xref: /freebsd/sys/net/iflib.c (revision dd41de95a84d979615a2ef11df6850622bf6184e)
1 /*-
2  * Copyright (c) 2014-2018, Matthew Macy <mmacy@mattmacy.io>
3  * All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions are met:
7  *
8  *  1. Redistributions of source code must retain the above copyright notice,
9  *     this list of conditions and the following disclaimer.
10  *
11  *  2. Neither the name of Matthew Macy nor the names of its
12  *     contributors may be used to endorse or promote products derived from
13  *     this software without specific prior written permission.
14  *
15  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
16  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
19  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
20  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
21  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
22  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
23  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
24  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
25  * POSSIBILITY OF SUCH DAMAGE.
26  */
27 
28 #include <sys/cdefs.h>
29 __FBSDID("$FreeBSD$");
30 
31 #include "opt_inet.h"
32 #include "opt_inet6.h"
33 #include "opt_acpi.h"
34 #include "opt_sched.h"
35 
36 #include <sys/param.h>
37 #include <sys/types.h>
38 #include <sys/bus.h>
39 #include <sys/eventhandler.h>
40 #include <sys/kernel.h>
41 #include <sys/lock.h>
42 #include <sys/mutex.h>
43 #include <sys/module.h>
44 #include <sys/kobj.h>
45 #include <sys/rman.h>
46 #include <sys/sbuf.h>
47 #include <sys/smp.h>
48 #include <sys/socket.h>
49 #include <sys/sockio.h>
50 #include <sys/sysctl.h>
51 #include <sys/syslog.h>
52 #include <sys/taskqueue.h>
53 #include <sys/limits.h>
54 
55 #include <net/if.h>
56 #include <net/if_var.h>
57 #include <net/if_types.h>
58 #include <net/if_media.h>
59 #include <net/bpf.h>
60 #include <net/ethernet.h>
61 #include <net/mp_ring.h>
62 #include <net/debugnet.h>
63 #include <net/pfil.h>
64 #include <net/vnet.h>
65 
66 #include <netinet/in.h>
67 #include <netinet/in_pcb.h>
68 #include <netinet/tcp_lro.h>
69 #include <netinet/in_systm.h>
70 #include <netinet/if_ether.h>
71 #include <netinet/ip.h>
72 #include <netinet/ip6.h>
73 #include <netinet/tcp.h>
74 #include <netinet/ip_var.h>
75 #include <netinet6/ip6_var.h>
76 
77 #include <machine/bus.h>
78 #include <machine/in_cksum.h>
79 
80 #include <vm/vm.h>
81 #include <vm/pmap.h>
82 
83 #include <dev/led/led.h>
84 #include <dev/pci/pcireg.h>
85 #include <dev/pci/pcivar.h>
86 #include <dev/pci/pci_private.h>
87 
88 #include <net/iflib.h>
89 #include <net/iflib_private.h>
90 
91 #include "ifdi_if.h"
92 
93 #ifdef PCI_IOV
94 #include <dev/pci/pci_iov.h>
95 #endif
96 
97 #include <sys/bitstring.h>
98 /*
99  * enable accounting of every mbuf as it comes in to and goes out of
100  * iflib's software descriptor references
101  */
102 #define MEMORY_LOGGING 0
103 /*
104  * Enable mbuf vectors for compressing long mbuf chains
105  */
106 
107 /*
108  * NB:
109  * - Prefetching in tx cleaning should perhaps be a tunable. The distance ahead
110  *   we prefetch needs to be determined by the time spent in m_free vis a vis
111  *   the cost of a prefetch. This will of course vary based on the workload:
112  *      - NFLX's m_free path is dominated by vm-based M_EXT manipulation which
113  *        is quite expensive, thus suggesting very little prefetch.
114  *      - small packet forwarding which is just returning a single mbuf to
115  *        UMA will typically be very fast vis a vis the cost of a memory
116  *        access.
117  */
118 
119 /*
120  * File organization:
121  *  - private structures
122  *  - iflib private utility functions
123  *  - ifnet functions
124  *  - vlan registry and other exported functions
125  *  - iflib public core functions
126  *
127  *
128  */
129 MALLOC_DEFINE(M_IFLIB, "iflib", "ifnet library");
130 
131 #define	IFLIB_RXEOF_MORE (1U << 0)
132 #define	IFLIB_RXEOF_EMPTY (2U << 0)
133 
134 struct iflib_txq;
135 typedef struct iflib_txq *iflib_txq_t;
136 struct iflib_rxq;
137 typedef struct iflib_rxq *iflib_rxq_t;
138 struct iflib_fl;
139 typedef struct iflib_fl *iflib_fl_t;
140 
141 struct iflib_ctx;
142 
143 static void iru_init(if_rxd_update_t iru, iflib_rxq_t rxq, uint8_t flid);
144 static void iflib_timer(void *arg);
145 static void iflib_tqg_detach(if_ctx_t ctx);
146 
147 typedef struct iflib_filter_info {
148 	driver_filter_t *ifi_filter;
149 	void *ifi_filter_arg;
150 	struct grouptask *ifi_task;
151 	void *ifi_ctx;
152 } *iflib_filter_info_t;
153 
154 struct iflib_ctx {
155 	KOBJ_FIELDS;
156 	/*
157 	 * Pointer to hardware driver's softc
158 	 */
159 	void *ifc_softc;
160 	device_t ifc_dev;
161 	if_t ifc_ifp;
162 
163 	cpuset_t ifc_cpus;
164 	if_shared_ctx_t ifc_sctx;
165 	struct if_softc_ctx ifc_softc_ctx;
166 
167 	struct sx ifc_ctx_sx;
168 	struct mtx ifc_state_mtx;
169 
170 	iflib_txq_t ifc_txqs;
171 	iflib_rxq_t ifc_rxqs;
172 	uint32_t ifc_if_flags;
173 	uint32_t ifc_flags;
174 	uint32_t ifc_max_fl_buf_size;
175 	uint32_t ifc_rx_mbuf_sz;
176 
177 	int ifc_link_state;
178 	int ifc_watchdog_events;
179 	struct cdev *ifc_led_dev;
180 	struct resource *ifc_msix_mem;
181 
182 	struct if_irq ifc_legacy_irq;
183 	struct grouptask ifc_admin_task;
184 	struct grouptask ifc_vflr_task;
185 	struct iflib_filter_info ifc_filter_info;
186 	struct ifmedia	ifc_media;
187 	struct ifmedia	*ifc_mediap;
188 
189 	struct sysctl_oid *ifc_sysctl_node;
190 	uint16_t ifc_sysctl_ntxqs;
191 	uint16_t ifc_sysctl_nrxqs;
192 	uint16_t ifc_sysctl_qs_eq_override;
193 	uint16_t ifc_sysctl_rx_budget;
194 	uint16_t ifc_sysctl_tx_abdicate;
195 	uint16_t ifc_sysctl_core_offset;
196 #define	CORE_OFFSET_UNSPECIFIED	0xffff
197 	uint8_t  ifc_sysctl_separate_txrx;
198 	uint8_t  ifc_sysctl_use_logical_cores;
199 	bool	 ifc_cpus_are_physical_cores;
200 
201 	qidx_t ifc_sysctl_ntxds[8];
202 	qidx_t ifc_sysctl_nrxds[8];
203 	struct if_txrx ifc_txrx;
204 #define isc_txd_encap  ifc_txrx.ift_txd_encap
205 #define isc_txd_flush  ifc_txrx.ift_txd_flush
206 #define isc_txd_credits_update  ifc_txrx.ift_txd_credits_update
207 #define isc_rxd_available ifc_txrx.ift_rxd_available
208 #define isc_rxd_pkt_get ifc_txrx.ift_rxd_pkt_get
209 #define isc_rxd_refill ifc_txrx.ift_rxd_refill
210 #define isc_rxd_flush ifc_txrx.ift_rxd_flush
211 #define isc_legacy_intr ifc_txrx.ift_legacy_intr
212 	eventhandler_tag ifc_vlan_attach_event;
213 	eventhandler_tag ifc_vlan_detach_event;
214 	struct ether_addr ifc_mac;
215 };
216 
217 void *
218 iflib_get_softc(if_ctx_t ctx)
219 {
220 
221 	return (ctx->ifc_softc);
222 }
223 
224 device_t
225 iflib_get_dev(if_ctx_t ctx)
226 {
227 
228 	return (ctx->ifc_dev);
229 }
230 
231 if_t
232 iflib_get_ifp(if_ctx_t ctx)
233 {
234 
235 	return (ctx->ifc_ifp);
236 }
237 
238 struct ifmedia *
239 iflib_get_media(if_ctx_t ctx)
240 {
241 
242 	return (ctx->ifc_mediap);
243 }
244 
245 uint32_t
246 iflib_get_flags(if_ctx_t ctx)
247 {
248 	return (ctx->ifc_flags);
249 }
250 
251 void
252 iflib_set_mac(if_ctx_t ctx, uint8_t mac[ETHER_ADDR_LEN])
253 {
254 
255 	bcopy(mac, ctx->ifc_mac.octet, ETHER_ADDR_LEN);
256 }
257 
258 if_softc_ctx_t
259 iflib_get_softc_ctx(if_ctx_t ctx)
260 {
261 
262 	return (&ctx->ifc_softc_ctx);
263 }
264 
265 if_shared_ctx_t
266 iflib_get_sctx(if_ctx_t ctx)
267 {
268 
269 	return (ctx->ifc_sctx);
270 }
271 
272 #define IP_ALIGNED(m) ((((uintptr_t)(m)->m_data) & 0x3) == 0x2)
273 #define CACHE_PTR_INCREMENT (CACHE_LINE_SIZE/sizeof(void*))
274 #define CACHE_PTR_NEXT(ptr) ((void *)(((uintptr_t)(ptr)+CACHE_LINE_SIZE-1) & (CACHE_LINE_SIZE-1)))
275 
276 #define LINK_ACTIVE(ctx) ((ctx)->ifc_link_state == LINK_STATE_UP)
277 #define CTX_IS_VF(ctx) ((ctx)->ifc_sctx->isc_flags & IFLIB_IS_VF)
278 
279 typedef struct iflib_sw_rx_desc_array {
280 	bus_dmamap_t	*ifsd_map;         /* bus_dma maps for packet */
281 	struct mbuf	**ifsd_m;           /* pkthdr mbufs */
282 	caddr_t		*ifsd_cl;          /* direct cluster pointer for rx */
283 	bus_addr_t	*ifsd_ba;          /* bus addr of cluster for rx */
284 } iflib_rxsd_array_t;
285 
286 typedef struct iflib_sw_tx_desc_array {
287 	bus_dmamap_t    *ifsd_map;         /* bus_dma maps for packet */
288 	bus_dmamap_t	*ifsd_tso_map;     /* bus_dma maps for TSO packet */
289 	struct mbuf    **ifsd_m;           /* pkthdr mbufs */
290 } if_txsd_vec_t;
291 
292 /* magic number that should be high enough for any hardware */
293 #define IFLIB_MAX_TX_SEGS		128
294 #define IFLIB_RX_COPY_THRESH		128
295 #define IFLIB_MAX_RX_REFRESH		32
296 /* The minimum descriptors per second before we start coalescing */
297 #define IFLIB_MIN_DESC_SEC		16384
298 #define IFLIB_DEFAULT_TX_UPDATE_FREQ	16
299 #define IFLIB_QUEUE_IDLE		0
300 #define IFLIB_QUEUE_HUNG		1
301 #define IFLIB_QUEUE_WORKING		2
302 /* maximum number of txqs that can share an rx interrupt */
303 #define IFLIB_MAX_TX_SHARED_INTR	4
304 
305 /* this should really scale with ring size - this is a fairly arbitrary value */
306 #define TX_BATCH_SIZE			32
307 
308 #define IFLIB_RESTART_BUDGET		8
309 
310 #define CSUM_OFFLOAD		(CSUM_IP_TSO|CSUM_IP6_TSO|CSUM_IP| \
311 				 CSUM_IP_UDP|CSUM_IP_TCP|CSUM_IP_SCTP| \
312 				 CSUM_IP6_UDP|CSUM_IP6_TCP|CSUM_IP6_SCTP)
313 
314 struct iflib_txq {
315 	qidx_t		ift_in_use;
316 	qidx_t		ift_cidx;
317 	qidx_t		ift_cidx_processed;
318 	qidx_t		ift_pidx;
319 	uint8_t		ift_gen;
320 	uint8_t		ift_br_offset;
321 	uint16_t	ift_npending;
322 	uint16_t	ift_db_pending;
323 	uint16_t	ift_rs_pending;
324 	/* implicit pad */
325 	uint8_t		ift_txd_size[8];
326 	uint64_t	ift_processed;
327 	uint64_t	ift_cleaned;
328 	uint64_t	ift_cleaned_prev;
329 #if MEMORY_LOGGING
330 	uint64_t	ift_enqueued;
331 	uint64_t	ift_dequeued;
332 #endif
333 	uint64_t	ift_no_tx_dma_setup;
334 	uint64_t	ift_no_desc_avail;
335 	uint64_t	ift_mbuf_defrag_failed;
336 	uint64_t	ift_mbuf_defrag;
337 	uint64_t	ift_map_failed;
338 	uint64_t	ift_txd_encap_efbig;
339 	uint64_t	ift_pullups;
340 	uint64_t	ift_last_timer_tick;
341 
342 	struct mtx	ift_mtx;
343 	struct mtx	ift_db_mtx;
344 
345 	/* constant values */
346 	if_ctx_t	ift_ctx;
347 	struct ifmp_ring        *ift_br;
348 	struct grouptask	ift_task;
349 	qidx_t		ift_size;
350 	uint16_t	ift_id;
351 	struct callout	ift_timer;
352 #ifdef DEV_NETMAP
353 	struct callout	ift_netmap_timer;
354 #endif /* DEV_NETMAP */
355 
356 	if_txsd_vec_t	ift_sds;
357 	uint8_t		ift_qstatus;
358 	uint8_t		ift_closed;
359 	uint8_t		ift_update_freq;
360 	struct iflib_filter_info ift_filter_info;
361 	bus_dma_tag_t	ift_buf_tag;
362 	bus_dma_tag_t	ift_tso_buf_tag;
363 	iflib_dma_info_t	ift_ifdi;
364 #define	MTX_NAME_LEN	32
365 	char                    ift_mtx_name[MTX_NAME_LEN];
366 	bus_dma_segment_t	ift_segs[IFLIB_MAX_TX_SEGS]  __aligned(CACHE_LINE_SIZE);
367 #ifdef IFLIB_DIAGNOSTICS
368 	uint64_t ift_cpu_exec_count[256];
369 #endif
370 } __aligned(CACHE_LINE_SIZE);
371 
372 struct iflib_fl {
373 	qidx_t		ifl_cidx;
374 	qidx_t		ifl_pidx;
375 	qidx_t		ifl_credits;
376 	uint8_t		ifl_gen;
377 	uint8_t		ifl_rxd_size;
378 #if MEMORY_LOGGING
379 	uint64_t	ifl_m_enqueued;
380 	uint64_t	ifl_m_dequeued;
381 	uint64_t	ifl_cl_enqueued;
382 	uint64_t	ifl_cl_dequeued;
383 #endif
384 	/* implicit pad */
385 	bitstr_t 	*ifl_rx_bitmap;
386 	qidx_t		ifl_fragidx;
387 	/* constant */
388 	qidx_t		ifl_size;
389 	uint16_t	ifl_buf_size;
390 	uint16_t	ifl_cltype;
391 	uma_zone_t	ifl_zone;
392 	iflib_rxsd_array_t	ifl_sds;
393 	iflib_rxq_t	ifl_rxq;
394 	uint8_t		ifl_id;
395 	bus_dma_tag_t	ifl_buf_tag;
396 	iflib_dma_info_t	ifl_ifdi;
397 	uint64_t	ifl_bus_addrs[IFLIB_MAX_RX_REFRESH] __aligned(CACHE_LINE_SIZE);
398 	qidx_t		ifl_rxd_idxs[IFLIB_MAX_RX_REFRESH];
399 }  __aligned(CACHE_LINE_SIZE);
400 
401 static inline qidx_t
402 get_inuse(int size, qidx_t cidx, qidx_t pidx, uint8_t gen)
403 {
404 	qidx_t used;
405 
406 	if (pidx > cidx)
407 		used = pidx - cidx;
408 	else if (pidx < cidx)
409 		used = size - cidx + pidx;
410 	else if (gen == 0 && pidx == cidx)
411 		used = 0;
412 	else if (gen == 1 && pidx == cidx)
413 		used = size;
414 	else
415 		panic("bad state");
416 
417 	return (used);
418 }
419 
420 #define TXQ_AVAIL(txq) (txq->ift_size - get_inuse(txq->ift_size, txq->ift_cidx, txq->ift_pidx, txq->ift_gen))
421 
422 #define IDXDIFF(head, tail, wrap) \
423 	((head) >= (tail) ? (head) - (tail) : (wrap) - (tail) + (head))
424 
425 struct iflib_rxq {
426 	if_ctx_t	ifr_ctx;
427 	iflib_fl_t	ifr_fl;
428 	uint64_t	ifr_rx_irq;
429 	struct pfil_head	*pfil;
430 	/*
431 	 * If there is a separate completion queue (IFLIB_HAS_RXCQ), this is
432 	 * the completion queue consumer index.  Otherwise it's unused.
433 	 */
434 	qidx_t		ifr_cq_cidx;
435 	uint16_t	ifr_id;
436 	uint8_t		ifr_nfl;
437 	uint8_t		ifr_ntxqirq;
438 	uint8_t		ifr_txqid[IFLIB_MAX_TX_SHARED_INTR];
439 	uint8_t		ifr_fl_offset;
440 	struct lro_ctrl			ifr_lc;
441 	struct grouptask        ifr_task;
442 	struct callout		ifr_watchdog;
443 	struct iflib_filter_info ifr_filter_info;
444 	iflib_dma_info_t		ifr_ifdi;
445 
446 	/* dynamically allocate if any drivers need a value substantially larger than this */
447 	struct if_rxd_frag	ifr_frags[IFLIB_MAX_RX_SEGS] __aligned(CACHE_LINE_SIZE);
448 #ifdef IFLIB_DIAGNOSTICS
449 	uint64_t ifr_cpu_exec_count[256];
450 #endif
451 }  __aligned(CACHE_LINE_SIZE);
452 
453 typedef struct if_rxsd {
454 	caddr_t *ifsd_cl;
455 	iflib_fl_t ifsd_fl;
456 } *if_rxsd_t;
457 
458 /* multiple of word size */
459 #ifdef __LP64__
460 #define PKT_INFO_SIZE	6
461 #define RXD_INFO_SIZE	5
462 #define PKT_TYPE uint64_t
463 #else
464 #define PKT_INFO_SIZE	11
465 #define RXD_INFO_SIZE	8
466 #define PKT_TYPE uint32_t
467 #endif
468 #define PKT_LOOP_BOUND  ((PKT_INFO_SIZE/3)*3)
469 #define RXD_LOOP_BOUND  ((RXD_INFO_SIZE/4)*4)
470 
471 typedef struct if_pkt_info_pad {
472 	PKT_TYPE pkt_val[PKT_INFO_SIZE];
473 } *if_pkt_info_pad_t;
474 typedef struct if_rxd_info_pad {
475 	PKT_TYPE rxd_val[RXD_INFO_SIZE];
476 } *if_rxd_info_pad_t;
477 
478 CTASSERT(sizeof(struct if_pkt_info_pad) == sizeof(struct if_pkt_info));
479 CTASSERT(sizeof(struct if_rxd_info_pad) == sizeof(struct if_rxd_info));
480 
481 static inline void
482 pkt_info_zero(if_pkt_info_t pi)
483 {
484 	if_pkt_info_pad_t pi_pad;
485 
486 	pi_pad = (if_pkt_info_pad_t)pi;
487 	pi_pad->pkt_val[0] = 0; pi_pad->pkt_val[1] = 0; pi_pad->pkt_val[2] = 0;
488 	pi_pad->pkt_val[3] = 0; pi_pad->pkt_val[4] = 0; pi_pad->pkt_val[5] = 0;
489 #ifndef __LP64__
490 	pi_pad->pkt_val[6] = 0; pi_pad->pkt_val[7] = 0; pi_pad->pkt_val[8] = 0;
491 	pi_pad->pkt_val[9] = 0; pi_pad->pkt_val[10] = 0;
492 #endif
493 }
494 
495 static device_method_t iflib_pseudo_methods[] = {
496 	DEVMETHOD(device_attach, noop_attach),
497 	DEVMETHOD(device_detach, iflib_pseudo_detach),
498 	DEVMETHOD_END
499 };
500 
501 driver_t iflib_pseudodriver = {
502 	"iflib_pseudo", iflib_pseudo_methods, sizeof(struct iflib_ctx),
503 };
504 
505 static inline void
506 rxd_info_zero(if_rxd_info_t ri)
507 {
508 	if_rxd_info_pad_t ri_pad;
509 	int i;
510 
511 	ri_pad = (if_rxd_info_pad_t)ri;
512 	for (i = 0; i < RXD_LOOP_BOUND; i += 4) {
513 		ri_pad->rxd_val[i] = 0;
514 		ri_pad->rxd_val[i+1] = 0;
515 		ri_pad->rxd_val[i+2] = 0;
516 		ri_pad->rxd_val[i+3] = 0;
517 	}
518 #ifdef __LP64__
519 	ri_pad->rxd_val[RXD_INFO_SIZE-1] = 0;
520 #endif
521 }
522 
523 /*
524  * Only allow a single packet to take up most 1/nth of the tx ring
525  */
526 #define MAX_SINGLE_PACKET_FRACTION 12
527 #define IF_BAD_DMA (bus_addr_t)-1
528 
529 #define CTX_ACTIVE(ctx) ((if_getdrvflags((ctx)->ifc_ifp) & IFF_DRV_RUNNING))
530 
531 #define CTX_LOCK_INIT(_sc)  sx_init(&(_sc)->ifc_ctx_sx, "iflib ctx lock")
532 #define CTX_LOCK(ctx) sx_xlock(&(ctx)->ifc_ctx_sx)
533 #define CTX_UNLOCK(ctx) sx_xunlock(&(ctx)->ifc_ctx_sx)
534 #define CTX_LOCK_DESTROY(ctx) sx_destroy(&(ctx)->ifc_ctx_sx)
535 
536 #define STATE_LOCK_INIT(_sc, _name)  mtx_init(&(_sc)->ifc_state_mtx, _name, "iflib state lock", MTX_DEF)
537 #define STATE_LOCK(ctx) mtx_lock(&(ctx)->ifc_state_mtx)
538 #define STATE_UNLOCK(ctx) mtx_unlock(&(ctx)->ifc_state_mtx)
539 #define STATE_LOCK_DESTROY(ctx) mtx_destroy(&(ctx)->ifc_state_mtx)
540 
541 #define CALLOUT_LOCK(txq)	mtx_lock(&txq->ift_mtx)
542 #define CALLOUT_UNLOCK(txq) 	mtx_unlock(&txq->ift_mtx)
543 
544 void
545 iflib_set_detach(if_ctx_t ctx)
546 {
547 	STATE_LOCK(ctx);
548 	ctx->ifc_flags |= IFC_IN_DETACH;
549 	STATE_UNLOCK(ctx);
550 }
551 
552 /* Our boot-time initialization hook */
553 static int	iflib_module_event_handler(module_t, int, void *);
554 
555 static moduledata_t iflib_moduledata = {
556 	"iflib",
557 	iflib_module_event_handler,
558 	NULL
559 };
560 
561 DECLARE_MODULE(iflib, iflib_moduledata, SI_SUB_INIT_IF, SI_ORDER_ANY);
562 MODULE_VERSION(iflib, 1);
563 
564 MODULE_DEPEND(iflib, pci, 1, 1, 1);
565 MODULE_DEPEND(iflib, ether, 1, 1, 1);
566 
567 TASKQGROUP_DEFINE(if_io_tqg, mp_ncpus, 1);
568 TASKQGROUP_DEFINE(if_config_tqg, 1, 1);
569 
570 #ifndef IFLIB_DEBUG_COUNTERS
571 #ifdef INVARIANTS
572 #define IFLIB_DEBUG_COUNTERS 1
573 #else
574 #define IFLIB_DEBUG_COUNTERS 0
575 #endif /* !INVARIANTS */
576 #endif
577 
578 static SYSCTL_NODE(_net, OID_AUTO, iflib, CTLFLAG_RD | CTLFLAG_MPSAFE, 0,
579     "iflib driver parameters");
580 
581 /*
582  * XXX need to ensure that this can't accidentally cause the head to be moved backwards
583  */
584 static int iflib_min_tx_latency = 0;
585 SYSCTL_INT(_net_iflib, OID_AUTO, min_tx_latency, CTLFLAG_RW,
586 		   &iflib_min_tx_latency, 0, "minimize transmit latency at the possible expense of throughput");
587 static int iflib_no_tx_batch = 0;
588 SYSCTL_INT(_net_iflib, OID_AUTO, no_tx_batch, CTLFLAG_RW,
589 		   &iflib_no_tx_batch, 0, "minimize transmit latency at the possible expense of throughput");
590 static int iflib_timer_default = 1000;
591 SYSCTL_INT(_net_iflib, OID_AUTO, timer_default, CTLFLAG_RW,
592 		   &iflib_timer_default, 0, "number of ticks between iflib_timer calls");
593 
594 
595 #if IFLIB_DEBUG_COUNTERS
596 
597 static int iflib_tx_seen;
598 static int iflib_tx_sent;
599 static int iflib_tx_encap;
600 static int iflib_rx_allocs;
601 static int iflib_fl_refills;
602 static int iflib_fl_refills_large;
603 static int iflib_tx_frees;
604 
605 SYSCTL_INT(_net_iflib, OID_AUTO, tx_seen, CTLFLAG_RD,
606 		   &iflib_tx_seen, 0, "# TX mbufs seen");
607 SYSCTL_INT(_net_iflib, OID_AUTO, tx_sent, CTLFLAG_RD,
608 		   &iflib_tx_sent, 0, "# TX mbufs sent");
609 SYSCTL_INT(_net_iflib, OID_AUTO, tx_encap, CTLFLAG_RD,
610 		   &iflib_tx_encap, 0, "# TX mbufs encapped");
611 SYSCTL_INT(_net_iflib, OID_AUTO, tx_frees, CTLFLAG_RD,
612 		   &iflib_tx_frees, 0, "# TX frees");
613 SYSCTL_INT(_net_iflib, OID_AUTO, rx_allocs, CTLFLAG_RD,
614 		   &iflib_rx_allocs, 0, "# RX allocations");
615 SYSCTL_INT(_net_iflib, OID_AUTO, fl_refills, CTLFLAG_RD,
616 		   &iflib_fl_refills, 0, "# refills");
617 SYSCTL_INT(_net_iflib, OID_AUTO, fl_refills_large, CTLFLAG_RD,
618 		   &iflib_fl_refills_large, 0, "# large refills");
619 
620 static int iflib_txq_drain_flushing;
621 static int iflib_txq_drain_oactive;
622 static int iflib_txq_drain_notready;
623 
624 SYSCTL_INT(_net_iflib, OID_AUTO, txq_drain_flushing, CTLFLAG_RD,
625 		   &iflib_txq_drain_flushing, 0, "# drain flushes");
626 SYSCTL_INT(_net_iflib, OID_AUTO, txq_drain_oactive, CTLFLAG_RD,
627 		   &iflib_txq_drain_oactive, 0, "# drain oactives");
628 SYSCTL_INT(_net_iflib, OID_AUTO, txq_drain_notready, CTLFLAG_RD,
629 		   &iflib_txq_drain_notready, 0, "# drain notready");
630 
631 static int iflib_encap_load_mbuf_fail;
632 static int iflib_encap_pad_mbuf_fail;
633 static int iflib_encap_txq_avail_fail;
634 static int iflib_encap_txd_encap_fail;
635 
636 SYSCTL_INT(_net_iflib, OID_AUTO, encap_load_mbuf_fail, CTLFLAG_RD,
637 		   &iflib_encap_load_mbuf_fail, 0, "# busdma load failures");
638 SYSCTL_INT(_net_iflib, OID_AUTO, encap_pad_mbuf_fail, CTLFLAG_RD,
639 		   &iflib_encap_pad_mbuf_fail, 0, "# runt frame pad failures");
640 SYSCTL_INT(_net_iflib, OID_AUTO, encap_txq_avail_fail, CTLFLAG_RD,
641 		   &iflib_encap_txq_avail_fail, 0, "# txq avail failures");
642 SYSCTL_INT(_net_iflib, OID_AUTO, encap_txd_encap_fail, CTLFLAG_RD,
643 		   &iflib_encap_txd_encap_fail, 0, "# driver encap failures");
644 
645 static int iflib_task_fn_rxs;
646 static int iflib_rx_intr_enables;
647 static int iflib_fast_intrs;
648 static int iflib_rx_unavail;
649 static int iflib_rx_ctx_inactive;
650 static int iflib_rx_if_input;
651 static int iflib_rxd_flush;
652 
653 static int iflib_verbose_debug;
654 
655 SYSCTL_INT(_net_iflib, OID_AUTO, task_fn_rx, CTLFLAG_RD,
656 		   &iflib_task_fn_rxs, 0, "# task_fn_rx calls");
657 SYSCTL_INT(_net_iflib, OID_AUTO, rx_intr_enables, CTLFLAG_RD,
658 		   &iflib_rx_intr_enables, 0, "# RX intr enables");
659 SYSCTL_INT(_net_iflib, OID_AUTO, fast_intrs, CTLFLAG_RD,
660 		   &iflib_fast_intrs, 0, "# fast_intr calls");
661 SYSCTL_INT(_net_iflib, OID_AUTO, rx_unavail, CTLFLAG_RD,
662 		   &iflib_rx_unavail, 0, "# times rxeof called with no available data");
663 SYSCTL_INT(_net_iflib, OID_AUTO, rx_ctx_inactive, CTLFLAG_RD,
664 		   &iflib_rx_ctx_inactive, 0, "# times rxeof called with inactive context");
665 SYSCTL_INT(_net_iflib, OID_AUTO, rx_if_input, CTLFLAG_RD,
666 		   &iflib_rx_if_input, 0, "# times rxeof called if_input");
667 SYSCTL_INT(_net_iflib, OID_AUTO, rxd_flush, CTLFLAG_RD,
668 	         &iflib_rxd_flush, 0, "# times rxd_flush called");
669 SYSCTL_INT(_net_iflib, OID_AUTO, verbose_debug, CTLFLAG_RW,
670 		   &iflib_verbose_debug, 0, "enable verbose debugging");
671 
672 #define DBG_COUNTER_INC(name) atomic_add_int(&(iflib_ ## name), 1)
673 static void
674 iflib_debug_reset(void)
675 {
676 	iflib_tx_seen = iflib_tx_sent = iflib_tx_encap = iflib_rx_allocs =
677 		iflib_fl_refills = iflib_fl_refills_large = iflib_tx_frees =
678 		iflib_txq_drain_flushing = iflib_txq_drain_oactive =
679 		iflib_txq_drain_notready =
680 		iflib_encap_load_mbuf_fail = iflib_encap_pad_mbuf_fail =
681 		iflib_encap_txq_avail_fail = iflib_encap_txd_encap_fail =
682 		iflib_task_fn_rxs = iflib_rx_intr_enables = iflib_fast_intrs =
683 		iflib_rx_unavail =
684 		iflib_rx_ctx_inactive = iflib_rx_if_input =
685 		iflib_rxd_flush = 0;
686 }
687 
688 #else
689 #define DBG_COUNTER_INC(name)
690 static void iflib_debug_reset(void) {}
691 #endif
692 
693 #define IFLIB_DEBUG 0
694 
695 static void iflib_tx_structures_free(if_ctx_t ctx);
696 static void iflib_rx_structures_free(if_ctx_t ctx);
697 static int iflib_queues_alloc(if_ctx_t ctx);
698 static int iflib_tx_credits_update(if_ctx_t ctx, iflib_txq_t txq);
699 static int iflib_rxd_avail(if_ctx_t ctx, iflib_rxq_t rxq, qidx_t cidx, qidx_t budget);
700 static int iflib_qset_structures_setup(if_ctx_t ctx);
701 static int iflib_msix_init(if_ctx_t ctx);
702 static int iflib_legacy_setup(if_ctx_t ctx, driver_filter_t filter, void *filterarg, int *rid, const char *str);
703 static void iflib_txq_check_drain(iflib_txq_t txq, int budget);
704 static uint32_t iflib_txq_can_drain(struct ifmp_ring *);
705 #ifdef ALTQ
706 static void iflib_altq_if_start(if_t ifp);
707 static int iflib_altq_if_transmit(if_t ifp, struct mbuf *m);
708 #endif
709 static int iflib_register(if_ctx_t);
710 static void iflib_deregister(if_ctx_t);
711 static void iflib_unregister_vlan_handlers(if_ctx_t ctx);
712 static uint16_t iflib_get_mbuf_size_for(unsigned int size);
713 static void iflib_init_locked(if_ctx_t ctx);
714 static void iflib_add_device_sysctl_pre(if_ctx_t ctx);
715 static void iflib_add_device_sysctl_post(if_ctx_t ctx);
716 static void iflib_ifmp_purge(iflib_txq_t txq);
717 static void _iflib_pre_assert(if_softc_ctx_t scctx);
718 static void iflib_if_init_locked(if_ctx_t ctx);
719 static void iflib_free_intr_mem(if_ctx_t ctx);
720 #ifndef __NO_STRICT_ALIGNMENT
721 static struct mbuf * iflib_fixup_rx(struct mbuf *m);
722 #endif
723 
724 static SLIST_HEAD(cpu_offset_list, cpu_offset) cpu_offsets =
725     SLIST_HEAD_INITIALIZER(cpu_offsets);
726 struct cpu_offset {
727 	SLIST_ENTRY(cpu_offset) entries;
728 	cpuset_t	set;
729 	unsigned int	refcount;
730 	uint16_t	next_cpuid;
731 };
732 static struct mtx cpu_offset_mtx;
733 MTX_SYSINIT(iflib_cpu_offset, &cpu_offset_mtx, "iflib_cpu_offset lock",
734     MTX_DEF);
735 
736 DEBUGNET_DEFINE(iflib);
737 
738 static int
739 iflib_num_rx_descs(if_ctx_t ctx)
740 {
741 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
742 	if_shared_ctx_t sctx = ctx->ifc_sctx;
743 	uint16_t first_rxq = (sctx->isc_flags & IFLIB_HAS_RXCQ) ? 1 : 0;
744 
745 	return scctx->isc_nrxd[first_rxq];
746 }
747 
748 static int
749 iflib_num_tx_descs(if_ctx_t ctx)
750 {
751 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
752 	if_shared_ctx_t sctx = ctx->ifc_sctx;
753 	uint16_t first_txq = (sctx->isc_flags & IFLIB_HAS_TXCQ) ? 1 : 0;
754 
755 	return scctx->isc_ntxd[first_txq];
756 }
757 
758 #ifdef DEV_NETMAP
759 #include <sys/selinfo.h>
760 #include <net/netmap.h>
761 #include <dev/netmap/netmap_kern.h>
762 
763 MODULE_DEPEND(iflib, netmap, 1, 1, 1);
764 
765 static int netmap_fl_refill(iflib_rxq_t rxq, struct netmap_kring *kring, bool init);
766 static void iflib_netmap_timer(void *arg);
767 
768 /*
769  * device-specific sysctl variables:
770  *
771  * iflib_crcstrip: 0: keep CRC in rx frames (default), 1: strip it.
772  *	During regular operations the CRC is stripped, but on some
773  *	hardware reception of frames not multiple of 64 is slower,
774  *	so using crcstrip=0 helps in benchmarks.
775  *
776  * iflib_rx_miss, iflib_rx_miss_bufs:
777  *	count packets that might be missed due to lost interrupts.
778  */
779 SYSCTL_DECL(_dev_netmap);
780 /*
781  * The xl driver by default strips CRCs and we do not override it.
782  */
783 
784 int iflib_crcstrip = 1;
785 SYSCTL_INT(_dev_netmap, OID_AUTO, iflib_crcstrip,
786     CTLFLAG_RW, &iflib_crcstrip, 1, "strip CRC on RX frames");
787 
788 int iflib_rx_miss, iflib_rx_miss_bufs;
789 SYSCTL_INT(_dev_netmap, OID_AUTO, iflib_rx_miss,
790     CTLFLAG_RW, &iflib_rx_miss, 0, "potentially missed RX intr");
791 SYSCTL_INT(_dev_netmap, OID_AUTO, iflib_rx_miss_bufs,
792     CTLFLAG_RW, &iflib_rx_miss_bufs, 0, "potentially missed RX intr bufs");
793 
794 /*
795  * Register/unregister. We are already under netmap lock.
796  * Only called on the first register or the last unregister.
797  */
798 static int
799 iflib_netmap_register(struct netmap_adapter *na, int onoff)
800 {
801 	if_t ifp = na->ifp;
802 	if_ctx_t ctx = ifp->if_softc;
803 	int status;
804 
805 	CTX_LOCK(ctx);
806 	if (!CTX_IS_VF(ctx))
807 		IFDI_CRCSTRIP_SET(ctx, onoff, iflib_crcstrip);
808 
809 	iflib_stop(ctx);
810 
811 	/*
812 	 * Enable (or disable) netmap flags, and intercept (or restore)
813 	 * ifp->if_transmit. This is done once the device has been stopped
814 	 * to prevent race conditions. Also, this must be done after
815 	 * calling netmap_disable_all_rings() and before calling
816 	 * netmap_enable_all_rings(), so that these two functions see the
817 	 * updated state of the NAF_NETMAP_ON bit.
818 	 */
819 	if (onoff) {
820 		nm_set_native_flags(na);
821 	} else {
822 		nm_clear_native_flags(na);
823 	}
824 
825 	iflib_init_locked(ctx);
826 	IFDI_CRCSTRIP_SET(ctx, onoff, iflib_crcstrip); // XXX why twice ?
827 	status = ifp->if_drv_flags & IFF_DRV_RUNNING ? 0 : 1;
828 	if (status)
829 		nm_clear_native_flags(na);
830 	CTX_UNLOCK(ctx);
831 	return (status);
832 }
833 
834 static int
835 iflib_netmap_config(struct netmap_adapter *na, struct nm_config_info *info)
836 {
837 	if_t ifp = na->ifp;
838 	if_ctx_t ctx = ifp->if_softc;
839 	iflib_rxq_t rxq = &ctx->ifc_rxqs[0];
840 	iflib_fl_t fl = &rxq->ifr_fl[0];
841 
842 	info->num_tx_rings = ctx->ifc_softc_ctx.isc_ntxqsets;
843 	info->num_rx_rings = ctx->ifc_softc_ctx.isc_nrxqsets;
844 	info->num_tx_descs = iflib_num_tx_descs(ctx);
845 	info->num_rx_descs = iflib_num_rx_descs(ctx);
846 	info->rx_buf_maxsize = fl->ifl_buf_size;
847 	nm_prinf("txr %u rxr %u txd %u rxd %u rbufsz %u",
848 		info->num_tx_rings, info->num_rx_rings, info->num_tx_descs,
849 		info->num_rx_descs, info->rx_buf_maxsize);
850 
851 	return 0;
852 }
853 
854 static int
855 netmap_fl_refill(iflib_rxq_t rxq, struct netmap_kring *kring, bool init)
856 {
857 	struct netmap_adapter *na = kring->na;
858 	u_int const lim = kring->nkr_num_slots - 1;
859 	struct netmap_ring *ring = kring->ring;
860 	bus_dmamap_t *map;
861 	struct if_rxd_update iru;
862 	if_ctx_t ctx = rxq->ifr_ctx;
863 	iflib_fl_t fl = &rxq->ifr_fl[0];
864 	u_int nic_i_first, nic_i;
865 	u_int nm_i;
866 	int i, n;
867 #if IFLIB_DEBUG_COUNTERS
868 	int rf_count = 0;
869 #endif
870 
871 	/*
872 	 * This function is used both at initialization and in rxsync.
