gfs2: Allow some glocks to be used during withdraw

We need to allow some glocks to be enqueued, dequeued, promoted, and demoted
when we're withdrawn. For example, to maintain metadata integrity, we should
disallow the use of inode and rgrp glocks when withdrawn. Other glocks, like
iopen or the transaction glocks may be safely used because none of their
metadata goes through the journal. So in general, we should disallow all
glocks with an address space, and allow all the others. One exception is:
we need to allow our active journal to be demoted so others may recover it.

Allowing glocks after withdraw gives us the ability to take appropriate
action (in a following patch) to have our journal properly replayed by
another node rather than just abandoning the current transactions and
pretending nothing bad happened, leaving the other nodes free to modify
the blocks we had in our journal, which may result in file system
corruption.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>

diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index 1cb471a8bc878e65421936bfea0de7d3e6aad2d0..454d94dd893386593e7b63e63e04fb4d6ceec399 100644 (file)
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -133,6 +133,33 @@ static void gfs2_glock_dealloc(struct rcu_head *rcu)
        }
 }
 
+/**
+ * glock_blocked_by_withdraw - determine if we can still use a glock
+ * @gl: the glock
+ *
+ * We need to allow some glocks to be enqueued, dequeued, promoted, and demoted
+ * when we're withdrawn. For example, to maintain metadata integrity, we should
+ * disallow the use of inode and rgrp glocks when withdrawn. Other glocks, like
+ * iopen or the transaction glocks may be safely used because none of their
+ * metadata goes through the journal. So in general, we should disallow all
+ * glocks that are journaled, and allow all the others. One exception is:
+ * we need to allow our active journal to be promoted and demoted so others
+ * may recover it and we can reacquire it when they're done.
+ */
+static bool glock_blocked_by_withdraw(struct gfs2_glock *gl)
+{
+       struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
+
+       if (likely(!gfs2_withdrawn(sdp)))
+               return false;
+       if (gl->gl_ops->go_flags & GLOF_NONDISK)
+               return false;
+       if (!sdp->sd_jdesc ||
+           gl->gl_name.ln_number == sdp->sd_jdesc->jd_no_addr)
+               return false;
+       return true;
+}
+
 void gfs2_glock_free(struct gfs2_glock *gl)
 {
        struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
@@ -549,8 +576,7 @@ __acquires(&gl->gl_lockref.lock)
        unsigned int lck_flags = (unsigned int)(gh ? gh->gh_flags : 0);
        int ret;
 
-       if (unlikely(gfs2_withdrawn(sdp)) &&
-           target != LM_ST_UNLOCKED)
+       if (target != LM_ST_UNLOCKED && glock_blocked_by_withdraw(gl))
                return;
        lck_flags &= (LM_FLAG_TRY | LM_FLAG_TRY_1CB | LM_FLAG_NOEXP |
                      LM_FLAG_PRIORITY);
@@ -1194,10 +1220,9 @@ trap_recursive:
 int gfs2_glock_nq(struct gfs2_holder *gh)
 {
        struct gfs2_glock *gl = gh->gh_gl;
-       struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
        int error = 0;
 
-       if (unlikely(gfs2_withdrawn(sdp)))
+       if (glock_blocked_by_withdraw(gl))
                return -EIO;
 
        if (test_bit(GLF_LRU, &gl->gl_flags))

diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 1c557457c7536030c91c8c95a84ff5f4c6687c40..3553ca939337b417bc2bbe2dd85a09f864b4077e 100644 (file)
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -579,6 +579,7 @@ static void iopen_go_callback(struct gfs2_glock *gl, bool remote)
 
 const struct gfs2_glock_operations gfs2_meta_glops = {
        .go_type = LM_TYPE_META,
+       .go_flags = GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations gfs2_inode_glops = {
@@ -605,30 +606,33 @@ const struct gfs2_glock_operations gfs2_freeze_glops = {
        .go_xmote_bh = freeze_go_xmote_bh,
        .go_demote_ok = freeze_go_demote_ok,
        .go_type = LM_TYPE_NONDISK,
+       .go_flags = GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations gfs2_iopen_glops = {
        .go_type = LM_TYPE_IOPEN,
        .go_callback = iopen_go_callback,
-       .go_flags = GLOF_LRU,
+       .go_flags = GLOF_LRU | GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations gfs2_flock_glops = {
        .go_type = LM_TYPE_FLOCK,
-       .go_flags = GLOF_LRU,
+       .go_flags = GLOF_LRU | GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations gfs2_nondisk_glops = {
        .go_type = LM_TYPE_NONDISK,
+       .go_flags = GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations gfs2_quota_glops = {
        .go_type = LM_TYPE_QUOTA,
-       .go_flags = GLOF_LVB | GLOF_LRU,
+       .go_flags = GLOF_LVB | GLOF_LRU | GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations gfs2_journal_glops = {
        .go_type = LM_TYPE_JOURNAL,
+       .go_flags = GLOF_NONDISK,
 };
 
 const struct gfs2_glock_operations *gfs2_glops_list[] = {

diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index ab89f746b3b6c74ae13c2a8132f1fca824c9ae5e..3072707aff7a7969dc3a5c8cf1e861281d7c8544 100644 (file)
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -244,9 +244,10 @@ struct gfs2_glock_operations {
        void (*go_callback)(struct gfs2_glock *gl, bool remote);
        const int go_type;
        const unsigned long go_flags;
-#define GLOF_ASPACE 1
-#define GLOF_LVB    2
-#define GLOF_LRU    4
+#define GLOF_ASPACE 1 /* address space attached */
+#define GLOF_LVB    2 /* Lock Value Block attached */
+#define GLOF_LRU    4 /* LRU managed */
+#define GLOF_NONDISK   8 /* not I/O related */
 };
 
 enum {
@@ -541,6 +542,7 @@ struct gfs2_jdesc {
        struct list_head jd_revoke_list;
        unsigned int jd_replay_tail;
 
+       u64 jd_no_addr;
 };
 
 struct gfs2_statfs_change_host {

diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index d19ee57c99ce56f8b5307c28691f5e8dfc69a99f..74389d856dd36f339d4a3757d4cf434064bb4530 100644 (file)
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -552,6 +552,8 @@ static int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh)
        mutex_lock(&sdp->sd_jindex_mutex);
 
        for (;;) {
+               struct gfs2_inode *jip;
+
                error = gfs2_glock_nq_init(dip->i_gl, LM_ST_SHARED, 0, ji_gh);
                if (error)
                        break;
@@ -591,6 +593,8 @@ static int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh)
 
                spin_lock(&sdp->sd_jindex_spin);
                jd->jd_jid = sdp->sd_journals++;
+               jip = GFS2_I(jd->jd_inode);
+               jd->jd_no_addr = jip->i_no_addr;
                list_add_tail(&jd->jd_list, &sdp->sd_jindex_list);
                spin_unlock(&sdp->sd_jindex_spin);
        }