]>
Commit | Line | Data |
---|---|---|
1da177e4 LT |
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 | |
88467055 | 10 | * (multiple copies of the same instance running on separate hosts) |
1da177e4 | 11 | * implement their own cache coherency protocol that subsumes file |
88467055 CL |
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 | |
1da177e4 LT |
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 | |
88467055 | 37 | * 04 May 2005 support O_DIRECT with aio --cel |
1da177e4 LT |
38 | * |
39 | */ | |
40 | ||
1da177e4 LT |
41 | #include <linux/errno.h> |
42 | #include <linux/sched.h> | |
43 | #include <linux/kernel.h> | |
1da177e4 LT |
44 | #include <linux/file.h> |
45 | #include <linux/pagemap.h> | |
46 | #include <linux/kref.h> | |
47 | ||
48 | #include <linux/nfs_fs.h> | |
49 | #include <linux/nfs_page.h> | |
50 | #include <linux/sunrpc/clnt.h> | |
51 | ||
52 | #include <asm/system.h> | |
53 | #include <asm/uaccess.h> | |
54 | #include <asm/atomic.h> | |
55 | ||
8d5658c9 | 56 | #include "internal.h" |
91d5b470 CL |
57 | #include "iostat.h" |
58 | ||
1da177e4 | 59 | #define NFSDBG_FACILITY NFSDBG_VFS |
1da177e4 | 60 | |
e18b890b | 61 | static struct kmem_cache *nfs_direct_cachep; |
1da177e4 LT |
62 | |
63 | /* | |
64 | * This represents a set of asynchronous requests that we're waiting on | |
65 | */ | |
66 | struct nfs_direct_req { | |
67 | struct kref kref; /* release manager */ | |
15ce4a0c CL |
68 | |
69 | /* I/O parameters */ | |
a8881f5a | 70 | struct nfs_open_context *ctx; /* file open context info */ |
99514f8f | 71 | struct kiocb * iocb; /* controlling i/o request */ |
88467055 | 72 | struct inode * inode; /* target file of i/o */ |
15ce4a0c CL |
73 | |
74 | /* completion state */ | |
607f31e8 | 75 | atomic_t io_count; /* i/os we're waiting for */ |
15ce4a0c | 76 | spinlock_t lock; /* protect completion state */ |
15ce4a0c | 77 | ssize_t count, /* bytes actually processed */ |
1da177e4 | 78 | error; /* any reported error */ |
d72b7a6b | 79 | struct completion completion; /* wait for i/o completion */ |
fad61490 TM |
80 | |
81 | /* commit state */ | |
607f31e8 | 82 | struct list_head rewrite_list; /* saved nfs_write_data structs */ |
fad61490 TM |
83 | struct nfs_write_data * commit_data; /* special write_data for commits */ |
84 | int flags; | |
85 | #define NFS_ODIRECT_DO_COMMIT (1) /* an unstable reply was received */ | |
86 | #define NFS_ODIRECT_RESCHED_WRITES (2) /* write verification failed */ | |
87 | struct nfs_writeverf verf; /* unstable write verifier */ | |
1da177e4 LT |
88 | }; |
89 | ||
fad61490 | 90 | static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode); |
607f31e8 TM |
91 | static const struct rpc_call_ops nfs_write_direct_ops; |
92 | ||
93 | static inline void get_dreq(struct nfs_direct_req *dreq) | |
94 | { | |
95 | atomic_inc(&dreq->io_count); | |
96 | } | |
97 | ||
98 | static inline int put_dreq(struct nfs_direct_req *dreq) | |
99 | { | |
100 | return atomic_dec_and_test(&dreq->io_count); | |
101 | } | |
102 | ||
1da177e4 | 103 | /** |
b8a32e2b CL |
104 | * nfs_direct_IO - NFS address space operation for direct I/O |
105 | * @rw: direction (read or write) | |
106 | * @iocb: target I/O control block | |
107 | * @iov: array of vectors that define I/O buffer | |
108 | * @pos: offset in file to begin the operation | |
109 | * @nr_segs: size of iovec array | |
110 | * | |
111 | * The presence of this routine in the address space ops vector means | |
112 | * the NFS client supports direct I/O. However, we shunt off direct | |
113 | * read and write requests before the VFS gets them, so this method | |
114 | * should never be called. | |
1da177e4 | 115 | */ |
b8a32e2b CL |
116 | ssize_t nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, loff_t pos, unsigned long nr_segs) |
117 | { | |
b8a32e2b | 118 | dprintk("NFS: nfs_direct_IO (%s) off/no(%Ld/%lu) EINVAL\n", |
01cce933 | 119 | iocb->ki_filp->f_path.dentry->d_name.name, |
e99170ff | 120 | (long long) pos, nr_segs); |
b8a32e2b CL |
121 | |
122 | return -EINVAL; | |
123 | } | |
124 | ||
d4a8f367 | 125 | static void nfs_direct_dirty_pages(struct page **pages, unsigned int pgbase, size_t count) |
6b45d858 | 126 | { |
d4a8f367 | 127 | unsigned int npages; |
749e146e | 128 | unsigned int i; |
d4a8f367 TM |
129 | |
130 | if (count == 0) | |
131 | return; | |
132 | pages += (pgbase >> PAGE_SHIFT); | |
133 | npages = (count + (pgbase & ~PAGE_MASK) + PAGE_SIZE - 1) >> PAGE_SHIFT; | |
6b45d858 TM |
134 | for (i = 0; i < npages; i++) { |
135 | struct page *page = pages[i]; | |
607f31e8 | 136 | if (!PageCompound(page)) |
d4a8f367 | 137 | set_page_dirty(page); |
6b45d858 | 138 | } |
9c93ab7d CL |
139 | } |
140 | ||
749e146e | 141 | static void nfs_direct_release_pages(struct page **pages, unsigned int npages) |
9c93ab7d | 142 | { |
749e146e | 143 | unsigned int i; |
607f31e8 TM |
