Subversion Repositories HelenOS

/*

 * 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    fat_ops.c

 * @brief   Implementation of VFS operations for the FAT file system server.

 */

#include "fat.h"

#include "fat_dentry.h"

#include "fat_fat.h"

#include "../../vfs/vfs.h"

#include <libfs.h>

#include <libblock.h>

#include <ipc/ipc.h>

#include <ipc/services.h>

#include <ipc/devmap.h>

#include <async.h>

#include <errno.h>

#include <string.h>

#include <byteorder.h>

#include <libadt/hash_table.h>

#include <libadt/list.h>

#include <assert.h>

#include <futex.h>

#include <sys/mman.h>

#include <align.h>

/** Futex protecting the list of cached free FAT nodes. */

static futex_t ffn_futex = FUTEX_INITIALIZER;

/** List of cached free FAT nodes. */

static LIST_INITIALIZE(ffn_head);

static void fat_node_initialize(fat_node_t *node)

{

    futex_initialize(&node->lock, 1);

    node->idx = NULL;

    node->type = 0;

    link_initialize(&node->ffn_link);

    node->size = 0;

    node->lnkcnt = 0;

    node->refcnt = 0;

    node->dirty = false;

}

static void fat_node_sync(fat_node_t *node)

{

    block_t *b;

    fat_bs_t *bs;

    fat_dentry_t *d;

    uint16_t bps;

    unsigned dps;

    assert(node->dirty);

    bs = block_bb_get(node->idx->dev_handle);

    bps = uint16_t_le2host(bs->bps);

    dps = bps / sizeof(fat_dentry_t);

    /* Read the block that contains the dentry of interest. */

    b = _fat_block_get(bs, node->idx->dev_handle, node->idx->pfc,

        (node->idx->pdi * sizeof(fat_dentry_t)) / bps, BLOCK_FLAGS_NONE);

    d = ((fat_dentry_t *)b->data) + (node->idx->pdi % dps);

    d->firstc = host2uint16_t_le(node->firstc);

    if (node->type == FAT_FILE)

        d->size = host2uint32_t_le(node->size);

    /* TODO: update other fields? (e.g time fields, attr field) */

    b->dirty = true;        /* need to sync block */

    block_put(b);

}

static fat_node_t *fat_node_get_new(void)

{

    fat_node_t *nodep;

    futex_down(&ffn_futex);

    if (!list_empty(&ffn_head)) {

        /* Try to use a cached free node structure. */

        fat_idx_t *idxp_tmp;

        nodep = list_get_instance(ffn_head.next, fat_node_t, ffn_link);

        if (futex_trydown(&nodep->lock) == ESYNCH_WOULD_BLOCK)

            goto skip_cache;

        idxp_tmp = nodep->idx;

        if (futex_trydown(&idxp_tmp->lock) == ESYNCH_WOULD_BLOCK) {

            futex_up(&nodep->lock);

            goto skip_cache;

        }

        list_remove(&nodep->ffn_link);

        futex_up(&ffn_futex);

        if (nodep->dirty)

            fat_node_sync(nodep);

        idxp_tmp->nodep = NULL;

        futex_up(&nodep->lock);

        futex_up(&idxp_tmp->lock);

    } else {

skip_cache:

        /* Try to allocate a new node structure. */

        futex_up(&ffn_futex);

        nodep = (fat_node_t *)malloc(sizeof(fat_node_t));

        if (!nodep)

            return NULL;

    }

    fat_node_initialize(nodep);

    return nodep;

}

/** Internal version of fat_node_get().

 *

 * @param idxp      Locked index structure.

 */

static void *fat_node_get_core(fat_idx_t *idxp)

{

    block_t *b;

    fat_bs_t *bs;

    fat_dentry_t *d;

    fat_node_t *nodep = NULL;

    unsigned bps;

    unsigned spc;

    unsigned dps;

    if (idxp->nodep) {

        /*

         * We are lucky.

