fs/nfs/direct.c

   1 /*
   2  * linux/fs/nfs/direct.c
   3  *
   4  * Copyright (C) 2003 by Chuck Lever <cel@netapp.com>
   5  *
   6  * High-performance uncached I/O for the Linux NFS client
   7  *
   8  * There are important applications whose performance or correctness
   9  * depends on uncached access to file data.  Database clusters
  10  * (multiple copies of the same instance running on separate hosts)
  11  * implement their own cache coherency protocol that subsumes file
  12  * system cache protocols.  Applications that process datasets
  13  * considerably larger than the client's memory do not always benefit
  14  * from a local cache.  A streaming video server, for instance, has no
  15  * need to cache the contents of a file.
  16  *
  17  * When an application requests uncached I/O, all read and write requests
  18  * are made directly to the server; data stored or fetched via these
  19  * requests is not cached in the Linux page cache.  The client does not
  20  * correct unaligned requests from applications.  All requested bytes are
  21  * held on permanent storage before a direct write system call returns to
  22  * an application.
  23  *
  24  * Solaris implements an uncached I/O facility called directio() that
  25  * is used for backups and sequential I/O to very large files.  Solaris
  26  * also supports uncaching whole NFS partitions with "-o forcedirectio,"
  27  * an undocumented mount option.
  28  *
  29  * Designed by Jeff Kimmel, Chuck Lever, and Trond Myklebust, with
  30  * help from Andrew Morton.
  31  *
  32  * 18 Dec 2001  Initial implementation for 2.4  --cel
  33  * 08 Jul 2002  Version for 2.4.19, with bug fixes --trondmy
  34  * 08 Jun 2003  Port to 2.5 APIs  --cel
  35  * 31 Mar 2004  Handle direct I/O without VFS support  --cel
  36  * 15 Sep 2004  Parallel async reads  --cel
  37  * 04 May 2005  support O_DIRECT with aio  --cel
  38  *
  39  */
  40
  41 #include <linux/errno.h>
  42 #include <linux/sched.h>
  43 #include <linux/kernel.h>
  44 #include <linux/file.h>
  45 #include <linux/pagemap.h>
  46 #include <linux/kref.h>
  47 #include <linux/slab.h>
  48 #include <linux/task_io_accounting_ops.h>
  49
  50 #include <linux/nfs_fs.h>
  51 #include <linux/nfs_page.h>
  52 #include <linux/sunrpc/clnt.h>
  53
  54 #include <asm/uaccess.h>
  55 #include <linux/atomic.h>
  56
  57 #include "internal.h"
  58 #include "iostat.h"
  59 #include "pnfs.h"
  60
  61 #define NFSDBG_FACILITY         NFSDBG_VFS
  62
  63 static struct kmem_cache *nfs_direct_cachep;
  64
  65 /*
  66  * This represents a set of asynchronous requests that we're waiting on
  67  */
  68 struct nfs_direct_req {
  69         struct kref             kref;           /* release manager */
  70
  71         /* I/O parameters */
  72         struct nfs_open_context *ctx;           /* file open context info */
  73         struct nfs_lock_context *l_ctx;         /* Lock context info */
  74         struct kiocb *          iocb;           /* controlling i/o request */
  75         struct inode *          inode;          /* target file of i/o */
  76
  77         /* completion state */
  78         atomic_t                io_count;       /* i/os we're waiting for */
  79         spinlock_t              lock;           /* protect completion state */
  80         ssize_t                 count,          /* bytes actually processed */
  81                                 error;          /* any reported error */
  82         struct completion       completion;     /* wait for i/o completion */
  83
  84         /* commit state */
  85         struct nfs_mds_commit_info mds_cinfo;   /* Storage for cinfo */
  86         struct pnfs_ds_commit_info ds_cinfo;    /* Storage for cinfo */
  87         struct work_struct      work;
  88         int                     flags;
  89 #define NFS_ODIRECT_DO_COMMIT           (1)     /* an unstable reply was received */
  90 #define NFS_ODIRECT_RESCHED_WRITES      (2)     /* write verification failed */
  91         struct nfs_writeverf    verf;           /* unstable write verifier */
  92 };
  93
  94 static const struct nfs_pgio_completion_ops nfs_direct_write_completion_ops;
  95 static const struct nfs_commit_completion_ops nfs_direct_commit_completion_ops;
  96 static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode);
  97 static void nfs_direct_write_schedule_work(struct work_struct *work);
  98
  99 static inline void get_dreq(struct nfs_direct_req *dreq)
 100 {
 101         atomic_inc(&dreq->io_count);
 102 }
 103
 104 static inline int put_dreq(struct nfs_direct_req *dreq)
 105 {
 106         return atomic_dec_and_test(&dreq->io_count);
 107 }
 108
 109 /**
 110  * nfs_direct_IO - NFS address space operation for direct I/O
 111  * @rw: direction (read or write)
 112  * @iocb: target I/O control block
 113  * @iov: array of vectors that define I/O buffer
 114  * @pos: offset in file to begin the operation
 115  * @nr_segs: size of iovec array
 116  *
 117  * The presence of this routine in the address space ops vector means
 118  * the NFS client supports direct I/O. However, for most direct IO, we
 119  * shunt off direct read and write requests before the VFS gets them,
 120  * so this method is only ever called for swap.