144 | for (i = 0; i < npages; i++) |
145 | page_cache_release(pages[i]); | |
6b45d858 TM |
146 | } |
147 | ||
93619e59 | 148 | static inline struct nfs_direct_req *nfs_direct_req_alloc(void) |
1da177e4 | 149 | { |
93619e59 CL |
150 | struct nfs_direct_req *dreq; |
151 | ||
e94b1766 | 152 | dreq = kmem_cache_alloc(nfs_direct_cachep, GFP_KERNEL); |
93619e59 CL |
153 | if (!dreq) |
154 | return NULL; | |
155 | ||
156 | kref_init(&dreq->kref); | |
607f31e8 | 157 | kref_get(&dreq->kref); |
d72b7a6b | 158 | init_completion(&dreq->completion); |
fad61490 | 159 | INIT_LIST_HEAD(&dreq->rewrite_list); |
93619e59 | 160 | dreq->iocb = NULL; |
a8881f5a | 161 | dreq->ctx = NULL; |
15ce4a0c | 162 | spin_lock_init(&dreq->lock); |
607f31e8 | 163 | atomic_set(&dreq->io_count, 0); |
15ce4a0c CL |
164 | dreq->count = 0; |
165 | dreq->error = 0; | |
fad61490 | 166 | dreq->flags = 0; |
93619e59 CL |
167 | |
168 | return dreq; | |
1da177e4 LT |
169 | } |
170 | ||
b4946ffb | 171 | static void nfs_direct_req_free(struct kref *kref) |
1da177e4 LT |
172 | { |
173 | struct nfs_direct_req *dreq = container_of(kref, struct nfs_direct_req, kref); | |
a8881f5a TM |
174 | |
175 | if (dreq->ctx != NULL) | |
176 | put_nfs_open_context(dreq->ctx); | |
1da177e4 LT |
177 | kmem_cache_free(nfs_direct_cachep, dreq); |
178 | } | |
179 | ||
b4946ffb TM |
180 | static void nfs_direct_req_release(struct nfs_direct_req *dreq) |
181 | { | |
182 | kref_put(&dreq->kref, nfs_direct_req_free); | |
183 | } | |
184 | ||
bc0fb201 CL |
185 | /* |
186 | * Collects and returns the final error value/byte-count. | |
187 | */ | |
188 | static ssize_t nfs_direct_wait(struct nfs_direct_req *dreq) | |
189 | { | |
15ce4a0c | 190 | ssize_t result = -EIOCBQUEUED; |
bc0fb201 CL |
191 | |
192 | /* Async requests don't wait here */ | |
193 | if (dreq->iocb) | |
194 | goto out; | |
195 | ||
150030b7 | 196 | result = wait_for_completion_killable(&dreq->completion); |
bc0fb201 CL |
197 | |
198 | if (!result) | |
15ce4a0c | 199 | result = dreq->error; |
bc0fb201 | 200 | if (!result) |
15ce4a0c | 201 | result = dreq->count; |
bc0fb201 CL |
202 | |
203 | out: | |
bc0fb201 CL |
204 | return (ssize_t) result; |
205 | } | |
206 | ||
63ab46ab | 207 | /* |
607f31e8 TM |
208 | * Synchronous I/O uses a stack-allocated iocb. Thus we can't trust |
209 | * the iocb is still valid here if this is a synchronous request. | |
63ab46ab CL |
210 | */ |
211 | static void nfs_direct_complete(struct nfs_direct_req *dreq) | |
212 | { | |
63ab46ab | 213 | if (dreq->iocb) { |
15ce4a0c | 214 | long res = (long) dreq->error; |
63ab46ab | 215 | if (!res) |
15ce4a0c | 216 | res = (long) dreq->count; |
63ab46ab | 217 | aio_complete(dreq->iocb, res, 0); |
d72b7a6b TM |
218 | } |
219 | complete_all(&dreq->completion); | |
63ab46ab | 220 | |
b4946ffb | 221 | nfs_direct_req_release(dreq); |
63ab46ab CL |
222 | } |
223 | ||
06cf6f2e | 224 | /* |
607f31e8 TM |
225 | * We must hold a reference to all the pages in this direct read request |
226 | * until the RPCs complete. This could be long *after* we are woken up in | |
227 | * nfs_direct_wait (for instance, if someone hits ^C on a slow server). | |
06cf6f2e | 228 | */ |
ec06c096 | 229 | static void nfs_direct_read_result(struct rpc_task *task, void *calldata) |
1da177e4 | 230 | { |
ec06c096 | 231 | struct nfs_read_data *data = calldata; |
1da177e4 | 232 | |
fdd1e74c TM |
233 | nfs_readpage_result(task, data); |
234 | } | |
235 | ||
236 | static void nfs_direct_read_release(void *calldata) | |
237 | { | |
238 | ||
239 | struct nfs_read_data *data = calldata; | |
240 | struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; | |
241 | int status = data->task.tk_status; | |
15ce4a0c CL |
242 | |
243 | spin_lock(&dreq->lock); | |
fdd1e74c TM |
244 | if (unlikely(status < 0)) { |
245 | dreq->error = status; | |
d4a8f367 TM |
246 | spin_unlock(&dreq->lock); |
247 | } else { | |
248 | dreq->count += data->res.count; | |
249 | spin_unlock(&dreq->lock); | |
250 | nfs_direct_dirty_pages(data->pagevec, | |
251 | data->args.pgbase, | |
252 | data->res.count); | |
253 | } | |
254 | nfs_direct_release_pages(data->pagevec, data->npages); | |
607f31e8 TM |
255 | |
256 | if (put_dreq(dreq)) | |
257 | nfs_direct_complete(dreq); | |
1ae88b2e | 258 | nfs_readdata_free(data); |
1da177e4 LT |
259 | } |
260 | ||
ec06c096 | 261 | static const struct rpc_call_ops nfs_read_direct_ops = { |
f11c88af AA |
262 | #if defined(CONFIG_NFS_V4_1) |
263 | .rpc_call_prepare = nfs_read_prepare, | |
264 | #endif /* CONFIG_NFS_V4_1 */ | |
ec06c096 | 265 | .rpc_call_done = nfs_direct_read_result, |
fdd1e74c | 266 | .rpc_release = nfs_direct_read_release, |
ec06c096 TM |
267 | }; |
268 | ||
d4cc948b | 269 | /* |
607f31e8 TM |
270 | * For each rsize'd chunk of the user's buffer, dispatch an NFS READ |
271 | * operation. If nfs_readdata_alloc() or get_user_pages() fails, | |
