fs/nfs/direct.c

   1 // SPDX-License-Identifier: GPL-2.0-only
   2 /*
   3  * linux/fs/nfs/direct.c
   4  *
   5  * Copyright (C) 2003 by Chuck Lever <cel@netapp.com>
   6  *
   7  * High-performance uncached I/O for the Linux NFS client
   8  *
   9  * There are important applications whose performance or correctness
  10  * depends on uncached access to file data.  Database clusters
  11  * (multiple copies of the same instance running on separate hosts)
  12  * implement their own cache coherency protocol that subsumes file
  13  * system cache protocols.  Applications that process datasets
  14  * considerably larger than the client's memory do not always benefit
  15  * from a local cache.  A streaming video server, for instance, has no
  16  * need to cache the contents of a file.
  17  *
  18  * When an application requests uncached I/O, all read and write requests
  19  * are made directly to the server; data stored or fetched via these
  20  * requests is not cached in the Linux page cache.  The client does not
  21  * correct unaligned requests from applications.  All requested bytes are
  22  * held on permanent storage before a direct write system call returns to
  23  * an application.
  24  *
  25  * Solaris implements an uncached I/O facility called directio() that
  26  * is used for backups and sequential I/O to very large files.  Solaris
  27  * also supports uncaching whole NFS partitions with "-o forcedirectio,"
  28  * an undocumented mount option.
  29  *
  30  * Designed by Jeff Kimmel, Chuck Lever, and Trond Myklebust, with
  31  * help from Andrew Morton.
  32  *
  33  * 18 Dec 2001  Initial implementation for 2.4  --cel
  34  * 08 Jul 2002  Version for 2.4.19, with bug fixes --trondmy
  35  * 08 Jun 2003  Port to 2.5 APIs  --cel
  36  * 31 Mar 2004  Handle direct I/O without VFS support  --cel
  37  * 15 Sep 2004  Parallel async reads  --cel
  38  * 04 May 2005  support O_DIRECT with aio  --cel
  39  *
  40  */
  41
  42 #include <linux/errno.h>
  43 #include <linux/sched.h>
  44 #include <linux/kernel.h>
  45 #include <linux/file.h>
  46 #include <linux/pagemap.h>
  47 #include <linux/kref.h>
  48 #include <linux/slab.h>
  49 #include <linux/task_io_accounting_ops.h>
  50 #include <linux/module.h>
  51
  52 #include <linux/nfs_fs.h>
  53 #include <linux/nfs_page.h>
  54 #include <linux/sunrpc/clnt.h>
  55
  56 #include <linux/uaccess.h>
  57 #include <linux/atomic.h>
  58
  59 #include "internal.h"
  60 #include "iostat.h"
  61 #include "pnfs.h"
  62 #include "fscache.h"
  63 #include "nfstrace.h"
  64
  65 #define NFSDBG_FACILITY         NFSDBG_VFS
  66
  67 static struct kmem_cache *nfs_direct_cachep;
  68
  69 static const struct nfs_pgio_completion_ops nfs_direct_write_completion_ops;
  70 static const struct nfs_commit_completion_ops nfs_direct_commit_completion_ops;
  71 static void nfs_direct_write_complete(struct nfs_direct_req *dreq);
  72 static void nfs_direct_write_schedule_work(struct work_struct *work);
  73
  74 static inline void get_dreq(struct nfs_direct_req *dreq)
  75 {
  76         atomic_inc(&dreq->io_count);
  77 }
  78
  79 static inline int put_dreq(struct nfs_direct_req *dreq)
  80 {
  81         return atomic_dec_and_test(&dreq->io_count);
  82 }
  83
  84 static void
  85 nfs_direct_handle_truncated(struct nfs_direct_req *dreq,
  86                             const struct nfs_pgio_header *hdr,
  87                             ssize_t dreq_len)
  88 {
  89         if (!(test_bit(NFS_IOHDR_ERROR, &hdr->flags) ||
  90               test_bit(NFS_IOHDR_EOF, &hdr->flags)))
  91                 return;
  92         if (dreq->max_count >= dreq_len) {
  93                 dreq->max_count = dreq_len;
  94                 if (dreq->count > dreq_len)
  95                         dreq->count = dreq_len;
  96
  97                 if (test_bit(NFS_IOHDR_ERROR, &hdr->flags))
  98                         dreq->error = hdr->error;
  99                 else /* Clear outstanding error if this is EOF */
 100                         dreq->error = 0;
 101         }
 102 }
 103
 104 static void
 105 nfs_direct_count_bytes(struct nfs_direct_req *dreq,
 106                        const struct nfs_pgio_header *hdr)
 107 {
 108         loff_t hdr_end = hdr->io_start + hdr->good_bytes;
 109         ssize_t dreq_len = 0;
 110
 111         if (hdr_end > dreq->io_start)
 112                 dreq_len = hdr_end - dreq->io_start;
 113
 114         nfs_direct_handle_truncated(dreq, hdr, dreq_len);
 115
 116         if (dreq_len > dreq->max_count)
 117                 dreq_len = dreq->max_count;
 118
 119         if (dreq->count < dreq_len)
 120                 dreq->count = dreq_len;
 121 }
 122
 123 /**
 124  * nfs_swap_rw - NFS address space operation for swap I/O
 125  * @iocb: target I/O control block
 126  * @iter: I/O buffer
 127  *
 128  * Perform IO to the swap-file.  This is much like direct IO.
