net_driver.h 47.5 KB
Newer Older
1
/****************************************************************************
B
Ben Hutchings 已提交
2
 * Driver for Solarflare network controllers and boards
3
 * Copyright 2005-2006 Fen Systems Ltd.
B
Ben Hutchings 已提交
4
 * Copyright 2005-2013 Solarflare Communications Inc.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License version 2 as published
 * by the Free Software Foundation, incorporated herein by reference.
 */

/* Common definitions for all Efx net driver code */

#ifndef EFX_NET_DRIVER_H
#define EFX_NET_DRIVER_H

#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
#include <linux/if_vlan.h>
20
#include <linux/timer.h>
21
#include <linux/mdio.h>
22 23 24 25 26
#include <linux/list.h>
#include <linux/pci.h>
#include <linux/device.h>
#include <linux/highmem.h>
#include <linux/workqueue.h>
27
#include <linux/mutex.h>
28
#include <linux/vmalloc.h>
29
#include <linux/i2c.h>
30
#include <linux/mtd/mtd.h>
31 32 33

#include "enum.h"
#include "bitfield.h"
34
#include "filter.h"
35 36 37 38 39 40

/**************************************************************************
 *
 * Build definitions
 *
 **************************************************************************/
41

42
#define EFX_DRIVER_VERSION	"4.0"
43

44
#ifdef DEBUG
45 46 47 48 49 50 51 52 53 54 55 56 57
#define EFX_BUG_ON_PARANOID(x) BUG_ON(x)
#define EFX_WARN_ON_PARANOID(x) WARN_ON(x)
#else
#define EFX_BUG_ON_PARANOID(x) do {} while (0)
#define EFX_WARN_ON_PARANOID(x) do {} while (0)
#endif

/**************************************************************************
 *
 * Efx data structures
 *
 **************************************************************************/

58
#define EFX_MAX_CHANNELS 32U
59
#define EFX_MAX_RX_QUEUES EFX_MAX_CHANNELS
60
#define EFX_EXTRA_CHANNEL_IOV	0
61 62
#define EFX_EXTRA_CHANNEL_PTP	1
#define EFX_MAX_EXTRA_CHANNELS	2U
63

B
Ben Hutchings 已提交
64 65 66
/* Checksum generation is a per-queue option in hardware, so each
 * queue visible to the networking core is backed by two hardware TX
 * queues. */
67 68 69 70 71 72
#define EFX_MAX_TX_TC		2
#define EFX_MAX_CORE_TX_QUEUES	(EFX_MAX_TX_TC * EFX_MAX_CHANNELS)
#define EFX_TXQ_TYPE_OFFLOAD	1	/* flag */
#define EFX_TXQ_TYPE_HIGHPRI	2	/* flag */
#define EFX_TXQ_TYPES		4
#define EFX_MAX_TX_QUEUES	(EFX_TXQ_TYPES * EFX_MAX_CHANNELS)
73

74 75 76
/* Maximum possible MTU the driver supports */
#define EFX_MAX_MTU (9 * 1024)

77 78 79 80 81 82 83 84 85 86 87 88 89 90
/* Size of an RX scatter buffer.  Small enough to pack 2 into a 4K page,
 * and should be a multiple of the cache line size.
 */
#define EFX_RX_USR_BUF_SIZE	(2048 - 256)

/* If possible, we should ensure cache line alignment at start and end
 * of every buffer.  Otherwise, we just need to ensure 4-byte
 * alignment of the network header.
 */
#if NET_IP_ALIGN == 0
#define EFX_RX_BUF_ALIGNMENT	L1_CACHE_BYTES
#else
#define EFX_RX_BUF_ALIGNMENT	4
#endif
91

92 93 94
/* Forward declare Precision Time Protocol (PTP) support structure. */
struct efx_ptp_data;

95 96
struct efx_self_tests;

97
/**
98 99
 * struct efx_buffer - A general-purpose DMA buffer
 * @addr: host base address of the buffer
100 101 102
 * @dma_addr: DMA base address of the buffer
 * @len: Buffer length, in bytes
 *
103 104
 * The NIC uses these buffers for its interrupt status registers and
 * MAC stats dumps.
105
 */
106
struct efx_buffer {
107 108 109
	void *addr;
	dma_addr_t dma_addr;
	unsigned int len;
110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
};

/**
 * struct efx_special_buffer - DMA buffer entered into buffer table
 * @buf: Standard &struct efx_buffer
 * @index: Buffer index within controller;s buffer table
 * @entries: Number of buffer table entries
 *
 * The NIC has a buffer table that maps buffers of size %EFX_BUF_SIZE.
 * Event and descriptor rings are addressed via one or more buffer
 * table entries (and so can be physically non-contiguous, although we
 * currently do not take advantage of that).  On Falcon and Siena we
 * have to take care of allocating and initialising the entries
 * ourselves.  On later hardware this is managed by the firmware and
 * @index and @entries are left as 0.
 */
struct efx_special_buffer {
	struct efx_buffer buf;
128 129
	unsigned int index;
	unsigned int entries;
130 131 132
};

/**
133 134 135
 * struct efx_tx_buffer - buffer state for a TX descriptor
 * @skb: When @flags & %EFX_TX_BUF_SKB, the associated socket buffer to be
 *	freed when descriptor completes
136 137
 * @heap_buf: When @flags & %EFX_TX_BUF_HEAP, the associated heap buffer to be
 *	freed when descriptor completes.
138
 * @option: When @flags & %EFX_TX_BUF_OPTION, a NIC-specific option descriptor.
139
 * @dma_addr: DMA address of the fragment.
140
 * @flags: Flags for allocation and DMA mapping type
141 142 143 144 145
 * @len: Length of this fragment.
 *	This field is zero when the queue slot is empty.
 * @unmap_len: Length of this fragment to unmap
 */
struct efx_tx_buffer {
146 147
	union {
		const struct sk_buff *skb;
148
		void *heap_buf;
149
	};
150 151 152 153
	union {
		efx_qword_t option;
		dma_addr_t dma_addr;
	};
154
	unsigned short flags;
155 156 157
	unsigned short len;
	unsigned short unmap_len;
};
158 159
#define EFX_TX_BUF_CONT		1	/* not last descriptor of packet */
#define EFX_TX_BUF_SKB		2	/* buffer is last part of skb */
160
#define EFX_TX_BUF_HEAP		4	/* buffer was allocated with kmalloc() */
161
#define EFX_TX_BUF_MAP_SINGLE	8	/* buffer was mapped with dma_map_single() */
162
#define EFX_TX_BUF_OPTION	0x10	/* empty buffer for option descriptor */
163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179

