linux/net/sunrpc/cache.c
<<
>>
Prefs
   1/*
   2 * net/sunrpc/cache.c
   3 *
   4 * Generic code for various authentication-related caches
   5 * used by sunrpc clients and servers.
   6 *
   7 * Copyright (C) 2002 Neil Brown <neilb@cse.unsw.edu.au>
   8 *
   9 * Released under terms in GPL version 2.  See COPYING.
  10 *
  11 */
  12
  13#include <linux/types.h>
  14#include <linux/fs.h>
  15#include <linux/file.h>
  16#include <linux/slab.h>
  17#include <linux/signal.h>
  18#include <linux/sched.h>
  19#include <linux/kmod.h>
  20#include <linux/list.h>
  21#include <linux/module.h>
  22#include <linux/ctype.h>
  23#include <asm/uaccess.h>
  24#include <linux/poll.h>
  25#include <linux/seq_file.h>
  26#include <linux/proc_fs.h>
  27#include <linux/net.h>
  28#include <linux/workqueue.h>
  29#include <linux/mutex.h>
  30#include <asm/ioctls.h>
  31#include <linux/sunrpc/types.h>
  32#include <linux/sunrpc/cache.h>
  33#include <linux/sunrpc/stats.h>
  34
  35#define  RPCDBG_FACILITY RPCDBG_CACHE
  36
  37static int cache_defer_req(struct cache_req *req, struct cache_head *item);
  38static void cache_revisit_request(struct cache_head *item);
  39
  40static void cache_init(struct cache_head *h)
  41{
  42        time_t now = get_seconds();
  43        h->next = NULL;
  44        h->flags = 0;
  45        kref_init(&h->ref);
  46        h->expiry_time = now + CACHE_NEW_EXPIRY;
  47        h->last_refresh = now;
  48}
  49
  50struct cache_head *sunrpc_cache_lookup(struct cache_detail *detail,
  51                                       struct cache_head *key, int hash)
  52{
  53        struct cache_head **head,  **hp;
  54        struct cache_head *new = NULL;
  55
  56        head = &detail->hash_table[hash];
  57
  58        read_lock(&detail->hash_lock);
  59
  60        for (hp=head; *hp != NULL ; hp = &(*hp)->next) {
  61                struct cache_head *tmp = *hp;
  62                if (detail->match(tmp, key)) {
  63                        cache_get(tmp);
  64                        read_unlock(&detail->hash_lock);
  65                        return tmp;
  66                }
  67        }
  68        read_unlock(&detail->hash_lock);
  69        /* Didn't find anything, insert an empty entry */
  70
  71        new = detail->alloc();
  72        if (!new)
  73                return NULL;
  74        /* must fully initialise 'new', else
  75         * we might get lose if we need to
  76         * cache_put it soon.
  77         */
  78        cache_init(new);
  79        detail->init(new, key);
  80
  81        write_lock(&detail->hash_lock);
  82
  83        /* check if entry appeared while we slept */
  84        for (hp=head; *hp != NULL ; hp = &(*hp)->next) {
  85                struct cache_head *tmp = *hp;
  86                if (detail->match(tmp, key)) {
  87                        cache_get(tmp);
  88                        write_unlock(&detail->hash_lock);
  89                        cache_put(new, detail);
  90                        return tmp;
  91                }
  92        }
  93        new->next = *head;
  94        *head = new;
  95        detail->entries++;
  96        cache_get(new);
  97        write_unlock(&detail->hash_lock);
  98
  99        return new;
 100}
 101EXPORT_SYMBOL_GPL(sunrpc_cache_lookup);
 102
 103
 104static void queue_loose(struct cache_detail *detail, struct cache_head *ch);
 105
 106static int cache_fresh_locked(struct cache_head *head, time_t expiry)
 107{
 108        head->expiry_time = expiry;
 109        head->last_refresh = get_seconds();
 110        return !test_and_set_bit(CACHE_VALID, &head->flags);
 111}
 112
 113static void cache_fresh_unlocked(struct cache_head *head,
 114                        struct cache_detail *detail, int new)
 115{
 116        if (new)
 117                cache_revisit_request(head);
 118        if (test_and_clear_bit(CACHE_PENDING, &head->flags)) {
 119                cache_revisit_request(head);
 120                queue_loose(detail, head);
 121        }
 122}
 123
 124struct cache_head *sunrpc_cache_update(struct cache_detail *detail,
 125                                       struct cache_head *new, struct cache_head *old, int hash)
 126{
 127        /* The 'old' entry is to be replaced by 'new'.
 128         * If 'old' is not VALID, we update it directly,
 129         * otherwise we need to replace it
 130         */
 131        struct cache_head **head;
 132        struct cache_head *tmp;
 133        int is_new;
 134
 135        if (!test_bit(CACHE_VALID, &old->flags)) {
 136                write_lock(&detail->hash_lock);
 137                if (!test_bit(CACHE_VALID, &old->flags)) {
 138                        if (test_bit(CACHE_NEGATIVE, &new->flags))
 139                                set_bit(CACHE_NEGATIVE, &old->flags);
 140                        else
 141                                detail->update(old, new);
 142                        is_new = cache_fresh_locked(old, new->expiry_time);
 143                        write_unlock(&detail->hash_lock);
 144                        cache_fresh_unlocked(old, detail, is_new);
 145                        return old;
 146                }
 147                write_unlock(&detail->hash_lock);
 148        }
 149        /* We need to insert a new entry */
 150        tmp = detail->alloc();
 151        if (!tmp) {
 152                cache_put(old, detail);
 153                return NULL;
 154        }
 155        cache_init(tmp);
 156        detail->init(tmp, old);
 157        head = &detail->hash_table[hash];
 158
 159        write_lock(&detail->hash_lock);
 160        if (test_bit(CACHE_NEGATIVE, &new->flags))
 161                set_bit(CACHE_NEGATIVE, &tmp->flags);
 162        else
 163                detail->update(tmp, new);
 164        tmp->next = *head;
 165        *head = tmp;
 166        detail->entries++;
 167        cache_get(tmp);
 168        is_new = cache_fresh_locked(tmp, new->expiry_time);
 169        cache_fresh_locked(old, 0);
 170        write_unlock(&detail->hash_lock);
 171        cache_fresh_unlocked(tmp, detail, is_new);
 172        cache_fresh_unlocked(old, detail, 0);
 173        cache_put(old, detail);
 174        return tmp;
 175}
 176EXPORT_SYMBOL_GPL(sunrpc_cache_update);
 177
 178static int cache_make_upcall(struct cache_detail *detail, struct cache_head *h);
 179/*
 180 * This is the generic cache management routine for all
 181 * the authentication caches.
