crypto: ccree - add ahash support

Add CryptoCell async. hash and HMAC support.

Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

drivers/crypto/ccree/Makefile

@@ -1,6 +1,6 @@
 # SPDX-License-Identifier: GPL-2.0
 
 obj-$(CONFIG_CRYPTO_DEV_CCREE) := ccree.o
-ccree-y := cc_driver.o cc_buffer_mgr.o cc_request_mgr.o cc_cipher.o cc_ivgen.o cc_sram_mgr.o
+ccree-y := cc_driver.o cc_buffer_mgr.o cc_request_mgr.o cc_cipher.o cc_hash.o cc_ivgen.o cc_sram_mgr.o
 ccree-$(CONFIG_DEBUG_FS) += cc_debugfs.o
 ccree-$(CONFIG_PM) += cc_pm.o

drivers/crypto/ccree/cc_buffer_mgr.c

@@ -9,6 +9,7 @@
 #include "cc_buffer_mgr.h"
 #include "cc_lli_defs.h"
 #include "cc_cipher.h"
+#include "cc_hash.h"
 
 enum dma_buffer_type {
 	DMA_NULL_TYPE = -1,
@@ -348,9 +349,33 @@ static int cc_map_sg(struct device *dev, struct scatterlist *sg,
 	return 0;
 }
 
+static int cc_set_hash_buf(struct device *dev, struct ahash_req_ctx *areq_ctx,
+			   u8 *curr_buff, u32 curr_buff_cnt,
+			   struct buffer_array *sg_data)
+{
+	dev_dbg(dev, " handle curr buff %x set to   DLLI\n", curr_buff_cnt);
+	/* create sg for the current buffer */
+	sg_init_one(areq_ctx->buff_sg, curr_buff, curr_buff_cnt);
+	if (dma_map_sg(dev, areq_ctx->buff_sg, 1, DMA_TO_DEVICE) != 1) {
+		dev_err(dev, "dma_map_sg() src buffer failed\n");
+		return -ENOMEM;
+	}
+	dev_dbg(dev, "Mapped curr_buff: dma_address=%pad page=%p addr=%pK offset=%u length=%u\n",
+		&sg_dma_address(areq_ctx->buff_sg), sg_page(areq_ctx->buff_sg),
+		sg_virt(areq_ctx->buff_sg), areq_ctx->buff_sg->offset,
+		areq_ctx->buff_sg->length);
+	areq_ctx->data_dma_buf_type = CC_DMA_BUF_DLLI;
+	areq_ctx->curr_sg = areq_ctx->buff_sg;
+	areq_ctx->in_nents = 0;
+	/* prepare for case of MLLI */
+	cc_add_sg_entry(dev, sg_data, 1, areq_ctx->buff_sg, curr_buff_cnt, 0,
+			false, NULL);
+	return 0;
+}
+
 void cc_unmap_cipher_request(struct device *dev, void *ctx,
-			     unsigned int ivsize, struct scatterlist *src,
-			     struct scatterlist *dst)
+				unsigned int ivsize, struct scatterlist *src,
+				struct scatterlist *dst)
 {
 	struct cipher_req_ctx *req_ctx = (struct cipher_req_ctx *)ctx;
 
