All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: Phillip Lougher <phillip@lougher.demon.co.uk>
Cc: linux-embedded@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org, phillip.lougher@gmail.com,
	tim.bird@am.sony.com, Stephen Rothwell <sfr@canb.auug.org.au>,
	Albin Tonnerre <albin.tonnerre@free-electrons.com>
Subject: Re: [PATCH V2 6/8] lzma: Make lzma available to non initramfs/initrd code
Date: Wed, 6 Jan 2010 14:04:09 -0800	[thread overview]
Message-ID: <20100106140409.ba0f63d1.akpm@linux-foundation.org> (raw)
In-Reply-To: <E1NIuOW-0001ol-RC@dylan.lougher.demon.co.uk>

On Fri, 11 Dec 2009 01:34:08 +0000
Phillip Lougher <phillip@lougher.demon.co.uk> wrote:

> Add a config option DECOMPRESS_LZMA_NEEDED which allows subsystems to
> specify they need the unlzma code.  Normally decompress_unlzma.c is
> compiled with __init and unlzma is not exported to modules.
> 
> Move INIT definition into separate header files for bzip2/lzma/inflate
> so it can be defined differently for each decompressor.
> 

This patch (which is in linux-next) breaks
lib-add-support-for-lzo-compressed-kernels.patch, below.  The
definition of INIT is no longer available in lib/decompress_unlzo.c, and

lib/decompress_unlzo.c: In function 'unlzo':
lib/decompress_unlzo.c:106: error: 'error' undeclared (first use in this function)
lib/decompress_unlzo.c:106: error: (Each undeclared identifier is reported only once
lib/decompress_unlzo.c:106: error: for each function it appears in.)
lib/decompress_unlzo.c:111: error: implicit declaration of function 'error'


I'm planning on merging 

zlib-optimize-inffast-when-copying-direct-from-output.patch
lib-add-support-for-lzo-compressed-kernels.patch
arm-add-support-for-lzo-compressed-kernels.patch
x86-add-support-for-lzo-compressed-kernels.patch
add-lzo-compression-support-for-initramfs-and-old-style-initrd.patch

into 2.6.33.  I don't immediately remember why I decided that - perhaps
because the patches did arrive in time for .33, but got stalled because
people were screwing around in other trees.

So if I go ahead with that merge, linux-next will need fixing.  And I
didn't get down and work what the appropriate fix is, and I don't want
to break linux-next in serious ways.


So what to do?  I guess I could go ahead with the mainline merge, and
Stephen drops <whatever that tree was> from linux-next until it has
been fixed up?

If we decide to hold the above five patches off until 2.6.34 then
they'll need to be fixed up to work against linux-next, and they will
be dependent upon your patch series.




From: Albin Tonnerre <albin.tonnerre@free-electrons.com>

This patch series adds generic support for creating and extracting
LZO-compressed kernel images, as well as support for using such images on
the x86 and ARM architectures, and support for creating and using
LZO-compressed initrd and initramfs images.


Russell King said:

: Testing on a Cortex A9 model:
: - lzo decompressor is 65% of the time gzip takes to decompress a kernel
: - lzo kernel is 9% larger than a gzip kernel
: 
: which I'm happy to say confirms your figures when comparing the two.
: 
: However, when comparing your new gzip code to the old gzip code:
: - new is 99% of the size of the old code
: - new takes 42% of the time to decompress than the old code
: 
: What this means is that for a proper comparison, the results get even better:
: - lzo is 7.5% larger than the old gzip'd kernel image
: - lzo takes 28% of the time that the old gzip code took
: 
: So the expense seems definitely worth the effort.  The only reason I
: can think of ever using gzip would be if you needed the additional
: compression (eg, because you have limited flash to store the image.)
: 
: I would argue that the default for ARM should therefore be LZO.


This patch:

The lzo compressor is worse than gzip at compression, but faster at
extraction.  Here are some figures for an ARM board I'm working on:

Uncompressed size: 3.24Mo
gzip  1.61Mo 0.72s
lzo   1.75Mo 0.48s

So for a compression ratio that is still relatively close to gzip, it's
much faster to extract, at least in that case.