272 | * bail and stop sending more reads. Read length accounting is | |
273 | * handled automatically by nfs_direct_read_result(). Otherwise, if | |
274 | * no requests have been sent, just return an error. | |
1da177e4 | 275 | */ |
02fe4946 CL |
276 | static ssize_t nfs_direct_read_schedule_segment(struct nfs_direct_req *dreq, |
277 | const struct iovec *iov, | |
278 | loff_t pos) | |
1da177e4 | 279 | { |
a8881f5a | 280 | struct nfs_open_context *ctx = dreq->ctx; |
88be9f99 | 281 | struct inode *inode = ctx->path.dentry->d_inode; |
02fe4946 CL |
282 | unsigned long user_addr = (unsigned long)iov->iov_base; |
283 | size_t count = iov->iov_len; | |
5dd602f2 | 284 | size_t rsize = NFS_SERVER(inode)->rsize; |
07737691 | 285 | struct rpc_task *task; |
bdc7f021 TM |
286 | struct rpc_message msg = { |
287 | .rpc_cred = ctx->cred, | |
288 | }; | |
84115e1c TM |
289 | struct rpc_task_setup task_setup_data = { |
290 | .rpc_client = NFS_CLIENT(inode), | |
bdc7f021 | 291 | .rpc_message = &msg, |
84115e1c | 292 | .callback_ops = &nfs_read_direct_ops, |
101070ca | 293 | .workqueue = nfsiod_workqueue, |
84115e1c TM |
294 | .flags = RPC_TASK_ASYNC, |
295 | }; | |
607f31e8 TM |
296 | unsigned int pgbase; |
297 | int result; | |
298 | ssize_t started = 0; | |
299 | ||
1da177e4 | 300 | do { |
82b145c5 | 301 | struct nfs_read_data *data; |
5dd602f2 | 302 | size_t bytes; |
1da177e4 | 303 | |
e9f7bee1 TM |
304 | pgbase = user_addr & ~PAGE_MASK; |
305 | bytes = min(rsize,count); | |
306 | ||
607f31e8 | 307 | result = -ENOMEM; |
8d5658c9 | 308 | data = nfs_readdata_alloc(nfs_page_array_len(pgbase, bytes)); |
607f31e8 TM |
309 | if (unlikely(!data)) |
310 | break; | |
311 | ||
607f31e8 TM |
312 | down_read(¤t->mm->mmap_sem); |
313 | result = get_user_pages(current, current->mm, user_addr, | |
314 | data->npages, 1, 0, data->pagevec, NULL); | |
315 | up_read(¤t->mm->mmap_sem); | |
749e146e | 316 | if (result < 0) { |
1ae88b2e | 317 | nfs_readdata_free(data); |
749e146e CL |
318 | break; |
319 | } | |
320 | if ((unsigned)result < data->npages) { | |
d9df8d6b TM |
321 | bytes = result * PAGE_SIZE; |
322 | if (bytes <= pgbase) { | |
323 | nfs_direct_release_pages(data->pagevec, result); | |
1ae88b2e | 324 | nfs_readdata_free(data); |
d9df8d6b TM |
325 | break; |
326 | } | |
327 | bytes -= pgbase; | |
328 | data->npages = result; | |
607f31e8 TM |
329 | } |
330 | ||
331 | get_dreq(dreq); | |
82b145c5 | 332 | |
607f31e8 | 333 | data->req = (struct nfs_page *) dreq; |
1da177e4 | 334 | data->inode = inode; |
bdc7f021 | 335 | data->cred = msg.rpc_cred; |
1da177e4 | 336 | data->args.fh = NFS_FH(inode); |
1ae88b2e | 337 | data->args.context = ctx; |
88467055 | 338 | data->args.offset = pos; |
1da177e4 | 339 | data->args.pgbase = pgbase; |
607f31e8 | 340 | data->args.pages = data->pagevec; |
1da177e4 LT |
341 | data->args.count = bytes; |
342 | data->res.fattr = &data->fattr; | |
343 | data->res.eof = 0; | |
344 | data->res.count = bytes; | |
bdc7f021 TM |
345 | msg.rpc_argp = &data->args; |
346 | msg.rpc_resp = &data->res; | |
1da177e4 | 347 | |
07737691 | 348 | task_setup_data.task = &data->task; |
84115e1c | 349 | task_setup_data.callback_data = data; |
bdc7f021 | 350 | NFS_PROTO(inode)->read_setup(data, &msg); |
1da177e4 | 351 | |
07737691 | 352 | task = rpc_run_task(&task_setup_data); |
dbae4c73 TM |
353 | if (IS_ERR(task)) |
354 | break; | |
355 | rpc_put_task(task); | |
1da177e4 | 356 | |
a3f565b1 CL |
357 | dprintk("NFS: %5u initiated direct read call " |
358 | "(req %s/%Ld, %zu bytes @ offset %Lu)\n", | |
1da177e4 LT |
359 | data->task.tk_pid, |
360 | inode->i_sb->s_id, | |
361 | (long long)NFS_FILEID(inode), | |
362 | bytes, | |
363 | (unsigned long long)data->args.offset); | |
364 | ||
607f31e8 TM |
365 | started += bytes; |
366 | user_addr += bytes; | |
88467055 | 367 | pos += bytes; |
e9f7bee1 | 368 | /* FIXME: Remove this unnecessary math from final patch */ |
1da177e4 | 369 | pgbase += bytes; |
1da177e4 | 370 | pgbase &= ~PAGE_MASK; |
e9f7bee1 | 371 | BUG_ON(pgbase != (user_addr & ~PAGE_MASK)); |
1da177e4 LT |
372 | |
373 | count -= bytes; | |
374 | } while (count != 0); | |
607f31e8 | 375 | |
607f31e8 | 376 | if (started) |
c216fd70 | 377 | return started; |
607f31e8 | 378 | return result < 0 ? (ssize_t) result : -EFAULT; |
1da177e4 LT |
379 | } |
380 | ||
19f73787 CL |
381 | static ssize_t nfs_direct_read_schedule_iovec(struct nfs_direct_req *dreq, |
382 | const struct iovec *iov, | |
383 | unsigned long nr_segs, | |
384 | loff_t pos) | |
385 | { | |
386 | ssize_t result = -EINVAL; | |
387 | size_t requested_bytes = 0; | |
388 | unsigned long seg; | |
389 | ||
390 | get_dreq(dreq); | |
391 | ||
392 | for (seg = 0; seg < nr_segs; seg++) { | |
393 | const struct iovec *vec = &iov[seg]; | |
02fe4946 | 394 | result = nfs_direct_read_schedule_segment(dreq, vec, pos); |
19f73787 CL |
395 | if (result < 0) |