@@ -472,6 +497,238 @@ cipher_exit:
 	return rc;
 }
 
+int cc_map_hash_request_final(struct cc_drvdata *drvdata, void *ctx,
+			      struct scatterlist *src, unsigned int nbytes,
+			      bool do_update, gfp_t flags)
+{
+	struct ahash_req_ctx *areq_ctx = (struct ahash_req_ctx *)ctx;
+	struct device *dev = drvdata_to_dev(drvdata);
+	u8 *curr_buff = cc_hash_buf(areq_ctx);
+	u32 *curr_buff_cnt = cc_hash_buf_cnt(areq_ctx);
+	struct mlli_params *mlli_params = &areq_ctx->mlli_params;
+	struct buffer_array sg_data;
+	struct buff_mgr_handle *buff_mgr = drvdata->buff_mgr_handle;
+	u32 dummy = 0;
+	u32 mapped_nents = 0;
+
+	dev_dbg(dev, "final params : curr_buff=%pK curr_buff_cnt=0x%X nbytes = 0x%X src=%pK curr_index=%u\n",
+		curr_buff, *curr_buff_cnt, nbytes, src, areq_ctx->buff_index);
+	/* Init the type of the dma buffer */
+	areq_ctx->data_dma_buf_type = CC_DMA_BUF_NULL;
+	mlli_params->curr_pool = NULL;
+	sg_data.num_of_buffers = 0;
+	areq_ctx->in_nents = 0;
+
+	if (nbytes == 0 && *curr_buff_cnt == 0) {
+		/* nothing to do */
+		return 0;
+	}
+
+	/*TODO: copy data in case that buffer is enough for operation */
+	/* map the previous buffer */
+	if (*curr_buff_cnt) {
+		if (cc_set_hash_buf(dev, areq_ctx, curr_buff, *curr_buff_cnt,
+				    &sg_data)) {
+			return -ENOMEM;
+		}
+	}
+
+	if (src && nbytes > 0 && do_update) {
+		if (cc_map_sg(dev, src, nbytes, DMA_TO_DEVICE,
+			      &areq_ctx->in_nents, LLI_MAX_NUM_OF_DATA_ENTRIES,
+			      &dummy, &mapped_nents)) {
+			goto unmap_curr_buff;
+		}
+		if (src && mapped_nents == 1 &&
+		    areq_ctx->data_dma_buf_type == CC_DMA_BUF_NULL) {
+			memcpy(areq_ctx->buff_sg, src,
+			       sizeof(struct scatterlist));
+			areq_ctx->buff_sg->length = nbytes;
+			areq_ctx->curr_sg = areq_ctx->buff_sg;
+			areq_ctx->data_dma_buf_type = CC_DMA_BUF_DLLI;
+		} else {
+			areq_ctx->data_dma_buf_type = CC_DMA_BUF_MLLI;
+		}
+	}
+
+	/*build mlli */
+	if (areq_ctx->data_dma_buf_type == CC_DMA_BUF_MLLI) {
+		mlli_params->curr_pool = buff_mgr->mlli_buffs_pool;
+		/* add the src data to the sg_data */
+		cc_add_sg_entry(dev, &sg_data, areq_ctx->in_nents, src, nbytes,
+				0, true, &areq_ctx->mlli_nents);
+		if (cc_generate_mlli(dev, &sg_data, mlli_params, flags))
+			goto fail_unmap_din;
+	}
+	/* change the buffer index for the unmap function */
+	areq_ctx->buff_index = (areq_ctx->buff_index ^ 1);
+	dev_dbg(dev, "areq_ctx->data_dma_buf_type = %s\n",
+		cc_dma_buf_type(areq_ctx->data_dma_buf_type));
+	return 0;
+
+fail_unmap_din:
+	dma_unmap_sg(dev, src, areq_ctx->in_nents, DMA_TO_DEVICE);
+
+unmap_curr_buff:
+	if (*curr_buff_cnt)
+		dma_unmap_sg(dev, areq_ctx->buff_sg, 1, DMA_TO_DEVICE);
+
+	return -ENOMEM;
+}
+
+int cc_map_hash_request_update(struct cc_drvdata *drvdata, void *ctx,
+			       struct scatterlist *src, unsigned int nbytes,
+			       unsigned int block_size, gfp_t flags)
+{
+	struct ahash_req_ctx *areq_ctx = (struct ahash_req_ctx *)ctx;
+	struct device *dev = drvdata_to_dev(drvdata);
+	u8 *curr_buff = cc_hash_buf(areq_ctx);
+	u32 *curr_buff_cnt = cc_hash_buf_cnt(areq_ctx);
+	u8 *next_buff = cc_next_buf(areq_ctx);
+	u32 *next_buff_cnt = cc_next_buf_cnt(areq_ctx);
+	struct mlli_params *mlli_params = &areq_ctx->mlli_params;
+	unsigned int update_data_len;
+	u32 total_in_len = nbytes + *curr_buff_cnt;
+	struct buffer_array sg_data;
+	struct buff_mgr_handle *buff_mgr = drvdata->buff_mgr_handle;
+	unsigned int swap_index = 0;
+	u32 dummy = 0;
+	u32 mapped_nents = 0;
+
+	dev_dbg(dev, " update params : curr_buff=%pK curr_buff_cnt=0x%X nbytes=0x%X src=%pK curr_index=%u\n",
+		curr_buff, *curr_buff_cnt, nbytes, src, areq_ctx->buff_index);
+	/* Init the type of the dma buffer */
+	areq_ctx->data_dma_buf_type = CC_DMA_BUF_NULL;
+	mlli_params->curr_pool = NULL;
+	areq_ctx->curr_sg = NULL;
+	sg_data.num_of_buffers = 0;
+	areq_ctx->in_nents = 0;
+
+	if (total_in_len < block_size) {
+		dev_dbg(dev, " less than one block: curr_buff=%pK *curr_buff_cnt=0x%X copy_to=%pK\n",
