2007-07-11 04:22:24 +04:00
/*
2012-08-13 19:25:44 +04:00
* LZO1X Decompressor from LZO
2007-07-11 04:22:24 +04:00
*
2012-08-13 19:25:44 +04:00
* Copyright ( C ) 1996 - 2012 Markus F . X . J . Oberhumer < markus @ oberhumer . com >
2007-07-11 04:22:24 +04:00
*
* The full LZO package can be found at :
* http : //www.oberhumer.com/opensource/lzo/
*
2012-08-13 19:25:44 +04:00
* Changed for Linux kernel use by :
2007-07-11 04:22:24 +04:00
* Nitin Gupta < nitingupta910 @ gmail . com >
* Richard Purdie < rpurdie @ openedhand . com >
*/
lib: add support for LZO-compressed kernels
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-09 01:42:42 +03:00
# ifndef STATIC
2007-07-11 04:22:24 +04:00
# include <linux/module.h>
# include <linux/kernel.h>
lib: add support for LZO-compressed kernels
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-09 01:42:42 +03:00
# endif
2007-07-11 04:22:24 +04:00
# include <asm/unaligned.h>
lib: add support for LZO-compressed kernels
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-09 01:42:42 +03:00
# include <linux/lzo.h>
2007-07-11 04:22:24 +04:00
# include "lzodefs.h"
2014-09-27 14:31:36 +04:00
# define HAVE_IP(x) ((size_t)(ip_end - ip) >= (size_t)(x))
# define HAVE_OP(x) ((size_t)(op_end - op) >= (size_t)(x))
# define NEED_IP(x) if (!HAVE_IP(x)) goto input_overrun
# define NEED_OP(x) if (!HAVE_OP(x)) goto output_overrun
# define TEST_LB(m_pos) if ((m_pos) < out) goto lookbehind_overrun
2007-07-11 04:22:24 +04:00
2014-09-27 14:31:37 +04:00
/* This MAX_255_COUNT is the maximum number of times we can add 255 to a base
* count without overflowing an integer . The multiply will overflow when
* multiplying 255 by more than MAXINT / 255. The sum will overflow earlier
* depending on the base count . Since the base count is taken from a u8
* and a few bits , it is safe to assume that it will always be lower than
* or equal to 2 * 255 , thus we can always prevent any overflow by accepting
* two less 255 steps . See Documentation / lzo . txt for more information .
*/
# define MAX_255_COUNT ((((size_t)~0) / 255) - 2)
2007-07-11 04:22:24 +04:00
int lzo1x_decompress_safe ( const unsigned char * in , size_t in_len ,
2012-08-13 19:25:44 +04:00
unsigned char * out , size_t * out_len )
2007-07-11 04:22:24 +04:00
{
2012-08-13 19:25:44 +04:00
unsigned char * op ;
const unsigned char * ip ;
size_t t , next ;
size_t state = 0 ;
const unsigned char * m_pos ;
2007-07-11 04:22:24 +04:00
const unsigned char * const ip_end = in + in_len ;
unsigned char * const op_end = out + * out_len ;
2012-08-13 19:25:44 +04:00
op = out ;
ip = in ;
2007-07-11 04:22:24 +04:00
2012-08-13 19:25:44 +04:00
if ( unlikely ( in_len < 3 ) )
goto input_overrun ;
2007-07-11 04:22:24 +04:00
if ( * ip > 17 ) {
t = * ip + + - 17 ;
2012-08-13 19:25:44 +04:00
if ( t < 4 ) {
next = t ;
2007-07-11 04:22:24 +04:00
goto match_next ;
}
2012-08-13 19:25:44 +04:00
goto copy_literal_run ;
}
2007-07-11 04:22:24 +04:00