This part contains:
 - Makefile routine to support lzo compression
 - Fixes to the existing lzo compressor so that it can be used in
   compressed kernels
 - wrapper around the existing lzo1x_decompress, as it only extracts one
   block at a time, while we need to extract a whole file here
 - config dialog for kernel compression

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: cleanup]
Signed-off-by: Albin Tonnerre <albin.tonnerre@free-electrons.com>
Tested-by: Wu Zhangjin <wuzhangjin@gmail.com>
Acked-by: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Russell King <rmk@arm.linux.org.uk>
Acked-by: Russell King <rmk@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 include/linux/decompress/unlzo.h |   10 +
 init/Kconfig                     |   18 +-
 lib/decompress_unlzo.c           |  209 +++++++++++++++++++++++++++++
 lib/lzo/lzo1x_decompress.c       |    9 -
 scripts/Makefile.lib             |    5 
 5 files changed, 244 insertions(+), 7 deletions(-)

diff -puN /dev/null include/linux/decompress/unlzo.h
--- /dev/null
+++ a/include/linux/decompress/unlzo.h
@@ -0,0 +1,10 @@
+#ifndef DECOMPRESS_UNLZO_H
+#define DECOMPRESS_UNLZO_H
+
+int unlzo(unsigned char *inbuf, int len,
+	int(*fill)(void*, unsigned int),
+	int(*flush)(void*, unsigned int),
+	unsigned char *output,
+	int *pos,
+	void(*error)(char *x));
+#endif
diff -puN init/Kconfig~lib-add-support-for-lzo-compressed-kernels init/Kconfig
--- a/init/Kconfig~lib-add-support-for-lzo-compressed-kernels
+++ a/init/Kconfig
@@ -123,10 +123,13 @@ config HAVE_KERNEL_BZIP2
 config HAVE_KERNEL_LZMA
 	bool
 
+config HAVE_KERNEL_LZO
+	bool
+
 choice
 	prompt "Kernel compression mode"
 	default KERNEL_GZIP
-	depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA
+	depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_LZO
 	help
 	  The linux kernel is a kind of self-extracting executable.
 	  Several compression algorithms are available, which differ
@@ -149,9 +152,8 @@ config KERNEL_GZIP
 	bool "Gzip"
 	depends on HAVE_KERNEL_GZIP
 	help
-	  The old and tried gzip compression. Its compression ratio is
-	  the poorest among the 3 choices; however its speed (both
-	  compression and decompression) is the fastest.
+	  The old and tried gzip compression. It provides a good balance
+	  between compression ratio and decompression speed.
 
 config KERNEL_BZIP2
 	bool "Bzip2"
@@ -172,6 +174,14 @@ config KERNEL_LZMA
 	  two. Compression is slowest.	The kernel size is about 33%
 	  smaller with LZMA in comparison to gzip.
 
+config KERNEL_LZO
+	bool "LZO"
+	depends on HAVE_KERNEL_LZO
+	help
+	  Its compression ratio is the poorest among the 4. The kernel
+	  size is about about 10% bigger than gzip; however its speed
+	  (both compression and decompression) is the fastest.
+
 endchoice
 
