fs/ceph/addr.c

   1 #include <linux/ceph/ceph_debug.h>
   2
   3 #include <linux/backing-dev.h>
   4 #include <linux/fs.h>
   5 #include <linux/mm.h>
   6 #include <linux/pagemap.h>
   7 #include <linux/writeback.h>    /* generic_writepages */
   8 #include <linux/slab.h>
   9 #include <linux/pagevec.h>
  10 #include <linux/task_io_accounting_ops.h>
  11
  12 #include "super.h"
  13 #include "mds_client.h"
  14 #include "cache.h"
  15 #include <linux/ceph/osd_client.h>
  16
  17 /*
  18  * Ceph address space ops.
  19  *
  20  * There are a few funny things going on here.
  21  *
  22  * The page->private field is used to reference a struct
  23  * ceph_snap_context for _every_ dirty page.  This indicates which
  24  * snapshot the page was logically dirtied in, and thus which snap
  25  * context needs to be associated with the osd write during writeback.
  26  *
  27  * Similarly, struct ceph_inode_info maintains a set of counters to
  28  * count dirty pages on the inode.  In the absence of snapshots,
  29  * i_wrbuffer_ref == i_wrbuffer_ref_head == the dirty page count.
  30  *
  31  * When a snapshot is taken (that is, when the client receives
  32  * notification that a snapshot was taken), each inode with caps and
  33  * with dirty pages (dirty pages implies there is a cap) gets a new
  34  * ceph_cap_snap in the i_cap_snaps list (which is sorted in ascending
  35  * order, new snaps go to the tail).  The i_wrbuffer_ref_head count is
  36  * moved to capsnap->dirty. (Unless a sync write is currently in
  37  * progress.  In that case, the capsnap is said to be "pending", new
  38  * writes cannot start, and the capsnap isn't "finalized" until the
  39  * write completes (or fails) and a final size/mtime for the inode for
  40  * that snap can be settled upon.)  i_wrbuffer_ref_head is reset to 0.
  41  *
  42  * On writeback, we must submit writes to the osd IN SNAP ORDER.  So,
  43  * we look for the first capsnap in i_cap_snaps and write out pages in
  44  * that snap context _only_.  Then we move on to the next capsnap,
  45  * eventually reaching the "live" or "head" context (i.e., pages that
  46  * are not yet snapped) and are writing the most recently dirtied
  47  * pages.
  48  *
  49  * Invalidate and so forth must take care to ensure the dirty page
  50  * accounting is preserved.
  51  */
  52
  53 #define CONGESTION_ON_THRESH(congestion_kb) (congestion_kb >> (PAGE_SHIFT-10))
  54 #define CONGESTION_OFF_THRESH(congestion_kb)                            \
  55         (CONGESTION_ON_THRESH(congestion_kb) -                          \
  56          (CONGESTION_ON_THRESH(congestion_kb) >> 2))
  57
  58 static inline struct ceph_snap_context *page_snap_context(struct page *page)
  59 {
  60         if (PagePrivate(page))
  61                 return (void *)page->private;
  62         return NULL;
  63 }
  64
  65 /*
  66  * Dirty a page.  Optimistically adjust accounting, on the assumption
  67  * that we won't race with invalidate.  If we do, readjust.
  68  */
  69 static int ceph_set_page_dirty(struct page *page)
  70 {
  71         struct address_space *mapping = page->mapping;
  72         struct inode *inode;
  73         struct ceph_inode_info *ci;
  74         struct ceph_snap_context *snapc;
  75         int ret;
  76
  77         if (unlikely(!mapping))
  78                 return !TestSetPageDirty(page);
  79
  80         if (PageDirty(page)) {
  81                 dout("%p set_page_dirty %p idx %lu -- already dirty\n",
  82                      mapping->host, page, page->index);
  83                 BUG_ON(!PagePrivate(page));
  84                 return 0;
  85         }
  86
  87         inode = mapping->host;
  88         ci = ceph_inode(inode);
  89
  90         /* dirty the head */
  91         spin_lock(&ci->i_ceph_lock);
  92         BUG_ON(ci->i_wr_ref == 0); // caller should hold Fw reference
  93         if (__ceph_have_pending_cap_snap(ci)) {
  94                 struct ceph_cap_snap *capsnap =
  95                                 list_last_entry(&ci->i_cap_snaps,
  96                                                 struct ceph_cap_snap,
  97                                                 ci_item);
  98                 snapc = ceph_get_snap_context(capsnap->context);
  99                 capsnap->dirty_pages++;
 100         } else {
 101                 BUG_ON(!ci->i_head_snapc);
 102                 snapc = ceph_get_snap_context(ci->i_head_snapc);
 103                 ++ci->i_wrbuffer_ref_head;
 104         }
 105         if (ci->i_wrbuffer_ref == 0)
 106                 ihold(inode);
 107         ++ci->i_wrbuffer_ref;
 108         dout("%p set_page_dirty %p idx %lu head %d/%d -> %d/%d "
 109              "snapc %p seq %lld (%d snaps)\n",
 110              mapping->host, page, page->index,
 111              ci->i_wrbuffer_ref-1, ci->i_wrbuffer_ref_head-1,
 112              ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head,
 113              snapc, snapc->seq, snapc->num_snaps);
 114         spin_unlock(&ci->i_ceph_lock);
 115
 116         /*
 117          * Reference snap context in page->private.  Also set
 118          * PagePrivate so that we get invalidatepage callback.
 119          */
 120         BUG_ON(PagePrivate(page));
 121         page->private = (unsigned long)snapc;
 122         SetPagePrivate(page);
 123
 124         ret = __set_page_dirty_nobuffers(page);
 125         WARN_ON(!PageLocked(page));
 126         WARN_ON(!page->mapping);
 127
 128         return ret;
 129 }
 130
 131 /*
 132  * If we are truncating the full page (i.e. offset == 0), adjust the
 133  * dirty page counters appropriately.  Only called if there is private
 134  * data on the page.
 135  */
 136 static void ceph_invalidatepage(struct page *page, unsigned int offset,
 137                                 unsigned int length)
 138 {
 139         struct inode *inode;
 140         struct ceph_inode_info *ci;
 141         struct ceph_snap_context *snapc = page_snap_context(page);
 142
 143         inode = page->mapping->host;
 144         ci = ceph_inode(inode);
 145
 146         if (offset != 0 || length != PAGE_SIZE) {
 147                 dout("%p invalidatepage %p idx %lu partial dirty page %u~%u\n",
 148                      inode, page, page->index, offset, length);
 149                 return;
 150         }
 151
 152         ceph_invalidate_fscache_page(inode, page);
 153
 154         if (!PagePrivate(page))
 155                 return;
 156
 157         /*
 158          * We can get non-dirty pages here due to races between
 159          * set_page_dirty and truncate_complete_page; just spit out a
 160          * warning, in case we end up with accounting problems later.
 161          */
 162         if (!PageDirty(page))
 163                 pr_err("%p invalidatepage %p page not dirty\n", inode, page);
 164
 165         ClearPageChecked(page);
 166
 167         dout("%p invalidatepage %p idx %lu full dirty page\n",
 168              inode, page, page->index);
 169
 170         ceph_put_wrbuffer_cap_refs(ci, 1, snapc);
 171         ceph_put_snap_context(snapc);
 172         page->private = 0;
 173         ClearPagePrivate(page);
 174 }
 175
 176 static int ceph_releasepage(struct page *page, gfp_t g)
 177 {
 178         dout("%p releasepage %p idx %lu\n", page->mapping->host,
 179              page, page->index);
 180         WARN_ON(PageDirty(page));
 181
 182         /* Can we release the page from the cache? */
 183         if (!ceph_release_fscache_page(page, g))
 184                 return 0;
 185
 186         return !PagePrivate(page);
 187 }
 188
 189 /*
 190  * read a single page, without unlocking it.
 191  */
 192 static int readpage_nounlock(struct file *filp, struct page *page)
 193 {
 194         struct inode *inode = file_inode(filp);
 195         struct ceph_inode_info *ci = ceph_inode(inode);
 196         struct ceph_osd_client *osdc =
 197                 &ceph_inode_to_client(inode)->client->osdc;
 198         int err = 0;
 199         u64 off = page_offset(page);
 200         u64 len = PAGE_SIZE;
 201
 202         if (off >= i_size_read(inode)) {
 203                 zero_user_segment(page, 0, PAGE_SIZE);
 204                 SetPageUptodate(page);
 205                 return 0;
 206         }
 207
 208         if (ci->i_inline_version != CEPH_INLINE_NONE) {
 209                 /*
 210                  * Uptodate inline data should have been added
 211                  * into page cache while getting Fcr caps.
 212                  */
 213                 if (off == 0)
 214                         return -EINVAL;
 215                 zero_user_segment(page, 0, PAGE_SIZE);
 216                 SetPageUptodate(page);
 217                 return 0;
 218         }
 219
 220         err = ceph_readpage_from_fscache(inode, page);
 221         if (err == 0)
 222                 goto out;
 223
 224         dout("readpage inode %p file %p page %p index %lu\n",
 225              inode, filp, page, page->index);
 226         err = ceph_osdc_readpages(osdc, ceph_vino(inode), &ci->i_layout,
 227                                   off, &len,
 228                                   ci->i_truncate_seq, ci->i_truncate_size,
 229                                   &page, 1, 0);
 230         if (err == -ENOENT)
 231                 err = 0;
 232         if (err < 0) {
 233                 SetPageError(page);
 234                 ceph_fscache_readpage_cancel(inode, page);
 235                 goto out;
 236         }
 237         if (err < PAGE_SIZE)
 238                 /* zero fill remainder of page */
 239                 zero_user_segment(page, err, PAGE_SIZE);
 240         else
 241                 flush_dcache_page(page);
 242
 243         SetPageUptodate(page);
 244         ceph_readpage_to_fscache(inode, page);
 245
 246 out:
 247         return err < 0 ? err : 0;
 248 }
 249
 250 static int ceph_readpage(struct file *filp, struct page *page)
 251 {
 252         int r = readpage_nounlock(filp, page);
 253         unlock_page(page);
 254         return r;
 255 }
 256
 257 /*
 258  * Finish an async read(ahead) op.