873 	 * At initialization we need to prepare (with isc_rxd_refill())
874 	 * all the netmap buffers currently owned by the kernel, in
875 	 * such a way to keep fl->ifl_pidx and kring->nr_hwcur in sync
876 	 * (except for kring->nkr_hwofs). These may be less than
877 	 * kring->nkr_num_slots if netmap_reset() was called while
878 	 * an application using the kring that still owned some
879 	 * buffers.
880 	 * At rxsync time, both indexes point to the next buffer to be
881 	 * refilled.
882 	 * In any case we publish (with isc_rxd_flush()) up to
883 	 * (fl->ifl_pidx - 1) % N (included), to avoid the NIC tail/prod
884 	 * pointer to overrun the head/cons pointer, although this is
885 	 * not necessary for some NICs (e.g. vmx).
886 	 */
887 	if (__predict_false(init)) {
888 		n = kring->nkr_num_slots - nm_kr_rxspace(kring);
889 	} else {
890 		n = kring->rhead - kring->nr_hwcur;
891 		if (n == 0)
892 			return (0); /* Nothing to do. */
893 		if (n < 0)
894 			n += kring->nkr_num_slots;
895 	}
896 
897 	iru_init(&iru, rxq, 0 /* flid */);
898 	map = fl->ifl_sds.ifsd_map;
899 	nic_i = fl->ifl_pidx;
900 	nm_i = netmap_idx_n2k(kring, nic_i);
901 	if (__predict_false(init)) {
902 		/*
903 		 * On init/reset, nic_i must be 0, and we must
904 		 * start to refill from hwtail (see netmap_reset()).
905 		 */
906 		MPASS(nic_i == 0);
907 		MPASS(nm_i == kring->nr_hwtail);
908 	} else
909 		MPASS(nm_i == kring->nr_hwcur);
910 	DBG_COUNTER_INC(fl_refills);
911 	while (n > 0) {
912 #if IFLIB_DEBUG_COUNTERS
913 		if (++rf_count == 9)
914 			DBG_COUNTER_INC(fl_refills_large);
915 #endif
916 		nic_i_first = nic_i;
917 		for (i = 0; n > 0 && i < IFLIB_MAX_RX_REFRESH; n--, i++) {
918 			struct netmap_slot *slot = &ring->slot[nm_i];
919 			uint64_t paddr;
920 			void *addr = PNMB(na, slot, &paddr);
921 
922 			MPASS(i < IFLIB_MAX_RX_REFRESH);
923 
924 			if (addr == NETMAP_BUF_BASE(na)) /* bad buf */
925 			        return netmap_ring_reinit(kring);
926 
927 			fl->ifl_bus_addrs[i] = paddr +
928 			    nm_get_offset(kring, slot);
929 			fl->ifl_rxd_idxs[i] = nic_i;
930 
931 			if (__predict_false(init)) {
932 				netmap_load_map(na, fl->ifl_buf_tag,
933 				    map[nic_i], addr);
934 			} else if (slot->flags & NS_BUF_CHANGED) {
935 				/* buffer has changed, reload map */
936 				netmap_reload_map(na, fl->ifl_buf_tag,
937 				    map[nic_i], addr);
938 			}
939 			bus_dmamap_sync(fl->ifl_buf_tag, map[nic_i],
940 			    BUS_DMASYNC_PREREAD);
941 			slot->flags &= ~NS_BUF_CHANGED;
942 
943 			nm_i = nm_next(nm_i, lim);
944 			nic_i = nm_next(nic_i, lim);
945 		}
946 
947 		iru.iru_pidx = nic_i_first;
948 		iru.iru_count = i;
949 		ctx->isc_rxd_refill(ctx->ifc_softc, &iru);
950 	}
951 	fl->ifl_pidx = nic_i;
952 	/*
953 	 * At the end of the loop we must have refilled everything
954 	 * we could possibly refill.
955 	 */
956 	MPASS(nm_i == kring->rhead);
957 	kring->nr_hwcur = nm_i;
958 
959 	bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
960 	    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
961 	ctx->isc_rxd_flush(ctx->ifc_softc, rxq->ifr_id, fl->ifl_id,
962 	    nm_prev(nic_i, lim));
963 	DBG_COUNTER_INC(rxd_flush);
964 
965 	return (0);
966 }
967 
968 #define NETMAP_TX_TIMER_US	90
969 
970 /*
971  * Reconcile kernel and user view of the transmit ring.
972  *
973  * All information is in the kring.
974  * Userspace wants to send packets up to the one before kring->rhead,
975  * kernel knows kring->nr_hwcur is the first unsent packet.
976  *
977  * Here we push packets out (as many as possible), and possibly
978  * reclaim buffers from previously completed transmission.
979  *
980  * The caller (netmap) guarantees that there is only one instance
981  * running at any time. Any interference with other driver
982  * methods should be handled by the individual drivers.
983  */
984 static int
985 iflib_netmap_txsync(struct netmap_kring *kring, int flags)
986 {
987 	struct netmap_adapter *na = kring->na;
988 	if_t ifp = na->ifp;
989 	struct netmap_ring *ring = kring->ring;
990 	u_int nm_i;	/* index into the netmap kring */
991 	u_int nic_i;	/* index into the NIC ring */
992 	u_int n;
993 	u_int const lim = kring->nkr_num_slots - 1;
994 	u_int const head = kring->rhead;
995 	struct if_pkt_info pi;
996 
997 	/*
998 	 * interrupts on every tx packet are expensive so request
999 	 * them every half ring, or where NS_REPORT is set
1000 	 */
1001 	u_int report_frequency = kring->nkr_num_slots >> 1;
1002 	/* device-specific */
1003 	if_ctx_t ctx = ifp->if_softc;
1004 	iflib_txq_t txq = &ctx->ifc_txqs[kring->ring_id];
1005 
1006 	bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
1007 	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
1008 
1009 	/*
1010 	 * First part: process new packets to send.
1011 	 * nm_i is the current index in the netmap kring,
1012 	 * nic_i is the corresponding index in the NIC ring.
1013 	 *
1014 	 * If we have packets to send (nm_i != head)
1015 	 * iterate over the netmap ring, fetch length and update
1016 	 * the corresponding slot in the NIC ring. Some drivers also
1017 	 * need to update the buffer's physical address in the NIC slot
1018 	 * even NS_BUF_CHANGED is not set (PNMB computes the addresses).
1019 	 *
1020 	 * The netmap_reload_map() calls is especially expensive,
1021 	 * even when (as in this case) the tag is 0, so do only
1022 	 * when the buffer has actually changed.
1023 	 *
1024 	 * If possible do not set the report/intr bit on all slots,
1025 	 * but only a few times per ring or when NS_REPORT is set.
1026 	 *
1027 	 * Finally, on 10G and faster drivers, it might be useful
1028 	 * to prefetch the next slot and txr entry.
1029 	 */
1030 
1031 	nm_i = kring->nr_hwcur;
1032 	if (nm_i != head) {	/* we have new packets to send */
1033 		uint32_t pkt_len = 0, seg_idx = 0;
1034 		int nic_i_start = -1, flags = 0;
1035 		pkt_info_zero(&pi);
1036 		pi.ipi_segs = txq->ift_segs;
1037 		pi.ipi_qsidx = kring->ring_id;
1038 		nic_i = netmap_idx_k2n(kring, nm_i);
1039 
1040 		__builtin_prefetch(&ring->slot[nm_i]);
1041 		__builtin_prefetch(&txq->ift_sds.ifsd_m[nic_i]);
1042 		__builtin_prefetch(&txq->ift_sds.ifsd_map[nic_i]);
1043 
1044 		for (n = 0; nm_i != head; n++) {
1045 			struct netmap_slot *slot = &ring->slot[nm_i];
1046 			uint64_t offset = nm_get_offset(kring, slot);
1047 			u_int len = slot->len;
1048 			uint64_t paddr;
1049 			void *addr = PNMB(na, slot, &paddr);
1050 
1051 			flags |= (slot->flags & NS_REPORT ||
1052 				nic_i == 0 || nic_i == report_frequency) ?
1053 				IPI_TX_INTR : 0;
1054 
1055 			/*
1056 			 * If this is the first packet fragment, save the
1057 			 * index of the first NIC slot for later.
1058 			 */
1059 			if (nic_i_start < 0)
1060 				nic_i_start = nic_i;
1061 
1062 			pi.ipi_segs[seg_idx].ds_addr = paddr + offset;
1063 			pi.ipi_segs[seg_idx].ds_len = len;
1064 			if (len) {
1065 				pkt_len += len;
1066 				seg_idx++;
1067 			}
1068 
1069 			if (!(slot->flags & NS_MOREFRAG)) {
1070 				pi.ipi_len = pkt_len;
1071 				pi.ipi_nsegs = seg_idx;
1072 				pi.ipi_pidx = nic_i_start;
1073 				pi.ipi_ndescs = 0;
1074 				pi.ipi_flags = flags;
1075 
1076 				/* Prepare the NIC TX ring. */
1077 				ctx->isc_txd_encap(ctx->ifc_softc, &pi);
1078 				DBG_COUNTER_INC(tx_encap);
1079 
1080 				/* Reinit per-packet info for the next one. */
1081 				flags = seg_idx = pkt_len = 0;
1082 				nic_i_start = -1;
1083 			}
1084 
1085 			/* prefetch for next round */
1086 			__builtin_prefetch(&ring->slot[nm_i + 1]);
1087 			__builtin_prefetch(&txq->ift_sds.ifsd_m[nic_i + 1]);
1088 			__builtin_prefetch(&txq->ift_sds.ifsd_map[nic_i + 1]);
1089 
1090 			NM_CHECK_ADDR_LEN_OFF(na, len, offset);
1091 
1092 			if (slot->flags & NS_BUF_CHANGED) {
1093 				/* buffer has changed, reload map */
1094 				netmap_reload_map(na, txq->ift_buf_tag,
1095 				    txq->ift_sds.ifsd_map[nic_i], addr);
1096 			}
1097 			/* make sure changes to the buffer are synced */
1098 			bus_dmamap_sync(txq->ift_buf_tag,
1099 			    txq->ift_sds.ifsd_map[nic_i],
1100 			    BUS_DMASYNC_PREWRITE);
1101 
1102 			slot->flags &= ~(NS_REPORT | NS_BUF_CHANGED | NS_MOREFRAG);
1103 			nm_i = nm_next(nm_i, lim);
1104 			nic_i = nm_next(nic_i, lim);
1105 		}
1106 		kring->nr_hwcur = nm_i;
1107 
1108 		/* synchronize the NIC ring */
1109 		bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
1110 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
1111 
1112 		/* (re)start the tx unit up to slot nic_i (excluded) */
1113 		ctx->isc_txd_flush(ctx->ifc_softc, txq->ift_id, nic_i);
1114 	}
1115 
1116 	/*
1117 	 * Second part: reclaim buffers for completed transmissions.
1118 	 *
1119 	 * If there are unclaimed buffers, attempt to reclaim them.
1120 	 * If we don't manage to reclaim them all, and TX IRQs are not in use,
1121 	 * trigger a per-tx-queue timer to try again later.
1122 	 */
1123 	if (kring->nr_hwtail != nm_prev(kring->nr_hwcur, lim)) {
1124 		if (iflib_tx_credits_update(ctx, txq)) {
1125 			/* some tx completed, increment avail */
1126 			nic_i = txq->ift_cidx_processed;
1127 			kring->nr_hwtail = nm_prev(netmap_idx_n2k(kring, nic_i), lim);
1128 		}
1129 	}
1130 
1131 	if (!(ctx->ifc_flags & IFC_NETMAP_TX_IRQ))
1132 		if (kring->nr_hwtail != nm_prev(kring->nr_hwcur, lim)) {
1133 			callout_reset_sbt_on(&txq->ift_netmap_timer,
1134 			    NETMAP_TX_TIMER_US * SBT_1US, SBT_1US,
1135 			    iflib_netmap_timer, txq,
1136 			    txq->ift_netmap_timer.c_cpu, 0);
1137 		}
1138 	return (0);
1139 }
1140 
1141 /*
1142  * Reconcile kernel and user view of the receive ring.
1143  * Same as for the txsync, this routine must be efficient.
1144  * The caller guarantees a single invocations, but races against
1145  * the rest of the driver should be handled here.
1146  *
1147  * On call, kring->rhead is the first packet that userspace wants
1148  * to keep, and kring->rcur is the wakeup point.
1149  * The kernel has previously reported packets up to kring->rtail.
1150  *
1151  * If (flags & NAF_FORCE_READ) also check for incoming packets irrespective
1152  * of whether or not we received an interrupt.
1153  */
1154 static int
1155 iflib_netmap_rxsync(struct netmap_kring *kring, int flags)
1156 {
1157 	struct netmap_adapter *na = kring->na;
1158 	struct netmap_ring *ring = kring->ring;
1159 	if_t ifp = na->ifp;
1160 	uint32_t nm_i;	/* index into the netmap ring */
1161 	uint32_t nic_i;	/* index into the NIC ring */
1162 	u_int n;
1163 	u_int const lim = kring->nkr_num_slots - 1;
1164 	int force_update = (flags & NAF_FORCE_READ) || kring->nr_kflags & NKR_PENDINTR;
1165 	int i = 0;
1166 
1167 	if_ctx_t ctx = ifp->if_softc;
1168 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1169 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1170 	iflib_rxq_t rxq = &ctx->ifc_rxqs[kring->ring_id];
1171 	iflib_fl_t fl = &rxq->ifr_fl[0];
1172 	struct if_rxd_info ri;
1173 	qidx_t *cidxp;
1174 
1175 	/*
1176 	 * netmap only uses free list 0, to avoid out of order consumption
1177 	 * of receive buffers
1178 	 */
1179 
1180 	bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
1181 	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
1182 
1183 	/*
1184 	 * First part: import newly received packets.
1185 	 *
1186 	 * nm_i is the index of the next free slot in the netmap ring,
1187 	 * nic_i is the index of the next received packet in the NIC ring
1188 	 * (or in the free list 0 if IFLIB_HAS_RXCQ is set), and they may
1189 	 * differ in case if_init() has been called while
1190 	 * in netmap mode. For the receive ring we have
1191 	 *
1192 	 *	nic_i = fl->ifl_cidx;
1193 	 *	nm_i = kring->nr_hwtail (previous)
1194 	 * and
1195 	 *	nm_i == (nic_i + kring->nkr_hwofs) % ring_size
1196 	 *
1197 	 * fl->ifl_cidx is set to 0 on a ring reinit
1198 	 */
1199 	if (netmap_no_pendintr || force_update) {
1200 		uint32_t hwtail_lim = nm_prev(kring->nr_hwcur, lim);
1201 		bool have_rxcq = sctx->isc_flags & IFLIB_HAS_RXCQ;
1202 		int crclen = iflib_crcstrip ? 0 : 4;
1203 		int error, avail;
1204 
1205 		/*
1206 		 * For the free list consumer index, we use the same
1207 		 * logic as in iflib_rxeof().
1208 		 */
1209 		if (have_rxcq)
1210 			cidxp = &rxq->ifr_cq_cidx;
1211 		else
1212 			cidxp = &fl->ifl_cidx;
1213 		avail = ctx->isc_rxd_available(ctx->ifc_softc,
1214 		    rxq->ifr_id, *cidxp, USHRT_MAX);
1215 
1216 		nic_i = fl->ifl_cidx;
1217 		nm_i = netmap_idx_n2k(kring, nic_i);
1218 		MPASS(nm_i == kring->nr_hwtail);
1219 		for (n = 0; avail > 0 && nm_i != hwtail_lim; n++, avail--) {
1220 			rxd_info_zero(&ri);
1221 			ri.iri_frags = rxq->ifr_frags;
1222 			ri.iri_qsidx = kring->ring_id;
1223 			ri.iri_ifp = ctx->ifc_ifp;
1224 			ri.iri_cidx = *cidxp;
1225 
1226 			error = ctx->isc_rxd_pkt_get(ctx->ifc_softc, &ri);
1227 			for (i = 0; i < ri.iri_nfrags; i++) {
1228 				if (error) {
1229 					ring->slot[nm_i].len = 0;
1230 					ring->slot[nm_i].flags = 0;
1231 				} else {
1232 					ring->slot[nm_i].len = ri.iri_frags[i].irf_len;
1233 					if (i == (ri.iri_nfrags - 1)) {
1234 						ring->slot[nm_i].len -= crclen;
1235 						ring->slot[nm_i].flags = 0;
1236 					} else
1237 						ring->slot[nm_i].flags = NS_MOREFRAG;
1238 				}
1239 
1240 				bus_dmamap_sync(fl->ifl_buf_tag,
1241 				    fl->ifl_sds.ifsd_map[nic_i], BUS_DMASYNC_POSTREAD);
1242 				nm_i = nm_next(nm_i, lim);
1243 				fl->ifl_cidx = nic_i = nm_next(nic_i, lim);
1244 			}
1245 
1246 			if (have_rxcq) {
1247 				*cidxp = ri.iri_cidx;
1248 				while (*cidxp >= scctx->isc_nrxd[0])
1249 					*cidxp -= scctx->isc_nrxd[0];
1250 			}
1251 
1252 		}
1253 		if (n) { /* update the state variables */
1254 			if (netmap_no_pendintr && !force_update) {
1255 				/* diagnostics */
1256 				iflib_rx_miss ++;
1257 				iflib_rx_miss_bufs += n;
1258 			}
1259 			kring->nr_hwtail = nm_i;
1260 		}
1261 		kring->nr_kflags &= ~NKR_PENDINTR;
1262 	}
1263 	/*
1264 	 * Second part: skip past packets that userspace has released.
1265 	 * (kring->nr_hwcur to head excluded),
1266 	 * and make the buffers available for reception.
1267 	 * As usual nm_i is the index in the netmap ring,
1268 	 * nic_i is the index in the NIC ring, and
1269 	 * nm_i == (nic_i + kring->nkr_hwofs) % ring_size
1270 	 */
1271 	netmap_fl_refill(rxq, kring, false);
1272 
1273 	return (0);
1274 }
1275 
1276 static void
1277 iflib_netmap_intr(struct netmap_adapter *na, int onoff)
1278 {
1279 	if_ctx_t ctx = na->ifp->if_softc;
1280 
1281 	CTX_LOCK(ctx);
1282 	if (onoff) {
1283 		IFDI_INTR_ENABLE(ctx);
1284 	} else {
1285 		IFDI_INTR_DISABLE(ctx);
1286 	}
1287 	CTX_UNLOCK(ctx);
1288 }
1289 
1290 static int
1291 iflib_netmap_attach(if_ctx_t ctx)
1292 {
1293 	struct netmap_adapter na;
1294 
1295 	bzero(&na, sizeof(na));
1296 
1297 	na.ifp = ctx->ifc_ifp;
1298 	na.na_flags = NAF_BDG_MAYSLEEP | NAF_MOREFRAG | NAF_OFFSETS;
1299 	MPASS(ctx->ifc_softc_ctx.isc_ntxqsets);
1300 	MPASS(ctx->ifc_softc_ctx.isc_nrxqsets);
1301 
1302 	na.num_tx_desc = iflib_num_tx_descs(ctx);
1303 	na.num_rx_desc = iflib_num_rx_descs(ctx);
1304 	na.nm_txsync = iflib_netmap_txsync;
1305 	na.nm_rxsync = iflib_netmap_rxsync;
1306 	na.nm_register = iflib_netmap_register;
1307 	na.nm_intr = iflib_netmap_intr;
1308 	na.nm_config = iflib_netmap_config;
1309 	na.num_tx_rings = ctx->ifc_softc_ctx.isc_ntxqsets;
1310 	na.num_rx_rings = ctx->ifc_softc_ctx.isc_nrxqsets;
1311 	return (netmap_attach(&na));
1312 }
1313 
1314 static int
1315 iflib_netmap_txq_init(if_ctx_t ctx, iflib_txq_t txq)
1316 {
1317 	struct netmap_adapter *na = NA(ctx->ifc_ifp);
1318 	struct netmap_slot *slot;
1319 
1320 	slot = netmap_reset(na, NR_TX, txq->ift_id, 0);
1321 	if (slot == NULL)
1322 		return (0);
1323 	for (int i = 0; i < ctx->ifc_softc_ctx.isc_ntxd[0]; i++) {
1324 		/*
1325 		 * In netmap mode, set the map for the packet buffer.
1326 		 * NOTE: Some drivers (not this one) also need to set
1327 		 * the physical buffer address in the NIC ring.
1328 		 * netmap_idx_n2k() maps a nic index, i, into the corresponding
1329 		 * netmap slot index, si
1330 		 */
1331 		int si = netmap_idx_n2k(na->tx_rings[txq->ift_id], i);
1332 		netmap_load_map(na, txq->ift_buf_tag, txq->ift_sds.ifsd_map[i],
1333 		    NMB(na, slot + si));
1334 	}
1335 	return (1);
1336 }
1337 
1338 static int
1339 iflib_netmap_rxq_init(if_ctx_t ctx, iflib_rxq_t rxq)
1340 {
1341 	struct netmap_adapter *na = NA(ctx->ifc_ifp);
1342 	struct netmap_kring *kring;
1343 	struct netmap_slot *slot;
1344 
1345 	slot = netmap_reset(na, NR_RX, rxq->ifr_id, 0);
1346 	if (slot == NULL)
1347 		return (0);
1348 	kring = na->rx_rings[rxq->ifr_id];
1349 	netmap_fl_refill(rxq, kring, true);
1350 	return (1);
1351 }
1352 
1353 static void
1354 iflib_netmap_timer(void *arg)
1355 {
1356 	iflib_txq_t txq = arg;
1357 	if_ctx_t ctx = txq->ift_ctx;
1358 
1359 	/*
1360 	 * Wake up the netmap application, to give it a chance to
1361 	 * call txsync and reclaim more completed TX buffers.
1362 	 */
1363 	netmap_tx_irq(ctx->ifc_ifp, txq->ift_id);
1364 }
1365 
1366 #define iflib_netmap_detach(ifp) netmap_detach(ifp)
1367 
1368 #else
1369 #define iflib_netmap_txq_init(ctx, txq) (0)
1370 #define iflib_netmap_rxq_init(ctx, rxq) (0)
1371 #define iflib_netmap_detach(ifp)
1372 #define netmap_enable_all_rings(ifp)
1373 #define netmap_disable_all_rings(ifp)
1374 
1375 #define iflib_netmap_attach(ctx) (0)
1376 #define netmap_rx_irq(ifp, qid, budget) (0)
1377 #endif
1378 
1379 #if defined(__i386__) || defined(__amd64__)
1380 static __inline void
1381 prefetch(void *x)
1382 {
1383 	__asm volatile("prefetcht0 %0" :: "m" (*(unsigned long *)x));
1384 }
1385 static __inline void
1386 prefetch2cachelines(void *x)
1387 {
1388 	__asm volatile("prefetcht0 %0" :: "m" (*(unsigned long *)x));
1389 #if (CACHE_LINE_SIZE < 128)
1390 	__asm volatile("prefetcht0 %0" :: "m" (*(((unsigned long *)x)+CACHE_LINE_SIZE/(sizeof(unsigned long)))));
1391 #endif
1392 }
1393 #else
1394 #define prefetch(x)
1395 #define prefetch2cachelines(x)
1396 #endif
1397 
1398 static void
1399 iru_init(if_rxd_update_t iru, iflib_rxq_t rxq, uint8_t flid)
1400 {
1401 	iflib_fl_t fl;
1402 
1403 	fl = &rxq->ifr_fl[flid];
1404 	iru->iru_paddrs = fl->ifl_bus_addrs;
1405 	iru->iru_idxs = fl->ifl_rxd_idxs;
1406 	iru->iru_qsidx = rxq->ifr_id;
1407 	iru->iru_buf_size = fl->ifl_buf_size;
1408 	iru->iru_flidx = fl->ifl_id;
1409 }
1410 
1411 static void
1412 _iflib_dmamap_cb(void *arg, bus_dma_segment_t *segs, int nseg, int err)
1413 {
1414 	if (err)
1415 		return;
1416 	*(bus_addr_t *) arg = segs[0].ds_addr;
1417 }
1418 
1419 #define	DMA_WIDTH_TO_BUS_LOWADDR(width)				\
1420 	(((width) == 0) || (width) == flsll(BUS_SPACE_MAXADDR) ?	\
1421 	    BUS_SPACE_MAXADDR : (1ULL << (width)) - 1ULL)
1422 
1423 int
1424 iflib_dma_alloc_align(if_ctx_t ctx, int size, int align, iflib_dma_info_t dma, int mapflags)
1425 {
1426 	int err;
1427 	device_t dev = ctx->ifc_dev;
1428 	bus_addr_t lowaddr;
1429 
1430 	lowaddr = DMA_WIDTH_TO_BUS_LOWADDR(ctx->ifc_softc_ctx.isc_dma_width);
1431 
1432 	err = bus_dma_tag_create(bus_get_dma_tag(dev),	/* parent */
1433 				align, 0,		/* alignment, bounds */
1434 				lowaddr,		/* lowaddr */
1435 				BUS_SPACE_MAXADDR,	/* highaddr */
1436 				NULL, NULL,		/* filter, filterarg */
1437 				size,			/* maxsize */
1438 				1,			/* nsegments */
1439 				size,			/* maxsegsize */
1440 				BUS_DMA_ALLOCNOW,	/* flags */
1441 				NULL,			/* lockfunc */
1442 				NULL,			/* lockarg */
1443 				&dma->idi_tag);
1444 	if (err) {
1445 		device_printf(dev,
1446 		    "%s: bus_dma_tag_create failed: %d\n",
1447 		    __func__, err);
1448 		goto fail_0;
1449 	}
1450 
1451 	err = bus_dmamem_alloc(dma->idi_tag, (void**) &dma->idi_vaddr,
1452 	    BUS_DMA_NOWAIT | BUS_DMA_COHERENT | BUS_DMA_ZERO, &dma->idi_map);
1453 	if (err) {
1454 		device_printf(dev,
1455 		    "%s: bus_dmamem_alloc(%ju) failed: %d\n",
1456 		    __func__, (uintmax_t)size, err);
1457 		goto fail_1;
1458 	}
1459 
1460 	dma->idi_paddr = IF_BAD_DMA;
1461 	err = bus_dmamap_load(dma->idi_tag, dma->idi_map, dma->idi_vaddr,
1462 	    size, _iflib_dmamap_cb, &dma->idi_paddr, mapflags | BUS_DMA_NOWAIT);
1463 	if (err || dma->idi_paddr == IF_BAD_DMA) {
1464 		device_printf(dev,
1465 		    "%s: bus_dmamap_load failed: %d\n",
1466 		    __func__, err);
1467 		goto fail_2;
1468 	}
1469 
1470 	dma->idi_size = size;
1471 	return (0);
1472 
1473 fail_2:
1474 	bus_dmamem_free(dma->idi_tag, dma->idi_vaddr, dma->idi_map);
1475 fail_1:
1476 	bus_dma_tag_destroy(dma->idi_tag);
1477 fail_0:
1478 	dma->idi_tag = NULL;
1479 
1480 	return (err);
1481 }
1482 
1483 int
1484 iflib_dma_alloc(if_ctx_t ctx, int size, iflib_dma_info_t dma, int mapflags)
1485 {
1486 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1487 
1488 	KASSERT(sctx->isc_q_align != 0, ("alignment value not initialized"));
1489 
1490 	return (iflib_dma_alloc_align(ctx, size, sctx->isc_q_align, dma, mapflags));
1491 }
1492 
1493 int
1494 iflib_dma_alloc_multi(if_ctx_t ctx, int *sizes, iflib_dma_info_t *dmalist, int mapflags, int count)
1495 {
1496 	int i, err;
1497 	iflib_dma_info_t *dmaiter;
1498 
1499 	dmaiter = dmalist;
1500 	for (i = 0; i < count; i++, dmaiter++) {
1501 		if ((err = iflib_dma_alloc(ctx, sizes[i], *dmaiter, mapflags)) != 0)
1502 			break;
1503 	}
1504 	if (err)
1505 		iflib_dma_free_multi(dmalist, i);
1506 	return (err);
1507 }
1508 
1509 void
1510 iflib_dma_free(iflib_dma_info_t dma)
1511 {
1512 	if (dma->idi_tag == NULL)
1513 		return;
1514 	if (dma->idi_paddr != IF_BAD_DMA) {
1515 		bus_dmamap_sync(dma->idi_tag, dma->idi_map,
1516 		    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
1517 		bus_dmamap_unload(dma->idi_tag, dma->idi_map);
1518 		dma->idi_paddr = IF_BAD_DMA;
1519 	}
1520 	if (dma->idi_vaddr != NULL) {
1521 		bus_dmamem_free(dma->idi_tag, dma->idi_vaddr, dma->idi_map);
1522 		dma->idi_vaddr = NULL;
1523 	}
1524 	bus_dma_tag_destroy(dma->idi_tag);
1525 	dma->idi_tag = NULL;
1526 }
1527 
1528 void
1529 iflib_dma_free_multi(iflib_dma_info_t *dmalist, int count)
1530 {
1531 	int i;
1532 	iflib_dma_info_t *dmaiter = dmalist;
1533 
1534 	for (i = 0; i < count; i++, dmaiter++)
1535 		iflib_dma_free(*dmaiter);
1536 }
1537 
1538 static int
1539 iflib_fast_intr(void *arg)
1540 {
1541 	iflib_filter_info_t info = arg;
1542 	struct grouptask *gtask = info->ifi_task;
1543 	int result;
1544 
1545 	DBG_COUNTER_INC(fast_intrs);
1546 	if (info->ifi_filter != NULL) {
1547 		result = info->ifi_filter(info->ifi_filter_arg);
1548 		if ((result & FILTER_SCHEDULE_THREAD) == 0)
1549 			return (result);
1550 	}
1551 
1552 	GROUPTASK_ENQUEUE(gtask);
1553 	return (FILTER_HANDLED);
1554 }
1555 
1556 static int
1557 iflib_fast_intr_rxtx(void *arg)
1558 {
1559 	iflib_filter_info_t info = arg;
1560 	struct grouptask *gtask = info->ifi_task;
1561 	if_ctx_t ctx;
1562 	iflib_rxq_t rxq = (iflib_rxq_t)info->ifi_ctx;
1563 	iflib_txq_t txq;
1564 	void *sc;
1565 	int i, cidx, result;
1566 	qidx_t txqid;
1567 	bool intr_enable, intr_legacy;
1568 
1569 	DBG_COUNTER_INC(fast_intrs);
1570 	if (info->ifi_filter != NULL) {
1571 		result = info->ifi_filter(info->ifi_filter_arg);
1572 		if ((result & FILTER_SCHEDULE_THREAD) == 0)
1573 			return (result);
1574 	}
1575 
1576 	ctx = rxq->ifr_ctx;
1577 	sc = ctx->ifc_softc;
1578 	intr_enable = false;
1579 	intr_legacy = !!(ctx->ifc_flags & IFC_LEGACY);
1580 	MPASS(rxq->ifr_ntxqirq);
1581 	for (i = 0; i < rxq->ifr_ntxqirq; i++) {
1582 		txqid = rxq->ifr_txqid[i];
1583 		txq = &ctx->ifc_txqs[txqid];
1584 		bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
1585 		    BUS_DMASYNC_POSTREAD);
1586 		if (!ctx->isc_txd_credits_update(sc, txqid, false)) {
1587 			if (intr_legacy)
1588 				intr_enable = true;
1589 			else
1590 				IFDI_TX_QUEUE_INTR_ENABLE(ctx, txqid);
1591 			continue;
1592 		}
1593 		GROUPTASK_ENQUEUE(&txq->ift_task);
1594 	}
1595 	if (ctx->ifc_sctx->isc_flags & IFLIB_HAS_RXCQ)
1596 		cidx = rxq->ifr_cq_cidx;
1597 	else
1598 		cidx = rxq->ifr_fl[0].ifl_cidx;
1599 	if (iflib_rxd_avail(ctx, rxq, cidx, 1))
1600 		GROUPTASK_ENQUEUE(gtask);
1601 	else {
1602 		if (intr_legacy)
1603 			intr_enable = true;
1604 		else
1605 			IFDI_RX_QUEUE_INTR_ENABLE(ctx, rxq->ifr_id);
1606 		DBG_COUNTER_INC(rx_intr_enables);
1607 	}
1608 	if (intr_enable)
1609 		IFDI_INTR_ENABLE(ctx);
1610 	return (FILTER_HANDLED);
1611 }
1612 
1613 static int
1614 iflib_fast_intr_ctx(void *arg)
1615 {
1616 	iflib_filter_info_t info = arg;
1617 	struct grouptask *gtask = info->ifi_task;
1618 	int result;
1619 
1620 	DBG_COUNTER_INC(fast_intrs);
1621 	if (info->ifi_filter != NULL) {
1622 		result = info->ifi_filter(info->ifi_filter_arg);
1623 		if ((result & FILTER_SCHEDULE_THREAD) == 0)
1624 			return (result);
1625 	}
1626 
1627 	GROUPTASK_ENQUEUE(gtask);
1628 	return (FILTER_HANDLED);
1629 }
1630 
1631 static int
1632 _iflib_irq_alloc(if_ctx_t ctx, if_irq_t irq, int rid,
1633 		 driver_filter_t filter, driver_intr_t handler, void *arg,
1634 		 const char *name)
1635 {
1636 	struct resource *res;
1637 	void *tag = NULL;
1638 	device_t dev = ctx->ifc_dev;
1639 	int flags, i, rc;
1640 
1641 	flags = RF_ACTIVE;
1642 	if (ctx->ifc_flags & IFC_LEGACY)
1643 		flags |= RF_SHAREABLE;
1644 	MPASS(rid < 512);
1645 	i = rid;
1646 	res = bus_alloc_resource_any(dev, SYS_RES_IRQ, &i, flags);
1647 	if (res == NULL) {
1648 		device_printf(dev,
1649 		    "failed to allocate IRQ for rid %d, name %s.\n", rid, name);
1650 		return (ENOMEM);
1651 	}
1652 	irq->ii_res = res;
1653 	KASSERT(filter == NULL || handler == NULL, ("filter and handler can't both be non-NULL"));
1654 	rc = bus_setup_intr(dev, res, INTR_MPSAFE | INTR_TYPE_NET,
1655 						filter, handler, arg, &tag);
1656 	if (rc != 0) {
1657 		device_printf(dev,
1658 		    "failed to setup interrupt for rid %d, name %s: %d\n",
1659 					  rid, name ? name : "unknown", rc);
1660 		return (rc);
1661 	} else if (name)
1662 		bus_describe_intr(dev, res, tag, "%s", name);
1663 
1664 	irq->ii_tag = tag;
1665 	return (0);
1666 }
1667 
1668 /*********************************************************************
1669  *
1670  *  Allocate DMA resources for TX buffers as well as memory for the TX
1671  *  mbuf map.  TX DMA maps (non-TSO/TSO) and TX mbuf map are kept in a
1672  *  iflib_sw_tx_desc_array structure, storing all the information that
1673  *  is needed to transmit a packet on the wire.  This is called only
1674  *  once at attach, setup is done every reset.
1675  *
1676  **********************************************************************/
1677 static int
1678 iflib_txsd_alloc(iflib_txq_t txq)
1679 {
1680 	if_ctx_t ctx = txq->ift_ctx;
1681 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1682 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1683 	device_t dev = ctx->ifc_dev;
1684 	bus_size_t tsomaxsize;
1685 	bus_addr_t lowaddr;
1686 	int err, nsegments, ntsosegments;
1687 	bool tso;
1688 
1689 	nsegments = scctx->isc_tx_nsegments;
1690 	ntsosegments = scctx->isc_tx_tso_segments_max;
1691 	tsomaxsize = scctx->isc_tx_tso_size_max;
1692 	if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_VLAN_MTU)
1693 		tsomaxsize += sizeof(struct ether_vlan_header);
1694 	MPASS(scctx->isc_ntxd[0] > 0);
1695 	MPASS(scctx->isc_ntxd[txq->ift_br_offset] > 0);
1696 	MPASS(nsegments > 0);
1697 	if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_TSO) {
1698 		MPASS(ntsosegments > 0);
1699 		MPASS(sctx->isc_tso_maxsize >= tsomaxsize);
1700 	}
1701 
1702 	lowaddr = DMA_WIDTH_TO_BUS_LOWADDR(scctx->isc_dma_width);
1703 
1704 	/*
1705 	 * Set up DMA tags for TX buffers.
1706 	 */
1707 	if ((err = bus_dma_tag_create(bus_get_dma_tag(dev),
1708 			       1, 0,			/* alignment, bounds */
1709 			       lowaddr,			/* lowaddr */
1710 			       BUS_SPACE_MAXADDR,	/* highaddr */
1711 			       NULL, NULL,		/* filter, filterarg */
1712 			       sctx->isc_tx_maxsize,		/* maxsize */
1713 			       nsegments,	/* nsegments */
1714 			       sctx->isc_tx_maxsegsize,	/* maxsegsize */
1715 			       0,			/* flags */
1716 			       NULL,			/* lockfunc */
1717 			       NULL,			/* lockfuncarg */
1718 			       &txq->ift_buf_tag))) {
1719 		device_printf(dev,"Unable to allocate TX DMA tag: %d\n", err);
1720 		device_printf(dev,"maxsize: %ju nsegments: %d maxsegsize: %ju\n",
1721 		    (uintmax_t)sctx->isc_tx_maxsize, nsegments, (uintmax_t)sctx->isc_tx_maxsegsize);
1722 		goto fail;
1723 	}
1724 	tso = (if_getcapabilities(ctx->ifc_ifp) & IFCAP_TSO) != 0;
1725 	if (tso && (err = bus_dma_tag_create(bus_get_dma_tag(dev),
1726 			       1, 0,			/* alignment, bounds */
1727 			       lowaddr,			/* lowaddr */
1728 			       BUS_SPACE_MAXADDR,	/* highaddr */
1729 			       NULL, NULL,		/* filter, filterarg */
1730 			       tsomaxsize,		/* maxsize */
1731 			       ntsosegments,	/* nsegments */
1732 			       sctx->isc_tso_maxsegsize,/* maxsegsize */
1733 			       0,			/* flags */
1734 			       NULL,			/* lockfunc */
1735 			       NULL,			/* lockfuncarg */
1736 			       &txq->ift_tso_buf_tag))) {
1737 		device_printf(dev, "Unable to allocate TSO TX DMA tag: %d\n",
1738 		    err);
1739 		goto fail;
1740 	}
1741 
1742 	/* Allocate memory for the TX mbuf map. */
1743 	if (!(txq->ift_sds.ifsd_m =
1744 	    (struct mbuf **) malloc(sizeof(struct mbuf *) *
1745 	    scctx->isc_ntxd[txq->ift_br_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1746 		device_printf(dev, "Unable to allocate TX mbuf map memory\n");
1747 		err = ENOMEM;
1748 		goto fail;
1749 	}
1750 
1751 	/*
1752 	 * Create the DMA maps for TX buffers.
