grub/patches/grub-0.97-ext4-3.patch

diff -Naur grub-0.97/stage2/fsys_ext2fs.c grub-0.97-magellan/stage2/fsys_ext2fs.c
--- grub-0.97/stage2/fsys_ext2fs.c	2004-08-08 20:19:18.000000000 +0200
+++ grub-0.97-magellan/stage2/fsys_ext2fs.c	2009-10-16 17:55:35.000000000 +0200
@@ -41,6 +41,7 @@
 typedef unsigned short __u16;
 typedef __signed__ int __s32;
 typedef unsigned int __u32;
+typedef unsigned long long __u64;
 
 /*
  * Constants relative to the data blocks, from ext2_fs.h
@@ -51,7 +52,7 @@
 #define EXT2_TIND_BLOCK                 (EXT2_DIND_BLOCK + 1)
 #define EXT2_N_BLOCKS                   (EXT2_TIND_BLOCK + 1)
 
-/* include/linux/ext2_fs.h */
+/* lib/ext2fs/ext2_fs.h from e2fsprogs */
 struct ext2_super_block
   {
     __u32 s_inodes_count;	/* Inodes count */
@@ -61,9 +62,9 @@
     __u32 s_free_inodes_count;	/* Free inodes count */
     __u32 s_first_data_block;	/* First Data Block */
     __u32 s_log_block_size;	/* Block size */
-    __s32 s_log_frag_size;	/* Fragment size */
+    __s32 s_obso_log_frag_size;	/* Obsoleted Fragment size */
     __u32 s_blocks_per_group;	/* # Blocks per group */
-    __u32 s_frags_per_group;	/* # Fragments per group */
+    __u32 s_obso_frags_per_group;	/* Obsoleted Fragments per group */
     __u32 s_inodes_per_group;	/* # Inodes per group */
     __u32 s_mtime;		/* Mount time */
     __u32 s_wtime;		/* Write time */
@@ -72,17 +73,77 @@
     __u16 s_magic;		/* Magic signature */
     __u16 s_state;		/* File system state */
     __u16 s_errors;		/* Behaviour when detecting errors */
-    __u16 s_pad;
+    __u16 s_minor_rev_level;	/* minor revision level */
+    __u16 s_desc_size;		/* size of group descriptor */
     __u32 s_lastcheck;		/* time of last check */
     __u32 s_checkinterval;	/* max. time between checks */
     __u32 s_creator_os;		/* OS */
     __u32 s_rev_level;		/* Revision level */
     __u16 s_def_resuid;		/* Default uid for reserved blocks */
     __u16 s_def_resgid;		/* Default gid for reserved blocks */
-    __u32 s_reserved[235];	/* Padding to the end of the block */
-  };
+    /*
+     * These fields are for EXT2_DYNAMIC_REV superblocks only.
+     *
+     * Note: the difference between the compatible feature set and
+     * the incompatible feature set is that if there is a bit set
+     * in the incompatible feature set that the kernel doesn't
+     * know about, it should refuse to mount the filesystem.
+     *
+     * e2fsck's requirements are more strict; if it doesn't know
+     * about a feature in either the compatible or incompatible
+     * feature set, it must abort and not try to meddle with
+     * things it doesn't understand...
+     */
+    __u32 s_first_ino;		/* First non-reserved inode */
+    __u16 s_inode_size;		/* size of inode structure */
+    __u16 s_block_group_nr;	/* block group # of this superblock */
+    __u32 s_feature_compat;	/* compatible feature set */
+    __u32 s_feature_incompat;	/* incompatible feature set */
+    __u32 s_feature_ro_compat;	/* readonly-compatible feature set */
+    __u8  s_uuid[16];		/* 128-bit uuid for volume */
+    char  s_volume_name[16];	/* volume name */
+    char  s_last_mounted[64];	/* directory where last mounted */
+    __u32 s_algorithm_usage_bitmap; /* For compression */
+    /*
+     * Performance hints.  Directory preallocation should only
+     * happen if the EXT2_FEATURE_COMPAT_DIR_PREALLOC flag is on.
+     */
+    __u8  s_prealloc_blocks;	/* Nr of blocks to try to preallocate*/
+    __u8  s_prealloc_dir_blocks;	/* Nr to preallocate for dirs */
+    __u16 s_reserved_gdt_blocks;/* Per group table for online growth */
+    /*
+     * Journaling support valid if EXT2_FEATURE_COMPAT_HAS_JOURNAL set.
