1 /************************************************************************** 2 3 Copyright (c) 2007-2009, Chelsio Inc. 4 All rights reserved. 5 6 Redistribution and use in source and binary forms, with or without 7 modification, are permitted provided that the following conditions are met: 8 9 1. Redistributions of source code must retain the above copyright notice, 10 this list of conditions and the following disclaimer. 11 12 2. Neither the name of the Chelsio Corporation nor the names of its 13 contributors may be used to endorse or promote products derived from 14 this software without specific prior written permission. 15 16 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 17 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 20 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 21 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 22 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 23 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 24 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 25 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 26 POSSIBILITY OF SUCH DAMAGE. 27 28 ***************************************************************************/ 29 30 #include <sys/cdefs.h> 31 __FBSDID("$FreeBSD$"); 32 33 #include "opt_inet6.h" 34 #include "opt_inet.h" 35 36 #include <sys/param.h> 37 #include <sys/systm.h> 38 #include <sys/kernel.h> 39 #include <sys/module.h> 40 #include <sys/bus.h> 41 #include <sys/conf.h> 42 #include <machine/bus.h> 43 #include <machine/resource.h> 44 #include <sys/bus_dma.h> 45 #include <sys/rman.h> 46 #include <sys/queue.h> 47 #include <sys/sysctl.h> 48 #include <sys/taskqueue.h> 49 50 #include <sys/proc.h> 51 #include <sys/sbuf.h> 52 #include <sys/sched.h> 53 #include <sys/smp.h> 54 #include <sys/systm.h> 55 #include <sys/syslog.h> 56 #include <sys/socket.h> 57 #include <sys/sglist.h> 58 59 #include <net/if.h> 60 #include <net/if_var.h> 61 #include <net/bpf.h> 62 #include <net/ethernet.h> 63 #include <net/if_vlan_var.h> 64 65 #include <netinet/in_systm.h> 66 #include <netinet/in.h> 67 #include <netinet/ip.h> 68 #include <netinet/ip6.h> 69 #include <netinet/tcp.h> 70 71 #include <dev/pci/pcireg.h> 72 #include <dev/pci/pcivar.h> 73 74 #include <vm/vm.h> 75 #include <vm/pmap.h> 76 77 #include <cxgb_include.h> 78 #include <sys/mvec.h> 79 80 int txq_fills = 0; 81 int multiq_tx_enable = 1; 82 83 #ifdef TCP_OFFLOAD 84 CTASSERT(NUM_CPL_HANDLERS >= NUM_CPL_CMDS); 85 #endif 86 87 extern struct sysctl_oid_list sysctl__hw_cxgb_children; 88 int cxgb_txq_buf_ring_size = TX_ETH_Q_SIZE; 89 SYSCTL_INT(_hw_cxgb, OID_AUTO, txq_mr_size, CTLFLAG_RDTUN, &cxgb_txq_buf_ring_size, 0, 90 "size of per-queue mbuf ring"); 91 92 static int cxgb_tx_coalesce_force = 0; 93 SYSCTL_INT(_hw_cxgb, OID_AUTO, tx_coalesce_force, CTLFLAG_RWTUN, 94 &cxgb_tx_coalesce_force, 0, 95 "coalesce small packets into a single work request regardless of ring state"); 96 97 #define COALESCE_START_DEFAULT TX_ETH_Q_SIZE>>1 98 #define COALESCE_START_MAX (TX_ETH_Q_SIZE-(TX_ETH_Q_SIZE>>3)) 99 #define COALESCE_STOP_DEFAULT TX_ETH_Q_SIZE>>2 100 #define COALESCE_STOP_MIN TX_ETH_Q_SIZE>>5 101 #define TX_RECLAIM_DEFAULT TX_ETH_Q_SIZE>>5 102 #define TX_RECLAIM_MAX TX_ETH_Q_SIZE>>2 103 #define TX_RECLAIM_MIN TX_ETH_Q_SIZE>>6 104 105 106 static int cxgb_tx_coalesce_enable_start = COALESCE_START_DEFAULT; 107 SYSCTL_INT(_hw_cxgb, OID_AUTO, tx_coalesce_enable_start, CTLFLAG_RWTUN, 108 &cxgb_tx_coalesce_enable_start, 0, 109 "coalesce enable threshold"); 110 static int cxgb_tx_coalesce_enable_stop = COALESCE_STOP_DEFAULT; 111 SYSCTL_INT(_hw_cxgb, OID_AUTO, tx_coalesce_enable_stop, CTLFLAG_RWTUN, 112 &cxgb_tx_coalesce_enable_stop, 0, 113 "coalesce disable threshold"); 114 static int cxgb_tx_reclaim_threshold = TX_RECLAIM_DEFAULT; 115 SYSCTL_INT(_hw_cxgb, OID_AUTO, tx_reclaim_threshold, CTLFLAG_RWTUN, 116 &cxgb_tx_reclaim_threshold, 0, 117 "tx cleaning minimum threshold"); 118 119 /* 120 * XXX don't re-enable this until TOE stops assuming 121 * we have an m_ext 122 */ 123 static int recycle_enable = 0; 124 125 extern int cxgb_use_16k_clusters; 126 extern int nmbjumbop; 127 extern int nmbjumbo9; 128 extern int nmbjumbo16; 129 130 #define USE_GTS 0 131 132 #define SGE_RX_SM_BUF_SIZE 1536 133 #define SGE_RX_DROP_THRES 16 134 #define SGE_RX_COPY_THRES 128 135 136 /* 137 * Period of the Tx buffer reclaim timer. This timer does not need to run 138 * frequently as Tx buffers are usually reclaimed by new Tx packets. 139 */ 140 #define TX_RECLAIM_PERIOD (hz >> 1) 141 142 /* 143 * Values for sge_txq.flags 144 */ 145 enum { 146 TXQ_RUNNING = 1 << 0, /* fetch engine is running */ 147 TXQ_LAST_PKT_DB = 1 << 1, /* last packet rang the doorbell */ 148 }; 149 150 struct tx_desc { 151 uint64_t flit[TX_DESC_FLITS]; 152 } __packed; 153 154 struct rx_desc { 155 uint32_t addr_lo; 156 uint32_t len_gen; 157 uint32_t gen2; 158 uint32_t addr_hi; 159 } __packed; 160 161 struct rsp_desc { /* response queue descriptor */ 162 struct rss_header rss_hdr; 163 uint32_t flags; 164 uint32_t len_cq; 165 uint8_t imm_data[47]; 166 uint8_t intr_gen; 167 } __packed; 168 169 #define RX_SW_DESC_MAP_CREATED (1 << 0) 170 #define TX_SW_DESC_MAP_CREATED (1 << 1) 171 #define RX_SW_DESC_INUSE (1 << 3) 172 #define TX_SW_DESC_MAPPED (1 << 4) 173 174 #define RSPQ_NSOP_NEOP G_RSPD_SOP_EOP(0) 175 #define RSPQ_EOP G_RSPD_SOP_EOP(F_RSPD_EOP) 176 #define RSPQ_SOP G_RSPD_SOP_EOP(F_RSPD_SOP) 177 #define RSPQ_SOP_EOP G_RSPD_SOP_EOP(F_RSPD_SOP|F_RSPD_EOP) 178 179 struct tx_sw_desc { /* SW state per Tx descriptor */ 180 struct mbuf *m; 181 bus_dmamap_t map; 182 int flags; 183 }; 184 185 struct rx_sw_desc { /* SW state per Rx descriptor */ 186 caddr_t rxsd_cl; 187 struct mbuf *m; 188 bus_dmamap_t map; 189 int flags; 190 }; 191 192 struct txq_state { 193 unsigned int compl; 194 unsigned int gen; 195 unsigned int pidx; 196 }; 197 198 struct refill_fl_cb_arg { 199 int error; 200 bus_dma_segment_t seg; 201 int nseg; 202 }; 203 204 205 /* 206 * Maps a number of flits to the number of Tx descriptors that can hold them. 207 * The formula is 208 * 209 * desc = 1 + (flits - 2) / (WR_FLITS - 1). 210 * 211 * HW allows up to 4 descriptors to be combined into a WR. 212 */ 213 static uint8_t flit_desc_map[] = { 214 0, 215 #if SGE_NUM_GENBITS == 1 216 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 217 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 218 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 219 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4 220 #elif SGE_NUM_GENBITS == 2 221 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 222 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 223 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 224 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 225 #else 226 # error "SGE_NUM_GENBITS must be 1 or 2" 227 #endif 228 }; 229 230 #define TXQ_LOCK_ASSERT(qs) mtx_assert(&(qs)->lock, MA_OWNED) 231 #define TXQ_TRYLOCK(qs) mtx_trylock(&(qs)->lock) 232 #define TXQ_LOCK(qs) mtx_lock(&(qs)->lock) 233 #define TXQ_UNLOCK(qs) mtx_unlock(&(qs)->lock) 234 #define TXQ_RING_EMPTY(qs) drbr_empty((qs)->port->ifp, (qs)->txq[TXQ_ETH].txq_mr) 235 #define TXQ_RING_NEEDS_ENQUEUE(qs) \ 236 drbr_needs_enqueue((qs)->port->ifp, (qs)->txq[TXQ_ETH].txq_mr) 237 #define TXQ_RING_FLUSH(qs) drbr_flush((qs)->port->ifp, (qs)->txq[TXQ_ETH].txq_mr) 238 #define TXQ_RING_DEQUEUE_COND(qs, func, arg) \ 239 drbr_dequeue_cond((qs)->port->ifp, (qs)->txq[TXQ_ETH].txq_mr, func, arg) 240 #define TXQ_RING_DEQUEUE(qs) \ 241 drbr_dequeue((qs)->port->ifp, (qs)->txq[TXQ_ETH].txq_mr) 242 243 int cxgb_debug = 0; 244 245 static void sge_timer_cb(void *arg); 246 static void sge_timer_reclaim(void *arg, int ncount); 247 static void sge_txq_reclaim_handler(void *arg, int ncount); 248 static void cxgb_start_locked(struct sge_qset *qs); 249 250 /* 251 * XXX need to cope with bursty scheduling by looking at a wider 252 * window than we are now for determining the need for coalescing 253 * 254 */ 255 static __inline uint64_t 256 check_pkt_coalesce(struct sge_qset *qs) 257 { 258 struct adapter *sc; 259 struct sge_txq *txq; 260 uint8_t *fill; 261 262 if (__predict_false(cxgb_tx_coalesce_force)) 263 return (1); 264 txq = &qs->txq[TXQ_ETH]; 265 sc = qs->port->adapter; 266 fill = &sc->tunq_fill[qs->idx]; 267 268 if (cxgb_tx_coalesce_enable_start > COALESCE_START_MAX) 269 cxgb_tx_coalesce_enable_start = COALESCE_START_MAX; 270 if (cxgb_tx_coalesce_enable_stop < COALESCE_STOP_MIN) 271 cxgb_tx_coalesce_enable_start = COALESCE_STOP_MIN; 272 /* 273 * if the hardware transmit queue is more than 1/8 full 274 * we mark it as coalescing - we drop back from coalescing 275 * when we go below 1/32 full and there are no packets enqueued, 276 * this provides us with some degree of hysteresis 277 */ 278 if (*fill != 0 && (txq->in_use <= cxgb_tx_coalesce_enable_stop) && 279 TXQ_RING_EMPTY(qs) && (qs->coalescing == 0)) 280 *fill = 0; 281 else if (*fill == 0 && (txq->in_use >= cxgb_tx_coalesce_enable_start)) 282 *fill = 1; 283 284 return (sc->tunq_coalesce); 285 } 286 287 #ifdef __LP64__ 288 static void 289 set_wr_hdr(struct work_request_hdr *wrp, uint32_t wr_hi, uint32_t wr_lo) 290 { 291 uint64_t wr_hilo; 292 #if _BYTE_ORDER == _LITTLE_ENDIAN 293 wr_hilo = wr_hi; 294 wr_hilo |= (((uint64_t)wr_lo)<<32); 295 #else 296 wr_hilo = wr_lo; 297 wr_hilo |= (((uint64_t)wr_hi)<<32); 298 #endif 299 wrp->wrh_hilo = wr_hilo; 300 } 301 #else 302 static void 303 set_wr_hdr(struct work_request_hdr *wrp, uint32_t wr_hi, uint32_t wr_lo) 304 { 305 306 wrp->wrh_hi = wr_hi; 307 wmb(); 308 wrp->wrh_lo = wr_lo; 309 } 310 #endif 311 312 struct coalesce_info { 313 int count; 314 int nbytes; 315 }; 316 317 static int 318 coalesce_check(struct mbuf *m, void *arg) 319 { 320 struct coalesce_info *ci = arg; 321 int *count = &ci->count; 322 int *nbytes = &ci->nbytes; 323 324 if ((*nbytes == 0) || ((*nbytes + m->m_len <= 10500) && 325 (*count < 7) && (m->m_next == NULL))) { 326 *count += 1; 327 *nbytes += m->m_len; 328 return (1); 329 } 330 return (0); 331 } 332 333 static struct mbuf * 334 cxgb_dequeue(struct sge_qset *qs) 335 { 336 struct mbuf *m, *m_head, *m_tail; 337 struct coalesce_info ci; 338 339 340 if (check_pkt_coalesce(qs) == 0) 341 return TXQ_RING_DEQUEUE(qs); 342 343 m_head = m_tail = NULL; 344 ci.count = ci.nbytes = 0; 345 do { 346 m = TXQ_RING_DEQUEUE_COND(qs, coalesce_check, &ci); 347 if (m_head == NULL) { 348 m_tail = m_head = m; 349 } else if (m != NULL) { 350 m_tail->m_nextpkt = m; 351 m_tail = m; 352 } 353 } while (m != NULL); 354 if (ci.count > 7) 355 panic("trying to coalesce %d packets in to one WR", ci.count); 356 return (m_head); 357 } 358 359 /** 360 * reclaim_completed_tx - reclaims completed Tx descriptors 361 * @adapter: the adapter 362 * @q: the Tx queue to reclaim completed descriptors from 363 * 364 * Reclaims Tx descriptors that the SGE has indicated it has processed, 365 * and frees the associated buffers if possible. Called with the Tx 366 * queue's lock held. 367 */ 368 static __inline int 369 reclaim_completed_tx(struct sge_qset *qs, int reclaim_min, int queue) 370 { 371 struct sge_txq *q = &qs->txq[queue]; 372 int reclaim = desc_reclaimable(q); 373 374 if ((cxgb_tx_reclaim_threshold > TX_RECLAIM_MAX) || 375 (cxgb_tx_reclaim_threshold < TX_RECLAIM_MIN)) 376 cxgb_tx_reclaim_threshold = TX_RECLAIM_DEFAULT; 377 378 if (reclaim < reclaim_min) 379 return (0); 380 381 mtx_assert(&qs->lock, MA_OWNED); 382 if (reclaim > 0) { 383 t3_free_tx_desc(qs, reclaim, queue); 384 q->cleaned += reclaim; 385 q->in_use -= reclaim; 386 } 387 if (isset(&qs->txq_stopped, TXQ_ETH)) 388 clrbit(&qs->txq_stopped, TXQ_ETH); 389 390 return (reclaim); 391 } 392 393 /** 394 * should_restart_tx - are there enough resources to restart a Tx queue? 395 * @q: the Tx queue 396 * 397 * Checks if there are enough descriptors to restart a suspended Tx queue. 398 */ 399 static __inline int 400 should_restart_tx(const struct sge_txq *q) 401 { 402 unsigned int r = q->processed - q->cleaned; 403 404 return q->in_use - r < (q->size >> 1); 405 } 406 407 /** 408 * t3_sge_init - initialize SGE 409 * @adap: the adapter 410 * @p: the SGE parameters 411 * 412 * Performs SGE initialization needed every time after a chip reset. 413 * We do not initialize any of the queue sets here, instead the driver 414 * top-level must request those individually. We also do not enable DMA 415 * here, that should be done after the queues have been set up. 416 */ 417 void 418 t3_sge_init(adapter_t *adap, struct sge_params *p) 419 { 420 u_int ctrl, ups; 421 422 ups = 0; /* = ffs(pci_resource_len(adap->pdev, 2) >> 12); */ 423 424 ctrl = F_DROPPKT | V_PKTSHIFT(2) | F_FLMODE | F_AVOIDCQOVFL | 425 F_CQCRDTCTRL | F_CONGMODE | F_TNLFLMODE | F_FATLPERREN | 426 V_HOSTPAGESIZE(PAGE_SHIFT - 11) | F_BIGENDIANINGRESS | 427 V_USERSPACESIZE(ups ? ups - 1 : 0) | F_ISCSICOALESCING; 428 #if SGE_NUM_GENBITS == 1 429 ctrl |= F_EGRGENCTRL; 430 #endif 431 if (adap->params.rev > 0) { 432 if (!(adap->flags & (USING_MSIX | USING_MSI))) 433 ctrl |= F_ONEINTMULTQ | F_OPTONEINTMULTQ; 434 } 435 t3_write_reg(adap, A_SG_CONTROL, ctrl); 436 t3_write_reg(adap, A_SG_EGR_RCQ_DRB_THRSH, V_HIRCQDRBTHRSH(512) | 437 V_LORCQDRBTHRSH(512)); 438 t3_write_reg(adap, A_SG_TIMER_TICK, core_ticks_per_usec(adap) / 10); 439 t3_write_reg(adap, A_SG_CMDQ_CREDIT_TH, V_THRESHOLD(32) | 440 V_TIMEOUT(200 * core_ticks_per_usec(adap))); 441 t3_write_reg(adap, A_SG_HI_DRB_HI_THRSH, 442 adap->params.rev < T3_REV_C ? 1000 : 500); 443 t3_write_reg(adap, A_SG_HI_DRB_LO_THRSH, 256); 444 t3_write_reg(adap, A_SG_LO_DRB_HI_THRSH, 1000); 445 t3_write_reg(adap, A_SG_LO_DRB_LO_THRSH, 256); 446 t3_write_reg(adap, A_SG_OCO_BASE, V_BASE1(0xfff)); 447 t3_write_reg(adap, A_SG_DRB_PRI_THRESH, 63 * 1024); 448 } 449 450 451 /** 452 * sgl_len - calculates the size of an SGL of the given capacity 453 * @n: the number of SGL entries 454 * 455 * Calculates the number of flits needed for a scatter/gather list that 456 * can hold the given number of entries. 457 */ 458 static __inline unsigned int 459 sgl_len(unsigned int n) 460 { 461 return ((3 * n) / 2 + (n & 1)); 462 } 463 464 /** 465 * get_imm_packet - return the next ingress packet buffer from a response 466 * @resp: the response descriptor containing the packet data 467 * 468 * Return a packet containing the immediate data of the given response. 