crypto: ccp - crypto API interface to the CCP device driver

These routines provide the support for the interface between the crypto API
and the AMD CCP. This includes insuring that requests associated with a
given tfm on the same cpu are processed in the order received.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

drivers/crypto/ccp/ccp-crypto-main.c

+/*
+ * AMD Cryptographic Coprocessor (CCP) crypto API support
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/ccp.h>
+#include <linux/scatterlist.h>
+#include <crypto/internal/hash.h>
+
+#include "ccp-crypto.h"
+
+MODULE_AUTHOR("Tom Lendacky <thomas.lendacky@amd.com>");
+MODULE_LICENSE("GPL");
+MODULE_VERSION("1.0.0");
+MODULE_DESCRIPTION("AMD Cryptographic Coprocessor crypto API support");
+
+
+/* List heads for the supported algorithms */
+static LIST_HEAD(hash_algs);
+static LIST_HEAD(cipher_algs);
+
+/* For any tfm, requests for that tfm on the same CPU must be returned
+ * in the order received.  With multiple queues available, the CCP can
+ * process more than one cmd at a time.  Therefore we must maintain
+ * a cmd list to insure the proper ordering of requests on a given tfm/cpu
+ * combination.
+ */
+struct ccp_crypto_cpu_queue {
+	struct list_head cmds;
+	struct list_head *backlog;
+	unsigned int cmd_count;
+};
+#define CCP_CRYPTO_MAX_QLEN	50
+
+struct ccp_crypto_percpu_queue {
+	struct ccp_crypto_cpu_queue __percpu *cpu_queue;
+};
+static struct ccp_crypto_percpu_queue req_queue;
+
+struct ccp_crypto_cmd {
+	struct list_head entry;
+
+	struct ccp_cmd *cmd;
+
+	/* Save the crypto_tfm and crypto_async_request addresses
+	 * separately to avoid any reference to a possibly invalid
+	 * crypto_async_request structure after invoking the request
+	 * callback
+	 */
+	struct crypto_async_request *req;
+	struct crypto_tfm *tfm;
+
+	/* Used for held command processing to determine state */
+	int ret;
+
+	int cpu;
+};
+
+struct ccp_crypto_cpu {
+	struct work_struct work;
+	struct completion completion;
+	struct ccp_crypto_cmd *crypto_cmd;
+	int err;
+};
+
+
+static inline bool ccp_crypto_success(int err)
+{
+	if (err && (err != -EINPROGRESS) && (err != -EBUSY))
+		return false;
+
+	return true;
+}
+
+/*
+ * ccp_crypto_cmd_complete must be called while running on the appropriate
+ * cpu and the caller must have done a get_cpu to disable preemption
+ */
+static struct ccp_crypto_cmd *ccp_crypto_cmd_complete(
+	struct ccp_crypto_cmd *crypto_cmd, struct ccp_crypto_cmd **backlog)
+{
+	struct ccp_crypto_cpu_queue *cpu_queue;
+	struct ccp_crypto_cmd *held = NULL, *tmp;
+
+	*backlog = NULL;
+
+	cpu_queue = this_cpu_ptr(req_queue.cpu_queue);
+
+	/* Held cmds will be after the current cmd in the queue so start
+	 * searching for a cmd with a matching tfm for submission.
+	 */
+	tmp = crypto_cmd;
+	list_for_each_entry_continue(tmp, &cpu_queue->cmds, entry) {
+		if (crypto_cmd->tfm != tmp->tfm)
+			continue;
+		held = tmp;
+		break;
+	}
+
+	/* Process the backlog:
+	 *   Because cmds can be executed from any point in the cmd list
+	 *   special precautions have to be taken when handling the backlog.
