Add independent LZMS decompressor

[wimlib] / src / resource.c

/*
 * resource.c
 *
 * Read uncompressed and compressed metadata and file resources from a WIM file.
 */

/*
 * Copyright (C) 2012, 2013 Eric Biggers
 *
 * This file is part of wimlib, a library for working with WIM files.
 *
 * wimlib is free software; you can redistribute it and/or modify it under the
 * terms of the GNU General Public License as published by the Free Software
 * Foundation; either version 3 of the License, or (at your option) any later
 * version.
 *
 * wimlib is distributed in the hope that it will be useful, but WITHOUT ANY
 * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
 * A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License along with
 * wimlib; if not, see http://www.gnu.org/licenses/.
 */

#ifdef HAVE_CONFIG_H
#  include "config.h"
#endif

#include "wimlib.h"
#include "wimlib/endianness.h"
#include "wimlib/error.h"
#include "wimlib/file_io.h"
#include "wimlib/lookup_table.h"
#include "wimlib/resource.h"
#include "wimlib/sha1.h"

#ifdef __WIN32__
/* for read_win32_file_prefix(), read_win32_encrypted_file_prefix() */
#  include "wimlib/win32.h"
#endif

#ifdef WITH_NTFS_3G
/* for read_ntfs_file_prefix() */
#  include "wimlib/ntfs_3g.h"
#endif

#ifdef HAVE_ALLOCA_H
#  include <alloca.h>
#endif
#include <errno.h>
#include <fcntl.h>
#include <stdlib.h>
#include <unistd.h>

/*
 *                         Compressed WIM resources
 *
 * A compressed resource in a WIM consists of a number of compressed chunks,
 * each of which decompresses to a fixed chunk size (given in the WIM header;
 * usually 32768) except possibly the last, which always decompresses to any
 * remaining bytes.  In addition, immediately before the chunks, a table (the
 * "chunk table") provides the offset, in bytes relative to the end of the chunk
 * table, of the start of each compressed chunk, except for the first chunk
 * which is omitted as it always has an offset of 0.  Therefore, a compressed
 * resource with N chunks will have a chunk table with N - 1 entries.
 *
 * Additional information:
 *
 * - Entries in the chunk table are 4 bytes each, except if the uncompressed
 *   size of the resource is greater than 4 GiB, in which case the entries in
 *   the chunk table are 8 bytes each.  In either case, the entries are unsigned
 *   little-endian integers.
 *
 * - The chunk table is included in the compressed size of the resource provided
 *   in the corresponding entry in the WIM's stream lookup table.
 *
 * - The compressed size of a chunk is never greater than the uncompressed size.
 *   From the compressor's point of view, chunks that would have compressed to a
 *   size greater than or equal to their original size are in fact stored
 *   uncompressed.  From the decompresser's point of view, chunks with
 *   compressed size equal to their uncompressed size are in fact uncompressed.
 *
 * Furthermore, wimlib supports its own "pipable" WIM format, and for this the
 * structure of compressed resources was modified to allow piped reading and
 * writing.  To make sequential writing possible, the chunk table is placed
 * after the chunks rather than before the chunks, and to make sequential
 * reading possible, each chunk is prefixed with a 4-byte header giving its
 * compressed size as a 32-bit, unsigned, little-endian integer.  Otherwise the
 * details are the same.
 */


/* Decompress the specified chunk that uses the specified compression type
 * @ctype, part of a WIM with default chunk size @wim_chunk_size.  For LZX the
 * separate @wim_chunk_size is needed because it determines the window size used
 * for LZX compression.  */
static int
decompress(const void *cchunk, unsigned clen, void *uchunk, unsigned ulen,
           int ctype, u32 wim_chunk_size)
{
        switch (ctype) {
        case WIMLIB_COMPRESSION_TYPE_LZX:
                return wimlib_lzx_decompress2(cchunk, clen,
                                              uchunk, ulen, wim_chunk_size);
        case WIMLIB_COMPRESSION_TYPE_XPRESS:
                return wimlib_xpress_decompress(cchunk, clen,
                                                uchunk, ulen);
        case WIMLIB_COMPRESSION_TYPE_LZMS:
                return wimlib_lzms_decompress(cchunk, clen, uchunk, ulen);
        default:
                wimlib_assert(0);
                return -1;
        }
}

/* Read data from a compressed WIM resource.  Assumes parameters were already
 * verified by read_partial_wim_resource().  */
static int
read_compressed_wim_resource(const struct wim_resource_spec * const rspec,
                             const u64 size, const consume_data_callback_t cb,
                             const u32 cb_chunk_size, void * const ctx_or_buf,
                             const int flags, const u64 offset)
{
        int ret;
        int errno_save;

        const u32 orig_chunk_size = rspec->cchunk_size;
        const u32 orig_chunk_order = bsr32(orig_chunk_size);

        wimlib_assert(is_power_of_2(orig_chunk_size));