         * The node is already instantiated in memory.

         */

        futex_down(&idxp->nodep->lock);

        if (!idxp->nodep->refcnt++)

            list_remove(&idxp->nodep->ffn_link);

        futex_up(&idxp->nodep->lock);

        return idxp->nodep;

    }

    /*

     * We must instantiate the node from the file system.

     */

    assert(idxp->pfc);

    nodep = fat_node_get_new();

    if (!nodep)

        return NULL;

    bs = block_bb_get(idxp->dev_handle);

    bps = uint16_t_le2host(bs->bps);

    spc = bs->spc;

    dps = bps / sizeof(fat_dentry_t);

    /* Read the block that contains the dentry of interest. */

    b = _fat_block_get(bs, idxp->dev_handle, idxp->pfc,

        (idxp->pdi * sizeof(fat_dentry_t)) / bps, BLOCK_FLAGS_NONE);

    assert(b);

    d = ((fat_dentry_t *)b->data) + (idxp->pdi % dps);

    if (d->attr & FAT_ATTR_SUBDIR) {

        /*

         * The only directory which does not have this bit set is the

         * root directory itself. The root directory node is handled

         * and initialized elsewhere.

         */

        nodep->type = FAT_DIRECTORY;

        /*

         * Unfortunately, the 'size' field of the FAT dentry is not

         * defined for the directory entry type. We must determine the

         * size of the directory by walking the FAT.

         */

        nodep->size = bps * spc * fat_clusters_get(bs, idxp->dev_handle,

            uint16_t_le2host(d->firstc));

    } else {

        nodep->type = FAT_FILE;

        nodep->size = uint32_t_le2host(d->size);

    }

    nodep->firstc = uint16_t_le2host(d->firstc);

    nodep->lnkcnt = 1;

    nodep->refcnt = 1;

    block_put(b);

    /* Link the idx structure with the node structure. */

    nodep->idx = idxp;

    idxp->nodep = nodep;

    return nodep;

}

/** Instantiate a FAT in-core node. */

static void *fat_node_get(dev_handle_t dev_handle, fs_index_t index)

{

    void *node;

    fat_idx_t *idxp;

    idxp = fat_idx_get_by_index(dev_handle, index);

    if (!idxp)

        return NULL;

    /* idxp->lock held */

    node = fat_node_get_core(idxp);

    futex_up(&idxp->lock);

    return node;

}

static void fat_node_put(void *node)

{

    fat_node_t *nodep = (fat_node_t *)node;

    bool destroy = false;

    futex_down(&nodep->lock);

    if (!--nodep->refcnt) {

        if (nodep->idx) {

            futex_down(&ffn_futex);

            list_append(&nodep->ffn_link, &ffn_head);

            futex_up(&ffn_futex);

        } else {

            /*

             * The node does not have any index structure associated

             * with itself. This can only mean that we are releasing

             * the node after a failed attempt to allocate the index

             * structure for it.

             */

            destroy = true;

        }

    }

    futex_up(&nodep->lock);

    if (destroy)

        free(node);

}

static void *fat_create_node(dev_handle_t dev_handle, int flags)

{

    fat_idx_t *idxp;

    fat_node_t *nodep;

    nodep = fat_node_get_new();

    if (!nodep)

        return NULL;

    idxp = fat_idx_get_new(dev_handle);

    if (!idxp) {

        fat_node_put(nodep);

        return NULL;

    }

    /* idxp->lock held */

    if (flags & L_DIRECTORY) {

        nodep->type = FAT_DIRECTORY;

    } else {

        nodep->type = FAT_FILE;

    }

    nodep->size = 0;

    nodep->firstc = FAT_CLST_RES0;

    nodep->lnkcnt = 0;  /* not linked anywhere */

    nodep->refcnt = 1;

    nodep->idx = idxp;

    idxp->nodep = nodep;

    futex_up(&idxp->lock);

    return nodep;

