xref: /freebsd/sys/net/iflib.c (revision 3110d4ebd6c0848cf5e25890d01791bb407e2a9b)
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 
199 	qidx_t ifc_sysctl_ntxds[8];
200 	qidx_t ifc_sysctl_nrxds[8];
201 	struct if_txrx ifc_txrx;
202 #define isc_txd_encap  ifc_txrx.ift_txd_encap
203 #define isc_txd_flush  ifc_txrx.ift_txd_flush
204 #define isc_txd_credits_update  ifc_txrx.ift_txd_credits_update
205 #define isc_rxd_available ifc_txrx.ift_rxd_available
206 #define isc_rxd_pkt_get ifc_txrx.ift_rxd_pkt_get
207 #define isc_rxd_refill ifc_txrx.ift_rxd_refill
208 #define isc_rxd_flush ifc_txrx.ift_rxd_flush
209 #define isc_legacy_intr ifc_txrx.ift_legacy_intr
210 	eventhandler_tag ifc_vlan_attach_event;
211 	eventhandler_tag ifc_vlan_detach_event;
212 	struct ether_addr ifc_mac;
213 };
214 
215 void *
216 iflib_get_softc(if_ctx_t ctx)
217 {
218 
219 	return (ctx->ifc_softc);
220 }
221 
222 device_t
223 iflib_get_dev(if_ctx_t ctx)
224 {
225 
226 	return (ctx->ifc_dev);
227 }
228 
229 if_t
230 iflib_get_ifp(if_ctx_t ctx)
231 {
232 
233 	return (ctx->ifc_ifp);
234 }
235 
236 struct ifmedia *
237 iflib_get_media(if_ctx_t ctx)
238 {
239 
240 	return (ctx->ifc_mediap);
241 }
242 
243 uint32_t
244 iflib_get_flags(if_ctx_t ctx)
245 {
246 	return (ctx->ifc_flags);
247 }
248 
249 void
250 iflib_set_mac(if_ctx_t ctx, uint8_t mac[ETHER_ADDR_LEN])
251 {
252 
253 	bcopy(mac, ctx->ifc_mac.octet, ETHER_ADDR_LEN);
254 }
255 
256 if_softc_ctx_t
257 iflib_get_softc_ctx(if_ctx_t ctx)
258 {
259 
260 	return (&ctx->ifc_softc_ctx);
261 }
262 
263 if_shared_ctx_t
264 iflib_get_sctx(if_ctx_t ctx)
265 {
266 
267 	return (ctx->ifc_sctx);
268 }
269 
270 #define IP_ALIGNED(m) ((((uintptr_t)(m)->m_data) & 0x3) == 0x2)
271 #define CACHE_PTR_INCREMENT (CACHE_LINE_SIZE/sizeof(void*))
272 #define CACHE_PTR_NEXT(ptr) ((void *)(((uintptr_t)(ptr)+CACHE_LINE_SIZE-1) & (CACHE_LINE_SIZE-1)))
273 
274 #define LINK_ACTIVE(ctx) ((ctx)->ifc_link_state == LINK_STATE_UP)
275 #define CTX_IS_VF(ctx) ((ctx)->ifc_sctx->isc_flags & IFLIB_IS_VF)
276 
277 typedef struct iflib_sw_rx_desc_array {
278 	bus_dmamap_t	*ifsd_map;         /* bus_dma maps for packet */
279 	struct mbuf	**ifsd_m;           /* pkthdr mbufs */
280 	caddr_t		*ifsd_cl;          /* direct cluster pointer for rx */
281 	bus_addr_t	*ifsd_ba;          /* bus addr of cluster for rx */
282 } iflib_rxsd_array_t;
283 
284 typedef struct iflib_sw_tx_desc_array {
285 	bus_dmamap_t    *ifsd_map;         /* bus_dma maps for packet */
286 	bus_dmamap_t	*ifsd_tso_map;     /* bus_dma maps for TSO packet */
287 	struct mbuf    **ifsd_m;           /* pkthdr mbufs */
288 } if_txsd_vec_t;
289 
290 /* magic number that should be high enough for any hardware */
291 #define IFLIB_MAX_TX_SEGS		128
292 #define IFLIB_RX_COPY_THRESH		128
293 #define IFLIB_MAX_RX_REFRESH		32
294 /* The minimum descriptors per second before we start coalescing */
295 #define IFLIB_MIN_DESC_SEC		16384
296 #define IFLIB_DEFAULT_TX_UPDATE_FREQ	16
297 #define IFLIB_QUEUE_IDLE		0
298 #define IFLIB_QUEUE_HUNG		1
299 #define IFLIB_QUEUE_WORKING		2
300 /* maximum number of txqs that can share an rx interrupt */
301 #define IFLIB_MAX_TX_SHARED_INTR	4
302 
303 /* this should really scale with ring size - this is a fairly arbitrary value */
304 #define TX_BATCH_SIZE			32
305 
306 #define IFLIB_RESTART_BUDGET		8
307 
308 #define CSUM_OFFLOAD		(CSUM_IP_TSO|CSUM_IP6_TSO|CSUM_IP| \
309 				 CSUM_IP_UDP|CSUM_IP_TCP|CSUM_IP_SCTP| \
310 				 CSUM_IP6_UDP|CSUM_IP6_TCP|CSUM_IP6_SCTP)
311 
312 struct iflib_txq {
313 	qidx_t		ift_in_use;
314 	qidx_t		ift_cidx;
315 	qidx_t		ift_cidx_processed;
316 	qidx_t		ift_pidx;
317 	uint8_t		ift_gen;
318 	uint8_t		ift_br_offset;
319 	uint16_t	ift_npending;
320 	uint16_t	ift_db_pending;
321 	uint16_t	ift_rs_pending;
322 	/* implicit pad */
323 	uint8_t		ift_txd_size[8];
324 	uint64_t	ift_processed;
325 	uint64_t	ift_cleaned;
326 	uint64_t	ift_cleaned_prev;
327 #if MEMORY_LOGGING
328 	uint64_t	ift_enqueued;
329 	uint64_t	ift_dequeued;
330 #endif
331 	uint64_t	ift_no_tx_dma_setup;
332 	uint64_t	ift_no_desc_avail;
333 	uint64_t	ift_mbuf_defrag_failed;
334 	uint64_t	ift_mbuf_defrag;
335 	uint64_t	ift_map_failed;
336 	uint64_t	ift_txd_encap_efbig;
337 	uint64_t	ift_pullups;
338 	uint64_t	ift_last_timer_tick;
339 
340 	struct mtx	ift_mtx;
341 	struct mtx	ift_db_mtx;
342 
343 	/* constant values */
344 	if_ctx_t	ift_ctx;
345 	struct ifmp_ring        *ift_br;
346 	struct grouptask	ift_task;
347 	qidx_t		ift_size;
348 	uint16_t	ift_id;
349 	struct callout	ift_timer;
350 #ifdef DEV_NETMAP
351 	struct callout	ift_netmap_timer;
352 #endif /* DEV_NETMAP */
353 
354 	if_txsd_vec_t	ift_sds;
355 	uint8_t		ift_qstatus;
356 	uint8_t		ift_closed;
357 	uint8_t		ift_update_freq;
358 	struct iflib_filter_info ift_filter_info;
359 	bus_dma_tag_t	ift_buf_tag;
360 	bus_dma_tag_t	ift_tso_buf_tag;
361 	iflib_dma_info_t	ift_ifdi;
362 #define	MTX_NAME_LEN	32
363 	char                    ift_mtx_name[MTX_NAME_LEN];
364 	bus_dma_segment_t	ift_segs[IFLIB_MAX_TX_SEGS]  __aligned(CACHE_LINE_SIZE);
365 #ifdef IFLIB_DIAGNOSTICS
366 	uint64_t ift_cpu_exec_count[256];
367 #endif
368 } __aligned(CACHE_LINE_SIZE);
369 
370 struct iflib_fl {
371 	qidx_t		ifl_cidx;
372 	qidx_t		ifl_pidx;
373 	qidx_t		ifl_credits;
374 	uint8_t		ifl_gen;
375 	uint8_t		ifl_rxd_size;
376 #if MEMORY_LOGGING
377 	uint64_t	ifl_m_enqueued;
378 	uint64_t	ifl_m_dequeued;
379 	uint64_t	ifl_cl_enqueued;
380 	uint64_t	ifl_cl_dequeued;
381 #endif
382 	/* implicit pad */
383 	bitstr_t 	*ifl_rx_bitmap;
384 	qidx_t		ifl_fragidx;
385 	/* constant */
386 	qidx_t		ifl_size;
387 	uint16_t	ifl_buf_size;
388 	uint16_t	ifl_cltype;
389 	uma_zone_t	ifl_zone;
390 	iflib_rxsd_array_t	ifl_sds;
391 	iflib_rxq_t	ifl_rxq;
392 	uint8_t		ifl_id;
393 	bus_dma_tag_t	ifl_buf_tag;
394 	iflib_dma_info_t	ifl_ifdi;
395 	uint64_t	ifl_bus_addrs[IFLIB_MAX_RX_REFRESH] __aligned(CACHE_LINE_SIZE);
396 	qidx_t		ifl_rxd_idxs[IFLIB_MAX_RX_REFRESH];
397 }  __aligned(CACHE_LINE_SIZE);
398 
399 static inline qidx_t
400 get_inuse(int size, qidx_t cidx, qidx_t pidx, uint8_t gen)
401 {
402 	qidx_t used;
403 
404 	if (pidx > cidx)
405 		used = pidx - cidx;
406 	else if (pidx < cidx)
407 		used = size - cidx + pidx;
408 	else if (gen == 0 && pidx == cidx)
409 		used = 0;
410 	else if (gen == 1 && pidx == cidx)
411 		used = size;
412 	else
413 		panic("bad state");
414 
415 	return (used);
416 }
417 
418 #define TXQ_AVAIL(txq) (txq->ift_size - get_inuse(txq->ift_size, txq->ift_cidx, txq->ift_pidx, txq->ift_gen))
419 
420 #define IDXDIFF(head, tail, wrap) \
421 	((head) >= (tail) ? (head) - (tail) : (wrap) - (tail) + (head))
422 
423 struct iflib_rxq {
424 	if_ctx_t	ifr_ctx;
425 	iflib_fl_t	ifr_fl;
426 	uint64_t	ifr_rx_irq;
427 	struct pfil_head	*pfil;
428 	/*
429 	 * If there is a separate completion queue (IFLIB_HAS_RXCQ), this is
430 	 * the completion queue consumer index.  Otherwise it's unused.
431 	 */
432 	qidx_t		ifr_cq_cidx;
433 	uint16_t	ifr_id;
434 	uint8_t		ifr_nfl;
435 	uint8_t		ifr_ntxqirq;
436 	uint8_t		ifr_txqid[IFLIB_MAX_TX_SHARED_INTR];
437 	uint8_t		ifr_fl_offset;
438 	struct lro_ctrl			ifr_lc;
439 	struct grouptask        ifr_task;
440 	struct callout		ifr_watchdog;
441 	struct iflib_filter_info ifr_filter_info;
442 	iflib_dma_info_t		ifr_ifdi;
443 
444 	/* dynamically allocate if any drivers need a value substantially larger than this */
445 	struct if_rxd_frag	ifr_frags[IFLIB_MAX_RX_SEGS] __aligned(CACHE_LINE_SIZE);
446 #ifdef IFLIB_DIAGNOSTICS
447 	uint64_t ifr_cpu_exec_count[256];
448 #endif
449 }  __aligned(CACHE_LINE_SIZE);
450 
451 typedef struct if_rxsd {
452 	caddr_t *ifsd_cl;
453 	iflib_fl_t ifsd_fl;
454 } *if_rxsd_t;
455 
456 /* multiple of word size */
457 #ifdef __LP64__
458 #define PKT_INFO_SIZE	6
459 #define RXD_INFO_SIZE	5
460 #define PKT_TYPE uint64_t
461 #else
462 #define PKT_INFO_SIZE	11
463 #define RXD_INFO_SIZE	8
464 #define PKT_TYPE uint32_t
465 #endif
466 #define PKT_LOOP_BOUND  ((PKT_INFO_SIZE/3)*3)
467 #define RXD_LOOP_BOUND  ((RXD_INFO_SIZE/4)*4)
468 
469 typedef struct if_pkt_info_pad {
470 	PKT_TYPE pkt_val[PKT_INFO_SIZE];
471 } *if_pkt_info_pad_t;
472 typedef struct if_rxd_info_pad {
473 	PKT_TYPE rxd_val[RXD_INFO_SIZE];
474 } *if_rxd_info_pad_t;
475 
476 CTASSERT(sizeof(struct if_pkt_info_pad) == sizeof(struct if_pkt_info));
477 CTASSERT(sizeof(struct if_rxd_info_pad) == sizeof(struct if_rxd_info));
478 
479 static inline void
480 pkt_info_zero(if_pkt_info_t pi)
481 {
482 	if_pkt_info_pad_t pi_pad;
483 
484 	pi_pad = (if_pkt_info_pad_t)pi;
485 	pi_pad->pkt_val[0] = 0; pi_pad->pkt_val[1] = 0; pi_pad->pkt_val[2] = 0;
486 	pi_pad->pkt_val[3] = 0; pi_pad->pkt_val[4] = 0; pi_pad->pkt_val[5] = 0;
487 #ifndef __LP64__
488 	pi_pad->pkt_val[6] = 0; pi_pad->pkt_val[7] = 0; pi_pad->pkt_val[8] = 0;
489 	pi_pad->pkt_val[9] = 0; pi_pad->pkt_val[10] = 0;
490 #endif
491 }
492 
493 static device_method_t iflib_pseudo_methods[] = {
494 	DEVMETHOD(device_attach, noop_attach),
495 	DEVMETHOD(device_detach, iflib_pseudo_detach),
496 	DEVMETHOD_END
497 };
498 
499 driver_t iflib_pseudodriver = {
500 	"iflib_pseudo", iflib_pseudo_methods, sizeof(struct iflib_ctx),
501 };
502 
503 static inline void
504 rxd_info_zero(if_rxd_info_t ri)
505 {
506 	if_rxd_info_pad_t ri_pad;
507 	int i;
508 
509 	ri_pad = (if_rxd_info_pad_t)ri;
510 	for (i = 0; i < RXD_LOOP_BOUND; i += 4) {
511 		ri_pad->rxd_val[i] = 0;
512 		ri_pad->rxd_val[i+1] = 0;
513 		ri_pad->rxd_val[i+2] = 0;
514 		ri_pad->rxd_val[i+3] = 0;
515 	}
516 #ifdef __LP64__
517 	ri_pad->rxd_val[RXD_INFO_SIZE-1] = 0;
518 #endif
519 }
520 
521 /*
522  * Only allow a single packet to take up most 1/nth of the tx ring
523  */
524 #define MAX_SINGLE_PACKET_FRACTION 12
525 #define IF_BAD_DMA (bus_addr_t)-1
526 
527 #define CTX_ACTIVE(ctx) ((if_getdrvflags((ctx)->ifc_ifp) & IFF_DRV_RUNNING))
528 
529 #define CTX_LOCK_INIT(_sc)  sx_init(&(_sc)->ifc_ctx_sx, "iflib ctx lock")
530 #define CTX_LOCK(ctx) sx_xlock(&(ctx)->ifc_ctx_sx)
531 #define CTX_UNLOCK(ctx) sx_xunlock(&(ctx)->ifc_ctx_sx)
532 #define CTX_LOCK_DESTROY(ctx) sx_destroy(&(ctx)->ifc_ctx_sx)
533 
534 #define STATE_LOCK_INIT(_sc, _name)  mtx_init(&(_sc)->ifc_state_mtx, _name, "iflib state lock", MTX_DEF)
535 #define STATE_LOCK(ctx) mtx_lock(&(ctx)->ifc_state_mtx)
536 #define STATE_UNLOCK(ctx) mtx_unlock(&(ctx)->ifc_state_mtx)
537 #define STATE_LOCK_DESTROY(ctx) mtx_destroy(&(ctx)->ifc_state_mtx)
538 
539 #define CALLOUT_LOCK(txq)	mtx_lock(&txq->ift_mtx)
540 #define CALLOUT_UNLOCK(txq) 	mtx_unlock(&txq->ift_mtx)
541 
542 void
543 iflib_set_detach(if_ctx_t ctx)
544 {
545 	STATE_LOCK(ctx);
546 	ctx->ifc_flags |= IFC_IN_DETACH;
547 	STATE_UNLOCK(ctx);
548 }
549 
550 /* Our boot-time initialization hook */
551 static int	iflib_module_event_handler(module_t, int, void *);
552 
553 static moduledata_t iflib_moduledata = {
554 	"iflib",
555 	iflib_module_event_handler,
556 	NULL
557 };
558 
559 DECLARE_MODULE(iflib, iflib_moduledata, SI_SUB_INIT_IF, SI_ORDER_ANY);
560 MODULE_VERSION(iflib, 1);
561 
562 MODULE_DEPEND(iflib, pci, 1, 1, 1);
563 MODULE_DEPEND(iflib, ether, 1, 1, 1);
564 
565 TASKQGROUP_DEFINE(if_io_tqg, mp_ncpus, 1);
566 TASKQGROUP_DEFINE(if_config_tqg, 1, 1);
567 
568 #ifndef IFLIB_DEBUG_COUNTERS
569 #ifdef INVARIANTS
570 #define IFLIB_DEBUG_COUNTERS 1
571 #else
572 #define IFLIB_DEBUG_COUNTERS 0
573 #endif /* !INVARIANTS */
574 #endif
575 
576 static SYSCTL_NODE(_net, OID_AUTO, iflib, CTLFLAG_RD | CTLFLAG_MPSAFE, 0,
577     "iflib driver parameters");
578 
579 /*
580  * XXX need to ensure that this can't accidentally cause the head to be moved backwards
581  */
582 static int iflib_min_tx_latency = 0;
583 SYSCTL_INT(_net_iflib, OID_AUTO, min_tx_latency, CTLFLAG_RW,
584 		   &iflib_min_tx_latency, 0, "minimize transmit latency at the possible expense of throughput");
585 static int iflib_no_tx_batch = 0;
586 SYSCTL_INT(_net_iflib, OID_AUTO, no_tx_batch, CTLFLAG_RW,
587 		   &iflib_no_tx_batch, 0, "minimize transmit latency at the possible expense of throughput");
588 static int iflib_timer_default = 1000;
589 SYSCTL_INT(_net_iflib, OID_AUTO, timer_default, CTLFLAG_RW,
590 		   &iflib_timer_default, 0, "number of ticks between iflib_timer calls");
591 
592 
593 #if IFLIB_DEBUG_COUNTERS
594 
595 static int iflib_tx_seen;
596 static int iflib_tx_sent;
597 static int iflib_tx_encap;
598 static int iflib_rx_allocs;
599 static int iflib_fl_refills;
600 static int iflib_fl_refills_large;
601 static int iflib_tx_frees;
602 
603 SYSCTL_INT(_net_iflib, OID_AUTO, tx_seen, CTLFLAG_RD,
604 		   &iflib_tx_seen, 0, "# TX mbufs seen");
605 SYSCTL_INT(_net_iflib, OID_AUTO, tx_sent, CTLFLAG_RD,
606 		   &iflib_tx_sent, 0, "# TX mbufs sent");
607 SYSCTL_INT(_net_iflib, OID_AUTO, tx_encap, CTLFLAG_RD,
608 		   &iflib_tx_encap, 0, "# TX mbufs encapped");
609 SYSCTL_INT(_net_iflib, OID_AUTO, tx_frees, CTLFLAG_RD,
610 		   &iflib_tx_frees, 0, "# TX frees");
611 SYSCTL_INT(_net_iflib, OID_AUTO, rx_allocs, CTLFLAG_RD,
612 		   &iflib_rx_allocs, 0, "# RX allocations");
613 SYSCTL_INT(_net_iflib, OID_AUTO, fl_refills, CTLFLAG_RD,
614 		   &iflib_fl_refills, 0, "# refills");
615 SYSCTL_INT(_net_iflib, OID_AUTO, fl_refills_large, CTLFLAG_RD,
616 		   &iflib_fl_refills_large, 0, "# large refills");
617 
618 static int iflib_txq_drain_flushing;
619 static int iflib_txq_drain_oactive;
620 static int iflib_txq_drain_notready;
621 
622 SYSCTL_INT(_net_iflib, OID_AUTO, txq_drain_flushing, CTLFLAG_RD,
623 		   &iflib_txq_drain_flushing, 0, "# drain flushes");
624 SYSCTL_INT(_net_iflib, OID_AUTO, txq_drain_oactive, CTLFLAG_RD,
625 		   &iflib_txq_drain_oactive, 0, "# drain oactives");
626 SYSCTL_INT(_net_iflib, OID_AUTO, txq_drain_notready, CTLFLAG_RD,
627 		   &iflib_txq_drain_notready, 0, "# drain notready");
628 
629 static int iflib_encap_load_mbuf_fail;
630 static int iflib_encap_pad_mbuf_fail;
631 static int iflib_encap_txq_avail_fail;
632 static int iflib_encap_txd_encap_fail;
633 
634 SYSCTL_INT(_net_iflib, OID_AUTO, encap_load_mbuf_fail, CTLFLAG_RD,
635 		   &iflib_encap_load_mbuf_fail, 0, "# busdma load failures");
636 SYSCTL_INT(_net_iflib, OID_AUTO, encap_pad_mbuf_fail, CTLFLAG_RD,
637 		   &iflib_encap_pad_mbuf_fail, 0, "# runt frame pad failures");
638 SYSCTL_INT(_net_iflib, OID_AUTO, encap_txq_avail_fail, CTLFLAG_RD,
639 		   &iflib_encap_txq_avail_fail, 0, "# txq avail failures");
640 SYSCTL_INT(_net_iflib, OID_AUTO, encap_txd_encap_fail, CTLFLAG_RD,
641 		   &iflib_encap_txd_encap_fail, 0, "# driver encap failures");
642 
643 static int iflib_task_fn_rxs;
644 static int iflib_rx_intr_enables;
645 static int iflib_fast_intrs;
646 static int iflib_rx_unavail;
647 static int iflib_rx_ctx_inactive;
648 static int iflib_rx_if_input;
649 static int iflib_rxd_flush;
650 
651 static int iflib_verbose_debug;
652 
653 SYSCTL_INT(_net_iflib, OID_AUTO, task_fn_rx, CTLFLAG_RD,
654 		   &iflib_task_fn_rxs, 0, "# task_fn_rx calls");
655 SYSCTL_INT(_net_iflib, OID_AUTO, rx_intr_enables, CTLFLAG_RD,
656 		   &iflib_rx_intr_enables, 0, "# RX intr enables");
657 SYSCTL_INT(_net_iflib, OID_AUTO, fast_intrs, CTLFLAG_RD,
658 		   &iflib_fast_intrs, 0, "# fast_intr calls");
659 SYSCTL_INT(_net_iflib, OID_AUTO, rx_unavail, CTLFLAG_RD,
660 		   &iflib_rx_unavail, 0, "# times rxeof called with no available data");
661 SYSCTL_INT(_net_iflib, OID_AUTO, rx_ctx_inactive, CTLFLAG_RD,
662 		   &iflib_rx_ctx_inactive, 0, "# times rxeof called with inactive context");
663 SYSCTL_INT(_net_iflib, OID_AUTO, rx_if_input, CTLFLAG_RD,
664 		   &iflib_rx_if_input, 0, "# times rxeof called if_input");
665 SYSCTL_INT(_net_iflib, OID_AUTO, rxd_flush, CTLFLAG_RD,
666 	         &iflib_rxd_flush, 0, "# times rxd_flush called");
667 SYSCTL_INT(_net_iflib, OID_AUTO, verbose_debug, CTLFLAG_RW,
668 		   &iflib_verbose_debug, 0, "enable verbose debugging");
669 
670 #define DBG_COUNTER_INC(name) atomic_add_int(&(iflib_ ## name), 1)
671 static void
672 iflib_debug_reset(void)
673 {
674 	iflib_tx_seen = iflib_tx_sent = iflib_tx_encap = iflib_rx_allocs =
675 		iflib_fl_refills = iflib_fl_refills_large = iflib_tx_frees =
676 		iflib_txq_drain_flushing = iflib_txq_drain_oactive =
677 		iflib_txq_drain_notready =
678 		iflib_encap_load_mbuf_fail = iflib_encap_pad_mbuf_fail =
679 		iflib_encap_txq_avail_fail = iflib_encap_txd_encap_fail =
680 		iflib_task_fn_rxs = iflib_rx_intr_enables = iflib_fast_intrs =
681 		iflib_rx_unavail =
682 		iflib_rx_ctx_inactive = iflib_rx_if_input =
683 		iflib_rxd_flush = 0;
684 }
685 
686 #else
687 #define DBG_COUNTER_INC(name)
688 static void iflib_debug_reset(void) {}
689 #endif
690 
691 #define IFLIB_DEBUG 0
692 
693 static void iflib_tx_structures_free(if_ctx_t ctx);
694 static void iflib_rx_structures_free(if_ctx_t ctx);
695 static int iflib_queues_alloc(if_ctx_t ctx);
696 static int iflib_tx_credits_update(if_ctx_t ctx, iflib_txq_t txq);
697 static int iflib_rxd_avail(if_ctx_t ctx, iflib_rxq_t rxq, qidx_t cidx, qidx_t budget);
698 static int iflib_qset_structures_setup(if_ctx_t ctx);
699 static int iflib_msix_init(if_ctx_t ctx);
700 static int iflib_legacy_setup(if_ctx_t ctx, driver_filter_t filter, void *filterarg, int *rid, const char *str);
701 static void iflib_txq_check_drain(iflib_txq_t txq, int budget);
702 static uint32_t iflib_txq_can_drain(struct ifmp_ring *);
703 #ifdef ALTQ
704 static void iflib_altq_if_start(if_t ifp);
705 static int iflib_altq_if_transmit(if_t ifp, struct mbuf *m);
706 #endif
707 static int iflib_register(if_ctx_t);
708 static void iflib_deregister(if_ctx_t);
709 static void iflib_unregister_vlan_handlers(if_ctx_t ctx);
710 static uint16_t iflib_get_mbuf_size_for(unsigned int size);
711 static void iflib_init_locked(if_ctx_t ctx);
712 static void iflib_add_device_sysctl_pre(if_ctx_t ctx);
713 static void iflib_add_device_sysctl_post(if_ctx_t ctx);
714 static void iflib_ifmp_purge(iflib_txq_t txq);
715 static void _iflib_pre_assert(if_softc_ctx_t scctx);
716 static void iflib_if_init_locked(if_ctx_t ctx);
717 static void iflib_free_intr_mem(if_ctx_t ctx);
718 #ifndef __NO_STRICT_ALIGNMENT
719 static struct mbuf * iflib_fixup_rx(struct mbuf *m);
720 #endif
721 
722 static SLIST_HEAD(cpu_offset_list, cpu_offset) cpu_offsets =
723     SLIST_HEAD_INITIALIZER(cpu_offsets);
724 struct cpu_offset {
725 	SLIST_ENTRY(cpu_offset) entries;
726 	cpuset_t	set;
727 	unsigned int	refcount;
728 	uint16_t	offset;
729 };
730 static struct mtx cpu_offset_mtx;
731 MTX_SYSINIT(iflib_cpu_offset, &cpu_offset_mtx, "iflib_cpu_offset lock",
732     MTX_DEF);
733 
734 DEBUGNET_DEFINE(iflib);
735 
736 static int
737 iflib_num_rx_descs(if_ctx_t ctx)
738 {
739 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
740 	if_shared_ctx_t sctx = ctx->ifc_sctx;
741 	uint16_t first_rxq = (sctx->isc_flags & IFLIB_HAS_RXCQ) ? 1 : 0;
742 
743 	return scctx->isc_nrxd[first_rxq];
744 }
745 
746 static int
747 iflib_num_tx_descs(if_ctx_t ctx)
748 {
749 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
750 	if_shared_ctx_t sctx = ctx->ifc_sctx;
751 	uint16_t first_txq = (sctx->isc_flags & IFLIB_HAS_TXCQ) ? 1 : 0;
752 
753 	return scctx->isc_ntxd[first_txq];
754 }
755 
756 #ifdef DEV_NETMAP
757 #include <sys/selinfo.h>
758 #include <net/netmap.h>
759 #include <dev/netmap/netmap_kern.h>
760 
761 MODULE_DEPEND(iflib, netmap, 1, 1, 1);
762 
763 static int netmap_fl_refill(iflib_rxq_t rxq, struct netmap_kring *kring, bool init);
764 static void iflib_netmap_timer(void *arg);
765 
766 /*
767  * device-specific sysctl variables:
768  *
769  * iflib_crcstrip: 0: keep CRC in rx frames (default), 1: strip it.
770  *	During regular operations the CRC is stripped, but on some
771  *	hardware reception of frames not multiple of 64 is slower,
772  *	so using crcstrip=0 helps in benchmarks.
773  *
774  * iflib_rx_miss, iflib_rx_miss_bufs:
775  *	count packets that might be missed due to lost interrupts.
776  */
777 SYSCTL_DECL(_dev_netmap);
778 /*
779  * The xl driver by default strips CRCs and we do not override it.
780  */
781 
782 int iflib_crcstrip = 1;
783 SYSCTL_INT(_dev_netmap, OID_AUTO, iflib_crcstrip,
784     CTLFLAG_RW, &iflib_crcstrip, 1, "strip CRC on RX frames");
785 
786 int iflib_rx_miss, iflib_rx_miss_bufs;
787 SYSCTL_INT(_dev_netmap, OID_AUTO, iflib_rx_miss,
788     CTLFLAG_RW, &iflib_rx_miss, 0, "potentially missed RX intr");
789 SYSCTL_INT(_dev_netmap, OID_AUTO, iflib_rx_miss_bufs,
790     CTLFLAG_RW, &iflib_rx_miss_bufs, 0, "potentially missed RX intr bufs");
791 
792 /*
793  * Register/unregister. We are already under netmap lock.
794  * Only called on the first register or the last unregister.
795  */
796 static int
797 iflib_netmap_register(struct netmap_adapter *na, int onoff)
798 {
799 	if_t ifp = na->ifp;
800 	if_ctx_t ctx = ifp->if_softc;
801 	int status;
802 
803 	CTX_LOCK(ctx);
804 	if (!CTX_IS_VF(ctx))
805 		IFDI_CRCSTRIP_SET(ctx, onoff, iflib_crcstrip);
806 
807 	iflib_stop(ctx);
808 
809 	/*
810 	 * Enable (or disable) netmap flags, and intercept (or restore)
811 	 * ifp->if_transmit. This is done once the device has been stopped
812 	 * to prevent race conditions. Also, this must be done after
813 	 * calling netmap_disable_all_rings() and before calling
814 	 * netmap_enable_all_rings(), so that these two functions see the
815 	 * updated state of the NAF_NETMAP_ON bit.
816 	 */
817 	if (onoff) {
818 		nm_set_native_flags(na);
819 	} else {
820 		nm_clear_native_flags(na);
821 	}
822 
823 	iflib_init_locked(ctx);
824 	IFDI_CRCSTRIP_SET(ctx, onoff, iflib_crcstrip); // XXX why twice ?
825 	status = ifp->if_drv_flags & IFF_DRV_RUNNING ? 0 : 1;
826 	if (status)
827 		nm_clear_native_flags(na);
828 	CTX_UNLOCK(ctx);
829 	return (status);
830 }
831 
832 static int
833 netmap_fl_refill(iflib_rxq_t rxq, struct netmap_kring *kring, bool init)
834 {
835 	struct netmap_adapter *na = kring->na;
836 	u_int const lim = kring->nkr_num_slots - 1;
837 	struct netmap_ring *ring = kring->ring;
838 	bus_dmamap_t *map;
839 	struct if_rxd_update iru;
840 	if_ctx_t ctx = rxq->ifr_ctx;
841 	iflib_fl_t fl = &rxq->ifr_fl[0];
842 	u_int nic_i_first, nic_i;
843 	u_int nm_i;
844 	int i, n;
845 #if IFLIB_DEBUG_COUNTERS
846 	int rf_count = 0;
847 #endif
848 
849 	/*
850 	 * This function is used both at initialization and in rxsync.
851 	 * At initialization we need to prepare (with isc_rxd_refill())
852 	 * all the netmap buffers currently owned by the kernel, in
853 	 * such a way to keep fl->ifl_pidx and kring->nr_hwcur in sync
854 	 * (except for kring->nkr_hwofs). These may be less than
855 	 * kring->nkr_num_slots if netmap_reset() was called while
856 	 * an application using the kring that still owned some
857 	 * buffers.
858 	 * At rxsync time, both indexes point to the next buffer to be
859 	 * refilled.
860 	 * In any case we publish (with isc_rxd_flush()) up to
861 	 * (fl->ifl_pidx - 1) % N (included), to avoid the NIC tail/prod
862 	 * pointer to overrun the head/cons pointer, although this is
863 	 * not necessary for some NICs (e.g. vmx).
864 	 */
865 	if (__predict_false(init)) {
866 		n = kring->nkr_num_slots - nm_kr_rxspace(kring);
867 	} else {
868 		n = kring->rhead - kring->nr_hwcur;
869 		if (n == 0)
870 			return (0); /* Nothing to do. */
871 		if (n < 0)
872 			n += kring->nkr_num_slots;
873 	}
874 
875 	iru_init(&iru, rxq, 0 /* flid */);
876 	map = fl->ifl_sds.ifsd_map;
877 	nic_i = fl->ifl_pidx;
878 	nm_i = netmap_idx_n2k(kring, nic_i);
879 	if (__predict_false(init)) {
880 		/*
881 		 * On init/reset, nic_i must be 0, and we must
882 		 * start to refill from hwtail (see netmap_reset()).
883 		 */
884 		MPASS(nic_i == 0);
885 		MPASS(nm_i == kring->nr_hwtail);
886 	} else
887 		MPASS(nm_i == kring->nr_hwcur);
888 	DBG_COUNTER_INC(fl_refills);
889 	while (n > 0) {
890 #if IFLIB_DEBUG_COUNTERS
891 		if (++rf_count == 9)
892 			DBG_COUNTER_INC(fl_refills_large);
893 #endif
894 		nic_i_first = nic_i;
895 		for (i = 0; n > 0 && i < IFLIB_MAX_RX_REFRESH; n--, i++) {
896 			struct netmap_slot *slot = &ring->slot[nm_i];
897 			void *addr = PNMB(na, slot, &fl->ifl_bus_addrs[i]);
898 
899 			MPASS(i < IFLIB_MAX_RX_REFRESH);
900 
901 			if (addr == NETMAP_BUF_BASE(na)) /* bad buf */
902 			        return netmap_ring_reinit(kring);
903 
904 			fl->ifl_rxd_idxs[i] = nic_i;
905 
906 			if (__predict_false(init)) {
907 				netmap_load_map(na, fl->ifl_buf_tag,
908 				    map[nic_i], addr);
909 			} else if (slot->flags & NS_BUF_CHANGED) {
910 				/* buffer has changed, reload map */
911 				netmap_reload_map(na, fl->ifl_buf_tag,
912 				    map[nic_i], addr);
913 			}
914 			bus_dmamap_sync(fl->ifl_buf_tag, map[nic_i],
915 			    BUS_DMASYNC_PREREAD);
916 			slot->flags &= ~NS_BUF_CHANGED;
917 
918 			nm_i = nm_next(nm_i, lim);
919 			nic_i = nm_next(nic_i, lim);
920 		}
921 
922 		iru.iru_pidx = nic_i_first;
923 		iru.iru_count = i;
924 		ctx->isc_rxd_refill(ctx->ifc_softc, &iru);
925 	}
926 	fl->ifl_pidx = nic_i;
927 	/*
928 	 * At the end of the loop we must have refilled everything
929 	 * we could possibly refill.
930 	 */
931 	MPASS(nm_i == kring->rhead);
932 	kring->nr_hwcur = nm_i;
933 
934 	bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
935 	    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
936 	ctx->isc_rxd_flush(ctx->ifc_softc, rxq->ifr_id, fl->ifl_id,
937 	    nm_prev(nic_i, lim));
938 	DBG_COUNTER_INC(rxd_flush);
939 
940 	return (0);
941 }
942 
943 #define NETMAP_TX_TIMER_US	90
944 
945 /*
946  * Reconcile kernel and user view of the transmit ring.
947  *
948  * All information is in the kring.
949  * Userspace wants to send packets up to the one before kring->rhead,
950  * kernel knows kring->nr_hwcur is the first unsent packet.
951  *
952  * Here we push packets out (as many as possible), and possibly
953  * reclaim buffers from previously completed transmission.
954  *
955  * The caller (netmap) guarantees that there is only one instance
956  * running at any time. Any interference with other driver
957  * methods should be handled by the individual drivers.
958  */
959 static int
960 iflib_netmap_txsync(struct netmap_kring *kring, int flags)
961 {
962 	struct netmap_adapter *na = kring->na;
963 	if_t ifp = na->ifp;
964 	struct netmap_ring *ring = kring->ring;
965 	u_int nm_i;	/* index into the netmap kring */
966 	u_int nic_i;	/* index into the NIC ring */
967 	u_int n;
968 	u_int const lim = kring->nkr_num_slots - 1;
969 	u_int const head = kring->rhead;
970 	struct if_pkt_info pi;
971 
972 	/*
973 	 * interrupts on every tx packet are expensive so request
974 	 * them every half ring, or where NS_REPORT is set
975 	 */
976 	u_int report_frequency = kring->nkr_num_slots >> 1;
977 	/* device-specific */
978 	if_ctx_t ctx = ifp->if_softc;
979 	iflib_txq_t txq = &ctx->ifc_txqs[kring->ring_id];
980 
981 	bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
982 	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
983 
984 	/*
985 	 * First part: process new packets to send.
986 	 * nm_i is the current index in the netmap kring,
987 	 * nic_i is the corresponding index in the NIC ring.
988 	 *
989 	 * If we have packets to send (nm_i != head)
990 	 * iterate over the netmap ring, fetch length and update
991 	 * the corresponding slot in the NIC ring. Some drivers also
992 	 * need to update the buffer's physical address in the NIC slot
993 	 * even NS_BUF_CHANGED is not set (PNMB computes the addresses).
994 	 *
995 	 * The netmap_reload_map() calls is especially expensive,
996 	 * even when (as in this case) the tag is 0, so do only
997 	 * when the buffer has actually changed.
