From c475146d8f3b97e79f9ef88521e28ad40ac07de6 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 12 Jan 2009 17:44:34 +0100 Subject: ocfs2: Remove ocfs2_dquot_initialize() and ocfs2_dquot_drop() Since ->acquire_dquot and ->release_dquot callbacks aren't called under dqptr_sem anymore, we don't have to start a transaction and obtain locks so early. So we can just remove all this complicated stuff. Signed-off-by: Jan Kara Acked-by: Mark Fasheh --- fs/ocfs2/quota_global.c | 169 +----------------------------------------------- 1 file changed, 2 insertions(+), 167 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index 6aff8f2d3e49..f4efa89baee5 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -810,171 +810,6 @@ out: return status; } -/* This is difficult. We have to lock quota inode and start transaction - * in this function but we don't want to take the penalty of exlusive - * quota file lock when we are just going to use cached structures. So - * we just take read lock check whether we have dquot cached and if so, - * we don't have to take the write lock... */ -static int ocfs2_dquot_initialize(struct inode *inode, int type) -{ - handle_t *handle = NULL; - int status = 0; - struct super_block *sb = inode->i_sb; - struct ocfs2_mem_dqinfo *oinfo; - int exclusive = 0; - int cnt; - qid_t id; - - mlog_entry_void(); - - for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (type != -1 && cnt != type) - continue; - if (!sb_has_quota_active(sb, cnt)) - continue; - oinfo = sb_dqinfo(sb, cnt)->dqi_priv; - status = ocfs2_lock_global_qf(oinfo, 0); - if (status < 0) - goto out; - /* This is just a performance optimization not a reliable test. - * Since we hold an inode lock, noone can actually release - * the structure until we are finished with initialization. */ - if (inode->i_dquot[cnt] != NODQUOT) { - ocfs2_unlock_global_qf(oinfo, 0); - continue; - } - /* When we have inode lock, we know that no dquot_release() can - * run and thus we can safely check whether we need to - * read+modify global file to get quota information or whether - * our node already has it. */ - if (cnt == USRQUOTA) - id = inode->i_uid; - else if (cnt == GRPQUOTA) - id = inode->i_gid; - else - BUG(); - /* Obtain exclusion from quota off... */ - down_write(&sb_dqopt(sb)->dqptr_sem); - exclusive = !dquot_is_cached(sb, id, cnt); - up_write(&sb_dqopt(sb)->dqptr_sem); - if (exclusive) { - status = ocfs2_lock_global_qf(oinfo, 1); - if (status < 0) { - exclusive = 0; - mlog_errno(status); - goto out_ilock; - } - handle = ocfs2_start_trans(OCFS2_SB(sb), - ocfs2_calc_qinit_credits(sb, cnt)); - if (IS_ERR(handle)) { - status = PTR_ERR(handle); - mlog_errno(status); - goto out_ilock; - } - } - dquot_initialize(inode, cnt); - if (exclusive) { - ocfs2_commit_trans(OCFS2_SB(sb), handle); - ocfs2_unlock_global_qf(oinfo, 1); - } - ocfs2_unlock_global_qf(oinfo, 0); - } - mlog_exit(0); - return 0; -out_ilock: - if (exclusive) - ocfs2_unlock_global_qf(oinfo, 1); - ocfs2_unlock_global_qf(oinfo, 0); -out: - mlog_exit(status); - return status; -} - -static int ocfs2_dquot_drop_slow(struct inode *inode) -{ - int status = 0; - int cnt; - int got_lock[MAXQUOTAS] = {0, 0}; - handle_t *handle; - struct super_block *sb = inode->i_sb; - struct ocfs2_mem_dqinfo *oinfo; - - for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (!sb_has_quota_active(sb, cnt)) - continue; - oinfo = sb_dqinfo(sb, cnt)->dqi_priv; - status = ocfs2_lock_global_qf(oinfo, 1); - if (status < 0) - goto out; - got_lock[cnt] = 1; - } - handle = ocfs2_start_trans(OCFS2_SB(sb), - ocfs2_calc_qinit_credits(sb, USRQUOTA) + - ocfs2_calc_qinit_credits(sb, GRPQUOTA)); - if (IS_ERR(handle)) { - status = PTR_ERR(handle); - mlog_errno(status); - goto out; - } - dquot_drop(inode); - ocfs2_commit_trans(OCFS2_SB(sb), handle); -out: - for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if (got_lock[cnt]) { - oinfo = sb_dqinfo(sb, cnt)->dqi_priv; - ocfs2_unlock_global_qf(oinfo, 1); - } - return status; -} - -/* See the comment before ocfs2_dquot_initialize. */ -static int ocfs2_dquot_drop(struct inode *inode) -{ - int status = 0; - struct super_block *sb = inode->i_sb; - struct ocfs2_mem_dqinfo *oinfo; - int exclusive = 0; - int cnt; - int got_lock[MAXQUOTAS] = {0, 0}; - - mlog_entry_void(); - for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (!sb_has_quota_active(sb, cnt)) - continue; - oinfo = sb_dqinfo(sb, cnt)->dqi_priv; - status = ocfs2_lock_global_qf(oinfo, 0); - if (status < 0) - goto out; - got_lock[cnt] = 1; - } - /* Lock against anyone releasing references so that when when we check - * we know we are not going to be last ones to release dquot */ - down_write(&sb_dqopt(sb)->dqptr_sem); - /* Urgh, this is a terrible hack :( */ - for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] != NODQUOT && - atomic_read(&inode->i_dquot[cnt]->dq_count) > 1) { - exclusive = 1; - break; - } - } - if (!exclusive) - dquot_drop_locked(inode); - up_write(&sb_dqopt(sb)->dqptr_sem); -out: - for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if (got_lock[cnt]) { - oinfo = sb_dqinfo(sb, cnt)->dqi_priv; - ocfs2_unlock_global_qf(oinfo, 0); - } - /* In case we bailed out because we had to do expensive locking - * do it now... */ - if (exclusive) - status = ocfs2_dquot_drop_slow(inode); - mlog_exit(status); - return status; -} - static struct dquot *ocfs2_alloc_dquot(struct super_block *sb, int type) { struct ocfs2_dquot *dquot = @@ -991,8 +826,8 @@ static void ocfs2_destroy_dquot(struct dquot *dquot) } struct dquot_operations ocfs2_quota_operations = { - .initialize = ocfs2_dquot_initialize, - .drop = ocfs2_dquot_drop, + .initialize = dquot_initialize, + .drop = dquot_drop, .alloc_space = dquot_alloc_space, .alloc_inode = dquot_alloc_inode, .free_space = dquot_free_space, -- cgit v1.2.3 From 2fe4371dff3f1a5a1f7d91f1b090076954f4d17e Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Thu, 22 Jan 2009 10:26:11 +0300 Subject: fs/Kconfig: move ocfs2 out Signed-off-by: Alexey Dobriyan --- fs/Kconfig | 87 +------------------------------------------------------- fs/ocfs2/Kconfig | 85 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 86 insertions(+), 86 deletions(-) create mode 100644 fs/ocfs2/Kconfig (limited to 'fs/ocfs2') diff --git a/fs/Kconfig b/fs/Kconfig index b39675cc0fc0..9fbc43f973d4 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -49,92 +49,7 @@ config FILE_LOCKING source "fs/xfs/Kconfig" source "fs/gfs2/Kconfig" - -config OCFS2_FS - tristate "OCFS2 file system support" - depends on NET && SYSFS - select CONFIGFS_FS - select JBD2 - select CRC32 - select QUOTA - select QUOTA_TREE - help - OCFS2 is a general purpose extent based shared disk cluster file - system with many similarities to ext3. It supports 64 bit inode - numbers, and has automatically extending metadata groups which may - also make it attractive for non-clustered use. - - You'll want to install the ocfs2-tools package in order to at least - get "mount.ocfs2". - - Project web page: http://oss.oracle.com/projects/ocfs2 - Tools web page: http://oss.oracle.com/projects/ocfs2-tools - OCFS2 mailing lists: http://oss.oracle.com/projects/ocfs2/mailman/ - - For more information on OCFS2, see the file - . - -config OCFS2_FS_O2CB - tristate "O2CB Kernelspace Clustering" - depends on OCFS2_FS - default y - help - OCFS2 includes a simple kernelspace clustering package, the OCFS2 - Cluster Base. It only requires a very small userspace component - to configure it. This comes with the standard ocfs2-tools package. - O2CB is limited to maintaining a cluster for OCFS2 file systems. - It cannot manage any other cluster applications. - - It is always safe to say Y here, as the clustering method is - run-time selectable. - -config OCFS2_FS_USERSPACE_CLUSTER - tristate "OCFS2 Userspace Clustering" - depends on OCFS2_FS && DLM - default y - help - This option will allow OCFS2 to use userspace clustering services - in conjunction with the DLM in fs/dlm. If you are using a - userspace cluster manager, say Y here. - - It is safe to say Y, as the clustering method is run-time - selectable. - -config OCFS2_FS_STATS - bool "OCFS2 statistics" - depends on OCFS2_FS - default y - help - This option allows some fs statistics to be captured. Enabling - this option may increase the memory consumption. - -config OCFS2_DEBUG_MASKLOG - bool "OCFS2 logging support" - depends on OCFS2_FS - default y - help - The ocfs2 filesystem has an extensive logging system. The system - allows selection of events to log via files in /sys/o2cb/logmask/. - This option will enlarge your kernel, but it allows debugging of - ocfs2 filesystem issues. - -config OCFS2_DEBUG_FS - bool "OCFS2 expensive checks" - depends on OCFS2_FS - default n - help - This option will enable expensive consistency checks. Enable - this option for debugging only as it is likely to decrease - performance of the filesystem. - -config OCFS2_FS_POSIX_ACL - bool "OCFS2 POSIX Access Control Lists" - depends on OCFS2_FS - select FS_POSIX_ACL - default n - help - Posix Access Control Lists (ACLs) support permissions for users and - groups beyond the owner/group/world scheme. +source "fs/ocfs2/Kconfig" config BTRFS_FS tristate "Btrfs filesystem (EXPERIMENTAL) Unstable disk format" diff --git a/fs/ocfs2/Kconfig b/fs/ocfs2/Kconfig new file mode 100644 index 000000000000..701b7a3a872e --- /dev/null +++ b/fs/ocfs2/Kconfig @@ -0,0 +1,85 @@ +config OCFS2_FS + tristate "OCFS2 file system support" + depends on NET && SYSFS + select CONFIGFS_FS + select JBD2 + select CRC32 + select QUOTA + select QUOTA_TREE + help + OCFS2 is a general purpose extent based shared disk cluster file + system with many similarities to ext3. It supports 64 bit inode + numbers, and has automatically extending metadata groups which may + also make it attractive for non-clustered use. + + You'll want to install the ocfs2-tools package in order to at least + get "mount.ocfs2". + + Project web page: http://oss.oracle.com/projects/ocfs2 + Tools web page: http://oss.oracle.com/projects/ocfs2-tools + OCFS2 mailing lists: http://oss.oracle.com/projects/ocfs2/mailman/ + + For more information on OCFS2, see the file + . + +config OCFS2_FS_O2CB + tristate "O2CB Kernelspace Clustering" + depends on OCFS2_FS + default y + help + OCFS2 includes a simple kernelspace clustering package, the OCFS2 + Cluster Base. It only requires a very small userspace component + to configure it. This comes with the standard ocfs2-tools package. + O2CB is limited to maintaining a cluster for OCFS2 file systems. + It cannot manage any other cluster applications. + + It is always safe to say Y here, as the clustering method is + run-time selectable. + +config OCFS2_FS_USERSPACE_CLUSTER + tristate "OCFS2 Userspace Clustering" + depends on OCFS2_FS && DLM + default y + help + This option will allow OCFS2 to use userspace clustering services + in conjunction with the DLM in fs/dlm. If you are using a + userspace cluster manager, say Y here. + + It is safe to say Y, as the clustering method is run-time + selectable. + +config OCFS2_FS_STATS + bool "OCFS2 statistics" + depends on OCFS2_FS + default y + help + This option allows some fs statistics to be captured. Enabling + this option may increase the memory consumption. + +config OCFS2_DEBUG_MASKLOG + bool "OCFS2 logging support" + depends on OCFS2_FS + default y + help + The ocfs2 filesystem has an extensive logging system. The system + allows selection of events to log via files in /sys/o2cb/logmask/. + This option will enlarge your kernel, but it allows debugging of + ocfs2 filesystem issues. + +config OCFS2_DEBUG_FS + bool "OCFS2 expensive checks" + depends on OCFS2_FS + default n + help + This option will enable expensive consistency checks. Enable + this option for debugging only as it is likely to decrease + performance of the filesystem. + +config OCFS2_FS_POSIX_ACL + bool "OCFS2 POSIX Access Control Lists" + depends on OCFS2_FS + select FS_POSIX_ACL + default n + help + Posix Access Control Lists (ACLs) support permissions for users and + groups beyond the owner/group/world scheme. -- cgit v1.2.3 From ea455f8ab68338ba69f5d3362b342c115bea8e13 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 12 Jan 2009 23:20:31 +0100 Subject: ocfs2: Push out dropping of dentry lock to ocfs2_wq Dropping of last reference to dentry lock is a complicated operation involving dropping of reference to inode. This can get complicated and quota code in particular needs to obtain some quota locks which leads to potential deadlock. Thus we defer dropping of inode reference to ocfs2_wq. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/ocfs2/dcache.c | 42 +++++++++++++++++++++++++++++++++++++++--- fs/ocfs2/dcache.h | 9 ++++++++- fs/ocfs2/ocfs2.h | 6 ++++++ fs/ocfs2/super.c | 3 +++ 4 files changed, 56 insertions(+), 4 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c index b1cc7c381e88..e9d7c2038c0f 100644 --- a/fs/ocfs2/dcache.c +++ b/fs/ocfs2/dcache.c @@ -38,6 +38,7 @@ #include "dlmglue.h" #include "file.h" #include "inode.h" +#include "super.h" static int ocfs2_dentry_revalidate(struct dentry *dentry, @@ -294,6 +295,34 @@ out_attach: return ret; } +static DEFINE_SPINLOCK(dentry_list_lock); + +/* We limit the number of dentry locks to drop in one go. We have + * this limit so that we don't starve other users of ocfs2_wq. */ +#define DL_INODE_DROP_COUNT 64 + +/* Drop inode references from dentry locks */ +void ocfs2_drop_dl_inodes(struct work_struct *work) +{ + struct ocfs2_super *osb = container_of(work, struct ocfs2_super, + dentry_lock_work); + struct ocfs2_dentry_lock *dl; + int drop_count = DL_INODE_DROP_COUNT; + + spin_lock(&dentry_list_lock); + while (osb->dentry_lock_list && drop_count--) { + dl = osb->dentry_lock_list; + osb->dentry_lock_list = dl->dl_next; + spin_unlock(&dentry_list_lock); + iput(dl->dl_inode); + kfree(dl); + spin_lock(&dentry_list_lock); + } + if (osb->dentry_lock_list) + queue_work(ocfs2_wq, &osb->dentry_lock_work); + spin_unlock(&dentry_list_lock); +} + /* * ocfs2_dentry_iput() and friends. * @@ -318,16 +347,23 @@ out_attach: static void ocfs2_drop_dentry_lock(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl) { - iput(dl->dl_inode); ocfs2_simple_drop_lockres(osb, &dl->dl_lockres); ocfs2_lock_res_free(&dl->dl_lockres); - kfree(dl); + + /* We leave dropping of inode reference to ocfs2_wq as that can + * possibly lead to inode deletion which gets tricky */ + spin_lock(&dentry_list_lock); + if (!osb->dentry_lock_list) + queue_work(ocfs2_wq, &osb->dentry_lock_work); + dl->dl_next = osb->dentry_lock_list; + osb->dentry_lock_list = dl; + spin_unlock(&dentry_list_lock); } void ocfs2_dentry_lock_put(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl) { - int unlock = 0; + int unlock; BUG_ON(dl->dl_count == 0); diff --git a/fs/ocfs2/dcache.h b/fs/ocfs2/dcache.h index c091c34d9883..d06e16c06640 100644 --- a/fs/ocfs2/dcache.h +++ b/fs/ocfs2/dcache.h @@ -29,8 +29,13 @@ extern struct dentry_operations ocfs2_dentry_ops; struct ocfs2_dentry_lock { + /* Use count of dentry lock */ unsigned int dl_count; - u64 dl_parent_blkno; + union { + /* Linked list of dentry locks to release */ + struct ocfs2_dentry_lock *dl_next; + u64 dl_parent_blkno; + }; /* * The ocfs2_dentry_lock keeps an inode reference until @@ -47,6 +52,8 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry, struct inode *inode, void ocfs2_dentry_lock_put(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl); +void ocfs2_drop_dl_inodes(struct work_struct *work); + struct dentry *ocfs2_find_local_alias(struct inode *inode, u64 parent_blkno, int skip_unhashed); diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index ad5c24a29edd..077384135f4e 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -210,6 +210,7 @@ struct ocfs2_journal; struct ocfs2_slot_info; struct ocfs2_recovery_map; struct ocfs2_quota_recovery; +struct ocfs2_dentry_lock; struct ocfs2_super { struct task_struct *commit_task; @@ -325,6 +326,11 @@ struct ocfs2_super struct list_head blocked_lock_list; unsigned long blocked_lock_count; + /* List of dentry locks to release. Anyone can add locks to + * the list, ocfs2_wq processes the list */ + struct ocfs2_dentry_lock *dentry_lock_list; + struct work_struct dentry_lock_work; + wait_queue_head_t osb_mount_event; /* Truncate log info */ diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 43ed11345b59..b1cb38fbe807 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1887,6 +1887,9 @@ static int ocfs2_initialize_super(struct super_block *sb, INIT_WORK(&journal->j_recovery_work, ocfs2_complete_recovery); journal->j_state = OCFS2_JOURNAL_FREE; + INIT_WORK(&osb->dentry_lock_work, ocfs2_drop_dl_inodes); + osb->dentry_lock_list = NULL; + /* get some pseudo constants for clustersize bits */ osb->s_clustersize_bits = le32_to_cpu(di->id2.i_super.s_clustersize_bits); -- cgit v1.2.3 From f8afead7169f0f28a4b421bcbdb510e52a2d094d Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 12 Jan 2009 23:20:32 +0100 Subject: ocfs2: Fix possible deadlock in ocfs2_write_dquot() It could happen that some limit has been set via quotactl() and in parallel ->mark_dirty() is called from another thread doing e.g. dquot_alloc_space(). In such case ocfs2_write_dquot() must not try to sync the dquot because that needs global quota lock but that ranks above transaction start. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/ocfs2/quota_global.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index f4efa89baee5..1ed0f7c86869 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -754,7 +754,9 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot) if (dquot->dq_flags & mask) sync = 1; spin_unlock(&dq_data_lock); - if (!sync) { + /* This is a slight hack but we can't afford getting global quota + * lock if we already have a transaction started. */ + if (!sync || journal_current_handle()) { status = ocfs2_write_dquot(dquot); goto out; } -- cgit v1.2.3 From 554e7f9e043e29da79c044f7a55efe4fad40701e Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Thu, 8 Jan 2009 08:21:43 +0800 Subject: ocfs2: Access the xattr bucket only before modifying it. In ocfs2_xattr_value_truncate, we may call b-tree codes which will extend the journal transaction. It has a potential problem that it may let the already-accessed-but-not-dirtied buffers gone. So we'd better access the bucket after we call ocfs2_xattr_value_truncate. And as for the root buffer for the xattr value, b-tree code will acess and dirty it, so we don't need to worry about it. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/xattr.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index e1d638af6ac3..915039fffe6e 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -4729,13 +4729,6 @@ static int ocfs2_xattr_bucket_value_truncate(struct inode *inode, vb.vb_xv = (struct ocfs2_xattr_value_root *) (vb.vb_bh->b_data + offset % blocksize); - ret = ocfs2_xattr_bucket_journal_access(ctxt->handle, bucket, - OCFS2_JOURNAL_ACCESS_WRITE); - if (ret) { - mlog_errno(ret); - goto out; - } - /* * From here on out we have to dirty the bucket. The generic * value calls only modify one of the bucket's bhs, but we need @@ -4748,12 +4741,18 @@ static int ocfs2_xattr_bucket_value_truncate(struct inode *inode, ret = ocfs2_xattr_value_truncate(inode, &vb, len, ctxt); if (ret) { mlog_errno(ret); - goto out_dirty; + goto out; + } + + ret = ocfs2_xattr_bucket_journal_access(ctxt->handle, bucket, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; } xe->xe_value_size = cpu_to_le64(len); -out_dirty: ocfs2_xattr_bucket_journal_dirty(ctxt->handle, bucket); out: -- cgit v1.2.3 From a4b91965d39d5d53b470d6aa62cba155a6f3ffe1 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Thu, 29 Jan 2009 17:12:31 -0800 Subject: ocfs2: Wakeup the downconvert thread after a successful cancel convert When two nodes holding PR locks on a resource concurrently attempt to upconvert the locks to EX, the master sends a BAST to one of the nodes. This message tells that node to first cancel convert the upconvert request, followed by downconvert to a NL. Only when this lock is downconverted to NL, can the master upconvert the first node's lock to EX. While the fs was doing the cancel convert, it was forgetting to wake up the dc thread after a successful cancel, leading to a deadlock. Reported-and-Tested-by: David Teigland Signed-off-by: Sunil Mushran Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index b0c4cadd4c45..206a2370876a 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -2860,6 +2860,10 @@ static void ocfs2_unlock_ast(void *opaque, int error) case OCFS2_UNLOCK_CANCEL_CONVERT: mlog(0, "Cancel convert success for %s\n", lockres->l_name); lockres->l_action = OCFS2_AST_INVALID; + /* Downconvert thread may have requeued this lock, we + * need to wake it. */ + if (lockres->l_flags & OCFS2_LOCK_BLOCKED) + ocfs2_wake_downconvert_thread(ocfs2_get_lockres_osb(lockres)); break; case OCFS2_UNLOCK_DROP_LOCK: lockres->l_level = DLM_LOCK_IV; -- cgit v1.2.3 From fd4ef231962ab44fd1004e87f9d7c6809f00cd64 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Thu, 29 Jan 2009 15:06:21 -0800 Subject: ocfs2: add quota call to ocfs2_remove_btree_range() We weren't reclaiming the clusters which get free'd from this function, so any user punching holes in a file would still have those bytes accounted against him/her. Add the call to vfs_dq_free_space_nodirty() to fix this. Interestingly enough, the journal credits calculation already took this into account. Signed-off-by: Mark Fasheh Acked-by: Jan Kara --- fs/ocfs2/alloc.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index d861096c9d81..60fe74035db5 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -5390,6 +5390,9 @@ int ocfs2_remove_btree_range(struct inode *inode, goto out; } + vfs_dq_free_space_nodirty(inode, + ocfs2_clusters_to_bytes(inode->i_sb, len)); + ret = ocfs2_remove_extent(inode, et, cpos, len, handle, meta_ac, dealloc); if (ret) { -- cgit v1.2.3 From 7f5aa215088b817add9c71914b83650bdd49f8a9 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 10 Feb 2009 11:15:34 -0500 Subject: jbd2: Avoid possible NULL dereference in jbd2_journal_begin_ordered_truncate() If we race with commit code setting i_transaction to NULL, we could possibly dereference it. Proper locking requires the journal pointer (to access journal->j_list_lock), which we don't have. So we have to change the prototype of the function so that filesystem passes us the journal pointer. Also add a more detailed comment about why the function jbd2_journal_begin_ordered_truncate() does what it does and how it should be used. Thanks to Dan Carpenter for pointing to the suspitious code. Signed-off-by: Jan Kara Signed-off-by: "Theodore Ts'o" Acked-by: Joel Becker CC: linux-ext4@vger.kernel.org CC: ocfs2-devel@oss.oracle.com CC: mfasheh@suse.de CC: Dan Carpenter --- fs/ext4/inode.c | 6 ++++-- fs/jbd2/transaction.c | 42 +++++++++++++++++++++++++++++++----------- fs/ocfs2/journal.h | 6 ++++-- include/linux/jbd2.h | 3 ++- 4 files changed, 41 insertions(+), 16 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 03ba20be1329..658c4a7f2578 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -47,8 +47,10 @@ static inline int ext4_begin_ordered_truncate(struct inode *inode, loff_t new_size) { - return jbd2_journal_begin_ordered_truncate(&EXT4_I(inode)->jinode, - new_size); + return jbd2_journal_begin_ordered_truncate( + EXT4_SB(inode->i_sb)->s_journal, + &EXT4_I(inode)->jinode, + new_size); } static void ext4_invalidatepage(struct page *page, unsigned long offset); diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 46b4e347ed7d..28ce21d8598e 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -2129,26 +2129,46 @@ done: } /* - * This function must be called when inode is journaled in ordered mode - * before truncation happens. It starts writeout of truncated part in - * case it is in the committing transaction so that we stand to ordered - * mode consistency guarantees. + * File truncate and transaction commit interact with each other in a + * non-trivial way. If a transaction writing data block A is + * committing, we cannot