1/* Copyright (c) 2013 Coraid, Inc. See COPYING for GPL terms. */ 2/* 3 * aoecmd.c 4 * Filesystem request handling methods 5 */ 6 7#include <linux/ata.h> 8#include <linux/slab.h> 9#include <linux/hdreg.h> 10#include <linux/blk-mq.h> 11#include <linux/skbuff.h> 12#include <linux/netdevice.h> 13#include <linux/moduleparam.h> 14#include <linux/workqueue.h> 15#include <linux/kthread.h> 16#include <net/net_namespace.h> 17#include <asm/unaligned.h> 18#include <linux/uio.h> 19#include "aoe.h" 20 21#define MAXIOC (8192) /* default meant to avoid most soft lockups */ 22 23static void ktcomplete(struct frame *, struct sk_buff *); 24static int count_targets(struct aoedev *d, int *untainted); 25 26static struct buf *nextbuf(struct aoedev *); 27 28static int aoe_deadsecs = 60 * 3; 29module_param(aoe_deadsecs, int, 0644); 30MODULE_PARM_DESC(aoe_deadsecs, "After aoe_deadsecs seconds, give up and fail dev."); 31 32static int aoe_maxout = 64; 33module_param(aoe_maxout, int, 0644); 34MODULE_PARM_DESC(aoe_maxout, 35 "Only aoe_maxout outstanding packets for every MAC on eX.Y."); 36 37/* The number of online cpus during module initialization gives us a 38 * convenient heuristic cap on the parallelism used for ktio threads 39 * doing I/O completion. It is not important that the cap equal the 40 * actual number of running CPUs at any given time, but because of CPU 41 * hotplug, we take care to use ncpus instead of using 42 * num_online_cpus() after module initialization. 43 */ 44static int ncpus; 45 46/* mutex lock used for synchronization while thread spawning */ 47static DEFINE_MUTEX(ktio_spawn_lock); 48 49static wait_queue_head_t *ktiowq; 50static struct ktstate *kts; 51 52/* io completion queue */ 53struct iocq_ktio { 54 struct list_head head; 55 spinlock_t lock; 56}; 57static struct iocq_ktio *iocq; 58 59static struct page *empty_page; 60 61static struct sk_buff * 62new_skb(ulong len) 63{ 64 struct sk_buff *skb; 65 66 skb = alloc_skb(len + MAX_HEADER, GFP_ATOMIC); 67 if (skb) { 68 skb_reserve(skb, MAX_HEADER); 69 skb_reset_mac_header(skb); 70 skb_reset_network_header(skb); 71 skb->protocol = __constant_htons(ETH_P_AOE); 72 skb_checksum_none_assert(skb); 73 } 74 return skb; 75} 76 77static struct frame * 78getframe_deferred(struct aoedev *d, u32 tag) 79{ 80 struct list_head *head, *pos, *nx; 81 struct frame *f; 82 83 head = &d->rexmitq; 84 list_for_each_safe(pos, nx, head) { 85 f = list_entry(pos, struct frame, head); 86 if (f->tag == tag) { 87 list_del(pos); 88 return f; 89 } 90 } 91 return NULL; 92} 93 94static struct frame * 95getframe(struct aoedev *d, u32 tag) 96{ 97 struct frame *f; 98 struct list_head *head, *pos, *nx; 99 u32 n; 100 101 n = tag % NFACTIVE; 102 head = &d->factive[n]; 103 list_for_each_safe(pos, nx, head) { 104 f = list_entry(pos, struct frame, head); 105 if (f->tag == tag) { 106 list_del(pos); 107 return f; 108 } 109 } 110 return NULL; 111} 112 113/* 114 * Leave the top bit clear so we have tagspace for userland. 115 * The bottom 16 bits are the xmit tick for rexmit/rttavg processing. 116 * This driver reserves tag -1 to mean "unused frame." 117 */ 118static int 119newtag(struct aoedev *d) 120{ 121 register ulong n; 122 123 n = jiffies & 0xffff; 124 return n | (++d->lasttag & 0x7fff) << 16; 125} 126 127static u32 128aoehdr_atainit(struct aoedev *d, struct aoetgt *t, struct aoe_hdr *h) 129{ 130 u32 host_tag = newtag(d); 131 132 memcpy(h->src, t->ifp->nd->dev_addr, sizeof h->src); 133 memcpy(h->dst, t->addr, sizeof h->dst); 134 h->type = __constant_cpu_to_be16(ETH_P_AOE); 135 h->verfl = AOE_HVER; 136 h->major = cpu_to_be16(d->aoemajor); 137 h->minor = d->aoeminor; 138 h->cmd = AOECMD_ATA; 139 h->tag = cpu_to_be32(host_tag); 140 141 return host_tag; 142} 143 144static inline void 145put_lba(struct aoe_atahdr *ah, sector_t lba) 146{ 147 ah->lba0 = lba; 148 ah->lba1 = lba >>= 8; 149 ah->lba2 = lba >>= 8; 150 ah->lba3 = lba >>= 8; 151 ah->lba4 = lba >>= 8; 152 ah->lba5 = lba >>= 8; 153} 154 155static struct aoeif * 156ifrotate(struct aoetgt *t) 157{ 158 struct aoeif *ifp; 159 160 ifp = t->ifp; 161 ifp++; 162 if (ifp >= &t->ifs[NAOEIFS] || ifp->nd == NULL) 163 ifp = t->ifs; 164 if (ifp->nd == NULL) 165 return NULL; 166 return t->ifp = ifp; 167} 168 169static void 170skb_pool_put(struct aoedev *d, struct sk_buff *skb) 171{ 172 __skb_queue_tail(&d->skbpool, skb); 173} 174 175static struct sk_buff * 176skb_pool_get(struct aoedev *d) 177{ 178 struct sk_buff *skb = skb_peek(&d->skbpool); 179 180 if (skb && atomic_read(&skb_shinfo(skb)->dataref) == 1) { 181 __skb_unlink(skb, &d->skbpool); 182 return skb; 183 } 184 if (skb_queue_len(&d->skbpool) < NSKBPOOLMAX && 185 (skb = new_skb(ETH_ZLEN))) 186 return skb; 187 188 return NULL; 189} 190 191void 192aoe_freetframe(struct frame *f) 193{ 194 struct aoetgt *t; 195 196 t = f->t; 197 f->buf = NULL; 198 memset(&f->iter, 0, sizeof(f->iter)); 199 f->r_skb = NULL; 200 f->flags = 0; 201 list_add(&f->head, &t->ffree); 202} 203 204static struct frame * 205newtframe(struct aoedev *d, struct aoetgt *t) 206{ 207 struct frame *f; 208 struct sk_buff *skb; 209 struct list_head *pos; 210 211 if (list_empty(&t->ffree)) { 