drivers/xen/balloon.c

   1 /******************************************************************************
   2  * Xen balloon driver - enables returning/claiming memory to/from Xen.
   3  *
   4  * Copyright (c) 2003, B Dragovic
   5  * Copyright (c) 2003-2004, M Williamson, K Fraser
   6  * Copyright (c) 2005 Dan M. Smith, IBM Corporation
   7  * Copyright (c) 2010 Daniel Kiper
   8  *
   9  * Memory hotplug support was written by Daniel Kiper. Work on
  10  * it was sponsored by Google under Google Summer of Code 2010
  11  * program. Jeremy Fitzhardinge from Citrix was the mentor for
  12  * this project.
  13  *
  14  * This program is free software; you can redistribute it and/or
  15  * modify it under the terms of the GNU General Public License version 2
  16  * as published by the Free Software Foundation; or, when distributed
  17  * separately from the Linux kernel or incorporated into other
  18  * software packages, subject to the following license:
  19  *
  20  * Permission is hereby granted, free of charge, to any person obtaining a copy
  21  * of this source file (the "Software"), to deal in the Software without
  22  * restriction, including without limitation the rights to use, copy, modify,
  23  * merge, publish, distribute, sublicense, and/or sell copies of the Software,
  24  * and to permit persons to whom the Software is furnished to do so, subject to
  25  * the following conditions:
  26  *
  27  * The above copyright notice and this permission notice shall be included in
  28  * all copies or substantial portions of the Software.
  29  *
  30  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  31  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  32  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  33  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  34  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  35  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
  36  * IN THE SOFTWARE.
  37  */
  38
  39 #define pr_fmt(fmt) "xen:" KBUILD_MODNAME ": " fmt
  40
  41 #include <linux/cpu.h>
  42 #include <linux/kernel.h>
  43 #include <linux/sched.h>
  44 #include <linux/errno.h>
  45 #include <linux/mm.h>
  46 #include <linux/bootmem.h>
  47 #include <linux/pagemap.h>
  48 #include <linux/highmem.h>
  49 #include <linux/mutex.h>
  50 #include <linux/list.h>
  51 #include <linux/gfp.h>
  52 #include <linux/notifier.h>
  53 #include <linux/memory.h>
  54 #include <linux/memory_hotplug.h>
  55 #include <linux/percpu-defs.h>
  56 #include <linux/slab.h>
  57 #include <linux/sysctl.h>
  58
  59 #include <asm/page.h>
  60 #include <asm/pgalloc.h>
  61 #include <asm/pgtable.h>
  62 #include <asm/tlb.h>
  63
  64 #include <asm/xen/hypervisor.h>
  65 #include <asm/xen/hypercall.h>
  66
  67 #include <xen/xen.h>
  68 #include <xen/interface/xen.h>
  69 #include <xen/interface/memory.h>
  70 #include <xen/balloon.h>
  71 #include <xen/features.h>
  72 #include <xen/page.h>
  73
  74 static int xen_hotplug_unpopulated;
  75
  76 #ifdef CONFIG_XEN_BALLOON_MEMORY_HOTPLUG
  77
  78 static int zero;
  79 static int one = 1;
  80
  81 static struct ctl_table balloon_table[] = {
  82         {
  83                 .procname       = "hotplug_unpopulated",
  84                 .data           = &xen_hotplug_unpopulated,
  85                 .maxlen         = sizeof(int),
  86                 .mode           = 0644,
  87                 .proc_handler   = proc_dointvec_minmax,
  88                 .extra1         = &zero,
  89                 .extra2         = &one,
  90         },
  91         { }
  92 };
  93
  94 static struct ctl_table balloon_root[] = {
  95         {
  96                 .procname       = "balloon",
  97                 .mode           = 0555,
  98                 .child          = balloon_table,
  99         },
 100         { }
 101 };
 102
 103 static struct ctl_table xen_root[] = {
 104         {
 105                 .procname       = "xen",
 106                 .mode           = 0555,
 107                 .child          = balloon_root,
 108         },
 109         { }
 110 };
 111
 112 #endif
 113
 114 /*
 115  * Use one extent per PAGE_SIZE to avoid to break down the page into
 116  * multiple frame.