469 */ 470 static int 471 get_imm_packet(adapter_t *sc, const struct rsp_desc *resp, struct mbuf *m) 472 { 473 474 if (resp->rss_hdr.opcode == CPL_RX_DATA) { 475 const struct cpl_rx_data *cpl = (const void *)&resp->imm_data[0]; 476 m->m_len = sizeof(*cpl) + ntohs(cpl->len); 477 } else if (resp->rss_hdr.opcode == CPL_RX_PKT) { 478 const struct cpl_rx_pkt *cpl = (const void *)&resp->imm_data[0]; 479 m->m_len = sizeof(*cpl) + ntohs(cpl->len); 480 } else 481 m->m_len = IMMED_PKT_SIZE; 482 m->m_ext.ext_buf = NULL; 483 m->m_ext.ext_type = 0; 484 memcpy(mtod(m, uint8_t *), resp->imm_data, m->m_len); 485 return (0); 486 } 487 488 static __inline u_int 489 flits_to_desc(u_int n) 490 { 491 return (flit_desc_map[n]); 492 } 493 494 #define SGE_PARERR (F_CPPARITYERROR | F_OCPARITYERROR | F_RCPARITYERROR | \ 495 F_IRPARITYERROR | V_ITPARITYERROR(M_ITPARITYERROR) | \ 496 V_FLPARITYERROR(M_FLPARITYERROR) | F_LODRBPARITYERROR | \ 497 F_HIDRBPARITYERROR | F_LORCQPARITYERROR | \ 498 F_HIRCQPARITYERROR) 499 #define SGE_FRAMINGERR (F_UC_REQ_FRAMINGERROR | F_R_REQ_FRAMINGERROR) 500 #define SGE_FATALERR (SGE_PARERR | SGE_FRAMINGERR | F_RSPQCREDITOVERFOW | \ 501 F_RSPQDISABLED) 502 503 /** 504 * t3_sge_err_intr_handler - SGE async event interrupt handler 505 * @adapter: the adapter 506 * 507 * Interrupt handler for SGE asynchronous (non-data) events. 508 */ 509 void 510 t3_sge_err_intr_handler(adapter_t *adapter) 511 { 512 unsigned int v, status; 513 514 status = t3_read_reg(adapter, A_SG_INT_CAUSE); 515 if (status & SGE_PARERR) 516 CH_ALERT(adapter, "SGE parity error (0x%x)\n", 517 status & SGE_PARERR); 518 if (status & SGE_FRAMINGERR) 519 CH_ALERT(adapter, "SGE framing error (0x%x)\n", 520 status & SGE_FRAMINGERR); 521 if (status & F_RSPQCREDITOVERFOW) 522 CH_ALERT(adapter, "SGE response queue credit overflow\n"); 523 524 if (status & F_RSPQDISABLED) { 525 v = t3_read_reg(adapter, A_SG_RSPQ_FL_STATUS); 526 527 CH_ALERT(adapter, 528 "packet delivered to disabled response queue (0x%x)\n", 529 (v >> S_RSPQ0DISABLED) & 0xff); 530 } 531 532 t3_write_reg(adapter, A_SG_INT_CAUSE, status); 533 if (status & SGE_FATALERR) 534 t3_fatal_err(adapter); 535 } 536 537 void 538 t3_sge_prep(adapter_t *adap, struct sge_params *p) 539 { 540 int i, nqsets, fl_q_size, jumbo_q_size, use_16k, jumbo_buf_size; 541 542 nqsets = min(SGE_QSETS / adap->params.nports, mp_ncpus); 543 nqsets *= adap->params.nports; 544 545 fl_q_size = min(nmbclusters/(3*nqsets), FL_Q_SIZE); 546 547 while (!powerof2(fl_q_size)) 548 fl_q_size--; 549 550 use_16k = cxgb_use_16k_clusters != -1 ? cxgb_use_16k_clusters : 551 is_offload(adap); 552 553 #if __FreeBSD_version >= 700111 554 if (use_16k) { 555 jumbo_q_size = min(nmbjumbo16/(3*nqsets), JUMBO_Q_SIZE); 556 jumbo_buf_size = MJUM16BYTES; 557 } else { 558 jumbo_q_size = min(nmbjumbo9/(3*nqsets), JUMBO_Q_SIZE); 559 jumbo_buf_size = MJUM9BYTES; 560 } 561 #else 562 jumbo_q_size = min(nmbjumbop/(3*nqsets), JUMBO_Q_SIZE); 563 jumbo_buf_size = MJUMPAGESIZE; 564 #endif 565 while (!powerof2(jumbo_q_size)) 566 jumbo_q_size--; 567 568 if (fl_q_size < (FL_Q_SIZE / 4) || jumbo_q_size < (JUMBO_Q_SIZE / 2)) 569 device_printf(adap->dev, 570 "Insufficient clusters and/or jumbo buffers.\n"); 571 572 p->max_pkt_size = jumbo_buf_size - sizeof(struct cpl_rx_data); 573 574 for (i = 0; i < SGE_QSETS; ++i) { 575 struct qset_params *q = p->qset + i; 576 577 if (adap->params.nports > 2) { 578 q->coalesce_usecs = 50; 579 } else { 580 #ifdef INVARIANTS 581 q->coalesce_usecs = 10; 582 #else 583 q->coalesce_usecs = 5; 584 #endif 585 } 586 q->polling = 0; 587 q->rspq_size = RSPQ_Q_SIZE; 588 q->fl_size = fl_q_size; 589 q->jumbo_size = jumbo_q_size; 590 q->jumbo_buf_size = jumbo_buf_size; 591 q->txq_size[TXQ_ETH] = TX_ETH_Q_SIZE; 592 q->txq_size[TXQ_OFLD] = is_offload(adap) ? TX_OFLD_Q_SIZE : 16; 593 q->txq_size[TXQ_CTRL] = TX_CTRL_Q_SIZE; 594 q->cong_thres = 0; 595 } 596 } 597 598 int 599 t3_sge_alloc(adapter_t *sc) 600 { 601 602 /* The parent tag. */ 603 if (bus_dma_tag_create( bus_get_dma_tag(sc->dev),/* PCI parent */ 604 1, 0, /* algnmnt, boundary */ 605 BUS_SPACE_MAXADDR, /* lowaddr */ 606 BUS_SPACE_MAXADDR, /* highaddr */ 607 NULL, NULL, /* filter, filterarg */ 608 BUS_SPACE_MAXSIZE_32BIT,/* maxsize */ 609 BUS_SPACE_UNRESTRICTED, /* nsegments */ 610 BUS_SPACE_MAXSIZE_32BIT,/* maxsegsize */ 611 0, /* flags */ 612 NULL, NULL, /* lock, lockarg */ 613 &sc->parent_dmat)) { 614 device_printf(sc->dev, "Cannot allocate parent DMA tag\n"); 615 return (ENOMEM); 616 } 617 618 /* 619 * DMA tag for normal sized RX frames 620 */ 621 if (bus_dma_tag_create(sc->parent_dmat, MCLBYTES, 0, BUS_SPACE_MAXADDR, 622 BUS_SPACE_MAXADDR, NULL, NULL, MCLBYTES, 1, 623 MCLBYTES, BUS_DMA_ALLOCNOW, NULL, NULL, &sc->rx_dmat)) { 624 device_printf(sc->dev, "Cannot allocate RX DMA tag\n"); 625 return (ENOMEM); 626 } 627 628 /* 629 * DMA tag for jumbo sized RX frames. 630 */ 631 if (bus_dma_tag_create(sc->parent_dmat, MJUM16BYTES, 0, BUS_SPACE_MAXADDR, 632 BUS_SPACE_MAXADDR, NULL, NULL, MJUM16BYTES, 1, MJUM16BYTES, 633 BUS_DMA_ALLOCNOW, NULL, NULL, &sc->rx_jumbo_dmat)) { 634 device_printf(sc->dev, "Cannot allocate RX jumbo DMA tag\n"); 635 return (ENOMEM); 636 } 637 638 /* 639 * DMA tag for TX frames. 640 */ 641 if (bus_dma_tag_create(sc->parent_dmat, 1, 0, BUS_SPACE_MAXADDR, 642 BUS_SPACE_MAXADDR, NULL, NULL, TX_MAX_SIZE, TX_MAX_SEGS, 643 TX_MAX_SIZE, BUS_DMA_ALLOCNOW, 644 NULL, NULL, &sc->tx_dmat)) { 645 device_printf(sc->dev, "Cannot allocate TX DMA tag\n"); 646 return (ENOMEM); 647 } 648 649 return (0); 650 } 651 652 int 653 t3_sge_free(struct adapter * sc) 654 { 655 656 if (sc->tx_dmat != NULL) 657 bus_dma_tag_destroy(sc->tx_dmat); 658 659 if (sc->rx_jumbo_dmat != NULL) 660 bus_dma_tag_destroy(sc->rx_jumbo_dmat); 661 662 if (sc->rx_dmat != NULL) 663 bus_dma_tag_destroy(sc->rx_dmat); 664 665 if (sc->parent_dmat != NULL) 666 bus_dma_tag_destroy(sc->parent_dmat); 667 668 return (0); 669 } 670 671 void 672 t3_update_qset_coalesce(struct sge_qset *qs, const struct qset_params *p) 673 { 674 675 qs->rspq.holdoff_tmr = max(p->coalesce_usecs * 10, 1U); 676 qs->rspq.polling = 0 /* p->polling */; 677 } 678 679 #if !defined(__i386__) && !defined(__amd64__) 680 static void 681 refill_fl_cb(void *arg, bus_dma_segment_t *segs, int nseg, int error) 682 { 683 struct refill_fl_cb_arg *cb_arg = arg; 684 685 cb_arg->error = error; 686 cb_arg->seg = segs[0]; 687 cb_arg->nseg = nseg; 688 689 } 690 #endif 691 /** 692 * refill_fl - refill an SGE free-buffer list 693 * @sc: the controller softc 694 * @q: the free-list to refill 695 * @n: the number of new buffers to allocate 696 * 697 * (Re)populate an SGE free-buffer list with up to @n new packet buffers. 698 * The caller must assure that @n does not exceed the queue's capacity. 699 */ 700 static void 701 refill_fl(adapter_t *sc, struct sge_fl *q, int n) 702 { 703 struct rx_sw_desc *sd = &q->sdesc[q->pidx]; 704 struct rx_desc *d = &q->desc[q->pidx]; 705 struct refill_fl_cb_arg cb_arg; 706 struct mbuf *m; 707 caddr_t cl; 708 int err; 709 710 cb_arg.error = 0; 711 while (n--) { 712 /* 713 * We allocate an uninitialized mbuf + cluster, mbuf is 714 * initialized after rx. 715 */ 716 if (q->zone == zone_pack) { 717 if ((m = m_getcl(M_NOWAIT, MT_NOINIT, M_PKTHDR)) == NULL) 718 break; 719 cl = m->m_ext.ext_buf; 720 } else { 721 if ((cl = m_cljget(NULL, M_NOWAIT, q->buf_size)) == NULL) 722 break; 723 if ((m = m_gethdr(M_NOWAIT, MT_NOINIT)) == NULL) { 724 uma_zfree(q->zone, cl); 725 break; 726 } 727 } 728 if ((sd->flags & RX_SW_DESC_MAP_CREATED) == 0) { 729 if ((err = bus_dmamap_create(q->entry_tag, 0, &sd->map))) { 730 log(LOG_WARNING, "bus_dmamap_create failed %d\n", err); 731 uma_zfree(q->zone, cl); 732 goto done; 733 } 734 sd->flags |= RX_SW_DESC_MAP_CREATED; 735 } 736 #if !defined(__i386__) && !defined(__amd64__) 737 err = bus_dmamap_load(q->entry_tag, sd->map, 738 cl, q->buf_size, refill_fl_cb, &cb_arg, 0); 739 740 if (err != 0 || cb_arg.error) { 741 if (q->zone != zone_pack) 742 uma_zfree(q->zone, cl); 743 m_free(m); 744 goto done; 745 } 746 #else 747 cb_arg.seg.ds_addr = pmap_kextract((vm_offset_t)cl); 748 #endif 749 sd->flags |= RX_SW_DESC_INUSE; 750 sd->rxsd_cl = cl; 751 sd->m = m; 752 d->addr_lo = htobe32(cb_arg.seg.ds_addr & 0xffffffff); 753 d->addr_hi = htobe32(((uint64_t)cb_arg.seg.ds_addr >>32) & 0xffffffff); 754 d->len_gen = htobe32(V_FLD_GEN1(q->gen)); 755 d->gen2 = htobe32(V_FLD_GEN2(q->gen)); 756 757 d++; 758 sd++; 759 760 if (++q->pidx == q->size) { 761 q->pidx = 0; 762 q->gen ^= 1; 763 sd = q->sdesc; 764 d = q->desc; 765 } 766 q->credits++; 767 q->db_pending++; 768 } 769 770 done: 771 if (q->db_pending >= 32) { 772 q->db_pending = 0; 773 t3_write_reg(sc, A_SG_KDOORBELL, V_EGRCNTX(q->cntxt_id)); 774 } 775 } 776 777 778 /** 779 * free_rx_bufs - free the Rx buffers on an SGE free list 780 * @sc: the controle softc 781 * @q: the SGE free list to clean up 782 * 783 * Release the buffers on an SGE free-buffer Rx queue. HW fetching from 784 * this queue should be stopped before calling this function. 785 */ 786 static void 787 free_rx_bufs(adapter_t *sc, struct sge_fl *q) 788 { 789 u_int cidx = q->cidx; 790 791 while (q->credits--) { 792 struct rx_sw_desc *d = &q->sdesc[cidx]; 793 794 if (d->flags & RX_SW_DESC_INUSE) { 795 bus_dmamap_unload(q->entry_tag, d->map); 796 bus_dmamap_destroy(q->entry_tag, d->map); 797 if (q->zone == zone_pack) { 798 m_init(d->m, M_NOWAIT, MT_DATA, M_EXT); 799 uma_zfree(zone_pack, d->m); 800 } else { 801 m_init(d->m, M_NOWAIT, MT_DATA, 0); 802 uma_zfree(zone_mbuf, d->m); 803 uma_zfree(q->zone, d->rxsd_cl); 804 } 805 } 806 807 d->rxsd_cl = NULL; 808 d->m = NULL; 809 if (++cidx == q->size) 810 cidx = 0; 811 } 812 } 813 814 static __inline void 815 __refill_fl(adapter_t *adap, struct sge_fl *fl) 816 { 817 refill_fl(adap, fl, min(16U, fl->size - fl->credits)); 818 } 819 820 static __inline void 821 __refill_fl_lt(adapter_t *adap, struct sge_fl *fl, int max) 822 { 823 uint32_t reclaimable = fl->size - fl->credits; 824 825 if (reclaimable > 0) 826 refill_fl(adap, fl, min(max, reclaimable)); 827 } 828 829 /** 830 * recycle_rx_buf - recycle a receive buffer 831 * @adapter: the adapter 832 * @q: the SGE free list 833 * @idx: index of buffer to recycle 834 * 835 * Recycles the specified buffer on the given free list by adding it at 836 * the next available slot on the list. 837 */ 838 static void 839 recycle_rx_buf(adapter_t *adap, struct sge_fl *q, unsigned int idx) 840 { 841 struct rx_desc *from = &q->desc[idx]; 842 struct rx_desc *to = &q->desc[q->pidx]; 843 844 q->sdesc[q->pidx] = q->sdesc[idx]; 845 to->addr_lo = from->addr_lo; // already big endian 846 to->addr_hi = from->addr_hi; // likewise 847 wmb(); /* necessary ? */ 848 to->len_gen = htobe32(V_FLD_GEN1(q->gen)); 849 to->gen2 = htobe32(V_FLD_GEN2(q->gen)); 850 q->credits++; 851 852 if (++q->pidx == q->size) { 853 q->pidx = 0; 854 q->gen ^= 1; 855 } 856 t3_write_reg(adap, A_SG_KDOORBELL, V_EGRCNTX(q->cntxt_id)); 857 } 858 859 static void 860 alloc_ring_cb(void *arg, bus_dma_segment_t *segs, int nsegs, int error) 861 { 862 uint32_t *addr; 863 864 addr = arg; 865 *addr = segs[0].ds_addr; 866 } 867 868 static int 869 alloc_ring(adapter_t *sc, size_t nelem, size_t elem_size, size_t sw_size, 870 bus_addr_t *phys, void *desc, void *sdesc, bus_dma_tag_t *tag, 871 bus_dmamap_t *map, bus_dma_tag_t parent_entry_tag, bus_dma_tag_t *entry_tag) 872 { 873 size_t len = nelem * elem_size; 874 void *s = NULL; 875 void *p = NULL; 876 int err; 877 878 if ((err = bus_dma_tag_create(sc->parent_dmat, PAGE_SIZE, 0, 879 BUS_SPACE_MAXADDR_32BIT, 880 BUS_SPACE_MAXADDR, NULL, NULL, len, 1, 881 len, 0, NULL, NULL, tag)) != 0) { 882 device_printf(sc->dev, "Cannot allocate descriptor tag\n"); 883 return (ENOMEM); 884 } 885 886 if ((err = bus_dmamem_alloc(*tag, (void **)&p, BUS_DMA_NOWAIT, 887 map)) != 0) { 888 device_printf(sc->dev, "Cannot allocate descriptor memory\n"); 889 return (ENOMEM); 890 } 891 892 bus_dmamap_load(*tag, *map, p, len, alloc_ring_cb, phys, 0); 893 bzero(p, len); 894 *(void **)desc = p; 895 896 if (sw_size) { 897 len = nelem * sw_size; 898 s = malloc(len, M_DEVBUF, M_WAITOK|M_ZERO); 899 *(void **)sdesc = s; 900 } 901 if (parent_entry_tag == NULL) 902 return (0); 903 904 if ((err = bus_dma_tag_create(parent_entry_tag, 1, 0, 905 BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR, 906 NULL, NULL, TX_MAX_SIZE, TX_MAX_SEGS, 907 TX_MAX_SIZE, BUS_DMA_ALLOCNOW, 908 NULL, NULL, entry_tag)) != 0) { 909 device_printf(sc->dev, "Cannot allocate descriptor entry tag\n"); 910 return (ENOMEM); 911 } 912 return (0); 913 } 914 915 static void 916 sge_slow_intr_handler(void *arg, int ncount) 917 { 918 adapter_t *sc = arg; 919 920 t3_slow_intr_handler(sc); 921 t3_write_reg(sc, A_PL_INT_ENABLE0, sc->slow_intr_mask); 922 (void) t3_read_reg(sc, A_PL_INT_ENABLE0); 923 } 924 925 /** 926 * sge_timer_cb - perform periodic maintenance of an SGE qset 927 * @data: the SGE queue set to maintain 928 * 929 * Runs periodically from a timer to perform maintenance of an SGE queue 930 * set. It performs two tasks: 931 * 932 * a) Cleans up any completed Tx descriptors that may still be pending. 933 * Normal descriptor cleanup happens when new packets are added to a Tx 934 * queue so this timer is relatively infrequent and does any cleanup only 935 * if the Tx queue has not seen any new packets in a while. We make a 936 * best effort attempt to reclaim descriptors, in that we don't wait 937 * around if we cannot get a queue's lock (which most likely is because 938 * someone else is queueing new packets and so will also handle the clean 939 * up). Since control queues use immediate data exclusively we don't 940 * bother cleaning them up here. 941 * 942 * b) Replenishes Rx queues that have run out due to memory shortage. 943 * Normally new Rx buffers are added when existing ones are consumed but 944 * when out of memory a queue can become empty. We try to add only a few 945 * buffers here, the queue will be replenished fully as these new buffers 946 * are used up if memory shortage has subsided. 947 * 948 * c) Return coalesced response queue credits in case a response queue is 949 * starved. 