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 #define pr_fmt(fmt) "xen-blkback: " fmt
  38
  39 #include <linux/spinlock.h>
  40 #include <linux/kthread.h>
  41 #include <linux/list.h>
  42 #include <linux/delay.h>
  43 #include <linux/freezer.h>
  44 #include <linux/bitmap.h>
  45
  46 #include <xen/events.h>
  47 #include <xen/page.h>
  48 #include <xen/xen.h>
  49 #include <asm/xen/hypervisor.h>
  50 #include <asm/xen/hypercall.h>
  51 #include <xen/balloon.h>
  52 #include <xen/grant_table.h>
  53 #include "common.h"
  54
  55 /*
  56  * Maximum number of unused free pages to keep in the internal buffer.
  57  * Setting this to a value too low will reduce memory used in each backend,
  58  * but can have a performance penalty.
  59  *
  60  * A sane value is xen_blkif_reqs * BLKIF_MAX_SEGMENTS_PER_REQUEST, but can
  61  * be set to a lower value that might degrade performance on some intensive
  62  * IO workloads.
  63  */
  64
  65 static int xen_blkif_max_buffer_pages = 1024;
  66 module_param_named(max_buffer_pages, xen_blkif_max_buffer_pages, int, 0644);
  67 MODULE_PARM_DESC(max_buffer_pages,
  68 "Maximum number of free pages to keep in each block backend buffer");
  69
  70 /*
  71  * Maximum number of grants to map persistently in blkback. For maximum
  72  * performance this should be the total numbers of grants that can be used
  73  * to fill the ring, but since this might become too high, specially with
  74  * the use of indirect descriptors, we set it to a value that provides good
  75  * performance without using too much memory.
  76  *
  77  * When the list of persistent grants is full we clean it up using a LRU
  78  * algorithm.
  79  */
  80
  81 static int xen_blkif_max_pgrants = 1056;
  82 module_param_named(max_persistent_grants, xen_blkif_max_pgrants, int, 0644);
  83 MODULE_PARM_DESC(max_persistent_grants,
  84                  "Maximum number of grants to map persistently");
  85
  86 /*
  87  * Maximum number of rings/queues blkback supports, allow as many queues as there
  88  * are CPUs if user has not specified a value.
  89  */
  90 unsigned int xenblk_max_queues;
  91 module_param_named(max_queues, xenblk_max_queues, uint, 0644);
  92 MODULE_PARM_DESC(max_queues,
  93                  "Maximum number of hardware queues per virtual disk." \
  94                  "By default it is the number of online CPUs.");
  95
  96 /*
  97  * Maximum order of pages to be used for the shared ring between front and
  98  * backend, 4KB page granularity is used.
  99  */
 100 unsigned int xen_blkif_max_ring_order = XENBUS_MAX_RING_GRANT_ORDER;
 101 module_param_named(max_ring_page_order, xen_blkif_max_ring_order, int, S_IRUGO);
 102 MODULE_PARM_DESC(max_ring_page_order, "Maximum order of pages to be used for the shared ring");
 103 /*
 104  * The LRU mechanism to clean the lists of persistent grants needs to
 105  * be executed periodically. The time interval between consecutive executions
 106  * of the purge mechanism is set in ms.
 107  */
 108 #define LRU_INTERVAL 100
 109
 110 /*
 111  * When the persistent grants list is full we will remove unused grants
 112  * from the list. The percent number of grants to be removed at each LRU
 113  * execution.
 114  */
 115 #define LRU_PERCENT_CLEAN 5
 116
 117 /* Run-time switchable: /sys/module/blkback/parameters/ */
 118 static unsigned int log_stats;
 119 module_param(log_stats, int, 0644);
 120
 121 #define BLKBACK_INVALID_HANDLE (~0)
 122
 123 /* Number of free pages to remove on each call to gnttab_free_pages */
 124 #define NUM_BATCH_FREE_PAGES 10
 125
 126 static inline int get_free_page(struct xen_blkif_ring *ring, struct page **page)
 127 {
 128         unsigned long flags;
 129
 130         spin_lock_irqsave(&ring->free_pages_lock, flags);
 131         if (list_empty(&ring->free_pages)) {
 132                 BUG_ON(ring->free_pages_num != 0);
 133                 spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 134                 return gnttab_alloc_pages(1, page);
 135         }
 136         BUG_ON(ring->free_pages_num == 0);
 137         page[0] = list_first_entry(&ring->free_pages, struct page, lru);
 138         list_del(&page[0]->lru);
 139         ring->free_pages_num--;
 140         spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 141
 142         return 0;
 143 }
 144
 145 static inline void put_free_pages(struct xen_blkif_ring *ring, struct page **page,
 146                                   int num)
 147 {
 148         unsigned long flags;
 149         int i;
 150
 151         spin_lock_irqsave(&ring->free_pages_lock, flags);
 152         for (i = 0; i < num; i++)
 153                 list_add(&page[i]->lru, &ring->free_pages);
 154         ring->free_pages_num += num;
 155         spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 156 }
 157
 158 static inline void shrink_free_pagepool(struct xen_blkif_ring *ring, int num)
 159 {
 160         /* Remove requested pages in batches of NUM_BATCH_FREE_PAGES */
 161         struct page *page[NUM_BATCH_FREE_PAGES];
 162         unsigned int num_pages = 0;
 163         unsigned long flags;
 164
 165         spin_lock_irqsave(&ring->free_pages_lock, flags);
 166         while (ring->free_pages_num > num) {
 167                 BUG_ON(list_empty(&ring->free_pages));
 168                 page[num_pages] = list_first_entry(&ring->free_pages,
 169                                                    struct page, lru);
 170                 list_del(&page[num_pages]->lru);
 171                 ring->free_pages_num--;
 172                 if (++num_pages == NUM_BATCH_FREE_PAGES) {
 173                         spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 174                         gnttab_free_pages(num_pages, page);
 175                         spin_lock_irqsave(&ring->free_pages_lock, flags);
 176                         num_pages = 0;
 177                 }
 178         }
 179         spin_unlock_irqrestore(&ring->free_pages_lock, flags);
 180         if (num_pages != 0)
 181                 gnttab_free_pages(num_pages, page);
 182 }
 183
 184 #define vaddr(page) ((unsigned long)pfn_to_kaddr(page_to_pfn(page)))
 185
 186 static int do_block_io_op(struct xen_blkif_ring *ring);
 187 static int dispatch_rw_block_io(struct xen_blkif_ring *ring,
 188                                 struct blkif_request *req,
 189                                 struct pending_req *pending_req);
 190 static void make_response(struct xen_blkif_ring *ring, u64 id,
 191                           unsigned short op, int st);
 192
 193 #define foreach_grant_safe(pos, n, rbtree, node) \
 194         for ((pos) = container_of(rb_first((rbtree)), typeof(*(pos)), node), \
 195              (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL; \
 196              &(pos)->node != NULL; \
 197              (pos) = container_of(n, typeof(*(pos)), node), \
 198              (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL)
 199
 200
 201 /*
 202  * We don't need locking around the persistent grant helpers
 203  * because blkback uses a single-thread for each backend, so we
 204  * can be sure that this functions will never be called recursively.
 205  *
 206  * The only exception to that is put_persistent_grant, that can be called
 207  * from interrupt context (by xen_blkbk_unmap), so we have to use atomic
 208  * bit operations to modify the flags of a persistent grant and to count
 209  * the number of used grants.
