main.c 23.2 KB
Newer Older
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
/*
 * Copyright (c) 2013, Mellanox Technologies inc.  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 <asm-generic/kmap_types.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/errno.h>
#include <linux/pci.h>
#include <linux/dma-mapping.h>
#include <linux/slab.h>
#include <linux/io-mapping.h>
#include <linux/mlx5/driver.h>
#include <linux/mlx5/cq.h>
#include <linux/mlx5/qp.h>
#include <linux/mlx5/srq.h>
#include <linux/debugfs.h>
46
#include <linux/mlx5/mlx5_ifc.h>
47
48
49
#include "mlx5_core.h"

#define DRIVER_NAME "mlx5_core"
50
51
#define DRIVER_VERSION "2.2-1"
#define DRIVER_RELDATE	"Feb 2014"
52
53
54
55
56
57
58
59
60
61

MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
MODULE_DESCRIPTION("Mellanox ConnectX-IB HCA core library");
MODULE_LICENSE("Dual BSD/GPL");
MODULE_VERSION(DRIVER_VERSION);

int mlx5_core_debug_mask;
module_param_named(debug_mask, mlx5_core_debug_mask, int, 0644);
MODULE_PARM_DESC(debug_mask, "debug mask: 1 = dump cmd data, 2 = dump cmd exec time, 3 = both. Default=0");

62
63
64
65
66
#define MLX5_DEFAULT_PROF	2
static int prof_sel = MLX5_DEFAULT_PROF;
module_param_named(prof_sel, prof_sel, int, 0444);
MODULE_PARM_DESC(prof_sel, "profile selector. Valid range 0 - 2");

67
struct workqueue_struct *mlx5_core_wq;
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
static LIST_HEAD(intf_list);
static LIST_HEAD(dev_list);
static DEFINE_MUTEX(intf_mutex);

struct mlx5_device_context {
	struct list_head	list;
	struct mlx5_interface  *intf;
	void		       *context;
};

static struct mlx5_profile profile[] = {
	[0] = {
		.mask           = 0,
	},
	[1] = {
		.mask		= MLX5_PROF_MASK_QP_SIZE,
		.log_max_qp	= 12,
	},
	[2] = {
		.mask		= MLX5_PROF_MASK_QP_SIZE |
				  MLX5_PROF_MASK_MR_CACHE,
		.log_max_qp	= 17,
		.mr_cache[0]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[1]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[2]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[3]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[4]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[5]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[6]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[7]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[8]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[9]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[10]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[11]	= {
			.size	= 500,
			.limit	= 250
		},
		.mr_cache[12]	= {
			.size	= 64,
			.limit	= 32
		},
		.mr_cache[13]	= {
			.size	= 32,
			.limit	= 16
		},
		.mr_cache[14]	= {
			.size	= 16,
			.limit	= 8
		},
		.mr_cache[15]	= {
			.size	= 8,
			.limit	= 4
		},
	},
};
156
157
158
159
160
161
162

static int set_dma_caps(struct pci_dev *pdev)
{
	int err;

	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
	if (err) {
163
		dev_warn(&pdev->dev, "Warning: couldn't set 64-bit PCI DMA mask\n");
164
165
		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
		if (err) {
166
			dev_err(&pdev->dev, "Can't set PCI DMA mask, aborting\n");
167
168
169
170
171
172
173
			return err;
		}
	}

	err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
	if (err) {
		dev_warn(&pdev->dev,
174
			 "Warning: couldn't set 64-bit consistent PCI DMA mask\n");
175
176
177
		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
		if (err) {
			dev_err(&pdev->dev,
178
				"Can't set consistent PCI DMA mask, aborting\n");
179
180
181
182
183
184
185
186
187
188
189
190
191
			return err;
		}
	}

	dma_set_max_seg_size(&pdev->dev, 2u * 1024 * 1024 * 1024);
	return err;
}

static int request_bar(struct pci_dev *pdev)
{
	int err = 0;