950 * 951 * d) Ring doorbells for T304 tunnel queues since we have seen doorbell 952 * fifo overflows and the FW doesn't implement any recovery scheme yet. 953 */ 954 static void 955 sge_timer_cb(void *arg) 956 { 957 adapter_t *sc = arg; 958 if ((sc->flags & USING_MSIX) == 0) { 959 960 struct port_info *pi; 961 struct sge_qset *qs; 962 struct sge_txq *txq; 963 int i, j; 964 int reclaim_ofl, refill_rx; 965 966 if (sc->open_device_map == 0) 967 return; 968 969 for (i = 0; i < sc->params.nports; i++) { 970 pi = &sc->port[i]; 971 for (j = 0; j < pi->nqsets; j++) { 972 qs = &sc->sge.qs[pi->first_qset + j]; 973 txq = &qs->txq[0]; 974 reclaim_ofl = txq[TXQ_OFLD].processed - txq[TXQ_OFLD].cleaned; 975 refill_rx = ((qs->fl[0].credits < qs->fl[0].size) || 976 (qs->fl[1].credits < qs->fl[1].size)); 977 if (reclaim_ofl || refill_rx) { 978 taskqueue_enqueue(sc->tq, &pi->timer_reclaim_task); 979 break; 980 } 981 } 982 } 983 } 984 985 if (sc->params.nports > 2) { 986 int i; 987 988 for_each_port(sc, i) { 989 struct port_info *pi = &sc->port[i]; 990 991 t3_write_reg(sc, A_SG_KDOORBELL, 992 F_SELEGRCNTX | 993 (FW_TUNNEL_SGEEC_START + pi->first_qset)); 994 } 995 } 996 if (((sc->flags & USING_MSIX) == 0 || sc->params.nports > 2) && 997 sc->open_device_map != 0) 998 callout_reset(&sc->sge_timer_ch, TX_RECLAIM_PERIOD, sge_timer_cb, sc); 999 } 1000 1001 /* 1002 * This is meant to be a catch-all function to keep sge state private 1003 * to sge.c 1004 * 1005 */ 1006 int 1007 t3_sge_init_adapter(adapter_t *sc) 1008 { 1009 callout_init(&sc->sge_timer_ch, 1); 1010 callout_reset(&sc->sge_timer_ch, TX_RECLAIM_PERIOD, sge_timer_cb, sc); 1011 TASK_INIT(&sc->slow_intr_task, 0, sge_slow_intr_handler, sc); 1012 return (0); 1013 } 1014 1015 int 1016 t3_sge_reset_adapter(adapter_t *sc) 1017 { 1018 callout_reset(&sc->sge_timer_ch, TX_RECLAIM_PERIOD, sge_timer_cb, sc); 1019 return (0); 1020 } 1021 1022 int 1023 t3_sge_init_port(struct port_info *pi) 1024 { 1025 TASK_INIT(&pi->timer_reclaim_task, 0, sge_timer_reclaim, pi); 1026 return (0); 1027 } 1028 1029 /** 1030 * refill_rspq - replenish an SGE response queue 1031 * @adapter: the adapter 1032 * @q: the response queue to replenish 1033 * @credits: how many new responses to make available 1034 * 1035 * Replenishes a response queue by making the supplied number of responses 1036 * available to HW. 1037 */ 1038 static __inline void 1039 refill_rspq(adapter_t *sc, const struct sge_rspq *q, u_int credits) 1040 { 1041 1042 /* mbufs are allocated on demand when a rspq entry is processed. */ 1043 t3_write_reg(sc, A_SG_RSPQ_CREDIT_RETURN, 1044 V_RSPQ(q->cntxt_id) | V_CREDITS(credits)); 1045 } 1046 1047 static void 1048 sge_txq_reclaim_handler(void *arg, int ncount) 1049 { 1050 struct sge_qset *qs = arg; 1051 int i; 1052 1053 for (i = 0; i < 3; i++) 1054 reclaim_completed_tx(qs, 16, i); 1055 } 1056 1057 static void 1058 sge_timer_reclaim(void *arg, int ncount) 1059 { 1060 struct port_info *pi = arg; 1061 int i, nqsets = pi->nqsets; 1062 adapter_t *sc = pi->adapter; 1063 struct sge_qset *qs; 1064 struct mtx *lock; 1065 1066 KASSERT((sc->flags & USING_MSIX) == 0, 1067 ("can't call timer reclaim for msi-x")); 1068 1069 for (i = 0; i < nqsets; i++) { 1070 qs = &sc->sge.qs[pi->first_qset + i]; 1071 1072 reclaim_completed_tx(qs, 16, TXQ_OFLD); 1073 lock = (sc->flags & USING_MSIX) ? &qs->rspq.lock : 1074 &sc->sge.qs[0].rspq.lock; 1075 1076 if (mtx_trylock(lock)) { 1077 /* XXX currently assume that we are *NOT* polling */ 1078 uint32_t status = t3_read_reg(sc, A_SG_RSPQ_FL_STATUS); 1079 1080 if (qs->fl[0].credits < qs->fl[0].size - 16) 1081 __refill_fl(sc, &qs->fl[0]); 1082 if (qs->fl[1].credits < qs->fl[1].size - 16) 1083 __refill_fl(sc, &qs->fl[1]); 1084 1085 if (status & (1 << qs->rspq.cntxt_id)) { 1086 if (qs->rspq.credits) { 1087 refill_rspq(sc, &qs->rspq, 1); 1088 qs->rspq.credits--; 1089 t3_write_reg(sc, A_SG_RSPQ_FL_STATUS, 1090 1 << qs->rspq.cntxt_id); 1091 } 1092 } 1093 mtx_unlock(lock); 1094 } 1095 } 1096 } 1097 1098 /** 1099 * init_qset_cntxt - initialize an SGE queue set context info 1100 * @qs: the queue set 1101 * @id: the queue set id 1102 * 1103 * Initializes the TIDs and context ids for the queues of a queue set. 1104 */ 1105 static void 1106 init_qset_cntxt(struct sge_qset *qs, u_int id) 1107 { 1108 1109 qs->rspq.cntxt_id = id; 1110 qs->fl[0].cntxt_id = 2 * id; 1111 qs->fl[1].cntxt_id = 2 * id + 1; 1112 qs->txq[TXQ_ETH].cntxt_id = FW_TUNNEL_SGEEC_START + id; 1113 qs->txq[TXQ_ETH].token = FW_TUNNEL_TID_START + id; 1114 qs->txq[TXQ_OFLD].cntxt_id = FW_OFLD_SGEEC_START + id; 1115 qs->txq[TXQ_CTRL].cntxt_id = FW_CTRL_SGEEC_START + id; 1116 qs->txq[TXQ_CTRL].token = FW_CTRL_TID_START + id; 1117 1118 /* XXX: a sane limit is needed instead of INT_MAX */ 1119 mbufq_init(&qs->txq[TXQ_ETH].sendq, INT_MAX); 1120 mbufq_init(&qs->txq[TXQ_OFLD].sendq, INT_MAX); 1121 mbufq_init(&qs->txq[TXQ_CTRL].sendq, INT_MAX); 1122 } 1123 1124 1125 static void 1126 txq_prod(struct sge_txq *txq, unsigned int ndesc, struct txq_state *txqs) 1127 { 1128 txq->in_use += ndesc; 1129 /* 1130 * XXX we don't handle stopping of queue 1131 * presumably start handles this when we bump against the end 1132 */ 1133 txqs->gen = txq->gen; 1134 txq->unacked += ndesc; 1135 txqs->compl = (txq->unacked & 32) << (S_WR_COMPL - 5); 1136 txq->unacked &= 31; 1137 txqs->pidx = txq->pidx; 1138 txq->pidx += ndesc; 1139 #ifdef INVARIANTS 1140 if (((txqs->pidx > txq->cidx) && 1141 (txq->pidx < txqs->pidx) && 1142 (txq->pidx >= txq->cidx)) || 1143 ((txqs->pidx < txq->cidx) && 1144 (txq->pidx >= txq-> cidx)) || 1145 ((txqs->pidx < txq->cidx) && 1146 (txq->cidx < txqs->pidx))) 1147 panic("txqs->pidx=%d txq->pidx=%d txq->cidx=%d", 1148 txqs->pidx, txq->pidx, txq->cidx); 1149 #endif 1150 if (txq->pidx >= txq->size) { 1151 txq->pidx -= txq->size; 1152 txq->gen ^= 1; 1153 } 1154 1155 } 1156 1157 /** 1158 * calc_tx_descs - calculate the number of Tx descriptors for a packet 1159 * @m: the packet mbufs 1160 * @nsegs: the number of segments 1161 * 1162 * Returns the number of Tx descriptors needed for the given Ethernet 1163 * packet. Ethernet packets require addition of WR and CPL headers. 1164 */ 1165 static __inline unsigned int 1166 calc_tx_descs(const struct mbuf *m, int nsegs) 1167 { 1168 unsigned int flits; 1169 1170 if (m->m_pkthdr.len <= PIO_LEN) 1171 return 1; 1172 1173 flits = sgl_len(nsegs) + 2; 1174 if (m->m_pkthdr.csum_flags & CSUM_TSO) 1175 flits++; 1176 1177 return flits_to_desc(flits); 1178 } 1179 1180 /** 1181 * make_sgl - populate a scatter/gather list for a packet 1182 * @sgp: the SGL to populate 1183 * @segs: the packet dma segments 1184 * @nsegs: the number of segments 1185 * 1186 * Generates a scatter/gather list for the buffers that make up a packet 1187 * and returns the SGL size in 8-byte words. The caller must size the SGL 1188 * appropriately. 1189 */ 1190 static __inline void 1191 make_sgl(struct sg_ent *sgp, bus_dma_segment_t *segs, int nsegs) 1192 { 1193 int i, idx; 1194 1195 for (idx = 0, i = 0; i < nsegs; i++) { 1196 /* 1197 * firmware doesn't like empty segments 1198 */ 1199 if (segs[i].ds_len == 0) 1200 continue; 1201 if (i && idx == 0) 1202 ++sgp; 1203 1204 sgp->len[idx] = htobe32(segs[i].ds_len); 1205 sgp->addr[idx] = htobe64(segs[i].ds_addr); 1206 idx ^= 1; 1207 } 1208 1209 if (idx) { 1210 sgp->len[idx] = 0; 1211 sgp->addr[idx] = 0; 1212 } 1213 } 1214 1215 /** 1216 * check_ring_tx_db - check and potentially ring a Tx queue's doorbell 1217 * @adap: the adapter 1218 * @q: the Tx queue 1219 * 1220 * Ring the doorbell if a Tx queue is asleep. There is a natural race, 1221 * where the HW is going to sleep just after we checked, however, 1222 * then the interrupt handler will detect the outstanding TX packet 1223 * and ring the doorbell for us. 1224 * 1225 * When GTS is disabled we unconditionally ring the doorbell. 1226 */ 1227 static __inline void 1228 check_ring_tx_db(adapter_t *adap, struct sge_txq *q, int mustring) 1229 { 1230 #if USE_GTS 1231 clear_bit(TXQ_LAST_PKT_DB, &q->flags); 1232 if (test_and_set_bit(TXQ_RUNNING, &q->flags) == 0) { 1233 set_bit(TXQ_LAST_PKT_DB, &q->flags); 1234 #ifdef T3_TRACE 1235 T3_TRACE1(adap->tb[q->cntxt_id & 7], "doorbell Tx, cntxt %d", 1236 q->cntxt_id); 1237 #endif 1238 t3_write_reg(adap, A_SG_KDOORBELL, 1239 F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id)); 1240 } 1241 #else 1242 if (mustring || ++q->db_pending >= 32) { 1243 wmb(); /* write descriptors before telling HW */ 1244 t3_write_reg(adap, A_SG_KDOORBELL, 1245 F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id)); 1246 q->db_pending = 0; 1247 } 1248 #endif 1249 } 1250 1251 static __inline void 1252 wr_gen2(struct tx_desc *d, unsigned int gen) 1253 { 1254 #if SGE_NUM_GENBITS == 2 1255 d->flit[TX_DESC_FLITS - 1] = htobe64(gen); 1256 #endif 1257 } 1258 1259 /** 1260 * write_wr_hdr_sgl - write a WR header and, optionally, SGL 1261 * @ndesc: number of Tx descriptors spanned by the SGL 1262 * @txd: first Tx descriptor to be written 1263 * @txqs: txq state (generation and producer index) 1264 * @txq: the SGE Tx queue 1265 * @sgl: the SGL 1266 * @flits: number of flits to the start of the SGL in the first descriptor 1267 * @sgl_flits: the SGL size in flits 1268 * @wr_hi: top 32 bits of WR header based on WR type (big endian) 1269 * @wr_lo: low 32 bits of WR header based on WR type (big endian) 1270 * 1271 * Write a work request header and an associated SGL. If the SGL is 1272 * small enough to fit into one Tx descriptor it has already been written 1273 * and we just need to write the WR header. Otherwise we distribute the 1274 * SGL across the number of descriptors it spans. 1275 */ 1276 static void 1277 write_wr_hdr_sgl(unsigned int ndesc, struct tx_desc *txd, struct txq_state *txqs, 1278 const struct sge_txq *txq, const struct sg_ent *sgl, unsigned int flits, 1279 unsigned int sgl_flits, unsigned int wr_hi, unsigned int wr_lo) 1280 { 1281 1282 struct work_request_hdr *wrp = (struct work_request_hdr *)txd; 1283 struct tx_sw_desc *txsd = &txq->sdesc[txqs->pidx]; 1284 1285 if (__predict_true(ndesc == 1)) { 1286 set_wr_hdr(wrp, htonl(F_WR_SOP | F_WR_EOP | V_WR_DATATYPE(1) | 1287 V_WR_SGLSFLT(flits)) | wr_hi, 1288 htonl(V_WR_LEN(flits + sgl_flits) | V_WR_GEN(txqs->gen)) | 1289 wr_lo); 1290 1291 wr_gen2(txd, txqs->gen); 1292 1293 } else { 1294 unsigned int ogen = txqs->gen; 1295 const uint64_t *fp = (const uint64_t *)sgl; 1296 struct work_request_hdr *wp = wrp; 1297 1298 wrp->wrh_hi = htonl(F_WR_SOP | V_WR_DATATYPE(1) | 1299 V_WR_SGLSFLT(flits)) | wr_hi; 1300 1301 while (sgl_flits) { 1302 unsigned int avail = WR_FLITS - flits; 1303 1304 if (avail > sgl_flits) 1305 avail = sgl_flits; 1306 memcpy(&txd->flit[flits], fp, avail * sizeof(*fp)); 1307 sgl_flits -= avail; 1308 ndesc--; 1309 if (!sgl_flits) 1310 break; 1311 1312 fp += avail; 1313 txd++; 1314 txsd++; 1315 if (++txqs->pidx == txq->size) { 1316 txqs->pidx = 0; 1317 txqs->gen ^= 1; 1318 txd = txq->desc; 1319 txsd = txq->sdesc; 1320 } 1321 1322 /* 1323 * when the head of the mbuf chain 1324 * is freed all clusters will be freed 1325 * with it 1326 */ 1327 wrp = (struct work_request_hdr *)txd; 1328 wrp->wrh_hi = htonl(V_WR_DATATYPE(1) | 1329 V_WR_SGLSFLT(1)) | wr_hi; 1330 wrp->wrh_lo = htonl(V_WR_LEN(min(WR_FLITS, 1331 sgl_flits + 1)) | 1332 V_WR_GEN(txqs->gen)) | wr_lo; 1333 wr_gen2(txd, txqs->gen); 1334 flits = 1; 1335 } 1336 wrp->wrh_hi |= htonl(F_WR_EOP); 1337 wmb(); 1338 wp->wrh_lo = htonl(V_WR_LEN(WR_FLITS) | V_WR_GEN(ogen)) | wr_lo; 1339 wr_gen2((struct tx_desc *)wp, ogen); 1340 } 1341 } 1342 1343 /* sizeof(*eh) + sizeof(*ip) + sizeof(*tcp) */ 1344 #define TCPPKTHDRSIZE (ETHER_HDR_LEN + 20 + 20) 1345 1346 #define GET_VTAG(cntrl, m) \ 1347 do { \ 1348 if ((m)->m_flags & M_VLANTAG) \ 1349 cntrl |= F_TXPKT_VLAN_VLD | V_TXPKT_VLAN((m)->m_pkthdr.ether_vtag); \ 1350 } while (0) 1351 1352 static int 1353 t3_encap(struct sge_qset *qs, struct mbuf **m) 1354 { 1355 adapter_t *sc; 1356 struct mbuf *m0; 1357 struct sge_txq *txq; 1358 struct txq_state txqs; 1359 struct port_info *pi; 1360 unsigned int ndesc, flits, cntrl, mlen; 1361 int err, nsegs, tso_info = 0; 1362 1363 struct work_request_hdr *wrp; 1364 struct tx_sw_desc *txsd; 1365 struct sg_ent *sgp, *sgl; 1366 uint32_t wr_hi, wr_lo, sgl_flits; 1367 bus_dma_segment_t segs[TX_MAX_SEGS]; 1368 1369 struct tx_desc *txd; 1370 1371 pi = qs->port; 1372 sc = pi->adapter; 1373 txq = &qs->txq[TXQ_ETH]; 1374 txd = &txq->desc[txq->pidx]; 1375 txsd = &txq->sdesc[txq->pidx]; 1376 sgl = txq->txq_sgl; 1377 1378 prefetch(txd); 1379 m0 = *m; 1380 1381 mtx_assert(&qs->lock, MA_OWNED); 1382 cntrl = V_TXPKT_INTF(pi->txpkt_intf); 1383 KASSERT(m0->m_flags & M_PKTHDR, ("not packet header\n")); 1384 1385 if (m0->m_nextpkt == NULL && m0->m_next != NULL && 1386 m0->m_pkthdr.csum_flags & (CSUM_TSO)) 1387 tso_info = V_LSO_MSS(m0->m_pkthdr.tso_segsz); 1388 1389 if (m0->m_nextpkt != NULL) { 1390 busdma_map_sg_vec(txq->entry_tag, txsd->map, m0, segs, &nsegs); 1391 ndesc = 1; 1392 mlen = 0; 1393 } else { 1394 if ((err = busdma_map_sg_collapse(txq->entry_tag, txsd->map, 1395 &m0, segs, &nsegs))) { 1396 if (cxgb_debug) 1397 printf("failed ... err=%d\n", err); 1398 return (err); 1399 } 1400 mlen = m0->m_pkthdr.len; 1401 ndesc = calc_tx_descs(m0, nsegs); 1402 } 1403 txq_prod(txq, ndesc, &txqs); 1404 1405 KASSERT(m0->m_pkthdr.len, ("empty packet nsegs=%d", nsegs)); 1406 txsd->m = m0; 1407 1408 if (m0->m_nextpkt != NULL) { 1409 struct cpl_tx_pkt_batch *cpl_batch = (struct cpl_tx_pkt_batch *)txd; 1410 int i, fidx; 1411 1412 if (nsegs > 7) 1413 panic("trying to coalesce %d packets in to one WR", nsegs); 1414 txq->txq_coalesced += nsegs; 1415 wrp = (struct work_request_hdr *)txd; 1416 flits = nsegs*2 + 1; 1417 1418 for (fidx = 1, i = 0; i < nsegs; i++, fidx += 2) { 1419 struct cpl_tx_pkt_batch_entry *cbe; 1420 uint64_t flit; 1421 uint32_t *hflit = (uint32_t *)&flit; 1422 int cflags = m0->m_pkthdr.csum_flags; 1423 1424 cntrl = V_TXPKT_INTF(pi->txpkt_intf); 1425 GET_VTAG(cntrl, m0); 1426 cntrl |= V_TXPKT_OPCODE(CPL_TX_PKT); 1427 if (__predict_false(!(cflags & CSUM_IP))) 1428 cntrl |= F_TXPKT_IPCSUM_DIS; 1429 if (__predict_false(!