1 /* 2 * Xen SCSI frontend driver 3 * 4 * Copyright (c) 2008, FUJITSU Limited 5 * 6 * This program is free software; you can redistribute it and/or 7 * modify it under the terms of the GNU General Public License version 2 8 * as published by the Free Software Foundation; or, when distributed 9 * separately from the Linux kernel or incorporated into other 10 * software packages, subject to the following license: 11 * 12 * Permission is hereby granted, free of charge, to any person obtaining a copy 13 * of this source file (the "Software"), to deal in the Software without 14 * restriction, including without limitation the rights to use, copy, modify, 15 * merge, publish, distribute, sublicense, and/or sell copies of the Software, 16 * and to permit persons to whom the Software is furnished to do so, subject to 17 * the following conditions: 18 * 19 * The above copyright notice and this permission notice shall be included in 20 * all copies or substantial portions of the Software. 21 * 22 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 23 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 24 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 25 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 26 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING 27 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS 28 * IN THE SOFTWARE. 29 */ 30 31 #include <linux/module.h> 32 #include <linux/kernel.h> 33 #include <linux/device.h> 34 #include <linux/wait.h> 35 #include <linux/interrupt.h> 36 #include <linux/mutex.h> 37 #include <linux/spinlock.h> 38 #include <linux/sched.h> 39 #include <linux/blkdev.h> 40 #include <linux/pfn.h> 41 #include <linux/slab.h> 42 #include <linux/bitops.h> 43 44 #include <scsi/scsi_cmnd.h> 45 #include <scsi/scsi_device.h> 46 #include <scsi/scsi.h> 47 #include <scsi/scsi_host.h> 48 49 #include <xen/xen.h> 50 #include <xen/xenbus.h> 51 #include <xen/grant_table.h> 52 #include <xen/events.h> 53 #include <xen/page.h> 54 55 #include <xen/interface/grant_table.h> 56 #include <xen/interface/io/vscsiif.h> 57 #include <xen/interface/io/protocols.h> 58 59 #include <asm/xen/hypervisor.h> 60 61 #define VSCSIFRONT_OP_ADD_LUN 1 62 #define VSCSIFRONT_OP_DEL_LUN 2 63 #define VSCSIFRONT_OP_READD_LUN 3 64 65 /* Tuning point. */ 66 #define VSCSIIF_DEFAULT_CMD_PER_LUN 10 67 #define VSCSIIF_MAX_TARGET 64 68 #define VSCSIIF_MAX_LUN 255 69 70 #define VSCSIIF_RING_SIZE __CONST_RING_SIZE(vscsiif, PAGE_SIZE) 71 #define VSCSIIF_MAX_REQS VSCSIIF_RING_SIZE 72 73 #define vscsiif_grants_sg(_sg) (PFN_UP((_sg) * \ 74 sizeof(struct scsiif_request_segment))) 75 76 struct vscsifrnt_shadow { 77 /* command between backend and frontend */ 78 unsigned char act; 79 uint8_t nr_segments; 80 uint16_t rqid; 81 uint16_t ref_rqid; 82 83 bool inflight; 84 85 unsigned int nr_grants; /* number of grants in gref[] */ 86 struct scsiif_request_segment *sg; /* scatter/gather elements */ 87 struct scsiif_request_segment seg[VSCSIIF_SG_TABLESIZE]; 88 89 /* Do reset or abort function. */ 90 wait_queue_head_t wq_reset; /* reset work queue */ 91 int wait_reset; /* reset work queue condition */ 92 int32_t rslt_reset; /* reset response status: */ 93 /* SUCCESS or FAILED or: */ 94 #define RSLT_RESET_WAITING 0 95 #define RSLT_RESET_ERR -1 96 97 /* Requested struct scsi_cmnd is stored from kernel. */ 98 struct scsi_cmnd *sc; 99 int gref[vscsiif_grants_sg(SG_ALL) + SG_ALL]; 100 }; 101 102 struct vscsifrnt_info { 103 struct xenbus_device *dev; 104 105 struct Scsi_Host *host; 106 enum { 107 STATE_INACTIVE, 108 STATE_ACTIVE, 109 STATE_ERROR 110 } host_active; 111 112 unsigned int evtchn; 113 unsigned int irq; 114 115 grant_ref_t ring_ref; 116 struct vscsiif_front_ring ring; 117 struct vscsiif_response ring_rsp; 118 119 spinlock_t shadow_lock; 120 DECLARE_BITMAP(shadow_free_bitmap, VSCSIIF_MAX_REQS); 121 struct vscsifrnt_shadow *shadow[VSCSIIF_MAX_REQS]; 122 123 /* Following items are protected by the host lock. */ 124 wait_queue_head_t wq_sync; 125 wait_queue_head_t wq_pause; 126 unsigned int wait_ring_available:1; 127 unsigned int waiting_pause:1; 128 unsigned int pause:1; 129 unsigned callers; 130 131 char dev_state_path[64]; 132 struct task_struct *curr; 133 }; 134 135 static DEFINE_MUTEX(scsifront_mutex); 136 137 static void scsifront_wake_up(struct vscsifrnt_info *info) 138 { 139 