Free Electrons

Embedded Linux Experts

  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
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
/*
 * Copyright (c) 2004, 2005, 2006 Voltaire, Inc. All rights reserved.
 * Copyright (c) 2013 Mellanox Technologies. All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *	- Redistributions of source code must retain the above
 *	  copyright notice, this list of conditions and the following
 *	  disclaimer.
 *
 *	- Redistributions in binary form must reproduce the above
 *	  copyright notice, this list of conditions and the following
 *	  disclaimer in the documentation and/or other materials
 *	  provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/highmem.h>
#include <linux/scatterlist.h>

#include "iscsi_iser.h"

#define ISER_KMALLOC_THRESHOLD 0x20000 /* 128K - kmalloc limit */

/**
 * iser_start_rdma_unaligned_sg
 */
static int iser_start_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
					enum iser_data_dir cmd_dir)
{
	int dma_nents;
	struct ib_device *dev;
	char *mem = NULL;
	struct iser_data_buf *data = &iser_task->data[cmd_dir];
	unsigned long  cmd_data_len = data->data_len;

	if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
		mem = (void *)__get_free_pages(GFP_ATOMIC,
		      ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
	else
		mem = kmalloc(cmd_data_len, GFP_ATOMIC);

	if (mem == NULL) {
		iser_err("Failed to allocate mem size %d %d for copying sglist\n",
			 data->size,(int)cmd_data_len);
		return -ENOMEM;
	}

	if (cmd_dir == ISER_DIR_OUT) {
		/* copy the unaligned sg the buffer which is used for RDMA */
		struct scatterlist *sgl = (struct scatterlist *)data->buf;
		struct scatterlist *sg;
		int i;
		char *p, *from;

		p = mem;
		for_each_sg(sgl, sg, data->size, i) {
			from = kmap_atomic(sg_page(sg));
			memcpy(p,
			       from + sg->offset,
			       sg->length);
			kunmap_atomic(from);
			p += sg->length;
		}
	}

	sg_init_one(&iser_task->data_copy[cmd_dir].sg_single, mem, cmd_data_len);
	iser_task->data_copy[cmd_dir].buf  =
		&iser_task->data_copy[cmd_dir].sg_single;
	iser_task->data_copy[cmd_dir].size = 1;

	iser_task->data_copy[cmd_dir].copy_buf  = mem;

	dev = iser_task->iser_conn->ib_conn->device->ib_device;
	dma_nents = ib_dma_map_sg(dev,
				  &iser_task->data_copy[cmd_dir].sg_single,
				  1,
				  (cmd_dir == ISER_DIR_OUT) ?
				  DMA_TO_DEVICE : DMA_FROM_DEVICE);
	BUG_ON(dma_nents == 0);

	iser_task->data_copy[cmd_dir].dma_nents = dma_nents;
	return 0;
}

/**
 * iser_finalize_rdma_unaligned_sg
 */
void iser_finalize_rdma_unaligned_sg(struct iscsi_iser_task *iser_task,
				     enum iser_data_dir         cmd_dir)
{
	struct ib_device *dev;
	struct iser_data_buf *mem_copy;
	unsigned long  cmd_data_len;

	dev = iser_task->iser_conn->ib_conn->device->ib_device;
	mem_copy = &iser_task->data_copy[cmd_dir];

	ib_dma_unmap_sg(dev, &mem_copy->sg_single, 1,
			(cmd_dir == ISER_DIR_OUT) ?
			DMA_TO_DEVICE : DMA_FROM_DEVICE);

	if (cmd_dir == ISER_DIR_IN) {
		char *mem;
		struct scatterlist *sgl, *sg;
		unsigned char *p, *to;
		unsigned int sg_size;
		int i;

		/* copy back read RDMA to unaligned sg */
		mem	= mem_copy->copy_buf;

		sgl	= (struct scatterlist *)iser_task->data[ISER_DIR_IN].buf;
		sg_size = iser_task->data[ISER_DIR_IN].size;

		p = mem;
		for_each_sg(sgl, sg, sg_size, i) {
			to = kmap_atomic(sg_page(sg));
			memcpy(to + sg->offset,
			       p,
			       sg->length);
			kunmap_atomic(to);
			p += sg->length;
		}
	}

	cmd_data_len = iser_task->data[cmd_dir].data_len;

	if (cmd_data_len > ISER_KMALLOC_THRESHOLD)
		free_pages((unsigned long)mem_copy->copy_buf,
			   ilog2(roundup_pow_of_two(cmd_data_len)) - PAGE_SHIFT);
	else
		kfree(mem_copy->copy_buf);

	mem_copy->copy_buf = NULL;
}

#define IS_4K_ALIGNED(addr)	((((unsigned long)addr) & ~MASK_4K) == 0)