/**
 * struct efx_tx_queue - An Efx TX queue
 *
 * This is a ring buffer of TX fragments.
 * Since the TX completion path always executes on the same
 * CPU and the xmit path can operate on different CPUs,
 * performance is increased by ensuring that the completion
 * path and the xmit path operate on different cache lines.
 * This is particularly important if the xmit path is always
 * executing on one CPU which is different from the completion
 * path.  There is also a cache line for members which are
 * read but not written on the fast path.
 *
 * @efx: The associated Efx NIC
 * @queue: DMA queue number
 * @channel: The associated channel
180
 * @core_txq: The networking core TX queue structure
181
 * @buffer: The software buffer ring
182
 * @tsoh_page: Array of pages of TSO header buffers
183
 * @txd: The hardware descriptor ring
184
 * @ptr_mask: The size of the ring minus 1.
185 186 187
 * @piobuf: PIO buffer region for this TX queue (shared with its partner).
 *	Size of the region is efx_piobuf_size.
 * @piobuf_offset: Buffer offset to be specified in PIO descriptors
188
 * @initialised: Has hardware queue been initialised?
189 190
 * @read_count: Current read pointer.
 *	This is the number of buffers that have been removed from both rings.
191 192 193 194 195 196
 * @old_write_count: The value of @write_count when last checked.
 *	This is here for performance reasons.  The xmit path will
 *	only get the up-to-date value of @write_count if this
 *	variable indicates that the queue is empty.  This is to
 *	avoid cache-line ping-pong between the xmit path and the
 *	completion path.
197
 * @merge_events: Number of TX merged completion events
198 199 200 201 202 203 204 205 206 207 208 209
 * @insert_count: Current insert pointer
 *	This is the number of buffers that have been added to the
 *	software ring.
 * @write_count: Current write pointer
 *	This is the number of buffers that have been added to the
 *	hardware ring.
 * @old_read_count: The value of read_count when last checked.
 *	This is here for performance reasons.  The xmit path will
 *	only get the up-to-date value of read_count if this
 *	variable indicates that the queue is full.  This is to
 *	avoid cache-line ping-pong between the xmit path and the
 *	completion path.
B
Ben Hutchings 已提交
210 211 212 213
 * @tso_bursts: Number of times TSO xmit invoked by kernel
 * @tso_long_headers: Number of packets with headers too long for standard
 *	blocks
 * @tso_packets: Number of packets via the TSO xmit path
214 215 216 217
 * @pushes: Number of times the TX push feature has been used
 * @empty_read_count: If the completion path has seen the queue as empty
 *	and the transmission path has not yet checked this, the value of
 *	@read_count bitwise-added to %EFX_EMPTY_COUNT_VALID; otherwise 0.
218 219 220 221
 */
struct efx_tx_queue {
	/* Members which don't change on the fast path */
	struct efx_nic *efx ____cacheline_aligned_in_smp;
B
Ben Hutchings 已提交
222
	unsigned queue;
223
	struct efx_channel *channel;
224
	struct netdev_queue *core_txq;
225
	struct efx_tx_buffer *buffer;
226
	struct efx_buffer *tsoh_page;
227
	struct efx_special_buffer txd;
228
	unsigned int ptr_mask;
229 230
	void __iomem *piobuf;
	unsigned int piobuf_offset;
231
	bool initialised;
232 233 234

	/* Members used mainly on the completion path */
	unsigned int read_count ____cacheline_aligned_in_smp;
235
	unsigned int old_write_count;
236
	unsigned int merge_events;
237 238 239 240 241

	/* Members used only on the xmit path */
	unsigned int insert_count ____cacheline_aligned_in_smp;
	unsigned int write_count;
	unsigned int old_read_count;
B
Ben Hutchings 已提交
242 243 244
	unsigned int tso_bursts;
	unsigned int tso_long_headers;
	unsigned int tso_packets;
245 246 247 248 249
	unsigned int pushes;

	/* Members shared between paths and sometimes updated */
	unsigned int empty_read_count ____cacheline_aligned_in_smp;
#define EFX_EMPTY_COUNT_VALID 0x80000000
250
	atomic_t flush_outstanding;
251 252 253 254 255
};

/**
 * struct efx_rx_buffer - An Efx RX data buffer
 * @dma_addr: DMA base address of the buffer
256
 * @page: The associated page buffer.
257
 *	Will be %NULL if the buffer slot is currently free.
258 259
 * @page_offset: If pending: offset in @page of DMA base address.
 *	If completed: offset in @page of Ethernet header.
260 261
 * @len: If pending: length for DMA descriptor.
 *	If completed: received length, excluding hash prefix.
262 263
 * @flags: Flags for buffer and packet state.  These are only set on the
 *	first buffer of a scattered packet.
264 265 266
 */
struct efx_rx_buffer {
	dma_addr_t dma_addr;
267
	struct page *page;
268 269
	u16 page_offset;
	u16 len;
270
	u16 flags;
271
};
272
#define EFX_RX_BUF_LAST_IN_PAGE	0x0001
273 274
#define EFX_RX_PKT_CSUMMED	0x0002
#define EFX_RX_PKT_DISCARD	0x0004
275
#define EFX_RX_PKT_TCP		0x0040
276
#define EFX_RX_PKT_PREFIX_LEN	0x0080	/* length is in prefix only */
277

278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
/**
 * struct efx_rx_page_state - Page-based rx buffer state
 *
 * Inserted at the start of every page allocated for receive buffers.
 * Used to facilitate sharing dma mappings between recycled rx buffers
 * and those passed up to the kernel.
 *
 * @refcnt: Number of struct efx_rx_buffer's referencing this page.
 *	When refcnt falls to zero, the page is unmapped for dma
 * @dma_addr: The dma address of this page.
 */
struct efx_rx_page_state {
	unsigned refcnt;
	dma_addr_t dma_addr;

	unsigned int __pad[0] ____cacheline_aligned;
};

296 297 298
/**
 * struct efx_rx_queue - An Efx RX queue
 * @efx: The associated Efx NIC
299 300
 * @core_index:  Index of network core RX queue.  Will be >= 0 iff this
 *	is associated with a real RX queue.
301 302
 * @buffer: The software buffer ring
 * @rxd: The hardware descriptor ring
303
 * @ptr_mask: The size of the ring minus 1.
304
 * @refill_enabled: Enable refill whenever fill level is low
305 306
 * @flush_pending: Set when a RX flush is pending. Has the same lifetime as
 *	@rxq_flush_pending.
307 308 309
 * @added_count: Number of buffers added to the receive queue.
 * @notified_count: Number of buffers given to NIC (<= @added_count).
 * @removed_count: Number of buffers removed from the receive queue.
J
Jon Cooper 已提交
310 311
 * @scatter_n: Used by NIC specific receive code.
 * @scatter_len: Used by NIC specific receive code.
312 313 314 315 316 317 318 319 320
 * @page_ring: The ring to store DMA mapped pages for reuse.
 * @page_add: Counter to calculate the write pointer for the recycle ring.
 * @page_remove: Counter to calculate the read pointer for the recycle ring.
 * @page_recycle_count: The number of pages that have been recycled.
 * @page_recycle_failed: The number of pages that couldn't be recycled because
 *      the kernel still held a reference to them.
 * @page_recycle_full: The number of pages that were released because the
 *      recycle ring was full.
 * @page_ptr_mask: The number of pages in the RX recycle ring minus 1.
321 322 323 324 325 326
 * @max_fill: RX descriptor maximum fill level (<= ring size)
 * @fast_fill_trigger: RX descriptor fill level that will trigger a fast fill
 *	(<= @max_fill)
 * @min_fill: RX descriptor minimum non-zero fill level.
 *	This records the minimum fill level observed when a ring
 *	refill was triggered.
327
 * @recycle_count: RX buffer recycle counter.
328
 * @slow_fill: Timer used to defer efx_nic_generate_fill_event().
329 330 331
 */
struct efx_rx_queue {
	struct efx_nic *efx;
332
	int core_index;
333 334
	struct efx_rx_buffer *buffer;
	struct efx_special_buffer rxd;
335
	unsigned int ptr_mask;
336
	bool refill_enabled;
337
	bool flush_pending;
338

339 340 341
	unsigned int added_count;
	unsigned int notified_count;
	unsigned int removed_count;
342
	unsigned int scatter_n;
J
Jon Cooper 已提交
343
	unsigned int scatter_len;
344 345 346 347 348 349 350
	struct page **page_ring;
	unsigned int page_add;
	unsigned int page_remove;
	unsigned int page_recycle_count;
	unsigned int page_recycle_failed;
	unsigned int page_recycle_full;
	unsigned int page_ptr_mask;
351 352 353 354
	unsigned int max_fill;
	unsigned int fast_fill_trigger;
	unsigned int min_fill;
	unsigned int min_overfill;
355
	unsigned int recycle_count;
356
	struct timer_list slow_fill;
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
	unsigned int slow_fill_count;
};

enum efx_rx_alloc_method {
	RX_ALLOC_METHOD_AUTO = 0,
	RX_ALLOC_METHOD_SKB = 1,
	RX_ALLOC_METHOD_PAGE = 2,
};

