main.c 37.3 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 46 47 48
/*
 * 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/sched.h>
#include <rdma/ib_user_verbs.h>
#include <rdma/ib_smi.h>
#include <rdma/ib_umem.h>
#include "user.h"
#include "mlx5_ib.h"

#define DRIVER_NAME "mlx5_ib"
49 50
#define DRIVER_VERSION "2.2-1"
#define DRIVER_RELDATE	"Feb 2014"
51 52 53 54 55 56

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

57 58 59
static int deprecated_prof_sel = 2;
module_param_named(prof_sel, deprecated_prof_sel, int, 0444);
MODULE_PARM_DESC(prof_sel, "profile selector. Deprecated here. Moved to module mlx5_core");
60 61 62 63 64 65 66

static char mlx5_version[] =
	DRIVER_NAME ": Mellanox Connect-IB Infiniband driver v"
	DRIVER_VERSION " (" DRIVER_RELDATE ")\n";

int mlx5_vector2eqn(struct mlx5_ib_dev *dev, int vector, int *eqn, int *irqn)
{
67
	struct mlx5_eq_table *table = &dev->mdev->priv.eq_table;
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
	struct mlx5_eq *eq, *n;
	int err = -ENOENT;

	spin_lock(&table->lock);
	list_for_each_entry_safe(eq, n, &dev->eqs_list, list) {
		if (eq->index == vector) {
			*eqn = eq->eqn;
			*irqn = eq->irqn;
			err = 0;
			break;
		}
	}
	spin_unlock(&table->lock);

	return err;
}

static int alloc_comp_eqs(struct mlx5_ib_dev *dev)
{
87
	struct mlx5_eq_table *table = &dev->mdev->priv.eq_table;
88
	char name[MLX5_MAX_EQ_NAME];
89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
	struct mlx5_eq *eq, *n;
	int ncomp_vec;
	int nent;
	int err;
	int i;

	INIT_LIST_HEAD(&dev->eqs_list);
	ncomp_vec = table->num_comp_vectors;
	nent = MLX5_COMP_EQ_SIZE;
	for (i = 0; i < ncomp_vec; i++) {
		eq = kzalloc(sizeof(*eq), GFP_KERNEL);
		if (!eq) {
			err = -ENOMEM;
			goto clean;
		}

105
		snprintf(name, MLX5_MAX_EQ_NAME, "mlx5_comp%d", i);
106
		err = mlx5_create_map_eq(dev->mdev, eq,
107
					 i + MLX5_EQ_VEC_COMP_BASE, nent, 0,
108
					 name, &dev->mdev->priv.uuari.uars[0]);
109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
		if (err) {
			kfree(eq);
			goto clean;
		}
		mlx5_ib_dbg(dev, "allocated completion EQN %d\n", eq->eqn);
		eq->index = i;
		spin_lock(&table->lock);
		list_add_tail(&eq->list, &dev->eqs_list);
		spin_unlock(&table->lock);
	}

	dev->num_comp_vectors = ncomp_vec;
	return 0;

clean:
	spin_lock(&table->lock);
	list_for_each_entry_safe(eq, n, &dev->eqs_list, list) {
		list_del(&eq->list);
		spin_unlock(&table->lock);
128
		if (mlx5_destroy_unmap_eq(dev->mdev, eq))
129 130 131 132 133 134 135 136 137 138
			mlx5_ib_warn(dev, "failed to destroy EQ 0x%x\n", eq->eqn);
		kfree(eq);
		spin_lock(&table->lock);
	}
	spin_unlock(&table->lock);
	return err;
}

static void free_comp_eqs(struct mlx5_ib_dev *dev)
{
139
	struct mlx5_eq_table *table = &dev->mdev->priv.eq_table;
140 141 142 143 144 145
	struct mlx5_eq *eq, *n;

	spin_lock(&table->lock);
	list_for_each_entry_safe(eq, n, &dev->eqs_list, list) {
		list_del(&eq->list);
		spin_unlock(&table->lock);
146
		if (mlx5_destroy_unmap_eq(dev->mdev, eq))
147 148 149 150 151 152 153 154 155 156 157 158 159
			mlx5_ib_warn(dev, "failed to destroy EQ 0x%x\n", eq->eqn);
		kfree(eq);
		spin_lock(&table->lock);
	}
	spin_unlock(&table->lock);
}

static int mlx5_ib_query_device(struct ib_device *ibdev,
				struct ib_device_attr *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct ib_smp *in_mad  = NULL;
	struct ib_smp *out_mad = NULL;
160
	struct mlx5_general_caps *gen;
161 162 163 164 165
	int err = -ENOMEM;
	int max_rq_sg;
	int max_sq_sg;
	u64 flags;

166
	gen = &dev->mdev->caps.gen;
167 168 169 170 171 172 173 174 175 176 177 178 179 180
	in_mad  = kzalloc(sizeof(*in_mad), GFP_KERNEL);
	out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
	if (!in_mad || !out_mad)
		goto out;

	init_query_mad(in_mad);
	in_mad->attr_id = IB_SMP_ATTR_NODE_INFO;

	err = mlx5_MAD_IFC(to_mdev(ibdev), 1, 1, 1, NULL, NULL, in_mad, out_mad);
	if (err)
		goto out;

	memset(props, 0, sizeof(*props));

181 182 183
	props->fw_ver = ((u64)fw_rev_maj(dev->mdev) << 32) |
		(fw_rev_min(dev->mdev) << 16) |
		fw_rev_sub(dev->mdev);
184 185 186
	props->device_cap_flags    = IB_DEVICE_CHANGE_PHY_PORT |
		IB_DEVICE_PORT_ACTIVE_EVENT		|
		IB_DEVICE_SYS_IMAGE_GUID		|
187
		IB_DEVICE_RC_RNR_NAK_GEN;
188
	flags = gen->flags;
189 190 191 192 193 194 195 196 197 198
	if (flags & MLX5_DEV_CAP_FLAG_BAD_PKEY_CNTR)
		props->device_cap_flags |= IB_DEVICE_BAD_PKEY_CNTR;
	if (flags & MLX5_DEV_CAP_FLAG_BAD_QKEY_CNTR)
		props->device_cap_flags |= IB_DEVICE_BAD_QKEY_CNTR;
	if (flags & MLX5_DEV_CAP_FLAG_APM)
		props->device_cap_flags |= IB_DEVICE_AUTO_PATH_MIG;
	props->device_cap_flags |= IB_DEVICE_LOCAL_DMA_LKEY;
	if (flags & MLX5_DEV_CAP_FLAG_XRC)
		props->device_cap_flags |= IB_DEVICE_XRC;
	props->device_cap_flags |= IB_DEVICE_MEM_MGT_EXTENSIONS;
199 200 201 202 203 204 205 206 207
	if (flags & MLX5_DEV_CAP_FLAG_SIG_HAND_OVER) {
		props->device_cap_flags |= IB_DEVICE_SIGNATURE_HANDOVER;
		/* At this stage no support for signature handover */
		props->sig_prot_cap = IB_PROT_T10DIF_TYPE_1 |
				      IB_PROT_T10DIF_TYPE_2 |
				      IB_PROT_T10DIF_TYPE_3;
		props->sig_guard_cap = IB_GUARD_T10DIF_CRC |
				       IB_GUARD_T10DIF_CSUM;
	}
208 209
	if (flags & MLX5_DEV_CAP_FLAG_BLOCK_MCAST)
		props->device_cap_flags |= IB_DEVICE_BLOCK_MULTICAST_LOOPBACK;
210 211 212 213 214 215 216 217

