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 <linux/ceph/osd_client.h>
  15
  16 /*
  17  * Ceph address space ops.
  18  *
  19  * There are a few funny things going on here.
  20  *
  21  * The page->private field is used to reference a struct
  22  * ceph_snap_context for _every_ dirty page.  This indicates which
  23  * snapshot the page was logically dirtied in, and thus which snap
  24  * context needs to be associated with the osd write during writeback.
  25  *
  26  * Similarly, struct ceph_inode_info maintains a set of counters to
  27  * count dirty pages on the inode.  In the absence of snapshots,
  28  * i_wrbuffer_ref == i_wrbuffer_ref_head == the dirty page count.
  29  *
  30  * When a snapshot is taken (that is, when the client receives
  31  * notification that a snapshot was taken), each inode with caps and
  32  * with dirty pages (dirty pages implies there is a cap) gets a new
  33  * ceph_cap_snap in the i_cap_snaps list (which is sorted in ascending
  34  * order, new snaps go to the tail).  The i_wrbuffer_ref_head count is
  35  * moved to capsnap->dirty. (Unless a sync write is currently in
  36  * progress.  In that case, the capsnap is said to be "pending", new
  37  * writes cannot start, and the capsnap isn't "finalized" until the
  38  * write completes (or fails) and a final size/mtime for the inode for
  39  * that snap can be settled upon.)  i_wrbuffer_ref_head is reset to 0.
  40  *
  41  * On writeback, we must submit writes to the osd IN SNAP ORDER.  So,
  42  * we look for the first capsnap in i_cap_snaps and write out pages in
  43  * that snap context _only_.  Then we move on to the next capsnap,
  44  * eventually reaching the "live" or "head" context (i.e., pages that
  45  * are not yet snapped) and are writing the most recently dirtied
  46  * pages.
  47  *
  48  * Invalidate and so forth must take care to ensure the dirty page
  49  * accounting is preserved.
  50  */
  51
  52 #define CONGESTION_ON_THRESH(congestion_kb) (congestion_kb >> (PAGE_SHIFT-10))
  53 #define CONGESTION_OFF_THRESH(congestion_kb)                            \
  54         (CONGESTION_ON_THRESH(congestion_kb) -                          \
  55          (CONGESTION_ON_THRESH(congestion_kb) >> 2))
  56
  57
  58
  59 /*
  60  * Dirty a page.  Optimistically adjust accounting, on the assumption
  61  * that we won't race with invalidate.  If we do, readjust.
  62  */
  63 static int ceph_set_page_dirty(struct page *page)
  64 {
  65         struct address_space *mapping = page->mapping;
  66         struct inode *inode;
  67         struct ceph_inode_info *ci;
  68         int undo = 0;
  69         struct ceph_snap_context *snapc;
  70
  71         if (unlikely(!mapping))
  72                 return !TestSetPageDirty(page);
  73
  74         if (TestSetPageDirty(page)) {
  75                 dout("%p set_page_dirty %p idx %lu -- already dirty\n",
  76                      mapping->host, page, page->index);
  77                 return 0;
  78         }
  79
  80         inode = mapping->host;
  81         ci = ceph_inode(inode);
  82
  83         /*
  84          * Note that we're grabbing a snapc ref here without holding
  85          * any locks!
  86          */
  87         snapc = ceph_get_snap_context(ci->i_snap_realm->cached_context);
  88
  89         /* dirty the head */
  90         spin_lock(&ci->i_ceph_lock);
  91         if (ci->i_head_snapc == NULL)
  92                 ci->i_head_snapc = ceph_get_snap_context(snapc);
  93         ++ci->i_wrbuffer_ref_head;
  94         if (ci->i_wrbuffer_ref == 0)
  95                 ihold(inode);
  96         ++ci->i_wrbuffer_ref;
  97         dout("%p set_page_dirty %p idx %lu head %d/%d -> %d/%d "
  98              "snapc %p seq %lld (%d snaps)\n",
  99              mapping->host, page, page->index,
 100              ci->i_wrbuffer_ref-1, ci->i_wrbuffer_ref_head-1,
 101              ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head,
 102              snapc, snapc->seq, snapc->num_snaps);
 103         spin_unlock(&ci->i_ceph_lock);
 104
 105         /* now adjust page */
 106         spin_lock_irq(&mapping->tree_lock);
 107         if (page->mapping) {    /* Race with truncate? */
 108                 WARN_ON_ONCE(!PageUptodate(page));
 109                 account_page_dirtied(page, page->mapping);
 110                 radix_tree_tag_set(&mapping->page_tree,
 111                                 page_index(page), PAGECACHE_TAG_DIRTY);
 112
 113                 /*
 114                  * Reference snap context in page->private.  Also set
 115                  * PagePrivate so that we get invalidatepage callback.
 116                  */
 117                 page->private = (unsigned long)snapc;
 118                 SetPagePrivate(page);
 119         } else {
 120                 dout("ANON set_page_dirty %p (raced truncate?)\n", page);
 121                 undo = 1;
 122         }
 123
 124         spin_unlock_irq(&mapping->tree_lock);
 125
 126         if (undo)
 127                 /* whoops, we failed to dirty the page */
 128                 ceph_put_wrbuffer_cap_refs(ci, 1, snapc);
 129
 130         __mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
 131
 132         BUG_ON(!PageDirty(page));
 133         return 1;
 134 }
 135
 136 /*
 137  * If we are truncating the full page (i.e. offset == 0), adjust the
 138  * dirty page counters appropriately.  Only called if there is private
 139  * data on the page.
 140  */
 141 static void ceph_invalidatepage(struct page *page, unsigned long offset)
 142 {
 143         struct inode *inode;
 144         struct ceph_inode_info *ci;
 145         struct ceph_snap_context *snapc = (void *)page->private;
 146
 147         BUG_ON(!PageLocked(page));
 148         BUG_ON(!page->private);
 149         BUG_ON(!PagePrivate(page));
 150         BUG_ON(!page->mapping);
 151
 152         inode = page->mapping->host;
 153
 154         /*
 155          * We can get non-dirty pages here due to races between
 156          * set_page_dirty and truncate_complete_page; just spit out a
 157          * warning, in case we end up with accounting problems later.