(cflags & (CSUM_TCP | CSUM_UDP | 1430 CSUM_UDP_IPV6 | CSUM_TCP_IPV6)))) 1431 cntrl |= F_TXPKT_L4CSUM_DIS; 1432 1433 hflit[0] = htonl(cntrl); 1434 hflit[1] = htonl(segs[i].ds_len | 0x80000000); 1435 flit |= htobe64(1 << 24); 1436 cbe = &cpl_batch->pkt_entry[i]; 1437 cbe->cntrl = hflit[0]; 1438 cbe->len = hflit[1]; 1439 cbe->addr = htobe64(segs[i].ds_addr); 1440 } 1441 1442 wr_hi = htonl(F_WR_SOP | F_WR_EOP | V_WR_DATATYPE(1) | 1443 V_WR_SGLSFLT(flits)) | 1444 htonl(V_WR_OP(FW_WROPCODE_TUNNEL_TX_PKT) | txqs.compl); 1445 wr_lo = htonl(V_WR_LEN(flits) | 1446 V_WR_GEN(txqs.gen)) | htonl(V_WR_TID(txq->token)); 1447 set_wr_hdr(wrp, wr_hi, wr_lo); 1448 wmb(); 1449 ETHER_BPF_MTAP(pi->ifp, m0); 1450 wr_gen2(txd, txqs.gen); 1451 check_ring_tx_db(sc, txq, 0); 1452 return (0); 1453 } else if (tso_info) { 1454 uint16_t eth_type; 1455 struct cpl_tx_pkt_lso *hdr = (struct cpl_tx_pkt_lso *)txd; 1456 struct ether_header *eh; 1457 void *l3hdr; 1458 struct tcphdr *tcp; 1459 1460 txd->flit[2] = 0; 1461 GET_VTAG(cntrl, m0); 1462 cntrl |= V_TXPKT_OPCODE(CPL_TX_PKT_LSO); 1463 hdr->cntrl = htonl(cntrl); 1464 hdr->len = htonl(mlen | 0x80000000); 1465 1466 if (__predict_false(mlen < TCPPKTHDRSIZE)) { 1467 printf("mbuf=%p,len=%d,tso_segsz=%d,csum_flags=%b,flags=%#x", 1468 m0, mlen, m0->m_pkthdr.tso_segsz, 1469 (int)m0->m_pkthdr.csum_flags, CSUM_BITS, m0->m_flags); 1470 panic("tx tso packet too small"); 1471 } 1472 1473 /* Make sure that ether, ip, tcp headers are all in m0 */ 1474 if (__predict_false(m0->m_len < TCPPKTHDRSIZE)) { 1475 m0 = m_pullup(m0, TCPPKTHDRSIZE); 1476 if (__predict_false(m0 == NULL)) { 1477 /* XXX panic probably an overreaction */ 1478 panic("couldn't fit header into mbuf"); 1479 } 1480 } 1481 1482 eh = mtod(m0, struct ether_header *); 1483 eth_type = eh->ether_type; 1484 if (eth_type == htons(ETHERTYPE_VLAN)) { 1485 struct ether_vlan_header *evh = (void *)eh; 1486 1487 tso_info |= V_LSO_ETH_TYPE(CPL_ETH_II_VLAN); 1488 l3hdr = evh + 1; 1489 eth_type = evh->evl_proto; 1490 } else { 1491 tso_info |= V_LSO_ETH_TYPE(CPL_ETH_II); 1492 l3hdr = eh + 1; 1493 } 1494 1495 if (eth_type == htons(ETHERTYPE_IP)) { 1496 struct ip *ip = l3hdr; 1497 1498 tso_info |= V_LSO_IPHDR_WORDS(ip->ip_hl); 1499 tcp = (struct tcphdr *)(ip + 1); 1500 } else if (eth_type == htons(ETHERTYPE_IPV6)) { 1501 struct ip6_hdr *ip6 = l3hdr; 1502 1503 KASSERT(ip6->ip6_nxt == IPPROTO_TCP, 1504 ("%s: CSUM_TSO with ip6_nxt %d", 1505 __func__, ip6->ip6_nxt)); 1506 1507 tso_info |= F_LSO_IPV6; 1508 tso_info |= V_LSO_IPHDR_WORDS(sizeof(*ip6) >> 2); 1509 tcp = (struct tcphdr *)(ip6 + 1); 1510 } else 1511 panic("%s: CSUM_TSO but neither ip nor ip6", __func__); 1512 1513 tso_info |= V_LSO_TCPHDR_WORDS(tcp->th_off); 1514 hdr->lso_info = htonl(tso_info); 1515 1516 if (__predict_false(mlen <= PIO_LEN)) { 1517 /* 1518 * pkt not undersized but fits in PIO_LEN 1519 * Indicates a TSO bug at the higher levels. 1520 */ 1521 txsd->m = NULL; 1522 m_copydata(m0, 0, mlen, (caddr_t)&txd->flit[3]); 1523 flits = (mlen + 7) / 8 + 3; 1524 wr_hi = htonl(V_WR_BCNTLFLT(mlen & 7) | 1525 V_WR_OP(FW_WROPCODE_TUNNEL_TX_PKT) | 1526 F_WR_SOP | F_WR_EOP | txqs.compl); 1527 wr_lo = htonl(V_WR_LEN(flits) | 1528 V_WR_GEN(txqs.gen) | V_WR_TID(txq->token)); 1529 set_wr_hdr(&hdr->wr, wr_hi, wr_lo); 1530 wmb(); 1531 ETHER_BPF_MTAP(pi->ifp, m0); 1532 wr_gen2(txd, txqs.gen); 1533 check_ring_tx_db(sc, txq, 0); 1534 m_freem(m0); 1535 return (0); 1536 } 1537 flits = 3; 1538 } else { 1539 struct cpl_tx_pkt *cpl = (struct cpl_tx_pkt *)txd; 1540 1541 GET_VTAG(cntrl, m0); 1542 cntrl |= V_TXPKT_OPCODE(CPL_TX_PKT); 1543 if (__predict_false(!(m0->m_pkthdr.csum_flags & CSUM_IP))) 1544 cntrl |= F_TXPKT_IPCSUM_DIS; 1545 if (__predict_false(!(m0->m_pkthdr.csum_flags & (CSUM_TCP | 1546 CSUM_UDP | CSUM_UDP_IPV6 | CSUM_TCP_IPV6)))) 1547 cntrl |= F_TXPKT_L4CSUM_DIS; 1548 cpl->cntrl = htonl(cntrl); 1549 cpl->len = htonl(mlen | 0x80000000); 1550 1551 if (mlen <= PIO_LEN) { 1552 txsd->m = NULL; 1553 m_copydata(m0, 0, mlen, (caddr_t)&txd->flit[2]); 1554 flits = (mlen + 7) / 8 + 2; 1555 1556 wr_hi = htonl(V_WR_BCNTLFLT(mlen & 7) | 1557 V_WR_OP(FW_WROPCODE_TUNNEL_TX_PKT) | 1558 F_WR_SOP | F_WR_EOP | txqs.compl); 1559 wr_lo = htonl(V_WR_LEN(flits) | 1560 V_WR_GEN(txqs.gen) | V_WR_TID(txq->token)); 1561 set_wr_hdr(&cpl->wr, wr_hi, wr_lo); 1562 wmb(); 1563 ETHER_BPF_MTAP(pi->ifp, m0); 1564 wr_gen2(txd, txqs.gen); 1565 check_ring_tx_db(sc, txq, 0); 1566 m_freem(m0); 1567 return (0); 1568 } 1569 flits = 2; 1570 } 1571 wrp = (struct work_request_hdr *)txd; 1572 sgp = (ndesc == 1) ? (struct sg_ent *)&txd->flit[flits] : sgl; 1573 make_sgl(sgp, segs, nsegs); 1574 1575 sgl_flits = sgl_len(nsegs); 1576 1577 ETHER_BPF_MTAP(pi->ifp, m0); 1578 1579 KASSERT(ndesc <= 4, ("ndesc too large %d", ndesc)); 1580 wr_hi = htonl(V_WR_OP(FW_WROPCODE_TUNNEL_TX_PKT) | txqs.compl); 1581 wr_lo = htonl(V_WR_TID(txq->token)); 1582 write_wr_hdr_sgl(ndesc, txd, &txqs, txq, sgl, flits, 1583 sgl_flits, wr_hi, wr_lo); 1584 check_ring_tx_db(sc, txq, 0); 1585 1586 return (0); 1587 } 1588 1589 void 1590 cxgb_tx_watchdog(void *arg) 1591 { 1592 struct sge_qset *qs = arg; 1593 struct sge_txq *txq = &qs->txq[TXQ_ETH]; 1594 1595 if (qs->coalescing != 0 && 1596 (txq->in_use <= cxgb_tx_coalesce_enable_stop) && 1597 TXQ_RING_EMPTY(qs)) 1598 qs->coalescing = 0; 1599 else if (qs->coalescing == 0 && 1600 (txq->in_use >= cxgb_tx_coalesce_enable_start)) 1601 qs->coalescing = 1; 1602 if (TXQ_TRYLOCK(qs)) { 1603 qs->qs_flags |= QS_FLUSHING; 1604 cxgb_start_locked(qs); 1605 qs->qs_flags &= ~QS_FLUSHING; 1606 TXQ_UNLOCK(qs); 1607 } 1608 if (qs->port->ifp->if_drv_flags & IFF_DRV_RUNNING) 1609 callout_reset_on(&txq->txq_watchdog, hz/4, cxgb_tx_watchdog, 1610 qs, txq->txq_watchdog.c_cpu); 1611 } 1612 1613 static void 1614 cxgb_tx_timeout(void *arg) 1615 { 1616 struct sge_qset *qs = arg; 1617 struct sge_txq *txq = &qs->txq[TXQ_ETH]; 1618 1619 if (qs->coalescing == 0 && (txq->in_use >= (txq->size>>3))) 1620 qs->coalescing = 1; 1621 if (TXQ_TRYLOCK(qs)) { 1622 qs->qs_flags |= QS_TIMEOUT; 1623 cxgb_start_locked(qs); 1624 qs->qs_flags &= ~QS_TIMEOUT; 1625 TXQ_UNLOCK(qs); 1626 } 1627 } 1628 1629 static void 1630 cxgb_start_locked(struct sge_qset *qs) 1631 { 1632 struct mbuf *m_head = NULL; 1633 struct sge_txq *txq = &qs->txq[TXQ_ETH]; 1634 struct port_info *pi = qs->port; 1635 struct ifnet *ifp = pi->ifp; 1636 1637 if (qs->qs_flags & (QS_FLUSHING|QS_TIMEOUT)) 1638 reclaim_completed_tx(qs, 0, TXQ_ETH); 1639 1640 if (!pi->link_config.link_ok) { 1641 TXQ_RING_FLUSH(qs); 1642 return; 1643 } 1644 TXQ_LOCK_ASSERT(qs); 1645 while (!TXQ_RING_EMPTY(qs) && (ifp->if_drv_flags & IFF_DRV_RUNNING) && 1646 pi->link_config.link_ok) { 1647 reclaim_completed_tx(qs, cxgb_tx_reclaim_threshold, TXQ_ETH); 1648 1649 if (txq->size - txq->in_use <= TX_MAX_DESC) 1650 break; 1651 1652 if ((m_head = cxgb_dequeue(qs)) == NULL) 1653 break; 1654 /* 1655 * Encapsulation can modify our pointer, and or make it 1656 * NULL on failure. In that event, we can't requeue. 1657 */ 1658 if (t3_encap(qs, &m_head) || m_head == NULL) 1659 break; 1660 1661 m_head = NULL; 1662 } 1663 1664 if (txq->db_pending) 1665 check_ring_tx_db(pi->adapter, txq, 1); 1666 1667 if (!TXQ_RING_EMPTY(qs) && callout_pending(&txq->txq_timer) == 0 && 1668 pi->link_config.link_ok) 1669 callout_reset_on(&txq->txq_timer, 1, cxgb_tx_timeout, 1670 qs, txq->txq_timer.c_cpu); 1671 if (m_head != NULL) 1672 m_freem(m_head); 1673 } 1674 1675 static int 1676 cxgb_transmit_locked(struct ifnet *ifp, struct sge_qset *qs, struct mbuf *m) 1677 { 1678 struct port_info *pi = qs->port; 1679 struct sge_txq *txq = &qs->txq[TXQ_ETH]; 1680 struct buf_ring *br = txq->txq_mr; 1681 int error, avail; 1682 1683 avail = txq->size - txq->in_use; 1684 TXQ_LOCK_ASSERT(qs); 1685 1686 /* 1687 * We can only do a direct transmit if the following are true: 1688 * - we aren't coalescing (ring < 3/4 full) 1689 * - the link is up -- checked in caller 1690 * - there are no packets enqueued already 1691 * - there is space in hardware transmit queue 1692 */ 1693 if (check_pkt_coalesce(qs) == 0 && 1694 !TXQ_RING_NEEDS_ENQUEUE(qs) && avail > TX_MAX_DESC) { 1695 if (t3_encap(qs, &m)) { 1696 if (m != NULL && 1697 (error = drbr_enqueue(ifp, br, m)) != 0) 1698 return (error); 1699 } else { 1700 if (txq->db_pending) 1701 check_ring_tx_db(pi->adapter, txq, 1); 1702 1703 /* 1704 * We've bypassed the buf ring so we need to update 1705 * the stats directly 1706 */ 1707 txq->txq_direct_packets++; 1708 txq->txq_direct_bytes += m->m_pkthdr.len; 1709 } 1710 } else if ((error = drbr_enqueue(ifp, br, m)) != 0) 1711 return (error); 1712 1713 reclaim_completed_tx(qs, cxgb_tx_reclaim_threshold, TXQ_ETH); 1714 if (!TXQ_RING_EMPTY(qs) && pi->link_config.link_ok && 1715 (!check_pkt_coalesce(qs) || (drbr_inuse(ifp, br) >= 7))) 1716 cxgb_start_locked(qs); 1717 else if (!TXQ_RING_EMPTY(qs) && !callout_pending(&txq->txq_timer)) 1718 callout_reset_on(&txq->txq_timer, 1, cxgb_tx_timeout, 1719 qs, txq->txq_timer.c_cpu); 1720 return (0); 1721 } 1722 1723 int 1724 cxgb_transmit(struct ifnet *ifp, struct mbuf *m) 1725 { 1726 struct sge_qset *qs; 1727 struct port_info *pi = ifp->if_softc; 1728 int error, qidx = pi->first_qset; 1729 1730 if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0 1731 ||(!pi->link_config.link_ok)) { 1732 m_freem(m); 1733 return (0); 1734 } 1735 1736 /* check if flowid is set */ 1737 if (M_HASHTYPE_GET(m) != M_HASHTYPE_NONE) 1738 qidx = (m->m_pkthdr.flowid % pi->nqsets) + pi->first_qset; 1739 1740 qs = &pi->adapter->sge.qs[qidx]; 1741 1742 if (TXQ_TRYLOCK(qs)) { 1743 /* XXX running */ 1744 error = cxgb_transmit_locked(ifp, qs, m); 1745 TXQ_UNLOCK(qs); 1746 } else 1747 error = drbr_enqueue(ifp, qs->txq[TXQ_ETH].txq_mr, m); 1748 return (error); 1749 } 1750 1751 void 1752 cxgb_qflush(struct ifnet *ifp) 1753 { 1754 /* 1755 * flush any enqueued mbufs in the buf_rings 1756 * and in the transmit queues 1757 * no-op for now 1758 */ 1759 return; 1760 } 1761 1762 /** 1763 * write_imm - write a packet into a Tx descriptor as immediate data 1764 * @d: the Tx descriptor to write 1765 * @m: the packet 1766 * @len: the length of packet data to write as immediate data 1767 * @gen: the generation bit value to write 1768 * 1769 * Writes a packet as immediate data into a Tx descriptor. The packet 1770 * contains a work request at its beginning. We must write the packet 1771 * carefully so the SGE doesn't read accidentally before it's written in 1772 * its entirety. 1773 */ 1774 static __inline void 1775 write_imm(struct tx_desc *d, caddr_t src, 1776 unsigned int len, unsigned int gen) 1777 { 1778 struct work_request_hdr *from = (struct work_request_hdr *)src; 1779 struct work_request_hdr *to = (struct work_request_hdr *)d; 1780 uint32_t wr_hi, wr_lo; 1781 1782 KASSERT(len <= WR_LEN && len >= sizeof(*from), 1783 ("%s: invalid len %d", __func__, len)); 1784 1785 memcpy(&to[1], &from[1], len - sizeof(*from)); 1786 wr_hi = from->wrh_hi | htonl(F_WR_SOP | F_WR_EOP | 1787 V_WR_BCNTLFLT(len & 7)); 1788 wr_lo = from->wrh_lo | htonl(V_WR_GEN(gen) | V_WR_LEN((len + 7) / 8)); 1789 set_wr_hdr(to, wr_hi, wr_lo); 1790 wmb(); 1791 wr_gen2(d, gen); 1792 } 1793 1794 /** 1795 * check_desc_avail - check descriptor availability on a send queue 1796 * @adap: the adapter 1797 * @q: the TX queue 1798 * @m: the packet needing the descriptors 1799 * @ndesc: the number of Tx descriptors needed 1800 * @qid: the Tx queue number in its queue set (TXQ_OFLD or TXQ_CTRL) 1801 * 1802 * Checks if the requested number of Tx descriptors is available on an 1803 * SGE send queue. If the queue is already suspended or not enough 1804 * descriptors are available the packet is queued for later transmission. 1805 * Must be called with the Tx queue locked. 1806 * 1807 * Returns 0 if enough descriptors are available, 1 if there aren't 1808 * enough descriptors and the packet has been queued, and 2 if the caller 1809 * needs to retry because there weren't enough descriptors at the 1810 * beginning of the call but some freed up in the mean time. 1811 */ 1812 static __inline int 1813 check_desc_avail(adapter_t *adap, struct sge_txq *q, 1814 struct mbuf *m, unsigned int ndesc, 1815 unsigned int qid) 1816 { 1817 /* 1818 * XXX We currently only use this for checking the control queue 1819 * the control queue is only used for binding qsets which happens 1820 * at init time so we are guaranteed enough descriptors 1821 */ 1822 if (__predict_false(mbufq_len(&q->sendq))) { 1823 addq_exit: (void )mbufq_enqueue(&q->sendq, m); 1824 return 1; 1825 } 1826 if (__predict_false(q->size - q->in_use < ndesc)) { 1827 1828 struct sge_qset *qs = txq_to_qset(q, qid); 1829 1830 setbit(&qs->txq_stopped, qid); 1831 if (should_restart_tx(q) && 1832 test_and_clear_bit(qid, &qs->txq_stopped)) 1833 return 2; 1834 1835 q->stops++; 1836 goto addq_exit; 1837 } 1838 return 0; 1839 } 1840 1841 1842 /** 1843 * reclaim_completed_tx_imm - reclaim completed control-queue Tx descs 1844 * @q: the SGE control Tx queue 1845 * 1846 * This is a variant of reclaim_completed_tx() that is used for Tx queues 1847 * that send only immediate data (presently just the control queues) and 1848 * thus do not have any mbufs 1849 */ 1850 static __inline void 1851 reclaim_completed_tx_imm(struct sge_txq *q) 1852 { 1853 unsigned int reclaim = q->processed - q->cleaned; 1854 1855 q->in_use -= reclaim; 1856 q->cleaned += reclaim; 1857 } 1858 1859 /** 1860 * ctrl_xmit - send a packet through an SGE control Tx queue 1861 * @adap: the adapter 1862 * @q: the control queue 1863 * @m: the packet 1864 * 1865 * Send a packet through an SGE control Tx queue. Packets sent through 1866 * a control queue must fit entirely as immediate data in a single Tx 1867 * descriptor and have no page fragments. 1868 */ 1869 static int 1870 ctrl_xmit(adapter_t *adap, struct sge_qset *qs, struct mbuf *m) 1871 { 1872 int ret; 1873 struct work_request_hdr *wrp = mtod(m, struct work_request_hdr *); 1874 struct sge_txq *q = &qs->txq[TXQ_CTRL]; 1875 1876 KASSERT(m->m_len <= WR_LEN, ("%s: bad tx data", __func__)); 1877 1878 wrp->wrh_hi |= htonl(F_WR_SOP | F_WR_EOP); 1879 wrp->wrh_lo = htonl(V_WR_TID(q->token)); 1880 1881 TXQ_LOCK(qs); 1882 again: reclaim_completed_tx_imm(q); 1883 1884 ret = check_desc_avail(adap, q, m, 1, TXQ_CTRL); 1885 if (__predict_false(ret)) { 1886 if (ret == 1) { 1887 TXQ_UNLOCK(qs); 1888 return (ENOSPC); 1889 } 1890 goto again; 1891 } 1892 write_imm(&q->desc[q->pidx], m->m_data, m->m_len, q->gen); 1893 1894 q->in_use++; 1895 if (++q->pidx >= q->size) { 1896 q->pidx = 0; 1897 q->gen ^= 1; 1898 } 1899 TXQ_UNLOCK(qs); 1900 wmb(); 1901 t3_write_reg(adap, A_SG_KDOORBELL, 1902 F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id)); 1903 1904 m_free(m); 1905 return (0); 1906 } 1907 1908 1909 /** 1910 * restart_ctrlq - restart a suspended control queue 1911 * @qs: the queue set cotaining the control queue 1912 * 1913 * Resumes transmission on a suspended Tx control queue. 