info->wait_ring_available = 0; 140 wake_up(&info->wq_sync); 141 } 142 143 static int scsifront_get_rqid(struct vscsifrnt_info *info) 144 { 145 unsigned long flags; 146 int free; 147 148 spin_lock_irqsave(&info->shadow_lock, flags); 149 150 free = find_first_bit(info->shadow_free_bitmap, VSCSIIF_MAX_REQS); 151 __clear_bit(free, info->shadow_free_bitmap); 152 153 spin_unlock_irqrestore(&info->shadow_lock, flags); 154 155 return free; 156 } 157 158 static int _scsifront_put_rqid(struct vscsifrnt_info *info, uint32_t id) 159 { 160 int empty = bitmap_empty(info->shadow_free_bitmap, VSCSIIF_MAX_REQS); 161 162 __set_bit(id, info->shadow_free_bitmap); 163 info->shadow[id] = NULL; 164 165 return empty || info->wait_ring_available; 166 } 167 168 static void scsifront_put_rqid(struct vscsifrnt_info *info, uint32_t id) 169 { 170 unsigned long flags; 171 int kick; 172 173 spin_lock_irqsave(&info->shadow_lock, flags); 174 kick = _scsifront_put_rqid(info, id); 175 spin_unlock_irqrestore(&info->shadow_lock, flags); 176 177 if (kick) 178 scsifront_wake_up(info); 179 } 180 181 static int scsifront_do_request(struct vscsifrnt_info *info, 182 struct vscsifrnt_shadow *shadow) 183 { 184 struct vscsiif_front_ring *ring = &(info->ring); 185 struct vscsiif_request *ring_req; 186 struct scsi_cmnd *sc = shadow->sc; 187 uint32_t id; 188 int i, notify; 189 190 if (RING_FULL(&info->ring)) 191 return -EBUSY; 192 193 id = scsifront_get_rqid(info); /* use id in response */ 194 if (id >= VSCSIIF_MAX_REQS) 195 return -EBUSY; 196 197 info->shadow[id] = shadow; 198 shadow->rqid = id; 199 200 ring_req = RING_GET_REQUEST(&(info->ring), ring->req_prod_pvt); 201 ring->req_prod_pvt++; 202 203 ring_req->rqid = id; 204 ring_req->act = shadow->act; 205 ring_req->ref_rqid = shadow->ref_rqid; 206 ring_req->nr_segments = shadow->nr_segments; 207 208 ring_req->id = sc->device->id; 209 ring_req->lun = sc->device->lun; 210 ring_req->channel = sc->device->channel; 211 ring_req->cmd_len = sc->cmd_len; 212 213 BUG_ON(sc->cmd_len > VSCSIIF_MAX_COMMAND_SIZE); 214 215 memcpy(ring_req->cmnd, sc->cmnd, sc->cmd_len); 216 217 ring_req->sc_data_direction = (uint8_t)sc->sc_data_direction; 218 ring_req->timeout_per_command = scsi_cmd_to_rq(sc)->timeout / HZ; 219 220 for (i = 0; i < (shadow->nr_segments & ~VSCSIIF_SG_GRANT); i++) 221 ring_req->seg[i] = shadow->seg[i]; 222 223 shadow->inflight = true; 224 225 RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(ring, notify); 226 if (notify) 227 notify_remote_via_irq(info->irq); 228 229 return 0; 230 } 231 232 static void scsifront_set_error(struct vscsifrnt_info *info, const char *msg) 233 { 234 shost_printk(KERN_ERR, info->host, KBUILD_MODNAME "%s\n" 235 "Disabling device for further use\n", msg); 236 info->host_active = STATE_ERROR; 237 } 238 239 static void scsifront_gnttab_done(struct vscsifrnt_info *info, 240 struct vscsifrnt_shadow *shadow) 241 { 242 int i; 243 244 if (shadow->sc->sc_data_direction == DMA_NONE) 245 return; 246 247 for (i = 0; i < shadow->nr_grants; i++) { 248 if (unlikely(!gnttab_try_end_foreign_access(shadow->gref[i]))) { 249 scsifront_set_error(info, "grant still in use by backend"); 250 return; 251 } 252 } 253 254 kfree(shadow->sg); 255 } 256 257 static unsigned int scsifront_host_byte(int32_t rslt) 258 { 259 switch (XEN_VSCSIIF_RSLT_HOST(rslt)) { 260 case XEN_VSCSIIF_RSLT_HOST_OK: 261 return DID_OK; 262 case XEN_VSCSIIF_RSLT_HOST_NO_CONNECT: 263 return DID_NO_CONNECT; 264 case XEN_VSCSIIF_RSLT_HOST_BUS_BUSY: 265 return DID_BUS_BUSY; 266 case XEN_VSCSIIF_RSLT_HOST_TIME_OUT: 267 return DID_TIME_OUT; 268 case XEN_VSCSIIF_RSLT_HOST_BAD_TARGET: 269 return DID_BAD_TARGET; 270 case XEN_VSCSIIF_RSLT_HOST_ABORT: 271 return DID_ABORT; 272 case XEN_VSCSIIF_RSLT_HOST_PARITY: 273 return DID_PARITY; 274 case XEN_VSCSIIF_RSLT_HOST_ERROR: 275 return DID_ERROR; 276 case XEN_VSCSIIF_RSLT_HOST_RESET: 277 return DID_RESET; 278 case XEN_VSCSIIF_RSLT_HOST_BAD_INTR: 279 return DID_BAD_INTR; 280 case XEN_VSCSIIF_RSLT_HOST_PASSTHROUGH: 281 return DID_PASSTHROUGH; 282 case XEN_VSCSIIF_RSLT_HOST_SOFT_ERROR: 283 return DID_SOFT_ERROR; 284 case XEN_VSCSIIF_RSLT_HOST_IMM_RETRY: 285 return DID_IMM_RETRY; 286 case XEN_VSCSIIF_RSLT_HOST_REQUEUE: 287 return DID_REQUEUE; 288 case XEN_VSCSIIF_RSLT_HOST_TRANSPORT_DISRUPTED: 289 return DID_TRANSPORT_DISRUPTED; 290 case