From 4227126f8558304b95ce1f9a1b6b6a267cd97814 Mon Sep 17 00:00:00 2001 From: alex Date: Thu, 25 Sep 2003 21:21:48 +0000 Subject: [PATCH] - rh-2.4.22 series for latest RH kernel - some patches were ported NOTE: it builds, it runs lustre and simple tests (cp, ls, stat) work --- .../patches/ext3-htree-2.4.22-rh.patch | 2569 ++++++++++ .../patches/ext3-orphan_lock-2.4.22-rh.patch | 158 + .../patches/extN-wantedi-2.4.22-rh.patch | 217 + .../patches/iod-stock-exports-2.4.22-rh.patch | 48 + .../patches/socket-exports-2.4.22-rh.patch | 41 + .../patches/tcp-zero-copy-2.4.22-rh.patch | 459 ++ .../patches/vfs_intent-2.4.22-rh.patch | 1809 +++++++ .../patches/xattr-0.8.54-2.4.22-rh.patch | 5403 ++++++++++++++++++++ lustre/kernel_patches/pc/ext3-htree-2.4.22-rh.pc | 11 + .../pc/ext3-orphan_lock-2.4.22-rh.pc | 5 + lustre/kernel_patches/pc/extN-wantedi-2.4.22-rh.pc | 6 + .../pc/iod-stock-exports-2.4.22-rh.pc | 3 + .../kernel_patches/pc/socket-exports-2.4.22-rh.pc | 3 + .../kernel_patches/pc/tcp-zero-copy-2.4.22-rh.pc | 5 + lustre/kernel_patches/pc/vfs_intent-2.4.22-rh.pc | 12 + lustre/kernel_patches/pc/xattr-0.8.54-2.4.22-rh.pc | 58 + lustre/kernel_patches/series/rh-2.4.22 | 23 + 17 files changed, 10830 insertions(+) create mode 100644 lustre/kernel_patches/patches/ext3-htree-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/ext3-orphan_lock-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/extN-wantedi-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/iod-stock-exports-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/socket-exports-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/tcp-zero-copy-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/vfs_intent-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/patches/xattr-0.8.54-2.4.22-rh.patch create mode 100644 lustre/kernel_patches/pc/ext3-htree-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/ext3-orphan_lock-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/extN-wantedi-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/iod-stock-exports-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/socket-exports-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/tcp-zero-copy-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/vfs_intent-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/pc/xattr-0.8.54-2.4.22-rh.pc create mode 100644 lustre/kernel_patches/series/rh-2.4.22 diff --git a/lustre/kernel_patches/patches/ext3-htree-2.4.22-rh.patch b/lustre/kernel_patches/patches/ext3-htree-2.4.22-rh.patch new file mode 100644 index 0000000..6fd0735 --- /dev/null +++ b/lustre/kernel_patches/patches/ext3-htree-2.4.22-rh.patch @@ -0,0 +1,2569 @@ + fs/ext3/Makefile | 2 + fs/ext3/dir.c | 302 +++++++++ + fs/ext3/file.c | 3 + fs/ext3/hash.c | 215 ++++++ + fs/ext3/namei.c | 1421 ++++++++++++++++++++++++++++++++++++++++----- + fs/ext3/super.c | 7 + include/linux/ext3_fs.h | 85 ++ + include/linux/ext3_fs_sb.h | 2 + include/linux/ext3_jbd.h | 2 + include/linux/rbtree.h | 2 + lib/rbtree.c | 42 + + 11 files changed, 1922 insertions(+), 161 deletions(-) + +--- linux-2.4.22-ac1/fs/ext3/dir.c~ext3-htree-2.4.22-rh 2001-11-10 01:25:04.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext3/dir.c 2003-09-25 14:58:30.000000000 +0400 +@@ -21,12 +21,16 @@ + #include + #include + #include ++#include ++#include + + static unsigned char ext3_filetype_table[] = { + DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK + }; + + static int ext3_readdir(struct file *, void *, filldir_t); ++static int ext3_dx_readdir(struct file * filp, ++ void * dirent, filldir_t filldir); + + struct file_operations ext3_dir_operations = { + read: generic_read_dir, +@@ -35,6 +39,17 @@ struct file_operations ext3_dir_operatio + fsync: ext3_sync_file, /* BKL held */ + }; + ++ ++static unsigned char get_dtype(struct super_block *sb, int filetype) ++{ ++ if (!EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_FILETYPE) || ++ (filetype >= EXT3_FT_MAX)) ++ return DT_UNKNOWN; ++ ++ return (ext3_filetype_table[filetype]); ++} ++ ++ + int ext3_check_dir_entry (const char * function, struct inode * dir, + struct ext3_dir_entry_2 * de, + struct buffer_head * bh, +@@ -79,6 +94,16 @@ static int ext3_readdir(struct file * fi + + sb = inode->i_sb; + ++ if (is_dx(inode)) { ++ err = ext3_dx_readdir(filp, dirent, filldir); ++ if (err != ERR_BAD_DX_DIR) ++ return err; ++ /* ++ * We don't set the inode dirty flag since it's not ++ * critical that it get flushed back to the disk. ++ */ ++ EXT3_I(filp->f_dentry->d_inode)->i_flags &= ~EXT3_INDEX_FL; ++ } + stored = 0; + bh = NULL; + offset = filp->f_pos & (sb->s_blocksize - 1); +@@ -162,18 +187,12 @@ revalidate: + * during the copy operation. + */ + unsigned long version = filp->f_version; +- unsigned char d_type = DT_UNKNOWN; + +- if (EXT3_HAS_INCOMPAT_FEATURE(sb, +- EXT3_FEATURE_INCOMPAT_FILETYPE) +- && de->file_type < EXT3_FT_MAX) +- d_type = +- ext3_filetype_table[de->file_type]; + error = filldir(dirent, de->name, + de->name_len, + filp->f_pos, + le32_to_cpu(de->inode), +- d_type); ++ get_dtype(sb, de->file_type)); + if (error) + break; + if (version != filp->f_version) +@@ -188,3 +207,272 @@ revalidate: + UPDATE_ATIME(inode); + return 0; + } ++ ++#ifdef CONFIG_EXT3_INDEX ++/* ++ * These functions convert from the major/minor hash to an f_pos ++ * value. ++ * ++ * Currently we only use major hash numer. This is unfortunate, but ++ * on 32-bit machines, the same VFS interface is used for lseek and ++ * llseek, so if we use the 64 bit offset, then the 32-bit versions of ++ * lseek/telldir/seekdir will blow out spectacularly, and from within ++ * the ext2 low-level routine, we don't know if we're being called by ++ * a 64-bit version of the system call or the 32-bit version of the ++ * system call. Worse yet, NFSv2 only allows for a 32-bit readdir ++ * cookie. Sigh. ++ */ ++#define hash2pos(major, minor) (major >> 1) ++#define pos2maj_hash(pos) ((pos << 1) & 0xffffffff) ++#define pos2min_hash(pos) (0) ++ ++/* ++ * This structure holds the nodes of the red-black tree used to store ++ * the directory entry in hash order. ++ */ ++struct fname { ++ __u32 hash; ++ __u32 minor_hash; ++ rb_node_t rb_hash; ++ struct fname *next; ++ __u32 inode; ++ __u8 name_len; ++ __u8 file_type; ++ char name[0]; ++}; ++ ++/* ++ * This functoin implements a non-recursive way of freeing all of the ++ * nodes in the red-black tree. ++ */ ++static void free_rb_tree_fname(rb_root_t *root) ++{ ++ rb_node_t *n = root->rb_node; ++ rb_node_t *parent; ++ struct fname *fname; ++ ++ while (n) { ++ /* Do the node's children first */ ++ if ((n)->rb_left) { ++ n = n->rb_left; ++ continue; ++ } ++ if (n->rb_right) { ++ n = n->rb_right; ++ continue; ++ } ++ /* ++ * The node has no children; free it, and then zero ++ * out parent's link to it. Finally go to the ++ * beginning of the loop and try to free the parent ++ * node. ++ */ ++ parent = n->rb_parent; ++ fname = rb_entry(n, struct fname, rb_hash); ++ kfree(fname); ++ if (!parent) ++ root->rb_node = 0; ++ else if (parent->rb_left == n) ++ parent->rb_left = 0; ++ else if (parent->rb_right == n) ++ parent->rb_right = 0; ++ n = parent; ++ } ++ root->rb_node = 0; ++} ++ ++ ++struct dir_private_info *create_dir_info(loff_t pos) ++{ ++ struct dir_private_info *p; ++ ++ p = kmalloc(sizeof(struct dir_private_info), GFP_KERNEL); ++ if (!p) ++ return NULL; ++ p->root.rb_node = 0; ++ p->curr_node = 0; ++ p->extra_fname = 0; ++ p->last_pos = 0; ++ p->curr_hash = pos2maj_hash(pos); ++ p->curr_minor_hash = pos2min_hash(pos); ++ p->next_hash = 0; ++ return p; ++} ++ ++void ext3_htree_free_dir_info(struct dir_private_info *p) ++{ ++ free_rb_tree_fname(&p->root); ++ kfree(p); ++} ++ ++/* ++ * Given a directory entry, enter it into the fname rb tree. ++ */ ++int ext3_htree_store_dirent(struct file *dir_file, __u32 hash, ++ __u32 minor_hash, ++ struct ext3_dir_entry_2 *dirent) ++{ ++ rb_node_t **p, *parent = NULL; ++ struct fname * fname, *new_fn; ++ struct dir_private_info *info; ++ int len; ++ ++ info = (struct dir_private_info *) dir_file->private_data; ++ p = &info->root.rb_node; ++ ++ /* Create and allocate the fname structure */ ++ len = sizeof(struct fname) + dirent->name_len + 1; ++ new_fn = kmalloc(len, GFP_KERNEL); ++ if (!new_fn) ++ return -ENOMEM; ++ memset(new_fn, 0, len); ++ new_fn->hash = hash; ++ new_fn->minor_hash = minor_hash; ++ new_fn->inode = le32_to_cpu(dirent->inode); ++ new_fn->name_len = dirent->name_len; ++ new_fn->file_type = dirent->file_type; ++ memcpy(new_fn->name, dirent->name, dirent->name_len); ++ new_fn->name[dirent->name_len] = 0; ++ ++ while (*p) { ++ parent = *p; ++ fname = rb_entry(parent, struct fname, rb_hash); ++ ++ /* ++ * If the hash and minor hash match up, then we put ++ * them on a linked list. This rarely happens... ++ */ ++ if ((new_fn->hash == fname->hash) && ++ (new_fn->minor_hash == fname->minor_hash)) { ++ new_fn->next = fname->next; ++ fname->next = new_fn; ++ return 0; ++ } ++ ++ if (new_fn->hash < fname->hash) ++ p = &(*p)->rb_left; ++ else if (new_fn->hash > fname->hash) ++ p = &(*p)->rb_right; ++ else if (new_fn->minor_hash < fname->minor_hash) ++ p = &(*p)->rb_left; ++ else /* if (new_fn->minor_hash > fname->minor_hash) */ ++ p = &(*p)->rb_right; ++ } ++ ++ rb_link_node(&new_fn->rb_hash, parent, p); ++ rb_insert_color(&new_fn->rb_hash, &info->root); ++ return 0; ++} ++ ++ ++ ++/* ++ * This is a helper function for ext3_dx_readdir. It calls filldir ++ * for all entres on the fname linked list. (Normally there is only ++ * one entry on the linked list, unless there are 62 bit hash collisions.) ++ */ ++static int call_filldir(struct file * filp, void * dirent, ++ filldir_t filldir, struct fname *fname) ++{ ++ struct dir_private_info *info = filp->private_data; ++ loff_t curr_pos; ++ struct inode *inode = filp->f_dentry->d_inode; ++ struct super_block * sb; ++ int error; ++ ++ sb = inode->i_sb; ++ ++ if (!fname) { ++ printk("call_filldir: called with null fname?!?\n"); ++ return 0; ++ } ++ curr_pos = hash2pos(fname->hash, fname->minor_hash); ++ while (fname) { ++ error = filldir(dirent, fname->name, ++ fname->name_len, curr_pos, ++ fname->inode, ++ get_dtype(sb, fname->file_type)); ++ if (error) { ++ filp->f_pos = curr_pos; ++ info->extra_fname = fname->next; ++ return error; ++ } ++ fname = fname->next; ++ } ++ return 0; ++} ++ ++static int ext3_dx_readdir(struct file * filp, ++ void * dirent, filldir_t filldir) ++{ ++ struct dir_private_info *info = filp->private_data; ++ struct inode *inode = filp->f_dentry->d_inode; ++ struct fname *fname; ++ int ret; ++ ++ if (!info) { ++ info = create_dir_info(filp->f_pos); ++ if (!info) ++ return -ENOMEM; ++ filp->private_data = info; ++ } ++ ++ /* Some one has messed with f_pos; reset the world */ ++ if (info->last_pos != filp->f_pos) { ++ free_rb_tree_fname(&info->root); ++ info->curr_node = 0; ++ info->extra_fname = 0; ++ info->curr_hash = pos2maj_hash(filp->f_pos); ++ info->curr_minor_hash = pos2min_hash(filp->f_pos); ++ } ++ ++ /* ++ * If there are any leftover names on the hash collision ++ * chain, return them first. ++ */ ++ if (info->extra_fname && ++ call_filldir(filp, dirent, filldir, info->extra_fname)) ++ goto finished; ++ ++ if (!info->curr_node) ++ info->curr_node = rb_get_first(&info->root); ++ ++ while (1) { ++ /* ++ * Fill the rbtree if we have no more entries, ++ * or the inode has changed since we last read in the ++ * cached entries. ++ */ ++ if ((!info->curr_node) || ++ (filp->f_version != inode->i_version)) { ++ info->curr_node = 0; ++ free_rb_tree_fname(&info->root); ++ filp->f_version = inode->i_version; ++ ret = ext3_htree_fill_tree(filp, info->curr_hash, ++ info->curr_minor_hash, ++ &info->next_hash); ++ if (ret < 0) ++ return ret; ++ if (ret == 0) ++ break; ++ info->curr_node = rb_get_first(&info->root); ++ } ++ ++ fname = rb_entry(info->curr_node, struct fname, rb_hash); ++ info->curr_hash = fname->hash; ++ info->curr_minor_hash = fname->minor_hash; ++ if (call_filldir(filp, dirent, filldir, fname)) ++ break; ++ ++ info->curr_node = rb_get_next(info->curr_node); ++ if (!info->curr_node) { ++ info->curr_hash = info->next_hash; ++ info->curr_minor_hash = 0; ++ } ++ } ++finished: ++ info->last_pos = filp->f_pos; ++ UPDATE_ATIME(inode); ++ return 0; ++} ++#endif +--- linux-2.4.22-ac1/fs/ext3/file.c~ext3-htree-2.4.22-rh 2003-08-25 15:44:43.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/file.c 2003-09-25 14:55:12.000000000 +0400 +@@ -35,6 +35,9 @@ static int ext3_release_file (struct ino + { + if (filp->f_mode & FMODE_WRITE) + ext3_discard_prealloc (inode); ++ if (is_dx(inode) && filp->private_data) ++ ext3_htree_free_dir_info(filp->private_data); ++ + return 0; + } + +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext3/hash.c 2003-09-25 14:55:12.000000000 +0400 +@@ -0,0 +1,215 @@ ++/* ++ * linux/fs/ext3/hash.c ++ * ++ * Copyright (C) 2002 by Theodore Ts'o ++ * ++ * This file is released under the GPL v2. ++ * ++ * This file may be redistributed under the terms of the GNU Public ++ * License. ++ */ ++ ++#include ++#include ++#include ++#include ++ ++#define DELTA 0x9E3779B9 ++ ++static void TEA_transform(__u32 buf[4], __u32 const in[]) ++{ ++ __u32 sum = 0; ++ __u32 b0 = buf[0], b1 = buf[1]; ++ __u32 a = in[0], b = in[1], c = in[2], d = in[3]; ++ int n = 16; ++ ++ do { ++ sum += DELTA; ++ b0 += ((b1 << 4)+a) ^ (b1+sum) ^ ((b1 >> 5)+b); ++ b1 += ((b0 << 4)+c) ^ (b0+sum) ^ ((b0 >> 5)+d); ++ } while(--n); ++ ++ buf[0] += b0; ++ buf[1] += b1; ++} ++ ++/* F, G and H are basic MD4 functions: selection, majority, parity */ ++#define F(x, y, z) ((z) ^ ((x) & ((y) ^ (z)))) ++#define G(x, y, z) (((x) & (y)) + (((x) ^ (y)) & (z))) ++#define H(x, y, z) ((x) ^ (y) ^ (z)) ++ ++/* ++ * The generic round function. The application is so specific that ++ * we don't bother protecting all the arguments with parens, as is generally ++ * good macro practice, in favor of extra legibility. ++ * Rotation is separate from addition to prevent recomputation ++ */ ++#define ROUND(f, a, b, c, d, x, s) \ ++ (a += f(b, c, d) + x, a = (a << s) | (a >> (32-s))) ++#define K1 0 ++#define K2 013240474631UL ++#define K3 015666365641UL ++ ++/* ++ * Basic cut-down MD4 transform. Returns only 32 bits of result. ++ */ ++static void halfMD4Transform (__u32 buf[4], __u32 const in[]) ++{ ++ __u32 a = buf[0], b = buf[1], c = buf[2], d = buf[3]; ++ ++ /* Round 1 */ ++ ROUND(F, a, b, c, d, in[0] + K1, 3); ++ ROUND(F, d, a, b, c, in[1] + K1, 7); ++ ROUND(F, c, d, a, b, in[2] + K1, 11); ++ ROUND(F, b, c, d, a, in[3] + K1, 19); ++ ROUND(F, a, b, c, d, in[4] + K1, 3); ++ ROUND(F, d, a, b, c, in[5] + K1, 7); ++ ROUND(F, c, d, a, b, in[6] + K1, 11); ++ ROUND(F, b, c, d, a, in[7] + K1, 19); ++ ++ /* Round 2 */ ++ ROUND(G, a, b, c, d, in[1] + K2, 3); ++ ROUND(G, d, a, b, c, in[3] + K2, 5); ++ ROUND(G, c, d, a, b, in[5] + K2, 9); ++ ROUND(G, b, c, d, a, in[7] + K2, 13); ++ ROUND(G, a, b, c, d, in[0] + K2, 3); ++ ROUND(G, d, a, b, c, in[2] + K2, 5); ++ ROUND(G, c, d, a, b, in[4] + K2, 9); ++ ROUND(G, b, c, d, a, in[6] + K2, 13); ++ ++ /* Round 3 */ ++ ROUND(H, a, b, c, d, in[3] + K3, 3); ++ ROUND(H, d, a, b, c, in[7] + K3, 9); ++ ROUND(H, c, d, a, b, in[2] + K3, 11); ++ ROUND(H, b, c, d, a, in[6] + K3, 15); ++ ROUND(H, a, b, c, d, in[1] + K3, 3); ++ ROUND(H, d, a, b, c, in[5] + K3, 9); ++ ROUND(H, c, d, a, b, in[0] + K3, 11); ++ ROUND(H, b, c, d, a, in[4] + K3, 15); ++ ++ buf[0] += a; ++ buf[1] += b; ++ buf[2] += c; ++ buf[3] += d; ++} ++ ++#undef ROUND ++#undef F ++#undef G ++#undef H ++#undef K1 ++#undef K2 ++#undef K3 ++ ++/* The old legacy hash */ ++static __u32 dx_hack_hash (const char *name, int len) ++{ ++ __u32 hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9; ++ while (len--) { ++ __u32 hash = hash1 + (hash0 ^ (*name++ * 7152373)); ++ ++ if (hash & 0x80000000) hash -= 0x7fffffff; ++ hash1 = hash0; ++ hash0 = hash; ++ } ++ return (hash0 << 1); ++} ++ ++static void str2hashbuf(const char *msg, int len, __u32 *buf, int num) ++{ ++ __u32 pad, val; ++ int i; ++ ++ pad = (__u32)len | ((__u32)len << 8); ++ pad |= pad << 16; ++ ++ val = pad; ++ if (len > num*4) ++ len = num * 4; ++ for (i=0; i < len; i++) { ++ if ((i % 4) == 0) ++ val = pad; ++ val = msg[i] + (val << 8); ++ if ((i % 4) == 3) { ++ *buf++ = val; ++ val = pad; ++ num--; ++ } ++ } ++ if (--num >= 0) ++ *buf++ = val; ++ while (--num >= 0) ++ *buf++ = pad; ++} ++ ++/* ++ * Returns the hash of a filename. If len is 0 and name is NULL, then ++ * this function can be used to test whether or not a hash version is ++ * supported. ++ * ++ * The seed is an 4 longword (32 bits) "secret" which can be used to ++ * uniquify a hash. If the seed is all zero's, then some default seed ++ * may be used. ++ * ++ * A particular hash version specifies whether or not the seed is ++ * represented, and whether or not the returned hash is 32 bits or 64 ++ * bits. 32 bit hashes will return 0 for the minor hash. ++ */ ++int ext3fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo) ++{ ++ __u32 hash; ++ __u32 minor_hash = 0; ++ const char *p; ++ int i; ++ __u32 in[8], buf[4]; ++ ++ /* Initialize the default seed for the hash checksum functions */ ++ buf[0] = 0x67452301; ++ buf[1] = 0xefcdab89; ++ buf[2] = 0x98badcfe; ++ buf[3] = 0x10325476; ++ ++ /* Check to see if the seed is all zero's */ ++ if (hinfo->seed) { ++ for (i=0; i < 4; i++) { ++ if (hinfo->seed[i]) ++ break; ++ } ++ if (i < 4) ++ memcpy(buf, hinfo->seed, sizeof(buf)); ++ } ++ ++ switch (hinfo->hash_version) { ++ case DX_HASH_LEGACY: ++ hash = dx_hack_hash(name, len); ++ break; ++ case DX_HASH_HALF_MD4: ++ p = name; ++ while (len > 0) { ++ str2hashbuf(p, len, in, 8); ++ halfMD4Transform(buf, in); ++ len -= 32; ++ p += 32; ++ } ++ minor_hash = buf[2]; ++ hash = buf[1]; ++ break; ++ case DX_HASH_TEA: ++ p = name; ++ while (len > 0) { ++ str2hashbuf(p, len, in, 4); ++ TEA_transform(buf, in); ++ len -= 16; ++ p += 16; ++ } ++ hash = buf[0]; ++ minor_hash = buf[1]; ++ break; ++ default: ++ hinfo->hash = 0; ++ return -1; ++ } ++ hinfo->hash = hash & ~1; ++ hinfo->minor_hash = minor_hash; ++ return 0; ++} +--- linux-2.4.22-ac1/fs/ext3/Makefile~ext3-htree-2.4.22-rh 2003-09-25 14:39:01.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/Makefile 2003-09-25 14:55:12.000000000 +0400 +@@ -12,7 +12,7 @@ O_TARGET := ext3.o + export-objs := super.o inode.o + + obj-y := balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o \ +- ioctl.o namei.o super.o symlink.o ++ ioctl.o namei.o super.o symlink.o hash.o + obj-m := $(O_TARGET) + + include $(TOPDIR)/Rules.make +--- linux-2.4.22-ac1/fs/ext3/namei.c~ext3-htree-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/namei.c 2003-09-25 14:58:37.000000000 +0400 +@@ -16,6 +16,12 @@ + * David S. Miller (davem@caip.rutgers.edu), 1995 + * Directory entry file type support and forward compatibility hooks + * for B-tree directories by Theodore Ts'o (tytso@mit.edu), 1998 ++ * Hash Tree Directory indexing (c) ++ * Daniel Phillips, 2001 ++ * Hash Tree Directory indexing porting ++ * Christopher Li, 2002 ++ * Hash Tree Directory indexing cleanup ++ * Theodore Ts'o, 2002 + */ + + #include +@@ -38,6 +44,642 @@ + #define NAMEI_RA_SIZE (NAMEI_RA_CHUNKS * NAMEI_RA_BLOCKS) + #define NAMEI_RA_INDEX(c,b) (((c) * NAMEI_RA_BLOCKS) + (b)) + ++static struct buffer_head *ext3_append(handle_t *handle, ++ struct inode *inode, ++ u32 *block, int *err) ++{ ++ struct buffer_head *bh; ++ ++ *block = inode->i_size >> inode->i_sb->s_blocksize_bits; ++ ++ if ((bh = ext3_bread(handle, inode, *block, 1, err))) { ++ inode->i_size += inode->i_sb->s_blocksize; ++ EXT3_I(inode)->i_disksize = inode->i_size; ++ ext3_journal_get_write_access(handle,bh); ++ } ++ return bh; ++} ++ ++#ifndef assert ++#define assert(test) J_ASSERT(test) ++#endif ++ ++#ifndef swap ++#define swap(x, y) do { typeof(x) z = x; x = y; y = z; } while (0) ++#endif ++ ++typedef struct { u32 v; } le_u32; ++typedef struct { u16 v; } le_u16; ++ ++#ifdef DX_DEBUG ++#define dxtrace(command) command ++#else ++#define dxtrace(command) ++#endif ++ ++struct fake_dirent ++{ ++ /*le*/u32 inode; ++ /*le*/u16 rec_len; ++ u8 name_len; ++ u8 file_type; ++}; ++ ++struct dx_countlimit ++{ ++ le_u16 limit; ++ le_u16 count; ++}; ++ ++struct dx_entry ++{ ++ le_u32 hash; ++ le_u32 block; ++}; ++ ++/* ++ * dx_root_info is laid out so that if it should somehow get overlaid by a ++ * dirent the two low bits of the hash version will be zero. Therefore, the ++ * hash version mod 4 should never be 0. Sincerely, the paranoia department. ++ */ ++ ++struct dx_root ++{ ++ struct fake_dirent dot; ++ char dot_name[4]; ++ struct fake_dirent dotdot; ++ char dotdot_name[4]; ++ struct dx_root_info ++ { ++ le_u32 reserved_zero; ++ u8 hash_version; ++ u8 info_length; /* 8 */ ++ u8 indirect_levels; ++ u8 unused_flags; ++ } ++ info; ++ struct dx_entry entries[0]; ++}; ++ ++struct dx_node ++{ ++ struct fake_dirent fake; ++ struct dx_entry entries[0]; ++}; ++ ++ ++struct dx_frame ++{ ++ struct buffer_head *bh; ++ struct dx_entry *entries; ++ struct dx_entry *at; ++}; ++ ++struct dx_map_entry ++{ ++ u32 hash; ++ u32 offs; ++}; ++ ++#ifdef CONFIG_EXT3_INDEX ++static inline unsigned dx_get_block (struct dx_entry *entry); ++static void dx_set_block (struct dx_entry *entry, unsigned value); ++static inline unsigned dx_get_hash (struct dx_entry *entry); ++static void dx_set_hash (struct dx_entry *entry, unsigned value); ++static unsigned dx_get_count (struct dx_entry *entries); ++static unsigned dx_get_limit (struct dx_entry *entries); ++static void dx_set_count (struct dx_entry *entries, unsigned value); ++static void dx_set_limit (struct dx_entry *entries, unsigned value); ++static unsigned dx_root_limit (struct inode *dir, unsigned infosize); ++static unsigned dx_node_limit (struct inode *dir); ++static struct dx_frame *dx_probe(struct dentry *dentry, ++ struct inode *dir, ++ struct dx_hash_info *hinfo, ++ struct dx_frame *frame, ++ int *err); ++static void dx_release (struct dx_frame *frames); ++static int dx_make_map (struct ext3_dir_entry_2 *de, int size, ++ struct dx_hash_info *hinfo, struct dx_map_entry map[]); ++static void dx_sort_map(struct dx_map_entry *map, unsigned count); ++static struct ext3_dir_entry_2 *dx_move_dirents (char *from, char *to, ++ struct dx_map_entry *offsets, int count); ++static struct ext3_dir_entry_2* dx_pack_dirents (char *base, int size); ++static void dx_insert_block (struct dx_frame *frame, u32 hash, u32 block); ++static int ext3_htree_next_block(struct inode *dir, __u32 hash, ++ struct dx_frame *frame, ++ struct dx_frame *frames, int *err, ++ __u32 *start_hash); ++static struct buffer_head * ext3_dx_find_entry(struct dentry *dentry, ++ struct ext3_dir_entry_2 **res_dir, int *err); ++static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry, ++ struct inode *inode); ++ ++/* ++ * Future: use high four bits of block for coalesce-on-delete flags ++ * Mask them off for now. ++ */ ++ ++static inline unsigned dx_get_block (struct dx_entry *entry) ++{ ++ return le32_to_cpu(entry->block.v) & 0x00ffffff; ++} ++ ++static inline void dx_set_block (struct dx_entry *entry, unsigned value) ++{ ++ entry->block.v = cpu_to_le32(value); ++} ++ ++static inline unsigned dx_get_hash (struct dx_entry *entry) ++{ ++ return le32_to_cpu(entry->hash.v); ++} ++ ++static inline void dx_set_hash (struct dx_entry *entry, unsigned value) ++{ ++ entry->hash.v = cpu_to_le32(value); ++} ++ ++static inline unsigned dx_get_count (struct dx_entry *entries) ++{ ++ return le16_to_cpu(((struct dx_countlimit *) entries)->count.v); ++} ++ ++static inline unsigned dx_get_limit (struct dx_entry *entries) ++{ ++ return le16_to_cpu(((struct dx_countlimit *) entries)->limit.v); ++} ++ ++static inline void dx_set_count (struct dx_entry *entries, unsigned value) ++{ ++ ((struct dx_countlimit *) entries)->count.v = cpu_to_le16(value); ++} ++ ++static inline void dx_set_limit (struct dx_entry *entries, unsigned value) ++{ ++ ((struct dx_countlimit *) entries)->limit.v = cpu_to_le16(value); ++} ++ ++static inline unsigned dx_root_limit (struct inode *dir, unsigned infosize) ++{ ++ unsigned entry_space = dir->i_sb->s_blocksize - EXT3_DIR_REC_LEN(1) - ++ EXT3_DIR_REC_LEN(2) - infosize; ++ return 0? 20: entry_space / sizeof(struct dx_entry); ++} ++ ++static inline unsigned dx_node_limit (struct inode *dir) ++{ ++ unsigned entry_space = dir->i_sb->s_blocksize - EXT3_DIR_REC_LEN(0); ++ return 0? 22: entry_space / sizeof(struct dx_entry); ++} ++ ++/* ++ * Debug ++ */ ++#ifdef DX_DEBUG ++struct stats ++{ ++ unsigned names; ++ unsigned space; ++ unsigned bcount; ++}; ++ ++static struct stats dx_show_leaf(struct dx_hash_info *hinfo, struct ext3_dir_entry_2 *de, ++ int size, int show_names) ++{ ++ unsigned names = 0, space = 0; ++ char *base = (char *) de; ++ struct dx_hash_info h = *hinfo; ++ ++ printk("names: "); ++ while ((char *) de < base + size) ++ { ++ if (de->inode) ++ { ++ if (show_names) ++ { ++ int len = de->name_len; ++ char *name = de->name; ++ while (len--) printk("%c", *name++); ++ ext3fs_dirhash(de->name, de->name_len, &h); ++ printk(":%x.%u ", h.hash, ++ ((char *) de - base)); ++ } ++ space += EXT3_DIR_REC_LEN(de->name_len); ++ names++; ++ } ++ de = (struct ext3_dir_entry_2 *) ((char *) de + le16_to_cpu(de->rec_len)); ++ } ++ printk("(%i)\n", names); ++ return (struct stats) { names, space, 1 }; ++} ++ ++struct stats dx_show_entries(struct dx_hash_info *hinfo, struct inode *dir, ++ struct dx_entry *entries, int levels) ++{ ++ unsigned blocksize = dir->i_sb->s_blocksize; ++ unsigned count = dx_get_count (entries), names = 0, space = 0, i; ++ unsigned bcount = 0; ++ struct buffer_head *bh; ++ int err; ++ printk("%i indexed blocks...\n", count); ++ for (i = 0; i < count; i++, entries++) ++ { ++ u32 block = dx_get_block(entries), hash = i? dx_get_hash(entries): 0; ++ u32 range = i < count - 1? (dx_get_hash(entries + 1) - hash): ~hash; ++ struct stats stats; ++ printk("%s%3u:%03u hash %8x/%8x ",levels?"":" ", i, block, hash, range); ++ if (!(bh = ext3_bread (NULL,dir, block, 0,&err))) continue; ++ stats = levels? ++ dx_show_entries(hinfo, dir, ((struct dx_node *) bh->b_data)->entries, levels - 1): ++ dx_show_leaf(hinfo, (struct ext3_dir_entry_2 *) bh->b_data, blocksize, 0); ++ names += stats.names; ++ space += stats.space; ++ bcount += stats.bcount; ++ brelse (bh); ++ } ++ if (bcount) ++ printk("%snames %u, fullness %u (%u%%)\n", levels?"":" ", ++ names, space/bcount,(space/bcount)*100/blocksize); ++ return (struct stats) { names, space, bcount}; ++} ++#endif /* DX_DEBUG */ ++ ++/* ++ * Probe for a directory leaf block to search. ++ * ++ * dx_probe can return ERR_BAD_DX_DIR, which means there was a format ++ * error in the directory index, and the caller should fall back to ++ * searching the directory normally. The callers of dx_probe **MUST** ++ * check for this error code, and make sure it never gets reflected ++ * back to userspace. ++ */ ++static struct dx_frame * ++dx_probe(struct dentry *dentry, struct inode *dir, ++ struct dx_hash_info *hinfo, struct dx_frame *frame_in, int *err) ++{ ++ unsigned count, indirect; ++ struct dx_entry *at, *entries, *p, *q, *m; ++ struct dx_root *root; ++ struct buffer_head *bh; ++ struct dx_frame *frame = frame_in; ++ u32 hash; ++ ++ frame->bh = NULL; ++ if (dentry) ++ dir = dentry->d_parent->d_inode; ++ if (!(bh = ext3_bread (NULL,dir, 0, 0, err))) ++ goto fail; ++ root = (struct dx_root *) bh->b_data; ++ if (root->info.hash_version != DX_HASH_TEA && ++ root->info.hash_version != DX_HASH_HALF_MD4 && ++ root->info.hash_version != DX_HASH_LEGACY) { ++ ext3_warning(dir->i_sb, __FUNCTION__, ++ "Unrecognised inode hash code %d", ++ root->info.hash_version); ++ brelse(bh); ++ *err = ERR_BAD_DX_DIR; ++ goto fail; ++ } ++ hinfo->hash_version = root->info.hash_version; ++ hinfo->seed = dir->i_sb->u.ext3_sb.s_hash_seed; ++ if (dentry) ++ ext3fs_dirhash(dentry->d_name.name, dentry->d_name.len, hinfo); ++ hash = hinfo->hash; ++ ++ if (root->info.unused_flags & 1) { ++ ext3_warning(dir->i_sb, __FUNCTION__, ++ "Unimplemented inode hash flags: %#06x", ++ root->info.unused_flags); ++ brelse(bh); ++ *err = ERR_BAD_DX_DIR; ++ goto fail; ++ } ++ ++ if ((indirect = root->info.indirect_levels) > 1) { ++ ext3_warning(dir->i_sb, __FUNCTION__, ++ "Unimplemented inode hash depth: %#06x", ++ root->info.indirect_levels); ++ brelse(bh); ++ *err = ERR_BAD_DX_DIR; ++ goto fail; ++ } ++ ++ entries = (struct dx_entry *) (((char *)&root->info) + ++ root->info.info_length); ++ assert(dx_get_limit(entries) == dx_root_limit(dir, ++ root->info.info_length)); ++ dxtrace (printk("Look up %x", hash)); ++ while (1) ++ { ++ count = dx_get_count(entries); ++ assert (count && count <= dx_get_limit(entries)); ++ p = entries + 1; ++ q = entries + count - 1; ++ while (p <= q) ++ { ++ m = p + (q - p)/2; ++ dxtrace(printk(".")); ++ if (dx_get_hash(m) > hash) ++ q = m - 1; ++ else ++ p = m + 1; ++ } ++ ++ if (0) // linear search cross check ++ { ++ unsigned n = count - 1; ++ at = entries; ++ while (n--) ++ { ++ dxtrace(printk(",")); ++ if (dx_get_hash(++at) > hash) ++ { ++ at--; ++ break; ++ } ++ } ++ assert (at == p - 1); ++ } ++ ++ at = p - 1; ++ dxtrace(printk(" %x->%u\n", at == entries? 0: dx_get_hash(at), dx_get_block(at))); ++ frame->bh = bh; ++ frame->entries = entries; ++ frame->at = at; ++ if (!indirect--) return frame; ++ if (!(bh = ext3_bread (NULL,dir, dx_get_block(at), 0, err))) ++ goto fail2; ++ at = entries = ((struct dx_node *) bh->b_data)->entries; ++ assert (dx_get_limit(entries) == dx_node_limit (dir)); ++ frame++; ++ } ++fail2: ++ while (frame >= frame_in) { ++ brelse(frame->bh); ++ frame--; ++ } ++fail: ++ return NULL; ++} ++ ++static void dx_release (struct dx_frame *frames) ++{ ++ if (frames[0].bh == NULL) ++ return; ++ ++ if (((struct dx_root *) frames[0].bh->b_data)->info.indirect_levels) ++ brelse(frames[1].bh); ++ brelse(frames[0].bh); ++} ++ ++/* ++ * This function increments the frame pointer to search the next leaf ++ * block, and reads in the necessary intervening nodes if the search ++ * should be necessary. Whether or not the search is necessary is ++ * controlled by the hash parameter. If the hash value is even, then ++ * the search is only continued if the next block starts with that ++ * hash value. This is used if we are searching for a specific file. ++ * ++ * If the hash value is HASH_NB_ALWAYS, then always go to the next block. ++ * ++ * This function returns 1 if the caller should continue to search, ++ * or 0 if it should not. If there is an error reading one of the ++ * index blocks, it will return -1. ++ * ++ * If start_hash is non-null, it will be filled in with the starting ++ * hash of the next page. ++ */ ++static int ext3_htree_next_block(struct inode *dir, __u32 hash, ++ struct dx_frame *frame, ++ struct dx_frame *frames, int *err, ++ __u32 *start_hash) ++{ ++ struct dx_frame *p; ++ struct buffer_head *bh; ++ int num_frames = 0; ++ __u32 bhash; ++ ++ *err = ENOENT; ++ p = frame; ++ /* ++ * Find the next leaf page by incrementing the frame pointer. ++ * If we run out of entries in the interior node, loop around and ++ * increment pointer in the parent node. When we break out of ++ * this loop, num_frames indicates the number of interior ++ * nodes need to be read. ++ */ ++ while (1) { ++ if (++(p->at) < p->entries + dx_get_count(p->entries)) ++ break; ++ if (p == frames) ++ return 0; ++ num_frames++; ++ p--; ++ } ++ ++ /* ++ * If the hash is 1, then continue only if the next page has a ++ * continuation hash of any value. This is used for readdir ++ * handling. Otherwise, check to see if the hash matches the ++ * desired contiuation hash. If it doesn't, return since ++ * there's no point to read in the successive index pages. ++ */ ++ bhash = dx_get_hash(p->at); ++ if (start_hash) ++ *start_hash = bhash; ++ if ((hash & 1) == 0) { ++ if ((bhash & ~1) != hash) ++ return 0; ++ } ++ /* ++ * If the hash is HASH_NB_ALWAYS, we always go to the next ++ * block so no check is necessary ++ */ ++ while (num_frames--) { ++ if (!(bh = ext3_bread(NULL, dir, dx_get_block(p->at), ++ 0, err))) ++ return -1; /* Failure */ ++ p++; ++ brelse (p->bh); ++ p->bh = bh; ++ p->at = p->entries = ((struct dx_node *) bh->b_data)->entries; ++ } ++ return 1; ++} ++ ++ ++/* ++ * p is at least 6 bytes before the end of page ++ */ ++static inline struct ext3_dir_entry_2 *ext3_next_entry(struct ext3_dir_entry_2 *p) ++{ ++ return (struct ext3_dir_entry_2 *)((char*)p + le16_to_cpu(p->rec_len)); ++} ++ ++/* ++ * This function fills a red-black tree with information from a ++ * directory. We start scanning the directory in hash order, starting ++ * at start_hash and start_minor_hash. ++ * ++ * This function returns the number of entries inserted into the tree, ++ * or a negative error code. ++ */ ++int ext3_htree_fill_tree(struct file *dir_file, __u32 start_hash, ++ __u32 start_minor_hash, __u32 *next_hash) ++{ ++ struct dx_hash_info hinfo; ++ struct buffer_head *bh; ++ struct ext3_dir_entry_2 *de, *top; ++ static struct dx_frame frames[2], *frame; ++ struct inode *dir; ++ int block, err; ++ int count = 0; ++ int ret; ++ __u32 hashval; ++ ++ dxtrace(printk("In htree_fill_tree, start hash: %x:%x\n", start_hash, ++ start_minor_hash)); ++ dir = dir_file->f_dentry->d_inode; ++ hinfo.hash = start_hash; ++ hinfo.minor_hash = 0; ++ frame = dx_probe(0, dir_file->f_dentry->d_inode, &hinfo, frames, &err); ++ if (!frame) ++ return err; ++ ++ /* Add '.' and '..' from the htree header */ ++ if (!start_hash && !start_minor_hash) { ++ de = (struct ext3_dir_entry_2 *) frames[0].bh->b_data; ++ if ((err = ext3_htree_store_dirent(dir_file, 0, 0, de)) != 0) ++ goto errout; ++ de = ext3_next_entry(de); ++ if ((err = ext3_htree_store_dirent(dir_file, 0, 0, de)) != 0) ++ goto errout; ++ count += 2; ++ } ++ ++ while (1) { ++ block = dx_get_block(frame->at); ++ dxtrace(printk("Reading block %d\n", block)); ++ if (!(bh = ext3_bread (NULL, dir, block, 0, &err))) ++ goto errout; ++ ++ de = (struct ext3_dir_entry_2 *) bh->b_data; ++ top = (struct ext3_dir_entry_2 *) ((char *) de + dir->i_sb->s_blocksize - ++ EXT3_DIR_REC_LEN(0)); ++ for (; de < top; de = ext3_next_entry(de)) { ++ ext3fs_dirhash(de->name, de->name_len, &hinfo); ++ if ((hinfo.hash < start_hash) || ++ ((hinfo.hash == start_hash) && ++ (hinfo.minor_hash < start_minor_hash))) ++ continue; ++ if ((err = ext3_htree_store_dirent(dir_file, ++ hinfo.hash, hinfo.minor_hash, de)) != 0) ++ goto errout; ++ count++; ++ } ++ brelse (bh); ++ hashval = ~1; ++ ret = ext3_htree_next_block(dir, HASH_NB_ALWAYS, ++ frame, frames, &err, &hashval); ++ if (next_hash) ++ *next_hash = hashval; ++ if (ret == -1) ++ goto errout; ++ /* ++ * Stop if: (a) there are no more entries, or ++ * (b) we have inserted at least one entry and the ++ * next hash value is not a continuation ++ */ ++ if ((ret == 0) || ++ (count && ((hashval & 1) == 0))) ++ break; ++ } ++ dx_release(frames); ++ dxtrace(printk("Fill tree: returned %d entries\n", count)); ++ return count; ++errout: ++ dx_release(frames); ++ return (err); ++} ++ ++ ++/* ++ * Directory block splitting, compacting ++ */ ++ ++static int dx_make_map (struct ext3_dir_entry_2 *de, int size, ++ struct dx_hash_info *hinfo, struct dx_map_entry *map_tail) ++{ ++ int count = 0; ++ char *base = (char *) de; ++ struct dx_hash_info h = *hinfo; ++ ++ while ((char *) de < base + size) ++ { ++ if (de->name_len && de->inode) { ++ ext3fs_dirhash(de->name, de->name_len, &h); ++ map_tail--; ++ map_tail->hash = h.hash; ++ map_tail->offs = (u32) ((char *) de - base); ++ count++; ++ } ++ /* XXX: do we need to check rec_len == 0 case? -Chris */ ++ de = (struct ext3_dir_entry_2 *) ((char *) de + le16_to_cpu(de->rec_len)); ++ } ++ return count; ++} ++ ++static void dx_sort_map (struct dx_map_entry *map, unsigned count) ++{ ++ struct dx_map_entry *p, *q, *top = map + count - 1; ++ int more; ++ /* Combsort until bubble sort doesn't suck */ ++ while (count > 2) ++ { ++ count = count*10/13; ++ if (count - 9 < 2) /* 9, 10 -> 11 */ ++ count = 11; ++ for (p = top, q = p - count; q >= map; p--, q--) ++ if (p->hash < q->hash) ++ swap(*p, *q); ++ } ++ /* Garden variety bubble sort */ ++ do { ++ more = 0; ++ q = top; ++ while (q-- > map) ++ { ++ if (q[1].hash >= q[0].hash) ++ continue; ++ swap(*(q+1), *q); ++ more = 1; ++ } ++ } while(more); ++} ++ ++static void dx_insert_block(struct dx_frame *frame, u32 hash, u32 block) ++{ ++ struct dx_entry *entries = frame->entries; ++ struct dx_entry *old = frame->at, *new = old + 1; ++ int count = dx_get_count(entries); ++ ++ assert(count < dx_get_limit(entries)); ++ assert(old < entries + count); ++ memmove(new + 1, new, (char *)(entries + count) - (char *)(new)); ++ dx_set_hash(new, hash); ++ dx_set_block(new, block); ++ dx_set_count(entries, count + 1); ++} ++#endif ++ ++ ++static void ext3_update_dx_flag(struct inode *inode) ++{ ++ if (!EXT3_HAS_COMPAT_FEATURE(inode->i_sb, ++ EXT3_FEATURE_COMPAT_DIR_INDEX)) ++ EXT3_I(inode)->i_flags &= ~EXT3_INDEX_FL; ++} ++ + /* + * NOTE! unlike strncmp, ext3_match returns 1 for success, 0 for failure. + * +@@ -94,6 +736,7 @@ static int inline search_dirblock(struct + return 0; + } + ++ + /* + * ext3_find_entry() + * +@@ -105,6 +748,8 @@ static int inline search_dirblock(struct + * The returned buffer_head has ->b_count elevated. The caller is expected + * to brelse() it when appropriate. + */ ++ ++ + static struct buffer_head * ext3_find_entry (struct dentry *dentry, + struct ext3_dir_entry_2 ** res_dir) + { +@@ -119,12 +764,32 @@ static struct buffer_head * ext3_find_en + int num = 0; + int nblocks, i, err; + struct inode *dir = dentry->d_parent->d_inode; ++ int namelen; ++ const u8 *name; ++ unsigned blocksize; + + *res_dir = NULL; + sb = dir->i_sb; +- ++ blocksize = sb->s_blocksize; ++ namelen = dentry->d_name.len; ++ name = dentry->d_name.name; ++ if (namelen > EXT3_NAME_LEN) ++ return NULL; ++#ifdef CONFIG_EXT3_INDEX ++ if (is_dx(dir)) { ++ bh = ext3_dx_find_entry(dentry, res_dir, &err); ++ /* ++ * On success, or if the error was file not found, ++ * return. Otherwise, fall back to doing a search the ++ * old fashioned way. ++ */ ++ if (bh || (err != ERR_BAD_DX_DIR)) ++ return bh; ++ dxtrace(printk("ext3_find_entry: dx failed, falling back\n")); ++ } ++#endif + nblocks = dir->i_size >> EXT3_BLOCK_SIZE_BITS(sb); +- start = dir->u.ext3_i.i_dir_start_lookup; ++ start = EXT3_I(dir)->i_dir_start_lookup; + if (start >= nblocks) + start = 0; + block = start; +@@ -166,7 +831,7 @@ restart: + i = search_dirblock(bh, dir, dentry, + block << EXT3_BLOCK_SIZE_BITS(sb), res_dir); + if (i == 1) { +- dir->u.ext3_i.i_dir_start_lookup = block; ++ EXT3_I(dir)->i_dir_start_lookup = block; + ret = bh; + goto cleanup_and_exit; + } else { +@@ -197,6 +862,66 @@ cleanup_and_exit: + return ret; + } + ++#ifdef CONFIG_EXT3_INDEX ++static struct buffer_head * ext3_dx_find_entry(struct dentry *dentry, ++ struct ext3_dir_entry_2 **res_dir, int *err) ++{ ++ struct super_block * sb; ++ struct dx_hash_info hinfo; ++ u32 hash; ++ struct dx_frame frames[2], *frame; ++ struct ext3_dir_entry_2 *de, *top; ++ struct buffer_head *bh; ++ unsigned long block; ++ int retval; ++ int namelen = dentry->d_name.len; ++ const u8 *name = dentry->d_name.name; ++ struct inode *dir = dentry->d_parent->d_inode; ++ ++ sb = dir->i_sb; ++ if (!(frame = dx_probe (dentry, 0, &hinfo, frames, err))) ++ return NULL; ++ hash = hinfo.hash; ++ do { ++ block = dx_get_block(frame->at); ++ if (!