 129  */
 130 int nfs_swap_rw(struct kiocb *iocb, struct iov_iter *iter)
 131 {
 132         ssize_t ret;
 133
 134         VM_BUG_ON(iov_iter_count(iter) != PAGE_SIZE);
 135
 136         if (iov_iter_rw(iter) == READ)
 137                 ret = nfs_file_direct_read(iocb, iter, true);
 138         else
 139                 ret = nfs_file_direct_write(iocb, iter, true);
 140         if (ret < 0)
 141                 return ret;
 142         return 0;
 143 }
 144
 145 static void nfs_direct_release_pages(struct page **pages, unsigned int npages)
 146 {
 147         unsigned int i;
 148         for (i = 0; i < npages; i++)
 149                 put_page(pages[i]);
 150 }
 151
 152 void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo,
 153                               struct nfs_direct_req *dreq)
 154 {
 155         cinfo->inode = dreq->inode;
 156         cinfo->mds = &dreq->mds_cinfo;
 157         cinfo->ds = &dreq->ds_cinfo;
 158         cinfo->dreq = dreq;
 159         cinfo->completion_ops = &nfs_direct_commit_completion_ops;
 160 }
 161
 162 static inline struct nfs_direct_req *nfs_direct_req_alloc(void)
 163 {
 164         struct nfs_direct_req *dreq;
 165
 166         dreq = kmem_cache_zalloc(nfs_direct_cachep, GFP_KERNEL);
 167         if (!dreq)
 168                 return NULL;
 169
 170         kref_init(&dreq->kref);
 171         kref_get(&dreq->kref);
 172         init_completion(&dreq->completion);
 173         INIT_LIST_HEAD(&dreq->mds_cinfo.list);
 174         pnfs_init_ds_commit_info(&dreq->ds_cinfo);
 175         INIT_WORK(&dreq->work, nfs_direct_write_schedule_work);
 176         spin_lock_init(&dreq->lock);
 177
 178         return dreq;
 179 }
 180
 181 static void nfs_direct_req_free(struct kref *kref)
 182 {
 183         struct nfs_direct_req *dreq = container_of(kref, struct nfs_direct_req, kref);
 184
 185         pnfs_release_ds_info(&dreq->ds_cinfo, dreq->inode);
 186         if (dreq->l_ctx != NULL)
 187                 nfs_put_lock_context(dreq->l_ctx);
 188         if (dreq->ctx != NULL)
 189                 put_nfs_open_context(dreq->ctx);
 190         kmem_cache_free(nfs_direct_cachep, dreq);
 191 }
 192
 193 static void nfs_direct_req_release(struct nfs_direct_req *dreq)
 194 {
 195         kref_put(&dreq->kref, nfs_direct_req_free);
 196 }
 197
 198 ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq)
 199 {
 200         return dreq->bytes_left;
 201 }
 202 EXPORT_SYMBOL_GPL(nfs_dreq_bytes_left);
 203
 204 /*
 205  * Collects and returns the final error value/byte-count.
 206  */
 207 static ssize_t nfs_direct_wait(struct nfs_direct_req *dreq)
 208 {
 209         ssize_t result = -EIOCBQUEUED;
 210
 211         /* Async requests don't wait here */
 212         if (dreq->iocb)
 213                 goto out;
 214
 215         result = wait_for_completion_killable(&dreq->completion);
 216
 217         if (!result) {
 218                 result = dreq->count;
 219                 WARN_ON_ONCE(dreq->count < 0);
 220         }
 221         if (!result)
 222                 result = dreq->error;
 223
 224 out:
 225         return (ssize_t) result;
 226 }
 227
 228 /*
 229  * Synchronous I/O uses a stack-allocated iocb.  Thus we can't trust
 230  * the iocb is still valid here if this is a synchronous request.
