Pileus Git - ~andy/linux/blob - drivers/block/xen-blkback/blkback.c

   1 /******************************************************************************
   2  *
   3  * Back-end of the driver for virtual block devices. This portion of the
   4  * driver exports a 'unified' block-device interface that can be accessed
   5  * by any operating system that implements a compatible front end. A
   6  * reference front-end implementation can be found in:
   7  *  drivers/block/xen-blkfront.c
   8  *
   9  * Copyright (c) 2003-2004, Keir Fraser & Steve Hand
  10  * Copyright (c) 2005, Christopher Clark
  11  *
  12  * This program is free software; you can redistribute it and/or
  13  * modify it under the terms of the GNU General Public License version 2
  14  * as published by the Free Software Foundation; or, when distributed
  15  * separately from the Linux kernel or incorporated into other
  16  * software packages, subject to the following license:
  17  *
  18  * Permission is hereby granted, free of charge, to any person obtaining a copy
  19  * of this source file (the "Software"), to deal in the Software without
  20  * restriction, including without limitation the rights to use, copy, modify,
  21  * merge, publish, distribute, sublicense, and/or sell copies of the Software,
  22  * and to permit persons to whom the Software is furnished to do so, subject to
  23  * the following conditions:
  24  *
  25  * The above copyright notice and this permission notice shall be included in
  26  * all copies or substantial portions of the Software.
  27  *
  28  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  29  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  30  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  31  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  32  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  33  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
  34  * IN THE SOFTWARE.
  35  */
  36
  37 #include <linux/spinlock.h>
  38 #include <linux/kthread.h>
  39 #include <linux/list.h>
  40 #include <linux/delay.h>
  41 #include <linux/freezer.h>
  42 #include <linux/bitmap.h>
  43
  44 #include <xen/events.h>
  45 #include <xen/page.h>
  46 #include <xen/xen.h>
  47 #include <asm/xen/hypervisor.h>
  48 #include <asm/xen/hypercall.h>
  49 #include <xen/balloon.h>
  50 #include "common.h"
  51
  52 /*
  53  * Maximum number of unused free pages to keep in the internal buffer.
  54  * Setting this to a value too low will reduce memory used in each backend,
  55  * but can have a performance penalty.
  56  *
  57  * A sane value is xen_blkif_reqs * BLKIF_MAX_SEGMENTS_PER_REQUEST, but can
  58  * be set to a lower value that might degrade performance on some intensive
  59  * IO workloads.
  60  */
  61
  62 static int xen_blkif_max_buffer_pages = 1024;
  63 module_param_named(max_buffer_pages, xen_blkif_max_buffer_pages, int, 0644);
  64 MODULE_PARM_DESC(max_buffer_pages,
  65 "Maximum number of free pages to keep in each block backend buffer");
  66
  67 /*
  68  * Maximum number of grants to map persistently in blkback. For maximum
  69  * performance this should be the total numbers of grants that can be used
  70  * to fill the ring, but since this might become too high, specially with
  71  * the use of indirect descriptors, we set it to a value that provides good
  72  * performance without using too much memory.
  73  *
  74  * When the list of persistent grants is full we clean it up using a LRU
  75  * algorithm.
  76  */
  77
  78 static int xen_blkif_max_pgrants = 1056;
  79 module_param_named(max_persistent_grants, xen_blkif_max_pgrants, int, 0644);
  80 MODULE_PARM_DESC(max_persistent_grants,
  81                  "Maximum number of grants to map persistently");
  82
  83 /*
  84  * The LRU mechanism to clean the lists of persistent grants needs to
  85  * be executed periodically. The time interval between consecutive executions
  86  * of the purge mechanism is set in ms.
  87  */
  88 #define LRU_INTERVAL 100
  89
  90 /*
  91  * When the persistent grants list is full we will remove unused grants
  92  * from the list. The percent number of grants to be removed at each LRU
  93  * execution.
  94  */
  95 #define LRU_PERCENT_CLEAN 5
  96
  97 /* Run-time switchable: /sys/module/blkback/parameters/ */
  98 static unsigned int log_stats;
  99 module_param(log_stats, int, 0644);
 100
 101 #define BLKBACK_INVALID_HANDLE (~0)
 102
 103 /* Number of free pages to remove on each call to free_xenballooned_pages */
 104 #define NUM_BATCH_FREE_PAGES 10
 105
 106 static inline int get_free_page(struct xen_blkif *blkif, struct page **page)
 107 {
 108         unsigned long flags;
 109
 110         spin_lock_irqsave(&blkif->free_pages_lock, flags);
 111         if (list_empty(&blkif->free_pages)) {
 112                 BUG_ON(blkif->free_pages_num != 0);
 113                 spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
 114                 return alloc_xenballooned_pages(1, page, false);
 115         }
 116         BUG_ON(blkif->free_pages_num == 0);
 117         page[0] = list_first_entry(&blkif->free_pages, struct page, lru);
 118         list_del(&page[0]->lru);
 119         blkif->free_pages_num--;
 120         spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
 121
 122         return 0;
 123 }
 124
 125 static inline void put_free_pages(struct xen_blkif *blkif, struct page **page,
 126                                   int num)
 127 {
 128         unsigned long flags;
 129         int i;
 130
 131         spin_lock_irqsave(&blkif->free_pages_lock, flags);
 132         for (i = 0; i < num; i++)
 133                 list_add(&page[i]->lru, &blkif->free_pages);
 134         blkif->free_pages_num += num;
 135         spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
 136 }
 137
 138 static inline void shrink_free_pagepool(struct xen_blkif *blkif, int num)
 139 {
 140         /* Remove requested pages in batches of NUM_BATCH_FREE_PAGES */
 141         struct page *page[NUM_BATCH_FREE_PAGES];
 142         unsigned int num_pages = 0;
 143         unsigned long flags;
 144
 145         spin_lock_irqsave(&blkif->free_pages_lock, flags);
 146         while (blkif->free_pages_num > num) {
 147                 BUG_ON(list_empty(&blkif->free_pages));
 148                 page[num_pages] = list_first_entry(&blkif->free_pages,
 149                                                    struct page, lru);
 150                 list_del(&page[num_pages]->lru);
 151                 blkif->free_pages_num--;
 152                 if (++num_pages == NUM_BATCH_FREE_PAGES) {
 153                         spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
 154                         free_xenballooned_pages(num_pages, page);
 155                         spin_lock_irqsave(&blkif->free_pages_lock, flags);
 156                         num_pages = 0;
 157                 }
 158         }
 159         spin_unlock_irqrestore(&blkif->free_pages_lock, flags);
 160         if (num_pages != 0)
 161                 free_xenballooned_pages(num_pages, page);
 162 }
 163
 164 #define vaddr(page) ((unsigned long)pfn_to_kaddr(page_to_pfn(page)))
 165
 166 static int do_block_io_op(struct xen_blkif *blkif);
 167 static int dispatch_rw_block_io(struct xen_blkif *blkif,
 168                                 struct blkif_request *req,
 169                                 struct pending_req *pending_req);
 170 static void make_response(struct xen_blkif *blkif, u64 id,
 171                           unsigned short op, int st);
 172
 173 #define foreach_grant_safe(pos, n, rbtree, node) \
 174         for ((pos) = container_of(rb_first((rbtree)), typeof(*(pos)), node), \
 175              (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL; \
 176              &(pos)->node != NULL; \
 177              (pos) = container_of(n, typeof(*(pos)), node), \
 178              (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL)
 179
 180
 181 /*
 182  * We don't need locking around the persistent grant helpers
 183  * because blkback uses a single-thread for each backed, so we
 184  * can be sure that this functions will never be called recursively.