1914 */ 1915 static void 1916 restart_ctrlq(void *data, int npending) 1917 { 1918 struct mbuf *m; 1919 struct sge_qset *qs = (struct sge_qset *)data; 1920 struct sge_txq *q = &qs->txq[TXQ_CTRL]; 1921 adapter_t *adap = qs->port->adapter; 1922 1923 TXQ_LOCK(qs); 1924 again: reclaim_completed_tx_imm(q); 1925 1926 while (q->in_use < q->size && 1927 (m = mbufq_dequeue(&q->sendq)) != NULL) { 1928 1929 write_imm(&q->desc[q->pidx], m->m_data, m->m_len, q->gen); 1930 m_free(m); 1931 1932 if (++q->pidx >= q->size) { 1933 q->pidx = 0; 1934 q->gen ^= 1; 1935 } 1936 q->in_use++; 1937 } 1938 if (mbufq_len(&q->sendq)) { 1939 setbit(&qs->txq_stopped, TXQ_CTRL); 1940 1941 if (should_restart_tx(q) && 1942 test_and_clear_bit(TXQ_CTRL, &qs->txq_stopped)) 1943 goto again; 1944 q->stops++; 1945 } 1946 TXQ_UNLOCK(qs); 1947 t3_write_reg(adap, A_SG_KDOORBELL, 1948 F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id)); 1949 } 1950 1951 1952 /* 1953 * Send a management message through control queue 0 1954 */ 1955 int 1956 t3_mgmt_tx(struct adapter *adap, struct mbuf *m) 1957 { 1958 return ctrl_xmit(adap, &adap->sge.qs[0], m); 1959 } 1960 1961 /** 1962 * free_qset - free the resources of an SGE queue set 1963 * @sc: the controller owning the queue set 1964 * @q: the queue set 1965 * 1966 * Release the HW and SW resources associated with an SGE queue set, such 1967 * as HW contexts, packet buffers, and descriptor rings. Traffic to the 1968 * queue set must be quiesced prior to calling this. 1969 */ 1970 static void 1971 t3_free_qset(adapter_t *sc, struct sge_qset *q) 1972 { 1973 int i; 1974 1975 reclaim_completed_tx(q, 0, TXQ_ETH); 1976 if (q->txq[TXQ_ETH].txq_mr != NULL) 1977 buf_ring_free(q->txq[TXQ_ETH].txq_mr, M_DEVBUF); 1978 if (q->txq[TXQ_ETH].txq_ifq != NULL) { 1979 ifq_delete(q->txq[TXQ_ETH].txq_ifq); 1980 free(q->txq[TXQ_ETH].txq_ifq, M_DEVBUF); 1981 } 1982 1983 for (i = 0; i < SGE_RXQ_PER_SET; ++i) { 1984 if (q->fl[i].desc) { 1985 mtx_lock_spin(&sc->sge.reg_lock); 1986 t3_sge_disable_fl(sc, q->fl[i].cntxt_id); 1987 mtx_unlock_spin(&sc->sge.reg_lock); 1988 bus_dmamap_unload(q->fl[i].desc_tag, q->fl[i].desc_map); 1989 bus_dmamem_free(q->fl[i].desc_tag, q->fl[i].desc, 1990 q->fl[i].desc_map); 1991 bus_dma_tag_destroy(q->fl[i].desc_tag); 1992 bus_dma_tag_destroy(q->fl[i].entry_tag); 1993 } 1994 if (q->fl[i].sdesc) { 1995 free_rx_bufs(sc, &q->fl[i]); 1996 free(q->fl[i].sdesc, M_DEVBUF); 1997 } 1998 } 1999 2000 mtx_unlock(&q->lock); 2001 MTX_DESTROY(&q->lock); 2002 for (i = 0; i < SGE_TXQ_PER_SET; i++) { 2003 if (q->txq[i].desc) { 2004 mtx_lock_spin(&sc->sge.reg_lock); 2005 t3_sge_enable_ecntxt(sc, q->txq[i].cntxt_id, 0); 2006 mtx_unlock_spin(&sc->sge.reg_lock); 2007 bus_dmamap_unload(q->txq[i].desc_tag, 2008 q->txq[i].desc_map); 2009 bus_dmamem_free(q->txq[i].desc_tag, q->txq[i].desc, 2010 q->txq[i].desc_map); 2011 bus_dma_tag_destroy(q->txq[i].desc_tag); 2012 bus_dma_tag_destroy(q->txq[i].entry_tag); 2013 } 2014 if (q->txq[i].sdesc) { 2015 free(q->txq[i].sdesc, M_DEVBUF); 2016 } 2017 } 2018 2019 if (q->rspq.desc) { 2020 mtx_lock_spin(&sc->sge.reg_lock); 2021 t3_sge_disable_rspcntxt(sc, q->rspq.cntxt_id); 2022 mtx_unlock_spin(&sc->sge.reg_lock); 2023 2024 bus_dmamap_unload(q->rspq.desc_tag, q->rspq.desc_map); 2025 bus_dmamem_free(q->rspq.desc_tag, q->rspq.desc, 2026 q->rspq.desc_map); 2027 bus_dma_tag_destroy(q->rspq.desc_tag); 2028 MTX_DESTROY(&q->rspq.lock); 2029 } 2030 2031 #if defined(INET6) || defined(INET) 2032 tcp_lro_free(&q->lro.ctrl); 2033 #endif 2034 2035 bzero(q, sizeof(*q)); 2036 } 2037 2038 /** 2039 * t3_free_sge_resources - free SGE resources 2040 * @sc: the adapter softc 2041 * 2042 * Frees resources used by the SGE queue sets. 2043 */ 2044 void 2045 t3_free_sge_resources(adapter_t *sc, int nqsets) 2046 { 2047 int i; 2048 2049 for (i = 0; i < nqsets; ++i) { 2050 TXQ_LOCK(&sc->sge.qs[i]); 2051 t3_free_qset(sc, &sc->sge.qs[i]); 2052 } 2053 } 2054 2055 /** 2056 * t3_sge_start - enable SGE 2057 * @sc: the controller softc 2058 * 2059 * Enables the SGE for DMAs. This is the last step in starting packet 2060 * transfers. 2061 */ 2062 void 2063 t3_sge_start(adapter_t *sc) 2064 { 2065 t3_set_reg_field(sc, A_SG_CONTROL, F_GLOBALENABLE, F_GLOBALENABLE); 2066 } 2067 2068 /** 2069 * t3_sge_stop - disable SGE operation 2070 * @sc: the adapter 2071 * 2072 * Disables the DMA engine. This can be called in emeregencies (e.g., 2073 * from error interrupts) or from normal process context. In the latter 2074 * case it also disables any pending queue restart tasklets. Note that 2075 * if it is called in interrupt context it cannot disable the restart 2076 * tasklets as it cannot wait, however the tasklets will have no effect 2077 * since the doorbells are disabled and the driver will call this again 2078 * later from process context, at which time the tasklets will be stopped 2079 * if they are still running. 2080 */ 2081 void 2082 t3_sge_stop(adapter_t *sc) 2083 { 2084 int i, nqsets; 2085 2086 t3_set_reg_field(sc, A_SG_CONTROL, F_GLOBALENABLE, 0); 2087 2088 if (sc->tq == NULL) 2089 return; 2090 2091 for (nqsets = i = 0; i < (sc)->params.nports; i++) 2092 nqsets += sc->port[i].nqsets; 2093 #ifdef notyet 2094 /* 2095 * 2096 * XXX 2097 */ 2098 for (i = 0; i < nqsets; ++i) { 2099 struct sge_qset *qs = &sc->sge.qs[i]; 2100 2101 taskqueue_drain(sc->tq, &qs->txq[TXQ_OFLD].qresume_task); 2102 taskqueue_drain(sc->tq, &qs->txq[TXQ_CTRL].qresume_task); 2103 } 2104 #endif 2105 } 2106 2107 /** 2108 * t3_free_tx_desc - reclaims Tx descriptors and their buffers 2109 * @adapter: the adapter 2110 * @q: the Tx queue to reclaim descriptors from 2111 * @reclaimable: the number of descriptors to reclaim 2112 * @m_vec_size: maximum number of buffers to reclaim 2113 * @desc_reclaimed: returns the number of descriptors reclaimed 2114 * 2115 * Reclaims Tx descriptors from an SGE Tx queue and frees the associated 2116 * Tx buffers. Called with the Tx queue lock held. 2117 * 2118 * Returns number of buffers of reclaimed 2119 */ 2120 void 2121 t3_free_tx_desc(struct sge_qset *qs, int reclaimable, int queue) 2122 { 2123 struct tx_sw_desc *txsd; 2124 unsigned int cidx, mask; 2125 struct sge_txq *q = &qs->txq[queue]; 2126 2127 #ifdef T3_TRACE 2128 T3_TRACE2(sc->tb[q->cntxt_id & 7], 2129 "reclaiming %u Tx descriptors at cidx %u", reclaimable, cidx); 2130 #endif 2131 cidx = q->cidx; 2132 mask = q->size - 1; 2133 txsd = &q->sdesc[cidx]; 2134 2135 mtx_assert(&qs->lock, MA_OWNED); 2136 while (reclaimable--) { 2137 prefetch(q->sdesc[(cidx + 1) & mask].m); 2138 prefetch(q->sdesc[(cidx + 2) & mask].m); 2139 2140 if (txsd->m != NULL) { 2141 if (txsd->flags & TX_SW_DESC_MAPPED) { 2142 bus_dmamap_unload(q->entry_tag, txsd->map); 2143 txsd->flags &= ~TX_SW_DESC_MAPPED; 2144 } 2145 m_freem_list(txsd->m); 2146 txsd->m = NULL; 2147 } else 2148 q->txq_skipped++; 2149 2150 ++txsd; 2151 if (++cidx == q->size) { 2152 cidx = 0; 2153 txsd = q->sdesc; 2154 } 2155 } 2156 q->cidx = cidx; 2157 2158 } 2159 2160 /** 2161 * is_new_response - check if a response is newly written 2162 * @r: the response descriptor 2163 * @q: the response queue 2164 * 2165 * Returns true if a response descriptor contains a yet unprocessed 2166 * response. 2167 */ 2168 static __inline int 2169 is_new_response(const struct rsp_desc *r, 2170 const struct sge_rspq *q) 2171 { 2172 return (r->intr_gen & F_RSPD_GEN2) == q->gen; 2173 } 2174 2175 #define RSPD_GTS_MASK (F_RSPD_TXQ0_GTS | F_RSPD_TXQ1_GTS) 2176 #define RSPD_CTRL_MASK (RSPD_GTS_MASK | \ 2177 V_RSPD_TXQ0_CR(M_RSPD_TXQ0_CR) | \ 2178 V_RSPD_TXQ1_CR(M_RSPD_TXQ1_CR) | \ 2179 V_RSPD_TXQ2_CR(M_RSPD_TXQ2_CR)) 2180 2181 /* How long to delay the next interrupt in case of memory shortage, in 0.1us. */ 2182 #define NOMEM_INTR_DELAY 2500 2183 2184 #ifdef TCP_OFFLOAD 2185 /** 2186 * write_ofld_wr - write an offload work request 2187 * @adap: the adapter 2188 * @m: the packet to send 2189 * @q: the Tx queue 2190 * @pidx: index of the first Tx descriptor to write 2191 * @gen: the generation value to use 2192 * @ndesc: number of descriptors the packet will occupy 2193 * 2194 * Write an offload work request to send the supplied packet. The packet 2195 * data already carry the work request with most fields populated. 2196 */ 2197 static void 2198 write_ofld_wr(adapter_t *adap, struct mbuf *m, struct sge_txq *q, 2199 unsigned int pidx, unsigned int gen, unsigned int ndesc) 2200 { 2201 unsigned int sgl_flits, flits; 2202 int i, idx, nsegs, wrlen; 2203 struct work_request_hdr *from; 2204 struct sg_ent *sgp, t3sgl[TX_MAX_SEGS / 2 + 1]; 2205 struct tx_desc *d = &q->desc[pidx]; 2206 struct txq_state txqs; 2207 struct sglist_seg *segs; 2208 struct ofld_hdr *oh = mtod(m, struct ofld_hdr *); 2209 struct sglist *sgl; 2210 2211 from = (void *)(oh + 1); /* Start of WR within mbuf */ 2212 wrlen = m->m_len - sizeof(*oh); 2213 2214 if (!(oh->flags & F_HDR_SGL)) { 2215 write_imm(d, (caddr_t)from, wrlen, gen); 2216 2217 /* 2218 * mbuf with "real" immediate tx data will be enqueue_wr'd by 2219 * t3_push_frames and freed in wr_ack. Others, like those sent 2220 * down by close_conn, t3_send_reset, etc. should be freed here. 2221 */ 2222 if (!(oh->flags & F_HDR_DF)) 2223 m_free(m); 2224 return; 2225 } 2226 2227 memcpy(&d->flit[1], &from[1], wrlen - sizeof(*from)); 2228 2229 sgl = oh->sgl; 2230 flits = wrlen / 8; 2231 sgp = (ndesc == 1) ? (struct sg_ent *)&d->flit[flits] : t3sgl; 2232 2233 nsegs = sgl->sg_nseg; 2234 segs = sgl->sg_segs; 2235 for (idx = 0, i = 0; i < nsegs; i++) { 2236 KASSERT(segs[i].ss_len, ("%s: 0 len in sgl", __func__)); 2237 if (i && idx == 0) 2238 ++sgp; 2239 sgp->len[idx] = htobe32(segs[i].ss_len); 2240 sgp->addr[idx] = htobe64(segs[i].ss_paddr); 2241 idx ^= 1; 2242 } 2243 if (idx) { 2244 sgp->len[idx] = 0; 2245 sgp->addr[idx] = 0; 2246 } 2247 2248 sgl_flits = sgl_len(nsegs); 2249 txqs.gen = gen; 2250 txqs.pidx = pidx; 2251 txqs.compl = 0; 2252 2253 write_wr_hdr_sgl(ndesc, d, &txqs, q, t3sgl, flits, sgl_flits, 2254 from->wrh_hi, from->wrh_lo); 2255 } 2256 2257 /** 2258 * ofld_xmit - send a packet through an offload queue 2259 * @adap: the adapter 2260 * @q: the Tx offload queue 2261 * @m: the packet 2262 * 2263 * Send an offload packet through an SGE offload queue. 2264 */ 2265 static int 2266 ofld_xmit(adapter_t *adap, struct sge_qset *qs, struct mbuf *m) 2267 { 2268 int ret; 2269 unsigned int ndesc; 2270 unsigned int pidx, gen; 2271 struct sge_txq *q = &qs->txq[TXQ_OFLD]; 2272 struct ofld_hdr *oh = mtod(m, struct ofld_hdr *); 2273 2274 ndesc = G_HDR_NDESC(oh->flags); 2275 2276 TXQ_LOCK(qs); 2277 again: reclaim_completed_tx(qs, 16, TXQ_OFLD); 2278 ret = check_desc_avail(adap, q, m, ndesc, TXQ_OFLD); 2279 if (__predict_false(ret)) { 2280 if (ret == 1) { 2281 TXQ_UNLOCK(qs); 2282 return (EINTR); 2283 } 2284 goto again; 2285 } 2286 2287 gen = q->gen; 2288 q->in_use += ndesc; 2289 pidx = q->pidx; 2290 q->pidx += ndesc; 2291 if (q->pidx >= q->size) { 2292 q->pidx -= q->size; 2293 q->gen ^= 1; 2294 } 2295 2296 write_ofld_wr(adap, m, q, pidx, gen, ndesc); 2297 check_ring_tx_db(adap, q, 1); 2298 TXQ_UNLOCK(qs); 2299 2300 return (0); 2301 } 2302 2303 /** 2304 * restart_offloadq - restart a suspended offload queue 2305 * @qs: the queue set cotaining the offload queue 2306 * 2307 * Resumes transmission on a suspended Tx offload queue. 2308 */ 2309 static void 2310 restart_offloadq(void *data, int npending) 2311 { 2312 struct mbuf *m; 2313 struct sge_qset *qs = data; 2314 struct sge_txq *q = &qs->txq[TXQ_OFLD]; 2315 adapter_t *adap = qs->port->adapter; 2316 int cleaned; 2317 2318 TXQ_LOCK(qs); 2319 again: cleaned = reclaim_completed_tx(qs, 16, TXQ_OFLD); 2320 2321 while ((m = mbufq_first(&q->sendq)) != NULL) { 2322 unsigned int gen, pidx; 2323 struct ofld_hdr *oh = mtod(m, struct ofld_hdr *); 2324 unsigned int ndesc = G_HDR_NDESC(oh->flags); 2325 2326 if (__predict_false(q->size - q->in_use < ndesc)) { 2327 setbit(&qs->txq_stopped, TXQ_OFLD); 2328 if (should_restart_tx(q) && 2329 test_and_clear_bit(TXQ_OFLD, &qs->txq_stopped)) 2330 goto again; 2331 q->stops++; 2332 break; 2333 } 2334 2335 gen = q->gen; 2336 q->in_use += ndesc; 2337 pidx = q->pidx; 2338 q->pidx += ndesc; 2339 if (q->pidx >= q->size) { 2340 q->pidx -= q->size; 2341 q->gen ^= 1; 2342 } 2343 2344 (void)mbufq_dequeue(&q->sendq); 2345 TXQ_UNLOCK(qs); 2346 write_ofld_wr(adap, m, q, pidx, gen, ndesc); 2347 TXQ_LOCK(qs); 2348 } 2349 #if USE_GTS 2350 set_bit(TXQ_RUNNING, &q->flags); 2351 set_bit(TXQ_LAST_PKT_DB, &q->flags); 2352 #endif 2353 TXQ_UNLOCK(qs); 2354 wmb(); 2355 t3_write_reg(adap, A_SG_KDOORBELL, 2356 F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id)); 2357 } 2358 2359 /** 2360 * t3_offload_tx - send an offload packet 2361 * @m: the packet 2362 * 2363 * Sends an offload packet. We use the packet priority to select the 2364 * appropriate Tx queue as follows: bit 0 indicates whether the packet 2365 * should be sent as regular or control, bits 1-3 select the queue set. 2366 */ 2367 int 2368 t3_offload_tx(struct adapter *sc, struct mbuf *m) 2369 { 2370 struct ofld_hdr *oh = mtod(m, struct ofld_hdr *); 2371 struct sge_qset *qs = &sc->sge.qs[G_HDR_QSET(oh->flags)]; 2372 2373 if (oh->flags & F_HDR_CTRL) { 2374 m_adj(m, sizeof (*oh)); /* trim ofld_hdr off */ 2375 return (ctrl_xmit(sc, qs, m)); 2376 } else 2377 return (ofld_xmit(sc, qs, m)); 2378 } 2379 #endif 2380 2381 static void 2382 restart_tx(struct sge_qset *qs) 2383 { 2384 struct adapter *sc = qs->port->adapter; 2385 2386 if (isset(&qs->txq_stopped, TXQ_OFLD) && 2387 should_restart_tx(&qs->txq[TXQ_OFLD]) && 2388 test_and_clear_bit(TXQ_OFLD, &qs->txq_stopped)) { 2389 qs->txq[TXQ_OFLD].restarts++; 2390 taskqueue_enqueue(sc->tq, &qs->txq[TXQ_OFLD].qresume_task); 2391 } 2392 2393 if (isset(&qs->txq_stopped, TXQ_CTRL) && 2394 should_restart_tx(&qs->txq[TXQ_CTRL]) && 2395 test_and_clear_bit(TXQ_CTRL, &qs->txq_stopped)) { 2396 qs->txq[TXQ_CTRL].restarts++; 2397 taskqueue_enqueue(sc->tq, &qs->txq[TXQ_CTRL].qresume_task); 2398 } 2399 } 2400 2401 /** 2402 * t3_sge_alloc_qset - initialize an SGE queue set 2403 * @sc: the controller softc 2404 * @id: the queue set id 2405 * @nports: how many Ethernet ports will be using this queue set 2406 * @irq_vec_idx: the IRQ vector index for response queue interrupts 2407 * @p: configuration parameters for this queue set 2408 * @ntxq: number of Tx queues for the queue set 2409 * @pi: port info for queue set 2410 * 2411 * Allocate resources and initialize an SGE queue set. A queue set 2412 * comprises a response queue, two Rx free-buffer queues, and up to 3 2413 * Tx queues. The Tx queues are assigned roles in the order Ethernet 2414 * queue, offload queue, and control queue. 