 121  */
 122 ssize_t nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, loff_t pos, unsigned long nr_segs)
 123 {
 124 #ifndef CONFIG_NFS_SWAP
 125         dprintk("NFS: nfs_direct_IO (%s) off/no(%Ld/%lu) EINVAL\n",
 126                         iocb->ki_filp->f_path.dentry->d_name.name,
 127                         (long long) pos, nr_segs);
 128
 129         return -EINVAL;
 130 #else
 131         VM_BUG_ON(iocb->ki_left != PAGE_SIZE);
 132         VM_BUG_ON(iocb->ki_nbytes != PAGE_SIZE);
 133
 134         if (rw == READ || rw == KERNEL_READ)
 135                 return nfs_file_direct_read(iocb, iov, nr_segs, pos,
 136                                 rw == READ ? true : false);
 137         return nfs_file_direct_write(iocb, iov, nr_segs, pos,
 138                                 rw == WRITE ? true : false);
 139 #endif /* CONFIG_NFS_SWAP */
 140 }
 141
 142 static void nfs_direct_release_pages(struct page **pages, unsigned int npages)
 143 {
 144         unsigned int i;
 145         for (i = 0; i < npages; i++)
 146                 page_cache_release(pages[i]);
 147 }
 148
 149 void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo,
 150                               struct nfs_direct_req *dreq)
 151 {
 152         cinfo->lock = &dreq->lock;
 153         cinfo->mds = &dreq->mds_cinfo;
 154         cinfo->ds = &dreq->ds_cinfo;
 155         cinfo->dreq = dreq;
 156         cinfo->completion_ops = &nfs_direct_commit_completion_ops;
 157 }
 158
 159 static inline struct nfs_direct_req *nfs_direct_req_alloc(void)
 160 {
 161         struct nfs_direct_req *dreq;
 162
 163         dreq = kmem_cache_zalloc(nfs_direct_cachep, GFP_KERNEL);
 164         if (!dreq)
 165                 return NULL;
 166
 167         kref_init(&dreq->kref);
 168         kref_get(&dreq->kref);
 169         init_completion(&dreq->completion);
 170         INIT_LIST_HEAD(&dreq->mds_cinfo.list);
 171         INIT_WORK(&dreq->work, nfs_direct_write_schedule_work);
 172         spin_lock_init(&dreq->lock);
 173
 174         return dreq;
 175 }
 176
 177 static void nfs_direct_req_free(struct kref *kref)
 178 {
 179         struct nfs_direct_req *dreq = container_of(kref, struct nfs_direct_req, kref);
 180
 181         if (dreq->l_ctx != NULL)
 182                 nfs_put_lock_context(dreq->l_ctx);
 183         if (dreq->ctx != NULL)
 184                 put_nfs_open_context(dreq->ctx);
 185         kmem_cache_free(nfs_direct_cachep, dreq);
 186 }
 187
 188 static void nfs_direct_req_release(struct nfs_direct_req *dreq)
 189 {
 190         kref_put(&dreq->kref, nfs_direct_req_free);
 191 }
 192
 193 /*
 194  * Collects and returns the final error value/byte-count.
 195  */
 196 static ssize_t nfs_direct_wait(struct nfs_direct_req *dreq)
 197 {
 198         ssize_t result = -EIOCBQUEUED;
 199
 200         /* Async requests don't wait here */
 201         if (dreq->iocb)
 202                 goto out;
 203
 204         result = wait_for_completion_killable(&dreq->completion);
 205
 206         if (!result)
 207                 result = dreq->error;
 208         if (!result)
 209                 result = dreq->count;
 210
 211 out:
 212         return (ssize_t) result;
 213 }
 214
 215 /*
 216  * Synchronous I/O uses a stack-allocated iocb.  Thus we can't trust
 217  * the iocb is still valid here if this is a synchronous request.
