hfi.h 72.6 KB
Newer Older
M
Mike Marciniszyn 已提交
1 2 3
#ifndef _HFI1_KERNEL_H
#define _HFI1_KERNEL_H
/*
4
 * Copyright(c) 2015-2017 Intel Corporation.
M
Mike Marciniszyn 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
 *
 * This file is provided under a dual BSD/GPLv2 license.  When using or
 * redistributing this file, you may do so under either license.
 *
 * GPL LICENSE SUMMARY
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of version 2 of the GNU General Public License as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * BSD LICENSE
 *
 * 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.
 *  - Neither the name of Intel Corporation nor the names of its
 *    contributors may be used to endorse or promote products derived
 *    from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 */

#include <linux/interrupt.h>
#include <linux/pci.h>
#include <linux/dma-mapping.h>
#include <linux/mutex.h>
#include <linux/list.h>
#include <linux/scatterlist.h>
#include <linux/slab.h>
57
#include <linux/idr.h>
M
Mike Marciniszyn 已提交
58 59 60 61 62 63 64 65
#include <linux/io.h>
#include <linux/fs.h>
#include <linux/completion.h>
#include <linux/kref.h>
#include <linux/sched.h>
#include <linux/cdev.h>
#include <linux/delay.h>
#include <linux/kthread.h>
66 67
#include <linux/i2c.h>
#include <linux/i2c-algo-bit.h>
68
#include <rdma/ib_hdrs.h>
69
#include <rdma/opa_addr.h>
70
#include <linux/rhashtable.h>
71
#include <linux/netdevice.h>
72
#include <rdma/rdma_vt.h>
73
#include <rdma/opa_addr.h>
M
Mike Marciniszyn 已提交
74 75 76 77 78 79 80 81

#include "chip_registers.h"
#include "common.h"
#include "verbs.h"
#include "pio.h"
#include "chip.h"
#include "mad.h"
#include "qsfp.h"
82
#include "platform.h"
83
#include "affinity.h"
M
Mike Marciniszyn 已提交
84 85 86 87 88 89 90 91 92 93 94 95 96 97

/* bumped 1 from s/w major version of TrueScale */
#define HFI1_CHIP_VERS_MAJ 3U

/* don't care about this except printing */
#define HFI1_CHIP_VERS_MIN 0U

/* The Organization Unique Identifier (Mfg code), and its position in GUID */
#define HFI1_OUI 0x001175
#define HFI1_OUI_LSB 40

#define DROP_PACKET_OFF		0
#define DROP_PACKET_ON		1

98 99 100
#define NEIGHBOR_TYPE_HFI		0
#define NEIGHBOR_TYPE_SWITCH	1

M
Mike Marciniszyn 已提交
101 102 103 104 105 106 107 108 109 110
extern unsigned long hfi1_cap_mask;
#define HFI1_CAP_KGET_MASK(mask, cap) ((mask) & HFI1_CAP_##cap)
#define HFI1_CAP_UGET_MASK(mask, cap) \
	(((mask) >> HFI1_CAP_USER_SHIFT) & HFI1_CAP_##cap)
#define HFI1_CAP_KGET(cap) (HFI1_CAP_KGET_MASK(hfi1_cap_mask, cap))
#define HFI1_CAP_UGET(cap) (HFI1_CAP_UGET_MASK(hfi1_cap_mask, cap))
#define HFI1_CAP_IS_KSET(cap) (!!HFI1_CAP_KGET(cap))
#define HFI1_CAP_IS_USET(cap) (!!HFI1_CAP_UGET(cap))
#define HFI1_MISC_GET() ((hfi1_cap_mask >> HFI1_CAP_MISC_SHIFT) & \
			HFI1_CAP_MISC_MASK)
111 112
/* Offline Disabled Reason is 4-bits */
#define HFI1_ODR_MASK(rsn) ((rsn) & OPA_PI_MASK_OFFLINE_REASON)
M
Mike Marciniszyn 已提交
113

114 115 116 117 118 119
/*
 * Control context is always 0 and handles the error packets.
 * It also handles the VL15 and multicast packets.
 */
#define HFI1_CTRL_CTXT    0

120 121 122 123 124 125 126 127 128 129 130 131 132 133
/*
 * Driver context will store software counters for each of the events
 * associated with these status registers
 */
#define NUM_CCE_ERR_STATUS_COUNTERS 41
#define NUM_RCV_ERR_STATUS_COUNTERS 64
#define NUM_MISC_ERR_STATUS_COUNTERS 13
#define NUM_SEND_PIO_ERR_STATUS_COUNTERS 36
#define NUM_SEND_DMA_ERR_STATUS_COUNTERS 4
#define NUM_SEND_EGRESS_ERR_STATUS_COUNTERS 64
#define NUM_SEND_ERR_STATUS_COUNTERS 3
#define NUM_SEND_CTXT_ERR_STATUS_COUNTERS 5
#define NUM_SEND_DMA_ENG_ERR_STATUS_COUNTERS 24

M
Mike Marciniszyn 已提交
134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
/*
 * per driver stats, either not device nor port-specific, or
 * summed over all of the devices and ports.
 * They are described by name via ipathfs filesystem, so layout
 * and number of elements can change without breaking compatibility.
 * If members are added or deleted hfi1_statnames[] in debugfs.c must
 * change to match.
 */
struct hfi1_ib_stats {
	__u64 sps_ints; /* number of interrupts handled */
	__u64 sps_errints; /* number of error interrupts */
	__u64 sps_txerrs; /* tx-related packet errors */
	__u64 sps_rcverrs; /* non-crc rcv packet errors */
	__u64 sps_hwerrs; /* hardware errors reported (parity, etc.) */
	__u64 sps_nopiobufs; /* no pio bufs avail from kernel */
	__u64 sps_ctxts; /* number of contexts currently open */
	__u64 sps_lenerrs; /* number of kernel packets where RHF != LRH len */
	__u64 sps_buffull;
	__u64 sps_hdrfull;
};

extern struct hfi1_ib_stats hfi1_stats;
extern const struct pci_error_handlers hfi1_pci_err_handler;

/*
 * First-cut criterion for "device is active" is
 * two thousand dwords combined Tx, Rx traffic per
 * 5-second interval. SMA packets are 64 dwords,
 * and occur "a few per second", presumably each way.
 */
#define HFI1_TRAFFIC_ACTIVE_THRESHOLD (2000)

/*
 * Below contains all data related to a single context (formerly called port).
 */

struct hfi1_opcode_stats_perctx;

struct ctxt_eager_bufs {
	ssize_t size;            /* total size of eager buffers */
	u32 count;               /* size of buffers array */
	u32 numbufs;             /* number of buffers allocated */
	u32 alloced;             /* number of rcvarray entries used */
	u32 rcvtid_size;         /* size of each eager rcv tid */
	u32 threshold;           /* head update threshold */
	struct eager_buffer {
		void *addr;
181
		dma_addr_t dma;
M
Mike Marciniszyn 已提交
182 183 184 185
		ssize_t len;
	} *buffers;
	struct {
		void *addr;
186
		dma_addr_t dma;
M
Mike Marciniszyn 已提交
187 188 189
	} *rcvtids;
};

190 191 192 193 194
struct exp_tid_set {
	struct list_head list;
	u32 count;
};

M
Mike Marciniszyn 已提交
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
struct hfi1_ctxtdata {
	/* shadow the ctxt's RcvCtrl register */
	u64 rcvctrl;
	/* rcvhdrq base, needs mmap before useful */
	void *rcvhdrq;
	/* kernel virtual address where hdrqtail is updated */
	volatile __le64 *rcvhdrtail_kvaddr;
	/* when waiting for rcv or pioavail */
	wait_queue_head_t wait;
	/* rcvhdrq size (for freeing) */
	size_t rcvhdrq_size;
	/* number of rcvhdrq entries */
	u16 rcvhdrq_cnt;
	/* size of each of the rcvhdrq entries */
	u16 rcvhdrqentsize;
	/* mmap of hdrq, must fit in 44 bits */
211 212
	dma_addr_t rcvhdrq_dma;
	dma_addr_t rcvhdrqtailaddr_dma;
M
Mike Marciniszyn 已提交
213 214 215 216 217 218
	struct ctxt_eager_bufs egrbufs;
	/* this receive context's assigned PIO ACK send context */
	struct send_context *sc;

	/* dynamic receive available interrupt timeout */
	u32 rcvavail_timeout;
219 220 221
	/* Reference count the base context usage */
	struct kref kref;

222
	/* Device context index */
223
	u16 ctxt;
M
Mike Marciniszyn 已提交
224
	/*
225
	 * non-zero if ctxt can be shared, and defines the maximum number of
226
	 * sub-contexts for this device context.
M
Mike Marciniszyn 已提交
227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
	 */
	u16 subctxt_cnt;
	/* non-zero if ctxt is being shared. */
	u16 subctxt_id;
	u8 uuid[16];
	/* job key */
	u16 jkey;
	/* number of RcvArray groups for this context. */
	u32 rcv_array_groups;
	/* index of first eager TID entry. */
	u32 eager_base;
	/* number of expected TID entries */
	u32 expected_count;
	/* index of first expected TID entry. */
	u32 expected_base;
242 243 244 245 246

	struct exp_tid_set tid_group_list;
	struct exp_tid_set tid_used_list;
	struct exp_tid_set tid_full_list;

M
Mike Marciniszyn 已提交
247
	/* lock protecting all Expected TID data */
248
	struct mutex exp_lock;
M
Mike Marciniszyn 已提交
249
	/* per-context configuration flags */
250
	unsigned long flags;
M
Mike Marciniszyn 已提交
251 252 253 254 255 256 257
	/* per-context event flags for fileops/intr communication */
	unsigned long event_flags;
	/* total number of polled urgent packets */
	u32 urgent;
	/* saved total number of polled urgent packets for poll edge trigger */
	u32 urgent_poll;
	/* same size as task_struct .comm[], command that opened context */
258
	char comm[TASK_COMM_LEN];
M
Mike Marciniszyn 已提交
259 260 261 262
	/* so file ops can get at unit */
	struct hfi1_devdata *dd;
	/* so functions that need physical port can get it easily */
	struct hfi1_pportdata *ppd;
263 264
	/* associated msix interrupt */
	u32 msix_intr;
M
Mike Marciniszyn 已提交
265 266 267 268 269 270
	/* A page of memory for rcvhdrhead, rcvegrhead, rcvegrtail * N */
	void *subctxt_uregbase;
	/* An array of pages for the eager receive buffers * N */
	void *subctxt_rcvegrbuf;
	/* An array of pages for the eager header queue entries * N */
	void *subctxt_rcvhdr_base;
271 272
	/* Bitmask of in use context(s) */
	DECLARE_BITMAP(in_use_ctxts, HFI1_MAX_SHARED_CTXTS);
M
Mike Marciniszyn 已提交
273 274 275 276 277 278 279 280 281 282 283 284 285 286
	/* The version of the library which opened this ctxt */
	u32 userversion;
	/* Type of packets or conditions we want to poll for */
	u16 poll_type;
	/* receive packet sequence counter */
	u8 seq_cnt;
	/* ctxt rcvhdrq head offset */
	u32 head;
	/* QPs waiting for context processing */
	struct list_head qp_wait_list;
	/* interrupt handling */
	u64 imask;	/* clear interrupt mask */
	int ireg;	/* clear interrupt register */
	unsigned numa_id; /* numa node of this context */
287
	/* verbs rx_stats per rcd */
M
Mike Marciniszyn 已提交
288 289
	struct hfi1_opcode_stats_perctx *opstats;

290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
	/* Is ASPM interrupt supported for this context */
	bool aspm_intr_supported;
	/* ASPM state (enabled/disabled) for this context */
	bool aspm_enabled;
	/* Timer for re-enabling ASPM if interrupt activity quietens down */
	struct timer_list aspm_timer;
	/* Lock to serialize between intr, timer intr and user threads */
	spinlock_t aspm_lock;
	/* Is ASPM processing enabled for this context (in intr context) */
	bool aspm_intr_enable;
	/* Last interrupt timestamp */
	ktime_t aspm_ts_last_intr;
	/* Last timestamp at which we scheduled a timer for this context */
	ktime_t aspm_ts_timer_sched;

M
Mike Marciniszyn 已提交
305 306 307 308 309 310 311
	/*
	 * The interrupt handler for a particular receive context can vary
	 * throughout it's lifetime. This is not a lock protected data member so
	 * it must be updated atomically and the prev and new value must always
	 * be valid. Worst case is we process an extra interrupt and up to 64
	 * packets with the wrong interrupt handler.
	 */
312
	int (*do_interrupt)(struct hfi1_ctxtdata *rcd, int threaded);
313 314 315 316 317 318

	/* Indicates that this is vnic context */
	bool is_vnic;

	/* vnic queue index this context is mapped to */
	u8 vnic_q_idx;
M
Mike Marciniszyn 已提交
319 320 321 322 323 324 325 326 327 328 329 330
};

/*
 * Represents a single packet at a high level. Put commonly computed things in
 * here so we do not have to keep doing them over and over. The rule of thumb is
 * if something is used one time to derive some value, store that something in
 * here. If it is used multiple times, then store the result of that derivation
 * in here.
 */
struct hfi1_packet {
	void *ebuf;
	void *hdr;
331
	void *payload;
M
Mike Marciniszyn 已提交
332 333
	struct hfi1_ctxtdata *rcd;
	__le32 *rhf_addr;
334
	struct rvt_qp *qp;
335
	struct ib_other_headers *ohdr;
336
	struct ib_grh *grh;
M
Mike Marciniszyn 已提交
337 338 339
	u64 rhf;
	u32 maxcnt;
	u32 rhqoff;
340 341
	u32 dlid;
	u32 slid;
M
Mike Marciniszyn 已提交
342 343
	u16 tlen;
	s16 etail;
344
	u16 pkey;
345 346 347
	u8 hlen;
	u8 numpkt;
	u8 rsize;
M
Mike Marciniszyn 已提交
348 349
	u8 updegr;
	u8 etype;
350 351 352 353 354
	u8 extra_byte;
	u8 pad;
	u8 sc;
	u8 sl;
	u8 opcode;
355
	bool migrated;
M
Mike Marciniszyn 已提交
356 357
};