+     */
+    __u8 s_journal_uuid[16];	/* uuid of journal superblock */
+    __u32 s_journal_inum;	/* inode number of journal file */
+    __u32 s_journal_dev;	/* device number of journal file */
+    __u32 s_last_orphan;	/* start of list of inodes to delete */
+    __u32 s_hash_seed[4];	/* HTREE hash seed */
+    __u8  s_def_hash_version;	/* Default hash version to use */
+    __u8  s_jnl_backup_type; 	/* Default type of journal backup */
+    __u16 s_reserved_word_pad;
+    __u32 s_default_mount_opts;
+    __u32 s_first_meta_bg;	/* First metablock group */
+    __u32 s_mkfs_time;		/* When the filesystem was created */
+    __u32 s_jnl_blocks[17]; 	/* Backup of the journal inode */
+    /* 64bit desc support valid if EXT4_FEATURE_INCOMPAT_64BIT */
+    __u32 s_blocks_count_hi;	/* Blocks count */
+    __u32 s_r_blocks_count_hi;	/* Reserved blocks count */
+    __u32 s_free_blocks_count_hi; /* Free blocks count */
+    __u16 s_min_extra_isize;	/* All inodes have at least # bytes */
+    __u16 s_max_extra_isize;	/* New inodes should reverve # bytes */
+    __u32 s_flags;		/* Miscellaneous flags */
+    __u16 s_raid_stride;	/* Raid stride */
+    __u16 s_mmp_interval;	/* # seconds to wait MMP checking */
+    __u64 s_mmp_block;		/* Block for multi-mount protection */
+    __u32 s_raid_stripe_width;	/* Blocks on all data disks (N*stride)*/
+    __u8  s_log_groups_per_flex;/* FLEX_BG group size*/
+    __u8  s_reserved_char_pad;
+    __u16 s_reserved_pad;
+    __u32 s_reserved[162];	/* Padding to the end of the block */
+};
 
-struct ext2_group_desc
+struct ext4_group_desc
   {
     __u32 bg_block_bitmap;	/* Blocks bitmap block */
     __u32 bg_inode_bitmap;	/* Inodes bitmap block */
@@ -90,8 +151,18 @@
     __u16 bg_free_blocks_count;	/* Free blocks count */
     __u16 bg_free_inodes_count;	/* Free inodes count */
     __u16 bg_used_dirs_count;	/* Directories count */
-    __u16 bg_pad;
-    __u32 bg_reserved[3];
+    __u16 bg_flags;		/* EXT4_BG_flags (INODE_UNINIT, etc) */
+    __u32 bg_reserved[2];		/* Likely block/inode bitmap checksum */
+    __u16 bg_itable_unused;	/* Unused inodes count */
+    __u16 bg_checksum;		/* crc16(sb_uuid+group+desc) */
+    __u32 bg_block_bitmap_hi;	/* Blocks bitmap block MSB */
+    __u32 bg_inode_bitmap_hi;	/* Inodes bitmap block MSB */
+    __u32 bg_inode_table_hi;	/* Inodes table block MSB */
+    __u16 bg_free_blocks_count_hi;/* Free blocks count MSB */
+    __u16 bg_free_inodes_count_hi;/* Free inodes count MSB */
+    __u16 bg_used_dirs_count_hi;	/* Directories count MSB */
+    __u16 bg_itable_unused_hi;	/* Unused inodes count MSB */
+    __u32 bg_reserved2[3];
   };
 
 struct ext2_inode
@@ -129,22 +200,22 @@
     __u32 i_block[EXT2_N_BLOCKS];	/* 40: Pointers to blocks */
     __u32 i_version;		/* File version (for NFS) */
     __u32 i_file_acl;		/* File ACL */
-    __u32 i_dir_acl;		/* Directory ACL */
-    __u32 i_faddr;		/* Fragment address */
+    __u32 i_size_high;
+    __u32 i_obso_faddr;		/* Obsoleted fragment address */
     union
       {
 	struct
 	  {
-	    __u8 l_i_frag;	/* Fragment number */
-	    __u8 l_i_fsize;	/* Fragment size */
-	    __u16 i_pad1;
-	    __u32 l_i_reserved2[2];
+		__u16	l_i_blocks_high; /* were l_i_reserved1 */
+		__u16	l_i_file_acl_high;
+		__u16	l_i_uid_high;	/* these 2 fields */
+		__u16	l_i_gid_high;	/* were reserved2[0] */
+		__u32	l_i_reserved2;
 	  }
 	linux2;
 	struct
 	  {
-	    __u8 h_i_frag;	/* Fragment number */
-	    __u8 h_i_fsize;	/* Fragment size */
+		__u16	h_i_reserved1;	/* Obsoleted fragment number/size which are removed in ext4 */
 	    __u16 h_i_mode_high;
 	    __u16 h_i_uid_high;
 	    __u16 h_i_gid_high;
@@ -153,16 +224,36 @@
 	hurd2;
 	struct
 	  {
-	    __u8 m_i_frag;	/* Fragment number */
-	    __u8 m_i_fsize;	/* Fragment size */
-	    __u16 m_pad1;
+		__u16	h_i_reserved1;	/* Obsoleted fragment number/size which are removed in ext4 */
+		__u16	m_i_file_acl_high;
 	    __u32 m_i_reserved2[2];
 	  }
 	masix2;
       }
     osd2;			/* OS dependent 2 */
+	__u16	i_extra_isize;
+	__u16	i_pad1;
+	__u32  i_ctime_extra;  /* extra Change time      (nsec << 2 | epoch) */
+	__u32  i_mtime_extra;  /* extra Modification time(nsec << 2 | epoch) */
+	__u32  i_atime_extra;  /* extra Access time      (nsec << 2 | epoch) */
+	__u32  i_crtime;       /* File Creation time */
+	__u32  i_crtime_extra; /* extra FileCreationtime (nsec << 2 | epoch) */
+	__u32  i_version_hi;	/* high 32 bits for 64-bit version */
   };
 
+#define EXT4_FEATURE_INCOMPAT_EXTENTS		0x0040 /* extents support */
+#define EXT4_FEATURE_INCOMPAT_64BIT			0x0080 /* grub not supported*/
+#define EXT4_FEATURE_INCOMPAT_MMP           0x0100
+#define EXT4_FEATURE_INCOMPAT_FLEX_BG		0x0200
+
+#define EXT4_HAS_INCOMPAT_FEATURE(sb,mask)			\
+	( sb->s_feature_incompat & mask )
+
+#define EXT4_EXTENTS_FL		0x00080000 /* Inode uses extents */
+#define EXT4_HUGE_FILE_FL	0x00040000 /* Set to each huge file */
+
+#define EXT4_MIN_DESC_SIZE			32
+
 /* linux/limits.h */
 #define NAME_MAX         255	/* # chars in a file name */
 
@@ -180,6 +271,56 @@
     char name[EXT2_NAME_LEN];	/* File name */
   };
 
+/* linux/ext4_fs_extents.h */
+/* This is the extent on-disk structure.
+ * It's used at the bottom of the tree.
+ */
+struct ext4_extent
+  {
+	__u32  ee_block;   /* first logical block extent covers */
+	__u16  ee_len;     /* number of blocks covered by extent */
+	__u16  ee_start_hi;    /* high 16 bits of physical block */
+	__u32  ee_start_lo;    /* low 32 bits of physical block */
+  };
+
+/*
+ * This is index on-disk structure.
+ * It's used at all the levels except the bottom.
+ */
+struct ext4_extent_idx
+  {
+    __u32  ei_block;   /* index covers logical blocks from 'block' */
+    __u32  ei_leaf_lo; /* pointer to the physical block of the next *
+	                     * level. leaf or next index could be there */
+    __u16  ei_leaf_hi; /* high 16 bits of physical block */
+    __u16  ei_unused;
+  };
+
+/*
+ * Each block (leaves and indexes), even inode-stored has header.