 259  */
 260 static void finish_read(struct ceph_osd_request *req)
 261 {
 262         struct inode *inode = req->r_inode;
 263         struct ceph_osd_data *osd_data;
 264         int rc = req->r_result <= 0 ? req->r_result : 0;
 265         int bytes = req->r_result >= 0 ? req->r_result : 0;
 266         int num_pages;
 267         int i;
 268
 269         dout("finish_read %p req %p rc %d bytes %d\n", inode, req, rc, bytes);
 270
 271         /* unlock all pages, zeroing any data we didn't read */
 272         osd_data = osd_req_op_extent_osd_data(req, 0);
 273         BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGES);
 274         num_pages = calc_pages_for((u64)osd_data->alignment,
 275                                         (u64)osd_data->length);
 276         for (i = 0; i < num_pages; i++) {
 277                 struct page *page = osd_data->pages[i];
 278
 279                 if (rc < 0 && rc != -ENOENT) {
 280                         ceph_fscache_readpage_cancel(inode, page);
 281                         goto unlock;
 282                 }
 283                 if (bytes < (int)PAGE_SIZE) {
 284                         /* zero (remainder of) page */
 285                         int s = bytes < 0 ? 0 : bytes;
 286                         zero_user_segment(page, s, PAGE_SIZE);
 287                 }
 288                 dout("finish_read %p uptodate %p idx %lu\n", inode, page,
 289                      page->index);
 290                 flush_dcache_page(page);
 291                 SetPageUptodate(page);
 292                 ceph_readpage_to_fscache(inode, page);
 293 unlock:
 294                 unlock_page(page);
 295                 put_page(page);
 296                 bytes -= PAGE_SIZE;
 297         }
 298         kfree(osd_data->pages);
 299 }
 300
 301 static void ceph_unlock_page_vector(struct page **pages, int num_pages)
 302 {
 303         int i;
 304
 305         for (i = 0; i < num_pages; i++)
 306                 unlock_page(pages[i]);
 307 }
 308
 309 /*
 310  * start an async read(ahead) operation.  return nr_pages we submitted
 311  * a read for on success, or negative error code.
 312  */
 313 static int start_read(struct inode *inode, struct list_head *page_list, int max)
 314 {
 315         struct ceph_osd_client *osdc =
 316                 &ceph_inode_to_client(inode)->client->osdc;
 317         struct ceph_inode_info *ci = ceph_inode(inode);
 318         struct page *page = list_entry(page_list->prev, struct page, lru);
 319         struct ceph_vino vino;
 320         struct ceph_osd_request *req;
 321         u64 off;
 322         u64 len;
 323         int i;
 324         struct page **pages;
 325         pgoff_t next_index;
 326         int nr_pages = 0;
 327         int ret;
 328
 329         off = (u64) page_offset(page);
 330
 331         /* count pages */
 332         next_index = page->index;
 333         list_for_each_entry_reverse(page, page_list, lru) {
 334                 if (page->index != next_index)
 335                         break;
 336                 nr_pages++;
 337                 next_index++;
 338                 if (max && nr_pages == max)
 339                         break;
 340         }
 341         len = nr_pages << PAGE_SHIFT;
 342         dout("start_read %p nr_pages %d is %lld~%lld\n", inode, nr_pages,
 343              off, len);
 344         vino = ceph_vino(inode);
 345         req = ceph_osdc_new_request(osdc, &ci->i_layout, vino, off, &len,
 346                                     0, 1, CEPH_OSD_OP_READ,
 347                                     CEPH_OSD_FLAG_READ, NULL,
 348                                     ci->i_truncate_seq, ci->i_truncate_size,
 349                                     false);
 350         if (IS_ERR(req))
 351                 return PTR_ERR(req);
 352
 353         /* build page vector */
 354         nr_pages = calc_pages_for(0, len);
 355         pages = kmalloc(sizeof(*pages) * nr_pages, GFP_KERNEL);
 356         ret = -ENOMEM;
 357         if (!pages)
 358                 goto out;
 359         for (i = 0; i < nr_pages; ++i) {
 360                 page = list_entry(page_list->prev, struct page, lru);
 361                 BUG_ON(PageLocked(page));
 362                 list_del(&page->lru);
 363
 364                 dout("start_read %p adding %p idx %lu\n", inode, page,
 365                      page->index);
 366                 if (add_to_page_cache_lru(page, &inode->i_data, page->index,
 367                                           GFP_KERNEL)) {
 368                         ceph_fscache_uncache_page(inode, page);
 369                         put_page(page);
 370                         dout("start_read %p add_to_page_cache failed %p\n",
 371                              inode, page);
 372                         nr_pages = i;
 373                         goto out_pages;
 374                 }
 375                 pages[i] = page;
 376         }
 377         osd_req_op_extent_osd_data_pages(req, 0, pages, len, 0, false, false);
 378         req->r_callback = finish_read;
 379         req->r_inode = inode;
 380
 381         dout("start_read %p starting %p %lld~%lld\n", inode, req, off, len);
 382         ret = ceph_osdc_start_request(osdc, req, false);
 383         if (ret < 0)
 384                 goto out_pages;
 385         ceph_osdc_put_request(req);
 386         return nr_pages;
 387
 388 out_pages:
 389         ceph_unlock_page_vector(pages, nr_pages);
 390         ceph_release_page_vector(pages, nr_pages);
 391 out:
 392         ceph_osdc_put_request(req);
 393         return ret;
 394 }
 395
 396
 397 /*
 398  * Read multiple pages.  Leave pages we don't read + unlock in page_list;
 399  * the caller (VM) cleans them up.
 400  */
 401 static int ceph_readpages(struct file *file, struct address_space *mapping,
 402                           struct list_head *page_list, unsigned nr_pages)
 403 {
 404         struct inode *inode = file_inode(file);
 405         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 406         int rc = 0;
 407         int max = 0;
 408
 409         if (ceph_inode(inode)->i_inline_version != CEPH_INLINE_NONE)
 410                 return -EINVAL;
 411
 412         rc = ceph_readpages_from_fscache(mapping->host, mapping, page_list,
 413                                          &nr_pages);
 414
 415         if (rc == 0)
 416                 goto out;
 417
 418         if (fsc->mount_options->rsize >= PAGE_SIZE)
 419                 max = (fsc->mount_options->rsize + PAGE_SIZE - 1)
 420                         >> PAGE_SHIFT;
 421
 422         dout("readpages %p file %p nr_pages %d max %d\n", inode,
 423                 file, nr_pages,
 424              max);
 425         while (!list_empty(page_list)) {
 426                 rc = start_read(inode, page_list, max);
 427                 if (rc < 0)
 428                         goto out;
 429                 BUG_ON(rc == 0);
 430         }
 431 out:
 432         ceph_fscache_readpages_cancel(inode, page_list);
 433
 434         dout("readpages %p file %p ret %d\n", inode, file, rc);
 435         return rc;
 436 }
 437
 438 /*
 439  * Get ref for the oldest snapc for an inode with dirty data... that is, the
 440  * only snap context we are allowed to write back.
 441  */
 442 static struct ceph_snap_context *get_oldest_context(struct inode *inode,
 443                                                     loff_t *snap_size)
 444 {
 445         struct ceph_inode_info *ci = ceph_inode(inode);
 446         struct ceph_snap_context *snapc = NULL;
 447         struct ceph_cap_snap *capsnap = NULL;
 448
 449         spin_lock(&ci->i_ceph_lock);
 450         list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) {
 451                 dout(" cap_snap %p snapc %p has %d dirty pages\n", capsnap,
 452                      capsnap->context, capsnap->dirty_pages);
 453                 if (capsnap->dirty_pages) {
 454                         snapc = ceph_get_snap_context(capsnap->context);
 455                         if (snap_size)
 456                                 *snap_size = capsnap->size;
 457                         break;
 458                 }
 459         }
 460         if (!snapc && ci->i_wrbuffer_ref_head) {
 461                 snapc = ceph_get_snap_context(ci->i_head_snapc);
 462                 dout(" head snapc %p has %d dirty pages\n",
 463                      snapc, ci->i_wrbuffer_ref_head);
 464         }
 465         spin_unlock(&ci->i_ceph_lock);
 466         return snapc;
 467 }
 468
 469 /*
 470  * Write a single page, but leave the page locked.
 471  *
 472  * If we get a write error, set the page error bit, but still adjust the
 473  * dirty page accounting (i.e., page is no longer dirty).
