Resguardo inicial

linux-7.2.3/fs/dlm/Kconfig

@@ -0,0 +1,16 @@
+# SPDX-License-Identifier: GPL-2.0-only
+menuconfig DLM
+	tristate "Distributed Lock Manager (DLM)"
+	depends on INET
+	depends on SYSFS && CONFIGFS_FS
+	help
+	A general purpose distributed lock manager for kernel or userspace
+	applications.
+
+config DLM_DEBUG
+	bool "DLM debugging"
+	depends on DLM
+	help
+	Under the debugfs mount point, the name of each lockspace will
+	appear as a file in the "dlm" directory.  The output is the
+	list of resource and locks the local node knows about.

linux-7.2.3/fs/dlm/Makefile

@@ -0,0 +1,21 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-$(CONFIG_DLM) +=		dlm.o
+dlm-y :=			ast.o \
+				config.o \
+				dir.o \
+				lock.o \
+				lockspace.o \
+				main.o \
+				member.o \
+				memory.o \
+				midcomms.o \
+				lowcomms.o \
+				plock.o \
+				rcom.o \
+				recover.o \
+				recoverd.o \
+				requestqueue.o \
+				user.o \
+				util.o 
+dlm-$(CONFIG_DLM_DEBUG) +=	debug_fs.o
+

linux-7.2.3/fs/dlm/ast.c

@@ -0,0 +1,275 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2010 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include <trace/events/dlm.h>
+
+#include "dlm_internal.h"
+#include "lvb_table.h"
+#include "memory.h"
+#include "lock.h"
+#include "user.h"
+#include "ast.h"
+
+static void dlm_run_callback(uint32_t ls_id, uint32_t lkb_id, int8_t mode,
+			     uint32_t flags, uint8_t sb_flags, int sb_status,
+			     struct dlm_lksb *lksb,
+			     void (*astfn)(void *astparam),
+			     void (*bastfn)(void *astparam, int mode),
+			     void *astparam, const char *res_name,
+			     size_t res_length)
+{
+	if (flags & DLM_CB_BAST) {
+		trace_dlm_bast(ls_id, lkb_id, mode, res_name, res_length);
+		bastfn(astparam, mode);
+	} else if (flags & DLM_CB_CAST) {
+		trace_dlm_ast(ls_id, lkb_id, sb_flags, sb_status, res_name,
+			      res_length);
+		lksb->sb_status = sb_status;
+		lksb->sb_flags = sb_flags;
+		astfn(astparam);
+	}
+}
+
+static void dlm_do_callback(struct dlm_callback *cb)
+{
+	dlm_run_callback(cb->ls_id, cb->lkb_id, cb->mode, cb->flags,
+			 cb->sb_flags, cb->sb_status, cb->lkb_lksb,
+			 cb->astfn, cb->bastfn, cb->astparam,
+			 cb->res_name, cb->res_length);
+	dlm_free_cb(cb);
+}
+
+static void dlm_callback_work(struct work_struct *work)
+{
+	struct dlm_callback *cb = container_of(work, struct dlm_callback, work);
+
+	dlm_do_callback(cb);
+}
+
+bool dlm_may_skip_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
+			   int status, uint32_t sbflags, int *copy_lvb)
+{
+	struct dlm_rsb *rsb = lkb->lkb_resource;
+	struct dlm_ls *ls = rsb->res_ls;
+	int prev_mode;
+
+	if (copy_lvb)
+		*copy_lvb = 0;
+
+	if (flags & DLM_CB_BAST) {
+		/* if cb is a bast, it should be skipped if the blocking mode is
+		 * compatible with the last granted mode
+		 */
+		if (lkb->lkb_last_cast_cb_mode != -1) {
+			if (dlm_modes_compat(mode, lkb->lkb_last_cast_cb_mode)) {
+				log_debug(ls, "skip %x bast mode %d for cast mode %d",
+					  lkb->lkb_id, mode,
+					  lkb->lkb_last_cast_cb_mode);
+				return true;
+			}
+		}
+
+		/*
+		 * Suppress some redundant basts here, do more on removal.
+		 * Don't even add a bast if the callback just before it
+		 * is a bast for the same mode or a more restrictive mode.
+		 * (the addional > PR check is needed for PR/CW inversion)
+		 */
+		if (lkb->lkb_last_cb_mode != -1 &&
+		    lkb->lkb_last_cb_flags & DLM_CB_BAST) {
+			prev_mode = lkb->lkb_last_cb_mode;
+
+			if ((prev_mode == mode) ||
+			    (prev_mode > mode && prev_mode > DLM_LOCK_PR)) {
+				log_debug(ls, "skip %x add bast mode %d for bast mode %d",
+					  lkb->lkb_id, mode, prev_mode);
+				return true;
+			}
+		}
+
+		lkb->lkb_last_bast_time = ktime_get();
+		lkb->lkb_last_bast_cb_mode = mode;
+	} else if (flags & DLM_CB_CAST) {
+		if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
+			prev_mode = lkb->lkb_last_cast_cb_mode;
+
+			if (!status && lkb->lkb_lksb->sb_lvbptr &&
+			    dlm_lvb_operations[prev_mode + 1][mode + 1]) {
+				if (copy_lvb)
+					*copy_lvb = 1;
+			}
+		}
+
+		lkb->lkb_last_cast_cb_mode = mode;
+		lkb->lkb_last_cast_time = ktime_get();
+	}
+
+	lkb->lkb_last_cb_mode = mode;
+	lkb->lkb_last_cb_flags = flags;
+
+	return false;
+}
+
+int dlm_get_cb(struct dlm_lkb *lkb, uint32_t flags, int mode,
+	       int status, uint32_t sbflags,
+	       struct dlm_callback **cb)
+{
+	struct dlm_rsb *rsb = lkb->lkb_resource;
+	struct dlm_ls *ls = rsb->res_ls;
+
+	*cb = dlm_allocate_cb();
+	if (WARN_ON_ONCE(!*cb))
+		return -ENOMEM;
+
+	/* for tracing */
+	(*cb)->lkb_id = lkb->lkb_id;
+	(*cb)->ls_id = ls->ls_global_id;
+	memcpy((*cb)->res_name, rsb->res_name, rsb->res_length);
+	(*cb)->res_length = rsb->res_length;
+
+	(*cb)->flags = flags;
+	(*cb)->mode = mode;
+	(*cb)->sb_status = status;
+	(*cb)->sb_flags = (sbflags & 0x000000FF);
+	(*cb)->lkb_lksb = lkb->lkb_lksb;
+
+	return 0;
+}
+
+static int dlm_get_queue_cb(struct dlm_lkb *lkb, uint32_t flags, int mode,
+			    int status, uint32_t sbflags,
+			    struct dlm_callback **cb)
+{
+	int rv;
+
+	rv = dlm_get_cb(lkb, flags, mode, status, sbflags, cb);
+	if (rv)
+		return rv;
+
+	(*cb)->astfn = lkb->lkb_astfn;
+	(*cb)->bastfn = lkb->lkb_bastfn;
+	(*cb)->astparam = lkb->lkb_astparam;
+	INIT_WORK(&(*cb)->work, dlm_callback_work);
+
+	return 0;
+}
+
+void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
+		uint32_t sbflags)
+{
+	struct dlm_rsb *rsb = lkb->lkb_resource;
+	struct dlm_ls *ls = rsb->res_ls;
+	struct dlm_callback *cb;
+	int rv;
+
+	if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
+		dlm_user_add_ast(lkb, flags, mode, status, sbflags);
+		return;
+	}
+
+	if (dlm_may_skip_callback(lkb, flags, mode, status, sbflags, NULL))
+		return;
+
+	spin_lock_bh(&ls->ls_cb_lock);
+	if (test_bit(LSFL_CB_DELAY, &ls->ls_flags)) {
+		rv = dlm_get_queue_cb(lkb, flags, mode, status, sbflags, &cb);
+		if (!rv)
+			list_add(&cb->list, &ls->ls_cb_delay);
+	} else {
+		if (test_bit(LSFL_SOFTIRQ, &ls->ls_flags)) {
+			dlm_run_callback(ls->ls_global_id, lkb->lkb_id, mode, flags,
+					 sbflags, status, lkb->lkb_lksb,
+					 lkb->lkb_astfn, lkb->lkb_bastfn,
+					 lkb->lkb_astparam, rsb->res_name,
+					 rsb->res_length);
+		} else {
+			rv = dlm_get_queue_cb(lkb, flags, mode, status, sbflags, &cb);
+			if (!rv)
+				queue_work(ls->ls_callback_wq, &cb->work);
+		}
+	}
+	spin_unlock_bh(&ls->ls_cb_lock);
+}
+
+int dlm_callback_start(struct dlm_ls *ls)
+{
+	if (!test_bit(LSFL_FS, &ls->ls_flags) ||
+	    test_bit(LSFL_SOFTIRQ, &ls->ls_flags))
+		return 0;
+
+	ls->ls_callback_wq = alloc_ordered_workqueue("dlm_callback",
+						     WQ_HIGHPRI | WQ_MEM_RECLAIM);
+	if (!ls->ls_callback_wq) {
+		log_print("can't start dlm_callback workqueue");
+		return -ENOMEM;
+	}
+	return 0;
+}
+
+void dlm_callback_stop(struct dlm_ls *ls)
+{
+	if (ls->ls_callback_wq)
+		destroy_workqueue(ls->ls_callback_wq);
+}
+
+void dlm_callback_suspend(struct dlm_ls *ls)
+{
+	if (!test_bit(LSFL_FS, &ls->ls_flags))
+		return;
+
+	spin_lock_bh(&ls->ls_cb_lock);
+	set_bit(LSFL_CB_DELAY, &ls->ls_flags);
+	spin_unlock_bh(&ls->ls_cb_lock);
+
+	if (ls->ls_callback_wq)
+		flush_workqueue(ls->ls_callback_wq);
+}
+
+#define MAX_CB_QUEUE 25
+
+void dlm_callback_resume(struct dlm_ls *ls)
+{
+	struct dlm_callback *cb, *safe;
+	int count = 0, sum = 0;
+	bool empty;
+
+	if (!test_bit(LSFL_FS, &ls->ls_flags))
+		return;
+
+more:
+	spin_lock_bh(&ls->ls_cb_lock);
+	list_for_each_entry_safe(cb, safe, &ls->ls_cb_delay, list) {
+		list_del(&cb->list);
+		if (test_bit(LSFL_SOFTIRQ, &ls->ls_flags))
+			dlm_do_callback(cb);
+		else
+			queue_work(ls->ls_callback_wq, &cb->work);
+
+		count++;
+		if (count == MAX_CB_QUEUE)
+			break;
+	}
+	empty = list_empty(&ls->ls_cb_delay);
+	if (empty)
+		clear_bit(LSFL_CB_DELAY, &ls->ls_flags);
+	spin_unlock_bh(&ls->ls_cb_lock);
+
+	sum += count;
+	if (!empty) {
+		count = 0;
+		cond_resched();
+		goto more;
+	}
+
+	if (sum)
+		log_rinfo(ls, "%s %d", __func__, sum);
+}
+

linux-7.2.3/fs/dlm/ast.h

@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2010 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __ASTD_DOT_H__
+#define __ASTD_DOT_H__
+
+bool dlm_may_skip_callback(struct dlm_lkb *lkb, uint32_t flags, int mode,
+			   int status, uint32_t sbflags, int *copy_lvb);
+int dlm_get_cb(struct dlm_lkb *lkb, uint32_t flags, int mode,
+	       int status, uint32_t sbflags,
+	       struct dlm_callback **cb);
+void dlm_add_cb(struct dlm_lkb *lkb, uint32_t flags, int mode, int status,
+                uint32_t sbflags);
+
+int dlm_callback_start(struct dlm_ls *ls);
+void dlm_callback_stop(struct dlm_ls *ls);
+void dlm_callback_suspend(struct dlm_ls *ls);
+void dlm_callback_resume(struct dlm_ls *ls);
+
+#endif
+
+

linux-7.2.3/fs/dlm/config.h

@@ -0,0 +1,60 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2011 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __CONFIG_DOT_H__
+#define __CONFIG_DOT_H__
+
+#define DLM_MAX_SOCKET_BUFSIZE	4096
+
+struct dlm_config_node {
+	int nodeid;
+	int weight;
+	bool gone;
+	int new;
+	uint32_t comm_seq;
+	unsigned int release_recover;
+};
+
+extern const struct rhashtable_params dlm_rhash_rsb_params;
+
+#define DLM_MAX_ADDR_COUNT 8
+
+#define DLM_PROTO_TCP	0
+#define DLM_PROTO_SCTP	1
+
+struct dlm_config_info {
+	__be16 ci_tcp_port;
+	unsigned int ci_buffer_size;
+	unsigned int ci_rsbtbl_size;
+	unsigned int ci_recover_timer;
+	unsigned int ci_toss_secs;
+	unsigned int ci_scan_secs;
+	unsigned int ci_log_debug;
+	unsigned int ci_log_info;
+	unsigned int ci_protocol;
+	unsigned int ci_mark;
+	unsigned int ci_new_rsb_count;
+	unsigned int ci_recover_callbacks;
+	char ci_cluster_name[DLM_LOCKSPACE_LEN];
+};
+
+extern struct dlm_config_info dlm_config;
+
+int dlm_config_init(void);
+void dlm_config_exit(void);
+int dlm_config_nodes(char *lsname, struct dlm_config_node **nodes_out,
+		     int *count_out);
+int dlm_comm_seq(int nodeid, uint32_t *seq, bool locked);
+int dlm_our_nodeid(void);
+int dlm_our_addr(struct sockaddr_storage *addr, int num);
+
+#endif				/* __CONFIG_DOT_H__ */
+

linux-7.2.3/fs/dlm/debug_fs.c

@@ -0,0 +1,825 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2009 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include <linux/pagemap.h>
+#include <linux/seq_file.h>
+#include <linux/init.h>
+#include <linux/ctype.h>
+#include <linux/debugfs.h>
+#include <linux/slab.h>
+
+#include "dlm_internal.h"
+#include "midcomms.h"
+#include "lock.h"
+#include "ast.h"
+
+#define DLM_DEBUG_BUF_LEN 4096
+static char debug_buf[DLM_DEBUG_BUF_LEN];
+static struct mutex debug_buf_lock;
+
+static struct dentry *dlm_root;
+static struct dentry *dlm_comms;
+
+static char *print_lockmode(int mode)
+{
+	switch (mode) {
+	case DLM_LOCK_IV:
+		return "--";
+	case DLM_LOCK_NL:
+		return "NL";
+	case DLM_LOCK_CR:
+		return "CR";
+	case DLM_LOCK_CW:
+		return "CW";
+	case DLM_LOCK_PR:
+		return "PR";
+	case DLM_LOCK_PW:
+		return "PW";
+	case DLM_LOCK_EX:
+		return "EX";
+	default:
+		return "??";
+	}
+}
+
+static void print_format1_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			       struct dlm_rsb *res)
+{
+	seq_printf(s, "%08x %s", lkb->lkb_id, print_lockmode(lkb->lkb_grmode));
+
+	if (lkb->lkb_status == DLM_LKSTS_CONVERT ||
+	    lkb->lkb_status == DLM_LKSTS_WAITING)
+		seq_printf(s, " (%s)", print_lockmode(lkb->lkb_rqmode));
+
+	if (lkb->lkb_nodeid) {
+		if (lkb->lkb_nodeid != res->res_nodeid)
+			seq_printf(s, " Remote: %3d %08x", lkb->lkb_nodeid,
+				   lkb->lkb_remid);
+		else
+			seq_printf(s, " Master:     %08x", lkb->lkb_remid);
+	}
+
+	if (lkb->lkb_wait_type)
+		seq_printf(s, " wait_type: %d", lkb->lkb_wait_type);
+
+	seq_putc(s, '\n');
+}
+
+static void print_format1(struct dlm_rsb *res, struct seq_file *s)
+{
+	struct dlm_lkb *lkb;
+	int i, lvblen = res->res_ls->ls_lvblen, recover_list, root_list;
+
+	lock_rsb(res);
+
+	seq_printf(s, "\nResource %p Name (len=%d) \"", res, res->res_length);
+
+	for (i = 0; i < res->res_length; i++) {
+		if (isprint(res->res_name[i]))
+			seq_printf(s, "%c", res->res_name[i]);
+		else
+			seq_printf(s, "%c", '.');
+	}
+
+	if (res->res_nodeid > 0)
+		seq_printf(s, "\"\nLocal Copy, Master is node %d\n",
+			   res->res_nodeid);
+	else if (res->res_nodeid == 0)
+		seq_puts(s, "\"\nMaster Copy\n");
+	else if (res->res_nodeid == -1)
+		seq_printf(s, "\"\nLooking up master (lkid %x)\n",
+			   res->res_first_lkid);
+	else
+		seq_printf(s, "\"\nInvalid master %d\n", res->res_nodeid);
+	if (seq_has_overflowed(s))
+		goto out;
+
+	/* Print the LVB: */
+	if (res->res_lvbptr) {
+		seq_puts(s, "LVB: ");
+		for (i = 0; i < lvblen; i++) {
+			if (i == lvblen / 2)
+				seq_puts(s, "\n     ");
+			seq_printf(s, "%02x ",
+				   (unsigned char) res->res_lvbptr[i]);
+		}
+		if (rsb_flag(res, RSB_VALNOTVALID))
+			seq_puts(s, " (INVALID)");
+		seq_putc(s, '\n');
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	root_list = !list_empty(&res->res_root_list);
+	recover_list = !list_empty(&res->res_recover_list);
+
+	if (root_list || recover_list) {
+		seq_printf(s, "Recovery: root %d recover %d flags %lx count %d\n",
+			   root_list, recover_list,
+			   res->res_flags, res->res_recover_locks_count);
+	}
+
+	/* Print the locks attached to this resource */
+	seq_puts(s, "Granted Queue\n");
+	list_for_each_entry(lkb, &res->res_grantqueue, lkb_statequeue) {
+		print_format1_lock(s, lkb, res);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	seq_puts(s, "Conversion Queue\n");
+	list_for_each_entry(lkb, &res->res_convertqueue, lkb_statequeue) {
+		print_format1_lock(s, lkb, res);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	seq_puts(s, "Waiting Queue\n");
+	list_for_each_entry(lkb, &res->res_waitqueue, lkb_statequeue) {
+		print_format1_lock(s, lkb, res);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	if (list_empty(&res->res_lookup))
+		goto out;
+
+	seq_puts(s, "Lookup Queue\n");
+	list_for_each_entry(lkb, &res->res_lookup, lkb_rsb_lookup) {
+		seq_printf(s, "%08x %s",
+			   lkb->lkb_id, print_lockmode(lkb->lkb_rqmode));
+		if (lkb->lkb_wait_type)
+			seq_printf(s, " wait_type: %d", lkb->lkb_wait_type);
+		seq_putc(s, '\n');
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+ out:
+	unlock_rsb(res);
+}
+
+static void print_format2_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			       struct dlm_rsb *r)
+{
+	u64 xid = 0;
+	u64 us;
+
+	if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
+		if (lkb->lkb_ua)
+			xid = lkb->lkb_ua->xid;
+	}
+
+	/* microseconds since lkb was added to current queue */
+	us = ktime_to_us(ktime_sub(ktime_get(), lkb->lkb_timestamp));
+
+	/* id nodeid remid pid xid exflags flags sts grmode rqmode time_us
+	   r_nodeid r_len r_name */
+
+	seq_printf(s, "%x %d %x %u %llu %x %x %d %d %d %llu %u %d \"%s\"\n",
+		   lkb->lkb_id,
+		   lkb->lkb_nodeid,
+		   lkb->lkb_remid,
+		   lkb->lkb_ownpid,
+		   (unsigned long long)xid,
+		   lkb->lkb_exflags,
+		   dlm_iflags_val(lkb),
+		   lkb->lkb_status,
+		   lkb->lkb_grmode,
+		   lkb->lkb_rqmode,
+		   (unsigned long long)us,
+		   r->res_nodeid,
+		   r->res_length,
+		   r->res_name);
+}
+
+static void print_format2(struct dlm_rsb *r, struct seq_file *s)
+{
+	struct dlm_lkb *lkb;
+
+	lock_rsb(r);
+
+	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
+		print_format2_lock(s, lkb, r);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue) {
+		print_format2_lock(s, lkb, r);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue) {
+		print_format2_lock(s, lkb, r);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+ out:
+	unlock_rsb(r);
+}
+
+static void print_format3_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			      int rsb_lookup)
+{
+	u64 xid = 0;
+
+	if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
+		if (lkb->lkb_ua)
+			xid = lkb->lkb_ua->xid;
+	}
+
+	seq_printf(s, "lkb %x %d %x %u %llu %x %x %d %d %d %d %d %d %u %llu %llu\n",
+		   lkb->lkb_id,
+		   lkb->lkb_nodeid,
+		   lkb->lkb_remid,
+		   lkb->lkb_ownpid,
+		   (unsigned long long)xid,
+		   lkb->lkb_exflags,
+		   dlm_iflags_val(lkb),
+		   lkb->lkb_status,
+		   lkb->lkb_grmode,
+		   lkb->lkb_rqmode,
+		   lkb->lkb_last_bast_cb_mode,
+		   rsb_lookup,
+		   lkb->lkb_wait_type,
+		   lkb->lkb_lvbseq,
+		   (unsigned long long)ktime_to_ns(lkb->lkb_timestamp),
+		   (unsigned long long)ktime_to_ns(lkb->lkb_last_bast_time));
+}
+
+static void print_format3(struct dlm_rsb *r, struct seq_file *s)
+{
+	struct dlm_lkb *lkb;
+	int i, lvblen = r->res_ls->ls_lvblen;
+	int print_name = 1;
+
+	lock_rsb(r);
+
+	seq_printf(s, "rsb %p %d %x %lx %d %d %u %d ",
+		   r,
+		   r->res_nodeid,
+		   r->res_first_lkid,
+		   r->res_flags,
+		   !list_empty(&r->res_root_list),
+		   !list_empty(&r->res_recover_list),
+		   r->res_recover_locks_count,
+		   r->res_length);
+	if (seq_has_overflowed(s))
+		goto out;
+
+	for (i = 0; i < r->res_length; i++) {
+		if (!isascii(r->res_name[i]) || !isprint(r->res_name[i]))
+			print_name = 0;
+	}
+
+	seq_puts(s, print_name ? "str " : "hex");
+
+	for (i = 0; i < r->res_length; i++) {
+		if (print_name)
+			seq_printf(s, "%c", r->res_name[i]);
+		else
+			seq_printf(s, " %02x", (unsigned char)r->res_name[i]);
+	}
+	seq_putc(s, '\n');
+	if (seq_has_overflowed(s))
+		goto out;
+
+	if (!r->res_lvbptr)
+		goto do_locks;
+
+	seq_printf(s, "lvb %u %d", r->res_lvbseq, lvblen);
+
+	for (i = 0; i < lvblen; i++)
+		seq_printf(s, " %02x", (unsigned char)r->res_lvbptr[i]);
+	seq_putc(s, '\n');
+	if (seq_has_overflowed(s))
+		goto out;
+
+ do_locks:
+	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
+		print_format3_lock(s, lkb, 0);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue) {
+		print_format3_lock(s, lkb, 0);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue) {
+		print_format3_lock(s, lkb, 0);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+
+	list_for_each_entry(lkb, &r->res_lookup, lkb_rsb_lookup) {
+		print_format3_lock(s, lkb, 1);
+		if (seq_has_overflowed(s))
+			goto out;
+	}
+ out:
+	unlock_rsb(r);
+}
+
+static void print_format4(struct dlm_rsb *r, struct seq_file *s)
+{
+	int our_nodeid = dlm_our_nodeid();
+	int print_name = 1;
+	int i;
+
+	lock_rsb(r);
+
+	seq_printf(s, "rsb %p %d %d %d %d %lu %lx %d ",
+		   r,
+		   r->res_nodeid,
+		   r->res_master_nodeid,
+		   r->res_dir_nodeid,
+		   our_nodeid,
+		   r->res_toss_time,
+		   r->res_flags,
+		   r->res_length);
+
+	for (i = 0; i < r->res_length; i++) {
+		if (!isascii(r->res_name[i]) || !isprint(r->res_name[i]))
+			print_name = 0;
+	}
+
+	seq_puts(s, print_name ? "str " : "hex");
+
+	for (i = 0; i < r->res_length; i++) {
+		if (print_name)
+			seq_printf(s, "%c", r->res_name[i]);
+		else
+			seq_printf(s, " %02x", (unsigned char)r->res_name[i]);
+	}
+	seq_putc(s, '\n');
+	unlock_rsb(r);
+}
+
+static const struct seq_operations format1_seq_ops;
+static const struct seq_operations format2_seq_ops;
+static const struct seq_operations format3_seq_ops;
+static const struct seq_operations format4_seq_ops;
+
+/*
+ * If the buffer is full, seq_printf can be called again, but it
+ * does nothing.  So, the these printing routines periodically check
+ * seq_has_overflowed to avoid wasting too much time trying to print to
+ * a full buffer.
+ */
+
+static int table_seq_show(struct seq_file *seq, void *iter_ptr)
+{
+	struct dlm_rsb *rsb = list_entry(iter_ptr, struct dlm_rsb, res_slow_list);
+
+	if (seq->op == &format1_seq_ops)
+		print_format1(rsb, seq);
+	else if (seq->op == &format2_seq_ops)
+		print_format2(rsb, seq);
+	else if (seq->op == &format3_seq_ops)
+		print_format3(rsb, seq);
+	else if (seq->op == &format4_seq_ops)
+		print_format4(rsb, seq);
+
+	return 0;
+}
+
+static void *table_seq_start(struct seq_file *seq, loff_t *pos)
+{
+	struct dlm_ls *ls = seq->private;
+	struct list_head *list;
+
+	if (!*pos) {
+		if (seq->op == &format2_seq_ops)
+			seq_puts(seq, "id nodeid remid pid xid exflags flags sts grmode rqmode time_ms r_nodeid r_len r_name\n");
+		else if (seq->op == &format3_seq_ops)
+			seq_puts(seq, "rsb ptr nodeid first_lkid flags !root_list_empty !recover_list_empty recover_locks_count len\n");
+		else if (seq->op == &format4_seq_ops)
+			seq_puts(seq, "rsb ptr nodeid master_nodeid dir_nodeid our_nodeid toss_time flags len str|hex name\n");
+	}
+
+	if (seq->op == &format4_seq_ops)
+		list = &ls->ls_slow_inactive;
+	else
+		list = &ls->ls_slow_active;
+
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	return seq_list_start(list, *pos);
+}
+
+static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
+{
+	struct dlm_ls *ls = seq->private;
+	struct list_head *list;
+
+	if (seq->op == &format4_seq_ops)
+		list = &ls->ls_slow_inactive;
+	else
+		list = &ls->ls_slow_active;
+
+	return seq_list_next(iter_ptr, list, pos);
+}
+
+static void table_seq_stop(struct seq_file *seq, void *iter_ptr)
+{
+	struct dlm_ls *ls = seq->private;
+
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+}
+
+static const struct seq_operations format1_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
+};
+
+static const struct seq_operations format2_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
+};
+
+static const struct seq_operations format3_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
+};
+
+static const struct seq_operations format4_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
+};
+
+static const struct file_operations format1_fops;
+static const struct file_operations format2_fops;
+static const struct file_operations format3_fops;
+static const struct file_operations format4_fops;
+
+static int table_open1(struct inode *inode, struct file *file)
+{
+	struct seq_file *seq;
+	int ret;
+
+	ret = seq_open(file, &format1_seq_ops);
+	if (ret)
+		return ret;
+
+	seq = file->private_data;
+	seq->private = inode->i_private; /* the dlm_ls */
+	return 0;
+}
+
+static int table_open2(struct inode *inode, struct file *file)
+{
+	struct seq_file *seq;
+	int ret;
+
+	ret = seq_open(file, &format2_seq_ops);
+	if (ret)
+		return ret;
+
+	seq = file->private_data;
+	seq->private = inode->i_private; /* the dlm_ls */
+	return 0;
+}
+
+static ssize_t table_write2(struct file *file, const char __user *user_buf,
+			    size_t count, loff_t *ppos)
+{
+	struct seq_file *seq = file->private_data;
+	int n, len, lkb_nodeid, lkb_status, error;
+	char name[DLM_RESNAME_MAXLEN + 1] = {};
+	struct dlm_ls *ls = seq->private;
+	unsigned int lkb_flags;
+	char buf[256] = {};
+	uint32_t lkb_id;
+
+	if (copy_from_user(buf, user_buf,
+			   min_t(size_t, sizeof(buf) - 1, count)))
+		return -EFAULT;
+
+	n = sscanf(buf, "%x %" __stringify(DLM_RESNAME_MAXLEN) "s %x %d %d",
+		   &lkb_id, name, &lkb_flags, &lkb_nodeid, &lkb_status);
+	if (n != 5)
+		return -EINVAL;
+
+	len = strnlen(name, DLM_RESNAME_MAXLEN);
+	error = dlm_debug_add_lkb(ls, lkb_id, name, len, lkb_flags,
+				  lkb_nodeid, lkb_status);
+	if (error)
+		return error;
+
+	return count;
+}
+
+static int table_open3(struct inode *inode, struct file *file)
+{
+	struct seq_file *seq;
+	int ret;
+
+	ret = seq_open(file, &format3_seq_ops);
+	if (ret)
+		return ret;
+
+	seq = file->private_data;
+	seq->private = inode->i_private; /* the dlm_ls */
+	return 0;
+}
+
+static int table_open4(struct inode *inode, struct file *file)
+{
+	struct seq_file *seq;
+	int ret;
+
+	ret = seq_open(file, &format4_seq_ops);
+	if (ret)
+		return ret;
+
+	seq = file->private_data;
+	seq->private = inode->i_private; /* the dlm_ls */
+	return 0;
+}
+
+static const struct file_operations format1_fops = {
+	.owner   = THIS_MODULE,
+	.open    = table_open1,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+static const struct file_operations format2_fops = {
+	.owner   = THIS_MODULE,
+	.open    = table_open2,
+	.read    = seq_read,
+	.write   = table_write2,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+static const struct file_operations format3_fops = {
+	.owner   = THIS_MODULE,
+	.open    = table_open3,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+static const struct file_operations format4_fops = {
+	.owner   = THIS_MODULE,
+	.open    = table_open4,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+/*
+ * dump lkb's on the ls_waiters list
+ */
+static ssize_t waiters_read(struct file *file, char __user *userbuf,
+			    size_t count, loff_t *ppos)
+{
+	struct dlm_ls *ls = file->private_data;
+	struct dlm_lkb *lkb;
+	size_t len = DLM_DEBUG_BUF_LEN, pos = 0, ret, rv;
+
+	mutex_lock(&debug_buf_lock);
+	ret = dlm_lock_recovery_try(ls);
+	if (!ret) {
+		rv = -EAGAIN;
+		goto out;
+	}
+
+	spin_lock_bh(&ls->ls_waiters_lock);
+	memset(debug_buf, 0, sizeof(debug_buf));
+
+	list_for_each_entry(lkb, &ls->ls_waiters, lkb_wait_reply) {
+		ret = snprintf(debug_buf + pos, len - pos, "%x %d %d %s\n",
+			       lkb->lkb_id, lkb->lkb_wait_type,
+			       lkb->lkb_nodeid, lkb->lkb_resource->res_name);
+		if (ret >= len - pos)
+			break;
+		pos += ret;
+	}
+	spin_unlock_bh(&ls->ls_waiters_lock);
+	dlm_unlock_recovery(ls);
+
+	rv = simple_read_from_buffer(userbuf, count, ppos, debug_buf, pos);
+out:
+	mutex_unlock(&debug_buf_lock);
+	return rv;
+}
+
+static ssize_t waiters_write(struct file *file, const char __user *user_buf,
+			     size_t count, loff_t *ppos)
+{
+	struct dlm_ls *ls = file->private_data;
+	int mstype, to_nodeid;
+	char buf[128] = {};
+	uint32_t lkb_id;
+	int n, error;
+
+	if (copy_from_user(buf, user_buf,
+			   min_t(size_t, sizeof(buf) - 1, count)))
+		return -EFAULT;
+
+	n = sscanf(buf, "%x %d %d", &lkb_id, &mstype, &to_nodeid);
+	if (n != 3)
+		return -EINVAL;
+
+	error = dlm_lock_recovery_try(ls);
+	if (!error)
+		return -EAGAIN;
+
+	error = dlm_debug_add_lkb_to_waiters(ls, lkb_id, mstype, to_nodeid);
+	dlm_unlock_recovery(ls);
+	if (error)
+		return error;
+
+	return count;
+}
+
+static const struct file_operations waiters_fops = {
+	.owner   = THIS_MODULE,
+	.open    = simple_open,
+	.read    = waiters_read,
+	.write   = waiters_write,
+	.llseek  = default_llseek,
+};
+
+void dlm_delete_debug_file(struct dlm_ls *ls)
+{
+	debugfs_remove(ls->ls_debug_rsb_dentry);
+	debugfs_remove(ls->ls_debug_waiters_dentry);
+	debugfs_remove(ls->ls_debug_locks_dentry);
+	debugfs_remove(ls->ls_debug_all_dentry);
+	debugfs_remove(ls->ls_debug_toss_dentry);
+	debugfs_remove(ls->ls_debug_queued_asts_dentry);
+}
+
+static int dlm_state_show(struct seq_file *file, void *offset)
+{
+	seq_printf(file, "%s\n", dlm_midcomms_state(file->private));
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dlm_state);
+
+static int dlm_flags_show(struct seq_file *file, void *offset)
+{
+	seq_printf(file, "%lu\n", dlm_midcomms_flags(file->private));
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dlm_flags);
+
+static int dlm_send_queue_cnt_show(struct seq_file *file, void *offset)
+{
+	seq_printf(file, "%d\n", dlm_midcomms_send_queue_cnt(file->private));
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dlm_send_queue_cnt);
+
+static int dlm_version_show(struct seq_file *file, void *offset)
+{
+	seq_printf(file, "0x%08x\n", dlm_midcomms_version(file->private));
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dlm_version);
+
+static ssize_t dlm_rawmsg_write(struct file *fp, const char __user *user_buf,
+				size_t count, loff_t *ppos)
+{
+	void *buf;
+	int ret;
+
+	if (count > PAGE_SIZE || count < sizeof(struct dlm_header))
+		return -EINVAL;
+
+	buf = kmalloc(PAGE_SIZE, GFP_NOFS);
+	if (!buf)
+		return -ENOMEM;
+
+	if (copy_from_user(buf, user_buf, count)) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	ret = dlm_midcomms_rawmsg_send(fp->private_data, buf, count);
+	if (ret)
+		goto out;
+
+	kfree(buf);
+	return count;
+
+out:
+	kfree(buf);
+	return ret;
+}
+
+static const struct file_operations dlm_rawmsg_fops = {
+	.open	= simple_open,
+	.write	= dlm_rawmsg_write,
+};
+
+void *dlm_create_debug_comms_file(int nodeid, void *data)
+{
+	struct dentry *d_node;
+	char name[256];
+
+	memset(name, 0, sizeof(name));
+	snprintf(name, 256, "%d", nodeid);
+
+	d_node = debugfs_create_dir(name, dlm_comms);
+	debugfs_create_file("state", 0444, d_node, data, &dlm_state_fops);
+	debugfs_create_file("flags", 0444, d_node, data, &dlm_flags_fops);
+	debugfs_create_file("send_queue_count", 0444, d_node, data,
+			    &dlm_send_queue_cnt_fops);
+	debugfs_create_file("version", 0444, d_node, data, &dlm_version_fops);
+	debugfs_create_file("rawmsg", 0200, d_node, data, &dlm_rawmsg_fops);
+
+	return d_node;
+}
+
+void dlm_delete_debug_comms_file(void *ctx)
+{
+	debugfs_remove(ctx);
+}
+
+void dlm_create_debug_file(struct dlm_ls *ls)
+{
+	/* Reserve enough space for the longest file name */
+	char name[DLM_LOCKSPACE_LEN + sizeof("_queued_asts")];
+
+	/* format 1 */
+
+	ls->ls_debug_rsb_dentry = debugfs_create_file(ls->ls_name,
+						      S_IFREG | S_IRUGO,
+						      dlm_root,
+						      ls,
+						      &format1_fops);
+
+	/* format 2 */
+
+	snprintf(name, sizeof(name), "%s_locks", ls->ls_name);
+
+	ls->ls_debug_locks_dentry = debugfs_create_file(name,
+							0644,
+							dlm_root,
+							ls,
+							&format2_fops);
+
+	/* format 3 */
+
+	snprintf(name, sizeof(name), "%s_all", ls->ls_name);
+
+	ls->ls_debug_all_dentry = debugfs_create_file(name,
+						      S_IFREG | S_IRUGO,
+						      dlm_root,
+						      ls,
+						      &format3_fops);
+
+	/* format 4 */
+
+	snprintf(name, sizeof(name), "%s_toss", ls->ls_name);
+
+	ls->ls_debug_toss_dentry = debugfs_create_file(name,
+						       S_IFREG | S_IRUGO,
+						       dlm_root,
+						       ls,
+						       &format4_fops);
+
+	snprintf(name, sizeof(name), "%s_waiters", ls->ls_name);
+
+	ls->ls_debug_waiters_dentry = debugfs_create_file(name,
+							  0644,
+							  dlm_root,
+							  ls,
+							  &waiters_fops);
+}
+
+void __init dlm_register_debugfs(void)
+{
+	mutex_init(&debug_buf_lock);
+	dlm_root = debugfs_create_dir("dlm", NULL);
+	dlm_comms = debugfs_create_dir("comms", dlm_root);
+}
+
+void dlm_unregister_debugfs(void)
+{
+	debugfs_remove(dlm_root);
+}
+