 218  */
 219 static void nfs_direct_complete(struct nfs_direct_req *dreq)
 220 {
 221         if (dreq->iocb) {
 222                 long res = (long) dreq->error;
 223                 if (!res)
 224                         res = (long) dreq->count;
 225                 aio_complete(dreq->iocb, res, 0);
 226         }
 227         complete_all(&dreq->completion);
 228
 229         nfs_direct_req_release(dreq);
 230 }
 231
 232 static void nfs_direct_readpage_release(struct nfs_page *req)
 233 {
 234         dprintk("NFS: direct read done (%s/%lld %d@%lld)\n",
 235                 req->wb_context->dentry->d_inode->i_sb->s_id,
 236                 (long long)NFS_FILEID(req->wb_context->dentry->d_inode),
 237                 req->wb_bytes,
 238                 (long long)req_offset(req));
 239         nfs_release_request(req);
 240 }
 241
 242 static void nfs_direct_read_completion(struct nfs_pgio_header *hdr)
 243 {
 244         unsigned long bytes = 0;
 245         struct nfs_direct_req *dreq = hdr->dreq;
 246
 247         if (test_bit(NFS_IOHDR_REDO, &hdr->flags))
 248                 goto out_put;
 249
 250         spin_lock(&dreq->lock);
 251         if (test_bit(NFS_IOHDR_ERROR, &hdr->flags) && (hdr->good_bytes == 0))
 252                 dreq->error = hdr->error;
 253         else
 254                 dreq->count += hdr->good_bytes;
 255         spin_unlock(&dreq->lock);
 256
 257         while (!list_empty(&hdr->pages)) {
 258                 struct nfs_page *req = nfs_list_entry(hdr->pages.next);
 259                 struct page *page = req->wb_page;
 260
 261                 if (test_bit(NFS_IOHDR_EOF, &hdr->flags)) {
 262                         if (bytes > hdr->good_bytes)
 263                                 zero_user(page, 0, PAGE_SIZE);
 264                         else if (hdr->good_bytes - bytes < PAGE_SIZE)
 265                                 zero_user_segment(page,
 266                                         hdr->good_bytes & ~PAGE_MASK,
 267                                         PAGE_SIZE);
 268                 }
 269                 if (!PageCompound(page)) {
 270                         if (test_bit(NFS_IOHDR_ERROR, &hdr->flags)) {
 271                                 if (bytes < hdr->good_bytes)
 272                                         set_page_dirty(page);
 273                         } else
 274                                 set_page_dirty(page);
 275                 }
 276                 bytes += req->wb_bytes;
 277                 nfs_list_remove_request(req);
 278                 nfs_direct_readpage_release(req);
 279         }
 280 out_put:
 281         if (put_dreq(dreq))
 282                 nfs_direct_complete(dreq);
 283         hdr->release(hdr);
 284 }
 285
 286 static void nfs_read_sync_pgio_error(struct list_head *head)
 287 {
 288         struct nfs_page *req;
 289
 290         while (!list_empty(head)) {
 291                 req = nfs_list_entry(head->next);
 292                 nfs_list_remove_request(req);
 293                 nfs_release_request(req);
 294         }
 295 }
 296
 297 static void nfs_direct_pgio_init(struct nfs_pgio_header *hdr)
 298 {
 299         get_dreq(hdr->dreq);
 300 }
 301
 302 static const struct nfs_pgio_completion_ops nfs_direct_read_completion_ops = {
 303         .error_cleanup = nfs_read_sync_pgio_error,
 304         .init_hdr = nfs_direct_pgio_init,
 305         .completion = nfs_direct_read_completion,
 306 };
 307
 308 /*
 309  * For each rsize'd chunk of the user's buffer, dispatch an NFS READ
 310  * operation.  If nfs_readdata_alloc() or get_user_pages() fails,
 311  * bail and stop sending more reads.  Read length accounting is
 312  * handled automatically by nfs_direct_read_result().  Otherwise, if
 313  * no requests have been sent, just return an error.
 314  */
 315 static ssize_t nfs_direct_read_schedule_segment(struct nfs_pageio_descriptor *desc,
 316                                                 const struct iovec *iov,
 317                                                 loff_t pos, bool uio)
 318 {
 319         struct nfs_direct_req *dreq = desc->pg_dreq;
 320         struct nfs_open_context *ctx = dreq->ctx;
 321         struct inode *inode = ctx->dentry->d_inode;
 322         unsigned long user_addr = (unsigned long)iov->iov_base;
 323         size_t count = iov->iov_len;
 324         size_t rsize = NFS_SERVER(inode)->rsize;
 325         unsigned int pgbase;
 326         int result;
 327         ssize_t started = 0;
 328         struct page **pagevec = NULL;
 329         unsigned int npages;
 330
 331         do {
 332                 size_t bytes;
 333                 int i;
 334
 335                 pgbase = user_addr & ~PAGE_MASK;
 336                 bytes = min(max_t(size_t, rsize, PAGE_SIZE), count);
 337
 338                 result = -ENOMEM;
 339                 npages = nfs_page_array_len(pgbase, bytes);
 340                 if (!pagevec)
 341                         pagevec = kmalloc(npages * sizeof(struct page *),
 342                                           GFP_KERNEL);
 343                 if (!pagevec)
 344                         break;
 345                 if (uio) {
 346                         down_read(&current->mm->mmap_sem);
 347                         result = get_user_pages(current, current->mm, user_addr,
 348                                         npages, 1, 0, pagevec, NULL);
 349                         up_read(&current->mm->mmap_sem);
 350                         if (result < 0)
 351                                 break;
 352                 } else {
 353                         WARN_ON(npages != 1);