/**
 * struct efx_channel - An Efx channel
 *
 * A channel comprises an event queue, at least one TX queue, at least
 * one RX queue, and an associated tasklet for processing the event
 * queue.
 *
 * @efx: Associated Efx NIC
 * @channel: Channel instance number
375
 * @type: Channel type definition
376
 * @eventq_init: Event queue initialised flag
377 378
 * @enabled: Channel enabled indicator
 * @irq: IRQ number (MSI and MSI-X only)
379
 * @irq_moderation: IRQ moderation value (in hardware ticks)
380 381 382
 * @napi_dev: Net device used with NAPI
 * @napi_str: NAPI control structure
 * @eventq: Event queue buffer
383
 * @eventq_mask: Event queue pointer mask
384
 * @eventq_read_ptr: Event queue read pointer
385
 * @event_test_cpu: Last CPU to handle interrupt or test event for this channel
386 387
 * @irq_count: Number of IRQs since last adaptive moderation decision
 * @irq_mod_score: IRQ moderation score
388 389 390
 * @n_rx_tobe_disc: Count of RX_TOBE_DISC errors
 * @n_rx_ip_hdr_chksum_err: Count of RX IP header checksum errors
 * @n_rx_tcp_udp_chksum_err: Count of RX TCP and UDP checksum errors
B
Ben Hutchings 已提交
391
 * @n_rx_mcast_mismatch: Count of unmatched multicast frames
392 393 394
 * @n_rx_frm_trunc: Count of RX_FRM_TRUNC errors
 * @n_rx_overlength: Count of RX_OVERLENGTH errors
 * @n_skbuff_leaks: Count of skbuffs leaked due to RX overrun
395 396
 * @n_rx_nodesc_trunc: Number of RX packets truncated and then dropped due to
 *	lack of descriptors
397 398
 * @n_rx_merge_events: Number of RX merged completion events
 * @n_rx_merge_packets: Number of RX packets completed by merged events
399 400 401 402
 * @rx_pkt_n_frags: Number of fragments in next packet to be delivered by
 *	__efx_rx_packet(), or zero if there is none
 * @rx_pkt_index: Ring index of first buffer for next packet to be delivered
 *	by __efx_rx_packet(), if @rx_pkt_n_frags != 0
403 404
 * @rx_queue: RX queue for this channel
 * @tx_queue: TX queues for this channel
405 406 407 408
 */
struct efx_channel {
	struct efx_nic *efx;
	int channel;
409
	const struct efx_channel_type *type;
410
	bool eventq_init;
411
	bool enabled;
412 413 414 415 416
	int irq;
	unsigned int irq_moderation;
	struct net_device *napi_dev;
	struct napi_struct napi_str;
	struct efx_special_buffer eventq;
417
	unsigned int eventq_mask;
418
	unsigned int eventq_read_ptr;
419
	int event_test_cpu;
420

421 422
	unsigned int irq_count;
	unsigned int irq_mod_score;
423 424 425
#ifdef CONFIG_RFS_ACCEL
	unsigned int rfs_filters_added;
#endif
426

427 428 429
	unsigned n_rx_tobe_disc;
	unsigned n_rx_ip_hdr_chksum_err;
	unsigned n_rx_tcp_udp_chksum_err;
B
Ben Hutchings 已提交
430
	unsigned n_rx_mcast_mismatch;
431 432 433
	unsigned n_rx_frm_trunc;
	unsigned n_rx_overlength;
	unsigned n_skbuff_leaks;
434
	unsigned int n_rx_nodesc_trunc;
435 436
	unsigned int n_rx_merge_events;
	unsigned int n_rx_merge_packets;
437

438 439
	unsigned int rx_pkt_n_frags;
	unsigned int rx_pkt_index;
440

441
	struct efx_rx_queue rx_queue;
442
	struct efx_tx_queue tx_queue[EFX_TXQ_TYPES];
443 444
};

B
Ben Hutchings 已提交
445 446 447 448 449 450 451 452 453 454 455 456 457 458 459
/**
 * struct efx_msi_context - Context for each MSI
 * @efx: The associated NIC
 * @index: Index of the channel/IRQ
 * @name: Name of the channel/IRQ
 *
 * Unlike &struct efx_channel, this is never reallocated and is always
 * safe for the IRQ handler to access.
 */
struct efx_msi_context {
	struct efx_nic *efx;
	unsigned int index;
	char name[IFNAMSIZ + 6];
};

460 461 462 463 464 465 466 467 468
/**
 * struct efx_channel_type - distinguishes traffic and extra channels
 * @handle_no_channel: Handle failure to allocate an extra channel
 * @pre_probe: Set up extra state prior to initialisation
 * @post_remove: Tear down extra state after finalisation, if allocated.
 *	May be called on channels that have not been probed.
 * @get_name: Generate the channel's name (used for its IRQ handler)
 * @copy: Copy the channel state prior to reallocation.  May be %NULL if
 *	reallocation is not supported.
469
 * @receive_skb: Handle an skb ready to be passed to netif_receive_skb()
470 471 472 473 474 475
 * @keep_eventq: Flag for whether event queue should be kept initialised
 *	while the device is stopped
 */
struct efx_channel_type {
	void (*handle_no_channel)(struct efx_nic *);
	int (*pre_probe)(struct efx_channel *);
476
	void (*post_remove)(struct efx_channel *);
477 478
	void (*get_name)(struct efx_channel *, char *buf, size_t len);
	struct efx_channel *(*copy)(const struct efx_channel *);
479
	bool (*receive_skb)(struct efx_channel *, struct sk_buff *);
480 481 482
	bool keep_eventq;
};

483 484 485 486 487 488
enum efx_led_mode {
	EFX_LED_OFF	= 0,
	EFX_LED_ON	= 1,
	EFX_LED_DEFAULT	= 2
};

489 490 491
#define STRING_TABLE_LOOKUP(val, member) \
	((val) < member ## _max) ? member ## _names[val] : "(invalid)"

492
extern const char *const efx_loopback_mode_names[];
493 494 495 496
extern const unsigned int efx_loopback_mode_max;
#define LOOPBACK_MODE(efx) \
	STRING_TABLE_LOOKUP((efx)->loopback_mode, efx_loopback_mode)

497
extern const char *const efx_reset_type_names[];
498 499 500
extern const unsigned int efx_reset_type_max;
#define RESET_TYPE(type) \
	STRING_TABLE_LOOKUP(type, efx_reset_type)
501

502 503 504 505 506 507 508 509 510 511
enum efx_int_mode {
	/* Be careful if altering to correct macro below */
	EFX_INT_MODE_MSIX = 0,
	EFX_INT_MODE_MSI = 1,
	EFX_INT_MODE_LEGACY = 2,
	EFX_INT_MODE_MAX	/* Insert any new items before this */
};
#define EFX_INT_MODE_USE_MSI(x) (((x)->interrupt_mode) <= EFX_INT_MODE_MSI)

enum nic_state {
512 513 514
	STATE_UNINIT = 0,	/* device being probed/removed or is frozen */
	STATE_READY = 1,	/* hardware ready and netdev registered */
	STATE_DISABLED = 2,	/* device disabled due to hardware errors */
515
	STATE_RECOVERY = 3,	/* device recovering from PCI error */
516 517 518 519 520 521 522
};

/*
 * Alignment of the skb->head which wraps a page-allocated RX buffer
 *
 * The skb allocated to wrap an rx_buffer can have this alignment. Since
 * the data is memcpy'd from the rx_buf, it does not need to be equal to
523
 * NET_IP_ALIGN.
524 525 526 527 528 529 530
 */
#define EFX_PAGE_SKB_ALIGN 2

/* Forward declaration */
struct efx_nic;

/* Pseudo bit-mask flow control field */
531 532 533
#define EFX_FC_RX	FLOW_CTRL_RX
#define EFX_FC_TX	FLOW_CTRL_TX
#define EFX_FC_AUTO	4
534

535 536 537 538 539 540 541 542 543 544
/**
 * struct efx_link_state - Current state of the link
 * @up: Link is up
 * @fd: Link is full-duplex
 * @fc: Actual flow control flags
 * @speed: Link speed (Mbps)
 */
struct efx_link_state {
	bool up;
	bool fd;
545
	u8 fc;
546 547 548
	unsigned int speed;
};