XEN_VSCSIIF_RSLT_HOST_TRANSPORT_FAILFAST: 291 return DID_TRANSPORT_FAILFAST; 292 case XEN_VSCSIIF_RSLT_HOST_TARGET_FAILURE: 293 return DID_TARGET_FAILURE; 294 case XEN_VSCSIIF_RSLT_HOST_NEXUS_FAILURE: 295 return DID_NEXUS_FAILURE; 296 case XEN_VSCSIIF_RSLT_HOST_ALLOC_FAILURE: 297 return DID_ALLOC_FAILURE; 298 case XEN_VSCSIIF_RSLT_HOST_MEDIUM_ERROR: 299 return DID_MEDIUM_ERROR; 300 case XEN_VSCSIIF_RSLT_HOST_TRANSPORT_MARGINAL: 301 return DID_TRANSPORT_MARGINAL; 302 default: 303 return DID_ERROR; 304 } 305 } 306 307 static void scsifront_cdb_cmd_done(struct vscsifrnt_info *info, 308 struct vscsiif_response *ring_rsp) 309 { 310 struct vscsifrnt_shadow *shadow; 311 struct scsi_cmnd *sc; 312 uint32_t id; 313 uint8_t sense_len; 314 315 id = ring_rsp->rqid; 316 shadow = info->shadow[id]; 317 sc = shadow->sc; 318 319 BUG_ON(sc == NULL); 320 321 scsifront_gnttab_done(info, shadow); 322 if (info->host_active == STATE_ERROR) 323 return; 324 scsifront_put_rqid(info, id); 325 326 set_host_byte(sc, scsifront_host_byte(ring_rsp->rslt)); 327 set_status_byte(sc, XEN_VSCSIIF_RSLT_STATUS(ring_rsp->rslt)); 328 scsi_set_resid(sc, ring_rsp->residual_len); 329 330 sense_len = min_t(uint8_t, VSCSIIF_SENSE_BUFFERSIZE, 331 ring_rsp->sense_len); 332 333 if (sense_len) 334 memcpy(sc->sense_buffer, ring_rsp->sense_buffer, sense_len); 335 336 scsi_done(sc); 337 } 338 339 static void scsifront_sync_cmd_done(struct vscsifrnt_info *info, 340 struct vscsiif_response *ring_rsp) 341 { 342 uint16_t id = ring_rsp->rqid; 343 unsigned long flags; 344 struct vscsifrnt_shadow *shadow = info->shadow[id]; 345 int kick; 346 347 spin_lock_irqsave(&info->shadow_lock, flags); 348 shadow->wait_reset = 1; 349 switch (shadow->rslt_reset) { 350 case RSLT_RESET_WAITING: 351 if (ring_rsp->rslt == XEN_VSCSIIF_RSLT_RESET_SUCCESS) 352 shadow->rslt_reset = SUCCESS; 353 else 354 shadow->rslt_reset = FAILED; 355 break; 356 case RSLT_RESET_ERR: 357 kick = _scsifront_put_rqid(info, id); 358 spin_unlock_irqrestore(&info->shadow_lock, flags); 359 kfree(shadow); 360 if (kick) 361 scsifront_wake_up(info); 362 return; 363 default: 364 scsifront_set_error(info, "bad reset state"); 365 break; 366 } 367 spin_unlock_irqrestore(&info->shadow_lock, flags); 368 369 wake_up(&shadow->wq_reset); 370 } 371 372 static void scsifront_do_response(struct vscsifrnt_info *info, 373 struct vscsiif_response *ring_rsp) 374 { 375 struct vscsifrnt_shadow *shadow; 376 377 if (ring_rsp->rqid >= VSCSIIF_MAX_REQS || 378 !info->shadow[ring_rsp->rqid]->inflight) { 379 scsifront_set_error(info, "illegal rqid returned by backend!"); 380 return; 381 } 382 shadow = info->shadow[ring_rsp->rqid]; 383 shadow->inflight = false; 384 385 if (shadow->act == VSCSIIF_ACT_SCSI_CDB) 386 scsifront_cdb_cmd_done(info, ring_rsp); 387 else 388 scsifront_sync_cmd_done(info, ring_rsp); 389 } 390 391 static int scsifront_ring_drain(struct vscsifrnt_info *info, 392 unsigned int *eoiflag) 393 { 394 struct vscsiif_response ring_rsp; 395 RING_IDX i, rp; 396 int more_to_do = 0; 397 398 rp = READ_ONCE(info->ring.sring->rsp_prod); 399 virt_rmb(); /* ordering required respective to backend */ 400 if (RING_RESPONSE_PROD_OVERFLOW(&info->ring, rp)) { 401 scsifront_set_error(info, "illegal number of responses"); 402 return 0; 403 } 404 for (i = info->ring.rsp_cons; i != rp; i++) { 405 RING_COPY_RESPONSE(&info->ring, i, &ring_rsp); 406 scsifront_do_response(info, &ring_rsp); 407 if (info->host_active == STATE_ERROR) 408 return 0; 409 *eoiflag &= ~XEN_EOI_FLAG_SPURIOUS; 410 } 411 412 info->ring.rsp_cons = i; 413 414 if (i != info->ring.req_prod_pvt) 415 RING_FINAL_CHECK_FOR_RESPONSES(&info->ring, more_to_do); 416 else 417 info->ring.sring->rsp_event = i + 1; 418 419 return more_to_do; 420 } 421 422 static int scsifront_cmd_done(struct vscsifrnt_info *info, 423 unsigned int *eoiflag) 424 { 425 int more_to_do; 426 unsigned long flags; 427 428 spin_lock_irqsave(info->host->host_lock, flags); 429 430 more_to_do = scsifront_ring_drain(info, eoiflag); 431 432 info->wait_ring_available = 0; 433 434 spin_unlock_irqrestore(info->host->host_lock, flags); 435 436 wake_up(&info->wq_sync); 437 438 return more_to_do; 439 } 440 441 static irqreturn_t scsifront_irq_fn(int irq, void *dev_id) 442 { 443 struct vscsifrnt_info *info = dev_id; 444 unsigned int eoiflag = XEN_EOI_FLAG_SPURIOUS; 445 446 if (info->host_active == STATE_ERROR) { 447 xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS); 448 return IRQ_HANDLED; 449 } 450 451 while (scsifront_cmd_done(info, &eoiflag)) 452 /* Yield point for this unbounded loop. */ 453 cond_resched(); 454 455 xen_irq_lateeoi(irq, eoiflag); 456 457 return IRQ_HANDLED; 458 } 459 460 static void scsifront_finish_all(struct vscsifrnt_info *info) 461 { 462 unsigned int i, dummy; 463 struct vscsiif_response resp; 464 465 scsifront_ring_drain(info, &dummy); 466 467 for (i = 0; i < VSCSIIF_MAX_REQS; i++) { 468 if (test_bit(i, info->shadow_free_bitmap)) 469 continue; 470 resp.rqid = i; 471 resp.sense_len = 0; 472 resp.rslt = DID_RESET << 16; 473 resp.residual_len = 0; 474 scsifront_do_response(info, &resp); 475 } 476 } 477 478 static int map_data_for_request(struct vscsifrnt_info *info, 479 struct scsi_cmnd *sc, 480 struct vscsifrnt_shadow *shadow) 481 { 482 grant_ref_t gref_head; 483 struct page *page; 484 int err, ref, ref_cnt = 0; 485 int grant_ro = (sc->sc_data_direction == DMA_TO_DEVICE); 486 unsigned int i, off, len, bytes; 487 unsigned int data_len = scsi_bufflen(sc); 488 unsigned int data_grants = 0, seg_grants = 0; 489 struct scatterlist *sg; 490 struct scsiif_request_segment *seg; 491 492 if (sc->sc_data_direction == DMA_NONE || !data_len) 493 return 0; 494 495 scsi_for_each_sg(sc, sg, scsi_sg_count(sc), i) 496 data_grants += PFN_UP(sg->offset + sg->length); 497 498 if (data_grants > VSCSIIF_SG_TABLESIZE) { 499 if (data_grants > info->host->sg_tablesize) { 500 shost_printk(KERN_ERR, info->host, KBUILD_MODNAME 501 "Unable to map request_buffer for command!\n"); 502 return -E2BIG; 503 } 504 seg_grants = vscsiif_grants_sg(data_grants); 505 shadow->sg = kcalloc(data_grants, 506 sizeof(struct scsiif_request_segment), GFP_ATOMIC); 507 if (!shadow->sg) 508 return -ENOMEM; 509 } 510 seg = shadow->sg ? : shadow->seg; 511 512 err = gnttab_alloc_grant_references(seg_grants + data_grants, 513 &gref_head); 514 if (err) { 515 kfree(shadow->sg); 516 shost_printk(KERN_ERR, info->host, KBUILD_MODNAME 517 "gnttab_alloc_grant_references() error\n"); 518 return -ENOMEM; 519 } 520 521 if (seg_grants) { 522 page = virt_to_page(seg); 523 off = offset_in_page(seg); 524 len = sizeof(struct scsiif_request_segment) * data_grants; 525 while (len > 0) { 526 bytes = min_t(unsigned int, len, PAGE_SIZE - off); 527 528 ref = gnttab_claim_grant_reference(&gref_head); 529 BUG_ON(ref == -ENOSPC); 530 531 gnttab_grant_foreign_access_ref(ref, 532 info->dev->otherend_id, 533 xen_page_to_gfn(page), 1); 534 shadow->gref[ref_cnt] = ref; 535 shadow->seg[ref_cnt].gref = ref; 536 shadow->seg[ref_cnt].offset = (uint16_t)off; 537 shadow->seg[ref_cnt].length = (uint16_t)bytes; 538 539 page++; 540 len -= bytes; 541 off = 0; 542 ref_cnt++; 543 } 544 BUG_ON(seg_grants < ref_cnt); 545 seg_grants = ref_cnt; 546 } 547 548 scsi_for_each_sg(sc, sg, scsi_sg_count(sc), i) { 549 page = sg_page(sg); 550 off = sg->offset; 551 len = sg->length; 552 553 while (len > 0 && data_len > 0) { 554 /* 555 * sg sends a scatterlist that is larger than 556 * the data_len it wants transferred for certain 557 * IO sizes. 558 */ 559 bytes = min_t(unsigned int, len, PAGE_SIZE - off); 560 bytes = min(bytes, data_len); 561 562 ref = gnttab_claim_grant_reference(&gref_head); 563 BUG_ON(ref == -ENOSPC); 564 565 gnttab_grant_foreign_access_ref(ref, 566 info->dev->otherend_id, 567 xen_page_to_gfn(page), 568 grant_ro); 569 570 shadow->gref[ref_cnt] = ref; 571 seg->gref = ref; 572 seg->offset = (uint16_t)off; 573 seg->length = (uint16_t)bytes; 574 575 page++; 576 seg++; 577 len -= bytes; 578 data_len -= bytes; 579 off = 0; 580 ref_cnt++; 581 } 582 } 583 584 if (seg_grants) 585 shadow->nr_segments = VSCSIIF_SG_GRANT | seg_grants; 586 else 587 shadow->nr_segments = (uint8_t)ref_cnt; 588 shadow->nr_grants = ref_cnt; 589 590 return 0; 591 } 592 593 static int scsifront_enter(struct vscsifrnt_info *info) 594 { 595 if (info->pause) 596 return 1; 597 info->callers++; 598 return 0; 599 } 600 601 static void scsifront_return(struct vscsifrnt_info *info) 602 { 603 info->callers--; 604 if (info->callers) 605 return; 606 607 if (!info->waiting_pause) 608 return; 609 610 info->waiting_pause = 0; 611 wake_up(&info->wq_pause); 612 } 613 614 static int scsifront_queuecommand(struct Scsi_Host *shost, 615 struct scsi_cmnd *sc) 616 { 617 struct vscsifrnt_info *info = shost_priv(shost); 618 struct vscsifrnt_shadow *shadow = scsi_cmd_priv(sc); 619 unsigned long flags; 620 int err; 621 622 if (info->host_active == STATE_ERROR) 623 return SCSI_MLQUEUE_HOST_BUSY; 624 625 sc->result = 0; 626 627 shadow->sc = sc; 628 shadow->act = VSCSIIF_ACT_SCSI_CDB; 629 630 spin_lock_irqsave(shost->host_lock, flags); 631 if (scsifront_enter(info)) { 632 spin_unlock_irqrestore(shost->host_lock, flags); 633 return SCSI_MLQUEUE_HOST_BUSY; 634 } 635 636 err = map_data_for_request(info, sc, shadow); 637 if (err < 0) { 638 pr_debug("%s: err %d\n", __func__, err); 639 scsifront_return(info); 640 spin_unlock_irqrestore(shost->host_lock, flags); 641 if (err == -ENOMEM) 642 return SCSI_MLQUEUE_HOST_BUSY; 643 sc->result = DID_ERROR << 16; 644 scsi_done(sc); 645 return 0; 646 } 647 648 if (scsifront_do_request(info, shadow)) { 649 scsifront_gnttab_done(info, shadow); 650 goto busy; 651 } 652 653 scsifront_return(info); 654 spin_unlock_irqrestore(shost->host_lock, flags); 655 656 return 0; 657 658 busy: 659 scsifront_return(info); 660 spin_unlock_irqrestore(shost->host_lock, flags); 661 pr_debug("%s: busy\n", __func__); 662 return SCSI_MLQUEUE_HOST_BUSY; 663 } 664 665 /* 666 * Any exception handling (reset or abort) must be forwarded to the backend. 667 * We have to wait until an answer is returned. This answer contains the 668 * result to be returned to the requestor. 669 */ 670 static int scsifront_action_handler(struct scsi_cmnd *sc, uint8_t act) 671 { 672 struct Scsi_Host *host = sc->device->host; 673 struct vscsifrnt_info *info = shost_priv(host); 674 struct vscsifrnt_shadow *shadow, *s = scsi_cmd_priv(sc); 675 int err = 0; 676 677 if (info->host_active == STATE_ERROR) 678 return FAILED; 679 680 shadow = kzalloc(sizeof(*shadow), GFP_NOIO); 681 if (!shadow) 682 return FAILED; 683 684 shadow->act = act; 685 shadow->rslt_reset = RSLT_RESET_WAITING; 686 shadow->sc = sc; 687 shadow->ref_rqid = s->rqid; 688 init_waitqueue_head(&shadow->wq_reset); 689 690 spin_lock_irq(host->host_lock); 691 692 for (;;) { 693 if (scsifront_enter(info)) 694 goto fail; 695 696 if (!scsifront_do_request(info, shadow)) 697 break; 698 699 scsifront_return(info); 700 if (err) 701 goto fail; 702 info->wait_ring_available = 1; 703 spin_unlock_irq(host->host_lock); 704 err = wait_event_interruptible(info->wq_sync, 705 !info->wait_ring_available); 706 spin_lock_irq(host->host_lock); 707 } 708 709 spin_unlock_irq(host->host_lock); 710 err = wait_event_interruptible(shadow->wq_reset, shadow->wait_reset); 711 spin_lock_irq(host->host_lock); 712 713 if (!err) { 714 err = shadow->rslt_reset; 715 scsifront_put_rqid(info, shadow->rqid); 716 kfree(shadow); 717 } else { 718 spin_lock(&info->shadow_lock); 719 shadow->rslt_reset = RSLT_RESET_ERR; 720 spin_unlock(&info->shadow_lock); 721 err = FAILED; 722 } 723 724 scsifront_return(info); 725 spin_unlock_irq(host->host_lock); 726 return err; 727 728 fail: 729 spin_unlock_irq(host->host_lock); 730 kfree(shadow); 731 return FAILED; 732 } 733 734 static int scsifront_eh_abort_handler(struct scsi_cmnd *sc) 735 { 736 pr_debug("%s\n", __func__); 737 return scsifront_action_handler(sc, VSCSIIF_ACT_SCSI_ABORT); 738 } 739 740 static int scsifront_dev_reset_handler(struct scsi_cmnd *sc) 741 { 742 pr_debug("%s\n", __func__); 743 return scsifront_action_handler(sc, VSCSIIF_ACT_SCSI_RESET); 744 } 745 746 static int scsifront_sdev_configure(struct scsi_device *sdev) 747 { 748 struct vscsifrnt_info *info = shost_priv(sdev->host); 749 int err; 750 751 if (info->host_active == STATE_ERROR) 752 return -EIO; 753 754 if (info && current == info->curr) { 755 err = xenbus_printf(XBT_NIL, info->dev->nodename, 756 info->dev_state_path, "%d", XenbusStateConnected); 757 if (err) { 758 xenbus_dev_error(info->dev, err, 759 "%s: writing dev_state_path", __func__); 760 return err; 761 } 762 } 763 764 return 0; 765 } 766 767 static void scsifront_sdev_destroy(struct scsi_device *sdev) 768 { 769 struct vscsifrnt_info *info = shost_priv(sdev->host); 770 int err; 771 772 if (info && current == info->curr) { 773 err = xenbus_printf(XBT_NIL, info->dev->nodename, 774 info->dev_state_path, "%d", XenbusStateClosed); 775 if (err) 776 xenbus_dev_error(info->dev, err, 777 "%s: writing dev_state_path", __func__); 778 } 779 } 780 781 static