 354                         result = get_kernel_page(user_addr, 1, pagevec);
 355                         if (WARN_ON(result != 1))
 356                                 break;
 357                 }
 358
 359                 if ((unsigned)result < npages) {
 360                         bytes = result * PAGE_SIZE;
 361                         if (bytes <= pgbase) {
 362                                 nfs_direct_release_pages(pagevec, result);
 363                                 break;
 364                         }
 365                         bytes -= pgbase;
 366                         npages = result;
 367                 }
 368
 369                 for (i = 0; i < npages; i++) {
 370                         struct nfs_page *req;
 371                         unsigned int req_len = min_t(size_t, bytes, PAGE_SIZE - pgbase);
 372                         /* XXX do we need to do the eof zeroing found in async_filler? */
 373                         req = nfs_create_request(dreq->ctx, dreq->inode,
 374                                                  pagevec[i],
 375                                                  pgbase, req_len);
 376                         if (IS_ERR(req)) {
 377                                 result = PTR_ERR(req);
 378                                 break;
 379                         }
 380                         req->wb_index = pos >> PAGE_SHIFT;
 381                         req->wb_offset = pos & ~PAGE_MASK;
 382                         if (!nfs_pageio_add_request(desc, req)) {
 383                                 result = desc->pg_error;
 384                                 nfs_release_request(req);
 385                                 break;
 386                         }
 387                         pgbase = 0;
 388                         bytes -= req_len;
 389                         started += req_len;
 390                         user_addr += req_len;
 391                         pos += req_len;
 392                         count -= req_len;
 393                 }
 394                 /* The nfs_page now hold references to these pages */
 395                 nfs_direct_release_pages(pagevec, npages);
 396         } while (count != 0 && result >= 0);
 397
 398         kfree(pagevec);
 399
 400         if (started)
 401                 return started;
 402         return result < 0 ? (ssize_t) result : -EFAULT;
 403 }
 404
 405 static ssize_t nfs_direct_read_schedule_iovec(struct nfs_direct_req *dreq,
 406                                               const struct iovec *iov,
 407                                               unsigned long nr_segs,
 408                                               loff_t pos, bool uio)
 409 {
 410         struct nfs_pageio_descriptor desc;
 411         ssize_t result = -EINVAL;
 412         size_t requested_bytes = 0;
 413         unsigned long seg;
 414
 415         NFS_PROTO(dreq->inode)->read_pageio_init(&desc, dreq->inode,
 416                              &nfs_direct_read_completion_ops);
 417         get_dreq(dreq);
 418         desc.pg_dreq = dreq;
 419
 420         for (seg = 0; seg < nr_segs; seg++) {
 421                 const struct iovec *vec = &iov[seg];
 422                 result = nfs_direct_read_schedule_segment(&desc, vec, pos, uio);
 423                 if (result < 0)
 424                         break;
 425                 requested_bytes += result;
 426                 if ((size_t)result < vec->iov_len)
 427                         break;
 428                 pos += vec->iov_len;
 429         }
 430
 431         nfs_pageio_complete(&desc);
 432
 433         /*
 434          * If no bytes were started, return the error, and let the
 435          * generic layer handle the completion.
 436          */
 437         if (requested_bytes == 0) {
 438                 nfs_direct_req_release(dreq);
 439                 return result < 0 ? result : -EIO;
 440         }
 441
 442         if (put_dreq(dreq))
 443                 nfs_direct_complete(dreq);
 444         return 0;
 445 }
 446
 447 static ssize_t nfs_direct_read(struct kiocb *iocb, const struct iovec *iov,
 448                                unsigned long nr_segs, loff_t pos, bool uio)
 449 {
 450         ssize_t result = -ENOMEM;
 451         struct inode *inode = iocb->ki_filp->f_mapping->host;
 452         struct nfs_direct_req *dreq;
 453
 454         dreq = nfs_direct_req_alloc();
 455         if (dreq == NULL)
 456                 goto out;
 457
 458         dreq->inode = inode;
 459         dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp));
 460         dreq->l_ctx = nfs_get_lock_context(dreq->ctx);
 461         if (dreq->l_ctx == NULL)
 462                 goto out_release;
 463         if (!is_sync_kiocb(iocb))
 464                 dreq->iocb = iocb;
 465
 466         result = nfs_direct_read_schedule_iovec(dreq, iov, nr_segs, pos, uio);
 467         if (!result)
 468                 result = nfs_direct_wait(dreq);
 469         NFS_I(inode)->read_io += result;
 470 out_release:
 471         nfs_direct_req_release(dreq);
 472 out:
 473         return result;
 474 }
 475
 476 static void nfs_inode_dio_write_done(struct inode *inode)
 477 {
 478         nfs_zap_mapping(inode, inode->i_mapping);
 479         inode_dio_done(inode);
 480 }
 481
 482 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
 483 static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq)
 484 {
 485         struct nfs_pageio_descriptor desc;
 486         struct nfs_page *req, *tmp;
 487         LIST_HEAD(reqs);
 488         struct nfs_commit_info cinfo;
 489         LIST_HEAD(failed);