/**
 * iser_sg_to_page_vec - Translates scatterlist entries to physical addresses
 * and returns the length of resulting physical address array (may be less than
 * the original due to possible compaction).
 *
 * we build a "page vec" under the assumption that the SG meets the RDMA
 * alignment requirements. Other then the first and last SG elements, all
 * the "internal" elements can be compacted into a list whose elements are
 * dma addresses of physical pages. The code supports also the weird case
 * where --few fragments of the same page-- are present in the SG as
 * consecutive elements. Also, it handles one entry SG.
 */

static int iser_sg_to_page_vec(struct iser_data_buf *data,
			       struct iser_page_vec *page_vec,
			       struct ib_device *ibdev)
{
	struct scatterlist *sg, *sgl = (struct scatterlist *)data->buf;
	u64 start_addr, end_addr, page, chunk_start = 0;
	unsigned long total_sz = 0;
	unsigned int dma_len;
	int i, new_chunk, cur_page, last_ent = data->dma_nents - 1;

	/* compute the offset of first element */
	page_vec->offset = (u64) sgl[0].offset & ~MASK_4K;

	new_chunk = 1;
	cur_page  = 0;
	for_each_sg(sgl, sg, data->dma_nents, i) {
		start_addr = ib_sg_dma_address(ibdev, sg);
		if (new_chunk)
			chunk_start = start_addr;
		dma_len = ib_sg_dma_len(ibdev, sg);
		end_addr = start_addr + dma_len;
		total_sz += dma_len;

		/* collect page fragments until aligned or end of SG list */
		if (!IS_4K_ALIGNED(end_addr) && i < last_ent) {
			new_chunk = 0;
			continue;
		}
		new_chunk = 1;

		/* address of the first page in the contiguous chunk;
		   masking relevant for the very first SG entry,
		   which might be unaligned */
		page = chunk_start & MASK_4K;
		do {
			page_vec->pages[cur_page++] = page;
			page += SIZE_4K;
		} while (page < end_addr);
	}

	page_vec->data_size = total_sz;
	iser_dbg("page_vec->data_size:%d cur_page %d\n", page_vec->data_size,cur_page);
	return cur_page;
}


/**
 * iser_data_buf_aligned_len - Tries to determine the maximal correctly aligned
 * for RDMA sub-list of a scatter-gather list of memory buffers, and  returns
 * the number of entries which are aligned correctly. Supports the case where
 * consecutive SG elements are actually fragments of the same physcial page.
 */
static int iser_data_buf_aligned_len(struct iser_data_buf *data,
				      struct ib_device *ibdev)
{
	struct scatterlist *sgl, *sg, *next_sg = NULL;
	u64 start_addr, end_addr;
	int i, ret_len, start_check = 0;

	if (data->dma_nents == 1)
		return 1;

	sgl = (struct scatterlist *)data->buf;
	start_addr  = ib_sg_dma_address(ibdev, sgl);

	for_each_sg(sgl, sg, data->dma_nents, i) {
		if (start_check && !IS_4K_ALIGNED(start_addr))
			break;

		next_sg = sg_next(sg);
		if (!next_sg)
			break;

		end_addr    = start_addr + ib_sg_dma_len(ibdev, sg);
		start_addr  = ib_sg_dma_address(ibdev, next_sg);

		if (end_addr == start_addr) {
			start_check = 0;
			continue;
		} else
			start_check = 1;

		if (!IS_4K_ALIGNED(end_addr))
			break;
	}
	ret_len = (next_sg) ? i : i+1;
	iser_dbg("Found %d aligned entries out of %d in sg:0x%p\n",
		 ret_len, data->dma_nents, data);
	return ret_len;
}

static void iser_data_buf_dump(struct iser_data_buf *data,
			       struct ib_device *ibdev)
{
	struct scatterlist *sgl = (struct scatterlist *)data->buf;
	struct scatterlist *sg;
	int i;

	if (iser_debug_level == 0)
		return;

	for_each_sg(sgl, sg, data->dma_nents, i)
		iser_warn("sg[%d] dma_addr:0x%lX page:0x%p "
			 "off:0x%x sz:0x%x dma_len:0x%x\n",
			 i, (unsigned long)ib_sg_dma_address(ibdev, sg),
			 sg_page(sg), sg->offset,
			 sg->length, ib_sg_dma_len(ibdev, sg));
}

static void iser_dump_page_vec(struct iser_page_vec *page_vec)
{
	int i;

	iser_err("page vec length %d data size %d\n",
		 page_vec->length, page_vec->data_size);
	for (i = 0; i < page_vec->length; i++)
		iser_err("%d %lx\n",i,(unsigned long)page_vec->pages[i]);
}

static void iser_page_vec_build(struct iser_data_buf *data,
				struct iser_page_vec *page_vec,
				struct ib_device *ibdev)
{
	int page_vec_len = 0;

	page_vec->length = 0;
	page_vec->offset = 0;

	iser_dbg("Translating sg sz: %d\n", data->dma_nents);
	page_vec_len = iser_sg_to_page_vec(data, page_vec, ibdev);
	iser_dbg("sg len %d page_vec_len %d\n", data->dma_nents,page_vec_len);

	page_vec->length = page_vec_len;