1753 	 */
1754 	if ((txq->ift_sds.ifsd_map = (bus_dmamap_t *)malloc(
1755 	    sizeof(bus_dmamap_t) * scctx->isc_ntxd[txq->ift_br_offset],
1756 	    M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
1757 		device_printf(dev,
1758 		    "Unable to allocate TX buffer DMA map memory\n");
1759 		err = ENOMEM;
1760 		goto fail;
1761 	}
1762 	if (tso && (txq->ift_sds.ifsd_tso_map = (bus_dmamap_t *)malloc(
1763 	    sizeof(bus_dmamap_t) * scctx->isc_ntxd[txq->ift_br_offset],
1764 	    M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
1765 		device_printf(dev,
1766 		    "Unable to allocate TSO TX buffer map memory\n");
1767 		err = ENOMEM;
1768 		goto fail;
1769 	}
1770 	for (int i = 0; i < scctx->isc_ntxd[txq->ift_br_offset]; i++) {
1771 		err = bus_dmamap_create(txq->ift_buf_tag, 0,
1772 		    &txq->ift_sds.ifsd_map[i]);
1773 		if (err != 0) {
1774 			device_printf(dev, "Unable to create TX DMA map\n");
1775 			goto fail;
1776 		}
1777 		if (!tso)
1778 			continue;
1779 		err = bus_dmamap_create(txq->ift_tso_buf_tag, 0,
1780 		    &txq->ift_sds.ifsd_tso_map[i]);
1781 		if (err != 0) {
1782 			device_printf(dev, "Unable to create TSO TX DMA map\n");
1783 			goto fail;
1784 		}
1785 	}
1786 	return (0);
1787 fail:
1788 	/* We free all, it handles case where we are in the middle */
1789 	iflib_tx_structures_free(ctx);
1790 	return (err);
1791 }
1792 
1793 static void
1794 iflib_txsd_destroy(if_ctx_t ctx, iflib_txq_t txq, int i)
1795 {
1796 	bus_dmamap_t map;
1797 
1798 	if (txq->ift_sds.ifsd_map != NULL) {
1799 		map = txq->ift_sds.ifsd_map[i];
1800 		bus_dmamap_sync(txq->ift_buf_tag, map, BUS_DMASYNC_POSTWRITE);
1801 		bus_dmamap_unload(txq->ift_buf_tag, map);
1802 		bus_dmamap_destroy(txq->ift_buf_tag, map);
1803 		txq->ift_sds.ifsd_map[i] = NULL;
1804 	}
1805 
1806 	if (txq->ift_sds.ifsd_tso_map != NULL) {
1807 		map = txq->ift_sds.ifsd_tso_map[i];
1808 		bus_dmamap_sync(txq->ift_tso_buf_tag, map,
1809 		    BUS_DMASYNC_POSTWRITE);
1810 		bus_dmamap_unload(txq->ift_tso_buf_tag, map);
1811 		bus_dmamap_destroy(txq->ift_tso_buf_tag, map);
1812 		txq->ift_sds.ifsd_tso_map[i] = NULL;
1813 	}
1814 }
1815 
1816 static void
1817 iflib_txq_destroy(iflib_txq_t txq)
1818 {
1819 	if_ctx_t ctx = txq->ift_ctx;
1820 
1821 	for (int i = 0; i < txq->ift_size; i++)
1822 		iflib_txsd_destroy(ctx, txq, i);
1823 
1824 	if (txq->ift_br != NULL) {
1825 		ifmp_ring_free(txq->ift_br);
1826 		txq->ift_br = NULL;
1827 	}
1828 
1829 	mtx_destroy(&txq->ift_mtx);
1830 
1831 	if (txq->ift_sds.ifsd_map != NULL) {
1832 		free(txq->ift_sds.ifsd_map, M_IFLIB);
1833 		txq->ift_sds.ifsd_map = NULL;
1834 	}
1835 	if (txq->ift_sds.ifsd_tso_map != NULL) {
1836 		free(txq->ift_sds.ifsd_tso_map, M_IFLIB);
1837 		txq->ift_sds.ifsd_tso_map = NULL;
1838 	}
1839 	if (txq->ift_sds.ifsd_m != NULL) {
1840 		free(txq->ift_sds.ifsd_m, M_IFLIB);
1841 		txq->ift_sds.ifsd_m = NULL;
1842 	}
1843 	if (txq->ift_buf_tag != NULL) {
1844 		bus_dma_tag_destroy(txq->ift_buf_tag);
1845 		txq->ift_buf_tag = NULL;
1846 	}
1847 	if (txq->ift_tso_buf_tag != NULL) {
1848 		bus_dma_tag_destroy(txq->ift_tso_buf_tag);
1849 		txq->ift_tso_buf_tag = NULL;
1850 	}
1851 	if (txq->ift_ifdi != NULL) {
1852 		free(txq->ift_ifdi, M_IFLIB);
1853 	}
1854 }
1855 
1856 static void
1857 iflib_txsd_free(if_ctx_t ctx, iflib_txq_t txq, int i)
1858 {
1859 	struct mbuf **mp;
1860 
1861 	mp = &txq->ift_sds.ifsd_m[i];
1862 	if (*mp == NULL)
1863 		return;
1864 
1865 	if (txq->ift_sds.ifsd_map != NULL) {
1866 		bus_dmamap_sync(txq->ift_buf_tag,
1867 		    txq->ift_sds.ifsd_map[i], BUS_DMASYNC_POSTWRITE);
1868 		bus_dmamap_unload(txq->ift_buf_tag, txq->ift_sds.ifsd_map[i]);
1869 	}
1870 	if (txq->ift_sds.ifsd_tso_map != NULL) {
1871 		bus_dmamap_sync(txq->ift_tso_buf_tag,
1872 		    txq->ift_sds.ifsd_tso_map[i], BUS_DMASYNC_POSTWRITE);
1873 		bus_dmamap_unload(txq->ift_tso_buf_tag,
1874 		    txq->ift_sds.ifsd_tso_map[i]);
1875 	}
1876 	m_freem(*mp);
1877 	DBG_COUNTER_INC(tx_frees);
1878 	*mp = NULL;
1879 }
1880 
1881 static int
1882 iflib_txq_setup(iflib_txq_t txq)
1883 {
1884 	if_ctx_t ctx = txq->ift_ctx;
1885 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1886 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1887 	iflib_dma_info_t di;
1888 	int i;
1889 
1890 	/* Set number of descriptors available */
1891 	txq->ift_qstatus = IFLIB_QUEUE_IDLE;
1892 	/* XXX make configurable */
1893 	txq->ift_update_freq = IFLIB_DEFAULT_TX_UPDATE_FREQ;
1894 
1895 	/* Reset indices */
1896 	txq->ift_cidx_processed = 0;
1897 	txq->ift_pidx = txq->ift_cidx = txq->ift_npending = 0;
1898 	txq->ift_size = scctx->isc_ntxd[txq->ift_br_offset];
1899 
1900 	for (i = 0, di = txq->ift_ifdi; i < sctx->isc_ntxqs; i++, di++)
1901 		bzero((void *)di->idi_vaddr, di->idi_size);
1902 
1903 	IFDI_TXQ_SETUP(ctx, txq->ift_id);
1904 	for (i = 0, di = txq->ift_ifdi; i < sctx->isc_ntxqs; i++, di++)
1905 		bus_dmamap_sync(di->idi_tag, di->idi_map,
1906 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
1907 	return (0);
1908 }
1909 
1910 /*********************************************************************
1911  *
1912  *  Allocate DMA resources for RX buffers as well as memory for the RX
1913  *  mbuf map, direct RX cluster pointer map and RX cluster bus address
1914  *  map.  RX DMA map, RX mbuf map, direct RX cluster pointer map and
1915  *  RX cluster map are kept in a iflib_sw_rx_desc_array structure.
1916  *  Since we use use one entry in iflib_sw_rx_desc_array per received
1917  *  packet, the maximum number of entries we'll need is equal to the
1918  *  number of hardware receive descriptors that we've allocated.
1919  *
1920  **********************************************************************/
1921 static int
1922 iflib_rxsd_alloc(iflib_rxq_t rxq)
1923 {
1924 	if_ctx_t ctx = rxq->ifr_ctx;
1925 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1926 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1927 	device_t dev = ctx->ifc_dev;
1928 	iflib_fl_t fl;
1929 	bus_addr_t lowaddr;
1930 	int			err;
1931 
1932 	MPASS(scctx->isc_nrxd[0] > 0);
1933 	MPASS(scctx->isc_nrxd[rxq->ifr_fl_offset] > 0);
1934 
1935 	lowaddr = DMA_WIDTH_TO_BUS_LOWADDR(scctx->isc_dma_width);
1936 
1937 	fl = rxq->ifr_fl;
1938 	for (int i = 0; i <  rxq->ifr_nfl; i++, fl++) {
1939 		fl->ifl_size = scctx->isc_nrxd[rxq->ifr_fl_offset]; /* this isn't necessarily the same */
1940 		/* Set up DMA tag for RX buffers. */
1941 		err = bus_dma_tag_create(bus_get_dma_tag(dev), /* parent */
1942 					 1, 0,			/* alignment, bounds */
1943 					 lowaddr,		/* lowaddr */
1944 					 BUS_SPACE_MAXADDR,	/* highaddr */
1945 					 NULL, NULL,		/* filter, filterarg */
1946 					 sctx->isc_rx_maxsize,	/* maxsize */
1947 					 sctx->isc_rx_nsegments,	/* nsegments */
1948 					 sctx->isc_rx_maxsegsize,	/* maxsegsize */
1949 					 0,			/* flags */
1950 					 NULL,			/* lockfunc */
1951 					 NULL,			/* lockarg */
1952 					 &fl->ifl_buf_tag);
1953 		if (err) {
1954 			device_printf(dev,
1955 			    "Unable to allocate RX DMA tag: %d\n", err);
1956 			goto fail;
1957 		}
1958 
1959 		/* Allocate memory for the RX mbuf map. */
1960 		if (!(fl->ifl_sds.ifsd_m =
1961 		      (struct mbuf **) malloc(sizeof(struct mbuf *) *
1962 					      scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1963 			device_printf(dev,
1964 			    "Unable to allocate RX mbuf map memory\n");
1965 			err = ENOMEM;
1966 			goto fail;
1967 		}
1968 
1969 		/* Allocate memory for the direct RX cluster pointer map. */
1970 		if (!(fl->ifl_sds.ifsd_cl =
1971 		      (caddr_t *) malloc(sizeof(caddr_t) *
1972 					      scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1973 			device_printf(dev,
1974 			    "Unable to allocate RX cluster map memory\n");
1975 			err = ENOMEM;
1976 			goto fail;
1977 		}
1978 
1979 		/* Allocate memory for the RX cluster bus address map. */
1980 		if (!(fl->ifl_sds.ifsd_ba =
1981 		      (bus_addr_t *) malloc(sizeof(bus_addr_t) *
1982 					      scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1983 			device_printf(dev,
1984 			    "Unable to allocate RX bus address map memory\n");
1985 			err = ENOMEM;
1986 			goto fail;
1987 		}
1988 
1989 		/*
1990 		 * Create the DMA maps for RX buffers.
1991 		 */
1992 		if (!(fl->ifl_sds.ifsd_map =
1993 		      (bus_dmamap_t *) malloc(sizeof(bus_dmamap_t) * scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1994 			device_printf(dev,
1995 			    "Unable to allocate RX buffer DMA map memory\n");
1996 			err = ENOMEM;
1997 			goto fail;
1998 		}
1999 		for (int i = 0; i < scctx->isc_nrxd[rxq->ifr_fl_offset]; i++) {
2000 			err = bus_dmamap_create(fl->ifl_buf_tag, 0,
2001 			    &fl->ifl_sds.ifsd_map[i]);
2002 			if (err != 0) {
2003 				device_printf(dev, "Unable to create RX buffer DMA map\n");
2004 				goto fail;
2005 			}
2006 		}
2007 	}
2008 	return (0);
2009 
2010 fail:
2011 	iflib_rx_structures_free(ctx);
2012 	return (err);
2013 }
2014 
2015 /*
2016  * Internal service routines
2017  */
2018 
2019 struct rxq_refill_cb_arg {
2020 	int               error;
2021 	bus_dma_segment_t seg;
2022 	int               nseg;
2023 };
2024 
2025 static void
2026 _rxq_refill_cb(void *arg, bus_dma_segment_t *segs, int nseg, int error)
2027 {
2028 	struct rxq_refill_cb_arg *cb_arg = arg;
2029 
2030 	cb_arg->error = error;
2031 	cb_arg->seg = segs[0];
2032 	cb_arg->nseg = nseg;
2033 }
2034 
2035 /**
2036  * iflib_fl_refill - refill an rxq free-buffer list
2037  * @ctx: the iflib context
2038  * @fl: the free list to refill
2039  * @count: the number of new buffers to allocate
2040  *
2041  * (Re)populate an rxq free-buffer list with up to @count new packet buffers.
2042  * The caller must assure that @count does not exceed the queue's capacity
2043  * minus one (since we always leave a descriptor unavailable).
2044  */
2045 static uint8_t
2046 iflib_fl_refill(if_ctx_t ctx, iflib_fl_t fl, int count)
2047 {
2048 	struct if_rxd_update iru;
2049 	struct rxq_refill_cb_arg cb_arg;
2050 	struct mbuf *m;
2051 	caddr_t cl, *sd_cl;
2052 	struct mbuf **sd_m;
2053 	bus_dmamap_t *sd_map;
2054 	bus_addr_t bus_addr, *sd_ba;
2055 	int err, frag_idx, i, idx, n, pidx;
2056 	qidx_t credits;
2057 
2058 	MPASS(count <= fl->ifl_size - fl->ifl_credits - 1);
2059 
2060 	sd_m = fl->ifl_sds.ifsd_m;
2061 	sd_map = fl->ifl_sds.ifsd_map;
2062 	sd_cl = fl->ifl_sds.ifsd_cl;
2063 	sd_ba = fl->ifl_sds.ifsd_ba;
2064 	pidx = fl->ifl_pidx;
2065 	idx = pidx;
2066 	frag_idx = fl->ifl_fragidx;
2067 	credits = fl->ifl_credits;
2068 
2069 	i = 0;
2070 	n = count;
2071 	MPASS(n > 0);
2072 	MPASS(credits + n <= fl->ifl_size);
2073 
2074 	if (pidx < fl->ifl_cidx)
2075 		MPASS(pidx + n <= fl->ifl_cidx);
2076 	if (pidx == fl->ifl_cidx && (credits < fl->ifl_size))
2077 		MPASS(fl->ifl_gen == 0);
2078 	if (pidx > fl->ifl_cidx)
2079 		MPASS(n <= fl->ifl_size - pidx + fl->ifl_cidx);
2080 
2081 	DBG_COUNTER_INC(fl_refills);
2082 	if (n > 8)
2083 		DBG_COUNTER_INC(fl_refills_large);
2084 	iru_init(&iru, fl->ifl_rxq, fl->ifl_id);
2085 	while (n-- > 0) {
2086 		/*
2087 		 * We allocate an uninitialized mbuf + cluster, mbuf is
2088 		 * initialized after rx.
2089 		 *
2090 		 * If the cluster is still set then we know a minimum sized
2091 		 * packet was received
2092 		 */
2093 		bit_ffc_at(fl->ifl_rx_bitmap, frag_idx, fl->ifl_size,
2094 		    &frag_idx);
2095 		if (frag_idx < 0)
2096 			bit_ffc(fl->ifl_rx_bitmap, fl->ifl_size, &frag_idx);
2097 		MPASS(frag_idx >= 0);
2098 		if ((cl = sd_cl[frag_idx]) == NULL) {
2099 			cl = uma_zalloc(fl->ifl_zone, M_NOWAIT);
2100 			if (__predict_false(cl == NULL))
2101 				break;
2102 
2103 			cb_arg.error = 0;
2104 			MPASS(sd_map != NULL);
2105 			err = bus_dmamap_load(fl->ifl_buf_tag, sd_map[frag_idx],
2106 			    cl, fl->ifl_buf_size, _rxq_refill_cb, &cb_arg,
2107 			    BUS_DMA_NOWAIT);
2108 			if (__predict_false(err != 0 || cb_arg.error)) {
2109 				uma_zfree(fl->ifl_zone, cl);
2110 				break;
2111 			}
2112 
2113 			sd_ba[frag_idx] = bus_addr = cb_arg.seg.ds_addr;
2114 			sd_cl[frag_idx] = cl;
2115 #if MEMORY_LOGGING
2116 			fl->ifl_cl_enqueued++;
2117 #endif
2118 		} else {
2119 			bus_addr = sd_ba[frag_idx];
2120 		}
2121 		bus_dmamap_sync(fl->ifl_buf_tag, sd_map[frag_idx],
2122 		    BUS_DMASYNC_PREREAD);
2123 
2124 		if (sd_m[frag_idx] == NULL) {
2125 			m = m_gethdr(M_NOWAIT, MT_NOINIT);
2126 			if (__predict_false(m == NULL))
2127 				break;
2128 			sd_m[frag_idx] = m;
2129 		}
2130 		bit_set(fl->ifl_rx_bitmap, frag_idx);
2131 #if MEMORY_LOGGING
2132 		fl->ifl_m_enqueued++;
2133 #endif
2134 
2135 		DBG_COUNTER_INC(rx_allocs);
2136 		fl->ifl_rxd_idxs[i] = frag_idx;
2137 		fl->ifl_bus_addrs[i] = bus_addr;
2138 		credits++;
2139 		i++;
2140 		MPASS(credits <= fl->ifl_size);
2141 		if (++idx == fl->ifl_size) {
2142 #ifdef INVARIANTS
2143 			fl->ifl_gen = 1;
2144 #endif
2145 			idx = 0;
2146 		}
2147 		if (n == 0 || i == IFLIB_MAX_RX_REFRESH) {
2148 			iru.iru_pidx = pidx;
2149 			iru.iru_count = i;
2150 			ctx->isc_rxd_refill(ctx->ifc_softc, &iru);
2151 			fl->ifl_pidx = idx;
2152 			fl->ifl_credits = credits;
2153 			pidx = idx;
2154 			i = 0;
2155 		}
2156 	}
2157 
2158 	if (n < count - 1) {
2159 		if (i != 0) {
2160 			iru.iru_pidx = pidx;
2161 			iru.iru_count = i;
2162 			ctx->isc_rxd_refill(ctx->ifc_softc, &iru);
2163 			fl->ifl_pidx = idx;
2164 			fl->ifl_credits = credits;
2165 		}
2166 		DBG_COUNTER_INC(rxd_flush);
2167 		bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
2168 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
2169 		ctx->isc_rxd_flush(ctx->ifc_softc, fl->ifl_rxq->ifr_id,
2170 		    fl->ifl_id, fl->ifl_pidx);
2171 		if (__predict_true(bit_test(fl->ifl_rx_bitmap, frag_idx))) {
2172 			fl->ifl_fragidx = frag_idx + 1;
2173 			if (fl->ifl_fragidx == fl->ifl_size)
2174 				fl->ifl_fragidx = 0;
2175 		} else {
2176 			fl->ifl_fragidx = frag_idx;
2177 		}
2178 	}
2179 
2180 	return (n == -1 ? 0 : IFLIB_RXEOF_EMPTY);
2181 }
2182 
2183 static inline uint8_t
2184 iflib_fl_refill_all(if_ctx_t ctx, iflib_fl_t fl)
2185 {
2186 	/*
2187 	 * We leave an unused descriptor to avoid pidx to catch up with cidx.
2188 	 * This is important as it confuses most NICs. For instance,
2189 	 * Intel NICs have (per receive ring) RDH and RDT registers, where
2190 	 * RDH points to the next receive descriptor to be used by the NIC,
2191 	 * and RDT for the next receive descriptor to be published by the
2192 	 * driver to the NIC (RDT - 1 is thus the last valid one).
2193 	 * The condition RDH == RDT means no descriptors are available to
2194 	 * the NIC, and thus it would be ambiguous if it also meant that
2195 	 * all the descriptors are available to the NIC.
2196 	 */
2197 	int32_t reclaimable = fl->ifl_size - fl->ifl_credits - 1;
2198 #ifdef INVARIANTS
2199 	int32_t delta = fl->ifl_size - get_inuse(fl->ifl_size, fl->ifl_cidx, fl->ifl_pidx, fl->ifl_gen) - 1;
2200 #endif
2201 
2202 	MPASS(fl->ifl_credits <= fl->ifl_size);
2203 	MPASS(reclaimable == delta);
2204 
2205 	if (reclaimable > 0)
2206 		return (iflib_fl_refill(ctx, fl, reclaimable));
2207 	return (0);
2208 }
2209 
2210 uint8_t
2211 iflib_in_detach(if_ctx_t ctx)
2212 {
2213 	bool in_detach;
2214 
2215 	STATE_LOCK(ctx);
2216 	in_detach = !!(ctx->ifc_flags & IFC_IN_DETACH);
2217 	STATE_UNLOCK(ctx);
2218 	return (in_detach);
2219 }
2220 
2221 static void
2222 iflib_fl_bufs_free(iflib_fl_t fl)
2223 {
2224 	iflib_dma_info_t idi = fl->ifl_ifdi;
2225 	bus_dmamap_t sd_map;
2226 	uint32_t i;
2227 
2228 	for (i = 0; i < fl->ifl_size; i++) {
2229 		struct mbuf **sd_m = &fl->ifl_sds.ifsd_m[i];
2230 		caddr_t *sd_cl = &fl->ifl_sds.ifsd_cl[i];
2231 
2232 		if (*sd_cl != NULL) {
2233 			sd_map = fl->ifl_sds.ifsd_map[i];
2234 			bus_dmamap_sync(fl->ifl_buf_tag, sd_map,
2235 			    BUS_DMASYNC_POSTREAD);
2236 			bus_dmamap_unload(fl->ifl_buf_tag, sd_map);
2237 			uma_zfree(fl->ifl_zone, *sd_cl);
2238 			*sd_cl = NULL;
2239 			if (*sd_m != NULL) {
2240 				m_init(*sd_m, M_NOWAIT, MT_DATA, 0);
2241 				uma_zfree(zone_mbuf, *sd_m);
2242 				*sd_m = NULL;
2243 			}
2244 		} else {
2245 			MPASS(*sd_m == NULL);
2246 		}
2247 #if MEMORY_LOGGING
2248 		fl->ifl_m_dequeued++;
2249 		fl->ifl_cl_dequeued++;
2250 #endif
2251 	}
2252 #ifdef INVARIANTS
2253 	for (i = 0; i < fl->ifl_size; i++) {
2254 		MPASS(fl->ifl_sds.ifsd_cl[i] == NULL);
2255 		MPASS(fl->ifl_sds.ifsd_m[i] == NULL);
2256 	}
2257 #endif
2258 	/*
2259 	 * Reset free list values
2260 	 */
2261 	fl->ifl_credits = fl->ifl_cidx = fl->ifl_pidx = fl->ifl_gen = fl->ifl_fragidx = 0;
2262 	bzero(idi->idi_vaddr, idi->idi_size);
2263 }
2264 
2265 /*********************************************************************
2266  *
2267  *  Initialize a free list and its buffers.
2268  *
2269  **********************************************************************/
2270 static int
2271 iflib_fl_setup(iflib_fl_t fl)
2272 {
2273 	iflib_rxq_t rxq = fl->ifl_rxq;
2274 	if_ctx_t ctx = rxq->ifr_ctx;
2275 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2276 	int qidx;
2277 
2278 	bit_nclear(fl->ifl_rx_bitmap, 0, fl->ifl_size - 1);
2279 	/*
2280 	** Free current RX buffer structs and their mbufs
2281 	*/
2282 	iflib_fl_bufs_free(fl);
2283 	/* Now replenish the mbufs */
2284 	MPASS(fl->ifl_credits == 0);
2285 	qidx = rxq->ifr_fl_offset + fl->ifl_id;
2286 	if (scctx->isc_rxd_buf_size[qidx] != 0)
2287 		fl->ifl_buf_size = scctx->isc_rxd_buf_size[qidx];
2288 	else
2289 		fl->ifl_buf_size = ctx->ifc_rx_mbuf_sz;
2290 	/*
2291 	 * ifl_buf_size may be a driver-supplied value, so pull it up
2292 	 * to the selected mbuf size.
2293 	 */
2294 	fl->ifl_buf_size = iflib_get_mbuf_size_for(fl->ifl_buf_size);
2295 	if (fl->ifl_buf_size > ctx->ifc_max_fl_buf_size)
2296 		ctx->ifc_max_fl_buf_size = fl->ifl_buf_size;
2297 	fl->ifl_cltype = m_gettype(fl->ifl_buf_size);
2298 	fl->ifl_zone = m_getzone(fl->ifl_buf_size);
2299 
2300 	/*
2301 	 * Avoid pre-allocating zillions of clusters to an idle card
2302 	 * potentially speeding up attach. In any case make sure
2303 	 * to leave a descriptor unavailable. See the comment in
2304 	 * iflib_fl_refill_all().
2305 	 */
2306 	MPASS(fl->ifl_size > 0);
2307 	(void)iflib_fl_refill(ctx, fl, min(128, fl->ifl_size - 1));
2308 	if (min(128, fl->ifl_size - 1) != fl->ifl_credits)
2309 		return (ENOBUFS);
2310 	/*
2311 	 * handle failure
2312 	 */
2313 	MPASS(rxq != NULL);
2314 	MPASS(fl->ifl_ifdi != NULL);
2315 	bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
2316 	    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
2317 	return (0);
2318 }
2319 
2320 /*********************************************************************
2321  *
2322  *  Free receive ring data structures
2323  *
2324  **********************************************************************/
2325 static void
2326 iflib_rx_sds_free(iflib_rxq_t rxq)
2327 {
2328 	iflib_fl_t fl;
2329 	int i, j;
2330 
2331 	if (rxq->ifr_fl != NULL) {
2332 		for (i = 0; i < rxq->ifr_nfl; i++) {
2333 			fl = &rxq->ifr_fl[i];
2334 			if (fl->ifl_buf_tag != NULL) {
2335 				if (fl->ifl_sds.ifsd_map != NULL) {
2336 					for (j = 0; j < fl->ifl_size; j++) {
2337 						bus_dmamap_sync(
2338 						    fl->ifl_buf_tag,
2339 						    fl->ifl_sds.ifsd_map[j],
2340 						    BUS_DMASYNC_POSTREAD);
2341 						bus_dmamap_unload(
2342 						    fl->ifl_buf_tag,
2343 						    fl->ifl_sds.ifsd_map[j]);
2344 						bus_dmamap_destroy(
2345 						    fl->ifl_buf_tag,
2346 						    fl->ifl_sds.ifsd_map[j]);
2347 					}
2348 				}
2349 				bus_dma_tag_destroy(fl->ifl_buf_tag);
2350 				fl->ifl_buf_tag = NULL;
2351 			}
2352 			free(fl->ifl_sds.ifsd_m, M_IFLIB);
2353 			free(fl->ifl_sds.ifsd_cl, M_IFLIB);
2354 			free(fl->ifl_sds.ifsd_ba, M_IFLIB);
2355 			free(fl->ifl_sds.ifsd_map, M_IFLIB);
2356 			free(fl->ifl_rx_bitmap, M_IFLIB);
2357 			fl->ifl_sds.ifsd_m = NULL;
2358 			fl->ifl_sds.ifsd_cl = NULL;
2359 			fl->ifl_sds.ifsd_ba = NULL;
2360 			fl->ifl_sds.ifsd_map = NULL;
2361 			fl->ifl_rx_bitmap = NULL;
2362 		}
2363 		free(rxq->ifr_fl, M_IFLIB);
2364 		rxq->ifr_fl = NULL;
2365 		free(rxq->ifr_ifdi, M_IFLIB);
2366 		rxq->ifr_ifdi = NULL;
2367 		rxq->ifr_cq_cidx = 0;
2368 	}
2369 }
2370 
2371 /*
2372  * Timer routine
2373  */
2374 static void
2375 iflib_timer(void *arg)
2376 {
2377 	iflib_txq_t txq = arg;
2378 	if_ctx_t ctx = txq->ift_ctx;
2379 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
2380 	uint64_t this_tick = ticks;
2381 
2382 	if (!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING))
2383 		return;
2384 
2385 	/*
2386 	** Check on the state of the TX queue(s), this
2387 	** can be done without the lock because its RO
2388 	** and the HUNG state will be static if set.
2389 	*/
2390 	if (this_tick - txq->ift_last_timer_tick >= iflib_timer_default) {
2391 		txq->ift_last_timer_tick = this_tick;
2392 		IFDI_TIMER(ctx, txq->ift_id);
2393 		if ((txq->ift_qstatus == IFLIB_QUEUE_HUNG) &&
2394 		    ((txq->ift_cleaned_prev == txq->ift_cleaned) ||
2395 		     (sctx->isc_pause_frames == 0)))
2396 			goto hung;
2397 
2398 		if (txq->ift_qstatus != IFLIB_QUEUE_IDLE &&
2399 		    ifmp_ring_is_stalled(txq->ift_br)) {
2400 			KASSERT(ctx->ifc_link_state == LINK_STATE_UP,
2401 			    ("queue can't be marked as hung if interface is down"));
2402 			txq->ift_qstatus = IFLIB_QUEUE_HUNG;
2403 		}
2404 		txq->ift_cleaned_prev = txq->ift_cleaned;
2405 	}
2406 	/* handle any laggards */
2407 	if (txq->ift_db_pending)
2408 		GROUPTASK_ENQUEUE(&txq->ift_task);
2409 
2410 	sctx->isc_pause_frames = 0;
2411 	if (if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING)
2412 		callout_reset_on(&txq->ift_timer, iflib_timer_default, iflib_timer,
2413 		    txq, txq->ift_timer.c_cpu);
2414 	return;
2415 
2416  hung:
2417 	device_printf(ctx->ifc_dev,
2418 	    "Watchdog timeout (TX: %d desc avail: %d pidx: %d) -- resetting\n",
2419 	    txq->ift_id, TXQ_AVAIL(txq), txq->ift_pidx);
2420 	STATE_LOCK(ctx);
2421 	if_setdrvflagbits(ctx->ifc_ifp, IFF_DRV_OACTIVE, IFF_DRV_RUNNING);
2422 	ctx->ifc_flags |= (IFC_DO_WATCHDOG|IFC_DO_RESET);
2423 	iflib_admin_intr_deferred(ctx);
2424 	STATE_UNLOCK(ctx);
2425 }
2426 
2427 static uint16_t
2428 iflib_get_mbuf_size_for(unsigned int size)
2429 {
2430 
2431 	if (size <= MCLBYTES)
2432 		return (MCLBYTES);
2433 	else
2434 		return (MJUMPAGESIZE);
2435 }
2436 
2437 static void
2438 iflib_calc_rx_mbuf_sz(if_ctx_t ctx)
2439 {
2440 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
2441 
2442 	/*
2443 	 * XXX don't set the max_frame_size to larger
2444 	 * than the hardware can handle
2445 	 */
2446 	ctx->ifc_rx_mbuf_sz =
2447 	    iflib_get_mbuf_size_for(sctx->isc_max_frame_size);
2448 }
2449 
2450 uint32_t
2451 iflib_get_rx_mbuf_sz(if_ctx_t ctx)
2452 {
2453 
2454 	return (ctx->ifc_rx_mbuf_sz);
2455 }
2456 
2457 static void
2458 iflib_init_locked(if_ctx_t ctx)
2459 {
2460 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
2461 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2462 	if_t ifp = ctx->ifc_ifp;
2463 	iflib_fl_t fl;
2464 	iflib_txq_t txq;
2465 	iflib_rxq_t rxq;
2466 	int i, j, tx_ip_csum_flags, tx_ip6_csum_flags;
2467 
2468 	if_setdrvflagbits(ifp, IFF_DRV_OACTIVE, IFF_DRV_RUNNING);
2469 	IFDI_INTR_DISABLE(ctx);
2470 
2471 	/*
2472 	 * See iflib_stop(). Useful in case iflib_init_locked() is
2473 	 * called without first calling iflib_stop().
2474 	 */
2475 	netmap_disable_all_rings(ifp);
2476 
2477 	tx_ip_csum_flags = scctx->isc_tx_csum_flags & (CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP);
2478 	tx_ip6_csum_flags = scctx->isc_tx_csum_flags & (CSUM_IP6_TCP | CSUM_IP6_UDP | CSUM_IP6_SCTP);
2479 	/* Set hardware offload abilities */
2480 	if_clearhwassist(ifp);
2481 	if (if_getcapenable(ifp) & IFCAP_TXCSUM)
2482 		if_sethwassistbits(ifp, tx_ip_csum_flags, 0);
2483 	if (if_getcapenable(ifp) & IFCAP_TXCSUM_IPV6)
2484 		if_sethwassistbits(ifp,  tx_ip6_csum_flags, 0);
2485 	if (if_getcapenable(ifp) & IFCAP_TSO4)
2486 		if_sethwassistbits(ifp, CSUM_IP_TSO, 0);
2487 	if (if_getcapenable(ifp) & IFCAP_TSO6)
2488 		if_sethwassistbits(ifp, CSUM_IP6_TSO, 0);
2489 
2490 	for (i = 0, txq = ctx->ifc_txqs; i < sctx->isc_ntxqsets; i++, txq++) {
2491 		CALLOUT_LOCK(txq);
2492 		callout_stop(&txq->ift_timer);
2493 #ifdef DEV_NETMAP
2494 		callout_stop(&txq->ift_netmap_timer);
2495 #endif /* DEV_NETMAP */
2496 		CALLOUT_UNLOCK(txq);
2497 		(void)iflib_netmap_txq_init(ctx, txq);
2498 	}
2499 
2500 	/*
2501 	 * Calculate a suitable Rx mbuf size prior to calling IFDI_INIT, so
2502 	 * that drivers can use the value when setting up the hardware receive
2503 	 * buffers.
2504 	 */
2505 	iflib_calc_rx_mbuf_sz(ctx);
2506 
2507 #ifdef INVARIANTS
2508 	i = if_getdrvflags(ifp);
2509 #endif
2510 	IFDI_INIT(ctx);
2511 	MPASS(if_getdrvflags(ifp) == i);
2512 	for (i = 0, rxq = ctx->ifc_rxqs; i < sctx->isc_nrxqsets; i++, rxq++) {
2513 		if (iflib_netmap_rxq_init(ctx, rxq) > 0) {
2514 			/* This rxq is in netmap mode. Skip normal init. */
2515 			continue;
2516 		}
2517 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++) {
2518 			if (iflib_fl_setup(fl)) {
2519 				device_printf(ctx->ifc_dev,
2520 				    "setting up free list %d failed - "
2521 				    "check cluster settings\n", j);
2522 				goto done;
2523 			}
2524 		}
2525 	}
2526 done:
2527 	if_setdrvflagbits(ctx->ifc_ifp, IFF_DRV_RUNNING, IFF_DRV_OACTIVE);
2528 	IFDI_INTR_ENABLE(ctx);
2529 	txq = ctx->ifc_txqs;
2530 	for (i = 0; i < sctx->isc_ntxqsets; i++, txq++)
2531 		callout_reset_on(&txq->ift_timer, iflib_timer_default, iflib_timer, txq,
2532 			txq->ift_timer.c_cpu);
2533 
2534         /* Re-enable txsync/rxsync. */
2535 	netmap_enable_all_rings(ifp);
2536 }
2537 
2538 static int
2539 iflib_media_change(if_t ifp)
2540 {
2541 	if_ctx_t ctx = if_getsoftc(ifp);
2542 	int err;
2543 
2544 	CTX_LOCK(ctx);
2545 	if ((err = IFDI_MEDIA_CHANGE(ctx)) == 0)
2546 		iflib_if_init_locked(ctx);
2547 	CTX_UNLOCK(ctx);
2548 	return (err);
2549 }
2550 
2551 static void
2552 iflib_media_status(if_t ifp, struct ifmediareq *ifmr)
2553 {
2554 	if_ctx_t ctx = if_getsoftc(ifp);
2555 
2556 	CTX_LOCK(ctx);
2557 	IFDI_UPDATE_ADMIN_STATUS(ctx);
2558 	IFDI_MEDIA_STATUS(ctx, ifmr);
2559 	CTX_UNLOCK(ctx);
2560 }
2561 
2562 void
2563 iflib_stop(if_ctx_t ctx)
2564 {
2565 	iflib_txq_t txq = ctx->ifc_txqs;
2566 	iflib_rxq_t rxq = ctx->ifc_rxqs;
2567 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2568 	if_shared_ctx_t sctx = ctx->ifc_sctx;
2569 	iflib_dma_info_t di;
2570 	iflib_fl_t fl;
2571 	int i, j;
2572 
2573 	/* Tell the stack that the interface is no longer active */
2574 	if_setdrvflagbits(ctx->ifc_ifp, IFF_DRV_OACTIVE, IFF_DRV_RUNNING);
2575 
2576 	IFDI_INTR_DISABLE(ctx);
2577 	DELAY(1000);
2578 	IFDI_STOP(ctx);
2579 	DELAY(1000);
2580 
2581 	/*
2582 	 * Stop any pending txsync/rxsync and prevent new ones
2583 	 * form starting. Processes blocked in poll() will get
2584 	 * POLLERR.
2585 	 */
2586 	netmap_disable_all_rings(ctx->ifc_ifp);
2587 
2588 	iflib_debug_reset();
2589 	/* Wait for current tx queue users to exit to disarm watchdog timer. */
2590 	for (i = 0; i < scctx->isc_ntxqsets; i++, txq++) {
2591 		/* make sure all transmitters have completed before proceeding XXX */
2592 
2593 		CALLOUT_LOCK(txq);
2594 		callout_stop(&txq->ift_timer);
2595 #ifdef DEV_NETMAP
2596 		callout_stop(&txq->ift_netmap_timer);
2597 #endif /* DEV_NETMAP */
2598 		CALLOUT_UNLOCK(txq);
2599 
2600 		/* clean any enqueued buffers */
2601 		iflib_ifmp_purge(txq);
2602 		/* Free any existing tx buffers. */
2603 		for (j = 0; j < txq->ift_size; j++) {
2604 			iflib_txsd_free(ctx, txq, j);
2605 		}
2606 		txq->ift_processed = txq->ift_cleaned = txq->ift_cidx_processed = 0;
2607 		txq->ift_in_use = txq->ift_gen = txq->ift_cidx = txq->ift_pidx = txq->ift_no_desc_avail = 0;
2608 		txq->ift_closed = txq->ift_mbuf_defrag = txq->ift_mbuf_defrag_failed = 0;
2609 		txq->ift_no_tx_dma_setup = txq->ift_txd_encap_efbig = txq->ift_map_failed = 0;
2610 		txq->ift_pullups = 0;
2611 		ifmp_ring_reset_stats(txq->ift_br);
2612 		for (j = 0, di = txq->ift_ifdi; j < sctx->isc_ntxqs; j++, di++)
2613 			bzero((void *)di->idi_vaddr, di->idi_size);
2614 	}
2615 	for (i = 0; i < scctx->isc_nrxqsets; i++, rxq++) {
2616 		/* make sure all transmitters have completed before proceeding XXX */
2617 
2618 		rxq->ifr_cq_cidx = 0;
2619 		for (j = 0, di = rxq->ifr_ifdi; j < sctx->isc_nrxqs; j++, di++)
2620 			bzero((void *)di->idi_vaddr, di->idi_size);
2621 		/* also resets the free lists pidx/cidx */
2622 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++)
2623 			iflib_fl_bufs_free(fl);
2624 	}
2625 }
2626 
2627 static inline caddr_t
2628 calc_next_rxd(iflib_fl_t fl, int cidx)
2629 {
2630 	qidx_t size;
2631 	int nrxd;
2632 	caddr_t start, end, cur, next;
2633 
2634 	nrxd = fl->ifl_size;
2635 	size = fl->ifl_rxd_size;
2636 	start = fl->ifl_ifdi->idi_vaddr;
2637 
2638 	if (__predict_false(size == 0))
2639 		return (start);
2640 	cur = start + size*cidx;
2641 	end = start + size*nrxd;
2642 	next = CACHE_PTR_NEXT(cur);
2643 	return (next < end ? next : start);
2644 }
2645 
2646 static inline void
2647 prefetch_pkts(iflib_fl_t fl, int cidx)
2648 {
2649 	int nextptr;
2650 	int nrxd = fl->ifl_size;
2651 	caddr_t next_rxd;
2652 
2653 	nextptr = (cidx + CACHE_PTR_INCREMENT) & (nrxd-1);
2654 	prefetch(&fl->ifl_sds.ifsd_m[nextptr]);
2655 	prefetch(&fl->ifl_sds.ifsd_cl[nextptr]);
2656 	next_rxd = calc_next_rxd(fl, cidx);
2657 	prefetch(next_rxd);
2658 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 1) & (nrxd-1)]);
2659 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 2) & (nrxd-1)]);
2660 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 3) & (nrxd-1)]);
2661 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 4) & (nrxd-1)]);
2662 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 1) & (nrxd-1)]);
2663 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 2) & (nrxd-1)]);
2664 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 3) & (nrxd-1)]);
2665 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 4) & (nrxd-1)]);
2666 }
2667 
2668 static struct mbuf *
2669 rxd_frag_to_sd(iflib_rxq_t rxq, if_rxd_frag_t irf, bool unload, if_rxsd_t sd,
2670     int *pf_rv, if_rxd_info_t ri)
2671 {
2672 	bus_dmamap_t map;
2673 	iflib_fl_t fl;
2674 	caddr_t payload;
2675 	struct mbuf *m;
2676 	int flid, cidx, len, next;
2677 
2678 	map = NULL;
2679 	flid = irf->irf_flid;
2680 	cidx = irf->irf_idx;
2681 	fl = &rxq->ifr_fl[flid];
2682 	sd->ifsd_fl = fl;
2683 	m = fl->ifl_sds.ifsd_m[cidx];
2684 	sd->ifsd_cl = &fl->ifl_sds.ifsd_cl[cidx];
2685 	fl->ifl_credits--;
2686 #if MEMORY_LOGGING
2687 	fl->ifl_m_dequeued++;
2688 #endif
2689 	if (rxq->ifr_ctx->ifc_flags & IFC_PREFETCH)
2690 		prefetch_pkts(fl, cidx);
2691 	next = (cidx + CACHE_PTR_INCREMENT) & (fl->ifl_size-1);
2692 	prefetch(&fl->ifl_sds.ifsd_map[next]);
2693 	map = fl->ifl_sds.ifsd_map[cidx];
2694 
2695 	bus_dmamap_sync(fl->ifl_buf_tag, map, BUS_DMASYNC_POSTREAD);
2696 
2697 	if (rxq->pfil != NULL && PFIL_HOOKED_IN(rxq->pfil) && pf_rv != NULL &&
2698 	    irf->irf_len != 0) {
2699 		payload  = *sd->ifsd_cl;
2700 		payload +=  ri->iri_pad;
2701 		len = ri->iri_len - ri->iri_pad;
2702 		*pf_rv = pfil_run_hooks(rxq->pfil, payload, ri->iri_ifp,
2703 		    len | PFIL_MEMPTR | PFIL_IN, NULL);
2704 		switch (*pf_rv) {
2705 		case PFIL_DROPPED:
2706 		case PFIL_CONSUMED:
2707 			/*
2708 			 * The filter ate it.  Everything is recycled.