 490
 491         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 492         pnfs_recover_commit_reqs(dreq->inode, &reqs, &cinfo);
 493         spin_lock(cinfo.lock);
 494         nfs_scan_commit_list(&cinfo.mds->list, &reqs, &cinfo, 0);
 495         spin_unlock(cinfo.lock);
 496
 497         dreq->count = 0;
 498         get_dreq(dreq);
 499
 500         NFS_PROTO(dreq->inode)->write_pageio_init(&desc, dreq->inode, FLUSH_STABLE,
 501                               &nfs_direct_write_completion_ops);
 502         desc.pg_dreq = dreq;
 503
 504         list_for_each_entry_safe(req, tmp, &reqs, wb_list) {
 505                 if (!nfs_pageio_add_request(&desc, req)) {
 506                         nfs_list_remove_request(req);
 507                         nfs_list_add_request(req, &failed);
 508                         spin_lock(cinfo.lock);
 509                         dreq->flags = 0;
 510                         dreq->error = -EIO;
 511                         spin_unlock(cinfo.lock);
 512                 }
 513                 nfs_release_request(req);
 514         }
 515         nfs_pageio_complete(&desc);
 516
 517         while (!list_empty(&failed)) {
 518                 req = nfs_list_entry(failed.next);
 519                 nfs_list_remove_request(req);
 520                 nfs_unlock_and_release_request(req);
 521         }
 522
 523         if (put_dreq(dreq))
 524                 nfs_direct_write_complete(dreq, dreq->inode);
 525 }
 526
 527 static void nfs_direct_commit_complete(struct nfs_commit_data *data)
 528 {
 529         struct nfs_direct_req *dreq = data->dreq;
 530         struct nfs_commit_info cinfo;
 531         struct nfs_page *req;
 532         int status = data->task.tk_status;
 533
 534         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 535         if (status < 0) {
 536                 dprintk("NFS: %5u commit failed with error %d.\n",
 537                         data->task.tk_pid, status);
 538                 dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 539         } else if (memcmp(&dreq->verf, &data->verf, sizeof(data->verf))) {
 540                 dprintk("NFS: %5u commit verify failed\n", data->task.tk_pid);
 541                 dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 542         }
 543
 544         dprintk("NFS: %5u commit returned %d\n", data->task.tk_pid, status);
 545         while (!list_empty(&data->pages)) {
 546                 req = nfs_list_entry(data->pages.next);
 547                 nfs_list_remove_request(req);
 548                 if (dreq->flags == NFS_ODIRECT_RESCHED_WRITES) {
 549                         /* Note the rewrite will go through mds */
 550                         nfs_mark_request_commit(req, NULL, &cinfo);
 551                 } else
 552                         nfs_release_request(req);
 553                 nfs_unlock_and_release_request(req);
 554         }
 555
 556         if (atomic_dec_and_test(&cinfo.mds->rpcs_out))
 557                 nfs_direct_write_complete(dreq, data->inode);
 558 }
 559
 560 static void nfs_direct_error_cleanup(struct nfs_inode *nfsi)
 561 {
 562         /* There is no lock to clear */
 563 }
 564
 565 static const struct nfs_commit_completion_ops nfs_direct_commit_completion_ops = {
 566         .completion = nfs_direct_commit_complete,
 567         .error_cleanup = nfs_direct_error_cleanup,
 568 };
 569
 570 static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq)
 571 {
 572         int res;
 573         struct nfs_commit_info cinfo;
 574         LIST_HEAD(mds_list);
 575
 576         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 577         nfs_scan_commit(dreq->inode, &mds_list, &cinfo);
 578         res = nfs_generic_commit_list(dreq->inode, &mds_list, 0, &cinfo);
 579         if (res < 0) /* res == -ENOMEM */
 580                 nfs_direct_write_reschedule(dreq);
 581 }
 582
 583 static void nfs_direct_write_schedule_work(struct work_struct *work)
 584 {
 585         struct nfs_direct_req *dreq = container_of(work, struct nfs_direct_req, work);
 586         int flags = dreq->flags;
 587
 588         dreq->flags = 0;
 589         switch (flags) {
 590                 case NFS_ODIRECT_DO_COMMIT:
 591                         nfs_direct_commit_schedule(dreq);
 592                         break;
 593                 case NFS_ODIRECT_RESCHED_WRITES:
 594                         nfs_direct_write_reschedule(dreq);
 595                         break;
 596                 default:
 597                         nfs_inode_dio_write_done(dreq->inode);
 598                         nfs_direct_complete(dreq);
 599         }
 600 }
 601
 602 static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode)
 603 {
 604         schedule_work(&dreq->work); /* Calls nfs_direct_write_schedule_work */
 605 }
 606
 607 #else
 608 static void nfs_direct_write_schedule_work(struct work_struct *work)
 609 {
 610 }
 611
 612 static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode)
 613 {
 614         nfs_inode_dio_write_done(inode);
 615         nfs_direct_complete(dreq);
 616 }
 617 #endif
 618
 619 /*
 620  * NB: Return the value of the first error return code.  Subsequent
 621  *     errors after the first one are ignored.
 622  */
 623 /*
 624  * For each wsize'd chunk of the user's buffer, dispatch an NFS WRITE
 625  * operation.  If nfs_writedata_alloc() or get_user_pages() fails,
 626  * bail and stop sending more writes.  Write length accounting is
 627  * handled automatically by nfs_direct_write_result().  Otherwise, if
 628  * no requests have been sent, just return an error.
