file.c 28.9 KB
Newer Older
1
#include <linux/ceph/ceph_debug.h>
S
Sage Weil 已提交
2

3
#include <linux/module.h>
S
Sage Weil 已提交
4
#include <linux/sched.h>
5
#include <linux/slab.h>
S
Sage Weil 已提交
6
#include <linux/file.h>
S
Sage Weil 已提交
7
#include <linux/mount.h>
S
Sage Weil 已提交
8 9
#include <linux/namei.h>
#include <linux/writeback.h>
10
#include <linux/aio.h>
L
Li Wang 已提交
11
#include <linux/falloc.h>
S
Sage Weil 已提交
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

#include "super.h"
#include "mds_client.h"

/*
 * Ceph file operations
 *
 * Implement basic open/close functionality, and implement
 * read/write.
 *
 * We implement three modes of file I/O:
 *  - buffered uses the generic_file_aio_{read,write} helpers
 *
 *  - synchronous is used when there is multi-client read/write
 *    sharing, avoids the page cache, and synchronously waits for an
 *    ack from the OSD.
 *
 *  - direct io takes the variant of the sync path that references
 *    user pages directly.
 *
 * fsync() flushes and waits on dirty pages, but just queues metadata
 * for writeback: since the MDS can recover size and mtime there is no
 * need to wait for MDS acknowledgement.
 */


/*
 * Prepare an open request.  Preallocate ceph_cap to avoid an
 * inopportune ENOMEM later.
 */
static struct ceph_mds_request *
prepare_open_request(struct super_block *sb, int flags, int create_mode)
{
45 46
	struct ceph_fs_client *fsc = ceph_sb_to_client(sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
	struct ceph_mds_request *req;
	int want_auth = USE_ANY_MDS;
	int op = (flags & O_CREAT) ? CEPH_MDS_OP_CREATE : CEPH_MDS_OP_OPEN;

	if (flags & (O_WRONLY|O_RDWR|O_CREAT|O_TRUNC))
		want_auth = USE_AUTH_MDS;

	req = ceph_mdsc_create_request(mdsc, op, want_auth);
	if (IS_ERR(req))
		goto out;
	req->r_fmode = ceph_flags_to_mode(flags);
	req->r_args.open.flags = cpu_to_le32(flags);
	req->r_args.open.mode = cpu_to_le32(create_mode);
out:
	return req;
}

/*
 * initialize private struct file data.
 * if we fail, clean up by dropping fmode reference on the ceph_inode
 */
static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
{
	struct ceph_file_info *cf;
	int ret = 0;

	switch (inode->i_mode & S_IFMT) {
	case S_IFREG:
	case S_IFDIR:
		dout("init_file %p %p 0%o (regular)\n", inode, file,
		     inode->i_mode);
		cf = kmem_cache_alloc(ceph_file_cachep, GFP_NOFS | __GFP_ZERO);
		if (cf == NULL) {
			ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */
			return -ENOMEM;
		}
		cf->fmode = fmode;
		cf->next_offset = 2;
		file->private_data = cf;
		BUG_ON(inode->i_fop->release != ceph_release);
		break;

	case S_IFLNK:
		dout("init_file %p %p 0%o (symlink)\n", inode, file,
		     inode->i_mode);
		ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */
		break;

	default:
		dout("init_file %p %p 0%o (special)\n", inode, file,
		     inode->i_mode);
		/*
		 * we need to drop the open ref now, since we don't
		 * have .release set to ceph_release.
		 */
		ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */
		BUG_ON(inode->i_fop->release == ceph_release);

		/* call the proper open fop */
		ret = inode->i_fop->open(inode, file);
	}
	return ret;
}

/*
 * If we already have the requisite capabilities, we can satisfy
 * the open request locally (no need to request new caps from the
 * MDS).  We do, however, need to inform the MDS (asynchronously)
 * if our wanted caps set expands.
 */
int ceph_open(struct inode *inode, struct file *file)
{
	struct ceph_inode_info *ci = ceph_inode(inode);
120 121
	struct ceph_fs_client *fsc = ceph_sb_to_client(inode->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
122 123
	struct ceph_mds_request *req;
	struct ceph_file_info *cf = file->private_data;
S
Sage Weil 已提交
124
	struct inode *parent_inode = NULL;
S
Sage Weil 已提交
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148
	int err;
	int flags, fmode, wanted;

	if (cf) {
		dout("open file %p is already opened\n", file);
		return 0;
	}

	/* filter out O_CREAT|O_EXCL; vfs did that already.  yuck. */
	flags = file->f_flags & ~(O_CREAT|O_EXCL);
	if (S_ISDIR(inode->i_mode))
		flags = O_DIRECTORY;  /* mds likes to know */

	dout("open inode %p ino %llx.%llx file %p flags %d (%d)\n", inode,
	     ceph_vinop(inode), file, flags, file->f_flags);
	fmode = ceph_flags_to_mode(flags);
	wanted = ceph_caps_for_mode(fmode);

	/* snapped files are read-only */
	if (ceph_snap(inode) != CEPH_NOSNAP && (file->f_mode & FMODE_WRITE))
		return -EROFS;

	/* trivially open snapdir */
	if (ceph_snap(inode) == CEPH_SNAPDIR) {
149
		spin_lock(&ci->i_ceph_lock);
S
Sage Weil 已提交
150
		__ceph_get_fmode(ci, fmode);
151
		spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
152 153 154 155
		return ceph_init_file(inode, file, fmode);
	}

