4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 1) // SPDX-License-Identifier: GPL-2.0
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 2)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 3) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 4) * Important notes about in-place decompression
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 5) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 6) * At least on x86, the kernel is decompressed in place: the compressed data
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 7) * is placed to the end of the output buffer, and the decompressor overwrites
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 8) * most of the compressed data. There must be enough safety margin to
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 9) * guarantee that the write position is always behind the read position.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 10) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 11) * The safety margin for ZSTD with a 128 KB block size is calculated below.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 12) * Note that the margin with ZSTD is bigger than with GZIP or XZ!
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 13) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 14) * The worst case for in-place decompression is that the beginning of
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 15) * the file is compressed extremely well, and the rest of the file is
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 16) * uncompressible. Thus, we must look for worst-case expansion when the
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 17) * compressor is encoding uncompressible data.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 18) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 19) * The structure of the .zst file in case of a compresed kernel is as follows.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 20) * Maximum sizes (as bytes) of the fields are in parenthesis.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 21) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 22) * Frame Header: (18)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 23) * Blocks: (N)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 24) * Checksum: (4)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 25) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 26) * The frame header and checksum overhead is at most 22 bytes.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 27) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 28) * ZSTD stores the data in blocks. Each block has a header whose size is
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 29) * a 3 bytes. After the block header, there is up to 128 KB of payload.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 30) * The maximum uncompressed size of the payload is 128 KB. The minimum
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 31) * uncompressed size of the payload is never less than the payload size
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 32) * (excluding the block header).
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 33) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 34) * The assumption, that the uncompressed size of the payload is never
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 35) * smaller than the payload itself, is valid only when talking about
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 36) * the payload as a whole. It is possible that the payload has parts where
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 37) * the decompressor consumes more input than it produces output. Calculating
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 38) * the worst case for this would be tricky. Instead of trying to do that,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 39) * let's simply make sure that the decompressor never overwrites any bytes
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 40) * of the payload which it is currently reading.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 41) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 42) * Now we have enough information to calculate the safety margin. We need
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 43) * - 22 bytes for the .zst file format headers;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 44) * - 3 bytes per every 128 KiB of uncompressed size (one block header per
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 45) * block); and
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 46) * - 128 KiB (biggest possible zstd block size) to make sure that the
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 47) * decompressor never overwrites anything from the block it is currently
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 48) * reading.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 49) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 50) * We get the following formula:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 51) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 52) * safety_margin = 22 + uncompressed_size * 3 / 131072 + 131072
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 53) * <= 22 + (uncompressed_size >> 15) + 131072
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 54) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 55)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 56) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 57) * Preboot environments #include "path/to/decompress_unzstd.c".
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 58) * All of the source files we depend on must be #included.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 59) * zstd's only source dependeny is xxhash, which has no source
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 60) * dependencies.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 61) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 62) * When UNZSTD_PREBOOT is defined we declare __decompress(), which is
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 63) * used for kernel decompression, instead of unzstd().
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 64) *
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 65) * Define __DISABLE_EXPORTS in preboot environments to prevent symbols
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 66) * from xxhash and zstd from being exported by the EXPORT_SYMBOL macro.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 67) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 68) #ifdef STATIC
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 69) # define UNZSTD_PREBOOT
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 70) # include "xxhash.c"
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 71) # include "zstd/entropy_common.c"
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 72) # include "zstd/fse_decompress.c"
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 73) # include "zstd/huf_decompress.c"
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 74) # include "zstd/zstd_common.c"
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 75) # include "zstd/decompress.c"
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 76) #endif
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 77)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 78) #include <linux/decompress/mm.h>
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 79) #include <linux/kernel.h>
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 80) #include <linux/zstd.h>
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 81)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 82) /* 128MB is the maximum window size supported by zstd. */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 83) #define ZSTD_WINDOWSIZE_MAX (1 << ZSTD_WINDOWLOG_MAX)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 84) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 85) * Size of the input and output buffers in multi-call mode.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 86) * Pick a larger size because it isn't used during kernel decompression,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 87) * since that is single pass, and we have to allocate a large buffer for