 210  */
 211 static int add_persistent_gnt(struct xen_blkif_ring *ring,
 212                                struct persistent_gnt *persistent_gnt)
 213 {
 214         struct rb_node **new = NULL, *parent = NULL;
 215         struct persistent_gnt *this;
 216         struct xen_blkif *blkif = ring->blkif;
 217
 218         if (ring->persistent_gnt_c >= xen_blkif_max_pgrants) {
 219                 if (!blkif->vbd.overflow_max_grants)
 220                         blkif->vbd.overflow_max_grants = 1;
 221                 return -EBUSY;
 222         }
 223         /* Figure out where to put new node */
 224         new = &ring->persistent_gnts.rb_node;
 225         while (*new) {
 226                 this = container_of(*new, struct persistent_gnt, node);
 227
 228                 parent = *new;
 229                 if (persistent_gnt->gnt < this->gnt)
 230                         new = &((*new)->rb_left);
 231                 else if (persistent_gnt->gnt > this->gnt)
 232                         new = &((*new)->rb_right);
 233                 else {
 234                         pr_alert_ratelimited("trying to add a gref that's already in the tree\n");
 235                         return -EINVAL;
 236                 }
 237         }
 238
 239         bitmap_zero(persistent_gnt->flags, PERSISTENT_GNT_FLAGS_SIZE);
 240         set_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags);
 241         /* Add new node and rebalance tree. */
 242         rb_link_node(&(persistent_gnt->node), parent, new);
 243         rb_insert_color(&(persistent_gnt->node), &ring->persistent_gnts);
 244         ring->persistent_gnt_c++;
 245         atomic_inc(&ring->persistent_gnt_in_use);
 246         return 0;
 247 }
 248
 249 static struct persistent_gnt *get_persistent_gnt(struct xen_blkif_ring *ring,
 250                                                  grant_ref_t gref)
 251 {
 252         struct persistent_gnt *data;
 253         struct rb_node *node = NULL;
 254
 255         node = ring->persistent_gnts.rb_node;
 256         while (node) {
 257                 data = container_of(node, struct persistent_gnt, node);
 258
 259                 if (gref < data->gnt)
 260                         node = node->rb_left;
 261                 else if (gref > data->gnt)
 262                         node = node->rb_right;
 263                 else {
 264                         if(test_bit(PERSISTENT_GNT_ACTIVE, data->flags)) {
 265                                 pr_alert_ratelimited("requesting a grant already in use\n");
 266                                 return NULL;
 267                         }
 268                         set_bit(PERSISTENT_GNT_ACTIVE, data->flags);
 269                         atomic_inc(&ring->persistent_gnt_in_use);
 270                         return data;
 271                 }
 272         }
 273         return NULL;
 274 }
 275
 276 static void put_persistent_gnt(struct xen_blkif_ring *ring,
 277                                struct persistent_gnt *persistent_gnt)
 278 {
 279         if(!test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags))
 280                 pr_alert_ratelimited("freeing a grant already unused\n");
 281         set_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags);
 282         clear_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags);
 283         atomic_dec(&ring->persistent_gnt_in_use);
 284 }
 285
 286 static void free_persistent_gnts(struct xen_blkif_ring *ring, struct rb_root *root,
 287                                  unsigned int num)
 288 {
 289         struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 290         struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 291         struct persistent_gnt *persistent_gnt;
 292         struct rb_node *n;
 293         int segs_to_unmap = 0;
 294         struct gntab_unmap_queue_data unmap_data;
 295
 296         unmap_data.pages = pages;
 297         unmap_data.unmap_ops = unmap;
 298         unmap_data.kunmap_ops = NULL;
 299
 300         foreach_grant_safe(persistent_gnt, n, root, node) {
 301                 BUG_ON(persistent_gnt->handle ==
 302                         BLKBACK_INVALID_HANDLE);
 303                 gnttab_set_unmap_op(&unmap[segs_to_unmap],
 304                         (unsigned long) pfn_to_kaddr(page_to_pfn(
 305                                 persistent_gnt->page)),
 306                         GNTMAP_host_map,
 307                         persistent_gnt->handle);
 308
 309                 pages[segs_to_unmap] = persistent_gnt->page;
 310
 311                 if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST ||
 312                         !rb_next(&persistent_gnt->node)) {
 313
 314                         unmap_data.count = segs_to_unmap;
 315                         BUG_ON(gnttab_unmap_refs_sync(&unmap_data));
 316
 317                         put_free_pages(ring, pages, segs_to_unmap);
 318                         segs_to_unmap = 0;
 319                 }
 320
 321                 rb_erase(&persistent_gnt->node, root);
 322                 kfree(persistent_gnt);
 323                 num--;
 324         }
 325         BUG_ON(num != 0);
 326 }
 327
 328 void xen_blkbk_unmap_purged_grants(struct work_struct *work)
 329 {
 330         struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 331         struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 332         struct persistent_gnt *persistent_gnt;
 333         int segs_to_unmap = 0;
 334         struct xen_blkif_ring *ring = container_of(work, typeof(*ring), persistent_purge_work);
 335         struct gntab_unmap_queue_data unmap_data;
 336
 337         unmap_data.pages = pages;
 338         unmap_data.unmap_ops = unmap;
 339         unmap_data.kunmap_ops = NULL;
 340
 341         while(!list_empty(&ring->persistent_purge_list)) {
 342                 persistent_gnt = list_first_entry(&ring->persistent_purge_list,
 343                                                   struct persistent_gnt,
 344                                                   remove_node);
 345                 list_del(&persistent_gnt->remove_node);
 346
 347                 gnttab_set_unmap_op(&unmap[segs_to_unmap],
 348                         vaddr(persistent_gnt->page),
 349                         GNTMAP_host_map,
 350                         persistent_gnt->handle);
 351
 352                 pages[segs_to_unmap] = persistent_gnt->page;
 353
 354                 if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST) {
 355                         unmap_data.count = segs_to_unmap;
 356                         BUG_ON(gnttab_unmap_refs_sync(&unmap_data));
 357                         put_free_pages(ring, pages, segs_to_unmap);
 358                         segs_to_unmap = 0;
 359                 }
 360                 kfree(persistent_gnt);
 361         }
 362         if (segs_to_unmap > 0) {
 363                 unmap_data.count = segs_to_unmap;
 364                 BUG_ON(gnttab_unmap_refs_sync(&unmap_data));
 365                 put_free_pages(ring, pages, segs_to_unmap);
 366         }
 367 }
 368
 369 static void purge_persistent_gnt(struct xen_blkif_ring *ring)
 370 {
 371         struct persistent_gnt *persistent_gnt;
 372         struct rb_node *n;
 373         unsigned int num_clean, total;
 374         bool scan_used = false, clean_used = false;
 375         struct rb_root *root;
 376
 377         if (ring->persistent_gnt_c < xen_blkif_max_pgrants ||
 378             (ring->persistent_gnt_c == xen_blkif_max_pgrants &&
 379             !ring->blkif->vbd.overflow_max_grants)) {
 380                 goto out;
 381         }
 382
 383         if (work_busy(&ring->persistent_purge_work)) {
 384                 pr_alert_ratelimited("Scheduled work from previous purge is still busy, cannot purge list\n");
 385                 goto out;
 386         }
 387
 388         num_clean = (xen_blkif_max_pgrants / 100) * LRU_PERCENT_CLEAN;
 389         num_clean = ring->persistent_gnt_c - xen_blkif_max_pgrants + num_clean;
 390         num_clean = min(ring->persistent_gnt_c, num_clean);
 391         if ((num_clean == 0) ||
 392             (num_clean > (ring->persistent_gnt_c - atomic_read(&ring->persistent_gnt_in_use))))
 393                 goto out;
 394
 395         /*
 396          * At this point, we can assure that there will be no calls
 397          * to get_persistent_grant (because we are executing this code from
 398          * xen_blkif_schedule), there can only be calls to put_persistent_gnt,
 399          * which means that the number of currently used grants will go down,
 400          * but never up, so we will always be able to remove the requested
 401          * number of grants.