396 | break; | |
397 | requested_bytes += result; | |
398 | if ((size_t)result < vec->iov_len) | |
399 | break; | |
400 | pos += vec->iov_len; | |
401 | } | |
402 | ||
403 | if (put_dreq(dreq)) | |
404 | nfs_direct_complete(dreq); | |
405 | ||
406 | if (requested_bytes != 0) | |
407 | return 0; | |
408 | ||
409 | if (result < 0) | |
410 | return result; | |
411 | return -EIO; | |
412 | } | |
413 | ||
c216fd70 CL |
414 | static ssize_t nfs_direct_read(struct kiocb *iocb, const struct iovec *iov, |
415 | unsigned long nr_segs, loff_t pos) | |
1da177e4 | 416 | { |
607f31e8 | 417 | ssize_t result = 0; |
99514f8f | 418 | struct inode *inode = iocb->ki_filp->f_mapping->host; |
1da177e4 LT |
419 | struct nfs_direct_req *dreq; |
420 | ||
607f31e8 | 421 | dreq = nfs_direct_req_alloc(); |
1da177e4 LT |
422 | if (!dreq) |
423 | return -ENOMEM; | |
424 | ||
91d5b470 | 425 | dreq->inode = inode; |
cd3758e3 | 426 | dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp)); |
487b8372 CL |
427 | if (!is_sync_kiocb(iocb)) |
428 | dreq->iocb = iocb; | |
1da177e4 | 429 | |
c216fd70 | 430 | result = nfs_direct_read_schedule_iovec(dreq, iov, nr_segs, pos); |
607f31e8 TM |
431 | if (!result) |
432 | result = nfs_direct_wait(dreq); | |
b4946ffb | 433 | nfs_direct_req_release(dreq); |
1da177e4 LT |
434 | |
435 | return result; | |
436 | } | |
437 | ||
fad61490 | 438 | static void nfs_direct_free_writedata(struct nfs_direct_req *dreq) |
1da177e4 | 439 | { |
607f31e8 TM |
440 | while (!list_empty(&dreq->rewrite_list)) { |
441 | struct nfs_write_data *data = list_entry(dreq->rewrite_list.next, struct nfs_write_data, pages); | |
fad61490 | 442 | list_del(&data->pages); |
607f31e8 | 443 | nfs_direct_release_pages(data->pagevec, data->npages); |
1ae88b2e | 444 | nfs_writedata_free(data); |
fad61490 TM |
445 | } |
446 | } | |
1da177e4 | 447 | |
fad61490 TM |
448 | #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4) |
449 | static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) | |
450 | { | |
607f31e8 TM |
451 | struct inode *inode = dreq->inode; |
452 | struct list_head *p; | |
453 | struct nfs_write_data *data; | |
07737691 | 454 | struct rpc_task *task; |
bdc7f021 TM |
455 | struct rpc_message msg = { |
456 | .rpc_cred = dreq->ctx->cred, | |
457 | }; | |
84115e1c TM |
458 | struct rpc_task_setup task_setup_data = { |
459 | .rpc_client = NFS_CLIENT(inode), | |
460 | .callback_ops = &nfs_write_direct_ops, | |
101070ca | 461 | .workqueue = nfsiod_workqueue, |
84115e1c TM |
462 | .flags = RPC_TASK_ASYNC, |
463 | }; | |
1da177e4 | 464 | |
fad61490 | 465 | dreq->count = 0; |
607f31e8 TM |
466 | get_dreq(dreq); |
467 | ||
468 | list_for_each(p, &dreq->rewrite_list) { | |
469 | data = list_entry(p, struct nfs_write_data, pages); | |
470 | ||
471 | get_dreq(dreq); | |
472 | ||
bdc7f021 TM |
473 | /* Use stable writes */ |
474 | data->args.stable = NFS_FILE_SYNC; | |
475 | ||
607f31e8 TM |
476 | /* |
477 | * Reset data->res. | |
478 | */ | |
479 | nfs_fattr_init(&data->fattr); | |
480 | data->res.count = data->args.count; | |
481 | memset(&data->verf, 0, sizeof(data->verf)); | |
482 | ||
483 | /* | |
484 | * Reuse data->task; data->args should not have changed | |
485 | * since the original request was sent. | |
486 | */ | |
07737691 | 487 | task_setup_data.task = &data->task; |
84115e1c | 488 | task_setup_data.callback_data = data; |
bdc7f021 TM |
489 | msg.rpc_argp = &data->args; |
490 | msg.rpc_resp = &data->res; | |
491 | NFS_PROTO(inode)->write_setup(data, &msg); | |
607f31e8 | 492 | |
607f31e8 TM |
493 | /* |
494 | * We're called via an RPC callback, so BKL is already held. | |
495 | */ | |
07737691 TM |
496 | task = rpc_run_task(&task_setup_data); |
497 | if (!IS_ERR(task)) | |
498 | rpc_put_task(task); | |
607f31e8 TM |
499 | |
500 | dprintk("NFS: %5u rescheduled direct write call (req %s/%Ld, %u bytes @ offset %Lu)\n", | |
501 | data->task.tk_pid, | |
502 | inode->i_sb->s_id, | |
503 | (long long)NFS_FILEID(inode), | |
504 | data->args.count, | |
505 | (unsigned long long)data->args.offset); | |
506 | } | |
fedb595c | 507 | |
607f31e8 TM |
508 | if (put_dreq(dreq)) |
509 | nfs_direct_write_complete(dreq, inode); | |
fad61490 TM |
510 | } |
511 | ||
512 | static void nfs_direct_commit_result(struct rpc_task *task, void *calldata) | |
513 | { | |
514 | struct nfs_write_data *data = calldata; | |
fad61490 TM |
515 | |
516 | /* Call the NFS version-specific code */ | |
c9d8f89d TM |
517 | NFS_PROTO(data->inode)->commit_done(task, data); |
518 | } | |
519 | ||
520 | static void nfs_direct_commit_release(void *calldata) | |
521 | { | |
522 | struct nfs_write_data *data = calldata; | |
523 | struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; | |
524 | int status = data->task.tk_status; | |
525 | ||
526 | if (status < 0) { | |
60fa3f76 | 527 | dprintk("NFS: %5u commit failed with error %d.\n", |