(bh = ext3_bread (NULL,dir, block, 0, err))) ++ goto errout; ++ de = (struct ext3_dir_entry_2 *) bh->b_data; ++ top = (struct ext3_dir_entry_2 *) ((char *) de + sb->s_blocksize - ++ EXT3_DIR_REC_LEN(0)); ++ for (; de < top; de = ext3_next_entry(de)) ++ if (ext3_match (namelen, name, de)) { ++ if (!ext3_check_dir_entry("ext3_find_entry", ++ dir, de, bh, ++ (block<b_data))) { ++ brelse (bh); ++ goto errout; ++ } ++ *res_dir = de; ++ dx_release (frames); ++ return bh; ++ } ++ brelse (bh); ++ /* Check to see if we should continue to search */ ++ retval = ext3_htree_next_block(dir, hash, frame, ++ frames, err, 0); ++ if (retval == -1) { ++ ext3_warning(sb, __FUNCTION__, ++ "error reading index page in directory #%lu", ++ dir->i_ino); ++ goto errout; ++ } ++ } while (retval == 1); ++ ++ *err = -ENOENT; ++errout: ++ dxtrace(printk("%s not found\n", name)); ++ dx_release (frames); ++ return NULL; ++} ++#endif ++ + static struct dentry *ext3_lookup(struct inode * dir, struct dentry *dentry) + { + struct inode * inode; +@@ -213,8 +938,9 @@ static struct dentry *ext3_lookup(struct + brelse (bh); + inode = iget(dir->i_sb, ino); + +- if (!inode) ++ if (!inode) { + return ERR_PTR(-EACCES); ++ } + } + d_add(dentry, inode); + return NULL; +@@ -238,6 +964,301 @@ static inline void ext3_set_de_type(stru + de->file_type = ext3_type_by_mode[(mode & S_IFMT)>>S_SHIFT]; + } + ++#ifdef CONFIG_EXT3_INDEX ++static struct ext3_dir_entry_2 * ++dx_move_dirents(char *from, char *to, struct dx_map_entry *map, int count) ++{ ++ unsigned rec_len = 0; ++ ++ while (count--) { ++ struct ext3_dir_entry_2 *de = (struct ext3_dir_entry_2 *) (from + map->offs); ++ rec_len = EXT3_DIR_REC_LEN(de->name_len); ++ memcpy (to, de, rec_len); ++ ((struct ext3_dir_entry_2 *) to)->rec_len = rec_len; ++ de->inode = 0; ++ map++; ++ to += rec_len; ++ } ++ return (struct ext3_dir_entry_2 *) (to - rec_len); ++} ++ ++static struct ext3_dir_entry_2* dx_pack_dirents(char *base, int size) ++{ ++ struct ext3_dir_entry_2 *next, *to, *prev, *de = (struct ext3_dir_entry_2 *) base; ++ unsigned rec_len = 0; ++ ++ prev = to = de; ++ while ((char*)de < base + size) { ++ next = (struct ext3_dir_entry_2 *) ((char *) de + ++ le16_to_cpu(de->rec_len)); ++ if (de->inode && de->name_len) { ++ rec_len = EXT3_DIR_REC_LEN(de->name_len); ++ if (de > to) ++ memmove(to, de, rec_len); ++ to->rec_len = rec_len; ++ prev = to; ++ to = (struct ext3_dir_entry_2 *) (((char *) to) + rec_len); ++ } ++ de = next; ++ } ++ return prev; ++} ++ ++static struct ext3_dir_entry_2 *do_split(handle_t *handle, struct inode *dir, ++ struct buffer_head **bh,struct dx_frame *frame, ++ struct dx_hash_info *hinfo, int *error) ++{ ++ unsigned blocksize = dir->i_sb->s_blocksize; ++ unsigned count, continued; ++ struct buffer_head *bh2; ++ u32 newblock; ++ u32 hash2; ++ struct dx_map_entry *map; ++ char *data1 = (*bh)->b_data, *data2; ++ unsigned split; ++ struct ext3_dir_entry_2 *de = NULL, *de2; ++ int err; ++ ++ bh2 = ext3_append (handle, dir, &newblock, error); ++ if (!(bh2)) { ++ brelse(*bh); ++ *bh = NULL; ++ goto errout; ++ } ++ ++ BUFFER_TRACE(*bh, "get_write_access"); ++ err = ext3_journal_get_write_access(handle, *bh); ++ if (err) { ++ journal_error: ++ brelse(*bh); ++ brelse(bh2); ++ *bh = NULL; ++ ext3_std_error(dir->i_sb, err); ++ goto errout; ++ } ++ BUFFER_TRACE(frame->bh, "get_write_access"); ++ err = ext3_journal_get_write_access(handle, frame->bh); ++ if (err) ++ goto journal_error; ++ ++ data2 = bh2->b_data; ++ ++ /* create map in the end of data2 block */ ++ map = (struct dx_map_entry *) (data2 + blocksize); ++ count = dx_make_map ((struct ext3_dir_entry_2 *) data1, ++ blocksize, hinfo, map); ++ map -= count; ++ split = count/2; // need to adjust to actual middle ++ dx_sort_map (map, count); ++ hash2 = map[split].hash; ++ continued = hash2 == map[split - 1].hash; ++ dxtrace(printk("Split block %i at %x, %i/%i\n", ++ dx_get_block(frame->at), hash2, split, count-split)); ++ ++ /* Fancy dance to stay within two buffers */ ++ de2 = dx_move_dirents(data1, data2, map + split, count - split); ++ de = dx_pack_dirents(data1,blocksize); ++ de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de); ++ de2->rec_len = cpu_to_le16(data2 + blocksize - (char *) de2); ++ dxtrace(dx_show_leaf (hinfo, (struct ext3_dir_entry_2 *) data1, blocksize, 1)); ++ dxtrace(dx_show_leaf (hinfo, (struct ext3_dir_entry_2 *) data2, blocksize, 1)); ++ ++ /* Which block gets the new entry? */ ++ if (hinfo->hash >= hash2) ++ { ++ swap(*bh, bh2); ++ de = de2; ++ } ++ dx_insert_block (frame, hash2 + continued, newblock); ++ err = ext3_journal_dirty_metadata (handle, bh2); ++ if (err) ++ goto journal_error; ++ err = ext3_journal_dirty_metadata (handle, frame->bh); ++ if (err) ++ goto journal_error; ++ brelse (bh2); ++ dxtrace(dx_show_index ("frame", frame->entries)); ++errout: ++ return de; ++} ++#endif ++ ++ ++/* ++ * Add a new entry into a directory (leaf) block. If de is non-NULL, ++ * it points to a directory entry which is guaranteed to be large ++ * enough for new directory entry. If de is NULL, then ++ * add_dirent_to_buf will attempt search the directory block for ++ * space. It will return -ENOSPC if no space is available, and -EIO ++ * and -EEXIST if directory entry already exists. ++ * ++ * NOTE! bh is NOT released in the case where ENOSPC is returned. In ++ * all other cases bh is released. ++ */ ++static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry, ++ struct inode *inode, struct ext3_dir_entry_2 *de, ++ struct buffer_head * bh) ++{ ++ struct inode *dir = dentry->d_parent->d_inode; ++ const char *name = dentry->d_name.name; ++ int namelen = dentry->d_name.len; ++ unsigned long offset = 0; ++ unsigned short reclen; ++ int nlen, rlen, err; ++ char *top; ++ ++ reclen = EXT3_DIR_REC_LEN(namelen); ++ if (!de) { ++ de = (struct ext3_dir_entry_2 *)bh->b_data; ++ top = bh->b_data + dir->i_sb->s_blocksize - reclen; ++ while ((char *) de <= top) { ++ if (!ext3_check_dir_entry("ext3_add_entry", dir, de, ++ bh, offset)) { ++ brelse (bh); ++ return -EIO; ++ } ++ if (ext3_match (namelen, name, de)) { ++ brelse (bh); ++ return -EEXIST; ++ } ++ nlen = EXT3_DIR_REC_LEN(de->name_len); ++ rlen = le16_to_cpu(de->rec_len); ++ if ((de->inode? rlen - nlen: rlen) >= reclen) ++ break; ++ de = (struct ext3_dir_entry_2 *)((char *)de + rlen); ++ offset += rlen; ++ } ++ if ((char *) de > top) ++ return -ENOSPC; ++ } ++ BUFFER_TRACE(bh, "get_write_access"); ++ err = ext3_journal_get_write_access(handle, bh); ++ if (err) { ++ ext3_std_error(dir->i_sb, err); ++ brelse(bh); ++ return err; ++ } ++ ++ /* By now the buffer is marked for journaling */ ++ nlen = EXT3_DIR_REC_LEN(de->name_len); ++ rlen = le16_to_cpu(de->rec_len); ++ if (de->inode) { ++ struct ext3_dir_entry_2 *de1 = (struct ext3_dir_entry_2 *)((char *)de + nlen); ++ de1->rec_len = cpu_to_le16(rlen - nlen); ++ de->rec_len = cpu_to_le16(nlen); ++ de = de1; ++ } ++ de->file_type = EXT3_FT_UNKNOWN; ++ if (inode) { ++ de->inode = cpu_to_le32(inode->i_ino); ++ ext3_set_de_type(dir->i_sb, de, inode->i_mode); ++ } else ++ de->inode = 0; ++ de->name_len = namelen; ++ memcpy (de->name, name, namelen); ++ /* ++ * XXX shouldn't update any times until successful ++ * completion of syscall, but too many callers depend ++ * on this. ++ * ++ * XXX similarly, too many callers depend on ++ * ext3_new_inode() setting the times, but error ++ * recovery deletes the inode, so the worst that can ++ * happen is that the times are slightly out of date ++ * and/or different from the directory change time. ++ */ ++ dir->i_mtime = dir->i_ctime = CURRENT_TIME; ++ ext3_update_dx_flag(dir); ++ dir->i_version = ++event; ++ ext3_mark_inode_dirty(handle, dir); ++ BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata"); ++ err = ext3_journal_dirty_metadata(handle, bh); ++ if (err) ++ ext3_std_error(dir->i_sb, err); ++ brelse(bh); ++ return 0; ++} ++ ++#ifdef CONFIG_EXT3_INDEX ++/* ++ * This converts a one block unindexed directory to a 3 block indexed ++ * directory, and adds the dentry to the indexed directory. ++ */ ++static int make_indexed_dir(handle_t *handle, struct dentry *dentry, ++ struct inode *inode, struct buffer_head *bh) ++{ ++ struct inode *dir = dentry->d_parent->d_inode; ++ const char *name = dentry->d_name.name; ++ int namelen = dentry->d_name.len; ++ struct buffer_head *bh2; ++ struct dx_root *root; ++ struct dx_frame frames[2], *frame; ++ struct dx_entry *entries; ++ struct ext3_dir_entry_2 *de, *de2; ++ char *data1, *top; ++ unsigned len; ++ int retval; ++ unsigned blocksize; ++ struct dx_hash_info hinfo; ++ u32 block; ++ ++ blocksize = dir->i_sb->s_blocksize; ++ dxtrace(printk("Creating index\n")); ++ retval = ext3_journal_get_write_access(handle, bh); ++ if (retval) { ++ ext3_std_error(dir->i_sb, retval); ++ brelse(bh); ++ return retval; ++ } ++ root = (struct dx_root *) bh->b_data; ++ ++ EXT3_I(dir)->i_flags |= EXT3_INDEX_FL; ++ bh2 = ext3_append (handle, dir, &block, &retval); ++ if (!(bh2)) { ++ brelse(bh); ++ return retval; ++ } ++ data1 = bh2->b_data; ++ ++ /* The 0th block becomes the root, move the dirents out */ ++ de = &root->dotdot; ++ de = (struct ext3_dir_entry_2 *) ((char *)de + de->rec_len); ++ len = ((char *) root) + blocksize - (char *) de; ++ memcpy (data1, de, len); ++ de = (struct ext3_dir_entry_2 *) data1; ++ top = data1 + len; ++ while (((char *) de2=(char*)de+le16_to_cpu(de->rec_len)) < top) ++ de = de2; ++ de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de); ++ /* Initialize the root; the dot dirents already exist */ ++ de = (struct ext3_dir_entry_2 *) (&root->dotdot); ++ de->rec_len = cpu_to_le16(blocksize - EXT3_DIR_REC_LEN(2)); ++ memset (&root->info, 0, sizeof(root->info)); ++ root->info.info_length = sizeof(root->info); ++ root->info.hash_version = dir->i_sb->u.ext3_sb.s_def_hash_version; ++ entries = root->entries; ++ dx_set_block (entries, 1); ++ dx_set_count (entries, 1); ++ dx_set_limit (entries, dx_root_limit(dir, sizeof(root->info))); ++ ++ /* Initialize as for dx_probe */ ++ hinfo.hash_version = root->info.hash_version; ++ hinfo.seed = dir->i_sb->u.ext3_sb.s_hash_seed; ++ ext3fs_dirhash(name, namelen, &hinfo); ++ frame = frames; ++ frame->entries = entries; ++ frame->at = entries; ++ frame->bh = bh; ++ bh = bh2; ++ de = do_split(handle,dir, &bh, frame, &hinfo, &retval); ++ dx_release (frames); ++ if (!(de)) ++ return retval; ++ ++ return add_dirent_to_buf(handle, dentry, inode, de, bh); ++} ++#endif ++ + /* + * ext3_add_entry() + * +@@ -248,127 +1269,198 @@ static inline void ext3_set_de_type(stru + * may not sleep between calling this and putting something into + * the entry, as someone else might have used it while you slept. + */ +- +-/* +- * AKPM: the journalling code here looks wrong on the error paths +- */ + static int ext3_add_entry (handle_t *handle, struct dentry *dentry, + struct inode *inode) + { + struct inode *dir = dentry->d_parent->d_inode; +- const char *name = dentry->d_name.name; +- int namelen = dentry->d_name.len; + unsigned long offset; +- unsigned short rec_len; + struct buffer_head * bh; +- struct ext3_dir_entry_2 * de, * de1; ++ struct ext3_dir_entry_2 *de; + struct super_block * sb; + int retval; ++#ifdef CONFIG_EXT3_INDEX ++ int dx_fallback=0; ++#endif ++ unsigned blocksize; ++ unsigned nlen, rlen; ++ u32 block, blocks; + + sb = dir->i_sb; +- +- if (!namelen) ++ blocksize = sb->s_blocksize; ++ if (!dentry->d_name.len) + return -EINVAL; +- bh = ext3_bread (handle, dir, 0, 0, &retval); ++#ifdef CONFIG_EXT3_INDEX ++ if (is_dx(dir)) { ++ retval = ext3_dx_add_entry(handle, dentry, inode); ++ if (!retval || (retval != ERR_BAD_DX_DIR)) ++ return retval; ++ EXT3_I(dir)->i_flags &= ~EXT3_INDEX_FL; ++ dx_fallback++; ++ ext3_mark_inode_dirty(handle, dir); ++ } ++#endif ++ blocks = dir->i_size >> sb->s_blocksize_bits; ++ for (block = 0, offset = 0; block < blocks; block++) { ++ bh = ext3_bread(handle, dir, block, 0, &retval); ++ if(!bh) ++ return retval; ++ retval = add_dirent_to_buf(handle, dentry, inode, 0, bh); ++ if (retval != -ENOSPC) ++ return retval; ++ ++#ifdef CONFIG_EXT3_INDEX ++ if (blocks == 1 && !dx_fallback && ++ EXT3_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_DIR_INDEX)) ++ return make_indexed_dir(handle, dentry, inode, bh); ++#endif ++ brelse(bh); ++ } ++ bh = ext3_append(handle, dir, &block, &retval); + if (!bh) + return retval; +- rec_len = EXT3_DIR_REC_LEN(namelen); +- offset = 0; + de = (struct ext3_dir_entry_2 *) bh->b_data; +- while (1) { +- if ((char *)de >= sb->s_blocksize + bh->b_data) { +- brelse (bh); +- bh = NULL; +- bh = ext3_bread (handle, dir, +- offset >> EXT3_BLOCK_SIZE_BITS(sb), 1, &retval); +- if (!bh) +- return retval; +- if (dir->i_size <= offset) { +- if (dir->i_size == 0) { +- brelse(bh); +- return -ENOENT; +- } ++ de->inode = 0; ++ de->rec_len = cpu_to_le16(rlen = blocksize); ++ nlen = 0; ++ return add_dirent_to_buf(handle, dentry, inode, de, bh); ++} + +- ext3_debug ("creating next block\n"); ++#ifdef CONFIG_EXT3_INDEX ++/* ++ * Returns 0 for success, or a negative error value ++ */ ++static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry, ++ struct inode *inode) ++{ ++ struct dx_frame frames[2], *frame; ++ struct dx_entry *entries, *at; ++ struct dx_hash_info hinfo; ++ struct buffer_head * bh; ++ struct inode *dir = dentry->d_parent->d_inode; ++ struct super_block * sb = dir->i_sb; ++ struct ext3_dir_entry_2 *de; ++ int err; + +- BUFFER_TRACE(bh, "get_write_access"); +- ext3_journal_get_write_access(handle, bh); +- de = (struct ext3_dir_entry_2 *) bh->b_data; +- de->inode = 0; +- de->rec_len = le16_to_cpu(sb->s_blocksize); +- dir->u.ext3_i.i_disksize = +- dir->i_size = offset + sb->s_blocksize; +- dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; +- ext3_mark_inode_dirty(handle, dir); +- } else { ++ frame = dx_probe(dentry, 0, &hinfo, frames, &err); ++ if (!frame) ++ return err; ++ entries = frame->entries; ++ at = frame->at; + +- ext3_debug ("skipping to next block\n"); ++ if (!(bh = ext3_bread(handle,dir, dx_get_block(frame->at), 0, &err))) ++ goto cleanup; + +- de = (struct ext3_dir_entry_2 *) bh->b_data; +- } +- } +- if (!ext3_check_dir_entry ("ext3_add_entry", dir, de, bh, +- offset)) { +- brelse (bh); +- return -ENOENT; +- } +- if (ext3_match (namelen, name, de)) { +- brelse (bh); +- return -EEXIST; ++ BUFFER_TRACE(bh, "get_write_access"); ++ err = ext3_journal_get_write_access(handle, bh); ++ if (err) ++ goto journal_error; ++ ++ err = add_dirent_to_buf(handle, dentry, inode, 0, bh); ++ if (err != -ENOSPC) { ++ bh = 0; ++ goto cleanup; ++ } ++ ++ /* Block full, should compress but for now just split */ ++ dxtrace(printk("using %u of %u node entries\n", ++ dx_get_count(entries), dx_get_limit(entries))); ++ /* Need to split index? */ ++ if (dx_get_count(entries) == dx_get_limit(entries)) { ++ u32 newblock; ++ unsigned icount = dx_get_count(entries); ++ int levels = frame - frames; ++ struct dx_entry *entries2; ++ struct dx_node *node2; ++ struct buffer_head *bh2; ++ ++ if (levels && (dx_get_count(frames->entries) == ++ dx_get_limit(frames->entries))) { ++ ext3_warning(sb, __FUNCTION__, ++ "Directory index full!\n"); ++ err = -ENOSPC; ++ goto cleanup; + } +- if ((le32_to_cpu(de->inode) == 0 && +- le16_to_cpu(de->rec_len) >= rec_len) || +- (le16_to_cpu(de->rec_len) >= +- EXT3_DIR_REC_LEN(de->name_len) + rec_len)) { +- BUFFER_TRACE(bh, "get_write_access"); +- ext3_journal_get_write_access(handle, bh); +- /* By now the buffer is marked for journaling */ +- offset += le16_to_cpu(de->rec_len); +- if (le32_to_cpu(de->inode)) { +- de1 = (struct ext3_dir_entry_2 *) ((char *) de + +- EXT3_DIR_REC_LEN(de->name_len)); +- de1->rec_len = +- cpu_to_le16(le16_to_cpu(de->rec_len) - +- EXT3_DIR_REC_LEN(de->name_len)); +- de->rec_len = cpu_to_le16( +- EXT3_DIR_REC_LEN(de->name_len)); +- de = de1; ++ bh2 = ext3_append (handle, dir, &newblock, &err); ++ if (!(bh2)) ++ goto cleanup; ++ node2 = (struct dx_node *)(bh2->b_data); ++ entries2 = node2->entries; ++ node2->fake.rec_len = cpu_to_le16(sb->s_blocksize); ++ node2->fake.inode = 0; ++ BUFFER_TRACE(frame->bh, "get_write_access"); ++ err = ext3_journal_get_write_access(handle, frame->bh); ++ if (err) ++ goto journal_error; ++ if (levels) { ++ unsigned icount1 = icount/2, icount2 = icount - icount1; ++ unsigned hash2 = dx_get_hash(entries + icount1); ++ dxtrace(printk("Split index %i/%i\n", icount1, icount2)); ++ ++ BUFFER_TRACE(frame->bh, "get_write_access"); /* index root */ ++ err = ext3_journal_get_write_access(handle, ++ frames[0].bh); ++ if (err) ++ goto journal_error; ++ ++ memcpy ((char *) entries2, (char *) (entries + icount1), ++ icount2 * sizeof(struct dx_entry)); ++ dx_set_count (entries, icount1); ++ dx_set_count (entries2, icount2); ++ dx_set_limit (entries2, dx_node_limit(dir)); ++ ++ /* Which index block gets the new entry? */ ++ if (at - entries >= icount1) { ++ frame->at = at = at - entries - icount1 + entries2; ++ frame->entries = entries = entries2; ++ swap(frame->bh, bh2); + } +- de->file_type = EXT3_FT_UNKNOWN; +- if (inode) { +- de->inode = cpu_to_le32(inode->i_ino); +- ext3_set_de_type(dir->i_sb, de, inode->i_mode); +- } else +- de->inode = 0; +- de->name_len = namelen; +- memcpy (de->name, name, namelen); +- /* +- * XXX shouldn't update any times until successful +- * completion of syscall, but too many callers depend +- * on this. +- * +- * XXX similarly, too many callers depend on +- * ext3_new_inode() setting the times, but error +- * recovery deletes the inode, so the worst that can +- * happen is that the times are slightly out of date +- * and/or different from the directory change time. +- */ +- dir->i_mtime = dir->i_ctime = CURRENT_TIME; +- dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; +- dir->i_version = ++event; +- ext3_mark_inode_dirty(handle, dir); +- BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata"); +- ext3_journal_dirty_metadata(handle, bh); +- brelse(bh); +- return 0; ++ dx_insert_block (frames + 0, hash2, newblock); ++ dxtrace(dx_show_index ("node", frames[1].entries)); ++ dxtrace(dx_show_index ("node", ++ ((struct dx_node *) bh2->b_data)->entries)); ++ err = ext3_journal_dirty_metadata(handle, bh2); ++ if (err) ++ goto journal_error; ++ brelse (bh2); ++ } else { ++ dxtrace(printk("Creating second level index...\n")); ++ memcpy((char *) entries2, (char *) entries, ++ icount * sizeof(struct dx_entry)); ++ dx_set_limit(entries2, dx_node_limit(dir)); ++ ++ /* Set up root */ ++ dx_set_count(entries, 1); ++ dx_set_block(entries + 0, newblock); ++ ((struct dx_root *) frames[0].bh->b_data)->info.indirect_levels = 1; ++ ++ /* Add new access path frame */ ++ frame = frames + 1; ++ frame->at = at = at - entries + entries2; ++ frame->entries = entries = entries2; ++ frame->bh = bh2; ++ err = ext3_journal_get_write_access(handle, ++ frame->bh); ++ if (err) ++ goto journal_error; + } +- offset += le16_to_cpu(de->rec_len); +- de = (struct ext3_dir_entry_2 *) +- ((char *) de + le16_to_cpu(de->rec_len)); ++ ext3_journal_dirty_metadata(handle, frames[0].bh); + } +- brelse (bh); +- return -ENOSPC; ++ de = do_split(handle, dir, &bh, frame, &hinfo, &err); ++ if (!de) ++ goto cleanup; ++ err = add_dirent_to_buf(handle, dentry, inode, de, bh); ++ bh = 0; ++ goto cleanup; ++ ++journal_error: ++ ext3_std_error(dir->i_sb, err); ++cleanup: ++ if (bh) ++ brelse(bh); ++ dx_release(frames); ++ return err; + } ++#endif + + /* + * ext3_delete_entry deletes a directory entry by merging it with the +@@ -455,9 +1547,11 @@ static int ext3_create (struct inode * d + struct inode * inode; + int err; + +- handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 3); +- if (IS_ERR(handle)) ++ handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + ++ EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3); ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(dir)) + handle->h_sync = 1; +@@ -481,9 +1575,11 @@ static int ext3_mknod (struct inode * di + struct inode *inode; + int err; + +- handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 3); +- if (IS_ERR(handle)) ++ handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + ++ EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3); ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(dir)) + handle->h_sync = 1; +@@ -509,9 +1605,11 @@ static int ext3_mkdir(struct inode * dir + if (dir->i_nlink >= EXT3_LINK_MAX) + return -EMLINK; + +- handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 3); +- if (IS_ERR(handle)) ++ handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + ++ EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3); ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(dir)) + handle->h_sync = 1; +@@ -523,7 +1621,7 @@ static int ext3_mkdir(struct inode * dir + + inode->i_op = &ext3_dir_inode_operations; + inode->i_fop = &ext3_dir_operations; +- inode->i_size = inode->u.ext3_i.i_disksize = inode->i_sb->s_blocksize; ++ inode->i_size = EXT3_I(inode)->i_disksize = inode->i_sb->s_blocksize; + inode->i_blocks = 0; + dir_block = ext3_bread (handle, inode, 0, 1, &err); + if (!dir_block) { +@@ -556,21 +1654,19 @@ static int ext3_mkdir(struct inode * dir + inode->i_mode |= S_ISGID; + ext3_mark_inode_dirty(handle, inode); + err = ext3_add_entry (handle, dentry, inode); +- if (err) +- goto out_no_entry; ++ if (err) { ++ inode->i_nlink = 0; ++ ext3_mark_inode_dirty(handle, inode); ++ iput (inode); ++ goto out_stop; ++ } + dir->i_nlink++; +- dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; ++ ext3_update_dx_flag(dir); + ext3_mark_inode_dirty(handle, dir); + d_instantiate(dentry, inode); + out_stop: + ext3_journal_stop(handle, dir); + return err; +- +-out_no_entry: +- inode->i_nlink = 0; +- ext3_mark_inode_dirty(handle, inode); +- iput (inode); +- goto out_stop; + } + + /* +@@ -657,7 +1753,7 @@ int ext3_orphan_add(handle_t *handle, st + int err = 0, rc; + + lock_super(sb); +- if (!list_empty(&inode->u.ext3_i.i_orphan)) ++ if (!list_empty(&EXT3_I(inode)->i_orphan)) + goto out_unlock; + + /* Orphan handling is only valid for files with data blocks +@@ -698,7 +1794,7 @@ int ext3_orphan_add(handle_t *handle, st + * This is safe: on error we're going to ignore the orphan list + * anyway on the next recovery. */ + if (!err) +- list_add(&inode->u.ext3_i.i_orphan, &EXT3_SB(sb)->s_orphan); ++ list_add(&EXT3_I(inode)->i_orphan, &EXT3_SB(sb)->s_orphan); + + jbd_debug(4, "superblock will point to %ld\n", inode->i_ino); + jbd_debug(4, "orphan inode %ld will point to %d\n", +@@ -716,25 +1812,26 @@ out_unlock: + int ext3_orphan_del(handle_t *handle, struct inode *inode) + { + struct list_head *prev; ++ struct ext3_inode_info *ei = EXT3_I(inode); + struct ext3_sb_info *sbi; + unsigned long ino_next; + struct ext3_iloc iloc; + int err = 0; + + lock_super(inode->i_sb); +- if (list_empty(&inode->u.ext3_i.i_orphan)) { ++ if (list_empty(&ei->i_orphan)) { + unlock_super(inode->i_sb); + return 0; + } + + ino_next = NEXT_ORPHAN(inode); +- prev = inode->u.ext3_i.i_orphan.prev; ++ prev = ei->i_orphan.prev; + sbi = EXT3_SB(inode->i_sb); + + jbd_debug(4, "remove inode %lu from orphan list\n", inode->i_ino); + +- list_del(&inode->u.ext3_i.i_orphan); +- INIT_LIST_HEAD(&inode->u.ext3_i.i_orphan); ++ list_del(&ei->i_orphan); ++ INIT_LIST_HEAD(&ei->i_orphan); + + /* If we're on an error path, we may not have a valid + * transaction handle with which to update the orphan list on +@@ -795,8 +1892,9 @@ static int ext3_rmdir (struct inode * di + handle_t *handle; + + handle = ext3_journal_start(dir, EXT3_DELETE_TRANS_BLOCKS); +- if (IS_ERR(handle)) ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + retval = -ENOENT; + bh = ext3_find_entry (dentry, &de); +@@ -834,7 +1932,7 @@ static int ext3_rmdir (struct inode * di + dir->i_nlink--; + inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME; + ext3_mark_inode_dirty(handle, inode); +- dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; ++ ext3_update_dx_flag(dir); + ext3_mark_inode_dirty(handle, dir); + + end_rmdir: +@@ -852,8 +1950,9 @@ static int ext3_unlink(struct inode * di + handle_t *handle; + + handle = ext3_journal_start(dir, EXT3_DELETE_TRANS_BLOCKS); +- if (IS_ERR(handle)) ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(dir)) + handle->h_sync = 1; +@@ -880,7 +1979,7 @@ static int ext3_unlink(struct inode * di + if (retval) + goto end_unlink; + dir->i_ctime = dir->i_mtime = CURRENT_TIME; +- dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; ++ ext3_update_dx_flag(dir); + ext3_mark_inode_dirty(handle, dir); + inode->i_nlink--; + if (!inode->i_nlink) +@@ -906,9 +2005,11 @@ static int ext3_symlink (struct inode * + if (l > dir->i_sb->s_blocksize) + return -ENAMETOOLONG; + +- handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 5); +- if (IS_ERR(handle)) ++ handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + ++ EXT3_INDEX_EXTRA_TRANS_BLOCKS + 5); ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(dir)) + handle->h_sync = 1; +@@ -918,7 +2019,7 @@ static int ext3_symlink (struct inode * + if (IS_ERR(inode)) + goto out_stop; + +- if (l > sizeof (inode->u.ext3_i.i_data)) { ++ if (l > sizeof (EXT3_I(inode)->i_data)) { + inode->i_op = &page_symlink_inode_operations; + inode->i_mapping->a_ops = &ext3_aops; + /* +@@ -927,24 +2028,22 @@ static int ext3_symlink (struct inode * + * i_size in generic_commit_write(). + */ + err = block_symlink(inode, symname, l); +- if (err) +- goto out_no_entry; ++ if (err) { ++ ext3_dec_count(handle, inode); ++ ext3_mark_inode_dirty(handle, inode); ++ iput (inode); ++ goto out_stop; ++ } + } else { + inode->i_op = &ext3_fast_symlink_inode_operations; +- memcpy((char*)&inode->u.ext3_i.i_data,symname,l); ++ memcpy((char*)&EXT3_I(inode)->i_data,symname,l); + inode->i_size = l-1; + } +- inode->u.ext3_i.i_disksize = inode->i_size; ++ EXT3_I(inode)->i_disksize = inode->i_size; + err = ext3_add_nondir(handle, dentry, inode); + out_stop: + ext3_journal_stop(handle, dir); + return err; +- +-out_no_entry: +- ext3_dec_count(handle, inode); +- ext3_mark_inode_dirty(handle, inode); +- iput (inode); +- goto out_stop; + } + + static int ext3_link (struct dentry * old_dentry, +@@ -957,12 +2056,15 @@ static int ext3_link (struct dentry * ol + if (S_ISDIR(inode->i_mode)) + return -EPERM; + +- if (inode->i_nlink >= EXT3_LINK_MAX) ++ if (inode->i_nlink >= EXT3_LINK_MAX) { + return -EMLINK; ++ } + +- handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS); +- if (IS_ERR(handle)) ++ handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + ++ EXT3_INDEX_EXTRA_TRANS_BLOCKS); ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(dir)) + handle->h_sync = 1; +@@ -995,9 +2097,11 @@ static int ext3_rename (struct inode * o + + old_bh = new_bh = dir_bh = NULL; + +- handle = ext3_journal_start(old_dir, 2 * EXT3_DATA_TRANS_BLOCKS + 2); +- if (IS_ERR(handle)) ++ handle = ext3_journal_start(old_dir, 2 * EXT3_DATA_TRANS_BLOCKS + ++ EXT3_INDEX_EXTRA_TRANS_BLOCKS + 2); ++ if (IS_ERR(handle)) { + return PTR_ERR(handle); ++ } + + if (IS_SYNC(old_dir) || IS_SYNC(new_dir)) + handle->h_sync = 1; +@@ -1070,14 +2174,33 @@ static int ext3_rename (struct inode * o + /* + * ok, that's it + */ +- ext3_delete_entry(handle, old_dir, old_de, old_bh); ++ retval = ext3_delete_entry(handle, old_dir, old_de, old_bh); ++ if (retval == -ENOENT) { ++ /* ++ * old_de could have moved out from under us. ++ */ ++ struct buffer_head *old_bh2; ++ struct ext3_dir_entry_2 *old_de2; ++ ++ old_bh2 = ext3_find_entry(old_dentry, &old_de2); ++ if (old_bh2) { ++ retval = ext3_delete_entry(handle, old_dir, ++ old_de2, old_bh2); ++ brelse(old_bh2); ++ } ++ } ++ if (retval) { ++ ext3_warning(old_dir->i_sb, "ext3_rename", ++ "Deleting old file (%lu), %d, error=%d", ++ old_dir->i_ino, old_dir->i_nlink, retval); ++ } + + if (new_inode) { + new_inode->i_nlink--; + new_inode->i_ctime = CURRENT_TIME; + } + old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME; +- old_dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; ++ ext3_update_dx_flag(old_dir); + if (dir_bh) { + BUFFER_TRACE(dir_bh, "get_write_access"); + ext3_journal_get_write_access(handle, dir_bh); +@@ -1089,7 +2212,7 @@ static int ext3_rename (struct inode * o + new_inode->i_nlink--; + } else { + new_dir->i_nlink++; +- new_dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; ++ ext3_update_dx_flag(new_dir); + ext3_mark_inode_dirty(handle, new_dir); + } + } +--- linux-2.4.22-ac1/fs/ext3/super.c~ext3-htree-2.4.22-rh 2003-09-25 14:39:01.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/super.c 2003-09-25 14:55:12.000000000 +0400 +@@ -714,6 +714,7 @@ static int ext3_setup_super(struct super + es->s_mtime = cpu_to_le32(CURRENT_TIME); + ext3_update_dynamic_rev(sb); + EXT3_SET_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER); ++ + ext3_commit_super (sb, es, 1); + if (test_opt (sb, DEBUG)) + printk (KERN_INFO +@@ -724,6 +725,7 @@ static int ext3_setup_super(struct super + EXT3_BLOCKS_PER_GROUP(sb), + EXT3_INODES_PER_GROUP(sb), + sbi->s_mount_opt); ++ + printk(KERN_INFO "EXT3 FS " EXT3FS_VERSION ", " EXT3FS_DATE " on %s, ", + bdevname(sb->s_dev)); + if (EXT3_SB(sb)->s_journal->j_inode == NULL) { +@@ -897,6 +899,7 @@ static loff_t ext3_max_size(int bits) + return res; + } + ++ + struct super_block * ext3_read_super (struct super_block * sb, void * data, + int silent) + { +@@ -1073,6 +1076,9 @@ struct super_block * ext3_read_super (st + sbi->s_mount_state = le16_to_cpu(es->s_state); + sbi->s_addr_per_block_bits = log2(EXT3_ADDR_PER_BLOCK(sb)); + sbi->s_desc_per_block_bits = log2(EXT3_DESC_PER_BLOCK(sb)); ++ for (i=0; i < 4; i++) ++ sbi->s_hash_seed[i] = le32_to_cpu(es->s_hash_seed[i]); ++ sbi->s_def_hash_version = es->s_def_hash_version; + + if (sbi->s_blocks_per_group > blocksize * 8) { + printk (KERN_ERR +@@ -1846,6 +1852,7 @@ static void __exit exit_ext3_fs(void) + unregister_filesystem(&ext3_fs_type); + } + ++EXPORT_SYMBOL(ext3_force_commit); + EXPORT_SYMBOL(ext3_bread); + + MODULE_AUTHOR("Remy Card, Stephen Tweedie, Andrew Morton, Andreas Dilger, Theodore Ts'o and others"); +--- linux-2.4.22-ac1/include/linux/ext3_fs.h~ext3-htree-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_fs.h 2003-09-25 14:58:30.000000000 +0400 +@@ -40,6 +40,11 @@ + #define EXT3FS_VERSION "2.4-0.9.19" + + /* ++ * Always enable hashed directories ++ */ ++#define CONFIG_EXT3_INDEX ++ ++/* + * Debug code + */ + #ifdef EXT3FS_DEBUG +@@ -440,8 +445,11 @@ struct ext3_super_block { + /*E0*/ __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 */ +- +-/*EC*/ __u32 s_reserved[197]; /* Padding to the end of the block */ ++ __u32 s_hash_seed[4]; /* HTREE hash seed */ ++ __u8 s_def_hash_version; /* Default hash version to use */ ++ __u8 s_reserved_char_pad; ++ __u16 s_reserved_word_pad; ++ __u32 s_reserved[192]; /* Padding to the end of the block */ + }; + + #ifdef __KERNEL__ +@@ -578,9 +586,46 @@ struct ext3_dir_entry_2 { + #define EXT3_DIR_ROUND (EXT3_DIR_PAD - 1) + #define EXT3_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT3_DIR_ROUND) & \ + ~EXT3_DIR_ROUND) ++/* ++ * Hash Tree Directory indexing ++ * (c) Daniel Phillips, 2001 ++ */ ++ ++#ifdef CONFIG_EXT3_INDEX ++ #define is_dx(dir) (EXT3_HAS_COMPAT_FEATURE(dir->i_sb, \ ++ EXT3_FEATURE_COMPAT_DIR_INDEX) && \ ++ (EXT3_I(dir)->i_flags & EXT3_INDEX_FL)) ++#define EXT3_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT3_LINK_MAX) ++#define EXT3_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1) ++#else ++ #define is_dx(dir) 0 ++#define EXT3_DIR_LINK_MAX(dir) ((dir)->i_nlink >= EXT3_LINK_MAX) ++#define EXT3_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2) ++#endif ++ ++/* Legal values for the dx_root hash_version field: */ ++ ++#define DX_HASH_LEGACY 0 ++#define DX_HASH_HALF_MD4 1 ++#define DX_HASH_TEA 2 ++ ++/* hash info structure used by the directory hash */ ++struct dx_hash_info ++{ ++ u32 hash; ++ u32 minor_hash; ++ int hash_version; ++ u32 *seed; ++}; + + #ifdef __KERNEL__ + /* ++ * Control parameters used by ext3_htree_next_block ++ */ ++#define HASH_NB_ALWAYS 1 ++ ++ ++/* + * Describe an inode's exact location on disk and in memory + */ + struct ext3_iloc +@@ -590,6 +635,27 @@ struct ext3_iloc + unsigned long block_group; + }; + ++ ++/* ++ * This structure is stuffed into the struct file's private_data field ++ * for directories. It is where we put information so that we can do ++ * readdir operations in hash tree order. ++ */ ++struct dir_private_info { ++ rb_root_t root; ++ rb_node_t *curr_node; ++ struct fname *extra_fname; ++ loff_t last_pos; ++ __u32 curr_hash; ++ __u32 curr_minor_hash; ++ __u32 next_hash; ++}; ++ ++/* ++ * Special error return code only used by dx_probe() and its callers. ++ */ ++#define ERR_BAD_DX_DIR -75000 ++ + /* + * Function prototypes + */ +@@ -617,11 +683,20 @@ extern struct ext3_group_desc * ext3_get + + /* dir.c */ + extern int ext3_check_dir_entry(const char *, struct inode *, +- struct ext3_dir_entry_2 *, struct buffer_head *, +- unsigned long); ++ struct ext3_dir_entry_2 *, ++ struct buffer_head *, unsigned long); ++extern int ext3_htree_store_dirent(struct file *dir_file, __u32 hash, ++ __u32 minor_hash, ++ struct ext3_dir_entry_2 *dirent); ++extern void ext3_htree_free_dir_info(struct dir_private_info *p); ++ + /* fsync.c */ + extern int ext3_sync_file (struct file *, struct dentry *, int); + ++/* hash.c */ ++extern int ext3fs_dirhash(const char *name, int len, struct ++ dx_hash_info *hinfo); ++ + /* ialloc.c */ + extern struct inode * ext3_new_inode (handle_t *, const struct inode *, int); + extern void ext3_free_inode (handle_t *, struct inode *); +@@ -655,6 +730,8 @@ extern int ext3_ioctl (struct inode *, s + /* namei.c */ + extern int ext3_orphan_add(handle_t *, struct inode *); + extern int ext3_orphan_del(handle_t *, struct inode *); ++extern int ext3_htree_fill_tree(struct file *dir_file, __u32 start_hash, ++ __u32 start_minor_hash, __u32 *next_hash); + + /* super.c */ + extern void ext3_error (struct super_block *, const char *, const char *, ...) +--- linux-2.4.22-ac1/include/linux/ext3_fs_sb.h~ext3-htree-2.4.22-rh 2003-09-25 14:16:34.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_fs_sb.h 2003-09-25 14:55:12.000000000 +0400 +@@ -62,6 +62,8 @@ struct ext3_sb_info { + int s_inode_size; + int s_first_ino; + u32 s_next_generation; ++ u32 s_hash_seed[4]; ++ int s_def_hash_version; + + unsigned long s_dir_count; + u8 *s_debts; +--- linux-2.4.22-ac1/include/linux/ext3_jbd.h~ext3-htree-2.4.22-rh 2003-06-13 18:51:38.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_jbd.h 2003-09-25 14:55:12.000000000 +0400 +@@ -63,6 +63,8 @@ extern int ext3_writepage_trans_blocks(s + + #define EXT3_RESERVE_TRANS_BLOCKS 12U + ++#define EXT3_INDEX_EXTRA_TRANS_BLOCKS 8 ++ + int + ext3_mark_iloc_dirty(handle_t *handle, + struct inode *inode, +--- linux-2.4.22-ac1/include/linux/rbtree.h~ext3-htree-2.4.22-rh 2001-11-22 22:46:18.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/include/linux/rbtree.h 2003-09-25 14:55:12.000000000 +0400 +@@ -120,6 +120,8 @@ rb_root_t; + + extern void rb_insert_color(rb_node_t *, rb_root_t *); + extern void rb_erase(rb_node_t *, rb_root_t *); ++extern rb_node_t *rb_get_first(rb_root_t *root); ++extern rb_node_t *rb_get_next(rb_node_t *n); + + static inline void rb_link_node(rb_node_t * node, rb_node_t * parent, rb_node_t ** rb_link) + { +--- linux-2.4.22-ac1/lib/rbtree.c~ext3-htree-2.4.22-rh 2002-08-03 04:39:46.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/lib/rbtree.c 2003-09-25 14:55:12.000000000 +0400 +@@ -17,6 +17,8 @@ + Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + + linux/lib/rbtree.c ++ ++ rb_get_first and rb_get_next written by Theodore Ts'o, 9/8/2002 + */ + + #include +@@ -294,3 +296,43 @@ void rb_erase(rb_node_t * node, rb_root_ + __rb_erase_color(child, parent, root); + } + EXPORT_SYMBOL(rb_erase); ++ ++/* ++ * This function returns the first node (in sort order) of the tree. ++ */ ++rb_node_t *rb_get_first(rb_root_t *root) ++{ ++ rb_node_t *n; ++ ++ n = root->rb_node; ++ if (!n) ++ return 0; ++ while (n->rb_left) ++ n = n->rb_left; ++ return n; ++} ++EXPORT_SYMBOL(rb_get_first); ++ ++/* ++ * Given a node, this function will return the next node in the tree. ++ */ ++rb_node_t *rb_get_next(rb_node_t *n) ++{ ++ rb_node_t *parent; ++ ++ if (n->rb_right) { ++ n = n->rb_right; ++ while (n->rb_left) ++ n = n->rb_left; ++ return n; ++ } else { ++ while ((parent = n->rb_parent)) { ++ if (n == parent->rb_left) ++ return parent; ++ n = parent; ++ } ++ return 0; ++ } ++} ++EXPORT_SYMBOL(rb_get_next); ++ + +_ diff --git a/lustre/kernel_patches/patches/ext3-orphan_lock-2.4.22-rh.patch b/lustre/kernel_patches/patches/ext3-orphan_lock-2.4.22-rh.patch new file mode 100644 index 0000000..46fc3d5 --- /dev/null +++ b/lustre/kernel_patches/patches/ext3-orphan_lock-2.4.22-rh.patch @@ -0,0 +1,158 @@ + fs/ext3/namei.c | 15 +++++++-------- + fs/ext3/namei.c.orig | 21 +++++++++++++++------ + fs/ext3/super.c | 1 + + include/linux/ext3_fs_sb.h | 1 + + include/linux/ext3_fs_sb.h.orig | 2 ++ + 5 files changed, 26 insertions(+), 14 deletions(-) + +--- linux-2.4.22-ac1/fs/ext3/namei.c~ext3-orphan_lock-2.4.22-rh 2003-09-26 00:24:09.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/namei.c 2003-09-26 00:26:36.000000000 +0400 +@@ -1748,8 +1748,8 @@ int ext3_orphan_add(handle_t *handle, st + struct super_block *sb = inode->i_sb; + struct ext3_iloc iloc; + int err = 0, rc; +- +- lock_super(sb); ++ ++ down(&EXT3_SB(sb)->s_orphan_lock); + if (!list_empty(&EXT3_I(inode)->i_orphan)) + goto out_unlock; + +@@ -1797,7 +1797,7 @@ int ext3_orphan_add(handle_t *handle, st + jbd_debug(4, "orphan inode %ld will point to %d\n", + inode->i_ino, NEXT_ORPHAN(inode)); + out_unlock: +- unlock_super(sb); ++ up(&EXT3_SB(sb)->s_orphan_lock); + ext3_std_error(inode->i_sb, err); + return err; + } +@@ -1810,20 +1810,19 @@ int ext3_orphan_del(handle_t *handle, st + { + struct list_head *prev; + struct ext3_inode_info *ei = EXT3_I(inode); +- struct ext3_sb_info *sbi; ++ struct ext3_sb_info *sbi = EXT3_SB(inode->i_sb); + unsigned long ino_next; + struct ext3_iloc iloc; + int err = 0; + +- lock_super(inode->i_sb); ++ down(&sbi->s_orphan_lock); + if (list_empty(&ei->i_orphan)) { +- unlock_super(inode->i_sb); ++ up(&sbi->s_orphan_lock); + return 0; + } + + ino_next = NEXT_ORPHAN(inode); + prev = ei->i_orphan.prev; +- sbi = EXT3_SB(inode->i_sb); + + jbd_debug(4, "remove inode %lu from orphan list\n", inode->i_ino); + +@@ -1872,7 +1871,7 @@ int ext3_orphan_del(handle_t *handle, st + out_err: + ext3_std_error(inode->i_sb, err); + out: +- unlock_super(inode->i_sb); ++ up(&sbi->s_orphan_lock); + return err; + + out_brelse: +--- linux-2.4.22-ac1/fs/ext3/namei.c.orig~ext3-orphan_lock-2.4.22-rh 2003-09-25 14:58:37.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/namei.c.orig 2003-09-26 00:24:09.000000000 +0400 +@@ -29,6 +29,7 @@ + #include + #include + #include ++#include + #include + #include + #include +@@ -1614,7 +1615,7 @@ static int ext3_mkdir(struct inode * dir + if (IS_SYNC(dir)) + handle->h_sync = 1; + +- inode = ext3_new_inode (handle, dir, S_IFDIR); ++ inode = ext3_new_inode (handle, dir, S_IFDIR | mode); + err = PTR_ERR(inode); + if (IS_ERR(inode)) + goto out_stop; +@@ -1622,7 +1623,6 @@ static int ext3_mkdir(struct inode * dir + inode->i_op = &ext3_dir_inode_operations; + inode->i_fop = &ext3_dir_operations; + inode->i_size = EXT3_I(inode)->i_disksize = inode->i_sb->s_blocksize; +- inode->i_blocks = 0; + dir_block = ext3_bread (handle, inode, 0, 1, &err); + if (!dir_block) { + inode->i_nlink--; /* is this nlink == 0? */ +@@ -1649,9 +1649,6 @@ static int ext3_mkdir(struct inode * dir + BUFFER_TRACE(dir_block, "call ext3_journal_dirty_metadata"); + ext3_journal_dirty_metadata(handle, dir_block); + brelse (dir_block); +- inode->i_mode = S_IFDIR | mode; +- if (dir->i_mode & S_ISGID) +- inode->i_mode |= S_ISGID; + ext3_mark_inode_dirty(handle, inode); + err = ext3_add_entry (handle, dentry, inode); + if (err) { +@@ -2020,7 +2017,7 @@ static int ext3_symlink (struct inode * + goto out_stop; + + if (l > sizeof (EXT3_I(inode)->i_data)) { +- inode->i_op = &page_symlink_inode_operations; ++ inode->i_op = &ext3_symlink_inode_operations; + inode->i_mapping->a_ops = &ext3_aops; + /* + * block_symlink() calls back into ext3_prepare/commit_write. +@@ -2245,4 +2242,16 @@ struct inode_operations ext3_dir_inode_o + rmdir: ext3_rmdir, /* BKL held */ + mknod: ext3_mknod, /* BKL held */ + rename: ext3_rename, /* BKL held */ ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ + }; ++ ++struct inode_operations ext3_special_inode_operations = { ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ ++}; ++ +--- linux-2.4.22-ac1/fs/ext3/super.c~ext3-orphan_lock-2.4.22-rh 2003-09-26 00:24:09.