 185  *
 186  * The only exception to that is put_persistent_grant, that can be called
 187  * from interrupt context (by xen_blkbk_unmap), so we have to use atomic
 188  * bit operations to modify the flags of a persistent grant and to count
 189  * the number of used grants.
 190  */
 191 static int add_persistent_gnt(struct xen_blkif *blkif,
 192                                struct persistent_gnt *persistent_gnt)
 193 {
 194         struct rb_node **new = NULL, *parent = NULL;
 195         struct persistent_gnt *this;
 196
 197         if (blkif->persistent_gnt_c >= xen_blkif_max_pgrants) {
 198                 if (!blkif->vbd.overflow_max_grants)
 199                         blkif->vbd.overflow_max_grants = 1;
 200                 return -EBUSY;
 201         }
 202         /* Figure out where to put new node */
 203         new = &blkif->persistent_gnts.rb_node;
 204         while (*new) {
 205                 this = container_of(*new, struct persistent_gnt, node);
 206
 207                 parent = *new;
 208                 if (persistent_gnt->gnt < this->gnt)
 209                         new = &((*new)->rb_left);
 210                 else if (persistent_gnt->gnt > this->gnt)
 211                         new = &((*new)->rb_right);
 212                 else {
 213                         pr_alert_ratelimited(DRV_PFX " trying to add a gref that's already in the tree\n");
 214                         return -EINVAL;
 215                 }
 216         }
 217
 218         bitmap_zero(persistent_gnt->flags, PERSISTENT_GNT_FLAGS_SIZE);
 219         set_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags);
 220         /* Add new node and rebalance tree. */
 221         rb_link_node(&(persistent_gnt->node), parent, new);
 222         rb_insert_color(&(persistent_gnt->node), &blkif->persistent_gnts);
 223         blkif->persistent_gnt_c++;
 224         atomic_inc(&blkif->persistent_gnt_in_use);
 225         return 0;
 226 }
 227
 228 static struct persistent_gnt *get_persistent_gnt(struct xen_blkif *blkif,
 229                                                  grant_ref_t gref)
 230 {
 231         struct persistent_gnt *data;
 232         struct rb_node *node = NULL;
 233
 234         node = blkif->persistent_gnts.rb_node;
 235         while (node) {
 236                 data = container_of(node, struct persistent_gnt, node);
 237
 238                 if (gref < data->gnt)
 239                         node = node->rb_left;
 240                 else if (gref > data->gnt)
 241                         node = node->rb_right;
 242                 else {
 243                         if(test_bit(PERSISTENT_GNT_ACTIVE, data->flags)) {
 244                                 pr_alert_ratelimited(DRV_PFX " requesting a grant already in use\n");
 245                                 return NULL;
 246                         }
 247                         set_bit(PERSISTENT_GNT_ACTIVE, data->flags);
 248                         atomic_inc(&blkif->persistent_gnt_in_use);
 249                         return data;
 250                 }
 251         }
 252         return NULL;
 253 }
 254
 255 static void put_persistent_gnt(struct xen_blkif *blkif,
 256                                struct persistent_gnt *persistent_gnt)
 257 {
 258         if(!test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags))
 259                   pr_alert_ratelimited(DRV_PFX " freeing a grant already unused");
 260         set_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags);
 261         clear_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags);
 262         atomic_dec(&blkif->persistent_gnt_in_use);
 263 }
 264
 265 static void free_persistent_gnts(struct xen_blkif *blkif, struct rb_root *root,
 266                                  unsigned int num)
 267 {
 268         struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 269         struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 270         struct persistent_gnt *persistent_gnt;
 271         struct rb_node *n;
 272         int ret = 0;
 273         int segs_to_unmap = 0;
 274
 275         foreach_grant_safe(persistent_gnt, n, root, node) {
 276                 BUG_ON(persistent_gnt->handle ==
 277                         BLKBACK_INVALID_HANDLE);
 278                 gnttab_set_unmap_op(&unmap[segs_to_unmap],
 279                         (unsigned long) pfn_to_kaddr(page_to_pfn(
 280                                 persistent_gnt->page)),
 281                         GNTMAP_host_map,
 282                         persistent_gnt->handle);
 283
 284                 pages[segs_to_unmap] = persistent_gnt->page;
 285
 286                 if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST ||
 287                         !rb_next(&persistent_gnt->node)) {
 288                         ret = gnttab_unmap_refs(unmap, NULL, pages,
 289                                 segs_to_unmap);
 290                         BUG_ON(ret);
 291                         put_free_pages(blkif, pages, segs_to_unmap);
 292                         segs_to_unmap = 0;
 293                 }
 294
 295                 rb_erase(&persistent_gnt->node, root);
 296                 kfree(persistent_gnt);
 297                 num--;
 298         }
 299         BUG_ON(num != 0);
 300 }
 301
 302 static void unmap_purged_grants(struct work_struct *work)
 303 {
 304         struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 305         struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 306         struct persistent_gnt *persistent_gnt;
 307         int ret, segs_to_unmap = 0;
 308         struct xen_blkif *blkif = container_of(work, typeof(*blkif), persistent_purge_work);
 309
 310         while(!list_empty(&blkif->persistent_purge_list)) {
 311                 persistent_gnt = list_first_entry(&blkif->persistent_purge_list,
 312                                                   struct persistent_gnt,
 313                                                   remove_node);
 314                 list_del(&persistent_gnt->remove_node);
 315
 316                 gnttab_set_unmap_op(&unmap[segs_to_unmap],
 317                         vaddr(persistent_gnt->page),
 318                         GNTMAP_host_map,
 319                         persistent_gnt->handle);
 320
 321                 pages[segs_to_unmap] = persistent_gnt->page;
 322
 323                 if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST) {
 324                         ret = gnttab_unmap_refs(unmap, NULL, pages,
 325                                 segs_to_unmap);
 326                         BUG_ON(ret);
 327                         put_free_pages(blkif, pages, segs_to_unmap);
 328                         segs_to_unmap = 0;
 329                 }
 330                 kfree(persistent_gnt);
 331         }
 332         if (segs_to_unmap > 0) {
 333                 ret = gnttab_unmap_refs(unmap, NULL, pages, segs_to_unmap);
 334                 BUG_ON(ret);
 335                 put_free_pages(blkif, pages, segs_to_unmap);
 336         }
 337 }
 338
 339 static void purge_persistent_gnt(struct xen_blkif *blkif)
 340 {
 341         struct persistent_gnt *persistent_gnt;
 342         struct rb_node *n;
 343         unsigned int num_clean, total;
 344         bool scan_used = false, clean_used = false;
 345         struct rb_root *root;
 346
 347         if (blkif->persistent_gnt_c < xen_blkif_max_pgrants ||
 348             (blkif->persistent_gnt_c == xen_blkif_max_pgrants &&
 349             !blkif->vbd.overflow_max_grants)) {
 350                 return;
 351         }
 352
 353         if (work_pending(&blkif->persistent_purge_work)) {
 354                 pr_alert_ratelimited(DRV_PFX "Scheduled work from previous purge is still pending, cannot purge list\n");
 355                 return;
 356         }
 357
 358         num_clean = (xen_blkif_max_pgrants / 100) * LRU_PERCENT_CLEAN;
 359         num_clean = blkif->persistent_gnt_c - xen_blkif_max_pgrants + num_clean;
 360         num_clean = min(blkif->persistent_gnt_c, num_clean);
 361         if ((num_clean == 0) ||
 362             (num_clean > (blkif->persistent_gnt_c - atomic_read(&blkif->persistent_gnt_in_use))))
 363                 return;
 364
 365         /*
 366          * At this point, we can assure that there will be no calls
 367          * to get_persistent_grant (because we are executing this code from
 368          * xen_blkif_schedule), there can only be calls to put_persistent_gnt,
 369          * which means that the number of currently used grants will go down,
 370          * but never up, so we will always be able to remove the requested
 371          * number of grants.
