source: mainline/uspace/srv/fs/exfat/exfat_idx.c@ fe61181

/*
 * Copyright (c) 2008 Jakub Jermar
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * - Redistributions of source code must retain the above copyright
 *   notice, this list of conditions and the following disclaimer.
 * - Redistributions in binary form must reproduce the above copyright
 *   notice, this list of conditions and the following disclaimer in the
 *   documentation and/or other materials provided with the distribution.
 * - The name of the author may not be used to endorse or promote products
 *   derived from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/** @addtogroup fs
 * @{
 */ 

/**
 * @file        exfat_idx.c
 * @brief       Layer for translating exFAT entities to VFS node indices.
 */

#include "exfat.h"
#include "../../vfs/vfs.h"
#include <errno.h>
#include <str.h>
#include <adt/hash_table.h>
#include <adt/list.h>
#include <assert.h>
#include <fibril_synch.h>
#include <malloc.h>

/** Each instance of this type describes one interval of freed VFS indices. */
typedef struct {
        link_t          link;
        fs_index_t      first;
        fs_index_t      last;
} freed_t;

/**
 * Each instance of this type describes state of all VFS indices that
 * are currently unused.
 */
typedef struct {
        link_t link;
        service_id_t service_id;

        /** Next unassigned index. */
        fs_index_t next;
        /** Number of remaining unassigned indices. */
        uint64_t remaining;

        /** Sorted list of intervals of freed indices. */
        list_t freed_list;
} unused_t;

/** Mutex protecting the list of unused structures. */
static FIBRIL_MUTEX_INITIALIZE(unused_lock);

/** List of unused structures. */
static LIST_INITIALIZE(unused_list);

static void unused_initialize(unused_t *u, service_id_t service_id)
{
        link_initialize(&u->link);
        u->service_id = service_id;
        u->next = 0;
        u->remaining = ((uint64_t)((fs_index_t)-1)) + 1;
        list_initialize(&u->freed_list);
}

static unused_t *unused_find(service_id_t service_id, bool lock)
{
        unused_t *u;

        if (lock)
                fibril_mutex_lock(&unused_lock);
        list_foreach(unused_list, l) {
                u = list_get_instance(l, unused_t, link);
                if (u->service_id == service_id) 
                        return u;
        }

        if (lock)
                fibril_mutex_unlock(&unused_lock);
        return NULL;
}

/** Mutex protecting the up_hash and ui_hash. */
static FIBRIL_MUTEX_INITIALIZE(used_lock);

/**
 * Global hash table of all used exfat_idx_t structures.
 * The index structures are hashed by the service_id, parent node's first
 * cluster and index within the parent directory.
 */ 
static hash_table_t up_hash;

#define UPH_BUCKETS_LOG 12
#define UPH_BUCKETS     (1 << UPH_BUCKETS_LOG)

#define UPH_SID_KEY     0
#define UPH_PFC_KEY     1
#define UPH_PDI_KEY     2

static hash_index_t pos_hash(unsigned long key[])
{
        service_id_t service_id = (service_id_t)key[UPH_SID_KEY];
        exfat_cluster_t pfc = (exfat_cluster_t)key[UPH_PFC_KEY];
        unsigned pdi = (unsigned)key[UPH_PDI_KEY];

        hash_index_t h;

        /*
         * The least significant half of all bits are the least significant bits
         * of the parent node's first cluster.
         *
         * The least significant half of the most significant half of all bits
         * are the least significant bits of the node's dentry index within the
         * parent directory node.
         *
         * The most significant half of the most significant half of all bits
         * are the least significant bits of the device handle.
         */
        h = pfc & ((1 << (UPH_BUCKETS_LOG / 2)) - 1);
        h |= (pdi & ((1 << (UPH_BUCKETS_LOG / 4)) - 1)) <<
            (UPH_BUCKETS_LOG / 2); 
        h |= (service_id & ((1 << (UPH_BUCKETS_LOG / 4)) - 1)) <<
            (3 * (UPH_BUCKETS_LOG / 4));

        return h;
}

static int pos_compare(unsigned long key[], hash_count_t keys, link_t *item)
{
        service_id_t service_id = (service_id_t)key[UPH_SID_KEY];
        exfat_cluster_t pfc;
        unsigned pdi;
        exfat_idx_t *fidx = list_get_instance(item, exfat_idx_t, uph_link);

        switch (keys) {
        case 1:
                return (service_id == fidx->service_id);
        case 3:
                pfc = (exfat_cluster_t) key[UPH_PFC_KEY];
                pdi = (unsigned) key[UPH_PDI_KEY];
                return (service_id == fidx->service_id) && (pfc == fidx->pfc) &&
                    (pdi == fidx->pdi);
        default:
                assert((keys == 1) || (keys == 3));
        }

        return 0;
}

static void pos_remove_callback(link_t *item)
{
        /* nothing to do */
}

static hash_table_operations_t uph_ops = {
        .hash = pos_hash,
        .compare = pos_compare,
        .remove_callback = pos_remove_callback,
};