+ */
+struct ext4_extent_header
+  {
+    __u16  eh_magic;   /* probably will support different formats */
+    __u16  eh_entries; /* number of valid entries */
+    __u16  eh_max;     /* capacity of store in entries */
+    __u16  eh_depth;   /* has tree real underlying blocks? */
+    __u32  eh_generation;  /* generation of the tree */
+  };
+
+#define EXT4_EXT_MAGIC      (0xf30a)
+#define EXT_FIRST_EXTENT(__hdr__) \
+    ((struct ext4_extent *) (((char *) (__hdr__)) +     \
+                 sizeof(struct ext4_extent_header)))
+#define EXT_FIRST_INDEX(__hdr__) \
+    ((struct ext4_extent_idx *) (((char *) (__hdr__)) + \
+                 sizeof(struct ext4_extent_header)))
+#define EXT_LAST_EXTENT(__hdr__) \
+    (EXT_FIRST_EXTENT((__hdr__)) + (__u16)((__hdr__)->eh_entries) - 1)
+#define EXT_LAST_INDEX(__hdr__) \
+    (EXT_FIRST_INDEX((__hdr__)) + (__u16)((__hdr__)->eh_entries) - 1)
+
+
 /* linux/ext2fs.h */
 /*
  * EXT2_DIR_PAD defines the directory entries boundaries
@@ -223,8 +364,17 @@
 /* kind of from ext2/super.c */
 #define EXT2_BLOCK_SIZE(s)	(1 << EXT2_BLOCK_SIZE_BITS(s))
 /* linux/ext2fs.h */
+/* sizeof(struct ext2_group_desc) is changed in ext4
+ * in kernel code, ext2/3 uses sizeof(struct ext2_group_desc) to calculate
+ * number of desc per block, while ext4 uses superblock->s_desc_size in stead
+ * superblock->s_desc_size is not available in ext2/3
+ * */
+#define EXT2_DESC_SIZE(s) \
+	(EXT4_HAS_INCOMPAT_FEATURE(s,EXT4_FEATURE_INCOMPAT_64BIT)? \
+	s->s_desc_size : EXT4_MIN_DESC_SIZE)
 #define EXT2_DESC_PER_BLOCK(s) \
-     (EXT2_BLOCK_SIZE(s) / sizeof (struct ext2_group_desc))
+	(EXT2_BLOCK_SIZE(s) / EXT2_DESC_SIZE(s))
+
 /* linux/stat.h */
 #define S_IFMT  00170000
 #define S_IFLNK  0120000
@@ -386,6 +536,122 @@
     [logical_block & (EXT2_ADDR_PER_BLOCK (SUPERBLOCK) - 1)];
 }
 
+/* extent binary search index
+ * find closest index in the current level extent tree
+ * kind of from ext4_ext_binsearch_idx in ext4/extents.c
+ */
+static struct ext4_extent_idx*
+ext4_ext_binsearch_idx(struct ext4_extent_header* eh, int logical_block)
+{
+  struct ext4_extent_idx *r, *l, *m;
+  l = EXT_FIRST_INDEX(eh) + 1;
+  r = EXT_LAST_INDEX(eh);
+  while (l <= r)
+    {
+	  m = l + (r - l) / 2;
+	  if (logical_block < m->ei_block)
+		  r = m - 1;
+	  else
+		  l = m + 1;
+	}
+  return (struct ext4_extent_idx*)(l - 1);
+}
+
+/* extent binary search
+ * find closest extent in the leaf level
+ * kind of from ext4_ext_binsearch in ext4/extents.c
+ */
+static struct ext4_extent*
+ext4_ext_binsearch(struct ext4_extent_header* eh, int logical_block)
+{
+  struct ext4_extent *r, *l, *m;
+  l = EXT_FIRST_EXTENT(eh) + 1;
+  r = EXT_LAST_EXTENT(eh);
+  while (l <= r)
+    {
+	  m = l + (r - l) / 2;
+	  if (logical_block < m->ee_block)
+		  r = m - 1;
+	  else
+		  l = m + 1;
+	}
+  return (struct ext4_extent*)(l - 1);
+}
+
+/* Maps extents enabled logical block into physical block via an inode.
+ * EXT4_HUGE_FILE_FL should be checked before calling this.