 372          */
 373
 374         total = num_clean;
 375
 376         pr_debug(DRV_PFX "Going to purge %u persistent grants\n", num_clean);
 377
 378         INIT_LIST_HEAD(&blkif->persistent_purge_list);
 379         root = &blkif->persistent_gnts;
 380 purge_list:
 381         foreach_grant_safe(persistent_gnt, n, root, node) {
 382                 BUG_ON(persistent_gnt->handle ==
 383                         BLKBACK_INVALID_HANDLE);
 384
 385                 if (clean_used) {
 386                         clear_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags);
 387                         continue;
 388                 }
 389
 390                 if (test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags))
 391                         continue;
 392                 if (!scan_used &&
 393                     (test_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags)))
 394                         continue;
 395
 396                 rb_erase(&persistent_gnt->node, root);
 397                 list_add(&persistent_gnt->remove_node,
 398                          &blkif->persistent_purge_list);
 399                 if (--num_clean == 0)
 400                         goto finished;
 401         }
 402         /*
 403          * If we get here it means we also need to start cleaning
 404          * grants that were used since last purge in order to cope
 405          * with the requested num
 406          */
 407         if (!scan_used && !clean_used) {
 408                 pr_debug(DRV_PFX "Still missing %u purged frames\n", num_clean);
 409                 scan_used = true;
 410                 goto purge_list;
 411         }
 412 finished:
 413         if (!clean_used) {
 414                 pr_debug(DRV_PFX "Finished scanning for grants to clean, removing used flag\n");
 415                 clean_used = true;
 416                 goto purge_list;
 417         }
 418
 419         blkif->persistent_gnt_c -= (total - num_clean);
 420         blkif->vbd.overflow_max_grants = 0;
 421
 422         /* We can defer this work */
 423         INIT_WORK(&blkif->persistent_purge_work, unmap_purged_grants);
 424         schedule_work(&blkif->persistent_purge_work);
 425         pr_debug(DRV_PFX "Purged %u/%u\n", (total - num_clean), total);
 426         return;
 427 }
 428
 429 /*
 430  * Retrieve from the 'pending_reqs' a free pending_req structure to be used.
 431  */
 432 static struct pending_req *alloc_req(struct xen_blkif *blkif)
 433 {
 434         struct pending_req *req = NULL;
 435         unsigned long flags;
 436
 437         spin_lock_irqsave(&blkif->pending_free_lock, flags);
 438         if (!list_empty(&blkif->pending_free)) {
 439                 req = list_entry(blkif->pending_free.next, struct pending_req,
 440                                  free_list);
 441                 list_del(&req->free_list);
 442         }
 443         spin_unlock_irqrestore(&blkif->pending_free_lock, flags);
 444         return req;
 445 }
 446
 447 /*
 448  * Return the 'pending_req' structure back to the freepool. We also
 449  * wake up the thread if it was waiting for a free page.
 450  */
 451 static void free_req(struct xen_blkif *blkif, struct pending_req *req)
 452 {
 453         unsigned long flags;
 454         int was_empty;
 455
 456         spin_lock_irqsave(&blkif->pending_free_lock, flags);
 457         was_empty = list_empty(&blkif->pending_free);
 458         list_add(&req->free_list, &blkif->pending_free);
 459         spin_unlock_irqrestore(&blkif->pending_free_lock, flags);
 460         if (was_empty)
 461                 wake_up(&blkif->pending_free_wq);
 462 }
 463
 464 /*
 465  * Routines for managing virtual block devices (vbds).
 466  */
 467 static int xen_vbd_translate(struct phys_req *req, struct xen_blkif *blkif,
 468                              int operation)
 469 {
 470         struct xen_vbd *vbd = &blkif->vbd;
 471         int rc = -EACCES;
 472
 473         if ((operation != READ) && vbd->readonly)
 474                 goto out;
 475
 476         if (likely(req->nr_sects)) {
 477                 blkif_sector_t end = req->sector_number + req->nr_sects;
 478
 479                 if (unlikely(end < req->sector_number))
 480                         goto out;
 481                 if (unlikely(end > vbd_sz(vbd)))
 482                         goto out;
 483         }
 484
 485         req->dev  = vbd->pdevice;
 486         req->bdev = vbd->bdev;
 487         rc = 0;
 488
 489  out:
 490         return rc;
 491 }
 492
 493 static void xen_vbd_resize(struct xen_blkif *blkif)
 494 {
 495         struct xen_vbd *vbd = &blkif->vbd;
 496         struct xenbus_transaction xbt;
 497         int err;
 498         struct xenbus_device *dev = xen_blkbk_xenbus(blkif->be);
 499         unsigned long long new_size = vbd_sz(vbd);
 500
 501         pr_info(DRV_PFX "VBD Resize: Domid: %d, Device: (%d, %d)\n",
 502                 blkif->domid, MAJOR(vbd->pdevice), MINOR(vbd->pdevice));
 503         pr_info(DRV_PFX "VBD Resize: new size %llu\n", new_size);
 504         vbd->size = new_size;
 505 again:
 506         err = xenbus_transaction_start(&xbt);
 507         if (err) {
 508                 pr_warn(DRV_PFX "Error starting transaction");
 509                 return;
 510         }
 511         err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu",
 512                             (unsigned long long)vbd_sz(vbd));
 513         if (err) {
 514                 pr_warn(DRV_PFX "Error writing new size");
 515                 goto abort;
 516         }
 517         /*
 518          * Write the current state; we will use this to synchronize
 519          * the front-end. If the current state is "connected" the
 520          * front-end will get the new size information online.