 402          */
 403
 404         total = num_clean;
 405
 406         pr_debug("Going to purge %u persistent grants\n", num_clean);
 407
 408         BUG_ON(!list_empty(&ring->persistent_purge_list));
 409         root = &ring->persistent_gnts;
 410 purge_list:
 411         foreach_grant_safe(persistent_gnt, n, root, node) {
 412                 BUG_ON(persistent_gnt->handle ==
 413                         BLKBACK_INVALID_HANDLE);
 414
 415                 if (clean_used) {
 416                         clear_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags);
 417                         continue;
 418                 }
 419
 420                 if (test_bit(PERSISTENT_GNT_ACTIVE, persistent_gnt->flags))
 421                         continue;
 422                 if (!scan_used &&
 423                     (test_bit(PERSISTENT_GNT_WAS_ACTIVE, persistent_gnt->flags)))
 424                         continue;
 425
 426                 rb_erase(&persistent_gnt->node, root);
 427                 list_add(&persistent_gnt->remove_node,
 428                          &ring->persistent_purge_list);
 429                 if (--num_clean == 0)
 430                         goto finished;
 431         }
 432         /*
 433          * If we get here it means we also need to start cleaning
 434          * grants that were used since last purge in order to cope
 435          * with the requested num
 436          */
 437         if (!scan_used && !clean_used) {
 438                 pr_debug("Still missing %u purged frames\n", num_clean);
 439                 scan_used = true;
 440                 goto purge_list;
 441         }
 442 finished:
 443         if (!clean_used) {
 444                 pr_debug("Finished scanning for grants to clean, removing used flag\n");
 445                 clean_used = true;
 446                 goto purge_list;
 447         }
 448
 449         ring->persistent_gnt_c -= (total - num_clean);
 450         ring->blkif->vbd.overflow_max_grants = 0;
 451
 452         /* We can defer this work */
 453         schedule_work(&ring->persistent_purge_work);
 454         pr_debug("Purged %u/%u\n", (total - num_clean), total);
 455
 456 out:
 457         return;
 458 }
 459
 460 /*
 461  * Retrieve from the 'pending_reqs' a free pending_req structure to be used.
 462  */
 463 static struct pending_req *alloc_req(struct xen_blkif_ring *ring)
 464 {
 465         struct pending_req *req = NULL;
 466         unsigned long flags;
 467
 468         spin_lock_irqsave(&ring->pending_free_lock, flags);
 469         if (!list_empty(&ring->pending_free)) {
 470                 req = list_entry(ring->pending_free.next, struct pending_req,
 471                                  free_list);
 472                 list_del(&req->free_list);
 473         }
 474         spin_unlock_irqrestore(&ring->pending_free_lock, flags);
 475         return req;
 476 }
 477
 478 /*
 479  * Return the 'pending_req' structure back to the freepool. We also
 480  * wake up the thread if it was waiting for a free page.
 481  */
 482 static void free_req(struct xen_blkif_ring *ring, struct pending_req *req)
 483 {
 484         unsigned long flags;
 485         int was_empty;
 486
 487         spin_lock_irqsave(&ring->pending_free_lock, flags);
 488         was_empty = list_empty(&ring->pending_free);
 489         list_add(&req->free_list, &ring->pending_free);
 490         spin_unlock_irqrestore(&ring->pending_free_lock, flags);
 491         if (was_empty)
 492                 wake_up(&ring->pending_free_wq);
 493 }
 494
 495 /*
 496  * Routines for managing virtual block devices (vbds).
 497  */
 498 static int xen_vbd_translate(struct phys_req *req, struct xen_blkif *blkif,
 499                              int operation)
 500 {
 501         struct xen_vbd *vbd = &blkif->vbd;
 502         int rc = -EACCES;
 503
 504         if ((operation != READ) && vbd->readonly)
 505                 goto out;
 506
 507         if (likely(req->nr_sects)) {
 508                 blkif_sector_t end = req->sector_number + req->nr_sects;
 509
 510                 if (unlikely(end < req->sector_number))
 511                         goto out;
 512                 if (unlikely(end > vbd_sz(vbd)))
 513                         goto out;
 514         }
 515
 516         req->dev  = vbd->pdevice;
 517         req->bdev = vbd->bdev;
 518         rc = 0;
 519
 520  out:
 521         return rc;
 522 }
 523
 524 static void xen_vbd_resize(struct xen_blkif *blkif)
 525 {
 526         struct xen_vbd *vbd = &blkif->vbd;
 527         struct xenbus_transaction xbt;
 528         int err;
 529         struct xenbus_device *dev = xen_blkbk_xenbus(blkif->be);
 530         unsigned long long new_size = vbd_sz(vbd);
 531
 532         pr_info("VBD Resize: Domid: %d, Device: (%d, %d)\n",
 533                 blkif->domid, MAJOR(vbd->pdevice), MINOR(vbd->pdevice));
 534         pr_info("VBD Resize: new size %llu\n", new_size);
 535         vbd->size = new_size;
 536 again:
 537         err = xenbus_transaction_start(&xbt);
 538         if (err) {
 539                 pr_warn("Error starting transaction\n");
 540                 return;
 541         }
 542         err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu",
 543                             (unsigned long long)vbd_sz(vbd));
 544         if (err) {
 545                 pr_warn("Error writing new size\n");
 546                 goto abort;
 547         }
 548         /*
 549          * Write the current state; we will use this to synchronize
 550          * the front-end. If the current state is "connected" the
 551          * front-end will get the new size information online.
 552          */
 553         err = xenbus_printf(xbt, dev->nodename, "state", "%d", dev->state);
 554         if (err) {
 555                 pr_warn("Error writing the state\n");
 556                 goto abort;
 557         }
 558
 559         err = xenbus_transaction_end(xbt, 0);
 560         if (err == -EAGAIN)
 561                 goto again;
 562         if (err)
 563                 pr_warn("Error ending transaction\n");
 564         return;
 565 abort:
 566         xenbus_transaction_end(xbt, 1);
 567 }
 568
 569 /*
 570  * Notification from the guest OS.