	props->vendor_id	   = be32_to_cpup((__be32 *)(out_mad->data + 36)) &
		0xffffff;
	props->vendor_part_id	   = be16_to_cpup((__be16 *)(out_mad->data + 30));
	props->hw_ver		   = be32_to_cpup((__be32 *)(out_mad->data + 32));
	memcpy(&props->sys_image_guid, out_mad->data +	4, 8);

	props->max_mr_size	   = ~0ull;
218 219 220 221 222
	props->page_size_cap	   = gen->min_page_sz;
	props->max_qp		   = 1 << gen->log_max_qp;
	props->max_qp_wr	   = gen->max_wqes;
	max_rq_sg = gen->max_rq_desc_sz / sizeof(struct mlx5_wqe_data_seg);
	max_sq_sg = (gen->max_sq_desc_sz - sizeof(struct mlx5_wqe_ctrl_seg)) /
223 224
		sizeof(struct mlx5_wqe_data_seg);
	props->max_sge = min(max_rq_sg, max_sq_sg);
225 226 227 228 229 230 231 232 233
	props->max_cq		   = 1 << gen->log_max_cq;
	props->max_cqe		   = gen->max_cqes - 1;
	props->max_mr		   = 1 << gen->log_max_mkey;
	props->max_pd		   = 1 << gen->log_max_pd;
	props->max_qp_rd_atom	   = 1 << gen->log_max_ra_req_qp;
	props->max_qp_init_rd_atom = 1 << gen->log_max_ra_res_qp;
	props->max_srq		   = 1 << gen->log_max_srq;
	props->max_srq_wr	   = gen->max_srq_wqes - 1;
	props->local_ca_ack_delay  = gen->local_ca_ack_delay;
234 235 236
	props->max_res_rd_atom	   = props->max_qp_rd_atom * props->max_qp;
	props->max_srq_sge	   = max_rq_sg - 1;
	props->max_fast_reg_page_list_len = (unsigned int)-1;
237
	props->local_ca_ack_delay  = gen->local_ca_ack_delay;
E
Eli Cohen 已提交
238 239
	props->atomic_cap	   = IB_ATOMIC_NONE;
	props->masked_atomic_cap   = IB_ATOMIC_NONE;
240
	props->max_pkeys	   = be16_to_cpup((__be16 *)(out_mad->data + 28));
241 242
	props->max_mcast_grp	   = 1 << gen->log_max_mcg;
	props->max_mcast_qp_attach = gen->max_qp_mcg;
243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
	props->max_total_mcast_qp_attach = props->max_mcast_qp_attach *
					   props->max_mcast_grp;
	props->max_map_per_fmr = INT_MAX; /* no limit in ConnectIB */

out:
	kfree(in_mad);
	kfree(out_mad);

	return err;
}

int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
		       struct ib_port_attr *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct ib_smp *in_mad  = NULL;
	struct ib_smp *out_mad = NULL;
260
	struct mlx5_general_caps *gen;
261 262 263
	int ext_active_speed;
	int err = -ENOMEM;

264 265
	gen = &dev->mdev->caps.gen;
	if (port < 1 || port > gen->num_ports) {
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
		mlx5_ib_warn(dev, "invalid port number %d\n", port);
		return -EINVAL;
	}

	in_mad  = kzalloc(sizeof(*in_mad), GFP_KERNEL);
	out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
	if (!in_mad || !out_mad)
		goto out;

	memset(props, 0, sizeof(*props));

	init_query_mad(in_mad);
	in_mad->attr_id  = IB_SMP_ATTR_PORT_INFO;
	in_mad->attr_mod = cpu_to_be32(port);

	err = mlx5_MAD_IFC(dev, 1, 1, port, NULL, NULL, in_mad, out_mad);
	if (err) {
		mlx5_ib_warn(dev, "err %d\n", err);
		goto out;
	}


	props->lid		= be16_to_cpup((__be16 *)(out_mad->data + 16));
	props->lmc		= out_mad->data[34] & 0x7;
	props->sm_lid		= be16_to_cpup((__be16 *)(out_mad->data + 18));
	props->sm_sl		= out_mad->data[36] & 0xf;
	props->state		= out_mad->data[32] & 0xf;
	props->phys_state	= out_mad->data[33] >> 4;
	props->port_cap_flags	= be32_to_cpup((__be32 *)(out_mad->data + 20));
	props->gid_tbl_len	= out_mad->data[50];
296 297
	props->max_msg_sz	= 1 << gen->log_max_msg;
	props->pkey_tbl_len	= gen->port[port - 1].pkey_table_len;
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
	props->bad_pkey_cntr	= be16_to_cpup((__be16 *)(out_mad->data + 46));
	props->qkey_viol_cntr	= be16_to_cpup((__be16 *)(out_mad->data + 48));
	props->active_width	= out_mad->data[31] & 0xf;
	props->active_speed	= out_mad->data[35] >> 4;
	props->max_mtu		= out_mad->data[41] & 0xf;
	props->active_mtu	= out_mad->data[36] >> 4;
	props->subnet_timeout	= out_mad->data[51] & 0x1f;
	props->max_vl_num	= out_mad->data[37] >> 4;
	props->init_type_reply	= out_mad->data[41] >> 4;

	/* Check if extended speeds (EDR/FDR/...) are supported */
	if (props->port_cap_flags & IB_PORT_EXTENDED_SPEEDS_SUP) {
		ext_active_speed = out_mad->data[62] >> 4;

		switch (ext_active_speed) {
		case 1:
			props->active_speed = 16; /* FDR */
			break;
		case 2:
			props->active_speed = 32; /* EDR */
			break;
		}
	}

	/* If reported active speed is QDR, check if is FDR-10 */
	if (props->active_speed == 4) {
324
		if (gen->ext_port_cap[port - 1] &
325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436
		    MLX_EXT_PORT_CAP_FLAG_EXTENDED_PORT_INFO) {
			init_query_mad(in_mad);
			in_mad->attr_id = MLX5_ATTR_EXTENDED_PORT_INFO;
			in_mad->attr_mod = cpu_to_be32(port);

			err = mlx5_MAD_IFC(dev, 1, 1, port,
					   NULL, NULL, in_mad, out_mad);
			if (err)
				goto out;