2709 			 */
2710 			m = NULL;
2711 			unload = 0;
2712 			break;
2713 		case PFIL_REALLOCED:
2714 			/*
2715 			 * The filter copied it.  Everything is recycled.
2716 			 */
2717 			m = pfil_mem2mbuf(payload);
2718 			unload = 0;
2719 			break;
2720 		case PFIL_PASS:
2721 			/*
2722 			 * Filter said it was OK, so receive like
2723 			 * normal
2724 			 */
2725 			fl->ifl_sds.ifsd_m[cidx] = NULL;
2726 			break;
2727 		default:
2728 			MPASS(0);
2729 		}
2730 	} else {
2731 		fl->ifl_sds.ifsd_m[cidx] = NULL;
2732 		if (pf_rv != NULL)
2733 			*pf_rv = PFIL_PASS;
2734 	}
2735 
2736 	if (unload && irf->irf_len != 0)
2737 		bus_dmamap_unload(fl->ifl_buf_tag, map);
2738 	fl->ifl_cidx = (fl->ifl_cidx + 1) & (fl->ifl_size-1);
2739 	if (__predict_false(fl->ifl_cidx == 0))
2740 		fl->ifl_gen = 0;
2741 	bit_clear(fl->ifl_rx_bitmap, cidx);
2742 	return (m);
2743 }
2744 
2745 static struct mbuf *
2746 assemble_segments(iflib_rxq_t rxq, if_rxd_info_t ri, if_rxsd_t sd, int *pf_rv)
2747 {
2748 	struct mbuf *m, *mh, *mt;
2749 	caddr_t cl;
2750 	int  *pf_rv_ptr, flags, i, padlen;
2751 	bool consumed;
2752 
2753 	i = 0;
2754 	mh = NULL;
2755 	consumed = false;
2756 	*pf_rv = PFIL_PASS;
2757 	pf_rv_ptr = pf_rv;
2758 	do {
2759 		m = rxd_frag_to_sd(rxq, &ri->iri_frags[i], !consumed, sd,
2760 		    pf_rv_ptr, ri);
2761 
2762 		MPASS(*sd->ifsd_cl != NULL);
2763 
2764 		/*
2765 		 * Exclude zero-length frags & frags from
2766 		 * packets the filter has consumed or dropped
2767 		 */
2768 		if (ri->iri_frags[i].irf_len == 0 || consumed ||
2769 		    *pf_rv == PFIL_CONSUMED || *pf_rv == PFIL_DROPPED) {
2770 			if (mh == NULL) {
2771 				/* everything saved here */
2772 				consumed = true;
2773 				pf_rv_ptr = NULL;
2774 				continue;
2775 			}
2776 			/* XXX we can save the cluster here, but not the mbuf */
2777 			m_init(m, M_NOWAIT, MT_DATA, 0);
2778 			m_free(m);
2779 			continue;
2780 		}
2781 		if (mh == NULL) {
2782 			flags = M_PKTHDR|M_EXT;
2783 			mh = mt = m;
2784 			padlen = ri->iri_pad;
2785 		} else {
2786 			flags = M_EXT;
2787 			mt->m_next = m;
2788 			mt = m;
2789 			/* assuming padding is only on the first fragment */
2790 			padlen = 0;
2791 		}
2792 		cl = *sd->ifsd_cl;
2793 		*sd->ifsd_cl = NULL;
2794 
2795 		/* Can these two be made one ? */
2796 		m_init(m, M_NOWAIT, MT_DATA, flags);
2797 		m_cljset(m, cl, sd->ifsd_fl->ifl_cltype);
2798 		/*
2799 		 * These must follow m_init and m_cljset
2800 		 */
2801 		m->m_data += padlen;
2802 		ri->iri_len -= padlen;
2803 		m->m_len = ri->iri_frags[i].irf_len;
2804 	} while (++i < ri->iri_nfrags);
2805 
2806 	return (mh);
2807 }
2808 
2809 /*
2810  * Process one software descriptor
2811  */
2812 static struct mbuf *
2813 iflib_rxd_pkt_get(iflib_rxq_t rxq, if_rxd_info_t ri)
2814 {
2815 	struct if_rxsd sd;
2816 	struct mbuf *m;
2817 	int pf_rv;
2818 
2819 	/* should I merge this back in now that the two paths are basically duplicated? */
2820 	if (ri->iri_nfrags == 1 &&
2821 	    ri->iri_frags[0].irf_len != 0 &&
2822 	    ri->iri_frags[0].irf_len <= MIN(IFLIB_RX_COPY_THRESH, MHLEN)) {
2823 		m = rxd_frag_to_sd(rxq, &ri->iri_frags[0], false, &sd,
2824 		    &pf_rv, ri);
2825 		if (pf_rv != PFIL_PASS && pf_rv != PFIL_REALLOCED)
2826 			return (m);
2827 		if (pf_rv == PFIL_PASS) {
2828 			m_init(m, M_NOWAIT, MT_DATA, M_PKTHDR);
2829 #ifndef __NO_STRICT_ALIGNMENT
2830 			if (!IP_ALIGNED(m) && ri->iri_pad == 0)
2831 				m->m_data += 2;
2832 #endif
2833 			memcpy(m->m_data, *sd.ifsd_cl, ri->iri_len);
2834 			m->m_len = ri->iri_frags[0].irf_len;
2835 			m->m_data += ri->iri_pad;
2836 			ri->iri_len -= ri->iri_pad;
2837 		}
2838 	} else {
2839 		m = assemble_segments(rxq, ri, &sd, &pf_rv);
2840 		if (m == NULL)
2841 			return (NULL);
2842 		if (pf_rv != PFIL_PASS && pf_rv != PFIL_REALLOCED)
2843 			return (m);
2844 	}
2845 	m->m_pkthdr.len = ri->iri_len;
2846 	m->m_pkthdr.rcvif = ri->iri_ifp;
2847 	m->m_flags |= ri->iri_flags;
2848 	m->m_pkthdr.ether_vtag = ri->iri_vtag;
2849 	m->m_pkthdr.flowid = ri->iri_flowid;
2850 	M_HASHTYPE_SET(m, ri->iri_rsstype);
2851 	m->m_pkthdr.csum_flags = ri->iri_csum_flags;
2852 	m->m_pkthdr.csum_data = ri->iri_csum_data;
2853 	return (m);
2854 }
2855 
2856 #if defined(INET6) || defined(INET)
2857 static void
2858 iflib_get_ip_forwarding(struct lro_ctrl *lc, bool *v4, bool *v6)
2859 {
2860 	CURVNET_SET(lc->ifp->if_vnet);
2861 #if defined(INET6)
2862 	*v6 = V_ip6_forwarding;
2863 #endif
2864 #if defined(INET)
2865 	*v4 = V_ipforwarding;
2866 #endif
2867 	CURVNET_RESTORE();
2868 }
2869 
2870 /*
2871  * Returns true if it's possible this packet could be LROed.
2872  * if it returns false, it is guaranteed that tcp_lro_rx()
2873  * would not return zero.
2874  */
2875 static bool
2876 iflib_check_lro_possible(struct mbuf *m, bool v4_forwarding, bool v6_forwarding)
2877 {
2878 	struct ether_header *eh;
2879 
2880 	eh = mtod(m, struct ether_header *);
2881 	switch (eh->ether_type) {
2882 #if defined(INET6)
2883 		case htons(ETHERTYPE_IPV6):
2884 			return (!v6_forwarding);
2885 #endif
2886 #if defined (INET)
2887 		case htons(ETHERTYPE_IP):
2888 			return (!v4_forwarding);
2889 #endif
2890 	}
2891 
2892 	return false;
2893 }
2894 #else
2895 static void
2896 iflib_get_ip_forwarding(struct lro_ctrl *lc __unused, bool *v4 __unused, bool *v6 __unused)
2897 {
2898 }
2899 #endif
2900 
2901 static void
2902 _task_fn_rx_watchdog(void *context)
2903 {
2904 	iflib_rxq_t rxq = context;
2905 
2906 	GROUPTASK_ENQUEUE(&rxq->ifr_task);
2907 }
2908 
2909 static uint8_t
2910 iflib_rxeof(iflib_rxq_t rxq, qidx_t budget)
2911 {
2912 	if_t ifp;
2913 	if_ctx_t ctx = rxq->ifr_ctx;
2914 	if_shared_ctx_t sctx = ctx->ifc_sctx;
2915 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2916 	int avail, i;
2917 	qidx_t *cidxp;
2918 	struct if_rxd_info ri;
2919 	int err, budget_left, rx_bytes, rx_pkts;
2920 	iflib_fl_t fl;
2921 	int lro_enabled;
2922 	bool v4_forwarding, v6_forwarding, lro_possible;
2923 	uint8_t retval = 0;
2924 
2925 	/*
2926 	 * XXX early demux data packets so that if_input processing only handles
2927 	 * acks in interrupt context
2928 	 */
2929 	struct mbuf *m, *mh, *mt, *mf;
2930 
2931 	NET_EPOCH_ASSERT();
2932 
2933 	lro_possible = v4_forwarding = v6_forwarding = false;
2934 	ifp = ctx->ifc_ifp;
2935 	mh = mt = NULL;
2936 	MPASS(budget > 0);
2937 	rx_pkts	= rx_bytes = 0;
2938 	if (sctx->isc_flags & IFLIB_HAS_RXCQ)
2939 		cidxp = &rxq->ifr_cq_cidx;
2940 	else
2941 		cidxp = &rxq->ifr_fl[0].ifl_cidx;
2942 	if ((avail = iflib_rxd_avail(ctx, rxq, *cidxp, budget)) == 0) {
2943 		for (i = 0, fl = &rxq->ifr_fl[0]; i < sctx->isc_nfl; i++, fl++)
2944 			retval |= iflib_fl_refill_all(ctx, fl);
2945 		DBG_COUNTER_INC(rx_unavail);
2946 		return (retval);
2947 	}
2948 
2949 	/* pfil needs the vnet to be set */
2950 	CURVNET_SET_QUIET(ifp->if_vnet);
2951 	for (budget_left = budget; budget_left > 0 && avail > 0;) {
2952 		if (__predict_false(!CTX_ACTIVE(ctx))) {
2953 			DBG_COUNTER_INC(rx_ctx_inactive);
2954 			break;
2955 		}
2956 		/*
2957 		 * Reset client set fields to their default values
2958 		 */
2959 		rxd_info_zero(&ri);
2960 		ri.iri_qsidx = rxq->ifr_id;
2961 		ri.iri_cidx = *cidxp;
2962 		ri.iri_ifp = ifp;
2963 		ri.iri_frags = rxq->ifr_frags;
2964 		err = ctx->isc_rxd_pkt_get(ctx->ifc_softc, &ri);
2965 
2966 		if (err)
2967 			goto err;
2968 		rx_pkts += 1;
2969 		rx_bytes += ri.iri_len;
2970 		if (sctx->isc_flags & IFLIB_HAS_RXCQ) {
2971 			*cidxp = ri.iri_cidx;
2972 			/* Update our consumer index */
2973 			/* XXX NB: shurd - check if this is still safe */
2974 			while (rxq->ifr_cq_cidx >= scctx->isc_nrxd[0])
2975 				rxq->ifr_cq_cidx -= scctx->isc_nrxd[0];
2976 			/* was this only a completion queue message? */
2977 			if (__predict_false(ri.iri_nfrags == 0))
2978 				continue;
2979 		}
2980 		MPASS(ri.iri_nfrags != 0);
2981 		MPASS(ri.iri_len != 0);
2982 
2983 		/* will advance the cidx on the corresponding free lists */
2984 		m = iflib_rxd_pkt_get(rxq, &ri);
2985 		avail--;
2986 		budget_left--;
2987 		if (avail == 0 && budget_left)
2988 			avail = iflib_rxd_avail(ctx, rxq, *cidxp, budget_left);
2989 
2990 		if (__predict_false(m == NULL))
2991 			continue;
2992 
2993 		/* imm_pkt: -- cxgb */
2994 		if (mh == NULL)
2995 			mh = mt = m;
2996 		else {
2997 			mt->m_nextpkt = m;
2998 			mt = m;
2999 		}
3000 	}
3001 	CURVNET_RESTORE();
3002 	/* make sure that we can refill faster than drain */
3003 	for (i = 0, fl = &rxq->ifr_fl[0]; i < sctx->isc_nfl; i++, fl++)
3004 		retval |= iflib_fl_refill_all(ctx, fl);
3005 
3006 	lro_enabled = (if_getcapenable(ifp) & IFCAP_LRO);
3007 	if (lro_enabled)
3008 		iflib_get_ip_forwarding(&rxq->ifr_lc, &v4_forwarding, &v6_forwarding);
3009 	mt = mf = NULL;
3010 	while (mh != NULL) {
3011 		m = mh;
3012 		mh = mh->m_nextpkt;
3013 		m->m_nextpkt = NULL;
3014 #ifndef __NO_STRICT_ALIGNMENT
3015 		if (!IP_ALIGNED(m) && (m = iflib_fixup_rx(m)) == NULL)
3016 			continue;
3017 #endif
3018 #if defined(INET6) || defined(INET)
3019 		if (lro_enabled) {
3020 			if (!lro_possible) {
3021 				lro_possible = iflib_check_lro_possible(m, v4_forwarding, v6_forwarding);
3022 				if (lro_possible && mf != NULL) {
3023 					ifp->if_input(ifp, mf);
3024 					DBG_COUNTER_INC(rx_if_input);
3025 					mt = mf = NULL;
3026 				}
3027 			}
3028 			if ((m->m_pkthdr.csum_flags & (CSUM_L4_CALC|CSUM_L4_VALID)) ==
3029 			    (CSUM_L4_CALC|CSUM_L4_VALID)) {
3030 				if (lro_possible && tcp_lro_rx(&rxq->ifr_lc, m, 0) == 0)
3031 					continue;
3032 			}
3033 		}
3034 #endif
3035 		if (lro_possible) {
3036 			ifp->if_input(ifp, m);
3037 			DBG_COUNTER_INC(rx_if_input);
3038 			continue;
3039 		}
3040 
3041 		if (mf == NULL)
3042 			mf = m;
3043 		if (mt != NULL)
3044 			mt->m_nextpkt = m;
3045 		mt = m;
3046 	}
3047 	if (mf != NULL) {
3048 		ifp->if_input(ifp, mf);
3049 		DBG_COUNTER_INC(rx_if_input);
3050 	}
3051 
3052 	if_inc_counter(ifp, IFCOUNTER_IBYTES, rx_bytes);
3053 	if_inc_counter(ifp, IFCOUNTER_IPACKETS, rx_pkts);
3054 
3055 	/*
3056 	 * Flush any outstanding LRO work
3057 	 */
3058 #if defined(INET6) || defined(INET)
3059 	tcp_lro_flush_all(&rxq->ifr_lc);
3060 #endif
3061 	if (avail != 0 || iflib_rxd_avail(ctx, rxq, *cidxp, 1) != 0)
3062 		retval |= IFLIB_RXEOF_MORE;
3063 	return (retval);
3064 err:
3065 	STATE_LOCK(ctx);
3066 	ctx->ifc_flags |= IFC_DO_RESET;
3067 	iflib_admin_intr_deferred(ctx);
3068 	STATE_UNLOCK(ctx);
3069 	return (0);
3070 }
3071 
3072 #define TXD_NOTIFY_COUNT(txq) (((txq)->ift_size / (txq)->ift_update_freq)-1)
3073 static inline qidx_t
3074 txq_max_db_deferred(iflib_txq_t txq, qidx_t in_use)
3075 {
3076 	qidx_t notify_count = TXD_NOTIFY_COUNT(txq);
3077 	qidx_t minthresh = txq->ift_size / 8;
3078 	if (in_use > 4*minthresh)
3079 		return (notify_count);
3080 	if (in_use > 2*minthresh)
3081 		return (notify_count >> 1);
3082 	if (in_use > minthresh)
3083 		return (notify_count >> 3);
3084 	return (0);
3085 }
3086 
3087 static inline qidx_t
3088 txq_max_rs_deferred(iflib_txq_t txq)
3089 {
3090 	qidx_t notify_count = TXD_NOTIFY_COUNT(txq);
3091 	qidx_t minthresh = txq->ift_size / 8;
3092 	if (txq->ift_in_use > 4*minthresh)
3093 		return (notify_count);
3094 	if (txq->ift_in_use > 2*minthresh)
3095 		return (notify_count >> 1);
3096 	if (txq->ift_in_use > minthresh)
3097 		return (notify_count >> 2);
3098 	return (2);
3099 }
3100 
3101 #define M_CSUM_FLAGS(m) ((m)->m_pkthdr.csum_flags)
3102 #define M_HAS_VLANTAG(m) (m->m_flags & M_VLANTAG)
3103 
3104 #define TXQ_MAX_DB_DEFERRED(txq, in_use) txq_max_db_deferred((txq), (in_use))
3105 #define TXQ_MAX_RS_DEFERRED(txq) txq_max_rs_deferred(txq)
3106 #define TXQ_MAX_DB_CONSUMED(size) (size >> 4)
3107 
3108 /* forward compatibility for cxgb */
3109 #define FIRST_QSET(ctx) 0
3110 #define NTXQSETS(ctx) ((ctx)->ifc_softc_ctx.isc_ntxqsets)
3111 #define NRXQSETS(ctx) ((ctx)->ifc_softc_ctx.isc_nrxqsets)
3112 #define QIDX(ctx, m) ((((m)->m_pkthdr.flowid & ctx->ifc_softc_ctx.isc_rss_table_mask) % NTXQSETS(ctx)) + FIRST_QSET(ctx))
3113 #define DESC_RECLAIMABLE(q) ((int)((q)->ift_processed - (q)->ift_cleaned - (q)->ift_ctx->ifc_softc_ctx.isc_tx_nsegments))
3114 
3115 /* XXX we should be setting this to something other than zero */
3116 #define RECLAIM_THRESH(ctx) ((ctx)->ifc_sctx->isc_tx_reclaim_thresh)
3117 #define	MAX_TX_DESC(ctx) MAX((ctx)->ifc_softc_ctx.isc_tx_tso_segments_max, \
3118     (ctx)->ifc_softc_ctx.isc_tx_nsegments)
3119 
3120 static inline bool
3121 iflib_txd_db_check(iflib_txq_t txq, int ring)
3122 {
3123 	if_ctx_t ctx = txq->ift_ctx;
3124 	qidx_t dbval, max;
3125 
3126 	max = TXQ_MAX_DB_DEFERRED(txq, txq->ift_in_use);
3127 
3128 	/* force || threshold exceeded || at the edge of the ring */
3129 	if (ring || (txq->ift_db_pending >= max) || (TXQ_AVAIL(txq) <= MAX_TX_DESC(ctx) + 2)) {
3130 
3131 		/*
3132 		 * 'npending' is used if the card's doorbell is in terms of the number of descriptors
3133 		 * pending flush (BRCM). 'pidx' is used in cases where the card's doorbeel uses the
3134 		 * producer index explicitly (INTC).
3135 		 */
3136 		dbval = txq->ift_npending ? txq->ift_npending : txq->ift_pidx;
3137 		bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
3138 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
3139 		ctx->isc_txd_flush(ctx->ifc_softc, txq->ift_id, dbval);
3140 
3141 		/*
3142 		 * Absent bugs there are zero packets pending so reset pending counts to zero.
3143 		 */
3144 		txq->ift_db_pending = txq->ift_npending = 0;
3145 		return (true);
3146 	}
3147 	return (false);
3148 }
3149 
3150 #ifdef PKT_DEBUG
3151 static void
3152 print_pkt(if_pkt_info_t pi)
3153 {
3154 	printf("pi len:  %d qsidx: %d nsegs: %d ndescs: %d flags: %x pidx: %d\n",
3155 	       pi->ipi_len, pi->ipi_qsidx, pi->ipi_nsegs, pi->ipi_ndescs, pi->ipi_flags, pi->ipi_pidx);
3156 	printf("pi new_pidx: %d csum_flags: %lx tso_segsz: %d mflags: %x vtag: %d\n",
3157 	       pi->ipi_new_pidx, pi->ipi_csum_flags, pi->ipi_tso_segsz, pi->ipi_mflags, pi->ipi_vtag);
3158 	printf("pi etype: %d ehdrlen: %d ip_hlen: %d ipproto: %d\n",
3159 	       pi->ipi_etype, pi->ipi_ehdrlen, pi->ipi_ip_hlen, pi->ipi_ipproto);
3160 }
3161 #endif
3162 
3163 #define IS_TSO4(pi) ((pi)->ipi_csum_flags & CSUM_IP_TSO)
3164 #define IS_TX_OFFLOAD4(pi) ((pi)->ipi_csum_flags & (CSUM_IP_TCP | CSUM_IP_TSO))
3165 #define IS_TSO6(pi) ((pi)->ipi_csum_flags & CSUM_IP6_TSO)
3166 #define IS_TX_OFFLOAD6(pi) ((pi)->ipi_csum_flags & (CSUM_IP6_TCP | CSUM_IP6_TSO))
3167 
3168 static int
3169 iflib_parse_header(iflib_txq_t txq, if_pkt_info_t pi, struct mbuf **mp)
3170 {
3171 	if_shared_ctx_t sctx = txq->ift_ctx->ifc_sctx;
3172 	struct ether_vlan_header *eh;
3173 	struct mbuf *m;
3174 
3175 	m = *mp;
3176 	if ((sctx->isc_flags & IFLIB_NEED_SCRATCH) &&
3177 	    M_WRITABLE(m) == 0) {
3178 		if ((m = m_dup(m, M_NOWAIT)) == NULL) {
3179 			return (ENOMEM);
3180 		} else {
3181 			m_freem(*mp);
3182 			DBG_COUNTER_INC(tx_frees);
3183 			*mp = m;
3184 		}
3185 	}
3186 
3187 	/*
3188 	 * Determine where frame payload starts.
3189 	 * Jump over vlan headers if already present,
3190 	 * helpful for QinQ too.
3191 	 */
3192 	if (__predict_false(m->m_len < sizeof(*eh))) {
3193 		txq->ift_pullups++;
3194 		if (__predict_false((m = m_pullup(m, sizeof(*eh))) == NULL))
3195 			return (ENOMEM);
3196 	}
3197 	eh = mtod(m, struct ether_vlan_header *);
3198 	if (eh->evl_encap_proto == htons(ETHERTYPE_VLAN)) {
3199 		pi->ipi_etype = ntohs(eh->evl_proto);
3200 		pi->ipi_ehdrlen = ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN;
3201 	} else {
3202 		pi->ipi_etype = ntohs(eh->evl_encap_proto);
3203 		pi->ipi_ehdrlen = ETHER_HDR_LEN;
3204 	}
3205 
3206 	switch (pi->ipi_etype) {
3207 #ifdef INET
3208 	case ETHERTYPE_IP:
3209 	{
3210 		struct mbuf *n;
3211 		struct ip *ip = NULL;
3212 		struct tcphdr *th = NULL;
3213 		int minthlen;
3214 
3215 		minthlen = min(m->m_pkthdr.len, pi->ipi_ehdrlen + sizeof(*ip) + sizeof(*th));
3216 		if (__predict_false(m->m_len < minthlen)) {
3217 			/*
3218 			 * if this code bloat is causing too much of a hit
3219 			 * move it to a separate function and mark it noinline
3220 			 */
3221 			if (m->m_len == pi->ipi_ehdrlen) {
3222 				n = m->m_next;
3223 				MPASS(n);
3224 				if (n->m_len >= sizeof(*ip))  {
3225 					ip = (struct ip *)n->m_data;
3226 					if (n->m_len >= (ip->ip_hl << 2) + sizeof(*th))
3227 						th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
3228 				} else {
3229 					txq->ift_pullups++;
3230 					if (__predict_false((m = m_pullup(m, minthlen)) == NULL))
3231 						return (ENOMEM);
3232 					ip = (struct ip *)(m->m_data + pi->ipi_ehdrlen);
3233 				}
3234 			} else {
3235 				txq->ift_pullups++;
3236 				if (__predict_false((m = m_pullup(m, minthlen)) == NULL))
3237 					return (ENOMEM);
3238 				ip = (struct ip *)(m->m_data + pi->ipi_ehdrlen);
3239 				if (m->m_len >= (ip->ip_hl << 2) + sizeof(*th))
3240 					th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
3241 			}
3242 		} else {
3243 			ip = (struct ip *)(m->m_data + pi->ipi_ehdrlen);
3244 			if (m->m_len >= (ip->ip_hl << 2) + sizeof(*th))
3245 				th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
3246 		}
3247 		pi->ipi_ip_hlen = ip->ip_hl << 2;
3248 		pi->ipi_ipproto = ip->ip_p;
3249 		pi->ipi_flags |= IPI_TX_IPV4;
3250 
3251 		/* TCP checksum offload may require TCP header length */
3252 		if (IS_TX_OFFLOAD4(pi)) {
3253 			if (__predict_true(pi->ipi_ipproto == IPPROTO_TCP)) {
3254 				if (__predict_false(th == NULL)) {
3255 					txq->ift_pullups++;
3256 					if (__predict_false((m = m_pullup(m, (ip->ip_hl << 2) + sizeof(*th))) == NULL))
3257 						return (ENOMEM);
3258 					th = (struct tcphdr *)((caddr_t)ip + pi->ipi_ip_hlen);
3259 				}
3260 				pi->ipi_tcp_hflags = th->th_flags;
3261 				pi->ipi_tcp_hlen = th->th_off << 2;
3262 				pi->ipi_tcp_seq = th->th_seq;
3263 			}
3264 			if (IS_TSO4(pi)) {
3265 				if (__predict_false(ip->ip_p != IPPROTO_TCP))
3266 					return (ENXIO);
3267 				/*
3268 				 * TSO always requires hardware checksum offload.
3269 				 */
3270 				pi->ipi_csum_flags |= (CSUM_IP_TCP | CSUM_IP);
3271 				th->th_sum = in_pseudo(ip->ip_src.s_addr,
3272 						       ip->ip_dst.s_addr, htons(IPPROTO_TCP));
3273 				pi->ipi_tso_segsz = m->m_pkthdr.tso_segsz;
3274 				if (sctx->isc_flags & IFLIB_TSO_INIT_IP) {
3275 					ip->ip_sum = 0;
3276 					ip->ip_len = htons(pi->ipi_ip_hlen + pi->ipi_tcp_hlen + pi->ipi_tso_segsz);
3277 				}
3278 			}
3279 		}
3280 		if ((sctx->isc_flags & IFLIB_NEED_ZERO_CSUM) && (pi->ipi_csum_flags & CSUM_IP))
3281                        ip->ip_sum = 0;
3282 
3283 		break;
3284 	}
3285 #endif
3286 #ifdef INET6
3287 	case ETHERTYPE_IPV6:
3288 	{
3289 		struct ip6_hdr *ip6 = (struct ip6_hdr *)(m->m_data + pi->ipi_ehdrlen);
3290 		struct tcphdr *th;
3291 		pi->ipi_ip_hlen = sizeof(struct ip6_hdr);
3292 
3293 		if (__predict_false(m->m_len < pi->ipi_ehdrlen + sizeof(struct ip6_hdr))) {
3294 			txq->ift_pullups++;
3295 			if (__predict_false((m = m_pullup(m, pi->ipi_ehdrlen + sizeof(struct ip6_hdr))) == NULL))
3296 				return (ENOMEM);
3297 		}
3298 		th = (struct tcphdr *)((caddr_t)ip6 + pi->ipi_ip_hlen);
3299 
3300 		/* XXX-BZ this will go badly in case of ext hdrs. */
3301 		pi->ipi_ipproto = ip6->ip6_nxt;
3302 		pi->ipi_flags |= IPI_TX_IPV6;
3303 
3304 		/* TCP checksum offload may require TCP header length */
3305 		if (IS_TX_OFFLOAD6(pi)) {
3306 			if (pi->ipi_ipproto == IPPROTO_TCP) {
3307 				if (__predict_false(m->m_len < pi->ipi_ehdrlen + sizeof(struct ip6_hdr) + sizeof(struct tcphdr))) {
3308 					txq->ift_pullups++;
3309 					if (__predict_false((m = m_pullup(m, pi->ipi_ehdrlen + sizeof(struct ip6_hdr) + sizeof(struct tcphdr))) == NULL))
3310 						return (ENOMEM);
3311 				}
3312 				pi->ipi_tcp_hflags = th->th_flags;
3313 				pi->ipi_tcp_hlen = th->th_off << 2;
3314 				pi->ipi_tcp_seq = th->th_seq;
3315 			}
3316 			if (IS_TSO6(pi)) {
3317 				if (__predict_false(ip6->ip6_nxt != IPPROTO_TCP))
3318 					return (ENXIO);
3319 				/*
3320 				 * TSO always requires hardware checksum offload.
3321 				 */
3322 				pi->ipi_csum_flags |= CSUM_IP6_TCP;
3323 				th->th_sum = in6_cksum_pseudo(ip6, 0, IPPROTO_TCP, 0);
3324 				pi->ipi_tso_segsz = m->m_pkthdr.tso_segsz;
3325 			}
3326 		}
3327 		break;
3328 	}
3329 #endif
3330 	default:
3331 		pi->ipi_csum_flags &= ~CSUM_OFFLOAD;
3332 		pi->ipi_ip_hlen = 0;
3333 		break;
3334 	}
3335 	*mp = m;
3336 
3337 	return (0);
3338 }
3339 
3340 /*
3341  * If dodgy hardware rejects the scatter gather chain we've handed it
3342  * we'll need to remove the mbuf chain from ifsg_m[] before we can add the
3343  * m_defrag'd mbufs
3344  */
3345 static __noinline struct mbuf *
3346 iflib_remove_mbuf(iflib_txq_t txq)
3347 {
3348 	int ntxd, pidx;
3349 	struct mbuf *m, **ifsd_m;
3350 
3351 	ifsd_m = txq->ift_sds.ifsd_m;
3352 	ntxd = txq->ift_size;
3353 	pidx = txq->ift_pidx & (ntxd - 1);
3354 	ifsd_m = txq->ift_sds.ifsd_m;
3355 	m = ifsd_m[pidx];
3356 	ifsd_m[pidx] = NULL;
3357 	bus_dmamap_unload(txq->ift_buf_tag, txq->ift_sds.ifsd_map[pidx]);
3358 	if (txq->ift_sds.ifsd_tso_map != NULL)
3359 		bus_dmamap_unload(txq->ift_tso_buf_tag,
3360 		    txq->ift_sds.ifsd_tso_map[pidx]);
3361 #if MEMORY_LOGGING
3362 	txq->ift_dequeued++;
3363 #endif
3364 	return (m);
3365 }
3366 
3367 static inline caddr_t
3368 calc_next_txd(iflib_txq_t txq, int cidx, uint8_t qid)
3369 {
3370 	qidx_t size;
3371 	int ntxd;
3372 	caddr_t start, end, cur, next;
3373 
3374 	ntxd = txq->ift_size;
3375 	size = txq->ift_txd_size[qid];
3376 	start = txq->ift_ifdi[qid].idi_vaddr;
3377 
3378 	if (__predict_false(size == 0))
3379 		return (start);
3380 	cur = start + size*cidx;
3381 	end = start + size*ntxd;
3382 	next = CACHE_PTR_NEXT(cur);
3383 	return (next < end ? next : start);
3384 }
3385 
3386 /*
3387  * Pad an mbuf to ensure a minimum ethernet frame size.
3388  * min_frame_size is the frame size (less CRC) to pad the mbuf to
3389  */
3390 static __noinline int
3391 iflib_ether_pad(device_t dev, struct mbuf **m_head, uint16_t min_frame_size)
3392 {
3393 	/*
3394 	 * 18 is enough bytes to pad an ARP packet to 46 bytes, and
3395 	 * and ARP message is the smallest common payload I can think of
3396 	 */
3397 	static char pad[18];	/* just zeros */
3398 	int n;
3399 	struct mbuf *new_head;
3400 
3401 	if (!M_WRITABLE(*m_head)) {
3402 		new_head = m_dup(*m_head, M_NOWAIT);
3403 		if (new_head == NULL) {
3404 			m_freem(*m_head);
3405 			device_printf(dev, "cannot pad short frame, m_dup() failed");
3406 			DBG_COUNTER_INC(encap_pad_mbuf_fail);
3407 			DBG_COUNTER_INC(tx_frees);
3408 			return ENOMEM;
3409 		}
3410 		m_freem(*m_head);
3411 		*m_head = new_head;
3412 	}
3413 
3414 	for (n = min_frame_size - (*m_head)->m_pkthdr.len;
3415 	     n > 0; n -= sizeof(pad))
3416 		if (!m_append(*m_head, min(n, sizeof(pad)), pad))
3417 			break;
3418 
3419 	if (n > 0) {
3420 		m_freem(*m_head);
3421 		device_printf(dev, "cannot pad short frame\n");
3422 		DBG_COUNTER_INC(encap_pad_mbuf_fail);
3423 		DBG_COUNTER_INC(tx_frees);
3424 		return (ENOBUFS);
3425 	}
3426 
3427 	return 0;
3428 }
3429 
3430 static int
3431 iflib_encap(iflib_txq_t txq, struct mbuf **m_headp)
3432 {
3433 	if_ctx_t		ctx;
3434 	if_shared_ctx_t		sctx;
3435 	if_softc_ctx_t		scctx;
3436 	bus_dma_tag_t		buf_tag;
3437 	bus_dma_segment_t	*segs;
3438 	struct mbuf		*m_head, **ifsd_m;
3439 	void			*next_txd;
3440 	bus_dmamap_t		map;
3441 	struct if_pkt_info	pi;
3442 	int remap = 0;
3443 	int err, nsegs, ndesc, max_segs, pidx, cidx, next, ntxd;
3444 
3445 	ctx = txq->ift_ctx;
3446 	sctx = ctx->ifc_sctx;
3447 	scctx = &ctx->ifc_softc_ctx;
3448 	segs = txq->ift_segs;
3449 	ntxd = txq->ift_size;
3450 	m_head = *m_headp;
3451 	map = NULL;
3452 
3453 	/*
3454 	 * If we're doing TSO the next descriptor to clean may be quite far ahead
3455 	 */
3456 	cidx = txq->ift_cidx;
3457 	pidx = txq->ift_pidx;
3458 	if (ctx->ifc_flags & IFC_PREFETCH) {
3459 		next = (cidx + CACHE_PTR_INCREMENT) & (ntxd-1);
3460 		if (!(ctx->ifc_flags & IFLIB_HAS_TXCQ)) {
3461 			next_txd = calc_next_txd(txq, cidx, 0);
3462 			prefetch(next_txd);
3463 		}
3464 
3465 		/* prefetch the next cache line of mbuf pointers and flags */
3466 		prefetch(&txq->ift_sds.ifsd_m[next]);
3467 		prefetch(&txq->ift_sds.ifsd_map[next]);
3468 		next = (cidx + CACHE_LINE_SIZE) & (ntxd-1);
3469 	}
3470 	map = txq->ift_sds.ifsd_map[pidx];
3471 	ifsd_m = txq->ift_sds.ifsd_m;
3472 
3473 	if (m_head->m_pkthdr.csum_flags & CSUM_TSO) {
3474 		buf_tag = txq->ift_tso_buf_tag;
3475 		max_segs = scctx->isc_tx_tso_segments_max;
3476 		map = txq->ift_sds.ifsd_tso_map[pidx];
3477 		MPASS(buf_tag != NULL);
3478 		MPASS(max_segs > 0);
3479 	} else {
3480 		buf_tag = txq->ift_buf_tag;
3481 		max_segs = scctx->isc_tx_nsegments;
3482 		map = txq->ift_sds.ifsd_map[pidx];
3483 	}
3484 	if ((sctx->isc_flags & IFLIB_NEED_ETHER_PAD) &&
3485 	    __predict_false(m_head->m_pkthdr.len < scctx->isc_min_frame_size)) {
3486 		err = iflib_ether_pad(ctx->ifc_dev, m_headp, scctx->isc_min_frame_size);
3487 		if (err) {
3488 			DBG_COUNTER_INC(encap_txd_encap_fail);
3489 			return err;
3490 		}
3491 	}
3492 	m_head = *m_headp;
3493 
3494 	pkt_info_zero(&pi);
3495 	pi.ipi_mflags = (m_head->m_flags & (M_VLANTAG|M_BCAST|M_MCAST));
3496 	pi.ipi_pidx = pidx;
3497 	pi.ipi_qsidx = txq->ift_id;
3498 	pi.ipi_len = m_head->m_pkthdr.len;
3499 	pi.ipi_csum_flags = m_head->m_pkthdr.csum_flags;
3500 	pi.ipi_vtag = M_HAS_VLANTAG(m_head) ? m_head->m_pkthdr.ether_vtag : 0;
3501 
3502 	/* deliberate bitwise OR to make one condition */
3503 	if (__predict_true((pi.ipi_csum_flags | pi.ipi_vtag))) {
3504 		if (__predict_false((err = iflib_parse_header(txq, &pi, m_headp)) != 0)) {
3505 			DBG_COUNTER_INC(encap_txd_encap_fail);
3506 			return (err);
3507 		}
3508 		m_head = *m_headp;
3509 	}
3510 
3511 retry:
3512 	err = bus_dmamap_load_mbuf_sg(buf_tag, map, m_head, segs, &nsegs,
3513 	    BUS_DMA_NOWAIT);
3514 defrag:
3515 	if (__predict_false(err)) {
3516 		switch (err) {
3517 		case EFBIG:
3518 			/* try collapse once and defrag once */
3519 			if (remap == 0) {
3520 				m_head = m_collapse(*m_headp, M_NOWAIT, max_segs);
3521 				/* try defrag if collapsing fails */
3522 				if (m_head == NULL)
3523 					remap++;
3524 			}
3525 			if (remap == 1) {
3526 				txq->ift_mbuf_defrag++;
3527 				m_head = m_defrag(*m_headp, M_NOWAIT);
3528 			}
3529 			/*
3530 			 * remap should never be >1 unless bus_dmamap_load_mbuf_sg
3531 			 * failed to map an mbuf that was run through m_defrag
3532 			 */
3533 			MPASS(remap <= 1);
3534 			if (__predict_false(m_head == NULL || remap > 1))
3535 				goto defrag_failed;
3536 			remap++;
3537 			*m_headp = m_head;
3538 			goto retry;
3539 			break;
3540 		case ENOMEM:
3541 			txq->ift_no_tx_dma_setup++;
3542 			break;
3543 		default:
3544 			txq->ift_no_tx_dma_setup++;
3545 			m_freem(*m_headp);
3546 			DBG_COUNTER_INC(tx_frees);
3547 			*m_headp = NULL;
3548 			break;
3549 		}
3550 		txq->ift_map_failed++;
3551 		DBG_COUNTER_INC(encap_load_mbuf_fail);
3552 		DBG_COUNTER_INC(encap_txd_encap_fail);
3553 		return (err);
3554 	}
3555 	ifsd_m[pidx] = m_head;
3556 	/*
3557 	 * XXX assumes a 1 to 1 relationship between segments and
3558 	 *        descriptors - this does not hold true on all drivers, e.g.