struct scsi_host_template scsifront_sht = { 782 .module = THIS_MODULE, 783 .name = "Xen SCSI frontend driver", 784 .queuecommand = scsifront_queuecommand, 785 .eh_abort_handler = scsifront_eh_abort_handler, 786 .eh_device_reset_handler = scsifront_dev_reset_handler, 787 .slave_configure = scsifront_sdev_configure, 788 .slave_destroy = scsifront_sdev_destroy, 789 .cmd_per_lun = VSCSIIF_DEFAULT_CMD_PER_LUN, 790 .can_queue = VSCSIIF_MAX_REQS, 791 .this_id = -1, 792 .cmd_size = sizeof(struct vscsifrnt_shadow), 793 .sg_tablesize = VSCSIIF_SG_TABLESIZE, 794 .proc_name = "scsifront", 795 }; 796 797 static int scsifront_alloc_ring(struct vscsifrnt_info *info) 798 { 799 struct xenbus_device *dev = info->dev; 800 struct vscsiif_sring *sring; 801 int err; 802 803 /***** Frontend to Backend ring start *****/ 804 err = xenbus_setup_ring(dev, GFP_KERNEL, (void **)&sring, 1, 805 &info->ring_ref); 806 if (err) 807 return err; 808 809 XEN_FRONT_RING_INIT(&info->ring, sring, PAGE_SIZE); 810 811 err = xenbus_alloc_evtchn(dev, &info->evtchn); 812 if (err) { 813 xenbus_dev_fatal(dev, err, "xenbus_alloc_evtchn"); 814 goto free_gnttab; 815 } 816 817 err = bind_evtchn_to_irq_lateeoi(info->evtchn); 818 if (err <= 0) { 819 xenbus_dev_fatal(dev, err, "bind_evtchn_to_irq"); 820 goto free_gnttab; 821 } 822 823 info->irq = err; 824 825 err = request_threaded_irq(info->irq, NULL, scsifront_irq_fn, 826 IRQF_ONESHOT, "scsifront", info); 827 if (err) { 828 xenbus_dev_fatal(dev, err, "request_threaded_irq"); 829 goto free_irq; 830 } 831 832 return 0; 833 834 /* free resource */ 835 free_irq: 836 unbind_from_irqhandler(info->irq, info); 837 free_gnttab: 838 xenbus_teardown_ring((void **)&sring, 1, &info->ring_ref); 839 840 return err; 841 } 842 843 static void scsifront_free_ring(struct vscsifrnt_info *info) 844 { 845 unbind_from_irqhandler(info->irq, info); 846 xenbus_teardown_ring((void **)&info->ring.sring, 1, &info->ring_ref); 847 } 848 849 static int scsifront_init_ring(struct vscsifrnt_info *info) 850 { 851 struct xenbus_device *dev = info->dev; 852 struct xenbus_transaction xbt; 853 int err; 854 855 pr_debug("%s\n", __func__); 856 857 err = scsifront_alloc_ring(info); 858 if (err) 859 return err; 860 pr_debug("%s: %u %u\n", __func__, info->ring_ref, info->evtchn); 861 862 again: 863 err = xenbus_transaction_start(&xbt); 864 if (err) 865 xenbus_dev_fatal(dev, err, "starting transaction"); 866 867 err = xenbus_printf(xbt, dev->nodename, "ring-ref", "%u", 868 info->ring_ref); 869 if (err) { 870 xenbus_dev_fatal(dev, err, "%s", "writing ring-ref"); 871 goto fail; 872 } 873 874 err = xenbus_printf(xbt, dev->nodename, "event-channel", "%u", 875 info->evtchn); 876 877 if (err) { 878 xenbus_dev_fatal(dev, err, "%s", "writing event-channel"); 879 goto fail; 880 } 881 882 err = xenbus_transaction_end(xbt, 0); 883 if (err) { 884 if (err == -EAGAIN) 885 goto again; 886 xenbus_dev_fatal(dev, err, "completing transaction"); 887 goto free_sring; 888 } 889 890 return 0; 891 892 fail: 893 xenbus_transaction_end(xbt, 1); 894 free_sring: 895 scsifront_free_ring(info); 896 897 return err; 898 } 899 900 901 static int scsifront_probe(struct xenbus_device *dev, 902 const struct xenbus_device_id *id) 903 { 904 struct vscsifrnt_info *info; 905 struct Scsi_Host *host; 906 int err = -ENOMEM; 907 char name[TASK_COMM_LEN]; 908 909 host = scsi_host_alloc(&scsifront_sht, sizeof(*info)); 910 if (!host) { 911 xenbus_dev_fatal(dev, err, "fail to allocate scsi host"); 912 return err; 913 } 914 info = (struct vscsifrnt_info *)host->hostdata; 915 916 dev_set_drvdata(&dev->dev, info); 917 info->dev = dev; 918 919 bitmap_fill(info->shadow_free_bitmap, VSCSIIF_MAX_REQS); 920 921 err = scsifront_init_ring(info); 922 if (err) { 923 scsi_host_put(host); 924 return err; 925 } 926 927 init_waitqueue_head(&info->wq_sync); 928 init_waitqueue_head(&info->wq_pause); 929 spin_lock_init(&info->shadow_lock); 930 931 snprintf(name, TASK_COMM_LEN, "vscsiif.