1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
|
/*
* caam - Freescale FSL CAAM support for hw_random
*
* Copyright 2011 Freescale Semiconductor, Inc.
*
* Based on caamalg.c crypto API driver.
*
* relationship between job descriptors to shared descriptors:
*
* --------------- --------------
* | JobDesc #0 |-------------------->| ShareDesc |
* | *(buffer 0) | |------------->| (generate) |
* --------------- | | (move) |
* | | (store) |
* --------------- | --------------
* | JobDesc #1 |------|
* | *(buffer 1) |
* ---------------
*
* A job desc looks like this:
*
* ---------------------
* | Header |
* | ShareDesc Pointer |
* | SEQ_OUT_PTR |
* | (output buffer) |
* ---------------------
*
* The SharedDesc never changes, and each job descriptor points to one of two
* buffers for each device, from which the data will be copied into the
* requested destination
*/
#include <linux/hw_random.h>
#include <linux/completion.h>
#include <linux/atomic.h>
#include "compat.h"
#include "regs.h"
#include "intern.h"
#include "desc_constr.h"
#include "jr.h"
#include "error.h"
/*
* Maximum buffer size: maximum number of random, cache-aligned bytes that
* will be generated and moved to seq out ptr (extlen not allowed)
*/
#define RN_BUF_SIZE (0xffff / L1_CACHE_BYTES * \
L1_CACHE_BYTES)
/* length of descriptors */
#define DESC_JOB_O_LEN (CAAM_CMD_SZ * 2 + CAAM_PTR_SZ * 2)
#define DESC_RNG_LEN (4 * CAAM_CMD_SZ)
/* Buffer, its dma address and lock */
struct buf_data {
u8 buf[RN_BUF_SIZE];
dma_addr_t addr;
struct completion filled;
u32 hw_desc[DESC_JOB_O_LEN];
#define BUF_NOT_EMPTY 0
#define BUF_EMPTY 1
#define BUF_PENDING 2 /* Empty, but with job pending --don't submit another */
atomic_t empty;
};
/* rng per-device context */
struct caam_rng_ctx {
struct device *jrdev;
dma_addr_t sh_desc_dma;
u32 sh_desc[DESC_RNG_LEN];
unsigned int cur_buf_idx;
int current_buf;
struct buf_data bufs[2];
};
static struct caam_rng_ctx *rng_ctx;
static inline void rng_unmap_buf(struct device *jrdev, struct buf_data *bd)
{
if (bd->addr)
dma_unmap_single(jrdev, bd->addr, RN_BUF_SIZE,
DMA_FROM_DEVICE);
}
static inline void rng_unmap_ctx(struct caam_rng_ctx *ctx)
{
struct device *jrdev = ctx->jrdev;
if (ctx->sh_desc_dma)
dma_unmap_single(jrdev, ctx->sh_desc_dma,
desc_bytes(ctx->sh_desc), DMA_TO_DEVICE);
rng_unmap_buf(jrdev, &ctx->bufs[0]);
rng_unmap_buf(jrdev, &ctx->bufs[1]);
}
static void rng_done(struct device *jrdev, u32 *desc, u32 err, void *context)
{
struct buf_data *bd;
bd = (struct buf_data *)((char *)desc -
offsetof(struct buf_data, hw_desc));
if (err)
caam_jr_strstatus(jrdev, err);
atomic_set(&bd->empty, BUF_NOT_EMPTY);
complete(&bd->filled);
#ifdef DEBUG
print_hex_dump(KERN_ERR, "rng refreshed buf@: ",
DUMP_PREFIX_ADDRESS, 16, 4, bd->buf, RN_BUF_SIZE, 1);
#endif
}
static inline int submit_job(struct caam_rng_ctx *ctx, int to_current)
{
struct buf_data *bd = &ctx->bufs[!(to_current ^ ctx->current_buf)];
struct device *jrdev = ctx->jrdev;
u32 *desc = bd->hw_desc;
int err;
dev_dbg(jrdev, "submitting job %d\n", !(to_current ^ ctx->current_buf));
init_completion(&bd->filled);
err = caam_jr_enqueue(jrdev, desc, rng_done, ctx);
if (err)
complete(&bd->filled); /* don't wait on failed job*/
else
atomic_inc(&bd->empty); /* note if pending */
return err;
}
static int caam_read(struct hwrng *rng, void *data, size_t max, bool wait)
{
struct caam_rng_ctx *ctx = rng_ctx;
struct buf_data *bd = &ctx->bufs[ctx->current_buf];
int next_buf_idx, copied_idx;
int err;
if (atomic_read(&bd->empty)) {
/* try to submit job if there wasn't one */
if (atomic_read(&bd->empty) == BUF_EMPTY) {
err = submit_job(ctx, 1);
/* if can't submit job, can't even wait */
if (err)
return 0;
}
/* no immediate data, so exit if not waiting */
if (!wait)
return 0;
/* waiting for pending job */
if (atomic_read(&bd->empty))
wait_for_completion(&bd->filled);
}
next_buf_idx = ctx->cur_buf_idx + max;
dev_dbg(ctx->jrdev, "%s: start reading at buffer %d, idx %d\n",
__func__, ctx->current_buf, ctx->cur_buf_idx);
/* if enough data in current buffer */
if (next_buf_idx < RN_BUF_SIZE) {
memcpy(data, bd->buf + ctx->cur_buf_idx, max);
ctx->cur_buf_idx = next_buf_idx;
return max;
}
/* else, copy what's left... */
