verbs.c 47.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * Copyright (c) 2004 Mellanox Technologies Ltd.  All rights reserved.
 * Copyright (c) 2004 Infinicon Corporation.  All rights reserved.
 * Copyright (c) 2004 Intel Corporation.  All rights reserved.
 * Copyright (c) 2004 Topspin Corporation.  All rights reserved.
 * Copyright (c) 2004 Voltaire Corporation.  All rights reserved.
7
 * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
8
 * Copyright (c) 2005, 2006 Cisco Systems.  All rights reserved.
L
Linus Torvalds 已提交
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
 *
 * 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/errno.h>
#include <linux/err.h>
41
#include <linux/export.h>
42
#include <linux/string.h>
43
#include <linux/slab.h>
44 45 46
#include <linux/in.h>
#include <linux/in6.h>
#include <net/addrconf.h>
L
Linus Torvalds 已提交
47

48 49
#include <rdma/ib_verbs.h>
#include <rdma/ib_cache.h>
50
#include <rdma/ib_addr.h>
51
#include <rdma/rw.h>
L
Linus Torvalds 已提交
52

53
#include "core_priv.h"
L
Linus Torvalds 已提交
54

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
static const char * const ib_events[] = {
	[IB_EVENT_CQ_ERR]		= "CQ error",
	[IB_EVENT_QP_FATAL]		= "QP fatal error",
	[IB_EVENT_QP_REQ_ERR]		= "QP request error",
	[IB_EVENT_QP_ACCESS_ERR]	= "QP access error",
	[IB_EVENT_COMM_EST]		= "communication established",
	[IB_EVENT_SQ_DRAINED]		= "send queue drained",
	[IB_EVENT_PATH_MIG]		= "path migration successful",
	[IB_EVENT_PATH_MIG_ERR]		= "path migration error",
	[IB_EVENT_DEVICE_FATAL]		= "device fatal error",
	[IB_EVENT_PORT_ACTIVE]		= "port active",
	[IB_EVENT_PORT_ERR]		= "port error",
	[IB_EVENT_LID_CHANGE]		= "LID change",
	[IB_EVENT_PKEY_CHANGE]		= "P_key change",
	[IB_EVENT_SM_CHANGE]		= "SM change",
	[IB_EVENT_SRQ_ERR]		= "SRQ error",
	[IB_EVENT_SRQ_LIMIT_REACHED]	= "SRQ limit reached",
	[IB_EVENT_QP_LAST_WQE_REACHED]	= "last WQE reached",
	[IB_EVENT_CLIENT_REREGISTER]	= "client reregister",
	[IB_EVENT_GID_CHANGE]		= "GID changed",
};

77
const char *__attribute_const__ ib_event_msg(enum ib_event_type event)
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
{
	size_t index = event;

	return (index < ARRAY_SIZE(ib_events) && ib_events[index]) ?
			ib_events[index] : "unrecognized event";
}
EXPORT_SYMBOL(ib_event_msg);

static const char * const wc_statuses[] = {
	[IB_WC_SUCCESS]			= "success",
	[IB_WC_LOC_LEN_ERR]		= "local length error",
	[IB_WC_LOC_QP_OP_ERR]		= "local QP operation error",
	[IB_WC_LOC_EEC_OP_ERR]		= "local EE context operation error",
	[IB_WC_LOC_PROT_ERR]		= "local protection error",
	[IB_WC_WR_FLUSH_ERR]		= "WR flushed",
	[IB_WC_MW_BIND_ERR]		= "memory management operation error",
	[IB_WC_BAD_RESP_ERR]		= "bad response error",
	[IB_WC_LOC_ACCESS_ERR]		= "local access error",
	[IB_WC_REM_INV_REQ_ERR]		= "invalid request error",
	[IB_WC_REM_ACCESS_ERR]		= "remote access error",
	[IB_WC_REM_OP_ERR]		= "remote operation error",
	[IB_WC_RETRY_EXC_ERR]		= "transport retry counter exceeded",
	[IB_WC_RNR_RETRY_EXC_ERR]	= "RNR retry counter exceeded",
	[IB_WC_LOC_RDD_VIOL_ERR]	= "local RDD violation error",
	[IB_WC_REM_INV_RD_REQ_ERR]	= "remote invalid RD request",
	[IB_WC_REM_ABORT_ERR]		= "operation aborted",
	[IB_WC_INV_EECN_ERR]		= "invalid EE context number",
	[IB_WC_INV_EEC_STATE_ERR]	= "invalid EE context state",
	[IB_WC_FATAL_ERR]		= "fatal error",
	[IB_WC_RESP_TIMEOUT_ERR]	= "response timeout error",
	[IB_WC_GENERAL_ERR]		= "general error",
};

111
const char *__attribute_const__ ib_wc_status_msg(enum ib_wc_status status)
112 113 114 115 116 117 118 119
{
	size_t index = status;

	return (index < ARRAY_SIZE(wc_statuses) && wc_statuses[index]) ?
			wc_statuses[index] : "unrecognized status";
}
EXPORT_SYMBOL(ib_wc_status_msg);

120
__attribute_const__ int ib_rate_to_mult(enum ib_rate rate)
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
{
	switch (rate) {
	case IB_RATE_2_5_GBPS: return  1;
	case IB_RATE_5_GBPS:   return  2;
	case IB_RATE_10_GBPS:  return  4;
	case IB_RATE_20_GBPS:  return  8;
	case IB_RATE_30_GBPS:  return 12;
	case IB_RATE_40_GBPS:  return 16;
	case IB_RATE_60_GBPS:  return 24;
	case IB_RATE_80_GBPS:  return 32;
	case IB_RATE_120_GBPS: return 48;
	default:	       return -1;
	}
}
EXPORT_SYMBOL(ib_rate_to_mult);

137
__attribute_const__ enum ib_rate mult_to_ib_rate(int mult)
138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
{
	switch (mult) {
	case 1:  return IB_RATE_2_5_GBPS;
	case 2:  return IB_RATE_5_GBPS;
	case 4:  return IB_RATE_10_GBPS;
	case 8:  return IB_RATE_20_GBPS;
	case 12: return IB_RATE_30_GBPS;
	case 16: return IB_RATE_40_GBPS;
	case 24: return IB_RATE_60_GBPS;
	case 32: return IB_RATE_80_GBPS;
	case 48: return IB_RATE_120_GBPS;
	default: return IB_RATE_PORT_CURRENT;
	}
}
EXPORT_SYMBOL(mult_to_ib_rate);

154
__attribute_const__ int ib_rate_to_mbps(enum ib_rate rate)
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178
{
	switch (rate) {
	case IB_RATE_2_5_GBPS: return 2500;
	case IB_RATE_5_GBPS:   return 5000;
	case IB_RATE_10_GBPS:  return 10000;
	case IB_RATE_20_GBPS:  return 20000;
	case IB_RATE_30_GBPS:  return 30000;
	case IB_RATE_40_GBPS:  return 40000;
	case IB_RATE_60_GBPS:  return 60000;
	case IB_RATE_80_GBPS:  return 80000;
	case IB_RATE_120_GBPS: return 120000;
	case IB_RATE_14_GBPS:  return 14062;
	case IB_RATE_56_GBPS:  return 56250;
	case IB_RATE_112_GBPS: return 112500;
	case IB_RATE_168_GBPS: return 168750;
	case IB_RATE_25_GBPS:  return 25781;
	case IB_RATE_100_GBPS: return 103125;
	case IB_RATE_200_GBPS: return 206250;
	case IB_RATE_300_GBPS: return 309375;
	default:	       return -1;
	}
}
EXPORT_SYMBOL(ib_rate_to_mbps);

179
__attribute_const__ enum rdma_transport_type
T
Tom Tucker 已提交
180 181 182 183 184 185 186 187 188
rdma_node_get_transport(enum rdma_node_type node_type)
{
	switch (node_type) {
	case RDMA_NODE_IB_CA:
	case RDMA_NODE_IB_SWITCH:
	case RDMA_NODE_IB_ROUTER:
		return RDMA_TRANSPORT_IB;
	case RDMA_NODE_RNIC:
		return RDMA_TRANSPORT_IWARP;
189
	case RDMA_NODE_USNIC:
190 191
		return RDMA_TRANSPORT_USNIC;
	case RDMA_NODE_USNIC_UDP:
192
		return RDMA_TRANSPORT_USNIC_UDP;
T
Tom Tucker 已提交
193 194 195 196 197 198 199
	default:
		BUG();
		return 0;
	}
}
EXPORT_SYMBOL(rdma_node_get_transport);

200 201 202 203 204 205 206 207 208
enum rdma_link_layer rdma_port_get_link_layer(struct ib_device *device, u8 port_num)
{
	if (device->get_link_layer)
		return device->get_link_layer(device, port_num);

	switch (rdma_node_get_transport(device->node_type)) {
	case RDMA_TRANSPORT_IB:
		return IB_LINK_LAYER_INFINIBAND;
	case RDMA_TRANSPORT_IWARP:
209
	case RDMA_TRANSPORT_USNIC:
210
	case RDMA_TRANSPORT_USNIC_UDP:
211 212 213 214 215 216 217
		return IB_LINK_LAYER_ETHERNET;
	default:
		return IB_LINK_LAYER_UNSPECIFIED;
	}
}
EXPORT_SYMBOL(rdma_port_get_link_layer);