	/*
156 157
	 * No need to block if we have caps on the auth MDS (for
	 * write) or any MDS (for read).  Update wanted set
S
Sage Weil 已提交
158 159
	 * asynchronously.
	 */
160
	spin_lock(&ci->i_ceph_lock);
161 162
	if (__ceph_is_any_real_caps(ci) &&
	    (((fmode & CEPH_FILE_MODE_WR) == 0) || ci->i_auth_cap)) {
S
Sage Weil 已提交
163 164 165 166 167 168 169
		int mds_wanted = __ceph_caps_mds_wanted(ci);
		int issued = __ceph_caps_issued(ci, NULL);

		dout("open %p fmode %d want %s issued %s using existing\n",
		     inode, fmode, ceph_cap_string(wanted),
		     ceph_cap_string(issued));
		__ceph_get_fmode(ci, fmode);
170
		spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
171 172 173 174 175 176 177 178 179 180 181

		/* adjust wanted? */
		if ((issued & wanted) != wanted &&
		    (mds_wanted & wanted) != wanted &&
		    ceph_snap(inode) != CEPH_SNAPDIR)
			ceph_check_caps(ci, 0, NULL);

		return ceph_init_file(inode, file, fmode);
	} else if (ceph_snap(inode) != CEPH_NOSNAP &&
		   (ci->i_snap_caps & wanted) == wanted) {
		__ceph_get_fmode(ci, fmode);
182
		spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
183 184
		return ceph_init_file(inode, file, fmode);
	}
185
	spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
186 187 188 189 190 191 192

	dout("open fmode %d wants %s\n", fmode, ceph_cap_string(wanted));
	req = prepare_open_request(inode->i_sb, flags, 0);
	if (IS_ERR(req)) {
		err = PTR_ERR(req);
		goto out;
	}
193 194
	req->r_inode = inode;
	ihold(inode);
S
Sage Weil 已提交
195
	req->r_num_caps = 1;
S
Sage Weil 已提交
196 197
	if (flags & (O_CREAT|O_TRUNC))
		parent_inode = ceph_get_dentry_parent_inode(file->f_dentry);
S
Sage Weil 已提交
198
	err = ceph_mdsc_do_request(mdsc, parent_inode, req);
S
Sage Weil 已提交
199
	iput(parent_inode);
S
Sage Weil 已提交
200 201 202 203 204 205 206 207 208 209
	if (!err)
		err = ceph_init_file(inode, file, req->r_fmode);
	ceph_mdsc_put_request(req);
	dout("open result=%d on %llx.%llx\n", err, ceph_vinop(inode));
out:
	return err;
}


/*
S
Sage Weil 已提交
210 211
 * Do a lookup + open with a single request.  If we get a non-existent
 * file or symlink, return 1 so the VFS can retry.
S
Sage Weil 已提交
212
 */
S
Sage Weil 已提交
213
int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
A
Al Viro 已提交
214
		     struct file *file, unsigned flags, umode_t mode,
A
Al Viro 已提交
215
		     int *opened)
S
Sage Weil 已提交
216
{
217 218
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
219
	struct ceph_mds_request *req;
S
Sage Weil 已提交
220
	struct dentry *dn;
S
Sage Weil 已提交
221 222
	int err;

S
Sage Weil 已提交
223 224 225 226 227 228 229 230 231 232
	dout("atomic_open %p dentry %p '%.*s' %s flags %d mode 0%o\n",
	     dir, dentry, dentry->d_name.len, dentry->d_name.name,
	     d_unhashed(dentry) ? "unhashed" : "hashed", flags, mode);

	if (dentry->d_name.len > NAME_MAX)
		return -ENAMETOOLONG;

	err = ceph_init_dentry(dentry);
	if (err < 0)
		return err;
S
Sage Weil 已提交
233 234 235 236

	/* do the open */
	req = prepare_open_request(dir->i_sb, flags, mode);
	if (IS_ERR(req))
A
Al Viro 已提交
237
		return PTR_ERR(req);
S
Sage Weil 已提交
238 239 240 241 242 243 244
	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
	if (flags & O_CREAT) {
		req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
		req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
	}
	req->r_locked_dir = dir;           /* caller holds dir->i_mutex */
245 246 247
	err = ceph_mdsc_do_request(mdsc,
				   (flags & (O_CREAT|O_TRUNC)) ? dir : NULL,
				   req);
248 249 250
	if (err)
		goto out_err;

251
	err = ceph_handle_snapdir(req, dentry, err);
S
Sage Weil 已提交
252
	if (err == 0 && (flags & O_CREAT) && !req->r_reply_info.head->is_dentry)
S
Sage Weil 已提交
253
		err = ceph_handle_notrace_create(dir, dentry);
254

S
Sage Weil 已提交
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
	if (d_unhashed(dentry)) {
		dn = ceph_finish_lookup(req, dentry, err);
		if (IS_ERR(dn))
			err = PTR_ERR(dn);
	} else {
		/* we were given a hashed negative dentry */
		dn = NULL;
	}
	if (err)
		goto out_err;
	if (dn || dentry->d_inode == NULL || S_ISLNK(dentry->d_inode->i_mode)) {
		/* make vfs retry on splice, ENOENT, or symlink */
		dout("atomic_open finish_no_open on dn %p\n", dn);
		err = finish_no_open(file, dn);
	} else {
		dout("atomic_open finish_open on dn %p\n", dn);
271 272 273
		if (req->r_op == CEPH_MDS_OP_CREATE && req->r_reply_info.has_create_ino) {
			*opened |= FILE_CREATED;
		}
S
Sage Weil 已提交
274 275
		err = finish_open(file, dentry, ceph_open, opened);
	}
276