/**
 * Global hash table of all used fat_idx_t structures.
 * The index structures are hashed by the service_id and index.
 */
static hash_table_t ui_hash;

#define UIH_BUCKETS_LOG 12
#define UIH_BUCKETS     (1 << UIH_BUCKETS_LOG)

#define UIH_SID_KEY     0
#define UIH_INDEX_KEY   1

static hash_index_t idx_hash(unsigned long key[])
{
        service_id_t service_id = (service_id_t)key[UIH_SID_KEY];
        fs_index_t index = (fs_index_t)key[UIH_INDEX_KEY];

        hash_index_t h;

        h = service_id & ((1 << (UIH_BUCKETS_LOG / 2)) - 1);
        h |= (index & ((1 << (UIH_BUCKETS_LOG / 2)) - 1)) <<
            (UIH_BUCKETS_LOG / 2);

        return h;
}

static int idx_compare(unsigned long key[], hash_count_t keys, link_t *item)
{
        service_id_t service_id = (service_id_t)key[UIH_SID_KEY];
        fs_index_t index;
        exfat_idx_t *fidx = list_get_instance(item, exfat_idx_t, uih_link);

        switch (keys) {
        case 1:
                return (service_id == fidx->service_id);
        case 2:
                index = (fs_index_t) key[UIH_INDEX_KEY];
                return (service_id == fidx->service_id) &&
                    (index == fidx->index);
        default:
                assert((keys == 1) || (keys == 2));
        }

        return 0;
}

static void idx_remove_callback(link_t *item)
{
        exfat_idx_t *fidx = list_get_instance(item, exfat_idx_t, uih_link);

        free(fidx);
}

static hash_table_operations_t uih_ops = {
        .hash = idx_hash,
        .compare = idx_compare,
        .remove_callback = idx_remove_callback,
};

/** Allocate a VFS index which is not currently in use. */
static bool exfat_index_alloc(service_id_t service_id, fs_index_t *index)
{
        unused_t *u;
        
        assert(index);
        u = unused_find(service_id, true);
        if (!u)
                return false;   

        if (list_empty(&u->freed_list)) {
                if (u->remaining) { 
                        /*
                         * There are no freed indices, allocate one directly
                         * from the counter.
                         */
                        *index = u->next++;
                        --u->remaining;
                        fibril_mutex_unlock(&unused_lock);
                        return true;
                }
        } else {
                /* There are some freed indices which we can reuse. */
                freed_t *f = list_get_instance(list_first(&u->freed_list),
                    freed_t, link);
                *index = f->first;
                if (f->first++ == f->last) {
                        /* Destroy the interval. */
                        list_remove(&f->link);
                        free(f);
                }
                fibril_mutex_unlock(&unused_lock);
                return true;
        }
        /*
         * We ran out of indices, which is extremely unlikely with FAT16, but
         * theoretically still possible (e.g. too many open unlinked nodes or
         * too many zero-sized nodes).
         */
        fibril_mutex_unlock(&unused_lock);
        return false;
}

/** If possible, coalesce two intervals of freed indices. */
static void try_coalesce_intervals(link_t *l, link_t *r, link_t *cur)
{
        freed_t *fl = list_get_instance(l, freed_t, link);
        freed_t *fr = list_get_instance(r, freed_t, link);

        if (fl->last + 1 == fr->first) {
                if (cur == l) {
                        fl->last = fr->last;
                        list_remove(r);
                        free(r);
                } else {
                        fr->first = fl->first;
                        list_remove(l);
                        free(l);
                }
        }
}