linux-7.2.3/fs/dlm/dir.c

@@ -0,0 +1,400 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "member.h"
+#include "lowcomms.h"
+#include "rcom.h"
+#include "config.h"
+#include "memory.h"
+#include "recover.h"
+#include "util.h"
+#include "lock.h"
+#include "dir.h"
+
+/*
+ * We use the upper 16 bits of the hash value to select the directory node.
+ * Low bits are used for distribution of rsb's among hash buckets on each node.
+ *
+ * To give the exact range wanted (0 to num_nodes-1), we apply a modulus of
+ * num_nodes to the hash value.  This value in the desired range is used as an
+ * offset into the sorted list of nodeid's to give the particular nodeid.
+ */
+
+int dlm_hash2nodeid(struct dlm_ls *ls, uint32_t hash)
+{
+	uint32_t node;
+
+	if (ls->ls_num_nodes == 1)
+		return dlm_our_nodeid();
+	else {
+		node = (hash >> 16) % ls->ls_total_weight;
+		return ls->ls_node_array[node];
+	}
+}
+
+int dlm_dir_nodeid(struct dlm_rsb *r)
+{
+	return r->res_dir_nodeid;
+}
+
+void dlm_recover_dir_nodeid(struct dlm_ls *ls, const struct list_head *root_list)
+{
+	struct dlm_rsb *r;
+
+	list_for_each_entry(r, root_list, res_root_list) {
+		r->res_dir_nodeid = dlm_hash2nodeid(ls, r->res_hash);
+	}
+}
+
+int dlm_recover_directory(struct dlm_ls *ls, uint64_t seq)
+{
+	struct dlm_member *memb;
+	char *b, *last_name = NULL;
+	int error = -ENOMEM, last_len, nodeid, result;
+	uint16_t namelen;
+	unsigned int count = 0, count_bad = 0, count_add = 0;
+
+	log_rinfo(ls, "dlm_recover_directory");
+
+	if (dlm_no_directory(ls))
+		goto out_status;
+
+	last_name = kmalloc(DLM_RESNAME_MAXLEN, GFP_NOFS);
+	if (!last_name)
+		goto out;
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (memb->nodeid == dlm_our_nodeid())
+			continue;
+
+		memset(last_name, 0, DLM_RESNAME_MAXLEN);
+		last_len = 0;
+
+		for (;;) {
+			int left;
+			if (dlm_recovery_stopped(ls)) {
+				error = -EINTR;
+				goto out_free;
+			}
+
+			error = dlm_rcom_names(ls, memb->nodeid,
+					       last_name, last_len, seq);
+			if (error)
+				goto out_free;
+
+			cond_resched();
+
+			/*
+			 * pick namelen/name pairs out of received buffer
+			 */
+
+			b = ls->ls_recover_buf->rc_buf;
+			left = le16_to_cpu(ls->ls_recover_buf->rc_header.h_length);
+			left -= sizeof(struct dlm_rcom);
+
+			for (;;) {
+				__be16 v;
+
+				error = -EINVAL;
+				if (left < sizeof(__be16))
+					goto out_free;
+
+				memcpy(&v, b, sizeof(__be16));
+				namelen = be16_to_cpu(v);
+				b += sizeof(__be16);
+				left -= sizeof(__be16);
+
+				/* namelen of 0xFFFFF marks end of names for
+				   this node; namelen of 0 marks end of the
+				   buffer */
+
+				if (namelen == 0xFFFF)
+					goto done;
+				if (!namelen)
+					break;
+
+				if (namelen > left)
+					goto out_free;
+
+				if (namelen > DLM_RESNAME_MAXLEN)
+					goto out_free;
+
+				error = dlm_master_lookup(ls, memb->nodeid,
+							  b, namelen,
+							  DLM_LU_RECOVER_DIR,
+							  &nodeid, &result);
+				if (error) {
+					log_error(ls, "recover_dir lookup %d",
+						  error);
+					goto out_free;
+				}
+
+				/* The name was found in rsbtbl, but the
+				 * master nodeid is different from
+				 * memb->nodeid which says it is the master.
+				 * This should not happen. */
+
+				if (result == DLM_LU_MATCH &&
+				    nodeid != memb->nodeid) {
+					count_bad++;
+					log_error(ls, "recover_dir lookup %d "
+						  "nodeid %d memb %d bad %u",
+						  result, nodeid, memb->nodeid,
+						  count_bad);
+					print_hex_dump_bytes("dlm_recover_dir ",
+							     DUMP_PREFIX_NONE,
+							     b, namelen);
+				}
+
+				/* The name was found in rsbtbl, and the
+				 * master nodeid matches memb->nodeid.
+
+				if (result == DLM_LU_MATCH &&
+				    nodeid == memb->nodeid) {
+					count_match++;
+				}*/
+
+				/* The name was not found in rsbtbl and was
+				 * added with memb->nodeid as the master. */
+
+				if (result == DLM_LU_ADD) {
+					count_add++;
+				}
+
+				last_len = namelen;
+				memcpy(last_name, b, namelen);
+				b += namelen;
+				left -= namelen;
+				count++;
+			}
+		}
+	 done:
+		;
+	}
+
+ out_status:
+	error = 0;
+	dlm_set_recover_status(ls, DLM_RS_DIR);
+
+	log_rinfo(ls, "dlm_recover_directory %u in %u new",
+		  count, count_add);
+ out_free:
+	kfree(last_name);
+ out:
+	return error;
+}
+
+static struct dlm_rsb *find_rsb_root(struct dlm_ls *ls, const char *name,
+				     int len)
+{
+	struct dlm_rsb *r;
+	int rv;
+
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	rv = dlm_search_rsb_tree(&ls->ls_rsbtbl, name, len, &r);
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+	if (!rv)
+		return r;
+
+	list_for_each_entry(r, &ls->ls_masters_list, res_masters_list) {
+		if (len == r->res_length && !memcmp(name, r->res_name, len)) {
+			log_debug(ls, "find_rsb_root revert to root_list %s",
+				  r->res_name);
+			return r;
+		}
+	}
+	return NULL;
+}
+
+struct dlm_dir_dump {
+	/* init values to match if whole
+	 * dump fits to one seq. Sanity check only.
+	 */
+	uint64_t seq_init;
+	uint64_t nodeid_init;
+	/* compare local pointer with last lookup,
+	 * just a sanity check.
+	 */
+	struct list_head *last;
+
+	unsigned int sent_res; /* for log info */
+	unsigned int sent_msg; /* for log info */
+
+	struct list_head list;
+};
+
+static void drop_dir_ctx(struct dlm_ls *ls, int nodeid)
+{
+	struct dlm_dir_dump *dd, *safe;
+
+	write_lock_bh(&ls->ls_dir_dump_lock);
+	list_for_each_entry_safe(dd, safe, &ls->ls_dir_dump_list, list) {
+		if (dd->nodeid_init == nodeid) {
+			log_error(ls, "drop dump seq %llu",
+				 (unsigned long long)dd->seq_init);
+			list_del(&dd->list);
+			kfree(dd);
+		}
+	}
+	write_unlock_bh(&ls->ls_dir_dump_lock);
+}
+
+static struct dlm_dir_dump *lookup_dir_dump(struct dlm_ls *ls, int nodeid)
+{
+	struct dlm_dir_dump *iter, *dd = NULL;
+
+	read_lock_bh(&ls->ls_dir_dump_lock);
+	list_for_each_entry(iter, &ls->ls_dir_dump_list, list) {
+		if (iter->nodeid_init == nodeid) {
+			dd = iter;
+			break;
+		}
+	}
+	read_unlock_bh(&ls->ls_dir_dump_lock);
+
+	return dd;
+}
+
+static struct dlm_dir_dump *init_dir_dump(struct dlm_ls *ls, int nodeid)
+{
+	struct dlm_dir_dump *dd;
+
+	dd = lookup_dir_dump(ls, nodeid);
+	if (dd) {
+		log_error(ls, "found ongoing dir dump for node %d, will drop it",
+			  nodeid);
+		drop_dir_ctx(ls, nodeid);
+	}
+
+	dd = kzalloc_obj(*dd, GFP_ATOMIC);
+	if (!dd)
+		return NULL;
+
+	dd->seq_init = ls->ls_recover_seq;
+	dd->nodeid_init = nodeid;
+
+	write_lock_bh(&ls->ls_dir_dump_lock);
+	list_add(&dd->list, &ls->ls_dir_dump_list);
+	write_unlock_bh(&ls->ls_dir_dump_lock);
+
+	return dd;
+}
+
+/* Find the rsb where we left off (or start again), then send rsb names
+   for rsb's we're master of and whose directory node matches the requesting
+   node.  inbuf is the rsb name last sent, inlen is the name's length */
+
+void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
+ 			   char *outbuf, int outlen, int nodeid)
+{
+	struct list_head *list;
+	struct dlm_rsb *r;
+	int offset = 0, dir_nodeid;
+	struct dlm_dir_dump *dd;
+	__be16 be_namelen;
+
+	read_lock_bh(&ls->ls_masters_lock);
+
+	if (inlen > 1) {
+		dd = lookup_dir_dump(ls, nodeid);
+		if (!dd) {
+			log_error(ls, "failed to lookup dir dump context nodeid: %d",
+				  nodeid);
+			goto out;
+		}
+
+		/* next chunk in dump */
+		r = find_rsb_root(ls, inbuf, inlen);
+		if (!r) {
+			log_error(ls, "copy_master_names from %d start %d %.*s",
+				  nodeid, inlen, inlen, inbuf);
+			goto out;
+		}
+		list = r->res_masters_list.next;
+
+		/* sanity checks */
+		if (dd->last != &r->res_masters_list ||
+		    dd->seq_init != ls->ls_recover_seq) {
+			log_error(ls, "failed dir dump sanity check seq_init: %llu seq: %llu",
+				  (unsigned long long)dd->seq_init,
+				  (unsigned long long)ls->ls_recover_seq);
+			goto out;
+		}
+	} else {
+		dd = init_dir_dump(ls, nodeid);
+		if (!dd) {
+			log_error(ls, "failed to allocate dir dump context");
+			goto out;
+		}
+
+		/* start dump */
+		list = ls->ls_masters_list.next;
+		dd->last = list;
+	}
+
+	for (offset = 0; list != &ls->ls_masters_list; list = list->next) {
+		r = list_entry(list, struct dlm_rsb, res_masters_list);
+		dir_nodeid = dlm_dir_nodeid(r);
+		if (dir_nodeid != nodeid)
+			continue;
+
+		/*
+		 * The block ends when we can't fit the following in the
+		 * remaining buffer space:
+		 * namelen (uint16_t) +
+		 * name (r->res_length) +
+		 * end-of-block record 0x0000 (uint16_t)
+		 */
+
+		if (offset + sizeof(uint16_t)*2 + r->res_length > outlen) {
+			/* Write end-of-block record */
+			be_namelen = cpu_to_be16(0);
+			memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
+			offset += sizeof(__be16);
+			dd->sent_msg++;
+			goto out;
+		}
+
+		be_namelen = cpu_to_be16(r->res_length);
+		memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
+		offset += sizeof(__be16);
+		memcpy(outbuf + offset, r->res_name, r->res_length);
+		offset += r->res_length;
+		dd->sent_res++;
+		dd->last = list;
+	}
+
+	/*
+	 * If we've reached the end of the list (and there's room) write a
+	 * terminating record.
+	 */
+
+	if ((list == &ls->ls_masters_list) &&
+	    (offset + sizeof(uint16_t) <= outlen)) {
+		/* end dump */
+		be_namelen = cpu_to_be16(0xFFFF);
+		memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
+		offset += sizeof(__be16);
+		dd->sent_msg++;
+		log_rinfo(ls, "dlm_recover_directory nodeid %d sent %u res out %u messages",
+			  nodeid, dd->sent_res, dd->sent_msg);
+
+		write_lock_bh(&ls->ls_dir_dump_lock);
+		list_del_init(&dd->list);
+		write_unlock_bh(&ls->ls_dir_dump_lock);
+		kfree(dd);
+	}
+ out:
+	read_unlock_bh(&ls->ls_masters_lock);
+}
+

linux-7.2.3/fs/dlm/dir.h

@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __DIR_DOT_H__
+#define __DIR_DOT_H__
+
+int dlm_dir_nodeid(struct dlm_rsb *rsb);
+int dlm_hash2nodeid(struct dlm_ls *ls, uint32_t hash);
+void dlm_recover_dir_nodeid(struct dlm_ls *ls,
+			    const struct list_head *root_list);
+int dlm_recover_directory(struct dlm_ls *ls, uint64_t seq);
+void dlm_copy_master_names(struct dlm_ls *ls, const char *inbuf, int inlen,
+			   char *outbuf, int outlen, int nodeid);
+
+#endif				/* __DIR_DOT_H__ */
+