L
Linus Torvalds 已提交
218 219
/* Protection domains */

220 221 222 223 224 225 226 227 228 229
/**
 * ib_alloc_pd - Allocates an unused protection domain.
 * @device: The device on which to allocate the protection domain.
 *
 * A protection domain object provides an association between QPs, shared
 * receive queues, address handles, memory regions, and memory windows.
 *
 * Every PD has a local_dma_lkey which can be used as the lkey value for local
 * memory operations.
 */
L
Linus Torvalds 已提交
230 231 232 233
struct ib_pd *ib_alloc_pd(struct ib_device *device)
{
	struct ib_pd *pd;

234
	pd = device->alloc_pd(device, NULL, NULL);
235 236
	if (IS_ERR(pd))
		return pd;
L
Linus Torvalds 已提交
237

238 239 240 241
	pd->device = device;
	pd->uobject = NULL;
	pd->local_mr = NULL;
	atomic_set(&pd->usecnt, 0);
L
Linus Torvalds 已提交
242

243
	if (device->attrs.device_cap_flags & IB_DEVICE_LOCAL_DMA_LKEY)
244 245 246 247 248 249 250 251 252
		pd->local_dma_lkey = device->local_dma_lkey;
	else {
		struct ib_mr *mr;

		mr = ib_get_dma_mr(pd, IB_ACCESS_LOCAL_WRITE);
		if (IS_ERR(mr)) {
			ib_dealloc_pd(pd);
			return (struct ib_pd *)mr;
		}
L
Linus Torvalds 已提交
253

254 255
		pd->local_mr = mr;
		pd->local_dma_lkey = pd->local_mr->lkey;
L
Linus Torvalds 已提交
256 257 258 259 260
	}
	return pd;
}
EXPORT_SYMBOL(ib_alloc_pd);

261 262 263 264 265 266 267 268 269
/**
 * ib_dealloc_pd - Deallocates a protection domain.
 * @pd: The protection domain to deallocate.
 *
 * It is an error to call this function while any resources in the pd still
 * exist.  The caller is responsible to synchronously destroy them and
 * guarantee no new allocations will happen.
 */
void ib_dealloc_pd(struct ib_pd *pd)
L
Linus Torvalds 已提交
270
{
271 272
	int ret;

273
	if (pd->local_mr) {
274 275
		ret = ib_dereg_mr(pd->local_mr);
		WARN_ON(ret);
276 277
		pd->local_mr = NULL;
	}
L
Linus Torvalds 已提交
278

279 280 281
	/* uverbs manipulates usecnt with proper locking, while the kabi
	   requires the caller to guarantee we can't race here. */
	WARN_ON(atomic_read(&pd->usecnt));
L
Linus Torvalds 已提交
282

283 284 285 286
	/* Making delalloc_pd a void return is a WIP, no driver should return
	   an error here. */
	ret = pd->device->dealloc_pd(pd);
	WARN_ONCE(ret, "Infiniband HW driver failed dealloc_pd");
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297 298
}
EXPORT_SYMBOL(ib_dealloc_pd);

/* Address handles */

struct ib_ah *ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr)
{
	struct ib_ah *ah;

	ah = pd->device->create_ah(pd, ah_attr);

	if (!IS_ERR(ah)) {
299 300 301
		ah->device  = pd->device;
		ah->pd      = pd;
		ah->uobject = NULL;
L
Linus Torvalds 已提交
302 303 304 305 306 307 308
		atomic_inc(&pd->usecnt);
	}

	return ah;
}
EXPORT_SYMBOL(ib_create_ah);

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
static int ib_get_header_version(const union rdma_network_hdr *hdr)
{
	const struct iphdr *ip4h = (struct iphdr *)&hdr->roce4grh;
	struct iphdr ip4h_checked;
	const struct ipv6hdr *ip6h = (struct ipv6hdr *)&hdr->ibgrh;

	/* If it's IPv6, the version must be 6, otherwise, the first
	 * 20 bytes (before the IPv4 header) are garbled.
	 */
	if (ip6h->version != 6)
		return (ip4h->version == 4) ? 4 : 0;
	/* version may be 6 or 4 because the first 20 bytes could be garbled */

	/* RoCE v2 requires no options, thus header length
	 * must be 5 words
	 */
	if (ip4h->ihl != 5)
		return 6;

	/* Verify checksum.
	 * We can't write on scattered buffers so we need to copy to
	 * temp buffer.
	 */
	memcpy(&ip4h_checked, ip4h, sizeof(ip4h_checked));
	ip4h_checked.check = 0;
	ip4h_checked.check = ip_fast_csum((u8 *)&ip4h_checked, 5);
	/* if IPv4 header checksum is OK, believe it */
	if (ip4h->check == ip4h_checked.check)
		return 4;
	return 6;
}

static enum rdma_network_type ib_get_net_type_by_grh(struct ib_device *device,
						     u8 port_num,
						     const struct ib_grh *grh)
{
	int grh_version;

	if (rdma_protocol_ib(device, port_num))
		return RDMA_NETWORK_IB;

	grh_version = ib_get_header_version((union rdma_network_hdr *)grh);

	if (grh_version == 4)
		return RDMA_NETWORK_IPV4;

	if (grh->next_hdr == IPPROTO_UDP)
		return RDMA_NETWORK_IPV6;

	return RDMA_NETWORK_ROCE_V1;
}

361 362
struct find_gid_index_context {
	u16 vlan_id;
363
	enum ib_gid_type gid_type;
364 365 366 367 368 369 370 371 372
};

static bool find_gid_index(const union ib_gid *gid,
			   const struct ib_gid_attr *gid_attr,
			   void *context)
{
	struct find_gid_index_context *ctx =
		(struct find_gid_index_context *)context;

373 374 375
	if (ctx->gid_type != gid_attr->gid_type)
		return false;

376 377 378 379 380 381 382 383 384 385
	if ((!!(ctx->vlan_id != 0xffff) == !is_vlan_dev(gid_attr->ndev)) ||
	    (is_vlan_dev(gid_attr->ndev) &&
	     vlan_dev_vlan_id(gid_attr->ndev) != ctx->vlan_id))
		return false;

	return true;
}

static int get_sgid_index_from_eth(struct ib_device *device, u8 port_num,
				   u16 vlan_id, const union ib_gid *sgid,
386
				   enum ib_gid_type gid_type,
387 388
				   u16 *gid_index)
{
389 390
	struct find_gid_index_context context = {.vlan_id = vlan_id,
						 .gid_type = gid_type};
391 392 393 394 395

	return ib_find_gid_by_filter(device, sgid, port_num, find_gid_index,
				     &context, gid_index);
}

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
static int get_gids_from_rdma_hdr(union rdma_network_hdr *hdr,
				  enum rdma_network_type net_type,
				  union ib_gid *sgid, union ib_gid *dgid)
{
	struct sockaddr_in  src_in;
	struct sockaddr_in  dst_in;
	__be32 src_saddr, dst_saddr;

	if (!sgid || !dgid)
		return -EINVAL;

	if (net_type == RDMA_NETWORK_IPV4) {
		memcpy(&src_in.sin_addr.s_addr,
		       &hdr->roce4grh.saddr, 4);
		memcpy(&dst_in.sin_addr.s_addr,
		       &hdr->roce4grh.daddr, 4);
		src_saddr = src_in.sin_addr.s_addr;
		dst_saddr = dst_in.sin_addr.s_addr;
		ipv6_addr_set_v4mapped(src_saddr,
				       (struct in6_addr *)sgid);
		ipv6_addr_set_v4mapped(dst_saddr,
				       (struct in6_addr *)dgid);
		return 0;
	} else if (net_type == RDMA_NETWORK_IPV6 ||
		   net_type == RDMA_NETWORK_IB) {
		*dgid = hdr->ibgrh.dgid;
		*sgid = hdr->ibgrh.sgid;
		return 0;
	} else {
		return -EINVAL;
	}
}

429 430 431
int ib_init_ah_from_wc(struct ib_device *device, u8 port_num,
		       const struct ib_wc *wc, const struct ib_grh *grh,
		       struct ib_ah_attr *ah_attr)
432 433 434 435
{
	u32 flow_class;
	u16 gid_index;
	int ret;
436 437
	enum rdma_network_type net_type = RDMA_NETWORK_IB;
	enum ib_gid_type gid_type = IB_GID_TYPE_IB;
438
	int hoplimit = 0xff;
439 440
	union ib_gid dgid;
	union ib_gid sgid;
441

S
Sean Hefty 已提交
442
	memset(ah_attr, 0, sizeof *ah_attr);
443
	if (rdma_cap_eth_ah(device, port_num)) {
444 445 446 447 448 449 450 451 452 453 454 455
		if (wc->wc_flags & IB_WC_WITH_NETWORK_HDR_TYPE)
			net_type = wc->network_hdr_type;
		else
			net_type = ib_get_net_type_by_grh(device, port_num, grh);
		gid_type = ib_network_to_gid_type(net_type);
	}
	ret = get_gids_from_rdma_hdr((union rdma_network_hdr *)grh, net_type,
				     &sgid, &dgid);
	if (ret)
		return ret;

	if (rdma_protocol_roce(device, port_num)) {
456
		int if_index = 0;
457 458
		u16 vlan_id = wc->wc_flags & IB_WC_WITH_VLAN ?
				wc->vlan_id : 0xffff;
459 460
		struct net_device *idev;
		struct net_device *resolved_dev;
461

