lustre/ptlrpc/sec.c

   1 /*
   2  * GPL HEADER START
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License version 2 only,
   8  * as published by the Free Software Foundation.
   9  *
  10  * This program is distributed in the hope that it will be useful, but
  11  * WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * General Public License version 2 for more details (a copy is included
  14  * in the LICENSE file that accompanied this code).
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * version 2 along with this program; If not, see
  18  * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
  19  *
  20  * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
  21  * CA 95054 USA or visit www.sun.com if you need additional information or
  22  * have any questions.
  23  *
  24  * GPL HEADER END
  25  */
  26 /*
  27  * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
  28  * Use is subject to license terms.
  29  *
  30  * Copyright (c) 2011, 2014, Intel Corporation.
  31  */
  32 /*
  33  * This file is part of Lustre, http://www.lustre.org/
  34  * Lustre is a trademark of Sun Microsystems, Inc.
  35  *
  36  * lustre/ptlrpc/sec.c
  37  *
  38  * Author: Eric Mei <ericm@clusterfs.com>
  39  */
  40
  41 #define DEBUG_SUBSYSTEM S_SEC
  42
  43 #include <linux/user_namespace.h>
  44 #ifdef HAVE_UIDGID_HEADER
  45 # include <linux/uidgid.h>
  46 #endif
  47 #include <linux/crypto.h>
  48 #include <linux/key.h>
  49
  50 #include <libcfs/libcfs.h>
  51 #include <obd.h>
  52 #include <obd_class.h>
  53 #include <obd_support.h>
  54 #include <lustre_net.h>
  55 #include <lustre_import.h>
  56 #include <lustre_dlm.h>
  57 #include <lustre_sec.h>
  58
  59 #include "ptlrpc_internal.h"
  60
  61 /***********************************************
  62  * policy registers                            *
  63  ***********************************************/
  64
  65 static rwlock_t policy_lock;
  66 static struct ptlrpc_sec_policy *policies[SPTLRPC_POLICY_MAX] = {
  67         NULL,
  68 };
  69
  70 int sptlrpc_register_policy(struct ptlrpc_sec_policy *policy)
  71 {
  72         __u16 number = policy->sp_policy;
  73
  74         LASSERT(policy->sp_name);
  75         LASSERT(policy->sp_cops);
  76         LASSERT(policy->sp_sops);
  77
  78         if (number >= SPTLRPC_POLICY_MAX)
  79                 return -EINVAL;
  80
  81         write_lock(&policy_lock);
  82         if (unlikely(policies[number])) {
  83                 write_unlock(&policy_lock);
  84                 return -EALREADY;
  85         }
  86         policies[number] = policy;
  87         write_unlock(&policy_lock);
  88
  89         CDEBUG(D_SEC, "%s: registered\n", policy->sp_name);
  90         return 0;
  91 }
  92 EXPORT_SYMBOL(sptlrpc_register_policy);
  93
  94 int sptlrpc_unregister_policy(struct ptlrpc_sec_policy *policy)
  95 {
  96         __u16 number = policy->sp_policy;
  97
  98         LASSERT(number < SPTLRPC_POLICY_MAX);
  99
 100         write_lock(&policy_lock);
 101         if (unlikely(policies[number] == NULL)) {
 102                 write_unlock(&policy_lock);
 103                 CERROR("%s: already unregistered\n", policy->sp_name);
 104                 return -EINVAL;
 105         }
 106
 107         LASSERT(policies[number] == policy);
 108         policies[number] = NULL;
 109         write_unlock(&policy_lock);
 110
 111         CDEBUG(D_SEC, "%s: unregistered\n", policy->sp_name);
 112         return 0;
 113 }
 114 EXPORT_SYMBOL(sptlrpc_unregister_policy);
 115
 116 static
 117 struct ptlrpc_sec_policy * sptlrpc_wireflavor2policy(__u32 flavor)
 118 {
 119         static DEFINE_MUTEX(load_mutex);
 120         static atomic_t           loaded = ATOMIC_INIT(0);
 121         struct ptlrpc_sec_policy *policy;
 122         __u16                     number = SPTLRPC_FLVR_POLICY(flavor);
 123         __u16                     flag = 0;
 124
 125         if (number >= SPTLRPC_POLICY_MAX)
 126                 return NULL;
 127
 128         while (1) {
 129                 read_lock(&policy_lock);
 130                 policy = policies[number];
 131                 if (policy && !try_module_get(policy->sp_owner))
 132                         policy = NULL;
 133                 if (policy == NULL)
 134                         flag = atomic_read(&loaded);
 135                 read_unlock(&policy_lock);
 136
 137                 if (policy != NULL || flag != 0 ||
 138                     number != SPTLRPC_POLICY_GSS)
 139                         break;
 140
 141                 /* try to load gss module, once */
 142                 mutex_lock(&load_mutex);
 143                 if (atomic_read(&loaded) == 0) {
 144                         if (request_module("ptlrpc_gss") == 0)
 145                                 CDEBUG(D_SEC,
 146                                        "module ptlrpc_gss loaded on demand\n");
 147                         else
 148                                 CERROR("Unable to load module ptlrpc_gss\n");
 149
 150                         atomic_set(&loaded, 1);
 151                 }
 152                 mutex_unlock(&load_mutex);
 153         }
 154
 155         return policy;
 156 }
 157
 158 __u32 sptlrpc_name2flavor_base(const char *name)
 159 {
 160         if (!strcmp(name, "null"))
 161                 return SPTLRPC_FLVR_NULL;
 162         if (!strcmp(name, "plain"))
 163                 return SPTLRPC_FLVR_PLAIN;
 164         if (!strcmp(name, "gssnull"))
 165                 return SPTLRPC_FLVR_GSSNULL;
 166         if (!strcmp(name, "krb5n"))
 167                 return SPTLRPC_FLVR_KRB5N;
 168         if (!strcmp(name, "krb5a"))
 169                 return SPTLRPC_FLVR_KRB5A;
 170         if (!strcmp(name, "krb5i"))
 171                 return SPTLRPC_FLVR_KRB5I;
 172         if (!strcmp(name, "krb5p"))
 173                 return SPTLRPC_FLVR_KRB5P;
 174         if (!strcmp(name, "ski"))
 175                 return SPTLRPC_FLVR_SKI;
 176         if (!strcmp(name, "skpi"))
 177                 return SPTLRPC_FLVR_SKPI;
 178
 179         return SPTLRPC_FLVR_INVALID;
 180 }
 181 EXPORT_SYMBOL(sptlrpc_name2flavor_base);
 182
 183 const char *sptlrpc_flavor2name_base(__u32 flvr)
 184 {
 185         __u32   base = SPTLRPC_FLVR_BASE(flvr);
 186
 187         if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_NULL))
 188                 return "null";
 189         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_PLAIN))
 190                 return "plain";
 191         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_GSSNULL))
 192                 return "gssnull";
 193         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_KRB5N))
 194                 return "krb5n";
 195         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_KRB5A))
 196                 return "krb5a";
 197         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_KRB5I))
 198                 return "krb5i";
 199         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_KRB5P))
 200                 return "krb5p";
 201         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_SKI))
 202                 return "ski";
 203         else if (base == SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_SKPI))
 204                 return "skpi";
 205
 206         CERROR("invalid wire flavor 0x%x\n", flvr);
 207         return "invalid";
 208 }
 209 EXPORT_SYMBOL(sptlrpc_flavor2name_base);
 210
 211 char *sptlrpc_flavor2name_bulk(struct sptlrpc_flavor *sf,
 212                                char *buf, int bufsize)
 213 {
 214         if (SPTLRPC_FLVR_POLICY(sf->sf_rpc) == SPTLRPC_POLICY_PLAIN)
 215                 snprintf(buf, bufsize, "hash:%s",
 216                          sptlrpc_get_hash_name(sf->u_bulk.hash.hash_alg));
 217         else
 218                 snprintf(buf, bufsize, "%s",
 219                          sptlrpc_flavor2name_base(sf->sf_rpc));
 220
 221         buf[bufsize - 1] = '\0';
 222         return buf;
 223 }
 224 EXPORT_SYMBOL(sptlrpc_flavor2name_bulk);
 225
 226 char *sptlrpc_flavor2name(struct sptlrpc_flavor *sf, char *buf, int bufsize)
 227 {
 228         snprintf(buf, bufsize, "%s", sptlrpc_flavor2name_base(sf->sf_rpc));
 229
 230         /*
 231          * currently we don't support customized bulk specification for
 232          * flavors other than plain
 233          */
 234         if (SPTLRPC_FLVR_POLICY(sf->sf_rpc) == SPTLRPC_POLICY_PLAIN) {
 235                 char bspec[16];
 236
 237                 bspec[0] = '-';
 238                 sptlrpc_flavor2name_bulk(sf, &bspec[1], sizeof(bspec) - 1);
 239                 strncat(buf, bspec, bufsize);
 240         }
 241
 242         buf[bufsize - 1] = '\0';
 243         return buf;
 244 }
 245 EXPORT_SYMBOL(sptlrpc_flavor2name);
 246
 247 char *sptlrpc_secflags2str(__u32 flags, char *buf, int bufsize)
 248 {
 249         buf[0] = '\0';
 250
 251         if (flags & PTLRPC_SEC_FL_REVERSE)
 252                 strlcat(buf, "reverse,", bufsize);
 253         if (flags & PTLRPC_SEC_FL_ROOTONLY)
 254                 strlcat(buf, "rootonly,", bufsize);
 255         if (flags & PTLRPC_SEC_FL_UDESC)
 256                 strlcat(buf, "udesc,", bufsize);
 257         if (flags & PTLRPC_SEC_FL_BULK)
 258                 strlcat(buf, "bulk,", bufsize);
 259         if (buf[0] == '\0')
 260                 strlcat(buf, "-,", bufsize);
 261
 262         return buf;
 263 }
 264 EXPORT_SYMBOL(sptlrpc_secflags2str);
 265
 266 /**************************************************
 267  * client context APIs                            *
 268  **************************************************/
 269
 270 static
 271 struct ptlrpc_cli_ctx *get_my_ctx(struct ptlrpc_sec *sec)
 272 {
 273         struct vfs_cred vcred;
 274         int create = 1, remove_dead = 1;
 275
 276         LASSERT(sec);
 277         LASSERT(sec->ps_policy->sp_cops->lookup_ctx);
 278
 279         if (sec->ps_flvr.sf_flags & (PTLRPC_SEC_FL_REVERSE |
 280                                      PTLRPC_SEC_FL_ROOTONLY)) {
 281                 vcred.vc_uid = 0;
 282                 vcred.vc_gid = 0;
 283                 if (sec->ps_flvr.sf_flags & PTLRPC_SEC_FL_REVERSE) {
 284                         create = 0;
 285                         remove_dead = 0;
 286                 }
 287         } else {
 288                 vcred.vc_uid = from_kuid(&init_user_ns, current_uid());
 289                 vcred.vc_gid = from_kgid(&init_user_ns, current_gid());
 290         }
 291
 292         return sec->ps_policy->sp_cops->lookup_ctx(sec, &vcred, create,
 293                                                    remove_dead);
 294 }
 295
 296 struct ptlrpc_cli_ctx *sptlrpc_cli_ctx_get(struct ptlrpc_cli_ctx *ctx)
 297 {
 298         atomic_inc(&ctx->cc_refcount);
 299         return ctx;
 300 }
 301 EXPORT_SYMBOL(sptlrpc_cli_ctx_get);
 302
 303 void sptlrpc_cli_ctx_put(struct ptlrpc_cli_ctx *ctx, int sync)
 304 {
 305         struct ptlrpc_sec *sec = ctx->cc_sec;
 306
 307         LASSERT(sec);
 308         LASSERT_ATOMIC_POS(&ctx->cc_refcount);
 309
 310         if (!atomic_dec_and_test(&ctx->cc_refcount))
 311                 return;
 312
 313         sec->ps_policy->sp_cops->release_ctx(sec, ctx, sync);
 314 }
 315 EXPORT_SYMBOL(sptlrpc_cli_ctx_put);
 316
 317 /**
 318  * Expire the client context immediately.
 319  *
 320  * \pre Caller must hold at least 1 reference on the \a ctx.
 321  */
 322 void sptlrpc_cli_ctx_expire(struct ptlrpc_cli_ctx *ctx)
 323 {
 324         LASSERT(ctx->cc_ops->die);
 325         ctx->cc_ops->die(ctx, 0);
 326 }
 327 EXPORT_SYMBOL(sptlrpc_cli_ctx_expire);
 328
 329 /**
 330  * To wake up the threads who are waiting for this client context. Called
 331  * after some status change happened on \a ctx.
