ext3-fiemap-2.6.18-vanilla.patch

非常经典的一个分布式系统

Index: linux-2.6.18/fs/ext3/ioctl.c
===================================================================
--- linux-2.6.18.orig/fs/ext3/ioctl.c
+++ linux-2.6.18/fs/ext3/ioctl.c
@@ -15,7 +15,7 @@
 #include <linux/time.h>
 #include <asm/uaccess.h>
 #include <linux/namei.h>
-
+#include "fiemap.h"
 
 int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd,
 		unsigned long arg)
@@ -272,6 +272,9 @@ flags_err:
 
 		return err;
 	}
+	case EXT3_IOC_FIEMAP: {
+		return ext3_fiemap(inode, filp, cmd, arg);
+	}
 
 
 	default:
Index: linux-2.6.18/include/linux/ext3_fs.h
===================================================================
--- linux-2.6.18.orig/include/linux/ext3_fs.h
+++ linux-2.6.18/include/linux/ext3_fs.h
@@ -249,7 +249,6 @@ struct ext3_new_group_data {
 	__u32 free_blocks_count;
 };
 
-
 /*
  * ioctl commands
  */
@@ -257,15 +256,16 @@ struct ext3_new_group_data {
 #define	EXT3_IOC_SETFLAGS		_IOW('f', 2, long)
 #define	EXT3_IOC_GETVERSION		_IOR('f', 3, long)
 #define	EXT3_IOC_SETVERSION		_IOW('f', 4, long)
-#define EXT3_IOC_GROUP_EXTEND		_IOW('f', 7, unsigned long)
+#define	EXT3_IOC_GETRSVSZ		_IOR('f', 5, long)
+#define	EXT3_IOC_SETRSVSZ		_IOW('f', 6, long)
+#define	EXT3_IOC_GROUP_EXTEND		_IOW('f', 7, unsigned long)
 #define EXT3_IOC_GROUP_ADD		_IOW('f', 8,struct ext3_new_group_input)
+#define	EXT3_IOC_FIEMAP			_IOWR('f', 10, struct fiemap)
 #define	EXT3_IOC_GETVERSION_OLD		_IOR('v', 1, long)
 #define	EXT3_IOC_SETVERSION_OLD		_IOW('v', 2, long)
 #ifdef CONFIG_JBD_DEBUG
 #define EXT3_IOC_WAIT_FOR_READONLY	_IOR('f', 99, long)
 #endif
-#define EXT3_IOC_GETRSVSZ		_IOR('f', 5, long)
-#define EXT3_IOC_SETRSVSZ		_IOW('f', 6, long)
 
 /*
  *  Mount options
@@ -1117,6 +1117,8 @@ ext3_get_blocks_wrap(handle_t *handle, s
 		bh->b_size = (ret << inode->i_blkbits);
 	return ret;
 }
+extern int ext3_fiemap(struct inode *, struct file *, unsigned int,
+		       unsigned long);
 
 
 #endif	/* __KERNEL__ */
Index: linux-2.6.18/include/linux/ext3_extents.h
===================================================================
--- linux-2.6.18.orig/include/linux/ext3_extents.h
+++ linux-2.6.18/include/linux/ext3_extents.h
@@ -142,8 +142,9 @@ struct ext3_ext_path {
  * callback must return valid extent (passed or newly created)
  */
 typedef int (*ext_prepare_callback)(struct inode *, struct ext3_ext_path *,
-					struct ext3_ext_cache *,
-					void *);
+				    struct ext3_ext_cache *,
+				    struct ext3_extent *, void *);
+#define HAVE_EXT_PREPARE_CB_EXTENT
 
 #define EXT_CONTINUE	0
 #define EXT_BREAK	1
@@ -152,6 +152,26 @@ typedef int (*ext_prepare_callback)(stru
 