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/super.c 2003-09-26 00:25:22.000000000 +0400 +@@ -1164,6 +1164,7 @@ struct super_block * ext3_read_super (st + sb->s_op = &ext3_sops; + sb->dq_op = &ext3_qops; + INIT_LIST_HEAD(&sbi->s_orphan); /* unlinked but open files */ ++ sema_init(&sbi->s_orphan_lock, 1); + + sb->s_root = 0; + +--- linux-2.4.22-ac1/include/linux/ext3_fs_sb.h~ext3-orphan_lock-2.4.22-rh 2003-09-26 00:24:08.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_fs_sb.h 2003-09-26 00:25:22.000000000 +0400 +@@ -72,6 +72,7 @@ struct ext3_sb_info { + struct inode * s_journal_inode; + struct journal_s * s_journal; + struct list_head s_orphan; ++ struct semaphore s_orphan_lock; + unsigned long s_commit_interval; + struct block_device *journal_bdev; + #ifdef CONFIG_JBD_DEBUG +--- linux-2.4.22-ac1/include/linux/ext3_fs_sb.h.orig~ext3-orphan_lock-2.4.22-rh 2003-09-25 14:16:34.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_fs_sb.h.orig 2003-09-26 00:24:08.000000000 +0400 +@@ -62,6 +62,8 @@ struct ext3_sb_info { + int s_inode_size; + int s_first_ino; + u32 s_next_generation; ++ u32 s_hash_seed[4]; ++ int s_def_hash_version; + + unsigned long s_dir_count; + u8 *s_debts; + +_ diff --git a/lustre/kernel_patches/patches/extN-wantedi-2.4.22-rh.patch b/lustre/kernel_patches/patches/extN-wantedi-2.4.22-rh.patch new file mode 100644 index 0000000..5ec4dab --- /dev/null +++ b/lustre/kernel_patches/patches/extN-wantedi-2.4.22-rh.patch @@ -0,0 +1,217 @@ + fs/ext3/ialloc.c | 41 +++++++++++++++++++++++++++++++++++++++-- + fs/ext3/inode.c | 2 +- + fs/ext3/ioctl.c | 25 +++++++++++++++++++++++++ + fs/ext3/namei.c | 21 +++++++++++++++++---- + include/linux/dcache.h | 5 +++++ + include/linux/ext3_fs.h | 5 ++++- + 6 files changed, 91 insertions(+), 8 deletions(-) + +--- linux-2.4.22-ac1/fs/ext3/ialloc.c~extN-wantedi-2.4.22-rh 2003-09-26 00:57:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/ialloc.c 2003-09-26 01:00:17.000000000 +0400 +@@ -524,7 +524,8 @@ static int find_group_other(struct super + * group to find a free inode. + */ + struct inode * ext3_new_inode (handle_t *handle, +- const struct inode * dir, int mode) ++ const struct inode * dir, int mode, ++ unsigned long goal) + { + struct super_block * sb; + struct buffer_head * bh; +@@ -549,7 +550,41 @@ struct inode * ext3_new_inode (handle_t + init_rwsem(&inode->u.ext3_i.truncate_sem); + + lock_super (sb); +- es = sb->u.ext3_sb.s_es; ++ es = EXT3_SB(sb)->s_es; ++ ++ if (goal) { ++ group = (goal - 1) / EXT3_INODES_PER_GROUP(sb); ++ ino = (goal - 1) % EXT3_INODES_PER_GROUP(sb); ++ gdp = ext3_get_group_desc(sb, group, &bh2); ++ ++ bitmap_nr = load_inode_bitmap (sb, group); ++ if (bitmap_nr < 0) { ++ err = bitmap_nr; ++ goto fail; ++ } ++ ++ bh = EXT3_SB(sb)->s_inode_bitmap[bitmap_nr]; ++ ++ BUFFER_TRACE(bh, "get_write_access"); ++ err = ext3_journal_get_write_access(handle, bh); ++ if (err) goto fail; ++ ++ if (ext3_set_bit(ino, bh->b_data)) { ++ printk(KERN_ERR "goal inode %lu unavailable\n", goal); ++ /* Oh well, we tried. */ ++ goto repeat; ++ } ++ ++ BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata"); ++ err = ext3_journal_dirty_metadata(handle, bh); ++ if (err) goto fail; ++ ++ /* We've shortcircuited the allocation system successfully, ++ * now finish filling in the inode. ++ */ ++ goto have_bit_and_group; ++ } ++ + repeat: + if (S_ISDIR(mode)) { + if (test_opt (sb, OLDALLOC)) +@@ -606,6 +641,8 @@ repeat: + } + goto repeat; + } ++ ++have_bit_and_group: + ino += group * EXT3_INODES_PER_GROUP(sb) + 1; + if (ino < EXT3_FIRST_INO(sb) || ino > le32_to_cpu(es->s_inodes_count)) { + ext3_error (sb, "ext3_new_inode", +--- linux-2.4.22-ac1/fs/ext3/inode.c~extN-wantedi-2.4.22-rh 2003-09-26 00:57:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/inode.c 2003-09-26 00:57:29.000000000 +0400 +@@ -2614,7 +2614,7 @@ void ext3_truncate_thread(struct inode * + if (IS_ERR(handle)) + goto out_truncate; + +- new_inode = ext3_new_inode(handle, old_inode, old_inode->i_mode); ++ new_inode = ext3_new_inode(handle, old_inode, old_inode->i_mode, 0); + if (IS_ERR(new_inode)) { + ext3_debug("truncate inode %lu directly (no new inodes)\n", + old_inode->i_ino); +--- linux-2.4.22-ac1/fs/ext3/ioctl.c~extN-wantedi-2.4.22-rh 2003-09-25 14:16:23.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/ioctl.c 2003-09-26 00:57:29.000000000 +0400 +@@ -23,6 +23,31 @@ int ext3_ioctl (struct inode * inode, st + ext3_debug ("cmd = %u, arg = %lu\n", cmd, arg); + + switch (cmd) { ++ case EXT3_IOC_CREATE_INUM: { ++ char name[32]; ++ struct dentry *dchild, *dparent; ++ int rc = 0; ++ ++ dparent = list_entry(inode->i_dentry.next, struct dentry, ++ d_alias); ++ snprintf(name, sizeof name, "%lu", arg); ++ dchild = lookup_one_len(name, dparent, strlen(name)); ++ if (dchild->d_inode) { ++ printk(KERN_ERR "%*s/%lu already exists (ino %lu)\n", ++ dparent->d_name.len, dparent->d_name.name, arg, ++ dchild->d_inode->i_ino); ++ rc = -EEXIST; ++ } else { ++ dchild->d_fsdata = (void *)arg; ++ rc = vfs_create(inode, dchild, 0644); ++ if (rc) ++ printk(KERN_ERR "vfs_create: %d\n", rc); ++ else if (dchild->d_inode->i_ino != arg) ++ rc = -EEXIST; ++ } ++ dput(dchild); ++ return rc; ++ } + case EXT3_IOC_GETFLAGS: + flags = inode->u.ext3_i.i_flags & EXT3_FL_USER_VISIBLE; + return put_user(flags, (int *) arg); +--- linux-2.4.22-ac1/fs/ext3/namei.c~extN-wantedi-2.4.22-rh 2003-09-26 00:57:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/namei.c 2003-09-26 00:57:29.000000000 +0400 +@@ -1534,6 +1534,19 @@ static int ext3_add_nondir(handle_t *han + return err; + } + ++static struct inode * ext3_new_inode_wantedi(handle_t *handle, struct inode *dir, ++ int mode, struct dentry *dentry) ++{ ++ unsigned long inum = 0; ++ ++ if (dentry->d_fsdata != NULL) { ++ struct dentry_params *param = ++ (struct dentry_params *) dentry->d_fsdata; ++ inum = param->p_inum; ++ } ++ return ext3_new_inode(handle, dir, mode, inum); ++} ++ + /* + * By the time this is called, we already have created + * the directory cache entry for the new file, but it +@@ -1557,7 +1570,7 @@ static int ext3_create (struct inode * d + if (IS_SYNC(dir)) + handle->h_sync = 1; + +- inode = ext3_new_inode (handle, dir, mode); ++ inode = ext3_new_inode_wantedi (handle, dir, mode, dentry); + err = PTR_ERR(inode); + if (!IS_ERR(inode)) { + inode->i_op = &ext3_file_inode_operations; +@@ -1585,7 +1598,7 @@ static int ext3_mknod (struct inode * di + if (IS_SYNC(dir)) + handle->h_sync = 1; + +- inode = ext3_new_inode (handle, dir, mode); ++ inode = ext3_new_inode_wantedi (handle, dir, mode, dentry); + err = PTR_ERR(inode); + if (!IS_ERR(inode)) { + init_special_inode(inode, mode, rdev); +@@ -1615,7 +1628,7 @@ static int ext3_mkdir(struct inode * dir + if (IS_SYNC(dir)) + handle->h_sync = 1; + +- inode = ext3_new_inode (handle, dir, S_IFDIR | mode); ++ inode = ext3_new_inode_wantedi (handle, dir, S_IFDIR | mode, dentry); + err = PTR_ERR(inode); + if (IS_ERR(inode)) + goto out_stop; +@@ -2010,7 +2023,7 @@ static int ext3_symlink (struct inode * + if (IS_SYNC(dir)) + handle->h_sync = 1; + +- inode = ext3_new_inode (handle, dir, S_IFLNK|S_IRWXUGO); ++ inode = ext3_new_inode_wantedi (handle, dir, S_IFLNK|S_IRWXUGO, dentry); + err = PTR_ERR(inode); + if (IS_ERR(inode)) + goto out_stop; +--- linux-2.4.22-ac1/include/linux/dcache.h~extN-wantedi-2.4.22-rh 2003-09-26 00:57:27.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/dcache.h 2003-09-26 00:57:29.000000000 +0400 +@@ -63,6 +63,11 @@ static inline void intent_init(struct lo + + #define IS_ROOT(x) ((x) == (x)->d_parent) + ++struct dentry_params { ++ unsigned long p_inum; ++ void *p_ptr; ++}; ++ + /* + * "quick string" -- eases parameter passing, but more importantly + * saves "metadata" about the string (ie length and the hash). +--- linux-2.4.22-ac1/include/linux/ext3_fs.h~extN-wantedi-2.4.22-rh 2003-09-26 00:57:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_fs.h 2003-09-26 00:57:29.000000000 +0400 +@@ -203,6 +203,7 @@ struct ext3_group_desc + #define EXT3_IOC_SETFLAGS _IOW('f', 2, long) + #define EXT3_IOC_GETVERSION _IOR('f', 3, long) + #define EXT3_IOC_SETVERSION _IOW('f', 4, long) ++/* EXT3_IOC_CREATE_INUM at bottom of file (visible to kernel and user). */ + #define EXT3_IOC_GETVERSION_OLD _IOR('v', 1, long) + #define EXT3_IOC_SETVERSION_OLD _IOW('v', 2, long) + #ifdef CONFIG_JBD_DEBUG +@@ -676,7 +677,8 @@ extern int ext3fs_dirhash(const char *na + dx_hash_info *hinfo); + + /* ialloc.c */ +-extern struct inode * ext3_new_inode (handle_t *, const struct inode *, int); ++extern struct inode * ext3_new_inode (handle_t *, const struct inode *, int, ++ unsigned long); + extern void ext3_free_inode (handle_t *, struct inode *); + extern struct inode * ext3_orphan_get (struct super_block *, unsigned long); + extern unsigned long ext3_count_free_inodes (struct super_block *); +@@ -769,4 +771,5 @@ extern struct inode_operations ext3_fast + + #endif /* __KERNEL__ */ + ++#define EXT3_IOC_CREATE_INUM _IOW('f', 5, long) + #endif /* _LINUX_EXT3_FS_H */ + +_ diff --git a/lustre/kernel_patches/patches/iod-stock-exports-2.4.22-rh.patch b/lustre/kernel_patches/patches/iod-stock-exports-2.4.22-rh.patch new file mode 100644 index 0000000..8d8790e --- /dev/null +++ b/lustre/kernel_patches/patches/iod-stock-exports-2.4.22-rh.patch @@ -0,0 +1,48 @@ + fs/Makefile | 2 +- + fs/inode.c | 4 +++- + mm/page_alloc.c | 1 + + 3 files changed, 5 insertions(+), 2 deletions(-) + +--- linux-2.4.22-ac1/fs/inode.c~iod-stock-exports-2.4.22-rh 2003-09-25 14:45:32.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/inode.c 2003-09-25 14:49:41.000000000 +0400 +@@ -5,6 +5,7 @@ + */ + + #include ++#include + #include + #include + #include +@@ -68,7 +69,8 @@ static LIST_HEAD(anon_hash_chain); /* fo + * NOTE! You also have to own the lock if you change + * the i_state of an inode while it is in use.. + */ +-static spinlock_t inode_lock = SPIN_LOCK_UNLOCKED; ++spinlock_t inode_lock = SPIN_LOCK_UNLOCKED; ++EXPORT_SYMBOL(inode_lock); + + /* + * Statistics gathering.. +--- linux-2.4.22-ac1/fs/Makefile~iod-stock-exports-2.4.22-rh 2003-09-25 14:16:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/Makefile 2003-09-25 14:50:00.000000000 +0400 +@@ -7,7 +7,7 @@ + + O_TARGET := fs.o + +-export-objs := filesystems.o open.o dcache.o buffer.o dquot.o dcookies.o ++export-objs := filesystems.o open.o dcache.o buffer.o dquot.o dcookies.o inode.o + mod-subdirs := nls xfs + + obj-y := open.o read_write.o devices.o file_table.o buffer.o \ +--- linux-2.4.22-ac1/mm/page_alloc.c~iod-stock-exports-2.4.22-rh 2003-09-25 14:16:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/mm/page_alloc.c 2003-09-25 14:49:41.000000000 +0400 +@@ -28,6 +28,7 @@ int nr_inactive_pages; + LIST_HEAD(inactive_list); + LIST_HEAD(active_list); + pg_data_t *pgdat_list; ++EXPORT_SYMBOL(pgdat_list); + + /* + * + +_ diff --git a/lustre/kernel_patches/patches/socket-exports-2.4.22-rh.patch b/lustre/kernel_patches/patches/socket-exports-2.4.22-rh.patch new file mode 100644 index 0000000..08ca1b2 --- /dev/null +++ b/lustre/kernel_patches/patches/socket-exports-2.4.22-rh.patch @@ -0,0 +1,41 @@ + include/linux/socket.h | 4 ++++ + net/netsyms.c | 1 + + net/socket.c | 2 +- + 3 files changed, 6 insertions(+), 1 deletion(-) + +--- linux-2.4.22-ac1/include/linux/socket.h~socket-exports-2.4.22-rh 2003-06-13 18:51:39.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/socket.h 2003-09-26 00:49:43.000000000 +0400 +@@ -275,6 +275,10 @@ extern void memcpy_tokerneliovec(struct + extern int move_addr_to_user(void *kaddr, int klen, void *uaddr, int *ulen); + extern int move_addr_to_kernel(void *uaddr, int ulen, void *kaddr); + extern int put_cmsg(struct msghdr*, int level, int type, int len, void *data); ++struct socket; ++extern int sock_map_fd(struct socket *sock); ++extern struct socket *sockfd_lookup(int fd, int *err); ++ + #endif + #endif /* not kernel and not glibc */ + #endif /* _LINUX_SOCKET_H */ +--- linux-2.4.22-ac1/net/netsyms.c~socket-exports-2.4.22-rh 2003-09-26 00:49:19.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/net/netsyms.c 2003-09-26 00:50:20.000000000 +0400 +@@ -163,6 +163,7 @@ EXPORT_SYMBOL(put_cmsg); + EXPORT_SYMBOL(sock_kmalloc); + EXPORT_SYMBOL(sock_kfree_s); + EXPORT_SYMBOL(sockfd_lookup); ++EXPORT_SYMBOL(sock_map_fd); + + #ifdef CONFIG_FILTER + EXPORT_SYMBOL(sk_run_filter); +--- linux-2.4.22-ac1/net/socket.c~socket-exports-2.4.22-rh 2003-08-25 15:44:44.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/net/socket.c 2003-09-26 00:49:43.000000000 +0400 +@@ -325,7 +325,7 @@ static struct dentry_operations sockfs_d + * but we take care of internal coherence yet. + */ + +-static int sock_map_fd(struct socket *sock) ++int sock_map_fd(struct socket *sock) + { + int fd; + struct qstr this; + +_ diff --git a/lustre/kernel_patches/patches/tcp-zero-copy-2.4.22-rh.patch b/lustre/kernel_patches/patches/tcp-zero-copy-2.4.22-rh.patch new file mode 100644 index 0000000..65baef5 --- /dev/null +++ b/lustre/kernel_patches/patches/tcp-zero-copy-2.4.22-rh.patch @@ -0,0 +1,459 @@ + include/linux/skbuff.h | 30 +++++ + include/net/tcp.h | 5 + net/core/skbuff.c | 25 ++++ + net/ipv4/tcp.c | 252 ++++++++++++++++++++++++++++++++++++++++++++++++- + net/netsyms.c | 2 + 5 files changed, 311 insertions(+), 3 deletions(-) + +--- linux-2.4.22-ac1/include/linux/skbuff.h~tcp-zero-copy-2.4.22-rh 2003-08-25 15:44:44.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/skbuff.h 2003-09-26 00:38:48.000000000 +0400 +@@ -116,6 +116,30 @@ struct skb_frag_struct + __u16 size; + }; + ++/* Support for callback when skb data has been released */ ++typedef struct zccd /* Zero Copy Callback Descriptor */ ++{ /* (embed as first member of custom struct) */ ++ atomic_t zccd_count; /* reference count */ ++ void (*zccd_destructor)(struct zccd *); /* callback when refcount reaches zero */ ++} zccd_t; ++ ++static inline void zccd_init (zccd_t *d, void (*callback)(zccd_t *)) ++{ ++ atomic_set (&d->zccd_count, 1); ++ d->zccd_destructor = callback; ++} ++ ++static inline void zccd_get (zccd_t *d) /* take a reference */ ++{ ++ atomic_inc (&d->zccd_count); ++} ++ ++static inline void zccd_put (zccd_t *d) /* release a reference */ ++{ ++ if (atomic_dec_and_test (&d->zccd_count)) ++ (d->zccd_destructor)(d); ++} ++ + /* This data is invariant across clones and lives at + * the end of the header data, ie. at skb->end. + */ +@@ -123,6 +147,12 @@ struct skb_shared_info { + atomic_t dataref; + unsigned int nr_frags; + struct sk_buff *frag_list; ++ zccd_t *zccd; /* zero copy descriptor */ ++ zccd_t *zccd2; /* 2nd zero copy descriptor */ ++ /* NB we expect zero-copy data to be at least 1 packet, so ++ * having 2 zccds means we don't unneccessarily split the packet ++ * where consecutive zero-copy sends abutt. ++ */ + skb_frag_t frags[MAX_SKB_FRAGS]; + }; + +--- linux-2.4.22-ac1/include/net/tcp.h~tcp-zero-copy-2.4.22-rh 2003-08-25 15:44:44.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/net/tcp.h 2003-09-26 00:38:48.000000000 +0400 +@@ -643,6 +643,8 @@ extern int tcp_v4_tw_remember_stam + + extern int tcp_sendmsg(struct sock *sk, struct msghdr *msg, int size); + extern ssize_t tcp_sendpage(struct socket *sock, struct page *page, int offset, size_t size, int flags); ++extern ssize_t tcp_sendpage_zccd(struct socket *sock, struct page *page, int offset, size_t size, ++ int flags, zccd_t *zccd); + + extern int tcp_ioctl(struct sock *sk, + int cmd, +@@ -737,6 +739,9 @@ extern int tcp_recvmsg(struct sock *sk + struct msghdr *msg, + int len, int nonblock, + int flags, int *addr_len); ++extern int tcp_recvpackets(struct sock *sk, ++ struct sk_buff_head *packets, ++ int len, int nonblock); + + extern int tcp_listen_start(struct sock *sk); + +--- linux-2.4.22-ac1/net/core/skbuff.c~tcp-zero-copy-2.4.22-rh 2003-08-25 15:44:44.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/net/core/skbuff.c 2003-09-26 00:38:48.000000000 +0400 +@@ -208,6 +208,8 @@ struct sk_buff *alloc_skb(unsigned int s + atomic_set(&(skb_shinfo(skb)->dataref), 1); + skb_shinfo(skb)->nr_frags = 0; + skb_shinfo(skb)->frag_list = NULL; ++ skb_shinfo(skb)->zccd = NULL; /* skbuffs kick off with NO user zero copy descriptors */ ++ skb_shinfo(skb)->zccd2 = NULL; + return skb; + + nodata: +@@ -277,6 +279,10 @@ static void skb_release_data(struct sk_b + { + if (!skb->cloned || + atomic_dec_and_test(&(skb_shinfo(skb)->dataref))) { ++ if (skb_shinfo(skb)->zccd != NULL) /* zero copy callback descriptor? */ ++ zccd_put (skb_shinfo(skb)->zccd); /* release hold */ ++ if (skb_shinfo(skb)->zccd2 != NULL) /* 2nd zero copy callback descriptor? */ ++ zccd_put (skb_shinfo(skb)->zccd2); /* release hold */ + if (skb_shinfo(skb)->nr_frags) { + int i; + for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) +@@ -535,6 +541,8 @@ int skb_linearize(struct sk_buff *skb, i + atomic_set(&(skb_shinfo(skb)->dataref), 1); + skb_shinfo(skb)->nr_frags = 0; + skb_shinfo(skb)->frag_list = NULL; ++ skb_shinfo(skb)->zccd = NULL; /* copied data => no user zero copy descriptor */ ++ skb_shinfo(skb)->zccd2 = NULL; + + /* We are no longer a clone, even if we were. */ + skb->cloned = 0; +@@ -581,6 +589,14 @@ struct sk_buff *pskb_copy(struct sk_buff + n->data_len = skb->data_len; + n->len = skb->len; + ++ if (skb_shinfo(skb)->zccd != NULL) /* user zero copy descriptor? */ ++ zccd_get (skb_shinfo(skb)->zccd); /* 1 more ref (pages are shared) */ ++ skb_shinfo(n)->zccd = skb_shinfo(skb)->zccd; ++ ++ if (skb_shinfo(skb)->zccd2 != NULL) /* 2nd user zero copy descriptor? */ ++ zccd_get (skb_shinfo(skb)->zccd2); /* 1 more ref (pages are shared) */ ++ skb_shinfo(n)->zccd2 = skb_shinfo(skb)->zccd2; ++ + if (skb_shinfo(skb)->nr_frags) { + int i; + +@@ -623,6 +639,8 @@ int pskb_expand_head(struct sk_buff *skb + u8 *data; + int size = nhead + (skb->end - skb->head) + ntail; + long off; ++ zccd_t *zccd = skb_shinfo(skb)->zccd; /* stash user zero copy descriptor */ ++ zccd_t *zccd2 = skb_shinfo(skb)->zccd2; /* stash 2nd user zero copy descriptor */ + + if (skb_shared(skb)) + BUG(); +@@ -644,6 +662,11 @@ int pskb_expand_head(struct sk_buff *skb + if (skb_shinfo(skb)->frag_list) + skb_clone_fraglist(skb); + ++ if (zccd != NULL) /* user zero copy descriptor? */ ++ zccd_get (zccd); /* extra ref (pages are shared) */ ++ if (zccd2 != NULL) /* 2nd user zero copy descriptor? */ ++ zccd_get (zccd2); /* extra ref (pages are shared) */ ++ + skb_release_data(skb); + + off = (data+nhead) - skb->head; +@@ -658,6 +681,8 @@ int pskb_expand_head(struct sk_buff *skb + skb->nh.raw += off; + skb->cloned = 0; + atomic_set(&skb_shinfo(skb)->dataref, 1); ++ skb_shinfo(skb)->zccd = zccd; ++ skb_shinfo(skb)->zccd2 = zccd2; + return 0; + + nodata: +--- linux-2.4.22-ac1/net/ipv4/tcp.c~tcp-zero-copy-2.4.22-rh 2003-08-25 15:44:44.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/net/ipv4/tcp.c 2003-09-26 00:38:48.000000000 +0400 +@@ -747,7 +747,7 @@ do_interrupted: + goto out; + } + +-ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffset, size_t psize, int flags); ++ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffset, size_t psize, int flags, zccd_t *zccd); + + static inline int + can_coalesce(struct sk_buff *skb, int i, struct page *page, int off) +@@ -826,7 +826,8 @@ static int tcp_error(struct sock *sk, in + return err; + } + +-ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffset, size_t psize, int flags) ++/* Extra parameter: user zero copy descriptor (or NULL if not doing that) */ ++ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffset, size_t psize, int flags, zccd_t *zccd) + { + struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp); + int mss_now; +@@ -874,6 +875,17 @@ new_segment: + copy = size; + + i = skb_shinfo(skb)->nr_frags; ++ ++ if (zccd != NULL && /* this is a zcc I/O */ ++ skb_shinfo(skb)->zccd != NULL && /* skb is part of a zcc I/O */ ++ skb_shinfo(skb)->zccd2 != NULL && ++ skb_shinfo(skb)->zccd != zccd && /* not the same one */ ++ skb_shinfo(skb)->zccd2 != zccd) ++ { ++ tcp_mark_push (tp, skb); ++ goto new_segment; ++ } ++ + if (can_coalesce(skb, i, page, offset)) { + skb_shinfo(skb)->frags[i-1].size += copy; + } else if (i < MAX_SKB_FRAGS) { +@@ -884,6 +896,20 @@ new_segment: + goto new_segment; + } + ++ if (zccd != NULL && /* this is a zcc I/O */ ++ skb_shinfo(skb)->zccd != zccd && /* not already referencing this zccd */ ++ skb_shinfo(skb)->zccd2 != zccd) ++ { ++ zccd_get (zccd); /* bump ref count */ ++ ++ BUG_TRAP (skb_shinfo(skb)->zccd2 == NULL); ++ ++ if (skb_shinfo(skb)->zccd == NULL) /* reference this zccd */ ++ skb_shinfo(skb)->zccd = zccd; ++ else ++ skb_shinfo(skb)->zccd2 = zccd; ++ } ++ + skb->len += copy; + skb->data_len += copy; + skb->ip_summed = CHECKSUM_HW; +@@ -947,7 +973,31 @@ ssize_t tcp_sendpage(struct socket *sock + + lock_sock(sk); + TCP_CHECK_TIMER(sk); +- res = do_tcp_sendpages(sk, &page, offset, size, flags); ++ res = do_tcp_sendpages(sk, &page, offset, size, flags, NULL); ++ TCP_CHECK_TIMER(sk); ++ release_sock(sk); ++ return res; ++} ++ ++ssize_t tcp_sendpage_zccd(struct socket *sock, struct page *page, int offset, size_t size, ++ int flags, zccd_t *zccd) ++{ ++ ssize_t res; ++ struct sock *sk = sock->sk; ++ ++#define TCP_ZC_CSUM_FLAGS (NETIF_F_IP_CSUM|NETIF_F_NO_CSUM|NETIF_F_HW_CSUM) ++ ++ if (!(sk->route_caps & NETIF_F_SG) || /* caller shouldn't waste her time */ ++ !(sk->route_caps & TCP_ZC_CSUM_FLAGS)) /* on double mapping */ ++ BUG (); ++ ++#undef TCP_ZC_CSUM_FLAGS ++ ++ lock_sock(sk); ++ TCP_CHECK_TIMER(sk); ++ ++ res = do_tcp_sendpages(sk, &page, offset, size, flags, zccd); ++ + TCP_CHECK_TIMER(sk); + release_sock(sk); + return res; +@@ -1771,6 +1821,202 @@ recv_urg: + goto out; + } + ++int tcp_recvpackets (struct sock *sk, struct sk_buff_head *packets, ++ int len, int nonblock) ++{ ++ struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp); ++ int copied; ++ long timeo; ++ ++ BUG_TRAP (len > 0); ++ /*BUG_TRAP ((flags & (MSG_OOB | MSG_PEEK | MSG_TRUNC)) == 0);*/ ++ ++ lock_sock(sk); ++ ++ TCP_CHECK_TIMER(sk); ++ ++ copied = -ENOTCONN; ++ if (sk->state == TCP_LISTEN) ++ goto out; ++ ++ copied = 0; ++ timeo = sock_rcvtimeo(sk, nonblock); ++ ++ do { ++ struct sk_buff * skb; ++ u32 offset; ++ unsigned long used; ++ int exhausted; ++ int eaten; ++ ++ /* Are we at urgent data? Stop if we have read anything. */ ++ if (copied && tp->urg_data && tp->urg_seq == tp->copied_seq) ++ break; ++ ++ /* We need to check signals first, to get correct SIGURG ++ * handling. FIXME: Need to check this doesnt impact 1003.1g ++ * and move it down to the bottom of the loop ++ */ ++ if (signal_pending(current)) { ++ if (copied) ++ break; ++ copied = timeo ? sock_intr_errno(timeo) : -EAGAIN; ++ break; ++ } ++ ++ /* Next get a buffer. */ ++ ++ skb = skb_peek(&sk->receive_queue); ++ ++ if (skb == NULL) /* nothing ready */ ++ { ++ if (copied) { ++ if (sk->err || ++ sk->state == TCP_CLOSE || ++ (sk->shutdown & RCV_SHUTDOWN) || ++ !timeo || ++ (0)) ++ break; ++ } else { ++ if (sk->done) ++ break; ++ ++ if (sk->err) { ++ copied = sock_error(sk); ++ break; ++ } ++ ++ if (sk->shutdown & RCV_SHUTDOWN) ++ break; ++ ++ if (sk->state == TCP_CLOSE) { ++ if (!sk->done) { ++ /* This occurs when user tries to read ++ * from never connected socket. ++ */ ++ copied = -ENOTCONN; ++ break; ++ } ++ break; ++ } ++ ++ if (!timeo) { ++ copied = -EAGAIN; ++ break; ++ } ++ } ++ ++ cleanup_rbuf(sk, copied); ++ timeo = tcp_data_wait(sk, timeo); ++ continue; ++ } ++ ++ BUG_TRAP (atomic_read (&skb->users) == 1); ++ ++ exhausted = eaten = 0; ++ ++ offset = tp->copied_seq - TCP_SKB_CB(skb)->seq; ++ if (skb->h.th->syn) ++ offset--; ++ ++ used = skb->len - offset; ++ ++ if (tp->urg_data) { ++ u32 urg_offset = tp->urg_seq - tp->copied_seq; ++ if (urg_offset < used) { ++ if (!urg_offset) { /* at urgent date */ ++ if (!sk->urginline) { ++ tp->copied_seq++; /* discard the single byte of urgent data */ ++ offset++; ++ used--; ++ } ++ } else /* truncate read */ ++ used = urg_offset; ++ } ++ } ++ ++ BUG_TRAP (used >= 0); ++ if (len < used) ++ used = len; ++ ++ if (used == 0) ++ exhausted = 1; ++ else ++ { ++ if (skb_is_nonlinear (skb)) ++ { ++ int rc = skb_linearize (skb, GFP_KERNEL); ++ ++ printk ("tcp_recvpackets(): linearising: %d\n", rc); ++ ++ if (rc) ++ { ++ if (!copied) ++ copied = rc; ++ break; ++ } ++ } ++ ++ if ((offset + used) == skb->len) /* consuming the whole packet */ ++ { ++ __skb_unlink (skb, &sk->receive_queue); ++ dst_release (skb->dst); ++ skb_orphan (skb); ++ __skb_pull (skb, offset); ++ __skb_queue_tail (packets, skb); ++ exhausted = eaten = 1; ++ } ++ else /* consuming only part of the packet */ ++ { ++ struct sk_buff *skb2 = skb_clone (skb, GFP_KERNEL); ++ ++ if (skb2 == NULL) ++ { ++ if (!copied) ++ copied = -ENOMEM; ++ break; ++ } ++ ++ dst_release (skb2->dst); ++ __skb_pull (skb2, offset); ++ __skb_trim (skb2, used); ++ __skb_queue_tail (packets, skb2); ++ } ++ ++ tp->copied_seq += used; ++ copied += used; ++ len -= used; ++ } ++ ++ if (tp->urg_data && after(tp->copied_seq,tp->urg_seq)) { ++ tp->urg_data = 0; ++ tcp_fast_path_check(sk, tp); ++ } ++ ++ if (!exhausted) ++ continue; ++ ++ if (skb->h.th->fin) ++ { ++ tp->copied_seq++; ++ if (!eaten) ++ tcp_eat_skb (sk, skb); ++ break; ++ } ++ ++ if (!eaten) ++ tcp_eat_skb (sk, skb); ++ ++ } while (len > 0); ++ ++ out: ++ /* Clean up data we have read: This will do ACK frames. */ ++ cleanup_rbuf(sk, copied); ++ TCP_CHECK_TIMER(sk); ++ release_sock(sk); ++ return copied; ++} ++ + /* + * State processing on a close. This implements the state shift for + * sending our FIN frame. Note that we only send a FIN for some +--- linux-2.4.22-ac1/net/netsyms.c~tcp-zero-copy-2.4.22-rh 2003-09-25 14:16:26.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/net/netsyms.c 2003-09-26 00:39:16.000000000 +0400 +@@ -396,6 +396,8 @@ EXPORT_SYMBOL(sysctl_tcp_wmem); + EXPORT_SYMBOL(sysctl_tcp_ecn); + EXPORT_SYMBOL(tcp_cwnd_application_limited); + EXPORT_SYMBOL(tcp_sendpage); ++EXPORT_SYMBOL(tcp_sendpage_zccd); ++EXPORT_SYMBOL(tcp_recvpackets); + EXPORT_SYMBOL(sysctl_tcp_low_latency); + + EXPORT_SYMBOL(tcp_write_xmit); + +_ diff --git a/lustre/kernel_patches/patches/vfs_intent-2.4.22-rh.patch b/lustre/kernel_patches/patches/vfs_intent-2.4.22-rh.patch new file mode 100644 index 0000000..5f266a8 --- /dev/null +++ b/lustre/kernel_patches/patches/vfs_intent-2.4.22-rh.patch @@ -0,0 +1,1809 @@ + fs/dcache.c | 19 ++ + fs/exec.c | 17 +- + fs/namei.c | 295 +++++++++++++++++++++++++++++++++++++++------- + fs/namespace.c | 28 +++- + fs/open.c | 172 +++++++++++++++++++------- + fs/stat.c | 52 +++++--- + include/linux/dcache.h | 60 +++++++++ + include/linux/fs.h | 32 ++++ + include/linux/fs_struct.h | 4 + kernel/exit.c | 3 + kernel/fork.c | 3 + kernel/ksyms.c | 1 + 12 files changed, 558 insertions(+), 128 deletions(-) + +--- linux-2.4.22-ac1/fs/dcache.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/dcache.c 2003-09-25 14:42:46.000000000 +0400 +@@ -181,6 +181,13 @@ int d_invalidate(struct dentry * dentry) + spin_unlock(&dcache_lock); + return 0; + } ++ ++ /* network invalidation by Lustre */ ++ if (dentry->d_flags & DCACHE_LUSTRE_INVALID) { ++ spin_unlock(&dcache_lock); ++ return 0; ++ } ++ + /* + * Check whether to do a partial shrink_dcache + * to get rid of unused child entries. +@@ -833,13 +840,19 @@ void d_delete(struct dentry * dentry) + * Adds a dentry to the hash according to its name. + */ + +-void d_rehash(struct dentry * entry) ++void __d_rehash(struct dentry * entry, int lock) + { + struct list_head *list = d_hash(entry->d_parent, entry->d_name.hash); + if (!list_empty(&entry->d_hash)) BUG(); +- spin_lock(&dcache_lock); ++ if (lock) spin_lock(&dcache_lock); + list_add(&entry->d_hash, list); +- spin_unlock(&dcache_lock); ++ if (lock) spin_unlock(&dcache_lock); ++} ++EXPORT_SYMBOL(__d_rehash); ++ ++void d_rehash(struct dentry * entry) ++{ ++ __d_rehash(entry, 1); + } + + #define do_switch(x,y) do { \ +--- linux-2.4.22-ac1/fs/exec.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/exec.c 2003-09-25 14:42:46.000000000 +0400 +@@ -115,8 +115,10 @@ asmlinkage long sys_uselib(const char * + struct file * file; + struct nameidata nd; + int error; ++ struct lookup_intent it = { .it_op = IT_OPEN, ++ .it_flags = FMODE_READ|FMODE_EXEC }; + +- error = user_path_walk(library, &nd); ++ error = user_path_walk_it(library, &nd, &it); + if (error) + goto out; + +@@ -128,7 +130,8 @@ asmlinkage long sys_uselib(const char * + if (error) + goto exit; + +- file = dentry_open(nd.dentry, nd.mnt, O_RDONLY); ++ file = dentry_open_it(nd.dentry, nd.mnt, O_RDONLY, &it); ++ intent_release(&it); + error = PTR_ERR(file); + if (IS_ERR(file)) + goto out; +@@ -390,8 +393,10 @@ struct file *open_exec(const char *name) + struct inode *inode; + struct file *file; + int err = 0; ++ struct lookup_intent it = { .it_op = IT_OPEN, ++ .it_flags = FMODE_READ|FMODE_EXEC }; + +- err = path_lookup(name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, &nd); ++ err = path_lookup_it(name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, &nd, &it); + file = ERR_PTR(err); + if (!err) { + inode = nd.dentry->d_inode; +@@ -403,7 +408,8 @@ struct file *open_exec(const char *name) + err = -EACCES; + file = ERR_PTR(err); + if (!err) { +- file = dentry_open(nd.dentry, nd.mnt, O_RDONLY); ++ file = dentry_open_it(nd.dentry, nd.mnt, O_RDONLY, &it); ++ intent_release(&it); + if (!IS_ERR(file)) { + err = deny_write_access(file); + if (err) { +@@ -415,6 +421,7 @@ out: + return file; + } + } ++ intent_release(&it); + path_release(&nd); + } + goto out; +@@ -1322,7 +1329,7 @@ int do_coredump(long signr, int exit_cod + goto close_fail; + if (!file->f_op->write) + goto close_fail; +- if (do_truncate(file->f_dentry, 0) != 0) ++ if (do_truncate(file->f_dentry, 0, 0) != 0) + goto close_fail; + + retval = binfmt->core_dump(signr, regs, file); +--- linux-2.4.22-ac1/fs/namei.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:23.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/namei.c 2003-09-25 14:44:40.000000000 +0400 +@@ -94,6 +94,13 @@ + * XEmacs seems to be relying on it... + */ + ++void intent_release(struct lookup_intent *it) ++{ ++ if (it && it->it_op_release) ++ it->it_op_release(it); ++ ++} ++ + /* In order to reduce some races, while at the same time doing additional + * checking and hopefully speeding things up, we copy filenames to the + * kernel data space before using them.. +@@ -260,10 +267,19 @@ void path_release(struct nameidata *nd) + * Internal lookup() using the new generic dcache. + * SMP-safe + */ +-static struct dentry * cached_lookup(struct dentry * parent, struct qstr * name, int flags) ++static struct dentry *cached_lookup(struct dentry *parent, struct qstr *name, ++ int flags, struct lookup_intent *it) + { + struct dentry * dentry = d_lookup(parent, name); + ++ if (dentry && dentry->d_op && dentry->d_op->d_revalidate_it) { ++ if (!dentry->d_op->d_revalidate_it(dentry, flags, it) && ++ !d_invalidate(dentry)) { ++ dput(dentry); ++ dentry = NULL; ++ } ++ return dentry; ++ } else + if (dentry && dentry->d_op && dentry->d_op->d_revalidate) { + if (!dentry->d_op->d_revalidate(dentry, flags) && !d_invalidate(dentry)) { + dput(dentry); +@@ -281,11 +297,15 @@ static struct dentry * cached_lookup(str + * make sure that nobody added the entry to the dcache in the meantime.. + * SMP-safe + */ +-static struct dentry * real_lookup(struct dentry * parent, struct qstr * name, int flags) ++static struct dentry *real_lookup(struct dentry *parent, struct qstr *name, ++ int flags, struct lookup_intent *it) + { + struct dentry * result; + struct inode *dir = parent->d_inode; ++ int counter = 0; + ++again: ++ counter++; + down(&dir->i_sem); + /* + * First re-do the cached lookup just in case it was created +@@ -300,6 +320,9 @@ static struct dentry * real_lookup(struc + result = ERR_PTR(-ENOMEM); + if (dentry) { + lock_kernel(); ++ if (dir->i_op->lookup_it) ++ result = dir->i_op->lookup_it(dir, dentry, it, flags); ++ else + result = dir->i_op->lookup(dir, dentry); + unlock_kernel(); + if (result) +@@ -321,6 +344,15 @@ static struct dentry * real_lookup(struc + dput(result); + result = ERR_PTR(-ENOENT); + } ++ } else if (result->d_op && result->d_op->d_revalidate_it) { ++ if (!result->d_op->d_revalidate_it(result, flags, it) && ++ !d_invalidate(result)) { ++ dput(result); ++ if (counter > 10) ++ result = ERR_PTR(-ESTALE); ++ if (!IS_ERR(result)) ++ goto again; ++ } + } + return result; + } +@@ -332,7 +364,8 @@ static struct dentry * real_lookup(struc + * Without that kind of total limit, nasty chains of consecutive + * symlinks can cause almost arbitrarily long lookups. + */ +-static inline int do_follow_link(struct dentry *dentry, struct nameidata *nd) ++static inline int do_follow_link(struct dentry *dentry, struct nameidata *nd, ++ struct lookup_intent *it) + { + int err; + if (current->link_count >= 5) +@@ -346,10 +379,12 @@ static inline int do_follow_link(struct + current->link_count++; + current->total_link_count++; + UPDATE_ATIME(dentry->d_inode); ++ nd->intent = it; + err = dentry->d_inode->i_op->follow_link(dentry, nd); + current->link_count--; + return err; + loop: ++ intent_release(it); + path_release(nd); + return -ELOOP; + } +@@ -447,7 +482,8 @@ static inline void follow_dotdot(struct + * + * We expect 'base' to be positive and a directory. + */ +-int link_path_walk(const char * name, struct nameidata *nd) ++int link_path_walk_it(const char *name, struct nameidata *nd, ++ struct lookup_intent *it) + { + struct dentry *dentry; + struct inode *inode; +@@ -520,9 +556,9 @@ int link_path_walk(const char * name, st + break; + } + /* This does the actual lookups.. */ +- dentry = cached_lookup(nd->dentry, &this, LOOKUP_CONTINUE); ++ dentry = cached_lookup(nd->dentry, &this, LOOKUP_CONTINUE, NULL); + if (!dentry) { +- dentry = real_lookup(nd->dentry, &this, LOOKUP_CONTINUE); ++ dentry = real_lookup(nd->dentry, &this, LOOKUP_CONTINUE, NULL); + err = PTR_ERR(dentry); + if (IS_ERR(dentry)) + break; +@@ -540,7 +576,7 @@ int link_path_walk(const char * name, st + goto out_dput; + + if (inode->i_op->follow_link) { +- err = do_follow_link(dentry, nd); ++ err = do_follow_link(dentry, nd, NULL); + dput(dentry); + if (err) + goto return_err; +@@ -556,7 +592,7 @@ int link_path_walk(const char * name, st + nd->dentry = dentry; + } + err = -ENOTDIR; +- if (!inode->i_op->lookup) ++ if (!inode->i_op->lookup && !inode->i_op->lookup_it) + break; + continue; + /* here ends the main loop */ +@@ -583,9 +619,9 @@ last_component: + if (err < 0) + break; + } +- dentry = cached_lookup(nd->dentry, &this, 0); ++ dentry = cached_lookup(nd->dentry, &this, 0, it); + if (!dentry) { +- dentry = real_lookup(nd->dentry, &this, 0); ++ dentry = real_lookup(nd->dentry, &this, 0, it); + err = PTR_ERR(dentry); + if (IS_ERR(dentry)) + break; +@@ -595,7 +631,7 @@ last_component: + inode = dentry->d_inode; + if ((lookup_flags & LOOKUP_FOLLOW) + && inode && inode->i_op && inode->i_op->follow_link) { +- err = do_follow_link(dentry, nd); ++ err = do_follow_link(dentry, nd, it); + dput(dentry); + if (err) + goto return_err; +@@ -609,7 +645,8 @@ last_component: + goto no_inode; + if (lookup_flags & LOOKUP_DIRECTORY) { + err = -ENOTDIR; +- if (!inode->i_op || !inode->i_op->lookup) ++ if (!inode->i_op || ++ (!inode->i_op->lookup && !inode->i_op->lookup_it)) + break; + } + goto return_base; +@@ -635,6 +672,25 @@ return_reval: + * Check the cached dentry for staleness. + */ + dentry = nd->dentry; ++ if (dentry && dentry->d_op && dentry->d_op->d_revalidate_it) { ++ err = -ESTALE; ++ if (!dentry->d_op->d_revalidate_it(dentry, 0, it)) { ++ struct dentry *new; ++ err = permission(dentry->d_parent->d_inode, ++ MAY_EXEC); ++ if (err) ++ break; ++ new = real_lookup(dentry->d_parent, ++ &dentry->d_name, 0, NULL); ++ d_invalidate(dentry); ++ dput(dentry); ++ if (IS_ERR(new)) { ++ err = PTR_ERR(new); ++ break; ++ } ++ nd->dentry = new; ++ } ++ } else + if (dentry && dentry->d_sb + && (dentry->d_sb->s_type->fs_flags & FS_ALWAYS_REVAL)) { + err = -ESTALE; +@@ -649,15 +705,28 @@ out_dput: + dput(dentry); + break; + } ++ if (err) ++ intent_release(it); + path_release(nd); + return_err: + return err; + } + ++int link_path_walk(const char * name, struct nameidata *nd) ++{ ++ return link_path_walk_it(name, nd, NULL); ++} ++ ++int path_walk_it(const char * name, struct nameidata *nd, struct lookup_intent *it) ++{ ++ current->total_link_count = 0; ++ return link_path_walk_it(name, nd, it); ++} ++ + int path_walk(const char * name, struct nameidata *nd) + { + current->total_link_count = 0; +- return link_path_walk(name, nd); ++ return link_path_walk_it(name, nd, NULL); + } + + /* SMP-safe */ +@@ -742,6 +811,17 @@ walk_init_root(const char *name, struct + } + + /* SMP-safe */ ++int path_lookup_it(const char *path, unsigned flags, struct nameidata *nd, ++ struct lookup_intent *it) ++{ ++ int error = 0; ++ if (path_init(path, flags, nd)) ++ error = path_walk_it(path, nd, it); ++ return error; ++} ++ ++ ++/* SMP-safe */ + int path_lookup(const char *path, unsigned flags, struct nameidata *nd) + { + int error = 0; +@@ -756,6 +836,7 @@ int path_init(const char *name, unsigned + { + nd->last_type = LAST_ROOT; /* if there are only slashes... */ + nd->flags = flags; ++ nd->intent = NULL; + if (*name=='/') + return walk_init_root(name,nd); + read_lock(¤t->fs->lock); +@@ -770,7 +851,8 @@ int path_init(const char *name, unsigned + * needs parent already locked. Doesn't follow mounts. + * SMP-safe. + */ +-struct dentry * lookup_hash(struct qstr *name, struct dentry * base) ++struct dentry * lookup_hash_it(struct qstr *name, struct dentry * base, ++ struct lookup_intent *it) + { + struct dentry * dentry; + struct inode *inode; +@@ -793,13 +875,16 @@ struct dentry * lookup_hash(struct qstr + goto out; + } + +- dentry = cached_lookup(base, name, 0); ++ dentry = cached_lookup(base, name, 0, it); + if (!dentry) { + struct dentry *new = d_alloc(base, name); + dentry = ERR_PTR(-ENOMEM); + if (!new) + goto out; + lock_kernel(); ++ if (inode->i_op->lookup_it) ++ dentry = inode->i_op->lookup_it(inode, new, it, 0); ++ else + dentry = inode->i_op->lookup(inode, new); + unlock_kernel(); + if (!dentry) +@@ -811,6 +896,12 @@ out: + return dentry; + } + ++struct dentry * lookup_hash(struct qstr *name, struct dentry * base) ++{ ++ return lookup_hash_it(name, base, NULL); ++} ++ ++ + /* SMP-safe */ + struct dentry * lookup_one_len(const char * name, struct dentry * base, int len) + { +@@ -832,7 +923,7 @@ struct dentry * lookup_one_len(const cha + } + this.hash = end_name_hash(hash); + +- return lookup_hash(&this, base); ++ return lookup_hash_it(&this, base, NULL); + access: + return ERR_PTR(-EACCES); + } +@@ -863,6 +954,23 @@ int __user_walk(const char *name, unsign + return err; + } + ++int __user_walk_it(const char *name, unsigned flags, struct nameidata *nd, ++ struct lookup_intent *it) ++{ ++ char *tmp; ++ int err; ++ ++ tmp = getname(name); ++ err = PTR_ERR(tmp); ++ if (!IS_ERR(tmp)) { ++ err = 0; ++ if (path_init(tmp, flags, nd)) ++ err = path_walk_it(tmp, nd, it); ++ putname(tmp); ++ } ++ return err; ++} ++ + /* + * It's inline, so penalty for filesystems that don't use sticky bit is + * minimal. +@@ -958,7 +1066,8 @@ static inline int lookup_flags(unsigned + return retval; + } + +-int vfs_create(struct inode *dir, struct dentry *dentry, int mode) ++static int vfs_create_it(struct inode *dir, struct dentry *dentry, int mode, ++ struct lookup_intent *it) + { + int error; + +@@ -971,12 +1080,15 @@ int vfs_create(struct inode *dir, struct + goto exit_lock; + + error = -EACCES; /* shouldn't it be ENOSYS? */ +- if (!dir->i_op || !dir->i_op->create) ++ if (!dir->i_op || (!dir->i_op->create && !dir->i_op->create_it)) + goto exit_lock; + + DQUOT_INIT(dir); + lock_kernel(); +- error = dir->i_op->create(dir, dentry, mode); ++ if (dir->i_op->create_it) ++ error = dir->i_op->create_it(dir, dentry, mode, it); ++ else ++ error = dir->i_op->create(dir, dentry, mode); + unlock_kernel(); + exit_lock: + up(&dir->i_zombie); +@@ -985,6 +1097,11 @@ exit_lock: + return error; + } + ++int vfs_create(struct inode *dir, struct dentry *dentry, int mode) ++{ ++ return vfs_create_it(dir, dentry, mode, NULL); ++} ++ + /* + * open_namei() + * +@@ -999,7 +1116,8 @@ exit_lock: + * for symlinks (where the permissions are checked later). + * SMP-safe + */ +-int open_namei(const char * pathname, int flag, int mode, struct nameidata *nd) ++int open_namei_it(const char *pathname, int flag, int mode, ++ struct nameidata *nd, struct lookup_intent *it) + { + int acc_mode, error = 0; + struct inode *inode; +@@ -1009,11 +1127,14 @@ int open_namei(const char * pathname, in + + acc_mode = ACC_MODE(flag); + ++ if (it) ++ it->it_flags = flag; ++ + /* + * The simplest case - just a plain lookup. + */ + if (!