 474  */
 475 static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 476 {
 477         struct inode *inode;
 478         struct ceph_inode_info *ci;
 479         struct ceph_fs_client *fsc;
 480         struct ceph_osd_client *osdc;
 481         struct ceph_snap_context *snapc, *oldest;
 482         loff_t page_off = page_offset(page);
 483         loff_t snap_size = -1;
 484         long writeback_stat;
 485         u64 truncate_size;
 486         u32 truncate_seq;
 487         int err = 0, len = PAGE_SIZE;
 488
 489         dout("writepage %p idx %lu\n", page, page->index);
 490
 491         if (!page->mapping || !page->mapping->host) {
 492                 dout("writepage %p - no mapping\n", page);
 493                 return -EFAULT;
 494         }
 495         inode = page->mapping->host;
 496         ci = ceph_inode(inode);
 497         fsc = ceph_inode_to_client(inode);
 498         osdc = &fsc->client->osdc;
 499
 500         /* verify this is a writeable snap context */
 501         snapc = page_snap_context(page);
 502         if (snapc == NULL) {
 503                 dout("writepage %p page %p not dirty?\n", inode, page);
 504                 goto out;
 505         }
 506         oldest = get_oldest_context(inode, &snap_size);
 507         if (snapc->seq > oldest->seq) {
 508                 dout("writepage %p page %p snapc %p not writeable - noop\n",
 509                      inode, page, snapc);
 510                 /* we should only noop if called by kswapd */
 511                 WARN_ON((current->flags & PF_MEMALLOC) == 0);
 512                 ceph_put_snap_context(oldest);
 513                 goto out;
 514         }
 515         ceph_put_snap_context(oldest);
 516
 517         spin_lock(&ci->i_ceph_lock);
 518         truncate_seq = ci->i_truncate_seq;
 519         truncate_size = ci->i_truncate_size;
 520         if (snap_size == -1)
 521                 snap_size = i_size_read(inode);
 522         spin_unlock(&ci->i_ceph_lock);
 523
 524         /* is this a partial page at end of file? */
 525         if (page_off >= snap_size) {
 526                 dout("%p page eof %llu\n", page, snap_size);
 527                 goto out;
 528         }
 529         if (snap_size < page_off + len)
 530                 len = snap_size - page_off;
 531
 532         dout("writepage %p page %p index %lu on %llu~%u snapc %p\n",
 533              inode, page, page->index, page_off, len, snapc);
 534
 535         writeback_stat = atomic_long_inc_return(&fsc->writeback_count);
 536         if (writeback_stat >
 537             CONGESTION_ON_THRESH(fsc->mount_options->congestion_kb))
 538                 set_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC);
 539
 540         set_page_writeback(page);
 541         err = ceph_osdc_writepages(osdc, ceph_vino(inode),
 542                                    &ci->i_layout, snapc,
 543                                    page_off, len,
 544                                    truncate_seq, truncate_size,
 545                                    &inode->i_mtime, &page, 1);
 546         if (err < 0) {
 547                 struct writeback_control tmp_wbc;
 548                 if (!wbc)
 549                         wbc = &tmp_wbc;
 550                 if (err == -ERESTARTSYS) {
 551                         /* killed by SIGKILL */
 552                         dout("writepage interrupted page %p\n", page);
 553                         redirty_page_for_writepage(wbc, page);
 554                         end_page_writeback(page);
 555                         goto out;
 556                 }
 557                 dout("writepage setting page/mapping error %d %p\n",
 558                      err, page);
 559                 SetPageError(page);
 560                 mapping_set_error(&inode->i_data, err);
 561                 wbc->pages_skipped++;
 562         } else {
 563                 dout("writepage cleaned page %p\n", page);
 564                 err = 0;  /* vfs expects us to return 0 */
 565         }
 566         page->private = 0;
 567         ClearPagePrivate(page);
 568         end_page_writeback(page);
 569         ceph_put_wrbuffer_cap_refs(ci, 1, snapc);
 570         ceph_put_snap_context(snapc);  /* page's reference */
 571 out:
 572         return err;
 573 }
 574
 575 static int ceph_writepage(struct page *page, struct writeback_control *wbc)
 576 {
 577         int err;
 578         struct inode *inode = page->mapping->host;
 579         BUG_ON(!inode);
 580         ihold(inode);
 581         err = writepage_nounlock(page, wbc);
 582         if (err == -ERESTARTSYS) {
 583                 /* direct memory reclaimer was killed by SIGKILL. return 0
 584                  * to prevent caller from setting mapping/page error */
 585                 err = 0;
 586         }
 587         unlock_page(page);
 588         iput(inode);
 589         return err;
 590 }
 591
 592 /*
 593  * lame release_pages helper.  release_pages() isn't exported to
 594  * modules.
 595  */
 596 static void ceph_release_pages(struct page **pages, int num)
 597 {
 598         struct pagevec pvec;
 599         int i;
 600
 601         pagevec_init(&pvec, 0);
 602         for (i = 0; i < num; i++) {
 603                 if (pagevec_add(&pvec, pages[i]) == 0)
 604                         pagevec_release(&pvec);
 605         }
 606         pagevec_release(&pvec);
 607 }
 608
 609 /*
 610  * async writeback completion handler.
 611  *
 612  * If we get an error, set the mapping error bit, but not the individual
 613  * page error bits.
 614  */
 615 static void writepages_finish(struct ceph_osd_request *req)
 616 {
 617         struct inode *inode = req->r_inode;
 618         struct ceph_inode_info *ci = ceph_inode(inode);
 619         struct ceph_osd_data *osd_data;
 620         struct page *page;
 621         int num_pages, total_pages = 0;
 622         int i, j;
 623         int rc = req->r_result;
 624         struct ceph_snap_context *snapc = req->r_snapc;
 625         struct address_space *mapping = inode->i_mapping;
 626         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 627         bool remove_page;
 628
 629         dout("writepages_finish %p rc %d\n", inode, rc);
 630         if (rc < 0)
 631                 mapping_set_error(mapping, rc);
 632
 633         /*
 634          * We lost the cache cap, need to truncate the page before
 635          * it is unlocked, otherwise we'd truncate it later in the
 636          * page truncation thread, possibly losing some data that
 637          * raced its way in
 638          */
 639         remove_page = !(ceph_caps_issued(ci) &
 640                         (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO));
 641
 642         /* clean all pages */
 643         for (i = 0; i < req->r_num_ops; i++) {
 644                 if (req->r_ops[i].op != CEPH_OSD_OP_WRITE)
 645                         break;
 646
 647                 osd_data = osd_req_op_extent_osd_data(req, i);
 648                 BUG_ON(osd_data->type != CEPH_OSD_DATA_TYPE_PAGES);
 649                 num_pages = calc_pages_for((u64)osd_data->alignment,
 650                                            (u64)osd_data->length);
 651                 total_pages += num_pages;
 652                 for (j = 0; j < num_pages; j++) {
 653                         page = osd_data->pages[j];
 654                         BUG_ON(!page);
 655                         WARN_ON(!PageUptodate(page));
 656
 657                         if (atomic_long_dec_return(&fsc->writeback_count) <
 658                              CONGESTION_OFF_THRESH(
 659                                         fsc->mount_options->congestion_kb))
 660                                 clear_bdi_congested(&fsc->backing_dev_info,
 661                                                     BLK_RW_ASYNC);
 662
 663                         if (rc < 0)
 664                                 SetPageError(page);
 665
 666                         ceph_put_snap_context(page_snap_context(page));
 667                         page->private = 0;
 668                         ClearPagePrivate(page);
 669                         dout("unlocking %p\n", page);
 670                         end_page_writeback(page);
 671
 672                         if (remove_page)
 673                                 generic_error_remove_page(inode->i_mapping,
 674                                                           page);
 675
 676                         unlock_page(page);
 677                 }
 678                 dout("writepages_finish %p wrote %llu bytes cleaned %d pages\n",
 679                      inode, osd_data->length, rc >= 0 ? num_pages : 0);
 680
 681                 ceph_release_pages(osd_data->pages, num_pages);
 682         }
 683
 684         ceph_put_wrbuffer_cap_refs(ci, total_pages, snapc);
 685
 686         osd_data = osd_req_op_extent_osd_data(req, 0);
 687         if (osd_data->pages_from_pool)
 688                 mempool_free(osd_data->pages,
 689                              ceph_sb_to_client(inode->i_sb)->wb_pagevec_pool);
 690         else
 691                 kfree(osd_data->pages);
 692         ceph_osdc_put_request(req);
 693 }
 694
 695 /*
 696  * initiate async writeback
 697  */
 698 static int ceph_writepages_start(struct address_space *mapping,
 699                                  struct writeback_control *wbc)
 700 {
 701         struct inode *inode = mapping->host;
 702         struct ceph_inode_info *ci = ceph_inode(inode);
 703         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 704         struct ceph_vino vino = ceph_vino(inode);
 705         pgoff_t index, start, end;
 706         int range_whole = 0;
 707         int should_loop = 1;
 708         pgoff_t max_pages = 0, max_pages_ever = 0;
 709         struct ceph_snap_context *snapc = NULL, *last_snapc = NULL, *pgsnapc;
 710         struct pagevec pvec;
 711         int done = 0;
 712         int rc = 0;
 713         unsigned wsize = 1 << inode->i_blkbits;
 714         struct ceph_osd_request *req = NULL;
 715         int do_sync = 0;
 716         loff_t snap_size, i_size;
 717         u64 truncate_size;
 718         u32 truncate_seq;
 719
 720         /*
 721          * Include a 'sync' in the OSD request if this is a data
 722          * integrity write (e.g., O_SYNC write or fsync()), or if our
 723          * cap is being revoked.