linux-7.2.3/fs/dlm/dlm_internal.h

@@ -0,0 +1,833 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2011 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __DLM_INTERNAL_DOT_H__
+#define __DLM_INTERNAL_DOT_H__
+
+/*
+ * This is the main header file to be included in each DLM source file.
+ */
+
+#include <uapi/linux/dlm_device.h>
+#include <linux/slab.h>
+#include <linux/sched.h>
+#include <linux/types.h>
+#include <linux/ctype.h>
+#include <linux/spinlock.h>
+#include <linux/vmalloc.h>
+#include <linux/list.h>
+#include <linux/errno.h>
+#include <linux/random.h>
+#include <linux/delay.h>
+#include <linux/socket.h>
+#include <linux/kthread.h>
+#include <linux/kobject.h>
+#include <linux/kref.h>
+#include <linux/kernel.h>
+#include <linux/jhash.h>
+#include <linux/miscdevice.h>
+#include <linux/rhashtable.h>
+#include <linux/mutex.h>
+#include <linux/xarray.h>
+#include <linux/ratelimit.h>
+#include <linux/uaccess.h>
+
+#include <linux/dlm.h>
+#include "config.h"
+
+struct dlm_ls;
+struct dlm_lkb;
+struct dlm_rsb;
+struct dlm_member;
+struct dlm_rsbtable;
+struct dlm_recover;
+struct dlm_header;
+struct dlm_message;
+struct dlm_rcom;
+struct dlm_mhandle;
+struct dlm_msg;
+
+#define log_print(fmt, args...) \
+	printk(KERN_ERR "dlm: "fmt"\n" , ##args)
+#define log_print_ratelimited(fmt, args...) \
+	printk_ratelimited(KERN_ERR "dlm: "fmt"\n", ##args)
+#define log_error(ls, fmt, args...) \
+	printk(KERN_ERR "dlm: %s: " fmt "\n", (ls)->ls_name , ##args)
+
+#define log_rinfo(ls, fmt, args...) \
+do { \
+	if (dlm_config.ci_log_info) \
+		printk(KERN_INFO "dlm: %s: " fmt "\n", \
+			(ls)->ls_name, ##args); \
+	else if (dlm_config.ci_log_debug) \
+		printk(KERN_DEBUG "dlm: %s: " fmt "\n", \
+		       (ls)->ls_name , ##args); \
+} while (0)
+
+#define log_debug(ls, fmt, args...) \
+do { \
+	if (dlm_config.ci_log_debug) \
+		printk(KERN_DEBUG "dlm: %s: " fmt "\n", \
+		       (ls)->ls_name , ##args); \
+} while (0)
+
+#define log_limit(ls, fmt, args...) \
+do { \
+	if (dlm_config.ci_log_debug) \
+		printk_ratelimited(KERN_DEBUG "dlm: %s: " fmt "\n", \
+			(ls)->ls_name , ##args); \
+} while (0)
+
+#define DLM_ASSERT(x, do) \
+{ \
+  if (!(x)) \
+  { \
+    printk(KERN_ERR "\nDLM:  Assertion failed on line %d of file %s\n" \
+               "DLM:  assertion:  \"%s\"\n" \
+               "DLM:  time = %lu\n", \
+               __LINE__, __FILE__, #x, jiffies); \
+    {do} \
+    printk("\n"); \
+    panic("DLM:  Record message above and reboot.\n"); \
+  } \
+}
+
+/*
+ * Lockspace member (per node in a ls)
+ */
+
+struct dlm_member {
+	struct list_head	list;
+	int			nodeid;
+	int			weight;
+	int			slot;
+	int			slot_prev;
+	int			comm_seq;
+	uint32_t		generation;
+};
+
+/*
+ * Save and manage recovery state for a lockspace.
+ */
+
+struct dlm_recover {
+	struct list_head	list;
+	struct dlm_config_node	*nodes;
+	int			nodes_count;
+	uint64_t		seq;
+};
+
+/*
+ * Pass input args to second stage locking function.
+ */
+
+struct dlm_args {
+	uint32_t		flags;
+	void			(*astfn) (void *astparam);
+	void			*astparam;
+	void			(*bastfn) (void *astparam, int mode);
+	int			mode;
+	struct dlm_lksb		*lksb;
+};
+
+
+/*
+ * Lock block
+ *
+ * A lock can be one of three types:
+ *
+ * local copy      lock is mastered locally
+ *                 (lkb_nodeid is zero and DLM_LKF_MSTCPY is not set)
+ * process copy    lock is mastered on a remote node
+ *                 (lkb_nodeid is non-zero and DLM_LKF_MSTCPY is not set)
+ * master copy     master node's copy of a lock owned by remote node
+ *                 (lkb_nodeid is non-zero and DLM_LKF_MSTCPY is set)
+ *
+ * lkb_exflags: a copy of the most recent flags arg provided to dlm_lock or
+ * dlm_unlock.  The dlm does not modify these or use any private flags in
+ * this field; it only contains DLM_LKF_ flags from dlm.h.  These flags
+ * are sent as-is to the remote master when the lock is remote.
+ *
+ * lkb_flags: internal dlm flags (DLM_IFL_ prefix) from dlm_internal.h.
+ * Some internal flags are shared between the master and process nodes;
+ * these shared flags are kept in the lower two bytes.  One of these
+ * flags set on the master copy will be propagated to the process copy
+ * and v.v.  Other internal flags are private to the master or process
+ * node (e.g. DLM_IFL_MSTCPY).  These are kept in the high two bytes.
+ *
+ * lkb_sbflags: status block flags.  These flags are copied directly into
+ * the caller's lksb.sb_flags prior to the dlm_lock/dlm_unlock completion
+ * ast.  All defined in dlm.h with DLM_SBF_ prefix.
+ *
+ * lkb_status: the lock status indicates which rsb queue the lock is
+ * on, grant, convert, or wait.  DLM_LKSTS_ WAITING/GRANTED/CONVERT
+ *
+ * lkb_wait_type: the dlm message type (DLM_MSG_ prefix) for which a
+ * reply is needed.  Only set when the lkb is on the lockspace waiters
+ * list awaiting a reply from a remote node.
+ *
+ * lkb_nodeid: when the lkb is a local copy, nodeid is 0; when the lkb
+ * is a master copy, nodeid specifies the remote lock holder, when the
+ * lkb is a process copy, the nodeid specifies the lock master.
+ */
+
+/* lkb_status */
+
+#define DLM_LKSTS_WAITING	1
+#define DLM_LKSTS_GRANTED	2
+#define DLM_LKSTS_CONVERT	3
+
+/* lkb_iflags */
+
+#define DLM_IFL_MSTCPY_BIT	16
+#define __DLM_IFL_MIN_BIT	DLM_IFL_MSTCPY_BIT
+#define DLM_IFL_RESEND_BIT	17
+#define DLM_IFL_DEAD_BIT	18
+#define DLM_IFL_OVERLAP_UNLOCK_BIT 19
+#define DLM_IFL_OVERLAP_CANCEL_BIT 20
+#define DLM_IFL_ENDOFLIFE_BIT	21
+#define DLM_IFL_DEADLOCK_CANCEL_BIT 24
+#define __DLM_IFL_MAX_BIT	DLM_IFL_DEADLOCK_CANCEL_BIT
+
+/* lkb_dflags */
+
+#define DLM_DFL_USER_BIT	0
+#define __DLM_DFL_MIN_BIT	DLM_DFL_USER_BIT
+#define DLM_DFL_ORPHAN_BIT	1
+#define __DLM_DFL_MAX_BIT	DLM_DFL_ORPHAN_BIT
+
+#define DLM_CB_CAST		0x00000001
+#define DLM_CB_BAST		0x00000002
+
+/* much of this is just saving user space pointers associated with the
+ * lock that we pass back to the user lib with an ast
+ */
+
+struct dlm_user_args {
+	struct dlm_user_proc	*proc; /* each process that opens the lockspace
+					* device has private data
+					* (dlm_user_proc) on the struct file,
+					* the process's locks point back to it
+					*/
+	struct dlm_lksb		lksb;
+	struct dlm_lksb __user	*user_lksb;
+	void __user		*castparam;
+	void __user		*castaddr;
+	void __user		*bastparam;
+	void __user		*bastaddr;
+	uint64_t		xid;
+};
+
+struct dlm_callback {
+	uint32_t		flags;		/* DLM_CBF_ */
+	int			sb_status;	/* copy to lksb status */
+	uint8_t			sb_flags;	/* copy to lksb flags */
+	int8_t			mode; /* rq mode of bast, gr mode of cast */
+	bool			copy_lvb;
+	struct dlm_lksb		*lkb_lksb;
+	unsigned char		lvbptr[DLM_USER_LVB_LEN];
+
+	union {
+		void			*astparam;	/* caller's ast arg */
+		struct dlm_user_args	ua;
+	};
+	struct work_struct	work;
+	void			(*bastfn)(void *astparam, int mode);
+	void			(*astfn)(void *astparam);
+	char			res_name[DLM_RESNAME_MAXLEN];
+	size_t			res_length;
+	uint32_t		ls_id;
+	uint32_t		lkb_id;
+
+	struct list_head	list;
+};
+
+struct dlm_lkb {
+	struct dlm_rsb		*lkb_resource;	/* the rsb */
+	struct kref		lkb_ref;
+	int			lkb_nodeid;	/* copied from rsb */
+	int			lkb_ownpid;	/* pid of lock owner */
+	uint32_t		lkb_id;		/* our lock ID */
+	uint32_t		lkb_remid;	/* lock ID on remote partner */
+	uint32_t		lkb_exflags;	/* external flags from caller */
+	unsigned long		lkb_sbflags;	/* lksb flags */
+	unsigned long		lkb_dflags;	/* distributed flags */
+	unsigned long		lkb_iflags;	/* internal flags */
+	uint32_t		lkb_lvbseq;	/* lvb sequence number */
+
+	int8_t			lkb_status;     /* granted, waiting, convert */
+	int8_t			lkb_rqmode;	/* requested lock mode */
+	int8_t			lkb_grmode;	/* granted lock mode */
+	int8_t			lkb_highbast;	/* highest mode bast sent for */
+
+	int8_t			lkb_wait_type;	/* type of reply waiting for */
+	int8_t			lkb_wait_count;
+	int			lkb_wait_nodeid; /* for debugging */
+
+	struct list_head	lkb_statequeue;	/* rsb g/c/w list */
+	struct list_head	lkb_rsb_lookup;	/* waiting for rsb lookup */
+	struct list_head	lkb_wait_reply;	/* waiting for remote reply */
+	struct list_head	lkb_ownqueue;	/* list of locks for a process */
+	ktime_t			lkb_timestamp;
+
+	int8_t			lkb_last_cast_cb_mode;
+	int8_t			lkb_last_bast_cb_mode;
+	int8_t			lkb_last_cb_mode;
+	uint8_t			lkb_last_cb_flags;
+	ktime_t			lkb_last_cast_time;	/* for debugging */
+	ktime_t			lkb_last_bast_time;	/* for debugging */
+
+	uint64_t		lkb_recover_seq; /* from ls_recover_seq */
+
+	char			*lkb_lvbptr;
+	struct dlm_lksb		*lkb_lksb;      /* caller's status block */
+	void			(*lkb_astfn) (void *astparam);
+	void			(*lkb_bastfn) (void *astparam, int mode);
+	union {
+		void			*lkb_astparam;	/* caller's ast arg */
+		struct dlm_user_args	*lkb_ua;
+	};
+	struct rcu_head		rcu;
+};
+
+/*
+ * res_master_nodeid is "normal": 0 is unset/invalid, non-zero is the real
+ * nodeid, even when nodeid is our_nodeid.
+ *
+ * res_nodeid is "odd": -1 is unset/invalid, zero means our_nodeid,
+ * greater than zero when another nodeid.
+ *
+ * (TODO: remove res_nodeid and only use res_master_nodeid)
+ */
+
+struct dlm_rsb {
+	struct dlm_ls		*res_ls;	/* the lockspace */
+	struct kref		res_ref;
+	spinlock_t		res_lock;
+	unsigned long		res_flags;
+	int			res_length;	/* length of rsb name */
+	int			res_nodeid;
+	int			res_master_nodeid;
+	int			res_dir_nodeid;
+	unsigned long		res_id;		/* for ls_recover_xa */
+	uint32_t                res_lvbseq;
+	uint32_t		res_hash;
+	unsigned long		res_toss_time;
+	uint32_t		res_first_lkid;
+	struct list_head	res_lookup;	/* lkbs waiting on first */
+	struct rhash_head	res_node;	/* rsbtbl */
+	struct list_head	res_grantqueue;
+	struct list_head	res_convertqueue;
+	struct list_head	res_waitqueue;
+
+	struct list_head	res_slow_list;      /* ls_slow_* */
+	struct list_head	res_scan_list;
+	struct list_head	res_root_list;	    /* used for recovery */
+	struct list_head	res_masters_list;   /* used for recovery */
+	struct list_head	res_recover_list;   /* used for recovery */
+	int			res_recover_locks_count;
+	struct rcu_head		rcu;
+
+	char			*res_lvbptr;
+	char			res_name[DLM_RESNAME_MAXLEN+1];
+};
+
+/* dlm_master_lookup() flags */
+
+#define DLM_LU_RECOVER_DIR	1
+#define DLM_LU_RECOVER_MASTER	2
+
+/* dlm_master_lookup() results */
+
+#define DLM_LU_MATCH		1
+#define DLM_LU_ADD		2
+
+/* find_rsb() flags */
+
+#define R_REQUEST		0x00000001
+#define R_RECEIVE_REQUEST	0x00000002
+#define R_RECEIVE_RECOVER	0x00000004
+
+/* rsb_flags */
+
+enum rsb_flags {
+	RSB_MASTER_UNCERTAIN,
+	RSB_VALNOTVALID,
+	RSB_VALNOTVALID_PREV,
+	RSB_NEW_MASTER,
+	RSB_NEW_MASTER2,
+	RSB_RECOVER_CONVERT,
+	RSB_RECOVER_GRANT,
+	RSB_RECOVER_LVB_INVAL,
+	RSB_INACTIVE,
+	RSB_HASHED, /* set while rsb is on ls_rsbtbl */
+};
+
+static inline void rsb_set_flag(struct dlm_rsb *r, enum rsb_flags flag)
+{
+	__set_bit(flag, &r->res_flags);
+}
+
+static inline void rsb_clear_flag(struct dlm_rsb *r, enum rsb_flags flag)
+{
+	__clear_bit(flag, &r->res_flags);
+}
+
+static inline int rsb_flag(struct dlm_rsb *r, enum rsb_flags flag)
+{
+	return test_bit(flag, &r->res_flags);
+}
+
+
+/* dlm_header is first element of all structs sent between nodes */
+
+#define DLM_HEADER_MAJOR	0x00030000
+#define DLM_HEADER_MINOR	0x00000002
+
+#define DLM_VERSION_3_1		0x00030001
+#define DLM_VERSION_3_2		0x00030002
+
+#define DLM_HEADER_SLOTS	0x00000001
+
+#define DLM_MSG			1
+#define DLM_RCOM		2
+#define DLM_OPTS		3
+#define DLM_ACK			4
+#define DLM_FIN			5
+
+struct dlm_header {
+	__le32			h_version;
+	union {
+		/* for DLM_MSG and DLM_RCOM */
+		__le32		h_lockspace;
+		/* for DLM_ACK and DLM_OPTS */
+		__le32		h_seq;
+	} u;
+	__le32			h_nodeid;	/* nodeid of sender */
+	__le16			h_length;
+	uint8_t			h_cmd;		/* DLM_MSG, DLM_RCOM */
+	uint8_t			h_pad;
+};
+
+#define DLM_MSG_REQUEST		1
+#define DLM_MSG_CONVERT		2
+#define DLM_MSG_UNLOCK		3
+#define DLM_MSG_CANCEL		4
+#define DLM_MSG_REQUEST_REPLY	5
+#define DLM_MSG_CONVERT_REPLY	6
+#define DLM_MSG_UNLOCK_REPLY	7
+#define DLM_MSG_CANCEL_REPLY	8
+#define DLM_MSG_GRANT		9
+#define DLM_MSG_BAST		10
+#define DLM_MSG_LOOKUP		11
+#define DLM_MSG_REMOVE		12
+#define DLM_MSG_LOOKUP_REPLY	13
+#define DLM_MSG_PURGE		14
+
+struct dlm_message {
+	struct dlm_header	m_header;
+	__le32			m_type;		/* DLM_MSG_ */
+	__le32			m_nodeid;
+	__le32			m_pid;
+	__le32			m_lkid;		/* lkid on sender */
+	__le32			m_remid;	/* lkid on receiver */
+	__le32			m_parent_lkid;
+	__le32			m_parent_remid;
+	__le32			m_exflags;
+	__le32			m_sbflags;
+	__le32			m_flags;
+	__le32			m_lvbseq;
+	__le32			m_hash;
+	__le32			m_status;
+	__le32			m_grmode;
+	__le32			m_rqmode;
+	__le32			m_bastmode;
+	__le32			m_asts;
+	__le32			m_result;	/* 0 or -EXXX */
+	char			m_extra[];	/* name or lvb */
+};
+
+
+#define DLM_RS_NODES		0x00000001
+#define DLM_RS_NODES_ALL	0x00000002
+#define DLM_RS_DIR		0x00000004
+#define DLM_RS_DIR_ALL		0x00000008
+#define DLM_RS_LOCKS		0x00000010
+#define DLM_RS_LOCKS_ALL	0x00000020
+#define DLM_RS_DONE		0x00000040
+#define DLM_RS_DONE_ALL		0x00000080
+
+#define DLM_RCOM_STATUS		1
+#define DLM_RCOM_NAMES		2
+#define DLM_RCOM_LOOKUP		3
+#define DLM_RCOM_LOCK		4
+#define DLM_RCOM_STATUS_REPLY	5
+#define DLM_RCOM_NAMES_REPLY	6
+#define DLM_RCOM_LOOKUP_REPLY	7
+#define DLM_RCOM_LOCK_REPLY	8
+
+struct dlm_rcom {
+	struct dlm_header	rc_header;
+	__le32			rc_type;	/* DLM_RCOM_ */
+	__le32			rc_result;	/* multi-purpose */
+	__le64			rc_id;		/* match reply with request */
+	__le64			rc_seq;		/* sender's ls_recover_seq */
+	__le64			rc_seq_reply;	/* remote ls_recover_seq */
+	char			rc_buf[];
+};
+
+struct dlm_opt_header {
+	__le16		t_type;
+	__le16		t_length;
+	__le32		t_pad;
+	/* need to be 8 byte aligned */
+	char		t_value[];
+};
+
+/* encapsulation header */
+struct dlm_opts {
+	struct dlm_header	o_header;
+	uint8_t			o_nextcmd;
+	uint8_t			o_pad;
+	__le16			o_optlen;
+	__le32			o_pad2;
+	char			o_opts[];
+};
+
+union dlm_packet {
+	struct dlm_header	header;		/* common to other two */
+	struct dlm_message	message;
+	struct dlm_rcom		rcom;
+	struct dlm_opts		opts;
+};
+
+#define DLM_RSF_NEED_SLOTS	0x00000001
+
+/* RCOM_STATUS data */
+struct rcom_status {
+	__le32			rs_flags;
+	__le32			rs_unused1;
+	__le64			rs_unused2;
+};
+
+/* RCOM_STATUS_REPLY data */
+struct rcom_config {
+	__le32			rf_lvblen;
+	__le32			rf_lsflags;
+
+	/* DLM_HEADER_SLOTS adds: */
+	__le32			rf_flags;
+	__le16			rf_our_slot;
+	__le16			rf_num_slots;
+	__le32			rf_generation;
+	__le32			rf_unused1;
+	__le64			rf_unused2;
+};
+
+struct rcom_slot {
+	__le32			ro_nodeid;
+	__le16			ro_slot;
+	__le16			ro_unused1;
+	__le64			ro_unused2;
+};
+
+struct rcom_lock {
+	__le32			rl_ownpid;
+	__le32			rl_lkid;
+	__le32			rl_remid;
+	__le32			rl_parent_lkid;
+	__le32			rl_parent_remid;
+	__le32			rl_exflags;
+	__le32			rl_flags;
+	__le32			rl_lvbseq;
+	__le32			rl_result;
+	int8_t			rl_rqmode;
+	int8_t			rl_grmode;
+	int8_t			rl_status;
+	int8_t			rl_asts;
+	__le16			rl_wait_type;
+	__le16			rl_namelen;
+	char			rl_name[DLM_RESNAME_MAXLEN];
+	char			rl_lvb[];
+};
+
+struct dlm_ls {
+	struct list_head	ls_list;	/* list of lockspaces */
+	uint32_t		ls_global_id;	/* global unique lockspace ID */
+	uint32_t		ls_generation;
+	uint32_t		ls_exflags;
+	int			ls_lvblen;
+	atomic_t		ls_count;	/* refcount of processes in
+						   the dlm using this ls */
+	wait_queue_head_t	ls_count_wait;
+	int			ls_create_count; /* create/release refcount */
+	unsigned long		ls_flags;	/* LSFL_ */
+	struct kobject		ls_kobj;
+
+	struct xarray		ls_lkbxa;
+	rwlock_t		ls_lkbxa_lock;
+
+	/* an rsb is on rsbtl for primary locking functions,
+	   and on a slow list for recovery/dump iteration  */
+	struct rhashtable	ls_rsbtbl;
+	rwlock_t		ls_rsbtbl_lock; /* for ls_rsbtbl and ls_slow */
+	struct list_head	ls_slow_inactive; /* to iterate rsbtbl */
+	struct list_head	ls_slow_active;   /* to iterate rsbtbl */
+
+	struct timer_list	ls_scan_timer; /* based on first scan_list rsb toss_time */
+	struct list_head	ls_scan_list;  /* rsbs ordered by res_toss_time */
+	spinlock_t		ls_scan_lock;
+
+	spinlock_t		ls_waiters_lock;
+	struct list_head	ls_waiters;	/* lkbs needing a reply */
+
+	spinlock_t		ls_orphans_lock;
+	struct list_head	ls_orphans;
+
+	struct list_head	ls_nodes;	/* current nodes in ls */
+	struct list_head	ls_nodes_gone;	/* dead node list, recovery */
+	int			ls_num_nodes;	/* number of nodes in ls */
+	int			ls_low_nodeid;
+	int			ls_total_weight;
+	int			*ls_node_array;
+
+	int			ls_slot;
+	int			ls_num_slots;
+	int			ls_slots_size;
+	struct dlm_slot		*ls_slots;
+
+	struct dlm_rsb		ls_local_rsb;	/* for returning errors */
+	struct dlm_lkb		ls_local_lkb;	/* for returning errors */
+
+	struct dentry		*ls_debug_rsb_dentry; /* debugfs */
+	struct dentry		*ls_debug_waiters_dentry; /* debugfs */
+	struct dentry		*ls_debug_locks_dentry; /* debugfs */
+	struct dentry		*ls_debug_all_dentry; /* debugfs */
+	struct dentry		*ls_debug_toss_dentry; /* debugfs */
+	struct dentry		*ls_debug_queued_asts_dentry; /* debugfs */
+
+	wait_queue_head_t	ls_uevent_wait;	/* user part of join/leave */
+	int			ls_uevent_result;
+	struct completion	ls_recovery_done;
+	int			ls_recovery_result;
+
+	struct miscdevice       ls_device;
+
+	struct workqueue_struct	*ls_callback_wq;
+
+	/* recovery related */
+
+	spinlock_t		ls_cb_lock;
+	struct list_head	ls_cb_delay; /* save for queue_work later */
+	struct task_struct	*ls_recoverd_task;
+	struct mutex		ls_recoverd_active;
+	spinlock_t		ls_recover_lock;
+	unsigned long		ls_recover_begin; /* jiffies timestamp */
+	uint32_t		ls_recover_status; /* DLM_RS_ */
+	uint64_t		ls_recover_seq;
+	struct dlm_recover	*ls_recover_args;
+	struct rw_semaphore	ls_in_recovery;	/* block local requests */
+	rwlock_t		ls_recv_active;	/* block dlm_recv */
+	struct list_head	ls_requestqueue;/* queue remote requests */
+	rwlock_t		ls_requestqueue_lock;
+	struct dlm_rcom		*ls_recover_buf;
+	int			ls_recover_nodeid; /* for debugging */
+	unsigned int		ls_recover_locks_in; /* for log info */
+	uint64_t		ls_rcom_seq;
+	spinlock_t		ls_rcom_spin;
+	struct list_head	ls_recover_list;
+	spinlock_t		ls_recover_list_lock;
+	int			ls_recover_list_count;
+	struct xarray		ls_recover_xa;
+	spinlock_t		ls_recover_xa_lock;
+	wait_queue_head_t	ls_wait_general;
+	wait_queue_head_t	ls_recover_lock_wait;
+	spinlock_t		ls_clear_proc_locks;
+
+	struct list_head	ls_masters_list; /* root resources */
+	rwlock_t		ls_masters_lock; /* protect root_list */
+	struct list_head	ls_dir_dump_list; /* root resources */
+	rwlock_t		ls_dir_dump_lock; /* protect root_list */
+
+	const struct dlm_lockspace_ops *ls_ops;
+	void			*ls_ops_arg;
+
+	struct work_struct	ls_free_work;
+
+	int			ls_namelen;
+	char			ls_name[DLM_LOCKSPACE_LEN + 1];
+
+	/* Must be last --ends in a flexible-array member.*/
+	struct dlm_message	ls_local_ms;	/* for faking a reply */
+};
+
+/*
+ * LSFL_RECOVER_STOP - dlm_ls_stop() sets this to tell dlm recovery routines
+ * that they should abort what they're doing so new recovery can be started.
+ *
+ * LSFL_RECOVER_DOWN - dlm_ls_stop() sets this to tell dlm_recoverd that it
+ * should do down_write() on the in_recovery rw_semaphore. (doing down_write
+ * within dlm_ls_stop causes complaints about the lock acquired/released
+ * in different contexts.)
+ *
+ * LSFL_RECOVER_LOCK - dlm_recoverd holds the in_recovery rw_semaphore.
+ * It sets this after it is done with down_write() on the in_recovery
+ * rw_semaphore and clears it after it has released the rw_semaphore.
+ *
+ * LSFL_RECOVER_WORK - dlm_ls_start() sets this to tell dlm_recoverd that it
+ * should begin recovery of the lockspace.
+ *
+ * LSFL_RUNNING - set when normal locking activity is enabled.
+ * dlm_ls_stop() clears this to tell dlm locking routines that they should
+ * quit what they are doing so recovery can run.  dlm_recoverd sets
+ * this after recovery is finished.
+ */
+
+#define LSFL_RECOVER_STOP	0
+#define LSFL_RECOVER_DOWN	1
+#define LSFL_RECOVER_LOCK	2
+#define LSFL_RECOVER_WORK	3
+#define LSFL_RUNNING		4
+
+#define LSFL_RCOM_READY		5
+#define LSFL_RCOM_WAIT		6
+#define LSFL_UEVENT_WAIT	7
+#define LSFL_CB_DELAY		9
+#define LSFL_NODIR		10
+#define LSFL_RECV_MSG_BLOCKED	11
+#define LSFL_FS			12
+#define LSFL_SOFTIRQ		13
+
+#define DLM_PROC_FLAGS_CLOSING 1
+#define DLM_PROC_FLAGS_COMPAT  2
+
+/* locks list is kept so we can remove all a process's locks when it
+   exits (or orphan those that are persistent) */
+
+struct dlm_user_proc {
+	dlm_lockspace_t		*lockspace;
+	unsigned long		flags; /* DLM_PROC_FLAGS */
+	struct list_head	asts;
+	spinlock_t		asts_spin;
+	struct list_head	locks;
+	spinlock_t		locks_spin;
+	struct list_head	unlocking;
+	wait_queue_head_t	wait;
+};
+
+static inline int dlm_locking_stopped(struct dlm_ls *ls)
+{
+	return !test_bit(LSFL_RUNNING, &ls->ls_flags);
+}
+
+static inline int dlm_recovery_stopped(struct dlm_ls *ls)
+{
+	return test_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
+}
+
+static inline int dlm_no_directory(struct dlm_ls *ls)
+{
+	return test_bit(LSFL_NODIR, &ls->ls_flags);
+}
+
+/* takes a snapshot from dlm atomic flags */
+static inline uint32_t dlm_flags_val(const unsigned long *addr,
+				     uint32_t min, uint32_t max)
+{
+	uint32_t bit = min, val = 0;
+
+	for_each_set_bit_from(bit, addr, max + 1) {
+		val |= BIT(bit);
+	}
+
+	return val;
+}
+
+static inline uint32_t dlm_iflags_val(const struct dlm_lkb *lkb)
+{
+	return dlm_flags_val(&lkb->lkb_iflags, __DLM_IFL_MIN_BIT,
+			     __DLM_IFL_MAX_BIT);
+}
+
+static inline uint32_t dlm_dflags_val(const struct dlm_lkb *lkb)
+{
+	return dlm_flags_val(&lkb->lkb_dflags, __DLM_DFL_MIN_BIT,
+			     __DLM_DFL_MAX_BIT);
+}
+
+/* coming from UAPI header
+ *
+ * TODO:
+ * Move this to UAPI header and let other values point to them and use BIT()
+ */
+#define DLM_SBF_DEMOTED_BIT	0
+#define __DLM_SBF_MIN_BIT	DLM_SBF_DEMOTED_BIT
+#define DLM_SBF_VALNOTVALID_BIT	1
+#define DLM_SBF_ALTMODE_BIT	2
+#define __DLM_SBF_MAX_BIT	DLM_SBF_ALTMODE_BIT
+
+static inline uint32_t dlm_sbflags_val(const struct dlm_lkb *lkb)
+{
+	/* be sure the next person updates this */
+	BUILD_BUG_ON(BIT(__DLM_SBF_MAX_BIT) != DLM_SBF_ALTMODE);
+
+	return dlm_flags_val(&lkb->lkb_sbflags, __DLM_SBF_MIN_BIT,
+			     __DLM_SBF_MAX_BIT);
+}
+
+static inline void dlm_set_flags_val(unsigned long *addr, uint32_t val,
+				     uint32_t min, uint32_t max)
+{
+	uint32_t bit;
+
+	for (bit = min; bit < (max + 1); bit++) {
+		if (val & BIT(bit))
+			set_bit(bit, addr);
+		else
+			clear_bit(bit, addr);
+	}
+}
+
+static inline void dlm_set_dflags_val(struct dlm_lkb *lkb, uint32_t val)
+{
+	dlm_set_flags_val(&lkb->lkb_dflags, val, __DLM_DFL_MIN_BIT,
+			  __DLM_DFL_MAX_BIT);
+}
+
+static inline void dlm_set_sbflags_val(struct dlm_lkb *lkb, uint32_t val)
+{
+	dlm_set_flags_val(&lkb->lkb_sbflags, val, __DLM_SBF_MIN_BIT,
+			  __DLM_SBF_MAX_BIT);
+}
+
+extern struct workqueue_struct *dlm_wq;
+
+int dlm_plock_init(void);
+void dlm_plock_exit(void);
+
+#ifdef CONFIG_DLM_DEBUG
+void dlm_register_debugfs(void);
+void dlm_unregister_debugfs(void);
+void dlm_create_debug_file(struct dlm_ls *ls);
+void dlm_delete_debug_file(struct dlm_ls *ls);
+void *dlm_create_debug_comms_file(int nodeid, void *data);
+void dlm_delete_debug_comms_file(void *ctx);
+#else
+static inline void dlm_register_debugfs(void) { }
+static inline void dlm_unregister_debugfs(void) { }
+static inline void dlm_create_debug_file(struct dlm_ls *ls) { }
+static inline void dlm_delete_debug_file(struct dlm_ls *ls) { }
+static inline void *dlm_create_debug_comms_file(int nodeid, void *data) { return NULL; }
+static inline void dlm_delete_debug_comms_file(void *ctx) { }
+#endif
+
+#endif				/* __DLM_INTERNAL_DOT_H__ */
+