 521          */
 522         err = xenbus_printf(xbt, dev->nodename, "state", "%d", dev->state);
 523         if (err) {
 524                 pr_warn(DRV_PFX "Error writing the state");
 525                 goto abort;
 526         }
 527
 528         err = xenbus_transaction_end(xbt, 0);
 529         if (err == -EAGAIN)
 530                 goto again;
 531         if (err)
 532                 pr_warn(DRV_PFX "Error ending transaction");
 533         return;
 534 abort:
 535         xenbus_transaction_end(xbt, 1);
 536 }
 537
 538 /*
 539  * Notification from the guest OS.
 540  */
 541 static void blkif_notify_work(struct xen_blkif *blkif)
 542 {
 543         blkif->waiting_reqs = 1;
 544         wake_up(&blkif->wq);
 545 }
 546
 547 irqreturn_t xen_blkif_be_int(int irq, void *dev_id)
 548 {
 549         blkif_notify_work(dev_id);
 550         return IRQ_HANDLED;
 551 }
 552
 553 /*
 554  * SCHEDULER FUNCTIONS
 555  */
 556
 557 static void print_stats(struct xen_blkif *blkif)
 558 {
 559         pr_info("xen-blkback (%s): oo %3llu  |  rd %4llu  |  wr %4llu  |  f %4llu"
 560                  "  |  ds %4llu | pg: %4u/%4d\n",
 561                  current->comm, blkif->st_oo_req,
 562                  blkif->st_rd_req, blkif->st_wr_req,
 563                  blkif->st_f_req, blkif->st_ds_req,
 564                  blkif->persistent_gnt_c,
 565                  xen_blkif_max_pgrants);
 566         blkif->st_print = jiffies + msecs_to_jiffies(10 * 1000);
 567         blkif->st_rd_req = 0;
 568         blkif->st_wr_req = 0;
 569         blkif->st_oo_req = 0;
 570         blkif->st_ds_req = 0;
 571 }
 572
 573 int xen_blkif_schedule(void *arg)
 574 {
 575         struct xen_blkif *blkif = arg;
 576         struct xen_vbd *vbd = &blkif->vbd;
 577         unsigned long timeout;
 578         int ret;
 579
 580         xen_blkif_get(blkif);
 581
 582         while (!kthread_should_stop()) {
 583                 if (try_to_freeze())
 584                         continue;
 585                 if (unlikely(vbd->size != vbd_sz(vbd)))
 586                         xen_vbd_resize(blkif);
 587
 588                 timeout = msecs_to_jiffies(LRU_INTERVAL);
 589
 590                 timeout = wait_event_interruptible_timeout(
 591                         blkif->wq,
 592                         blkif->waiting_reqs || kthread_should_stop(),
 593                         timeout);
 594                 if (timeout == 0)
 595                         goto purge_gnt_list;
 596                 timeout = wait_event_interruptible_timeout(
 597                         blkif->pending_free_wq,
 598                         !list_empty(&blkif->pending_free) ||
 599                         kthread_should_stop(),
 600                         timeout);
 601                 if (timeout == 0)
 602                         goto purge_gnt_list;
 603
 604                 blkif->waiting_reqs = 0;
 605                 smp_mb(); /* clear flag *before* checking for work */
 606
 607                 ret = do_block_io_op(blkif);
 608                 if (ret > 0)
 609                         blkif->waiting_reqs = 1;
 610                 if (ret == -EACCES)
 611                         wait_event_interruptible(blkif->shutdown_wq,
 612                                                  kthread_should_stop());
 613
 614 purge_gnt_list:
 615                 if (blkif->vbd.feature_gnt_persistent &&
 616                     time_after(jiffies, blkif->next_lru)) {
 617                         purge_persistent_gnt(blkif);
 618                         blkif->next_lru = jiffies + msecs_to_jiffies(LRU_INTERVAL);
 619                 }
 620
 621                 /* Shrink if we have more than xen_blkif_max_buffer_pages */
 622                 shrink_free_pagepool(blkif, xen_blkif_max_buffer_pages);
 623
 624                 if (log_stats && time_after(jiffies, blkif->st_print))
 625                         print_stats(blkif);
 626         }
 627
 628         /* Since we are shutting down remove all pages from the buffer */
 629         shrink_free_pagepool(blkif, 0 /* All */);
 630
 631         /* Free all persistent grant pages */
 632         if (!RB_EMPTY_ROOT(&blkif->persistent_gnts))
 633                 free_persistent_gnts(blkif, &blkif->persistent_gnts,
 634                         blkif->persistent_gnt_c);
 635
 636         BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts));
 637         blkif->persistent_gnt_c = 0;
 638
 639         if (log_stats)
 640                 print_stats(blkif);
 641
 642         blkif->xenblkd = NULL;
 643         xen_blkif_put(blkif);
 644
 645         return 0;
 646 }
 647
 648 /*
 649  * Unmap the grant references, and also remove the M2P over-rides
 650  * used in the 'pending_req'.
 651  */
 652 static void xen_blkbk_unmap(struct xen_blkif *blkif,
 653                             struct grant_page *pages[],
 654                             int num)
 655 {
 656         struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 657         struct page *unmap_pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 658         unsigned int i, invcount = 0;
 659         int ret;
 660
 661         for (i = 0; i < num; i++) {
 662                 if (pages[i]->persistent_gnt != NULL) {
 663                         put_persistent_gnt(blkif, pages[i]->persistent_gnt);
 664                         continue;
 665                 }
 666                 if (pages[i]->handle == BLKBACK_INVALID_HANDLE)
 667                         continue;
 668                 unmap_pages[invcount] = pages[i]->page;
 669                 gnttab_set_unmap_op(&unmap[invcount], vaddr(pages[i]->page),
 670                                     GNTMAP_host_map, pages[i]->handle);
 671                 pages[i]->handle = BLKBACK_INVALID_HANDLE;
 672                 if (++invcount == BLKIF_MAX_SEGMENTS_PER_REQUEST) {
 673                         ret = gnttab_unmap_refs(unmap, NULL, unmap_pages,
 674                                                 invcount);
 675                         BUG_ON(ret);
 676                         put_free_pages(blkif, unmap_pages, invcount);
 677                         invcount = 0;
 678                 }
 679         }
 680         if (invcount) {
 681                 ret = gnttab_unmap_refs(unmap, NULL, unmap_pages, invcount);
 682                 BUG_ON(ret);
 683                 put_free_pages(blkif, unmap_pages, invcount);
 684         }
 685 }
 686
 687 static int xen_blkbk_map(struct xen_blkif *blkif,
 688                          struct grant_page *pages[],
 689                          int num, bool ro)
 690 {
 691         struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 692         struct page *pages_to_gnt[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 693         struct persistent_gnt *persistent_gnt = NULL;
 694         phys_addr_t addr = 0;
 695         int i, seg_idx, new_map_idx;
 696         int segs_to_map = 0;
 697         int ret = 0;
 698         int last_map = 0, map_until = 0;
 699         int use_persistent_gnts;
 700
 701         use_persistent_gnts = (blkif->vbd.feature_gnt_persistent);
 702
 703         /*
 704          * Fill out preq.nr_sects with proper amount of sectors, and setup
 705          * assign map[..] with the PFN of the page in our domain with the
 706          * corresponding grant reference for each page.