			/* Checking LinkSpeedActive for FDR-10 */
			if (out_mad->data[15] & 0x1)
				props->active_speed = 8;
		}
	}

out:
	kfree(in_mad);
	kfree(out_mad);

	return err;
}

static int mlx5_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
			     union ib_gid *gid)
{
	struct ib_smp *in_mad  = NULL;
	struct ib_smp *out_mad = NULL;
	int err = -ENOMEM;

	in_mad  = kzalloc(sizeof(*in_mad), GFP_KERNEL);
	out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
	if (!in_mad || !out_mad)
		goto out;

	init_query_mad(in_mad);
	in_mad->attr_id  = IB_SMP_ATTR_PORT_INFO;
	in_mad->attr_mod = cpu_to_be32(port);

	err = mlx5_MAD_IFC(to_mdev(ibdev), 1, 1, port, NULL, NULL, in_mad, out_mad);
	if (err)
		goto out;

	memcpy(gid->raw, out_mad->data + 8, 8);

	init_query_mad(in_mad);
	in_mad->attr_id  = IB_SMP_ATTR_GUID_INFO;
	in_mad->attr_mod = cpu_to_be32(index / 8);

	err = mlx5_MAD_IFC(to_mdev(ibdev), 1, 1, port, NULL, NULL, in_mad, out_mad);
	if (err)
		goto out;

	memcpy(gid->raw + 8, out_mad->data + (index % 8) * 8, 8);

out:
	kfree(in_mad);
	kfree(out_mad);
	return err;
}

static int mlx5_ib_query_pkey(struct ib_device *ibdev, u8 port, u16 index,
			      u16 *pkey)
{
	struct ib_smp *in_mad  = NULL;
	struct ib_smp *out_mad = NULL;
	int err = -ENOMEM;

	in_mad  = kzalloc(sizeof(*in_mad), GFP_KERNEL);
	out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
	if (!in_mad || !out_mad)
		goto out;

	init_query_mad(in_mad);
	in_mad->attr_id  = IB_SMP_ATTR_PKEY_TABLE;
	in_mad->attr_mod = cpu_to_be32(index / 32);

	err = mlx5_MAD_IFC(to_mdev(ibdev), 1, 1, port, NULL, NULL, in_mad, out_mad);
	if (err)
		goto out;

	*pkey = be16_to_cpu(((__be16 *)out_mad->data)[index % 32]);

out:
	kfree(in_mad);
	kfree(out_mad);
	return err;
}

struct mlx5_reg_node_desc {
	u8	desc[64];
};

static int mlx5_ib_modify_device(struct ib_device *ibdev, int mask,
				 struct ib_device_modify *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct mlx5_reg_node_desc in;
	struct mlx5_reg_node_desc out;
	int err;

	if (mask & ~IB_DEVICE_MODIFY_NODE_DESC)
		return -EOPNOTSUPP;

	if (!(mask & IB_DEVICE_MODIFY_NODE_DESC))
		return 0;

	/*
	 * If possible, pass node desc to FW, so it can generate
	 * a 144 trap.  If cmd fails, just ignore.
	 */
	memcpy(&in, props->node_desc, 64);
437
	err = mlx5_core_access_reg(dev->mdev, &in, sizeof(in), &out,
438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
				   sizeof(out), MLX5_REG_NODE_DESC, 0, 1);
	if (err)
		return err;

	memcpy(ibdev->node_desc, props->node_desc, 64);

	return err;
}

static int mlx5_ib_modify_port(struct ib_device *ibdev, u8 port, int mask,
			       struct ib_port_modify *props)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
	struct ib_port_attr attr;
	u32 tmp;
	int err;

	mutex_lock(&dev->cap_mask_mutex);

	err = mlx5_ib_query_port(ibdev, port, &attr);
	if (err)
		goto out;

	tmp = (attr.port_cap_flags | props->set_port_cap_mask) &
		~props->clr_port_cap_mask;

464
	err = mlx5_set_port_caps(dev->mdev, port, tmp);
465 466 467 468 469 470 471 472 473 474

out:
	mutex_unlock(&dev->cap_mask_mutex);
	return err;
}

static struct ib_ucontext *mlx5_ib_alloc_ucontext(struct ib_device *ibdev,
						  struct ib_udata *udata)
{
	struct mlx5_ib_dev *dev = to_mdev(ibdev);
475
	struct mlx5_ib_alloc_ucontext_req_v2 req;
476 477
	struct mlx5_ib_alloc_ucontext_resp resp;
	struct mlx5_ib_ucontext *context;
478
	struct mlx5_general_caps *gen;
479 480
	struct mlx5_uuar_info *uuari;
	struct mlx5_uar *uars;
E
Eli Cohen 已提交
481
	int gross_uuars;
482
	int num_uars;
483
	int ver;
484 485 486
	int uuarn;
	int err;
	int i;
487
	size_t reqlen;
488

489
	gen = &dev->mdev->caps.gen;
490 491 492
	if (!dev->ib_active)
		return ERR_PTR(-EAGAIN);

493 494 495 496 497 498 499 500 501 502
	memset(&req, 0, sizeof(req));
	reqlen = udata->inlen - sizeof(struct ib_uverbs_cmd_hdr);
	if (reqlen == sizeof(struct mlx5_ib_alloc_ucontext_req))
		ver = 0;
	else if (reqlen == sizeof(struct mlx5_ib_alloc_ucontext_req_v2))
		ver = 2;
	else
		return ERR_PTR(-EINVAL);

	err = ib_copy_from_udata(&req, udata, reqlen);
503 504 505
	if (err)
		return ERR_PTR(err);

506 507 508
	if (req.flags || req.reserved)
		return ERR_PTR(-EINVAL);

509 510 511 512 513 514
	if (req.total_num_uuars > MLX5_MAX_UUARS)
		return ERR_PTR(-ENOMEM);

	if (req.total_num_uuars == 0)
		return ERR_PTR(-EINVAL);

E
Eli Cohen 已提交
515 516
	req.total_num_uuars = ALIGN(req.total_num_uuars,
				    MLX5_NON_FP_BF_REGS_PER_PAGE);
517 518 519
	if (req.num_low_latency_uuars > req.total_num_uuars - 1)
		return ERR_PTR(-EINVAL);

E
Eli Cohen 已提交
520 521
	num_uars = req.total_num_uuars / MLX5_NON_FP_BF_REGS_PER_PAGE;
	gross_uuars = num_uars * MLX5_BF_REGS_PER_PAGE;
522 523
	resp.qp_tab_size      = 1 << gen->log_max_qp;
	resp.bf_reg_size      = gen->bf_reg_size;
524
	resp.cache_line_size  = L1_CACHE_BYTES;
525 526 527 528 529
	resp.max_sq_desc_sz = gen->max_sq_desc_sz;
	resp.max_rq_desc_sz = gen->max_rq_desc_sz;
	resp.max_send_wqebb = gen->max_wqes;
	resp.max_recv_wr = gen->max_wqes;
	resp.max_srq_recv_wr = gen->max_srq_wqes;
530 531 532 533 534 535 536 537 538 539 540 541 542

	context = kzalloc(sizeof(*context), GFP_KERNEL);
	if (!context)
		return ERR_PTR(-ENOMEM);

	uuari = &context->uuari;
	mutex_init(&uuari->lock);
	uars = kcalloc(num_uars, sizeof(*uars), GFP_KERNEL);
	if (!uars) {
		err = -ENOMEM;
		goto out_ctx;
	}