 158          */
 159         if (!PageDirty(page))
 160                 pr_err("%p invalidatepage %p page not dirty\n", inode, page);
 161
 162         if (offset == 0)
 163                 ClearPageChecked(page);
 164
 165         ci = ceph_inode(inode);
 166         if (offset == 0) {
 167                 dout("%p invalidatepage %p idx %lu full dirty page %lu\n",
 168                      inode, page, page->index, offset);
 169                 ceph_put_wrbuffer_cap_refs(ci, 1, snapc);
 170                 ceph_put_snap_context(snapc);
 171                 page->private = 0;
 172                 ClearPagePrivate(page);
 173         } else {
 174                 dout("%p invalidatepage %p idx %lu partial dirty page\n",
 175                      inode, page, page->index);
 176         }
 177 }
 178
 179 /* just a sanity check */
 180 static int ceph_releasepage(struct page *page, gfp_t g)
 181 {
 182         struct inode *inode = page->mapping ? page->mapping->host : NULL;
 183         dout("%p releasepage %p idx %lu\n", inode, page, page->index);
 184         WARN_ON(PageDirty(page));
 185         WARN_ON(page->private);
 186         WARN_ON(PagePrivate(page));
 187         return 0;
 188 }
 189
 190 /*
 191  * read a single page, without unlocking it.
 192  */
 193 static int readpage_nounlock(struct file *filp, struct page *page)
 194 {
 195         struct inode *inode = filp->f_dentry->d_inode;
 196         struct ceph_inode_info *ci = ceph_inode(inode);
 197         struct ceph_osd_client *osdc =
 198                 &ceph_inode_to_client(inode)->client->osdc;
 199         int err = 0;
 200         u64 len = PAGE_CACHE_SIZE;
 201
 202         dout("readpage inode %p file %p page %p index %lu\n",
 203              inode, filp, page, page->index);
 204         err = ceph_osdc_readpages(osdc, ceph_vino(inode), &ci->i_layout,
 205                                   page->index << PAGE_CACHE_SHIFT, &len,
 206                                   ci->i_truncate_seq, ci->i_truncate_size,
 207                                   &page, 1, 0);
 208         if (err == -ENOENT)
 209                 err = 0;
 210         if (err < 0) {
 211                 SetPageError(page);
 212                 goto out;
 213         } else if (err < PAGE_CACHE_SIZE) {
 214                 /* zero fill remainder of page */
 215                 zero_user_segment(page, err, PAGE_CACHE_SIZE);
 216         }
 217         SetPageUptodate(page);
 218
 219 out:
 220         return err < 0 ? err : 0;
 221 }
 222
 223 static int ceph_readpage(struct file *filp, struct page *page)
 224 {
 225         int r = readpage_nounlock(filp, page);
 226         unlock_page(page);
 227         return r;
 228 }
 229
 230 /*
 231  * Finish an async read(ahead) op.
 232  */
 233 static void finish_read(struct ceph_osd_request *req, struct ceph_msg *msg)
 234 {
 235         struct inode *inode = req->r_inode;
 236         struct ceph_osd_reply_head *replyhead;
 237         int rc, bytes;
 238         int i;
 239
 240         /* parse reply */
 241         replyhead = msg->front.iov_base;
 242         WARN_ON(le32_to_cpu(replyhead->num_ops) == 0);
 243         rc = le32_to_cpu(replyhead->result);
 244         bytes = le32_to_cpu(msg->hdr.data_len);
 245
 246         dout("finish_read %p req %p rc %d bytes %d\n", inode, req, rc, bytes);
 247
 248         /* unlock all pages, zeroing any data we didn't read */
 249         for (i = 0; i < req->r_num_pages; i++, bytes -= PAGE_CACHE_SIZE) {
 250                 struct page *page = req->r_pages[i];
 251
 252                 if (bytes < (int)PAGE_CACHE_SIZE) {
 253                         /* zero (remainder of) page */
 254                         int s = bytes < 0 ? 0 : bytes;
 255                         zero_user_segment(page, s, PAGE_CACHE_SIZE);
 256                 }
 257                 dout("finish_read %p uptodate %p idx %lu\n", inode, page,
 258                      page->index);
 259                 flush_dcache_page(page);
 260                 SetPageUptodate(page);
 261                 unlock_page(page);
 262                 page_cache_release(page);
 263         }
 264         kfree(req->r_pages);
 265 }
 266
 267 /*
 268  * start an async read(ahead) operation.  return nr_pages we submitted
 269  * a read for on success, or negative error code.