358 359 360 361
/* Packet types */
#define HFI1_PKT_TYPE_9B  0
#define HFI1_PKT_TYPE_16B 1

362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
/*
 * OPA 16B Header
 */
#define OPA_16B_L4_MASK		0xFFull
#define OPA_16B_SC_MASK		0x1F00000ull
#define OPA_16B_SC_SHIFT	20
#define OPA_16B_LID_MASK	0xFFFFFull
#define OPA_16B_DLID_MASK	0xF000ull
#define OPA_16B_DLID_SHIFT	20
#define OPA_16B_DLID_HIGH_SHIFT	12
#define OPA_16B_SLID_MASK	0xF00ull
#define OPA_16B_SLID_SHIFT	20
#define OPA_16B_SLID_HIGH_SHIFT	8
#define OPA_16B_BECN_MASK       0x80000000ull
#define OPA_16B_BECN_SHIFT      31
#define OPA_16B_FECN_MASK       0x10000000ull
#define OPA_16B_FECN_SHIFT      28
#define OPA_16B_L2_MASK		0x60000000ull
#define OPA_16B_L2_SHIFT	29
381 382 383 384
#define OPA_16B_PKEY_MASK	0xFFFF0000ull
#define OPA_16B_PKEY_SHIFT	16
#define OPA_16B_LEN_MASK	0x7FF00000ull
#define OPA_16B_LEN_SHIFT	20
D
Don Hiatt 已提交
385 386 387 388 389
#define OPA_16B_RC_MASK		0xE000000ull
#define OPA_16B_RC_SHIFT	25
#define OPA_16B_AGE_MASK	0xFF0000ull
#define OPA_16B_AGE_SHIFT	16
#define OPA_16B_ENTROPY_MASK	0xFFFFull
390 391 392 393

/*
 * OPA 16B L2/L4 Encodings
 */
394
#define OPA_16B_L4_9B		0x00
395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
#define OPA_16B_L2_TYPE		0x02
#define OPA_16B_L4_IB_LOCAL	0x09
#define OPA_16B_L4_IB_GLOBAL	0x0A
#define OPA_16B_L4_ETHR		OPA_VNIC_L4_ETHR

static inline u8 hfi1_16B_get_l4(struct hfi1_16b_header *hdr)
{
	return (u8)(hdr->lrh[2] & OPA_16B_L4_MASK);
}

static inline u8 hfi1_16B_get_sc(struct hfi1_16b_header *hdr)
{
	return (u8)((hdr->lrh[1] & OPA_16B_SC_MASK) >> OPA_16B_SC_SHIFT);
}

static inline u32 hfi1_16B_get_dlid(struct hfi1_16b_header *hdr)
{
	return (u32)((hdr->lrh[1] & OPA_16B_LID_MASK) |
		     (((hdr->lrh[2] & OPA_16B_DLID_MASK) >>
		     OPA_16B_DLID_HIGH_SHIFT) << OPA_16B_DLID_SHIFT));
}

static inline u32 hfi1_16B_get_slid(struct hfi1_16b_header *hdr)
{
	return (u32)((hdr->lrh[0] & OPA_16B_LID_MASK) |
		     (((hdr->lrh[2] & OPA_16B_SLID_MASK) >>
		     OPA_16B_SLID_HIGH_SHIFT) << OPA_16B_SLID_SHIFT));
}

static inline u8 hfi1_16B_get_becn(struct hfi1_16b_header *hdr)
{
	return (u8)((hdr->lrh[0] & OPA_16B_BECN_MASK) >> OPA_16B_BECN_SHIFT);
}

static inline u8 hfi1_16B_get_fecn(struct hfi1_16b_header *hdr)
{
	return (u8)((hdr->lrh[1] & OPA_16B_FECN_MASK) >> OPA_16B_FECN_SHIFT);
}

static inline u8 hfi1_16B_get_l2(struct hfi1_16b_header *hdr)
{
	return (u8)((hdr->lrh[1] & OPA_16B_L2_MASK) >> OPA_16B_L2_SHIFT);
}

439 440 441 442 443
static inline u16 hfi1_16B_get_pkey(struct hfi1_16b_header *hdr)
{
	return (u16)((hdr->lrh[2] & OPA_16B_PKEY_MASK) >> OPA_16B_PKEY_SHIFT);
}

D
Don Hiatt 已提交
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
static inline u8 hfi1_16B_get_rc(struct hfi1_16b_header *hdr)
{
	return (u8)((hdr->lrh[1] & OPA_16B_RC_MASK) >> OPA_16B_RC_SHIFT);
}

static inline u8 hfi1_16B_get_age(struct hfi1_16b_header *hdr)
{
	return (u8)((hdr->lrh[3] & OPA_16B_AGE_MASK) >> OPA_16B_AGE_SHIFT);
}

static inline u16 hfi1_16B_get_len(struct hfi1_16b_header *hdr)
{
	return (u16)((hdr->lrh[0] & OPA_16B_LEN_MASK) >> OPA_16B_LEN_SHIFT);
}

static inline u16 hfi1_16B_get_entropy(struct hfi1_16b_header *hdr)
{
	return (u16)(hdr->lrh[3] & OPA_16B_ENTROPY_MASK);
}

D
Don Hiatt 已提交
464 465
#define OPA_16B_MAKE_QW(low_dw, high_dw) (((u64)(high_dw) << 32) | (low_dw))

466 467 468 469 470 471 472 473 474 475
/*
 * BTH
 */
#define OPA_16B_BTH_PAD_MASK	7
static inline u8 hfi1_16B_bth_get_pad(struct ib_other_headers *ohdr)
{
	return (u8)((be32_to_cpu(ohdr->bth[0]) >> IB_BTH_PAD_SHIFT) &
		   OPA_16B_BTH_PAD_MASK);
}

476
struct rvt_sge_state;
M
Mike Marciniszyn 已提交
477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524

/*
 * Get/Set IB link-level config parameters for f_get/set_ib_cfg()
 * Mostly for MADs that set or query link parameters, also ipath
 * config interfaces
 */
#define HFI1_IB_CFG_LIDLMC 0 /* LID (LS16b) and Mask (MS16b) */
#define HFI1_IB_CFG_LWID_DG_ENB 1 /* allowed Link-width downgrade */
#define HFI1_IB_CFG_LWID_ENB 2 /* allowed Link-width */
#define HFI1_IB_CFG_LWID 3 /* currently active Link-width */
#define HFI1_IB_CFG_SPD_ENB 4 /* allowed Link speeds */
#define HFI1_IB_CFG_SPD 5 /* current Link spd */
#define HFI1_IB_CFG_RXPOL_ENB 6 /* Auto-RX-polarity enable */
#define HFI1_IB_CFG_LREV_ENB 7 /* Auto-Lane-reversal enable */
#define HFI1_IB_CFG_LINKLATENCY 8 /* Link Latency (IB1.2 only) */
#define HFI1_IB_CFG_HRTBT 9 /* IB heartbeat off/enable/auto; DDR/QDR only */
#define HFI1_IB_CFG_OP_VLS 10 /* operational VLs */
#define HFI1_IB_CFG_VL_HIGH_CAP 11 /* num of VL high priority weights */
#define HFI1_IB_CFG_VL_LOW_CAP 12 /* num of VL low priority weights */
#define HFI1_IB_CFG_OVERRUN_THRESH 13 /* IB overrun threshold */
#define HFI1_IB_CFG_PHYERR_THRESH 14 /* IB PHY error threshold */
#define HFI1_IB_CFG_LINKDEFAULT 15 /* IB link default (sleep/poll) */
#define HFI1_IB_CFG_PKEYS 16 /* update partition keys */
#define HFI1_IB_CFG_MTU 17 /* update MTU in IBC */
#define HFI1_IB_CFG_VL_HIGH_LIMIT 19
#define HFI1_IB_CFG_PMA_TICKS 20 /* PMA sample tick resolution */
#define HFI1_IB_CFG_PORT 21 /* switch port we are connected to */

/*
 * HFI or Host Link States
 *
 * These describe the states the driver thinks the logical and physical
 * states are in.  Used as an argument to set_link_state().  Implemented
 * as bits for easy multi-state checking.  The actual state can only be
 * one.
 */
#define __HLS_UP_INIT_BP	0
#define __HLS_UP_ARMED_BP	1
#define __HLS_UP_ACTIVE_BP	2
#define __HLS_DN_DOWNDEF_BP	3	/* link down default */
#define __HLS_DN_POLL_BP	4
#define __HLS_DN_DISABLE_BP	5
#define __HLS_DN_OFFLINE_BP	6
#define __HLS_VERIFY_CAP_BP	7
#define __HLS_GOING_UP_BP	8
#define __HLS_GOING_OFFLINE_BP  9
#define __HLS_LINK_COOLDOWN_BP 10

525 526 527 528 529 530 531 532 533 534 535
#define HLS_UP_INIT	  BIT(__HLS_UP_INIT_BP)
#define HLS_UP_ARMED	  BIT(__HLS_UP_ARMED_BP)
#define HLS_UP_ACTIVE	  BIT(__HLS_UP_ACTIVE_BP)
#define HLS_DN_DOWNDEF	  BIT(__HLS_DN_DOWNDEF_BP) /* link down default */
#define HLS_DN_POLL	  BIT(__HLS_DN_POLL_BP)
#define HLS_DN_DISABLE	  BIT(__HLS_DN_DISABLE_BP)
#define HLS_DN_OFFLINE	  BIT(__HLS_DN_OFFLINE_BP)
#define HLS_VERIFY_CAP	  BIT(__HLS_VERIFY_CAP_BP)
#define HLS_GOING_UP	  BIT(__HLS_GOING_UP_BP)
#define HLS_GOING_OFFLINE BIT(__HLS_GOING_OFFLINE_BP)
#define HLS_LINK_COOLDOWN BIT(__HLS_LINK_COOLDOWN_BP)
M
Mike Marciniszyn 已提交
536 537

#define HLS_UP (HLS_UP_INIT | HLS_UP_ARMED | HLS_UP_ACTIVE)
538
#define HLS_DOWN ~(HLS_UP)
M
Mike Marciniszyn 已提交
539

540 541
#define HLS_DEFAULT HLS_DN_POLL

M
Mike Marciniszyn 已提交
542
/* use this MTU size if none other is given */
543
#define HFI1_DEFAULT_ACTIVE_MTU 10240
M
Mike Marciniszyn 已提交
544
/* use this MTU size as the default maximum */
545
#define HFI1_DEFAULT_MAX_MTU 10240
M
Mike Marciniszyn 已提交
546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
/* default partition key */
#define DEFAULT_PKEY 0xffff

/*
 * Possible fabric manager config parameters for fm_{get,set}_table()
 */
#define FM_TBL_VL_HIGH_ARB		1 /* Get/set VL high prio weights */
#define FM_TBL_VL_LOW_ARB		2 /* Get/set VL low prio weights */
#define FM_TBL_BUFFER_CONTROL		3 /* Get/set Buffer Control */
#define FM_TBL_SC2VLNT			4 /* Get/set SC->VLnt */
#define FM_TBL_VL_PREEMPT_ELEMS		5 /* Get (no set) VL preempt elems */
#define FM_TBL_VL_PREEMPT_MATRIX	6 /* Get (no set) VL preempt matrix */

/*
 * Possible "operations" for f_rcvctrl(ppd, op, ctxt)
 * these are bits so they can be combined, e.g.
 * HFI1_RCVCTRL_INTRAVAIL_ENB | HFI1_RCVCTRL_CTXT_ENB
 */
#define HFI1_RCVCTRL_TAILUPD_ENB 0x01
#define HFI1_RCVCTRL_TAILUPD_DIS 0x02
#define HFI1_RCVCTRL_CTXT_ENB 0x04
#define HFI1_RCVCTRL_CTXT_DIS 0x08
#define HFI1_RCVCTRL_INTRAVAIL_ENB 0x10
#define HFI1_RCVCTRL_INTRAVAIL_DIS 0x20
#define HFI1_RCVCTRL_PKEY_ENB 0x40  /* Note, default is enabled */
#define HFI1_RCVCTRL_PKEY_DIS 0x80
#define HFI1_RCVCTRL_TIDFLOW_ENB 0x0400
#define HFI1_RCVCTRL_TIDFLOW_DIS 0x0800
#define HFI1_RCVCTRL_ONE_PKT_EGR_ENB 0x1000
#define HFI1_RCVCTRL_ONE_PKT_EGR_DIS 0x2000
#define HFI1_RCVCTRL_NO_RHQ_DROP_ENB 0x4000
#define HFI1_RCVCTRL_NO_RHQ_DROP_DIS 0x8000
#define HFI1_RCVCTRL_NO_EGR_DROP_ENB 0x10000
#define HFI1_RCVCTRL_NO_EGR_DROP_DIS 0x20000

/* partition enforcement flags */
#define HFI1_PART_ENFORCE_IN	0x1
#define HFI1_PART_ENFORCE_OUT	0x2

/* how often we check for synthetic counter wrap around */
T
Tadeusz Struk 已提交
586
#define SYNTH_CNT_TIME 3
M
Mike Marciniszyn 已提交
587 588 589 590 591 592 593