 231  */
 232 static void nfs_direct_complete(struct nfs_direct_req *dreq)
 233 {
 234         struct inode *inode = dreq->inode;
 235
 236         inode_dio_end(inode);
 237
 238         if (dreq->iocb) {
 239                 long res = (long) dreq->error;
 240                 if (dreq->count != 0) {
 241                         res = (long) dreq->count;
 242                         WARN_ON_ONCE(dreq->count < 0);
 243                 }
 244                 dreq->iocb->ki_complete(dreq->iocb, res);
 245         }
 246
 247         complete(&dreq->completion);
 248
 249         nfs_direct_req_release(dreq);
 250 }
 251
 252 static void nfs_direct_read_completion(struct nfs_pgio_header *hdr)
 253 {
 254         unsigned long bytes = 0;
 255         struct nfs_direct_req *dreq = hdr->dreq;
 256
 257         spin_lock(&dreq->lock);
 258         if (test_bit(NFS_IOHDR_REDO, &hdr->flags)) {
 259                 spin_unlock(&dreq->lock);
 260                 goto out_put;
 261         }
 262
 263         nfs_direct_count_bytes(dreq, hdr);
 264         spin_unlock(&dreq->lock);
 265
 266         while (!list_empty(&hdr->pages)) {
 267                 struct nfs_page *req = nfs_list_entry(hdr->pages.next);
 268                 struct page *page = req->wb_page;
 269
 270                 if (!PageCompound(page) && bytes < hdr->good_bytes &&
 271                     (dreq->flags == NFS_ODIRECT_SHOULD_DIRTY))
 272                         set_page_dirty(page);
 273                 bytes += req->wb_bytes;
 274                 nfs_list_remove_request(req);
 275                 nfs_release_request(req);
 276         }
 277 out_put:
 278         if (put_dreq(dreq))
 279                 nfs_direct_complete(dreq);
 280         hdr->release(hdr);
 281 }
 282
 283 static void nfs_read_sync_pgio_error(struct list_head *head, int error)
 284 {
 285         struct nfs_page *req;
 286
 287         while (!list_empty(head)) {
 288                 req = nfs_list_entry(head->next);
 289                 nfs_list_remove_request(req);
 290                 nfs_release_request(req);
 291         }
 292 }
 293
 294 static void nfs_direct_pgio_init(struct nfs_pgio_header *hdr)
 295 {
 296         get_dreq(hdr->dreq);
 297 }
 298
 299 static const struct nfs_pgio_completion_ops nfs_direct_read_completion_ops = {
 300         .error_cleanup = nfs_read_sync_pgio_error,
 301         .init_hdr = nfs_direct_pgio_init,
 302         .completion = nfs_direct_read_completion,
 303 };
 304
 305 /*
 306  * For each rsize'd chunk of the user's buffer, dispatch an NFS READ
 307  * operation.  If nfs_readdata_alloc() or get_user_pages() fails,
 308  * bail and stop sending more reads.  Read length accounting is
 309  * handled automatically by nfs_direct_read_result().  Otherwise, if
 310  * no requests have been sent, just return an error.
 311  */
 312
 313 static ssize_t nfs_direct_read_schedule_iovec(struct nfs_direct_req *dreq,
 314                                               struct iov_iter *iter,
 315                                               loff_t pos)
 316 {
 317         struct nfs_pageio_descriptor desc;
 318         struct inode *inode = dreq->inode;
 319         ssize_t result = -EINVAL;
 320         size_t requested_bytes = 0;
 321         size_t rsize = max_t(size_t, NFS_SERVER(inode)->rsize, PAGE_SIZE);
 322
 323         nfs_pageio_init_read(&desc, dreq->inode, false,
 324                              &nfs_direct_read_completion_ops);
 325         get_dreq(dreq);
 326         desc.pg_dreq = dreq;
 327         inode_dio_begin(inode);
 328
 329         while (iov_iter_count(iter)) {
 330                 struct page **pagevec;
 331                 size_t bytes;
 332                 size_t pgbase;
 333                 unsigned npages, i;
 334
 335                 result = iov_iter_get_pages_alloc2(iter, &pagevec,
 336                                                   rsize, &pgbase);
 337                 if (result < 0)
 338                         break;
 339
 340                 bytes = result;
 341                 npages = (result + pgbase + PAGE_SIZE - 1) / PAGE_SIZE;
 342                 for (i = 0; i < npages; i++) {
 343                         struct nfs_page *req;
 344                         unsigned int req_len = min_t(size_t, bytes, PAGE_SIZE - pgbase);
 345                         /* XXX do we need to do the eof zeroing found in async_filler? */
 346                         req = nfs_create_request(dreq->ctx, pagevec[i],
 347                                                  pgbase, req_len);
 348                         if (IS_ERR(req)) {
 349                                 result = PTR_ERR(req);
 350                                 break;
 351                         }
 352                         req->wb_index = pos >> PAGE_SHIFT;
 353                         req->wb_offset = pos & ~PAGE_MASK;
 354                         if (!nfs_pageio_add_request(&desc, req)) {
 355                                 result = desc.pg_error;
 356                                 nfs_release_request(req);
 357                                 break;
 358                         }
 359                         pgbase = 0;
 360                         bytes -= req_len;
 361                         requested_bytes += req_len;
 362                         pos += req_len;
 363                         dreq->bytes_left -= req_len;
 364                 }
 365                 nfs_direct_release_pages(pagevec, npages);
 366                 kvfree(pagevec);
 367                 if (result < 0)
 368                         break;
 369         }
 370
 371         nfs_pageio_complete(&desc);
 372
 373         /*
 374          * If no bytes were started, return the error, and let the
 375          * generic layer handle the completion.
