]> wimlib.net Git - wimlib/blobdiff - src/resource.c
bitops: rename bit scan functions
[wimlib] / src / resource.c
index a49507f3778df55b652b54df4c0e00b5e480cf3a..385f9f65ba6836aa594c5bb34e9b4823eabff79c 100644 (file)
 #include "wimlib/endianness.h"
 #include "wimlib/error.h"
 #include "wimlib/file_io.h"
+#include "wimlib/ntfs_3g.h"
 #include "wimlib/resource.h"
 #include "wimlib/sha1.h"
 #include "wimlib/wim.h"
-
-#ifdef __WIN32__
-/* for read_winnt_stream_prefix(), read_win32_encrypted_file_prefix() */
-#  include "wimlib/win32.h"
-#endif
-
-#ifdef WITH_NTFS_3G
-/* for read_ntfs_attribute_prefix() */
-#  include "wimlib/ntfs_3g.h"
-#endif
-
+#include "wimlib/win32.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:
+ * A compressed resource in a WIM consists of a sequence of chunks.  Each chunk
+ * decompresses to the same size except possibly for the last, which
+ * decompresses to the remaining size.  Chunks that did not compress to less
+ * than their original size are stored uncompressed.
  *
- * - 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.
+ * We support three variations on this resource format, independently of the
+ * compression type and chunk size which can vary as well:
  *
- * - The chunk table is included in the compressed size of the resource provided
- *   in the corresponding entry in the WIM's blob table.
+ * - Original resource format: immediately before the compressed chunks, 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.  Chunk table
+ *   entries are 32-bit for resources < 4 GiB uncompressed and 64-bit for
+ *   resources >= 4 GiB uncompressed.
  *
- * - 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.
+ * - Solid resource format (distinguished by the use of WIM_RESHDR_FLAG_SOLID
+ *   instead of WIM_RESHDR_FLAG_COMPRESSED): similar to the original format, but
+ *   the resource begins with a 16-byte header which specifies the uncompressed
+ *   size of the resource, the compression type, and the chunk size.  (In the
+ *   original format, these values were instead determined from outside the
+ *   resource itself, from the blob table and the WIM file header.) In addition,
+ *   in this format the entries in the chunk table contain compressed chunk
+ *   sizes rather than offsets.  As a consequence of this, the chunk table
+ *   entries are always 32-bit and there is an entry for chunk 0.
  *
- * 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.
+ * - Pipable resource format (wimlib extension; all resources in a pipable WIM
+ *   have this format): similar to the original format, but the chunk table is
+ *   at the end of the resource rather than the beginning, and each compressed
+ *   chunk is prefixed with its compressed size as a 32-bit integer.  This
+ *   format allows a resource to be written without rewinding.
  */
 
 
@@ -95,8 +84,6 @@ struct data_range {
 };
 
 /*
- * read_compressed_wim_resource() -
- *
  * Read data from a compressed WIM resource.
  *
  * @rdesc
@@ -106,34 +93,30 @@ struct data_range {
  *     read, sorted by increasing offset.
  * @num_ranges
  *     Number of ranges in @ranges; must be at least 1.
- * @cb
- *     Callback function to feed the data being read.  Each call provides the
- *     next chunk of the requested data, uncompressed.  Each chunk will be of
- *     nonzero size and will not cross range boundaries, but otherwise will be
- *     of unspecified size.
- * @cb_ctx
- *     Parameter to pass to @cb_ctx.
+ * @cbs
+ *     Structure which provides the consume_chunk() callback to feed the data
+ *     being read.  Each call provides the next chunk of the requested data,
+ *     uncompressed.  Each chunk will be nonempty and will not cross range
+ *     boundaries but otherwise will be of unspecified size.
  *
  * Possible return values:
  *
  *     WIMLIB_ERR_SUCCESS (0)
  *     WIMLIB_ERR_READ                   (errno set)
- *     WIMLIB_ERR_UNEXPECTED_END_OF_FILE (errno set to 0)
+ *     WIMLIB_ERR_UNEXPECTED_END_OF_FILE (errno set to EINVAL)
  *     WIMLIB_ERR_NOMEM                  (errno set to ENOMEM)
  *     WIMLIB_ERR_DECOMPRESSION          (errno set to EINVAL)
+ *     WIMLIB_ERR_INVALID_CHUNK_SIZE     (errno set to EINVAL)
  *
- *     or other error code returned by the @cb function.
+ *     or other error code returned by the cbs->consume_chunk() function.
  */
 static int
 read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                             const struct data_range * const ranges,
                             const size_t num_ranges,
-                            const consume_data_callback_t cb,
-                            void * const cb_ctx)
+                            const struct read_blob_callbacks *cbs)
 {
        int ret;
-       int errno_save;
-
        u64 *chunk_offsets = NULL;
        u8 *ubuf = NULL;
        void *cbuf = NULL;
@@ -143,17 +126,10 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
        struct wimlib_decompressor *decompressor = NULL;
 
        /* Sanity checks  */
-       wimlib_assert(rdesc != NULL);
-       wimlib_assert(resource_is_compressed(rdesc));
-       wimlib_assert(cb != NULL);
        wimlib_assert(num_ranges != 0);
        for (size_t i = 0; i < num_ranges; i++) {
-               DEBUG("Range %zu/%zu: %"PRIu64"@+%"PRIu64" / %"PRIu64,
-                     i + 1, num_ranges, ranges[i].size, ranges[i].offset,
-                     rdesc->uncompressed_size);
-               wimlib_assert(ranges[i].size != 0);
-               wimlib_assert(ranges[i].offset + ranges[i].size >= ranges[i].size);
-               wimlib_assert(ranges[i].offset + ranges[i].size <= rdesc->uncompressed_size);
+               wimlib_assert(ranges[i].offset + ranges[i].size > ranges[i].offset &&
+                             ranges[i].offset + ranges[i].size <= rdesc->uncompressed_size);
        }
        for (size_t i = 0; i < num_ranges - 1; i++)
                wimlib_assert(ranges[i].offset + ranges[i].size <= ranges[i + 1].offset);
@@ -185,18 +161,18 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                cur_read_offset += sizeof(struct alt_chunk_table_header_disk);
        }
 
-       if (!is_power_of_2(chunk_size)) {
+       if (unlikely(!is_power_of_2(chunk_size))) {
                ERROR("Invalid compressed resource: "
                      "expected power-of-2 chunk size (got %"PRIu32")",
                      chunk_size);
                ret = WIMLIB_ERR_INVALID_CHUNK_SIZE;
                errno = EINVAL;
-               goto out_free_memory;
+               goto out_cleanup;
        }
 
        /* Get valid decompressor.  */
-       if (ctype == rdesc->wim->decompressor_ctype &&
-           chunk_size == rdesc->wim->decompressor_max_block_size)
+       if (likely(ctype == rdesc->wim->decompressor_ctype &&
+                  chunk_size == rdesc->wim->decompressor_max_block_size))
        {
                /* Cached decompressor.  */
                decompressor = rdesc->wim->decompressor;
@@ -205,14 +181,14 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
        } else {
                ret = wimlib_create_decompressor(ctype, chunk_size,
                                                 &decompressor);
-               if (ret) {
+               if (unlikely(ret)) {
                        if (ret != WIMLIB_ERR_NOMEM)
                                errno = EINVAL;
-                       goto out_free_memory;
+                       goto out_cleanup;
                }
        }
 