S
Sage Weil 已提交
277 278 279
out_err:
	ceph_mdsc_put_request(req);
	dout("atomic_open result=%d\n", err);
A
Al Viro 已提交
280
	return err;
S
Sage Weil 已提交
281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
}

int ceph_release(struct inode *inode, struct file *file)
{
	struct ceph_inode_info *ci = ceph_inode(inode);
	struct ceph_file_info *cf = file->private_data;

	dout("release inode %p file %p\n", inode, file);
	ceph_put_fmode(ci, cf->fmode);
	if (cf->last_readdir)
		ceph_mdsc_put_request(cf->last_readdir);
	kfree(cf->last_name);
	kfree(cf->dir_info);
	dput(cf->dentry);
	kmem_cache_free(ceph_file_cachep, cf);
296 297

	/* wake up anyone waiting for caps on this inode */
298
	wake_up_all(&ci->i_cap_wq);
S
Sage Weil 已提交
299 300 301 302 303 304 305 306 307 308 309 310
	return 0;
}

/*
 * Read a range of bytes striped over one or more objects.  Iterate over
 * objects we stripe over.  (That's not atomic, but good enough for now.)
 *
 * If we get a short result from the OSD, check against i_size; we need to
 * only return a short read to the caller if we hit EOF.
 */
static int striped_read(struct inode *inode,
			u64 off, u64 len,
311
			struct page **pages, int num_pages,
312
			int *checkeof, bool o_direct,
313
			unsigned long buf_align)
S
Sage Weil 已提交
314
{
315
	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
S
Sage Weil 已提交
316
	struct ceph_inode_info *ci = ceph_inode(inode);
317
	u64 pos, this_len, left;
318
	int io_align, page_align;
319
	int pages_left;
S
Sage Weil 已提交
320 321 322 323 324 325 326 327 328 329 330 331 332
	int read;
	struct page **page_pos;
	int ret;
	bool hit_stripe, was_short;

	/*
	 * we may need to do multiple reads.  not atomic, unfortunately.
	 */
	pos = off;
	left = len;
	page_pos = pages;
	pages_left = num_pages;
	read = 0;
333
	io_align = off & ~PAGE_MASK;
S
Sage Weil 已提交
334 335

more:
336
	if (o_direct)
337
		page_align = (pos - io_align + buf_align) & ~PAGE_MASK;
338 339
	else
		page_align = pos & ~PAGE_MASK;
S
Sage Weil 已提交
340
	this_len = left;
341
	ret = ceph_osdc_readpages(&fsc->client->osdc, ceph_vino(inode),
S
Sage Weil 已提交
342 343 344
				  &ci->i_layout, pos, &this_len,
				  ci->i_truncate_seq,
				  ci->i_truncate_size,
345
				  page_pos, pages_left, page_align);
S
Sage Weil 已提交
346 347
	if (ret == -ENOENT)
		ret = 0;
348 349
	hit_stripe = this_len < left;
	was_short = ret >= 0 && ret < this_len;
350
	dout("striped_read %llu~%llu (read %u) got %d%s%s\n", pos, left, read,
S
Sage Weil 已提交
351 352
	     ret, hit_stripe ? " HITSTRIPE" : "", was_short ? " SHORT" : "");

353 354 355 356 357 358 359 360 361 362
	if (ret >= 0) {
		int didpages;
		if (was_short && (pos + ret < inode->i_size)) {
			u64 tmp = min(this_len - ret,
					inode->i_size - pos - ret);
			dout(" zero gap %llu to %llu\n",
				pos + ret, pos + ret + tmp);
			ceph_zero_page_vector_range(page_align + read + ret,
							tmp, pages);
			ret += tmp;
S
Sage Weil 已提交
363
		}
364 365

		didpages = (page_align + ret) >> PAGE_CACHE_SHIFT;
S
Sage Weil 已提交
366 367 368 369 370 371
		pos += ret;
		read = pos - off;
		left -= ret;
		page_pos += didpages;
		pages_left -= didpages;

372 373
		/* hit stripe and need continue*/
		if (left && hit_stripe && pos < inode->i_size)
S
Sage Weil 已提交
374 375 376
			goto more;
	}

377
	if (read > 0) {
378
		ret = read;
379 380 381
		/* did we bounce off eof? */
		if (pos + left > inode->i_size)
			*checkeof = 1;
S
Sage Weil 已提交
382 383 384 385 386 387 388 389 390 391 392 393 394
	}

	dout("striped_read returns %d\n", ret);
	return ret;
}

/*
 * Completely synchronous read and write methods.  Direct from __user
 * buffer to osd, or directly to user pages (if O_DIRECT).
 *
 * If the read spans object boundary, just do multiple reads.
 */
static ssize_t ceph_sync_read(struct file *file, char __user *data,
395
			      unsigned len, loff_t *poff, int *checkeof)
S
Sage Weil 已提交
396
{
A
Al Viro 已提交
397
	struct inode *inode = file_inode(file);
S
Sage Weil 已提交
398 399
	struct page **pages;
	u64 off = *poff;
400
	int num_pages, ret;
S
Sage Weil 已提交
401 402 403 404

	dout("sync_read on file %p %llu~%u %s\n", file, off, len,
	     (file->f_flags & O_DIRECT) ? "O_DIRECT" : "");