/* Counter flags */
#define CNTR_NORMAL		0x0 /* Normal counters, just read register */
#define CNTR_SYNTH		0x1 /* Synthetic counters, saturate at all 1s */
#define CNTR_DISABLED		0x2 /* Disable this counter */
#define CNTR_32BIT		0x4 /* Simulate 64 bits for this counter */
#define CNTR_VL			0x8 /* Per VL counter */
594
#define CNTR_SDMA              0x10
M
Mike Marciniszyn 已提交
595 596 597 598 599 600 601 602
#define CNTR_INVALID_VL		-1  /* Specifies invalid VL */
#define CNTR_MODE_W		0x0
#define CNTR_MODE_R		0x1

/* VLs Supported/Operational */
#define HFI1_MIN_VLS_SUPPORTED 1
#define HFI1_MAX_VLS_SUPPORTED 8

603 604 605
#define HFI1_GUIDS_PER_PORT  5
#define HFI1_PORT_GUID_INDEX 0

M
Mike Marciniszyn 已提交
606 607 608 609 610 611 612 613 614 615 616 617 618 619
static inline void incr_cntr64(u64 *cntr)
{
	if (*cntr < (u64)-1LL)
		(*cntr)++;
}

static inline void incr_cntr32(u32 *cntr)
{
	if (*cntr < (u32)-1LL)
		(*cntr)++;
}

#define MAX_NAME_SIZE 64
struct hfi1_msix_entry {
620
	enum irq_type type;
621
	int irq;
M
Mike Marciniszyn 已提交
622
	void *arg;
623
	cpumask_t mask;
624
	struct irq_affinity_notify notify;
M
Mike Marciniszyn 已提交
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650
};

/* per-SL CCA information */
struct cca_timer {
	struct hrtimer hrtimer;
	struct hfi1_pportdata *ppd; /* read-only */
	int sl; /* read-only */
	u16 ccti; /* read/write - current value of CCTI */
};

struct link_down_reason {
	/*
	 * SMA-facing value.  Should be set from .latest when
	 * HLS_UP_* -> HLS_DN_* transition actually occurs.
	 */
	u8 sma;
	u8 latest;
};

enum {
	LO_PRIO_TABLE,
	HI_PRIO_TABLE,
	MAX_PRIO_TABLE
};

struct vl_arb_cache {
651
	/* protect vl arb cache */
M
Mike Marciniszyn 已提交
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670
	spinlock_t lock;
	struct ib_vl_weight_elem table[VL_ARB_TABLE_SIZE];
};

/*
 * The structure below encapsulates data relevant to a physical IB Port.
 * Current chips support only one such port, but the separation
 * clarifies things a bit. Note that to conform to IB conventions,
 * port-numbers are one-based. The first or only port is port1.
 */
struct hfi1_pportdata {
	struct hfi1_ibport ibport_data;

	struct hfi1_devdata *dd;
	struct kobject pport_cc_kobj;
	struct kobject sc2vl_kobj;
	struct kobject sl2sc_kobj;
	struct kobject vl2mtu_kobj;

671
	/* PHY support */
M
Mike Marciniszyn 已提交
672
	struct qsfp_data qsfp_info;
673 674 675 676 677 678 679 680 681
	/* Values for SI tuning of SerDes */
	u32 port_type;
	u32 tx_preset_eq;
	u32 tx_preset_noeq;
	u32 rx_preset;
	u8  local_atten;
	u8  remote_atten;
	u8  default_atten;
	u8  max_power_class;
M
Mike Marciniszyn 已提交
682

683 684 685
	/* did we read platform config from scratch registers? */
	bool config_from_scratch;

686 687
	/* GUIDs for this interface, in host order, guids[0] is a port guid */
	u64 guids[HFI1_GUIDS_PER_PORT];
M
Mike Marciniszyn 已提交
688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703

	/* GUID for peer interface, in host order */
	u64 neighbor_guid;

	/* up or down physical link state */
	u32 linkup;

	/*
	 * this address is mapped read-only into user processes so they can
	 * get status cheaply, whenever they want.  One qword of status per port
	 */
	u64 *statusp;

	/* SendDMA related entries */

	struct workqueue_struct *hfi1_wq;
704
	struct workqueue_struct *link_wq;
M
Mike Marciniszyn 已提交
705 706 707 708 709 710 711 712 713

	/* move out of interrupt context */
	struct work_struct link_vc_work;
	struct work_struct link_up_work;
	struct work_struct link_down_work;
	struct work_struct sma_message_work;
	struct work_struct freeze_work;
	struct work_struct link_downgrade_work;
	struct work_struct link_bounce_work;
D
Dean Luick 已提交
714
	struct delayed_work start_link_work;
M
Mike Marciniszyn 已提交
715 716 717 718 719 720 721 722 723 724 725 726 727 728
	/* host link state variables */
	struct mutex hls_lock;
	u32 host_link_state;

	/* these are the "32 bit" regs */

	u32 ibmtu; /* The MTU programmed for this unit */
	/*
	 * Current max size IB packet (in bytes) including IB headers, that
	 * we can send. Changes when ibmtu changes.
	 */
	u32 ibmaxlen;
	u32 current_egress_rate; /* units [10^6 bits/sec] */
	/* LID programmed for this instance */
729
	u32 lid;
M
Mike Marciniszyn 已提交
730 731 732 733 734 735 736 737 738 739 740 741 742 743
	/* list of pkeys programmed; 0 if not set */
	u16 pkeys[MAX_PKEY_VALUES];
	u16 link_width_supported;
	u16 link_width_downgrade_supported;
	u16 link_speed_supported;
	u16 link_width_enabled;
	u16 link_width_downgrade_enabled;
	u16 link_speed_enabled;
	u16 link_width_active;
	u16 link_width_downgrade_tx_active;
	u16 link_width_downgrade_rx_active;
	u16 link_speed_active;
	u8 vls_supported;
	u8 vls_operational;
744
	u8 actual_vls_operational;
M
Mike Marciniszyn 已提交
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763
	/* LID mask control */
	u8 lmc;
	/* Rx Polarity inversion (compensate for ~tx on partner) */
	u8 rx_pol_inv;

	u8 hw_pidx;     /* physical port index */
	u8 port;        /* IB port number and index into dd->pports - 1 */
	/* type of neighbor node */
	u8 neighbor_type;
	u8 neighbor_normal;
	u8 neighbor_fm_security; /* 1 if firmware checking is disabled */
	u8 neighbor_port_number;
	u8 is_sm_config_started;
	u8 offline_disabled_reason;
	u8 is_active_optimize_enabled;
	u8 driver_link_ready;	/* driver ready for active link */
	u8 link_enabled;	/* link enabled? */
	u8 linkinit_reason;
	u8 local_tx_rate;	/* rate given to 8051 firmware */
D
Dean Luick 已提交
764
	u8 qsfp_retry_count;
M
Mike Marciniszyn 已提交
765 766 767 768

	/* placeholders for IB MAD packet settings */
	u8 overrun_threshold;
	u8 phy_error_threshold;
769
	unsigned int is_link_down_queued;
M
Mike Marciniszyn 已提交
770

771 772 773 774 775 776 777
	/* Used to override LED behavior for things like maintenance beaconing*/
	/*
	 * Alternates per phase of blink
	 * [0] holds LED off duration, [1] holds LED on duration
	 */
	unsigned long led_override_vals[2];
	u8 led_override_phase; /* LSB picks from vals[] */
M
Mike Marciniszyn 已提交
778 779 780
	atomic_t led_override_timer_active;
	/* Used to flash LEDs in override mode */
	struct timer_list led_override_timer;
781

M
Mike Marciniszyn 已提交
782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
	u32 sm_trap_qp;
	u32 sa_qp;

	/*
	 * cca_timer_lock protects access to the per-SL cca_timer
	 * structures (specifically the ccti member).
	 */
	spinlock_t cca_timer_lock ____cacheline_aligned_in_smp;
	struct cca_timer cca_timer[OPA_MAX_SLS];

	/* List of congestion control table entries */
	struct ib_cc_table_entry_shadow ccti_entries[CC_TABLE_SHADOW_MAX];

	/* congestion entries, each entry corresponding to a SL */
	struct opa_congestion_setting_entry_shadow
		congestion_entries[OPA_MAX_SLS];

	/*
	 * cc_state_lock protects (write) access to the per-port
	 * struct cc_state.
	 */
	spinlock_t cc_state_lock ____cacheline_aligned_in_smp;

	struct cc_state __rcu *cc_state;

	/* Total number of congestion control table entries */
	u16 total_cct_entry;

	/* Bit map identifying service level */
	u32 cc_sl_control_map;

	/* CA's max number of 64 entry units in the congestion control table */
	u8 cc_max_table_entries;

816 817 818 819
	/*
	 * begin congestion log related entries
	 * cc_log_lock protects all congestion log related data
	 */
M
Mike Marciniszyn 已提交
820
	spinlock_t cc_log_lock ____cacheline_aligned_in_smp;
821
	u8 threshold_cong_event_map[OPA_MAX_SLS / 8];
M
Mike Marciniszyn 已提交
822 823 824 825 826 827 828 829 830 831 832 833
	u16 threshold_event_counter;
	struct opa_hfi1_cong_log_event_internal cc_events[OPA_CONG_LOG_ELEMS];
	int cc_log_idx; /* index for logging events */
	int cc_mad_idx; /* index for reporting events */
	/* end congestion log related entries */

	struct vl_arb_cache vl_arb_cache[MAX_PRIO_TABLE];

	/* port relative counter buffer */
	u64 *cntrs;
	/* port relative synthetic counter buffer */
	u64 *scntrs;
834
	/* port_xmit_discards are synthesized from different egress errors */
M
Mike Marciniszyn 已提交
835
	u64 port_xmit_discards;
836
	u64 port_xmit_discards_vl[C_VL_COUNT];
M
Mike Marciniszyn 已提交
837 838 839 840 841 842
	u64 port_xmit_constraint_errors;
	u64 port_rcv_constraint_errors;
	/* count of 'link_err' interrupts from DC */
	u64 link_downed;
	/* number of times link retrained successfully */
	u64 link_up;
843 844
	/* number of times a link unknown frame was reported */
	u64 unknown_frame_count;
M
Mike Marciniszyn 已提交
845 846 847 848 849 850 851 852 853 854 855 856 857
	/* port_ltp_crc_mode is returned in 'portinfo' MADs */
	u16 port_ltp_crc_mode;
	/* port_crc_mode_enabled is the crc we support */
	u8 port_crc_mode_enabled;
	/* mgmt_allowed is also returned in 'portinfo' MADs */
	u8 mgmt_allowed;
	u8 part_enforce; /* partition enforcement flags */
	struct link_down_reason local_link_down_reason;
	struct link_down_reason neigh_link_down_reason;
	/* Value to be sent to link peer on LinkDown .*/
	u8 remote_link_down_reason;
	/* Error events that will cause a port bounce. */
	u32 port_error_action;
858
	struct work_struct linkstate_active_work;
859 860
	/* Does this port need to prescan for FECNs */
	bool cc_prescan;
861 862 863 864 865 866 867
	/*
	 * Sample sendWaitCnt & sendWaitVlCnt during link transition
	 * and counter request.
	 */
	u64 port_vl_xmit_wait_last[C_VL_COUNT + 1];
	u16 prev_link_width;
	u64 vl_xmit_flit_cnt[C_VL_COUNT + 1];
M
Mike Marciniszyn 已提交
868 869 870 871 872
};

typedef int (*rhf_rcv_function_ptr)(struct hfi1_packet *packet);

typedef void (*opcode_handler)(struct hfi1_packet *packet);
D
Don Hiatt 已提交
873 874 875 876
typedef void (*hfi1_make_req)(struct rvt_qp *qp,
			      struct hfi1_pkt_state *ps,
			      struct rvt_swqe *wqe);

M
Mike Marciniszyn 已提交
877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916

/* return values for the RHF receive functions */
#define RHF_RCV_CONTINUE  0	/* keep going */
#define RHF_RCV_DONE	  1	/* stop, this packet processed */
#define RHF_RCV_REPROCESS 2	/* stop. retain this packet */

struct rcv_array_data {
	u8 group_size;
	u16 ngroups;
	u16 nctxt_extra;
};

struct per_vl_data {
	u16 mtu;
	struct send_context *sc;
};

/* 16 to directly index */
#define PER_VL_SEND_CONTEXTS 16

struct err_info_rcvport {
	u8 status_and_code;
	u64 packet_flit1;
	u64 packet_flit2;
};

struct err_info_constraint {
	u8 status;
	u16 pkey;
	u32 slid;
};

struct hfi1_temp {
	unsigned int curr;       /* current temperature */
	unsigned int lo_lim;     /* low temperature limit */
	unsigned int hi_lim;     /* high temperature limit */
	unsigned int crit_lim;   /* critical temperature limit */
	u8 triggers;      /* temperature triggers */
};

917 918 919 920 921 922 923
struct hfi1_i2c_bus {
	struct hfi1_devdata *controlling_dd; /* current controlling device */
	struct i2c_adapter adapter;	/* bus details */
	struct i2c_algo_bit_data algo;	/* bus algorithm details */
	int num;			/* bus number, 0 or 1 */
};

D
Dean Luick 已提交
924 925 926 927
/* common data between shared ASIC HFIs */
struct hfi1_asic_data {
	struct hfi1_devdata *dds[2];	/* back pointers */
	struct mutex asic_resource_mutex;
928 929
	struct hfi1_i2c_bus *i2c_bus0;
	struct hfi1_i2c_bus *i2c_bus1;
D
Dean Luick 已提交
930 931
};

932 933 934 935
/* sizes for both the QP and RSM map tables */
#define NUM_MAP_ENTRIES	 256
#define NUM_MAP_REGS      32