(flag & O_CREAT)) { +- error = path_lookup(pathname, lookup_flags(flag), nd); ++ error = path_lookup_it(pathname, lookup_flags(flag), nd, it); + if (error) + return error; + dentry = nd->dentry; +@@ -1023,6 +1144,10 @@ int open_namei(const char * pathname, in + /* + * Create - we need to know the parent. + */ ++ if (it) { ++ it->it_create_mode = mode; ++ it->it_op |= IT_CREAT; ++ } + error = path_lookup(pathname, LOOKUP_PARENT, nd); + if (error) + return error; +@@ -1038,7 +1163,7 @@ int open_namei(const char * pathname, in + + dir = nd->dentry; + down(&dir->d_inode->i_sem); +- dentry = lookup_hash(&nd->last, nd->dentry); ++ dentry = lookup_hash_it(&nd->last, nd->dentry, it); + + do_last: + error = PTR_ERR(dentry); +@@ -1047,10 +1172,11 @@ do_last: + goto exit; + } + ++ it->it_create_mode = mode; + /* Negative dentry, just create the file */ + if (!dentry->d_inode) { +- error = vfs_create(dir->d_inode, dentry, +- mode & ~current->fs->umask); ++ error = vfs_create_it(dir->d_inode, dentry, ++ mode & ~current->fs->umask, it); + up(&dir->d_inode->i_sem); + dput(nd->dentry); + nd->dentry = dentry; +@@ -1154,7 +1280,7 @@ ok: + if (!error) { + DQUOT_INIT(inode); + +- error = do_truncate(dentry, 0); ++ error = do_truncate(dentry, 0, 1); + } + put_write_access(inode); + if (error) +@@ -1166,8 +1292,10 @@ ok: + return 0; + + exit_dput: ++ intent_release(it); + dput(dentry); + exit: ++ intent_release(it); + path_release(nd); + return error; + +@@ -1186,7 +1314,10 @@ do_link: + * are done. Procfs-like symlinks just set LAST_BIND. + */ + UPDATE_ATIME(dentry->d_inode); ++ nd->intent = it; + error = dentry->d_inode->i_op->follow_link(dentry, nd); ++ if (error) ++ intent_release(it); + dput(dentry); + if (error) + return error; +@@ -1208,13 +1339,20 @@ do_link: + } + dir = nd->dentry; + down(&dir->d_inode->i_sem); +- dentry = lookup_hash(&nd->last, nd->dentry); ++ dentry = lookup_hash_it(&nd->last, nd->dentry, it); + putname(nd->last.name); + goto do_last; + } + ++int open_namei(const char *pathname, int flag, int mode, struct nameidata *nd) ++{ ++ return open_namei_it(pathname, flag, mode, nd, NULL); ++} ++ ++ + /* SMP-safe */ +-static struct dentry *lookup_create(struct nameidata *nd, int is_dir) ++static struct dentry *lookup_create(struct nameidata *nd, int is_dir, ++ struct lookup_intent *it) + { + struct dentry *dentry; + +@@ -1222,7 +1360,7 @@ static struct dentry *lookup_create(stru + dentry = ERR_PTR(-EEXIST); + if (nd->last_type != LAST_NORM) + goto fail; +- dentry = lookup_hash(&nd->last, nd->dentry); ++ dentry = lookup_hash_it(&nd->last, nd->dentry, it); + if (IS_ERR(dentry)) + goto fail; + if (!is_dir && nd->last.name[nd->last.len] && !dentry->d_inode) +@@ -1278,7 +1416,16 @@ asmlinkage long sys_mknod(const char * f + error = path_lookup(tmp, LOOKUP_PARENT, &nd); + if (error) + goto out; +- dentry = lookup_create(&nd, 0); ++ ++ if (nd.dentry->d_inode->i_op->mknod_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ error = op->mknod_raw(&nd, mode, dev); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto out2; ++ } ++ ++ dentry = lookup_create(&nd, 0, NULL); + error = PTR_ERR(dentry); + + mode &= ~current->fs->umask; +@@ -1299,6 +1446,7 @@ asmlinkage long sys_mknod(const char * f + dput(dentry); + } + up(&nd.dentry->d_inode->i_sem); ++out2: + path_release(&nd); + out: + putname(tmp); +@@ -1346,7 +1494,14 @@ asmlinkage long sys_mkdir(const char * p + error = path_lookup(tmp, LOOKUP_PARENT, &nd); + if (error) + goto out; +- dentry = lookup_create(&nd, 1); ++ if (nd.dentry->d_inode->i_op->mkdir_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ error = op->mkdir_raw(&nd, mode); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto out2; ++ } ++ dentry = lookup_create(&nd, 1, NULL); + error = PTR_ERR(dentry); + if (!IS_ERR(dentry)) { + error = vfs_mkdir(nd.dentry->d_inode, dentry, +@@ -1354,6 +1509,7 @@ asmlinkage long sys_mkdir(const char * p + dput(dentry); + } + up(&nd.dentry->d_inode->i_sem); ++out2: + path_release(&nd); + out: + putname(tmp); +@@ -1454,8 +1610,16 @@ asmlinkage long sys_rmdir(const char * p + error = -EBUSY; + goto exit1; + } ++ if (nd.dentry->d_inode->i_op->rmdir_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ ++ error = op->rmdir_raw(&nd); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto exit1; ++ } + down(&nd.dentry->d_inode->i_sem); +- dentry = lookup_hash(&nd.last, nd.dentry); ++ dentry = lookup_hash_it(&nd.last, nd.dentry, NULL); + error = PTR_ERR(dentry); + if (!IS_ERR(dentry)) { + error = vfs_rmdir(nd.dentry->d_inode, dentry); +@@ -1513,8 +1677,15 @@ asmlinkage long sys_unlink(const char * + error = -EISDIR; + if (nd.last_type != LAST_NORM) + goto exit1; ++ if (nd.dentry->d_inode->i_op->unlink_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ error = op->unlink_raw(&nd); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto exit1; ++ } + down(&nd.dentry->d_inode->i_sem); +- dentry = lookup_hash(&nd.last, nd.dentry); ++ dentry = lookup_hash_it(&nd.last, nd.dentry, NULL); + error = PTR_ERR(dentry); + if (!IS_ERR(dentry)) { + /* Why not before? Because we want correct error value */ +@@ -1581,15 +1752,23 @@ asmlinkage long sys_symlink(const char * + error = path_lookup(to, LOOKUP_PARENT, &nd); + if (error) + goto out; +- dentry = lookup_create(&nd, 0); ++ if (nd.dentry->d_inode->i_op->symlink_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ error = op->symlink_raw(&nd, from); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto out2; ++ } ++ dentry = lookup_create(&nd, 0, NULL); + error = PTR_ERR(dentry); + if (!IS_ERR(dentry)) { + error = vfs_symlink(nd.dentry->d_inode, dentry, from); + dput(dentry); + } + up(&nd.dentry->d_inode->i_sem); ++ out2: + path_release(&nd); +-out: ++ out: + putname(to); + } + putname(from); +@@ -1665,7 +1844,14 @@ asmlinkage long sys_link(const char * ol + error = -EXDEV; + if (old_nd.mnt != nd.mnt) + goto out_release; +- new_dentry = lookup_create(&nd, 0); ++ if (nd.dentry->d_inode->i_op->link_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ error = op->link_raw(&old_nd, &nd); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto out_release; ++ } ++ new_dentry = lookup_create(&nd, 0, NULL); + error = PTR_ERR(new_dentry); + if (!IS_ERR(new_dentry)) { + error = vfs_link(old_nd.dentry, nd.dentry->d_inode, new_dentry); +@@ -1709,7 +1895,7 @@ exit: + * locking]. + */ + int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry, +- struct inode *new_dir, struct dentry *new_dentry) ++ struct inode *new_dir, struct dentry *new_dentry) + { + int error; + struct inode *target; +@@ -1788,7 +1974,7 @@ out_unlock: + } + + int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry, +- struct inode *new_dir, struct dentry *new_dentry) ++ struct inode *new_dir, struct dentry *new_dentry) + { + int error; + +@@ -1876,9 +2062,18 @@ static inline int do_rename(const char * + if (newnd.last_type != LAST_NORM) + goto exit2; + ++ if (old_dir->d_inode->i_op->rename_raw) { ++ lock_kernel(); ++ error = old_dir->d_inode->i_op->rename_raw(&oldnd, &newnd); ++ unlock_kernel(); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto exit2; ++ } ++ + double_lock(new_dir, old_dir); + +- old_dentry = lookup_hash(&oldnd.last, old_dir); ++ old_dentry = lookup_hash_it(&oldnd.last, old_dir, NULL); + error = PTR_ERR(old_dentry); + if (IS_ERR(old_dentry)) + goto exit3; +@@ -1894,16 +2089,16 @@ static inline int do_rename(const char * + if (newnd.last.name[newnd.last.len]) + goto exit4; + } +- new_dentry = lookup_hash(&newnd.last, new_dir); ++ new_dentry = lookup_hash_it(&newnd.last, new_dir, NULL); + error = PTR_ERR(new_dentry); + if (IS_ERR(new_dentry)) + goto exit4; + ++ + lock_kernel(); + error = vfs_rename(old_dir->d_inode, old_dentry, + new_dir->d_inode, new_dentry); + unlock_kernel(); +- + dput(new_dentry); + exit4: + dput(old_dentry); +@@ -1954,20 +2149,26 @@ out: + } + + static inline int +-__vfs_follow_link(struct nameidata *nd, const char *link) ++__vfs_follow_link(struct nameidata *nd, const char *link, ++ struct lookup_intent *it) + { + int res = 0; + char *name; + if (IS_ERR(link)) + goto fail; + ++ if (it == NULL) ++ it = nd->intent; ++ else if (it != nd->intent) ++ printk("it != nd->intent: tell phil@clusterfs.com\n"); ++ + if (*link == '/') { + path_release(nd); + if (!walk_init_root(link, nd)) + /* weird __emul_prefix() stuff did it */ + goto out; + } +- res = link_path_walk(link, nd); ++ res = link_path_walk_it(link, nd, it); + out: + if (current->link_count || res || nd->last_type!=LAST_NORM) + return res; +@@ -1991,7 +2192,13 @@ fail: + + int vfs_follow_link(struct nameidata *nd, const char *link) + { +- return __vfs_follow_link(nd, link); ++ return __vfs_follow_link(nd, link, NULL); ++} ++ ++int vfs_follow_link_it(struct nameidata *nd, const char *link, ++ struct lookup_intent *it) ++{ ++ return __vfs_follow_link(nd, link, it); + } + + /* get the link contents into pagecache */ +@@ -2033,7 +2240,7 @@ int page_follow_link(struct dentry *dent + { + struct page *page = NULL; + char *s = page_getlink(dentry, &page); +- int res = __vfs_follow_link(nd, s); ++ int res = __vfs_follow_link(nd, s, NULL); + if (page) { + kunmap(page); + page_cache_release(page); +--- linux-2.4.22-ac1/fs/namespace.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/namespace.c 2003-09-25 14:42:46.000000000 +0400 +@@ -98,6 +98,7 @@ static void detach_mnt(struct vfsmount * + { + old_nd->dentry = mnt->mnt_mountpoint; + old_nd->mnt = mnt->mnt_parent; ++ UNPIN(old_nd->dentry, old_nd->mnt, 1); + mnt->mnt_parent = mnt; + mnt->mnt_mountpoint = mnt->mnt_root; + list_del_init(&mnt->mnt_child); +@@ -109,6 +110,7 @@ static void attach_mnt(struct vfsmount * + { + mnt->mnt_parent = mntget(nd->mnt); + mnt->mnt_mountpoint = dget(nd->dentry); ++ PIN(nd->dentry, nd->mnt, 1); + list_add(&mnt->mnt_hash, mount_hashtable+hash(nd->mnt, nd->dentry)); + list_add(&mnt->mnt_child, &nd->mnt->mnt_mounts); + nd->dentry->d_mounted++; +@@ -488,14 +490,17 @@ static int do_loopback(struct nameidata + { + struct nameidata old_nd; + struct vfsmount *mnt = NULL; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + int err = mount_is_safe(nd); + if (err) + return err; + if (!old_name || !*old_name) + return -EINVAL; +- err = path_lookup(old_name, LOOKUP_POSITIVE|LOOKUP_FOLLOW, &old_nd); +- if (err) ++ err = path_lookup_it(old_name, LOOKUP_POSITIVE|LOOKUP_FOLLOW, &old_nd, &it); ++ if (err) { ++ intent_release(&it); + return err; ++ } + + down_write(¤t->namespace->sem); + err = -EINVAL; +@@ -518,6 +523,7 @@ static int do_loopback(struct nameidata + } + + up_write(¤t->namespace->sem); ++ intent_release(&it); + path_release(&old_nd); + return err; + } +@@ -701,6 +707,7 @@ long do_mount(char * dev_name, char * di + unsigned long flags, void *data_page) + { + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + int retval = 0; + int mnt_flags = 0; + +@@ -725,10 +732,11 @@ long do_mount(char * dev_name, char * di + flags &= ~(MS_NOSUID|MS_NOEXEC|MS_NODEV); + + /* ... and get the mountpoint */ +- retval = path_lookup(dir_name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, &nd); +- if (retval) ++ retval = path_lookup_it(dir_name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, &nd, &it); ++ if (retval) { ++ intent_release(&it); + return retval; +- ++ } + if (flags & MS_REMOUNT) + retval = do_remount(&nd, flags & ~MS_REMOUNT, mnt_flags, + data_page); +@@ -739,6 +747,8 @@ long do_mount(char * dev_name, char * di + else + retval = do_add_mount(&nd, type_page, flags, mnt_flags, + dev_name, data_page); ++ ++ intent_release(&it); + path_release(&nd); + return retval; + } +@@ -904,6 +914,8 @@ asmlinkage long sys_pivot_root(const cha + { + struct vfsmount *tmp; + struct nameidata new_nd, old_nd, parent_nd, root_parent, user_nd; ++ struct lookup_intent new_it = { .it_op = IT_GETATTR }; ++ struct lookup_intent old_it = { .it_op = IT_GETATTR }; + int error; + + if (!capable(CAP_SYS_ADMIN)) +@@ -911,14 +923,14 @@ asmlinkage long sys_pivot_root(const cha + + lock_kernel(); + +- error = __user_walk(new_root, LOOKUP_POSITIVE|LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &new_nd); ++ error = __user_walk_it(new_root, LOOKUP_POSITIVE|LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &new_nd, &new_it); + if (error) + goto out0; + error = -EINVAL; + if (!check_mnt(new_nd.mnt)) + goto out1; + +- error = __user_walk(put_old, LOOKUP_POSITIVE|LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &old_nd); ++ error = __user_walk_it(put_old, LOOKUP_POSITIVE|LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &old_nd, &old_it); + if (error) + goto out1; + +@@ -973,8 +985,10 @@ out2: + up(&old_nd.dentry->d_inode->i_zombie); + up_write(¤t->namespace->sem); + path_release(&user_nd); ++ intent_release(&old_it); + path_release(&old_nd); + out1: ++ intent_release(&new_it); + path_release(&new_nd); + out0: + unlock_kernel(); +--- linux-2.4.22-ac1/fs/open.c~vfs_intent-2.4.22-rh 2003-08-25 15:44:43.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/open.c 2003-09-25 14:42:46.000000000 +0400 +@@ -19,6 +19,8 @@ + #include + + #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m)) ++extern int path_walk_it(const char *name, struct nameidata *nd, ++ struct lookup_intent *it); + + int vfs_statfs(struct super_block *sb, struct statfs *buf) + { +@@ -95,9 +97,10 @@ void fd_install(unsigned int fd, struct + write_unlock(&files->file_lock); + } + +-int do_truncate(struct dentry *dentry, loff_t length) ++int do_truncate(struct dentry *dentry, loff_t length, int called_from_open) + { + struct inode *inode = dentry->d_inode; ++ struct inode_operations *op = dentry->d_inode->i_op; + int error; + struct iattr newattrs; + +@@ -109,7 +112,13 @@ int do_truncate(struct dentry *dentry, l + down(&inode->i_sem); + newattrs.ia_size = length; + newattrs.ia_valid = ATTR_SIZE | ATTR_CTIME; +- error = notify_change(dentry, &newattrs); ++ if (called_from_open) ++ newattrs.ia_valid |= ATTR_FROM_OPEN; ++ if (op->setattr_raw) { ++ newattrs.ia_valid |= ATTR_RAW; ++ error = op->setattr_raw(inode, &newattrs); ++ } else ++ error = notify_change(dentry, &newattrs); + up(&inode->i_sem); + up_write(&inode->i_alloc_sem); + return error; +@@ -120,12 +129,13 @@ static inline long do_sys_truncate(const + struct nameidata nd; + struct inode * inode; + int error; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + + error = -EINVAL; + if (length < 0) /* sorry, but loff_t says... */ + goto out; + +- error = user_path_walk(path, &nd); ++ error = user_path_walk_it(path, &nd, &it); + if (error) + goto out; + inode = nd.dentry->d_inode; +@@ -165,11 +175,13 @@ static inline long do_sys_truncate(const + error = locks_verify_truncate(inode, NULL, length); + if (!error) { + DQUOT_INIT(inode); +- error = do_truncate(nd.dentry, length); ++ intent_release(&it); ++ error = do_truncate(nd.dentry, length, 0); + } + put_write_access(inode); + + dput_and_out: ++ intent_release(&it); + path_release(&nd); + out: + return error; +@@ -217,7 +229,7 @@ static inline long do_sys_ftruncate(unsi + + error = locks_verify_truncate(inode, file, length); + if (!error) +- error = do_truncate(dentry, length); ++ error = do_truncate(dentry, length, 0); + out_putf: + fput(file); + out: +@@ -262,11 +274,13 @@ asmlinkage long sys_utime(char * filenam + struct inode * inode; + struct iattr newattrs; + +- error = user_path_walk(filename, &nd); ++ error = user_path_walk_it(filename, &nd, NULL); + if (error) + goto out; + inode = nd.dentry->d_inode; + ++ /* this is safe without a Lustre lock because it only depends ++ on the super block */ + error = -EROFS; + if (IS_RDONLY(inode)) + goto dput_and_out; +@@ -281,11 +295,25 @@ asmlinkage long sys_utime(char * filenam + goto dput_and_out; + + newattrs.ia_valid |= ATTR_ATIME_SET | ATTR_MTIME_SET; +- } else { ++ } ++ ++ if (inode->i_op->setattr_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ ++ newattrs.ia_valid |= ATTR_RAW; ++ error = op->setattr_raw(inode, &newattrs); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto dput_and_out; ++ } ++ ++ error = -EPERM; ++ if (!times) { + if (current->fsuid != inode->i_uid && + (error = permission(inode,MAY_WRITE)) != 0) + goto dput_and_out; + } ++ + error = notify_change(nd.dentry, &newattrs); + dput_and_out: + path_release(&nd); +@@ -306,12 +334,14 @@ asmlinkage long sys_utimes(char * filena + struct inode * inode; + struct iattr newattrs; + +- error = user_path_walk(filename, &nd); ++ error = user_path_walk_it(filename, &nd, NULL); + + if (error) + goto out; + inode = nd.dentry->d_inode; + ++ /* this is safe without a Lustre lock because it only depends ++ on the super block */ + error = -EROFS; + if (IS_RDONLY(inode)) + goto dput_and_out; +@@ -326,7 +356,20 @@ asmlinkage long sys_utimes(char * filena + newattrs.ia_atime = times[0].tv_sec; + newattrs.ia_mtime = times[1].tv_sec; + newattrs.ia_valid |= ATTR_ATIME_SET | ATTR_MTIME_SET; +- } else { ++ } ++ ++ if (inode->i_op->setattr_raw) { ++ struct inode_operations *op = nd.dentry->d_inode->i_op; ++ ++ newattrs.ia_valid |= ATTR_RAW; ++ error = op->setattr_raw(inode, &newattrs); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ goto dput_and_out; ++ } ++ ++ error = -EPERM; ++ if (!utimes) { + if (current->fsuid != inode->i_uid && + (error = permission(inode,MAY_WRITE)) != 0) + goto dput_and_out; +@@ -349,6 +392,7 @@ asmlinkage long sys_access(const char * + int old_fsuid, old_fsgid; + kernel_cap_t old_cap; + int res; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + + if (mode & ~S_IRWXO) /* where's F_OK, X_OK, W_OK, R_OK? */ + return -EINVAL; +@@ -366,13 +410,14 @@ asmlinkage long sys_access(const char * + else + current->cap_effective = current->cap_permitted; + +- res = user_path_walk(filename, &nd); ++ res = user_path_walk_it(filename, &nd, &it); + if (!res) { + res = permission(nd.dentry->d_inode, mode); + /* SuS v2 requires we report a read only fs too */ + if(!res && (mode & S_IWOTH) && IS_RDONLY(nd.dentry->d_inode) + && !special_file(nd.dentry->d_inode->i_mode)) + res = -EROFS; ++ intent_release(&it); + path_release(&nd); + } + +@@ -387,8 +432,9 @@ asmlinkage long sys_chdir(const char * f + { + int error; + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + +- error = __user_walk(filename,LOOKUP_POSITIVE|LOOKUP_FOLLOW|LOOKUP_DIRECTORY,&nd); ++ error = __user_walk_it(filename,LOOKUP_POSITIVE|LOOKUP_FOLLOW|LOOKUP_DIRECTORY,&nd, &it); + if (error) + goto out; + +@@ -399,6 +445,7 @@ asmlinkage long sys_chdir(const char * f + set_fs_pwd(current->fs, nd.mnt, nd.dentry); + + dput_and_out: ++ intent_release(&it); + path_release(&nd); + out: + return error; +@@ -438,9 +485,10 @@ asmlinkage long sys_chroot(const char * + { + int error; + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + +- error = __user_walk(filename, LOOKUP_POSITIVE | LOOKUP_FOLLOW | +- LOOKUP_DIRECTORY | LOOKUP_NOALT, &nd); ++ error = __user_walk_it(filename, LOOKUP_POSITIVE | LOOKUP_FOLLOW | ++ LOOKUP_DIRECTORY | LOOKUP_NOALT, &nd, &it); + if (error) + goto out; + +@@ -456,39 +504,56 @@ asmlinkage long sys_chroot(const char * + set_fs_altroot(); + error = 0; + dput_and_out: ++ intent_release(&it); + path_release(&nd); + out: + return error; + } + +-asmlinkage long sys_fchmod(unsigned int fd, mode_t mode) ++int chmod_common(struct dentry *dentry, mode_t mode) + { +- struct inode * inode; +- struct dentry * dentry; +- struct file * file; +- int err = -EBADF; ++ struct inode *inode = dentry->d_inode; + struct iattr newattrs; ++ int err = -EROFS; + +- file = fget(fd); +- if (!file) ++ if (IS_RDONLY(inode)) + goto out; + +- dentry = file->f_dentry; +- inode = dentry->d_inode; ++ if (inode->i_op->setattr_raw) { ++ newattrs.ia_mode = mode; ++ newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; ++ newattrs.ia_valid |= ATTR_RAW; ++ err = inode->i_op->setattr_raw(inode, &newattrs); ++ /* the file system wants to use normal vfs path now */ ++ if (err != -EOPNOTSUPP) ++ goto out; ++ } + +- err = -EROFS; +- if (IS_RDONLY(inode)) +- goto out_putf; + err = -EPERM; + if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) +- goto out_putf; ++ goto out; ++ + if (mode == (mode_t) -1) + mode = inode->i_mode; + newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO); + newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; + err = notify_change(dentry, &newattrs); + +-out_putf: ++out: ++ return err; ++} ++ ++asmlinkage long sys_fchmod(unsigned int fd, mode_t mode) ++{ ++ struct file * file; ++ int err = -EBADF; ++ ++ file = fget(fd); ++ if (!file) ++ goto out; ++ ++ err = chmod_common(file->f_dentry, mode); ++ + fput(file); + out: + return err; +@@ -497,30 +562,14 @@ out: + asmlinkage long sys_chmod(const char * filename, mode_t mode) + { + struct nameidata nd; +- struct inode * inode; + int error; +- struct iattr newattrs; + + error = user_path_walk(filename, &nd); + if (error) + goto out; +- inode = nd.dentry->d_inode; +- +- error = -EROFS; +- if (IS_RDONLY(inode)) +- goto dput_and_out; + +- error = -EPERM; +- if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) +- goto dput_and_out; ++ error = chmod_common(nd.dentry, mode); + +- if (mode == (mode_t) -1) +- mode = inode->i_mode; +- newattrs.ia_mode = (mode & S_IALLUGO) | (inode->i_mode & ~S_IALLUGO); +- newattrs.ia_valid = ATTR_MODE | ATTR_CTIME; +- error = notify_change(nd.dentry, &newattrs); +- +-dput_and_out: + path_release(&nd); + out: + return error; +@@ -540,6 +589,20 @@ static int chown_common(struct dentry * + error = -EROFS; + if (IS_RDONLY(inode)) + goto out; ++ ++ if (inode->i_op->setattr_raw) { ++ struct inode_operations *op = dentry->d_inode->i_op; ++ ++ newattrs.ia_uid = user; ++ newattrs.ia_gid = group; ++ newattrs.ia_valid = ATTR_UID | ATTR_GID | ATTR_CTIME; ++ newattrs.ia_valid |= ATTR_RAW; ++ error = op->setattr_raw(inode, &newattrs); ++ /* the file system wants to use normal vfs path now */ ++ if (error != -EOPNOTSUPP) ++ return error; ++ } ++ + error = -EPERM; + if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) + goto out; +@@ -644,6 +707,7 @@ struct file *filp_open(const char * file + { + int namei_flags, error; + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_OPEN }; + + namei_flags = flags; + if ((namei_flags+1) & O_ACCMODE) +@@ -651,14 +715,15 @@ struct file *filp_open(const char * file + if (namei_flags & O_TRUNC) + namei_flags |= 2; + +- error = open_namei(filename, namei_flags, mode, &nd); +- if (!error) +- return dentry_open(nd.dentry, nd.mnt, flags); ++ error = open_namei_it(filename, namei_flags, mode, &nd, &it); ++ if (error) ++ return ERR_PTR(error); + +- return ERR_PTR(error); ++ return dentry_open_it(nd.dentry, nd.mnt, flags, &it); + } + +-struct file *dentry_open(struct dentry *dentry, struct vfsmount *mnt, int flags) ++struct file *dentry_open_it(struct dentry *dentry, struct vfsmount *mnt, ++ int flags, struct lookup_intent *it) + { + struct file * f; + struct inode *inode; +@@ -695,12 +760,15 @@ struct file *dentry_open(struct dentry * + } + + if (f->f_op && f->f_op->open) { ++ f->f_it = it; + error = f->f_op->open(inode,f); ++ f->f_it = NULL; + if (error) + goto cleanup_all; + } + f->f_flags &= ~(O_CREAT | O_EXCL | O_NOCTTY | O_TRUNC); + ++ intent_release(it); + return f; + + cleanup_all: +@@ -715,11 +783,17 @@ cleanup_all: + cleanup_file: + put_filp(f); + cleanup_dentry: ++ intent_release(it); + dput(dentry); + mntput(mnt); + return ERR_PTR(error); + } + ++struct file *dentry_open(struct dentry *dentry, struct vfsmount *mnt, int flags) ++{ ++ return dentry_open_it(dentry, mnt, flags, NULL); ++} ++ + /* + * Find an empty file descriptor entry, and mark it busy. + */ +--- linux-2.4.22-ac1/fs/stat.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:27.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/stat.c 2003-09-25 14:42:46.000000000 +0400 +@@ -17,10 +17,14 @@ + * Revalidate the inode. This is required for proper NFS attribute caching. + */ + static __inline__ int +-do_revalidate(struct dentry *dentry) ++do_revalidate(struct dentry *dentry, struct lookup_intent *it) + { + struct inode * inode = dentry->d_inode; +- if (inode->i_op && inode->i_op->revalidate) ++ if (!inode) ++ return -ENOENT; ++ if (inode->i_op && inode->i_op->revalidate_it) ++ return inode->i_op->revalidate_it(dentry, it); ++ else if (inode->i_op && inode->i_op->revalidate) + return inode->i_op->revalidate(dentry); + return 0; + } +@@ -143,13 +147,15 @@ static int cp_new_stat(struct inode * in + asmlinkage long sys_stat(char * filename, struct __old_kernel_stat * statbuf) + { + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + int error; + +- error = user_path_walk(filename, &nd); ++ error = user_path_walk_it(filename, &nd, &it); + if (!error) { +- error = do_revalidate(nd.dentry); ++ error = do_revalidate(nd.dentry, &it); + if (!error) + error = cp_old_stat(nd.dentry->d_inode, statbuf); ++ intent_release(&it); + path_release(&nd); + } + return error; +@@ -159,13 +165,15 @@ asmlinkage long sys_stat(char * filename + asmlinkage long sys_newstat(char * filename, struct stat * statbuf) + { + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + int error; + +- error = user_path_walk(filename, &nd); ++ error = user_path_walk_it(filename, &nd, &it); + if (!error) { +- error = do_revalidate(nd.dentry); ++ error = do_revalidate(nd.dentry, &it); + if (!error) + error = cp_new_stat(nd.dentry->d_inode, statbuf); ++ intent_release(&it); + path_release(&nd); + } + return error; +@@ -180,13 +188,15 @@ asmlinkage long sys_newstat(char * filen + asmlinkage long sys_lstat(char * filename, struct __old_kernel_stat * statbuf) + { + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + int error; + +- error = user_path_walk_link(filename, &nd); ++ error = user_path_walk_link_it(filename, &nd, &it); + if (!error) { +- error = do_revalidate(nd.dentry); ++ error = do_revalidate(nd.dentry, &it); + if (!error) + error = cp_old_stat(nd.dentry->d_inode, statbuf); ++ intent_release(&it); + path_release(&nd); + } + return error; +@@ -197,13 +207,15 @@ asmlinkage long sys_lstat(char * filenam + asmlinkage long sys_newlstat(char * filename, struct stat * statbuf) + { + struct nameidata nd; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + int error; + +- error = user_path_walk_link(filename, &nd); ++ error = user_path_walk_link_it(filename, &nd, &it); + if (!error) { +- error = do_revalidate(nd.dentry); ++ error = do_revalidate(nd.dentry, &it); + if (!error) + error = cp_new_stat(nd.dentry->d_inode, statbuf); ++ intent_release(&it); + path_release(&nd); + } + return error; +@@ -224,7 +236,7 @@ asmlinkage long sys_fstat(unsigned int f + if (f) { + struct dentry * dentry = f->f_dentry; + +- err = do_revalidate(dentry); ++ err = do_revalidate(dentry, NULL); + if (!err) + err = cp_old_stat(dentry->d_inode, statbuf); + fput(f); +@@ -243,7 +255,7 @@ asmlinkage long sys_newfstat(unsigned in + if (f) { + struct dentry * dentry = f->f_dentry; + +- err = do_revalidate(dentry); ++ err = do_revalidate(dentry, NULL); + if (!err) + err = cp_new_stat(dentry->d_inode, statbuf); + fput(f); +@@ -265,7 +277,7 @@ asmlinkage long sys_readlink(const char + + error = -EINVAL; + if (inode->i_op && inode->i_op->readlink && +- !(error = do_revalidate(nd.dentry))) { ++ !(error = do_revalidate(nd.dentry, NULL))) { + UPDATE_ATIME(inode); + error = inode->i_op->readlink(nd.dentry, buf, bufsiz); + } +@@ -341,12 +353,14 @@ asmlinkage long sys_stat64(char * filena + { + struct nameidata nd; + int error; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + +- error = user_path_walk(filename, &nd); ++ error = user_path_walk_it(filename, &nd, &it); + if (!error) { +- error = do_revalidate(nd.dentry); ++ error = do_revalidate(nd.dentry, &it); + if (!error) + error = cp_new_stat64(nd.dentry->d_inode, statbuf); ++ intent_release(&it); + path_release(&nd); + } + return error; +@@ -356,12 +370,14 @@ asmlinkage long sys_lstat64(char * filen + { + struct nameidata nd; + int error; ++ struct lookup_intent it = { .it_op = IT_GETATTR }; + +- error = user_path_walk_link(filename, &nd); ++ error = user_path_walk_link_it(filename, &nd, &it); + if (!error) { +- error = do_revalidate(nd.dentry); ++ error = do_revalidate(nd.dentry, &it); + if (!error) + error = cp_new_stat64(nd.dentry->d_inode, statbuf); ++ intent_release(&it); + path_release(&nd); + } + return error; +@@ -376,7 +392,7 @@ asmlinkage long sys_fstat64(unsigned lon + if (f) { + struct dentry * dentry = f->f_dentry; + +- err = do_revalidate(dentry); ++ err = do_revalidate(dentry, NULL); + if (!err) + err = cp_new_stat64(dentry->d_inode, statbuf); + fput(f); +--- linux-2.4.22-ac1/include/linux/dcache.h~vfs_intent-2.4.22-rh 2003-09-25 14:16:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/dcache.h 2003-09-25 14:42:46.000000000 +0400 +@@ -6,6 +6,51 @@ + #include + #include + #include ++#include ++ ++#define IT_OPEN 0x0001 ++#define IT_CREAT 0x0002 ++#define IT_READDIR 0x0004 ++#define IT_GETATTR 0x0008 ++#define IT_LOOKUP 0x0010 ++#define IT_UNLINK 0x0020 ++#define IT_GETXATTR 0x0040 ++#define IT_EXEC 0x0080 ++#define IT_PIN 0x0100 ++ ++#define IT_FL_LOCKED 0x0001 ++#define IT_FL_FOLLOWED 0x0002 /* set by vfs_follow_link */ ++ ++#define INTENT_MAGIC 0x19620323 ++ ++ ++struct lustre_intent_data { ++ int it_disposition; ++ int it_status; ++ __u64 it_lock_handle; ++ void *it_data; ++ int it_lock_mode; ++ int it_int_flags; ++}; ++struct lookup_intent { ++ int it_magic; ++ void (*it_op_release)(struct lookup_intent *); ++ int it_op; ++ int it_flags; ++ int it_create_mode; ++ union { ++ struct lustre_intent_data lustre; ++ } d; ++}; ++ ++static inline void intent_init(struct lookup_intent *it, int op, int flags) ++{ ++ memset(it, 0, sizeof(*it)); ++ it->it_magic = INTENT_MAGIC; ++ it->it_op = op; ++ it->it_flags = flags; ++} ++ + + /* + * linux/include/linux/dcache.h +@@ -95,8 +140,22 @@ struct dentry_operations { + int (*d_delete)(struct dentry *); + void (*d_release)(struct dentry *); + void (*d_iput)(struct dentry *, struct inode *); ++ int (*d_revalidate_it)(struct dentry *, int, struct lookup_intent *); ++ void (*d_pin)(struct dentry *, struct vfsmount * , int); ++ void (*d_unpin)(struct dentry *, struct vfsmount *, int); + }; + ++#define PIN(de,mnt,flag) if (de->d_op && de->d_op->d_pin) \ ++ de->d_op->d_pin(de, mnt, flag); ++#define UNPIN(de,mnt,flag) if (de->d_op && de->d_op->d_unpin) \ ++ de->d_op->d_unpin(de, mnt, flag); ++ ++ ++/* defined in fs/namei.c */ ++extern void intent_release(struct lookup_intent *it); ++/* defined in fs/dcache.c */ ++extern void __d_rehash(struct dentry * entry, int lock); ++ + /* the dentry parameter passed to d_hash and d_compare is the parent + * directory of the entries to be compared. It is used in case these + * functions need any directory specific information for determining +@@ -128,6 +187,7 @@ d_iput: no no yes + * s_nfsd_free_path semaphore will be down + */ + #define DCACHE_REFERENCED 0x0008 /* Recently used, don't discard. */ ++#define DCACHE_LUSTRE_INVALID 0x0010 /* Lustre invalidated */ + + extern spinlock_t dcache_lock; + +--- linux-2.4.22-ac1/include/linux/fs.h~vfs_intent-2.4.22-rh 2003-09-25 14:39:01.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/fs.h 2003-09-25 14:42:46.000000000 +0400 +@@ -73,6 +73,7 @@ extern int leases_enable, dir_notify_ena + + #define FMODE_READ 1 + #define FMODE_WRITE 2 ++#define FMODE_EXEC 4 + + #define READ 0 + #define WRITE 1 +@@ -343,6 +344,9 @@ extern void set_bh_page(struct buffer_he + #define ATTR_MTIME_SET 256 + #define ATTR_FORCE 512 /* Not a change, but a change it */ + #define ATTR_ATTR_FLAG 1024 ++#define ATTR_RAW 0x0800 /* file system, not vfs will massage attrs */ ++#define ATTR_FROM_OPEN 0x1000 /* called from open path, ie O_TRUNC */ ++#define ATTR_CTIME_SET 0x2000 + + /* + * This is the Inode Attributes structure, used for notify_change(). It +@@ -481,6 +485,7 @@ struct inode { + struct pipe_inode_info *i_pipe; + struct block_device *i_bdev; + struct char_device *i_cdev; ++ void *i_filterdata; + + unsigned long i_dnotify_mask; /* Directory notify events */ + struct dnotify_struct *i_dnotify; /* for directory notifications */ +@@ -583,6 +588,7 @@ struct file { + + /* needed for tty driver, and maybe others */ + void *private_data; ++ struct lookup_intent *f_it; + + /* preallocated helper kiobuf to speedup O_DIRECT */ + struct kiobuf *f_iobuf; +@@ -703,6 +709,7 @@ struct nameidata { + struct qstr last; + unsigned int flags; + int last_type; ++ struct lookup_intent *intent; + }; + + /* +@@ -823,7 +830,8 @@ extern int vfs_symlink(struct inode *, s + extern int vfs_link(struct dentry *, struct inode *, struct dentry *); + extern int vfs_rmdir(struct inode *, struct dentry *); + extern int vfs_unlink(struct inode *, struct dentry *); +-extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); ++int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, ++ struct inode *new_dir, struct dentry *new_dentry); + + /* + * File types +@@ -883,21 +891,32 @@ struct file_operations { + + struct inode_operations { + int (*create) (struct inode *,struct dentry *,int); ++ int (*create_it) (struct inode *,struct dentry *,int, struct lookup_intent *); + struct dentry * (*lookup) (struct inode *,struct dentry *); ++ struct dentry * (*lookup_it) (struct inode *,struct dentry *, struct lookup_intent *, int flags); + int (*link) (struct dentry *,struct inode *,struct dentry *); ++ int (*link_raw) (struct nameidata *,struct nameidata *); + int (*unlink) (struct inode *,struct dentry *); ++ int (*unlink_raw) (struct nameidata *); + int (*symlink) (struct inode *,struct dentry *,const char *); ++ int (*symlink_raw) (struct nameidata *,const char *); + int (*mkdir) (struct inode *,struct dentry *,int); ++ int (*mkdir_raw) (struct nameidata *,int); + int (*rmdir) (struct inode *,struct dentry *); ++ int (*rmdir_raw) (struct nameidata *); + int (*mknod) (struct inode *,struct dentry *,int,int); ++ int (*mknod_raw) (struct nameidata *,int,dev_t); + int (*rename) (struct inode *, struct dentry *, + struct inode *, struct dentry *); ++ int (*rename_raw) (struct nameidata *, struct nameidata *); + int (*readlink) (struct dentry *, char *,int); + int (*follow_link) (struct dentry *, struct nameidata *); + void (*truncate) (struct inode *); + int (*permission) (struct inode *, int); + int (*revalidate) (struct dentry *); ++ int (*revalidate_it) (struct dentry *, struct lookup_intent *); + int (*setattr) (struct dentry *, struct iattr *); ++ int (*setattr_raw) (struct inode *, struct iattr *); + int (*getattr) (struct dentry *, struct iattr *); + int (*setxattr) (struct dentry *, const char *, void *, size_t, int); + ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t); +@@ -1094,10 +1113,14 @@ static inline int get_lease(struct inode + + asmlinkage long sys_open(const char *, int, int); + asmlinkage long sys_close(unsigned int); /* yes, it's really unsigned */ +-extern int do_truncate(struct dentry *, loff_t start); ++extern int do_truncate(struct dentry *, loff_t start, int called_from_open); + + extern struct file *filp_open(const char *, int, int); + extern struct file * dentry_open(struct dentry *, struct vfsmount *, int); ++extern int open_namei_it(const char *filename, int namei_flags, int mode, ++ struct nameidata *nd, struct lookup_intent *it); ++extern struct file *dentry_open_it(struct dentry *dentry, struct vfsmount *mnt, ++ int flags, struct lookup_intent *it); + extern int filp_close(struct file *, fl_owner_t id); + extern char * getname(const char *); + +@@ -1388,6 +1411,7 @@ typedef int (*read_actor_t)(read_descrip + extern loff_t default_llseek(struct file *file, loff_t offset, int origin); + + extern int FASTCALL(__user_walk(const char *, unsigned, struct nameidata *)); ++extern int FASTCALL(__user_walk_it(const char *, unsigned, struct nameidata *, struct lookup_intent *it)); + extern int FASTCALL(path_init(const char *, unsigned, struct nameidata *)); + extern int FASTCALL(path_walk(const char *, struct nameidata *)); + extern int FASTCALL(path_lookup(const char *, unsigned, struct nameidata *)); +@@ -1399,6 +1423,8 @@ extern struct dentry * lookup_one_len(co + extern struct dentry * lookup_hash(struct qstr *, struct dentry *); + #define user_path_walk(name,nd) __user_walk(name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, nd) + #define user_path_walk_link(name,nd) __user_walk(name, LOOKUP_POSITIVE, nd) ++#define user_path_walk_it(name,nd,it) __user_walk_it(name, LOOKUP_FOLLOW|LOOKUP_POSITIVE, nd, it) ++#define user_path_walk_link_it(name,nd,it) __user_walk_it(name, LOOKUP_POSITIVE, nd, it) + + extern void inode_init_once(struct inode *); + extern void iput(struct inode *); +@@ -1538,6 +1564,8 @@ extern struct file_operations generic_ro + + extern int vfs_readlink(struct dentry *, char *, int, const char *); + extern int vfs_follow_link(struct nameidata *, const char *); ++extern int vfs_follow_link_it(struct nameidata *, const char *, ++ struct lookup_intent *it); + extern int page_readlink(struct dentry *, char *, int); + extern int page_follow_link(struct dentry *, struct nameidata *); + extern struct inode_operations page_symlink_inode_operations; +--- linux-2.4.22-ac1/include/linux/fs_struct.h~vfs_intent-2.4.22-rh 2003-09-25 14:16:24.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/fs_struct.h 2003-09-25 14:42:46.000000000 +0400 +@@ -37,10 +37,12 @@ static inline void set_fs_root(struct fs + write_lock(&fs->lock); + old_root = fs->root; + old_rootmnt = fs->rootmnt; ++ PIN(dentry, mnt, 1); + fs->rootmnt = mntget(mnt); + fs->root = dget(dentry); + write_unlock(&fs->lock); + if (old_root) { ++ UNPIN(old_root, old_rootmnt, 1); + dput(old_root); + mntput(old_rootmnt); + } +@@ -60,10 +62,12 @@ static inline void set_fs_pwd(struct fs_ + write_lock(&fs->lock); + old_pwd = fs->pwd; + old_pwdmnt = fs->pwdmnt; ++ PIN(dentry, mnt, 0); + fs->pwdmnt = mntget(mnt); + fs->pwd = dget(dentry); + write_unlock(&fs->lock); + if (old_pwd) { ++ UNPIN(old_pwd, old_pwdmnt, 0); + dput(old_pwd); + mntput(old_pwdmnt); + } +--- linux-2.4.22-ac1/kernel/exit.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/kernel/exit.c 2003-09-25 14:42:46.000000000 +0400 +@@ -342,11 +342,14 @@ static inline void __put_fs_struct(struc + { + /* No need to hold fs->lock if we are killing it */ + if (atomic_dec_and_test(&fs->count)) { ++ UNPIN(fs->pwd, fs->pwdmnt, 0); ++ UNPIN(fs->root, fs->rootmnt, 1); + dput(fs->root); + mntput(fs->rootmnt); + dput(fs->pwd); + mntput(fs->pwdmnt); + if (fs->altroot) { ++ UNPIN(fs->altroot, fs->altrootmnt, 1); + dput(fs->altroot); + mntput(fs->altrootmnt); + } +--- linux-2.4.22-ac1/kernel/fork.c~vfs_intent-2.4.22-rh 2003-09-25 14:16:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/kernel/fork.c 2003-09-25 14:42:46.000000000 +0400 +@@ -457,10 +457,13 @@ static inline struct fs_struct *__copy_f + fs->umask = old->umask; + read_lock(&old->lock); + fs->rootmnt = mntget(old->rootmnt); ++ PIN(old->pwd, old->pwdmnt, 0); ++ PIN(old->root, old->rootmnt, 1); + fs->root = dget(old->root); + fs->pwdmnt = mntget(old->pwdmnt); + fs->pwd = dget(old->pwd); + if (old->altroot) { ++ PIN(old->altroot, old->altrootmnt, 1); + fs->altrootmnt = mntget(old->altrootmnt); + fs->altroot = dget(old->altroot); + } else { +--- linux-2.4.22-ac1/kernel/ksyms.c~vfs_intent-2.4.22-rh 2003-09-25 14:39:02.