 270  */
 271 static int start_read(struct inode *inode, struct list_head *page_list, int max)
 272 {
 273         struct ceph_osd_client *osdc =
 274                 &ceph_inode_to_client(inode)->client->osdc;
 275         struct ceph_inode_info *ci = ceph_inode(inode);
 276         struct page *page = list_entry(page_list->prev, struct page, lru);
 277         struct ceph_osd_request *req;
 278         u64 off;
 279         u64 len;
 280         int i;
 281         struct page **pages;
 282         pgoff_t next_index;
 283         int nr_pages = 0;
 284         int ret;
 285
 286         off = page->index << PAGE_CACHE_SHIFT;
 287
 288         /* count pages */
 289         next_index = page->index;
 290         list_for_each_entry_reverse(page, page_list, lru) {
 291                 if (page->index != next_index)
 292                         break;
 293                 nr_pages++;
 294                 next_index++;
 295                 if (max && nr_pages == max)
 296                         break;
 297         }
 298         len = nr_pages << PAGE_CACHE_SHIFT;
 299         dout("start_read %p nr_pages %d is %lld~%lld\n", inode, nr_pages,
 300              off, len);
 301
 302         req = ceph_osdc_new_request(osdc, &ci->i_layout, ceph_vino(inode),
 303                                     off, &len,
 304                                     CEPH_OSD_OP_READ, CEPH_OSD_FLAG_READ,
 305                                     NULL, 0,
 306                                     ci->i_truncate_seq, ci->i_truncate_size,
 307                                     NULL, false, 1, 0);
 308         if (!req)
 309                 return -ENOMEM;
 310
 311         /* build page vector */
 312         nr_pages = len >> PAGE_CACHE_SHIFT;
 313         pages = kmalloc(sizeof(*pages) * nr_pages, GFP_NOFS);
 314         ret = -ENOMEM;
 315         if (!pages)
 316                 goto out;
 317         for (i = 0; i < nr_pages; ++i) {
 318                 page = list_entry(page_list->prev, struct page, lru);
 319                 BUG_ON(PageLocked(page));
 320                 list_del(&page->lru);
 321
 322                 dout("start_read %p adding %p idx %lu\n", inode, page,
 323                      page->index);
 324                 if (add_to_page_cache_lru(page, &inode->i_data, page->index,
 325                                           GFP_NOFS)) {
 326                         page_cache_release(page);
 327                         dout("start_read %p add_to_page_cache failed %p\n",
 328                              inode, page);
 329                         nr_pages = i;
 330                         goto out_pages;
 331                 }
 332                 pages[i] = page;
 333         }
 334         req->r_pages = pages;
 335         req->r_num_pages = nr_pages;
 336         req->r_callback = finish_read;
 337         req->r_inode = inode;
 338
 339         dout("start_read %p starting %p %lld~%lld\n", inode, req, off, len);
 340         ret = ceph_osdc_start_request(osdc, req, false);
 341         if (ret < 0)
 342                 goto out_pages;
 343         ceph_osdc_put_request(req);
 344         return nr_pages;
 345
 346 out_pages:
 347         ceph_release_page_vector(pages, nr_pages);
 348 out:
 349         ceph_osdc_put_request(req);
 350         return ret;
 351 }
 352
 353
 354 /*
 355  * Read multiple pages.  Leave pages we don't read + unlock in page_list;
 356  * the caller (VM) cleans them up.
 357  */
 358 static int ceph_readpages(struct file *file, struct address_space *mapping,
 359                           struct list_head *page_list, unsigned nr_pages)
 360 {
 361         struct inode *inode = file->f_dentry->d_inode;
 362         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 363         int rc = 0;
 364         int max = 0;
 365
 366         if (fsc->mount_options->rsize >= PAGE_CACHE_SIZE)
 367                 max = (fsc->mount_options->rsize + PAGE_CACHE_SIZE - 1)
 368                         >> PAGE_SHIFT;
 369
 370         dout("readpages %p file %p nr_pages %d max %d\n", inode, file, nr_pages,
 371              max);
 372         while (!list_empty(page_list)) {
 373                 rc = start_read(inode, page_list, max);
 374                 if (rc < 0)
 375                         goto out;
 376                 BUG_ON(rc == 0);
 377         }
 378 out:
 379         dout("readpages %p file %p ret %d\n", inode, file, rc);
 380         return rc;
 381 }
 382
 383 /*
 384  * Get ref for the oldest snapc for an inode with dirty data... that is, the
 385  * only snap context we are allowed to write back.
 386  */
 387 static struct ceph_snap_context *get_oldest_context(struct inode *inode,
 388                                                     u64 *snap_size)
 389 {
 390         struct ceph_inode_info *ci = ceph_inode(inode);
 391         struct ceph_snap_context *snapc = NULL;
 392         struct ceph_cap_snap *capsnap = NULL;
 393
 394         spin_lock(&ci->i_ceph_lock);
 395         list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) {
 396                 dout(" cap_snap %p snapc %p has %d dirty pages\n", capsnap,
 397                      capsnap->context, capsnap->dirty_pages);
 398                 if (capsnap->dirty_pages) {
 399                         snapc = ceph_get_snap_context(capsnap->context);
 400                         if (snap_size)
 401                                 *snap_size = capsnap->size;
 402                         break;
 403                 }
 404         }
 405         if (!snapc && ci->i_wrbuffer_ref_head) {
 406                 snapc = ceph_get_snap_context(ci->i_head_snapc);
 407                 dout(" head snapc %p has %d dirty pages\n",
 408                      snapc, ci->i_wrbuffer_ref_head);
 409         }
 410         spin_unlock(&ci->i_ceph_lock);
 411         return snapc;
 412 }
 413
 414 /*
 415  * Write a single page, but leave the page locked.
 416  *
 417  * If we get a write error, set the page error bit, but still adjust the
 418  * dirty page accounting (i.e., page is no longer dirty).