 182 * It checks the currency of a cache item and will (later)
 183 * initiate an upcall to fill it if needed.
 184 *
 185 *
 186 * Returns 0 if the cache_head can be used, or cache_puts it and returns
 187 * -EAGAIN if upcall is pending,
 188 * -ETIMEDOUT if upcall failed and should be retried,
 189 * -ENOENT if cache entry was negative
 190 */
 191int cache_check(struct cache_detail *detail,
 192                    struct cache_head *h, struct cache_req *rqstp)
 193{
 194        int rv;
 195        long refresh_age, age;
 196
 197        /* First decide return status as best we can */
 198        if (!test_bit(CACHE_VALID, &h->flags) ||
 199            h->expiry_time < get_seconds())
 200                rv = -EAGAIN;
 201        else if (detail->flush_time > h->last_refresh)
 202                rv = -EAGAIN;
 203        else {
 204                /* entry is valid */
 205                if (test_bit(CACHE_NEGATIVE, &h->flags))
 206                        rv = -ENOENT;
 207                else rv = 0;
 208        }
 209
 210        /* now see if we want to start an upcall */
 211        refresh_age = (h->expiry_time - h->last_refresh);
 212        age = get_seconds() - h->last_refresh;
 213
 214        if (rqstp == NULL) {
 215                if (rv == -EAGAIN)
 216                        rv = -ENOENT;
 217        } else if (rv == -EAGAIN || age > refresh_age/2) {
 218                dprintk("RPC:       Want update, refage=%ld, age=%ld\n",
 219                                refresh_age, age);
 220                if (!test_and_set_bit(CACHE_PENDING, &h->flags)) {
 221                        switch (cache_make_upcall(detail, h)) {
 222                        case -EINVAL:
 223                                clear_bit(CACHE_PENDING, &h->flags);
 224                                if (rv == -EAGAIN) {
 225                                        set_bit(CACHE_NEGATIVE, &h->flags);
 226                                        cache_fresh_unlocked(h, detail,
 227                                             cache_fresh_locked(h, get_seconds()+CACHE_NEW_EXPIRY));
 228                                        rv = -ENOENT;
 229                                }
 230                                break;
 231
 232                        case -EAGAIN:
 233                                clear_bit(CACHE_PENDING, &h->flags);
 234                                cache_revisit_request(h);
 235                                break;
 236                        }
 237                }
 238        }
 239
 240        if (rv == -EAGAIN)
 241                if (cache_defer_req(rqstp, h) != 0)
 242                        rv = -ETIMEDOUT;
 243
 244        if (rv)
 245                cache_put(h, detail);
 246        return rv;
 247}
 248EXPORT_SYMBOL_GPL(cache_check);
 249
 250/*
 251 * caches need to be periodically cleaned.
 252 * For this we maintain a list of cache_detail and
 253 * a current pointer into that list and into the table
 254 * for that entry.
 255 *
 256 * Each time clean_cache is called it finds the next non-empty entry
 257 * in the current table and walks the list in that entry
 258 * looking for entries that can be removed.
 259 *
 260 * An entry gets removed if:
 261 * - The expiry is before current time
 262 * - The last_refresh time is before the flush_time for that cache
 263 *
 264 * later we might drop old entries with non-NEVER expiry if that table
 265 * is getting 'full' for some definition of 'full'
 266 *
 267 * The question of "how often to scan a table" is an interesting one
 268 * and is answered in part by the use of the "nextcheck" field in the
 269 * cache_detail.
 270 * When a scan of a table begins, the nextcheck field is set to a time
 271 * that is well into the future.
 272 * While scanning, if an expiry time is found that is earlier than the
 273 * current nextcheck time, nextcheck is set to that expiry time.
 274 * If the flush_time is ever set to a time earlier than the nextcheck
 275 * time, the nextcheck time is then set to that flush_time.
 276 *
 277 * A table is then only scanned if the current time is at least
 278 * the nextcheck time.