3559 	 *        cxgb
3560 	 */
3561 	if (__predict_false(nsegs + 2 > TXQ_AVAIL(txq))) {
3562 		txq->ift_no_desc_avail++;
3563 		bus_dmamap_unload(buf_tag, map);
3564 		DBG_COUNTER_INC(encap_txq_avail_fail);
3565 		DBG_COUNTER_INC(encap_txd_encap_fail);
3566 		if ((txq->ift_task.gt_task.ta_flags & TASK_ENQUEUED) == 0)
3567 			GROUPTASK_ENQUEUE(&txq->ift_task);
3568 		return (ENOBUFS);
3569 	}
3570 	/*
3571 	 * On Intel cards we can greatly reduce the number of TX interrupts
3572 	 * we see by only setting report status on every Nth descriptor.
3573 	 * However, this also means that the driver will need to keep track
3574 	 * of the descriptors that RS was set on to check them for the DD bit.
3575 	 */
3576 	txq->ift_rs_pending += nsegs + 1;
3577 	if (txq->ift_rs_pending > TXQ_MAX_RS_DEFERRED(txq) ||
3578 	     iflib_no_tx_batch || (TXQ_AVAIL(txq) - nsegs) <= MAX_TX_DESC(ctx) + 2) {
3579 		pi.ipi_flags |= IPI_TX_INTR;
3580 		txq->ift_rs_pending = 0;
3581 	}
3582 
3583 	pi.ipi_segs = segs;
3584 	pi.ipi_nsegs = nsegs;
3585 
3586 	MPASS(pidx >= 0 && pidx < txq->ift_size);
3587 #ifdef PKT_DEBUG
3588 	print_pkt(&pi);
3589 #endif
3590 	if ((err = ctx->isc_txd_encap(ctx->ifc_softc, &pi)) == 0) {
3591 		bus_dmamap_sync(buf_tag, map, BUS_DMASYNC_PREWRITE);
3592 		DBG_COUNTER_INC(tx_encap);
3593 		MPASS(pi.ipi_new_pidx < txq->ift_size);
3594 
3595 		ndesc = pi.ipi_new_pidx - pi.ipi_pidx;
3596 		if (pi.ipi_new_pidx < pi.ipi_pidx) {
3597 			ndesc += txq->ift_size;
3598 			txq->ift_gen = 1;
3599 		}
3600 		/*
3601 		 * drivers can need as many as
3602 		 * two sentinels
3603 		 */
3604 		MPASS(ndesc <= pi.ipi_nsegs + 2);
3605 		MPASS(pi.ipi_new_pidx != pidx);
3606 		MPASS(ndesc > 0);
3607 		txq->ift_in_use += ndesc;
3608 		txq->ift_db_pending += ndesc;
3609 
3610 		/*
3611 		 * We update the last software descriptor again here because there may
3612 		 * be a sentinel and/or there may be more mbufs than segments
3613 		 */
3614 		txq->ift_pidx = pi.ipi_new_pidx;
3615 		txq->ift_npending += pi.ipi_ndescs;
3616 	} else {
3617 		*m_headp = m_head = iflib_remove_mbuf(txq);
3618 		if (err == EFBIG) {
3619 			txq->ift_txd_encap_efbig++;
3620 			if (remap < 2) {
3621 				remap = 1;
3622 				goto defrag;
3623 			}
3624 		}
3625 		goto defrag_failed;
3626 	}
3627 	/*
3628 	 * err can't possibly be non-zero here, so we don't neet to test it
3629 	 * to see if we need to DBG_COUNTER_INC(encap_txd_encap_fail).
3630 	 */
3631 	return (err);
3632 
3633 defrag_failed:
3634 	txq->ift_mbuf_defrag_failed++;
3635 	txq->ift_map_failed++;
3636 	m_freem(*m_headp);
3637 	DBG_COUNTER_INC(tx_frees);
3638 	*m_headp = NULL;
3639 	DBG_COUNTER_INC(encap_txd_encap_fail);
3640 	return (ENOMEM);
3641 }
3642 
3643 static void
3644 iflib_tx_desc_free(iflib_txq_t txq, int n)
3645 {
3646 	uint32_t qsize, cidx, mask, gen;
3647 	struct mbuf *m, **ifsd_m;
3648 	bool do_prefetch;
3649 
3650 	cidx = txq->ift_cidx;
3651 	gen = txq->ift_gen;
3652 	qsize = txq->ift_size;
3653 	mask = qsize-1;
3654 	ifsd_m = txq->ift_sds.ifsd_m;
3655 	do_prefetch = (txq->ift_ctx->ifc_flags & IFC_PREFETCH);
3656 
3657 	while (n-- > 0) {
3658 		if (do_prefetch) {
3659 			prefetch(ifsd_m[(cidx + 3) & mask]);
3660 			prefetch(ifsd_m[(cidx + 4) & mask]);
3661 		}
3662 		if ((m = ifsd_m[cidx]) != NULL) {
3663 			prefetch(&ifsd_m[(cidx + CACHE_PTR_INCREMENT) & mask]);
3664 			if (m->m_pkthdr.csum_flags & CSUM_TSO) {
3665 				bus_dmamap_sync(txq->ift_tso_buf_tag,
3666 				    txq->ift_sds.ifsd_tso_map[cidx],
3667 				    BUS_DMASYNC_POSTWRITE);
3668 				bus_dmamap_unload(txq->ift_tso_buf_tag,
3669 				    txq->ift_sds.ifsd_tso_map[cidx]);
3670 			} else {
3671 				bus_dmamap_sync(txq->ift_buf_tag,
3672 				    txq->ift_sds.ifsd_map[cidx],
3673 				    BUS_DMASYNC_POSTWRITE);
3674 				bus_dmamap_unload(txq->ift_buf_tag,
3675 				    txq->ift_sds.ifsd_map[cidx]);
3676 			}
3677 			/* XXX we don't support any drivers that batch packets yet */
3678 			MPASS(m->m_nextpkt == NULL);
3679 			m_freem(m);
3680 			ifsd_m[cidx] = NULL;
3681 #if MEMORY_LOGGING
3682 			txq->ift_dequeued++;
3683 #endif
3684 			DBG_COUNTER_INC(tx_frees);
3685 		}
3686 		if (__predict_false(++cidx == qsize)) {
3687 			cidx = 0;
3688 			gen = 0;
3689 		}
3690 	}
3691 	txq->ift_cidx = cidx;
3692 	txq->ift_gen = gen;
3693 }
3694 
3695 static __inline int
3696 iflib_completed_tx_reclaim(iflib_txq_t txq, int thresh)
3697 {
3698 	int reclaim;
3699 	if_ctx_t ctx = txq->ift_ctx;
3700 
3701 	KASSERT(thresh >= 0, ("invalid threshold to reclaim"));
3702 	MPASS(thresh /*+ MAX_TX_DESC(txq->ift_ctx) */ < txq->ift_size);
3703 
3704 	/*
3705 	 * Need a rate-limiting check so that this isn't called every time
3706 	 */
3707 	iflib_tx_credits_update(ctx, txq);
3708 	reclaim = DESC_RECLAIMABLE(txq);
3709 
3710 	if (reclaim <= thresh /* + MAX_TX_DESC(txq->ift_ctx) */) {
3711 #ifdef INVARIANTS
3712 		if (iflib_verbose_debug) {
3713 			printf("%s processed=%ju cleaned=%ju tx_nsegments=%d reclaim=%d thresh=%d\n", __FUNCTION__,
3714 			       txq->ift_processed, txq->ift_cleaned, txq->ift_ctx->ifc_softc_ctx.isc_tx_nsegments,
3715 			       reclaim, thresh);
3716 		}
3717 #endif
3718 		return (0);
3719 	}
3720 	iflib_tx_desc_free(txq, reclaim);
3721 	txq->ift_cleaned += reclaim;
3722 	txq->ift_in_use -= reclaim;
3723 
3724 	return (reclaim);
3725 }
3726 
3727 static struct mbuf **
3728 _ring_peek_one(struct ifmp_ring *r, int cidx, int offset, int remaining)
3729 {
3730 	int next, size;
3731 	struct mbuf **items;
3732 
3733 	size = r->size;
3734 	next = (cidx + CACHE_PTR_INCREMENT) & (size-1);
3735 	items = __DEVOLATILE(struct mbuf **, &r->items[0]);
3736 
3737 	prefetch(items[(cidx + offset) & (size-1)]);
3738 	if (remaining > 1) {
3739 		prefetch2cachelines(&items[next]);
3740 		prefetch2cachelines(items[(cidx + offset + 1) & (size-1)]);
3741 		prefetch2cachelines(items[(cidx + offset + 2) & (size-1)]);
3742 		prefetch2cachelines(items[(cidx + offset + 3) & (size-1)]);
3743 	}
3744 	return (__DEVOLATILE(struct mbuf **, &r->items[(cidx + offset) & (size-1)]));
3745 }
3746 
3747 static void
3748 iflib_txq_check_drain(iflib_txq_t txq, int budget)
3749 {
3750 
3751 	ifmp_ring_check_drainage(txq->ift_br, budget);
3752 }
3753 
3754 static uint32_t
3755 iflib_txq_can_drain(struct ifmp_ring *r)
3756 {
3757 	iflib_txq_t txq = r->cookie;
3758 	if_ctx_t ctx = txq->ift_ctx;
3759 
3760 	if (TXQ_AVAIL(txq) > MAX_TX_DESC(ctx) + 2)
3761 		return (1);
3762 	bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
3763 	    BUS_DMASYNC_POSTREAD);
3764 	return (ctx->isc_txd_credits_update(ctx->ifc_softc, txq->ift_id,
3765 	    false));
3766 }
3767 
3768 static uint32_t
3769 iflib_txq_drain(struct ifmp_ring *r, uint32_t cidx, uint32_t pidx)
3770 {
3771 	iflib_txq_t txq = r->cookie;
3772 	if_ctx_t ctx = txq->ift_ctx;
3773 	if_t ifp = ctx->ifc_ifp;
3774 	struct mbuf *m, **mp;
3775 	int avail, bytes_sent, skipped, count, err, i;
3776 	int mcast_sent, pkt_sent, reclaimed;
3777 	bool do_prefetch, rang, ring;
3778 
3779 	if (__predict_false(!(if_getdrvflags(ifp) & IFF_DRV_RUNNING) ||
3780 			    !LINK_ACTIVE(ctx))) {
3781 		DBG_COUNTER_INC(txq_drain_notready);
3782 		return (0);
3783 	}
3784 	reclaimed = iflib_completed_tx_reclaim(txq, RECLAIM_THRESH(ctx));
3785 	rang = iflib_txd_db_check(txq, reclaimed && txq->ift_db_pending);
3786 	avail = IDXDIFF(pidx, cidx, r->size);
3787 
3788 	if (__predict_false(ctx->ifc_flags & IFC_QFLUSH)) {
3789 		/*
3790 		 * The driver is unloading so we need to free all pending packets.
3791 		 */
3792 		DBG_COUNTER_INC(txq_drain_flushing);
3793 		for (i = 0; i < avail; i++) {
3794 			if (__predict_true(r->items[(cidx + i) & (r->size-1)] != (void *)txq))
3795 				m_freem(r->items[(cidx + i) & (r->size-1)]);
3796 			r->items[(cidx + i) & (r->size-1)] = NULL;
3797 		}
3798 		return (avail);
3799 	}
3800 
3801 	if (__predict_false(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_OACTIVE)) {
3802 		txq->ift_qstatus = IFLIB_QUEUE_IDLE;
3803 		CALLOUT_LOCK(txq);
3804 		callout_stop(&txq->ift_timer);
3805 		CALLOUT_UNLOCK(txq);
3806 		DBG_COUNTER_INC(txq_drain_oactive);
3807 		return (0);
3808 	}
3809 
3810 	/*
3811 	 * If we've reclaimed any packets this queue cannot be hung.
3812 	 */
3813 	if (reclaimed)
3814 		txq->ift_qstatus = IFLIB_QUEUE_IDLE;
3815 	skipped = mcast_sent = bytes_sent = pkt_sent = 0;
3816 	count = MIN(avail, TX_BATCH_SIZE);
3817 #ifdef INVARIANTS
3818 	if (iflib_verbose_debug)
3819 		printf("%s avail=%d ifc_flags=%x txq_avail=%d ", __FUNCTION__,
3820 		       avail, ctx->ifc_flags, TXQ_AVAIL(txq));
3821 #endif
3822 	do_prefetch = (ctx->ifc_flags & IFC_PREFETCH);
3823 	err = 0;
3824 	for (i = 0; i < count && TXQ_AVAIL(txq) >= MAX_TX_DESC(ctx) + 2; i++) {
3825 		int rem = do_prefetch ? count - i : 0;
3826 
3827 		mp = _ring_peek_one(r, cidx, i, rem);
3828 		MPASS(mp != NULL && *mp != NULL);
3829 
3830 		/*
3831 		 * Completion interrupts will use the address of the txq
3832 		 * as a sentinel to enqueue _something_ in order to acquire
3833 		 * the lock on the mp_ring (there's no direct lock call).
3834 		 * We obviously whave to check for these sentinel cases
3835 		 * and skip them.
3836 		 */
3837 		if (__predict_false(*mp == (struct mbuf *)txq)) {
3838 			skipped++;
3839 			continue;
3840 		}
3841 		err = iflib_encap(txq, mp);
3842 		if (__predict_false(err)) {
3843 			/* no room - bail out */
3844 			if (err == ENOBUFS)
3845 				break;
3846 			skipped++;
3847 			/* we can't send this packet - skip it */
3848 			continue;
3849 		}
3850 		pkt_sent++;
3851 		m = *mp;
3852 		DBG_COUNTER_INC(tx_sent);
3853 		bytes_sent += m->m_pkthdr.len;
3854 		mcast_sent += !!(m->m_flags & M_MCAST);
3855 
3856 		if (__predict_false(!(ifp->if_drv_flags & IFF_DRV_RUNNING)))
3857 			break;
3858 		ETHER_BPF_MTAP(ifp, m);
3859 		rang = iflib_txd_db_check(txq, false);
3860 	}
3861 
3862 	/* deliberate use of bitwise or to avoid gratuitous short-circuit */
3863 	ring = rang ? false  : (iflib_min_tx_latency | err);
3864 	iflib_txd_db_check(txq, ring);
3865 	if_inc_counter(ifp, IFCOUNTER_OBYTES, bytes_sent);
3866 	if_inc_counter(ifp, IFCOUNTER_OPACKETS, pkt_sent);
3867 	if (mcast_sent)
3868 		if_inc_counter(ifp, IFCOUNTER_OMCASTS, mcast_sent);
3869 #ifdef INVARIANTS
3870 	if (iflib_verbose_debug)
3871 		printf("consumed=%d\n", skipped + pkt_sent);
3872 #endif
3873 	return (skipped + pkt_sent);
3874 }
3875 
3876 static uint32_t
3877 iflib_txq_drain_always(struct ifmp_ring *r)
3878 {
3879 	return (1);
3880 }
3881 
3882 static uint32_t
3883 iflib_txq_drain_free(struct ifmp_ring *r, uint32_t cidx, uint32_t pidx)
3884 {
3885 	int i, avail;
3886 	struct mbuf **mp;
3887 	iflib_txq_t txq;
3888 
3889 	txq = r->cookie;
3890 
3891 	txq->ift_qstatus = IFLIB_QUEUE_IDLE;
3892 	CALLOUT_LOCK(txq);
3893 	callout_stop(&txq->ift_timer);
3894 	CALLOUT_UNLOCK(txq);
3895 
3896 	avail = IDXDIFF(pidx, cidx, r->size);
3897 	for (i = 0; i < avail; i++) {
3898 		mp = _ring_peek_one(r, cidx, i, avail - i);
3899 		if (__predict_false(*mp == (struct mbuf *)txq))
3900 			continue;
3901 		m_freem(*mp);
3902 		DBG_COUNTER_INC(tx_frees);
3903 	}
3904 	MPASS(ifmp_ring_is_stalled(r) == 0);
3905 	return (avail);
3906 }
3907 
3908 static void
3909 iflib_ifmp_purge(iflib_txq_t txq)
3910 {
3911 	struct ifmp_ring *r;
3912 
3913 	r = txq->ift_br;
3914 	r->drain = iflib_txq_drain_free;
3915 	r->can_drain = iflib_txq_drain_always;
3916 
3917 	ifmp_ring_check_drainage(r, r->size);
3918 
3919 	r->drain = iflib_txq_drain;
3920 	r->can_drain = iflib_txq_can_drain;
3921 }
3922 
3923 static void
3924 _task_fn_tx(void *context)
3925 {
3926 	iflib_txq_t txq = context;
3927 	if_ctx_t ctx = txq->ift_ctx;
3928 	if_t ifp = ctx->ifc_ifp;
3929 	int abdicate = ctx->ifc_sysctl_tx_abdicate;
3930 
3931 #ifdef IFLIB_DIAGNOSTICS
3932 	txq->ift_cpu_exec_count[curcpu]++;
3933 #endif
3934 	if (!(if_getdrvflags(ifp) & IFF_DRV_RUNNING))
3935 		return;
3936 #ifdef DEV_NETMAP
3937 	if ((if_getcapenable(ifp) & IFCAP_NETMAP) &&
3938 	    netmap_tx_irq(ifp, txq->ift_id))
3939 		goto skip_ifmp;
3940 #endif
3941 #ifdef ALTQ
3942 	if (ALTQ_IS_ENABLED(&ifp->if_snd))
3943 		iflib_altq_if_start(ifp);
3944 #endif
3945 	if (txq->ift_db_pending)
3946 		ifmp_ring_enqueue(txq->ift_br, (void **)&txq, 1, TX_BATCH_SIZE, abdicate);
3947 	else if (!abdicate)
3948 		ifmp_ring_check_drainage(txq->ift_br, TX_BATCH_SIZE);
3949 	/*
3950 	 * When abdicating, we always need to check drainage, not just when we don't enqueue
3951 	 */
3952 	if (abdicate)
3953 		ifmp_ring_check_drainage(txq->ift_br, TX_BATCH_SIZE);
3954 #ifdef DEV_NETMAP
3955 skip_ifmp:
3956 #endif
3957 	if (ctx->ifc_flags & IFC_LEGACY)
3958 		IFDI_INTR_ENABLE(ctx);
3959 	else
3960 		IFDI_TX_QUEUE_INTR_ENABLE(ctx, txq->ift_id);
3961 }
3962 
3963 static void
3964 _task_fn_rx(void *context)
3965 {
3966 	iflib_rxq_t rxq = context;
3967 	if_ctx_t ctx = rxq->ifr_ctx;
3968 	uint8_t more;
3969 	uint16_t budget;
3970 #ifdef DEV_NETMAP
3971 	u_int work = 0;
3972 	int nmirq;
3973 #endif
3974 
3975 #ifdef IFLIB_DIAGNOSTICS
3976 	rxq->ifr_cpu_exec_count[curcpu]++;
3977 #endif
3978 	DBG_COUNTER_INC(task_fn_rxs);
3979 	if (__predict_false(!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING)))
3980 		return;
3981 #ifdef DEV_NETMAP
3982 	nmirq = netmap_rx_irq(ctx->ifc_ifp, rxq->ifr_id, &work);
3983 	if (nmirq != NM_IRQ_PASS) {
3984 		more = (nmirq == NM_IRQ_RESCHED) ? IFLIB_RXEOF_MORE : 0;
3985 		goto skip_rxeof;
3986 	}
3987 #endif
3988 	budget = ctx->ifc_sysctl_rx_budget;
3989 	if (budget == 0)
3990 		budget = 16;	/* XXX */
3991 	more = iflib_rxeof(rxq, budget);
3992 #ifdef DEV_NETMAP
3993 skip_rxeof:
3994 #endif
3995 	if ((more & IFLIB_RXEOF_MORE) == 0) {
3996 		if (ctx->ifc_flags & IFC_LEGACY)
3997 			IFDI_INTR_ENABLE(ctx);
3998 		else
3999 			IFDI_RX_QUEUE_INTR_ENABLE(ctx, rxq->ifr_id);
4000 		DBG_COUNTER_INC(rx_intr_enables);
4001 	}
4002 	if (__predict_false(!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING)))
4003 		return;
4004 
4005 	if (more & IFLIB_RXEOF_MORE)
4006 		GROUPTASK_ENQUEUE(&rxq->ifr_task);
4007 	else if (more & IFLIB_RXEOF_EMPTY)
4008 		callout_reset_curcpu(&rxq->ifr_watchdog, 1, &_task_fn_rx_watchdog, rxq);
4009 }
4010 
4011 static void
4012 _task_fn_admin(void *context)
4013 {
4014 	if_ctx_t ctx = context;
4015 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
4016 	iflib_txq_t txq;
4017 	int i;
4018 	bool oactive, running, do_reset, do_watchdog, in_detach;
4019 
4020 	STATE_LOCK(ctx);
4021 	running = (if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING);
4022 	oactive = (if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_OACTIVE);
4023 	do_reset = (ctx->ifc_flags & IFC_DO_RESET);
4024 	do_watchdog = (ctx->ifc_flags & IFC_DO_WATCHDOG);
4025 	in_detach = (ctx->ifc_flags & IFC_IN_DETACH);
4026 	ctx->ifc_flags &= ~(IFC_DO_RESET|IFC_DO_WATCHDOG);
4027 	STATE_UNLOCK(ctx);
4028 
4029 	if ((!running && !oactive) && !(ctx->ifc_sctx->isc_flags & IFLIB_ADMIN_ALWAYS_RUN))
4030 		return;
4031 	if (in_detach)
4032 		return;
4033 
4034 	CTX_LOCK(ctx);
4035 	for (txq = ctx->ifc_txqs, i = 0; i < sctx->isc_ntxqsets; i++, txq++) {
4036 		CALLOUT_LOCK(txq);
4037 		callout_stop(&txq->ift_timer);
4038 		CALLOUT_UNLOCK(txq);
4039 	}
4040 	if (ctx->ifc_sctx->isc_flags & IFLIB_HAS_ADMINCQ)
4041 		IFDI_ADMIN_COMPLETION_HANDLE(ctx);
4042 	if (do_watchdog) {
4043 		ctx->ifc_watchdog_events++;
4044 		IFDI_WATCHDOG_RESET(ctx);
4045 	}
4046 	IFDI_UPDATE_ADMIN_STATUS(ctx);
4047 	for (txq = ctx->ifc_txqs, i = 0; i < sctx->isc_ntxqsets; i++, txq++) {
4048 		callout_reset_on(&txq->ift_timer, iflib_timer_default, iflib_timer, txq,
4049 		    txq->ift_timer.c_cpu);
4050 	}
4051 	IFDI_LINK_INTR_ENABLE(ctx);
4052 	if (do_reset)
4053 		iflib_if_init_locked(ctx);
4054 	CTX_UNLOCK(ctx);
4055 
4056 	if (LINK_ACTIVE(ctx) == 0)
4057 		return;
4058 	for (txq = ctx->ifc_txqs, i = 0; i < sctx->isc_ntxqsets; i++, txq++)
4059 		iflib_txq_check_drain(txq, IFLIB_RESTART_BUDGET);
4060 }
4061 
4062 static void
4063 _task_fn_iov(void *context)
4064 {
4065 	if_ctx_t ctx = context;
4066 
4067 	if (!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING) &&
4068 	    !(ctx->ifc_sctx->isc_flags & IFLIB_ADMIN_ALWAYS_RUN))
4069 		return;
4070 
4071 	CTX_LOCK(ctx);
4072 	IFDI_VFLR_HANDLE(ctx);
4073 	CTX_UNLOCK(ctx);
4074 }
4075 
4076 static int
4077 iflib_sysctl_int_delay(SYSCTL_HANDLER_ARGS)
4078 {
4079 	int err;
4080 	if_int_delay_info_t info;
4081 	if_ctx_t ctx;
4082 
4083 	info = (if_int_delay_info_t)arg1;
4084 	ctx = info->iidi_ctx;
4085 	info->iidi_req = req;
4086 	info->iidi_oidp = oidp;
4087 	CTX_LOCK(ctx);
4088 	err = IFDI_SYSCTL_INT_DELAY(ctx, info);
4089 	CTX_UNLOCK(ctx);
4090 	return (err);
4091 }
4092 
4093 /*********************************************************************
4094  *
4095  *  IFNET FUNCTIONS
4096  *
4097  **********************************************************************/
4098 
4099 static void
4100 iflib_if_init_locked(if_ctx_t ctx)
4101 {
4102 	iflib_stop(ctx);
4103 	iflib_init_locked(ctx);
4104 }
4105 
4106 static void
4107 iflib_if_init(void *arg)
4108 {
4109 	if_ctx_t ctx = arg;
4110 
4111 	CTX_LOCK(ctx);
4112 	iflib_if_init_locked(ctx);
4113 	CTX_UNLOCK(ctx);
4114 }
4115 
4116 static int
4117 iflib_if_transmit(if_t ifp, struct mbuf *m)
4118 {
4119 	if_ctx_t	ctx = if_getsoftc(ifp);
4120 
4121 	iflib_txq_t txq;
4122 	int err, qidx;
4123 	int abdicate = ctx->ifc_sysctl_tx_abdicate;
4124 
4125 	if (__predict_false((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 || !LINK_ACTIVE(ctx))) {
4126 		DBG_COUNTER_INC(tx_frees);
4127 		m_freem(m);
4128 		return (ENETDOWN);
4129 	}
4130 
4131 	MPASS(m->m_nextpkt == NULL);
4132 	/* ALTQ-enabled interfaces always use queue 0. */
4133 	qidx = 0;
4134 	if ((NTXQSETS(ctx) > 1) && M_HASHTYPE_GET(m) && !ALTQ_IS_ENABLED(&ifp->if_snd))
4135 		qidx = QIDX(ctx, m);
4136 	/*
4137 	 * XXX calculate buf_ring based on flowid (divvy up bits?)
4138 	 */
4139 	txq = &ctx->ifc_txqs[qidx];
4140 
4141 #ifdef DRIVER_BACKPRESSURE
4142 	if (txq->ift_closed) {
4143 		while (m != NULL) {
4144 			next = m->m_nextpkt;
4145 			m->m_nextpkt = NULL;
4146 			m_freem(m);
4147 			DBG_COUNTER_INC(tx_frees);
4148 			m = next;
4149 		}
4150 		return (ENOBUFS);
4151 	}
4152 #endif
4153 #ifdef notyet
4154 	qidx = count = 0;
4155 	mp = marr;
4156 	next = m;
4157 	do {
4158 		count++;
4159 		next = next->m_nextpkt;
4160 	} while (next != NULL);
4161 
4162 	if (count > nitems(marr))
4163 		if ((mp = malloc(count*sizeof(struct mbuf *), M_IFLIB, M_NOWAIT)) == NULL) {
4164 			/* XXX check nextpkt */
4165 			m_freem(m);
4166 			/* XXX simplify for now */
4167 			DBG_COUNTER_INC(tx_frees);
4168 			return (ENOBUFS);
4169 		}
4170 	for (next = m, i = 0; next != NULL; i++) {
4171 		mp[i] = next;
4172 		next = next->m_nextpkt;
4173 		mp[i]->m_nextpkt = NULL;
4174 	}
4175 #endif
4176 	DBG_COUNTER_INC(tx_seen);
4177 	err = ifmp_ring_enqueue(txq->ift_br, (void **)&m, 1, TX_BATCH_SIZE, abdicate);
4178 
4179 	if (abdicate)
4180 		GROUPTASK_ENQUEUE(&txq->ift_task);
4181  	if (err) {
4182 		if (!abdicate)
4183 			GROUPTASK_ENQUEUE(&txq->ift_task);
4184 		/* support forthcoming later */
4185 #ifdef DRIVER_BACKPRESSURE
4186 		txq->ift_closed = TRUE;
4187 #endif
4188 		ifmp_ring_check_drainage(txq->ift_br, TX_BATCH_SIZE);
4189 		m_freem(m);
4190 		DBG_COUNTER_INC(tx_frees);
4191 	}
4192 
4193 	return (err);
4194 }
4195 
4196 #ifdef ALTQ
4197 /*
4198  * The overall approach to integrating iflib with ALTQ is to continue to use
4199  * the iflib mp_ring machinery between the ALTQ queue(s) and the hardware
4200  * ring.  Technically, when using ALTQ, queueing to an intermediate mp_ring
4201  * is redundant/unnecessary, but doing so minimizes the amount of
4202  * ALTQ-specific code required in iflib.  It is assumed that the overhead of
4203  * redundantly queueing to an intermediate mp_ring is swamped by the
4204  * performance limitations inherent in using ALTQ.
4205  *
4206  * When ALTQ support is compiled in, all iflib drivers will use a transmit
4207  * routine, iflib_altq_if_transmit(), that checks if ALTQ is enabled for the
4208  * given interface.  If ALTQ is enabled for an interface, then all
4209  * transmitted packets for that interface will be submitted to the ALTQ
4210  * subsystem via IFQ_ENQUEUE().  We don't use the legacy if_transmit()
4211  * implementation because it uses IFQ_HANDOFF(), which will duplicatively
4212  * update stats that the iflib machinery handles, and which is sensitve to
4213  * the disused IFF_DRV_OACTIVE flag.  Additionally, iflib_altq_if_start()
4214  * will be installed as the start routine for use by ALTQ facilities that
4215  * need to trigger queue drains on a scheduled basis.
4216  *
4217  */
4218 static void
4219 iflib_altq_if_start(if_t ifp)
4220 {
4221 	struct ifaltq *ifq = &ifp->if_snd;
4222 	struct mbuf *m;
4223 
4224 	IFQ_LOCK(ifq);
4225 	IFQ_DEQUEUE_NOLOCK(ifq, m);
4226 	while (m != NULL) {
4227 		iflib_if_transmit(ifp, m);
4228 		IFQ_DEQUEUE_NOLOCK(ifq, m);
4229 	}
4230 	IFQ_UNLOCK(ifq);
4231 }
4232 
4233 static int
4234 iflib_altq_if_transmit(if_t ifp, struct mbuf *m)
4235 {
4236 	int err;
4237 
4238 	if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
4239 		IFQ_ENQUEUE(&ifp->if_snd, m, err);
4240 		if (err == 0)
4241 			iflib_altq_if_start(ifp);
4242 	} else
4243 		err = iflib_if_transmit(ifp, m);
4244 
4245 	return (err);
4246 }
4247 #endif /* ALTQ */
4248 
4249 static void
4250 iflib_if_qflush(if_t ifp)
4251 {
4252 	if_ctx_t ctx = if_getsoftc(ifp);
4253 	iflib_txq_t txq = ctx->ifc_txqs;
4254 	int i;
4255 
4256 	STATE_LOCK(ctx);
4257 	ctx->ifc_flags |= IFC_QFLUSH;
4258 	STATE_UNLOCK(ctx);
4259 	for (i = 0; i < NTXQSETS(ctx); i++, txq++)
4260 		while (!(ifmp_ring_is_idle(txq->ift_br) || ifmp_ring_is_stalled(txq->ift_br)))
4261 			iflib_txq_check_drain(txq, 0);
4262 	STATE_LOCK(ctx);
4263 	ctx->ifc_flags &= ~IFC_QFLUSH;
4264 	STATE_UNLOCK(ctx);
4265 
4266 	/*
4267 	 * When ALTQ is enabled, this will also take care of purging the
4268 	 * ALTQ queue(s).
4269 	 */
4270 	if_qflush(ifp);
4271 }
4272 
4273 #define IFCAP_FLAGS (IFCAP_HWCSUM_IPV6 | IFCAP_HWCSUM | IFCAP_LRO | \
4274 		     IFCAP_TSO | IFCAP_VLAN_HWTAGGING | IFCAP_HWSTATS | \
4275 		     IFCAP_VLAN_MTU | IFCAP_VLAN_HWFILTER | \
4276 		     IFCAP_VLAN_HWTSO | IFCAP_VLAN_HWCSUM | IFCAP_MEXTPG)
4277 
4278 static int
4279 iflib_if_ioctl(if_t ifp, u_long command, caddr_t data)
4280 {
4281 	if_ctx_t ctx = if_getsoftc(ifp);
4282 	struct ifreq	*ifr = (struct ifreq *)data;
4283 #if defined(INET) || defined(INET6)
4284 	struct ifaddr	*ifa = (struct ifaddr *)data;
4285 #endif
4286 	bool		avoid_reset = false;
4287 	int		err = 0, reinit = 0, bits;
4288 
4289 	switch (command) {
4290 	case SIOCSIFADDR:
4291 #ifdef INET
4292 		if (ifa->ifa_addr->sa_family == AF_INET)
4293 			avoid_reset = true;
4294 #endif
4295 #ifdef INET6
4296 		if (ifa->ifa_addr->sa_family == AF_INET6)
4297 			avoid_reset = true;
4298 #endif
4299 		/*
4300 		** Calling init results in link renegotiation,
4301 		** so we avoid doing it when possible.
4302 		*/
4303 		if (avoid_reset) {
4304 			if_setflagbits(ifp, IFF_UP,0);
4305 			if (!(if_getdrvflags(ifp) & IFF_DRV_RUNNING))
4306 				reinit = 1;
4307 #ifdef INET
4308 			if (!(if_getflags(ifp) & IFF_NOARP))
4309 				arp_ifinit(ifp, ifa);
4310 #endif
4311 		} else
4312 			err = ether_ioctl(ifp, command, data);
4313 		break;
4314 	case SIOCSIFMTU:
4315 		CTX_LOCK(ctx);
4316 		if (ifr->ifr_mtu == if_getmtu(ifp)) {
4317 			CTX_UNLOCK(ctx);
4318 			break;
4319 		}
4320 		bits = if_getdrvflags(ifp);
4321 		/* stop the driver and free any clusters before proceeding */
4322 		iflib_stop(ctx);
4323 
4324 		if ((err = IFDI_MTU_SET(ctx, ifr->ifr_mtu)) == 0) {
4325 			STATE_LOCK(ctx);
4326 			if (ifr->ifr_mtu > ctx->ifc_max_fl_buf_size)
4327 				ctx->ifc_flags |= IFC_MULTISEG;
4328 			else
4329 				ctx->ifc_flags &= ~IFC_MULTISEG;
4330 			STATE_UNLOCK(ctx);
4331 			err = if_setmtu(ifp, ifr->ifr_mtu);
4332 		}
4333 		iflib_init_locked(ctx);
4334 		STATE_LOCK(ctx);
4335 		if_setdrvflags(ifp, bits);
4336 		STATE_UNLOCK(ctx);
4337 		CTX_UNLOCK(ctx);
4338 		break;
4339 	case SIOCSIFFLAGS:
4340 		CTX_LOCK(ctx);
4341 		if (if_getflags(ifp) & IFF_UP) {
4342 			if (if_getdrvflags(ifp) & IFF_DRV_RUNNING) {
4343 				if ((if_getflags(ifp) ^ ctx->ifc_if_flags) &
4344 				    (IFF_PROMISC | IFF_ALLMULTI)) {
4345 					CTX_UNLOCK(ctx);
4346 					err = IFDI_PROMISC_SET(ctx, if_getflags(ifp));
4347 					CTX_LOCK(ctx);
4348 				}
4349 			} else
4350 				reinit = 1;
4351 		} else if (if_getdrvflags(ifp) & IFF_DRV_RUNNING) {
4352 			iflib_stop(ctx);
4353 		}
4354 		ctx->ifc_if_flags = if_getflags(ifp);
4355 		CTX_UNLOCK(ctx);
4356 		break;
4357 	case SIOCADDMULTI:
4358 	case SIOCDELMULTI:
4359 		if (if_getdrvflags(ifp) & IFF_DRV_RUNNING) {
4360 			CTX_LOCK(ctx);
4361 			IFDI_INTR_DISABLE(ctx);
4362 			IFDI_MULTI_SET(ctx);
4363 			IFDI_INTR_ENABLE(ctx);
4364 			CTX_UNLOCK(ctx);
4365 		}
4366 		break;
4367 	case SIOCSIFMEDIA:
4368 		CTX_LOCK(ctx);
4369 		IFDI_MEDIA_SET(ctx);
4370 		CTX_UNLOCK(ctx);
4371 		/* FALLTHROUGH */
4372 	case SIOCGIFMEDIA:
4373 	case SIOCGIFXMEDIA:
4374 		err = ifmedia_ioctl(ifp, ifr, ctx->ifc_mediap, command);
4375 		break;
4376 	case SIOCGI2C:
4377 	{
4378 		struct ifi2creq i2c;
4379 
4380 		err = copyin(ifr_data_get_ptr(ifr), &i2c, sizeof(i2c));
4381 		if (err != 0)
4382 			break;
4383 		if (i2c.dev_addr != 0xA0 && i2c.dev_addr != 0xA2) {
4384 			err = EINVAL;
4385 			break;
4386 		}
4387 		if (i2c.len > sizeof(i2c.data)) {
4388 			err = EINVAL;
4389 			break;
4390 		}
4391 
4392 		if ((err = IFDI_I2C_REQ(ctx, &i2c)) == 0)
4393 			err = copyout(&i2c, ifr_data_get_ptr(ifr),
4394 			    sizeof(i2c));
4395 		break;
4396 	}
4397 	case SIOCSIFCAP:
4398 	{
4399 		int mask, setmask, oldmask;
4400 
4401 		oldmask = if_getcapenable(ifp);
4402 		mask = ifr->ifr_reqcap ^ oldmask;
4403 		mask &= ctx->ifc_softc_ctx.isc_capabilities | IFCAP_MEXTPG;
4404 		setmask = 0;
4405 #ifdef TCP_OFFLOAD
4406 		setmask |= mask & (IFCAP_TOE4|IFCAP_TOE6);
4407 #endif
4408 		setmask |= (mask & IFCAP_FLAGS);
4409 		setmask |= (mask & IFCAP_WOL);
4410 
4411 		/*
4412 		 * If any RX csum has changed, change all the ones that
4413 		 * are supported by the driver.