 #define EXT_MAX_BLOCK	0xffffffff
 
+/*
+ * EXT_INIT_MAX_LEN is the maximum number of blocks we can have in an
+ * initialized extent. This is 2^15 and not (2^16 - 1), since we use the
+ * MSB of ee_len field in the extent datastructure to signify if this
+ * particular extent is an initialized extent or an uninitialized (i.e.
+ * preallocated).
+ * EXT_UNINIT_MAX_LEN is the maximum number of blocks we can have in an
+ * uninitialized extent.
+ * If ee_len is <= 0x8000, it is an initialized extent. Otherwise, it is an
+ * uninitialized one. In other words, if MSB of ee_len is set, it is an
+ * uninitialized extent with only one special scenario when ee_len = 0x8000.
+ * In this case we can not have an uninitialized extent of zero length and
+ * thus we make it as a special case of initialized extent with 0x8000 length.
+ * This way we get better extent-to-group alignment for initialized extents.
+ * Hence, the maximum number of blocks we can have in an *initialized*
+ * extent is 2^15 (32768) and in an *uninitialized* extent is 2^15-1 (32767).
+ */
+#define EXT_INIT_MAX_LEN	(1UL << 15)
+#define EXT_UNINIT_MAX_LEN	(EXT_INIT_MAX_LEN - 1)
+
 #define EXT_FLAGS_CLR_UNKNOWN	0x7  /* Flags cleared on modification */
 #define EXT_HDR_GEN_BITS	24
 #define EXT_HDR_GEN_MASK	((1 << EXT_HDR_GEN_BITS) - 1)
@@ -219,6 +239,13 @@ ext3_ext_invalidate_cache(struct inode *
 	EXT3_I(inode)->i_cached_extent.ec_type = EXT3_EXT_CACHE_NO;
 }
 
+static inline int ext3_ext_is_uninitialized(struct ext3_extent *ext)
+{
+	/* Extent with ee_len of 0x8000 is treated as an initialized extent */
+	return (le16_to_cpu(ext->ee_len) > EXT_INIT_MAX_LEN);
+}
+
+
 extern int ext3_ext_search_left(struct inode *, struct ext3_ext_path *, unsigned long *, unsigned long *);
 extern int ext3_ext_search_right(struct inode *, struct ext3_ext_path *, unsigned long *, unsigned long *);
 extern int ext3_extent_tree_init(handle_t *, struct inode *);
Index: linux-2.6.18/fs/ext3/extents.c
===================================================================
--- linux-2.6.18.orig/fs/ext3/extents.c
+++ linux-2.6.18/fs/ext3/extents.c
@@ -42,7 +42,7 @@
 #include <linux/slab.h>
 #include <linux/ext3_extents.h>
 #include <asm/uaccess.h>
-
+#include "fiemap.h"
 
 static handle_t *ext3_ext_journal_restart(handle_t *handle, int needed)
 {
@@ -1477,7 +1477,7 @@ int ext3_ext_walk_space(struct inode *in
 		}
 
 		BUG_ON(cbex.ec_len == 0);
-		err = func(inode, path, &cbex, cbdata);
+		err = func(inode, path, &cbex, ex, cbdata);
 		ext3_ext_drop_refs(path);
 
 		if (err < 0)
@@ -2289,6 +2289,143 @@ int ext3_ext_writepage_trans_blocks(stru
 	return needed;
 }
 