 279 *
 280 */
 281
 282static LIST_HEAD(cache_list);
 283static DEFINE_SPINLOCK(cache_list_lock);
 284static struct cache_detail *current_detail;
 285static int current_index;
 286
 287static const struct file_operations cache_file_operations;
 288static const struct file_operations content_file_operations;
 289static const struct file_operations cache_flush_operations;
 290
 291static void do_cache_clean(struct work_struct *work);
 292static DECLARE_DELAYED_WORK(cache_cleaner, do_cache_clean);
 293
 294static void remove_cache_proc_entries(struct cache_detail *cd)
 295{
 296        if (cd->proc_ent == NULL)
 297                return;
 298        if (cd->flush_ent)
 299                remove_proc_entry("flush", cd->proc_ent);
 300        if (cd->channel_ent)
 301                remove_proc_entry("channel", cd->proc_ent);
 302        if (cd->content_ent)
 303                remove_proc_entry("content", cd->proc_ent);
 304        cd->proc_ent = NULL;
 305        remove_proc_entry(cd->name, proc_net_rpc);
 306}
 307
 308#ifdef CONFIG_PROC_FS
 309static int create_cache_proc_entries(struct cache_detail *cd)
 310{
 311        struct proc_dir_entry *p;
 312
 313        cd->proc_ent = proc_mkdir(cd->name, proc_net_rpc);
 314        if (cd->proc_ent == NULL)
 315                goto out_nomem;
 316        cd->channel_ent = cd->content_ent = NULL;
 317
 318        p = proc_create_data("flush", S_IFREG|S_IRUSR|S_IWUSR,
 319                             cd->proc_ent, &cache_flush_operations, cd);
 320        cd->flush_ent = p;
 321        if (p == NULL)
 322                goto out_nomem;
 323
 324        if (cd->cache_request || cd->cache_parse) {
 325                p = proc_create_data("channel", S_IFREG|S_IRUSR|S_IWUSR,
 326                                     cd->proc_ent, &cache_file_operations, cd);
 327                cd->channel_ent = p;
 328                if (p == NULL)
 329                        goto out_nomem;
 330        }
 331        if (cd->cache_show) {
 332                p = proc_create_data("content", S_IFREG|S_IRUSR|S_IWUSR,
 333                                cd->proc_ent, &content_file_operations, cd);
 334                cd->content_ent = p;
 335                if (p == NULL)
 336                        goto out_nomem;
 337        }
 338        return 0;
 339out_nomem:
 340        remove_cache_proc_entries(cd);
 341        return -ENOMEM;
 342}
 343#else /* CONFIG_PROC_FS */
 344static int create_cache_proc_entries(struct cache_detail *cd)
 345{
 346        return 0;
 347}
 348#endif
 349
 350int cache_register(struct cache_detail *cd)
 351{
 352        int ret;
 353
 354        ret = create_cache_proc_entries(cd);
 355        if (ret)
 356                return ret;
 357        rwlock_init(&cd->hash_lock);
 358        INIT_LIST_HEAD(&cd->queue);
 359        spin_lock(&cache_list_lock);
 360        cd->nextcheck = 0;
 361        cd->entries = 0;
 362        atomic_set(&cd->readers, 0);
 363        cd->last_close = 0;
 364        cd->last_warn = -1;
 365        list_add(&cd->others, &cache_list);
 366        spin_unlock(&cache_list_lock);
 367
 368        /* start the cleaning process */
 369        schedule_delayed_work(&cache_cleaner, 0);
 370        return 0;
 371}
 372EXPORT_SYMBOL_GPL(cache_register);
 373
 374void cache_unregister(struct cache_detail *cd)
 375{
 376        cache_purge(cd);
 377        spin_lock(&cache_list_lock);
 378        write_lock(&cd->hash_lock);
 379        if (cd->entries || atomic_read(&cd->inuse)) {
 380                write_unlock(&cd->hash_lock);
 381                spin_unlock(&cache_list_lock);
 382                goto out;
 383        }
 384        if (current_detail == cd)
 385                current_detail = NULL;
 386        list_del_init(&cd->others);
 387        write_unlock(&cd->hash_lock);
 388        spin_unlock(&cache_list_lock);
 389        remove_cache_proc_entries(cd);
 390        if (list_empty(&cache_list)) {
 391                /* module must be being unloaded so its safe to kill the worker */
 392                cancel_delayed_work_sync(&cache_cleaner);
 393        }
 394        return;
 395out:
 396        printk(KERN_ERR "nfsd: failed to unregister %s cache\n", cd->name);
 397}
 398EXPORT_SYMBOL_GPL(cache_unregister);
 399
 400/* clean cache tries to find something to clean
 401 * and cleans it.
 402 * It returns 1 if it cleaned something,
 403 *            0 if it didn't find anything this time
 404 *           -1 if it fell off the end of the list.
 405 */
 406static int cache_clean(void)
 407{
 408        int rv = 0;
 409        struct list_head *next;
 410
 411        spin_lock(&cache_list_lock);
 412
 413        /* find a suitable table if we don't already have one */
 414        while (current_detail == NULL ||
 415            current_index >= current_detail->hash_size) {
 416                if (current_detail)
 417                        next = current_detail->others.next;
 418                else
 419                        next = cache_list.next;
 420                if (next == &cache_list) {
 421                        current_detail = NULL;
 422                        spin_unlock(&cache_list_lock);
 423                        return -1;
 424                }
 425                current_detail = list_entry(next, struct cache_detail, others);
 426                if (current_detail->nextcheck > get_seconds())
 427                        current_index = current_detail->hash_size;
 428                else {
 429                        current_index = 0;
 430                        current_detail->nextcheck = get_seconds()+30*60;
 431                }
 432        }
 433
 434        /* find a non-empty bucket in the table */
 435        while (current_detail &&
 436               current_index < current_detail->hash_size &&
 437               current_detail->hash_table[current_index] == NULL)
 438                current_index++;
 439
 440        /* find a cleanable entry in the bucket and clean it, or set to next bucket */
 441
 442        if (current_detail && current_index < current_detail->hash_size) {
 443                struct cache_head *ch, **cp;
 444                struct cache_detail *d;
 445
 446                write_lock(&current_detail->hash_lock);
 447
 448                /* Ok, now to clean this strand */
 449
 450                cp = & current_detail->hash_table[current_index];
 451                ch = *cp;
 452                for (; ch; cp= & ch->next, ch= *cp) {
 453                        if (current_detail->nextcheck > ch->expiry_time)
 454                                current_detail->nextcheck = ch->expiry_time+1;
 455                        if (ch->expiry_time >= get_seconds()
 456                            && ch->last_refresh >= current_detail->flush_time
 457                                )
 458                                continue;
 459                        if (test_and_clear_bit(CACHE_PENDING, &ch->flags))
 460                                queue_loose(current_detail, ch);
 461
 462                        if (atomic_read(&ch->ref.refcount) == 1)
 463                                break;
 464                }
 465                if (ch) {
 466                        *cp = ch->next;
 467                        ch->next = NULL;
 468                        current_detail->entries--;
 469                        rv = 1;
 470                }
 471                write_unlock(&current_detail->hash_lock);
 472                d = current_detail;
 473                if (!ch)
 474                        current_index ++;
 475                spin_unlock(&cache_list_lock);
 476                if (ch)
 477                        cache_put(ch, d);
 478        } else
 479                spin_unlock(&cache_list_lock);
 480
 481        return rv;
 482}
 483
 484/*
 485 * We want to regularly clean the cache, so we need to schedule some work ...