 117  */
 118 #define EXTENT_ORDER (fls(XEN_PFN_PER_PAGE) - 1)
 119
 120 /*
 121  * balloon_process() state:
 122  *
 123  * BP_DONE: done or nothing to do,
 124  * BP_WAIT: wait to be rescheduled,
 125  * BP_EAGAIN: error, go to sleep,
 126  * BP_ECANCELED: error, balloon operation canceled.
 127  */
 128
 129 enum bp_state {
 130         BP_DONE,
 131         BP_WAIT,
 132         BP_EAGAIN,
 133         BP_ECANCELED
 134 };
 135
 136
 137 static DEFINE_MUTEX(balloon_mutex);
 138
 139 struct balloon_stats balloon_stats;
 140 EXPORT_SYMBOL_GPL(balloon_stats);
 141
 142 /* We increase/decrease in batches which fit in a page */
 143 static xen_pfn_t frame_list[PAGE_SIZE / sizeof(xen_pfn_t)];
 144
 145
 146 /* List of ballooned pages, threaded through the mem_map array. */
 147 static LIST_HEAD(ballooned_pages);
 148 static DECLARE_WAIT_QUEUE_HEAD(balloon_wq);
 149
 150 /* Main work function, always executed in process context. */
 151 static void balloon_process(struct work_struct *work);
 152 static DECLARE_DELAYED_WORK(balloon_worker, balloon_process);
 153
 154 static void release_memory_resource(struct resource *resource);
 155
 156 /* When ballooning out (allocating memory to return to Xen) we don't really
 157    want the kernel to try too hard since that can trigger the oom killer. */
 158 #define GFP_BALLOON \
 159         (GFP_HIGHUSER | __GFP_NOWARN | __GFP_NORETRY | __GFP_NOMEMALLOC)
 160
 161 static void scrub_page(struct page *page)
 162 {
 163 #ifdef CONFIG_XEN_SCRUB_PAGES
 164         clear_highpage(page);
 165 #endif
 166 }
 167
 168 /* balloon_append: add the given page to the balloon. */
 169 static void __balloon_append(struct page *page)
 170 {
 171         /* Lowmem is re-populated first, so highmem pages go at list tail. */
 172         if (PageHighMem(page)) {
 173                 list_add_tail(&page->lru, &ballooned_pages);
 174                 balloon_stats.balloon_high++;
 175         } else {
 176                 list_add(&page->lru, &ballooned_pages);
 177                 balloon_stats.balloon_low++;
 178         }
 179         wake_up(&balloon_wq);
 180 }
 181
 182 static void balloon_append(struct page *page)
 183 {
 184         __balloon_append(page);
 185         adjust_managed_page_count(page, -1);
 186 }
 187
 188 /* balloon_retrieve: rescue a page from the balloon, if it is not empty. */
 189 static struct page *balloon_retrieve(bool require_lowmem)
 190 {
 191         struct page *page;
 192
 193         if (list_empty(&ballooned_pages))
 194                 return NULL;
 195
 196         page = list_entry(ballooned_pages.next, struct page, lru);
 197         if (require_lowmem && PageHighMem(page))
 198                 return NULL;
 199         list_del(&page->lru);
 200
 201         if (PageHighMem(page))
 202                 balloon_stats.balloon_high--;
 203         else
 204                 balloon_stats.balloon_low--;
 205
 206         adjust_managed_page_count(page, 1);
 207
 208         return page;
 209 }
 210
 211 static struct page *balloon_next_page(struct page *page)
 212 {
 213         struct list_head *next = page->lru.next;
 214         if (next == &ballooned_pages)
 215                 return NULL;
 216         return list_entry(next, struct page, lru);
 217 }
 218
 219 static enum bp_state update_schedule(enum bp_state state)
 220 {
 221         if (state == BP_WAIT)
 222                 return BP_WAIT;
 223
 224         if (state == BP_ECANCELED)
 225                 return BP_ECANCELED;
 226
 227         if (state == BP_DONE) {
 228                 balloon_stats.schedule_delay = 1;
 229                 balloon_stats.retry_count = 1;
 230                 return BP_DONE;
 231         }