S
Steve Hodgson 已提交
549 550 551 552 553 554 555
static inline bool efx_link_state_equal(const struct efx_link_state *left,
					const struct efx_link_state *right)
{
	return left->up == right->up && left->fd == right->fd &&
		left->fc == right->fc && left->speed == right->speed;
}

556 557
/**
 * struct efx_phy_operations - Efx PHY operations table
558 559
 * @probe: Probe PHY and initialise efx->mdio.mode_support, efx->mdio.mmds,
 *	efx->loopback_modes.
560 561 562
 * @init: Initialise PHY
 * @fini: Shut down PHY
 * @reconfigure: Reconfigure PHY (e.g. for new link parameters)
S
Steve Hodgson 已提交
563 564
 * @poll: Update @link_state and report whether it changed.
 *	Serialised by the mac_lock.
565 566
 * @get_settings: Get ethtool settings. Serialised by the mac_lock.
 * @set_settings: Set ethtool settings. Serialised by the mac_lock.
567
 * @set_npage_adv: Set abilities advertised in (Extended) Next Page
B
Ben Hutchings 已提交
568
 *	(only needed where AN bit is set in mmds)
569
 * @test_alive: Test that PHY is 'alive' (online)
570
 * @test_name: Get the name of a PHY-specific test/result
571
 * @run_tests: Run tests and record results as appropriate (offline).
572
 *	Flags are the ethtool tests flags.
573 574
 */
struct efx_phy_operations {
575
	int (*probe) (struct efx_nic *efx);
576 577
	int (*init) (struct efx_nic *efx);
	void (*fini) (struct efx_nic *efx);
578
	void (*remove) (struct efx_nic *efx);
B
Ben Hutchings 已提交
579
	int (*reconfigure) (struct efx_nic *efx);
S
Steve Hodgson 已提交
580
	bool (*poll) (struct efx_nic *efx);
581 582 583 584
	void (*get_settings) (struct efx_nic *efx,
			      struct ethtool_cmd *ecmd);
	int (*set_settings) (struct efx_nic *efx,
			     struct ethtool_cmd *ecmd);
585
	void (*set_npage_adv) (struct efx_nic *efx, u32);
586
	int (*test_alive) (struct efx_nic *efx);
587
	const char *(*test_name) (struct efx_nic *efx, unsigned int index);
588
	int (*run_tests) (struct efx_nic *efx, int *results, unsigned flags);
589 590 591 592 593
	int (*get_module_eeprom) (struct efx_nic *efx,
			       struct ethtool_eeprom *ee,
			       u8 *data);
	int (*get_module_info) (struct efx_nic *efx,
				struct ethtool_modinfo *modinfo);
594 595
};

596
/**
597
 * enum efx_phy_mode - PHY operating mode flags
598 599
 * @PHY_MODE_NORMAL: on and should pass traffic
 * @PHY_MODE_TX_DISABLED: on with TX disabled
600 601
 * @PHY_MODE_LOW_POWER: set to low power through MDIO
 * @PHY_MODE_OFF: switched off through external control
602 603 604 605 606
 * @PHY_MODE_SPECIAL: on but will not pass traffic
 */
enum efx_phy_mode {
	PHY_MODE_NORMAL		= 0,
	PHY_MODE_TX_DISABLED	= 1,
607 608
	PHY_MODE_LOW_POWER	= 2,
	PHY_MODE_OFF		= 4,
609 610 611 612 613
	PHY_MODE_SPECIAL	= 8,
};

static inline bool efx_phy_mode_disabled(enum efx_phy_mode mode)
{
B
Ben Hutchings 已提交
614
	return !!(mode & ~PHY_MODE_TX_DISABLED);
615 616
}

617 618 619 620 621 622
/**
 * struct efx_hw_stat_desc - Description of a hardware statistic
 * @name: Name of the statistic as visible through ethtool, or %NULL if
 *	it should not be exposed
 * @dma_width: Width in bits (0 for non-DMA statistics)
 * @offset: Offset within stats (ignored for non-DMA statistics)
623
 */
624 625 626 627
struct efx_hw_stat_desc {
	const char *name;
	u16 dma_width;
	u16 offset;
628 629 630 631 632 633 634 635 636 637 638 639 640 641
};

/* Number of bits used in a multicast filter hash address */
#define EFX_MCAST_HASH_BITS 8

/* Number of (single-bit) entries in a multicast filter hash */
#define EFX_MCAST_HASH_ENTRIES (1 << EFX_MCAST_HASH_BITS)

/* An Efx multicast filter hash */
union efx_multicast_hash {
	u8 byte[EFX_MCAST_HASH_ENTRIES / 8];
	efx_oword_t oword[EFX_MCAST_HASH_ENTRIES / sizeof(efx_oword_t) / 8];
};