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/kernel/ksyms.c 2003-09-25 14:42:46.000000000 +0400 +@@ -295,6 +295,7 @@ EXPORT_SYMBOL(read_cache_page); + EXPORT_SYMBOL(set_page_dirty); + EXPORT_SYMBOL(vfs_readlink); + EXPORT_SYMBOL(vfs_follow_link); ++EXPORT_SYMBOL(vfs_follow_link_it); + EXPORT_SYMBOL(page_readlink); + EXPORT_SYMBOL(page_follow_link); + EXPORT_SYMBOL(page_symlink_inode_operations); + +_ diff --git a/lustre/kernel_patches/patches/xattr-0.8.54-2.4.22-rh.patch b/lustre/kernel_patches/patches/xattr-0.8.54-2.4.22-rh.patch new file mode 100644 index 0000000..a213a48 --- /dev/null +++ b/lustre/kernel_patches/patches/xattr-0.8.54-2.4.22-rh.patch @@ -0,0 +1,5403 @@ + Documentation/Configure.help | 66 ++ + arch/alpha/defconfig | 7 + arch/alpha/kernel/entry.S | 12 + arch/arm/defconfig | 7 + arch/arm/kernel/calls.S | 24 + arch/i386/defconfig | 7 + arch/ia64/defconfig | 7 + arch/m68k/defconfig | 7 + arch/mips/defconfig | 7 + arch/mips64/defconfig | 7 + arch/ppc/defconfig | 14 + arch/ppc64/kernel/misc.S | 2 + arch/s390/defconfig | 7 + arch/s390/kernel/entry.S | 24 + arch/s390x/defconfig | 7 + arch/s390x/kernel/entry.S | 24 + arch/s390x/kernel/wrapper32.S | 90 +++ + arch/sparc/defconfig | 7 + arch/sparc64/defconfig | 7 + fs/Config.in | 14 + fs/Makefile | 3 + fs/ext2/Makefile | 4 + fs/ext2/file.c | 5 + fs/ext2/ialloc.c | 2 + fs/ext2/inode.c | 34 - + fs/ext2/namei.c | 14 + fs/ext2/super.c | 29 + fs/ext2/symlink.c | 14 + fs/ext2/xattr.c | 1212 +++++++++++++++++++++++++++++++++++++++++ + fs/ext2/xattr_user.c | 103 +++ + fs/ext3/Makefile | 10 + fs/ext3/ext3-exports.c | 13 + fs/ext3/file.c | 5 + fs/ext3/ialloc.c | 2 + fs/ext3/inode.c | 35 - + fs/ext3/namei.c | 21 + fs/ext3/super.c | 37 + + fs/ext3/symlink.c | 14 + fs/ext3/xattr.c | 1225 ++++++++++++++++++++++++++++++++++++++++++ + fs/ext3/xattr_user.c | 111 +++ + fs/jfs/jfs_xattr.h | 6 + fs/jfs/xattr.c | 6 + fs/mbcache.c | 648 ++++++++++++++++++++++ + include/asm-arm/unistd.h | 2 + include/asm-ppc64/unistd.h | 2 + include/asm-s390/unistd.h | 13 + include/asm-s390x/unistd.h | 13 + include/linux/cache_def.h | 15 + include/linux/errno.h | 4 + include/linux/ext2_fs.h | 31 - + include/linux/ext2_xattr.h | 157 +++++ + include/linux/ext3_fs.h | 31 - + include/linux/ext3_jbd.h | 8 + include/linux/ext3_xattr.h | 157 +++++ + include/linux/fs.h | 2 + include/linux/mbcache.h | 69 ++ + kernel/ksyms.c | 4 + mm/vmscan.c | 35 + + 58 files changed, 4306 insertions(+), 137 deletions(-) + +--- linux-2.4.22-ac1/arch/alpha/defconfig~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/alpha/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_ALPHA=y + # CONFIG_UID16 is not set + # CONFIG_RWSEM_GENERIC_SPINLOCK is not set +--- linux-2.4.22-ac1/arch/alpha/kernel/entry.S~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/alpha/kernel/entry.S 2003-09-25 23:57:02.000000000 +0400 +@@ -1158,6 +1158,18 @@ sys_call_table: + .quad sys_readahead + .quad sys_ni_syscall /* 380, sys_security */ + .quad sys_tkill ++ .quad sys_setxattr ++ .quad sys_lsetxattr ++ .quad sys_fsetxattr ++ .quad sys_getxattr /* 385 */ ++ .quad sys_lgetxattr ++ .quad sys_fgetxattr ++ .quad sys_listxattr ++ .quad sys_llistxattr ++ .quad sys_flistxattr /* 390 */ ++ .quad sys_removexattr ++ .quad sys_lremovexattr ++ .quad sys_fremovexattr + + /* Remember to update everything, kids. */ + .ifne (. - sys_call_table) - (NR_SYSCALLS * 8) +--- linux-2.4.22-ac1/arch/arm/defconfig~xattr-0.8.54-2.4.22-rh 2001-05-20 04:43:05.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/arm/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_ARM=y + # CONFIG_EISA is not set + # CONFIG_SBUS is not set +--- linux-2.4.22-ac1/arch/arm/kernel/calls.S~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:39.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/arm/kernel/calls.S 2003-09-26 00:00:10.000000000 +0400 +@@ -240,18 +240,18 @@ __syscall_start: + .long SYMBOL_NAME(sys_ni_syscall) /* Security */ + .long SYMBOL_NAME(sys_gettid) + /* 225 */ .long SYMBOL_NAME(sys_readahead) +- .long SYMBOL_NAME(sys_ni_syscall) /* setxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* lsetxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* fsetxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* getxattr */ +-/* 230 */ .long SYMBOL_NAME(sys_ni_syscall) /* lgetxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* fgetxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* listxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* llistxattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* flistxattr */ +-/* 235 */ .long SYMBOL_NAME(sys_ni_syscall) /* removexattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* lremovexattr */ +- .long SYMBOL_NAME(sys_ni_syscall) /* fremovexattr */ ++ .long SYMBOL_NAME(sys_setxattr) ++ .long SYMBOL_NAME(sys_lsetxattr) ++ .long SYMBOL_NAME(sys_fsetxattr) ++ .long SYMBOL_NAME(sys_getxattr) ++/* 230 */ .long SYMBOL_NAME(sys_lgetxattr) ++ .long SYMBOL_NAME(sys_fgetxattr) ++ .long SYMBOL_NAME(sys_listxattr) ++ .long SYMBOL_NAME(sys_llistxattr) ++ .long SYMBOL_NAME(sys_flistxattr) ++/* 235 */ .long SYMBOL_NAME(sys_removexattr) ++ .long SYMBOL_NAME(sys_lremovexattr) ++ .long SYMBOL_NAME(sys_fremovexattr) + .long SYMBOL_NAME(sys_tkill) + .long SYMBOL_NAME(sys_ni_syscall) /* sendfile64 */ + /* 240 */ .long SYMBOL_NAME(sys_ni_syscall) /* futex */ +--- linux-2.4.22-ac1/arch/i386/defconfig~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/i386/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_X86=y + CONFIG_ISA=y + # CONFIG_SBUS is not set +--- linux-2.4.22-ac1/arch/ia64/defconfig~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:39.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/ia64/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + + # + # Code maturity level options +--- linux-2.4.22-ac1/arch/m68k/defconfig~xattr-0.8.54-2.4.22-rh 2000-06-19 23:56:08.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/m68k/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_UID16=y + + # +--- linux-2.4.22-ac1/arch/mips64/defconfig~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:40.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/mips64/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_MIPS=y + # CONFIG_MIPS32 is not set + CONFIG_MIPS64=y +--- linux-2.4.22-ac1/arch/mips/defconfig~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:39.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/mips/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_MIPS=y + CONFIG_MIPS32=y + # CONFIG_MIPS64 is not set +--- linux-2.4.22-ac1/arch/ppc64/kernel/misc.S~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:40.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/ppc64/kernel/misc.S 2003-09-25 23:57:02.000000000 +0400 +@@ -805,6 +805,7 @@ _GLOBAL(sys_call_table32) + .llong .sys_gettid /* 207 */ + #if 0 /* Reserved syscalls */ + .llong .sys_tkill /* 208 */ ++#endif + .llong .sys_setxattr + .llong .sys_lsetxattr /* 210 */ + .llong .sys_fsetxattr +@@ -817,6 +818,7 @@ _GLOBAL(sys_call_table32) + .llong .sys_removexattr + .llong .sys_lremovexattr + .llong .sys_fremovexattr /* 220 */ ++#if 0 /* Reserved syscalls */ + .llong .sys_futex + #endif + .llong .sys_perfmonctl /* Put this here for now ... */ +--- linux-2.4.22-ac1/arch/ppc/defconfig~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:31.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/ppc/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,20 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + # CONFIG_UID16 is not set + # CONFIG_RWSEM_GENERIC_SPINLOCK is not set + CONFIG_RWSEM_XCHGADD_ALGORITHM=y +--- linux-2.4.22-ac1/arch/s390/defconfig~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/s390/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + # CONFIG_ISA is not set + # CONFIG_EISA is not set + # CONFIG_MCA is not set +--- linux-2.4.22-ac1/arch/s390/kernel/entry.S~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/s390/kernel/entry.S 2003-09-25 23:57:02.000000000 +0400 +@@ -559,18 +559,18 @@ sys_call_table: + .long sys_fcntl64 + .long sys_readahead + .long sys_ni_syscall +- .long sys_ni_syscall /* 224 - reserved for setxattr */ +- .long sys_ni_syscall /* 225 - reserved for lsetxattr */ +- .long sys_ni_syscall /* 226 - reserved for fsetxattr */ +- .long sys_ni_syscall /* 227 - reserved for getxattr */ +- .long sys_ni_syscall /* 228 - reserved for lgetxattr */ +- .long sys_ni_syscall /* 229 - reserved for fgetxattr */ +- .long sys_ni_syscall /* 230 - reserved for listxattr */ +- .long sys_ni_syscall /* 231 - reserved for llistxattr */ +- .long sys_ni_syscall /* 232 - reserved for flistxattr */ +- .long sys_ni_syscall /* 233 - reserved for removexattr */ +- .long sys_ni_syscall /* 234 - reserved for lremovexattr */ +- .long sys_ni_syscall /* 235 - reserved for fremovexattr */ ++ .long sys_setxattr ++ .long sys_lsetxattr /* 225 */ ++ .long sys_fsetxattr ++ .long sys_getxattr ++ .long sys_lgetxattr ++ .long sys_fgetxattr ++ .long sys_listxattr /* 230 */ ++ .long sys_llistxattr ++ .long sys_flistxattr ++ .long sys_removexattr ++ .long sys_lremovexattr ++ .long sys_fremovexattr /* 235 */ + .long sys_gettid + .long sys_tkill + .rept 255-237 +--- linux-2.4.22-ac1/arch/s390x/defconfig~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/s390x/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + # CONFIG_ISA is not set + # CONFIG_EISA is not set + # CONFIG_MCA is not set +--- linux-2.4.22-ac1/arch/s390x/kernel/entry.S~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/s390x/kernel/entry.S 2003-09-25 23:57:02.000000000 +0400 +@@ -591,18 +591,18 @@ sys_call_table: + .long SYSCALL(sys_ni_syscall,sys32_fcntl64_wrapper) + .long SYSCALL(sys_readahead,sys32_readahead) + .long SYSCALL(sys_ni_syscall,sys_ni_syscall) +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 224 - reserved for setxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 225 - reserved for lsetxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 226 - reserved for fsetxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 227 - reserved for getxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 228 - reserved for lgetxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 229 - reserved for fgetxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 230 - reserved for listxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 231 - reserved for llistxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 232 - reserved for flistxattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 233 - reserved for removexattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 234 - reserved for lremovexattr */ +- .long SYSCALL(sys_ni_syscall,sys_ni_syscall) /* 235 - reserved for fremovexattr */ ++ .long SYSCALL(sys_setxattr,sys32_setxattr_wrapper) ++ .long SYSCALL(sys_lsetxattr,sys32_lsetxattr_wrapper) /* 225 */ ++ .long SYSCALL(sys_fsetxattr,sys32_fsetxattr_wrapper) ++ .long SYSCALL(sys_getxattr,sys32_getxattr_wrapper) ++ .long SYSCALL(sys_lgetxattr,sys32_lgetxattr_wrapper) ++ .long SYSCALL(sys_fgetxattr,sys32_fgetxattr_wrapper) ++ .long SYSCALL(sys_listxattr,sys32_listxattr_wrapper) /* 230 */ ++ .long SYSCALL(sys_llistxattr,sys32_llistxattr_wrapper) ++ .long SYSCALL(sys_flistxattr,sys32_flistxattr_wrapper) ++ .long SYSCALL(sys_removexattr,sys32_removexattr_wrapper) ++ .long SYSCALL(sys_lremovexattr,sys32_lremovexattr_wrapper) ++ .long SYSCALL(sys_fremovexattr,sys32_fremovexattr_wrapper)/* 235 */ + .long SYSCALL(sys_gettid,sys_gettid) + .long SYSCALL(sys_tkill,sys_tkill) + .rept 255-237 +--- linux-2.4.22-ac1/arch/s390x/kernel/wrapper32.S~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:40.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/s390x/kernel/wrapper32.S 2003-09-26 00:05:14.000000000 +0400 +@@ -1097,6 +1097,96 @@ sys32_fstat64_wrapper: + llgtr %r3,%r3 # struct stat64 * + llgfr %r4,%r4 # long + jg sys32_fstat64 # branch to system call ++ ++ .globl sys32_setxattr_wrapper ++sys32_setxattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ llgtr %r4,%r4 # void * ++ llgfr %r5,%r5 # size_t ++ lgfr %r6,%r6 # int ++ jg sys_setxattr ++ ++ .globl sys32_lsetxattr_wrapper ++sys32_lsetxattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ llgtr %r4,%r4 # void * ++ llgfr %r5,%r5 # size_t ++ lgfr %r6,%r6 # int ++ jg sys_lsetxattr ++ ++ .globl sys32_fsetxattr_wrapper ++sys32_fsetxattr_wrapper: ++ lgfr %r2,%r2 # int ++ llgtr %r3,%r3 # char * ++ llgtr %r4,%r4 # void * ++ llgfr %r5,%r5 # size_t ++ lgfr %r6,%r6 # int ++ jg sys_fsetxattr ++ ++ .globl sys32_getxattr_wrapper ++sys32_getxattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ llgtr %r4,%r4 # void * ++ llgfr %r5,%r5 # size_t ++ jg sys_getxattr ++ ++ .globl sys32_lgetxattr_wrapper ++sys32_lgetxattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ llgtr %r4,%r4 # void * ++ llgfr %r5,%r5 # size_t ++ jg sys_lgetxattr ++ ++ .globl sys32_fgetxattr_wrapper ++sys32_fgetxattr_wrapper: ++ lgfr %r2,%r2 # int ++ llgtr %r3,%r3 # char * ++ llgtr %r4,%r4 # void * ++ llgfr %r5,%r5 # size_t ++ jg sys_fgetxattr ++ ++ .globl sys32_listxattr_wrapper ++sys32_listxattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ llgfr %r4,%r4 # size_t ++ jg sys_listxattr ++ ++ .globl sys32_llistxattr_wrapper ++sys32_llistxattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ llgfr %r4,%r4 # size_t ++ jg sys_llistxattr ++ ++ .globl sys32_flistxattr_wrapper ++sys32_flistxattr_wrapper: ++ lgfr %r2,%r2 # int ++ llgtr %r3,%r3 # char * ++ llgfr %r4,%r4 # size_t ++ jg sys_flistxattr ++ ++ .globl sys32_removexattr_wrapper ++sys32_removexattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ jg sys_removexattr ++ ++ .globl sys32_lremovexattr_wrapper ++sys32_lremovexattr_wrapper: ++ llgtr %r2,%r2 # char * ++ llgtr %r3,%r3 # char * ++ jg sys_lremovexattr ++ ++ .globl sys32_fremovexattr_wrapper ++sys32_fremovexattr_wrapper: ++ lgfr %r2,%r2 # int ++ llgtr %r3,%r3 # char * ++ jg sys_fremovexattr + + .globl sys32_stime_wrapper + sys32_stime_wrapper: +--- linux-2.4.22-ac1/arch/sparc64/defconfig~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:40.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/sparc64/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + + # + # Code maturity level options +--- linux-2.4.22-ac1/arch/sparc/defconfig~xattr-0.8.54-2.4.22-rh 2002-08-03 04:39:43.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/arch/sparc/defconfig 2003-09-25 23:57:02.000000000 +0400 +@@ -1,6 +1,13 @@ + # + # Automatically generated make config: don't edit + # ++# CONFIG_EXT3_FS_XATTR is not set ++# CONFIG_EXT3_FS_XATTR_SHARING is not set ++# CONFIG_EXT3_FS_XATTR_USER is not set ++# CONFIG_EXT2_FS_XATTR is not set ++# CONFIG_EXT2_FS_XATTR_SHARING is not set ++# CONFIG_EXT2_FS_XATTR_USER is not set ++# CONFIG_FS_MBCACHE is not set + CONFIG_UID16=y + CONFIG_HIGHMEM=y + +--- linux-2.4.22-ac1/Documentation/Configure.help~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:30.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/Documentation/Configure.help 2003-09-25 23:57:02.000000000 +0400 +@@ -16145,6 +16145,39 @@ CONFIG_EXT2_FS + be compiled as a module, and so this could be dangerous. Most + everyone wants to say Y here. + ++Ext2 extended attributes ++CONFIG_EXT2_FS_XATTR ++ Extended attributes are name:value pairs associated with inodes by ++ the kernel or by users (see the attr(5) manual page, or visit ++ for details). ++ ++ If unsure, say N. ++ ++Ext2 extended attribute block sharing ++CONFIG_EXT2_FS_XATTR_SHARING ++ This options enables code for sharing identical extended attribute ++ blocks among multiple inodes. ++ ++ Usually, say Y. ++ ++Ext2 extended user attributes ++CONFIG_EXT2_FS_XATTR_USER ++ This option enables extended user attributes on ext2. Processes can ++ associate extended user attributes with inodes to store additional ++ information such as the character encoding of files, etc. (see the ++ attr(5) manual page, or visit for details). ++ ++ If unsure, say N. ++ ++Ext2 trusted extended attributes ++CONFIG_EXT2_FS_XATTR_TRUSTED ++ This option enables extended attributes on ext2 that are accessible ++ (and visible) only to users capable of CAP_SYS_ADMIN. Usually this ++ is only the super user. Trusted extended attributes are meant for ++ implementing system/security services. ++ ++ If unsure, say N. ++ + Ext3 journalling file system support (EXPERIMENTAL) + CONFIG_EXT3_FS + This is the journalling version of the Second extended file system +@@ -16177,6 +16210,39 @@ CONFIG_EXT3_FS + of your root partition (the one containing the directory /) cannot + be compiled as a module, and so this may be dangerous. + ++Ext3 extended attributes ++CONFIG_EXT3_FS_XATTR ++ Extended attributes are name:value pairs associated with inodes by ++ the kernel or by users (see the attr(5) manual page, or visit ++ for details). ++ ++ If unsure, say N. ++ ++Ext3 extended attribute block sharing ++CONFIG_EXT3_FS_XATTR_SHARING ++ This options enables code for sharing identical extended attribute ++ blocks among multiple inodes. ++ ++ Usually, say Y. ++ ++Ext3 extended user attributes ++CONFIG_EXT3_FS_XATTR_USER ++ This option enables extended user attributes on ext3. Processes can ++ associate extended user attributes with inodes to store additional ++ information such as the character encoding of files, etc. (see the ++ attr(5) manual page, or visit for details). ++ ++ If unsure, say N. ++ ++Ext3 trusted extended attributes ++CONFIG_EXT3_FS_XATTR_TRUSTED ++ This option enables extended attributes on ext3 that are accessible ++ (and visible) only to users capable of CAP_SYS_ADMIN. Usually this ++ is only the super user. Trusted extended attributes are meant for ++ implementing system/security services. ++ ++ If unsure, say N. ++ + Journal Block Device support (JBD for ext3) (EXPERIMENTAL) + CONFIG_JBD + This is a generic journalling layer for block devices. It is +--- linux-2.4.22-ac1/fs/Config.in~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:23.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/Config.in 2003-09-25 23:57:02.000000000 +0400 +@@ -29,6 +29,11 @@ dep_mbool ' Debug Befs' CONFIG_BEFS_DEB + dep_tristate 'BFS file system support (EXPERIMENTAL)' CONFIG_BFS_FS $CONFIG_EXPERIMENTAL + + tristate 'Ext3 journalling file system support' CONFIG_EXT3_FS ++dep_mbool ' Ext3 extended attributes' CONFIG_EXT3_FS_XATTR $CONFIG_EXT3_FS ++dep_bool ' Ext3 extended attribute block sharing' \ ++ CONFIG_EXT3_FS_XATTR_SHARING $CONFIG_EXT3_FS_XATTR ++dep_bool ' Ext3 extended user attributes' \ ++ CONFIG_EXT3_FS_XATTR_USER $CONFIG_EXT3_FS_XATTR + # CONFIG_JBD could be its own option (even modular), but until there are + # other users than ext3, we will simply make it be the same as CONFIG_EXT3_FS + # dep_tristate ' Journal Block Device support (JBD for ext3)' CONFIG_JBD $CONFIG_EXT3_FS +@@ -88,6 +93,11 @@ dep_mbool ' QNX4FS write support (DANGE + tristate 'ROM file system support' CONFIG_ROMFS_FS + + tristate 'Second extended fs support' CONFIG_EXT2_FS ++dep_mbool ' Ext2 extended attributes' CONFIG_EXT2_FS_XATTR $CONFIG_EXT2_FS ++dep_bool ' Ext2 extended attribute block sharing' \ ++ CONFIG_EXT2_FS_XATTR_SHARING $CONFIG_EXT2_FS_XATTR ++dep_bool ' Ext2 extended user attributes' \ ++ CONFIG_EXT2_FS_XATTR_USER $CONFIG_EXT2_FS_XATTR + + tristate 'System V/Xenix/V7/Coherent file system support' CONFIG_SYSV_FS + +@@ -164,6 +174,10 @@ else + define_tristate CONFIG_ZISOFS_FS n + fi + ++# Meta block cache for Extended Attributes (ext2/ext3) ++#tristate 'Meta block cache' CONFIG_FS_MBCACHE ++define_tristate CONFIG_FS_MBCACHE y ++ + mainmenu_option next_comment + comment 'Partition Types' + source fs/partitions/Config.in +--- linux-2.4.22-ac1/fs/ext2/file.c~xattr-0.8.54-2.4.22-rh 2001-10-11 19:05:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext2/file.c 2003-09-25 23:57:02.000000000 +0400 +@@ -20,6 +20,7 @@ + + #include + #include ++#include + #include + + /* +@@ -51,4 +52,8 @@ struct file_operations ext2_file_operati + + struct inode_operations ext2_file_inode_operations = { + truncate: ext2_truncate, ++ setxattr: ext2_setxattr, ++ getxattr: ext2_getxattr, ++ listxattr: ext2_listxattr, ++ removexattr: ext2_removexattr, + }; +--- linux-2.4.22-ac1/fs/ext2/ialloc.c~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:37.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext2/ialloc.c 2003-09-25 23:57:02.000000000 +0400 +@@ -15,6 +15,7 @@ + #include + #include + #include ++#include + #include + #include + +@@ -167,6 +168,7 @@ void ext2_free_inode (struct inode * ino + */ + if (!is_bad_inode(inode)) { + /* Quota is already initialized in iput() */ ++ ext2_xattr_delete_inode(inode); + DQUOT_FREE_INODE(inode); + DQUOT_DROP(inode); + } +--- linux-2.4.22-ac1/fs/ext2/inode.c~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:37.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext2/inode.c 2003-09-25 23:57:02.000000000 +0400 +@@ -39,6 +39,18 @@ MODULE_LICENSE("GPL"); + static int ext2_update_inode(struct inode * inode, int do_sync); + + /* ++ * Test whether an inode is a fast symlink. ++ */ ++static inline int ext2_inode_is_fast_symlink(struct inode *inode) ++{ ++ int ea_blocks = inode->u.ext2_i.i_file_acl ? ++ (inode->i_sb->s_blocksize >> 9) : 0; ++ ++ return (S_ISLNK(inode->i_mode) && ++ inode->i_blocks - ea_blocks == 0); ++} ++ ++/* + * Called at each iput() + */ + void ext2_put_inode (struct inode * inode) +@@ -53,9 +65,7 @@ void ext2_delete_inode (struct inode * i + { + lock_kernel(); + +- if (is_bad_inode(inode) || +- inode->i_ino == EXT2_ACL_IDX_INO || +- inode->i_ino == EXT2_ACL_DATA_INO) ++ if (is_bad_inode(inode)) + goto no_delete; + inode->u.ext2_i.i_dtime = CURRENT_TIME; + mark_inode_dirty(inode); +@@ -801,6 +811,8 @@ void ext2_truncate (struct inode * inode + if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || + S_ISLNK(inode->i_mode))) + return; ++ if (ext2_inode_is_fast_symlink(inode)) ++ return; + if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + return; + +@@ -903,8 +915,7 @@ void ext2_read_inode (struct inode * ino + unsigned long offset; + struct ext2_group_desc * gdp; + +- if ((inode->i_ino != EXT2_ROOT_INO && inode->i_ino != EXT2_ACL_IDX_INO && +- inode->i_ino != EXT2_ACL_DATA_INO && ++ if ((inode->i_ino != EXT2_ROOT_INO && + inode->i_ino < EXT2_FIRST_INO(inode->i_sb)) || + inode->i_ino > le32_to_cpu(inode->i_sb->u.ext2_sb.s_es->s_inodes_count)) { + ext2_error (inode->i_sb, "ext2_read_inode", +@@ -989,10 +1000,7 @@ void ext2_read_inode (struct inode * ino + for (block = 0; block < EXT2_N_BLOCKS; block++) + inode->u.ext2_i.i_data[block] = raw_inode->i_block[block]; + +- if (inode->i_ino == EXT2_ACL_IDX_INO || +- inode->i_ino == EXT2_ACL_DATA_INO) +- /* Nothing to do */ ; +- else if (S_ISREG(inode->i_mode)) { ++ if (S_ISREG(inode->i_mode)) { + inode->i_op = &ext2_file_inode_operations; + inode->i_fop = &ext2_file_operations; + inode->i_mapping->a_ops = &ext2_aops; +@@ -1001,15 +1009,17 @@ void ext2_read_inode (struct inode * ino + inode->i_fop = &ext2_dir_operations; + inode->i_mapping->a_ops = &ext2_aops; + } else if (S_ISLNK(inode->i_mode)) { +- if (!inode->i_blocks) ++ if (ext2_inode_is_fast_symlink(inode)) + inode->i_op = &ext2_fast_symlink_inode_operations; + else { +- inode->i_op = &page_symlink_inode_operations; ++ inode->i_op = &ext2_symlink_inode_operations; + inode->i_mapping->a_ops = &ext2_aops; + } +- } else ++ } else { ++ inode->i_op = &ext2_special_inode_operations; + init_special_inode(inode, inode->i_mode, + le32_to_cpu(raw_inode->i_block[0])); ++ } + brelse (bh); + inode->i_attr_flags = 0; + ext2_set_inode_flags(inode); +--- linux-2.4.22-ac1/fs/ext2/Makefile~xattr-0.8.54-2.4.22-rh 2001-10-11 19:05:18.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext2/Makefile 2003-09-25 23:57:02.000000000 +0400 +@@ -13,4 +13,8 @@ obj-y := balloc.o bitmap.o dir.o file + ioctl.o namei.o super.o symlink.o + obj-m := $(O_TARGET) + ++export-objs += xattr.o ++obj-$(CONFIG_EXT2_FS_XATTR) += xattr.o ++obj-$(CONFIG_EXT2_FS_XATTR_USER) += xattr_user.o ++ + include $(TOPDIR)/Rules.make +--- linux-2.4.22-ac1/fs/ext2/namei.c~xattr-0.8.54-2.4.22-rh 2001-10-04 09:57:36.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext2/namei.c 2003-09-25 23:57:02.000000000 +0400 +@@ -31,6 +31,7 @@ + + #include + #include ++#include + #include + + /* +@@ -136,7 +137,7 @@ static int ext2_symlink (struct inode * + + if (l > sizeof (inode->u.ext2_i.i_data)) { + /* slow symlink */ +- inode->i_op = &page_symlink_inode_operations; ++ inode->i_op = &ext2_symlink_inode_operations; + inode->i_mapping->a_ops = &ext2_aops; + err = block_symlink(inode, symname, l); + if (err) +@@ -345,4 +346,15 @@ struct inode_operations ext2_dir_inode_o + rmdir: ext2_rmdir, + mknod: ext2_mknod, + rename: ext2_rename, ++ setxattr: ext2_setxattr, ++ getxattr: ext2_getxattr, ++ listxattr: ext2_listxattr, ++ removexattr: ext2_removexattr, ++}; ++ ++struct inode_operations ext2_special_inode_operations = { ++ setxattr: ext2_setxattr, ++ getxattr: ext2_getxattr, ++ listxattr: ext2_listxattr, ++ removexattr: ext2_removexattr, + }; +--- linux-2.4.22-ac1/fs/ext2/super.c~xattr-0.8.54-2.4.22-rh 2002-11-29 02:53:15.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext2/super.c 2003-09-25 23:57:02.000000000 +0400 +@@ -21,6 +21,7 @@ + #include + #include + #include ++#include + #include + #include + #include +@@ -125,6 +126,7 @@ void ext2_put_super (struct super_block + int db_count; + int i; + ++ ext2_xattr_put_super(sb); + if (!(sb->s_flags & MS_RDONLY)) { + struct ext2_super_block *es = EXT2_SB(sb)->s_es; + +@@ -175,6 +177,13 @@ static int parse_options (char * options + this_char = strtok (NULL, ",")) { + if ((value = strchr (this_char, '=')) != NULL) + *value++ = 0; ++#ifdef CONFIG_EXT2_FS_XATTR_USER ++ if (!strcmp (this_char, "user_xattr")) ++ set_opt (*mount_options, XATTR_USER); ++ else if (!strcmp (this_char, "nouser_xattr")) ++ clear_opt (*mount_options, XATTR_USER); ++ else ++#endif + if (!strcmp (this_char, "bsddf")) + clear_opt (*mount_options, MINIX_DF); + else if (!strcmp (this_char, "nouid32")) { +@@ -424,6 +433,9 @@ struct super_block * ext2_read_super (st + blocksize = BLOCK_SIZE; + + sb->u.ext2_sb.s_mount_opt = 0; ++#ifdef CONFIG_EXT2_FS_XATTR_USER ++ /* set_opt (sb->u.ext2_sb.s_mount_opt, XATTR_USER); */ ++#endif + if (!parse_options ((char *) data, &sb_block, &resuid, &resgid, + &sb->u.ext2_sb.s_mount_opt)) { + return NULL; +@@ -813,12 +825,27 @@ static DECLARE_FSTYPE_DEV(ext2_fs_type, + + static int __init init_ext2_fs(void) + { +- return register_filesystem(&ext2_fs_type); ++ int error = init_ext2_xattr(); ++ if (error) ++ return error; ++ error = init_ext2_xattr_user(); ++ if (error) ++ goto fail; ++ error = register_filesystem(&ext2_fs_type); ++ if (!error) ++ return 0; ++ ++ exit_ext2_xattr_user(); ++fail: ++ exit_ext2_xattr(); ++ return error; + } + + static void __exit exit_ext2_fs(void) + { + unregister_filesystem(&ext2_fs_type); ++ exit_ext2_xattr_user(); ++ exit_ext2_xattr(); + } + + EXPORT_NO_SYMBOLS; +--- linux-2.4.22-ac1/fs/ext2/symlink.c~xattr-0.8.54-2.4.22-rh 2000-09-28 00:41:33.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext2/symlink.c 2003-09-25 23:57:02.000000000 +0400 +@@ -19,6 +19,7 @@ + + #include + #include ++#include + + static int ext2_readlink(struct dentry *dentry, char *buffer, int buflen) + { +@@ -32,7 +33,20 @@ static int ext2_follow_link(struct dentr + return vfs_follow_link(nd, s); + } + ++struct inode_operations ext2_symlink_inode_operations = { ++ readlink: page_readlink, ++ follow_link: page_follow_link, ++ setxattr: ext2_setxattr, ++ getxattr: ext2_getxattr, ++ listxattr: ext2_listxattr, ++ removexattr: ext2_removexattr, ++}; ++ + struct inode_operations ext2_fast_symlink_inode_operations = { + readlink: ext2_readlink, + follow_link: ext2_follow_link, ++ setxattr: ext2_setxattr, ++ getxattr: ext2_getxattr, ++ listxattr: ext2_listxattr, ++ removexattr: ext2_removexattr, + }; +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext2/xattr.c 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,1212 @@ ++/* ++ * linux/fs/ext2/xattr.c ++ * ++ * Copyright (C) 2001 by Andreas Gruenbacher, ++ * ++ * Fix by Harrison Xing . ++ * Extended attributes for symlinks and special files added per ++ * suggestion of Luka Renko . ++ */ ++ ++/* ++ * Extended attributes are stored on disk blocks allocated outside of ++ * any inode. The i_file_acl field is then made to point to this allocated ++ * block. If all extended attributes of an inode are identical, these ++ * inodes may share the same extended attribute block. Such situations ++ * are automatically detected by keeping a cache of recent attribute block ++ * numbers and hashes over the block's contents in memory. ++ * ++ * ++ * Extended attribute block layout: ++ * ++ * +------------------+ ++ * | header | ++ * | entry 1 | | ++ * | entry 2 | | growing downwards ++ * | entry 3 | v ++ * | four null bytes | ++ * | . . . | ++ * | value 1 | ^ ++ * | value 3 | | growing upwards ++ * | value 2 | | ++ * +------------------+ ++ * ++ * The block header is followed by multiple entry descriptors. These entry ++ * descriptors are variable in size, and alligned to EXT2_XATTR_PAD ++ * byte boundaries. The entry descriptors are sorted by attribute name, ++ * so that two extended attribute blocks can be compared efficiently. ++ * ++ * Attribute values are aligned to the end of the block, stored in ++ * no specific order. They are also padded to EXT2_XATTR_PAD byte ++ * boundaries. No additional gaps are left between them. ++ * ++ * Locking strategy ++ * ---------------- ++ * The VFS already holds the BKL and the inode->i_sem semaphore when any of ++ * the xattr inode operations are called, so we are guaranteed that only one ++ * processes accesses extended attributes of an inode at any time. ++ * ++ * For writing we also grab the ext2_xattr_sem semaphore. This ensures that ++ * only a single process is modifying an extended attribute block, even ++ * if the block is shared among inodes. ++ * ++ * Note for porting to 2.5 ++ * ----------------------- ++ * The BKL will no longer be held in the xattr inode operations. ++ */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++ ++/* These symbols may be needed by a module. */ ++EXPORT_SYMBOL(ext2_xattr_register); ++EXPORT_SYMBOL(ext2_xattr_unregister); ++EXPORT_SYMBOL(ext2_xattr_get); ++EXPORT_SYMBOL(ext2_xattr_list); ++EXPORT_SYMBOL(ext2_xattr_set); ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,0) ++# define mark_buffer_dirty(bh) mark_buffer_dirty(bh, 1) ++#endif ++ ++#define HDR(bh) ((struct ext2_xattr_header *)((bh)->b_data)) ++#define ENTRY(ptr) ((struct ext2_xattr_entry *)(ptr)) ++#define FIRST_ENTRY(bh) ENTRY(HDR(bh)+1) ++#define IS_LAST_ENTRY(entry) (*(__u32 *)(entry) == 0) ++ ++#ifdef EXT2_XATTR_DEBUG ++# define ea_idebug(inode, f...) do { \ ++ printk(KERN_DEBUG "inode %s:%ld: ", \ ++ kdevname(inode->i_dev), inode->i_ino); \ ++ printk(f); \ ++ printk("\n"); \ ++ } while (0) ++# define ea_bdebug(bh, f...) do { \ ++ printk(KERN_DEBUG "block %s:%ld: ", \ ++ kdevname(bh->b_dev), bh->b_blocknr); \ ++ printk(f); \ ++ printk("\n"); \ ++ } while (0) ++#else ++# define ea_idebug(f...) ++# define ea_bdebug(f...) ++#endif ++ ++static int ext2_xattr_set2(struct inode *, struct buffer_head *, ++ struct ext2_xattr_header *); ++ ++#ifdef CONFIG_EXT2_FS_XATTR_SHARING ++ ++static int ext2_xattr_cache_insert(struct buffer_head *); ++static struct buffer_head *ext2_xattr_cache_find(struct inode *, ++ struct ext2_xattr_header *); ++static void ext2_xattr_cache_remove(struct buffer_head *); ++static void ext2_xattr_rehash(struct ext2_xattr_header *, ++ struct ext2_xattr_entry *); ++ ++static struct mb_cache *ext2_xattr_cache; ++ ++#else ++# define ext2_xattr_cache_insert(bh) 0 ++# define ext2_xattr_cache_find(inode, header) NULL ++# define ext2_xattr_cache_remove(bh) while(0) {} ++# define ext2_xattr_rehash(header, entry) while(0) {} ++#endif ++ ++/* ++ * If a file system does not share extended attributes among inodes, ++ * we should not need the ext2_xattr_sem semaphore. However, the ++ * filesystem may still contain shared blocks, so we always take ++ * the lock. ++ */ ++ ++DECLARE_MUTEX(ext2_xattr_sem); ++ ++static inline int ++ext2_xattr_new_block(struct inode *inode, int * errp, int force) ++{ ++ struct super_block *sb = inode->i_sb; ++ int goal = le32_to_cpu(EXT2_SB(sb)->s_es->s_first_data_block) + ++ EXT2_I(inode)->i_block_group * EXT2_BLOCKS_PER_GROUP(sb); ++ ++ /* How can we enforce the allocation? */ ++ int block = ext2_new_block(inode, goal, 0, 0, errp); ++#ifdef OLD_QUOTAS ++ if (!*errp) ++ inode->i_blocks += inode->i_sb->s_blocksize >> 9; ++#endif ++ return block; ++} ++ ++static inline int ++ext2_xattr_quota_alloc(struct inode *inode, int force) ++{ ++ /* How can we enforce the allocation? */ ++#ifdef OLD_QUOTAS ++ int error = DQUOT_ALLOC_BLOCK(inode->i_sb, inode, 1); ++ if (!error) ++ inode->i_blocks += inode->i_sb->s_blocksize >> 9; ++#else ++ int error = DQUOT_ALLOC_BLOCK(inode, 1); ++#endif ++ return error; ++} ++ ++#ifdef OLD_QUOTAS ++ ++static inline void ++ext2_xattr_quota_free(struct inode *inode) ++{ ++ DQUOT_FREE_BLOCK(inode->i_sb, inode, 1); ++ inode->i_blocks -= inode->i_sb->s_blocksize >> 9; ++} ++ ++static inline void ++ext2_xattr_free_block(struct inode * inode, unsigned long block) ++{ ++ ext2_free_blocks(inode, block, 1); ++ inode->i_blocks -= inode->i_sb->s_blocksize >> 9; ++} ++ ++#else ++# define ext2_xattr_quota_free(inode) \ ++ DQUOT_FREE_BLOCK(inode, 1) ++# define ext2_xattr_free_block(inode, block) \ ++ ext2_free_blocks(inode, block, 1) ++#endif ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,18) ++ ++static inline struct buffer_head * ++sb_bread(struct super_block *sb, int block) ++{ ++ return bread(sb->s_dev, block, sb->s_blocksize); ++} ++ ++static inline struct buffer_head * ++sb_getblk(struct super_block *sb, int block) ++{ ++ return getblk(sb->s_dev, block, sb->s_blocksize); ++} ++ ++#endif ++ ++struct ext2_xattr_handler *ext2_xattr_handlers[EXT2_XATTR_INDEX_MAX]; ++rwlock_t ext2_handler_lock = RW_LOCK_UNLOCKED; ++ ++int ++ext2_xattr_register(int name_index, struct ext2_xattr_handler *handler) ++{ ++ int error = -EINVAL; ++ ++ if (name_index > 0 && name_index <= EXT2_XATTR_INDEX_MAX) { ++ write_lock(&ext2_handler_lock); ++ if (!ext2_xattr_handlers[name_index-1]) { ++ ext2_xattr_handlers[name_index-1] = handler; ++ error = 0; ++ } ++ write_unlock(&ext2_handler_lock); ++ } ++ return error; ++} ++ ++void ++ext2_xattr_unregister(int name_index, struct ext2_xattr_handler *handler) ++{ ++ if (name_index > 0 || name_index <= EXT2_XATTR_INDEX_MAX) { ++ write_lock(&ext2_handler_lock); ++ ext2_xattr_handlers[name_index-1] = NULL; ++ write_unlock(&ext2_handler_lock); ++ } ++} ++ ++static inline const char * ++strcmp_prefix(const char *a, const char *a_prefix) ++{ ++ while (*a_prefix && *a == *a_prefix) { ++ a++; ++ a_prefix++; ++ } ++ return *a_prefix ? NULL : a; ++} ++ ++/* ++ * Decode the extended attribute name, and translate it into ++ * the name_index and name suffix. ++ */ ++static struct ext2_xattr_handler * ++ext2_xattr_resolve_name(const char **name) ++{ ++ struct ext2_xattr_handler *handler = NULL; ++ int i; ++ ++ if (!*name) ++ return NULL; ++ read_lock(&ext2_handler_lock); ++ for (i=0; iprefix); ++ if (n) { ++ handler = ext2_xattr_handlers[i]; ++ *name = n; ++ break; ++ } ++ } ++ } ++ read_unlock(&ext2_handler_lock); ++ return handler; ++} ++ ++static inline struct ext2_xattr_handler * ++ext2_xattr_handler(int name_index) ++{ ++ struct ext2_xattr_handler *handler = NULL; ++ if (name_index > 0 && name_index <= EXT2_XATTR_INDEX_MAX) { ++ read_lock(&ext2_handler_lock); ++ handler = ext2_xattr_handlers[name_index-1]; ++ read_unlock(&ext2_handler_lock); ++ } ++ return handler; ++} ++ ++/* ++ * Inode operation getxattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++ssize_t ++ext2_getxattr(struct dentry *dentry, const char *name, ++ void *buffer, size_t size) ++{ ++ struct ext2_xattr_handler *handler; ++ struct inode *inode = dentry->d_inode; ++ ++ handler = ext2_xattr_resolve_name(&name); ++ if (!handler) ++ return -ENOTSUP; ++ return handler->get(inode, name, buffer, size); ++} ++ ++/* ++ * Inode operation listxattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++ssize_t ++ext2_listxattr(struct dentry *dentry, char *buffer, size_t size) ++{ ++ return ext2_xattr_list(dentry->d_inode, buffer, size); ++} ++ ++/* ++ * Inode operation setxattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++int ++ext2_setxattr(struct dentry *dentry, const char *name, ++ const void *value, size_t size, int flags) ++{ ++ struct ext2_xattr_handler *handler; ++ struct inode *inode = dentry->d_inode; ++ ++ if (size == 0) ++ value = ""; /* empty EA, do not remove */ ++ handler = ext2_xattr_resolve_name(&name); ++ if (!handler) ++ return -ENOTSUP; ++ return handler->set(inode, name, value, size, flags); ++} ++ ++/* ++ * Inode operation removexattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++int ++ext2_removexattr(struct dentry *dentry, const char *name) ++{ ++ struct ext2_xattr_handler *handler; ++ struct inode *inode = dentry->d_inode; ++ ++ handler = ext2_xattr_resolve_name(&name); ++ if (!handler) ++ return -ENOTSUP; ++ return handler->set(inode, name, NULL, 0, XATTR_REPLACE); ++} ++ ++/* ++ * ext2_xattr_get() ++ * ++ * Copy an extended attribute into the buffer ++ * provided, or compute the buffer size required. ++ * Buffer is NULL to compute the size of the buffer required. ++ * ++ * Returns a negative error number on failure, or the number of bytes ++ * used / required on success. ++ */ ++int ++ext2_xattr_get(struct inode *inode, int name_index, const char *name, ++ void *buffer, size_t buffer_size) ++{ ++ struct buffer_head *bh = NULL; ++ struct ext2_xattr_entry *entry; ++ unsigned int block, size; ++ char *end; ++ int name_len, error; ++ ++ ea_idebug(inode, "name=%d.