 232
 233         ++balloon_stats.retry_count;
 234
 235         if (balloon_stats.max_retry_count != RETRY_UNLIMITED &&
 236                         balloon_stats.retry_count > balloon_stats.max_retry_count) {
 237                 balloon_stats.schedule_delay = 1;
 238                 balloon_stats.retry_count = 1;
 239                 return BP_ECANCELED;
 240         }
 241
 242         balloon_stats.schedule_delay <<= 1;
 243
 244         if (balloon_stats.schedule_delay > balloon_stats.max_schedule_delay)
 245                 balloon_stats.schedule_delay = balloon_stats.max_schedule_delay;
 246
 247         return BP_EAGAIN;
 248 }
 249
 250 #ifdef CONFIG_XEN_BALLOON_MEMORY_HOTPLUG
 251 static struct resource *additional_memory_resource(phys_addr_t size)
 252 {
 253         struct resource *res;
 254         int ret;
 255
 256         res = kzalloc(sizeof(*res), GFP_KERNEL);
 257         if (!res)
 258                 return NULL;
 259
 260         res->name = "System RAM";
 261         res->flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
 262
 263         ret = allocate_resource(&iomem_resource, res,
 264                                 size, 0, -1,
 265                                 PAGES_PER_SECTION * PAGE_SIZE, NULL, NULL);
 266         if (ret < 0) {
 267                 pr_err("Cannot allocate new System RAM resource\n");
 268                 kfree(res);
 269                 return NULL;
 270         }
 271
 272 #ifdef CONFIG_SPARSEMEM
 273         {
 274                 unsigned long limit = 1UL << (MAX_PHYSMEM_BITS - PAGE_SHIFT);
 275                 unsigned long pfn = res->start >> PAGE_SHIFT;
 276
 277                 if (pfn > limit) {
 278                         pr_err("New System RAM resource outside addressable RAM (%lu > %lu)\n",
 279                                pfn, limit);
 280                         release_memory_resource(res);
 281                         return NULL;
 282                 }
 283         }
 284 #endif
 285
 286         return res;
 287 }
 288
 289 static void release_memory_resource(struct resource *resource)
 290 {
 291         if (!resource)
 292                 return;
 293
 294         /*
 295          * No need to reset region to identity mapped since we now
 296          * know that no I/O can be in this region
 297          */
 298         release_resource(resource);
 299         kfree(resource);
 300 }
 301
 302 static enum bp_state reserve_additional_memory(void)
 303 {
 304         long credit;
 305         struct resource *resource;
 306         int nid, rc;
 307         unsigned long balloon_hotplug;
 308
 309         credit = balloon_stats.target_pages + balloon_stats.target_unpopulated
 310                 - balloon_stats.total_pages;
 311
 312         /*
 313          * Already hotplugged enough pages?  Wait for them to be
 314          * onlined.
 315          */
 316         if (credit <= 0)
 317                 return BP_WAIT;
 318
 319         balloon_hotplug = round_up(credit, PAGES_PER_SECTION);
 320
 321         resource = additional_memory_resource(balloon_hotplug * PAGE_SIZE);
 322         if (!resource)
 323                 goto err;
 324
 325         nid = memory_add_physaddr_to_nid(resource->start);
 326
 327 #ifdef CONFIG_XEN_HAVE_PVMMU
 328         /*
 329          * We don't support PV MMU when Linux and Xen is using
 330          * different page granularity.
 331          */
 332         BUILD_BUG_ON(XEN_PAGE_SIZE != PAGE_SIZE);
 333
 334         /*
 335          * add_memory() will build page tables for the new memory so
 336          * the p2m must contain invalid entries so the correct
 337          * non-present PTEs will be written.