4414 		 */
4415 		if (setmask & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6)) {
4416 			setmask |= ctx->ifc_softc_ctx.isc_capabilities &
4417 			    (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6);
4418 		}
4419 
4420 		/*
4421 		 * want to ensure that traffic has stopped before we change any of the flags
4422 		 */
4423 		if (setmask) {
4424 			CTX_LOCK(ctx);
4425 			bits = if_getdrvflags(ifp);
4426 			if (bits & IFF_DRV_RUNNING && setmask & ~IFCAP_WOL)
4427 				iflib_stop(ctx);
4428 			STATE_LOCK(ctx);
4429 			if_togglecapenable(ifp, setmask);
4430 			STATE_UNLOCK(ctx);
4431 			if (bits & IFF_DRV_RUNNING && setmask & ~IFCAP_WOL)
4432 				iflib_init_locked(ctx);
4433 			STATE_LOCK(ctx);
4434 			if_setdrvflags(ifp, bits);
4435 			STATE_UNLOCK(ctx);
4436 			CTX_UNLOCK(ctx);
4437 		}
4438 		if_vlancap(ifp);
4439 		break;
4440 	}
4441 	case SIOCGPRIVATE_0:
4442 	case SIOCSDRVSPEC:
4443 	case SIOCGDRVSPEC:
4444 		CTX_LOCK(ctx);
4445 		err = IFDI_PRIV_IOCTL(ctx, command, data);
4446 		CTX_UNLOCK(ctx);
4447 		break;
4448 	default:
4449 		err = ether_ioctl(ifp, command, data);
4450 		break;
4451 	}
4452 	if (reinit)
4453 		iflib_if_init(ctx);
4454 	return (err);
4455 }
4456 
4457 static uint64_t
4458 iflib_if_get_counter(if_t ifp, ift_counter cnt)
4459 {
4460 	if_ctx_t ctx = if_getsoftc(ifp);
4461 
4462 	return (IFDI_GET_COUNTER(ctx, cnt));
4463 }
4464 
4465 /*********************************************************************
4466  *
4467  *  OTHER FUNCTIONS EXPORTED TO THE STACK
4468  *
4469  **********************************************************************/
4470 
4471 static void
4472 iflib_vlan_register(void *arg, if_t ifp, uint16_t vtag)
4473 {
4474 	if_ctx_t ctx = if_getsoftc(ifp);
4475 
4476 	if ((void *)ctx != arg)
4477 		return;
4478 
4479 	if ((vtag == 0) || (vtag > 4095))
4480 		return;
4481 
4482 	if (iflib_in_detach(ctx))
4483 		return;
4484 
4485 	CTX_LOCK(ctx);
4486 	/* Driver may need all untagged packets to be flushed */
4487 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4488 		iflib_stop(ctx);
4489 	IFDI_VLAN_REGISTER(ctx, vtag);
4490 	/* Re-init to load the changes, if required */
4491 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4492 		iflib_init_locked(ctx);
4493 	CTX_UNLOCK(ctx);
4494 }
4495 
4496 static void
4497 iflib_vlan_unregister(void *arg, if_t ifp, uint16_t vtag)
4498 {
4499 	if_ctx_t ctx = if_getsoftc(ifp);
4500 
4501 	if ((void *)ctx != arg)
4502 		return;
4503 
4504 	if ((vtag == 0) || (vtag > 4095))
4505 		return;
4506 
4507 	CTX_LOCK(ctx);
4508 	/* Driver may need all tagged packets to be flushed */
4509 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4510 		iflib_stop(ctx);
4511 	IFDI_VLAN_UNREGISTER(ctx, vtag);
4512 	/* Re-init to load the changes, if required */
4513 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4514 		iflib_init_locked(ctx);
4515 	CTX_UNLOCK(ctx);
4516 }
4517 
4518 static void
4519 iflib_led_func(void *arg, int onoff)
4520 {
4521 	if_ctx_t ctx = arg;
4522 
4523 	CTX_LOCK(ctx);
4524 	IFDI_LED_FUNC(ctx, onoff);
4525 	CTX_UNLOCK(ctx);
4526 }
4527 
4528 /*********************************************************************
4529  *
4530  *  BUS FUNCTION DEFINITIONS
4531  *
4532  **********************************************************************/
4533 
4534 int
4535 iflib_device_probe(device_t dev)
4536 {
4537 	const pci_vendor_info_t *ent;
4538 	if_shared_ctx_t sctx;
4539 	uint16_t pci_device_id, pci_rev_id, pci_subdevice_id, pci_subvendor_id;
4540 	uint16_t pci_vendor_id;
4541 
4542 	if ((sctx = DEVICE_REGISTER(dev)) == NULL || sctx->isc_magic != IFLIB_MAGIC)
4543 		return (ENOTSUP);
4544 
4545 	pci_vendor_id = pci_get_vendor(dev);
4546 	pci_device_id = pci_get_device(dev);
4547 	pci_subvendor_id = pci_get_subvendor(dev);
4548 	pci_subdevice_id = pci_get_subdevice(dev);
4549 	pci_rev_id = pci_get_revid(dev);
4550 	if (sctx->isc_parse_devinfo != NULL)
4551 		sctx->isc_parse_devinfo(&pci_device_id, &pci_subvendor_id, &pci_subdevice_id, &pci_rev_id);
4552 
4553 	ent = sctx->isc_vendor_info;
4554 	while (ent->pvi_vendor_id != 0) {
4555 		if (pci_vendor_id != ent->pvi_vendor_id) {
4556 			ent++;
4557 			continue;
4558 		}
4559 		if ((pci_device_id == ent->pvi_device_id) &&
4560 		    ((pci_subvendor_id == ent->pvi_subvendor_id) ||
4561 		     (ent->pvi_subvendor_id == 0)) &&
4562 		    ((pci_subdevice_id == ent->pvi_subdevice_id) ||
4563 		     (ent->pvi_subdevice_id == 0)) &&
4564 		    ((pci_rev_id == ent->pvi_rev_id) ||
4565 		     (ent->pvi_rev_id == 0))) {
4566 			device_set_desc_copy(dev, ent->pvi_name);
4567 			/* this needs to be changed to zero if the bus probing code
4568 			 * ever stops re-probing on best match because the sctx
4569 			 * may have its values over written by register calls
4570 			 * in subsequent probes
4571 			 */
4572 			return (BUS_PROBE_DEFAULT);
4573 		}
4574 		ent++;
4575 	}
4576 	return (ENXIO);
4577 }
4578 
4579 int
4580 iflib_device_probe_vendor(device_t dev)
4581 {
4582 	int probe;
4583 
4584 	probe = iflib_device_probe(dev);
4585 	if (probe == BUS_PROBE_DEFAULT)
4586 		return (BUS_PROBE_VENDOR);
4587 	else
4588 		return (probe);
4589 }
4590 
4591 static void
4592 iflib_reset_qvalues(if_ctx_t ctx)
4593 {
4594 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
4595 	if_shared_ctx_t sctx = ctx->ifc_sctx;
4596 	device_t dev = ctx->ifc_dev;
4597 	int i;
4598 
4599 	if (ctx->ifc_sysctl_ntxqs != 0)
4600 		scctx->isc_ntxqsets = ctx->ifc_sysctl_ntxqs;
4601 	if (ctx->ifc_sysctl_nrxqs != 0)
4602 		scctx->isc_nrxqsets = ctx->ifc_sysctl_nrxqs;
4603 
4604 	for (i = 0; i < sctx->isc_ntxqs; i++) {
4605 		if (ctx->ifc_sysctl_ntxds[i] != 0)
4606 			scctx->isc_ntxd[i] = ctx->ifc_sysctl_ntxds[i];
4607 		else
4608 			scctx->isc_ntxd[i] = sctx->isc_ntxd_default[i];
4609 	}
4610 
4611 	for (i = 0; i < sctx->isc_nrxqs; i++) {
4612 		if (ctx->ifc_sysctl_nrxds[i] != 0)
4613 			scctx->isc_nrxd[i] = ctx->ifc_sysctl_nrxds[i];
4614 		else
4615 			scctx->isc_nrxd[i] = sctx->isc_nrxd_default[i];
4616 	}
4617 
4618 	for (i = 0; i < sctx->isc_nrxqs; i++) {
4619 		if (scctx->isc_nrxd[i] < sctx->isc_nrxd_min[i]) {
4620 			device_printf(dev, "nrxd%d: %d less than nrxd_min %d - resetting to min\n",
4621 				      i, scctx->isc_nrxd[i], sctx->isc_nrxd_min[i]);
4622 			scctx->isc_nrxd[i] = sctx->isc_nrxd_min[i];
4623 		}
4624 		if (scctx->isc_nrxd[i] > sctx->isc_nrxd_max[i]) {
4625 			device_printf(dev, "nrxd%d: %d greater than nrxd_max %d - resetting to max\n",
4626 				      i, scctx->isc_nrxd[i], sctx->isc_nrxd_max[i]);
4627 			scctx->isc_nrxd[i] = sctx->isc_nrxd_max[i];
4628 		}
4629 		if (!powerof2(scctx->isc_nrxd[i])) {
4630 			device_printf(dev, "nrxd%d: %d is not a power of 2 - using default value of %d\n",
4631 				      i, scctx->isc_nrxd[i], sctx->isc_nrxd_default[i]);
4632 			scctx->isc_nrxd[i] = sctx->isc_nrxd_default[i];
4633 		}
4634 	}
4635 
4636 	for (i = 0; i < sctx->isc_ntxqs; i++) {
4637 		if (scctx->isc_ntxd[i] < sctx->isc_ntxd_min[i]) {
4638 			device_printf(dev, "ntxd%d: %d less than ntxd_min %d - resetting to min\n",
4639 				      i, scctx->isc_ntxd[i], sctx->isc_ntxd_min[i]);
4640 			scctx->isc_ntxd[i] = sctx->isc_ntxd_min[i];
4641 		}
4642 		if (scctx->isc_ntxd[i] > sctx->isc_ntxd_max[i]) {
4643 			device_printf(dev, "ntxd%d: %d greater than ntxd_max %d - resetting to max\n",
4644 				      i, scctx->isc_ntxd[i], sctx->isc_ntxd_max[i]);
4645 			scctx->isc_ntxd[i] = sctx->isc_ntxd_max[i];
4646 		}
4647 		if (!powerof2(scctx->isc_ntxd[i])) {
4648 			device_printf(dev, "ntxd%d: %d is not a power of 2 - using default value of %d\n",
4649 				      i, scctx->isc_ntxd[i], sctx->isc_ntxd_default[i]);
4650 			scctx->isc_ntxd[i] = sctx->isc_ntxd_default[i];
4651 		}
4652 	}
4653 }
4654 
4655 static void
4656 iflib_add_pfil(if_ctx_t ctx)
4657 {
4658 	struct pfil_head *pfil;
4659 	struct pfil_head_args pa;
4660 	iflib_rxq_t rxq;
4661 	int i;
4662 
4663 	pa.pa_version = PFIL_VERSION;
4664 	pa.pa_flags = PFIL_IN;
4665 	pa.pa_type = PFIL_TYPE_ETHERNET;
4666 	pa.pa_headname = ctx->ifc_ifp->if_xname;
4667 	pfil = pfil_head_register(&pa);
4668 
4669 	for (i = 0, rxq = ctx->ifc_rxqs; i < NRXQSETS(ctx); i++, rxq++) {
4670 		rxq->pfil = pfil;
4671 	}
4672 }
4673 
4674 static void
4675 iflib_rem_pfil(if_ctx_t ctx)
4676 {
4677 	struct pfil_head *pfil;
4678 	iflib_rxq_t rxq;
4679 	int i;
4680 
4681 	rxq = ctx->ifc_rxqs;
4682 	pfil = rxq->pfil;
4683 	for (i = 0; i < NRXQSETS(ctx); i++, rxq++) {
4684 		rxq->pfil = NULL;
4685 	}
4686 	pfil_head_unregister(pfil);
4687 }
4688 
4689 
4690 /*
4691  * Advance forward by n members of the cpuset ctx->ifc_cpus starting from
4692  * cpuid and wrapping as necessary.
4693  */
4694 static unsigned int
4695 cpuid_advance(if_ctx_t ctx, unsigned int cpuid, unsigned int n)
4696 {
4697 	unsigned int first_valid;
4698 	unsigned int last_valid;
4699 
4700 	/* cpuid should always be in the valid set */
4701 	MPASS(CPU_ISSET(cpuid, &ctx->ifc_cpus));
4702 
4703 	/* valid set should never be empty */
4704 	MPASS(!CPU_EMPTY(&ctx->ifc_cpus));
4705 
4706 	first_valid = CPU_FFS(&ctx->ifc_cpus) - 1;
4707 	last_valid = CPU_FLS(&ctx->ifc_cpus) - 1;
4708 	n = n % CPU_COUNT(&ctx->ifc_cpus);
4709 	while (n > 0) {
4710 		do {
4711 			cpuid++;
4712 			if (cpuid > last_valid)
4713 				cpuid = first_valid;
4714 		} while (!CPU_ISSET(cpuid, &ctx->ifc_cpus));
4715 		n--;
4716 	}
4717 
4718 	return (cpuid);
4719 }
4720 
4721 #if defined(SMP) && defined(SCHED_ULE)
4722 extern struct cpu_group *cpu_top;              /* CPU topology */
4723 
4724 static int
4725 find_child_with_core(int cpu, struct cpu_group *grp)
4726 {
4727 	int i;
4728 
4729 	if (grp->cg_children == 0)
4730 		return -1;
4731 
4732 	MPASS(grp->cg_child);
4733 	for (i = 0; i < grp->cg_children; i++) {
4734 		if (CPU_ISSET(cpu, &grp->cg_child[i].cg_mask))
4735 			return i;
4736 	}
4737 
4738 	return -1;
4739 }
4740 
4741 
4742 /*
4743  * Find an L2 neighbor of the given CPU or return -1 if none found.  This
4744  * does not distinguish among multiple L2 neighbors if the given CPU has
4745  * more than one (it will always return the same result in that case).
4746  */
4747 static int
4748 find_l2_neighbor(int cpu)
4749 {
4750 	struct cpu_group *grp;
4751 	int i;
4752 
4753 	grp = cpu_top;
4754 	if (grp == NULL)
4755 		return -1;
4756 
4757 	/*
4758 	 * Find the smallest CPU group that contains the given core.
4759 	 */
4760 	i = 0;
4761 	while ((i = find_child_with_core(cpu, grp)) != -1) {
4762 		/*
4763 		 * If the smallest group containing the given CPU has less
4764 		 * than two members, we conclude the given CPU has no
4765 		 * L2 neighbor.
4766 		 */
4767 		if (grp->cg_child[i].cg_count <= 1)
4768 			return (-1);
4769 		grp = &grp->cg_child[i];
4770 	}
4771 
4772 	/* Must share L2. */
4773 	if (grp->cg_level > CG_SHARE_L2 || grp->cg_level == CG_SHARE_NONE)
4774 		return -1;
4775 
4776 	/*
4777 	 * Select the first member of the set that isn't the reference
4778 	 * CPU, which at this point is guaranteed to exist.
4779 	 */
4780 	for (i = 0; i < CPU_SETSIZE; i++) {
4781 		if (CPU_ISSET(i, &grp->cg_mask) && i != cpu)
4782 			return (i);
4783 	}
4784 
4785 	/* Should never be reached */
4786 	return (-1);
4787 }
4788 
4789 #else
4790 static int
4791 find_l2_neighbor(int cpu)
4792 {
4793 
4794 	return (-1);
4795 }
4796 #endif
4797 
4798 /*
4799  * CPU mapping behaviors
4800  * ---------------------
4801  * 'separate txrx' refers to the separate_txrx sysctl
4802  * 'use logical' refers to the use_logical_cores sysctl
4803  * 'INTR CPUS' indicates whether bus_get_cpus(INTR_CPUS) succeeded
4804  *
4805  *  separate     use     INTR
4806  *    txrx     logical   CPUS   result
4807  * ---------- --------- ------ ------------------------------------------------
4808  *     -          -       X     RX and TX queues mapped to consecutive physical
4809  *                              cores with RX/TX pairs on same core and excess
4810  *                              of either following
4811  *     -          X       X     RX and TX queues mapped to consecutive cores
4812  *                              of any type with RX/TX pairs on same core and
4813  *                              excess of either following
4814  *     X          -       X     RX and TX queues mapped to consecutive physical
4815  *                              cores; all RX then all TX
4816  *     X          X       X     RX queues mapped to consecutive physical cores
4817  *                              first, then TX queues mapped to L2 neighbor of
4818  *                              the corresponding RX queue if one exists,
4819  *                              otherwise to consecutive physical cores
4820  *     -         n/a      -     RX and TX queues mapped to consecutive cores of
4821  *                              any type with RX/TX pairs on same core and excess
4822  *                              of either following
4823  *     X         n/a      -     RX and TX queues mapped to consecutive cores of
4824  *                              any type; all RX then all TX
4825  */
4826 static unsigned int
4827 get_cpuid_for_queue(if_ctx_t ctx, unsigned int base_cpuid, unsigned int qid,
4828     bool is_tx)
4829 {
4830 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
4831 	unsigned int core_index;
4832 
4833 	if (ctx->ifc_sysctl_separate_txrx) {
4834 		/*
4835 		 * When using separate CPUs for TX and RX, the assignment
4836 		 * will always be of a consecutive CPU out of the set of
4837 		 * context CPUs, except for the specific case where the
4838 		 * context CPUs are phsyical cores, the use of logical cores
4839 		 * has been enabled, the assignment is for TX, the TX qid
4840 		 * corresponds to an RX qid, and the CPU assigned to the
4841 		 * corresponding RX queue has an L2 neighbor.
4842 		 */
4843 		if (ctx->ifc_sysctl_use_logical_cores &&
4844 		    ctx->ifc_cpus_are_physical_cores &&
4845 		    is_tx && qid < scctx->isc_nrxqsets) {
4846 			int l2_neighbor;
4847 			unsigned int rx_cpuid;
4848 
4849 			rx_cpuid = cpuid_advance(ctx, base_cpuid, qid);
4850 			l2_neighbor = find_l2_neighbor(rx_cpuid);
4851 			if (l2_neighbor != -1) {
4852 				return (l2_neighbor);
4853 			}
4854 			/*
4855 			 * ... else fall through to the normal
4856 			 * consecutive-after-RX assignment scheme.
4857 			 *
4858 			 * Note that we are assuming that all RX queue CPUs
4859 			 * have an L2 neighbor, or all do not.  If a mixed
4860 			 * scenario is possible, we will have to keep track
4861 			 * separately of how many queues prior to this one
4862 			 * were not able to be assigned to an L2 neighbor.
4863 			 */
4864 		}
4865 		if (is_tx)
4866 			core_index = scctx->isc_nrxqsets + qid;
4867 		else
4868 			core_index = qid;
4869 	} else {
4870 		core_index = qid;
4871 	}
4872 
4873 	return (cpuid_advance(ctx, base_cpuid, core_index));
4874 }
4875 
4876 static uint16_t
4877 get_ctx_core_offset(if_ctx_t ctx)
4878 {
4879 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
4880 	struct cpu_offset *op;
4881 	cpuset_t assigned_cpus;
4882 	unsigned int cores_consumed;
4883 	unsigned int base_cpuid = ctx->ifc_sysctl_core_offset;
4884 	unsigned int first_valid;
4885 	unsigned int last_valid;
4886 	unsigned int i;
4887 
4888 	first_valid = CPU_FFS(&ctx->ifc_cpus) - 1;
4889 	last_valid = CPU_FLS(&ctx->ifc_cpus) - 1;
4890 
4891 	if (base_cpuid != CORE_OFFSET_UNSPECIFIED) {
4892 		/*
4893 		 * Align the user-chosen base CPU ID to the next valid CPU
4894 		 * for this device.  If the chosen base CPU ID is smaller
4895 		 * than the first valid CPU or larger than the last valid
4896 		 * CPU, we assume the user does not know what the valid
4897 		 * range is for this device and is thinking in terms of a
4898 		 * zero-based reference frame, and so we shift the given
4899 		 * value into the valid range (and wrap accordingly) so the
4900 		 * intent is translated to the proper frame of reference.
4901 		 * If the base CPU ID is within the valid first/last, but
4902 		 * does not correspond to a valid CPU, it is advanced to the
4903 		 * next valid CPU (wrapping if necessary).
4904 		 */
4905 		if (base_cpuid < first_valid || base_cpuid > last_valid) {
4906 			/* shift from zero-based to first_valid-based */
4907 			base_cpuid += first_valid;
4908 			/* wrap to range [first_valid, last_valid] */
4909 			base_cpuid = (base_cpuid - first_valid) %
4910 			    (last_valid - first_valid + 1);
4911 		}
4912 		if (!CPU_ISSET(base_cpuid, &ctx->ifc_cpus)) {
4913 			/*
4914 			 * base_cpuid is in [first_valid, last_valid], but
4915 			 * not a member of the valid set.  In this case,
4916 			 * there will always be a member of the valid set
4917 			 * with a CPU ID that is greater than base_cpuid,
4918 			 * and we simply advance to it.
4919 			 */
4920 			while (!CPU_ISSET(base_cpuid, &ctx->ifc_cpus))
4921 				base_cpuid++;
4922 		}
4923 		return (base_cpuid);
4924 	}
4925 
4926 	/*
4927 	 * Determine how many cores will be consumed by performing the CPU
4928 	 * assignments and counting how many of the assigned CPUs correspond
4929 	 * to CPUs in the set of context CPUs.  This is done using the CPU
4930 	 * ID first_valid as the base CPU ID, as the base CPU must be within
4931 	 * the set of context CPUs.
4932 	 *
4933 	 * Note not all assigned CPUs will be in the set of context CPUs
4934 	 * when separate CPUs are being allocated to TX and RX queues,
4935 	 * assignment to logical cores has been enabled, the set of context
4936 	 * CPUs contains only physical CPUs, and TX queues are mapped to L2
4937 	 * neighbors of CPUs that RX queues have been mapped to - in this
4938 	 * case we do only want to count how many CPUs in the set of context
4939 	 * CPUs have been consumed, as that determines the next CPU in that
4940 	 * set to start allocating at for the next device for which
4941 	 * core_offset is not set.
4942 	 */
4943 	CPU_ZERO(&assigned_cpus);
4944 	for (i = 0; i < scctx->isc_ntxqsets; i++)
4945 		CPU_SET(get_cpuid_for_queue(ctx, first_valid, i, true),
4946 		    &assigned_cpus);
4947 	for (i = 0; i < scctx->isc_nrxqsets; i++)
4948 		CPU_SET(get_cpuid_for_queue(ctx, first_valid, i, false),
4949 		    &assigned_cpus);
4950 	CPU_AND(&assigned_cpus, &ctx->ifc_cpus);
4951 	cores_consumed = CPU_COUNT(&assigned_cpus);
4952 
4953 	mtx_lock(&cpu_offset_mtx);
4954 	SLIST_FOREACH(op, &cpu_offsets, entries) {
4955 		if (CPU_CMP(&ctx->ifc_cpus, &op->set) == 0) {
4956 			base_cpuid = op->next_cpuid;
4957 			op->next_cpuid = cpuid_advance(ctx, op->next_cpuid,
4958 			    cores_consumed);
4959 			MPASS(op->refcount < UINT_MAX);
4960 			op->refcount++;
4961 			break;
4962 		}
4963 	}
4964 	if (base_cpuid == CORE_OFFSET_UNSPECIFIED) {
4965 		base_cpuid = first_valid;
4966 		op = malloc(sizeof(struct cpu_offset), M_IFLIB,
4967 		    M_NOWAIT | M_ZERO);
4968 		if (op == NULL) {
4969 			device_printf(ctx->ifc_dev,
4970 			    "allocation for cpu offset failed.\n");
4971 		} else {
4972 			op->next_cpuid = cpuid_advance(ctx, base_cpuid,
4973 			    cores_consumed);
4974 			op->refcount = 1;
4975 			CPU_COPY(&ctx->ifc_cpus, &op->set);
4976 			SLIST_INSERT_HEAD(&cpu_offsets, op, entries);
4977 		}
4978 	}
4979 	mtx_unlock(&cpu_offset_mtx);
4980 
4981 	return (base_cpuid);
4982 }
4983 
4984 static void
4985 unref_ctx_core_offset(if_ctx_t ctx)
4986 {
4987 	struct cpu_offset *op, *top;
4988 
4989 	mtx_lock(&cpu_offset_mtx);
4990 	SLIST_FOREACH_SAFE(op, &cpu_offsets, entries, top) {
4991 		if (CPU_CMP(&ctx->ifc_cpus, &op->set) == 0) {
4992 			MPASS(op->refcount > 0);
4993 			op->refcount--;
4994 			if (op->refcount == 0) {
4995 				SLIST_REMOVE(&cpu_offsets, op, cpu_offset, entries);
4996 				free(op, M_IFLIB);
4997 			}
4998 			break;
4999 		}
5000 	}
5001 	mtx_unlock(&cpu_offset_mtx);
5002 }
5003 
5004 int
5005 iflib_device_register(device_t dev, void *sc, if_shared_ctx_t sctx, if_ctx_t *ctxp)
5006 {
5007 	if_ctx_t ctx;
5008 	if_t ifp;
5009 	if_softc_ctx_t scctx;
5010 	kobjop_desc_t kobj_desc;
5011 	kobj_method_t *kobj_method;
5012 	int err, msix, rid;
5013 	int num_txd, num_rxd;
5014 
5015 	ctx = malloc(sizeof(* ctx), M_IFLIB, M_WAITOK|M_ZERO);
5016 
5017 	if (sc == NULL) {
5018 		sc = malloc(sctx->isc_driver->size, M_IFLIB, M_WAITOK|M_ZERO);
5019 		device_set_softc(dev, ctx);
5020 		ctx->ifc_flags |= IFC_SC_ALLOCATED;
5021 	}
5022 
5023 	ctx->ifc_sctx = sctx;
5024 	ctx->ifc_dev = dev;
5025 	ctx->ifc_softc = sc;
5026 
5027 	if ((err = iflib_register(ctx)) != 0) {
5028 		device_printf(dev, "iflib_register failed %d\n", err);
5029 		goto fail_ctx_free;
5030 	}
5031 	iflib_add_device_sysctl_pre(ctx);
5032 
5033 	scctx = &ctx->ifc_softc_ctx;
5034 	ifp = ctx->ifc_ifp;
5035 
5036 	iflib_reset_qvalues(ctx);
5037 	CTX_LOCK(ctx);
5038 	if ((err = IFDI_ATTACH_PRE(ctx)) != 0) {
5039 		device_printf(dev, "IFDI_ATTACH_PRE failed %d\n", err);
5040 		goto fail_unlock;
5041 	}
5042 	_iflib_pre_assert(scctx);
5043 	ctx->ifc_txrx = *scctx->isc_txrx;
5044 
5045 	MPASS(scctx->isc_dma_width <= flsll(BUS_SPACE_MAXADDR));
5046 
5047 	if (sctx->isc_flags & IFLIB_DRIVER_MEDIA)
5048 		ctx->ifc_mediap = scctx->isc_media;
5049 
5050 #ifdef INVARIANTS
5051 	if (scctx->isc_capabilities & IFCAP_TXCSUM)
5052 		MPASS(scctx->isc_tx_csum_flags);
5053 #endif
5054 
5055 	if_setcapabilities(ifp,
5056 	    scctx->isc_capabilities | IFCAP_HWSTATS | IFCAP_MEXTPG);
5057 	if_setcapenable(ifp,
5058 	    scctx->isc_capenable | IFCAP_HWSTATS | IFCAP_MEXTPG);
5059 
5060 	if (scctx->isc_ntxqsets == 0 || (scctx->isc_ntxqsets_max && scctx->isc_ntxqsets_max < scctx->isc_ntxqsets))
5061 		scctx->isc_ntxqsets = scctx->isc_ntxqsets_max;
5062 	if (scctx->isc_nrxqsets == 0 || (scctx->isc_nrxqsets_max && scctx->isc_nrxqsets_max < scctx->isc_nrxqsets))
5063 		scctx->isc_nrxqsets = scctx->isc_nrxqsets_max;
5064 
5065 	num_txd = iflib_num_tx_descs(ctx);
5066 	num_rxd = iflib_num_rx_descs(ctx);
5067 
5068 	/* XXX change for per-queue sizes */
5069 	device_printf(dev, "Using %d TX descriptors and %d RX descriptors\n",
5070 	    num_txd, num_rxd);
5071 
5072 	if (scctx->isc_tx_nsegments > num_txd / MAX_SINGLE_PACKET_FRACTION)
5073 		scctx->isc_tx_nsegments = max(1, num_txd /
5074 		    MAX_SINGLE_PACKET_FRACTION);
5075 	if (scctx->isc_tx_tso_segments_max > num_txd /
5076 	    MAX_SINGLE_PACKET_FRACTION)
5077 		scctx->isc_tx_tso_segments_max = max(1,
5078 		    num_txd / MAX_SINGLE_PACKET_FRACTION);
5079 
5080 	/* TSO parameters - dig these out of the data sheet - simply correspond to tag setup */
5081 	if (if_getcapabilities(ifp) & IFCAP_TSO) {
5082 		/*
5083 		 * The stack can't handle a TSO size larger than IP_MAXPACKET,
5084 		 * but some MACs do.
5085 		 */
5086 		if_sethwtsomax(ifp, min(scctx->isc_tx_tso_size_max,
5087 		    IP_MAXPACKET));
5088 		/*
5089 		 * Take maximum number of m_pullup(9)'s in iflib_parse_header()
5090 		 * into account.  In the worst case, each of these calls will
5091 		 * add another mbuf and, thus, the requirement for another DMA
5092 		 * segment.  So for best performance, it doesn't make sense to
5093 		 * advertize a maximum of TSO segments that typically will
5094 		 * require defragmentation in iflib_encap().
5095 		 */
5096 		if_sethwtsomaxsegcount(ifp, scctx->isc_tx_tso_segments_max - 3);
5097 		if_sethwtsomaxsegsize(ifp, scctx->isc_tx_tso_segsize_max);
5098 	}
5099 	if (scctx->isc_rss_table_size == 0)
5100 		scctx->isc_rss_table_size = 64;
5101 	scctx->isc_rss_table_mask = scctx->isc_rss_table_size-1;
5102 
5103 	GROUPTASK_INIT(&ctx->ifc_admin_task, 0, _task_fn_admin, ctx);
5104 	/* XXX format name */
5105 	taskqgroup_attach(qgroup_if_config_tqg, &ctx->ifc_admin_task, ctx,
5106 	    NULL, NULL, "admin");
5107 
5108 	/* Set up cpu set.  If it fails, use the set of all CPUs. */
5109 	if (bus_get_cpus(dev, INTR_CPUS, sizeof(ctx->ifc_cpus), &ctx->ifc_cpus) != 0) {
5110 		device_printf(dev, "Unable to fetch CPU list\n");
5111 		CPU_COPY(&all_cpus, &ctx->ifc_cpus);
5112 		ctx->ifc_cpus_are_physical_cores = false;
5113 	} else
5114 		ctx->ifc_cpus_are_physical_cores = true;
5115 	MPASS(CPU_COUNT(&ctx->ifc_cpus) > 0);
5116 
5117 	/*
5118 	** Now set up MSI or MSI-X, should return us the number of supported
5119 	** vectors (will be 1 for a legacy interrupt and MSI).
5120 	*/
5121 	if (sctx->isc_flags & IFLIB_SKIP_MSIX) {
5122 		msix = scctx->isc_vectors;
5123 	} else if (scctx->isc_msix_bar != 0)
5124 	       /*
5125 		* The simple fact that isc_msix_bar is not 0 does not mean we
5126 		* we have a good value there that is known to work.
5127 		*/
5128 		msix = iflib_msix_init(ctx);
5129 	else {
5130 		scctx->isc_vectors = 1;
5131 		scctx->isc_ntxqsets = 1;
5132 		scctx->isc_nrxqsets = 1;
5133 		scctx->isc_intr = IFLIB_INTR_LEGACY;
5134 		msix = 0;
5135 	}
5136 	/* Get memory for the station queues */
5137 	if ((err = iflib_queues_alloc(ctx))) {
5138 		device_printf(dev, "Unable to allocate queue memory\n");
5139 		goto fail_intr_free;
5140 	}
5141 
5142 	if ((err = iflib_qset_structures_setup(ctx)))
5143 		goto fail_queues;
5144 
5145 	/*
5146 	 * Now that we know how many queues there are, get the core offset.
5147 	 */
5148 	ctx->ifc_sysctl_core_offset = get_ctx_core_offset(ctx);
5149 
5150 	if (msix > 1) {
5151 		/*
5152 		 * When using MSI-X, ensure that ifdi_{r,t}x_queue_intr_enable
5153 		 * aren't the default NULL implementation.
5154 		 */
5155 		kobj_desc = &ifdi_rx_queue_intr_enable_desc;
5156 		kobj_method = kobj_lookup_method(((kobj_t)ctx)->ops->cls, NULL,
5157 		    kobj_desc);
5158 		if (kobj_method == &kobj_desc->deflt) {
5159 			device_printf(dev,
5160 			    "MSI-X requires ifdi_rx_queue_intr_enable method");
5161 			err = EOPNOTSUPP;
5162 			goto fail_queues;
5163 		}
5164 		kobj_desc = &ifdi_tx_queue_intr_enable_desc;
5165 		kobj_method = kobj_lookup_method(((kobj_t)ctx)->ops->cls, NULL,
5166 		    kobj_desc);
5167 		if (kobj_method == &kobj_desc->deflt) {
5168 			device_printf(dev,
5169 			    "MSI-X requires ifdi_tx_queue_intr_enable method");
5170 			err = EOPNOTSUPP;
5171 			goto fail_queues;
5172 		}
5173 
5174 		/*
5175 		 * Assign the MSI-X vectors.
5176 		 * Note that the default NULL ifdi_msix_intr_assign method will
5177 		 * fail here, too.
5178 		 */
5179 		err = IFDI_MSIX_INTR_ASSIGN(ctx, msix);
5180 		if (err != 0) {
5181 			device_printf(dev, "IFDI_MSIX_INTR_ASSIGN failed %d\n",
5182 			    err);
5183 			goto fail_queues;
5184 		}
5185 	} else if (scctx->isc_intr != IFLIB_INTR_MSIX) {
5186 		rid = 0;
5187 		if (scctx->isc_intr == IFLIB_INTR_MSI) {
5188 			MPASS(msix == 1);
5189 			rid = 1;
5190 		}
5191 		if ((err = iflib_legacy_setup(ctx, ctx->isc_legacy_intr, ctx->ifc_softc, &rid, "irq0")) != 0) {
5192 			device_printf(dev, "iflib_legacy_setup failed %d\n", err);
5193 			goto fail_queues;
5194 		}
5195 	} else {
5196 		device_printf(dev,
5197 		    "Cannot use iflib with only 1 MSI-X interrupt!\n");
5198 		err = ENODEV;
5199 		goto fail_queues;
5200 	}
5201 
5202 	ether_ifattach(ctx->ifc_ifp, ctx->ifc_mac.octet);
5203 
5204 	if ((err = IFDI_ATTACH_POST(ctx)) != 0) {
5205 		device_printf(dev, "IFDI_ATTACH_POST failed %d\n", err);
5206 		goto fail_detach;
5207 	}
5208 
5209 	/*
5210 	 * Tell the upper layer(s) if IFCAP_VLAN_MTU is supported.
5211 	 * This must appear after the call to ether_ifattach() because
5212 	 * ether_ifattach() sets if_hdrlen to the default value.
5213 	 */
5214 	if (if_getcapabilities(ifp) & IFCAP_VLAN_MTU)
5215 		if_setifheaderlen(ifp, sizeof(struct ether_vlan_header));
5216 
5217 	if ((err = iflib_netmap_attach(ctx))) {
5218 		device_printf(ctx->ifc_dev, "netmap attach failed: %d\n", err);
5219 		goto fail_detach;
5220 	}
5221 	*ctxp = ctx;
5222 
5223 	DEBUGNET_SET(ctx->ifc_ifp, iflib);
5224 
5225 	if_setgetcounterfn(ctx->ifc_ifp, iflib_if_get_counter);
5226 	iflib_add_device_sysctl_post(ctx);
5227 	iflib_add_pfil(ctx);
5228 	ctx->ifc_flags |= IFC_INIT_DONE;
5229 	CTX_UNLOCK(ctx);
5230 
5231 	return (0);
5232 
5233 fail_detach:
5234 	ether_ifdetach(ctx->ifc_ifp);
5235 fail_queues:
5236 	iflib_tqg_detach(ctx);
5237 	iflib_tx_structures_free(ctx);
5238 	iflib_rx_structures_free(ctx);
5239 	IFDI_DETACH(ctx);
5240 	IFDI_QUEUES_FREE(ctx);
5241 fail_intr_free:
5242 	iflib_free_intr_mem(ctx);
5243 fail_unlock:
5244 	CTX_UNLOCK(ctx);
5245 	iflib_deregister(ctx);
5246 fail_ctx_free:
5247 	device_set_softc(ctx->ifc_dev, NULL);
5248         if (ctx->ifc_flags & IFC_SC_ALLOCATED)
5249                 free(ctx->ifc_softc, M_IFLIB);
5250         free(ctx, M_IFLIB);
5251 	return (err);
5252 }
5253 
5254 int
5255 iflib_pseudo_register(device_t dev, if_shared_ctx_t sctx, if_ctx_t *ctxp,
5256 					  struct iflib_cloneattach_ctx *clctx)
5257 {
5258 	int num_txd, num_rxd;
5259 	int err;
5260 	if_ctx_t ctx;
5261 	if_t ifp;
5262 	if_softc_ctx_t scctx;
5263 	int i;
5264 	void *sc;
5265 
5266 	ctx = malloc(sizeof(*ctx), M_IFLIB, M_WAITOK|M_ZERO);
5267 	sc = malloc(sctx->isc_driver->size, M_IFLIB, M_WAITOK|M_ZERO);
5268 	ctx->ifc_flags |= IFC_SC_ALLOCATED;
5269 	if (sctx->isc_flags & (IFLIB_PSEUDO|IFLIB_VIRTUAL))
5270 		ctx->ifc_flags |= IFC_PSEUDO;
5271 
5272 	ctx->ifc_sctx = sctx;
5273 	ctx->ifc_softc = sc;
5274 	ctx->ifc_dev = dev;
5275 
5276 	if ((err = iflib_register(ctx)) != 0) {
5277 		device_printf(dev, "%s: iflib_register failed %d\n", __func__, err);
5278 		goto fail_ctx_free;
5279 	}
5280 	iflib_add_device_sysctl_pre(ctx);
5281 
5282 	scctx = &ctx->ifc_softc_ctx;
5283 	ifp = ctx->ifc_ifp;
5284 
5285 	iflib_reset_qvalues(ctx);
5286 	CTX_LOCK(ctx);
5287 	if ((err = IFDI_ATTACH_PRE(ctx)) != 0) {
5288 		device_printf(dev, "IFDI_ATTACH_PRE failed %d\n", err);
5289 		goto fail_unlock;
5290 	}
5291 	if (sctx->isc_flags & IFLIB_GEN_MAC)
5292 		ether_gen_addr(ifp, &ctx->ifc_mac);
5293 	if ((err = IFDI_CLONEATTACH(ctx, clctx->cc_ifc, clctx->cc_name,
5294 								clctx->cc_params)) != 0) {
5295 		device_printf(dev, "IFDI_CLONEATTACH failed %d\n", err);
5296 		goto fail_unlock;
5297 	}
5298 #ifdef INVARIANTS
5299 	if (scctx->isc_capabilities & IFCAP_TXCSUM)
5300 		MPASS(scctx->isc_tx_csum_flags);
5301 #endif
5302 
5303 	if_setcapabilities(ifp, scctx->isc_capabilities | IFCAP_HWSTATS | IFCAP_LINKSTATE);
5304 	if_setcapenable(ifp, scctx->isc_capenable | IFCAP_HWSTATS | IFCAP_LINKSTATE);
5305 
5306 	ifp->if_flags |= IFF_NOGROUP;
5307 	if (sctx->isc_flags & IFLIB_PSEUDO) {
5308 		ifmedia_add(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO, 0, NULL);
5309 		ifmedia_set(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO);
5310 		if (sctx->isc_flags & IFLIB_PSEUDO_ETHER) {
5311 			ether_ifattach(ctx->ifc_ifp, ctx->ifc_mac.octet);
5312 		} else {
5313 			if_attach(ctx->ifc_ifp);
5314 			bpfattach(ctx->ifc_ifp, DLT_NULL, sizeof(u_int32_t));
5315 		}
5316 
5317 		if ((err = IFDI_ATTACH_POST(ctx)) != 0) {
5318 			device_printf(dev, "IFDI_ATTACH_POST failed %d\n", err);
5319 			goto fail_detach;
5320 		}
5321 		*ctxp = ctx;
5322 
5323 		/*
5324 		 * Tell the upper layer(s) if IFCAP_VLAN_MTU is supported.
5325 		 * This must appear after the call to ether_ifattach() because
5326 		 * ether_ifattach() sets if_hdrlen to the default value.
