Created
July 22, 2024 17:45
-
-
Save alexshpilkin/a86a34f23744303678e9db140f152680 to your computer and use it in GitHub Desktop.
Pack sparse files into Zstandard-compatible byte streams
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
/* SPDX-License-Identifier: CC0-1.0 */ | |
#define _GNU_SOURCE /* for copy_file_range */ | |
#include <errno.h> | |
#include <fcntl.h> | |
#include <stdarg.h> | |
#include <stdlib.h> | |
#include <string.h> | |
#include <sysexits.h> | |
#include <unistd.h> | |
static void die(int r, ...) { | |
static const char colon[2] = ": ", newline[1] = "\n"; | |
int e; va_list ap; const char *s; | |
e = errno; | |
va_start(ap, r); | |
while ((s = va_arg(ap, const char *))) { | |
write(STDERR_FILENO, s, strlen(s)); | |
write(STDERR_FILENO, colon, sizeof colon); | |
} | |
va_end(ap); | |
s = strerror(e); | |
write(STDERR_FILENO, s, strlen(s)); | |
write(STDERR_FILENO, newline, sizeof newline); | |
exit(r); | |
} | |
#define SUFFIX ".zst" | |
static const char *progname = "<unknown>"; | |
int main(int argc, char **argv) { | |
static struct { | |
unsigned char magic[4]; | |
unsigned char fdesc; | |
unsigned char wdesc; | |
unsigned char size[8]; | |
} header = { | |
{ 0x28, 0xB5, 0x2F, 0xFD }, | |
}; | |
char *filename = NULL; | |
int ifd, ofd; | |
off_t off, len, nxt = 0; | |
size_t n; | |
if (argc > 0) { | |
const char *p = argv[0]; | |
while (*p) if (*p++ == '/') progname = p; | |
} | |
if (argc > 1 && (argv[1][0] != '-' || argv[1][1] != 0)) | |
filename = argv[1]; | |
if (argc > 2) { | |
static const char prefix[7] = "usage: "; | |
static const char suffix[8] = " [FILE]\n"; | |
write(STDERR_FILENO, prefix, sizeof prefix); | |
write(STDERR_FILENO, progname, strlen(progname)); | |
write(STDERR_FILENO, suffix, sizeof suffix); | |
return EX_USAGE; | |
} | |
/* Open input and output files */ | |
if (!filename) { | |
filename = "<stdin>"; | |
ifd = STDIN_FILENO; | |
ofd = STDOUT_FILENO; | |
} else { | |
size_t slash = -1; | |
int dfd = AT_FDCWD; | |
char c, *basename; | |
for (n = 0; filename[n]; n++) if (filename[n] == '/') slash = n; | |
if (slash != -1) { | |
c = filename[slash+1]; filename[slash+1] = 0; | |
#if defined O_SEARCH /* POSIX */ | |
dfd = open(filename, O_SEARCH | O_DIRECTORY); | |
#elif defined O_PATH /* Linux/Glibc */ | |
dfd = open(filename, O_PATH | O_DIRECTORY); | |
#else | |
#error "no *at() functions" | |
#endif | |
if (dfd < 0) | |
die(EX_NOINPUT, filename, "open", NULL); | |
filename[slash+1] = c; | |
} | |
filename += slash + 1; n -= slash + 1; | |
if ((ifd = openat(dfd, filename, O_RDONLY)) < 0) | |
die(EX_NOINPUT, filename, "openat", NULL); | |
basename = malloc(n + sizeof SUFFIX); | |
if (!basename) | |
die(EX_OSERR, progname, "malloc", NULL); | |
memcpy(basename, filename, n); | |
memcpy(basename + n, SUFFIX, sizeof SUFFIX); | |
if ((ofd = openat(dfd, basename, O_WRONLY | O_CREAT | O_EXCL, 0666)) < 0) | |
