1/* 2 * Device operations for the pnfs nfs4 file layout driver. 3 * 4 * Copyright (c) 2014, Primary Data, Inc. All rights reserved. 5 * 6 * Tao Peng <bergwolf@primarydata.com> 7 */ 8 9#include <linux/nfs_fs.h> 10#include <linux/vmalloc.h> 11#include <linux/module.h> 12#include <linux/sunrpc/addr.h> 13 14#include "../internal.h" 15#include "../nfs4session.h" 16#include "flexfilelayout.h" 17 18#define NFSDBG_FACILITY NFSDBG_PNFS_LD 19 20static unsigned int dataserver_timeo = NFS4_DEF_DS_TIMEO; 21static unsigned int dataserver_retrans = NFS4_DEF_DS_RETRANS; 22 23void nfs4_ff_layout_put_deviceid(struct nfs4_ff_layout_ds *mirror_ds) 24{ 25 if (mirror_ds) 26 nfs4_put_deviceid_node(&mirror_ds->id_node); 27} 28 29void nfs4_ff_layout_free_deviceid(struct nfs4_ff_layout_ds *mirror_ds) 30{ 31 nfs4_print_deviceid(&mirror_ds->id_node.deviceid); 32 nfs4_pnfs_ds_put(mirror_ds->ds); 33 kfree_rcu(mirror_ds, id_node.rcu); 34} 35 36/* Decode opaque device data and construct new_ds using it */ 37struct nfs4_ff_layout_ds * 38nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev, 39 gfp_t gfp_flags) 40{ 41 struct xdr_stream stream; 42 struct xdr_buf buf; 43 struct page *scratch; 44 struct list_head dsaddrs; 45 struct nfs4_pnfs_ds_addr *da; 46 struct nfs4_ff_layout_ds *new_ds = NULL; 47 struct nfs4_ff_ds_version *ds_versions = NULL; 48 u32 mp_count; 49 u32 version_count; 50 __be32 *p; 51 int i, ret = -ENOMEM; 52 53 /* set up xdr stream */ 54 scratch = alloc_page(gfp_flags); 55 if (!scratch) 56 goto out_err; 57 58 new_ds = kzalloc(sizeof(struct nfs4_ff_layout_ds), gfp_flags); 59 if (!new_ds) 60 goto out_scratch; 61 62 nfs4_init_deviceid_node(&new_ds->id_node, 63 server, 64 &pdev->dev_id); 65 INIT_LIST_HEAD(&dsaddrs); 66 67 xdr_init_decode_pages(&stream, &buf, pdev->pages, pdev->pglen); 68 xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE); 69 70 /* multipath count */ 71 p = xdr_inline_decode(&stream, 4); 72 if (unlikely(!p)) 73 goto out_err_drain_dsaddrs; 74 mp_count = be32_to_cpup(p); 75 dprintk("%s: multipath ds count %d\n", __func__, mp_count); 76 77 for (i = 0; i < mp_count; i++) { 78 /* multipath ds */ 79 da = nfs4_decode_mp_ds_addr(server->nfs_client->cl_net, 80 &stream, gfp_flags); 81 if (da) 82 list_add_tail(&da->da_node, &dsaddrs); 83 } 84 if (list_empty(&dsaddrs)) { 85 dprintk("%s: no suitable DS addresses found\n", 86 __func__); 87 ret = -ENOMEDIUM; 88 goto out_err_drain_dsaddrs; 89 } 90 91 /* version count */ 92 p = xdr_inline_decode(&stream, 4); 93 if (unlikely(!p)) 94 goto out_err_drain_dsaddrs; 95 version_count = be32_to_cpup(p); 96 dprintk("%s: version count %d\n", __func__, version_count); 97 98 ds_versions = kzalloc(version_count * sizeof(struct nfs4_ff_ds_version), 99 gfp_flags); 100 if (!ds_versions) 101 goto out_scratch; 102 103 for (i = 0; i < version_count; i++) { 104 /* 20 = version(4) + minor_version(4) + rsize(4) + wsize(4) + 105 * tightly_coupled(4) */ 106 p = xdr_inline_decode(&stream, 20); 107 if (unlikely(!p)) 108 goto out_err_drain_dsaddrs; 109 ds_versions[i].version = be32_to_cpup(p++); 110 ds_versions[i].minor_version = be32_to_cpup(p++); 111 ds_versions[i].rsize = nfs_block_size(be32_to_cpup(p++), NULL); 112 ds_versions[i].wsize = nfs_block_size(be32_to_cpup(p++), NULL); 113 ds_versions[i].tightly_coupled = be32_to_cpup(p); 114 115 if (ds_versions[i].rsize > NFS_MAX_FILE_IO_SIZE) 116 ds_versions[i].rsize = NFS_MAX_FILE_IO_SIZE; 117 if (ds_versions[i].wsize > NFS_MAX_FILE_IO_SIZE) 118 ds_versions[i].wsize = NFS_MAX_FILE_IO_SIZE; 119 120 if (ds_versions[i].version != 3 || ds_versions[i].minor_version != 0) { 121 dprintk("%s: [%d] unsupported ds version %d-%d\n", __func__, 122 i, ds_versions[i].version, 123 ds_versions[i].minor_version); 124 ret = -EPROTONOSUPPORT; 125 goto out_err_drain_dsaddrs; 126 } 127 128 dprintk("%s: [%d] vers %u minor_ver %u rsize %u wsize %u coupled %d\n", 129 __func__, i, ds_versions[i].version, 130 ds_versions[i].minor_version, 131 ds_versions[i].rsize, 132 ds_versions[i].wsize, 133 ds_versions[i].tightly_coupled); 134 } 135 136 new_ds->ds_versions = ds_versions; 137 new_ds->ds_versions_cnt = version_count; 138 139 new_ds->ds = nfs4_pnfs_ds_add(&dsaddrs, gfp_flags); 140 if (!new_ds->ds) 141 goto out_err_drain_dsaddrs; 142 143 /* If DS was already in cache, free ds addrs */ 144 while (!list_empty(&dsaddrs)) { 145 da = list_first_entry(&dsaddrs, 146 struct nfs4_pnfs_ds_addr, 147 da_node); 148 list_del_init(&da->da_node); 149 kfree(da->da_remotestr); 150 kfree(da); 151 } 152 153 __free_page(scratch); 154 return new_ds; 155 156out_err_drain_dsaddrs: 157 while (!list_empty(&dsaddrs)) { 158 da = list_first_entry(&dsaddrs, struct nfs4_pnfs_ds_addr, 159 da_node); 160 list_del_init(&da->da_node); 161 kfree(da->da_remotestr); 162 kfree(da); 163 } 164 165 kfree(ds_versions); 166out_scratch: 167 __free_page(scratch); 168out_err: 169 kfree(new_ds); 170 171 dprintk("%s ERROR: returning %d\n", __func__, ret); 172 return NULL; 173} 174 175static void ff_layout_mark_devid_invalid(struct pnfs_layout_segment *lseg, 176 struct nfs4_deviceid_node *devid) 177{ 178 nfs4_mark_deviceid_unavailable(devid); 179 if (!ff_layout_has_available_ds(lseg)) 180 pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode, 181 lseg); 182} 183 184static bool ff_layout_mirror_valid(struct pnfs_layout_segment *lseg, 185 struct nfs4_ff_layout_mirror *mirror) 186{ 187 if (mirror == NULL || mirror->mirror_ds == NULL) { 188 pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode, 189 lseg); 190 return false; 191 } 192 if (mirror->mirror_ds->ds == NULL) { 193 struct nfs4_deviceid_node *devid; 194 devid = &mirror->mirror_ds->id_node; 195 ff_layout_mark_devid_invalid(lseg, devid); 196 return false; 197 } 198 return true; 199} 200 201static u64 202end_offset(u64 start, u64 len) 203{ 204 u64 end; 205 206 end = start + len; 207 return end >= start ? end : NFS4_MAX_UINT64; 208} 209 210static void extend_ds_error(struct nfs4_ff_layout_ds_err *err, 211 u64 offset, u64 length) 212{ 213 u64 end; 214 215 end = max_t(u64, end_offset(err->offset, err->length), 216 end_offset(offset, length)); 217 err->offset = min_t(u64, err->offset, offset); 218 err->length = end - err->offset; 219} 220 221static bool ds_error_can_merge(struct nfs4_ff_layout_ds_err *err, u64 offset, 222 u64 length, int status, enum nfs_opnum4 opnum, 223 nfs4_stateid *stateid, 224 struct nfs4_deviceid *deviceid) 225{ 226 return err->status == status && err->opnum == opnum && 227 nfs4_stateid_match(&err->stateid, stateid) && 228 !memcmp(&err->deviceid, deviceid, sizeof(*deviceid)) && 229 end_offset(err->offset, err->length) >= offset && 230 err->offset <= end_offset(offset, length); 231} 232 233static bool merge_ds_error(struct nfs4_ff_layout_ds_err *old, 234 struct nfs4_ff_layout_ds_err *new) 235{ 236 if (!ds_error_can_merge(old, new->offset, new->length, new->status, 237 new->opnum, &new->stateid, &new->deviceid)) 238 return false; 239 240 extend_ds_error(old, new->offset, new->length); 241 return true; 242} 243 244static bool 245ff_layout_add_ds_error_locked(struct