+			curr_buff, *curr_buff_cnt, &curr_buff[*curr_buff_cnt]);
+		areq_ctx->in_nents =
+			cc_get_sgl_nents(dev, src, nbytes, &dummy, NULL);
+		sg_copy_to_buffer(src, areq_ctx->in_nents,
+				  &curr_buff[*curr_buff_cnt], nbytes);
+		*curr_buff_cnt += nbytes;
+		return 1;
+	}
+
+	/* Calculate the residue size*/
+	*next_buff_cnt = total_in_len & (block_size - 1);
+	/* update data len */
+	update_data_len = total_in_len - *next_buff_cnt;
+
+	dev_dbg(dev, " temp length : *next_buff_cnt=0x%X update_data_len=0x%X\n",
+		*next_buff_cnt, update_data_len);
+
+	/* Copy the new residue to next buffer */
+	if (*next_buff_cnt) {
+		dev_dbg(dev, " handle residue: next buff %pK skip data %u residue %u\n",
+			next_buff, (update_data_len - *curr_buff_cnt),
+			*next_buff_cnt);
+		cc_copy_sg_portion(dev, next_buff, src,
+				   (update_data_len - *curr_buff_cnt),
+				   nbytes, CC_SG_TO_BUF);
+		/* change the buffer index for next operation */
+		swap_index = 1;
+	}
+
+	if (*curr_buff_cnt) {
+		if (cc_set_hash_buf(dev, areq_ctx, curr_buff, *curr_buff_cnt,
+				    &sg_data)) {
+			return -ENOMEM;
+		}
+		/* change the buffer index for next operation */
+		swap_index = 1;
+	}
+
+	if (update_data_len > *curr_buff_cnt) {
+		if (cc_map_sg(dev, src, (update_data_len - *curr_buff_cnt),
+			      DMA_TO_DEVICE, &areq_ctx->in_nents,
+			      LLI_MAX_NUM_OF_DATA_ENTRIES, &dummy,
+			      &mapped_nents)) {
+			goto unmap_curr_buff;
+		}
+		if (mapped_nents == 1 &&
+		    areq_ctx->data_dma_buf_type == CC_DMA_BUF_NULL) {
+			/* only one entry in the SG and no previous data */
+			memcpy(areq_ctx->buff_sg, src,
+			       sizeof(struct scatterlist));
+			areq_ctx->buff_sg->length = update_data_len;
+			areq_ctx->data_dma_buf_type = CC_DMA_BUF_DLLI;
+			areq_ctx->curr_sg = areq_ctx->buff_sg;
+		} else {
+			areq_ctx->data_dma_buf_type = CC_DMA_BUF_MLLI;
+		}
+	}
+
+	if (areq_ctx->data_dma_buf_type == CC_DMA_BUF_MLLI) {
+		mlli_params->curr_pool = buff_mgr->mlli_buffs_pool;
+		/* add the src data to the sg_data */
+		cc_add_sg_entry(dev, &sg_data, areq_ctx->in_nents, src,
+				(update_data_len - *curr_buff_cnt), 0, true,
+				&areq_ctx->mlli_nents);
+		if (cc_generate_mlli(dev, &sg_data, mlli_params, flags))
+			goto fail_unmap_din;
+	}
+	areq_ctx->buff_index = (areq_ctx->buff_index ^ swap_index);
+
+	return 0;
+
+fail_unmap_din:
+	dma_unmap_sg(dev, src, areq_ctx->in_nents, DMA_TO_DEVICE);
+
+unmap_curr_buff:
+	if (*curr_buff_cnt)
+		dma_unmap_sg(dev, areq_ctx->buff_sg, 1, DMA_TO_DEVICE);
+
+	return -ENOMEM;
+}
+
+void cc_unmap_hash_request(struct device *dev, void *ctx,
+			   struct scatterlist *src, bool do_revert)
+{
+	struct ahash_req_ctx *areq_ctx = (struct ahash_req_ctx *)ctx;
+	u32 *prev_len = cc_next_buf_cnt(areq_ctx);
+
+	/*In case a pool was set, a table was
+	 *allocated and should be released
+	 */
+	if (areq_ctx->mlli_params.curr_pool) {
+		dev_dbg(dev, "free MLLI buffer: dma=%pad virt=%pK\n",
+			&areq_ctx->mlli_params.mlli_dma_addr,
+			areq_ctx->mlli_params.mlli_virt_addr);
+		dma_pool_free(areq_ctx->mlli_params.curr_pool,
+			      areq_ctx->mlli_params.mlli_virt_addr,
+			      areq_ctx->mlli_params.mlli_dma_addr);
+	}
+
+	if (src && areq_ctx->in_nents) {
+		dev_dbg(dev, "Unmapped sg src: virt=%pK dma=%pad len=0x%X\n",
+			sg_virt(src), &sg_dma_address(src), sg_dma_len(src));
+		dma_unmap_sg(dev, src,
+			     areq_ctx->in_nents, DMA_TO_DEVICE);
+	}
+
+	if (*prev_len) {
+		dev_dbg(dev, "Unmapped buffer: areq_ctx->buff_sg=%pK dma=%pad len 0x%X\n",
+			sg_virt(areq_ctx->buff_sg),
+			&sg_dma_address(areq_ctx->buff_sg),
+			sg_dma_len(areq_ctx->buff_sg));
+		dma_unmap_sg(dev, areq_ctx->buff_sg, 1, DMA_TO_DEVICE);
+		if (!do_revert) {
+			/* clean the previous data length for update
+			 * operation
+			 */
+			*prev_len = 0;
+		} else {
+			areq_ctx->buff_index ^= 1;
+		}
+	}
+}
+
 int cc_buffer_mgr_init(struct cc_drvdata *drvdata)
 {
 	struct buff_mgr_handle *buff_mgr_handle;