998 	 *
999 	 * If possible do not set the report/intr bit on all slots,
1000 	 * but only a few times per ring or when NS_REPORT is set.
1001 	 *
1002 	 * Finally, on 10G and faster drivers, it might be useful
1003 	 * to prefetch the next slot and txr entry.
1004 	 */
1005 
1006 	nm_i = kring->nr_hwcur;
1007 	if (nm_i != head) {	/* we have new packets to send */
1008 		pkt_info_zero(&pi);
1009 		pi.ipi_segs = txq->ift_segs;
1010 		pi.ipi_qsidx = kring->ring_id;
1011 		nic_i = netmap_idx_k2n(kring, nm_i);
1012 
1013 		__builtin_prefetch(&ring->slot[nm_i]);
1014 		__builtin_prefetch(&txq->ift_sds.ifsd_m[nic_i]);
1015 		__builtin_prefetch(&txq->ift_sds.ifsd_map[nic_i]);
1016 
1017 		for (n = 0; nm_i != head; n++) {
1018 			struct netmap_slot *slot = &ring->slot[nm_i];
1019 			u_int len = slot->len;
1020 			uint64_t paddr;
1021 			void *addr = PNMB(na, slot, &paddr);
1022 			int flags = (slot->flags & NS_REPORT ||
1023 				nic_i == 0 || nic_i == report_frequency) ?
1024 				IPI_TX_INTR : 0;
1025 
1026 			/* device-specific */
1027 			pi.ipi_len = len;
1028 			pi.ipi_segs[0].ds_addr = paddr;
1029 			pi.ipi_segs[0].ds_len = len;
1030 			pi.ipi_nsegs = 1;
1031 			pi.ipi_ndescs = 0;
1032 			pi.ipi_pidx = nic_i;
1033 			pi.ipi_flags = flags;
1034 
1035 			/* Fill the slot in the NIC ring. */
1036 			ctx->isc_txd_encap(ctx->ifc_softc, &pi);
1037 			DBG_COUNTER_INC(tx_encap);
1038 
1039 			/* prefetch for next round */
1040 			__builtin_prefetch(&ring->slot[nm_i + 1]);
1041 			__builtin_prefetch(&txq->ift_sds.ifsd_m[nic_i + 1]);
1042 			__builtin_prefetch(&txq->ift_sds.ifsd_map[nic_i + 1]);
1043 
1044 			NM_CHECK_ADDR_LEN(na, addr, len);
1045 
1046 			if (slot->flags & NS_BUF_CHANGED) {
1047 				/* buffer has changed, reload map */
1048 				netmap_reload_map(na, txq->ift_buf_tag,
1049 				    txq->ift_sds.ifsd_map[nic_i], addr);
1050 			}
1051 			/* make sure changes to the buffer are synced */
1052 			bus_dmamap_sync(txq->ift_buf_tag,
1053 			    txq->ift_sds.ifsd_map[nic_i],
1054 			    BUS_DMASYNC_PREWRITE);
1055 
1056 			slot->flags &= ~(NS_REPORT | NS_BUF_CHANGED);
1057 			nm_i = nm_next(nm_i, lim);
1058 			nic_i = nm_next(nic_i, lim);
1059 		}
1060 		kring->nr_hwcur = nm_i;
1061 
1062 		/* synchronize the NIC ring */
1063 		bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
1064 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
1065 
1066 		/* (re)start the tx unit up to slot nic_i (excluded) */
1067 		ctx->isc_txd_flush(ctx->ifc_softc, txq->ift_id, nic_i);
1068 	}
1069 
1070 	/*
1071 	 * Second part: reclaim buffers for completed transmissions.
1072 	 *
1073 	 * If there are unclaimed buffers, attempt to reclaim them.
1074 	 * If we don't manage to reclaim them all, and TX IRQs are not in use,
1075 	 * trigger a per-tx-queue timer to try again later.
1076 	 */
1077 	if (kring->nr_hwtail != nm_prev(kring->nr_hwcur, lim)) {
1078 		if (iflib_tx_credits_update(ctx, txq)) {
1079 			/* some tx completed, increment avail */
1080 			nic_i = txq->ift_cidx_processed;
1081 			kring->nr_hwtail = nm_prev(netmap_idx_n2k(kring, nic_i), lim);
1082 		}
1083 	}
1084 
1085 	if (!(ctx->ifc_flags & IFC_NETMAP_TX_IRQ))
1086 		if (kring->nr_hwtail != nm_prev(kring->nr_hwcur, lim)) {
1087 			callout_reset_sbt_on(&txq->ift_netmap_timer,
1088 			    NETMAP_TX_TIMER_US * SBT_1US, SBT_1US,
1089 			    iflib_netmap_timer, txq,
1090 			    txq->ift_netmap_timer.c_cpu, 0);
1091 		}
1092 	return (0);
1093 }
1094 
1095 /*
1096  * Reconcile kernel and user view of the receive ring.
1097  * Same as for the txsync, this routine must be efficient.
1098  * The caller guarantees a single invocations, but races against
1099  * the rest of the driver should be handled here.
1100  *
1101  * On call, kring->rhead is the first packet that userspace wants
1102  * to keep, and kring->rcur is the wakeup point.
1103  * The kernel has previously reported packets up to kring->rtail.
1104  *
1105  * If (flags & NAF_FORCE_READ) also check for incoming packets irrespective
1106  * of whether or not we received an interrupt.
1107  */
1108 static int
1109 iflib_netmap_rxsync(struct netmap_kring *kring, int flags)
1110 {
1111 	struct netmap_adapter *na = kring->na;
1112 	struct netmap_ring *ring = kring->ring;
1113 	if_t ifp = na->ifp;
1114 	uint32_t nm_i;	/* index into the netmap ring */
1115 	uint32_t nic_i;	/* index into the NIC ring */
1116 	u_int n;
1117 	u_int const lim = kring->nkr_num_slots - 1;
1118 	int force_update = (flags & NAF_FORCE_READ) || kring->nr_kflags & NKR_PENDINTR;
1119 
1120 	if_ctx_t ctx = ifp->if_softc;
1121 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1122 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1123 	iflib_rxq_t rxq = &ctx->ifc_rxqs[kring->ring_id];
1124 	iflib_fl_t fl = &rxq->ifr_fl[0];
1125 	struct if_rxd_info ri;
1126 	qidx_t *cidxp;
1127 
1128 	/*
1129 	 * netmap only uses free list 0, to avoid out of order consumption
1130 	 * of receive buffers
1131 	 */
1132 
1133 	bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
1134 	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
1135 
1136 	/*
1137 	 * First part: import newly received packets.
1138 	 *
1139 	 * nm_i is the index of the next free slot in the netmap ring,
1140 	 * nic_i is the index of the next received packet in the NIC ring
1141 	 * (or in the free list 0 if IFLIB_HAS_RXCQ is set), and they may
1142 	 * differ in case if_init() has been called while
1143 	 * in netmap mode. For the receive ring we have
1144 	 *
1145 	 *	nic_i = fl->ifl_cidx;
1146 	 *	nm_i = kring->nr_hwtail (previous)
1147 	 * and
1148 	 *	nm_i == (nic_i + kring->nkr_hwofs) % ring_size
1149 	 *
1150 	 * fl->ifl_cidx is set to 0 on a ring reinit
1151 	 */
1152 	if (netmap_no_pendintr || force_update) {
1153 		uint32_t hwtail_lim = nm_prev(kring->nr_hwcur, lim);
1154 		bool have_rxcq = sctx->isc_flags & IFLIB_HAS_RXCQ;
1155 		int crclen = iflib_crcstrip ? 0 : 4;
1156 		int error, avail;
1157 
1158 		/*
1159 		 * For the free list consumer index, we use the same
1160 		 * logic as in iflib_rxeof().
1161 		 */
1162 		if (have_rxcq)
1163 			cidxp = &rxq->ifr_cq_cidx;
1164 		else
1165 			cidxp = &fl->ifl_cidx;
1166 		avail = ctx->isc_rxd_available(ctx->ifc_softc,
1167 		    rxq->ifr_id, *cidxp, USHRT_MAX);
1168 
1169 		nic_i = fl->ifl_cidx;
1170 		nm_i = netmap_idx_n2k(kring, nic_i);
1171 		MPASS(nm_i == kring->nr_hwtail);
1172 		for (n = 0; avail > 0 && nm_i != hwtail_lim; n++, avail--) {
1173 			rxd_info_zero(&ri);
1174 			ri.iri_frags = rxq->ifr_frags;
1175 			ri.iri_qsidx = kring->ring_id;
1176 			ri.iri_ifp = ctx->ifc_ifp;
1177 			ri.iri_cidx = *cidxp;
1178 
1179 			error = ctx->isc_rxd_pkt_get(ctx->ifc_softc, &ri);
1180 			ring->slot[nm_i].len = error ? 0 : ri.iri_len - crclen;
1181 			ring->slot[nm_i].flags = 0;
1182 			if (have_rxcq) {
1183 				*cidxp = ri.iri_cidx;
1184 				while (*cidxp >= scctx->isc_nrxd[0])
1185 					*cidxp -= scctx->isc_nrxd[0];
1186 			}
1187 			bus_dmamap_sync(fl->ifl_buf_tag,
1188 			    fl->ifl_sds.ifsd_map[nic_i], BUS_DMASYNC_POSTREAD);
1189 			nm_i = nm_next(nm_i, lim);
1190 			fl->ifl_cidx = nic_i = nm_next(nic_i, lim);
1191 		}
1192 		if (n) { /* update the state variables */
1193 			if (netmap_no_pendintr && !force_update) {
1194 				/* diagnostics */
1195 				iflib_rx_miss ++;
1196 				iflib_rx_miss_bufs += n;
1197 			}
1198 			kring->nr_hwtail = nm_i;
1199 		}
1200 		kring->nr_kflags &= ~NKR_PENDINTR;
1201 	}
1202 	/*
1203 	 * Second part: skip past packets that userspace has released.
1204 	 * (kring->nr_hwcur to head excluded),
1205 	 * and make the buffers available for reception.
1206 	 * As usual nm_i is the index in the netmap ring,
1207 	 * nic_i is the index in the NIC ring, and
1208 	 * nm_i == (nic_i + kring->nkr_hwofs) % ring_size
1209 	 */
1210 	netmap_fl_refill(rxq, kring, false);
1211 
1212 	return (0);
1213 }
1214 
1215 static void
1216 iflib_netmap_intr(struct netmap_adapter *na, int onoff)
1217 {
1218 	if_ctx_t ctx = na->ifp->if_softc;
1219 
1220 	CTX_LOCK(ctx);
1221 	if (onoff) {
1222 		IFDI_INTR_ENABLE(ctx);
1223 	} else {
1224 		IFDI_INTR_DISABLE(ctx);
1225 	}
1226 	CTX_UNLOCK(ctx);
1227 }
1228 
1229 static int
1230 iflib_netmap_attach(if_ctx_t ctx)
1231 {
1232 	struct netmap_adapter na;
1233 
1234 	bzero(&na, sizeof(na));
1235 
1236 	na.ifp = ctx->ifc_ifp;
1237 	na.na_flags = NAF_BDG_MAYSLEEP;
1238 	MPASS(ctx->ifc_softc_ctx.isc_ntxqsets);
1239 	MPASS(ctx->ifc_softc_ctx.isc_nrxqsets);
1240 
1241 	na.num_tx_desc = iflib_num_tx_descs(ctx);
1242 	na.num_rx_desc = iflib_num_rx_descs(ctx);
1243 	na.nm_txsync = iflib_netmap_txsync;
1244 	na.nm_rxsync = iflib_netmap_rxsync;
1245 	na.nm_register = iflib_netmap_register;
1246 	na.nm_intr = iflib_netmap_intr;
1247 	na.num_tx_rings = ctx->ifc_softc_ctx.isc_ntxqsets;
1248 	na.num_rx_rings = ctx->ifc_softc_ctx.isc_nrxqsets;
1249 	return (netmap_attach(&na));
1250 }
1251 
1252 static int
1253 iflib_netmap_txq_init(if_ctx_t ctx, iflib_txq_t txq)
1254 {
1255 	struct netmap_adapter *na = NA(ctx->ifc_ifp);
1256 	struct netmap_slot *slot;
1257 
1258 	slot = netmap_reset(na, NR_TX, txq->ift_id, 0);
1259 	if (slot == NULL)
1260 		return (0);
1261 	for (int i = 0; i < ctx->ifc_softc_ctx.isc_ntxd[0]; i++) {
1262 		/*
1263 		 * In netmap mode, set the map for the packet buffer.
1264 		 * NOTE: Some drivers (not this one) also need to set
1265 		 * the physical buffer address in the NIC ring.
1266 		 * netmap_idx_n2k() maps a nic index, i, into the corresponding
1267 		 * netmap slot index, si
1268 		 */
1269 		int si = netmap_idx_n2k(na->tx_rings[txq->ift_id], i);
1270 		netmap_load_map(na, txq->ift_buf_tag, txq->ift_sds.ifsd_map[i],
1271 		    NMB(na, slot + si));
1272 	}
1273 	return (1);
1274 }
1275 
1276 static int
1277 iflib_netmap_rxq_init(if_ctx_t ctx, iflib_rxq_t rxq)
1278 {
1279 	struct netmap_adapter *na = NA(ctx->ifc_ifp);
1280 	struct netmap_kring *kring;
1281 	struct netmap_slot *slot;
1282 
1283 	slot = netmap_reset(na, NR_RX, rxq->ifr_id, 0);
1284 	if (slot == NULL)
1285 		return (0);
1286 	kring = na->rx_rings[rxq->ifr_id];
1287 	netmap_fl_refill(rxq, kring, true);
1288 	return (1);
1289 }
1290 
1291 static void
1292 iflib_netmap_timer(void *arg)
1293 {
1294 	iflib_txq_t txq = arg;
1295 	if_ctx_t ctx = txq->ift_ctx;
1296 
1297 	/*
1298 	 * Wake up the netmap application, to give it a chance to
1299 	 * call txsync and reclaim more completed TX buffers.
1300 	 */
1301 	netmap_tx_irq(ctx->ifc_ifp, txq->ift_id);
1302 }
1303 
1304 #define iflib_netmap_detach(ifp) netmap_detach(ifp)
1305 
1306 #else
1307 #define iflib_netmap_txq_init(ctx, txq) (0)
1308 #define iflib_netmap_rxq_init(ctx, rxq) (0)
1309 #define iflib_netmap_detach(ifp)
1310 #define netmap_enable_all_rings(ifp)
1311 #define netmap_disable_all_rings(ifp)
1312 
1313 #define iflib_netmap_attach(ctx) (0)
1314 #define netmap_rx_irq(ifp, qid, budget) (0)
1315 #endif
1316 
1317 #if defined(__i386__) || defined(__amd64__)
1318 static __inline void
1319 prefetch(void *x)
1320 {
1321 	__asm volatile("prefetcht0 %0" :: "m" (*(unsigned long *)x));
1322 }
1323 static __inline void
1324 prefetch2cachelines(void *x)
1325 {
1326 	__asm volatile("prefetcht0 %0" :: "m" (*(unsigned long *)x));
1327 #if (CACHE_LINE_SIZE < 128)
1328 	__asm volatile("prefetcht0 %0" :: "m" (*(((unsigned long *)x)+CACHE_LINE_SIZE/(sizeof(unsigned long)))));
1329 #endif
1330 }
1331 #else
1332 #define prefetch(x)
1333 #define prefetch2cachelines(x)
1334 #endif
1335 
1336 static void
1337 iru_init(if_rxd_update_t iru, iflib_rxq_t rxq, uint8_t flid)
1338 {
1339 	iflib_fl_t fl;
1340 
1341 	fl = &rxq->ifr_fl[flid];
1342 	iru->iru_paddrs = fl->ifl_bus_addrs;
1343 	iru->iru_idxs = fl->ifl_rxd_idxs;
1344 	iru->iru_qsidx = rxq->ifr_id;
1345 	iru->iru_buf_size = fl->ifl_buf_size;
1346 	iru->iru_flidx = fl->ifl_id;
1347 }
1348 
1349 static void
1350 _iflib_dmamap_cb(void *arg, bus_dma_segment_t *segs, int nseg, int err)
1351 {
1352 	if (err)
1353 		return;
1354 	*(bus_addr_t *) arg = segs[0].ds_addr;
1355 }
1356 
1357 int
1358 iflib_dma_alloc_align(if_ctx_t ctx, int size, int align, iflib_dma_info_t dma, int mapflags)
1359 {
1360 	int err;
1361 	device_t dev = ctx->ifc_dev;
1362 
1363 	err = bus_dma_tag_create(bus_get_dma_tag(dev),	/* parent */
1364 				align, 0,		/* alignment, bounds */
1365 				BUS_SPACE_MAXADDR,	/* lowaddr */
1366 				BUS_SPACE_MAXADDR,	/* highaddr */
1367 				NULL, NULL,		/* filter, filterarg */
1368 				size,			/* maxsize */
1369 				1,			/* nsegments */
1370 				size,			/* maxsegsize */
1371 				BUS_DMA_ALLOCNOW,	/* flags */
1372 				NULL,			/* lockfunc */
1373 				NULL,			/* lockarg */
1374 				&dma->idi_tag);
1375 	if (err) {
1376 		device_printf(dev,
1377 		    "%s: bus_dma_tag_create failed: %d\n",
1378 		    __func__, err);
1379 		goto fail_0;
1380 	}
1381 
1382 	err = bus_dmamem_alloc(dma->idi_tag, (void**) &dma->idi_vaddr,
1383 	    BUS_DMA_NOWAIT | BUS_DMA_COHERENT | BUS_DMA_ZERO, &dma->idi_map);
1384 	if (err) {
1385 		device_printf(dev,
1386 		    "%s: bus_dmamem_alloc(%ju) failed: %d\n",
1387 		    __func__, (uintmax_t)size, err);
1388 		goto fail_1;
1389 	}
1390 
1391 	dma->idi_paddr = IF_BAD_DMA;
1392 	err = bus_dmamap_load(dma->idi_tag, dma->idi_map, dma->idi_vaddr,
1393 	    size, _iflib_dmamap_cb, &dma->idi_paddr, mapflags | BUS_DMA_NOWAIT);
1394 	if (err || dma->idi_paddr == IF_BAD_DMA) {
1395 		device_printf(dev,
1396 		    "%s: bus_dmamap_load failed: %d\n",
1397 		    __func__, err);
1398 		goto fail_2;
1399 	}
1400 
1401 	dma->idi_size = size;
1402 	return (0);
1403 
1404 fail_2:
1405 	bus_dmamem_free(dma->idi_tag, dma->idi_vaddr, dma->idi_map);
1406 fail_1:
1407 	bus_dma_tag_destroy(dma->idi_tag);
1408 fail_0:
1409 	dma->idi_tag = NULL;
1410 
1411 	return (err);
1412 }
1413 
1414 int
1415 iflib_dma_alloc(if_ctx_t ctx, int size, iflib_dma_info_t dma, int mapflags)
1416 {
1417 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1418 
1419 	KASSERT(sctx->isc_q_align != 0, ("alignment value not initialized"));
1420 
1421 	return (iflib_dma_alloc_align(ctx, size, sctx->isc_q_align, dma, mapflags));
1422 }
1423 
1424 int
1425 iflib_dma_alloc_multi(if_ctx_t ctx, int *sizes, iflib_dma_info_t *dmalist, int mapflags, int count)
1426 {
1427 	int i, err;
1428 	iflib_dma_info_t *dmaiter;
1429 
1430 	dmaiter = dmalist;
1431 	for (i = 0; i < count; i++, dmaiter++) {
1432 		if ((err = iflib_dma_alloc(ctx, sizes[i], *dmaiter, mapflags)) != 0)
1433 			break;
1434 	}
1435 	if (err)
1436 		iflib_dma_free_multi(dmalist, i);
1437 	return (err);
1438 }
1439 
1440 void
1441 iflib_dma_free(iflib_dma_info_t dma)
1442 {
1443 	if (dma->idi_tag == NULL)
1444 		return;
1445 	if (dma->idi_paddr != IF_BAD_DMA) {
1446 		bus_dmamap_sync(dma->idi_tag, dma->idi_map,
1447 		    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
1448 		bus_dmamap_unload(dma->idi_tag, dma->idi_map);
1449 		dma->idi_paddr = IF_BAD_DMA;
1450 	}
1451 	if (dma->idi_vaddr != NULL) {
1452 		bus_dmamem_free(dma->idi_tag, dma->idi_vaddr, dma->idi_map);
1453 		dma->idi_vaddr = NULL;
1454 	}
1455 	bus_dma_tag_destroy(dma->idi_tag);
1456 	dma->idi_tag = NULL;
1457 }
1458 
1459 void
1460 iflib_dma_free_multi(iflib_dma_info_t *dmalist, int count)
1461 {
1462 	int i;
1463 	iflib_dma_info_t *dmaiter = dmalist;
1464 
1465 	for (i = 0; i < count; i++, dmaiter++)
1466 		iflib_dma_free(*dmaiter);
1467 }
1468 
1469 static int
1470 iflib_fast_intr(void *arg)
1471 {
1472 	iflib_filter_info_t info = arg;
1473 	struct grouptask *gtask = info->ifi_task;
1474 	int result;
1475 
1476 	DBG_COUNTER_INC(fast_intrs);
1477 	if (info->ifi_filter != NULL) {
1478 		result = info->ifi_filter(info->ifi_filter_arg);
1479 		if ((result & FILTER_SCHEDULE_THREAD) == 0)
1480 			return (result);
1481 	}
1482 
1483 	GROUPTASK_ENQUEUE(gtask);
1484 	return (FILTER_HANDLED);
1485 }
1486 
1487 static int
1488 iflib_fast_intr_rxtx(void *arg)
1489 {
1490 	iflib_filter_info_t info = arg;
1491 	struct grouptask *gtask = info->ifi_task;
1492 	if_ctx_t ctx;
1493 	iflib_rxq_t rxq = (iflib_rxq_t)info->ifi_ctx;
1494 	iflib_txq_t txq;
1495 	void *sc;
1496 	int i, cidx, result;
1497 	qidx_t txqid;
1498 	bool intr_enable, intr_legacy;
1499 
1500 	DBG_COUNTER_INC(fast_intrs);
1501 	if (info->ifi_filter != NULL) {
1502 		result = info->ifi_filter(info->ifi_filter_arg);
1503 		if ((result & FILTER_SCHEDULE_THREAD) == 0)
1504 			return (result);
1505 	}
1506 
1507 	ctx = rxq->ifr_ctx;
1508 	sc = ctx->ifc_softc;
1509 	intr_enable = false;
1510 	intr_legacy = !!(ctx->ifc_flags & IFC_LEGACY);
1511 	MPASS(rxq->ifr_ntxqirq);
1512 	for (i = 0; i < rxq->ifr_ntxqirq; i++) {
1513 		txqid = rxq->ifr_txqid[i];
1514 		txq = &ctx->ifc_txqs[txqid];
1515 		bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
1516 		    BUS_DMASYNC_POSTREAD);
1517 		if (!ctx->isc_txd_credits_update(sc, txqid, false)) {
1518 			if (intr_legacy)
1519 				intr_enable = true;
1520 			else
1521 				IFDI_TX_QUEUE_INTR_ENABLE(ctx, txqid);
1522 			continue;
1523 		}
1524 		GROUPTASK_ENQUEUE(&txq->ift_task);
1525 	}
1526 	if (ctx->ifc_sctx->isc_flags & IFLIB_HAS_RXCQ)
1527 		cidx = rxq->ifr_cq_cidx;
1528 	else
1529 		cidx = rxq->ifr_fl[0].ifl_cidx;
1530 	if (iflib_rxd_avail(ctx, rxq, cidx, 1))
1531 		GROUPTASK_ENQUEUE(gtask);
1532 	else {
1533 		if (intr_legacy)
1534 			intr_enable = true;
1535 		else
1536 			IFDI_RX_QUEUE_INTR_ENABLE(ctx, rxq->ifr_id);
1537 		DBG_COUNTER_INC(rx_intr_enables);
1538 	}
1539 	if (intr_enable)
1540 		IFDI_INTR_ENABLE(ctx);
1541 	return (FILTER_HANDLED);
1542 }
1543 
1544 static int
1545 iflib_fast_intr_ctx(void *arg)
1546 {
1547 	iflib_filter_info_t info = arg;
1548 	struct grouptask *gtask = info->ifi_task;
1549 	int result;
1550 
1551 	DBG_COUNTER_INC(fast_intrs);
1552 	if (info->ifi_filter != NULL) {
1553 		result = info->ifi_filter(info->ifi_filter_arg);
1554 		if ((result & FILTER_SCHEDULE_THREAD) == 0)
1555 			return (result);
1556 	}
1557 
1558 	GROUPTASK_ENQUEUE(gtask);
1559 	return (FILTER_HANDLED);
1560 }
1561 
1562 static int
1563 _iflib_irq_alloc(if_ctx_t ctx, if_irq_t irq, int rid,
1564 		 driver_filter_t filter, driver_intr_t handler, void *arg,
1565 		 const char *name)
1566 {
1567 	struct resource *res;
1568 	void *tag = NULL;
1569 	device_t dev = ctx->ifc_dev;
1570 	int flags, i, rc;
1571 
1572 	flags = RF_ACTIVE;
1573 	if (ctx->ifc_flags & IFC_LEGACY)
1574 		flags |= RF_SHAREABLE;
1575 	MPASS(rid < 512);
1576 	i = rid;
1577 	res = bus_alloc_resource_any(dev, SYS_RES_IRQ, &i, flags);
1578 	if (res == NULL) {
1579 		device_printf(dev,
1580 		    "failed to allocate IRQ for rid %d, name %s.\n", rid, name);
1581 		return (ENOMEM);
1582 	}
1583 	irq->ii_res = res;
1584 	KASSERT(filter == NULL || handler == NULL, ("filter and handler can't both be non-NULL"));
1585 	rc = bus_setup_intr(dev, res, INTR_MPSAFE | INTR_TYPE_NET,
1586 						filter, handler, arg, &tag);
1587 	if (rc != 0) {
1588 		device_printf(dev,
1589 		    "failed to setup interrupt for rid %d, name %s: %d\n",
1590 					  rid, name ? name : "unknown", rc);
1591 		return (rc);
1592 	} else if (name)
1593 		bus_describe_intr(dev, res, tag, "%s", name);
1594 
1595 	irq->ii_tag = tag;
1596 	return (0);
1597 }
1598 
1599 /*********************************************************************
1600  *
1601  *  Allocate DMA resources for TX buffers as well as memory for the TX
1602  *  mbuf map.  TX DMA maps (non-TSO/TSO) and TX mbuf map are kept in a
1603  *  iflib_sw_tx_desc_array structure, storing all the information that
1604  *  is needed to transmit a packet on the wire.  This is called only
1605  *  once at attach, setup is done every reset.
1606  *
1607  **********************************************************************/
1608 static int
1609 iflib_txsd_alloc(iflib_txq_t txq)
1610 {
1611 	if_ctx_t ctx = txq->ift_ctx;
1612 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1613 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1614 	device_t dev = ctx->ifc_dev;
1615 	bus_size_t tsomaxsize;
1616 	int err, nsegments, ntsosegments;
1617 	bool tso;
1618 
1619 	nsegments = scctx->isc_tx_nsegments;
1620 	ntsosegments = scctx->isc_tx_tso_segments_max;
1621 	tsomaxsize = scctx->isc_tx_tso_size_max;
1622 	if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_VLAN_MTU)
1623 		tsomaxsize += sizeof(struct ether_vlan_header);
1624 	MPASS(scctx->isc_ntxd[0] > 0);
1625 	MPASS(scctx->isc_ntxd[txq->ift_br_offset] > 0);
1626 	MPASS(nsegments > 0);
1627 	if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_TSO) {
1628 		MPASS(ntsosegments > 0);
1629 		MPASS(sctx->isc_tso_maxsize >= tsomaxsize);
1630 	}
1631 
1632 	/*
1633 	 * Set up DMA tags for TX buffers.
1634 	 */
1635 	if ((err = bus_dma_tag_create(bus_get_dma_tag(dev),
1636 			       1, 0,			/* alignment, bounds */
1637 			       BUS_SPACE_MAXADDR,	/* lowaddr */
1638 			       BUS_SPACE_MAXADDR,	/* highaddr */
1639 			       NULL, NULL,		/* filter, filterarg */
1640 			       sctx->isc_tx_maxsize,		/* maxsize */
1641 			       nsegments,	/* nsegments */
1642 			       sctx->isc_tx_maxsegsize,	/* maxsegsize */
1643 			       0,			/* flags */
1644 			       NULL,			/* lockfunc */
1645 			       NULL,			/* lockfuncarg */
1646 			       &txq->ift_buf_tag))) {
1647 		device_printf(dev,"Unable to allocate TX DMA tag: %d\n", err);
1648 		device_printf(dev,"maxsize: %ju nsegments: %d maxsegsize: %ju\n",
1649 		    (uintmax_t)sctx->isc_tx_maxsize, nsegments, (uintmax_t)sctx->isc_tx_maxsegsize);
1650 		goto fail;
1651 	}
1652 	tso = (if_getcapabilities(ctx->ifc_ifp) & IFCAP_TSO) != 0;
1653 	if (tso && (err = bus_dma_tag_create(bus_get_dma_tag(dev),
1654 			       1, 0,			/* alignment, bounds */
1655 			       BUS_SPACE_MAXADDR,	/* lowaddr */
1656 			       BUS_SPACE_MAXADDR,	/* highaddr */
1657 			       NULL, NULL,		/* filter, filterarg */
1658 			       tsomaxsize,		/* maxsize */
1659 			       ntsosegments,	/* nsegments */
1660 			       sctx->isc_tso_maxsegsize,/* maxsegsize */
1661 			       0,			/* flags */
1662 			       NULL,			/* lockfunc */
1663 			       NULL,			/* lockfuncarg */
1664 			       &txq->ift_tso_buf_tag))) {
1665 		device_printf(dev, "Unable to allocate TSO TX DMA tag: %d\n",
1666 		    err);
1667 		goto fail;
1668 	}
1669 
1670 	/* Allocate memory for the TX mbuf map. */
1671 	if (!(txq->ift_sds.ifsd_m =
1672 	    (struct mbuf **) malloc(sizeof(struct mbuf *) *
1673 	    scctx->isc_ntxd[txq->ift_br_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1674 		device_printf(dev, "Unable to allocate TX mbuf map memory\n");
1675 		err = ENOMEM;
1676 		goto fail;
1677 	}
1678 
1679 	/*
1680 	 * Create the DMA maps for TX buffers.
1681 	 */
1682 	if ((txq->ift_sds.ifsd_map = (bus_dmamap_t *)malloc(
1683 	    sizeof(bus_dmamap_t) * scctx->isc_ntxd[txq->ift_br_offset],
1684 	    M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
1685 		device_printf(dev,
1686 		    "Unable to allocate TX buffer DMA map memory\n");
1687 		err = ENOMEM;
1688 		goto fail;
1689 	}
1690 	if (tso && (txq->ift_sds.ifsd_tso_map = (bus_dmamap_t *)malloc(
1691 	    sizeof(bus_dmamap_t) * scctx->isc_ntxd[txq->ift_br_offset],
1692 	    M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
1693 		device_printf(dev,
1694 		    "Unable to allocate TSO TX buffer map memory\n");
1695 		err = ENOMEM;
1696 		goto fail;
1697 	}
1698 	for (int i = 0; i < scctx->isc_ntxd[txq->ift_br_offset]; i++) {
1699 		err = bus_dmamap_create(txq->ift_buf_tag, 0,
1700 		    &txq->ift_sds.ifsd_map[i]);
1701 		if (err != 0) {
1702 			device_printf(dev, "Unable to create TX DMA map\n");
1703 			goto fail;
1704 		}
1705 		if (!tso)
1706 			continue;
1707 		err = bus_dmamap_create(txq->ift_tso_buf_tag, 0,
1708 		    &txq->ift_sds.ifsd_tso_map[i]);
1709 		if (err != 0) {
1710 			device_printf(dev, "Unable to create TSO TX DMA map\n");
1711 			goto fail;
1712 		}
1713 	}
1714 	return (0);
1715 fail:
1716 	/* We free all, it handles case where we are in the middle */
1717 	iflib_tx_structures_free(ctx);
1718 	return (err);
1719 }
1720 
1721 static void
1722 iflib_txsd_destroy(if_ctx_t ctx, iflib_txq_t txq, int i)
1723 {
1724 	bus_dmamap_t map;
1725 
1726 	if (txq->ift_sds.ifsd_map != NULL) {
1727 		map = txq->ift_sds.ifsd_map[i];
1728 		bus_dmamap_sync(txq->ift_buf_tag, map, BUS_DMASYNC_POSTWRITE);
1729 		bus_dmamap_unload(txq->ift_buf_tag, map);
1730 		bus_dmamap_destroy(txq->ift_buf_tag, map);
1731 		txq->ift_sds.ifsd_map[i] = NULL;
1732 	}
1733 
1734 	if (txq->ift_sds.ifsd_tso_map != NULL) {
1735 		map = txq->ift_sds.ifsd_tso_map[i];
1736 		bus_dmamap_sync(txq->ift_tso_buf_tag, map,
1737 		    BUS_DMASYNC_POSTWRITE);
1738 		bus_dmamap_unload(txq->ift_tso_buf_tag, map);
1739 		bus_dmamap_destroy(txq->ift_tso_buf_tag, map);
1740 		txq->ift_sds.ifsd_tso_map[i] = NULL;
1741 	}
1742 }
1743 
1744 static void
1745 iflib_txq_destroy(iflib_txq_t txq)
1746 {
1747 	if_ctx_t ctx = txq->ift_ctx;
1748 
1749 	for (int i = 0; i < txq->ift_size; i++)
1750 		iflib_txsd_destroy(ctx, txq, i);
1751 
1752 	if (txq->ift_br != NULL) {
1753 		ifmp_ring_free(txq->ift_br);
1754 		txq->ift_br = NULL;
1755 	}
1756 
1757 	mtx_destroy(&txq->ift_mtx);
1758 
1759 	if (txq->ift_sds.ifsd_map != NULL) {
1760 		free(txq->ift_sds.ifsd_map, M_IFLIB);
1761 		txq->ift_sds.ifsd_map = NULL;
1762 	}
1763 	if (txq->ift_sds.ifsd_tso_map != NULL) {
1764 		free(txq->ift_sds.ifsd_tso_map, M_IFLIB);
1765 		txq->ift_sds.ifsd_tso_map = NULL;
1766 	}
1767 	if (txq->ift_sds.ifsd_m != NULL) {
1768 		free(txq->ift_sds.ifsd_m, M_IFLIB);
1769 		txq->ift_sds.ifsd_m = NULL;
1770 	}
1771 	if (txq->ift_buf_tag != NULL) {
1772 		bus_dma_tag_destroy(txq->ift_buf_tag);
1773 		txq->ift_buf_tag = NULL;
1774 	}
1775 	if (txq->ift_tso_buf_tag != NULL) {
1776 		bus_dma_tag_destroy(txq->ift_tso_buf_tag);
1777 		txq->ift_tso_buf_tag = NULL;
1778 	}
1779 	if (txq->ift_ifdi != NULL) {
1780 		free(txq->ift_ifdi, M_IFLIB);
1781 	}
1782 }
1783 
1784 static void
1785 iflib_txsd_free(if_ctx_t ctx, iflib_txq_t txq, int i)
1786 {
1787 	struct mbuf **mp;
1788 
1789 	mp = &txq->ift_sds.ifsd_m[i];
1790 	if (*mp == NULL)
1791 		return;
1792 
1793 	if (txq->ift_sds.ifsd_map != NULL) {
1794 		bus_dmamap_sync(txq->ift_buf_tag,
1795 		    txq->ift_sds.ifsd_map[i], BUS_DMASYNC_POSTWRITE);
1796 		bus_dmamap_unload(txq->ift_buf_tag, txq->ift_sds.ifsd_map[i]);
1797 	}
1798 	if (txq->ift_sds.ifsd_tso_map != NULL) {
1799 		bus_dmamap_sync(txq->ift_tso_buf_tag,
1800 		    txq->ift_sds.ifsd_tso_map[i], BUS_DMASYNC_POSTWRITE);
1801 		bus_dmamap_unload(txq->ift_tso_buf_tag,
1802 		    txq->ift_sds.ifsd_tso_map[i]);
1803 	}
1804 	m_freem(*mp);
1805 	DBG_COUNTER_INC(tx_frees);
1806 	*mp = NULL;
1807 }
1808 
1809 static int
1810 iflib_txq_setup(iflib_txq_t txq)
1811 {
1812 	if_ctx_t ctx = txq->ift_ctx;
1813 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1814 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1815 	iflib_dma_info_t di;
1816 	int i;
1817 
1818 	/* Set number of descriptors available */
1819 	txq->ift_qstatus = IFLIB_QUEUE_IDLE;
1820 	/* XXX make configurable */
1821 	txq->ift_update_freq = IFLIB_DEFAULT_TX_UPDATE_FREQ;
1822 
1823 	/* Reset indices */
1824 	txq->ift_cidx_processed = 0;
1825 	txq->ift_pidx = txq->ift_cidx = txq->ift_npending = 0;
1826 	txq->ift_size = scctx->isc_ntxd[txq->ift_br_offset];
1827 
1828 	for (i = 0, di = txq->ift_ifdi; i < sctx->isc_ntxqs; i++, di++)
1829 		bzero((void *)di->idi_vaddr, di->idi_size);
1830 
1831 	IFDI_TXQ_SETUP(ctx, txq->ift_id);
1832 	for (i = 0, di = txq->ift_ifdi; i < sctx->isc_ntxqs; i++, di++)
1833 		bus_dmamap_sync(di->idi_tag, di->idi_map,
1834 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
1835 	return (0);
1836 }
1837 
1838 /*********************************************************************
1839  *
1840  *  Allocate DMA resources for RX buffers as well as memory for the RX
1841  *  mbuf map, direct RX cluster pointer map and RX cluster bus address
1842  *  map.  RX DMA map, RX mbuf map, direct RX cluster pointer map and
1843  *  RX cluster map are kept in a iflib_sw_rx_desc_array structure.
1844  *  Since we use use one entry in iflib_sw_rx_desc_array per received
1845  *  packet, the maximum number of entries we'll need is equal to the
1846  *  number of hardware receive descriptors that we've allocated.