 486 */
 487static void do_cache_clean(struct work_struct *work)
 488{
 489        int delay = 5;
 490        if (cache_clean() == -1)
 491                delay = 30*HZ;
 492
 493        if (list_empty(&cache_list))
 494                delay = 0;
 495
 496        if (delay)
 497                schedule_delayed_work(&cache_cleaner, delay);
 498}
 499
 500
 501/*
 502 * Clean all caches promptly.  This just calls cache_clean
 503 * repeatedly until we are sure that every cache has had a chance to
 504 * be fully cleaned
 505 */
 506void cache_flush(void)
 507{
 508        while (cache_clean() != -1)
 509                cond_resched();
 510        while (cache_clean() != -1)
 511                cond_resched();
 512}
 513EXPORT_SYMBOL_GPL(cache_flush);
 514
 515void cache_purge(struct cache_detail *detail)
 516{
 517        detail->flush_time = LONG_MAX;
 518        detail->nextcheck = get_seconds();
 519        cache_flush();
 520        detail->flush_time = 1;
 521}
 522EXPORT_SYMBOL_GPL(cache_purge);
 523
 524
 525/*
 526 * Deferral and Revisiting of Requests.
 527 *
 528 * If a cache lookup finds a pending entry, we
 529 * need to defer the request and revisit it later.
 530 * All deferred requests are stored in a hash table,
 531 * indexed by "struct cache_head *".
 532 * As it may be wasteful to store a whole request
 533 * structure, we allow the request to provide a
 534 * deferred form, which must contain a
 535 * 'struct cache_deferred_req'
 536 * This cache_deferred_req contains a method to allow
 537 * it to be revisited when cache info is available
 538 */
 539
 540#define DFR_HASHSIZE    (PAGE_SIZE/sizeof(struct list_head))
 541#define DFR_HASH(item)  ((((long)item)>>4 ^ (((long)item)>>13)) % DFR_HASHSIZE)
 542
 543#define DFR_MAX 300     /* ??? */
 544
 545static DEFINE_SPINLOCK(cache_defer_lock);
 546static LIST_HEAD(cache_defer_list);
 547static struct list_head cache_defer_hash[DFR_HASHSIZE];
 548static int cache_defer_cnt;
 549
 550static int cache_defer_req(struct cache_req *req, struct cache_head *item)
 551{
 552        struct cache_deferred_req *dreq;
 553        int hash = DFR_HASH(item);
 554
 555        if (cache_defer_cnt >= DFR_MAX) {
 556                /* too much in the cache, randomly drop this one,
 557                 * or continue and drop the oldest below
 558                 */
 559                if (net_random()&1)
 560                        return -ETIMEDOUT;
 561        }
 562        dreq = req->defer(req);
 563        if (dreq == NULL)
 564                return -ETIMEDOUT;
 565
 566        dreq->item = item;
 567
 568        spin_lock(&cache_defer_lock);
 569
 570        list_add(&dreq->recent, &cache_defer_list);
 571
 572        if (cache_defer_hash[hash].next == NULL)
 573                INIT_LIST_HEAD(&cache_defer_hash[hash]);
 574        list_add(&dreq->hash, &cache_defer_hash[hash]);
 575
 576        /* it is in, now maybe clean up */
 577        dreq = NULL;
 578        if (++cache_defer_cnt > DFR_MAX) {
 579                dreq = list_entry(cache_defer_list.prev,
 580                                  struct cache_deferred_req, recent);
 581                list_del(&dreq->recent);
 582                list_del(&dreq->hash);
 583                cache_defer_cnt--;
 584        }
 585        spin_unlock(&cache_defer_lock);
 586
 587        if (dreq) {
 588                /* there was one too many */
 589                dreq->revisit(dreq, 1);
 590        }
 591        if (!test_bit(CACHE_PENDING, &item->flags)) {
 592                /* must have just been validated... */
 593                cache_revisit_request(item);
 594        }
 595        return 0;
 596}
 597
 598static void cache_revisit_request(struct cache_head *item)
 599{
 600        struct cache_deferred_req *dreq;
 601        struct list_head pending;
 602
 603        struct list_head *lp;
 604        int hash = DFR_HASH(item);
 605
 606        INIT_LIST_HEAD(&pending);
 607        spin_lock(&cache_defer_lock);
 608
 609        lp = cache_defer_hash[hash].next;
 610        if (lp) {
 611                while (lp != &cache_defer_hash[hash]) {
 612                        dreq = list_entry(lp, struct cache_deferred_req, hash);
 613                        lp = lp->next;
 614                        if (dreq->item == item) {
 615                                list_del(&dreq->hash);
 616                                list_move(&dreq->recent, &pending);
 617                                cache_defer_cnt--;
 618                        }
 619                }
 620        }
 621        spin_unlock(&cache_defer_lock);
 622
 623        while (!list_empty(&pending)) {
 624                dreq = list_entry(pending.next, struct cache_deferred_req, recent);
 625                list_del_init(&dreq->recent);
 626                dreq->revisit(dreq, 0);
 627        }
 628}
 629
 630void cache_clean_deferred(void *owner)
 631{
 632        struct cache_deferred_req *dreq, *tmp;
 633        struct list_head pending;
 634
 635
 636        INIT_LIST_HEAD(&pending);
 637        spin_lock(&cache_defer_lock);
 638
 639        list_for_each_entry_safe(dreq, tmp, &cache_defer_list, recent) {
 640                if (dreq->owner == owner) {
 641                        list_del(&dreq->hash);
 642                        list_move(&dreq->recent, &pending);
 643                        cache_defer_cnt--;
 644                }
 645        }
 646        spin_unlock(&cache_defer_lock);
 647
 648        while (!list_empty(&pending)) {
 649                dreq = list_entry(pending.next, struct cache_deferred_req, recent);
 650                list_del_init(&dreq->recent);
 651                dreq->revisit(dreq, 1);
 652        }
 653}
 654
 655/*
 656 * communicate with user-space
 657 *
 658 * We have a magic /proc file - /proc/sunrpc/<cachename>/channel.