 707          */
 708 again:
 709         for (i = map_until; i < num; i++) {
 710                 uint32_t flags;
 711
 712                 if (use_persistent_gnts)
 713                         persistent_gnt = get_persistent_gnt(
 714                                 blkif,
 715                                 pages[i]->gref);
 716
 717                 if (persistent_gnt) {
 718                         /*
 719                          * We are using persistent grants and
 720                          * the grant is already mapped
 721                          */
 722                         pages[i]->page = persistent_gnt->page;
 723                         pages[i]->persistent_gnt = persistent_gnt;
 724                 } else {
 725                         if (get_free_page(blkif, &pages[i]->page))
 726                                 goto out_of_memory;
 727                         addr = vaddr(pages[i]->page);
 728                         pages_to_gnt[segs_to_map] = pages[i]->page;
 729                         pages[i]->persistent_gnt = NULL;
 730                         flags = GNTMAP_host_map;
 731                         if (!use_persistent_gnts && ro)
 732                                 flags |= GNTMAP_readonly;
 733                         gnttab_set_map_op(&map[segs_to_map++], addr,
 734                                           flags, pages[i]->gref,
 735                                           blkif->domid);
 736                 }
 737                 map_until = i + 1;
 738                 if (segs_to_map == BLKIF_MAX_SEGMENTS_PER_REQUEST)
 739                         break;
 740         }
 741
 742         if (segs_to_map) {
 743                 ret = gnttab_map_refs(map, NULL, pages_to_gnt, segs_to_map);
 744                 BUG_ON(ret);
 745         }
 746
 747         /*
 748          * Now swizzle the MFN in our domain with the MFN from the other domain
 749          * so that when we access vaddr(pending_req,i) it has the contents of
 750          * the page from the other domain.
 751          */
 752         for (seg_idx = last_map, new_map_idx = 0; seg_idx < map_until; seg_idx++) {
 753                 if (!pages[seg_idx]->persistent_gnt) {
 754                         /* This is a newly mapped grant */
 755                         BUG_ON(new_map_idx >= segs_to_map);
 756                         if (unlikely(map[new_map_idx].status != 0)) {
 757                                 pr_debug(DRV_PFX "invalid buffer -- could not remap it\n");
 758                                 pages[seg_idx]->handle = BLKBACK_INVALID_HANDLE;
 759                                 ret |= 1;
 760                                 goto next;
 761                         }
 762                         pages[seg_idx]->handle = map[new_map_idx].handle;
 763                 } else {
 764                         continue;
 765                 }
 766                 if (use_persistent_gnts &&
 767                     blkif->persistent_gnt_c < xen_blkif_max_pgrants) {
 768                         /*
 769                          * We are using persistent grants, the grant is
 770                          * not mapped but we might have room for it.
 771                          */
 772                         persistent_gnt = kmalloc(sizeof(struct persistent_gnt),
 773                                                  GFP_KERNEL);
 774                         if (!persistent_gnt) {
 775                                 /*
 776                                  * If we don't have enough memory to
 777                                  * allocate the persistent_gnt struct
 778                                  * map this grant non-persistenly
 779                                  */
 780                                 goto next;
 781                         }
 782                         persistent_gnt->gnt = map[new_map_idx].ref;
 783                         persistent_gnt->handle = map[new_map_idx].handle;
 784                         persistent_gnt->page = pages[seg_idx]->page;
 785                         if (add_persistent_gnt(blkif,
 786                                                persistent_gnt)) {
 787                                 kfree(persistent_gnt);
 788                                 persistent_gnt = NULL;
 789                                 goto next;
 790                         }
 791                         pages[seg_idx]->persistent_gnt = persistent_gnt;
 792                         pr_debug(DRV_PFX " grant %u added to the tree of persistent grants, using %u/%u\n",
 793                                  persistent_gnt->gnt, blkif->persistent_gnt_c,
 794                                  xen_blkif_max_pgrants);
 795                         goto next;
 796                 }
 797                 if (use_persistent_gnts && !blkif->vbd.overflow_max_grants) {
 798                         blkif->vbd.overflow_max_grants = 1;
 799                         pr_debug(DRV_PFX " domain %u, device %#x is using maximum number of persistent grants\n",
 800                                  blkif->domid, blkif->vbd.handle);
 801                 }
 802                 /*
 803                  * We could not map this grant persistently, so use it as
 804                  * a non-persistent grant.
 805                  */
 806 next:
 807                 new_map_idx++;
 808         }
 809         segs_to_map = 0;
 810         last_map = map_until;
 811         if (map_until != num)
 812                 goto again;
 813
 814         return ret;
 815
 816 out_of_memory:
 817         pr_alert(DRV_PFX "%s: out of memory\n", __func__);
 818         put_free_pages(blkif, pages_to_gnt, segs_to_map);
 819         return -ENOMEM;
 820 }
 821
 822 static int xen_blkbk_map_seg(struct pending_req *pending_req)
 823 {
 824         int rc;
 825
 826         rc = xen_blkbk_map(pending_req->blkif, pending_req->segments,
 827                            pending_req->nr_pages,
 828                            (pending_req->operation != BLKIF_OP_READ));
 829
 830         return rc;
 831 }
 832
 833 static int xen_blkbk_parse_indirect(struct blkif_request *req,
 834                                     struct pending_req *pending_req,
 835                                     struct seg_buf seg[],
 836                                     struct phys_req *preq)
 837 {
 838         struct grant_page **pages = pending_req->indirect_pages;
 839         struct xen_blkif *blkif = pending_req->blkif;
 840         int indirect_grefs, rc, n, nseg, i;
 841         struct blkif_request_segment_aligned *segments = NULL;
 842
 843         nseg = pending_req->nr_pages;
 844         indirect_grefs = INDIRECT_PAGES(nseg);
 845         BUG_ON(indirect_grefs > BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST);
 846
 847         for (i = 0; i < indirect_grefs; i++)
 848                 pages[i]->gref = req->u.indirect.indirect_grefs[i];
 849
 850         rc = xen_blkbk_map(blkif, pages, indirect_grefs, true);
 851         if (rc)
 852                 goto unmap;
 853
 854         for (n = 0, i = 0; n < nseg; n++) {
 855                 if ((n % SEGS_PER_INDIRECT_FRAME) == 0) {
 856                         /* Map indirect segments */
 857                         if (segments)
 858                                 kunmap_atomic(segments);
 859                         segments = kmap_atomic(pages[n/SEGS_PER_INDIRECT_FRAME]->page);
 860                 }
 861                 i = n % SEGS_PER_INDIRECT_FRAME;
 862                 pending_req->segments[n]->gref = segments[i].gref;
 863                 seg[n].nsec = segments[i].last_sect -
 864                         segments[i].first_sect + 1;
 865                 seg[n].offset = (segments[i].first_sect << 9);
 866                 if ((segments[i].last_sect >= (PAGE_SIZE >> 9)) ||
 867                     (segments[i].last_sect < segments[i].first_sect)) {
 868                         rc = -EINVAL;
 869                         goto unmap;
 870                 }
 871                 preq->nr_sects += seg[n].nsec;
 872         }
 873
 874 unmap:
 875         if (segments)
 876                 kunmap_atomic(segments);
 877         xen_blkbk_unmap(blkif, pages, indirect_grefs);
 878         return rc;
 879 }
 880
 881 static int dispatch_discard_io(struct xen_blkif *blkif,
 882                                 struct blkif_request *req)
 883 {
 884         int err = 0;
 885         int status = BLKIF_RSP_OKAY;
 886         struct block_device *bdev = blkif->vbd.bdev;
 887         unsigned long secure;
 888         struct phys_req preq;
 889
 890         xen_blkif_get(blkif);
 891
 892         preq.sector_number = req->u.discard.sector_number;
 893         preq.nr_sects      = req->u.discard.nr_sectors;
 894
 895         err = xen_vbd_translate(&preq, blkif, WRITE);
 896         if (err) {
 897                 pr_warn(DRV_PFX "access denied: DISCARD [%llu->%llu] on dev=%04x\n",
 898                         preq.sector_number,
 899                         preq.sector_number + preq.nr_sects, blkif->vbd.pdevice);
 900                 goto fail_response;
 901         }
 902         blkif->st_ds_req++;
 903
 904         secure = (blkif->vbd.discard_secure &&
 905                  (req->u.discard.flag & BLKIF_DISCARD_SECURE)) ?
