freebsd-dev/usr.bin/mkuzip/mkuzip.c

385 lines
8.8 KiB
C
Raw Normal View History

/*
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
* Copyright (c) 2004-2016 Maxim Sobolev <sobomax@FreeBSD.org>
* All rights reserved.
*
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
*/
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/types.h>
#include <sys/disk.h>
#include <sys/endian.h>
#include <sys/param.h>
#include <sys/stat.h>
#include <sys/uio.h>
#include <netinet/in.h>
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
#include <ctype.h>
#include <err.h>
#include <fcntl.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
#include "mkuzip.h"
#include "mkuz_cloop.h"
#include "mkuz_blockcache.h"
#include "mkuz_zlib.h"
#include "mkuz_lzma.h"
#define DEFINE_RAW_METHOD(func, rval, args...) typedef rval (*func##_t)(args)
#define DEFAULT_CLSTSIZE 16384
DEFINE_RAW_METHOD(f_init, void *, uint32_t);
DEFINE_RAW_METHOD(f_compress, void, const char *, uint32_t *);
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
struct mkuz_format {
const char *magic;
const char *default_sufx;
f_init_t f_init;
f_compress_t f_compress;
};
static struct mkuz_format uzip_fmt = {
.magic = CLOOP_MAGIC_ZLIB,
.default_sufx = DEFAULT_SUFX_ZLIB,
.f_init = &mkuz_zlib_init,
.f_compress = &mkuz_zlib_compress
};
static struct mkuz_format ulzma_fmt = {
.magic = CLOOP_MAGIC_LZMA,
.default_sufx = DEFAULT_SUFX_LZMA,
.f_init = &mkuz_lzma_init,
.f_compress = &mkuz_lzma_compress
};
static char *readblock(int, char *, u_int32_t);
static void usage(void);
static void cleanup(void);
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
static int memvcmp(const void *, unsigned char, size_t);
static char *cleanfile = NULL;
int main(int argc, char **argv)
{
char *iname, *oname, *obuf, *ibuf;
uint64_t *toc;
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
int fdr, fdw, i, opt, verbose, no_zcomp, tmp, en_dedup;
struct {
int en;
FILE *f;
} summary;
struct iovec iov[2];
struct stat sb;
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
uint32_t destlen;
uint64_t offset, last_offset;
struct cloop_header hdr;
struct mkuz_blkcache_hit *chit;
const struct mkuz_format *handler;
memset(&hdr, 0, sizeof(hdr));
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
hdr.blksz = DEFAULT_CLSTSIZE;
oname = NULL;
verbose = 0;
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
no_zcomp = 0;
en_dedup = 0;
summary.en = 0;
summary.f = stderr;
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
handler = &uzip_fmt;
while((opt = getopt(argc, argv, "o:s:vZdLS")) != -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 */
}
hdr.blksz = tmp;
break;
case 'v':
verbose = 1;
break;
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
case 'Z':
no_zcomp = 1;
break;
case 'd':
en_dedup = 1;
break;
case 'L':
handler = &ulzma_fmt;
break;
case 'S':
summary.en = 1;
summary.f = stdout;
break;
default:
usage();
/* Not reached */
}
}
argc -= optind;
argv += optind;
if (argc != 1) {
usage();
/* Not reached */
}
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
strcpy(hdr.magic, handler->magic);
if (en_dedup != 0) {
hdr.magic[CLOOP_OFS_VERSN] = CLOOP_MAJVER_3;
hdr.magic[CLOOP_OFS_COMPR] =
tolower(hdr.magic[CLOOP_OFS_COMPR]);
}
obuf = handler->f_init(hdr.blksz);
iname = argv[0];
if (oname == NULL) {
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
asprintf(&oname, "%s%s", iname, handler->default_sufx);
if (oname == NULL) {
err(1, "can't allocate memory");
/* Not reached */
}
}
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
ibuf = mkuz_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++;
}
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
toc = mkuz_safe_malloc((hdr.nblocks + 1) * sizeof(*toc));
fdw = open(oname, (en_dedup ? O_RDWR : 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);
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
last_offset = 0;
for(i = 0; i == 0 || ibuf != NULL; i++) {
ibuf = readblock(fdr, ibuf, hdr.blksz);
if (ibuf != NULL) {
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
if (no_zcomp == 0 && \
memvcmp(ibuf, '\0', hdr.blksz) != 0) {
/* All zeroes block */
destlen = 0;
} else {
handler->f_compress(ibuf, &destlen);
}
} else {
destlen = DEV_BSIZE - (offset % DEV_BSIZE);
memset(obuf, 0, destlen);
if (verbose != 0)
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
fprintf(stderr, "padding data with %lu bytes "
"so that file size is multiple of %d\n",
(u_long)destlen, DEV_BSIZE);
}
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
if (destlen > 0 && en_dedup != 0) {
chit = mkuz_blkcache_regblock(fdw, i, offset, destlen,
obuf);
/*
* There should be at least one non-empty block
* between us and the backref'ed offset, otherwise
* we won't be able to parse that sequence correctly
* as it would be indistinguishible from another
* empty block.