 659 * On read, you get a full request, or block.
 660 * On write, an update request is processed.
 661 * Poll works if anything to read, and always allows write.
 662 *
 663 * Implemented by linked list of requests.  Each open file has
 664 * a ->private that also exists in this list.  New requests are added
 665 * to the end and may wakeup and preceding readers.
 666 * New readers are added to the head.  If, on read, an item is found with
 667 * CACHE_UPCALLING clear, we free it from the list.
 668 *
 669 */
 670
 671static DEFINE_SPINLOCK(queue_lock);
 672static DEFINE_MUTEX(queue_io_mutex);
 673
 674struct cache_queue {
 675        struct list_head        list;
 676        int                     reader; /* if 0, then request */
 677};
 678struct cache_request {
 679        struct cache_queue      q;
 680        struct cache_head       *item;
 681        char                    * buf;
 682        int                     len;
 683        int                     readers;
 684};
 685struct cache_reader {
 686        struct cache_queue      q;
 687        int                     offset; /* if non-0, we have a refcnt on next request */
 688};
 689
 690static ssize_t
 691cache_read(struct file *filp, char __user *buf, size_t count, loff_t *ppos)
 692{
 693        struct cache_reader *rp = filp->private_data;
 694        struct cache_request *rq;
 695        struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
 696        int err;
 697
 698        if (count == 0)
 699                return 0;
 700
 701        mutex_lock(&queue_io_mutex); /* protect against multiple concurrent
 702                              * readers on this file */
 703 again:
 704        spin_lock(&queue_lock);
 705        /* need to find next request */
 706        while (rp->q.list.next != &cd->queue &&
 707               list_entry(rp->q.list.next, struct cache_queue, list)
 708               ->reader) {
 709                struct list_head *next = rp->q.list.next;
 710                list_move(&rp->q.list, next);
 711        }
 712        if (rp->q.list.next == &cd->queue) {
 713                spin_unlock(&queue_lock);
 714                mutex_unlock(&queue_io_mutex);
 715                BUG_ON(rp->offset);
 716                return 0;
 717        }
 718        rq = container_of(rp->q.list.next, struct cache_request, q.list);
 719        BUG_ON(rq->q.reader);
 720        if (rp->offset == 0)
 721                rq->readers++;
 722        spin_unlock(&queue_lock);
 723
 724        if (rp->offset == 0 && !test_bit(CACHE_PENDING, &rq->item->flags)) {
 725                err = -EAGAIN;
 726                spin_lock(&queue_lock);
 727                list_move(&rp->q.list, &rq->q.list);
 728                spin_unlock(&queue_lock);
 729        } else {
 730                if (rp->offset + count > rq->len)
 731                        count = rq->len - rp->offset;
 732                err = -EFAULT;
 733                if (copy_to_user(buf, rq->buf + rp->offset, count))
 734                        goto out;
 735                rp->offset += count;
 736                if (rp->offset >= rq->len) {
 737                        rp->offset = 0;
 738                        spin_lock(&queue_lock);
 739                        list_move(&rp->q.list, &rq->q.list);
 740                        spin_unlock(&queue_lock);
 741                }
 742                err = 0;
 743        }
 744 out:
 745        if (rp->offset == 0) {
 746                /* need to release rq */
 747                spin_lock(&queue_lock);
 748                rq->readers--;
 749                if (rq->readers == 0 &&
 750                    !test_bit(CACHE_PENDING, &rq->item->flags)) {
 751                        list_del(&rq->q.list);
 752                        spin_unlock(&queue_lock);
 753                        cache_put(rq->item, cd);
 754                        kfree(rq->buf);
 755                        kfree(rq);
 756                } else
 757                        spin_unlock(&queue_lock);
 758        }
 759        if (err == -EAGAIN)
 760                goto again;
 761        mutex_unlock(&queue_io_mutex);
 762        return err ? err :  count;
 763}
 764
 765static char write_buf[8192]; /* protected by queue_io_mutex */
 766
 767static ssize_t
 768cache_write(struct file *filp, const char __user *buf, size_t count,
 769            loff_t *ppos)
 770{
 771        int err;
 772        struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
 773
 774        if (count == 0)
 775                return 0;
 776        if (count >= sizeof(write_buf))
 777                return -EINVAL;
 778
 779        mutex_lock(&queue_io_mutex);
 780
 781        if (copy_from_user(write_buf, buf, count)) {
 782                mutex_unlock(&queue_io_mutex);
 783                return -EFAULT;
 784        }
 785        write_buf[count] = '\0';
 786        if (cd->cache_parse)
 787                err = cd->cache_parse(cd, write_buf, count);
 788        else
 789                err = -EINVAL;
 790
 791        mutex_unlock(&queue_io_mutex);
 792        return err ? err : count;
 793}
 794
 795static DECLARE_WAIT_QUEUE_HEAD(queue_wait);
 796
 797static unsigned int
 798cache_poll(struct file *filp, poll_table *wait)
 799{
 800        unsigned int mask;
 801        struct cache_reader *rp = filp->private_data;