 376          */
 377         if (requested_bytes == 0) {
 378                 inode_dio_end(inode);
 379                 nfs_direct_req_release(dreq);
 380                 return result < 0 ? result : -EIO;
 381         }
 382
 383         if (put_dreq(dreq))
 384                 nfs_direct_complete(dreq);
 385         return requested_bytes;
 386 }
 387
 388 /**
 389  * nfs_file_direct_read - file direct read operation for NFS files
 390  * @iocb: target I/O control block
 391  * @iter: vector of user buffers into which to read data
 392  * @swap: flag indicating this is swap IO, not O_DIRECT IO
 393  *
 394  * We use this function for direct reads instead of calling
 395  * generic_file_aio_read() in order to avoid gfar's check to see if
 396  * the request starts before the end of the file.  For that check
 397  * to work, we must generate a GETATTR before each direct read, and
 398  * even then there is a window between the GETATTR and the subsequent
 399  * READ where the file size could change.  Our preference is simply
 400  * to do all reads the application wants, and the server will take
 401  * care of managing the end of file boundary.
 402  *
 403  * This function also eliminates unnecessarily updating the file's
 404  * atime locally, as the NFS server sets the file's atime, and this
 405  * client must read the updated atime from the server back into its
 406  * cache.
 407  */
 408 ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter,
 409                              bool swap)
 410 {
 411         struct file *file = iocb->ki_filp;
 412         struct address_space *mapping = file->f_mapping;
 413         struct inode *inode = mapping->host;
 414         struct nfs_direct_req *dreq;
 415         struct nfs_lock_context *l_ctx;
 416         ssize_t result, requested;
 417         size_t count = iov_iter_count(iter);
 418         nfs_add_stats(mapping->host, NFSIOS_DIRECTREADBYTES, count);
 419
 420         dfprintk(FILE, "NFS: direct read(%pD2, %zd@%Ld)\n",
 421                 file, count, (long long) iocb->ki_pos);
 422
 423         result = 0;
 424         if (!count)
 425                 goto out;
 426
 427         task_io_account_read(count);
 428
 429         result = -ENOMEM;
 430         dreq = nfs_direct_req_alloc();
 431         if (dreq == NULL)
 432                 goto out;
 433
 434         dreq->inode = inode;
 435         dreq->bytes_left = dreq->max_count = count;
 436         dreq->io_start = iocb->ki_pos;
 437         dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp));
 438         l_ctx = nfs_get_lock_context(dreq->ctx);
 439         if (IS_ERR(l_ctx)) {
 440                 result = PTR_ERR(l_ctx);
 441                 nfs_direct_req_release(dreq);
 442                 goto out_release;
 443         }
 444         dreq->l_ctx = l_ctx;
 445         if (!is_sync_kiocb(iocb))
 446                 dreq->iocb = iocb;
 447
 448         if (user_backed_iter(iter))
 449                 dreq->flags = NFS_ODIRECT_SHOULD_DIRTY;
 450
 451         if (!swap)
 452                 nfs_start_io_direct(inode);
 453
 454         NFS_I(inode)->read_io += count;
 455         requested = nfs_direct_read_schedule_iovec(dreq, iter, iocb->ki_pos);
 456
 457         if (!swap)
 458                 nfs_end_io_direct(inode);
 459
 460         if (requested > 0) {
 461                 result = nfs_direct_wait(dreq);
 462                 if (result > 0) {
 463                         requested -= result;
 464                         iocb->ki_pos += result;
 465                 }
 466                 iov_iter_revert(iter, requested);
 467         } else {
 468                 result = requested;
 469         }
 470
 471 out_release:
 472         nfs_direct_req_release(dreq);
 473 out:
 474         return result;
 475 }
 476
 477 static void
 478 nfs_direct_join_group(struct list_head *list, struct inode *inode)
 479 {
 480         struct nfs_page *req, *next;
 481
 482         list_for_each_entry(req, list, wb_list) {
 483                 if (req->wb_head != req || req->wb_this_page == req)
 484                         continue;
 485                 for (next = req->wb_this_page;