936 937 938 939 940 941
/*
 * Number of VNIC contexts used. Ensure it is less than or equal to
 * max queues supported by VNIC (HFI1_VNIC_MAX_QUEUE).
 */
#define HFI1_NUM_VNIC_CTXT   8

942 943 944
/* Number of VNIC RSM entries */
#define NUM_VNIC_MAP_ENTRIES 8

945 946
/* Virtual NIC information */
struct hfi1_vnic_data {
947
	struct hfi1_ctxtdata *ctxt[HFI1_NUM_VNIC_CTXT];
948
	struct kmem_cache *txreq_cache;
949
	u8 num_vports;
950
	struct idr vesw_idr;
951 952 953
	u8 rmt_start;
	u8 num_ctxt;
	u32 msix_idx;
954 955 956 957
};

struct hfi1_vnic_vport_info;

M
Mike Marciniszyn 已提交
958 959 960 961 962 963 964 965 966
/* device data struct now contains only "general per-device" info.
 * fields related to a physical IB port are in a hfi1_pportdata struct.
 */
struct sdma_engine;
struct sdma_vl_map;

#define BOARD_VERS_MAX 96 /* how long the version string can be */
#define SERIAL_MAX 16 /* length of the serial number */

967
typedef int (*send_routine)(struct rvt_qp *, struct hfi1_pkt_state *, u64);
M
Mike Marciniszyn 已提交
968 969 970 971 972 973 974 975 976 977 978 979 980
struct hfi1_devdata {
	struct hfi1_ibdev verbs_dev;     /* must be first */
	struct list_head list;
	/* pointers to related structs for this device */
	/* pci access data structure */
	struct pci_dev *pcidev;
	struct cdev user_cdev;
	struct cdev diag_cdev;
	struct cdev ui_cdev;
	struct device *user_device;
	struct device *diag_device;
	struct device *ui_device;

981 982
	/* first mapping up to RcvArray */
	u8 __iomem *kregbase1;
M
Mike Marciniszyn 已提交
983
	resource_size_t physaddr;
984 985 986 987 988 989

	/* second uncached mapping from RcvArray to pio send buffers */
	u8 __iomem *kregbase2;
	/* for detecting offset above kregbase2 address */
	u32 base2_start;

990 991
	/* Per VL data. Enough for all VLs but not all elements are set/used. */
	struct per_vl_data vld[PER_VL_SEND_CONTEXTS];
M
Mike Marciniszyn 已提交
992 993 994 995 996 997
	/* send context data */
	struct send_context_info *send_contexts;
	/* map hardware send contexts to software index */
	u8 *hw_to_sw;
	/* spinlock for allocating and releasing send context resources */
	spinlock_t sc_lock;
998 999
	/* lock for pio_map */
	spinlock_t pio_map_lock;
1000 1001 1002 1003
	/* Send Context initialization lock. */
	spinlock_t sc_init_lock;
	/* lock for sdma_map */
	spinlock_t                          sde_map_lock;
1004 1005 1006 1007
	/* array of kernel send contexts */
	struct send_context **kernel_send_context;
	/* array of vl maps */
	struct pio_vl_map __rcu *pio_map;
1008 1009
	/* default flags to last descriptor */
	u64 default_desc1;
M
Mike Marciniszyn 已提交
1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030

	/* fields common to all SDMA engines */

	volatile __le64                    *sdma_heads_dma; /* DMA'ed by chip */
	dma_addr_t                          sdma_heads_phys;
	void                               *sdma_pad_dma; /* DMA'ed by chip */
	dma_addr_t                          sdma_pad_phys;
	/* for deallocation */
	size_t                              sdma_heads_size;
	/* number from the chip */
	u32                                 chip_sdma_engines;
	/* num used */
	u32                                 num_sdma;
	/* array of engines sized by num_sdma */
	struct sdma_engine                 *per_sdma;
	/* array of vl maps */
	struct sdma_vl_map __rcu           *sdma_map;
	/* SPC freeze waitqueue and variable */
	wait_queue_head_t		  sdma_unfreeze_wq;
	atomic_t			  sdma_unfreeze_count;

1031 1032
	u32 lcb_access_count;		/* count of LCB users */

D
Dean Luick 已提交
1033 1034 1035
	/* common data between shared ASIC HFIs in this OS */
	struct hfi1_asic_data *asic_data;

M
Mike Marciniszyn 已提交
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056
	/* mem-mapped pointer to base of PIO buffers */
	void __iomem *piobase;
	/*
	 * write-combining mem-mapped pointer to base of RcvArray
	 * memory.
	 */
	void __iomem *rcvarray_wc;
	/*
	 * credit return base - a per-NUMA range of DMA address that
	 * the chip will use to update the per-context free counter
	 */
	struct credit_return_base *cr_base;

	/* send context numbers and sizes for each type */
	struct sc_config_sizes sc_sizes[SC_MAX];

	char *boardname; /* human readable board info */

	/* reset value */
	u64 z_int_counter;
	u64 z_rcv_limit;
1057
	u64 z_send_schedule;
1058

1059
	u64 __percpu *send_schedule;
1060 1061
	/* number of reserved contexts for VNIC usage */
	u16 num_vnic_contexts;
M
Mike Marciniszyn 已提交
1062 1063 1064 1065 1066 1067 1068 1069
	/* number of receive contexts in use by the driver */
	u32 num_rcv_contexts;
	/* number of pio send contexts in use by the driver */
	u32 num_send_contexts;
	/*
	 * number of ctxts available for PSM open
	 */
	u32 freectxts;
1070 1071
	/* total number of available user/PSM contexts */
	u32 num_user_contexts;
M
Mike Marciniszyn 已提交
1072 1073 1074
	/* base receive interrupt timeout, in CSR units */
	u32 rcv_intr_timeout_csr;

1075
	u32 freezelen; /* max length of freezemsg */
M
Mike Marciniszyn 已提交
1076 1077 1078
	u64 __iomem *egrtidbase;
	spinlock_t sendctrl_lock; /* protect changes to SendCtrl */
	spinlock_t rcvctrl_lock; /* protect changes to RcvCtrl */
1079
	spinlock_t uctxt_lock; /* protect rcd changes */
T
Tadeusz Struk 已提交
1080 1081 1082
	struct mutex dc8051_lock; /* exclusive access to 8051 */
	struct workqueue_struct *update_cntr_wq;
	struct work_struct update_cntr_work;
M
Mike Marciniszyn 已提交
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
	/* exclusive access to 8051 memory */
	spinlock_t dc8051_memlock;
	int dc8051_timed_out;	/* remember if the 8051 timed out */
	/*
	 * A page that will hold event notification bitmaps for all
	 * contexts. This page will be mapped into all processes.
	 */
	unsigned long *events;
	/*
	 * per unit status, see also portdata statusp
	 * mapped read-only into user processes so they can get unit and
	 * IB link status cheaply
	 */
	struct hfi1_status *status;

	/* revision register shadow */
	u64 revision;
	/* Base GUID for device (network order) */
	u64 base_guid;

	/* these are the "32 bit" regs */

	/* value we put in kr_rcvhdrsize */
	u32 rcvhdrsize;
	/* number of receive contexts the chip supports */
	u32 chip_rcv_contexts;
	/* number of receive array entries */
	u32 chip_rcv_array_count;
	/* number of PIO send contexts the chip supports */
	u32 chip_send_contexts;
	/* number of bytes in the PIO memory buffer */
	u32 chip_pio_mem_size;
	/* number of bytes in the SDMA memory buffer */
	u32 chip_sdma_mem_size;

	/* size of each rcvegrbuffer */
	u32 rcvegrbufsize;
	/* log2 of above */
	u16 rcvegrbufsize_shift;
	/* both sides of the PCIe link are gen3 capable */
	u8 link_gen3_capable;
1124
	u8 dc_shutdown;
M
Mike Marciniszyn 已提交
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
	/* localbus width (1, 2,4,8,16,32) from config space  */
	u32 lbus_width;
	/* localbus speed in MHz */
	u32 lbus_speed;
	int unit; /* unit # of this chip */
	int node; /* home node of this chip */

	/* save these PCI fields to restore after a reset */
	u32 pcibar0;
	u32 pcibar1;
	u32 pci_rom;
	u16 pci_command;
	u16 pcie_devctl;
	u16 pcie_lnkctl;
	u16 pcie_devctl2;
	u32 pci_msix0;
	u32 pci_tph2;

	/*
	 * ASCII serial number, from flash, large enough for original
	 * all digit strings, and longer serial number format
	 */
	u8 serial[SERIAL_MAX];
	/* human readable board version */
	u8 boardversion[BOARD_VERS_MAX];
	u8 lbus_info[32]; /* human readable localbus info */
	/* chip major rev, from CceRevision */
	u8 majrev;
	/* chip minor rev, from CceRevision */
	u8 minrev;
	/* hardware ID */
	u8 hfi1_id;
	/* implementation code */
	u8 icode;
	/* vAU of this device */
	u8 vau;
	/* vCU of this device */
	u8 vcu;
	/* link credits of this device */
	u16 link_credits;
	/* initial vl15 credits to use */
	u16 vl15_init;

1168 1169 1170 1171 1172 1173 1174 1175
	/*
	 * Cached value for vl15buf, read during verify cap interrupt. VL15
	 * credits are to be kept at 0 and set when handling the link-up
	 * interrupt. This removes the possibility of receiving VL15 MAD
	 * packets before this HFI is ready.
	 */
	u16 vl15buf_cached;

M
Mike Marciniszyn 已提交
1176 1177 1178 1179 1180
	/* Misc small ints */
	u8 n_krcv_queues;
	u8 qos_shift;

	u16 irev;	/* implementation revision */
1181
	u32 dc8051_ver; /* 8051 firmware version */
M
Mike Marciniszyn 已提交
1182

1183
	spinlock_t hfi1_diag_trans_lock; /* protect diag observer ops */
1184
	struct platform_config platform_config;
M
Mike Marciniszyn 已提交
1185 1186 1187 1188 1189 1190 1191
	struct platform_config_cache pcfg_cache;

	struct diag_client *diag_client;

	/* MSI-X information */
	struct hfi1_msix_entry *msix_entries;
	u32 num_msix_entries;
1192
	u32 first_dyn_msix_idx;
M
Mike Marciniszyn 已提交
1193 1194 1195 1196 1197 1198 1199 1200 1201

	/* INTx information */
	u32 requested_intx_irq;		/* did we request one? */

	/* general interrupt: mask of handled interrupts */
	u64 gi_mask[CCE_NUM_INT_CSRS];

	struct rcv_array_data rcv_entries;

1202 1203 1204
	/* cycle length of PS* counters in HW (in picoseconds) */
	u16 psxmitwait_check_rate;

M
Mike Marciniszyn 已提交
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
	/*
	 * 64 bit synthetic counters
	 */
	struct timer_list synth_stats_timer;

	/*
	 * device counters
	 */
	char *cntrnames;
	size_t cntrnameslen;
	size_t ndevcntrs;
	u64 *cntrs;
	u64 *scntrs;

	/*
	 * remembered values for synthetic counters
	 */
	u64 last_tx;
	u64 last_rx;

	/*
	 * per-port counters
	 */
	size_t nportcntrs;
	char *portcntrnames;
	size_t portcntrnameslen;

	struct err_info_rcvport err_info_rcvport;
	struct err_info_constraint err_info_rcv_constraint;
	struct err_info_constraint err_info_xmit_constraint;

	atomic_t drop_packet;
	u8 do_drop;
1238 1239
	u8 err_info_uncorrectable;
	u8 err_info_fmconfig;
M
Mike Marciniszyn 已提交
1240

1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
	/*
	 * Software counters for the status bits defined by the
	 * associated error status registers
	 */
	u64 cce_err_status_cnt[NUM_CCE_ERR_STATUS_COUNTERS];
	u64 rcv_err_status_cnt[NUM_RCV_ERR_STATUS_COUNTERS];
	u64 misc_err_status_cnt[NUM_MISC_ERR_STATUS_COUNTERS];
	u64 send_pio_err_status_cnt[NUM_SEND_PIO_ERR_STATUS_COUNTERS];
	u64 send_dma_err_status_cnt[NUM_SEND_DMA_ERR_STATUS_COUNTERS];
	u64 send_egress_err_status_cnt[NUM_SEND_EGRESS_ERR_STATUS_COUNTERS];
	u64 send_err_status_cnt[NUM_SEND_ERR_STATUS_COUNTERS];

	/* Software counter that spans all contexts */
	u64 sw_ctxt_err_status_cnt[NUM_SEND_CTXT_ERR_STATUS_COUNTERS];
	/* Software counter that spans all DMA engines */
	u64 sw_send_dma_eng_err_status_cnt[
		NUM_SEND_DMA_ENG_ERR_STATUS_COUNTERS];
	/* Software counter that aggregates all cce_err_status errors */
	u64 sw_cce_err_status_aggregate;
1260 1261
	/* Software counter that aggregates all bypass packet rcv errors */
	u64 sw_rcv_bypass_packet_errors;
1262
	/* receive interrupt function */
M
Mike Marciniszyn 已提交
1263 1264
	rhf_rcv_function_ptr normal_rhf_rcv_functions[8];

1265 1266 1267
	/* Save the enabled LCB error bits */
	u64 lcb_err_en;

