Merge: dlm: fix recovery of middle conversions

MR: https://gitlab.com/redhat/centos-stream/src/kernel/centos-stream-9/-/merge_requests/5792

JIRA: https://issues.redhat.com/browse/RHEL-67852
Upstream Status: https://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm.git

Conflicts: ms_stub rename to ms_local and some conversions to __le16

Signed-off-by: Alexander Aring <aahringo@redhat.com>

Approved-by: David Teigland <teigland@redhat.com>
Approved-by: Chris von Recklinghausen <crecklin@redhat.com>
Approved-by: CKI KWF Bot <cki-ci-bot+kwf-gitlab-com@redhat.com>

Merged-by: Rado Vrbovsky <rvrbovsk@redhat.com>

Author: emmil

fs/dlm/lock.c

@@ -5084,17 +5084,20 @@ static void recover_convert_waiter(struct dlm_ls *ls, struct dlm_lkb *lkb,
 				   struct dlm_message *ms_stub)
 {
 	if (middle_conversion(lkb)) {
+		log_rinfo(ls, "%s %x middle convert in progress", __func__,
+			 lkb->lkb_id);
+
+		/* We sent this lock to the new master. The new master will
+		 * tell us when it's granted.  We no longer need a reply, so
+		 * use a fake reply to put the lkb into the right state.
+		 */
 		hold_lkb(lkb);
 		memset(ms_stub, 0, sizeof(struct dlm_message));
 		ms_stub->m_flags = DLM_IFL_STUB_MS;
 		ms_stub->m_type = DLM_MSG_CONVERT_REPLY;
 		ms_stub->m_result = -EINPROGRESS;
 		ms_stub->m_header.h_nodeid = lkb->lkb_nodeid;
 		_receive_convert_reply(lkb, ms_stub);
-
-		/* Same special case as in receive_rcom_lock_args() */
-		lkb->lkb_grmode = DLM_LOCK_IV;
-		rsb_set_flag(lkb->lkb_resource, RSB_RECOVER_CONVERT);
 		unhold_lkb(lkb);
 
 	} else if (lkb->lkb_rqmode >= lkb->lkb_grmode) {
@@ -5618,10 +5621,11 @@ static int receive_rcom_lock_args(struct dlm_ls *ls, struct dlm_lkb *lkb,
 	   The real granted mode of these converting locks cannot be determined
 	   until all locks have been rebuilt on the rsb (recover_conversion) */
 
-	if (rl->rl_wait_type == cpu_to_le16(DLM_MSG_CONVERT) &&
-	    middle_conversion(lkb)) {
-		rl->rl_status = DLM_LKSTS_CONVERT;
-		lkb->lkb_grmode = DLM_LOCK_IV;
+	if (rl->rl_status == DLM_LKSTS_CONVERT && middle_conversion(lkb)) {
+		/* We may need to adjust grmode depending on other granted locks. */
+		log_limit(ls, "%s %x middle convert gr %d rq %d remote %d %x",
+			  __func__, lkb->lkb_id, lkb->lkb_grmode,
+			  lkb->lkb_rqmode, lkb->lkb_nodeid, lkb->lkb_remid);
 		rsb_set_flag(r, RSB_RECOVER_CONVERT);
 	}
 

fs/dlm/recover.c

@@ -813,33 +813,42 @@ static void recover_lvb(struct dlm_rsb *r)
 }
 
 /* All master rsb's flagged RECOVER_CONVERT need to be looked at.  The locks
-   converting PR->CW or CW->PR need to have their lkb_grmode set. */
+ * converting PR->CW or CW->PR may need to have their lkb_grmode changed.
+ */
 
 static void recover_conversion(struct dlm_rsb *r)
 {
 	struct dlm_ls *ls = r->res_ls;
+	uint32_t other_lkid = 0;
+	int other_grmode = -1;
 	struct dlm_lkb *lkb;
-	int grmode = -1;
 
 	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
 		if (lkb->lkb_grmode == DLM_LOCK_PR ||
 		    lkb->lkb_grmode == DLM_LOCK_CW) {
-			grmode = lkb->lkb_grmode;
+			other_grmode = lkb->lkb_grmode;
+			other_lkid = lkb->lkb_id;
 			break;
 		}
 	}
 
+	if (other_grmode == -1)
+		return;
+
 	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue) {
-		if (lkb->lkb_grmode != DLM_LOCK_IV)
-			continue;
-		if (grmode == -1) {
-			log_debug(ls, "recover_conversion %x set gr to rq %d",
-				  lkb->lkb_id, lkb->lkb_rqmode);
-			lkb->lkb_grmode = lkb->lkb_rqmode;
-		} else {
-			log_debug(ls, "recover_conversion %x set gr %d",
-				  lkb->lkb_id, grmode);
-			lkb->lkb_grmode = grmode;
+		/* Lock recovery created incompatible granted modes, so
+		 * change the granted mode of the converting lock to
+		 * NL. The rqmode of the converting lock should be CW,
+		 * which means the converting lock should be granted at
+		 * the end of recovery.
+		 */
+		if (((lkb->lkb_grmode == DLM_LOCK_PR) && (other_grmode == DLM_LOCK_CW)) ||
+		    ((lkb->lkb_grmode == DLM_LOCK_CW) && (other_grmode == DLM_LOCK_PR))) {
+			log_limit(ls, "%s %x gr %d rq %d, remote %d %x, other_lkid %u, other gr %d, set gr=NL",
+				  __func__, lkb->lkb_id, lkb->lkb_grmode,
+				  lkb->lkb_rqmode, lkb->lkb_nodeid,
+				  lkb->lkb_remid, other_lkid, other_grmode);
+			lkb->lkb_grmode = DLM_LOCK_NL;
 		}
 	}
 }