 802        struct cache_queue *cq;
 803        struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
 804
 805        poll_wait(filp, &queue_wait, wait);
 806
 807        /* alway allow write */
 808        mask = POLL_OUT | POLLWRNORM;
 809
 810        if (!rp)
 811                return mask;
 812
 813        spin_lock(&queue_lock);
 814
 815        for (cq= &rp->q; &cq->list != &cd->queue;
 816             cq = list_entry(cq->list.next, struct cache_queue, list))
 817                if (!cq->reader) {
 818                        mask |= POLLIN | POLLRDNORM;
 819                        break;
 820                }
 821        spin_unlock(&queue_lock);
 822        return mask;
 823}
 824
 825static int
 826cache_ioctl(struct inode *ino, struct file *filp,
 827            unsigned int cmd, unsigned long arg)
 828{
 829        int len = 0;
 830        struct cache_reader *rp = filp->private_data;
 831        struct cache_queue *cq;
 832        struct cache_detail *cd = PDE(ino)->data;
 833
 834        if (cmd != FIONREAD || !rp)
 835                return -EINVAL;
 836
 837        spin_lock(&queue_lock);
 838
 839        /* only find the length remaining in current request,
 840         * or the length of the next request
 841         */
 842        for (cq= &rp->q; &cq->list != &cd->queue;
 843             cq = list_entry(cq->list.next, struct cache_queue, list))
 844                if (!cq->reader) {
 845                        struct cache_request *cr =
 846                                container_of(cq, struct cache_request, q);
 847                        len = cr->len - rp->offset;
 848                        break;
 849                }
 850        spin_unlock(&queue_lock);
 851
 852        return put_user(len, (int __user *)arg);
 853}
 854
 855static int
 856cache_open(struct inode *inode, struct file *filp)
 857{
 858        struct cache_reader *rp = NULL;
 859
 860        nonseekable_open(inode, filp);
 861        if (filp->f_mode & FMODE_READ) {
 862                struct cache_detail *cd = PDE(inode)->data;
 863
 864                rp = kmalloc(sizeof(*rp), GFP_KERNEL);
 865                if (!rp)
 866                        return -ENOMEM;
 867                rp->offset = 0;
 868                rp->q.reader = 1;
 869                atomic_inc(&cd->readers);
 870                spin_lock(&queue_lock);
 871                list_add(&rp->q.list, &cd->queue);
 872                spin_unlock(&queue_lock);
 873        }
 874        filp->private_data = rp;
 875        return 0;
 876}
 877
 878static int
 879cache_release(struct inode *inode, struct file *filp)
 880{
 881        struct cache_reader *rp = filp->private_data;
 882        struct cache_detail *cd = PDE(inode)->data;
 883
 884        if (rp) {
 885                spin_lock(&queue_lock);
 886                if (rp->offset) {
 887                        struct cache_queue *cq;
 888                        for (cq= &rp->q; &cq->list != &cd->queue;
 889                             cq = list_entry(cq->list.next, struct cache_queue, list))
 890                                if (!cq->reader) {
 891                                        container_of(cq, struct cache_request, q)
 892                                                ->readers--;
 893                                        break;
 894                                }
 895                        rp->offset = 0;
 896                }
 897                list_del(&rp->q.list);
 898                spin_unlock(&queue_lock);
 899
 900                filp->private_data = NULL;
 901                kfree(rp);
 902
 903                cd->last_close = get_seconds();
 904                atomic_dec(&cd->readers);
 905        }
 906        return 0;
 907}
 908
 909
 910
 911static const struct file_operations cache_file_operations = {
 912        .owner          = THIS_MODULE,
 913        .llseek         = no_llseek,
 914        .read           = cache_read,
 915        .write          = cache_write,
 916        .poll           = cache_poll,
 917        .ioctl          = cache_ioctl, /* for FIONREAD */
 918        .open           = cache_open,
 919        .release        = cache_release,
 920};
 921
 922
 923static void queue_loose(struct cache_detail *detail, struct cache_head *ch)
 924{
 925        struct cache_queue *cq;
 926        spin_lock(&queue_lock);
 927        list_for_each_entry(cq, &detail->queue, list)
 928                if (!cq->reader) {
 929                        struct cache_request *cr = container_of(cq, struct cache_request, q);
 930                        if (cr->item != ch)
 931                                continue;
 932                        if (cr->readers != 0)
 933                                continue;
 934                        list_del(&cr->q.list);
 935                        spin_unlock(&queue_lock);
 936                        cache_put(cr->item, detail);
 937                        kfree(cr->buf);
 938                        kfree(cr);
 939                        return;
 940                }
 941        spin_unlock(&queue_lock);
 942}
 943
 944/*
 945 * Support routines for text-based upcalls.
 946 * Fields are separated by spaces.
 947 * Fields are either mangled to quote space tab newline slosh with slosh
 948 * or a hexified with a leading \x
 949 * Record is terminated with newline.