 571  */
 572 static void blkif_notify_work(struct xen_blkif_ring *ring)
 573 {
 574         ring->waiting_reqs = 1;
 575         wake_up(&ring->wq);
 576 }
 577
 578 irqreturn_t xen_blkif_be_int(int irq, void *dev_id)
 579 {
 580         blkif_notify_work(dev_id);
 581         return IRQ_HANDLED;
 582 }
 583
 584 /*
 585  * SCHEDULER FUNCTIONS
 586  */
 587
 588 static void print_stats(struct xen_blkif_ring *ring)
 589 {
 590         struct xen_blkif *blkif = ring->blkif;
 591
 592         pr_info("(%s): oo %3llu  |  rd %4llu  |  wr %4llu  |  f %4llu"
 593                  "  |  ds %4llu | pg: %4u/%4d\n",
 594                  current->comm, blkif->st_oo_req,
 595                  blkif->st_rd_req, blkif->st_wr_req,
 596                  blkif->st_f_req, blkif->st_ds_req,
 597                  ring->persistent_gnt_c,
 598                  xen_blkif_max_pgrants);
 599         blkif->st_print = jiffies + msecs_to_jiffies(10 * 1000);
 600         blkif->st_rd_req = 0;
 601         blkif->st_wr_req = 0;
 602         blkif->st_oo_req = 0;
 603         blkif->st_ds_req = 0;
 604 }
 605
 606 int xen_blkif_schedule(void *arg)
 607 {
 608         struct xen_blkif_ring *ring = arg;
 609         struct xen_blkif *blkif = ring->blkif;
 610         struct xen_vbd *vbd = &blkif->vbd;
 611         unsigned long timeout;
 612         int ret;
 613
 614         xen_blkif_get(blkif);
 615
 616         set_freezable();
 617         while (!kthread_should_stop()) {
 618                 if (try_to_freeze())
 619                         continue;
 620                 if (unlikely(vbd->size != vbd_sz(vbd)))
 621                         xen_vbd_resize(blkif);
 622
 623                 timeout = msecs_to_jiffies(LRU_INTERVAL);
 624
 625                 timeout = wait_event_interruptible_timeout(
 626                         ring->wq,
 627                         ring->waiting_reqs || kthread_should_stop(),
 628                         timeout);
 629                 if (timeout == 0)
 630                         goto purge_gnt_list;
 631                 timeout = wait_event_interruptible_timeout(
 632                         ring->pending_free_wq,
 633                         !list_empty(&ring->pending_free) ||
 634                         kthread_should_stop(),
 635                         timeout);
 636                 if (timeout == 0)
 637                         goto purge_gnt_list;
 638
 639                 ring->waiting_reqs = 0;
 640                 smp_mb(); /* clear flag *before* checking for work */
 641
 642                 ret = do_block_io_op(ring);
 643                 if (ret > 0)
 644                         ring->waiting_reqs = 1;
 645                 if (ret == -EACCES)
 646                         wait_event_interruptible(ring->shutdown_wq,
 647                                                  kthread_should_stop());
 648
 649 purge_gnt_list:
 650                 if (blkif->vbd.feature_gnt_persistent &&
 651                     time_after(jiffies, ring->next_lru)) {
 652                         purge_persistent_gnt(ring);
 653                         ring->next_lru = jiffies + msecs_to_jiffies(LRU_INTERVAL);
 654                 }
 655
 656                 /* Shrink if we have more than xen_blkif_max_buffer_pages */
 657                 shrink_free_pagepool(ring, xen_blkif_max_buffer_pages);
 658
 659                 if (log_stats && time_after(jiffies, ring->blkif->st_print))
 660                         print_stats(ring);
 661         }
 662
 663         /* Drain pending purge work */
 664         flush_work(&ring->persistent_purge_work);
 665
 666         if (log_stats)
 667                 print_stats(ring);
 668
 669         ring->xenblkd = NULL;
 670         xen_blkif_put(blkif);
 671
 672         return 0;
 673 }
 674
 675 /*
 676  * Remove persistent grants and empty the pool of free pages
 677  */
 678 void xen_blkbk_free_caches(struct xen_blkif_ring *ring)
 679 {
 680         /* Free all persistent grant pages */
 681         if (!RB_EMPTY_ROOT(&ring->persistent_gnts))
 682                 free_persistent_gnts(ring, &ring->persistent_gnts,
 683                         ring->persistent_gnt_c);
 684
 685         BUG_ON(!RB_EMPTY_ROOT(&ring->persistent_gnts));
 686         ring->persistent_gnt_c = 0;
 687
 688         /* Since we are shutting down remove all pages from the buffer */
 689         shrink_free_pagepool(ring, 0 /* All */);
 690 }
 691
 692 static unsigned int xen_blkbk_unmap_prepare(
 693         struct xen_blkif_ring *ring,
 694         struct grant_page **pages,
 695         unsigned int num,
 696         struct gnttab_unmap_grant_ref *unmap_ops,
 697         struct page **unmap_pages)
 698 {
 699         unsigned int i, invcount = 0;
 700
 701         for (i = 0; i < num; i++) {
 702                 if (pages[i]->persistent_gnt != NULL) {
 703                         put_persistent_gnt(ring, pages[i]->persistent_gnt);
 704                         continue;
 705                 }
 706                 if (pages[i]->handle == BLKBACK_INVALID_HANDLE)
 707                         continue;
 708                 unmap_pages[invcount] = pages[i]->page;
 709                 gnttab_set_unmap_op(&unmap_ops[invcount], vaddr(pages[i]->page),
 710                                     GNTMAP_host_map, pages[i]->handle);
 711                 pages[i]->handle = BLKBACK_INVALID_HANDLE;
 712                 invcount++;
 713        }
 714
 715        return invcount;
 716 }
 717
 718 static void xen_blkbk_unmap_and_respond_callback(int result, struct gntab_unmap_queue_data *data)
 719 {
 720         struct pending_req *pending_req = (struct pending_req *)(data->data);
 721         struct xen_blkif_ring *ring = pending_req->ring;
 722         struct xen_blkif *blkif = ring->blkif;
 723
 724         /* BUG_ON used to reproduce existing behaviour,
 725            but is this the best way to deal with this? */
 726         BUG_ON(result);
 727
 728         put_free_pages(ring, data->pages, data->count);
 729         make_response(ring, pending_req->id,
 730                       pending_req->operation, pending_req->status);
 731         free_req(ring, pending_req);
 732         /*
 733          * Make sure the request is freed before releasing blkif,
 734          * or there could be a race between free_req and the
 735          * cleanup done in xen_blkif_free during shutdown.
 736          *
 737          * NB: The fact that we might try to wake up pending_free_wq
 738          * before drain_complete (in case there's a drain going on)
 739          * it's not a problem with our current implementation
 740          * because we can assure there's no thread waiting on
 741          * pending_free_wq if there's a drain going on, but it has
 742          * to be taken into account if the current model is changed.
 743          */
 744         if (atomic_dec_and_test(&ring->inflight) && atomic_read(&blkif->drain)) {
 745                 complete(&blkif->drain_complete);
 746         }
 747         xen_blkif_put(blkif);
 748 }
 749
 750 static void xen_blkbk_unmap_and_respond(struct pending_req *req)
 751 {
 752         struct gntab_unmap_queue_data* work = &req->gnttab_unmap_data;
 753         struct xen_blkif_ring *ring = req->ring;
 754         struct grant_page **pages = req->segments;
 755         unsigned int invcount;
 756
 757         invcount = xen_blkbk_unmap_prepare(ring, pages, req->nr_segs,
 758                                            req->unmap, req->unmap_pages);
 759
 760         work->data = req;
 761         work->done = xen_blkbk_unmap_and_respond_callback;
 762         work->unmap_ops = req->unmap;
 763         work->kunmap_ops = NULL;
 764         work->pages = req->unmap_pages;
 765         work->count = invcount;
 766
 767         gnttab_unmap_refs_async(&req->gnttab_unmap_data);
 768 }
 769
 770
 771 /*
 772  * Unmap the grant references.
