Illumos #3035 LZ4 compression support in ZFS and GRUB

3035 LZ4 compression support in ZFS and GRUB

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Approved by: Christopher Siden <csiden@delphix.com>

References:
  illumos/illumos-gate@a6f561b
  https://www.illumos.org/issues/3035
  http://wiki.illumos.org/display/illumos/LZ4+Compression+In+ZFS

This patch has been slightly modified from the upstream Illumos
version to be compatible with Linux.  Due to the very limited
stack space in the kernel a lz4 workspace kmem cache is used.
Since we are using gcc we are also able to take advantage of the
gcc optimized __builtin_ctz functions.

Support for GRUB has been dropped from this patch.  That code
is available but those changes will need to made to the upstream
GRUB package.

Lastly, several hunks of dead code were dropped for clarity.  They
include the functions real_LZ4_uncompress(), LZ4_compressBound()
and the Visual Studio specific hunks wrapped in _MSC_VER.

Ported-by: Eric Dillmann <eric@jave.fr>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #1217

Author: edillmann

include/sys/zio.h

@@ -25,6 +25,7 @@
 /*
  * Copyright 2011 Nexenta Systems, Inc. All rights reserved.
  * Copyright (c) 2012 by Delphix. All rights reserved.
+ * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
  */
 
 #ifndef _ZIO_H
@@ -108,6 +109,7 @@ enum zio_compress {
 	ZIO_COMPRESS_GZIP_8,
 	ZIO_COMPRESS_GZIP_9,
 	ZIO_COMPRESS_ZLE,
+	ZIO_COMPRESS_LZ4,
 	ZIO_COMPRESS_FUNCTIONS
 };
 
@@ -116,6 +118,7 @@ enum zio_compress {
 
 #define	BOOTFS_COMPRESS_VALID(compress)			\
 	((compress) == ZIO_COMPRESS_LZJB ||		\
+	(compress) == ZIO_COMPRESS_LZ4 ||		\
 	((compress) == ZIO_COMPRESS_ON &&		\
 	ZIO_COMPRESS_ON_VALUE == ZIO_COMPRESS_LZJB) ||	\
 	(compress) == ZIO_COMPRESS_OFF)

include/sys/zio_compress.h

@@ -53,6 +53,12 @@ typedef struct zio_compress_info {
 
 extern zio_compress_info_t zio_compress_table[ZIO_COMPRESS_FUNCTIONS];
 
+/*
+ * lz4 compression init & free
+ */
+extern void lz4_init(void);
+extern void lz4_fini(void);
+
 /*
  * Compression routines.
  */
@@ -68,6 +74,10 @@ extern size_t zle_compress(void *src, void *dst, size_t s_len, size_t d_len,
     int level);
 extern int zle_decompress(void *src, void *dst, size_t s_len, size_t d_len,
     int level);
+extern size_t lz4_compress(void *src, void *dst, size_t s_len, size_t d_len,
+    int level);
+extern int lz4_decompress(void *src, void *dst, size_t s_len, size_t d_len,
+    int level);
 
 /*
  * Compress and decompress data if necessary.

include/zfeature_common.h

@@ -21,6 +21,7 @@
 
 /*
  * Copyright (c) 2012 by Delphix. All rights reserved.
+ * Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
  */
 
 #ifndef _ZFEATURE_COMMON_H
@@ -52,6 +53,7 @@ typedef int (zfeature_func_t)(zfeature_info_t *fi, void *arg);
 typedef enum spa_feature {
 	SPA_FEATURE_ASYNC_DESTROY,
 	SPA_FEATURE_EMPTY_BPOBJ,
+	SPA_FEATURE_LZ4_COMPRESS,
 	SPA_FEATURES
 } spa_feature_t;
 

lib/libzpool/Makefile.am

@@ -48,6 +48,7 @@ libzpool_la_SOURCES = \
 	$(top_srcdir)/module/zfs/fm.c \
 	$(top_srcdir)/module/zfs/gzip.c \
 	$(top_srcdir)/module/zfs/lzjb.c \
+	$(top_srcdir)/module/zfs/lz4.c \
 	$(top_srcdir)/module/zfs/metaslab.c \
 	$(top_srcdir)/module/zfs/refcount.c \
 	$(top_srcdir)/module/zfs/rrwlock.c \

man/man5/zpool-features.5

@@ -1,5 +1,6 @@
 '\" te
 .\" Copyright (c) 2012 by Delphix. All rights reserved.
+.\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
 .\" The contents of this file are subject to the terms of the Common Development
 .\" and Distribution License (the "License").  You may not use this file except
 .\" in compliance with the License. You can obtain a copy of the license at
@@ -197,5 +198,38 @@ This feature is \fBactive\fR while there are any filesystems, volumes,
 or snapshots which were created after enabling this feature.
 .RE
 