462 463 464
		if (!(wc->wc_flags & IB_WC_GRH))
			return -EPROTOTYPE;

465 466 467 468 469 470 471
		if (!device->get_netdev)
			return -EOPNOTSUPP;

		idev = device->get_netdev(device, port_num);
		if (!idev)
			return -ENODEV;

472 473 474 475
		ret = rdma_addr_find_l2_eth_by_grh(&dgid, &sgid,
						   ah_attr->dmac,
						   wc->wc_flags & IB_WC_WITH_VLAN ?
						   NULL : &vlan_id,
476
						   &if_index, &hoplimit);
477 478 479
		if (ret) {
			dev_put(idev);
			return ret;
480
		}
481

482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
		resolved_dev = dev_get_by_index(&init_net, if_index);
		if (resolved_dev->flags & IFF_LOOPBACK) {
			dev_put(resolved_dev);
			resolved_dev = idev;
			dev_hold(resolved_dev);
		}
		rcu_read_lock();
		if (resolved_dev != idev && !rdma_is_upper_dev_rcu(idev,
								   resolved_dev))
			ret = -EHOSTUNREACH;
		rcu_read_unlock();
		dev_put(idev);
		dev_put(resolved_dev);
		if (ret)
			return ret;

498
		ret = get_sgid_index_from_eth(device, port_num, vlan_id,
499
					      &dgid, gid_type, &gid_index);
500 501
		if (ret)
			return ret;
502 503
	}

S
Sean Hefty 已提交
504 505 506 507
	ah_attr->dlid = wc->slid;
	ah_attr->sl = wc->sl;
	ah_attr->src_path_bits = wc->dlid_path_bits;
	ah_attr->port_num = port_num;
508 509

	if (wc->wc_flags & IB_WC_GRH) {
S
Sean Hefty 已提交
510
		ah_attr->ah_flags = IB_AH_GRH;
511
		ah_attr->grh.dgid = sgid;
512

513
		if (!rdma_cap_eth_ah(device, port_num)) {
514
			ret = ib_find_cached_gid_by_port(device, &dgid,
515
							 IB_GID_TYPE_IB,
516 517 518 519 520
							 port_num, NULL,
							 &gid_index);
			if (ret)
				return ret;
		}
521

S
Sean Hefty 已提交
522
		ah_attr->grh.sgid_index = (u8) gid_index;
523
		flow_class = be32_to_cpu(grh->version_tclass_flow);
S
Sean Hefty 已提交
524
		ah_attr->grh.flow_label = flow_class & 0xFFFFF;
525
		ah_attr->grh.hop_limit = hoplimit;
S
Sean Hefty 已提交
526
		ah_attr->grh.traffic_class = (flow_class >> 20) & 0xFF;
527
	}
S
Sean Hefty 已提交
528 529 530 531
	return 0;
}
EXPORT_SYMBOL(ib_init_ah_from_wc);

532 533
struct ib_ah *ib_create_ah_from_wc(struct ib_pd *pd, const struct ib_wc *wc,
				   const struct ib_grh *grh, u8 port_num)
S
Sean Hefty 已提交
534 535 536 537 538 539 540
{
	struct ib_ah_attr ah_attr;
	int ret;

	ret = ib_init_ah_from_wc(pd->device, port_num, wc, grh, &ah_attr);
	if (ret)
		return ERR_PTR(ret);
541 542 543 544 545

	return ib_create_ah(pd, &ah_attr);
}
EXPORT_SYMBOL(ib_create_ah_from_wc);

L
Linus Torvalds 已提交
546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575
int ib_modify_ah(struct ib_ah *ah, struct ib_ah_attr *ah_attr)
{
	return ah->device->modify_ah ?
		ah->device->modify_ah(ah, ah_attr) :
		-ENOSYS;
}
EXPORT_SYMBOL(ib_modify_ah);

int ib_query_ah(struct ib_ah *ah, struct ib_ah_attr *ah_attr)
{
	return ah->device->query_ah ?
		ah->device->query_ah(ah, ah_attr) :
		-ENOSYS;
}
EXPORT_SYMBOL(ib_query_ah);

int ib_destroy_ah(struct ib_ah *ah)
{
	struct ib_pd *pd;
	int ret;

	pd = ah->pd;
	ret = ah->device->destroy_ah(ah);
	if (!ret)
		atomic_dec(&pd->usecnt);

	return ret;
}
EXPORT_SYMBOL(ib_destroy_ah);

576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593
/* Shared receive queues */

struct ib_srq *ib_create_srq(struct ib_pd *pd,
			     struct ib_srq_init_attr *srq_init_attr)
{
	struct ib_srq *srq;

	if (!pd->device->create_srq)
		return ERR_PTR(-ENOSYS);

	srq = pd->device->create_srq(pd, srq_init_attr, NULL);

	if (!IS_ERR(srq)) {
		srq->device    	   = pd->device;
		srq->pd        	   = pd;
		srq->uobject       = NULL;
		srq->event_handler = srq_init_attr->event_handler;
		srq->srq_context   = srq_init_attr->srq_context;
S
Sean Hefty 已提交
594
		srq->srq_type      = srq_init_attr->srq_type;
S
Sean Hefty 已提交
595 596 597 598 599 600
		if (srq->srq_type == IB_SRQT_XRC) {
			srq->ext.xrc.xrcd = srq_init_attr->ext.xrc.xrcd;
			srq->ext.xrc.cq   = srq_init_attr->ext.xrc.cq;
			atomic_inc(&srq->ext.xrc.xrcd->usecnt);
			atomic_inc(&srq->ext.xrc.cq->usecnt);
		}
601 602 603 604 605 606 607 608 609 610 611 612
		atomic_inc(&pd->usecnt);
		atomic_set(&srq->usecnt, 0);
	}

	return srq;
}
EXPORT_SYMBOL(ib_create_srq);

int ib_modify_srq(struct ib_srq *srq,
		  struct ib_srq_attr *srq_attr,
		  enum ib_srq_attr_mask srq_attr_mask)
{
613 614 615
	return srq->device->modify_srq ?
		srq->device->modify_srq(srq, srq_attr, srq_attr_mask, NULL) :
		-ENOSYS;
616 617 618 619 620 621 622 623 624 625 626 627 628 629
}
EXPORT_SYMBOL(ib_modify_srq);

int ib_query_srq(struct ib_srq *srq,
		 struct ib_srq_attr *srq_attr)
{
	return srq->device->query_srq ?
		srq->device->query_srq(srq, srq_attr) : -ENOSYS;
}
EXPORT_SYMBOL(ib_query_srq);

int ib_destroy_srq(struct ib_srq *srq)
{
	struct ib_pd *pd;
S
Sean Hefty 已提交
630 631 632
	enum ib_srq_type srq_type;
	struct ib_xrcd *uninitialized_var(xrcd);
	struct ib_cq *uninitialized_var(cq);
633 634 635 636 637 638
	int ret;

	if (atomic_read(&srq->usecnt))
		return -EBUSY;

	pd = srq->pd;
S
Sean Hefty 已提交
639 640 641 642 643
	srq_type = srq->srq_type;
	if (srq_type == IB_SRQT_XRC) {
		xrcd = srq->ext.xrc.xrcd;
		cq = srq->ext.xrc.cq;
	}
644 645

	ret = srq->device->destroy_srq(srq);
S
Sean Hefty 已提交
646
	if (!ret) {
647
		atomic_dec(&pd->usecnt);
S
Sean Hefty 已提交
648 649 650 651 652
		if (srq_type == IB_SRQT_XRC) {
			atomic_dec(&xrcd->usecnt);
			atomic_dec(&cq->usecnt);
		}
	}
653 654 655 656 657

	return ret;
}
EXPORT_SYMBOL(ib_destroy_srq);

L
Linus Torvalds 已提交
658 659
/* Queue pairs */

660 661 662
static void __ib_shared_qp_event_handler(struct ib_event *event, void *context)
{
	struct ib_qp *qp = context;
663
	unsigned long flags;
664

665
	spin_lock_irqsave(&qp->device->event_handler_lock, flags);
666
	list_for_each_entry(event->element.qp, &qp->open_list, open_list)
667 668
		if (event->element.qp->event_handler)
			event->element.qp->event_handler(event, event->element.qp->qp_context);
669
	spin_unlock_irqrestore(&qp->device->event_handler_lock, flags);
670 671
}

672 673 674 675 676 677 678
static void __ib_insert_xrcd_qp(struct ib_xrcd *xrcd, struct ib_qp *qp)
{
	mutex_lock(&xrcd->tgt_qp_mutex);
	list_add(&qp->xrcd_list, &xrcd->tgt_qp_list);
	mutex_unlock(&xrcd->tgt_qp_mutex);
}

679 680 681
static struct ib_qp *__ib_open_qp(struct ib_qp *real_qp,
				  void (*event_handler)(struct ib_event *, void *),
				  void *qp_context)
682
{
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
	struct ib_qp *qp;
	unsigned long flags;