discard the data by truncate until we have + * written them. Otherwise if we crashed after the transaction with + * write has committed but before the transaction with truncate has + * committed, we could see stale data in block A. This function is a + * helper to solve this problem. It starts writeout of the truncated + * part in case it is in the committing transaction. + * + * Filesystem code must call this function when inode is journaled in + * ordered mode before truncation happens and after the inode has been + * placed on orphan list with the new inode size. The second condition + * avoids the race that someone writes new data and we start + * committing the transaction after this function has been called but + * before a transaction for truncate is started (and furthermore it + * allows us to optimize the case where the addition to orphan list + * happens in the same transaction as write --- we don't have to write + * any data in such case). */ -int jbd2_journal_begin_ordered_truncate(struct jbd2_inode *inode, +int jbd2_journal_begin_ordered_truncate(journal_t *journal, + struct jbd2_inode *jinode, loff_t new_size) { - journal_t *journal; - transaction_t *commit_trans; + transaction_t *inode_trans, *commit_trans; int ret = 0; - if (!inode->i_transaction && !inode->i_next_transaction) + /* This is a quick check to avoid locking if not necessary */ + if (!jinode->i_transaction) goto out; - journal = inode->i_transaction->t_journal; + /* Locks are here just to force reading of recent values, it is + * enough that the transaction was not committing before we started + * a transaction adding the inode to orphan list */ spin_lock(&journal->j_state_lock); commit_trans = journal->j_committing_transaction; spin_unlock(&journal->j_state_lock); - if (inode->i_transaction == commit_trans) { - ret = filemap_fdatawrite_range(inode->i_vfs_inode->i_mapping, + spin_lock(&journal->j_list_lock); + inode_trans = jinode->i_transaction; + spin_unlock(&journal->j_list_lock); + if (inode_trans == commit_trans) { + ret = filemap_fdatawrite_range(jinode->i_vfs_inode->i_mapping, new_size, LLONG_MAX); if (ret) jbd2_journal_abort(journal, ret); diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index 3c3532e1307c..172850a9a12a 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -513,8 +513,10 @@ static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode) static inline int ocfs2_begin_ordered_truncate(struct inode *inode, loff_t new_size) { - return jbd2_journal_begin_ordered_truncate(&OCFS2_I(inode)->ip_jinode, - new_size); + return jbd2_journal_begin_ordered_truncate( + OCFS2_SB(inode->i_sb)->journal->j_journal, + &OCFS2_I(inode)->ip_jinode, + new_size); } #endif /* OCFS2_JOURNAL_H */ diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index b28b37eb11c6..4d248b3f1323 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -1150,7 +1150,8 @@ extern int jbd2_journal_clear_err (journal_t *); extern int jbd2_journal_bmap(journal_t *, unsigned long, unsigned long long *); extern int jbd2_journal_force_commit(journal_t *); extern int jbd2_journal_file_inode(handle_t *handle, struct jbd2_inode *inode); -extern int jbd2_journal_begin_ordered_truncate(struct jbd2_inode *inode, loff_t new_size); +extern int jbd2_journal_begin_ordered_truncate(journal_t *journal, + struct jbd2_inode *inode, loff_t new_size); extern void jbd2_journal_init_jbd_inode(struct jbd2_inode *jinode, struct inode *inode); extern void jbd2_journal_release_jbd_inode(journal_t *journal, struct jbd2_inode *jinode); -- cgit v1.2.3 From 47be12e4eec84c1846f29af64fe25a396b57a026 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Fri, 9 Jan 2009 07:32:48 +0800 Subject: ocfs2: Access and dirty the buffer_head in mark_written. In __ocfs2_mark_extent_written, when we meet with the situation of c_split_covers_rec, the old solution just replace the extent record and forget to access and dirty the buffer_head. This will cause a problem when the unwritten extent is in an extent block. So access and dirty it. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/alloc.c | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 60fe74035db5..3a9e5deed74d 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -4796,6 +4796,29 @@ out: return ret; } +static int ocfs2_replace_extent_rec(struct inode *inode, + handle_t *handle, + struct ocfs2_path *path, + struct ocfs2_extent_list *el, + int split_index, + struct ocfs2_extent_rec *split_rec) +{ + int ret; + + ret = ocfs2_path_bh_journal_access(handle, inode, path, + path_num_items(path) - 1); + if (ret) { + mlog_errno(ret); + goto out; + } + + el->l_recs[split_index] = *split_rec; + + ocfs2_journal_dirty(handle, path_leaf_bh(path)); +out: + return ret; +} + /* * Mark part or all of the extent record at split_index in the leaf * pointed to by path as written. This removes the unwritten @@ -4885,7 +4908,9 @@ static int __ocfs2_mark_extent_written(struct inode *inode, if (ctxt.c_contig_type == CONTIG_NONE) { if (ctxt.c_split_covers_rec) - el->l_recs[split_index] = *split_rec; + ret = ocfs2_replace_extent_rec(inode, handle, + path, el, + split_index, split_rec); else ret = ocfs2_split_and_insert(inode, handle, path, et, &last_eb_bh, split_index, -- cgit v1.2.3 From 7dc102b737e9f49dac426161294cb2d326a97d8e Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Tue, 3 Feb 2009 12:37:13 -0800 Subject: ocfs2/dlm: Retract fix for race between purge and migrate Mainline commit d4f7e650e55af6b235871126f747da88600e8040 attempts to delay the dlm_thread from sending the drop ref message if the lockres is being migrated. The problem is that we make the dlm_thread wait for the migration to complete. This causes a deadlock as dlm_thread also participates in the lockres migration process. A better fix for the original oss bugzilla#1012 is in testing. Signed-off-by: Sunil Mushran Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmthread.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index d1295203029f..4060bb328bc8 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -181,8 +181,7 @@ static int dlm_purge_lockres(struct dlm_ctxt *dlm, spin_lock(&res->spinlock); /* This ensures that clear refmap is sent after the set */ - __dlm_wait_on_lockres_flags(res, (DLM_LOCK_RES_SETREF_INPROG | - DLM_LOCK_RES_MIGRATING)); + __dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG); spin_unlock(&res->spinlock); /* clear our bit from the master's refmap, ignore errors */ -- cgit v1.2.3 From c74ff8bb2235d848beb67fcfddae71ecbe3f92b1 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Tue, 3 Feb 2009 12:37:14 -0800 Subject: ocfs2: Cleanup the lockname print in dlmglue.c The dentry lock has a different format than other locks. This patch fixes ocfs2_log_dlm_error() macro to make it print the dentry lock correctly. Signed-off-by: Sunil Mushran Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 206a2370876a..7219a86d34cc 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -320,9 +320,14 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres); static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres, int convert); -#define ocfs2_log_dlm_error(_func, _err, _lockres) do { \ - mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \ - _err, _func, _lockres->l_name); \ +#define ocfs2_log_dlm_error(_func, _err, _lockres) do { \ + if ((_lockres)->l_type != OCFS2_LOCK_TYPE_DENTRY) \ + mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \ + _err, _func, _lockres->l_name); \ + else \ + mlog(ML_ERROR, "DLM error %d while calling %s on resource %.*s%08x\n", \ + _err, _func, OCFS2_DENTRY_LOCK_INO_START - 1, (_lockres)->l_name, \ + (unsigned int)ocfs2_get_dentry_lock_ino(_lockres)); \ } while (0) static int ocfs2_downconvert_thread(void *arg); static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb, -- cgit v1.2.3 From dabc47de7a23f57522dc762d9d2ad875700d3497 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Tue, 3 Feb 2009 12:37:15 -0800 Subject: ocfs2/dlm: Use ast_lock to protect ast_list The code was using dlm->spinlock instead of dlm->ast_lock to protect the ast_list. This patch fixes the issue. Signed-off-by: Sunil Mushran Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmunlock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c index 86ca085ef324..fcf879ed6930 100644 --- a/fs/ocfs2/dlm/dlmunlock.c +++ b/fs/ocfs2/dlm/dlmunlock.c @@ -117,11 +117,11 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm, else BUG_ON(res->owner == dlm->node_num); - spin_lock(&dlm->spinlock); + spin_lock(&dlm->ast_lock); /* We want to be sure that we're not freeing a lock * that still has AST's pending... */ in_use = !list_empty(&lock->ast_list); - spin_unlock(&dlm->spinlock); + spin_unlock(&dlm->ast_lock); if (in_use) { mlog(ML_ERROR, "lockres %.*s: Someone is calling dlmunlock " "while waiting for an ast!", res->lockname.len, -- cgit v1.2.3 From 53ecd25e148615e0ed2a72635cc76f4773f97f90 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Tue, 3 Feb 2009 12:37:16 -0800 Subject: ocfs2/dlm: Make dlm_assert_master_handler() kill itself instead of the asserter In dlm_assert_master_handler(), if we get an incorrect assert master from a node that, we reply with EINVAL asking the asserter to die. The problem is that an assert is sent after so many hoops, it is invariably the node that thinks the asserter is wrong, is actually wrong. So instead of killing the asserter, this patch kills the assertee. This patch papers over a race that is still being addressed. Signed-off-by: Sunil Mushran Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmmaster.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 54e182a27caf..0a2813947853 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -1849,12 +1849,12 @@ int dlm_assert_master_handler(struct o2net_msg *msg, u32 len, void *data, if (!mle) { if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN && res->owner != assert->node_idx) { - mlog(ML_ERROR, "assert_master from " - "%u, but current owner is " - "%u! (%.*s)\n", - assert->node_idx, res->owner, - namelen, name); - goto kill; + mlog(ML_ERROR, "DIE! Mastery assert from %u, " + "but current owner is %u! (%.*s)\n", + assert->node_idx, res->owner, namelen, + name); + __dlm_print_one_lock_resource(res); + BUG(); } } else if (mle->type != DLM_MLE_MIGRATION) { if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN) { -- cgit v1.2.3 From 89a907afe073b8971a83d0ad54f391542b64d327 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Tue, 17 Feb 2009 04:39:28 +0800 Subject: ocfs2: Use the right access_* method in ctime update of xattr. In ctime updating of xattr, it use the wrong type of access for inode, so use ocfs2_journal_access_di instead. Reported-and-Tested-by: Tristan Ye Signed-off-by: Tao Ma Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/xattr.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index 915039fffe6e..e3933158e1d7 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -2592,8 +2592,9 @@ static int __ocfs2_xattr_set_handle(struct inode *inode, if (!ret) { /* Update inode ctime. */ - ret = ocfs2_journal_access(ctxt->handle, inode, xis->inode_bh, - OCFS2_JOURNAL_ACCESS_WRITE); + ret = ocfs2_journal_access_di(ctxt->handle, inode, + xis->inode_bh, + OCFS2_JOURNAL_ACCESS_WRITE); if (ret) { mlog_errno(ret); goto out; -- cgit v1.2.3 From c8b9cf9a7cd25ba65166116d0a958f0bc709f0a7 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Tue, 24 Feb 2009 17:40:26 -0800 Subject: ocfs2: lock the metaecc process for xattr bucket For other metadata in ocfs2, metaecc is checked in ocfs2_read_blocks with io_mutex held. While for xattr bucket, it is calculated by the whole buckets. So we have to add a spin_lock to prevent multiple processes calculating metaecc. Signed-off-by: Tao Ma Tested-by: Tristan Ye Signed-off-by: Mark Fasheh --- fs/ocfs2/ocfs2.h | 3 +++ fs/ocfs2/super.c | 1 + fs/ocfs2/xattr.c | 4 ++++ 3 files changed, 8 insertions(+) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 077384135f4e..946d3c34b90b 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -341,6 +341,9 @@ struct ocfs2_super struct