405 406
	if (file->f_flags & O_DIRECT) {
		num_pages = calc_pages_for((unsigned long)data, len);
407
		pages = ceph_get_direct_page_vector(data, num_pages, true);
408 409
	} else {
		num_pages = calc_pages_for(off, len);
410
		pages = ceph_alloc_page_vector(num_pages, GFP_NOFS);
411
	}
S
Sage Weil 已提交
412 413 414
	if (IS_ERR(pages))
		return PTR_ERR(pages);

415 416 417 418 419 420
	/*
	 * flush any page cache pages in this range.  this
	 * will make concurrent normal and sync io slow,
	 * but it will at least behave sensibly when they are
	 * in sequence.
	 */
421 422 423 424
	ret = filemap_write_and_wait(inode->i_mapping);
	if (ret < 0)
		goto done;

425
	ret = striped_read(inode, off, len, pages, num_pages, checkeof,
426 427
			   file->f_flags & O_DIRECT,
			   (unsigned long)data & ~PAGE_MASK);
S
Sage Weil 已提交
428 429

	if (ret >= 0 && (file->f_flags & O_DIRECT) == 0)
430
		ret = ceph_copy_page_vector_to_user(pages, data, off, ret);
S
Sage Weil 已提交
431 432 433
	if (ret >= 0)
		*poff = off + ret;

434
done:
S
Sage Weil 已提交
435
	if (file->f_flags & O_DIRECT)
436
		ceph_put_page_vector(pages, num_pages, true);
S
Sage Weil 已提交
437 438 439 440 441 442 443
	else
		ceph_release_page_vector(pages, num_pages);
	dout("sync_read result %d\n", ret);
	return ret;
}

/*
444 445 446 447 448 449 450 451 452 453
 * Write commit request unsafe callback, called to tell us when a
 * request is unsafe (that is, in flight--has been handed to the
 * messenger to send to its target osd).  It is called again when
 * we've received a response message indicating the request is
 * "safe" (its CEPH_OSD_FLAG_ONDISK flag is set), or when a request
 * is completed early (and unsuccessfully) due to a timeout or
 * interrupt.
 *
 * This is used if we requested both an ACK and ONDISK commit reply
 * from the OSD.
S
Sage Weil 已提交
454
 */
455
static void ceph_sync_write_unsafe(struct ceph_osd_request *req, bool unsafe)
S
Sage Weil 已提交
456 457 458
{
	struct ceph_inode_info *ci = ceph_inode(req->r_inode);

459 460 461 462 463 464 465 466 467 468 469 470 471 472
	dout("%s %p tid %llu %ssafe\n", __func__, req, req->r_tid,
		unsafe ? "un" : "");
	if (unsafe) {
		ceph_get_cap_refs(ci, CEPH_CAP_FILE_WR);
		spin_lock(&ci->i_unsafe_lock);
		list_add_tail(&req->r_unsafe_item,
			      &ci->i_unsafe_writes);
		spin_unlock(&ci->i_unsafe_lock);
	} else {
		spin_lock(&ci->i_unsafe_lock);
		list_del_init(&req->r_unsafe_item);
		spin_unlock(&ci->i_unsafe_lock);
		ceph_put_cap_refs(ci, CEPH_CAP_FILE_WR);
	}
S
Sage Weil 已提交
473 474 475 476 477 478 479 480 481 482 483
}

/*
 * Synchronous write, straight from __user pointer or user pages (if
 * O_DIRECT).
 *
 * If write spans object boundary, just do multiple writes.  (For a
 * correct atomic write, we should e.g. take write locks on all
 * objects, rollback on failure, etc.)
 */
static ssize_t ceph_sync_write(struct file *file, const char __user *data,
484
			       size_t left, loff_t pos, loff_t *ppos)
S
Sage Weil 已提交
485
{
A
Al Viro 已提交
486
	struct inode *inode = file_inode(file);
S
Sage Weil 已提交
487
	struct ceph_inode_info *ci = ceph_inode(inode);
488
	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
489 490
	struct ceph_snap_context *snapc;
	struct ceph_vino vino;
S
Sage Weil 已提交
491
	struct ceph_osd_request *req;
492
	int num_ops = 1;
S
Sage Weil 已提交
493 494 495 496 497 498
	struct page **pages;
	int num_pages;
	u64 len;
	int written = 0;
	int flags;
	int check_caps = 0;
499
	int page_align, io_align;
500
	unsigned long buf_align;
S
Sage Weil 已提交
501 502
	int ret;
	struct timespec mtime = CURRENT_TIME;
503
	bool own_pages = false;
S
Sage Weil 已提交
504

A
Al Viro 已提交
505
	if (ceph_snap(file_inode(file)) != CEPH_NOSNAP)
S
Sage Weil 已提交
506 507
		return -EROFS;

508
	dout("sync_write on file %p %lld~%u %s\n", file, pos,
S
Sage Weil 已提交
509 510
	     (unsigned)left, (file->f_flags & O_DIRECT) ? "O_DIRECT" : "");

511 512 513 514 515 516 517 518 519 520
	ret = filemap_write_and_wait_range(inode->i_mapping, pos, pos + left);
	if (ret < 0)
		return ret;

	ret = invalidate_inode_pages2_range(inode->i_mapping,
					    pos >> PAGE_CACHE_SHIFT,
					    (pos + left) >> PAGE_CACHE_SHIFT);
	if (ret < 0)
		dout("invalidate_inode_pages2_range returned %d\n", ret);

S
Sage Weil 已提交
521 522 523 524 525 526
	flags = CEPH_OSD_FLAG_ORDERSNAP |
		CEPH_OSD_FLAG_ONDISK |
		CEPH_OSD_FLAG_WRITE;
	if ((file->f_flags & (O_SYNC|O_DIRECT)) == 0)
		flags |= CEPH_OSD_FLAG_ACK;
	else
527
		num_ops++;	/* Also include a 'startsync' command. */
S
Sage Weil 已提交
528 529 530 531 532 533