 332  */
 333 void sptlrpc_cli_ctx_wakeup(struct ptlrpc_cli_ctx *ctx)
 334 {
 335         struct ptlrpc_request *req, *next;
 336
 337         spin_lock(&ctx->cc_lock);
 338         list_for_each_entry_safe(req, next, &ctx->cc_req_list,
 339                                      rq_ctx_chain) {
 340                 list_del_init(&req->rq_ctx_chain);
 341                 ptlrpc_client_wake_req(req);
 342         }
 343         spin_unlock(&ctx->cc_lock);
 344 }
 345 EXPORT_SYMBOL(sptlrpc_cli_ctx_wakeup);
 346
 347 int sptlrpc_cli_ctx_display(struct ptlrpc_cli_ctx *ctx, char *buf, int bufsize)
 348 {
 349         LASSERT(ctx->cc_ops);
 350
 351         if (ctx->cc_ops->display == NULL)
 352                 return 0;
 353
 354         return ctx->cc_ops->display(ctx, buf, bufsize);
 355 }
 356
 357 static int import_sec_check_expire(struct obd_import *imp)
 358 {
 359         int     adapt = 0;
 360
 361         spin_lock(&imp->imp_lock);
 362         if (imp->imp_sec_expire &&
 363             imp->imp_sec_expire < cfs_time_current_sec()) {
 364                 adapt = 1;
 365                 imp->imp_sec_expire = 0;
 366         }
 367         spin_unlock(&imp->imp_lock);
 368
 369         if (!adapt)
 370                 return 0;
 371
 372         CDEBUG(D_SEC, "found delayed sec adapt expired, do it now\n");
 373         return sptlrpc_import_sec_adapt(imp, NULL, NULL);
 374 }
 375
 376 /**
 377  * Get and validate the client side ptlrpc security facilities from
 378  * \a imp. There is a race condition on client reconnect when the import is
 379  * being destroyed while there are outstanding client bound requests. In
 380  * this case do not output any error messages if import secuity is not
 381  * found.
 382  *
 383  * \param[in] imp obd import associated with client
 384  * \param[out] sec client side ptlrpc security
 385  *
 386  * \retval 0 if security retrieved successfully
 387  * \retval -ve errno if there was a problem
 388  */
 389 static int import_sec_validate_get(struct obd_import *imp,
 390                                    struct ptlrpc_sec **sec)
 391 {
 392         int     rc;
 393
 394         if (unlikely(imp->imp_sec_expire)) {
 395                 rc = import_sec_check_expire(imp);
 396                 if (rc)
 397                         return rc;
 398         }
 399
 400         *sec = sptlrpc_import_sec_ref(imp);
 401         /* Only output an error when the import is still active */
 402         if (*sec == NULL) {
 403                 if (list_empty(&imp->imp_zombie_chain))
 404                         CERROR("import %p (%s) with no sec\n",
 405                                 imp, ptlrpc_import_state_name(imp->imp_state));
 406                 return -EACCES;
 407         }
 408
 409         if (unlikely((*sec)->ps_dying)) {
 410                 CERROR("attempt to use dying sec %p\n", sec);
 411                 sptlrpc_sec_put(*sec);
 412                 return -EACCES;
 413         }
 414
 415         return 0;
 416 }
 417
 418 /**
 419  * Given a \a req, find or allocate an appropriate context for it.
 420  * \pre req->rq_cli_ctx == NULL.
 421  *
 422  * \retval 0 succeed, and req->rq_cli_ctx is set.
 423  * \retval -ev error number, and req->rq_cli_ctx == NULL.
 424  */
 425 int sptlrpc_req_get_ctx(struct ptlrpc_request *req)
 426 {
 427         struct obd_import *imp = req->rq_import;
 428         struct ptlrpc_sec *sec;
 429         int                rc;
 430         ENTRY;
 431
 432         LASSERT(!req->rq_cli_ctx);
 433         LASSERT(imp);
 434
 435         rc = import_sec_validate_get(imp, &sec);
 436         if (rc)
 437                 RETURN(rc);
 438
 439         req->rq_cli_ctx = get_my_ctx(sec);
 440
 441         sptlrpc_sec_put(sec);
 442
 443         if (!req->rq_cli_ctx) {
 444                 CERROR("req %p: fail to get context\n", req);
 445                 RETURN(-ECONNREFUSED);
 446         }
 447
 448         RETURN(0);
 449 }
 450
 451 /**
 452  * Drop the context for \a req.
 453  * \pre req->rq_cli_ctx != NULL.
 454  * \post req->rq_cli_ctx == NULL.
 455  *
 456  * If \a sync == 0, this function should return quickly without sleep;
 457  * otherwise it might trigger and wait for the whole process of sending
 458  * an context-destroying rpc to server.
 459  */
 460 void sptlrpc_req_put_ctx(struct ptlrpc_request *req, int sync)
 461 {
 462         ENTRY;
 463
 464         LASSERT(req);
 465         LASSERT(req->rq_cli_ctx);
 466
 467         /* request might be asked to release earlier while still
 468          * in the context waiting list.
 469          */
 470         if (!list_empty(&req->rq_ctx_chain)) {
 471                 spin_lock(&req->rq_cli_ctx->cc_lock);
 472                 list_del_init(&req->rq_ctx_chain);
 473                 spin_unlock(&req->rq_cli_ctx->cc_lock);
 474         }
 475
 476         sptlrpc_cli_ctx_put(req->rq_cli_ctx, sync);
 477         req->rq_cli_ctx = NULL;
 478         EXIT;
 479 }
 480
 481 static
 482 int sptlrpc_req_ctx_switch(struct ptlrpc_request *req,
 483                            struct ptlrpc_cli_ctx *oldctx,
 484                            struct ptlrpc_cli_ctx *newctx)
 485 {
 486         struct sptlrpc_flavor   old_flvr;
 487         char                   *reqmsg = NULL; /* to workaround old gcc */
 488         int                     reqmsg_size;
 489         int                     rc = 0;
 490
 491         LASSERT(req->rq_reqmsg);
 492         LASSERT(req->rq_reqlen);
 493         LASSERT(req->rq_replen);
 494
 495         CDEBUG(D_SEC, "req %p: switch ctx %p(%u->%s) -> %p(%u->%s), "
 496                "switch sec %p(%s) -> %p(%s)\n", req,
 497                oldctx, oldctx->cc_vcred.vc_uid, sec2target_str(oldctx->cc_sec),
 498                newctx, newctx->cc_vcred.vc_uid, sec2target_str(newctx->cc_sec),
 499                oldctx->cc_sec, oldctx->cc_sec->ps_policy->sp_name,
 500                newctx->cc_sec, newctx->cc_sec->ps_policy->sp_name);
 501
 502         /* save flavor */
 503         old_flvr = req->rq_flvr;
 504
 505         /* save request message */
 506         reqmsg_size = req->rq_reqlen;
 507         if (reqmsg_size != 0) {
 508                 OBD_ALLOC_LARGE(reqmsg, reqmsg_size);
 509                 if (reqmsg == NULL)
 510                         return -ENOMEM;
 511                 memcpy(reqmsg, req->rq_reqmsg, reqmsg_size);
 512         }
 513
 514         /* release old req/rep buf */
 515         req->rq_cli_ctx = oldctx;
 516         sptlrpc_cli_free_reqbuf(req);
 517         sptlrpc_cli_free_repbuf(req);
 518         req->rq_cli_ctx = newctx;
 519
 520         /* recalculate the flavor */
 521         sptlrpc_req_set_flavor(req, 0);
 522
 523         /* alloc new request buffer
 524          * we don't need to alloc reply buffer here, leave it to the
 525          * rest procedure of ptlrpc */
 526         if (reqmsg_size != 0) {
 527                 rc = sptlrpc_cli_alloc_reqbuf(req, reqmsg_size);
 528                 if (!rc) {
 529                         LASSERT(req->rq_reqmsg);
 530                         memcpy(req->rq_reqmsg, reqmsg, reqmsg_size);
 531                 } else {
 532                         CWARN("failed to alloc reqbuf: %d\n", rc);
 533                         req->rq_flvr = old_flvr;
 534                 }
 535
 536                 OBD_FREE_LARGE(reqmsg, reqmsg_size);
 537         }
 538         return rc;
 539 }
 540
 541 /**
 542  * If current context of \a req is dead somehow, e.g. we just switched flavor
 543  * thus marked original contexts dead, we'll find a new context for it. if
 544  * no switch is needed, \a req will end up with the same context.
 545  *
 546  * \note a request must have a context, to keep other parts of code happy.
 547  * In any case of failure during the switching, we must restore the old one.
 548  */
 549 int sptlrpc_req_replace_dead_ctx(struct ptlrpc_request *req)
 550 {
 551         struct ptlrpc_cli_ctx *oldctx = req->rq_cli_ctx;
 552         struct ptlrpc_cli_ctx *newctx;
 553         int                    rc;
 554         ENTRY;
 555
 556         LASSERT(oldctx);
 557
 558         sptlrpc_cli_ctx_get(oldctx);
 559         sptlrpc_req_put_ctx(req, 0);
 560
 561         rc = sptlrpc_req_get_ctx(req);
 562         if (unlikely(rc)) {
 563                 LASSERT(!req->rq_cli_ctx);
 564
 565                 /* restore old ctx */
 566                 req->rq_cli_ctx = oldctx;
 567                 RETURN(rc);
 568         }
 569
 570         newctx = req->rq_cli_ctx;
 571         LASSERT(newctx);
 572
 573         if (unlikely(newctx == oldctx &&
 574                      test_bit(PTLRPC_CTX_DEAD_BIT, &oldctx->cc_flags))) {
 575                 /*
 576                  * still get the old dead ctx, usually means system too busy
 577                  */
 578                 CDEBUG(D_SEC,
 579                        "ctx (%p, fl %lx) doesn't switch, relax a little bit\n",
 580                        newctx, newctx->cc_flags);
 581
 582                 set_current_state(TASK_INTERRUPTIBLE);
 583                 schedule_timeout(msecs_to_jiffies(MSEC_PER_SEC));
 584         } else if (unlikely(test_bit(PTLRPC_CTX_UPTODATE_BIT, &newctx->cc_flags)
 585                             == 0)) {
 586                 /*
 587                  * new ctx not up to date yet
 588                  */
 589                 CDEBUG(D_SEC,
 590                        "ctx (%p, fl %lx) doesn't switch, not up to date yet\n",
 591                        newctx, newctx->cc_flags);
 592         } else {
 593                 /*
 594                  * it's possible newctx == oldctx if we're switching
 595                  * subflavor with the same sec.
 596                  */
 597                 rc = sptlrpc_req_ctx_switch(req, oldctx, newctx);
 598                 if (rc) {
 599                         /* restore old ctx */
 600                         sptlrpc_req_put_ctx(req, 0);
 601                         req->rq_cli_ctx = oldctx;
 602                         RETURN(rc);
 603                 }
 604
 605                 LASSERT(req->rq_cli_ctx == newctx);
 606         }
 607
 608         sptlrpc_cli_ctx_put(oldctx, 1);
 609         RETURN(0);
 610 }
 611 EXPORT_SYMBOL(sptlrpc_req_replace_dead_ctx);
 612
 613 static
 614 int ctx_check_refresh(struct ptlrpc_cli_ctx *ctx)
 615 {
 616         if (cli_ctx_is_refreshed(ctx))
 617                 return 1;
 618         return 0;
 619 }
 620
 621 static
 622 int ctx_refresh_timeout(void *data)
 623 {
 624         struct ptlrpc_request *req = data;
 625         int rc;
 626
 627         /* conn_cnt is needed in expire_one_request */
 628         lustre_msg_set_conn_cnt(req->rq_reqmsg, req->rq_import->imp_conn_cnt);
 629
 630         rc = ptlrpc_expire_one_request(req, 1);
 631         /* if we started recovery, we should mark this ctx dead; otherwise
 632          * in case of lgssd died nobody would retire this ctx, following
 633          * connecting will still find the same ctx thus cause deadlock.
 634          * there's an assumption that expire time of the request should be
 635          * later than the context refresh expire time.
 636          */
 637         if (rc == 0)
 638                 req->rq_cli_ctx->cc_ops->die(req->rq_cli_ctx, 0);
 639         return rc;
 640 }
 641
 642 static
 643 void ctx_refresh_interrupt(void *data)
 644 {
 645         struct ptlrpc_request *req = data;
 646
 647         spin_lock(&req->rq_lock);
 648         req->rq_intr = 1;
 649         spin_unlock(&req->rq_lock);
 650 }
 651
 652 static
 653 void req_off_ctx_list(struct ptlrpc_request *req, struct ptlrpc_cli_ctx *ctx)
 654 {
 655         spin_lock(&ctx->cc_lock);
 656         if (!list_empty(&req->rq_ctx_chain))
 657                 list_del_init(&req->rq_ctx_chain);
 658         spin_unlock(&ctx->cc_lock);
 659 }
 660
 661 /**
 662  * To refresh the context of \req, if it's not up-to-date.
 663  * \param timeout
 664  * - < 0: don't wait
 665  * - = 0: wait until success or fatal error occur
 666  * - > 0: timeout value (in seconds)
 667  *
 668  * The status of the context could be subject to be changed by other threads
 669  * at any time. We allow this race, but once we return with 0, the caller will
 670  * suppose it's uptodated and keep using it until the owning rpc is done.
 671  *
 672  * \retval 0 only if the context is uptodated.
 673  * \retval -ev error number.