ocfs2_node_map osb_recovering_orphan_dirs; unsigned int *osb_orphan_wipes; wait_queue_head_t osb_wipe_event; + + /* used to protect metaecc calculation check of xattr. */ + spinlock_t osb_xattr_lock; }; #define OCFS2_SB(sb) ((struct ocfs2_super *)(sb)->s_fs_info) diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index b1cb38fbe807..1c3acc4654d8 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1747,6 +1747,7 @@ static int ocfs2_initialize_super(struct super_block *sb, INIT_LIST_HEAD(&osb->blocked_lock_list); osb->blocked_lock_count = 0; spin_lock_init(&osb->osb_lock); + spin_lock_init(&osb->osb_xattr_lock); ocfs2_init_inode_steal_slot(osb); atomic_set(&osb->alloc_stats.moves, 0); diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index e3933158e1d7..a7c167905c56 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -274,10 +274,12 @@ static int ocfs2_read_xattr_bucket(struct ocfs2_xattr_bucket *bucket, bucket->bu_blocks, bucket->bu_bhs, 0, NULL); if (!rc) { + spin_lock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock); rc = ocfs2_validate_meta_ecc_bhs(bucket->bu_inode->i_sb, bucket->bu_bhs, bucket->bu_blocks, &bucket_xh(bucket)->xh_check); + spin_unlock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock); if (rc) mlog_errno(rc); } @@ -310,9 +312,11 @@ static void ocfs2_xattr_bucket_journal_dirty(handle_t *handle, { int i; + spin_lock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock); ocfs2_compute_meta_ecc_bhs(bucket->bu_inode->i_sb, bucket->bu_bhs, bucket->bu_blocks, &bucket_xh(bucket)->xh_check); + spin_unlock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock); for (i = 0; i < bucket->bu_blocks; i++) ocfs2_journal_dirty(handle, bucket->bu_bhs[i]); -- cgit v1.2.3 From 4442f518269c6b3686fcbcadad22dc4475309b16 Mon Sep 17 00:00:00 2001 From: Tiger Yang Date: Fri, 20 Feb 2009 11:11:50 +0800 Subject: ocfs2: set gap to seperate entry and value when xattr in bucket This patch set a gap (4 bytes) between xattr entry and name/value when xattr in bucket. This gap use to seperate entry and name/value when a bucket is full. It had already been set when xattr in inode/block. Signed-off-by: Tiger Yang Signed-off-by: Mark Fasheh --- fs/ocfs2/xattr.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index a7c167905c56..4ddd788add67 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -82,13 +82,14 @@ struct ocfs2_xattr_set_ctxt { #define OCFS2_XATTR_ROOT_SIZE (sizeof(struct ocfs2_xattr_def_value_root)) #define OCFS2_XATTR_INLINE_SIZE 80 +#define OCFS2_XATTR_HEADER_GAP 4 #define OCFS2_XATTR_FREE_IN_IBODY (OCFS2_MIN_XATTR_INLINE_SIZE \ - sizeof(struct ocfs2_xattr_header) \ - - sizeof(__u32)) + - OCFS2_XATTR_HEADER_GAP) #define OCFS2_XATTR_FREE_IN_BLOCK(ptr) ((ptr)->i_sb->s_blocksize \ - sizeof(struct ocfs2_xattr_block) \ - sizeof(struct ocfs2_xattr_header) \ - - sizeof(__u32)) + - OCFS2_XATTR_HEADER_GAP) static struct ocfs2_xattr_def_value_root def_xv = { .xv.xr_list.l_count = cpu_to_le16(1), @@ -1511,7 +1512,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode, last += 1; } - free = min_offs - ((void *)last - xs->base) - sizeof(__u32); + free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP; if (free < 0) return -EIO; @@ -2194,7 +2195,7 @@ static int ocfs2_xattr_can_be_in_inode(struct inode *inode, last += 1; } - free = min_offs - ((void *)last - xs->base) - sizeof(__u32); + free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP; if (free < 0) return 0; @@ -5065,8 +5066,8 @@ try_again: xh_free_start = le16_to_cpu(xh->xh_free_start); header_size = sizeof(struct ocfs2_xattr_header) + count * sizeof(struct ocfs2_xattr_entry); - max_free = OCFS2_XATTR_BUCKET_SIZE - - le16_to_cpu(xh->xh_name_value_len) - header_size; + max_free = OCFS2_XATTR_BUCKET_SIZE - header_size - + le16_to_cpu(xh->xh_name_value_len) - OCFS2_XATTR_HEADER_GAP; mlog_bug_on_msg(header_size > blocksize, "bucket %llu has header size " "of %u which exceed block size\n", @@ -5099,7 +5100,7 @@ try_again: need = 0; } - free = xh_free_start - header_size; + free = xh_free_start - header_size - OCFS2_XATTR_HEADER_GAP; /* * We need to make sure the new name/value pair * can exist in the same block. @@ -5132,7 +5133,8 @@ try_again: } xh_free_start = le16_to_cpu(xh->xh_free_start); - free = xh_free_start - header_size; + free = xh_free_start - header_size + - OCFS2_XATTR_HEADER_GAP; if (xh_free_start % blocksize < need) free -= xh_free_start % blocksize; -- cgit v1.2.3 From 28d57d437786eb3e44f1ca3f0f41e7cfe29c6dd4 Mon Sep 17 00:00:00 2001 From: wengang wang Date: Fri, 13 Feb 2009 10:11:47 +0800 Subject: ocfs2: add IO error check in ocfs2_get_sector() Check for IO error in ocfs2_get_sector(). Signed-off-by: Wengang Wang Signed-off-by: Mark Fasheh --- fs/ocfs2/super.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 1c3acc4654d8..7ac83a81ee55 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1537,6 +1537,13 @@ static int ocfs2_get_sector(struct super_block *sb, unlock_buffer(*bh); ll_rw_block(READ, 1, bh); wait_on_buffer(*bh); + if (!buffer_uptodate(*bh)) { + mlog_errno(-EIO); + brelse(*bh); + *bh = NULL; + return -EIO; + } + return 0; } -- cgit v1.2.3 From 6c9fd1dc0a597e575617a7de7086c8a3efa8f524 Mon Sep 17 00:00:00 2001 From: Tiger Yang Date: Fri, 6 Mar 2009 10:19:30 +0800 Subject: ocfs2: reserve xattr block for new directory with inline data If this is a new directory with inline data, we choose to reserve the entire inline area for directory contents and force an external xattr block. Signed-off-by: Tiger Yang Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/xattr.