 950 *
 951 */
 952
 953void qword_add(char **bpp, int *lp, char *str)
 954{
 955        char *bp = *bpp;
 956        int len = *lp;
 957        char c;
 958
 959        if (len < 0) return;
 960
 961        while ((c=*str++) && len)
 962                switch(c) {
 963                case ' ':
 964                case '\t':
 965                case '\n':
 966                case '\\':
 967                        if (len >= 4) {
 968                                *bp++ = '\\';
 969                                *bp++ = '0' + ((c & 0300)>>6);
 970                                *bp++ = '0' + ((c & 0070)>>3);
 971                                *bp++ = '0' + ((c & 0007)>>0);
 972                        }
 973                        len -= 4;
 974                        break;
 975                default:
 976                        *bp++ = c;
 977                        len--;
 978                }
 979        if (c || len <1) len = -1;
 980        else {
 981                *bp++ = ' ';
 982                len--;
 983        }
 984        *bpp = bp;
 985        *lp = len;
 986}
 987EXPORT_SYMBOL_GPL(qword_add);
 988
 989void qword_addhex(char **bpp, int *lp, char *buf, int blen)
 990{
 991        char *bp = *bpp;
 992        int len = *lp;
 993
 994        if (len < 0) return;
 995
 996        if (len > 2) {
 997                *bp++ = '\\';
 998                *bp++ = 'x';
 999                len -= 2;
1000                while (blen && len >= 2) {
1001                        unsigned char c = *buf++;
1002                        *bp++ = '0' + ((c&0xf0)>>4) + (c>=0xa0)*('a'-'9'-1);
1003                        *bp++ = '0' + (c&0x0f) + ((c&0x0f)>=0x0a)*('a'-'9'-1);
1004                        len -= 2;
1005                        blen--;
1006                }
1007        }
1008        if (blen || len<1) len = -1;
1009        else {
1010                *bp++ = ' ';
1011                len--;
1012        }
1013        *bpp = bp;
1014        *lp = len;
1015}
1016EXPORT_SYMBOL_GPL(qword_addhex);
1017
1018static void warn_no_listener(struct cache_detail *detail)
1019{
1020        if (detail->last_warn != detail->last_close) {
1021                detail->last_warn = detail->last_close;
1022                if (detail->warn_no_listener)
1023                        detail->warn_no_listener(detail);
1024        }
1025}
1026
1027/*
1028 * register an upcall request to user-space.
1029 * Each request is at most one page long.
1030 */
1031static int cache_make_upcall(struct cache_detail *detail, struct cache_head *h)
1032{
1033
1034        char *buf;
1035        struct cache_request *crq;
1036        char *bp;
1037        int len;
1038
1039        if (detail->cache_request == NULL)
1040                return -EINVAL;
1041
1042        if (atomic_read(&detail->readers) == 0 &&
1043            detail->last_close < get_seconds() - 30) {
1044                        warn_no_listener(detail);
1045                        return -EINVAL;
1046        }
1047
1048        buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
1049        if (!buf)
1050                return -EAGAIN;
1051
1052        crq = kmalloc(sizeof (*crq), GFP_KERNEL);
1053        if (!crq) {
1054                kfree(buf);
1055                return -EAGAIN;
1056        }
1057
1058        bp = buf; len = PAGE_SIZE;
1059
1060        detail->cache_request(detail, h, &bp, &len);
1061
1062        if (len < 0) {
1063                kfree(buf);
1064                kfree(crq);
1065                return -EAGAIN;
1066        }
1067        crq->q.reader = 0;
1068        crq->item = cache_get(h);
1069        crq->buf = buf;
1070        crq->len = PAGE_SIZE - len;
1071        crq->readers = 0;
1072        spin_lock(&queue_lock);
1073        list_add_tail(&crq->q.list, &detail->queue);
1074        spin_unlock(&queue_lock);
1075        wake_up(&queue_wait);
1076        return 0;
1077}
1078
1079/*
1080 * parse a message from user-space and pass it
1081 * to an appropriate cache
1082 * Messages are, like requests, separated into fields by
1083 * spaces and dequotes as \xHEXSTRING or embedded \nnn octal
1084 *
1085 * Message is
1086 *   reply cachename expiry key ... content....
1087 *
1088 * key and content are both parsed by cache
1089 */
1090
1091#define isodigit(c) (isdigit(c) && c <= '7')
1092int qword_get(char **bpp, char *dest, int bufsize)
1093{
1094        /* return bytes copied, or -1 on error */
1095        char *bp = *bpp;
1096        int len = 0;
1097
1098        while (*bp == ' ') bp++;
1099
1100        if (bp[0] == '\\' && bp[1] == 'x') {
1101                /* HEX STRING */
1102                bp += 2;
1103                while (isxdigit(bp[0]) && isxdigit(bp[1]) && len < bufsize) {
1104                        int byte = isdigit(*bp) ? *bp-'0' : toupper(*bp)-'A'+10;
1105                        bp++;
1106                        byte <<= 4;
1107                        byte |= isdigit(*bp) ? *bp-'0' : toupper(*bp)-'A'+10;
1108                        *dest++ = byte;
1109                        bp++;
1110                        len++;
1111                }
1112        } else {
1113                /* text with \nnn octal quoting */
1114                while (*bp != ' ' && *bp != '\n' && *bp && len < bufsize-1) {
1115                        if (*bp == '\\' &&
1116                            isodigit(bp[1]) && (bp[1] <= '3') &&
1117                            isodigit(bp[2]) &&
1118                            isodigit(bp[3])) {
1119                                int byte = (*++bp -'0');
1120                                bp++;
1121                                byte = (byte << 3) | (*bp++ - '0');
1122                                byte = (byte << 3) | (*bp++ - '0');
1123                                *dest++ = byte;
1124                                len++;
1125                        } else {
1126                                *dest++ = *bp++;
1127                                len++;
1128                        }
1129                }
1130        }
1131
1132        if (*bp != ' ' && *bp != '\n' && *bp != '\0')
1133                return -1;
1134        while (*bp == ' ') bp++;
1135        *bpp = bp;
1136        *dest = '\0';
1137        return len;
1138}
1139EXPORT_SYMBOL_GPL(qword_get);
1140
1141
1142/*
1143 * support /proc/sunrpc/cache/$CACHENAME/content
1144 * as a seqfile.
