
index 006350b8a1fdc5f6f71bf9ab92c10ca85e93ce6d..d8cca441db27d7bbeedf68783d15a3f425c4cdc7 100644
--- a/drivers/crypto/hisilicon/sec2/sec_crypto.c
+++ b/drivers/crypto/hisilicon/sec2/sec_crypto.c
@@ -54,7 +54,6 @@
 #define SEC_AUTH_CIPHER_V3	0x40
 #define SEC_FLAG_OFFSET		7
 #define SEC_FLAG_MASK		0x0780
-#define SEC_TYPE_MASK		0x0F
 #define SEC_DONE_MASK		0x0001
 #define SEC_ICV_MASK		0x000E
 
@@ -148,7 +147,7 @@ static void sec_free_req_id(struct sec_req *req)
 	spin_unlock_bh(&qp_ctx->id_lock);
 }
 
-static u8 pre_parse_finished_bd(struct bd_status *status, void *resp)
+static void pre_parse_finished_bd(struct bd_status *status, void *resp)
 {
 	struct sec_sqe *bd = resp;
 
@@ -158,11 +157,9 @@ static u8 pre_parse_finished_bd(struct bd_status *status, void *resp)
 					SEC_FLAG_MASK) >> SEC_FLAG_OFFSET;
 	status->tag = le16_to_cpu(bd->type2.tag);
 	status->err_type = bd->type2.error_type;
-
-	return bd->type_cipher_auth & SEC_TYPE_MASK;
 }
 
-static u8 pre_parse_finished_bd3(struct bd_status *status, void *resp)
+static void pre_parse_finished_bd3(struct bd_status *status, void *resp)
 {
 	struct sec_sqe3 *bd3 = resp;
 
@@ -172,8 +169,6 @@ static u8 pre_parse_finished_bd3(struct bd_status *status, void *resp)
 					SEC_FLAG_MASK) >> SEC_FLAG_OFFSET;
 	status->tag = le64_to_cpu(bd3->tag);
 	status->err_type = bd3->error_type;
-
-	return le32_to_cpu(bd3->bd_param) & SEC_TYPE_MASK;
 }
 
 static int sec_cb_status_check(struct sec_req *req,
@@ -244,7 +239,7 @@ static void sec_alg_send_backlog_soft(struct sec_ctx *ctx, struct sec_qp_ctx *qp
 	struct sec_req *req, *tmp;
 	int ret;
 
-	list_for_each_entry_safe(req, tmp, &qp_ctx->backlog.list, list) {
+	list_for_each_entry_safe(req, tmp, &qp_ctx->qp->backlog.list, list) {
 		list_del(&req->list);
 		ctx->req_op->buf_unmap(ctx, req);
 		if (req->req_id >= 0)
@@ -265,11 +260,12 @@ static void sec_alg_send_backlog_soft(struct sec_ctx *ctx, struct sec_qp_ctx *qp
 
 static void sec_alg_send_backlog(struct sec_ctx *ctx, struct sec_qp_ctx *qp_ctx)
 {
+	struct hisi_qp *qp = qp_ctx->qp;
 	struct sec_req *req, *tmp;
 	int ret;
 
-	spin_lock_bh(&qp_ctx->backlog.lock);
-	list_for_each_entry_safe(req, tmp, &qp_ctx->backlog.list, list) {
+	spin_lock_bh(&qp->backlog.lock);
+	list_for_each_entry_safe(req, tmp, &qp->backlog.list, list) {
 		ret = qp_send_message(req);
 		switch (ret) {
 		case -EINPROGRESS:
@@ -287,42 +283,46 @@ static void sec_alg_send_backlog(struct sec_ctx *ctx, struct sec_qp_ctx *qp_ctx)
 	}
 
 unlock:
-	spin_unlock_bh(&qp_ctx->backlog.lock);
+	spin_unlock_bh(&qp->backlog.lock);
 }
 
 static void sec_req_cb(struct hisi_qp *qp, void *resp)
 {
-	struct sec_qp_ctx *qp_ctx = qp->qp_ctx;
-	struct sec_dfx *dfx = &qp_ctx->ctx->sec->debug.dfx;
-	u8 type_supported = qp_ctx->ctx->type_supported;
+	const struct sec_sqe *sqe = qp->msg[qp->qp_status.cq_head];
+	struct sec_req *req = container_of(sqe, struct sec_req, sec_sqe);
+	struct sec_ctx *ctx = req->ctx;
+	struct sec_dfx *dfx = &ctx->sec->debug.dfx;
 	struct bd_status status;
-	struct sec_ctx *ctx;
-	struct sec_req *req;
 	int err;
-	u8 type;
 
-	if (type_supported == SEC_BD_TYPE2) {
-		type = pre_parse_finished_bd(&status, resp);
-		req = qp_ctx->req_list[status.tag];
-	} else {
-		type = pre_parse_finished_bd3(&status, resp);
-		req = (void *)(uintptr_t)status.tag;
-	}
+	pre_parse_finished_bd(&status, resp);
 
-	if (unlikely(type != type_supported)) {
-		atomic64_inc(&dfx->err_bd_cnt);
-		pr_err("err bd type [%u]\n", type);
-		return;
-	}
+	req->err_type = status.err_type;
+	err = sec_cb_status_check(req, &status);
+	if (err)
+		atomic64_inc(&dfx->done_flag_cnt);
 