%s, buffer=%p, buffer_size=%ld", ++ name_index, name, buffer, (long)buffer_size); ++ ++ if (name == NULL) ++ return -EINVAL; ++ if (!EXT2_I(inode)->i_file_acl) ++ return -ENOATTR; ++ block = EXT2_I(inode)->i_file_acl; ++ ea_idebug(inode, "reading block %d", block); ++ bh = sb_bread(inode->i_sb, block); ++ if (!bh) ++ return -EIO; ++ ea_bdebug(bh, "b_count=%d, refcount=%d", ++ atomic_read(&(bh->b_count)), le32_to_cpu(HDR(bh)->h_refcount)); ++ end = bh->b_data + bh->b_size; ++ if (HDR(bh)->h_magic != cpu_to_le32(EXT2_XATTR_MAGIC) || ++ HDR(bh)->h_blocks != cpu_to_le32(1)) { ++bad_block: ext2_error(inode->i_sb, "ext2_xattr_get", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ /* find named attribute */ ++ name_len = strlen(name); ++ ++ error = -ERANGE; ++ if (name_len > 255) ++ goto cleanup; ++ entry = FIRST_ENTRY(bh); ++ while (!IS_LAST_ENTRY(entry)) { ++ struct ext2_xattr_entry *next = ++ EXT2_XATTR_NEXT(entry); ++ if ((char *)next >= end) ++ goto bad_block; ++ if (name_index == entry->e_name_index && ++ name_len == entry->e_name_len && ++ memcmp(name, entry->e_name, name_len) == 0) ++ goto found; ++ entry = next; ++ } ++ /* Check the remaining name entries */ ++ while (!IS_LAST_ENTRY(entry)) { ++ struct ext2_xattr_entry *next = ++ EXT2_XATTR_NEXT(entry); ++ if ((char *)next >= end) ++ goto bad_block; ++ entry = next; ++ } ++ if (ext2_xattr_cache_insert(bh)) ++ ea_idebug(inode, "cache insert failed"); ++ error = -ENOATTR; ++ goto cleanup; ++found: ++ /* check the buffer size */ ++ if (entry->e_value_block != 0) ++ goto bad_block; ++ size = le32_to_cpu(entry->e_value_size); ++ if (size > inode->i_sb->s_blocksize || ++ le16_to_cpu(entry->e_value_offs) + size > inode->i_sb->s_blocksize) ++ goto bad_block; ++ ++ if (ext2_xattr_cache_insert(bh)) ++ ea_idebug(inode, "cache insert failed"); ++ if (buffer) { ++ error = -ERANGE; ++ if (size > buffer_size) ++ goto cleanup; ++ /* return value of attribute */ ++ memcpy(buffer, bh->b_data + le16_to_cpu(entry->e_value_offs), ++ size); ++ } ++ error = size; ++ ++cleanup: ++ brelse(bh); ++ ++ return error; ++} ++ ++/* ++ * ext2_xattr_list() ++ * ++ * Copy a list of attribute names into the buffer ++ * provided, or compute the buffer size required. ++ * Buffer is NULL to compute the size of the buffer required. ++ * ++ * Returns a negative error number on failure, or the number of bytes ++ * used / required on success. ++ */ ++int ++ext2_xattr_list(struct inode *inode, char *buffer, size_t buffer_size) ++{ ++ struct buffer_head *bh = NULL; ++ struct ext2_xattr_entry *entry; ++ unsigned int block, size = 0; ++ char *buf, *end; ++ int error; ++ ++ ea_idebug(inode, "buffer=%p, buffer_size=%ld", ++ buffer, (long)buffer_size); ++ ++ if (!EXT2_I(inode)->i_file_acl) ++ return 0; ++ block = EXT2_I(inode)->i_file_acl; ++ ea_idebug(inode, "reading block %d", block); ++ bh = sb_bread(inode->i_sb, block); ++ if (!bh) ++ return -EIO; ++ ea_bdebug(bh, "b_count=%d, refcount=%d", ++ atomic_read(&(bh->b_count)), le32_to_cpu(HDR(bh)->h_refcount)); ++ end = bh->b_data + bh->b_size; ++ if (HDR(bh)->h_magic != cpu_to_le32(EXT2_XATTR_MAGIC) || ++ HDR(bh)->h_blocks != cpu_to_le32(1)) { ++bad_block: ext2_error(inode->i_sb, "ext2_xattr_list", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ /* compute the size required for the list of attribute names */ ++ for (entry = FIRST_ENTRY(bh); !IS_LAST_ENTRY(entry); ++ entry = EXT2_XATTR_NEXT(entry)) { ++ struct ext2_xattr_handler *handler; ++ struct ext2_xattr_entry *next = ++ EXT2_XATTR_NEXT(entry); ++ if ((char *)next >= end) ++ goto bad_block; ++ ++ handler = ext2_xattr_handler(entry->e_name_index); ++ if (handler) ++ size += handler->list(NULL, inode, entry->e_name, ++ entry->e_name_len); ++ } ++ ++ if (ext2_xattr_cache_insert(bh)) ++ ea_idebug(inode, "cache insert failed"); ++ if (!buffer) { ++ error = size; ++ goto cleanup; ++ } else { ++ error = -ERANGE; ++ if (size > buffer_size) ++ goto cleanup; ++ } ++ ++ /* list the attribute names */ ++ buf = buffer; ++ for (entry = FIRST_ENTRY(bh); !IS_LAST_ENTRY(entry); ++ entry = EXT2_XATTR_NEXT(entry)) { ++ struct ext2_xattr_handler *handler; ++ ++ handler = ext2_xattr_handler(entry->e_name_index); ++ if (handler) ++ buf += handler->list(buf, inode, entry->e_name, ++ entry->e_name_len); ++ } ++ error = size; ++ ++cleanup: ++ brelse(bh); ++ ++ return error; ++} ++ ++/* ++ * If the EXT2_FEATURE_COMPAT_EXT_ATTR feature of this file system is ++ * not set, set it. ++ */ ++static void ext2_xattr_update_super_block(struct super_block *sb) ++{ ++ if (EXT2_HAS_COMPAT_FEATURE(sb, EXT2_FEATURE_COMPAT_EXT_ATTR)) ++ return; ++ ++ lock_super(sb); ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,0) ++ EXT2_SB(sb)->s_feature_compat |= EXT2_FEATURE_COMPAT_EXT_ATTR; ++#endif ++ EXT2_SB(sb)->s_es->s_feature_compat |= ++ cpu_to_le32(EXT2_FEATURE_COMPAT_EXT_ATTR); ++ sb->s_dirt = 1; ++ mark_buffer_dirty(EXT2_SB(sb)->s_sbh); ++ unlock_super(sb); ++} ++ ++/* ++ * ext2_xattr_set() ++ * ++ * Create, replace or remove an extended attribute for this inode. Buffer ++ * is NULL to remove an existing extended attribute, and non-NULL to ++ * either replace an existing extended attribute, or create a new extended ++ * attribute. The flags XATTR_REPLACE and XATTR_CREATE ++ * specify that an extended attribute must exist and must not exist ++ * previous to the call, respectively. ++ * ++ * Returns 0, or a negative error number on failure. ++ */ ++int ++ext2_xattr_set(struct inode *inode, int name_index, const char *name, ++ const void *value, size_t value_len, int flags) ++{ ++ struct super_block *sb = inode->i_sb; ++ struct buffer_head *bh = NULL; ++ struct ext2_xattr_header *header = NULL; ++ struct ext2_xattr_entry *here, *last; ++ unsigned int name_len; ++ int block = EXT2_I(inode)->i_file_acl; ++ int min_offs = sb->s_blocksize, not_found = 1, free, error; ++ char *end; ++ ++ /* ++ * header -- Points either into bh, or to a temporarily ++ * allocated buffer. ++ * here -- The named entry found, or the place for inserting, within ++ * the block pointed to by header. ++ * last -- Points right after the last named entry within the block ++ * pointed to by header. ++ * min_offs -- The offset of the first value (values are aligned ++ * towards the end of the block). ++ * end -- Points right after the block pointed to by header. ++ */ ++ ++ ea_idebug(inode, "name=%d.%s, value=%p, value_len=%ld", ++ name_index, name, value, (long)value_len); ++ ++ if (IS_RDONLY(inode)) ++ return -EROFS; ++ if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) ++ return -EPERM; ++ if (value == NULL) ++ value_len = 0; ++ if (name == NULL) ++ return -EINVAL; ++ name_len = strlen(name); ++ if (name_len > 255 || value_len > sb->s_blocksize) ++ return -ERANGE; ++ down(&ext2_xattr_sem); ++ ++ if (block) { ++ /* The inode already has an extended attribute block. */ ++ ++ bh = sb_bread(sb, block); ++ error = -EIO; ++ if (!bh) ++ goto cleanup; ++ ea_bdebug(bh, "b_count=%d, refcount=%d", ++ atomic_read(&(bh->b_count)), ++ le32_to_cpu(HDR(bh)->h_refcount)); ++ header = HDR(bh); ++ end = bh->b_data + bh->b_size; ++ if (header->h_magic != cpu_to_le32(EXT2_XATTR_MAGIC) || ++ header->h_blocks != cpu_to_le32(1)) { ++bad_block: ext2_error(sb, "ext2_xattr_set", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ /* Find the named attribute. */ ++ here = FIRST_ENTRY(bh); ++ while (!IS_LAST_ENTRY(here)) { ++ struct ext2_xattr_entry *next = EXT2_XATTR_NEXT(here); ++ if ((char *)next >= end) ++ goto bad_block; ++ if (!here->e_value_block && here->e_value_size) { ++ int offs = le16_to_cpu(here->e_value_offs); ++ if (offs < min_offs) ++ min_offs = offs; ++ } ++ not_found = name_index - here->e_name_index; ++ if (!not_found) ++ not_found = name_len - here->e_name_len; ++ if (!not_found) ++ not_found = memcmp(name, here->e_name,name_len); ++ if (not_found <= 0) ++ break; ++ here = next; ++ } ++ last = here; ++ /* We still need to compute min_offs and last. */ ++ while (!IS_LAST_ENTRY(last)) { ++ struct ext2_xattr_entry *next = EXT2_XATTR_NEXT(last); ++ if ((char *)next >= end) ++ goto bad_block; ++ if (!last->e_value_block && last->e_value_size) { ++ int offs = le16_to_cpu(last->e_value_offs); ++ if (offs < min_offs) ++ min_offs = offs; ++ } ++ last = next; ++ } ++ ++ /* Check whether we have enough space left. */ ++ free = min_offs - ((char*)last - (char*)header) - sizeof(__u32); ++ } else { ++ /* We will use a new extended attribute block. */ ++ free = sb->s_blocksize - ++ sizeof(struct ext2_xattr_header) - sizeof(__u32); ++ here = last = NULL; /* avoid gcc uninitialized warning. */ ++ } ++ ++ if (not_found) { ++ /* Request to remove a nonexistent attribute? */ ++ error = -ENOATTR; ++ if (flags & XATTR_REPLACE) ++ goto cleanup; ++ error = 0; ++ if (value == NULL) ++ goto cleanup; ++ else ++ free -= EXT2_XATTR_LEN(name_len); ++ } else { ++ /* Request to create an existing attribute? */ ++ error = -EEXIST; ++ if (flags & XATTR_CREATE) ++ goto cleanup; ++ if (!here->e_value_block && here->e_value_size) { ++ unsigned int size = le32_to_cpu(here->e_value_size); ++ ++ if (le16_to_cpu(here->e_value_offs) + size > ++ sb->s_blocksize || size > sb->s_blocksize) ++ goto bad_block; ++ free += EXT2_XATTR_SIZE(size); ++ } ++ } ++ free -= EXT2_XATTR_SIZE(value_len); ++ error = -ENOSPC; ++ if (free < 0) ++ goto cleanup; ++ ++ /* Here we know that we can set the new attribute. */ ++ ++ if (header) { ++ if (header->h_refcount == cpu_to_le32(1)) { ++ ea_bdebug(bh, "modifying in-place"); ++ ext2_xattr_cache_remove(bh); ++ } else { ++ int offset; ++ ++ ea_bdebug(bh, "cloning"); ++ header = kmalloc(bh->b_size, GFP_KERNEL); ++ error = -ENOMEM; ++ if (header == NULL) ++ goto cleanup; ++ memcpy(header, HDR(bh), bh->b_size); ++ header->h_refcount = cpu_to_le32(1); ++ offset = (char *)header - bh->b_data; ++ here = ENTRY((char *)here + offset); ++ last = ENTRY((char *)last + offset); ++ } ++ } else { ++ /* Allocate a buffer where we construct the new block. */ ++ header = kmalloc(sb->s_blocksize, GFP_KERNEL); ++ error = -ENOMEM; ++ if (header == NULL) ++ goto cleanup; ++ memset(header, 0, sb->s_blocksize); ++ end = (char *)header + sb->s_blocksize; ++ header->h_magic = cpu_to_le32(EXT2_XATTR_MAGIC); ++ header->h_blocks = header->h_refcount = cpu_to_le32(1); ++ last = here = ENTRY(header+1); ++ } ++ ++ if (not_found) { ++ /* Insert the new name. */ ++ int size = EXT2_XATTR_LEN(name_len); ++ int rest = (char *)last - (char *)here; ++ memmove((char *)here + size, here, rest); ++ memset(here, 0, size); ++ here->e_name_index = name_index; ++ here->e_name_len = name_len; ++ memcpy(here->e_name, name, name_len); ++ } else { ++ /* Remove the old value. */ ++ if (!here->e_value_block && here->e_value_size) { ++ char *first_val = (char *)header + min_offs; ++ int offs = le16_to_cpu(here->e_value_offs); ++ char *val = (char *)header + offs; ++ size_t size = EXT2_XATTR_SIZE( ++ le32_to_cpu(here->e_value_size)); ++ memmove(first_val + size, first_val, val - first_val); ++ memset(first_val, 0, size); ++ here->e_value_offs = 0; ++ min_offs += size; ++ ++ /* Adjust all value offsets. */ ++ last = ENTRY(header+1); ++ while (!IS_LAST_ENTRY(last)) { ++ int o = le16_to_cpu(last->e_value_offs); ++ if (!last->e_value_block && o < offs) ++ last->e_value_offs = ++ cpu_to_le16(o + size); ++ last = EXT2_XATTR_NEXT(last); ++ } ++ } ++ if (value == NULL) { ++ /* Remove this attribute. */ ++ if (EXT2_XATTR_NEXT(ENTRY(header+1)) == last) { ++ /* This block is now empty. */ ++ error = ext2_xattr_set2(inode, bh, NULL); ++ goto cleanup; ++ } else { ++ /* Remove the old name. */ ++ int size = EXT2_XATTR_LEN(name_len); ++ last = ENTRY((char *)last - size); ++ memmove(here, (char*)here + size, ++ (char*)last - (char*)here); ++ memset(last, 0, size); ++ } ++ } ++ } ++ ++ if (value != NULL) { ++ /* Insert the new value. */ ++ here->e_value_size = cpu_to_le32(value_len); ++ if (value_len) { ++ size_t size = EXT2_XATTR_SIZE(value_len); ++ char *val = (char *)header + min_offs - size; ++ here->e_value_offs = ++ cpu_to_le16((char *)val - (char *)header); ++ memset(val + size - EXT2_XATTR_PAD, 0, ++ EXT2_XATTR_PAD); /* Clear the pad bytes. */ ++ memcpy(val, value, value_len); ++ } ++ } ++ ext2_xattr_rehash(header, here); ++ ++ error = ext2_xattr_set2(inode, bh, header); ++ ++cleanup: ++ brelse(bh); ++ if (!(bh && header == HDR(bh))) ++ kfree(header); ++ up(&ext2_xattr_sem); ++ ++ return error; ++} ++ ++/* ++ * Second half of ext2_xattr_set(): Update the file system. ++ */ ++static int ++ext2_xattr_set2(struct inode *inode, struct buffer_head *old_bh, ++ struct ext2_xattr_header *header) ++{ ++ struct super_block *sb = inode->i_sb; ++ struct buffer_head *new_bh = NULL; ++ int error; ++ ++ if (header) { ++ new_bh = ext2_xattr_cache_find(inode, header); ++ if (new_bh) { ++ /* ++ * We found an identical block in the cache. ++ * The old block will be released after updating ++ * the inode. ++ */ ++ ea_bdebug(old_bh, "reusing block %ld", ++ new_bh->b_blocknr); ++ ++ error = -EDQUOT; ++ if (ext2_xattr_quota_alloc(inode, 1)) ++ goto cleanup; ++ ++ HDR(new_bh)->h_refcount = cpu_to_le32( ++ le32_to_cpu(HDR(new_bh)->h_refcount) + 1); ++ ea_bdebug(new_bh, "refcount now=%d", ++ le32_to_cpu(HDR(new_bh)->h_refcount)); ++ } else if (old_bh && header == HDR(old_bh)) { ++ /* Keep this block. */ ++ new_bh = old_bh; ++ ext2_xattr_cache_insert(new_bh); ++ } else { ++ /* We need to allocate a new block */ ++ int force = EXT2_I(inode)->i_file_acl != 0; ++ int block = ext2_xattr_new_block(inode, &error, force); ++ if (error) ++ goto cleanup; ++ ea_idebug(inode, "creating block %d", block); ++ ++ new_bh = sb_getblk(sb, block); ++ if (!new_bh) { ++ ext2_xattr_free_block(inode, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ lock_buffer(new_bh); ++ memcpy(new_bh->b_data, header, new_bh->b_size); ++ mark_buffer_uptodate(new_bh, 1); ++ unlock_buffer(new_bh); ++ ext2_xattr_cache_insert(new_bh); ++ ++ ext2_xattr_update_super_block(sb); ++ } ++ mark_buffer_dirty(new_bh); ++ if (IS_SYNC(inode)) { ++ ll_rw_block(WRITE, 1, &new_bh); ++ wait_on_buffer(new_bh); ++ error = -EIO; ++ if (buffer_req(new_bh) && !buffer_uptodate(new_bh)) ++ goto cleanup; ++ } ++ } ++ ++ /* Update the inode. */ ++ EXT2_I(inode)->i_file_acl = new_bh ? new_bh->b_blocknr : 0; ++ inode->i_ctime = CURRENT_TIME; ++ if (IS_SYNC(inode)) { ++ error = ext2_sync_inode (inode); ++ if (error) ++ goto cleanup; ++ } else ++ mark_inode_dirty(inode); ++ ++ error = 0; ++ if (old_bh && old_bh != new_bh) { ++ /* ++ * If there was an old block, and we are not still using it, ++ * we now release the old block. ++ */ ++ unsigned int refcount = le32_to_cpu(HDR(old_bh)->h_refcount); ++ ++ if (refcount == 1) { ++ /* Free the old block. */ ++ ea_bdebug(old_bh, "freeing"); ++ ext2_xattr_free_block(inode, old_bh->b_blocknr); ++ mark_buffer_clean(old_bh); ++ } else { ++ /* Decrement the refcount only. */ ++ refcount--; ++ HDR(old_bh)->h_refcount = cpu_to_le32(refcount); ++ ext2_xattr_quota_free(inode); ++ mark_buffer_dirty(old_bh); ++ ea_bdebug(old_bh, "refcount now=%d", refcount); ++ } ++ } ++ ++cleanup: ++ if (old_bh != new_bh) ++ brelse(new_bh); ++ ++ return error; ++} ++ ++/* ++ * ext2_xattr_delete_inode() ++ * ++ * Free extended attribute resources associated with this inode. This ++ * is called immediately before an inode is freed. ++ */ ++void ++ext2_xattr_delete_inode(struct inode *inode) ++{ ++ struct buffer_head *bh; ++ unsigned int block = EXT2_I(inode)->i_file_acl; ++ ++ if (!block) ++ return; ++ down(&ext2_xattr_sem); ++ ++ bh = sb_bread(inode->i_sb, block); ++ if (!bh) { ++ ext2_error(inode->i_sb, "ext2_xattr_delete_inode", ++ "inode %ld: block %d read error", inode->i_ino, block); ++ goto cleanup; ++ } ++ ea_bdebug(bh, "b_count=%d", atomic_read(&(bh->b_count))); ++ if (HDR(bh)->h_magic != cpu_to_le32(EXT2_XATTR_MAGIC) || ++ HDR(bh)->h_blocks != cpu_to_le32(1)) { ++ ext2_error(inode->i_sb, "ext2_xattr_delete_inode", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ goto cleanup; ++ } ++ ea_bdebug(bh, "refcount now=%d", le32_to_cpu(HDR(bh)->h_refcount) - 1); ++ if (HDR(bh)->h_refcount == cpu_to_le32(1)) { ++ ext2_xattr_cache_remove(bh); ++ ext2_xattr_free_block(inode, block); ++ bforget(bh); ++ bh = NULL; ++ } else { ++ HDR(bh)->h_refcount = cpu_to_le32( ++ le32_to_cpu(HDR(bh)->h_refcount) - 1); ++ mark_buffer_dirty(bh); ++ if (IS_SYNC(inode)) { ++ ll_rw_block(WRITE, 1, &bh); ++ wait_on_buffer(bh); ++ } ++ ext2_xattr_quota_free(inode); ++ } ++ EXT2_I(inode)->i_file_acl = 0; ++ ++cleanup: ++ brelse(bh); ++ up(&ext2_xattr_sem); ++} ++ ++/* ++ * ext2_xattr_put_super() ++ * ++ * This is called when a file system is unmounted. ++ */ ++void ++ext2_xattr_put_super(struct super_block *sb) ++{ ++#ifdef CONFIG_EXT2_FS_XATTR_SHARING ++ mb_cache_shrink(ext2_xattr_cache, sb->s_dev); ++#endif ++} ++ ++#ifdef CONFIG_EXT2_FS_XATTR_SHARING ++ ++/* ++ * ext2_xattr_cache_insert() ++ * ++ * Create a new entry in the extended attribute cache, and insert ++ * it unless such an entry is already in the cache. ++ * ++ * Returns 0, or a negative error number on failure. ++ */ ++static int ++ext2_xattr_cache_insert(struct buffer_head *bh) ++{ ++ __u32 hash = le32_to_cpu(HDR(bh)->h_hash); ++ struct mb_cache_entry *ce; ++ int error; ++ ++ ce = mb_cache_entry_alloc(ext2_xattr_cache); ++ if (!ce) ++ return -ENOMEM; ++ error = mb_cache_entry_insert(ce, bh->b_dev, bh->b_blocknr, &hash); ++ if (error) { ++ mb_cache_entry_free(ce); ++ if (error == -EBUSY) { ++ ea_bdebug(bh, "already in cache (%d cache entries)", ++ atomic_read(&ext2_xattr_cache->c_entry_count)); ++ error = 0; ++ } ++ } else { ++ ea_bdebug(bh, "inserting [%x] (%d cache entries)", (int)hash, ++ atomic_read(&ext2_xattr_cache->c_entry_count)); ++ mb_cache_entry_release(ce); ++ } ++ return error; ++} ++ ++/* ++ * ext2_xattr_cmp() ++ * ++ * Compare two extended attribute blocks for equality. ++ * ++ * Returns 0 if the blocks are equal, 1 if they differ, and ++ * a negative error number on errors. ++ */ ++static int ++ext2_xattr_cmp(struct ext2_xattr_header *header1, ++ struct ext2_xattr_header *header2) ++{ ++ struct ext2_xattr_entry *entry1, *entry2; ++ ++ entry1 = ENTRY(header1+1); ++ entry2 = ENTRY(header2+1); ++ while (!IS_LAST_ENTRY(entry1)) { ++ if (IS_LAST_ENTRY(entry2)) ++ return 1; ++ if (entry1->e_hash != entry2->e_hash || ++ entry1->e_name_len != entry2->e_name_len || ++ entry1->e_value_size != entry2->e_value_size || ++ memcmp(entry1->e_name, entry2->e_name, entry1->e_name_len)) ++ return 1; ++ if (entry1->e_value_block != 0 || entry2->e_value_block != 0) ++ return -EIO; ++ if (memcmp((char *)header1 + le16_to_cpu(entry1->e_value_offs), ++ (char *)header2 + le16_to_cpu(entry2->e_value_offs), ++ le32_to_cpu(entry1->e_value_size))) ++ return 1; ++ ++ entry1 = EXT2_XATTR_NEXT(entry1); ++ entry2 = EXT2_XATTR_NEXT(entry2); ++ } ++ if (!IS_LAST_ENTRY(entry2)) ++ return 1; ++ return 0; ++} ++ ++/* ++ * ext2_xattr_cache_find() ++ * ++ * Find an identical extended attribute block. ++ * ++ * Returns a pointer to the block found, or NULL if such a block was ++ * not found or an error occurred. ++ */ ++static struct buffer_head * ++ext2_xattr_cache_find(struct inode *inode, struct ext2_xattr_header *header) ++{ ++ __u32 hash = le32_to_cpu(header->h_hash); ++ struct mb_cache_entry *ce; ++ ++ if (!header->h_hash) ++ return NULL; /* never share */ ++ ea_idebug(inode, "looking for cached blocks [%x]", (int)hash); ++ ce = mb_cache_entry_find_first(ext2_xattr_cache, 0, inode->i_dev, hash); ++ while (ce) { ++ struct buffer_head *bh = sb_bread(inode->i_sb, ce->e_block); ++ ++ if (!bh) { ++ ext2_error(inode->i_sb, "ext2_xattr_cache_find", ++ "inode %ld: block %ld read error", ++ inode->i_ino, ce->e_block); ++ } else if (le32_to_cpu(HDR(bh)->h_refcount) > ++ EXT2_XATTR_REFCOUNT_MAX) { ++ ea_idebug(inode, "block %ld refcount %d>%d",ce->e_block, ++ le32_to_cpu(HDR(bh)->h_refcount), ++ EXT2_XATTR_REFCOUNT_MAX); ++ } else if (!ext2_xattr_cmp(header, HDR(bh))) { ++ ea_bdebug(bh, "b_count=%d",atomic_read(&(bh->b_count))); ++ mb_cache_entry_release(ce); ++ return bh; ++ } ++ brelse(bh); ++ ce = mb_cache_entry_find_next(ce, 0, inode->i_dev, hash); ++ } ++ return NULL; ++} ++ ++/* ++ * ext2_xattr_cache_remove() ++ * ++ * Remove the cache entry of a block from the cache. Called when a ++ * block becomes invalid. ++ */ ++static void ++ext2_xattr_cache_remove(struct buffer_head *bh) ++{ ++ struct mb_cache_entry *ce; ++ ++ ce = mb_cache_entry_get(ext2_xattr_cache, bh->b_dev, bh->b_blocknr); ++ if (ce) { ++ ea_bdebug(bh, "removing (%d cache entries remaining)", ++ atomic_read(&ext2_xattr_cache->c_entry_count)-1); ++ mb_cache_entry_free(ce); ++ } else ++ ea_bdebug(bh, "no cache entry"); ++} ++ ++#define NAME_HASH_SHIFT 5 ++#define VALUE_HASH_SHIFT 16 ++ ++/* ++ * ext2_xattr_hash_entry() ++ * ++ * Compute the hash of an extended attribute. ++ */ ++static inline void ext2_xattr_hash_entry(struct ext2_xattr_header *header, ++ struct ext2_xattr_entry *entry) ++{ ++ __u32 hash = 0; ++ char *name = entry->e_name; ++ int n; ++ ++ for (n=0; n < entry->e_name_len; n++) { ++ hash = (hash << NAME_HASH_SHIFT) ^ ++ (hash >> (8*sizeof(hash) - NAME_HASH_SHIFT)) ^ ++ *name++; ++ } ++ ++ if (entry->e_value_block == 0 && entry->e_value_size != 0) { ++ __u32 *value = (__u32 *)((char *)header + ++ le16_to_cpu(entry->e_value_offs)); ++ for (n = (le32_to_cpu(entry->e_value_size) + ++ EXT2_XATTR_ROUND) >> EXT2_XATTR_PAD_BITS; n; n--) { ++ hash = (hash << VALUE_HASH_SHIFT) ^ ++ (hash >> (8*sizeof(hash) - VALUE_HASH_SHIFT)) ^ ++ le32_to_cpu(*value++); ++ } ++ } ++ entry->e_hash = cpu_to_le32(hash); ++} ++ ++#undef NAME_HASH_SHIFT ++#undef VALUE_HASH_SHIFT ++ ++#define BLOCK_HASH_SHIFT 16 ++ ++/* ++ * ext2_xattr_rehash() ++ * ++ * Re-compute the extended attribute hash value after an entry has changed. ++ */ ++static void ext2_xattr_rehash(struct ext2_xattr_header *header, ++ struct ext2_xattr_entry *entry) ++{ ++ struct ext2_xattr_entry *here; ++ __u32 hash = 0; ++ ++ ext2_xattr_hash_entry(header, entry); ++ here = ENTRY(header+1); ++ while (!IS_LAST_ENTRY(here)) { ++ if (!here->e_hash) { ++ /* Block is not shared if an entry's hash value == 0 */ ++ hash = 0; ++ break; ++ } ++ hash = (hash << BLOCK_HASH_SHIFT) ^ ++ (hash >> (8*sizeof(hash) - BLOCK_HASH_SHIFT)) ^ ++ le32_to_cpu(here->e_hash); ++ here = EXT2_XATTR_NEXT(here); ++ } ++ header->h_hash = cpu_to_le32(hash); ++} ++ ++#undef BLOCK_HASH_SHIFT ++ ++int __init ++init_ext2_xattr(void) ++{ ++ ext2_xattr_cache = mb_cache_create("ext2_xattr", NULL, ++ sizeof(struct mb_cache_entry) + ++ sizeof(struct mb_cache_entry_index), 1, 61); ++ if (!ext2_xattr_cache) ++ return -ENOMEM; ++ ++ return 0; ++} ++ ++void ++exit_ext2_xattr(void) ++{ ++ mb_cache_destroy(ext2_xattr_cache); ++} ++ ++#else /* CONFIG_EXT2_FS_XATTR_SHARING */ ++ ++int __init ++init_ext2_xattr(void) ++{ ++ return 0; ++} ++ ++void ++exit_ext2_xattr(void) ++{ ++} ++ ++#endif /* CONFIG_EXT2_FS_XATTR_SHARING */ +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext2/xattr_user.c 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,103 @@ ++/* ++ * linux/fs/ext2/xattr_user.c ++ * Handler for extended user attributes. ++ * ++ * Copyright (C) 2001 by Andreas Gruenbacher, ++ */ ++ ++#include ++#include ++#include ++#include ++#include ++ ++#ifdef CONFIG_EXT2_FS_POSIX_ACL ++# include ++#endif ++ ++#define XATTR_USER_PREFIX "user." ++ ++static size_t ++ext2_xattr_user_list(char *list, struct inode *inode, ++ const char *name, int name_len) ++{ ++ const int prefix_len = sizeof(XATTR_USER_PREFIX)-1; ++ ++ if (!test_opt(inode->i_sb, XATTR_USER)) ++ return 0; ++ ++ if (list) { ++ memcpy(list, XATTR_USER_PREFIX, prefix_len); ++ memcpy(list+prefix_len, name, name_len); ++ list[prefix_len + name_len] = '\0'; ++ } ++ return prefix_len + name_len + 1; ++} ++ ++static int ++ext2_xattr_user_get(struct inode *inode, const char *name, ++ void *buffer, size_t size) ++{ ++ int error; ++ ++ if (strcmp(name, "") == 0) ++ return -EINVAL; ++ if (!test_opt(inode->i_sb, XATTR_USER)) ++ return -ENOTSUP; ++#ifdef CONFIG_EXT2_FS_POSIX_ACL ++ error = ext2_permission_locked(inode, MAY_READ); ++#else ++ error = permission(inode, MAY_READ); ++#endif ++ if (error) ++ return error; ++ ++ return ext2_xattr_get(inode, EXT2_XATTR_INDEX_USER, name, ++ buffer, size); ++} ++ ++static int ++ext2_xattr_user_set(struct inode *inode, const char *name, ++ const void *value, size_t size, int flags) ++{ ++ int error; ++ ++ if (strcmp(name, "") == 0) ++ return -EINVAL; ++ if (!test_opt(inode->i_sb, XATTR_USER)) ++ return -ENOTSUP; ++ if ( !S_ISREG(inode->i_mode) && ++ (!S_ISDIR(inode->i_mode) || inode->i_mode & S_ISVTX)) ++ return -EPERM; ++#ifdef CONFIG_EXT2_FS_POSIX_ACL ++ error = ext2_permission_locked(inode, MAY_WRITE); ++#else ++ error = permission(inode, MAY_WRITE); ++#endif ++ if (error) ++ return error; ++ ++ return ext2_xattr_set(inode, EXT2_XATTR_INDEX_USER, name, ++ value, size, flags); ++} ++ ++struct ext2_xattr_handler ext2_xattr_user_handler = { ++ prefix: XATTR_USER_PREFIX, ++ list: ext2_xattr_user_list, ++ get: ext2_xattr_user_get, ++ set: ext2_xattr_user_set, ++}; ++ ++int __init ++init_ext2_xattr_user(void) ++{ ++ return ext2_xattr_register(EXT2_XATTR_INDEX_USER, ++ &ext2_xattr_user_handler); ++} ++ ++void ++exit_ext2_xattr_user(void) ++{ ++ ext2_xattr_unregister(EXT2_XATTR_INDEX_USER, ++ &ext2_xattr_user_handler); ++} +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext3/ext3-exports.c 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,13 @@ ++#include ++#include ++#include ++#include ++#include ++ ++EXPORT_SYMBOL(ext3_force_commit); ++EXPORT_SYMBOL(ext3_bread); ++EXPORT_SYMBOL(ext3_xattr_register); ++EXPORT_SYMBOL(ext3_xattr_unregister); ++EXPORT_SYMBOL(ext3_xattr_get); ++EXPORT_SYMBOL(ext3_xattr_list); ++EXPORT_SYMBOL(ext3_xattr_set); +--- linux-2.4.22-ac1/fs/ext3/file.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:55:12.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/file.c 2003-09-25 23:57:02.000000000 +0400 +@@ -23,6 +23,7 @@ + #include + #include + #include ++#include + #include + #include + +@@ -127,5 +128,9 @@ struct file_operations ext3_file_operati + struct inode_operations ext3_file_inode_operations = { + truncate: ext3_truncate, /* BKL held */ + setattr: ext3_setattr, /* BKL held */ ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ + }; + +--- linux-2.4.22-ac1/fs/ext3/ialloc.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/ialloc.c 2003-09-25 23:57:02.000000000 +0400 +@@ -17,6 +17,7 @@ + #include + #include + #include ++#include + #include + #include + #include +@@ -217,6 +218,7 @@ void ext3_free_inode (handle_t *handle, + * as writing the quota to disk may need the lock as well. + */ + DQUOT_INIT(inode); ++ ext3_xattr_delete_inode(handle, inode); + DQUOT_FREE_INODE(inode); + DQUOT_DROP(inode); + +--- linux-2.4.22-ac1/fs/ext3/inode.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:29.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/inode.c 2003-09-26 00:10:09.000000000 +0400 +@@ -39,6 +39,18 @@ + */ + #undef SEARCH_FROM_ZERO + ++/* ++ * Test whether an inode is a fast symlink. ++ */ ++static inline int ext3_inode_is_fast_symlink(struct inode *inode) ++{ ++ int ea_blocks = inode->u.ext3_i.i_file_acl ? ++ (inode->i_sb->s_blocksize >> 9) : 0; ++ ++ return (S_ISLNK(inode->i_mode) && ++ inode->i_blocks - ea_blocks == 0); ++} ++ + /* The ext3 forget function must perform a revoke if we are freeing data + * which has been journaled. Metadata (eg. indirect blocks) must be + * revoked in all cases. +@@ -48,7 +60,7 @@ + * still needs to be revoked. + */ + +-static int ext3_forget(handle_t *handle, int is_metadata, ++int ext3_forget(handle_t *handle, int is_metadata, + struct inode *inode, struct buffer_head *bh, + int blocknr) + { +@@ -179,9 +191,7 @@ void ext3_delete_inode (struct inode * i + { + handle_t *handle; + +- if (is_bad_inode(inode) || +- inode->i_ino == EXT3_ACL_IDX_INO || +- inode->i_ino == EXT3_ACL_DATA_INO) ++ if (is_bad_inode(inode)) + goto no_delete; + + lock_kernel(); +@@ -1874,6 +1884,8 @@ void ext3_truncate(struct inode * inode) + if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || + S_ISLNK(inode->i_mode))) + return; ++ if (ext3_inode_is_fast_symlink(inode)) ++ return; + if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) + return; + +@@ -2021,8 +2033,6 @@ int ext3_get_inode_loc (struct inode *in + struct ext3_group_desc * gdp; + + if ((inode->i_ino != EXT3_ROOT_INO && +- inode->i_ino != EXT3_ACL_IDX_INO && +- inode->i_ino != EXT3_ACL_DATA_INO && + inode->i_ino != EXT3_JOURNAL_INO && + inode->i_ino < EXT3_FIRST_INO(inode->i_sb)) || + inode->i_ino > le32_to_cpu( +@@ -2163,10 +2173,7 @@ void ext3_read_inode(struct inode * inod + inode->u.ext3_i.i_data[block] = iloc.raw_inode->i_block[block]; + INIT_LIST_HEAD(&inode->u.ext3_i.i_orphan); + +- if (inode->i_ino == EXT3_ACL_IDX_INO || +- inode->i_ino == EXT3_ACL_DATA_INO) +- /* Nothing to do */ ; +- else if (S_ISREG(inode->i_mode)) { ++ if (S_ISREG(inode->i_mode)) { + inode->i_op = &ext3_file_inode_operations; + inode->i_fop = &ext3_file_operations; + inode->i_mapping->a_ops = &ext3_aops; +@@ -2174,15 +2181,17 @@ void ext3_read_inode(struct inode * inod + inode->i_op = &ext3_dir_inode_operations; + inode->i_fop = &ext3_dir_operations; + } else if (S_ISLNK(inode->i_mode)) { +- if (!inode->i_blocks) ++ if (ext3_inode_is_fast_symlink(inode)) + inode->i_op = &ext3_fast_symlink_inode_operations; + else { +- inode->i_op = &page_symlink_inode_operations; ++ inode->i_op = &ext3_symlink_inode_operations; + inode->i_mapping->a_ops = &ext3_aops; + } +- } else ++ } else { ++ inode->i_op = &ext3_special_inode_operations; + init_special_inode(inode, inode->i_mode, + le32_to_cpu(iloc.raw_inode->i_block[0])); ++ } + brelse(iloc.bh); + ext3_set_inode_flags(inode); + return; +--- linux-2.4.22-ac1/fs/ext3/Makefile~xattr-0.8.54-2.4.22-rh 2003-09-25 14:55:12.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/Makefile 2003-09-25 23:57:02.000000000 +0400 +@@ -1,5 +1,5 @@ + # +-# Makefile for the linux ext2-filesystem routines. ++# Makefile for the linux ext3-filesystem routines. + # + # Note! Dependencies are done automagically by 'make dep', which also + # removes any old dependencies. DON'T put your own dependencies here +@@ -9,10 +9,14 @@ + + O_TARGET := ext3.o + +-export-objs := super.o inode.o ++export-objs := ext3-exports.o + + obj-y := balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o \ +- ioctl.o namei.o super.o symlink.o hash.o ++ ioctl.o namei.o super.o symlink.o hash.o ext3-exports.o + obj-m := $(O_TARGET) + ++export-objs += xattr.o ++obj-$(CONFIG_EXT3_FS_XATTR) += xattr.o ++obj-$(CONFIG_EXT3_FS_XATTR_USER) += xattr_user.o ++ + include $(TOPDIR)/Rules.make +--- linux-2.4.22-ac1/fs/ext3/namei.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:58:37.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/namei.c 2003-09-25 23:57:02.000000000 +0400 +@@ -29,6 +29,7 @@ + #include + #include + #include ++#include + #include + #include + #include +@@ -1614,7 +1615,7 @@ static int ext3_mkdir(struct inode * dir + if (IS_SYNC(dir)) + handle->h_sync = 1; + +- inode = ext3_new_inode (handle, dir, S_IFDIR); ++ inode = ext3_new_inode (handle, dir, S_IFDIR | mode); + err = PTR_ERR(inode); + if (IS_ERR(inode)) + goto out_stop; +@@ -1622,7 +1623,6 @@ static int ext3_mkdir(struct inode * dir + inode->i_op = &ext3_dir_inode_operations; + inode->i_fop = &ext3_dir_operations; + inode->i_size = EXT3_I(inode)->i_disksize = inode->i_sb->s_blocksize; +- inode->i_blocks = 0; + dir_block = ext3_bread (handle, inode, 0, 1, &err); + if (!dir_block) { + inode->i_nlink--; /* is this nlink == 0? */ +@@ -1649,9 +1649,6 @@ static int ext3_mkdir(struct inode * dir + BUFFER_TRACE(dir_block, "call ext3_journal_dirty_metadata"); + ext3_journal_dirty_metadata(handle, dir_block); + brelse (dir_block); +- inode->i_mode = S_IFDIR | mode; +- if (dir->i_mode & S_ISGID) +- inode->i_mode |= S_ISGID; + ext3_mark_inode_dirty(handle, inode); + err = ext3_add_entry (handle, dentry, inode); + if (err) { +@@ -2020,7 +2017,7 @@ static int ext3_symlink (struct inode * + goto out_stop; + + if (l > sizeof (EXT3_I(inode)->i_data)) { +- inode->i_op = &page_symlink_inode_operations; ++ inode->i_op = &ext3_symlink_inode_operations; + inode->i_mapping->a_ops = &ext3_aops; + /* + * block_symlink() calls back into ext3_prepare/commit_write. +@@ -2245,4 +2242,16 @@ struct inode_operations ext3_dir_inode_o + rmdir: ext3_rmdir, /* BKL held */ + mknod: ext3_mknod, /* BKL held */ + rename: ext3_rename, /* BKL held */ ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ + }; ++ ++struct inode_operations ext3_special_inode_operations = { ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ ++}; ++ +--- linux-2.4.22-ac1/fs/ext3/super.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:55:12.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/ext3/super.c 2003-09-26 00:12:23.000000000 +0400 +@@ -24,6 +24,7 @@ + #include + #include + #include ++#include + #include + #include + #include +@@ -406,6 +407,7 @@ void ext3_put_super (struct super_block + kdev_t j_dev = sbi->s_journal->j_dev; + int i; + ++ ext3_xattr_put_super(sb); + journal_destroy(sbi->s_journal); + if (!(sb->s_flags & MS_RDONLY)) { + EXT3_CLEAR_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER); +@@ -506,6 +508,7 @@ static int parse_options (char * options + int is_remount) + { + unsigned long *mount_options = &sbi->s_mount_opt; ++ + uid_t *resuid = &sbi->s_resuid; + gid_t *resgid = &sbi->s_resgid; + char * this_char; +@@ -518,6 +521,13 @@ static int parse_options (char * options + this_char = strtok (NULL, ",")) { + if ((value = strchr (this_char, '=')) != NULL) + *value++ = 0; ++#ifdef CONFIG_EXT3_FS_XATTR_USER ++ if (!strcmp (this_char, "user_xattr")) ++ set_opt (*mount_options, XATTR_USER); ++ else if (!strcmp (this_char, "nouser_xattr")) ++ clear_opt (*mount_options, XATTR_USER); ++ else ++#endif + if (!strcmp (this_char, "bsddf")) + clear_opt (*mount_options, MINIX_DF); + else if (!strcmp (this_char, "nouid32")) { +@@ -935,6 +945,12 @@ struct super_block * ext3_read_super (st + sbi->s_mount_opt = 0; + sbi->s_resuid = EXT3_DEF_RESUID; + sbi->s_resgid = EXT3_DEF_RESGID; ++ ++ /* Default extended attribute flags */ ++#ifdef CONFIG_EXT3_FS_XATTR_USER ++ /* set_opt(sbi->s_mount_opt, XATTR_USER); */ ++#endif ++ + if (!parse_options ((char *) data, &sb_block, sbi, &journal_inum, 0)) { + sb->s_dev = 0; + goto out_fail; +@@ -1839,22 +1855,35 @@ static DECLARE_FSTYPE_DEV(ext3_fs_type, + + static int __init init_ext3_fs(void) + { ++ int error; + #ifdef CONFIG_QUOTA + init_dquot_operations(&ext3_qops); + old_sync_dquot = ext3_qops.sync_dquot; + ext3_qops.sync_dquot = ext3_sync_dquot; + #endif +- return register_filesystem(&ext3_fs_type); ++ error = init_ext3_xattr(); ++ if (error) ++ return error; ++ error = init_ext3_xattr_user(); ++ if (error) ++ goto fail; ++ error = register_filesystem(&ext3_fs_type); ++ if (!error) ++ return 0; ++ ++ exit_ext3_xattr_user(); ++fail: ++ exit_ext3_xattr(); ++ return error; + } + + static void __exit exit_ext3_fs(void) + { + unregister_filesystem(&ext3_fs_type); ++ exit_ext3_xattr_user(); ++ exit_ext3_xattr(); + } + +-EXPORT_SYMBOL(ext3_force_commit); +-EXPORT_SYMBOL(ext3_bread); +- + MODULE_AUTHOR("Remy Card, Stephen Tweedie, Andrew Morton, Andreas Dilger, Theodore Ts'o and others"); + MODULE_DESCRIPTION("Second Extended Filesystem with journaling extensions"); + MODULE_LICENSE("GPL"); +--- linux-2.4.22-ac1/fs/ext3/symlink.c~xattr-0.8.54-2.4.22-rh 2001-11-10 01:25:04.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext3/symlink.c 2003-09-25 23:57:02.000000000 +0400 +@@ -20,6 +20,7 @@ + #include + #include + #include ++#include + + static int ext3_readlink(struct dentry *dentry, char *buffer, int buflen) + { +@@ -33,7 +34,20 @@ static int ext3_follow_link(struct dentr + return vfs_follow_link(nd, s); + } + ++struct inode_operations ext3_symlink_inode_operations = { ++ readlink: page_readlink, /* BKL not held. Don't need */ ++ follow_link: page_follow_link, /* BKL not held. Don't need */ ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ ++}; ++ + struct inode_operations ext3_fast_symlink_inode_operations = { + readlink: ext3_readlink, /* BKL not held. Don't need */ + follow_link: ext3_follow_link, /* BKL not held. Don't need */ ++ setxattr: ext3_setxattr, /* BKL held */ ++ getxattr: ext3_getxattr, /* BKL held */ ++ listxattr: ext3_listxattr, /* BKL held */ ++ removexattr: ext3_removexattr, /* BKL held */ + }; +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext3/xattr.c 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,1225 @@ ++/* ++ * linux/fs/ext3/xattr.c ++ * ++ * Copyright (C) 2001 by Andreas Gruenbacher, ++ * ++ * Fix by Harrison Xing . ++ * Ext3 code with a lot of help from Eric Jarman . ++ * Extended attributes for symlinks and special files added per ++ * suggestion of Luka Renko . ++ */ ++ ++/* ++ * Extended attributes are stored on disk blocks allocated outside of ++ * any inode. The i_file_acl field is then made to point to this allocated ++ * block. If all extended attributes of an inode are identical, these ++ * inodes may share the same extended attribute block. Such situations ++ * are automatically detected by keeping a cache of recent attribute block ++ * numbers and hashes over the block's contents in memory. ++ * ++ * ++ * Extended attribute block layout: ++ * ++ * +------------------+ ++ * | header | ++ * | entry 1 | | ++ * | entry 2 | | growing downwards ++ * | entry 3 | v ++ * | four null bytes | ++ * | . . . | ++ * | value 1 | ^ ++ * | value 3 | | growing upwards ++ * | value 2 | | ++ * +------------------+ ++ * ++ * The block header is followed by multiple entry descriptors. These entry ++ * descriptors are variable in size, and alligned to EXT3_XATTR_PAD ++ * byte boundaries. The entry descriptors are sorted by attribute name, ++ * so that two extended attribute blocks can be compared efficiently. ++ * ++ * Attribute values are aligned to the end of the block, stored in ++ * no specific order. They are also padded to EXT3_XATTR_PAD byte ++ * boundaries. No additional gaps are left between them. ++ * ++ * Locking strategy ++ * ---------------- ++ * The VFS already holds the BKL and the inode->i_sem semaphore when any of ++ * the xattr inode operations are called, so we are guaranteed that only one ++ * processes accesses extended attributes of an inode at any time. ++ * ++ * For writing we also grab the ext3_xattr_sem semaphore. This ensures that ++ * only a single process is modifying an extended attribute block, even ++ * if the block is shared among inodes. ++ * ++ * Note for porting to 2.5 ++ * ----------------------- ++ * The BKL will no longer be held in the xattr inode operations. ++ */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++ ++#define EXT3_EA_USER "user." ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,0) ++# define mark_buffer_dirty(bh) mark_buffer_dirty(bh, 1) ++#endif ++ ++#define HDR(bh) ((struct ext3_xattr_header *)((bh)->b_data)) ++#define ENTRY(ptr) ((struct ext3_xattr_entry *)(ptr)) ++#define FIRST_ENTRY(bh) ENTRY(HDR(bh)+1) ++#define IS_LAST_ENTRY(entry) (*(__u32 *)(entry) == 0) ++ ++#ifdef EXT3_XATTR_DEBUG ++# define ea_idebug(inode, f...) do { \ ++ printk(KERN_DEBUG "inode %s:%ld: ", \ ++ kdevname(inode->i_dev), inode->i_ino); \ ++ printk(f); \ ++ printk("\n"); \ ++ } while (0) ++# define ea_bdebug(bh, f...) do { \ ++ printk(KERN_DEBUG "block %s:%ld: ", \ ++ kdevname(bh->b_dev), bh->b_blocknr); \ ++ printk(f); \ ++ printk("\n"); \ ++ } while (0) ++#else ++# define ea_idebug(f...) ++# define ea_bdebug(f...) ++#endif ++ ++static int ext3_xattr_set2(handle_t *, struct inode *, struct buffer_head *, ++ struct ext3_xattr_header *); ++ ++#ifdef CONFIG_EXT3_FS_XATTR_SHARING ++ ++static int ext3_xattr_cache_insert(struct buffer_head *); ++static struct buffer_head *ext3_xattr_cache_find(struct inode *, ++ struct ext3_xattr_header *); ++static void ext3_xattr_cache_remove(struct buffer_head *); ++static void ext3_xattr_rehash(struct ext3_xattr_header *, ++ struct ext3_xattr_entry *); ++ ++static struct mb_cache *ext3_xattr_cache; ++ ++#else ++# define ext3_xattr_cache_insert(bh) 0 ++# define ext3_xattr_cache_find(inode, header) NULL ++# define ext3_xattr_cache_remove(bh) while(0) {} ++# define ext3_xattr_rehash(header, entry) while(0) {} ++#endif ++ ++/* ++ * If a file system does not share extended attributes among inodes, ++ * we should not need the ext3_xattr_sem semaphore. However, the ++ * filesystem may still contain shared blocks, so we always take ++ * the lock. ++ */ ++ ++DECLARE_MUTEX(ext3_xattr_sem); ++ ++static inline int ++ext3_xattr_new_block(handle_t *handle, struct inode *inode, ++ int * errp, int force) ++{ ++ struct super_block *sb = inode->i_sb; ++ int goal = le32_to_cpu(EXT3_SB(sb)->s_es->s_first_data_block) + ++ EXT3_I(inode)->i_block_group * EXT3_BLOCKS_PER_GROUP(sb); ++ ++ /* How can we enforce the allocation? */ ++ int block = ext3_new_block(handle, inode, goal, 0, 0, errp); ++#ifdef OLD_QUOTAS ++ if (!*errp) ++ inode->i_blocks += inode->i_sb->s_blocksize >> 9; ++#endif ++ return block; ++} ++ ++static inline int ++ext3_xattr_quota_alloc(struct inode *inode, int force) ++{ ++ /* How can we enforce the allocation? */ ++#ifdef OLD_QUOTAS ++ int error = DQUOT_ALLOC_BLOCK(inode->i_sb, inode, 1); ++ if (!error) ++ inode->i_blocks += inode->i_sb->s_blocksize >> 9; ++#else ++ int error = DQUOT_ALLOC_BLOCK(inode, 1); ++#endif ++ return error; ++} ++ ++#ifdef OLD_QUOTAS ++ ++static inline void ++ext3_xattr_quota_free(struct inode *inode) ++{ ++ DQUOT_FREE_BLOCK(inode->i_sb, inode, 1); ++ inode->i_blocks -= inode->i_sb->s_blocksize >> 9; ++} ++ ++static inline void ++ext3_xattr_free_block(handle_t *handle, struct inode * inode, ++ unsigned long block) ++{ ++ ext3_free_blocks(handle, inode, block, 1); ++ inode->i_blocks -= inode->i_sb->s_blocksize >> 9; ++} ++ ++#else ++# define ext3_xattr_quota_free(inode) \ ++ DQUOT_FREE_BLOCK(inode, 1) ++# define ext3_xattr_free_block(handle, inode, block) \ ++ ext3_free_blocks(handle, inode, block, 1) ++#endif ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,18) ++ ++static inline struct buffer_head * ++sb_bread(struct super_block *sb, int block) ++{ ++ return bread(sb->s_dev, block, sb->s_blocksize); ++} ++ ++static inline struct buffer_head * ++sb_getblk(struct super_block *sb, int block) ++{ ++ return getblk(sb->s_dev, block, sb->s_blocksize); ++} ++ ++#endif ++ ++struct ext3_xattr_handler *ext3_xattr_handlers[EXT3_XATTR_INDEX_MAX]; ++rwlock_t ext3_handler_lock = RW_LOCK_UNLOCKED; ++ ++int ++ext3_xattr_register(int name_index, struct ext3_xattr_handler *handler) ++{ ++ int error = -EINVAL; ++ ++ if (name_index > 0 && name_index <= EXT3_XATTR_INDEX_MAX) { ++ write_lock(&ext3_handler_lock); ++ if (!ext3_xattr_handlers[name_index-1]) { ++ ext3_xattr_handlers[name_index-1] = handler; ++ error = 0; ++ } ++ write_unlock(&ext3_handler_lock); ++ } ++ return error; ++} ++ ++void ++ext3_xattr_unregister(int name_index, struct ext3_xattr_handler *handler) ++{ ++ if (name_index > 0 || name_index <= EXT3_XATTR_INDEX_MAX) { ++ write_lock(&ext3_handler_lock); ++ ext3_xattr_handlers[name_index-1] = NULL; ++ write_unlock(&ext3_handler_lock); ++ } ++} ++ ++static inline const char * ++strcmp_prefix(const char *a, const char *a_prefix) ++{ ++ while (*a_prefix && *a == *a_prefix) { ++ a++; ++ a_prefix++; ++ } ++ return *a_prefix ? NULL : a; ++} ++ ++/* ++ * Decode the extended attribute name, and translate it into ++ * the name_index and name suffix. ++ */ ++static inline struct ext3_xattr_handler * ++ext3_xattr_resolve_name(const char **name) ++{ ++ struct ext3_xattr_handler *handler = NULL; ++ int i; ++ ++ if (!