1145 * We call ->cache_show passing NULL for the item to
1146 * get a header, then pass each real item in the cache
1147 */
1148
1149struct handle {
1150        struct cache_detail *cd;
1151};
1152
1153static void *c_start(struct seq_file *m, loff_t *pos)
1154        __acquires(cd->hash_lock)
1155{
1156        loff_t n = *pos;
1157        unsigned hash, entry;
1158        struct cache_head *ch;
1159        struct cache_detail *cd = ((struct handle*)m->private)->cd;
1160
1161
1162        read_lock(&cd->hash_lock);
1163        if (!n--)
1164                return SEQ_START_TOKEN;
1165        hash = n >> 32;
1166        entry = n & ((1LL<<32) - 1);
1167
1168        for (ch=cd->hash_table[hash]; ch; ch=ch->next)
1169                if (!entry--)
1170                        return ch;
1171        n &= ~((1LL<<32) - 1);
1172        do {
1173                hash++;
1174                n += 1LL<<32;
1175        } while(hash < cd->hash_size &&
1176                cd->hash_table[hash]==NULL);
1177        if (hash >= cd->hash_size)
1178                return NULL;
1179        *pos = n+1;
1180        return cd->hash_table[hash];
1181}
1182
1183static void *c_next(struct seq_file *m, void *p, loff_t *pos)
1184{
1185        struct cache_head *ch = p;
1186        int hash = (*pos >> 32);
1187        struct cache_detail *cd = ((struct handle*)m->private)->cd;
1188
1189        if (p == SEQ_START_TOKEN)
1190                hash = 0;
1191        else if (ch->next == NULL) {
1192                hash++;
1193                *pos += 1LL<<32;
1194        } else {
1195                ++*pos;
1196                return ch->next;
1197        }
1198        *pos &= ~((1LL<<32) - 1);
1199        while (hash < cd->hash_size &&
1200               cd->hash_table[hash] == NULL) {
1201                hash++;
1202                *pos += 1LL<<32;
1203        }
1204        if (hash >= cd->hash_size)
1205                return NULL;
1206        ++*pos;
1207        return cd->hash_table[hash];
1208}
1209
1210static void c_stop(struct seq_file *m, void *p)
1211        __releases(cd->hash_lock)
1212{
1213        struct cache_detail *cd = ((struct handle*)m->private)->cd;
1214        read_unlock(&cd->hash_lock);
1215}
1216
1217static int c_show(struct seq_file *m, void *p)
1218{
1219        struct cache_head *cp = p;
1220        struct cache_detail *cd = ((struct handle*)m->private)->cd;
1221
1222        if (p == SEQ_START_TOKEN)
1223                return cd->cache_show(m, cd, NULL);
1224
1225        ifdebug(CACHE)
1226                seq_printf(m, "# expiry=%ld refcnt=%d flags=%lx\n",
1227                           cp->expiry_time, atomic_read(&cp->ref.refcount), cp->flags);
1228        cache_get(cp);
1229        if (cache_check(cd, cp, NULL))
1230                /* cache_check does a cache_put on failure */
1231                seq_printf(m, "# ");
1232        else
1233                cache_put(cp, cd);
1234
1235        return cd->cache_show(m, cd, cp);
1236}
1237
1238static const struct seq_operations cache_content_op = {
1239        .start  = c_start,
1240        .next   = c_next,
1241        .stop   = c_stop,
1242        .show   = c_show,
1243};
1244
1245static int content_open(struct inode *inode, struct file *file)
1246{
1247        struct handle *han;
1248        struct cache_detail *cd = PDE(inode)->data;
1249
1250        han = __seq_open_private(file, &cache_content_op, sizeof(*han));
1251        if (han == NULL)
1252                return -ENOMEM;
1253
1254        han->cd = cd;
1255        return 0;
1256}
1257
1258static const struct file_operations content_file_operations = {
1259        .open           = content_open,
1260        .read           = seq_read,
1261        .llseek         = seq_lseek,
1262        .release        = seq_release_private,
1263};
1264
1265static ssize_t read_flush(struct file *file, char __user *buf,
1266                            size_t count, loff_t *ppos)
1267{
1268        struct cache_detail *cd = PDE(file->f_path.dentry->d_inode)->data;
1269        char tbuf[20];
1270        unsigned long p = *ppos;
1271        size_t len;
1272
1273        sprintf(tbuf, "%lu\n", cd->flush_time);
1274        len = strlen(tbuf);
1275        if (p >= len)
1276                return 0;
1277        len -= p;
1278        if (len > count)
1279                len = count;
1280        if (copy_to_user(buf, (void*)(tbuf+p), len))
1281                return -EFAULT;
1282        *ppos += len;
1283        return len;
1284}
1285
1286static ssize_t write_flush(struct file * file, const char __user * buf,
1287                             size_t count, loff_t *ppos)
1288{
1289        struct cache_detail *cd = PDE(file->f_path.dentry->d_inode)->data;
1290        char tbuf[20];
1291        char *ep;
1292        long flushtime;
1293        if (*ppos || count > sizeof(tbuf)-1)
1294                return -EINVAL;
1295        if (copy_from_user(tbuf, buf, count))
1296                return -EFAULT;
1297        tbuf[count] = 0;
1298        flushtime = simple_strtoul(tbuf, &ep, 0);
1299        if (*ep && *ep != '\n')
1300                return -EINVAL;
1301
1302        cd->flush_time = flushtime;
1303        cd->nextcheck = get_seconds();
1304        cache_flush();
1305
1306        *ppos += count;
1307        return count;
1308}
1309
1310static const struct file_operations cache_flush_operations = {
1311        .open           = nonseekable_open,
1312        .read           = read_flush,
1313        .write          = write_flush,
1314};
1315