        /* Handle the trivial case.  */
        if (size == 0)
                return 0;

        if (rspec->ctype == WIMLIB_COMPRESSION_TYPE_LZMS) {
                /* TODO */

                unsigned clen = rspec->size_in_wim;
                unsigned ulen = rspec->uncompressed_size;
                unsigned lzms_offset;

                fprintf(stderr, "clen=%u, ulen=%u, offset=%lu\n", clen, ulen,
                        rspec->offset_in_wim);

                u8 *cbuf = MALLOC(clen);
                u8 *ubuf = MALLOC(ulen);


                ret = full_pread(&rspec->wim->in_fd,
                                 cbuf, clen, rspec->offset_in_wim);
                if (ret) {
                        ERROR_WITH_ERRNO("Can't read compressed data");
                        goto out_free_bufs;
                }

                if (clen <= rspec->cchunk_size)
                        lzms_offset = 0;
                else
                        lzms_offset = 20;

                ret = wimlib_lzms_decompress(cbuf + lzms_offset,
                                             clen - lzms_offset,
                                             ubuf, ulen);
                if (ret) {
                        ERROR("LZMS decompression error.");
                        errno = EINVAL;
                        ret = WIMLIB_ERR_DECOMPRESSION;
                        goto out_free_bufs;
                }
                if (cb) {
                        u32 chunk_size;
                        for (u64 i = offset; i < offset + size; i += chunk_size) {
                                chunk_size = min(offset + size - i, cb_chunk_size);
                                ret = cb(&ubuf[i], chunk_size, ctx_or_buf);
                                if (ret)
                                        goto out_free_bufs;
                        }
                } else {
                        memcpy(ctx_or_buf, &ubuf[offset], size);
                }
                ret = 0;
        out_free_bufs:
                FREE(ubuf);
                FREE(cbuf);
                return ret;
        }

        u64 *chunk_offsets = NULL;
        u8 *out_buf = NULL;
        u8 *tmp_buf = NULL;
        void *compressed_buf = NULL;
        bool chunk_offsets_malloced = false;
        bool out_buf_malloced = false;
        bool tmp_buf_malloced = false;
        bool compressed_buf_malloced = false;

        /* Get the file descriptor for the WIM.  */
        struct filedes * const in_fd = &rspec->wim->in_fd;

        /* Determine if we're reading a pipable resource from a pipe or not.  */
        const bool is_pipe_read = !filedes_is_seekable(in_fd);

        /* Calculate the number of chunks the resource is divided into.  */
        const u64 num_chunks = (rspec->uncompressed_size + orig_chunk_size - 1) >> orig_chunk_order;

        /* Calculate the 0-based index of the chunk at which the read starts.
         */
        const u64 start_chunk = offset >> orig_chunk_order;

        /* For pipe reads, we always must start from the 0th chunk.  */
        const u64 actual_start_chunk = (is_pipe_read ? 0 : start_chunk);

        /* Calculate the offset, within the start chunk, of the first byte of
         * the read.  */
        const u32 start_offset_in_chunk = offset & (orig_chunk_size - 1);

        /* Calculate the index of the chunk that contains the last byte of the
         * read.  */
        const u64 end_chunk = (offset + size - 1) >> orig_chunk_order;

        /* Calculate the offset, within the end chunk, of the last byte of the
         * read.  */
        const u32 end_offset_in_chunk = (offset + size - 1) & (orig_chunk_size - 1);

        /* Calculate the number of entries in the chunk table; it's one less
         * than the number of chunks, since the first chunk has no entry.  */
        const u64 num_chunk_entries = num_chunks - 1;

        /* Set the size of each chunk table entry based on the resource's
         * uncompressed size.  */
        const u64 chunk_entry_size = (rspec->uncompressed_size > (1ULL << 32)) ? 8 : 4;

        /* Calculate the size, in bytes, of the full chunk table.  */
        const u64 chunk_table_size = num_chunk_entries * chunk_entry_size;

        /* Current offset to read from.  */
        u64 cur_read_offset = rspec->offset_in_wim;
        if (!is_pipe_read) {
                /* Read the chunk table into memory.  */

                /* Calculate the number of chunk entries are actually needed to
                 * read the requested part of the resource.  Include an entry
                 * for the first chunk even though that doesn't exist in the
                 * on-disk table, but take into account that if the last chunk
                 * required for the read is not the last chunk of the resource,
                 * an extra chunk entry is needed so that the compressed size of
                 * the last chunk of the read can be determined.  */
                const u64 num_alloc_chunk_entries = end_chunk - start_chunk +
                                                    1 + (end_chunk != num_chunks - 1);

                /* Allocate a buffer to hold a subset of the chunk table.  It
                 * will only contain offsets for the chunks that are actually
                 * needed for this read.  For speed, allocate the buffer on the
                 * stack unless it's too large.  */
                if ((size_t)(num_alloc_chunk_entries * sizeof(u64)) !=
                            (num_alloc_chunk_entries * sizeof(u64)))
                        goto oom;

                if (num_alloc_chunk_entries <= STACK_MAX / sizeof(u64)) {
                        chunk_offsets = alloca(num_alloc_chunk_entries * sizeof(u64));
                } else {
                        chunk_offsets = MALLOC(num_alloc_chunk_entries * sizeof(u64));
                        if (chunk_offsets == NULL)
                                goto oom;
                        chunk_offsets_malloced = true;
                }

                /* Set the implicit offset of the first chunk if it's included
                 * in the needed chunks.  */
                if (start_chunk == 0)
                        chunk_offsets[0] = 0;

                /* Calculate the index of the first needed entry in the chunk
                 * table.  */
                const u64 start_table_idx = (start_chunk == 0) ?
                                0 : start_chunk - 1;

                /* Calculate the number of entries that need to be read from the
                 * chunk table.  */
                const u64 num_needed_chunk_entries = (start_chunk == 0) ?
                                num_alloc_chunk_entries - 1 : num_alloc_chunk_entries;