-       const u32 chunk_order = fls32(chunk_size);
+       const u32 chunk_order = bsr32(chunk_size);
 
        /* Calculate the total number of chunks the resource is divided into.  */
        const u64 num_chunks = (rdesc->uncompressed_size + chunk_size - 1) >> chunk_order;
@@ -258,47 +234,48 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                 * to initialize the chunk_offsets array.  */
 
                u64 first_chunk_entry_to_read;
-               u64 last_chunk_entry_to_read;
+               u64 num_chunk_entries_to_read;
 
                if (alt_chunk_table) {
                        /* The alternate chunk table contains chunk sizes, not
                         * offsets, so we always must read all preceding entries
                         * in order to determine offsets.  */
                        first_chunk_entry_to_read = 0;
-                       last_chunk_entry_to_read = last_needed_chunk;
+                       num_chunk_entries_to_read = last_needed_chunk + 1;
                } else {
-                       /* Here we must account for the fact that the first
-                        * chunk has no explicit chunk table entry.  */
 
-                       if (read_start_chunk == 0)
+                       num_chunk_entries_to_read = last_needed_chunk - read_start_chunk + 1;
+
+                       /* The first chunk has no explicit chunk table entry.  */
+                       if (read_start_chunk == 0) {
+                               num_chunk_entries_to_read--;
                                first_chunk_entry_to_read = 0;
-                       else
+                       } else {
                                first_chunk_entry_to_read = read_start_chunk - 1;
+                       }
 
-                       if (last_needed_chunk == 0)
-                               last_chunk_entry_to_read = 0;
-                       else
-                               last_chunk_entry_to_read = last_needed_chunk - 1;
-
+                       /* Unless we're reading the final chunk of the resource,
+                        * we need the offset of the chunk following the last
+                        * needed chunk so that the compressed size of the last
+                        * needed chunk can be computed.  */
                        if (last_needed_chunk < num_chunks - 1)
-                               last_chunk_entry_to_read++;
+                               num_chunk_entries_to_read++;
                }
 
-               const u64 num_chunk_entries_to_read =
-                       last_chunk_entry_to_read - first_chunk_entry_to_read + 1;
-
                const u64 chunk_offsets_alloc_size =
                        max(num_chunk_entries_to_read,
                            num_needed_chunk_offsets) * sizeof(chunk_offsets[0]);
 
-               if ((size_t)chunk_offsets_alloc_size != chunk_offsets_alloc_size)
+               if (unlikely((size_t)chunk_offsets_alloc_size != chunk_offsets_alloc_size)) {
+                       errno = ENOMEM;
                        goto oom;
+               }
 
-               if (chunk_offsets_alloc_size <= STACK_MAX) {
+               if (likely(chunk_offsets_alloc_size <= STACK_MAX)) {
                        chunk_offsets = alloca(chunk_offsets_alloc_size);
                } else {
                        chunk_offsets = MALLOC(chunk_offsets_alloc_size);
-                       if (chunk_offsets == NULL)
+                       if (unlikely(!chunk_offsets))
                                goto oom;
                        chunk_offsets_malloced = true;
                }
@@ -318,7 +295,7 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
 
                ret = full_pread(in_fd, chunk_table_data, chunk_table_size_to_read,
                                 file_offset_of_needed_chunk_entries);
-               if (ret)
+               if (unlikely(ret))
                        goto read_error;
 
                /* Now fill in chunk_offsets from the entries we have read in
@@ -368,7 +345,7 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                ubuf = alloca(chunk_size);
        } else {
                ubuf = MALLOC(chunk_size);
-               if (ubuf == NULL)
+               if (unlikely(!ubuf))
                        goto oom;
                ubuf_malloced = true;
        }
@@ -381,7 +358,7 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                cbuf = alloca(chunk_size - 1);
        } else {
                cbuf = MALLOC(chunk_size - 1);
-               if (cbuf == NULL)
+               if (unlikely(!cbuf))
                        goto oom;
                cbuf_malloced = true;
        }
@@ -409,7 +386,7 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
 
                        ret = full_pread(in_fd, &chunk_hdr,
                                         sizeof(chunk_hdr), cur_read_offset);
-                       if (ret)
+                       if (unlikely(ret))
                                goto read_error;
                        chunk_csize = le32_to_cpu(chunk_hdr.compressed_size);
                } else {
@@ -424,11 +401,11 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                                              chunk_offsets[i - read_start_chunk];
                        }
                }
-               if (chunk_csize == 0 || chunk_csize > chunk_usize) {
+               if (unlikely(chunk_csize == 0 || chunk_csize > chunk_usize)) {
                        ERROR("Invalid chunk size in compressed resource!");
                        errno = EINVAL;
                        ret = WIMLIB_ERR_DECOMPRESSION;
-                       goto out_free_memory;
+                       goto out_cleanup;
                }
                if (rdesc->is_pipable)
                        cur_read_offset += sizeof(struct pwm_chunk_hdr);
@@ -447,7 +424,7 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                                u8 dummy;
 
                                ret = full_pread(in_fd, &dummy, 1, cur_read_offset - 1);
-                               if (ret)
+                               if (unlikely(ret))
                                        goto read_error;
                        }
                } else {
@@ -465,23 +442,20 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                                         read_buf,
                                         chunk_csize,
                                         cur_read_offset);
-                       if (ret)
+                       if (unlikely(ret))
                                goto read_error;
 
                        if (read_buf == cbuf) {
-                               DEBUG("Decompressing chunk %"PRIu64" "
-                                     "(csize=%"PRIu32" usize=%"PRIu32")",
-                                     i, chunk_csize, chunk_usize);
                                ret = wimlib_decompress(cbuf,
                                                        chunk_csize,
                                                        ubuf,
                                                        chunk_usize,
                                                        decompressor);
-                               if (ret) {
+                               if (unlikely(ret)) {
                                        ERROR("Failed to decompress data!");
                                        ret = WIMLIB_ERR_DECOMPRESSION;
                                        errno = EINVAL;
-                                       goto out_free_memory;
+                                       goto out_cleanup;
                                }
                        }
                        cur_read_offset += chunk_csize;
@@ -499,10 +473,9 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                                end = min(cur_range_end, chunk_end_offset) - chunk_start_offset;
                                size = end - start;
 
-                               ret = (*cb)(&ubuf[start], size, cb_ctx);
-
-                               if (ret)
-                                       goto out_free_memory;
+                               ret = call_consume_chunk(&ubuf[start], size, cbs);
+                               if (unlikely(ret))
+                                       goto out_cleanup;
 
                                cur_range_pos += size;
                                if (cur_range_pos == cur_range_end) {
@@ -529,13 +502,12 @@ read_compressed_wim_resource(const struct wim_resource_descriptor * const rdesc,
                 * returns.  */
                cur_read_offset += chunk_table_size;
                ret = full_pread(in_fd, &dummy, 1, cur_read_offset - 1);
-               if (ret)
+               if (unlikely(ret))
                        goto read_error;
        }
        ret = 0;
 