642 643
struct efx_vf;
struct vfdi_status;
B
Ben Hutchings 已提交
644

645 646 647 648 649 650
/**
 * struct efx_nic - an Efx NIC
 * @name: Device name (net device name or bus id before net device registered)
 * @pci_dev: The PCI device
 * @type: Controller type attributes
 * @legacy_irq: IRQ number
651 652
 * @workqueue: Workqueue for port reconfigures and the HW monitor.
 *	Work items do not hold and must not acquire RTNL.
653
 * @workqueue_name: Name of workqueue
654 655 656 657
 * @reset_work: Scheduled reset workitem
 * @membase_phys: Memory BAR value as physical address
 * @membase: Memory BAR value
 * @interrupt_mode: Interrupt mode
658
 * @timer_quantum_ns: Interrupt timer quantum, in nanoseconds
659 660
 * @irq_rx_adaptive: Adaptive IRQ moderation enabled for RX event queues
 * @irq_rx_moderation: IRQ moderation time for RX event queues
661
 * @msg_enable: Log message enable flags
662
 * @state: Device state number (%STATE_*). Serialised by the rtnl_lock.
663
 * @reset_pending: Bitmask for pending resets
664 665 666
 * @tx_queue: TX DMA queues
 * @rx_queue: RX DMA queues
 * @channel: Channels
B
Ben Hutchings 已提交
667
 * @msi_context: Context for each MSI
668 669
 * @extra_channel_types: Types of extra (non-traffic) channels that
 *	should be allocated for this NIC
670 671
 * @rxq_entries: Size of receive queues requested by user.
 * @txq_entries: Size of transmit queues requested by user.
672 673
 * @txq_stop_thresh: TX queue fill level at or above which we stop it.
 * @txq_wake_thresh: TX queue fill level at or below which we wake it.
674 675 676
 * @tx_dc_base: Base qword address in SRAM of TX queue descriptor caches
 * @rx_dc_base: Base qword address in SRAM of RX queue descriptor caches
 * @sram_lim_qw: Qword address limit of SRAM
677
 * @next_buffer_table: First available buffer table id
678
 * @n_channels: Number of channels in use
B
Ben Hutchings 已提交
679 680
 * @n_rx_channels: Number of channels used for RX (= number of RX queues)
 * @n_tx_channels: Number of channels used for TX
681
 * @rx_dma_len: Current maximum RX DMA length
682
 * @rx_buffer_order: Order (log2) of number of pages for each RX buffer
683 684
 * @rx_buffer_truesize: Amortised allocation size of an RX buffer,
 *	for use in sk_buff::truesize
685 686 687
 * @rx_prefix_size: Size of RX prefix before packet data
 * @rx_packet_hash_offset: Offset of RX flow hash from start of packet data
 *	(valid only if @rx_prefix_size != 0; always negative)
688 689
 * @rx_packet_len_offset: Offset of RX packet length from start of packet data
 *	(valid only for NICs that set %EFX_RX_PKT_PREFIX_LEN; always negative)
690
 * @rx_hash_key: Toeplitz hash key for RSS
691
 * @rx_indir_table: Indirection table for RSS
692
 * @rx_scatter: Scatter mode enabled for receives
693 694
 * @int_error_count: Number of internal errors seen recently
 * @int_error_expire: Time at which error count will be expired
B
Ben Hutchings 已提交
695 696
 * @irq_soft_enabled: Are IRQs soft-enabled? If not, IRQ handler will
 *	acknowledge but do nothing else.
697
 * @irq_status: Interrupt status buffer
698
 * @irq_zero_count: Number of legacy IRQs seen with queue flags == 0
699
 * @irq_level: IRQ level/index for IRQs not triggered by an event queue
700
 * @selftest_work: Work item for asynchronous self-test
701
 * @mtd_list: List of MTDs attached to the NIC
L
Lucas De Marchi 已提交
702
 * @nic_data: Hardware dependent state
703
 * @mcdi: Management-Controller-to-Driver Interface state
B
Ben Hutchings 已提交
704
 * @mac_lock: MAC access lock. Protects @port_enabled, @phy_mode,
705
 *	efx_monitor() and efx_reconfigure_port()
706
 * @port_enabled: Port enabled indicator.
S
Steve Hodgson 已提交
707 708 709 710
 *	Serialises efx_stop_all(), efx_start_all(), efx_monitor() and
 *	efx_mac_work() with kernel interfaces. Safe to read under any
 *	one of the rtnl_lock, mac_lock, or netif_tx_lock, but all three must
 *	be held to modify it.
711 712 713 714 715 716
 * @port_initialized: Port initialized?
 * @net_dev: Operating system network device. Consider holding the rtnl lock
 * @stats_buffer: DMA buffer for statistics
 * @phy_type: PHY type
 * @phy_op: PHY interface
 * @phy_data: PHY private data (including PHY-specific stats)
717
 * @mdio: PHY MDIO interface
718
 * @mdio_bus: PHY MDIO bus ID (only used by Siena)
B
Ben Hutchings 已提交
719
 * @phy_mode: PHY operating mode. Serialised by @mac_lock.
B
Ben Hutchings 已提交
720
 * @link_advertising: Autonegotiation advertising flags
721
 * @link_state: Current state of the link
722
 * @n_link_state_changes: Number of times the link has changed state
723 724 725 726
 * @unicast_filter: Flag for Falcon-arch simple unicast filter.
 *	Protected by @mac_lock.
 * @multicast_hash: Multicast hash table for Falcon-arch.
 *	Protected by @mac_lock.
B
Ben Hutchings 已提交
727
 * @wanted_fc: Wanted flow control flags
728 729 730
 * @fc_disable: When non-zero flow control is disabled. Typically used to
 *	ensure that network back pressure doesn't delay dma queue flushes.
 *	Serialised by the rtnl lock.
731
 * @mac_work: Work item for changing MAC promiscuity and multicast hash
732 733 734
 * @loopback_mode: Loopback status
 * @loopback_modes: Supported loopback mode bitmask
 * @loopback_selftest: Offline self-test private state
735 736 737 738 739
 * @filter_lock: Filter table lock
 * @filter_state: Architecture-dependent filter table state
 * @rps_flow_id: Flow IDs of filters allocated for accelerated RFS,
 *	indexed by filter ID
 * @rps_expire_index: Next index to check for expiry in @rps_flow_id
740
 * @active_queues: Count of RX and TX queues that haven't been flushed and drained.
741 742 743 744 745 746
 * @rxq_flush_pending: Count of number of receive queues that need to be flushed.
 *	Decremented when the efx_flush_rx_queue() is called.
 * @rxq_flush_outstanding: Count of number of RX flushes started but not yet
 *	completed (either success or failure). Not used when MCDI is used to
 *	flush receive queues.
 * @flush_wq: wait queue used by efx_nic_flush_queues() to wait for flush completions.
747 748 749 750 751 752 753 754 755 756 757
 * @vf: Array of &struct efx_vf objects.
 * @vf_count: Number of VFs intended to be enabled.
 * @vf_init_count: Number of VFs that have been fully initialised.
 * @vi_scale: log2 number of vnics per VF.
 * @vf_buftbl_base: The zeroth buffer table index used to back VF queues.
 * @vfdi_status: Common VFDI status page to be dmad to VF address space.
 * @local_addr_list: List of local addresses. Protected by %local_lock.
 * @local_page_list: List of DMA addressable pages used to broadcast
 *	%local_addr_list. Protected by %local_lock.
 * @local_lock: Mutex protecting %local_addr_list and %local_page_list.
 * @peer_work: Work item to broadcast peer addresses to VMs.
758
 * @ptp_data: PTP state data
759 760
 * @monitor_work: Hardware monitor workitem
 * @biu_lock: BIU (bus interface unit) lock
761 762 763
 * @last_irq_cpu: Last CPU to handle a possible test interrupt.  This
 *	field is used by efx_test_interrupts() to verify that an
 *	interrupt has occurred.
764 765
 * @stats_lock: Statistics update lock. Must be held when calling
 *	efx_nic_type::{update,start,stop}_stats.
766
 *
767
 * This is stored in the private area of the &struct net_device.
768 769
 */
struct efx_nic {
770 771
	/* The following fields should be written very rarely */

772 773
	char name[IFNAMSIZ];
	struct pci_dev *pci_dev;
774
	unsigned int port_num;
775 776
	const struct efx_nic_type *type;
	int legacy_irq;
777
	bool eeh_disabled_legacy_irq;
778
	struct workqueue_struct *workqueue;
779
	char workqueue_name[16];
780
	struct work_struct reset_work;
781
	resource_size_t membase_phys;
782
	void __iomem *membase;
783

784
	enum efx_int_mode interrupt_mode;
785
	unsigned int timer_quantum_ns;
786 787
	bool irq_rx_adaptive;
	unsigned int irq_rx_moderation;
788
	u32 msg_enable;
789 790

	enum nic_state state;
791
	unsigned long reset_pending;
792

793
	struct efx_channel *channel[EFX_MAX_CHANNELS];
B
Ben Hutchings 已提交
794
	struct efx_msi_context msi_context[EFX_MAX_CHANNELS];
795 796
	const struct efx_channel_type *
	extra_channel_type[EFX_MAX_EXTRA_CHANNELS];
797

798 799
	unsigned rxq_entries;
	unsigned txq_entries;
800 801 802
	unsigned int txq_stop_thresh;
	unsigned int txq_wake_thresh;

803 804 805
	unsigned tx_dc_base;
	unsigned rx_dc_base;
	unsigned sram_lim_qw;
806
	unsigned next_buffer_table;
807 808

	unsigned int max_channels;
B
Ben Hutchings 已提交
809 810
	unsigned n_channels;
	unsigned n_rx_channels;
811
	unsigned rss_spread;
812
	unsigned tx_channel_offset;
B
Ben Hutchings 已提交
813
	unsigned n_tx_channels;
814
	unsigned int rx_dma_len;
815
	unsigned int rx_buffer_order;
816
	unsigned int rx_buffer_truesize;
817
	unsigned int rx_page_buf_step;
818
	unsigned int rx_bufs_per_page;
819
	unsigned int rx_pages_per_batch;
820 821
	unsigned int rx_prefix_size;
	int rx_packet_hash_offset;
822
	int rx_packet_len_offset;
823
	u8 rx_hash_key[40];
824
	u32 rx_indir_table[128];
825
	bool rx_scatter;
826

827 828 829
	unsigned int_error_count;
	unsigned long int_error_expire;

B
Ben Hutchings 已提交
830
	bool irq_soft_enabled;
831
	struct efx_buffer irq_status;
832
	unsigned irq_zero_count;
833
	unsigned irq_level;
834
	struct delayed_work selftest_work;
835

836 837 838
#ifdef CONFIG_SFC_MTD
	struct list_head mtd_list;
#endif
839

840
	void *nic_data;
841
	struct efx_mcdi_data *mcdi;
842 843

	struct mutex mac_lock;
844
	struct work_struct mac_work;
845
	bool port_enabled;
846

847
	bool port_initialized;
848 849 850 851
	struct net_device *net_dev;

	struct efx_buffer stats_buffer;