                /* Calculate the number of bytes of data that need to be read
                 * from the chunk table.  */
                const size_t chunk_table_needed_size =
                                num_needed_chunk_entries * chunk_entry_size;

                /* Calculate the byte offset, in the WIM file, of the first
                 * chunk table entry to read.  Take into account that if the WIM
                 * file is in the special "pipable" format, then the chunk table
                 * is at the end of the resource, not the beginning.  */
                const u64 file_offset_of_needed_chunk_entries =
                        rspec->offset_in_wim
                        + (start_table_idx * chunk_entry_size)
                        + (rspec->is_pipable ? (rspec->size_in_wim - chunk_table_size) : 0);

                /* Read the needed chunk table entries into the end of the
                 * chunk_offsets buffer.  */
                void * const chunk_tab_data = (u8*)&chunk_offsets[num_alloc_chunk_entries] -
                                              chunk_table_needed_size;
                ret = full_pread(in_fd, chunk_tab_data, chunk_table_needed_size,
                                 file_offset_of_needed_chunk_entries);
                if (ret)
                        goto read_error;

                /* Now fill in chunk_offsets from the entries we have read in
                 * chunk_tab_data.  Careful: chunk_offsets aliases
                 * chunk_tab_data, which breaks C's aliasing rules when we read
                 * 32-bit integers and store 64-bit integers.  But since the
                 * operations are safe as long as the compiler doesn't mess with
                 * their order, we use the gcc may_alias extension to tell the
                 * compiler that loads from the 32-bit integers may alias stores
                 * to the 64-bit integers.  */
                {
                        typedef le64 __attribute__((may_alias)) aliased_le64_t;
                        typedef le32 __attribute__((may_alias)) aliased_le32_t;
                        u64 * const chunk_offsets_p = chunk_offsets + (start_chunk == 0);
                        u64 i;

                        if (chunk_entry_size == 4) {
                                aliased_le32_t *raw_entries = (aliased_le32_t*)chunk_tab_data;
                                for (i = 0; i < num_needed_chunk_entries; i++)
                                        chunk_offsets_p[i] = le32_to_cpu(raw_entries[i]);
                        } else {
                                aliased_le64_t *raw_entries = (aliased_le64_t*)chunk_tab_data;
                                for (i = 0; i < num_needed_chunk_entries; i++)
                                        chunk_offsets_p[i] = le64_to_cpu(raw_entries[i]);
                        }
                }

                /* Set offset to beginning of first chunk to read.  */
                cur_read_offset += chunk_offsets[0];
                if (rspec->is_pipable)
                        cur_read_offset += start_chunk * sizeof(struct pwm_chunk_hdr);
                else
                        cur_read_offset += chunk_table_size;
        }

        /* If using a callback function, allocate a temporary buffer that will
         * hold data being passed to it.  If writing directly to a buffer
         * instead, arrange to write data directly into it.  */
        size_t out_buf_size;
        u8 *out_buf_end, *out_p;
        if (cb) {
                out_buf_size = max(cb_chunk_size, orig_chunk_size);
                if (out_buf_size <= STACK_MAX) {
                        out_buf = alloca(out_buf_size);
                } else {
                        out_buf = MALLOC(out_buf_size);
                        if (out_buf == NULL)
                                goto oom;
                        out_buf_malloced = true;
                }
        } else {
                out_buf_size = size;
                out_buf = ctx_or_buf;
        }
        out_buf_end = out_buf + out_buf_size;
        out_p = out_buf;

        /* Unless the raw compressed data was requested, allocate a temporary
         * buffer for reading compressed chunks, each of which can be at most
         * @orig_chunk_size - 1 bytes.  This excludes compressed chunks that are
         * a full @orig_chunk_size bytes, which are actually stored
         * uncompressed.  */
        if (!(flags & WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS)) {
                if (orig_chunk_size - 1 <= STACK_MAX) {
                        compressed_buf = alloca(orig_chunk_size - 1);
                } else {
                        compressed_buf = MALLOC(orig_chunk_size - 1);
                        if (compressed_buf == NULL)
                                goto oom;
                        compressed_buf_malloced = true;
                }
        }

        /* Allocate yet another temporary buffer, this one for decompressing
         * chunks for which only part of the data is needed.  */
        if (start_offset_in_chunk != 0 ||
            (end_offset_in_chunk != orig_chunk_size - 1 &&
             offset + size != rspec->uncompressed_size))
        {
                if (orig_chunk_size <= STACK_MAX) {
                        tmp_buf = alloca(orig_chunk_size);
                } else {
                        tmp_buf = MALLOC(orig_chunk_size);
                        if (tmp_buf == NULL)
                                goto oom;
                        tmp_buf_malloced = true;
                }
        }

        /* Read, and possibly decompress, each needed chunk, either writing the
         * data directly into the @ctx_or_buf buffer or passing it to the @cb
         * callback function.  */
        for (u64 i = actual_start_chunk; i <= end_chunk; i++) {

                /* Calculate uncompressed size of next chunk.  */
                u32 chunk_usize;
                if ((i == num_chunks - 1) && (rspec->uncompressed_size & (orig_chunk_size - 1)))
                        chunk_usize = (rspec->uncompressed_size & (orig_chunk_size - 1));
                else
                        chunk_usize = orig_chunk_size;

