main.c 69 KB
Newer Older
1
2
3
/*
 * Copyright (c) 2004, 2005 Topspin Communications.  All rights reserved.
 * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
4
 * Copyright (c) 2005, 2006, 2007, 2008 Mellanox Technologies. All rights reserved.
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
 * Copyright (c) 2006, 2007 Cisco Systems, 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 <linux/module.h>
#include <linux/init.h>
#include <linux/errno.h>
#include <linux/pci.h>
#include <linux/dma-mapping.h>
41
#include <linux/slab.h>
42
#include <linux/io-mapping.h>
43
#include <linux/delay.h>
44
#include <linux/netdevice.h>
45
46
47
48
49
50
51
52
53
54
55
56
57

#include <linux/mlx4/device.h>
#include <linux/mlx4/doorbell.h>

#include "mlx4.h"
#include "fw.h"
#include "icm.h"

MODULE_AUTHOR("Roland Dreier");
MODULE_DESCRIPTION("Mellanox ConnectX HCA low-level driver");
MODULE_LICENSE("Dual BSD/GPL");
MODULE_VERSION(DRV_VERSION);

58
59
struct workqueue_struct *mlx4_wq;

60
61
62
63
64
65
66
67
68
69
#ifdef CONFIG_MLX4_DEBUG

int mlx4_debug_level = 0;
module_param_named(debug_level, mlx4_debug_level, int, 0644);
MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0");

#endif /* CONFIG_MLX4_DEBUG */

#ifdef CONFIG_PCI_MSI

70
static int msi_x = 1;
71
72
73
74
75
76
77
78
79
module_param(msi_x, int, 0444);
MODULE_PARM_DESC(msi_x, "attempt to use MSI-X if nonzero");

#else /* CONFIG_PCI_MSI */

#define msi_x (0)

#endif /* CONFIG_PCI_MSI */

80
81
82
83
84
85
86
87
static int num_vfs;
module_param(num_vfs, int, 0444);
MODULE_PARM_DESC(num_vfs, "enable #num_vfs functions if num_vfs > 0");

static int probe_vf;
module_param(probe_vf, int, 0644);
MODULE_PARM_DESC(probe_vf, "number of vfs to probe by pf driver (num_vfs > 0)");

88
int mlx4_log_num_mgm_entry_size = MLX4_DEFAULT_MGM_LOG_ENTRY_SIZE;
89
90
91
92
module_param_named(log_num_mgm_entry_size,
			mlx4_log_num_mgm_entry_size, int, 0444);
MODULE_PARM_DESC(log_num_mgm_entry_size, "log mgm size, that defines the num"
					 " of qp per mcg, for example:"
93
					 " 10 gives 248.range: 7 <="
94
					 " log_num_mgm_entry_size <= 12."
95
96
					 " To activate device managed"
					 " flow steering when available, set to -1");
97

Or Gerlitz's avatar
Or Gerlitz committed
98
99
100
101
102
static bool enable_64b_cqe_eqe;
module_param(enable_64b_cqe_eqe, bool, 0444);
MODULE_PARM_DESC(enable_64b_cqe_eqe,
		 "Enable 64 byte CQEs/EQEs when the the FW supports this");

103
#define HCA_GLOBAL_CAP_MASK            0
Or Gerlitz's avatar
Or Gerlitz committed
104
105

#define PF_CONTEXT_BEHAVIOUR_MASK	MLX4_FUNC_CAP_64B_EQE_CQE
106

107
static char mlx4_version[] =
108
109
110
111
	DRV_NAME ": Mellanox ConnectX core driver v"
	DRV_VERSION " (" DRV_RELDATE ")\n";

static struct mlx4_profile default_profile = {
112
	.num_qp		= 1 << 18,
113
	.num_srq	= 1 << 16,
114
	.rdmarc_per_qp	= 1 << 4,
115
116
	.num_cq		= 1 << 16,
	.num_mcg	= 1 << 13,
117
	.num_mpt	= 1 << 19,
118
	.num_mtt	= 1 << 20, /* It is really num mtt segements */
119
120
};

121
static int log_num_mac = 7;
122
123
124
125
126
127
module_param_named(log_num_mac, log_num_mac, int, 0444);
MODULE_PARM_DESC(log_num_mac, "Log2 max number of MACs per ETH port (1-7)");

static int log_num_vlan;
module_param_named(log_num_vlan, log_num_vlan, int, 0444);
MODULE_PARM_DESC(log_num_vlan, "Log2 max number of VLANs per ETH port (0-7)");
128
129
/* Log2 max number of VLANs per ETH port (0-7) */
#define MLX4_LOG_NUM_VLANS 7
130

131
static bool use_prio;
132
133
134
135
module_param_named(use_prio, use_prio, bool, 0444);
MODULE_PARM_DESC(use_prio, "Enable steering by VLAN priority on ETH ports "
		  "(0/1, default 0)");

136
int log_mtts_per_seg = ilog2(MLX4_MTT_ENTRY_PER_SEG);
137
module_param_named(log_mtts_per_seg, log_mtts_per_seg, int, 0444);
138
MODULE_PARM_DESC(log_mtts_per_seg, "Log2 number of MTT entries per segment (1-7)");
139