copied_idx = RN_BUF_SIZE - ctx->cur_buf_idx;
memcpy(data, bd->buf + ctx->cur_buf_idx, copied_idx);
ctx->cur_buf_idx = 0;
atomic_set(&bd->empty, BUF_EMPTY);
/* ...refill... */
submit_job(ctx, 1);
/* and use next buffer */
ctx->current_buf = !ctx->current_buf;
dev_dbg(ctx->jrdev, "switched to buffer %d\n", ctx->current_buf);
/* since there already is some data read, don't wait */
return copied_idx + caam_read(rng, data + copied_idx,
max - copied_idx, false);
}
static inline int rng_create_sh_desc(struct caam_rng_ctx *ctx)
{
struct device *jrdev = ctx->jrdev;
u32 *desc = ctx->sh_desc;
init_sh_desc(desc, HDR_SHARE_SERIAL);
/* Propagate errors from shared to job descriptor */
append_cmd(desc, SET_OK_NO_PROP_ERRORS | CMD_LOAD);
/* Generate random bytes */
append_operation(desc, OP_ALG_ALGSEL_RNG | OP_TYPE_CLASS1_ALG);
/* Store bytes */
append_seq_fifo_store(desc, RN_BUF_SIZE, FIFOST_TYPE_RNGSTORE);
ctx->sh_desc_dma = dma_map_single(jrdev, desc, desc_bytes(desc),
DMA_TO_DEVICE);
if (dma_mapping_error(jrdev, ctx->sh_desc_dma)) {
dev_err(jrdev, "unable to map shared descriptor\n");
return -ENOMEM;
}
#ifdef DEBUG
print_hex_dump(KERN_ERR, "rng shdesc@: ", DUMP_PREFIX_ADDRESS, 16, 4,
desc, desc_bytes(desc), 1);
#endif
return 0;
}
static inline int rng_create_job_desc(struct caam_rng_ctx *ctx, int buf_id)
{
struct device *jrdev = ctx->jrdev;
struct buf_data *bd = &ctx->bufs[buf_id];
u32 *desc = bd->hw_desc;
int sh_len = desc_len(ctx->sh_desc);
init_job_desc_shared(desc, ctx->sh_desc_dma, sh_len, HDR_SHARE_DEFER |
HDR_REVERSE);
bd->addr = dma_map_single(jrdev, bd->buf, RN_BUF_SIZE, DMA_FROM_DEVICE);
if (dma_mapping_error(jrdev, bd->addr)) {
dev_err(jrdev, "unable to map dst\n");
return -ENOMEM;
}
append_seq_out_ptr_intlen(desc, bd->addr, RN_BUF_SIZE, 0);
#ifdef DEBUG
print_hex_dump(KERN_ERR, "rng job desc@: ", DUMP_PREFIX_ADDRESS, 16, 4,
desc, desc_bytes(desc), 1);
#endif
return 0;
}
static void caam_cleanup(struct hwrng *rng)
{
int i;
struct buf_data *bd;
for (i = 0; i < 2; i++) {
bd = &rng_ctx->bufs[i];
if (atomic_read(&bd->empty) == BUF_PENDING)
wait_for_completion(&bd->filled);
}
rng_unmap_ctx(rng_ctx);
}
static int caam_init_buf(struct caam_rng_ctx *ctx, int buf_id)
{
struct buf_data *bd = &ctx->bufs[buf_id];
int err;
err = rng_create_job_desc(ctx, buf_id);
if (err)
return err;
atomic_set(&bd->empty, BUF_EMPTY);
submit_job(ctx, buf_id == ctx->current_buf);
wait_for_completion(&bd->filled);
return 0;
}
static int caam_init_rng(struct caam_rng_ctx *ctx, struct device *jrdev)
{
int err;
ctx->jrdev = jrdev;
err = rng_create_sh_desc(ctx);
if (err)
return err;
ctx->current_buf = 0;
ctx->cur_buf_idx = 0;
err = caam_init_buf(ctx, 0);
if (err)
return err;
err = caam_init_buf(ctx, 1);
if (err)
return err;
return 0;
}
static struct hwrng caam_rng = {
.name = "rng-caam",
.cleanup = caam_cleanup,
.read = caam_read,
};
static void __exit caam_rng_exit(void)
{
caam_jr_free(rng_ctx->jrdev);
hwrng_unregister(&caam_rng);
kfree(rng_ctx);
}
static int __init caam_rng_init(void)
{
struct device *dev;
struct device_node *dev_node;
struct platform_device *pdev;
struct device *ctrldev;
void *priv;
int err;
dev_node = of_find_compatible_node(NULL, NULL, "fsl,sec-v4.0");
if (!dev_node) {
dev_node = of_find_compatible_node(NULL, NULL, "fsl,sec4.0");
if (!dev_node)
return -ENODEV;
}
pdev = of_find_device_by_node(dev_node);
if (!pdev) {
of_node_put(dev_node);
return -ENODEV;
}
ctrldev = &pdev->dev;
priv = dev_get_drvdata(ctrldev);
of_node_put(dev_node);
/*
* If priv is NULL, it's probably because the caam driver wasn't
* properly initialized (e.g. RNG4 init failed). Thus, bail out here.
*/
if (!priv)
return -ENODEV;
dev = caam_jr_alloc();
if (IS_ERR(dev)) {
pr_err("Job Ring Device allocation for transform failed\n");
return PTR_ERR(dev);
}
rng_ctx = kmalloc(sizeof(struct caam_rng_ctx), GFP_DMA);
if (!rng_ctx)
return -ENOMEM;
err = caam_init_rng(rng_ctx, dev);
if (err)
return err;
dev_info(dev, "registering rng-caam\n");
return hwrng_register(&caam_rng);
}
module_init(caam_rng_init);
module_exit(caam_rng_exit);
MODULE_LICENSE("GPL");
MODULE_DESCRIPTION("FSL CAAM support for hw_random API");
MODULE_AUTHOR("Freescale Semiconductor - NMG");
|