 config SWAP
diff -puN /dev/null lib/decompress_unlzo.c
--- /dev/null
+++ a/lib/decompress_unlzo.c
@@ -0,0 +1,209 @@
+/*
+ * LZO decompressor for the Linux kernel. Code borrowed from the lzo
+ * implementation by Markus Franz Xaver Johannes Oberhumer.
+ *
+ * Linux kernel adaptation:
+ * Copyright (C) 2009
+ * Albin Tonnerre, Free Electrons <albin.tonnerre@free-electrons.com>
+ *
+ * Original code:
+ * Copyright (C) 1996-2005 Markus Franz Xaver Johannes Oberhumer
+ * All Rights Reserved.
+ *
+ * lzop and the LZO library are free software; you can redistribute them
+ * and/or modify them under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; see the file COPYING.
+ * If not, write to the Free Software Foundation, Inc.,
+ * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Markus F.X.J. Oberhumer
+ * <markus@oberhumer.com>
+ * http://www.oberhumer.com/opensource/lzop/
+ */
+
+#ifdef STATIC
+#include "lzo/lzo1x_decompress.c"
+#else
+#include <linux/slab.h>
+#include <linux/decompress/unlzo.h>
+#endif
+
+#include <linux/types.h>
+#include <linux/lzo.h>
+#include <linux/decompress/mm.h>
+
+#include <linux/compiler.h>
+#include <asm/unaligned.h>
+
+static const unsigned char lzop_magic[] = {
+	0x89, 0x4c, 0x5a, 0x4f, 0x00, 0x0d, 0x0a, 0x1a, 0x0a };
+
+#define LZO_BLOCK_SIZE        (256*1024l)
+#define HEADER_HAS_FILTER      0x00000800L
+
+STATIC inline int INIT parse_header(u8 *input, u8 *skip)
+{
+	int l;
+	u8 *parse = input;
+	u8 level = 0;
+	u16 version;
+
+	/* read magic: 9 first bits */
+	for (l = 0; l < 9; l++) {
+		if (*parse++ != lzop_magic[l])
+			return 0;
+	}
+	/* get version (2bytes), skip library version (2),
+	 * 'need to be extracted' version (2) and
+	 * method (1) */
+	version = get_unaligned_be16(parse);
+	parse += 7;
+	if (version >= 0x0940)
+		level = *parse++;
+	if (get_unaligned_be32(parse) & HEADER_HAS_FILTER)
+		parse += 8; /* flags + filter info */
+	else
+		parse += 4; /* flags */
+
+	/* skip mode and mtime_low */
+	parse += 8;
+	if (version >= 0x0940)
+		parse += 4;	/* skip mtime_high */
+
+	l = *parse++;
+	/* don't care about the file name, and skip checksum */
+	parse += l + 4;
+
+	*skip = parse - input;
+	return 1;
+}
+
+STATIC inline int INIT unlzo(u8 *input, int in_len,
+				int (*fill) (void *, unsigned int),
+				int (*flush) (void *, unsigned int),
+				u8 *output, int *posp,
+				void (*error_fn) (char *x))
+{
+	u8 skip = 0, r = 0;
+	u32 src_len, dst_len;
+	size_t tmp;
+	u8 *in_buf, *in_buf_save, *out_buf;
+	int obytes_processed = 0;
+
+	set_error_fn(error_fn);
+
+	if (output) {
+		out_buf = output;
+	} else if (!flush) {
+		error("NULL output pointer and no flush function provided");
+		goto exit;
+	} else {
+		out_buf = malloc(LZO_BLOCK_SIZE);
+		if (!out_buf) {
+			error("Could not allocate output buffer");
+			goto exit;
+		}
+	}
+
+	if (input && fill) {
+		error("Both input pointer and fill function provided, don't know what to do");
+		goto exit_1;
+	} else if (input) {
+		in_buf = input;
+	} else if (!fill || !posp) {
+		error("NULL input pointer and missing position pointer or fill function");
+		goto exit_1;
+	} else {
+		in_buf = malloc(lzo1x_worst_compress(LZO_BLOCK_SIZE));
+		if (!in_buf) {
+			error("Could not allocate input buffer");
+			goto exit_1;
+		}
+	}
+	in_buf_save = in_buf;
+
+	if (posp)
+		*posp = 0;
+
+	if (fill)
+		fill(in_buf, lzo1x_worst_compress(LZO_BLOCK_SIZE));
+
+	if (!parse_header(input, &skip)) {
+		error("invalid header");
+		goto exit_2;
+	}
+	in_buf += skip;
+
+	if (posp)
+		*posp = skip;
+
+	for (;;) {
+		/* read uncompressed block size */
+		dst_len = get_unaligned_be32(in_buf);
+		in_buf += 4;
+
+		/* exit if last block */
+		if (dst_len == 0) {
+			if (posp)
+				*posp += 4;
+			break;
+		}
+
+		if (dst_len > LZO_BLOCK_SIZE) {
+			error("dest len longer than block size");
+			goto exit_2;
+		}
+
+		/* read compressed block size, and skip block checksum info */
+		src_len = get_unaligned_be32(in_buf);
+		in_buf += 8;
+
+		if (src_len <= 0 || src_len > dst_len) {
+			error("file corrupted");
+			goto exit_2;
+		}
+
+		/* decompress */
+		tmp = dst_len;
+		r = lzo1x_decompress_safe((u8 *) in_buf, src_len,
+						out_buf, &tmp);
+
+		if (r != LZO_E_OK || dst_len != tmp) {
+			error("Compressed data violation");
+			goto exit_2;
+		}
+
+		obytes_processed += dst_len;
+		if (flush)
+			flush(out_buf, dst_len);
+		if (output)
+			out_buf += dst_len;
+		if (posp)
+			*posp += src_len + 12;
+		if (fill) {
+			in_buf = in_buf_save;
+			fill(in_buf, lzo1x_worst_compress(LZO_BLOCK_SIZE));
+		} else
+			in_buf += src_len;
+	}
+
+exit_2:
+	if (!input)
+		free(in_buf);
+exit_1:
+	if (!output)
+		free(out_buf);
+exit:
+	return obytes_processed;
+}
+
+#define decompress unlzo
diff -puN lib/lzo/lzo1x_decompress.c~lib-add-support-for-lzo-compressed-kernels lib/lzo/lzo1x_decompress.c
--- a/lib/lzo/lzo1x_decompress.c~lib-add-support-for-lzo-compressed-kernels
+++ a/lib/lzo/lzo1x_decompress.c
@@ -11,11 +11,13 @@
  *  Richard Purdie <rpurdie@openedhand.com>
  */
 