linux-7.2.3/fs/dlm/lock.h

@@ -0,0 +1,85 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __LOCK_DOT_H__
+#define __LOCK_DOT_H__
+
+void dlm_dump_rsb(struct dlm_rsb *r);
+void dlm_dump_rsb_name(struct dlm_ls *ls, const char *name, int len);
+void dlm_print_lkb(struct dlm_lkb *lkb);
+void dlm_receive_message_saved(struct dlm_ls *ls, const struct dlm_message *ms,
+			       uint32_t saved_seq);
+void dlm_receive_buffer(const union dlm_packet *p, int nodeid);
+int dlm_modes_compat(int mode1, int mode2);
+void free_inactive_rsb(struct dlm_rsb *r);
+void dlm_put_rsb(struct dlm_rsb *r);
+void dlm_hold_rsb(struct dlm_rsb *r);
+int dlm_put_lkb(struct dlm_lkb *lkb);
+int dlm_lock_recovery_try(struct dlm_ls *ls);
+void dlm_lock_recovery(struct dlm_ls *ls);
+void dlm_unlock_recovery(struct dlm_ls *ls);
+void dlm_rsb_scan(struct timer_list *timer);
+void resume_scan_timer(struct dlm_ls *ls);
+
+int dlm_master_lookup(struct dlm_ls *ls, int from_nodeid, const char *name,
+		      int len, unsigned int flags, int *r_nodeid, int *result);
+
+int dlm_search_rsb_tree(struct rhashtable *rhash, const void *name, int len,
+			struct dlm_rsb **r_ret);
+
+void dlm_recover_purge(struct dlm_ls *ls, const struct list_head *root_list);
+void dlm_purge_mstcpy_locks(struct dlm_rsb *r);
+void dlm_recover_grant(struct dlm_ls *ls);
+int dlm_recover_waiters_post(struct dlm_ls *ls);
+void dlm_recover_waiters_pre(struct dlm_ls *ls);
+int dlm_recover_master_copy(struct dlm_ls *ls, const struct dlm_rcom *rc,
+			    __le32 *rl_remid, __le32 *rl_result);
+int dlm_recover_process_copy(struct dlm_ls *ls, const struct dlm_rcom *rc,
+			     uint64_t seq);
+
+int dlm_user_request(struct dlm_ls *ls, struct dlm_user_args *ua, int mode,
+	uint32_t flags, void *name, unsigned int namelen);
+int dlm_user_convert(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
+	int mode, uint32_t flags, uint32_t lkid, char *lvb_in);
+int dlm_user_adopt_orphan(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
+	int mode, uint32_t flags, void *name, unsigned int namelen,
+	uint32_t *lkid);
+int dlm_user_unlock(struct dlm_ls *ls, struct dlm_user_args *ua_tmp,
+	uint32_t flags, uint32_t lkid, char *lvb_in);
+int dlm_user_cancel(struct dlm_ls *ls,  struct dlm_user_args *ua_tmp,
+	uint32_t flags, uint32_t lkid);
+int dlm_user_purge(struct dlm_ls *ls, struct dlm_user_proc *proc,
+	int nodeid, int pid);
+int dlm_user_deadlock(struct dlm_ls *ls, uint32_t flags, uint32_t lkid);
+void dlm_clear_proc_locks(struct dlm_ls *ls, struct dlm_user_proc *proc);
+int dlm_debug_add_lkb(struct dlm_ls *ls, uint32_t lkb_id, char *name, int len,
+		      int lkb_nodeid, unsigned int lkb_flags, int lkb_status);
+int dlm_debug_add_lkb_to_waiters(struct dlm_ls *ls, uint32_t lkb_id,
+				 int mstype, int to_nodeid);
+
+static inline int is_master(struct dlm_rsb *r)
+{
+	WARN_ON_ONCE(r->res_nodeid == -1);
+
+	return !r->res_nodeid;
+}
+
+static inline void lock_rsb(struct dlm_rsb *r)
+{
+	spin_lock_bh(&r->res_lock);
+}
+
+static inline void unlock_rsb(struct dlm_rsb *r)
+{
+	spin_unlock_bh(&r->res_lock);
+}
+
+#endif
+

linux-7.2.3/fs/dlm/lockspace.c

@@ -0,0 +1,841 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2011 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include <linux/module.h>
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "member.h"
+#include "recoverd.h"
+#include "dir.h"
+#include "midcomms.h"
+#include "config.h"
+#include "memory.h"
+#include "lock.h"
+#include "recover.h"
+#include "requestqueue.h"
+#include "user.h"
+#include "ast.h"
+
+static int			ls_count;
+static struct mutex		ls_lock;
+static struct list_head		lslist;
+static spinlock_t		lslist_lock;
+
+static ssize_t dlm_control_store(struct dlm_ls *ls, const char *buf, size_t len)
+{
+	ssize_t ret = len;
+	int n;
+	int rc = kstrtoint(buf, 0, &n);
+
+	if (rc)
+		return rc;
+	ls = dlm_find_lockspace_local(ls);
+	if (!ls)
+		return -EINVAL;
+
+	switch (n) {
+	case 0:
+		dlm_ls_stop(ls);
+		break;
+	case 1:
+		dlm_ls_start(ls);
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	dlm_put_lockspace(ls);
+	return ret;
+}
+
+static ssize_t dlm_event_store(struct dlm_ls *ls, const char *buf, size_t len)
+{
+	int rc = kstrtoint(buf, 0, &ls->ls_uevent_result);
+
+	if (rc)
+		return rc;
+	set_bit(LSFL_UEVENT_WAIT, &ls->ls_flags);
+	wake_up(&ls->ls_uevent_wait);
+	return len;
+}
+
+static ssize_t dlm_id_show(struct dlm_ls *ls, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%u\n", ls->ls_global_id);
+}
+
+static ssize_t dlm_id_store(struct dlm_ls *ls, const char *buf, size_t len)
+{
+	int rc = kstrtouint(buf, 0, &ls->ls_global_id);
+
+	if (rc)
+		return rc;
+	return len;
+}
+
+static ssize_t dlm_nodir_show(struct dlm_ls *ls, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%u\n", dlm_no_directory(ls));
+}
+
+static ssize_t dlm_nodir_store(struct dlm_ls *ls, const char *buf, size_t len)
+{
+	int val;
+	int rc = kstrtoint(buf, 0, &val);
+
+	if (rc)
+		return rc;
+	if (val == 1)
+		set_bit(LSFL_NODIR, &ls->ls_flags);
+	return len;
+}
+
+static ssize_t dlm_recover_status_show(struct dlm_ls *ls, char *buf)
+{
+	uint32_t status = dlm_recover_status(ls);
+	return snprintf(buf, PAGE_SIZE, "%x\n", status);
+}
+
+static ssize_t dlm_recover_nodeid_show(struct dlm_ls *ls, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%d\n", ls->ls_recover_nodeid);
+}
+
+struct dlm_attr {
+	struct attribute attr;
+	ssize_t (*show)(struct dlm_ls *, char *);
+	ssize_t (*store)(struct dlm_ls *, const char *, size_t);
+};
+
+static struct dlm_attr dlm_attr_control = {
+	.attr  = {.name = "control", .mode = S_IWUSR},
+	.store = dlm_control_store
+};
+
+static struct dlm_attr dlm_attr_event = {
+	.attr  = {.name = "event_done", .mode = S_IWUSR},
+	.store = dlm_event_store
+};
+
+static struct dlm_attr dlm_attr_id = {
+	.attr  = {.name = "id", .mode = S_IRUGO | S_IWUSR},
+	.show  = dlm_id_show,
+	.store = dlm_id_store
+};
+
+static struct dlm_attr dlm_attr_nodir = {
+	.attr  = {.name = "nodir", .mode = S_IRUGO | S_IWUSR},
+	.show  = dlm_nodir_show,
+	.store = dlm_nodir_store
+};
+
+static struct dlm_attr dlm_attr_recover_status = {
+	.attr  = {.name = "recover_status", .mode = S_IRUGO},
+	.show  = dlm_recover_status_show
+};
+
+static struct dlm_attr dlm_attr_recover_nodeid = {
+	.attr  = {.name = "recover_nodeid", .mode = S_IRUGO},
+	.show  = dlm_recover_nodeid_show
+};
+
+static struct attribute *dlm_attrs[] = {
+	&dlm_attr_control.attr,
+	&dlm_attr_event.attr,
+	&dlm_attr_id.attr,
+	&dlm_attr_nodir.attr,
+	&dlm_attr_recover_status.attr,
+	&dlm_attr_recover_nodeid.attr,
+	NULL,
+};
+ATTRIBUTE_GROUPS(dlm);
+
+static ssize_t dlm_attr_show(struct kobject *kobj, struct attribute *attr,
+			     char *buf)
+{
+	struct dlm_ls *ls  = container_of(kobj, struct dlm_ls, ls_kobj);
+	struct dlm_attr *a = container_of(attr, struct dlm_attr, attr);
+	return a->show ? a->show(ls, buf) : 0;
+}
+
+static ssize_t dlm_attr_store(struct kobject *kobj, struct attribute *attr,
+			      const char *buf, size_t len)
+{
+	struct dlm_ls *ls  = container_of(kobj, struct dlm_ls, ls_kobj);
+	struct dlm_attr *a = container_of(attr, struct dlm_attr, attr);
+	return a->store ? a->store(ls, buf, len) : len;
+}
+
+static const struct sysfs_ops dlm_attr_ops = {
+	.show  = dlm_attr_show,
+	.store = dlm_attr_store,
+};
+
+static struct kobj_type dlm_ktype = {
+	.default_groups = dlm_groups,
+	.sysfs_ops     = &dlm_attr_ops,
+};
+
+static struct kset *dlm_kset;
+
+static int do_uevent(struct dlm_ls *ls, int in, unsigned int release_recover)
+{
+	char message[512] = {};
+	char *envp[] = { message, NULL };
+
+	if (in) {
+		kobject_uevent(&ls->ls_kobj, KOBJ_ONLINE);
+	} else {
+		snprintf(message, 511, "RELEASE_RECOVER=%u", release_recover);
+		kobject_uevent_env(&ls->ls_kobj, KOBJ_OFFLINE, envp);
+	}
+
+	log_rinfo(ls, "%s the lockspace group...", in ? "joining" : "leaving");
+
+	/* dlm_controld will see the uevent, do the necessary group management
+	   and then write to sysfs to wake us */
+
+	wait_event(ls->ls_uevent_wait,
+		   test_and_clear_bit(LSFL_UEVENT_WAIT, &ls->ls_flags));
+
+	log_rinfo(ls, "group event done %d", ls->ls_uevent_result);
+
+	return ls->ls_uevent_result;
+}
+
+static int dlm_uevent(const struct kobject *kobj, struct kobj_uevent_env *env)
+{
+	const struct dlm_ls *ls = container_of(kobj, struct dlm_ls, ls_kobj);
+
+	add_uevent_var(env, "LOCKSPACE=%s", ls->ls_name);
+	return 0;
+}
+
+static const struct kset_uevent_ops dlm_uevent_ops = {
+	.uevent = dlm_uevent,
+};
+
+int __init dlm_lockspace_init(void)
+{
+	ls_count = 0;
+	mutex_init(&ls_lock);
+	INIT_LIST_HEAD(&lslist);
+	spin_lock_init(&lslist_lock);
+
+	dlm_kset = kset_create_and_add("dlm", &dlm_uevent_ops, kernel_kobj);
+	if (!dlm_kset) {
+		printk(KERN_WARNING "%s: can not create kset\n", __func__);
+		return -ENOMEM;
+	}
+	return 0;
+}
+
+void dlm_lockspace_exit(void)
+{
+	kset_unregister(dlm_kset);
+}
+
+struct dlm_ls *dlm_find_lockspace_global(uint32_t id)
+{
+	struct dlm_ls *ls;
+
+	spin_lock_bh(&lslist_lock);
+
+	list_for_each_entry(ls, &lslist, ls_list) {
+		if (ls->ls_global_id == id) {
+			atomic_inc(&ls->ls_count);
+			goto out;
+		}
+	}
+	ls = NULL;
+ out:
+	spin_unlock_bh(&lslist_lock);
+	return ls;
+}
+
+struct dlm_ls *dlm_find_lockspace_local(dlm_lockspace_t *lockspace)
+{
+	struct dlm_ls *ls = lockspace;
+
+	atomic_inc(&ls->ls_count);
+	return ls;
+}
+
+struct dlm_ls *dlm_find_lockspace_device(int minor)
+{
+	struct dlm_ls *ls;
+
+	spin_lock_bh(&lslist_lock);
+	list_for_each_entry(ls, &lslist, ls_list) {
+		if (ls->ls_device.minor == minor) {
+			atomic_inc(&ls->ls_count);
+			goto out;
+		}
+	}
+	ls = NULL;
+ out:
+	spin_unlock_bh(&lslist_lock);
+	return ls;
+}
+
+void dlm_put_lockspace(struct dlm_ls *ls)
+{
+	if (atomic_dec_and_test(&ls->ls_count))
+		wake_up(&ls->ls_count_wait);
+}
+
+static void remove_lockspace(struct dlm_ls *ls)
+{
+retry:
+	wait_event(ls->ls_count_wait, atomic_read(&ls->ls_count) == 0);
+
+	spin_lock_bh(&lslist_lock);
+	if (atomic_read(&ls->ls_count) != 0) {
+		spin_unlock_bh(&lslist_lock);
+		goto retry;
+	}
+
+	WARN_ON(ls->ls_create_count != 0);
+	list_del(&ls->ls_list);
+	spin_unlock_bh(&lslist_lock);
+}
+
+static int threads_start(void)
+{
+	int error;
+
+	/* Thread for sending/receiving messages for all lockspace's */
+	error = dlm_midcomms_start();
+	if (error)
+		log_print("cannot start dlm midcomms %d", error);
+
+	return error;
+}
+
+static int lkb_idr_free(struct dlm_lkb *lkb)
+{
+	if (lkb->lkb_lvbptr && test_bit(DLM_IFL_MSTCPY_BIT, &lkb->lkb_iflags))
+		dlm_free_lvb(lkb->lkb_lvbptr);
+
+	dlm_free_lkb(lkb);
+	return 0;
+}
+
+static void rhash_free_rsb(void *ptr, void *arg)
+{
+	struct dlm_rsb *rsb = ptr;
+
+	dlm_free_rsb(rsb);
+}
+
+static void free_lockspace(struct work_struct *work)
+{
+	struct dlm_ls *ls  = container_of(work, struct dlm_ls, ls_free_work);
+	struct dlm_lkb *lkb;
+	unsigned long id;
+
+	/*
+	 * Free all lkb's in xa
+	 */
+	xa_for_each(&ls->ls_lkbxa, id, lkb) {
+		lkb_idr_free(lkb);
+	}
+	xa_destroy(&ls->ls_lkbxa);
+
+	/*
+	 * Free all rsb's on rsbtbl
+	 */
+	rhashtable_free_and_destroy(&ls->ls_rsbtbl, rhash_free_rsb, NULL);
+
+	kfree(ls);
+}
+
+static int new_lockspace(const char *name, const char *cluster,
+			 uint32_t flags, int lvblen,
+			 const struct dlm_lockspace_ops *ops, void *ops_arg,
+			 int *ops_result, dlm_lockspace_t **lockspace)
+{
+	struct dlm_ls *ls;
+	int namelen = strlen(name);
+	int error;
+
+	if (namelen > DLM_LOCKSPACE_LEN || namelen == 0)
+		return -EINVAL;
+
+	if (lvblen % 8)
+		return -EINVAL;
+
+	if (!try_module_get(THIS_MODULE))
+		return -EINVAL;
+
+	if (!dlm_user_daemon_available()) {
+		log_print("dlm user daemon not available");
+		error = -EUNATCH;
+		goto out;
+	}
+
+	if (ops && ops_result) {
+	       	if (!dlm_config.ci_recover_callbacks)
+			*ops_result = -EOPNOTSUPP;
+		else
+			*ops_result = 0;
+	}
+
+	if (!cluster)
+		log_print("dlm cluster name '%s' is being used without an application provided cluster name",
+			  dlm_config.ci_cluster_name);
+
+	if (dlm_config.ci_recover_callbacks && cluster &&
+	    strncmp(cluster, dlm_config.ci_cluster_name, DLM_LOCKSPACE_LEN)) {
+		log_print("dlm cluster name '%s' does not match "
+			  "the application cluster name '%s'",
+			  dlm_config.ci_cluster_name, cluster);
+		error = -EBADR;
+		goto out;
+	}
+
+	error = 0;
+
+	spin_lock_bh(&lslist_lock);
+	list_for_each_entry(ls, &lslist, ls_list) {
+		WARN_ON(ls->ls_create_count <= 0);
+		if (ls->ls_namelen != namelen)
+			continue;
+		if (memcmp(ls->ls_name, name, namelen))
+			continue;
+		if (flags & DLM_LSFL_NEWEXCL) {
+			error = -EEXIST;
+			break;
+		}
+		ls->ls_create_count++;
+		*lockspace = ls;
+		error = 1;
+		break;
+	}
+	spin_unlock_bh(&lslist_lock);
+
+	if (error)
+		goto out;
+
+	error = -ENOMEM;
+
+	ls = kzalloc_obj(*ls, GFP_NOFS);
+	if (!ls)
+		goto out;
+	memcpy(ls->ls_name, name, namelen);
+	ls->ls_namelen = namelen;
+	ls->ls_lvblen = lvblen;
+	atomic_set(&ls->ls_count, 0);
+	init_waitqueue_head(&ls->ls_count_wait);
+	ls->ls_flags = 0;
+
+	if (ops && dlm_config.ci_recover_callbacks) {
+		ls->ls_ops = ops;
+		ls->ls_ops_arg = ops_arg;
+	}
+
+	if (flags & DLM_LSFL_SOFTIRQ)
+		set_bit(LSFL_SOFTIRQ, &ls->ls_flags);
+
+	/* ls_exflags are forced to match among nodes, and we don't
+	 * need to require all nodes to have some flags set
+	 */
+	ls->ls_exflags = (flags & ~(DLM_LSFL_FS | DLM_LSFL_NEWEXCL |
+				    DLM_LSFL_SOFTIRQ));
+
+	INIT_LIST_HEAD(&ls->ls_slow_inactive);
+	INIT_LIST_HEAD(&ls->ls_slow_active);
+	rwlock_init(&ls->ls_rsbtbl_lock);
+
+	error = rhashtable_init(&ls->ls_rsbtbl, &dlm_rhash_rsb_params);
+	if (error)
+		goto out_lsfree;
+
+	xa_init_flags(&ls->ls_lkbxa, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_BH);
+	rwlock_init(&ls->ls_lkbxa_lock);
+
+	INIT_LIST_HEAD(&ls->ls_waiters);
+	spin_lock_init(&ls->ls_waiters_lock);
+	INIT_LIST_HEAD(&ls->ls_orphans);
+	spin_lock_init(&ls->ls_orphans_lock);
+
+	INIT_LIST_HEAD(&ls->ls_nodes);
+	INIT_LIST_HEAD(&ls->ls_nodes_gone);
+	ls->ls_num_nodes = 0;
+	ls->ls_low_nodeid = 0;
+	ls->ls_total_weight = 0;
+	ls->ls_node_array = NULL;
+
+	memset(&ls->ls_local_rsb, 0, sizeof(struct dlm_rsb));
+	ls->ls_local_rsb.res_ls = ls;
+
+	ls->ls_debug_rsb_dentry = NULL;
+	ls->ls_debug_waiters_dentry = NULL;
+
+	init_waitqueue_head(&ls->ls_uevent_wait);
+	ls->ls_uevent_result = 0;
+	init_completion(&ls->ls_recovery_done);
+	ls->ls_recovery_result = -1;
+
+	spin_lock_init(&ls->ls_cb_lock);
+	INIT_LIST_HEAD(&ls->ls_cb_delay);
+
+	INIT_WORK(&ls->ls_free_work, free_lockspace);
+
+	ls->ls_recoverd_task = NULL;
+	mutex_init(&ls->ls_recoverd_active);
+	spin_lock_init(&ls->ls_recover_lock);
+	spin_lock_init(&ls->ls_rcom_spin);
+	get_random_bytes(&ls->ls_rcom_seq, sizeof(uint64_t));
+	ls->ls_recover_status = 0;
+	ls->ls_recover_seq = get_random_u64();
+	ls->ls_recover_args = NULL;
+	init_rwsem(&ls->ls_in_recovery);
+	rwlock_init(&ls->ls_recv_active);
+	INIT_LIST_HEAD(&ls->ls_requestqueue);
+	rwlock_init(&ls->ls_requestqueue_lock);
+	spin_lock_init(&ls->ls_clear_proc_locks);
+
+	/* Due backwards compatibility with 3.1 we need to use maximum
+	 * possible dlm message size to be sure the message will fit and
+	 * not having out of bounds issues. However on sending side 3.2
+	 * might send less.
+	 */
+	ls->ls_recover_buf = kmalloc(DLM_MAX_SOCKET_BUFSIZE, GFP_NOFS);
+	if (!ls->ls_recover_buf) {
+		error = -ENOMEM;
+		goto out_lkbxa;
+	}
+
+	ls->ls_slot = 0;
+	ls->ls_num_slots = 0;
+	ls->ls_slots_size = 0;
+	ls->ls_slots = NULL;
+
+	INIT_LIST_HEAD(&ls->ls_recover_list);
+	spin_lock_init(&ls->ls_recover_list_lock);
+	xa_init_flags(&ls->ls_recover_xa, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_BH);
+	spin_lock_init(&ls->ls_recover_xa_lock);
+	ls->ls_recover_list_count = 0;
+	init_waitqueue_head(&ls->ls_wait_general);
+	INIT_LIST_HEAD(&ls->ls_masters_list);
+	rwlock_init(&ls->ls_masters_lock);
+	INIT_LIST_HEAD(&ls->ls_dir_dump_list);
+	rwlock_init(&ls->ls_dir_dump_lock);
+
+	INIT_LIST_HEAD(&ls->ls_scan_list);
+	spin_lock_init(&ls->ls_scan_lock);
+	timer_setup(&ls->ls_scan_timer, dlm_rsb_scan, TIMER_DEFERRABLE);
+
+	spin_lock_bh(&lslist_lock);
+	ls->ls_create_count = 1;
+	list_add(&ls->ls_list, &lslist);
+	spin_unlock_bh(&lslist_lock);
+
+	if (flags & DLM_LSFL_FS)
+		set_bit(LSFL_FS, &ls->ls_flags);
+
+	error = dlm_callback_start(ls);
+	if (error) {
+		log_error(ls, "can't start dlm_callback %d", error);
+		goto out_delist;
+	}
+
+	init_waitqueue_head(&ls->ls_recover_lock_wait);
+
+	/*
+	 * Once started, dlm_recoverd first looks for ls in lslist, then
+	 * initializes ls_in_recovery as locked in "down" mode.  We need
+	 * to wait for the wakeup from dlm_recoverd because in_recovery
+	 * has to start out in down mode.
+	 */
+
+	error = dlm_recoverd_start(ls);
+	if (error) {
+		log_error(ls, "can't start dlm_recoverd %d", error);
+		goto out_callback;
+	}
+
+	wait_event(ls->ls_recover_lock_wait,
+		   test_bit(LSFL_RECOVER_LOCK, &ls->ls_flags));
+
+	ls->ls_kobj.kset = dlm_kset;
+	error = kobject_init_and_add(&ls->ls_kobj, &dlm_ktype, NULL,
+				     "%s", ls->ls_name);
+	if (error)
+		goto out_recoverd;
+	kobject_uevent(&ls->ls_kobj, KOBJ_ADD);
+
+	/* This uevent triggers dlm_controld in userspace to add us to the
+	   group of nodes that are members of this lockspace (managed by the
+	   cluster infrastructure.)  Once it's done that, it tells us who the
+	   current lockspace members are (via configfs) and then tells the
+	   lockspace to start running (via sysfs) in dlm_ls_start(). */
+
+	error = do_uevent(ls, 1, 0);
+	if (error < 0)
+		goto out_recoverd;
+
+	/* wait until recovery is successful or failed */
+	wait_for_completion(&ls->ls_recovery_done);
+	error = ls->ls_recovery_result;
+	if (error)
+		goto out_members;
+
+	dlm_create_debug_file(ls);
+
+	log_rinfo(ls, "join complete");
+	*lockspace = ls;
+	return 0;
+
+ out_members:
+	do_uevent(ls, 0, 0);
+	dlm_clear_members(ls);
+	kfree(ls->ls_node_array);
+ out_recoverd:
+	dlm_recoverd_stop(ls);
+ out_callback:
+	dlm_callback_stop(ls);
+ out_delist:
+	spin_lock_bh(&lslist_lock);
+	list_del(&ls->ls_list);
+	spin_unlock_bh(&lslist_lock);
+	xa_destroy(&ls->ls_recover_xa);
+	kfree(ls->ls_recover_buf);
+ out_lkbxa:
+	xa_destroy(&ls->ls_lkbxa);
+	rhashtable_destroy(&ls->ls_rsbtbl);
+ out_lsfree:
+	kobject_put(&ls->ls_kobj);
+	kfree(ls);
+ out:
+	module_put(THIS_MODULE);
+	return error;
+}
+
+static int __dlm_new_lockspace(const char *name, const char *cluster,
+			       uint32_t flags, int lvblen,
+			       const struct dlm_lockspace_ops *ops,
+			       void *ops_arg, int *ops_result,
+			       dlm_lockspace_t **lockspace)
+{
+	int error = 0;
+
+	mutex_lock(&ls_lock);
+	if (!ls_count)
+		error = threads_start();
+	if (error)
+		goto out;
+
+	error = new_lockspace(name, cluster, flags, lvblen, ops, ops_arg,
+			      ops_result, lockspace);
+	if (!error)
+		ls_count++;
+	if (error > 0)
+		error = 0;
+	if (!ls_count) {
+		dlm_midcomms_shutdown();
+		dlm_midcomms_stop();
+	}
+ out:
+	mutex_unlock(&ls_lock);
+	return error;
+}
+
+int dlm_new_lockspace(const char *name, const char *cluster, uint32_t flags,
+		      int lvblen, const struct dlm_lockspace_ops *ops,
+		      void *ops_arg, int *ops_result,
+		      dlm_lockspace_t **lockspace)
+{
+	return __dlm_new_lockspace(name, cluster, flags | DLM_LSFL_FS, lvblen,
+				   ops, ops_arg, ops_result, lockspace);
+}
+
+int dlm_new_user_lockspace(const char *name, const char *cluster,
+			   uint32_t flags, int lvblen,
+			   const struct dlm_lockspace_ops *ops,
+			   void *ops_arg, int *ops_result,
+			   dlm_lockspace_t **lockspace)
+{
+	if (flags & DLM_LSFL_SOFTIRQ)
+		return -EINVAL;
+
+	return __dlm_new_lockspace(name, cluster, flags, lvblen, ops,
+				   ops_arg, ops_result, lockspace);
+}
+
+/* NOTE: We check the lkbxa here rather than the resource table.
+   This is because there may be LKBs queued as ASTs that have been unlinked
+   from their RSBs and are pending deletion once the AST has been delivered */
+
+static int lockspace_busy(struct dlm_ls *ls, unsigned int release_option)
+{
+	struct dlm_lkb *lkb;
+	unsigned long id;
+	int rv = 0;
+
+	read_lock_bh(&ls->ls_lkbxa_lock);
+	if (release_option == DLM_RELEASE_NO_LOCKS) {
+		xa_for_each(&ls->ls_lkbxa, id, lkb) {
+			rv = 1;
+			break;
+		}
+	} else if (release_option == DLM_RELEASE_UNUSED) {
+		/* TODO: handle this UNUSED option as NO_LOCKS in later patch */
+		xa_for_each(&ls->ls_lkbxa, id, lkb) {
+			if (lkb->lkb_nodeid == 0 &&
+			    lkb->lkb_grmode != DLM_LOCK_IV) {
+				rv = 1;
+				break;
+			}
+		}
+	} else {
+		rv = 0;
+	}
+	read_unlock_bh(&ls->ls_lkbxa_lock);
+	return rv;
+}
+
+static int release_lockspace(struct dlm_ls *ls, unsigned int release_option)
+{
+	int busy, rv;
+
+	busy = lockspace_busy(ls, release_option);
+
+	spin_lock_bh(&lslist_lock);
+	if (ls->ls_create_count == 1) {
+		if (busy) {
+			rv = -EBUSY;
+		} else {
+			/* remove_lockspace takes ls off lslist */
+			ls->ls_create_count = 0;
+			rv = 0;
+		}
+	} else if (ls->ls_create_count > 1) {
+		rv = --ls->ls_create_count;
+	} else {
+		rv = -EINVAL;
+	}
+	spin_unlock_bh(&lslist_lock);
+
+	if (rv) {
+		log_debug(ls, "release_lockspace no remove %d", rv);
+		return rv;
+	}
+
+	if (ls_count == 1)
+		dlm_midcomms_version_wait();
+
+	dlm_device_deregister(ls);
+
+	if (release_option != DLM_RELEASE_NO_EVENT &&
+	    dlm_user_daemon_available())
+		do_uevent(ls, 0, (release_option == DLM_RELEASE_RECOVER));
+
+	dlm_recoverd_stop(ls);
+
+	/* clear the LSFL_RUNNING flag to fast up
+	 * time_shutdown_sync(), we don't care anymore
+	 */
+	clear_bit(LSFL_RUNNING, &ls->ls_flags);
+	timer_shutdown_sync(&ls->ls_scan_timer);
+
+	if (ls_count == 1) {
+		dlm_clear_members(ls);
+		dlm_midcomms_shutdown();
+	}
+
+	dlm_callback_stop(ls);
+
+	remove_lockspace(ls);
+
+	dlm_delete_debug_file(ls);
+
+	kobject_put(&ls->ls_kobj);
+
+	xa_destroy(&ls->ls_recover_xa);
+	kfree(ls->ls_recover_buf);
+
+	/*
+	 * Free structures on any other lists
+	 */
+
+	dlm_purge_requestqueue(ls);
+	kfree(ls->ls_recover_args);
+	dlm_clear_members(ls);
+	dlm_clear_members_gone(ls);
+	kfree(ls->ls_node_array);
+
+	log_rinfo(ls, "%s final free", __func__);
+
+	/* delayed free of data structures see free_lockspace() */
+	queue_work(dlm_wq, &ls->ls_free_work);
+	module_put(THIS_MODULE);
+	return 0;
+}
+
+/*
+ * Called when a system has released all its locks and is not going to use the
+ * lockspace any longer.  We free everything we're managing for this lockspace.
+ * Remaining nodes will go through the recovery process as if we'd died.  The
+ * lockspace must continue to function as usual, participating in recoveries,
+ * until this returns.
+ *
+ * See DLM_RELEASE defines for release_option values and their meaning.
+ */
+
+int dlm_release_lockspace(void *lockspace, unsigned int release_option)
+{
+	struct dlm_ls *ls;
+	int error;
+
+	if (release_option > __DLM_RELEASE_MAX)
+		return -EINVAL;
+
+	ls = dlm_find_lockspace_local(lockspace);
+	if (!ls)
+		return -EINVAL;
+	dlm_put_lockspace(ls);
+
+	mutex_lock(&ls_lock);
+	error = release_lockspace(ls, release_option);
+	if (!error)
+		ls_count--;
+	if (!ls_count)
+		dlm_midcomms_stop();
+	mutex_unlock(&ls_lock);
+
+	return error;
+}
+
+void dlm_stop_lockspaces(void)
+{
+	struct dlm_ls *ls;
+	int count;
+
+ restart:
+	count = 0;
+	spin_lock_bh(&lslist_lock);
+	list_for_each_entry(ls, &lslist, ls_list) {
+		if (!test_bit(LSFL_RUNNING, &ls->ls_flags)) {
+			count++;
+			continue;
+		}
+		spin_unlock_bh(&lslist_lock);
+		log_error(ls, "no userland control daemon, stopping lockspace");
+		dlm_ls_stop(ls);
+		goto restart;
+	}
+	spin_unlock_bh(&lslist_lock);
+
+	if (count)
+		log_print("dlm user daemon left %d lockspaces", count);
+}