	if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM)) {
192
		dev_err(&pdev->dev, "Missing registers BAR, aborting\n");
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
		return -ENODEV;
	}

	err = pci_request_regions(pdev, DRIVER_NAME);
	if (err)
		dev_err(&pdev->dev, "Couldn't get PCI resources, aborting\n");

	return err;
}

static void release_bar(struct pci_dev *pdev)
{
	pci_release_regions(pdev);
}

static int mlx5_enable_msix(struct mlx5_core_dev *dev)
{
	struct mlx5_eq_table *table = &dev->priv.eq_table;
211
	int num_eqs = 1 << dev->caps.gen.log_max_eq;
212
213
214
	int nvec;
	int i;

215
	nvec = dev->caps.gen.num_ports * num_online_cpus() + MLX5_EQ_VEC_COMP_BASE;
216
217
218
219
220
221
222
223
224
225
226
	nvec = min_t(int, nvec, num_eqs);
	if (nvec <= MLX5_EQ_VEC_COMP_BASE)
		return -ENOMEM;

	table->msix_arr = kzalloc(nvec * sizeof(*table->msix_arr), GFP_KERNEL);
	if (!table->msix_arr)
		return -ENOMEM;

	for (i = 0; i < nvec; i++)
		table->msix_arr[i].entry = i;

227
228
229
230
	nvec = pci_enable_msix_range(dev->pdev, table->msix_arr,
				     MLX5_EQ_VEC_COMP_BASE, nvec);
	if (nvec < 0)
		return nvec;
231

232
	table->num_comp_vectors = nvec - MLX5_EQ_VEC_COMP_BASE;
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249

	return 0;
}

static void mlx5_disable_msix(struct mlx5_core_dev *dev)
{
	struct mlx5_eq_table *table = &dev->priv.eq_table;

	pci_disable_msix(dev->pdev);
	kfree(table->msix_arr);
}

struct mlx5_reg_host_endianess {
	u8	he;
	u8      rsvd[15];
};

250
251
252
253

#define CAP_MASK(pos, size) ((u64)((1 << (size)) - 1) << (pos))

enum {
254
255
	MLX5_CAP_BITS_RW_MASK = CAP_MASK(MLX5_CAP_OFF_CMDIF_CSUM, 2) |
				MLX5_DEV_CAP_FLAG_DCT,
256
257
};

258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
static u16 to_fw_pkey_sz(u32 size)
{
	switch (size) {
	case 128:
		return 0;
	case 256:
		return 1;
	case 512:
		return 2;
	case 1024:
		return 3;
	case 2048:
		return 4;
	case 4096:
		return 5;
	default:
		pr_warn("invalid pkey table size %d\n", size);
		return 0;
	}
}

279
280
/* selectively copy writable fields clearing any reserved area
 */
281
static void copy_rw_fields(void *to, struct mlx5_caps *from)
282
{
283
	__be64 *flags_off = (__be64 *)MLX5_ADDR_OF(cmd_hca_cap, to, reserved_22);
284
285
	u64 v64;

286
287
288
289
290
291
292
293
294
	MLX5_SET(cmd_hca_cap, to, log_max_qp, from->gen.log_max_qp);
	MLX5_SET(cmd_hca_cap, to, log_max_ra_req_qp, from->gen.log_max_ra_req_qp);
	MLX5_SET(cmd_hca_cap, to, log_max_ra_res_qp, from->gen.log_max_ra_res_qp);
	MLX5_SET(cmd_hca_cap, to, pkey_table_size, from->gen.pkey_table_size);
	MLX5_SET(cmd_hca_cap, to, log_max_ra_req_dc, from->gen.log_max_ra_req_dc);
	MLX5_SET(cmd_hca_cap, to, log_max_ra_res_dc, from->gen.log_max_ra_res_dc);
	MLX5_SET(cmd_hca_cap, to, pkey_table_size, to_fw_pkey_sz(from->gen.pkey_table_size));
	v64 = from->gen.flags & MLX5_CAP_BITS_RW_MASK;
	*flags_off = cpu_to_be64(v64);
295
296
}