-out_free_memory:
-       errno_save = errno;
+out_cleanup:
        if (decompressor) {
                wimlib_free_decompressor(rdesc->wim->decompressor);
                rdesc->wim->decompressor = decompressor;
@@ -548,48 +520,24 @@ out_free_memory:
                FREE(ubuf);
        if (cbuf_malloced)
                FREE(cbuf);
-       errno = errno_save;
        return ret;
 
 oom:
-       ERROR("Not enough memory available to read size=%"PRIu64" bytes "
-             "from compressed WIM resource!", last_offset - first_offset + 1);
-       errno = ENOMEM;
+       ERROR("Out of memory while reading compressed WIM resource");
        ret = WIMLIB_ERR_NOMEM;
-       goto out_free_memory;
+       goto out_cleanup;
 
 read_error:
-       ERROR_WITH_ERRNO("Error reading compressed WIM resource!");
-       goto out_free_memory;
-}
-
-static int
-fill_zeroes(u64 size, consume_data_callback_t cb, void *cb_ctx)
-{
-       if (unlikely(size)) {
-               u8 buf[min(size, BUFFER_SIZE)];
-
-               memset(buf, 0, sizeof(buf));
-
-               do {
-                       size_t len;
-                       int ret;
-
-                       len = min(size, BUFFER_SIZE);
-                       ret = cb(buf, len, cb_ctx);
-                       if (ret)
-                               return ret;
-                       size -= len;
-               } while (size);
-       }
-       return 0;
+       ERROR_WITH_ERRNO("Error reading data from WIM file");
+       goto out_cleanup;
 }
 
 /* Read raw data from a file descriptor at the specified offset, feeding the
- * data it in chunks into the specified callback function.  */
+ * data in nonempty chunks into the cbs->consume_chunk() function.  */
 static int
 read_raw_file_data(struct filedes *in_fd, u64 offset, u64 size,
-                  consume_data_callback_t cb, void *cb_ctx)
+                  const struct read_blob_callbacks *cbs,
+                  const tchar *filename)
 {
        u8 buf[BUFFER_SIZE];
        size_t bytes_to_read;
@@ -598,159 +546,105 @@ read_raw_file_data(struct filedes *in_fd, u64 offset, u64 size,
        while (size) {
                bytes_to_read = min(sizeof(buf), size);
                ret = full_pread(in_fd, buf, bytes_to_read, offset);
-               if (ret) {
-                       ERROR_WITH_ERRNO("Read error");
-                       return ret;
-               }
-               ret = cb(buf, bytes_to_read, cb_ctx);
-               if (ret)
+               if (unlikely(ret))
+                       goto read_error;
+               ret = call_consume_chunk(buf, bytes_to_read, cbs);
+               if (unlikely(ret))
                        return ret;
                size -= bytes_to_read;
                offset += bytes_to_read;
        }
        return 0;
+
+read_error:
+       if (!filename) {
+               ERROR_WITH_ERRNO("Error reading data from WIM file");
+       } else if (ret == WIMLIB_ERR_UNEXPECTED_END_OF_FILE) {
+               ERROR("\"%"TS"\": File was concurrently truncated", filename);
+               ret = WIMLIB_ERR_CONCURRENT_MODIFICATION_DETECTED;
+       } else {
+               ERROR_WITH_ERRNO("\"%"TS"\": Error reading data", filename);
+       }
+       return ret;
 }
 
-/* A consume_data_callback_t implementation that simply concatenates all chunks
- * into a buffer.  */
+/* A consume_chunk() implementation that simply concatenates all chunks into an
+ * in-memory buffer.  */
 static int
 bufferer_cb(const void *chunk, size_t size, void *_ctx)
 {
-       u8 **buf_p = _ctx;
+       void **buf_p = _ctx;
 
        *buf_p = mempcpy(*buf_p, chunk, size);
        return 0;
 }
 
 /*
- * read_partial_wim_resource()-
+ * Read @size bytes at @offset in the WIM resource described by @rdesc and feed
+ * the data into the @cbs->consume_chunk callback function.
  *
- * Read a range of data from an uncompressed or compressed resource in a WIM
- * file.
+ * @offset and @size are assumed to have already been validated against the
+ * resource's uncompressed size.
  *
- * @rdesc
- *     Description of the WIM resource to read from.
- * @offset
- *     Offset within the uncompressed resource at which to start reading.
- * @size
- *     Number of bytes to read.
- * @cb
- *     Callback function to feed the data being read.  Each call provides the
- *     next chunk of the requested data, uncompressed.  Each chunk will be of
- *     nonzero size and will not cross range boundaries, but otherwise will be
- *     of unspecified size.
- * @cb_ctx
- *     Parameter to pass to @cb_ctx.
- *
- * 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.
+ * Returns 0 on success; or the first nonzero value returned by the callback
+ * function; or a nonzero wimlib error code with errno set as well.
  */
 static int
 read_partial_wim_resource(const struct wim_resource_descriptor *rdesc,
-                         u64 offset, u64 size,
-                         consume_data_callback_t cb, void *cb_ctx)
+                         const u64 offset, const u64 size,
+                         const struct read_blob_callbacks *cbs)
 {
-       /* Sanity checks.  */
-       wimlib_assert(offset + size >= offset);
-       wimlib_assert(offset + size <= rdesc->uncompressed_size);
-
-       DEBUG("Reading %"PRIu64" @ %"PRIu64" from WIM resource  "
-             "%"PRIu64" => %"PRIu64" @ %"PRIu64,
-             size, offset, rdesc->uncompressed_size,
-             rdesc->size_in_wim, rdesc->offset_in_wim);
-
-       /* Trivial case.  */
-       if (size == 0)
-               return 0;
-
-       if (resource_is_compressed(rdesc)) {
+       if (rdesc->flags & (WIM_RESHDR_FLAG_COMPRESSED |
+                           WIM_RESHDR_FLAG_SOLID))
+       {
+               /* Compressed resource  */
+               if (unlikely(!size))
+                       return 0;
                struct data_range range = {
                        .offset = offset,
                        .size = size,
                };
-               return read_compressed_wim_resource(rdesc, &range, 1,
-                                                   cb, cb_ctx);
-       } else {
-               /* Reading uncompressed resource.  For completeness, handle the
-                * weird case where size_in_wim < uncompressed_size.  */
-
-               u64 read_size;
-               u64 zeroes_size;
-               int ret;
-
-               if (likely(offset + size <= rdesc->size_in_wim) ||
-                   rdesc->is_pipable)
-               {
-                       read_size = size;
-                       zeroes_size = 0;
-               } else {
-                       if (offset >= rdesc->size_in_wim) {
-                               read_size = 0;
-                               zeroes_size = size;
-                       } else {
-                               read_size = rdesc->size_in_wim - offset;
-                               zeroes_size = offset + size - rdesc->size_in_wim;
-                       }
-               }
-
-               ret = read_raw_file_data(&rdesc->wim->in_fd,
-                                        rdesc->offset_in_wim + offset,
-                                        read_size,
-                                        cb,
-                                        cb_ctx);
-               if (ret)
-                       return ret;
-
-               return fill_zeroes(zeroes_size, cb, cb_ctx);
+               return read_compressed_wim_resource(rdesc, &range, 1, cbs);
        }
+
+       /* Uncompressed resource  */
+       return read_raw_file_data(&rdesc->wim->in_fd,
+                                 rdesc->offset_in_wim + offset,
+                                 size, cbs, NULL);
 }
 
