Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 4 Jan 2012 23:45:11 +0000 (UTC)
From:      Aleksandr Rybalko <ray@FreeBSD.org>
To:        src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org
Subject:   svn commit: r229538 - head/usr.bin/mkulzma
Message-ID:  <201201042345.q04NjBBD004255@svn.freebsd.org>

next in thread | raw e-mail | index | archive | help
Author: ray
Date: Wed Jan  4 23:45:10 2012
New Revision: 229538
URL: http://svn.freebsd.org/changeset/base/229538

Log:
  mkulzma used to create lzma compressed images, just like mkuzip do.
  
  Approved by:	adrian (mentor)

Added:
  head/usr.bin/mkulzma/
  head/usr.bin/mkulzma/Makefile   (contents, props changed)
  head/usr.bin/mkulzma/mkulzma.8   (contents, props changed)
  head/usr.bin/mkulzma/mkulzma.c   (contents, props changed)

Added: head/usr.bin/mkulzma/Makefile
==============================================================================
--- /dev/null	00:00:00 1970	(empty, because file is newly added)
+++ head/usr.bin/mkulzma/Makefile	Wed Jan  4 23:45:10 2012	(r229538)
@@ -0,0 +1,8 @@
+# $FreeBSD$
+
+PROG=	mkulzma
+MAN=	mkulzma.8
+DPADD=	${LIBLZMA}
+LDADD=	-llzma
+
+.include <bsd.prog.mk>

Added: head/usr.bin/mkulzma/mkulzma.8
==============================================================================
--- /dev/null	00:00:00 1970	(empty, because file is newly added)
+++ head/usr.bin/mkulzma/mkulzma.8	Wed Jan  4 23:45:10 2012	(r229538)
@@ -0,0 +1,107 @@
+.\" ----------------------------------------------------------------------------
+.\" Derived from mkuzip.8 by Aleksandr Rybalko <ray@ddteam.net>
+.\" ----------------------------------------------------------------------------
+.\" "THE BEER-WARE LICENSE" (Revision 42):
+.\" <sobomax@FreeBSD.ORG> wrote this file. As long as you retain this notice you
+.\" can do whatever you want with this stuff. If we meet some day, and you think
+.\" this stuff is worth it, you can buy me a beer in return.       Maxim Sobolev
+.\" ----------------------------------------------------------------------------
+.\"
+.\" $FreeBSD$
+.\"
+.Dd March 17, 2006
+.Dt mkulzma 8
+.Os
+.Sh NAME
+.Nm mkulzma
+.Nd compress disk image for use with
+.Xr geom_uncompress 4
+class
+.Sh SYNOPSIS
+.Nm
+.Op Fl v
+.Op Fl o Ar outfile
+.Op Fl s Ar cluster_size
+.Ar infile
+.Sh DESCRIPTION
+The
+.Nm
+utility compresses a disk image file so that the
+.Xr geom_uncompress 4
+class will be able to decompress the resulting image at run-time.
+This allows for a significant reduction of size of disk image at
+the expense of some CPU time required to decompress the data each
+time it is read.
+The
+.Nm
+utility
+works in two phases:
+.Bl -enum
+.It
+An
+.Ar infile
+image is split into clusters; each cluster is compressed using liblzma.
+.It
+The resulting set of compressed clusters along with headers that allow
+locating each individual cluster is written to the output file.
+.El
+.Pp
+The options are:
+.Bl -tag -width indent
+.It Fl o Ar outfile
+Name of the output file
+.Ar outfile .
+The default is to use the input name with the suffix
+.Pa .ulzma .
+.It Fl s Ar cluster_size
+Split the image into clusters of
+.Ar cluster_size
+bytes, 16384 bytes by default.
+The
+.Ar cluster_size
+should be a multiple of 512 bytes.
+.It Fl v
+Display verbose messages.
+.El
+.Sh NOTES
+The compression ratio largely depends on the cluster size used.
+.\" The following two sentences are unclear: how can xz(1) be
+.\" used in a comparable fashion, and wouldn't a lzma-compressed
+.\" image suffer from larger cluster sizes as well?
+For large cluster sizes (16K and higher), typical compression ratios
+are only 1-2% less than those achieved with
+.Xr lzma 1 .
+However, it should be kept in mind that larger cluster
+sizes lead to higher overhead in the
+.Xr geom_uncompress 4
+class, as the class has to decompress the whole cluster even if
+only a few bytes from that cluster have to be read.
+.Pp
+The
+.Nm
+utility
+inserts a short shell script at the beginning of the generated image,
+which makes it possible to
+.Dq run
+the image just like any other shell script.
+The script tries to load the
+.Xr geom_uncompress 4
+class if it is not loaded, configure the image as an
+.Xr md 4
+disk device using
+.Xr mdconfig 8 ,
+and automatically mount it using
+.Xr mount_cd9660 8
+on the mount point provided as the first argument to the script.
+.Sh EXIT STATUS
+.Ex -std
+.Sh SEE ALSO
+.Xr lzma 1 ,
+.Xr geom 4 ,
+.Xr geom_uncompress 4 ,
+.Xr md 4 ,
+.Xr mdconfig 8 ,
+.Xr mount_cd9660 8
+.Sh AUTHORS
+.An Maxim Sobolev Aq sobomax@FreeBSD.org
+.An Aleksandr Rybalko Aq ray@ddteam.net