140
static int port_type_array[2] = {MLX4_PORT_TYPE_NONE, MLX4_PORT_TYPE_NONE};
141
142
static int arr_argc = 2;
module_param_array(port_type_array, int, &arr_argc, 0444);
143
144
MODULE_PARM_DESC(port_type_array, "Array of port types: HW_DEFAULT (0) is default "
				"1 for IB, 2 for Ethernet");
145
146
147
148
149
150
151

struct mlx4_port_config {
	struct list_head list;
	enum mlx4_port_type port_type[MLX4_MAX_PORTS + 1];
	struct pci_dev *pdev;
};

152
153
int mlx4_check_port_params(struct mlx4_dev *dev,
			   enum mlx4_port_type *port_type)
154
155
156
157
{
	int i;

	for (i = 0; i < dev->caps.num_ports - 1; i++) {
158
159
160
161
162
163
		if (port_type[i] != port_type[i + 1]) {
			if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP)) {
				mlx4_err(dev, "Only same port types supported "
					 "on this HCA, aborting.\n");
				return -EINVAL;
			}
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
		}
	}

	for (i = 0; i < dev->caps.num_ports; i++) {
		if (!(port_type[i] & dev->caps.supported_type[i+1])) {
			mlx4_err(dev, "Requested port type for port %d is not "
				      "supported on this HCA\n", i + 1);
			return -EINVAL;
		}
	}
	return 0;
}

static void mlx4_set_port_mask(struct mlx4_dev *dev)
{
	int i;

	for (i = 1; i <= dev->caps.num_ports; ++i)
182
		dev->caps.port_mask[i] = dev->caps.port_type[i];
183
}
184

185
static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
186
187
{
	int err;
188
	int i;
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

	err = mlx4_QUERY_DEV_CAP(dev, dev_cap);
	if (err) {
		mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
		return err;
	}

	if (dev_cap->min_page_sz > PAGE_SIZE) {
		mlx4_err(dev, "HCA minimum page size of %d bigger than "
			 "kernel PAGE_SIZE of %ld, aborting.\n",
			 dev_cap->min_page_sz, PAGE_SIZE);
		return -ENODEV;
	}
	if (dev_cap->num_ports > MLX4_MAX_PORTS) {
		mlx4_err(dev, "HCA has %d ports, but we only support %d, "
			 "aborting.\n",
			 dev_cap->num_ports, MLX4_MAX_PORTS);
		return -ENODEV;
	}

	if (dev_cap->uar_size > pci_resource_len(dev->pdev, 2)) {
		mlx4_err(dev, "HCA reported UAR size of 0x%x bigger than "
			 "PCI resource 2 size of 0x%llx, aborting.\n",
			 dev_cap->uar_size,
			 (unsigned long long) pci_resource_len(dev->pdev, 2));
		return -ENODEV;
	}

	dev->caps.num_ports	     = dev_cap->num_ports;
218
	dev->phys_caps.num_phys_eqs  = MLX4_MAX_EQ_NUM;
219
220
	for (i = 1; i <= dev->caps.num_ports; ++i) {
		dev->caps.vl_cap[i]	    = dev_cap->max_vl[i];
221
		dev->caps.ib_mtu_cap[i]	    = dev_cap->ib_mtu[i];
222
223
224
225
		dev->phys_caps.gid_phys_table_len[i]  = dev_cap->max_gids[i];
		dev->phys_caps.pkey_phys_table_len[i] = dev_cap->max_pkeys[i];
		/* set gid and pkey table operating lengths by default
		 * to non-sriov values */
226
227
228
		dev->caps.gid_table_len[i]  = dev_cap->max_gids[i];
		dev->caps.pkey_table_len[i] = dev_cap->max_pkeys[i];
		dev->caps.port_width_cap[i] = dev_cap->max_port_width[i];
229
230
		dev->caps.eth_mtu_cap[i]    = dev_cap->eth_mtu[i];
		dev->caps.def_mac[i]        = dev_cap->def_mac[i];
231
		dev->caps.supported_type[i] = dev_cap->supported_port_types[i];
232
233
		dev->caps.suggested_type[i] = dev_cap->suggested_type[i];
		dev->caps.default_sense[i] = dev_cap->default_sense[i];
234
235
236
237
		dev->caps.trans_type[i]	    = dev_cap->trans_type[i];
		dev->caps.vendor_oui[i]     = dev_cap->vendor_oui[i];
		dev->caps.wavelength[i]     = dev_cap->wavelength[i];
		dev->caps.trans_code[i]     = dev_cap->trans_code[i];
238
239
	}

240
	dev->caps.uar_page_size	     = PAGE_SIZE;
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
	dev->caps.num_uars	     = dev_cap->uar_size / PAGE_SIZE;
	dev->caps.local_ca_ack_delay = dev_cap->local_ca_ack_delay;
	dev->caps.bf_reg_size	     = dev_cap->bf_reg_size;
	dev->caps.bf_regs_per_page   = dev_cap->bf_regs_per_page;
	dev->caps.max_sq_sg	     = dev_cap->max_sq_sg;
	dev->caps.max_rq_sg	     = dev_cap->max_rq_sg;
	dev->caps.max_wqes	     = dev_cap->max_qp_sz;
	dev->caps.max_qp_init_rdma   = dev_cap->max_requester_per_qp;
	dev->caps.max_srq_wqes	     = dev_cap->max_srq_sz;
	dev->caps.max_srq_sge	     = dev_cap->max_rq_sg - 1;
	dev->caps.reserved_srqs	     = dev_cap->reserved_srqs;
	dev->caps.max_sq_desc_sz     = dev_cap->max_sq_desc_sz;
	dev->caps.max_rq_desc_sz     = dev_cap->max_rq_desc_sz;
	/*
	 * Subtract 1 from the limit because we need to allocate a
	 * spare CQE so the HCA HW can tell the difference between an
	 * empty CQ and a full CQ.
	 */
	dev->caps.max_cqes	     = dev_cap->max_cq_sz - 1;
	dev->caps.reserved_cqs	     = dev_cap->reserved_cqs;
	dev->caps.reserved_eqs	     = dev_cap->reserved_eqs;
262
	dev->caps.reserved_mtts      = dev_cap->reserved_mtts;
263
	dev->caps.reserved_mrws	     = dev_cap->reserved_mrws;
264
265
266