E
Eli Cohen 已提交
543
	uuari->bitmap = kcalloc(BITS_TO_LONGS(gross_uuars),
544 545 546 547 548 549 550 551 552
				sizeof(*uuari->bitmap),
				GFP_KERNEL);
	if (!uuari->bitmap) {
		err = -ENOMEM;
		goto out_uar_ctx;
	}
	/*
	 * clear all fast path uuars
	 */
E
Eli Cohen 已提交
553
	for (i = 0; i < gross_uuars; i++) {
554 555 556 557 558
		uuarn = i & 3;
		if (uuarn == 2 || uuarn == 3)
			set_bit(i, uuari->bitmap);
	}

E
Eli Cohen 已提交
559
	uuari->count = kcalloc(gross_uuars, sizeof(*uuari->count), GFP_KERNEL);
560 561 562 563 564 565
	if (!uuari->count) {
		err = -ENOMEM;
		goto out_bitmap;
	}

	for (i = 0; i < num_uars; i++) {
566
		err = mlx5_cmd_alloc_uar(dev->mdev, &uars[i].index);
567 568 569 570 571 572 573 574
		if (err)
			goto out_count;
	}

	INIT_LIST_HEAD(&context->db_page_list);
	mutex_init(&context->db_page_mutex);

	resp.tot_uuars = req.total_num_uuars;
575
	resp.num_ports = gen->num_ports;
576 577
	err = ib_copy_to_udata(udata, &resp,
			       sizeof(resp) - sizeof(resp.reserved));
578 579 580
	if (err)
		goto out_uars;

581
	uuari->ver = ver;
582 583 584 585 586 587 588
	uuari->num_low_latency_uuars = req.num_low_latency_uuars;
	uuari->uars = uars;
	uuari->num_uars = num_uars;
	return &context->ibucontext;

out_uars:
	for (i--; i >= 0; i--)
589
		mlx5_cmd_free_uar(dev->mdev, uars[i].index);
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
out_count:
	kfree(uuari->count);

out_bitmap:
	kfree(uuari->bitmap);

out_uar_ctx:
	kfree(uars);

out_ctx:
	kfree(context);
	return ERR_PTR(err);
}

static int mlx5_ib_dealloc_ucontext(struct ib_ucontext *ibcontext)
{
	struct mlx5_ib_ucontext *context = to_mucontext(ibcontext);
	struct mlx5_ib_dev *dev = to_mdev(ibcontext->device);
	struct mlx5_uuar_info *uuari = &context->uuari;
	int i;

	for (i = 0; i < uuari->num_uars; i++) {
612
		if (mlx5_cmd_free_uar(dev->mdev, uuari->uars[i].index))
613 614 615 616 617 618 619 620 621 622 623 624 625
			mlx5_ib_warn(dev, "failed to free UAR 0x%x\n", uuari->uars[i].index);
	}

	kfree(uuari->count);
	kfree(uuari->bitmap);
	kfree(uuari->uars);
	kfree(context);

	return 0;
}

static phys_addr_t uar_index2pfn(struct mlx5_ib_dev *dev, int index)
{
626
	return (pci_resource_start(dev->mdev->pdev, 0) >> PAGE_SHIFT) + index;
627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
}

static int get_command(unsigned long offset)
{
	return (offset >> MLX5_IB_MMAP_CMD_SHIFT) & MLX5_IB_MMAP_CMD_MASK;
}

static int get_arg(unsigned long offset)
{
	return offset & ((1 << MLX5_IB_MMAP_CMD_SHIFT) - 1);
}

static int get_index(unsigned long offset)
{
	return get_arg(offset);
}

static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma)
{
	struct mlx5_ib_ucontext *context = to_mucontext(ibcontext);
	struct mlx5_ib_dev *dev = to_mdev(ibcontext->device);
	struct mlx5_uuar_info *uuari = &context->uuari;
	unsigned long command;
	unsigned long idx;
	phys_addr_t pfn;

	command = get_command(vma->vm_pgoff);
	switch (command) {
	case MLX5_IB_MMAP_REGULAR_PAGE:
		if (vma->vm_end - vma->vm_start != PAGE_SIZE)
			return -EINVAL;

		idx = get_index(vma->vm_pgoff);
E
Eli Cohen 已提交
660 661 662
		if (idx >= uuari->num_uars)
			return -EINVAL;

663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
		pfn = uar_index2pfn(dev, uuari->uars[idx].index);
		mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn 0x%llx\n", idx,
			    (unsigned long long)pfn);

		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
		if (io_remap_pfn_range(vma, vma->vm_start, pfn,
				       PAGE_SIZE, vma->vm_page_prot))
			return -EAGAIN;

		mlx5_ib_dbg(dev, "mapped WC at 0x%lx, PA 0x%llx\n",
			    vma->vm_start,
			    (unsigned long long)pfn << PAGE_SHIFT);
		break;

	case MLX5_IB_MMAP_GET_CONTIGUOUS_PAGES:
		return -ENOSYS;

	default:
		return -EINVAL;
	}

	return 0;
}

static int alloc_pa_mkey(struct mlx5_ib_dev *dev, u32 *key, u32 pdn)
{
	struct mlx5_create_mkey_mbox_in *in;
	struct mlx5_mkey_seg *seg;
	struct mlx5_core_mr mr;
	int err;

	in = kzalloc(sizeof(*in), GFP_KERNEL);
	if (!in)
		return -ENOMEM;

	seg = &in->seg;
	seg->flags = MLX5_PERM_LOCAL_READ | MLX5_ACCESS_MODE_PA;
	seg->flags_pd = cpu_to_be32(pdn | MLX5_MKEY_LEN64);
	seg->qpn_mkey7_0 = cpu_to_be32(0xffffff << 8);
	seg->start_addr = 0;

704
	err = mlx5_core_create_mkey(dev->mdev, &mr, in, sizeof(*in),
E
Eli Cohen 已提交
705
				    NULL, NULL, NULL);
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
	if (err) {
		mlx5_ib_warn(dev, "failed to create mkey, %d\n", err);
		goto err_in;
	}

	kfree(in);
	*key = mr.key;

	return 0;

err_in:
	kfree(in);

	return err;
}

static void free_pa_mkey(struct mlx5_ib_dev *dev, u32 key)
{
	struct mlx5_core_mr mr;
	int err;

	memset(&mr, 0, sizeof(mr));
	mr.key = key;
729
	err = mlx5_core_destroy_mkey(dev->mdev, &mr);
730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745
	if (err)
		mlx5_ib_warn(dev, "failed to destroy mkey 0x%x\n", key);
}

static struct ib_pd *mlx5_ib_alloc_pd(struct ib_device *ibdev,
				      struct ib_ucontext *context,
				      struct ib_udata *udata)
{
	struct mlx5_ib_alloc_pd_resp resp;
	struct mlx5_ib_pd *pd;
	int err;

	pd = kmalloc(sizeof(*pd), GFP_KERNEL);
	if (!pd)
		return ERR_PTR(-ENOMEM);