nfs4_flexfile_layout *flo, 246 struct nfs4_ff_layout_ds_err *dserr) 247{ 248 struct nfs4_ff_layout_ds_err *err; 249 250 list_for_each_entry(err, &flo->error_list, list) { 251 if (merge_ds_error(err, dserr)) { 252 return true; 253 } 254 } 255 256 list_add(&dserr->list, &flo->error_list); 257 return false; 258} 259 260static bool 261ff_layout_update_ds_error(struct nfs4_flexfile_layout *flo, u64 offset, 262 u64 length, int status, enum nfs_opnum4 opnum, 263 nfs4_stateid *stateid, struct nfs4_deviceid *deviceid) 264{ 265 bool found = false; 266 struct nfs4_ff_layout_ds_err *err; 267 268 list_for_each_entry(err, &flo->error_list, list) { 269 if (ds_error_can_merge(err, offset, length, status, opnum, 270 stateid, deviceid)) { 271 found = true; 272 extend_ds_error(err, offset, length); 273 break; 274 } 275 } 276 277 return found; 278} 279 280int ff_layout_track_ds_error(struct nfs4_flexfile_layout *flo, 281 struct nfs4_ff_layout_mirror *mirror, u64 offset, 282 u64 length, int status, enum nfs_opnum4 opnum, 283 gfp_t gfp_flags) 284{ 285 struct nfs4_ff_layout_ds_err *dserr; 286 bool needfree; 287 288 if (status == 0) 289 return 0; 290 291 if (mirror->mirror_ds == NULL) 292 return -EINVAL; 293 294 spin_lock(&flo->generic_hdr.plh_inode->i_lock); 295 if (ff_layout_update_ds_error(flo, offset, length, status, opnum, 296 &mirror->stateid, 297 &mirror->mirror_ds->id_node.deviceid)) { 298 spin_unlock(&flo->generic_hdr.plh_inode->i_lock); 299 return 0; 300 } 301 spin_unlock(&flo->generic_hdr.plh_inode->i_lock); 302 dserr = kmalloc(sizeof(*dserr), gfp_flags); 303 if (!dserr) 304 return -ENOMEM; 305 306 INIT_LIST_HEAD(&dserr->list); 307 dserr->offset = offset; 308 dserr->length = length; 309 dserr->status = status; 310 dserr->opnum = opnum; 311 nfs4_stateid_copy(&dserr->stateid, &mirror->stateid); 312 memcpy(&dserr->deviceid, &mirror->mirror_ds->id_node.deviceid, 313 NFS4_DEVICEID4_SIZE); 314 315 spin_lock(&flo->generic_hdr.plh_inode->i_lock); 316 needfree = ff_layout_add_ds_error_locked(flo, dserr); 317 spin_unlock(&flo->generic_hdr.plh_inode->i_lock); 318 if (needfree) 319 kfree(dserr); 320 321 return 0; 322} 323 324/* currently we only support AUTH_NONE and AUTH_SYS */ 325static rpc_authflavor_t 326nfs4_ff_layout_choose_authflavor(struct nfs4_ff_layout_mirror *mirror) 327{ 328 if (mirror->uid == (u32)-1) 329 return RPC_AUTH_NULL; 330 return RPC_AUTH_UNIX; 331} 332 333/* fetch cred for NFSv3 DS */ 334static int ff_layout_update_mirror_cred(struct nfs4_ff_layout_mirror *mirror, 335 struct nfs4_pnfs_ds *ds) 336{ 337 if (ds->ds_clp && !mirror->cred && 338 mirror->mirror_ds->ds_versions[0].version == 3) { 339 struct rpc_auth *auth = ds->ds_clp->cl_rpcclient->cl_auth; 340 struct rpc_cred *cred; 341 struct auth_cred acred = { 342 .uid = make_kuid(&init_user_ns, mirror->uid), 343 .gid = make_kgid(&init_user_ns, mirror->gid), 344 }; 345 346 /* AUTH_NULL ignores acred */ 347 cred = auth->au_ops->lookup_cred(auth, &acred, 0); 348 if (IS_ERR(cred)) { 349 dprintk("%s: lookup_cred failed with %ld\n", 350 __func__, PTR_ERR(cred)); 351 return PTR_ERR(cred); 352 } else { 353 if (cmpxchg(&mirror->cred, NULL, cred)) 354 put_rpccred(cred); 355 } 356 } 357 return 0; 358} 359 360struct nfs_fh * 361nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx) 362{ 363 struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, mirror_idx); 364 struct nfs_fh *fh = NULL; 365 366 if (!ff_layout_mirror_valid(lseg, mirror)) { 367 pr_err_ratelimited("NFS: %s: No data server for mirror offset index %d\n", 