	/* The first 128 UARs are used for EQ doorbells */
	dev->caps.reserved_uars	     = max_t(int, 128, dev_cap->reserved_uars);
267
	dev->caps.reserved_pds	     = dev_cap->reserved_pds;
268
269
270
271
	dev->caps.reserved_xrcds     = (dev->caps.flags & MLX4_DEV_CAP_FLAG_XRC) ?
					dev_cap->reserved_xrcds : 0;
	dev->caps.max_xrcds          = (dev->caps.flags & MLX4_DEV_CAP_FLAG_XRC) ?
					dev_cap->max_xrcds : 0;
272
273
	dev->caps.mtt_entry_sz       = dev_cap->mtt_entry_sz;

274
	dev->caps.max_msg_sz         = dev_cap->max_msg_sz;
275
276
	dev->caps.page_size_cap	     = ~(u32) (dev_cap->min_page_sz - 1);
	dev->caps.flags		     = dev_cap->flags;
277
	dev->caps.flags2	     = dev_cap->flags2;
278
279
	dev->caps.bmme_flags	     = dev_cap->bmme_flags;
	dev->caps.reserved_lkey	     = dev_cap->reserved_lkey;
280
	dev->caps.stat_rate_support  = dev_cap->stat_rate_support;
Eli Cohen's avatar
Eli Cohen committed
281
	dev->caps.max_gso_sz	     = dev_cap->max_gso_sz;
282
	dev->caps.max_rss_tbl_sz     = dev_cap->max_rss_tbl_sz;
283

284
285
	/* Sense port always allowed on supported devices for ConnectX-1 and -2 */
	if (mlx4_priv(dev)->pci_dev_data & MLX4_PCI_DEV_FORCE_SENSE_PORT)
286
		dev->caps.flags |= MLX4_DEV_CAP_FLAG_SENSE_SUPPORT;
287
288
289
	/* Don't do sense port on multifunction devices (for now at least) */
	if (mlx4_is_mfunc(dev))
		dev->caps.flags &= ~MLX4_DEV_CAP_FLAG_SENSE_SUPPORT;
290

291
	dev->caps.log_num_macs  = log_num_mac;
292
	dev->caps.log_num_vlans = MLX4_LOG_NUM_VLANS;
293
294
295
	dev->caps.log_num_prios = use_prio ? 3 : 0;

	for (i = 1; i <= dev->caps.num_ports; ++i) {
296
297
298
299
300
		dev->caps.port_type[i] = MLX4_PORT_TYPE_NONE;
		if (dev->caps.supported_type[i]) {
			/* if only ETH is supported - assign ETH */
			if (dev->caps.supported_type[i] == MLX4_PORT_TYPE_ETH)
				dev->caps.port_type[i] = MLX4_PORT_TYPE_ETH;
301
			/* if only IB is supported, assign IB */
302
			else if (dev->caps.supported_type[i] ==
303
304
				 MLX4_PORT_TYPE_IB)
				dev->caps.port_type[i] = MLX4_PORT_TYPE_IB;
305
			else {
306
307
308
309
				/* if IB and ETH are supported, we set the port
				 * type according to user selection of port type;
				 * if user selected none, take the FW hint */
				if (port_type_array[i - 1] == MLX4_PORT_TYPE_NONE)
310
311
					dev->caps.port_type[i] = dev->caps.suggested_type[i] ?
						MLX4_PORT_TYPE_ETH : MLX4_PORT_TYPE_IB;
312
				else
313
					dev->caps.port_type[i] = port_type_array[i - 1];
314
315
			}
		}
316
317
318
319
320
321
		/*
		 * Link sensing is allowed on the port if 3 conditions are true:
		 * 1. Both protocols are supported on the port.
		 * 2. Different types are supported on the port
		 * 3. FW declared that it supports link sensing
		 */
322
		mlx4_priv(dev)->sense.sense_allowed[i] =
323
			((dev->caps.supported_type[i] == MLX4_PORT_TYPE_AUTO) &&
324
			 (dev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP) &&
325
			 (dev->caps.flags & MLX4_DEV_CAP_FLAG_SENSE_SUPPORT));
326

327
328
329
330
331
		/*
		 * If "default_sense" bit is set, we move the port to "AUTO" mode
		 * and perform sense_port FW command to try and set the correct
		 * port type from beginning
		 */
332
		if (mlx4_priv(dev)->sense.sense_allowed[i] && dev->caps.default_sense[i]) {
333
334
335
336
337
338
339
340
341
			enum mlx4_port_type sensed_port = MLX4_PORT_TYPE_NONE;
			dev->caps.possible_type[i] = MLX4_PORT_TYPE_AUTO;
			mlx4_SENSE_PORT(dev, i, &sensed_port);
			if (sensed_port != MLX4_PORT_TYPE_NONE)
				dev->caps.port_type[i] = sensed_port;
		} else {
			dev->caps.possible_type[i] = dev->caps.port_type[i];
		}