}

static int fat_destroy_node(void *node)

{

    return ENOTSUP; /* not supported at the moment */

}

static bool fat_link(void *prnt, void *chld, const char *name)

{

    return false;   /* not supported at the moment */

}

static int fat_unlink(void *prnt, void *chld)

{

    return ENOTSUP; /* not supported at the moment */

}

static void *fat_match(void *prnt, const char *component)

{

    fat_bs_t *bs;

    fat_node_t *parentp = (fat_node_t *)prnt;

    char name[FAT_NAME_LEN + 1 + FAT_EXT_LEN + 1];

    unsigned i, j;

    unsigned bps;       /* bytes per sector */

    unsigned dps;       /* dentries per sector */

    unsigned blocks;

    fat_dentry_t *d;

    block_t *b;

    futex_down(&parentp->idx->lock);

    bs = block_bb_get(parentp->idx->dev_handle);

    bps = uint16_t_le2host(bs->bps);

    dps = bps / sizeof(fat_dentry_t);

    blocks = parentp->size / bps;

    for (i = 0; i < blocks; i++) {

        b = fat_block_get(bs, parentp, i, BLOCK_FLAGS_NONE);

        for (j = 0; j < dps; j++) {

            d = ((fat_dentry_t *)b->data) + j;

            switch (fat_classify_dentry(d)) {

            case FAT_DENTRY_SKIP:

                continue;

            case FAT_DENTRY_LAST:

                block_put(b);

                futex_up(&parentp->idx->lock);

                return NULL;

            default:

            case FAT_DENTRY_VALID:

                dentry_name_canonify(d, name);

                break;

            }

            if (stricmp(name, component) == 0) {

                /* hit */

                void *node;

                /*

                 * Assume tree hierarchy for locking.  We

                 * already have the parent and now we are going

                 * to lock the child.  Never lock in the oposite

                 * order.

                 */

                fat_idx_t *idx = fat_idx_get_by_pos(

                    parentp->idx->dev_handle, parentp->firstc,

                    i * dps + j);

                futex_up(&parentp->idx->lock);

                if (!idx) {

                    /*

                     * Can happen if memory is low or if we

                     * run out of 32-bit indices.

                     */

                    block_put(b);

                    return NULL;

                }

                node = fat_node_get_core(idx);

                futex_up(&idx->lock);

                block_put(b);

                return node;

            }

        }

        block_put(b);

    }

    futex_up(&parentp->idx->lock);

    return NULL;

}

static fs_index_t fat_index_get(void *node)

{

    fat_node_t *fnodep = (fat_node_t *)node;

    if (!fnodep)

        return 0;

    return fnodep->idx->index;

}

static size_t fat_size_get(void *node)

{

    return ((fat_node_t *)node)->size;

}

static unsigned fat_lnkcnt_get(void *node)

{

    return ((fat_node_t *)node)->lnkcnt;

}

static bool fat_has_children(void *node)

{

    fat_bs_t *bs;

    fat_node_t *nodep = (fat_node_t *)node;

    unsigned bps;

    unsigned dps;

    unsigned blocks;

    block_t *b;

    unsigned i, j;

    if (nodep->type != FAT_DIRECTORY)

        return false;

    futex_down(&nodep->idx->lock);

    bs = block_bb_get(nodep->idx->dev_handle);

    bps = uint16_t_le2host(bs->bps);

    dps = bps / sizeof(fat_dentry_t);

    blocks = nodep->size / bps;

    for (i = 0; i < blocks; i++) {

        fat_dentry_t *d;

        b = fat_block_get(bs, nodep, i, BLOCK_FLAGS_NONE);

        for (j = 0; j < dps; j++) {

            d = ((fat_dentry_t *)b->data) + j;

            switch (fat_classify_dentry(d)) {

            case FAT_DENTRY_SKIP:

                continue;

            case FAT_DENTRY_LAST:

                block_put(b);

                futex_up(&nodep->idx->lock);

                return false;

            default:

            case FAT_DENTRY_VALID:

                block_put(b);

                futex_up(&nodep->idx->lock);

                return true;