212 if (t->falloc >= NSKBPOOLMAX*2) 213 return NULL; 214 f = kcalloc(1, sizeof(*f), GFP_ATOMIC); 215 if (f == NULL) 216 return NULL; 217 t->falloc++; 218 f->t = t; 219 } else { 220 pos = t->ffree.next; 221 list_del(pos); 222 f = list_entry(pos, struct frame, head); 223 } 224 225 skb = f->skb; 226 if (skb == NULL) { 227 f->skb = skb = new_skb(ETH_ZLEN); 228 if (!skb) { 229bail: aoe_freetframe(f); 230 return NULL; 231 } 232 } 233 234 if (atomic_read(&skb_shinfo(skb)->dataref) != 1) { 235 skb = skb_pool_get(d); 236 if (skb == NULL) 237 goto bail; 238 skb_pool_put(d, f->skb); 239 f->skb = skb; 240 } 241 242 skb->truesize -= skb->data_len; 243 skb_shinfo(skb)->nr_frags = skb->data_len = 0; 244 skb_trim(skb, 0); 245 return f; 246} 247 248static struct frame * 249newframe(struct aoedev *d) 250{ 251 struct frame *f; 252 struct aoetgt *t, **tt; 253 int totout = 0; 254 int use_tainted; 255 int has_untainted; 256 257 if (!d->targets || !d->targets[0]) { 258 printk(KERN_ERR "aoe: NULL TARGETS!\n"); 259 return NULL; 260 } 261 tt = d->tgt; /* last used target */ 262 for (use_tainted = 0, has_untainted = 0;;) { 263 tt++; 264 if (tt >= &d->targets[d->ntargets] || !*tt) 265 tt = d->targets; 266 t = *tt; 267 if (!t->taint) { 268 has_untainted = 1; 269 totout += t->nout; 270 } 271 if (t->nout < t->maxout 272 && (use_tainted || !t->taint) 273 && t->ifp->nd) { 274 f = newtframe(d, t); 275 if (f) { 276 ifrotate(t); 277 d->tgt = tt; 278 return f; 279 } 280 } 281 if (tt == d->tgt) { /* we've looped and found nada */ 282 if (!use_tainted && !has_untainted) 283 use_tainted = 1; 284 else 285 break; 286 } 287 } 288 if (totout == 0) { 289 d->kicked++; 290 d->flags |= DEVFL_KICKME; 291 } 292 return NULL; 293} 294 295static void 296skb_fillup(struct sk_buff *skb, struct bio *bio, struct bvec_iter iter) 297{ 298 int frag = 0; 299 struct bio_vec bv; 300 301 __bio_for_each_segment(bv, bio, iter, iter) 302 skb_fill_page_desc(skb, frag++, bv.bv_page, 303 bv.bv_offset, bv.bv_len); 304} 305 306static void 307fhash(struct frame *f) 308{ 309 struct aoedev *d = f->t->d; 310 u32 n; 311 312 n = f->tag % NFACTIVE; 313 list_add_tail(&f->head, &d->factive[n]); 314} 315 316static void 317ata_rw_frameinit(struct frame *f) 318{ 319 struct aoetgt *t; 320 struct aoe_hdr *h; 321 struct aoe_atahdr *ah; 322 struct sk_buff *skb; 323 char writebit, extbit; 324 325 skb = f->skb; 326 h = (struct aoe_hdr *) skb_mac_header(skb); 327 ah = (struct aoe_atahdr *) (h + 1); 328 skb_put(skb, sizeof(*h) + sizeof(*ah)); 329 memset(h, 0, skb->len); 330 331 writebit = 0x10; 332 extbit = 0x4; 333 334 t = f->t; 335 f->tag = aoehdr_atainit(t->d, t, h); 336 fhash(f); 337 t->nout++; 338 f->waited = 0; 339 f->waited_total = 0; 340 341 /* set up ata header */ 342 ah->scnt = f->iter.bi_size >> 9; 343 put_lba(ah, f->iter.bi_sector); 344 if (t->d->flags & DEVFL_EXT) { 345 ah->aflags |= AOEAFL_EXT; 346 } else { 347 extbit = 0; 348 ah->lba3 &= 0x0f; 349 ah->lba3 |= 0xe0; /* LBA bit + obsolete 0xa0 */ 350 } 351 if (f->buf && bio_data_dir(f->buf->bio) == WRITE) { 352 skb_fillup(skb, f->buf->bio, f->iter); 353 ah->aflags |= AOEAFL_WRITE; 354 skb->len += f->iter.bi_size; 355 skb->data_len = f->iter.bi_size; 356 skb->truesize += f->iter.bi_size; 357 t->wpkts++; 358 } else { 359 t->rpkts++; 360 writebit = 0; 361 } 362 363 ah->cmdstat = ATA_CMD_PIO_READ | writebit | extbit; 364 skb->dev = t->ifp->nd; 365} 366 367static int 368aoecmd_ata_rw(struct aoedev *d) 369{ 370 struct frame *f; 371 struct buf *buf; 372 struct sk_buff *skb; 373 struct sk_buff_head queue; 374 375 buf = nextbuf(d); 376 if (buf == NULL) 377 return 0; 378 f = newframe(d); 379 if (f == NULL) 380 return 0; 381 382 /* initialize the headers & frame */ 383 f->buf = buf; 384 f->iter = buf->iter; 385 f->iter.bi_size = min_t(unsigned long, 386 d->maxbcnt ?: DEFAULTBCNT, 387 f->iter.bi_size); 388 bio_advance_iter(buf->bio, &buf->iter, f->iter.bi_size); 389 390 if (!buf->iter.bi_size) 391 d->ip.buf = NULL; 392 393 /* mark all tracking fields and load out */ 394 buf->nframesout += 1; 395 396 ata_rw_frameinit(f); 397 398 skb = skb_clone(f->skb, GFP_ATOMIC); 399 if (skb) { 400 f->sent = ktime_get(); 401 __skb_queue_head_init(&queue); 402 __skb_queue_tail(&queue, skb); 403 aoenet_xmit(&queue); 404 } 405 return 1; 406} 407 408/* some callers cannot sleep, and they can call this function, 409 * transmitting the packets later, when interrupts are on 410 */ 411static void 412aoecmd_cfg_pkts(ushort aoemajor, unsigned char aoeminor, struct sk_buff_head *queue) 413{ 414 struct aoe_hdr *h; 415 struct aoe_cfghdr *ch; 416 struct sk_buff *skb; 417 struct net_device *ifp; 418 419 rcu_read_lock(); 420 for_each_netdev_rcu(&init_net, ifp) { 421 dev_hold(ifp); 422 if (!is_aoe_netif(ifp)) { 423 dev_put(ifp); 424 continue; 425 } 426 427 skb = new_skb(sizeof *h + sizeof *ch); 428 if (skb == NULL) { 429 printk(KERN_INFO "aoe: skb alloc failure\n"); 430 dev_put(ifp); 431 continue; 432 } 433 skb_put(skb, sizeof *h + sizeof *ch); 434 skb->dev = ifp; 435 __skb_queue_tail(queue, skb); 436 h = (struct aoe_hdr *) skb_mac_header(skb); 437 memset(h, 0, sizeof *h + sizeof *ch); 438 439 memset(h->dst, 0xff, sizeof h->dst); 440 memcpy(h->src, ifp->dev_addr, sizeof h->src); 441 h->type = __constant_cpu_to_be16(ETH_P_AOE); 442 h->verfl = AOE_HVER; 443 h->major = cpu_to_be16(aoemajor); 444 h->minor = aoeminor; 445 h->cmd = AOECMD_CFG; 446 } 447 rcu_read_unlock(); 448} 449 450static void 451resend(struct aoedev *d, struct frame *f) 452{ 453 struct sk_buff *skb; 454 struct sk_buff_head queue; 455 struct aoe_hdr *h; 456 struct aoetgt *t; 457 char buf[128]; 458 u32 n; 459 460 t = f->t; 461 n = newtag(d); 462 skb = f->skb; 463 if (ifrotate(t) == NULL) { 464 /* probably can't happen, but set it up to fail anyway */ 465 pr_info("aoe: resend: no interfaces to rotate to.