                /* Calculate compressed size of next chunk.  */
                u32 chunk_csize;
                if (is_pipe_read) {
                        struct pwm_chunk_hdr chunk_hdr;

                        ret = full_pread(in_fd, &chunk_hdr,
                                         sizeof(chunk_hdr), cur_read_offset);
                        if (ret)
                                goto read_error;
                        chunk_csize = le32_to_cpu(chunk_hdr.compressed_size);
                } else {
                        if (i == num_chunks - 1) {
                                chunk_csize = rspec->size_in_wim -
                                              chunk_table_size -
                                              chunk_offsets[i - start_chunk];
                                if (rspec->is_pipable)
                                        chunk_csize -= num_chunks * sizeof(struct pwm_chunk_hdr);
                        } else {
                                chunk_csize = chunk_offsets[i + 1 - start_chunk] -
                                              chunk_offsets[i - start_chunk];
                        }
                }
                if (chunk_csize == 0 || chunk_csize > chunk_usize) {
                        ERROR("Invalid chunk size in compressed resource!");
                        errno = EINVAL;
                        ret = WIMLIB_ERR_DECOMPRESSION;
                        goto out_free_memory;
                }
                if (rspec->is_pipable)
                        cur_read_offset += sizeof(struct pwm_chunk_hdr);

                if (i >= start_chunk) {
                        /* Calculate how much of this chunk needs to be read.  */
                        u32 chunk_needed_size;
                        u32 start_offset = 0;
                        u32 end_offset = orig_chunk_size - 1;

                        if (flags & WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS) {
                                chunk_needed_size = chunk_csize;
                        } else {
                                if (i == start_chunk)
                                        start_offset = start_offset_in_chunk;

                                if (i == end_chunk)
                                        end_offset = end_offset_in_chunk;

                                chunk_needed_size = end_offset + 1 - start_offset;
                        }

                        if (chunk_csize == chunk_usize ||
                            (flags & WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS))
                        {
                                /* Read the raw chunk data.  */

                                ret = full_pread(in_fd,
                                                 out_p,
                                                 chunk_needed_size,
                                                 cur_read_offset + start_offset);
                                if (ret)
                                        goto read_error;
                        } else {
                                /* Read and decompress the chunk.  */

                                u8 *target;

                                ret = full_pread(in_fd,
                                                 compressed_buf,
                                                 chunk_csize,
                                                 cur_read_offset);
                                if (ret)
                                        goto read_error;

                                if (chunk_needed_size == chunk_usize)
                                        target = out_p;
                                else
                                        target = tmp_buf;

                                ret = decompress(compressed_buf,
                                                 chunk_csize,
                                                 target,
                                                 chunk_usize,
                                                 rspec->ctype,
                                                 orig_chunk_size);
                                if (ret) {
                                        ERROR("Failed to decompress data!");
                                        ret = WIMLIB_ERR_DECOMPRESSION;
                                        errno = EINVAL;
                                        goto out_free_memory;
                                }
                                if (chunk_needed_size != chunk_usize)
                                        memcpy(out_p, tmp_buf + start_offset,
                                               chunk_needed_size);
                        }

                        out_p += chunk_needed_size;

                        if (cb) {
                                /* Feed the data to the callback function.  */

                                if (flags & WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS) {
                                        ret = cb(out_buf, out_p - out_buf, ctx_or_buf);
                                        if (ret)
                                                goto out_free_memory;
                                        out_p = out_buf;
                                } else if (i == end_chunk || out_p == out_buf_end) {
                                        size_t bytes_sent;
                                        const u8 *p;

                                        for (p = out_buf; p != out_p; p += bytes_sent) {
                                                bytes_sent = min(cb_chunk_size, out_p - p);
                                                ret = cb(p, bytes_sent, ctx_or_buf);
                                                if (ret)
                                                        goto out_free_memory;
                                        }
                                        out_p = out_buf;
                                }
                        }
                        cur_read_offset += chunk_csize;
                } else {
                        u8 dummy;

                        /* Skip data only.  */
                        cur_read_offset += chunk_csize;
                        ret = full_pread(in_fd, &dummy, 1, cur_read_offset - 1);
                        if (ret)
                                goto read_error;
                }
        }

        if (is_pipe_read
            && size == rspec->uncompressed_size
            && chunk_table_size)
        {
                u8 dummy;
                /* Skip chunk table at end of pipable resource.  */

                cur_read_offset += chunk_table_size;
                ret = full_pread(in_fd, &dummy, 1, cur_read_offset - 1);
                if (ret)
                        goto read_error;
        }
        ret = 0;
out_free_memory:
        errno_save = errno;
        if (chunk_offsets_malloced)
                FREE(chunk_offsets);
        if (out_buf_malloced)
                FREE(out_buf);
        if (compressed_buf_malloced)
                FREE(compressed_buf);
        if (tmp_buf_malloced)
                FREE(tmp_buf);
        errno = errno_save;
        return ret;

oom:
        ERROR("Not enough memory available to read size=%"PRIu64" bytes "
              "from compressed resource!", size);
        errno = ENOMEM;
        ret = WIMLIB_ERR_NOMEM;
        goto out_free_memory;

read_error:
        ERROR_WITH_ERRNO("Error reading compressed file resource!");
        goto out_free_memory;
}