-	if (unlikely(!req)) {
-		atomic64_inc(&dfx->invalid_req_cnt);
-		atomic_inc(&qp->qp_status.used);
-		return;
-	}
+	atomic64_inc(&dfx->recv_cnt);
 
+	ctx->req_op->buf_unmap(ctx, req);
+	ctx->req_op->callback(ctx, req, err);
+}
+
+static void sec_req_cb3(struct hisi_qp *qp, void *resp)
+{
+	struct bd_status status;
+	struct sec_ctx *ctx;
+	struct sec_dfx *dfx;
+	struct sec_req *req;
+	int err;
+
+	pre_parse_finished_bd3(&status, resp);
+
+	req = (void *)(uintptr_t)status.tag;
 	req->err_type = status.err_type;
 	ctx = req->ctx;
+	dfx = &ctx->sec->debug.dfx;
+
 	err = sec_cb_status_check(req, &status);
 	if (err)
 		atomic64_inc(&dfx->done_flag_cnt);
@@ -330,7 +330,6 @@ static void sec_req_cb(struct hisi_qp *qp, void *resp)
 	atomic64_inc(&dfx->recv_cnt);
 
 	ctx->req_op->buf_unmap(ctx, req);
-
 	ctx->req_op->callback(ctx, req, err);
 }
 
@@ -348,8 +347,10 @@ static int sec_alg_send_message_retry(struct sec_req *req)
 
 static int sec_alg_try_enqueue(struct sec_req *req)
 {
+	struct hisi_qp *qp = req->qp_ctx->qp;
+
 	/* Check if any request is already backlogged */
-	if (!list_empty(&req->backlog->list))
+	if (!list_empty(&qp->backlog.list))
 		return -EBUSY;
 
 	/* Try to enqueue to HW ring */
@@ -359,17 +360,18 @@ static int sec_alg_try_enqueue(struct sec_req *req)
 
 static int sec_alg_send_message_maybacklog(struct sec_req *req)
 {
+	struct hisi_qp *qp = req->qp_ctx->qp;
 	int ret;
 
 	ret = sec_alg_try_enqueue(req);
 	if (ret != -EBUSY)
 		return ret;
 
-	spin_lock_bh(&req->backlog->lock);
+	spin_lock_bh(&qp->backlog.lock);
 	ret = sec_alg_try_enqueue(req);
 	if (ret == -EBUSY)
-		list_add_tail(&req->list, &req->backlog->list);
-	spin_unlock_bh(&req->backlog->lock);
+		list_add_tail(&req->list, &qp->backlog.list);
+	spin_unlock_bh(&qp->backlog.lock);
 
 	return ret;
 }
@@ -624,32 +626,25 @@ static int sec_create_qp_ctx(struct sec_ctx *ctx, int qp_ctx_id)
 
 	qp_ctx = &ctx->qp_ctx[qp_ctx_id];
 	qp = ctx->qps[qp_ctx_id];
-	qp->req_type = 0;
-	qp->qp_ctx = qp_ctx;
 	qp_ctx->qp = qp;
 	qp_ctx->ctx = ctx;
 
-	qp->req_cb = sec_req_cb;
+	if (ctx->type_supported == SEC_BD_TYPE3)
+		qp->req_cb = sec_req_cb3;
+	else
+		qp->req_cb = sec_req_cb;
 
 	spin_lock_init(&qp_ctx->req_lock);
 	idr_init(&qp_ctx->req_idr);
-	spin_lock_init(&qp_ctx->backlog.lock);
 	spin_lock_init(&qp_ctx->id_lock);
-	INIT_LIST_HEAD(&qp_ctx->backlog.list);
 	qp_ctx->send_head = 0;
 
 	ret = sec_alloc_qp_ctx_resource(ctx, qp_ctx);
 	if (ret)
 		goto err_destroy_idr;
 
-	ret = hisi_qm_start_qp(qp, 0);
-	if (ret < 0)
-		goto err_resource_free;
-
 	return 0;
 
-err_resource_free:
-	sec_free_qp_ctx_resource(ctx, qp_ctx);
 err_destroy_idr:
 	idr_destroy(&qp_ctx->req_idr);
 	return ret;
@@ -658,7 +653,6 @@ static int sec_create_qp_ctx(struct sec_ctx *ctx, int qp_ctx_id)
 static void sec_release_qp_ctx(struct sec_ctx *ctx,
 			       struct sec_qp_ctx *qp_ctx)
 {
-	hisi_qm_stop_qp(qp_ctx->qp);
 	sec_free_qp_ctx_resource(ctx, qp_ctx);
 	idr_destroy(&qp_ctx->req_idr);
 }
@@ -669,10 +663,8 @@ static int sec_ctx_base_init(struct sec_ctx *ctx)
 	int i, ret;
 
 	ctx->qps = sec_create_qps();
-	if (!ctx->qps) {
-		pr_err("Can not create sec qps!\n");
+	if (!ctx->qps)
 		return -ENODEV;
-	}
 
 	sec = container_of(ctx->qps[0]->qm, struct sec_dev, qm);
 	ctx->sec = sec;
@@ -708,6 +700,9 @@ static void sec_ctx_base_uninit(struct sec_ctx *ctx)
 {
 	int i;
 
