Fix zlib compression

While portions of the code needed to support z_compress_level() and
z_uncompress() where in place.  In reality the current implementation
was non-functional, it just was compilable.

The critical missing component was to setup a workspace for the
compress/uncompress stream structures to use.  A kmem_cache was
added for the workspace area because we require a large chunk
of memory.  This avoids to need to continually alloc/free this
memory and vmap() the pages which is very slow.  Several objects
will reside in the per-cpu kmem_cache making them quick to acquire
and release.  A further optimization would be to adjust the
implementation to additional ensure the memory is local to the cpu.
Currently that may not be the case.

6 files changed, 238 insertions, 100 deletions

diff --git a/include/spl-debug.h b/include/spl-debug.h
index 0028c29bd..8bd4c4495 100644
--- a/include/spl-debug.h
+++ b/include/spl-debug.h
@@ -64,6 +64,7 @@
 #define SS_KSTAT	0x00020000
 #define SS_XDR		0x00040000
 #define SS_TSD		0x00080000
+#define SS_ZLIB		0x00100000
 #define SS_USER1	0x01000000
 #define SS_USER2	0x02000000
 #define SS_USER3	0x04000000
diff --git a/include/sys/zmod.h b/include/sys/zmod.h
index f1a63174a..246aa2aec 100644
--- a/include/sys/zmod.h
+++ b/include/sys/zmod.h
@@ -55,105 +55,15 @@
 #ifndef _SPL_ZMOD_H
 #define _SPL_ZMOD_H
 
+#include <sys/types.h>
 #include <linux/zlib.h>
 
-/*
- * Compresses the source buffer into the destination buffer. The level
- * parameter has the same meaning as in deflateInit.  sourceLen is the byte
- * length of the source buffer. Upon entry, destLen is the total size of the
- * destination buffer, which must be at least 0.1% larger than sourceLen plus
- * 12 bytes. Upon exit, destLen is the actual size of the compressed buffer.
- *
- * compress2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
- * memory, Z_BUF_ERROR if there was not enough room in the output buffer,
- * Z_STREAM_ERROR if the level parameter is invalid.
- */
-static __inline__ int
-z_compress_level(void *dest, size_t *destLen, const void *source,
-                 size_t sourceLen, int level)
-{
-	z_stream stream;
-	int err;
-
-	stream.next_in = (Byte *)source;
-	stream.avail_in = (uInt)sourceLen;
-#ifdef MAXSEG_64K
-	/* Check for source > 64K on 16-bit machine: */
-	if ((size_t)stream.avail_in != sourceLen)
-		return Z_BUF_ERROR;
-#endif
-	stream.next_out = dest;
-	stream.avail_out = (uInt)*destLen;
-
-	if ((size_t)stream.avail_out != *destLen)
-		return Z_BUF_ERROR;
-
-	err = zlib_deflateInit(&stream, level);
-	if (err != Z_OK)
-		return err;
-
-	err = zlib_deflate(&stream, Z_FINISH);
-	if (err != Z_STREAM_END) {
-		zlib_deflateEnd(&stream);
-		return err == Z_OK ? Z_BUF_ERROR : err;
-	}
-	*destLen = stream.total_out;
-
-	err = zlib_deflateEnd(&stream);
-	return err;
-} /* z_compress_level() */
-
-/*
- * Decompresses the source buffer into the destination buffer.  sourceLen is
- * the byte length of the source buffer. Upon entry, destLen is the total
- * size of the destination buffer, which must be large enough to hold the
- * entire uncompressed data. (The size of the uncompressed data must have
- * been saved previously by the compressor and transmitted to the decompressor
- * by some mechanism outside the scope of this compression library.)
- * Upon exit, destLen is the actual size of the compressed buffer.
- * This function can be used to decompress a whole file at once if the
- * input file is mmap'ed.
- *
- * uncompress returns Z_OK if success, Z_MEM_ERROR if there was not
- * enough memory, Z_BUF_ERROR if there was not enough room in the output
- * buffer, or Z_DATA_ERROR if the input data was corrupted.
- */
-static __inline__ int
-z_uncompress(void *dest, size_t *destLen, const void *source, size_t sourceLen)
-{
-	z_stream stream;
-	int err;
-
-	stream.next_in = (Byte *)source;
-	stream.avail_in = (uInt)sourceLen;
-	/* Check for source > 64K on 16-bit machine: */
-	if ((size_t)stream.avail_in != sourceLen)
-		return Z_BUF_ERROR;
-
-	stream.next_out = dest;
-	stream.avail_out = (uInt)*destLen;
-
-	if ((size_t)stream.avail_out != *destLen)
-		return Z_BUF_ERROR;
-
-	err = zlib_inflateInit(&stream);
-	if (err != Z_OK)
-		return err;
-
-	err = zlib_inflate(&stream, Z_FINISH);
-	if (err != Z_STREAM_END) {
-		zlib_inflateEnd(&stream);
-
-		if (err == Z_NEED_DICT ||
-		   (err == Z_BUF_ERROR && stream.avail_in == 0))
-			return Z_DATA_ERROR;
-
-		return err;
-	}
-	*destLen = stream.total_out;
+extern int z_compress_level(void *dest, size_t *destLen, const void *source,
+    size_t sourceLen, int level);
+extern int z_uncompress(void *dest, size_t *destLen, const void *source,
+    size_t sourceLen);
 