2415 */ 2416 int 2417 t3_sge_alloc_qset(adapter_t *sc, u_int id, int nports, int irq_vec_idx, 2418 const struct qset_params *p, int ntxq, struct port_info *pi) 2419 { 2420 struct sge_qset *q = &sc->sge.qs[id]; 2421 int i, ret = 0; 2422 2423 MTX_INIT(&q->lock, q->namebuf, NULL, MTX_DEF); 2424 q->port = pi; 2425 q->adap = sc; 2426 2427 if ((q->txq[TXQ_ETH].txq_mr = buf_ring_alloc(cxgb_txq_buf_ring_size, 2428 M_DEVBUF, M_WAITOK, &q->lock)) == NULL) { 2429 device_printf(sc->dev, "failed to allocate mbuf ring\n"); 2430 goto err; 2431 } 2432 if ((q->txq[TXQ_ETH].txq_ifq = malloc(sizeof(struct ifaltq), M_DEVBUF, 2433 M_NOWAIT | M_ZERO)) == NULL) { 2434 device_printf(sc->dev, "failed to allocate ifq\n"); 2435 goto err; 2436 } 2437 ifq_init(q->txq[TXQ_ETH].txq_ifq, pi->ifp); 2438 callout_init(&q->txq[TXQ_ETH].txq_timer, 1); 2439 callout_init(&q->txq[TXQ_ETH].txq_watchdog, 1); 2440 q->txq[TXQ_ETH].txq_timer.c_cpu = id % mp_ncpus; 2441 q->txq[TXQ_ETH].txq_watchdog.c_cpu = id % mp_ncpus; 2442 2443 init_qset_cntxt(q, id); 2444 q->idx = id; 2445 if ((ret = alloc_ring(sc, p->fl_size, sizeof(struct rx_desc), 2446 sizeof(struct rx_sw_desc), &q->fl[0].phys_addr, 2447 &q->fl[0].desc, &q->fl[0].sdesc, 2448 &q->fl[0].desc_tag, &q->fl[0].desc_map, 2449 sc->rx_dmat, &q->fl[0].entry_tag)) != 0) { 2450 printf("error %d from alloc ring fl0\n", ret); 2451 goto err; 2452 } 2453 2454 if ((ret = alloc_ring(sc, p->jumbo_size, sizeof(struct rx_desc), 2455 sizeof(struct rx_sw_desc), &q->fl[1].phys_addr, 2456 &q->fl[1].desc, &q->fl[1].sdesc, 2457 &q->fl[1].desc_tag, &q->fl[1].desc_map, 2458 sc->rx_jumbo_dmat, &q->fl[1].entry_tag)) != 0) { 2459 printf("error %d from alloc ring fl1\n", ret); 2460 goto err; 2461 } 2462 2463 if ((ret = alloc_ring(sc, p->rspq_size, sizeof(struct rsp_desc), 0, 2464 &q->rspq.phys_addr, &q->rspq.desc, NULL, 2465 &q->rspq.desc_tag, &q->rspq.desc_map, 2466 NULL, NULL)) != 0) { 2467 printf("error %d from alloc ring rspq\n", ret); 2468 goto err; 2469 } 2470 2471 snprintf(q->rspq.lockbuf, RSPQ_NAME_LEN, "t3 rspq lock %d:%d", 2472 device_get_unit(sc->dev), irq_vec_idx); 2473 MTX_INIT(&q->rspq.lock, q->rspq.lockbuf, NULL, MTX_DEF); 2474 2475 for (i = 0; i < ntxq; ++i) { 2476 size_t sz = i == TXQ_CTRL ? 0 : sizeof(struct tx_sw_desc); 2477 2478 if ((ret = alloc_ring(sc, p->txq_size[i], 2479 sizeof(struct tx_desc), sz, 2480 &q->txq[i].phys_addr, &q->txq[i].desc, 2481 &q->txq[i].sdesc, &q->txq[i].desc_tag, 2482 &q->txq[i].desc_map, 2483 sc->tx_dmat, &q->txq[i].entry_tag)) != 0) { 2484 printf("error %d from alloc ring tx %i\n", ret, i); 2485 goto err; 2486 } 2487 mbufq_init(&q->txq[i].sendq, INT_MAX); 2488 q->txq[i].gen = 1; 2489 q->txq[i].size = p->txq_size[i]; 2490 } 2491 2492 #ifdef TCP_OFFLOAD 2493 TASK_INIT(&q->txq[TXQ_OFLD].qresume_task, 0, restart_offloadq, q); 2494 #endif 2495 TASK_INIT(&q->txq[TXQ_CTRL].qresume_task, 0, restart_ctrlq, q); 2496 TASK_INIT(&q->txq[TXQ_ETH].qreclaim_task, 0, sge_txq_reclaim_handler, q); 2497 TASK_INIT(&q->txq[TXQ_OFLD].qreclaim_task, 0, sge_txq_reclaim_handler, q); 2498 2499 q->fl[0].gen = q->fl[1].gen = 1; 2500 q->fl[0].size = p->fl_size; 2501 q->fl[1].size = p->jumbo_size; 2502 2503 q->rspq.gen = 1; 2504 q->rspq.cidx = 0; 2505 q->rspq.size = p->rspq_size; 2506 2507 q->txq[TXQ_ETH].stop_thres = nports * 2508 flits_to_desc(sgl_len(TX_MAX_SEGS + 1) + 3); 2509 2510 q->fl[0].buf_size = MCLBYTES; 2511 q->fl[0].zone = zone_pack; 2512 q->fl[0].type = EXT_PACKET; 2513 2514 if (p->jumbo_buf_size == MJUM16BYTES) { 2515 q->fl[1].zone = zone_jumbo16; 2516 q->fl[1].type = EXT_JUMBO16; 2517 } else if (p->jumbo_buf_size == MJUM9BYTES) { 2518 q->fl[1].zone = zone_jumbo9; 2519 q->fl[1].type = EXT_JUMBO9; 2520 } else if (p->jumbo_buf_size == MJUMPAGESIZE) { 2521 q->fl[1].zone = zone_jumbop; 2522 q->fl[1].type = EXT_JUMBOP; 2523 } else { 2524 KASSERT(0, ("can't deal with jumbo_buf_size %d.", p->jumbo_buf_size)); 2525 ret = EDOOFUS; 2526 goto err; 2527 } 2528 q->fl[1].buf_size = p->jumbo_buf_size; 2529 2530 /* Allocate and setup the lro_ctrl structure */ 2531 q->lro.enabled = !!(pi->ifp->if_capenable & IFCAP_LRO); 2532 #if defined(INET6) || defined(INET) 2533 ret = tcp_lro_init(&q->lro.ctrl); 2534 if (ret) { 2535 printf("error %d from tcp_lro_init\n", ret); 2536 goto err; 2537 } 2538 #endif 2539 q->lro.ctrl.ifp = pi->ifp; 2540 2541 mtx_lock_spin(&sc->sge.reg_lock); 2542 ret = -t3_sge_init_rspcntxt(sc, q->rspq.cntxt_id, irq_vec_idx, 2543 q->rspq.phys_addr, q->rspq.size, 2544 q->fl[0].buf_size, 1, 0); 2545 if (ret) { 2546 printf("error %d from t3_sge_init_rspcntxt\n", ret); 2547 goto err_unlock; 2548 } 2549 2550 for (i = 0; i < SGE_RXQ_PER_SET; ++i) { 2551 ret = -t3_sge_init_flcntxt(sc, q->fl[i].cntxt_id, 0, 2552 q->fl[i].phys_addr, q->fl[i].size, 2553 q->fl[i].buf_size, p->cong_thres, 1, 2554 0); 2555 if (ret) { 2556 printf("error %d from t3_sge_init_flcntxt for index i=%d\n", ret, i); 2557 goto err_unlock; 2558 } 2559 } 2560 2561 ret = -t3_sge_init_ecntxt(sc, q->txq[TXQ_ETH].cntxt_id, USE_GTS, 2562 SGE_CNTXT_ETH, id, q->txq[TXQ_ETH].phys_addr, 2563 q->txq[TXQ_ETH].size, q->txq[TXQ_ETH].token, 2564 1, 0); 2565 if (ret) { 2566 printf("error %d from t3_sge_init_ecntxt\n", ret); 2567 goto err_unlock; 2568 } 2569 2570 if (ntxq > 1) { 2571 ret = -t3_sge_init_ecntxt(sc, q->txq[TXQ_OFLD].cntxt_id, 2572 USE_GTS, SGE_CNTXT_OFLD, id, 2573 q->txq[TXQ_OFLD].phys_addr, 2574 q->txq[TXQ_OFLD].size, 0, 1, 0); 2575 if (ret) { 2576 printf("error %d from t3_sge_init_ecntxt\n", ret); 2577 goto err_unlock; 2578 } 2579 } 2580 2581 if (ntxq > 2) { 2582 ret = -t3_sge_init_ecntxt(sc, q->txq[TXQ_CTRL].cntxt_id, 0, 2583 SGE_CNTXT_CTRL, id, 2584 q->txq[TXQ_CTRL].phys_addr, 2585 q->txq[TXQ_CTRL].size, 2586 q->txq[TXQ_CTRL].token, 1, 0); 2587 if (ret) { 2588 printf("error %d from t3_sge_init_ecntxt\n", ret); 2589 goto err_unlock; 2590 } 2591 } 2592 2593 mtx_unlock_spin(&sc->sge.reg_lock); 2594 t3_update_qset_coalesce(q, p); 2595 2596 refill_fl(sc, &q->fl[0], q->fl[0].size); 2597 refill_fl(sc, &q->fl[1], q->fl[1].size); 2598 refill_rspq(sc, &q->rspq, q->rspq.size - 1); 2599 2600 t3_write_reg(sc, A_SG_GTS, V_RSPQ(q->rspq.cntxt_id) | 2601 V_NEWTIMER(q->rspq.holdoff_tmr)); 2602 2603 return (0); 2604 2605 err_unlock: 2606 mtx_unlock_spin(&sc->sge.reg_lock); 2607 err: 2608 TXQ_LOCK(q); 2609 t3_free_qset(sc, q); 2610 2611 return (ret); 2612 } 2613 2614 /* 2615 * Remove CPL_RX_PKT headers from the mbuf and reduce it to a regular mbuf with 2616 * ethernet data. Hardware assistance with various checksums and any vlan tag 2617 * will also be taken into account here. 2618 */ 2619 void 2620 t3_rx_eth(struct adapter *adap, struct mbuf *m, int ethpad) 2621 { 2622 struct cpl_rx_pkt *cpl = (struct cpl_rx_pkt *)(mtod(m, uint8_t *) + ethpad); 2623 struct port_info *pi = &adap->port[adap->rxpkt_map[cpl->iff]]; 2624 struct ifnet *ifp = pi->ifp; 2625 2626 if (cpl->vlan_valid) { 2627 m->m_pkthdr.ether_vtag = ntohs(cpl->vlan); 2628 m->m_flags |= M_VLANTAG; 2629 } 2630 2631 m->m_pkthdr.rcvif = ifp; 2632 /* 2633 * adjust after conversion to mbuf chain 2634 */ 2635 m->m_pkthdr.len -= (sizeof(*cpl) + ethpad); 2636 m->m_len -= (sizeof(*cpl) + ethpad); 2637 m->m_data += (sizeof(*cpl) + ethpad); 2638 2639 if (!cpl->fragment && cpl->csum_valid && cpl->csum == 0xffff) { 2640 struct ether_header *eh = mtod(m, void *); 2641 uint16_t eh_type; 2642 2643 if (eh->ether_type == htons(ETHERTYPE_VLAN)) { 2644 struct ether_vlan_header *evh = mtod(m, void *); 2645 2646 eh_type = evh->evl_proto; 2647 } else 2648 eh_type = eh->ether_type; 2649 2650 if (ifp->if_capenable & IFCAP_RXCSUM && 2651 eh_type == htons(ETHERTYPE_IP)) { 2652 m->m_pkthdr.csum_flags = (CSUM_IP_CHECKED | 2653 CSUM_IP_VALID | CSUM_DATA_VALID | CSUM_PSEUDO_HDR); 2654 m->m_pkthdr.csum_data = 0xffff; 2655 } else if (ifp->if_capenable & IFCAP_RXCSUM_IPV6 && 2656 eh_type == htons(ETHERTYPE_IPV6)) { 2657 m->m_pkthdr.csum_flags = (CSUM_DATA_VALID_IPV6 | 2658 CSUM_PSEUDO_HDR); 2659 m->m_pkthdr.csum_data = 0xffff; 2660 } 2661 } 2662 } 2663 2664 /** 2665 * get_packet - return the next ingress packet buffer from a free list 2666 * @adap: the adapter that received the packet 2667 * @drop_thres: # of remaining buffers before we start dropping packets 2668 * @qs: the qset that the SGE free list holding the packet belongs to 2669 * @mh: the mbuf header, contains a pointer to the head and tail of the mbuf chain 2670 * @r: response descriptor 2671 * 2672 * Get the next packet from a free list and complete setup of the 2673 * sk_buff. If the packet is small we make a copy and recycle the 2674 * original buffer, otherwise we use the original buffer itself. If a 2675 * positive drop threshold is supplied packets are dropped and their 2676 * buffers recycled if (a) the number of remaining buffers is under the 2677 * threshold and the packet is too big to copy, or (b) the packet should 2678 * be copied but there is no memory for the copy. 2679 */ 2680 static int 2681 get_packet(adapter_t *adap, unsigned int drop_thres, struct sge_qset *qs, 2682 struct t3_mbuf_hdr *mh, struct rsp_desc *r) 2683 { 2684 2685 unsigned int len_cq = ntohl(r->len_cq); 2686 struct sge_fl *fl = (len_cq & F_RSPD_FLQ) ? &qs->fl[1] : &qs->fl[0]; 2687 int mask, cidx = fl->cidx; 2688 struct rx_sw_desc *sd = &fl->sdesc[cidx]; 2689 uint32_t len = G_RSPD_LEN(len_cq); 2690 uint32_t flags = M_EXT; 2691 uint8_t sopeop = G_RSPD_SOP_EOP(ntohl(r->flags)); 2692 caddr_t cl; 2693 struct mbuf *m; 2694 int ret = 0; 2695 2696 mask = fl->size - 1; 2697 prefetch(fl->sdesc[(cidx + 1) & mask].m); 2698 prefetch(fl->sdesc[(cidx + 2) & mask].m); 2699 prefetch(fl->sdesc[(cidx + 1) & mask].rxsd_cl); 2700 prefetch(fl->sdesc[(cidx + 2) & mask].rxsd_cl); 2701 2702 fl->credits--; 2703 bus_dmamap_sync(fl->entry_tag, sd->map, BUS_DMASYNC_POSTREAD); 2704 2705 if (recycle_enable && len <= SGE_RX_COPY_THRES && 2706 sopeop == RSPQ_SOP_EOP) { 2707 if ((m = m_gethdr(M_NOWAIT, MT_DATA)) == NULL) 2708 goto skip_recycle; 2709 cl = mtod(m, void *); 2710 memcpy(cl, sd->rxsd_cl, len); 2711 recycle_rx_buf(adap, fl, fl->cidx); 2712 m->m_pkthdr.len = m->m_len = len; 2713 m->m_flags = 0; 2714 mh->mh_head = mh->mh_tail = m; 2715 ret = 1; 2716 goto done; 2717 } else { 2718 skip_recycle: 2719 bus_dmamap_unload(fl->entry_tag, sd->map); 2720 cl = sd->rxsd_cl; 2721 m = sd->m; 2722 2723 if ((sopeop == RSPQ_SOP_EOP) || 2724 (sopeop == RSPQ_SOP)) 2725 flags |= M_PKTHDR; 2726 m_init(m, M_NOWAIT, MT_DATA, flags); 2727 if (fl->zone == zone_pack) { 2728 /* 2729 * restore clobbered data pointer 2730 */ 2731 m->m_data = m->m_ext.ext_buf; 2732 } else { 2733 m_cljset(m, cl, fl->type); 2734 } 2735 m->m_len = len; 2736 } 2737 switch(sopeop) { 2738 case RSPQ_SOP_EOP: 2739 ret = 1; 2740 /* FALLTHROUGH */ 2741 case RSPQ_SOP: 2742 mh->mh_head = mh->mh_tail = m; 2743 m->m_pkthdr.len = len; 2744 break; 2745 case RSPQ_EOP: 2746 ret = 1; 2747 /* FALLTHROUGH */ 2748 case RSPQ_NSOP_NEOP: 2749 if (mh->mh_tail == NULL) { 2750 log(LOG_ERR, "discarding intermediate descriptor entry\n"); 2751 m_freem(m); 2752 break; 2753 } 2754 mh->mh_tail->m_next = m; 2755 mh->mh_tail = m; 2756 mh->mh_head->m_pkthdr.len += len; 2757 break; 2758 } 2759 if (cxgb_debug) 2760 printf("len=%d pktlen=%d\n", m->m_len, m->m_pkthdr.len); 2761 done: 2762 if (++fl->cidx == fl->size) 2763 fl->cidx = 0; 2764 2765 return (ret); 2766 } 2767 2768 /** 2769 * handle_rsp_cntrl_info - handles control information in a response 2770 * @qs: the queue set corresponding to the response 2771 * @flags: the response control flags 2772 * 2773 * Handles the control information of an SGE response, such as GTS 2774 * indications and completion credits for the queue set's Tx queues. 2775 * HW coalesces credits, we don't do any extra SW coalescing. 2776 */ 2777 static __inline void 2778 handle_rsp_cntrl_info(struct sge_qset *qs, uint32_t flags) 2779 { 2780 unsigned int credits; 2781 2782 #if USE_GTS 2783 if (flags & F_RSPD_TXQ0_GTS) 2784 clear_bit(TXQ_RUNNING, &qs->txq[TXQ_ETH].flags); 2785 #endif 2786 credits = G_RSPD_TXQ0_CR(flags); 2787 if (credits) 2788 qs->txq[TXQ_ETH].processed += credits; 2789 2790 credits = G_RSPD_TXQ2_CR(flags); 2791 if (credits) 2792 qs->txq[TXQ_CTRL].processed += credits; 2793 2794 # if USE_GTS 2795 if (flags & F_RSPD_TXQ1_GTS) 2796 clear_bit(TXQ_RUNNING, &qs->txq[TXQ_OFLD].flags); 2797 # endif 2798 credits = G_RSPD_TXQ1_CR(flags); 2799 if (credits) 2800 qs->txq[TXQ_OFLD].processed += credits; 2801 2802 } 2803 2804 static void 2805 check_ring_db(adapter_t *adap, struct sge_qset *qs, 2806 unsigned int sleeping) 2807 { 2808 ; 2809 } 2810 2811 /** 2812 * process_responses - process responses from an SGE response queue 2813 * @adap: the adapter 2814 * @qs: the queue set to which the response queue belongs 2815 * @budget: how many responses can be processed in this round 2816 * 2817 * Process responses from an SGE response queue up to the supplied budget. 2818 * Responses include received packets as well as credits and other events 2819 * for the queues that belong to the response queue's queue set. 2820 * A negative budget is effectively unlimited. 2821 * 2822 * Additionally choose the interrupt holdoff time for the next interrupt 2823 * on this queue. If the system is under memory shortage use a fairly 2824 * long delay to help recovery. 