1847  *
1848  **********************************************************************/
1849 static int
1850 iflib_rxsd_alloc(iflib_rxq_t rxq)
1851 {
1852 	if_ctx_t ctx = rxq->ifr_ctx;
1853 	if_shared_ctx_t sctx = ctx->ifc_sctx;
1854 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
1855 	device_t dev = ctx->ifc_dev;
1856 	iflib_fl_t fl;
1857 	int			err;
1858 
1859 	MPASS(scctx->isc_nrxd[0] > 0);
1860 	MPASS(scctx->isc_nrxd[rxq->ifr_fl_offset] > 0);
1861 
1862 	fl = rxq->ifr_fl;
1863 	for (int i = 0; i <  rxq->ifr_nfl; i++, fl++) {
1864 		fl->ifl_size = scctx->isc_nrxd[rxq->ifr_fl_offset]; /* this isn't necessarily the same */
1865 		/* Set up DMA tag for RX buffers. */
1866 		err = bus_dma_tag_create(bus_get_dma_tag(dev), /* parent */
1867 					 1, 0,			/* alignment, bounds */
1868 					 BUS_SPACE_MAXADDR,	/* lowaddr */
1869 					 BUS_SPACE_MAXADDR,	/* highaddr */
1870 					 NULL, NULL,		/* filter, filterarg */
1871 					 sctx->isc_rx_maxsize,	/* maxsize */
1872 					 sctx->isc_rx_nsegments,	/* nsegments */
1873 					 sctx->isc_rx_maxsegsize,	/* maxsegsize */
1874 					 0,			/* flags */
1875 					 NULL,			/* lockfunc */
1876 					 NULL,			/* lockarg */
1877 					 &fl->ifl_buf_tag);
1878 		if (err) {
1879 			device_printf(dev,
1880 			    "Unable to allocate RX DMA tag: %d\n", err);
1881 			goto fail;
1882 		}
1883 
1884 		/* Allocate memory for the RX mbuf map. */
1885 		if (!(fl->ifl_sds.ifsd_m =
1886 		      (struct mbuf **) malloc(sizeof(struct mbuf *) *
1887 					      scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1888 			device_printf(dev,
1889 			    "Unable to allocate RX mbuf map memory\n");
1890 			err = ENOMEM;
1891 			goto fail;
1892 		}
1893 
1894 		/* Allocate memory for the direct RX cluster pointer map. */
1895 		if (!(fl->ifl_sds.ifsd_cl =
1896 		      (caddr_t *) malloc(sizeof(caddr_t) *
1897 					      scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1898 			device_printf(dev,
1899 			    "Unable to allocate RX cluster map memory\n");
1900 			err = ENOMEM;
1901 			goto fail;
1902 		}
1903 
1904 		/* Allocate memory for the RX cluster bus address map. */
1905 		if (!(fl->ifl_sds.ifsd_ba =
1906 		      (bus_addr_t *) malloc(sizeof(bus_addr_t) *
1907 					      scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1908 			device_printf(dev,
1909 			    "Unable to allocate RX bus address map memory\n");
1910 			err = ENOMEM;
1911 			goto fail;
1912 		}
1913 
1914 		/*
1915 		 * Create the DMA maps for RX buffers.
1916 		 */
1917 		if (!(fl->ifl_sds.ifsd_map =
1918 		      (bus_dmamap_t *) malloc(sizeof(bus_dmamap_t) * scctx->isc_nrxd[rxq->ifr_fl_offset], M_IFLIB, M_NOWAIT | M_ZERO))) {
1919 			device_printf(dev,
1920 			    "Unable to allocate RX buffer DMA map memory\n");
1921 			err = ENOMEM;
1922 			goto fail;
1923 		}
1924 		for (int i = 0; i < scctx->isc_nrxd[rxq->ifr_fl_offset]; i++) {
1925 			err = bus_dmamap_create(fl->ifl_buf_tag, 0,
1926 			    &fl->ifl_sds.ifsd_map[i]);
1927 			if (err != 0) {
1928 				device_printf(dev, "Unable to create RX buffer DMA map\n");
1929 				goto fail;
1930 			}
1931 		}
1932 	}
1933 	return (0);
1934 
1935 fail:
1936 	iflib_rx_structures_free(ctx);
1937 	return (err);
1938 }
1939 
1940 /*
1941  * Internal service routines
1942  */
1943 
1944 struct rxq_refill_cb_arg {
1945 	int               error;
1946 	bus_dma_segment_t seg;
1947 	int               nseg;
1948 };
1949 
1950 static void
1951 _rxq_refill_cb(void *arg, bus_dma_segment_t *segs, int nseg, int error)
1952 {
1953 	struct rxq_refill_cb_arg *cb_arg = arg;
1954 
1955 	cb_arg->error = error;
1956 	cb_arg->seg = segs[0];
1957 	cb_arg->nseg = nseg;
1958 }
1959 
1960 /**
1961  * iflib_fl_refill - refill an rxq free-buffer list
1962  * @ctx: the iflib context
1963  * @fl: the free list to refill
1964  * @count: the number of new buffers to allocate
1965  *
1966  * (Re)populate an rxq free-buffer list with up to @count new packet buffers.
1967  * The caller must assure that @count does not exceed the queue's capacity
1968  * minus one (since we always leave a descriptor unavailable).
1969  */
1970 static uint8_t
1971 iflib_fl_refill(if_ctx_t ctx, iflib_fl_t fl, int count)
1972 {
1973 	struct if_rxd_update iru;
1974 	struct rxq_refill_cb_arg cb_arg;
1975 	struct mbuf *m;
1976 	caddr_t cl, *sd_cl;
1977 	struct mbuf **sd_m;
1978 	bus_dmamap_t *sd_map;
1979 	bus_addr_t bus_addr, *sd_ba;
1980 	int err, frag_idx, i, idx, n, pidx;
1981 	qidx_t credits;
1982 
1983 	MPASS(count <= fl->ifl_size - fl->ifl_credits - 1);
1984 
1985 	sd_m = fl->ifl_sds.ifsd_m;
1986 	sd_map = fl->ifl_sds.ifsd_map;
1987 	sd_cl = fl->ifl_sds.ifsd_cl;
1988 	sd_ba = fl->ifl_sds.ifsd_ba;
1989 	pidx = fl->ifl_pidx;
1990 	idx = pidx;
1991 	frag_idx = fl->ifl_fragidx;
1992 	credits = fl->ifl_credits;
1993 
1994 	i = 0;
1995 	n = count;
1996 	MPASS(n > 0);
1997 	MPASS(credits + n <= fl->ifl_size);
1998 
1999 	if (pidx < fl->ifl_cidx)
2000 		MPASS(pidx + n <= fl->ifl_cidx);
2001 	if (pidx == fl->ifl_cidx && (credits < fl->ifl_size))
2002 		MPASS(fl->ifl_gen == 0);
2003 	if (pidx > fl->ifl_cidx)
2004 		MPASS(n <= fl->ifl_size - pidx + fl->ifl_cidx);
2005 
2006 	DBG_COUNTER_INC(fl_refills);
2007 	if (n > 8)
2008 		DBG_COUNTER_INC(fl_refills_large);
2009 	iru_init(&iru, fl->ifl_rxq, fl->ifl_id);
2010 	while (n-- > 0) {
2011 		/*
2012 		 * We allocate an uninitialized mbuf + cluster, mbuf is
2013 		 * initialized after rx.
2014 		 *
2015 		 * If the cluster is still set then we know a minimum sized
2016 		 * packet was received
2017 		 */
2018 		bit_ffc_at(fl->ifl_rx_bitmap, frag_idx, fl->ifl_size,
2019 		    &frag_idx);
2020 		if (frag_idx < 0)
2021 			bit_ffc(fl->ifl_rx_bitmap, fl->ifl_size, &frag_idx);
2022 		MPASS(frag_idx >= 0);
2023 		if ((cl = sd_cl[frag_idx]) == NULL) {
2024 			cl = uma_zalloc(fl->ifl_zone, M_NOWAIT);
2025 			if (__predict_false(cl == NULL))
2026 				break;
2027 
2028 			cb_arg.error = 0;
2029 			MPASS(sd_map != NULL);
2030 			err = bus_dmamap_load(fl->ifl_buf_tag, sd_map[frag_idx],
2031 			    cl, fl->ifl_buf_size, _rxq_refill_cb, &cb_arg,
2032 			    BUS_DMA_NOWAIT);
2033 			if (__predict_false(err != 0 || cb_arg.error)) {
2034 				uma_zfree(fl->ifl_zone, cl);
2035 				break;
2036 			}
2037 
2038 			sd_ba[frag_idx] = bus_addr = cb_arg.seg.ds_addr;
2039 			sd_cl[frag_idx] = cl;
2040 #if MEMORY_LOGGING
2041 			fl->ifl_cl_enqueued++;
2042 #endif
2043 		} else {
2044 			bus_addr = sd_ba[frag_idx];
2045 		}
2046 		bus_dmamap_sync(fl->ifl_buf_tag, sd_map[frag_idx],
2047 		    BUS_DMASYNC_PREREAD);
2048 
2049 		if (sd_m[frag_idx] == NULL) {
2050 			m = m_gethdr(M_NOWAIT, MT_NOINIT);
2051 			if (__predict_false(m == NULL))
2052 				break;
2053 			sd_m[frag_idx] = m;
2054 		}
2055 		bit_set(fl->ifl_rx_bitmap, frag_idx);
2056 #if MEMORY_LOGGING
2057 		fl->ifl_m_enqueued++;
2058 #endif
2059 
2060 		DBG_COUNTER_INC(rx_allocs);
2061 		fl->ifl_rxd_idxs[i] = frag_idx;
2062 		fl->ifl_bus_addrs[i] = bus_addr;
2063 		credits++;
2064 		i++;
2065 		MPASS(credits <= fl->ifl_size);
2066 		if (++idx == fl->ifl_size) {
2067 #ifdef INVARIANTS
2068 			fl->ifl_gen = 1;
2069 #endif
2070 			idx = 0;
2071 		}
2072 		if (n == 0 || i == IFLIB_MAX_RX_REFRESH) {
2073 			iru.iru_pidx = pidx;
2074 			iru.iru_count = i;
2075 			ctx->isc_rxd_refill(ctx->ifc_softc, &iru);
2076 			fl->ifl_pidx = idx;
2077 			fl->ifl_credits = credits;
2078 			pidx = idx;
2079 			i = 0;
2080 		}
2081 	}
2082 
2083 	if (n < count - 1) {
2084 		if (i != 0) {
2085 			iru.iru_pidx = pidx;
2086 			iru.iru_count = i;
2087 			ctx->isc_rxd_refill(ctx->ifc_softc, &iru);
2088 			fl->ifl_pidx = idx;
2089 			fl->ifl_credits = credits;
2090 		}
2091 		DBG_COUNTER_INC(rxd_flush);
2092 		bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
2093 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
2094 		ctx->isc_rxd_flush(ctx->ifc_softc, fl->ifl_rxq->ifr_id,
2095 		    fl->ifl_id, fl->ifl_pidx);
2096 		if (__predict_true(bit_test(fl->ifl_rx_bitmap, frag_idx))) {
2097 			fl->ifl_fragidx = frag_idx + 1;
2098 			if (fl->ifl_fragidx == fl->ifl_size)
2099 				fl->ifl_fragidx = 0;
2100 		} else {
2101 			fl->ifl_fragidx = frag_idx;
2102 		}
2103 	}
2104 
2105 	return (n == -1 ? 0 : IFLIB_RXEOF_EMPTY);
2106 }
2107 
2108 static inline uint8_t
2109 iflib_fl_refill_all(if_ctx_t ctx, iflib_fl_t fl)
2110 {
2111 	/*
2112 	 * We leave an unused descriptor to avoid pidx to catch up with cidx.
2113 	 * This is important as it confuses most NICs. For instance,
2114 	 * Intel NICs have (per receive ring) RDH and RDT registers, where
2115 	 * RDH points to the next receive descriptor to be used by the NIC,
2116 	 * and RDT for the next receive descriptor to be published by the
2117 	 * driver to the NIC (RDT - 1 is thus the last valid one).
2118 	 * The condition RDH == RDT means no descriptors are available to
2119 	 * the NIC, and thus it would be ambiguous if it also meant that
2120 	 * all the descriptors are available to the NIC.
2121 	 */
2122 	int32_t reclaimable = fl->ifl_size - fl->ifl_credits - 1;
2123 #ifdef INVARIANTS
2124 	int32_t delta = fl->ifl_size - get_inuse(fl->ifl_size, fl->ifl_cidx, fl->ifl_pidx, fl->ifl_gen) - 1;
2125 #endif
2126 
2127 	MPASS(fl->ifl_credits <= fl->ifl_size);
2128 	MPASS(reclaimable == delta);
2129 
2130 	if (reclaimable > 0)
2131 		return (iflib_fl_refill(ctx, fl, reclaimable));
2132 	return (0);
2133 }
2134 
2135 uint8_t
2136 iflib_in_detach(if_ctx_t ctx)
2137 {
2138 	bool in_detach;
2139 
2140 	STATE_LOCK(ctx);
2141 	in_detach = !!(ctx->ifc_flags & IFC_IN_DETACH);
2142 	STATE_UNLOCK(ctx);
2143 	return (in_detach);
2144 }
2145 
2146 static void
2147 iflib_fl_bufs_free(iflib_fl_t fl)
2148 {
2149 	iflib_dma_info_t idi = fl->ifl_ifdi;
2150 	bus_dmamap_t sd_map;
2151 	uint32_t i;
2152 
2153 	for (i = 0; i < fl->ifl_size; i++) {
2154 		struct mbuf **sd_m = &fl->ifl_sds.ifsd_m[i];
2155 		caddr_t *sd_cl = &fl->ifl_sds.ifsd_cl[i];
2156 
2157 		if (*sd_cl != NULL) {
2158 			sd_map = fl->ifl_sds.ifsd_map[i];
2159 			bus_dmamap_sync(fl->ifl_buf_tag, sd_map,
2160 			    BUS_DMASYNC_POSTREAD);
2161 			bus_dmamap_unload(fl->ifl_buf_tag, sd_map);
2162 			uma_zfree(fl->ifl_zone, *sd_cl);
2163 			*sd_cl = NULL;
2164 			if (*sd_m != NULL) {
2165 				m_init(*sd_m, M_NOWAIT, MT_DATA, 0);
2166 				uma_zfree(zone_mbuf, *sd_m);
2167 				*sd_m = NULL;
2168 			}
2169 		} else {
2170 			MPASS(*sd_m == NULL);
2171 		}
2172 #if MEMORY_LOGGING
2173 		fl->ifl_m_dequeued++;
2174 		fl->ifl_cl_dequeued++;
2175 #endif
2176 	}
2177 #ifdef INVARIANTS
2178 	for (i = 0; i < fl->ifl_size; i++) {
2179 		MPASS(fl->ifl_sds.ifsd_cl[i] == NULL);
2180 		MPASS(fl->ifl_sds.ifsd_m[i] == NULL);
2181 	}
2182 #endif
2183 	/*
2184 	 * Reset free list values
2185 	 */
2186 	fl->ifl_credits = fl->ifl_cidx = fl->ifl_pidx = fl->ifl_gen = fl->ifl_fragidx = 0;
2187 	bzero(idi->idi_vaddr, idi->idi_size);
2188 }
2189 
2190 /*********************************************************************
2191  *
2192  *  Initialize a free list and its buffers.
2193  *
2194  **********************************************************************/
2195 static int
2196 iflib_fl_setup(iflib_fl_t fl)
2197 {
2198 	iflib_rxq_t rxq = fl->ifl_rxq;
2199 	if_ctx_t ctx = rxq->ifr_ctx;
2200 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2201 	int qidx;
2202 
2203 	bit_nclear(fl->ifl_rx_bitmap, 0, fl->ifl_size - 1);
2204 	/*
2205 	** Free current RX buffer structs and their mbufs
2206 	*/
2207 	iflib_fl_bufs_free(fl);
2208 	/* Now replenish the mbufs */
2209 	MPASS(fl->ifl_credits == 0);
2210 	qidx = rxq->ifr_fl_offset + fl->ifl_id;
2211 	if (scctx->isc_rxd_buf_size[qidx] != 0)
2212 		fl->ifl_buf_size = scctx->isc_rxd_buf_size[qidx];
2213 	else
2214 		fl->ifl_buf_size = ctx->ifc_rx_mbuf_sz;
2215 	/*
2216 	 * ifl_buf_size may be a driver-supplied value, so pull it up
2217 	 * to the selected mbuf size.
2218 	 */
2219 	fl->ifl_buf_size = iflib_get_mbuf_size_for(fl->ifl_buf_size);
2220 	if (fl->ifl_buf_size > ctx->ifc_max_fl_buf_size)
2221 		ctx->ifc_max_fl_buf_size = fl->ifl_buf_size;
2222 	fl->ifl_cltype = m_gettype(fl->ifl_buf_size);
2223 	fl->ifl_zone = m_getzone(fl->ifl_buf_size);
2224 
2225 	/*
2226 	 * Avoid pre-allocating zillions of clusters to an idle card
2227 	 * potentially speeding up attach. In any case make sure
2228 	 * to leave a descriptor unavailable. See the comment in
2229 	 * iflib_fl_refill_all().
2230 	 */
2231 	MPASS(fl->ifl_size > 0);
2232 	(void)iflib_fl_refill(ctx, fl, min(128, fl->ifl_size - 1));
2233 	if (min(128, fl->ifl_size - 1) != fl->ifl_credits)
2234 		return (ENOBUFS);
2235 	/*
2236 	 * handle failure
2237 	 */
2238 	MPASS(rxq != NULL);
2239 	MPASS(fl->ifl_ifdi != NULL);
2240 	bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
2241 	    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
2242 	return (0);
2243 }
2244 
2245 /*********************************************************************
2246  *
2247  *  Free receive ring data structures
2248  *
2249  **********************************************************************/
2250 static void
2251 iflib_rx_sds_free(iflib_rxq_t rxq)
2252 {
2253 	iflib_fl_t fl;
2254 	int i, j;
2255 
2256 	if (rxq->ifr_fl != NULL) {
2257 		for (i = 0; i < rxq->ifr_nfl; i++) {
2258 			fl = &rxq->ifr_fl[i];
2259 			if (fl->ifl_buf_tag != NULL) {
2260 				if (fl->ifl_sds.ifsd_map != NULL) {
2261 					for (j = 0; j < fl->ifl_size; j++) {
2262 						bus_dmamap_sync(
2263 						    fl->ifl_buf_tag,
2264 						    fl->ifl_sds.ifsd_map[j],
2265 						    BUS_DMASYNC_POSTREAD);
2266 						bus_dmamap_unload(
2267 						    fl->ifl_buf_tag,
2268 						    fl->ifl_sds.ifsd_map[j]);
2269 						bus_dmamap_destroy(
2270 						    fl->ifl_buf_tag,
2271 						    fl->ifl_sds.ifsd_map[j]);
2272 					}
2273 				}
2274 				bus_dma_tag_destroy(fl->ifl_buf_tag);
2275 				fl->ifl_buf_tag = NULL;
2276 			}
2277 			free(fl->ifl_sds.ifsd_m, M_IFLIB);
2278 			free(fl->ifl_sds.ifsd_cl, M_IFLIB);
2279 			free(fl->ifl_sds.ifsd_ba, M_IFLIB);
2280 			free(fl->ifl_sds.ifsd_map, M_IFLIB);
2281 			free(fl->ifl_rx_bitmap, M_IFLIB);
2282 			fl->ifl_sds.ifsd_m = NULL;
2283 			fl->ifl_sds.ifsd_cl = NULL;
2284 			fl->ifl_sds.ifsd_ba = NULL;
2285 			fl->ifl_sds.ifsd_map = NULL;
2286 			fl->ifl_rx_bitmap = NULL;
2287 		}
2288 		free(rxq->ifr_fl, M_IFLIB);
2289 		rxq->ifr_fl = NULL;
2290 		free(rxq->ifr_ifdi, M_IFLIB);
2291 		rxq->ifr_ifdi = NULL;
2292 		rxq->ifr_cq_cidx = 0;
2293 	}
2294 }
2295 
2296 /*
2297  * Timer routine
2298  */
2299 static void
2300 iflib_timer(void *arg)
2301 {
2302 	iflib_txq_t txq = arg;
2303 	if_ctx_t ctx = txq->ift_ctx;
2304 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
2305 	uint64_t this_tick = ticks;
2306 
2307 	if (!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING))
2308 		return;
2309 
2310 	/*
2311 	** Check on the state of the TX queue(s), this
2312 	** can be done without the lock because its RO
2313 	** and the HUNG state will be static if set.
2314 	*/
2315 	if (this_tick - txq->ift_last_timer_tick >= iflib_timer_default) {
2316 		txq->ift_last_timer_tick = this_tick;
2317 		IFDI_TIMER(ctx, txq->ift_id);
2318 		if ((txq->ift_qstatus == IFLIB_QUEUE_HUNG) &&
2319 		    ((txq->ift_cleaned_prev == txq->ift_cleaned) ||
2320 		     (sctx->isc_pause_frames == 0)))
2321 			goto hung;
2322 
2323 		if (txq->ift_qstatus != IFLIB_QUEUE_IDLE &&
2324 		    ifmp_ring_is_stalled(txq->ift_br)) {
2325 			KASSERT(ctx->ifc_link_state == LINK_STATE_UP,
2326 			    ("queue can't be marked as hung if interface is down"));
2327 			txq->ift_qstatus = IFLIB_QUEUE_HUNG;
2328 		}
2329 		txq->ift_cleaned_prev = txq->ift_cleaned;
2330 	}
2331 	/* handle any laggards */
2332 	if (txq->ift_db_pending)
2333 		GROUPTASK_ENQUEUE(&txq->ift_task);
2334 
2335 	sctx->isc_pause_frames = 0;
2336 	if (if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING)
2337 		callout_reset_on(&txq->ift_timer, iflib_timer_default, iflib_timer,
2338 		    txq, txq->ift_timer.c_cpu);
2339 	return;
2340 
2341  hung:
2342 	device_printf(ctx->ifc_dev,
2343 	    "Watchdog timeout (TX: %d desc avail: %d pidx: %d) -- resetting\n",
2344 	    txq->ift_id, TXQ_AVAIL(txq), txq->ift_pidx);
2345 	STATE_LOCK(ctx);
2346 	if_setdrvflagbits(ctx->ifc_ifp, IFF_DRV_OACTIVE, IFF_DRV_RUNNING);
2347 	ctx->ifc_flags |= (IFC_DO_WATCHDOG|IFC_DO_RESET);
2348 	iflib_admin_intr_deferred(ctx);
2349 	STATE_UNLOCK(ctx);
2350 }
2351 
2352 static uint16_t
2353 iflib_get_mbuf_size_for(unsigned int size)
2354 {
2355 
2356 	if (size <= MCLBYTES)
2357 		return (MCLBYTES);
2358 	else
2359 		return (MJUMPAGESIZE);
2360 }
2361 
2362 static void
2363 iflib_calc_rx_mbuf_sz(if_ctx_t ctx)
2364 {
2365 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
2366 
2367 	/*
2368 	 * XXX don't set the max_frame_size to larger
2369 	 * than the hardware can handle
2370 	 */
2371 	ctx->ifc_rx_mbuf_sz =
2372 	    iflib_get_mbuf_size_for(sctx->isc_max_frame_size);
2373 }
2374 
2375 uint32_t
2376 iflib_get_rx_mbuf_sz(if_ctx_t ctx)
2377 {
2378 
2379 	return (ctx->ifc_rx_mbuf_sz);
2380 }
2381 
2382 static void
2383 iflib_init_locked(if_ctx_t ctx)
2384 {
2385 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
2386 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2387 	if_t ifp = ctx->ifc_ifp;
2388 	iflib_fl_t fl;
2389 	iflib_txq_t txq;
2390 	iflib_rxq_t rxq;
2391 	int i, j, tx_ip_csum_flags, tx_ip6_csum_flags;
2392 
2393 	if_setdrvflagbits(ifp, IFF_DRV_OACTIVE, IFF_DRV_RUNNING);
2394 	IFDI_INTR_DISABLE(ctx);
2395 
2396 	/*
2397 	 * See iflib_stop(). Useful in case iflib_init_locked() is
2398 	 * called without first calling iflib_stop().
2399 	 */
2400 	netmap_disable_all_rings(ifp);
2401 
2402 	tx_ip_csum_flags = scctx->isc_tx_csum_flags & (CSUM_IP | CSUM_TCP | CSUM_UDP | CSUM_SCTP);
2403 	tx_ip6_csum_flags = scctx->isc_tx_csum_flags & (CSUM_IP6_TCP | CSUM_IP6_UDP | CSUM_IP6_SCTP);
2404 	/* Set hardware offload abilities */
2405 	if_clearhwassist(ifp);
2406 	if (if_getcapenable(ifp) & IFCAP_TXCSUM)
2407 		if_sethwassistbits(ifp, tx_ip_csum_flags, 0);
2408 	if (if_getcapenable(ifp) & IFCAP_TXCSUM_IPV6)
2409 		if_sethwassistbits(ifp,  tx_ip6_csum_flags, 0);
2410 	if (if_getcapenable(ifp) & IFCAP_TSO4)
2411 		if_sethwassistbits(ifp, CSUM_IP_TSO, 0);
2412 	if (if_getcapenable(ifp) & IFCAP_TSO6)
2413 		if_sethwassistbits(ifp, CSUM_IP6_TSO, 0);
2414 
2415 	for (i = 0, txq = ctx->ifc_txqs; i < sctx->isc_ntxqsets; i++, txq++) {
2416 		CALLOUT_LOCK(txq);
2417 		callout_stop(&txq->ift_timer);
2418 #ifdef DEV_NETMAP
2419 		callout_stop(&txq->ift_netmap_timer);
2420 #endif /* DEV_NETMAP */
2421 		CALLOUT_UNLOCK(txq);
2422 		iflib_netmap_txq_init(ctx, txq);
2423 	}
2424 
2425 	/*
2426 	 * Calculate a suitable Rx mbuf size prior to calling IFDI_INIT, so
2427 	 * that drivers can use the value when setting up the hardware receive
2428 	 * buffers.
2429 	 */
2430 	iflib_calc_rx_mbuf_sz(ctx);
2431 
2432 #ifdef INVARIANTS
2433 	i = if_getdrvflags(ifp);
2434 #endif
2435 	IFDI_INIT(ctx);
2436 	MPASS(if_getdrvflags(ifp) == i);
2437 	for (i = 0, rxq = ctx->ifc_rxqs; i < sctx->isc_nrxqsets; i++, rxq++) {
2438 		if (iflib_netmap_rxq_init(ctx, rxq) > 0) {
2439 			/* This rxq is in netmap mode. Skip normal init. */
2440 			continue;
2441 		}
2442 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++) {
2443 			if (iflib_fl_setup(fl)) {
2444 				device_printf(ctx->ifc_dev,
2445 				    "setting up free list %d failed - "
2446 				    "check cluster settings\n", j);
2447 				goto done;
2448 			}
2449 		}
2450 	}
2451 done:
2452 	if_setdrvflagbits(ctx->ifc_ifp, IFF_DRV_RUNNING, IFF_DRV_OACTIVE);
2453 	IFDI_INTR_ENABLE(ctx);
2454 	txq = ctx->ifc_txqs;
2455 	for (i = 0; i < sctx->isc_ntxqsets; i++, txq++)
2456 		callout_reset_on(&txq->ift_timer, iflib_timer_default, iflib_timer, txq,
2457 			txq->ift_timer.c_cpu);
2458 
2459         /* Re-enable txsync/rxsync. */
2460 	netmap_enable_all_rings(ifp);
2461 }
2462 
2463 static int
2464 iflib_media_change(if_t ifp)
2465 {
2466 	if_ctx_t ctx = if_getsoftc(ifp);
2467 	int err;
2468 
2469 	CTX_LOCK(ctx);
2470 	if ((err = IFDI_MEDIA_CHANGE(ctx)) == 0)
2471 		iflib_init_locked(ctx);
2472 	CTX_UNLOCK(ctx);
2473 	return (err);
2474 }
2475 
2476 static void
2477 iflib_media_status(if_t ifp, struct ifmediareq *ifmr)
2478 {
2479 	if_ctx_t ctx = if_getsoftc(ifp);
2480 
2481 	CTX_LOCK(ctx);
2482 	IFDI_UPDATE_ADMIN_STATUS(ctx);
2483 	IFDI_MEDIA_STATUS(ctx, ifmr);
2484 	CTX_UNLOCK(ctx);
2485 }
2486 
2487 void
2488 iflib_stop(if_ctx_t ctx)
2489 {
2490 	iflib_txq_t txq = ctx->ifc_txqs;
2491 	iflib_rxq_t rxq = ctx->ifc_rxqs;
2492 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2493 	if_shared_ctx_t sctx = ctx->ifc_sctx;
2494 	iflib_dma_info_t di;
2495 	iflib_fl_t fl;
2496 	int i, j;
2497 
2498 	/* Tell the stack that the interface is no longer active */
2499 	if_setdrvflagbits(ctx->ifc_ifp, IFF_DRV_OACTIVE, IFF_DRV_RUNNING);
2500 
2501 	IFDI_INTR_DISABLE(ctx);
2502 	DELAY(1000);
2503 	IFDI_STOP(ctx);
2504 	DELAY(1000);
2505 
2506 	/*
2507 	 * Stop any pending txsync/rxsync and prevent new ones
2508 	 * form starting. Processes blocked in poll() will get
2509 	 * POLLERR.
2510 	 */
2511 	netmap_disable_all_rings(ctx->ifc_ifp);
2512 
2513 	iflib_debug_reset();
2514 	/* Wait for current tx queue users to exit to disarm watchdog timer. */
2515 	for (i = 0; i < scctx->isc_ntxqsets; i++, txq++) {
2516 		/* make sure all transmitters have completed before proceeding XXX */
2517 
2518 		CALLOUT_LOCK(txq);
2519 		callout_stop(&txq->ift_timer);
2520 #ifdef DEV_NETMAP
2521 		callout_stop(&txq->ift_netmap_timer);
2522 #endif /* DEV_NETMAP */
2523 		CALLOUT_UNLOCK(txq);
2524 
2525 		/* clean any enqueued buffers */
2526 		iflib_ifmp_purge(txq);
2527 		/* Free any existing tx buffers. */
2528 		for (j = 0; j < txq->ift_size; j++) {
2529 			iflib_txsd_free(ctx, txq, j);
2530 		}
2531 		txq->ift_processed = txq->ift_cleaned = txq->ift_cidx_processed = 0;
2532 		txq->ift_in_use = txq->ift_gen = txq->ift_cidx = txq->ift_pidx = txq->ift_no_desc_avail = 0;
2533 		txq->ift_closed = txq->ift_mbuf_defrag = txq->ift_mbuf_defrag_failed = 0;
2534 		txq->ift_no_tx_dma_setup = txq->ift_txd_encap_efbig = txq->ift_map_failed = 0;
2535 		txq->ift_pullups = 0;
2536 		ifmp_ring_reset_stats(txq->ift_br);
2537 		for (j = 0, di = txq->ift_ifdi; j < sctx->isc_ntxqs; j++, di++)
2538 			bzero((void *)di->idi_vaddr, di->idi_size);
2539 	}
2540 	for (i = 0; i < scctx->isc_nrxqsets; i++, rxq++) {
2541 		/* make sure all transmitters have completed before proceeding XXX */
2542 
2543 		rxq->ifr_cq_cidx = 0;
2544 		for (j = 0, di = rxq->ifr_ifdi; j < sctx->isc_nrxqs; j++, di++)
2545 			bzero((void *)di->idi_vaddr, di->idi_size);
2546 		/* also resets the free lists pidx/cidx */
2547 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++)
2548 			iflib_fl_bufs_free(fl);
2549 	}
2550 }
2551 
2552 static inline caddr_t
2553 calc_next_rxd(iflib_fl_t fl, int cidx)
2554 {
2555 	qidx_t size;
2556 	int nrxd;
2557 	caddr_t start, end, cur, next;
2558 
2559 	nrxd = fl->ifl_size;
2560 	size = fl->ifl_rxd_size;
2561 	start = fl->ifl_ifdi->idi_vaddr;
2562 
2563 	if (__predict_false(size == 0))
2564 		return (start);
2565 	cur = start + size*cidx;
2566 	end = start + size*nrxd;
2567 	next = CACHE_PTR_NEXT(cur);
2568 	return (next < end ? next : start);
2569 }
2570 
2571 static inline void
2572 prefetch_pkts(iflib_fl_t fl, int cidx)
2573 {
2574 	int nextptr;
2575 	int nrxd = fl->ifl_size;
2576 	caddr_t next_rxd;
2577 
2578 	nextptr = (cidx + CACHE_PTR_INCREMENT) & (nrxd-1);
2579 	prefetch(&fl->ifl_sds.ifsd_m[nextptr]);
2580 	prefetch(&fl->ifl_sds.ifsd_cl[nextptr]);
2581 	next_rxd = calc_next_rxd(fl, cidx);
2582 	prefetch(next_rxd);
2583 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 1) & (nrxd-1)]);
2584 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 2) & (nrxd-1)]);
2585 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 3) & (nrxd-1)]);
2586 	prefetch(fl->ifl_sds.ifsd_m[(cidx + 4) & (nrxd-1)]);
2587 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 1) & (nrxd-1)]);
2588 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 2) & (nrxd-1)]);
2589 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 3) & (nrxd-1)]);
2590 	prefetch(fl->ifl_sds.ifsd_cl[(cidx + 4) & (nrxd-1)]);
2591 }
2592 
2593 static struct mbuf *
2594 rxd_frag_to_sd(iflib_rxq_t rxq, if_rxd_frag_t irf, bool unload, if_rxsd_t sd,
2595     int *pf_rv, if_rxd_info_t ri)
2596 {
2597 	bus_dmamap_t map;
2598 	iflib_fl_t fl;
2599 	caddr_t payload;
2600 	struct mbuf *m;
2601 	int flid, cidx, len, next;
2602 
2603 	map = NULL;
2604 	flid = irf->irf_flid;
2605 	cidx = irf->irf_idx;
2606 	fl = &rxq->ifr_fl[flid];
2607 	sd->ifsd_fl = fl;
2608 	m = fl->ifl_sds.ifsd_m[cidx];
2609 	sd->ifsd_cl = &fl->ifl_sds.ifsd_cl[cidx];
2610 	fl->ifl_credits--;
2611 #if MEMORY_LOGGING
2612 	fl->ifl_m_dequeued++;
2613 #endif
2614 	if (rxq->ifr_ctx->ifc_flags & IFC_PREFETCH)
2615 		prefetch_pkts(fl, cidx);
2616 	next = (cidx + CACHE_PTR_INCREMENT) & (fl->ifl_size-1);
2617 	prefetch(&fl->ifl_sds.ifsd_map[next]);
2618 	map = fl->ifl_sds.ifsd_map[cidx];
2619 
2620 	bus_dmamap_sync(fl->ifl_buf_tag, map, BUS_DMASYNC_POSTREAD);
2621 
2622 	if (rxq->pfil != NULL && PFIL_HOOKED_IN(rxq->pfil) && pf_rv != NULL &&
2623 	    irf->irf_len != 0) {
2624 		payload  = *sd->ifsd_cl;
2625 		payload +=  ri->iri_pad;
2626 		len = ri->iri_len - ri->iri_pad;
2627 		*pf_rv = pfil_run_hooks(rxq->pfil, payload, ri->iri_ifp,
2628 		    len | PFIL_MEMPTR | PFIL_IN, NULL);
2629 		switch (*pf_rv) {
2630 		case PFIL_DROPPED:
2631 		case PFIL_CONSUMED:
2632 			/*
2633 			 * The filter ate it.  Everything is recycled.
2634 			 */
2635 			m = NULL;
2636 			unload = 0;
2637 			break;
2638 		case PFIL_REALLOCED:
2639 			/*
2640 			 * The filter copied it.  Everything is recycled.