M
Mike Marciniszyn 已提交
1268
	/*
1269 1270
	 * Capability to have different send engines simply by changing a
	 * pointer value.
M
Mike Marciniszyn 已提交
1271
	 */
1272
	send_routine process_pio_send ____cacheline_aligned_in_smp;
1273
	send_routine process_dma_send;
M
Mike Marciniszyn 已提交
1274 1275
	void (*pio_inline_send)(struct hfi1_devdata *dd, struct pio_buf *pbuf,
				u64 pbc, const void *from, size_t count);
1276 1277 1278
	int (*process_vnic_dma_send)(struct hfi1_devdata *dd, u8 q_idx,
				     struct hfi1_vnic_vport_info *vinfo,
				     struct sk_buff *skb, u64 pbc, u8 plen);
1279 1280 1281 1282 1283 1284 1285
	/* hfi1_pportdata, points to array of (physical) port-specific
	 * data structs, indexed by pidx (0..n-1)
	 */
	struct hfi1_pportdata *pport;
	/* receive context data */
	struct hfi1_ctxtdata **rcd;
	u64 __percpu *int_counter;
1286 1287
	/* verbs tx opcode stats */
	struct hfi1_opcode_stats_perctx __percpu *tx_opstats;
1288 1289 1290 1291
	/* device (not port) flags, basically device capabilities */
	u16 flags;
	/* Number of physical ports available */
	u8 num_pports;
1292 1293
	/* Lowest context number which can be used by user processes or VNIC */
	u8 first_dyn_alloc_ctxt;
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
	/* adding a new field here would make it part of this cacheline */

	/* seqlock for sc2vl */
	seqlock_t sc2vl_lock ____cacheline_aligned_in_smp;
	u64 sc2vl[4];
	/* receive interrupt functions */
	rhf_rcv_function_ptr *rhf_rcv_function_map;
	u64 __percpu *rcv_limit;
	u16 rhf_offset; /* offset of RHF within receive header entry */
	/* adding a new field here would make it part of this cacheline */
M
Mike Marciniszyn 已提交
1304 1305 1306 1307 1308

	/* OUI comes from the HW. Used everywhere as 3 separate bytes. */
	u8 oui1;
	u8 oui2;
	u8 oui3;
1309

M
Mike Marciniszyn 已提交
1310 1311 1312 1313 1314
	/* Timer and counter used to detect RcvBufOvflCnt changes */
	struct timer_list rcverr_timer;

	wait_queue_head_t event_queue;

1315 1316
	/* receive context tail dummy address */
	__le64 *rcvhdrtail_dummy_kvaddr;
1317
	dma_addr_t rcvhdrtail_dummy_dma;
1318

1319
	u32 rcv_ovfl_cnt;
1320 1321 1322 1323
	/* Serialize ASPM enable/disable between multiple verbs contexts */
	spinlock_t aspm_lock;
	/* Number of verbs contexts which have disabled ASPM */
	atomic_t aspm_disabled_cnt;
1324 1325 1326 1327
	/* Keeps track of user space clients */
	atomic_t user_refcount;
	/* Used to wait for outstanding user space clients before dev removal */
	struct completion user_comp;
1328

1329 1330 1331
	bool eprom_available;	/* true if EPROM is available for this device */
	bool aspm_supported;	/* Does HW support ASPM */
	bool aspm_enabled;	/* ASPM state: enabled/disabled */
1332
	struct rhashtable *sdma_rht;
1333

1334
	struct kobject kobj;
1335 1336 1337

	/* vnic data */
	struct hfi1_vnic_data vnic;
M
Mike Marciniszyn 已提交
1338 1339
};

1340 1341 1342 1343 1344
static inline bool hfi1_vnic_is_rsm_full(struct hfi1_devdata *dd, int spare)
{
	return (dd->vnic.rmt_start + spare) > NUM_MAP_ENTRIES;
}

M
Mike Marciniszyn 已提交
1345
/* 8051 firmware version helper */
1346 1347 1348 1349
#define dc8051_ver(a, b, c) ((a) << 16 | (b) << 8 | (c))
#define dc8051_ver_maj(a) (((a) & 0xff0000) >> 16)
#define dc8051_ver_min(a) (((a) & 0x00ff00) >> 8)
#define dc8051_ver_patch(a) ((a) & 0x0000ff)
M
Mike Marciniszyn 已提交
1350 1351

/* f_put_tid types */
1352 1353 1354 1355
#define PT_EXPECTED       0
#define PT_EAGER          1
#define PT_INVALID_FLUSH  2
#define PT_INVALID        3
M
Mike Marciniszyn 已提交
1356

1357
struct tid_rb_node;
1358
struct mmu_rb_node;
1359
struct mmu_rb_handler;
1360

M
Mike Marciniszyn 已提交
1361 1362
/* Private data for file operations */
struct hfi1_filedata {
1363
	struct hfi1_devdata *dd;
M
Mike Marciniszyn 已提交
1364 1365 1366
	struct hfi1_ctxtdata *uctxt;
	struct hfi1_user_sdma_comp_q *cq;
	struct hfi1_user_sdma_pkt_q *pq;
1367
	u16 subctxt;
M
Mike Marciniszyn 已提交
1368 1369
	/* for cpu affinity; -1 if none */
	int rec_cpu_num;
1370
	u32 tid_n_pinned;
1371
	struct mmu_rb_handler *handler;
1372
	struct tid_rb_node **entry_to_rb;
1373 1374 1375 1376 1377
	spinlock_t tid_lock; /* protect tid_[limit,used] counters */
	u32 tid_limit;
	u32 tid_used;
	u32 *invalid_tids;
	u32 invalid_tid_idx;
1378 1379
	/* protect invalid_tids array and invalid_tid_idx */
	spinlock_t invalid_lock;
I
Ira Weiny 已提交
1380
	struct mm_struct *mm;
M
Mike Marciniszyn 已提交
1381 1382 1383 1384 1385 1386
};

extern struct list_head hfi1_dev_list;
extern spinlock_t hfi1_devs_lock;
struct hfi1_devdata *hfi1_lookup(int unit);

1387 1388 1389 1390 1391 1392
static inline unsigned long uctxt_offset(struct hfi1_ctxtdata *uctxt)
{
	return (uctxt->ctxt - uctxt->dd->first_dyn_alloc_ctxt) *
		HFI1_MAX_SHARED_CTXTS;
}

1393
int hfi1_init(struct hfi1_devdata *dd, int reinit);
M
Mike Marciniszyn 已提交
1394 1395
int hfi1_count_active_units(void);

1396 1397
int hfi1_diag_add(struct hfi1_devdata *dd);
void hfi1_diag_remove(struct hfi1_devdata *dd);
M
Mike Marciniszyn 已提交
1398 1399 1400 1401
void handle_linkup_change(struct hfi1_devdata *dd, u32 linkup);

void handle_user_interrupt(struct hfi1_ctxtdata *rcd);

1402 1403
int hfi1_create_rcvhdrq(struct hfi1_devdata *dd, struct hfi1_ctxtdata *rcd);
int hfi1_setup_eagerbufs(struct hfi1_ctxtdata *rcd);
1404 1405 1406
int hfi1_create_kctxts(struct hfi1_devdata *dd);
int hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, int numa,
			 struct hfi1_ctxtdata **rcd);
1407
void hfi1_free_ctxt(struct hfi1_ctxtdata *rcd);
1408 1409 1410
void hfi1_init_pportdata(struct pci_dev *pdev, struct hfi1_pportdata *ppd,
			 struct hfi1_devdata *dd, u8 hw_pidx, u8 port);
void hfi1_free_ctxtdata(struct hfi1_devdata *dd, struct hfi1_ctxtdata *rcd);
1411 1412
int hfi1_rcd_put(struct hfi1_ctxtdata *rcd);
void hfi1_rcd_get(struct hfi1_ctxtdata *rcd);
1413 1414
struct hfi1_ctxtdata *hfi1_rcd_get_by_index_safe(struct hfi1_devdata *dd,
						 u16 ctxt);
1415
struct hfi1_ctxtdata *hfi1_rcd_get_by_index(struct hfi1_devdata *dd, u16 ctxt);
1416 1417 1418
int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread);
int handle_receive_interrupt_nodma_rtail(struct hfi1_ctxtdata *rcd, int thread);
int handle_receive_interrupt_dma_rtail(struct hfi1_ctxtdata *rcd, int thread);
1419
void set_all_slowpath(struct hfi1_devdata *dd);
1420 1421 1422
void hfi1_vnic_synchronize_irq(struct hfi1_devdata *dd);
void hfi1_set_vnic_msix_info(struct hfi1_ctxtdata *rcd);
void hfi1_reset_vnic_msix_info(struct hfi1_ctxtdata *rcd);
1423

1424
extern const struct pci_device_id hfi1_pci_tbl[];
D
Don Hiatt 已提交
1425 1426 1427 1428 1429 1430 1431
void hfi1_make_ud_req_9B(struct rvt_qp *qp,
			 struct hfi1_pkt_state *ps,
			 struct rvt_swqe *wqe);

void hfi1_make_ud_req_16B(struct rvt_qp *qp,
			  struct hfi1_pkt_state *ps,
			  struct rvt_swqe *wqe);
1432

1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443
/* receive packet handler dispositions */
#define RCV_PKT_OK      0x0 /* keep going */
#define RCV_PKT_LIMIT   0x1 /* stop, hit limit, start thread */
#define RCV_PKT_DONE    0x2 /* stop, no more packets detected */

/* calculate the current RHF address */
static inline __le32 *get_rhf_addr(struct hfi1_ctxtdata *rcd)
{
	return (__le32 *)rcd->rcvhdrq + rcd->head + rcd->dd->rhf_offset;
}

M
Mike Marciniszyn 已提交
1444 1445
int hfi1_reset_device(int);

1446 1447 1448
void receive_interrupt_work(struct work_struct *work);

/* extract service channel from header and rhf */
1449
static inline int hfi1_9B_get_sc5(struct ib_header *hdr, u64 rhf)
1450
{
1451
	return ib_get_sc(hdr) | ((!!(rhf_dc_info(rhf))) << 4);
1452 1453
}

1454 1455 1456 1457 1458 1459 1460 1461 1462 1463
#define HFI1_JKEY_WIDTH       16
#define HFI1_JKEY_MASK        (BIT(16) - 1)
#define HFI1_ADMIN_JKEY_RANGE 32

/*
 * J_KEYs are split and allocated in the following groups:
 *   0 - 31    - users with administrator privileges
 *  32 - 63    - kernel protocols using KDETH packets
 *  64 - 65535 - all other users using KDETH packets
 */
M
Mike Marciniszyn 已提交
1464 1465
static inline u16 generate_jkey(kuid_t uid)
{
1466 1467 1468 1469 1470 1471 1472 1473
	u16 jkey = from_kuid(current_user_ns(), uid) & HFI1_JKEY_MASK;

	if (capable(CAP_SYS_ADMIN))
		jkey &= HFI1_ADMIN_JKEY_RANGE - 1;
	else if (jkey < 64)
		jkey |= BIT(HFI1_JKEY_WIDTH - 1);

	return jkey;
M
Mike Marciniszyn 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
}

/*
 * active_egress_rate
 *
 * returns the active egress rate in units of [10^6 bits/sec]
 */
static inline u32 active_egress_rate(struct hfi1_pportdata *ppd)
{
	u16 link_speed = ppd->link_speed_active;
	u16 link_width = ppd->link_width_active;
	u32 egress_rate;

	if (link_speed == OPA_LINK_SPEED_25G)
		egress_rate = 25000;
	else /* assume OPA_LINK_SPEED_12_5G */
		egress_rate = 12500;

	switch (link_width) {
	case OPA_LINK_WIDTH_4X:
		egress_rate *= 4;
		break;
	case OPA_LINK_WIDTH_3X:
		egress_rate *= 3;
		break;
	case OPA_LINK_WIDTH_2X:
		egress_rate *= 2;
		break;
	default:
		/* assume IB_WIDTH_1X */
		break;
	}

	return egress_rate;
}

/*
 * egress_cycles
 *
 * Returns the number of 'fabric clock cycles' to egress a packet
 * of length 'len' bytes, at 'rate' Mbit/s. Since the fabric clock
 * rate is (approximately) 805 MHz, the units of the returned value
 * are (1/805 MHz).
 */
static inline u32 egress_cycles(u32 len, u32 rate)
{
	u32 cycles;

	/*
	 * cycles is:
	 *
	 *          (length) [bits] / (rate) [bits/sec]
	 *  ---------------------------------------------------
	 *  fabric_clock_period == 1 /(805 * 10^6) [cycles/sec]
	 */

	cycles = len * 8; /* bits */
	cycles *= 805;
	cycles /= rate;

	return cycles;
}

void set_link_ipg(struct hfi1_pportdata *ppd);
D
Don Hiatt 已提交
1538
void process_becn(struct hfi1_pportdata *ppd, u8 sl, u32 rlid, u32 lqpn,
M
Mike Marciniszyn 已提交
1539
		  u32 rqpn, u8 svc_type);
1540
void return_cnp(struct hfi1_ibport *ibp, struct rvt_qp *qp, u32 remote_qpn,
M
Mike Marciniszyn 已提交
1541 1542
		u32 pkey, u32 slid, u32 dlid, u8 sc5,
		const struct ib_grh *old_grh);
D
Don Hiatt 已提交
1543 1544 1545 1546 1547 1548 1549
void return_cnp_16B(struct hfi1_ibport *ibp, struct rvt_qp *qp,
		    u32 remote_qpn, u32 pkey, u32 slid, u32 dlid,
		    u8 sc5, const struct ib_grh *old_grh);
typedef void (*hfi1_handle_cnp)(struct hfi1_ibport *ibp, struct rvt_qp *qp,
				u32 remote_qpn, u32 pkey, u32 slid, u32 dlid,
				u8 sc5, const struct ib_grh *old_grh);

1550
#define PKEY_CHECK_INVALID -1
1551
int egress_pkey_check(struct hfi1_pportdata *ppd, u32 slid, u16 pkey,
1552
		      u8 sc5, int8_t s_pkey_index);
M
Mike Marciniszyn 已提交
1553 1554 1555 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 1583 1584 1585 1586 1587 1588 1589 1590 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 1632