 486                                 next != req->wb_head;
 487                                 next = next->wb_this_page) {
 488                         nfs_list_remove_request(next);
 489                         nfs_release_request(next);
 490                 }
 491                 nfs_join_page_group(req, inode);
 492         }
 493 }
 494
 495 static void
 496 nfs_direct_write_scan_commit_list(struct inode *inode,
 497                                   struct list_head *list,
 498                                   struct nfs_commit_info *cinfo)
 499 {
 500         mutex_lock(&NFS_I(cinfo->inode)->commit_mutex);
 501         pnfs_recover_commit_reqs(list, cinfo);
 502         nfs_scan_commit_list(&cinfo->mds->list, list, cinfo, 0);
 503         mutex_unlock(&NFS_I(cinfo->inode)->commit_mutex);
 504 }
 505
 506 static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq)
 507 {
 508         struct nfs_pageio_descriptor desc;
 509         struct nfs_page *req, *tmp;
 510         LIST_HEAD(reqs);
 511         struct nfs_commit_info cinfo;
 512         LIST_HEAD(failed);
 513
 514         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 515         nfs_direct_write_scan_commit_list(dreq->inode, &reqs, &cinfo);
 516
 517         nfs_direct_join_group(&reqs, dreq->inode);
 518
 519         dreq->count = 0;
 520         dreq->max_count = 0;
 521         list_for_each_entry(req, &reqs, wb_list)
 522                 dreq->max_count += req->wb_bytes;
 523         nfs_clear_pnfs_ds_commit_verifiers(&dreq->ds_cinfo);
 524         get_dreq(dreq);
 525
 526         nfs_pageio_init_write(&desc, dreq->inode, FLUSH_STABLE, false,
 527                               &nfs_direct_write_completion_ops);
 528         desc.pg_dreq = dreq;
 529
 530         list_for_each_entry_safe(req, tmp, &reqs, wb_list) {
 531                 /* Bump the transmission count */
 532                 req->wb_nio++;
 533                 if (!nfs_pageio_add_request(&desc, req)) {
 534                         nfs_list_move_request(req, &failed);
 535                         spin_lock(&cinfo.inode->i_lock);
 536                         dreq->flags = 0;
 537                         if (desc.pg_error < 0)
 538                                 dreq->error = desc.pg_error;
 539                         else
 540                                 dreq->error = -EIO;
 541                         spin_unlock(&cinfo.inode->i_lock);
 542                 }
 543                 nfs_release_request(req);
 544         }
 545         nfs_pageio_complete(&desc);
 546
 547         while (!list_empty(&failed)) {
 548                 req = nfs_list_entry(failed.next);
 549                 nfs_list_remove_request(req);
 550                 nfs_unlock_and_release_request(req);
 551         }
 552
 553         if (put_dreq(dreq))
 554                 nfs_direct_write_complete(dreq);
 555 }
 556
 557 static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 558 {
 559         const struct nfs_writeverf *verf = data->res.verf;
 560         struct nfs_direct_req *dreq = data->dreq;
 561         struct nfs_commit_info cinfo;
 562         struct nfs_page *req;
 563         int status = data->task.tk_status;
 564
 565         trace_nfs_direct_commit_complete(dreq);
 566
 567         if (status < 0) {
 568                 /* Errors in commit are fatal */
 569                 dreq->error = status;
 570                 dreq->max_count = 0;
 571                 dreq->count = 0;
 572                 dreq->flags = NFS_ODIRECT_DONE;
 573         } else {
 574                 status = dreq->error;
 575         }
 576
 577         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 578
 579         while (!list_empty(&data->pages)) {
 580                 req = nfs_list_entry(data->pages.next);
 581                 nfs_list_remove_request(req);
 582                 if (status >= 0 && !nfs_write_match_verf(verf, req)) {
 583                         dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 584                         /*
 585                          * Despite the reboot, the write was successful,
 586                          * so reset wb_nio.