+#ifndef STATIC
 #include <linux/module.h>
 #include <linux/kernel.h>
-#include <linux/lzo.h>
-#include <asm/byteorder.h>
+#endif
+
 #include <asm/unaligned.h>
+#include <linux/lzo.h>
 #include "lzodefs.h"
 
 #define HAVE_IP(x, ip_end, ip) ((size_t)(ip_end - ip) < (x))
@@ -244,9 +246,10 @@ lookbehind_overrun:
 	*out_len = op - out;
 	return LZO_E_LOOKBEHIND_OVERRUN;
 }
-
+#ifndef STATIC
 EXPORT_SYMBOL_GPL(lzo1x_decompress_safe);
 
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("LZO1X Decompressor");
 
+#endif
diff -puN scripts/Makefile.lib~lib-add-support-for-lzo-compressed-kernels scripts/Makefile.lib
--- a/scripts/Makefile.lib~lib-add-support-for-lzo-compressed-kernels
+++ a/scripts/Makefile.lib
@@ -235,3 +235,8 @@ quiet_cmd_lzma = LZMA    $@
 cmd_lzma = (cat $(filter-out FORCE,$^) | \
 	lzma -9 && $(call size_append, $(filter-out FORCE,$^))) > $@ || \
 	(rm -f $@ ; false)
+
+quiet_cmd_lzo = LZO    $@
+cmd_lzo = (cat $(filter-out FORCE,$^) | \
+	lzop -9 && $(call size_append, $(filter-out FORCE,$^))) > $@ || \
+	(rm -f $@ ; false)
_

  reply	other threads:[~2010-01-06 22:04 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-12-11  1:34 [PATCH V2 6/8] lzma: Make lzma available to non initramfs/initrd code Phillip Lougher
2010-01-06 22:04 ` Andrew Morton [this message]
2010-01-06 22:45   ` Albin Tonnerre
2010-01-07  0:04   ` Stephen Rothwell
2010-01-07  0:20   ` Phillip Lougher
2010-01-07  0:31     ` Andrew Morton
2010-01-07  0:37       ` Phillip Lougher
2010-01-07  0:37         ` Phillip Lougher
2010-01-07  7:43         ` [PATCH V2 6/8] lzma: Make lzma ... thunderbird ok Hein_Tibosch
2010-01-07 17:26           ` Marco Stornelli
  -- strict thread matches above, loose matches on Subject: below --
2010-01-07  0:49 [PATCH V2 6/8] lzma: Make lzma available to non initramfs/initrd code Phillip Lougher
2010-01-07  0:49 ` Phillip Lougher

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100106140409.ba0f63d1.akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=albin.tonnerre@free-electrons.com \
    --cc=linux-embedded@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=phillip.lougher@gmail.com \
    --cc=phillip@lougher.demon.co.uk \
    --cc=sfr@canb.auug.org.au \
    --cc=tim.bird@am.sony.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.