+struct fiemap_internal {
+	struct fiemap	   	*fiemap_s;
+	struct fiemap_extent    fm_extent;
+	size_t		  	tot_mapping_len;
+	char		    	*cur_ext_ptr;
+	int		     	current_extent;
+	int		     	err;
+};
+
+/*
+ * Callback function called for each extent to gather fiemap information.
+ */
+int ext3_ext_fiemap_cb(struct inode *inode, struct ext3_ext_path *path,
+		       struct ext3_ext_cache *newex, struct ext3_extent *ex,
+		       void *data)
+{
+	struct fiemap_internal *fiemap_i = data;
+	struct fiemap *fiemap_s = fiemap_i->fiemap_s;
+	struct fiemap_extent *fm_extent = &fiemap_i->fm_extent;
+	int current_extent = fiemap_i->current_extent;
+	unsigned long blksize_bits = inode->i_sb->s_blocksize_bits;
+
+	/*
+	 * ext3_ext_walk_space returns a hole for extents that have not been
+	 * allocated yet.
+	 */
+	if (((u64)(newex->ec_block + newex->ec_len) << blksize_bits >=
+	     inode->i_size) && !ext3_ext_is_uninitialized(ex) &&
+	    newex->ec_type == EXT3_EXT_CACHE_GAP)
+		return EXT_BREAK;
+
+	/*
+	 * We only need to return number of extents.
+	 */
+	if (fiemap_s->fm_flags & FIEMAP_FLAG_NUM_EXTENTS)
+		goto count_extents;
+
+	if (current_extent >= fiemap_s->fm_extent_count)
+		return EXT_BREAK;
+
+	memset(fm_extent, 0, sizeof(*fm_extent));
+	fm_extent->fe_offset = (__u64)newex->ec_start << blksize_bits;
+	fm_extent->fe_length = (__u64)newex->ec_len << blksize_bits;
+	fiemap_i->tot_mapping_len += fm_extent->fe_length;
+
+	if (newex->ec_type == EXT3_EXT_CACHE_GAP)
+		fm_extent->fe_flags |= FIEMAP_EXTENT_HOLE;
+
+	if (ext3_ext_is_uninitialized(ex))
+		fm_extent->fe_flags |= (FIEMAP_EXTENT_DELALLOC |
+					FIEMAP_EXTENT_UNMAPPED);
+
+	/*
+	 * Mark this fiemap_extent as FIEMAP_EXTENT_EOF if it's past the end
+	 * of file.
+	 */
+	if ((u64)(newex->ec_block + newex->ec_len) << blksize_bits >=
+								inode->i_size)
+		fm_extent->fe_flags |= FIEMAP_EXTENT_EOF;
+
+	if (!copy_to_user(fiemap_i->cur_ext_ptr, fm_extent,
+			  sizeof(struct fiemap_extent))) {
+		fiemap_i->cur_ext_ptr += sizeof(struct fiemap_extent);
+	} else {
+		fiemap_i->err = -EFAULT;
+		return EXT_BREAK;
+	}
+
+count_extents:
+	fiemap_i->current_extent++;
+
+	/*
+	 * Stop if we are beyond requested mapping size but return complete last
+	 * extent.
+	 */
+	if ((u64)(newex->ec_block + newex->ec_len) << blksize_bits >=
+	    fiemap_s->fm_length)
+		return EXT_BREAK;
+
+	return EXT_CONTINUE;
+}
+
+int ext3_fiemap(struct inode *inode, struct file *filp, unsigned int cmd,
+		unsigned long arg)
+{
+	struct fiemap *fiemap_s;
+	struct fiemap_internal fiemap_i;
+	struct fiemap_extent *last_extent;
+	ext3_fsblk_t start_blk;
+	int err = 0;
+
+	if (!(EXT3_I(inode)->i_flags & EXT3_EXTENTS_FL))
+		return -EOPNOTSUPP;
+
+	fiemap_s = kmalloc(sizeof(*fiemap_s), GFP_KERNEL);
+	if (fiemap_s == NULL)
+		return -ENOMEM;
+	if (copy_from_user(fiemap_s, (struct fiemap __user *)arg,
+			   sizeof(*fiemap_s)))
+		return -EFAULT;
+
+	if (fiemap_s->fm_flags & FIEMAP_FLAG_INCOMPAT)
+		return -EOPNOTSUPP;
+
+	if (fiemap_s->fm_flags & FIEMAP_FLAG_SYNC)
+		ext3_sync_file(filp, filp->f_dentry, 1);
+
+	start_blk = (fiemap_s->fm_start + inode->i_sb->s_blocksize - 1) >>
+						inode->i_sb->s_blocksize_bits;