#define PACKET_EGRESS_TIMEOUT 350
static inline void pause_for_credit_return(struct hfi1_devdata *dd)
{
	/* Pause at least 1us, to ensure chip returns all credits */
	u32 usec = cclock_to_ns(dd, PACKET_EGRESS_TIMEOUT) / 1000;

	udelay(usec ? usec : 1);
}

/**
 * sc_to_vlt() reverse lookup sc to vl
 * @dd - devdata
 * @sc5 - 5 bit sc
 */
static inline u8 sc_to_vlt(struct hfi1_devdata *dd, u8 sc5)
{
	unsigned seq;
	u8 rval;

	if (sc5 >= OPA_MAX_SCS)
		return (u8)(0xff);

	do {
		seq = read_seqbegin(&dd->sc2vl_lock);
		rval = *(((u8 *)dd->sc2vl) + sc5);
	} while (read_seqretry(&dd->sc2vl_lock, seq));

	return rval;
}

#define PKEY_MEMBER_MASK 0x8000
#define PKEY_LOW_15_MASK 0x7fff

/*
 * ingress_pkey_matches_entry - return 1 if the pkey matches ent (ent
 * being an entry from the ingress partition key table), return 0
 * otherwise. Use the matching criteria for ingress partition keys
 * specified in the OPAv1 spec., section 9.10.14.
 */
static inline int ingress_pkey_matches_entry(u16 pkey, u16 ent)
{
	u16 mkey = pkey & PKEY_LOW_15_MASK;
	u16 ment = ent & PKEY_LOW_15_MASK;

	if (mkey == ment) {
		/*
		 * If pkey[15] is clear (limited partition member),
		 * is bit 15 in the corresponding table element
		 * clear (limited member)?
		 */
		if (!(pkey & PKEY_MEMBER_MASK))
			return !!(ent & PKEY_MEMBER_MASK);
		return 1;
	}
	return 0;
}

/*
 * ingress_pkey_table_search - search the entire pkey table for
 * an entry which matches 'pkey'. return 0 if a match is found,
 * and 1 otherwise.
 */
static int ingress_pkey_table_search(struct hfi1_pportdata *ppd, u16 pkey)
{
	int i;

	for (i = 0; i < MAX_PKEY_VALUES; i++) {
		if (ingress_pkey_matches_entry(pkey, ppd->pkeys[i]))
			return 0;
	}
	return 1;
}

/*
 * ingress_pkey_table_fail - record a failure of ingress pkey validation,
 * i.e., increment port_rcv_constraint_errors for the port, and record
 * the 'error info' for this failure.
 */
static void ingress_pkey_table_fail(struct hfi1_pportdata *ppd, u16 pkey,
1633
				    u32 slid)
M
Mike Marciniszyn 已提交
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
{
	struct hfi1_devdata *dd = ppd->dd;

	incr_cntr64(&ppd->port_rcv_constraint_errors);
	if (!(dd->err_info_rcv_constraint.status & OPA_EI_STATUS_SMASK)) {
		dd->err_info_rcv_constraint.status |= OPA_EI_STATUS_SMASK;
		dd->err_info_rcv_constraint.slid = slid;
		dd->err_info_rcv_constraint.pkey = pkey;
	}
}

/*
 * ingress_pkey_check - Return 0 if the ingress pkey is valid, return 1
 * otherwise. Use the criteria in the OPAv1 spec, section 9.10.14. idx
 * is a hint as to the best place in the partition key table to begin
 * searching. This function should not be called on the data path because
 * of performance reasons. On datapath pkey check is expected to be done
 * by HW and rcv_pkey_check function should be called instead.
 */
static inline int ingress_pkey_check(struct hfi1_pportdata *ppd, u16 pkey,
1654
				     u8 sc5, u8 idx, u32 slid, bool force)
M
Mike Marciniszyn 已提交
1655
{
1656
	if (!(force) && !(ppd->part_enforce & HFI1_PART_ENFORCE_IN))
M
Mike Marciniszyn 已提交
1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
		return 0;

	/* If SC15, pkey[0:14] must be 0x7fff */
	if ((sc5 == 0xf) && ((pkey & PKEY_LOW_15_MASK) != PKEY_LOW_15_MASK))
		goto bad;

	/* Is the pkey = 0x0, or 0x8000? */
	if ((pkey & PKEY_LOW_15_MASK) == 0)
		goto bad;

	/* The most likely matching pkey has index 'idx' */
	if (ingress_pkey_matches_entry(pkey, ppd->pkeys[idx]))
		return 0;

	/* no match - try the whole table */
	if (!ingress_pkey_table_search(ppd, pkey))
		return 0;

bad:
	ingress_pkey_table_fail(ppd, pkey, slid);
	return 1;
}

/*
 * rcv_pkey_check - Return 0 if the ingress pkey is valid, return 1
 * otherwise. It only ensures pkey is vlid for QP0. This function
 * should be called on the data path instead of ingress_pkey_check
 * as on data path, pkey check is done by HW (except for QP0).
 */
static inline int rcv_pkey_check(struct hfi1_pportdata *ppd, u16 pkey,
				 u8 sc5, u16 slid)
{
	if (!(ppd->part_enforce & HFI1_PART_ENFORCE_IN))
		return 0;

	/* If SC15, pkey[0:14] must be 0x7fff */
	if ((sc5 == 0xf) && ((pkey & PKEY_LOW_15_MASK) != PKEY_LOW_15_MASK))
		goto bad;

	return 0;
bad:
	ingress_pkey_table_fail(ppd, pkey, slid);
	return 1;
}

/* MTU handling */

/* MTU enumeration, 256-4k match IB */
#define OPA_MTU_0     0
#define OPA_MTU_256   1
#define OPA_MTU_512   2
#define OPA_MTU_1024  3
#define OPA_MTU_2048  4
#define OPA_MTU_4096  5

u32 lrh_max_header_bytes(struct hfi1_devdata *dd);
int mtu_to_enum(u32 mtu, int default_if_bad);
1714
u16 enum_to_mtu(int mtu);
M
Mike Marciniszyn 已提交
1715 1716 1717 1718 1719 1720
static inline int valid_ib_mtu(unsigned int mtu)
{
	return mtu == 256 || mtu == 512 ||
		mtu == 1024 || mtu == 2048 ||
		mtu == 4096;
}
1721

M
Mike Marciniszyn 已提交
1722 1723 1724 1725 1726 1727
static inline int valid_opa_max_mtu(unsigned int mtu)
{
	return mtu >= 2048 &&
		(valid_ib_mtu(mtu) || mtu == 8192 || mtu == 10240);
}

1728
int set_mtu(struct hfi1_pportdata *ppd);
M
Mike Marciniszyn 已提交
1729

1730 1731 1732 1733
int hfi1_set_lid(struct hfi1_pportdata *ppd, u32 lid, u8 lmc);
void hfi1_disable_after_error(struct hfi1_devdata *dd);
int hfi1_set_uevent_bits(struct hfi1_pportdata *ppd, const int evtbit);
int hfi1_rcvbuf_validate(u32 size, u8 type, u16 *encode);
M
Mike Marciniszyn 已提交
1734

1735 1736
int fm_get_table(struct hfi1_pportdata *ppd, int which, void *t);
int fm_set_table(struct hfi1_pportdata *ppd, int which, void *t);
M
Mike Marciniszyn 已提交
1737

1738 1739
void set_up_vau(struct hfi1_devdata *dd, u8 vau);
void set_up_vl15(struct hfi1_devdata *dd, u16 vl15buf);
M
Mike Marciniszyn 已提交
1740 1741 1742
void reset_link_credits(struct hfi1_devdata *dd);
void assign_remote_cm_au_table(struct hfi1_devdata *dd, u8 vcu);

1743
int set_buffer_control(struct hfi1_pportdata *ppd, struct buffer_control *bc);
M
Mike Marciniszyn 已提交
1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764

static inline struct hfi1_devdata *dd_from_ppd(struct hfi1_pportdata *ppd)
{
	return ppd->dd;
}

static inline struct hfi1_devdata *dd_from_dev(struct hfi1_ibdev *dev)
{
	return container_of(dev, struct hfi1_devdata, verbs_dev);
}

static inline struct hfi1_devdata *dd_from_ibdev(struct ib_device *ibdev)
{
	return dd_from_dev(to_idev(ibdev));
}

static inline struct hfi1_pportdata *ppd_from_ibp(struct hfi1_ibport *ibp)
{
	return container_of(ibp, struct hfi1_pportdata, ibport_data);
}

1765 1766 1767 1768 1769
static inline struct hfi1_ibdev *dev_from_rdi(struct rvt_dev_info *rdi)
{
	return container_of(rdi, struct hfi1_ibdev, rdi);
}

M
Mike Marciniszyn 已提交
1770 1771 1772 1773 1774 1775 1776 1777 1778
static inline struct hfi1_ibport *to_iport(struct ib_device *ibdev, u8 port)
{
	struct hfi1_devdata *dd = dd_from_ibdev(ibdev);
	unsigned pidx = port - 1; /* IB number port from 1, hdw from 0 */

	WARN_ON(pidx >= dd->num_pports);
	return &dd->pport[pidx].ibport_data;
}

1779 1780 1781 1782 1783
static inline struct hfi1_ibport *rcd_to_iport(struct hfi1_ctxtdata *rcd)
{
	return &rcd->ppd->ibport_data;
}

1784 1785 1786 1787 1788
void hfi1_process_ecn_slowpath(struct rvt_qp *qp, struct hfi1_packet *pkt,
			       bool do_cnp);
static inline bool process_ecn(struct rvt_qp *qp, struct hfi1_packet *pkt,
			       bool do_cnp)
{
1789 1790
	bool becn;
	bool fecn;
D
Don Hiatt 已提交
1791 1792 1793 1794 1795

	if (pkt->etype == RHF_RCV_TYPE_BYPASS) {
		fecn = hfi1_16B_get_fecn(pkt->hdr);
		becn = hfi1_16B_get_becn(pkt->hdr);
	} else {
1796 1797
		fecn = ib_bth_get_fecn(pkt->ohdr);
		becn = ib_bth_get_becn(pkt->ohdr);
D
Don Hiatt 已提交
1798 1799
	}
	if (unlikely(fecn || becn)) {
1800
		hfi1_process_ecn_slowpath(qp, pkt, do_cnp);
D
Don Hiatt 已提交
1801
		return fecn;
1802 1803 1804 1805
	}
	return false;
}

M
Mike Marciniszyn 已提交
1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821
/*
 * Return the indexed PKEY from the port PKEY table.
 */
static inline u16 hfi1_get_pkey(struct hfi1_ibport *ibp, unsigned index)
{
	struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);
	u16 ret;

	if (index >= ARRAY_SIZE(ppd->pkeys))
		ret = 0;
	else
		ret = ppd->pkeys[index];

	return ret;
}

1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832
/*
 * Return the indexed GUID from the port GUIDs table.
 */
static inline __be64 get_sguid(struct hfi1_ibport *ibp, unsigned int index)
{
	struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);

	WARN_ON(index >= HFI1_GUIDS_PER_PORT);
	return cpu_to_be64(ppd->guids[index]);
}

M
Mike Marciniszyn 已提交
1833
/*
1834
 * Called by readers of cc_state only, must call under rcu_read_lock().
M
Mike Marciniszyn 已提交
1835 1836 1837 1838 1839 1840
 */
static inline struct cc_state *get_cc_state(struct hfi1_pportdata *ppd)
{
	return rcu_dereference(ppd->cc_state);
}

1841 1842 1843 1844 1845 1846 1847 1848 1849 1850
/*
 * Called by writers of cc_state only,  must call under cc_state_lock.
 */
static inline
struct cc_state *get_cc_state_protected(struct hfi1_pportdata *ppd)
{
	return rcu_dereference_protected(ppd->cc_state,
					 lockdep_is_held(&ppd->cc_state_lock));
}

M
Mike Marciniszyn 已提交
1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864
/*
 * values for dd->flags (_device_ related flags)
 */
#define HFI1_INITTED           0x1    /* chip and driver up and initted */
#define HFI1_PRESENT           0x2    /* chip accesses can be done */
#define HFI1_FROZEN            0x4    /* chip in SPC freeze */
#define HFI1_HAS_SDMA_TIMEOUT  0x8
#define HFI1_HAS_SEND_DMA      0x10   /* Supports Send DMA */
#define HFI1_FORCED_FREEZE     0x80   /* driver forced freeze mode */

/* IB dword length mask in PBC (lower 11 bits); same for all chips */
#define HFI1_PBC_LENGTH_MASK                     ((1 << 11) - 1)

/* ctxt_flag bit offsets */
1865 1866 1867 1868
		/* base context has not finished initializing */
#define HFI1_CTXT_BASE_UNINIT 1
		/* base context initaliation failed */
#define HFI1_CTXT_BASE_FAILED 2
M
Mike Marciniszyn 已提交
1869
		/* waiting for a packet to arrive */
1870
#define HFI1_CTXT_WAITING_RCV 3
M
Mike Marciniszyn 已提交
1871
		/* waiting for an urgent packet to arrive */
1872
#define HFI1_CTXT_WAITING_URG 4
M
Mike Marciniszyn 已提交
1873 1874

/* free up any allocated data at closes */
1875 1876 1877
struct hfi1_devdata *hfi1_init_dd(struct pci_dev *pdev,
				  const struct pci_device_id *ent);
void hfi1_free_devdata(struct hfi1_devdata *dd);
M
Mike Marciniszyn 已提交
1878 1879
struct hfi1_devdata *hfi1_alloc_devdata(struct pci_dev *pdev, size_t extra);