2641 			 */
2642 			m = pfil_mem2mbuf(payload);
2643 			unload = 0;
2644 			break;
2645 		case PFIL_PASS:
2646 			/*
2647 			 * Filter said it was OK, so receive like
2648 			 * normal
2649 			 */
2650 			fl->ifl_sds.ifsd_m[cidx] = NULL;
2651 			break;
2652 		default:
2653 			MPASS(0);
2654 		}
2655 	} else {
2656 		fl->ifl_sds.ifsd_m[cidx] = NULL;
2657 		if (pf_rv != NULL)
2658 			*pf_rv = PFIL_PASS;
2659 	}
2660 
2661 	if (unload && irf->irf_len != 0)
2662 		bus_dmamap_unload(fl->ifl_buf_tag, map);
2663 	fl->ifl_cidx = (fl->ifl_cidx + 1) & (fl->ifl_size-1);
2664 	if (__predict_false(fl->ifl_cidx == 0))
2665 		fl->ifl_gen = 0;
2666 	bit_clear(fl->ifl_rx_bitmap, cidx);
2667 	return (m);
2668 }
2669 
2670 static struct mbuf *
2671 assemble_segments(iflib_rxq_t rxq, if_rxd_info_t ri, if_rxsd_t sd, int *pf_rv)
2672 {
2673 	struct mbuf *m, *mh, *mt;
2674 	caddr_t cl;
2675 	int  *pf_rv_ptr, flags, i, padlen;
2676 	bool consumed;
2677 
2678 	i = 0;
2679 	mh = NULL;
2680 	consumed = false;
2681 	*pf_rv = PFIL_PASS;
2682 	pf_rv_ptr = pf_rv;
2683 	do {
2684 		m = rxd_frag_to_sd(rxq, &ri->iri_frags[i], !consumed, sd,
2685 		    pf_rv_ptr, ri);
2686 
2687 		MPASS(*sd->ifsd_cl != NULL);
2688 
2689 		/*
2690 		 * Exclude zero-length frags & frags from
2691 		 * packets the filter has consumed or dropped
2692 		 */
2693 		if (ri->iri_frags[i].irf_len == 0 || consumed ||
2694 		    *pf_rv == PFIL_CONSUMED || *pf_rv == PFIL_DROPPED) {
2695 			if (mh == NULL) {
2696 				/* everything saved here */
2697 				consumed = true;
2698 				pf_rv_ptr = NULL;
2699 				continue;
2700 			}
2701 			/* XXX we can save the cluster here, but not the mbuf */
2702 			m_init(m, M_NOWAIT, MT_DATA, 0);
2703 			m_free(m);
2704 			continue;
2705 		}
2706 		if (mh == NULL) {
2707 			flags = M_PKTHDR|M_EXT;
2708 			mh = mt = m;
2709 			padlen = ri->iri_pad;
2710 		} else {
2711 			flags = M_EXT;
2712 			mt->m_next = m;
2713 			mt = m;
2714 			/* assuming padding is only on the first fragment */
2715 			padlen = 0;
2716 		}
2717 		cl = *sd->ifsd_cl;
2718 		*sd->ifsd_cl = NULL;
2719 
2720 		/* Can these two be made one ? */
2721 		m_init(m, M_NOWAIT, MT_DATA, flags);
2722 		m_cljset(m, cl, sd->ifsd_fl->ifl_cltype);
2723 		/*
2724 		 * These must follow m_init and m_cljset
2725 		 */
2726 		m->m_data += padlen;
2727 		ri->iri_len -= padlen;
2728 		m->m_len = ri->iri_frags[i].irf_len;
2729 	} while (++i < ri->iri_nfrags);
2730 
2731 	return (mh);
2732 }
2733 
2734 /*
2735  * Process one software descriptor
2736  */
2737 static struct mbuf *
2738 iflib_rxd_pkt_get(iflib_rxq_t rxq, if_rxd_info_t ri)
2739 {
2740 	struct if_rxsd sd;
2741 	struct mbuf *m;
2742 	int pf_rv;
2743 
2744 	/* should I merge this back in now that the two paths are basically duplicated? */
2745 	if (ri->iri_nfrags == 1 &&
2746 	    ri->iri_frags[0].irf_len != 0 &&
2747 	    ri->iri_frags[0].irf_len <= MIN(IFLIB_RX_COPY_THRESH, MHLEN)) {
2748 		m = rxd_frag_to_sd(rxq, &ri->iri_frags[0], false, &sd,
2749 		    &pf_rv, ri);
2750 		if (pf_rv != PFIL_PASS && pf_rv != PFIL_REALLOCED)
2751 			return (m);
2752 		if (pf_rv == PFIL_PASS) {
2753 			m_init(m, M_NOWAIT, MT_DATA, M_PKTHDR);
2754 #ifndef __NO_STRICT_ALIGNMENT
2755 			if (!IP_ALIGNED(m))
2756 				m->m_data += 2;
2757 #endif
2758 			memcpy(m->m_data, *sd.ifsd_cl, ri->iri_len);
2759 			m->m_len = ri->iri_frags[0].irf_len;
2760 		}
2761 	} else {
2762 		m = assemble_segments(rxq, ri, &sd, &pf_rv);
2763 		if (m == NULL)
2764 			return (NULL);
2765 		if (pf_rv != PFIL_PASS && pf_rv != PFIL_REALLOCED)
2766 			return (m);
2767 	}
2768 	m->m_pkthdr.len = ri->iri_len;
2769 	m->m_pkthdr.rcvif = ri->iri_ifp;
2770 	m->m_flags |= ri->iri_flags;
2771 	m->m_pkthdr.ether_vtag = ri->iri_vtag;
2772 	m->m_pkthdr.flowid = ri->iri_flowid;
2773 	M_HASHTYPE_SET(m, ri->iri_rsstype);
2774 	m->m_pkthdr.csum_flags = ri->iri_csum_flags;
2775 	m->m_pkthdr.csum_data = ri->iri_csum_data;
2776 	return (m);
2777 }
2778 
2779 #if defined(INET6) || defined(INET)
2780 static void
2781 iflib_get_ip_forwarding(struct lro_ctrl *lc, bool *v4, bool *v6)
2782 {
2783 	CURVNET_SET(lc->ifp->if_vnet);
2784 #if defined(INET6)
2785 	*v6 = V_ip6_forwarding;
2786 #endif
2787 #if defined(INET)
2788 	*v4 = V_ipforwarding;
2789 #endif
2790 	CURVNET_RESTORE();
2791 }
2792 
2793 /*
2794  * Returns true if it's possible this packet could be LROed.
2795  * if it returns false, it is guaranteed that tcp_lro_rx()
2796  * would not return zero.
2797  */
2798 static bool
2799 iflib_check_lro_possible(struct mbuf *m, bool v4_forwarding, bool v6_forwarding)
2800 {
2801 	struct ether_header *eh;
2802 
2803 	eh = mtod(m, struct ether_header *);
2804 	switch (eh->ether_type) {
2805 #if defined(INET6)
2806 		case htons(ETHERTYPE_IPV6):
2807 			return (!v6_forwarding);
2808 #endif
2809 #if defined (INET)
2810 		case htons(ETHERTYPE_IP):
2811 			return (!v4_forwarding);
2812 #endif
2813 	}
2814 
2815 	return false;
2816 }
2817 #else
2818 static void
2819 iflib_get_ip_forwarding(struct lro_ctrl *lc __unused, bool *v4 __unused, bool *v6 __unused)
2820 {
2821 }
2822 #endif
2823 
2824 static void
2825 _task_fn_rx_watchdog(void *context)
2826 {
2827 	iflib_rxq_t rxq = context;
2828 
2829 	GROUPTASK_ENQUEUE(&rxq->ifr_task);
2830 }
2831 
2832 static uint8_t
2833 iflib_rxeof(iflib_rxq_t rxq, qidx_t budget)
2834 {
2835 	if_t ifp;
2836 	if_ctx_t ctx = rxq->ifr_ctx;
2837 	if_shared_ctx_t sctx = ctx->ifc_sctx;
2838 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
2839 	int avail, i;
2840 	qidx_t *cidxp;
2841 	struct if_rxd_info ri;
2842 	int err, budget_left, rx_bytes, rx_pkts;
2843 	iflib_fl_t fl;
2844 	int lro_enabled;
2845 	bool v4_forwarding, v6_forwarding, lro_possible;
2846 	uint8_t retval = 0;
2847 
2848 	/*
2849 	 * XXX early demux data packets so that if_input processing only handles
2850 	 * acks in interrupt context
2851 	 */
2852 	struct mbuf *m, *mh, *mt, *mf;
2853 
2854 	NET_EPOCH_ASSERT();
2855 
2856 	lro_possible = v4_forwarding = v6_forwarding = false;
2857 	ifp = ctx->ifc_ifp;
2858 	mh = mt = NULL;
2859 	MPASS(budget > 0);
2860 	rx_pkts	= rx_bytes = 0;
2861 	if (sctx->isc_flags & IFLIB_HAS_RXCQ)
2862 		cidxp = &rxq->ifr_cq_cidx;
2863 	else
2864 		cidxp = &rxq->ifr_fl[0].ifl_cidx;
2865 	if ((avail = iflib_rxd_avail(ctx, rxq, *cidxp, budget)) == 0) {
2866 		for (i = 0, fl = &rxq->ifr_fl[0]; i < sctx->isc_nfl; i++, fl++)
2867 			retval |= iflib_fl_refill_all(ctx, fl);
2868 		DBG_COUNTER_INC(rx_unavail);
2869 		return (retval);
2870 	}
2871 
2872 	/* pfil needs the vnet to be set */
2873 	CURVNET_SET_QUIET(ifp->if_vnet);
2874 	for (budget_left = budget; budget_left > 0 && avail > 0;) {
2875 		if (__predict_false(!CTX_ACTIVE(ctx))) {
2876 			DBG_COUNTER_INC(rx_ctx_inactive);
2877 			break;
2878 		}
2879 		/*
2880 		 * Reset client set fields to their default values
2881 		 */
2882 		rxd_info_zero(&ri);
2883 		ri.iri_qsidx = rxq->ifr_id;
2884 		ri.iri_cidx = *cidxp;
2885 		ri.iri_ifp = ifp;
2886 		ri.iri_frags = rxq->ifr_frags;
2887 		err = ctx->isc_rxd_pkt_get(ctx->ifc_softc, &ri);
2888 
2889 		if (err)
2890 			goto err;
2891 		rx_pkts += 1;
2892 		rx_bytes += ri.iri_len;
2893 		if (sctx->isc_flags & IFLIB_HAS_RXCQ) {
2894 			*cidxp = ri.iri_cidx;
2895 			/* Update our consumer index */
2896 			/* XXX NB: shurd - check if this is still safe */
2897 			while (rxq->ifr_cq_cidx >= scctx->isc_nrxd[0])
2898 				rxq->ifr_cq_cidx -= scctx->isc_nrxd[0];
2899 			/* was this only a completion queue message? */
2900 			if (__predict_false(ri.iri_nfrags == 0))
2901 				continue;
2902 		}
2903 		MPASS(ri.iri_nfrags != 0);
2904 		MPASS(ri.iri_len != 0);
2905 
2906 		/* will advance the cidx on the corresponding free lists */
2907 		m = iflib_rxd_pkt_get(rxq, &ri);
2908 		avail--;
2909 		budget_left--;
2910 		if (avail == 0 && budget_left)
2911 			avail = iflib_rxd_avail(ctx, rxq, *cidxp, budget_left);
2912 
2913 		if (__predict_false(m == NULL))
2914 			continue;
2915 
2916 		/* imm_pkt: -- cxgb */
2917 		if (mh == NULL)
2918 			mh = mt = m;
2919 		else {
2920 			mt->m_nextpkt = m;
2921 			mt = m;
2922 		}
2923 	}
2924 	CURVNET_RESTORE();
2925 	/* make sure that we can refill faster than drain */
2926 	for (i = 0, fl = &rxq->ifr_fl[0]; i < sctx->isc_nfl; i++, fl++)
2927 		retval |= iflib_fl_refill_all(ctx, fl);
2928 
2929 	lro_enabled = (if_getcapenable(ifp) & IFCAP_LRO);
2930 	if (lro_enabled)
2931 		iflib_get_ip_forwarding(&rxq->ifr_lc, &v4_forwarding, &v6_forwarding);
2932 	mt = mf = NULL;
2933 	while (mh != NULL) {
2934 		m = mh;
2935 		mh = mh->m_nextpkt;
2936 		m->m_nextpkt = NULL;
2937 #ifndef __NO_STRICT_ALIGNMENT
2938 		if (!IP_ALIGNED(m) && (m = iflib_fixup_rx(m)) == NULL)
2939 			continue;
2940 #endif
2941 		rx_bytes += m->m_pkthdr.len;
2942 		rx_pkts++;
2943 #if defined(INET6) || defined(INET)
2944 		if (lro_enabled) {
2945 			if (!lro_possible) {
2946 				lro_possible = iflib_check_lro_possible(m, v4_forwarding, v6_forwarding);
2947 				if (lro_possible && mf != NULL) {
2948 					ifp->if_input(ifp, mf);
2949 					DBG_COUNTER_INC(rx_if_input);
2950 					mt = mf = NULL;
2951 				}
2952 			}
2953 			if ((m->m_pkthdr.csum_flags & (CSUM_L4_CALC|CSUM_L4_VALID)) ==
2954 			    (CSUM_L4_CALC|CSUM_L4_VALID)) {
2955 				if (lro_possible && tcp_lro_rx(&rxq->ifr_lc, m, 0) == 0)
2956 					continue;
2957 			}
2958 		}
2959 #endif
2960 		if (lro_possible) {
2961 			ifp->if_input(ifp, m);
2962 			DBG_COUNTER_INC(rx_if_input);
2963 			continue;
2964 		}
2965 
2966 		if (mf == NULL)
2967 			mf = m;
2968 		if (mt != NULL)
2969 			mt->m_nextpkt = m;
2970 		mt = m;
2971 	}
2972 	if (mf != NULL) {
2973 		ifp->if_input(ifp, mf);
2974 		DBG_COUNTER_INC(rx_if_input);
2975 	}
2976 
2977 	if_inc_counter(ifp, IFCOUNTER_IBYTES, rx_bytes);
2978 	if_inc_counter(ifp, IFCOUNTER_IPACKETS, rx_pkts);
2979 
2980 	/*
2981 	 * Flush any outstanding LRO work
2982 	 */
2983 #if defined(INET6) || defined(INET)
2984 	tcp_lro_flush_all(&rxq->ifr_lc);
2985 #endif
2986 	if (avail != 0 || iflib_rxd_avail(ctx, rxq, *cidxp, 1) != 0)
2987 		retval |= IFLIB_RXEOF_MORE;
2988 	return (retval);
2989 err:
2990 	STATE_LOCK(ctx);
2991 	ctx->ifc_flags |= IFC_DO_RESET;
2992 	iflib_admin_intr_deferred(ctx);
2993 	STATE_UNLOCK(ctx);
2994 	return (0);
2995 }
2996 
2997 #define TXD_NOTIFY_COUNT(txq) (((txq)->ift_size / (txq)->ift_update_freq)-1)
2998 static inline qidx_t
2999 txq_max_db_deferred(iflib_txq_t txq, qidx_t in_use)
3000 {
3001 	qidx_t notify_count = TXD_NOTIFY_COUNT(txq);
3002 	qidx_t minthresh = txq->ift_size / 8;
3003 	if (in_use > 4*minthresh)
3004 		return (notify_count);
3005 	if (in_use > 2*minthresh)
3006 		return (notify_count >> 1);
3007 	if (in_use > minthresh)
3008 		return (notify_count >> 3);
3009 	return (0);
3010 }
3011 
3012 static inline qidx_t
3013 txq_max_rs_deferred(iflib_txq_t txq)
3014 {
3015 	qidx_t notify_count = TXD_NOTIFY_COUNT(txq);
3016 	qidx_t minthresh = txq->ift_size / 8;
3017 	if (txq->ift_in_use > 4*minthresh)
3018 		return (notify_count);
3019 	if (txq->ift_in_use > 2*minthresh)
3020 		return (notify_count >> 1);
3021 	if (txq->ift_in_use > minthresh)
3022 		return (notify_count >> 2);
3023 	return (2);
3024 }
3025 
3026 #define M_CSUM_FLAGS(m) ((m)->m_pkthdr.csum_flags)
3027 #define M_HAS_VLANTAG(m) (m->m_flags & M_VLANTAG)
3028 
3029 #define TXQ_MAX_DB_DEFERRED(txq, in_use) txq_max_db_deferred((txq), (in_use))
3030 #define TXQ_MAX_RS_DEFERRED(txq) txq_max_rs_deferred(txq)
3031 #define TXQ_MAX_DB_CONSUMED(size) (size >> 4)
3032 
3033 /* forward compatibility for cxgb */
3034 #define FIRST_QSET(ctx) 0
3035 #define NTXQSETS(ctx) ((ctx)->ifc_softc_ctx.isc_ntxqsets)
3036 #define NRXQSETS(ctx) ((ctx)->ifc_softc_ctx.isc_nrxqsets)
3037 #define QIDX(ctx, m) ((((m)->m_pkthdr.flowid & ctx->ifc_softc_ctx.isc_rss_table_mask) % NTXQSETS(ctx)) + FIRST_QSET(ctx))
3038 #define DESC_RECLAIMABLE(q) ((int)((q)->ift_processed - (q)->ift_cleaned - (q)->ift_ctx->ifc_softc_ctx.isc_tx_nsegments))
3039 
3040 /* XXX we should be setting this to something other than zero */
3041 #define RECLAIM_THRESH(ctx) ((ctx)->ifc_sctx->isc_tx_reclaim_thresh)
3042 #define	MAX_TX_DESC(ctx) MAX((ctx)->ifc_softc_ctx.isc_tx_tso_segments_max, \
3043     (ctx)->ifc_softc_ctx.isc_tx_nsegments)
3044 
3045 static inline bool
3046 iflib_txd_db_check(iflib_txq_t txq, int ring)
3047 {
3048 	if_ctx_t ctx = txq->ift_ctx;
3049 	qidx_t dbval, max;
3050 
3051 	max = TXQ_MAX_DB_DEFERRED(txq, txq->ift_in_use);
3052 
3053 	/* force || threshold exceeded || at the edge of the ring */
3054 	if (ring || (txq->ift_db_pending >= max) || (TXQ_AVAIL(txq) <= MAX_TX_DESC(ctx) + 2)) {
3055 
3056 		/*
3057 		 * 'npending' is used if the card's doorbell is in terms of the number of descriptors
3058 		 * pending flush (BRCM). 'pidx' is used in cases where the card's doorbeel uses the
3059 		 * producer index explicitly (INTC).
3060 		 */
3061 		dbval = txq->ift_npending ? txq->ift_npending : txq->ift_pidx;
3062 		bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
3063 		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
3064 		ctx->isc_txd_flush(ctx->ifc_softc, txq->ift_id, dbval);
3065 
3066 		/*
3067 		 * Absent bugs there are zero packets pending so reset pending counts to zero.
3068 		 */
3069 		txq->ift_db_pending = txq->ift_npending = 0;
3070 		return (true);
3071 	}
3072 	return (false);
3073 }
3074 
3075 #ifdef PKT_DEBUG
3076 static void
3077 print_pkt(if_pkt_info_t pi)
3078 {
3079 	printf("pi len:  %d qsidx: %d nsegs: %d ndescs: %d flags: %x pidx: %d\n",
3080 	       pi->ipi_len, pi->ipi_qsidx, pi->ipi_nsegs, pi->ipi_ndescs, pi->ipi_flags, pi->ipi_pidx);
3081 	printf("pi new_pidx: %d csum_flags: %lx tso_segsz: %d mflags: %x vtag: %d\n",
3082 	       pi->ipi_new_pidx, pi->ipi_csum_flags, pi->ipi_tso_segsz, pi->ipi_mflags, pi->ipi_vtag);
3083 	printf("pi etype: %d ehdrlen: %d ip_hlen: %d ipproto: %d\n",
3084 	       pi->ipi_etype, pi->ipi_ehdrlen, pi->ipi_ip_hlen, pi->ipi_ipproto);
3085 }
3086 #endif
3087 
3088 #define IS_TSO4(pi) ((pi)->ipi_csum_flags & CSUM_IP_TSO)
3089 #define IS_TX_OFFLOAD4(pi) ((pi)->ipi_csum_flags & (CSUM_IP_TCP | CSUM_IP_TSO))
3090 #define IS_TSO6(pi) ((pi)->ipi_csum_flags & CSUM_IP6_TSO)
3091 #define IS_TX_OFFLOAD6(pi) ((pi)->ipi_csum_flags & (CSUM_IP6_TCP | CSUM_IP6_TSO))
3092 
3093 static int
3094 iflib_parse_header(iflib_txq_t txq, if_pkt_info_t pi, struct mbuf **mp)
3095 {
3096 	if_shared_ctx_t sctx = txq->ift_ctx->ifc_sctx;
3097 	struct ether_vlan_header *eh;
3098 	struct mbuf *m;
3099 
3100 	m = *mp;
3101 	if ((sctx->isc_flags & IFLIB_NEED_SCRATCH) &&
3102 	    M_WRITABLE(m) == 0) {
3103 		if ((m = m_dup(m, M_NOWAIT)) == NULL) {
3104 			return (ENOMEM);
3105 		} else {
3106 			m_freem(*mp);
3107 			DBG_COUNTER_INC(tx_frees);
3108 			*mp = m;
3109 		}
3110 	}
3111 
3112 	/*
3113 	 * Determine where frame payload starts.
3114 	 * Jump over vlan headers if already present,
3115 	 * helpful for QinQ too.
3116 	 */
3117 	if (__predict_false(m->m_len < sizeof(*eh))) {
3118 		txq->ift_pullups++;
3119 		if (__predict_false((m = m_pullup(m, sizeof(*eh))) == NULL))
3120 			return (ENOMEM);
3121 	}
3122 	eh = mtod(m, struct ether_vlan_header *);
3123 	if (eh->evl_encap_proto == htons(ETHERTYPE_VLAN)) {
3124 		pi->ipi_etype = ntohs(eh->evl_proto);
3125 		pi->ipi_ehdrlen = ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN;
3126 	} else {
3127 		pi->ipi_etype = ntohs(eh->evl_encap_proto);
3128 		pi->ipi_ehdrlen = ETHER_HDR_LEN;
3129 	}
3130 
3131 	switch (pi->ipi_etype) {
3132 #ifdef INET
3133 	case ETHERTYPE_IP:
3134 	{
3135 		struct mbuf *n;
3136 		struct ip *ip = NULL;
3137 		struct tcphdr *th = NULL;
3138 		int minthlen;
3139 
3140 		minthlen = min(m->m_pkthdr.len, pi->ipi_ehdrlen + sizeof(*ip) + sizeof(*th));
3141 		if (__predict_false(m->m_len < minthlen)) {
3142 			/*
3143 			 * if this code bloat is causing too much of a hit
3144 			 * move it to a separate function and mark it noinline
3145 			 */
3146 			if (m->m_len == pi->ipi_ehdrlen) {
3147 				n = m->m_next;
3148 				MPASS(n);
3149 				if (n->m_len >= sizeof(*ip))  {
3150 					ip = (struct ip *)n->m_data;
3151 					if (n->m_len >= (ip->ip_hl << 2) + sizeof(*th))
3152 						th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
3153 				} else {
3154 					txq->ift_pullups++;
3155 					if (__predict_false((m = m_pullup(m, minthlen)) == NULL))
3156 						return (ENOMEM);
3157 					ip = (struct ip *)(m->m_data + pi->ipi_ehdrlen);
3158 				}
3159 			} else {
3160 				txq->ift_pullups++;
3161 				if (__predict_false((m = m_pullup(m, minthlen)) == NULL))
3162 					return (ENOMEM);
3163 				ip = (struct ip *)(m->m_data + pi->ipi_ehdrlen);
3164 				if (m->m_len >= (ip->ip_hl << 2) + sizeof(*th))
3165 					th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
3166 			}
3167 		} else {
3168 			ip = (struct ip *)(m->m_data + pi->ipi_ehdrlen);
3169 			if (m->m_len >= (ip->ip_hl << 2) + sizeof(*th))
3170 				th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
3171 		}
3172 		pi->ipi_ip_hlen = ip->ip_hl << 2;
3173 		pi->ipi_ipproto = ip->ip_p;
3174 		pi->ipi_flags |= IPI_TX_IPV4;
3175 
3176 		/* TCP checksum offload may require TCP header length */
3177 		if (IS_TX_OFFLOAD4(pi)) {
3178 			if (__predict_true(pi->ipi_ipproto == IPPROTO_TCP)) {
3179 				if (__predict_false(th == NULL)) {
3180 					txq->ift_pullups++;
3181 					if (__predict_false((m = m_pullup(m, (ip->ip_hl << 2) + sizeof(*th))) == NULL))
3182 						return (ENOMEM);
3183 					th = (struct tcphdr *)((caddr_t)ip + pi->ipi_ip_hlen);
3184 				}
3185 				pi->ipi_tcp_hflags = th->th_flags;
3186 				pi->ipi_tcp_hlen = th->th_off << 2;
3187 				pi->ipi_tcp_seq = th->th_seq;
3188 			}
3189 			if (IS_TSO4(pi)) {
3190 				if (__predict_false(ip->ip_p != IPPROTO_TCP))
3191 					return (ENXIO);
3192 				/*
3193 				 * TSO always requires hardware checksum offload.
3194 				 */
3195 				pi->ipi_csum_flags |= (CSUM_IP_TCP | CSUM_IP);
3196 				th->th_sum = in_pseudo(ip->ip_src.s_addr,
3197 						       ip->ip_dst.s_addr, htons(IPPROTO_TCP));
3198 				pi->ipi_tso_segsz = m->m_pkthdr.tso_segsz;
3199 				if (sctx->isc_flags & IFLIB_TSO_INIT_IP) {
3200 					ip->ip_sum = 0;
3201 					ip->ip_len = htons(pi->ipi_ip_hlen + pi->ipi_tcp_hlen + pi->ipi_tso_segsz);
3202 				}
3203 			}
3204 		}
3205 		if ((sctx->isc_flags & IFLIB_NEED_ZERO_CSUM) && (pi->ipi_csum_flags & CSUM_IP))
3206                        ip->ip_sum = 0;
3207 
3208 		break;
3209 	}
3210 #endif
3211 #ifdef INET6
3212 	case ETHERTYPE_IPV6:
3213 	{
3214 		struct ip6_hdr *ip6 = (struct ip6_hdr *)(m->m_data + pi->ipi_ehdrlen);
3215 		struct tcphdr *th;
3216 		pi->ipi_ip_hlen = sizeof(struct ip6_hdr);
3217 
3218 		if (__predict_false(m->m_len < pi->ipi_ehdrlen + sizeof(struct ip6_hdr))) {
3219 			txq->ift_pullups++;
3220 			if (__predict_false((m = m_pullup(m, pi->ipi_ehdrlen + sizeof(struct ip6_hdr))) == NULL))
3221 				return (ENOMEM);
3222 		}
3223 		th = (struct tcphdr *)((caddr_t)ip6 + pi->ipi_ip_hlen);
3224 
3225 		/* XXX-BZ this will go badly in case of ext hdrs. */
3226 		pi->ipi_ipproto = ip6->ip6_nxt;
3227 		pi->ipi_flags |= IPI_TX_IPV6;
3228 
3229 		/* TCP checksum offload may require TCP header length */
3230 		if (IS_TX_OFFLOAD6(pi)) {
3231 			if (pi->ipi_ipproto == IPPROTO_TCP) {
3232 				if (__predict_false(m->m_len < pi->ipi_ehdrlen + sizeof(struct ip6_hdr) + sizeof(struct tcphdr))) {
3233 					txq->ift_pullups++;
3234 					if (__predict_false((m = m_pullup(m, pi->ipi_ehdrlen + sizeof(struct ip6_hdr) + sizeof(struct tcphdr))) == NULL))
3235 						return (ENOMEM);
3236 				}
3237 				pi->ipi_tcp_hflags = th->th_flags;
3238 				pi->ipi_tcp_hlen = th->th_off << 2;
3239 				pi->ipi_tcp_seq = th->th_seq;
3240 			}
3241 			if (IS_TSO6(pi)) {
3242 				if (__predict_false(ip6->ip6_nxt != IPPROTO_TCP))
3243 					return (ENXIO);
3244 				/*
3245 				 * TSO always requires hardware checksum offload.
3246 				 */
3247 				pi->ipi_csum_flags |= CSUM_IP6_TCP;
3248 				th->th_sum = in6_cksum_pseudo(ip6, 0, IPPROTO_TCP, 0);
3249 				pi->ipi_tso_segsz = m->m_pkthdr.tso_segsz;
3250 			}
3251 		}
3252 		break;
3253 	}
3254 #endif
3255 	default:
3256 		pi->ipi_csum_flags &= ~CSUM_OFFLOAD;
3257 		pi->ipi_ip_hlen = 0;
3258 		break;
3259 	}
3260 	*mp = m;
3261 
3262 	return (0);
3263 }
3264 
3265 /*
3266  * If dodgy hardware rejects the scatter gather chain we've handed it
3267  * we'll need to remove the mbuf chain from ifsg_m[] before we can add the
3268  * m_defrag'd mbufs
3269  */
3270 static __noinline struct mbuf *
3271 iflib_remove_mbuf(iflib_txq_t txq)
3272 {
3273 	int ntxd, pidx;
3274 	struct mbuf *m, **ifsd_m;
3275 
3276 	ifsd_m = txq->ift_sds.ifsd_m;
3277 	ntxd = txq->ift_size;
3278 	pidx = txq->ift_pidx & (ntxd - 1);
3279 	ifsd_m = txq->ift_sds.ifsd_m;
3280 	m = ifsd_m[pidx];
3281 	ifsd_m[pidx] = NULL;
3282 	bus_dmamap_unload(txq->ift_buf_tag, txq->ift_sds.ifsd_map[pidx]);
3283 	if (txq->ift_sds.ifsd_tso_map != NULL)
3284 		bus_dmamap_unload(txq->ift_tso_buf_tag,
3285 		    txq->ift_sds.ifsd_tso_map[pidx]);
3286 #if MEMORY_LOGGING
3287 	txq->ift_dequeued++;
3288 #endif
3289 	return (m);
3290 }
3291 
3292 static inline caddr_t
3293 calc_next_txd(iflib_txq_t txq, int cidx, uint8_t qid)
3294 {
3295 	qidx_t size;
3296 	int ntxd;
3297 	caddr_t start, end, cur, next;
3298 
3299 	ntxd = txq->ift_size;
3300 	size = txq->ift_txd_size[qid];
3301 	start = txq->ift_ifdi[qid].idi_vaddr;
3302 
3303 	if (__predict_false(size == 0))
3304 		return (start);
3305 	cur = start + size*cidx;
3306 	end = start + size*ntxd;
3307 	next = CACHE_PTR_NEXT(cur);
3308 	return (next < end ? next : start);
3309 }
3310 
3311 /*
3312  * Pad an mbuf to ensure a minimum ethernet frame size.
3313  * min_frame_size is the frame size (less CRC) to pad the mbuf to
3314  */
3315 static __noinline int
3316 iflib_ether_pad(device_t dev, struct mbuf **m_head, uint16_t min_frame_size)
3317 {
3318 	/*
3319 	 * 18 is enough bytes to pad an ARP packet to 46 bytes, and
3320 	 * and ARP message is the smallest common payload I can think of
3321 	 */
3322 	static char pad[18];	/* just zeros */
3323 	int n;
3324 	struct mbuf *new_head;
3325 
3326 	if (!M_WRITABLE(*m_head)) {
3327 		new_head = m_dup(*m_head, M_NOWAIT);
3328 		if (new_head == NULL) {
3329 			m_freem(*m_head);
3330 			device_printf(dev, "cannot pad short frame, m_dup() failed");
3331 			DBG_COUNTER_INC(encap_pad_mbuf_fail);
3332 			DBG_COUNTER_INC(tx_frees);
3333 			return ENOMEM;
3334 		}
3335 		m_freem(*m_head);
3336 		*m_head = new_head;
3337 	}
3338 
3339 	for (n = min_frame_size - (*m_head)->m_pkthdr.len;
3340 	     n > 0; n -= sizeof(pad))
3341 		if (!m_append(*m_head, min(n, sizeof(pad)), pad))
3342 			break;
3343 
3344 	if (n > 0) {
3345 		m_freem(*m_head);
3346 		device_printf(dev, "cannot pad short frame\n");
3347 		DBG_COUNTER_INC(encap_pad_mbuf_fail);
3348 		DBG_COUNTER_INC(tx_frees);
3349 		return (ENOBUFS);
3350 	}
3351 
3352 	return 0;
3353 }
3354 
3355 static int
3356 iflib_encap(iflib_txq_t txq, struct mbuf **m_headp)
3357 {
3358 	if_ctx_t		ctx;
3359 	if_shared_ctx_t		sctx;
3360 	if_softc_ctx_t		scctx;
3361 	bus_dma_tag_t		buf_tag;
3362 	bus_dma_segment_t	*segs;
3363 	struct mbuf		*m_head, **ifsd_m;
3364 	void			*next_txd;
3365 	bus_dmamap_t		map;
3366 	struct if_pkt_info	pi;
3367 	int remap = 0;
3368 	int err, nsegs, ndesc, max_segs, pidx, cidx, next, ntxd;
3369 
3370 	ctx = txq->ift_ctx;
3371 	sctx = ctx->ifc_sctx;
3372 	scctx = &ctx->ifc_softc_ctx;
3373 	segs = txq->ift_segs;
3374 	ntxd = txq->ift_size;
3375 	m_head = *m_headp;
3376 	map = NULL;
3377 
3378 	/*
3379 	 * If we're doing TSO the next descriptor to clean may be quite far ahead
3380 	 */
3381 	cidx = txq->ift_cidx;
3382 	pidx = txq->ift_pidx;
3383 	if (ctx->ifc_flags & IFC_PREFETCH) {
3384 		next = (cidx + CACHE_PTR_INCREMENT) & (ntxd-1);
3385 		if (!(ctx->ifc_flags & IFLIB_HAS_TXCQ)) {
3386 			next_txd = calc_next_txd(txq, cidx, 0);
3387 			prefetch(next_txd);
3388 		}
3389 
3390 		/* prefetch the next cache line of mbuf pointers and flags */
3391 		prefetch(&txq->ift_sds.ifsd_m[next]);
3392 		prefetch(&txq->ift_sds.ifsd_map[next]);
3393 		next = (cidx + CACHE_LINE_SIZE) & (ntxd-1);
3394 	}
3395 	map = txq->ift_sds.ifsd_map[pidx];
3396 	ifsd_m = txq->ift_sds.ifsd_m;
3397 
3398 	if (m_head->m_pkthdr.csum_flags & CSUM_TSO) {
3399 		buf_tag = txq->ift_tso_buf_tag;
3400 		max_segs = scctx->isc_tx_tso_segments_max;
3401 		map = txq->ift_sds.ifsd_tso_map[pidx];
3402 		MPASS(buf_tag != NULL);
3403 		MPASS(max_segs > 0);
3404 	} else {
3405 		buf_tag = txq->ift_buf_tag;
3406 		max_segs = scctx->isc_tx_nsegments;
3407 		map = txq->ift_sds.ifsd_map[pidx];
3408 	}
3409 	if ((sctx->isc_flags & IFLIB_NEED_ETHER_PAD) &&
3410 	    __predict_false(m_head->m_pkthdr.len < scctx->isc_min_frame_size)) {
3411 		err = iflib_ether_pad(ctx->ifc_dev, m_headp, scctx->isc_min_frame_size);
3412 		if (err) {
3413 			DBG_COUNTER_INC(encap_txd_encap_fail);
3414 			return err;
3415 		}
3416 	}
3417 	m_head = *m_headp;
3418 
3419 	pkt_info_zero(&pi);
3420 	pi.ipi_mflags = (m_head->m_flags & (M_VLANTAG|M_BCAST|M_MCAST));
3421 	pi.ipi_pidx = pidx;
3422 	pi.ipi_qsidx = txq->ift_id;
3423 	pi.ipi_len = m_head->m_pkthdr.len;
3424 	pi.ipi_csum_flags = m_head->m_pkthdr.csum_flags;
3425 	pi.ipi_vtag = M_HAS_VLANTAG(m_head) ? m_head->m_pkthdr.ether_vtag : 0;
3426 
3427 	/* deliberate bitwise OR to make one condition */
3428 	if (__predict_true((pi.ipi_csum_flags | pi.ipi_vtag))) {
3429 		if (__predict_false((err = iflib_parse_header(txq, &pi, m_headp)) != 0)) {
3430 			DBG_COUNTER_INC(encap_txd_encap_fail);
3431 			return (err);
3432 		}
3433 		m_head = *m_headp;
3434 	}
3435 
3436 retry:
3437 	err = bus_dmamap_load_mbuf_sg(buf_tag, map, m_head, segs, &nsegs,
3438 	    BUS_DMA_NOWAIT);
3439 defrag:
3440 	if (__predict_false(err)) {
3441 		switch (err) {
3442 		case EFBIG:
3443 			/* try collapse once and defrag once */
3444 			if (remap == 0) {
3445 				m_head = m_collapse(*m_headp, M_NOWAIT, max_segs);
3446 				/* try defrag if collapsing fails */
3447 				if (m_head == NULL)
3448 					remap++;
3449 			}
3450 			if (remap == 1) {
3451 				txq->ift_mbuf_defrag++;
3452 				m_head = m_defrag(*m_headp, M_NOWAIT);
3453 			}
3454 			/*
3455 			 * remap should never be >1 unless bus_dmamap_load_mbuf_sg
3456 			 * failed to map an mbuf that was run through m_defrag
3457 			 */
3458 			MPASS(remap <= 1);
3459 			if (__predict_false(m_head == NULL || remap > 1))
3460 				goto defrag_failed;
3461 			remap++;
3462 			*m_headp = m_head;
3463 			goto retry;
3464 			break;
3465 		case ENOMEM:
3466 			txq->ift_no_tx_dma_setup++;
3467 			break;
3468 		default:
3469 			txq->ift_no_tx_dma_setup++;
3470 			m_freem(*m_headp);
3471 			DBG_COUNTER_INC(tx_frees);
3472 			*m_headp = NULL;
3473 			break;
3474 		}
3475 		txq->ift_map_failed++;
3476 		DBG_COUNTER_INC(encap_load_mbuf_fail);
3477 		DBG_COUNTER_INC(encap_txd_encap_fail);
3478 		return (err);
3479 	}
3480 	ifsd_m[pidx] = m_head;
3481 	/*
3482 	 * XXX assumes a 1 to 1 relationship between segments and
3483 	 *        descriptors - this does not hold true on all drivers, e.g.
3484 	 *        cxgb
3485 	 */
3486 	if (__predict_false(nsegs + 2 > TXQ_AVAIL(txq))) {
3487 		txq->ift_no_desc_avail++;
3488 		bus_dmamap_unload(buf_tag, map);
3489 		DBG_COUNTER_INC(encap_txq_avail_fail);
3490 		DBG_COUNTER_INC(encap_txd_encap_fail);
3491 		if ((txq->ift_task.gt_task.ta_flags & TASK_ENQUEUED) == 0)
3492 			GROUPTASK_ENQUEUE(&txq->ift_task);
3493 		return (ENOBUFS);
3494 	}
3495 	/*
3496 	 * On Intel cards we can greatly reduce the number of TX interrupts
3497 	 * we see by only setting report status on every Nth descriptor.
3498 	 * However, this also means that the driver will need to keep track
3499 	 * of the descriptors that RS was set on to check them for the DD bit.