297
298
299
300
static u16 get_pkey_table_size(int pkey)
{
	if (pkey > MLX5_MAX_LOG_PKEY_TABLE)
		return 0;
301

302
303
304
	return MLX5_MIN_PKEY_TABLE_SIZE << pkey;
}

305
static void fw2drv_caps(struct mlx5_caps *caps, void *out)
306
{
307
	struct mlx5_general_caps *gen = &caps->gen;
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

	gen->max_srq_wqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_srq_sz);
	gen->max_wqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_qp_sz);
	gen->log_max_qp = MLX5_GET_PR(cmd_hca_cap, out, log_max_qp);
	gen->log_max_strq = MLX5_GET_PR(cmd_hca_cap, out, log_max_strq_sz);
	gen->log_max_srq = MLX5_GET_PR(cmd_hca_cap, out, log_max_srqs);
	gen->max_cqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_cq_sz);
	gen->log_max_cq = MLX5_GET_PR(cmd_hca_cap, out, log_max_cq);
	gen->max_eqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_eq_sz);
	gen->log_max_mkey = MLX5_GET_PR(cmd_hca_cap, out, log_max_mkey);
	gen->log_max_eq = MLX5_GET_PR(cmd_hca_cap, out, log_max_eq);
	gen->max_indirection = MLX5_GET_PR(cmd_hca_cap, out, max_indirection);
	gen->log_max_mrw_sz = MLX5_GET_PR(cmd_hca_cap, out, log_max_mrw_sz);
	gen->log_max_bsf_list_size = MLX5_GET_PR(cmd_hca_cap, out, log_max_bsf_list_size);
	gen->log_max_klm_list_size = MLX5_GET_PR(cmd_hca_cap, out, log_max_klm_list_size);
	gen->log_max_ra_req_dc = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_req_dc);
	gen->log_max_ra_res_dc = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_res_dc);
	gen->log_max_ra_req_qp = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_req_qp);
	gen->log_max_ra_res_qp = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_res_qp);
	gen->max_qp_counters = MLX5_GET_PR(cmd_hca_cap, out, max_qp_cnt);
	gen->pkey_table_size = get_pkey_table_size(MLX5_GET_PR(cmd_hca_cap, out, pkey_table_size));
	gen->local_ca_ack_delay = MLX5_GET_PR(cmd_hca_cap, out, local_ca_ack_delay);
	gen->num_ports = MLX5_GET_PR(cmd_hca_cap, out, num_ports);
	gen->log_max_msg = MLX5_GET_PR(cmd_hca_cap, out, log_max_msg);
	gen->stat_rate_support = MLX5_GET_PR(cmd_hca_cap, out, stat_rate_support);
	gen->flags = be64_to_cpu(*(__be64 *)MLX5_ADDR_OF(cmd_hca_cap, out, reserved_22));
334
	pr_debug("flags = 0x%llx\n", gen->flags);
335
336
337
338
339
340
341
342
343
344
345
	gen->uar_sz = MLX5_GET_PR(cmd_hca_cap, out, uar_sz);
	gen->min_log_pg_sz = MLX5_GET_PR(cmd_hca_cap, out, log_pg_sz);
	gen->bf_reg_size = MLX5_GET_PR(cmd_hca_cap, out, bf);
	gen->bf_reg_size = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_bf_reg_size);
	gen->max_sq_desc_sz = MLX5_GET_PR(cmd_hca_cap, out, max_wqe_sz_sq);
	gen->max_rq_desc_sz = MLX5_GET_PR(cmd_hca_cap, out, max_wqe_sz_rq);
	gen->max_dc_sq_desc_sz = MLX5_GET_PR(cmd_hca_cap, out, max_wqe_sz_sq_dc);
	gen->max_qp_mcg = MLX5_GET_PR(cmd_hca_cap, out, max_qp_mcg);
	gen->log_max_pd = MLX5_GET_PR(cmd_hca_cap, out, log_max_pd);
	gen->log_max_xrcd = MLX5_GET_PR(cmd_hca_cap, out, log_max_xrcd);
	gen->log_uar_page_sz = MLX5_GET_PR(cmd_hca_cap, out, log_uar_page_sz);
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
}

