Subject: + ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace.patch added to -mm tree To: rgoldwyn@xxxxxxx,dan.carpenter@xxxxxxxxxx,jlbec@xxxxxxxxxxxx,mfasheh@xxxxxxx,rgoldwyn@xxxxxxxx From: akpm@xxxxxxxxxxxxxxxxxxxx Date: Tue, 10 Dec 2013 14:43:20 -0800 The patch titled Subject: ocfs2: use the new DLM operation callbacks while requesting new lockspace has been added to the -mm tree. Its filename is ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace.patch This patch should soon appear at http://ozlabs.org/~akpm/mmots/broken-out/ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace.patch and later at http://ozlabs.org/~akpm/mmotm/broken-out/ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace.patch Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/SubmitChecklist when testing your code *** The -mm tree is included into linux-next and is updated there every 3-4 working days ------------------------------------------------------ From: Goldwyn Rodrigues <rgoldwyn@xxxxxxx> Subject: ocfs2: use the new DLM operation callbacks while requesting new lockspace Attempt to use the new DLM operations. If it is not supported, use the traditional ocfs2_controld. To exchange ocfs2 versioning, we use the LVB of the version dlm lock. It first attempts to take the lock in EX mode (non-blocking). If successful (which means it is the first mount), it writes the version number and downconverts to PR lock. If it is unsuccessful, it reads the version from the lock. If this becomes the standard (with o2cb as well), it could simplify userspace tools to check if the filesystem is mounted on other nodes. Dan: Since ocfs2_protocol_version are two u8 values, the additional checks with LONG* don't make sense. Signed-off-by: Goldwyn Rodrigues <rgoldwyn@xxxxxxxx> Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> Reviewed-by: Mark Fasheh <mfasheh@xxxxxxx> Cc: Joel Becker <jlbec@xxxxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- fs/ocfs2/stack_user.c | 130 ++++++++++++++++++++++++++++++++-------- 1 file changed, 106 insertions(+), 24 deletions(-) diff -puN fs/ocfs2/stack_user.c~ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace fs/ocfs2/stack_user.c --- a/fs/ocfs2/stack_user.c~ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace +++ a/fs/ocfs2/stack_user.c @@ -23,6 +23,7 @@ #include <linux/mutex.h> #include <linux/slab.h> #include <linux/reboot.h> +#include <linux/sched.h> #include <asm/uaccess.h> #include "stackglue.h" @@ -122,6 +123,7 @@ struct ocfs2_live_connection { struct dlm_lksb oc_version_lksb; char oc_lvb[DLM_LVB_LEN]; struct completion oc_sync_wait; + wait_queue_head_t oc_wait; }; struct ocfs2_control_private { @@ -218,7 +220,7 @@ static int ocfs2_live_connection_attach( mutex_lock(&ocfs2_control_lock); c->oc_conn = conn; - if (atomic_read(&ocfs2_control_opened)) + if ((c->oc_type == NO_CONTROLD) || atomic_read(&ocfs2_control_opened)) list_add(&c->oc_list, &ocfs2_live_connection_list); else { printk(KERN_ERR @@ -897,6 +899,59 @@ static int version_unlock(struct ocfs2_c return sync_unlock(conn, &lc->oc_version_lksb, VERSION_LOCK); } +/* get_protocol_version() + * + * To exchange ocfs2 versioning, we use the LVB of the version dlm lock. + * The algorithm is: + * 1. Attempt to take the lock in EX mode (non-blocking). + * 2. If successful (which means it is the first mount), write the + * version number and downconvert to PR lock. + * 3. If unsuccessful (returns -EAGAIN), read the version from the LVB after + * taking the PR lock. + */ + +static int get_protocol_version(struct ocfs2_cluster_connection *conn) +{ + int ret; + struct ocfs2_live_connection *lc = conn->cc_private; + struct ocfs2_protocol_version pv; + + running_proto.pv_major = + ocfs2_user_plugin.sp_max_proto.pv_major; + running_proto.pv_minor = + ocfs2_user_plugin.sp_max_proto.pv_minor; + + lc->oc_version_lksb.sb_lvbptr = lc->oc_lvb; + ret = version_lock(conn, DLM_LOCK_EX, + DLM_LKF_VALBLK|DLM_LKF_NOQUEUE); + if (!ret) { + conn->cc_version.pv_major = running_proto.pv_major; + conn->cc_version.pv_minor = running_proto.pv_minor; + version_to_lvb(&running_proto, lc->oc_lvb); + version_lock(conn, DLM_LOCK_PR, DLM_LKF_CONVERT|DLM_LKF_VALBLK); + } else if (ret == -EAGAIN) { + ret = version_lock(conn, DLM_LOCK_PR, DLM_LKF_VALBLK); + if (ret) + goto out; + lvb_to_version(lc->oc_lvb, &pv); + + if (pv.pv_minor < 0) { + ret = -ERANGE; + goto out; + } + if ((pv.pv_major != running_proto.pv_major) || + (pv.pv_minor > running_proto.pv_minor)) { + ret = -EINVAL; + goto out; + } + + conn->cc_version.pv_major = pv.pv_major; + conn->cc_version.pv_minor = pv.pv_minor; + } +out: + return ret; +} + static void user_recover_prep(void *arg) { } @@ -925,6 +980,7 @@ static void user_recover_done(void *arg, } lc->oc_our_slot = our_slot; + wake_up(&lc->oc_wait); } const struct dlm_lockspace_ops ocfs2_ls_ops = { @@ -933,11 +989,21 @@ const struct dlm_lockspace_ops ocfs2_ls_ .recover_done = user_recover_done, }; +static int user_cluster_disconnect(struct ocfs2_cluster_connection *conn) +{ + version_unlock(conn); + dlm_release_lockspace(conn->cc_lockspace, 2); + conn->cc_lockspace = NULL; + ocfs2_live_connection_drop(conn->cc_private); + conn->cc_private = NULL; + return 0; +} + static int user_cluster_connect(struct ocfs2_cluster_connection *conn) { dlm_lockspace_t *fsdlm; struct ocfs2_live_connection *lc; - int rc; + int rc, ops_rv; BUG_ON(conn == NULL); @@ -947,11 +1013,44 @@ static int user_cluster_connect(struct o goto out; } - lc->oc_type = WITH_CONTROLD; + init_waitqueue_head(&lc->oc_wait); + init_completion(&lc->oc_sync_wait); + atomic_set(&lc->oc_this_node, 0); + conn->cc_private = lc; + lc->oc_type = NO_CONTROLD; + + rc = dlm_new_lockspace(conn->cc_name, conn->cc_cluster_name, + DLM_LSFL_FS, DLM_LVB_LEN, + &ocfs2_ls_ops, conn, &ops_rv, &fsdlm); + if (rc) + goto out; + + if (ops_rv == -EOPNOTSUPP) { + lc->oc_type = WITH_CONTROLD; + printk(KERN_NOTICE "ocfs2: You seem to be using an older " + "version of dlm_controld and/or ocfs2-tools." + " Please consider upgrading.\n"); + } else if (ops_rv) { + rc = ops_rv; + goto out; + } + conn->cc_lockspace = fsdlm; + rc = ocfs2_live_connection_attach(conn, lc); if (rc) goto out; + if (lc->oc_type == NO_CONTROLD) { + rc = get_protocol_version(conn); + if (rc) { + printk(KERN_ERR "ocfs2: Could not determine" + " locking version\n"); + user_cluster_disconnect(conn); + goto out; + } + wait_event(lc->oc_wait, (atomic_read(&lc->oc_this_node) > 0)); + } + /* * running_proto must have been set before we allowed any mounts * to proceed. @@ -959,40 +1058,20 @@ static int user_cluster_connect(struct o if (fs_protocol_compare(&running_proto, &conn->cc_version)) { printk(KERN_ERR "Unable to mount with fs locking protocol version " - "%u.%u because the userspace control daemon has " - "negotiated %u.%u\n", + "%u.%u because negotiated protocol is %u.%u\n", conn->cc_version.pv_major, conn->cc_version.pv_minor, running_proto.pv_major, running_proto.pv_minor); rc = -EPROTO; ocfs2_live_connection_drop(lc); lc = NULL; - goto out; } - rc = dlm_new_lockspace(conn->cc_name, NULL, DLM_LSFL_FS, DLM_LVB_LEN, - NULL, NULL, NULL, &fsdlm); - if (rc) { - ocfs2_live_connection_drop(lc); - lc = NULL; - goto out; - } - - conn->cc_private = lc; - conn->cc_lockspace = fsdlm; out: if (rc && lc) kfree(lc); return rc; } -static int user_cluster_disconnect(struct ocfs2_cluster_connection *conn) -{ - dlm_release_lockspace(conn->cc_lockspace, 2); - conn->cc_lockspace = NULL; - ocfs2_live_connection_drop(conn->cc_private); - conn->cc_private = NULL; - return 0; -} static int user_cluster_this_node(struct ocfs2_cluster_connection *conn, unsigned int *this_node) @@ -1002,8 +1081,11 @@ static int user_cluster_this_node(struct if (lc->oc_type == WITH_CONTROLD) rc = ocfs2_control_get_this_node(); + else if (lc->oc_type == NO_CONTROLD) + rc = atomic_read(&lc->oc_this_node); else rc = -EINVAL; + if (rc < 0) return rc; _ Patches currently in -mm which might be from rgoldwyn@xxxxxxx are ocfs2-remove-versioning-information.patch ocfs2-add-clustername-to-cluster-connection.patch ocfs2-add-dlm-recovery-callbacks.patch ocfs2-shift-allocation-ocfs2_live_connection-to-user_connect.patch ocfs2-pass-ocfs2_cluster_connection-to-ocfs2_this_node.patch ocfs2-framework-for-version-lvb.patch ocfs2-use-the-new-dlm-operation-callbacks-while-requesting-new-lockspace.patch -- To unsubscribe from this list: send the line "unsubscribe mm-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html