/* Read raw data from a file descriptor at the specified offset.  */
static int
read_raw_file_data(struct filedes *in_fd, u64 size, consume_data_callback_t cb,
                   u32 cb_chunk_size, void *ctx_or_buf, u64 offset)
{
        int ret;
        u8 *tmp_buf;
        bool tmp_buf_malloced = false;

        if (cb) {
                /* Send data to callback function in chunks.  */
                if (cb_chunk_size <= STACK_MAX) {
                        tmp_buf = alloca(cb_chunk_size);
                } else {
                        tmp_buf = MALLOC(cb_chunk_size);
                        if (tmp_buf == NULL) {
                                ret = WIMLIB_ERR_NOMEM;
                                goto out;
                        }
                        tmp_buf_malloced = true;
                }

                while (size) {
                        size_t bytes_to_read = min(cb_chunk_size, size);
                        ret = full_pread(in_fd, tmp_buf, bytes_to_read,
                                         offset);
                        if (ret)
                                goto read_error;
                        ret = cb(tmp_buf, bytes_to_read, ctx_or_buf);
                        if (ret)
                                goto out;
                        size -= bytes_to_read;
                        offset += bytes_to_read;
                }
        } else {
                /* Read data directly into buffer.  */
                ret = full_pread(in_fd, ctx_or_buf, size, offset);
                if (ret)
                        goto read_error;
        }
        ret = 0;
        goto out;

read_error:
        ERROR_WITH_ERRNO("Read error");
out:
        if (tmp_buf_malloced)
                FREE(tmp_buf);
        return ret;
}

/*
 * read_partial_wim_resource()-
 *
 * Read a range of data from an uncompressed or compressed resource in a WIM
 * file.  Data is written into a buffer or fed into a callback function, as
 * documented in read_stream_prefix().
 *
 * By default, this function provides the uncompressed data of the resource, and
 * @size and @offset and interpreted relative to the uncompressed contents of
 * the resource.  This behavior can be modified by either of the following
 * flags:
 *
 * WIMLIB_READ_RESOURCE_FLAG_RAW_FULL:
 *      Read @size bytes at @offset of the raw contents of the compressed
 *      resource.  In the case of pipable resources, this excludes the stream
 *      header.  Exclusive with WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS.
 *
 * WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS:
 *      Read the raw compressed chunks of the compressed resource.  @size must
 *      be the full uncompressed size, @offset must be 0, and @cb_chunk_size
 *      must be the resource chunk size.
 *
 * Return values:
 *      WIMLIB_ERR_SUCCESS (0)
 *      WIMLIB_ERR_READ                   (errno set)
 *      WIMLIB_ERR_UNEXPECTED_END_OF_FILE (errno set to 0)
 *      WIMLIB_ERR_NOMEM                  (errno set to ENOMEM)
 *      WIMLIB_ERR_DECOMPRESSION          (errno set to EINVAL)
 *
 *      or other error code returned by the @cb function.
 */
int
read_partial_wim_resource(const struct wim_lookup_table_entry *lte,
                          u64 size, consume_data_callback_t cb,
                          u32 cb_chunk_size, void *ctx_or_buf,
                          int flags, u64 offset)
{
        const struct wim_resource_spec *rspec;
        struct filedes *in_fd;

        /* Verify parameters.  */
        wimlib_assert(lte->resource_location == RESOURCE_IN_WIM);
        rspec = lte->rspec;
        in_fd = &rspec->wim->in_fd;
        if (cb)
                wimlib_assert(is_power_of_2(cb_chunk_size));
        if (flags & WIMLIB_READ_RESOURCE_FLAG_RAW_CHUNKS) {
                /* Raw chunks mode is subject to the restrictions noted.  */
                wimlib_assert(!lte_is_partial(lte));
                wimlib_assert(!(flags & WIMLIB_READ_RESOURCE_FLAG_RAW_FULL));
                wimlib_assert(cb_chunk_size == rspec->cchunk_size);
                wimlib_assert(size == rspec->uncompressed_size);
                wimlib_assert(offset == 0);
        } else if (flags & WIMLIB_READ_RESOURCE_FLAG_RAW_FULL) {
                /* Raw full mode:  read must not overrun end of store size.  */
                wimlib_assert(!lte_is_partial(lte));
                wimlib_assert(offset + size >= size &&
                              offset + size <= rspec->size_in_wim);
        } else {
                /* Normal mode:  read must not overrun end of original size.  */
                wimlib_assert(offset + size >= size &&
                              lte->offset_in_res + offset + size <= rspec->uncompressed_size);
        }

        DEBUG("Reading WIM resource: %"PRIu64" @ +%"PRIu64"[+%"PRIu64"] "
              "from %"PRIu64"(%"PRIu64") @ +%"PRIu64" "
              "(readflags 0x%08x, resflags 0x%02x%s)",
              size, offset, lte->offset_in_res,
              rspec->size_in_wim,
              rspec->uncompressed_size,
              rspec->offset_in_wim,
              flags, lte->flags,
              (rspec->is_pipable ? ", pipable" : ""));

        if ((flags & WIMLIB_READ_RESOURCE_FLAG_RAW_FULL) ||
            rspec->ctype == WIMLIB_COMPRESSION_TYPE_NONE)
        {
                return read_raw_file_data(in_fd,
                                          size,
                                          cb,
                                          cb_chunk_size,
                                          ctx_or_buf,
                                          offset + rspec->offset_in_wim);
        } else {
                return read_compressed_wim_resource(rspec, size, cb,
                                                    cb_chunk_size,
                                                    ctx_or_buf, flags, offset + lte->offset_in_res);
        }
}

int
read_partial_wim_stream_into_buf(const struct wim_lookup_table_entry *lte,
                                 size_t size, u64 offset, void *buf)
{
        return read_partial_wim_resource(lte, size, NULL, 0, buf, 0, offset);
}

static int
read_wim_stream_prefix(const struct wim_lookup_table_entry *lte, u64 size,
                       consume_data_callback_t cb, u32 cb_chunk_size,
                       void *ctx_or_buf, int flags)
{
        return read_partial_wim_resource(lte, size, cb, cb_chunk_size,
                                         ctx_or_buf, flags, 0);
}