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

	qp->real_qp = real_qp;
	atomic_inc(&real_qp->usecnt);
	qp->device = real_qp->device;
	qp->event_handler = event_handler;
	qp->qp_context = qp_context;
	qp->qp_num = real_qp->qp_num;
	qp->qp_type = real_qp->qp_type;

	spin_lock_irqsave(&real_qp->device->event_handler_lock, flags);
	list_add(&qp->open_list, &real_qp->open_list);
	spin_unlock_irqrestore(&real_qp->device->event_handler_lock, flags);

	return qp;
}

struct ib_qp *ib_open_qp(struct ib_xrcd *xrcd,
			 struct ib_qp_open_attr *qp_open_attr)
{
	struct ib_qp *qp, *real_qp;

	if (qp_open_attr->qp_type != IB_QPT_XRC_TGT)
		return ERR_PTR(-EINVAL);

	qp = ERR_PTR(-EINVAL);
714
	mutex_lock(&xrcd->tgt_qp_mutex);
715 716 717 718 719 720 721
	list_for_each_entry(real_qp, &xrcd->tgt_qp_list, xrcd_list) {
		if (real_qp->qp_num == qp_open_attr->qp_num) {
			qp = __ib_open_qp(real_qp, qp_open_attr->event_handler,
					  qp_open_attr->qp_context);
			break;
		}
	}
722
	mutex_unlock(&xrcd->tgt_qp_mutex);
723
	return qp;
724
}
725
EXPORT_SYMBOL(ib_open_qp);
726

727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
static struct ib_qp *ib_create_xrc_qp(struct ib_qp *qp,
		struct ib_qp_init_attr *qp_init_attr)
{
	struct ib_qp *real_qp = qp;

	qp->event_handler = __ib_shared_qp_event_handler;
	qp->qp_context = qp;
	qp->pd = NULL;
	qp->send_cq = qp->recv_cq = NULL;
	qp->srq = NULL;
	qp->xrcd = qp_init_attr->xrcd;
	atomic_inc(&qp_init_attr->xrcd->usecnt);
	INIT_LIST_HEAD(&qp->open_list);

	qp = __ib_open_qp(real_qp, qp_init_attr->event_handler,
			  qp_init_attr->qp_context);
	if (!IS_ERR(qp))
		__ib_insert_xrcd_qp(qp_init_attr->xrcd, real_qp);
	else
		real_qp->device->destroy_qp(real_qp);
	return qp;
}

L
Linus Torvalds 已提交
750 751 752
struct ib_qp *ib_create_qp(struct ib_pd *pd,
			   struct ib_qp_init_attr *qp_init_attr)
{
753 754
	struct ib_device *device = pd ? pd->device : qp_init_attr->xrcd->device;
	struct ib_qp *qp;
755 756 757 758 759 760 761 762 763 764
	int ret;

	/*
	 * If the callers is using the RDMA API calculate the resources
	 * needed for the RDMA READ/WRITE operations.
	 *
	 * Note that these callers need to pass in a port number.
	 */
	if (qp_init_attr->cap.max_rdma_ctxs)
		rdma_rw_init_qp(device, qp_init_attr);
L
Linus Torvalds 已提交
765

S
Sean Hefty 已提交
766
	qp = device->create_qp(pd, qp_init_attr, NULL);
767 768 769 770 771 772 773 774 775
	if (IS_ERR(qp))
		return qp;

	qp->device     = device;
	qp->real_qp    = qp;
	qp->uobject    = NULL;
	qp->qp_type    = qp_init_attr->qp_type;

	atomic_set(&qp->usecnt, 0);
776 777
	qp->mrs_used = 0;
	spin_lock_init(&qp->mr_lock);
778
	INIT_LIST_HEAD(&qp->rdma_mrs);
779
	INIT_LIST_HEAD(&qp->sig_mrs);
780

781 782 783 784 785 786 787 788 789 790 791 792 793 794
	if (qp_init_attr->qp_type == IB_QPT_XRC_TGT)
		return ib_create_xrc_qp(qp, qp_init_attr);

	qp->event_handler = qp_init_attr->event_handler;
	qp->qp_context = qp_init_attr->qp_context;
	if (qp_init_attr->qp_type == IB_QPT_XRC_INI) {
		qp->recv_cq = NULL;
		qp->srq = NULL;
	} else {
		qp->recv_cq = qp_init_attr->recv_cq;
		atomic_inc(&qp_init_attr->recv_cq->usecnt);
		qp->srq = qp_init_attr->srq;
		if (qp->srq)
			atomic_inc(&qp_init_attr->srq->usecnt);
L
Linus Torvalds 已提交
795 796
	}

797 798 799 800 801 802
	qp->pd	    = pd;
	qp->send_cq = qp_init_attr->send_cq;
	qp->xrcd    = NULL;

	atomic_inc(&pd->usecnt);
	atomic_inc(&qp_init_attr->send_cq->usecnt);
803 804 805 806 807 808 809 810 811 812

	if (qp_init_attr->cap.max_rdma_ctxs) {
		ret = rdma_rw_init_mrs(qp, qp_init_attr);
		if (ret) {
			pr_err("failed to init MR pool ret= %d\n", ret);
			ib_destroy_qp(qp);
			qp = ERR_PTR(ret);
		}
	}

L
Linus Torvalds 已提交
813 814 815 816
	return qp;
}
EXPORT_SYMBOL(ib_create_qp);