+	if (!ctx->qps)
+		return;
+
 	for (i = 0; i < ctx->sec->ctx_q_num; i++)
 		sec_release_qp_ctx(ctx, &ctx->qp_ctx[i]);
 
@@ -719,6 +714,9 @@ static int sec_cipher_init(struct sec_ctx *ctx)
 {
 	struct sec_cipher_ctx *c_ctx = &ctx->c_ctx;
 
+	if (!ctx->qps)
+		return 0;
+
 	c_ctx->c_key = dma_alloc_coherent(ctx->dev, SEC_MAX_KEY_SIZE,
 					  &c_ctx->c_key_dma, GFP_KERNEL);
 	if (!c_ctx->c_key)
@@ -731,6 +729,9 @@ static void sec_cipher_uninit(struct sec_ctx *ctx)
 {
 	struct sec_cipher_ctx *c_ctx = &ctx->c_ctx;
 
+	if (!ctx->qps)
+		return;
+
 	memzero_explicit(c_ctx->c_key, SEC_MAX_KEY_SIZE);
 	dma_free_coherent(ctx->dev, SEC_MAX_KEY_SIZE,
 			  c_ctx->c_key, c_ctx->c_key_dma);
@@ -752,6 +753,9 @@ static void sec_auth_uninit(struct sec_ctx *ctx)
 {
 	struct sec_auth_ctx *a_ctx = &ctx->a_ctx;
 
+	if (!ctx->qps)
+		return;
+
 	memzero_explicit(a_ctx->a_key, SEC_MAX_AKEY_SIZE);
 	dma_free_coherent(ctx->dev, SEC_MAX_AKEY_SIZE,
 			  a_ctx->a_key, a_ctx->a_key_dma);
@@ -789,7 +793,7 @@ static int sec_skcipher_init(struct crypto_skcipher *tfm)
 	}
 
 	ret = sec_ctx_base_init(ctx);
-	if (ret)
+	if (ret && ret != -ENODEV)
 		return ret;
 
 	ret = sec_cipher_init(ctx);
@@ -898,6 +902,9 @@ static int sec_skcipher_setkey(struct crypto_skcipher *tfm, const u8 *key,
 	struct device *dev = ctx->dev;
 	int ret;
 
+	if (!ctx->qps)
+		goto set_soft_key;
+
 	if (c_mode == SEC_CMODE_XTS) {
 		ret = xts_verify_key(tfm, key, keylen);
 		if (ret) {
@@ -928,13 +935,14 @@ static int sec_skcipher_setkey(struct crypto_skcipher *tfm, const u8 *key,
 	}
 
 	memcpy(c_ctx->c_key, key, keylen);
-	if (c_ctx->fbtfm) {
-		ret = crypto_sync_skcipher_setkey(c_ctx->fbtfm, key, keylen);
-		if (ret) {
-			dev_err(dev, "failed to set fallback skcipher key!\n");
-			return ret;
-		}
+
+set_soft_key:
+	ret = crypto_sync_skcipher_setkey(c_ctx->fbtfm, key, keylen);
+	if (ret) {
+		dev_err(dev, "failed to set fallback skcipher key!\n");
+		return ret;
 	}
+
 	return 0;
 }
 
@@ -1398,6 +1406,9 @@ static int sec_aead_setkey(struct crypto_aead *tfm, const u8 *key,
 	struct crypto_authenc_keys keys;
 	int ret;
 
+	if (!ctx->qps)
+		return sec_aead_fallback_setkey(a_ctx, tfm, key, keylen);
+
 	ctx->a_ctx.a_alg = a_alg;
 	ctx->c_ctx.c_alg = c_alg;
 	c_ctx->c_mode = c_mode;
@@ -1952,7 +1963,6 @@ static int sec_request_init(struct sec_ctx *ctx, struct sec_req *req)
 	} while (req->req_id < 0 && ++i < ctx->sec->ctx_q_num);
 
 	req->qp_ctx = qp_ctx;
-	req->backlog = &qp_ctx->backlog;
 
 	return 0;
 }
@@ -2055,6 +2065,9 @@ static int sec_skcipher_ctx_init(struct crypto_skcipher *tfm)
 	if (ret)
 		return ret;
 
+	if (!ctx->qps)
+		return 0;
+
 	if (ctx->sec->qm.ver < QM_HW_V3) {
 		ctx->type_supported = SEC_BD_TYPE2;
 		ctx->req_op = &sec_skcipher_req_ops;
@@ -2063,7 +2076,7 @@ static int sec_skcipher_ctx_init(struct crypto_skcipher *tfm)
 		ctx->req_op = &sec_skcipher_req_ops_v3;
 	}
 