342
343
344
345
346
347
348
349
350
351
352
353
354
355
		if (dev->caps.log_num_macs > dev_cap->log_max_macs[i]) {
			dev->caps.log_num_macs = dev_cap->log_max_macs[i];
			mlx4_warn(dev, "Requested number of MACs is too much "
				  "for port %d, reducing to %d.\n",
				  i, 1 << dev->caps.log_num_macs);
		}
		if (dev->caps.log_num_vlans > dev_cap->log_max_vlans[i]) {
			dev->caps.log_num_vlans = dev_cap->log_max_vlans[i];
			mlx4_warn(dev, "Requested number of VLANs is too much "
				  "for port %d, reducing to %d.\n",
				  i, 1 << dev->caps.log_num_vlans);
		}
	}

356
357
	dev->caps.max_counters = 1 << ilog2(dev_cap->max_counters);

358
359
360
361
362
363
364
365
366
367
368
369
370
371
	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] = dev_cap->reserved_qps;
	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] =
		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] =
		(1 << dev->caps.log_num_macs) *
		(1 << dev->caps.log_num_vlans) *
		(1 << dev->caps.log_num_prios) *
		dev->caps.num_ports;
	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH] = MLX4_NUM_FEXCH;

	dev->caps.reserved_qps = dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] +
		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] +
		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] +
		dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH];

372
	dev->caps.sqp_demux = (mlx4_is_master(dev)) ? MLX4_MAX_NUM_SLAVES : 0;
Or Gerlitz's avatar
Or Gerlitz committed
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387

	if (!enable_64b_cqe_eqe) {
		if (dev_cap->flags &
		    (MLX4_DEV_CAP_FLAG_64B_CQE | MLX4_DEV_CAP_FLAG_64B_EQE)) {
			mlx4_warn(dev, "64B EQEs/CQEs supported by the device but not enabled\n");
			dev->caps.flags &= ~MLX4_DEV_CAP_FLAG_64B_CQE;
			dev->caps.flags &= ~MLX4_DEV_CAP_FLAG_64B_EQE;
		}
	}

	if ((dev_cap->flags &
	    (MLX4_DEV_CAP_FLAG_64B_CQE | MLX4_DEV_CAP_FLAG_64B_EQE)) &&
	    mlx4_is_master(dev))
		dev->caps.function_caps |= MLX4_FUNC_CAP_64B_EQE_CQE;

388
389
	return 0;
}
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
/*The function checks if there are live vf, return the num of them*/
static int mlx4_how_many_lives_vf(struct mlx4_dev *dev)
{
	struct mlx4_priv *priv = mlx4_priv(dev);
	struct mlx4_slave_state *s_state;
	int i;
	int ret = 0;

	for (i = 1/*the ppf is 0*/; i < dev->num_slaves; ++i) {
		s_state = &priv->mfunc.master.slave_state[i];
		if (s_state->active && s_state->last_cmd !=
		    MLX4_COMM_CMD_RESET) {
			mlx4_warn(dev, "%s: slave: %d is still active\n",
				  __func__, i);
			ret++;
		}
	}
	return ret;
}

410
411
412
int mlx4_get_parav_qkey(struct mlx4_dev *dev, u32 qpn, u32 *qkey)
{
	u32 qk = MLX4_RESERVED_QKEY_BASE;
413
414
415

	if (qpn >= dev->phys_caps.base_tunnel_sqpn + 8 * MLX4_MFUNC_MAX ||
	    qpn < dev->phys_caps.base_proxy_sqpn)
416
417
		return -EINVAL;

418
	if (qpn >= dev->phys_caps.base_tunnel_sqpn)
419
		/* tunnel qp */
420
		qk += qpn - dev->phys_caps.base_tunnel_sqpn;
421
	else
422
		qk += qpn - dev->phys_caps.base_proxy_sqpn;
423
424
425
426
427
	*qkey = qk;
	return 0;
}
EXPORT_SYMBOL(mlx4_get_parav_qkey);

428
429
430
431
432
433
434
435
436
437
438
void mlx4_sync_pkey_table(struct mlx4_dev *dev, int slave, int port, int i, int val)
{
	struct mlx4_priv *priv = container_of(dev, struct mlx4_priv, dev);

	if (!mlx4_is_master(dev))
		return;

	priv->virt2phys_pkey[slave][port - 1][i] = val;
}
EXPORT_SYMBOL(mlx4_sync_pkey_table);

439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
void mlx4_put_slave_node_guid(struct mlx4_dev *dev, int slave, __be64 guid)
{
	struct mlx4_priv *priv = container_of(dev, struct mlx4_priv, dev);

	if (!mlx4_is_master(dev))
		return;

	priv->slave_node_guids[slave] = guid;
}
EXPORT_SYMBOL(mlx4_put_slave_node_guid);

__be64 mlx4_get_slave_node_guid(struct mlx4_dev *dev, int slave)
{
	struct mlx4_priv *priv = container_of(dev, struct mlx4_priv, dev);

	if (!mlx4_is_master(dev))
		return 0;

	return priv->slave_node_guids[slave];
}
EXPORT_SYMBOL(mlx4_get_slave_node_guid);