            }

            block_put(b);

            futex_up(&nodep->idx->lock);

            return true;

        }

        block_put(b);

    }

    futex_up(&nodep->idx->lock);

    return false;

}

static void *fat_root_get(dev_handle_t dev_handle)

{

    return fat_node_get(dev_handle, 0);

}

static char fat_plb_get_char(unsigned pos)

{

    return fat_reg.plb_ro[pos % PLB_SIZE];

}

static bool fat_is_directory(void *node)

{

    return ((fat_node_t *)node)->type == FAT_DIRECTORY;

}

static bool fat_is_file(void *node)

{

    return ((fat_node_t *)node)->type == FAT_FILE;

}

/** libfs operations */

libfs_ops_t fat_libfs_ops = {

    .match = fat_match,

    .node_get = fat_node_get,

    .node_put = fat_node_put,

    .create = fat_create_node,

    .destroy = fat_destroy_node,

    .link = fat_link,

    .unlink = fat_unlink,

    .index_get = fat_index_get,

    .size_get = fat_size_get,

    .lnkcnt_get = fat_lnkcnt_get,

    .has_children = fat_has_children,

    .root_get = fat_root_get,

    .plb_get_char = fat_plb_get_char,

    .is_directory = fat_is_directory,

    .is_file = fat_is_file

};

void fat_mounted(ipc_callid_t rid, ipc_call_t *request)

{

    dev_handle_t dev_handle = (dev_handle_t) IPC_GET_ARG1(*request);

    fat_bs_t *bs;

    uint16_t bps;

    uint16_t rde;

    int rc;

    /* initialize libblock */

    rc = block_init(dev_handle, BS_SIZE);

    if (rc != EOK) {

        ipc_answer_0(rid, rc);

        return;

    }

    /* prepare the boot block */

    rc = block_bb_read(dev_handle, BS_BLOCK * BS_SIZE, BS_SIZE);

    if (rc != EOK) {

        block_fini(dev_handle);

        ipc_answer_0(rid, rc);

        return;

    }

    /* get the buffer with the boot sector */

    bs = block_bb_get(dev_handle);

    /* Read the number of root directory entries. */

    bps = uint16_t_le2host(bs->bps);

    rde = uint16_t_le2host(bs->root_ent_max);

    if (bps != BS_SIZE) {

        block_fini(dev_handle);

        ipc_answer_0(rid, ENOTSUP);

        return;

    }

    /* Initialize the block cache */

    rc = block_cache_init(dev_handle, bps, 0 /* XXX */);

    if (rc != EOK) {

        block_fini(dev_handle);

        ipc_answer_0(rid, rc);

        return;

    }

    rc = fat_idx_init_by_dev_handle(dev_handle);

    if (rc != EOK) {

        block_fini(dev_handle);

        ipc_answer_0(rid, rc);

        return;

    }

    /* Initialize the root node. */

    fat_node_t *rootp = (fat_node_t *)malloc(sizeof(fat_node_t));

    if (!rootp) {

        block_fini(dev_handle);

        fat_idx_fini_by_dev_handle(dev_handle);

        ipc_answer_0(rid, ENOMEM);

        return;

    }

    fat_node_initialize(rootp);

    fat_idx_t *ridxp = fat_idx_get_by_pos(dev_handle, FAT_CLST_ROOTPAR, 0);

    if (!ridxp) {

        block_fini(dev_handle);

        free(rootp);

        fat_idx_fini_by_dev_handle(dev_handle);

        ipc_answer_0(rid, ENOMEM);

        return;

    }

    assert(ridxp->index == 0);