 419  */
 420 static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 421 {
 422         struct inode *inode;
 423         struct ceph_inode_info *ci;
 424         struct ceph_fs_client *fsc;
 425         struct ceph_osd_client *osdc;
 426         loff_t page_off = page->index << PAGE_CACHE_SHIFT;
 427         int len = PAGE_CACHE_SIZE;
 428         loff_t i_size;
 429         int err = 0;
 430         struct ceph_snap_context *snapc, *oldest;
 431         u64 snap_size = 0;
 432         long writeback_stat;
 433
 434         dout("writepage %p idx %lu\n", page, page->index);
 435
 436         if (!page->mapping || !page->mapping->host) {
 437                 dout("writepage %p - no mapping\n", page);
 438                 return -EFAULT;
 439         }
 440         inode = page->mapping->host;
 441         ci = ceph_inode(inode);
 442         fsc = ceph_inode_to_client(inode);
 443         osdc = &fsc->client->osdc;
 444
 445         /* verify this is a writeable snap context */
 446         snapc = (void *)page->private;
 447         if (snapc == NULL) {
 448                 dout("writepage %p page %p not dirty?\n", inode, page);
 449                 goto out;
 450         }
 451         oldest = get_oldest_context(inode, &snap_size);
 452         if (snapc->seq > oldest->seq) {
 453                 dout("writepage %p page %p snapc %p not writeable - noop\n",
 454                      inode, page, (void *)page->private);
 455                 /* we should only noop if called by kswapd */
 456                 WARN_ON((current->flags & PF_MEMALLOC) == 0);
 457                 ceph_put_snap_context(oldest);
 458                 goto out;
 459         }
 460         ceph_put_snap_context(oldest);
 461
 462         /* is this a partial page at end of file? */
 463         if (snap_size)
 464                 i_size = snap_size;
 465         else
 466                 i_size = i_size_read(inode);
 467         if (i_size < page_off + len)
 468                 len = i_size - page_off;
 469
 470         dout("writepage %p page %p index %lu on %llu~%u snapc %p\n",
 471              inode, page, page->index, page_off, len, snapc);
 472
 473         writeback_stat = atomic_long_inc_return(&fsc->writeback_count);
 474         if (writeback_stat >
 475             CONGESTION_ON_THRESH(fsc->mount_options->congestion_kb))
 476                 set_bdi_congested(&fsc->backing_dev_info, BLK_RW_ASYNC);
 477
 478         set_page_writeback(page);
 479         err = ceph_osdc_writepages(osdc, ceph_vino(inode),
 480                                    &ci->i_layout, snapc,
 481                                    page_off, len,
 482                                    ci->i_truncate_seq, ci->i_truncate_size,
 483                                    &inode->i_mtime,
 484                                    &page, 1, 0, 0, true);
 485         if (err < 0) {
 486                 dout("writepage setting page/mapping error %d %p\n", err, page);
 487                 SetPageError(page);
 488                 mapping_set_error(&inode->i_data, err);
 489                 if (wbc)
 490                         wbc->pages_skipped++;
 491         } else {
 492                 dout("writepage cleaned page %p\n", page);
 493                 err = 0;  /* vfs expects us to return 0 */
 494         }
 495         page->private = 0;
 496         ClearPagePrivate(page);
 497         end_page_writeback(page);
 498         ceph_put_wrbuffer_cap_refs(ci, 1, snapc);
 499         ceph_put_snap_context(snapc);  /* page's reference */
 500 out:
 501         return err;
 502 }
 503
 504 static int ceph_writepage(struct page *page, struct writeback_control *wbc)
 505 {
 506         int err;
 507         struct inode *inode = page->mapping->host;
 508         BUG_ON(!inode);
 509         ihold(inode);
 510         err = writepage_nounlock(page, wbc);
 511         unlock_page(page);
 512         iput(inode);
 513         return err;
 514 }
 515
 516
 517 /*
 518  * lame release_pages helper.  release_pages() isn't exported to
 519  * modules.
 520  */
 521 static void ceph_release_pages(struct page **pages, int num)
 522 {
 523         struct pagevec pvec;
 524         int i;
 525
 526         pagevec_init(&pvec, 0);
 527         for (i = 0; i < num; i++) {
 528                 if (pagevec_add(&pvec, pages[i]) == 0)
 529                         pagevec_release(&pvec);
 530         }
 531         pagevec_release(&pvec);
 532 }
 533
 534
 535 /*
 536  * async writeback completion handler.
 537  *
 538  * If we get an error, set the mapping error bit, but not the individual
 539  * page error bits.
 540  */
 541 static void writepages_finish(struct ceph_osd_request *req,
 542                               struct ceph_msg *msg)
 543 {
 544         struct inode *inode = req->r_inode;
 545         struct ceph_osd_reply_head *replyhead;
 546         struct ceph_osd_op *op;
 547         struct ceph_inode_info *ci = ceph_inode(inode);
 548         unsigned wrote;
 549         struct page *page;
 550         int i;
 551         struct ceph_snap_context *snapc = req->r_snapc;
 552         struct address_space *mapping = inode->i_mapping;
 553         __s32 rc = -EIO;
 554         u64 bytes = 0;
 555         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
 556         long writeback_stat;
 557         unsigned issued = ceph_caps_issued(ci);
 558
 559         /* parse reply */
 560         replyhead = msg->front.iov_base;
 561         WARN_ON(le32_to_cpu(replyhead->num_ops) == 0);
 562         op = (void *)(replyhead + 1);
 563         rc = le32_to_cpu(replyhead->result);
 564         bytes = le64_to_cpu(op->extent.length);
 565
 566         if (rc >= 0) {
 567                 /*
 568                  * Assume we wrote the pages we originally sent.  The
 569                  * osd might reply with fewer pages if our writeback
 570                  * raced with a truncation and was adjusted at the osd,
 571                  * so don't believe the reply.
 572                  */
 573                 wrote = req->r_num_pages;
 574         } else {
 575                 wrote = 0;
 576                 mapping_set_error(mapping, rc);
 577         }
 578         dout("writepages_finish %p rc %d bytes %llu wrote %d (pages)\n",
 579              inode, rc, bytes, wrote);
 580
 581         /* clean all pages */
 582         for (i = 0; i < req->r_num_pages; i++) {
 583                 page = req->r_pages[i];
 584                 BUG_ON(!page);
 585                 WARN_ON(!PageUptodate(page));
 586
 587                 writeback_stat =
 588                         atomic_long_dec_return(&fsc->writeback_count);
 589                 if (writeback_stat <
 590                     CONGESTION_OFF_THRESH(fsc->mount_options->congestion_kb))
 591                         clear_bdi_congested(&fsc->backing_dev_info,
 592                                             BLK_RW_ASYNC);
 593
 594                 ceph_put_snap_context((void *)page->private);
 595                 page->private = 0;
 596                 ClearPagePrivate(page);
 597                 dout("unlocking %d %p\n", i, page);
 598                 end_page_writeback(page);
 599
 600                 /*
 601                  * We lost the cache cap, need to truncate the page before
 602                  * it is unlocked, otherwise we'd truncate it later in the
 603                  * page truncation thread, possibly losing some data that
 604                  * raced its way in
 605                  */
 606                 if ((issued & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) == 0)
 607                         generic_error_remove_page(inode->i_mapping, page);
 608
 609                 unlock_page(page);
 610         }
 611         dout("%p wrote+cleaned %d pages\n", inode, wrote);
 612         ceph_put_wrbuffer_cap_refs(ci, req->r_num_pages, snapc);
 613
 614         ceph_release_pages(req->r_pages, req->r_num_pages);
 615         if (req->r_pages_from_pool)
 616                 mempool_free(req->r_pages,
 617                              ceph_sb_to_client(inode->i_sb)->wb_pagevec_pool);
 618         else
 619                 kfree(req->r_pages);
 620         ceph_osdc_put_request(req);
 621 }
 622
 623 /*
 624  * allocate a page vec, either directly, or if necessary, via a the
 625  * mempool.  we avoid the mempool if we can because req->r_num_pages
 626  * may be less than the maximum write size.