linux-7.2.3/fs/dlm/lockspace.h

@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __LOCKSPACE_DOT_H__
+#define __LOCKSPACE_DOT_H__
+
+/* DLM_LSFL_FS
+ *   The lockspace user is in the kernel (i.e. filesystem).  Enables
+ *   direct bast/cast callbacks.
+ *
+ * internal lockspace flag - will be removed in future
+ */
+#define DLM_LSFL_FS	0x00000004
+
+int dlm_lockspace_init(void);
+void dlm_lockspace_exit(void);
+struct dlm_ls *dlm_find_lockspace_global(uint32_t id);
+struct dlm_ls *dlm_find_lockspace_local(void *id);
+struct dlm_ls *dlm_find_lockspace_device(int minor);
+void dlm_put_lockspace(struct dlm_ls *ls);
+void dlm_stop_lockspaces(void);
+int dlm_new_user_lockspace(const char *name, const char *cluster,
+			   uint32_t flags, int lvblen,
+			   const struct dlm_lockspace_ops *ops,
+			   void *ops_arg, int *ops_result,
+			   dlm_lockspace_t **lockspace);
+
+#endif				/* __LOCKSPACE_DOT_H__ */
+

linux-7.2.3/fs/dlm/lowcomms.h

@@ -0,0 +1,55 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2009 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __LOWCOMMS_DOT_H__
+#define __LOWCOMMS_DOT_H__
+
+#include "dlm_internal.h"
+
+#define DLM_MIDCOMMS_OPT_LEN		sizeof(struct dlm_opts)
+#define DLM_MAX_APP_BUFSIZE		(DLM_MAX_SOCKET_BUFSIZE - \
+					 DLM_MIDCOMMS_OPT_LEN)
+
+#define CONN_HASH_SIZE 32
+
+/* This is deliberately very simple because most clusters have simple
+ * sequential nodeids, so we should be able to go straight to a connection
+ * struct in the array
+ */
+static inline int nodeid_hash(int nodeid)
+{
+	return nodeid & (CONN_HASH_SIZE-1);
+}
+
+/* check if dlm is running */
+bool dlm_lowcomms_is_running(void);
+
+int dlm_lowcomms_start(void);
+void dlm_lowcomms_shutdown(void);
+void dlm_lowcomms_shutdown_node(int nodeid, bool force);
+void dlm_lowcomms_stop(void);
+void dlm_lowcomms_init(void);
+void dlm_lowcomms_exit(void);
+int dlm_lowcomms_close(int nodeid);
+struct dlm_msg *dlm_lowcomms_new_msg(int nodeid, int len, char **ppc,
+				     void (*cb)(void *data), void *data);
+void dlm_lowcomms_commit_msg(struct dlm_msg *msg);
+void dlm_lowcomms_put_msg(struct dlm_msg *msg);
+int dlm_lowcomms_resend_msg(struct dlm_msg *msg);
+int dlm_lowcomms_connect_node(int nodeid);
+int dlm_lowcomms_nodes_set_mark(int nodeid, unsigned int mark);
+int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr);
+void dlm_midcomms_receive_done(int nodeid);
+struct kmem_cache *dlm_lowcomms_writequeue_cache_create(void);
+struct kmem_cache *dlm_lowcomms_msg_cache_create(void);
+
+#endif				/* __LOWCOMMS_DOT_H__ */
+

linux-7.2.3/fs/dlm/lvb_table.h

@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __LVB_TABLE_DOT_H__
+#define __LVB_TABLE_DOT_H__
+
+extern const int dlm_lvb_operations[8][8];
+
+#endif

linux-7.2.3/fs/dlm/main.c

@@ -0,0 +1,105 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include <linux/module.h>
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "lock.h"
+#include "user.h"
+#include "memory.h"
+#include "config.h"
+#include "midcomms.h"
+
+#define CREATE_TRACE_POINTS
+#include <trace/events/dlm.h>
+
+struct workqueue_struct *dlm_wq;
+
+static int __init init_dlm(void)
+{
+	int error;
+
+	error = dlm_memory_init();
+	if (error)
+		goto out;
+
+	dlm_midcomms_init();
+
+	error = dlm_lockspace_init();
+	if (error)
+		goto out_mem;
+
+	error = dlm_config_init();
+	if (error)
+		goto out_lockspace;
+
+	dlm_register_debugfs();
+
+	error = dlm_user_init();
+	if (error)
+		goto out_debug;
+
+	error = dlm_plock_init();
+	if (error)
+		goto out_user;
+
+	dlm_wq = alloc_workqueue("dlm_wq", WQ_PERCPU, 0);
+	if (!dlm_wq) {
+		error = -ENOMEM;
+		goto out_plock;
+	}
+
+	printk("DLM installed\n");
+
+	return 0;
+
+ out_plock:
+	dlm_plock_exit();
+ out_user:
+	dlm_user_exit();
+ out_debug:
+	dlm_unregister_debugfs();
+	dlm_config_exit();
+ out_lockspace:
+	dlm_lockspace_exit();
+ out_mem:
+	dlm_midcomms_exit();
+	dlm_memory_exit();
+ out:
+	return error;
+}
+
+static void __exit exit_dlm(void)
+{
+	/* be sure every pending work e.g. freeing is done */
+	destroy_workqueue(dlm_wq);
+	dlm_plock_exit();
+	dlm_user_exit();
+	dlm_config_exit();
+	dlm_lockspace_exit();
+	dlm_midcomms_exit();
+	dlm_unregister_debugfs();
+	dlm_memory_exit();
+}
+
+module_init(init_dlm);
+module_exit(exit_dlm);
+
+MODULE_DESCRIPTION("Distributed Lock Manager");
+MODULE_AUTHOR("Red Hat, Inc.");
+MODULE_LICENSE("GPL");
+
+EXPORT_SYMBOL_GPL(dlm_new_lockspace);
+EXPORT_SYMBOL_GPL(dlm_release_lockspace);
+EXPORT_SYMBOL_GPL(dlm_lock);
+EXPORT_SYMBOL_GPL(dlm_unlock);
+

linux-7.2.3/fs/dlm/member.c

@@ -0,0 +1,772 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2011 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "member.h"
+#include "recoverd.h"
+#include "recover.h"
+#include "rcom.h"
+#include "config.h"
+#include "midcomms.h"
+#include "lowcomms.h"
+
+int dlm_slots_version(const struct dlm_header *h)
+{
+	if ((le32_to_cpu(h->h_version) & 0x0000FFFF) < DLM_HEADER_SLOTS)
+		return 0;
+	return 1;
+}
+
+void dlm_slot_save(struct dlm_ls *ls, struct dlm_rcom *rc,
+		   struct dlm_member *memb)
+{
+	struct rcom_config *rf = (struct rcom_config *)rc->rc_buf;
+
+	if (!dlm_slots_version(&rc->rc_header))
+		return;
+
+	memb->slot = le16_to_cpu(rf->rf_our_slot);
+	memb->generation = le32_to_cpu(rf->rf_generation);
+}
+
+void dlm_slots_copy_out(struct dlm_ls *ls, struct dlm_rcom *rc)
+{
+	struct dlm_slot *slot;
+	struct rcom_slot *ro;
+	int i;
+
+	ro = (struct rcom_slot *)(rc->rc_buf + sizeof(struct rcom_config));
+
+	/* ls_slots array is sparse, but not rcom_slots */
+
+	for (i = 0; i < ls->ls_slots_size; i++) {
+		slot = &ls->ls_slots[i];
+		if (!slot->nodeid)
+			continue;
+		ro->ro_nodeid = cpu_to_le32(slot->nodeid);
+		ro->ro_slot = cpu_to_le16(slot->slot);
+		ro++;
+	}
+}
+
+#define SLOT_DEBUG_LINE 128
+
+static void log_slots(struct dlm_ls *ls, uint32_t gen, int num_slots,
+		      struct rcom_slot *ro0, struct dlm_slot *array,
+		      int array_size)
+{
+	char line[SLOT_DEBUG_LINE];
+	int len = SLOT_DEBUG_LINE - 1;
+	int pos = 0;
+	int ret, i;
+
+	memset(line, 0, sizeof(line));
+
+	if (array) {
+		for (i = 0; i < array_size; i++) {
+			if (!array[i].nodeid)
+				continue;
+
+			ret = snprintf(line + pos, len - pos, " %d:%d",
+				       array[i].slot, array[i].nodeid);
+			if (ret >= len - pos)
+				break;
+			pos += ret;
+		}
+	} else if (ro0) {
+		for (i = 0; i < num_slots; i++) {
+			ret = snprintf(line + pos, len - pos, " %d:%d",
+				       ro0[i].ro_slot, ro0[i].ro_nodeid);
+			if (ret >= len - pos)
+				break;
+			pos += ret;
+		}
+	}
+
+	log_rinfo(ls, "generation %u slots %d%s", gen, num_slots, line);
+}
+
+int dlm_slots_copy_in(struct dlm_ls *ls)
+{
+	struct dlm_member *memb;
+	struct dlm_rcom *rc = ls->ls_recover_buf;
+	struct rcom_config *rf = (struct rcom_config *)rc->rc_buf;
+	struct rcom_slot *ro0, *ro;
+	int our_nodeid = dlm_our_nodeid();
+	int i, num_slots;
+	uint32_t gen;
+
+	if (!dlm_slots_version(&rc->rc_header))
+		return -1;
+
+	gen = le32_to_cpu(rf->rf_generation);
+	if (gen <= ls->ls_generation) {
+		log_error(ls, "dlm_slots_copy_in gen %u old %u",
+			  gen, ls->ls_generation);
+	}
+	ls->ls_generation = gen;
+
+	num_slots = le16_to_cpu(rf->rf_num_slots);
+	if (!num_slots)
+		return -1;
+
+	ro0 = (struct rcom_slot *)(rc->rc_buf + sizeof(struct rcom_config));
+
+	log_slots(ls, gen, num_slots, ro0, NULL, 0);
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		for (i = 0, ro = ro0; i < num_slots; i++, ro++) {
+			if (le32_to_cpu(ro->ro_nodeid) != memb->nodeid)
+				continue;
+			memb->slot = le16_to_cpu(ro->ro_slot);
+			memb->slot_prev = memb->slot;
+			break;
+		}
+
+		if (memb->nodeid == our_nodeid) {
+			if (ls->ls_slot && ls->ls_slot != memb->slot) {
+				log_error(ls, "dlm_slots_copy_in our slot "
+					  "changed %d %d", ls->ls_slot,
+					  memb->slot);
+				return -1;
+			}
+
+			if (!ls->ls_slot)
+				ls->ls_slot = memb->slot;
+		}
+
+		if (!memb->slot) {
+			log_error(ls, "dlm_slots_copy_in nodeid %d no slot",
+				   memb->nodeid);
+			return -1;
+		}
+	}
+
+	return 0;
+}
+
+/* for any nodes that do not support slots, we will not have set memb->slot
+   in wait_status_all(), so memb->slot will remain -1, and we will not
+   assign slots or set ls_num_slots here */
+
+int dlm_slots_assign(struct dlm_ls *ls, int *num_slots, int *slots_size,
+		     struct dlm_slot **slots_out, uint32_t *gen_out)
+{
+	struct dlm_member *memb;
+	struct dlm_slot *array;
+	int our_nodeid = dlm_our_nodeid();
+	int array_size, max_slots, i;
+	int need = 0;
+	int max = 0;
+	int num = 0;
+	uint32_t gen = 0;
+
+	/* our own memb struct will have slot -1 gen 0 */
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (memb->nodeid == our_nodeid) {
+			memb->slot = ls->ls_slot;
+			memb->generation = ls->ls_generation;
+			break;
+		}
+	}
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (memb->generation > gen)
+			gen = memb->generation;
+
+		/* node doesn't support slots */
+
+		if (memb->slot == -1)
+			return -1;
+
+		/* node needs a slot assigned */
+
+		if (!memb->slot)
+			need++;
+
+		/* node has a slot assigned */
+
+		num++;
+
+		if (!max || max < memb->slot)
+			max = memb->slot;
+
+		/* sanity check, once slot is assigned it shouldn't change */
+
+		if (memb->slot_prev && memb->slot && memb->slot_prev != memb->slot) {
+			log_error(ls, "nodeid %d slot changed %d %d",
+				  memb->nodeid, memb->slot_prev, memb->slot);
+			return -1;
+		}
+		memb->slot_prev = memb->slot;
+	}
+
+	array_size = max + need;
+	array = kzalloc_objs(*array, array_size, GFP_NOFS);
+	if (!array)
+		return -ENOMEM;
+
+	num = 0;
+
+	/* fill in slots (offsets) that are used */
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (!memb->slot)
+			continue;
+
+		if (memb->slot > array_size) {
+			log_error(ls, "invalid slot number %d", memb->slot);
+			kfree(array);
+			return -1;
+		}
+
+		array[memb->slot - 1].nodeid = memb->nodeid;
+		array[memb->slot - 1].slot = memb->slot;
+		num++;
+	}
+
+	/* assign new slots from unused offsets */
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (memb->slot)
+			continue;
+
+		for (i = 0; i < array_size; i++) {
+			if (array[i].nodeid)
+				continue;
+
+			memb->slot = i + 1;
+			memb->slot_prev = memb->slot;
+			array[i].nodeid = memb->nodeid;
+			array[i].slot = memb->slot;
+			num++;
+
+			if (!ls->ls_slot && memb->nodeid == our_nodeid)
+				ls->ls_slot = memb->slot;
+			break;
+		}
+
+		if (!memb->slot) {
+			log_error(ls, "no free slot found");
+			kfree(array);
+			return -1;
+		}
+	}
+
+	gen++;
+
+	log_slots(ls, gen, num, NULL, array, array_size);
+
+	max_slots = (DLM_MAX_APP_BUFSIZE - sizeof(struct dlm_rcom) -
+		     sizeof(struct rcom_config)) / sizeof(struct rcom_slot);
+
+	if (num > max_slots) {
+		log_error(ls, "num_slots %d exceeds max_slots %d",
+			  num, max_slots);
+		kfree(array);
+		return -1;
+	}
+
+	*gen_out = gen;
+	*slots_out = array;
+	*slots_size = array_size;
+	*num_slots = num;
+	return 0;
+}
+
+static void add_ordered_member(struct dlm_ls *ls, struct dlm_member *new)
+{
+	struct dlm_member *memb = NULL;
+	struct list_head *tmp;
+	struct list_head *newlist = &new->list;
+	struct list_head *head = &ls->ls_nodes;
+
+	list_for_each(tmp, head) {
+		memb = list_entry(tmp, struct dlm_member, list);
+		if (new->nodeid < memb->nodeid)
+			break;
+	}
+
+	if (!memb)
+		list_add_tail(newlist, head);
+	else {
+		list_add_tail(newlist, tmp);
+	}
+}
+
+static int add_remote_member(int nodeid)
+{
+	int error;
+
+	if (nodeid == dlm_our_nodeid())
+		return 0;
+
+	error = dlm_lowcomms_connect_node(nodeid);
+	if (error < 0)
+		return error;
+
+	dlm_midcomms_add_member(nodeid);
+	return 0;
+}
+
+static int dlm_add_member(struct dlm_ls *ls, struct dlm_config_node *node)
+{
+	struct dlm_member *memb;
+	int error;
+
+	memb = kzalloc_obj(*memb, GFP_NOFS);
+	if (!memb)
+		return -ENOMEM;
+
+	memb->nodeid = node->nodeid;
+	memb->weight = node->weight;
+	memb->comm_seq = node->comm_seq;
+
+	error = add_remote_member(node->nodeid);
+	if (error < 0) {
+		kfree(memb);
+		return error;
+	}
+
+	add_ordered_member(ls, memb);
+	ls->ls_num_nodes++;
+	return 0;
+}
+
+static struct dlm_member *find_memb(struct list_head *head, int nodeid)
+{
+	struct dlm_member *memb;
+
+	list_for_each_entry(memb, head, list) {
+		if (memb->nodeid == nodeid)
+			return memb;
+	}
+	return NULL;
+}
+
+int dlm_is_member(struct dlm_ls *ls, int nodeid)
+{
+	if (find_memb(&ls->ls_nodes, nodeid))
+		return 1;
+	return 0;
+}
+
+int dlm_is_removed(struct dlm_ls *ls, int nodeid)
+{
+	WARN_ON_ONCE(!nodeid || nodeid == -1);
+
+	if (find_memb(&ls->ls_nodes_gone, nodeid))
+		return 1;
+	return 0;
+}
+
+static void clear_memb_list(struct list_head *head,
+			    void (*after_del)(int nodeid))
+{
+	struct dlm_member *memb;
+
+	while (!list_empty(head)) {
+		memb = list_entry(head->next, struct dlm_member, list);
+		list_del(&memb->list);
+		if (after_del)
+			after_del(memb->nodeid);
+		kfree(memb);
+	}
+}
+
+static void remove_remote_member(int nodeid)
+{
+	if (nodeid == dlm_our_nodeid())
+		return;
+
+	dlm_midcomms_remove_member(nodeid);
+}
+
+void dlm_clear_members(struct dlm_ls *ls)
+{
+	clear_memb_list(&ls->ls_nodes, remove_remote_member);
+	ls->ls_num_nodes = 0;
+}
+
+void dlm_clear_members_gone(struct dlm_ls *ls)
+{
+	clear_memb_list(&ls->ls_nodes_gone, NULL);
+}
+
+static void make_member_array(struct dlm_ls *ls)
+{
+	struct dlm_member *memb;
+	int i, w, x = 0, total = 0, all_zero = 0, *array;
+
+	kfree(ls->ls_node_array);
+	ls->ls_node_array = NULL;
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (memb->weight)
+			total += memb->weight;
+	}
+
+	/* all nodes revert to weight of 1 if all have weight 0 */
+
+	if (!total) {
+		total = ls->ls_num_nodes;
+		all_zero = 1;
+	}
+
+	ls->ls_total_weight = total;
+	array = kmalloc_objs(*array, total, GFP_NOFS);
+	if (!array)
+		return;
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (!all_zero && !memb->weight)
+			continue;
+
+		if (all_zero)
+			w = 1;
+		else
+			w = memb->weight;
+
+		DLM_ASSERT(x < total, printk("total %d x %d\n", total, x););
+
+		for (i = 0; i < w; i++)
+			array[x++] = memb->nodeid;
+	}
+
+	ls->ls_node_array = array;
+}
+
+/* send a status request to all members just to establish comms connections */
+
+static int ping_members(struct dlm_ls *ls, uint64_t seq)
+{
+	struct dlm_member *memb;
+	int error = 0;
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (dlm_recovery_stopped(ls)) {
+			error = -EINTR;
+			break;
+		}
+		error = dlm_rcom_status(ls, memb->nodeid, 0, seq);
+		if (error)
+			break;
+	}
+	if (error)
+		log_rinfo(ls, "ping_members aborted %d last nodeid %d",
+			  error, ls->ls_recover_nodeid);
+	return error;
+}
+
+static void dlm_lsop_recover_prep(struct dlm_ls *ls)
+{
+	if (!ls->ls_ops || !ls->ls_ops->recover_prep)
+		return;
+	ls->ls_ops->recover_prep(ls->ls_ops_arg);
+}
+
+static void dlm_lsop_recover_slot(struct dlm_ls *ls, struct dlm_member *memb,
+				  unsigned int release_recover)
+{
+	struct dlm_slot slot;
+	uint32_t seq;
+	int error;
+
+	if (!ls->ls_ops || !ls->ls_ops->recover_slot)
+		return;
+
+	/* if there is no comms connection with this node
+	   or the present comms connection is newer
+	   than the one when this member was added, then
+	   we consider the node to have failed (versus
+	   being removed due to dlm_release_lockspace) */
+
+	error = dlm_comm_seq(memb->nodeid, &seq, false);
+
+	if (!release_recover && !error && seq == memb->comm_seq)
+		return;
+
+	slot.nodeid = memb->nodeid;
+	slot.slot = memb->slot;
+
+	ls->ls_ops->recover_slot(ls->ls_ops_arg, &slot);
+}
+
+void dlm_lsop_recover_done(struct dlm_ls *ls)
+{
+	struct dlm_member *memb;
+	struct dlm_slot *slots;
+	int i, num;
+
+	if (!ls->ls_ops || !ls->ls_ops->recover_done)
+		return;
+
+	num = ls->ls_num_nodes;
+	slots = kzalloc_objs(*slots, num);
+	if (!slots)
+		return;
+
+	i = 0;
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (i == num) {
+			log_error(ls, "dlm_lsop_recover_done bad num %d", num);
+			goto out;
+		}
+		slots[i].nodeid = memb->nodeid;
+		slots[i].slot = memb->slot;
+		i++;
+	}
+
+	ls->ls_ops->recover_done(ls->ls_ops_arg, slots, num,
+				 ls->ls_slot, ls->ls_generation);
+ out:
+	kfree(slots);
+}
+
+static struct dlm_config_node *find_config_node(struct dlm_recover *rv,
+						int nodeid)
+{
+	int i;
+
+	for (i = 0; i < rv->nodes_count; i++) {
+		if (rv->nodes[i].nodeid == nodeid)
+			return &rv->nodes[i];
+	}
+	return NULL;
+}
+
+int dlm_recover_members(struct dlm_ls *ls, struct dlm_recover *rv, int *neg_out)
+{
+	struct dlm_member *memb, *safe;
+	struct dlm_config_node *node;
+	int i, error, neg = 0, low = -1;
+	unsigned int release_recover;
+
+	/* previously removed members that we've not finished removing need to
+	 * count as a negative change so the "neg" recovery steps will happen
+	 *
+	 * This functionality must report all member changes to lsops or
+	 * midcomms layer and must never return before.
+	 */
+
+	list_for_each_entry(memb, &ls->ls_nodes_gone, list) {
+		log_rinfo(ls, "prev removed member %d", memb->nodeid);
+		neg++;
+	}
+
+	/* move departed members from ls_nodes to ls_nodes_gone */
+
+	list_for_each_entry_safe(memb, safe, &ls->ls_nodes, list) {
+		node = find_config_node(rv, memb->nodeid);
+		if (!node) {
+			log_error(ls, "remove member %d invalid",
+				  memb->nodeid);
+			return -EFAULT;
+		}
+
+		if (!node->new && !node->gone)
+			continue;
+
+		release_recover = 0;
+
+		if (node->gone) {
+			release_recover = node->release_recover;
+			log_rinfo(ls, "remove member %d%s", memb->nodeid,
+				  release_recover ? " (release_recover)" : "");
+		} else {
+			/* removed and re-added */
+			log_rinfo(ls, "remove member %d comm_seq %u %u",
+				  memb->nodeid, memb->comm_seq, node->comm_seq);
+		}
+
+		neg++;
+		list_move(&memb->list, &ls->ls_nodes_gone);
+		remove_remote_member(memb->nodeid);
+		ls->ls_num_nodes--;
+		dlm_lsop_recover_slot(ls, memb, release_recover);
+	}
+
+	/* add new members to ls_nodes */
+
+	for (i = 0; i < rv->nodes_count; i++) {
+		node = &rv->nodes[i];
+		if (node->gone)
+			continue;
+
+		if (dlm_is_member(ls, node->nodeid))
+			continue;
+		error = dlm_add_member(ls, node);
+		if (error)
+			return error;
+
+		log_rinfo(ls, "add member %d", node->nodeid);
+	}
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		if (low == -1 || memb->nodeid < low)
+			low = memb->nodeid;
+	}
+	ls->ls_low_nodeid = low;
+
+	make_member_array(ls);
+	*neg_out = neg;
+
+	error = ping_members(ls, rv->seq);
+	log_rinfo(ls, "dlm_recover_members %d nodes", ls->ls_num_nodes);
+	return error;
+}
+
+/* Userspace guarantees that dlm_ls_stop() has completed on all nodes before
+   dlm_ls_start() is called on any of them to start the new recovery. */
+
+int dlm_ls_stop(struct dlm_ls *ls)
+{
+	int new;
+
+	/*
+	 * Prevent dlm_recv from being in the middle of something when we do
+	 * the stop.  This includes ensuring dlm_recv isn't processing a
+	 * recovery message (rcom), while dlm_recoverd is aborting and
+	 * resetting things from an in-progress recovery.  i.e. we want
+	 * dlm_recoverd to abort its recovery without worrying about dlm_recv
+	 * processing an rcom at the same time.  Stopping dlm_recv also makes
+	 * it easy for dlm_receive_message() to check locking stopped and add a
+	 * message to the requestqueue without races.
+	 */
+
+	write_lock_bh(&ls->ls_recv_active);
+
+	/*
+	 * Abort any recovery that's in progress (see RECOVER_STOP,
+	 * dlm_recovery_stopped()) and tell any other threads running in the
+	 * dlm to quit any processing (see RUNNING, dlm_locking_stopped()).
+	 */
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	set_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
+	new = test_and_clear_bit(LSFL_RUNNING, &ls->ls_flags);
+	if (new)
+		timer_delete_sync(&ls->ls_scan_timer);
+	ls->ls_recover_seq++;
+
+	/* activate requestqueue and stop processing */
+	write_lock_bh(&ls->ls_requestqueue_lock);
+	set_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags);
+	write_unlock_bh(&ls->ls_requestqueue_lock);
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+	/*
+	 * Let dlm_recv run again, now any normal messages will be saved on the
+	 * requestqueue for later.
+	 */
+
+	write_unlock_bh(&ls->ls_recv_active);
+
+	/*
+	 * This in_recovery lock does two things:
+	 * 1) Keeps this function from returning until all threads are out
+	 *    of locking routines and locking is truly stopped.
+	 * 2) Keeps any new requests from being processed until it's unlocked
+	 *    when recovery is complete.
+	 */
+
+	if (new) {
+		set_bit(LSFL_RECOVER_DOWN, &ls->ls_flags);
+		wake_up_process(ls->ls_recoverd_task);
+		wait_event(ls->ls_recover_lock_wait,
+			   test_bit(LSFL_RECOVER_LOCK, &ls->ls_flags));
+	}
+
+	/*
+	 * The recoverd suspend/resume makes sure that dlm_recoverd (if
+	 * running) has noticed RECOVER_STOP above and quit processing the
+	 * previous recovery.
+	 */
+
+	dlm_recoverd_suspend(ls);
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	kfree(ls->ls_slots);
+	ls->ls_slots = NULL;
+	ls->ls_num_slots = 0;
+	ls->ls_slots_size = 0;
+	ls->ls_recover_status = 0;
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+	dlm_recoverd_resume(ls);
+
+	if (!ls->ls_recover_begin)
+		ls->ls_recover_begin = jiffies;
+
+	/* call recover_prep ops only once and not multiple times
+	 * for each possible dlm_ls_stop() when recovery is already
+	 * stopped.
+	 *
+	 * If we successful was able to clear LSFL_RUNNING bit and
+	 * it was set we know it is the first dlm_ls_stop() call.
+	 */
+	if (new)
+		dlm_lsop_recover_prep(ls);
+
+	return 0;
+}
+
+int dlm_ls_start(struct dlm_ls *ls)
+{
+	struct dlm_recover *rv, *rv_old;
+	struct dlm_config_node *nodes = NULL;
+	int error, count;
+
+	rv = kzalloc_obj(*rv, GFP_NOFS);
+	if (!rv)
+		return -ENOMEM;
+
+	error = dlm_config_nodes(ls->ls_name, &nodes, &count);
+	if (error < 0)
+		goto fail_rv;
+
+	spin_lock_bh(&ls->ls_recover_lock);
+
+	/* the lockspace needs to be stopped before it can be started */
+
+	if (!dlm_locking_stopped(ls)) {
+		spin_unlock_bh(&ls->ls_recover_lock);
+		log_error(ls, "start ignored: lockspace running");
+		error = -EINVAL;
+		goto fail;
+	}
+
+	rv->nodes = nodes;
+	rv->nodes_count = count;
+	rv->seq = ++ls->ls_recover_seq;
+	rv_old = ls->ls_recover_args;
+	ls->ls_recover_args = rv;
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+	if (rv_old) {
+		log_error(ls, "unused recovery %llx %d",
+			  (unsigned long long)rv_old->seq, rv_old->nodes_count);
+		kfree(rv_old->nodes);
+		kfree(rv_old);
+	}
+
+	set_bit(LSFL_RECOVER_WORK, &ls->ls_flags);
+	wake_up_process(ls->ls_recoverd_task);
+	return 0;
+
+ fail:
+	kfree(nodes);
+ fail_rv:
+	kfree(rv);
+	return error;
+}
+