+ */
+static int
+ext4fs_block_map (int logical_block)
+{
+  struct ext4_extent_header *eh;
+  struct ext4_extent *ex, *extent;
+  struct ext4_extent_idx *ei, *index;
+  int depth;
+  int i;
+
+#ifdef E2DEBUG
+  unsigned char *i;
+  for (i = (unsigned char *) INODE;
+       i < ((unsigned char *) INODE + sizeof (struct ext2_inode));
+       i++)
+    {
+      printf ("%c", "0123456789abcdef"[*i >> 4]);
+      printf ("%c", "0123456789abcdef"[*i % 16]);
+      if (!((i + 1 - (unsigned char *) INODE) % 16))
+	{
+	  printf ("\n");
+	}
+      else
+	{
+	  printf (" ");
+	}
+    }
+  printf ("logical block %d\n", logical_block);
+#endif /* E2DEBUG */
+  eh = (struct ext4_extent_header*)INODE->i_block;
+  if (eh->eh_magic != EXT4_EXT_MAGIC)
+  {
+          errnum = ERR_FSYS_CORRUPT;
+          return -1;
+  }
+  while((depth = eh->eh_depth) != 0)
+  	{ /* extent index */
+	  if (eh->eh_magic != EXT4_EXT_MAGIC)
+	  {
+	          errnum = ERR_FSYS_CORRUPT;
+		  return -1;
+	  }
+	  ei = ext4_ext_binsearch_idx(eh, logical_block);
+	  if (ei->ei_leaf_hi)
+	{/* 64bit physical block number not supported */
+	  errnum = ERR_FILELENGTH;
+	  return -1;
+	}
+	  if (!ext2_rdfsb(ei->ei_leaf_lo, DATABLOCK1))
+	{
+	  errnum = ERR_FSYS_CORRUPT;
+	  return -1;
+	}
+	  eh = (struct ext4_extent_header*)DATABLOCK1;
+  	}
+
+  /* depth==0, we come to the leaf */
+  ex = ext4_ext_binsearch(eh, logical_block);
+  if (ex->ee_start_hi)
+	{/* 64bit physical block number not supported */
+	  errnum = ERR_FILELENGTH;
+	  return -1;
+	}
+  if ((ex->ee_block + ex->ee_len) < logical_block)
+	{
+	  errnum = ERR_FSYS_CORRUPT;
+	  return -1;
+	}
+  return ex->ee_start_lo + logical_block - ex->ee_block;
+
+}
+
 /* preconditions: all preconds of ext2fs_block_map */
 int
 ext2fs_read (char *buf, int len)
@@ -420,6 +686,11 @@
       /* find the (logical) block component of our location */
       logical_block = filepos >> EXT2_BLOCK_SIZE_BITS (SUPERBLOCK);
       offset = filepos & (EXT2_BLOCK_SIZE (SUPERBLOCK) - 1);
+      /* map extents enabled logical block number to physical fs on-disk block number */
+      if (EXT4_HAS_INCOMPAT_FEATURE(SUPERBLOCK,EXT4_FEATURE_INCOMPAT_EXTENTS)
+                    && INODE->i_flags & EXT4_EXTENTS_FL)
+          map = ext4fs_block_map (logical_block);
+      else
       map = ext2fs_block_map (logical_block);
 #ifdef E2DEBUG
       printf ("map=%d\n", map);
@@ -504,7 +775,7 @@
   int desc;			/* index within that group */
   int ino_blk;			/* fs pointer of the inode's information */
   int str_chk = 0;		/* used to hold the results of a string compare */
-  struct ext2_group_desc *gdp;
+  struct ext4_group_desc *ext4_gdp;
   struct ext2_inode *raw_inode;	/* inode info corresponding to current_ino */
 
   char linkbuf[PATH_MAX];	/* buffer for following symbolic links */
@@ -550,8 +821,15 @@
 	{
 	  return 0;
 	}
-      gdp = GROUP_DESC;
-      ino_blk = gdp[desc].bg_inode_table +
+	  ext4_gdp = (struct ext4_group_desc *)( (__u8*)GROUP_DESC +
+			  		desc * EXT2_DESC_SIZE(SUPERBLOCK));
+	  if (EXT4_HAS_INCOMPAT_FEATURE(SUPERBLOCK, EXT4_FEATURE_INCOMPAT_64BIT)
+		&& (! ext4_gdp->bg_inode_table_hi))
+	{/* 64bit itable not supported */
+	  errnum = ERR_FILELENGTH;
+	  return -1;
+	}
+      ino_blk = ext4_gdp->bg_inode_table +
 	(((current_ino - 1) % (SUPERBLOCK->s_inodes_per_group))
 	 >> log2 (EXT2_BLOCK_SIZE (SUPERBLOCK) / sizeof (struct ext2_inode)));
 #ifdef E2DEBUG
@@ -629,7 +907,10 @@
 	    }
 	  linkbuf[filemax + len] = '\0';
 
-	  /* Read the symlink data. */
+	  /* Read the symlink data.
+	   * Slow symlink is extents enabled
+	   * But since grub_read invokes ext2fs_read, nothing to change here
+	   */
 	  if (! ext2_is_fast_symlink ())
 	    {
 	      /* Read the necessary blocks, and reset the file pointer. */
@@ -640,7 +921,9 @@
 	    }
 	  else
 	    {
-	      /* Copy the data directly from the inode. */
+	      /* Copy the data directly from the inode.
+	       * Fast symlink is not extents enabled
+	       */
 	      len = filemax;
 	      memmove (linkbuf, (char *) INODE->i_block, len);
 	    }
@@ -674,7 +957,13 @@
 	      errnum = ERR_BAD_FILETYPE;
 	      return 0;
 	    }
-
+	  /* if file is too large, just stop and report an error*/
+	  if ( (INODE->i_flags & EXT4_HUGE_FILE_FL) && !(INODE->i_size_high))
+	    {
+		  /* file too large, stop reading */
+		  errnum = ERR_FILELENGTH;
+		  return 0;
+	    }
 	  filemax = (INODE->i_size);
 	  return 1;
 	}
@@ -728,17 +1017,28 @@
 	    }
 
 	  /* else, find the (logical) block component of our location */
+	  /* ext4 logical block number the same as ext2/3 */
 	  blk = loc >> EXT2_BLOCK_SIZE_BITS (SUPERBLOCK);
 
 	  /* we know which logical block of the directory entry we are looking
 	     for, now we have to translate that to the physical (fs) block on
 	     the disk */
+	  /* map extents enabled logical block number to physical fs on-disk block number */
+	  if (EXT4_HAS_INCOMPAT_FEATURE(SUPERBLOCK,EXT4_FEATURE_INCOMPAT_EXTENTS)
+                        && INODE->i_flags & EXT4_EXTENTS_FL)