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 88) * zstd's window anyway. The larger size speeds up initramfs decompression.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 89) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 90) #define ZSTD_IOBUF_SIZE (1 << 17)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 91)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 92) static int INIT handle_zstd_error(size_t ret, void (*error)(char *x))
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 93) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 94) const int err = ZSTD_getErrorCode(ret);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 95)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 96) if (!ZSTD_isError(ret))
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 97) return 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 98)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 99) switch (err) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 100) case ZSTD_error_memory_allocation:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 101) error("ZSTD decompressor ran out of memory");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 102) break;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 103) case ZSTD_error_prefix_unknown:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 104) error("Input is not in the ZSTD format (wrong magic bytes)");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 105) break;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 106) case ZSTD_error_dstSize_tooSmall:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 107) case ZSTD_error_corruption_detected:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 108) case ZSTD_error_checksum_wrong:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 109) error("ZSTD-compressed data is corrupt");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 110) break;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 111) default:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 112) error("ZSTD-compressed data is probably corrupt");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 113) break;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 114) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 115) return -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 116) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 117)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 118) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 119) * Handle the case where we have the entire input and output in one segment.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 120) * We can allocate less memory (no circular buffer for the sliding window),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 121) * and avoid some memcpy() calls.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 122) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 123) static int INIT decompress_single(const u8 *in_buf, long in_len, u8 *out_buf,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 124) long out_len, long *in_pos,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 125) void (*error)(char *x))
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 126) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 127) const size_t wksp_size = ZSTD_DCtxWorkspaceBound();
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 128) void *wksp = large_malloc(wksp_size);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 129) ZSTD_DCtx *dctx = ZSTD_initDCtx(wksp, wksp_size);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 130) int err;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 131) size_t ret;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 132)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 133) if (dctx == NULL) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 134) error("Out of memory while allocating ZSTD_DCtx");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 135) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 136) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 137) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 138) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 139) * Find out how large the frame actually is, there may be junk at
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 140) * the end of the frame that ZSTD_decompressDCtx() can't handle.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 141) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 142) ret = ZSTD_findFrameCompressedSize(in_buf, in_len);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 143) err = handle_zstd_error(ret, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 144) if (err)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 145) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 146) in_len = (long)ret;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 147)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 148) ret = ZSTD_decompressDCtx(dctx, out_buf, out_len, in_buf, in_len);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 149) err = handle_zstd_error(ret, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 150) if (err)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 151) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 152)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 153) if (in_pos != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 154) *in_pos = in_len;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 155)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 156) err = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 157) out:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 158) if (wksp != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 159) large_free(wksp);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 160) return err;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 161) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 162)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 163) static int INIT __unzstd(unsigned char *in_buf, long in_len,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 164) long (*fill)(void*, unsigned long),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 165) long (*flush)(void*, unsigned long),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 166) unsigned char *out_buf, long out_len,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 167) long *in_pos,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 168) void (*error)(char *x))
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 169) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 170) ZSTD_inBuffer in;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 171) ZSTD_outBuffer out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 172) ZSTD_frameParams params;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 173) void *in_allocated = NULL;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 174) void *out_allocated = NULL;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 175) void *wksp = NULL;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 176) size_t wksp_size;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 177) ZSTD_DStream *dstream;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 178) int err;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 179) size_t ret;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 180)
1c4dd334df3a0 (Paul Cercueil 2020-09-01 16:26:50 +0200 181) /*
1c4dd334df3a0 (Paul Cercueil 2020-09-01 16:26:50 +0200 182) * ZSTD decompression code won't be happy if the buffer size is so big
1c4dd334df3a0 (Paul Cercueil 2020-09-01 16:26:50 +0200 183) * that its end address overflows. When the size is not provided, make
1c4dd334df3a0 (Paul Cercueil 2020-09-01 16:26:50 +0200 184) * it as big as possible without having the end address overflow.