 724          */
 725         if ((wbc->sync_mode == WB_SYNC_ALL) ||
 726                 ceph_caps_revoking(ci, CEPH_CAP_FILE_BUFFER))
 727                 do_sync = 1;
 728         dout("writepages_start %p dosync=%d (mode=%s)\n",
 729              inode, do_sync,
 730              wbc->sync_mode == WB_SYNC_NONE ? "NONE" :
 731              (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD"));
 732
 733         if (ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
 734                 if (ci->i_wrbuffer_ref > 0) {
 735                         pr_warn_ratelimited(
 736                                 "writepage_start %p %lld forced umount\n",
 737                                 inode, ceph_ino(inode));
 738                 }
 739                 mapping_set_error(mapping, -EIO);
 740                 return -EIO; /* we're in a forced umount, don't write! */
 741         }
 742         if (fsc->mount_options->wsize && fsc->mount_options->wsize < wsize)
 743                 wsize = fsc->mount_options->wsize;
 744         if (wsize < PAGE_SIZE)
 745                 wsize = PAGE_SIZE;
 746         max_pages_ever = wsize >> PAGE_SHIFT;
 747
 748         pagevec_init(&pvec, 0);
 749
 750         /* where to start/end? */
 751         if (wbc->range_cyclic) {
 752                 start = mapping->writeback_index; /* Start from prev offset */
 753                 end = -1;
 754                 dout(" cyclic, start at %lu\n", start);
 755         } else {
 756                 start = wbc->range_start >> PAGE_SHIFT;
 757                 end = wbc->range_end >> PAGE_SHIFT;
 758                 if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
 759                         range_whole = 1;
 760                 should_loop = 0;
 761                 dout(" not cyclic, %lu to %lu\n", start, end);
 762         }
 763         index = start;
 764
 765 retry:
 766         /* find oldest snap context with dirty data */
 767         ceph_put_snap_context(snapc);
 768         snap_size = -1;
 769         snapc = get_oldest_context(inode, &snap_size);
 770         if (!snapc) {
 771                 /* hmm, why does writepages get called when there
 772                    is no dirty data? */
 773                 dout(" no snap context with dirty data?\n");
 774                 goto out;
 775         }
 776         dout(" oldest snapc is %p seq %lld (%d snaps)\n",
 777              snapc, snapc->seq, snapc->num_snaps);
 778
 779         spin_lock(&ci->i_ceph_lock);
 780         truncate_seq = ci->i_truncate_seq;
 781         truncate_size = ci->i_truncate_size;
 782         i_size = i_size_read(inode);
 783         spin_unlock(&ci->i_ceph_lock);
 784
 785         if (last_snapc && snapc != last_snapc) {
 786                 /* if we switched to a newer snapc, restart our scan at the
 787                  * start of the original file range. */
 788                 dout("  snapc differs from last pass, restarting at %lu\n",
 789                      index);
 790                 index = start;
 791         }
 792         last_snapc = snapc;
 793
 794         while (!done && index <= end) {
 795                 unsigned i;
 796                 int first;
 797                 pgoff_t strip_unit_end = 0;
 798                 int num_ops = 0, op_idx;
 799                 int pvec_pages, locked_pages = 0;
 800                 struct page **pages = NULL, **data_pages;
 801                 mempool_t *pool = NULL; /* Becomes non-null if mempool used */
 802                 struct page *page;
 803                 int want;
 804                 u64 offset = 0, len = 0;
 805
 806                 max_pages = max_pages_ever;
 807
 808 get_more_pages:
 809                 first = -1;
 810                 want = min(end - index,
 811                            min((pgoff_t)PAGEVEC_SIZE,
 812                                max_pages - (pgoff_t)locked_pages) - 1)
 813                         + 1;
 814                 pvec_pages = pagevec_lookup_tag(&pvec, mapping, &index,
 815                                                 PAGECACHE_TAG_DIRTY,
 816                                                 want);
 817                 dout("pagevec_lookup_tag got %d\n", pvec_pages);
 818                 if (!pvec_pages && !locked_pages)
 819                         break;
 820                 for (i = 0; i < pvec_pages && locked_pages < max_pages; i++) {
 821                         page = pvec.pages[i];
 822                         dout("? %p idx %lu\n", page, page->index);
 823                         if (locked_pages == 0)
 824                                 lock_page(page);  /* first page */
 825                         else if (!trylock_page(page))
 826                                 break;
 827
 828                         /* only dirty pages, or our accounting breaks */
 829                         if (unlikely(!PageDirty(page)) ||
 830                             unlikely(page->mapping != mapping)) {
 831                                 dout("!dirty or !mapping %p\n", page);
 832                                 unlock_page(page);
 833                                 break;
 834                         }
 835                         if (!wbc->range_cyclic && page->index > end) {
 836                                 dout("end of range %p\n", page);
 837                                 done = 1;
 838                                 unlock_page(page);
 839                                 break;
 840                         }
 841                         if (strip_unit_end && (page->index > strip_unit_end)) {
 842                                 dout("end of strip unit %p\n", page);
 843                                 unlock_page(page);
 844                                 break;
 845                         }
 846                         if (wbc->sync_mode != WB_SYNC_NONE) {
 847                                 dout("waiting on writeback %p\n", page);
 848                                 wait_on_page_writeback(page);
 849                         }
 850                         if (page_offset(page) >=
 851                             (snap_size == -1 ? i_size : snap_size)) {
 852                                 dout("%p page eof %llu\n", page,
 853                                      (snap_size == -1 ? i_size : snap_size));
 854                                 done = 1;
 855                                 unlock_page(page);
 856                                 break;
 857                         }
 858                         if (PageWriteback(page)) {
 859                                 dout("%p under writeback\n", page);
 860                                 unlock_page(page);
 861                                 break;
 862                         }
 863
 864                         /* only if matching snap context */
 865                         pgsnapc = page_snap_context(page);
 866                         if (pgsnapc->seq > snapc->seq) {
 867                                 dout("page snapc %p %lld > oldest %p %lld\n",
 868                                      pgsnapc, pgsnapc->seq, snapc, snapc->seq);
 869                                 unlock_page(page);
 870                                 if (!locked_pages)
 871                                         continue; /* keep looking for snap */
 872                                 break;
 873                         }
 874
 875                         if (!clear_page_dirty_for_io(page)) {
 876                                 dout("%p !clear_page_dirty_for_io\n", page);
 877                                 unlock_page(page);
 878                                 break;
 879                         }
 880
 881                         /*
 882                          * We have something to write.  If this is
 883                          * the first locked page this time through,
 884                          * calculate max possinle write size and
 885                          * allocate a page array
 886                          */
 887                         if (locked_pages == 0) {
 888                                 u64 objnum;
 889                                 u64 objoff;
 890
 891                                 /* prepare async write request */
 892                                 offset = (u64)page_offset(page);
 893                                 len = wsize;
 894
 895                                 rc = ceph_calc_file_object_mapping(&ci->i_layout,
 896                                                                 offset, len,
 897                                                                 &objnum, &objoff,
 898                                                                 &len);
 899                                 if (rc < 0) {
 900                                         unlock_page(page);
 901                                         break;
 902                                 }
 903
 904                                 num_ops = 1 + do_sync;
 905                                 strip_unit_end = page->index +
 906                                         ((len - 1) >> PAGE_SHIFT);
 907
 908                                 BUG_ON(pages);
 909                                 max_pages = calc_pages_for(0, (u64)len);
 910                                 pages = kmalloc(max_pages * sizeof (*pages),
 911                                                 GFP_NOFS);
 912                                 if (!pages) {
 913                                         pool = fsc->wb_pagevec_pool;
 914                                         pages = mempool_alloc(pool, GFP_NOFS);
 915                                         BUG_ON(!pages);
 916                                 }
 917
 918                                 len = 0;
 919                         } else if (page->index !=
 920                                    (offset + len) >> PAGE_SHIFT) {
 921                                 if (num_ops >= (pool ?  CEPH_OSD_SLAB_OPS :
 922                                                         CEPH_OSD_MAX_OPS)) {
 923                                         redirty_page_for_writepage(wbc, page);
 924                                         unlock_page(page);
 925                                         break;
 926                                 }
 927
 928                                 num_ops++;
 929                                 offset = (u64)page_offset(page);
 930                                 len = 0;
 931                         }
 932
 933                         /* note position of first page in pvec */
 934                         if (first < 0)
 935                                 first = i;
 936                         dout("%p will write page %p idx %lu\n",
 937                              inode, page, page->index);
 938
 939                         if (atomic_long_inc_return(&fsc->writeback_count) >
 940                             CONGESTION_ON_THRESH(
 941                                     fsc->mount_options->congestion_kb)) {
 942                                 set_bdi_congested(&fsc->backing_dev_info,
 943                                                   BLK_RW_ASYNC);
 944                         }
 945
 946                         pages[locked_pages] = page;
 947                         locked_pages++;
 948                         len += PAGE_SIZE;
 949                 }
 950
 951                 /* did we get anything? */
 952                 if (!locked_pages)
 953                         goto release_pvec_pages;
 954                 if (i) {
 955                         int j;
 956                         BUG_ON(!locked_pages || first < 0);
 957
 958                         if (pvec_pages && i == pvec_pages &&
 959                             locked_pages < max_pages) {
 960                                 dout("reached end pvec, trying for more\n");
 961                                 pagevec_reinit(&pvec);
 962                                 goto get_more_pages;
 963                         }
 964
 965                         /* shift unused pages over in the pvec...  we
 966                          * will need to release them below. */
 967                         for (j = i; j < pvec_pages; j++) {
 968                                 dout(" pvec leftover page %p\n", pvec.pages[j]);