c9d8f89d | 528 | data->task.tk_pid, status); |
fad61490 | 529 | dreq->flags = NFS_ODIRECT_RESCHED_WRITES; |
60fa3f76 | 530 | } else if (memcmp(&dreq->verf, &data->verf, sizeof(data->verf))) { |
c9d8f89d | 531 | dprintk("NFS: %5u commit verify failed\n", data->task.tk_pid); |
fad61490 | 532 | dreq->flags = NFS_ODIRECT_RESCHED_WRITES; |
1da177e4 LT |
533 | } |
534 | ||
c9d8f89d | 535 | dprintk("NFS: %5u commit returned %d\n", data->task.tk_pid, status); |
fad61490 | 536 | nfs_direct_write_complete(dreq, data->inode); |
1ae88b2e | 537 | nfs_commit_free(data); |
1da177e4 LT |
538 | } |
539 | ||
fad61490 | 540 | static const struct rpc_call_ops nfs_commit_direct_ops = { |
21d9a851 AA |
541 | #if defined(CONFIG_NFS_V4_1) |
542 | .rpc_call_prepare = nfs_write_prepare, | |
543 | #endif /* CONFIG_NFS_V4_1 */ | |
fad61490 | 544 | .rpc_call_done = nfs_direct_commit_result, |
c9d8f89d | 545 | .rpc_release = nfs_direct_commit_release, |
fad61490 TM |
546 | }; |
547 | ||
548 | static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq) | |
1da177e4 | 549 | { |
fad61490 | 550 | struct nfs_write_data *data = dreq->commit_data; |
07737691 | 551 | struct rpc_task *task; |
bdc7f021 TM |
552 | struct rpc_message msg = { |
553 | .rpc_argp = &data->args, | |
554 | .rpc_resp = &data->res, | |
555 | .rpc_cred = dreq->ctx->cred, | |
556 | }; | |
84115e1c | 557 | struct rpc_task_setup task_setup_data = { |
07737691 | 558 | .task = &data->task, |
84115e1c | 559 | .rpc_client = NFS_CLIENT(dreq->inode), |
bdc7f021 | 560 | .rpc_message = &msg, |
84115e1c TM |
561 | .callback_ops = &nfs_commit_direct_ops, |
562 | .callback_data = data, | |
101070ca | 563 | .workqueue = nfsiod_workqueue, |
84115e1c TM |
564 | .flags = RPC_TASK_ASYNC, |
565 | }; | |
1da177e4 | 566 | |
fad61490 | 567 | data->inode = dreq->inode; |
bdc7f021 | 568 | data->cred = msg.rpc_cred; |
1da177e4 | 569 | |
fad61490 | 570 | data->args.fh = NFS_FH(data->inode); |
607f31e8 TM |
571 | data->args.offset = 0; |
572 | data->args.count = 0; | |
1ae88b2e | 573 | data->args.context = dreq->ctx; |
fad61490 TM |
574 | data->res.count = 0; |
575 | data->res.fattr = &data->fattr; | |
576 | data->res.verf = &data->verf; | |
1da177e4 | 577 | |
bdc7f021 | 578 | NFS_PROTO(data->inode)->commit_setup(data, &msg); |
1da177e4 | 579 | |
fad61490 TM |
580 | /* Note: task.tk_ops->rpc_release will free dreq->commit_data */ |
581 | dreq->commit_data = NULL; | |
1da177e4 | 582 | |
e99170ff | 583 | dprintk("NFS: %5u initiated commit call\n", data->task.tk_pid); |
1da177e4 | 584 | |
07737691 TM |
585 | task = rpc_run_task(&task_setup_data); |
586 | if (!IS_ERR(task)) | |
587 | rpc_put_task(task); | |
fad61490 | 588 | } |
1da177e4 | 589 | |
fad61490 TM |
590 | static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode) |
591 | { | |
592 | int flags = dreq->flags; | |
1da177e4 | 593 | |
fad61490 TM |
594 | dreq->flags = 0; |
595 | switch (flags) { | |
596 | case NFS_ODIRECT_DO_COMMIT: | |
597 | nfs_direct_commit_schedule(dreq); | |
1da177e4 | 598 | break; |
fad61490 TM |
599 | case NFS_ODIRECT_RESCHED_WRITES: |
600 | nfs_direct_write_reschedule(dreq); | |
601 | break; | |
602 | default: | |
fad61490 TM |
603 | if (dreq->commit_data != NULL) |
604 | nfs_commit_free(dreq->commit_data); | |
605 | nfs_direct_free_writedata(dreq); | |
cd9ae2b6 | 606 | nfs_zap_mapping(inode, inode->i_mapping); |
fad61490 TM |
607 | nfs_direct_complete(dreq); |
608 | } | |
609 | } | |
1da177e4 | 610 | |
fad61490 TM |
611 | static void nfs_alloc_commit_data(struct nfs_direct_req *dreq) |
612 | { | |
c9d8f89d | 613 | dreq->commit_data = nfs_commitdata_alloc(); |
fad61490 TM |
614 | if (dreq->commit_data != NULL) |
615 | dreq->commit_data->req = (struct nfs_page *) dreq; | |
616 | } | |
617 | #else | |
618 | static inline void nfs_alloc_commit_data(struct nfs_direct_req *dreq) | |
619 | { | |
620 | dreq->commit_data = NULL; | |
621 | } | |
1da177e4 | 622 | |
fad61490 TM |
623 | static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode) |
624 | { | |
fad61490 | 625 | nfs_direct_free_writedata(dreq); |
cd9ae2b6 | 626 | nfs_zap_mapping(inode, inode->i_mapping); |
fad61490 TM |
627 | nfs_direct_complete(dreq); |
628 | } | |
629 | #endif | |
1da177e4 | 630 | |
462d5b32 | 631 | static void nfs_direct_write_result(struct rpc_task *task, void *calldata) |
1da177e4 | 632 | { |
462d5b32 | 633 | struct nfs_write_data *data = calldata; |
462d5b32 CL |
634 | |
635 | if (nfs_writeback_done(task, data) != 0) | |
636 | return; | |
c9d8f89d TM |
637 | } |
638 | ||
639 | /* | |
640 | * NB: Return the value of the first error return code. Subsequent | |
641 | * errors after the first one are ignored. | |
642 | */ | |
643 | static void nfs_direct_write_release(void *calldata) | |
644 | { | |
645 | struct nfs_write_data *data = calldata; | |
646 | struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; | |