 773  *
 774  * This could accumulate ops up to the batch size to reduce the number
 775  * of hypercalls, but since this is only used in error paths there's
 776  * no real need.
 777  */
 778 static void xen_blkbk_unmap(struct xen_blkif_ring *ring,
 779                             struct grant_page *pages[],
 780                             int num)
 781 {
 782         struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 783         struct page *unmap_pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 784         unsigned int invcount = 0;
 785         int ret;
 786
 787         while (num) {
 788                 unsigned int batch = min(num, BLKIF_MAX_SEGMENTS_PER_REQUEST);
 789
 790                 invcount = xen_blkbk_unmap_prepare(ring, pages, batch,
 791                                                    unmap, unmap_pages);
 792                 if (invcount) {
 793                         ret = gnttab_unmap_refs(unmap, NULL, unmap_pages, invcount);
 794                         BUG_ON(ret);
 795                         put_free_pages(ring, unmap_pages, invcount);
 796                 }
 797                 pages += batch;
 798                 num -= batch;
 799         }
 800 }
 801
 802 static int xen_blkbk_map(struct xen_blkif_ring *ring,
 803                          struct grant_page *pages[],
 804                          int num, bool ro)
 805 {
 806         struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 807         struct page *pages_to_gnt[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 808         struct persistent_gnt *persistent_gnt = NULL;
 809         phys_addr_t addr = 0;
 810         int i, seg_idx, new_map_idx;
 811         int segs_to_map = 0;
 812         int ret = 0;
 813         int last_map = 0, map_until = 0;
 814         int use_persistent_gnts;
 815         struct xen_blkif *blkif = ring->blkif;
 816
 817         use_persistent_gnts = (blkif->vbd.feature_gnt_persistent);
 818
 819         /*
 820          * Fill out preq.nr_sects with proper amount of sectors, and setup
 821          * assign map[..] with the PFN of the page in our domain with the
 822          * corresponding grant reference for each page.
 823          */
 824 again:
 825         for (i = map_until; i < num; i++) {
 826                 uint32_t flags;
 827
 828                 if (use_persistent_gnts) {
 829                         persistent_gnt = get_persistent_gnt(
 830                                 ring,
 831                                 pages[i]->gref);
 832                 }
 833
 834                 if (persistent_gnt) {
 835                         /*
 836                          * We are using persistent grants and
 837                          * the grant is already mapped
 838                          */
 839                         pages[i]->page = persistent_gnt->page;
 840                         pages[i]->persistent_gnt = persistent_gnt;
 841                 } else {
 842                         if (get_free_page(ring, &pages[i]->page))
 843                                 goto out_of_memory;
 844                         addr = vaddr(pages[i]->page);
 845                         pages_to_gnt[segs_to_map] = pages[i]->page;
 846                         pages[i]->persistent_gnt = NULL;
 847                         flags = GNTMAP_host_map;
 848                         if (!use_persistent_gnts && ro)
 849                                 flags |= GNTMAP_readonly;
 850                         gnttab_set_map_op(&map[segs_to_map++], addr,
 851                                           flags, pages[i]->gref,
 852                                           blkif->domid);
 853                 }
 854                 map_until = i + 1;
 855                 if (segs_to_map == BLKIF_MAX_SEGMENTS_PER_REQUEST)
 856                         break;
 857         }
 858
 859         if (segs_to_map) {
 860                 ret = gnttab_map_refs(map, NULL, pages_to_gnt, segs_to_map);
 861                 BUG_ON(ret);
 862         }
 863
 864         /*
 865          * Now swizzle the MFN in our domain with the MFN from the other domain
 866          * so that when we access vaddr(pending_req,i) it has the contents of
 867          * the page from the other domain.
 868          */
 869         for (seg_idx = last_map, new_map_idx = 0; seg_idx < map_until; seg_idx++) {
 870                 if (!pages[seg_idx]->persistent_gnt) {
 871                         /* This is a newly mapped grant */
 872                         BUG_ON(new_map_idx >= segs_to_map);
 873                         if (unlikely(map[new_map_idx].status != 0)) {
 874                                 pr_debug("invalid buffer -- could not remap it\n");
 875                                 put_free_pages(ring, &pages[seg_idx]->page, 1);
 876                                 pages[seg_idx]->handle = BLKBACK_INVALID_HANDLE;
 877                                 ret |= 1;
 878                                 goto next;
 879                         }
 880                         pages[seg_idx]->handle = map[new_map_idx].handle;
 881                 } else {
 882                         continue;
 883                 }
 884                 if (use_persistent_gnts &&
 885                     ring->persistent_gnt_c < xen_blkif_max_pgrants) {
 886                         /*
 887                          * We are using persistent grants, the grant is
 888                          * not mapped but we might have room for it.
 889                          */
 890                         persistent_gnt = kmalloc(sizeof(struct persistent_gnt),
 891                                                  GFP_KERNEL);
 892                         if (!persistent_gnt) {
 893                                 /*
 894                                  * If we don't have enough memory to
 895                                  * allocate the persistent_gnt struct
 896                                  * map this grant non-persistenly
 897                                  */
 898                                 goto next;
 899                         }
 900                         persistent_gnt->gnt = map[new_map_idx].ref;
 901                         persistent_gnt->handle = map[new_map_idx].handle;
 902                         persistent_gnt->page = pages[seg_idx]->page;
 903                         if (add_persistent_gnt(ring,
 904                                                persistent_gnt)) {
 905                                 kfree(persistent_gnt);
 906                                 persistent_gnt = NULL;
 907                                 goto next;
 908                         }
 909                         pages[seg_idx]->persistent_gnt = persistent_gnt;
 910                         pr_debug("grant %u added to the tree of persistent grants, using %u/%u\n",
 911                                  persistent_gnt->gnt, ring->persistent_gnt_c,
 912                                  xen_blkif_max_pgrants);
 913                         goto next;
 914                 }
 915                 if (use_persistent_gnts && !blkif->vbd.overflow_max_grants) {
 916                         blkif->vbd.overflow_max_grants = 1;
 917                         pr_debug("domain %u, device %#x is using maximum number of persistent grants\n",
 918                                  blkif->domid, blkif->vbd.handle);
 919                 }
 920                 /*
 921                  * We could not map this grant persistently, so use it as
 922                  * a non-persistent grant.