 674  */
 675 int sptlrpc_req_refresh_ctx(struct ptlrpc_request *req, long timeout)
 676 {
 677         struct ptlrpc_cli_ctx  *ctx = req->rq_cli_ctx;
 678         struct ptlrpc_sec      *sec;
 679         struct l_wait_info      lwi;
 680         int                     rc;
 681         ENTRY;
 682
 683         LASSERT(ctx);
 684
 685         if (req->rq_ctx_init || req->rq_ctx_fini)
 686                 RETURN(0);
 687
 688         /*
 689          * during the process a request's context might change type even
 690          * (e.g. from gss ctx to null ctx), so each loop we need to re-check
 691          * everything
 692          */
 693 again:
 694         rc = import_sec_validate_get(req->rq_import, &sec);
 695         if (rc)
 696                 RETURN(rc);
 697
 698         if (sec->ps_flvr.sf_rpc != req->rq_flvr.sf_rpc) {
 699                 CDEBUG(D_SEC, "req %p: flavor has changed %x -> %x\n",
 700                       req, req->rq_flvr.sf_rpc, sec->ps_flvr.sf_rpc);
 701                 req_off_ctx_list(req, ctx);
 702                 sptlrpc_req_replace_dead_ctx(req);
 703                 ctx = req->rq_cli_ctx;
 704         }
 705         sptlrpc_sec_put(sec);
 706
 707         if (cli_ctx_is_eternal(ctx))
 708                 RETURN(0);
 709
 710         if (unlikely(test_bit(PTLRPC_CTX_NEW_BIT, &ctx->cc_flags))) {
 711                 LASSERT(ctx->cc_ops->refresh);
 712                 ctx->cc_ops->refresh(ctx);
 713         }
 714         LASSERT(test_bit(PTLRPC_CTX_NEW_BIT, &ctx->cc_flags) == 0);
 715
 716         LASSERT(ctx->cc_ops->validate);
 717         if (ctx->cc_ops->validate(ctx) == 0) {
 718                 req_off_ctx_list(req, ctx);
 719                 RETURN(0);
 720         }
 721
 722         if (unlikely(test_bit(PTLRPC_CTX_ERROR_BIT, &ctx->cc_flags))) {
 723                 spin_lock(&req->rq_lock);
 724                 req->rq_err = 1;
 725                 spin_unlock(&req->rq_lock);
 726                 req_off_ctx_list(req, ctx);
 727                 RETURN(-EPERM);
 728         }
 729
 730         /*
 731          * There's a subtle issue for resending RPCs, suppose following
 732          * situation:
 733          *  1. the request was sent to server.
 734          *  2. recovery was kicked start, after finished the request was
 735          *     marked as resent.
 736          *  3. resend the request.
 737          *  4. old reply from server received, we accept and verify the reply.
 738          *     this has to be success, otherwise the error will be aware
 739          *     by application.
 740          *  5. new reply from server received, dropped by LNet.
 741          *
 742          * Note the xid of old & new request is the same. We can't simply
 743          * change xid for the resent request because the server replies on
 744          * it for reply reconstruction.
 745          *
 746          * Commonly the original context should be uptodate because we
 747          * have an expiry nice time; server will keep its context because
 748          * we at least hold a ref of old context which prevent context
 749          * from destroying RPC being sent. So server still can accept the
 750          * request and finish the RPC. But if that's not the case:
 751          *  1. If server side context has been trimmed, a NO_CONTEXT will
 752          *     be returned, gss_cli_ctx_verify/unseal will switch to new
 753          *     context by force.
 754          *  2. Current context never be refreshed, then we are fine: we
 755          *     never really send request with old context before.
 756          */
 757         if (test_bit(PTLRPC_CTX_UPTODATE_BIT, &ctx->cc_flags) &&
 758             unlikely(req->rq_reqmsg) &&
 759             lustre_msg_get_flags(req->rq_reqmsg) & MSG_RESENT) {
 760                 req_off_ctx_list(req, ctx);
 761                 RETURN(0);
 762         }
 763
 764         if (unlikely(test_bit(PTLRPC_CTX_DEAD_BIT, &ctx->cc_flags))) {
 765                 req_off_ctx_list(req, ctx);
 766                 /*
 767                  * don't switch ctx if import was deactivated
 768                  */
 769                 if (req->rq_import->imp_deactive) {
 770                         spin_lock(&req->rq_lock);
 771                         req->rq_err = 1;
 772                         spin_unlock(&req->rq_lock);
 773                         RETURN(-EINTR);
 774                 }
 775
 776                 rc = sptlrpc_req_replace_dead_ctx(req);
 777                 if (rc) {
 778                         LASSERT(ctx == req->rq_cli_ctx);
 779                         CERROR("req %p: failed to replace dead ctx %p: %d\n",
 780                                req, ctx, rc);
 781                         spin_lock(&req->rq_lock);
 782                         req->rq_err = 1;
 783                         spin_unlock(&req->rq_lock);
 784                         RETURN(rc);
 785                 }
 786
 787                 ctx = req->rq_cli_ctx;
 788                 goto again;
 789         }
 790
 791         /*
 792          * Now we're sure this context is during upcall, add myself into
 793          * waiting list
 794          */
 795         spin_lock(&ctx->cc_lock);
 796         if (list_empty(&req->rq_ctx_chain))
 797                 list_add(&req->rq_ctx_chain, &ctx->cc_req_list);
 798         spin_unlock(&ctx->cc_lock);
 799
 800         if (timeout < 0)
 801                 RETURN(-EWOULDBLOCK);
 802
 803         /* Clear any flags that may be present from previous sends */
 804         LASSERT(req->rq_receiving_reply == 0);
 805         spin_lock(&req->rq_lock);
 806         req->rq_err = 0;
 807         req->rq_timedout = 0;
 808         req->rq_resend = 0;
 809         req->rq_restart = 0;
 810         spin_unlock(&req->rq_lock);
 811
 812         lwi = LWI_TIMEOUT_INTR(msecs_to_jiffies(timeout * MSEC_PER_SEC),
 813                                ctx_refresh_timeout,
 814                                ctx_refresh_interrupt, req);
 815         rc = l_wait_event(req->rq_reply_waitq, ctx_check_refresh(ctx), &lwi);
 816
 817         /*
 818          * following cases could lead us here:
 819          * - successfully refreshed;
 820          * - interrupted;
 821          * - timedout, and we don't want recover from the failure;
 822          * - timedout, and waked up upon recovery finished;
 823          * - someone else mark this ctx dead by force;
 824          * - someone invalidate the req and call ptlrpc_client_wake_req(),
 825          *   e.g. ptlrpc_abort_inflight();
 826          */
 827         if (!cli_ctx_is_refreshed(ctx)) {
 828                 /* timed out or interruptted */
 829                 req_off_ctx_list(req, ctx);
 830
 831                 LASSERT(rc != 0);
 832                 RETURN(rc);
 833         }
 834
 835         goto again;
 836 }
 837
 838 /**
 839  * Initialize flavor settings for \a req, according to \a opcode.
 840  *
 841  * \note this could be called in two situations:
 842  * - new request from ptlrpc_pre_req(), with proper @opcode
 843  * - old request which changed ctx in the middle, with @opcode == 0
 844  */
 845 void sptlrpc_req_set_flavor(struct ptlrpc_request *req, int opcode)
 846 {
 847         struct ptlrpc_sec *sec;
 848
 849         LASSERT(req->rq_import);
 850         LASSERT(req->rq_cli_ctx);
 851         LASSERT(req->rq_cli_ctx->cc_sec);
 852         LASSERT(req->rq_bulk_read == 0 || req->rq_bulk_write == 0);
 853
 854         /* special security flags accoding to opcode */
 855         switch (opcode) {
 856         case OST_READ:
 857         case MDS_READPAGE:
 858         case MGS_CONFIG_READ:
 859         case OBD_IDX_READ:
 860                 req->rq_bulk_read = 1;
 861                 break;
 862         case OST_WRITE:
 863         case MDS_WRITEPAGE:
 864                 req->rq_bulk_write = 1;
 865                 break;
 866         case SEC_CTX_INIT:
 867                 req->rq_ctx_init = 1;
 868                 break;
 869         case SEC_CTX_FINI:
 870                 req->rq_ctx_fini = 1;
 871                 break;
 872         case 0:
 873                 /* init/fini rpc won't be resend, so can't be here */
 874                 LASSERT(req->rq_ctx_init == 0);
 875                 LASSERT(req->rq_ctx_fini == 0);
 876
 877                 /* cleanup flags, which should be recalculated */
 878                 req->rq_pack_udesc = 0;
 879                 req->rq_pack_bulk = 0;
 880                 break;
 881         }
 882
 883         sec = req->rq_cli_ctx->cc_sec;
 884
 885         spin_lock(&sec->ps_lock);
 886         req->rq_flvr = sec->ps_flvr;
 887         spin_unlock(&sec->ps_lock);
 888
 889         /* force SVC_NULL for context initiation rpc, SVC_INTG for context
 890          * destruction rpc */
 891         if (unlikely(req->rq_ctx_init))
 892                 flvr_set_svc(&req->rq_flvr.sf_rpc, SPTLRPC_SVC_NULL);
 893         else if (unlikely(req->rq_ctx_fini))
 894                 flvr_set_svc(&req->rq_flvr.sf_rpc, SPTLRPC_SVC_INTG);
 895
 896         /* user descriptor flag, null security can't do it anyway */
 897         if ((sec->ps_flvr.sf_flags & PTLRPC_SEC_FL_UDESC) &&
 898             (req->rq_flvr.sf_rpc != SPTLRPC_FLVR_NULL))
 899                 req->rq_pack_udesc = 1;
 900
 901         /* bulk security flag */
 902         if ((req->rq_bulk_read || req->rq_bulk_write) &&
 903             sptlrpc_flavor_has_bulk(&req->rq_flvr))
 904                 req->rq_pack_bulk = 1;
 905 }
 906
 907 void sptlrpc_request_out_callback(struct ptlrpc_request *req)
 908 {
 909         if (SPTLRPC_FLVR_SVC(req->rq_flvr.sf_rpc) != SPTLRPC_SVC_PRIV)
 910                 return;
 911
 912         LASSERT(req->rq_clrbuf);
 913         if (req->rq_pool || !req->rq_reqbuf)
 914                 return;
 915
 916         OBD_FREE(req->rq_reqbuf, req->rq_reqbuf_len);
 917         req->rq_reqbuf = NULL;
 918         req->rq_reqbuf_len = 0;
 919 }
 920
 921 /**
 922  * Given an import \a imp, check whether current user has a valid context
 923  * or not. We may create a new context and try to refresh it, and try
 924  * repeatedly try in case of non-fatal errors. Return 0 means success.
 925  */
 926 int sptlrpc_import_check_ctx(struct obd_import *imp)
 927 {
 928         struct ptlrpc_sec     *sec;
 929         struct ptlrpc_cli_ctx *ctx;
 930         struct ptlrpc_request *req = NULL;
 931         int rc;
 932         ENTRY;
 933
 934         might_sleep();
 935
 936         sec = sptlrpc_import_sec_ref(imp);
 937         ctx = get_my_ctx(sec);
 938         sptlrpc_sec_put(sec);
 939
 940         if (!ctx)
 941                 RETURN(-ENOMEM);
 942
 943         if (cli_ctx_is_eternal(ctx) ||
 944             ctx->cc_ops->validate(ctx) == 0) {
 945                 sptlrpc_cli_ctx_put(ctx, 1);
 946                 RETURN(0);
 947         }
 948
 949         if (cli_ctx_is_error(ctx)) {
 950                 sptlrpc_cli_ctx_put(ctx, 1);
 951                 RETURN(-EACCES);
 952         }
 953
 954         req = ptlrpc_request_cache_alloc(GFP_NOFS);
 955         if (!req)
 956                 RETURN(-ENOMEM);
 957
 958         ptlrpc_cli_req_init(req);
 959         atomic_set(&req->rq_refcount, 10000);
 960
 961         req->rq_import = imp;
 962         req->rq_flvr = sec->ps_flvr;
 963         req->rq_cli_ctx = ctx;
 964
 965         rc = sptlrpc_req_refresh_ctx(req, 0);
 966         LASSERT(list_empty(&req->rq_ctx_chain));
 967         sptlrpc_cli_ctx_put(req->rq_cli_ctx, 1);
 968         ptlrpc_request_cache_free(req);
 969
 970         RETURN(rc);
 971 }
 972
 973 /**
 974  * Used by ptlrpc client, to perform the pre-defined security transformation
 975  * upon the request message of \a req. After this function called,
 976  * req->rq_reqmsg is still accessible as clear text.
 977  */
 978 int sptlrpc_cli_wrap_request(struct ptlrpc_request *req)
 979 {
 980         struct ptlrpc_cli_ctx *ctx = req->rq_cli_ctx;
 981         int rc = 0;
 982         ENTRY;
 983
 984         LASSERT(ctx);
 985         LASSERT(ctx->cc_sec);
 986         LASSERT(req->rq_reqbuf || req->rq_clrbuf);
 987
 988         /* we wrap bulk request here because now we can be sure
 989          * the context is uptodate.