3500 	 */
3501 	txq->ift_rs_pending += nsegs + 1;
3502 	if (txq->ift_rs_pending > TXQ_MAX_RS_DEFERRED(txq) ||
3503 	     iflib_no_tx_batch || (TXQ_AVAIL(txq) - nsegs) <= MAX_TX_DESC(ctx) + 2) {
3504 		pi.ipi_flags |= IPI_TX_INTR;
3505 		txq->ift_rs_pending = 0;
3506 	}
3507 
3508 	pi.ipi_segs = segs;
3509 	pi.ipi_nsegs = nsegs;
3510 
3511 	MPASS(pidx >= 0 && pidx < txq->ift_size);
3512 #ifdef PKT_DEBUG
3513 	print_pkt(&pi);
3514 #endif
3515 	if ((err = ctx->isc_txd_encap(ctx->ifc_softc, &pi)) == 0) {
3516 		bus_dmamap_sync(buf_tag, map, BUS_DMASYNC_PREWRITE);
3517 		DBG_COUNTER_INC(tx_encap);
3518 		MPASS(pi.ipi_new_pidx < txq->ift_size);
3519 
3520 		ndesc = pi.ipi_new_pidx - pi.ipi_pidx;
3521 		if (pi.ipi_new_pidx < pi.ipi_pidx) {
3522 			ndesc += txq->ift_size;
3523 			txq->ift_gen = 1;
3524 		}
3525 		/*
3526 		 * drivers can need as many as
3527 		 * two sentinels
3528 		 */
3529 		MPASS(ndesc <= pi.ipi_nsegs + 2);
3530 		MPASS(pi.ipi_new_pidx != pidx);
3531 		MPASS(ndesc > 0);
3532 		txq->ift_in_use += ndesc;
3533 		txq->ift_db_pending += ndesc;
3534 
3535 		/*
3536 		 * We update the last software descriptor again here because there may
3537 		 * be a sentinel and/or there may be more mbufs than segments
3538 		 */
3539 		txq->ift_pidx = pi.ipi_new_pidx;
3540 		txq->ift_npending += pi.ipi_ndescs;
3541 	} else {
3542 		*m_headp = m_head = iflib_remove_mbuf(txq);
3543 		if (err == EFBIG) {
3544 			txq->ift_txd_encap_efbig++;
3545 			if (remap < 2) {
3546 				remap = 1;
3547 				goto defrag;
3548 			}
3549 		}
3550 		goto defrag_failed;
3551 	}
3552 	/*
3553 	 * err can't possibly be non-zero here, so we don't neet to test it
3554 	 * to see if we need to DBG_COUNTER_INC(encap_txd_encap_fail).
3555 	 */
3556 	return (err);
3557 
3558 defrag_failed:
3559 	txq->ift_mbuf_defrag_failed++;
3560 	txq->ift_map_failed++;
3561 	m_freem(*m_headp);
3562 	DBG_COUNTER_INC(tx_frees);
3563 	*m_headp = NULL;
3564 	DBG_COUNTER_INC(encap_txd_encap_fail);
3565 	return (ENOMEM);
3566 }
3567 
3568 static void
3569 iflib_tx_desc_free(iflib_txq_t txq, int n)
3570 {
3571 	uint32_t qsize, cidx, mask, gen;
3572 	struct mbuf *m, **ifsd_m;
3573 	bool do_prefetch;
3574 
3575 	cidx = txq->ift_cidx;
3576 	gen = txq->ift_gen;
3577 	qsize = txq->ift_size;
3578 	mask = qsize-1;
3579 	ifsd_m = txq->ift_sds.ifsd_m;
3580 	do_prefetch = (txq->ift_ctx->ifc_flags & IFC_PREFETCH);
3581 
3582 	while (n-- > 0) {
3583 		if (do_prefetch) {
3584 			prefetch(ifsd_m[(cidx + 3) & mask]);
3585 			prefetch(ifsd_m[(cidx + 4) & mask]);
3586 		}
3587 		if ((m = ifsd_m[cidx]) != NULL) {
3588 			prefetch(&ifsd_m[(cidx + CACHE_PTR_INCREMENT) & mask]);
3589 			if (m->m_pkthdr.csum_flags & CSUM_TSO) {
3590 				bus_dmamap_sync(txq->ift_tso_buf_tag,
3591 				    txq->ift_sds.ifsd_tso_map[cidx],
3592 				    BUS_DMASYNC_POSTWRITE);
3593 				bus_dmamap_unload(txq->ift_tso_buf_tag,
3594 				    txq->ift_sds.ifsd_tso_map[cidx]);
3595 			} else {
3596 				bus_dmamap_sync(txq->ift_buf_tag,
3597 				    txq->ift_sds.ifsd_map[cidx],
3598 				    BUS_DMASYNC_POSTWRITE);
3599 				bus_dmamap_unload(txq->ift_buf_tag,
3600 				    txq->ift_sds.ifsd_map[cidx]);
3601 			}
3602 			/* XXX we don't support any drivers that batch packets yet */
3603 			MPASS(m->m_nextpkt == NULL);
3604 			m_freem(m);
3605 			ifsd_m[cidx] = NULL;
3606 #if MEMORY_LOGGING
3607 			txq->ift_dequeued++;
3608 #endif
3609 			DBG_COUNTER_INC(tx_frees);
3610 		}
3611 		if (__predict_false(++cidx == qsize)) {
3612 			cidx = 0;
3613 			gen = 0;
3614 		}
3615 	}
3616 	txq->ift_cidx = cidx;
3617 	txq->ift_gen = gen;
3618 }
3619 
3620 static __inline int
3621 iflib_completed_tx_reclaim(iflib_txq_t txq, int thresh)
3622 {
3623 	int reclaim;
3624 	if_ctx_t ctx = txq->ift_ctx;
3625 
3626 	KASSERT(thresh >= 0, ("invalid threshold to reclaim"));
3627 	MPASS(thresh /*+ MAX_TX_DESC(txq->ift_ctx) */ < txq->ift_size);
3628 
3629 	/*
3630 	 * Need a rate-limiting check so that this isn't called every time
3631 	 */
3632 	iflib_tx_credits_update(ctx, txq);
3633 	reclaim = DESC_RECLAIMABLE(txq);
3634 
3635 	if (reclaim <= thresh /* + MAX_TX_DESC(txq->ift_ctx) */) {
3636 #ifdef INVARIANTS
3637 		if (iflib_verbose_debug) {
3638 			printf("%s processed=%ju cleaned=%ju tx_nsegments=%d reclaim=%d thresh=%d\n", __FUNCTION__,
3639 			       txq->ift_processed, txq->ift_cleaned, txq->ift_ctx->ifc_softc_ctx.isc_tx_nsegments,
3640 			       reclaim, thresh);
3641 		}
3642 #endif
3643 		return (0);
3644 	}
3645 	iflib_tx_desc_free(txq, reclaim);
3646 	txq->ift_cleaned += reclaim;
3647 	txq->ift_in_use -= reclaim;
3648 
3649 	return (reclaim);
3650 }
3651 
3652 static struct mbuf **
3653 _ring_peek_one(struct ifmp_ring *r, int cidx, int offset, int remaining)
3654 {
3655 	int next, size;
3656 	struct mbuf **items;
3657 
3658 	size = r->size;
3659 	next = (cidx + CACHE_PTR_INCREMENT) & (size-1);
3660 	items = __DEVOLATILE(struct mbuf **, &r->items[0]);
3661 
3662 	prefetch(items[(cidx + offset) & (size-1)]);
3663 	if (remaining > 1) {
3664 		prefetch2cachelines(&items[next]);
3665 		prefetch2cachelines(items[(cidx + offset + 1) & (size-1)]);
3666 		prefetch2cachelines(items[(cidx + offset + 2) & (size-1)]);
3667 		prefetch2cachelines(items[(cidx + offset + 3) & (size-1)]);
3668 	}
3669 	return (__DEVOLATILE(struct mbuf **, &r->items[(cidx + offset) & (size-1)]));
3670 }
3671 
3672 static void
3673 iflib_txq_check_drain(iflib_txq_t txq, int budget)
3674 {
3675 
3676 	ifmp_ring_check_drainage(txq->ift_br, budget);
3677 }
3678 
3679 static uint32_t
3680 iflib_txq_can_drain(struct ifmp_ring *r)
3681 {
3682 	iflib_txq_t txq = r->cookie;
3683 	if_ctx_t ctx = txq->ift_ctx;
3684 
3685 	if (TXQ_AVAIL(txq) > MAX_TX_DESC(ctx) + 2)
3686 		return (1);
3687 	bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
3688 	    BUS_DMASYNC_POSTREAD);
3689 	return (ctx->isc_txd_credits_update(ctx->ifc_softc, txq->ift_id,
3690 	    false));
3691 }
3692 
3693 static uint32_t
3694 iflib_txq_drain(struct ifmp_ring *r, uint32_t cidx, uint32_t pidx)
3695 {
3696 	iflib_txq_t txq = r->cookie;
3697 	if_ctx_t ctx = txq->ift_ctx;
3698 	if_t ifp = ctx->ifc_ifp;
3699 	struct mbuf *m, **mp;
3700 	int avail, bytes_sent, skipped, count, err, i;
3701 	int mcast_sent, pkt_sent, reclaimed;
3702 	bool do_prefetch, rang, ring;
3703 
3704 	if (__predict_false(!(if_getdrvflags(ifp) & IFF_DRV_RUNNING) ||
3705 			    !LINK_ACTIVE(ctx))) {
3706 		DBG_COUNTER_INC(txq_drain_notready);
3707 		return (0);
3708 	}
3709 	reclaimed = iflib_completed_tx_reclaim(txq, RECLAIM_THRESH(ctx));
3710 	rang = iflib_txd_db_check(txq, reclaimed && txq->ift_db_pending);
3711 	avail = IDXDIFF(pidx, cidx, r->size);
3712 
3713 	if (__predict_false(ctx->ifc_flags & IFC_QFLUSH)) {
3714 		/*
3715 		 * The driver is unloading so we need to free all pending packets.
3716 		 */
3717 		DBG_COUNTER_INC(txq_drain_flushing);
3718 		for (i = 0; i < avail; i++) {
3719 			if (__predict_true(r->items[(cidx + i) & (r->size-1)] != (void *)txq))
3720 				m_freem(r->items[(cidx + i) & (r->size-1)]);
3721 			r->items[(cidx + i) & (r->size-1)] = NULL;
3722 		}
3723 		return (avail);
3724 	}
3725 
3726 	if (__predict_false(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_OACTIVE)) {
3727 		txq->ift_qstatus = IFLIB_QUEUE_IDLE;
3728 		CALLOUT_LOCK(txq);
3729 		callout_stop(&txq->ift_timer);
3730 		CALLOUT_UNLOCK(txq);
3731 		DBG_COUNTER_INC(txq_drain_oactive);
3732 		return (0);
3733 	}
3734 
3735 	/*
3736 	 * If we've reclaimed any packets this queue cannot be hung.
3737 	 */
3738 	if (reclaimed)
3739 		txq->ift_qstatus = IFLIB_QUEUE_IDLE;
3740 	skipped = mcast_sent = bytes_sent = pkt_sent = 0;
3741 	count = MIN(avail, TX_BATCH_SIZE);
3742 #ifdef INVARIANTS
3743 	if (iflib_verbose_debug)
3744 		printf("%s avail=%d ifc_flags=%x txq_avail=%d ", __FUNCTION__,
3745 		       avail, ctx->ifc_flags, TXQ_AVAIL(txq));
3746 #endif
3747 	do_prefetch = (ctx->ifc_flags & IFC_PREFETCH);
3748 	err = 0;
3749 	for (i = 0; i < count && TXQ_AVAIL(txq) >= MAX_TX_DESC(ctx) + 2; i++) {
3750 		int rem = do_prefetch ? count - i : 0;
3751 
3752 		mp = _ring_peek_one(r, cidx, i, rem);
3753 		MPASS(mp != NULL && *mp != NULL);
3754 
3755 		/*
3756 		 * Completion interrupts will use the address of the txq
3757 		 * as a sentinel to enqueue _something_ in order to acquire
3758 		 * the lock on the mp_ring (there's no direct lock call).
3759 		 * We obviously whave to check for these sentinel cases
3760 		 * and skip them.
3761 		 */
3762 		if (__predict_false(*mp == (struct mbuf *)txq)) {
3763 			skipped++;
3764 			continue;
3765 		}
3766 		err = iflib_encap(txq, mp);
3767 		if (__predict_false(err)) {
3768 			/* no room - bail out */
3769 			if (err == ENOBUFS)
3770 				break;
3771 			skipped++;
3772 			/* we can't send this packet - skip it */
3773 			continue;
3774 		}
3775 		pkt_sent++;
3776 		m = *mp;
3777 		DBG_COUNTER_INC(tx_sent);
3778 		bytes_sent += m->m_pkthdr.len;
3779 		mcast_sent += !!(m->m_flags & M_MCAST);
3780 
3781 		if (__predict_false(!(ifp->if_drv_flags & IFF_DRV_RUNNING)))
3782 			break;
3783 		ETHER_BPF_MTAP(ifp, m);
3784 		rang = iflib_txd_db_check(txq, false);
3785 	}
3786 
3787 	/* deliberate use of bitwise or to avoid gratuitous short-circuit */
3788 	ring = rang ? false  : (iflib_min_tx_latency | err);
3789 	iflib_txd_db_check(txq, ring);
3790 	if_inc_counter(ifp, IFCOUNTER_OBYTES, bytes_sent);
3791 	if_inc_counter(ifp, IFCOUNTER_OPACKETS, pkt_sent);
3792 	if (mcast_sent)
3793 		if_inc_counter(ifp, IFCOUNTER_OMCASTS, mcast_sent);
3794 #ifdef INVARIANTS
3795 	if (iflib_verbose_debug)
3796 		printf("consumed=%d\n", skipped + pkt_sent);
3797 #endif
3798 	return (skipped + pkt_sent);
3799 }
3800 
3801 static uint32_t
3802 iflib_txq_drain_always(struct ifmp_ring *r)
3803 {
3804 	return (1);
3805 }
3806 
3807 static uint32_t
3808 iflib_txq_drain_free(struct ifmp_ring *r, uint32_t cidx, uint32_t pidx)
3809 {
3810 	int i, avail;
3811 	struct mbuf **mp;
3812 	iflib_txq_t txq;
3813 
3814 	txq = r->cookie;
3815 
3816 	txq->ift_qstatus = IFLIB_QUEUE_IDLE;
3817 	CALLOUT_LOCK(txq);
3818 	callout_stop(&txq->ift_timer);
3819 	CALLOUT_UNLOCK(txq);
3820 
3821 	avail = IDXDIFF(pidx, cidx, r->size);
3822 	for (i = 0; i < avail; i++) {
3823 		mp = _ring_peek_one(r, cidx, i, avail - i);
3824 		if (__predict_false(*mp == (struct mbuf *)txq))
3825 			continue;
3826 		m_freem(*mp);
3827 		DBG_COUNTER_INC(tx_frees);
3828 	}
3829 	MPASS(ifmp_ring_is_stalled(r) == 0);
3830 	return (avail);
3831 }
3832 
3833 static void
3834 iflib_ifmp_purge(iflib_txq_t txq)
3835 {
3836 	struct ifmp_ring *r;
3837 
3838 	r = txq->ift_br;
3839 	r->drain = iflib_txq_drain_free;
3840 	r->can_drain = iflib_txq_drain_always;
3841 
3842 	ifmp_ring_check_drainage(r, r->size);
3843 
3844 	r->drain = iflib_txq_drain;
3845 	r->can_drain = iflib_txq_can_drain;
3846 }
3847 
3848 static void
3849 _task_fn_tx(void *context)
3850 {
3851 	iflib_txq_t txq = context;
3852 	if_ctx_t ctx = txq->ift_ctx;
3853 	if_t ifp = ctx->ifc_ifp;
3854 	int abdicate = ctx->ifc_sysctl_tx_abdicate;
3855 
3856 #ifdef IFLIB_DIAGNOSTICS
3857 	txq->ift_cpu_exec_count[curcpu]++;
3858 #endif
3859 	if (!(if_getdrvflags(ifp) & IFF_DRV_RUNNING))
3860 		return;
3861 #ifdef DEV_NETMAP
3862 	if ((if_getcapenable(ifp) & IFCAP_NETMAP) &&
3863 	    netmap_tx_irq(ifp, txq->ift_id))
3864 		goto skip_ifmp;
3865 #endif
3866 #ifdef ALTQ
3867 	if (ALTQ_IS_ENABLED(&ifp->if_snd))
3868 		iflib_altq_if_start(ifp);
3869 #endif
3870 	if (txq->ift_db_pending)
3871 		ifmp_ring_enqueue(txq->ift_br, (void **)&txq, 1, TX_BATCH_SIZE, abdicate);
3872 	else if (!abdicate)
3873 		ifmp_ring_check_drainage(txq->ift_br, TX_BATCH_SIZE);
3874 	/*
3875 	 * When abdicating, we always need to check drainage, not just when we don't enqueue
3876 	 */
3877 	if (abdicate)
3878 		ifmp_ring_check_drainage(txq->ift_br, TX_BATCH_SIZE);
3879 #ifdef DEV_NETMAP
3880 skip_ifmp:
3881 #endif
3882 	if (ctx->ifc_flags & IFC_LEGACY)
3883 		IFDI_INTR_ENABLE(ctx);
3884 	else
3885 		IFDI_TX_QUEUE_INTR_ENABLE(ctx, txq->ift_id);
3886 }
3887 
3888 static void
3889 _task_fn_rx(void *context)
3890 {
3891 	iflib_rxq_t rxq = context;
3892 	if_ctx_t ctx = rxq->ifr_ctx;
3893 	uint8_t more;
3894 	uint16_t budget;
3895 #ifdef DEV_NETMAP
3896 	u_int work = 0;
3897 	int nmirq;
3898 #endif
3899 
3900 #ifdef IFLIB_DIAGNOSTICS
3901 	rxq->ifr_cpu_exec_count[curcpu]++;
3902 #endif
3903 	DBG_COUNTER_INC(task_fn_rxs);
3904 	if (__predict_false(!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING)))
3905 		return;
3906 #ifdef DEV_NETMAP
3907 	nmirq = netmap_rx_irq(ctx->ifc_ifp, rxq->ifr_id, &work);
3908 	if (nmirq != NM_IRQ_PASS) {
3909 		more = (nmirq == NM_IRQ_RESCHED) ? IFLIB_RXEOF_MORE : 0;
3910 		goto skip_rxeof;
3911 	}
3912 #endif
3913 	budget = ctx->ifc_sysctl_rx_budget;
3914 	if (budget == 0)
3915 		budget = 16;	/* XXX */
3916 	more = iflib_rxeof(rxq, budget);
3917 #ifdef DEV_NETMAP
3918 skip_rxeof:
3919 #endif
3920 	if ((more & IFLIB_RXEOF_MORE) == 0) {
3921 		if (ctx->ifc_flags & IFC_LEGACY)
3922 			IFDI_INTR_ENABLE(ctx);
3923 		else
3924 			IFDI_RX_QUEUE_INTR_ENABLE(ctx, rxq->ifr_id);
3925 		DBG_COUNTER_INC(rx_intr_enables);
3926 	}
3927 	if (__predict_false(!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING)))
3928 		return;
3929 
3930 	if (more & IFLIB_RXEOF_MORE)
3931 		GROUPTASK_ENQUEUE(&rxq->ifr_task);
3932 	else if (more & IFLIB_RXEOF_EMPTY)
3933 		callout_reset_curcpu(&rxq->ifr_watchdog, 1, &_task_fn_rx_watchdog, rxq);
3934 }
3935 
3936 static void
3937 _task_fn_admin(void *context)
3938 {
3939 	if_ctx_t ctx = context;
3940 	if_softc_ctx_t sctx = &ctx->ifc_softc_ctx;
3941 	iflib_txq_t txq;
3942 	int i;
3943 	bool oactive, running, do_reset, do_watchdog, in_detach;
3944 
3945 	STATE_LOCK(ctx);
3946 	running = (if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING);
3947 	oactive = (if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_OACTIVE);
3948 	do_reset = (ctx->ifc_flags & IFC_DO_RESET);
3949 	do_watchdog = (ctx->ifc_flags & IFC_DO_WATCHDOG);
3950 	in_detach = (ctx->ifc_flags & IFC_IN_DETACH);
3951 	ctx->ifc_flags &= ~(IFC_DO_RESET|IFC_DO_WATCHDOG);
3952 	STATE_UNLOCK(ctx);
3953 
3954 	if ((!running && !oactive) && !(ctx->ifc_sctx->isc_flags & IFLIB_ADMIN_ALWAYS_RUN))
3955 		return;
3956 	if (in_detach)
3957 		return;
3958 
3959 	CTX_LOCK(ctx);
3960 	for (txq = ctx->ifc_txqs, i = 0; i < sctx->isc_ntxqsets; i++, txq++) {
3961 		CALLOUT_LOCK(txq);
3962 		callout_stop(&txq->ift_timer);
3963 		CALLOUT_UNLOCK(txq);
3964 	}
3965 	if (do_watchdog) {
3966 		ctx->ifc_watchdog_events++;
3967 		IFDI_WATCHDOG_RESET(ctx);
3968 	}
3969 	IFDI_UPDATE_ADMIN_STATUS(ctx);
3970 	for (txq = ctx->ifc_txqs, i = 0; i < sctx->isc_ntxqsets; i++, txq++) {
3971 		callout_reset_on(&txq->ift_timer, iflib_timer_default, iflib_timer, txq,
3972 		    txq->ift_timer.c_cpu);
3973 	}
3974 	IFDI_LINK_INTR_ENABLE(ctx);
3975 	if (do_reset)
3976 		iflib_if_init_locked(ctx);
3977 	CTX_UNLOCK(ctx);
3978 
3979 	if (LINK_ACTIVE(ctx) == 0)
3980 		return;
3981 	for (txq = ctx->ifc_txqs, i = 0; i < sctx->isc_ntxqsets; i++, txq++)
3982 		iflib_txq_check_drain(txq, IFLIB_RESTART_BUDGET);
3983 }
3984 
3985 static void
3986 _task_fn_iov(void *context)
3987 {
3988 	if_ctx_t ctx = context;
3989 
3990 	if (!(if_getdrvflags(ctx->ifc_ifp) & IFF_DRV_RUNNING) &&
3991 	    !(ctx->ifc_sctx->isc_flags & IFLIB_ADMIN_ALWAYS_RUN))
3992 		return;
3993 
3994 	CTX_LOCK(ctx);
3995 	IFDI_VFLR_HANDLE(ctx);
3996 	CTX_UNLOCK(ctx);
3997 }
3998 
3999 static int
4000 iflib_sysctl_int_delay(SYSCTL_HANDLER_ARGS)
4001 {
4002 	int err;
4003 	if_int_delay_info_t info;
4004 	if_ctx_t ctx;
4005 
4006 	info = (if_int_delay_info_t)arg1;
4007 	ctx = info->iidi_ctx;
4008 	info->iidi_req = req;
4009 	info->iidi_oidp = oidp;
4010 	CTX_LOCK(ctx);
4011 	err = IFDI_SYSCTL_INT_DELAY(ctx, info);
4012 	CTX_UNLOCK(ctx);
4013 	return (err);
4014 }
4015 
4016 /*********************************************************************
4017  *
4018  *  IFNET FUNCTIONS
4019  *
4020  **********************************************************************/
4021 
4022 static void
4023 iflib_if_init_locked(if_ctx_t ctx)
4024 {
4025 	iflib_stop(ctx);
4026 	iflib_init_locked(ctx);
4027 }
4028 
4029 static void
4030 iflib_if_init(void *arg)
4031 {
4032 	if_ctx_t ctx = arg;
4033 
4034 	CTX_LOCK(ctx);
4035 	iflib_if_init_locked(ctx);
4036 	CTX_UNLOCK(ctx);
4037 }
4038 
4039 static int
4040 iflib_if_transmit(if_t ifp, struct mbuf *m)
4041 {
4042 	if_ctx_t	ctx = if_getsoftc(ifp);
4043 
4044 	iflib_txq_t txq;
4045 	int err, qidx;
4046 	int abdicate = ctx->ifc_sysctl_tx_abdicate;
4047 
4048 	if (__predict_false((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 || !LINK_ACTIVE(ctx))) {
4049 		DBG_COUNTER_INC(tx_frees);
4050 		m_freem(m);
4051 		return (ENETDOWN);
4052 	}
4053 
4054 	MPASS(m->m_nextpkt == NULL);
4055 	/* ALTQ-enabled interfaces always use queue 0. */
4056 	qidx = 0;
4057 	if ((NTXQSETS(ctx) > 1) && M_HASHTYPE_GET(m) && !ALTQ_IS_ENABLED(&ifp->if_snd))
4058 		qidx = QIDX(ctx, m);
4059 	/*
4060 	 * XXX calculate buf_ring based on flowid (divvy up bits?)
4061 	 */
4062 	txq = &ctx->ifc_txqs[qidx];
4063 
4064 #ifdef DRIVER_BACKPRESSURE
4065 	if (txq->ift_closed) {
4066 		while (m != NULL) {
4067 			next = m->m_nextpkt;
4068 			m->m_nextpkt = NULL;
4069 			m_freem(m);
4070 			DBG_COUNTER_INC(tx_frees);
4071 			m = next;
4072 		}
4073 		return (ENOBUFS);
4074 	}
4075 #endif
4076 #ifdef notyet
4077 	qidx = count = 0;
4078 	mp = marr;
4079 	next = m;
4080 	do {
4081 		count++;
4082 		next = next->m_nextpkt;
4083 	} while (next != NULL);
4084 
4085 	if (count > nitems(marr))
4086 		if ((mp = malloc(count*sizeof(struct mbuf *), M_IFLIB, M_NOWAIT)) == NULL) {
4087 			/* XXX check nextpkt */
4088 			m_freem(m);
4089 			/* XXX simplify for now */
4090 			DBG_COUNTER_INC(tx_frees);
4091 			return (ENOBUFS);
4092 		}
4093 	for (next = m, i = 0; next != NULL; i++) {
4094 		mp[i] = next;
4095 		next = next->m_nextpkt;
4096 		mp[i]->m_nextpkt = NULL;
4097 	}
4098 #endif
4099 	DBG_COUNTER_INC(tx_seen);
4100 	err = ifmp_ring_enqueue(txq->ift_br, (void **)&m, 1, TX_BATCH_SIZE, abdicate);
4101 
4102 	if (abdicate)
4103 		GROUPTASK_ENQUEUE(&txq->ift_task);
4104  	if (err) {
4105 		if (!abdicate)
4106 			GROUPTASK_ENQUEUE(&txq->ift_task);
4107 		/* support forthcoming later */
4108 #ifdef DRIVER_BACKPRESSURE
4109 		txq->ift_closed = TRUE;
4110 #endif
4111 		ifmp_ring_check_drainage(txq->ift_br, TX_BATCH_SIZE);
4112 		m_freem(m);
4113 		DBG_COUNTER_INC(tx_frees);
4114 	}
4115 
4116 	return (err);
4117 }
4118 
4119 #ifdef ALTQ
4120 /*
4121  * The overall approach to integrating iflib with ALTQ is to continue to use
4122  * the iflib mp_ring machinery between the ALTQ queue(s) and the hardware
4123  * ring.  Technically, when using ALTQ, queueing to an intermediate mp_ring
4124  * is redundant/unnecessary, but doing so minimizes the amount of
4125  * ALTQ-specific code required in iflib.  It is assumed that the overhead of
4126  * redundantly queueing to an intermediate mp_ring is swamped by the
4127  * performance limitations inherent in using ALTQ.
4128  *
4129  * When ALTQ support is compiled in, all iflib drivers will use a transmit
4130  * routine, iflib_altq_if_transmit(), that checks if ALTQ is enabled for the
4131  * given interface.  If ALTQ is enabled for an interface, then all
4132  * transmitted packets for that interface will be submitted to the ALTQ
4133  * subsystem via IFQ_ENQUEUE().  We don't use the legacy if_transmit()
4134  * implementation because it uses IFQ_HANDOFF(), which will duplicatively
4135  * update stats that the iflib machinery handles, and which is sensitve to
4136  * the disused IFF_DRV_OACTIVE flag.  Additionally, iflib_altq_if_start()
4137  * will be installed as the start routine for use by ALTQ facilities that
4138  * need to trigger queue drains on a scheduled basis.
4139  *
4140  */
4141 static void
4142 iflib_altq_if_start(if_t ifp)
4143 {
4144 	struct ifaltq *ifq = &ifp->if_snd;
4145 	struct mbuf *m;
4146 
4147 	IFQ_LOCK(ifq);
4148 	IFQ_DEQUEUE_NOLOCK(ifq, m);
4149 	while (m != NULL) {
4150 		iflib_if_transmit(ifp, m);
4151 		IFQ_DEQUEUE_NOLOCK(ifq, m);
4152 	}
4153 	IFQ_UNLOCK(ifq);
4154 }
4155 
4156 static int
4157 iflib_altq_if_transmit(if_t ifp, struct mbuf *m)
4158 {
4159 	int err;
4160 
4161 	if (ALTQ_IS_ENABLED(&ifp->if_snd)) {
4162 		IFQ_ENQUEUE(&ifp->if_snd, m, err);
4163 		if (err == 0)
4164 			iflib_altq_if_start(ifp);
4165 	} else
4166 		err = iflib_if_transmit(ifp, m);
4167 
4168 	return (err);
4169 }
4170 #endif /* ALTQ */
4171 
4172 static void
4173 iflib_if_qflush(if_t ifp)
4174 {
4175 	if_ctx_t ctx = if_getsoftc(ifp);
4176 	iflib_txq_t txq = ctx->ifc_txqs;
4177 	int i;
4178 
4179 	STATE_LOCK(ctx);
4180 	ctx->ifc_flags |= IFC_QFLUSH;
4181 	STATE_UNLOCK(ctx);
4182 	for (i = 0; i < NTXQSETS(ctx); i++, txq++)
4183 		while (!(ifmp_ring_is_idle(txq->ift_br) || ifmp_ring_is_stalled(txq->ift_br)))
4184 			iflib_txq_check_drain(txq, 0);
4185 	STATE_LOCK(ctx);
4186 	ctx->ifc_flags &= ~IFC_QFLUSH;
4187 	STATE_UNLOCK(ctx);
4188 
4189 	/*
4190 	 * When ALTQ is enabled, this will also take care of purging the
4191 	 * ALTQ queue(s).
4192 	 */
4193 	if_qflush(ifp);
4194 }
4195 
4196 #define IFCAP_FLAGS (IFCAP_HWCSUM_IPV6 | IFCAP_HWCSUM | IFCAP_LRO | \
4197 		     IFCAP_TSO | IFCAP_VLAN_HWTAGGING | IFCAP_HWSTATS | \
4198 		     IFCAP_VLAN_MTU | IFCAP_VLAN_HWFILTER | \
4199 		     IFCAP_VLAN_HWTSO | IFCAP_VLAN_HWCSUM | IFCAP_NOMAP)
4200 
4201 static int
4202 iflib_if_ioctl(if_t ifp, u_long command, caddr_t data)
4203 {
4204 	if_ctx_t ctx = if_getsoftc(ifp);
4205 	struct ifreq	*ifr = (struct ifreq *)data;
4206 #if defined(INET) || defined(INET6)
4207 	struct ifaddr	*ifa = (struct ifaddr *)data;
4208 #endif
4209 	bool		avoid_reset = false;
4210 	int		err = 0, reinit = 0, bits;
4211 
4212 	switch (command) {
4213 	case SIOCSIFADDR:
4214 #ifdef INET
4215 		if (ifa->ifa_addr->sa_family == AF_INET)
4216 			avoid_reset = true;
4217 #endif
4218 #ifdef INET6
4219 		if (ifa->ifa_addr->sa_family == AF_INET6)
4220 			avoid_reset = true;
4221 #endif
4222 		/*
4223 		** Calling init results in link renegotiation,
4224 		** so we avoid doing it when possible.
4225 		*/
4226 		if (avoid_reset) {
4227 			if_setflagbits(ifp, IFF_UP,0);
4228 			if (!(if_getdrvflags(ifp) & IFF_DRV_RUNNING))
4229 				reinit = 1;
4230 #ifdef INET
4231 			if (!(if_getflags(ifp) & IFF_NOARP))
4232 				arp_ifinit(ifp, ifa);
4233 #endif
4234 		} else
4235 			err = ether_ioctl(ifp, command, data);
4236 		break;
4237 	case SIOCSIFMTU:
4238 		CTX_LOCK(ctx);
4239 		if (ifr->ifr_mtu == if_getmtu(ifp)) {
4240 			CTX_UNLOCK(ctx);
4241 			break;
4242 		}
4243 		bits = if_getdrvflags(ifp);
4244 		/* stop the driver and free any clusters before proceeding */
4245 		iflib_stop(ctx);
4246 
4247 		if ((err = IFDI_MTU_SET(ctx, ifr->ifr_mtu)) == 0) {
4248 			STATE_LOCK(ctx);
4249 			if (ifr->ifr_mtu > ctx->ifc_max_fl_buf_size)
4250 				ctx->ifc_flags |= IFC_MULTISEG;
4251 			else
4252 				ctx->ifc_flags &= ~IFC_MULTISEG;
4253 			STATE_UNLOCK(ctx);
4254 			err = if_setmtu(ifp, ifr->ifr_mtu);
4255 		}
4256 		iflib_init_locked(ctx);
4257 		STATE_LOCK(ctx);
4258 		if_setdrvflags(ifp, bits);
4259 		STATE_UNLOCK(ctx);
4260 		CTX_UNLOCK(ctx);
4261 		break;
4262 	case SIOCSIFFLAGS:
4263 		CTX_LOCK(ctx);
4264 		if (if_getflags(ifp) & IFF_UP) {
4265 			if (if_getdrvflags(ifp) & IFF_DRV_RUNNING) {
4266 				if ((if_getflags(ifp) ^ ctx->ifc_if_flags) &
4267 				    (IFF_PROMISC | IFF_ALLMULTI)) {
4268 					CTX_UNLOCK(ctx);
4269 					err = IFDI_PROMISC_SET(ctx, if_getflags(ifp));
4270 					CTX_LOCK(ctx);
4271 				}
4272 			} else
4273 				reinit = 1;
4274 		} else if (if_getdrvflags(ifp) & IFF_DRV_RUNNING) {
4275 			iflib_stop(ctx);
4276 		}
4277 		ctx->ifc_if_flags = if_getflags(ifp);
4278 		CTX_UNLOCK(ctx);
4279 		break;
4280 	case SIOCADDMULTI:
4281 	case SIOCDELMULTI:
4282 		if (if_getdrvflags(ifp) & IFF_DRV_RUNNING) {
4283 			CTX_LOCK(ctx);
4284 			IFDI_INTR_DISABLE(ctx);
4285 			IFDI_MULTI_SET(ctx);
4286 			IFDI_INTR_ENABLE(ctx);
4287 			CTX_UNLOCK(ctx);
4288 		}
4289 		break;
4290 	case SIOCSIFMEDIA:
4291 		CTX_LOCK(ctx);
4292 		IFDI_MEDIA_SET(ctx);
4293 		CTX_UNLOCK(ctx);
4294 		/* FALLTHROUGH */
4295 	case SIOCGIFMEDIA:
4296 	case SIOCGIFXMEDIA:
4297 		err = ifmedia_ioctl(ifp, ifr, ctx->ifc_mediap, command);
4298 		break;
4299 	case SIOCGI2C:
4300 	{
4301 		struct ifi2creq i2c;
4302 
4303 		err = copyin(ifr_data_get_ptr(ifr), &i2c, sizeof(i2c));
4304 		if (err != 0)
4305 			break;
4306 		if (i2c.dev_addr != 0xA0 && i2c.dev_addr != 0xA2) {
4307 			err = EINVAL;
4308 			break;
4309 		}
4310 		if (i2c.len > sizeof(i2c.data)) {
4311 			err = EINVAL;
4312 			break;
4313 		}
4314 
4315 		if ((err = IFDI_I2C_REQ(ctx, &i2c)) == 0)
4316 			err = copyout(&i2c, ifr_data_get_ptr(ifr),
4317 			    sizeof(i2c));
4318 		break;
4319 	}
4320 	case SIOCSIFCAP:
4321 	{
4322 		int mask, setmask, oldmask;
4323 
4324 		oldmask = if_getcapenable(ifp);
4325 		mask = ifr->ifr_reqcap ^ oldmask;
4326 		mask &= ctx->ifc_softc_ctx.isc_capabilities | IFCAP_NOMAP;
4327 		setmask = 0;
4328 #ifdef TCP_OFFLOAD
4329 		setmask |= mask & (IFCAP_TOE4|IFCAP_TOE6);
4330 #endif
4331 		setmask |= (mask & IFCAP_FLAGS);
4332 		setmask |= (mask & IFCAP_WOL);
4333 
4334 		/*
4335 		 * If any RX csum has changed, change all the ones that
4336 		 * are supported by the driver.