 923                  */
 924 next:
 925                 new_map_idx++;
 926         }
 927         segs_to_map = 0;
 928         last_map = map_until;
 929         if (map_until != num)
 930                 goto again;
 931
 932         return ret;
 933
 934 out_of_memory:
 935         pr_alert("%s: out of memory\n", __func__);
 936         put_free_pages(ring, pages_to_gnt, segs_to_map);
 937         return -ENOMEM;
 938 }
 939
 940 static int xen_blkbk_map_seg(struct pending_req *pending_req)
 941 {
 942         int rc;
 943
 944         rc = xen_blkbk_map(pending_req->ring, pending_req->segments,
 945                            pending_req->nr_segs,
 946                            (pending_req->operation != BLKIF_OP_READ));
 947
 948         return rc;
 949 }
 950
 951 static int xen_blkbk_parse_indirect(struct blkif_request *req,
 952                                     struct pending_req *pending_req,
 953                                     struct seg_buf seg[],
 954                                     struct phys_req *preq)
 955 {
 956         struct grant_page **pages = pending_req->indirect_pages;
 957         struct xen_blkif_ring *ring = pending_req->ring;
 958         int indirect_grefs, rc, n, nseg, i;
 959         struct blkif_request_segment *segments = NULL;
 960
 961         nseg = pending_req->nr_segs;
 962         indirect_grefs = INDIRECT_PAGES(nseg);
 963         BUG_ON(indirect_grefs > BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST);
 964
 965         for (i = 0; i < indirect_grefs; i++)
 966                 pages[i]->gref = req->u.indirect.indirect_grefs[i];
 967
 968         rc = xen_blkbk_map(ring, pages, indirect_grefs, true);
 969         if (rc)
 970                 goto unmap;
 971
 972         for (n = 0, i = 0; n < nseg; n++) {
 973                 if ((n % SEGS_PER_INDIRECT_FRAME) == 0) {
 974                         /* Map indirect segments */
 975                         if (segments)
 976                                 kunmap_atomic(segments);
 977                         segments = kmap_atomic(pages[n/SEGS_PER_INDIRECT_FRAME]->page);
 978                 }
 979                 i = n % SEGS_PER_INDIRECT_FRAME;
 980                 pending_req->segments[n]->gref = segments[i].gref;
 981                 seg[n].nsec = segments[i].last_sect -
 982                         segments[i].first_sect + 1;
 983                 seg[n].offset = (segments[i].first_sect << 9);
 984                 if ((segments[i].last_sect >= (XEN_PAGE_SIZE >> 9)) ||
 985                     (segments[i].last_sect < segments[i].first_sect)) {
 986                         rc = -EINVAL;
 987                         goto unmap;
 988                 }
 989                 preq->nr_sects += seg[n].nsec;
 990         }
 991
 992 unmap:
 993         if (segments)
 994                 kunmap_atomic(segments);
 995         xen_blkbk_unmap(ring, pages, indirect_grefs);
 996         return rc;
 997 }
 998
 999 static int dispatch_discard_io(struct xen_blkif_ring *ring,
1000                                 struct blkif_request *req)
1001 {
1002         int err = 0;
1003         int status = BLKIF_RSP_OKAY;
1004         struct xen_blkif *blkif = ring->blkif;
1005         struct block_device *bdev = blkif->vbd.bdev;
1006         unsigned long secure;
1007         struct phys_req preq;
1008
1009         xen_blkif_get(blkif);
1010
1011         preq.sector_number = req->u.discard.sector_number;
1012         preq.nr_sects      = req->u.discard.nr_sectors;
1013
1014         err = xen_vbd_translate(&preq, blkif, WRITE);
1015         if (err) {
1016                 pr_warn("access denied: DISCARD [%llu->%llu] on dev=%04x\n",
1017                         preq.sector_number,
1018                         preq.sector_number + preq.nr_sects, blkif->vbd.pdevice);
1019                 goto fail_response;
1020         }
1021         blkif->st_ds_req++;
1022
1023         secure = (blkif->vbd.discard_secure &&
1024                  (req->u.discard.flag & BLKIF_DISCARD_SECURE)) ?
1025                  BLKDEV_DISCARD_SECURE : 0;
1026
1027         err = blkdev_issue_discard(bdev, req->u.discard.sector_number,
1028                                    req->u.discard.nr_sectors,
1029                                    GFP_KERNEL, secure);
1030 fail_response:
1031         if (err == -EOPNOTSUPP) {
1032                 pr_debug("discard op failed, not supported\n");
1033                 status = BLKIF_RSP_EOPNOTSUPP;
1034         } else if (err)
1035                 status = BLKIF_RSP_ERROR;
1036
1037         make_response(ring, req->u.discard.id, req->operation, status);
1038         xen_blkif_put(blkif);
1039         return err;
1040 }
1041
1042 static int dispatch_other_io(struct xen_blkif_ring *ring,
1043                              struct blkif_request *req,
1044                              struct pending_req *pending_req)
1045 {
1046         free_req(ring, pending_req);
1047         make_response(ring, req->u.other.id, req->operation,
1048                       BLKIF_RSP_EOPNOTSUPP);
1049         return -EIO;
1050 }
1051
1052 static void xen_blk_drain_io(struct xen_blkif_ring *ring)
1053 {
1054         struct xen_blkif *blkif = ring->blkif;
1055
1056         atomic_set(&blkif->drain, 1);
1057         do {
1058                 if (atomic_read(&ring->inflight) == 0)
1059                         break;
1060                 wait_for_completion_interruptible_timeout(
1061                                 &blkif->drain_complete, HZ);
1062
1063                 if (!atomic_read(&blkif->drain))
1064                         break;
1065         } while (!kthread_should_stop());
1066         atomic_set(&blkif->drain, 0);
1067 }
1068
1069 /*
1070  * Completion callback on the bio's. Called as bh->b_end_io()
1071  */
1072
1073 static void __end_block_io_op(struct pending_req *pending_req, int error)
1074 {
1075         /* An error fails the entire request. */
1076         if ((pending_req->operation == BLKIF_OP_FLUSH_DISKCACHE) &&
1077             (error == -EOPNOTSUPP)) {
1078                 pr_debug("flush diskcache op failed, not supported\n");
1079                 xen_blkbk_flush_diskcache(XBT_NIL, pending_req->ring->blkif->be, 0);
1080                 pending_req->status = BLKIF_RSP_EOPNOTSUPP;
1081         } else if ((pending_req->operation == BLKIF_OP_WRITE_BARRIER) &&
1082                     (error == -EOPNOTSUPP)) {
1083                 pr_debug("write barrier op failed, not supported\n");
1084                 xen_blkbk_barrier(XBT_NIL, pending_req->ring->blkif->be, 0);
1085                 pending_req->status = BLKIF_RSP_EOPNOTSUPP;
1086         } else if (error) {
1087                 pr_debug("Buffer not up-to-date at end of operation,"
1088                          " error=%d\n", error);
1089                 pending_req->status = BLKIF_RSP_ERROR;
1090         }
1091
1092         /*
1093          * If all of the bio's have completed it is time to unmap
1094          * the grant references associated with 'request' and provide
1095          * the proper response on the ring.
1096          */
1097         if (atomic_dec_and_test(&pending_req->pendcnt))
1098                 xen_blkbk_unmap_and_respond(pending_req);
1099 }
1100
1101 /*
1102  * bio callback.
1103  */
1104 static void end_block_io_op(struct bio *bio)
1105 {
1106         __end_block_io_op(bio->bi_private, bio->bi_error);
1107         bio_put(bio);
1108 }
1109
1110
1111
1112 /*
1113  * Function to copy the from the ring buffer the 'struct blkif_request'
1114  * (which has the sectors we want, number of them, grant references, etc),
1115  * and transmute  it to the block API to hand it over to the proper block disk.