+	 */
+	if (cpu_queue->backlog != &cpu_queue->cmds) {
+		/* Skip over this cmd if it is the next backlog cmd */
+		if (cpu_queue->backlog == &crypto_cmd->entry)
+			cpu_queue->backlog = crypto_cmd->entry.next;
+
+		*backlog = container_of(cpu_queue->backlog,
+					struct ccp_crypto_cmd, entry);
+		cpu_queue->backlog = cpu_queue->backlog->next;
+
+		/* Skip over this cmd if it is now the next backlog cmd */
+		if (cpu_queue->backlog == &crypto_cmd->entry)
+			cpu_queue->backlog = crypto_cmd->entry.next;
+	}
+
+	/* Remove the cmd entry from the list of cmds */
+	cpu_queue->cmd_count--;
+	list_del(&crypto_cmd->entry);
+
+	return held;
+}
+
+static void ccp_crypto_complete_on_cpu(struct work_struct *work)
+{
+	struct ccp_crypto_cpu *cpu_work =
+		container_of(work, struct ccp_crypto_cpu, work);
+	struct ccp_crypto_cmd *crypto_cmd = cpu_work->crypto_cmd;
+	struct ccp_crypto_cmd *held, *next, *backlog;
+	struct crypto_async_request *req = crypto_cmd->req;
+	struct ccp_ctx *ctx = crypto_tfm_ctx(req->tfm);
+	int cpu, ret;
+
+	cpu = get_cpu();
+
+	if (cpu_work->err == -EINPROGRESS) {
+		/* Only propogate the -EINPROGRESS if necessary */
+		if (crypto_cmd->ret == -EBUSY) {
+			crypto_cmd->ret = -EINPROGRESS;
+			req->complete(req, -EINPROGRESS);
+		}
+
+		goto e_cpu;
+	}
+
+	/* Operation has completed - update the queue before invoking
+	 * the completion callbacks and retrieve the next cmd (cmd with
+	 * a matching tfm) that can be submitted to the CCP.
+	 */
+	held = ccp_crypto_cmd_complete(crypto_cmd, &backlog);
+	if (backlog) {
+		backlog->ret = -EINPROGRESS;
+		backlog->req->complete(backlog->req, -EINPROGRESS);
+	}
+
+	/* Transition the state from -EBUSY to -EINPROGRESS first */
+	if (crypto_cmd->ret == -EBUSY)
+		req->complete(req, -EINPROGRESS);
+
+	/* Completion callbacks */
+	ret = cpu_work->err;
+	if (ctx->complete)
+		ret = ctx->complete(req, ret);
+	req->complete(req, ret);
+
+	/* Submit the next cmd */
+	while (held) {
+		ret = ccp_enqueue_cmd(held->cmd);
+		if (ccp_crypto_success(ret))
+			break;
+
+		/* Error occurred, report it and get the next entry */
+		held->req->complete(held->req, ret);
+
+		next = ccp_crypto_cmd_complete(held, &backlog);
+		if (backlog) {
+			backlog->ret = -EINPROGRESS;
+			backlog->req->complete(backlog->req, -EINPROGRESS);
+		}
+
+		kfree(held);
+		held = next;
+	}
+
+	kfree(crypto_cmd);
+
+e_cpu:
+	put_cpu();
+
+	complete(&cpu_work->completion);
+}
+
+static void ccp_crypto_complete(void *data, int err)
+{
+	struct ccp_crypto_cmd *crypto_cmd = data;
+	struct ccp_crypto_cpu cpu_work;
+
+	INIT_WORK(&cpu_work.work, ccp_crypto_complete_on_cpu);
+	init_completion(&cpu_work.completion);
+	cpu_work.crypto_cmd = crypto_cmd;
+	cpu_work.err = err;
+
+	schedule_work_on(crypto_cmd->cpu, &cpu_work.work);
+
+	/* Keep the completion call synchronous */
+	wait_for_completion(&cpu_work.completion);
+}
+
+static int ccp_crypto_enqueue_cmd(struct ccp_crypto_cmd *crypto_cmd)
+{
+	struct ccp_crypto_cpu_queue *cpu_queue;
+	struct ccp_crypto_cmd *active = NULL, *tmp;
+	int cpu, ret;
+
+	cpu = get_cpu();
+	crypto_cmd->cpu = cpu;
+
+	cpu_queue = this_cpu_ptr(req_queue.cpu_queue);
+
+	/* Check if the cmd can/should be queued */
+	if (cpu_queue->cmd_count >= CCP_CRYPTO_MAX_QLEN) {
+		ret = -EBUSY;
+		if (!(crypto_cmd->cmd->flags & CCP_CMD_MAY_BACKLOG))
+			goto e_cpu;
+	}
+
+	/* Look for an entry with the same tfm.  If there is a cmd
+	 * with the same tfm in the list for this cpu then the current
+	 * cmd cannot be submitted to the CCP yet.