461
int mlx4_is_slave_active(struct mlx4_dev *dev, int slave)
462
463
464
465
466
467
468
469
470
471
472
473
{
	struct mlx4_priv *priv = mlx4_priv(dev);
	struct mlx4_slave_state *s_slave;

	if (!mlx4_is_master(dev))
		return 0;

	s_slave = &priv->mfunc.master.slave_state[slave];
	return !!s_slave->active;
}
EXPORT_SYMBOL(mlx4_is_slave_active);

474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
static void slave_adjust_steering_mode(struct mlx4_dev *dev,
				       struct mlx4_dev_cap *dev_cap,
				       struct mlx4_init_hca_param *hca_param)
{
	dev->caps.steering_mode = hca_param->steering_mode;
	if (dev->caps.steering_mode == MLX4_STEERING_MODE_DEVICE_MANAGED) {
		dev->caps.num_qp_per_mgm = dev_cap->fs_max_num_qp_per_entry;
		dev->caps.fs_log_max_ucast_qp_range_size =
			dev_cap->fs_log_max_ucast_qp_range_size;
	} else
		dev->caps.num_qp_per_mgm =
			4 * ((1 << hca_param->log_mc_entry_sz)/16 - 2);

	mlx4_dbg(dev, "Steering mode is: %s\n",
		 mlx4_steering_mode_str(dev->caps.steering_mode));
}

491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
static int mlx4_slave_cap(struct mlx4_dev *dev)
{
	int			   err;
	u32			   page_size;
	struct mlx4_dev_cap	   dev_cap;
	struct mlx4_func_cap	   func_cap;
	struct mlx4_init_hca_param hca_param;
	int			   i;

	memset(&hca_param, 0, sizeof(hca_param));
	err = mlx4_QUERY_HCA(dev, &hca_param);
	if (err) {
		mlx4_err(dev, "QUERY_HCA command failed, aborting.\n");
		return err;
	}

	/*fail if the hca has an unknown capability */
	if ((hca_param.global_caps | HCA_GLOBAL_CAP_MASK) !=
	    HCA_GLOBAL_CAP_MASK) {
		mlx4_err(dev, "Unknown hca global capabilities\n");
		return -ENOSYS;
	}

	mlx4_log_num_mgm_entry_size = hca_param.log_mc_entry_sz;

	memset(&dev_cap, 0, sizeof(dev_cap));
517
	dev->caps.max_qp_dest_rdma = 1 << hca_param.log_rd_per_qp;
518
519
520
521
522
523
	err = mlx4_dev_cap(dev, &dev_cap);
	if (err) {
		mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
		return err;
	}

524
525
526
527
	err = mlx4_QUERY_FW(dev);
	if (err)
		mlx4_err(dev, "QUERY_FW command failed: could not get FW version.\n");

528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
	page_size = ~dev->caps.page_size_cap + 1;
	mlx4_warn(dev, "HCA minimum page size:%d\n", page_size);
	if (page_size > PAGE_SIZE) {
		mlx4_err(dev, "HCA minimum page size of %d bigger than "
			 "kernel PAGE_SIZE of %ld, aborting.\n",
			 page_size, PAGE_SIZE);
		return -ENODEV;
	}

	/* slave gets uar page size from QUERY_HCA fw command */
	dev->caps.uar_page_size = 1 << (hca_param.uar_page_sz + 12);

	/* TODO: relax this assumption */
	if (dev->caps.uar_page_size != PAGE_SIZE) {
		mlx4_err(dev, "UAR size:%d != kernel PAGE_SIZE of %ld\n",
			 dev->caps.uar_page_size, PAGE_SIZE);
		return -ENODEV;
	}

	memset(&func_cap, 0, sizeof(func_cap));
548
	err = mlx4_QUERY_FUNC_CAP(dev, 0, &func_cap);
549
	if (err) {
550
551
		mlx4_err(dev, "QUERY_FUNC_CAP general command failed, aborting (%d).\n",
			  err);
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
		return err;
	}

	if ((func_cap.pf_context_behaviour | PF_CONTEXT_BEHAVIOUR_MASK) !=
	    PF_CONTEXT_BEHAVIOUR_MASK) {
		mlx4_err(dev, "Unknown pf context behaviour\n");
		return -ENOSYS;
	}

	dev->caps.num_ports		= func_cap.num_ports;
	dev->caps.num_qps		= func_cap.qp_quota;
	dev->caps.num_srqs		= func_cap.srq_quota;
	dev->caps.num_cqs		= func_cap.cq_quota;
	dev->caps.num_eqs               = func_cap.max_eq;
	dev->caps.reserved_eqs          = func_cap.reserved_eq;
	dev->caps.num_mpts		= func_cap.mpt_quota;
	dev->caps.num_mtts		= func_cap.mtt_quota;
	dev->caps.num_pds               = MLX4_NUM_PDS;
	dev->caps.num_mgms              = 0;
	dev->caps.num_amgms             = 0;

	if (dev->caps.num_ports > MLX4_MAX_PORTS) {
		mlx4_err(dev, "HCA has %d ports, but we only support %d, "
			 "aborting.\n", dev->caps.num_ports, MLX4_MAX_PORTS);
		return -ENODEV;
	}

579
580
581
582
583
584
585
586
587
588
589
	dev->caps.qp0_tunnel = kcalloc(dev->caps.num_ports, sizeof (u32), GFP_KERNEL);
	dev->caps.qp0_proxy = kcalloc(dev->caps.num_ports, sizeof (u32), GFP_KERNEL);
	dev->caps.qp1_tunnel = kcalloc(dev->caps.num_ports, sizeof (u32), GFP_KERNEL);
	dev->caps.qp1_proxy = kcalloc(dev->caps.num_ports, sizeof (u32), GFP_KERNEL);

	if (!dev->caps.qp0_tunnel || !dev->caps.qp0_proxy ||
	    !dev->caps.qp1_tunnel || !dev->caps.qp1_proxy) {
		err = -ENOMEM;
		goto err_mem;
	}