+              map = ext4fs_block_map (blk);
+	  else
 	  map = ext2fs_block_map (blk);
 #ifdef E2DEBUG
 	  printf ("fs block=%d\n", map);
 #endif /* E2DEBUG */
 	  mapblock2 = -1;
-	  if ((map < 0) || !ext2_rdfsb (map, DATABLOCK2))
+	  if (map < 0)
+	  {
+	      *rest = ch;
+	      return 0;
+	  }
+          if (!ext2_rdfsb (map, DATABLOCK2))
 	    {
 	      errnum = ERR_FSYS_CORRUPT;
 	      *rest = ch;
1	diff -Naur grub-0.97/stage2/fsys_ext2fs.c grub-0.97-magellan/stage2/fsys_ext2fs.c
2	--- grub-0.97/stage2/fsys_ext2fs.c 2004-08-08 20:19:18.000000000 +0200
3	+++ grub-0.97-magellan/stage2/fsys_ext2fs.c 2009-10-16 17:55:35.000000000 +0200
4	@@ -41,6 +41,7 @@
5	typedef unsigned short __u16;
6	typedef __signed__ int __s32;
7	typedef unsigned int __u32;
8	+typedef unsigned long long __u64;
9
10	/*
11	* Constants relative to the data blocks, from ext2_fs.h
12	@@ -51,7 +52,7 @@
13	#define EXT2_TIND_BLOCK (EXT2_DIND_BLOCK + 1)
14	#define EXT2_N_BLOCKS (EXT2_TIND_BLOCK + 1)
15
16	-/* include/linux/ext2_fs.h */
17	+/* lib/ext2fs/ext2_fs.h from e2fsprogs */
18	struct ext2_super_block
19	{
20	__u32 s_inodes_count; /* Inodes count */
21	@@ -61,9 +62,9 @@
22	__u32 s_free_inodes_count; /* Free inodes count */
23	__u32 s_first_data_block; /* First Data Block */
24	__u32 s_log_block_size; /* Block size */
25	- __s32 s_log_frag_size; /* Fragment size */
26	+ __s32 s_obso_log_frag_size; /* Obsoleted Fragment size */
27	__u32 s_blocks_per_group; /* # Blocks per group */
28	- __u32 s_frags_per_group; /* # Fragments per group */
29	+ __u32 s_obso_frags_per_group; /* Obsoleted Fragments per group */
30	__u32 s_inodes_per_group; /* # Inodes per group */
31	__u32 s_mtime; /* Mount time */
32	__u32 s_wtime; /* Write time */
33	@@ -72,17 +73,77 @@
34	__u16 s_magic; /* Magic signature */
35	__u16 s_state; /* File system state */
36	__u16 s_errors; /* Behaviour when detecting errors */
37	- __u16 s_pad;
38	+ __u16 s_minor_rev_level; /* minor revision level */
39	+ __u16 s_desc_size; /* size of group descriptor */
40	__u32 s_lastcheck; /* time of last check */
41	__u32 s_checkinterval; /* max. time between checks */
42	__u32 s_creator_os; /* OS */
43	__u32 s_rev_level; /* Revision level */
44	__u16 s_def_resuid; /* Default uid for reserved blocks */
45	__u16 s_def_resgid; /* Default gid for reserved blocks */
46	- __u32 s_reserved[235]; /* Padding to the end of the block */
47	- };
48	+ /*
49	+ * These fields are for EXT2_DYNAMIC_REV superblocks only.
50	+ *
51	+ * Note: the difference between the compatible feature set and
52	+ * the incompatible feature set is that if there is a bit set
53	+ * in the incompatible feature set that the kernel doesn't
54	+ * know about, it should refuse to mount the filesystem.
55	+ *
56	+ * e2fsck's requirements are more strict; if it doesn't know
57	+ * about a feature in either the compatible or incompatible
58	+ * feature set, it must abort and not try to meddle with
59	+ * things it doesn't understand...
60	+ */
61	+ __u32 s_first_ino; /* First non-reserved inode */
62	+ __u16 s_inode_size; /* size of inode structure */
63	+ __u16 s_block_group_nr; /* block group # of this superblock */
64	+ __u32 s_feature_compat; /* compatible feature set */
65	+ __u32 s_feature_incompat; /* incompatible feature set */
66	+ __u32 s_feature_ro_compat; /* readonly-compatible feature set */
67	+ __u8 s_uuid[16]; /* 128-bit uuid for volume */
68	+ char s_volume_name[16]; /* volume name */
69	+ char s_last_mounted[64]; /* directory where last mounted */
70	+ __u32 s_algorithm_usage_bitmap; /* For compression */
71	+ /*
72	+ * Performance hints. Directory preallocation should only
73	+ * happen if the EXT2_FEATURE_COMPAT_DIR_PREALLOC flag is on.
74	+ */
75	+ __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/
76	+ __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */
77	+ __u16 s_reserved_gdt_blocks;/* Per group table for online growth */
78	+ /*
79	+ * Journaling support valid if EXT2_FEATURE_COMPAT_HAS_JOURNAL set.
80	+ */
81	+ __u8 s_journal_uuid[16]; /* uuid of journal superblock */
82	+ __u32 s_journal_inum; /* inode number of journal file */
83	+ __u32 s_journal_dev; /* device number of journal file */
84	+ __u32 s_last_orphan; /* start of list of inodes to delete */
85	+ __u32 s_hash_seed[4]; /* HTREE hash seed */
86	+ __u8 s_def_hash_version; /* Default hash version to use */
87	+ __u8 s_jnl_backup_type; /* Default type of journal backup */
88	+ __u16 s_reserved_word_pad;
89	+ __u32 s_default_mount_opts;
90	+ __u32 s_first_meta_bg; /* First metablock group */
91	+ __u32 s_mkfs_time; /* When the filesystem was created */
92	+ __u32 s_jnl_blocks[17]; /* Backup of the journal inode */
93	+ /* 64bit desc support valid if EXT4_FEATURE_INCOMPAT_64BIT */
94	+ __u32 s_blocks_count_hi; /* Blocks count */
95	+ __u32 s_r_blocks_count_hi; /* Reserved blocks count */
96	+ __u32 s_free_blocks_count_hi; /* Free blocks count */
97	+ __u16 s_min_extra_isize; /* All inodes have at least # bytes */
98	+ __u16 s_max_extra_isize; /* New inodes should reverve # bytes */