817 818
static const struct {
	int			valid;
S
Sean Hefty 已提交
819 820
	enum ib_qp_attr_mask	req_param[IB_QPT_MAX];
	enum ib_qp_attr_mask	opt_param[IB_QPT_MAX];
821 822 823 824 825 826 827 828 829
} qp_state_table[IB_QPS_ERR + 1][IB_QPS_ERR + 1] = {
	[IB_QPS_RESET] = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_INIT]  = {
			.valid = 1,
			.req_param = {
				[IB_QPT_UD]  = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_QKEY),
O
Or Gerlitz 已提交
830
				[IB_QPT_RAW_PACKET] = IB_QP_PORT,
831 832 833 834 835 836
				[IB_QPT_UC]  = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
				[IB_QPT_RC]  = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
S
Sean Hefty 已提交
837 838 839 840 841 842
				[IB_QPT_XRC_INI] = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
				[IB_QPT_XRC_TGT] = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864
				[IB_QPT_SMI] = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
				[IB_QPT_GSI] = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
			}
		},
	},
	[IB_QPS_INIT]  = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_ERR] =   { .valid = 1 },
		[IB_QPS_INIT]  = {
			.valid = 1,
			.opt_param = {
				[IB_QPT_UD]  = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_QKEY),
				[IB_QPT_UC]  = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
				[IB_QPT_RC]  = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
S
Sean Hefty 已提交
865 866 867 868 869 870
				[IB_QPT_XRC_INI] = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
				[IB_QPT_XRC_TGT] = (IB_QP_PKEY_INDEX		|
						IB_QP_PORT			|
						IB_QP_ACCESS_FLAGS),
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889
				[IB_QPT_SMI] = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
				[IB_QPT_GSI] = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
			}
		},
		[IB_QPS_RTR]   = {
			.valid = 1,
			.req_param = {
				[IB_QPT_UC]  = (IB_QP_AV			|
						IB_QP_PATH_MTU			|
						IB_QP_DEST_QPN			|
						IB_QP_RQ_PSN),
				[IB_QPT_RC]  = (IB_QP_AV			|
						IB_QP_PATH_MTU			|
						IB_QP_DEST_QPN			|
						IB_QP_RQ_PSN			|
						IB_QP_MAX_DEST_RD_ATOMIC	|
						IB_QP_MIN_RNR_TIMER),
S
Sean Hefty 已提交
890 891 892 893 894 895 896 897 898 899
				[IB_QPT_XRC_INI] = (IB_QP_AV			|
						IB_QP_PATH_MTU			|
						IB_QP_DEST_QPN			|
						IB_QP_RQ_PSN),
				[IB_QPT_XRC_TGT] = (IB_QP_AV			|
						IB_QP_PATH_MTU			|
						IB_QP_DEST_QPN			|
						IB_QP_RQ_PSN			|
						IB_QP_MAX_DEST_RD_ATOMIC	|
						IB_QP_MIN_RNR_TIMER),
900 901 902 903 904 905 906 907 908 909
			},
			.opt_param = {
				 [IB_QPT_UD]  = (IB_QP_PKEY_INDEX		|
						 IB_QP_QKEY),
				 [IB_QPT_UC]  = (IB_QP_ALT_PATH			|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_PKEY_INDEX),
				 [IB_QPT_RC]  = (IB_QP_ALT_PATH			|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_PKEY_INDEX),
S
Sean Hefty 已提交
910 911 912 913 914 915
				 [IB_QPT_XRC_INI] = (IB_QP_ALT_PATH		|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_PKEY_INDEX),
				 [IB_QPT_XRC_TGT] = (IB_QP_ALT_PATH		|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_PKEY_INDEX),
916 917 918 919
				 [IB_QPT_SMI] = (IB_QP_PKEY_INDEX		|
						 IB_QP_QKEY),
				 [IB_QPT_GSI] = (IB_QP_PKEY_INDEX		|
						 IB_QP_QKEY),
920
			 },
921
		},
922 923 924 925 926 927 928 929 930 931 932 933 934 935
	},
	[IB_QPS_RTR]   = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_ERR] =   { .valid = 1 },
		[IB_QPS_RTS]   = {
			.valid = 1,
			.req_param = {
				[IB_QPT_UD]  = IB_QP_SQ_PSN,
				[IB_QPT_UC]  = IB_QP_SQ_PSN,
				[IB_QPT_RC]  = (IB_QP_TIMEOUT			|
						IB_QP_RETRY_CNT			|
						IB_QP_RNR_RETRY			|
						IB_QP_SQ_PSN			|
						IB_QP_MAX_QP_RD_ATOMIC),
S
Sean Hefty 已提交
936 937 938 939 940 941 942
				[IB_QPT_XRC_INI] = (IB_QP_TIMEOUT		|
						IB_QP_RETRY_CNT			|
						IB_QP_RNR_RETRY			|
						IB_QP_SQ_PSN			|
						IB_QP_MAX_QP_RD_ATOMIC),
				[IB_QPT_XRC_TGT] = (IB_QP_TIMEOUT		|
						IB_QP_SQ_PSN),
943 944 945 946 947 948 949 950 951 952 953 954 955 956 957
				[IB_QPT_SMI] = IB_QP_SQ_PSN,
				[IB_QPT_GSI] = IB_QP_SQ_PSN,
			},
			.opt_param = {
				 [IB_QPT_UD]  = (IB_QP_CUR_STATE		|
						 IB_QP_QKEY),
				 [IB_QPT_UC]  = (IB_QP_CUR_STATE		|
						 IB_QP_ALT_PATH			|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_PATH_MIG_STATE),
				 [IB_QPT_RC]  = (IB_QP_CUR_STATE		|
						 IB_QP_ALT_PATH			|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_MIN_RNR_TIMER		|
						 IB_QP_PATH_MIG_STATE),
S
Sean Hefty 已提交
958 959 960 961 962 963 964 965 966
				 [IB_QPT_XRC_INI] = (IB_QP_CUR_STATE		|
						 IB_QP_ALT_PATH			|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_PATH_MIG_STATE),
				 [IB_QPT_XRC_TGT] = (IB_QP_CUR_STATE		|
						 IB_QP_ALT_PATH			|
						 IB_QP_ACCESS_FLAGS		|
						 IB_QP_MIN_RNR_TIMER		|
						 IB_QP_PATH_MIG_STATE),
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981
				 [IB_QPT_SMI] = (IB_QP_CUR_STATE		|
						 IB_QP_QKEY),
				 [IB_QPT_GSI] = (IB_QP_CUR_STATE		|
						 IB_QP_QKEY),
			 }
		}
	},
	[IB_QPS_RTS]   = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_ERR] =   { .valid = 1 },
		[IB_QPS_RTS]   = {
			.valid = 1,
			.opt_param = {
				[IB_QPT_UD]  = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
982 983
				[IB_QPT_UC]  = (IB_QP_CUR_STATE			|
						IB_QP_ACCESS_FLAGS		|
984 985
						IB_QP_ALT_PATH			|
						IB_QP_PATH_MIG_STATE),
986 987
				[IB_QPT_RC]  = (IB_QP_CUR_STATE			|
						IB_QP_ACCESS_FLAGS		|
988 989 990
						IB_QP_ALT_PATH			|
						IB_QP_PATH_MIG_STATE		|
						IB_QP_MIN_RNR_TIMER),
S
Sean Hefty 已提交
991 992 993 994 995 996 997 998 999
				[IB_QPT_XRC_INI] = (IB_QP_CUR_STATE		|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_ALT_PATH			|
						IB_QP_PATH_MIG_STATE),
				[IB_QPT_XRC_TGT] = (IB_QP_CUR_STATE		|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_ALT_PATH			|
						IB_QP_PATH_MIG_STATE		|
						IB_QP_MIN_RNR_TIMER),
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
				[IB_QPT_SMI] = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
				[IB_QPT_GSI] = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
			}
		},
		[IB_QPS_SQD]   = {
			.valid = 1,
			.opt_param = {
				[IB_QPT_UD]  = IB_QP_EN_SQD_ASYNC_NOTIFY,
				[IB_QPT_UC]  = IB_QP_EN_SQD_ASYNC_NOTIFY,
				[IB_QPT_RC]  = IB_QP_EN_SQD_ASYNC_NOTIFY,
S
Sean Hefty 已提交
1012 1013
				[IB_QPT_XRC_INI] = IB_QP_EN_SQD_ASYNC_NOTIFY,
				[IB_QPT_XRC_TGT] = IB_QP_EN_SQD_ASYNC_NOTIFY, /* ??? */
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
				[IB_QPT_SMI] = IB_QP_EN_SQD_ASYNC_NOTIFY,
				[IB_QPT_GSI] = IB_QP_EN_SQD_ASYNC_NOTIFY
			}
		},
	},
	[IB_QPS_SQD]   = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_ERR] =   { .valid = 1 },
		[IB_QPS_RTS]   = {
			.valid = 1,
			.opt_param = {
				[IB_QPT_UD]  = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
				[IB_QPT_UC]  = (IB_QP_CUR_STATE			|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_PATH_MIG_STATE),
				[IB_QPT_RC]  = (IB_QP_CUR_STATE			|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_MIN_RNR_TIMER		|
						IB_QP_PATH_MIG_STATE),
S
Sean Hefty 已提交
1036 1037 1038 1039 1040 1041 1042 1043 1044
				[IB_QPT_XRC_INI] = (IB_QP_CUR_STATE		|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_PATH_MIG_STATE),
				[IB_QPT_XRC_TGT] = (IB_QP_CUR_STATE		|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_MIN_RNR_TIMER		|
						IB_QP_PATH_MIG_STATE),
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
				[IB_QPT_SMI] = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
				[IB_QPT_GSI] = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
			}
		},
		[IB_QPS_SQD]   = {
			.valid = 1,
			.opt_param = {
				[IB_QPT_UD]  = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
				[IB_QPT_UC]  = (IB_QP_AV			|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_PKEY_INDEX		|
						IB_QP_PATH_MIG_STATE),
				[IB_QPT_RC]  = (IB_QP_PORT			|
						IB_QP_AV			|
						IB_QP_TIMEOUT			|
						IB_QP_RETRY_CNT			|
						IB_QP_RNR_RETRY			|
						IB_QP_MAX_QP_RD_ATOMIC		|
						IB_QP_MAX_DEST_RD_ATOMIC	|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_PKEY_INDEX		|
						IB_QP_MIN_RNR_TIMER		|
						IB_QP_PATH_MIG_STATE),
S
Sean Hefty 已提交
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
				[IB_QPT_XRC_INI] = (IB_QP_PORT			|
						IB_QP_AV			|
						IB_QP_TIMEOUT			|
						IB_QP_RETRY_CNT			|
						IB_QP_RNR_RETRY			|
						IB_QP_MAX_QP_RD_ATOMIC		|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_PKEY_INDEX		|
						IB_QP_PATH_MIG_STATE),
				[IB_QPT_XRC_TGT] = (IB_QP_PORT			|
						IB_QP_AV			|
						IB_QP_TIMEOUT			|
						IB_QP_MAX_DEST_RD_ATOMIC	|
						IB_QP_ALT_PATH			|
						IB_QP_ACCESS_FLAGS		|
						IB_QP_PKEY_INDEX		|
						IB_QP_MIN_RNR_TIMER		|
						IB_QP_PATH_MIG_STATE),
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
				[IB_QPT_SMI] = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
				[IB_QPT_GSI] = (IB_QP_PKEY_INDEX		|
						IB_QP_QKEY),
			}
		}
	},
	[IB_QPS_SQE]   = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_ERR] =   { .valid = 1 },
		[IB_QPS_RTS]   = {
			.valid = 1,
			.opt_param = {
				[IB_QPT_UD]  = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
				[IB_QPT_UC]  = (IB_QP_CUR_STATE			|
						IB_QP_ACCESS_FLAGS),
				[IB_QPT_SMI] = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
				[IB_QPT_GSI] = (IB_QP_CUR_STATE			|
						IB_QP_QKEY),
			}
		}
	},
	[IB_QPS_ERR] = {
		[IB_QPS_RESET] = { .valid = 1 },
		[IB_QPS_ERR] =   { .valid = 1 }
	}
};

int ib_modify_qp_is_ok(enum ib_qp_state cur_state, enum ib_qp_state next_state,
1123 1124
		       enum ib_qp_type type, enum ib_qp_attr_mask mask,
		       enum rdma_link_layer ll)
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
{
	enum ib_qp_attr_mask req_param, opt_param;

	if (cur_state  < 0 || cur_state  > IB_QPS_ERR ||
	    next_state < 0 || next_state > IB_QPS_ERR)
		return 0;

	if (mask & IB_QP_CUR_STATE  &&
	    cur_state != IB_QPS_RTR && cur_state != IB_QPS_RTS &&
	    cur_state != IB_QPS_SQD && cur_state != IB_QPS_SQE)
		return 0;

	if (!qp_state_table[cur_state][next_state].valid)
		return 0;

	req_param = qp_state_table[cur_state][next_state].req_param[type];
	opt_param = qp_state_table[cur_state][next_state].opt_param[type];

	if ((mask & req_param) != req_param)
		return 0;

	if (mask & ~(req_param | opt_param | IB_QP_STATE))
		return 0;

	return 1;
}
EXPORT_SYMBOL(ib_modify_qp_is_ok);