5327 		 */
5328 		if (if_getcapabilities(ifp) & IFCAP_VLAN_MTU)
5329 			if_setifheaderlen(ifp,
5330 			    sizeof(struct ether_vlan_header));
5331 
5332 		if_setgetcounterfn(ctx->ifc_ifp, iflib_if_get_counter);
5333 		iflib_add_device_sysctl_post(ctx);
5334 		ctx->ifc_flags |= IFC_INIT_DONE;
5335 		CTX_UNLOCK(ctx);
5336 		return (0);
5337 	}
5338 	ifmedia_add(ctx->ifc_mediap, IFM_ETHER | IFM_1000_T | IFM_FDX, 0, NULL);
5339 	ifmedia_add(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO, 0, NULL);
5340 	ifmedia_set(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO);
5341 
5342 	_iflib_pre_assert(scctx);
5343 	ctx->ifc_txrx = *scctx->isc_txrx;
5344 
5345 	if (scctx->isc_ntxqsets == 0 || (scctx->isc_ntxqsets_max && scctx->isc_ntxqsets_max < scctx->isc_ntxqsets))
5346 		scctx->isc_ntxqsets = scctx->isc_ntxqsets_max;
5347 	if (scctx->isc_nrxqsets == 0 || (scctx->isc_nrxqsets_max && scctx->isc_nrxqsets_max < scctx->isc_nrxqsets))
5348 		scctx->isc_nrxqsets = scctx->isc_nrxqsets_max;
5349 
5350 	num_txd = iflib_num_tx_descs(ctx);
5351 	num_rxd = iflib_num_rx_descs(ctx);
5352 
5353 	/* XXX change for per-queue sizes */
5354 	device_printf(dev, "Using %d TX descriptors and %d RX descriptors\n",
5355 	    num_txd, num_rxd);
5356 
5357 	if (scctx->isc_tx_nsegments > num_txd / MAX_SINGLE_PACKET_FRACTION)
5358 		scctx->isc_tx_nsegments = max(1, num_txd /
5359 		    MAX_SINGLE_PACKET_FRACTION);
5360 	if (scctx->isc_tx_tso_segments_max > num_txd /
5361 	    MAX_SINGLE_PACKET_FRACTION)
5362 		scctx->isc_tx_tso_segments_max = max(1,
5363 		    num_txd / MAX_SINGLE_PACKET_FRACTION);
5364 
5365 	/* TSO parameters - dig these out of the data sheet - simply correspond to tag setup */
5366 	if (if_getcapabilities(ifp) & IFCAP_TSO) {
5367 		/*
5368 		 * The stack can't handle a TSO size larger than IP_MAXPACKET,
5369 		 * but some MACs do.
5370 		 */
5371 		if_sethwtsomax(ifp, min(scctx->isc_tx_tso_size_max,
5372 		    IP_MAXPACKET));
5373 		/*
5374 		 * Take maximum number of m_pullup(9)'s in iflib_parse_header()
5375 		 * into account.  In the worst case, each of these calls will
5376 		 * add another mbuf and, thus, the requirement for another DMA
5377 		 * segment.  So for best performance, it doesn't make sense to
5378 		 * advertize a maximum of TSO segments that typically will
5379 		 * require defragmentation in iflib_encap().
5380 		 */
5381 		if_sethwtsomaxsegcount(ifp, scctx->isc_tx_tso_segments_max - 3);
5382 		if_sethwtsomaxsegsize(ifp, scctx->isc_tx_tso_segsize_max);
5383 	}
5384 	if (scctx->isc_rss_table_size == 0)
5385 		scctx->isc_rss_table_size = 64;
5386 	scctx->isc_rss_table_mask = scctx->isc_rss_table_size-1;
5387 
5388 	GROUPTASK_INIT(&ctx->ifc_admin_task, 0, _task_fn_admin, ctx);
5389 	/* XXX format name */
5390 	taskqgroup_attach(qgroup_if_config_tqg, &ctx->ifc_admin_task, ctx,
5391 	    NULL, NULL, "admin");
5392 
5393 	/* XXX --- can support > 1 -- but keep it simple for now */
5394 	scctx->isc_intr = IFLIB_INTR_LEGACY;
5395 
5396 	/* Get memory for the station queues */
5397 	if ((err = iflib_queues_alloc(ctx))) {
5398 		device_printf(dev, "Unable to allocate queue memory\n");
5399 		goto fail_iflib_detach;
5400 	}
5401 
5402 	if ((err = iflib_qset_structures_setup(ctx))) {
5403 		device_printf(dev, "qset structure setup failed %d\n", err);
5404 		goto fail_queues;
5405 	}
5406 
5407 	/*
5408 	 * XXX What if anything do we want to do about interrupts?
5409 	 */
5410 	ether_ifattach(ctx->ifc_ifp, ctx->ifc_mac.octet);
5411 	if ((err = IFDI_ATTACH_POST(ctx)) != 0) {
5412 		device_printf(dev, "IFDI_ATTACH_POST failed %d\n", err);
5413 		goto fail_detach;
5414 	}
5415 
5416 	/*
5417 	 * Tell the upper layer(s) if IFCAP_VLAN_MTU is supported.
5418 	 * This must appear after the call to ether_ifattach() because
5419 	 * ether_ifattach() sets if_hdrlen to the default value.
5420 	 */
5421 	if (if_getcapabilities(ifp) & IFCAP_VLAN_MTU)
5422 		if_setifheaderlen(ifp, sizeof(struct ether_vlan_header));
5423 
5424 	/* XXX handle more than one queue */
5425 	for (i = 0; i < scctx->isc_nrxqsets; i++)
5426 		IFDI_RX_CLSET(ctx, 0, i, ctx->ifc_rxqs[i].ifr_fl[0].ifl_sds.ifsd_cl);
5427 
5428 	*ctxp = ctx;
5429 
5430 	if_setgetcounterfn(ctx->ifc_ifp, iflib_if_get_counter);
5431 	iflib_add_device_sysctl_post(ctx);
5432 	ctx->ifc_flags |= IFC_INIT_DONE;
5433 	CTX_UNLOCK(ctx);
5434 
5435 	return (0);
5436 fail_detach:
5437 	ether_ifdetach(ctx->ifc_ifp);
5438 fail_queues:
5439 	iflib_tqg_detach(ctx);
5440 	iflib_tx_structures_free(ctx);
5441 	iflib_rx_structures_free(ctx);
5442 fail_iflib_detach:
5443 	IFDI_DETACH(ctx);
5444 	IFDI_QUEUES_FREE(ctx);
5445 fail_unlock:
5446 	CTX_UNLOCK(ctx);
5447 	iflib_deregister(ctx);
5448 fail_ctx_free:
5449 	free(ctx->ifc_softc, M_IFLIB);
5450 	free(ctx, M_IFLIB);
5451 	return (err);
5452 }
5453 
5454 int
5455 iflib_pseudo_deregister(if_ctx_t ctx)
5456 {
5457 	if_t ifp = ctx->ifc_ifp;
5458 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5459 
5460 	/* Unregister VLAN event handlers early */
5461 	iflib_unregister_vlan_handlers(ctx);
5462 
5463 	if ((sctx->isc_flags & IFLIB_PSEUDO)  &&
5464 		(sctx->isc_flags & IFLIB_PSEUDO_ETHER) == 0) {
5465 		bpfdetach(ifp);
5466 		if_detach(ifp);
5467 	} else {
5468 		ether_ifdetach(ifp);
5469 	}
5470 
5471 	iflib_tqg_detach(ctx);
5472 	iflib_tx_structures_free(ctx);
5473 	iflib_rx_structures_free(ctx);
5474 	IFDI_DETACH(ctx);
5475 	IFDI_QUEUES_FREE(ctx);
5476 
5477 	iflib_deregister(ctx);
5478 
5479 	if (ctx->ifc_flags & IFC_SC_ALLOCATED)
5480 		free(ctx->ifc_softc, M_IFLIB);
5481 	free(ctx, M_IFLIB);
5482 	return (0);
5483 }
5484 
5485 int
5486 iflib_device_attach(device_t dev)
5487 {
5488 	if_ctx_t ctx;
5489 	if_shared_ctx_t sctx;
5490 
5491 	if ((sctx = DEVICE_REGISTER(dev)) == NULL || sctx->isc_magic != IFLIB_MAGIC)
5492 		return (ENOTSUP);
5493 
5494 	pci_enable_busmaster(dev);
5495 
5496 	return (iflib_device_register(dev, NULL, sctx, &ctx));
5497 }
5498 
5499 int
5500 iflib_device_deregister(if_ctx_t ctx)
5501 {
5502 	if_t ifp = ctx->ifc_ifp;
5503 	device_t dev = ctx->ifc_dev;
5504 
5505 	/* Make sure VLANS are not using driver */
5506 	if (if_vlantrunkinuse(ifp)) {
5507 		device_printf(dev, "Vlan in use, detach first\n");
5508 		return (EBUSY);
5509 	}
5510 #ifdef PCI_IOV
5511 	if (!CTX_IS_VF(ctx) && pci_iov_detach(dev) != 0) {
5512 		device_printf(dev, "SR-IOV in use; detach first.\n");
5513 		return (EBUSY);
5514 	}
5515 #endif
5516 
5517 	STATE_LOCK(ctx);
5518 	ctx->ifc_flags |= IFC_IN_DETACH;
5519 	STATE_UNLOCK(ctx);
5520 
5521 	/* Unregister VLAN handlers before calling iflib_stop() */
5522 	iflib_unregister_vlan_handlers(ctx);
5523 
5524 	iflib_netmap_detach(ifp);
5525 	ether_ifdetach(ifp);
5526 
5527 	CTX_LOCK(ctx);
5528 	iflib_stop(ctx);
5529 	CTX_UNLOCK(ctx);
5530 
5531 	iflib_rem_pfil(ctx);
5532 	if (ctx->ifc_led_dev != NULL)
5533 		led_destroy(ctx->ifc_led_dev);
5534 
5535 	iflib_tqg_detach(ctx);
5536 	iflib_tx_structures_free(ctx);
5537 	iflib_rx_structures_free(ctx);
5538 
5539 	CTX_LOCK(ctx);
5540 	IFDI_DETACH(ctx);
5541 	IFDI_QUEUES_FREE(ctx);
5542 	CTX_UNLOCK(ctx);
5543 
5544 	/* ether_ifdetach calls if_qflush - lock must be destroy afterwards*/
5545 	iflib_free_intr_mem(ctx);
5546 
5547 	bus_generic_detach(dev);
5548 
5549 	iflib_deregister(ctx);
5550 
5551 	device_set_softc(ctx->ifc_dev, NULL);
5552 	if (ctx->ifc_flags & IFC_SC_ALLOCATED)
5553 		free(ctx->ifc_softc, M_IFLIB);
5554 	unref_ctx_core_offset(ctx);
5555 	free(ctx, M_IFLIB);
5556 	return (0);
5557 }
5558 
5559 static void
5560 iflib_tqg_detach(if_ctx_t ctx)
5561 {
5562 	iflib_txq_t txq;
5563 	iflib_rxq_t rxq;
5564 	int i;
5565 	struct taskqgroup *tqg;
5566 
5567 	/* XXX drain any dependent tasks */
5568 	tqg = qgroup_if_io_tqg;
5569 	for (txq = ctx->ifc_txqs, i = 0; i < NTXQSETS(ctx); i++, txq++) {
5570 		callout_drain(&txq->ift_timer);
5571 #ifdef DEV_NETMAP
5572 		callout_drain(&txq->ift_netmap_timer);
5573 #endif /* DEV_NETMAP */
5574 		if (txq->ift_task.gt_uniq != NULL)
5575 			taskqgroup_detach(tqg, &txq->ift_task);
5576 	}
5577 	for (i = 0, rxq = ctx->ifc_rxqs; i < NRXQSETS(ctx); i++, rxq++) {
5578 		if (rxq->ifr_task.gt_uniq != NULL)
5579 			taskqgroup_detach(tqg, &rxq->ifr_task);
5580 	}
5581 	tqg = qgroup_if_config_tqg;
5582 	if (ctx->ifc_admin_task.gt_uniq != NULL)
5583 		taskqgroup_detach(tqg, &ctx->ifc_admin_task);
5584 	if (ctx->ifc_vflr_task.gt_uniq != NULL)
5585 		taskqgroup_detach(tqg, &ctx->ifc_vflr_task);
5586 }
5587 
5588 static void
5589 iflib_free_intr_mem(if_ctx_t ctx)
5590 {
5591 
5592 	if (ctx->ifc_softc_ctx.isc_intr != IFLIB_INTR_MSIX) {
5593 		iflib_irq_free(ctx, &ctx->ifc_legacy_irq);
5594 	}
5595 	if (ctx->ifc_softc_ctx.isc_intr != IFLIB_INTR_LEGACY) {
5596 		pci_release_msi(ctx->ifc_dev);
5597 	}
5598 	if (ctx->ifc_msix_mem != NULL) {
5599 		bus_release_resource(ctx->ifc_dev, SYS_RES_MEMORY,
5600 		    rman_get_rid(ctx->ifc_msix_mem), ctx->ifc_msix_mem);
5601 		ctx->ifc_msix_mem = NULL;
5602 	}
5603 }
5604 
5605 int
5606 iflib_device_detach(device_t dev)
5607 {
5608 	if_ctx_t ctx = device_get_softc(dev);
5609 
5610 	return (iflib_device_deregister(ctx));
5611 }
5612 
5613 int
5614 iflib_device_suspend(device_t dev)
5615 {
5616 	if_ctx_t ctx = device_get_softc(dev);
5617 
5618 	CTX_LOCK(ctx);
5619 	IFDI_SUSPEND(ctx);
5620 	CTX_UNLOCK(ctx);
5621 
5622 	return bus_generic_suspend(dev);
5623 }
5624 int
5625 iflib_device_shutdown(device_t dev)
5626 {
5627 	if_ctx_t ctx = device_get_softc(dev);
5628 
5629 	CTX_LOCK(ctx);
5630 	IFDI_SHUTDOWN(ctx);
5631 	CTX_UNLOCK(ctx);
5632 
5633 	return bus_generic_suspend(dev);
5634 }
5635 
5636 int
5637 iflib_device_resume(device_t dev)
5638 {
5639 	if_ctx_t ctx = device_get_softc(dev);
5640 	iflib_txq_t txq = ctx->ifc_txqs;
5641 
5642 	CTX_LOCK(ctx);
5643 	IFDI_RESUME(ctx);
5644 	iflib_if_init_locked(ctx);
5645 	CTX_UNLOCK(ctx);
5646 	for (int i = 0; i < NTXQSETS(ctx); i++, txq++)
5647 		iflib_txq_check_drain(txq, IFLIB_RESTART_BUDGET);
5648 
5649 	return (bus_generic_resume(dev));
5650 }
5651 
5652 int
5653 iflib_device_iov_init(device_t dev, uint16_t num_vfs, const nvlist_t *params)
5654 {
5655 	int error;
5656 	if_ctx_t ctx = device_get_softc(dev);
5657 
5658 	CTX_LOCK(ctx);
5659 	error = IFDI_IOV_INIT(ctx, num_vfs, params);
5660 	CTX_UNLOCK(ctx);
5661 
5662 	return (error);
5663 }
5664 
5665 void
5666 iflib_device_iov_uninit(device_t dev)
5667 {
5668 	if_ctx_t ctx = device_get_softc(dev);
5669 
5670 	CTX_LOCK(ctx);
5671 	IFDI_IOV_UNINIT(ctx);
5672 	CTX_UNLOCK(ctx);
5673 }
5674 
5675 int
5676 iflib_device_iov_add_vf(device_t dev, uint16_t vfnum, const nvlist_t *params)
5677 {
5678 	int error;
5679 	if_ctx_t ctx = device_get_softc(dev);
5680 
5681 	CTX_LOCK(ctx);
5682 	error = IFDI_IOV_VF_ADD(ctx, vfnum, params);
5683 	CTX_UNLOCK(ctx);
5684 
5685 	return (error);
5686 }
5687 
5688 /*********************************************************************
5689  *
5690  *  MODULE FUNCTION DEFINITIONS
5691  *
5692  **********************************************************************/
5693 
5694 /*
5695  * - Start a fast taskqueue thread for each core
5696  * - Start a taskqueue for control operations
5697  */
5698 static int
5699 iflib_module_init(void)
5700 {
5701 	iflib_timer_default = hz / 2;
5702 	return (0);
5703 }
5704 
5705 static int
5706 iflib_module_event_handler(module_t mod, int what, void *arg)
5707 {
5708 	int err;
5709 
5710 	switch (what) {
5711 	case MOD_LOAD:
5712 		if ((err = iflib_module_init()) != 0)
5713 			return (err);
5714 		break;
5715 	case MOD_UNLOAD:
5716 		return (EBUSY);
5717 	default:
5718 		return (EOPNOTSUPP);
5719 	}
5720 
5721 	return (0);
5722 }
5723 
5724 /*********************************************************************
5725  *
5726  *  PUBLIC FUNCTION DEFINITIONS
5727  *     ordered as in iflib.h
5728  *
5729  **********************************************************************/
5730 
5731 static void
5732 _iflib_assert(if_shared_ctx_t sctx)
5733 {
5734 	int i;
5735 
5736 	MPASS(sctx->isc_tx_maxsize);
5737 	MPASS(sctx->isc_tx_maxsegsize);
5738 
5739 	MPASS(sctx->isc_rx_maxsize);
5740 	MPASS(sctx->isc_rx_nsegments);
5741 	MPASS(sctx->isc_rx_maxsegsize);
5742 
5743 	MPASS(sctx->isc_nrxqs >= 1 && sctx->isc_nrxqs <= 8);
5744 	for (i = 0; i < sctx->isc_nrxqs; i++) {
5745 		MPASS(sctx->isc_nrxd_min[i]);
5746 		MPASS(powerof2(sctx->isc_nrxd_min[i]));
5747 		MPASS(sctx->isc_nrxd_max[i]);
5748 		MPASS(powerof2(sctx->isc_nrxd_max[i]));
5749 		MPASS(sctx->isc_nrxd_default[i]);
5750 		MPASS(powerof2(sctx->isc_nrxd_default[i]));
5751 	}
5752 
5753 	MPASS(sctx->isc_ntxqs >= 1 && sctx->isc_ntxqs <= 8);
5754 	for (i = 0; i < sctx->isc_ntxqs; i++) {
5755 		MPASS(sctx->isc_ntxd_min[i]);
5756 		MPASS(powerof2(sctx->isc_ntxd_min[i]));
5757 		MPASS(sctx->isc_ntxd_max[i]);
5758 		MPASS(powerof2(sctx->isc_ntxd_max[i]));
5759 		MPASS(sctx->isc_ntxd_default[i]);
5760 		MPASS(powerof2(sctx->isc_ntxd_default[i]));
5761 	}
5762 }
5763 
5764 static void
5765 _iflib_pre_assert(if_softc_ctx_t scctx)
5766 {
5767 
5768 	MPASS(scctx->isc_txrx->ift_txd_encap);
5769 	MPASS(scctx->isc_txrx->ift_txd_flush);
5770 	MPASS(scctx->isc_txrx->ift_txd_credits_update);
5771 	MPASS(scctx->isc_txrx->ift_rxd_available);
5772 	MPASS(scctx->isc_txrx->ift_rxd_pkt_get);
5773 	MPASS(scctx->isc_txrx->ift_rxd_refill);
5774 	MPASS(scctx->isc_txrx->ift_rxd_flush);
5775 }
5776 
5777 static int
5778 iflib_register(if_ctx_t ctx)
5779 {
5780 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5781 	driver_t *driver = sctx->isc_driver;
5782 	device_t dev = ctx->ifc_dev;
5783 	if_t ifp;
5784 	u_char type;
5785 	int iflags;
5786 
5787 	if ((sctx->isc_flags & IFLIB_PSEUDO) == 0)
5788 		_iflib_assert(sctx);
5789 
5790 	CTX_LOCK_INIT(ctx);
5791 	STATE_LOCK_INIT(ctx, device_get_nameunit(ctx->ifc_dev));
5792 	if (sctx->isc_flags & IFLIB_PSEUDO) {
5793 		if (sctx->isc_flags & IFLIB_PSEUDO_ETHER)
5794 			type = IFT_ETHER;
5795 		else
5796 			type = IFT_PPP;
5797 	} else
5798 		type = IFT_ETHER;
5799 	ifp = ctx->ifc_ifp = if_alloc(type);
5800 	if (ifp == NULL) {
5801 		device_printf(dev, "can not allocate ifnet structure\n");
5802 		return (ENOMEM);
5803 	}
5804 
5805 	/*
5806 	 * Initialize our context's device specific methods
5807 	 */
5808 	kobj_init((kobj_t) ctx, (kobj_class_t) driver);
5809 	kobj_class_compile((kobj_class_t) driver);
5810 
5811 	if_initname(ifp, device_get_name(dev), device_get_unit(dev));
5812 	if_setsoftc(ifp, ctx);
5813 	if_setdev(ifp, dev);
5814 	if_setinitfn(ifp, iflib_if_init);
5815 	if_setioctlfn(ifp, iflib_if_ioctl);
5816 #ifdef ALTQ
5817 	if_setstartfn(ifp, iflib_altq_if_start);
5818 	if_settransmitfn(ifp, iflib_altq_if_transmit);
5819 	if_setsendqready(ifp);
5820 #else
5821 	if_settransmitfn(ifp, iflib_if_transmit);
5822 #endif
5823 	if_setqflushfn(ifp, iflib_if_qflush);
5824 	iflags = IFF_MULTICAST | IFF_KNOWSEPOCH;
5825 
5826 	if ((sctx->isc_flags & IFLIB_PSEUDO) &&
5827 		(sctx->isc_flags & IFLIB_PSEUDO_ETHER) == 0)
5828 		iflags |= IFF_POINTOPOINT;
5829 	else
5830 		iflags |= IFF_BROADCAST | IFF_SIMPLEX;
5831 	if_setflags(ifp, iflags);
5832 	ctx->ifc_vlan_attach_event =
5833 		EVENTHANDLER_REGISTER(vlan_config, iflib_vlan_register, ctx,
5834 							  EVENTHANDLER_PRI_FIRST);
5835 	ctx->ifc_vlan_detach_event =
5836 		EVENTHANDLER_REGISTER(vlan_unconfig, iflib_vlan_unregister, ctx,
5837 							  EVENTHANDLER_PRI_FIRST);
5838 
5839 	if ((sctx->isc_flags & IFLIB_DRIVER_MEDIA) == 0) {
5840 		ctx->ifc_mediap = &ctx->ifc_media;
5841 		ifmedia_init(ctx->ifc_mediap, IFM_IMASK,
5842 		    iflib_media_change, iflib_media_status);
5843 	}
5844 	return (0);
5845 }
5846 
5847 static void
5848 iflib_unregister_vlan_handlers(if_ctx_t ctx)
5849 {
5850 	/* Unregister VLAN events */
5851 	if (ctx->ifc_vlan_attach_event != NULL) {
5852 		EVENTHANDLER_DEREGISTER(vlan_config, ctx->ifc_vlan_attach_event);
5853 		ctx->ifc_vlan_attach_event = NULL;
5854 	}
5855 	if (ctx->ifc_vlan_detach_event != NULL) {
5856 		EVENTHANDLER_DEREGISTER(vlan_unconfig, ctx->ifc_vlan_detach_event);
5857 		ctx->ifc_vlan_detach_event = NULL;
5858 	}
5859 
5860 }
5861 
5862 static void
5863 iflib_deregister(if_ctx_t ctx)
5864 {
5865 	if_t ifp = ctx->ifc_ifp;
5866 
5867 	/* Remove all media */
5868 	ifmedia_removeall(&ctx->ifc_media);
5869 
5870 	/* Ensure that VLAN event handlers are unregistered */
5871 	iflib_unregister_vlan_handlers(ctx);
5872 
5873 	/* Release kobject reference */
5874 	kobj_delete((kobj_t) ctx, NULL);
5875 
5876 	/* Free the ifnet structure */
5877 	if_free(ifp);
5878 
5879 	STATE_LOCK_DESTROY(ctx);
5880 
5881 	/* ether_ifdetach calls if_qflush - lock must be destroy afterwards*/
5882 	CTX_LOCK_DESTROY(ctx);
5883 }
5884 
5885 static int
5886 iflib_queues_alloc(if_ctx_t ctx)
5887 {
5888 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5889 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
5890 	device_t dev = ctx->ifc_dev;
5891 	int nrxqsets = scctx->isc_nrxqsets;
5892 	int ntxqsets = scctx->isc_ntxqsets;
5893 	iflib_txq_t txq;
5894 	iflib_rxq_t rxq;
5895 	iflib_fl_t fl = NULL;
5896 	int i, j, cpu, err, txconf, rxconf;
5897 	iflib_dma_info_t ifdip;
5898 	uint32_t *rxqsizes = scctx->isc_rxqsizes;
5899 	uint32_t *txqsizes = scctx->isc_txqsizes;
5900 	uint8_t nrxqs = sctx->isc_nrxqs;
5901 	uint8_t ntxqs = sctx->isc_ntxqs;
5902 	int nfree_lists = sctx->isc_nfl ? sctx->isc_nfl : 1;
5903 	int fl_offset = (sctx->isc_flags & IFLIB_HAS_RXCQ ? 1 : 0);
5904 	caddr_t *vaddrs;
5905 	uint64_t *paddrs;
5906 
5907 	KASSERT(ntxqs > 0, ("number of queues per qset must be at least 1"));
5908 	KASSERT(nrxqs > 0, ("number of queues per qset must be at least 1"));
5909 	KASSERT(nrxqs >= fl_offset + nfree_lists,
5910            ("there must be at least a rxq for each free list"));
5911 
5912 	/* Allocate the TX ring struct memory */
5913 	if (!(ctx->ifc_txqs =
5914 	    (iflib_txq_t) malloc(sizeof(struct iflib_txq) *
5915 	    ntxqsets, M_IFLIB, M_NOWAIT | M_ZERO))) {
5916 		device_printf(dev, "Unable to allocate TX ring memory\n");
5917 		err = ENOMEM;
5918 		goto fail;
5919 	}
5920 
5921 	/* Now allocate the RX */
5922 	if (!(ctx->ifc_rxqs =
5923 	    (iflib_rxq_t) malloc(sizeof(struct iflib_rxq) *
5924 	    nrxqsets, M_IFLIB, M_NOWAIT | M_ZERO))) {
5925 		device_printf(dev, "Unable to allocate RX ring memory\n");
5926 		err = ENOMEM;
5927 		goto rx_fail;
5928 	}
5929 
5930 	txq = ctx->ifc_txqs;
5931 	rxq = ctx->ifc_rxqs;
5932 
5933 	/*
5934 	 * XXX handle allocation failure
5935 	 */
5936 	for (txconf = i = 0, cpu = CPU_FIRST(); i < ntxqsets; i++, txconf++, txq++, cpu = CPU_NEXT(cpu)) {
5937 		/* Set up some basics */
5938 
5939 		if ((ifdip = malloc(sizeof(struct iflib_dma_info) * ntxqs,
5940 		    M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
5941 			device_printf(dev,
5942 			    "Unable to allocate TX DMA info memory\n");
5943 			err = ENOMEM;
5944 			goto err_tx_desc;
5945 		}
5946 		txq->ift_ifdi = ifdip;
5947 		for (j = 0; j < ntxqs; j++, ifdip++) {
5948 			if (iflib_dma_alloc(ctx, txqsizes[j], ifdip, 0)) {
5949 				device_printf(dev,
5950 				    "Unable to allocate TX descriptors\n");
5951 				err = ENOMEM;
5952 				goto err_tx_desc;
5953 			}
5954 			txq->ift_txd_size[j] = scctx->isc_txd_size[j];
5955 			bzero((void *)ifdip->idi_vaddr, txqsizes[j]);
5956 		}
5957 		txq->ift_ctx = ctx;
5958 		txq->ift_id = i;
5959 		if (sctx->isc_flags & IFLIB_HAS_TXCQ) {
5960 			txq->ift_br_offset = 1;
5961 		} else {
5962 			txq->ift_br_offset = 0;
5963 		}
5964 
5965 		if (iflib_txsd_alloc(txq)) {
5966 			device_printf(dev, "Critical Failure setting up TX buffers\n");
5967 			err = ENOMEM;
5968 			goto err_tx_desc;
5969 		}
5970 
5971 		/* Initialize the TX lock */
5972 		snprintf(txq->ift_mtx_name, MTX_NAME_LEN, "%s:TX(%d):callout",
5973 		    device_get_nameunit(dev), txq->ift_id);
5974 		mtx_init(&txq->ift_mtx, txq->ift_mtx_name, NULL, MTX_DEF);
5975 		callout_init_mtx(&txq->ift_timer, &txq->ift_mtx, 0);
5976 		txq->ift_timer.c_cpu = cpu;
5977 #ifdef DEV_NETMAP
5978 		callout_init_mtx(&txq->ift_netmap_timer, &txq->ift_mtx, 0);
5979 		txq->ift_netmap_timer.c_cpu = cpu;
5980 #endif /* DEV_NETMAP */
5981 
5982 		err = ifmp_ring_alloc(&txq->ift_br, 2048, txq, iflib_txq_drain,
5983 				      iflib_txq_can_drain, M_IFLIB, M_WAITOK);
5984 		if (err) {
5985 			/* XXX free any allocated rings */
5986 			device_printf(dev, "Unable to allocate buf_ring\n");
5987 			goto err_tx_desc;
5988 		}
5989 	}
5990 
5991 	for (rxconf = i = 0; i < nrxqsets; i++, rxconf++, rxq++) {
5992 		/* Set up some basics */
5993 		callout_init(&rxq->ifr_watchdog, 1);
5994 
5995 		if ((ifdip = malloc(sizeof(struct iflib_dma_info) * nrxqs,
5996 		   M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
5997 			device_printf(dev,
5998 			    "Unable to allocate RX DMA info memory\n");
5999 			err = ENOMEM;
6000 			goto err_tx_desc;
6001 		}
6002 
6003 		rxq->ifr_ifdi = ifdip;
6004 		/* XXX this needs to be changed if #rx queues != #tx queues */
6005 		rxq->ifr_ntxqirq = 1;
6006 		rxq->ifr_txqid[0] = i;
6007 		for (j = 0; j < nrxqs; j++, ifdip++) {
6008 			if (iflib_dma_alloc(ctx, rxqsizes[j], ifdip, 0)) {
6009 				device_printf(dev,
6010 				    "Unable to allocate RX descriptors\n");
6011 				err = ENOMEM;
6012 				goto err_tx_desc;
6013 			}
6014 			bzero((void *)ifdip->idi_vaddr, rxqsizes[j]);
6015 		}
6016 		rxq->ifr_ctx = ctx;
6017 		rxq->ifr_id = i;
6018 		rxq->ifr_fl_offset = fl_offset;
6019 		rxq->ifr_nfl = nfree_lists;
6020 		if (!(fl =
6021 			  (iflib_fl_t) malloc(sizeof(struct iflib_fl) * nfree_lists, M_IFLIB, M_NOWAIT | M_ZERO))) {
6022 			device_printf(dev, "Unable to allocate free list memory\n");
6023 			err = ENOMEM;
6024 			goto err_tx_desc;
6025 		}
6026 		rxq->ifr_fl = fl;
6027 		for (j = 0; j < nfree_lists; j++) {
6028 			fl[j].ifl_rxq = rxq;
6029 			fl[j].ifl_id = j;
6030 			fl[j].ifl_ifdi = &rxq->ifr_ifdi[j + rxq->ifr_fl_offset];
6031 			fl[j].ifl_rxd_size = scctx->isc_rxd_size[j];
6032 		}
6033 		/* Allocate receive buffers for the ring */
6034 		if (iflib_rxsd_alloc(rxq)) {
6035 			device_printf(dev,
6036 			    "Critical Failure setting up receive buffers\n");
6037 			err = ENOMEM;
6038 			goto err_rx_desc;
6039 		}
6040 
6041 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++)
6042 			fl->ifl_rx_bitmap = bit_alloc(fl->ifl_size, M_IFLIB,
6043 			    M_WAITOK);
6044 	}
6045 
6046 	/* TXQs */
6047 	vaddrs = malloc(sizeof(caddr_t)*ntxqsets*ntxqs, M_IFLIB, M_WAITOK);
6048 	paddrs = malloc(sizeof(uint64_t)*ntxqsets*ntxqs, M_IFLIB, M_WAITOK);
6049 	for (i = 0; i < ntxqsets; i++) {
6050 		iflib_dma_info_t di = ctx->ifc_txqs[i].ift_ifdi;
6051 
6052 		for (j = 0; j < ntxqs; j++, di++) {
6053 			vaddrs[i*ntxqs + j] = di->idi_vaddr;
6054 			paddrs[i*ntxqs + j] = di->idi_paddr;
6055 		}
6056 	}
6057 	if ((err = IFDI_TX_QUEUES_ALLOC(ctx, vaddrs, paddrs, ntxqs, ntxqsets)) != 0) {
6058 		device_printf(ctx->ifc_dev,
6059 		    "Unable to allocate device TX queue\n");
6060 		iflib_tx_structures_free(ctx);
6061 		free(vaddrs, M_IFLIB);
6062 		free(paddrs, M_IFLIB);
6063 		goto err_rx_desc;
6064 	}
6065 	free(vaddrs, M_IFLIB);
6066 	free(paddrs, M_IFLIB);
6067 
6068 	/* RXQs */
6069 	vaddrs = malloc(sizeof(caddr_t)*nrxqsets*nrxqs, M_IFLIB, M_WAITOK);
6070 	paddrs = malloc(sizeof(uint64_t)*nrxqsets*nrxqs, M_IFLIB, M_WAITOK);
6071 	for (i = 0; i < nrxqsets; i++) {
6072 		iflib_dma_info_t di = ctx->ifc_rxqs[i].ifr_ifdi;
6073 
6074 		for (j = 0; j < nrxqs; j++, di++) {
6075 			vaddrs[i*nrxqs + j] = di->idi_vaddr;
6076 			paddrs[i*nrxqs + j] = di->idi_paddr;
6077 		}
6078 	}
6079 	if ((err = IFDI_RX_QUEUES_ALLOC(ctx, vaddrs, paddrs, nrxqs, nrxqsets)) != 0) {
6080 		device_printf(ctx->ifc_dev,
6081 		    "Unable to allocate device RX queue\n");
6082 		iflib_tx_structures_free(ctx);
6083 		free(vaddrs, M_IFLIB);
6084 		free(paddrs, M_IFLIB);
6085 		goto err_rx_desc;
6086 	}
6087 	free(vaddrs, M_IFLIB);
6088 	free(paddrs, M_IFLIB);
6089 
6090 	return (0);
6091 
6092 /* XXX handle allocation failure changes */
6093 err_rx_desc:
6094 err_tx_desc:
6095 rx_fail:
6096 	if (ctx->ifc_rxqs != NULL)
6097 		free(ctx->ifc_rxqs, M_IFLIB);
6098 	ctx->ifc_rxqs = NULL;
6099 	if (ctx->ifc_txqs != NULL)
6100 		free(ctx->ifc_txqs, M_IFLIB);
6101 	ctx->ifc_txqs = NULL;
6102 fail:
6103 	return (err);
6104 }
6105 
6106 static int
6107 iflib_tx_structures_setup(if_ctx_t ctx)
6108 {
6109 	iflib_txq_t txq = ctx->ifc_txqs;
6110 	int i;
6111 
6112 	for (i = 0; i < NTXQSETS(ctx); i++, txq++)
6113 		iflib_txq_setup(txq);
6114 
6115 	return (0);
6116 }
6117 
6118 static void
6119 iflib_tx_structures_free(if_ctx_t ctx)
6120 {
6121 	iflib_txq_t txq = ctx->ifc_txqs;
6122 	if_shared_ctx_t sctx = ctx->ifc_sctx;
6123 	int i, j;
6124 
6125 	for (i = 0; i < NTXQSETS(ctx); i++, txq++) {
6126 		for (j = 0; j < sctx->isc_ntxqs; j++)
6127 			iflib_dma_free(&txq->ift_ifdi[j]);
6128 		iflib_txq_destroy(txq);
6129 	}
6130 	free(ctx->ifc_txqs, M_IFLIB);
6131 	ctx->ifc_txqs = NULL;
6132 }
6133 
6134 /*********************************************************************
6135  *
6136  *  Initialize all receive rings.
6137  *
6138  **********************************************************************/
6139 static int
6140 iflib_rx_structures_setup(if_ctx_t ctx)
6141 {
6142 	iflib_rxq_t rxq = ctx->ifc_rxqs;
6143 	int q;
6144 #if defined(INET6) || defined(INET)
6145 	int err, i;
6146 #endif
6147 
6148 	for (q = 0; q < ctx->ifc_softc_ctx.isc_nrxqsets; q++, rxq++) {
6149 #if defined(INET6) || defined(INET)
6150 		err = tcp_lro_init_args(&rxq->ifr_lc, ctx->ifc_ifp,
6151 		    TCP_LRO_ENTRIES, min(1024,
6152 		    ctx->ifc_softc_ctx.isc_nrxd[rxq->ifr_fl_offset]));
6153 		if (err != 0) {
6154 			device_printf(ctx->ifc_dev,
6155 			    "LRO Initialization failed!\n");
6156 			goto fail;
6157 		}
6158 #endif
6159 		IFDI_RXQ_SETUP(ctx, rxq->ifr_id);
6160 	}
6161 	return (0);
6162 #if defined(INET6) || defined(INET)
6163 fail:
6164 	/*
6165 	 * Free LRO resources allocated so far, we will only handle
6166 	 * the rings that completed, the failing case will have
6167 	 * cleaned up for itself.  'q' failed, so its the terminus.
6168 	 */
6169 	rxq = ctx->ifc_rxqs;
6170 	for (i = 0; i < q; ++i, rxq++) {
6171 		tcp_lro_free(&rxq->ifr_lc);
6172 	}
6173 	return (err);
6174 #endif
6175 }
6176 
6177 /*********************************************************************
6178  *
6179  *  Free all receive rings.
6180  *
6181  **********************************************************************/
6182 static void
6183 iflib_rx_structures_free(if_ctx_t ctx)
6184 {
6185 	iflib_rxq_t rxq = ctx->ifc_rxqs;
6186 	if_shared_ctx_t sctx = ctx->ifc_sctx;
6187 	int i, j;
6188 
6189 	for (i = 0; i < ctx->ifc_softc_ctx.isc_nrxqsets; i++, rxq++) {
6190 		for (j = 0; j < sctx->isc_nrxqs; j++)
6191 			iflib_dma_free(&rxq->ifr_ifdi[j]);
6192 		iflib_rx_sds_free(rxq);
6193 #if defined(INET6) || defined(INET)
6194 		tcp_lro_free(&rxq->ifr_lc);
6195 #endif
6196 	}
6197 	free(ctx->ifc_rxqs, M_IFLIB);
6198 	ctx->ifc_rxqs = NULL;
6199 }
6200 
6201 static int
6202 iflib_qset_structures_setup(if_ctx_t ctx)
6203 {
6204 	int err;
6205 
6206 	/*
6207 	 * It is expected that the caller takes care of freeing queues if this
6208 	 * fails.