746
	err = mlx5_core_alloc_pd(to_mdev(ibdev)->mdev, &pd->pdn);
747 748 749 750 751 752 753 754
	if (err) {
		kfree(pd);
		return ERR_PTR(err);
	}

	if (context) {
		resp.pdn = pd->pdn;
		if (ib_copy_to_udata(udata, &resp, sizeof(resp))) {
755
			mlx5_core_dealloc_pd(to_mdev(ibdev)->mdev, pd->pdn);
756 757 758 759 760 761
			kfree(pd);
			return ERR_PTR(-EFAULT);
		}
	} else {
		err = alloc_pa_mkey(to_mdev(ibdev), &pd->pa_lkey, pd->pdn);
		if (err) {
762
			mlx5_core_dealloc_pd(to_mdev(ibdev)->mdev, pd->pdn);
763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778
			kfree(pd);
			return ERR_PTR(err);
		}
	}

	return &pd->ibpd;
}

static int mlx5_ib_dealloc_pd(struct ib_pd *pd)
{
	struct mlx5_ib_dev *mdev = to_mdev(pd->device);
	struct mlx5_ib_pd *mpd = to_mpd(pd);

	if (!pd->uobject)
		free_pa_mkey(mdev, mpd->pa_lkey);

779
	mlx5_core_dealloc_pd(mdev->mdev, mpd->pdn);
780 781 782 783 784 785 786 787 788 789
	kfree(mpd);

	return 0;
}

static int mlx5_ib_mcg_attach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
{
	struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
	int err;

790
	err = mlx5_core_attach_mcg(dev->mdev, gid, ibqp->qp_num);
791 792 793 794 795 796 797 798 799 800 801 802
	if (err)
		mlx5_ib_warn(dev, "failed attaching QPN 0x%x, MGID %pI6\n",
			     ibqp->qp_num, gid->raw);

	return err;
}

static int mlx5_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
{
	struct mlx5_ib_dev *dev = to_mdev(ibqp->device);
	int err;

803
	err = mlx5_core_detach_mcg(dev->mdev, gid, ibqp->qp_num);
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
	if (err)
		mlx5_ib_warn(dev, "failed detaching QPN 0x%x, MGID %pI6\n",
			     ibqp->qp_num, gid->raw);

	return err;
}

static int init_node_data(struct mlx5_ib_dev *dev)
{
	struct ib_smp *in_mad  = NULL;
	struct ib_smp *out_mad = NULL;
	int err = -ENOMEM;

	in_mad  = kzalloc(sizeof(*in_mad), GFP_KERNEL);
	out_mad = kmalloc(sizeof(*out_mad), GFP_KERNEL);
	if (!in_mad || !out_mad)
		goto out;

	init_query_mad(in_mad);
	in_mad->attr_id = IB_SMP_ATTR_NODE_DESC;

	err = mlx5_MAD_IFC(dev, 1, 1, 1, NULL, NULL, in_mad, out_mad);
	if (err)
		goto out;

	memcpy(dev->ib_dev.node_desc, out_mad->data, 64);

	in_mad->attr_id = IB_SMP_ATTR_NODE_INFO;

	err = mlx5_MAD_IFC(dev, 1, 1, 1, NULL, NULL, in_mad, out_mad);
	if (err)
		goto out;

837
	dev->mdev->rev_id = be32_to_cpup((__be32 *)(out_mad->data + 32));
838 839 840 841 842 843 844 845 846 847 848 849 850 851
	memcpy(&dev->ib_dev.node_guid, out_mad->data + 12, 8);

out:
	kfree(in_mad);
	kfree(out_mad);
	return err;
}

static ssize_t show_fw_pages(struct device *device, struct device_attribute *attr,
			     char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);

852
	return sprintf(buf, "%d\n", dev->mdev->priv.fw_pages);
853 854 855 856 857 858 859 860
}

static ssize_t show_reg_pages(struct device *device,
			      struct device_attribute *attr, char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);

861
	return sprintf(buf, "%d\n", dev->mdev->priv.reg_pages);
862 863 864 865 866 867 868
}

static ssize_t show_hca(struct device *device, struct device_attribute *attr,
			char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
869
	return sprintf(buf, "MT%d\n", dev->mdev->pdev->device);
870 871 872 873 874 875 876
}

static ssize_t show_fw_ver(struct device *device, struct device_attribute *attr,
			   char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
877 878
	return sprintf(buf, "%d.%d.%d\n", fw_rev_maj(dev->mdev),
		       fw_rev_min(dev->mdev), fw_rev_sub(dev->mdev));
879 880 881 882 883 884 885
}

static ssize_t show_rev(struct device *device, struct device_attribute *attr,
			char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
886
	return sprintf(buf, "%x\n", dev->mdev->rev_id);
887 888 889 890 891 892 893 894
}

static ssize_t show_board(struct device *device, struct device_attribute *attr,
			  char *buf)
{
	struct mlx5_ib_dev *dev =
		container_of(device, struct mlx5_ib_dev, ib_dev.dev);
	return sprintf(buf, "%.*s\n", MLX5_BOARD_ID_LEN,
895
		       dev->mdev->board_id);
896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913
}

static DEVICE_ATTR(hw_rev,   S_IRUGO, show_rev,    NULL);
static DEVICE_ATTR(fw_ver,   S_IRUGO, show_fw_ver, NULL);
static DEVICE_ATTR(hca_type, S_IRUGO, show_hca,    NULL);
static DEVICE_ATTR(board_id, S_IRUGO, show_board,  NULL);
static DEVICE_ATTR(fw_pages, S_IRUGO, show_fw_pages, NULL);
static DEVICE_ATTR(reg_pages, S_IRUGO, show_reg_pages, NULL);

static struct device_attribute *mlx5_class_attributes[] = {
	&dev_attr_hw_rev,
	&dev_attr_fw_ver,
	&dev_attr_hca_type,
	&dev_attr_board_id,
	&dev_attr_fw_pages,
	&dev_attr_reg_pages,
};

914
static void mlx5_ib_event(struct mlx5_core_dev *dev, void *context,
915
			  enum mlx5_dev_event event, unsigned long param)
916
{
917
	struct mlx5_ib_dev *ibdev = (struct mlx5_ib_dev *)context;
918
	struct ib_event ibev;
919