%d", host->host_no); 932 933 host->max_id = VSCSIIF_MAX_TARGET; 934 host->max_channel = 0; 935 host->max_lun = VSCSIIF_MAX_LUN; 936 host->max_sectors = (host->sg_tablesize - 1) * PAGE_SIZE / 512; 937 host->max_cmd_len = VSCSIIF_MAX_COMMAND_SIZE; 938 939 err = scsi_add_host(host, &dev->dev); 940 if (err) { 941 dev_err(&dev->dev, "fail to add scsi host %d\n", err); 942 goto free_sring; 943 } 944 info->host = host; 945 info->host_active = STATE_ACTIVE; 946 947 xenbus_switch_state(dev, XenbusStateInitialised); 948 949 return 0; 950 951 free_sring: 952 scsifront_free_ring(info); 953 scsi_host_put(host); 954 return err; 955 } 956 957 static int scsifront_resume(struct xenbus_device *dev) 958 { 959 struct vscsifrnt_info *info = dev_get_drvdata(&dev->dev); 960 struct Scsi_Host *host = info->host; 961 int err; 962 963 spin_lock_irq(host->host_lock); 964 965 /* Finish all still pending commands. */ 966 scsifront_finish_all(info); 967 968 spin_unlock_irq(host->host_lock); 969 970 /* Reconnect to dom0. */ 971 scsifront_free_ring(info); 972 err = scsifront_init_ring(info); 973 if (err) { 974 dev_err(&dev->dev, "fail to resume %d\n", err); 975 scsi_host_put(host); 976 return err; 977 } 978 979 xenbus_switch_state(dev, XenbusStateInitialised); 980 981 return 0; 982 } 983 984 static int scsifront_suspend(struct xenbus_device *dev) 985 { 986 struct vscsifrnt_info *info = dev_get_drvdata(&dev->dev); 987 struct Scsi_Host *host = info->host; 988 int err = 0; 989 990 /* No new commands for the backend. */ 991 spin_lock_irq(host->host_lock); 992 info->pause = 1; 993 while (info->callers && !err) { 994 info->waiting_pause = 1; 995 info->wait_ring_available = 0; 996 spin_unlock_irq(host->host_lock); 997 wake_up(&info->wq_sync); 998 err = wait_event_interruptible(info->wq_pause, 999 !info->waiting_pause); 1000 spin_lock_irq(host->host_lock); 1001 } 1002 spin_unlock_irq(host->host_lock); 1003 return err; 1004 } 1005 1006 static int scsifront_remove(struct xenbus_device *dev) 1007 { 1008 struct vscsifrnt_info *info = dev_get_drvdata(&dev->dev); 1009 1010 pr_debug("%s: %s removed\n", __func__, dev->nodename); 1011 1012 mutex_lock(&scsifront_mutex); 1013 if (info->host_active != STATE_INACTIVE) { 1014 /* Scsi_host not yet removed */ 1015 scsi_remove_host(info->host); 1016 info->host_active = STATE_INACTIVE; 1017 } 1018 mutex_unlock(&scsifront_mutex); 1019 1020 scsifront_free_ring(info); 1021 scsi_host_put(info->host); 1022 1023 return 0; 1024 } 1025 1026 static void scsifront_disconnect(struct vscsifrnt_info *info) 1027 { 1028 struct xenbus_device *dev = info->dev; 1029 struct Scsi_Host *host = info->host; 1030 1031 pr_debug("%s: %s disconnect\n", __func__, dev->nodename); 1032 1033 /* 1034 * When this function is executed, all devices of 1035 * Frontend have been deleted. 1036 * Therefore, it need not block I/O before remove_host. 1037 */ 1038 1039 mutex_lock(&scsifront_mutex); 1040 if (info->host_active != STATE_INACTIVE) { 1041 scsi_remove_host(host); 1042 info->host_active = STATE_INACTIVE; 1043 } 1044 mutex_unlock(&scsifront_mutex); 1045 1046 xenbus_frontend_closed(dev); 1047 } 1048 1049 static void scsifront_do_lun_hotplug(struct vscsifrnt_info *info, int op) 1050 { 1051 struct xenbus_device *dev = info->dev; 1052 int i, err = 0; 1053 char str[64]; 1054 char **dir; 1055 unsigned int dir_n = 0; 1056 unsigned int device_state; 1057 unsigned int hst, chn, tgt, lun; 1058 struct scsi_device *sdev; 1059 1060 if (info->host_active == STATE_ERROR) 1061 return; 1062 1063 dir = xenbus_directory(XBT_NIL, dev->otherend, "vscsi-devs", &dir_n); 1064 if (IS_ERR(dir)) 1065 return; 1066 1067 /* mark current task as the one allowed to modify device states */ 1068 BUG_ON(info->curr); 1069 info->curr = current; 1070 1071 for (i = 0; i < dir_n; i++) { 1072 /* read status */ 1073 snprintf(str, sizeof(str), "vscsi-devs/%s/state", dir[i]); 1074 err = xenbus_scanf(XBT_NIL, dev->otherend, str, "%u", 1075 &device_state); 1076 if (XENBUS_EXIST_ERR(err)) 1077 continue; 1078 1079 /* virtual SCSI device */ 1080 snprintf(str, sizeof(str), "vscsi-devs/%s/v-dev", dir[i]); 1081 err = xenbus_scanf(XBT_NIL, dev->otherend, str, 1082 "%u:%u:%u:%u", &hst, &chn, &tgt, &lun); 1083 if (XENBUS_EXIST_ERR(err)) 1084 continue; 1085 1086 /* 1087 * Front device state path, used in slave_configure called 1088 * on successfull scsi_add_device, and in slave_destroy called 1089 * on remove of a device. 