+	 */
+	list_for_each_entry(tmp, &cpu_queue->cmds, entry) {
+		if (crypto_cmd->tfm != tmp->tfm)
+			continue;
+		active = tmp;
+		break;
+	}
+
+	ret = -EINPROGRESS;
+	if (!active) {
+		ret = ccp_enqueue_cmd(crypto_cmd->cmd);
+		if (!ccp_crypto_success(ret))
+			goto e_cpu;
+	}
+
+	if (cpu_queue->cmd_count >= CCP_CRYPTO_MAX_QLEN) {
+		ret = -EBUSY;
+		if (cpu_queue->backlog == &cpu_queue->cmds)
+			cpu_queue->backlog = &crypto_cmd->entry;
+	}
+	crypto_cmd->ret = ret;
+
+	cpu_queue->cmd_count++;
+	list_add_tail(&crypto_cmd->entry, &cpu_queue->cmds);
+
+e_cpu:
+	put_cpu();
+
+	return ret;
+}
+
+/**
+ * ccp_crypto_enqueue_request - queue an crypto async request for processing
+ *				by the CCP
+ *
+ * @req: crypto_async_request struct to be processed
+ * @cmd: ccp_cmd struct to be sent to the CCP
+ */
+int ccp_crypto_enqueue_request(struct crypto_async_request *req,
+			       struct ccp_cmd *cmd)
+{
+	struct ccp_crypto_cmd *crypto_cmd;
+	gfp_t gfp;
+	int ret;
+
+	gfp = req->flags & CRYPTO_TFM_REQ_MAY_SLEEP ? GFP_KERNEL : GFP_ATOMIC;
+
+	crypto_cmd = kzalloc(sizeof(*crypto_cmd), gfp);
+	if (!crypto_cmd)
+		return -ENOMEM;
+
+	/* The tfm pointer must be saved and not referenced from the
+	 * crypto_async_request (req) pointer because it is used after
+	 * completion callback for the request and the req pointer
+	 * might not be valid anymore.
+	 */
+	crypto_cmd->cmd = cmd;
+	crypto_cmd->req = req;
+	crypto_cmd->tfm = req->tfm;
+
+	cmd->callback = ccp_crypto_complete;
+	cmd->data = crypto_cmd;
+
+	if (req->flags & CRYPTO_TFM_REQ_MAY_BACKLOG)
+		cmd->flags |= CCP_CMD_MAY_BACKLOG;
+	else
+		cmd->flags &= ~CCP_CMD_MAY_BACKLOG;
+
+	ret = ccp_crypto_enqueue_cmd(crypto_cmd);
+	if (!ccp_crypto_success(ret))
+		kfree(crypto_cmd);
+
+	return ret;
+}
+
+struct scatterlist *ccp_crypto_sg_table_add(struct sg_table *table,
+					    struct scatterlist *sg_add)
+{
+	struct scatterlist *sg, *sg_last = NULL;
+
+	for (sg = table->sgl; sg; sg = sg_next(sg))
+		if (!sg_page(sg))
+			break;
+	BUG_ON(!sg);
+
+	for (; sg && sg_add; sg = sg_next(sg), sg_add = sg_next(sg_add)) {
+		sg_set_page(sg, sg_page(sg_add), sg_add->length,
+			    sg_add->offset);
+		sg_last = sg;
+	}
+	BUG_ON(sg_add);
+
+	return sg_last;
+}
+
+static int ccp_register_algs(void)
+{
+	int ret;
+
+	ret = ccp_register_aes_algs(&cipher_algs);
+	if (ret)
+		return ret;
+
+	ret = ccp_register_aes_cmac_algs(&hash_algs);
+	if (ret)
+		return ret;
+
+	ret = ccp_register_aes_xts_algs(&cipher_algs);
+	if (ret)
+		return ret;
+
+	ret = ccp_register_sha_algs(&hash_algs);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void ccp_unregister_algs(void)
+{
+	struct ccp_crypto_ahash_alg *ahash_alg, *ahash_tmp;
+	struct ccp_crypto_ablkcipher_alg *ablk_alg, *ablk_tmp;
+