 969                                 pvec.pages[j-i+first] = pvec.pages[j];
 970                         }
 971                         pvec.nr -= i-first;
 972                 }
 973
 974 new_request:
 975                 offset = page_offset(pages[0]);
 976                 len = wsize;
 977
 978                 req = ceph_osdc_new_request(&fsc->client->osdc,
 979                                         &ci->i_layout, vino,
 980                                         offset, &len, 0, num_ops,
 981                                         CEPH_OSD_OP_WRITE,
 982                                         CEPH_OSD_FLAG_WRITE |
 983                                         CEPH_OSD_FLAG_ONDISK,
 984                                         snapc, truncate_seq,
 985                                         truncate_size, false);
 986                 if (IS_ERR(req)) {
 987                         req = ceph_osdc_new_request(&fsc->client->osdc,
 988                                                 &ci->i_layout, vino,
 989                                                 offset, &len, 0,
 990                                                 min(num_ops,
 991                                                     CEPH_OSD_SLAB_OPS),
 992                                                 CEPH_OSD_OP_WRITE,
 993                                                 CEPH_OSD_FLAG_WRITE |
 994                                                 CEPH_OSD_FLAG_ONDISK,
 995                                                 snapc, truncate_seq,
 996                                                 truncate_size, true);
 997                         BUG_ON(IS_ERR(req));
 998                 }
 999                 BUG_ON(len < page_offset(pages[locked_pages - 1]) +
1000                              PAGE_SIZE - offset);
1001
1002                 req->r_callback = writepages_finish;
1003                 req->r_inode = inode;
1004
1005                 /* Format the osd request message and submit the write */
1006                 len = 0;
1007                 data_pages = pages;
1008                 op_idx = 0;
1009                 for (i = 0; i < locked_pages; i++) {
1010                         u64 cur_offset = page_offset(pages[i]);
1011                         if (offset + len != cur_offset) {
1012                                 if (op_idx + do_sync + 1 == req->r_num_ops)
1013                                         break;
1014                                 osd_req_op_extent_dup_last(req, op_idx,
1015                                                            cur_offset - offset);
1016                                 dout("writepages got pages at %llu~%llu\n",
1017                                      offset, len);
1018                                 osd_req_op_extent_osd_data_pages(req, op_idx,
1019                                                         data_pages, len, 0,
1020                                                         !!pool, false);
1021                                 osd_req_op_extent_update(req, op_idx, len);
1022
1023                                 len = 0;
1024                                 offset = cur_offset;
1025                                 data_pages = pages + i;
1026                                 op_idx++;
1027                         }
1028
1029                         set_page_writeback(pages[i]);
1030                         len += PAGE_SIZE;
1031                 }
1032
1033                 if (snap_size != -1) {
1034                         len = min(len, snap_size - offset);
1035                 } else if (i == locked_pages) {
1036                         /* writepages_finish() clears writeback pages
1037                          * according to the data length, so make sure
1038                          * data length covers all locked pages */
1039                         u64 min_len = len + 1 - PAGE_SIZE;
1040                         len = min(len, (u64)i_size_read(inode) - offset);
1041                         len = max(len, min_len);
1042                 }
1043                 dout("writepages got pages at %llu~%llu\n", offset, len);
1044
1045                 osd_req_op_extent_osd_data_pages(req, op_idx, data_pages, len,
1046                                                  0, !!pool, false);
1047                 osd_req_op_extent_update(req, op_idx, len);
1048
1049                 if (do_sync) {
1050                         op_idx++;
1051                         osd_req_op_init(req, op_idx, CEPH_OSD_OP_STARTSYNC, 0);
1052                 }
1053                 BUG_ON(op_idx + 1 != req->r_num_ops);
1054
1055                 pool = NULL;
1056                 if (i < locked_pages) {
1057                         BUG_ON(num_ops <= req->r_num_ops);
1058                         num_ops -= req->r_num_ops;
1059                         num_ops += do_sync;
1060                         locked_pages -= i;
1061
1062                         /* allocate new pages array for next request */
1063                         data_pages = pages;
1064                         pages = kmalloc(locked_pages * sizeof (*pages),
1065                                         GFP_NOFS);
1066                         if (!pages) {
1067                                 pool = fsc->wb_pagevec_pool;
1068                                 pages = mempool_alloc(pool, GFP_NOFS);
1069                                 BUG_ON(!pages);
1070                         }
1071                         memcpy(pages, data_pages + i,
1072                                locked_pages * sizeof(*pages));
1073                         memset(data_pages + i, 0,
1074                                locked_pages * sizeof(*pages));
1075                 } else {
1076                         BUG_ON(num_ops != req->r_num_ops);
1077                         index = pages[i - 1]->index + 1;
1078                         /* request message now owns the pages array */
1079                         pages = NULL;
1080                 }
1081
1082                 req->r_mtime = inode->i_mtime;
1083                 rc = ceph_osdc_start_request(&fsc->client->osdc, req, true);
1084                 BUG_ON(rc);
1085                 req = NULL;
1086
1087                 wbc->nr_to_write -= i;
1088                 if (pages)
1089                         goto new_request;
1090
1091                 if (wbc->nr_to_write <= 0)
1092                         done = 1;
1093
1094 release_pvec_pages:
1095                 dout("pagevec_release on %d pages (%p)\n", (int)pvec.nr,
1096                      pvec.nr ? pvec.pages[0] : NULL);
1097                 pagevec_release(&pvec);
1098
1099                 if (locked_pages && !done)
1100                         goto retry;
1101         }
1102
1103         if (should_loop && !done) {
1104                 /* more to do; loop back to beginning of file */
1105                 dout("writepages looping back to beginning of file\n");
1106                 should_loop = 0;
1107                 index = 0;
1108                 goto retry;
1109         }
1110
1111         if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
1112                 mapping->writeback_index = index;
1113
1114 out:
1115         ceph_osdc_put_request(req);
1116         ceph_put_snap_context(snapc);
1117         dout("writepages done, rc = %d\n", rc);
1118         return rc;
1119 }
1120
1121
1122
1123 /*
1124  * See if a given @snapc is either writeable, or already written.
1125  */
1126 static int context_is_writeable_or_written(struct inode *inode,
1127                                            struct ceph_snap_context *snapc)
1128 {
1129         struct ceph_snap_context *oldest = get_oldest_context(inode, NULL);
1130         int ret = !oldest || snapc->seq <= oldest->seq;
1131
1132         ceph_put_snap_context(oldest);
1133         return ret;
1134 }
1135
1136 /*
1137  * We are only allowed to write into/dirty the page if the page is
1138  * clean, or already dirty within the same snap context.
1139  *
1140  * called with page locked.
1141  * return success with page locked,
1142  * or any failure (incl -EAGAIN) with page unlocked.
1143  */
1144 static int ceph_update_writeable_page(struct file *file,
1145                             loff_t pos, unsigned len,
1146                             struct page *page)
1147 {
1148         struct inode *inode = file_inode(file);
1149         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
1150         struct ceph_inode_info *ci = ceph_inode(inode);
1151         loff_t page_off = pos & PAGE_MASK;
1152         int pos_in_page = pos & ~PAGE_MASK;
1153         int end_in_page = pos_in_page + len;
1154         loff_t i_size;
1155         int r;
1156         struct ceph_snap_context *snapc, *oldest;
1157
1158         if (ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
1159                 dout(" page %p forced umount\n", page);
1160                 unlock_page(page);
1161                 return -EIO;
1162         }
1163
1164 retry_locked:
1165         /* writepages currently holds page lock, but if we change that later, */
1166         wait_on_page_writeback(page);
1167
1168         snapc = page_snap_context(page);
1169         if (snapc && snapc != ci->i_head_snapc) {
1170                 /*
1171                  * this page is already dirty in another (older) snap
1172                  * context!  is it writeable now?
1173                  */
1174                 oldest = get_oldest_context(inode, NULL);
1175
1176                 if (snapc->seq > oldest->seq) {
1177                         ceph_put_snap_context(oldest);
1178                         dout(" page %p snapc %p not current or oldest\n",
1179                              page, snapc);
1180                         /*
1181                          * queue for writeback, and wait for snapc to
1182                          * be writeable or written
1183                          */
1184                         snapc = ceph_get_snap_context(snapc);
1185                         unlock_page(page);
1186                         ceph_queue_writeback(inode);
1187                         r = wait_event_killable(ci->i_cap_wq,
1188                                context_is_writeable_or_written(inode, snapc));
1189                         ceph_put_snap_context(snapc);
1190                         if (r == -ERESTARTSYS)
1191                                 return r;
1192                         return -EAGAIN;
1193                 }
1194                 ceph_put_snap_context(oldest);
1195
1196                 /* yay, writeable, do it now (without dropping page lock) */
1197                 dout(" page %p snapc %p not current, but oldest\n",
1198                      page, snapc);
1199                 if (!clear_page_dirty_for_io(page))
1200                         goto retry_locked;
1201                 r = writepage_nounlock(page, NULL);
1202                 if (r < 0)
1203                         goto fail_nosnap;
1204                 goto retry_locked;
1205         }
1206
1207         if (PageUptodate(page)) {
1208                 dout(" page %p already uptodate\n", page);
1209                 return 0;
1210         }
1211
1212         /* full page? */
1213         if (pos_in_page == 0 && len == PAGE_SIZE)
1214                 return 0;
1215
1216         /* past end of file? */
1217         i_size = i_size_read(inode);
1218
1219         if (page_off >= i_size ||
1220             (pos_in_page == 0 && (pos+len) >= i_size &&
1221              end_in_page - pos_in_page != PAGE_SIZE)) {
1222                 dout(" zeroing %p 0 - %d and %d - %d\n",
1223                      page, pos_in_page, end_in_page, (int)PAGE_SIZE);
1224                 zero_user_segments(page,
1225                                    0, pos_in_page,
1226                                    end_in_page, PAGE_SIZE);
1227                 return 0;
1228         }
1229
1230         /* we need to read it. */
1231         r = readpage_nounlock(file, page);
1232         if (r < 0)
1233                 goto fail_nosnap;
1234         goto retry_locked;
1235 fail_nosnap:
1236         unlock_page(page);
1237         return r;
1238 }
1239
1240 /*
1241  * We are only allowed to write into/dirty the page if the page is
1242  * clean, or already dirty within the same snap context.