\n"); 466 ktcomplete(f, NULL); 467 return; 468 } 469 h = (struct aoe_hdr *) skb_mac_header(skb); 470 471 if (!(f->flags & FFL_PROBE)) { 472 snprintf(buf, sizeof(buf), 473 "%15s e%ld.%d oldtag=%08x@%08lx newtag=%08x s=%pm d=%pm nout=%d\n", 474 "retransmit", d->aoemajor, d->aoeminor, 475 f->tag, jiffies, n, 476 h->src, h->dst, t->nout); 477 aoechr_error(buf); 478 } 479 480 f->tag = n; 481 fhash(f); 482 h->tag = cpu_to_be32(n); 483 memcpy(h->dst, t->addr, sizeof h->dst); 484 memcpy(h->src, t->ifp->nd->dev_addr, sizeof h->src); 485 486 skb->dev = t->ifp->nd; 487 skb = skb_clone(skb, GFP_ATOMIC); 488 if (skb == NULL) 489 return; 490 f->sent = ktime_get(); 491 __skb_queue_head_init(&queue); 492 __skb_queue_tail(&queue, skb); 493 aoenet_xmit(&queue); 494} 495 496static int 497tsince_hr(struct frame *f) 498{ 499 u64 delta = ktime_to_ns(ktime_sub(ktime_get(), f->sent)); 500 501 /* delta is normally under 4.2 seconds, avoid 64-bit division */ 502 if (likely(delta <= UINT_MAX)) 503 return (u32)delta / NSEC_PER_USEC; 504 505 /* avoid overflow after 71 minutes */ 506 if (delta > ((u64)INT_MAX * NSEC_PER_USEC)) 507 return INT_MAX; 508 509 return div_u64(delta, NSEC_PER_USEC); 510} 511 512static int 513tsince(u32 tag) 514{ 515 int n; 516 517 n = jiffies & 0xffff; 518 n -= tag & 0xffff; 519 if (n < 0) 520 n += 1<<16; 521 return jiffies_to_usecs(n + 1); 522} 523 524static struct aoeif * 525getif(struct aoetgt *t, struct net_device *nd) 526{ 527 struct aoeif *p, *e; 528 529 p = t->ifs; 530 e = p + NAOEIFS; 531 for (; p < e; p++) 532 if (p->nd == nd) 533 return p; 534 return NULL; 535} 536 537static void 538ejectif(struct aoetgt *t, struct aoeif *ifp) 539{ 540 struct aoeif *e; 541 struct net_device *nd; 542 ulong n; 543 544 nd = ifp->nd; 545 e = t->ifs + NAOEIFS - 1; 546 n = (e - ifp) * sizeof *ifp; 547 memmove(ifp, ifp+1, n); 548 e->nd = NULL; 549 dev_put(nd); 550} 551 552static struct frame * 553reassign_frame(struct frame *f) 554{ 555 struct frame *nf; 556 struct sk_buff *skb; 557 558 nf = newframe(f->t->d); 559 if (!nf) 560 return NULL; 561 if (nf->t == f->t) { 562 aoe_freetframe(nf); 563 return NULL; 564 } 565 566 skb = nf->skb; 567 nf->skb = f->skb; 568 nf->buf = f->buf; 569 nf->iter = f->iter; 570 nf->waited = 0; 571 nf->waited_total = f->waited_total; 572 nf->sent = f->sent; 573 f->skb = skb; 574 575 return nf; 576} 577 578static void 579probe(struct aoetgt *t) 580{ 581 struct aoedev *d; 582 struct frame *f; 583 struct sk_buff *skb; 584 struct sk_buff_head queue; 585 size_t n, m; 586 int frag; 587 588 d = t->d; 589 f = newtframe(d, t); 590 if (!f) { 591 pr_err("%s %pm for e%ld.%d: %s\n", 592 "aoe: cannot probe remote address", 593 t->addr, 594 (long) d->aoemajor, d->aoeminor, 595 "no frame available"); 596 return; 597 } 598 f->flags |= FFL_PROBE; 599 ifrotate(t); 600 f->iter.bi_size = t->d->maxbcnt ? t->d->maxbcnt : DEFAULTBCNT; 601 ata_rw_frameinit(f); 602 skb = f->skb; 603 for (frag = 0, n = f->iter.bi_size; n > 0; ++frag, n -= m) { 604 if (n < PAGE_SIZE) 605 m = n; 606 else 607 m = PAGE_SIZE; 608 skb_fill_page_desc(skb, frag, empty_page, 0, m); 609 } 610 skb->len += f->iter.bi_size; 611 skb->data_len = f->iter.bi_size; 612 skb->truesize += f->iter.bi_size; 613 614 skb = skb_clone(f->skb, GFP_ATOMIC); 615 if (skb) { 616 f->sent = ktime_get(); 617 __skb_queue_head_init(&queue); 618 __skb_queue_tail(&queue, skb); 619 aoenet_xmit(&queue); 620 } 621} 622 623static long 624rto(struct aoedev *d) 625{ 626 long t; 627 628 t = 2 * d->rttavg >> RTTSCALE; 629 t += 8 * d->rttdev >> RTTDSCALE; 630 if (t == 0) 631 t = 1; 632 633 return t; 634} 635 636static void 637rexmit_deferred(struct aoedev *d) 638{ 639 struct aoetgt *t; 640 struct frame *f; 641 struct frame *nf; 642 struct list_head *pos, *nx, *head; 643 int since; 644 int untainted; 645 646 count_targets(d, &untainted); 647 648 head = &d->rexmitq; 649 list_for_each_safe(pos, nx, head) { 650 f = list_entry(pos, struct frame, head); 651 t = f->t; 652 if (t->taint) { 653 if (!