drivers/crypto/ccree/cc_driver.c

@@ -20,6 +20,7 @@
 #include "cc_buffer_mgr.h"
 #include "cc_debugfs.h"
 #include "cc_cipher.h"
+#include "cc_hash.h"
 #include "cc_ivgen.h"
 #include "cc_sram_mgr.h"
 #include "cc_pm.h"
@@ -286,8 +287,17 @@ static int init_cc_resources(struct platform_device *plat_dev)
 		goto post_ivgen_err;
 	}
 
+	/* hash must be allocated before aead since hash exports APIs */
+	rc = cc_hash_alloc(new_drvdata);
+	if (rc) {
+		dev_err(dev, "cc_hash_alloc failed\n");
+		goto post_cipher_err;
+	}
+
 	return 0;
 
+post_cipher_err:
+	cc_cipher_free(new_drvdata);
 post_ivgen_err:
 	cc_ivgen_fini(new_drvdata);
 post_power_mgr_err:
@@ -318,6 +328,7 @@ static void cleanup_cc_resources(struct platform_device *plat_dev)
 	struct cc_drvdata *drvdata =
 		(struct cc_drvdata *)platform_get_drvdata(plat_dev);
 
+	cc_hash_free(drvdata);
 	cc_cipher_free(drvdata);
 	cc_ivgen_fini(drvdata);
 	cc_pm_fini(drvdata);
@@ -406,6 +417,8 @@ static int __init ccree_init(void)
 {
 	int ret;
 
+	cc_hash_global_init();
+
 	ret = cc_debugfs_global_init();
 	if (ret)
 		return ret;

drivers/crypto/ccree/cc_driver.h

@@ -113,6 +113,7 @@ struct cc_drvdata {
 	cc_sram_addr_t mlli_sram_addr;
 	void *buff_mgr_handle;
 	void *cipher_handle;
+	void *hash_handle;
 	void *request_mgr_handle;
 	void *ivgen_handle;
 	void *sram_mgr_handle;