	/*
	 * we may need to do multiple writes here if we span an object
	 * boundary.  this isn't atomic, unfortunately.  :(
	 */
more:
534 535
	io_align = pos & ~PAGE_MASK;
	buf_align = (unsigned long)data & ~PAGE_MASK;
S
Sage Weil 已提交
536
	len = left;
537

538 539
	snapc = ci->i_snap_realm->cached_context;
	vino = ceph_vino(inode);
540
	req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
541
				    vino, pos, &len, num_ops,
542
				    CEPH_OSD_OP_WRITE, flags, snapc,
S
Sage Weil 已提交
543
				    ci->i_truncate_seq, ci->i_truncate_size,
544
				    false);
545 546
	if (IS_ERR(req))
		return PTR_ERR(req);
S
Sage Weil 已提交
547

548 549 550
	/* write from beginning of first page, regardless of io alignment */
	page_align = file->f_flags & O_DIRECT ? buf_align : io_align;
	num_pages = calc_pages_for(page_align, len);
S
Sage Weil 已提交
551
	if (file->f_flags & O_DIRECT) {
552
		pages = ceph_get_direct_page_vector(data, num_pages, false);
S
Sage Weil 已提交
553 554 555 556 557 558 559 560 561
		if (IS_ERR(pages)) {
			ret = PTR_ERR(pages);
			goto out;
		}

		/*
		 * throw out any page cache pages in this range. this
		 * may block.
		 */
S
Sage Weil 已提交
562
		truncate_inode_pages_range(inode->i_mapping, pos,
S
Sage Weil 已提交
563
					   (pos+len) | (PAGE_CACHE_SIZE-1));
S
Sage Weil 已提交
564
	} else {
565
		pages = ceph_alloc_page_vector(num_pages, GFP_NOFS);
S
Sage Weil 已提交
566 567 568 569
		if (IS_ERR(pages)) {
			ret = PTR_ERR(pages);
			goto out;
		}
570
		ret = ceph_copy_user_to_page_vector(pages, data, pos, len);
S
Sage Weil 已提交
571 572 573 574 575 576 577
		if (ret < 0) {
			ceph_release_page_vector(pages, num_pages);
			goto out;
		}

		if ((file->f_flags & O_SYNC) == 0) {
			/* get a second commit callback */
578 579
			req->r_unsafe_callback = ceph_sync_write_unsafe;
			req->r_inode = inode;
580
			own_pages = true;
S
Sage Weil 已提交
581 582
		}
	}
583 584
	osd_req_op_extent_osd_data_pages(req, 0, pages, len, page_align,
					false, own_pages);
S
Sage Weil 已提交
585

586
	/* BUG_ON(vino.snap != CEPH_NOSNAP); */
587
	ceph_osdc_build_request(req, pos, snapc, vino.snap, &mtime);
588

589
	ret = ceph_osdc_start_request(&fsc->client->osdc, req, false);
590
	if (!ret)
591
		ret = ceph_osdc_wait_request(&fsc->client->osdc, req);
S
Sage Weil 已提交
592 593

	if (file->f_flags & O_DIRECT)
594
		ceph_put_page_vector(pages, num_pages, false);
S
Sage Weil 已提交
595 596 597 598 599 600 601 602 603
	else if (file->f_flags & O_SYNC)
		ceph_release_page_vector(pages, num_pages);

out:
	ceph_osdc_put_request(req);
	if (ret == 0) {
		pos += len;
		written += len;
		left -= len;
604
		data += len;
S
Sage Weil 已提交
605 606 607 608
		if (left)
			goto more;

		ret = written;
609
		*ppos = pos;
S
Sage Weil 已提交
610 611 612 613 614
		if (pos > i_size_read(inode))
			check_caps = ceph_inode_set_size(inode, pos);
		if (check_caps)
			ceph_check_caps(ceph_inode(inode), CHECK_CAPS_AUTHONLY,
					NULL);
615 616
	} else if (ret != -EOLDSNAPC && written > 0) {
		ret = written;
S
Sage Weil 已提交
617 618 619 620 621 622 623 624 625 626 627 628 629 630 631
	}
	return ret;
}

/*
 * Wrap generic_file_aio_read with checks for cap bits on the inode.
 * Atomically grab references, so that those bits are not released
 * back to the MDS mid-read.
 *
 * Hmm, the sync read case isn't actually async... should it be?
 */
static ssize_t ceph_aio_read(struct kiocb *iocb, const struct iovec *iov,
			     unsigned long nr_segs, loff_t pos)
{
	struct file *filp = iocb->ki_filp;
632
	struct ceph_file_info *fi = filp->private_data;
S
Sage Weil 已提交
633 634
	loff_t *ppos = &iocb->ki_pos;
	size_t len = iov->iov_len;
A
Al Viro 已提交
635
	struct inode *inode = file_inode(filp);
S
Sage Weil 已提交
636
	struct ceph_inode_info *ci = ceph_inode(inode);
Y
Yehuda Sadeh 已提交
637
	void __user *base = iov->iov_base;
S
Sage Weil 已提交
638
	ssize_t ret;
639
	int want, got = 0;
640
	int checkeof = 0, read = 0;
S
Sage Weil 已提交
641 642 643

	dout("aio_read %p %llx.%llx %llu~%u trying to get caps on %p\n",
	     inode, ceph_vinop(inode), pos, (unsigned)len, inode);
644
again:
645 646 647 648 649
	if (fi->fmode & CEPH_FILE_MODE_LAZY)
		want = CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO;
	else
		want = CEPH_CAP_FILE_CACHE;
	ret = ceph_get_caps(ci, CEPH_CAP_FILE_RD, want, &got, -1);
S
Sage Weil 已提交
650 651 652 653 654 655
	if (ret < 0)
		goto out;
	dout("aio_read %p %llx.%llx %llu~%u got cap refs on %s\n",
	     inode, ceph_vinop(inode), pos, (unsigned)len,
	     ceph_cap_string(got));