#ifndef __WIN32__
/* This function handles reading stream data that is located in an external
 * file,  such as a file that has been added to the WIM image through execution
 * of a wimlib_add_command.
 *
 * This assumes the file can be accessed using the standard POSIX open(),
 * read(), and close().  On Windows this will not necessarily be the case (since
 * the file may need FILE_FLAG_BACKUP_SEMANTICS to be opened, or the file may be
 * encrypted), so Windows uses its own code for its equivalent case.
 */
static int
read_file_on_disk_prefix(const struct wim_lookup_table_entry *lte, u64 size,
                         consume_data_callback_t cb, u32 cb_chunk_size,
                         void *ctx_or_buf, int _ignored_flags)
{
        int ret;
        int raw_fd;
        struct filedes fd;

        wimlib_assert(size <= lte->size);
        DEBUG("Reading %"PRIu64" bytes from \"%"TS"\"", size, lte->file_on_disk);

        raw_fd = open(lte->file_on_disk, O_BINARY | O_RDONLY);
        if (raw_fd < 0) {
                ERROR_WITH_ERRNO("Can't open \"%"TS"\"", lte->file_on_disk);
                return WIMLIB_ERR_OPEN;
        }
        filedes_init(&fd, raw_fd);
        ret = read_raw_file_data(&fd, size, cb, cb_chunk_size, ctx_or_buf, 0);
        filedes_close(&fd);
        return ret;
}
#endif /* !__WIN32__ */

/* This function handles the trivial case of reading stream data that is, in
 * fact, already located in an in-memory buffer.  */
static int
read_buffer_prefix(const struct wim_lookup_table_entry *lte,
                   u64 size, consume_data_callback_t cb,
                   u32 cb_chunk_size, void *ctx_or_buf, int _ignored_flags)
{
        wimlib_assert(size <= lte->size);

        if (cb) {
                /* Feed the data into the callback function in
                 * appropriately-sized chunks.  */
                int ret;
                u32 chunk_size;

                for (u64 offset = 0; offset < size; offset += chunk_size) {
                        chunk_size = min(cb_chunk_size, size - offset);
                        ret = cb((const u8*)lte->attached_buffer + offset,
                                 chunk_size, ctx_or_buf);
                        if (ret)
                                return ret;
                }
        } else {
                /* Copy the data directly into the specified buffer.  */
                memcpy(ctx_or_buf, lte->attached_buffer, size);
        }
        return 0;
}

typedef int (*read_stream_prefix_handler_t)(const struct wim_lookup_table_entry *lte,
                                            u64 size, consume_data_callback_t cb,
                                            u32 cb_chunk_size, void *ctx_or_buf,
                                            int flags);

/*
 * read_stream_prefix()-
 *
 * Reads the first @size bytes from a generic "stream", which may be located in
 * any one of several locations, such as in a WIM file (compressed or
 * uncompressed), in an external file, or directly in an in-memory buffer.
 *
 * This function feeds the data either to a callback function (@cb != NULL,
 * passing it @ctx_or_buf), or write it directly into a buffer (@cb == NULL,
 * @ctx_or_buf specifies the buffer, which must have room for at least @size
 * bytes).
 *
 * When (@cb != NULL), @cb_chunk_size specifies the maximum size of data chunks
 * to feed the callback function.  @cb_chunk_size must be positive, and if the
 * stream is in a WIM file, must be a power of 2.  All chunks, except possibly
 * the last one, will be this size.  If (@cb == NULL), @cb_chunk_size is
 * ignored.
 *
 * If the stream is located in a WIM file, @flags can be set as documented in
 * read_partial_wim_resource().  Otherwise @flags are ignored.
 *
 * Returns 0 on success; nonzero on error.  A nonzero value will be returned if
 * the stream data cannot be successfully read (for a number of different
 * reasons, depending on the stream location), or if a callback function was
 * specified and it returned nonzero.
 */
int
read_stream_prefix(const struct wim_lookup_table_entry *lte, u64 size,
                   consume_data_callback_t cb, u32 cb_chunk_size,
                   void *ctx_or_buf, int flags)
{
        /* This function merely verifies several preconditions, then passes
         * control to an appropriate function for understanding each possible
         * stream location.  */
        static const read_stream_prefix_handler_t handlers[] = {
                [RESOURCE_IN_WIM]             = read_wim_stream_prefix,
        #ifdef __WIN32__
                [RESOURCE_IN_FILE_ON_DISK]    = read_win32_file_prefix,
        #else
                [RESOURCE_IN_FILE_ON_DISK]    = read_file_on_disk_prefix,
        #endif
                [RESOURCE_IN_ATTACHED_BUFFER] = read_buffer_prefix,
        #ifdef WITH_FUSE
                [RESOURCE_IN_STAGING_FILE]    = read_file_on_disk_prefix,
        #endif
        #ifdef WITH_NTFS_3G
                [RESOURCE_IN_NTFS_VOLUME]     = read_ntfs_file_prefix,
        #endif
        #ifdef __WIN32__
                [RESOURCE_WIN32_ENCRYPTED]    = read_win32_encrypted_file_prefix,
        #endif
        };
        wimlib_assert(lte->resource_location < ARRAY_LEN(handlers)
                      && handlers[lte->resource_location] != NULL);
        wimlib_assert(cb == NULL || cb_chunk_size > 0);
        return handlers[lte->resource_location](lte, size, cb, cb_chunk_size,
                                                ctx_or_buf, flags);
}