 990          */
 991         if (req->rq_bulk) {
 992                 rc = sptlrpc_cli_wrap_bulk(req, req->rq_bulk);
 993                 if (rc)
 994                         RETURN(rc);
 995         }
 996
 997         switch (SPTLRPC_FLVR_SVC(req->rq_flvr.sf_rpc)) {
 998         case SPTLRPC_SVC_NULL:
 999         case SPTLRPC_SVC_AUTH:
1000         case SPTLRPC_SVC_INTG:
1001                 LASSERT(ctx->cc_ops->sign);
1002                 rc = ctx->cc_ops->sign(ctx, req);
1003                 break;
1004         case SPTLRPC_SVC_PRIV:
1005                 LASSERT(ctx->cc_ops->seal);
1006                 rc = ctx->cc_ops->seal(ctx, req);
1007                 break;
1008         default:
1009                 LBUG();
1010         }
1011
1012         if (rc == 0) {
1013                 LASSERT(req->rq_reqdata_len);
1014                 LASSERT(req->rq_reqdata_len % 8 == 0);
1015                 LASSERT(req->rq_reqdata_len <= req->rq_reqbuf_len);
1016         }
1017
1018         RETURN(rc);
1019 }
1020
1021 static int do_cli_unwrap_reply(struct ptlrpc_request *req)
1022 {
1023         struct ptlrpc_cli_ctx *ctx = req->rq_cli_ctx;
1024         int                    rc;
1025         ENTRY;
1026
1027         LASSERT(ctx);
1028         LASSERT(ctx->cc_sec);
1029         LASSERT(req->rq_repbuf);
1030         LASSERT(req->rq_repdata);
1031         LASSERT(req->rq_repmsg == NULL);
1032
1033         req->rq_rep_swab_mask = 0;
1034
1035         rc = __lustre_unpack_msg(req->rq_repdata, req->rq_repdata_len);
1036         switch (rc) {
1037         case 1:
1038                 lustre_set_rep_swabbed(req, MSG_PTLRPC_HEADER_OFF);
1039         case 0:
1040                 break;
1041         default:
1042                 CERROR("failed unpack reply: x"LPU64"\n", req->rq_xid);
1043                 RETURN(-EPROTO);
1044         }
1045
1046         if (req->rq_repdata_len < sizeof(struct lustre_msg)) {
1047                 CERROR("replied data length %d too small\n",
1048                        req->rq_repdata_len);
1049                 RETURN(-EPROTO);
1050         }
1051
1052         if (SPTLRPC_FLVR_POLICY(req->rq_repdata->lm_secflvr) !=
1053             SPTLRPC_FLVR_POLICY(req->rq_flvr.sf_rpc)) {
1054                 CERROR("reply policy %u doesn't match request policy %u\n",
1055                        SPTLRPC_FLVR_POLICY(req->rq_repdata->lm_secflvr),
1056                        SPTLRPC_FLVR_POLICY(req->rq_flvr.sf_rpc));
1057                 RETURN(-EPROTO);
1058         }
1059
1060         switch (SPTLRPC_FLVR_SVC(req->rq_flvr.sf_rpc)) {
1061         case SPTLRPC_SVC_NULL:
1062         case SPTLRPC_SVC_AUTH:
1063         case SPTLRPC_SVC_INTG:
1064                 LASSERT(ctx->cc_ops->verify);
1065                 rc = ctx->cc_ops->verify(ctx, req);
1066                 break;
1067         case SPTLRPC_SVC_PRIV:
1068                 LASSERT(ctx->cc_ops->unseal);
1069                 rc = ctx->cc_ops->unseal(ctx, req);
1070                 break;
1071         default:
1072                 LBUG();
1073         }
1074         LASSERT(rc || req->rq_repmsg || req->rq_resend);
1075
1076         if (SPTLRPC_FLVR_POLICY(req->rq_flvr.sf_rpc) != SPTLRPC_POLICY_NULL &&
1077             !req->rq_ctx_init)
1078                 req->rq_rep_swab_mask = 0;
1079         RETURN(rc);
1080 }
1081
1082 /**
1083  * Used by ptlrpc client, to perform security transformation upon the reply
1084  * message of \a req. After return successfully, req->rq_repmsg points to
1085  * the reply message in clear text.
1086  *
1087  * \pre the reply buffer should have been un-posted from LNet, so nothing is
1088  * going to change.
1089  */
1090 int sptlrpc_cli_unwrap_reply(struct ptlrpc_request *req)
1091 {
1092         LASSERT(req->rq_repbuf);
1093         LASSERT(req->rq_repdata == NULL);
1094         LASSERT(req->rq_repmsg == NULL);
1095         LASSERT(req->rq_reply_off + req->rq_nob_received <= req->rq_repbuf_len);
1096
1097         if (req->rq_reply_off == 0 &&
1098             (lustre_msghdr_get_flags(req->rq_reqmsg) & MSGHDR_AT_SUPPORT)) {
1099                 CERROR("real reply with offset 0\n");
1100                 return -EPROTO;
1101         }
1102
1103         if (req->rq_reply_off % 8 != 0) {
1104                 CERROR("reply at odd offset %u\n", req->rq_reply_off);
1105                 return -EPROTO;
1106         }
1107
1108         req->rq_repdata = (struct lustre_msg *)
1109                                 (req->rq_repbuf + req->rq_reply_off);
1110         req->rq_repdata_len = req->rq_nob_received;
1111
1112         return do_cli_unwrap_reply(req);
1113 }
1114
1115 /**
1116  * Used by ptlrpc client, to perform security transformation upon the early
1117  * reply message of \a req. We expect the rq_reply_off is 0, and
1118  * rq_nob_received is the early reply size.
1119  *
1120  * Because the receive buffer might be still posted, the reply data might be
1121  * changed at any time, no matter we're holding rq_lock or not. For this reason
1122  * we allocate a separate ptlrpc_request and reply buffer for early reply
1123  * processing.
1124  *
1125  * \retval 0 success, \a req_ret is filled with a duplicated ptlrpc_request.
1126  * Later the caller must call sptlrpc_cli_finish_early_reply() on the returned
1127  * \a *req_ret to release it.
1128  * \retval -ev error number, and \a req_ret will not be set.
1129  */
1130 int sptlrpc_cli_unwrap_early_reply(struct ptlrpc_request *req,
1131                                    struct ptlrpc_request **req_ret)
1132 {
1133         struct ptlrpc_request  *early_req;
1134         char                   *early_buf;
1135         int                     early_bufsz, early_size;
1136         int                     rc;
1137         ENTRY;
1138
1139         early_req = ptlrpc_request_cache_alloc(GFP_NOFS);
1140         if (early_req == NULL)
1141                 RETURN(-ENOMEM);
1142
1143         ptlrpc_cli_req_init(early_req);
1144
1145         early_size = req->rq_nob_received;
1146         early_bufsz = size_roundup_power2(early_size);
1147         OBD_ALLOC_LARGE(early_buf, early_bufsz);
1148         if (early_buf == NULL)
1149                 GOTO(err_req, rc = -ENOMEM);
1150
1151         /* sanity checkings and copy data out, do it inside spinlock */
1152         spin_lock(&req->rq_lock);
1153
1154         if (req->rq_replied) {
1155                 spin_unlock(&req->rq_lock);
1156                 GOTO(err_buf, rc = -EALREADY);
1157         }
1158
1159         LASSERT(req->rq_repbuf);
1160         LASSERT(req->rq_repdata == NULL);
1161         LASSERT(req->rq_repmsg == NULL);
1162
1163         if (req->rq_reply_off != 0) {
1164                 CERROR("early reply with offset %u\n", req->rq_reply_off);
1165                 spin_unlock(&req->rq_lock);
1166                 GOTO(err_buf, rc = -EPROTO);
1167         }
1168
1169         if (req->rq_nob_received != early_size) {
1170                 /* even another early arrived the size should be the same */
1171                 CERROR("data size has changed from %u to %u\n",
1172                        early_size, req->rq_nob_received);
1173                 spin_unlock(&req->rq_lock);
1174                 GOTO(err_buf, rc = -EINVAL);
1175         }
1176
1177         if (req->rq_nob_received < sizeof(struct lustre_msg)) {
1178                 CERROR("early reply length %d too small\n",
1179                        req->rq_nob_received);
1180                 spin_unlock(&req->rq_lock);
1181                 GOTO(err_buf, rc = -EALREADY);
1182         }
1183
1184         memcpy(early_buf, req->rq_repbuf, early_size);
1185         spin_unlock(&req->rq_lock);
1186
1187         early_req->rq_cli_ctx = sptlrpc_cli_ctx_get(req->rq_cli_ctx);
1188         early_req->rq_flvr = req->rq_flvr;
1189         early_req->rq_repbuf = early_buf;
1190         early_req->rq_repbuf_len = early_bufsz;
1191         early_req->rq_repdata = (struct lustre_msg *) early_buf;
1192         early_req->rq_repdata_len = early_size;
1193         early_req->rq_early = 1;
1194         early_req->rq_reqmsg = req->rq_reqmsg;
1195
1196         rc = do_cli_unwrap_reply(early_req);
1197         if (rc) {
1198                 DEBUG_REQ(D_ADAPTTO, early_req,
1199                           "error %d unwrap early reply", rc);
1200                 GOTO(err_ctx, rc);
1201         }
1202
1203         LASSERT(early_req->rq_repmsg);
1204         *req_ret = early_req;
1205         RETURN(0);
1206
1207 err_ctx:
1208         sptlrpc_cli_ctx_put(early_req->rq_cli_ctx, 1);
1209 err_buf:
1210         OBD_FREE_LARGE(early_buf, early_bufsz);
1211 err_req:
1212         ptlrpc_request_cache_free(early_req);
1213         RETURN(rc);
1214 }
1215
1216 /**
1217  * Used by ptlrpc client, to release a processed early reply \a early_req.
1218  *
1219  * \pre \a early_req was obtained from calling sptlrpc_cli_unwrap_early_reply().
1220  */
1221 void sptlrpc_cli_finish_early_reply(struct ptlrpc_request *early_req)
1222 {
1223         LASSERT(early_req->rq_repbuf);
1224         LASSERT(early_req->rq_repdata);
1225         LASSERT(early_req->rq_repmsg);
1226
1227         sptlrpc_cli_ctx_put(early_req->rq_cli_ctx, 1);
1228         OBD_FREE_LARGE(early_req->rq_repbuf, early_req->rq_repbuf_len);
1229         ptlrpc_request_cache_free(early_req);
1230 }
1231
1232 /**************************************************
1233  * sec ID                                         *
1234  **************************************************/
1235
1236 /*
1237  * "fixed" sec (e.g. null) use sec_id < 0
1238  */
1239 static atomic_t sptlrpc_sec_id = ATOMIC_INIT(1);
1240
1241 int sptlrpc_get_next_secid(void)
1242 {
1243         return atomic_inc_return(&sptlrpc_sec_id);
1244 }
1245 EXPORT_SYMBOL(sptlrpc_get_next_secid);
1246
1247 /**************************************************
1248  * client side high-level security APIs           *
1249  **************************************************/
1250
1251 static int sec_cop_flush_ctx_cache(struct ptlrpc_sec *sec, uid_t uid,
1252                                    int grace, int force)
1253 {
1254         struct ptlrpc_sec_policy *policy = sec->ps_policy;
1255
1256         LASSERT(policy->sp_cops);
1257         LASSERT(policy->sp_cops->flush_ctx_cache);
1258
1259         return policy->sp_cops->flush_ctx_cache(sec, uid, grace, force);
1260 }
1261
1262 static void sec_cop_destroy_sec(struct ptlrpc_sec *sec)
1263 {
1264         struct ptlrpc_sec_policy *policy = sec->ps_policy;
1265
1266         LASSERT_ATOMIC_ZERO(&sec->ps_refcount);
1267         LASSERT_ATOMIC_ZERO(&sec->ps_nctx);
1268         LASSERT(policy->sp_cops->destroy_sec);
1269
1270         CDEBUG(D_SEC, "%s@%p: being destroied\n", sec->ps_policy->sp_name, sec);
1271
1272         policy->sp_cops->destroy_sec(sec);
1273         sptlrpc_policy_put(policy);
1274 }
1275
1276 void sptlrpc_sec_destroy(struct ptlrpc_sec *sec)
1277 {
1278         sec_cop_destroy_sec(sec);
1279 }
1280 EXPORT_SYMBOL(sptlrpc_sec_destroy);
1281
1282 static void sptlrpc_sec_kill(struct ptlrpc_sec *sec)
1283 {
1284         LASSERT_ATOMIC_POS(&sec->ps_refcount);
1285
1286         if (sec->ps_policy->sp_cops->kill_sec) {
1287                 sec->ps_policy->sp_cops->kill_sec(sec);
1288
1289                 sec_cop_flush_ctx_cache(sec, -1, 1, 1);
1290         }
1291 }
1292
1293 struct ptlrpc_sec *sptlrpc_sec_get(struct ptlrpc_sec *sec)
1294 {
1295         if (sec)
1296                 atomic_inc(&sec->ps_refcount);
1297
1298         return sec;
1299 }
1300 EXPORT_SYMBOL(sptlrpc_sec_get);
1301
1302 void sptlrpc_sec_put(struct ptlrpc_sec *sec)
1303 {
1304         if (sec) {
1305                 LASSERT_ATOMIC_POS(&sec->ps_refcount);
1306
1307                 if (atomic_dec_and_test(&sec->ps_refcount)) {
1308                         sptlrpc_gc_del_sec(sec);
1309                         sec_cop_destroy_sec(sec);
1310                 }
1311         }
1312 }
1313 EXPORT_SYMBOL(sptlrpc_sec_put);
1314
1315 /*
1316  * policy module is responsible for taking refrence of import
1317  */
1318 static
1319 struct ptlrpc_sec * sptlrpc_sec_create(struct obd_import *imp,
1320                                        struct ptlrpc_svc_ctx *svc_ctx,
1321                                        struct sptlrpc_flavor *sf,
1322                                        enum lustre_sec_part sp)
1323 {
1324         struct ptlrpc_sec_policy *policy;
1325         struct ptlrpc_sec        *sec;
1326         char                      str[32];
1327         ENTRY;
1328
1329         if (svc_ctx) {
1330                 LASSERT(imp->imp_dlm_fake == 1);
1331
1332                 CDEBUG(D_SEC, "%s %s: reverse sec using flavor %s\n",
1333                        imp->imp_obd->obd_type->typ_name,
1334                        imp->imp_obd->obd_name,
1335                        sptlrpc_flavor2name(sf, str, sizeof(str)));
1336
1337                 policy = sptlrpc_policy_get(svc_ctx->sc_policy);
1338                 sf->sf_flags |= PTLRPC_SEC_FL_REVERSE | PTLRPC_SEC_FL_ROOTONLY;
1339         } else {
1340                 LASSERT(imp->imp_dlm_fake == 0);
1341
1342                 CDEBUG(D_SEC, "%s %s: select security flavor %s\n",
1343                        imp->imp_obd->obd_type->typ_name,
1344                        imp->imp_obd->obd_name,
1345                        sptlrpc_flavor2name(sf, str, sizeof(str)));
1346
1347                 policy = sptlrpc_wireflavor2policy(sf->sf_rpc);
1348                 if (!policy) {
1349                         CERROR("invalid flavor 0x%x\n", sf->sf_rpc);
1350                         RETURN(NULL);
1351                 }
1352         }
1353
1354         sec = policy->sp_cops->create_sec(imp, svc_ctx, sf);
1355         if (sec) {
1356                 atomic_inc(&sec->ps_refcount);
1357
1358                 sec->ps_part = sp;
1359
1360                 if (sec->ps_gc_interval && policy->sp_cops->gc_ctx)
1361                         sptlrpc_gc_add_sec(sec);
1362         } else {
1363                 sptlrpc_policy_put(policy);
1364         }
1365
1366         RETURN(sec);
1367 }
1368
1369 struct ptlrpc_sec *sptlrpc_import_sec_ref(struct obd_import *imp)
1370 {
1371         struct ptlrpc_sec *sec;
1372
1373         spin_lock(&imp->imp_lock);
1374         sec = sptlrpc_sec_get(imp->imp_sec);
1375         spin_unlock(&imp->imp_lock);
1376
1377         return sec;
1378 }
1379 EXPORT_SYMBOL(sptlrpc_import_sec_ref);
1380
1381 static void sptlrpc_import_sec_install(struct obd_import *imp,
1382                                        struct ptlrpc_sec *sec)
1383 {
1384         struct ptlrpc_sec *old_sec;
1385
1386         LASSERT_ATOMIC_POS(&sec->ps_refcount);
1387
1388         spin_lock(&imp->imp_lock);
1389         old_sec = imp->imp_sec;
1390         imp->imp_sec = sec;
1391         spin_unlock(&imp->imp_lock);
1392
1393         if (old_sec) {
1394                 sptlrpc_sec_kill(old_sec);
1395
1396                 /* balance the ref taken by this import */
1397                 sptlrpc_sec_put(old_sec);
1398         }
1399 }
1400
1401 static inline
1402 int flavor_equal(struct sptlrpc_flavor *sf1, struct sptlrpc_flavor *sf2)
1403 {
1404         return (memcmp(sf1, sf2, sizeof(*sf1)) == 0);
1405 }
1406
1407 static inline
1408 void flavor_copy(struct sptlrpc_flavor *dst, struct sptlrpc_flavor *src)
1409 {
1410         *dst = *src;
1411 }
1412
1413 static void sptlrpc_import_sec_adapt_inplace(struct obd_import *imp,
1414                                              struct ptlrpc_sec *sec,
1415                                              struct sptlrpc_flavor *sf)
1416 {
1417         char    str1[32], str2[32];
1418
1419         if (sec->ps_flvr.sf_flags != sf->sf_flags)
1420                 CDEBUG(D_SEC, "changing sec flags: %s -> %s\n",
1421                        sptlrpc_secflags2str(sec->ps_flvr.sf_flags,
1422                                             str1, sizeof(str1)),
1423                        sptlrpc_secflags2str(sf->sf_flags,
1424                                             str2, sizeof(str2)));
1425
1426         spin_lock(&sec->ps_lock);
1427         flavor_copy(&sec->ps_flvr, sf);
1428         spin_unlock(&sec->ps_lock);
1429 }
1430
1431 /**
1432  * To get an appropriate ptlrpc_sec for the \a imp, according to the current
1433  * configuration. Upon called, imp->imp_sec may or may not be NULL.