1243  */
1244 static int ceph_write_begin(struct file *file, struct address_space *mapping,
1245                             loff_t pos, unsigned len, unsigned flags,
1246                             struct page **pagep, void **fsdata)
1247 {
1248         struct inode *inode = file_inode(file);
1249         struct page *page;
1250         pgoff_t index = pos >> PAGE_SHIFT;
1251         int r;
1252
1253         do {
1254                 /* get a page */
1255                 page = grab_cache_page_write_begin(mapping, index, 0);
1256                 if (!page)
1257                         return -ENOMEM;
1258
1259                 dout("write_begin file %p inode %p page %p %d~%d\n", file,
1260                      inode, page, (int)pos, (int)len);
1261
1262                 r = ceph_update_writeable_page(file, pos, len, page);
1263                 if (r < 0)
1264                         put_page(page);
1265                 else
1266                         *pagep = page;
1267         } while (r == -EAGAIN);
1268
1269         return r;
1270 }
1271
1272 /*
1273  * we don't do anything in here that simple_write_end doesn't do
1274  * except adjust dirty page accounting
1275  */
1276 static int ceph_write_end(struct file *file, struct address_space *mapping,
1277                           loff_t pos, unsigned len, unsigned copied,
1278                           struct page *page, void *fsdata)
1279 {
1280         struct inode *inode = file_inode(file);
1281         unsigned from = pos & (PAGE_SIZE - 1);
1282         int check_cap = 0;
1283
1284         dout("write_end file %p inode %p page %p %d~%d (%d)\n", file,
1285              inode, page, (int)pos, (int)copied, (int)len);
1286
1287         /* zero the stale part of the page if we did a short copy */
1288         if (copied < len)
1289                 zero_user_segment(page, from+copied, len);
1290
1291         /* did file size increase? */
1292         if (pos+copied > i_size_read(inode))
1293                 check_cap = ceph_inode_set_size(inode, pos+copied);
1294
1295         if (!PageUptodate(page))
1296                 SetPageUptodate(page);
1297
1298         set_page_dirty(page);
1299
1300         unlock_page(page);
1301         put_page(page);
1302
1303         if (check_cap)
1304                 ceph_check_caps(ceph_inode(inode), CHECK_CAPS_AUTHONLY, NULL);
1305
1306         return copied;
1307 }
1308
1309 /*
1310  * we set .direct_IO to indicate direct io is supported, but since we
1311  * intercept O_DIRECT reads and writes early, this function should
1312  * never get called.
1313  */
1314 static ssize_t ceph_direct_io(struct kiocb *iocb, struct iov_iter *iter)
1315 {
1316         WARN_ON(1);
1317         return -EINVAL;
1318 }
1319
1320 const struct address_space_operations ceph_aops = {
1321         .readpage = ceph_readpage,
1322         .readpages = ceph_readpages,
1323         .writepage = ceph_writepage,
1324         .writepages = ceph_writepages_start,
1325         .write_begin = ceph_write_begin,
1326         .write_end = ceph_write_end,
1327         .set_page_dirty = ceph_set_page_dirty,
1328         .invalidatepage = ceph_invalidatepage,
1329         .releasepage = ceph_releasepage,
1330         .direct_IO = ceph_direct_io,
1331 };
1332
1333 static void ceph_block_sigs(sigset_t *oldset)
1334 {
1335         sigset_t mask;
1336         siginitsetinv(&mask, sigmask(SIGKILL));
1337         sigprocmask(SIG_BLOCK, &mask, oldset);
1338 }
1339
1340 static void ceph_restore_sigs(sigset_t *oldset)
1341 {
1342         sigprocmask(SIG_SETMASK, oldset, NULL);
1343 }
1344
1345 /*
1346  * vm ops
1347  */
1348 static int ceph_filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
1349 {
1350         struct inode *inode = file_inode(vma->vm_file);
1351         struct ceph_inode_info *ci = ceph_inode(inode);
1352         struct ceph_file_info *fi = vma->vm_file->private_data;
1353         struct page *pinned_page = NULL;
1354         loff_t off = vmf->pgoff << PAGE_SHIFT;
1355         int want, got, ret;
1356         sigset_t oldset;
1357
1358         ceph_block_sigs(&oldset);
1359
1360         dout("filemap_fault %p %llx.%llx %llu~%zd trying to get caps\n",
1361              inode, ceph_vinop(inode), off, (size_t)PAGE_SIZE);
1362         if (fi->fmode & CEPH_FILE_MODE_LAZY)
1363                 want = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO;
1364         else
1365                 want = CEPH_CAP_FILE_CACHE;
1366
1367         got = 0;
1368         ret = ceph_get_caps(ci, CEPH_CAP_FILE_RD, want, -1, &got, &pinned_page);
1369         if (ret < 0)
1370                 goto out_restore;
1371
1372         dout("filemap_fault %p %llu~%zd got cap refs on %s\n",
1373              inode, off, (size_t)PAGE_SIZE, ceph_cap_string(got));
1374
1375         if ((got & (CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO)) ||
1376             ci->i_inline_version == CEPH_INLINE_NONE)
1377                 ret = filemap_fault(vma, vmf);
1378         else
1379                 ret = -EAGAIN;
1380
1381         dout("filemap_fault %p %llu~%zd dropping cap refs on %s ret %d\n",
1382              inode, off, (size_t)PAGE_SIZE, ceph_cap_string(got), ret);
1383         if (pinned_page)
1384                 put_page(pinned_page);
1385         ceph_put_cap_refs(ci, got);
1386
1387         if (ret != -EAGAIN)
1388                 goto out_restore;
1389
1390         /* read inline data */
1391         if (off >= PAGE_SIZE) {
1392                 /* does not support inline data > PAGE_SIZE */
1393                 ret = VM_FAULT_SIGBUS;
1394         } else {
1395                 int ret1;
1396                 struct address_space *mapping = inode->i_mapping;
1397                 struct page *page = find_or_create_page(mapping, 0,
1398                                                 mapping_gfp_constraint(mapping,
1399                                                 ~__GFP_FS));
1400                 if (!page) {
1401                         ret = VM_FAULT_OOM;
1402                         goto out_inline;
1403                 }
1404                 ret1 = __ceph_do_getattr(inode, page,
1405                                          CEPH_STAT_CAP_INLINE_DATA, true);
1406                 if (ret1 < 0 || off >= i_size_read(inode)) {
1407                         unlock_page(page);
1408                         put_page(page);
1409                         if (ret1 < 0)
1410                                 ret = ret1;
1411                         else
1412                                 ret = VM_FAULT_SIGBUS;
1413                         goto out_inline;
1414                 }
1415                 if (ret1 < PAGE_SIZE)
1416                         zero_user_segment(page, ret1, PAGE_SIZE);
1417                 else
1418                         flush_dcache_page(page);
1419                 SetPageUptodate(page);
1420                 vmf->page = page;
1421                 ret = VM_FAULT_MAJOR | VM_FAULT_LOCKED;
1422 out_inline:
1423                 dout("filemap_fault %p %llu~%zd read inline data ret %d\n",
1424                      inode, off, (size_t)PAGE_SIZE, ret);
1425         }
1426 out_restore:
1427         ceph_restore_sigs(&oldset);
1428         if (ret < 0)
1429                 ret = (ret == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS;
1430
1431         return ret;
1432 }
1433
1434 /*
1435  * Reuse write_begin here for simplicity.