-	err = zlib_inflateEnd(&stream);
-	return err;
-} /* z_uncompress() */
+int zlib_init(void);
+void zlib_fini(void);
 
 #endif /* SPL_ZMOD_H */
diff --git a/module/spl/Makefile.in b/module/spl/Makefile.in
index 483933b64..a0211d2db 100644
--- a/module/spl/Makefile.in
+++ b/module/spl/Makefile.in
@@ -27,3 +27,4 @@ spl-objs += @top_srcdir@/module/spl/spl-condvar.o
 spl-objs += @top_srcdir@/module/spl/spl-xdr.o
 spl-objs += @top_srcdir@/module/spl/spl-cred.o
 spl-objs += @top_srcdir@/module/spl/spl-tsd.o
+spl-objs += @top_srcdir@/module/spl/spl-zlib.o
diff --git a/module/spl/spl-debug.c b/module/spl/spl-debug.c
index 2c76c7964..21e8c5d3d 100644
--- a/module/spl/spl-debug.c
+++ b/module/spl/spl-debug.c
@@ -160,6 +160,8 @@ spl_debug_subsys2str(int subsys)
                 return "xdr";
         case SS_TSD:
                 return "tsd";
+	case SS_ZLIB:
+		return "zlib";
         case SS_USER1:
                 return "user1";
         case SS_USER2:
diff --git a/module/spl/spl-generic.c b/module/spl/spl-generic.c
index b83d753d8..2b43f0c33 100644
--- a/module/spl/spl-generic.c
+++ b/module/spl/spl-generic.c
@@ -33,6 +33,7 @@
 #include <sys/rwlock.h>
 #include <sys/taskq.h>
 #include <sys/tsd.h>
+#include <sys/zmod.h>
 #include <sys/debug.h>
 #include <sys/proc.h>
 #include <sys/kstat.h>
@@ -471,20 +472,25 @@ __init spl_init(void)
 	if ((rc = tsd_init()))
 		SGOTO(out8, rc);
 
+	if ((rc = zlib_init()))
+		SGOTO(out9, rc);
+
 	if ((rc = set_hostid()))
-		SGOTO(out9, rc = -EADDRNOTAVAIL);
+		SGOTO(out10, rc = -EADDRNOTAVAIL);
 
 #ifndef HAVE_KALLSYMS_LOOKUP_NAME
 	if ((rc = set_kallsyms_lookup_name()))
-		SGOTO(out9, rc = -EADDRNOTAVAIL);
+		SGOTO(out10, rc = -EADDRNOTAVAIL);
 #endif /* HAVE_KALLSYMS_LOOKUP_NAME */
 
 	if ((rc = spl_kmem_init_kallsyms_lookup()))
-		SGOTO(out9, rc);
+		SGOTO(out10, rc);
 
 	printk(KERN_NOTICE "SPL: Loaded Solaris Porting Layer v%s%s\n",
 	       SPL_META_VERSION, SPL_DEBUG_STR);
 	SRETURN(rc);
+out10:
+	zlib_fini();
 out9:
 	tsd_fini();
 out8:
@@ -516,6 +522,7 @@ spl_fini(void)
 