 906                  BLKDEV_DISCARD_SECURE : 0;
 907
 908         err = blkdev_issue_discard(bdev, req->u.discard.sector_number,
 909                                    req->u.discard.nr_sectors,
 910                                    GFP_KERNEL, secure);
 911 fail_response:
 912         if (err == -EOPNOTSUPP) {
 913                 pr_debug(DRV_PFX "discard op failed, not supported\n");
 914                 status = BLKIF_RSP_EOPNOTSUPP;
 915         } else if (err)
 916                 status = BLKIF_RSP_ERROR;
 917
 918         make_response(blkif, req->u.discard.id, req->operation, status);
 919         xen_blkif_put(blkif);
 920         return err;
 921 }
 922
 923 static int dispatch_other_io(struct xen_blkif *blkif,
 924                              struct blkif_request *req,
 925                              struct pending_req *pending_req)
 926 {
 927         free_req(blkif, pending_req);
 928         make_response(blkif, req->u.other.id, req->operation,
 929                       BLKIF_RSP_EOPNOTSUPP);
 930         return -EIO;
 931 }
 932
 933 static void xen_blk_drain_io(struct xen_blkif *blkif)
 934 {
 935         atomic_set(&blkif->drain, 1);
 936         do {
 937                 /* The initial value is one, and one refcnt taken at the
 938                  * start of the xen_blkif_schedule thread. */
 939                 if (atomic_read(&blkif->refcnt) <= 2)
 940                         break;
 941                 wait_for_completion_interruptible_timeout(
 942                                 &blkif->drain_complete, HZ);
 943
 944                 if (!atomic_read(&blkif->drain))
 945                         break;
 946         } while (!kthread_should_stop());
 947         atomic_set(&blkif->drain, 0);
 948 }
 949
 950 /*
 951  * Completion callback on the bio's. Called as bh->b_end_io()
 952  */
 953
 954 static void __end_block_io_op(struct pending_req *pending_req, int error)
 955 {
 956         /* An error fails the entire request. */
 957         if ((pending_req->operation == BLKIF_OP_FLUSH_DISKCACHE) &&
 958             (error == -EOPNOTSUPP)) {
 959                 pr_debug(DRV_PFX "flush diskcache op failed, not supported\n");
 960                 xen_blkbk_flush_diskcache(XBT_NIL, pending_req->blkif->be, 0);
 961                 pending_req->status = BLKIF_RSP_EOPNOTSUPP;
 962         } else if ((pending_req->operation == BLKIF_OP_WRITE_BARRIER) &&
 963                     (error == -EOPNOTSUPP)) {
 964                 pr_debug(DRV_PFX "write barrier op failed, not supported\n");
 965                 xen_blkbk_barrier(XBT_NIL, pending_req->blkif->be, 0);
 966                 pending_req->status = BLKIF_RSP_EOPNOTSUPP;
 967         } else if (error) {
 968                 pr_debug(DRV_PFX "Buffer not up-to-date at end of operation,"
 969                          " error=%d\n", error);
 970                 pending_req->status = BLKIF_RSP_ERROR;
 971         }
 972
 973         /*
 974          * If all of the bio's have completed it is time to unmap
 975          * the grant references associated with 'request' and provide
 976          * the proper response on the ring.
 977          */
 978         if (atomic_dec_and_test(&pending_req->pendcnt)) {
 979                 xen_blkbk_unmap(pending_req->blkif,
 980                                 pending_req->segments,
 981                                 pending_req->nr_pages);
 982                 make_response(pending_req->blkif, pending_req->id,
 983                               pending_req->operation, pending_req->status);
 984                 xen_blkif_put(pending_req->blkif);
 985                 if (atomic_read(&pending_req->blkif->refcnt) <= 2) {
 986                         if (atomic_read(&pending_req->blkif->drain))
 987                                 complete(&pending_req->blkif->drain_complete);
 988                 }
 989                 free_req(pending_req->blkif, pending_req);
 990         }
 991 }
 992
 993 /*
 994  * bio callback.
 995  */
 996 static void end_block_io_op(struct bio *bio, int error)
 997 {
 998         __end_block_io_op(bio->bi_private, error);
 999         bio_put(bio);
1000 }
1001
1002
1003
1004 /*
1005  * Function to copy the from the ring buffer the 'struct blkif_request'
1006  * (which has the sectors we want, number of them, grant references, etc),
1007  * and transmute  it to the block API to hand it over to the proper block disk.