6209 	 */
6210 	if ((err = iflib_tx_structures_setup(ctx)) != 0) {
6211 		device_printf(ctx->ifc_dev, "iflib_tx_structures_setup failed: %d\n", err);
6212 		return (err);
6213 	}
6214 
6215 	if ((err = iflib_rx_structures_setup(ctx)) != 0)
6216 		device_printf(ctx->ifc_dev, "iflib_rx_structures_setup failed: %d\n", err);
6217 
6218 	return (err);
6219 }
6220 
6221 int
6222 iflib_irq_alloc(if_ctx_t ctx, if_irq_t irq, int rid,
6223 		driver_filter_t filter, void *filter_arg, driver_intr_t handler, void *arg, const char *name)
6224 {
6225 
6226 	return (_iflib_irq_alloc(ctx, irq, rid, filter, handler, arg, name));
6227 }
6228 
6229 /* Just to avoid copy/paste */
6230 static inline int
6231 iflib_irq_set_affinity(if_ctx_t ctx, if_irq_t irq, iflib_intr_type_t type,
6232     int qid, struct grouptask *gtask, struct taskqgroup *tqg, void *uniq,
6233     const char *name)
6234 {
6235 	device_t dev;
6236 	unsigned int base_cpuid, cpuid;
6237 	int err;
6238 
6239 	dev = ctx->ifc_dev;
6240 	base_cpuid = ctx->ifc_sysctl_core_offset;
6241 	cpuid = get_cpuid_for_queue(ctx, base_cpuid, qid, type == IFLIB_INTR_TX);
6242 	err = taskqgroup_attach_cpu(tqg, gtask, uniq, cpuid, dev,
6243 	    irq ? irq->ii_res : NULL, name);
6244 	if (err) {
6245 		device_printf(dev, "taskqgroup_attach_cpu failed %d\n", err);
6246 		return (err);
6247 	}
6248 #ifdef notyet
6249 	if (cpuid > ctx->ifc_cpuid_highest)
6250 		ctx->ifc_cpuid_highest = cpuid;
6251 #endif
6252 	return (0);
6253 }
6254 
6255 int
6256 iflib_irq_alloc_generic(if_ctx_t ctx, if_irq_t irq, int rid,
6257 			iflib_intr_type_t type, driver_filter_t *filter,
6258 			void *filter_arg, int qid, const char *name)
6259 {
6260 	device_t dev;
6261 	struct grouptask *gtask;
6262 	struct taskqgroup *tqg;
6263 	iflib_filter_info_t info;
6264 	gtask_fn_t *fn;
6265 	int tqrid, err;
6266 	driver_filter_t *intr_fast;
6267 	void *q;
6268 
6269 	info = &ctx->ifc_filter_info;
6270 	tqrid = rid;
6271 
6272 	switch (type) {
6273 	/* XXX merge tx/rx for netmap? */
6274 	case IFLIB_INTR_TX:
6275 		q = &ctx->ifc_txqs[qid];
6276 		info = &ctx->ifc_txqs[qid].ift_filter_info;
6277 		gtask = &ctx->ifc_txqs[qid].ift_task;
6278 		tqg = qgroup_if_io_tqg;
6279 		fn = _task_fn_tx;
6280 		intr_fast = iflib_fast_intr;
6281 		GROUPTASK_INIT(gtask, 0, fn, q);
6282 		ctx->ifc_flags |= IFC_NETMAP_TX_IRQ;
6283 		break;
6284 	case IFLIB_INTR_RX:
6285 		q = &ctx->ifc_rxqs[qid];
6286 		info = &ctx->ifc_rxqs[qid].ifr_filter_info;
6287 		gtask = &ctx->ifc_rxqs[qid].ifr_task;
6288 		tqg = qgroup_if_io_tqg;
6289 		fn = _task_fn_rx;
6290 		intr_fast = iflib_fast_intr;
6291 		NET_GROUPTASK_INIT(gtask, 0, fn, q);
6292 		break;
6293 	case IFLIB_INTR_RXTX:
6294 		q = &ctx->ifc_rxqs[qid];
6295 		info = &ctx->ifc_rxqs[qid].ifr_filter_info;
6296 		gtask = &ctx->ifc_rxqs[qid].ifr_task;
6297 		tqg = qgroup_if_io_tqg;
6298 		fn = _task_fn_rx;
6299 		intr_fast = iflib_fast_intr_rxtx;
6300 		NET_GROUPTASK_INIT(gtask, 0, fn, q);
6301 		break;
6302 	case IFLIB_INTR_ADMIN:
6303 		q = ctx;
6304 		tqrid = -1;
6305 		info = &ctx->ifc_filter_info;
6306 		gtask = &ctx->ifc_admin_task;
6307 		tqg = qgroup_if_config_tqg;
6308 		fn = _task_fn_admin;
6309 		intr_fast = iflib_fast_intr_ctx;
6310 		break;
6311 	default:
6312 		device_printf(ctx->ifc_dev, "%s: unknown net intr type\n",
6313 		    __func__);
6314 		return (EINVAL);
6315 	}
6316 
6317 	info->ifi_filter = filter;
6318 	info->ifi_filter_arg = filter_arg;
6319 	info->ifi_task = gtask;
6320 	info->ifi_ctx = q;
6321 
6322 	dev = ctx->ifc_dev;
6323 	err = _iflib_irq_alloc(ctx, irq, rid, intr_fast, NULL, info,  name);
6324 	if (err != 0) {
6325 		device_printf(dev, "_iflib_irq_alloc failed %d\n", err);
6326 		return (err);
6327 	}
6328 	if (type == IFLIB_INTR_ADMIN)
6329 		return (0);
6330 
6331 	if (tqrid != -1) {
6332 		err = iflib_irq_set_affinity(ctx, irq, type, qid, gtask, tqg, q,
6333 		    name);
6334 		if (err)
6335 			return (err);
6336 	} else {
6337 		taskqgroup_attach(tqg, gtask, q, dev, irq->ii_res, name);
6338 	}
6339 
6340 	return (0);
6341 }
6342 
6343 void
6344 iflib_softirq_alloc_generic(if_ctx_t ctx, if_irq_t irq, iflib_intr_type_t type, void *arg, int qid, const char *name)
6345 {
6346 	device_t dev;
6347 	struct grouptask *gtask;
6348 	struct taskqgroup *tqg;
6349 	gtask_fn_t *fn;
6350 	void *q;
6351 	int err;
6352 
6353 	switch (type) {
6354 	case IFLIB_INTR_TX:
6355 		q = &ctx->ifc_txqs[qid];
6356 		gtask = &ctx->ifc_txqs[qid].ift_task;
6357 		tqg = qgroup_if_io_tqg;
6358 		fn = _task_fn_tx;
6359 		GROUPTASK_INIT(gtask, 0, fn, q);
6360 		break;
6361 	case IFLIB_INTR_RX:
6362 		q = &ctx->ifc_rxqs[qid];
6363 		gtask = &ctx->ifc_rxqs[qid].ifr_task;
6364 		tqg = qgroup_if_io_tqg;
6365 		fn = _task_fn_rx;
6366 		NET_GROUPTASK_INIT(gtask, 0, fn, q);
6367 		break;
6368 	case IFLIB_INTR_IOV:
6369 		q = ctx;
6370 		gtask = &ctx->ifc_vflr_task;
6371 		tqg = qgroup_if_config_tqg;
6372 		fn = _task_fn_iov;
6373 		GROUPTASK_INIT(gtask, 0, fn, q);
6374 		break;
6375 	default:
6376 		panic("unknown net intr type");
6377 	}
6378 	err = iflib_irq_set_affinity(ctx, irq, type, qid, gtask, tqg, q, name);
6379 	if (err) {
6380 		dev = ctx->ifc_dev;
6381 		taskqgroup_attach(tqg, gtask, q, dev, irq ? irq->ii_res : NULL,
6382 		    name);
6383 	}
6384 }
6385 
6386 void
6387 iflib_irq_free(if_ctx_t ctx, if_irq_t irq)
6388 {
6389 
6390 	if (irq->ii_tag)
6391 		bus_teardown_intr(ctx->ifc_dev, irq->ii_res, irq->ii_tag);
6392 
6393 	if (irq->ii_res)
6394 		bus_release_resource(ctx->ifc_dev, SYS_RES_IRQ,
6395 		    rman_get_rid(irq->ii_res), irq->ii_res);
6396 }
6397 
6398 static int
6399 iflib_legacy_setup(if_ctx_t ctx, driver_filter_t filter, void *filter_arg, int *rid, const char *name)
6400 {
6401 	iflib_txq_t txq = ctx->ifc_txqs;
6402 	iflib_rxq_t rxq = ctx->ifc_rxqs;
6403 	if_irq_t irq = &ctx->ifc_legacy_irq;
6404 	iflib_filter_info_t info;
6405 	device_t dev;
6406 	struct grouptask *gtask;
6407 	struct resource *res;
6408 	struct taskqgroup *tqg;
6409 	void *q;
6410 	int err, tqrid;
6411 	bool rx_only;
6412 
6413 	q = &ctx->ifc_rxqs[0];
6414 	info = &rxq[0].ifr_filter_info;
6415 	gtask = &rxq[0].ifr_task;
6416 	tqg = qgroup_if_io_tqg;
6417 	tqrid = *rid;
6418 	rx_only = (ctx->ifc_sctx->isc_flags & IFLIB_SINGLE_IRQ_RX_ONLY) != 0;
6419 
6420 	ctx->ifc_flags |= IFC_LEGACY;
6421 	info->ifi_filter = filter;
6422 	info->ifi_filter_arg = filter_arg;
6423 	info->ifi_task = gtask;
6424 	info->ifi_ctx = rx_only ? ctx : q;
6425 
6426 	dev = ctx->ifc_dev;
6427 	/* We allocate a single interrupt resource */
6428 	err = _iflib_irq_alloc(ctx, irq, tqrid, rx_only ? iflib_fast_intr_ctx :
6429 	    iflib_fast_intr_rxtx, NULL, info, name);
6430 	if (err != 0)
6431 		return (err);
6432 	NET_GROUPTASK_INIT(gtask, 0, _task_fn_rx, q);
6433 	res = irq->ii_res;
6434 	taskqgroup_attach(tqg, gtask, q, dev, res, name);
6435 
6436 	GROUPTASK_INIT(&txq->ift_task, 0, _task_fn_tx, txq);
6437 	taskqgroup_attach(qgroup_if_io_tqg, &txq->ift_task, txq, dev, res,
6438 	    "tx");
6439 	return (0);
6440 }
6441 
6442 void
6443 iflib_led_create(if_ctx_t ctx)
6444 {
6445 
6446 	ctx->ifc_led_dev = led_create(iflib_led_func, ctx,
6447 	    device_get_nameunit(ctx->ifc_dev));
6448 }
6449 
6450 void
6451 iflib_tx_intr_deferred(if_ctx_t ctx, int txqid)
6452 {
6453 
6454 	GROUPTASK_ENQUEUE(&ctx->ifc_txqs[txqid].ift_task);
6455 }
6456 
6457 void
6458 iflib_rx_intr_deferred(if_ctx_t ctx, int rxqid)
6459 {
6460 
6461 	GROUPTASK_ENQUEUE(&ctx->ifc_rxqs[rxqid].ifr_task);
6462 }
6463 
6464 void
6465 iflib_admin_intr_deferred(if_ctx_t ctx)
6466 {
6467 
6468 	MPASS(ctx->ifc_admin_task.gt_taskqueue != NULL);
6469 	GROUPTASK_ENQUEUE(&ctx->ifc_admin_task);
6470 }
6471 
6472 void
6473 iflib_iov_intr_deferred(if_ctx_t ctx)
6474 {
6475 
6476 	GROUPTASK_ENQUEUE(&ctx->ifc_vflr_task);
6477 }
6478 
6479 void
6480 iflib_io_tqg_attach(struct grouptask *gt, void *uniq, int cpu, const char *name)
6481 {
6482 
6483 	taskqgroup_attach_cpu(qgroup_if_io_tqg, gt, uniq, cpu, NULL, NULL,
6484 	    name);
6485 }
6486 
6487 void
6488 iflib_config_gtask_init(void *ctx, struct grouptask *gtask, gtask_fn_t *fn,
6489 	const char *name)
6490 {
6491 
6492 	GROUPTASK_INIT(gtask, 0, fn, ctx);
6493 	taskqgroup_attach(qgroup_if_config_tqg, gtask, gtask, NULL, NULL,
6494 	    name);
6495 }
6496 
6497 void
6498 iflib_config_gtask_deinit(struct grouptask *gtask)
6499 {
6500 
6501 	taskqgroup_detach(qgroup_if_config_tqg, gtask);
6502 }
6503 
6504 void
6505 iflib_link_state_change(if_ctx_t ctx, int link_state, uint64_t baudrate)
6506 {
6507 	if_t ifp = ctx->ifc_ifp;
6508 	iflib_txq_t txq = ctx->ifc_txqs;
6509 
6510 	if_setbaudrate(ifp, baudrate);
6511 	if (baudrate >= IF_Gbps(10)) {
6512 		STATE_LOCK(ctx);
6513 		ctx->ifc_flags |= IFC_PREFETCH;
6514 		STATE_UNLOCK(ctx);
6515 	}
6516 	/* If link down, disable watchdog */
6517 	if ((ctx->ifc_link_state == LINK_STATE_UP) && (link_state == LINK_STATE_DOWN)) {
6518 		for (int i = 0; i < ctx->ifc_softc_ctx.isc_ntxqsets; i++, txq++)
6519 			txq->ift_qstatus = IFLIB_QUEUE_IDLE;
6520 	}
6521 	ctx->ifc_link_state = link_state;
6522 	if_link_state_change(ifp, link_state);
6523 }
6524 
6525 static int
6526 iflib_tx_credits_update(if_ctx_t ctx, iflib_txq_t txq)
6527 {
6528 	int credits;
6529 #ifdef INVARIANTS
6530 	int credits_pre = txq->ift_cidx_processed;
6531 #endif
6532 
6533 	bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
6534 	    BUS_DMASYNC_POSTREAD);
6535 	if ((credits = ctx->isc_txd_credits_update(ctx->ifc_softc, txq->ift_id, true)) == 0)
6536 		return (0);
6537 
6538 	txq->ift_processed += credits;
6539 	txq->ift_cidx_processed += credits;
6540 
6541 	MPASS(credits_pre + credits == txq->ift_cidx_processed);
6542 	if (txq->ift_cidx_processed >= txq->ift_size)
6543 		txq->ift_cidx_processed -= txq->ift_size;
6544 	return (credits);
6545 }
6546 
6547 static int
6548 iflib_rxd_avail(if_ctx_t ctx, iflib_rxq_t rxq, qidx_t cidx, qidx_t budget)
6549 {
6550 	iflib_fl_t fl;
6551 	u_int i;
6552 
6553 	for (i = 0, fl = &rxq->ifr_fl[0]; i < rxq->ifr_nfl; i++, fl++)
6554 		bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
6555 		    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
6556 	return (ctx->isc_rxd_available(ctx->ifc_softc, rxq->ifr_id, cidx,
6557 	    budget));
6558 }
6559 
6560 void
6561 iflib_add_int_delay_sysctl(if_ctx_t ctx, const char *name,
6562 	const char *description, if_int_delay_info_t info,
6563 	int offset, int value)
6564 {
6565 	info->iidi_ctx = ctx;
6566 	info->iidi_offset = offset;
6567 	info->iidi_value = value;
6568 	SYSCTL_ADD_PROC(device_get_sysctl_ctx(ctx->ifc_dev),
6569 	    SYSCTL_CHILDREN(device_get_sysctl_tree(ctx->ifc_dev)),
6570 	    OID_AUTO, name, CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_MPSAFE,
6571 	    info, 0, iflib_sysctl_int_delay, "I", description);
6572 }
6573 
6574 struct sx *
6575 iflib_ctx_lock_get(if_ctx_t ctx)
6576 {
6577 
6578 	return (&ctx->ifc_ctx_sx);
6579 }
6580 
6581 static int
6582 iflib_msix_init(if_ctx_t ctx)
6583 {
6584 	device_t dev = ctx->ifc_dev;
6585 	if_shared_ctx_t sctx = ctx->ifc_sctx;
6586 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
6587 	int admincnt, bar, err, iflib_num_rx_queues, iflib_num_tx_queues;
6588 	int msgs, queuemsgs, queues, rx_queues, tx_queues, vectors;
6589 
6590 	iflib_num_tx_queues = ctx->ifc_sysctl_ntxqs;
6591 	iflib_num_rx_queues = ctx->ifc_sysctl_nrxqs;
6592 
6593 	if (bootverbose)
6594 		device_printf(dev, "msix_init qsets capped at %d\n",
6595 		    imax(scctx->isc_ntxqsets, scctx->isc_nrxqsets));
6596 
6597 	/* Override by tuneable */
6598 	if (scctx->isc_disable_msix)
6599 		goto msi;
6600 
6601 	/* First try MSI-X */
6602 	if ((msgs = pci_msix_count(dev)) == 0) {
6603 		if (bootverbose)
6604 			device_printf(dev, "MSI-X not supported or disabled\n");
6605 		goto msi;
6606 	}
6607 
6608 	bar = ctx->ifc_softc_ctx.isc_msix_bar;
6609 	/*
6610 	 * bar == -1 => "trust me I know what I'm doing"
6611 	 * Some drivers are for hardware that is so shoddily
6612 	 * documented that no one knows which bars are which
6613 	 * so the developer has to map all bars. This hack
6614 	 * allows shoddy garbage to use MSI-X in this framework.
6615 	 */
6616 	if (bar != -1) {
6617 		ctx->ifc_msix_mem = bus_alloc_resource_any(dev,
6618 	            SYS_RES_MEMORY, &bar, RF_ACTIVE);
6619 		if (ctx->ifc_msix_mem == NULL) {
6620 			device_printf(dev, "Unable to map MSI-X table\n");
6621 			goto msi;
6622 		}
6623 	}
6624 
6625 	admincnt = sctx->isc_admin_intrcnt;
6626 #if IFLIB_DEBUG
6627 	/* use only 1 qset in debug mode */
6628 	queuemsgs = min(msgs - admincnt, 1);
6629 #else
6630 	queuemsgs = msgs - admincnt;
6631 #endif
6632 #ifdef RSS
6633 	queues = imin(queuemsgs, rss_getnumbuckets());
6634 #else
6635 	queues = queuemsgs;
6636 #endif
6637 	queues = imin(CPU_COUNT(&ctx->ifc_cpus), queues);
6638 	if (bootverbose)
6639 		device_printf(dev,
6640 		    "intr CPUs: %d queue msgs: %d admincnt: %d\n",
6641 		    CPU_COUNT(&ctx->ifc_cpus), queuemsgs, admincnt);
6642 #ifdef  RSS
6643 	/* If we're doing RSS, clamp at the number of RSS buckets */
6644 	if (queues > rss_getnumbuckets())
6645 		queues = rss_getnumbuckets();
6646 #endif
6647 	if (iflib_num_rx_queues > 0 && iflib_num_rx_queues < queuemsgs - admincnt)
6648 		rx_queues = iflib_num_rx_queues;
6649 	else
6650 		rx_queues = queues;
6651 
6652 	if (rx_queues > scctx->isc_nrxqsets)
6653 		rx_queues = scctx->isc_nrxqsets;
6654 
6655 	/*
6656 	 * We want this to be all logical CPUs by default
6657 	 */
6658 	if (iflib_num_tx_queues > 0 && iflib_num_tx_queues < queues)
6659 		tx_queues = iflib_num_tx_queues;
6660 	else
6661 		tx_queues = mp_ncpus;
6662 
6663 	if (tx_queues > scctx->isc_ntxqsets)
6664 		tx_queues = scctx->isc_ntxqsets;
6665 
6666 	if (ctx->ifc_sysctl_qs_eq_override == 0) {
6667 #ifdef INVARIANTS
6668 		if (tx_queues != rx_queues)
6669 			device_printf(dev,
6670 			    "queue equality override not set, capping rx_queues at %d and tx_queues at %d\n",
6671 			    min(rx_queues, tx_queues), min(rx_queues, tx_queues));
6672 #endif
6673 		tx_queues = min(rx_queues, tx_queues);
6674 		rx_queues = min(rx_queues, tx_queues);
6675 	}
6676 
6677 	vectors = rx_queues + admincnt;
6678 	if (msgs < vectors) {
6679 		device_printf(dev,
6680 		    "insufficient number of MSI-X vectors "
6681 		    "(supported %d, need %d)\n", msgs, vectors);
6682 		goto msi;
6683 	}
6684 
6685 	device_printf(dev, "Using %d RX queues %d TX queues\n", rx_queues,
6686 	    tx_queues);
6687 	msgs = vectors;
6688 	if ((err = pci_alloc_msix(dev, &vectors)) == 0) {
6689 		if (vectors != msgs) {
6690 			device_printf(dev,
6691 			    "Unable to allocate sufficient MSI-X vectors "
6692 			    "(got %d, need %d)\n", vectors, msgs);
6693 			pci_release_msi(dev);
6694 			if (bar != -1) {
6695 				bus_release_resource(dev, SYS_RES_MEMORY, bar,
6696 				    ctx->ifc_msix_mem);
6697 				ctx->ifc_msix_mem = NULL;
6698 			}
6699 			goto msi;
6700 		}
6701 		device_printf(dev, "Using MSI-X interrupts with %d vectors\n",
6702 		    vectors);
6703 		scctx->isc_vectors = vectors;
6704 		scctx->isc_nrxqsets = rx_queues;
6705 		scctx->isc_ntxqsets = tx_queues;
6706 		scctx->isc_intr = IFLIB_INTR_MSIX;
6707 
6708 		return (vectors);
6709 	} else {
6710 		device_printf(dev,
6711 		    "failed to allocate %d MSI-X vectors, err: %d\n", vectors,
6712 		    err);
6713 		if (bar != -1) {
6714 			bus_release_resource(dev, SYS_RES_MEMORY, bar,
6715 			    ctx->ifc_msix_mem);
6716 			ctx->ifc_msix_mem = NULL;
6717 		}
6718 	}
6719 
6720 msi:
6721 	vectors = pci_msi_count(dev);
6722 	scctx->isc_nrxqsets = 1;
6723 	scctx->isc_ntxqsets = 1;
6724 	scctx->isc_vectors = vectors;
6725 	if (vectors == 1 && pci_alloc_msi(dev, &vectors) == 0) {
6726 		device_printf(dev,"Using an MSI interrupt\n");
6727 		scctx->isc_intr = IFLIB_INTR_MSI;
6728 	} else {
6729 		scctx->isc_vectors = 1;
6730 		device_printf(dev,"Using a Legacy interrupt\n");
6731 		scctx->isc_intr = IFLIB_INTR_LEGACY;
6732 	}
6733 
6734 	return (vectors);
6735 }
6736 
6737 static const char *ring_states[] = { "IDLE", "BUSY", "STALLED", "ABDICATED" };
6738 
6739 static int
6740 mp_ring_state_handler(SYSCTL_HANDLER_ARGS)
6741 {
6742 	int rc;
6743 	uint16_t *state = ((uint16_t *)oidp->oid_arg1);
6744 	struct sbuf *sb;
6745 	const char *ring_state = "UNKNOWN";
6746 
6747 	/* XXX needed ? */
6748 	rc = sysctl_wire_old_buffer(req, 0);
6749 	MPASS(rc == 0);
6750 	if (rc != 0)
6751 		return (rc);
6752 	sb = sbuf_new_for_sysctl(NULL, NULL, 80, req);
6753 	MPASS(sb != NULL);
6754 	if (sb == NULL)
6755 		return (ENOMEM);
6756 	if (state[3] <= 3)
6757 		ring_state = ring_states[state[3]];
6758 
6759 	sbuf_printf(sb, "pidx_head: %04hd pidx_tail: %04hd cidx: %04hd state: %s",
6760 		    state[0], state[1], state[2], ring_state);
6761 	rc = sbuf_finish(sb);
6762 	sbuf_delete(sb);
6763         return(rc);
6764 }
6765 
6766 enum iflib_ndesc_handler {
6767 	IFLIB_NTXD_HANDLER,
6768 	IFLIB_NRXD_HANDLER,
6769 };
6770 
6771 static int
6772 mp_ndesc_handler(SYSCTL_HANDLER_ARGS)
6773 {
6774 	if_ctx_t ctx = (void *)arg1;
6775 	enum iflib_ndesc_handler type = arg2;
6776 	char buf[256] = {0};
6777 	qidx_t *ndesc;
6778 	char *p, *next;
6779 	int nqs, rc, i;
6780 
6781 	nqs = 8;
6782 	switch(type) {
6783 	case IFLIB_NTXD_HANDLER:
6784 		ndesc = ctx->ifc_sysctl_ntxds;
6785 		if (ctx->ifc_sctx)
6786 			nqs = ctx->ifc_sctx->isc_ntxqs;
6787 		break;
6788 	case IFLIB_NRXD_HANDLER:
6789 		ndesc = ctx->ifc_sysctl_nrxds;
6790 		if (ctx->ifc_sctx)
6791 			nqs = ctx->ifc_sctx->isc_nrxqs;
6792 		break;
6793 	default:
6794 		printf("%s: unhandled type\n", __func__);
6795 		return (EINVAL);
6796 	}
6797 	if (nqs == 0)
6798 		nqs = 8;
6799 
6800 	for (i=0; i<8; i++) {
6801 		if (i >= nqs)
6802 			break;
6803 		if (i)
6804 			strcat(buf, ",");
6805 		sprintf(strchr(buf, 0), "%d", ndesc[i]);
6806 	}
6807 
6808 	rc = sysctl_handle_string(oidp, buf, sizeof(buf), req);
6809 	if (rc || req->newptr == NULL)
6810 		return rc;
6811 
6812 	for (i = 0, next = buf, p = strsep(&next, " ,"); i < 8 && p;
6813 	    i++, p = strsep(&next, " ,")) {
6814 		ndesc[i] = strtoul(p, NULL, 10);
6815 	}
6816 
6817 	return(rc);
6818 }
6819 
6820 #define NAME_BUFLEN 32
6821 static void
6822 iflib_add_device_sysctl_pre(if_ctx_t ctx)
6823 {
6824         device_t dev = iflib_get_dev(ctx);
6825 	struct sysctl_oid_list *child, *oid_list;
6826 	struct sysctl_ctx_list *ctx_list;
6827 	struct sysctl_oid *node;
6828 
6829 	ctx_list = device_get_sysctl_ctx(dev);
6830 	child = SYSCTL_CHILDREN(device_get_sysctl_tree(dev));
6831 	ctx->ifc_sysctl_node = node = SYSCTL_ADD_NODE(ctx_list, child, OID_AUTO, "iflib",
6832 	    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "IFLIB fields");
6833 	oid_list = SYSCTL_CHILDREN(node);
6834 
6835 	SYSCTL_ADD_CONST_STRING(ctx_list, oid_list, OID_AUTO, "driver_version",
6836 		       CTLFLAG_RD, ctx->ifc_sctx->isc_driver_version,
6837 		       "driver version");
6838 
6839 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "override_ntxqs",
6840 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_ntxqs, 0,
6841 			"# of txqs to use, 0 => use default #");
6842 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "override_nrxqs",
6843 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_nrxqs, 0,
6844 			"# of rxqs to use, 0 => use default #");
6845 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "override_qs_enable",
6846 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_qs_eq_override, 0,
6847                        "permit #txq != #rxq");
6848 	SYSCTL_ADD_INT(ctx_list, oid_list, OID_AUTO, "disable_msix",
6849                       CTLFLAG_RWTUN, &ctx->ifc_softc_ctx.isc_disable_msix, 0,
6850                       "disable MSI-X (default 0)");
6851 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "rx_budget",
6852 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_rx_budget, 0,
6853 		       "set the RX budget");
6854 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "tx_abdicate",
6855 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_tx_abdicate, 0,
6856 		       "cause TX to abdicate instead of running to completion");
6857 	ctx->ifc_sysctl_core_offset = CORE_OFFSET_UNSPECIFIED;
6858 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "core_offset",
6859 		       CTLFLAG_RDTUN, &ctx->ifc_sysctl_core_offset, 0,
6860 		       "offset to start using cores at");
6861 	SYSCTL_ADD_U8(ctx_list, oid_list, OID_AUTO, "separate_txrx",
6862 		       CTLFLAG_RDTUN, &ctx->ifc_sysctl_separate_txrx, 0,
6863 		       "use separate cores for TX and RX");
6864 	SYSCTL_ADD_U8(ctx_list, oid_list, OID_AUTO, "use_logical_cores",
6865 		      CTLFLAG_RDTUN, &ctx->ifc_sysctl_use_logical_cores, 0,
6866 		      "try to make use of logical cores for TX and RX");
6867 
6868 	/* XXX change for per-queue sizes */
6869 	SYSCTL_ADD_PROC(ctx_list, oid_list, OID_AUTO, "override_ntxds",
6870 	    CTLTYPE_STRING | CTLFLAG_RWTUN | CTLFLAG_NEEDGIANT, ctx,
6871 	    IFLIB_NTXD_HANDLER, mp_ndesc_handler, "A",
6872 	    "list of # of TX descriptors to use, 0 = use default #");
6873 	SYSCTL_ADD_PROC(ctx_list, oid_list, OID_AUTO, "override_nrxds",
6874 	    CTLTYPE_STRING | CTLFLAG_RWTUN | CTLFLAG_NEEDGIANT, ctx,
6875 	    IFLIB_NRXD_HANDLER, mp_ndesc_handler, "A",
6876 	    "list of # of RX descriptors to use, 0 = use default #");
6877 }
6878 
6879 static void
6880 iflib_add_device_sysctl_post(if_ctx_t ctx)
6881 {
6882 	if_shared_ctx_t sctx = ctx->ifc_sctx;
6883 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
6884         device_t dev = iflib_get_dev(ctx);
6885 	struct sysctl_oid_list *child;
6886 	struct sysctl_ctx_list *ctx_list;
6887 	iflib_fl_t fl;
6888 	iflib_txq_t txq;
6889 	iflib_rxq_t rxq;
6890 	int i, j;
6891 	char namebuf[NAME_BUFLEN];
6892 	char *qfmt;
6893 	struct sysctl_oid *queue_node, *fl_node, *node;
6894 	struct sysctl_oid_list *queue_list, *fl_list;
6895 	ctx_list = device_get_sysctl_ctx(dev);
6896 
6897 	node = ctx->ifc_sysctl_node;
6898 	child = SYSCTL_CHILDREN(node);
6899 
6900 	if (scctx->isc_ntxqsets > 100)
6901 		qfmt = "txq%03d";
6902 	else if (scctx->isc_ntxqsets > 10)
6903 		qfmt = "txq%02d";
6904 	else
6905 		qfmt = "txq%d";
6906 	for (i = 0, txq = ctx->ifc_txqs; i < scctx->isc_ntxqsets; i++, txq++) {
6907 		snprintf(namebuf, NAME_BUFLEN, qfmt, i);
6908 		queue_node = SYSCTL_ADD_NODE(ctx_list, child, OID_AUTO, namebuf,
6909 		    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "Queue Name");
6910 		queue_list = SYSCTL_CHILDREN(queue_node);
6911 		SYSCTL_ADD_INT(ctx_list, queue_list, OID_AUTO, "cpu",
6912 			       CTLFLAG_RD,
6913 			       &txq->ift_task.gt_cpu, 0, "cpu this queue is bound to");
6914 #if MEMORY_LOGGING
6915 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_dequeued",
6916 				CTLFLAG_RD,
6917 				&txq->ift_dequeued, "total mbufs freed");
6918 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_enqueued",
6919 				CTLFLAG_RD,
6920 				&txq->ift_enqueued, "total mbufs enqueued");
6921 #endif
6922 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "mbuf_defrag",
6923 				   CTLFLAG_RD,
6924 				   &txq->ift_mbuf_defrag, "# of times m_defrag was called");
6925 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "m_pullups",
6926 				   CTLFLAG_RD,
6927 				   &txq->ift_pullups, "# of times m_pullup was called");
6928 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "mbuf_defrag_failed",
6929 				   CTLFLAG_RD,
6930 				   &txq->ift_mbuf_defrag_failed, "# of times m_defrag failed");
6931 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "no_desc_avail",
6932 				   CTLFLAG_RD,
6933 				   &txq->ift_no_desc_avail, "# of times no descriptors were available");
6934 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "tx_map_failed",
6935 				   CTLFLAG_RD,
6936 				   &txq->ift_map_failed, "# of times DMA map failed");
6937 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txd_encap_efbig",
6938 				   CTLFLAG_RD,
6939 				   &txq->ift_txd_encap_efbig, "# of times txd_encap returned EFBIG");
6940 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "no_tx_dma_setup",
6941 				   CTLFLAG_RD,
6942 				   &txq->ift_no_tx_dma_setup, "# of times map failed for other than EFBIG");
6943 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_pidx",
6944 				   CTLFLAG_RD,
6945 				   &txq->ift_pidx, 1, "Producer Index");
6946 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_cidx",
6947 				   CTLFLAG_RD,
6948 				   &txq->ift_cidx, 1, "Consumer Index");
6949 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_cidx_processed",
6950 				   CTLFLAG_RD,
6951 				   &txq->ift_cidx_processed, 1, "Consumer Index seen by credit update");
6952 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_in_use",
6953 				   CTLFLAG_RD,
6954 				   &txq->ift_in_use, 1, "descriptors in use");
6955 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_processed",
6956 				   CTLFLAG_RD,
6957 				   &txq->ift_processed, "descriptors procesed for clean");
6958 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_cleaned",
6959 				   CTLFLAG_RD,
6960 				   &txq->ift_cleaned, "total cleaned");
6961 		SYSCTL_ADD_PROC(ctx_list, queue_list, OID_AUTO, "ring_state",
6962 		    CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_NEEDGIANT,
6963 		    __DEVOLATILE(uint64_t *, &txq->ift_br->state), 0,
6964 		    mp_ring_state_handler, "A", "soft ring state");
6965 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_enqueues",
6966 				       CTLFLAG_RD, &txq->ift_br->enqueues,
6967 				       "# of enqueues to the mp_ring for this queue");
6968 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_drops",
6969 				       CTLFLAG_RD, &txq->ift_br->drops,
6970 				       "# of drops in the mp_ring for this queue");
6971 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_starts",
6972 				       CTLFLAG_RD, &txq->ift_br->starts,
6973 				       "# of normal consumer starts in the mp_ring for this queue");
6974 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_stalls",
6975 				       CTLFLAG_RD, &txq->ift_br->stalls,
6976 					       "# of consumer stalls in the mp_ring for this queue");
6977 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_restarts",
6978 			       CTLFLAG_RD, &txq->ift_br->restarts,
6979 				       "# of consumer restarts in the mp_ring for this queue");
6980 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_abdications",
6981 				       CTLFLAG_RD, &txq->ift_br->abdications,
6982 				       "# of consumer abdications in the mp_ring for this queue");
6983 	}
6984 
6985 	if (scctx->isc_nrxqsets > 100)
6986 		qfmt = "rxq%03d";
6987 	else if (scctx->isc_nrxqsets > 10)
6988 		qfmt = "rxq%02d";
6989 	else
6990 		qfmt = "rxq%d";
6991 	for (i = 0, rxq = ctx->ifc_rxqs; i < scctx->isc_nrxqsets; i++, rxq++) {
6992 		snprintf(namebuf, NAME_BUFLEN, qfmt, i);
6993 		queue_node = SYSCTL_ADD_NODE(ctx_list, child, OID_AUTO, namebuf,
6994 		    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "Queue Name");
6995 		queue_list = SYSCTL_CHILDREN(queue_node);
6996 		SYSCTL_ADD_INT(ctx_list, queue_list, OID_AUTO, "cpu",
6997 			       CTLFLAG_RD,
6998 			       &rxq->ifr_task.gt_cpu, 0, "cpu this queue is bound to");
6999 		if (sctx->isc_flags & IFLIB_HAS_RXCQ) {
7000 			SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "rxq_cq_cidx",
7001 				       CTLFLAG_RD,
7002 				       &rxq->ifr_cq_cidx, 1, "Consumer Index");
7003 		}
7004 
7005 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++) {
7006 			snprintf(namebuf, NAME_BUFLEN, "rxq_fl%d", j);
7007 			fl_node = SYSCTL_ADD_NODE(ctx_list, queue_list, OID_AUTO, namebuf,
7008 			    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "freelist Name");
7009 			fl_list = SYSCTL_CHILDREN(fl_node);
7010 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "pidx",
7011 				       CTLFLAG_RD,
7012 				       &fl->ifl_pidx, 1, "Producer Index");
7013 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "cidx",
7014 				       CTLFLAG_RD,
7015 				       &fl->ifl_cidx, 1, "Consumer Index");
7016 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "credits",
7017 				       CTLFLAG_RD,
7018 				       &fl->ifl_credits, 1, "credits available");
7019 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "buf_size",
7020 				       CTLFLAG_RD,
7021 				       &fl->ifl_buf_size, 1, "buffer size");
7022 #if MEMORY_LOGGING
7023 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_m_enqueued",
7024 					CTLFLAG_RD,
7025 					&fl->ifl_m_enqueued, "mbufs allocated");
7026 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_m_dequeued",
7027 					CTLFLAG_RD,
7028 					&fl->ifl_m_dequeued, "mbufs freed");
7029 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_cl_enqueued",
7030 					CTLFLAG_RD,
7031 					&fl->ifl_cl_enqueued, "clusters allocated");
7032 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_cl_dequeued",
7033 					CTLFLAG_RD,
7034 					&fl->ifl_cl_dequeued, "clusters freed");
7035 #endif
7036 		}
7037 	}
7038 
7039 }
7040 
7041 void
7042 iflib_request_reset(if_ctx_t ctx)
7043 {
7044 
7045 	STATE_LOCK(ctx);
7046 	ctx->ifc_flags |= IFC_DO_RESET;
7047 	STATE_UNLOCK(ctx);
7048 }
7049 
7050 #ifndef __NO_STRICT_ALIGNMENT
7051 static struct mbuf *
7052 iflib_fixup_rx(struct mbuf *m)
7053 {
7054 	struct mbuf *n;
7055 
7056 	if (m->m_len <= (MCLBYTES - ETHER_HDR_LEN)) {
7057 		bcopy(m->m_data, m->m_data + ETHER_HDR_LEN, m->m_len);
7058 		m->m_data += ETHER_HDR_LEN;
7059 		n = m;
7060 	} else {
7061 		MGETHDR(n, M_NOWAIT, MT_DATA);
7062 		if (n == NULL) {
7063 			m_freem(m);
7064 			return (NULL);
7065 		}
7066 		bcopy(m->m_data, n->m_data, ETHER_HDR_LEN);
7067 		m->m_data += ETHER_HDR_LEN;
7068 		m->m_len -= ETHER_HDR_LEN;
7069 		n->m_len = ETHER_HDR_LEN;
7070 		M_MOVE_PKTHDR(n, m);
7071 		n->m_next = m;
7072 	}
7073 	return (n);
7074 }
7075 #endif
7076 
7077 #ifdef DEBUGNET
7078 static void
7079 iflib_debugnet_init(if_t ifp, int *nrxr, int *ncl, int *clsize)
7080 {
7081 	if_ctx_t ctx;
7082 
7083 	ctx = if_getsoftc(ifp);
7084 	CTX_LOCK(ctx);
7085 	*nrxr = NRXQSETS(ctx);
7086 	*ncl = ctx->ifc_rxqs[0].ifr_fl->ifl_size;
7087 	*clsize = ctx->ifc_rxqs[0].ifr_fl->ifl_buf_size;
7088 	CTX_UNLOCK(ctx);
7089 }
7090 
7091 static void
7092 iflib_debugnet_event(if_t ifp, enum debugnet_ev event)
7093 {
7094 	if_ctx_t ctx;
7095 	if_softc_ctx_t scctx;
7096 	iflib_fl_t fl;
7097 	iflib_rxq_t rxq;
7098 	int i, j;
7099 
7100 	ctx = if_getsoftc(ifp);
7101 	scctx = &ctx->ifc_softc_ctx;
7102 
7103 	switch (event) {
7104 	case DEBUGNET_START:
7105 		for (i = 0; i < scctx->isc_nrxqsets; i++) {
7106 			rxq = &ctx->ifc_rxqs[i];
7107 			for (j = 0; j < rxq->ifr_nfl; j++) {
7108 				fl = rxq->ifr_fl;
7109 				fl->ifl_zone = m_getzone(fl->ifl_buf_size);
7110 			}
7111 		}
7112 		iflib_no_tx_batch = 1;
7113 		break;
7114 	default:
7115 		break;
7116 	}
7117 }
7118 
7119 static int
7120 iflib_debugnet_transmit(if_t ifp, struct mbuf *m)
7121 {
7122 	if_ctx_t ctx;
7123 	iflib_txq_t txq;
7124 	int error;
7125 
7126 	ctx = if_getsoftc(ifp);
7127 	if ((if_getdrvflags(ifp) & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
7128 	    IFF_DRV_RUNNING)
7129 		return (EBUSY);
7130 
7131 	txq = &ctx->ifc_txqs[0];
7132 	error = iflib_encap(txq, &m);
7133 	if (error == 0)
7134 		(void)iflib_txd_db_check(txq, true);
7135 	return (error);
7136 }
7137 
7138 static int
7139 iflib_debugnet_poll(if_t ifp, int count)
7140 {
7141 	struct epoch_tracker et;
7142 	if_ctx_t ctx;
7143 	if_softc_ctx_t scctx;
7144 	iflib_txq_t txq;
7145 	int i;
7146 
7147 	ctx = if_getsoftc(ifp);
7148 	scctx = &ctx->ifc_softc_ctx;
7149 
7150 	if ((if_getdrvflags(ifp) & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
7151 	    IFF_DRV_RUNNING)
7152 		return (EBUSY);
7153 
7154 	txq = &ctx->ifc_txqs[0];
7155 	(void)iflib_completed_tx_reclaim(txq, RECLAIM_THRESH(ctx));
7156 
7157 	NET_EPOCH_ENTER(et);
7158 	for (i = 0; i < scctx->isc_nrxqsets; i++)
7159 		(void)iflib_rxeof(&ctx->ifc_rxqs[i], 16 /* XXX */);
7160 	NET_EPOCH_EXIT(et);
7161 	return (0);
7162 }
7163 #endif /* DEBUGNET */
7164