2825 */ 2826 static int 2827 process_responses(adapter_t *adap, struct sge_qset *qs, int budget) 2828 { 2829 struct sge_rspq *rspq = &qs->rspq; 2830 struct rsp_desc *r = &rspq->desc[rspq->cidx]; 2831 int budget_left = budget; 2832 unsigned int sleeping = 0; 2833 #if defined(INET6) || defined(INET) 2834 int lro_enabled = qs->lro.enabled; 2835 int skip_lro; 2836 struct lro_ctrl *lro_ctrl = &qs->lro.ctrl; 2837 #endif 2838 struct t3_mbuf_hdr *mh = &rspq->rspq_mh; 2839 #ifdef DEBUG 2840 static int last_holdoff = 0; 2841 if (cxgb_debug && rspq->holdoff_tmr != last_holdoff) { 2842 printf("next_holdoff=%d\n", rspq->holdoff_tmr); 2843 last_holdoff = rspq->holdoff_tmr; 2844 } 2845 #endif 2846 rspq->next_holdoff = rspq->holdoff_tmr; 2847 2848 while (__predict_true(budget_left && is_new_response(r, rspq))) { 2849 int eth, eop = 0, ethpad = 0; 2850 uint32_t flags = ntohl(r->flags); 2851 uint32_t rss_hash = be32toh(r->rss_hdr.rss_hash_val); 2852 uint8_t opcode = r->rss_hdr.opcode; 2853 2854 eth = (opcode == CPL_RX_PKT); 2855 2856 if (__predict_false(flags & F_RSPD_ASYNC_NOTIF)) { 2857 struct mbuf *m; 2858 2859 if (cxgb_debug) 2860 printf("async notification\n"); 2861 2862 if (mh->mh_head == NULL) { 2863 mh->mh_head = m_gethdr(M_NOWAIT, MT_DATA); 2864 m = mh->mh_head; 2865 } else { 2866 m = m_gethdr(M_NOWAIT, MT_DATA); 2867 } 2868 if (m == NULL) 2869 goto no_mem; 2870 2871 memcpy(mtod(m, char *), r, AN_PKT_SIZE); 2872 m->m_len = m->m_pkthdr.len = AN_PKT_SIZE; 2873 *mtod(m, char *) = CPL_ASYNC_NOTIF; 2874 opcode = CPL_ASYNC_NOTIF; 2875 eop = 1; 2876 rspq->async_notif++; 2877 goto skip; 2878 } else if (flags & F_RSPD_IMM_DATA_VALID) { 2879 struct mbuf *m = m_gethdr(M_NOWAIT, MT_DATA); 2880 2881 if (m == NULL) { 2882 no_mem: 2883 rspq->next_holdoff = NOMEM_INTR_DELAY; 2884 budget_left--; 2885 break; 2886 } 2887 if (mh->mh_head == NULL) 2888 mh->mh_head = m; 2889 else 2890 mh->mh_tail->m_next = m; 2891 mh->mh_tail = m; 2892 2893 get_imm_packet(adap, r, m); 2894 mh->mh_head->m_pkthdr.len += m->m_len; 2895 eop = 1; 2896 rspq->imm_data++; 2897 } else if (r->len_cq) { 2898 int drop_thresh = eth ? SGE_RX_DROP_THRES : 0; 2899 2900 eop = get_packet(adap, drop_thresh, qs, mh, r); 2901 if (eop) { 2902 if (r->rss_hdr.hash_type && !adap->timestamp) { 2903 M_HASHTYPE_SET(mh->mh_head, M_HASHTYPE_OPAQUE); 2904 mh->mh_head->m_pkthdr.flowid = rss_hash; 2905 } 2906 } 2907 2908 ethpad = 2; 2909 } else { 2910 rspq->pure_rsps++; 2911 } 2912 skip: 2913 if (flags & RSPD_CTRL_MASK) { 2914 sleeping |= flags & RSPD_GTS_MASK; 2915 handle_rsp_cntrl_info(qs, flags); 2916 } 2917 2918 if (!eth && eop) { 2919 rspq->offload_pkts++; 2920 #ifdef TCP_OFFLOAD 2921 adap->cpl_handler[opcode](qs, r, mh->mh_head); 2922 #else 2923 m_freem(mh->mh_head); 2924 #endif 2925 mh->mh_head = NULL; 2926 } else if (eth && eop) { 2927 struct mbuf *m = mh->mh_head; 2928 2929 t3_rx_eth(adap, m, ethpad); 2930 2931 /* 2932 * The T304 sends incoming packets on any qset. If LRO 2933 * is also enabled, we could end up sending packet up 2934 * lro_ctrl->ifp's input. That is incorrect. 2935 * 2936 * The mbuf's rcvif was derived from the cpl header and 2937 * is accurate. Skip LRO and just use that. 2938 */ 2939 #if defined(INET6) || defined(INET) 2940 skip_lro = __predict_false(qs->port->ifp != m->m_pkthdr.rcvif); 2941 2942 if (lro_enabled && lro_ctrl->lro_cnt && !skip_lro 2943 && (tcp_lro_rx(lro_ctrl, m, 0) == 0) 2944 ) { 2945 /* successfully queue'd for LRO */ 2946 } else 2947 #endif 2948 { 2949 /* 2950 * LRO not enabled, packet unsuitable for LRO, 2951 * or unable to queue. Pass it up right now in 2952 * either case. 2953 */ 2954 struct ifnet *ifp = m->m_pkthdr.rcvif; 2955 (*ifp->if_input)(ifp, m); 2956 } 2957 mh->mh_head = NULL; 2958 2959 } 2960 2961 r++; 2962 if (__predict_false(++rspq->cidx == rspq->size)) { 2963 rspq->cidx = 0; 2964 rspq->gen ^= 1; 2965 r = rspq->desc; 2966 } 2967 2968 if (++rspq->credits >= 64) { 2969 refill_rspq(adap, rspq, rspq->credits); 2970 rspq->credits = 0; 2971 } 2972 __refill_fl_lt(adap, &qs->fl[0], 32); 2973 __refill_fl_lt(adap, &qs->fl[1], 32); 2974 --budget_left; 2975 } 2976 2977 #if defined(INET6) || defined(INET) 2978 /* Flush LRO */ 2979 while (!SLIST_EMPTY(&lro_ctrl->lro_active)) { 2980 struct lro_entry *queued = SLIST_FIRST(&lro_ctrl->lro_active); 2981 SLIST_REMOVE_HEAD(&lro_ctrl->lro_active, next); 2982 tcp_lro_flush(lro_ctrl, queued); 2983 } 2984 #endif 2985 2986 if (sleeping) 2987 check_ring_db(adap, qs, sleeping); 2988 2989 mb(); /* commit Tx queue processed updates */ 2990 if (__predict_false(qs->txq_stopped > 1)) 2991 restart_tx(qs); 2992 2993 __refill_fl_lt(adap, &qs->fl[0], 512); 2994 __refill_fl_lt(adap, &qs->fl[1], 512); 2995 budget -= budget_left; 2996 return (budget); 2997 } 2998 2999 /* 3000 * A helper function that processes responses and issues GTS. 3001 */ 3002 static __inline int 3003 process_responses_gts(adapter_t *adap, struct sge_rspq *rq) 3004 { 3005 int work; 3006 static int last_holdoff = 0; 3007 3008 work = process_responses(adap, rspq_to_qset(rq), -1); 3009 3010 if (cxgb_debug && (rq->next_holdoff != last_holdoff)) { 3011 printf("next_holdoff=%d\n", rq->next_holdoff); 3012 last_holdoff = rq->next_holdoff; 3013 } 3014 t3_write_reg(adap, A_SG_GTS, V_RSPQ(rq->cntxt_id) | 3015 V_NEWTIMER(rq->next_holdoff) | V_NEWINDEX(rq->cidx)); 3016 3017 return (work); 3018 } 3019 3020 3021 /* 3022 * Interrupt handler for legacy INTx interrupts for T3B-based cards. 3023 * Handles data events from SGE response queues as well as error and other 3024 * async events as they all use the same interrupt pin. We use one SGE 3025 * response queue per port in this mode and protect all response queues with 3026 * queue 0's lock. 3027 */ 3028 void 3029 t3b_intr(void *data) 3030 { 3031 uint32_t i, map; 3032 adapter_t *adap = data; 3033 struct sge_rspq *q0 = &adap->sge.qs[0].rspq; 3034 3035 t3_write_reg(adap, A_PL_CLI, 0); 3036 map = t3_read_reg(adap, A_SG_DATA_INTR); 3037 3038 if (!map) 3039 return; 3040 3041 if (__predict_false(map & F_ERRINTR)) { 3042 t3_write_reg(adap, A_PL_INT_ENABLE0, 0); 3043 (void) t3_read_reg(adap, A_PL_INT_ENABLE0); 3044 taskqueue_enqueue(adap->tq, &adap->slow_intr_task); 3045 } 3046 3047 mtx_lock(&q0->lock); 3048 for_each_port(adap, i) 3049 if (map & (1 << i)) 3050 process_responses_gts(adap, &adap->sge.qs[i].rspq); 3051 mtx_unlock(&q0->lock); 3052 } 3053 3054 /* 3055 * The MSI interrupt handler. This needs to handle data events from SGE 3056 * response queues as well as error and other async events as they all use 3057 * the same MSI vector. We use one SGE response queue per port in this mode 3058 * and protect all response queues with queue 0's lock. 3059 */ 3060 void 3061 t3_intr_msi(void *data) 3062 { 3063 adapter_t *adap = data; 3064 struct sge_rspq *q0 = &adap->sge.qs[0].rspq; 3065 int i, new_packets = 0; 3066 3067 mtx_lock(&q0->lock); 3068 3069 for_each_port(adap, i) 3070 if (process_responses_gts(adap, &adap->sge.qs[i].rspq)) 3071 new_packets = 1; 3072 mtx_unlock(&q0->lock); 3073 if (new_packets == 0) { 3074 t3_write_reg(adap, A_PL_INT_ENABLE0, 0); 3075 (void) t3_read_reg(adap, A_PL_INT_ENABLE0); 3076 taskqueue_enqueue(adap->tq, &adap->slow_intr_task); 3077 } 3078 } 3079 3080 void 3081 t3_intr_msix(void *data) 3082 { 3083 struct sge_qset *qs = data; 3084 adapter_t *adap = qs->port->adapter; 3085 struct sge_rspq *rspq = &qs->rspq; 3086 3087 if (process_responses_gts(adap, rspq) == 0) 3088 rspq->unhandled_irqs++; 3089 } 3090 3091 #define QDUMP_SBUF_SIZE 32 * 400 3092 static int 3093 t3_dump_rspq(SYSCTL_HANDLER_ARGS) 3094 { 3095 struct sge_rspq *rspq; 3096 struct sge_qset *qs; 3097 int i, err, dump_end, idx; 3098 struct sbuf *sb; 3099 struct rsp_desc *rspd; 3100 uint32_t data[4]; 3101 3102 rspq = arg1; 3103 qs = rspq_to_qset(rspq); 3104 if (rspq->rspq_dump_count == 0) 3105 return (0); 3106 if (rspq->rspq_dump_count > RSPQ_Q_SIZE) { 3107 log(LOG_WARNING, 3108 "dump count is too large %d\n", rspq->rspq_dump_count); 3109 rspq->rspq_dump_count = 0; 3110 return (EINVAL); 3111 } 3112 if (rspq->rspq_dump_start > (RSPQ_Q_SIZE-1)) { 3113 log(LOG_WARNING, 3114 "dump start of %d is greater than queue size\n", 3115 rspq->rspq_dump_start); 3116 rspq->rspq_dump_start = 0; 3117 return (EINVAL); 3118 } 3119 err = t3_sge_read_rspq(qs->port->adapter, rspq->cntxt_id, data); 3120 if (err) 3121 return (err); 3122 err = sysctl_wire_old_buffer(req, 0); 3123 if (err) 3124 return (err); 3125 sb = sbuf_new_for_sysctl(NULL, NULL, QDUMP_SBUF_SIZE, req); 3126 3127 sbuf_printf(sb, " \n index=%u size=%u MSI-X/RspQ=%u intr enable=%u intr armed=%u\n", 3128 (data[0] & 0xffff), data[0] >> 16, ((data[2] >> 20) & 0x3f), 3129 ((data[2] >> 26) & 1), ((data[2] >> 27) & 1)); 3130 sbuf_printf(sb, " generation=%u CQ mode=%u FL threshold=%u\n", 3131 ((data[2] >> 28) & 1), ((data[2] >> 31) & 1), data[3]); 3132 3133 sbuf_printf(sb, " start=%d -> end=%d\n", rspq->rspq_dump_start, 3134 (rspq->rspq_dump_start + rspq->rspq_dump_count) & (RSPQ_Q_SIZE-1)); 3135 3136 dump_end = rspq->rspq_dump_start + rspq->rspq_dump_count; 3137 for (i = rspq->rspq_dump_start; i < dump_end; i++) { 3138 idx = i & (RSPQ_Q_SIZE-1); 3139 3140 rspd = &rspq->desc[idx]; 3141 sbuf_printf(sb, "\tidx=%04d opcode=%02x cpu_idx=%x hash_type=%x cq_idx=%x\n", 3142 idx, rspd->rss_hdr.opcode, rspd->rss_hdr.cpu_idx, 3143 rspd->rss_hdr.hash_type, be16toh(rspd->rss_hdr.cq_idx)); 3144 sbuf_printf(sb, "\trss_hash_val=%x flags=%08x len_cq=%x intr_gen=%x\n", 3145 rspd->rss_hdr.rss_hash_val, be32toh(rspd->flags), 3146 be32toh(rspd->len_cq), rspd->intr_gen); 3147 } 3148 3149 err = sbuf_finish(sb); 3150 sbuf_delete(sb); 3151 return (err); 3152 } 3153 3154 static int 3155 t3_dump_txq_eth(SYSCTL_HANDLER_ARGS) 3156 { 3157 struct sge_txq *txq; 3158 struct sge_qset *qs; 3159 int i, j, err, dump_end; 3160 struct sbuf *sb; 3161 struct tx_desc *txd; 3162 uint32_t *WR, wr_hi, wr_lo, gen; 3163 uint32_t data[4]; 3164 3165 txq = arg1; 3166 qs = txq_to_qset(txq, TXQ_ETH); 3167 if (txq->txq_dump_count == 0) { 3168 return (0); 3169 } 3170 if (txq->txq_dump_count > TX_ETH_Q_SIZE) { 3171 log(LOG_WARNING, 3172 "dump count is too large %d\n", txq->txq_dump_count); 3173 txq->txq_dump_count = 1; 3174 return (EINVAL); 3175 } 3176 if (txq->txq_dump_start > (TX_ETH_Q_SIZE-1)) { 3177 log(LOG_WARNING, 3178 "dump start of %d is greater than queue size\n", 3179 txq->txq_dump_start); 3180 txq->txq_dump_start = 0; 3181 return (EINVAL); 3182 } 3183 err = t3_sge_read_ecntxt(qs->port->adapter, qs->rspq.cntxt_id, data); 3184 if (err) 3185 return (err); 3186 err = sysctl_wire_old_buffer(req, 0); 3187 if (err) 3188 return (err); 3189 sb = sbuf_new_for_sysctl(NULL, NULL, QDUMP_SBUF_SIZE, req); 3190 3191 sbuf_printf(sb, " \n credits=%u GTS=%u index=%u size=%u rspq#=%u cmdq#=%u\n", 3192 (data[0] & 0x7fff), ((data[0] >> 15) & 1), (data[0] >> 16), 3193 (data[1] & 0xffff), ((data[3] >> 4) & 7), ((data[3] >> 7) & 1)); 3194 sbuf_printf(sb, " TUN=%u TOE=%u generation%u uP token=%u valid=%u\n", 3195 ((data[3] >> 8) & 1), ((data[3] >> 9) & 1), ((data[3] >> 10) & 1), 3196 ((data[3] >> 11) & 0xfffff), ((data[3] >> 31) & 1)); 3197 sbuf_printf(sb, " qid=%d start=%d -> end=%d\n", qs->idx, 3198 txq->txq_dump_start, 3199 (txq->txq_dump_start + txq->txq_dump_count) & (TX_ETH_Q_SIZE-1)); 3200 3201 dump_end = txq->txq_dump_start + txq->txq_dump_count; 3202 for (i = txq->txq_dump_start; i < dump_end; i++) { 3203 txd = &txq->desc[i & (TX_ETH_Q_SIZE-1)]; 3204 WR = (uint32_t *)txd->flit; 3205 wr_hi = ntohl(WR[0]); 3206 wr_lo = ntohl(WR[1]); 3207 gen = G_WR_GEN(wr_lo); 3208 3209 sbuf_printf(sb," wr_hi %08x wr_lo %08x gen %d\n", 3210 wr_hi, wr_lo, gen); 3211 for (j = 2; j < 30; j += 4) 3212 sbuf_printf(sb, "\t%08x %08x %08x %08x \n", 3213 WR[j], WR[j + 1], WR[j + 2], WR[j + 3]); 3214 3215 } 3216 err = sbuf_finish(sb); 3217 sbuf_delete(sb); 3218 return (err); 3219 } 3220 3221 static int 3222 t3_dump_txq_ctrl(SYSCTL_HANDLER_ARGS) 3223 { 3224 struct sge_txq *txq; 3225 struct sge_qset *qs; 3226 int i, j, err, dump_end; 3227 struct sbuf *sb; 3228 struct tx_desc *txd; 3229 uint32_t *WR, wr_hi, wr_lo, gen; 3230 3231 txq = arg1; 3232 qs = txq_to_qset(txq, TXQ_CTRL); 3233 if (txq->txq_dump_count == 0) { 3234 return (0); 3235 } 3236 if (txq->txq_dump_count > 256) { 3237 log(LOG_WARNING, 3238 "dump count is too large %d\n", txq->txq_dump_count); 3239 txq->txq_dump_count = 1; 3240 return (EINVAL); 3241 } 3242 if (txq->txq_dump_start > 255) { 3243 log(LOG_WARNING, 3244 "dump start of %d is greater than queue size\n", 3245 txq->txq_dump_start); 3246 txq->txq_dump_start = 0; 3247 return (EINVAL); 3248 } 3249 3250 err = sysctl_wire_old_buffer(req, 0); 3251 if (err != 0) 3252 return (err); 3253 sb = sbuf_new_for_sysctl(NULL, NULL, QDUMP_SBUF_SIZE, req); 3254 sbuf_printf(sb, " qid=%d start=%d -> end=%d\n", qs->idx, 3255 txq->txq_dump_start, 3256 (txq->txq_dump_start + txq->txq_dump_count) & 255); 3257 3258 dump_end = txq->txq_dump_start + txq->txq_dump_count; 3259 for (i = txq->txq_dump_start; i < dump_end; i++) { 3260 txd = &txq->desc[i & (255)]; 3261 WR = (uint32_t *)txd->flit; 3262 wr_hi = ntohl(WR[0]); 3263 wr_lo = ntohl(WR[1]); 3264 gen = G_WR_GEN(wr_lo); 3265 3266 sbuf_printf(sb," wr_hi %08x wr_lo %08x gen %d\n", 3267 wr_hi, wr_lo, gen); 3268 for (j = 2; j < 30; j += 4) 3269 sbuf_printf(sb, "\t%08x %08x %08x %08x \n", 3270 WR[j], WR[j + 1], WR[j + 2], WR[j + 3]); 3271 3272 } 3273 err = sbuf_finish(sb); 3274 sbuf_delete(sb); 3275 return (err); 3276 } 3277 3278 static int 3279 t3_set_coalesce_usecs(SYSCTL_HANDLER_ARGS) 3280 { 3281 adapter_t *sc = arg1; 3282 struct qset_params *qsp = &sc->params.sge.qset[0]; 3283 int coalesce_usecs; 3284 struct sge_qset *qs; 3285 int i, j, err, nqsets = 0; 3286 struct mtx *lock; 3287 3288 if ((sc->flags & FULL_INIT_DONE) == 0) 3289 return (ENXIO); 3290 3291 coalesce_usecs = qsp->coalesce_usecs; 3292 err = sysctl_handle_int(oidp, &coalesce_usecs, arg2, req); 3293 3294 if (err != 0) { 3295 return (err); 3296 } 3297 if (coalesce_usecs == qsp->coalesce_usecs) 3298 return (0); 3299 3300 for (i = 0; i < sc->params.nports; i++) 3301 for (j = 0; j < sc->port[i].nqsets; j++) 3302 nqsets++; 3303 3304 coalesce_usecs = max(1, coalesce_usecs); 3305 3306 for (i = 0; i < nqsets; i++) { 3307 qs = &sc->sge.qs[i]; 3308 qsp = &sc->params.sge.qset[i]; 3309 qsp->coalesce_usecs = coalesce_usecs; 3310 3311 lock = (sc->flags & USING_MSIX) ? &qs->rspq.lock : 3312 &sc->sge.qs[0].rspq.lock; 3313 3314 mtx_lock(lock); 3315 t3_update_qset_coalesce(qs, qsp); 3316 t3_write_reg(sc, A_SG_GTS, V_RSPQ(qs->rspq.cntxt_id) | 3317 V_NEWTIMER(qs->rspq.holdoff_tmr)); 3318 mtx_unlock(lock); 3319 } 3320 3321 return (0); 3322 } 3323 3324 static int 3325 t3_pkt_timestamp(SYSCTL_HANDLER_ARGS) 3326 { 3327 adapter_t *sc = arg1; 3328 int rc, timestamp; 3329 3330 if ((sc->flags & FULL_INIT_DONE) == 0) 3331 return (ENXIO); 3332 3333 timestamp = sc->timestamp; 3334 rc = sysctl_handle_int(oidp, ×tamp, arg2, req); 3335 3336 if (rc != 0) 3337 return (rc); 3338 3339 if (timestamp != sc->timestamp) { 