1436  */
1437 static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
1438 {
1439         struct inode *inode = file_inode(vma->vm_file);
1440         struct ceph_inode_info *ci = ceph_inode(inode);
1441         struct ceph_file_info *fi = vma->vm_file->private_data;
1442         struct ceph_cap_flush *prealloc_cf;
1443         struct page *page = vmf->page;
1444         loff_t off = page_offset(page);
1445         loff_t size = i_size_read(inode);
1446         size_t len;
1447         int want, got, ret;
1448         sigset_t oldset;
1449
1450         prealloc_cf = ceph_alloc_cap_flush();
1451         if (!prealloc_cf)
1452                 return VM_FAULT_OOM;
1453
1454         ceph_block_sigs(&oldset);
1455
1456         if (ci->i_inline_version != CEPH_INLINE_NONE) {
1457                 struct page *locked_page = NULL;
1458                 if (off == 0) {
1459                         lock_page(page);
1460                         locked_page = page;
1461                 }
1462                 ret = ceph_uninline_data(vma->vm_file, locked_page);
1463                 if (locked_page)
1464                         unlock_page(locked_page);
1465                 if (ret < 0)
1466                         goto out_free;
1467         }
1468
1469         if (off + PAGE_SIZE <= size)
1470                 len = PAGE_SIZE;
1471         else
1472                 len = size & ~PAGE_MASK;
1473
1474         dout("page_mkwrite %p %llx.%llx %llu~%zd getting caps i_size %llu\n",
1475              inode, ceph_vinop(inode), off, len, size);
1476         if (fi->fmode & CEPH_FILE_MODE_LAZY)
1477                 want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO;
1478         else
1479                 want = CEPH_CAP_FILE_BUFFER;
1480
1481         got = 0;
1482         ret = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, off + len,
1483                             &got, NULL);
1484         if (ret < 0)
1485                 goto out_free;
1486
1487         dout("page_mkwrite %p %llu~%zd got cap refs on %s\n",
1488              inode, off, len, ceph_cap_string(got));
1489
1490         /* Update time before taking page lock */
1491         file_update_time(vma->vm_file);
1492
1493         do {
1494                 lock_page(page);
1495
1496                 if ((off > size) || (page->mapping != inode->i_mapping)) {
1497                         unlock_page(page);
1498                         ret = VM_FAULT_NOPAGE;
1499                         break;
1500                 }
1501
1502                 ret = ceph_update_writeable_page(vma->vm_file, off, len, page);
1503                 if (ret >= 0) {
1504                         /* success.  we'll keep the page locked. */
1505                         set_page_dirty(page);
1506                         ret = VM_FAULT_LOCKED;
1507                 }
1508         } while (ret == -EAGAIN);
1509
1510         if (ret == VM_FAULT_LOCKED ||
1511             ci->i_inline_version != CEPH_INLINE_NONE) {
1512                 int dirty;
1513                 spin_lock(&ci->i_ceph_lock);
1514                 ci->i_inline_version = CEPH_INLINE_NONE;
1515                 dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR,
1516                                                &prealloc_cf);
1517                 spin_unlock(&ci->i_ceph_lock);
1518                 if (dirty)
1519                         __mark_inode_dirty(inode, dirty);
1520         }
1521
1522         dout("page_mkwrite %p %llu~%zd dropping cap refs on %s ret %d\n",
1523              inode, off, len, ceph_cap_string(got), ret);
1524         ceph_put_cap_refs(ci, got);
1525 out_free:
1526         ceph_restore_sigs(&oldset);
1527         ceph_free_cap_flush(prealloc_cf);
1528         if (ret < 0)
1529                 ret = (ret == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS;
1530         return ret;
1531 }
1532
1533 void ceph_fill_inline_data(struct inode *inode, struct page *locked_page,
1534                            char *data, size_t len)
1535 {
1536         struct address_space *mapping = inode->i_mapping;
1537         struct page *page;
1538
1539         if (locked_page) {
1540                 page = locked_page;
1541         } else {
1542                 if (i_size_read(inode) == 0)
1543                         return;
1544                 page = find_or_create_page(mapping, 0,
1545                                            mapping_gfp_constraint(mapping,
1546                                            ~__GFP_FS));
1547                 if (!page)
1548                         return;
1549                 if (PageUptodate(page)) {
1550                         unlock_page(page);
1551                         put_page(page);
1552                         return;
1553                 }
1554         }
1555
1556         dout("fill_inline_data %p %llx.%llx len %zu locked_page %p\n",
1557              inode, ceph_vinop(inode), len, locked_page);
1558
1559         if (len > 0) {
1560                 void *kaddr = kmap_atomic(page);
1561                 memcpy(kaddr, data, len);
1562                 kunmap_atomic(kaddr);
1563         }
1564
1565         if (page != locked_page) {
1566                 if (len < PAGE_SIZE)
1567                         zero_user_segment(page, len, PAGE_SIZE);
1568                 else
1569                         flush_dcache_page(page);
1570
1571                 SetPageUptodate(page);
1572                 unlock_page(page);
1573                 put_page(page);
1574         }
1575 }
1576
1577 int ceph_uninline_data(struct file *filp, struct page *locked_page)
1578 {
1579         struct inode *inode = file_inode(filp);
1580         struct ceph_inode_info *ci = ceph_inode(inode);
1581         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
1582         struct ceph_osd_request *req;
1583         struct page *page = NULL;
1584         u64 len, inline_version;
1585         int err = 0;
1586         bool from_pagecache = false;
1587
1588         spin_lock(&ci->i_ceph_lock);
1589         inline_version = ci->i_inline_version;
1590         spin_unlock(&ci->i_ceph_lock);
1591
1592         dout("uninline_data %p %llx.%llx inline_version %llu\n",
1593              inode, ceph_vinop(inode), inline_version);
1594
1595         if (inline_version == 1 || /* initial version, no data */
1596             inline_version == CEPH_INLINE_NONE)
1597                 goto out;
1598
1599         if (locked_page) {
1600                 page = locked_page;
1601                 WARN_ON(!PageUptodate(page));
1602         } else if (ceph_caps_issued(ci) &
1603                    (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) {
1604                 page = find_get_page(inode->i_mapping, 0);
1605                 if (page) {
1606                         if (PageUptodate(page)) {
1607                                 from_pagecache = true;
1608                                 lock_page(page);
1609                         } else {
1610                                 put_page(page);
1611                                 page = NULL;
1612                         }
1613                 }
1614         }
1615
1616         if (page) {
1617                 len = i_size_read(inode);
1618                 if (len > PAGE_SIZE)
1619                         len = PAGE_SIZE;
1620         } else {
1621                 page = __page_cache_alloc(GFP_NOFS);
1622                 if (!page) {
1623                         err = -ENOMEM;
1624                         goto out;
1625                 }
1626                 err = __ceph_do_getattr(inode, page,
1627                                         CEPH_STAT_CAP_INLINE_DATA, true);
1628                 if (err < 0) {
1629                         /* no inline data */
1630                         if (err == -ENODATA)
1631                                 err = 0;
1632                         goto out;
1633                 }
1634                 len = err;
1635         }
1636
1637         req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
1638                                     ceph_vino(inode), 0, &len, 0, 1,
1639                                     CEPH_OSD_OP_CREATE,
1640                                     CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
1641                                     NULL, 0, 0, false);
1642         if (IS_ERR(req)) {
1643                 err = PTR_ERR(req);
1644                 goto out;
1645         }
1646
1647         req->r_mtime = inode->i_mtime;
1648         err = ceph_osdc_start_request(&fsc->client->osdc, req, false);
1649         if (!err)
1650                 err = ceph_osdc_wait_request(&fsc->client->osdc, req);
1651         ceph_osdc_put_request(req);
1652         if (err < 0)
1653                 goto out;
1654
1655         req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
1656                                     ceph_vino(inode), 0, &len, 1, 3,
1657                                     CEPH_OSD_OP_WRITE,
1658                                     CEPH_OSD_FLAG_ONDISK | CEPH_OSD_FLAG_WRITE,
1659                                     NULL, ci->i_truncate_seq,
1660                                     ci->i_truncate_size, false);
1661         if (IS_ERR(req)) {
1662                 err = PTR_ERR(req);
1663                 goto out;
1664         }
1665
1666         osd_req_op_extent_osd_data_pages(req, 1, &page, len, 0, false, false);
1667
1668         {
1669                 __le64 xattr_buf = cpu_to_le64(inline_version);
1670                 err = osd_req_op_xattr_init(req, 0, CEPH_OSD_OP_CMPXATTR,
1671                                             "inline_version", &xattr_buf,
1672                                             sizeof(xattr_buf),
1673                                             CEPH_OSD_CMPXATTR_OP_GT,
1674                                             CEPH_OSD_CMPXATTR_MODE_U64);
1675                 if (err)
1676                         goto out_put;
1677         }
1678
1679         {
1680                 char xattr_buf[32];
1681                 int xattr_len = snprintf(xattr_buf, sizeof(xattr_buf),
1682                                          "%llu", inline_version);
1683                 err = osd_req_op_xattr_init(req, 2, CEPH_OSD_OP_SETXATTR,
1684                                             "inline_version",
1685                                             xattr_buf, xattr_len, 0, 0);
1686                 if (err)
1687                         goto out_put;
1688         }
1689
1690         req->r_mtime = inode->i_mtime;
1691         err = ceph_osdc_start_request(&fsc->client->osdc, req, false);
1692         if (!err)
1693                 err = ceph_osdc_wait_request(&fsc->client->osdc, req);