 627  */
 628 static void alloc_page_vec(struct ceph_fs_client *fsc,
 629                            struct ceph_osd_request *req)
 630 {
 631         req->r_pages = kmalloc(sizeof(struct page *) * req->r_num_pages,
 632                                GFP_NOFS);
 633         if (!req->r_pages) {
 634                 req->r_pages = mempool_alloc(fsc->wb_pagevec_pool, GFP_NOFS);
 635                 req->r_pages_from_pool = 1;
 636                 WARN_ON(!req->r_pages);
 637         }
 638 }
 639
 640 /*
 641  * initiate async writeback
 642  */
 643 static int ceph_writepages_start(struct address_space *mapping,
 644                                  struct writeback_control *wbc)
 645 {
 646         struct inode *inode = mapping->host;
 647         struct ceph_inode_info *ci = ceph_inode(inode);
 648         struct ceph_fs_client *fsc;
 649         pgoff_t index, start, end;
 650         int range_whole = 0;
 651         int should_loop = 1;
 652         pgoff_t max_pages = 0, max_pages_ever = 0;
 653         struct ceph_snap_context *snapc = NULL, *last_snapc = NULL, *pgsnapc;
 654         struct pagevec pvec;
 655         int done = 0;
 656         int rc = 0;
 657         unsigned wsize = 1 << inode->i_blkbits;
 658         struct ceph_osd_request *req = NULL;
 659         int do_sync;
 660         u64 snap_size = 0;
 661
 662         /*
 663          * Include a 'sync' in the OSD request if this is a data
 664          * integrity write (e.g., O_SYNC write or fsync()), or if our
 665          * cap is being revoked.
 666          */
 667         do_sync = wbc->sync_mode == WB_SYNC_ALL;
 668         if (ceph_caps_revoking(ci, CEPH_CAP_FILE_BUFFER))
 669                 do_sync = 1;
 670         dout("writepages_start %p dosync=%d (mode=%s)\n",
 671              inode, do_sync,
 672              wbc->sync_mode == WB_SYNC_NONE ? "NONE" :
 673              (wbc->sync_mode == WB_SYNC_ALL ? "ALL" : "HOLD"));
 674
 675         fsc = ceph_inode_to_client(inode);
 676         if (fsc->mount_state == CEPH_MOUNT_SHUTDOWN) {
 677                 pr_warning("writepage_start %p on forced umount\n", inode);
 678                 return -EIO; /* we're in a forced umount, don't write! */
 679         }
 680         if (fsc->mount_options->wsize && fsc->mount_options->wsize < wsize)
 681                 wsize = fsc->mount_options->wsize;
 682         if (wsize < PAGE_CACHE_SIZE)
 683                 wsize = PAGE_CACHE_SIZE;
 684         max_pages_ever = wsize >> PAGE_CACHE_SHIFT;
 685
 686         pagevec_init(&pvec, 0);
 687
 688         /* where to start/end? */
 689         if (wbc->range_cyclic) {
 690                 start = mapping->writeback_index; /* Start from prev offset */
 691                 end = -1;
 692                 dout(" cyclic, start at %lu\n", start);
 693         } else {
 694                 start = wbc->range_start >> PAGE_CACHE_SHIFT;
 695                 end = wbc->range_end >> PAGE_CACHE_SHIFT;
 696                 if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
 697                         range_whole = 1;
 698                 should_loop = 0;
 699                 dout(" not cyclic, %lu to %lu\n", start, end);
 700         }
 701         index = start;
 702
 703 retry:
 704         /* find oldest snap context with dirty data */
 705         ceph_put_snap_context(snapc);
 706         snapc = get_oldest_context(inode, &snap_size);
 707         if (!snapc) {
 708                 /* hmm, why does writepages get called when there
 709                    is no dirty data? */
 710                 dout(" no snap context with dirty data?\n");
 711                 goto out;
 712         }
 713         dout(" oldest snapc is %p seq %lld (%d snaps)\n",
 714              snapc, snapc->seq, snapc->num_snaps);
 715         if (last_snapc && snapc != last_snapc) {
 716                 /* if we switched to a newer snapc, restart our scan at the
 717                  * start of the original file range. */
 718                 dout("  snapc differs from last pass, restarting at %lu\n",
 719                      index);
 720                 index = start;
 721         }
 722         last_snapc = snapc;
 723
 724         while (!done && index <= end) {
 725                 unsigned i;
 726                 int first;
 727                 pgoff_t next;
 728                 int pvec_pages, locked_pages;
 729                 struct page *page;
 730                 int want;
 731                 u64 offset, len;
 732                 struct ceph_osd_request_head *reqhead;
 733                 struct ceph_osd_op *op;
 734                 long writeback_stat;
 735
 736                 next = 0;
 737                 locked_pages = 0;
 738                 max_pages = max_pages_ever;
 739
 740 get_more_pages:
 741                 first = -1;
 742                 want = min(end - index,
 743                            min((pgoff_t)PAGEVEC_SIZE,
 744                                max_pages - (pgoff_t)locked_pages) - 1)
 745                         + 1;
 746                 pvec_pages = pagevec_lookup_tag(&pvec, mapping, &index,
 747                                                 PAGECACHE_TAG_DIRTY,
 748                                                 want);
 749                 dout("pagevec_lookup_tag got %d\n", pvec_pages);
 750                 if (!pvec_pages && !locked_pages)
 751                         break;
 752                 for (i = 0; i < pvec_pages && locked_pages < max_pages; i++) {
 753                         page = pvec.pages[i];
 754                         dout("? %p idx %lu\n", page, page->index);
 755                         if (locked_pages == 0)
 756                                 lock_page(page);  /* first page */
 757                         else if (!trylock_page(page))
 758                                 break;
 759
 760                         /* only dirty pages, or our accounting breaks */
 761                         if (unlikely(!PageDirty(page)) ||
 762                             unlikely(page->mapping != mapping)) {
 763                                 dout("!dirty or !mapping %p\n", page);
 764                                 unlock_page(page);
 765                                 break;
 766                         }
 767                         if (!wbc->range_cyclic && page->index > end) {
 768                                 dout("end of range %p\n", page);
 769                                 done = 1;
 770                                 unlock_page(page);
 771                                 break;
 772                         }
 773                         if (next && (page->index != next)) {
 774                                 dout("not consecutive %p\n", page);
 775                                 unlock_page(page);
 776                                 break;
 777                         }
 778                         if (wbc->sync_mode != WB_SYNC_NONE) {
 779                                 dout("waiting on writeback %p\n", page);
 780                                 wait_on_page_writeback(page);
 781                         }
 782                         if ((snap_size && page_offset(page) > snap_size) ||
 783                             (!snap_size &&
 784                              page_offset(page) > i_size_read(inode))) {
 785                                 dout("%p page eof %llu\n", page, snap_size ?