static const char *caps_opmod_str(u16 opmod)
{
	switch (opmod) {
	case HCA_CAP_OPMOD_GET_MAX:
		return "GET_MAX";
	case HCA_CAP_OPMOD_GET_CUR:
		return "GET_CUR";
	default:
		return "Invalid";
	}
}

int mlx5_core_get_caps(struct mlx5_core_dev *dev, struct mlx5_caps *caps,
		       u16 opmod)
{
363
364
365
	u8 in[MLX5_ST_SZ_BYTES(query_hca_cap_in)];
	int out_sz = MLX5_ST_SZ_BYTES(query_hca_cap_out);
	void *out;
366
367
	int err;

368
369
	memset(in, 0, sizeof(in));
	out = kzalloc(out_sz, GFP_KERNEL);
370
	if (!out)
371
		return -ENOMEM;
372
373
374
375
376
	MLX5_SET(query_hca_cap_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_CAP);
	MLX5_SET(query_hca_cap_in, in, op_mod, opmod);
	err = mlx5_cmd_exec(dev, in, sizeof(in), out, out_sz);
	if (err)
		goto query_ex;
377

378
	err = mlx5_cmd_status_to_err_v2(out);
379
380
	if (err) {
		mlx5_core_warn(dev, "query max hca cap failed, %d\n", err);
381
382
		goto query_ex;
	}
383
	mlx5_core_dbg(dev, "%s\n", caps_opmod_str(opmod));
384
	fw2drv_caps(caps, MLX5_ADDR_OF(query_hca_cap_out, out, capability_struct));
385
386
387
388
389
390

query_ex:
	kfree(out);
	return err;
}

391
static int set_caps(struct mlx5_core_dev *dev, void *in, int in_sz)
392
{
393
	u32 out[MLX5_ST_SZ_DW(set_hca_cap_out)];
394
395
	int err;

396
	memset(out, 0, sizeof(out));
397

398
399
	MLX5_SET(set_hca_cap_in, in, opcode, MLX5_CMD_OP_SET_HCA_CAP);
	err = mlx5_cmd_exec(dev, in, in_sz, out, sizeof(out));
400
	if (err)
401
		return err;
402

403
	err = mlx5_cmd_status_to_err_v2(out);
404
405
406
407
408
409

	return err;
}

static int handle_hca_cap(struct mlx5_core_dev *dev)
{
410
	void *set_ctx = NULL;
411
412
413
414
	struct mlx5_profile *prof = dev->profile;
	struct mlx5_caps *cur_caps = NULL;
	struct mlx5_caps *max_caps = NULL;
	int err = -ENOMEM;
415
	int set_sz = MLX5_ST_SZ_BYTES(set_hca_cap_in);
416

417
	set_ctx = kzalloc(set_sz, GFP_KERNEL);
418
	if (!set_ctx)
419
420
		goto query_ex;

421
422
423
	max_caps = kzalloc(sizeof(*max_caps), GFP_KERNEL);
	if (!max_caps)
		goto query_ex;
424

425
426
427
	cur_caps = kzalloc(sizeof(*cur_caps), GFP_KERNEL);
	if (!cur_caps)
		goto query_ex;
428

429
430
	err = mlx5_core_get_caps(dev, max_caps, HCA_CAP_OPMOD_GET_MAX);
	if (err)
431
432
		goto query_ex;

433
	err = mlx5_core_get_caps(dev, cur_caps, HCA_CAP_OPMOD_GET_CUR);
434
435
436
	if (err)
		goto query_ex;

437
438
439
440
441
442
443
444
445
	/* we limit the size of the pkey table to 128 entries for now */
	cur_caps->gen.pkey_table_size = 128;

	if (prof->mask & MLX5_PROF_MASK_QP_SIZE)
		cur_caps->gen.log_max_qp = prof->log_max_qp;