1694 out_put:
1695         ceph_osdc_put_request(req);
1696         if (err == -ECANCELED)
1697                 err = 0;
1698 out:
1699         if (page && page != locked_page) {
1700                 if (from_pagecache) {
1701                         unlock_page(page);
1702                         put_page(page);
1703                 } else
1704                         __free_pages(page, 0);
1705         }
1706
1707         dout("uninline_data %p %llx.%llx inline_version %llu = %d\n",
1708              inode, ceph_vinop(inode), inline_version, err);
1709         return err;
1710 }
1711
1712 static const struct vm_operations_struct ceph_vmops = {
1713         .fault          = ceph_filemap_fault,
1714         .page_mkwrite   = ceph_page_mkwrite,
1715 };
1716
1717 int ceph_mmap(struct file *file, struct vm_area_struct *vma)
1718 {
1719         struct address_space *mapping = file->f_mapping;
1720
1721         if (!mapping->a_ops->readpage)
1722                 return -ENOEXEC;
1723         file_accessed(file);
1724         vma->vm_ops = &ceph_vmops;
1725         return 0;
1726 }
1727
1728 enum {
1729         POOL_READ       = 1,
1730         POOL_WRITE      = 2,
1731 };
1732
1733 static int __ceph_pool_perm_get(struct ceph_inode_info *ci,
1734                                 s64 pool, struct ceph_string *pool_ns)
1735 {
1736         struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode);
1737         struct ceph_mds_client *mdsc = fsc->mdsc;
1738         struct ceph_osd_request *rd_req = NULL, *wr_req = NULL;
1739         struct rb_node **p, *parent;
1740         struct ceph_pool_perm *perm;
1741         struct page **pages;
1742         size_t pool_ns_len;
1743         int err = 0, err2 = 0, have = 0;
1744
1745         down_read(&mdsc->pool_perm_rwsem);
1746         p = &mdsc->pool_perm_tree.rb_node;
1747         while (*p) {
1748                 perm = rb_entry(*p, struct ceph_pool_perm, node);
1749                 if (pool < perm->pool)
1750                         p = &(*p)->rb_left;
1751                 else if (pool > perm->pool)
1752                         p = &(*p)->rb_right;
1753                 else {
1754                         int ret = ceph_compare_string(pool_ns,
1755                                                 perm->pool_ns,
1756                                                 perm->pool_ns_len);
1757                         if (ret < 0)
1758                                 p = &(*p)->rb_left;
1759                         else if (ret > 0)
1760                                 p = &(*p)->rb_right;
1761                         else {
1762                                 have = perm->perm;
1763                                 break;
1764                         }
1765                 }
1766         }
1767         up_read(&mdsc->pool_perm_rwsem);
1768         if (*p)
1769                 goto out;
1770
1771         if (pool_ns)
1772                 dout("__ceph_pool_perm_get pool %lld ns %.*s no perm cached\n",
1773                      pool, (int)pool_ns->len, pool_ns->str);
1774         else
1775                 dout("__ceph_pool_perm_get pool %lld no perm cached\n", pool);
1776
1777         down_write(&mdsc->pool_perm_rwsem);
1778         p = &mdsc->pool_perm_tree.rb_node;
1779         parent = NULL;
1780         while (*p) {
1781                 parent = *p;
1782                 perm = rb_entry(parent, struct ceph_pool_perm, node);
1783                 if (pool < perm->pool)
1784                         p = &(*p)->rb_left;
1785                 else if (pool > perm->pool)
1786                         p = &(*p)->rb_right;
1787                 else {
1788                         int ret = ceph_compare_string(pool_ns,
1789                                                 perm->pool_ns,
1790                                                 perm->pool_ns_len);
1791                         if (ret < 0)
1792                                 p = &(*p)->rb_left;
1793                         else if (ret > 0)
1794                                 p = &(*p)->rb_right;
1795                         else {
1796                                 have = perm->perm;
1797                                 break;
1798                         }
1799                 }
1800         }
1801         if (*p) {
1802                 up_write(&mdsc->pool_perm_rwsem);
1803                 goto out;
1804         }
1805
1806         rd_req = ceph_osdc_alloc_request(&fsc->client->osdc, NULL,
1807                                          1, false, GFP_NOFS);
1808         if (!rd_req) {
1809                 err = -ENOMEM;
1810                 goto out_unlock;
1811         }
1812
1813         rd_req->r_flags = CEPH_OSD_FLAG_READ;
1814         osd_req_op_init(rd_req, 0, CEPH_OSD_OP_STAT, 0);
1815         rd_req->r_base_oloc.pool = pool;
1816         if (pool_ns)
1817                 rd_req->r_base_oloc.pool_ns = ceph_get_string(pool_ns);
1818         ceph_oid_printf(&rd_req->r_base_oid, "%llx.00000000", ci->i_vino.ino);
1819
1820         err = ceph_osdc_alloc_messages(rd_req, GFP_NOFS);
1821         if (err)
1822                 goto out_unlock;
1823
1824         wr_req = ceph_osdc_alloc_request(&fsc->client->osdc, NULL,
1825                                          1, false, GFP_NOFS);
1826         if (!wr_req) {
1827                 err = -ENOMEM;
1828                 goto out_unlock;
1829         }
1830
1831         wr_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ACK;
1832         osd_req_op_init(wr_req, 0, CEPH_OSD_OP_CREATE, CEPH_OSD_OP_FLAG_EXCL);
1833         ceph_oloc_copy(&wr_req->r_base_oloc, &rd_req->r_base_oloc);
1834         ceph_oid_copy(&wr_req->r_base_oid, &rd_req->r_base_oid);
1835
1836         err = ceph_osdc_alloc_messages(wr_req, GFP_NOFS);
1837         if (err)
1838                 goto out_unlock;
1839
1840         /* one page should be large enough for STAT data */
1841         pages = ceph_alloc_page_vector(1, GFP_KERNEL);
1842         if (IS_ERR(pages)) {
1843                 err = PTR_ERR(pages);
1844                 goto out_unlock;
1845         }
1846
1847         osd_req_op_raw_data_in_pages(rd_req, 0, pages, PAGE_SIZE,
1848                                      0, false, true);
1849         err = ceph_osdc_start_request(&fsc->client->osdc, rd_req, false);
1850
1851         wr_req->r_mtime = ci->vfs_inode.i_mtime;
1852         err2 = ceph_osdc_start_request(&fsc->client->osdc, wr_req, false);
1853
1854         if (!err)
1855                 err = ceph_osdc_wait_request(&fsc->client->osdc, rd_req);
1856         if (!err2)
1857                 err2 = ceph_osdc_wait_request(&fsc->client->osdc, wr_req);
1858
1859         if (err >= 0 || err == -ENOENT)
1860                 have |= POOL_READ;
1861         else if (err != -EPERM)
1862                 goto out_unlock;
1863
1864         if (err2 == 0 || err2 == -EEXIST)
1865                 have |= POOL_WRITE;
1866         else if (err2 != -EPERM) {
1867                 err = err2;
1868                 goto out_unlock;
1869         }
1870
1871         pool_ns_len = pool_ns ? pool_ns->len : 0;
1872         perm = kmalloc(sizeof(*perm) + pool_ns_len + 1, GFP_NOFS);
1873         if (!perm) {
1874                 err = -ENOMEM;
1875                 goto out_unlock;
1876         }
1877
1878         perm->pool = pool;
1879         perm->perm = have;
1880         perm->pool_ns_len = pool_ns_len;
1881         if (pool_ns_len > 0)
1882                 memcpy(perm->pool_ns, pool_ns->str, pool_ns_len);
1883         perm->pool_ns[pool_ns_len] = 0;
1884
1885         rb_link_node(&perm->node, parent, p);
1886         rb_insert_color(&perm->node, &mdsc->pool_perm_tree);
1887         err = 0;
1888 out_unlock:
1889         up_write(&mdsc->pool_perm_rwsem);
1890
1891         ceph_osdc_put_request(rd_req);
1892         ceph_osdc_put_request(wr_req);
1893 out:
1894         if (!err)
1895                 err = have;
1896         if (pool_ns)
1897                 dout("__ceph_pool_perm_get pool %lld ns %.*s result = %d\n",
1898                      pool, (int)pool_ns->len, pool_ns->str, err);
1899         else
1900                 dout("__ceph_pool_perm_get pool %lld result = %d\n", pool, err);
1901         return err;
1902 }
1903
1904 int ceph_pool_perm_check(struct ceph_inode_info *ci, int need)
1905 {
1906         s64 pool;
1907         struct ceph_string *pool_ns;
1908         int ret, flags;
1909
1910         if (ceph_test_mount_opt(ceph_inode_to_client(&ci->vfs_inode),
1911                                 NOPOOLPERM))
1912                 return 0;
1913
1914         spin_lock(&ci->i_ceph_lock);
1915         flags = ci->i_ceph_flags;
1916         pool = ci->i_layout.pool_id;
1917         spin_unlock(&ci->i_ceph_lock);
1918 check:
1919         if (flags & CEPH_I_POOL_PERM) {
1920                 if ((need & CEPH_CAP_FILE_RD) && !(flags & CEPH_I_POOL_RD)) {
1921                         dout("ceph_pool_perm_check pool %lld no read perm\n",
1922                              pool);
1923                         return -EPERM;
1924                 }
1925                 if ((need & CEPH_CAP_FILE_WR) && !(flags & CEPH_I_POOL_WR)) {
1926                         dout("ceph_pool_perm_check pool %lld no write perm\n",
1927                              pool);
1928                         return -EPERM;
1929                 }
1930                 return 0;
1931         }
1932
1933         pool_ns = ceph_try_get_string(ci->i_layout.pool_ns);
1934         ret = __ceph_pool_perm_get(ci, pool, pool_ns);
1935         ceph_put_string(pool_ns);
1936         if (ret < 0)
1937                 return ret;
1938
1939         flags = CEPH_I_POOL_PERM;
1940         if (ret & POOL_READ)
1941                 flags |= CEPH_I_POOL_RD;
1942         if (ret & POOL_WRITE)
1943                 flags |= CEPH_I_POOL_WR;
1944
1945         spin_lock(&ci->i_ceph_lock);
1946         if (pool == ci->i_layout.pool_id &&
1947             pool_ns == rcu_dereference_raw(ci->i_layout.pool_ns)) {
1948                 ci->i_ceph_flags |= flags;
1949         } else {
1950                 pool = ci->i_layout.pool_id;
1951                 flags = ci->i_ceph_flags;
1952         }
1953         spin_unlock(&ci->i_ceph_lock);
1954         goto check;
1955 }
1956
1957 void ceph_pool_perm_destroy(struct ceph_mds_client *mdsc)
1958 {
1959         struct ceph_pool_perm *perm;
1960         struct rb_node *n;
1961
1962         while (!RB_EMPTY_ROOT(&mdsc->pool_perm_tree)) {
1963                 n = rb_first(&mdsc->pool_perm_tree);
1964                 perm = rb_entry(n, struct ceph_pool_perm, node);
1965                 rb_erase(n, &mdsc->pool_perm_tree);
1966                 kfree(perm);
1967         }
1968 }