*name) ++ return NULL; ++ read_lock(&ext3_handler_lock); ++ for (i=0; iprefix); ++ if (n) { ++ handler = ext3_xattr_handlers[i]; ++ *name = n; ++ break; ++ } ++ } ++ } ++ read_unlock(&ext3_handler_lock); ++ return handler; ++} ++ ++static inline struct ext3_xattr_handler * ++ext3_xattr_handler(int name_index) ++{ ++ struct ext3_xattr_handler *handler = NULL; ++ if (name_index > 0 && name_index <= EXT3_XATTR_INDEX_MAX) { ++ read_lock(&ext3_handler_lock); ++ handler = ext3_xattr_handlers[name_index-1]; ++ read_unlock(&ext3_handler_lock); ++ } ++ return handler; ++} ++ ++/* ++ * Inode operation getxattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++ssize_t ++ext3_getxattr(struct dentry *dentry, const char *name, ++ void *buffer, size_t size) ++{ ++ struct ext3_xattr_handler *handler; ++ struct inode *inode = dentry->d_inode; ++ ++ handler = ext3_xattr_resolve_name(&name); ++ if (!handler) ++ return -ENOTSUP; ++ return handler->get(inode, name, buffer, size); ++} ++ ++/* ++ * Inode operation listxattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++ssize_t ++ext3_listxattr(struct dentry *dentry, char *buffer, size_t size) ++{ ++ return ext3_xattr_list(dentry->d_inode, buffer, size); ++} ++ ++/* ++ * Inode operation setxattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++int ++ext3_setxattr(struct dentry *dentry, const char *name, ++ const void *value, size_t size, int flags) ++{ ++ struct ext3_xattr_handler *handler; ++ struct inode *inode = dentry->d_inode; ++ ++ if (size == 0) ++ value = ""; /* empty EA, do not remove */ ++ handler = ext3_xattr_resolve_name(&name); ++ if (!handler) ++ return -ENOTSUP; ++ return handler->set(inode, name, value, size, flags); ++} ++ ++/* ++ * Inode operation removexattr() ++ * ++ * dentry->d_inode->i_sem down ++ * BKL held [before 2.5.x] ++ */ ++int ++ext3_removexattr(struct dentry *dentry, const char *name) ++{ ++ struct ext3_xattr_handler *handler; ++ struct inode *inode = dentry->d_inode; ++ ++ handler = ext3_xattr_resolve_name(&name); ++ if (!handler) ++ return -ENOTSUP; ++ return handler->set(inode, name, NULL, 0, XATTR_REPLACE); ++} ++ ++/* ++ * ext3_xattr_get() ++ * ++ * Copy an extended attribute into the buffer ++ * provided, or compute the buffer size required. ++ * Buffer is NULL to compute the size of the buffer required. ++ * ++ * Returns a negative error number on failure, or the number of bytes ++ * used / required on success. ++ */ ++int ++ext3_xattr_get(struct inode *inode, int name_index, const char *name, ++ void *buffer, size_t buffer_size) ++{ ++ struct buffer_head *bh = NULL; ++ struct ext3_xattr_entry *entry; ++ unsigned int block, size; ++ char *end; ++ int name_len, error; ++ ++ ea_idebug(inode, "name=%d.%s, buffer=%p, buffer_size=%ld", ++ name_index, name, buffer, (long)buffer_size); ++ ++ if (name == NULL) ++ return -EINVAL; ++ if (!EXT3_I(inode)->i_file_acl) ++ return -ENOATTR; ++ block = EXT3_I(inode)->i_file_acl; ++ ea_idebug(inode, "reading block %d", block); ++ bh = sb_bread(inode->i_sb, block); ++ if (!bh) ++ return -EIO; ++ ea_bdebug(bh, "b_count=%d, refcount=%d", ++ atomic_read(&(bh->b_count)), le32_to_cpu(HDR(bh)->h_refcount)); ++ end = bh->b_data + bh->b_size; ++ if (HDR(bh)->h_magic != cpu_to_le32(EXT3_XATTR_MAGIC) || ++ HDR(bh)->h_blocks != cpu_to_le32(1)) { ++bad_block: ext3_error(inode->i_sb, "ext3_xattr_get", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ /* find named attribute */ ++ name_len = strlen(name); ++ ++ error = -ERANGE; ++ if (name_len > 255) ++ goto cleanup; ++ entry = FIRST_ENTRY(bh); ++ while (!IS_LAST_ENTRY(entry)) { ++ struct ext3_xattr_entry *next = ++ EXT3_XATTR_NEXT(entry); ++ if ((char *)next >= end) ++ goto bad_block; ++ if (name_index == entry->e_name_index && ++ name_len == entry->e_name_len && ++ memcmp(name, entry->e_name, name_len) == 0) ++ goto found; ++ entry = next; ++ } ++ /* Check the remaining name entries */ ++ while (!IS_LAST_ENTRY(entry)) { ++ struct ext3_xattr_entry *next = ++ EXT3_XATTR_NEXT(entry); ++ if ((char *)next >= end) ++ goto bad_block; ++ entry = next; ++ } ++ if (ext3_xattr_cache_insert(bh)) ++ ea_idebug(inode, "cache insert failed"); ++ error = -ENOATTR; ++ goto cleanup; ++found: ++ /* check the buffer size */ ++ if (entry->e_value_block != 0) ++ goto bad_block; ++ size = le32_to_cpu(entry->e_value_size); ++ if (size > inode->i_sb->s_blocksize || ++ le16_to_cpu(entry->e_value_offs) + size > inode->i_sb->s_blocksize) ++ goto bad_block; ++ ++ if (ext3_xattr_cache_insert(bh)) ++ ea_idebug(inode, "cache insert failed"); ++ if (buffer) { ++ error = -ERANGE; ++ if (size > buffer_size) ++ goto cleanup; ++ /* return value of attribute */ ++ memcpy(buffer, bh->b_data + le16_to_cpu(entry->e_value_offs), ++ size); ++ } ++ error = size; ++ ++cleanup: ++ brelse(bh); ++ ++ return error; ++} ++ ++/* ++ * ext3_xattr_list() ++ * ++ * Copy a list of attribute names into the buffer ++ * provided, or compute the buffer size required. ++ * Buffer is NULL to compute the size of the buffer required. ++ * ++ * Returns a negative error number on failure, or the number of bytes ++ * used / required on success. ++ */ ++int ++ext3_xattr_list(struct inode *inode, char *buffer, size_t buffer_size) ++{ ++ struct buffer_head *bh = NULL; ++ struct ext3_xattr_entry *entry; ++ unsigned int block, size = 0; ++ char *buf, *end; ++ int error; ++ ++ ea_idebug(inode, "buffer=%p, buffer_size=%ld", ++ buffer, (long)buffer_size); ++ ++ if (!EXT3_I(inode)->i_file_acl) ++ return 0; ++ block = EXT3_I(inode)->i_file_acl; ++ ea_idebug(inode, "reading block %d", block); ++ bh = sb_bread(inode->i_sb, block); ++ if (!bh) ++ return -EIO; ++ ea_bdebug(bh, "b_count=%d, refcount=%d", ++ atomic_read(&(bh->b_count)), le32_to_cpu(HDR(bh)->h_refcount)); ++ end = bh->b_data + bh->b_size; ++ if (HDR(bh)->h_magic != cpu_to_le32(EXT3_XATTR_MAGIC) || ++ HDR(bh)->h_blocks != cpu_to_le32(1)) { ++bad_block: ext3_error(inode->i_sb, "ext3_xattr_list", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ /* compute the size required for the list of attribute names */ ++ for (entry = FIRST_ENTRY(bh); !IS_LAST_ENTRY(entry); ++ entry = EXT3_XATTR_NEXT(entry)) { ++ struct ext3_xattr_handler *handler; ++ struct ext3_xattr_entry *next = ++ EXT3_XATTR_NEXT(entry); ++ if ((char *)next >= end) ++ goto bad_block; ++ ++ handler = ext3_xattr_handler(entry->e_name_index); ++ if (handler) ++ size += handler->list(NULL, inode, entry->e_name, ++ entry->e_name_len); ++ } ++ ++ if (ext3_xattr_cache_insert(bh)) ++ ea_idebug(inode, "cache insert failed"); ++ if (!buffer) { ++ error = size; ++ goto cleanup; ++ } else { ++ error = -ERANGE; ++ if (size > buffer_size) ++ goto cleanup; ++ } ++ ++ /* list the attribute names */ ++ buf = buffer; ++ for (entry = FIRST_ENTRY(bh); !IS_LAST_ENTRY(entry); ++ entry = EXT3_XATTR_NEXT(entry)) { ++ struct ext3_xattr_handler *handler; ++ ++ handler = ext3_xattr_handler(entry->e_name_index); ++ if (handler) ++ buf += handler->list(buf, inode, entry->e_name, ++ entry->e_name_len); ++ } ++ error = size; ++ ++cleanup: ++ brelse(bh); ++ ++ return error; ++} ++ ++/* ++ * If the EXT3_FEATURE_COMPAT_EXT_ATTR feature of this file system is ++ * not set, set it. ++ */ ++static void ext3_xattr_update_super_block(handle_t *handle, ++ struct super_block *sb) ++{ ++ if (EXT3_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_EXT_ATTR)) ++ return; ++ ++ lock_super(sb); ++ ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh); ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,4,0) ++ EXT3_SB(sb)->s_feature_compat |= EXT3_FEATURE_COMPAT_EXT_ATTR; ++#endif ++ EXT3_SB(sb)->s_es->s_feature_compat |= ++ cpu_to_le32(EXT3_FEATURE_COMPAT_EXT_ATTR); ++ sb->s_dirt = 1; ++ ext3_journal_dirty_metadata(handle, EXT3_SB(sb)->s_sbh); ++ unlock_super(sb); ++} ++ ++/* ++ * ext3_xattr_set() ++ * ++ * Create, replace or remove an extended attribute for this inode. Buffer ++ * is NULL to remove an existing extended attribute, and non-NULL to ++ * either replace an existing extended attribute, or create a new extended ++ * attribute. The flags XATTR_REPLACE and XATTR_CREATE ++ * specify that an extended attribute must exist and must not exist ++ * previous to the call, respectively. ++ * ++ * Returns 0, or a negative error number on failure. ++ */ ++int ++ext3_xattr_set(handle_t *handle, struct inode *inode, int name_index, ++ const char *name, const void *value, size_t value_len, int flags) ++{ ++ struct super_block *sb = inode->i_sb; ++ struct buffer_head *bh = NULL; ++ struct ext3_xattr_header *header = NULL; ++ struct ext3_xattr_entry *here, *last; ++ unsigned int name_len; ++ int block = EXT3_I(inode)->i_file_acl; ++ int min_offs = sb->s_blocksize, not_found = 1, free, error; ++ char *end; ++ ++ /* ++ * header -- Points either into bh, or to a temporarily ++ * allocated buffer. ++ * here -- The named entry found, or the place for inserting, within ++ * the block pointed to by header. ++ * last -- Points right after the last named entry within the block ++ * pointed to by header. ++ * min_offs -- The offset of the first value (values are aligned ++ * towards the end of the block). ++ * end -- Points right after the block pointed to by header. ++ */ ++ ++ ea_idebug(inode, "name=%d.%s, value=%p, value_len=%ld", ++ name_index, name, value, (long)value_len); ++ ++ if (IS_RDONLY(inode)) ++ return -EROFS; ++ if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) ++ return -EPERM; ++ if (value == NULL) ++ value_len = 0; ++ if (name == NULL) ++ return -EINVAL; ++ name_len = strlen(name); ++ if (name_len > 255 || value_len > sb->s_blocksize) ++ return -ERANGE; ++ down(&ext3_xattr_sem); ++ ++ if (block) { ++ /* The inode already has an extended attribute block. */ ++ bh = sb_bread(sb, block); ++ error = -EIO; ++ if (!bh) ++ goto cleanup; ++ ea_bdebug(bh, "b_count=%d, refcount=%d", ++ atomic_read(&(bh->b_count)), ++ le32_to_cpu(HDR(bh)->h_refcount)); ++ header = HDR(bh); ++ end = bh->b_data + bh->b_size; ++ if (header->h_magic != cpu_to_le32(EXT3_XATTR_MAGIC) || ++ header->h_blocks != cpu_to_le32(1)) { ++bad_block: ext3_error(sb, "ext3_xattr_set", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ /* Find the named attribute. */ ++ here = FIRST_ENTRY(bh); ++ while (!IS_LAST_ENTRY(here)) { ++ struct ext3_xattr_entry *next = EXT3_XATTR_NEXT(here); ++ if ((char *)next >= end) ++ goto bad_block; ++ if (!here->e_value_block && here->e_value_size) { ++ int offs = le16_to_cpu(here->e_value_offs); ++ if (offs < min_offs) ++ min_offs = offs; ++ } ++ not_found = name_index - here->e_name_index; ++ if (!not_found) ++ not_found = name_len - here->e_name_len; ++ if (!not_found) ++ not_found = memcmp(name, here->e_name,name_len); ++ if (not_found <= 0) ++ break; ++ here = next; ++ } ++ last = here; ++ /* We still need to compute min_offs and last. */ ++ while (!IS_LAST_ENTRY(last)) { ++ struct ext3_xattr_entry *next = EXT3_XATTR_NEXT(last); ++ if ((char *)next >= end) ++ goto bad_block; ++ if (!last->e_value_block && last->e_value_size) { ++ int offs = le16_to_cpu(last->e_value_offs); ++ if (offs < min_offs) ++ min_offs = offs; ++ } ++ last = next; ++ } ++ ++ /* Check whether we have enough space left. */ ++ free = min_offs - ((char*)last - (char*)header) - sizeof(__u32); ++ } else { ++ /* We will use a new extended attribute block. */ ++ free = sb->s_blocksize - ++ sizeof(struct ext3_xattr_header) - sizeof(__u32); ++ here = last = NULL; /* avoid gcc uninitialized warning. */ ++ } ++ ++ if (not_found) { ++ /* Request to remove a nonexistent attribute? */ ++ error = -ENOATTR; ++ if (flags & XATTR_REPLACE) ++ goto cleanup; ++ error = 0; ++ if (value == NULL) ++ goto cleanup; ++ else ++ free -= EXT3_XATTR_LEN(name_len); ++ } else { ++ /* Request to create an existing attribute? */ ++ error = -EEXIST; ++ if (flags & XATTR_CREATE) ++ goto cleanup; ++ if (!here->e_value_block && here->e_value_size) { ++ unsigned int size = le32_to_cpu(here->e_value_size); ++ ++ if (le16_to_cpu(here->e_value_offs) + size > ++ sb->s_blocksize || size > sb->s_blocksize) ++ goto bad_block; ++ free += EXT3_XATTR_SIZE(size); ++ } ++ } ++ free -= EXT3_XATTR_SIZE(value_len); ++ error = -ENOSPC; ++ if (free < 0) ++ goto cleanup; ++ ++ /* Here we know that we can set the new attribute. */ ++ ++ if (header) { ++ if (header->h_refcount == cpu_to_le32(1)) { ++ ea_bdebug(bh, "modifying in-place"); ++ ext3_xattr_cache_remove(bh); ++ error = ext3_journal_get_write_access(handle, bh); ++ if (error) ++ goto cleanup; ++ } else { ++ int offset; ++ ++ ea_bdebug(bh, "cloning"); ++ header = kmalloc(bh->b_size, GFP_KERNEL); ++ error = -ENOMEM; ++ if (header == NULL) ++ goto cleanup; ++ memcpy(header, HDR(bh), bh->b_size); ++ header->h_refcount = cpu_to_le32(1); ++ offset = (char *)header - bh->b_data; ++ here = ENTRY((char *)here + offset); ++ last = ENTRY((char *)last + offset); ++ } ++ } else { ++ /* Allocate a buffer where we construct the new block. */ ++ header = kmalloc(sb->s_blocksize, GFP_KERNEL); ++ error = -ENOMEM; ++ if (header == NULL) ++ goto cleanup; ++ memset(header, 0, sb->s_blocksize); ++ end = (char *)header + sb->s_blocksize; ++ header->h_magic = cpu_to_le32(EXT3_XATTR_MAGIC); ++ header->h_blocks = header->h_refcount = cpu_to_le32(1); ++ last = here = ENTRY(header+1); ++ } ++ ++ if (not_found) { ++ /* Insert the new name. */ ++ int size = EXT3_XATTR_LEN(name_len); ++ int rest = (char *)last - (char *)here; ++ memmove((char *)here + size, here, rest); ++ memset(here, 0, size); ++ here->e_name_index = name_index; ++ here->e_name_len = name_len; ++ memcpy(here->e_name, name, name_len); ++ } else { ++ /* Remove the old value. */ ++ if (!here->e_value_block && here->e_value_size) { ++ char *first_val = (char *)header + min_offs; ++ int offs = le16_to_cpu(here->e_value_offs); ++ char *val = (char *)header + offs; ++ size_t size = EXT3_XATTR_SIZE( ++ le32_to_cpu(here->e_value_size)); ++ memmove(first_val + size, first_val, val - first_val); ++ memset(first_val, 0, size); ++ here->e_value_offs = 0; ++ min_offs += size; ++ ++ /* Adjust all value offsets. */ ++ last = ENTRY(header+1); ++ while (!IS_LAST_ENTRY(last)) { ++ int o = le16_to_cpu(last->e_value_offs); ++ if (!last->e_value_block && o < offs) ++ last->e_value_offs = ++ cpu_to_le16(o + size); ++ last = EXT3_XATTR_NEXT(last); ++ } ++ } ++ if (value == NULL) { ++ /* Remove this attribute. */ ++ if (EXT3_XATTR_NEXT(ENTRY(header+1)) == last) { ++ /* This block is now empty. */ ++ error = ext3_xattr_set2(handle, inode, bh,NULL); ++ goto cleanup; ++ } else { ++ /* Remove the old name. */ ++ int size = EXT3_XATTR_LEN(name_len); ++ last = ENTRY((char *)last - size); ++ memmove(here, (char*)here + size, ++ (char*)last - (char*)here); ++ memset(last, 0, size); ++ } ++ } ++ } ++ ++ if (value != NULL) { ++ /* Insert the new value. */ ++ here->e_value_size = cpu_to_le32(value_len); ++ if (value_len) { ++ size_t size = EXT3_XATTR_SIZE(value_len); ++ char *val = (char *)header + min_offs - size; ++ here->e_value_offs = ++ cpu_to_le16((char *)val - (char *)header); ++ memset(val + size - EXT3_XATTR_PAD, 0, ++ EXT3_XATTR_PAD); /* Clear the pad bytes. */ ++ memcpy(val, value, value_len); ++ } ++ } ++ ext3_xattr_rehash(header, here); ++ ++ error = ext3_xattr_set2(handle, inode, bh, header); ++ ++cleanup: ++ brelse(bh); ++ if (!(bh && header == HDR(bh))) ++ kfree(header); ++ up(&ext3_xattr_sem); ++ ++ return error; ++} ++ ++/* ++ * Second half of ext3_xattr_set(): Update the file system. ++ */ ++static int ++ext3_xattr_set2(handle_t *handle, struct inode *inode, ++ struct buffer_head *old_bh, struct ext3_xattr_header *header) ++{ ++ struct super_block *sb = inode->i_sb; ++ struct buffer_head *new_bh = NULL; ++ int error; ++ ++ if (header) { ++ new_bh = ext3_xattr_cache_find(inode, header); ++ if (new_bh) { ++ /* ++ * We found an identical block in the cache. ++ * The old block will be released after updating ++ * the inode. ++ */ ++ ea_bdebug(old_bh, "reusing block %ld", ++ new_bh->b_blocknr); ++ ++ error = -EDQUOT; ++ if (ext3_xattr_quota_alloc(inode, 1)) ++ goto cleanup; ++ ++ error = ext3_journal_get_write_access(handle, new_bh); ++ if (error) ++ goto cleanup; ++ HDR(new_bh)->h_refcount = cpu_to_le32( ++ le32_to_cpu(HDR(new_bh)->h_refcount) + 1); ++ ea_bdebug(new_bh, "refcount now=%d", ++ le32_to_cpu(HDR(new_bh)->h_refcount)); ++ } else if (old_bh && header == HDR(old_bh)) { ++ /* Keep this block. */ ++ new_bh = old_bh; ++ ext3_xattr_cache_insert(new_bh); ++ } else { ++ /* We need to allocate a new block */ ++ int force = EXT3_I(inode)->i_file_acl != 0; ++ int block = ext3_xattr_new_block(handle, inode, ++ &error, force); ++ if (error) ++ goto cleanup; ++ ea_idebug(inode, "creating block %d", block); ++ ++ new_bh = sb_getblk(sb, block); ++ if (!new_bh) { ++getblk_failed: ext3_xattr_free_block(handle, inode, block); ++ error = -EIO; ++ goto cleanup; ++ } ++ lock_buffer(new_bh); ++ error = ext3_journal_get_create_access(handle, new_bh); ++ if (error) { ++ unlock_buffer(new_bh); ++ goto getblk_failed; ++ } ++ memcpy(new_bh->b_data, header, new_bh->b_size); ++ mark_buffer_uptodate(new_bh, 1); ++ unlock_buffer(new_bh); ++ ext3_xattr_cache_insert(new_bh); ++ ++ ext3_xattr_update_super_block(handle, sb); ++ } ++ error = ext3_journal_dirty_metadata(handle, new_bh); ++ if (error) ++ goto cleanup; ++ } ++ ++ /* Update the inode. */ ++ EXT3_I(inode)->i_file_acl = new_bh ? new_bh->b_blocknr : 0; ++ inode->i_ctime = CURRENT_TIME; ++ ext3_mark_inode_dirty(handle, inode); ++ if (IS_SYNC(inode)) ++ handle->h_sync = 1; ++ ++ error = 0; ++ if (old_bh && old_bh != new_bh) { ++ /* ++ * If there was an old block, and we are not still using it, ++ * we now release the old block. ++ */ ++ unsigned int refcount = le32_to_cpu(HDR(old_bh)->h_refcount); ++ ++ error = ext3_journal_get_write_access(handle, old_bh); ++ if (error) ++ goto cleanup; ++ if (refcount == 1) { ++ /* Free the old block. */ ++ ea_bdebug(old_bh, "freeing"); ++ ext3_xattr_free_block(handle, inode, old_bh->b_blocknr); ++ ++ /* ext3_forget() calls bforget() for us, but we ++ let our caller release old_bh, so we need to ++ duplicate the handle before. */ ++ get_bh(old_bh); ++ ext3_forget(handle, 1, inode, old_bh,old_bh->b_blocknr); ++ } else { ++ /* Decrement the refcount only. */ ++ refcount--; ++ HDR(old_bh)->h_refcount = cpu_to_le32(refcount); ++ ext3_xattr_quota_free(inode); ++ ext3_journal_dirty_metadata(handle, old_bh); ++ ea_bdebug(old_bh, "refcount now=%d", refcount); ++ } ++ } ++ ++cleanup: ++ if (old_bh != new_bh) ++ brelse(new_bh); ++ ++ return error; ++} ++ ++/* ++ * ext3_xattr_delete_inode() ++ * ++ * Free extended attribute resources associated with this inode. This ++ * is called immediately before an inode is freed. ++ */ ++void ++ext3_xattr_delete_inode(handle_t *handle, struct inode *inode) ++{ ++ struct buffer_head *bh; ++ unsigned int block = EXT3_I(inode)->i_file_acl; ++ ++ if (!block) ++ return; ++ down(&ext3_xattr_sem); ++ ++ bh = sb_bread(inode->i_sb, block); ++ if (!bh) { ++ ext3_error(inode->i_sb, "ext3_xattr_delete_inode", ++ "inode %ld: block %d read error", inode->i_ino, block); ++ goto cleanup; ++ } ++ ea_bdebug(bh, "b_count=%d", atomic_read(&(bh->b_count))); ++ if (HDR(bh)->h_magic != cpu_to_le32(EXT3_XATTR_MAGIC) || ++ HDR(bh)->h_blocks != cpu_to_le32(1)) { ++ ext3_error(inode->i_sb, "ext3_xattr_delete_inode", ++ "inode %ld: bad block %d", inode->i_ino, block); ++ goto cleanup; ++ } ++ ext3_journal_get_write_access(handle, bh); ++ ea_bdebug(bh, "refcount now=%d", le32_to_cpu(HDR(bh)->h_refcount) - 1); ++ if (HDR(bh)->h_refcount == cpu_to_le32(1)) { ++ ext3_xattr_cache_remove(bh); ++ ext3_xattr_free_block(handle, inode, block); ++ ext3_forget(handle, 1, inode, bh, block); ++ bh = NULL; ++ } else { ++ HDR(bh)->h_refcount = cpu_to_le32( ++ le32_to_cpu(HDR(bh)->h_refcount) - 1); ++ ext3_journal_dirty_metadata(handle, bh); ++ if (IS_SYNC(inode)) ++ handle->h_sync = 1; ++ ext3_xattr_quota_free(inode); ++ } ++ EXT3_I(inode)->i_file_acl = 0; ++ ++cleanup: ++ brelse(bh); ++ up(&ext3_xattr_sem); ++} ++ ++/* ++ * ext3_xattr_put_super() ++ * ++ * This is called when a file system is unmounted. ++ */ ++void ++ext3_xattr_put_super(struct super_block *sb) ++{ ++#ifdef CONFIG_EXT3_FS_XATTR_SHARING ++ mb_cache_shrink(ext3_xattr_cache, sb->s_dev); ++#endif ++} ++ ++#ifdef CONFIG_EXT3_FS_XATTR_SHARING ++ ++/* ++ * ext3_xattr_cache_insert() ++ * ++ * Create a new entry in the extended attribute cache, and insert ++ * it unless such an entry is already in the cache. ++ * ++ * Returns 0, or a negative error number on failure. ++ */ ++static int ++ext3_xattr_cache_insert(struct buffer_head *bh) ++{ ++ __u32 hash = le32_to_cpu(HDR(bh)->h_hash); ++ struct mb_cache_entry *ce; ++ int error; ++ ++ ce = mb_cache_entry_alloc(ext3_xattr_cache); ++ if (!ce) ++ return -ENOMEM; ++ error = mb_cache_entry_insert(ce, bh->b_dev, bh->b_blocknr, &hash); ++ if (error) { ++ mb_cache_entry_free(ce); ++ if (error == -EBUSY) { ++ ea_bdebug(bh, "already in cache (%d cache entries)", ++ atomic_read(&ext3_xattr_cache->c_entry_count)); ++ error = 0; ++ } ++ } else { ++ ea_bdebug(bh, "inserting [%x] (%d cache entries)", (int)hash, ++ atomic_read(&ext3_xattr_cache->c_entry_count)); ++ mb_cache_entry_release(ce); ++ } ++ return error; ++} ++ ++/* ++ * ext3_xattr_cmp() ++ * ++ * Compare two extended attribute blocks for equality. ++ * ++ * Returns 0 if the blocks are equal, 1 if they differ, and ++ * a negative error number on errors. ++ */ ++static int ++ext3_xattr_cmp(struct ext3_xattr_header *header1, ++ struct ext3_xattr_header *header2) ++{ ++ struct ext3_xattr_entry *entry1, *entry2; ++ ++ entry1 = ENTRY(header1+1); ++ entry2 = ENTRY(header2+1); ++ while (!IS_LAST_ENTRY(entry1)) { ++ if (IS_LAST_ENTRY(entry2)) ++ return 1; ++ if (entry1->e_hash != entry2->e_hash || ++ entry1->e_name_len != entry2->e_name_len || ++ entry1->e_value_size != entry2->e_value_size || ++ memcmp(entry1->e_name, entry2->e_name, entry1->e_name_len)) ++ return 1; ++ if (entry1->e_value_block != 0 || entry2->e_value_block != 0) ++ return -EIO; ++ if (memcmp((char *)header1 + le16_to_cpu(entry1->e_value_offs), ++ (char *)header2 + le16_to_cpu(entry2->e_value_offs), ++ le32_to_cpu(entry1->e_value_size))) ++ return 1; ++ ++ entry1 = EXT3_XATTR_NEXT(entry1); ++ entry2 = EXT3_XATTR_NEXT(entry2); ++ } ++ if (!IS_LAST_ENTRY(entry2)) ++ return 1; ++ return 0; ++} ++ ++/* ++ * ext3_xattr_cache_find() ++ * ++ * Find an identical extended attribute block. ++ * ++ * Returns a pointer to the block found, or NULL if such a block was ++ * not found or an error occurred. ++ */ ++static struct buffer_head * ++ext3_xattr_cache_find(struct inode *inode, struct ext3_xattr_header *header) ++{ ++ __u32 hash = le32_to_cpu(header->h_hash); ++ struct mb_cache_entry *ce; ++ ++ if (!header->h_hash) ++ return NULL; /* never share */ ++ ea_idebug(inode, "looking for cached blocks [%x]", (int)hash); ++ ce = mb_cache_entry_find_first(ext3_xattr_cache, 0, inode->i_dev, hash); ++ while (ce) { ++ struct buffer_head *bh = sb_bread(inode->i_sb, ce->e_block); ++ ++ if (!bh) { ++ ext3_error(inode->i_sb, "ext3_xattr_cache_find", ++ "inode %ld: block %ld read error", ++ inode->i_ino, ce->e_block); ++ } else if (le32_to_cpu(HDR(bh)->h_refcount) > ++ EXT3_XATTR_REFCOUNT_MAX) { ++ ea_idebug(inode, "block %ld refcount %d>%d",ce->e_block, ++ le32_to_cpu(HDR(bh)->h_refcount), ++ EXT3_XATTR_REFCOUNT_MAX); ++ } else if (!ext3_xattr_cmp(header, HDR(bh))) { ++ ea_bdebug(bh, "b_count=%d",atomic_read(&(bh->b_count))); ++ mb_cache_entry_release(ce); ++ return bh; ++ } ++ brelse(bh); ++ ce = mb_cache_entry_find_next(ce, 0, inode->i_dev, hash); ++ } ++ return NULL; ++} ++ ++/* ++ * ext3_xattr_cache_remove() ++ * ++ * Remove the cache entry of a block from the cache. Called when a ++ * block becomes invalid. ++ */ ++static void ++ext3_xattr_cache_remove(struct buffer_head *bh) ++{ ++ struct mb_cache_entry *ce; ++ ++ ce = mb_cache_entry_get(ext3_xattr_cache, bh->b_dev, bh->b_blocknr); ++ if (ce) { ++ ea_bdebug(bh, "removing (%d cache entries remaining)", ++ atomic_read(&ext3_xattr_cache->c_entry_count)-1); ++ mb_cache_entry_free(ce); ++ } else ++ ea_bdebug(bh, "no cache entry"); ++} ++ ++#define NAME_HASH_SHIFT 5 ++#define VALUE_HASH_SHIFT 16 ++ ++/* ++ * ext3_xattr_hash_entry() ++ * ++ * Compute the hash of an extended attribute. ++ */ ++static inline void ext3_xattr_hash_entry(struct ext3_xattr_header *header, ++ struct ext3_xattr_entry *entry) ++{ ++ __u32 hash = 0; ++ char *name = entry->e_name; ++ int n; ++ ++ for (n=0; n < entry->e_name_len; n++) { ++ hash = (hash << NAME_HASH_SHIFT) ^ ++ (hash >> (8*sizeof(hash) - NAME_HASH_SHIFT)) ^ ++ *name++; ++ } ++ ++ if (entry->e_value_block == 0 && entry->e_value_size != 0) { ++ __u32 *value = (__u32 *)((char *)header + ++ le16_to_cpu(entry->e_value_offs)); ++ for (n = (le32_to_cpu(entry->e_value_size) + ++ EXT3_XATTR_ROUND) >> EXT3_XATTR_PAD_BITS; n; n--) { ++ hash = (hash << VALUE_HASH_SHIFT) ^ ++ (hash >> (8*sizeof(hash) - VALUE_HASH_SHIFT)) ^ ++ le32_to_cpu(*value++); ++ } ++ } ++ entry->e_hash = cpu_to_le32(hash); ++} ++ ++#undef NAME_HASH_SHIFT ++#undef VALUE_HASH_SHIFT ++ ++#define BLOCK_HASH_SHIFT 16 ++ ++/* ++ * ext3_xattr_rehash() ++ * ++ * Re-compute the extended attribute hash value after an entry has changed. ++ */ ++static void ext3_xattr_rehash(struct ext3_xattr_header *header, ++ struct ext3_xattr_entry *entry) ++{ ++ struct ext3_xattr_entry *here; ++ __u32 hash = 0; ++ ++ ext3_xattr_hash_entry(header, entry); ++ here = ENTRY(header+1); ++ while (!IS_LAST_ENTRY(here)) { ++ if (!here->e_hash) { ++ /* Block is not shared if an entry's hash value == 0 */ ++ hash = 0; ++ break; ++ } ++ hash = (hash << BLOCK_HASH_SHIFT) ^ ++ (hash >> (8*sizeof(hash) - BLOCK_HASH_SHIFT)) ^ ++ le32_to_cpu(here->e_hash); ++ here = EXT3_XATTR_NEXT(here); ++ } ++ header->h_hash = cpu_to_le32(hash); ++} ++ ++#undef BLOCK_HASH_SHIFT ++ ++int __init ++init_ext3_xattr(void) ++{ ++ ext3_xattr_cache = mb_cache_create("ext3_xattr", NULL, ++ sizeof(struct mb_cache_entry) + ++ sizeof(struct mb_cache_entry_index), 1, 61); ++ if (!ext3_xattr_cache) ++ return -ENOMEM; ++ ++ return 0; ++} ++ ++void ++exit_ext3_xattr(void) ++{ ++ if (ext3_xattr_cache) ++ mb_cache_destroy(ext3_xattr_cache); ++ ext3_xattr_cache = NULL; ++} ++ ++#else /* CONFIG_EXT3_FS_XATTR_SHARING */ ++ ++int __init ++init_ext3_xattr(void) ++{ ++ return 0; ++} ++ ++void ++exit_ext3_xattr(void) ++{ ++} ++ ++#endif /* CONFIG_EXT3_FS_XATTR_SHARING */ +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/ext3/xattr_user.c 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,111 @@ ++/* ++ * linux/fs/ext3/xattr_user.c ++ * Handler for extended user attributes. ++ * ++ * Copyright (C) 2001 by Andreas Gruenbacher, ++ */ ++ ++#include ++#include ++#include ++#include ++#include ++#include ++ ++#ifdef CONFIG_EXT3_FS_POSIX_ACL ++# include ++#endif ++ ++#define XATTR_USER_PREFIX "user." ++ ++static size_t ++ext3_xattr_user_list(char *list, struct inode *inode, ++ const char *name, int name_len) ++{ ++ const int prefix_len = sizeof(XATTR_USER_PREFIX)-1; ++ ++ if (!test_opt(inode->i_sb, XATTR_USER)) ++ return 0; ++ ++ if (list) { ++ memcpy(list, XATTR_USER_PREFIX, prefix_len); ++ memcpy(list+prefix_len, name, name_len); ++ list[prefix_len + name_len] = '\0'; ++ } ++ return prefix_len + name_len + 1; ++} ++ ++static int ++ext3_xattr_user_get(struct inode *inode, const char *name, ++ void *buffer, size_t size) ++{ ++ int error; ++ ++ if (strcmp(name, "") == 0) ++ return -EINVAL; ++ if (!test_opt(inode->i_sb, XATTR_USER)) ++ return -ENOTSUP; ++#ifdef CONFIG_EXT3_FS_POSIX_ACL ++ error = ext3_permission_locked(inode, MAY_READ); ++#else ++ error = permission(inode, MAY_READ); ++#endif ++ if (error) ++ return error; ++ ++ return ext3_xattr_get(inode, EXT3_XATTR_INDEX_USER, name, ++ buffer, size); ++} ++ ++static int ++ext3_xattr_user_set(struct inode *inode, const char *name, ++ const void *value, size_t size, int flags) ++{ ++ handle_t *handle; ++ int error; ++ ++ if (strcmp(name, "") == 0) ++ return -EINVAL; ++ if (!test_opt(inode->i_sb, XATTR_USER)) ++ return -ENOTSUP; ++ if ( !S_ISREG(inode->i_mode) && ++ (!S_ISDIR(inode->i_mode) || inode->i_mode & S_ISVTX)) ++ return -EPERM; ++#ifdef CONFIG_EXT3_FS_POSIX_ACL ++ error = ext3_permission_locked(inode, MAY_WRITE); ++#else ++ error = permission(inode, MAY_WRITE); ++#endif ++ if (error) ++ return error; ++ ++ handle = ext3_journal_start(inode, EXT3_XATTR_TRANS_BLOCKS); ++ if (IS_ERR(handle)) ++ return PTR_ERR(handle); ++ error = ext3_xattr_set(handle, inode, EXT3_XATTR_INDEX_USER, name, ++ value, size, flags); ++ ext3_journal_stop(handle, inode); ++ ++ return error; ++} ++ ++struct ext3_xattr_handler ext3_xattr_user_handler = { ++ prefix: XATTR_USER_PREFIX, ++ list: ext3_xattr_user_list, ++ get: ext3_xattr_user_get, ++ set: ext3_xattr_user_set, ++}; ++ ++int __init ++init_ext3_xattr_user(void) ++{ ++ return ext3_xattr_register(EXT3_XATTR_INDEX_USER, ++ &ext3_xattr_user_handler); ++} ++ ++void ++exit_ext3_xattr_user(void) ++{ ++ ext3_xattr_unregister(EXT3_XATTR_INDEX_USER, ++ &ext3_xattr_user_handler); ++} +--- linux-2.4.22-ac1/fs/jfs/jfs_xattr.h~xattr-0.8.54-2.4.22-rh 2002-11-29 02:53:15.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/jfs/jfs_xattr.h 2003-09-25 23:57:02.000000000 +0400 +@@ -52,8 +52,10 @@ struct jfs_ea_list { + #define END_EALIST(ealist) \ + ((struct jfs_ea *) (((char *) (ealist)) + EALIST_SIZE(ealist))) + +-extern int __jfs_setxattr(struct inode *, const char *, void *, size_t, int); +-extern int jfs_setxattr(struct dentry *, const char *, void *, size_t, int); ++extern int __jfs_setxattr(struct inode *, const char *, const void *, size_t, ++ int); ++extern int jfs_setxattr(struct dentry *, const char *, const void *, size_t, ++ int); + extern ssize_t __jfs_getxattr(struct inode *, const char *, void *, size_t); + extern ssize_t jfs_getxattr(struct dentry *, const char *, void *, size_t); + extern ssize_t jfs_listxattr(struct dentry *, char *, size_t); +--- linux-2.4.22-ac1/fs/jfs/xattr.c~xattr-0.8.54-2.4.22-rh 2002-11-29 02:53:15.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/jfs/xattr.c 2003-09-25 23:57:02.000000000 +0400 +@@ -641,7 +641,7 @@ static int ea_put(struct inode *inode, s + } + + static int can_set_xattr(struct inode *inode, const char *name, +- void *value, size_t value_len) ++ const void *value, size_t value_len) + { + if (IS_RDONLY(inode)) + return -EROFS; +@@ -660,7 +660,7 @@ static int can_set_xattr(struct inode *i + return permission(inode, MAY_WRITE); + } + +-int __jfs_setxattr(struct inode *inode, const char *name, void *value, ++int __jfs_setxattr(struct inode *inode, const char *name, const void *value, + size_t value_len, int flags) + { + struct jfs_ea_list *ealist; +@@ -799,7 +799,7 @@ int __jfs_setxattr(struct inode *inode, + return rc; + } + +-int jfs_setxattr(struct dentry *dentry, const char *name, void *value, ++int jfs_setxattr(struct dentry *dentry, const char *name, const void *value, + size_t value_len, int flags) + { + if (value == NULL) { /* empty EA, do not remove */ +--- linux-2.4.22-ac1/fs/Makefile~xattr-0.8.54-2.4.22-rh 2003-09-25 14:50:00.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/fs/Makefile 2003-09-25 23:57:02.000000000 +0400 +@@ -82,6 +82,9 @@ obj-y += binfmt_script.o + + obj-$(CONFIG_BINFMT_ELF) += binfmt_elf.o + ++export-objs += mbcache.o ++obj-$(CONFIG_FS_MBCACHE) += mbcache.o ++ + # persistent filesystems + obj-y += $(join $(subdir-y),$(subdir-y:%=/%.o)) + +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/fs/mbcache.c 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,648 @@ ++/* ++ * linux/fs/mbcache.c ++ * (C) 2001-2002 Andreas Gruenbacher, ++ */ ++ ++/* ++ * Filesystem Meta Information Block Cache (mbcache) ++ * ++ * The mbcache caches blocks of block devices that need to be located ++ * by their device/block number, as well as by other criteria (such ++ * as the block's contents). ++ * ++ * There can only be one cache entry in a cache per device and block number. ++ * Additional indexes need not be unique in this sense. The number of ++ * additional indexes (=other criteria) can be hardwired at compile time ++ * or specified at cache create time. ++ * ++ * Each cache entry is of fixed size. An entry may be `valid' or `invalid' ++ * in the cache. A valid entry is in the main hash tables of the cache, ++ * and may also be in the lru list. An invalid entry is not in any hashes ++ * or lists. ++ * ++ * A valid cache entry is only in the lru list if no handles refer to it. ++ * Invalid cache entries will be freed when the last handle to the cache ++ * entry is released. Entries that cannot be freed immediately are put ++ * back on the lru list. ++ */ ++ ++#include ++#include ++ ++#include ++#include ++#include ++#include ++#include ++#include ++#include ++ ++ ++#ifdef MB_CACHE_DEBUG ++# define mb_debug(f...) do { \ ++ printk(KERN_DEBUG f); \ ++ printk("\n"); \ ++ } while (0) ++#define mb_assert(c) do { if (!