4337 		 */
4338 		if (setmask & (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6)) {
4339 			setmask |= ctx->ifc_softc_ctx.isc_capabilities &
4340 			    (IFCAP_RXCSUM | IFCAP_RXCSUM_IPV6);
4341 		}
4342 
4343 		/*
4344 		 * want to ensure that traffic has stopped before we change any of the flags
4345 		 */
4346 		if (setmask) {
4347 			CTX_LOCK(ctx);
4348 			bits = if_getdrvflags(ifp);
4349 			if (bits & IFF_DRV_RUNNING && setmask & ~IFCAP_WOL)
4350 				iflib_stop(ctx);
4351 			STATE_LOCK(ctx);
4352 			if_togglecapenable(ifp, setmask);
4353 			STATE_UNLOCK(ctx);
4354 			if (bits & IFF_DRV_RUNNING && setmask & ~IFCAP_WOL)
4355 				iflib_init_locked(ctx);
4356 			STATE_LOCK(ctx);
4357 			if_setdrvflags(ifp, bits);
4358 			STATE_UNLOCK(ctx);
4359 			CTX_UNLOCK(ctx);
4360 		}
4361 		if_vlancap(ifp);
4362 		break;
4363 	}
4364 	case SIOCGPRIVATE_0:
4365 	case SIOCSDRVSPEC:
4366 	case SIOCGDRVSPEC:
4367 		CTX_LOCK(ctx);
4368 		err = IFDI_PRIV_IOCTL(ctx, command, data);
4369 		CTX_UNLOCK(ctx);
4370 		break;
4371 	default:
4372 		err = ether_ioctl(ifp, command, data);
4373 		break;
4374 	}
4375 	if (reinit)
4376 		iflib_if_init(ctx);
4377 	return (err);
4378 }
4379 
4380 static uint64_t
4381 iflib_if_get_counter(if_t ifp, ift_counter cnt)
4382 {
4383 	if_ctx_t ctx = if_getsoftc(ifp);
4384 
4385 	return (IFDI_GET_COUNTER(ctx, cnt));
4386 }
4387 
4388 /*********************************************************************
4389  *
4390  *  OTHER FUNCTIONS EXPORTED TO THE STACK
4391  *
4392  **********************************************************************/
4393 
4394 static void
4395 iflib_vlan_register(void *arg, if_t ifp, uint16_t vtag)
4396 {
4397 	if_ctx_t ctx = if_getsoftc(ifp);
4398 
4399 	if ((void *)ctx != arg)
4400 		return;
4401 
4402 	if ((vtag == 0) || (vtag > 4095))
4403 		return;
4404 
4405 	if (iflib_in_detach(ctx))
4406 		return;
4407 
4408 	CTX_LOCK(ctx);
4409 	/* Driver may need all untagged packets to be flushed */
4410 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4411 		iflib_stop(ctx);
4412 	IFDI_VLAN_REGISTER(ctx, vtag);
4413 	/* Re-init to load the changes, if required */
4414 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4415 		iflib_init_locked(ctx);
4416 	CTX_UNLOCK(ctx);
4417 }
4418 
4419 static void
4420 iflib_vlan_unregister(void *arg, if_t ifp, uint16_t vtag)
4421 {
4422 	if_ctx_t ctx = if_getsoftc(ifp);
4423 
4424 	if ((void *)ctx != arg)
4425 		return;
4426 
4427 	if ((vtag == 0) || (vtag > 4095))
4428 		return;
4429 
4430 	CTX_LOCK(ctx);
4431 	/* Driver may need all tagged packets to be flushed */
4432 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4433 		iflib_stop(ctx);
4434 	IFDI_VLAN_UNREGISTER(ctx, vtag);
4435 	/* Re-init to load the changes, if required */
4436 	if (IFDI_NEEDS_RESTART(ctx, IFLIB_RESTART_VLAN_CONFIG))
4437 		iflib_init_locked(ctx);
4438 	CTX_UNLOCK(ctx);
4439 }
4440 
4441 static void
4442 iflib_led_func(void *arg, int onoff)
4443 {
4444 	if_ctx_t ctx = arg;
4445 
4446 	CTX_LOCK(ctx);
4447 	IFDI_LED_FUNC(ctx, onoff);
4448 	CTX_UNLOCK(ctx);
4449 }
4450 
4451 /*********************************************************************
4452  *
4453  *  BUS FUNCTION DEFINITIONS
4454  *
4455  **********************************************************************/
4456 
4457 int
4458 iflib_device_probe(device_t dev)
4459 {
4460 	const pci_vendor_info_t *ent;
4461 	if_shared_ctx_t sctx;
4462 	uint16_t pci_device_id, pci_rev_id, pci_subdevice_id, pci_subvendor_id;
4463 	uint16_t pci_vendor_id;
4464 
4465 	if ((sctx = DEVICE_REGISTER(dev)) == NULL || sctx->isc_magic != IFLIB_MAGIC)
4466 		return (ENOTSUP);
4467 
4468 	pci_vendor_id = pci_get_vendor(dev);
4469 	pci_device_id = pci_get_device(dev);
4470 	pci_subvendor_id = pci_get_subvendor(dev);
4471 	pci_subdevice_id = pci_get_subdevice(dev);
4472 	pci_rev_id = pci_get_revid(dev);
4473 	if (sctx->isc_parse_devinfo != NULL)
4474 		sctx->isc_parse_devinfo(&pci_device_id, &pci_subvendor_id, &pci_subdevice_id, &pci_rev_id);
4475 
4476 	ent = sctx->isc_vendor_info;
4477 	while (ent->pvi_vendor_id != 0) {
4478 		if (pci_vendor_id != ent->pvi_vendor_id) {
4479 			ent++;
4480 			continue;
4481 		}
4482 		if ((pci_device_id == ent->pvi_device_id) &&
4483 		    ((pci_subvendor_id == ent->pvi_subvendor_id) ||
4484 		     (ent->pvi_subvendor_id == 0)) &&
4485 		    ((pci_subdevice_id == ent->pvi_subdevice_id) ||
4486 		     (ent->pvi_subdevice_id == 0)) &&
4487 		    ((pci_rev_id == ent->pvi_rev_id) ||
4488 		     (ent->pvi_rev_id == 0))) {
4489 			device_set_desc_copy(dev, ent->pvi_name);
4490 			/* this needs to be changed to zero if the bus probing code
4491 			 * ever stops re-probing on best match because the sctx
4492 			 * may have its values over written by register calls
4493 			 * in subsequent probes
4494 			 */
4495 			return (BUS_PROBE_DEFAULT);
4496 		}
4497 		ent++;
4498 	}
4499 	return (ENXIO);
4500 }
4501 
4502 int
4503 iflib_device_probe_vendor(device_t dev)
4504 {
4505 	int probe;
4506 
4507 	probe = iflib_device_probe(dev);
4508 	if (probe == BUS_PROBE_DEFAULT)
4509 		return (BUS_PROBE_VENDOR);
4510 	else
4511 		return (probe);
4512 }
4513 
4514 static void
4515 iflib_reset_qvalues(if_ctx_t ctx)
4516 {
4517 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
4518 	if_shared_ctx_t sctx = ctx->ifc_sctx;
4519 	device_t dev = ctx->ifc_dev;
4520 	int i;
4521 
4522 	if (ctx->ifc_sysctl_ntxqs != 0)
4523 		scctx->isc_ntxqsets = ctx->ifc_sysctl_ntxqs;
4524 	if (ctx->ifc_sysctl_nrxqs != 0)
4525 		scctx->isc_nrxqsets = ctx->ifc_sysctl_nrxqs;
4526 
4527 	for (i = 0; i < sctx->isc_ntxqs; i++) {
4528 		if (ctx->ifc_sysctl_ntxds[i] != 0)
4529 			scctx->isc_ntxd[i] = ctx->ifc_sysctl_ntxds[i];
4530 		else
4531 			scctx->isc_ntxd[i] = sctx->isc_ntxd_default[i];
4532 	}
4533 
4534 	for (i = 0; i < sctx->isc_nrxqs; i++) {
4535 		if (ctx->ifc_sysctl_nrxds[i] != 0)
4536 			scctx->isc_nrxd[i] = ctx->ifc_sysctl_nrxds[i];
4537 		else
4538 			scctx->isc_nrxd[i] = sctx->isc_nrxd_default[i];
4539 	}
4540 
4541 	for (i = 0; i < sctx->isc_nrxqs; i++) {
4542 		if (scctx->isc_nrxd[i] < sctx->isc_nrxd_min[i]) {
4543 			device_printf(dev, "nrxd%d: %d less than nrxd_min %d - resetting to min\n",
4544 				      i, scctx->isc_nrxd[i], sctx->isc_nrxd_min[i]);
4545 			scctx->isc_nrxd[i] = sctx->isc_nrxd_min[i];
4546 		}
4547 		if (scctx->isc_nrxd[i] > sctx->isc_nrxd_max[i]) {
4548 			device_printf(dev, "nrxd%d: %d greater than nrxd_max %d - resetting to max\n",
4549 				      i, scctx->isc_nrxd[i], sctx->isc_nrxd_max[i]);
4550 			scctx->isc_nrxd[i] = sctx->isc_nrxd_max[i];
4551 		}
4552 		if (!powerof2(scctx->isc_nrxd[i])) {
4553 			device_printf(dev, "nrxd%d: %d is not a power of 2 - using default value of %d\n",
4554 				      i, scctx->isc_nrxd[i], sctx->isc_nrxd_default[i]);
4555 			scctx->isc_nrxd[i] = sctx->isc_nrxd_default[i];
4556 		}
4557 	}
4558 
4559 	for (i = 0; i < sctx->isc_ntxqs; i++) {
4560 		if (scctx->isc_ntxd[i] < sctx->isc_ntxd_min[i]) {
4561 			device_printf(dev, "ntxd%d: %d less than ntxd_min %d - resetting to min\n",
4562 				      i, scctx->isc_ntxd[i], sctx->isc_ntxd_min[i]);
4563 			scctx->isc_ntxd[i] = sctx->isc_ntxd_min[i];
4564 		}
4565 		if (scctx->isc_ntxd[i] > sctx->isc_ntxd_max[i]) {
4566 			device_printf(dev, "ntxd%d: %d greater than ntxd_max %d - resetting to max\n",
4567 				      i, scctx->isc_ntxd[i], sctx->isc_ntxd_max[i]);
4568 			scctx->isc_ntxd[i] = sctx->isc_ntxd_max[i];
4569 		}
4570 		if (!powerof2(scctx->isc_ntxd[i])) {
4571 			device_printf(dev, "ntxd%d: %d is not a power of 2 - using default value of %d\n",
4572 				      i, scctx->isc_ntxd[i], sctx->isc_ntxd_default[i]);
4573 			scctx->isc_ntxd[i] = sctx->isc_ntxd_default[i];
4574 		}
4575 	}
4576 }
4577 
4578 static void
4579 iflib_add_pfil(if_ctx_t ctx)
4580 {
4581 	struct pfil_head *pfil;
4582 	struct pfil_head_args pa;
4583 	iflib_rxq_t rxq;
4584 	int i;
4585 
4586 	pa.pa_version = PFIL_VERSION;
4587 	pa.pa_flags = PFIL_IN;
4588 	pa.pa_type = PFIL_TYPE_ETHERNET;
4589 	pa.pa_headname = ctx->ifc_ifp->if_xname;
4590 	pfil = pfil_head_register(&pa);
4591 
4592 	for (i = 0, rxq = ctx->ifc_rxqs; i < NRXQSETS(ctx); i++, rxq++) {
4593 		rxq->pfil = pfil;
4594 	}
4595 }
4596 
4597 static void
4598 iflib_rem_pfil(if_ctx_t ctx)
4599 {
4600 	struct pfil_head *pfil;
4601 	iflib_rxq_t rxq;
4602 	int i;
4603 
4604 	rxq = ctx->ifc_rxqs;
4605 	pfil = rxq->pfil;
4606 	for (i = 0; i < NRXQSETS(ctx); i++, rxq++) {
4607 		rxq->pfil = NULL;
4608 	}
4609 	pfil_head_unregister(pfil);
4610 }
4611 
4612 static uint16_t
4613 get_ctx_core_offset(if_ctx_t ctx)
4614 {
4615 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
4616 	struct cpu_offset *op;
4617 	uint16_t qc;
4618 	uint16_t ret = ctx->ifc_sysctl_core_offset;
4619 
4620 	if (ret != CORE_OFFSET_UNSPECIFIED)
4621 		return (ret);
4622 
4623 	if (ctx->ifc_sysctl_separate_txrx)
4624 		qc = scctx->isc_ntxqsets + scctx->isc_nrxqsets;
4625 	else
4626 		qc = max(scctx->isc_ntxqsets, scctx->isc_nrxqsets);
4627 
4628 	mtx_lock(&cpu_offset_mtx);
4629 	SLIST_FOREACH(op, &cpu_offsets, entries) {
4630 		if (CPU_CMP(&ctx->ifc_cpus, &op->set) == 0) {
4631 			ret = op->offset;
4632 			op->offset += qc;
4633 			MPASS(op->refcount < UINT_MAX);
4634 			op->refcount++;
4635 			break;
4636 		}
4637 	}
4638 	if (ret == CORE_OFFSET_UNSPECIFIED) {
4639 		ret = 0;
4640 		op = malloc(sizeof(struct cpu_offset), M_IFLIB,
4641 		    M_NOWAIT | M_ZERO);
4642 		if (op == NULL) {
4643 			device_printf(ctx->ifc_dev,
4644 			    "allocation for cpu offset failed.\n");
4645 		} else {
4646 			op->offset = qc;
4647 			op->refcount = 1;
4648 			CPU_COPY(&ctx->ifc_cpus, &op->set);
4649 			SLIST_INSERT_HEAD(&cpu_offsets, op, entries);
4650 		}
4651 	}
4652 	mtx_unlock(&cpu_offset_mtx);
4653 
4654 	return (ret);
4655 }
4656 
4657 static void
4658 unref_ctx_core_offset(if_ctx_t ctx)
4659 {
4660 	struct cpu_offset *op, *top;
4661 
4662 	mtx_lock(&cpu_offset_mtx);
4663 	SLIST_FOREACH_SAFE(op, &cpu_offsets, entries, top) {
4664 		if (CPU_CMP(&ctx->ifc_cpus, &op->set) == 0) {
4665 			MPASS(op->refcount > 0);
4666 			op->refcount--;
4667 			if (op->refcount == 0) {
4668 				SLIST_REMOVE(&cpu_offsets, op, cpu_offset, entries);
4669 				free(op, M_IFLIB);
4670 			}
4671 			break;
4672 		}
4673 	}
4674 	mtx_unlock(&cpu_offset_mtx);
4675 }
4676 
4677 int
4678 iflib_device_register(device_t dev, void *sc, if_shared_ctx_t sctx, if_ctx_t *ctxp)
4679 {
4680 	if_ctx_t ctx;
4681 	if_t ifp;
4682 	if_softc_ctx_t scctx;
4683 	kobjop_desc_t kobj_desc;
4684 	kobj_method_t *kobj_method;
4685 	int err, msix, rid;
4686 	int num_txd, num_rxd;
4687 
4688 	ctx = malloc(sizeof(* ctx), M_IFLIB, M_WAITOK|M_ZERO);
4689 
4690 	if (sc == NULL) {
4691 		sc = malloc(sctx->isc_driver->size, M_IFLIB, M_WAITOK|M_ZERO);
4692 		device_set_softc(dev, ctx);
4693 		ctx->ifc_flags |= IFC_SC_ALLOCATED;
4694 	}
4695 
4696 	ctx->ifc_sctx = sctx;
4697 	ctx->ifc_dev = dev;
4698 	ctx->ifc_softc = sc;
4699 
4700 	if ((err = iflib_register(ctx)) != 0) {
4701 		device_printf(dev, "iflib_register failed %d\n", err);
4702 		goto fail_ctx_free;
4703 	}
4704 	iflib_add_device_sysctl_pre(ctx);
4705 
4706 	scctx = &ctx->ifc_softc_ctx;
4707 	ifp = ctx->ifc_ifp;
4708 
4709 	iflib_reset_qvalues(ctx);
4710 	CTX_LOCK(ctx);
4711 	if ((err = IFDI_ATTACH_PRE(ctx)) != 0) {
4712 		device_printf(dev, "IFDI_ATTACH_PRE failed %d\n", err);
4713 		goto fail_unlock;
4714 	}
4715 	_iflib_pre_assert(scctx);
4716 	ctx->ifc_txrx = *scctx->isc_txrx;
4717 
4718 	if (sctx->isc_flags & IFLIB_DRIVER_MEDIA)
4719 		ctx->ifc_mediap = scctx->isc_media;
4720 
4721 #ifdef INVARIANTS
4722 	if (scctx->isc_capabilities & IFCAP_TXCSUM)
4723 		MPASS(scctx->isc_tx_csum_flags);
4724 #endif
4725 
4726 	if_setcapabilities(ifp,
4727 	    scctx->isc_capabilities | IFCAP_HWSTATS | IFCAP_NOMAP);
4728 	if_setcapenable(ifp,
4729 	    scctx->isc_capenable | IFCAP_HWSTATS | IFCAP_NOMAP);
4730 
4731 	if (scctx->isc_ntxqsets == 0 || (scctx->isc_ntxqsets_max && scctx->isc_ntxqsets_max < scctx->isc_ntxqsets))
4732 		scctx->isc_ntxqsets = scctx->isc_ntxqsets_max;
4733 	if (scctx->isc_nrxqsets == 0 || (scctx->isc_nrxqsets_max && scctx->isc_nrxqsets_max < scctx->isc_nrxqsets))
4734 		scctx->isc_nrxqsets = scctx->isc_nrxqsets_max;
4735 
4736 	num_txd = iflib_num_tx_descs(ctx);
4737 	num_rxd = iflib_num_rx_descs(ctx);
4738 
4739 	/* XXX change for per-queue sizes */
4740 	device_printf(dev, "Using %d TX descriptors and %d RX descriptors\n",
4741 	    num_txd, num_rxd);
4742 
4743 	if (scctx->isc_tx_nsegments > num_txd / MAX_SINGLE_PACKET_FRACTION)
4744 		scctx->isc_tx_nsegments = max(1, num_txd /
4745 		    MAX_SINGLE_PACKET_FRACTION);
4746 	if (scctx->isc_tx_tso_segments_max > num_txd /
4747 	    MAX_SINGLE_PACKET_FRACTION)
4748 		scctx->isc_tx_tso_segments_max = max(1,
4749 		    num_txd / MAX_SINGLE_PACKET_FRACTION);
4750 
4751 	/* TSO parameters - dig these out of the data sheet - simply correspond to tag setup */
4752 	if (if_getcapabilities(ifp) & IFCAP_TSO) {
4753 		/*
4754 		 * The stack can't handle a TSO size larger than IP_MAXPACKET,
4755 		 * but some MACs do.
4756 		 */
4757 		if_sethwtsomax(ifp, min(scctx->isc_tx_tso_size_max,
4758 		    IP_MAXPACKET));
4759 		/*
4760 		 * Take maximum number of m_pullup(9)'s in iflib_parse_header()
4761 		 * into account.  In the worst case, each of these calls will
4762 		 * add another mbuf and, thus, the requirement for another DMA
4763 		 * segment.  So for best performance, it doesn't make sense to
4764 		 * advertize a maximum of TSO segments that typically will
4765 		 * require defragmentation in iflib_encap().
4766 		 */
4767 		if_sethwtsomaxsegcount(ifp, scctx->isc_tx_tso_segments_max - 3);
4768 		if_sethwtsomaxsegsize(ifp, scctx->isc_tx_tso_segsize_max);
4769 	}
4770 	if (scctx->isc_rss_table_size == 0)
4771 		scctx->isc_rss_table_size = 64;
4772 	scctx->isc_rss_table_mask = scctx->isc_rss_table_size-1;
4773 
4774 	GROUPTASK_INIT(&ctx->ifc_admin_task, 0, _task_fn_admin, ctx);
4775 	/* XXX format name */
4776 	taskqgroup_attach(qgroup_if_config_tqg, &ctx->ifc_admin_task, ctx,
4777 	    NULL, NULL, "admin");
4778 
4779 	/* Set up cpu set.  If it fails, use the set of all CPUs. */
4780 	if (bus_get_cpus(dev, INTR_CPUS, sizeof(ctx->ifc_cpus), &ctx->ifc_cpus) != 0) {
4781 		device_printf(dev, "Unable to fetch CPU list\n");
4782 		CPU_COPY(&all_cpus, &ctx->ifc_cpus);
4783 	}
4784 	MPASS(CPU_COUNT(&ctx->ifc_cpus) > 0);
4785 
4786 	/*
4787 	** Now set up MSI or MSI-X, should return us the number of supported
4788 	** vectors (will be 1 for a legacy interrupt and MSI).
4789 	*/
4790 	if (sctx->isc_flags & IFLIB_SKIP_MSIX) {
4791 		msix = scctx->isc_vectors;
4792 	} else if (scctx->isc_msix_bar != 0)
4793 	       /*
4794 		* The simple fact that isc_msix_bar is not 0 does not mean we
4795 		* we have a good value there that is known to work.
4796 		*/
4797 		msix = iflib_msix_init(ctx);
4798 	else {
4799 		scctx->isc_vectors = 1;
4800 		scctx->isc_ntxqsets = 1;
4801 		scctx->isc_nrxqsets = 1;
4802 		scctx->isc_intr = IFLIB_INTR_LEGACY;
4803 		msix = 0;
4804 	}
4805 	/* Get memory for the station queues */
4806 	if ((err = iflib_queues_alloc(ctx))) {
4807 		device_printf(dev, "Unable to allocate queue memory\n");
4808 		goto fail_intr_free;
4809 	}
4810 
4811 	if ((err = iflib_qset_structures_setup(ctx)))
4812 		goto fail_queues;
4813 
4814 	/*
4815 	 * Now that we know how many queues there are, get the core offset.
4816 	 */
4817 	ctx->ifc_sysctl_core_offset = get_ctx_core_offset(ctx);
4818 
4819 	if (msix > 1) {
4820 		/*
4821 		 * When using MSI-X, ensure that ifdi_{r,t}x_queue_intr_enable
4822 		 * aren't the default NULL implementation.
4823 		 */
4824 		kobj_desc = &ifdi_rx_queue_intr_enable_desc;
4825 		kobj_method = kobj_lookup_method(((kobj_t)ctx)->ops->cls, NULL,
4826 		    kobj_desc);
4827 		if (kobj_method == &kobj_desc->deflt) {
4828 			device_printf(dev,
4829 			    "MSI-X requires ifdi_rx_queue_intr_enable method");
4830 			err = EOPNOTSUPP;
4831 			goto fail_queues;
4832 		}
4833 		kobj_desc = &ifdi_tx_queue_intr_enable_desc;
4834 		kobj_method = kobj_lookup_method(((kobj_t)ctx)->ops->cls, NULL,
4835 		    kobj_desc);
4836 		if (kobj_method == &kobj_desc->deflt) {
4837 			device_printf(dev,
4838 			    "MSI-X requires ifdi_tx_queue_intr_enable method");
4839 			err = EOPNOTSUPP;
4840 			goto fail_queues;
4841 		}
4842 
4843 		/*
4844 		 * Assign the MSI-X vectors.
4845 		 * Note that the default NULL ifdi_msix_intr_assign method will
4846 		 * fail here, too.
4847 		 */
4848 		err = IFDI_MSIX_INTR_ASSIGN(ctx, msix);
4849 		if (err != 0) {
4850 			device_printf(dev, "IFDI_MSIX_INTR_ASSIGN failed %d\n",
4851 			    err);
4852 			goto fail_queues;
4853 		}
4854 	} else if (scctx->isc_intr != IFLIB_INTR_MSIX) {
4855 		rid = 0;
4856 		if (scctx->isc_intr == IFLIB_INTR_MSI) {
4857 			MPASS(msix == 1);
4858 			rid = 1;
4859 		}
4860 		if ((err = iflib_legacy_setup(ctx, ctx->isc_legacy_intr, ctx->ifc_softc, &rid, "irq0")) != 0) {
4861 			device_printf(dev, "iflib_legacy_setup failed %d\n", err);
4862 			goto fail_queues;
4863 		}
4864 	} else {
4865 		device_printf(dev,
4866 		    "Cannot use iflib with only 1 MSI-X interrupt!\n");
4867 		err = ENODEV;
4868 		goto fail_intr_free;
4869 	}
4870 
4871 	ether_ifattach(ctx->ifc_ifp, ctx->ifc_mac.octet);
4872 
4873 	if ((err = IFDI_ATTACH_POST(ctx)) != 0) {
4874 		device_printf(dev, "IFDI_ATTACH_POST failed %d\n", err);
4875 		goto fail_detach;
4876 	}
4877 
4878 	/*
4879 	 * Tell the upper layer(s) if IFCAP_VLAN_MTU is supported.
4880 	 * This must appear after the call to ether_ifattach() because
4881 	 * ether_ifattach() sets if_hdrlen to the default value.
4882 	 */
4883 	if (if_getcapabilities(ifp) & IFCAP_VLAN_MTU)
4884 		if_setifheaderlen(ifp, sizeof(struct ether_vlan_header));
4885 
4886 	if ((err = iflib_netmap_attach(ctx))) {
4887 		device_printf(ctx->ifc_dev, "netmap attach failed: %d\n", err);
4888 		goto fail_detach;
4889 	}
4890 	*ctxp = ctx;
4891 
4892 	DEBUGNET_SET(ctx->ifc_ifp, iflib);
4893 
4894 	if_setgetcounterfn(ctx->ifc_ifp, iflib_if_get_counter);
4895 	iflib_add_device_sysctl_post(ctx);
4896 	iflib_add_pfil(ctx);
4897 	ctx->ifc_flags |= IFC_INIT_DONE;
4898 	CTX_UNLOCK(ctx);
4899 
4900 	return (0);
4901 
4902 fail_detach:
4903 	ether_ifdetach(ctx->ifc_ifp);
4904 fail_intr_free:
4905 	iflib_free_intr_mem(ctx);
4906 fail_queues:
4907 	iflib_tx_structures_free(ctx);
4908 	iflib_rx_structures_free(ctx);
4909 	iflib_tqg_detach(ctx);
4910 	IFDI_DETACH(ctx);
4911 fail_unlock:
4912 	CTX_UNLOCK(ctx);
4913 	iflib_deregister(ctx);
4914 fail_ctx_free:
4915 	device_set_softc(ctx->ifc_dev, NULL);
4916         if (ctx->ifc_flags & IFC_SC_ALLOCATED)
4917                 free(ctx->ifc_softc, M_IFLIB);
4918         free(ctx, M_IFLIB);
4919 	return (err);
4920 }
4921 
4922 int
4923 iflib_pseudo_register(device_t dev, if_shared_ctx_t sctx, if_ctx_t *ctxp,
4924 					  struct iflib_cloneattach_ctx *clctx)
4925 {
4926 	int num_txd, num_rxd;
4927 	int err;
4928 	if_ctx_t ctx;
4929 	if_t ifp;
4930 	if_softc_ctx_t scctx;
4931 	int i;
4932 	void *sc;
4933 
4934 	ctx = malloc(sizeof(*ctx), M_IFLIB, M_WAITOK|M_ZERO);
4935 	sc = malloc(sctx->isc_driver->size, M_IFLIB, M_WAITOK|M_ZERO);
4936 	ctx->ifc_flags |= IFC_SC_ALLOCATED;
4937 	if (sctx->isc_flags & (IFLIB_PSEUDO|IFLIB_VIRTUAL))
4938 		ctx->ifc_flags |= IFC_PSEUDO;
4939 
4940 	ctx->ifc_sctx = sctx;
4941 	ctx->ifc_softc = sc;
4942 	ctx->ifc_dev = dev;
4943 
4944 	if ((err = iflib_register(ctx)) != 0) {
4945 		device_printf(dev, "%s: iflib_register failed %d\n", __func__, err);
4946 		goto fail_ctx_free;
4947 	}
4948 	iflib_add_device_sysctl_pre(ctx);
4949 
4950 	scctx = &ctx->ifc_softc_ctx;
4951 	ifp = ctx->ifc_ifp;
4952 
4953 	iflib_reset_qvalues(ctx);
4954 	CTX_LOCK(ctx);
4955 	if ((err = IFDI_ATTACH_PRE(ctx)) != 0) {
4956 		device_printf(dev, "IFDI_ATTACH_PRE failed %d\n", err);
4957 		goto fail_unlock;
4958 	}
4959 	if (sctx->isc_flags & IFLIB_GEN_MAC)
4960 		ether_gen_addr(ifp, &ctx->ifc_mac);
4961 	if ((err = IFDI_CLONEATTACH(ctx, clctx->cc_ifc, clctx->cc_name,
4962 								clctx->cc_params)) != 0) {
4963 		device_printf(dev, "IFDI_CLONEATTACH failed %d\n", err);
4964 		goto fail_unlock;
4965 	}
4966 #ifdef INVARIANTS
4967 	if (scctx->isc_capabilities & IFCAP_TXCSUM)
4968 		MPASS(scctx->isc_tx_csum_flags);
4969 #endif
4970 
4971 	if_setcapabilities(ifp, scctx->isc_capabilities | IFCAP_HWSTATS | IFCAP_LINKSTATE);
4972 	if_setcapenable(ifp, scctx->isc_capenable | IFCAP_HWSTATS | IFCAP_LINKSTATE);
4973 
4974 	ifp->if_flags |= IFF_NOGROUP;
4975 	if (sctx->isc_flags & IFLIB_PSEUDO) {
4976 		ifmedia_add(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO, 0, NULL);
4977 		ifmedia_set(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO);
4978 		if (sctx->isc_flags & IFLIB_PSEUDO_ETHER) {
4979 			ether_ifattach(ctx->ifc_ifp, ctx->ifc_mac.octet);
4980 		} else {
4981 			if_attach(ctx->ifc_ifp);
4982 			bpfattach(ctx->ifc_ifp, DLT_NULL, sizeof(u_int32_t));
4983 		}
4984 
4985 		if ((err = IFDI_ATTACH_POST(ctx)) != 0) {
4986 			device_printf(dev, "IFDI_ATTACH_POST failed %d\n", err);
4987 			goto fail_detach;
4988 		}
4989 		*ctxp = ctx;
4990 
4991 		/*
4992 		 * Tell the upper layer(s) if IFCAP_VLAN_MTU is supported.
4993 		 * This must appear after the call to ether_ifattach() because
4994 		 * ether_ifattach() sets if_hdrlen to the default value.
4995 		 */
4996 		if (if_getcapabilities(ifp) & IFCAP_VLAN_MTU)
4997 			if_setifheaderlen(ifp,
4998 			    sizeof(struct ether_vlan_header));
4999 
5000 		if_setgetcounterfn(ctx->ifc_ifp, iflib_if_get_counter);
5001 		iflib_add_device_sysctl_post(ctx);
5002 		ctx->ifc_flags |= IFC_INIT_DONE;
5003 		CTX_UNLOCK(ctx);
5004 		return (0);
5005 	}
5006 	ifmedia_add(ctx->ifc_mediap, IFM_ETHER | IFM_1000_T | IFM_FDX, 0, NULL);
5007 	ifmedia_add(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO, 0, NULL);
5008 	ifmedia_set(ctx->ifc_mediap, IFM_ETHER | IFM_AUTO);
5009 
5010 	_iflib_pre_assert(scctx);
5011 	ctx->ifc_txrx = *scctx->isc_txrx;
5012 
5013 	if (scctx->isc_ntxqsets == 0 || (scctx->isc_ntxqsets_max && scctx->isc_ntxqsets_max < scctx->isc_ntxqsets))
5014 		scctx->isc_ntxqsets = scctx->isc_ntxqsets_max;
5015 	if (scctx->isc_nrxqsets == 0 || (scctx->isc_nrxqsets_max && scctx->isc_nrxqsets_max < scctx->isc_nrxqsets))
5016 		scctx->isc_nrxqsets = scctx->isc_nrxqsets_max;
5017 
5018 	num_txd = iflib_num_tx_descs(ctx);
5019 	num_rxd = iflib_num_rx_descs(ctx);
5020 
5021 	/* XXX change for per-queue sizes */
5022 	device_printf(dev, "Using %d TX descriptors and %d RX descriptors\n",
5023 	    num_txd, num_rxd);
5024 
5025 	if (scctx->isc_tx_nsegments > num_txd / MAX_SINGLE_PACKET_FRACTION)
5026 		scctx->isc_tx_nsegments = max(1, num_txd /
5027 		    MAX_SINGLE_PACKET_FRACTION);
5028 	if (scctx->isc_tx_tso_segments_max > num_txd /
5029 	    MAX_SINGLE_PACKET_FRACTION)
5030 		scctx->isc_tx_tso_segments_max = max(1,
5031 		    num_txd / MAX_SINGLE_PACKET_FRACTION);
5032 
5033 	/* TSO parameters - dig these out of the data sheet - simply correspond to tag setup */
5034 	if (if_getcapabilities(ifp) & IFCAP_TSO) {
5035 		/*
5036 		 * The stack can't handle a TSO size larger than IP_MAXPACKET,
5037 		 * but some MACs do.
5038 		 */
5039 		if_sethwtsomax(ifp, min(scctx->isc_tx_tso_size_max,
5040 		    IP_MAXPACKET));
5041 		/*
5042 		 * Take maximum number of m_pullup(9)'s in iflib_parse_header()
5043 		 * into account.  In the worst case, each of these calls will
5044 		 * add another mbuf and, thus, the requirement for another DMA
5045 		 * segment.  So for best performance, it doesn't make sense to
5046 		 * advertize a maximum of TSO segments that typically will
5047 		 * require defragmentation in iflib_encap().
5048 		 */
5049 		if_sethwtsomaxsegcount(ifp, scctx->isc_tx_tso_segments_max - 3);
5050 		if_sethwtsomaxsegsize(ifp, scctx->isc_tx_tso_segsize_max);
5051 	}
5052 	if (scctx->isc_rss_table_size == 0)
5053 		scctx->isc_rss_table_size = 64;
5054 	scctx->isc_rss_table_mask = scctx->isc_rss_table_size-1;
5055 
5056 	GROUPTASK_INIT(&ctx->ifc_admin_task, 0, _task_fn_admin, ctx);
5057 	/* XXX format name */
5058 	taskqgroup_attach(qgroup_if_config_tqg, &ctx->ifc_admin_task, ctx,
5059 	    NULL, NULL, "admin");
5060 
5061 	/* XXX --- can support > 1 -- but keep it simple for now */
5062 	scctx->isc_intr = IFLIB_INTR_LEGACY;
5063 
5064 	/* Get memory for the station queues */
5065 	if ((err = iflib_queues_alloc(ctx))) {
5066 		device_printf(dev, "Unable to allocate queue memory\n");
5067 		goto fail_iflib_detach;
5068 	}
5069 
5070 	if ((err = iflib_qset_structures_setup(ctx))) {
5071 		device_printf(dev, "qset structure setup failed %d\n", err);
5072 		goto fail_queues;
5073 	}
5074 
5075 	/*
5076 	 * XXX What if anything do we want to do about interrupts?
5077 	 */
5078 	ether_ifattach(ctx->ifc_ifp, ctx->ifc_mac.octet);
5079 	if ((err = IFDI_ATTACH_POST(ctx)) != 0) {
5080 		device_printf(dev, "IFDI_ATTACH_POST failed %d\n", err);
5081 		goto fail_detach;
5082 	}
5083 
5084 	/*
5085 	 * Tell the upper layer(s) if IFCAP_VLAN_MTU is supported.
5086 	 * This must appear after the call to ether_ifattach() because
5087 	 * ether_ifattach() sets if_hdrlen to the default value.