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index 4ddd788add67..c63efb5ef136 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -547,8 +547,12 @@ int ocfs2_calc_xattr_init(struct inode *dir, * when blocksize = 512, may reserve one more cluser for * xattr bucket, otherwise reserve one metadata block * for them is ok. + * If this is a new directory with inline data, + * we choose to reserve the entire inline area for + * directory contents and force an external xattr block. */ if (dir->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE || + (S_ISDIR(mode) && ocfs2_supports_inline_data(osb)) || (s_size + a_size) > OCFS2_XATTR_FREE_IN_IBODY) { ret = ocfs2_reserve_new_metadata_blocks(osb, 1, xattr_ac); if (ret) { -- cgit v1.2.3 From d9ae49d6e2b1ac9166e58ae3c9345135604beaa6 Mon Sep 17 00:00:00 2001 From: Tiger Yang Date: Thu, 5 Mar 2009 11:06:15 +0800 Subject: ocfs2: tweak to get the maximum inline data size with xattr Replace max_inline_data with max_inline_data_with_xattr to ensure it correct when xattr inlined. Signed-off-by: Tiger Yang Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/aops.c | 7 +++++-- fs/ocfs2/namei.c | 3 ++- fs/ocfs2/ocfs2_fs.h | 6 ------ 3 files changed, 7 insertions(+), 9 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index a067a6cffb01..8e1709a679b7 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -227,7 +227,7 @@ int ocfs2_read_inline_data(struct inode *inode, struct page *page, size = i_size_read(inode); if (size > PAGE_CACHE_SIZE || - size > ocfs2_max_inline_data(inode->i_sb)) { + size > ocfs2_max_inline_data_with_xattr(inode->i_sb, di)) { ocfs2_error(inode->i_sb, "Inode %llu has with inline data has bad size: %Lu", (unsigned long long)OCFS2_I(inode)->ip_blkno, @@ -1555,6 +1555,7 @@ static int ocfs2_try_to_write_inline_data(struct address_space *mapping, int ret, written = 0; loff_t end = pos + len; struct ocfs2_inode_info *oi = OCFS2_I(inode); + struct ocfs2_dinode *di = NULL; mlog(0, "Inode %llu, write of %u bytes at off %llu. features: 0x%x\n", (unsigned long long)oi->ip_blkno, len, (unsigned long long)pos, @@ -1587,7 +1588,9 @@ static int ocfs2_try_to_write_inline_data(struct address_space *mapping, /* * Check whether the write can fit. */ - if (mmap_page || end > ocfs2_max_inline_data(inode->i_sb)) + di = (struct ocfs2_dinode *)wc->w_di_bh->b_data; + if (mmap_page || + end > ocfs2_max_inline_data_with_xattr(inode->i_sb, di)) return 0; do_inline_write: diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index 084aba86c3b2..4b11762f249e 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -532,7 +532,8 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb, fe->i_dyn_features = cpu_to_le16(feat | OCFS2_INLINE_DATA_FL); - fe->id2.i_data.id_count = cpu_to_le16(ocfs2_max_inline_data(osb->sb)); + fe->id2.i_data.id_count = cpu_to_le16( + ocfs2_max_inline_data_with_xattr(osb->sb, fe)); } else { fel = &fe->id2.i_list; fel->l_tree_depth = 0; diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index c7ae45aaa36c..2332ef740f4f 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -1070,12 +1070,6 @@ static inline int ocfs2_fast_symlink_chars(struct super_block *sb) offsetof(struct ocfs2_dinode, id2.i_symlink); } -static inline int ocfs2_max_inline_data(struct super_block *sb) -{ - return sb->s_blocksize - - offsetof(struct ocfs2_dinode, id2.i_data.id_data); -} - static inline int ocfs2_max_inline_data_with_xattr(struct super_block *sb, struct ocfs2_dinode *di) { -- cgit v1.2.3 From 74e77eb30d0ecbb12964d005b439c8b84a505b84 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Thu, 12 Mar 2009 06:24:23 +0800 Subject: ocfs2: Fix a bug found by sparse check. We need to use le32_to_cpu to test rec->e_cpos in ocfs2_dinode_insert_check. Signed-off-by: Tao Ma Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/alloc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 3a9e5deed74d..19e3a96aa02c 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -176,7 +176,8 @@ static int ocfs2_dinode_insert_check(struct inode *inode, BUG_ON(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL); mlog_bug_on_msg(!ocfs2_sparse_alloc(osb) && - (OCFS2_I(inode)->ip_clusters != rec->e_cpos), + (OCFS2_I(inode)->ip_clusters != + le32_to_cpu(rec->e_cpos)), "Device %s, asking for sparse allocation: inode %llu, " "cpos %u, clusters %u\n", osb->dev_str, -- cgit v1.2.3 From 712e53e46a1da35fcd88c05aa0c675b10f7c0e9d Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Thu, 12 Mar 2009 08:37:34 +0800 Subject: ocfs2: Use xs->bucket to set xattr value outside A long time ago, xs->base is allocated a 4K size and all the contents in the bucket are copied to the it. Now we use ocfs2_xattr_bucket to abstract xattr bucket and xs->base is initialized to the start of the bu_bhs[0]. So xs->base + offset will overflow when the value root is stored outside the first block. Then why we can survive the xattr test by now? It is because we always read the bucket contiguously now and kernel mm allocate continguous memory for us. We are lucky, but we should fix it. So just get the right value root as other callers do. Signed-off-by: Tao Ma Acked-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/xattr.c | 26 ++++++++++++++++++++------ 1 file changed, 20 insertions(+), 6 deletions(-) (limited to 'fs/ocfs2') diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index c63efb5ef136..2563df89fc2a 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -4795,19 +4795,33 @@ static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode, char *val, int value_len) { - int offset; + int ret, offset, block_off; struct ocfs2_xattr_value_root *xv; struct ocfs2_xattr_entry *xe = xs->here; + struct ocfs2_xattr_header *xh = bucket_xh(xs->bucket); + void *base; BUG_ON(!xs->base || !xe || ocfs2_xattr_is_local(xe)); - offset = le16_to_cpu(xe->xe_name_offset) + - OCFS2_XATTR_SIZE(xe->xe_name_len); + ret = ocfs2_xattr_bucket_get_name_value(inode, xh, + xe - xh->xh_entries, + &block_off, + &offset); + if (ret) { + mlog_errno(ret); + goto out; + } - xv = (struct ocfs2_xattr_value_root *)(xs->base + offset); + base = bucket_block(xs->bucket, block_off); + xv = (struct ocfs2_xattr_value_root *)(base + offset + + OCFS2_XATTR_SIZE(xe->xe_name_len)); - return __ocfs2_xattr_set_value_outside(inode, handle, - xv, val, value_len); + ret = __ocfs2_xattr_set_value_outside(inode, handle, + xv, val, value_len); + if (ret) + mlog_errno(ret); +out: + return ret; } static int ocfs2_rm_xattr_cluster(struct inode *inode, -- cgit v1.2.3