diff options
author | Benjamin Marzinski <bmarzins@redhat.com> | 2014-05-02 03:26:55 (GMT) |
---|---|---|
committer | Steven Whitehouse <swhiteho@redhat.com> | 2014-05-14 09:04:34 (GMT) |
commit | 24972557b12ce8fd5b6c6847d0e2ee1837ddc13b (patch) | |
tree | fb82edff8bfab9d4d1c4df6dd7784dd3c9f85c1b /fs/gfs2/super.c | |
parent | 5a7c6690c2759d686d2c299402327e92ed92ab6c (diff) | |
download | linux-24972557b12ce8fd5b6c6847d0e2ee1837ddc13b.tar.xz |
GFS2: remove transaction glock
GFS2 has a transaction glock, which must be grabbed for every
transaction, whose purpose is to deal with freezing the filesystem.
Aside from this involving a large amount of locking, it is very easy to
make the current fsfreeze code hang on unfreezing.
This patch rewrites how gfs2 handles freezing the filesystem. The
transaction glock is removed. In it's place is a freeze glock, which is
cached (but not held) in a shared state by every node in the cluster
when the filesystem is mounted. This lock only needs to be grabbed on
freezing, and actions which need to be safe from freezing, like
recovery.
When a node wants to freeze the filesystem, it grabs this glock
exclusively. When the freeze glock state changes on the nodes (either
from shared to unlocked, or shared to exclusive), the filesystem does a
special log flush. gfs2_log_flush() does all the work for flushing out
the and shutting down the incore log, and then it tries to grab the
freeze glock in a shared state again. Since the filesystem is stuck in
gfs2_log_flush, no new transaction can start, and nothing can be written
to disk. Unfreezing the filesytem simply involes dropping the freeze
glock, allowing gfs2_log_flush() to grab and then release the shared
lock, so it is cached for next time.
However, in order for the unfreezing ioctl to occur, gfs2 needs to get a
shared lock on the filesystem root directory inode to check permissions.
If that glock has already been grabbed exclusively, fsfreeze will be
unable to get the shared lock and unfreeze the filesystem.
In order to allow the unfreeze, this patch makes gfs2 grab a shared lock
on the filesystem root directory during the freeze, and hold it until it
unfreezes the filesystem. The functions which need to grab a shared
lock in order to allow the unfreeze ioctl to be issued now use the lock
grabbed by the freeze code instead.
The freeze and unfreeze code take care to make sure that this shared
lock will not be dropped while another process is using it.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Diffstat (limited to 'fs/gfs2/super.c')
-rw-r--r-- | fs/gfs2/super.c | 69 |
1 files changed, 43 insertions, 26 deletions
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index de8afad..1319b5c 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -399,7 +399,7 @@ int gfs2_make_fs_rw(struct gfs2_sbd *sdp) { struct gfs2_inode *ip = GFS2_I(sdp->sd_jdesc->jd_inode); struct gfs2_glock *j_gl = ip->i_gl; - struct gfs2_holder t_gh; + struct gfs2_holder thaw_gh; struct gfs2_log_header_host head; int error; @@ -407,7 +407,8 @@ int gfs2_make_fs_rw(struct gfs2_sbd *sdp) if (error) return error; - error = gfs2_glock_nq_init(sdp->sd_trans_gl, LM_ST_SHARED, 0, &t_gh); + error = gfs2_glock_nq_init(sdp->sd_freeze_gl, LM_ST_SHARED, 0, + &thaw_gh); if (error) goto fail_threads; @@ -433,13 +434,13 @@ int gfs2_make_fs_rw(struct gfs2_sbd *sdp) set_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags); - gfs2_glock_dq_uninit(&t_gh); + gfs2_glock_dq_uninit(&thaw_gh); return 0; fail: - t_gh.gh_flags |= GL_NOCACHE; - gfs2_glock_dq_uninit(&t_gh); + thaw_gh.gh_flags |= GL_NOCACHE; + gfs2_glock_dq_uninit(&thaw_gh); fail_threads: kthread_stop(sdp->sd_quotad_process); kthread_stop(sdp->sd_logd_process); @@ -635,15 +636,21 @@ struct lfcc { */ static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp, - struct gfs2_holder *t_gh) + struct gfs2_holder *freeze_gh) { struct gfs2_inode *ip; struct gfs2_jdesc *jd; struct lfcc *lfcc; LIST_HEAD(list); struct gfs2_log_header_host lh; + struct gfs2_inode *dip = GFS2_I(sdp->sd_root_dir->d_inode); int error; + error = gfs2_glock_nq_init(dip->i_gl, LM_ST_SHARED, 0, + &sdp->sd_freeze_root_gh); + if (error) + return error; + atomic_set(&sdp->sd_frozen_root, 1); list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) { lfcc = kmalloc(sizeof(struct lfcc), GFP_KERNEL); if (!lfcc) { @@ -659,8 +666,8 @@ static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp, list_add(&lfcc->list, &list); } - error = gfs2_glock_nq_init(sdp->sd_trans_gl, LM_ST_DEFERRED, - GL_NOCACHE, t_gh); + error = gfs2_glock_nq_init(sdp->sd_freeze_gl, LM_ST_EXCLUSIVE, + GL_NOCACHE, freeze_gh); list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) { error = gfs2_jdesc_check(jd); @@ -676,7 +683,7 @@ static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp, } if (error) - gfs2_glock_dq_uninit(t_gh); + gfs2_glock_dq_uninit(freeze_gh); out: while (!list_empty(&list)) { @@ -685,6 +692,11 @@ out: gfs2_glock_dq_uninit(&lfcc->gh); kfree(lfcc); } + if (error) { + atomic_dec(&sdp->sd_frozen_root); + wait_event(sdp->sd_frozen_root_wait, atomic_read(&sdp->sd_frozen_root) == 0); + gfs2_glock_dq_uninit(&sdp->sd_freeze_root_gh); + } return error; } @@ -742,7 +754,7 @@ static int gfs2_write_inode(struct inode *inode, struct writeback_control *wbc) int ret = 0; if (wbc->sync_mode == WB_SYNC_ALL) - gfs2_log_flush(GFS2_SB(inode), ip->i_gl); + gfs2_log_flush(GFS2_SB(inode), ip->i_gl, NORMAL_FLUSH); if (bdi->dirty_exceeded) gfs2_ail1_flush(sdp, wbc); else @@ -822,9 +834,18 @@ out: static int gfs2_make_fs_ro(struct gfs2_sbd *sdp) { - struct gfs2_holder t_gh; + struct gfs2_holder thaw_gh; int error; + error = gfs2_glock_nq_init(sdp->sd_freeze_gl, LM_ST_SHARED, GL_NOCACHE, + &thaw_gh); + if (error && !test_bit(SDF_SHUTDOWN, &sdp->sd_flags)) + return error; + + down_write(&sdp->sd_log_flush_lock); + clear_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags); + up_write(&sdp->sd_log_flush_lock); + kthread_stop(sdp->sd_quotad_process); kthread_stop(sdp->sd_logd_process); @@ -832,18 +853,11 @@ static int gfs2_make_fs_ro(struct gfs2_sbd *sdp) gfs2_quota_sync(sdp->sd_vfs, 0); gfs2_statfs_sync(sdp->sd_vfs, 0); - error = gfs2_glock_nq_init(sdp->sd_trans_gl, LM_ST_SHARED, GL_NOCACHE, - &t_gh); - if (error && !test_bit(SDF_SHUTDOWN, &sdp->sd_flags)) - return error; - - gfs2_meta_syncfs(sdp); - gfs2_log_shutdown(sdp); - - clear_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags); + gfs2_log_flush(sdp, NULL, SHUTDOWN_FLUSH); + gfs2_assert_warn(sdp, atomic_read(&sdp->sd_log_blks_free) == sdp->sd_jdesc->jd_blocks); - if (t_gh.gh_gl) - gfs2_glock_dq_uninit(&t_gh); + if (thaw_gh.gh_gl) + gfs2_glock_dq_uninit(&thaw_gh); gfs2_quota_cleanup(sdp); @@ -900,7 +914,7 @@ restart: iput(sdp->sd_quota_inode); gfs2_glock_put(sdp->sd_rename_gl); - gfs2_glock_put(sdp->sd_trans_gl); + gfs2_glock_put(sdp->sd_freeze_gl); if (!sdp->sd_args.ar_spectator) { gfs2_glock_dq_uninit(&sdp->sd_journal_gh); @@ -935,8 +949,8 @@ static int gfs2_sync_fs(struct super_block *sb, int wait) struct gfs2_sbd *sdp = sb->s_fs_info; gfs2_quota_sync(sb, -1); - if (wait && sdp) - gfs2_log_flush(sdp, NULL); + if (wait && sdp && !atomic_read(&sdp->sd_log_freeze)) + gfs2_log_flush(sdp, NULL, NORMAL_FLUSH); return 0; } @@ -986,6 +1000,9 @@ static int gfs2_unfreeze(struct super_block *sb) struct gfs2_sbd *sdp = sb->s_fs_info; gfs2_glock_dq_uninit(&sdp->sd_freeze_gh); + atomic_dec(&sdp->sd_frozen_root); + wait_event(sdp->sd_frozen_root_wait, atomic_read(&sdp->sd_frozen_root) == 0); + gfs2_glock_dq_uninit(&sdp->sd_freeze_root_gh); return 0; } @@ -1525,7 +1542,7 @@ static void gfs2_evict_inode(struct inode *inode) goto out_unlock; out_truncate: - gfs2_log_flush(sdp, ip->i_gl); + gfs2_log_flush(sdp, ip->i_gl, NORMAL_FLUSH); if (test_bit(GLF_DIRTY, &ip->i_gl->gl_flags)) { struct address_space *metamapping = gfs2_glock2aspace(ip->i_gl); filemap_fdatawrite(metamapping); |