1880 1881 1882
/* LED beaconing functions */
void hfi1_start_led_override(struct hfi1_pportdata *ppd, unsigned int timeon,
			     unsigned int timeoff);
1883
void shutdown_led_override(struct hfi1_pportdata *ppd);
M
Mike Marciniszyn 已提交
1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924

#define HFI1_CREDIT_RETURN_RATE (100)

/*
 * The number of words for the KDETH protocol field.  If this is
 * larger then the actual field used, then part of the payload
 * will be in the header.
 *
 * Optimally, we want this sized so that a typical case will
 * use full cache lines.  The typical local KDETH header would
 * be:
 *
 *	Bytes	Field
 *	  8	LRH
 *	 12	BHT
 *	 ??	KDETH
 *	  8	RHF
 *	---
 *	 28 + KDETH
 *
 * For a 64-byte cache line, KDETH would need to be 36 bytes or 9 DWORDS
 */
#define DEFAULT_RCVHDRSIZE 9

/*
 * Maximal header byte count:
 *
 *	Bytes	Field
 *	  8	LRH
 *	 40	GRH (optional)
 *	 12	BTH
 *	 ??	KDETH
 *	  8	RHF
 *	---
 *	 68 + KDETH
 *
 * We also want to maintain a cache line alignment to assist DMA'ing
 * of the header bytes.  Round up to a good size.
 */
#define DEFAULT_RCVHDR_ENTSIZE 32

I
Ira Weiny 已提交
1925 1926 1927 1928
bool hfi1_can_pin_pages(struct hfi1_devdata *dd, struct mm_struct *mm,
			u32 nlocked, u32 npages);
int hfi1_acquire_user_pages(struct mm_struct *mm, unsigned long vaddr,
			    size_t npages, bool writable, struct page **pages);
1929 1930
void hfi1_release_user_pages(struct mm_struct *mm, struct page **p,
			     size_t npages, bool dirty);
M
Mike Marciniszyn 已提交
1931 1932 1933

static inline void clear_rcvhdrtail(const struct hfi1_ctxtdata *rcd)
{
1934
	*((u64 *)rcd->rcvhdrtail_kvaddr) = 0ULL;
M
Mike Marciniszyn 已提交
1935 1936 1937 1938 1939 1940 1941 1942
}

static inline u32 get_rcvhdrtail(const struct hfi1_ctxtdata *rcd)
{
	/*
	 * volatile because it's a DMA target from the chip, routine is
	 * inlined, and don't want register caching or reordering.
	 */
1943
	return (u32)le64_to_cpu(*rcd->rcvhdrtail_kvaddr);
M
Mike Marciniszyn 已提交
1944 1945 1946 1947 1948 1949 1950 1951
}

/*
 * sysfs interface.
 */

extern const char ib_hfi1_version[];

1952 1953
int hfi1_device_create(struct hfi1_devdata *dd);
void hfi1_device_remove(struct hfi1_devdata *dd);
M
Mike Marciniszyn 已提交
1954 1955 1956

int hfi1_create_port_files(struct ib_device *ibdev, u8 port_num,
			   struct kobject *kobj);
1957 1958
int hfi1_verbs_register_sysfs(struct hfi1_devdata *dd);
void hfi1_verbs_unregister_sysfs(struct hfi1_devdata *dd);
M
Mike Marciniszyn 已提交
1959 1960 1961
/* Hook for sysfs read of QSFP */
int qsfp_dump(struct hfi1_pportdata *ppd, char *buf, int len);

1962
int hfi1_pcie_init(struct pci_dev *pdev, const struct pci_device_id *ent);
1963
void hfi1_clean_up_interrupts(struct hfi1_devdata *dd);
1964 1965
void hfi1_pcie_cleanup(struct pci_dev *pdev);
int hfi1_pcie_ddinit(struct hfi1_devdata *dd, struct pci_dev *pdev);
M
Mike Marciniszyn 已提交
1966
void hfi1_pcie_ddcleanup(struct hfi1_devdata *);
1967
int pcie_speeds(struct hfi1_devdata *dd);
1968
int request_msix(struct hfi1_devdata *dd, u32 msireq);
1969
int restore_pci_variables(struct hfi1_devdata *dd);
1970
int save_pci_variables(struct hfi1_devdata *dd);
M
Mike Marciniszyn 已提交
1971 1972 1973
int do_pcie_gen3_transition(struct hfi1_devdata *dd);
int parse_platform_config(struct hfi1_devdata *dd);
int get_platform_config_field(struct hfi1_devdata *dd,
1974 1975 1976
			      enum platform_config_table_type_encoding
			      table_type, int table_index, int field_index,
			      u32 *data, u32 len);
M
Mike Marciniszyn 已提交
1977

1978
struct pci_dev *get_pci_dev(struct rvt_dev_info *rdi);
M
Mike Marciniszyn 已提交
1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

/*
 * Flush write combining store buffers (if present) and perform a write
 * barrier.
 */
static inline void flush_wc(void)
{
	asm volatile("sfence" : : : "memory");
}

void handle_eflags(struct hfi1_packet *packet);
int process_receive_ib(struct hfi1_packet *packet);
int process_receive_bypass(struct hfi1_packet *packet);
int process_receive_error(struct hfi1_packet *packet);
int kdeth_process_expected(struct hfi1_packet *packet);
int kdeth_process_eager(struct hfi1_packet *packet);
int process_receive_invalid(struct hfi1_packet *packet);
1996
void seqfile_dump_rcd(struct seq_file *s, struct hfi1_ctxtdata *rcd);
M
Mike Marciniszyn 已提交
1997 1998 1999 2000 2001

/* global module parameter variables */
extern unsigned int hfi1_max_mtu;
extern unsigned int hfi1_cu;
extern unsigned int user_credit_return_threshold;
2002
extern int num_user_contexts;
2003
extern unsigned long n_krcvqs;
2004
extern uint krcvqs[];
M
Mike Marciniszyn 已提交
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
extern int krcvqsset;
extern uint kdeth_qp;
extern uint loopback;
extern uint quick_linkup;
extern uint rcv_intr_timeout;
extern uint rcv_intr_count;
extern uint rcv_intr_dynamic;
extern ushort link_crc_mask;

extern struct mutex hfi1_mutex;

/* Number of seconds before our card status check...  */
#define STATUS_TIMEOUT 60

#define DRIVER_NAME		"hfi1"
#define HFI1_USER_MINOR_BASE     0
#define HFI1_TRACE_MINOR         127
#define HFI1_NMINORS             255

#define PCI_VENDOR_ID_INTEL 0x8086
#define PCI_DEVICE_ID_INTEL0 0x24f0
#define PCI_DEVICE_ID_INTEL1 0x24f1

#define HFI1_PKT_USER_SC_INTEGRITY					    \
	(SEND_CTXT_CHECK_ENABLE_DISALLOW_NON_KDETH_PACKETS_SMASK	    \
2030
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_KDETH_PACKETS_SMASK		\
M
Mike Marciniszyn 已提交
2031 2032 2033 2034 2035 2036 2037 2038 2039
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_BYPASS_SMASK		    \
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_GRH_SMASK)

#define HFI1_PKT_KERNEL_SC_INTEGRITY					    \
	(SEND_CTXT_CHECK_ENABLE_DISALLOW_KDETH_PACKETS_SMASK)

static inline u64 hfi1_pkt_default_send_ctxt_mask(struct hfi1_devdata *dd,
						  u16 ctxt_type)
{
2040 2041 2042 2043 2044 2045 2046
	u64 base_sc_integrity;

	/* No integrity checks if HFI1_CAP_NO_INTEGRITY is set */
	if (HFI1_CAP_IS_KSET(NO_INTEGRITY))
		return 0;

	base_sc_integrity =
M
Mike Marciniszyn 已提交
2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068
	SEND_CTXT_CHECK_ENABLE_DISALLOW_BYPASS_BAD_PKT_LEN_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_PBC_STATIC_RATE_CONTROL_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_TOO_LONG_BYPASS_PACKETS_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_TOO_LONG_IB_PACKETS_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_BAD_PKT_LEN_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_PBC_TEST_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_TOO_SMALL_BYPASS_PACKETS_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_TOO_SMALL_IB_PACKETS_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_RAW_IPV6_SMASK
	| SEND_CTXT_CHECK_ENABLE_DISALLOW_RAW_SMASK
	| SEND_CTXT_CHECK_ENABLE_CHECK_BYPASS_VL_MAPPING_SMASK
	| SEND_CTXT_CHECK_ENABLE_CHECK_VL_MAPPING_SMASK
	| SEND_CTXT_CHECK_ENABLE_CHECK_OPCODE_SMASK
	| SEND_CTXT_CHECK_ENABLE_CHECK_SLID_SMASK
	| SEND_CTXT_CHECK_ENABLE_CHECK_VL_SMASK
	| SEND_CTXT_CHECK_ENABLE_CHECK_ENABLE_SMASK;

	if (ctxt_type == SC_USER)
		base_sc_integrity |= HFI1_PKT_USER_SC_INTEGRITY;
	else
		base_sc_integrity |= HFI1_PKT_KERNEL_SC_INTEGRITY;

2069 2070 2071 2072
	/* turn on send-side job key checks if !A0 */
	if (!is_ax(dd))
		base_sc_integrity |= SEND_CTXT_CHECK_ENABLE_CHECK_JOB_KEY_SMASK;

M
Mike Marciniszyn 已提交
2073 2074 2075 2076 2077
	return base_sc_integrity;
}

static inline u64 hfi1_pkt_base_sdma_integrity(struct hfi1_devdata *dd)
{
2078 2079 2080 2081 2082 2083 2084
	u64 base_sdma_integrity;

	/* No integrity checks if HFI1_CAP_NO_INTEGRITY is set */
	if (HFI1_CAP_IS_KSET(NO_INTEGRITY))
		return 0;

	base_sdma_integrity =
M
Mike Marciniszyn 已提交
2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099
	SEND_DMA_CHECK_ENABLE_DISALLOW_BYPASS_BAD_PKT_LEN_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_TOO_LONG_BYPASS_PACKETS_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_TOO_LONG_IB_PACKETS_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_BAD_PKT_LEN_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_TOO_SMALL_BYPASS_PACKETS_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_TOO_SMALL_IB_PACKETS_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_RAW_IPV6_SMASK
	| SEND_DMA_CHECK_ENABLE_DISALLOW_RAW_SMASK
	| SEND_DMA_CHECK_ENABLE_CHECK_BYPASS_VL_MAPPING_SMASK
	| SEND_DMA_CHECK_ENABLE_CHECK_VL_MAPPING_SMASK
	| SEND_DMA_CHECK_ENABLE_CHECK_OPCODE_SMASK
	| SEND_DMA_CHECK_ENABLE_CHECK_SLID_SMASK
	| SEND_DMA_CHECK_ENABLE_CHECK_VL_SMASK
	| SEND_DMA_CHECK_ENABLE_CHECK_ENABLE_SMASK;

2100 2101 2102 2103 2104 2105 2106 2107 2108
	if (!HFI1_CAP_IS_KSET(STATIC_RATE_CTRL))
		base_sdma_integrity |=
		SEND_DMA_CHECK_ENABLE_DISALLOW_PBC_STATIC_RATE_CONTROL_SMASK;

	/* turn on send-side job key checks if !A0 */
	if (!is_ax(dd))
		base_sdma_integrity |=
			SEND_DMA_CHECK_ENABLE_CHECK_JOB_KEY_SMASK;

M
Mike Marciniszyn 已提交
2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126
	return base_sdma_integrity;
}

/*
 * hfi1_early_err is used (only!) to print early errors before devdata is
 * allocated, or when dd->pcidev may not be valid, and at the tail end of
 * cleanup when devdata may have been freed, etc.  hfi1_dev_porterr is
 * the same as dd_dev_err, but is used when the message really needs
 * the IB port# to be definitive as to what's happening..
 */
#define hfi1_early_err(dev, fmt, ...) \
	dev_err(dev, fmt, ##__VA_ARGS__)

#define hfi1_early_info(dev, fmt, ...) \
	dev_info(dev, fmt, ##__VA_ARGS__)

#define dd_dev_emerg(dd, fmt, ...) \
	dev_emerg(&(dd)->pcidev->dev, "%s: " fmt, \
2127
		  rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), ##__VA_ARGS__)
2128

M
Mike Marciniszyn 已提交
2129 2130
#define dd_dev_err(dd, fmt, ...) \
	dev_err(&(dd)->pcidev->dev, "%s: " fmt, \
2131
		rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), ##__VA_ARGS__)
2132 2133 2134

#define dd_dev_err_ratelimited(dd, fmt, ...) \
	dev_err_ratelimited(&(dd)->pcidev->dev, "%s: " fmt, \
2135 2136
			    rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), \
			    ##__VA_ARGS__)
2137

M
Mike Marciniszyn 已提交
2138 2139
#define dd_dev_warn(dd, fmt, ...) \
	dev_warn(&(dd)->pcidev->dev, "%s: " fmt, \
2140
		 rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), ##__VA_ARGS__)
M
Mike Marciniszyn 已提交
2141 2142 2143

#define dd_dev_warn_ratelimited(dd, fmt, ...) \
	dev_warn_ratelimited(&(dd)->pcidev->dev, "%s: " fmt, \
2144 2145
			     rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), \
			     ##__VA_ARGS__)
M
Mike Marciniszyn 已提交
2146 2147 2148

#define dd_dev_info(dd, fmt, ...) \
	dev_info(&(dd)->pcidev->dev, "%s: " fmt, \
2149
		 rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), ##__VA_ARGS__)
M
Mike Marciniszyn 已提交
2150

2151 2152
#define dd_dev_info_ratelimited(dd, fmt, ...) \
	dev_info_ratelimited(&(dd)->pcidev->dev, "%s: " fmt, \
2153 2154
			     rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), \
			     ##__VA_ARGS__)
2155