(f->flags & FFL_PROBE)) { 654 nf = reassign_frame(f); 655 if (nf) { 656 if (t->nout_probes == 0 657 && untainted > 0) { 658 probe(t); 659 t->nout_probes++; 660 } 661 list_replace(&f->head, &nf->head); 662 pos = &nf->head; 663 aoe_freetframe(f); 664 f = nf; 665 t = f->t; 666 } 667 } else if (untainted < 1) { 668 /* don't probe w/o other untainted aoetgts */ 669 goto stop_probe; 670 } else if (tsince_hr(f) < t->taint * rto(d)) { 671 /* reprobe slowly when taint is high */ 672 continue; 673 } 674 } else if (f->flags & FFL_PROBE) { 675stop_probe: /* don't probe untainted aoetgts */ 676 list_del(pos); 677 aoe_freetframe(f); 678 /* leaving d->kicked, because this is routine */ 679 f->t->d->flags |= DEVFL_KICKME; 680 continue; 681 } 682 if (t->nout >= t->maxout) 683 continue; 684 list_del(pos); 685 t->nout++; 686 if (f->flags & FFL_PROBE) 687 t->nout_probes++; 688 since = tsince_hr(f); 689 f->waited += since; 690 f->waited_total += since; 691 resend(d, f); 692 } 693} 694 695/* An aoetgt accumulates demerits quickly, and successful 696 * probing redeems the aoetgt slowly. 697 */ 698static void 699scorn(struct aoetgt *t) 700{ 701 int n; 702 703 n = t->taint++; 704 t->taint += t->taint * 2; 705 if (n > t->taint) 706 t->taint = n; 707 if (t->taint > MAX_TAINT) 708 t->taint = MAX_TAINT; 709} 710 711static int 712count_targets(struct aoedev *d, int *untainted) 713{ 714 int i, good; 715 716 for (i = good = 0; i < d->ntargets && d->targets[i]; ++i) 717 if (d->targets[i]->taint == 0) 718 good++; 719 720 if (untainted) 721 *untainted = good; 722 return i; 723} 724 725static void 726rexmit_timer(struct timer_list *timer) 727{ 728 struct aoedev *d; 729 struct aoetgt *t; 730 struct aoeif *ifp; 731 struct frame *f; 732 struct list_head *head, *pos, *nx; 733 LIST_HEAD(flist); 734 register long timeout; 735 ulong flags, n; 736 int i; 737 int utgts; /* number of aoetgt descriptors (not slots) */ 738 int since; 739 740 d = from_timer(d, timer, timer); 741 742 spin_lock_irqsave(&d->lock, flags); 743 744 /* timeout based on observed timings and variations */ 745 timeout = rto(d); 746 747 utgts = count_targets(d, NULL); 748 749 if (d->flags & DEVFL_TKILL) { 750 spin_unlock_irqrestore(&d->lock, flags); 751 return; 752 } 753 754 /* collect all frames to rexmit into flist */ 755 for (i = 0; i < NFACTIVE; i++) { 756 head = &d->factive[i]; 757 list_for_each_safe(pos, nx, head) { 758 f = list_entry(pos, struct frame, head); 759 if (tsince_hr(f) < timeout) 760 break; /* end of expired frames */ 761 /* move to flist for later processing */ 762 list_move_tail(pos, &flist); 763 } 764 } 765 766 /* process expired frames */ 767 while (!list_empty(&flist)) { 768 pos = flist.next; 769 f = list_entry(pos, struct frame, head); 770 since = tsince_hr(f); 771 n = f->waited_total + since; 772 n /= USEC_PER_SEC; 773 if (aoe_deadsecs 774 && n > aoe_deadsecs 775 && !(f->flags & FFL_PROBE)) { 776 /* Waited too long. Device failure. 777 * Hang all frames on first hash bucket for downdev 778 * to clean up. 779 */ 780 list_splice(&flist, &d->factive[0]); 781 aoedev_downdev(d); 782 goto out; 783 } 784 785 t = f->t; 786 n = f->waited + since; 787 n /= USEC_PER_SEC; 788 if (aoe_deadsecs && utgts > 0 789 && (n > aoe_deadsecs / utgts || n > HARD_SCORN_SECS)) 790 scorn(t); /* avoid this target */ 791 792 if (t->maxout != 1) { 793 t->ssthresh = t->maxout / 2; 794 t->maxout = 1; 795 } 796 797 if (f->flags & FFL_PROBE) { 798 t->nout_probes--; 799 } else { 800 ifp = getif(t, f->skb->dev); 801 if (ifp && ++ifp->lost > (t->nframes << 1) 802 && (ifp != t->ifs || t->ifs[1].nd)) { 803 ejectif(t, ifp); 804 ifp = NULL; 805 } 806 } 807 list_move_tail(pos, &d->rexmitq); 808 t->nout--; 809 } 810 rexmit_deferred(d); 811 812out: 813 if ((d->flags & DEVFL_KICKME) && d->blkq) { 814 d->flags &= ~DEVFL_KICKME; 815 blk_mq_run_hw_queues(d->blkq, true); 816 } 817 818 d->timer.expires = jiffies + TIMERTICK; 819 add_timer(&d->timer); 820 821 spin_unlock_irqrestore(&d->lock, flags); 822} 823 824static void 825bufinit(struct buf *buf, struct request *rq, struct bio *bio) 826{ 827 memset(buf, 0, sizeof(*buf)); 828 buf->rq = rq; 829 buf->bio = bio; 830 buf->iter = bio->bi_iter; 831} 832 833static struct buf * 834nextbuf(struct aoedev *d) 835{ 836 struct request *rq; 837 struct request_queue *q; 838 struct aoe_req *req; 839 struct buf *buf; 840 struct bio *bio; 841 842 q = d->blkq; 843 if (q == NULL) 844 return NULL; /* initializing */ 845 if (d->ip.buf) 846 return d->ip.buf; 847 rq = d->ip.rq; 848 if (rq == NULL) { 849 rq = list_first_entry_or_null(&d->rq_list, struct request, 850 queuelist); 851 if (rq == NULL) 852 return NULL; 853 list_del_init(&rq->queuelist); 854 blk_mq_start_request(rq); 855 d->ip.rq = rq; 856 d->ip.nxbio = rq->bio; 857 858 req = blk_mq_rq_to_pdu(rq); 859 req->nr_bios = 0; 860 __rq_for_each_bio(bio, rq) 861 req->nr_bios++; 862 } 863 buf = mempool_alloc(d->bufpool, GFP_ATOMIC); 864 if (buf == NULL) { 865 pr_err("aoe: nextbuf: unable to mempool_alloc!