 	printk(KERN_NOTICE "SPL: Unloaded Solaris Porting Layer v%s%s\n",
 	       SPL_META_VERSION, SPL_DEBUG_STR);
+	zlib_fini();
 	tsd_fini();
 	kstat_fini();
 	proc_fini();
diff --git a/module/spl/spl-zlib.c b/module/spl/spl-zlib.c
new file mode 100644
index 000000000..02825b461
--- /dev/null
+++ b/module/spl/spl-zlib.c
@@ -0,0 +1,217 @@
+/*****************************************************************************\
+ *  Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
+ *  Copyright (C) 2007 The Regents of the University of California.
+ *  Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ *  Written by Brian Behlendorf <[email protected]>.
+ *  UCRL-CODE-235197
+ *
+ *  This file is part of the SPL, Solaris Porting Layer.
+ *  For details, see <http://github.com/behlendorf/spl/>.
+ *
+ *  The SPL is free software; you can redistribute it and/or modify it
+ *  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.
+ *
+ *  The SPL 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 the SPL.  If not, see <http://www.gnu.org/licenses/>.
+ *****************************************************************************
+ *  z_compress_level/z_uncompress are nearly identical copies of the
+ *  compress2/uncompress functions provided by the official zlib package
+ *  available at http://zlib.net/.  The only changes made we to slightly
+ *  adapt the functions called to match the linux kernel implementation
+ *  of zlib.  The full zlib license follows:
+ *
+ *  zlib.h -- interface of the 'zlib' general purpose compression library
+ *  version 1.2.5, April 19th, 2010
+ *
+ *  Copyright (C) 1995-2010 Jean-loup Gailly and Mark Adler
+ *
+ *  This software is provided 'as-is', without any express or implied
+ *  warranty.  In no event will the authors be held liable for any damages
+ *  arising from the use of this software.
+ *
+ *  Permission is granted to anyone to use this software for any purpose,
+ *  including commercial applications, and to alter it and redistribute it
+ *  freely, subject to the following restrictions:
+ *
+ *  1. The origin of this software must not be misrepresented; you must not
+ *     claim that you wrote the original software. If you use this software
+ *     in a product, an acknowledgment in the product documentation would be
+ *     appreciated but is not required.
+ *  2. Altered source versions must be plainly marked as such, and must not be
+ *     misrepresented as being the original software.
+ *  3. This notice may not be removed or altered from any source distribution.
+ *
+ *  Jean-loup Gailly
+ *  Mark Adler
+\*****************************************************************************/
+
+
+#include <sys/kmem.h>
+#include <sys/zmod.h>
+#include <spl-debug.h>
+
+#ifdef DEBUG_SUBSYSTEM
+#undef DEBUG_SUBSYSTEM
+#endif
+
+#define DEBUG_SUBSYSTEM SS_ZLIB
+
+static spl_kmem_cache_t *zlib_workspace_cache;
+
+/*
+ * A kmem_cache is used for the zlib workspaces to avoid having to vmalloc
+ * and vfree for every call.  Using a kmem_cache also has the advantage
+ * that improves the odds that the memory used will be local to this cpu.
+ * To further improve things it might be wise to create a dedicated per-cpu
+ * workspace for use.  This would take some additional care because we then
+ * must disable preemption around the critical section, and verify that
+ * zlib_deflate* and zlib_inflate* never internally call schedule().
+ */
+static void *
+zlib_workspace_alloc(int flags)
+{
+	return kmem_cache_alloc(zlib_workspace_cache, flags & ~(__GFP_FS));
+}
+
+static void
+zlib_workspace_free(void *workspace)
+{
+	kmem_cache_free(zlib_workspace_cache, workspace);
+}
+
+/*
+ * Compresses the source buffer into the destination buffer. The level