1c4dd334df3a0 (Paul Cercueil 2020-09-01 16:26:50 +0200 185) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 186) if (out_len == 0)
1c4dd334df3a0 (Paul Cercueil 2020-09-01 16:26:50 +0200 187) out_len = UINTPTR_MAX - (uintptr_t)out_buf;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 188)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 189) if (fill == NULL && flush == NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 190) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 191) * We can decompress faster and with less memory when we have a
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 192) * single chunk.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 193) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 194) return decompress_single(in_buf, in_len, out_buf, out_len,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 195) in_pos, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 196)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 197) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 198) * If in_buf is not provided, we must be using fill(), so allocate
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 199) * a large enough buffer. If it is provided, it must be at least
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 200) * ZSTD_IOBUF_SIZE large.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 201) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 202) if (in_buf == NULL) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 203) in_allocated = large_malloc(ZSTD_IOBUF_SIZE);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 204) if (in_allocated == NULL) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 205) error("Out of memory while allocating input buffer");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 206) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 207) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 208) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 209) in_buf = in_allocated;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 210) in_len = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 211) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 212) /* Read the first chunk, since we need to decode the frame header. */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 213) if (fill != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 214) in_len = fill(in_buf, ZSTD_IOBUF_SIZE);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 215) if (in_len < 0) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 216) error("ZSTD-compressed data is truncated");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 217) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 218) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 219) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 220) /* Set the first non-empty input buffer. */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 221) in.src = in_buf;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 222) in.pos = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 223) in.size = in_len;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 224) /* Allocate the output buffer if we are using flush(). */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 225) if (flush != NULL) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 226) out_allocated = large_malloc(ZSTD_IOBUF_SIZE);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 227) if (out_allocated == NULL) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 228) error("Out of memory while allocating output buffer");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 229) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 230) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 231) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 232) out_buf = out_allocated;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 233) out_len = ZSTD_IOBUF_SIZE;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 234) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 235) /* Set the output buffer. */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 236) out.dst = out_buf;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 237) out.pos = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 238) out.size = out_len;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 239)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 240) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 241) * We need to know the window size to allocate the ZSTD_DStream.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 242) * Since we are streaming, we need to allocate a buffer for the sliding
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 243) * window. The window size varies from 1 KB to ZSTD_WINDOWSIZE_MAX
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 244) * (8 MB), so it is important to use the actual value so as not to
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 245) * waste memory when it is smaller.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 246) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 247) ret = ZSTD_getFrameParams(¶ms, in.src, in.size);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 248) err = handle_zstd_error(ret, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 249) if (err)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 250) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 251) if (ret != 0) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 252) error("ZSTD-compressed data has an incomplete frame header");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 253) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 254) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 255) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 256) if (params.windowSize > ZSTD_WINDOWSIZE_MAX) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 257) error("ZSTD-compressed data has too large a window size");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 258) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 259) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 260) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 261)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 262) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 263) * Allocate the ZSTD_DStream now that we know how much memory is
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 264) * required.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 265) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 266) wksp_size = ZSTD_DStreamWorkspaceBound(params.windowSize);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 267) wksp = large_malloc(wksp_size);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 268) dstream = ZSTD_initDStream(params.windowSize, wksp, wksp_size);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 269) if (dstream == NULL) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 270) error("Out of memory while allocating ZSTD_DStream");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 271) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 272) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 273) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 274)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 275) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 276) * Decompression loop:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 277) * Read more data if necessary (error if no more data can be read).
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 278) * Call the decompression function, which returns 0 when finished.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 279) * Flush any data produced if using flush().
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 280) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 281) if (in_pos != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 282) *in_pos = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 283) do {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 284) /*
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 285) * If we need to reload data, either we have fill() and can
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 286) * try to get more data, or we don't and the input is truncated.
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 287) */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 288) if (in.pos == in.size) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 289) if (in_pos != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 290) *in_pos += in.pos;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 291) in_len = fill ? fill(in_buf, ZSTD_IOBUF_SIZE) : -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 292) if (in_len < 0) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 293) error("ZSTD-compressed data is truncated");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 294) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 295) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 296) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 297) in.pos = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 298) in.size = in_len;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 299) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 300) /* Returns zero when the frame is complete. */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 301) ret = ZSTD_decompressStream(dstream, &out, &in);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 302) err = handle_zstd_error(ret, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 303) if (err)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 304) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 305) /* Flush all of the data produced if using flush(). */
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 306) if (flush != NULL && out.pos > 0) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 307) if (out.pos != flush(out.dst, out.pos)) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 308) error("Failed to flush()");
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 309) err = -1;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 310) goto out;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 311) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 312) out.pos = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 313) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 314) } while (ret != 0);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 315)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 316) if (in_pos != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 317) *in_pos += in.pos;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 318)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 319) err = 0;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 320) out:
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 321) if (in_allocated != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 322) large_free(in_allocated);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 323) if (out_allocated != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 324) large_free(out_allocated);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 325) if (wksp != NULL)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 326) large_free(wksp);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 327) return err;
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 328) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 329)
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 330) #ifndef UNZSTD_PREBOOT
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 331) STATIC int INIT unzstd(unsigned char *buf, long len,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 332) long (*fill)(void*, unsigned long),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 333) long (*flush)(void*, unsigned long),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 334) unsigned char *out_buf,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 335) long *pos,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 336) void (*error)(char *x))
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 337) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 338) return __unzstd(buf, len, fill, flush, out_buf, 0, pos, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 339) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 340) #else
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 341) STATIC int INIT __decompress(unsigned char *buf, long len,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 342) long (*fill)(void*, unsigned long),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 343) long (*flush)(void*, unsigned long),
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 344) unsigned char *out_buf, long out_len,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 345) long *pos,
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 346) void (*error)(char *x))
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 347) {
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 348) return __unzstd(buf, len, fill, flush, out_buf, out_len, pos, error);
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 349) }
4963bb2b89884 (Nick Terrell 2020-07-30 12:08:35 -0700 350) #endif