99	+ __u32 s_flags; /* Miscellaneous flags */
100	+ __u16 s_raid_stride; /* Raid stride */
101	+ __u16 s_mmp_interval; /* # seconds to wait MMP checking */
102	+ __u64 s_mmp_block; /* Block for multi-mount protection */
103	+ __u32 s_raid_stripe_width; /* Blocks on all data disks (Nstride)/
104	+ __u8 s_log_groups_per_flex;/* FLEX_BG group size*/
105	+ __u8 s_reserved_char_pad;
106	+ __u16 s_reserved_pad;
107	+ __u32 s_reserved[162]; /* Padding to the end of the block */
108	+};
109
110	-struct ext2_group_desc
111	+struct ext4_group_desc
112	{
113	__u32 bg_block_bitmap; /* Blocks bitmap block */
114	__u32 bg_inode_bitmap; /* Inodes bitmap block */
115	@@ -90,8 +151,18 @@
116	__u16 bg_free_blocks_count; /* Free blocks count */
117	__u16 bg_free_inodes_count; /* Free inodes count */
118	__u16 bg_used_dirs_count; /* Directories count */
119	- __u16 bg_pad;
120	- __u32 bg_reserved[3];
121	+ __u16 bg_flags; /* EXT4_BG_flags (INODE_UNINIT, etc) */
122	+ __u32 bg_reserved[2]; /* Likely block/inode bitmap checksum */
123	+ __u16 bg_itable_unused; /* Unused inodes count */
124	+ __u16 bg_checksum; /* crc16(sb_uuid+group+desc) */
125	+ __u32 bg_block_bitmap_hi; /* Blocks bitmap block MSB */
126	+ __u32 bg_inode_bitmap_hi; /* Inodes bitmap block MSB */
127	+ __u32 bg_inode_table_hi; /* Inodes table block MSB */
128	+ __u16 bg_free_blocks_count_hi;/* Free blocks count MSB */
129	+ __u16 bg_free_inodes_count_hi;/* Free inodes count MSB */
130	+ __u16 bg_used_dirs_count_hi; /* Directories count MSB */
131	+ __u16 bg_itable_unused_hi; /* Unused inodes count MSB */
132	+ __u32 bg_reserved2[3];
133	};
134
135	struct ext2_inode
136	@@ -129,22 +200,22 @@
137	__u32 i_block[EXT2_N_BLOCKS]; /* 40: Pointers to blocks */
138	__u32 i_version; /* File version (for NFS) */
139	__u32 i_file_acl; /* File ACL */
140	- __u32 i_dir_acl; /* Directory ACL */
141	- __u32 i_faddr; /* Fragment address */
142	+ __u32 i_size_high;
143	+ __u32 i_obso_faddr; /* Obsoleted fragment address */
144	union
145	{
146	struct
147	{
148	- __u8 l_i_frag; /* Fragment number */
149	- __u8 l_i_fsize; /* Fragment size */
150	- __u16 i_pad1;
151	- __u32 l_i_reserved2[2];
152	+ __u16 l_i_blocks_high; /* were l_i_reserved1 */
153	+ __u16 l_i_file_acl_high;
154	+ __u16 l_i_uid_high; /* these 2 fields */
155	+ __u16 l_i_gid_high; /* were reserved2[0] */
156	+ __u32 l_i_reserved2;
157	}
158	linux2;
159	struct
160	{
161	- __u8 h_i_frag; /* Fragment number */
162	- __u8 h_i_fsize; /* Fragment size */
163	+ __u16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
164	__u16 h_i_mode_high;
165	__u16 h_i_uid_high;
166	__u16 h_i_gid_high;
167	@@ -153,16 +224,36 @@
168	hurd2;
169	struct
170	{
171	- __u8 m_i_frag; /* Fragment number */
172	- __u8 m_i_fsize; /* Fragment size */
173	- __u16 m_pad1;
174	+ __u16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
175	+ __u16 m_i_file_acl_high;
176	__u32 m_i_reserved2[2];
177	}
178	masix2;
179	}
180	osd2; /* OS dependent 2 */
181	+ __u16 i_extra_isize;
182	+ __u16 i_pad1;
183	+ __u32 i_ctime_extra; /* extra Change time (nsec << 2 \| epoch) */
184	+ __u32 i_mtime_extra; /* extra Modification time(nsec << 2 \| epoch) */
185	+ __u32 i_atime_extra; /* extra Access time (nsec << 2 \| epoch) */
186	+ __u32 i_crtime; /* File Creation time */
187	+ __u32 i_crtime_extra; /* extra FileCreationtime (nsec << 2 \| epoch) */
188	+ __u32 i_version_hi; /* high 32 bits for 64-bit version */
189	};
190
191	+#define EXT4_FEATURE_INCOMPAT_EXTENTS 0x0040 /* extents support */
192	+#define EXT4_FEATURE_INCOMPAT_64BIT 0x0080 /* grub not supported*/
193	+#define EXT4_FEATURE_INCOMPAT_MMP 0x0100
194	+#define EXT4_FEATURE_INCOMPAT_FLEX_BG 0x0200
195	+
196	+#define EXT4_HAS_INCOMPAT_FEATURE(sb,mask) \
197	+ ( sb->s_feature_incompat & mask )
198	+
199	+#define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */
200	+#define EXT4_HUGE_FILE_FL 0x00040000 /* Set to each huge file */
201	+
202	+#define EXT4_MIN_DESC_SIZE 32
203	+
204	/* linux/limits.h */
205	#define NAME_MAX 255 /* # chars in a file name */
206
207	@@ -180,6 +271,56 @@
208	char name[EXT2_NAME_LEN]; /* File name */
209	};
210
211	+/* linux/ext4_fs_extents.h */
212	+/* This is the extent on-disk structure.
213	+ * It's used at the bottom of the tree.
214	+ */
215	+struct ext4_extent
216	+ {
217	+ __u32 ee_block; /* first logical block extent covers */
218	+ __u16 ee_len; /* number of blocks covered by extent */
219	+ __u16 ee_start_hi; /* high 16 bits of physical block */
220	+ __u32 ee_start_lo; /* low 32 bits of physical block */
221	+ };
222	+
223	+/*
224	+ * This is index on-disk structure.
225	+ * It's used at all the levels except the bottom.
226	+ */
227	+struct ext4_extent_idx
228	+ {
229	+ __u32 ei_block; /* index covers logical blocks from 'block' */
230	+ __u32 ei_leaf_lo; /* pointer to the physical block of the next *
231	+ * level. leaf or next index could be there */
232	+ __u16 ei_leaf_hi; /* high 16 bits of physical block */
233	+ __u16 ei_unused;
234	+ };
235	+
236	+/*
237	+ * Each block (leaves and indexes), even inode-stored has header.