 338          *
 339          * If a failure occurs, the original (identity) p2m entries
 340          * are not restored since this region is now known not to
 341          * conflict with any devices.
 342          */
 343         if (!xen_feature(XENFEAT_auto_translated_physmap)) {
 344                 unsigned long pfn, i;
 345
 346                 pfn = PFN_DOWN(resource->start);
 347                 for (i = 0; i < balloon_hotplug; i++) {
 348                         if (!set_phys_to_machine(pfn + i, INVALID_P2M_ENTRY)) {
 349                                 pr_warn("set_phys_to_machine() failed, no memory added\n");
 350                                 goto err;
 351                         }
 352                 }
 353         }
 354 #endif
 355
 356         /*
 357          * add_memory_resource() will call online_pages() which in its turn
 358          * will call xen_online_page() callback causing deadlock if we don't
 359          * release balloon_mutex here. Unlocking here is safe because the
 360          * callers drop the mutex before trying again.
 361          */
 362         mutex_unlock(&balloon_mutex);
 363         rc = add_memory_resource(nid, resource, memhp_auto_online);
 364         mutex_lock(&balloon_mutex);
 365
 366         if (rc) {
 367                 pr_warn("Cannot add additional memory (%i)\n", rc);
 368                 goto err;
 369         }
 370
 371         balloon_stats.total_pages += balloon_hotplug;
 372
 373         return BP_WAIT;
 374   err:
 375         release_memory_resource(resource);
 376         return BP_ECANCELED;
 377 }
 378
 379 static void xen_online_page(struct page *page)
 380 {
 381         __online_page_set_limits(page);
 382
 383         mutex_lock(&balloon_mutex);
 384
 385         __balloon_append(page);
 386
 387         mutex_unlock(&balloon_mutex);
 388 }
 389
 390 static int xen_memory_notifier(struct notifier_block *nb, unsigned long val, void *v)
 391 {
 392         if (val == MEM_ONLINE)
 393                 schedule_delayed_work(&balloon_worker, 0);
 394
 395         return NOTIFY_OK;
 396 }
 397
 398 static struct notifier_block xen_memory_nb = {
 399         .notifier_call = xen_memory_notifier,
 400         .priority = 0
 401 };
 402 #else
 403 static enum bp_state reserve_additional_memory(void)
 404 {
 405         balloon_stats.target_pages = balloon_stats.current_pages;
 406         return BP_ECANCELED;
 407 }
 408 #endif /* CONFIG_XEN_BALLOON_MEMORY_HOTPLUG */
 409
 410 static long current_credit(void)
 411 {
 412         return balloon_stats.target_pages - balloon_stats.current_pages;
 413 }
 414
 415 static bool balloon_is_inflated(void)
 416 {
 417         return balloon_stats.balloon_low || balloon_stats.balloon_high;
 418 }
 419
 420 static enum bp_state increase_reservation(unsigned long nr_pages)
 421 {
 422         int rc;
 423         unsigned long i;
 424         struct page   *page;
 425         struct xen_memory_reservation reservation = {
 426                 .address_bits = 0,
 427                 .extent_order = EXTENT_ORDER,
 428                 .domid        = DOMID_SELF
 429         };
 430
 431         if (nr_pages > ARRAY_SIZE(frame_list))
 432                 nr_pages = ARRAY_SIZE(frame_list);
 433
 434         page = list_first_entry_or_null(&ballooned_pages, struct page, lru);
 435         for (i = 0; i < nr_pages; i++) {
 436                 if (!page) {
 437                         nr_pages = i;
 438                         break;
 439                 }
 440
 441                 /* XENMEM_populate_physmap requires a PFN based on Xen
 442                  * granularity.