linux-7.2.3/fs/dlm/member.h

@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2011 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __MEMBER_DOT_H__
+#define __MEMBER_DOT_H__
+
+int dlm_ls_stop(struct dlm_ls *ls);
+int dlm_ls_start(struct dlm_ls *ls);
+void dlm_clear_members(struct dlm_ls *ls);
+void dlm_clear_members_gone(struct dlm_ls *ls);
+int dlm_recover_members(struct dlm_ls *ls, struct dlm_recover *rv,int *neg_out);
+int dlm_is_removed(struct dlm_ls *ls, int nodeid);
+int dlm_is_member(struct dlm_ls *ls, int nodeid);
+int dlm_slots_version(const struct dlm_header *h);
+void dlm_slot_save(struct dlm_ls *ls, struct dlm_rcom *rc,
+		   struct dlm_member *memb);
+void dlm_slots_copy_out(struct dlm_ls *ls, struct dlm_rcom *rc);
+int dlm_slots_copy_in(struct dlm_ls *ls);
+int dlm_slots_assign(struct dlm_ls *ls, int *num_slots, int *slots_size,
+		     struct dlm_slot **slots_out, uint32_t *gen_out);
+void dlm_lsop_recover_done(struct dlm_ls *ls);
+
+#endif                          /* __MEMBER_DOT_H__ */
+

linux-7.2.3/fs/dlm/memory.c

@@ -0,0 +1,179 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "midcomms.h"
+#include "lowcomms.h"
+#include "config.h"
+#include "memory.h"
+#include "ast.h"
+
+static struct kmem_cache *writequeue_cache;
+static struct kmem_cache *mhandle_cache;
+static struct kmem_cache *msg_cache;
+static struct kmem_cache *lkb_cache;
+static struct kmem_cache *rsb_cache;
+static struct kmem_cache *cb_cache;
+
+
+int __init dlm_memory_init(void)
+{
+	writequeue_cache = dlm_lowcomms_writequeue_cache_create();
+	if (!writequeue_cache)
+		goto out;
+
+	mhandle_cache = dlm_midcomms_cache_create();
+	if (!mhandle_cache)
+		goto mhandle;
+
+	lkb_cache = kmem_cache_create("dlm_lkb", sizeof(struct dlm_lkb),
+				__alignof__(struct dlm_lkb), 0, NULL);
+	if (!lkb_cache)
+		goto lkb;
+
+	msg_cache = dlm_lowcomms_msg_cache_create();
+	if (!msg_cache)
+		goto msg;
+
+	rsb_cache = kmem_cache_create("dlm_rsb", sizeof(struct dlm_rsb),
+				__alignof__(struct dlm_rsb), 0, NULL);
+	if (!rsb_cache)
+		goto rsb;
+
+	cb_cache = kmem_cache_create_usercopy("dlm_cb", sizeof(struct dlm_callback),
+				     __alignof__(struct dlm_callback), 0,
+					 offsetof(struct dlm_callback, lvbptr),
+					 sizeof_field(struct dlm_callback, lvbptr),
+				     NULL);
+	if (!cb_cache)
+		goto cb;
+
+	return 0;
+
+cb:
+	kmem_cache_destroy(rsb_cache);
+rsb:
+	kmem_cache_destroy(msg_cache);
+msg:
+	kmem_cache_destroy(lkb_cache);
+lkb:
+	kmem_cache_destroy(mhandle_cache);
+mhandle:
+	kmem_cache_destroy(writequeue_cache);
+out:
+	return -ENOMEM;
+}
+
+void dlm_memory_exit(void)
+{
+	rcu_barrier();
+
+	kmem_cache_destroy(writequeue_cache);
+	kmem_cache_destroy(mhandle_cache);
+	kmem_cache_destroy(msg_cache);
+	kmem_cache_destroy(lkb_cache);
+	kmem_cache_destroy(rsb_cache);
+	kmem_cache_destroy(cb_cache);
+}
+
+char *dlm_allocate_lvb(struct dlm_ls *ls)
+{
+	return kzalloc(ls->ls_lvblen, GFP_ATOMIC);
+}
+
+void dlm_free_lvb(char *p)
+{
+	kfree(p);
+}
+
+struct dlm_rsb *dlm_allocate_rsb(void)
+{
+	return kmem_cache_zalloc(rsb_cache, GFP_ATOMIC);
+}
+
+static void __free_rsb_rcu(struct rcu_head *rcu)
+{
+	struct dlm_rsb *r = container_of(rcu, struct dlm_rsb, rcu);
+	if (r->res_lvbptr)
+		dlm_free_lvb(r->res_lvbptr);
+	kmem_cache_free(rsb_cache, r);
+}
+
+void dlm_free_rsb(struct dlm_rsb *r)
+{
+	call_rcu(&r->rcu, __free_rsb_rcu);
+}
+
+struct dlm_lkb *dlm_allocate_lkb(void)
+{
+	return kmem_cache_zalloc(lkb_cache, GFP_ATOMIC);
+}
+
+static void __free_lkb_rcu(struct rcu_head *rcu)
+{
+	struct dlm_lkb *lkb = container_of(rcu, struct dlm_lkb, rcu);
+
+	if (test_bit(DLM_DFL_USER_BIT, &lkb->lkb_dflags)) {
+		struct dlm_user_args *ua;
+		ua = lkb->lkb_ua;
+		if (ua) {
+			kfree(ua->lksb.sb_lvbptr);
+			kfree(ua);
+		}
+	}
+
+	kmem_cache_free(lkb_cache, lkb);
+}
+
+void dlm_free_lkb(struct dlm_lkb *lkb)
+{
+	call_rcu(&lkb->rcu, __free_lkb_rcu);
+}
+
+struct dlm_mhandle *dlm_allocate_mhandle(void)
+{
+	return kmem_cache_alloc(mhandle_cache, GFP_ATOMIC);
+}
+
+void dlm_free_mhandle(struct dlm_mhandle *mhandle)
+{
+	kmem_cache_free(mhandle_cache, mhandle);
+}
+
+struct writequeue_entry *dlm_allocate_writequeue(void)
+{
+	return kmem_cache_alloc(writequeue_cache, GFP_ATOMIC);
+}
+
+void dlm_free_writequeue(struct writequeue_entry *writequeue)
+{
+	kmem_cache_free(writequeue_cache, writequeue);
+}
+
+struct dlm_msg *dlm_allocate_msg(void)
+{
+	return kmem_cache_alloc(msg_cache, GFP_ATOMIC);
+}
+
+void dlm_free_msg(struct dlm_msg *msg)
+{
+	kmem_cache_free(msg_cache, msg);
+}
+
+struct dlm_callback *dlm_allocate_cb(void)
+{
+	return kmem_cache_alloc(cb_cache, GFP_ATOMIC);
+}
+
+void dlm_free_cb(struct dlm_callback *cb)
+{
+	kmem_cache_free(cb_cache, cb);
+}

linux-7.2.3/fs/dlm/memory.h

@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __MEMORY_DOT_H__
+#define __MEMORY_DOT_H__
+
+int dlm_memory_init(void);
+void dlm_memory_exit(void);
+struct dlm_rsb *dlm_allocate_rsb(void);
+void dlm_free_rsb(struct dlm_rsb *r);
+struct dlm_lkb *dlm_allocate_lkb(void);
+void dlm_free_lkb(struct dlm_lkb *l);
+char *dlm_allocate_lvb(struct dlm_ls *ls);
+void dlm_free_lvb(char *l);
+struct dlm_mhandle *dlm_allocate_mhandle(void);
+void dlm_free_mhandle(struct dlm_mhandle *mhandle);
+struct writequeue_entry *dlm_allocate_writequeue(void);
+void dlm_free_writequeue(struct writequeue_entry *writequeue);
+struct dlm_msg *dlm_allocate_msg(void);
+void dlm_free_msg(struct dlm_msg *msg);
+struct dlm_callback *dlm_allocate_cb(void);
+void dlm_free_cb(struct dlm_callback *cb);
+
+#endif		/* __MEMORY_DOT_H__ */
+

linux-7.2.3/fs/dlm/midcomms.h

@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __MIDCOMMS_DOT_H__
+#define __MIDCOMMS_DOT_H__
+
+struct midcomms_node;
+
+int dlm_validate_incoming_buffer(int nodeid, unsigned char *buf, int len);
+int dlm_process_incoming_buffer(int nodeid, unsigned char *buf, int buflen);
+struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len, char **ppc);
+void dlm_midcomms_commit_mhandle(struct dlm_mhandle *mh, const void *name,
+				 int namelen);
+int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr);
+void dlm_midcomms_version_wait(void);
+int dlm_midcomms_close(int nodeid);
+int dlm_midcomms_start(void);
+void dlm_midcomms_stop(void);
+void dlm_midcomms_init(void);
+void dlm_midcomms_exit(void);
+void dlm_midcomms_shutdown(void);
+void dlm_midcomms_add_member(int nodeid);
+void dlm_midcomms_remove_member(int nodeid);
+void dlm_midcomms_unack_msg_resend(int nodeid);
+const char *dlm_midcomms_state(struct midcomms_node *node);
+unsigned long dlm_midcomms_flags(struct midcomms_node *node);
+int dlm_midcomms_send_queue_cnt(struct midcomms_node *node);
+uint32_t dlm_midcomms_version(struct midcomms_node *node);
+int dlm_midcomms_rawmsg_send(struct midcomms_node *node, void *buf,
+			     int buflen);
+struct kmem_cache *dlm_midcomms_cache_create(void);
+
+#endif				/* __MIDCOMMS_DOT_H__ */
+

linux-7.2.3/fs/dlm/plock.c

@@ -0,0 +1,630 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
+ */
+
+#include <linux/fs.h>
+#include <linux/filelock.h>
+#include <linux/miscdevice.h>
+#include <linux/poll.h>
+#include <linux/dlm.h>
+#include <linux/dlm_plock.h>
+#include <linux/slab.h>
+
+#include <trace/events/dlm.h>
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+
+static DEFINE_SPINLOCK(ops_lock);
+static LIST_HEAD(send_list);
+static LIST_HEAD(recv_list);
+static DECLARE_WAIT_QUEUE_HEAD(send_wq);
+static DECLARE_WAIT_QUEUE_HEAD(recv_wq);
+
+struct plock_async_data {
+	void *fl;
+	void *file;
+	struct file_lock flc;
+	int (*callback)(struct file_lock *fl, int result);
+};
+
+struct plock_op {
+	struct list_head list;
+	int done;
+	struct dlm_plock_info info;
+	/* if set indicates async handling */
+	struct plock_async_data *data;
+};
+
+static inline void set_version(struct dlm_plock_info *info)
+{
+	info->version[0] = DLM_PLOCK_VERSION_MAJOR;
+	info->version[1] = DLM_PLOCK_VERSION_MINOR;
+	info->version[2] = DLM_PLOCK_VERSION_PATCH;
+}
+
+static struct plock_op *plock_lookup_waiter(const struct dlm_plock_info *info)
+{
+	struct plock_op *op = NULL, *iter;
+
+	list_for_each_entry(iter, &recv_list, list) {
+		if (iter->info.fsid == info->fsid &&
+		    iter->info.number == info->number &&
+		    iter->info.owner == info->owner &&
+		    iter->info.pid == info->pid &&
+		    iter->info.start == info->start &&
+		    iter->info.end == info->end &&
+		    iter->info.ex == info->ex &&
+		    iter->info.wait) {
+			op = iter;
+			break;
+		}
+	}
+
+	return op;
+}
+
+static int check_version(struct dlm_plock_info *info)
+{
+	if ((DLM_PLOCK_VERSION_MAJOR != info->version[0]) ||
+	    (DLM_PLOCK_VERSION_MINOR < info->version[1])) {
+		log_print("plock device version mismatch: "
+			  "kernel (%u.%u.%u), user (%u.%u.%u)",
+			  DLM_PLOCK_VERSION_MAJOR,
+			  DLM_PLOCK_VERSION_MINOR,
+			  DLM_PLOCK_VERSION_PATCH,
+			  info->version[0],
+			  info->version[1],
+			  info->version[2]);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static void dlm_release_plock_op(struct plock_op *op)
+{
+	kfree(op->data);
+	kfree(op);
+}
+
+static void send_op(struct plock_op *op)
+{
+	set_version(&op->info);
+	spin_lock(&ops_lock);
+	list_add_tail(&op->list, &send_list);
+	spin_unlock(&ops_lock);
+	wake_up(&send_wq);
+}
+
+static int do_lock_cancel(const struct dlm_plock_info *orig_info)
+{
+	struct plock_op *op;
+	int rv;
+
+	op = kzalloc_obj(*op, GFP_NOFS);
+	if (!op)
+		return -ENOMEM;
+
+	op->info = *orig_info;
+	op->info.optype = DLM_PLOCK_OP_CANCEL;
+	op->info.wait = 0;
+
+	send_op(op);
+	wait_event(recv_wq, (op->done != 0));
+
+	rv = op->info.rv;
+
+	dlm_release_plock_op(op);
+	return rv;
+}
+
+int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
+		   int cmd, struct file_lock *fl)
+{
+	struct plock_async_data *op_data;
+	struct dlm_ls *ls;
+	struct plock_op *op;
+	int rv;
+
+	ls = dlm_find_lockspace_local(lockspace);
+	if (!ls)
+		return -EINVAL;
+
+	op = kzalloc_obj(*op, GFP_NOFS);
+	if (!op) {
+		rv = -ENOMEM;
+		goto out;
+	}
+
+	op->info.optype		= DLM_PLOCK_OP_LOCK;
+	op->info.pid		= fl->c.flc_pid;
+	op->info.ex		= lock_is_write(fl);
+	op->info.wait		= !!(fl->c.flc_flags & FL_SLEEP);
+	op->info.fsid		= ls->ls_global_id;
+	op->info.number		= number;
+	op->info.start		= fl->fl_start;
+	op->info.end		= fl->fl_end;
+	op->info.owner = (__u64)(long) fl->c.flc_owner;
+	/* async handling */
+	if (fl->fl_lmops && fl->fl_lmops->lm_grant) {
+		op_data = kzalloc_obj(*op_data, GFP_NOFS);
+		if (!op_data) {
+			dlm_release_plock_op(op);
+			rv = -ENOMEM;
+			goto out;
+		}
+
+		op_data->callback = fl->fl_lmops->lm_grant;
+		locks_init_lock(&op_data->flc);
+		locks_copy_lock(&op_data->flc, fl);
+		op_data->fl		= fl;
+		op_data->file	= file;
+
+		op->data = op_data;
+
+		send_op(op);
+		rv = FILE_LOCK_DEFERRED;
+		goto out;
+	}
+
+	send_op(op);
+
+	if (op->info.wait) {
+		rv = wait_event_interruptible(recv_wq, (op->done != 0));
+		if (rv == -ERESTARTSYS) {
+			spin_lock(&ops_lock);
+			/* recheck under ops_lock if we got a done != 0,
+			 * if so this interrupt case should be ignored
+			 */
+			if (op->done != 0) {
+				spin_unlock(&ops_lock);
+				goto do_lock_wait;
+			}
+			spin_unlock(&ops_lock);
+
+			rv = do_lock_cancel(&op->info);
+			switch (rv) {
+			case 0:
+				/* waiter was deleted in user space, answer will never come
+				 * remove original request. The original request must be
+				 * on recv_list because the answer of do_lock_cancel()
+				 * synchronized it.
+				 */
+				spin_lock(&ops_lock);
+				list_del(&op->list);
+				spin_unlock(&ops_lock);
+				rv = -EINTR;
+				break;
+			case -ENOENT:
+				/* cancellation wasn't successful but op should be done */
+				fallthrough;
+			default:
+				/* internal error doing cancel we need to wait */
+				goto wait;
+			}
+
+			log_debug(ls, "%s: wait interrupted %x %llx pid %d",
+				  __func__, ls->ls_global_id,
+				  (unsigned long long)number, op->info.pid);
+			dlm_release_plock_op(op);
+			goto out;
+		}
+	} else {
+wait:
+		wait_event(recv_wq, (op->done != 0));
+	}
+
+do_lock_wait:
+
+	WARN_ON(!list_empty(&op->list));
+
+	rv = op->info.rv;
+
+	if (!rv) {
+		if (locks_lock_file_wait(file, fl) < 0)
+			log_error(ls, "dlm_posix_lock: vfs lock error %llx",
+				  (unsigned long long)number);
+	}
+
+	dlm_release_plock_op(op);
+out:
+	dlm_put_lockspace(ls);
+	return rv;
+}
+EXPORT_SYMBOL_GPL(dlm_posix_lock);
+
+/* Returns failure iff a successful lock operation should be canceled */
+static int dlm_plock_callback(struct plock_op *op)
+{
+	struct plock_async_data *op_data = op->data;
+	struct file *file;
+	struct file_lock *fl;
+	struct file_lock *flc;
+	int (*notify)(struct file_lock *fl, int result) = NULL;
+	int rv = 0;
+
+	WARN_ON(!list_empty(&op->list));
+
+	/* check if the following 2 are still valid or make a copy */
+	file = op_data->file;
+	flc = &op_data->flc;
+	fl = op_data->fl;
+	notify = op_data->callback;
+
+	if (op->info.rv) {
+		notify(fl, op->info.rv);
+		goto out;
+	}
+
+	/* got fs lock; bookkeep locally as well: */
+	flc->c.flc_flags &= ~FL_SLEEP;
+	if (posix_lock_file(file, flc, NULL)) {
+		/*
+		 * This can only happen in the case of kmalloc() failure.
+		 * The filesystem's own lock is the authoritative lock,
+		 * so a failure to get the lock locally is not a disaster.
+		 * As long as the fs cannot reliably cancel locks (especially
+		 * in a low-memory situation), we're better off ignoring
+		 * this failure than trying to recover.
+		 */
+		log_print("dlm_plock_callback: vfs lock error %llx file %p fl %p",
+			  (unsigned long long)op->info.number, file, fl);
+	}
+
+	rv = notify(fl, 0);
+	if (rv) {
+		/* XXX: We need to cancel the fs lock here: */
+		log_print("%s: lock granted after lock request failed; dangling lock!",
+			  __func__);
+		goto out;
+	}
+
+out:
+	dlm_release_plock_op(op);
+	return rv;
+}
+
+int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
+		     struct file_lock *fl)
+{
+	struct dlm_ls *ls;
+	struct plock_op *op;
+	int rv;
+	unsigned char saved_flags = fl->c.flc_flags;
+
+	ls = dlm_find_lockspace_local(lockspace);
+	if (!ls)
+		return -EINVAL;
+
+	op = kzalloc_obj(*op, GFP_NOFS);
+	if (!op) {
+		rv = -ENOMEM;
+		goto out;
+	}
+
+	/* cause the vfs unlock to return ENOENT if lock is not found */
+	fl->c.flc_flags |= FL_EXISTS;
+
+	rv = locks_lock_file_wait(file, fl);
+	if (rv == -ENOENT) {
+		rv = 0;
+		goto out_free;
+	}
+	if (rv < 0) {
+		log_error(ls, "dlm_posix_unlock: vfs unlock error %d %llx",
+			  rv, (unsigned long long)number);
+	}
+
+	op->info.optype		= DLM_PLOCK_OP_UNLOCK;
+	op->info.pid		= fl->c.flc_pid;
+	op->info.fsid		= ls->ls_global_id;
+	op->info.number		= number;
+	op->info.start		= fl->fl_start;
+	op->info.end		= fl->fl_end;
+	op->info.owner = (__u64)(long) fl->c.flc_owner;
+
+	if (fl->c.flc_flags & FL_CLOSE) {
+		op->info.flags |= DLM_PLOCK_FL_CLOSE;
+		send_op(op);
+		rv = 0;
+		goto out;
+	}
+
+	send_op(op);
+	wait_event(recv_wq, (op->done != 0));
+
+	WARN_ON(!list_empty(&op->list));
+
+	rv = op->info.rv;
+
+	if (rv == -ENOENT)
+		rv = 0;
+
+out_free:
+	dlm_release_plock_op(op);
+out:
+	dlm_put_lockspace(ls);
+	fl->c.flc_flags = saved_flags;
+	return rv;
+}
+EXPORT_SYMBOL_GPL(dlm_posix_unlock);
+
+/*
+ * NOTE: This implementation can only handle async lock requests as nfs
+ * do it. It cannot handle cancellation of a pending lock request sitting
+ * in wait_event(), but for now only nfs is the only user local kernel
+ * user.
+ */
+int dlm_posix_cancel(dlm_lockspace_t *lockspace, u64 number, struct file *file,
+		     struct file_lock *fl)
+{
+	struct dlm_plock_info info;
+	struct plock_op *op;
+	struct dlm_ls *ls;
+	int rv;
+
+	/* this only works for async request for now and nfs is the only
+	 * kernel user right now.
+	 */
+	if (WARN_ON_ONCE(!fl->fl_lmops || !fl->fl_lmops->lm_grant))
+		return -EOPNOTSUPP;
+
+	ls = dlm_find_lockspace_local(lockspace);
+	if (!ls)
+		return -EINVAL;
+
+	memset(&info, 0, sizeof(info));
+	info.pid = fl->c.flc_pid;
+	info.ex = lock_is_write(fl);
+	info.fsid = ls->ls_global_id;
+	dlm_put_lockspace(ls);
+	info.number = number;
+	info.start = fl->fl_start;
+	info.end = fl->fl_end;
+	info.owner = (__u64)(long) fl->c.flc_owner;
+
+	rv = do_lock_cancel(&info);
+	switch (rv) {
+	case 0:
+		spin_lock(&ops_lock);
+		/* lock request to cancel must be on recv_list because
+		 * do_lock_cancel() synchronizes it.
+		 */
+		op = plock_lookup_waiter(&info);
+		if (WARN_ON_ONCE(!op)) {
+			spin_unlock(&ops_lock);
+			rv = -ENOLCK;
+			break;
+		}
+
+		list_del(&op->list);
+		spin_unlock(&ops_lock);
+		WARN_ON(op->info.optype != DLM_PLOCK_OP_LOCK);
+		op->data->callback(op->data->fl, -EINTR);
+		dlm_release_plock_op(op);
+		rv = -EINTR;
+		break;
+	case -ENOENT:
+		/* if cancel wasn't successful we probably were to late
+		 * or it was a non-blocking lock request, so just unlock it.
+		 */
+		rv = dlm_posix_unlock(lockspace, number, file, fl);
+		break;
+	default:
+		break;
+	}
+
+	return rv;
+}
+EXPORT_SYMBOL_GPL(dlm_posix_cancel);
+
+int dlm_posix_get(dlm_lockspace_t *lockspace, u64 number, struct file *file,
+		  struct file_lock *fl)
+{
+	struct dlm_ls *ls;
+	struct plock_op *op;
+	int rv;
+
+	ls = dlm_find_lockspace_local(lockspace);
+	if (!ls)
+		return -EINVAL;
+
+	op = kzalloc_obj(*op, GFP_NOFS);
+	if (!op) {
+		rv = -ENOMEM;
+		goto out;
+	}
+
+	op->info.optype		= DLM_PLOCK_OP_GET;
+	op->info.pid		= fl->c.flc_pid;
+	op->info.ex		= lock_is_write(fl);
+	op->info.fsid		= ls->ls_global_id;
+	op->info.number		= number;
+	op->info.start		= fl->fl_start;
+	op->info.end		= fl->fl_end;
+	op->info.owner = (__u64)(long) fl->c.flc_owner;
+
+	send_op(op);
+	wait_event(recv_wq, (op->done != 0));
+
+	WARN_ON(!list_empty(&op->list));
+
+	/* info.rv from userspace is 1 for conflict, 0 for no-conflict,
+	   -ENOENT if there are no locks on the file */
+
+	rv = op->info.rv;
+
+	fl->c.flc_type = F_UNLCK;
+	if (rv == -ENOENT)
+		rv = 0;
+	else if (rv > 0) {
+		locks_init_lock(fl);
+		fl->c.flc_type = (op->info.ex) ? F_WRLCK : F_RDLCK;
+		fl->c.flc_flags = FL_POSIX;
+		fl->c.flc_pid = op->info.pid;
+		if (op->info.nodeid != dlm_our_nodeid())
+			fl->c.flc_pid = -fl->c.flc_pid;
+		fl->fl_start = op->info.start;
+		fl->fl_end = op->info.end;
+		rv = 0;
+	}
+
+	dlm_release_plock_op(op);
+out:
+	dlm_put_lockspace(ls);
+	return rv;
+}
+EXPORT_SYMBOL_GPL(dlm_posix_get);
+
+/* a read copies out one plock request from the send list */
+static ssize_t dev_read(struct file *file, char __user *u, size_t count,
+			loff_t *ppos)
+{
+	struct dlm_plock_info info;
+	struct plock_op *op = NULL;
+
+	if (count < sizeof(info))
+		return -EINVAL;
+
+	spin_lock(&ops_lock);
+	if (!list_empty(&send_list)) {
+		op = list_first_entry(&send_list, struct plock_op, list);
+		if (op->info.flags & DLM_PLOCK_FL_CLOSE)
+			list_del(&op->list);
+		else
+			list_move_tail(&op->list, &recv_list);
+		memcpy(&info, &op->info, sizeof(info));
+	}
+	spin_unlock(&ops_lock);
+
+	if (!op)
+		return -EAGAIN;
+
+	trace_dlm_plock_read(&info);
+
+	/* there is no need to get a reply from userspace for unlocks
+	   that were generated by the vfs cleaning up for a close
+	   (the process did not make an unlock call). */
+
+	if (op->info.flags & DLM_PLOCK_FL_CLOSE)
+		dlm_release_plock_op(op);
+
+	if (copy_to_user(u, &info, sizeof(info)))
+		return -EFAULT;
+	return sizeof(info);
+}
+
+/* a write copies in one plock result that should match a plock_op
+   on the recv list */
+static ssize_t dev_write(struct file *file, const char __user *u, size_t count,
+			 loff_t *ppos)
+{
+	struct plock_op *op = NULL, *iter;
+	struct dlm_plock_info info;
+	int do_callback = 0;
+
+	if (count != sizeof(info))
+		return -EINVAL;
+
+	if (copy_from_user(&info, u, sizeof(info)))
+		return -EFAULT;
+
+	trace_dlm_plock_write(&info);
+
+	if (check_version(&info))
+		return -EINVAL;
+
+	/*
+	 * The results for waiting ops (SETLKW) can be returned in any
+	 * order, so match all fields to find the op.  The results for
+	 * non-waiting ops are returned in the order that they were sent
+	 * to userspace, so match the result with the first non-waiting op.
+	 */
+	spin_lock(&ops_lock);
+	if (info.wait) {
+		op = plock_lookup_waiter(&info);
+	} else {
+		list_for_each_entry(iter, &recv_list, list) {
+			if (!iter->info.wait &&
+			    iter->info.fsid == info.fsid) {
+				op = iter;
+				break;
+			}
+		}
+	}
+
+	if (op) {
+		/* Sanity check that op and info match. */
+		if (info.wait)
+			WARN_ON(op->info.optype != DLM_PLOCK_OP_LOCK);
+		else
+			WARN_ON(op->info.number != info.number ||
+				op->info.owner != info.owner ||
+				op->info.optype != info.optype);
+
+		list_del_init(&op->list);
+		memcpy(&op->info, &info, sizeof(info));
+		if (op->data)
+			do_callback = 1;
+		else
+			op->done = 1;
+	}
+	spin_unlock(&ops_lock);
+
+	if (op) {
+		if (do_callback)
+			dlm_plock_callback(op);
+		else
+			wake_up(&recv_wq);
+	} else
+		pr_debug("%s: no op %x %llx", __func__,
+			 info.fsid, (unsigned long long)info.number);
+	return count;
+}
+
+static __poll_t dev_poll(struct file *file, poll_table *wait)
+{
+	__poll_t mask = 0;
+
+	poll_wait(file, &send_wq, wait);
+
+	spin_lock(&ops_lock);
+	if (!list_empty(&send_list))
+		mask = EPOLLIN | EPOLLRDNORM;
+	spin_unlock(&ops_lock);
+
+	return mask;
+}
+
+static const struct file_operations dev_fops = {
+	.read    = dev_read,
+	.write   = dev_write,
+	.poll    = dev_poll,
+	.owner   = THIS_MODULE,
+	.llseek  = noop_llseek,
+};
+
+static struct miscdevice plock_dev_misc = {
+	.minor = MISC_DYNAMIC_MINOR,
+	.name = DLM_PLOCK_MISC_NAME,
+	.fops = &dev_fops
+};
+
+int dlm_plock_init(void)
+{
+	int rv;
+
+	rv = misc_register(&plock_dev_misc);
+	if (rv)
+		log_print("dlm_plock_init: misc_register failed %d", rv);
+	return rv;
+}
+
+void dlm_plock_exit(void)
+{
+	misc_deregister(&plock_dev_misc);
+	WARN_ON(!list_empty(&send_list));
+	WARN_ON(!list_empty(&recv_list));
+}
+