1153 1154
int ib_resolve_eth_dmac(struct ib_qp *qp,
			struct ib_qp_attr *qp_attr, int *qp_attr_mask)
1155 1156 1157
{
	int           ret = 0;

1158 1159 1160 1161 1162 1163 1164 1165
	if (*qp_attr_mask & IB_QP_AV) {
		if (qp_attr->ah_attr.port_num < rdma_start_port(qp->device) ||
		    qp_attr->ah_attr.port_num > rdma_end_port(qp->device))
			return -EINVAL;

		if (!rdma_cap_eth_ah(qp->device, qp_attr->ah_attr.port_num))
			return 0;

1166
		if (rdma_link_local_addr((struct in6_addr *)qp_attr->ah_attr.grh.dgid.raw)) {
1167 1168
			rdma_get_ll_mac((struct in6_addr *)qp_attr->ah_attr.grh.dgid.raw,
					qp_attr->ah_attr.dmac);
1169
		} else {
1170 1171 1172
			union ib_gid		sgid;
			struct ib_gid_attr	sgid_attr;
			int			ifindex;
1173
			int			hop_limit;
1174 1175 1176 1177 1178 1179 1180 1181 1182

			ret = ib_query_gid(qp->device,
					   qp_attr->ah_attr.port_num,
					   qp_attr->ah_attr.grh.sgid_index,
					   &sgid, &sgid_attr);

			if (ret || !sgid_attr.ndev) {
				if (!ret)
					ret = -ENXIO;
1183
				goto out;
1184 1185 1186 1187
			}

			ifindex = sgid_attr.ndev->ifindex;

1188 1189 1190
			ret = rdma_addr_find_l2_eth_by_grh(&sgid,
							   &qp_attr->ah_attr.grh.dgid,
							   qp_attr->ah_attr.dmac,
1191
							   NULL, &ifindex, &hop_limit);
1192 1193

			dev_put(sgid_attr.ndev);
1194 1195

			qp_attr->ah_attr.grh.hop_limit = hop_limit;
1196 1197 1198 1199 1200
		}
	}
out:
	return ret;
}
1201
EXPORT_SYMBOL(ib_resolve_eth_dmac);
1202 1203


L
Linus Torvalds 已提交
1204 1205 1206 1207
int ib_modify_qp(struct ib_qp *qp,
		 struct ib_qp_attr *qp_attr,
		 int qp_attr_mask)
{
1208 1209
	int ret;

1210
	ret = ib_resolve_eth_dmac(qp, qp_attr, &qp_attr_mask);
1211 1212 1213
	if (ret)
		return ret;

1214
	return qp->device->modify_qp(qp->real_qp, qp_attr, qp_attr_mask, NULL);
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219 1220 1221 1222 1223
}
EXPORT_SYMBOL(ib_modify_qp);

int ib_query_qp(struct ib_qp *qp,
		struct ib_qp_attr *qp_attr,
		int qp_attr_mask,
		struct ib_qp_init_attr *qp_init_attr)
{
	return qp->device->query_qp ?
1224
		qp->device->query_qp(qp->real_qp, qp_attr, qp_attr_mask, qp_init_attr) :
L
Linus Torvalds 已提交
1225 1226 1227 1228
		-ENOSYS;
}
EXPORT_SYMBOL(ib_query_qp);

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 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
int ib_close_qp(struct ib_qp *qp)
{
	struct ib_qp *real_qp;
	unsigned long flags;

	real_qp = qp->real_qp;
	if (real_qp == qp)
		return -EINVAL;

	spin_lock_irqsave(&real_qp->device->event_handler_lock, flags);
	list_del(&qp->open_list);
	spin_unlock_irqrestore(&real_qp->device->event_handler_lock, flags);

	atomic_dec(&real_qp->usecnt);
	kfree(qp);

	return 0;
}
EXPORT_SYMBOL(ib_close_qp);

static int __ib_destroy_shared_qp(struct ib_qp *qp)
{
	struct ib_xrcd *xrcd;
	struct ib_qp *real_qp;
	int ret;

	real_qp = qp->real_qp;
	xrcd = real_qp->xrcd;

	mutex_lock(&xrcd->tgt_qp_mutex);
	ib_close_qp(qp);
	if (atomic_read(&real_qp->usecnt) == 0)
		list_del(&real_qp->xrcd_list);
	else
		real_qp = NULL;
	mutex_unlock(&xrcd->tgt_qp_mutex);

	if (real_qp) {
		ret = ib_destroy_qp(real_qp);
		if (!ret)
			atomic_dec(&xrcd->usecnt);
		else
			__ib_insert_xrcd_qp(xrcd, real_qp);
	}

	return 0;
}

L
Linus Torvalds 已提交
1277 1278 1279 1280 1281 1282 1283
int ib_destroy_qp(struct ib_qp *qp)
{
	struct ib_pd *pd;
	struct ib_cq *scq, *rcq;
	struct ib_srq *srq;
	int ret;

1284 1285
	WARN_ON_ONCE(qp->mrs_used > 0);

1286 1287 1288 1289 1290 1291
	if (atomic_read(&qp->usecnt))
		return -EBUSY;

	if (qp->real_qp != qp)
		return __ib_destroy_shared_qp(qp);

S
Sean Hefty 已提交
1292 1293 1294 1295
	pd   = qp->pd;
	scq  = qp->send_cq;
	rcq  = qp->recv_cq;
	srq  = qp->srq;
L
Linus Torvalds 已提交
1296

1297 1298 1299
	if (!qp->uobject)
		rdma_rw_cleanup_mrs(qp);

L
Linus Torvalds 已提交
1300 1301
	ret = qp->device->destroy_qp(qp);
	if (!ret) {
S
Sean Hefty 已提交
1302 1303 1304 1305 1306 1307
		if (pd)
			atomic_dec(&pd->usecnt);
		if (scq)
			atomic_dec(&scq->usecnt);
		if (rcq)
			atomic_dec(&rcq->usecnt);
L
Linus Torvalds 已提交
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
		if (srq)
			atomic_dec(&srq->usecnt);
	}

	return ret;
}
EXPORT_SYMBOL(ib_destroy_qp);

/* Completion queues */

struct ib_cq *ib_create_cq(struct ib_device *device,
			   ib_comp_handler comp_handler,
			   void (*event_handler)(struct ib_event *, void *),
1321 1322
			   void *cq_context,
			   const struct ib_cq_init_attr *cq_attr)
L
Linus Torvalds 已提交
1323 1324 1325
{
	struct ib_cq *cq;

1326
	cq = device->create_cq(device, cq_attr, NULL, NULL);
L
Linus Torvalds 已提交
1327 1328 1329

	if (!IS_ERR(cq)) {
		cq->device        = device;
1330
		cq->uobject       = NULL;
L
Linus Torvalds 已提交
1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
		cq->comp_handler  = comp_handler;
		cq->event_handler = event_handler;
		cq->cq_context    = cq_context;
		atomic_set(&cq->usecnt, 0);
	}

	return cq;
}
EXPORT_SYMBOL(ib_create_cq);

E
Eli Cohen 已提交
1341 1342 1343 1344 1345 1346 1347
int ib_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period)
{
	return cq->device->modify_cq ?
		cq->device->modify_cq(cq, cq_count, cq_period) : -ENOSYS;
}
EXPORT_SYMBOL(ib_modify_cq);

L
Linus Torvalds 已提交
1348 1349 1350 1351 1352 1353 1354 1355 1356
int ib_destroy_cq(struct ib_cq *cq)
{
	if (atomic_read(&cq->usecnt))
		return -EBUSY;

	return cq->device->destroy_cq(cq);
}
EXPORT_SYMBOL(ib_destroy_cq);

R
Roland Dreier 已提交
1357
int ib_resize_cq(struct ib_cq *cq, int cqe)
L
Linus Torvalds 已提交
1358
{
1359
	return cq->device->resize_cq ?
1360
		cq->device->resize_cq(cq, cqe, NULL) : -ENOSYS;
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367 1368
}
EXPORT_SYMBOL(ib_resize_cq);

/* Memory regions */

struct ib_mr *ib_get_dma_mr(struct ib_pd *pd, int mr_access_flags)
{
	struct ib_mr *mr;
1369 1370 1371 1372 1373
	int err;

	err = ib_check_mr_access(mr_access_flags);
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
1374 1375 1376 1377

	mr = pd->device->get_dma_mr(pd, mr_access_flags);

	if (!IS_ERR(mr)) {
1378 1379 1380
		mr->device  = pd->device;
		mr->pd      = pd;
		mr->uobject = NULL;
L
Linus Torvalds 已提交
1381
		atomic_inc(&pd->usecnt);
1382
		mr->need_inval = false;
L
Linus Torvalds 已提交
1383 1384 1385 1386 1387 1388 1389 1390
	}

	return mr;
}
EXPORT_SYMBOL(ib_get_dma_mr);

int ib_dereg_mr(struct ib_mr *mr)
{
1391
	struct ib_pd *pd = mr->pd;
L
Linus Torvalds 已提交
1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
	int ret;

	ret = mr->device->dereg_mr(mr);
	if (!ret)
		atomic_dec(&pd->usecnt);

	return ret;
}
EXPORT_SYMBOL(ib_dereg_mr);