 443                  */
 444                 frame_list[i] = page_to_xen_pfn(page);
 445                 page = balloon_next_page(page);
 446         }
 447
 448         set_xen_guest_handle(reservation.extent_start, frame_list);
 449         reservation.nr_extents = nr_pages;
 450         rc = HYPERVISOR_memory_op(XENMEM_populate_physmap, &reservation);
 451         if (rc <= 0)
 452                 return BP_EAGAIN;
 453
 454         for (i = 0; i < rc; i++) {
 455                 page = balloon_retrieve(false);
 456                 BUG_ON(page == NULL);
 457
 458 #ifdef CONFIG_XEN_HAVE_PVMMU
 459                 /*
 460                  * We don't support PV MMU when Linux and Xen is using
 461                  * different page granularity.
 462                  */
 463                 BUILD_BUG_ON(XEN_PAGE_SIZE != PAGE_SIZE);
 464
 465                 if (!xen_feature(XENFEAT_auto_translated_physmap)) {
 466                         unsigned long pfn = page_to_pfn(page);
 467
 468                         set_phys_to_machine(pfn, frame_list[i]);
 469
 470                         /* Link back into the page tables if not highmem. */
 471                         if (!PageHighMem(page)) {
 472                                 int ret;
 473                                 ret = HYPERVISOR_update_va_mapping(
 474                                                 (unsigned long)__va(pfn << PAGE_SHIFT),
 475                                                 mfn_pte(frame_list[i], PAGE_KERNEL),
 476                                                 0);
 477                                 BUG_ON(ret);
 478                         }
 479                 }
 480 #endif
 481
 482                 /* Relinquish the page back to the allocator. */
 483                 __free_reserved_page(page);
 484         }
 485
 486         balloon_stats.current_pages += rc;
 487
 488         return BP_DONE;
 489 }
 490
 491 static enum bp_state decrease_reservation(unsigned long nr_pages, gfp_t gfp)
 492 {
 493         enum bp_state state = BP_DONE;
 494         unsigned long i;
 495         struct page *page, *tmp;
 496         int ret;
 497         struct xen_memory_reservation reservation = {
 498                 .address_bits = 0,
 499                 .extent_order = EXTENT_ORDER,
 500                 .domid        = DOMID_SELF
 501         };
 502         LIST_HEAD(pages);
 503
 504         if (nr_pages > ARRAY_SIZE(frame_list))
 505                 nr_pages = ARRAY_SIZE(frame_list);
 506
 507         for (i = 0; i < nr_pages; i++) {
 508                 page = alloc_page(gfp);
 509                 if (page == NULL) {
 510                         nr_pages = i;
 511                         state = BP_EAGAIN;
 512                         break;
 513                 }
 514                 scrub_page(page);
 515                 list_add(&page->lru, &pages);
 516         }
 517
 518         /*
 519          * Ensure that ballooned highmem pages don't have kmaps.
 520          *
 521          * Do this before changing the p2m as kmap_flush_unused()
 522          * reads PTEs to obtain pages (and hence needs the original
 523          * p2m entry).
 524          */
 525         kmap_flush_unused();
 526
 527         /*
 528          * Setup the frame, update direct mapping, invalidate P2M,
 529          * and add to balloon.
 530          */
 531         i = 0;
 532         list_for_each_entry_safe(page, tmp, &pages, lru) {
 533                 /* XENMEM_decrease_reservation requires a GFN */
 534                 frame_list[i++] = xen_page_to_gfn(page);
 535
 536 #ifdef CONFIG_XEN_HAVE_PVMMU
 537                 /*
 538                  * We don't support PV MMU when Linux and Xen is using
 539                  * different page granularity.
 540                  */
 541                 BUILD_BUG_ON(XEN_PAGE_SIZE != PAGE_SIZE);
 542
 543                 if (!xen_feature(XENFEAT_auto_translated_physmap)) {
 544                         unsigned long pfn = page_to_pfn(page);
 545
 546                         if (!PageHighMem(page)) {
 547                                 ret = HYPERVISOR_update_va_mapping(
 548                                                 (unsigned long)__va(pfn << PAGE_SHIFT),
 549                                                 __pte_ma(0), 0);
 550                                 BUG_ON(ret);
 551                         }
 552                         __set_phys_to_machine(pfn, INVALID_P2M_ENTRY);
 553                 }
 554 #endif
 555                 list_del(&page->lru);
 556
 557                 balloon_append(page);
 558         }
 559
 560         flush_tlb_all();
 561
 562         set_xen_guest_handle(reservation.extent_start, frame_list);
 563         reservation.nr_extents   = nr_pages;
 564         ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation, &reservation);
 565         BUG_ON(ret != nr_pages);
 566
 567         balloon_stats.current_pages -= nr_pages;
 568
 569         return state;
 570 }
 571
 572 /*
 573  * As this is a work item it is guaranteed to run as a single instance only.