1434  *
1435  *  - regular import: \a svc_ctx should be NULL and \a flvr is ignored;
1436  *  - reverse import: \a svc_ctx and \a flvr are obtained from incoming request.
1437  */
1438 int sptlrpc_import_sec_adapt(struct obd_import *imp,
1439                              struct ptlrpc_svc_ctx *svc_ctx,
1440                              struct sptlrpc_flavor *flvr)
1441 {
1442         struct ptlrpc_connection   *conn;
1443         struct sptlrpc_flavor       sf;
1444         struct ptlrpc_sec          *sec, *newsec;
1445         enum lustre_sec_part        sp;
1446         char                        str[24];
1447         int                         rc = 0;
1448         ENTRY;
1449
1450         might_sleep();
1451
1452         if (imp == NULL)
1453                 RETURN(0);
1454
1455         conn = imp->imp_connection;
1456
1457         if (svc_ctx == NULL) {
1458                 struct client_obd *cliobd = &imp->imp_obd->u.cli;
1459                 /*
1460                  * normal import, determine flavor from rule set, except
1461                  * for mgc the flavor is predetermined.
1462                  */
1463                 if (cliobd->cl_sp_me == LUSTRE_SP_MGC)
1464                         sf = cliobd->cl_flvr_mgc;
1465                 else
1466                         sptlrpc_conf_choose_flavor(cliobd->cl_sp_me,
1467                                                    cliobd->cl_sp_to,
1468                                                    &cliobd->cl_target_uuid,
1469                                                    conn->c_self, &sf);
1470
1471                 sp = imp->imp_obd->u.cli.cl_sp_me;
1472         } else {
1473                 /* reverse import, determine flavor from incoming reqeust */
1474                 sf = *flvr;
1475
1476                 if (sf.sf_rpc != SPTLRPC_FLVR_NULL)
1477                         sf.sf_flags = PTLRPC_SEC_FL_REVERSE |
1478                                       PTLRPC_SEC_FL_ROOTONLY;
1479
1480                 sp = sptlrpc_target_sec_part(imp->imp_obd);
1481         }
1482
1483         sec = sptlrpc_import_sec_ref(imp);
1484         if (sec) {
1485                 char    str2[24];
1486
1487                 if (flavor_equal(&sf, &sec->ps_flvr))
1488                         GOTO(out, rc);
1489
1490                 CDEBUG(D_SEC, "import %s->%s: changing flavor %s -> %s\n",
1491                        imp->imp_obd->obd_name,
1492                        obd_uuid2str(&conn->c_remote_uuid),
1493                        sptlrpc_flavor2name(&sec->ps_flvr, str, sizeof(str)),
1494                        sptlrpc_flavor2name(&sf, str2, sizeof(str2)));
1495
1496                 if (SPTLRPC_FLVR_POLICY(sf.sf_rpc) ==
1497                     SPTLRPC_FLVR_POLICY(sec->ps_flvr.sf_rpc) &&
1498                     SPTLRPC_FLVR_MECH(sf.sf_rpc) ==
1499                     SPTLRPC_FLVR_MECH(sec->ps_flvr.sf_rpc)) {
1500                         sptlrpc_import_sec_adapt_inplace(imp, sec, &sf);
1501                         GOTO(out, rc);
1502                 }
1503         } else if (SPTLRPC_FLVR_BASE(sf.sf_rpc) !=
1504                    SPTLRPC_FLVR_BASE(SPTLRPC_FLVR_NULL)) {
1505                 CDEBUG(D_SEC, "import %s->%s netid %x: select flavor %s\n",
1506                        imp->imp_obd->obd_name,
1507                        obd_uuid2str(&conn->c_remote_uuid),
1508                        LNET_NIDNET(conn->c_self),
1509                        sptlrpc_flavor2name(&sf, str, sizeof(str)));
1510         }
1511
1512         mutex_lock(&imp->imp_sec_mutex);
1513
1514         newsec = sptlrpc_sec_create(imp, svc_ctx, &sf, sp);
1515         if (newsec) {
1516                 sptlrpc_import_sec_install(imp, newsec);
1517         } else {
1518                 CERROR("import %s->%s: failed to create new sec\n",
1519                        imp->imp_obd->obd_name,
1520                        obd_uuid2str(&conn->c_remote_uuid));
1521                 rc = -EPERM;
1522         }
1523
1524         mutex_unlock(&imp->imp_sec_mutex);
1525 out:
1526         sptlrpc_sec_put(sec);
1527         RETURN(rc);
1528 }
1529
1530 void sptlrpc_import_sec_put(struct obd_import *imp)
1531 {
1532         if (imp->imp_sec) {
1533                 sptlrpc_sec_kill(imp->imp_sec);
1534
1535                 sptlrpc_sec_put(imp->imp_sec);
1536                 imp->imp_sec = NULL;
1537         }
1538 }
1539
1540 static void import_flush_ctx_common(struct obd_import *imp,
1541                                     uid_t uid, int grace, int force)
1542 {
1543         struct ptlrpc_sec *sec;
1544
1545         if (imp == NULL)
1546                 return;
1547
1548         sec = sptlrpc_import_sec_ref(imp);
1549         if (sec == NULL)
1550                 return;
1551
1552         sec_cop_flush_ctx_cache(sec, uid, grace, force);
1553         sptlrpc_sec_put(sec);
1554 }
1555
1556 void sptlrpc_import_flush_root_ctx(struct obd_import *imp)
1557 {
1558         /* it's important to use grace mode, see explain in
1559          * sptlrpc_req_refresh_ctx() */
1560         import_flush_ctx_common(imp, 0, 1, 1);
1561 }
1562
1563 void sptlrpc_import_flush_my_ctx(struct obd_import *imp)
1564 {
1565         import_flush_ctx_common(imp, from_kuid(&init_user_ns, current_uid()),
1566                                 1, 1);
1567 }
1568 EXPORT_SYMBOL(sptlrpc_import_flush_my_ctx);
1569
1570 void sptlrpc_import_flush_all_ctx(struct obd_import *imp)
1571 {
1572         import_flush_ctx_common(imp, -1, 1, 1);
1573 }
1574 EXPORT_SYMBOL(sptlrpc_import_flush_all_ctx);
1575
1576 /**
1577  * Used by ptlrpc client to allocate request buffer of \a req. Upon return
1578  * successfully, req->rq_reqmsg points to a buffer with size \a msgsize.
1579  */
1580 int sptlrpc_cli_alloc_reqbuf(struct ptlrpc_request *req, int msgsize)
1581 {
1582         struct ptlrpc_cli_ctx *ctx = req->rq_cli_ctx;
1583         struct ptlrpc_sec_policy *policy;
1584         int rc;
1585
1586         LASSERT(ctx);
1587         LASSERT(ctx->cc_sec);
1588         LASSERT(ctx->cc_sec->ps_policy);
1589         LASSERT(req->rq_reqmsg == NULL);
1590         LASSERT_ATOMIC_POS(&ctx->cc_refcount);
1591
1592         policy = ctx->cc_sec->ps_policy;
1593         rc = policy->sp_cops->alloc_reqbuf(ctx->cc_sec, req, msgsize);
1594         if (!rc) {
1595                 LASSERT(req->rq_reqmsg);
1596                 LASSERT(req->rq_reqbuf || req->rq_clrbuf);
1597
1598                 /* zeroing preallocated buffer */
1599                 if (req->rq_pool)
1600                         memset(req->rq_reqmsg, 0, msgsize);
1601         }
1602
1603         return rc;
1604 }
1605
1606 /**
1607  * Used by ptlrpc client to free request buffer of \a req. After this
1608  * req->rq_reqmsg is set to NULL and should not be accessed anymore.
1609  */
1610 void sptlrpc_cli_free_reqbuf(struct ptlrpc_request *req)
1611 {
1612         struct ptlrpc_cli_ctx *ctx = req->rq_cli_ctx;
1613         struct ptlrpc_sec_policy *policy;
1614
1615         LASSERT(ctx);
1616         LASSERT(ctx->cc_sec);
1617         LASSERT(ctx->cc_sec->ps_policy);
1618         LASSERT_ATOMIC_POS(&ctx->cc_refcount);
1619
1620         if (req->rq_reqbuf == NULL && req->rq_clrbuf == NULL)
1621                 return;
1622
1623         policy = ctx->cc_sec->ps_policy;
1624         policy->sp_cops->free_reqbuf(ctx->cc_sec, req);
1625         req->rq_reqmsg = NULL;
1626 }
1627
1628 /*
1629  * NOTE caller must guarantee the buffer size is enough for the enlargement
1630  */
1631 void _sptlrpc_enlarge_msg_inplace(struct lustre_msg *msg,
1632                                   int segment, int newsize)
1633 {
1634         void   *src, *dst;
1635         int     oldsize, oldmsg_size, movesize;
1636
1637         LASSERT(segment < msg->lm_bufcount);
1638         LASSERT(msg->lm_buflens[segment] <= newsize);
1639
1640         if (msg->lm_buflens[segment] == newsize)
1641                 return;
1642
1643         /* nothing to do if we are enlarging the last segment */
1644         if (segment == msg->lm_bufcount - 1) {
1645                 msg->lm_buflens[segment] = newsize;
1646                 return;
1647         }
1648
1649         oldsize = msg->lm_buflens[segment];
1650
1651         src = lustre_msg_buf(msg, segment + 1, 0);
1652         msg->lm_buflens[segment] = newsize;
1653         dst = lustre_msg_buf(msg, segment + 1, 0);
1654         msg->lm_buflens[segment] = oldsize;
1655
1656         /* move from segment + 1 to end segment */
1657         LASSERT(msg->lm_magic == LUSTRE_MSG_MAGIC_V2);
1658         oldmsg_size = lustre_msg_size_v2(msg->lm_bufcount, msg->lm_buflens);
1659         movesize = oldmsg_size - ((unsigned long) src - (unsigned long) msg);
1660         LASSERT(movesize >= 0);
1661
1662         if (movesize)
1663                 memmove(dst, src, movesize);
1664
1665         /* note we don't clear the ares where old data live, not secret */
1666
1667         /* finally set new segment size */
1668         msg->lm_buflens[segment] = newsize;
1669 }
1670 EXPORT_SYMBOL(_sptlrpc_enlarge_msg_inplace);
1671
1672 /**
1673  * Used by ptlrpc client to enlarge the \a segment of request message pointed
1674  * by req->rq_reqmsg to size \a newsize, all previously filled-in data will be
1675  * preserved after the enlargement. this must be called after original request
1676  * buffer being allocated.