+	list_for_each_entry_safe(ahash_alg, ahash_tmp, &hash_algs, entry) {
+		crypto_unregister_ahash(&ahash_alg->alg);
+		list_del(&ahash_alg->entry);
+		kfree(ahash_alg);
+	}
+
+	list_for_each_entry_safe(ablk_alg, ablk_tmp, &cipher_algs, entry) {
+		crypto_unregister_alg(&ablk_alg->alg);
+		list_del(&ablk_alg->entry);
+		kfree(ablk_alg);
+	}
+}
+
+static int ccp_init_queues(void)
+{
+	struct ccp_crypto_cpu_queue *cpu_queue;
+	int cpu;
+
+	req_queue.cpu_queue = alloc_percpu(struct ccp_crypto_cpu_queue);
+	if (!req_queue.cpu_queue)
+		return -ENOMEM;
+
+	for_each_possible_cpu(cpu) {
+		cpu_queue = per_cpu_ptr(req_queue.cpu_queue, cpu);
+		INIT_LIST_HEAD(&cpu_queue->cmds);
+		cpu_queue->backlog = &cpu_queue->cmds;
+		cpu_queue->cmd_count = 0;
+	}
+
+	return 0;
+}
+
+static void ccp_fini_queue(void)
+{
+	struct ccp_crypto_cpu_queue *cpu_queue;
+	int cpu;
+
+	for_each_possible_cpu(cpu) {
+		cpu_queue = per_cpu_ptr(req_queue.cpu_queue, cpu);
+		BUG_ON(!list_empty(&cpu_queue->cmds));
+	}
+	free_percpu(req_queue.cpu_queue);
+}
+
+static int ccp_crypto_init(void)
+{
+	int ret;
+
+	ret = ccp_init_queues();
+	if (ret)
+		return ret;
+
+	ret = ccp_register_algs();
+	if (ret) {
+		ccp_unregister_algs();
+		ccp_fini_queue();
+	}
+
+	return ret;
+}
+
+static void ccp_crypto_exit(void)
+{
+	ccp_unregister_algs();
+	ccp_fini_queue();
+}
+
+module_init(ccp_crypto_init);
+module_exit(ccp_crypto_exit);

drivers/crypto/ccp/ccp-crypto.h

+/*
+ * AMD Cryptographic Coprocessor (CCP) crypto API support
+ *
+ * Copyright (C) 2013 Advanced Micro Devices, Inc.
+ *
+ * Author: Tom Lendacky <thomas.lendacky@amd.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef __CCP_CRYPTO_H__
+#define __CCP_CRYPTO_H__
+
+
+#include <linux/list.h>
+#include <linux/wait.h>
+#include <linux/pci.h>
+#include <linux/ccp.h>
+#include <linux/crypto.h>
+#include <crypto/algapi.h>
+#include <crypto/aes.h>
+#include <crypto/ctr.h>
+#include <crypto/hash.h>
+#include <crypto/sha.h>
+
+
+#define CCP_CRA_PRIORITY	300
+
+struct ccp_crypto_ablkcipher_alg {
+	struct list_head entry;
+
+	u32 mode;
+
+	struct crypto_alg alg;
+};
+
+struct ccp_crypto_ahash_alg {
+	struct list_head entry;
+
+	const u32 *init;
+	u32 type;
+	u32 mode;
+
+	/* Child algorithm used for HMAC, CMAC, etc */
+	char child_alg[CRYPTO_MAX_ALG_NAME];
+
+	struct ahash_alg alg;
+};
+
+static inline struct ccp_crypto_ablkcipher_alg *
+	ccp_crypto_ablkcipher_alg(struct crypto_tfm *tfm)
+{
+	struct crypto_alg *alg = tfm->__crt_alg;
+
+	return container_of(alg, struct ccp_crypto_ablkcipher_alg, alg);
+}
+
+static inline struct ccp_crypto_ahash_alg *
+	ccp_crypto_ahash_alg(struct crypto_tfm *tfm)
+{
+	struct crypto_alg *alg = tfm->__crt_alg;
+	struct ahash_alg *ahash_alg;
+