 /* Read the specified range of uncompressed data from the specified blob, which
- * must be located into a WIM file, into the specified buffer.  */
+ * must be located in a WIM file, into the specified buffer.  */
 int
 read_partial_wim_blob_into_buf(const struct blob_descriptor *blob,
-                              size_t size, u64 offset, void *_buf)
+                              u64 offset, size_t size, void *buf)
 {
-       u8 *buf = _buf;
-
-       wimlib_assert(blob->blob_location == BLOB_IN_WIM);
-
+       struct read_blob_callbacks cbs = {
+               .consume_chunk  = bufferer_cb,
+               .ctx            = &buf,
+       };
        return read_partial_wim_resource(blob->rdesc,
                                         blob->offset_in_res + offset,
                                         size,
-                                        bufferer_cb,
-                                        &buf);
-}
-
-/* A consume_data_callback_t implementation that simply ignores the data
- * received.  */
-static int
-skip_chunk_cb(const void *chunk, size_t size, void *_ctx)
-{
-       return 0;
+                                        &cbs);
 }
 
 /* Skip over the data of the specified WIM resource.  */
 int
-skip_wim_resource(struct wim_resource_descriptor *rdesc)
+skip_wim_resource(const struct wim_resource_descriptor *rdesc)
 {
-       DEBUG("Skipping resource (size=%"PRIu64")", rdesc->uncompressed_size);
-       return read_partial_wim_resource(rdesc, 0, rdesc->uncompressed_size,
-                                        skip_chunk_cb, NULL);
+       struct read_blob_callbacks cbs = {
+       };
+       return read_partial_wim_resource(rdesc, 0,
+                                        rdesc->uncompressed_size, &cbs);
 }
 
 static int
 read_wim_blob_prefix(const struct blob_descriptor *blob, u64 size,
-                    consume_data_callback_t cb, void *cb_ctx)
+                    const struct read_blob_callbacks *cbs)
 {
-       return read_partial_wim_resource(blob->rdesc, blob->offset_in_res, size,
-                                        cb, cb_ctx);
+       return read_partial_wim_resource(blob->rdesc, blob->offset_in_res,
+                                        size, cbs);
 }
 
 /* This function handles reading blob data that is located in an external file,
@@ -763,23 +657,19 @@ read_wim_blob_prefix(const struct blob_descriptor *blob, u64 size,
  * encrypted), so Windows uses its own code for its equivalent case.  */
 static int
 read_file_on_disk_prefix(const struct blob_descriptor *blob, u64 size,
-                        consume_data_callback_t cb, void *cb_ctx)
+                        const struct read_blob_callbacks *cbs)
 {
        int ret;
        int raw_fd;
        struct filedes fd;
 
-       wimlib_assert(size <= blob->size);
-
-       DEBUG("Reading %"PRIu64" bytes from \"%"TS"\"", size, blob->file_on_disk);
-
        raw_fd = topen(blob->file_on_disk, O_BINARY | O_RDONLY);
-       if (raw_fd < 0) {
+       if (unlikely(raw_fd < 0)) {
                ERROR_WITH_ERRNO("Can't open \"%"TS"\"", blob->file_on_disk);
                return WIMLIB_ERR_OPEN;
        }
        filedes_init(&fd, raw_fd);
-       ret = read_raw_file_data(&fd, 0, size, cb, cb_ctx);
+       ret = read_raw_file_data(&fd, 0, size, cbs, blob->file_on_disk);
        filedes_close(&fd);
        return ret;
 }
@@ -787,26 +677,21 @@ read_file_on_disk_prefix(const struct blob_descriptor *blob, u64 size,
 #ifdef WITH_FUSE
 static int
 read_staging_file_prefix(const struct blob_descriptor *blob, u64 size,
-                        consume_data_callback_t cb, void *cb_ctx)
+                        const struct read_blob_callbacks *cbs)
 {
        int raw_fd;
        struct filedes fd;
        int ret;
 
-       wimlib_assert(size <= blob->size);
-
-       DEBUG("Reading %"PRIu64" bytes from staging file \"%s\"",
-             size, blob->staging_file_name);
-
        raw_fd = openat(blob->staging_dir_fd, blob->staging_file_name,
                        O_RDONLY | O_NOFOLLOW);
-       if (raw_fd < 0) {
+       if (unlikely(raw_fd < 0)) {
                ERROR_WITH_ERRNO("Can't open staging file \"%s\"",
                                 blob->staging_file_name);
                return WIMLIB_ERR_OPEN;
        }
        filedes_init(&fd, raw_fd);
-       ret = read_raw_file_data(&fd, 0, size, cb, cb_ctx);
+       ret = read_raw_file_data(&fd, 0, size, cbs, blob->staging_file_name);
        filedes_close(&fd);
        return ret;
 }
@@ -816,35 +701,32 @@ read_staging_file_prefix(const struct blob_descriptor *blob, u64 size,
  * already located in an in-memory buffer.  */
 static int
 read_buffer_prefix(const struct blob_descriptor *blob,
-                  u64 size, consume_data_callback_t cb, void *cb_ctx)
+                  u64 size, const struct read_blob_callbacks *cbs)
 {
-       wimlib_assert(size <= blob->size);
-       return (*cb)(blob->attached_buffer, size, cb_ctx);
+       if (unlikely(!size))
+               return 0;
+       return call_consume_chunk(blob->attached_buffer, size, cbs);
 }
 
 typedef int (*read_blob_prefix_handler_t)(const struct blob_descriptor *blob,
                                          u64 size,
-                                         consume_data_callback_t cb,
-                                         void *cb_ctx);
+                                         const struct read_blob_callbacks *cbs);
 
 /*
- * read_blob_prefix()-
- *
- * Reads the first @size bytes from a generic "blob", 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 to a callback function @cb in chunks of
- * unspecified size.
+ * Read the first @size bytes from a generic "blob", which may be located in any
+ * one of several locations, such as in a WIM resource (possibly compressed), in
+ * an external file, or directly in an in-memory buffer.  The blob data will be
+ * fed to the cbs->consume_chunk() callback function in chunks that are nonempty
+ * but otherwise are of unspecified size.
  *
  * Returns 0 on success; nonzero on error.  A nonzero value will be returned if
  * the blob data cannot be successfully read (for a number of different reasons,
- * depending on the blob location), or if @cb returned nonzero in which case
- * that error code will be returned.
+ * depending on the blob location), or if cbs->consume_chunk() returned nonzero
+ * in which case that error code will be returned.
  */
 static int
 read_blob_prefix(const struct blob_descriptor *blob, u64 size,
-                consume_data_callback_t cb, void *cb_ctx)
+                const struct read_blob_callbacks *cbs)
 {
        static const read_blob_prefix_handler_t handlers[] = {
                [BLOB_IN_WIM] = read_wim_blob_prefix,
@@ -857,43 +739,65 @@ read_blob_prefix(const struct blob_descriptor *blob, u64 size,
                [BLOB_IN_NTFS_VOLUME] = read_ntfs_attribute_prefix,
        #endif
        #ifdef __WIN32__
-               [BLOB_IN_WINNT_FILE_ON_DISK] = read_winnt_stream_prefix,
-               [BLOB_WIN32_ENCRYPTED] = read_win32_encrypted_file_prefix,
+               [BLOB_IN_WINDOWS_FILE] = read_windows_file_prefix,
        #endif
        };
        wimlib_assert(blob->blob_location < ARRAY_LEN(handlers)
                      && handlers[blob->blob_location] != NULL);
-       return handlers[blob->blob_location](blob, size, cb, cb_ctx);
+       wimlib_assert(size <= blob->size);
+       return handlers[blob->blob_location](blob, size, cbs);
+}
+
+/* Read the full data of the specified blob, passing the data into the specified
+ * callbacks (all of which are optional).  */
+int
+read_blob_with_cbs(struct blob_descriptor *blob,
+                  const struct read_blob_callbacks *cbs)
+{
+       int ret;
+
+       ret = call_begin_blob(blob, cbs);
+       if (unlikely(ret))
+               return ret;
+
+       ret = read_blob_prefix(blob, blob->size, cbs);
+
+       return call_end_blob(blob, ret, cbs);
 }
 