 786                                      snap_size : i_size_read(inode));
 787                                 done = 1;
 788                                 unlock_page(page);
 789                                 break;
 790                         }
 791                         if (PageWriteback(page)) {
 792                                 dout("%p under writeback\n", page);
 793                                 unlock_page(page);
 794                                 break;
 795                         }
 796
 797                         /* only if matching snap context */
 798                         pgsnapc = (void *)page->private;
 799                         if (pgsnapc->seq > snapc->seq) {
 800                                 dout("page snapc %p %lld > oldest %p %lld\n",
 801                                      pgsnapc, pgsnapc->seq, snapc, snapc->seq);
 802                                 unlock_page(page);
 803                                 if (!locked_pages)
 804                                         continue; /* keep looking for snap */
 805                                 break;
 806                         }
 807
 808                         if (!clear_page_dirty_for_io(page)) {
 809                                 dout("%p !clear_page_dirty_for_io\n", page);
 810                                 unlock_page(page);
 811                                 break;
 812                         }
 813
 814                         /* ok */
 815                         if (locked_pages == 0) {
 816                                 /* prepare async write request */
 817                                 offset = (unsigned long long)page->index
 818                                         << PAGE_CACHE_SHIFT;
 819                                 len = wsize;
 820                                 req = ceph_osdc_new_request(&fsc->client->osdc,
 821                                             &ci->i_layout,
 822                                             ceph_vino(inode),
 823                                             offset, &len,
 824                                             CEPH_OSD_OP_WRITE,
 825                                             CEPH_OSD_FLAG_WRITE |
 826                                                     CEPH_OSD_FLAG_ONDISK,
 827                                             snapc, do_sync,
 828                                             ci->i_truncate_seq,
 829                                             ci->i_truncate_size,
 830                                             &inode->i_mtime, true, 1, 0);
 831
 832                                 if (!req) {
 833                                         rc = -ENOMEM;
 834                                         unlock_page(page);
 835                                         break;
 836                                 }
 837
 838                                 max_pages = req->r_num_pages;
 839
 840                                 alloc_page_vec(fsc, req);
 841                                 req->r_callback = writepages_finish;
 842                                 req->r_inode = inode;
 843                         }
 844
 845                         /* note position of first page in pvec */
 846                         if (first < 0)
 847                                 first = i;
 848                         dout("%p will write page %p idx %lu\n",
 849                              inode, page, page->index);
 850
 851                         writeback_stat =
 852                                atomic_long_inc_return(&fsc->writeback_count);
 853                         if (writeback_stat > CONGESTION_ON_THRESH(
 854                                     fsc->mount_options->congestion_kb)) {
 855                                 set_bdi_congested(&fsc->backing_dev_info,
 856                                                   BLK_RW_ASYNC);
 857                         }
 858
 859                         set_page_writeback(page);
 860                         req->r_pages[locked_pages] = page;
 861                         locked_pages++;
 862                         next = page->index + 1;
 863                 }
 864
 865                 /* did we get anything? */
 866                 if (!locked_pages)
 867                         goto release_pvec_pages;
 868                 if (i) {
 869                         int j;
 870                         BUG_ON(!locked_pages || first < 0);
 871
 872                         if (pvec_pages && i == pvec_pages &&
 873                             locked_pages < max_pages) {
 874                                 dout("reached end pvec, trying for more\n");
 875                                 pagevec_reinit(&pvec);
 876                                 goto get_more_pages;
 877                         }
 878
 879                         /* shift unused pages over in the pvec...  we
 880                          * will need to release them below. */
 881                         for (j = i; j < pvec_pages; j++) {
 882                                 dout(" pvec leftover page %p\n",
 883                                      pvec.pages[j]);
 884                                 pvec.pages[j-i+first] = pvec.pages[j];
 885                         }
 886                         pvec.nr -= i-first;
 887                 }
 888
 889                 /* submit the write */
 890                 offset = req->r_pages[0]->index << PAGE_CACHE_SHIFT;
 891                 len = min((snap_size ? snap_size : i_size_read(inode)) - offset,
 892                           (u64)locked_pages << PAGE_CACHE_SHIFT);
 893                 dout("writepages got %d pages at %llu~%llu\n",
 894                      locked_pages, offset, len);
 895
 896                 /* revise final length, page count */
 897                 req->r_num_pages = locked_pages;
 898                 reqhead = req->r_request->front.iov_base;
 899                 op = (void *)(reqhead + 1);
 900                 op->extent.length = cpu_to_le64(len);
 901                 op->payload_len = cpu_to_le32(len);
 902                 req->r_request->hdr.data_len = cpu_to_le32(len);
 903
 904                 rc = ceph_osdc_start_request(&fsc->client->osdc, req, true);
 905                 BUG_ON(rc);
 906                 req = NULL;
 907
 908                 /* continue? */
 909                 index = next;
 910                 wbc->nr_to_write -= locked_pages;
 911                 if (wbc->nr_to_write <= 0)
 912                         done = 1;
 913
 914 release_pvec_pages:
 915                 dout("pagevec_release on %d pages (%p)\n", (int)pvec.nr,
 916                      pvec.nr ? pvec.pages[0] : NULL);
 917                 pagevec_release(&pvec);
 918
 919                 if (locked_pages && !done)
 920                         goto retry;
 921         }
 922
 923         if (should_loop && !done) {
 924                 /* more to do; loop back to beginning of file */
 925                 dout("writepages looping back to beginning of file\n");
 926                 should_loop = 0;
 927                 index = 0;
 928                 goto retry;
 929         }
 930
 931         if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
 932                 mapping->writeback_index = index;
 933
 934 out:
 935         if (req)
 936                 ceph_osdc_put_request(req);
 937         ceph_put_snap_context(snapc);
 938         dout("writepages done, rc = %d\n", rc);
 939         return rc;
 940 }
 941
 942
 943
 944 /*
 945  * See if a given @snapc is either writeable, or already written.