    /* ridxp->lock held */

    rootp->type = FAT_DIRECTORY;

    rootp->firstc = FAT_CLST_ROOT;

    rootp->refcnt = 1;

    rootp->lnkcnt = 0;  /* FS root is not linked */

    rootp->size = rde * sizeof(fat_dentry_t);

    rootp->idx = ridxp;

    ridxp->nodep = rootp;

    futex_up(&ridxp->lock);

    ipc_answer_3(rid, EOK, ridxp->index, rootp->size, rootp->lnkcnt);

}

void fat_mount(ipc_callid_t rid, ipc_call_t *request)

{

    ipc_answer_0(rid, ENOTSUP);

}

void fat_lookup(ipc_callid_t rid, ipc_call_t *request)

{

    libfs_lookup(&fat_libfs_ops, fat_reg.fs_handle, rid, request);

}

void fat_read(ipc_callid_t rid, ipc_call_t *request)

{

    dev_handle_t dev_handle = (dev_handle_t)IPC_GET_ARG1(*request);

    fs_index_t index = (fs_index_t)IPC_GET_ARG2(*request);

    off_t pos = (off_t)IPC_GET_ARG3(*request);

    fat_node_t *nodep = (fat_node_t *)fat_node_get(dev_handle, index);

    fat_bs_t *bs;

    uint16_t bps;

    size_t bytes;

    block_t *b;

    if (!nodep) {

        ipc_answer_0(rid, ENOENT);

        return;

    }

    ipc_callid_t callid;

    size_t len;

    if (!ipc_data_read_receive(&callid, &len)) {

        fat_node_put(nodep);

        ipc_answer_0(callid, EINVAL);

        ipc_answer_0(rid, EINVAL);

        return;

    }

    bs = block_bb_get(dev_handle);

    bps = uint16_t_le2host(bs->bps);

    if (nodep->type == FAT_FILE) {

        /*

         * Our strategy for regular file reads is to read one block at

         * most and make use of the possibility to return less data than

         * requested. This keeps the code very simple.

         */

        if (pos >= nodep->size) {

            /* reading beyond the EOF */

            bytes = 0;

            (void) ipc_data_read_finalize(callid, NULL, 0);

        } else {

            bytes = min(len, bps - pos % bps);

            bytes = min(bytes, nodep->size - pos);

            b = fat_block_get(bs, nodep, pos / bps,

                BLOCK_FLAGS_NONE);

            (void) ipc_data_read_finalize(callid, b->data + pos % bps,

                bytes);

            block_put(b);

        }

    } else {

        unsigned bnum;

        off_t spos = pos;

        char name[FAT_NAME_LEN + 1 + FAT_EXT_LEN + 1];

        fat_dentry_t *d;

        assert(nodep->type == FAT_DIRECTORY);

        assert(nodep->size % bps == 0);

        assert(bps % sizeof(fat_dentry_t) == 0);

        /*

         * Our strategy for readdir() is to use the position pointer as

         * an index into the array of all dentries. On entry, it points

         * to the first unread dentry. If we skip any dentries, we bump

         * the position pointer accordingly.

         */

        bnum = (pos * sizeof(fat_dentry_t)) / bps;

        while (bnum < nodep->size / bps) {

            off_t o;

            b = fat_block_get(bs, nodep, bnum, BLOCK_FLAGS_NONE);

            for (o = pos % (bps / sizeof(fat_dentry_t));

                o < bps / sizeof(fat_dentry_t);

                o++, pos++) {

                d = ((fat_dentry_t *)b->data) + o;

                switch (fat_classify_dentry(d)) {

                case FAT_DENTRY_SKIP:

                    continue;

                case FAT_DENTRY_LAST:

                    block_put(b);

                    goto miss;

                default:

                case FAT_DENTRY_VALID:

                    dentry_name_canonify(d, name);

                    block_put(b);

                    goto hit;