1677  *
1678  * \note after this be called, rq_reqmsg and rq_reqlen might have been changed,
1679  * so caller should refresh its local pointers if needed.
1680  */
1681 int sptlrpc_cli_enlarge_reqbuf(struct ptlrpc_request *req,
1682                                int segment, int newsize)
1683 {
1684         struct ptlrpc_cli_ctx    *ctx = req->rq_cli_ctx;
1685         struct ptlrpc_sec_cops   *cops;
1686         struct lustre_msg        *msg = req->rq_reqmsg;
1687
1688         LASSERT(ctx);
1689         LASSERT(msg);
1690         LASSERT(msg->lm_bufcount > segment);
1691         LASSERT(msg->lm_buflens[segment] <= newsize);
1692
1693         if (msg->lm_buflens[segment] == newsize)
1694                 return 0;
1695
1696         cops = ctx->cc_sec->ps_policy->sp_cops;
1697         LASSERT(cops->enlarge_reqbuf);
1698         return cops->enlarge_reqbuf(ctx->cc_sec, req, segment, newsize);
1699 }
1700 EXPORT_SYMBOL(sptlrpc_cli_enlarge_reqbuf);
1701
1702 /**
1703  * Used by ptlrpc client to allocate reply buffer of \a req.
1704  *
1705  * \note After this, req->rq_repmsg is still not accessible.
1706  */
1707 int sptlrpc_cli_alloc_repbuf(struct ptlrpc_request *req, int msgsize)
1708 {
1709         struct ptlrpc_cli_ctx *ctx = req->rq_cli_ctx;
1710         struct ptlrpc_sec_policy *policy;
1711         ENTRY;
1712
1713         LASSERT(ctx);
1714         LASSERT(ctx->cc_sec);
1715         LASSERT(ctx->cc_sec->ps_policy);
1716
1717         if (req->rq_repbuf)
1718                 RETURN(0);
1719
1720         policy = ctx->cc_sec->ps_policy;
1721         RETURN(policy->sp_cops->alloc_repbuf(ctx->cc_sec, req, msgsize));
1722 }
1723
1724 /**
1725  * Used by ptlrpc client to free reply buffer of \a req. After this
1726  * req->rq_repmsg is set to NULL and should not be accessed anymore.
1727  */
1728 void sptlrpc_cli_free_repbuf(struct ptlrpc_request *req)
1729 {
1730         struct ptlrpc_cli_ctx *ctx = req->rq_cli_ctx;
1731         struct ptlrpc_sec_policy *policy;
1732         ENTRY;
1733
1734         LASSERT(ctx);
1735         LASSERT(ctx->cc_sec);
1736         LASSERT(ctx->cc_sec->ps_policy);
1737         LASSERT_ATOMIC_POS(&ctx->cc_refcount);
1738
1739         if (req->rq_repbuf == NULL)
1740                 return;
1741         LASSERT(req->rq_repbuf_len);
1742
1743         policy = ctx->cc_sec->ps_policy;
1744         policy->sp_cops->free_repbuf(ctx->cc_sec, req);
1745         req->rq_repmsg = NULL;
1746         EXIT;
1747 }
1748
1749 int sptlrpc_cli_install_rvs_ctx(struct obd_import *imp,
1750                                 struct ptlrpc_cli_ctx *ctx)
1751 {
1752         struct ptlrpc_sec_policy *policy = ctx->cc_sec->ps_policy;
1753
1754         if (!policy->sp_cops->install_rctx)
1755                 return 0;
1756         return policy->sp_cops->install_rctx(imp, ctx->cc_sec, ctx);
1757 }
1758
1759 int sptlrpc_svc_install_rvs_ctx(struct obd_import *imp,
1760                                 struct ptlrpc_svc_ctx *ctx)
1761 {
1762         struct ptlrpc_sec_policy *policy = ctx->sc_policy;
1763
1764         if (!policy->sp_sops->install_rctx)
1765                 return 0;
1766         return policy->sp_sops->install_rctx(imp, ctx);
1767 }
1768
1769 /****************************************
1770  * server side security                 *
1771  ****************************************/
1772
1773 static int flavor_allowed(struct sptlrpc_flavor *exp,
1774                           struct ptlrpc_request *req)
1775 {
1776         struct sptlrpc_flavor *flvr = &req->rq_flvr;
1777
1778         if (exp->sf_rpc == SPTLRPC_FLVR_ANY || exp->sf_rpc == flvr->sf_rpc)
1779                 return 1;
1780
1781         if ((req->rq_ctx_init || req->rq_ctx_fini) &&
1782             SPTLRPC_FLVR_POLICY(exp->sf_rpc) ==
1783             SPTLRPC_FLVR_POLICY(flvr->sf_rpc) &&
1784             SPTLRPC_FLVR_MECH(exp->sf_rpc) == SPTLRPC_FLVR_MECH(flvr->sf_rpc))
1785                 return 1;
1786
1787         return 0;
1788 }
1789
1790 #define EXP_FLVR_UPDATE_EXPIRE      (OBD_TIMEOUT_DEFAULT + 10)
1791
1792 /**
1793  * Given an export \a exp, check whether the flavor of incoming \a req
1794  * is allowed by the export \a exp. Main logic is about taking care of
1795  * changing configurations. Return 0 means success.
1796  */
1797 int sptlrpc_target_export_check(struct obd_export *exp,
1798                                 struct ptlrpc_request *req)
1799 {
1800         struct sptlrpc_flavor   flavor;
1801
1802         if (exp == NULL)
1803                 return 0;
1804
1805         /* client side export has no imp_reverse, skip
1806          * FIXME maybe we should check flavor this as well??? */
1807         if (exp->exp_imp_reverse == NULL)
1808                 return 0;
1809
1810         /* don't care about ctx fini rpc */
1811         if (req->rq_ctx_fini)
1812                 return 0;
1813
1814         spin_lock(&exp->exp_lock);
1815
1816         /* if flavor just changed (exp->exp_flvr_changed != 0), we wait for
1817          * the first req with the new flavor, then treat it as current flavor,
1818          * adapt reverse sec according to it.
1819          * note the first rpc with new flavor might not be with root ctx, in
1820          * which case delay the sec_adapt by leaving exp_flvr_adapt == 1. */
1821         if (unlikely(exp->exp_flvr_changed) &&
1822             flavor_allowed(&exp->exp_flvr_old[1], req)) {
1823                 /* make the new flavor as "current", and old ones as
1824                  * about-to-expire */
1825                 CDEBUG(D_SEC, "exp %p: just changed: %x->%x\n", exp,
1826                        exp->exp_flvr.sf_rpc, exp->exp_flvr_old[1].sf_rpc);
1827                 flavor = exp->exp_flvr_old[1];
1828                 exp->exp_flvr_old[1] = exp->exp_flvr_old[0];
1829                 exp->exp_flvr_expire[1] = exp->exp_flvr_expire[0];
1830                 exp->exp_flvr_old[0] = exp->exp_flvr;
1831                 exp->exp_flvr_expire[0] = cfs_time_current_sec() +
1832                                           EXP_FLVR_UPDATE_EXPIRE;
1833                 exp->exp_flvr = flavor;
1834
1835                 /* flavor change finished */
1836                 exp->exp_flvr_changed = 0;
1837                 LASSERT(exp->exp_flvr_adapt == 1);
1838
1839                 /* if it's gss, we only interested in root ctx init */
1840                 if (req->rq_auth_gss &&
1841                     !(req->rq_ctx_init &&
1842                       (req->rq_auth_usr_root || req->rq_auth_usr_mdt ||
1843                        req->rq_auth_usr_ost))) {
1844                         spin_unlock(&exp->exp_lock);
1845                         CDEBUG(D_SEC, "is good but not root(%d:%d:%d:%d:%d)\n",
1846                                req->rq_auth_gss, req->rq_ctx_init,
1847                                req->rq_auth_usr_root, req->rq_auth_usr_mdt,
1848                                req->rq_auth_usr_ost);
1849                         return 0;
1850                 }
1851
1852                 exp->exp_flvr_adapt = 0;
1853                 spin_unlock(&exp->exp_lock);
1854
1855                 return sptlrpc_import_sec_adapt(exp->exp_imp_reverse,
1856                                                 req->rq_svc_ctx, &flavor);
1857         }
1858
1859         /* if it equals to the current flavor, we accept it, but need to
1860          * dealing with reverse sec/ctx */
1861         if (likely(flavor_allowed(&exp->exp_flvr, req))) {
1862                 /* most cases should return here, we only interested in
1863                  * gss root ctx init */
1864                 if (!req->rq_auth_gss || !req->rq_ctx_init ||
1865                     (!req->rq_auth_usr_root && !req->rq_auth_usr_mdt &&
1866                      !req->rq_auth_usr_ost)) {
1867                         spin_unlock(&exp->exp_lock);
1868                         return 0;
1869                 }
1870
1871                 /* if flavor just changed, we should not proceed, just leave
1872                  * it and current flavor will be discovered and replaced
1873                  * shortly, and let _this_ rpc pass through */
1874                 if (exp->exp_flvr_changed) {
1875                         LASSERT(exp->exp_flvr_adapt);
1876                         spin_unlock(&exp->exp_lock);
1877                         return 0;
1878                 }
1879
1880                 if (exp->exp_flvr_adapt) {
1881                         exp->exp_flvr_adapt = 0;
1882                         CDEBUG(D_SEC, "exp %p (%x|%x|%x): do delayed adapt\n",
1883                                exp, exp->exp_flvr.sf_rpc,
1884                                exp->exp_flvr_old[0].sf_rpc,
1885                                exp->exp_flvr_old[1].sf_rpc);
1886                         flavor = exp->exp_flvr;
1887                         spin_unlock(&exp->exp_lock);
1888
1889                         return sptlrpc_import_sec_adapt(exp->exp_imp_reverse,
1890                                                         req->rq_svc_ctx,
1891                                                         &flavor);
1892                 } else {
1893                         CDEBUG(D_SEC, "exp %p (%x|%x|%x): is current flavor, "
1894                                "install rvs ctx\n", exp, exp->exp_flvr.sf_rpc,
1895                                exp->exp_flvr_old[0].sf_rpc,
1896                                exp->exp_flvr_old[1].sf_rpc);
1897                         spin_unlock(&exp->exp_lock);
1898
1899                         return sptlrpc_svc_install_rvs_ctx(exp->exp_imp_reverse,
1900                                                            req->rq_svc_ctx);
1901                 }
1902         }
1903
1904         if (exp->exp_flvr_expire[0]) {
1905                 if (exp->exp_flvr_expire[0] >= cfs_time_current_sec()) {
1906                         if (flavor_allowed(&exp->exp_flvr_old[0], req)) {
1907                                 CDEBUG(D_SEC, "exp %p (%x|%x|%x): match the "
1908                                        "middle one ("CFS_DURATION_T")\n", exp,
1909                                        exp->exp_flvr.sf_rpc,
1910                                        exp->exp_flvr_old[0].sf_rpc,
1911                                        exp->exp_flvr_old[1].sf_rpc,
1912                                        exp->exp_flvr_expire[0] -
1913                                                 cfs_time_current_sec());
1914                                 spin_unlock(&exp->exp_lock);
1915                                 return 0;
1916                         }
1917                 } else {
1918                         CDEBUG(D_SEC, "mark middle expired\n");
1919                         exp->exp_flvr_expire[0] = 0;
1920                 }
1921                 CDEBUG(D_SEC, "exp %p (%x|%x|%x): %x not match middle\n", exp,
1922                        exp->exp_flvr.sf_rpc,
1923                        exp->exp_flvr_old[0].sf_rpc, exp->exp_flvr_old[1].sf_rpc,
1924                        req->rq_flvr.sf_rpc);
1925         }
1926
1927         /* now it doesn't match the current flavor, the only chance we can
1928          * accept it is match the old flavors which is not expired. */
1929         if (exp->exp_flvr_changed == 0 && exp->exp_flvr_expire[1]) {
1930                 if (exp->exp_flvr_expire[1] >= cfs_time_current_sec()) {
1931                         if (flavor_allowed(&exp->exp_flvr_old[1], req)) {
1932                                 CDEBUG(D_SEC, "exp %p (%x|%x|%x): match the "
1933                                        "oldest one ("CFS_DURATION_T")\n", exp,
1934                                        exp->exp_flvr.sf_rpc,
1935                                        exp->exp_flvr_old[0].sf_rpc,
1936                                        exp->exp_flvr_old[1].sf_rpc,
1937                                        exp->exp_flvr_expire[1] -
1938                                                 cfs_time_current_sec());
1939                                 spin_unlock(&exp->exp_lock);
1940                                 return 0;
1941                         }
1942                 } else {
1943                         CDEBUG(D_SEC, "mark oldest expired\n");
1944                         exp->exp_flvr_expire[1] = 0;
1945                 }
1946                 CDEBUG(D_SEC, "exp %p (%x|%x|%x): %x not match found\n",
1947                        exp, exp->exp_flvr.sf_rpc,
1948                        exp->exp_flvr_old[0].sf_rpc, exp->exp_flvr_old[1].sf_rpc,
1949                        req->rq_flvr.sf_rpc);
1950         } else {
1951                 CDEBUG(D_SEC, "exp %p (%x|%x|%x): skip the last one\n",
1952                        exp, exp->exp_flvr.sf_rpc, exp->exp_flvr_old[0].sf_rpc,
1953                        exp->exp_flvr_old[1].sf_rpc);
1954         }
1955
1956         spin_unlock(&exp->exp_lock);
1957
1958         CWARN("exp %p(%s): req %p (%u|%u|%u|%u|%u|%u) with "
1959               "unauthorized flavor %x, expect %x|%x(%+ld)|%x(%+ld)\n",
1960               exp, exp->exp_obd->obd_name,
1961               req, req->rq_auth_gss, req->rq_ctx_init, req->rq_ctx_fini,
1962               req->rq_auth_usr_root, req->rq_auth_usr_mdt, req->rq_auth_usr_ost,
1963               req->rq_flvr.sf_rpc,
1964               exp->exp_flvr.sf_rpc,
1965               exp->exp_flvr_old[0].sf_rpc,
1966               exp->exp_flvr_expire[0] ?