647 | int status = data->task.tk_status; | |
462d5b32 | 648 | |
15ce4a0c | 649 | spin_lock(&dreq->lock); |
1da177e4 | 650 | |
eda3cef8 | 651 | if (unlikely(status < 0)) { |
432409ee | 652 | /* An error has occurred, so we should not commit */ |
60fa3f76 | 653 | dreq->flags = 0; |
eda3cef8 | 654 | dreq->error = status; |
eda3cef8 | 655 | } |
432409ee NB |
656 | if (unlikely(dreq->error != 0)) |
657 | goto out_unlock; | |
eda3cef8 TM |
658 | |
659 | dreq->count += data->res.count; | |
1da177e4 | 660 | |
fad61490 TM |
661 | if (data->res.verf->committed != NFS_FILE_SYNC) { |
662 | switch (dreq->flags) { | |
663 | case 0: | |
664 | memcpy(&dreq->verf, &data->verf, sizeof(dreq->verf)); | |
665 | dreq->flags = NFS_ODIRECT_DO_COMMIT; | |
1da177e4 | 666 | break; |
fad61490 TM |
667 | case NFS_ODIRECT_DO_COMMIT: |
668 | if (memcmp(&dreq->verf, &data->verf, sizeof(dreq->verf))) { | |
c9d8f89d | 669 | dprintk("NFS: %5u write verify failed\n", data->task.tk_pid); |
fad61490 TM |
670 | dreq->flags = NFS_ODIRECT_RESCHED_WRITES; |
671 | } | |
1da177e4 | 672 | } |
1da177e4 | 673 | } |
eda3cef8 | 674 | out_unlock: |
fad61490 | 675 | spin_unlock(&dreq->lock); |
1da177e4 | 676 | |
607f31e8 TM |
677 | if (put_dreq(dreq)) |
678 | nfs_direct_write_complete(dreq, data->inode); | |
462d5b32 CL |
679 | } |
680 | ||
681 | static const struct rpc_call_ops nfs_write_direct_ops = { | |
def6ed7e AA |
682 | #if defined(CONFIG_NFS_V4_1) |
683 | .rpc_call_prepare = nfs_write_prepare, | |
684 | #endif /* CONFIG_NFS_V4_1 */ | |
462d5b32 | 685 | .rpc_call_done = nfs_direct_write_result, |
fad61490 | 686 | .rpc_release = nfs_direct_write_release, |
462d5b32 CL |
687 | }; |
688 | ||
689 | /* | |
607f31e8 TM |
690 | * For each wsize'd chunk of the user's buffer, dispatch an NFS WRITE |
691 | * operation. If nfs_writedata_alloc() or get_user_pages() fails, | |
692 | * bail and stop sending more writes. Write length accounting is | |
693 | * handled automatically by nfs_direct_write_result(). Otherwise, if | |
694 | * no requests have been sent, just return an error. | |
462d5b32 | 695 | */ |
02fe4946 CL |
696 | static ssize_t nfs_direct_write_schedule_segment(struct nfs_direct_req *dreq, |
697 | const struct iovec *iov, | |
698 | loff_t pos, int sync) | |
462d5b32 | 699 | { |
a8881f5a | 700 | struct nfs_open_context *ctx = dreq->ctx; |
88be9f99 | 701 | struct inode *inode = ctx->path.dentry->d_inode; |
02fe4946 CL |
702 | unsigned long user_addr = (unsigned long)iov->iov_base; |
703 | size_t count = iov->iov_len; | |
07737691 | 704 | struct rpc_task *task; |
bdc7f021 TM |
705 | struct rpc_message msg = { |
706 | .rpc_cred = ctx->cred, | |
707 | }; | |
84115e1c TM |
708 | struct rpc_task_setup task_setup_data = { |
709 | .rpc_client = NFS_CLIENT(inode), | |
bdc7f021 | 710 | .rpc_message = &msg, |
84115e1c | 711 | .callback_ops = &nfs_write_direct_ops, |
101070ca | 712 | .workqueue = nfsiod_workqueue, |
84115e1c TM |
713 | .flags = RPC_TASK_ASYNC, |
714 | }; | |
462d5b32 | 715 | size_t wsize = NFS_SERVER(inode)->wsize; |
607f31e8 TM |
716 | unsigned int pgbase; |
717 | int result; | |
718 | ssize_t started = 0; | |
82b145c5 | 719 | |
1da177e4 | 720 | do { |
82b145c5 | 721 | struct nfs_write_data *data; |
462d5b32 CL |
722 | size_t bytes; |
723 | ||
e9f7bee1 TM |
724 | pgbase = user_addr & ~PAGE_MASK; |
725 | bytes = min(wsize,count); | |
726 | ||
607f31e8 | 727 | result = -ENOMEM; |
8d5658c9 | 728 | data = nfs_writedata_alloc(nfs_page_array_len(pgbase, bytes)); |
607f31e8 TM |
729 | if (unlikely(!data)) |
730 | break; | |
731 | ||
607f31e8 TM |
732 | down_read(¤t->mm->mmap_sem); |
733 | result = get_user_pages(current, current->mm, user_addr, | |
734 | data->npages, 0, 0, data->pagevec, NULL); | |
735 | up_read(¤t->mm->mmap_sem); | |
749e146e | 736 | if (result < 0) { |
1ae88b2e | 737 | nfs_writedata_free(data); |
749e146e CL |
738 | break; |
739 | } | |
740 | if ((unsigned)result < data->npages) { | |
d9df8d6b TM |
741 | bytes = result * PAGE_SIZE; |
742 | if (bytes <= pgbase) { | |
743 | nfs_direct_release_pages(data->pagevec, result); | |
1ae88b2e | 744 | nfs_writedata_free(data); |
d9df8d6b TM |
745 | break; |
746 | } | |
747 | bytes -= pgbase; | |
748 | data->npages = result; | |
607f31e8 TM |
749 | } |
750 | ||
751 | get_dreq(dreq); | |
752 | ||
fad61490 | 753 | list_move_tail(&data->pages, &dreq->rewrite_list); |
462d5b32 | 754 | |
607f31e8 | 755 | data->req = (struct nfs_page *) dreq; |
462d5b32 | 756 | data->inode = inode; |
bdc7f021 | 757 | data->cred = msg.rpc_cred; |
462d5b32 | 758 | data->args.fh = NFS_FH(inode); |
1ae88b2e | 759 | data->args.context = ctx; |
88467055 | 760 | data->args.offset = pos; |
462d5b32 | 761 | data->args.pgbase = pgbase; |
607f31e8 | 762 | data->args.pages = data->pagevec; |
462d5b32 | 763 | data->args.count = bytes; |
bdc7f021 | 764 | data->args.stable = sync; |
462d5b32 CL |