(c)) \ ++ printk(KERN_ERR "assertion " #c " failed\n"); \ ++ } while(0) ++#else ++# define mb_debug(f...) do { } while(0) ++# define mb_assert(c) do { } while(0) ++#endif ++#define mb_error(f...) do { \ ++ printk(KERN_ERR f); \ ++ printk("\n"); \ ++ } while(0) ++ ++MODULE_AUTHOR("Andreas Gruenbacher "); ++MODULE_DESCRIPTION("Meta block cache (for extended attributes)"); ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,4,0) ++MODULE_LICENSE("GPL"); ++#endif ++ ++EXPORT_SYMBOL(mb_cache_create); ++EXPORT_SYMBOL(mb_cache_shrink); ++EXPORT_SYMBOL(mb_cache_destroy); ++EXPORT_SYMBOL(mb_cache_entry_alloc); ++EXPORT_SYMBOL(mb_cache_entry_insert); ++EXPORT_SYMBOL(mb_cache_entry_release); ++EXPORT_SYMBOL(mb_cache_entry_takeout); ++EXPORT_SYMBOL(mb_cache_entry_free); ++EXPORT_SYMBOL(mb_cache_entry_dup); ++EXPORT_SYMBOL(mb_cache_entry_get); ++#if !defined(MB_CACHE_INDEXES_COUNT) || (MB_CACHE_INDEXES_COUNT > 0) ++EXPORT_SYMBOL(mb_cache_entry_find_first); ++EXPORT_SYMBOL(mb_cache_entry_find_next); ++#endif ++ ++ ++/* ++ * Global data: list of all mbcache's, lru list, and a spinlock for ++ * accessing cache data structures on SMP machines. The lru list is ++ * global across all mbcaches. ++ */ ++ ++static LIST_HEAD(mb_cache_list); ++static LIST_HEAD(mb_cache_lru_list); ++static spinlock_t mb_cache_spinlock = SPIN_LOCK_UNLOCKED; ++ ++static inline int ++mb_cache_indexes(struct mb_cache *cache) ++{ ++#ifdef MB_CACHE_INDEXES_COUNT ++ return MB_CACHE_INDEXES_COUNT; ++#else ++ return cache->c_indexes_count; ++#endif ++} ++ ++/* ++ * What the mbcache registers as to get shrunk dynamically. ++ */ ++ ++static void ++mb_cache_memory_pressure(int priority, unsigned int gfp_mask); ++ ++static struct cache_definition mb_cache_definition = { ++ "mb_cache", ++ mb_cache_memory_pressure ++}; ++ ++ ++static inline int ++__mb_cache_entry_is_hashed(struct mb_cache_entry *ce) ++{ ++ return !list_empty(&ce->e_block_list); ++} ++ ++ ++static inline void ++__mb_cache_entry_unhash(struct mb_cache_entry *ce) ++{ ++ int n; ++ ++ if (__mb_cache_entry_is_hashed(ce)) { ++ list_del_init(&ce->e_block_list); ++ for (n=0; ne_cache); n++) ++ list_del(&ce->e_indexes[n].o_list); ++ } ++} ++ ++ ++static inline void ++__mb_cache_entry_forget(struct mb_cache_entry *ce, int gfp_mask) ++{ ++ struct mb_cache *cache = ce->e_cache; ++ ++ mb_assert(atomic_read(&ce->e_used) == 0); ++ if (cache->c_op.free && cache->c_op.free(ce, gfp_mask)) { ++ /* free failed -- put back on the lru list ++ for freeing later. */ ++ spin_lock(&mb_cache_spinlock); ++ list_add(&ce->e_lru_list, &mb_cache_lru_list); ++ spin_unlock(&mb_cache_spinlock); ++ } else { ++ kmem_cache_free(cache->c_entry_cache, ce); ++ atomic_dec(&cache->c_entry_count); ++ } ++} ++ ++ ++static inline void ++__mb_cache_entry_release_unlock(struct mb_cache_entry *ce) ++{ ++ if (atomic_dec_and_test(&ce->e_used)) { ++ if (__mb_cache_entry_is_hashed(ce)) ++ list_add_tail(&ce->e_lru_list, &mb_cache_lru_list); ++ else { ++ spin_unlock(&mb_cache_spinlock); ++ __mb_cache_entry_forget(ce, GFP_KERNEL); ++ return; ++ } ++ } ++ spin_unlock(&mb_cache_spinlock); ++} ++ ++ ++/* ++ * mb_cache_memory_pressure() memory pressure callback ++ * ++ * This function is called by the kernel memory management when memory ++ * gets low. ++ * ++ * @priority: Amount by which to shrink the cache (0 = highes priority) ++ * @gfp_mask: (ignored) ++ */ ++static void ++mb_cache_memory_pressure(int priority, unsigned int gfp_mask) ++{ ++ LIST_HEAD(free_list); ++ struct list_head *l, *ltmp; ++ int count = 0; ++ ++ spin_lock(&mb_cache_spinlock); ++ list_for_each(l, &mb_cache_list) { ++ struct mb_cache *cache = ++ list_entry(l, struct mb_cache, c_cache_list); ++ mb_debug("cache %s (%d)", cache->c_name, ++ atomic_read(&cache->c_entry_count)); ++ count += atomic_read(&cache->c_entry_count); ++ } ++ mb_debug("trying to free %d of %d entries", ++ count / (priority ? priority : 1), count); ++ if (priority) ++ count /= priority; ++ while (count-- && !list_empty(&mb_cache_lru_list)) { ++ struct mb_cache_entry *ce = ++ list_entry(mb_cache_lru_list.next, ++ struct mb_cache_entry, e_lru_list); ++ list_del(&ce->e_lru_list); ++ __mb_cache_entry_unhash(ce); ++ list_add_tail(&ce->e_lru_list, &free_list); ++ } ++ spin_unlock(&mb_cache_spinlock); ++ list_for_each_safe(l, ltmp, &free_list) { ++ __mb_cache_entry_forget(list_entry(l, struct mb_cache_entry, ++ e_lru_list), gfp_mask); ++ } ++} ++ ++ ++/* ++ * mb_cache_create() create a new cache ++ * ++ * All entries in one cache are equal size. Cache entries may be from ++ * multiple devices. If this is the first mbcache created, registers ++ * the cache with kernel memory management. Returns NULL if no more ++ * memory was available. ++ * ++ * @name: name of the cache (informal) ++ * @cache_op: contains the callback called when freeing a cache entry ++ * @entry_size: The size of a cache entry, including ++ * struct mb_cache_entry ++ * @indexes_count: number of additional indexes in the cache. Must equal ++ * MB_CACHE_INDEXES_COUNT if the number of indexes is ++ * hardwired. ++ * @bucket_count: number of hash buckets ++ */ ++struct mb_cache * ++mb_cache_create(const char *name, struct mb_cache_op *cache_op, ++ size_t entry_size, int indexes_count, int bucket_count) ++{ ++ int m=0, n; ++ struct mb_cache *cache = NULL; ++ ++ if(entry_size < sizeof(struct mb_cache_entry) + ++ indexes_count * sizeof(struct mb_cache_entry_index)) ++ return NULL; ++ ++ MOD_INC_USE_COUNT; ++ cache = kmalloc(sizeof(struct mb_cache) + ++ indexes_count * sizeof(struct list_head), GFP_KERNEL); ++ if (!cache) ++ goto fail; ++ cache->c_name = name; ++ cache->c_op.free = NULL; ++ if (cache_op) ++ cache->c_op.free = cache_op->free; ++ atomic_set(&cache->c_entry_count, 0); ++ cache->c_bucket_count = bucket_count; ++#ifdef MB_CACHE_INDEXES_COUNT ++ mb_assert(indexes_count == MB_CACHE_INDEXES_COUNT); ++#else ++ cache->c_indexes_count = indexes_count; ++#endif ++ cache->c_block_hash = kmalloc(bucket_count * sizeof(struct list_head), ++ GFP_KERNEL); ++ if (!cache->c_block_hash) ++ goto fail; ++ for (n=0; nc_block_hash[n]); ++ for (m=0; mc_indexes_hash[m] = kmalloc(bucket_count * ++ sizeof(struct list_head), ++ GFP_KERNEL); ++ if (!cache->c_indexes_hash[m]) ++ goto fail; ++ for (n=0; nc_indexes_hash[m][n]); ++ } ++ cache->c_entry_cache = kmem_cache_create(name, entry_size, 0, ++ 0 /*SLAB_POISON | SLAB_RED_ZONE*/, NULL, NULL); ++ if (!cache->c_entry_cache) ++ goto fail; ++ ++ spin_lock(&mb_cache_spinlock); ++ list_add(&cache->c_cache_list, &mb_cache_list); ++ spin_unlock(&mb_cache_spinlock); ++ return cache; ++ ++fail: ++ if (cache) { ++ while (--m >= 0) ++ kfree(cache->c_indexes_hash[m]); ++ if (cache->c_block_hash) ++ kfree(cache->c_block_hash); ++ kfree(cache); ++ } ++ MOD_DEC_USE_COUNT; ++ return NULL; ++} ++ ++ ++/* ++ * mb_cache_shrink() ++ * ++ * Removes all cache entires of a device from the cache. All cache entries ++ * currently in use cannot be freed, and thus remain in the cache. ++ * ++ * @cache: which cache to shrink ++ * @dev: which device's cache entries to shrink ++ */ ++void ++mb_cache_shrink(struct mb_cache *cache, kdev_t dev) ++{ ++ LIST_HEAD(free_list); ++ struct list_head *l, *ltmp; ++ ++ spin_lock(&mb_cache_spinlock); ++ list_for_each_safe(l, ltmp, &mb_cache_lru_list) { ++ struct mb_cache_entry *ce = ++ list_entry(l, struct mb_cache_entry, e_lru_list); ++ if (ce->e_dev == dev) { ++ list_del(&ce->e_lru_list); ++ list_add_tail(&ce->e_lru_list, &free_list); ++ __mb_cache_entry_unhash(ce); ++ } ++ } ++ spin_unlock(&mb_cache_spinlock); ++ list_for_each_safe(l, ltmp, &free_list) { ++ __mb_cache_entry_forget(list_entry(l, struct mb_cache_entry, ++ e_lru_list), GFP_KERNEL); ++ } ++} ++ ++ ++/* ++ * mb_cache_destroy() ++ * ++ * Shrinks the cache to its minimum possible size (hopefully 0 entries), ++ * and then destroys it. If this was the last mbcache, un-registers the ++ * mbcache from kernel memory management. ++ */ ++void ++mb_cache_destroy(struct mb_cache *cache) ++{ ++ LIST_HEAD(free_list); ++ struct list_head *l, *ltmp; ++ int n; ++ ++ spin_lock(&mb_cache_spinlock); ++ list_for_each_safe(l, ltmp, &mb_cache_lru_list) { ++ struct mb_cache_entry *ce = ++ list_entry(l, struct mb_cache_entry, e_lru_list); ++ if (ce->e_cache == cache) { ++ list_del(&ce->e_lru_list); ++ list_add_tail(&ce->e_lru_list, &free_list); ++ __mb_cache_entry_unhash(ce); ++ } ++ } ++ list_del(&cache->c_cache_list); ++ spin_unlock(&mb_cache_spinlock); ++ list_for_each_safe(l, ltmp, &free_list) { ++ __mb_cache_entry_forget(list_entry(l, struct mb_cache_entry, ++ e_lru_list), GFP_KERNEL); ++ } ++ ++ if (atomic_read(&cache->c_entry_count) > 0) { ++ mb_error("cache %s: %d orphaned entries", ++ cache->c_name, ++ atomic_read(&cache->c_entry_count)); ++ } ++ ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,3,0)) ++ /* We don't have kmem_cache_destroy() in 2.2.x */ ++ kmem_cache_shrink(cache->c_entry_cache); ++#else ++ kmem_cache_destroy(cache->c_entry_cache); ++#endif ++ for (n=0; n < mb_cache_indexes(cache); n++) ++ kfree(cache->c_indexes_hash[n]); ++ kfree(cache->c_block_hash); ++ kfree(cache); ++ ++ MOD_DEC_USE_COUNT; ++} ++ ++ ++/* ++ * mb_cache_entry_alloc() ++ * ++ * Allocates a new cache entry. The new entry will not be valid initially, ++ * and thus cannot be looked up yet. It should be filled with data, and ++ * then inserted into the cache using mb_cache_entry_insert(). Returns NULL ++ * if no more memory was available. ++ */ ++struct mb_cache_entry * ++mb_cache_entry_alloc(struct mb_cache *cache) ++{ ++ struct mb_cache_entry *ce; ++ ++ atomic_inc(&cache->c_entry_count); ++ ce = kmem_cache_alloc(cache->c_entry_cache, GFP_KERNEL); ++ if (ce) { ++ INIT_LIST_HEAD(&ce->e_lru_list); ++ INIT_LIST_HEAD(&ce->e_block_list); ++ ce->e_cache = cache; ++ atomic_set(&ce->e_used, 1); ++ } ++ return ce; ++} ++ ++ ++/* ++ * mb_cache_entry_insert() ++ * ++ * Inserts an entry that was allocated using mb_cache_entry_alloc() into ++ * the cache. After this, the cache entry can be looked up, but is not yet ++ * in the lru list as the caller still holds a handle to it. Returns 0 on ++ * success, or -EBUSY if a cache entry for that device + inode exists ++ * already (this may happen after a failed lookup, if another process has ++ * inserted the same cache entry in the meantime). ++ * ++ * @dev: device the cache entry belongs to ++ * @block: block number ++ * @keys: array of additional keys. There must be indexes_count entries ++ * in the array (as specified when creating the cache). ++ */ ++int ++mb_cache_entry_insert(struct mb_cache_entry *ce, kdev_t dev, ++ unsigned long block, unsigned int keys[]) ++{ ++ struct mb_cache *cache = ce->e_cache; ++ unsigned int bucket = (HASHDEV(dev) + block) % cache->c_bucket_count; ++ struct list_head *l; ++ int error = -EBUSY, n; ++ ++ spin_lock(&mb_cache_spinlock); ++ list_for_each(l, &cache->c_block_hash[bucket]) { ++ struct mb_cache_entry *ce = ++ list_entry(l, struct mb_cache_entry, e_block_list); ++ if (ce->e_dev == dev && ce->e_block == block) ++ goto out; ++ } ++ __mb_cache_entry_unhash(ce); ++ ce->e_dev = dev; ++ ce->e_block = block; ++ list_add(&ce->e_block_list, &cache->c_block_hash[bucket]); ++ for (n=0; ne_indexes[n].o_key = keys[n]; ++ bucket = keys[n] % cache->c_bucket_count; ++ list_add(&ce->e_indexes[n].o_list, ++ &cache->c_indexes_hash[n][bucket]); ++ } ++out: ++ spin_unlock(&mb_cache_spinlock); ++ return error; ++} ++ ++ ++/* ++ * mb_cache_entry_release() ++ * ++ * Release a handle to a cache entry. When the last handle to a cache entry ++ * is released it is either freed (if it is invalid) or otherwise inserted ++ * in to the lru list. ++ */ ++void ++mb_cache_entry_release(struct mb_cache_entry *ce) ++{ ++ spin_lock(&mb_cache_spinlock); ++ __mb_cache_entry_release_unlock(ce); ++} ++ ++ ++/* ++ * mb_cache_entry_takeout() ++ * ++ * Take a cache entry out of the cache, making it invalid. The entry can later ++ * be re-inserted using mb_cache_entry_insert(), or released using ++ * mb_cache_entry_release(). ++ */ ++void ++mb_cache_entry_takeout(struct mb_cache_entry *ce) ++{ ++ spin_lock(&mb_cache_spinlock); ++ mb_assert(list_empty(&ce->e_lru_list)); ++ __mb_cache_entry_unhash(ce); ++ spin_unlock(&mb_cache_spinlock); ++} ++ ++ ++/* ++ * mb_cache_entry_free() ++ * ++ * This is equivalent to the sequence mb_cache_entry_takeout() -- ++ * mb_cache_entry_release(). ++ */ ++void ++mb_cache_entry_free(struct mb_cache_entry *ce) ++{ ++ spin_lock(&mb_cache_spinlock); ++ mb_assert(list_empty(&ce->e_lru_list)); ++ __mb_cache_entry_unhash(ce); ++ __mb_cache_entry_release_unlock(ce); ++} ++ ++ ++/* ++ * mb_cache_entry_dup() ++ * ++ * Duplicate a handle to a cache entry (does not duplicate the cache entry ++ * itself). After the call, both the old and the new handle must be released. ++ */ ++struct mb_cache_entry * ++mb_cache_entry_dup(struct mb_cache_entry *ce) ++{ ++ atomic_inc(&ce->e_used); ++ return ce; ++} ++ ++ ++/* ++ * mb_cache_entry_get() ++ * ++ * Get a cache entry by device / block number. (There can only be one entry ++ * in the cache per device and block.) Returns NULL if no such cache entry ++ * exists. ++ */ ++struct mb_cache_entry * ++mb_cache_entry_get(struct mb_cache *cache, kdev_t dev, unsigned long block) ++{ ++ unsigned int bucket = (HASHDEV(dev) + block) % cache->c_bucket_count; ++ struct list_head *l; ++ struct mb_cache_entry *ce; ++ ++ spin_lock(&mb_cache_spinlock); ++ list_for_each(l, &cache->c_block_hash[bucket]) { ++ ce = list_entry(l, struct mb_cache_entry, e_block_list); ++ if (ce->e_dev == dev && ce->e_block == block) { ++ if (!list_empty(&ce->e_lru_list)) ++ list_del_init(&ce->e_lru_list); ++ atomic_inc(&ce->e_used); ++ goto cleanup; ++ } ++ } ++ ce = NULL; ++ ++cleanup: ++ spin_unlock(&mb_cache_spinlock); ++ return ce; ++} ++ ++#if !defined(MB_CACHE_INDEXES_COUNT) || (MB_CACHE_INDEXES_COUNT > 0) ++ ++static struct mb_cache_entry * ++__mb_cache_entry_find(struct list_head *l, struct list_head *head, ++ int index, kdev_t dev, unsigned int key) ++{ ++ while (l != head) { ++ struct mb_cache_entry *ce = ++ list_entry(l, struct mb_cache_entry, ++ e_indexes[index].o_list); ++ if (ce->e_dev == dev && ce->e_indexes[index].o_key == key) { ++ if (!list_empty(&ce->e_lru_list)) ++ list_del_init(&ce->e_lru_list); ++ atomic_inc(&ce->e_used); ++ return ce; ++ } ++ l = l->next; ++ } ++ return NULL; ++} ++ ++ ++/* ++ * mb_cache_entry_find_first() ++ * ++ * Find the first cache entry on a given device with a certain key in ++ * an additional index. Additonal matches can be found with ++ * mb_cache_entry_find_next(). Returns NULL if no match was found. ++ * ++ * @cache: the cache to search ++ * @index: the number of the additonal index to search (0<=indexc_bucket_count; ++ struct list_head *l; ++ struct mb_cache_entry *ce; ++ ++ mb_assert(index < mb_cache_indexes(cache)); ++ spin_lock(&mb_cache_spinlock); ++ l = cache->c_indexes_hash[index][bucket].next; ++ ce = __mb_cache_entry_find(l, &cache->c_indexes_hash[index][bucket], ++ index, dev, key); ++ spin_unlock(&mb_cache_spinlock); ++ return ce; ++} ++ ++ ++/* ++ * mb_cache_entry_find_next() ++ * ++ * Find the next cache entry on a given device with a certain key in an ++ * additional index. Returns NULL if no match could be found. The previous ++ * entry is atomatically released, so that mb_cache_entry_find_next() can ++ * be called like this: ++ * ++ * entry = mb_cache_entry_find_first(); ++ * while (entry) { ++ * ... ++ * entry = mb_cache_entry_find_next(entry, ...); ++ * } ++ * ++ * @prev: The previous match ++ * @index: the number of the additonal index to search (0<=indexe_cache; ++ unsigned int bucket = key % cache->c_bucket_count; ++ struct list_head *l; ++ struct mb_cache_entry *ce; ++ ++ mb_assert(index < mb_cache_indexes(cache)); ++ spin_lock(&mb_cache_spinlock); ++ l = prev->e_indexes[index].o_list.next; ++ ce = __mb_cache_entry_find(l, &cache->c_indexes_hash[index][bucket], ++ index, dev, key); ++ __mb_cache_entry_release_unlock(prev); ++ return ce; ++} ++ ++#endif /* !defined(MB_CACHE_INDEXES_COUNT) || (MB_CACHE_INDEXES_COUNT > 0) */ ++ ++static int __init init_mbcache(void) ++{ ++ register_cache(&mb_cache_definition); ++ return 0; ++} ++ ++static void __exit exit_mbcache(void) ++{ ++ unregister_cache(&mb_cache_definition); ++} ++ ++module_init(init_mbcache) ++module_exit(exit_mbcache) ++ +--- linux-2.4.22-ac1/include/asm-arm/unistd.h~xattr-0.8.54-2.4.22-rh 2003-08-25 15:44:43.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/asm-arm/unistd.h 2003-09-25 23:57:02.000000000 +0400 +@@ -250,7 +250,6 @@ + #define __NR_security (__NR_SYSCALL_BASE+223) + #define __NR_gettid (__NR_SYSCALL_BASE+224) + #define __NR_readahead (__NR_SYSCALL_BASE+225) +-#if 0 /* allocated in 2.5 */ + #define __NR_setxattr (__NR_SYSCALL_BASE+226) + #define __NR_lsetxattr (__NR_SYSCALL_BASE+227) + #define __NR_fsetxattr (__NR_SYSCALL_BASE+228) +@@ -263,7 +262,6 @@ + #define __NR_removexattr (__NR_SYSCALL_BASE+235) + #define __NR_lremovexattr (__NR_SYSCALL_BASE+236) + #define __NR_fremovexattr (__NR_SYSCALL_BASE+237) +-#endif + #define __NR_tkill (__NR_SYSCALL_BASE+238) + #if 0 /* allocated in 2.5 */ + #define __NR_sendfile64 (__NR_SYSCALL_BASE+239) +--- linux-2.4.22-ac1/include/asm-ppc64/unistd.h~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:38.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/asm-ppc64/unistd.h 2003-09-25 23:57:02.000000000 +0400 +@@ -218,6 +218,7 @@ + #define __NR_mincore 206 + #define __NR_gettid 207 + #define __NR_tkill 208 ++#endif + #define __NR_setxattr 209 + #define __NR_lsetxattr 210 + #define __NR_fsetxattr 211 +@@ -230,6 +231,7 @@ + #define __NR_removexattr 218 + #define __NR_lremovexattr 219 + #define __NR_fremovexattr 220 ++#if 0 /* Reserved syscalls */ + #define __NR_futex 221 + #define __NR_sched_setaffinity 222 + #define __NR_sched_getaffinity 223 +--- linux-2.4.22-ac1/include/asm-s390/unistd.h~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:38.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/asm-s390/unistd.h 2003-09-26 00:14:23.000000000 +0400 +@@ -213,6 +213,19 @@ + #define __NR_getdents64 220 + #define __NR_fcntl64 221 + #define __NR_readahead 222 ++#define __NR_setxattr 224 ++#define __NR_lsetxattr 225 ++#define __NR_fsetxattr 226 ++#define __NR_getxattr 227 ++#define __NR_lgetxattr 228 ++#define __NR_fgetxattr 229 ++#define __NR_listxattr 230 ++#define __NR_llistxattr 231 ++#define __NR_flistxattr 232 ++#define __NR_removexattr 233 ++#define __NR_lremovexattr 234 ++#define __NR_fremovexattr 235 ++ + /* + * Numbers 224-235 are reserved for posix acl + */ +--- linux-2.4.22-ac1/include/asm-s390x/unistd.h~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:38.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/asm-s390x/unistd.h 2003-09-26 00:15:11.000000000 +0400 +@@ -181,6 +181,19 @@ + #define __NR_mincore 218 + #define __NR_madvise 219 + #define __NR_readahead 222 ++#define __NR_setxattr 224 ++#define __NR_lsetxattr 225 ++#define __NR_fsetxattr 226 ++#define __NR_getxattr 227 ++#define __NR_lgetxattr 228 ++#define __NR_fgetxattr 229 ++#define __NR_listxattr 230 ++#define __NR_llistxattr 231 ++#define __NR_flistxattr 232 ++#define __NR_removexattr 233 ++#define __NR_lremovexattr 234 ++#define __NR_fremovexattr 235 ++ + /* + * Numbers 224-235 are reserved for posix acl + */ +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/include/linux/cache_def.h 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,15 @@ ++/* ++ * linux/cache_def.h ++ * Handling of caches defined in drivers, filesystems, ... ++ * ++ * Copyright (C) 2002 by Andreas Gruenbacher, ++ */ ++ ++struct cache_definition { ++ const char *name; ++ void (*shrink)(int, unsigned int); ++ struct list_head link; ++}; ++ ++extern void register_cache(struct cache_definition *); ++extern void unregister_cache(struct cache_definition *); +--- linux-2.4.22-ac1/include/linux/errno.h~xattr-0.8.54-2.4.22-rh 2001-02-10 01:46:13.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/include/linux/errno.h 2003-09-25 23:57:02.000000000 +0400 +@@ -23,4 +23,8 @@ + + #endif + ++/* Defined for extended attributes */ ++#define ENOATTR ENODATA /* No such attribute */ ++#define ENOTSUP EOPNOTSUPP /* Operation not supported */ ++ + #endif +--- linux-2.4.22-ac1/include/linux/ext2_fs.h~xattr-0.8.54-2.4.22-rh 2003-06-13 18:51:38.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext2_fs.h 2003-09-25 23:57:02.000000000 +0400 +@@ -57,8 +57,6 @@ + */ + #define EXT2_BAD_INO 1 /* Bad blocks inode */ + #define EXT2_ROOT_INO 2 /* Root inode */ +-#define EXT2_ACL_IDX_INO 3 /* ACL inode */ +-#define EXT2_ACL_DATA_INO 4 /* ACL inode */ + #define EXT2_BOOT_LOADER_INO 5 /* Boot loader inode */ + #define EXT2_UNDEL_DIR_INO 6 /* Undelete directory inode */ + +@@ -86,7 +84,6 @@ + #else + # define EXT2_BLOCK_SIZE(s) (EXT2_MIN_BLOCK_SIZE << (s)->s_log_block_size) + #endif +-#define EXT2_ACLE_PER_BLOCK(s) (EXT2_BLOCK_SIZE(s) / sizeof (struct ext2_acl_entry)) + #define EXT2_ADDR_PER_BLOCK(s) (EXT2_BLOCK_SIZE(s) / sizeof (__u32)) + #ifdef __KERNEL__ + # define EXT2_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits) +@@ -121,28 +118,6 @@ + #endif + + /* +- * ACL structures +- */ +-struct ext2_acl_header /* Header of Access Control Lists */ +-{ +- __u32 aclh_size; +- __u32 aclh_file_count; +- __u32 aclh_acle_count; +- __u32 aclh_first_acle; +-}; +- +-struct ext2_acl_entry /* Access Control List Entry */ +-{ +- __u32 acle_size; +- __u16 acle_perms; /* Access permissions */ +- __u16 acle_type; /* Type of entry */ +- __u16 acle_tag; /* User or group identity */ +- __u16 acle_pad1; +- __u32 acle_next; /* Pointer on next entry for the */ +- /* same inode or on next free entry */ +-}; +- +-/* + * Structure of a blocks group descriptor + */ + struct ext2_group_desc +@@ -314,6 +289,7 @@ struct ext2_inode { + #define EXT2_MOUNT_ERRORS_PANIC 0x0040 /* Panic on errors */ + #define EXT2_MOUNT_MINIX_DF 0x0080 /* Mimics the Minix statfs */ + #define EXT2_MOUNT_NO_UID32 0x0200 /* Disable 32-bit UIDs */ ++#define EXT2_MOUNT_XATTR_USER 0x4000 /* Extended user attributes */ + + #define clear_opt(o, opt) o &= ~EXT2_MOUNT_##opt + #define set_opt(o, opt) o |= EXT2_MOUNT_##opt +@@ -397,6 +373,7 @@ struct ext2_super_block { + + #ifdef __KERNEL__ + #define EXT2_SB(sb) (&((sb)->u.ext2_sb)) ++#define EXT2_I(inode) (&((inode)->u.ext2_i)) + #else + /* Assume that user mode programs are passing in an ext2fs superblock, not + * a kernel struct super_block. This will allow us to call the feature-test +@@ -466,7 +443,7 @@ struct ext2_super_block { + #define EXT3_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 + #define EXT2_FEATURE_INCOMPAT_ANY 0xffffffff + +-#define EXT2_FEATURE_COMPAT_SUPP 0 ++#define EXT2_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR + #define EXT2_FEATURE_INCOMPAT_SUPP EXT2_FEATURE_INCOMPAT_FILETYPE + #define EXT2_FEATURE_RO_COMPAT_SUPP (EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER| \ + EXT2_FEATURE_RO_COMPAT_LARGE_FILE| \ +@@ -624,8 +601,10 @@ extern struct address_space_operations e + + /* namei.c */ + extern struct inode_operations ext2_dir_inode_operations; ++extern struct inode_operations ext2_special_inode_operations; + + /* symlink.c */ ++extern struct inode_operations ext2_symlink_inode_operations; + extern struct inode_operations ext2_fast_symlink_inode_operations; + + #endif /* __KERNEL__ */ +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/include/linux/ext2_xattr.h 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,157 @@ ++/* ++ File: linux/ext2_xattr.h ++ ++ On-disk format of extended attributes for the ext2 filesystem. ++ ++ (C) 2001 Andreas Gruenbacher, ++*/ ++ ++#include ++#include ++#include ++ ++/* Magic value in attribute blocks */ ++#define EXT2_XATTR_MAGIC 0xEA020000 ++ ++/* Maximum number of references to one attribute block */ ++#define EXT2_XATTR_REFCOUNT_MAX 1024 ++ ++/* Name indexes */ ++#define EXT2_XATTR_INDEX_MAX 10 ++#define EXT2_XATTR_INDEX_USER 1 ++#define EXT2_XATTR_INDEX_POSIX_ACL_ACCESS 2 ++#define EXT2_XATTR_INDEX_POSIX_ACL_DEFAULT 3 ++ ++struct ext2_xattr_header { ++ __u32 h_magic; /* magic number for identification */ ++ __u32 h_refcount; /* reference count */ ++ __u32 h_blocks; /* number of disk blocks used */ ++ __u32 h_hash; /* hash value of all attributes */ ++ __u32 h_reserved[4]; /* zero right now */ ++}; ++ ++struct ext2_xattr_entry { ++ __u8 e_name_len; /* length of name */ ++ __u8 e_name_index; /* attribute name index */ ++ __u16 e_value_offs; /* offset in disk block of value */ ++ __u32 e_value_block; /* disk block attribute is stored on (n/i) */ ++ __u32 e_value_size; /* size of attribute value */ ++ __u32 e_hash; /* hash value of name and value */ ++ char e_name[0]; /* attribute name */ ++}; ++ ++#define EXT2_XATTR_PAD_BITS 2 ++#define EXT2_XATTR_PAD (1<e_name_len)) ) ++#define EXT2_XATTR_SIZE(size) \ ++ (((size) + EXT2_XATTR_ROUND) & ~EXT2_XATTR_ROUND) ++ ++#ifdef __KERNEL__ ++ ++# ifdef CONFIG_EXT2_FS_XATTR ++ ++struct ext2_xattr_handler { ++ char *prefix; ++ size_t (*list)(char *list, struct inode *inode, const char *name, ++ int name_len); ++ int (*get)(struct inode *inode, const char *name, void *buffer, ++ size_t size); ++ int (*set)(struct inode *inode, const char *name, const void *buffer, ++ size_t size, int flags); ++}; ++ ++extern int ext2_xattr_register(int, struct ext2_xattr_handler *); ++extern void ext2_xattr_unregister(int, struct ext2_xattr_handler *); ++ ++extern int ext2_setxattr(struct dentry *, const char *, const void *, size_t, int); ++extern ssize_t ext2_getxattr(struct dentry *, const char *, void *, size_t); ++extern ssize_t ext2_listxattr(struct dentry *, char *, size_t); ++extern int ext2_removexattr(struct dentry *, const char *); ++ ++extern int ext2_xattr_get(struct inode *, int, const char *, void *, size_t); ++extern int ext2_xattr_list(struct inode *, char *, size_t); ++extern int ext2_xattr_set(struct inode *, int, const char *, const void *, size_t, int); ++ ++extern void ext2_xattr_delete_inode(struct inode *); ++extern void ext2_xattr_put_super(struct super_block *); ++ ++extern int init_ext2_xattr(void) __init; ++extern void exit_ext2_xattr(void); ++ ++# else /* CONFIG_EXT2_FS_XATTR */ ++# define ext2_setxattr NULL ++# define ext2_getxattr NULL ++# define ext2_listxattr NULL ++# define ext2_removexattr NULL ++ ++static inline int ++ext2_xattr_get(struct inode *inode, int name_index, ++ const char *name, void *buffer, size_t size) ++{ ++ return -ENOTSUP; ++} ++ ++static inline int ++ext2_xattr_list(struct inode *inode, char *buffer, size_t size) ++{ ++ return -ENOTSUP; ++} ++ ++static inline int ++ext2_xattr_set(struct inode *inode, int name_index, const char *name, ++ const void *value, size_t size, int flags) ++{ ++ return -ENOTSUP; ++} ++ ++static inline void ++ext2_xattr_delete_inode(struct inode *inode) ++{ ++} ++ ++static inline void ++ext2_xattr_put_super(struct super_block *sb) ++{ ++} ++ ++static inline int ++init_ext2_xattr(void) ++{ ++ return 0; ++} ++ ++static inline void ++exit_ext2_xattr(void) ++{ ++} ++ ++# endif /* CONFIG_EXT2_FS_XATTR */ ++ ++# ifdef CONFIG_EXT2_FS_XATTR_USER ++ ++extern int init_ext2_xattr_user(void) __init; ++extern void exit_ext2_xattr_user(void); ++ ++# else /* CONFIG_EXT2_FS_XATTR_USER */ ++ ++static inline int ++init_ext2_xattr_user(void) ++{ ++ return 0; ++} ++ ++static inline void ++exit_ext2_xattr_user(void) ++{ ++} ++ ++# endif /* CONFIG_EXT2_FS_XATTR_USER */ ++ ++#endif /* __KERNEL__ */ ++ +--- linux-2.4.22-ac1/include/linux/ext3_fs.h~xattr-0.8.54-2.4.22-rh 2003-09-25 14:58:30.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_fs.h 2003-09-25 23:57:02.000000000 +0400 +@@ -63,8 +63,6 @@ + */ + #define EXT3_BAD_INO 1 /* Bad blocks inode */ + #define EXT3_ROOT_INO 2 /* Root inode */ +-#define EXT3_ACL_IDX_INO 3 /* ACL inode */ +-#define EXT3_ACL_DATA_INO 4 /* ACL inode */ + #define EXT3_BOOT_LOADER_INO 5 /* Boot loader inode */ + #define EXT3_UNDEL_DIR_INO 6 /* Undelete directory inode */ + #define EXT3_RESIZE_INO 7 /* Reserved group descriptors inode */ +@@ -94,7 +92,6 @@ + #else + # define EXT3_BLOCK_SIZE(s) (EXT3_MIN_BLOCK_SIZE << (s)->s_log_block_size) + #endif +-#define EXT3_ACLE_PER_BLOCK(s) (EXT3_BLOCK_SIZE(s) / sizeof (struct ext3_acl_entry)) + #define EXT3_ADDR_PER_BLOCK(s) (EXT3_BLOCK_SIZE(s) / sizeof (__u32)) + #ifdef __KERNEL__ + # define EXT3_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits) +@@ -129,28 +126,6 @@ + #endif + + /* +- * ACL structures +- */ +-struct ext3_acl_header /* Header of Access Control Lists */ +-{ +- __u32 aclh_size; +- __u32 aclh_file_count; +- __u32 aclh_acle_count; +- __u32 aclh_first_acle; +-}; +- +-struct ext3_acl_entry /* Access Control List Entry */ +-{ +- __u32 acle_size; +- __u16 acle_perms; /* Access permissions */ +- __u16 acle_type; /* Type of entry */ +- __u16 acle_tag; /* User or group identity */ +- __u16 acle_pad1; +- __u32 acle_next; /* Pointer on next entry for the */ +- /* same inode or on next free entry */ +-}; +- +-/* + * Structure of a blocks group descriptor + */ + struct ext3_group_desc +@@ -346,6 +321,7 @@ struct ext3_inode { + #define EXT3_MOUNT_WRITEBACK_DATA 0x0C00 /* No data ordering */ + #define EXT3_MOUNT_UPDATE_JOURNAL 0x1000 /* Update the journal format */ + #define EXT3_MOUNT_NO_UID32 0x2000 /* Disable 32-bit UIDs */ ++#define EXT3_MOUNT_XATTR_USER 0x4000 /* Extended user attributes */ + + /* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */ + #ifndef _LINUX_EXT2_FS_H +@@ -523,7 +499,7 @@ struct ext3_super_block { + #define EXT3_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */ + #define EXT3_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */ + +-#define EXT3_FEATURE_COMPAT_SUPP 0 ++#define EXT3_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR + #define EXT3_FEATURE_INCOMPAT_SUPP (EXT3_FEATURE_INCOMPAT_FILETYPE| \ + EXT3_FEATURE_INCOMPAT_RECOVER) + #define EXT3_FEATURE_RO_COMPAT_SUPP (EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER| \ +@@ -707,6 +683,7 @@ extern void ext3_check_inodes_bitmap (st + extern unsigned long ext3_count_free (struct buffer_head *, unsigned); + + /* inode.c */ ++extern int ext3_forget(handle_t *, int, struct inode *, struct buffer_head *, int); + extern struct buffer_head * ext3_getblk (handle_t *, struct inode *, long, int, int *); + extern struct buffer_head * ext3_bread (handle_t *, struct inode *, int, int, int *); + +@@ -776,8 +753,10 @@ extern struct address_space_operations e + + /* namei.c */ + extern struct inode_operations ext3_dir_inode_operations; ++extern struct inode_operations ext3_special_inode_operations; + + /* symlink.c */ ++extern struct inode_operations ext3_symlink_inode_operations; + extern struct inode_operations ext3_fast_symlink_inode_operations; + + +--- linux-2.4.22-ac1/include/linux/ext3_jbd.h~xattr-0.8.54-2.4.22-rh 2003-09-25 14:55:12.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_jbd.h 2003-09-25 23:57:02.000000000 +0400 +@@ -30,13 +30,19 @@ + + #define EXT3_SINGLEDATA_TRANS_BLOCKS 8U + ++/* Extended attributes may touch two data buffers, two bitmap buffers, ++ * and two group and summaries. */ ++ ++#define EXT3_XATTR_TRANS_BLOCKS 8 ++ + /* Define the minimum size for a transaction which modifies data. This + * needs to take into account the fact that we may end up modifying two + * quota files too (one for the group, one for the user quota). The + * superblock only gets updated once, of course, so don't bother + * counting that again for the quota updates. */ + +-#define EXT3_DATA_TRANS_BLOCKS (3 * EXT3_SINGLEDATA_TRANS_BLOCKS - 2) ++#define EXT3_DATA_TRANS_BLOCKS (3 * EXT3_SINGLEDATA_TRANS_BLOCKS + \ ++ EXT3_XATTR_TRANS_BLOCKS - 2) + + extern int ext3_writepage_trans_blocks(struct inode *inode); + +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/include/linux/ext3_xattr.h 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,157 @@ ++/* ++ File: linux/ext3_xattr.h ++ ++ On-disk format of extended attributes for the ext3 filesystem. ++ ++ (C) 2001 Andreas Gruenbacher, ++*/ ++ ++#include ++#include ++#include ++ ++/* Magic value in attribute blocks */ ++#define EXT3_XATTR_MAGIC 0xEA020000 ++ ++/* Maximum number of references to one attribute block */ ++#define EXT3_XATTR_REFCOUNT_MAX 1024 ++ ++/* Name indexes */ ++#define EXT3_XATTR_INDEX_MAX 10 ++#define EXT3_XATTR_INDEX_USER 1 ++#define EXT3_XATTR_INDEX_POSIX_ACL_ACCESS 2 ++#define EXT3_XATTR_INDEX_POSIX_ACL_DEFAULT 3 ++ ++struct ext3_xattr_header { ++ __u32 h_magic; /* magic number for identification */ ++ __u32 h_refcount; /* reference count */ ++ __u32 h_blocks; /* number of disk blocks used */ ++ __u32 h_hash; /* hash value of all attributes */ ++ __u32 h_reserved[4]; /* zero right now */ ++}; ++ ++struct ext3_xattr_entry { ++ __u8 e_name_len; /* length of name */ ++ __u8 e_name_index; /* attribute name index */ ++ __u16 e_value_offs; /* offset in disk block of value */ ++ __u32 e_value_block; /* disk block attribute is stored on (n/i) */ ++ __u32 e_value_size; /* size of attribute value */ ++ __u32 e_hash; /* hash value of name and value */ ++ char e_name[0]; /* attribute name */ ++}; ++ ++#define EXT3_XATTR_PAD_BITS 2 ++#define EXT3_XATTR_PAD (1<e_name_len)) ) ++#define EXT3_XATTR_SIZE(size) \ ++ (((size) + EXT3_XATTR_ROUND) & ~EXT3_XATTR_ROUND) ++ ++#ifdef __KERNEL__ ++ ++# ifdef CONFIG_EXT3_FS_XATTR ++ ++struct ext3_xattr_handler { ++ char *prefix; ++ size_t (*list)(char *list, struct inode *inode, const char *name, ++ int name_len); ++ int (*get)(struct inode *inode, const char *name, void *buffer, ++ size_t size); ++ int (*set)(struct inode *inode, const char *name, const void *buffer, ++ size_t size, int flags); ++}; ++ ++extern int ext3_xattr_register(int, struct ext3_xattr_handler *); ++extern void ext3_xattr_unregister(int, struct ext3_xattr_handler *); ++ ++extern int ext3_setxattr(struct dentry *, const char *, const void *, size_t, int); ++extern ssize_t ext3_getxattr(struct dentry *, const char *, void *, size_t); ++extern ssize_t ext3_listxattr(struct dentry *, char *, size_t); ++extern int ext3_removexattr(struct dentry *, const char *); ++ ++extern int ext3_xattr_get(struct inode *, int, const char *, void *, size_t); ++extern int ext3_xattr_list(struct inode *, char *, size_t); ++extern int ext3_xattr_set(handle_t *handle, struct inode *, int, const char *, const void *, size_t, int); ++ ++extern void ext3_xattr_delete_inode(handle_t *, struct inode *); ++extern void ext3_xattr_put_super(struct super_block *); ++ ++extern int init_ext3_xattr(void) __init; ++extern void exit_ext3_xattr(void); ++ ++# else /* CONFIG_EXT3_FS_XATTR */ ++# define ext3_setxattr NULL ++# define ext3_getxattr NULL ++# define ext3_listxattr NULL ++# define ext3_removexattr NULL ++ ++static inline int ++ext3_xattr_get(struct inode *inode, int name_index, const char *name, ++ void *buffer, size_t size) ++{ ++ return -ENOTSUP; ++} ++ ++static inline int ++ext3_xattr_list(struct inode *inode, void *buffer, size_t size) ++{ ++ return -ENOTSUP; ++} ++ ++static inline int ++ext3_xattr_set(handle_t *handle, struct inode *inode, int name_index, ++ const char *name, const void *value, size_t size, int flags) ++{ ++ return -ENOTSUP; ++} ++ ++static inline void ++ext3_xattr_delete_inode(handle_t *handle, struct inode *inode) ++{ ++} ++ ++static inline void ++ext3_xattr_put_super(struct super_block *sb) ++{ ++} ++ ++static inline int ++init_ext3_xattr(void) ++{ ++ return 0; ++} ++ ++static inline void ++exit_ext3_xattr(void) ++{ ++} ++ ++# endif /* CONFIG_EXT3_FS_XATTR */ ++ ++# ifdef CONFIG_EXT3_FS_XATTR_USER ++ ++extern int init_ext3_xattr_user(void) __init; ++extern void exit_ext3_xattr_user(void); ++ ++# else /* CONFIG_EXT3_FS_XATTR_USER */ ++ ++static inline int ++init_ext3_xattr_user(void) ++{ ++ return 0; ++} ++ ++static inline void ++exit_ext3_xattr_user(void) ++{ ++} ++ ++#endif /* CONFIG_EXT3_FS_XATTR_USER */ ++ ++#endif /* __KERNEL__ */ ++ +--- linux-2.4.22-ac1/include/linux/fs.h~xattr-0.8.54-2.4.22-rh 2003-09-25 14:45:32.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/include/linux/fs.h 2003-09-25 23:57:02.000000000 +0400 +@@ -918,7 +918,7 @@ struct inode_operations { + int (*setattr) (struct dentry *, struct iattr *); + int (*setattr_raw) (struct inode *, struct iattr *); + int (*getattr) (struct dentry *, struct iattr *); +- int (*setxattr) (struct dentry *, const char *, void *, size_t, int); ++ int (*setxattr) (struct dentry *, const char *, const void *, size_t, int); + ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t); + ssize_t (*listxattr) (struct dentry *, char *, size_t); + int (*removexattr) (struct dentry *, const char *); +--- /dev/null 2003-01-30 13:24:37.000000000 +0300 ++++ linux-2.4.22-ac1-alexey/include/linux/mbcache.h 2003-09-25 23:57:02.000000000 +0400 +@@ -0,0 +1,69 @@ ++/* ++ File: linux/mbcache.h ++ ++ (C) 2001 by Andreas Gruenbacher, ++*/ ++ ++/* Hardwire the number of additional indexes */ ++#define MB_CACHE_INDEXES_COUNT 1 ++ ++struct mb_cache_entry; ++ ++struct mb_cache_op { ++ int (*free)(struct mb_cache_entry *, int); ++}; ++ ++struct mb_cache { ++ struct list_head c_cache_list; ++ const char *c_name; ++ struct mb_cache_op c_op; ++ atomic_t c_entry_count; ++ int c_bucket_count; ++#ifndef MB_CACHE_INDEXES_COUNT ++ int c_indexes_count; ++#endif ++ kmem_cache_t *c_entry_cache; ++ struct list_head *c_block_hash; ++ struct list_head *c_indexes_hash[0]; ++}; ++ ++struct mb_cache_entry_index { ++ struct list_head o_list; ++ unsigned int o_key; ++}; ++ ++struct mb_cache_entry { ++ struct list_head e_lru_list; ++ struct mb_cache *e_cache; ++ atomic_t e_used; ++ kdev_t e_dev; ++ unsigned long e_block; ++ struct list_head e_block_list; ++ struct mb_cache_entry_index e_indexes[0]; ++}; ++ ++/* Functions on caches */ ++ ++struct mb_cache * mb_cache_create(const char *, struct mb_cache_op *, size_t, ++ int, int); ++void mb_cache_shrink(struct mb_cache *, kdev_t); ++void mb_cache_destroy(struct mb_cache *); ++ ++/* Functions on cache entries */ ++ ++struct mb_cache_entry *mb_cache_entry_alloc(struct mb_cache *); ++int mb_cache_entry_insert(struct mb_cache_entry *, kdev_t, unsigned long, ++ unsigned int[]); ++void mb_cache_entry_rehash(struct mb_cache_entry *, unsigned int[]); ++void mb_cache_entry_release(struct mb_cache_entry *); ++void mb_cache_entry_takeout(struct mb_cache_entry *); ++void mb_cache_entry_free(struct mb_cache_entry *); ++struct mb_cache_entry *mb_cache_entry_dup(struct mb_cache_entry *); ++struct mb_cache_entry *mb_cache_entry_get(struct mb_cache *, kdev_t, ++ unsigned long); ++#if !defined(MB_CACHE_INDEXES_COUNT) || (MB_CACHE_INDEXES_COUNT > 0) ++struct mb_cache_entry *mb_cache_entry_find_first(struct mb_cache *cache, int, ++ kdev_t, unsigned int); ++struct mb_cache_entry *mb_cache_entry_find_next(struct mb_cache_entry *, int, ++ kdev_t, unsigned int); ++#endif +--- linux-2.4.22-ac1/kernel/ksyms.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:42:46.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/kernel/ksyms.c 2003-09-26 00:19:05.000000000 +0400 +@@ -11,6 +11,7 @@ + + #include + #include ++#include + #include + #include + #include +@@ -106,6 +107,7 @@ EXPORT_SYMBOL(exit_files); + EXPORT_SYMBOL(exit_fs); + EXPORT_SYMBOL(exit_sighand); + EXPORT_SYMBOL(unshare_files); ++EXPORT_SYMBOL(copy_fs_struct); + + /* internal kernel memory management */ + EXPORT_SYMBOL(_alloc_pages); +@@ -124,6 +126,8 @@ EXPORT_SYMBOL(kmem_cache_validate); + EXPORT_SYMBOL(kmem_cache_alloc); + EXPORT_SYMBOL(kmem_cache_free); + EXPORT_SYMBOL(kmem_cache_size); ++EXPORT_SYMBOL(register_cache); ++EXPORT_SYMBOL(unregister_cache); + EXPORT_SYMBOL(kmalloc); + EXPORT_SYMBOL(kfree); + EXPORT_SYMBOL(vfree); +--- linux-2.4.22-ac1/mm/vmscan.c~xattr-0.8.54-2.4.22-rh 2003-09-25 14:16:28.000000000 +0400 ++++ linux-2.4.22-ac1-alexey/mm/vmscan.c 2003-09-25 23:57:02.000000000 +0400 +@@ -18,6 +18,7 @@ + #include + #include + #include ++#include + #include + #include + #include +@@ -34,6 +35,39 @@ + */ + #define DEF_PRIORITY (6) + ++static DECLARE_MUTEX(other_caches_sem); ++static LIST_HEAD(cache_definitions); ++ ++void register_cache(struct cache_definition *cache) ++{ ++ down(&other_caches_sem); ++ list_add(&cache->link, &cache_definitions); ++ up(&other_caches_sem); ++} ++ ++void unregister_cache(struct cache_definition *cache) ++{ ++ down(&other_caches_sem); ++ list_del(&cache->link); ++ up(&other_caches_sem); ++} ++ ++static void shrink_other_caches(unsigned int priority, int gfp_mask) ++{ ++ struct list_head *p; ++ ++ if (down_trylock(&other_caches_sem)) ++ return; ++ ++ list_for_each_prev(p, &cache_definitions) { ++ struct cache_definition *cache = ++ list_entry(p, struct cache_definition, link); ++ ++ cache->shrink(priority, gfp_mask); ++ } ++ up(&other_caches_sem); ++} ++ + /* + * The swap-out function returns 1 if it successfully + * scanned all the pages it was asked to (`count'). +@@ -577,6 +611,7 @@ static int shrink_caches(zone_t * classz + + shrink_dcache_memory(priority, gfp_mask); + shrink_icache_memory(priority, gfp_mask); ++ shrink_other_caches(priority, gfp_mask); + #ifdef CONFIG_QUOTA + shrink_dqcache_memory(DEF_PRIORITY, gfp_mask); + #endif + +_ diff --git a/lustre/kernel_patches/pc/ext3-htree-2.4.22-rh.pc b/lustre/kernel_patches/pc/ext3-htree-2.4.22-rh.pc new file mode 100644 index 0000000..9734b6d --- /dev/null +++ b/lustre/kernel_patches/pc/ext3-htree-2.4.22-rh.pc @@ -0,0 +1,11 @@ +fs/ext3/dir.c +fs/ext3/file.c +fs/ext3/hash.c +fs/ext3/Makefile +fs/ext3/namei.c +fs/ext3/super.c +include/linux/ext3_fs.h +include/linux/ext3_fs_sb.h +include/linux/ext3_jbd.h +include/linux/rbtree.h +lib/rbtree.c diff --git a/lustre/kernel_patches/pc/ext3-orphan_lock-2.4.22-rh.pc b/lustre/kernel_patches/pc/ext3-orphan_lock-2.4.22-rh.pc new file mode 100644 index 0000000..1b6b9ea --- /dev/null +++ b/lustre/kernel_patches/pc/ext3-orphan_lock-2.4.22-rh.pc @@ -0,0 +1,5 @@ +fs/ext3/namei.c +fs/ext3/namei.c.orig +fs/ext3/super.c +include/linux/ext3_fs_sb.h +include/linux/ext3_fs_sb.h.orig diff --git a/lustre/kernel_patches/pc/extN-wantedi-2.4.22-rh.pc b/lustre/kernel_patches/pc/extN-wantedi-2.4.22-rh.pc new file mode 100644 index 0000000..a9b845a --- /dev/null +++ b/lustre/kernel_patches/pc/extN-wantedi-2.4.22-rh.pc @@ -0,0 +1,6 @@ +fs/ext3/ialloc.c +fs/ext3/inode.c +fs/ext3/ioctl.c +fs/ext3/namei.c +include/linux/dcache.h +include/linux/ext3_fs.h diff --git a/lustre/kernel_patches/pc/iod-stock-exports-2.4.22-rh.pc b/lustre/kernel_patches/pc/iod-stock-exports-2.4.22-rh.pc new file mode 100644 index 0000000..e4eceee --- /dev/null +++ b/lustre/kernel_patches/pc/iod-stock-exports-2.4.22-rh.pc @@ -0,0 +1,3 @@ +fs/inode.c +fs/Makefile +mm/page_alloc.c diff --git a/lustre/kernel_patches/pc/socket-exports-2.4.22-rh.pc b/lustre/kernel_patches/pc/socket-exports-2.4.22-rh.pc new file mode 100644 index 0000000..3277003 --- /dev/null +++ b/lustre/kernel_patches/pc/socket-exports-2.4.22-rh.pc @@ -0,0 +1,3 @@ +include/linux/socket.h +net/netsyms.c +net/socket.c diff --git a/lustre/kernel_patches/pc/tcp-zero-copy-2.4.22-rh.pc b/lustre/kernel_patches/pc/tcp-zero-copy-2.4.22-rh.pc new file mode 100644 index 0000000..556327b --- /dev/null +++ b/lustre/kernel_patches/pc/tcp-zero-copy-2.4.22-rh.pc @@ -0,0 +1,5 @@ +include/linux/skbuff.h +include/net/tcp.h +net/core/skbuff.c +net/ipv4/tcp.c +net/netsyms.c diff --git a/lustre/kernel_patches/pc/vfs_intent-2.4.22-rh.pc b/lustre/kernel_patches/pc/vfs_intent-2.4.22-rh.pc new file mode 100644 index 0000000..d177a30 --- /dev/null +++ b/lustre/kernel_patches/pc/vfs_intent-2.4.22-rh.pc @@ -0,0 +1,12 @@ +fs/dcache.c +fs/exec.c +fs/namei.c +fs/namespace.c +fs/open.c +fs/stat.c +include/linux/dcache.h +include/linux/fs.h +include/linux/fs_struct.h +kernel/exit.c +kernel/fork.c +kernel/ksyms.c diff --git a/lustre/kernel_patches/pc/xattr-0.8.54-2.4.22-rh.pc b/lustre/kernel_patches/pc/xattr-0.8.54-2.4.22-rh.pc new file mode 100644 index 0000000..e6edcc6 --- /dev/null +++ b/lustre/kernel_patches/pc/xattr-0.8.54-2.4.22-rh.pc @@ -0,0 +1,58 @@ +arch/alpha/defconfig +arch/alpha/kernel/entry.S +arch/arm/defconfig +arch/arm/kernel/calls.S +arch/i386/defconfig +arch/ia64/defconfig +arch/m68k/defconfig +arch/mips64/defconfig +arch/mips/defconfig +arch/ppc64/kernel/misc.S +arch/ppc/defconfig +arch/s390/defconfig +arch/s390/kernel/entry.S +arch/s390x/defconfig +arch/s390x/kernel/entry.S +arch/s390x/kernel/wrapper32.S +arch/sparc64/defconfig +arch/sparc/defconfig +Documentation/Configure.help +fs/Config.in +fs/ext2/file.c +fs/ext2/ialloc.c +fs/ext2/inode.c +fs/ext2/Makefile +fs/ext2/namei.c +fs/ext2/super.c +fs/ext2/symlink.c +fs/ext2/xattr.c +fs/ext2/xattr_user.c +fs/ext3/ext3-exports.c +fs/ext3/file.c +fs/ext3/ialloc.c +fs/ext3/inode.c +fs/ext3/Makefile +fs/ext3/namei.c +fs/ext3/super.c +fs/ext3/symlink.c +fs/ext3/xattr.c +fs/ext3/xattr_user.c +fs/jfs/jfs_xattr.h +fs/jfs/xattr.c +fs/Makefile +fs/mbcache.c +include/asm-arm/unistd.h +include/asm-ppc64/unistd.h +include/asm-s390/unistd.h +include/asm-s390x/unistd.h +include/linux/cache_def.h +include/linux/errno.h +include/linux/ext2_fs.h +include/linux/ext2_xattr.h +include/linux/ext3_fs.h +include/linux/ext3_jbd.h +include/linux/ext3_xattr.h +include/linux/fs.h +include/linux/mbcache.h +kernel/ksyms.c +mm/vmscan.c diff --git a/lustre/kernel_patches/series/rh-2.4.22 b/lustre/kernel_patches/series/rh-2.4.22 new file mode 100644 index 0000000..0c364b5 --- /dev/null +++ b/lustre/kernel_patches/series/rh-2.4.22 @@ -0,0 +1,23 @@ +dev_read_only_2.4.20-rh.patch +exports_2.4.20-rh-hp.patch +kmem_cache_validate_2.4.20.patch +lustre_version.patch +vfs_intent-2.4.22-rh.patch +invalidate_show-2.4.20-rh.patch +export-truncate.patch +iod-stock-exports-2.4.22-rh.patch +ext3-htree-2.4.22-rh.patch +xattr-0.8.54-2.4.22-rh.patch +ext3-orphan_lock-2.4.22-rh.patch +ext3-noread-2.4.20.patch +ext3_delete_thread_2.4.20_chaos.patch +extN-wantedi-2.4.22-rh.patch +ext3-san-2.4.20.patch +ext3-map_inode_page.patch +ext3-error-export.patch +iopen-2.4.20.patch +tcp-zero-copy-2.4.22-rh.patch +jbd-dont-account-blocks-twice.patch +jbd-commit-tricks.patch +add_page_private.patch +socket-exports-2.4.22-rh.patch -- 1.8.3.1