3340 t3_set_reg_field(sc, A_TP_PC_CONFIG2, F_ENABLERXPKTTMSTPRSS, 3341 timestamp ? F_ENABLERXPKTTMSTPRSS : 0); 3342 sc->timestamp = timestamp; 3343 } 3344 3345 return (0); 3346 } 3347 3348 void 3349 t3_add_attach_sysctls(adapter_t *sc) 3350 { 3351 struct sysctl_ctx_list *ctx; 3352 struct sysctl_oid_list *children; 3353 3354 ctx = device_get_sysctl_ctx(sc->dev); 3355 children = SYSCTL_CHILDREN(device_get_sysctl_tree(sc->dev)); 3356 3357 /* random information */ 3358 SYSCTL_ADD_STRING(ctx, children, OID_AUTO, 3359 "firmware_version", 3360 CTLFLAG_RD, sc->fw_version, 3361 0, "firmware version"); 3362 SYSCTL_ADD_UINT(ctx, children, OID_AUTO, 3363 "hw_revision", 3364 CTLFLAG_RD, &sc->params.rev, 3365 0, "chip model"); 3366 SYSCTL_ADD_STRING(ctx, children, OID_AUTO, 3367 "port_types", 3368 CTLFLAG_RD, sc->port_types, 3369 0, "type of ports"); 3370 SYSCTL_ADD_INT(ctx, children, OID_AUTO, 3371 "enable_debug", 3372 CTLFLAG_RW, &cxgb_debug, 3373 0, "enable verbose debugging output"); 3374 SYSCTL_ADD_UQUAD(ctx, children, OID_AUTO, "tunq_coalesce", 3375 CTLFLAG_RD, &sc->tunq_coalesce, 3376 "#tunneled packets freed"); 3377 SYSCTL_ADD_INT(ctx, children, OID_AUTO, 3378 "txq_overrun", 3379 CTLFLAG_RD, &txq_fills, 3380 0, "#times txq overrun"); 3381 SYSCTL_ADD_UINT(ctx, children, OID_AUTO, 3382 "core_clock", 3383 CTLFLAG_RD, &sc->params.vpd.cclk, 3384 0, "core clock frequency (in KHz)"); 3385 } 3386 3387 3388 static const char *rspq_name = "rspq"; 3389 static const char *txq_names[] = 3390 { 3391 "txq_eth", 3392 "txq_ofld", 3393 "txq_ctrl" 3394 }; 3395 3396 static int 3397 sysctl_handle_macstat(SYSCTL_HANDLER_ARGS) 3398 { 3399 struct port_info *p = arg1; 3400 uint64_t *parg; 3401 3402 if (!p) 3403 return (EINVAL); 3404 3405 cxgb_refresh_stats(p); 3406 parg = (uint64_t *) ((uint8_t *)&p->mac.stats + arg2); 3407 3408 return (sysctl_handle_64(oidp, parg, 0, req)); 3409 } 3410 3411 void 3412 t3_add_configured_sysctls(adapter_t *sc) 3413 { 3414 struct sysctl_ctx_list *ctx; 3415 struct sysctl_oid_list *children; 3416 int i, j; 3417 3418 ctx = device_get_sysctl_ctx(sc->dev); 3419 children = SYSCTL_CHILDREN(device_get_sysctl_tree(sc->dev)); 3420 3421 SYSCTL_ADD_PROC(ctx, children, OID_AUTO, 3422 "intr_coal", 3423 CTLTYPE_INT|CTLFLAG_RW, sc, 3424 0, t3_set_coalesce_usecs, 3425 "I", "interrupt coalescing timer (us)"); 3426 3427 SYSCTL_ADD_PROC(ctx, children, OID_AUTO, 3428 "pkt_timestamp", 3429 CTLTYPE_INT | CTLFLAG_RW, sc, 3430 0, t3_pkt_timestamp, 3431 "I", "provide packet timestamp instead of connection hash"); 3432 3433 for (i = 0; i < sc->params.nports; i++) { 3434 struct port_info *pi = &sc->port[i]; 3435 struct sysctl_oid *poid; 3436 struct sysctl_oid_list *poidlist; 3437 struct mac_stats *mstats = &pi->mac.stats; 3438 3439 snprintf(pi->namebuf, PORT_NAME_LEN, "port%d", i); 3440 poid = SYSCTL_ADD_NODE(ctx, children, OID_AUTO, 3441 pi->namebuf, CTLFLAG_RD, NULL, "port statistics"); 3442 poidlist = SYSCTL_CHILDREN(poid); 3443 SYSCTL_ADD_UINT(ctx, poidlist, OID_AUTO, 3444 "nqsets", CTLFLAG_RD, &pi->nqsets, 3445 0, "#queue sets"); 3446 3447 for (j = 0; j < pi->nqsets; j++) { 3448 struct sge_qset *qs = &sc->sge.qs[pi->first_qset + j]; 3449 struct sysctl_oid *qspoid, *rspqpoid, *txqpoid, 3450 *ctrlqpoid, *lropoid; 3451 struct sysctl_oid_list *qspoidlist, *rspqpoidlist, 3452 *txqpoidlist, *ctrlqpoidlist, 3453 *lropoidlist; 3454 struct sge_txq *txq = &qs->txq[TXQ_ETH]; 3455 3456 snprintf(qs->namebuf, QS_NAME_LEN, "qs%d", j); 3457 3458 qspoid = SYSCTL_ADD_NODE(ctx, poidlist, OID_AUTO, 3459 qs->namebuf, CTLFLAG_RD, NULL, "qset statistics"); 3460 qspoidlist = SYSCTL_CHILDREN(qspoid); 3461 3462 SYSCTL_ADD_UINT(ctx, qspoidlist, OID_AUTO, "fl0_empty", 3463 CTLFLAG_RD, &qs->fl[0].empty, 0, 3464 "freelist #0 empty"); 3465 SYSCTL_ADD_UINT(ctx, qspoidlist, OID_AUTO, "fl1_empty", 3466 CTLFLAG_RD, &qs->fl[1].empty, 0, 3467 "freelist #1 empty"); 3468 3469 rspqpoid = SYSCTL_ADD_NODE(ctx, qspoidlist, OID_AUTO, 3470 rspq_name, CTLFLAG_RD, NULL, "rspq statistics"); 3471 rspqpoidlist = SYSCTL_CHILDREN(rspqpoid); 3472 3473 txqpoid = SYSCTL_ADD_NODE(ctx, qspoidlist, OID_AUTO, 3474 txq_names[0], CTLFLAG_RD, NULL, "txq statistics"); 3475 txqpoidlist = SYSCTL_CHILDREN(txqpoid); 3476 3477 ctrlqpoid = SYSCTL_ADD_NODE(ctx, qspoidlist, OID_AUTO, 3478 txq_names[2], CTLFLAG_RD, NULL, "ctrlq statistics"); 3479 ctrlqpoidlist = SYSCTL_CHILDREN(ctrlqpoid); 3480 3481 lropoid = SYSCTL_ADD_NODE(ctx, qspoidlist, OID_AUTO, 3482 "lro_stats", CTLFLAG_RD, NULL, "LRO statistics"); 3483 lropoidlist = SYSCTL_CHILDREN(lropoid); 3484 3485 SYSCTL_ADD_UINT(ctx, rspqpoidlist, OID_AUTO, "size", 3486 CTLFLAG_RD, &qs->rspq.size, 3487 0, "#entries in response queue"); 3488 SYSCTL_ADD_UINT(ctx, rspqpoidlist, OID_AUTO, "cidx", 3489 CTLFLAG_RD, &qs->rspq.cidx, 3490 0, "consumer index"); 3491 SYSCTL_ADD_UINT(ctx, rspqpoidlist, OID_AUTO, "credits", 3492 CTLFLAG_RD, &qs->rspq.credits, 3493 0, "#credits"); 3494 SYSCTL_ADD_UINT(ctx, rspqpoidlist, OID_AUTO, "starved", 3495 CTLFLAG_RD, &qs->rspq.starved, 3496 0, "#times starved"); 3497 SYSCTL_ADD_UAUTO(ctx, rspqpoidlist, OID_AUTO, "phys_addr", 3498 CTLFLAG_RD, &qs->rspq.phys_addr, 3499 "physical_address_of the queue"); 3500 SYSCTL_ADD_UINT(ctx, rspqpoidlist, OID_AUTO, "dump_start", 3501 CTLFLAG_RW, &qs->rspq.rspq_dump_start, 3502 0, "start rspq dump entry"); 3503 SYSCTL_ADD_UINT(ctx, rspqpoidlist, OID_AUTO, "dump_count", 3504 CTLFLAG_RW, &qs->rspq.rspq_dump_count, 3505 0, "#rspq entries to dump"); 3506 SYSCTL_ADD_PROC(ctx, rspqpoidlist, OID_AUTO, "qdump", 3507 CTLTYPE_STRING | CTLFLAG_RD, &qs->rspq, 3508 0, t3_dump_rspq, "A", "dump of the response queue"); 3509 3510 SYSCTL_ADD_UQUAD(ctx, txqpoidlist, OID_AUTO, "dropped", 3511 CTLFLAG_RD, &qs->txq[TXQ_ETH].txq_mr->br_drops, 3512 "#tunneled packets dropped"); 3513 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "sendqlen", 3514 CTLFLAG_RD, &qs->txq[TXQ_ETH].sendq.mq_len, 3515 0, "#tunneled packets waiting to be sent"); 3516 #if 0 3517 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "queue_pidx", 3518 CTLFLAG_RD, (uint32_t *)(uintptr_t)&qs->txq[TXQ_ETH].txq_mr.br_prod, 3519 0, "#tunneled packets queue producer index"); 3520 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "queue_cidx", 3521 CTLFLAG_RD, (uint32_t *)(uintptr_t)&qs->txq[TXQ_ETH].txq_mr.br_cons, 3522 0, "#tunneled packets queue consumer index"); 3523 #endif 3524 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "processed", 3525 CTLFLAG_RD, &qs->txq[TXQ_ETH].processed, 3526 0, "#tunneled packets processed by the card"); 3527 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "cleaned", 3528 CTLFLAG_RD, &txq->cleaned, 3529 0, "#tunneled packets cleaned"); 3530 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "in_use", 3531 CTLFLAG_RD, &txq->in_use, 3532 0, "#tunneled packet slots in use"); 3533 SYSCTL_ADD_UQUAD(ctx, txqpoidlist, OID_AUTO, "frees", 3534 CTLFLAG_RD, &txq->txq_frees, 3535 "#tunneled packets freed"); 3536 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "skipped", 3537 CTLFLAG_RD, &txq->txq_skipped, 3538 0, "#tunneled packet descriptors skipped"); 3539 SYSCTL_ADD_UQUAD(ctx, txqpoidlist, OID_AUTO, "coalesced", 3540 CTLFLAG_RD, &txq->txq_coalesced, 3541 "#tunneled packets coalesced"); 3542 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "enqueued", 3543 CTLFLAG_RD, &txq->txq_enqueued, 3544 0, "#tunneled packets enqueued to hardware"); 3545 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "stopped_flags", 3546 CTLFLAG_RD, &qs->txq_stopped, 3547 0, "tx queues stopped"); 3548 SYSCTL_ADD_UAUTO(ctx, txqpoidlist, OID_AUTO, "phys_addr", 3549 CTLFLAG_RD, &txq->phys_addr, 3550 "physical_address_of the queue"); 3551 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "qgen", 3552 CTLFLAG_RW, &qs->txq[TXQ_ETH].gen, 3553 0, "txq generation"); 3554 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "hw_cidx", 3555 CTLFLAG_RD, &txq->cidx, 3556 0, "hardware queue cidx"); 3557 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "hw_pidx", 3558 CTLFLAG_RD, &txq->pidx, 3559 0, "hardware queue pidx"); 3560 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "dump_start", 3561 CTLFLAG_RW, &qs->txq[TXQ_ETH].txq_dump_start, 3562 0, "txq start idx for dump"); 3563 SYSCTL_ADD_UINT(ctx, txqpoidlist, OID_AUTO, "dump_count", 3564 CTLFLAG_RW, &qs->txq[TXQ_ETH].txq_dump_count, 3565 0, "txq #entries to dump"); 3566 SYSCTL_ADD_PROC(ctx, txqpoidlist, OID_AUTO, "qdump", 3567 CTLTYPE_STRING | CTLFLAG_RD, &qs->txq[TXQ_ETH], 3568 0, t3_dump_txq_eth, "A", "dump of the transmit queue"); 3569 3570 SYSCTL_ADD_UINT(ctx, ctrlqpoidlist, OID_AUTO, "dump_start", 3571 CTLFLAG_RW, &qs->txq[TXQ_CTRL].txq_dump_start, 3572 0, "ctrlq start idx for dump"); 3573 SYSCTL_ADD_UINT(ctx, ctrlqpoidlist, OID_AUTO, "dump_count", 3574 CTLFLAG_RW, &qs->txq[TXQ_CTRL].txq_dump_count, 3575 0, "ctrl #entries to dump"); 3576 SYSCTL_ADD_PROC(ctx, ctrlqpoidlist, OID_AUTO, "qdump", 3577 CTLTYPE_STRING | CTLFLAG_RD, &qs->txq[TXQ_CTRL], 3578 0, t3_dump_txq_ctrl, "A", "dump of the transmit queue"); 3579 3580 SYSCTL_ADD_U64(ctx, lropoidlist, OID_AUTO, "lro_queued", 3581 CTLFLAG_RD, &qs->lro.ctrl.lro_queued, 0, NULL); 3582 SYSCTL_ADD_U64(ctx, lropoidlist, OID_AUTO, "lro_flushed", 3583 CTLFLAG_RD, &qs->lro.ctrl.lro_flushed, 0, NULL); 3584 SYSCTL_ADD_U64(ctx, lropoidlist, OID_AUTO, "lro_bad_csum", 3585 CTLFLAG_RD, &qs->lro.ctrl.lro_bad_csum, 0, NULL); 3586 SYSCTL_ADD_INT(ctx, lropoidlist, OID_AUTO, "lro_cnt", 3587 CTLFLAG_RD, &qs->lro.ctrl.lro_cnt, 0, NULL); 3588 } 3589 3590 /* Now add a node for mac stats. */ 3591 poid = SYSCTL_ADD_NODE(ctx, poidlist, OID_AUTO, "mac_stats", 3592 CTLFLAG_RD, NULL, "MAC statistics"); 3593 poidlist = SYSCTL_CHILDREN(poid); 3594 3595 /* 3596 * We (ab)use the length argument (arg2) to pass on the offset 3597 * of the data that we are interested in. This is only required 3598 * for the quad counters that are updated from the hardware (we 3599 * make sure that we return the latest value). 3600 * sysctl_handle_macstat first updates *all* the counters from 3601 * the hardware, and then returns the latest value of the 3602 * requested counter. Best would be to update only the 3603 * requested counter from hardware, but t3_mac_update_stats() 3604 * hides all the register details and we don't want to dive into 3605 * all that here. 3606 */ 3607 #define CXGB_SYSCTL_ADD_QUAD(a) SYSCTL_ADD_OID(ctx, poidlist, OID_AUTO, #a, \ 3608 (CTLTYPE_U64 | CTLFLAG_RD), pi, offsetof(struct mac_stats, a), \ 3609 sysctl_handle_macstat, "QU", 0) 3610 CXGB_SYSCTL_ADD_QUAD(tx_octets); 3611 CXGB_SYSCTL_ADD_QUAD(tx_octets_bad); 3612 CXGB_SYSCTL_ADD_QUAD(tx_frames); 3613 CXGB_SYSCTL_ADD_QUAD(tx_mcast_frames); 3614 CXGB_SYSCTL_ADD_QUAD(tx_bcast_frames); 3615 CXGB_SYSCTL_ADD_QUAD(tx_pause); 3616 CXGB_SYSCTL_ADD_QUAD(tx_deferred); 3617 CXGB_SYSCTL_ADD_QUAD(tx_late_collisions); 3618 CXGB_SYSCTL_ADD_QUAD(tx_total_collisions); 3619 CXGB_SYSCTL_ADD_QUAD(tx_excess_collisions); 3620 CXGB_SYSCTL_ADD_QUAD(tx_underrun); 3621 CXGB_SYSCTL_ADD_QUAD(tx_len_errs); 3622 CXGB_SYSCTL_ADD_QUAD(tx_mac_internal_errs); 3623 CXGB_SYSCTL_ADD_QUAD(tx_excess_deferral); 3624 CXGB_SYSCTL_ADD_QUAD(tx_fcs_errs); 3625 CXGB_SYSCTL_ADD_QUAD(tx_frames_64); 3626 CXGB_SYSCTL_ADD_QUAD(tx_frames_65_127); 3627 CXGB_SYSCTL_ADD_QUAD(tx_frames_128_255); 3628 CXGB_SYSCTL_ADD_QUAD(tx_frames_256_511); 3629 CXGB_SYSCTL_ADD_QUAD(tx_frames_512_1023); 3630 CXGB_SYSCTL_ADD_QUAD(tx_frames_1024_1518); 3631 CXGB_SYSCTL_ADD_QUAD(tx_frames_1519_max); 3632 CXGB_SYSCTL_ADD_QUAD(rx_octets); 3633 CXGB_SYSCTL_ADD_QUAD(rx_octets_bad); 3634 CXGB_SYSCTL_ADD_QUAD(rx_frames); 3635 CXGB_SYSCTL_ADD_QUAD(rx_mcast_frames); 3636 CXGB_SYSCTL_ADD_QUAD(rx_bcast_frames); 3637 CXGB_SYSCTL_ADD_QUAD(rx_pause); 3638 CXGB_SYSCTL_ADD_QUAD(rx_fcs_errs); 3639 CXGB_SYSCTL_ADD_QUAD(rx_align_errs); 3640 CXGB_SYSCTL_ADD_QUAD(rx_symbol_errs); 3641 CXGB_SYSCTL_ADD_QUAD(rx_data_errs); 3642 CXGB_SYSCTL_ADD_QUAD(rx_sequence_errs); 3643 CXGB_SYSCTL_ADD_QUAD(rx_runt); 3644 CXGB_SYSCTL_ADD_QUAD(rx_jabber); 3645 CXGB_SYSCTL_ADD_QUAD(rx_short); 3646 CXGB_SYSCTL_ADD_QUAD(rx_too_long); 3647 CXGB_SYSCTL_ADD_QUAD(rx_mac_internal_errs); 3648 CXGB_SYSCTL_ADD_QUAD(rx_cong_drops); 3649 CXGB_SYSCTL_ADD_QUAD(rx_frames_64); 3650 CXGB_SYSCTL_ADD_QUAD(rx_frames_65_127); 3651 CXGB_SYSCTL_ADD_QUAD(rx_frames_128_255); 3652 CXGB_SYSCTL_ADD_QUAD(rx_frames_256_511); 3653 CXGB_SYSCTL_ADD_QUAD(rx_frames_512_1023); 3654 CXGB_SYSCTL_ADD_QUAD(rx_frames_1024_1518); 3655 CXGB_SYSCTL_ADD_QUAD(rx_frames_1519_max); 3656 #undef CXGB_SYSCTL_ADD_QUAD 3657 3658 #define CXGB_SYSCTL_ADD_ULONG(a) SYSCTL_ADD_ULONG(ctx, poidlist, OID_AUTO, #a, \ 3659 CTLFLAG_RD, &mstats->a, 0) 3660 CXGB_SYSCTL_ADD_ULONG(tx_fifo_parity_err); 3661 CXGB_SYSCTL_ADD_ULONG(rx_fifo_parity_err); 3662 CXGB_SYSCTL_ADD_ULONG(tx_fifo_urun); 3663 CXGB_SYSCTL_ADD_ULONG(rx_fifo_ovfl); 3664 CXGB_SYSCTL_ADD_ULONG(serdes_signal_loss); 3665 CXGB_SYSCTL_ADD_ULONG(xaui_pcs_ctc_err); 3666 CXGB_SYSCTL_ADD_ULONG(xaui_pcs_align_change); 3667 CXGB_SYSCTL_ADD_ULONG(num_toggled); 3668 CXGB_SYSCTL_ADD_ULONG(num_resets); 3669 CXGB_SYSCTL_ADD_ULONG(link_faults); 3670 #undef CXGB_SYSCTL_ADD_ULONG 3671 } 3672 } 3673 3674 /** 3675 * t3_get_desc - dump an SGE descriptor for debugging purposes 3676 * @qs: the queue set 3677 * @qnum: identifies the specific queue (0..2: Tx, 3:response, 4..5: Rx) 3678 * @idx: the descriptor index in the queue 3679 * @data: where to dump the descriptor contents 3680 * 3681 * Dumps the contents of a HW descriptor of an SGE queue. Returns the 3682 * size of the descriptor. 3683 */ 3684 int 3685 t3_get_desc(const struct sge_qset *qs, unsigned int qnum, unsigned int idx, 3686 unsigned char *data) 3687 { 3688 if (qnum >= 6) 3689 return (EINVAL); 3690 3691 if (qnum < 3) { 3692 if (!qs->txq[qnum].desc || idx >= qs->txq[qnum].size) 3693 return -EINVAL; 3694 memcpy(data, &qs->txq[qnum].desc[idx], sizeof(struct tx_desc)); 3695 return sizeof(struct tx_desc); 3696 } 3697 3698 if (qnum == 3) { 3699 if (!qs->rspq.desc || idx >= qs->rspq.size) 3700 return (EINVAL); 3701 memcpy(data, &qs->rspq.desc[idx], sizeof(struct rsp_desc)); 3702 return sizeof(struct rsp_desc); 3703 } 3704 3705 qnum -= 4; 3706 if (!qs->fl[qnum].desc || idx >= qs->fl[qnum].size) 3707 return (EINVAL); 3708 memcpy(data, &qs->fl[qnum].desc[idx], sizeof(struct rx_desc)); 3709 return sizeof(struct rx_desc); 3710 } 3711