 /* Read the full uncompressed data of the specified blob into the specified
- * buffer, which must have space for at least blob->size bytes.  */
+ * buffer, which must have space for at least blob->size bytes.  The SHA-1
+ * message digest is *not* checked.  */
 int
-read_full_blob_into_buf(const struct blob_descriptor *blob, void *_buf)
+read_blob_into_buf(const struct blob_descriptor *blob, void *buf)
 {
-       u8 *buf = _buf;
-       return read_blob_prefix(blob, blob->size, bufferer_cb, &buf);
+       struct read_blob_callbacks cbs = {
+               .consume_chunk  = bufferer_cb,
+               .ctx            = &buf,
+       };
+       return read_blob_prefix(blob, blob->size, &cbs);
 }
 
 /* Retrieve the full uncompressed data of the specified blob.  A buffer large
- * enough hold the data is allocated and returned in @buf_ret.  */
+ * enough hold the data is allocated and returned in @buf_ret.  The SHA-1
+ * message digest is *not* checked.  */
 int
-read_full_blob_into_alloc_buf(const struct blob_descriptor *blob, void **buf_ret)
+read_blob_into_alloc_buf(const struct blob_descriptor *blob, void **buf_ret)
 {
        int ret;
        void *buf;
 
-       if ((size_t)blob->size != blob->size) {
+       if (unlikely((size_t)blob->size != blob->size)) {
                ERROR("Can't read %"PRIu64" byte blob into memory", blob->size);
                return WIMLIB_ERR_NOMEM;
        }
 
        buf = MALLOC(blob->size);
-       if (buf == NULL)
+       if (unlikely(!buf))
                return WIMLIB_ERR_NOMEM;
 
-       ret = read_full_blob_into_buf(blob, buf);
-       if (ret) {
+       ret = read_blob_into_buf(blob, buf);
+       if (unlikely(ret)) {
                FREE(buf);
                return ret;
        }
@@ -906,17 +810,19 @@ read_full_blob_into_alloc_buf(const struct blob_descriptor *blob, void **buf_ret
  * `wim_reshdr' and the corresponding WIM file.  A buffer large enough hold the
  * data is allocated and returned in @buf_ret.  */
 int
-wim_reshdr_to_data(const struct wim_reshdr *reshdr, WIMStruct *wim, void **buf_ret)
+wim_reshdr_to_data(const struct wim_reshdr *reshdr, WIMStruct *wim,
+                  void **buf_ret)
 {
        struct wim_resource_descriptor rdesc;
        struct blob_descriptor blob;
 
-       wim_res_hdr_to_desc(reshdr, wim, &rdesc);
-       blob_set_is_located_in_nonsolid_wim_resource(&blob, &rdesc);
+       wim_reshdr_to_desc_and_blob(reshdr, wim, &rdesc, &blob);
 
-       return read_full_blob_into_alloc_buf(&blob, buf_ret);
+       return read_blob_into_alloc_buf(&blob, buf_ret);
 }
 
+/* Calculate the SHA-1 message digest of the uncompressed data of the specified
+ * WIM resource.  */
 int
 wim_reshdr_to_hash(const struct wim_reshdr *reshdr, WIMStruct *wim,
                   u8 hash[SHA1_HASH_SIZE])
@@ -925,19 +831,19 @@ wim_reshdr_to_hash(const struct wim_reshdr *reshdr, WIMStruct *wim,
        struct blob_descriptor blob;
        int ret;
 
-       wim_res_hdr_to_desc(reshdr, wim, &rdesc);
-       blob_set_is_located_in_nonsolid_wim_resource(&blob, &rdesc);
+       wim_reshdr_to_desc_and_blob(reshdr, wim, &rdesc, &blob);
        blob.unhashed = 1;
 
        ret = sha1_blob(&blob);
-       if (ret)
+       if (unlikely(ret))
                return ret;
+
        copy_hash(hash, blob.hash);
        return 0;
 }
 
 struct blobifier_context {
-       struct read_blob_list_callbacks cbs;
+       struct read_blob_callbacks cbs;
        struct blob_descriptor *cur_blob;
        struct blob_descriptor *next_blob;
        u64 cur_blob_offset;
@@ -955,35 +861,28 @@ next_blob(struct blob_descriptor *blob, size_t list_head_offset)
        return (struct blob_descriptor*)((u8*)cur->next - list_head_offset);
 }
 
-/* A consume_data_callback_t implementation that translates raw resource data
- * into blobs, calling the begin_blob, consume_chunk, and end_blob callback
- * functions as appropriate.  */
+/* A consume_chunk() implementation that translates raw resource data into
+ * blobs, calling the begin_blob, consume_chunk, and end_blob callbacks as
+ * appropriate.  */
 static int
 blobifier_cb(const void *chunk, size_t size, void *_ctx)
 {
        struct blobifier_context *ctx = _ctx;
        int ret;
 
-       DEBUG("%zu bytes passed to blobifier", size);
-
        wimlib_assert(ctx->cur_blob != NULL);
        wimlib_assert(size <= ctx->cur_blob->size - ctx->cur_blob_offset);
 
        if (ctx->cur_blob_offset == 0) {
-
                /* Starting a new blob.  */
-               DEBUG("Begin new blob (size=%"PRIu64").", ctx->cur_blob->size);
-
-               ret = (*ctx->cbs.begin_blob)(ctx->cur_blob,
-                                            ctx->cbs.begin_blob_ctx);
+               ret = call_begin_blob(ctx->cur_blob, &ctx->cbs);
                if (ret)
                        return ret;
        }
 
-       /* Consume the chunk.  */
-       ret = (*ctx->cbs.consume_chunk)(chunk, size,
-                                       ctx->cbs.consume_chunk_ctx);
        ctx->cur_blob_offset += size;
+
+       ret = call_consume_chunk(chunk, size, &ctx->cbs);
        if (ret)
                return ret;
 
@@ -992,9 +891,7 @@ blobifier_cb(const void *chunk, size_t size, void *_ctx)
 
                ctx->cur_blob_offset = 0;
 
-               DEBUG("End blob (size=%"PRIu64").", ctx->cur_blob->size);
-               ret = (*ctx->cbs.end_blob)(ctx->cur_blob, 0,
-                                          ctx->cbs.end_blob_ctx);
+               ret = call_end_blob(ctx->cur_blob, 0, &ctx->cbs);
                if (ret)
                        return ret;
 
@@ -1014,7 +911,7 @@ blobifier_cb(const void *chunk, size_t size, void *_ctx)
 struct hasher_context {
        SHA_CTX sha_ctx;
        int flags;
-       struct read_blob_list_callbacks cbs;
+       struct read_blob_callbacks cbs;
 };
 