 629  */
 630 static ssize_t nfs_direct_write_schedule_segment(struct nfs_pageio_descriptor *desc,
 631                                                  const struct iovec *iov,
 632                                                  loff_t pos, bool uio)
 633 {
 634         struct nfs_direct_req *dreq = desc->pg_dreq;
 635         struct nfs_open_context *ctx = dreq->ctx;
 636         struct inode *inode = ctx->dentry->d_inode;
 637         unsigned long user_addr = (unsigned long)iov->iov_base;
 638         size_t count = iov->iov_len;
 639         size_t wsize = NFS_SERVER(inode)->wsize;
 640         unsigned int pgbase;
 641         int result;
 642         ssize_t started = 0;
 643         struct page **pagevec = NULL;
 644         unsigned int npages;
 645
 646         do {
 647                 size_t bytes;
 648                 int i;
 649
 650                 pgbase = user_addr & ~PAGE_MASK;
 651                 bytes = min(max_t(size_t, wsize, PAGE_SIZE), count);
 652
 653                 result = -ENOMEM;
 654                 npages = nfs_page_array_len(pgbase, bytes);
 655                 if (!pagevec)
 656                         pagevec = kmalloc(npages * sizeof(struct page *), GFP_KERNEL);
 657                 if (!pagevec)
 658                         break;
 659
 660                 if (uio) {
 661                         down_read(&current->mm->mmap_sem);
 662                         result = get_user_pages(current, current->mm, user_addr,
 663                                                 npages, 0, 0, pagevec, NULL);
 664                         up_read(&current->mm->mmap_sem);
 665                         if (result < 0)
 666                                 break;
 667                 } else {
 668                         WARN_ON(npages != 1);
 669                         result = get_kernel_page(user_addr, 0, pagevec);
 670                         if (WARN_ON(result != 1))
 671                                 break;
 672                 }
 673
 674                 if ((unsigned)result < npages) {
 675                         bytes = result * PAGE_SIZE;
 676                         if (bytes <= pgbase) {
 677                                 nfs_direct_release_pages(pagevec, result);
 678                                 break;
 679                         }
 680                         bytes -= pgbase;
 681                         npages = result;
 682                 }
 683
 684                 for (i = 0; i < npages; i++) {
 685                         struct nfs_page *req;
 686                         unsigned int req_len = min_t(size_t, bytes, PAGE_SIZE - pgbase);
 687
 688                         req = nfs_create_request(dreq->ctx, dreq->inode,
 689                                                  pagevec[i],
 690                                                  pgbase, req_len);
 691                         if (IS_ERR(req)) {
 692                                 result = PTR_ERR(req);
 693                                 break;
 694                         }
 695                         nfs_lock_request(req);
 696                         req->wb_index = pos >> PAGE_SHIFT;
 697                         req->wb_offset = pos & ~PAGE_MASK;
 698                         if (!nfs_pageio_add_request(desc, req)) {
 699                                 result = desc->pg_error;
 700                                 nfs_unlock_and_release_request(req);
 701                                 break;
 702                         }
 703                         pgbase = 0;
 704                         bytes -= req_len;
 705                         started += req_len;
 706                         user_addr += req_len;
 707                         pos += req_len;
 708                         count -= req_len;
 709                 }
 710                 /* The nfs_page now hold references to these pages */
 711                 nfs_direct_release_pages(pagevec, npages);
 712         } while (count != 0 && result >= 0);
 713
 714         kfree(pagevec);
 715
 716         if (started)
 717                 return started;
 718         return result < 0 ? (ssize_t) result : -EFAULT;
 719 }
 720
 721 static void nfs_direct_write_completion(struct nfs_pgio_header *hdr)
 722 {
 723         struct nfs_direct_req *dreq = hdr->dreq;
 724         struct nfs_commit_info cinfo;
 725         int bit = -1;
 726         struct nfs_page *req = nfs_list_entry(hdr->pages.next);
 727
 728         if (test_bit(NFS_IOHDR_REDO, &hdr->flags))
 729                 goto out_put;
 730
 731         nfs_init_cinfo_from_dreq(&cinfo, dreq);
 732
 733         spin_lock(&dreq->lock);
 734
 735         if (test_bit(NFS_IOHDR_ERROR, &hdr->flags)) {
 736                 dreq->flags = 0;
 737                 dreq->error = hdr->error;
 738         }
 739         if (dreq->error != 0)
 740                 bit = NFS_IOHDR_ERROR;
 741         else {
 742                 dreq->count += hdr->good_bytes;
 743                 if (test_bit(NFS_IOHDR_NEED_RESCHED, &hdr->flags)) {
 744                         dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 745                         bit = NFS_IOHDR_NEED_RESCHED;
 746                 } else if (test_bit(NFS_IOHDR_NEED_COMMIT, &hdr->flags)) {
 747                         if (dreq->flags == NFS_ODIRECT_RESCHED_WRITES)
 748                                 bit = NFS_IOHDR_NEED_RESCHED;
 749                         else if (dreq->flags == 0) {
 750                                 memcpy(&dreq->verf, hdr->verf,
 751                                        sizeof(dreq->verf));
 752                                 bit = NFS_IOHDR_NEED_COMMIT;