1090 */ 1091 snprintf(info->dev_state_path, sizeof(info->dev_state_path), 1092 "vscsi-devs/%s/state", dir[i]); 1093 1094 switch (op) { 1095 case VSCSIFRONT_OP_ADD_LUN: 1096 if (device_state != XenbusStateInitialised) 1097 break; 1098 1099 if (scsi_add_device(info->host, chn, tgt, lun)) { 1100 dev_err(&dev->dev, "scsi_add_device\n"); 1101 err = xenbus_printf(XBT_NIL, dev->nodename, 1102 info->dev_state_path, 1103 "%d", XenbusStateClosed); 1104 if (err) 1105 xenbus_dev_error(dev, err, 1106 "%s: writing dev_state_path", __func__); 1107 } 1108 break; 1109 case VSCSIFRONT_OP_DEL_LUN: 1110 if (device_state != XenbusStateClosing) 1111 break; 1112 1113 sdev = scsi_device_lookup(info->host, chn, tgt, lun); 1114 if (sdev) { 1115 scsi_remove_device(sdev); 1116 scsi_device_put(sdev); 1117 } 1118 break; 1119 case VSCSIFRONT_OP_READD_LUN: 1120 if (device_state == XenbusStateConnected) { 1121 err = xenbus_printf(XBT_NIL, dev->nodename, 1122 info->dev_state_path, 1123 "%d", XenbusStateConnected); 1124 if (err) 1125 xenbus_dev_error(dev, err, 1126 "%s: writing dev_state_path", __func__); 1127 } 1128 break; 1129 default: 1130 break; 1131 } 1132 } 1133 1134 info->curr = NULL; 1135 1136 kfree(dir); 1137 } 1138 1139 static void scsifront_read_backend_params(struct xenbus_device *dev, 1140 struct vscsifrnt_info *info) 1141 { 1142 unsigned int sg_grant, nr_segs; 1143 struct Scsi_Host *host = info->host; 1144 1145 sg_grant = xenbus_read_unsigned(dev->otherend, "feature-sg-grant", 0); 1146 nr_segs = min_t(unsigned int, sg_grant, SG_ALL); 1147 nr_segs = max_t(unsigned int, nr_segs, VSCSIIF_SG_TABLESIZE); 1148 nr_segs = min_t(unsigned int, nr_segs, 1149 VSCSIIF_SG_TABLESIZE * PAGE_SIZE / 1150 sizeof(struct scsiif_request_segment)); 1151 1152 if (!info->pause && sg_grant) 1153 dev_info(&dev->dev, "using up to %d SG entries\n", nr_segs); 1154 else if (info->pause && nr_segs < host->sg_tablesize) 1155 dev_warn(&dev->dev, 1156 "SG entries decreased from %d to %u - device may not work properly anymore\n", 1157 host->sg_tablesize, nr_segs); 1158 1159 host->sg_tablesize = nr_segs; 1160 host->max_sectors = (nr_segs - 1) * PAGE_SIZE / 512; 1161 } 1162 1163 static void scsifront_backend_changed(struct xenbus_device *dev, 1164 enum xenbus_state backend_state) 1165 { 1166 struct vscsifrnt_info *info = dev_get_drvdata(&dev->dev); 1167 1168 pr_debug("%s: %p %u %u\n", __func__, dev, dev->state, backend_state); 1169 1170 switch (backend_state) { 1171 case XenbusStateUnknown: 1172 case XenbusStateInitialising: 1173 case XenbusStateInitWait: 1174 case XenbusStateInitialised: 1175 break; 1176 1177 case XenbusStateConnected: 1178 scsifront_read_backend_params(dev, info); 1179 1180 if (info->pause) { 1181 scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_READD_LUN); 1182 xenbus_switch_state(dev, XenbusStateConnected); 1183 info->pause = 0; 1184 return; 1185 } 1186 1187 if (xenbus_read_driver_state(dev->nodename) == 1188 XenbusStateInitialised) 1189 scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_ADD_LUN); 1190 1191 if (dev->state != XenbusStateConnected) 1192 xenbus_switch_state(dev, XenbusStateConnected); 1193 break; 1194 1195 case XenbusStateClosed: 1196 if (dev->state == XenbusStateClosed) 1197 break; 1198 fallthrough; /* Missed the backend's Closing state */ 1199 case XenbusStateClosing: 1200 scsifront_disconnect(info); 1201 break; 1202 1203 case XenbusStateReconfiguring: 1204 scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_DEL_LUN); 1205 xenbus_switch_state(dev, XenbusStateReconfiguring); 1206 break; 1207 1208 case XenbusStateReconfigured: 1209 scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_ADD_LUN); 1210 xenbus_switch_state(dev, XenbusStateConnected); 1211 break; 1212 } 1213 } 1214 1215 static const struct xenbus_device_id scsifront_ids[] = { 1216 { "vscsi" }, 1217 { "" } 1218 }; 1219 1220 static struct xenbus_driver scsifront_driver = { 1221 .ids = scsifront_ids, 1222 .probe = scsifront_probe, 1223 .remove = scsifront_remove, 1224 .resume = scsifront_resume, 1225 .suspend = scsifront_suspend, 1226 .otherend_changed = scsifront_backend_changed, 1227 }; 1228 1229 static int __init scsifront_init(void) 1230 { 1231 if (!xen_domain()) 1232 return -ENODEV; 1233 1234 return xenbus_register_frontend(&scsifront_driver); 1235 } 1236 module_init(scsifront_init); 1237 1238 static void __exit scsifront_exit(void) 1239 { 1240 xenbus_unregister_driver(&scsifront_driver); 1241 } 1242 module_exit(scsifront_exit); 1243 1244 MODULE_DESCRIPTION("Xen SCSI frontend driver"); 1245 MODULE_LICENSE("GPL"); 1246 MODULE_ALIAS("xen:vscsi"); 1247 MODULE_AUTHOR("Juergen Gross <jgross@suse.com>"); 1248