-	return ret;
+	return 0;
 }
 
 static void sec_skcipher_ctx_exit(struct crypto_skcipher *tfm)
@@ -2131,7 +2144,7 @@ static int sec_aead_ctx_init(struct crypto_aead *tfm, const char *hash_name)
 	int ret;
 
 	ret = sec_aead_init(tfm);
-	if (ret) {
+	if (ret && ret != -ENODEV) {
 		pr_err("hisi_sec2: aead init error!\n");
 		return ret;
 	}
@@ -2173,7 +2186,7 @@ static int sec_aead_xcm_ctx_init(struct crypto_aead *tfm)
 	int ret;
 
 	ret = sec_aead_init(tfm);
-	if (ret) {
+	if (ret && ret != -ENODEV) {
 		dev_err(ctx->dev, "hisi_sec2: aead xcm init error!\n");
 		return ret;
 	}
@@ -2318,6 +2331,9 @@ static int sec_skcipher_crypto(struct skcipher_request *sk_req, bool encrypt)
 	bool need_fallback = false;
 	int ret;
 
+	if (!ctx->qps)
+		goto soft_crypto;
+
 	if (!sk_req->cryptlen) {
 		if (ctx->c_ctx.c_mode == SEC_CMODE_XTS)
 			return -EINVAL;
@@ -2335,9 +2351,12 @@ static int sec_skcipher_crypto(struct skcipher_request *sk_req, bool encrypt)
 		return -EINVAL;
 
 	if (unlikely(ctx->c_ctx.fallback || need_fallback))
-		return sec_skcipher_soft_crypto(ctx, sk_req, encrypt);
+		goto soft_crypto;
 
 	return ctx->req_op->process(ctx, req);
+
+soft_crypto:
+	return sec_skcipher_soft_crypto(ctx, sk_req, encrypt);
 }
 
 static int sec_skcipher_encrypt(struct skcipher_request *sk_req)
@@ -2545,6 +2564,9 @@ static int sec_aead_crypto(struct aead_request *a_req, bool encrypt)
 	bool need_fallback = false;
 	int ret;
 
+	if (!ctx->qps)
+		goto soft_crypto;
+
 	req->flag = a_req->base.flags;
 	req->aead_req.aead_req = a_req;
 	req->c_req.encrypt = encrypt;
@@ -2555,11 +2577,14 @@ static int sec_aead_crypto(struct aead_request *a_req, bool encrypt)
 	ret = sec_aead_param_check(ctx, req, &need_fallback);
 	if (unlikely(ret)) {
 		if (need_fallback)
-			return sec_aead_soft_crypto(ctx, a_req, encrypt);
+			goto soft_crypto;
 		return -EINVAL;
 	}
 
 	return ctx->req_op->process(ctx, req);
+
+soft_crypto:
+	return sec_aead_soft_crypto(ctx, a_req, encrypt);
 }
 
 static int sec_aead_encrypt(struct aead_request *a_req)
diff --git a/drivers/crypto/hisilicon/sec2/sec_main.c b/drivers/crypto/hisilicon/sec2/sec_main.c
index 1e79b7471d1b61d9f790a0d68f734fce8cfd37b8..30b6e29bb4671a27f2a7624325079740fd8e5e89 100644
--- a/drivers/crypto/hisilicon/sec2/sec_main.c
+++ b/drivers/crypto/hisilicon/sec2/sec_main.c
@@ -417,18 +417,29 @@ struct hisi_qp **sec_create_qps(void)
 	int node = cpu_to_node(raw_smp_processor_id());
 	u32 ctx_num = ctx_q_num;
 	struct hisi_qp **qps;
+	u8 *type;
 	int ret;
 
 	qps = kcalloc(ctx_num, sizeof(struct hisi_qp *), GFP_KERNEL);
 	if (!qps)
 		return NULL;
 
-	ret = hisi_qm_alloc_qps_node(&sec_devices, ctx_num, 0, node, qps);
-	if (!ret)
-		return qps;
+	/* The type of SEC is all 0, so just allocated by kcalloc */
+	type = kcalloc(ctx_num, sizeof(u8), GFP_KERNEL);
+	if (!type) {
+		kfree(qps);
+		return NULL;
+	}
 
-	kfree(qps);
-	return NULL;
+	ret = hisi_qm_alloc_qps_node(&sec_devices, ctx_num, type, node, qps);
+	if (ret) {
+		kfree(type);
+		kfree(qps);
+		return NULL;
+	}
+
+	kfree(type);
+	return qps;
 }
 
 u64 sec_get_alg_bitmap(struct hisi_qm *qm, u32 high, u32 low)
diff --git a/drivers/crypto/hisilicon/zip/zip.h b/drivers/crypto/hisilicon/zip/zip.h
index 9fb2a9c01132b48a1fa7cea14b99c80454e69491..b83f228281ab1eb8bdd5c6c9281db1bce28536f7 100644
--- a/drivers/crypto/hisilicon/zip/zip.h
+++ b/drivers/crypto/hisilicon/zip/zip.h
@@ -99,7 +99,7 @@ enum zip_cap_table_type {
 	ZIP_CORE5_BITMAP,
 };
 