drivers/crypto/ccree/cc_hash.h

@@ -0,0 +1,113 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (C) 2012-2018 ARM Limited or its affiliates. */
+
+/* \file cc_hash.h
+ * ARM CryptoCell Hash Crypto API
+ */
+
+#ifndef __CC_HASH_H__
+#define __CC_HASH_H__
+
+#include "cc_buffer_mgr.h"
+
+#define HMAC_IPAD_CONST	0x36363636
+#define HMAC_OPAD_CONST	0x5C5C5C5C
+#if (CC_DEV_SHA_MAX > 256)
+#define HASH_LEN_SIZE 16
+#define CC_MAX_HASH_DIGEST_SIZE	SHA512_DIGEST_SIZE
+#define CC_MAX_HASH_BLCK_SIZE SHA512_BLOCK_SIZE
+#else
+#define HASH_LEN_SIZE 8
+#define CC_MAX_HASH_DIGEST_SIZE	SHA256_DIGEST_SIZE
+#define CC_MAX_HASH_BLCK_SIZE SHA256_BLOCK_SIZE
+#endif
+
+#define XCBC_MAC_K1_OFFSET 0
+#define XCBC_MAC_K2_OFFSET 16
+#define XCBC_MAC_K3_OFFSET 32
+
+#define CC_EXPORT_MAGIC 0xC2EE1070U
+
+/* this struct was taken from drivers/crypto/nx/nx-aes-xcbc.c and it is used
+ * for xcbc/cmac statesize
+ */
+struct aeshash_state {
+	u8 state[AES_BLOCK_SIZE];
+	unsigned int count;
+	u8 buffer[AES_BLOCK_SIZE];
+};
+
+/* ahash state */
+struct ahash_req_ctx {
+	u8 buffers[2][CC_MAX_HASH_BLCK_SIZE] ____cacheline_aligned;
+	u8 digest_result_buff[CC_MAX_HASH_DIGEST_SIZE] ____cacheline_aligned;
+	u8 digest_buff[CC_MAX_HASH_DIGEST_SIZE] ____cacheline_aligned;
+	u8 opad_digest_buff[CC_MAX_HASH_DIGEST_SIZE] ____cacheline_aligned;
+	u8 digest_bytes_len[HASH_LEN_SIZE] ____cacheline_aligned;
+	struct async_gen_req_ctx gen_ctx ____cacheline_aligned;
+	enum cc_req_dma_buf_type data_dma_buf_type;
+	dma_addr_t opad_digest_dma_addr;
+	dma_addr_t digest_buff_dma_addr;
+	dma_addr_t digest_bytes_len_dma_addr;
+	dma_addr_t digest_result_dma_addr;
+	u32 buf_cnt[2];
+	u32 buff_index;
+	u32 xcbc_count; /* count xcbc update operatations */
+	struct scatterlist buff_sg[2];
+	struct scatterlist *curr_sg;
+	u32 in_nents;
+	u32 mlli_nents;
+	struct mlli_params mlli_params;
+};
+
+static inline u32 *cc_hash_buf_cnt(struct ahash_req_ctx *state)
+{
+	return &state->buf_cnt[state->buff_index];
+}
+
+static inline u8 *cc_hash_buf(struct ahash_req_ctx *state)
+{
+	return state->buffers[state->buff_index];
+}
+
+static inline u32 *cc_next_buf_cnt(struct ahash_req_ctx *state)
+{
+	return &state->buf_cnt[state->buff_index ^ 1];
+}
+
+static inline u8 *cc_next_buf(struct ahash_req_ctx *state)
+{
+	return state->buffers[state->buff_index ^ 1];
+}
+
+int cc_hash_alloc(struct cc_drvdata *drvdata);
+int cc_init_hash_sram(struct cc_drvdata *drvdata);
+int cc_hash_free(struct cc_drvdata *drvdata);
+
+/*!
+ * Gets the initial digest length
+ *
+ * \param drvdata
+ * \param mode The Hash mode. Supported modes:
+ *             MD5/SHA1/SHA224/SHA256/SHA384/SHA512
+ *
+ * \return u32 returns the address of the initial digest length in SRAM
+ */
+cc_sram_addr_t
+cc_digest_len_addr(void *drvdata, u32 mode);
+
+/*!
+ * Gets the address of the initial digest in SRAM
+ * according to the given hash mode
+ *
+ * \param drvdata
+ * \param mode The Hash mode. Supported modes:
+ *             MD5/SHA1/SHA224/SHA256/SHA384/SHA512
+ *
+ * \return u32 The address of the initial digest in SRAM
+ */
+cc_sram_addr_t cc_larval_digest_addr(void *drvdata, u32 mode);
+
+void cc_hash_global_init(void);
+
+#endif /*__CC_HASH_H__*/

drivers/crypto/ccree/cc_pm.c

@@ -9,6 +9,7 @@
 #include "cc_request_mgr.h"
 #include "cc_sram_mgr.h"
 #include "cc_ivgen.h"
+#include "cc_hash.h"
 #include "cc_pm.h"
 
 #define POWER_DOWN_ENABLE 0x01
@@ -61,6 +62,9 @@ int cc_pm_resume(struct device *dev)
 		return rc;
 	}
 
+	/* must be after the queue resuming as it uses the HW queue*/
+	cc_init_hash_sram(drvdata);
+
 	cc_init_iv_sram(drvdata);
 	return 0;
 }