+.sp
+.ne 2
+.na
+\fB\fBlz4_compress\fR\fR
+.ad
+.RS 4n
+.TS
+l l .
+GUID	org.illumos:lz4_compress
+READ\-ONLY COMPATIBLE	no
+DEPENDENCIES	none
+.TE
+
+\fBlz4\fR is a high-performance real-time compression algorithm that
+features significantly faster compression and decompression as well as a
+higher compression ratio than the older \fBlzjb\fR compression.
+Typically, \fBlz4\fR compression is approximately 50% faster on
+compressible data and 200% faster on incompressible data than
+\fBlzjb\fR. It is also approximately 80% faster on decompression, while
+giving approximately 10% better compression ratio.
+
+When the \fBlz4_compress\fR feature is set to \fBenabled\fR, the
+administrator can turn on \fBlz4\fR compression on any dataset on the
+pool using the \fBzfs\fR(1M) command. Please note that doing so will
+immediately activate the \fBlz4_compress\fR feature on the underlying
+pool (even before any data is written). Since this feature is not
+read-only compatible, this operation will render the pool unimportable
+on systems without support for the \fBlz4_compress\fR feature. At the
+moment, this operation cannot be reversed. Booting off of
+\fBlz4\fR-compressed root pools is supported.
+
+.RE
+
 .SH "SEE ALSO"
 \fBzpool\fR(1M)

man/man8/zfs.8

@@ -25,6 +25,7 @@
 .\" Copyright (c) 2012 by Delphix. All rights reserved.
 .\" Copyright (c) 2012, Joyent, Inc. All rights reserved.
 .\" Copyright 2012 Nexenta Systems, Inc. All Rights Reserved.
+.\" Copyright (c) 2013 by Saso Kiselkov. All rights reserved.
 .\"
 .TH zfs 8 "Jan 10, 2013" "ZFS pool 28, filesystem 5" "System Administration Commands"
 .SH NAME
@@ -731,7 +732,7 @@ Changing this property affects only newly-written data.
 .ne 2
 .mk
 .na
-\fBcompression\fR=\fBon\fR | \fBoff\fR | \fBlzjb\fR | \fBgzip\fR | \fBgzip-\fR\fIN\fR | \fBzle\fR
+\fBcompression\fR=\fBon\fR | \fBoff\fR | \fBlzjb\fR | \fBgzip\fR | \fBgzip-\fR\fIN\fR | \fBzle\fR | \fBlz4\fR
 .ad
 .sp .6
 .RS 4n
@@ -741,6 +742,14 @@ The \fBgzip\fR compression algorithm uses the same compression as the \fBgzip\fR
 .sp
 The \fBzle\fR (zero-length encoding) compression algorithm is a fast and simple algorithm to eliminate runs of zeroes.
 .sp
+The \fBlz4\fR compression algorithm is a high-performance replacement
+for the \fBlzjb\fR algorithm. It features significantly faster
+compression and decompression, as well as a moderately higher
+compression ratio than \fBlzjb\fR, but can only be used on pools with
+the \fBlz4_compress\fR feature set to \fIenabled\fR. See
+\fBzpool-features\fR(5) for details on ZFS feature flags and the
+\fBlz4_compress\fR feature.
+.sp
 This property can also be referred to by its shortened column name \fBcompress\fR. Changing this property affects only newly-written data.
 .RE
 

���module/zcommon/zfs_prop.c

@@ -96,6 +96,7 @@ zfs_prop_init(void)
 		{ "gzip-8",	ZIO_COMPRESS_GZIP_8 },
 		{ "gzip-9",	ZIO_COMPRESS_GZIP_9 },
 		{ "zle",	ZIO_COMPRESS_ZLE },
+		{ "lz4",	ZIO_COMPRESS_LZ4 },
 		{ NULL }
 	};
 
@@ -211,7 +212,7 @@ zfs_prop_init(void)
 	zprop_register_index(ZFS_PROP_COMPRESSION, "compression",
 	    ZIO_COMPRESS_DEFAULT, PROP_INHERIT,
 	    ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME,
-	    "on | off | lzjb | gzip | gzip-[1-9] | zle", "COMPRESS",
+	    "on | off | lzjb | gzip | gzip-[1-9] | zle | lz4", "COMPRESS",
 	    compress_table);
 	zprop_register_index(ZFS_PROP_SNAPDIR, "snapdir", ZFS_SNAPDIR_HIDDEN,
 	    PROP_INHERIT, ZFS_TYPE_FILESYSTEM,

module/zfs/Makefile.in

@@ -32,6 +32,7 @@ $(MODULE)-objs += @top_srcdir@/module/zfs/dsl_synctask.o
 $(MODULE)-objs += @top_srcdir@/module/zfs/fm.o
 $(MODULE)-objs += @top_srcdir@/module/zfs/gzip.o
 $(MODULE)-objs += @top_srcdir@/module/zfs/lzjb.o
+$(MODULE)-objs += @top_srcdir@/module/zfs/lz4.o
 $(MODULE)-objs += @top_srcdir@/module/zfs/metaslab.o
 $(MODULE)-objs += @top_srcdir@/module/zfs/refcount.o
 $(MODULE)-objs += @top_srcdir@/module/zfs/rrwlock.o