 587                          */
 588                         req->wb_nio = 0;
 589                         nfs_mark_request_commit(req, NULL, &cinfo, 0);
 590                 } else /* Error or match */
 591                         nfs_release_request(req);
 592                 nfs_unlock_and_release_request(req);
 593         }
 594
 595         if (nfs_commit_end(cinfo.mds))
 596                 nfs_direct_write_complete(dreq);
 597 }
 598
 599 static void nfs_direct_resched_write(struct nfs_commit_info *cinfo,
 600                 struct nfs_page *req)
 601 {
 602         struct nfs_direct_req *dreq = cinfo->dreq;
 603
 604         trace_nfs_direct_resched_write(dreq);
 605
 606         spin_lock(&dreq->lock);
 607         if (dreq->flags != NFS_ODIRECT_DONE)
 608                 dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 609         spin_unlock(&dreq->lock);
 610         nfs_mark_request_commit(req, NULL, cinfo, 0);
 611 }
 612
 613 static const struct nfs_commit_completion_ops nfs_direct_commit_completion_ops = {
 614         .completion = nfs_direct_commit_complete,
 615         .resched_write = nfs_direct_resched_write,
 616 };
 617
 618 static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq)
 619 {
 620         int res;
 621         struct nfs_commit_info cinfo;
 622         LIST_HEAD(mds_list);
 623
 624         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 625         nfs_scan_commit(dreq->inode, &mds_list, &cinfo);
 626         res = nfs_generic_commit_list(dreq->inode, &mds_list, 0, &cinfo);
 627         if (res < 0) /* res == -ENOMEM */
 628                 nfs_direct_write_reschedule(dreq);
 629 }
 630
 631 static void nfs_direct_write_clear_reqs(struct nfs_direct_req *dreq)
 632 {
 633         struct nfs_commit_info cinfo;
 634         struct nfs_page *req;
 635         LIST_HEAD(reqs);
 636
 637         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 638         nfs_direct_write_scan_commit_list(dreq->inode, &reqs, &cinfo);
 639
 640         while (!list_empty(&reqs)) {
 641                 req = nfs_list_entry(reqs.next);
 642                 nfs_list_remove_request(req);
 643                 nfs_release_request(req);
 644                 nfs_unlock_and_release_request(req);
 645         }
 646 }
 647
 648 static void nfs_direct_write_schedule_work(struct work_struct *work)
 649 {
 650         struct nfs_direct_req *dreq = container_of(work, struct nfs_direct_req, work);
 651         int flags = dreq->flags;
 652
 653         dreq->flags = 0;
 654         switch (flags) {
 655                 case NFS_ODIRECT_DO_COMMIT:
 656                         nfs_direct_commit_schedule(dreq);
 657                         break;
 658                 case NFS_ODIRECT_RESCHED_WRITES:
 659                         nfs_direct_write_reschedule(dreq);
 660                         break;
 661                 default:
 662                         nfs_direct_write_clear_reqs(dreq);
 663                         nfs_zap_mapping(dreq->inode, dreq->inode->i_mapping);
 664                         nfs_direct_complete(dreq);
 665         }
 666 }
 667
 668 static void nfs_direct_write_complete(struct nfs_direct_req *dreq)
 669 {
 670         trace_nfs_direct_write_complete(dreq);
 671         queue_work(nfsiod_workqueue, &dreq->work); /* Calls nfs_direct_write_schedule_work */
 672 }
 673
 674 static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 675 {
 676         struct nfs_direct_req *dreq = hdr->dreq;
 677         struct nfs_commit_info cinfo;
 678         struct nfs_page *req = nfs_list_entry(hdr->pages.next);
 679         int flags = NFS_ODIRECT_DONE;
 680
 681         trace_nfs_direct_write_completion(dreq);
 682
 683         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 684
 685         spin_lock(&dreq->lock);
 686         if (test_bit(NFS_IOHDR_REDO, &hdr->flags)) {
 687                 spin_unlock(&dreq->lock);
 688                 goto out_put;
 689         }
 690
 691         nfs_direct_count_bytes(dreq, hdr);
 692         if (test_bit(NFS_IOHDR_UNSTABLE_WRITES, &hdr->flags)) {
 693                 if (!dreq->flags)
 694                         dreq->flags = NFS_ODIRECT_DO_COMMIT;
 695                 flags = dreq->flags;
 696         }
 697         spin_unlock(&dreq->lock);
 698
 699         while (!list_empty(&hdr->pages)) {
 700
 701                 req = nfs_list_entry(hdr->pages.next);
 702                 nfs_list_remove_request(req);
 703                 if (flags == NFS_ODIRECT_DO_COMMIT) {
 704                         kref_get(&req->wb_kref);
 705                         memcpy(&req->wb_verf, &hdr->verf.verifier,
 706                                sizeof(req->wb_verf));
 707                         nfs_mark_request_commit(req, hdr->lseg, &cinfo,
 708                                 hdr->ds_commit_idx);
 709                 } else if (flags == NFS_ODIRECT_RESCHED_WRITES) {
 710                         kref_get(&req->wb_kref);
 711                         nfs_mark_request_commit(req, NULL, &cinfo, 0);
 712                 }
 713                 nfs_unlock_and_release_request(req);
 714         }
 715
 716 out_put:
 717         if (put_dreq(dreq))
 718                 nfs_direct_write_complete(dreq);
 719         hdr->release(hdr);
 720 }
 721
 722 static void nfs_write_sync_pgio_error(struct list_head *head, int error)
 723 {
 724         struct nfs_page *req;
 725
 726         while (!list_empty(head)) {
 727                 req = nfs_list_entry(head->next);
 728                 nfs_list_remove_request(req);
 729                 nfs_unlock_and_release_request(req);
 730         }
 731 }
 732
 733 static void nfs_direct_write_reschedule_io(struct nfs_pgio_header *hdr)
 734 {
 735         struct nfs_direct_req *dreq = hdr->dreq;
 736
 737         trace_nfs_direct_write_reschedule_io(dreq);
 738
 739         spin_lock(&dreq->lock);
 740         if (dreq->error == 0) {
 741                 dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 742                 /* fake unstable write to let common nfs resend pages */
 743                 hdr->verf.committed = NFS_UNSTABLE;
 744                 hdr->good_bytes = hdr->args.offset + hdr->args.count -
 745                         hdr->io_start;
 746         }
 747         spin_unlock(&dreq->lock);
 748 }
 749
 750 static const struct nfs_pgio_completion_ops nfs_direct_write_completion_ops = {
 751         .error_cleanup = nfs_write_sync_pgio_error,
 752         .init_hdr = nfs_direct_pgio_init,
 753         .completion = nfs_direct_write_completion,
 754         .reschedule_io = nfs_direct_write_reschedule_io,
 755 };
 756
 757
 758 /*
 759  * NB: Return the value of the first error return code.  Subsequent
 760  *     errors after the first one are ignored.