\n"); 866 return NULL; 867 } 868 bio = d->ip.nxbio; 869 bufinit(buf, rq, bio); 870 bio = bio->bi_next; 871 d->ip.nxbio = bio; 872 if (bio == NULL) 873 d->ip.rq = NULL; 874 return d->ip.buf = buf; 875} 876 877/* enters with d->lock held */ 878void 879aoecmd_work(struct aoedev *d) 880{ 881 rexmit_deferred(d); 882 while (aoecmd_ata_rw(d)) 883 ; 884} 885 886/* this function performs work that has been deferred until sleeping is OK 887 */ 888void 889aoecmd_sleepwork(struct work_struct *work) 890{ 891 struct aoedev *d = container_of(work, struct aoedev, work); 892 893 if (d->flags & DEVFL_GDALLOC) 894 aoeblk_gdalloc(d); 895 896 if (d->flags & DEVFL_NEWSIZE) { 897 set_capacity_and_notify(d->gd, d->ssize); 898 899 spin_lock_irq(&d->lock); 900 d->flags |= DEVFL_UP; 901 d->flags &= ~DEVFL_NEWSIZE; 902 spin_unlock_irq(&d->lock); 903 } 904} 905 906static void 907ata_ident_fixstring(u16 *id, int ns) 908{ 909 u16 s; 910 911 while (ns-- > 0) { 912 s = *id; 913 *id++ = s >> 8 | s << 8; 914 } 915} 916 917static void 918ataid_complete(struct aoedev *d, struct aoetgt *t, unsigned char *id) 919{ 920 u64 ssize; 921 u16 n; 922 923 /* word 83: command set supported */ 924 n = get_unaligned_le16(&id[83 << 1]); 925 926 /* word 86: command set/feature enabled */ 927 n |= get_unaligned_le16(&id[86 << 1]); 928 929 if (n & (1<<10)) { /* bit 10: LBA 48 */ 930 d->flags |= DEVFL_EXT; 931 932 /* word 100: number lba48 sectors */ 933 ssize = get_unaligned_le64(&id[100 << 1]); 934 935 /* set as in ide-disk.c:init_idedisk_capacity */ 936 d->geo.cylinders = ssize; 937 d->geo.cylinders /= (255 * 63); 938 d->geo.heads = 255; 939 d->geo.sectors = 63; 940 } else { 941 d->flags &= ~DEVFL_EXT; 942 943 /* number lba28 sectors */ 944 ssize = get_unaligned_le32(&id[60 << 1]); 945 946 /* NOTE: obsolete in ATA 6 */ 947 d->geo.cylinders = get_unaligned_le16(&id[54 << 1]); 948 d->geo.heads = get_unaligned_le16(&id[55 << 1]); 949 d->geo.sectors = get_unaligned_le16(&id[56 << 1]); 950 } 951 952 ata_ident_fixstring((u16 *) &id[10<<1], 10); /* serial */ 953 ata_ident_fixstring((u16 *) &id[23<<1], 4); /* firmware */ 954 ata_ident_fixstring((u16 *) &id[27<<1], 20); /* model */ 955 memcpy(d->ident, id, sizeof(d->ident)); 956 957 if (d->ssize != ssize) 958 printk(KERN_INFO 959 "aoe: %pm e%ld.%d v%04x has %llu sectors\n", 960 t->addr, 961 d->aoemajor, d->aoeminor, 962 d->fw_ver, (long long)ssize); 963 d->ssize = ssize; 964 d->geo.start = 0; 965 if (d->flags & (DEVFL_GDALLOC|DEVFL_NEWSIZE)) 966 return; 967 if (d->gd != NULL) 968 d->flags |= DEVFL_NEWSIZE; 969 else 970 d->flags |= DEVFL_GDALLOC; 971 queue_work(aoe_wq, &d->work); 972} 973 974static void 975calc_rttavg(struct aoedev *d, struct aoetgt *t, int rtt) 976{ 977 register long n; 978 979 n = rtt; 980 981 /* cf. Congestion Avoidance and Control, Jacobson & Karels, 1988 */ 982 n -= d->rttavg >> RTTSCALE; 983 d->rttavg += n; 984 if (n < 0) 985 n = -n; 986 n -= d->rttdev >> RTTDSCALE; 987 d->rttdev += n; 988 989 if (!t || t->maxout >= t->nframes) 990 return; 991 if (t->maxout < t->ssthresh) 992 t->maxout += 1; 993 else if (t->nout == t->maxout && t->next_cwnd-- == 0) { 994 t->maxout += 1; 995 t->next_cwnd = t->maxout; 996 } 997} 998 999static struct aoetgt * 1000gettgt(struct aoedev *d, char *addr) 1001{ 1002 struct aoetgt **t, **e; 1003 1004 t = d->targets; 1005 e = t + d->ntargets; 1006 for (; t < e && *t; t++) 1007 if (memcmp((*t)->addr, addr, sizeof((*t)->addr)) == 0) 1008 return *t; 1009 return NULL; 1010} 1011 1012static void 1013bvcpy(struct sk_buff *skb, struct bio *bio, struct bvec_iter iter, long cnt) 1014{ 1015 int soff = 0; 1016 struct bio_vec bv; 1017 1018 iter.bi_size = cnt; 1019 1020 __bio_for_each_segment(bv, bio, iter, iter) { 1021 char *p = bvec_kmap_local(&bv); 1022 skb_copy_bits(skb, soff, p, bv.bv_len); 1023 kunmap_local(p); 1024 soff += bv.bv_len; 1025 } 1026} 1027 1028void 1029aoe_end_request(struct aoedev *d, struct request *rq, int fastfail) 1030{ 1031 struct bio *bio; 1032 int bok; 1033 struct request_queue *q; 1034 blk_status_t err = BLK_STS_OK; 1035 1036 q = d->blkq; 1037 if (rq == d->ip.rq) 1038 d->ip.rq = NULL; 1039 do { 1040 bio = rq->bio; 1041 bok = !fastfail && !bio->bi_status; 1042 if (!bok) 1043 err = BLK_STS_IOERR; 1044 } while (blk_update_request(rq, bok ? BLK_STS_OK : BLK_STS_IOERR, bio->bi_iter.bi_size)); 1045 1046 __blk_mq_end_request(rq, err); 1047 1048 /* cf. https://lore.kernel.org/lkml/20061031071040.GS14055@kernel.dk/ */ 1049 if (!fastfail) 1050 blk_mq_run_hw_queues(q, true); 1051} 1052 1053static void 1054aoe_end_buf(struct aoedev *d, struct buf *buf) 1055{ 1056 struct request *rq = buf->rq; 1057 struct aoe_req *req = blk_mq_rq_to_pdu(rq); 1058 1059 if (buf == d->ip.buf) 1060 d->ip.buf = NULL; 1061 mempool_free(buf, d->bufpool); 1062 if (--req->nr_bios == 0) 1063 aoe_end_request(d, rq, 0); 1064} 1065 1066static void 1067ktiocomplete(struct frame *f) 1068{ 1069 struct aoe_hdr *hin, *hout; 1070 struct aoe_atahdr *ahin, *ahout; 1071 struct buf *buf; 1072 struct sk_buff *skb; 1073 struct aoetgt *t; 1074 struct aoeif *ifp; 1075 struct aoedev *d; 1076 long n; 1077 int untainted; 1078 1079 if (f == NULL) 1080 return; 1081 1082 t = f->t; 1083 d = t->d; 1084 skb = f->r_skb; 1085 buf = f->buf; 1086 if (f->flags & FFL_PROBE) 1087 goto out; 1088 if (!skb) /* just fail the buf. */ 1089 goto noskb; 1090 1091 hout = (struct aoe_hdr *) skb_mac_header(f->skb); 1092 ahout = (struct aoe_atahdr *) (hout+1); 1093 1094 hin = (struct aoe_hdr *) skb->data; 1095 skb_pull(skb, sizeof(*hin)); 1096 ahin = (struct aoe_atahdr *) skb->data; 1097 skb_pull(skb, sizeof(*ahin)); 1098 if (ahin->cmdstat & 0xa9) { /* these bits cleared on success */ 1099 pr_err("aoe: ata error cmd=%2.2Xh stat=%2.2Xh from e%ld.