Added: head/usr.bin/mkulzma/mkulzma.c
==============================================================================
--- /dev/null	00:00:00 1970	(empty, because file is newly added)
+++ head/usr.bin/mkulzma/mkulzma.c	Wed Jan  4 23:45:10 2012	(r229538)
@@ -0,0 +1,330 @@
+/*
+ * ----------------------------------------------------------------------------
+ * Derived from mkuzip.c by Aleksandr Rybalko <ray@ddteam.net>
+ * ----------------------------------------------------------------------------
+ * "THE BEER-WARE LICENSE" (Revision 42):
+ * <sobomax@FreeBSD.ORG> wrote this file. As long as you retain this notice you
+ * can do whatever you want with this stuff. If we meet some day, and you think
+ * this stuff is worth it, you can buy me a beer in return.       Maxim Sobolev
+ * ----------------------------------------------------------------------------
+ *
+ * $FreeBSD$
+ *
+ */
+
+#include <sys/disk.h>
+#include <sys/endian.h>
+#include <sys/param.h>
+#include <sys/stat.h>
+#include <sys/uio.h>
+#include <netinet/in.h>
+#include <err.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+#include <lzma.h>
+
+#define CLSTSIZE	16384
+#define DEFAULT_SUFX	".ulzma"
+
+#define USED_BLOCKSIZE DEV_BSIZE
+
+#define CLOOP_MAGIC_LEN 128
+/* Format L3.0, since we move to XZ API */
+static char CLOOP_MAGIC_START[] =
+    "#!/bin/sh\n"
+    "#L3.0\n"
+    "n=uncompress\n"
+    "m=geom_$n\n"
+    "(kldstat -m $m 2>&-||kldload $m)>&-&&"
+	"mount_cd9660 /dev/`mdconfig -af $0`.$n $1\n"
+    "exit $?\n";
+
+static char *readblock(int, char *, u_int32_t);
+static void usage(void);
+static void *safe_malloc(size_t);
+static void cleanup(void);
+
+static char *cleanfile = NULL;
+
+int main(int argc, char **argv)
+{
+	char *iname, *oname, *obuf, *ibuf;
+	int fdr, fdw, i, opt, verbose, tmp;
+	struct iovec iov[2];
+	struct stat sb;
+	uint32_t destlen;
+	uint64_t offset;
+	uint64_t *toc;
+	lzma_filter filters[2];
+	lzma_options_lzma opt_lzma;
+	lzma_ret ret;
+	lzma_stream strm = LZMA_STREAM_INIT;
+	struct cloop_header {
+		char magic[CLOOP_MAGIC_LEN];    /* cloop magic */
+		uint32_t blksz;                 /* block size */
+		uint32_t nblocks;               /* number of blocks */
+	} hdr;
+
+	memset(&hdr, 0, sizeof(hdr));
+	hdr.blksz = CLSTSIZE;
+	strcpy(hdr.magic, CLOOP_MAGIC_START);
+	oname = NULL;
+	verbose = 0;
+
+	while((opt = getopt(argc, argv, "o:s:v")) != -1) {
+		switch(opt) {
+		case 'o':
+			oname = optarg;
+			break;
+
+		case 's':
+			tmp = atoi(optarg);
+			if (tmp <= 0) {
+				errx(1,
+				    "invalid cluster size specified: %s",
+				    optarg);
+				/* Not reached */
+			}
+			if (tmp % USED_BLOCKSIZE != 0) {
+				errx(1,