+ * parameter has the same meaning as in deflateInit.  sourceLen is the byte
+ * length of the source buffer. Upon entry, destLen is the total size of the
+ * destination buffer, which must be at least 0.1% larger than sourceLen plus
+ * 12 bytes. Upon exit, destLen is the actual size of the compressed buffer.
+ *
+ * compress2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
+ * memory, Z_BUF_ERROR if there was not enough room in the output buffer,
+ * Z_STREAM_ERROR if the level parameter is invalid.
+ */
+int
+z_compress_level(void *dest, size_t *destLen, const void *source,
+                 size_t sourceLen, int level)
+{
+	z_stream stream;
+	int err;
+
+	stream.next_in = (Byte *)source;
+	stream.avail_in = (uInt)sourceLen;
+	stream.next_out = dest;
+	stream.avail_out = (uInt)*destLen;
+
+	if ((size_t)stream.avail_out != *destLen)
+		return Z_BUF_ERROR;
+
+	stream.workspace = zlib_workspace_alloc(KM_SLEEP);
+	if (!stream.workspace)
+		return Z_MEM_ERROR;
+
+	err = zlib_deflateInit(&stream, level);
+	if (err != Z_OK) {
+		zlib_workspace_free(stream.workspace);
+		return err;
+	}
+
+	err = zlib_deflate(&stream, Z_FINISH);
+	if (err != Z_STREAM_END) {
+		zlib_deflateEnd(&stream);
+		zlib_workspace_free(stream.workspace);
+		return err == Z_OK ? Z_BUF_ERROR : err;
+	}
+	*destLen = stream.total_out;
+
+	err = zlib_deflateEnd(&stream);
+	zlib_workspace_free(stream.workspace);
+
+	return err;
+}
+EXPORT_SYMBOL(z_compress_level);
+
+/*
+ * Decompresses the source buffer into the destination buffer.  sourceLen is
+ * the byte length of the source buffer. Upon entry, destLen is the total
+ * size of the destination buffer, which must be large enough to hold the
+ * entire uncompressed data. (The size of the uncompressed data must have
+ * been saved previously by the compressor and transmitted to the decompressor
+ * by some mechanism outside the scope of this compression library.)
+ * Upon exit, destLen is the actual size of the compressed buffer.
+ * This function can be used to decompress a whole file at once if the
+ * input file is mmap'ed.
+ *
+ * uncompress returns Z_OK if success, Z_MEM_ERROR if there was not
+ * enough memory, Z_BUF_ERROR if there was not enough room in the output
+ * buffer, or Z_DATA_ERROR if the input data was corrupted.
+ */
+int
+z_uncompress(void *dest, size_t *destLen, const void *source, size_t sourceLen)
+{
+	z_stream stream;
+	int err;
+
+	stream.next_in = (Byte *)source;
+	stream.avail_in = (uInt)sourceLen;
+	stream.next_out = dest;
+	stream.avail_out = (uInt)*destLen;
+
+	if ((size_t)stream.avail_out != *destLen)
+		return Z_BUF_ERROR;
+
+	stream.workspace = zlib_workspace_alloc(KM_SLEEP);
+	if (!stream.workspace)
+		return Z_MEM_ERROR;
+
+	err = zlib_inflateInit(&stream);
+	if (err != Z_OK) {
+		zlib_workspace_free(stream.workspace);
+		return err;
+	}
+
+	err = zlib_inflate(&stream, Z_FINISH);
+	if (err != Z_STREAM_END) {
+		zlib_inflateEnd(&stream);
+		zlib_workspace_free(stream.workspace);
+
+		if (err == Z_NEED_DICT ||
+		   (err == Z_BUF_ERROR && stream.avail_in == 0))
+			return Z_DATA_ERROR;
+
+		return err;
+	}
+	*destLen = stream.total_out;
+
+	err = zlib_inflateEnd(&stream);
+	zlib_workspace_free(stream.workspace);
+
+	return err;
+}
+EXPORT_SYMBOL(z_uncompress);
+
+int zlib_init(void)
+{
+        SENTRY;
+	zlib_workspace_cache = kmem_cache_create("spl_zlib_workspace_cache",
+	    max(zlib_deflate_workspacesize(), zlib_inflate_workspacesize()),
+	    0, NULL, NULL, NULL, NULL, NULL, KMC_VMEM);
+        if (!zlib_workspace_cache)
+		SRETURN(1);
+
+        SRETURN(0);
+}
+
+void zlib_fini(void)
+{
+        SENTRY;
+	kmem_cache_destroy(zlib_workspace_cache);
+        zlib_workspace_cache = NULL;
+        SEXIT;
+}