	/* disable checksum */
	cur_caps->gen.flags &= ~MLX5_DEV_CAP_FLAG_CMDIF_CSUM;

446
447
448
	copy_rw_fields(MLX5_ADDR_OF(set_hca_cap_in, set_ctx, hca_capability_struct),
		       cur_caps);
	err = set_caps(dev, set_ctx, set_sz);
449

450
query_ex:
451
452
	kfree(cur_caps);
	kfree(max_caps);
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
	kfree(set_ctx);

	return err;
}

static int set_hca_ctrl(struct mlx5_core_dev *dev)
{
	struct mlx5_reg_host_endianess he_in;
	struct mlx5_reg_host_endianess he_out;
	int err;

	memset(&he_in, 0, sizeof(he_in));
	he_in.he = MLX5_SET_HOST_ENDIANNESS;
	err = mlx5_core_access_reg(dev, &he_in,  sizeof(he_in),
					&he_out, sizeof(he_out),
					MLX5_REG_HOST_ENDIANNESS, 0, 1);
	return err;
}

472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
static int mlx5_core_enable_hca(struct mlx5_core_dev *dev)
{
	int err;
	struct mlx5_enable_hca_mbox_in in;
	struct mlx5_enable_hca_mbox_out out;

	memset(&in, 0, sizeof(in));
	memset(&out, 0, sizeof(out));
	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_ENABLE_HCA);
	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
	if (err)
		return err;

	if (out.hdr.status)
		return mlx5_cmd_status_to_err(&out.hdr);

	return 0;
}

static int mlx5_core_disable_hca(struct mlx5_core_dev *dev)
{
	int err;
	struct mlx5_disable_hca_mbox_in in;
	struct mlx5_disable_hca_mbox_out out;

	memset(&in, 0, sizeof(in));
	memset(&out, 0, sizeof(out));
	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DISABLE_HCA);
	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
	if (err)
		return err;

	if (out.hdr.status)
		return mlx5_cmd_status_to_err(&out.hdr);

	return 0;
}

510
static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
{
	struct mlx5_priv *priv = &dev->priv;
	int err;

	dev->pdev = pdev;
	pci_set_drvdata(dev->pdev, dev);
	strncpy(priv->name, dev_name(&pdev->dev), MLX5_MAX_NAME_LEN);
	priv->name[MLX5_MAX_NAME_LEN - 1] = 0;

	mutex_init(&priv->pgdir_mutex);
	INIT_LIST_HEAD(&priv->pgdir_list);
	spin_lock_init(&priv->mkey_lock);

	priv->dbg_root = debugfs_create_dir(dev_name(&pdev->dev), mlx5_debugfs_root);
	if (!priv->dbg_root)
		return -ENOMEM;

	err = pci_enable_device(pdev);
	if (err) {
530
		dev_err(&pdev->dev, "Cannot enable PCI device, aborting\n");
531
532
533
534
535
		goto err_dbg;
	}

	err = request_bar(pdev);
	if (err) {
536
		dev_err(&pdev->dev, "error requesting BARs, aborting\n");
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
		goto err_disable;
	}

	pci_set_master(pdev);

	err = set_dma_caps(pdev);
	if (err) {
		dev_err(&pdev->dev, "Failed setting DMA capabilities mask, aborting\n");
		goto err_clr_master;
	}

	dev->iseg_base = pci_resource_start(dev->pdev, 0);
	dev->iseg = ioremap(dev->iseg_base, sizeof(*dev->iseg));
	if (!dev->iseg) {
		err = -ENOMEM;
		dev_err(&pdev->dev, "Failed mapping initialization segment, aborting\n");
		goto err_clr_master;
	}
	dev_info(&pdev->dev, "firmware version: %d.%d.%d\n", fw_rev_maj(dev),
		 fw_rev_min(dev), fw_rev_sub(dev));

	err = mlx5_cmd_init(dev);
	if (err) {
		dev_err(&pdev->dev, "Failed initializing command interface, aborting\n");
		goto err_unmap;
	}

	mlx5_pagealloc_init(dev);
565
566
567
568
569
570
571
572
573
574
575
576
577

	err = mlx5_core_enable_hca(dev);
	if (err) {
		dev_err(&pdev->dev, "enable hca failed\n");
		goto err_pagealloc_cleanup;
	}

	err = mlx5_satisfy_startup_pages(dev, 1);
	if (err) {
		dev_err(&pdev->dev, "failed to allocate boot pages\n");
		goto err_disable_hca;
	}