*/
if (chit != NULL && chit->offset == last_offset) {
chit = NULL;
}
} else {
chit = NULL;
}
if (chit != NULL) {
toc[i] = htobe64(chit->offset);
} else {
if (destlen > 0 && write(fdw, obuf, destlen) < 0) {
err(1, "write(%s)", oname);
/* Not reached */
}
toc[i] = htobe64(offset);
last_offset = offset;
offset += destlen;
}
if (ibuf != NULL && verbose != 0) {
fprintf(stderr, "cluster #%d, in %u bytes, "
"out len=%lu offset=%lu", i, hdr.blksz,
chit == NULL ? (u_long)destlen : 0,
(u_long)be64toh(toc[i]));
if (chit != NULL) {
fprintf(stderr, " (backref'ed to #%d)",
chit->blkno);
}
fprintf(stderr, "\n");
}
}
close(fdr);
if (verbose != 0 || summary.en != 0)
fprintf(summary.f, "compressed data to %ju bytes, saved %lld "
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
"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: mkuzip [-vZdLS] [-o outfile] [-s cluster_size] "
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
"infile\n");
exit(1);
}
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
void *
mkuz_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);
}
Improve mkuzip(8) and geom_uzip(4), merge in LZMA support from mkulzma(8) and geom_uncompress(4): 1. mkuzip(8): - Proper support for eliminating all-zero blocks when compressing an image. This feature is already supported by the geom_uzip(4) module and CLOOP format in general, so it's just a matter of making mkuzip(8) match. It should be noted, however that this feature while it sounds great, results in very slight improvement in the overall compression ratio, since compressing default 16k all-zero block produces only 39 bytes compressed output block, which is 99.8% compression ratio. With typical average compression ratio of amd64 binaries and data being around 60-70% the difference between 99.8% and 100.0% is not that great further diluted by the ratio of number of zero blocks in the uncompressed image to the overall number of blocks being less than 0.5 (typically). However, this may be important from performance standpoint, so that kernel are not spinning its wheels decompressing those empty blocks every time this zero region is read. It could also be important when you create huge image mostly filled with zero blocks for testing purposes. - New feature allowing to de-duplicate output image. It turns out that if you twist CLOOP format a bit you can do that as well. And unlike zero-blocks elimination, this gives a noticeable improvement in the overall compression ratio, reducing output image by something like 3-4% on my test UFS2 3GB image consisting of full FreeBSD base system plus some of the packages (openjdk, apache etc), about 2.3GB worth of file data (800+MB compressed). The only caveat is that images created with this feature "on" would not work on older versions of FeeBSDxi kernel, hence it's turned off by default. - provide options to control both features and document them in manual page. - merge in all relevant LZMA compression support from the mkulzma(8), add new option to select between both. - switch license from ad-hoc beerware into standard 2-clause BSD. 2. geom_uzip(4): - implement support for de-duplicated images; - optimize some code paths to handle "all-zero" blocks without reading any compressed data; - beef up manual page to explain that geom_uzip(4) is not limited only to md(4) images. The compressed data can be written to the block device and accessed directly via magic of GEOM(4) and devfs(4), including to mount root fs from a compressed drive. - convert debug log code from being compiled in conditionally into being present all the time and provide two sysctls to turn it on or off. Due to intended use of the module, it can be used in environments where there may not be a luxury to put new kernel with debug code enabled. Having those options handy allows debug issues without as much problem by just having access to serial console or network shell access to a box/appliance. The resulting additional CPU cycles are just few int comparisons and branches, and those are minuscule when compared to data decompression which is the main feature of the module. - hopefully improve robustness and resiliency of the geom_uzip(4) by performing some of the data validation / range checking on the TOC entries and rejecting to attach to an image if those checks fail. - merge in all relevant LZMA decompression support from the geom_uncompress(4), enable automatically when appropriate format is indicated in the header. - move compilation work into its own worker thread so that it does not clog g_up. This allows multiple instances work in parallel utilizing smp cores. - document new knobs in the manual page. Reviewed by: adrian MFC after: 1 month Differential Revision: https://reviews.freebsd.org/D5333
2016-02-23 23:59:08 +00:00
static int
memvcmp(const void *memory, unsigned char val, size_t size)
{
const u_char *mm;
mm = (const u_char *)memory;
return (*mm == val) && memcmp(mm, mm + 1, size - 1) == 0;
}