590
	for (i = 1; i <= dev->caps.num_ports; ++i) {
591
592
593
594
595
596
597
598
599
600
		err = mlx4_QUERY_FUNC_CAP(dev, (u32) i, &func_cap);
		if (err) {
			mlx4_err(dev, "QUERY_FUNC_CAP port command failed for"
				 " port %d, aborting (%d).\n", i, err);
			goto err_mem;
		}
		dev->caps.qp0_tunnel[i - 1] = func_cap.qp0_tunnel_qpn;
		dev->caps.qp0_proxy[i - 1] = func_cap.qp0_proxy_qpn;
		dev->caps.qp1_tunnel[i - 1] = func_cap.qp1_tunnel_qpn;
		dev->caps.qp1_proxy[i - 1] = func_cap.qp1_proxy_qpn;
601
		dev->caps.port_mask[i] = dev->caps.port_type[i];
602
603
604
		if (mlx4_get_slave_pkey_gid_tbl_len(dev, i,
						    &dev->caps.gid_table_len[i],
						    &dev->caps.pkey_table_len[i]))
605
			goto err_mem;
606
	}
607

608
609
610
611
612
613
614
	if (dev->caps.uar_page_size * (dev->caps.num_uars -
				       dev->caps.reserved_uars) >
				       pci_resource_len(dev->pdev, 2)) {
		mlx4_err(dev, "HCA reported UAR region size of 0x%x bigger than "
			 "PCI resource 2 size of 0x%llx, aborting.\n",
			 dev->caps.uar_page_size * dev->caps.num_uars,
			 (unsigned long long) pci_resource_len(dev->pdev, 2));
615
		goto err_mem;
616
617
	}

Or Gerlitz's avatar
Or Gerlitz committed
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
	if (hca_param.dev_cap_enabled & MLX4_DEV_CAP_64B_EQE_ENABLED) {
		dev->caps.eqe_size   = 64;
		dev->caps.eqe_factor = 1;
	} else {
		dev->caps.eqe_size   = 32;
		dev->caps.eqe_factor = 0;
	}

	if (hca_param.dev_cap_enabled & MLX4_DEV_CAP_64B_CQE_ENABLED) {
		dev->caps.cqe_size   = 64;
		dev->caps.userspace_caps |= MLX4_USER_DEV_CAP_64B_CQE;
	} else {
		dev->caps.cqe_size   = 32;
	}

633
634
	slave_adjust_steering_mode(dev, &dev_cap, &hca_param);

635
	return 0;
636
637
638
639
640
641
642
643
644
645

err_mem:
	kfree(dev->caps.qp0_tunnel);
	kfree(dev->caps.qp0_proxy);
	kfree(dev->caps.qp1_tunnel);
	kfree(dev->caps.qp1_proxy);
	dev->caps.qp0_tunnel = dev->caps.qp0_proxy =
		dev->caps.qp1_tunnel = dev->caps.qp1_proxy = NULL;

	return err;
646
}
647

648
649
650
651
/*
 * Change the port configuration of the device.
 * Every user of this function must hold the port mutex.
 */
652
653
int mlx4_change_port_types(struct mlx4_dev *dev,
			   enum mlx4_port_type *port_types)
654
655
656
657
658
659
{
	int err = 0;
	int change = 0;
	int port;

	for (port = 0; port <  dev->caps.num_ports; port++) {
660
661
		/* Change the port type only if the new type is different
		 * from the current, and not set to Auto */
662
		if (port_types[port] != dev->caps.port_type[port + 1])
663
664
665
666
667
668
			change = 1;
	}
	if (change) {
		mlx4_unregister_device(dev);
		for (port = 1; port <= dev->caps.num_ports; port++) {
			mlx4_CLOSE_PORT(dev, port);
669
			dev->caps.port_type[port] = port_types[port - 1];
670
			err = mlx4_SET_PORT(dev, port, -1);
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
			if (err) {
				mlx4_err(dev, "Failed to set port %d, "
					      "aborting\n", port);
				goto out;
			}
		}
		mlx4_set_port_mask(dev);
		err = mlx4_register_device(dev);
	}

out:
	return err;
}

static ssize_t show_port_type(struct device *dev,
			      struct device_attribute *attr,
			      char *buf)
{
	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
						   port_attr);
	struct mlx4_dev *mdev = info->dev;
692
693
694
695
696
697
698
699
700
	char type[8];

	sprintf(type, "%s",
		(mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_IB) ?
		"ib" : "eth");
	if (mdev->caps.possible_type[info->port] == MLX4_PORT_TYPE_AUTO)
		sprintf(buf, "auto (%s)\n", type);
	else
		sprintf(buf, "%s\n", type);
701