368 __func__, mirror_idx); 369 goto out; 370 } 371 372 /* FIXME: For now assume there is only 1 version available for the DS */ 373 fh = &mirror->fh_versions[0]; 374out: 375 return fh; 376} 377 378/* Upon return, either ds is connected, or ds is NULL */ 379struct nfs4_pnfs_ds * 380nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, 381 bool fail_return) 382{ 383 struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx); 384 struct nfs4_pnfs_ds *ds = NULL; 385 struct nfs4_deviceid_node *devid; 386 struct inode *ino = lseg->pls_layout->plh_inode; 387 struct nfs_server *s = NFS_SERVER(ino); 388 unsigned int max_payload; 389 rpc_authflavor_t flavor; 390 391 if (!ff_layout_mirror_valid(lseg, mirror)) { 392 pr_err_ratelimited("NFS: %s: No data server for offset index %d\n", 393 __func__, ds_idx); 394 goto out; 395 } 396 397 devid = &mirror->mirror_ds->id_node; 398 if (ff_layout_test_devid_unavailable(devid)) 399 goto out; 400 401 ds = mirror->mirror_ds->ds; 402 /* matching smp_wmb() in _nfs4_pnfs_v3/4_ds_connect */ 403 smp_rmb(); 404 if (ds->ds_clp) 405 goto out_update_creds; 406 407 flavor = nfs4_ff_layout_choose_authflavor(mirror); 408 409 /* FIXME: For now we assume the server sent only one version of NFS 410 * to use for the DS. 411 */ 412 nfs4_pnfs_ds_connect(s, ds, devid, dataserver_timeo, 413 dataserver_retrans, 414 mirror->mirror_ds->ds_versions[0].version, 415 mirror->mirror_ds->ds_versions[0].minor_version, 416 flavor); 417 418 /* connect success, check rsize/wsize limit */ 419 if (ds->ds_clp) { 420 max_payload = 421 nfs_block_size(rpc_max_payload(ds->ds_clp->cl_rpcclient), 422 NULL); 423 if (mirror->mirror_ds->ds_versions[0].rsize > max_payload) 424 mirror->mirror_ds->ds_versions[0].rsize = max_payload; 425 if (mirror->mirror_ds->ds_versions[0].wsize > max_payload) 426 mirror->mirror_ds->ds_versions[0].wsize = max_payload; 427 } else { 428 ff_layout_track_ds_error(FF_LAYOUT_FROM_HDR(lseg->pls_layout), 429 mirror, lseg->pls_range.offset, 430 lseg->pls_range.length, NFS4ERR_NXIO, 431 OP_ILLEGAL, GFP_NOIO); 432 if (fail_return) { 433 pnfs_error_mark_layout_for_return(ino, lseg); 434 if (ff_layout_has_available_ds(lseg)) 435 pnfs_set_retry_layoutget(lseg->pls_layout); 436 else 437 pnfs_clear_retry_layoutget(lseg->pls_layout); 438 439 } else { 440 if (ff_layout_has_available_ds(lseg)) 441 set_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE, 442 &lseg->pls_layout->plh_flags); 443 else { 444 pnfs_error_mark_layout_for_return(ino, lseg); 445 pnfs_clear_retry_layoutget(lseg->pls_layout); 446 } 447 } 448 } 449out_update_creds: 450 if (ff_layout_update_mirror_cred(mirror, ds)) 451 ds = NULL; 452out: 453 return ds; 454} 455 456struct rpc_cred * 457ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx, 458 struct rpc_cred *mdscred) 459{ 460 struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx); 461 struct rpc_cred *cred = ERR_PTR(-EINVAL); 462 463 if (!nfs4_ff_layout_prepare_ds(lseg, ds_idx, true)) 464 goto out; 465 466 if (mirror && mirror->cred) 467 cred = mirror->cred; 468 else 469 cred = mdscred; 470out: 471 return cred; 472} 473 474/** 475* Find or create a DS rpc client with th MDS server rpc client auth flavor 476* in the nfs_client cl_ds_clients list. 477*/ 478struct rpc_clnt * 479nfs4_ff_find_or_create_ds_client(struct pnfs_layout_segment *lseg, u32 ds_idx, 480 struct nfs_client *ds_clp, struct inode *inode) 481{ 482 struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx); 483 484 switch (mirror->mirror_ds->ds_versions[0].version) { 485 