 574  * We may of course race updates of the target counts (which are protected
 575  * by the balloon lock), or with changes to the Xen hard limit, but we will
 576  * recover from these in time.
 577  */
 578 static void balloon_process(struct work_struct *work)
 579 {
 580         enum bp_state state = BP_DONE;
 581         long credit;
 582
 583
 584         do {
 585                 mutex_lock(&balloon_mutex);
 586
 587                 credit = current_credit();
 588
 589                 if (credit > 0) {
 590                         if (balloon_is_inflated())
 591                                 state = increase_reservation(credit);
 592                         else
 593                                 state = reserve_additional_memory();
 594                 }
 595
 596                 if (credit < 0)
 597                         state = decrease_reservation(-credit, GFP_BALLOON);
 598
 599                 state = update_schedule(state);
 600
 601                 mutex_unlock(&balloon_mutex);
 602
 603                 cond_resched();
 604
 605         } while (credit && state == BP_DONE);
 606
 607         /* Schedule more work if there is some still to be done. */
 608         if (state == BP_EAGAIN)
 609                 schedule_delayed_work(&balloon_worker, balloon_stats.schedule_delay * HZ);
 610 }
 611
 612 /* Resets the Xen limit, sets new target, and kicks off processing. */
 613 void balloon_set_new_target(unsigned long target)
 614 {
 615         /* No need for lock. Not read-modify-write updates. */
 616         balloon_stats.target_pages = target;
 617         schedule_delayed_work(&balloon_worker, 0);
 618 }
 619 EXPORT_SYMBOL_GPL(balloon_set_new_target);
 620
 621 static int add_ballooned_pages(int nr_pages)
 622 {
 623         enum bp_state st;
 624
 625         if (xen_hotplug_unpopulated) {
 626                 st = reserve_additional_memory();
 627                 if (st != BP_ECANCELED) {
 628                         mutex_unlock(&balloon_mutex);
 629                         wait_event(balloon_wq,
 630                                    !list_empty(&ballooned_pages));
 631                         mutex_lock(&balloon_mutex);
 632                         return 0;
 633                 }
 634         }
 635
 636         st = decrease_reservation(nr_pages, GFP_USER);
 637         if (st != BP_DONE)
 638                 return -ENOMEM;
 639
 640         return 0;
 641 }
 642
 643 /**
 644  * alloc_xenballooned_pages - get pages that have been ballooned out
 645  * @nr_pages: Number of pages to get
 646  * @pages: pages returned
 647  * @return 0 on success, error otherwise
 648  */
 649 int alloc_xenballooned_pages(int nr_pages, struct page **pages)
 650 {
 651         int pgno = 0;
 652         struct page *page;
 653         int ret;
 654
 655         mutex_lock(&balloon_mutex);
 656
 657         balloon_stats.target_unpopulated += nr_pages;
 658
 659         while (pgno < nr_pages) {
 660                 page = balloon_retrieve(true);
 661                 if (page) {
 662                         pages[pgno++] = page;
 663 #ifdef CONFIG_XEN_HAVE_PVMMU
 664                         /*
 665                          * We don't support PV MMU when Linux and Xen is using
 666                          * different page granularity.