578
579
580
	err = set_hca_ctrl(dev);
	if (err) {
		dev_err(&pdev->dev, "set_hca_ctrl failed\n");
581
		goto reclaim_boot_pages;
582
583
584
585
586
	}

	err = handle_hca_cap(dev);
	if (err) {
		dev_err(&pdev->dev, "handle_hca_cap failed\n");
587
		goto reclaim_boot_pages;
588
589
	}

590
	err = mlx5_satisfy_startup_pages(dev, 0);
591
	if (err) {
592
593
		dev_err(&pdev->dev, "failed to allocate init pages\n");
		goto reclaim_boot_pages;
594
595
596
597
598
	}

	err = mlx5_pagealloc_start(dev);
	if (err) {
		dev_err(&pdev->dev, "mlx5_pagealloc_start failed\n");
599
		goto reclaim_boot_pages;
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
	}

	err = mlx5_cmd_init_hca(dev);
	if (err) {
		dev_err(&pdev->dev, "init hca failed\n");
		goto err_pagealloc_stop;
	}

	mlx5_start_health_poll(dev);

	err = mlx5_cmd_query_hca_cap(dev, &dev->caps);
	if (err) {
		dev_err(&pdev->dev, "query hca failed\n");
		goto err_stop_poll;
	}

	err = mlx5_cmd_query_adapter(dev);
	if (err) {
		dev_err(&pdev->dev, "query adapter failed\n");
		goto err_stop_poll;
	}

	err = mlx5_enable_msix(dev);
	if (err) {
		dev_err(&pdev->dev, "enable msix failed\n");
		goto err_stop_poll;
	}

	err = mlx5_eq_init(dev);
	if (err) {
		dev_err(&pdev->dev, "failed to initialize eq\n");
		goto disable_msix;
	}

	err = mlx5_alloc_uuars(dev, &priv->uuari);
	if (err) {
		dev_err(&pdev->dev, "Failed allocating uar, aborting\n");
		goto err_eq_cleanup;
	}

	err = mlx5_start_eqs(dev);
	if (err) {
		dev_err(&pdev->dev, "Failed to start pages and async EQs\n");
		goto err_free_uar;
	}

	MLX5_INIT_DOORBELL_LOCK(&priv->cq_uar_lock);

	mlx5_init_cq_table(dev);
	mlx5_init_qp_table(dev);
	mlx5_init_srq_table(dev);
651
	mlx5_init_mr_table(dev);
652
653
654
655
656
657
658
659
660
661
662
663
664
665

	return 0;

err_free_uar:
	mlx5_free_uuars(dev, &priv->uuari);

err_eq_cleanup:
	mlx5_eq_cleanup(dev);

disable_msix:
	mlx5_disable_msix(dev);

err_stop_poll:
	mlx5_stop_health_poll(dev);
666
667
668
669
	if (mlx5_cmd_teardown_hca(dev)) {
		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
		return err;
	}
670
671
672
673

err_pagealloc_stop:
	mlx5_pagealloc_stop(dev);

674
reclaim_boot_pages:
675
676
	mlx5_reclaim_startup_pages(dev);

677
678
679
err_disable_hca:
	mlx5_core_disable_hca(dev);

680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
err_pagealloc_cleanup:
	mlx5_pagealloc_cleanup(dev);
	mlx5_cmd_cleanup(dev);

err_unmap:
	iounmap(dev->iseg);

err_clr_master:
	pci_clear_master(dev->pdev);
	release_bar(dev->pdev);

err_disable:
	pci_disable_device(dev->pdev);

err_dbg:
	debugfs_remove(priv->dbg_root);
	return err;
}
EXPORT_SYMBOL(mlx5_dev_init);