%d\n", 1100 ahout->cmdstat, ahin->cmdstat, 1101 d->aoemajor, d->aoeminor); 1102noskb: if (buf) 1103 buf->bio->bi_status = BLK_STS_IOERR; 1104 goto out; 1105 } 1106 1107 n = ahout->scnt << 9; 1108 switch (ahout->cmdstat) { 1109 case ATA_CMD_PIO_READ: 1110 case ATA_CMD_PIO_READ_EXT: 1111 if (skb->len < n) { 1112 pr_err("%s e%ld.%d. skb->len=%d need=%ld\n", 1113 "aoe: runt data size in read from", 1114 (long) d->aoemajor, d->aoeminor, 1115 skb->len, n); 1116 buf->bio->bi_status = BLK_STS_IOERR; 1117 break; 1118 } 1119 if (n > f->iter.bi_size) { 1120 pr_err_ratelimited("%s e%ld.%d. bytes=%ld need=%u\n", 1121 "aoe: too-large data size in read from", 1122 (long) d->aoemajor, d->aoeminor, 1123 n, f->iter.bi_size); 1124 buf->bio->bi_status = BLK_STS_IOERR; 1125 break; 1126 } 1127 bvcpy(skb, f->buf->bio, f->iter, n); 1128 fallthrough; 1129 case ATA_CMD_PIO_WRITE: 1130 case ATA_CMD_PIO_WRITE_EXT: 1131 spin_lock_irq(&d->lock); 1132 ifp = getif(t, skb->dev); 1133 if (ifp) 1134 ifp->lost = 0; 1135 spin_unlock_irq(&d->lock); 1136 break; 1137 case ATA_CMD_ID_ATA: 1138 if (skb->len < 512) { 1139 pr_info("%s e%ld.%d. skb->len=%d need=512\n", 1140 "aoe: runt data size in ataid from", 1141 (long) d->aoemajor, d->aoeminor, 1142 skb->len); 1143 break; 1144 } 1145 if (skb_linearize(skb)) 1146 break; 1147 spin_lock_irq(&d->lock); 1148 ataid_complete(d, t, skb->data); 1149 spin_unlock_irq(&d->lock); 1150 break; 1151 default: 1152 pr_info("aoe: unrecognized ata command %2.2Xh for %d.%d\n", 1153 ahout->cmdstat, 1154 be16_to_cpu(get_unaligned(&hin->major)), 1155 hin->minor); 1156 } 1157out: 1158 spin_lock_irq(&d->lock); 1159 if (t->taint > 0 1160 && --t->taint > 0 1161 && t->nout_probes == 0) { 1162 count_targets(d, &untainted); 1163 if (untainted > 0) { 1164 probe(t); 1165 t->nout_probes++; 1166 } 1167 } 1168 1169 aoe_freetframe(f); 1170 1171 if (buf && --buf->nframesout == 0 && buf->iter.bi_size == 0) 1172 aoe_end_buf(d, buf); 1173 1174 spin_unlock_irq(&d->lock); 1175 aoedev_put(d); 1176 dev_kfree_skb(skb); 1177} 1178 1179/* Enters with iocq.lock held. 1180 * Returns true iff responses needing processing remain. 1181 */ 1182static int 1183ktio(int id) 1184{ 1185 struct frame *f; 1186 struct list_head *pos; 1187 int i; 1188 int actual_id; 1189 1190 for (i = 0; ; ++i) { 1191 if (i == MAXIOC) 1192 return 1; 1193 if (list_empty(&iocq[id].head)) 1194 return 0; 1195 pos = iocq[id].head.next; 1196 list_del(pos); 1197 f = list_entry(pos, struct frame, head); 1198 spin_unlock_irq(&iocq[id].lock); 1199 ktiocomplete(f); 1200 1201 /* Figure out if extra threads are required. */ 1202 actual_id = f->t->d->aoeminor % ncpus; 1203 1204 if (!kts[actual_id].active) { 1205 BUG_ON(id != 0); 1206 mutex_lock(&ktio_spawn_lock); 1207 if (!kts[actual_id].active 1208 && aoe_ktstart(&kts[actual_id]) == 0) 1209 kts[actual_id].active = 1; 1210 mutex_unlock(&ktio_spawn_lock); 1211 } 1212 spin_lock_irq(&iocq[id].lock); 1213 } 1214} 1215 1216static int 1217kthread(void *vp) 1218{ 1219 struct ktstate *k; 1220 DECLARE_WAITQUEUE(wait, current); 1221 int more; 1222 1223 k = vp; 1224 current->flags |= PF_NOFREEZE; 1225 set_user_nice(current, -10); 1226 complete(&k->rendez); /* tell spawner we're running */ 1227 do { 1228 spin_lock_irq(k->lock); 1229 more = k->fn(k->id); 1230 if (!more) { 1231 add_wait_queue(k->waitq, &wait); 1232 __set_current_state(TASK_INTERRUPTIBLE); 1233 } 1234 spin_unlock_irq(k->lock); 1235 if (!more) { 1236 schedule(); 1237 remove_wait_queue(k->waitq, &wait); 1238 } else 1239 cond_resched(); 1240 } while (!kthread_should_stop()); 1241 complete(&k->rendez); /* tell spawner we're stopping */ 1242 return 0; 1243} 1244 1245void 1246aoe_ktstop(struct ktstate *k) 1247{ 1248 kthread_stop(k->task); 1249 wait_for_completion(&k->rendez); 1250} 1251 1252int 1253aoe_ktstart(struct ktstate *k) 1254{ 1255 struct task_struct *task; 1256 1257 init_completion(&k->rendez); 1258 task = kthread_run(kthread, k, "%s", k->name); 1259 if (task == NULL || IS_ERR(task)) 1260 return -ENOMEM; 1261 k->task = task; 1262 wait_for_completion(&k->rendez); /* allow kthread to start */ 1263 init_completion(&k->rendez); /* for waiting for exit later */ 1264 return 0; 1265} 1266 1267/* pass it off to kthreads for processing */ 1268static void 1269ktcomplete(struct frame *f, struct sk_buff *skb) 1270{ 1271 int id; 1272 ulong flags; 1273 1274 f->r_skb = skb; 1275 id = f->t->d->aoeminor % ncpus; 1276 spin_lock_irqsave(&iocq[id].lock, flags); 1277 if (!kts[id].active) { 1278 spin_unlock_irqrestore(&iocq[id].lock, flags); 1279 /* The thread with id has not been spawned yet, 1280 * so delegate the work to the main thread and 1281 * try spawning a new thread. 1282 */ 1283 id = 0; 1284 spin_lock_irqsave(&iocq[id].lock, flags); 1285 } 1286 list_add_tail(&f->head, &iocq[id].head); 1287 spin_unlock_irqrestore(&iocq[id].lock, flags); 1288 wake_up(&ktiowq[id]); 1289} 1290 1291struct sk_buff * 1292aoecmd_ata_rsp(struct sk_buff *skb) 1293{ 1294 struct aoedev *d; 1295 struct aoe_hdr *h; 1296 struct frame *f; 1297 u32 n; 1298 ulong flags; 1299 char ebuf[128]; 1300 u16 aoemajor; 1301 1302 h = (struct aoe_hdr *) skb->data; 1303 aoemajor = be16_to_cpu(get_unaligned(&h->major)); 1304 d = aoedev_by_aoeaddr(aoemajor, h->minor, 0); 1305 if (d == NULL) { 1306 snprintf(ebuf, sizeof ebuf, "aoecmd_ata_rsp: ata response " 1307 "for unknown device %d.