 946  */
 947 static int context_is_writeable_or_written(struct inode *inode,
 948                                            struct ceph_snap_context *snapc)
 949 {
 950         struct ceph_snap_context *oldest = get_oldest_context(inode, NULL);
 951         int ret = !oldest || snapc->seq <= oldest->seq;
 952
 953         ceph_put_snap_context(oldest);
 954         return ret;
 955 }
 956
 957 /*
 958  * We are only allowed to write into/dirty the page if the page is
 959  * clean, or already dirty within the same snap context.
 960  *
 961  * called with page locked.
 962  * return success with page locked,
 963  * or any failure (incl -EAGAIN) with page unlocked.
 964  */
 965 static int ceph_update_writeable_page(struct file *file,
 966                             loff_t pos, unsigned len,
 967                             struct page *page)
 968 {
 969         struct inode *inode = file->f_dentry->d_inode;
 970         struct ceph_inode_info *ci = ceph_inode(inode);
 971         struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
 972         loff_t page_off = pos & PAGE_CACHE_MASK;
 973         int pos_in_page = pos & ~PAGE_CACHE_MASK;
 974         int end_in_page = pos_in_page + len;
 975         loff_t i_size;
 976         int r;
 977         struct ceph_snap_context *snapc, *oldest;
 978
 979 retry_locked:
 980         /* writepages currently holds page lock, but if we change that later, */
 981         wait_on_page_writeback(page);
 982
 983         /* check snap context */
 984         BUG_ON(!ci->i_snap_realm);
 985         down_read(&mdsc->snap_rwsem);
 986         BUG_ON(!ci->i_snap_realm->cached_context);
 987         snapc = (void *)page->private;
 988         if (snapc && snapc != ci->i_head_snapc) {
 989                 /*
 990                  * this page is already dirty in another (older) snap
 991                  * context!  is it writeable now?
 992                  */
 993                 oldest = get_oldest_context(inode, NULL);
 994                 up_read(&mdsc->snap_rwsem);
 995
 996                 if (snapc->seq > oldest->seq) {
 997                         ceph_put_snap_context(oldest);
 998                         dout(" page %p snapc %p not current or oldest\n",
 999                              page, snapc);
1000                         /*
1001                          * queue for writeback, and wait for snapc to
1002                          * be writeable or written
1003                          */
1004                         snapc = ceph_get_snap_context(snapc);
1005                         unlock_page(page);
1006                         ceph_queue_writeback(inode);
1007                         r = wait_event_interruptible(ci->i_cap_wq,
1008                                context_is_writeable_or_written(inode, snapc));
1009                         ceph_put_snap_context(snapc);
1010                         if (r == -ERESTARTSYS)
1011                                 return r;
1012                         return -EAGAIN;
1013                 }
1014                 ceph_put_snap_context(oldest);
1015
1016                 /* yay, writeable, do it now (without dropping page lock) */
1017                 dout(" page %p snapc %p not current, but oldest\n",
1018                      page, snapc);
1019                 if (!clear_page_dirty_for_io(page))
1020                         goto retry_locked;
1021                 r = writepage_nounlock(page, NULL);
1022                 if (r < 0)
1023                         goto fail_nosnap;
1024                 goto retry_locked;
1025         }
1026
1027         if (PageUptodate(page)) {
1028                 dout(" page %p already uptodate\n", page);
1029                 return 0;
1030         }
1031
1032         /* full page? */
1033         if (pos_in_page == 0 && len == PAGE_CACHE_SIZE)
1034                 return 0;
1035
1036         /* past end of file? */
1037         i_size = inode->i_size;   /* caller holds i_mutex */
1038
1039         if (i_size + len > inode->i_sb->s_maxbytes) {
1040                 /* file is too big */
1041                 r = -EINVAL;
1042                 goto fail;
1043         }
1044
1045         if (page_off >= i_size ||
1046             (pos_in_page == 0 && (pos+len) >= i_size &&
1047              end_in_page - pos_in_page != PAGE_CACHE_SIZE)) {
1048                 dout(" zeroing %p 0 - %d and %d - %d\n",
1049                      page, pos_in_page, end_in_page, (int)PAGE_CACHE_SIZE);
1050                 zero_user_segments(page,
1051                                    0, pos_in_page,
1052                                    end_in_page, PAGE_CACHE_SIZE);
1053                 return 0;
1054         }
1055
1056         /* we need to read it. */
1057         up_read(&mdsc->snap_rwsem);
1058         r = readpage_nounlock(file, page);
1059         if (r < 0)
1060                 goto fail_nosnap;
1061         goto retry_locked;
1062
1063 fail:
1064         up_read(&mdsc->snap_rwsem);
1065 fail_nosnap:
1066         unlock_page(page);
1067         return r;
1068 }
1069
1070 /*
1071  * We are only allowed to write into/dirty the page if the page is
1072  * clean, or already dirty within the same snap context.