238	+ */
239	+struct ext4_extent_header
240	+ {
241	+ __u16 eh_magic; /* probably will support different formats */
242	+ __u16 eh_entries; /* number of valid entries */
243	+ __u16 eh_max; /* capacity of store in entries */
244	+ __u16 eh_depth; /* has tree real underlying blocks? */
245	+ __u32 eh_generation; /* generation of the tree */
246	+ };
247	+
248	+#define EXT4_EXT_MAGIC (0xf30a)
249	+#define EXT_FIRST_EXTENT(__hdr__) \
250	+ ((struct ext4_extent ) (((char ) (__hdr__)) + \
251	+ sizeof(struct ext4_extent_header)))
252	+#define EXT_FIRST_INDEX(__hdr__) \
253	+ ((struct ext4_extent_idx ) (((char ) (__hdr__)) + \
254	+ sizeof(struct ext4_extent_header)))
255	+#define EXT_LAST_EXTENT(__hdr__) \
256	+ (EXT_FIRST_EXTENT((__hdr__)) + (__u16)((__hdr__)->eh_entries) - 1)
257	+#define EXT_LAST_INDEX(__hdr__) \
258	+ (EXT_FIRST_INDEX((__hdr__)) + (__u16)((__hdr__)->eh_entries) - 1)
259	+
260	+
261	/* linux/ext2fs.h */
262	/*
263	* EXT2_DIR_PAD defines the directory entries boundaries
264	@@ -223,8 +364,17 @@
265	/* kind of from ext2/super.c */
266	#define EXT2_BLOCK_SIZE(s) (1 << EXT2_BLOCK_SIZE_BITS(s))
267	/* linux/ext2fs.h */
268	+/* sizeof(struct ext2_group_desc) is changed in ext4
269	+ * in kernel code, ext2/3 uses sizeof(struct ext2_group_desc) to calculate
270	+ * number of desc per block, while ext4 uses superblock->s_desc_size in stead
271	+ * superblock->s_desc_size is not available in ext2/3
272	+ * */
273	+#define EXT2_DESC_SIZE(s) \
274	+ (EXT4_HAS_INCOMPAT_FEATURE(s,EXT4_FEATURE_INCOMPAT_64BIT)? \
275	+ s->s_desc_size : EXT4_MIN_DESC_SIZE)
276	#define EXT2_DESC_PER_BLOCK(s) \
277	- (EXT2_BLOCK_SIZE(s) / sizeof (struct ext2_group_desc))
278	+ (EXT2_BLOCK_SIZE(s) / EXT2_DESC_SIZE(s))
279	+
280	/* linux/stat.h */
281	#define S_IFMT 00170000
282	#define S_IFLNK 0120000
283	@@ -386,6 +536,122 @@
284	[logical_block & (EXT2_ADDR_PER_BLOCK (SUPERBLOCK) - 1)];
285	}
286
287	+/* extent binary search index
288	+ * find closest index in the current level extent tree
289	+ * kind of from ext4_ext_binsearch_idx in ext4/extents.c
290	+ */
291	+static struct ext4_extent_idx*
292	+ext4_ext_binsearch_idx(struct ext4_extent_header* eh, int logical_block)
293	+{
294	+ struct ext4_extent_idx r, l, *m;
295	+ l = EXT_FIRST_INDEX(eh) + 1;
296	+ r = EXT_LAST_INDEX(eh);
297	+ while (l <= r)
298	+ {
299	+ m = l + (r - l) / 2;
300	+ if (logical_block < m->ei_block)
301	+ r = m - 1;
302	+ else
303	+ l = m + 1;
304	+ }
305	+ return (struct ext4_extent_idx*)(l - 1);
306	+}
307	+
308	+/* extent binary search
309	+ * find closest extent in the leaf level
310	+ * kind of from ext4_ext_binsearch in ext4/extents.c
311	+ */
312	+static struct ext4_extent*
313	+ext4_ext_binsearch(struct ext4_extent_header* eh, int logical_block)
314	+{
315	+ struct ext4_extent r, l, *m;
316	+ l = EXT_FIRST_EXTENT(eh) + 1;
317	+ r = EXT_LAST_EXTENT(eh);
318	+ while (l <= r)
319	+ {
320	+ m = l + (r - l) / 2;
321	+ if (logical_block < m->ee_block)
322	+ r = m - 1;
323	+ else
324	+ l = m + 1;
325	+ }
326	+ return (struct ext4_extent*)(l - 1);
327	+}
328	+
329	+/* Maps extents enabled logical block into physical block via an inode.
330	+ * EXT4_HUGE_FILE_FL should be checked before calling this.