1967               (unsigned long) (exp->exp_flvr_expire[0] -
1968                                cfs_time_current_sec()) : 0,
1969               exp->exp_flvr_old[1].sf_rpc,
1970               exp->exp_flvr_expire[1] ?
1971               (unsigned long) (exp->exp_flvr_expire[1] -
1972                                cfs_time_current_sec()) : 0);
1973         return -EACCES;
1974 }
1975 EXPORT_SYMBOL(sptlrpc_target_export_check);
1976
1977 void sptlrpc_target_update_exp_flavor(struct obd_device *obd,
1978                                       struct sptlrpc_rule_set *rset)
1979 {
1980         struct obd_export       *exp;
1981         struct sptlrpc_flavor    new_flvr;
1982
1983         LASSERT(obd);
1984
1985         spin_lock(&obd->obd_dev_lock);
1986
1987         list_for_each_entry(exp, &obd->obd_exports, exp_obd_chain) {
1988                 if (exp->exp_connection == NULL)
1989                         continue;
1990
1991                 /* note if this export had just been updated flavor
1992                  * (exp_flvr_changed == 1), this will override the
1993                  * previous one. */
1994                 spin_lock(&exp->exp_lock);
1995                 sptlrpc_target_choose_flavor(rset, exp->exp_sp_peer,
1996                                              exp->exp_connection->c_peer.nid,
1997                                              &new_flvr);
1998                 if (exp->exp_flvr_changed ||
1999                     !flavor_equal(&new_flvr, &exp->exp_flvr)) {
2000                         exp->exp_flvr_old[1] = new_flvr;
2001                         exp->exp_flvr_expire[1] = 0;
2002                         exp->exp_flvr_changed = 1;
2003                         exp->exp_flvr_adapt = 1;
2004
2005                         CDEBUG(D_SEC, "exp %p (%s): updated flavor %x->%x\n",
2006                                exp, sptlrpc_part2name(exp->exp_sp_peer),
2007                                exp->exp_flvr.sf_rpc,
2008                                exp->exp_flvr_old[1].sf_rpc);
2009                 }
2010                 spin_unlock(&exp->exp_lock);
2011         }
2012
2013         spin_unlock(&obd->obd_dev_lock);
2014 }
2015 EXPORT_SYMBOL(sptlrpc_target_update_exp_flavor);
2016
2017 static int sptlrpc_svc_check_from(struct ptlrpc_request *req, int svc_rc)
2018 {
2019         /* peer's claim is unreliable unless gss is being used */
2020         if (!req->rq_auth_gss || svc_rc == SECSVC_DROP)
2021                 return svc_rc;
2022
2023         switch (req->rq_sp_from) {
2024         case LUSTRE_SP_CLI:
2025                 if (req->rq_auth_usr_mdt || req->rq_auth_usr_ost) {
2026                         DEBUG_REQ(D_ERROR, req, "faked source CLI");
2027                         svc_rc = SECSVC_DROP;
2028                 }
2029                 break;
2030         case LUSTRE_SP_MDT:
2031                 if (!req->rq_auth_usr_mdt) {
2032                         DEBUG_REQ(D_ERROR, req, "faked source MDT");
2033                         svc_rc = SECSVC_DROP;
2034                 }
2035                 break;
2036         case LUSTRE_SP_OST:
2037                 if (!req->rq_auth_usr_ost) {
2038                         DEBUG_REQ(D_ERROR, req, "faked source OST");
2039                         svc_rc = SECSVC_DROP;
2040                 }
2041                 break;
2042         case LUSTRE_SP_MGS:
2043         case LUSTRE_SP_MGC:
2044                 if (!req->rq_auth_usr_root && !req->rq_auth_usr_mdt &&
2045                     !req->rq_auth_usr_ost) {
2046                         DEBUG_REQ(D_ERROR, req, "faked source MGC/MGS");
2047                         svc_rc = SECSVC_DROP;
2048                 }
2049                 break;
2050         case LUSTRE_SP_ANY:
2051         default:
2052                 DEBUG_REQ(D_ERROR, req, "invalid source %u", req->rq_sp_from);
2053                 svc_rc = SECSVC_DROP;
2054         }
2055
2056         return svc_rc;
2057 }
2058
2059 /**
2060  * Used by ptlrpc server, to perform transformation upon request message of
2061  * incoming \a req. This must be the first thing to do with an incoming
2062  * request in ptlrpc layer.
2063  *
2064  * \retval SECSVC_OK success, and req->rq_reqmsg point to request message in
2065  * clear text, size is req->rq_reqlen; also req->rq_svc_ctx is set.
2066  * \retval SECSVC_COMPLETE success, the request has been fully processed, and
2067  * reply message has been prepared.
2068  * \retval SECSVC_DROP failed, this request should be dropped.
2069  */
2070 int sptlrpc_svc_unwrap_request(struct ptlrpc_request *req)
2071 {
2072         struct ptlrpc_sec_policy *policy;
2073         struct lustre_msg        *msg = req->rq_reqbuf;
2074         int                       rc;
2075         ENTRY;
2076
2077         LASSERT(msg);
2078         LASSERT(req->rq_reqmsg == NULL);
2079         LASSERT(req->rq_repmsg == NULL);
2080         LASSERT(req->rq_svc_ctx == NULL);
2081
2082         req->rq_req_swab_mask = 0;
2083
2084         rc = __lustre_unpack_msg(msg, req->rq_reqdata_len);
2085         switch (rc) {
2086         case 1:
2087                 lustre_set_req_swabbed(req, MSG_PTLRPC_HEADER_OFF);
2088         case 0:
2089                 break;
2090         default:
2091                 CERROR("error unpacking request from %s x"LPU64"\n",
2092                        libcfs_id2str(req->rq_peer), req->rq_xid);
2093                 RETURN(SECSVC_DROP);
2094         }
2095
2096         req->rq_flvr.sf_rpc = WIRE_FLVR(msg->lm_secflvr);
2097         req->rq_sp_from = LUSTRE_SP_ANY;
2098         req->rq_auth_uid = -1;          /* set to INVALID_UID */
2099         req->rq_auth_mapped_uid = -1;
2100
2101         policy = sptlrpc_wireflavor2policy(req->rq_flvr.sf_rpc);
2102         if (!policy) {
2103                 CERROR("unsupported rpc flavor %x\n", req->rq_flvr.sf_rpc);
2104                 RETURN(SECSVC_DROP);
2105         }
2106
2107         LASSERT(policy->sp_sops->accept);
2108         rc = policy->sp_sops->accept(req);
2109         sptlrpc_policy_put(policy);
2110         LASSERT(req->rq_reqmsg || rc != SECSVC_OK);
2111         LASSERT(req->rq_svc_ctx || rc == SECSVC_DROP);
2112
2113         /*
2114          * if it's not null flavor (which means embedded packing msg),
2115          * reset the swab mask for the comming inner msg unpacking.
2116          */
2117         if (SPTLRPC_FLVR_POLICY(req->rq_flvr.sf_rpc) != SPTLRPC_POLICY_NULL)
2118                 req->rq_req_swab_mask = 0;
2119
2120         /* sanity check for the request source */
2121         rc = sptlrpc_svc_check_from(req, rc);
2122         RETURN(rc);
2123 }
2124
2125 /**
2126  * Used by ptlrpc server, to allocate reply buffer for \a req. If succeed,
2127  * req->rq_reply_state is set, and req->rq_reply_state->rs_msg point to
2128  * a buffer of \a msglen size.
2129  */
2130 int sptlrpc_svc_alloc_rs(struct ptlrpc_request *req, int msglen)
2131 {
2132         struct ptlrpc_sec_policy *policy;
2133         struct ptlrpc_reply_state *rs;
2134         int rc;
2135         ENTRY;
2136
2137         LASSERT(req->rq_svc_ctx);
2138         LASSERT(req->rq_svc_ctx->sc_policy);
2139
2140         policy = req->rq_svc_ctx->sc_policy;
2141         LASSERT(policy->sp_sops->alloc_rs);
2142
2143         rc = policy->sp_sops->alloc_rs(req, msglen);
2144         if (unlikely(rc == -ENOMEM)) {
2145                 struct ptlrpc_service_part *svcpt = req->rq_rqbd->rqbd_svcpt;
2146                 if (svcpt->scp_service->srv_max_reply_size <
2147                    msglen + sizeof(struct ptlrpc_reply_state)) {
2148                         /* Just return failure if the size is too big */
2149                         CERROR("size of message is too big (%zd), %d allowed\n",
2150                                 msglen + sizeof(struct ptlrpc_reply_state),
2151                                 svcpt->scp_service->srv_max_reply_size);
2152                         RETURN(-ENOMEM);
2153                 }
2154
2155                 /* failed alloc, try emergency pool */
2156                 rs = lustre_get_emerg_rs(svcpt);
2157                 if (rs == NULL)
2158                         RETURN(-ENOMEM);
2159
2160                 req->rq_reply_state = rs;
2161                 rc = policy->sp_sops->alloc_rs(req, msglen);
2162                 if (rc) {
2163                         lustre_put_emerg_rs(rs);
2164                         req->rq_reply_state = NULL;
2165                 }
2166         }
2167
2168         LASSERT(rc != 0 ||
2169                 (req->rq_reply_state && req->rq_reply_state->rs_msg));
2170
2171         RETURN(rc);
2172 }
2173
2174 /**
2175  * Used by ptlrpc server, to perform transformation upon reply message.
2176  *
2177  * \post req->rq_reply_off is set to approriate server-controlled reply offset.
2178  * \post req->rq_repmsg and req->rq_reply_state->rs_msg becomes inaccessible.
2179  */
2180 int sptlrpc_svc_wrap_reply(struct ptlrpc_request *req)
2181 {
2182         struct ptlrpc_sec_policy *policy;
2183         int rc;
2184         ENTRY;
2185
2186         LASSERT(req->rq_svc_ctx);
2187         LASSERT(req->rq_svc_ctx->sc_policy);
2188
2189         policy = req->rq_svc_ctx->sc_policy;
2190         LASSERT(policy->sp_sops->authorize);
2191
2192         rc = policy->sp_sops->authorize(req);
2193         LASSERT(rc || req->rq_reply_state->rs_repdata_len);
2194
2195         RETURN(rc);
2196 }
2197
2198 /**
2199  * Used by ptlrpc server, to free reply_state.
2200  */
2201 void sptlrpc_svc_free_rs(struct ptlrpc_reply_state *rs)
2202 {
2203         struct ptlrpc_sec_policy *policy;
2204         unsigned int prealloc;
2205         ENTRY;
2206
2207         LASSERT(rs->rs_svc_ctx);
2208         LASSERT(rs->rs_svc_ctx->sc_policy);
2209
2210         policy = rs->rs_svc_ctx->sc_policy;
2211         LASSERT(policy->sp_sops->free_rs);
2212
2213         prealloc = rs->rs_prealloc;
2214         policy->sp_sops->free_rs(rs);
2215
2216         if (prealloc)
2217                 lustre_put_emerg_rs(rs);
2218         EXIT;
2219 }
2220
2221 void sptlrpc_svc_ctx_addref(struct ptlrpc_request *req)
2222 {
2223         struct ptlrpc_svc_ctx *ctx = req->rq_svc_ctx;
2224
2225         if (ctx != NULL)
2226                 atomic_inc(&ctx->sc_refcount);
2227 }
2228
2229 void sptlrpc_svc_ctx_decref(struct ptlrpc_request *req)
2230 {
2231         struct ptlrpc_svc_ctx *ctx = req->rq_svc_ctx;
2232
2233         if (ctx == NULL)
2234                 return;
2235
2236         LASSERT_ATOMIC_POS(&ctx->sc_refcount);
2237         if (atomic_dec_and_test(&ctx->sc_refcount)) {
2238                 if (ctx->sc_policy->sp_sops->free_ctx)
2239                         ctx->sc_policy->sp_sops->free_ctx(ctx);
2240         }
2241         req->rq_svc_ctx = NULL;
2242 }
2243
2244 void sptlrpc_svc_ctx_invalidate(struct ptlrpc_request *req)
2245 {
2246         struct ptlrpc_svc_ctx *ctx = req->rq_svc_ctx;
2247
2248         if (ctx == NULL)
2249                 return;
2250
2251         LASSERT_ATOMIC_POS(&ctx->sc_refcount);
2252         if (ctx->sc_policy->sp_sops->invalidate_ctx)
2253                 ctx->sc_policy->sp_sops->invalidate_ctx(ctx);
2254 }
2255 EXPORT_SYMBOL(sptlrpc_svc_ctx_invalidate);
2256
2257 /****************************************
2258  * bulk security                        *
2259  ****************************************/
2260
2261 /**
2262  * Perform transformation upon bulk data pointed by \a desc. This is called
2263  * before transforming the request message.