 /* Callback for starting to read a blob while calculating its SHA-1 message
@@ -1026,26 +923,77 @@ hasher_begin_blob(struct blob_descriptor *blob, void *_ctx)
 
        sha1_init(&ctx->sha_ctx);
 
-       if (ctx->cbs.begin_blob == NULL)
-               return 0;
-       else
-               return (*ctx->cbs.begin_blob)(blob, ctx->cbs.begin_blob_ctx);
+       return call_begin_blob(blob, &ctx->cbs);
 }
 
-/* A consume_data_callback_t implementation that continues calculating the SHA-1
- * message digest of the blob being read, then optionally passes the data on to
- * another consume_data_callback_t implementation.  This allows checking the
- * SHA-1 message digest of a blob being extracted, for example.  */
+/* A consume_chunk() implementation that continues calculating the SHA-1 message
+ * digest of the blob being read, then optionally passes the data on to another
+ * consume_chunk() implementation.  This allows checking the SHA-1 message
+ * digest of a blob being extracted, for example.  */
 static int
 hasher_consume_chunk(const void *chunk, size_t size, void *_ctx)
 {
        struct hasher_context *ctx = _ctx;
 
        sha1_update(&ctx->sha_ctx, chunk, size);
-       if (ctx->cbs.consume_chunk == NULL)
-               return 0;
-       else
-               return (*ctx->cbs.consume_chunk)(chunk, size, ctx->cbs.consume_chunk_ctx);
+
+       return call_consume_chunk(chunk, size, &ctx->cbs);
+}
+
+static int
+report_sha1_mismatch_error(const struct blob_descriptor *blob,
+                          const u8 actual_hash[SHA1_HASH_SIZE])
+{
+       tchar expected_hashstr[SHA1_HASH_SIZE * 2 + 1];
+       tchar actual_hashstr[SHA1_HASH_SIZE * 2 + 1];
+
+       wimlib_assert(blob->blob_location != BLOB_NONEXISTENT);
+       wimlib_assert(blob->blob_location != BLOB_IN_ATTACHED_BUFFER);
+
+       sprint_hash(blob->hash, expected_hashstr);
+       sprint_hash(actual_hash, actual_hashstr);
+
+       if (blob_is_in_file(blob)) {
+               ERROR("A file was concurrently modified!\n"
+                     "        Path: \"%"TS"\"\n"
+                     "        Expected SHA-1: %"TS"\n"
+                     "        Actual SHA-1: %"TS"\n",
+                     blob_file_path(blob), expected_hashstr, actual_hashstr);
+               return WIMLIB_ERR_CONCURRENT_MODIFICATION_DETECTED;
+       } else if (blob->blob_location == BLOB_IN_WIM) {
+               const struct wim_resource_descriptor *rdesc = blob->rdesc;
+               ERROR("A WIM resource is corrupted!\n"
+                     "        WIM file: \"%"TS"\"\n"
+                     "        Blob uncompressed size: %"PRIu64"\n"
+                     "        Resource offset in WIM: %"PRIu64"\n"
+                     "        Resource uncompressed size: %"PRIu64"\n"
+                     "        Resource size in WIM: %"PRIu64"\n"
+                     "        Resource flags: 0x%x%"TS"\n"
+                     "        Resource compression type: %"TS"\n"
+                     "        Resource compression chunk size: %"PRIu32"\n"
+                     "        Expected SHA-1: %"TS"\n"
+                     "        Actual SHA-1: %"TS"\n",
+                     rdesc->wim->filename,
+                     blob->size,
+                     rdesc->offset_in_wim,
+                     rdesc->uncompressed_size,
+                     rdesc->size_in_wim,
+                     (unsigned int)rdesc->flags,
+                     (rdesc->is_pipable ? T(", pipable") : T("")),
+                     wimlib_get_compression_type_string(
+                                               rdesc->compression_type),
+                     rdesc->chunk_size,
+                     expected_hashstr, actual_hashstr);
+               return WIMLIB_ERR_INVALID_RESOURCE_HASH;
+       } else {
+               ERROR("File data was concurrently modified!\n"
+                     "        Location ID: %d\n"
+                     "        Expected SHA-1: %"TS"\n"
+                     "        Actual SHA-1: %"TS"\n",
+                     (int)blob->blob_location,
+                     expected_hashstr, actual_hashstr);
+               return WIMLIB_ERR_CONCURRENT_MODIFICATION_DETECTED;
+       }
 }
 
 /* Callback for finishing reading a blob while calculating its SHA-1 message
@@ -1057,7 +1005,7 @@ hasher_end_blob(struct blob_descriptor *blob, int status, void *_ctx)
        u8 hash[SHA1_HASH_SIZE];
        int ret;
 
-       if (status) {
+       if (unlikely(status)) {
                /* Error occurred; the full blob may not have been read.  */
                ret = status;
                goto out_next_cb;
@@ -1066,90 +1014,48 @@ hasher_end_blob(struct blob_descriptor *blob, int status, void *_ctx)
        /* Retrieve the final SHA-1 message digest.  */
        sha1_final(hash, &ctx->sha_ctx);
 
+       /* Set the SHA-1 message digest of the blob, or compare the calculated
+        * value with stored value.  */
        if (blob->unhashed) {
-               if (ctx->flags & COMPUTE_MISSING_BLOB_HASHES) {
-                       /* No SHA-1 message digest was previously present for the
-                        * blob.  Set it to the one just calculated.  */
-                       DEBUG("Set SHA-1 message digest for blob "
-                             "(size=%"PRIu64").", blob->size);
+               if (ctx->flags & COMPUTE_MISSING_BLOB_HASHES)
                        copy_hash(blob->hash, hash);
-               }
-       } else {
-               if (ctx->flags & VERIFY_BLOB_HASHES) {
-                       /* The blob already had a SHA-1 message digest present.
-                        * Verify that it is the same as the calculated value.
-                        */
-                       if (!hashes_equal(hash, blob->hash)) {
-                               if (wimlib_print_errors) {
-                                       tchar expected_hashstr[SHA1_HASH_SIZE * 2 + 1];
-                                       tchar actual_hashstr[SHA1_HASH_SIZE * 2 + 1];
-                                       sprint_hash(blob->hash, expected_hashstr);
-                                       sprint_hash(hash, actual_hashstr);
-                                       ERROR("The blob is corrupted!\n"
-                                             "        (Expected SHA-1=%"TS",\n"
-                                             "              got SHA-1=%"TS")",
-                                             expected_hashstr, actual_hashstr);
-                               }
-                               ret = WIMLIB_ERR_INVALID_RESOURCE_HASH;
-                               errno = EINVAL;
-                               goto out_next_cb;
-                       }
-                       DEBUG("SHA-1 message digest okay for "
-                             "blob (size=%"PRIu64").", blob->size);
-               }
+       } else if ((ctx->flags & VERIFY_BLOB_HASHES) &&
+                  unlikely(!hashes_equal(hash, blob->hash)))
+       {
+               ret = report_sha1_mismatch_error(blob, hash);
+               goto out_next_cb;
        }
        ret = 0;
 out_next_cb:
-       if (ctx->cbs.end_blob == NULL)
-               return ret;
-       else
-               return (*ctx->cbs.end_blob)(blob, ret, ctx->cbs.end_blob_ctx);
-}
-
-static int
-read_full_blob_with_cbs(struct blob_descriptor *blob,
-                       const struct read_blob_list_callbacks *cbs)
-{
-       int ret;
-
-       ret = (*cbs->begin_blob)(blob, cbs->begin_blob_ctx);
-       if (ret)
-               return ret;
-
-       ret = read_blob_prefix(blob, blob->size, cbs->consume_chunk,
-                              cbs->consume_chunk_ctx);
-
-       return (*cbs->end_blob)(blob, ret, cbs->end_blob_ctx);
+       return call_end_blob(blob, ret, &ctx->cbs);
 }
 