 761  */
 762 /*
 763  * For each wsize'd chunk of the user's buffer, dispatch an NFS WRITE
 764  * operation.  If nfs_writedata_alloc() or get_user_pages() fails,
 765  * bail and stop sending more writes.  Write length accounting is
 766  * handled automatically by nfs_direct_write_result().  Otherwise, if
 767  * no requests have been sent, just return an error.
 768  */
 769 static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq,
 770                                                struct iov_iter *iter,
 771                                                loff_t pos, int ioflags)
 772 {
 773         struct nfs_pageio_descriptor desc;
 774         struct inode *inode = dreq->inode;
 775         ssize_t result = 0;
 776         size_t requested_bytes = 0;
 777         size_t wsize = max_t(size_t, NFS_SERVER(inode)->wsize, PAGE_SIZE);
 778
 779         trace_nfs_direct_write_schedule_iovec(dreq);
 780
 781         nfs_pageio_init_write(&desc, inode, ioflags, false,
 782                               &nfs_direct_write_completion_ops);
 783         desc.pg_dreq = dreq;
 784         get_dreq(dreq);
 785         inode_dio_begin(inode);
 786
 787         NFS_I(inode)->write_io += iov_iter_count(iter);
 788         while (iov_iter_count(iter)) {
 789                 struct page **pagevec;
 790                 size_t bytes;
 791                 size_t pgbase;
 792                 unsigned npages, i;
 793
 794                 result = iov_iter_get_pages_alloc2(iter, &pagevec,
 795                                                   wsize, &pgbase);
 796                 if (result < 0)
 797                         break;
 798
 799                 bytes = result;
 800                 npages = (result + pgbase + PAGE_SIZE - 1) / PAGE_SIZE;
 801                 for (i = 0; i < npages; i++) {
 802                         struct nfs_page *req;
 803                         unsigned int req_len = min_t(size_t, bytes, PAGE_SIZE - pgbase);
 804
 805                         req = nfs_create_request(dreq->ctx, pagevec[i],
 806                                                  pgbase, req_len);
 807                         if (IS_ERR(req)) {
 808                                 result = PTR_ERR(req);
 809                                 break;
 810                         }
 811
 812                         if (desc.pg_error < 0) {
 813                                 nfs_free_request(req);
 814                                 result = desc.pg_error;
 815                                 break;
 816                         }
 817
 818                         nfs_lock_request(req);
 819                         req->wb_index = pos >> PAGE_SHIFT;
 820                         req->wb_offset = pos & ~PAGE_MASK;
 821                         if (!nfs_pageio_add_request(&desc, req)) {
 822                                 result = desc.pg_error;
 823                                 nfs_unlock_and_release_request(req);
 824                                 break;
 825                         }
 826                         pgbase = 0;
 827                         bytes -= req_len;
 828                         requested_bytes += req_len;
 829                         pos += req_len;
 830                         dreq->bytes_left -= req_len;
 831                 }
 832                 nfs_direct_release_pages(pagevec, npages);
 833                 kvfree(pagevec);
 834                 if (result < 0)
 835                         break;
 836         }
 837         nfs_pageio_complete(&desc);
 838
 839         /*
 840          * If no bytes were started, return the error, and let the
 841          * generic layer handle the completion.
 842          */
 843         if (requested_bytes == 0) {
 844                 inode_dio_end(inode);
 845                 nfs_direct_req_release(dreq);
 846                 return result < 0 ? result : -EIO;
 847         }
 848
 849         if (put_dreq(dreq))
 850                 nfs_direct_write_complete(dreq);
 851         return requested_bytes;
 852 }
 853
 854 /**
 855  * nfs_file_direct_write - file direct write operation for NFS files
 856  * @iocb: target I/O control block
 857  * @iter: vector of user buffers from which to write data
 858  * @swap: flag indicating this is swap IO, not O_DIRECT IO
 859  *
 860  * We use this function for direct writes instead of calling
 861  * generic_file_aio_write() in order to avoid taking the inode
 862  * semaphore and updating the i_size.  The NFS server will set
 863  * the new i_size and this client must read the updated size
 864  * back into its cache.  We let the server do generic write
 865  * parameter checking and report problems.