852
	unsigned int phy_type;
853
	const struct efx_phy_operations *phy_op;
854
	void *phy_data;
855
	struct mdio_if_info mdio;
856
	unsigned int mdio_bus;
857
	enum efx_phy_mode phy_mode;
858

B
Ben Hutchings 已提交
859
	u32 link_advertising;
860
	struct efx_link_state link_state;
861 862
	unsigned int n_link_state_changes;

863
	bool unicast_filter;
864
	union efx_multicast_hash multicast_hash;
865
	u8 wanted_fc;
866
	unsigned fc_disable;
867 868

	atomic_t rx_reset;
869
	enum efx_loopback_mode loopback_mode;
870
	u64 loopback_modes;
871 872

	void *loopback_selftest;
B
Ben Hutchings 已提交
873

874 875 876 877 878 879
	spinlock_t filter_lock;
	void *filter_state;
#ifdef CONFIG_RFS_ACCEL
	u32 *rps_flow_id;
	unsigned int rps_expire_index;
#endif
880

881
	atomic_t active_queues;
882 883 884 885
	atomic_t rxq_flush_pending;
	atomic_t rxq_flush_outstanding;
	wait_queue_head_t flush_wq;

886 887 888 889 890 891 892 893 894 895 896 897 898 899
#ifdef CONFIG_SFC_SRIOV
	struct efx_channel *vfdi_channel;
	struct efx_vf *vf;
	unsigned vf_count;
	unsigned vf_init_count;
	unsigned vi_scale;
	unsigned vf_buftbl_base;
	struct efx_buffer vfdi_status;
	struct list_head local_addr_list;
	struct list_head local_page_list;
	struct mutex local_lock;
	struct work_struct peer_work;
#endif

900 901
	struct efx_ptp_data *ptp_data;

902 903 904 905
	/* The following fields may be written more often */

	struct delayed_work monitor_work ____cacheline_aligned_in_smp;
	spinlock_t biu_lock;
906
	int last_irq_cpu;
907
	spinlock_t stats_lock;
908 909
};

910 911 912 913 914
static inline int efx_dev_registered(struct efx_nic *efx)
{
	return efx->net_dev->reg_state == NETREG_REGISTERED;
}

915 916
static inline unsigned int efx_port_num(struct efx_nic *efx)
{
917
	return efx->port_num;
918 919
}

920 921 922 923 924 925 926 927
struct efx_mtd_partition {
	struct list_head node;
	struct mtd_info mtd;
	const char *dev_type_name;
	const char *type_name;
	char name[IFNAMSIZ + 20];
};