 753                                 dreq->flags = NFS_ODIRECT_DO_COMMIT;
 754                         } else if (dreq->flags == NFS_ODIRECT_DO_COMMIT) {
 755                                 if (memcmp(&dreq->verf, hdr->verf, sizeof(dreq->verf))) {
 756                                         dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 757                                         bit = NFS_IOHDR_NEED_RESCHED;
 758                                 } else
 759                                         bit = NFS_IOHDR_NEED_COMMIT;
 760                         }
 761                 }
 762         }
 763         spin_unlock(&dreq->lock);
 764
 765         while (!list_empty(&hdr->pages)) {
 766                 req = nfs_list_entry(hdr->pages.next);
 767                 nfs_list_remove_request(req);
 768                 switch (bit) {
 769                 case NFS_IOHDR_NEED_RESCHED:
 770                 case NFS_IOHDR_NEED_COMMIT:
 771                         kref_get(&req->wb_kref);
 772                         nfs_mark_request_commit(req, hdr->lseg, &cinfo);
 773                 }
 774                 nfs_unlock_and_release_request(req);
 775         }
 776
 777 out_put:
 778         if (put_dreq(dreq))
 779                 nfs_direct_write_complete(dreq, hdr->inode);
 780         hdr->release(hdr);
 781 }
 782
 783 static void nfs_write_sync_pgio_error(struct list_head *head)
 784 {
 785         struct nfs_page *req;
 786
 787         while (!list_empty(head)) {
 788                 req = nfs_list_entry(head->next);
 789                 nfs_list_remove_request(req);
 790                 nfs_unlock_and_release_request(req);
 791         }
 792 }
 793
 794 static const struct nfs_pgio_completion_ops nfs_direct_write_completion_ops = {
 795         .error_cleanup = nfs_write_sync_pgio_error,
 796         .init_hdr = nfs_direct_pgio_init,
 797         .completion = nfs_direct_write_completion,
 798 };
 799
 800 static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq,
 801                                                const struct iovec *iov,
 802                                                unsigned long nr_segs,
 803                                                loff_t pos, bool uio)
 804 {
 805         struct nfs_pageio_descriptor desc;
 806         struct inode *inode = dreq->inode;
 807         ssize_t result = 0;
 808         size_t requested_bytes = 0;
 809         unsigned long seg;
 810
 811         NFS_PROTO(inode)->write_pageio_init(&desc, inode, FLUSH_COND_STABLE,
 812                               &nfs_direct_write_completion_ops);
 813         desc.pg_dreq = dreq;
 814         get_dreq(dreq);
 815         atomic_inc(&inode->i_dio_count);
 816
 817         for (seg = 0; seg < nr_segs; seg++) {
 818                 const struct iovec *vec = &iov[seg];
 819                 result = nfs_direct_write_schedule_segment(&desc, vec, pos, uio);
 820                 if (result < 0)
 821                         break;
 822                 requested_bytes += result;
 823                 if ((size_t)result < vec->iov_len)
 824                         break;
 825                 pos += vec->iov_len;
 826         }
 827         nfs_pageio_complete(&desc);
 828         NFS_I(dreq->inode)->write_io += desc.pg_bytes_written;
 829
 830         /*
 831          * If no bytes were started, return the error, and let the
 832          * generic layer handle the completion.
 833          */
 834         if (requested_bytes == 0) {
 835                 inode_dio_done(inode);
 836                 nfs_direct_req_release(dreq);
 837                 return result < 0 ? result : -EIO;
 838         }
 839
 840         if (put_dreq(dreq))
 841                 nfs_direct_write_complete(dreq, dreq->inode);
 842         return 0;
 843 }
 844
 845 static ssize_t nfs_direct_write(struct kiocb *iocb, const struct iovec *iov,
 846                                 unsigned long nr_segs, loff_t pos,
 847                                 size_t count, bool uio)
 848 {
 849         ssize_t result = -ENOMEM;
 850         struct inode *inode = iocb->ki_filp->f_mapping->host;
 851         struct nfs_direct_req *dreq;
 852
 853         dreq = nfs_direct_req_alloc();
 854         if (!dreq)
 855                 goto out;
 856
 857         dreq->inode = inode;
 858         dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp));
 859         dreq->l_ctx = nfs_get_lock_context(dreq->ctx);
 860         if (dreq->l_ctx == NULL)
 861                 goto out_release;
 862         if (!is_sync_kiocb(iocb))
 863                 dreq->iocb = iocb;
 864
 865         result = nfs_direct_write_schedule_iovec(dreq, iov, nr_segs, pos, uio);
 866         if (!result)
 867                 result = nfs_direct_wait(dreq);
 868 out_release:
 869         nfs_direct_req_release(dreq);
 870 out:
 871         return result;
 872 }
 873
 874 /**
 875  * nfs_file_direct_read - file direct read operation for NFS files
 876  * @iocb: target I/O control block
 877  * @iov: vector of user buffers into which to read data
 878  * @nr_segs: size of iov vector
 879  * @pos: byte offset in file where reading starts
 880  *
 881  * We use this function for direct reads instead of calling
 882  * generic_file_aio_read() in order to avoid gfar's check to see if
 883  * the request starts before the end of the file.  For that check
 884  * to work, we must generate a GETATTR before each direct read, and
 885  * even then there is a window between the GETATTR and the subsequent
 886  * READ where the file size could change.  Our preference is simply
 887  * to do all reads the application wants, and the server will take
 888  * care of managing the end of file boundary.