case 3: 486 /* For NFSv3 DS, flavor is set when creating DS connections */ 487 return ds_clp->cl_rpcclient; 488 case 4: 489 return nfs4_find_or_create_ds_client(ds_clp, inode); 490 default: 491 BUG(); 492 } 493} 494 495static bool is_range_intersecting(u64 offset1, u64 length1, 496 u64 offset2, u64 length2) 497{ 498 u64 end1 = end_offset(offset1, length1); 499 u64 end2 = end_offset(offset2, length2); 500 501 return (end1 == NFS4_MAX_UINT64 || end1 > offset2) && 502 (end2 == NFS4_MAX_UINT64 || end2 > offset1); 503} 504 505/* called with inode i_lock held */ 506int ff_layout_encode_ds_ioerr(struct nfs4_flexfile_layout *flo, 507 struct xdr_stream *xdr, int *count, 508 const struct pnfs_layout_range *range) 509{ 510 struct nfs4_ff_layout_ds_err *err, *n; 511 __be32 *p; 512 513 list_for_each_entry_safe(err, n, &flo->error_list, list) { 514 if (!is_range_intersecting(err->offset, err->length, 515 range->offset, range->length)) 516 continue; 517 /* offset(8) + length(8) + stateid(NFS4_STATEID_SIZE) 518 * + array length + deviceid(NFS4_DEVICEID4_SIZE) 519 * + status(4) + opnum(4) 520 */ 521 p = xdr_reserve_space(xdr, 522 28 + NFS4_STATEID_SIZE + NFS4_DEVICEID4_SIZE); 523 if (unlikely(!p)) 524 return -ENOBUFS; 525 p = xdr_encode_hyper(p, err->offset); 526 p = xdr_encode_hyper(p, err->length); 527 p = xdr_encode_opaque_fixed(p, &err->stateid, 528 NFS4_STATEID_SIZE); 529 /* Encode 1 error */ 530 *p++ = cpu_to_be32(1); 531 p = xdr_encode_opaque_fixed(p, &err->deviceid, 532 NFS4_DEVICEID4_SIZE); 533 *p++ = cpu_to_be32(err->status); 534 *p++ = cpu_to_be32(err->opnum); 535 *count += 1; 536 list_del(&err->list); 537 dprintk("%s: offset %llu length %llu status %d op %d count %d\n", 538 __func__, err->offset, err->length, err->status, 539 err->opnum, *count); 540 kfree(err); 541 } 542 543 return 0; 544} 545 546static bool ff_read_layout_has_available_ds(struct pnfs_layout_segment *lseg) 547{ 548 struct nfs4_ff_layout_mirror *mirror; 549 struct nfs4_deviceid_node *devid; 550 u32 idx; 551 552 for (idx = 0; idx < FF_LAYOUT_MIRROR_COUNT(lseg); idx++) { 553 mirror = FF_LAYOUT_COMP(lseg, idx); 554 if (mirror && mirror->mirror_ds) { 555 devid = &mirror->mirror_ds->id_node; 556 if (!ff_layout_test_devid_unavailable(devid)) 557 return true; 558 } 559 } 560 561 return false; 562} 563 564static bool ff_rw_layout_has_available_ds(struct pnfs_layout_segment *lseg) 565{ 566 struct nfs4_ff_layout_mirror *mirror; 567 struct nfs4_deviceid_node *devid; 568 u32 idx; 569 570 for (idx = 0; idx < FF_LAYOUT_MIRROR_COUNT(lseg); idx++) { 571 mirror = FF_LAYOUT_COMP(lseg, idx); 572 if (!mirror || !mirror->mirror_ds) 573 return false; 574 devid = &mirror->mirror_ds->id_node; 575 if (ff_layout_test_devid_unavailable(devid)) 576 return false; 577 } 578 579 return FF_LAYOUT_MIRROR_COUNT(lseg) != 0; 580} 581 582bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg) 583{ 584 if (lseg->pls_range.iomode == IOMODE_READ) 585 return ff_read_layout_has_available_ds(lseg); 586 /* Note: RW layout needs all mirrors available */ 587 return ff_rw_layout_has_available_ds(lseg); 588} 589 590module_param(dataserver_retrans, uint, 0644); 591MODULE_PARM_DESC(dataserver_retrans, "The number of times the NFSv4.1 client " 592 "retries a request before it attempts further " 593 " recovery action."); 594module_param(dataserver_timeo, uint, 0644); 595MODULE_PARM_DESC(dataserver_timeo, "The time (in tenths of a second) the " 596 "NFSv4.1 client waits for a response from a " 597 " data server before it retries an NFS request."); 598