702
	return strlen(buf);
703
704
705
706
707
708
709
710
711
712
713
}

static ssize_t set_port_type(struct device *dev,
			     struct device_attribute *attr,
			     const char *buf, size_t count)
{
	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
						   port_attr);
	struct mlx4_dev *mdev = info->dev;
	struct mlx4_priv *priv = mlx4_priv(mdev);
	enum mlx4_port_type types[MLX4_MAX_PORTS];
714
	enum mlx4_port_type new_types[MLX4_MAX_PORTS];
715
716
717
718
719
720
721
	int i;
	int err = 0;

	if (!strcmp(buf, "ib\n"))
		info->tmp_type = MLX4_PORT_TYPE_IB;
	else if (!strcmp(buf, "eth\n"))
		info->tmp_type = MLX4_PORT_TYPE_ETH;
722
723
	else if (!strcmp(buf, "auto\n"))
		info->tmp_type = MLX4_PORT_TYPE_AUTO;
724
725
726
727
728
	else {
		mlx4_err(mdev, "%s is not supported port type\n", buf);
		return -EINVAL;
	}

729
	mlx4_stop_sense(mdev);
730
	mutex_lock(&priv->port_mutex);
731
732
733
734
	/* Possible type is always the one that was delivered */
	mdev->caps.possible_type[info->port] = info->tmp_type;

	for (i = 0; i < mdev->caps.num_ports; i++) {
735
		types[i] = priv->port[i+1].tmp_type ? priv->port[i+1].tmp_type :
736
737
738
739
					mdev->caps.possible_type[i+1];
		if (types[i] == MLX4_PORT_TYPE_AUTO)
			types[i] = mdev->caps.port_type[i+1];
	}
740

741
742
	if (!(mdev->caps.flags & MLX4_DEV_CAP_FLAG_DPDP) &&
	    !(mdev->caps.flags & MLX4_DEV_CAP_FLAG_SENSE_SUPPORT)) {
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
		for (i = 1; i <= mdev->caps.num_ports; i++) {
			if (mdev->caps.possible_type[i] == MLX4_PORT_TYPE_AUTO) {
				mdev->caps.possible_type[i] = mdev->caps.port_type[i];
				err = -EINVAL;
			}
		}
	}
	if (err) {
		mlx4_err(mdev, "Auto sensing is not supported on this HCA. "
			       "Set only 'eth' or 'ib' for both ports "
			       "(should be the same)\n");
		goto out;
	}

	mlx4_do_sense_ports(mdev, new_types, types);

	err = mlx4_check_port_params(mdev, new_types);
760
761
762
	if (err)
		goto out;

763
764
765
766
767
	/* We are about to apply the changes after the configuration
	 * was verified, no need to remember the temporary types
	 * any more */
	for (i = 0; i < mdev->caps.num_ports; i++)
		priv->port[i + 1].tmp_type = 0;
768

769
	err = mlx4_change_port_types(mdev, new_types);
770
771

out:
772
	mlx4_start_sense(mdev);
773
774
775
776
	mutex_unlock(&priv->port_mutex);
	return err ? err : count;
}

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
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
enum ibta_mtu {
	IB_MTU_256  = 1,
	IB_MTU_512  = 2,
	IB_MTU_1024 = 3,
	IB_MTU_2048 = 4,
	IB_MTU_4096 = 5
};

static inline int int_to_ibta_mtu(int mtu)
{
	switch (mtu) {
	case 256:  return IB_MTU_256;
	case 512:  return IB_MTU_512;
	case 1024: return IB_MTU_1024;
	case 2048: return IB_MTU_2048;
	case 4096: return IB_MTU_4096;
	default: return -1;
	}
}

static inline int ibta_mtu_to_int(enum ibta_mtu mtu)
{
	switch (mtu) {
	case IB_MTU_256:  return  256;
	case IB_MTU_512:  return  512;
	case IB_MTU_1024: return 1024;
	case IB_MTU_2048: return 2048;
	case IB_MTU_4096: return 4096;
	default: return -1;
	}
}

static ssize_t show_port_ib_mtu(struct device *dev,
			     struct device_attribute *attr,
			     char *buf)
{
	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
						   port_mtu_attr);
	struct mlx4_dev *mdev = info->dev;

	if (mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_ETH)
		mlx4_warn(mdev, "port level mtu is only used for IB ports\n");

	sprintf(buf, "%d\n",
			ibta_mtu_to_int(mdev->caps.port_ib_mtu[info->port]));
	return strlen(buf);
}

static ssize_t set_port_ib_mtu(struct device *dev,
			     struct device_attribute *attr,
			     const char *buf, size_t count)
{
	struct mlx4_port_info *info = container_of(attr, struct mlx4_port_info,
						   port_mtu_attr);
	struct mlx4_dev *mdev = info->dev;
	struct mlx4_priv *priv = mlx4_priv(mdev);
	int err, port, mtu, ibta_mtu = -1;

	if (mdev->caps.port_type[info->port] == MLX4_PORT_TYPE_ETH) {
		mlx4_warn(mdev, "port level mtu is only used for IB ports\n");
		return -EINVAL;
	}

	err = sscanf(buf, "%d", &mtu);
	if (err > 0)
		ibta_mtu = int_to_ibta_mtu(mtu);

	if (err <= 0 || ibta_mtu < 0) {
		mlx4_err(mdev, "%s is invalid IBTA mtu\n", buf);
		return -EINVAL;
	}

	mdev->caps.port_ib_mtu[info->port] = ibta_mtu;

	mlx4_stop_sense(mdev);
	mutex_lock(&priv->port_mutex);
	mlx4_unregister_device(mdev);
	for (port = 1; port <= mdev->caps.num_ports; port++) {
		mlx4_CLOSE_PORT(mdev, port);
856
		err = mlx4_SET_PORT(mdev, port, -1);
857
858
859
860
861
862
863
864
865
866
867
868
869
		if (err) {
			mlx4_err(mdev, "Failed to set port %d, "
				      "aborting\n", port);
			goto err_set_port;
		}
	}
	err = mlx4_register_device(mdev);
err_set_port:
	mutex_unlock(&priv->port_mutex);
	mlx4_start_sense(mdev);
	return err ? err : count;
}