700
static void mlx5_dev_cleanup(struct mlx5_core_dev *dev)
701
702
703
704
705
706
707
708
709
710
711
{
	struct mlx5_priv *priv = &dev->priv;

	mlx5_cleanup_srq_table(dev);
	mlx5_cleanup_qp_table(dev);
	mlx5_cleanup_cq_table(dev);
	mlx5_stop_eqs(dev);
	mlx5_free_uuars(dev, &priv->uuari);
	mlx5_eq_cleanup(dev);
	mlx5_disable_msix(dev);
	mlx5_stop_health_poll(dev);
712
713
714
715
	if (mlx5_cmd_teardown_hca(dev)) {
		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
		return;
	}
716
717
	mlx5_pagealloc_stop(dev);
	mlx5_reclaim_startup_pages(dev);
718
	mlx5_core_disable_hca(dev);
719
720
721
722
723
724
725
726
	mlx5_pagealloc_cleanup(dev);
	mlx5_cmd_cleanup(dev);
	iounmap(dev->iseg);
	pci_clear_master(dev->pdev);
	release_bar(dev->pdev);
	pci_disable_device(dev->pdev);
	debugfs_remove(priv->dbg_root);
}
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821

static void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
{
	struct mlx5_device_context *dev_ctx;
	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);

	dev_ctx = kmalloc(sizeof(*dev_ctx), GFP_KERNEL);
	if (!dev_ctx) {
		pr_warn("mlx5_add_device: alloc context failed\n");
		return;
	}

	dev_ctx->intf    = intf;
	dev_ctx->context = intf->add(dev);

	if (dev_ctx->context) {
		spin_lock_irq(&priv->ctx_lock);
		list_add_tail(&dev_ctx->list, &priv->ctx_list);
		spin_unlock_irq(&priv->ctx_lock);
	} else {
		kfree(dev_ctx);
	}
}

static void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
{
	struct mlx5_device_context *dev_ctx;
	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);

	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
		if (dev_ctx->intf == intf) {
			spin_lock_irq(&priv->ctx_lock);
			list_del(&dev_ctx->list);
			spin_unlock_irq(&priv->ctx_lock);

			intf->remove(dev, dev_ctx->context);
			kfree(dev_ctx);
			return;
		}
}
static int mlx5_register_device(struct mlx5_core_dev *dev)
{
	struct mlx5_priv *priv = &dev->priv;
	struct mlx5_interface *intf;

	mutex_lock(&intf_mutex);
	list_add_tail(&priv->dev_list, &dev_list);
	list_for_each_entry(intf, &intf_list, list)
		mlx5_add_device(intf, priv);
	mutex_unlock(&intf_mutex);

	return 0;
}
static void mlx5_unregister_device(struct mlx5_core_dev *dev)
{
	struct mlx5_priv *priv = &dev->priv;
	struct mlx5_interface *intf;

	mutex_lock(&intf_mutex);
	list_for_each_entry(intf, &intf_list, list)
		mlx5_remove_device(intf, priv);
	list_del(&priv->dev_list);
	mutex_unlock(&intf_mutex);
}

int mlx5_register_interface(struct mlx5_interface *intf)
{
	struct mlx5_priv *priv;

	if (!intf->add || !intf->remove)
		return -EINVAL;

	mutex_lock(&intf_mutex);
	list_add_tail(&intf->list, &intf_list);
	list_for_each_entry(priv, &dev_list, dev_list)
		mlx5_add_device(intf, priv);
	mutex_unlock(&intf_mutex);

	return 0;
}
EXPORT_SYMBOL(mlx5_register_interface);

void mlx5_unregister_interface(struct mlx5_interface *intf)
{
	struct mlx5_priv *priv;

	mutex_lock(&intf_mutex);
	list_for_each_entry(priv, &dev_list, dev_list)
	       mlx5_remove_device(intf, priv);
	list_del(&intf->list);
	mutex_unlock(&intf_mutex);
}
EXPORT_SYMBOL(mlx5_unregister_interface);

static void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
822
			    unsigned long param)
823
824
825
826
827
828
829
830
831
{
	struct mlx5_priv *priv = &dev->priv;
	struct mlx5_device_context *dev_ctx;
	unsigned long flags;

	spin_lock_irqsave(&priv->ctx_lock, flags);