+				    "cluster size should be multiple of %d",
+				    USED_BLOCKSIZE);
+				/* Not reached */
+			}
+			if ( tmp > MAXPHYS) {
+				errx(1, "cluster size is too large");
+				    /* Not reached */
+			}
+			hdr.blksz = tmp;
+			break;
+
+		case 'v':
+			verbose = 1;
+			break;
+
+		default:
+			usage();
+			/* Not reached */
+		}
+	}
+	argc -= optind;
+	argv += optind;
+
+	if (argc != 1) {
+		usage();
+		/* Not reached */
+	}
+
+	iname = argv[0];
+	if (oname == NULL) {
+		asprintf(&oname, "%s%s", iname, DEFAULT_SUFX);
+		if (oname == NULL) {
+			err(1, "can't allocate memory");
+			/* Not reached */
+		}
+	}
+
+	obuf = safe_malloc(hdr.blksz*2);
+	ibuf = safe_malloc(hdr.blksz);
+
+	signal(SIGHUP, exit);
+	signal(SIGINT, exit);
+	signal(SIGTERM, exit);
+	signal(SIGXCPU, exit);
+	signal(SIGXFSZ, exit);
+	atexit(cleanup);
+
+	fdr = open(iname, O_RDONLY);
+	if (fdr < 0) {
+		err(1, "open(%s)", iname);
+		/* Not reached */
+	}
+	if (fstat(fdr, &sb) != 0) {
+		err(1, "fstat(%s)", iname);
+		/* Not reached */
+	}
+	if (S_ISCHR(sb.st_mode)) {
+		off_t ms;
+
+		if (ioctl(fdr, DIOCGMEDIASIZE, &ms) < 0) {
+			err(1, "ioctl(DIOCGMEDIASIZE)");
+			/* Not reached */
+		}
+		sb.st_size = ms;
+	} else if (!S_ISREG(sb.st_mode)) {
+		fprintf(stderr,
+		    "%s: not a character device or regular file\n",
+		    iname);
+		exit(1);
+	}
+	hdr.nblocks = sb.st_size / hdr.blksz;
+	if ((sb.st_size % hdr.blksz) != 0) {
+		if (verbose != 0)
+			fprintf(stderr, "file size is not multiple "
+			"of %d, padding data\n", hdr.blksz);
+		hdr.nblocks++;
+	}
+	toc = safe_malloc((hdr.nblocks + 1) * sizeof(*toc));
+
+	fdw = open(oname, O_WRONLY | O_TRUNC | O_CREAT,
+		   S_IRWXU | S_IRGRP | S_IXGRP | S_IROTH | S_IXOTH);
+	if (fdw < 0) {
+		err(1, "open(%s)", oname);
+		/* Not reached */
+	}
+	cleanfile = oname;
+
+	/*
+	 * Prepare header that we will write later when we have index ready.
+	 */
+	iov[0].iov_base = (char *)&hdr;
+	iov[0].iov_len = sizeof(hdr);
+	iov[1].iov_base = (char *)toc;
+	iov[1].iov_len = (hdr.nblocks + 1) * sizeof(*toc);
+	offset = iov[0].iov_len + iov[1].iov_len;
+
+	/* Reserve space for header */
+	lseek(fdw, offset, SEEK_SET);
+
+	if (verbose != 0)
+		fprintf(stderr, "data size %ju bytes, number of clusters "
+		    "%u, index length %zu bytes\n", sb.st_size,
+		    hdr.nblocks, iov[1].iov_len);
+
+	/* Init lzma encoder */
+	if (lzma_lzma_preset(&opt_lzma, LZMA_PRESET_DEFAULT))
+		errx(1, "Error loading LZMA preset");
+
+	filters[0].id = LZMA_FILTER_LZMA2;
+	filters[0].options = &opt_lzma;
+	filters[1].id = LZMA_VLI_UNKNOWN;
+
+	for(i = 0; i == 0 || ibuf != NULL; i++) {
+		ibuf = readblock(fdr, ibuf, hdr.blksz);
+		if (ibuf != NULL) {
