search for: ip_alloc_sem

Hi, I've been playing lately with lockdep annotations of OCFS2. I seem to have most of the false positives sorted out and currently I hit the report below. I've analyzed that ocfs2_extend_dir() does first lock local alloc inode in ocfs2_reserve_clusters() and then acquires ip_alloc_sem from the directory. The usual ordering e.g. in ocfs2_write_begin() is to first acquire ip_alloc_sem and then lock local alloc. Now these two paths obviously cannot deadlock against each other because one works only for directories and another only for files. But there are some code paths that are s...

Hi, here is the second version of OCFS2 quota fixes with Joel's comments fixed. Also I've added a patch defining counts of credits for quota operations as Joel asked. Honza

Hi, I did some more in-depth testing of OCFS2 quota code, especially with ECC feature enabled and spotted some problems. First four patches fix them. I need the fifth patch to be able to mount OCFS2 filesystem with 2.6.31-rc3. The sixth patch fixes a potential problem when quota syncing interval is updated while the cluster is running (which is not possible currently). Joel, could you please

Hi, the first four patches in this series fix locking problems in OCFS2 quota code (three of them can lead to potential deadlocks). The fifth patch reorders ip_alloc_sem for directories to be acquired before localalloc locks. Mark would you please merge these? The last two patches implement lockdep annotations for OCFS2 cluster locks. We annotate all the cluster locks except for special ones like open lock or local quota file lock - these are not held by any part...

All, Please review the patches backported to 1.4 from mainline. Sunil

Hi, I'm resending this patch series. It's rediffed against linux-next branch of Joel's git tree. The first four patches are obvious fixes of deadlocks in quota code and should go in as soon as possible. The other three patches implement lockdep support for OCFS2 cluster locks. So you can have a look whether the code make sence to you and possibly merge them. They should be NOP when

...m two places: 1) ocfs2_truncate_for_delete() - easy to fix, just move the call just after the write of the inode. 2) ocfs2_setattr() - we can move the call into ocfs2_truncate_file() but that would mean calling jbd2_journal_begin_ordered_truncate() and consequently ocfs2_write_page() under ip_alloc_sem - not too nice. Furthermore ocfs2_orphan_for_truncate() zeros the last cluster beyond i_size and we cannot do that before writing out previous content... Not sure how to solve that yet. Any ideas welcome. Honza -- Jan Kara <jack at suse.cz> SUSE Labs, CR

...-------------- next part -------------- Index: fs/ocfs2/file.c =================================================================== --- fs/ocfs2/file.c (revision 3001) +++ fs/ocfs2/file.c (working copy) @@ -368,8 +368,6 @@ static ssize_t ocfs2_file_sendfile(struc down_read(&OCFS2_I(inode)->ip_alloc_sem); ret = generic_file_sendfile(in_file, ppos, count, actor, target); - if (ret < 0) - mlog_errno(ret); up_read(&OCFS2_I(inode)->ip_alloc_sem);

...fs/ocfs2/quota_global.c @@ -210,7 +210,7 @@ ssize_t ocfs2_quota_write(struct super_block *sb, int type, } mutex_lock_nested(&gqinode->i_mutex, I_MUTEX_QUOTA); - if (gqinode->i_size < off + len) { + if (i_size_read(gqinode) < off + len) { down_write(&OCFS2_I(gqinode)->ip_alloc_sem); err = ocfs2_extend_no_holes(gqinode, off + len, off); up_write(&OCFS2_I(gqinode)->ip_alloc_sem); diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index 07deec5..cf9deab 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -938,14 +938,14 @@ static struct oc...

Two patches follow this message. One fixes the default implementation of SEEK_HOLE/DATA. This patch applies atop Josef's last posted patch. The second patch implements the same on ocfs2. The test tool for the same is available here. http://oss.oracle.com/~smushran/seek_data/seek_test.c It is improved since the last post. It runs cleanly on zfs, ocfs2 and ext3 (default behavior). Users

Two patches follow this message. One fixes the default implementation of SEEK_HOLE/DATA. This patch applies atop Josef's last posted patch. The second patch implements the same on ocfs2. The test tool for the same is available here. http://oss.oracle.com/~smushran/seek_data/seek_test.c It is improved since the last post. It runs cleanly on zfs, ocfs2 and ext3 (default behavior). Users

...e the nonblocking flag for the dlm code to avoid page * lock inversion, but don't bother with retrying. */ ret = ocfs2_inode_lock_full(inode, NULL, 0, OCFS2_LOCK_NONBLOCK); - if (ret) - return err; + if (ret) { + mlog_errno(ret); + goto bail; + } if (down_read_trylock(&oi->ip_alloc_sem) == 0) { ocfs2_inode_unlock(inode, 0); - return err; + goto bail; } /* @@ -378,7 +395,8 @@ static int ocfs2_readpages(struct file *filp, struct address_space *mapping, out_unlock: up_read(&oi->ip_alloc_sem); ocfs2_inode_unlock(inode, 0); - +bail: + mlog_exit(err); return er...