 889  *
 890  * This function also eliminates unnecessarily updating the file's
 891  * atime locally, as the NFS server sets the file's atime, and this
 892  * client must read the updated atime from the server back into its
 893  * cache.
 894  */
 895 ssize_t nfs_file_direct_read(struct kiocb *iocb, const struct iovec *iov,
 896                                 unsigned long nr_segs, loff_t pos, bool uio)
 897 {
 898         ssize_t retval = -EINVAL;
 899         struct file *file = iocb->ki_filp;
 900         struct address_space *mapping = file->f_mapping;
 901         size_t count;
 902
 903         count = iov_length(iov, nr_segs);
 904         nfs_add_stats(mapping->host, NFSIOS_DIRECTREADBYTES, count);
 905
 906         dfprintk(FILE, "NFS: direct read(%s/%s, %zd@%Ld)\n",
 907                 file->f_path.dentry->d_parent->d_name.name,
 908                 file->f_path.dentry->d_name.name,
 909                 count, (long long) pos);
 910
 911         retval = 0;
 912         if (!count)
 913                 goto out;
 914
 915         retval = nfs_sync_mapping(mapping);
 916         if (retval)
 917                 goto out;
 918
 919         task_io_account_read(count);
 920
 921         retval = nfs_direct_read(iocb, iov, nr_segs, pos, uio);
 922         if (retval > 0)
 923                 iocb->ki_pos = pos + retval;
 924
 925 out:
 926         return retval;
 927 }
 928
 929 /**
 930  * nfs_file_direct_write - file direct write operation for NFS files
 931  * @iocb: target I/O control block
 932  * @iov: vector of user buffers from which to write data
 933  * @nr_segs: size of iov vector
 934  * @pos: byte offset in file where writing starts
 935  *
 936  * We use this function for direct writes instead of calling
 937  * generic_file_aio_write() in order to avoid taking the inode
 938  * semaphore and updating the i_size.  The NFS server will set
 939  * the new i_size and this client must read the updated size
 940  * back into its cache.  We let the server do generic write
 941  * parameter checking and report problems.
 942  *
 943  * We eliminate local atime updates, see direct read above.
 944  *
 945  * We avoid unnecessary page cache invalidations for normal cached
 946  * readers of this file.
 947  *
 948  * Note that O_APPEND is not supported for NFS direct writes, as there
 949  * is no atomic O_APPEND write facility in the NFS protocol.
 950  */
 951 ssize_t nfs_file_direct_write(struct kiocb *iocb, const struct iovec *iov,
 952                                 unsigned long nr_segs, loff_t pos, bool uio)
 953 {
 954         ssize_t retval = -EINVAL;
 955         struct file *file = iocb->ki_filp;
 956         struct address_space *mapping = file->f_mapping;
 957         size_t count;
 958
 959         count = iov_length(iov, nr_segs);
 960         nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, count);
 961
 962         dfprintk(FILE, "NFS: direct write(%s/%s, %zd@%Ld)\n",
 963                 file->f_path.dentry->d_parent->d_name.name,
 964                 file->f_path.dentry->d_name.name,
 965                 count, (long long) pos);
 966
 967         retval = generic_write_checks(file, &pos, &count, 0);
 968         if (retval)
 969                 goto out;
 970
 971         retval = -EINVAL;
 972         if ((ssize_t) count < 0)
 973                 goto out;
 974         retval = 0;
 975         if (!count)
 976                 goto out;
 977
 978         retval = nfs_sync_mapping(mapping);
 979         if (retval)
 980                 goto out;
 981
 982         task_io_account_write(count);
 983
 984         retval = nfs_direct_write(iocb, iov, nr_segs, pos, count, uio);
 985         if (retval > 0) {
 986                 struct inode *inode = mapping->host;
 987
 988                 iocb->ki_pos = pos + retval;
 989                 spin_lock(&inode->i_lock);
 990                 if (i_size_read(inode) < iocb->ki_pos)
 991                         i_size_write(inode, iocb->ki_pos);
 992                 spin_unlock(&inode->i_lock);
 993         }
 994 out:
 995         return retval;
 996 }
 997
 998 /**
 999  * nfs_init_directcache - create a slab cache for nfs_direct_req structures
1000  *
1001  */
1002 int __init nfs_init_directcache(void)
1003 {
1004         nfs_direct_cachep = kmem_cache_create("nfs_direct_cache",
1005                                                 sizeof(struct nfs_direct_req),
1006                                                 0, (SLAB_RECLAIM_ACCOUNT|
1007                                                         SLAB_MEM_SPREAD),
1008                                                 NULL);
1009         if (nfs_direct_cachep == NULL)
1010                 return -ENOMEM;
1011
1012         return 0;
1013 }
1014
1015 /**
1016  * nfs_destroy_directcache - destroy the slab cache for nfs_direct_req structures
1017  *
1018  */
1019 void nfs_destroy_directcache(void)
1020 {
1021         kmem_cache_destroy(nfs_direct_cachep);
1022 }