 667                          */
 668                         BUILD_BUG_ON(XEN_PAGE_SIZE != PAGE_SIZE);
 669
 670                         ret = xen_alloc_p2m_entry(page_to_pfn(page));
 671                         if (ret < 0)
 672                                 goto out_undo;
 673 #endif
 674                 } else {
 675                         ret = add_ballooned_pages(nr_pages - pgno);
 676                         if (ret < 0)
 677                                 goto out_undo;
 678                 }
 679         }
 680         mutex_unlock(&balloon_mutex);
 681         return 0;
 682  out_undo:
 683         mutex_unlock(&balloon_mutex);
 684         free_xenballooned_pages(pgno, pages);
 685         return ret;
 686 }
 687 EXPORT_SYMBOL(alloc_xenballooned_pages);
 688
 689 /**
 690  * free_xenballooned_pages - return pages retrieved with get_ballooned_pages
 691  * @nr_pages: Number of pages
 692  * @pages: pages to return
 693  */
 694 void free_xenballooned_pages(int nr_pages, struct page **pages)
 695 {
 696         int i;
 697
 698         mutex_lock(&balloon_mutex);
 699
 700         for (i = 0; i < nr_pages; i++) {
 701                 if (pages[i])
 702                         balloon_append(pages[i]);
 703         }
 704
 705         balloon_stats.target_unpopulated -= nr_pages;
 706
 707         /* The balloon may be too large now. Shrink it if needed. */
 708         if (current_credit())
 709                 schedule_delayed_work(&balloon_worker, 0);
 710
 711         mutex_unlock(&balloon_mutex);
 712 }
 713 EXPORT_SYMBOL(free_xenballooned_pages);
 714
 715 static void __init balloon_add_region(unsigned long start_pfn,
 716                                       unsigned long pages)
 717 {
 718         unsigned long pfn, extra_pfn_end;
 719         struct page *page;
 720
 721         /*
 722          * If the amount of usable memory has been limited (e.g., with
 723          * the 'mem' command line parameter), don't add pages beyond
 724          * this limit.
 725          */
 726         extra_pfn_end = min(max_pfn, start_pfn + pages);
 727
 728         for (pfn = start_pfn; pfn < extra_pfn_end; pfn++) {
 729                 page = pfn_to_page(pfn);
 730                 /* totalram_pages and totalhigh_pages do not
 731                    include the boot-time balloon extension, so
 732                    don't subtract from it. */
 733                 __balloon_append(page);
 734         }
 735
 736         balloon_stats.total_pages += extra_pfn_end - start_pfn;
 737 }
 738
 739 static int __init balloon_init(void)
 740 {
 741         int i;
 742
 743         if (!xen_domain())
 744                 return -ENODEV;
 745
 746         pr_info("Initialising balloon driver\n");
 747
 748         balloon_stats.current_pages = xen_pv_domain()
 749                 ? min(xen_start_info->nr_pages - xen_released_pages, max_pfn)
 750                 : get_num_physpages();
 751         balloon_stats.target_pages  = balloon_stats.current_pages;
 752         balloon_stats.balloon_low   = 0;
 753         balloon_stats.balloon_high  = 0;
 754         balloon_stats.total_pages   = balloon_stats.current_pages;
 755
 756         balloon_stats.schedule_delay = 1;
 757         balloon_stats.max_schedule_delay = 32;
 758         balloon_stats.retry_count = 1;
 759         balloon_stats.max_retry_count = RETRY_UNLIMITED;
 760
 761 #ifdef CONFIG_XEN_BALLOON_MEMORY_HOTPLUG
 762         set_online_page_callback(&xen_online_page);
 763         register_memory_notifier(&xen_memory_nb);
 764         register_sysctl_table(xen_root);
 765 #endif
 766
 767         /*
 768          * Initialize the balloon with pages from the extra memory
 769          * regions (see arch/x86/xen/setup.c).
 770          */
 771         for (i = 0; i < XEN_EXTRA_MEM_MAX_REGIONS; i++)
 772                 if (xen_extra_mem[i].n_pfns)
 773                         balloon_add_region(xen_extra_mem[i].start_pfn,
 774                                            xen_extra_mem[i].n_pfns);
 775
 776         return 0;
 777 }
 778 subsys_initcall(balloon_init);