die(EX_CANTCREAT, basename, "openat", NULL); | |
} | |
/* Write Zstandard header */ | |
if ((off = len = lseek(ifd, 0, SEEK_END)) < 0) | |
die(EX_IOERR, filename, "lseek", NULL); | |
n = (char *)&header.size[0] - (char *)&header; // FIXME | |
if (off < 0x100) { | |
header.fdesc = 1 << 5; /* Single_Segment_Flag */ | |
header.wdesc = off; /* 1-byte length instead of Window_Size */ | |
} else if (len <= 0x10000) { | |
off -= 0x100; /* 2-byte lengths (only) are biased */ | |
} | |
if (off >> 8) { | |
n += 2; | |
header.fdesc = 0x40; /* Frame_Content_Size_Flag */ | |
header.wdesc = (16 - 10) << 3; /* Window_Size = 2**16 */ | |
header.size[0] = off & 0xFF; | |
header.size[1] = off >> 8 & 0xFF; | |
} | |
if (off >> 16) { | |
n += 2; | |
header.fdesc += 0x40; | |
header.wdesc = (17 - 10) << 3; /* Window_Size = 2**17 */ | |
header.size[2] = off >> 16 & 0xFF; | |
header.size[3] = off >> 24 & 0xFF; | |
} | |
if (off >> 32) { | |
n += 4; | |
header.fdesc += 0x40; | |
header.size[4] = off >> 32 & 0xFF; | |
header.size[5] = off >> 40 & 0xFF; | |
header.size[6] = off >> 48 & 0xFF; | |
header.size[7] = off >> 56 & 0xFF; | |
} | |
if (write(ofd, &header, n) != n) | |
die(EX_IOERR, "write", NULL); // FIXME basename? | |
/* Write alternating data and holes */ | |
if ((off = lseek(ifd, 0, SEEK_SET)) < 0) | |
die(EX_IOERR, filename, "lseek", NULL); | |
if (off == len) | |
goto writedata; /* write a single zero-length data chunk */ | |
for (;;) { | |
long chunk; | |
unsigned long hdata; | |
unsigned char hbyte[4]; | |
/* Write raw block until start of hole */ | |
nxt = lseek(ifd, 0, SEEK_HOLE); | |
if (nxt < 0) | |
die(EX_IOERR, filename, "lseek", NULL); | |
while (nxt != off) writedata: { | |
chunk = nxt - off < 1L << 17 ? nxt - off : 1L << 17; | |
hdata = chunk << 3; | |
if (off + chunk == len) hdata |= 1; | |
hbyte[0] = hdata & 0xFF; | |
hbyte[1] = hdata >> 8 & 0xFF; | |
hbyte[2] = hdata >> 16 & 0xFF; | |
if (write(ofd, hbyte, 3) != 3) | |
die(EX_IOERR, "write", NULL); | |
if (lseek(ifd, off, SEEK_SET) < 0) | |
die(EX_IOERR, filename, "lseek", NULL); | |
if (copy_file_range(ifd, NULL, ofd, NULL, chunk, 0) != chunk) | |
die(EX_IOERR, "copy_file_range", NULL); | |
off += chunk; | |
} | |
if (off == len) break; | |
/* Write RLE block until end of hole */ | |
nxt = lseek(ifd, 0, SEEK_DATA); | |
if (nxt < 0) { | |
if (errno != ENXIO) | |
die(EX_IOERR, filename, "lseek", NULL); | |
nxt = len; | |
} | |
while (nxt != off) { | |
chunk = nxt - off < 1L << 17 ? nxt - off : 1L << 17; | |
hdata = chunk << 3 | 2; | |
if (off + chunk == len) hdata |= 1; | |
hbyte[0] = hdata & 0xFF; | |
hbyte[1] = hdata >> 8 & 0xFF; | |
hbyte[2] = hdata >> 16 & 0xFF; | |
hbyte[3] = 0; /* RLE payload */ | |
if (write(ofd, hbyte, 4) != 4) | |
die(EX_IOERR, "write", NULL); | |
off += chunk; | |
} | |
if (off == len) break; | |
} | |
return 0; | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment