/usr/src/kernel-patches/lustre/patches/quota-deadlock-on-pagelock-ext3.patch is in linux-patch-lustre 1.8.5+dfsg-3ubuntu1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 | Index: linux-2.6.9/fs/ext3/inode.c
===================================================================
--- linux-2.6.9.orig/fs/ext3/inode.c 2006-08-25 16:39:10.000000000 +0800
+++ linux-2.6.9/fs/ext3/inode.c 2006-09-14 11:44:29.000000000 +0800
@@ -1028,7 +1028,7 @@
return ret;
}
-static int
+int
ext3_journal_dirty_data(handle_t *handle, struct buffer_head *bh)
{
int err = journal_dirty_data(handle, bh);
Index: linux-2.6.9/fs/ext3/super.c
===================================================================
--- linux-2.6.9.orig/fs/ext3/super.c 2006-08-25 16:39:48.000000000 +0800
+++ linux-2.6.9/fs/ext3/super.c 2006-09-14 11:51:48.000000000 +0800
@@ -529,7 +529,10 @@
static int ext3_write_info(struct super_block *sb, int type);
static int ext3_quota_on(struct super_block *sb, int type, int format_id, char *path);
static int ext3_quota_on_mount(struct super_block *sb, int type);
-static int ext3_quota_off_mount(struct super_block *sb, int type);
+static ssize_t ext3_quota_read(struct super_block *sb, int type, char *data,
+ size_t len, loff_t off);
+static ssize_t ext3_quota_write(struct super_block *sb, int type,
+ const char *data, size_t len, loff_t off);
static struct dquot_operations ext3_quota_operations = {
.initialize = ext3_dquot_initialize,
@@ -572,6 +575,10 @@
.statfs = ext3_statfs,
.remount_fs = ext3_remount,
.clear_inode = ext3_clear_inode,
+#ifdef CONFIG_QUOTA
+ .quota_read = ext3_quota_read,
+ .quota_write = ext3_quota_write,
+#endif
};
static struct dentry *ext3_get_dentry(struct super_block *sb, void *vobjp)
@@ -713,6 +720,7 @@
int option;
#ifdef CONFIG_QUOTA
int qtype;
+ char *qname;
#endif
if (!options)
@@ -891,19 +899,22 @@
"quota options when quota turned on.\n");
return 0;
}
- if (sbi->s_qf_names[qtype]) {
+ qname = match_strdup(&args[0]);
+ if (!qname) {
printk(KERN_ERR
- "EXT3-fs: %s quota file already "
- "specified.\n", QTYPE2NAME(qtype));
+ "EXT3-fs: not enough memory for "
+ "storing quotafile name.\n");
return 0;
}
- sbi->s_qf_names[qtype] = match_strdup(&args[0]);
- if (!sbi->s_qf_names[qtype]) {
+ if (sbi->s_qf_names[qtype] &&
+ strcmp(sbi->s_qf_names[qtype], qname)) {
printk(KERN_ERR
- "EXT3-fs: not enough memory for "
- "storing quotafile name.\n");
+ "EXT3-fs: %s quota file already "
+ "specified.\n", QTYPE2NAME(qtype));
+ kfree(qname);
return 0;
}
+ sbi->s_qf_names[qtype] = qname;
if (strchr(sbi->s_qf_names[qtype], '/')) {
printk(KERN_ERR
"EXT3-fs: quotafile must be on "
@@ -1223,7 +1234,7 @@
/* Turn quotas off */
for (i = 0; i < MAXQUOTAS; i++) {
if (sb_dqopt(sb)->files[i])
- ext3_quota_off_mount(sb, i);
+ vfs_quota_off(sb, i);
}
#endif
sb->s_flags = s_flags; /* Restore MS_RDONLY status */
@@ -2240,7 +2251,7 @@
static inline struct inode *dquot_to_inode(struct dquot *dquot)
{
- return sb_dqopt(dquot->dq_sb)->files[dquot->dq_type]->f_dentry->d_inode;
+ return sb_dqopt(dquot->dq_sb)->files[dquot->dq_type];
}
static int ext3_dquot_initialize(struct inode *inode, int type)
@@ -2279,8 +2290,10 @@
{
int ret, err;
handle_t *handle;
+ struct inode *inode;
- handle = ext3_journal_start(dquot_to_inode(dquot),
+ inode = dquot_to_inode(dquot);
+ handle = ext3_journal_start(inode,
EXT3_QUOTA_TRANS_BLOCKS);
if (IS_ERR(handle))
return PTR_ERR(handle);
@@ -2367,22 +2380,9 @@
if (IS_ERR(dentry))
return PTR_ERR(dentry);
err = vfs_quota_on_mount(type, EXT3_SB(sb)->s_jquota_fmt, dentry);
- if (err)
- dput(dentry);
- /* We keep the dentry reference if everything went ok - we drop it
- * on quota_off time */
- return err;
-}
-
-/* Turn quotas off during mount time */
-static int ext3_quota_off_mount(struct super_block *sb, int type)
-{
- int err;
- struct dentry *dentry;
-
- dentry = sb_dqopt(sb)->files[type]->f_dentry;
- err = vfs_quota_off_mount(sb, type);
- /* We invalidate dentry - it has at least wrong hash... */
+ /* Now invalidate and put the dentry - quota got its own reference
+ * to inode and dentry has at least wrong hash so we had better
+ * throw it away */
d_invalidate(dentry);
dput(dentry);
return err;
@@ -2405,20 +2405,121 @@
if (err)
return err;
/* Quotafile not on the same filesystem? */
- if (nd.mnt->mnt_sb != sb)
+ if (nd.mnt->mnt_sb != sb) {
+ path_release(&nd);
return -EXDEV;
+ }
/* Quotafile not of fs root? */
if (nd.dentry->d_parent->d_inode != sb->s_root->d_inode)
printk(KERN_WARNING
"EXT3-fs: Quota file not on filesystem root. "
"Journalled quota will not work.\n");
- if (!ext3_should_journal_data(nd.dentry->d_inode))
- printk(KERN_WARNING "EXT3-fs: Quota file does not have "
- "data-journalling. Journalled quota will not work.\n");
path_release(&nd);
return vfs_quota_on(sb, type, format_id, path);
}
+/* Read data from quotafile - avoid pagecache and such because we cannot afford
+ * acquiring the locks... As quota files are never truncated and quota code
+ * itself serializes the operations (and noone else should touch the files)
+ * we don't have to be afraid of races */
+static ssize_t ext3_quota_read(struct super_block *sb, int type, char *data,
+ size_t len, loff_t off)
+{
+ struct inode *inode = sb_dqopt(sb)->files[type];
+ sector_t blk = off >> EXT3_BLOCK_SIZE_BITS(sb);
+ int err = 0;
+ int offset = off & (sb->s_blocksize - 1);
+ int tocopy;
+ size_t toread;
+ struct buffer_head *bh;
+ loff_t i_size = i_size_read(inode);
+
+ if (off > i_size)
+ return 0;
+ if (off+len > i_size)
+ len = i_size-off;
+ toread = len;
+ while (toread > 0) {
+ tocopy = sb->s_blocksize - offset < toread ?
+ sb->s_blocksize - offset : toread;
+ bh = ext3_bread(NULL, inode, blk, 0, &err);
+ if (err)
+ return err;
+ if (!bh) /* A hole? */
+ memset(data, 0, tocopy);
+ else
+ memcpy(data, bh->b_data+offset, tocopy);
+ brelse(bh);
+ offset = 0;
+ toread -= tocopy;
+ data += tocopy;
+ blk++;
+ }
+ return len;
+}
+
+/* Write to quotafile (we know the transaction is already started and has
+ * enough credits) */
+static ssize_t ext3_quota_write(struct super_block *sb, int type,
+ const char *data, size_t len, loff_t off)
+{
+ struct inode *inode = sb_dqopt(sb)->files[type];
+ sector_t blk = off >> EXT3_BLOCK_SIZE_BITS(sb);
+ int err = 0;
+ int offset = off & (sb->s_blocksize - 1);
+ int tocopy;
+ int journal_quota = EXT3_SB(sb)->s_qf_names[type] != NULL;
+ size_t towrite = len;
+ struct buffer_head *bh;
+ handle_t *handle = journal_current_handle();
+
+ down(&inode->i_sem);
+ while (towrite > 0) {
+ tocopy = sb->s_blocksize - offset < towrite ?
+ sb->s_blocksize - offset : towrite;
+ bh = ext3_bread(handle, inode, blk, 1, &err);
+ if (!bh)
+ goto out;
+ if (journal_quota) {
+ err = ext3_journal_get_write_access(handle, bh);
+ if (err) {
+ brelse(bh);
+ goto out;
+ }
+ }
+ lock_buffer(bh);
+ memcpy(bh->b_data+offset, data, tocopy);
+ flush_dcache_page(bh->b_page);
+ unlock_buffer(bh);
+ if (journal_quota)
+ err = ext3_journal_dirty_metadata(handle, bh);
+ else {
+ /* Always do at least ordered writes for quotas */
+ err = ext3_journal_dirty_data(handle, bh);
+ mark_buffer_dirty(bh);
+ }
+ brelse(bh);
+ if (err)
+ goto out;
+ offset = 0;
+ towrite -= tocopy;
+ data += tocopy;
+ blk++;
+ }
+out:
+ if (len == towrite)
+ return err;
+ if (inode->i_size < off+len-towrite) {
+ i_size_write(inode, off+len-towrite);
+ EXT3_I(inode)->i_disksize = inode->i_size;
+ }
+ inode->i_version++;
+ inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+ ext3_mark_inode_dirty(handle, inode);
+ up(&inode->i_sem);
+ return len - towrite;
+}
+
#endif
static struct super_block *ext3_get_sb(struct file_system_type *fs_type,
Index: linux-2.6.9/include/linux/ext3_jbd.h
===================================================================
--- linux-2.6.9.orig/include/linux/ext3_jbd.h 2006-08-25 16:39:09.000000000 +0800
+++ linux-2.6.9/include/linux/ext3_jbd.h 2006-09-14 11:44:29.000000000 +0800
@@ -193,6 +193,8 @@
#define ext3_journal_forget(handle, bh) \
__ext3_journal_forget(__FUNCTION__, (handle), (bh))
+int ext3_journal_dirty_data(handle_t *handle, struct buffer_head *bh);
+
handle_t *ext3_journal_start_sb(struct super_block *sb, int nblocks);
int __ext3_journal_stop(const char *where, handle_t *handle);
|