I
Ira Weiny 已提交
2156 2157
#define dd_dev_dbg(dd, fmt, ...) \
	dev_dbg(&(dd)->pcidev->dev, "%s: " fmt, \
2158
		rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), ##__VA_ARGS__)
I
Ira Weiny 已提交
2159

M
Mike Marciniszyn 已提交
2160
#define hfi1_dev_porterr(dd, port, fmt, ...) \
2161
	dev_err(&(dd)->pcidev->dev, "%s: port %u: " fmt, \
2162
		rvt_get_ibdev_name(&(dd)->verbs_dev.rdi), (port), ##__VA_ARGS__)
M
Mike Marciniszyn 已提交
2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189

/*
 * this is used for formatting hw error messages...
 */
struct hfi1_hwerror_msgs {
	u64 mask;
	const char *msg;
	size_t sz;
};

/* in intr.c... */
void hfi1_format_hwerrors(u64 hwerrs,
			  const struct hfi1_hwerror_msgs *hwerrmsgs,
			  size_t nhwerrmsgs, char *msg, size_t lmsg);

#define USER_OPCODE_CHECK_VAL 0xC0
#define USER_OPCODE_CHECK_MASK 0xC0
#define OPCODE_CHECK_VAL_DISABLED 0x0
#define OPCODE_CHECK_MASK_DISABLED 0x0

static inline void hfi1_reset_cpu_counters(struct hfi1_devdata *dd)
{
	struct hfi1_pportdata *ppd;
	int i;

	dd->z_int_counter = get_all_cpu_total(dd->int_counter);
	dd->z_rcv_limit = get_all_cpu_total(dd->rcv_limit);
2190
	dd->z_send_schedule = get_all_cpu_total(dd->send_schedule);
M
Mike Marciniszyn 已提交
2191 2192 2193

	ppd = (struct hfi1_pportdata *)(dd + 1);
	for (i = 0; i < dd->num_pports; i++, ppd++) {
2194 2195 2196 2197
		ppd->ibport_data.rvp.z_rc_acks =
			get_all_cpu_total(ppd->ibport_data.rvp.rc_acks);
		ppd->ibport_data.rvp.z_rc_qacks =
			get_all_cpu_total(ppd->ibport_data.rvp.rc_qacks);
M
Mike Marciniszyn 已提交
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209
	}
}

/* Control LED state */
static inline void setextled(struct hfi1_devdata *dd, u32 on)
{
	if (on)
		write_csr(dd, DCC_CFG_LED_CNTRL, 0x1F);
	else
		write_csr(dd, DCC_CFG_LED_CNTRL, 0x10);
}

2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221
/* return the i2c resource given the target */
static inline u32 i2c_target(u32 target)
{
	return target ? CR_I2C2 : CR_I2C1;
}

/* return the i2c chain chip resource that this HFI uses for QSFP */
static inline u32 qsfp_resource(struct hfi1_devdata *dd)
{
	return i2c_target(dd->hfi1_id);
}

2222 2223 2224 2225 2226 2227
/* Is this device integrated or discrete? */
static inline bool is_integrated(struct hfi1_devdata *dd)
{
	return dd->pcidev->device == PCI_DEVICE_ID_INTEL1;
}

M
Mike Marciniszyn 已提交
2228 2229
int hfi1_tempsense_rd(struct hfi1_devdata *dd, struct hfi1_temp *temp);

2230 2231
#define DD_DEV_ENTRY(dd)       __string(dev, dev_name(&(dd)->pcidev->dev))
#define DD_DEV_ASSIGN(dd)      __assign_str(dev, dev_name(&(dd)->pcidev->dev))
2232

2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
static inline void hfi1_update_ah_attr(struct ib_device *ibdev,
				       struct rdma_ah_attr *attr)
{
	struct hfi1_pportdata *ppd;
	struct hfi1_ibport *ibp;
	u32 dlid = rdma_ah_get_dlid(attr);

	/*
	 * Kernel clients may not have setup GRH information
	 * Set that here.
	 */
	ibp = to_iport(ibdev, rdma_ah_get_port_num(attr));
	ppd = ppd_from_ibp(ibp);
	if ((((dlid >= be16_to_cpu(IB_MULTICAST_LID_BASE)) ||
	      (ppd->lid >= be16_to_cpu(IB_MULTICAST_LID_BASE))) &&
	    (dlid != be32_to_cpu(OPA_LID_PERMISSIVE)) &&
	    (dlid != be16_to_cpu(IB_LID_PERMISSIVE)) &&
	    (!(rdma_ah_get_ah_flags(attr) & IB_AH_GRH))) ||
	    (rdma_ah_get_make_grd(attr))) {
		rdma_ah_set_ah_flags(attr, IB_AH_GRH);
		rdma_ah_set_interface_id(attr, OPA_MAKE_ID(dlid));
		rdma_ah_set_subnet_prefix(attr, ibp->rvp.gid_prefix);
	}
}

2258 2259
/*
 * hfi1_check_mcast- Check if the given lid is
2260 2261 2262 2263 2264
 * in the OPA multicast range.
 *
 * The LID might either reside in ah.dlid or might be
 * in the GRH of the address handle as DGID if extended
 * addresses are in use.
2265
 */
2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305
static inline bool hfi1_check_mcast(u32 lid)
{
	return ((lid >= opa_get_mcast_base(OPA_MCAST_NR)) &&
		(lid != be32_to_cpu(OPA_LID_PERMISSIVE)));
}

#define opa_get_lid(lid, format)	\
	__opa_get_lid(lid, OPA_PORT_PACKET_FORMAT_##format)

/* Convert a lid to a specific lid space */
static inline u32 __opa_get_lid(u32 lid, u8 format)
{
	bool is_mcast = hfi1_check_mcast(lid);

	switch (format) {
	case OPA_PORT_PACKET_FORMAT_8B:
	case OPA_PORT_PACKET_FORMAT_10B:
		if (is_mcast)
			return (lid - opa_get_mcast_base(OPA_MCAST_NR) +
				0xF0000);
		return lid & 0xFFFFF;
	case OPA_PORT_PACKET_FORMAT_16B:
		if (is_mcast)
			return (lid - opa_get_mcast_base(OPA_MCAST_NR) +
				0xF00000);
		return lid & 0xFFFFFF;
	case OPA_PORT_PACKET_FORMAT_9B:
		if (is_mcast)
			return (lid -
				opa_get_mcast_base(OPA_MCAST_NR) +
				be16_to_cpu(IB_MULTICAST_LID_BASE));
		else
			return lid & 0xFFFF;
	default:
		return lid;
	}
}

/* Return true if the given lid is the OPA 16B multicast range */
static inline bool hfi1_is_16B_mcast(u32 lid)
2306
{
2307 2308 2309
	return ((lid >=
		opa_get_lid(opa_get_mcast_base(OPA_MCAST_NR), 16B)) &&
		(lid != opa_get_lid(be32_to_cpu(OPA_LID_PERMISSIVE), 16B)));
2310
}
2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372

static inline void hfi1_make_opa_lid(struct rdma_ah_attr *attr)
{
	const struct ib_global_route *grh = rdma_ah_read_grh(attr);
	u32 dlid = rdma_ah_get_dlid(attr);

	/* Modify ah_attr.dlid to be in the 32 bit LID space.
	 * This is how the address will be laid out:
	 * Assuming MCAST_NR to be 4,
	 * 32 bit permissive LID = 0xFFFFFFFF
	 * Multicast LID range = 0xFFFFFFFE to 0xF0000000
	 * Unicast LID range = 0xEFFFFFFF to 1
	 * Invalid LID = 0
	 */
	if (ib_is_opa_gid(&grh->dgid))
		dlid = opa_get_lid_from_gid(&grh->dgid);
	else if ((dlid >= be16_to_cpu(IB_MULTICAST_LID_BASE)) &&
		 (dlid != be16_to_cpu(IB_LID_PERMISSIVE)) &&
		 (dlid != be32_to_cpu(OPA_LID_PERMISSIVE)))
		dlid = dlid - be16_to_cpu(IB_MULTICAST_LID_BASE) +
			opa_get_mcast_base(OPA_MCAST_NR);
	else if (dlid == be16_to_cpu(IB_LID_PERMISSIVE))
		dlid = be32_to_cpu(OPA_LID_PERMISSIVE);

	rdma_ah_set_dlid(attr, dlid);
}

static inline u8 hfi1_get_packet_type(u32 lid)
{
	/* 9B if lid > 0xF0000000 */
	if (lid >= opa_get_mcast_base(OPA_MCAST_NR))
		return HFI1_PKT_TYPE_9B;

	/* 16B if lid > 0xC000 */
	if (lid >= opa_get_lid(opa_get_mcast_base(OPA_MCAST_NR), 9B))
		return HFI1_PKT_TYPE_16B;

	return HFI1_PKT_TYPE_9B;
}

static inline bool hfi1_get_hdr_type(u32 lid, struct rdma_ah_attr *attr)
{
	/*
	 * If there was an incoming 16B packet with permissive
	 * LIDs, OPA GIDs would have been programmed when those
	 * packets were received. A 16B packet will have to
	 * be sent in response to that packet. Return a 16B
	 * header type if that's the case.
	 */
	if (rdma_ah_get_dlid(attr) == be32_to_cpu(OPA_LID_PERMISSIVE))
		return (ib_is_opa_gid(&rdma_ah_read_grh(attr)->dgid)) ?
			HFI1_PKT_TYPE_16B : HFI1_PKT_TYPE_9B;

	/*
	 * Return a 16B header type if either the the destination
	 * or source lid is extended.
	 */
	if (hfi1_get_packet_type(rdma_ah_get_dlid(attr)) == HFI1_PKT_TYPE_16B)
		return HFI1_PKT_TYPE_16B;

	return hfi1_get_packet_type(lid);
}
D
Don Hiatt 已提交
2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422

static inline void hfi1_make_ext_grh(struct hfi1_packet *packet,
				     struct ib_grh *grh, u32 slid,
				     u32 dlid)
{
	struct hfi1_ibport *ibp = &packet->rcd->ppd->ibport_data;
	struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);

	if (!ibp)
		return;

	grh->hop_limit = 1;
	grh->sgid.global.subnet_prefix = ibp->rvp.gid_prefix;
	if (slid == opa_get_lid(be32_to_cpu(OPA_LID_PERMISSIVE), 16B))
		grh->sgid.global.interface_id =
			OPA_MAKE_ID(be32_to_cpu(OPA_LID_PERMISSIVE));
	else
		grh->sgid.global.interface_id = OPA_MAKE_ID(slid);

	/*
	 * Upper layers (like mad) may compare the dgid in the
	 * wc that is obtained here with the sgid_index in
	 * the wr. Since sgid_index in wr is always 0 for
	 * extended lids, set the dgid here to the default
	 * IB gid.
	 */
	grh->dgid.global.subnet_prefix = ibp->rvp.gid_prefix;
	grh->dgid.global.interface_id =
		cpu_to_be64(ppd->guids[HFI1_PORT_GUID_INDEX]);
}

static inline int hfi1_get_16b_padding(u32 hdr_size, u32 payload)
{
	return -(hdr_size + payload + (SIZE_OF_CRC << 2) +
		     SIZE_OF_LT) & 0x7;
}

static inline void hfi1_make_ib_hdr(struct ib_header *hdr,
				    u16 lrh0, u16 len,
				    u16 dlid, u16 slid)
{
	hdr->lrh[0] = cpu_to_be16(lrh0);
	hdr->lrh[1] = cpu_to_be16(dlid);
	hdr->lrh[2] = cpu_to_be16(len);
	hdr->lrh[3] = cpu_to_be16(slid);
}

static inline void hfi1_make_16b_hdr(struct hfi1_16b_header *hdr,
				     u32 slid, u32 dlid,
				     u16 len, u16 pkey,
2423
				     bool becn, bool fecn, u8 l4,
D
Don Hiatt 已提交
2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448
				     u8 sc)
{
	u32 lrh0 = 0;
	u32 lrh1 = 0x40000000;
	u32 lrh2 = 0;
	u32 lrh3 = 0;

	lrh0 = (lrh0 & ~OPA_16B_BECN_MASK) | (becn << OPA_16B_BECN_SHIFT);
	lrh0 = (lrh0 & ~OPA_16B_LEN_MASK) | (len << OPA_16B_LEN_SHIFT);
	lrh0 = (lrh0 & ~OPA_16B_LID_MASK)  | (slid & OPA_16B_LID_MASK);
	lrh1 = (lrh1 & ~OPA_16B_FECN_MASK) | (fecn << OPA_16B_FECN_SHIFT);
	lrh1 = (lrh1 & ~OPA_16B_SC_MASK) | (sc << OPA_16B_SC_SHIFT);
	lrh1 = (lrh1 & ~OPA_16B_LID_MASK) | (dlid & OPA_16B_LID_MASK);
	lrh2 = (lrh2 & ~OPA_16B_SLID_MASK) |
		((slid >> OPA_16B_SLID_SHIFT) << OPA_16B_SLID_HIGH_SHIFT);
	lrh2 = (lrh2 & ~OPA_16B_DLID_MASK) |
		((dlid >> OPA_16B_DLID_SHIFT) << OPA_16B_DLID_HIGH_SHIFT);
	lrh2 = (lrh2 & ~OPA_16B_PKEY_MASK) | (pkey << OPA_16B_PKEY_SHIFT);
	lrh2 = (lrh2 & ~OPA_16B_L4_MASK) | l4;

	hdr->lrh[0] = lrh0;
	hdr->lrh[1] = lrh1;
	hdr->lrh[2] = lrh2;
	hdr->lrh[3] = lrh3;
}
M
Mike Marciniszyn 已提交
2449
#endif                          /* _HFI1_KERNEL_H */