 866  *
 867  * We eliminate local atime updates, see direct read above.
 868  *
 869  * We avoid unnecessary page cache invalidations for normal cached
 870  * readers of this file.
 871  *
 872  * Note that O_APPEND is not supported for NFS direct writes, as there
 873  * is no atomic O_APPEND write facility in the NFS protocol.
 874  */
 875 ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter,
 876                               bool swap)
 877 {
 878         ssize_t result, requested;
 879         size_t count;
 880         struct file *file = iocb->ki_filp;
 881         struct address_space *mapping = file->f_mapping;
 882         struct inode *inode = mapping->host;
 883         struct nfs_direct_req *dreq;
 884         struct nfs_lock_context *l_ctx;
 885         loff_t pos, end;
 886
 887         dfprintk(FILE, "NFS: direct write(%pD2, %zd@%Ld)\n",
 888                 file, iov_iter_count(iter), (long long) iocb->ki_pos);
 889
 890         if (swap)
 891                 /* bypass generic checks */
 892                 result =  iov_iter_count(iter);
 893         else
 894                 result = generic_write_checks(iocb, iter);
 895         if (result <= 0)
 896                 return result;
 897         count = result;
 898         nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, count);
 899
 900         pos = iocb->ki_pos;
 901         end = (pos + iov_iter_count(iter) - 1) >> PAGE_SHIFT;
 902
 903         task_io_account_write(count);
 904
 905         result = -ENOMEM;
 906         dreq = nfs_direct_req_alloc();
 907         if (!dreq)
 908                 goto out;
 909
 910         dreq->inode = inode;
 911         dreq->bytes_left = dreq->max_count = count;
 912         dreq->io_start = pos;
 913         dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp));
 914         l_ctx = nfs_get_lock_context(dreq->ctx);
 915         if (IS_ERR(l_ctx)) {
 916                 result = PTR_ERR(l_ctx);
 917                 nfs_direct_req_release(dreq);
 918                 goto out_release;
 919         }
 920         dreq->l_ctx = l_ctx;
 921         if (!is_sync_kiocb(iocb))
 922                 dreq->iocb = iocb;
 923         pnfs_init_ds_commit_info_ops(&dreq->ds_cinfo, inode);
 924
 925         if (swap) {
 926                 requested = nfs_direct_write_schedule_iovec(dreq, iter, pos,
 927                                                             FLUSH_STABLE);
 928         } else {
 929                 nfs_start_io_direct(inode);
 930
 931                 requested = nfs_direct_write_schedule_iovec(dreq, iter, pos,
 932                                                             FLUSH_COND_STABLE);
 933
 934                 if (mapping->nrpages) {
 935                         invalidate_inode_pages2_range(mapping,
 936                                                       pos >> PAGE_SHIFT, end);
 937                 }
 938
 939                 nfs_end_io_direct(inode);
 940         }
 941
 942         if (requested > 0) {
 943                 result = nfs_direct_wait(dreq);
 944                 if (result > 0) {
 945                         requested -= result;
 946                         iocb->ki_pos = pos + result;
 947                         /* XXX: should check the generic_write_sync retval */
 948                         generic_write_sync(iocb, result);
 949                 }
 950                 iov_iter_revert(iter, requested);
 951         } else {
 952                 result = requested;
 953         }
 954         nfs_fscache_invalidate(inode, FSCACHE_INVAL_DIO_WRITE);
 955 out_release:
 956         nfs_direct_req_release(dreq);
 957 out:
 958         return result;
 959 }
 960
 961 /**
 962  * nfs_init_directcache - create a slab cache for nfs_direct_req structures
 963  *
 964  */
 965 int __init nfs_init_directcache(void)
 966 {
 967         nfs_direct_cachep = kmem_cache_create("nfs_direct_cache",
 968                                                 sizeof(struct nfs_direct_req),
 969                                                 0, (SLAB_RECLAIM_ACCOUNT|
 970                                                         SLAB_MEM_SPREAD),
 971                                                 NULL);
 972         if (nfs_direct_cachep == NULL)
 973                 return -ENOMEM;
 974
 975         return 0;
 976 }
 977
 978 /**
 979  * nfs_destroy_directcache - destroy the slab cache for nfs_direct_req structures
 980  *
 981  */
 982 void nfs_destroy_directcache(void)
 983 {
 984         kmem_cache_destroy(nfs_direct_cachep);
 985 }