1008  */
1009 static int
1010 __do_block_io_op(struct xen_blkif *blkif)
1011 {
1012         union blkif_back_rings *blk_rings = &blkif->blk_rings;
1013         struct blkif_request req;
1014         struct pending_req *pending_req;
1015         RING_IDX rc, rp;
1016         int more_to_do = 0;
1017
1018         rc = blk_rings->common.req_cons;
1019         rp = blk_rings->common.sring->req_prod;
1020         rmb(); /* Ensure we see queued requests up to 'rp'. */
1021
1022         if (RING_REQUEST_PROD_OVERFLOW(&blk_rings->common, rp)) {
1023                 rc = blk_rings->common.rsp_prod_pvt;
1024                 pr_warn(DRV_PFX "Frontend provided bogus ring requests (%d - %d = %d). Halting ring processing on dev=%04x\n",
1025                         rp, rc, rp - rc, blkif->vbd.pdevice);
1026                 return -EACCES;
1027         }
1028         while (rc != rp) {
1029
1030                 if (RING_REQUEST_CONS_OVERFLOW(&blk_rings->common, rc))
1031                         break;
1032
1033                 if (kthread_should_stop()) {
1034                         more_to_do = 1;
1035                         break;
1036                 }
1037
1038                 pending_req = alloc_req(blkif);
1039                 if (NULL == pending_req) {
1040                         blkif->st_oo_req++;
1041                         more_to_do = 1;
1042                         break;
1043                 }
1044
1045                 switch (blkif->blk_protocol) {
1046                 case BLKIF_PROTOCOL_NATIVE:
1047                         memcpy(&req, RING_GET_REQUEST(&blk_rings->native, rc), sizeof(req));
1048                         break;
1049                 case BLKIF_PROTOCOL_X86_32:
1050                         blkif_get_x86_32_req(&req, RING_GET_REQUEST(&blk_rings->x86_32, rc));
1051                         break;
1052                 case BLKIF_PROTOCOL_X86_64:
1053                         blkif_get_x86_64_req(&req, RING_GET_REQUEST(&blk_rings->x86_64, rc));
1054                         break;
1055                 default:
1056                         BUG();
1057                 }
1058                 blk_rings->common.req_cons = ++rc; /* before make_response() */
1059
1060                 /* Apply all sanity checks to /private copy/ of request. */
1061                 barrier();
1062
1063                 switch (req.operation) {
1064                 case BLKIF_OP_READ:
1065                 case BLKIF_OP_WRITE:
1066                 case BLKIF_OP_WRITE_BARRIER:
1067                 case BLKIF_OP_FLUSH_DISKCACHE:
1068                 case BLKIF_OP_INDIRECT:
1069                         if (dispatch_rw_block_io(blkif, &req, pending_req))
1070                                 goto done;
1071                         break;
1072                 case BLKIF_OP_DISCARD:
1073                         free_req(blkif, pending_req);
1074                         if (dispatch_discard_io(blkif, &req))
1075                                 goto done;
1076                         break;
1077                 default:
1078                         if (dispatch_other_io(blkif, &req, pending_req))
1079                                 goto done;
1080                         break;
1081                 }
1082
1083                 /* Yield point for this unbounded loop. */
1084                 cond_resched();
1085         }
1086 done:
1087         return more_to_do;
1088 }
1089
1090 static int
1091 do_block_io_op(struct xen_blkif *blkif)
1092 {
1093         union blkif_back_rings *blk_rings = &blkif->blk_rings;
1094         int more_to_do;
1095
1096         do {
1097                 more_to_do = __do_block_io_op(blkif);
1098                 if (more_to_do)
1099                         break;
1100
1101                 RING_FINAL_CHECK_FOR_REQUESTS(&blk_rings->common, more_to_do);
1102         } while (more_to_do);
1103
1104         return more_to_do;
1105 }
1106 /*
1107  * Transmutation of the 'struct blkif_request' to a proper 'struct bio'
1108  * and call the 'submit_bio' to pass it to the underlying storage.
1109  */
1110 static int dispatch_rw_block_io(struct xen_blkif *blkif,
1111                                 struct blkif_request *req,
1112                                 struct pending_req *pending_req)
1113 {
1114         struct phys_req preq;
1115         struct seg_buf *seg = pending_req->seg;
1116         unsigned int nseg;
1117         struct bio *bio = NULL;
1118         struct bio **biolist = pending_req->biolist;
1119         int i, nbio = 0;
1120         int operation;
1121         struct blk_plug plug;
1122         bool drain = false;
1123         struct grant_page **pages = pending_req->segments;
1124         unsigned short req_operation;
1125
1126         req_operation = req->operation == BLKIF_OP_INDIRECT ?
1127                         req->u.indirect.indirect_op : req->operation;
1128         if ((req->operation == BLKIF_OP_INDIRECT) &&
1129             (req_operation != BLKIF_OP_READ) &&
1130             (req_operation != BLKIF_OP_WRITE)) {
1131                 pr_debug(DRV_PFX "Invalid indirect operation (%u)\n",
1132                          req_operation);
1133                 goto fail_response;
1134         }
1135
1136         switch (req_operation) {
1137         case BLKIF_OP_READ:
1138                 blkif->st_rd_req++;
1139                 operation = READ;
1140                 break;
1141         case BLKIF_OP_WRITE:
1142                 blkif->st_wr_req++;
1143                 operation = WRITE_ODIRECT;
1144                 break;
1145         case BLKIF_OP_WRITE_BARRIER:
1146                 drain = true;
1147         case BLKIF_OP_FLUSH_DISKCACHE:
1148                 blkif->st_f_req++;
1149                 operation = WRITE_FLUSH;
1150                 break;
1151         default:
1152                 operation = 0; /* make gcc happy */
1153                 goto fail_response;
1154                 break;
1155         }
1156
1157         /* Check that the number of segments is sane. */
1158         nseg = req->operation == BLKIF_OP_INDIRECT ?
1159                req->u.indirect.nr_segments : req->u.rw.nr_segments;
1160
1161         if (unlikely(nseg == 0 && operation != WRITE_FLUSH) ||
1162             unlikely((req->operation != BLKIF_OP_INDIRECT) &&
1163                      (nseg > BLKIF_MAX_SEGMENTS_PER_REQUEST)) ||
1164             unlikely((req->operation == BLKIF_OP_INDIRECT) &&
1165                      (nseg > MAX_INDIRECT_SEGMENTS))) {
1166                 pr_debug(DRV_PFX "Bad number of segments in request (%d)\n",
1167                          nseg);
1168                 /* Haven't submitted any bio's yet. */
1169                 goto fail_response;
1170         }
1171
1172         preq.nr_sects      = 0;
1173
1174         pending_req->blkif     = blkif;
1175         pending_req->id        = req->u.rw.id;
1176         pending_req->operation = req_operation;
1177         pending_req->status    = BLKIF_RSP_OKAY;
1178         pending_req->nr_pages  = nseg;
1179
1180         if (req->operation != BLKIF_OP_INDIRECT) {
1181                 preq.dev               = req->u.rw.handle;
1182                 preq.sector_number     = req->u.rw.sector_number;
1183                 for (i = 0; i < nseg; i++) {
1184                         pages[i]->gref = req->u.rw.seg[i].gref;
1185                         seg[i].nsec = req->u.rw.seg[i].last_sect -
1186                                 req->u.rw.seg[i].first_sect + 1;
1187                         seg[i].offset = (req->u.rw.seg[i].first_sect << 9);
1188                         if ((req->u.rw.seg[i].last_sect >= (PAGE_SIZE >> 9)) ||
1189                             (req->u.rw.seg[i].last_sect <
1190                              req->u.rw.seg[i].first_sect))
1191                                 goto fail_response;
1192                         preq.nr_sects += seg[i].nsec;
1193                 }
1194         } else {
1195                 preq.dev               = req->u.indirect.handle;
1196                 preq.sector_number     = req->u.indirect.sector_number;
1197                 if (xen_blkbk_parse_indirect(req, pending_req, seg, &preq))
1198                         goto fail_response;
1199         }
1200
1201         if (xen_vbd_translate(&preq, blkif, operation) != 0) {
1202                 pr_debug(DRV_PFX "access denied: %s of [%llu,%llu] on dev=%04x\n",
1203                          operation == READ ? "read" : "write",
1204                          preq.sector_number,
1205                          preq.sector_number + preq.nr_sects,
1206                          blkif->vbd.pdevice);
1207                 goto fail_response;
1208         }
1209
1210         /*
1211          * This check _MUST_ be done after xen_vbd_translate as the preq.bdev
1212          * is set there.