928 929
/**
 * struct efx_nic_type - Efx device type definition
930
 * @mem_map_size: Get memory BAR mapped size
931 932 933
 * @probe: Probe the controller
 * @remove: Free resources allocated by probe()
 * @init: Initialise the controller
934 935
 * @dimension_resources: Dimension controller resources (buffer table,
 *	and VIs once the available interrupt resources are clear)
936 937
 * @fini: Shut down the controller
 * @monitor: Periodic function for polling link state and hardware monitor
938 939
 * @map_reset_reason: Map ethtool reset reason to a reset method
 * @map_reset_flags: Map ethtool reset flags to a reset method, if possible
940 941 942 943
 * @reset: Reset the controller hardware and possibly the PHY.  This will
 *	be called while the controller is uninitialised.
 * @probe_port: Probe the MAC and PHY
 * @remove_port: Free resources allocated by probe_port()
944
 * @handle_global_event: Handle a "global" event (may be %NULL)
945
 * @fini_dmaq: Flush and finalise DMA queues (RX and TX queues)
946
 * @prepare_flush: Prepare the hardware for flushing the DMA queues
947 948 949
 *	(for Falcon architecture)
 * @finish_flush: Clean up after flushing the DMA queues (for Falcon
 *	architecture)
950 951 952
 * @describe_stats: Describe statistics for ethtool
 * @update_stats: Update statistics not provided by event handling.
 *	Either argument may be %NULL.
953 954
 * @start_stats: Start the regular fetching of statistics
 * @stop_stats: Stop the regular fetching of statistics
955
 * @set_id_led: Set state of identifying LED or revert to automatic function
956
 * @push_irq_moderation: Apply interrupt moderation value
B
Ben Hutchings 已提交
957
 * @reconfigure_port: Push loopback/power/txdis changes to the MAC and PHY
958
 * @prepare_enable_fc_tx: Prepare MAC to enable pause frame TX (may be %NULL)
959 960
 * @reconfigure_mac: Push MAC address, MTU, flow control and filter settings
 *	to the hardware.  Serialised by the mac_lock.
961
 * @check_mac_fault: Check MAC fault state. True if fault present.
962 963 964
 * @get_wol: Get WoL configuration from driver state
 * @set_wol: Push WoL configuration to the NIC
 * @resume_wol: Synchronise WoL state between driver and MC (e.g. after resume)
965
 * @test_chip: Test registers.  May use efx_farch_test_registers(), and is
966
 *	expected to reset the NIC.
967
 * @test_nvram: Test validity of NVRAM contents
968 969 970 971 972 973 974 975 976 977 978
 * @mcdi_request: Send an MCDI request with the given header and SDU.
 *	The SDU length may be any value from 0 up to the protocol-
 *	defined maximum, but its buffer will be padded to a multiple
 *	of 4 bytes.
 * @mcdi_poll_response: Test whether an MCDI response is available.
 * @mcdi_read_response: Read the MCDI response PDU.  The offset will
 *	be a multiple of 4.  The length may not be, but the buffer
 *	will be padded so it is safe to round up.
 * @mcdi_poll_reboot: Test whether the MCDI has rebooted.  If so,
 *	return an appropriate error code for aborting any current
 *	request; otherwise return 0.
979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004
 * @irq_enable_master: Enable IRQs on the NIC.  Each event queue must
 *	be separately enabled after this.
 * @irq_test_generate: Generate a test IRQ
 * @irq_disable_non_ev: Disable non-event IRQs on the NIC.  Each event
 *	queue must be separately disabled before this.
 * @irq_handle_msi: Handle MSI for a channel.  The @dev_id argument is
 *	a pointer to the &struct efx_msi_context for the channel.
 * @irq_handle_legacy: Handle legacy interrupt.  The @dev_id argument
 *	is a pointer to the &struct efx_nic.
 * @tx_probe: Allocate resources for TX queue
 * @tx_init: Initialise TX queue on the NIC
 * @tx_remove: Free resources for TX queue
 * @tx_write: Write TX descriptors and doorbell
 * @rx_push_indir_table: Write RSS indirection table to the NIC
 * @rx_probe: Allocate resources for RX queue
 * @rx_init: Initialise RX queue on the NIC
 * @rx_remove: Free resources for RX queue
 * @rx_write: Write RX descriptors and doorbell
 * @rx_defer_refill: Generate a refill reminder event
 * @ev_probe: Allocate resources for event queue
 * @ev_init: Initialise event queue on the NIC
 * @ev_fini: Deinitialise event queue on the NIC
 * @ev_remove: Free resources for event queue
 * @ev_process: Process events for a queue, up to the given NAPI quota
 * @ev_read_ack: Acknowledge read events on a queue, rearming its IRQ
 * @ev_test_generate: Generate a test event
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022
 * @filter_table_probe: Probe filter capabilities and set up filter software state
 * @filter_table_restore: Restore filters removed from hardware
 * @filter_table_remove: Remove filters from hardware and tear down software state
 * @filter_update_rx_scatter: Update filters after change to rx scatter setting
 * @filter_insert: add or replace a filter
 * @filter_remove_safe: remove a filter by ID, carefully
 * @filter_get_safe: retrieve a filter by ID, carefully
 * @filter_clear_rx: remove RX filters by priority
 * @filter_count_rx_used: Get the number of filters in use at a given priority
 * @filter_get_rx_id_limit: Get maximum value of a filter id, plus 1
 * @filter_get_rx_ids: Get list of RX filters at a given priority
 * @filter_rfs_insert: Add or replace a filter for RFS.  This must be
 *	atomic.  The hardware change may be asynchronous but should
 *	not be delayed for long.  It may fail if this can't be done
 *	atomically.
 * @filter_rfs_expire_one: Consider expiring a filter inserted for RFS.
 *	This must check whether the specified table entry is used by RFS
 *	and that rps_may_expire_flow() returns true for it.
1023 1024 1025 1026 1027 1028 1029 1030 1031
 * @mtd_probe: Probe and add MTD partitions associated with this net device,
 *	 using efx_mtd_add()
 * @mtd_rename: Set an MTD partition name using the net device name
 * @mtd_read: Read from an MTD partition
 * @mtd_erase: Erase part of an MTD partition
 * @mtd_write: Write to an MTD partition
 * @mtd_sync: Wait for write-back to complete on MTD partition.  This
 *	also notifies the driver that a writer has finished using this
 *	partition.
1032
 * @revision: Hardware architecture revision
1033 1034 1035 1036 1037 1038
 * @txd_ptr_tbl_base: TX descriptor ring base address
 * @rxd_ptr_tbl_base: RX descriptor ring base address
 * @buf_tbl_base: Buffer table base address
 * @evq_ptr_tbl_base: Event queue pointer table base address
 * @evq_rptr_tbl_base: Event queue read-pointer table base address
 * @max_dma_mask: Maximum possible DMA mask
1039 1040
 * @rx_prefix_size: Size of RX prefix before packet data
 * @rx_hash_offset: Offset of RX flow hash within prefix
1041
 * @rx_buffer_padding: Size of padding at end of RX packet
J
Jon Cooper 已提交
1042 1043
 * @can_rx_scatter: NIC is able to scatter packets to multiple buffers
 * @always_rx_scatter: NIC will always scatter packets to multiple buffers
1044 1045
 * @max_interrupt_mode: Highest capability interrupt mode supported
 *	from &enum efx_init_mode.
1046
 * @timer_period_max: Maximum period of interrupt timer (in ticks)
1047 1048
 * @offload_features: net_device feature flags for protocol offload
 *	features implemented in hardware
B
Ben Hutchings 已提交
1049
 * @mcdi_max_ver: Maximum MCDI version supported
1050 1051
 */
struct efx_nic_type {
1052
	unsigned int (*mem_map_size)(struct efx_nic *efx);
1053 1054 1055
	int (*probe)(struct efx_nic *efx);
	void (*remove)(struct efx_nic *efx);
	int (*init)(struct efx_nic *efx);
1056
	int (*dimension_resources)(struct efx_nic *efx);
1057 1058
	void (*fini)(struct efx_nic *efx);
	void (*monitor)(struct efx_nic *efx);
1059 1060
	enum reset_type (*map_reset_reason)(enum reset_type reason);
	int (*map_reset_flags)(u32 *flags);
1061 1062 1063
	int (*reset)(struct efx_nic *efx, enum reset_type method);
	int (*probe_port)(struct efx_nic *efx);
	void (*remove_port)(struct efx_nic *efx);
1064
	bool (*handle_global_event)(struct efx_channel *channel, efx_qword_t *);
1065
	int (*fini_dmaq)(struct efx_nic *efx);
1066
	void (*prepare_flush)(struct efx_nic *efx);
1067
	void (*finish_flush)(struct efx_nic *efx);
1068 1069 1070
	size_t (*describe_stats)(struct efx_nic *efx, u8 *names);
	size_t (*update_stats)(struct efx_nic *efx, u64 *full_stats,
			       struct rtnl_link_stats64 *core_stats);
1071 1072
	void (*start_stats)(struct efx_nic *efx);
	void (*stop_stats)(struct efx_nic *efx);
1073
	void (*set_id_led)(struct efx_nic *efx, enum efx_led_mode mode);
1074
	void (*push_irq_moderation)(struct efx_channel *channel);
B
Ben Hutchings 已提交
1075
	int (*reconfigure_port)(struct efx_nic *efx);
1076
	void (*prepare_enable_fc_tx)(struct efx_nic *efx);
1077 1078
	int (*reconfigure_mac)(struct efx_nic *efx);
	bool (*check_mac_fault)(struct efx_nic *efx);
1079 1080 1081
	void (*get_wol)(struct efx_nic *efx, struct ethtool_wolinfo *wol);
	int (*set_wol)(struct efx_nic *efx, u32 type);
	void (*resume_wol)(struct efx_nic *efx);
1082
	int (*test_chip)(struct efx_nic *efx, struct efx_self_tests *tests);
1083
	int (*test_nvram)(struct efx_nic *efx);
1084 1085 1086 1087 1088 1089 1090
	void (*mcdi_request)(struct efx_nic *efx,
			     const efx_dword_t *hdr, size_t hdr_len,
			     const efx_dword_t *sdu, size_t sdu_len);
	bool (*mcdi_poll_response)(struct efx_nic *efx);
	void (*mcdi_read_response)(struct efx_nic *efx, efx_dword_t *pdu,
				   size_t pdu_offset, size_t pdu_len);
	int (*mcdi_poll_reboot)(struct efx_nic *efx);
1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
	void (*irq_enable_master)(struct efx_nic *efx);
	void (*irq_test_generate)(struct efx_nic *efx);
	void (*irq_disable_non_ev)(struct efx_nic *efx);
	irqreturn_t (*irq_handle_msi)(int irq, void *dev_id);
	irqreturn_t (*irq_handle_legacy)(int irq, void *dev_id);
	int (*tx_probe)(struct efx_tx_queue *tx_queue);
	void (*tx_init)(struct efx_tx_queue *tx_queue);
	void (*tx_remove)(struct efx_tx_queue *tx_queue);
	void (*tx_write)(struct efx_tx_queue *tx_queue);
	void (*rx_push_indir_table)(struct efx_nic *efx);
	int (*rx_probe)(struct efx_rx_queue *rx_queue);
	void (*rx_init)(struct efx_rx_queue *rx_queue);
	void (*rx_remove)(struct efx_rx_queue *rx_queue);
	void (*rx_write)(struct efx_rx_queue *rx_queue);
	void (*rx_defer_refill)(struct efx_rx_queue *rx_queue);
	int (*ev_probe)(struct efx_channel *channel);
1107
	int (*ev_init)(struct efx_channel *channel);
1108 1109 1110 1111 1112
	void (*ev_fini)(struct efx_channel *channel);
	void (*ev_remove)(struct efx_channel *channel);
	int (*ev_process)(struct efx_channel *channel, int quota);
	void (*ev_read_ack)(struct efx_channel *channel);
	void (*ev_test_generate)(struct efx_channel *channel);
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
	int (*filter_table_probe)(struct efx_nic *efx);
	void (*filter_table_restore)(struct efx_nic *efx);
	void (*filter_table_remove)(struct efx_nic *efx);
	void (*filter_update_rx_scatter)(struct efx_nic *efx);
	s32 (*filter_insert)(struct efx_nic *efx,
			     struct efx_filter_spec *spec, bool replace);
	int (*filter_remove_safe)(struct efx_nic *efx,
				  enum efx_filter_priority priority,
				  u32 filter_id);
	int (*filter_get_safe)(struct efx_nic *efx,
			       enum efx_filter_priority priority,
			       u32 filter_id, struct efx_filter_spec *);
	void (*filter_clear_rx)(struct efx_nic *efx,
				enum efx_filter_priority priority);
	u32 (*filter_count_rx_used)(struct efx_nic *efx,
				    enum efx_filter_priority priority);
	u32 (*filter_get_rx_id_limit)(struct efx_nic *efx);
	s32 (*filter_get_rx_ids)(struct efx_nic *efx,
				 enum efx_filter_priority priority,
				 u32 *buf, u32 size);
#ifdef CONFIG_RFS_ACCEL
	s32 (*filter_rfs_insert)(struct efx_nic *efx,
				 struct efx_filter_spec *spec);
	bool (*filter_rfs_expire_one)(struct efx_nic *efx, u32 flow_id,
				      unsigned int index);
#endif
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
#ifdef CONFIG_SFC_MTD
	int (*mtd_probe)(struct efx_nic *efx);
	void (*mtd_rename)(struct efx_mtd_partition *part);
	int (*mtd_read)(struct mtd_info *mtd, loff_t start, size_t len,
			size_t *retlen, u8 *buffer);
	int (*mtd_erase)(struct mtd_info *mtd, loff_t start, size_t len);
	int (*mtd_write)(struct mtd_info *mtd, loff_t start, size_t len,
			 size_t *retlen, const u8 *buffer);
	int (*mtd_sync)(struct mtd_info *mtd);
#endif
1149
	void (*ptp_write_host_time)(struct efx_nic *efx, u32 host_time);
1150