	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
		if (dev_ctx->intf->event)
832
			dev_ctx->intf->event(dev, dev_ctx->context, event, param);
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866

	spin_unlock_irqrestore(&priv->ctx_lock, flags);
}

struct mlx5_core_event_handler {
	void (*event)(struct mlx5_core_dev *dev,
		      enum mlx5_dev_event event,
		      void *data);
};

static int init_one(struct pci_dev *pdev,
		    const struct pci_device_id *id)
{
	struct mlx5_core_dev *dev;
	struct mlx5_priv *priv;
	int err;

	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
	if (!dev) {
		dev_err(&pdev->dev, "kzalloc failed\n");
		return -ENOMEM;
	}
	priv = &dev->priv;

	pci_set_drvdata(pdev, dev);

	if (prof_sel < 0 || prof_sel >= ARRAY_SIZE(profile)) {
		pr_warn("selected profile out of range, selecting default (%d)\n",
			MLX5_DEFAULT_PROF);
		prof_sel = MLX5_DEFAULT_PROF;
	}
	dev->profile = &profile[prof_sel];
	dev->event = mlx5_core_event;

867
868
	INIT_LIST_HEAD(&priv->ctx_list);
	spin_lock_init(&priv->ctx_lock);
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
	err = mlx5_dev_init(dev, pdev);
	if (err) {
		dev_err(&pdev->dev, "mlx5_dev_init failed %d\n", err);
		goto out;
	}

	err = mlx5_register_device(dev);
	if (err) {
		dev_err(&pdev->dev, "mlx5_register_device failed %d\n", err);
		goto out_init;
	}

	return 0;

out_init:
	mlx5_dev_cleanup(dev);
out:
	kfree(dev);
	return err;
}
static void remove_one(struct pci_dev *pdev)
{
	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);

	mlx5_unregister_device(dev);
	mlx5_dev_cleanup(dev);
	kfree(dev);
}

static const struct pci_device_id mlx5_core_pci_table[] = {
	{ PCI_VDEVICE(MELLANOX, 4113) }, /* MT4113 Connect-IB */
900
	{ PCI_VDEVICE(MELLANOX, 4115) }, /* ConnectX-4 */
901
902
903
904
905
906
907
908
909
910
911
	{ 0, }
};

MODULE_DEVICE_TABLE(pci, mlx5_core_pci_table);

static struct pci_driver mlx5_core_driver = {
	.name           = DRIVER_NAME,
	.id_table       = mlx5_core_pci_table,
	.probe          = init_one,
	.remove         = remove_one
};
912
913
914
915
916
917
918
919
920
921
922
923
924

static int __init init(void)
{
	int err;

	mlx5_register_debugfs();
	mlx5_core_wq = create_singlethread_workqueue("mlx5_core_wq");
	if (!mlx5_core_wq) {
		err = -ENOMEM;
		goto err_debug;
	}
	mlx5_health_init();

925
926
927
928
	err = pci_register_driver(&mlx5_core_driver);
	if (err)
		goto err_health;

929
930
	return 0;

931
932
933
err_health:
	mlx5_health_cleanup();
	destroy_workqueue(mlx5_core_wq);
934
935
936
937
938
939
940
err_debug:
	mlx5_unregister_debugfs();
	return err;
}

static void __exit cleanup(void)
{
941
	pci_unregister_driver(&mlx5_core_driver);
942
943
944
945
946
947
948
	mlx5_health_cleanup();
	destroy_workqueue(mlx5_core_wq);
	mlx5_unregister_debugfs();
}

module_init(init);
module_exit(cleanup);