-int zip_create_qps(struct hisi_qp **qps, int qp_num, int node);
+int zip_create_qps(struct hisi_qp **qps, int qp_num, int node, u8 *alg_type);
 int hisi_zip_register_to_crypto(struct hisi_qm *qm);
 void hisi_zip_unregister_from_crypto(struct hisi_qm *qm);
 bool hisi_zip_alg_support(struct hisi_qm *qm, u32 alg);
diff --git a/drivers/crypto/hisilicon/zip/zip_crypto.c b/drivers/crypto/hisilicon/zip/zip_crypto.c
index 9df9a97633c6fea4259d9068e332a342311a473c..25493821ac2b8d9e01e32a8b9273a8c4610514e8 100644
--- a/drivers/crypto/hisilicon/zip/zip_crypto.c
+++ b/drivers/crypto/hisilicon/zip/zip_crypto.c
@@ -4,6 +4,7 @@
 #include <linux/bitfield.h>
 #include <linux/bitmap.h>
 #include <linux/dma-mapping.h>
+#include <linux/highmem.h>
 #include <linux/scatterlist.h>
 #include "zip.h"
 
@@ -39,6 +40,7 @@ enum {
 	HZIP_CTX_Q_NUM
 };
 
+#define GET_REQ_FROM_SQE(sqe)	((u64)(sqe)->dw26 | (u64)(sqe)->dw27 << 32)
 #define COMP_NAME_TO_TYPE(alg_name)					\
 	(!strcmp((alg_name), "deflate") ? HZIP_ALG_TYPE_DEFLATE : 0)
 
@@ -48,6 +50,7 @@ struct hisi_zip_req {
 	struct hisi_acc_hw_sgl *hw_dst;
 	dma_addr_t dma_src;
 	dma_addr_t dma_dst;
+	struct hisi_zip_qp_ctx *qp_ctx;
 	u16 req_id;
 };
 
@@ -64,6 +67,7 @@ struct hisi_zip_qp_ctx {
 	struct hisi_acc_sgl_pool *sgl_pool;
 	struct hisi_zip *zip_dev;
 	struct hisi_zip_ctx *ctx;
+	u8 req_type;
 };
 
 struct hisi_zip_sqe_ops {
@@ -74,7 +78,6 @@ struct hisi_zip_sqe_ops {
 	void (*fill_req_type)(struct hisi_zip_sqe *sqe, u8 req_type);
 	void (*fill_tag)(struct hisi_zip_sqe *sqe, struct hisi_zip_req *req);
 	void (*fill_sqe_type)(struct hisi_zip_sqe *sqe, u8 sqe_type);
-	u32 (*get_tag)(struct hisi_zip_sqe *sqe);
 	u32 (*get_status)(struct hisi_zip_sqe *sqe);
 	u32 (*get_dstlen)(struct hisi_zip_sqe *sqe);
 };
@@ -82,6 +85,8 @@ struct hisi_zip_sqe_ops {
 struct hisi_zip_ctx {
 	struct hisi_zip_qp_ctx qp_ctx[HZIP_CTX_Q_NUM];
 	const struct hisi_zip_sqe_ops *ops;
+	struct crypto_comp *soft_tfm;
+	bool fallback;
 };
 
 static int sgl_sge_nr_set(const char *val, const struct kernel_param *kp)
@@ -108,6 +113,38 @@ static u16 sgl_sge_nr = HZIP_SGL_SGE_NR;
 module_param_cb(sgl_sge_nr, &sgl_sge_nr_ops, &sgl_sge_nr, 0444);
 MODULE_PARM_DESC(sgl_sge_nr, "Number of sge in sgl(1-255)");
 
+static int hisi_zip_fallback_do_work(struct crypto_comp *tfm, struct acomp_req *acomp_req,
+			  bool is_decompress)
+{
+	void *input, *output;
+	const char *algo;
+	int err = 0;
+	int ret;
+
+	input = kmap_local_page(sg_page(acomp_req->src)) + acomp_req->src->offset;
+	output = kmap_local_page(sg_page(acomp_req->dst)) + acomp_req->dst->offset;
+
+	if (!is_decompress)
+		ret = crypto_comp_compress(tfm, input, acomp_req->slen,
+					   output, &acomp_req->dlen);
+	else
+		ret = crypto_comp_decompress(tfm, input, acomp_req->slen,
+					     output, &acomp_req->dlen);
+	if (ret) {
+		algo = crypto_tfm_alg_driver_name(crypto_comp_tfm(tfm));
+		pr_err("failed to do fallback %s work, ret=%d\n", algo, ret);
+		err = -EIO;
+	}
+
+	kunmap_local(output);
+	kunmap_local(input);
+
+	if (acomp_req->base.complete)
+		acomp_request_complete(acomp_req, err);
+
+	return ret;
+}
+
 static struct hisi_zip_req *hisi_zip_create_req(struct hisi_zip_qp_ctx *qp_ctx,
 						struct acomp_req *req)
 {
@@ -131,6 +168,7 @@ static struct hisi_zip_req *hisi_zip_create_req(struct hisi_zip_qp_ctx *qp_ctx,
 	req_cache = q + req_id;
 	req_cache->req_id = req_id;
 	req_cache->req = req;
+	req_cache->qp_ctx = qp_ctx;
 
 	return req_cache;
 }
@@ -181,7 +219,8 @@ static void hisi_zip_fill_req_type(struct hisi_zip_sqe *sqe, u8 req_type)
 
 static void hisi_zip_fill_tag(struct hisi_zip_sqe *sqe, struct hisi_zip_req *req)
 {
-	sqe->dw26 = req->req_id;
+	sqe->dw26 = lower_32_bits((u64)req);
+	sqe->dw27 = upper_32_bits((u64)req);
 }
 
