bsync: a write-minimizing block device synchronizing utility

bsync.c

#include <stdbool.h>
#include <stdint.h>
#include <stdio.h>
#include <string.h>

#include <errno.h>
#include <error.h>
#include <fcntl.h>
#include <sysexits.h>
#include <unistd.h>
#include <sys/mman.h>


static bool all_balls(const void *buf, size_t size) {
	for (const char *p = buf, *q = p + size; p < q; ++p) {
		if (*p) {
			return false;
		}
	}
	return true;
}

int main(int argc, char *argv[]) {
	if (argc != 3) {
		fprintf(stderr, "usage: %s <input> <output>\n", argv[0]);
		return -1;
	}
	close(STDIN_FILENO);
	close(STDOUT_FILENO);
	int in_fd = open(argv[1], O_RDONLY);
	if (in_fd < 0) {
		error(EX_NOINPUT, errno, "%s: open", argv[1]);
	}
	int out_fd = open(argv[2], O_RDWR);
	if (out_fd < 0) {
		error(EX_NOINPUT, errno, "%s: open", argv[2]);
	}
	off_t size = lseek(in_fd, 0, SEEK_END);
	if (size < 0) {
		error(EX_OSERR, errno, "%s: lseek", argv[1]);
	}
	off_t out_size = lseek(out_fd, 0, SEEK_END);
	if (out_size < 0) {
		error(EX_OSERR, errno, "%s: lseek", argv[2]);
	}
	if (size != out_size) {
		fputs("input and output sizes must match\n", stderr);
		return EX_DATAERR;
	}
	long page_size = sysconf(_SC_PAGESIZE);
	if (page_size < 0) {
		perror("sysconf");
		return EX_OSERR;
	}
	if (size % page_size != 0) {
		fputs("size must be a multiple of page size\n", stderr);
		return EX_DATAERR;
	}
	const void *in = mmap(NULL, size, PROT_READ, MAP_SHARED, in_fd, 0);
	if (!~(uintptr_t) in) {
		error(EX_OSERR, errno, "%s: mmap", argv[1]);
	}
	void *out = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, out_fd, 0);
	if (!~(uintptr_t) out) {
		error(EX_OSERR, errno, "%s: mmap", argv[2]);
	}
	madvise((void *) in, size, MADV_SEQUENTIAL);
	madvise(out, size, MADV_SEQUENTIAL);
	const void *in_tail = in;
	void *out_tail = out;
	bool try_discard = true;
	void *discard_begin = out, *discard_end = out;
	for (off_t rem = size; rem > 0;) {
		if (memcmp(out, in, page_size) != 0) {
			if (try_discard && all_balls(in, page_size)) {
				if (discard_end != out) {
					size_t discard_size = discard_end - discard_begin;
					if (discard_size > 0 && madvise(discard_begin, discard_size, MADV_REMOVE) != 0) {
						memset(discard_begin, 0, discard_size);
						try_discard = false;
					}
					discard_begin = out;
				}
				discard_end = (char *) out + page_size;
			}
			else {
				memcpy(out, in, page_size);
			}
		}
		in = (const char *) in + page_size;
		out = (char *) out + page_size;
		rem -= page_size;
		if ((rem & (1 << 25) - 1) == 0) {
			madvise((void *) in_tail, in - in_tail, MADV_DONTNEED), in_tail = in;
			madvise(out_tail, out - out_tail, MADV_DONTNEED), out_tail = out;
			size_t advice_len = rem < 2 << 25 ? (size_t) rem : 2 << 25;
			madvise((void *) in, advice_len, MADV_WILLNEED);
			madvise(out, advice_len, MADV_WILLNEED);
			fprintf(stderr, "\r%.1f%%", (double) (size - rem) / (double) size * 100);
			fflush(stderr);
		}
	}
	size_t discard_size = discard_end - discard_begin;
	if (discard_size > 0 && madvise(discard_begin, discard_size, MADV_REMOVE) != 0) {
		memset(discard_begin, 0, discard_size);
	}
	putc('\n', stderr);
	return 0;
}