* Re: ext4 support on pvgrub
2010-10-12 20:13 ` Jeremy Fitzhardinge
@ 2010-10-12 20:57 ` M A Young
2010-10-14 13:57 ` Stefano Stabellini
0 siblings, 1 reply; 9+ messages in thread
From: M A Young @ 2010-10-12 20:57 UTC (permalink / raw)
To: Jeremy Fitzhardinge; +Cc: xen-devel, Keir Fraser, Stefano Stabellini
[-- Attachment #1: Type: TEXT/PLAIN, Size: 1120 bytes --]
On Tue, 12 Oct 2010, Jeremy Fitzhardinge wrote:
>
> On 10/07/2010 02:48 PM, M A Young wrote:
>> I have this working on Fedora. It turned out to be very easy to add
>> this as I simply took the grub-ext4-support.patch from the Fedora grub
>> package and put it in xen-4.0.1/stubdom/grub.patches/ . Using pvgrub
>> (without a vif on the virtual machine as I was having problems if I
>> did give it a network interface) I could read and boot from ext4
>> partition with this patch but not without it.
>>
>> Is this a good way to add this feature to xen more generally?
>
> Sounds like it just needs to be committed. Post the patch?
I am attaching both the raw patch (taken from grub-0.97-66.fc14, it
traces back to the attachment at
https://bugzilla.redhat.com/show_bug.cgi?id=486284#c7 with only spelling
fixes) and also one that could be applied to the xen source.
Use whichever is more convenient or you can fetch it directly from the
Fedora source repository at
http://pkgs.fedoraproject.org/gitweb/?p=grub.git;a=blob_plain;f=grub-ext4-support.patch;hb=3bcdb10fc21d8e94efa70fd91d21224f13f01433
Michael Young
[-- Attachment #2: patch from grub-0.97-66.fc14 --]
[-- Type: TEXT/PLAIN, Size: 16944 bytes --]
Index: grub-0.97/stage2/fsys_ext2fs.c
===================================================================
--- grub-0.97.orig/stage2/fsys_ext2fs.c
+++ grub-0.97/stage2/fsys_ext2fs.c
@@ -41,6 +41,7 @@ typedef __signed__ short __s16;
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 @@ typedef unsigned int __u32;
#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 @@ struct ext2_super_block
__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 */
@@ -72,7 +73,7 @@ struct ext2_super_block
__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 */
__u32 s_lastcheck; /* time of last check */
__u32 s_checkinterval; /* max. time between checks */
__u32 s_creator_os; /* OS */
@@ -119,15 +120,29 @@ struct ext2_super_block
__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;
+ __u16 s_desc_size; /* size of group descriptor */
__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 */
- __u32 s_reserved[172]; /* Padding to the end of the block */
- };
+ /* 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 */
@@ -135,8 +150,18 @@ struct ext2_group_desc
__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
@@ -174,22 +199,22 @@ struct ext2_inode
__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;
@@ -198,16 +223,36 @@ struct ext2_inode
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 */
@@ -225,6 +270,57 @@ struct ext2_dir_entry
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
@@ -271,8 +367,17 @@ struct ext2_dir_entry
/* 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
@@ -434,6 +539,122 @@ ext2fs_block_map (int logical_block)
[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)
@@ -468,6 +689,11 @@ ext2fs_read (char *buf, int len)
/* 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);
@@ -552,7 +778,7 @@ ext2fs_dir (char *dirname)
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 */
@@ -598,8 +824,15 @@ ext2fs_dir (char *dirname)
{
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_INODES_PER_BLOCK (SUPERBLOCK)));
#ifdef E2DEBUG
@@ -676,7 +909,10 @@ ext2fs_dir (char *dirname)
}
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. */
@@ -687,7 +923,9 @@ ext2fs_dir (char *dirname)
}
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);
}
@@ -721,6 +959,13 @@ ext2fs_dir (char *dirname)
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;
@@ -775,17 +1020,28 @@ ext2fs_dir (char *dirname)
}
/* 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;
[-- Attachment #3: patch to xen-4.0.1 --]
[-- Type: TEXT/PLAIN, Size: 17586 bytes --]
--- /dev/null 2010-10-12 22:17:49.295779369 +0100
+++ xen-4.0.1/stubdom/grub.patches/99grub-ext4-support.patch 2010-10-12 21:36:24.000000000 +0100
@@ -0,0 +1,474 @@
+Index: grub-0.97/stage2/fsys_ext2fs.c
+===================================================================
+--- grub-0.97.orig/stage2/fsys_ext2fs.c
++++ grub-0.97/stage2/fsys_ext2fs.c
+@@ -41,6 +41,7 @@ typedef __signed__ short __s16;
+ 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 @@ typedef unsigned int __u32;
+ #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 @@ struct ext2_super_block
+ __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 */
+@@ -72,7 +73,7 @@ struct ext2_super_block
+ __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 */
+ __u32 s_lastcheck; /* time of last check */
+ __u32 s_checkinterval; /* max. time between checks */
+ __u32 s_creator_os; /* OS */
+@@ -119,15 +120,29 @@ struct ext2_super_block
+ __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;
++ __u16 s_desc_size; /* size of group descriptor */
+ __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 */
+- __u32 s_reserved[172]; /* Padding to the end of the block */
+- };
++ /* 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 */
+@@ -135,8 +150,18 @@ struct ext2_group_desc
+ __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
+@@ -174,22 +199,22 @@ struct ext2_inode
+ __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;
+@@ -198,16 +223,36 @@ struct ext2_inode
+ 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 */
+
+@@ -225,6 +270,57 @@ struct ext2_dir_entry
+ 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
+@@ -271,8 +367,17 @@ struct ext2_dir_entry
+ /* 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
+@@ -434,6 +539,122 @@ ext2fs_block_map (int logical_block)
+ [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)
+@@ -468,6 +689,11 @@ ext2fs_read (char *buf, int len)
+ /* 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);
+@@ -552,7 +778,7 @@ ext2fs_dir (char *dirname)
+ 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 */
+@@ -598,8 +824,15 @@ ext2fs_dir (char *dirname)
+ {
+ 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_INODES_PER_BLOCK (SUPERBLOCK)));
+ #ifdef E2DEBUG
+@@ -676,7 +909,10 @@ ext2fs_dir (char *dirname)
+ }
+ 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. */
+@@ -687,7 +923,9 @@ ext2fs_dir (char *dirname)
+ }
+ 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);
+ }
+@@ -721,6 +959,13 @@ ext2fs_dir (char *dirname)
+ 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;
+@@ -775,17 +1020,28 @@ ext2fs_dir (char *dirname)
+ }
+
+ /* 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;
[-- Attachment #4: Type: text/plain, Size: 138 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
^ permalink raw reply [flat|nested] 9+ messages in thread