2012-08-13 19:25:44 +04:00
for ( ; ; ) {
2007-07-11 04:22:24 +04:00
t = * ip + + ;
2012-08-13 19:25:44 +04:00
if ( t < 16 ) {
if ( likely ( state = = 0 ) ) {
if ( unlikely ( t = = 0 ) ) {
2014-09-27 14:31:37 +04:00
size_t offset ;
const unsigned char * ip_last = ip ;
2012-08-13 19:25:44 +04:00
while ( unlikely ( * ip = = 0 ) ) {
2007-07-11 04:22:24 +04:00
ip + + ;
2014-09-27 14:31:36 +04:00
NEED_IP ( 1 ) ;
2007-07-11 04:22:24 +04:00
}
2014-09-27 14:31:37 +04:00
offset = ip - ip_last ;
if ( unlikely ( offset > MAX_255_COUNT ) )
return LZO_E_ERROR ;
offset = ( offset < < 8 ) - offset ;
t + = offset + 15 + * ip + + ;
2007-07-11 04:22:24 +04:00
}
2012-08-13 19:25:44 +04:00
t + = 3 ;
copy_literal_run :
# if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
2014-09-27 14:31:36 +04:00
if ( likely ( HAVE_IP ( t + 15 ) & & HAVE_OP ( t + 15 ) ) ) {
2012-08-13 19:25:44 +04:00
const unsigned char * ie = ip + t ;
unsigned char * oe = op + t ;
do {
COPY8 ( op , ip ) ;
op + = 8 ;
ip + = 8 ;
COPY8 ( op , ip ) ;
op + = 8 ;
ip + = 8 ;
} while ( ip < ie ) ;
ip = ie ;
op = oe ;
} else
# endif
{
2014-09-27 14:31:36 +04:00
NEED_OP ( t ) ;
NEED_IP ( t + 3 ) ;
2012-08-13 19:25:44 +04:00
do {
* op + + = * ip + + ;
} while ( - - t > 0 ) ;
2007-07-11 04:22:24 +04:00
}
2012-08-13 19:25:44 +04:00
state = 4 ;
continue ;
} else if ( state ! = 4 ) {
next = t & 3 ;
2007-07-11 04:22:24 +04:00
m_pos = op - 1 ;
m_pos - = t > > 2 ;
m_pos - = * ip + + < < 2 ;
2012-08-13 19:25:44 +04:00
TEST_LB ( m_pos ) ;
2014-09-27 14:31:36 +04:00
NEED_OP ( 2 ) ;
2012-08-13 19:25:44 +04:00
op [ 0 ] = m_pos [ 0 ] ;
op [ 1 ] = m_pos [ 1 ] ;
op + = 2 ;
goto match_next ;
} else {
next = t & 3 ;
m_pos = op - ( 1 + M2_MAX_OFFSET ) ;
m_pos - = t > > 2 ;
m_pos - = * ip + + < < 2 ;
t = 3 ;
2007-07-11 04:22:24 +04:00
}
2012-08-13 19:25:44 +04:00
} else if ( t > = 64 ) {
next = t & 3 ;
m_pos = op - 1 ;
m_pos - = ( t > > 2 ) & 7 ;
m_pos - = * ip + + < < 3 ;
t = ( t > > 5 ) - 1 + ( 3 - 1 ) ;
} else if ( t > = 32 ) {
t = ( t & 31 ) + ( 3 - 1 ) ;
if ( unlikely ( t = = 2 ) ) {
2014-09-27 14:31:37 +04:00
size_t offset ;
const unsigned char * ip_last = ip ;
2012-08-13 19:25:44 +04:00
while ( unlikely ( * ip = = 0 ) ) {
ip + + ;
2014-09-27 14:31:36 +04:00
NEED_IP ( 1 ) ;
2012-08-13 19:25:44 +04:00
}
2014-09-27 14:31:37 +04:00
offset = ip - ip_last ;
if ( unlikely ( offset > MAX_255_COUNT ) )
return LZO_E_ERROR ;
offset = ( offset < < 8 ) - offset ;
t + = offset + 31 + * ip + + ;
2014-09-27 14:31:36 +04:00
NEED_IP ( 2 ) ;
2012-08-13 19:25:44 +04:00
}
m_pos = op - 1 ;
next = get_unaligned_le16 ( ip ) ;
ip + = 2 ;
m_pos - = next > > 2 ;
next & = 3 ;
} else {
m_pos = op ;
m_pos - = ( t & 8 ) < < 11 ;
t = ( t & 7 ) + ( 3 - 1 ) ;
if ( unlikely ( t = = 2 ) ) {
2014-09-27 14:31:37 +04:00
size_t offset ;
const unsigned char * ip_last = ip ;
2012-08-13 19:25:44 +04:00
while ( unlikely ( * ip = = 0 ) ) {
ip + + ;
2014-09-27 14:31:36 +04:00
NEED_IP ( 1 ) ;