/* Read the full uncompressed data of the specified stream into the specified
 * buffer, which must have space for at least lte->size bytes.  */
int
read_full_stream_into_buf(const struct wim_lookup_table_entry *lte, void *buf)
{
        return read_stream_prefix(lte, lte->size, NULL, 0, buf, 0);
}

/* Read the full uncompressed data of the specified stream.  A buffer sufficient
 * to hold the data is allocated and returned in @buf_ret.  */
int
read_full_stream_into_alloc_buf(const struct wim_lookup_table_entry *lte,
                                void **buf_ret)
{
        int ret;
        void *buf;

        if ((size_t)lte->size != lte->size) {
                ERROR("Can't read %"PRIu64" byte stream into "
                      "memory", lte->size);
                return WIMLIB_ERR_NOMEM;
        }

        buf = MALLOC(lte->size);
        if (buf == NULL)
                return WIMLIB_ERR_NOMEM;

        ret = read_full_stream_into_buf(lte, buf);
        if (ret) {
                FREE(buf);
                return ret;
        }

        *buf_ret = buf;
        return 0;
}

/* Retrieve the full uncompressed data of the specified WIM resource.  */
static int
wim_resource_spec_to_data(struct wim_resource_spec *rspec, void **buf_ret)
{
        int ret;
        struct wim_lookup_table_entry *lte;

        lte = new_lookup_table_entry();
        if (lte == NULL)
                return WIMLIB_ERR_NOMEM;

        lte->unhashed = 1;
        lte_bind_wim_resource_spec(lte, rspec);
        lte->flags = rspec->flags;
        lte->size = rspec->uncompressed_size;
        lte->offset_in_res = 0;

        ret = read_full_stream_into_alloc_buf(lte, buf_ret);

        lte_unbind_wim_resource_spec(lte);
        free_lookup_table_entry(lte);
        return ret;
}

/* Retrieve the full uncompressed data of the specified WIM resource.  */
int
wim_reshdr_to_data(const struct wim_reshdr *reshdr, WIMStruct *wim, void **buf_ret)
{
        DEBUG("offset_in_wim=%"PRIu64", size_in_wim=%"PRIu64", "
              "uncompressed_size=%"PRIu64,
              reshdr->offset_in_wim, reshdr->size_in_wim, reshdr->uncompressed_size);

        struct wim_resource_spec rspec;
        wim_res_hdr_to_spec(reshdr, wim, &rspec);
        return wim_resource_spec_to_data(&rspec, buf_ret);
}

struct extract_ctx {
        SHA_CTX sha_ctx;
        consume_data_callback_t extract_chunk;
        void *extract_chunk_arg;
};

static int
extract_chunk_sha1_wrapper(const void *chunk, size_t chunk_size, void *_ctx)
{
        struct extract_ctx *ctx = _ctx;

        sha1_update(&ctx->sha_ctx, chunk, chunk_size);
        return ctx->extract_chunk(chunk, chunk_size, ctx->extract_chunk_arg);
}

/* Extracts the first @size bytes of a stream to somewhere.  In the process, the
 * SHA1 message digest of the uncompressed stream is checked if the full stream
 * is being extracted.
 *
 * @extract_chunk is a function that will be called to extract each chunk of the
 * stream.  */
int
extract_stream(const struct wim_lookup_table_entry *lte, u64 size,
               consume_data_callback_t extract_chunk, void *extract_chunk_arg)
{
        int ret;
        if (size == lte->size) {
                /* Do SHA1 */
                struct extract_ctx ctx;
                ctx.extract_chunk = extract_chunk;
                ctx.extract_chunk_arg = extract_chunk_arg;
                sha1_init(&ctx.sha_ctx);
                ret = read_stream_prefix(lte, size,
                                         extract_chunk_sha1_wrapper,
                                         lte_cchunk_size(lte),
                                         &ctx, 0);
                if (ret == 0) {
                        u8 hash[SHA1_HASH_SIZE];
                        sha1_final(hash, &ctx.sha_ctx);
                        if (!hashes_equal(hash, lte->hash)) {
                                if (wimlib_print_errors) {
                                        ERROR("Invalid SHA1 message digest "
                                              "on the following WIM stream:");
                                        print_lookup_table_entry(lte, stderr);
                                        if (lte->resource_location == RESOURCE_IN_WIM)
                                                ERROR("The WIM file appears to be corrupt!");
                                }
                                ret = WIMLIB_ERR_INVALID_RESOURCE_HASH;
                        }
                }
        } else {
                /* Don't do SHA1 */
                ret = read_stream_prefix(lte, size, extract_chunk,
                                         lte_cchunk_size(lte),
                                         extract_chunk_arg, 0);
        }
        return ret;
}

static int
extract_wim_chunk_to_fd(const void *buf, size_t len, void *_fd_p)
{
        struct filedes *fd = _fd_p;
        int ret = full_write(fd, buf, len);
        if (ret)
                ERROR_WITH_ERRNO("Error writing to file descriptor");
        return ret;
}