+	fiemap_i.fiemap_s = fiemap_s;
+	fiemap_i.tot_mapping_len = 0;
+	fiemap_i.cur_ext_ptr = (char *)(arg + sizeof(*fiemap_s));
+	fiemap_i.current_extent = 0;
+	fiemap_i.err = 0;
+
+	/*
+	 * Walk the extent tree gathering extent information
+	 */
+	mutex_lock(&EXT3_I(inode)->truncate_mutex);
+	err = ext3_ext_walk_space(inode, start_blk , EXT_MAX_BLOCK - start_blk,
+				  ext3_ext_fiemap_cb, &fiemap_i);
+	mutex_unlock(&EXT3_I(inode)->truncate_mutex);
+	if (err)
+		return err;
+
+	fiemap_s->fm_extent_count = fiemap_i.current_extent;
+	fiemap_s->fm_length = fiemap_i.tot_mapping_len;
+	if (fiemap_i.current_extent != 0 &&
+	    !(fiemap_s->fm_flags & FIEMAP_FLAG_NUM_EXTENTS)) {
+		last_extent = &fiemap_i.fm_extent;
+		last_extent->fe_flags |= FIEMAP_EXTENT_LAST;
+	}
+	err = copy_to_user((void *)arg, fiemap_s, sizeof(*fiemap_s));
+
+	return err;
+}
+
 EXPORT_SYMBOL(ext3_mark_inode_dirty);
 EXPORT_SYMBOL(ext3_ext_invalidate_cache);
 EXPORT_SYMBOL(ext3_ext_insert_extent);
Index: linux-2.6.18/fs/ext3/fiemap.h
===================================================================
--- /dev/null
+++ linux-2.6.18/fs/ext3/fiemap.h
@@ -0,0 +1,49 @@
+/*
+ * linux/fs/ext3/fiemap.h
+ *
+ * Copyright (C) 2007 Cluster File Systems, Inc
+ *
+ * Author: Kalpak Shah <[e-mail]>
+ */
+
+#ifndef _LINUX_EXT3_FIEMAP_H
+#define _LINUX_EXT3_FIEMAP_H
+
+struct fiemap_extent {
+	__u64	fe_offset; /* offset in bytes for the start of the extent */
+	__u64	fe_length; /* length in bytes for the extent */
+	__u32	fe_flags;  /* returned FIEMAP_EXTENT_* flags for the extent */
+	__u32	fe_lun;	   /* logical device number for extent (starting at 0)*/
+};
+
+/*
+ * fiemap is not ext3-specific and should be moved into fs.h eventually.
+ */
+
+struct fiemap {
+	__u64	fm_start;	 /* logical starting byte offset (in/out) */
+	__u64	fm_length;	 /* logical length of map (in/out) */
+	__u32	fm_flags;	 /* FIEMAP_FLAG_* flags for request (in/out) */
+	__u32	fm_extent_count; /* number of extents in fm_extents (in/out) */
+	__u64	fm_unused;
+	struct fiemap_extent	fm_extents[0];
+};
+
+#define	FIEMAP_FLAG_SYNC	0x00000001 /* sync file data before map */
+#define	FIEMAP_FLAG_HSM_READ	0x00000002 /* get data from HSM before map */
+#define	FIEMAP_FLAG_NUM_EXTENTS	0x00000004 /* return only number of extents */
+#define	FIEMAP_FLAG_INCOMPAT	0xff000000 /* error for unknown flags in here */
+
+#define	FIEMAP_EXTENT_HOLE	0x00000001 /* has no data or space allocation */
+#define	FIEMAP_EXTENT_UNWRITTEN	0x00000002 /* space allocated, but no data */
+#define	FIEMAP_EXTENT_UNMAPPED	0x00000004 /* has data but no space allocation*/
+#define	FIEMAP_EXTENT_ERROR	0x00000008 /* mapping error, errno in fe_start*/
+#define	FIEMAP_EXTENT_NO_DIRECT	0x00000010 /* cannot access data directly */
+#define	FIEMAP_EXTENT_LAST	0x00000020 /* last extent in the file */
+#define	FIEMAP_EXTENT_DELALLOC	0x00000040 /* has data but not yet written,
+					    * must have EXTENT_UNKNOWN set */
+#define	FIEMAP_EXTENT_SECONDARY	0x00000080 /* data (also) in secondary storage,
+					    * not in primary if EXTENT_UNKNOWN*/
+#define	FIEMAP_EXTENT_EOF	0x00000100 /* if fm_start+fm_len is beyond EOF*/
+
+#endif /* _LINUX_EXT3_FIEMAP_H */