	if (page_vec_len * SIZE_4K < page_vec->data_size) {
		iser_err("page_vec too short to hold this SG\n");
		iser_data_buf_dump(data, ibdev);
		iser_dump_page_vec(page_vec);
		BUG();
	}
}

int iser_dma_map_task_data(struct iscsi_iser_task *iser_task,
			    struct iser_data_buf *data,
			    enum iser_data_dir iser_dir,
			    enum dma_data_direction dma_dir)
{
	struct ib_device *dev;

	iser_task->dir[iser_dir] = 1;
	dev = iser_task->iser_conn->ib_conn->device->ib_device;

	data->dma_nents = ib_dma_map_sg(dev, data->buf, data->size, dma_dir);
	if (data->dma_nents == 0) {
		iser_err("dma_map_sg failed!!!\n");
		return -EINVAL;
	}
	return 0;
}

void iser_dma_unmap_task_data(struct iscsi_iser_task *iser_task)
{
	struct ib_device *dev;
	struct iser_data_buf *data;

	dev = iser_task->iser_conn->ib_conn->device->ib_device;

	if (iser_task->dir[ISER_DIR_IN]) {
		data = &iser_task->data[ISER_DIR_IN];
		ib_dma_unmap_sg(dev, data->buf, data->size, DMA_FROM_DEVICE);
	}

	if (iser_task->dir[ISER_DIR_OUT]) {
		data = &iser_task->data[ISER_DIR_OUT];
		ib_dma_unmap_sg(dev, data->buf, data->size, DMA_TO_DEVICE);
	}
}

/**
 * iser_reg_rdma_mem - Registers memory intended for RDMA,
 * obtaining rkey and va
 *
 * returns 0 on success, errno code on failure
 */
int iser_reg_rdma_mem(struct iscsi_iser_task *iser_task,
		      enum   iser_data_dir        cmd_dir)
{
	struct iscsi_conn    *iscsi_conn = iser_task->iser_conn->iscsi_conn;
	struct iser_conn     *ib_conn = iser_task->iser_conn->ib_conn;
	struct iser_device   *device = ib_conn->device;
	struct ib_device     *ibdev = device->ib_device;
	struct iser_data_buf *mem = &iser_task->data[cmd_dir];
	struct iser_regd_buf *regd_buf;
	int aligned_len;
	int err;
	int i;
	struct scatterlist *sg;

	regd_buf = &iser_task->rdma_regd[cmd_dir];

	aligned_len = iser_data_buf_aligned_len(mem, ibdev);
	if (aligned_len != mem->dma_nents ||
	    (!ib_conn->fmr_pool && mem->dma_nents > 1)) {
		iscsi_conn->fmr_unalign_cnt++;
		iser_warn("rdma alignment violation (%d/%d aligned) or FMR not supported\n",
			  aligned_len, mem->size);
		iser_data_buf_dump(mem, ibdev);

		/* unmap the command data before accessing it */
		iser_dma_unmap_task_data(iser_task);

		/* allocate copy buf, if we are writing, copy the */
		/* unaligned scatterlist, dma map the copy        */
		if (iser_start_rdma_unaligned_sg(iser_task, cmd_dir) != 0)
				return -ENOMEM;
		mem = &iser_task->data_copy[cmd_dir];
	}

	/* if there a single dma entry, FMR is not needed */
	if (mem->dma_nents == 1) {
		sg = (struct scatterlist *)mem->buf;

		regd_buf->reg.lkey = device->mr->lkey;
		regd_buf->reg.rkey = device->mr->rkey;
		regd_buf->reg.len  = ib_sg_dma_len(ibdev, &sg[0]);
		regd_buf->reg.va   = ib_sg_dma_address(ibdev, &sg[0]);
		regd_buf->reg.is_fmr = 0;

		iser_dbg("PHYSICAL Mem.register: lkey: 0x%08X rkey: 0x%08X  "
			 "va: 0x%08lX sz: %ld]\n",
			 (unsigned int)regd_buf->reg.lkey,
			 (unsigned int)regd_buf->reg.rkey,
			 (unsigned long)regd_buf->reg.va,
			 (unsigned long)regd_buf->reg.len);
	} else { /* use FMR for multiple dma entries */
		iser_page_vec_build(mem, ib_conn->page_vec, ibdev);
		err = iser_reg_page_vec(ib_conn, ib_conn->page_vec, &regd_buf->reg);
		if (err && err != -EAGAIN) {
			iser_data_buf_dump(mem, ibdev);
			iser_err("mem->dma_nents = %d (dlength = 0x%x)\n",
				 mem->dma_nents,
				 ntoh24(iser_task->desc.iscsi_header.dlength));
			iser_err("page_vec: data_size = 0x%x, length = %d, offset = 0x%x\n",
				 ib_conn->page_vec->data_size, ib_conn->page_vec->length,
				 ib_conn->page_vec->offset);
			for (i=0 ; i<ib_conn->page_vec->length ; i++)
				iser_err("page_vec[%d] = 0x%llx\n", i,
					 (unsigned long long) ib_conn->page_vec->pages[i]);
		}
		if (err)
			return err;
	}
	return 0;
}