 static void hisi_zip_fill_sqe_type(struct hisi_zip_sqe *sqe, u8 sqe_type)
@@ -213,7 +252,6 @@ static int hisi_zip_do_work(struct hisi_zip_qp_ctx *qp_ctx,
 {
 	struct hisi_acc_sgl_pool *pool = qp_ctx->sgl_pool;
 	struct hisi_zip_dfx *dfx = &qp_ctx->zip_dev->dfx;
-	struct hisi_zip_req_q *req_q = &qp_ctx->req_q;
 	struct acomp_req *a_req = req->req;
 	struct hisi_qp *qp = qp_ctx->qp;
 	struct device *dev = &qp->qm->pdev->dev;
@@ -242,13 +280,11 @@ static int hisi_zip_do_work(struct hisi_zip_qp_ctx *qp_ctx,
 		goto err_unmap_input;
 	}
 
-	hisi_zip_fill_sqe(qp_ctx->ctx, &zip_sqe, qp->req_type, req);
+	hisi_zip_fill_sqe(qp_ctx->ctx, &zip_sqe, qp_ctx->req_type, req);
 
 	/* send command to start a task */
 	atomic64_inc(&dfx->send_cnt);
-	spin_lock_bh(&req_q->req_lock);
 	ret = hisi_qp_send(qp, &zip_sqe);
-	spin_unlock_bh(&req_q->req_lock);
 	if (unlikely(ret < 0)) {
 		atomic64_inc(&dfx->send_busy_cnt);
 		ret = -EAGAIN;
@@ -265,11 +301,6 @@ static int hisi_zip_do_work(struct hisi_zip_qp_ctx *qp_ctx,
 	return ret;
 }
 
-static u32 hisi_zip_get_tag(struct hisi_zip_sqe *sqe)
-{
-	return sqe->dw26;
-}
-
 static u32 hisi_zip_get_status(struct hisi_zip_sqe *sqe)
 {
 	return sqe->dw3 & HZIP_BD_STATUS_M;
@@ -282,14 +313,12 @@ static u32 hisi_zip_get_dstlen(struct hisi_zip_sqe *sqe)
 
 static void hisi_zip_acomp_cb(struct hisi_qp *qp, void *data)
 {
-	struct hisi_zip_qp_ctx *qp_ctx = qp->qp_ctx;
+	struct hisi_zip_sqe *sqe = data;
+	struct hisi_zip_req *req = (struct hisi_zip_req *)GET_REQ_FROM_SQE(sqe);
+	struct hisi_zip_qp_ctx *qp_ctx = req->qp_ctx;
 	const struct hisi_zip_sqe_ops *ops = qp_ctx->ctx->ops;
 	struct hisi_zip_dfx *dfx = &qp_ctx->zip_dev->dfx;
-	struct hisi_zip_req_q *req_q = &qp_ctx->req_q;
 	struct device *dev = &qp->qm->pdev->dev;
-	struct hisi_zip_sqe *sqe = data;
-	u32 tag = ops->get_tag(sqe);
-	struct hisi_zip_req *req = req_q->q + tag;
 	struct acomp_req *acomp_req = req->req;
 	int err = 0;
 	u32 status;
@@ -323,6 +352,9 @@ static int hisi_zip_acompress(struct acomp_req *acomp_req)
 	struct hisi_zip_req *req;
 	int ret;
 
+	if (ctx->fallback)
+		return hisi_zip_fallback_do_work(ctx->soft_tfm, acomp_req, 0);
+
 	req = hisi_zip_create_req(qp_ctx, acomp_req);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
@@ -344,6 +376,9 @@ static int hisi_zip_adecompress(struct acomp_req *acomp_req)
 	struct hisi_zip_req *req;
 	int ret;
 
+	if (ctx->fallback)
+		return hisi_zip_fallback_do_work(ctx->soft_tfm, acomp_req, 1);
+
 	req = hisi_zip_create_req(qp_ctx, acomp_req);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
@@ -358,33 +393,6 @@ static int hisi_zip_adecompress(struct acomp_req *acomp_req)
 	return ret;
 }
 
-static int hisi_zip_start_qp(struct hisi_qp *qp, struct hisi_zip_qp_ctx *qp_ctx,
-			     int alg_type, int req_type)
-{
-	struct device *dev = &qp->qm->pdev->dev;
-	int ret;
-
-	qp->req_type = req_type;
-	qp->alg_type = alg_type;
-	qp->qp_ctx = qp_ctx;
-
-	ret = hisi_qm_start_qp(qp, 0);
-	if (ret < 0) {
-		dev_err(dev, "failed to start qp (%d)!\n", ret);
-		return ret;
-	}
-
-	qp_ctx->qp = qp;
-
-	return 0;
-}
-
-static void hisi_zip_release_qp(struct hisi_zip_qp_ctx *qp_ctx)
-{
-	hisi_qm_stop_qp(qp_ctx->qp);
-	hisi_qm_free_qps(&qp_ctx->qp, 1);
-}
-
 static const struct hisi_zip_sqe_ops hisi_zip_ops = {
 	.sqe_type		= 0x3,
 	.fill_addr		= hisi_zip_fill_addr,
@@ -393,7 +401,6 @@ static const struct hisi_zip_sqe_ops hisi_zip_ops = {
 	.fill_req_type		= hisi_zip_fill_req_type,
 	.fill_tag		= hisi_zip_fill_tag,
 	.fill_sqe_type		= hisi_zip_fill_sqe_type,
-	.get_tag		= hisi_zip_get_tag,
 	.get_status		= hisi_zip_get_status,
 	.get_dstlen		= hisi_zip_get_dstlen,
 };
@@ -402,10 +409,15 @@ static int hisi_zip_ctx_init(struct hisi_zip_ctx *hisi_zip_ctx, u8 req_type, int
 {
 	struct hisi_qp *qps[HZIP_CTX_Q_NUM] = { NULL };
 	struct hisi_zip_qp_ctx *qp_ctx;
+	u8 alg_type[HZIP_CTX_Q_NUM];
 	struct hisi_zip *hisi_zip;
-	int ret, i, j;
+	int ret, i;
+
+	/* alg_type = 0 for compress, 1 for decompress in hw sqe */
+	for (i = 0; i < HZIP_CTX_Q_NUM; i++)
+		alg_type[i] = i;
 