2012-08-13 19:25:44 +04:00
}
2014-09-27 14:31:37 +04:00
offset = ip - ip_last ;
if ( unlikely ( offset > MAX_255_COUNT ) )
return LZO_E_ERROR ;
offset = ( offset < < 8 ) - offset ;
t + = offset + 7 + * ip + + ;
2014-09-27 14:31:36 +04:00
NEED_IP ( 2 ) ;
2012-08-13 19:25:44 +04:00
}
next = get_unaligned_le16 ( ip ) ;
ip + = 2 ;
m_pos - = next > > 2 ;
next & = 3 ;
if ( m_pos = = op )
goto eof_found ;
m_pos - = 0x4000 ;
}
TEST_LB ( m_pos ) ;
# if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
if ( op - m_pos > = 8 ) {
unsigned char * oe = op + t ;
2014-09-27 14:31:36 +04:00
if ( likely ( HAVE_OP ( t + 15 ) ) ) {
2007-07-11 04:22:24 +04:00
do {
2012-08-13 19:25:44 +04:00
COPY8 ( op , m_pos ) ;
op + = 8 ;
m_pos + = 8 ;
COPY8 ( op , m_pos ) ;
op + = 8 ;
m_pos + = 8 ;
} while ( op < oe ) ;
op = oe ;
2014-09-27 14:31:36 +04:00
if ( HAVE_IP ( 6 ) ) {
2012-08-13 19:25:44 +04:00
state = next ;
COPY4 ( op , ip ) ;
op + = next ;
ip + = next ;
continue ;
}
2007-07-11 04:22:24 +04:00
} else {
2014-09-27 14:31:36 +04:00
NEED_OP ( t ) ;
2007-07-11 04:22:24 +04:00
do {
* op + + = * m_pos + + ;
2012-08-13 19:25:44 +04:00
} while ( op < oe ) ;
2007-07-11 04:22:24 +04:00
}
2012-08-13 19:25:44 +04:00
} else
# endif
{
unsigned char * oe = op + t ;
2014-09-27 14:31:36 +04:00
NEED_OP ( t ) ;
2012-08-13 19:25:44 +04:00
op [ 0 ] = m_pos [ 0 ] ;
op [ 1 ] = m_pos [ 1 ] ;
op + = 2 ;
m_pos + = 2 ;
do {
* op + + = * m_pos + + ;
} while ( op < oe ) ;
}
2007-07-11 04:22:24 +04:00
match_next :
2012-08-13 19:25:44 +04:00
state = next ;
t = next ;
# if defined(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS)
2014-09-27 14:31:36 +04:00
if ( likely ( HAVE_IP ( 6 ) & & HAVE_OP ( 4 ) ) ) {
2012-08-13 19:25:44 +04:00
COPY4 ( op , ip ) ;
op + = t ;
ip + = t ;
} else
# endif
{
2014-09-27 14:31:36 +04:00
NEED_IP ( t + 3 ) ;
NEED_OP ( t ) ;
2012-08-13 19:25:44 +04:00
while ( t > 0 ) {
2007-07-11 04:22:24 +04:00
* op + + = * ip + + ;
2012-08-13 19:25:44 +04:00
t - - ;
2007-07-11 04:22:24 +04:00
}
2012-08-13 19:25:44 +04:00
}
2007-07-11 04:22:24 +04:00
}
eof_found :
* out_len = op - out ;
2012-08-13 19:25:44 +04:00
return ( t ! = 3 ? LZO_E_ERROR :
ip = = ip_end ? LZO_E_OK :
ip < ip_end ? LZO_E_INPUT_NOT_CONSUMED : LZO_E_INPUT_OVERRUN ) ;
2007-07-11 04:22:24 +04:00
input_overrun :
* out_len = op - out ;
return LZO_E_INPUT_OVERRUN ;
output_overrun :
* out_len = op - out ;
return LZO_E_OUTPUT_OVERRUN ;
lookbehind_overrun :
* out_len = op - out ;
return LZO_E_LOOKBEHIND_OVERRUN ;
}
lib: add support for LZO-compressed kernels
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-09 01:42:42 +03:00
# ifndef STATIC
2007-07-11 04:22:24 +04:00
EXPORT_SYMBOL_GPL ( lzo1x_decompress_safe ) ;
MODULE_LICENSE ( " GPL " ) ;
MODULE_DESCRIPTION ( " LZO1X Decompressor " ) ;
lib: add support for LZO-compressed kernels
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>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-09 01:42:42 +03:00
# endif