5088 	 */
5089 	if (if_getcapabilities(ifp) & IFCAP_VLAN_MTU)
5090 		if_setifheaderlen(ifp, sizeof(struct ether_vlan_header));
5091 
5092 	/* XXX handle more than one queue */
5093 	for (i = 0; i < scctx->isc_nrxqsets; i++)
5094 		IFDI_RX_CLSET(ctx, 0, i, ctx->ifc_rxqs[i].ifr_fl[0].ifl_sds.ifsd_cl);
5095 
5096 	*ctxp = ctx;
5097 
5098 	if_setgetcounterfn(ctx->ifc_ifp, iflib_if_get_counter);
5099 	iflib_add_device_sysctl_post(ctx);
5100 	ctx->ifc_flags |= IFC_INIT_DONE;
5101 	CTX_UNLOCK(ctx);
5102 
5103 	return (0);
5104 fail_detach:
5105 	ether_ifdetach(ctx->ifc_ifp);
5106 fail_queues:
5107 	iflib_tx_structures_free(ctx);
5108 	iflib_rx_structures_free(ctx);
5109 	iflib_tqg_detach(ctx);
5110 fail_iflib_detach:
5111 	IFDI_DETACH(ctx);
5112 fail_unlock:
5113 	CTX_UNLOCK(ctx);
5114 	iflib_deregister(ctx);
5115 fail_ctx_free:
5116 	free(ctx->ifc_softc, M_IFLIB);
5117 	free(ctx, M_IFLIB);
5118 	return (err);
5119 }
5120 
5121 int
5122 iflib_pseudo_deregister(if_ctx_t ctx)
5123 {
5124 	if_t ifp = ctx->ifc_ifp;
5125 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5126 
5127 	/* Unregister VLAN event handlers early */
5128 	iflib_unregister_vlan_handlers(ctx);
5129 
5130 	if ((sctx->isc_flags & IFLIB_PSEUDO)  &&
5131 		(sctx->isc_flags & IFLIB_PSEUDO_ETHER) == 0) {
5132 		bpfdetach(ifp);
5133 		if_detach(ifp);
5134 	} else {
5135 		ether_ifdetach(ifp);
5136 	}
5137 
5138 	iflib_tqg_detach(ctx);
5139 	iflib_tx_structures_free(ctx);
5140 	iflib_rx_structures_free(ctx);
5141 
5142 	iflib_deregister(ctx);
5143 
5144 	if (ctx->ifc_flags & IFC_SC_ALLOCATED)
5145 		free(ctx->ifc_softc, M_IFLIB);
5146 	free(ctx, M_IFLIB);
5147 	return (0);
5148 }
5149 
5150 int
5151 iflib_device_attach(device_t dev)
5152 {
5153 	if_ctx_t ctx;
5154 	if_shared_ctx_t sctx;
5155 
5156 	if ((sctx = DEVICE_REGISTER(dev)) == NULL || sctx->isc_magic != IFLIB_MAGIC)
5157 		return (ENOTSUP);
5158 
5159 	pci_enable_busmaster(dev);
5160 
5161 	return (iflib_device_register(dev, NULL, sctx, &ctx));
5162 }
5163 
5164 int
5165 iflib_device_deregister(if_ctx_t ctx)
5166 {
5167 	if_t ifp = ctx->ifc_ifp;
5168 	device_t dev = ctx->ifc_dev;
5169 
5170 	/* Make sure VLANS are not using driver */
5171 	if (if_vlantrunkinuse(ifp)) {
5172 		device_printf(dev, "Vlan in use, detach first\n");
5173 		return (EBUSY);
5174 	}
5175 #ifdef PCI_IOV
5176 	if (!CTX_IS_VF(ctx) && pci_iov_detach(dev) != 0) {
5177 		device_printf(dev, "SR-IOV in use; detach first.\n");
5178 		return (EBUSY);
5179 	}
5180 #endif
5181 
5182 	STATE_LOCK(ctx);
5183 	ctx->ifc_flags |= IFC_IN_DETACH;
5184 	STATE_UNLOCK(ctx);
5185 
5186 	/* Unregister VLAN handlers before calling iflib_stop() */
5187 	iflib_unregister_vlan_handlers(ctx);
5188 
5189 	iflib_netmap_detach(ifp);
5190 	ether_ifdetach(ifp);
5191 
5192 	CTX_LOCK(ctx);
5193 	iflib_stop(ctx);
5194 	CTX_UNLOCK(ctx);
5195 
5196 	iflib_rem_pfil(ctx);
5197 	if (ctx->ifc_led_dev != NULL)
5198 		led_destroy(ctx->ifc_led_dev);
5199 
5200 	iflib_tqg_detach(ctx);
5201 	CTX_LOCK(ctx);
5202 	IFDI_DETACH(ctx);
5203 	CTX_UNLOCK(ctx);
5204 
5205 	/* ether_ifdetach calls if_qflush - lock must be destroy afterwards*/
5206 	iflib_free_intr_mem(ctx);
5207 
5208 	bus_generic_detach(dev);
5209 
5210 	iflib_tx_structures_free(ctx);
5211 	iflib_rx_structures_free(ctx);
5212 
5213 	iflib_deregister(ctx);
5214 
5215 	device_set_softc(ctx->ifc_dev, NULL);
5216 	if (ctx->ifc_flags & IFC_SC_ALLOCATED)
5217 		free(ctx->ifc_softc, M_IFLIB);
5218 	unref_ctx_core_offset(ctx);
5219 	free(ctx, M_IFLIB);
5220 	return (0);
5221 }
5222 
5223 static void
5224 iflib_tqg_detach(if_ctx_t ctx)
5225 {
5226 	iflib_txq_t txq;
5227 	iflib_rxq_t rxq;
5228 	int i;
5229 	struct taskqgroup *tqg;
5230 
5231 	/* XXX drain any dependent tasks */
5232 	tqg = qgroup_if_io_tqg;
5233 	for (txq = ctx->ifc_txqs, i = 0; i < NTXQSETS(ctx); i++, txq++) {
5234 		callout_drain(&txq->ift_timer);
5235 #ifdef DEV_NETMAP
5236 		callout_drain(&txq->ift_netmap_timer);
5237 #endif /* DEV_NETMAP */
5238 		if (txq->ift_task.gt_uniq != NULL)
5239 			taskqgroup_detach(tqg, &txq->ift_task);
5240 	}
5241 	for (i = 0, rxq = ctx->ifc_rxqs; i < NRXQSETS(ctx); i++, rxq++) {
5242 		if (rxq->ifr_task.gt_uniq != NULL)
5243 			taskqgroup_detach(tqg, &rxq->ifr_task);
5244 	}
5245 	tqg = qgroup_if_config_tqg;
5246 	if (ctx->ifc_admin_task.gt_uniq != NULL)
5247 		taskqgroup_detach(tqg, &ctx->ifc_admin_task);
5248 	if (ctx->ifc_vflr_task.gt_uniq != NULL)
5249 		taskqgroup_detach(tqg, &ctx->ifc_vflr_task);
5250 }
5251 
5252 static void
5253 iflib_free_intr_mem(if_ctx_t ctx)
5254 {
5255 
5256 	if (ctx->ifc_softc_ctx.isc_intr != IFLIB_INTR_MSIX) {
5257 		iflib_irq_free(ctx, &ctx->ifc_legacy_irq);
5258 	}
5259 	if (ctx->ifc_softc_ctx.isc_intr != IFLIB_INTR_LEGACY) {
5260 		pci_release_msi(ctx->ifc_dev);
5261 	}
5262 	if (ctx->ifc_msix_mem != NULL) {
5263 		bus_release_resource(ctx->ifc_dev, SYS_RES_MEMORY,
5264 		    rman_get_rid(ctx->ifc_msix_mem), ctx->ifc_msix_mem);
5265 		ctx->ifc_msix_mem = NULL;
5266 	}
5267 }
5268 
5269 int
5270 iflib_device_detach(device_t dev)
5271 {
5272 	if_ctx_t ctx = device_get_softc(dev);
5273 
5274 	return (iflib_device_deregister(ctx));
5275 }
5276 
5277 int
5278 iflib_device_suspend(device_t dev)
5279 {
5280 	if_ctx_t ctx = device_get_softc(dev);
5281 
5282 	CTX_LOCK(ctx);
5283 	IFDI_SUSPEND(ctx);
5284 	CTX_UNLOCK(ctx);
5285 
5286 	return bus_generic_suspend(dev);
5287 }
5288 int
5289 iflib_device_shutdown(device_t dev)
5290 {
5291 	if_ctx_t ctx = device_get_softc(dev);
5292 
5293 	CTX_LOCK(ctx);
5294 	IFDI_SHUTDOWN(ctx);
5295 	CTX_UNLOCK(ctx);
5296 
5297 	return bus_generic_suspend(dev);
5298 }
5299 
5300 int
5301 iflib_device_resume(device_t dev)
5302 {
5303 	if_ctx_t ctx = device_get_softc(dev);
5304 	iflib_txq_t txq = ctx->ifc_txqs;
5305 
5306 	CTX_LOCK(ctx);
5307 	IFDI_RESUME(ctx);
5308 	iflib_if_init_locked(ctx);
5309 	CTX_UNLOCK(ctx);
5310 	for (int i = 0; i < NTXQSETS(ctx); i++, txq++)
5311 		iflib_txq_check_drain(txq, IFLIB_RESTART_BUDGET);
5312 
5313 	return (bus_generic_resume(dev));
5314 }
5315 
5316 int
5317 iflib_device_iov_init(device_t dev, uint16_t num_vfs, const nvlist_t *params)
5318 {
5319 	int error;
5320 	if_ctx_t ctx = device_get_softc(dev);
5321 
5322 	CTX_LOCK(ctx);
5323 	error = IFDI_IOV_INIT(ctx, num_vfs, params);
5324 	CTX_UNLOCK(ctx);
5325 
5326 	return (error);
5327 }
5328 
5329 void
5330 iflib_device_iov_uninit(device_t dev)
5331 {
5332 	if_ctx_t ctx = device_get_softc(dev);
5333 
5334 	CTX_LOCK(ctx);
5335 	IFDI_IOV_UNINIT(ctx);
5336 	CTX_UNLOCK(ctx);
5337 }
5338 
5339 int
5340 iflib_device_iov_add_vf(device_t dev, uint16_t vfnum, const nvlist_t *params)
5341 {
5342 	int error;
5343 	if_ctx_t ctx = device_get_softc(dev);
5344 
5345 	CTX_LOCK(ctx);
5346 	error = IFDI_IOV_VF_ADD(ctx, vfnum, params);
5347 	CTX_UNLOCK(ctx);
5348 
5349 	return (error);
5350 }
5351 
5352 /*********************************************************************
5353  *
5354  *  MODULE FUNCTION DEFINITIONS
5355  *
5356  **********************************************************************/
5357 
5358 /*
5359  * - Start a fast taskqueue thread for each core
5360  * - Start a taskqueue for control operations
5361  */
5362 static int
5363 iflib_module_init(void)
5364 {
5365 	iflib_timer_default = hz / 2;
5366 	return (0);
5367 }
5368 
5369 static int
5370 iflib_module_event_handler(module_t mod, int what, void *arg)
5371 {
5372 	int err;
5373 
5374 	switch (what) {
5375 	case MOD_LOAD:
5376 		if ((err = iflib_module_init()) != 0)
5377 			return (err);
5378 		break;
5379 	case MOD_UNLOAD:
5380 		return (EBUSY);
5381 	default:
5382 		return (EOPNOTSUPP);
5383 	}
5384 
5385 	return (0);
5386 }
5387 
5388 /*********************************************************************
5389  *
5390  *  PUBLIC FUNCTION DEFINITIONS
5391  *     ordered as in iflib.h
5392  *
5393  **********************************************************************/
5394 
5395 static void
5396 _iflib_assert(if_shared_ctx_t sctx)
5397 {
5398 	int i;
5399 
5400 	MPASS(sctx->isc_tx_maxsize);
5401 	MPASS(sctx->isc_tx_maxsegsize);
5402 
5403 	MPASS(sctx->isc_rx_maxsize);
5404 	MPASS(sctx->isc_rx_nsegments);
5405 	MPASS(sctx->isc_rx_maxsegsize);
5406 
5407 	MPASS(sctx->isc_nrxqs >= 1 && sctx->isc_nrxqs <= 8);
5408 	for (i = 0; i < sctx->isc_nrxqs; i++) {
5409 		MPASS(sctx->isc_nrxd_min[i]);
5410 		MPASS(powerof2(sctx->isc_nrxd_min[i]));
5411 		MPASS(sctx->isc_nrxd_max[i]);
5412 		MPASS(powerof2(sctx->isc_nrxd_max[i]));
5413 		MPASS(sctx->isc_nrxd_default[i]);
5414 		MPASS(powerof2(sctx->isc_nrxd_default[i]));
5415 	}
5416 
5417 	MPASS(sctx->isc_ntxqs >= 1 && sctx->isc_ntxqs <= 8);
5418 	for (i = 0; i < sctx->isc_ntxqs; i++) {
5419 		MPASS(sctx->isc_ntxd_min[i]);
5420 		MPASS(powerof2(sctx->isc_ntxd_min[i]));
5421 		MPASS(sctx->isc_ntxd_max[i]);
5422 		MPASS(powerof2(sctx->isc_ntxd_max[i]));
5423 		MPASS(sctx->isc_ntxd_default[i]);
5424 		MPASS(powerof2(sctx->isc_ntxd_default[i]));
5425 	}
5426 }
5427 
5428 static void
5429 _iflib_pre_assert(if_softc_ctx_t scctx)
5430 {
5431 
5432 	MPASS(scctx->isc_txrx->ift_txd_encap);
5433 	MPASS(scctx->isc_txrx->ift_txd_flush);
5434 	MPASS(scctx->isc_txrx->ift_txd_credits_update);
5435 	MPASS(scctx->isc_txrx->ift_rxd_available);
5436 	MPASS(scctx->isc_txrx->ift_rxd_pkt_get);
5437 	MPASS(scctx->isc_txrx->ift_rxd_refill);
5438 	MPASS(scctx->isc_txrx->ift_rxd_flush);
5439 }
5440 
5441 static int
5442 iflib_register(if_ctx_t ctx)
5443 {
5444 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5445 	driver_t *driver = sctx->isc_driver;
5446 	device_t dev = ctx->ifc_dev;
5447 	if_t ifp;
5448 	u_char type;
5449 	int iflags;
5450 
5451 	if ((sctx->isc_flags & IFLIB_PSEUDO) == 0)
5452 		_iflib_assert(sctx);
5453 
5454 	CTX_LOCK_INIT(ctx);
5455 	STATE_LOCK_INIT(ctx, device_get_nameunit(ctx->ifc_dev));
5456 	if (sctx->isc_flags & IFLIB_PSEUDO) {
5457 		if (sctx->isc_flags & IFLIB_PSEUDO_ETHER)
5458 			type = IFT_ETHER;
5459 		else
5460 			type = IFT_PPP;
5461 	} else
5462 		type = IFT_ETHER;
5463 	ifp = ctx->ifc_ifp = if_alloc(type);
5464 	if (ifp == NULL) {
5465 		device_printf(dev, "can not allocate ifnet structure\n");
5466 		return (ENOMEM);
5467 	}
5468 
5469 	/*
5470 	 * Initialize our context's device specific methods
5471 	 */
5472 	kobj_init((kobj_t) ctx, (kobj_class_t) driver);
5473 	kobj_class_compile((kobj_class_t) driver);
5474 
5475 	if_initname(ifp, device_get_name(dev), device_get_unit(dev));
5476 	if_setsoftc(ifp, ctx);
5477 	if_setdev(ifp, dev);
5478 	if_setinitfn(ifp, iflib_if_init);
5479 	if_setioctlfn(ifp, iflib_if_ioctl);
5480 #ifdef ALTQ
5481 	if_setstartfn(ifp, iflib_altq_if_start);
5482 	if_settransmitfn(ifp, iflib_altq_if_transmit);
5483 	if_setsendqready(ifp);
5484 #else
5485 	if_settransmitfn(ifp, iflib_if_transmit);
5486 #endif
5487 	if_setqflushfn(ifp, iflib_if_qflush);
5488 	iflags = IFF_MULTICAST | IFF_KNOWSEPOCH;
5489 
5490 	if ((sctx->isc_flags & IFLIB_PSEUDO) &&
5491 		(sctx->isc_flags & IFLIB_PSEUDO_ETHER) == 0)
5492 		iflags |= IFF_POINTOPOINT;
5493 	else
5494 		iflags |= IFF_BROADCAST | IFF_SIMPLEX;
5495 	if_setflags(ifp, iflags);
5496 	ctx->ifc_vlan_attach_event =
5497 		EVENTHANDLER_REGISTER(vlan_config, iflib_vlan_register, ctx,
5498 							  EVENTHANDLER_PRI_FIRST);
5499 	ctx->ifc_vlan_detach_event =
5500 		EVENTHANDLER_REGISTER(vlan_unconfig, iflib_vlan_unregister, ctx,
5501 							  EVENTHANDLER_PRI_FIRST);
5502 
5503 	if ((sctx->isc_flags & IFLIB_DRIVER_MEDIA) == 0) {
5504 		ctx->ifc_mediap = &ctx->ifc_media;
5505 		ifmedia_init(ctx->ifc_mediap, IFM_IMASK,
5506 		    iflib_media_change, iflib_media_status);
5507 	}
5508 	return (0);
5509 }
5510 
5511 static void
5512 iflib_unregister_vlan_handlers(if_ctx_t ctx)
5513 {
5514 	/* Unregister VLAN events */
5515 	if (ctx->ifc_vlan_attach_event != NULL) {
5516 		EVENTHANDLER_DEREGISTER(vlan_config, ctx->ifc_vlan_attach_event);
5517 		ctx->ifc_vlan_attach_event = NULL;
5518 	}
5519 	if (ctx->ifc_vlan_detach_event != NULL) {
5520 		EVENTHANDLER_DEREGISTER(vlan_unconfig, ctx->ifc_vlan_detach_event);
5521 		ctx->ifc_vlan_detach_event = NULL;
5522 	}
5523 
5524 }
5525 
5526 static void
5527 iflib_deregister(if_ctx_t ctx)
5528 {
5529 	if_t ifp = ctx->ifc_ifp;
5530 
5531 	/* Remove all media */
5532 	ifmedia_removeall(&ctx->ifc_media);
5533 
5534 	/* Ensure that VLAN event handlers are unregistered */
5535 	iflib_unregister_vlan_handlers(ctx);
5536 
5537 	/* Release kobject reference */
5538 	kobj_delete((kobj_t) ctx, NULL);
5539 
5540 	/* Free the ifnet structure */
5541 	if_free(ifp);
5542 
5543 	STATE_LOCK_DESTROY(ctx);
5544 
5545 	/* ether_ifdetach calls if_qflush - lock must be destroy afterwards*/
5546 	CTX_LOCK_DESTROY(ctx);
5547 }
5548 
5549 static int
5550 iflib_queues_alloc(if_ctx_t ctx)
5551 {
5552 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5553 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
5554 	device_t dev = ctx->ifc_dev;
5555 	int nrxqsets = scctx->isc_nrxqsets;
5556 	int ntxqsets = scctx->isc_ntxqsets;
5557 	iflib_txq_t txq;
5558 	iflib_rxq_t rxq;
5559 	iflib_fl_t fl = NULL;
5560 	int i, j, cpu, err, txconf, rxconf;
5561 	iflib_dma_info_t ifdip;
5562 	uint32_t *rxqsizes = scctx->isc_rxqsizes;
5563 	uint32_t *txqsizes = scctx->isc_txqsizes;
5564 	uint8_t nrxqs = sctx->isc_nrxqs;
5565 	uint8_t ntxqs = sctx->isc_ntxqs;
5566 	int nfree_lists = sctx->isc_nfl ? sctx->isc_nfl : 1;
5567 	int fl_offset = (sctx->isc_flags & IFLIB_HAS_RXCQ ? 1 : 0);
5568 	caddr_t *vaddrs;
5569 	uint64_t *paddrs;
5570 
5571 	KASSERT(ntxqs > 0, ("number of queues per qset must be at least 1"));
5572 	KASSERT(nrxqs > 0, ("number of queues per qset must be at least 1"));
5573 	KASSERT(nrxqs >= fl_offset + nfree_lists,
5574            ("there must be at least a rxq for each free list"));
5575 
5576 	/* Allocate the TX ring struct memory */
5577 	if (!(ctx->ifc_txqs =
5578 	    (iflib_txq_t) malloc(sizeof(struct iflib_txq) *
5579 	    ntxqsets, M_IFLIB, M_NOWAIT | M_ZERO))) {
5580 		device_printf(dev, "Unable to allocate TX ring memory\n");
5581 		err = ENOMEM;
5582 		goto fail;
5583 	}
5584 
5585 	/* Now allocate the RX */
5586 	if (!(ctx->ifc_rxqs =
5587 	    (iflib_rxq_t) malloc(sizeof(struct iflib_rxq) *
5588 	    nrxqsets, M_IFLIB, M_NOWAIT | M_ZERO))) {
5589 		device_printf(dev, "Unable to allocate RX ring memory\n");
5590 		err = ENOMEM;
5591 		goto rx_fail;
5592 	}
5593 
5594 	txq = ctx->ifc_txqs;
5595 	rxq = ctx->ifc_rxqs;
5596 
5597 	/*
5598 	 * XXX handle allocation failure
5599 	 */
5600 	for (txconf = i = 0, cpu = CPU_FIRST(); i < ntxqsets; i++, txconf++, txq++, cpu = CPU_NEXT(cpu)) {
5601 		/* Set up some basics */
5602 
5603 		if ((ifdip = malloc(sizeof(struct iflib_dma_info) * ntxqs,
5604 		    M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
5605 			device_printf(dev,
5606 			    "Unable to allocate TX DMA info memory\n");
5607 			err = ENOMEM;
5608 			goto err_tx_desc;
5609 		}
5610 		txq->ift_ifdi = ifdip;
5611 		for (j = 0; j < ntxqs; j++, ifdip++) {
5612 			if (iflib_dma_alloc(ctx, txqsizes[j], ifdip, 0)) {
5613 				device_printf(dev,
5614 				    "Unable to allocate TX descriptors\n");
5615 				err = ENOMEM;
5616 				goto err_tx_desc;
5617 			}
5618 			txq->ift_txd_size[j] = scctx->isc_txd_size[j];
5619 			bzero((void *)ifdip->idi_vaddr, txqsizes[j]);
5620 		}
5621 		txq->ift_ctx = ctx;
5622 		txq->ift_id = i;
5623 		if (sctx->isc_flags & IFLIB_HAS_TXCQ) {
5624 			txq->ift_br_offset = 1;
5625 		} else {
5626 			txq->ift_br_offset = 0;
5627 		}
5628 
5629 		if (iflib_txsd_alloc(txq)) {
5630 			device_printf(dev, "Critical Failure setting up TX buffers\n");
5631 			err = ENOMEM;
5632 			goto err_tx_desc;
5633 		}
5634 
5635 		/* Initialize the TX lock */
5636 		snprintf(txq->ift_mtx_name, MTX_NAME_LEN, "%s:TX(%d):callout",
5637 		    device_get_nameunit(dev), txq->ift_id);
5638 		mtx_init(&txq->ift_mtx, txq->ift_mtx_name, NULL, MTX_DEF);
5639 		callout_init_mtx(&txq->ift_timer, &txq->ift_mtx, 0);
5640 		txq->ift_timer.c_cpu = cpu;
5641 #ifdef DEV_NETMAP
5642 		callout_init_mtx(&txq->ift_netmap_timer, &txq->ift_mtx, 0);
5643 		txq->ift_netmap_timer.c_cpu = cpu;
5644 #endif /* DEV_NETMAP */
5645 
5646 		err = ifmp_ring_alloc(&txq->ift_br, 2048, txq, iflib_txq_drain,
5647 				      iflib_txq_can_drain, M_IFLIB, M_WAITOK);
5648 		if (err) {
5649 			/* XXX free any allocated rings */
5650 			device_printf(dev, "Unable to allocate buf_ring\n");
5651 			goto err_tx_desc;
5652 		}
5653 	}
5654 
5655 	for (rxconf = i = 0; i < nrxqsets; i++, rxconf++, rxq++) {
5656 		/* Set up some basics */
5657 		callout_init(&rxq->ifr_watchdog, 1);
5658 
5659 		if ((ifdip = malloc(sizeof(struct iflib_dma_info) * nrxqs,
5660 		   M_IFLIB, M_NOWAIT | M_ZERO)) == NULL) {
5661 			device_printf(dev,
5662 			    "Unable to allocate RX DMA info memory\n");
5663 			err = ENOMEM;
5664 			goto err_tx_desc;
5665 		}
5666 
5667 		rxq->ifr_ifdi = ifdip;
5668 		/* XXX this needs to be changed if #rx queues != #tx queues */
5669 		rxq->ifr_ntxqirq = 1;
5670 		rxq->ifr_txqid[0] = i;
5671 		for (j = 0; j < nrxqs; j++, ifdip++) {
5672 			if (iflib_dma_alloc(ctx, rxqsizes[j], ifdip, 0)) {
5673 				device_printf(dev,
5674 				    "Unable to allocate RX descriptors\n");
5675 				err = ENOMEM;
5676 				goto err_tx_desc;
5677 			}
5678 			bzero((void *)ifdip->idi_vaddr, rxqsizes[j]);
5679 		}
5680 		rxq->ifr_ctx = ctx;
5681 		rxq->ifr_id = i;
5682 		rxq->ifr_fl_offset = fl_offset;
5683 		rxq->ifr_nfl = nfree_lists;
5684 		if (!(fl =
5685 			  (iflib_fl_t) malloc(sizeof(struct iflib_fl) * nfree_lists, M_IFLIB, M_NOWAIT | M_ZERO))) {
5686 			device_printf(dev, "Unable to allocate free list memory\n");
5687 			err = ENOMEM;
5688 			goto err_tx_desc;
5689 		}
5690 		rxq->ifr_fl = fl;
5691 		for (j = 0; j < nfree_lists; j++) {
5692 			fl[j].ifl_rxq = rxq;
5693 			fl[j].ifl_id = j;
5694 			fl[j].ifl_ifdi = &rxq->ifr_ifdi[j + rxq->ifr_fl_offset];
5695 			fl[j].ifl_rxd_size = scctx->isc_rxd_size[j];
5696 		}
5697 		/* Allocate receive buffers for the ring */
5698 		if (iflib_rxsd_alloc(rxq)) {
5699 			device_printf(dev,
5700 			    "Critical Failure setting up receive buffers\n");
5701 			err = ENOMEM;
5702 			goto err_rx_desc;
5703 		}
5704 
5705 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++)
5706 			fl->ifl_rx_bitmap = bit_alloc(fl->ifl_size, M_IFLIB,
5707 			    M_WAITOK);
5708 	}
5709 
5710 	/* TXQs */
5711 	vaddrs = malloc(sizeof(caddr_t)*ntxqsets*ntxqs, M_IFLIB, M_WAITOK);
5712 	paddrs = malloc(sizeof(uint64_t)*ntxqsets*ntxqs, M_IFLIB, M_WAITOK);
5713 	for (i = 0; i < ntxqsets; i++) {
5714 		iflib_dma_info_t di = ctx->ifc_txqs[i].ift_ifdi;
5715 
5716 		for (j = 0; j < ntxqs; j++, di++) {
5717 			vaddrs[i*ntxqs + j] = di->idi_vaddr;
5718 			paddrs[i*ntxqs + j] = di->idi_paddr;
5719 		}
5720 	}
5721 	if ((err = IFDI_TX_QUEUES_ALLOC(ctx, vaddrs, paddrs, ntxqs, ntxqsets)) != 0) {
5722 		device_printf(ctx->ifc_dev,
5723 		    "Unable to allocate device TX queue\n");
5724 		iflib_tx_structures_free(ctx);
5725 		free(vaddrs, M_IFLIB);
5726 		free(paddrs, M_IFLIB);
5727 		goto err_rx_desc;
5728 	}
5729 	free(vaddrs, M_IFLIB);
5730 	free(paddrs, M_IFLIB);
5731 
5732 	/* RXQs */
5733 	vaddrs = malloc(sizeof(caddr_t)*nrxqsets*nrxqs, M_IFLIB, M_WAITOK);
5734 	paddrs = malloc(sizeof(uint64_t)*nrxqsets*nrxqs, M_IFLIB, M_WAITOK);
5735 	for (i = 0; i < nrxqsets; i++) {
5736 		iflib_dma_info_t di = ctx->ifc_rxqs[i].ifr_ifdi;
5737 
5738 		for (j = 0; j < nrxqs; j++, di++) {
5739 			vaddrs[i*nrxqs + j] = di->idi_vaddr;
5740 			paddrs[i*nrxqs + j] = di->idi_paddr;
5741 		}
5742 	}
5743 	if ((err = IFDI_RX_QUEUES_ALLOC(ctx, vaddrs, paddrs, nrxqs, nrxqsets)) != 0) {
5744 		device_printf(ctx->ifc_dev,
5745 		    "Unable to allocate device RX queue\n");
5746 		iflib_tx_structures_free(ctx);
5747 		free(vaddrs, M_IFLIB);
5748 		free(paddrs, M_IFLIB);
5749 		goto err_rx_desc;
5750 	}
5751 	free(vaddrs, M_IFLIB);
5752 	free(paddrs, M_IFLIB);
5753 
5754 	return (0);
5755 
5756 /* XXX handle allocation failure changes */
5757 err_rx_desc:
5758 err_tx_desc:
5759 rx_fail:
5760 	if (ctx->ifc_rxqs != NULL)
5761 		free(ctx->ifc_rxqs, M_IFLIB);
5762 	ctx->ifc_rxqs = NULL;
5763 	if (ctx->ifc_txqs != NULL)
5764 		free(ctx->ifc_txqs, M_IFLIB);
5765 	ctx->ifc_txqs = NULL;
5766 fail:
5767 	return (err);
5768 }
5769 
5770 static int
5771 iflib_tx_structures_setup(if_ctx_t ctx)
5772 {
5773 	iflib_txq_t txq = ctx->ifc_txqs;
5774 	int i;
5775 
5776 	for (i = 0; i < NTXQSETS(ctx); i++, txq++)
5777 		iflib_txq_setup(txq);
5778 
5779 	return (0);
5780 }
5781 
5782 static void
5783 iflib_tx_structures_free(if_ctx_t ctx)
5784 {
5785 	iflib_txq_t txq = ctx->ifc_txqs;
5786 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5787 	int i, j;
5788 
5789 	for (i = 0; i < NTXQSETS(ctx); i++, txq++) {
5790 		for (j = 0; j < sctx->isc_ntxqs; j++)
5791 			iflib_dma_free(&txq->ift_ifdi[j]);
5792 		iflib_txq_destroy(txq);
5793 	}
5794 	free(ctx->ifc_txqs, M_IFLIB);
5795 	ctx->ifc_txqs = NULL;
5796 	IFDI_QUEUES_FREE(ctx);
5797 }
5798 
5799 /*********************************************************************
5800  *
5801  *  Initialize all receive rings.
5802  *
5803  **********************************************************************/
5804 static int
5805 iflib_rx_structures_setup(if_ctx_t ctx)
5806 {
5807 	iflib_rxq_t rxq = ctx->ifc_rxqs;
5808 	int q;
5809 #if defined(INET6) || defined(INET)
5810 	int err, i;
5811 #endif
5812 
5813 	for (q = 0; q < ctx->ifc_softc_ctx.isc_nrxqsets; q++, rxq++) {
5814 #if defined(INET6) || defined(INET)
5815 		if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_LRO) {
5816 			err = tcp_lro_init_args(&rxq->ifr_lc, ctx->ifc_ifp,
5817 			    TCP_LRO_ENTRIES, min(1024,
5818 			    ctx->ifc_softc_ctx.isc_nrxd[rxq->ifr_fl_offset]));
5819 			if (err != 0) {
5820 				device_printf(ctx->ifc_dev,
5821 				    "LRO Initialization failed!\n");
5822 				goto fail;
5823 			}
5824 		}
5825 #endif
5826 		IFDI_RXQ_SETUP(ctx, rxq->ifr_id);
5827 	}
5828 	return (0);
5829 #if defined(INET6) || defined(INET)
5830 fail:
5831 	/*
5832 	 * Free LRO resources allocated so far, we will only handle
5833 	 * the rings that completed, the failing case will have
5834 	 * cleaned up for itself.  'q' failed, so its the terminus.
5835 	 */
5836 	rxq = ctx->ifc_rxqs;
5837 	for (i = 0; i < q; ++i, rxq++) {
5838 		if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_LRO)
5839 			tcp_lro_free(&rxq->ifr_lc);
5840 	}
5841 	return (err);
5842 #endif
5843 }
5844 
5845 /*********************************************************************
5846  *
5847  *  Free all receive rings.
5848  *
5849  **********************************************************************/
5850 static void
5851 iflib_rx_structures_free(if_ctx_t ctx)
5852 {
5853 	iflib_rxq_t rxq = ctx->ifc_rxqs;
5854 	if_shared_ctx_t sctx = ctx->ifc_sctx;
5855 	int i, j;
5856 
5857 	for (i = 0; i < ctx->ifc_softc_ctx.isc_nrxqsets; i++, rxq++) {
5858 		for (j = 0; j < sctx->isc_nrxqs; j++)
5859 			iflib_dma_free(&rxq->ifr_ifdi[j]);
5860 		iflib_rx_sds_free(rxq);
5861 #if defined(INET6) || defined(INET)
5862 		if (if_getcapabilities(ctx->ifc_ifp) & IFCAP_LRO)
5863 			tcp_lro_free(&rxq->ifr_lc);
5864 #endif
5865 	}
5866 	free(ctx->ifc_rxqs, M_IFLIB);
5867 	ctx->ifc_rxqs = NULL;
5868 }
5869 
5870 static int
5871 iflib_qset_structures_setup(if_ctx_t ctx)
5872 {
5873 	int err;
5874 
5875 	/*
5876 	 * It is expected that the caller takes care of freeing queues if this
5877 	 * fails.
5878 	 */
5879 	if ((err = iflib_tx_structures_setup(ctx)) != 0) {
5880 		device_printf(ctx->ifc_dev, "iflib_tx_structures_setup failed: %d\n", err);
5881 		return (err);
5882 	}
5883 
5884 	if ((err = iflib_rx_structures_setup(ctx)) != 0)
5885 		device_printf(ctx->ifc_dev, "iflib_rx_structures_setup failed: %d\n", err);
5886 
5887 	return (err);
5888 }
5889 
5890 int
5891 iflib_irq_alloc(if_ctx_t ctx, if_irq_t irq, int rid,
5892 		driver_filter_t filter, void *filter_arg, driver_intr_t handler, void *arg, const char *name)
5893 {
5894 
5895 	return (_iflib_irq_alloc(ctx, irq, rid, filter, handler, arg, name));
5896 }
5897 
5898 #ifdef SMP
5899 static int
5900 find_nth(if_ctx_t ctx, int qid)
5901 {
5902 	cpuset_t cpus;
5903 	int i, cpuid, eqid, count;
5904 
5905 	CPU_COPY(&ctx->ifc_cpus, &cpus);
5906 	count = CPU_COUNT(&cpus);
5907 	eqid = qid % count;
5908 	/* clear up to the qid'th bit */
5909 	for (i = 0; i < eqid; i++) {
5910 		cpuid = CPU_FFS(&cpus);
5911 		MPASS(cpuid != 0);
5912 		CPU_CLR(cpuid-1, &cpus);
5913 	}
5914 	cpuid = CPU_FFS(&cpus);
5915 	MPASS(cpuid != 0);
5916 	return (cpuid-1);
5917 }
5918 
5919 #ifdef SCHED_ULE
5920 extern struct cpu_group *cpu_top;              /* CPU topology */
5921 
5922 static int
5923 find_child_with_core(int cpu, struct cpu_group *grp)
5924 {
5925 	int i;
5926 
5927 	if (grp->cg_children == 0)
5928 		return -1;
5929 
5930 	MPASS(grp->cg_child);
5931 	for (i = 0; i < grp->cg_children; i++) {
5932 		if (CPU_ISSET(cpu, &grp->cg_child[i].cg_mask))
5933 			return i;
5934 	}
5935 
5936 	return -1;
5937 }
5938 
5939 /*
5940  * Find the nth "close" core to the specified core
5941  * "close" is defined as the deepest level that shares
5942  * at least an L2 cache.  With threads, this will be
5943  * threads on the same core.  If the shared cache is L3
5944  * or higher, simply returns the same core.