1151
	int revision;
1152 1153 1154 1155 1156
	unsigned int txd_ptr_tbl_base;
	unsigned int rxd_ptr_tbl_base;
	unsigned int buf_tbl_base;
	unsigned int evq_ptr_tbl_base;
	unsigned int evq_rptr_tbl_base;
1157
	u64 max_dma_mask;
1158 1159
	unsigned int rx_prefix_size;
	unsigned int rx_hash_offset;
1160
	unsigned int rx_buffer_padding;
1161
	bool can_rx_scatter;
J
Jon Cooper 已提交
1162
	bool always_rx_scatter;
1163
	unsigned int max_interrupt_mode;
1164
	unsigned int timer_period_max;
1165
	netdev_features_t offload_features;
B
Ben Hutchings 已提交
1166
	int mcdi_max_ver;
1167
	unsigned int max_rx_ip_filters;
1168 1169 1170 1171 1172 1173 1174 1175
};

/**************************************************************************
 *
 * Prototypes and inline functions
 *
 *************************************************************************/

1176 1177 1178 1179
static inline struct efx_channel *
efx_get_channel(struct efx_nic *efx, unsigned index)
{
	EFX_BUG_ON_PARANOID(index >= efx->n_channels);
1180
	return efx->channel[index];
1181 1182
}

1183 1184
/* Iterate over all used channels */
#define efx_for_each_channel(_channel, _efx)				\
1185 1186 1187 1188
	for (_channel = (_efx)->channel[0];				\
	     _channel;							\
	     _channel = (_channel->channel + 1 < (_efx)->n_channels) ?	\
		     (_efx)->channel[_channel->channel + 1] : NULL)
1189

1190 1191 1192 1193 1194 1195 1196
/* Iterate over all used channels in reverse */
#define efx_for_each_channel_rev(_channel, _efx)			\
	for (_channel = (_efx)->channel[(_efx)->n_channels - 1];	\
	     _channel;							\
	     _channel = _channel->channel ?				\
		     (_efx)->channel[_channel->channel - 1] : NULL)

1197 1198 1199 1200 1201 1202 1203
static inline struct efx_tx_queue *
efx_get_tx_queue(struct efx_nic *efx, unsigned index, unsigned type)
{
	EFX_BUG_ON_PARANOID(index >= efx->n_tx_channels ||
			    type >= EFX_TXQ_TYPES);
	return &efx->channel[efx->tx_channel_offset + index]->tx_queue[type];
}
1204

1205 1206 1207 1208 1209 1210
static inline bool efx_channel_has_tx_queues(struct efx_channel *channel)
{
	return channel->channel - channel->efx->tx_channel_offset <
		channel->efx->n_tx_channels;
}

1211 1212 1213
static inline struct efx_tx_queue *
efx_channel_get_tx_queue(struct efx_channel *channel, unsigned type)
{
1214 1215 1216
	EFX_BUG_ON_PARANOID(!efx_channel_has_tx_queues(channel) ||
			    type >= EFX_TXQ_TYPES);
	return &channel->tx_queue[type];
1217
}
1218

1219 1220 1221 1222 1223 1224
static inline bool efx_tx_queue_used(struct efx_tx_queue *tx_queue)
{
	return !(tx_queue->efx->net_dev->num_tc < 2 &&
		 tx_queue->queue & EFX_TXQ_TYPE_HIGHPRI);
}

1225 1226
/* Iterate over all TX queues belonging to a channel */
#define efx_for_each_channel_tx_queue(_tx_queue, _channel)		\
1227 1228 1229 1230
	if (!efx_channel_has_tx_queues(_channel))			\
		;							\
	else								\
		for (_tx_queue = (_channel)->tx_queue;			\
1231 1232
		     _tx_queue < (_channel)->tx_queue + EFX_TXQ_TYPES && \
			     efx_tx_queue_used(_tx_queue);		\
1233
		     _tx_queue++)
1234

1235 1236
/* Iterate over all possible TX queues belonging to a channel */
#define efx_for_each_possible_channel_tx_queue(_tx_queue, _channel)	\
1237 1238 1239 1240 1241 1242
	if (!efx_channel_has_tx_queues(_channel))			\
		;							\
	else								\
		for (_tx_queue = (_channel)->tx_queue;			\
		     _tx_queue < (_channel)->tx_queue + EFX_TXQ_TYPES;	\
		     _tx_queue++)
1243

1244 1245
static inline bool efx_channel_has_rx_queue(struct efx_channel *channel)
{
1246
	return channel->rx_queue.core_index >= 0;
1247 1248
}

1249 1250 1251
static inline struct efx_rx_queue *
efx_channel_get_rx_queue(struct efx_channel *channel)
{
1252 1253
	EFX_BUG_ON_PARANOID(!efx_channel_has_rx_queue(channel));
	return &channel->rx_queue;
1254 1255
}

1256 1257
/* Iterate over all RX queues belonging to a channel */
#define efx_for_each_channel_rx_queue(_rx_queue, _channel)		\
1258 1259 1260 1261 1262 1263
	if (!efx_channel_has_rx_queue(_channel))			\
		;							\
	else								\
		for (_rx_queue = &(_channel)->rx_queue;			\
		     _rx_queue;						\
		     _rx_queue = NULL)
1264

1265 1266 1267
static inline struct efx_channel *
efx_rx_queue_channel(struct efx_rx_queue *rx_queue)
{
1268
	return container_of(rx_queue, struct efx_channel, rx_queue);
1269 1270 1271 1272
}

static inline int efx_rx_queue_index(struct efx_rx_queue *rx_queue)
{
1273
	return efx_rx_queue_channel(rx_queue)->channel;
1274 1275
}

1276 1277 1278 1279 1280 1281
/* Returns a pointer to the specified receive buffer in the RX
 * descriptor queue.
 */
static inline struct efx_rx_buffer *efx_rx_buffer(struct efx_rx_queue *rx_queue,
						  unsigned int index)
{
1282
	return &rx_queue->buffer[index];
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
}


/**
 * EFX_MAX_FRAME_LEN - calculate maximum frame length
 *
 * This calculates the maximum frame length that will be used for a
 * given MTU.  The frame length will be equal to the MTU plus a
 * constant amount of header space and padding.  This is the quantity
 * that the net driver will program into the MAC as the maximum frame
 * length.
 *
1295
 * The 10G MAC requires 8-byte alignment on the frame
1296
 * length, so we round up to the nearest 8.
1297 1298 1299 1300 1301
 *
 * Re-clocking by the XGXS on RX can reduce an IPG to 32 bits (half an
 * XGMII cycle).  If the frame length reaches the maximum value in the
 * same cycle, the XMAC can miss the IPG altogether.  We work around
 * this by adding a further 16 bytes.
1302 1303
 */
#define EFX_MAX_FRAME_LEN(mtu) \
1304
	((((mtu) + ETH_HLEN + VLAN_HLEN + 4/* FCS */ + 7) & ~7) + 16)
1305

1306 1307 1308 1309 1310 1311 1312 1313
static inline bool efx_xmit_with_hwtstamp(struct sk_buff *skb)
{
	return skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP;
}
static inline void efx_xmit_hwtstamp_pending(struct sk_buff *skb)
{
	skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
}
1314 1315

#endif /* EFX_NET_DRIVER_H */