1116  */
1117 static int
1118 __do_block_io_op(struct xen_blkif_ring *ring)
1119 {
1120         union blkif_back_rings *blk_rings = &ring->blk_rings;
1121         struct blkif_request req;
1122         struct pending_req *pending_req;
1123         RING_IDX rc, rp;
1124         int more_to_do = 0;
1125
1126         rc = blk_rings->common.req_cons;
1127         rp = blk_rings->common.sring->req_prod;
1128         rmb(); /* Ensure we see queued requests up to 'rp'. */
1129
1130         if (RING_REQUEST_PROD_OVERFLOW(&blk_rings->common, rp)) {
1131                 rc = blk_rings->common.rsp_prod_pvt;
1132                 pr_warn("Frontend provided bogus ring requests (%d - %d = %d). Halting ring processing on dev=%04x\n",
1133                         rp, rc, rp - rc, ring->blkif->vbd.pdevice);
1134                 return -EACCES;
1135         }
1136         while (rc != rp) {
1137
1138                 if (RING_REQUEST_CONS_OVERFLOW(&blk_rings->common, rc))
1139                         break;
1140
1141                 if (kthread_should_stop()) {
1142                         more_to_do = 1;
1143                         break;
1144                 }
1145
1146                 pending_req = alloc_req(ring);
1147                 if (NULL == pending_req) {
1148                         ring->blkif->st_oo_req++;
1149                         more_to_do = 1;
1150                         break;
1151                 }
1152
1153                 switch (ring->blkif->blk_protocol) {
1154                 case BLKIF_PROTOCOL_NATIVE:
1155                         memcpy(&req, RING_GET_REQUEST(&blk_rings->native, rc), sizeof(req));
1156                         break;
1157                 case BLKIF_PROTOCOL_X86_32:
1158                         blkif_get_x86_32_req(&req, RING_GET_REQUEST(&blk_rings->x86_32, rc));
1159                         break;
1160                 case BLKIF_PROTOCOL_X86_64:
1161                         blkif_get_x86_64_req(&req, RING_GET_REQUEST(&blk_rings->x86_64, rc));
1162                         break;
1163                 default:
1164                         BUG();
1165                 }
1166                 blk_rings->common.req_cons = ++rc; /* before make_response() */
1167
1168                 /* Apply all sanity checks to /private copy/ of request. */
1169                 barrier();
1170
1171                 switch (req.operation) {
1172                 case BLKIF_OP_READ:
1173                 case BLKIF_OP_WRITE:
1174                 case BLKIF_OP_WRITE_BARRIER:
1175                 case BLKIF_OP_FLUSH_DISKCACHE:
1176                 case BLKIF_OP_INDIRECT:
1177                         if (dispatch_rw_block_io(ring, &req, pending_req))
1178                                 goto done;
1179                         break;
1180                 case BLKIF_OP_DISCARD:
1181                         free_req(ring, pending_req);
1182                         if (dispatch_discard_io(ring, &req))
1183                                 goto done;
1184                         break;
1185                 default:
1186                         if (dispatch_other_io(ring, &req, pending_req))
1187                                 goto done;
1188                         break;
1189                 }
1190
1191                 /* Yield point for this unbounded loop. */
1192                 cond_resched();
1193         }
1194 done:
1195         return more_to_do;
1196 }
1197
1198 static int
1199 do_block_io_op(struct xen_blkif_ring *ring)
1200 {
1201         union blkif_back_rings *blk_rings = &ring->blk_rings;
1202         int more_to_do;
1203
1204         do {
1205                 more_to_do = __do_block_io_op(ring);
1206                 if (more_to_do)
1207                         break;
1208
1209                 RING_FINAL_CHECK_FOR_REQUESTS(&blk_rings->common, more_to_do);
1210         } while (more_to_do);
1211
1212         return more_to_do;
1213 }
1214 /*
1215  * Transmutation of the 'struct blkif_request' to a proper 'struct bio'
1216  * and call the 'submit_bio' to pass it to the underlying storage.
1217  */
1218 static int dispatch_rw_block_io(struct xen_blkif_ring *ring,
1219                                 struct blkif_request *req,
1220                                 struct pending_req *pending_req)
1221 {
1222         struct phys_req preq;
1223         struct seg_buf *seg = pending_req->seg;
1224         unsigned int nseg;
1225         struct bio *bio = NULL;
1226         struct bio **biolist = pending_req->biolist;
1227         int i, nbio = 0;
1228         int operation;
1229         struct blk_plug plug;
1230         bool drain = false;
1231         struct grant_page **pages = pending_req->segments;
1232         unsigned short req_operation;
1233
1234         req_operation = req->operation == BLKIF_OP_INDIRECT ?
1235                         req->u.indirect.indirect_op : req->operation;
1236
1237         if ((req->operation == BLKIF_OP_INDIRECT) &&
1238             (req_operation != BLKIF_OP_READ) &&
1239             (req_operation != BLKIF_OP_WRITE)) {
1240                 pr_debug("Invalid indirect operation (%u)\n", req_operation);
1241                 goto fail_response;
1242         }
1243
1244         switch (req_operation) {
1245         case BLKIF_OP_READ:
1246                 ring->blkif->st_rd_req++;
1247                 operation = READ;
1248                 break;
1249         case BLKIF_OP_WRITE:
1250                 ring->blkif->st_wr_req++;
1251                 operation = WRITE_ODIRECT;
1252                 break;
1253         case BLKIF_OP_WRITE_BARRIER:
1254                 drain = true;
1255         case BLKIF_OP_FLUSH_DISKCACHE:
1256                 ring->blkif->st_f_req++;
1257                 operation = WRITE_FLUSH;
1258                 break;
1259         default:
1260                 operation = 0; /* make gcc happy */
1261                 goto fail_response;
1262                 break;
1263         }
1264
1265         /* Check that the number of segments is sane. */
1266         nseg = req->operation == BLKIF_OP_INDIRECT ?
1267                req->u.indirect.nr_segments : req->u.rw.nr_segments;
1268
1269         if (unlikely(nseg == 0 && operation != WRITE_FLUSH) ||
1270             unlikely((req->operation != BLKIF_OP_INDIRECT) &&
1271                      (nseg > BLKIF_MAX_SEGMENTS_PER_REQUEST)) ||
1272             unlikely((req->operation == BLKIF_OP_INDIRECT) &&
1273                      (nseg > MAX_INDIRECT_SEGMENTS))) {
1274                 pr_debug("Bad number of segments in request (%d)\n", nseg);
1275                 /* Haven't submitted any bio's yet. */
1276                 goto fail_response;
1277         }
1278
1279         preq.nr_sects      = 0;
1280
1281         pending_req->ring      = ring;
1282         pending_req->id        = req->u.rw.id;
1283         pending_req->operation = req_operation;
1284         pending_req->status    = BLKIF_RSP_OKAY;
1285         pending_req->nr_segs   = nseg;
1286
1287         if (req->operation != BLKIF_OP_INDIRECT) {
1288                 preq.dev               = req->u.rw.handle;
1289                 preq.sector_number     = req->u.rw.sector_number;
1290                 for (i = 0; i < nseg; i++) {
1291                         pages[i]->gref = req->u.rw.seg[i].gref;
1292                         seg[i].nsec = req->u.rw.seg[i].last_sect -
1293                                 req->u.rw.seg[i].first_sect + 1;
1294                         seg[i].offset = (req->u.rw.seg[i].first_sect << 9);
1295                         if ((req->u.rw.seg[i].last_sect >= (XEN_PAGE_SIZE >> 9)) ||
1296                             (req->u.rw.seg[i].last_sect <
1297                              req->u.rw.seg[i].first_sect))
1298                                 goto fail_response;
1299                         preq.nr_sects += seg[i].nsec;
1300                 }
1301         } else {
1302                 preq.dev               = req->u.indirect.handle;
1303                 preq.sector_number     = req->u.indirect.sector_number;
1304                 if (xen_blkbk_parse_indirect(req, pending_req, seg, &preq))
1305                         goto fail_response;
1306         }
1307
1308         if (xen_vbd_translate(&preq, ring->blkif, operation) != 0) {
1309                 pr_debug("access denied: %s of [%llu,%llu] on dev=%04x\n",
1310                          operation == READ ? "read" : "write",
1311                          preq.sector_number,
1312                          preq.sector_number + preq.nr_sects,
1313                          ring->blkif->vbd.pdevice);
1314                 goto fail_response;
1315         }
1316
1317         /*
1318          * This check _MUST_ be done after xen_vbd_translate as the preq.bdev
1319          * is set there.