2264  */
2265 int sptlrpc_cli_wrap_bulk(struct ptlrpc_request *req,
2266                           struct ptlrpc_bulk_desc *desc)
2267 {
2268         struct ptlrpc_cli_ctx *ctx;
2269
2270         LASSERT(req->rq_bulk_read || req->rq_bulk_write);
2271
2272         if (!req->rq_pack_bulk)
2273                 return 0;
2274
2275         ctx = req->rq_cli_ctx;
2276         if (ctx->cc_ops->wrap_bulk)
2277                 return ctx->cc_ops->wrap_bulk(ctx, req, desc);
2278         return 0;
2279 }
2280 EXPORT_SYMBOL(sptlrpc_cli_wrap_bulk);
2281
2282 /**
2283  * This is called after unwrap the reply message.
2284  * return nob of actual plain text size received, or error code.
2285  */
2286 int sptlrpc_cli_unwrap_bulk_read(struct ptlrpc_request *req,
2287                                  struct ptlrpc_bulk_desc *desc,
2288                                  int nob)
2289 {
2290         struct ptlrpc_cli_ctx  *ctx;
2291         int                     rc;
2292
2293         LASSERT(req->rq_bulk_read && !req->rq_bulk_write);
2294
2295         if (!req->rq_pack_bulk)
2296                 return desc->bd_nob_transferred;
2297
2298         ctx = req->rq_cli_ctx;
2299         if (ctx->cc_ops->unwrap_bulk) {
2300                 rc = ctx->cc_ops->unwrap_bulk(ctx, req, desc);
2301                 if (rc < 0)
2302                         return rc;
2303         }
2304         return desc->bd_nob_transferred;
2305 }
2306 EXPORT_SYMBOL(sptlrpc_cli_unwrap_bulk_read);
2307
2308 /**
2309  * This is called after unwrap the reply message.
2310  * return 0 for success or error code.
2311  */
2312 int sptlrpc_cli_unwrap_bulk_write(struct ptlrpc_request *req,
2313                                   struct ptlrpc_bulk_desc *desc)
2314 {
2315         struct ptlrpc_cli_ctx  *ctx;
2316         int                     rc;
2317
2318         LASSERT(!req->rq_bulk_read && req->rq_bulk_write);
2319
2320         if (!req->rq_pack_bulk)
2321                 return 0;
2322
2323         ctx = req->rq_cli_ctx;
2324         if (ctx->cc_ops->unwrap_bulk) {
2325                 rc = ctx->cc_ops->unwrap_bulk(ctx, req, desc);
2326                 if (rc < 0)
2327                         return rc;
2328         }
2329
2330         /*
2331          * if everything is going right, nob should equals to nob_transferred.
2332          * in case of privacy mode, nob_transferred needs to be adjusted.
2333          */
2334         if (desc->bd_nob != desc->bd_nob_transferred) {
2335                 CERROR("nob %d doesn't match transferred nob %d\n",
2336                        desc->bd_nob, desc->bd_nob_transferred);
2337                 return -EPROTO;
2338         }
2339
2340         return 0;
2341 }
2342 EXPORT_SYMBOL(sptlrpc_cli_unwrap_bulk_write);
2343
2344 #ifdef HAVE_SERVER_SUPPORT
2345 /**
2346  * Performe transformation upon outgoing bulk read.
2347  */
2348 int sptlrpc_svc_wrap_bulk(struct ptlrpc_request *req,
2349                           struct ptlrpc_bulk_desc *desc)
2350 {
2351         struct ptlrpc_svc_ctx *ctx;
2352
2353         LASSERT(req->rq_bulk_read);
2354
2355         if (!req->rq_pack_bulk)
2356                 return 0;
2357
2358         ctx = req->rq_svc_ctx;
2359         if (ctx->sc_policy->sp_sops->wrap_bulk)
2360                 return ctx->sc_policy->sp_sops->wrap_bulk(req, desc);
2361
2362         return 0;
2363 }
2364 EXPORT_SYMBOL(sptlrpc_svc_wrap_bulk);
2365
2366 /**
2367  * Performe transformation upon incoming bulk write.
2368  */
2369 int sptlrpc_svc_unwrap_bulk(struct ptlrpc_request *req,
2370                             struct ptlrpc_bulk_desc *desc)
2371 {
2372         struct ptlrpc_svc_ctx *ctx;
2373         int                    rc;
2374
2375         LASSERT(req->rq_bulk_write);
2376
2377         /*
2378          * if it's in privacy mode, transferred should >= expected; otherwise
2379          * transferred should == expected.
2380          */
2381         if (desc->bd_nob_transferred < desc->bd_nob ||
2382             (desc->bd_nob_transferred > desc->bd_nob &&
2383              SPTLRPC_FLVR_BULK_SVC(req->rq_flvr.sf_rpc) !=
2384              SPTLRPC_BULK_SVC_PRIV)) {
2385                 DEBUG_REQ(D_ERROR, req, "truncated bulk GET %d(%d)",
2386                           desc->bd_nob_transferred, desc->bd_nob);
2387                 return -ETIMEDOUT;
2388         }
2389
2390         if (!req->rq_pack_bulk)
2391                 return 0;
2392
2393         ctx = req->rq_svc_ctx;
2394         if (ctx->sc_policy->sp_sops->unwrap_bulk) {
2395                 rc = ctx->sc_policy->sp_sops->unwrap_bulk(req, desc);
2396                 if (rc)
2397                         CERROR("error unwrap bulk: %d\n", rc);
2398         }
2399
2400         /* return 0 to allow reply be sent */
2401         return 0;
2402 }
2403 EXPORT_SYMBOL(sptlrpc_svc_unwrap_bulk);
2404
2405 /**
2406  * Prepare buffers for incoming bulk write.
2407  */
2408 int sptlrpc_svc_prep_bulk(struct ptlrpc_request *req,
2409                           struct ptlrpc_bulk_desc *desc)
2410 {
2411         struct ptlrpc_svc_ctx *ctx;
2412
2413         LASSERT(req->rq_bulk_write);
2414
2415         if (!req->rq_pack_bulk)
2416                 return 0;
2417
2418         ctx = req->rq_svc_ctx;
2419         if (ctx->sc_policy->sp_sops->prep_bulk)
2420                 return ctx->sc_policy->sp_sops->prep_bulk(req, desc);
2421
2422         return 0;
2423 }
2424 EXPORT_SYMBOL(sptlrpc_svc_prep_bulk);
2425
2426 #endif /* HAVE_SERVER_SUPPORT */
2427
2428 /****************************************
2429  * user descriptor helpers              *
2430  ****************************************/
2431
2432 int sptlrpc_current_user_desc_size(void)
2433 {
2434         int ngroups;
2435
2436         ngroups = current_ngroups;
2437
2438         if (ngroups > LUSTRE_MAX_GROUPS)
2439                 ngroups = LUSTRE_MAX_GROUPS;
2440         return sptlrpc_user_desc_size(ngroups);
2441 }
2442 EXPORT_SYMBOL(sptlrpc_current_user_desc_size);
2443
2444 int sptlrpc_pack_user_desc(struct lustre_msg *msg, int offset)
2445 {
2446         struct ptlrpc_user_desc *pud;
2447
2448         pud = lustre_msg_buf(msg, offset, 0);
2449
2450         pud->pud_uid = from_kuid(&init_user_ns, current_uid());
2451         pud->pud_gid = from_kgid(&init_user_ns, current_gid());
2452         pud->pud_fsuid = from_kuid(&init_user_ns, current_fsuid());
2453         pud->pud_fsgid = from_kgid(&init_user_ns, current_fsgid());
2454         pud->pud_cap = cfs_curproc_cap_pack();
2455         pud->pud_ngroups = (msg->lm_buflens[offset] - sizeof(*pud)) / 4;
2456
2457         task_lock(current);
2458         if (pud->pud_ngroups > current_ngroups)
2459                 pud->pud_ngroups = current_ngroups;
2460         memcpy(pud->pud_groups, current_cred()->group_info->blocks[0],
2461                pud->pud_ngroups * sizeof(__u32));
2462         task_unlock(current);
2463
2464         return 0;
2465 }
2466 EXPORT_SYMBOL(sptlrpc_pack_user_desc);
2467
2468 int sptlrpc_unpack_user_desc(struct lustre_msg *msg, int offset, int swabbed)
2469 {
2470         struct ptlrpc_user_desc *pud;
2471         int                      i;
2472
2473         pud = lustre_msg_buf(msg, offset, sizeof(*pud));
2474         if (!pud)
2475                 return -EINVAL;
2476
2477         if (swabbed) {
2478                 __swab32s(&pud->pud_uid);
2479                 __swab32s(&pud->pud_gid);
2480                 __swab32s(&pud->pud_fsuid);
2481                 __swab32s(&pud->pud_fsgid);
2482                 __swab32s(&pud->pud_cap);
2483                 __swab32s(&pud->pud_ngroups);
2484         }
2485
2486         if (pud->pud_ngroups > LUSTRE_MAX_GROUPS) {
2487                 CERROR("%u groups is too large\n", pud->pud_ngroups);
2488                 return -EINVAL;
2489         }
2490
2491         if (sizeof(*pud) + pud->pud_ngroups * sizeof(__u32) >
2492             msg->lm_buflens[offset]) {
2493                 CERROR("%u groups are claimed but bufsize only %u\n",
2494                        pud->pud_ngroups, msg->lm_buflens[offset]);
2495                 return -EINVAL;
2496         }
2497
2498         if (swabbed) {
2499                 for (i = 0; i < pud->pud_ngroups; i++)
2500                         __swab32s(&pud->pud_groups[i]);
2501         }
2502
2503         return 0;
2504 }
2505 EXPORT_SYMBOL(sptlrpc_unpack_user_desc);
2506
2507 /****************************************
2508  * misc helpers                         *
2509  ****************************************/
2510
2511 const char * sec2target_str(struct ptlrpc_sec *sec)
2512 {
2513         if (!sec || !sec->ps_import || !sec->ps_import->imp_obd)
2514                 return "*";
2515         if (sec_is_reverse(sec))
2516                 return "c";
2517         return obd_uuid2str(&sec->ps_import->imp_obd->u.cli.cl_target_uuid);
2518 }
2519 EXPORT_SYMBOL(sec2target_str);
2520
2521 /*
2522  * return true if the bulk data is protected
2523  */
2524 int sptlrpc_flavor_has_bulk(struct sptlrpc_flavor *flvr)
2525 {
2526         switch (SPTLRPC_FLVR_BULK_SVC(flvr->sf_rpc)) {
2527         case SPTLRPC_BULK_SVC_INTG:
2528         case SPTLRPC_BULK_SVC_PRIV:
2529                 return 1;
2530         default:
2531                 return 0;
2532         }
2533 }
2534 EXPORT_SYMBOL(sptlrpc_flavor_has_bulk);
2535
2536 /****************************************
2537  * crypto API helper/alloc blkciper     *
2538  ****************************************/
2539
2540 /****************************************
2541  * initialize/finalize                  *
2542  ****************************************/
2543
2544 int sptlrpc_init(void)
2545 {
2546         int rc;
2547
2548         rwlock_init(&policy_lock);
2549
2550         rc = sptlrpc_gc_init();
2551         if (rc)
2552                 goto out;
2553
2554         rc = sptlrpc_conf_init();
2555         if (rc)
2556                 goto out_gc;
2557
2558         rc = sptlrpc_enc_pool_init();
2559         if (rc)
2560                 goto out_conf;
2561
2562         rc = sptlrpc_null_init();
2563         if (rc)
2564                 goto out_pool;
2565
2566         rc = sptlrpc_plain_init();
2567         if (rc)
2568                 goto out_null;
2569
2570         rc = sptlrpc_lproc_init();
2571         if (rc)
2572                 goto out_plain;
2573
2574         return 0;
2575
2576 out_plain:
2577         sptlrpc_plain_fini();
2578 out_null:
2579         sptlrpc_null_fini();
2580 out_pool:
2581         sptlrpc_enc_pool_fini();
2582 out_conf:
2583         sptlrpc_conf_fini();
2584 out_gc:
2585         sptlrpc_gc_fini();
2586 out:
2587         return rc;
2588 }
2589
2590 void sptlrpc_fini(void)
2591 {
2592         sptlrpc_lproc_fini();
2593         sptlrpc_plain_fini();
2594         sptlrpc_null_fini();
2595         sptlrpc_enc_pool_fini();
2596         sptlrpc_conf_fini();
2597         sptlrpc_gc_fini();
2598 }