...e the nonblocking flag for the dlm code to avoid page * lock inversion, but don't bother with retrying. */ ret = ocfs2_inode_lock_full(inode, NULL, 0, OCFS2_LOCK_NONBLOCK); - if (ret) - return err; + if (ret) { + mlog_errno(ret); + goto bail; + } if (down_read_trylock(&oi->ip_alloc_sem) == 0) { ocfs2_inode_unlock(inode, 0); - return err; + goto bail; } /* @@ -378,7 +397,8 @@ static int ocfs2_readpages(struct file *filp, struct address_space *mapping, out_unlock: up_read(&oi->ip_alloc_sem); ocfs2_inode_unlock(inode, 0); - +bail: + mlog_exit(err); return er...

...e the nonblocking flag for the dlm code to avoid page * lock inversion, but don't bother with retrying. */ ret = ocfs2_inode_lock_full(inode, NULL, 0, OCFS2_LOCK_NONBLOCK); - if (ret) - return err; + if (ret) { + mlog_errno(ret); + goto bail; + } if (down_read_trylock(&oi->ip_alloc_sem) == 0) { ocfs2_inode_unlock(inode, 0); - return err; + goto bail; } /* @@ -378,7 +397,8 @@ static int ocfs2_readpages(struct file *filp, struct address_space *mapping, out_unlock: up_read(&oi->ip_alloc_sem); ocfs2_inode_unlock(inode, 0); - +bail: + mlog_exit(err); return er...

...writeout Jan Kara (5): ocfs2: Fix possible deadlock in ocfs2_global_read_dquot() ocfs2: Fix lock inversion in ocfs2_local_read_info() ocfs2: Fix possible deadlock with quotas in ocfs2_setattr() ocfs2: Fix possible deadlock in quota recovery ocfs2: Correct ordering of ip_alloc_sem and localloc locks for directories Joel Becker (1): ocfs2: Add statistics for the checksum and ecc operations. Srinivas Eeda (2): ocfs2: timer to queue scan of all orphan slots ocfs2 patch to track delayed orphan scan timer statistics Sunil Mushran (1): ocfs2/net: Use wai...

Hi, here comes the next version of OCFS2 lockdep support. I've dropped patches with fixes from the series since they were already merged. As Joel suggested, I've simplified the main patch a bit so that we don't have ifdefs around lock declarations and there are also a few other minor improvements. Honza

We're currently support two paths from VFS to preallocate unwritten extents(from FS_IOC_RESVSP, or fallocate()), likewise, behavior of punching-hole should be treated as the same, this patch tries to teach file_ioctl() to handle FS_IOC_UNRESVSP, underlying filesystem like ocfs2 is wise enough to do the rest of work;-) Signed-off-by: Tristan Ye <tristan.ye at oracle.com> --- fs/ioctl.c

We're currently support two paths from VFS to preallocate unwritten extents(from FS_IOC_RESVSP, or fallocate()), likewise, behavior of punching-hole should be treated as the same, this patch tries to teach file_ioctl() to handle FS_IOC_UNRESVSP, underlying filesystem like ocfs2 is wise enough to do the rest of work;-) Signed-off-by: Tristan Ye <tristan.ye at oracle.com> --- fs/ioctl.c

Hi all, So I have finally finished the v1 of reflink for ocfs2. It has some bugs that I am still investigating, but the schema is almost there. So I'd like to send it out first for review. And Tristan and I will continue to work on the stability of the code. The general information for reflink, please see http://oss.oracle.com/osswiki/OCFS2/DesignDocs/Reflink. For the design doc, please

Hi all, Change from v1 to v2: bug fix and metadata/credits reservation improvement. The general information for reflink, please see http://oss.oracle.com/osswiki/OCFS2/DesignDocs/Reflink. For the design doc, please see http://oss.oracle.com/osswiki/OCFS2/DesignDocs/RefcountTrees http://oss.oracle.com/osswiki/OCFS2/DesignDocs/ReflinkOperation