                }

            }

            block_put(b);

            bnum++;

        }

miss:

        fat_node_put(nodep);

        ipc_answer_0(callid, ENOENT);

        ipc_answer_1(rid, ENOENT, 0);

        return;

hit:

        (void) ipc_data_read_finalize(callid, name, strlen(name) + 1);

        bytes = (pos - spos) + 1;

    }

    fat_node_put(nodep);

    ipc_answer_1(rid, EOK, (ipcarg_t)bytes);

}

void fat_write(ipc_callid_t rid, ipc_call_t *request)

{

    dev_handle_t dev_handle = (dev_handle_t)IPC_GET_ARG1(*request);

    fs_index_t index = (fs_index_t)IPC_GET_ARG2(*request);

    off_t pos = (off_t)IPC_GET_ARG3(*request);

    fat_node_t *nodep = (fat_node_t *)fat_node_get(dev_handle, index);

    fat_bs_t *bs;

    size_t bytes;

    block_t *b;

    uint16_t bps;

    unsigned spc;

    unsigned bpc;       /* bytes per cluster */

    off_t boundary;

    int flags = BLOCK_FLAGS_NONE;

    if (!nodep) {

        ipc_answer_0(rid, ENOENT);

        return;

    }

    ipc_callid_t callid;

    size_t len;

    if (!ipc_data_write_receive(&callid, &len)) {

        fat_node_put(nodep);

        ipc_answer_0(callid, EINVAL);

        ipc_answer_0(rid, EINVAL);

        return;

    }

    bs = block_bb_get(dev_handle);

    bps = uint16_t_le2host(bs->bps);

    spc = bs->spc;

    bpc = bps * spc;

    /*

     * In all scenarios, we will attempt to write out only one block worth

     * of data at maximum. There might be some more efficient approaches,

     * but this one greatly simplifies fat_write(). Note that we can afford

     * to do this because the client must be ready to handle the return

     * value signalizing a smaller number of bytes written.

     */

    bytes = min(len, bps - pos % bps);

    if (bytes == bps)

        flags |= BLOCK_FLAGS_NOREAD;

    boundary = ROUND_UP(nodep->size, bpc);

    if (pos < boundary) {

        /*

         * This is the easier case - we are either overwriting already

         * existing contents or writing behind the EOF, but still within

         * the limits of the last cluster. The node size may grow to the

         * next block size boundary.

         */

        fat_fill_gap(bs, nodep, FAT_CLST_RES0, pos);

        b = fat_block_get(bs, nodep, pos / bps, flags);

        (void) ipc_data_write_finalize(callid, b->data + pos % bps,

            bytes);

        b->dirty = true;        /* need to sync block */

        block_put(b);

        if (pos + bytes > nodep->size) {

            nodep->size = pos + bytes;

            nodep->dirty = true;    /* need to sync node */

        }

        ipc_answer_2(rid, EOK, bytes, nodep->size);

        fat_node_put(nodep);

        return;

    } else {

        /*

         * This is the more difficult case. We must allocate new

         * clusters for the node and zero them out.

         */

        int status;

        unsigned nclsts;

        fat_cluster_t mcl, lcl;

        nclsts = (ROUND_UP(pos + bytes, bpc) - boundary) / bpc;

        /* create an independent chain of nclsts clusters in all FATs */

        status = fat_alloc_clusters(bs, dev_handle, nclsts, &mcl, &lcl);

        if (status != EOK) {

            /* could not allocate a chain of nclsts clusters */

            fat_node_put(nodep);

            ipc_answer_0(callid, status);

            ipc_answer_0(rid, status);

            return;

        }

        /* zero fill any gaps */

        fat_fill_gap(bs, nodep, mcl, pos);

        b = _fat_block_get(bs, dev_handle, lcl, (pos / bps) % spc,

            flags);

        (void) ipc_data_write_finalize(callid, b->data + pos % bps,

            bytes);