765 | data->res.fattr = &data->fattr; |
766 | data->res.count = bytes; | |
47989d74 | 767 | data->res.verf = &data->verf; |
462d5b32 | 768 | |
07737691 | 769 | task_setup_data.task = &data->task; |
84115e1c | 770 | task_setup_data.callback_data = data; |
bdc7f021 TM |
771 | msg.rpc_argp = &data->args; |
772 | msg.rpc_resp = &data->res; | |
773 | NFS_PROTO(inode)->write_setup(data, &msg); | |
1da177e4 | 774 | |
07737691 | 775 | task = rpc_run_task(&task_setup_data); |
dbae4c73 TM |
776 | if (IS_ERR(task)) |
777 | break; | |
778 | rpc_put_task(task); | |
1da177e4 | 779 | |
a3f565b1 CL |
780 | dprintk("NFS: %5u initiated direct write call " |
781 | "(req %s/%Ld, %zu bytes @ offset %Lu)\n", | |
462d5b32 CL |
782 | data->task.tk_pid, |
783 | inode->i_sb->s_id, | |
784 | (long long)NFS_FILEID(inode), | |
785 | bytes, | |
786 | (unsigned long long)data->args.offset); | |
1da177e4 | 787 | |
607f31e8 TM |
788 | started += bytes; |
789 | user_addr += bytes; | |
88467055 | 790 | pos += bytes; |
e9f7bee1 TM |
791 | |
792 | /* FIXME: Remove this useless math from the final patch */ | |
462d5b32 | 793 | pgbase += bytes; |
462d5b32 | 794 | pgbase &= ~PAGE_MASK; |
e9f7bee1 | 795 | BUG_ON(pgbase != (user_addr & ~PAGE_MASK)); |
1da177e4 | 796 | |
462d5b32 CL |
797 | count -= bytes; |
798 | } while (count != 0); | |
607f31e8 | 799 | |
607f31e8 | 800 | if (started) |
c216fd70 | 801 | return started; |
607f31e8 | 802 | return result < 0 ? (ssize_t) result : -EFAULT; |
462d5b32 | 803 | } |
1da177e4 | 804 | |
19f73787 CL |
805 | static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq, |
806 | const struct iovec *iov, | |
807 | unsigned long nr_segs, | |
808 | loff_t pos, int sync) | |
809 | { | |
810 | ssize_t result = 0; | |
811 | size_t requested_bytes = 0; | |
812 | unsigned long seg; | |
813 | ||
814 | get_dreq(dreq); | |
815 | ||
816 | for (seg = 0; seg < nr_segs; seg++) { | |
817 | const struct iovec *vec = &iov[seg]; | |
02fe4946 CL |
818 | result = nfs_direct_write_schedule_segment(dreq, vec, |
819 | pos, sync); | |
19f73787 CL |
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 | ||
828 | if (put_dreq(dreq)) | |
829 | nfs_direct_write_complete(dreq, dreq->inode); | |
830 | ||
831 | if (requested_bytes != 0) | |
832 | return 0; | |
833 | ||
834 | if (result < 0) | |
835 | return result; | |
836 | return -EIO; | |
837 | } | |
838 | ||
c216fd70 CL |
839 | static ssize_t nfs_direct_write(struct kiocb *iocb, const struct iovec *iov, |
840 | unsigned long nr_segs, loff_t pos, | |
841 | size_t count) | |
462d5b32 | 842 | { |
607f31e8 | 843 | ssize_t result = 0; |
c89f2ee5 | 844 | struct inode *inode = iocb->ki_filp->f_mapping->host; |
462d5b32 | 845 | struct nfs_direct_req *dreq; |
fad61490 | 846 | size_t wsize = NFS_SERVER(inode)->wsize; |
bdc7f021 | 847 | int sync = NFS_UNSTABLE; |
1da177e4 | 848 | |
607f31e8 | 849 | dreq = nfs_direct_req_alloc(); |
462d5b32 CL |
850 | if (!dreq) |
851 | return -ENOMEM; | |
607f31e8 TM |
852 | nfs_alloc_commit_data(dreq); |
853 | ||
fad61490 | 854 | if (dreq->commit_data == NULL || count < wsize) |
bdc7f021 | 855 | sync = NFS_FILE_SYNC; |
1da177e4 | 856 | |
c89f2ee5 | 857 | dreq->inode = inode; |
cd3758e3 | 858 | dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp)); |
c89f2ee5 CL |
859 | if (!is_sync_kiocb(iocb)) |
860 | dreq->iocb = iocb; | |
1da177e4 | 861 | |
c216fd70 | 862 | result = nfs_direct_write_schedule_iovec(dreq, iov, nr_segs, pos, sync); |
607f31e8 TM |
863 | if (!result) |
864 | result = nfs_direct_wait(dreq); | |
b4946ffb | 865 | nfs_direct_req_release(dreq); |
1da177e4 | 866 | |
1da177e4 LT |
867 | return result; |
868 | } | |
869 | ||
870 | /** | |
871 | * nfs_file_direct_read - file direct read operation for NFS files | |
872 | * @iocb: target I/O control block | |
027445c3 BP |
873 | * @iov: vector of user buffers into which to read data |
874 | * @nr_segs: size of iov vector | |
88467055 | 875 | * @pos: byte offset in file where reading starts |
1da177e4 LT |
876 | * |
877 | * We use this function for direct reads instead of calling | |
878 | * generic_file_aio_read() in order to avoid gfar's check to see if | |
879 | * the request starts before the end of the file. For that check | |
880 | * to work, we must generate a GETATTR before each direct read, and | |
881 | * even then there is a window between the GETATTR and the subsequent | |
88467055 | 882 | * READ where the file size could change. Our preference is simply |
1da177e4 LT |
883 | * to do all reads the application wants, and the server will take |
884 | * care of managing the end of file boundary. | |
88467055 | 885 | * |
1da177e4 LT |
886 | * This function also eliminates unnecessarily updating the file's |
887 | * atime locally, as the NFS server sets the file's atime, and this | |
888 | * client must read the updated atime from the server back into its | |
889 | * cache. | |