+	ahash_alg = container_of(alg, struct ahash_alg, halg.base);
+
+	return container_of(ahash_alg, struct ccp_crypto_ahash_alg, alg);
+}
+
+
+/***** AES related defines *****/
+struct ccp_aes_ctx {
+	/* Fallback cipher for XTS with unsupported unit sizes */
+	struct crypto_ablkcipher *tfm_ablkcipher;
+
+	/* Cipher used to generate CMAC K1/K2 keys */
+	struct crypto_cipher *tfm_cipher;
+
+	enum ccp_engine engine;
+	enum ccp_aes_type type;
+	enum ccp_aes_mode mode;
+
+	struct scatterlist key_sg;
+	unsigned int key_len;
+	u8 key[AES_MAX_KEY_SIZE];
+
+	u8 nonce[CTR_RFC3686_NONCE_SIZE];
+
+	/* CMAC key structures */
+	struct scatterlist k1_sg;
+	struct scatterlist k2_sg;
+	unsigned int kn_len;
+	u8 k1[AES_BLOCK_SIZE];
+	u8 k2[AES_BLOCK_SIZE];
+};
+
+struct ccp_aes_req_ctx {
+	struct scatterlist iv_sg;
+	u8 iv[AES_BLOCK_SIZE];
+
+	/* Fields used for RFC3686 requests */
+	u8 *rfc3686_info;
+	u8 rfc3686_iv[AES_BLOCK_SIZE];
+
+	struct ccp_cmd cmd;
+};
+
+struct ccp_aes_cmac_req_ctx {
+	unsigned int null_msg;
+	unsigned int final;
+
+	unsigned int hash_cnt;
+	unsigned int hash_rem;
+
+	struct sg_table data_sg;
+
+	struct scatterlist iv_sg;
+	u8 iv[AES_BLOCK_SIZE];
+
+	struct scatterlist buf_sg;
+	unsigned int buf_count;
+	u8 buf[AES_BLOCK_SIZE];
+
+	struct scatterlist pad_sg;
+	unsigned int pad_count;
+	u8 pad[AES_BLOCK_SIZE];
+
+	struct ccp_cmd cmd;
+};
+
+/***** SHA related defines *****/
+#define MAX_SHA_CONTEXT_SIZE	SHA256_DIGEST_SIZE
+#define MAX_SHA_BLOCK_SIZE	SHA256_BLOCK_SIZE
+
+struct ccp_sha_ctx {
+	unsigned int key_len;
+	u8 key[MAX_SHA_BLOCK_SIZE];
+	u8 ipad[MAX_SHA_BLOCK_SIZE];
+	u8 opad[MAX_SHA_BLOCK_SIZE];
+	struct crypto_ahash *hmac_tfm;
+};
+
+struct ccp_sha_req_ctx {
+	enum ccp_sha_type type;
+
+	u64 msg_bits;
+
+	unsigned int first;
+	unsigned int final;
+
+	unsigned int hash_cnt;
+	unsigned int hash_rem;
+
+	struct sg_table data_sg;
+
+	struct scatterlist ctx_sg;
+	u8 ctx[MAX_SHA_CONTEXT_SIZE];
+
+	struct scatterlist buf_sg;
+	unsigned int buf_count;
+	u8 buf[MAX_SHA_BLOCK_SIZE];
+
+	/* HMAC support field */
+	struct scatterlist pad_sg;
+
+	/* CCP driver command */
+	struct ccp_cmd cmd;
+};
+
+/***** Common Context Structure *****/
+struct ccp_ctx {
+	int (*complete)(struct crypto_async_request *req, int ret);
+
+	union {
+		struct ccp_aes_ctx aes;
+		struct ccp_sha_ctx sha;
+	} u;
+};
+
+int ccp_crypto_enqueue_request(struct crypto_async_request *req,
+			       struct ccp_cmd *cmd);
+struct scatterlist *ccp_crypto_sg_table_add(struct sg_table *table,
+					    struct scatterlist *sg_add);
+
+int ccp_register_aes_algs(struct list_head *head);
+int ccp_register_aes_cmac_algs(struct list_head *head);
+int ccp_register_aes_xts_algs(struct list_head *head);
+int ccp_register_sha_algs(struct list_head *head);
+
+#endif