/* Extract the first @size bytes of the specified stream to the specified file
 * descriptor.  If @size is the full size of the stream, its SHA1 message digest
 * is also checked.  */
int
extract_stream_to_fd(const struct wim_lookup_table_entry *lte,
                     struct filedes *fd, u64 size)
{
        return extract_stream(lte, size, extract_wim_chunk_to_fd, fd);
}


static int
sha1_chunk(const void *buf, size_t len, void *ctx)
{
        sha1_update(ctx, buf, len);
        return 0;
}

/* Calculate the SHA1 message digest of a stream, storing it in @lte->hash.  */
int
sha1_stream(struct wim_lookup_table_entry *lte)
{
        int ret;
        SHA_CTX sha_ctx;

        sha1_init(&sha_ctx);
        ret = read_stream_prefix(lte, lte->size,
                                 sha1_chunk, lte_cchunk_size(lte),
                                 &sha_ctx, 0);
        if (ret == 0)
                sha1_final(lte->hash, &sha_ctx);

        return ret;
}

/* Convert a WIM resource header to a stand-alone resource specification.  */
void
wim_res_hdr_to_spec(const struct wim_reshdr *reshdr, WIMStruct *wim,
                    struct wim_resource_spec *spec)
{
        spec->wim = wim;
        spec->offset_in_wim = reshdr->offset_in_wim;
        spec->size_in_wim = reshdr->size_in_wim;
        spec->uncompressed_size = reshdr->uncompressed_size;
        INIT_LIST_HEAD(&spec->lte_list);
        spec->flags = reshdr->flags;
        spec->is_pipable = wim_is_pipable(wim);
        if (spec->flags & (WIM_RESHDR_FLAG_COMPRESSED | WIM_RESHDR_FLAG_CONCAT)) {
                spec->ctype = wim->compression_type;
                spec->cchunk_size = wim->chunk_size;
        } else {
                spec->ctype = WIMLIB_COMPRESSION_TYPE_NONE;
                spec->cchunk_size = 0;
        }
}

/* Convert a stand-alone resource specification to a WIM resource header.  */
void
wim_res_spec_to_hdr(const struct wim_resource_spec *rspec,
                    struct wim_reshdr *reshdr)
{
        reshdr->offset_in_wim     = rspec->offset_in_wim;
        reshdr->size_in_wim       = rspec->size_in_wim;
        reshdr->flags             = rspec->flags;
        reshdr->uncompressed_size = rspec->uncompressed_size;
}

/* Translates a WIM resource header from the on-disk format into an in-memory
 * format.  */
int
get_wim_reshdr(const struct wim_reshdr_disk *disk_reshdr,
               struct wim_reshdr *reshdr)
{
        reshdr->offset_in_wim = le64_to_cpu(disk_reshdr->offset_in_wim);
        reshdr->size_in_wim = (((u64)disk_reshdr->size_in_wim[0] <<  0) |
                              ((u64)disk_reshdr->size_in_wim[1] <<  8) |
                              ((u64)disk_reshdr->size_in_wim[2] << 16) |
                              ((u64)disk_reshdr->size_in_wim[3] << 24) |
                              ((u64)disk_reshdr->size_in_wim[4] << 32) |
                              ((u64)disk_reshdr->size_in_wim[5] << 40) |
                              ((u64)disk_reshdr->size_in_wim[6] << 48));
        reshdr->uncompressed_size = le64_to_cpu(disk_reshdr->uncompressed_size);
        reshdr->flags = disk_reshdr->flags;

        /* Truncate numbers to 62 bits to avoid possible overflows.  */
        if (reshdr->offset_in_wim & 0xc000000000000000ULL)
                return WIMLIB_ERR_INVALID_LOOKUP_TABLE_ENTRY;

        if (reshdr->uncompressed_size & 0xc000000000000000ULL)
                return WIMLIB_ERR_INVALID_LOOKUP_TABLE_ENTRY;

        return 0;
}

/* Translates a WIM resource header from an in-memory format into the on-disk
 * format.  */
void
put_wim_reshdr(const struct wim_reshdr *reshdr,
               struct wim_reshdr_disk *disk_reshdr)
{
        disk_reshdr->size_in_wim[0] = reshdr->size_in_wim  >>  0;
        disk_reshdr->size_in_wim[1] = reshdr->size_in_wim  >>  8;
        disk_reshdr->size_in_wim[2] = reshdr->size_in_wim  >> 16;
        disk_reshdr->size_in_wim[3] = reshdr->size_in_wim  >> 24;
        disk_reshdr->size_in_wim[4] = reshdr->size_in_wim  >> 32;
        disk_reshdr->size_in_wim[5] = reshdr->size_in_wim  >> 40;
        disk_reshdr->size_in_wim[6] = reshdr->size_in_wim  >> 48;
        disk_reshdr->flags = reshdr->flags;
        disk_reshdr->offset_in_wim = cpu_to_le64(reshdr->offset_in_wim);
        disk_reshdr->uncompressed_size = cpu_to_le64(reshdr->uncompressed_size);
}