 /* Read the full data of the specified blob, passing the data into the specified
  * callbacks (all of which are optional) and either checking or computing the
  * SHA-1 message digest of the blob.  */
-static int
-read_full_blob_with_sha1(struct blob_descriptor *blob,
-                        const struct read_blob_list_callbacks *cbs)
+int
+read_blob_with_sha1(struct blob_descriptor *blob,
+                   const struct read_blob_callbacks *cbs)
 {
        struct hasher_context hasher_ctx = {
                .flags = VERIFY_BLOB_HASHES | COMPUTE_MISSING_BLOB_HASHES,
                .cbs = *cbs,
        };
-       struct read_blob_list_callbacks hasher_cbs = {
-               .begin_blob             = hasher_begin_blob,
-               .begin_blob_ctx         = &hasher_ctx,
-               .consume_chunk          = hasher_consume_chunk,
-               .consume_chunk_ctx      = &hasher_ctx,
-               .end_blob               = hasher_end_blob,
-               .end_blob_ctx           = &hasher_ctx,
+       struct read_blob_callbacks hasher_cbs = {
+               .begin_blob     = hasher_begin_blob,
+               .consume_chunk  = hasher_consume_chunk,
+               .end_blob       = hasher_end_blob,
+               .ctx            = &hasher_ctx,
        };
-       return read_full_blob_with_cbs(blob, &hasher_cbs);
+       return read_blob_with_cbs(blob, &hasher_cbs);
 }
 
 static int
 read_blobs_in_solid_resource(struct blob_descriptor *first_blob,
                             struct blob_descriptor *last_blob,
-                            u64 blob_count,
+                            size_t blob_count,
                             size_t list_head_offset,
-                            const struct read_blob_list_callbacks *sink_cbs)
+                            const struct read_blob_callbacks *sink_cbs)
 {
        struct data_range *ranges;
        bool ranges_malloced;
@@ -1158,28 +1064,22 @@ read_blobs_in_solid_resource(struct blob_descriptor *first_blob,
        int ret;
        u64 ranges_alloc_size;
 
-       DEBUG("Reading %"PRIu64" blobs combined in same WIM resource",
-             blob_count);
-
        /* Setup data ranges array (one range per blob to read); this way
         * read_compressed_wim_resource() does not need to be aware of blobs.
         */
 
-       ranges_alloc_size = blob_count * sizeof(ranges[0]);
+       ranges_alloc_size = (u64)blob_count * sizeof(ranges[0]);
 
-       if (unlikely((size_t)ranges_alloc_size != ranges_alloc_size)) {
-               ERROR("Too many blobs in one resource!");
-               return WIMLIB_ERR_NOMEM;
-       }
-       if (likely(ranges_alloc_size <= STACK_MAX)) {
+       if (unlikely((size_t)ranges_alloc_size != ranges_alloc_size))
+               goto oom;
+
+       if (ranges_alloc_size <= STACK_MAX) {
                ranges = alloca(ranges_alloc_size);
                ranges_malloced = false;
        } else {
                ranges = MALLOC(ranges_alloc_size);
-               if (ranges == NULL) {
-                       ERROR("Too many blobs in one resource!");
-                       return WIMLIB_ERR_NOMEM;
-               }
+               if (unlikely(!ranges))
+                       goto oom;
                ranges_malloced = true;
        }
 
@@ -1199,25 +1099,26 @@ read_blobs_in_solid_resource(struct blob_descriptor *first_blob,
                .final_blob             = last_blob,
                .list_head_offset       = list_head_offset,
        };
+       struct read_blob_callbacks cbs = {
+               .consume_chunk  = blobifier_cb,
+               .ctx            = &blobifier_ctx,
+       };
 
-       ret = read_compressed_wim_resource(first_blob->rdesc,
-                                          ranges,
-                                          blob_count,
-                                          blobifier_cb,
-                                          &blobifier_ctx);
+       ret = read_compressed_wim_resource(first_blob->rdesc, ranges,
+                                          blob_count, &cbs);
 
        if (ranges_malloced)
                FREE(ranges);
 
-       if (ret) {
-               if (blobifier_ctx.cur_blob_offset != 0) {
-                       ret = (*blobifier_ctx.cbs.end_blob)
-                               (blobifier_ctx.cur_blob,
-                                ret,
-                                blobifier_ctx.cbs.end_blob_ctx);
-               }
+       if (unlikely(ret && blobifier_ctx.cur_blob_offset != 0)) {
+               ret = call_end_blob(blobifier_ctx.cur_blob, ret,
+                                   &blobifier_ctx.cbs);
        }
        return ret;
+
+oom:
+       ERROR("Too many blobs in one resource!");
+       return WIMLIB_ERR_NOMEM;
 }
 
 /*
@@ -1230,8 +1131,8 @@ read_blobs_in_solid_resource(struct blob_descriptor *first_blob,
  * @blob_list
  *     List of blobs to read.
  * @list_head_offset
- *     Offset of the `struct list_head' within each `struct blob_descriptor' that makes up
- *     the @blob_list.
+ *     Offset of the `struct list_head' within each `struct blob_descriptor'
+ *     that makes up the @blob_list.
  * @cbs
  *     Callback functions to accept the blob data.
  * @flags
@@ -1259,19 +1160,18 @@ read_blobs_in_solid_resource(struct blob_descriptor *first_blob,
  * of the callback functions.
  */
 int
-read_blob_list(struct list_head *blob_list,
-              size_t list_head_offset,
-              const struct read_blob_list_callbacks *cbs,
-              int flags)
+read_blob_list(struct list_head *blob_list, size_t list_head_offset,
+              const struct read_blob_callbacks *cbs, int flags)
 {
        int ret;
        struct list_head *cur, *next;
        struct blob_descriptor *blob;
        struct hasher_context *hasher_ctx;
-       struct read_blob_list_callbacks *sink_cbs;
+       struct read_blob_callbacks *sink_cbs;
 
        if (!(flags & BLOB_LIST_ALREADY_SORTED)) {
-               ret = sort_blob_list_by_sequential_order(blob_list, list_head_offset);
+               ret = sort_blob_list_by_sequential_order(blob_list,
+                                                        list_head_offset);
                if (ret)
                        return ret;
        }
@@ -1283,16 +1183,14 @@ read_blob_list(struct list_head *blob_list,
                        .cbs    = *cbs,
                };
                sink_cbs = alloca(sizeof(*sink_cbs));
-               *sink_cbs = (struct read_blob_list_callbacks) {
-                       .begin_blob             = hasher_begin_blob,
-                       .begin_blob_ctx         = hasher_ctx,
-                       .consume_chunk          = hasher_consume_chunk,
-                       .consume_chunk_ctx      = hasher_ctx,
-                       .end_blob               = hasher_end_blob,
-                       .end_blob_ctx           = hasher_ctx,
+               *sink_cbs = (struct read_blob_callbacks) {
+                       .begin_blob     = hasher_begin_blob,
+                       .consume_chunk  = hasher_consume_chunk,
+                       .end_blob       = hasher_end_blob,
+                       .ctx            = hasher_ctx,
                };
        } else {
-               sink_cbs = (struct read_blob_list_callbacks*)cbs;
+               sink_cbs = (struct read_blob_callbacks *)cbs;
        }
 
        for (cur = blob_list->next, next = cur->next;
@@ -1306,7 +1204,7 @@ read_blob_list(struct list_head *blob_list,
                {
                        struct blob_descriptor *blob_next, *blob_last;
                        struct list_head *next2;
-                       u64 blob_count;
+                       size_t blob_count;
 