-	ret = zip_create_qps(qps, HZIP_CTX_Q_NUM, node);
+	ret = zip_create_qps(qps, HZIP_CTX_Q_NUM, node, alg_type);
 	if (ret) {
 		pr_err("failed to create zip qps (%d)!\n", ret);
 		return -ENODEV;
@@ -414,19 +426,11 @@ static int hisi_zip_ctx_init(struct hisi_zip_ctx *hisi_zip_ctx, u8 req_type, int
 	hisi_zip = container_of(qps[0]->qm, struct hisi_zip, qm);
 
 	for (i = 0; i < HZIP_CTX_Q_NUM; i++) {
-		/* alg_type = 0 for compress, 1 for decompress in hw sqe */
 		qp_ctx = &hisi_zip_ctx->qp_ctx[i];
 		qp_ctx->ctx = hisi_zip_ctx;
-		ret = hisi_zip_start_qp(qps[i], qp_ctx, i, req_type);
-		if (ret) {
-			for (j = i - 1; j >= 0; j--)
-				hisi_qm_stop_qp(hisi_zip_ctx->qp_ctx[j].qp);
-
-			hisi_qm_free_qps(qps, HZIP_CTX_Q_NUM);
-			return ret;
-		}
-
 		qp_ctx->zip_dev = hisi_zip;
+		qp_ctx->req_type = req_type;
+		qp_ctx->qp = qps[i];
 	}
 
 	hisi_zip_ctx->ops = &hisi_zip_ops;
@@ -436,10 +440,13 @@ static int hisi_zip_ctx_init(struct hisi_zip_ctx *hisi_zip_ctx, u8 req_type, int
 
 static void hisi_zip_ctx_exit(struct hisi_zip_ctx *hisi_zip_ctx)
 {
+	struct hisi_qp *qps[HZIP_CTX_Q_NUM] = { NULL };
 	int i;
 
 	for (i = 0; i < HZIP_CTX_Q_NUM; i++)
-		hisi_zip_release_qp(&hisi_zip_ctx->qp_ctx[i]);
+		qps[i] = hisi_zip_ctx->qp_ctx[i].qp;
+
+	hisi_qm_free_qps(qps, HZIP_CTX_Q_NUM);
 }
 
 static int hisi_zip_create_req_q(struct hisi_zip_ctx *ctx)
@@ -539,6 +546,29 @@ static void hisi_zip_set_acomp_cb(struct hisi_zip_ctx *ctx,
 		ctx->qp_ctx[i].qp->req_cb = fn;
 }
 
+static int hisi_zip_fallback_init(struct hisi_zip_ctx *ctx, const char *alg_name)
+{
+	if (!IS_ERR_OR_NULL(ctx->soft_tfm))
+		return 0;
+
+	ctx->soft_tfm = crypto_alloc_comp(alg_name, 0, 0);
+	if (IS_ERR_OR_NULL(ctx->soft_tfm)) {
+		pr_err("could not alloc soft tfm %s\n", alg_name);
+		return PTR_ERR(ctx->soft_tfm);
+	}
+
+	return 0;
+}
+
+static void hisi_zip_fallback_uninit(struct hisi_zip_ctx *ctx)
+{
+	if (IS_ERR_OR_NULL(ctx->soft_tfm))
+		return;
+
+	crypto_free_comp(ctx->soft_tfm);
+	ctx->soft_tfm = NULL;
+}
+
 static int hisi_zip_acomp_init(struct crypto_acomp *tfm)
 {
 	const char *alg_name = crypto_tfm_alg_name(&tfm->base);
@@ -549,7 +579,7 @@ static int hisi_zip_acomp_init(struct crypto_acomp *tfm)
 	ret = hisi_zip_ctx_init(ctx, COMP_NAME_TO_TYPE(alg_name), tfm->base.node);
 	if (ret) {
 		pr_err("failed to init ctx (%d)!\n", ret);
-		return ret;
+		goto switch_to_soft;
 	}
 
 	dev = &ctx->qp_ctx[0].qp->qm->pdev->dev;
@@ -574,17 +604,22 @@ static int hisi_zip_acomp_init(struct crypto_acomp *tfm)
 	hisi_zip_release_req_q(ctx);
 err_ctx_exit:
 	hisi_zip_ctx_exit(ctx);
-	return ret;
+switch_to_soft:
+	ctx->fallback = true;
+	return hisi_zip_fallback_init(ctx, alg_name);
 }
 
 static void hisi_zip_acomp_exit(struct crypto_acomp *tfm)
 {
 	struct hisi_zip_ctx *ctx = crypto_tfm_ctx(&tfm->base);
 