656
	if ((got & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) == 0 ||
S
Sage Weil 已提交
657
	    (iocb->ki_filp->f_flags & O_DIRECT) ||
658
	    (fi->flags & CEPH_F_SYNC))
S
Sage Weil 已提交
659
		/* hmm, this isn't really async... */
660
		ret = ceph_sync_read(filp, base, len, ppos, &checkeof);
S
Sage Weil 已提交
661 662 663 664 665 666 667
	else
		ret = generic_file_aio_read(iocb, iov, nr_segs, pos);

out:
	dout("aio_read %p %llx.%llx dropping cap refs on %s = %d\n",
	     inode, ceph_vinop(inode), ceph_cap_string(got), (int)ret);
	ceph_put_cap_refs(ci, got);
668 669 670 671 672 673

	if (checkeof && ret >= 0) {
		int statret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE);

		/* hit EOF or hole? */
		if (statret == 0 && *ppos < inode->i_size) {
674
			dout("aio_read sync_read hit hole, ppos %lld < size %lld, reading more\n", *ppos, inode->i_size);
675 676 677 678 679 680 681 682 683 684
			read += ret;
			base += ret;
			len -= ret;
			checkeof = 0;
			goto again;
		}
	}
	if (ret >= 0)
		ret += read;

S
Sage Weil 已提交
685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701
	return ret;
}

/*
 * Take cap references to avoid releasing caps to MDS mid-write.
 *
 * If we are synchronous, and write with an old snap context, the OSD
 * may return EOLDSNAPC.  In that case, retry the write.. _after_
 * dropping our cap refs and allowing the pending snap to logically
 * complete _before_ this write occurs.
 *
 * If we are near ENOSPC, write synchronously.
 */
static ssize_t ceph_aio_write(struct kiocb *iocb, const struct iovec *iov,
		       unsigned long nr_segs, loff_t pos)
{
	struct file *file = iocb->ki_filp;
702
	struct ceph_file_info *fi = file->private_data;
A
Al Viro 已提交
703
	struct inode *inode = file_inode(file);
S
Sage Weil 已提交
704
	struct ceph_inode_info *ci = ceph_inode(inode);
705 706
	struct ceph_osd_client *osdc =
		&ceph_sb_to_client(inode->i_sb)->client->osdc;
707 708
	ssize_t count, written = 0;
	int err, want, got;
S
Sage Weil 已提交
709 710 711 712

	if (ceph_snap(inode) != CEPH_NOSNAP)
		return -EROFS;

713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736
	mutex_lock(&inode->i_mutex);

	err = generic_segment_checks(iov, &nr_segs, &count, VERIFY_READ);
	if (err)
		goto out;

	/* We can write back this queue in page reclaim */
	current->backing_dev_info = file->f_mapping->backing_dev_info;

	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
	if (err)
		goto out;

	if (count == 0)
		goto out;

	err = file_remove_suid(file);
	if (err)
		goto out;

	err = file_update_time(file);
	if (err)
		goto out;

S
Sage Weil 已提交
737
retry_snap:
Y
Yan, Zheng 已提交
738
	if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL)) {
739
		err = -ENOSPC;
Y
Yan, Zheng 已提交
740 741
		goto out;
	}
742

743
	dout("aio_write %p %llx.%llx %llu~%zd getting caps. i_size %llu\n",
744
	     inode, ceph_vinop(inode), pos, count, inode->i_size);
S
Sage Weil 已提交
745 746 747 748
	if (fi->fmode & CEPH_FILE_MODE_LAZY)
		want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO;
	else
		want = CEPH_CAP_FILE_BUFFER;
749 750 751
	got = 0;
	err = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, &got, pos + count);
	if (err < 0)
752
		goto out;
S
Sage Weil 已提交
753

754
	dout("aio_write %p %llx.%llx %llu~%zd got cap refs on %s\n",
755
	     inode, ceph_vinop(inode), pos, count, ceph_cap_string(got));
S
Sage Weil 已提交
756 757 758 759

	if ((got & (CEPH_CAP_FILE_BUFFER|CEPH_CAP_FILE_LAZYIO)) == 0 ||
	    (iocb->ki_filp->f_flags & O_DIRECT) ||
	    (fi->flags & CEPH_F_SYNC)) {
760
		mutex_unlock(&inode->i_mutex);
761 762
		written = ceph_sync_write(file, iov->iov_base, count,
					  pos, &iocb->ki_pos);
763 764 765 766 767 768 769 770
		if (written == -EOLDSNAPC) {
			dout("aio_write %p %llx.%llx %llu~%u"
				"got EOLDSNAPC, retrying\n",
				inode, ceph_vinop(inode),
				pos, (unsigned)iov->iov_len);
			mutex_lock(&inode->i_mutex);
			goto retry_snap;
		}
S
Sage Weil 已提交
771
	} else {
Y
Yan, Zheng 已提交
772 773 774 775 776 777 778
		/*
		 * No need to acquire the i_truncate_mutex. Because
		 * the MDS revokes Fwb caps before sending truncate
		 * message to us. We can't get Fwb cap while there
		 * are pending vmtruncate. So write and vmtruncate
		 * can not run at the same time
		 */
779 780 781
		written = generic_file_buffered_write(iocb, iov, nr_segs,
						      pos, &iocb->ki_pos,
						      count, 0);
Y
Yan, Zheng 已提交
782
		mutex_unlock(&inode->i_mutex);
S
Sage Weil 已提交
783
	}
784