S
Sagi Grimberg 已提交
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416
/**
 * ib_alloc_mr() - Allocates a memory region
 * @pd:            protection domain associated with the region
 * @mr_type:       memory region type
 * @max_num_sg:    maximum sg entries available for registration.
 *
 * Notes:
 * Memory registeration page/sg lists must not exceed max_num_sg.
 * For mr_type IB_MR_TYPE_MEM_REG, the total length cannot exceed
 * max_num_sg * used_page_size.
 *
 */
struct ib_mr *ib_alloc_mr(struct ib_pd *pd,
			  enum ib_mr_type mr_type,
			  u32 max_num_sg)
1417 1418 1419
{
	struct ib_mr *mr;

1420
	if (!pd->device->alloc_mr)
1421 1422
		return ERR_PTR(-ENOSYS);

1423
	mr = pd->device->alloc_mr(pd, mr_type, max_num_sg);
1424 1425 1426 1427 1428
	if (!IS_ERR(mr)) {
		mr->device  = pd->device;
		mr->pd      = pd;
		mr->uobject = NULL;
		atomic_inc(&pd->usecnt);
1429
		mr->need_inval = false;
1430 1431 1432 1433
	}

	return mr;
}
1434
EXPORT_SYMBOL(ib_alloc_mr);
1435

L
Linus Torvalds 已提交
1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
/* "Fast" memory regions */

struct ib_fmr *ib_alloc_fmr(struct ib_pd *pd,
			    int mr_access_flags,
			    struct ib_fmr_attr *fmr_attr)
{
	struct ib_fmr *fmr;

	if (!pd->device->alloc_fmr)
		return ERR_PTR(-ENOSYS);

	fmr = pd->device->alloc_fmr(pd, mr_access_flags, fmr_attr);
	if (!IS_ERR(fmr)) {
		fmr->device = pd->device;
		fmr->pd     = pd;
		atomic_inc(&pd->usecnt);
	}

	return fmr;
}
EXPORT_SYMBOL(ib_alloc_fmr);

int ib_unmap_fmr(struct list_head *fmr_list)
{
	struct ib_fmr *fmr;

	if (list_empty(fmr_list))
		return 0;

	fmr = list_entry(fmr_list->next, struct ib_fmr, list);
	return fmr->device->unmap_fmr(fmr_list);
}
EXPORT_SYMBOL(ib_unmap_fmr);

int ib_dealloc_fmr(struct ib_fmr *fmr)
{
	struct ib_pd *pd;
	int ret;

	pd = fmr->pd;
	ret = fmr->device->dealloc_fmr(fmr);
	if (!ret)
		atomic_dec(&pd->usecnt);

	return ret;
}
EXPORT_SYMBOL(ib_dealloc_fmr);

/* Multicast groups */

int ib_attach_mcast(struct ib_qp *qp, union ib_gid *gid, u16 lid)
{
1488 1489
	int ret;

1490 1491 1492 1493 1494
	if (!qp->device->attach_mcast)
		return -ENOSYS;
	if (gid->raw[0] != 0xff || qp->qp_type != IB_QPT_UD)
		return -EINVAL;

1495 1496 1497 1498
	ret = qp->device->attach_mcast(qp, gid, lid);
	if (!ret)
		atomic_inc(&qp->usecnt);
	return ret;
L
Linus Torvalds 已提交
1499 1500 1501 1502 1503
}
EXPORT_SYMBOL(ib_attach_mcast);

int ib_detach_mcast(struct ib_qp *qp, union ib_gid *gid, u16 lid)
{
1504 1505
	int ret;

1506 1507 1508 1509 1510
	if (!qp->device->detach_mcast)
		return -ENOSYS;
	if (gid->raw[0] != 0xff || qp->qp_type != IB_QPT_UD)
		return -EINVAL;

1511 1512 1513 1514
	ret = qp->device->detach_mcast(qp, gid, lid);
	if (!ret)
		atomic_dec(&qp->usecnt);
	return ret;
L
Linus Torvalds 已提交
1515 1516
}
EXPORT_SYMBOL(ib_detach_mcast);
S
Sean Hefty 已提交
1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527

struct ib_xrcd *ib_alloc_xrcd(struct ib_device *device)
{
	struct ib_xrcd *xrcd;

	if (!device->alloc_xrcd)
		return ERR_PTR(-ENOSYS);

	xrcd = device->alloc_xrcd(device, NULL, NULL);
	if (!IS_ERR(xrcd)) {
		xrcd->device = device;
1528
		xrcd->inode = NULL;
S
Sean Hefty 已提交
1529
		atomic_set(&xrcd->usecnt, 0);
1530 1531
		mutex_init(&xrcd->tgt_qp_mutex);
		INIT_LIST_HEAD(&xrcd->tgt_qp_list);
S
Sean Hefty 已提交
1532 1533 1534 1535 1536 1537 1538 1539
	}

	return xrcd;
}
EXPORT_SYMBOL(ib_alloc_xrcd);

int ib_dealloc_xrcd(struct ib_xrcd *xrcd)
{
1540 1541 1542
	struct ib_qp *qp;
	int ret;

S
Sean Hefty 已提交
1543 1544 1545
	if (atomic_read(&xrcd->usecnt))
		return -EBUSY;

1546 1547 1548 1549 1550 1551 1552
	while (!list_empty(&xrcd->tgt_qp_list)) {
		qp = list_entry(xrcd->tgt_qp_list.next, struct ib_qp, xrcd_list);
		ret = ib_destroy_qp(qp);
		if (ret)
			return ret;
	}

S
Sean Hefty 已提交
1553 1554 1555
	return xrcd->device->dealloc_xrcd(xrcd);
}
EXPORT_SYMBOL(ib_dealloc_xrcd);
1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582

struct ib_flow *ib_create_flow(struct ib_qp *qp,
			       struct ib_flow_attr *flow_attr,
			       int domain)
{
	struct ib_flow *flow_id;
	if (!qp->device->create_flow)
		return ERR_PTR(-ENOSYS);

	flow_id = qp->device->create_flow(qp, flow_attr, domain);
	if (!IS_ERR(flow_id))
		atomic_inc(&qp->usecnt);
	return flow_id;
}
EXPORT_SYMBOL(ib_create_flow);

int ib_destroy_flow(struct ib_flow *flow_id)
{
	int err;
	struct ib_qp *qp = flow_id->qp;

	err = qp->device->destroy_flow(flow_id);
	if (!err)
		atomic_dec(&qp->usecnt);
	return err;
}
EXPORT_SYMBOL(ib_destroy_flow);
1583 1584 1585 1586 1587 1588 1589 1590

int ib_check_mr_status(struct ib_mr *mr, u32 check_mask,
		       struct ib_mr_status *mr_status)
{
	return mr->device->check_mr_status ?
		mr->device->check_mr_status(mr, check_mask, mr_status) : -ENOSYS;
}
EXPORT_SYMBOL(ib_check_mr_status);
1591

1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631
int ib_set_vf_link_state(struct ib_device *device, int vf, u8 port,
			 int state)
{
	if (!device->set_vf_link_state)
		return -ENOSYS;

	return device->set_vf_link_state(device, vf, port, state);
}
EXPORT_SYMBOL(ib_set_vf_link_state);

int ib_get_vf_config(struct ib_device *device, int vf, u8 port,
		     struct ifla_vf_info *info)
{
	if (!device->get_vf_config)
		return -ENOSYS;

	return device->get_vf_config(device, vf, port, info);
}
EXPORT_SYMBOL(ib_get_vf_config);

int ib_get_vf_stats(struct ib_device *device, int vf, u8 port,
		    struct ifla_vf_stats *stats)
{
	if (!device->get_vf_stats)
		return -ENOSYS;

	return device->get_vf_stats(device, vf, port, stats);
}
EXPORT_SYMBOL(ib_get_vf_stats);

int ib_set_vf_guid(struct ib_device *device, int vf, u8 port, u64 guid,
		   int type)
{
	if (!device->set_vf_guid)
		return -ENOSYS;

	return device->set_vf_guid(device, vf, port, guid, type);
}
EXPORT_SYMBOL(ib_set_vf_guid);