890 | */ | |
027445c3 BP |
891 | ssize_t nfs_file_direct_read(struct kiocb *iocb, const struct iovec *iov, |
892 | unsigned long nr_segs, loff_t pos) | |
1da177e4 LT |
893 | { |
894 | ssize_t retval = -EINVAL; | |
1da177e4 | 895 | struct file *file = iocb->ki_filp; |
1da177e4 | 896 | struct address_space *mapping = file->f_mapping; |
c216fd70 CL |
897 | size_t count; |
898 | ||
899 | count = iov_length(iov, nr_segs); | |
900 | nfs_add_stats(mapping->host, NFSIOS_DIRECTREADBYTES, count); | |
1da177e4 | 901 | |
6da24bc9 | 902 | dfprintk(FILE, "NFS: direct read(%s/%s, %zd@%Ld)\n", |
01cce933 JJS |
903 | file->f_path.dentry->d_parent->d_name.name, |
904 | file->f_path.dentry->d_name.name, | |
c216fd70 | 905 | count, (long long) pos); |
1da177e4 | 906 | |
1da177e4 LT |
907 | retval = 0; |
908 | if (!count) | |
909 | goto out; | |
910 | ||
29884df0 TM |
911 | retval = nfs_sync_mapping(mapping); |
912 | if (retval) | |
913 | goto out; | |
1da177e4 | 914 | |
c216fd70 | 915 | retval = nfs_direct_read(iocb, iov, nr_segs, pos); |
1da177e4 | 916 | if (retval > 0) |
0cdd80d0 | 917 | iocb->ki_pos = pos + retval; |
1da177e4 LT |
918 | |
919 | out: | |
920 | return retval; | |
921 | } | |
922 | ||
923 | /** | |
924 | * nfs_file_direct_write - file direct write operation for NFS files | |
925 | * @iocb: target I/O control block | |
027445c3 BP |
926 | * @iov: vector of user buffers from which to write data |
927 | * @nr_segs: size of iov vector | |
88467055 | 928 | * @pos: byte offset in file where writing starts |
1da177e4 LT |
929 | * |
930 | * We use this function for direct writes instead of calling | |
931 | * generic_file_aio_write() in order to avoid taking the inode | |
932 | * semaphore and updating the i_size. The NFS server will set | |
933 | * the new i_size and this client must read the updated size | |
934 | * back into its cache. We let the server do generic write | |
935 | * parameter checking and report problems. | |
936 | * | |
937 | * We also avoid an unnecessary invocation of generic_osync_inode(), | |
938 | * as it is fairly meaningless to sync the metadata of an NFS file. | |
939 | * | |
940 | * We eliminate local atime updates, see direct read above. | |
941 | * | |
942 | * We avoid unnecessary page cache invalidations for normal cached | |
943 | * readers of this file. | |
944 | * | |
945 | * Note that O_APPEND is not supported for NFS direct writes, as there | |
946 | * is no atomic O_APPEND write facility in the NFS protocol. | |
947 | */ | |
027445c3 BP |
948 | ssize_t nfs_file_direct_write(struct kiocb *iocb, const struct iovec *iov, |
949 | unsigned long nr_segs, loff_t pos) | |
1da177e4 | 950 | { |
070ea602 | 951 | ssize_t retval = -EINVAL; |
1da177e4 | 952 | struct file *file = iocb->ki_filp; |
1da177e4 | 953 | struct address_space *mapping = file->f_mapping; |
c216fd70 | 954 | size_t count; |
1da177e4 | 955 | |
c216fd70 CL |
956 | count = iov_length(iov, nr_segs); |
957 | nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, count); | |
958 | ||
6da24bc9 | 959 | dfprintk(FILE, "NFS: direct write(%s/%s, %zd@%Ld)\n", |
01cce933 JJS |
960 | file->f_path.dentry->d_parent->d_name.name, |
961 | file->f_path.dentry->d_name.name, | |
c216fd70 | 962 | count, (long long) pos); |
027445c3 | 963 | |
ce1a8e67 CL |
964 | retval = generic_write_checks(file, &pos, &count, 0); |
965 | if (retval) | |
1da177e4 | 966 | goto out; |
ce1a8e67 CL |
967 | |
968 | retval = -EINVAL; | |
969 | if ((ssize_t) count < 0) | |
1da177e4 | 970 | goto out; |
1da177e4 LT |
971 | retval = 0; |
972 | if (!count) | |
973 | goto out; | |
ce1a8e67 | 974 | |
29884df0 TM |
975 | retval = nfs_sync_mapping(mapping); |
976 | if (retval) | |
977 | goto out; | |
1da177e4 | 978 | |
c216fd70 | 979 | retval = nfs_direct_write(iocb, iov, nr_segs, pos, count); |
9eafa8cc | 980 | |
1da177e4 | 981 | if (retval > 0) |
ce1a8e67 | 982 | iocb->ki_pos = pos + retval; |
1da177e4 LT |
983 | |
984 | out: | |
985 | return retval; | |
986 | } | |
987 | ||
88467055 CL |
988 | /** |
989 | * nfs_init_directcache - create a slab cache for nfs_direct_req structures | |
990 | * | |
991 | */ | |
f7b422b1 | 992 | int __init nfs_init_directcache(void) |
1da177e4 LT |
993 | { |
994 | nfs_direct_cachep = kmem_cache_create("nfs_direct_cache", | |
995 | sizeof(struct nfs_direct_req), | |
fffb60f9 PJ |
996 | 0, (SLAB_RECLAIM_ACCOUNT| |
997 | SLAB_MEM_SPREAD), | |
20c2df83 | 998 | NULL); |
1da177e4 LT |
999 | if (nfs_direct_cachep == NULL) |
1000 | return -ENOMEM; | |
1001 | ||
1002 | return 0; | |
1003 | } | |
1004 | ||
88467055 | 1005 | /** |
f7b422b1 | 1006 | * nfs_destroy_directcache - destroy the slab cache for nfs_direct_req structures |
88467055 CL |
1007 | * |
1008 | */ | |
266bee88 | 1009 | void nfs_destroy_directcache(void) |
1da177e4 | 1010 | { |
1a1d92c1 | 1011 | kmem_cache_destroy(nfs_direct_cachep); |
1da177e4 | 1012 | } |