785
	if (written >= 0) {
786
		int dirty;
787
		spin_lock(&ci->i_ceph_lock);
788
		dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR);
789
		spin_unlock(&ci->i_ceph_lock);
790 791
		if (dirty)
			__mark_inode_dirty(inode, dirty);
S
Sage Weil 已提交
792
	}
S
Sage Weil 已提交
793

S
Sage Weil 已提交
794
	dout("aio_write %p %llx.%llx %llu~%u  dropping cap refs on %s\n",
S
Sage Weil 已提交
795 796
	     inode, ceph_vinop(inode), pos, (unsigned)iov->iov_len,
	     ceph_cap_string(got));
S
Sage Weil 已提交
797
	ceph_put_cap_refs(ci, got);
S
Sage Weil 已提交
798

799
	if (written >= 0 &&
Y
Yan, Zheng 已提交
800 801
	    ((file->f_flags & O_SYNC) || IS_SYNC(file->f_mapping->host) ||
	     ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_NEARFULL))) {
802
		err = vfs_fsync_range(file, pos, pos + written - 1, 1);
Y
Yan, Zheng 已提交
803
		if (err < 0)
804
			written = err;
Y
Yan, Zheng 已提交
805
	}
806

807 808
	goto out_unlocked;

809
out:
810 811
	mutex_unlock(&inode->i_mutex);
out_unlocked:
812 813
	current->backing_dev_info = NULL;
	return written ? written : err;
S
Sage Weil 已提交
814 815 816 817 818
}

/*
 * llseek.  be sure to verify file size on SEEK_END.
 */
819
static loff_t ceph_llseek(struct file *file, loff_t offset, int whence)
S
Sage Weil 已提交
820 821 822 823 824
{
	struct inode *inode = file->f_mapping->host;
	int ret;

	mutex_lock(&inode->i_mutex);
825

826
	if (whence == SEEK_END || whence == SEEK_DATA || whence == SEEK_HOLE) {
S
Sage Weil 已提交
827 828 829 830 831
		ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE);
		if (ret < 0) {
			offset = ret;
			goto out;
		}
832 833
	}

834
	switch (whence) {
835
	case SEEK_END:
S
Sage Weil 已提交
836 837 838 839 840 841 842 843 844 845 846 847 848 849 850
		offset += inode->i_size;
		break;
	case SEEK_CUR:
		/*
		 * Here we special-case the lseek(fd, 0, SEEK_CUR)
		 * position-querying operation.  Avoid rewriting the "same"
		 * f_pos value back to the file because a concurrent read(),
		 * write() or lseek() might have altered it
		 */
		if (offset == 0) {
			offset = file->f_pos;
			goto out;
		}
		offset += file->f_pos;
		break;
851 852 853 854 855 856 857 858 859 860 861 862 863
	case SEEK_DATA:
		if (offset >= inode->i_size) {
			ret = -ENXIO;
			goto out;
		}
		break;
	case SEEK_HOLE:
		if (offset >= inode->i_size) {
			ret = -ENXIO;
			goto out;
		}
		offset = inode->i_size;
		break;
S
Sage Weil 已提交
864 865
	}

J
Jie Liu 已提交
866
	offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
S
Sage Weil 已提交
867 868 869 870 871 872

out:
	mutex_unlock(&inode->i_mutex);
	return offset;
}

L
Li Wang 已提交
873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957
static inline void ceph_zero_partial_page(
	struct inode *inode, loff_t offset, unsigned size)
{
	struct page *page;
	pgoff_t index = offset >> PAGE_CACHE_SHIFT;

	page = find_lock_page(inode->i_mapping, index);
	if (page) {
		wait_on_page_writeback(page);
		zero_user(page, offset & (PAGE_CACHE_SIZE - 1), size);
		unlock_page(page);
		page_cache_release(page);
	}
}

static void ceph_zero_pagecache_range(struct inode *inode, loff_t offset,
				      loff_t length)
{
	loff_t nearly = round_up(offset, PAGE_CACHE_SIZE);
	if (offset < nearly) {
		loff_t size = nearly - offset;
		if (length < size)
			size = length;
		ceph_zero_partial_page(inode, offset, size);
		offset += size;
		length -= size;
	}
	if (length >= PAGE_CACHE_SIZE) {
		loff_t size = round_down(length, PAGE_CACHE_SIZE);
		truncate_pagecache_range(inode, offset, offset + size - 1);
		offset += size;
		length -= size;
	}
	if (length)
		ceph_zero_partial_page(inode, offset, length);
}

static int ceph_zero_partial_object(struct inode *inode,
				    loff_t offset, loff_t *length)
{
	struct ceph_inode_info *ci = ceph_inode(inode);
	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
	struct ceph_osd_request *req;
	int ret = 0;
	loff_t zero = 0;
	int op;

	if (!length) {
		op = offset ? CEPH_OSD_OP_DELETE : CEPH_OSD_OP_TRUNCATE;
		length = &zero;
	} else {
		op = CEPH_OSD_OP_ZERO;
	}