                        /* The next blob is a proper sub-sequence of a WIM
                         * resource.  See if there are other blobs in the same
@@ -1346,80 +1244,55 @@ read_blob_list(struct list_head *blob_list,
                        }
                }
 
-               ret = read_full_blob_with_cbs(blob, sink_cbs);
-               if (ret && ret != BEGIN_BLOB_STATUS_SKIP_BLOB)
+               ret = read_blob_with_cbs(blob, sink_cbs);
+               if (unlikely(ret && ret != BEGIN_BLOB_STATUS_SKIP_BLOB))
                        return ret;
        }
        return 0;
 }
 
-/*
- * Extract the first @size bytes of the specified blob.
- *
- * If @size specifies the full uncompressed size of the blob, then the SHA-1
- * message digest of the uncompressed blob is checked while being extracted.
- *
- * The uncompressed data of the blob is passed in chunks of unspecified size to
- * the @extract_chunk function, passing it @extract_chunk_arg.
- */
-int
-extract_blob(struct blob_descriptor *blob, u64 size,
-            consume_data_callback_t extract_chunk, void *extract_chunk_arg)
-{
-       wimlib_assert(size <= blob->size);
-       if (size == blob->size) {
-               /* Do SHA-1.  */
-               struct read_blob_list_callbacks cbs = {
-                       .consume_chunk          = extract_chunk,
-                       .consume_chunk_ctx      = extract_chunk_arg,
-               };
-               return read_full_blob_with_sha1(blob, &cbs);
-       } else {
-               /* Don't do SHA-1.  */
-               return read_blob_prefix(blob, size, extract_chunk,
-                                       extract_chunk_arg);
-       }
-}
-
-/* A consume_data_callback_t implementation that writes the chunk of data to a
- * file descriptor.  */
 static int
-extract_chunk_to_fd(const void *chunk, size_t size, void *_fd_p)
+extract_chunk_to_fd(const void *chunk, size_t size, void *_fd)
 {
-       struct filedes *fd = _fd_p;
-
+       struct filedes *fd = _fd;
        int ret = full_write(fd, chunk, size);
-       if (ret) {
+       if (unlikely(ret))
                ERROR_WITH_ERRNO("Error writing to file descriptor");
-               return ret;
-       }
-       return 0;
+       return ret;
 }
 
 /* Extract the first @size bytes of the specified blob to the specified file
- * descriptor.  */
+ * descriptor.  This does *not* check the SHA-1 message digest.  */
 int
-extract_blob_to_fd(struct blob_descriptor *blob, struct filedes *fd, u64 size)
+extract_blob_prefix_to_fd(struct blob_descriptor *blob, u64 size,
+                         struct filedes *fd)
 {
-       return extract_blob(blob, size, extract_chunk_to_fd, fd);
+       struct read_blob_callbacks cbs = {
+               .consume_chunk  = extract_chunk_to_fd,
+               .ctx            = fd,
+       };
+       return read_blob_prefix(blob, size, &cbs);
 }
 
 /* Extract the full uncompressed contents of the specified blob to the specified
- * file descriptor.  */
+ * file descriptor.  This checks the SHA-1 message digest.  */
 int
-extract_full_blob_to_fd(struct blob_descriptor *blob, struct filedes *fd)
+extract_blob_to_fd(struct blob_descriptor *blob, struct filedes *fd)
 {
-       return extract_blob_to_fd(blob, fd, blob->size);
+       struct read_blob_callbacks cbs = {
+               .consume_chunk  = extract_chunk_to_fd,
+               .ctx            = fd,
+       };
+       return read_blob_with_sha1(blob, &cbs);
 }
 
 /* Calculate the SHA-1 message digest of a blob and store it in @blob->hash.  */
 int
 sha1_blob(struct blob_descriptor *blob)
 {
-       wimlib_assert(blob->unhashed);
-       struct read_blob_list_callbacks cbs = {
+       struct read_blob_callbacks cbs = {
        };
-       return read_full_blob_with_sha1(blob, &cbs);
+       return read_blob_with_sha1(blob, &cbs);
 }
 
 /*
@@ -1428,8 +1301,8 @@ sha1_blob(struct blob_descriptor *blob)
  * Note: for solid resources some fields still need to be overridden.
  */
 void
-wim_res_hdr_to_desc(const struct wim_reshdr *reshdr, WIMStruct *wim,
-                   struct wim_resource_descriptor *rdesc)
+wim_reshdr_to_desc(const struct wim_reshdr *reshdr, WIMStruct *wim,
+                  struct wim_resource_descriptor *rdesc)
 {
        rdesc->wim = wim;
        rdesc->offset_in_wim = reshdr->offset_in_wim;
@@ -1447,19 +1320,22 @@ wim_res_hdr_to_desc(const struct wim_reshdr *reshdr, WIMStruct *wim,
        }
 }
 
-/* Convert a stand-alone resource descriptor to a WIM resource header.  */
+/*
+ * Convert the short WIM resource header @reshdr to a stand-alone WIM resource
+ * descriptor @rdesc, then set @blob to consist of that entire resource.  This
+ * should only be used for non-solid resources!
+ */
 void
-wim_res_desc_to_hdr(const struct wim_resource_descriptor *rdesc,
-                   struct wim_reshdr *reshdr)
+wim_reshdr_to_desc_and_blob(const struct wim_reshdr *reshdr, WIMStruct *wim,
+                           struct wim_resource_descriptor *rdesc,
+                           struct blob_descriptor *blob)
 {
-       reshdr->offset_in_wim     = rdesc->offset_in_wim;
-       reshdr->size_in_wim       = rdesc->size_in_wim;
-       reshdr->flags             = rdesc->flags;
-       reshdr->uncompressed_size = rdesc->uncompressed_size;
+       wim_reshdr_to_desc(reshdr, wim, rdesc);
+       blob->size = rdesc->uncompressed_size;
+       blob_set_is_located_in_wim_resource(blob, rdesc, 0);
 }
 
-/* Translates a WIM resource header from the on-disk format into an in-memory
- * format.  */
+/* Import a WIM resource header from the on-disk format.  */
 void
 get_wim_reshdr(const struct wim_reshdr_disk *disk_reshdr,
               struct wim_reshdr *reshdr)
@@ -1476,8 +1352,7 @@ get_wim_reshdr(const struct wim_reshdr_disk *disk_reshdr,
        reshdr->flags = disk_reshdr->flags;
 }
 
-/* Translates a WIM resource header from an in-memory format into the on-disk
- * format.  */
+/* Export a WIM resource header to the on-disk format.  */
 void
 put_wim_reshdr(const struct wim_reshdr *reshdr,
               struct wim_reshdr_disk *disk_reshdr)