1632 1633 1634 1635 1636 1637
/**
 * ib_map_mr_sg() - Map the largest prefix of a dma mapped SG list
 *     and set it the memory region.
 * @mr:            memory region
 * @sg:            dma mapped scatterlist
 * @sg_nents:      number of entries in sg
1638
 * @sg_offset:     offset in bytes into sg
1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
 * @page_size:     page vector desired page size
 *
 * Constraints:
 * - The first sg element is allowed to have an offset.
 * - Each sg element must be aligned to page_size (or physically
 *   contiguous to the previous element). In case an sg element has a
 *   non contiguous offset, the mapping prefix will not include it.
 * - The last sg element is allowed to have length less than page_size.
 * - If sg_nents total byte length exceeds the mr max_num_sge * page_size
 *   then only max_num_sg entries will be mapped.
1649 1650
 * - If the MR was allocated with type IB_MR_TYPE_SG_GAPS_REG, non of these
 *   constraints holds and the page_size argument is ignored.
1651 1652 1653 1654 1655 1656
 *
 * Returns the number of sg elements that were mapped to the memory region.
 *
 * After this completes successfully, the  memory region
 * is ready for registration.
 */
1657 1658
int ib_map_mr_sg(struct ib_mr *mr, struct scatterlist *sg, int sg_nents,
		unsigned int sg_offset, unsigned int page_size)
1659 1660 1661 1662 1663 1664
{
	if (unlikely(!mr->device->map_mr_sg))
		return -ENOSYS;

	mr->page_size = page_size;

1665
	return mr->device->map_mr_sg(mr, sg, sg_nents, sg_offset);
1666 1667 1668 1669 1670 1671 1672 1673 1674
}
EXPORT_SYMBOL(ib_map_mr_sg);

/**
 * ib_sg_to_pages() - Convert the largest prefix of a sg list
 *     to a page vector
 * @mr:            memory region
 * @sgl:           dma mapped scatterlist
 * @sg_nents:      number of entries in sg
1675
 * @sg_offset:     offset in bytes into sg
1676 1677
 * @set_page:      driver page assignment function pointer
 *
B
Bart Van Assche 已提交
1678
 * Core service helper for drivers to convert the largest
1679 1680 1681 1682 1683 1684 1685
 * prefix of given sg list to a page vector. The sg list
 * prefix converted is the prefix that meet the requirements
 * of ib_map_mr_sg.
 *
 * Returns the number of sg elements that were assigned to
 * a page vector.
 */
1686 1687
int ib_sg_to_pages(struct ib_mr *mr, struct scatterlist *sgl, int sg_nents,
		unsigned int sg_offset, int (*set_page)(struct ib_mr *, u64))
1688 1689
{
	struct scatterlist *sg;
1690
	u64 last_end_dma_addr = 0;
1691 1692
	unsigned int last_page_off = 0;
	u64 page_mask = ~((u64)mr->page_size - 1);
B
Bart Van Assche 已提交
1693
	int i, ret;
1694

1695
	mr->iova = sg_dma_address(&sgl[0]) + sg_offset;
1696 1697 1698
	mr->length = 0;

	for_each_sg(sgl, sg, sg_nents, i) {
1699 1700
		u64 dma_addr = sg_dma_address(sg) + sg_offset;
		unsigned int dma_len = sg_dma_len(sg) - sg_offset;
1701 1702 1703
		u64 end_dma_addr = dma_addr + dma_len;
		u64 page_addr = dma_addr & page_mask;

B
Bart Van Assche 已提交
1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
		/*
		 * For the second and later elements, check whether either the
		 * end of element i-1 or the start of element i is not aligned
		 * on a page boundary.
		 */
		if (i && (last_page_off != 0 || page_addr != dma_addr)) {
			/* Stop mapping if there is a gap. */
			if (last_end_dma_addr != dma_addr)
				break;

			/*
			 * Coalesce this element with the last. If it is small
			 * enough just update mr->length. Otherwise start
			 * mapping from the next page.
			 */
			goto next_page;
1720 1721 1722
		}

		do {
B
Bart Van Assche 已提交
1723 1724 1725 1726
			ret = set_page(mr, page_addr);
			if (unlikely(ret < 0))
				return i ? : ret;
next_page:
1727 1728 1729 1730 1731 1732
			page_addr += mr->page_size;
		} while (page_addr < end_dma_addr);

		mr->length += dma_len;
		last_end_dma_addr = end_dma_addr;
		last_page_off = end_dma_addr & ~page_mask;
1733 1734

		sg_offset = 0;
1735 1736 1737 1738 1739
	}

	return i;
}
EXPORT_SYMBOL(ib_sg_to_pages);
1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900

struct ib_drain_cqe {
	struct ib_cqe cqe;
	struct completion done;
};

static void ib_drain_qp_done(struct ib_cq *cq, struct ib_wc *wc)
{
	struct ib_drain_cqe *cqe = container_of(wc->wr_cqe, struct ib_drain_cqe,
						cqe);

	complete(&cqe->done);
}

/*
 * Post a WR and block until its completion is reaped for the SQ.
 */
static void __ib_drain_sq(struct ib_qp *qp)
{
	struct ib_qp_attr attr = { .qp_state = IB_QPS_ERR };
	struct ib_drain_cqe sdrain;
	struct ib_send_wr swr = {}, *bad_swr;
	int ret;

	if (qp->send_cq->poll_ctx == IB_POLL_DIRECT) {
		WARN_ONCE(qp->send_cq->poll_ctx == IB_POLL_DIRECT,
			  "IB_POLL_DIRECT poll_ctx not supported for drain\n");
		return;
	}

	swr.wr_cqe = &sdrain.cqe;
	sdrain.cqe.done = ib_drain_qp_done;
	init_completion(&sdrain.done);

	ret = ib_modify_qp(qp, &attr, IB_QP_STATE);
	if (ret) {
		WARN_ONCE(ret, "failed to drain send queue: %d\n", ret);
		return;
	}

	ret = ib_post_send(qp, &swr, &bad_swr);
	if (ret) {
		WARN_ONCE(ret, "failed to drain send queue: %d\n", ret);
		return;
	}

	wait_for_completion(&sdrain.done);
}

/*
 * Post a WR and block until its completion is reaped for the RQ.
 */
static void __ib_drain_rq(struct ib_qp *qp)
{
	struct ib_qp_attr attr = { .qp_state = IB_QPS_ERR };
	struct ib_drain_cqe rdrain;
	struct ib_recv_wr rwr = {}, *bad_rwr;
	int ret;

	if (qp->recv_cq->poll_ctx == IB_POLL_DIRECT) {
		WARN_ONCE(qp->recv_cq->poll_ctx == IB_POLL_DIRECT,
			  "IB_POLL_DIRECT poll_ctx not supported for drain\n");
		return;
	}

	rwr.wr_cqe = &rdrain.cqe;
	rdrain.cqe.done = ib_drain_qp_done;
	init_completion(&rdrain.done);

	ret = ib_modify_qp(qp, &attr, IB_QP_STATE);
	if (ret) {
		WARN_ONCE(ret, "failed to drain recv queue: %d\n", ret);
		return;
	}

	ret = ib_post_recv(qp, &rwr, &bad_rwr);
	if (ret) {
		WARN_ONCE(ret, "failed to drain recv queue: %d\n", ret);
		return;
	}

	wait_for_completion(&rdrain.done);
}

/**
 * ib_drain_sq() - Block until all SQ CQEs have been consumed by the
 *		   application.
 * @qp:            queue pair to drain
 *
 * If the device has a provider-specific drain function, then
 * call that.  Otherwise call the generic drain function
 * __ib_drain_sq().
 *
 * The caller must:
 *
 * ensure there is room in the CQ and SQ for the drain work request and
 * completion.
 *
 * allocate the CQ using ib_alloc_cq() and the CQ poll context cannot be
 * IB_POLL_DIRECT.
 *
 * ensure that there are no other contexts that are posting WRs concurrently.
 * Otherwise the drain is not guaranteed.
 */
void ib_drain_sq(struct ib_qp *qp)
{
	if (qp->device->drain_sq)
		qp->device->drain_sq(qp);
	else
		__ib_drain_sq(qp);
}
EXPORT_SYMBOL(ib_drain_sq);

/**
 * ib_drain_rq() - Block until all RQ CQEs have been consumed by the
 *		   application.
 * @qp:            queue pair to drain
 *
 * If the device has a provider-specific drain function, then
 * call that.  Otherwise call the generic drain function
 * __ib_drain_rq().
 *
 * The caller must:
 *
 * ensure there is room in the CQ and RQ for the drain work request and
 * completion.
 *
 * allocate the CQ using ib_alloc_cq() and the CQ poll context cannot be
 * IB_POLL_DIRECT.
 *
 * ensure that there are no other contexts that are posting WRs concurrently.
 * Otherwise the drain is not guaranteed.
 */
void ib_drain_rq(struct ib_qp *qp)
{
	if (qp->device->drain_rq)
		qp->device->drain_rq(qp);
	else
		__ib_drain_rq(qp);
}
EXPORT_SYMBOL(ib_drain_rq);

/**
 * ib_drain_qp() - Block until all CQEs have been consumed by the
 *		   application on both the RQ and SQ.
 * @qp:            queue pair to drain
 *
 * The caller must:
 *
 * ensure there is room in the CQ(s), SQ, and RQ for drain work requests
 * and completions.
 *
 * allocate the CQs using ib_alloc_cq() and the CQ poll context cannot be
 * IB_POLL_DIRECT.
 *
 * ensure that there are no other contexts that are posting WRs concurrently.
 * Otherwise the drain is not guaranteed.
 */
void ib_drain_qp(struct ib_qp *qp)
{
	ib_drain_sq(qp);
1901 1902
	if (!qp->srq)
		ib_drain_rq(qp);
1903 1904
}
EXPORT_SYMBOL(ib_drain_qp);