920 921 922 923 924 925 926 927 928 929
	u8 port = 0;

	switch (event) {
	case MLX5_DEV_EVENT_SYS_ERROR:
		ibdev->ib_active = false;
		ibev.event = IB_EVENT_DEVICE_FATAL;
		break;

	case MLX5_DEV_EVENT_PORT_UP:
		ibev.event = IB_EVENT_PORT_ACTIVE;
930
		port = (u8)param;
931 932 933 934
		break;

	case MLX5_DEV_EVENT_PORT_DOWN:
		ibev.event = IB_EVENT_PORT_ERR;
935
		port = (u8)param;
936 937 938 939 940 941 942 943
		break;

	case MLX5_DEV_EVENT_PORT_INITIALIZED:
		/* not used by ULPs */
		return;

	case MLX5_DEV_EVENT_LID_CHANGE:
		ibev.event = IB_EVENT_LID_CHANGE;
944
		port = (u8)param;
945 946 947 948
		break;

	case MLX5_DEV_EVENT_PKEY_CHANGE:
		ibev.event = IB_EVENT_PKEY_CHANGE;
949
		port = (u8)param;
950 951 952 953
		break;

	case MLX5_DEV_EVENT_GUID_CHANGE:
		ibev.event = IB_EVENT_GID_CHANGE;
954
		port = (u8)param;
955 956 957 958
		break;

	case MLX5_DEV_EVENT_CLIENT_REREG:
		ibev.event = IB_EVENT_CLIENT_REREGISTER;
959
		port = (u8)param;
960 961 962 963 964 965
		break;
	}

	ibev.device	      = &ibdev->ib_dev;
	ibev.element.port_num = port;

966 967 968 969 970
	if (port < 1 || port > ibdev->num_ports) {
		mlx5_ib_warn(ibdev, "warning: event on port %d\n", port);
		return;
	}

971 972 973 974 975 976
	if (ibdev->ib_active)
		ib_dispatch_event(&ibev);
}

static void get_ext_port_caps(struct mlx5_ib_dev *dev)
{
977
	struct mlx5_general_caps *gen;
978 979
	int port;

980 981
	gen = &dev->mdev->caps.gen;
	for (port = 1; port <= gen->num_ports; port++)
982 983 984 985 986 987 988
		mlx5_query_ext_port_caps(dev, port);
}

static int get_port_caps(struct mlx5_ib_dev *dev)
{
	struct ib_device_attr *dprops = NULL;
	struct ib_port_attr *pprops = NULL;
989
	struct mlx5_general_caps *gen;
990 991 992
	int err = 0;
	int port;

993
	gen = &dev->mdev->caps.gen;
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007
	pprops = kmalloc(sizeof(*pprops), GFP_KERNEL);
	if (!pprops)
		goto out;

	dprops = kmalloc(sizeof(*dprops), GFP_KERNEL);
	if (!dprops)
		goto out;

	err = mlx5_ib_query_device(&dev->ib_dev, dprops);
	if (err) {
		mlx5_ib_warn(dev, "query_device failed %d\n", err);
		goto out;
	}

1008
	for (port = 1; port <= gen->num_ports; port++) {
1009 1010 1011 1012 1013
		err = mlx5_ib_query_port(&dev->ib_dev, port, pprops);
		if (err) {
			mlx5_ib_warn(dev, "query_port %d failed %d\n", port, err);
			break;
		}
1014 1015
		gen->port[port - 1].pkey_table_len = dprops->max_pkeys;
		gen->port[port - 1].gid_table_len = pprops->gid_tbl_len;
1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
		mlx5_ib_dbg(dev, "pkey_table_len %d, gid_table_len %d\n",
			    dprops->max_pkeys, pprops->gid_tbl_len);
	}

out:
	kfree(pprops);
	kfree(dprops);

	return err;
}

static void destroy_umrc_res(struct mlx5_ib_dev *dev)
{
	int err;

	err = mlx5_mr_cache_cleanup(dev);
	if (err)
		mlx5_ib_warn(dev, "mr cache cleanup failed\n");

	mlx5_ib_destroy_qp(dev->umrc.qp);
	ib_destroy_cq(dev->umrc.cq);
	ib_dereg_mr(dev->umrc.mr);
	ib_dealloc_pd(dev->umrc.pd);
}

enum {
	MAX_UMR_WR = 128,
};

static int create_umr_res(struct mlx5_ib_dev *dev)
{
	struct ib_qp_init_attr *init_attr = NULL;
	struct ib_qp_attr *attr = NULL;
	struct ib_pd *pd;
	struct ib_cq *cq;
	struct ib_qp *qp;
	struct ib_mr *mr;
	int ret;

	attr = kzalloc(sizeof(*attr), GFP_KERNEL);
	init_attr = kzalloc(sizeof(*init_attr), GFP_KERNEL);
	if (!attr || !init_attr) {
		ret = -ENOMEM;
		goto error_0;
	}

	pd = ib_alloc_pd(&dev->ib_dev);
	if (IS_ERR(pd)) {
		mlx5_ib_dbg(dev, "Couldn't create PD for sync UMR QP\n");
		ret = PTR_ERR(pd);
		goto error_0;
	}

	mr = ib_get_dma_mr(pd,  IB_ACCESS_LOCAL_WRITE);
	if (IS_ERR(mr)) {
		mlx5_ib_dbg(dev, "Couldn't create DMA MR for sync UMR QP\n");
		ret = PTR_ERR(mr);
		goto error_1;
	}

	cq = ib_create_cq(&dev->ib_dev, mlx5_umr_cq_handler, NULL, NULL, 128,
			  0);
	if (IS_ERR(cq)) {
		mlx5_ib_dbg(dev, "Couldn't create CQ for sync UMR QP\n");
		ret = PTR_ERR(cq);
		goto error_2;
	}
	ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);

	init_attr->send_cq = cq;
	init_attr->recv_cq = cq;
	init_attr->sq_sig_type = IB_SIGNAL_ALL_WR;
	init_attr->cap.max_send_wr = MAX_UMR_WR;
	init_attr->cap.max_send_sge = 1;
	init_attr->qp_type = MLX5_IB_QPT_REG_UMR;
	init_attr->port_num = 1;
	qp = mlx5_ib_create_qp(pd, init_attr, NULL);
	if (IS_ERR(qp)) {
		mlx5_ib_dbg(dev, "Couldn't create sync UMR QP\n");
		ret = PTR_ERR(qp);
		goto error_3;
	}
	qp->device     = &dev->ib_dev;
	qp->real_qp    = qp;
	qp->uobject    = NULL;
	qp->qp_type    = MLX5_IB_QPT_REG_UMR;

	attr->qp_state = IB_QPS_INIT;
	attr->port_num = 1;
	ret = mlx5_ib_modify_qp(qp, attr, IB_QP_STATE | IB_QP_PKEY_INDEX |
				IB_QP_PORT, NULL);
	if (ret) {
		mlx5_ib_dbg(dev, "Couldn't modify UMR QP\n");
		goto error_4;
	}

	memset(attr, 0, sizeof(*attr));
	attr->qp_state = IB_QPS_RTR;
	attr->path_mtu = IB_MTU_256;

	ret = mlx5_ib_modify_qp(qp, attr, IB_QP_STATE, NULL);
	if (ret) {
		mlx5_ib_dbg(dev, "Couldn't modify umr QP to rtr\n");
		goto error_4;
	}

	memset(attr, 0, sizeof(*attr));
	attr->qp_state = IB_QPS_RTS;
	ret = mlx5_ib_modify_qp(qp, attr, IB_QP_STATE, NULL);
	if (ret) {
		mlx5_ib_dbg(dev, "Couldn't modify umr QP to rts\n");
		goto error_4;
	}

	dev->umrc.qp = qp;
	dev->umrc.cq = cq;
	dev->umrc.mr = mr;
	dev->umrc.pd = pd;