1073  */
1074 static int ceph_write_begin(struct file *file, struct address_space *mapping,
1075                             loff_t pos, unsigned len, unsigned flags,
1076                             struct page **pagep, void **fsdata)
1077 {
1078         struct inode *inode = file->f_dentry->d_inode;
1079         struct page *page;
1080         pgoff_t index = pos >> PAGE_CACHE_SHIFT;
1081         int r;
1082
1083         do {
1084                 /* get a page */
1085                 page = grab_cache_page_write_begin(mapping, index, 0);
1086                 if (!page)
1087                         return -ENOMEM;
1088                 *pagep = page;
1089
1090                 dout("write_begin file %p inode %p page %p %d~%d\n", file,
1091                      inode, page, (int)pos, (int)len);
1092
1093                 r = ceph_update_writeable_page(file, pos, len, page);
1094         } while (r == -EAGAIN);
1095
1096         return r;
1097 }
1098
1099 /*
1100  * we don't do anything in here that simple_write_end doesn't do
1101  * except adjust dirty page accounting and drop read lock on
1102  * mdsc->snap_rwsem.
1103  */
1104 static int ceph_write_end(struct file *file, struct address_space *mapping,
1105                           loff_t pos, unsigned len, unsigned copied,
1106                           struct page *page, void *fsdata)
1107 {
1108         struct inode *inode = file->f_dentry->d_inode;
1109         struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
1110         struct ceph_mds_client *mdsc = fsc->mdsc;
1111         unsigned from = pos & (PAGE_CACHE_SIZE - 1);
1112         int check_cap = 0;
1113
1114         dout("write_end file %p inode %p page %p %d~%d (%d)\n", file,
1115              inode, page, (int)pos, (int)copied, (int)len);
1116
1117         /* zero the stale part of the page if we did a short copy */
1118         if (copied < len)
1119                 zero_user_segment(page, from+copied, len);
1120
1121         /* did file size increase? */
1122         /* (no need for i_size_read(); we caller holds i_mutex */
1123         if (pos+copied > inode->i_size)
1124                 check_cap = ceph_inode_set_size(inode, pos+copied);
1125
1126         if (!PageUptodate(page))
1127                 SetPageUptodate(page);
1128
1129         set_page_dirty(page);
1130
1131         unlock_page(page);
1132         up_read(&mdsc->snap_rwsem);
1133         page_cache_release(page);
1134
1135         if (check_cap)
1136                 ceph_check_caps(ceph_inode(inode), CHECK_CAPS_AUTHONLY, NULL);
1137
1138         return copied;
1139 }
1140
1141 /*
1142  * we set .direct_IO to indicate direct io is supported, but since we
1143  * intercept O_DIRECT reads and writes early, this function should
1144  * never get called.
1145  */
1146 static ssize_t ceph_direct_io(int rw, struct kiocb *iocb,
1147                               const struct iovec *iov,
1148                               loff_t pos, unsigned long nr_segs)
1149 {
1150         WARN_ON(1);
1151         return -EINVAL;
1152 }
1153
1154 const struct address_space_operations ceph_aops = {
1155         .readpage = ceph_readpage,
1156         .readpages = ceph_readpages,
1157         .writepage = ceph_writepage,
1158         .writepages = ceph_writepages_start,
1159         .write_begin = ceph_write_begin,
1160         .write_end = ceph_write_end,
1161         .set_page_dirty = ceph_set_page_dirty,
1162         .invalidatepage = ceph_invalidatepage,
1163         .releasepage = ceph_releasepage,
1164         .direct_IO = ceph_direct_io,
1165 };
1166
1167
1168 /*
1169  * vm ops
1170  */
1171
1172 /*
1173  * Reuse write_begin here for simplicity.
1174  */
1175 static int ceph_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
1176 {
1177         struct inode *inode = vma->vm_file->f_dentry->d_inode;
1178         struct page *page = vmf->page;
1179         struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
1180         loff_t off = page->index << PAGE_CACHE_SHIFT;
1181         loff_t size, len;
1182         int ret;
1183
1184         size = i_size_read(inode);
1185         if (off + PAGE_CACHE_SIZE <= size)
1186                 len = PAGE_CACHE_SIZE;
1187         else
1188                 len = size & ~PAGE_CACHE_MASK;
1189
1190         dout("page_mkwrite %p %llu~%llu page %p idx %lu\n", inode,
1191              off, len, page, page->index);
1192
1193         lock_page(page);
1194
1195         ret = VM_FAULT_NOPAGE;
1196         if ((off > size) ||
1197             (page->mapping != inode->i_mapping))
1198                 goto out;
1199
1200         ret = ceph_update_writeable_page(vma->vm_file, off, len, page);
1201         if (ret == 0) {
1202                 /* success.  we'll keep the page locked. */
1203                 set_page_dirty(page);
1204                 up_read(&mdsc->snap_rwsem);
1205                 ret = VM_FAULT_LOCKED;
1206         } else {
1207                 if (ret == -ENOMEM)
1208                         ret = VM_FAULT_OOM;
1209                 else
1210                         ret = VM_FAULT_SIGBUS;
1211         }
1212 out:
1213         dout("page_mkwrite %p %llu~%llu = %d\n", inode, off, len, ret);
1214         if (ret != VM_FAULT_LOCKED)
1215                 unlock_page(page);
1216         return ret;
1217 }
1218
1219 static struct vm_operations_struct ceph_vmops = {
1220         .fault          = filemap_fault,
1221         .page_mkwrite   = ceph_page_mkwrite,
1222 };
1223
1224 int ceph_mmap(struct file *file, struct vm_area_struct *vma)
1225 {
1226         struct address_space *mapping = file->f_mapping;
1227
1228         if (!mapping->a_ops->readpage)
1229                 return -ENOEXEC;
1230         file_accessed(file);
1231         vma->vm_ops = &ceph_vmops;
1232         vma->vm_flags |= VM_CAN_NONLINEAR;
1233         return 0;
1234 }