870
static int mlx4_load_fw(struct mlx4_dev *dev)
871
872
873
874
875
{
	struct mlx4_priv *priv = mlx4_priv(dev);
	int err;

	priv->fw.fw_icm = mlx4_alloc_icm(dev, priv->fw.fw_pages,
876
					 GFP_HIGHUSER | __GFP_NOWARN, 0);
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
	if (!priv->fw.fw_icm) {
		mlx4_err(dev, "Couldn't allocate FW area, aborting.\n");
		return -ENOMEM;
	}

	err = mlx4_MAP_FA(dev, priv->fw.fw_icm);
	if (err) {
		mlx4_err(dev, "MAP_FA command failed, aborting.\n");
		goto err_free;
	}

	err = mlx4_RUN_FW(dev);
	if (err) {
		mlx4_err(dev, "RUN_FW command failed, aborting.\n");
		goto err_unmap_fa;
	}

	return 0;

err_unmap_fa:
	mlx4_UNMAP_FA(dev);

err_free:
900
	mlx4_free_icm(dev, priv->fw.fw_icm, 0);
901
902
903
	return err;
}

904
905
static int mlx4_init_cmpt_table(struct mlx4_dev *dev, u64 cmpt_base,
				int cmpt_entry_sz)
906
907
908
{
	struct mlx4_priv *priv = mlx4_priv(dev);
	int err;
909
	int num_eqs;
910
911
912
913
914
915

	err = mlx4_init_icm_table(dev, &priv->qp_table.cmpt_table,
				  cmpt_base +
				  ((u64) (MLX4_CMPT_TYPE_QP *
					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
				  cmpt_entry_sz, dev->caps.num_qps,
916
917
				  dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW],
				  0, 0);
918
919
920
921
922
923
924
925
	if (err)
		goto err;

	err = mlx4_init_icm_table(dev, &priv->srq_table.cmpt_table,
				  cmpt_base +
				  ((u64) (MLX4_CMPT_TYPE_SRQ *
					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
				  cmpt_entry_sz, dev->caps.num_srqs,
926
				  dev->caps.reserved_srqs, 0, 0);
927
928
929
930
931
932
933
934
	if (err)
		goto err_qp;

	err = mlx4_init_icm_table(dev, &priv->cq_table.cmpt_table,
				  cmpt_base +
				  ((u64) (MLX4_CMPT_TYPE_CQ *
					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
				  cmpt_entry_sz, dev->caps.num_cqs,
935
				  dev->caps.reserved_cqs, 0, 0);
936
937
938
	if (err)
		goto err_srq;

939
940
	num_eqs = (mlx4_is_master(dev)) ? dev->phys_caps.num_phys_eqs :
		  dev->caps.num_eqs;
941
942
943
944
	err = mlx4_init_icm_table(dev, &priv->eq_table.cmpt_table,
				  cmpt_base +
				  ((u64) (MLX4_CMPT_TYPE_EQ *
					  cmpt_entry_sz) << MLX4_CMPT_SHIFT),
945
				  cmpt_entry_sz, num_eqs, num_eqs, 0, 0);
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
	if (err)
		goto err_cq;

	return 0;

err_cq:
	mlx4_cleanup_icm_table(dev, &priv->cq_table.cmpt_table);

err_srq:
	mlx4_cleanup_icm_table(dev, &priv->srq_table.cmpt_table);

err_qp:
	mlx4_cleanup_icm_table(dev, &priv->qp_table.cmpt_table);

err:
	return err;
}

964
965
static int mlx4_init_icm(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap,
			 struct mlx4_init_hca_param *init_hca, u64 icm_size)
966
967
968
{
	struct mlx4_priv *priv = mlx4_priv(dev);
	u64 aux_pages;
969
	int num_eqs;
970
971
972
973
974
975
976
977
978
979
980
981
982
	int err;

	err = mlx4_SET_ICM_SIZE(dev, icm_size, &aux_pages);
	if (err) {
		mlx4_err(dev, "SET_ICM_SIZE command failed, aborting.\n");
		return err;
	}

	mlx4_dbg(dev, "%lld KB of HCA context requires %lld KB aux memory.\n",
		 (unsigned long long) icm_size >> 10,
		 (unsigned long long) aux_pages << 2);

	priv->fw.aux_icm = mlx4_alloc_icm(dev, aux_pages,
983
					  GFP_HIGHUSER | __GFP_NOWARN, 0);
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
	if (!priv->fw.aux_icm) {
		mlx4_err(dev, "Couldn't allocate aux memory, aborting.\n");
		return -ENOMEM;
	}

	err = mlx4_MAP_ICM_AUX(dev, priv->fw.aux_icm);
	if (err) {
		mlx4_err(dev, "MAP_ICM_AUX command failed, aborting.\n");
		goto err_free_aux;
	}

	err = mlx4_init_cmpt_table(dev, init_hca->cmpt_base, dev_cap->cmpt_entry_sz);
	if (err) {
		mlx4_err(dev, "Failed to map cMPT context memory, aborting.\n");
		goto err_unmap_aux;
	}