-	hisi_zip_set_acomp_cb(ctx, NULL);
-	hisi_zip_release_sgl_pool(ctx);
-	hisi_zip_release_req_q(ctx);
-	hisi_zip_ctx_exit(ctx);
+	if (!ctx->fallback) {
+		hisi_zip_release_sgl_pool(ctx);
+		hisi_zip_release_req_q(ctx);
+		hisi_zip_ctx_exit(ctx);
+	}
+
+	hisi_zip_fallback_uninit(ctx);
 }
 
 static struct acomp_alg hisi_zip_acomp_deflate = {
@@ -595,7 +630,8 @@ static struct acomp_alg hisi_zip_acomp_deflate = {
 	.base			= {
 		.cra_name		= "deflate",
 		.cra_driver_name	= "hisi-deflate-acomp",
-		.cra_flags		= CRYPTO_ALG_ASYNC,
+		.cra_flags		= CRYPTO_ALG_ASYNC |
+					  CRYPTO_ALG_NEED_FALLBACK,
 		.cra_module		= THIS_MODULE,
 		.cra_priority		= HZIP_ALG_PRIORITY,
 		.cra_ctxsize		= sizeof(struct hisi_zip_ctx),
diff --git a/drivers/crypto/hisilicon/zip/zip_main.c b/drivers/crypto/hisilicon/zip/zip_main.c
index 956539f3a49470ce91a567738111b1ef23d75bbf..cb55ae8e6ce31d7cadd9f7c7800061a3ca2d466e 100644
--- a/drivers/crypto/hisilicon/zip/zip_main.c
+++ b/drivers/crypto/hisilicon/zip/zip_main.c
@@ -445,12 +445,12 @@ static const struct pci_device_id hisi_zip_dev_ids[] = {
 };
 MODULE_DEVICE_TABLE(pci, hisi_zip_dev_ids);
 
-int zip_create_qps(struct hisi_qp **qps, int qp_num, int node)
+int zip_create_qps(struct hisi_qp **qps, int qp_num, int node, u8 *alg_type)
 {
 	if (node == NUMA_NO_NODE)
 		node = cpu_to_node(raw_smp_processor_id());
 
-	return hisi_qm_alloc_qps_node(&zip_devices, qp_num, 0, node, qps);
+	return hisi_qm_alloc_qps_node(&zip_devices, qp_num, alg_type, node, qps);
 }
 
 bool hisi_zip_alg_support(struct hisi_qm *qm, u32 alg)
diff --git a/include/linux/hisi_acc_qm.h b/include/linux/hisi_acc_qm.h
index 9e5b360cc44766fb248749b4a58509e0152feff2..d0141e045ccc21013400bcd54cea79c7cfe25a63 100644
--- a/include/linux/hisi_acc_qm.h
+++ b/include/linux/hisi_acc_qm.h
@@ -448,12 +448,16 @@ struct hisi_qp_ops {
 	int (*fill_sqe)(void *sqe, void *q_parm, void *d_parm);
 };
 
+struct instance_backlog {
+	struct list_head list;
+	spinlock_t lock;
+};
+
 struct hisi_qp {
 	u32 qp_id;
 	u16 sq_depth;
 	u16 cq_depth;
 	u8 alg_type;
-	u8 req_type;
 
 	struct qm_dma qdma;
 	void *sqe;
@@ -463,7 +467,6 @@ struct hisi_qp {
 
 	struct hisi_qp_status qp_status;
 	struct hisi_qp_ops *hw_ops;
-	void *qp_ctx;
 	void (*req_cb)(struct hisi_qp *qp, void *data);
 	void (*event_cb)(struct hisi_qp *qp);
 
@@ -472,6 +475,11 @@ struct hisi_qp {
 	bool is_in_kernel;
 	u16 pasid;
 	struct uacce_queue *uacce_q;
+
+	u32 ref_count;
+	spinlock_t qp_lock;
+	struct instance_backlog backlog;
+	const void **msg;
 };
 
 static inline int vfs_num_set(const char *val, const struct kernel_param *kp)
@@ -575,7 +583,7 @@ struct hisi_acc_sgl_pool *hisi_acc_create_sgl_pool(struct device *dev,
 void hisi_acc_free_sgl_pool(struct device *dev,
 			    struct hisi_acc_sgl_pool *pool);
 int hisi_qm_alloc_qps_node(struct hisi_qm_list *qm_list, int qp_num,
-			   u8 alg_type, int node, struct hisi_qp **qps);
+			   u8 *alg_type, int node, struct hisi_qp **qps);
 void hisi_qm_free_qps(struct hisi_qp **qps, int qp_num);
 void hisi_qm_dev_shutdown(struct pci_dev *pdev);
 void hisi_qm_wait_task_finish(struct hisi_qm *qm, struct hisi_qm_list *qm_list);