%d\n", 1308 aoemajor, h->minor); 1309 aoechr_error(ebuf); 1310 return skb; 1311 } 1312 1313 spin_lock_irqsave(&d->lock, flags); 1314 1315 n = be32_to_cpu(get_unaligned(&h->tag)); 1316 f = getframe(d, n); 1317 if (f) { 1318 calc_rttavg(d, f->t, tsince_hr(f)); 1319 f->t->nout--; 1320 if (f->flags & FFL_PROBE) 1321 f->t->nout_probes--; 1322 } else { 1323 f = getframe_deferred(d, n); 1324 if (f) { 1325 calc_rttavg(d, NULL, tsince_hr(f)); 1326 } else { 1327 calc_rttavg(d, NULL, tsince(n)); 1328 spin_unlock_irqrestore(&d->lock, flags); 1329 aoedev_put(d); 1330 snprintf(ebuf, sizeof(ebuf), 1331 "%15s e%d.%d tag=%08x@%08lx s=%pm d=%pm\n", 1332 "unexpected rsp", 1333 get_unaligned_be16(&h->major), 1334 h->minor, 1335 get_unaligned_be32(&h->tag), 1336 jiffies, 1337 h->src, 1338 h->dst); 1339 aoechr_error(ebuf); 1340 return skb; 1341 } 1342 } 1343 aoecmd_work(d); 1344 1345 spin_unlock_irqrestore(&d->lock, flags); 1346 1347 ktcomplete(f, skb); 1348 1349 /* 1350 * Note here that we do not perform an aoedev_put, as we are 1351 * leaving this reference for the ktio to release. 1352 */ 1353 return NULL; 1354} 1355 1356void 1357aoecmd_cfg(ushort aoemajor, unsigned char aoeminor) 1358{ 1359 struct sk_buff_head queue; 1360 1361 __skb_queue_head_init(&queue); 1362 aoecmd_cfg_pkts(aoemajor, aoeminor, &queue); 1363 aoenet_xmit(&queue); 1364} 1365 1366struct sk_buff * 1367aoecmd_ata_id(struct aoedev *d) 1368{ 1369 struct aoe_hdr *h; 1370 struct aoe_atahdr *ah; 1371 struct frame *f; 1372 struct sk_buff *skb; 1373 struct aoetgt *t; 1374 1375 f = newframe(d); 1376 if (f == NULL) 1377 return NULL; 1378 1379 t = *d->tgt; 1380 1381 /* initialize the headers & frame */ 1382 skb = f->skb; 1383 h = (struct aoe_hdr *) skb_mac_header(skb); 1384 ah = (struct aoe_atahdr *) (h+1); 1385 skb_put(skb, sizeof *h + sizeof *ah); 1386 memset(h, 0, skb->len); 1387 f->tag = aoehdr_atainit(d, t, h); 1388 fhash(f); 1389 t->nout++; 1390 f->waited = 0; 1391 f->waited_total = 0; 1392 1393 /* set up ata header */ 1394 ah->scnt = 1; 1395 ah->cmdstat = ATA_CMD_ID_ATA; 1396 ah->lba3 = 0xa0; 1397 1398 skb->dev = t->ifp->nd; 1399 1400 d->rttavg = RTTAVG_INIT; 1401 d->rttdev = RTTDEV_INIT; 1402 d->timer.function = rexmit_timer; 1403 1404 skb = skb_clone(skb, GFP_ATOMIC); 1405 if (skb) 1406 f->sent = ktime_get(); 1407 1408 return skb; 1409} 1410 1411static struct aoetgt ** 1412grow_targets(struct aoedev *d) 1413{ 1414 ulong oldn, newn; 1415 struct aoetgt **tt; 1416 1417 oldn = d->ntargets; 1418 newn = oldn * 2; 1419 tt = kcalloc(newn, sizeof(*d->targets), GFP_ATOMIC); 1420 if (!tt) 1421 return NULL; 1422 memmove(tt, d->targets, sizeof(*d->targets) * oldn); 1423 d->tgt = tt + (d->tgt - d->targets); 1424 kfree(d->targets); 1425 d->targets = tt; 1426 d->ntargets = newn; 1427 1428 return &d->targets[oldn]; 1429} 1430 1431static struct aoetgt * 1432addtgt(struct aoedev *d, char *addr, ulong nframes) 1433{ 1434 struct aoetgt *t, **tt, **te; 1435 1436 tt = d->targets; 1437 te = tt + d->ntargets; 1438 for (; tt < te && *tt; tt++) 1439 ; 1440 1441 if (tt == te) { 1442 tt = grow_targets(d); 1443 if (!tt) 1444 goto nomem; 1445 } 1446 t = kzalloc(sizeof(*t), GFP_ATOMIC); 1447 if (!t) 1448 goto nomem; 1449 t->nframes = nframes; 1450 t->d = d; 1451 memcpy(t->addr, addr, sizeof t->addr); 1452 t->ifp = t->ifs; 1453 aoecmd_wreset(t); 1454 t->maxout = t->nframes / 2; 1455 INIT_LIST_HEAD(&t->ffree); 1456 return *tt = t; 1457 1458 nomem: 1459 pr_info("aoe: cannot allocate memory to add target\n"); 1460 return NULL; 1461} 1462 1463static void 1464setdbcnt(struct aoedev *d) 1465{ 1466 struct aoetgt **t, **e; 1467 int bcnt = 0; 1468 1469 t = d->targets; 1470 e = t + d->ntargets; 1471 for (; t < e && *t; t++) 1472 if (bcnt == 0 || bcnt > (*t)->minbcnt) 1473 bcnt = (*t)->minbcnt; 1474 if (bcnt != d->maxbcnt) { 1475 d->maxbcnt = bcnt; 1476 pr_info("aoe: e%ld.%d: setting %d byte data frames\n", 1477 d->aoemajor, d->aoeminor, bcnt); 1478 } 1479} 1480 1481static void 1482setifbcnt(struct aoetgt *t, struct net_device *nd, int bcnt) 1483{ 1484 struct aoedev *d; 1485 struct aoeif *p, *e; 1486 int minbcnt; 1487 1488 d = t->d; 1489 minbcnt = bcnt; 1490 p = t->ifs; 1491 e = p + NAOEIFS; 1492 for (; p < e; p++) { 1493 if (p->nd == NULL) 1494 break; /* end of the valid interfaces */ 1495 if (p->nd == nd) { 1496 p->bcnt = bcnt; /* we're updating */ 1497 nd = NULL; 1498 } else if (minbcnt > p->bcnt) 1499 minbcnt = p->bcnt; /* find the min interface */ 1500 } 1501 if (nd) { 1502 if (p == e) { 1503 pr_err("aoe: device setifbcnt failure; too many interfaces.