/** Free a VFS index, which is no longer in use. */
static void exfat_index_free(service_id_t service_id, fs_index_t index)
{
        unused_t *u;

        u = unused_find(service_id, true);
        assert(u);

        if (u->next == index + 1) {
                /* The index can be returned directly to the counter. */
                u->next--;
                u->remaining++;
        } else {
                /*
                 * The index must be returned either to an existing freed
                 * interval or a new interval must be created.
                 */
                link_t *lnk;
                freed_t *n;
                for (lnk = u->freed_list.head.next; lnk != &u->freed_list.head;
                    lnk = lnk->next) {
                        freed_t *f = list_get_instance(lnk, freed_t, link);
                        if (f->first == index + 1) {
                                f->first--;
                                if (lnk->prev != &u->freed_list.head)
                                        try_coalesce_intervals(lnk->prev, lnk,
                                            lnk);
                                fibril_mutex_unlock(&unused_lock);
                                return;
                        }
                        if (f->last == index - 1) {
                                f->last++;
                                if (lnk->next != &u->freed_list.head)
                                        try_coalesce_intervals(lnk, lnk->next,
                                            lnk);
                                fibril_mutex_unlock(&unused_lock);
                                return;
                        }
                        if (index > f->first) {
                                n = malloc(sizeof(freed_t));
                                /* TODO: sleep until allocation succeeds */
                                assert(n);
                                link_initialize(&n->link);
                                n->first = index;
                                n->last = index;
                                list_insert_before(&n->link, lnk);
                                fibril_mutex_unlock(&unused_lock);
                                return;
                        }

                }
                /* The index will form the last interval. */
                n = malloc(sizeof(freed_t));
                /* TODO: sleep until allocation succeeds */
                assert(n);
                link_initialize(&n->link);
                n->first = index;
                n->last = index;
                list_append(&n->link, &u->freed_list);
        }
        fibril_mutex_unlock(&unused_lock);
}

static int exfat_idx_create(exfat_idx_t **fidxp, service_id_t service_id)
{
        exfat_idx_t *fidx;

        fidx = (exfat_idx_t *) malloc(sizeof(exfat_idx_t));
        if (!fidx) 
                return ENOMEM;
        if (!exfat_index_alloc(service_id, &fidx->index)) {
                free(fidx);
                return ENOSPC;
        }
                
        link_initialize(&fidx->uph_link);
        link_initialize(&fidx->uih_link);
        fibril_mutex_initialize(&fidx->lock);
        fidx->service_id = service_id;
        fidx->pfc = 0;  /* no parent yet */
        fidx->pdi = 0;
        fidx->nodep = NULL;

        *fidxp = fidx;
        return EOK;
}

int exfat_idx_get_new(exfat_idx_t **fidxp, service_id_t service_id)
{
        exfat_idx_t *fidx;
        int rc;

        fibril_mutex_lock(&used_lock);
        rc = exfat_idx_create(&fidx, service_id);
        if (rc != EOK) {
                fibril_mutex_unlock(&used_lock);
                return rc;
        }
                
        unsigned long ikey[] = {
                [UIH_SID_KEY] = service_id,
                [UIH_INDEX_KEY] = fidx->index,
        };
        
        hash_table_insert(&ui_hash, ikey, &fidx->uih_link);
        fibril_mutex_lock(&fidx->lock);
        fibril_mutex_unlock(&used_lock);

        *fidxp = fidx;
        return EOK;
}

exfat_idx_t *
exfat_idx_get_by_pos(service_id_t service_id, exfat_cluster_t pfc, unsigned pdi)
{
        exfat_idx_t *fidx;
        link_t *l;
        unsigned long pkey[] = {
                [UPH_SID_KEY] = service_id,
                [UPH_PFC_KEY] = pfc,
                [UPH_PDI_KEY] = pdi,
        };

        fibril_mutex_lock(&used_lock);
        l = hash_table_find(&up_hash, pkey);
        if (l) {
                fidx = hash_table_get_instance(l, exfat_idx_t, uph_link);
        } else {
                int rc;

                rc = exfat_idx_create(&fidx, service_id);
                if (rc != EOK) {
                        fibril_mutex_unlock(&used_lock);
                        return NULL;
                }
                
                unsigned long ikey[] = {
                        [UIH_SID_KEY] = service_id,
                        [UIH_INDEX_KEY] = fidx->index,
                };
        
                fidx->pfc = pfc;
                fidx->pdi = pdi;

                hash_table_insert(&up_hash, pkey, &fidx->uph_link);
                hash_table_insert(&ui_hash, ikey, &fidx->uih_link);
        }
        fibril_mutex_lock(&fidx->lock);
        fibril_mutex_unlock(&used_lock);

        return fidx;
}

void exfat_idx_hashin(exfat_idx_t *idx)
{
        unsigned long pkey[] = {
                [UPH_SID_KEY] = idx->service_id,
                [UPH_PFC_KEY] = idx->pfc,
                [UPH_PDI_KEY] = idx->pdi,
        };