	sema_init(&dev->umrc.sem, MAX_UMR_WR);
	ret = mlx5_mr_cache_init(dev);
	if (ret) {
		mlx5_ib_warn(dev, "mr cache init failed %d\n", ret);
		goto error_4;
	}

	kfree(attr);
	kfree(init_attr);

	return 0;

error_4:
	mlx5_ib_destroy_qp(qp);

error_3:
	ib_destroy_cq(cq);

error_2:
	ib_dereg_mr(mr);

error_1:
	ib_dealloc_pd(pd);

error_0:
	kfree(attr);
	kfree(init_attr);
	return ret;
}

static int create_dev_resources(struct mlx5_ib_resources *devr)
{
	struct ib_srq_init_attr attr;
	struct mlx5_ib_dev *dev;
	int ret = 0;

	dev = container_of(devr, struct mlx5_ib_dev, devr);

	devr->p0 = mlx5_ib_alloc_pd(&dev->ib_dev, NULL, NULL);
	if (IS_ERR(devr->p0)) {
		ret = PTR_ERR(devr->p0);
		goto error0;
	}
	devr->p0->device  = &dev->ib_dev;
	devr->p0->uobject = NULL;
	atomic_set(&devr->p0->usecnt, 0);

	devr->c0 = mlx5_ib_create_cq(&dev->ib_dev, 1, 0, NULL, NULL);
	if (IS_ERR(devr->c0)) {
		ret = PTR_ERR(devr->c0);
		goto error1;
	}
	devr->c0->device        = &dev->ib_dev;
	devr->c0->uobject       = NULL;
	devr->c0->comp_handler  = NULL;
	devr->c0->event_handler = NULL;
	devr->c0->cq_context    = NULL;
	atomic_set(&devr->c0->usecnt, 0);

	devr->x0 = mlx5_ib_alloc_xrcd(&dev->ib_dev, NULL, NULL);
	if (IS_ERR(devr->x0)) {
		ret = PTR_ERR(devr->x0);
		goto error2;
	}
	devr->x0->device = &dev->ib_dev;
	devr->x0->inode = NULL;
	atomic_set(&devr->x0->usecnt, 0);
	mutex_init(&devr->x0->tgt_qp_mutex);
	INIT_LIST_HEAD(&devr->x0->tgt_qp_list);

	devr->x1 = mlx5_ib_alloc_xrcd(&dev->ib_dev, NULL, NULL);
	if (IS_ERR(devr->x1)) {
		ret = PTR_ERR(devr->x1);
		goto error3;
	}
	devr->x1->device = &dev->ib_dev;
	devr->x1->inode = NULL;
	atomic_set(&devr->x1->usecnt, 0);
	mutex_init(&devr->x1->tgt_qp_mutex);
	INIT_LIST_HEAD(&devr->x1->tgt_qp_list);

	memset(&attr, 0, sizeof(attr));
	attr.attr.max_sge = 1;
	attr.attr.max_wr = 1;
	attr.srq_type = IB_SRQT_XRC;
	attr.ext.xrc.cq = devr->c0;
	attr.ext.xrc.xrcd = devr->x0;

	devr->s0 = mlx5_ib_create_srq(devr->p0, &attr, NULL);
	if (IS_ERR(devr->s0)) {
		ret = PTR_ERR(devr->s0);
		goto error4;
	}
	devr->s0->device	= &dev->ib_dev;
	devr->s0->pd		= devr->p0;
	devr->s0->uobject       = NULL;
	devr->s0->event_handler = NULL;
	devr->s0->srq_context   = NULL;
	devr->s0->srq_type      = IB_SRQT_XRC;
	devr->s0->ext.xrc.xrcd	= devr->x0;
	devr->s0->ext.xrc.cq	= devr->c0;
	atomic_inc(&devr->s0->ext.xrc.xrcd->usecnt);
	atomic_inc(&devr->s0->ext.xrc.cq->usecnt);
	atomic_inc(&devr->p0->usecnt);
	atomic_set(&devr->s0->usecnt, 0);

	return 0;

error4:
	mlx5_ib_dealloc_xrcd(devr->x1);
error3:
	mlx5_ib_dealloc_xrcd(devr->x0);
error2:
	mlx5_ib_destroy_cq(devr->c0);
error1:
	mlx5_ib_dealloc_pd(devr->p0);
error0:
	return ret;
}

static void destroy_dev_resources(struct mlx5_ib_resources *devr)
{
	mlx5_ib_destroy_srq(devr->s0);
	mlx5_ib_dealloc_xrcd(devr->x0);
	mlx5_ib_dealloc_xrcd(devr->x1);
	mlx5_ib_destroy_cq(devr->c0);
	mlx5_ib_dealloc_pd(devr->p0);
}

1264
static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
1265 1266 1267 1268 1269 1270 1271 1272 1273
{
	struct mlx5_ib_dev *dev;
	int err;
	int i;

	printk_once(KERN_INFO "%s", mlx5_version);

	dev = (struct mlx5_ib_dev *)ib_alloc_device(sizeof(*dev));
	if (!dev)
1274
		return NULL;
1275

1276
	dev->mdev = mdev;
1277 1278 1279

	err = get_port_caps(dev);
	if (err)
1280
		goto err_dealloc;
1281 1282 1283 1284 1285

	get_ext_port_caps(dev);

	err = alloc_comp_eqs(dev);
	if (err)
1286
		goto err_dealloc;
1287 1288 1289 1290 1291 1292

	MLX5_INIT_DOORBELL_LOCK(&dev->uar_lock);

	strlcpy(dev->ib_dev.name, "mlx5_%d", IB_DEVICE_NAME_MAX);
	dev->ib_dev.owner		= THIS_MODULE;
	dev->ib_dev.node_type		= RDMA_NODE_IB_CA;
1293 1294
	dev->ib_dev.local_dma_lkey	= mdev->caps.gen.reserved_lkey;
	dev->num_ports		= mdev->caps.gen.num_ports;
1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
	dev->ib_dev.phys_port_cnt     = dev->num_ports;
	dev->ib_dev.num_comp_vectors	= dev->num_comp_vectors;
	dev->ib_dev.dma_device	= &mdev->pdev->dev;

	dev->ib_dev.uverbs_abi_ver	= MLX5_IB_UVERBS_ABI_VERSION;
	dev->ib_dev.uverbs_cmd_mask	=
		(1ull << IB_USER_VERBS_CMD_GET_CONTEXT)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_DEVICE)	|
		(1ull << IB_USER_VERBS_CMD_QUERY_PORT)		|
		(1ull << IB_USER_VERBS_CMD_ALLOC_PD)		|
		(1ull << IB_USER_VERBS_CMD_DEALLOC_PD)		|
		(1ull << IB_USER_VERBS_CMD_REG_MR)		|
		(1ull << IB_USER_VERBS_CMD_DEREG_MR)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL)	|
		(1ull << IB_USER_VERBS_CMD_CREATE_CQ)		|
		(1ull << IB_USER_VERBS_CMD_RESIZE_CQ)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_CQ)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_QP)		|
		(1ull << IB_USER_VERBS_CMD_MODIFY_QP)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_QP)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_QP)		|
		(1ull << IB_USER_VERBS_CMD_ATTACH_MCAST)	|
		(1ull << IB_USER_VERBS_CMD_DETACH_MCAST)	|
		(1ull << IB_USER_VERBS_CMD_CREATE_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_MODIFY_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_SRQ)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_XSRQ)		|
		(1ull << IB_USER_VERBS_CMD_OPEN_QP);