5945  */
5946 static int
5947 find_close_core(int cpu, int core_offset)
5948 {
5949 	struct cpu_group *grp;
5950 	int i;
5951 	int fcpu;
5952 	cpuset_t cs;
5953 
5954 	grp = cpu_top;
5955 	if (grp == NULL)
5956 		return cpu;
5957 	i = 0;
5958 	while ((i = find_child_with_core(cpu, grp)) != -1) {
5959 		/* If the child only has one cpu, don't descend */
5960 		if (grp->cg_child[i].cg_count <= 1)
5961 			break;
5962 		grp = &grp->cg_child[i];
5963 	}
5964 
5965 	/* If they don't share at least an L2 cache, use the same CPU */
5966 	if (grp->cg_level > CG_SHARE_L2 || grp->cg_level == CG_SHARE_NONE)
5967 		return cpu;
5968 
5969 	/* Now pick one */
5970 	CPU_COPY(&grp->cg_mask, &cs);
5971 
5972 	/* Add the selected CPU offset to core offset. */
5973 	for (i = 0; (fcpu = CPU_FFS(&cs)) != 0; i++) {
5974 		if (fcpu - 1 == cpu)
5975 			break;
5976 		CPU_CLR(fcpu - 1, &cs);
5977 	}
5978 	MPASS(fcpu);
5979 
5980 	core_offset += i;
5981 
5982 	CPU_COPY(&grp->cg_mask, &cs);
5983 	for (i = core_offset % grp->cg_count; i > 0; i--) {
5984 		MPASS(CPU_FFS(&cs));
5985 		CPU_CLR(CPU_FFS(&cs) - 1, &cs);
5986 	}
5987 	MPASS(CPU_FFS(&cs));
5988 	return CPU_FFS(&cs) - 1;
5989 }
5990 #else
5991 static int
5992 find_close_core(int cpu, int core_offset __unused)
5993 {
5994 	return cpu;
5995 }
5996 #endif
5997 
5998 static int
5999 get_core_offset(if_ctx_t ctx, iflib_intr_type_t type, int qid)
6000 {
6001 	switch (type) {
6002 	case IFLIB_INTR_TX:
6003 		/* TX queues get cores which share at least an L2 cache with the corresponding RX queue */
6004 		/* XXX handle multiple RX threads per core and more than two core per L2 group */
6005 		return qid / CPU_COUNT(&ctx->ifc_cpus) + 1;
6006 	case IFLIB_INTR_RX:
6007 	case IFLIB_INTR_RXTX:
6008 		/* RX queues get the specified core */
6009 		return qid / CPU_COUNT(&ctx->ifc_cpus);
6010 	default:
6011 		return -1;
6012 	}
6013 }
6014 #else
6015 #define get_core_offset(ctx, type, qid)	CPU_FIRST()
6016 #define find_close_core(cpuid, tid)	CPU_FIRST()
6017 #define find_nth(ctx, gid)		CPU_FIRST()
6018 #endif
6019 
6020 /* Just to avoid copy/paste */
6021 static inline int
6022 iflib_irq_set_affinity(if_ctx_t ctx, if_irq_t irq, iflib_intr_type_t type,
6023     int qid, struct grouptask *gtask, struct taskqgroup *tqg, void *uniq,
6024     const char *name)
6025 {
6026 	device_t dev;
6027 	int co, cpuid, err, tid;
6028 
6029 	dev = ctx->ifc_dev;
6030 	co = ctx->ifc_sysctl_core_offset;
6031 	if (ctx->ifc_sysctl_separate_txrx && type == IFLIB_INTR_TX)
6032 		co += ctx->ifc_softc_ctx.isc_nrxqsets;
6033 	cpuid = find_nth(ctx, qid + co);
6034 	tid = get_core_offset(ctx, type, qid);
6035 	if (tid < 0) {
6036 		device_printf(dev, "get_core_offset failed\n");
6037 		return (EOPNOTSUPP);
6038 	}
6039 	cpuid = find_close_core(cpuid, tid);
6040 	err = taskqgroup_attach_cpu(tqg, gtask, uniq, cpuid, dev, irq->ii_res,
6041 	    name);
6042 	if (err) {
6043 		device_printf(dev, "taskqgroup_attach_cpu failed %d\n", err);
6044 		return (err);
6045 	}
6046 #ifdef notyet
6047 	if (cpuid > ctx->ifc_cpuid_highest)
6048 		ctx->ifc_cpuid_highest = cpuid;
6049 #endif
6050 	return (0);
6051 }
6052 
6053 int
6054 iflib_irq_alloc_generic(if_ctx_t ctx, if_irq_t irq, int rid,
6055 			iflib_intr_type_t type, driver_filter_t *filter,
6056 			void *filter_arg, int qid, const char *name)
6057 {
6058 	device_t dev;
6059 	struct grouptask *gtask;
6060 	struct taskqgroup *tqg;
6061 	iflib_filter_info_t info;
6062 	gtask_fn_t *fn;
6063 	int tqrid, err;
6064 	driver_filter_t *intr_fast;
6065 	void *q;
6066 
6067 	info = &ctx->ifc_filter_info;
6068 	tqrid = rid;
6069 
6070 	switch (type) {
6071 	/* XXX merge tx/rx for netmap? */
6072 	case IFLIB_INTR_TX:
6073 		q = &ctx->ifc_txqs[qid];
6074 		info = &ctx->ifc_txqs[qid].ift_filter_info;
6075 		gtask = &ctx->ifc_txqs[qid].ift_task;
6076 		tqg = qgroup_if_io_tqg;
6077 		fn = _task_fn_tx;
6078 		intr_fast = iflib_fast_intr;
6079 		GROUPTASK_INIT(gtask, 0, fn, q);
6080 		ctx->ifc_flags |= IFC_NETMAP_TX_IRQ;
6081 		break;
6082 	case IFLIB_INTR_RX:
6083 		q = &ctx->ifc_rxqs[qid];
6084 		info = &ctx->ifc_rxqs[qid].ifr_filter_info;
6085 		gtask = &ctx->ifc_rxqs[qid].ifr_task;
6086 		tqg = qgroup_if_io_tqg;
6087 		fn = _task_fn_rx;
6088 		intr_fast = iflib_fast_intr;
6089 		NET_GROUPTASK_INIT(gtask, 0, fn, q);
6090 		break;
6091 	case IFLIB_INTR_RXTX:
6092 		q = &ctx->ifc_rxqs[qid];
6093 		info = &ctx->ifc_rxqs[qid].ifr_filter_info;
6094 		gtask = &ctx->ifc_rxqs[qid].ifr_task;
6095 		tqg = qgroup_if_io_tqg;
6096 		fn = _task_fn_rx;
6097 		intr_fast = iflib_fast_intr_rxtx;
6098 		NET_GROUPTASK_INIT(gtask, 0, fn, q);
6099 		break;
6100 	case IFLIB_INTR_ADMIN:
6101 		q = ctx;
6102 		tqrid = -1;
6103 		info = &ctx->ifc_filter_info;
6104 		gtask = &ctx->ifc_admin_task;
6105 		tqg = qgroup_if_config_tqg;
6106 		fn = _task_fn_admin;
6107 		intr_fast = iflib_fast_intr_ctx;
6108 		break;
6109 	default:
6110 		device_printf(ctx->ifc_dev, "%s: unknown net intr type\n",
6111 		    __func__);
6112 		return (EINVAL);
6113 	}
6114 
6115 	info->ifi_filter = filter;
6116 	info->ifi_filter_arg = filter_arg;
6117 	info->ifi_task = gtask;
6118 	info->ifi_ctx = q;
6119 
6120 	dev = ctx->ifc_dev;
6121 	err = _iflib_irq_alloc(ctx, irq, rid, intr_fast, NULL, info,  name);
6122 	if (err != 0) {
6123 		device_printf(dev, "_iflib_irq_alloc failed %d\n", err);
6124 		return (err);
6125 	}
6126 	if (type == IFLIB_INTR_ADMIN)
6127 		return (0);
6128 
6129 	if (tqrid != -1) {
6130 		err = iflib_irq_set_affinity(ctx, irq, type, qid, gtask, tqg,
6131 		    q, name);
6132 		if (err)
6133 			return (err);
6134 	} else {
6135 		taskqgroup_attach(tqg, gtask, q, dev, irq->ii_res, name);
6136 	}
6137 
6138 	return (0);
6139 }
6140 
6141 void
6142 iflib_softirq_alloc_generic(if_ctx_t ctx, if_irq_t irq, iflib_intr_type_t type, void *arg, int qid, const char *name)
6143 {
6144 	struct grouptask *gtask;
6145 	struct taskqgroup *tqg;
6146 	gtask_fn_t *fn;
6147 	void *q;
6148 	int err;
6149 
6150 	switch (type) {
6151 	case IFLIB_INTR_TX:
6152 		q = &ctx->ifc_txqs[qid];
6153 		gtask = &ctx->ifc_txqs[qid].ift_task;
6154 		tqg = qgroup_if_io_tqg;
6155 		fn = _task_fn_tx;
6156 		GROUPTASK_INIT(gtask, 0, fn, q);
6157 		break;
6158 	case IFLIB_INTR_RX:
6159 		q = &ctx->ifc_rxqs[qid];
6160 		gtask = &ctx->ifc_rxqs[qid].ifr_task;
6161 		tqg = qgroup_if_io_tqg;
6162 		fn = _task_fn_rx;
6163 		NET_GROUPTASK_INIT(gtask, 0, fn, q);
6164 		break;
6165 	case IFLIB_INTR_IOV:
6166 		q = ctx;
6167 		gtask = &ctx->ifc_vflr_task;
6168 		tqg = qgroup_if_config_tqg;
6169 		fn = _task_fn_iov;
6170 		GROUPTASK_INIT(gtask, 0, fn, q);
6171 		break;
6172 	default:
6173 		panic("unknown net intr type");
6174 	}
6175 	if (irq != NULL) {
6176 		err = iflib_irq_set_affinity(ctx, irq, type, qid, gtask, tqg,
6177 		    q, name);
6178 		if (err)
6179 			taskqgroup_attach(tqg, gtask, q, ctx->ifc_dev,
6180 			    irq->ii_res, name);
6181 	} else {
6182 		taskqgroup_attach(tqg, gtask, q, NULL, NULL, name);
6183 	}
6184 }
6185 
6186 void
6187 iflib_irq_free(if_ctx_t ctx, if_irq_t irq)
6188 {
6189 
6190 	if (irq->ii_tag)
6191 		bus_teardown_intr(ctx->ifc_dev, irq->ii_res, irq->ii_tag);
6192 
6193 	if (irq->ii_res)
6194 		bus_release_resource(ctx->ifc_dev, SYS_RES_IRQ,
6195 		    rman_get_rid(irq->ii_res), irq->ii_res);
6196 }
6197 
6198 static int
6199 iflib_legacy_setup(if_ctx_t ctx, driver_filter_t filter, void *filter_arg, int *rid, const char *name)
6200 {
6201 	iflib_txq_t txq = ctx->ifc_txqs;
6202 	iflib_rxq_t rxq = ctx->ifc_rxqs;
6203 	if_irq_t irq = &ctx->ifc_legacy_irq;
6204 	iflib_filter_info_t info;
6205 	device_t dev;
6206 	struct grouptask *gtask;
6207 	struct resource *res;
6208 	struct taskqgroup *tqg;
6209 	void *q;
6210 	int err, tqrid;
6211 	bool rx_only;
6212 
6213 	q = &ctx->ifc_rxqs[0];
6214 	info = &rxq[0].ifr_filter_info;
6215 	gtask = &rxq[0].ifr_task;
6216 	tqg = qgroup_if_io_tqg;
6217 	tqrid = *rid;
6218 	rx_only = (ctx->ifc_sctx->isc_flags & IFLIB_SINGLE_IRQ_RX_ONLY) != 0;
6219 
6220 	ctx->ifc_flags |= IFC_LEGACY;
6221 	info->ifi_filter = filter;
6222 	info->ifi_filter_arg = filter_arg;
6223 	info->ifi_task = gtask;
6224 	info->ifi_ctx = rx_only ? ctx : q;
6225 
6226 	dev = ctx->ifc_dev;
6227 	/* We allocate a single interrupt resource */
6228 	err = _iflib_irq_alloc(ctx, irq, tqrid, rx_only ? iflib_fast_intr_ctx :
6229 	    iflib_fast_intr_rxtx, NULL, info, name);
6230 	if (err != 0)
6231 		return (err);
6232 	NET_GROUPTASK_INIT(gtask, 0, _task_fn_rx, q);
6233 	res = irq->ii_res;
6234 	taskqgroup_attach(tqg, gtask, q, dev, res, name);
6235 
6236 	GROUPTASK_INIT(&txq->ift_task, 0, _task_fn_tx, txq);
6237 	taskqgroup_attach(qgroup_if_io_tqg, &txq->ift_task, txq, dev, res,
6238 	    "tx");
6239 	return (0);
6240 }
6241 
6242 void
6243 iflib_led_create(if_ctx_t ctx)
6244 {
6245 
6246 	ctx->ifc_led_dev = led_create(iflib_led_func, ctx,
6247 	    device_get_nameunit(ctx->ifc_dev));
6248 }
6249 
6250 void
6251 iflib_tx_intr_deferred(if_ctx_t ctx, int txqid)
6252 {
6253 
6254 	GROUPTASK_ENQUEUE(&ctx->ifc_txqs[txqid].ift_task);
6255 }
6256 
6257 void
6258 iflib_rx_intr_deferred(if_ctx_t ctx, int rxqid)
6259 {
6260 
6261 	GROUPTASK_ENQUEUE(&ctx->ifc_rxqs[rxqid].ifr_task);
6262 }
6263 
6264 void
6265 iflib_admin_intr_deferred(if_ctx_t ctx)
6266 {
6267 
6268 	MPASS(ctx->ifc_admin_task.gt_taskqueue != NULL);
6269 	GROUPTASK_ENQUEUE(&ctx->ifc_admin_task);
6270 }
6271 
6272 void
6273 iflib_iov_intr_deferred(if_ctx_t ctx)
6274 {
6275 
6276 	GROUPTASK_ENQUEUE(&ctx->ifc_vflr_task);
6277 }
6278 
6279 void
6280 iflib_io_tqg_attach(struct grouptask *gt, void *uniq, int cpu, const char *name)
6281 {
6282 
6283 	taskqgroup_attach_cpu(qgroup_if_io_tqg, gt, uniq, cpu, NULL, NULL,
6284 	    name);
6285 }
6286 
6287 void
6288 iflib_config_gtask_init(void *ctx, struct grouptask *gtask, gtask_fn_t *fn,
6289 	const char *name)
6290 {
6291 
6292 	GROUPTASK_INIT(gtask, 0, fn, ctx);
6293 	taskqgroup_attach(qgroup_if_config_tqg, gtask, gtask, NULL, NULL,
6294 	    name);
6295 }
6296 
6297 void
6298 iflib_config_gtask_deinit(struct grouptask *gtask)
6299 {
6300 
6301 	taskqgroup_detach(qgroup_if_config_tqg, gtask);
6302 }
6303 
6304 void
6305 iflib_link_state_change(if_ctx_t ctx, int link_state, uint64_t baudrate)
6306 {
6307 	if_t ifp = ctx->ifc_ifp;
6308 	iflib_txq_t txq = ctx->ifc_txqs;
6309 
6310 	if_setbaudrate(ifp, baudrate);
6311 	if (baudrate >= IF_Gbps(10)) {
6312 		STATE_LOCK(ctx);
6313 		ctx->ifc_flags |= IFC_PREFETCH;
6314 		STATE_UNLOCK(ctx);
6315 	}
6316 	/* If link down, disable watchdog */
6317 	if ((ctx->ifc_link_state == LINK_STATE_UP) && (link_state == LINK_STATE_DOWN)) {
6318 		for (int i = 0; i < ctx->ifc_softc_ctx.isc_ntxqsets; i++, txq++)
6319 			txq->ift_qstatus = IFLIB_QUEUE_IDLE;
6320 	}
6321 	ctx->ifc_link_state = link_state;
6322 	if_link_state_change(ifp, link_state);
6323 }
6324 
6325 static int
6326 iflib_tx_credits_update(if_ctx_t ctx, iflib_txq_t txq)
6327 {
6328 	int credits;
6329 #ifdef INVARIANTS
6330 	int credits_pre = txq->ift_cidx_processed;
6331 #endif
6332 
6333 	bus_dmamap_sync(txq->ift_ifdi->idi_tag, txq->ift_ifdi->idi_map,
6334 	    BUS_DMASYNC_POSTREAD);
6335 	if ((credits = ctx->isc_txd_credits_update(ctx->ifc_softc, txq->ift_id, true)) == 0)
6336 		return (0);
6337 
6338 	txq->ift_processed += credits;
6339 	txq->ift_cidx_processed += credits;
6340 
6341 	MPASS(credits_pre + credits == txq->ift_cidx_processed);
6342 	if (txq->ift_cidx_processed >= txq->ift_size)
6343 		txq->ift_cidx_processed -= txq->ift_size;
6344 	return (credits);
6345 }
6346 
6347 static int
6348 iflib_rxd_avail(if_ctx_t ctx, iflib_rxq_t rxq, qidx_t cidx, qidx_t budget)
6349 {
6350 	iflib_fl_t fl;
6351 	u_int i;
6352 
6353 	for (i = 0, fl = &rxq->ifr_fl[0]; i < rxq->ifr_nfl; i++, fl++)
6354 		bus_dmamap_sync(fl->ifl_ifdi->idi_tag, fl->ifl_ifdi->idi_map,
6355 		    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
6356 	return (ctx->isc_rxd_available(ctx->ifc_softc, rxq->ifr_id, cidx,
6357 	    budget));
6358 }
6359 
6360 void
6361 iflib_add_int_delay_sysctl(if_ctx_t ctx, const char *name,
6362 	const char *description, if_int_delay_info_t info,
6363 	int offset, int value)
6364 {
6365 	info->iidi_ctx = ctx;
6366 	info->iidi_offset = offset;
6367 	info->iidi_value = value;
6368 	SYSCTL_ADD_PROC(device_get_sysctl_ctx(ctx->ifc_dev),
6369 	    SYSCTL_CHILDREN(device_get_sysctl_tree(ctx->ifc_dev)),
6370 	    OID_AUTO, name, CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_MPSAFE,
6371 	    info, 0, iflib_sysctl_int_delay, "I", description);
6372 }
6373 
6374 struct sx *
6375 iflib_ctx_lock_get(if_ctx_t ctx)
6376 {
6377 
6378 	return (&ctx->ifc_ctx_sx);
6379 }
6380 
6381 static int
6382 iflib_msix_init(if_ctx_t ctx)
6383 {
6384 	device_t dev = ctx->ifc_dev;
6385 	if_shared_ctx_t sctx = ctx->ifc_sctx;
6386 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
6387 	int admincnt, bar, err, iflib_num_rx_queues, iflib_num_tx_queues;
6388 	int msgs, queuemsgs, queues, rx_queues, tx_queues, vectors;
6389 
6390 	iflib_num_tx_queues = ctx->ifc_sysctl_ntxqs;
6391 	iflib_num_rx_queues = ctx->ifc_sysctl_nrxqs;
6392 
6393 	if (bootverbose)
6394 		device_printf(dev, "msix_init qsets capped at %d\n",
6395 		    imax(scctx->isc_ntxqsets, scctx->isc_nrxqsets));
6396 
6397 	/* Override by tuneable */
6398 	if (scctx->isc_disable_msix)
6399 		goto msi;
6400 
6401 	/* First try MSI-X */
6402 	if ((msgs = pci_msix_count(dev)) == 0) {
6403 		if (bootverbose)
6404 			device_printf(dev, "MSI-X not supported or disabled\n");
6405 		goto msi;
6406 	}
6407 
6408 	bar = ctx->ifc_softc_ctx.isc_msix_bar;
6409 	/*
6410 	 * bar == -1 => "trust me I know what I'm doing"
6411 	 * Some drivers are for hardware that is so shoddily
6412 	 * documented that no one knows which bars are which
6413 	 * so the developer has to map all bars. This hack
6414 	 * allows shoddy garbage to use MSI-X in this framework.
6415 	 */
6416 	if (bar != -1) {
6417 		ctx->ifc_msix_mem = bus_alloc_resource_any(dev,
6418 	            SYS_RES_MEMORY, &bar, RF_ACTIVE);
6419 		if (ctx->ifc_msix_mem == NULL) {
6420 			device_printf(dev, "Unable to map MSI-X table\n");
6421 			goto msi;
6422 		}
6423 	}
6424 
6425 	admincnt = sctx->isc_admin_intrcnt;
6426 #if IFLIB_DEBUG
6427 	/* use only 1 qset in debug mode */
6428 	queuemsgs = min(msgs - admincnt, 1);
6429 #else
6430 	queuemsgs = msgs - admincnt;
6431 #endif
6432 #ifdef RSS
6433 	queues = imin(queuemsgs, rss_getnumbuckets());
6434 #else
6435 	queues = queuemsgs;
6436 #endif
6437 	queues = imin(CPU_COUNT(&ctx->ifc_cpus), queues);
6438 	if (bootverbose)
6439 		device_printf(dev,
6440 		    "intr CPUs: %d queue msgs: %d admincnt: %d\n",
6441 		    CPU_COUNT(&ctx->ifc_cpus), queuemsgs, admincnt);
6442 #ifdef  RSS
6443 	/* If we're doing RSS, clamp at the number of RSS buckets */
6444 	if (queues > rss_getnumbuckets())
6445 		queues = rss_getnumbuckets();
6446 #endif
6447 	if (iflib_num_rx_queues > 0 && iflib_num_rx_queues < queuemsgs - admincnt)
6448 		rx_queues = iflib_num_rx_queues;
6449 	else
6450 		rx_queues = queues;
6451 
6452 	if (rx_queues > scctx->isc_nrxqsets)
6453 		rx_queues = scctx->isc_nrxqsets;
6454 
6455 	/*
6456 	 * We want this to be all logical CPUs by default
6457 	 */
6458 	if (iflib_num_tx_queues > 0 && iflib_num_tx_queues < queues)
6459 		tx_queues = iflib_num_tx_queues;
6460 	else
6461 		tx_queues = mp_ncpus;
6462 
6463 	if (tx_queues > scctx->isc_ntxqsets)
6464 		tx_queues = scctx->isc_ntxqsets;
6465 
6466 	if (ctx->ifc_sysctl_qs_eq_override == 0) {
6467 #ifdef INVARIANTS
6468 		if (tx_queues != rx_queues)
6469 			device_printf(dev,
6470 			    "queue equality override not set, capping rx_queues at %d and tx_queues at %d\n",
6471 			    min(rx_queues, tx_queues), min(rx_queues, tx_queues));
6472 #endif
6473 		tx_queues = min(rx_queues, tx_queues);
6474 		rx_queues = min(rx_queues, tx_queues);
6475 	}
6476 
6477 	vectors = rx_queues + admincnt;
6478 	if (msgs < vectors) {
6479 		device_printf(dev,
6480 		    "insufficient number of MSI-X vectors "
6481 		    "(supported %d, need %d)\n", msgs, vectors);
6482 		goto msi;
6483 	}
6484 
6485 	device_printf(dev, "Using %d RX queues %d TX queues\n", rx_queues,
6486 	    tx_queues);
6487 	msgs = vectors;
6488 	if ((err = pci_alloc_msix(dev, &vectors)) == 0) {
6489 		if (vectors != msgs) {
6490 			device_printf(dev,
6491 			    "Unable to allocate sufficient MSI-X vectors "
6492 			    "(got %d, need %d)\n", vectors, msgs);
6493 			pci_release_msi(dev);
6494 			if (bar != -1) {
6495 				bus_release_resource(dev, SYS_RES_MEMORY, bar,
6496 				    ctx->ifc_msix_mem);
6497 				ctx->ifc_msix_mem = NULL;
6498 			}
6499 			goto msi;
6500 		}
6501 		device_printf(dev, "Using MSI-X interrupts with %d vectors\n",
6502 		    vectors);
6503 		scctx->isc_vectors = vectors;
6504 		scctx->isc_nrxqsets = rx_queues;
6505 		scctx->isc_ntxqsets = tx_queues;
6506 		scctx->isc_intr = IFLIB_INTR_MSIX;
6507 
6508 		return (vectors);
6509 	} else {
6510 		device_printf(dev,
6511 		    "failed to allocate %d MSI-X vectors, err: %d\n", vectors,
6512 		    err);
6513 		if (bar != -1) {
6514 			bus_release_resource(dev, SYS_RES_MEMORY, bar,
6515 			    ctx->ifc_msix_mem);
6516 			ctx->ifc_msix_mem = NULL;
6517 		}
6518 	}
6519 
6520 msi:
6521 	vectors = pci_msi_count(dev);
6522 	scctx->isc_nrxqsets = 1;
6523 	scctx->isc_ntxqsets = 1;
6524 	scctx->isc_vectors = vectors;
6525 	if (vectors == 1 && pci_alloc_msi(dev, &vectors) == 0) {
6526 		device_printf(dev,"Using an MSI interrupt\n");
6527 		scctx->isc_intr = IFLIB_INTR_MSI;
6528 	} else {
6529 		scctx->isc_vectors = 1;
6530 		device_printf(dev,"Using a Legacy interrupt\n");
6531 		scctx->isc_intr = IFLIB_INTR_LEGACY;
6532 	}
6533 
6534 	return (vectors);
6535 }
6536 
6537 static const char *ring_states[] = { "IDLE", "BUSY", "STALLED", "ABDICATED" };
6538 
6539 static int
6540 mp_ring_state_handler(SYSCTL_HANDLER_ARGS)
6541 {
6542 	int rc;
6543 	uint16_t *state = ((uint16_t *)oidp->oid_arg1);
6544 	struct sbuf *sb;
6545 	const char *ring_state = "UNKNOWN";
6546 
6547 	/* XXX needed ? */
6548 	rc = sysctl_wire_old_buffer(req, 0);
6549 	MPASS(rc == 0);
6550 	if (rc != 0)
6551 		return (rc);
6552 	sb = sbuf_new_for_sysctl(NULL, NULL, 80, req);
6553 	MPASS(sb != NULL);
6554 	if (sb == NULL)
6555 		return (ENOMEM);
6556 	if (state[3] <= 3)
6557 		ring_state = ring_states[state[3]];
6558 
6559 	sbuf_printf(sb, "pidx_head: %04hd pidx_tail: %04hd cidx: %04hd state: %s",
6560 		    state[0], state[1], state[2], ring_state);
6561 	rc = sbuf_finish(sb);
6562 	sbuf_delete(sb);
6563         return(rc);
6564 }
6565 
6566 enum iflib_ndesc_handler {
6567 	IFLIB_NTXD_HANDLER,
6568 	IFLIB_NRXD_HANDLER,
6569 };
6570 
6571 static int
6572 mp_ndesc_handler(SYSCTL_HANDLER_ARGS)
6573 {
6574 	if_ctx_t ctx = (void *)arg1;
6575 	enum iflib_ndesc_handler type = arg2;
6576 	char buf[256] = {0};
6577 	qidx_t *ndesc;
6578 	char *p, *next;
6579 	int nqs, rc, i;
6580 
6581 	nqs = 8;
6582 	switch(type) {
6583 	case IFLIB_NTXD_HANDLER:
6584 		ndesc = ctx->ifc_sysctl_ntxds;
6585 		if (ctx->ifc_sctx)
6586 			nqs = ctx->ifc_sctx->isc_ntxqs;
6587 		break;
6588 	case IFLIB_NRXD_HANDLER:
6589 		ndesc = ctx->ifc_sysctl_nrxds;
6590 		if (ctx->ifc_sctx)
6591 			nqs = ctx->ifc_sctx->isc_nrxqs;
6592 		break;
6593 	default:
6594 		printf("%s: unhandled type\n", __func__);
6595 		return (EINVAL);
6596 	}
6597 	if (nqs == 0)
6598 		nqs = 8;
6599 
6600 	for (i=0; i<8; i++) {
6601 		if (i >= nqs)
6602 			break;
6603 		if (i)
6604 			strcat(buf, ",");
6605 		sprintf(strchr(buf, 0), "%d", ndesc[i]);
6606 	}
6607 
6608 	rc = sysctl_handle_string(oidp, buf, sizeof(buf), req);
6609 	if (rc || req->newptr == NULL)
6610 		return rc;
6611 
6612 	for (i = 0, next = buf, p = strsep(&next, " ,"); i < 8 && p;
6613 	    i++, p = strsep(&next, " ,")) {
6614 		ndesc[i] = strtoul(p, NULL, 10);
6615 	}
6616 
6617 	return(rc);
6618 }
6619 
6620 #define NAME_BUFLEN 32
6621 static void
6622 iflib_add_device_sysctl_pre(if_ctx_t ctx)
6623 {
6624         device_t dev = iflib_get_dev(ctx);
6625 	struct sysctl_oid_list *child, *oid_list;
6626 	struct sysctl_ctx_list *ctx_list;
6627 	struct sysctl_oid *node;
6628 
6629 	ctx_list = device_get_sysctl_ctx(dev);
6630 	child = SYSCTL_CHILDREN(device_get_sysctl_tree(dev));
6631 	ctx->ifc_sysctl_node = node = SYSCTL_ADD_NODE(ctx_list, child, OID_AUTO, "iflib",
6632 	    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "IFLIB fields");
6633 	oid_list = SYSCTL_CHILDREN(node);
6634 
6635 	SYSCTL_ADD_CONST_STRING(ctx_list, oid_list, OID_AUTO, "driver_version",
6636 		       CTLFLAG_RD, ctx->ifc_sctx->isc_driver_version,
6637 		       "driver version");
6638 
6639 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "override_ntxqs",
6640 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_ntxqs, 0,
6641 			"# of txqs to use, 0 => use default #");
6642 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "override_nrxqs",
6643 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_nrxqs, 0,
6644 			"# of rxqs to use, 0 => use default #");
6645 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "override_qs_enable",
6646 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_qs_eq_override, 0,
6647                        "permit #txq != #rxq");
6648 	SYSCTL_ADD_INT(ctx_list, oid_list, OID_AUTO, "disable_msix",
6649                       CTLFLAG_RWTUN, &ctx->ifc_softc_ctx.isc_disable_msix, 0,
6650                       "disable MSI-X (default 0)");
6651 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "rx_budget",
6652 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_rx_budget, 0,
6653 		       "set the RX budget");
6654 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "tx_abdicate",
6655 		       CTLFLAG_RWTUN, &ctx->ifc_sysctl_tx_abdicate, 0,
6656 		       "cause TX to abdicate instead of running to completion");
6657 	ctx->ifc_sysctl_core_offset = CORE_OFFSET_UNSPECIFIED;
6658 	SYSCTL_ADD_U16(ctx_list, oid_list, OID_AUTO, "core_offset",
6659 		       CTLFLAG_RDTUN, &ctx->ifc_sysctl_core_offset, 0,
6660 		       "offset to start using cores at");
6661 	SYSCTL_ADD_U8(ctx_list, oid_list, OID_AUTO, "separate_txrx",
6662 		       CTLFLAG_RDTUN, &ctx->ifc_sysctl_separate_txrx, 0,
6663 		       "use separate cores for TX and RX");
6664 
6665 	/* XXX change for per-queue sizes */
6666 	SYSCTL_ADD_PROC(ctx_list, oid_list, OID_AUTO, "override_ntxds",
6667 	    CTLTYPE_STRING | CTLFLAG_RWTUN | CTLFLAG_NEEDGIANT, ctx,
6668 	    IFLIB_NTXD_HANDLER, mp_ndesc_handler, "A",
6669 	    "list of # of TX descriptors to use, 0 = use default #");
6670 	SYSCTL_ADD_PROC(ctx_list, oid_list, OID_AUTO, "override_nrxds",
6671 	    CTLTYPE_STRING | CTLFLAG_RWTUN | CTLFLAG_NEEDGIANT, ctx,
6672 	    IFLIB_NRXD_HANDLER, mp_ndesc_handler, "A",
6673 	    "list of # of RX descriptors to use, 0 = use default #");
6674 }
6675 
6676 static void
6677 iflib_add_device_sysctl_post(if_ctx_t ctx)
6678 {
6679 	if_shared_ctx_t sctx = ctx->ifc_sctx;
6680 	if_softc_ctx_t scctx = &ctx->ifc_softc_ctx;
6681         device_t dev = iflib_get_dev(ctx);
6682 	struct sysctl_oid_list *child;
6683 	struct sysctl_ctx_list *ctx_list;
6684 	iflib_fl_t fl;
6685 	iflib_txq_t txq;
6686 	iflib_rxq_t rxq;
6687 	int i, j;
6688 	char namebuf[NAME_BUFLEN];
6689 	char *qfmt;
6690 	struct sysctl_oid *queue_node, *fl_node, *node;
6691 	struct sysctl_oid_list *queue_list, *fl_list;
6692 	ctx_list = device_get_sysctl_ctx(dev);
6693 
6694 	node = ctx->ifc_sysctl_node;
6695 	child = SYSCTL_CHILDREN(node);
6696 
6697 	if (scctx->isc_ntxqsets > 100)
6698 		qfmt = "txq%03d";
6699 	else if (scctx->isc_ntxqsets > 10)
6700 		qfmt = "txq%02d";
6701 	else
6702 		qfmt = "txq%d";
6703 	for (i = 0, txq = ctx->ifc_txqs; i < scctx->isc_ntxqsets; i++, txq++) {
6704 		snprintf(namebuf, NAME_BUFLEN, qfmt, i);
6705 		queue_node = SYSCTL_ADD_NODE(ctx_list, child, OID_AUTO, namebuf,
6706 		    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "Queue Name");
6707 		queue_list = SYSCTL_CHILDREN(queue_node);
6708 #if MEMORY_LOGGING
6709 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_dequeued",
6710 				CTLFLAG_RD,
6711 				&txq->ift_dequeued, "total mbufs freed");
6712 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_enqueued",
6713 				CTLFLAG_RD,
6714 				&txq->ift_enqueued, "total mbufs enqueued");
6715 #endif
6716 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "mbuf_defrag",
6717 				   CTLFLAG_RD,
6718 				   &txq->ift_mbuf_defrag, "# of times m_defrag was called");
6719 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "m_pullups",
6720 				   CTLFLAG_RD,
6721 				   &txq->ift_pullups, "# of times m_pullup was called");
6722 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "mbuf_defrag_failed",
6723 				   CTLFLAG_RD,
6724 				   &txq->ift_mbuf_defrag_failed, "# of times m_defrag failed");
6725 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "no_desc_avail",
6726 				   CTLFLAG_RD,
6727 				   &txq->ift_no_desc_avail, "# of times no descriptors were available");
6728 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "tx_map_failed",
6729 				   CTLFLAG_RD,
6730 				   &txq->ift_map_failed, "# of times DMA map failed");
6731 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txd_encap_efbig",
6732 				   CTLFLAG_RD,
6733 				   &txq->ift_txd_encap_efbig, "# of times txd_encap returned EFBIG");
6734 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "no_tx_dma_setup",
6735 				   CTLFLAG_RD,
6736 				   &txq->ift_no_tx_dma_setup, "# of times map failed for other than EFBIG");
6737 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_pidx",
6738 				   CTLFLAG_RD,
6739 				   &txq->ift_pidx, 1, "Producer Index");
6740 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_cidx",
6741 				   CTLFLAG_RD,
6742 				   &txq->ift_cidx, 1, "Consumer Index");
6743 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_cidx_processed",
6744 				   CTLFLAG_RD,
6745 				   &txq->ift_cidx_processed, 1, "Consumer Index seen by credit update");
6746 		SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "txq_in_use",
6747 				   CTLFLAG_RD,
6748 				   &txq->ift_in_use, 1, "descriptors in use");
6749 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_processed",
6750 				   CTLFLAG_RD,
6751 				   &txq->ift_processed, "descriptors procesed for clean");
6752 		SYSCTL_ADD_QUAD(ctx_list, queue_list, OID_AUTO, "txq_cleaned",
6753 				   CTLFLAG_RD,
6754 				   &txq->ift_cleaned, "total cleaned");
6755 		SYSCTL_ADD_PROC(ctx_list, queue_list, OID_AUTO, "ring_state",
6756 		    CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_NEEDGIANT,
6757 		    __DEVOLATILE(uint64_t *, &txq->ift_br->state), 0,
6758 		    mp_ring_state_handler, "A", "soft ring state");
6759 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_enqueues",
6760 				       CTLFLAG_RD, &txq->ift_br->enqueues,
6761 				       "# of enqueues to the mp_ring for this queue");
6762 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_drops",
6763 				       CTLFLAG_RD, &txq->ift_br->drops,
6764 				       "# of drops in the mp_ring for this queue");
6765 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_starts",
6766 				       CTLFLAG_RD, &txq->ift_br->starts,
6767 				       "# of normal consumer starts in the mp_ring for this queue");
6768 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_stalls",
6769 				       CTLFLAG_RD, &txq->ift_br->stalls,
6770 					       "# of consumer stalls in the mp_ring for this queue");
6771 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_restarts",
6772 			       CTLFLAG_RD, &txq->ift_br->restarts,
6773 				       "# of consumer restarts in the mp_ring for this queue");
6774 		SYSCTL_ADD_COUNTER_U64(ctx_list, queue_list, OID_AUTO, "r_abdications",
6775 				       CTLFLAG_RD, &txq->ift_br->abdications,
6776 				       "# of consumer abdications in the mp_ring for this queue");
6777 	}
6778 
6779 	if (scctx->isc_nrxqsets > 100)
6780 		qfmt = "rxq%03d";
6781 	else if (scctx->isc_nrxqsets > 10)
6782 		qfmt = "rxq%02d";
6783 	else
6784 		qfmt = "rxq%d";
6785 	for (i = 0, rxq = ctx->ifc_rxqs; i < scctx->isc_nrxqsets; i++, rxq++) {
6786 		snprintf(namebuf, NAME_BUFLEN, qfmt, i);
6787 		queue_node = SYSCTL_ADD_NODE(ctx_list, child, OID_AUTO, namebuf,
6788 		    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "Queue Name");
6789 		queue_list = SYSCTL_CHILDREN(queue_node);
6790 		if (sctx->isc_flags & IFLIB_HAS_RXCQ) {
6791 			SYSCTL_ADD_U16(ctx_list, queue_list, OID_AUTO, "rxq_cq_cidx",
6792 				       CTLFLAG_RD,
6793 				       &rxq->ifr_cq_cidx, 1, "Consumer Index");
6794 		}
6795 
6796 		for (j = 0, fl = rxq->ifr_fl; j < rxq->ifr_nfl; j++, fl++) {
6797 			snprintf(namebuf, NAME_BUFLEN, "rxq_fl%d", j);
6798 			fl_node = SYSCTL_ADD_NODE(ctx_list, queue_list, OID_AUTO, namebuf,
6799 			    CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, "freelist Name");
6800 			fl_list = SYSCTL_CHILDREN(fl_node);
6801 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "pidx",
6802 				       CTLFLAG_RD,
6803 				       &fl->ifl_pidx, 1, "Producer Index");
6804 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "cidx",
6805 				       CTLFLAG_RD,
6806 				       &fl->ifl_cidx, 1, "Consumer Index");
6807 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "credits",
6808 				       CTLFLAG_RD,
6809 				       &fl->ifl_credits, 1, "credits available");
6810 			SYSCTL_ADD_U16(ctx_list, fl_list, OID_AUTO, "buf_size",
6811 				       CTLFLAG_RD,
6812 				       &fl->ifl_buf_size, 1, "buffer size");
6813 #if MEMORY_LOGGING
6814 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_m_enqueued",
6815 					CTLFLAG_RD,
6816 					&fl->ifl_m_enqueued, "mbufs allocated");
6817 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_m_dequeued",
6818 					CTLFLAG_RD,
6819 					&fl->ifl_m_dequeued, "mbufs freed");
6820 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_cl_enqueued",
6821 					CTLFLAG_RD,
6822 					&fl->ifl_cl_enqueued, "clusters allocated");
6823 			SYSCTL_ADD_QUAD(ctx_list, fl_list, OID_AUTO, "fl_cl_dequeued",
6824 					CTLFLAG_RD,
6825 					&fl->ifl_cl_dequeued, "clusters freed");
6826 #endif
6827 		}
6828 	}
6829 
6830 }
6831 
6832 void
6833 iflib_request_reset(if_ctx_t ctx)
6834 {
6835 
6836 	STATE_LOCK(ctx);
6837 	ctx->ifc_flags |= IFC_DO_RESET;
6838 	STATE_UNLOCK(ctx);
6839 }
6840 
6841 #ifndef __NO_STRICT_ALIGNMENT
6842 static struct mbuf *
6843 iflib_fixup_rx(struct mbuf *m)
6844 {
6845 	struct mbuf *n;
6846 
6847 	if (m->m_len <= (MCLBYTES - ETHER_HDR_LEN)) {
6848 		bcopy(m->m_data, m->m_data + ETHER_HDR_LEN, m->m_len);
6849 		m->m_data += ETHER_HDR_LEN;
6850 		n = m;
6851 	} else {
6852 		MGETHDR(n, M_NOWAIT, MT_DATA);
6853 		if (n == NULL) {
6854 			m_freem(m);
6855 			return (NULL);
6856 		}
6857 		bcopy(m->m_data, n->m_data, ETHER_HDR_LEN);
6858 		m->m_data += ETHER_HDR_LEN;
6859 		m->m_len -= ETHER_HDR_LEN;
6860 		n->m_len = ETHER_HDR_LEN;
6861 		M_MOVE_PKTHDR(n, m);
6862 		n->m_next = m;
6863 	}
6864 	return (n);
6865 }
6866 #endif
6867 
6868 #ifdef DEBUGNET
6869 static void
6870 iflib_debugnet_init(if_t ifp, int *nrxr, int *ncl, int *clsize)
6871 {
6872 	if_ctx_t ctx;
6873 
6874 	ctx = if_getsoftc(ifp);
6875 	CTX_LOCK(ctx);
6876 	*nrxr = NRXQSETS(ctx);
6877 	*ncl = ctx->ifc_rxqs[0].ifr_fl->ifl_size;
6878 	*clsize = ctx->ifc_rxqs[0].ifr_fl->ifl_buf_size;
6879 	CTX_UNLOCK(ctx);
6880 }
6881 
6882 static void
6883 iflib_debugnet_event(if_t ifp, enum debugnet_ev event)
6884 {
6885 	if_ctx_t ctx;
6886 	if_softc_ctx_t scctx;
6887 	iflib_fl_t fl;
6888 	iflib_rxq_t rxq;
6889 	int i, j;
6890 
6891 	ctx = if_getsoftc(ifp);
6892 	scctx = &ctx->ifc_softc_ctx;
6893 
6894 	switch (event) {
6895 	case DEBUGNET_START:
6896 		for (i = 0; i < scctx->isc_nrxqsets; i++) {
6897 			rxq = &ctx->ifc_rxqs[i];
6898 			for (j = 0; j < rxq->ifr_nfl; j++) {
6899 				fl = rxq->ifr_fl;
6900 				fl->ifl_zone = m_getzone(fl->ifl_buf_size);
6901 			}
6902 		}
6903 		iflib_no_tx_batch = 1;
6904 		break;
6905 	default:
6906 		break;
6907 	}
6908 }
6909 
6910 static int
6911 iflib_debugnet_transmit(if_t ifp, struct mbuf *m)
6912 {
6913 	if_ctx_t ctx;
6914 	iflib_txq_t txq;
6915 	int error;
6916 
6917 	ctx = if_getsoftc(ifp);
6918 	if ((if_getdrvflags(ifp) & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
6919 	    IFF_DRV_RUNNING)
6920 		return (EBUSY);
6921 
6922 	txq = &ctx->ifc_txqs[0];
6923 	error = iflib_encap(txq, &m);
6924 	if (error == 0)
6925 		(void)iflib_txd_db_check(txq, true);
6926 	return (error);
6927 }
6928 
6929 static int
6930 iflib_debugnet_poll(if_t ifp, int count)
6931 {
6932 	struct epoch_tracker et;
6933 	if_ctx_t ctx;
6934 	if_softc_ctx_t scctx;
6935 	iflib_txq_t txq;
6936 	int i;
6937 
6938 	ctx = if_getsoftc(ifp);
6939 	scctx = &ctx->ifc_softc_ctx;
6940 
6941 	if ((if_getdrvflags(ifp) & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
6942 	    IFF_DRV_RUNNING)
6943 		return (EBUSY);
6944 
6945 	txq = &ctx->ifc_txqs[0];
6946 	(void)iflib_completed_tx_reclaim(txq, RECLAIM_THRESH(ctx));
6947 
6948 	NET_EPOCH_ENTER(et);
6949 	for (i = 0; i < scctx->isc_nrxqsets; i++)
6950 		(void)iflib_rxeof(&ctx->ifc_rxqs[i], 16 /* XXX */);
6951 	NET_EPOCH_EXIT(et);
6952 	return (0);
6953 }
6954 #endif /* DEBUGNET */
6955