        fibril_mutex_lock(&used_lock);
        hash_table_insert(&up_hash, pkey, &idx->uph_link);
        fibril_mutex_unlock(&used_lock);
}

void exfat_idx_hashout(exfat_idx_t *idx)
{
        unsigned long pkey[] = {
                [UPH_SID_KEY] = idx->service_id,
                [UPH_PFC_KEY] = idx->pfc,
                [UPH_PDI_KEY] = idx->pdi,
        };

        fibril_mutex_lock(&used_lock);
        hash_table_remove(&up_hash, pkey, 3);
        fibril_mutex_unlock(&used_lock);
}

exfat_idx_t *
exfat_idx_get_by_index(service_id_t service_id, fs_index_t index)
{
        exfat_idx_t *fidx = NULL;
        link_t *l;
        unsigned long ikey[] = {
                [UIH_SID_KEY] = service_id,
                [UIH_INDEX_KEY] = index,
        };

        fibril_mutex_lock(&used_lock);
        l = hash_table_find(&ui_hash, ikey);
        if (l) {
                fidx = hash_table_get_instance(l, exfat_idx_t, uih_link);
                fibril_mutex_lock(&fidx->lock);
        }
        fibril_mutex_unlock(&used_lock);

        return fidx;
}

/** Destroy the index structure.
 *
 * @param idx           The index structure to be destroyed.
 */
void exfat_idx_destroy(exfat_idx_t *idx)
{
        unsigned long ikey[] = {
                [UIH_SID_KEY] = idx->service_id,
                [UIH_INDEX_KEY] = idx->index,
        };
        service_id_t service_id = idx->service_id;
        fs_index_t index = idx->index;

        /* TODO: assert(idx->pfc == FAT_CLST_RES0); */
        assert(idx->pfc == 0);

        fibril_mutex_lock(&used_lock);
        /*
         * Since we can only free unlinked nodes, the index structure is not
         * present in the position hash (uph). We therefore hash it out from
         * the index hash only.
         */
        hash_table_remove(&ui_hash, ikey, 2);
        fibril_mutex_unlock(&used_lock);
        /* Release the VFS index. */
        exfat_index_free(service_id, index);
        /* The index structure itself is freed in idx_remove_callback(). */
}

int exfat_idx_init(void)
{
        if (!hash_table_create(&up_hash, UPH_BUCKETS, 3, &uph_ops)) 
                return ENOMEM;
        if (!hash_table_create(&ui_hash, UIH_BUCKETS, 2, &uih_ops)) {
                hash_table_destroy(&up_hash);
                return ENOMEM;
        }
        return EOK;
}

void exfat_idx_fini(void)
{
        /* We assume the hash tables are empty. */
        hash_table_destroy(&up_hash);
        hash_table_destroy(&ui_hash);
}

int exfat_idx_init_by_service_id(service_id_t service_id)
{
        unused_t *u;
        int rc = EOK;

        u = (unused_t *) malloc(sizeof(unused_t));
        if (!u)
                return ENOMEM;
        unused_initialize(u, service_id);
        fibril_mutex_lock(&unused_lock);
        if (!unused_find(service_id, false)) {
                list_append(&u->link, &unused_list);
        } else {
                free(u);
                rc = EEXIST;
        }
        fibril_mutex_unlock(&unused_lock);
        return rc;
}

void exfat_idx_fini_by_service_id(service_id_t service_id)
{
        unsigned long ikey[] = {
                [UIH_SID_KEY] = service_id 
        };
        unsigned long pkey[] = {
                [UPH_SID_KEY] = service_id 
        };

        /*
         * Remove this instance's index structure from up_hash and ui_hash.
         * Process up_hash first and ui_hash second because the index structure
         * is actually removed in idx_remove_callback(). 
         */
        fibril_mutex_lock(&used_lock);
        hash_table_remove(&up_hash, pkey, 1);
        hash_table_remove(&ui_hash, ikey, 1);
        fibril_mutex_unlock(&used_lock);

        /*
         * Free the unused and freed structures for this instance.
         */
        unused_t *u = unused_find(service_id, true);
        assert(u);
        list_remove(&u->link);
        fibril_mutex_unlock(&unused_lock);

        while (!list_empty(&u->freed_list)) {
                freed_t *f;
                f = list_get_instance(list_first(&u->freed_list), freed_t, link);
                list_remove(&f->link);
                free(f);
        }
        free(u); 
}

/**
 * @}
 */