	dev->ib_dev.query_device	= mlx5_ib_query_device;
	dev->ib_dev.query_port		= mlx5_ib_query_port;
	dev->ib_dev.query_gid		= mlx5_ib_query_gid;
	dev->ib_dev.query_pkey		= mlx5_ib_query_pkey;
	dev->ib_dev.modify_device	= mlx5_ib_modify_device;
	dev->ib_dev.modify_port		= mlx5_ib_modify_port;
	dev->ib_dev.alloc_ucontext	= mlx5_ib_alloc_ucontext;
	dev->ib_dev.dealloc_ucontext	= mlx5_ib_dealloc_ucontext;
	dev->ib_dev.mmap		= mlx5_ib_mmap;
	dev->ib_dev.alloc_pd		= mlx5_ib_alloc_pd;
	dev->ib_dev.dealloc_pd		= mlx5_ib_dealloc_pd;
	dev->ib_dev.create_ah		= mlx5_ib_create_ah;
	dev->ib_dev.query_ah		= mlx5_ib_query_ah;
	dev->ib_dev.destroy_ah		= mlx5_ib_destroy_ah;
	dev->ib_dev.create_srq		= mlx5_ib_create_srq;
	dev->ib_dev.modify_srq		= mlx5_ib_modify_srq;
	dev->ib_dev.query_srq		= mlx5_ib_query_srq;
	dev->ib_dev.destroy_srq		= mlx5_ib_destroy_srq;
	dev->ib_dev.post_srq_recv	= mlx5_ib_post_srq_recv;
	dev->ib_dev.create_qp		= mlx5_ib_create_qp;
	dev->ib_dev.modify_qp		= mlx5_ib_modify_qp;
	dev->ib_dev.query_qp		= mlx5_ib_query_qp;
	dev->ib_dev.destroy_qp		= mlx5_ib_destroy_qp;
	dev->ib_dev.post_send		= mlx5_ib_post_send;
	dev->ib_dev.post_recv		= mlx5_ib_post_recv;
	dev->ib_dev.create_cq		= mlx5_ib_create_cq;
	dev->ib_dev.modify_cq		= mlx5_ib_modify_cq;
	dev->ib_dev.resize_cq		= mlx5_ib_resize_cq;
	dev->ib_dev.destroy_cq		= mlx5_ib_destroy_cq;
	dev->ib_dev.poll_cq		= mlx5_ib_poll_cq;
	dev->ib_dev.req_notify_cq	= mlx5_ib_arm_cq;
	dev->ib_dev.get_dma_mr		= mlx5_ib_get_dma_mr;
	dev->ib_dev.reg_user_mr		= mlx5_ib_reg_user_mr;
	dev->ib_dev.dereg_mr		= mlx5_ib_dereg_mr;
1359
	dev->ib_dev.destroy_mr		= mlx5_ib_destroy_mr;
1360 1361 1362
	dev->ib_dev.attach_mcast	= mlx5_ib_mcg_attach;
	dev->ib_dev.detach_mcast	= mlx5_ib_mcg_detach;
	dev->ib_dev.process_mad		= mlx5_ib_process_mad;
1363
	dev->ib_dev.create_mr		= mlx5_ib_create_mr;
1364 1365 1366
	dev->ib_dev.alloc_fast_reg_mr	= mlx5_ib_alloc_fast_reg_mr;
	dev->ib_dev.alloc_fast_reg_page_list = mlx5_ib_alloc_fast_reg_page_list;
	dev->ib_dev.free_fast_reg_page_list  = mlx5_ib_free_fast_reg_page_list;
1367
	dev->ib_dev.check_mr_status	= mlx5_ib_check_mr_status;
1368

1369
	if (mdev->caps.gen.flags & MLX5_DEV_CAP_FLAG_XRC) {
1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387
		dev->ib_dev.alloc_xrcd = mlx5_ib_alloc_xrcd;
		dev->ib_dev.dealloc_xrcd = mlx5_ib_dealloc_xrcd;
		dev->ib_dev.uverbs_cmd_mask |=
			(1ull << IB_USER_VERBS_CMD_OPEN_XRCD) |
			(1ull << IB_USER_VERBS_CMD_CLOSE_XRCD);
	}

	err = init_node_data(dev);
	if (err)
		goto err_eqs;

	mutex_init(&dev->cap_mask_mutex);
	spin_lock_init(&dev->mr_lock);

	err = create_dev_resources(&dev->devr);
	if (err)
		goto err_eqs;

1388 1389
	err = ib_register_device(&dev->ib_dev, NULL);
	if (err)
1390 1391 1392 1393 1394 1395 1396
		goto err_rsrc;

	err = create_umr_res(dev);
	if (err)
		goto err_dev;

	for (i = 0; i < ARRAY_SIZE(mlx5_class_attributes); i++) {
1397 1398 1399
		err = device_create_file(&dev->ib_dev.dev,
					 mlx5_class_attributes[i]);
		if (err)
1400 1401 1402 1403 1404
			goto err_umrc;
	}

	dev->ib_active = true;

1405
	return dev;
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418

err_umrc:
	destroy_umrc_res(dev);

err_dev:
	ib_unregister_device(&dev->ib_dev);

err_rsrc:
	destroy_dev_resources(&dev->devr);

err_eqs:
	free_comp_eqs(dev);

1419
err_dealloc:
1420 1421
	ib_dealloc_device((struct ib_device *)dev);

1422
	return NULL;
1423 1424
}

1425
static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
1426
{
1427
	struct mlx5_ib_dev *dev = context;
1428
	ib_unregister_device(&dev->ib_dev);
1429
	destroy_umrc_res(dev);
1430 1431 1432 1433 1434
	destroy_dev_resources(&dev->devr);
	free_comp_eqs(dev);
	ib_dealloc_device(&dev->ib_dev);
}

1435 1436 1437 1438
static struct mlx5_interface mlx5_ib_interface = {
	.add            = mlx5_ib_add,
	.remove         = mlx5_ib_remove,
	.event          = mlx5_ib_event,
1439 1440 1441 1442
};

static int __init mlx5_ib_init(void)
{
1443 1444 1445 1446
	if (deprecated_prof_sel != 2)
		pr_warn("prof_sel is deprecated for mlx5_ib, set it for mlx5_core\n");

	return mlx5_register_interface(&mlx5_ib_interface);
1447 1448 1449 1450
}

static void __exit mlx5_ib_cleanup(void)
{
1451
	mlx5_unregister_interface(&mlx5_ib_interface);
1452 1453 1454 1455
}

module_init(mlx5_ib_init);
module_exit(mlx5_ib_cleanup);