\n"); 1504 return; 1505 } 1506 dev_hold(nd); 1507 p->nd = nd; 1508 p->bcnt = bcnt; 1509 } 1510 t->minbcnt = minbcnt; 1511 setdbcnt(d); 1512} 1513 1514void 1515aoecmd_cfg_rsp(struct sk_buff *skb) 1516{ 1517 struct aoedev *d; 1518 struct aoe_hdr *h; 1519 struct aoe_cfghdr *ch; 1520 struct aoetgt *t; 1521 ulong flags, aoemajor; 1522 struct sk_buff *sl; 1523 struct sk_buff_head queue; 1524 u16 n; 1525 1526 sl = NULL; 1527 h = (struct aoe_hdr *) skb_mac_header(skb); 1528 ch = (struct aoe_cfghdr *) (h+1); 1529 1530 /* 1531 * Enough people have their dip switches set backwards to 1532 * warrant a loud message for this special case. 1533 */ 1534 aoemajor = get_unaligned_be16(&h->major); 1535 if (aoemajor == 0xfff) { 1536 printk(KERN_ERR "aoe: Warning: shelf address is all ones. " 1537 "Check shelf dip switches.\n"); 1538 return; 1539 } 1540 if (aoemajor == 0xffff) { 1541 pr_info("aoe: e%ld.%d: broadcast shelf number invalid\n", 1542 aoemajor, (int) h->minor); 1543 return; 1544 } 1545 if (h->minor == 0xff) { 1546 pr_info("aoe: e%ld.%d: broadcast slot number invalid\n", 1547 aoemajor, (int) h->minor); 1548 return; 1549 } 1550 1551 n = be16_to_cpu(ch->bufcnt); 1552 if (n > aoe_maxout) /* keep it reasonable */ 1553 n = aoe_maxout; 1554 1555 d = aoedev_by_aoeaddr(aoemajor, h->minor, 1); 1556 if (d == NULL) { 1557 pr_info("aoe: device allocation failure\n"); 1558 return; 1559 } 1560 1561 spin_lock_irqsave(&d->lock, flags); 1562 1563 t = gettgt(d, h->src); 1564 if (t) { 1565 t->nframes = n; 1566 if (n < t->maxout) 1567 aoecmd_wreset(t); 1568 } else { 1569 t = addtgt(d, h->src, n); 1570 if (!t) 1571 goto bail; 1572 } 1573 n = skb->dev->mtu; 1574 n -= sizeof(struct aoe_hdr) + sizeof(struct aoe_atahdr); 1575 n /= 512; 1576 if (n > ch->scnt) 1577 n = ch->scnt; 1578 n = n ? n * 512 : DEFAULTBCNT; 1579 setifbcnt(t, skb->dev, n); 1580 1581 /* don't change users' perspective */ 1582 if (d->nopen == 0) { 1583 d->fw_ver = be16_to_cpu(ch->fwver); 1584 sl = aoecmd_ata_id(d); 1585 } 1586bail: 1587 spin_unlock_irqrestore(&d->lock, flags); 1588 aoedev_put(d); 1589 if (sl) { 1590 __skb_queue_head_init(&queue); 1591 __skb_queue_tail(&queue, sl); 1592 aoenet_xmit(&queue); 1593 } 1594} 1595 1596void 1597aoecmd_wreset(struct aoetgt *t) 1598{ 1599 t->maxout = 1; 1600 t->ssthresh = t->nframes / 2; 1601 t->next_cwnd = t->nframes; 1602} 1603 1604void 1605aoecmd_cleanslate(struct aoedev *d) 1606{ 1607 struct aoetgt **t, **te; 1608 1609 d->rttavg = RTTAVG_INIT; 1610 d->rttdev = RTTDEV_INIT; 1611 d->maxbcnt = 0; 1612 1613 t = d->targets; 1614 te = t + d->ntargets; 1615 for (; t < te && *t; t++) 1616 aoecmd_wreset(*t); 1617} 1618 1619void 1620aoe_failbuf(struct aoedev *d, struct buf *buf) 1621{ 1622 if (buf == NULL) 1623 return; 1624 buf->iter.bi_size = 0; 1625 buf->bio->bi_status = BLK_STS_IOERR; 1626 if (buf->nframesout == 0) 1627 aoe_end_buf(d, buf); 1628} 1629 1630void 1631aoe_flush_iocq(void) 1632{ 1633 int i; 1634 1635 for (i = 0; i < ncpus; i++) { 1636 if (kts[i].active) 1637 aoe_flush_iocq_by_index(i); 1638 } 1639} 1640 1641void 1642aoe_flush_iocq_by_index(int id) 1643{ 1644 struct frame *f; 1645 struct aoedev *d; 1646 LIST_HEAD(flist); 1647 struct list_head *pos; 1648 struct sk_buff *skb; 1649 ulong flags; 1650 1651 spin_lock_irqsave(&iocq[id].lock, flags); 1652 list_splice_init(&iocq[id].head, &flist); 1653 spin_unlock_irqrestore(&iocq[id].lock, flags); 1654 while (!list_empty(&flist)) { 1655 pos = flist.next; 1656 list_del(pos); 1657 f = list_entry(pos, struct frame, head); 1658 d = f->t->d; 1659 skb = f->r_skb; 1660 spin_lock_irqsave(&d->lock, flags); 1661 if (f->buf) { 1662 f->buf->nframesout--; 1663 aoe_failbuf(d, f->buf); 1664 } 1665 aoe_freetframe(f); 1666 spin_unlock_irqrestore(&d->lock, flags); 1667 dev_kfree_skb(skb); 1668 aoedev_put(d); 1669 } 1670} 1671 1672int __init 1673aoecmd_init(void) 1674{ 1675 void *p; 1676 int i; 1677 int ret; 1678 1679 /* get_zeroed_page returns page with ref count 1 */ 1680 p = (void *) get_zeroed_page(GFP_KERNEL); 1681 if (!p) 1682 return -ENOMEM; 1683 empty_page = virt_to_page(p); 1684 1685 ncpus = num_online_cpus(); 1686 1687 iocq = kcalloc(ncpus, sizeof(struct iocq_ktio), GFP_KERNEL); 1688 if (!iocq) 1689 return -ENOMEM; 1690 1691 kts = kcalloc(ncpus, sizeof(struct ktstate), GFP_KERNEL); 1692 if (!kts) { 1693 ret = -ENOMEM; 1694 goto kts_fail; 1695 } 1696 1697 ktiowq = kcalloc(ncpus, sizeof(wait_queue_head_t), GFP_KERNEL); 1698 if (!ktiowq) { 1699 ret = -ENOMEM; 1700 goto ktiowq_fail; 1701 } 1702 1703 for (i = 0; i < ncpus; i++) { 1704 INIT_LIST_HEAD(&iocq[i].head); 1705 spin_lock_init(&iocq[i].lock); 1706 init_waitqueue_head(&ktiowq[i]); 1707 snprintf(kts[i].name, sizeof(kts[i].name), "aoe_ktio%d", i); 1708 kts[i].fn = ktio; 1709 kts[i].waitq = &ktiowq[i]; 1710 kts[i].lock = &iocq[i].lock; 1711 kts[i].id = i; 1712 kts[i].active = 0; 1713 } 1714 kts[0].active = 1; 1715 if (aoe_ktstart(&kts[0])) { 1716 ret = -ENOMEM; 1717 goto ktstart_fail; 1718 } 1719 return 0; 1720 1721ktstart_fail: 1722 kfree(ktiowq); 1723ktiowq_fail: 1724 kfree(kts); 1725kts_fail: 1726 kfree(iocq); 1727 1728 return ret; 1729} 1730 1731void 1732aoecmd_exit(void) 1733{ 1734 int i; 1735 1736 for (i = 0; i < ncpus; i++) 1737 if (kts[i].active) 1738 aoe_ktstop(&kts[i]); 1739 1740 aoe_flush_iocq(); 1741 1742 /* Free up the iocq and thread speicific configuration 1743 * allocated during startup. 1744 */ 1745 kfree(iocq); 1746 kfree(kts); 1747 kfree(ktiowq); 1748 1749 free_page((unsigned long) page_address(empty_page)); 1750 empty_page = NULL; 1751} 1752