	req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
					ceph_vino(inode),
					offset, length,
					1, op,
					CEPH_OSD_FLAG_WRITE |
					CEPH_OSD_FLAG_ONDISK,
					NULL, 0, 0, false);
	if (IS_ERR(req)) {
		ret = PTR_ERR(req);
		goto out;
	}

	ceph_osdc_build_request(req, offset, NULL, ceph_vino(inode).snap,
				&inode->i_mtime);

	ret = ceph_osdc_start_request(&fsc->client->osdc, req, false);
	if (!ret) {
		ret = ceph_osdc_wait_request(&fsc->client->osdc, req);
		if (ret == -ENOENT)
			ret = 0;
	}
	ceph_osdc_put_request(req);

out:
	return ret;
}

static int ceph_zero_objects(struct inode *inode, loff_t offset, loff_t length)
{
	int ret = 0;
	struct ceph_inode_info *ci = ceph_inode(inode);
S
Sage Weil 已提交
958 959 960 961 962 963 964 965 966 967
	s32 stripe_unit = ceph_file_layout_su(ci->i_layout);
	s32 stripe_count = ceph_file_layout_stripe_count(ci->i_layout);
	s32 object_size = ceph_file_layout_object_size(ci->i_layout);
	u64 object_set_size = object_size * stripe_count;
	u64 nearly, t;

	/* round offset up to next period boundary */
	nearly = offset + object_set_size - 1;
	t = nearly;
	nearly -= do_div(t, object_set_size);
L
Li Wang 已提交
968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070

	while (length && offset < nearly) {
		loff_t size = length;
		ret = ceph_zero_partial_object(inode, offset, &size);
		if (ret < 0)
			return ret;
		offset += size;
		length -= size;
	}
	while (length >= object_set_size) {
		int i;
		loff_t pos = offset;
		for (i = 0; i < stripe_count; ++i) {
			ret = ceph_zero_partial_object(inode, pos, NULL);
			if (ret < 0)
				return ret;
			pos += stripe_unit;
		}
		offset += object_set_size;
		length -= object_set_size;
	}
	while (length) {
		loff_t size = length;
		ret = ceph_zero_partial_object(inode, offset, &size);
		if (ret < 0)
			return ret;
		offset += size;
		length -= size;
	}
	return ret;
}

static long ceph_fallocate(struct file *file, int mode,
				loff_t offset, loff_t length)
{
	struct ceph_file_info *fi = file->private_data;
	struct inode *inode = file->f_dentry->d_inode;
	struct ceph_inode_info *ci = ceph_inode(inode);
	struct ceph_osd_client *osdc =
		&ceph_inode_to_client(inode)->client->osdc;
	int want, got = 0;
	int dirty;
	int ret = 0;
	loff_t endoff = 0;
	loff_t size;

	if (!S_ISREG(inode->i_mode))
		return -EOPNOTSUPP;

	if (IS_SWAPFILE(inode))
		return -ETXTBSY;

	mutex_lock(&inode->i_mutex);

	if (ceph_snap(inode) != CEPH_NOSNAP) {
		ret = -EROFS;
		goto unlock;
	}

	if (ceph_osdmap_flag(osdc->osdmap, CEPH_OSDMAP_FULL) &&
		!(mode & FALLOC_FL_PUNCH_HOLE)) {
		ret = -ENOSPC;
		goto unlock;
	}

	size = i_size_read(inode);
	if (!(mode & FALLOC_FL_KEEP_SIZE))
		endoff = offset + length;

	if (fi->fmode & CEPH_FILE_MODE_LAZY)
		want = CEPH_CAP_FILE_BUFFER | CEPH_CAP_FILE_LAZYIO;
	else
		want = CEPH_CAP_FILE_BUFFER;

	ret = ceph_get_caps(ci, CEPH_CAP_FILE_WR, want, &got, endoff);
	if (ret < 0)
		goto unlock;

	if (mode & FALLOC_FL_PUNCH_HOLE) {
		if (offset < size)
			ceph_zero_pagecache_range(inode, offset, length);
		ret = ceph_zero_objects(inode, offset, length);
	} else if (endoff > size) {
		truncate_pagecache_range(inode, size, -1);
		if (ceph_inode_set_size(inode, endoff))
			ceph_check_caps(ceph_inode(inode),
				CHECK_CAPS_AUTHONLY, NULL);
	}

	if (!ret) {
		spin_lock(&ci->i_ceph_lock);
		dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR);
		spin_unlock(&ci->i_ceph_lock);
		if (dirty)
			__mark_inode_dirty(inode, dirty);
	}

	ceph_put_cap_refs(ci, got);
unlock:
	mutex_unlock(&inode->i_mutex);
	return ret;
}

S
Sage Weil 已提交
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
const struct file_operations ceph_file_fops = {
	.open = ceph_open,
	.release = ceph_release,
	.llseek = ceph_llseek,
	.read = do_sync_read,
	.write = do_sync_write,
	.aio_read = ceph_aio_read,
	.aio_write = ceph_aio_write,
	.mmap = ceph_mmap,
	.fsync = ceph_fsync,
G
Greg Farnum 已提交
1081 1082
	.lock = ceph_lock,
	.flock = ceph_flock,
S
Sage Weil 已提交
1083 1084 1085 1086
	.splice_read = generic_file_splice_read,
	.splice_write = generic_file_splice_write,
	.unlocked_ioctl = ceph_ioctl,
	.compat_ioctl	= ceph_ioctl,
L
Li Wang 已提交
1087
	.fallocate	= ceph_fallocate,
S
Sage Weil 已提交
1088 1089
};