1320          */
1321         for (i = 0; i < nseg; i++) {
1322                 if (((int)preq.sector_number|(int)seg[i].nsec) &
1323                     ((bdev_logical_block_size(preq.bdev) >> 9) - 1)) {
1324                         pr_debug("Misaligned I/O request from domain %d\n",
1325                                  ring->blkif->domid);
1326                         goto fail_response;
1327                 }
1328         }
1329
1330         /* Wait on all outstanding I/O's and once that has been completed
1331          * issue the WRITE_FLUSH.
1332          */
1333         if (drain)
1334                 xen_blk_drain_io(pending_req->ring);
1335
1336         /*
1337          * If we have failed at this point, we need to undo the M2P override,
1338          * set gnttab_set_unmap_op on all of the grant references and perform
1339          * the hypercall to unmap the grants - that is all done in
1340          * xen_blkbk_unmap.
1341          */
1342         if (xen_blkbk_map_seg(pending_req))
1343                 goto fail_flush;
1344
1345         /*
1346          * This corresponding xen_blkif_put is done in __end_block_io_op, or
1347          * below (in "!bio") if we are handling a BLKIF_OP_DISCARD.
1348          */
1349         xen_blkif_get(ring->blkif);
1350         atomic_inc(&ring->inflight);
1351
1352         for (i = 0; i < nseg; i++) {
1353                 while ((bio == NULL) ||
1354                        (bio_add_page(bio,
1355                                      pages[i]->page,
1356                                      seg[i].nsec << 9,
1357                                      seg[i].offset) == 0)) {
1358
1359                         int nr_iovecs = min_t(int, (nseg-i), BIO_MAX_PAGES);
1360                         bio = bio_alloc(GFP_KERNEL, nr_iovecs);
1361                         if (unlikely(bio == NULL))
1362                                 goto fail_put_bio;
1363
1364                         biolist[nbio++] = bio;
1365                         bio->bi_bdev    = preq.bdev;
1366                         bio->bi_private = pending_req;
1367                         bio->bi_end_io  = end_block_io_op;
1368                         bio->bi_iter.bi_sector  = preq.sector_number;
1369                 }
1370
1371                 preq.sector_number += seg[i].nsec;
1372         }
1373
1374         /* This will be hit if the operation was a flush or discard. */
1375         if (!bio) {
1376                 BUG_ON(operation != WRITE_FLUSH);
1377
1378                 bio = bio_alloc(GFP_KERNEL, 0);
1379                 if (unlikely(bio == NULL))
1380                         goto fail_put_bio;
1381
1382                 biolist[nbio++] = bio;
1383                 bio->bi_bdev    = preq.bdev;
1384                 bio->bi_private = pending_req;
1385                 bio->bi_end_io  = end_block_io_op;
1386         }
1387
1388         atomic_set(&pending_req->pendcnt, nbio);
1389         blk_start_plug(&plug);
1390
1391         for (i = 0; i < nbio; i++)
1392                 submit_bio(operation, biolist[i]);
1393
1394         /* Let the I/Os go.. */
1395         blk_finish_plug(&plug);
1396
1397         if (operation == READ)
1398                 ring->blkif->st_rd_sect += preq.nr_sects;
1399         else if (operation & WRITE)
1400                 ring->blkif->st_wr_sect += preq.nr_sects;
1401
1402         return 0;
1403
1404  fail_flush:
1405         xen_blkbk_unmap(ring, pending_req->segments,
1406                         pending_req->nr_segs);
1407  fail_response:
1408         /* Haven't submitted any bio's yet. */
1409         make_response(ring, req->u.rw.id, req_operation, BLKIF_RSP_ERROR);
1410         free_req(ring, pending_req);
1411         msleep(1); /* back off a bit */
1412         return -EIO;
1413
1414  fail_put_bio:
1415         for (i = 0; i < nbio; i++)
1416                 bio_put(biolist[i]);
1417         atomic_set(&pending_req->pendcnt, 1);
1418         __end_block_io_op(pending_req, -EINVAL);
1419         msleep(1); /* back off a bit */
1420         return -EIO;
1421 }
1422
1423
1424
1425 /*
1426  * Put a response on the ring on how the operation fared.
1427  */
1428 static void make_response(struct xen_blkif_ring *ring, u64 id,
1429                           unsigned short op, int st)
1430 {
1431         struct blkif_response  resp;
1432         unsigned long     flags;
1433         union blkif_back_rings *blk_rings;
1434         int notify;
1435
1436         resp.id        = id;
1437         resp.operation = op;
1438         resp.status    = st;
1439
1440         spin_lock_irqsave(&ring->blk_ring_lock, flags);
1441         blk_rings = &ring->blk_rings;
1442         /* Place on the response ring for the relevant domain. */
1443         switch (ring->blkif->blk_protocol) {
1444         case BLKIF_PROTOCOL_NATIVE:
1445                 memcpy(RING_GET_RESPONSE(&blk_rings->native, blk_rings->native.rsp_prod_pvt),
1446                        &resp, sizeof(resp));
1447                 break;
1448         case BLKIF_PROTOCOL_X86_32:
1449                 memcpy(RING_GET_RESPONSE(&blk_rings->x86_32, blk_rings->x86_32.rsp_prod_pvt),
1450                        &resp, sizeof(resp));
1451                 break;
1452         case BLKIF_PROTOCOL_X86_64:
1453                 memcpy(RING_GET_RESPONSE(&blk_rings->x86_64, blk_rings->x86_64.rsp_prod_pvt),
1454                        &resp, sizeof(resp));
1455                 break;
1456         default:
1457                 BUG();
1458         }
1459         blk_rings->common.rsp_prod_pvt++;
1460         RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&blk_rings->common, notify);
1461         spin_unlock_irqrestore(&ring->blk_ring_lock, flags);
1462         if (notify)
1463                 notify_remote_via_irq(ring->irq);
1464 }
1465
1466 static int __init xen_blkif_init(void)
1467 {
1468         int rc = 0;
1469
1470         if (!xen_domain())
1471                 return -ENODEV;
1472
1473         if (xen_blkif_max_ring_order > XENBUS_MAX_RING_GRANT_ORDER) {
1474                 pr_info("Invalid max_ring_order (%d), will use default max: %d.\n",
1475                         xen_blkif_max_ring_order, XENBUS_MAX_RING_GRANT_ORDER);
1476                 xen_blkif_max_ring_order = XENBUS_MAX_RING_GRANT_ORDER;
1477         }
1478
1479         if (xenblk_max_queues == 0)
1480                 xenblk_max_queues = num_online_cpus();
1481
1482         rc = xen_blkif_interface_init();
1483         if (rc)
1484                 goto failed_init;
1485
1486         rc = xen_blkif_xenbus_init();
1487         if (rc)
1488                 goto failed_init;
1489
1490  failed_init:
1491         return rc;
1492 }
1493
1494 module_init(xen_blkif_init);
1495
1496 MODULE_LICENSE("Dual BSD/GPL");
1497 MODULE_ALIAS("xen-backend:vbd");