1213          */
1214         for (i = 0; i < nseg; i++) {
1215                 if (((int)preq.sector_number|(int)seg[i].nsec) &
1216                     ((bdev_logical_block_size(preq.bdev) >> 9) - 1)) {
1217                         pr_debug(DRV_PFX "Misaligned I/O request from domain %d",
1218                                  blkif->domid);
1219                         goto fail_response;
1220                 }
1221         }
1222
1223         /* Wait on all outstanding I/O's and once that has been completed
1224          * issue the WRITE_FLUSH.
1225          */
1226         if (drain)
1227                 xen_blk_drain_io(pending_req->blkif);
1228
1229         /*
1230          * If we have failed at this point, we need to undo the M2P override,
1231          * set gnttab_set_unmap_op on all of the grant references and perform
1232          * the hypercall to unmap the grants - that is all done in
1233          * xen_blkbk_unmap.
1234          */
1235         if (xen_blkbk_map_seg(pending_req))
1236                 goto fail_flush;
1237
1238         /*
1239          * This corresponding xen_blkif_put is done in __end_block_io_op, or
1240          * below (in "!bio") if we are handling a BLKIF_OP_DISCARD.
1241          */
1242         xen_blkif_get(blkif);
1243
1244         for (i = 0; i < nseg; i++) {
1245                 while ((bio == NULL) ||
1246                        (bio_add_page(bio,
1247                                      pages[i]->page,
1248                                      seg[i].nsec << 9,
1249                                      seg[i].offset) == 0)) {
1250
1251                         int nr_iovecs = min_t(int, (nseg-i), BIO_MAX_PAGES);
1252                         bio = bio_alloc(GFP_KERNEL, nr_iovecs);
1253                         if (unlikely(bio == NULL))
1254                                 goto fail_put_bio;
1255
1256                         biolist[nbio++] = bio;
1257                         bio->bi_bdev    = preq.bdev;
1258                         bio->bi_private = pending_req;
1259                         bio->bi_end_io  = end_block_io_op;
1260                         bio->bi_iter.bi_sector  = preq.sector_number;
1261                 }
1262
1263                 preq.sector_number += seg[i].nsec;
1264         }
1265
1266         /* This will be hit if the operation was a flush or discard. */
1267         if (!bio) {
1268                 BUG_ON(operation != WRITE_FLUSH);
1269
1270                 bio = bio_alloc(GFP_KERNEL, 0);
1271                 if (unlikely(bio == NULL))
1272                         goto fail_put_bio;
1273
1274                 biolist[nbio++] = bio;
1275                 bio->bi_bdev    = preq.bdev;
1276                 bio->bi_private = pending_req;
1277                 bio->bi_end_io  = end_block_io_op;
1278         }
1279
1280         atomic_set(&pending_req->pendcnt, nbio);
1281         blk_start_plug(&plug);
1282
1283         for (i = 0; i < nbio; i++)
1284                 submit_bio(operation, biolist[i]);
1285
1286         /* Let the I/Os go.. */
1287         blk_finish_plug(&plug);
1288
1289         if (operation == READ)
1290                 blkif->st_rd_sect += preq.nr_sects;
1291         else if (operation & WRITE)
1292                 blkif->st_wr_sect += preq.nr_sects;
1293
1294         return 0;
1295
1296  fail_flush:
1297         xen_blkbk_unmap(blkif, pending_req->segments,
1298                         pending_req->nr_pages);
1299  fail_response:
1300         /* Haven't submitted any bio's yet. */
1301         make_response(blkif, req->u.rw.id, req_operation, BLKIF_RSP_ERROR);
1302         free_req(blkif, pending_req);
1303         msleep(1); /* back off a bit */
1304         return -EIO;
1305
1306  fail_put_bio:
1307         for (i = 0; i < nbio; i++)
1308                 bio_put(biolist[i]);
1309         atomic_set(&pending_req->pendcnt, 1);
1310         __end_block_io_op(pending_req, -EINVAL);
1311         msleep(1); /* back off a bit */
1312         return -EIO;
1313 }
1314
1315
1316
1317 /*
1318  * Put a response on the ring on how the operation fared.
1319  */
1320 static void make_response(struct xen_blkif *blkif, u64 id,
1321                           unsigned short op, int st)
1322 {
1323         struct blkif_response  resp;
1324         unsigned long     flags;
1325         union blkif_back_rings *blk_rings = &blkif->blk_rings;
1326         int notify;
1327
1328         resp.id        = id;
1329         resp.operation = op;
1330         resp.status    = st;
1331
1332         spin_lock_irqsave(&blkif->blk_ring_lock, flags);
1333         /* Place on the response ring for the relevant domain. */
1334         switch (blkif->blk_protocol) {
1335         case BLKIF_PROTOCOL_NATIVE:
1336                 memcpy(RING_GET_RESPONSE(&blk_rings->native, blk_rings->native.rsp_prod_pvt),
1337                        &resp, sizeof(resp));
1338                 break;
1339         case BLKIF_PROTOCOL_X86_32:
1340                 memcpy(RING_GET_RESPONSE(&blk_rings->x86_32, blk_rings->x86_32.rsp_prod_pvt),
1341                        &resp, sizeof(resp));
1342                 break;
1343         case BLKIF_PROTOCOL_X86_64:
1344                 memcpy(RING_GET_RESPONSE(&blk_rings->x86_64, blk_rings->x86_64.rsp_prod_pvt),
1345                        &resp, sizeof(resp));
1346                 break;
1347         default:
1348                 BUG();
1349         }
1350         blk_rings->common.rsp_prod_pvt++;
1351         RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&blk_rings->common, notify);
1352         spin_unlock_irqrestore(&blkif->blk_ring_lock, flags);
1353         if (notify)
1354                 notify_remote_via_irq(blkif->irq);
1355 }
1356
1357 static int __init xen_blkif_init(void)
1358 {
1359         int rc = 0;
1360
1361         if (!xen_domain())
1362                 return -ENODEV;
1363
1364         rc = xen_blkif_interface_init();
1365         if (rc)
1366                 goto failed_init;
1367
1368         rc = xen_blkif_xenbus_init();
1369         if (rc)
1370                 goto failed_init;
1371
1372  failed_init:
1373         return rc;
1374 }
1375
1376 module_init(xen_blkif_init);
1377
1378 MODULE_LICENSE("Dual BSD/GPL");
1379 MODULE_ALIAS("xen-backend:vbd");