331	+ */
332	+static int
333	+ext4fs_block_map (int logical_block)
334	+{
335	+ struct ext4_extent_header *eh;
336	+ struct ext4_extent ex, extent;
337	+ struct ext4_extent_idx ei, index;
338	+ int depth;
339	+ int i;
340	+
341	+#ifdef E2DEBUG
342	+ unsigned char *i;
343	+ for (i = (unsigned char *) INODE;
344	+ i < ((unsigned char *) INODE + sizeof (struct ext2_inode));
345	+ i++)
346	+ {
347	+ printf ("%c", "0123456789abcdef"[*i >> 4]);
348	+ printf ("%c", "0123456789abcdef"[*i % 16]);
349	+ if (!((i + 1 - (unsigned char *) INODE) % 16))
350	+ {
351	+ printf ("\n");
352	+ }
353	+ else
354	+ {
355	+ printf (" ");
356	+ }
357	+ }
358	+ printf ("logical block %d\n", logical_block);
359	+#endif /* E2DEBUG */
360	+ eh = (struct ext4_extent_header*)INODE->i_block;
361	+ if (eh->eh_magic != EXT4_EXT_MAGIC)
362	+ {
363	+ errnum = ERR_FSYS_CORRUPT;
364	+ return -1;
365	+ }
366	+ while((depth = eh->eh_depth) != 0)
367	+ { /* extent index */
368	+ if (eh->eh_magic != EXT4_EXT_MAGIC)
369	+ {
370	+ errnum = ERR_FSYS_CORRUPT;
371	+ return -1;
372	+ }
373	+ ei = ext4_ext_binsearch_idx(eh, logical_block);
374	+ if (ei->ei_leaf_hi)
375	+ {/* 64bit physical block number not supported */
376	+ errnum = ERR_FILELENGTH;
377	+ return -1;
378	+ }
379	+ if (!ext2_rdfsb(ei->ei_leaf_lo, DATABLOCK1))
380	+ {
381	+ errnum = ERR_FSYS_CORRUPT;
382	+ return -1;
383	+ }
384	+ eh = (struct ext4_extent_header*)DATABLOCK1;
385	+ }
386	+
387	+ /* depth==0, we come to the leaf */
388	+ ex = ext4_ext_binsearch(eh, logical_block);
389	+ if (ex->ee_start_hi)
390	+ {/* 64bit physical block number not supported */
391	+ errnum = ERR_FILELENGTH;
392	+ return -1;
393	+ }
394	+ if ((ex->ee_block + ex->ee_len) < logical_block)
395	+ {
396	+ errnum = ERR_FSYS_CORRUPT;
397	+ return -1;
398	+ }
399	+ return ex->ee_start_lo + logical_block - ex->ee_block;
400	+
401	+}
402	+
403	/* preconditions: all preconds of ext2fs_block_map */
404	int
405	ext2fs_read (char *buf, int len)
406	@@ -420,6 +686,11 @@
407	/* find the (logical) block component of our location */
408	logical_block = filepos >> EXT2_BLOCK_SIZE_BITS (SUPERBLOCK);
409	offset = filepos & (EXT2_BLOCK_SIZE (SUPERBLOCK) - 1);
410	+ /* map extents enabled logical block number to physical fs on-disk block number */
411	+ if (EXT4_HAS_INCOMPAT_FEATURE(SUPERBLOCK,EXT4_FEATURE_INCOMPAT_EXTENTS)
412	+ && INODE->i_flags & EXT4_EXTENTS_FL)
413	+ map = ext4fs_block_map (logical_block);
414	+ else
415	map = ext2fs_block_map (logical_block);
416	#ifdef E2DEBUG
417	printf ("map=%d\n", map);
418	@@ -504,7 +775,7 @@
419	int desc; /* index within that group */
420	int ino_blk; /* fs pointer of the inode's information */
421	int str_chk = 0; /* used to hold the results of a string compare */
422	- struct ext2_group_desc *gdp;
423	+ struct ext4_group_desc *ext4_gdp;
424	struct ext2_inode raw_inode; / inode info corresponding to current_ino */
425
426	char linkbuf[PATH_MAX]; /* buffer for following symbolic links */
427	@@ -550,8 +821,15 @@
428	{
429	return 0;
430	}
431	- gdp = GROUP_DESC;
432	- ino_blk = gdp[desc].bg_inode_table +
433	+ ext4_gdp = (struct ext4_group_desc )( (__u8)GROUP_DESC +
434	+ desc * EXT2_DESC_SIZE(SUPERBLOCK));
435	+ if (EXT4_HAS_INCOMPAT_FEATURE(SUPERBLOCK, EXT4_FEATURE_INCOMPAT_64BIT)
436	+ && (! ext4_gdp->bg_inode_table_hi))
437	+ {/* 64bit itable not supported */
438	+ errnum = ERR_FILELENGTH;
439	+ return -1;
440	+ }
441	+ ino_blk = ext4_gdp->bg_inode_table +
442	(((current_ino - 1) % (SUPERBLOCK->s_inodes_per_group))
443	>> log2 (EXT2_BLOCK_SIZE (SUPERBLOCK) / sizeof (struct ext2_inode)));
444	#ifdef E2DEBUG
445	@@ -629,7 +907,10 @@
446	}
447	linkbuf[filemax + len] = '\0';
448
449	- /* Read the symlink data. */
450	+ /* Read the symlink data.
451	+ * Slow symlink is extents enabled
452	+ * But since grub_read invokes ext2fs_read, nothing to change here
453	+ */
454	if (! ext2_is_fast_symlink ())
455	{
456	/* Read the necessary blocks, and reset the file pointer. */
457	@@ -640,7 +921,9 @@
458	}
459	else
460	{
461	- /* Copy the data directly from the inode. */
462	+ /* Copy the data directly from the inode.
463	+ * Fast symlink is not extents enabled
464	+ */
465	len = filemax;
466	memmove (linkbuf, (char *) INODE->i_block, len);
467	}
468	@@ -674,7 +957,13 @@
469	errnum = ERR_BAD_FILETYPE;
470	return 0;
471	}
472	-
473	+ /* if file is too large, just stop and report an error*/
474	+ if ( (INODE->i_flags & EXT4_HUGE_FILE_FL) && !(INODE->i_size_high))
475	+ {
476	+ /* file too large, stop reading */
477	+ errnum = ERR_FILELENGTH;
478	+ return 0;
479	+ }
480	filemax = (INODE->i_size);
481	return 1;
482	}
483	@@ -728,17 +1017,28 @@
484	}
485
486	/* else, find the (logical) block component of our location */
487	+ /* ext4 logical block number the same as ext2/3 */
488	blk = loc >> EXT2_BLOCK_SIZE_BITS (SUPERBLOCK);
489
490	/* we know which logical block of the directory entry we are looking
491	for, now we have to translate that to the physical (fs) block on
492	the disk */
493	+ /* map extents enabled logical block number to physical fs on-disk block number */
494	+ if (EXT4_HAS_INCOMPAT_FEATURE(SUPERBLOCK,EXT4_FEATURE_INCOMPAT_EXTENTS)
495	+ && INODE->i_flags & EXT4_EXTENTS_FL)
496	+ map = ext4fs_block_map (blk);
497	+ else
498	map = ext2fs_block_map (blk);
499	#ifdef E2DEBUG
500	printf ("fs block=%d\n", map);
501	#endif /* E2DEBUG */
502	mapblock2 = -1;
503	- if ((map < 0) \|\| !ext2_rdfsb (map, DATABLOCK2))
504	+ if (map < 0)
505	+ {
506	+ *rest = ch;
507	+ return 0;
508	+ }
509	+ if (!ext2_rdfsb (map, DATABLOCK2))
510	{
511	errnum = ERR_FSYS_CORRUPT;
512	*rest = ch;