+			destlen = hdr.blksz*2;
+
+			ret = lzma_stream_encoder(&strm, filters,
+			    LZMA_CHECK_CRC32);
+			if (ret != LZMA_OK) {
+				if (ret == LZMA_MEMLIMIT_ERROR)
+					errx(1, "can't compress data: "
+					    "LZMA_MEMLIMIT_ERROR");
+
+				errx(1, "can't compress data: "
+				    "LZMA compressor ERROR");
+			}
+
+			strm.next_in = ibuf;
+			strm.avail_in = hdr.blksz;
+			strm.next_out = obuf;
+			strm.avail_out = hdr.blksz*2;
+
+			ret = lzma_code(&strm, LZMA_FINISH);
+
+			if (ret != LZMA_STREAM_END) {
+				/* Error */
+				errx(1, "lzma_code FINISH failed, code=%d, "
+				    "pos(in=%zd, out=%zd)",
+				    ret,
+				    (hdr.blksz - strm.avail_in),
+				    (hdr.blksz*2 - strm.avail_out));
+			}
+
+			destlen -= strm.avail_out;
+
+			lzma_end(&strm);
+
+			if (verbose != 0)
+				fprintf(stderr, "cluster #%d, in %u bytes, "
+				    "out %u bytes\n", i, hdr.blksz, destlen);
+		} else {
+			destlen = USED_BLOCKSIZE - (offset % USED_BLOCKSIZE);
+			memset(obuf, 0, destlen);
+			if (verbose != 0)
+				fprintf(stderr, "padding data with %u bytes"
+				    " so that file size is multiple of %d\n",
+				    destlen,
+				    USED_BLOCKSIZE);
+		}
+		if (write(fdw, obuf, destlen) < 0) {
+			err(1, "write(%s)", oname);
+			/* Not reached */
+		}
+		toc[i] = htobe64(offset);
+		offset += destlen;
+	}
+	close(fdr);
+
+	if (verbose != 0)
+		fprintf(stderr, "compressed data to %ju bytes, saved %lld "
+		    "bytes, %.2f%% decrease.\n", offset,
+		    (long long)(sb.st_size - offset),
+		    100.0 * (long long)(sb.st_size - offset) /
+		    (float)sb.st_size);
+
+	/* Convert to big endian */
+	hdr.blksz = htonl(hdr.blksz);
+	hdr.nblocks = htonl(hdr.nblocks);
+	/* Write headers into pre-allocated space */
+	lseek(fdw, 0, SEEK_SET);
+	if (writev(fdw, iov, 2) < 0) {
+		err(1, "writev(%s)", oname);
+		/* Not reached */
+	}
+	cleanfile = NULL;
+	close(fdw);
+
+	exit(0);
+}
+
+static char *
+readblock(int fd, char *ibuf, u_int32_t clstsize)
+{
+	int numread;
+
+	bzero(ibuf, clstsize);
+	numread = read(fd, ibuf, clstsize);
+	if (numread < 0) {
+		err(1, "read() failed");
+		/* Not reached */
+	}
+	if (numread == 0) {
+		return NULL;
+	}
+	return ibuf;
+}
+
+static void
+usage(void)
+{
+
+	fprintf(stderr, "usage: mkulzma [-v] [-o outfile] [-s cluster_size] "
+	    "infile\n");
+	exit(1);
+}
+
+static void *
+safe_malloc(size_t size)
+{
+	void *retval;
+
+	retval = malloc(size);
+	if (retval == NULL) {
+		err(1, "can't allocate memory");
+		/* Not reached */
+	}
+	return retval;
+}
+
+static void
+cleanup(void)
+{
+
+	if (cleanfile != NULL)
+		unlink(cleanfile);
+}



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201201042345.q04NjBBD004255>