linux-7.2.3/fs/dlm/rcom.c

@@ -0,0 +1,691 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "member.h"
+#include "lowcomms.h"
+#include "midcomms.h"
+#include "rcom.h"
+#include "recover.h"
+#include "dir.h"
+#include "config.h"
+#include "memory.h"
+#include "lock.h"
+#include "util.h"
+
+static int rcom_response(struct dlm_ls *ls)
+{
+	return test_bit(LSFL_RCOM_READY, &ls->ls_flags);
+}
+
+static void _create_rcom(struct dlm_ls *ls, int to_nodeid, int type, int len,
+			 struct dlm_rcom **rc_ret, char *mb, int mb_len,
+			 uint64_t seq)
+{
+	struct dlm_rcom *rc;
+
+	rc = (struct dlm_rcom *) mb;
+
+	rc->rc_header.h_version = cpu_to_le32(DLM_HEADER_MAJOR | DLM_HEADER_MINOR);
+	rc->rc_header.u.h_lockspace = cpu_to_le32(ls->ls_global_id);
+	rc->rc_header.h_nodeid = cpu_to_le32(dlm_our_nodeid());
+	rc->rc_header.h_length = cpu_to_le16(mb_len);
+	rc->rc_header.h_cmd = DLM_RCOM;
+
+	rc->rc_type = cpu_to_le32(type);
+	rc->rc_seq = cpu_to_le64(seq);
+
+	*rc_ret = rc;
+}
+
+static int create_rcom(struct dlm_ls *ls, int to_nodeid, int type, int len,
+		       struct dlm_rcom **rc_ret, struct dlm_mhandle **mh_ret,
+		       uint64_t seq)
+{
+	int mb_len = sizeof(struct dlm_rcom) + len;
+	struct dlm_mhandle *mh;
+	char *mb;
+
+	mh = dlm_midcomms_get_mhandle(to_nodeid, mb_len, &mb);
+	if (!mh) {
+		log_print("%s to %d type %d len %d ENOBUFS",
+			  __func__, to_nodeid, type, len);
+		return -ENOBUFS;
+	}
+
+	_create_rcom(ls, to_nodeid, type, len, rc_ret, mb, mb_len, seq);
+	*mh_ret = mh;
+	return 0;
+}
+
+static int create_rcom_stateless(struct dlm_ls *ls, int to_nodeid, int type,
+				 int len, struct dlm_rcom **rc_ret,
+				 struct dlm_msg **msg_ret, uint64_t seq)
+{
+	int mb_len = sizeof(struct dlm_rcom) + len;
+	struct dlm_msg *msg;
+	char *mb;
+
+	msg = dlm_lowcomms_new_msg(to_nodeid, mb_len, &mb, NULL, NULL);
+	if (!msg) {
+		log_print("create_rcom to %d type %d len %d ENOBUFS",
+			  to_nodeid, type, len);
+		return -ENOBUFS;
+	}
+
+	_create_rcom(ls, to_nodeid, type, len, rc_ret, mb, mb_len, seq);
+	*msg_ret = msg;
+	return 0;
+}
+
+static void send_rcom(struct dlm_mhandle *mh, struct dlm_rcom *rc)
+{
+	dlm_midcomms_commit_mhandle(mh, NULL, 0);
+}
+
+static void send_rcom_stateless(struct dlm_msg *msg, struct dlm_rcom *rc)
+{
+	dlm_lowcomms_commit_msg(msg);
+	dlm_lowcomms_put_msg(msg);
+}
+
+static void set_rcom_status(struct dlm_ls *ls, struct rcom_status *rs,
+			    uint32_t flags)
+{
+	rs->rs_flags = cpu_to_le32(flags);
+}
+
+/* When replying to a status request, a node also sends back its
+   configuration values.  The requesting node then checks that the remote
+   node is configured the same way as itself. */
+
+static void set_rcom_config(struct dlm_ls *ls, struct rcom_config *rf,
+			    uint32_t num_slots)
+{
+	rf->rf_lvblen = cpu_to_le32(ls->ls_lvblen);
+	rf->rf_lsflags = cpu_to_le32(ls->ls_exflags);
+
+	rf->rf_our_slot = cpu_to_le16(ls->ls_slot);
+	rf->rf_num_slots = cpu_to_le16(num_slots);
+	rf->rf_generation =  cpu_to_le32(ls->ls_generation);
+}
+
+static int check_rcom_config(struct dlm_ls *ls, struct dlm_rcom *rc, int nodeid)
+{
+	struct rcom_config *rf = (struct rcom_config *) rc->rc_buf;
+
+	if ((le32_to_cpu(rc->rc_header.h_version) & 0xFFFF0000) != DLM_HEADER_MAJOR) {
+		log_error(ls, "version mismatch: %x nodeid %d: %x",
+			  DLM_HEADER_MAJOR | DLM_HEADER_MINOR, nodeid,
+			  le32_to_cpu(rc->rc_header.h_version));
+		return -EPROTO;
+	}
+
+	if (le32_to_cpu(rf->rf_lvblen) != ls->ls_lvblen ||
+	    le32_to_cpu(rf->rf_lsflags) != ls->ls_exflags) {
+		log_error(ls, "config mismatch: %d,%x nodeid %d: %d,%x",
+			  ls->ls_lvblen, ls->ls_exflags, nodeid,
+			  le32_to_cpu(rf->rf_lvblen),
+			  le32_to_cpu(rf->rf_lsflags));
+		return -EPROTO;
+	}
+	return 0;
+}
+
+static void allow_sync_reply(struct dlm_ls *ls, __le64 *new_seq)
+{
+	spin_lock_bh(&ls->ls_rcom_spin);
+	*new_seq = cpu_to_le64(++ls->ls_rcom_seq);
+	set_bit(LSFL_RCOM_WAIT, &ls->ls_flags);
+	spin_unlock_bh(&ls->ls_rcom_spin);
+}
+
+static void disallow_sync_reply(struct dlm_ls *ls)
+{
+	spin_lock_bh(&ls->ls_rcom_spin);
+	clear_bit(LSFL_RCOM_WAIT, &ls->ls_flags);
+	clear_bit(LSFL_RCOM_READY, &ls->ls_flags);
+	spin_unlock_bh(&ls->ls_rcom_spin);
+}
+
+/*
+ * low nodeid gathers one slot value at a time from each node.
+ * it sets need_slots=0, and saves rf_our_slot returned from each
+ * rcom_config.
+ *
+ * other nodes gather all slot values at once from the low nodeid.
+ * they set need_slots=1, and ignore the rf_our_slot returned from each
+ * rcom_config.  they use the rf_num_slots returned from the low
+ * node's rcom_config.
+ */
+
+int dlm_rcom_status(struct dlm_ls *ls, int nodeid, uint32_t status_flags,
+		    uint64_t seq)
+{
+	struct dlm_rcom *rc;
+	struct dlm_msg *msg;
+	int error = 0;
+
+	ls->ls_recover_nodeid = nodeid;
+
+	if (nodeid == dlm_our_nodeid()) {
+		rc = ls->ls_recover_buf;
+		rc->rc_result = cpu_to_le32(dlm_recover_status(ls));
+		goto out;
+	}
+
+retry:
+	error = create_rcom_stateless(ls, nodeid, DLM_RCOM_STATUS,
+				      sizeof(struct rcom_status), &rc, &msg,
+				      seq);
+	if (error)
+		goto out;
+
+	set_rcom_status(ls, (struct rcom_status *)rc->rc_buf, status_flags);
+
+	allow_sync_reply(ls, &rc->rc_id);
+	memset(ls->ls_recover_buf, 0, DLM_MAX_SOCKET_BUFSIZE);
+
+	send_rcom_stateless(msg, rc);
+
+	error = dlm_wait_function(ls, &rcom_response);
+	disallow_sync_reply(ls);
+	if (error == -ETIMEDOUT)
+		goto retry;
+	if (error)
+		goto out;
+
+	rc = ls->ls_recover_buf;
+
+	if (rc->rc_result == cpu_to_le32(-ESRCH)) {
+		/* we pretend the remote lockspace exists with 0 status */
+		log_debug(ls, "remote node %d not ready", nodeid);
+		rc->rc_result = 0;
+		error = 0;
+	} else {
+		error = check_rcom_config(ls, rc, nodeid);
+	}
+
+	/* the caller looks at rc_result for the remote recovery status */
+ out:
+	return error;
+}
+
+static void receive_rcom_status(struct dlm_ls *ls,
+				const struct dlm_rcom *rc_in,
+				uint64_t seq)
+{
+	struct dlm_rcom *rc;
+	struct rcom_status *rs;
+	uint32_t status;
+	int nodeid = le32_to_cpu(rc_in->rc_header.h_nodeid);
+	int len = sizeof(struct rcom_config);
+	struct dlm_msg *msg;
+	int num_slots = 0;
+	int error;
+
+	if (!dlm_slots_version(&rc_in->rc_header)) {
+		status = dlm_recover_status(ls);
+		goto do_create;
+	}
+
+	rs = (struct rcom_status *)rc_in->rc_buf;
+
+	if (!(le32_to_cpu(rs->rs_flags) & DLM_RSF_NEED_SLOTS)) {
+		status = dlm_recover_status(ls);
+		goto do_create;
+	}
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	status = ls->ls_recover_status;
+	num_slots = ls->ls_num_slots;
+	spin_unlock_bh(&ls->ls_recover_lock);
+	len += num_slots * sizeof(struct rcom_slot);
+
+ do_create:
+	error = create_rcom_stateless(ls, nodeid, DLM_RCOM_STATUS_REPLY,
+				      len, &rc, &msg, seq);
+	if (error)
+		return;
+
+	rc->rc_id = rc_in->rc_id;
+	rc->rc_seq_reply = rc_in->rc_seq;
+	rc->rc_result = cpu_to_le32(status);
+
+	set_rcom_config(ls, (struct rcom_config *)rc->rc_buf, num_slots);
+
+	if (!num_slots)
+		goto do_send;
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	if (ls->ls_num_slots != num_slots) {
+		spin_unlock_bh(&ls->ls_recover_lock);
+		log_debug(ls, "receive_rcom_status num_slots %d to %d",
+			  num_slots, ls->ls_num_slots);
+		rc->rc_result = 0;
+		set_rcom_config(ls, (struct rcom_config *)rc->rc_buf, 0);
+		goto do_send;
+	}
+
+	dlm_slots_copy_out(ls, rc);
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+ do_send:
+	send_rcom_stateless(msg, rc);
+}
+
+static void receive_sync_reply(struct dlm_ls *ls, const struct dlm_rcom *rc_in)
+{
+	spin_lock_bh(&ls->ls_rcom_spin);
+	if (!test_bit(LSFL_RCOM_WAIT, &ls->ls_flags) ||
+	    le64_to_cpu(rc_in->rc_id) != ls->ls_rcom_seq) {
+		log_debug(ls, "reject reply %d from %d seq %llx expect %llx",
+			  le32_to_cpu(rc_in->rc_type),
+			  le32_to_cpu(rc_in->rc_header.h_nodeid),
+			  (unsigned long long)le64_to_cpu(rc_in->rc_id),
+			  (unsigned long long)ls->ls_rcom_seq);
+		goto out;
+	}
+	memcpy(ls->ls_recover_buf, rc_in,
+	       le16_to_cpu(rc_in->rc_header.h_length));
+	set_bit(LSFL_RCOM_READY, &ls->ls_flags);
+	clear_bit(LSFL_RCOM_WAIT, &ls->ls_flags);
+	wake_up(&ls->ls_wait_general);
+ out:
+	spin_unlock_bh(&ls->ls_rcom_spin);
+}
+
+int dlm_rcom_names(struct dlm_ls *ls, int nodeid, char *last_name,
+		   int last_len, uint64_t seq)
+{
+	struct dlm_mhandle *mh;
+	struct dlm_rcom *rc;
+	int error = 0;
+
+	ls->ls_recover_nodeid = nodeid;
+
+retry:
+	error = create_rcom(ls, nodeid, DLM_RCOM_NAMES, last_len,
+			    &rc, &mh, seq);
+	if (error)
+		goto out;
+	memcpy(rc->rc_buf, last_name, last_len);
+
+	allow_sync_reply(ls, &rc->rc_id);
+	memset(ls->ls_recover_buf, 0, DLM_MAX_SOCKET_BUFSIZE);
+
+	send_rcom(mh, rc);
+
+	error = dlm_wait_function(ls, &rcom_response);
+	disallow_sync_reply(ls);
+	if (error == -ETIMEDOUT)
+		goto retry;
+ out:
+	return error;
+}
+
+static void receive_rcom_names(struct dlm_ls *ls, const struct dlm_rcom *rc_in,
+			       uint64_t seq)
+{
+	struct dlm_mhandle *mh;
+	struct dlm_rcom *rc;
+	int error, inlen, outlen, nodeid;
+
+	nodeid = le32_to_cpu(rc_in->rc_header.h_nodeid);
+	inlen = le16_to_cpu(rc_in->rc_header.h_length) -
+		sizeof(struct dlm_rcom);
+	outlen = DLM_MAX_APP_BUFSIZE - sizeof(struct dlm_rcom);
+
+	error = create_rcom(ls, nodeid, DLM_RCOM_NAMES_REPLY, outlen,
+			    &rc, &mh, seq);
+	if (error)
+		return;
+	rc->rc_id = rc_in->rc_id;
+	rc->rc_seq_reply = rc_in->rc_seq;
+
+	dlm_copy_master_names(ls, rc_in->rc_buf, inlen, rc->rc_buf, outlen,
+			      nodeid);
+	send_rcom(mh, rc);
+}
+
+int dlm_send_rcom_lookup(struct dlm_rsb *r, int dir_nodeid, uint64_t seq)
+{
+	struct dlm_rcom *rc;
+	struct dlm_mhandle *mh;
+	struct dlm_ls *ls = r->res_ls;
+	int error;
+
+	error = create_rcom(ls, dir_nodeid, DLM_RCOM_LOOKUP, r->res_length,
+			    &rc, &mh, seq);
+	if (error)
+		goto out;
+	memcpy(rc->rc_buf, r->res_name, r->res_length);
+	rc->rc_id = cpu_to_le64(r->res_id);
+
+	send_rcom(mh, rc);
+ out:
+	return error;
+}
+
+static void receive_rcom_lookup(struct dlm_ls *ls,
+				const struct dlm_rcom *rc_in, uint64_t seq)
+{
+	struct dlm_rcom *rc;
+	struct dlm_mhandle *mh;
+	int error, ret_nodeid, nodeid = le32_to_cpu(rc_in->rc_header.h_nodeid);
+	int len = le16_to_cpu(rc_in->rc_header.h_length) -
+		sizeof(struct dlm_rcom);
+
+	/* Old code would send this special id to trigger a debug dump. */
+	if (rc_in->rc_id == cpu_to_le64(0xFFFFFFFF)) {
+		log_error(ls, "receive_rcom_lookup dump from %d", nodeid);
+		dlm_dump_rsb_name(ls, rc_in->rc_buf, len);
+		return;
+	}
+
+	error = create_rcom(ls, nodeid, DLM_RCOM_LOOKUP_REPLY, 0, &rc, &mh,
+			    seq);
+	if (error)
+		return;
+
+	error = dlm_master_lookup(ls, nodeid, rc_in->rc_buf, len,
+				  DLM_LU_RECOVER_MASTER, &ret_nodeid, NULL);
+	if (error)
+		ret_nodeid = error;
+	rc->rc_result = cpu_to_le32(ret_nodeid);
+	rc->rc_id = rc_in->rc_id;
+	rc->rc_seq_reply = rc_in->rc_seq;
+
+	send_rcom(mh, rc);
+}
+
+static void receive_rcom_lookup_reply(struct dlm_ls *ls,
+				      const struct dlm_rcom *rc_in)
+{
+	dlm_recover_master_reply(ls, rc_in);
+}
+
+static void pack_rcom_lock(struct dlm_rsb *r, struct dlm_lkb *lkb,
+			   struct rcom_lock *rl)
+{
+	memset(rl, 0, sizeof(*rl));
+
+	rl->rl_ownpid = cpu_to_le32(lkb->lkb_ownpid);
+	rl->rl_lkid = cpu_to_le32(lkb->lkb_id);
+	rl->rl_exflags = cpu_to_le32(lkb->lkb_exflags);
+	rl->rl_flags = cpu_to_le32(dlm_dflags_val(lkb));
+	rl->rl_lvbseq = cpu_to_le32(lkb->lkb_lvbseq);
+	rl->rl_rqmode = lkb->lkb_rqmode;
+	rl->rl_grmode = lkb->lkb_grmode;
+	rl->rl_status = lkb->lkb_status;
+	rl->rl_wait_type = cpu_to_le16(lkb->lkb_wait_type);
+
+	if (lkb->lkb_bastfn)
+		rl->rl_asts |= DLM_CB_BAST;
+	if (lkb->lkb_astfn)
+		rl->rl_asts |= DLM_CB_CAST;
+
+	rl->rl_namelen = cpu_to_le16(r->res_length);
+	memcpy(rl->rl_name, r->res_name, r->res_length);
+
+	/* FIXME: might we have an lvb without DLM_LKF_VALBLK set ?
+	   If so, receive_rcom_lock_args() won't take this copy. */
+
+	if (lkb->lkb_lvbptr)
+		memcpy(rl->rl_lvb, lkb->lkb_lvbptr, r->res_ls->ls_lvblen);
+}
+
+int dlm_send_rcom_lock(struct dlm_rsb *r, struct dlm_lkb *lkb, uint64_t seq)
+{
+	struct dlm_ls *ls = r->res_ls;
+	struct dlm_rcom *rc;
+	struct dlm_mhandle *mh;
+	struct rcom_lock *rl;
+	int error, len = sizeof(struct rcom_lock);
+
+	if (lkb->lkb_lvbptr)
+		len += ls->ls_lvblen;
+
+	error = create_rcom(ls, r->res_nodeid, DLM_RCOM_LOCK, len, &rc, &mh,
+			    seq);
+	if (error)
+		goto out;
+
+	rl = (struct rcom_lock *) rc->rc_buf;
+	pack_rcom_lock(r, lkb, rl);
+	rc->rc_id = cpu_to_le64((uintptr_t)r);
+
+	send_rcom(mh, rc);
+ out:
+	return error;
+}
+
+/* needs at least dlm_rcom + rcom_lock */
+static void receive_rcom_lock(struct dlm_ls *ls, const struct dlm_rcom *rc_in,
+			      uint64_t seq)
+{
+	__le32 rl_remid, rl_result;
+	struct rcom_lock *rl;
+	struct dlm_rcom *rc;
+	struct dlm_mhandle *mh;
+	int error, nodeid = le32_to_cpu(rc_in->rc_header.h_nodeid);
+
+	dlm_recover_master_copy(ls, rc_in, &rl_remid, &rl_result);
+
+	error = create_rcom(ls, nodeid, DLM_RCOM_LOCK_REPLY,
+			    sizeof(struct rcom_lock), &rc, &mh, seq);
+	if (error)
+		return;
+
+	memcpy(rc->rc_buf, rc_in->rc_buf, sizeof(struct rcom_lock));
+	rl = (struct rcom_lock *)rc->rc_buf;
+	/* set rl_remid and rl_result from dlm_recover_master_copy() */
+	rl->rl_remid = rl_remid;
+	rl->rl_result = rl_result;
+
+	rc->rc_id = rc_in->rc_id;
+	rc->rc_seq_reply = rc_in->rc_seq;
+
+	send_rcom(mh, rc);
+}
+
+/* If the lockspace doesn't exist then still send a status message
+   back; it's possible that it just doesn't have its global_id yet. */
+
+int dlm_send_ls_not_ready(int nodeid, const struct dlm_rcom *rc_in)
+{
+	struct dlm_rcom *rc;
+	struct rcom_config *rf;
+	struct dlm_mhandle *mh;
+	char *mb;
+	int mb_len = sizeof(struct dlm_rcom) + sizeof(struct rcom_config);
+
+	mh = dlm_midcomms_get_mhandle(nodeid, mb_len, &mb);
+	if (!mh)
+		return -ENOBUFS;
+
+	rc = (struct dlm_rcom *) mb;
+
+	rc->rc_header.h_version = cpu_to_le32(DLM_HEADER_MAJOR | DLM_HEADER_MINOR);
+	rc->rc_header.u.h_lockspace = rc_in->rc_header.u.h_lockspace;
+	rc->rc_header.h_nodeid = cpu_to_le32(dlm_our_nodeid());
+	rc->rc_header.h_length = cpu_to_le16(mb_len);
+	rc->rc_header.h_cmd = DLM_RCOM;
+
+	rc->rc_type = cpu_to_le32(DLM_RCOM_STATUS_REPLY);
+	rc->rc_id = rc_in->rc_id;
+	rc->rc_seq_reply = rc_in->rc_seq;
+	rc->rc_result = cpu_to_le32(-ESRCH);
+
+	rf = (struct rcom_config *) rc->rc_buf;
+	rf->rf_lvblen = cpu_to_le32(~0U);
+
+	dlm_midcomms_commit_mhandle(mh, NULL, 0);
+
+	return 0;
+}
+
+/*
+ * Ignore messages for stage Y before we set
+ * recover_status bit for stage X:
+ *
+ * recover_status = 0
+ *
+ * dlm_recover_members()
+ * - send nothing
+ * - recv nothing
+ * - ignore NAMES, NAMES_REPLY
+ * - ignore LOOKUP, LOOKUP_REPLY
+ * - ignore LOCK, LOCK_REPLY
+ *
+ * recover_status |= NODES
+ *
+ * dlm_recover_members_wait()
+ *
+ * dlm_recover_directory()
+ * - send NAMES
+ * - recv NAMES_REPLY
+ * - ignore LOOKUP, LOOKUP_REPLY
+ * - ignore LOCK, LOCK_REPLY
+ *
+ * recover_status |= DIR
+ *
+ * dlm_recover_directory_wait()
+ *
+ * dlm_recover_masters()
+ * - send LOOKUP
+ * - recv LOOKUP_REPLY
+ *
+ * dlm_recover_locks()
+ * - send LOCKS
+ * - recv LOCKS_REPLY
+ *
+ * recover_status |= LOCKS
+ *
+ * dlm_recover_locks_wait()
+ *
+ * recover_status |= DONE
+ */
+
+/* Called by dlm_recv; corresponds to dlm_receive_message() but special
+   recovery-only comms are sent through here. */
+
+void dlm_receive_rcom(struct dlm_ls *ls, const struct dlm_rcom *rc, int nodeid)
+{
+	int lock_size = sizeof(struct dlm_rcom) + sizeof(struct rcom_lock);
+	int stop, reply = 0, names = 0, lookup = 0, lock = 0;
+	uint32_t status;
+	uint64_t seq;
+
+	switch (rc->rc_type) {
+	case cpu_to_le32(DLM_RCOM_STATUS_REPLY):
+		reply = 1;
+		break;
+	case cpu_to_le32(DLM_RCOM_NAMES):
+		names = 1;
+		break;
+	case cpu_to_le32(DLM_RCOM_NAMES_REPLY):
+		names = 1;
+		reply = 1;
+		break;
+	case cpu_to_le32(DLM_RCOM_LOOKUP):
+		lookup = 1;
+		break;
+	case cpu_to_le32(DLM_RCOM_LOOKUP_REPLY):
+		lookup = 1;
+		reply = 1;
+		break;
+	case cpu_to_le32(DLM_RCOM_LOCK):
+		lock = 1;
+		break;
+	case cpu_to_le32(DLM_RCOM_LOCK_REPLY):
+		lock = 1;
+		reply = 1;
+		break;
+	}
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	status = ls->ls_recover_status;
+	stop = dlm_recovery_stopped(ls);
+	seq = ls->ls_recover_seq;
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+	if (stop && (rc->rc_type != cpu_to_le32(DLM_RCOM_STATUS)))
+		goto ignore;
+
+	if (reply && (le64_to_cpu(rc->rc_seq_reply) != seq))
+		goto ignore;
+
+	if (!(status & DLM_RS_NODES) && (names || lookup || lock))
+		goto ignore;
+
+	if (!(status & DLM_RS_DIR) && (lookup || lock))
+		goto ignore;
+
+	switch (rc->rc_type) {
+	case cpu_to_le32(DLM_RCOM_STATUS):
+		receive_rcom_status(ls, rc, seq);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_NAMES):
+		receive_rcom_names(ls, rc, seq);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_LOOKUP):
+		receive_rcom_lookup(ls, rc, seq);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_LOCK):
+		if (le16_to_cpu(rc->rc_header.h_length) < lock_size)
+			goto Eshort;
+		receive_rcom_lock(ls, rc, seq);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_STATUS_REPLY):
+		receive_sync_reply(ls, rc);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_NAMES_REPLY):
+		receive_sync_reply(ls, rc);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_LOOKUP_REPLY):
+		receive_rcom_lookup_reply(ls, rc);
+		break;
+
+	case cpu_to_le32(DLM_RCOM_LOCK_REPLY):
+		if (le16_to_cpu(rc->rc_header.h_length) < lock_size)
+			goto Eshort;
+		dlm_recover_process_copy(ls, rc, seq);
+		break;
+
+	default:
+		log_error(ls, "receive_rcom bad type %d",
+			  le32_to_cpu(rc->rc_type));
+	}
+	return;
+
+ignore:
+	log_limit(ls, "dlm_receive_rcom ignore msg %d "
+		  "from %d %llu %llu recover seq %llu sts %x gen %u",
+		   le32_to_cpu(rc->rc_type),
+		   nodeid,
+		   (unsigned long long)le64_to_cpu(rc->rc_seq),
+		   (unsigned long long)le64_to_cpu(rc->rc_seq_reply),
+		   (unsigned long long)seq,
+		   status, ls->ls_generation);
+	return;
+Eshort:
+	log_error(ls, "recovery message %d from %d is too short",
+		  le32_to_cpu(rc->rc_type), nodeid);
+}
+

linux-7.2.3/fs/dlm/rcom.h

@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2005-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __RCOM_DOT_H__
+#define __RCOM_DOT_H__
+
+int dlm_rcom_status(struct dlm_ls *ls, int nodeid, uint32_t status_flags,
+		    uint64_t seq);
+int dlm_rcom_names(struct dlm_ls *ls, int nodeid, char *last_name,
+		   int last_len, uint64_t seq);
+int dlm_send_rcom_lookup(struct dlm_rsb *r, int dir_nodeid, uint64_t seq);
+int dlm_send_rcom_lock(struct dlm_rsb *r, struct dlm_lkb *lkb, uint64_t seq);
+void dlm_receive_rcom(struct dlm_ls *ls, const struct dlm_rcom *rc,
+		      int nodeid);
+int dlm_send_ls_not_ready(int nodeid, const struct dlm_rcom *rc_in);
+
+#endif
+

linux-7.2.3/fs/dlm/recover.c

@@ -0,0 +1,917 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "dir.h"
+#include "config.h"
+#include "ast.h"
+#include "memory.h"
+#include "rcom.h"
+#include "lock.h"
+#include "lowcomms.h"
+#include "member.h"
+#include "recover.h"
+
+
+/*
+ * Recovery waiting routines: these functions wait for a particular reply from
+ * a remote node, or for the remote node to report a certain status.  They need
+ * to abort if the lockspace is stopped indicating a node has failed (perhaps
+ * the one being waited for).
+ */
+
+/*
+ * Wait until given function returns non-zero or lockspace is stopped
+ * (LS_RECOVERY_STOP set due to failure of a node in ls_nodes).  When another
+ * function thinks it could have completed the waited-on task, they should wake
+ * up ls_wait_general to get an immediate response rather than waiting for the
+ * timeout.  This uses a timeout so it can check periodically if the wait
+ * should abort due to node failure (which doesn't cause a wake_up).
+ * This should only be called by the dlm_recoverd thread.
+ */
+
+int dlm_wait_function(struct dlm_ls *ls, int (*testfn) (struct dlm_ls *ls))
+{
+	int error = 0;
+	int rv;
+
+	while (1) {
+		rv = wait_event_timeout(ls->ls_wait_general,
+					testfn(ls) || dlm_recovery_stopped(ls),
+					dlm_config.ci_recover_timer * HZ);
+		if (rv)
+			break;
+		if (test_bit(LSFL_RCOM_WAIT, &ls->ls_flags)) {
+			log_debug(ls, "dlm_wait_function timed out");
+			return -ETIMEDOUT;
+		}
+	}
+
+	if (dlm_recovery_stopped(ls)) {
+		log_debug(ls, "dlm_wait_function aborted");
+		error = -EINTR;
+	}
+	return error;
+}
+
+/*
+ * An efficient way for all nodes to wait for all others to have a certain
+ * status.  The node with the lowest nodeid polls all the others for their
+ * status (wait_status_all) and all the others poll the node with the low id
+ * for its accumulated result (wait_status_low).  When all nodes have set
+ * status flag X, then status flag X_ALL will be set on the low nodeid.
+ */
+
+uint32_t dlm_recover_status(struct dlm_ls *ls)
+{
+	uint32_t status;
+	spin_lock_bh(&ls->ls_recover_lock);
+	status = ls->ls_recover_status;
+	spin_unlock_bh(&ls->ls_recover_lock);
+	return status;
+}
+
+static void _set_recover_status(struct dlm_ls *ls, uint32_t status)
+{
+	ls->ls_recover_status |= status;
+}
+
+void dlm_set_recover_status(struct dlm_ls *ls, uint32_t status)
+{
+	spin_lock_bh(&ls->ls_recover_lock);
+	_set_recover_status(ls, status);
+	spin_unlock_bh(&ls->ls_recover_lock);
+}
+
+static int wait_status_all(struct dlm_ls *ls, uint32_t wait_status,
+			   int save_slots, uint64_t seq)
+{
+	struct dlm_rcom *rc = ls->ls_recover_buf;
+	struct dlm_member *memb;
+	int error = 0, delay;
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		delay = 0;
+		for (;;) {
+			if (dlm_recovery_stopped(ls)) {
+				error = -EINTR;
+				goto out;
+			}
+
+			error = dlm_rcom_status(ls, memb->nodeid, 0, seq);
+			if (error)
+				goto out;
+
+			if (save_slots)
+				dlm_slot_save(ls, rc, memb);
+
+			if (le32_to_cpu(rc->rc_result) & wait_status)
+				break;
+			if (delay < 1000)
+				delay += 20;
+			msleep(delay);
+		}
+	}
+ out:
+	return error;
+}
+
+static int wait_status_low(struct dlm_ls *ls, uint32_t wait_status,
+			   uint32_t status_flags, uint64_t seq)
+{
+	struct dlm_rcom *rc = ls->ls_recover_buf;
+	int error = 0, delay = 0, nodeid = ls->ls_low_nodeid;
+
+	for (;;) {
+		if (dlm_recovery_stopped(ls)) {
+			error = -EINTR;
+			goto out;
+		}
+
+		error = dlm_rcom_status(ls, nodeid, status_flags, seq);
+		if (error)
+			break;
+
+		if (le32_to_cpu(rc->rc_result) & wait_status)
+			break;
+		if (delay < 1000)
+			delay += 20;
+		msleep(delay);
+	}
+ out:
+	return error;
+}
+
+static int wait_status(struct dlm_ls *ls, uint32_t status, uint64_t seq)
+{
+	uint32_t status_all = status << 1;
+	int error;
+
+	if (ls->ls_low_nodeid == dlm_our_nodeid()) {
+		error = wait_status_all(ls, status, 0, seq);
+		if (!error)
+			dlm_set_recover_status(ls, status_all);
+	} else
+		error = wait_status_low(ls, status_all, 0, seq);
+
+	return error;
+}
+
+int dlm_recover_members_wait(struct dlm_ls *ls, uint64_t seq)
+{
+	struct dlm_member *memb;
+	struct dlm_slot *slots;
+	int num_slots, slots_size;
+	int error, rv;
+	uint32_t gen;
+
+	list_for_each_entry(memb, &ls->ls_nodes, list) {
+		memb->slot = -1;
+		memb->generation = 0;
+	}
+
+	if (ls->ls_low_nodeid == dlm_our_nodeid()) {
+		error = wait_status_all(ls, DLM_RS_NODES, 1, seq);
+		if (error)
+			goto out;
+
+		/* slots array is sparse, slots_size may be > num_slots */
+
+		rv = dlm_slots_assign(ls, &num_slots, &slots_size, &slots, &gen);
+		if (!rv) {
+			spin_lock_bh(&ls->ls_recover_lock);
+			_set_recover_status(ls, DLM_RS_NODES_ALL);
+			ls->ls_num_slots = num_slots;
+			ls->ls_slots_size = slots_size;
+			ls->ls_slots = slots;
+			ls->ls_generation = gen;
+			spin_unlock_bh(&ls->ls_recover_lock);
+		} else {
+			dlm_set_recover_status(ls, DLM_RS_NODES_ALL);
+		}
+	} else {
+		error = wait_status_low(ls, DLM_RS_NODES_ALL,
+					DLM_RSF_NEED_SLOTS, seq);
+		if (error)
+			goto out;
+
+		dlm_slots_copy_in(ls);
+	}
+ out:
+	return error;
+}
+
+int dlm_recover_directory_wait(struct dlm_ls *ls, uint64_t seq)
+{
+	return wait_status(ls, DLM_RS_DIR, seq);
+}
+
+int dlm_recover_locks_wait(struct dlm_ls *ls, uint64_t seq)
+{
+	return wait_status(ls, DLM_RS_LOCKS, seq);
+}
+
+int dlm_recover_done_wait(struct dlm_ls *ls, uint64_t seq)
+{
+	return wait_status(ls, DLM_RS_DONE, seq);
+}
+
+/*
+ * The recover_list contains all the rsb's for which we've requested the new
+ * master nodeid.  As replies are returned from the resource directories the
+ * rsb's are removed from the list.  When the list is empty we're done.
+ *
+ * The recover_list is later similarly used for all rsb's for which we've sent
+ * new lkb's and need to receive new corresponding lkid's.
+ *
+ * We use the address of the rsb struct as a simple local identifier for the
+ * rsb so we can match an rcom reply with the rsb it was sent for.
+ */
+
+static int recover_list_empty(struct dlm_ls *ls)
+{
+	int empty;
+
+	spin_lock_bh(&ls->ls_recover_list_lock);
+	empty = list_empty(&ls->ls_recover_list);
+	spin_unlock_bh(&ls->ls_recover_list_lock);
+
+	return empty;
+}
+
+static void recover_list_add(struct dlm_rsb *r)
+{
+	struct dlm_ls *ls = r->res_ls;
+
+	spin_lock_bh(&ls->ls_recover_list_lock);
+	if (list_empty(&r->res_recover_list)) {
+		list_add_tail(&r->res_recover_list, &ls->ls_recover_list);
+		ls->ls_recover_list_count++;
+		dlm_hold_rsb(r);
+	}
+	spin_unlock_bh(&ls->ls_recover_list_lock);
+}
+
+static void recover_list_del(struct dlm_rsb *r)
+{
+	struct dlm_ls *ls = r->res_ls;
+
+	spin_lock_bh(&ls->ls_recover_list_lock);
+	list_del_init(&r->res_recover_list);
+	ls->ls_recover_list_count--;
+	spin_unlock_bh(&ls->ls_recover_list_lock);
+
+	dlm_put_rsb(r);
+}
+
+static void recover_list_clear(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r, *s;
+
+	spin_lock_bh(&ls->ls_recover_list_lock);
+	list_for_each_entry_safe(r, s, &ls->ls_recover_list, res_recover_list) {
+		list_del_init(&r->res_recover_list);
+		r->res_recover_locks_count = 0;
+		dlm_put_rsb(r);
+		ls->ls_recover_list_count--;
+	}
+
+	if (ls->ls_recover_list_count != 0) {
+		log_error(ls, "warning: recover_list_count %d",
+			  ls->ls_recover_list_count);
+		ls->ls_recover_list_count = 0;
+	}
+	spin_unlock_bh(&ls->ls_recover_list_lock);
+}
+
+static int recover_xa_empty(struct dlm_ls *ls)
+{
+	int empty = 1;
+
+	spin_lock_bh(&ls->ls_recover_xa_lock);
+	if (ls->ls_recover_list_count)
+		empty = 0;
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
+
+	return empty;
+}
+
+static int recover_xa_add(struct dlm_rsb *r)
+{
+	struct dlm_ls *ls = r->res_ls;
+	struct xa_limit limit = {
+		.min = 1,
+		.max = UINT_MAX,
+	};
+	uint32_t id;
+	int rv;
+
+	spin_lock_bh(&ls->ls_recover_xa_lock);
+	if (r->res_id) {
+		rv = -1;
+		goto out_unlock;
+	}
+	rv = xa_alloc(&ls->ls_recover_xa, &id, r, limit, GFP_ATOMIC);
+	if (rv < 0)
+		goto out_unlock;
+
+	r->res_id = id;
+	ls->ls_recover_list_count++;
+	dlm_hold_rsb(r);
+	rv = 0;
+out_unlock:
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
+	return rv;
+}
+
+static void recover_xa_del(struct dlm_rsb *r)
+{
+	struct dlm_ls *ls = r->res_ls;
+
+	spin_lock_bh(&ls->ls_recover_xa_lock);
+	xa_erase_bh(&ls->ls_recover_xa, r->res_id);
+	r->res_id = 0;
+	ls->ls_recover_list_count--;
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
+
+	dlm_put_rsb(r);
+}
+
+static struct dlm_rsb *recover_xa_find(struct dlm_ls *ls, uint64_t id)
+{
+	struct dlm_rsb *r;
+
+	spin_lock_bh(&ls->ls_recover_xa_lock);
+	r = xa_load(&ls->ls_recover_xa, (int)id);
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
+	return r;
+}
+
+static void recover_xa_clear(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r;
+	unsigned long id;
+
+	spin_lock_bh(&ls->ls_recover_xa_lock);
+
+	xa_for_each(&ls->ls_recover_xa, id, r) {
+		xa_erase_bh(&ls->ls_recover_xa, id);
+		r->res_id = 0;
+		r->res_recover_locks_count = 0;
+		ls->ls_recover_list_count--;
+
+		dlm_put_rsb(r);
+	}
+
+	if (ls->ls_recover_list_count != 0) {
+		log_error(ls, "warning: recover_list_count %d",
+			  ls->ls_recover_list_count);
+		ls->ls_recover_list_count = 0;
+	}
+	spin_unlock_bh(&ls->ls_recover_xa_lock);
+}
+
+
+/* Master recovery: find new master node for rsb's that were
+   mastered on nodes that have been removed.
+
+   dlm_recover_masters
+   recover_master
+   dlm_send_rcom_lookup            ->  receive_rcom_lookup
+                                       dlm_dir_lookup
+   receive_rcom_lookup_reply       <-
+   dlm_recover_master_reply
+   set_new_master
+   set_master_lkbs
+   set_lock_master
+*/
+
+/*
+ * Set the lock master for all LKBs in a lock queue
+ * If we are the new master of the rsb, we may have received new
+ * MSTCPY locks from other nodes already which we need to ignore
+ * when setting the new nodeid.
+ */
+
+static void set_lock_master(struct list_head *queue, int nodeid)
+{
+	struct dlm_lkb *lkb;
+
+	list_for_each_entry(lkb, queue, lkb_statequeue) {
+		if (!test_bit(DLM_IFL_MSTCPY_BIT, &lkb->lkb_iflags)) {
+			lkb->lkb_nodeid = nodeid;
+			lkb->lkb_remid = 0;
+		}
+	}
+}
+
+static void set_master_lkbs(struct dlm_rsb *r)
+{
+	set_lock_master(&r->res_grantqueue, r->res_nodeid);
+	set_lock_master(&r->res_convertqueue, r->res_nodeid);
+	set_lock_master(&r->res_waitqueue, r->res_nodeid);
+}
+
+/*
+ * Propagate the new master nodeid to locks
+ * The NEW_MASTER flag tells dlm_recover_locks() which rsb's to consider.
+ * The NEW_MASTER2 flag tells recover_lvb() and recover_grant() which
+ * rsb's to consider.
+ */
+
+static void set_new_master(struct dlm_rsb *r)
+{
+	set_master_lkbs(r);
+	rsb_set_flag(r, RSB_NEW_MASTER);
+	rsb_set_flag(r, RSB_NEW_MASTER2);
+}
+
+/*
+ * We do async lookups on rsb's that need new masters.  The rsb's
+ * waiting for a lookup reply are kept on the recover_list.
+ *
+ * Another node recovering the master may have sent us a rcom lookup,
+ * and our dlm_master_lookup() set it as the new master, along with
+ * NEW_MASTER so that we'll recover it here (this implies dir_nodeid
+ * equals our_nodeid below).
+ */
+
+static int recover_master(struct dlm_rsb *r, unsigned int *count, uint64_t seq)
+{
+	struct dlm_ls *ls = r->res_ls;
+	int our_nodeid, dir_nodeid;
+	int is_removed = 0;
+	int error;
+
+	if (r->res_nodeid != -1 && is_master(r))
+		return 0;
+
+	if (r->res_nodeid != -1)
+		is_removed = dlm_is_removed(ls, r->res_nodeid);
+
+	if (!is_removed && !rsb_flag(r, RSB_NEW_MASTER))
+		return 0;
+
+	our_nodeid = dlm_our_nodeid();
+	dir_nodeid = dlm_dir_nodeid(r);
+
+	if (dir_nodeid == our_nodeid) {
+		if (is_removed) {
+			r->res_master_nodeid = our_nodeid;
+			r->res_nodeid = 0;
+		}
+
+		/* set master of lkbs to ourself when is_removed, or to
+		   another new master which we set along with NEW_MASTER
+		   in dlm_master_lookup */
+		set_new_master(r);
+		error = 0;
+	} else {
+		recover_xa_add(r);
+		error = dlm_send_rcom_lookup(r, dir_nodeid, seq);
+	}
+
+	(*count)++;
+	return error;
+}
+
+/*
+ * All MSTCPY locks are purged and rebuilt, even if the master stayed the same.
+ * This is necessary because recovery can be started, aborted and restarted,
+ * causing the master nodeid to briefly change during the aborted recovery, and
+ * change back to the original value in the second recovery.  The MSTCPY locks
+ * may or may not have been purged during the aborted recovery.  Another node
+ * with an outstanding request in waiters list and a request reply saved in the
+ * requestqueue, cannot know whether it should ignore the reply and resend the
+ * request, or accept the reply and complete the request.  It must do the
+ * former if the remote node purged MSTCPY locks, and it must do the later if
+ * the remote node did not.  This is solved by always purging MSTCPY locks, in
+ * which case, the request reply would always be ignored and the request
+ * resent.
+ */
+
+static int recover_master_static(struct dlm_rsb *r, unsigned int *count)
+{
+	int dir_nodeid = dlm_dir_nodeid(r);
+	int new_master = dir_nodeid;
+
+	if (dir_nodeid == dlm_our_nodeid())
+		new_master = 0;
+
+	dlm_purge_mstcpy_locks(r);
+	r->res_master_nodeid = dir_nodeid;
+	r->res_nodeid = new_master;
+	set_new_master(r);
+	(*count)++;
+	return 0;
+}
+
+/*
+ * Go through local root resources and for each rsb which has a master which
+ * has departed, get the new master nodeid from the directory.  The dir will
+ * assign mastery to the first node to look up the new master.  That means
+ * we'll discover in this lookup if we're the new master of any rsb's.
+ *
+ * We fire off all the dir lookup requests individually and asynchronously to
+ * the correct dir node.
+ */
+
+int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq,
+			const struct list_head *root_list)
+{
+	struct dlm_rsb *r;
+	unsigned int total = 0;
+	unsigned int count = 0;
+	int nodir = dlm_no_directory(ls);
+	int error;
+
+	log_rinfo(ls, "dlm_recover_masters");
+
+	list_for_each_entry(r, root_list, res_root_list) {
+		if (dlm_recovery_stopped(ls)) {
+			error = -EINTR;
+			goto out;
+		}
+
+		lock_rsb(r);
+		if (nodir)
+			error = recover_master_static(r, &count);
+		else
+			error = recover_master(r, &count, seq);
+		unlock_rsb(r);
+		cond_resched();
+		total++;
+
+		if (error)
+			goto out;
+	}
+
+	log_rinfo(ls, "dlm_recover_masters %u of %u", count, total);
+
+	error = dlm_wait_function(ls, &recover_xa_empty);
+ out:
+	if (error)
+		recover_xa_clear(ls);
+	return error;
+}
+
+int dlm_recover_master_reply(struct dlm_ls *ls, const struct dlm_rcom *rc)
+{
+	struct dlm_rsb *r;
+	int ret_nodeid, new_master;
+
+	r = recover_xa_find(ls, le64_to_cpu(rc->rc_id));
+	if (!r) {
+		log_error(ls, "dlm_recover_master_reply no id %llx",
+			  (unsigned long long)le64_to_cpu(rc->rc_id));
+		goto out;
+	}
+
+	ret_nodeid = le32_to_cpu(rc->rc_result);
+
+	if (ret_nodeid == dlm_our_nodeid())
+		new_master = 0;
+	else
+		new_master = ret_nodeid;
+
+	lock_rsb(r);
+	r->res_master_nodeid = ret_nodeid;
+	r->res_nodeid = new_master;
+	set_new_master(r);
+	unlock_rsb(r);
+	recover_xa_del(r);
+
+	if (recover_xa_empty(ls))
+		wake_up(&ls->ls_wait_general);
+ out:
+	return 0;
+}
+
+
+/* Lock recovery: rebuild the process-copy locks we hold on a
+   remastered rsb on the new rsb master.
+
+   dlm_recover_locks
+   recover_locks
+   recover_locks_queue
+   dlm_send_rcom_lock              ->  receive_rcom_lock
+                                       dlm_recover_master_copy
+   receive_rcom_lock_reply         <-
+   dlm_recover_process_copy
+*/
+
+
+/*
+ * keep a count of the number of lkb's we send to the new master; when we get
+ * an equal number of replies then recovery for the rsb is done
+ */
+
+static int recover_locks_queue(struct dlm_rsb *r, struct list_head *head,
+			       uint64_t seq)
+{
+	struct dlm_lkb *lkb;
+	int error = 0;
+
+	list_for_each_entry(lkb, head, lkb_statequeue) {
+		error = dlm_send_rcom_lock(r, lkb, seq);
+		if (error)
+			break;
+		r->res_recover_locks_count++;
+	}
+
+	return error;
+}
+
+static int recover_locks(struct dlm_rsb *r, uint64_t seq)
+{
+	int error = 0;
+
+	lock_rsb(r);
+
+	DLM_ASSERT(!r->res_recover_locks_count, dlm_dump_rsb(r););
+
+	error = recover_locks_queue(r, &r->res_grantqueue, seq);
+	if (error)
+		goto out;
+	error = recover_locks_queue(r, &r->res_convertqueue, seq);
+	if (error)
+		goto out;
+	error = recover_locks_queue(r, &r->res_waitqueue, seq);
+	if (error)
+		goto out;
+
+	if (r->res_recover_locks_count)
+		recover_list_add(r);
+	else
+		rsb_clear_flag(r, RSB_NEW_MASTER);
+ out:
+	unlock_rsb(r);
+	return error;
+}
+
+int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq,
+		      const struct list_head *root_list)
+{
+	struct dlm_rsb *r;
+	int error, count = 0;
+
+	list_for_each_entry(r, root_list, res_root_list) {
+		if (r->res_nodeid != -1 && is_master(r)) {
+			rsb_clear_flag(r, RSB_NEW_MASTER);
+			continue;
+		}
+
+		if (!rsb_flag(r, RSB_NEW_MASTER))
+			continue;
+
+		if (dlm_recovery_stopped(ls)) {
+			error = -EINTR;
+			goto out;
+		}
+
+		error = recover_locks(r, seq);
+		if (error)
+			goto out;
+
+		count += r->res_recover_locks_count;
+	}
+
+	log_rinfo(ls, "dlm_recover_locks %d out", count);
+
+	error = dlm_wait_function(ls, &recover_list_empty);
+ out:
+	if (error)
+		recover_list_clear(ls);
+	return error;
+}
+
+void dlm_recovered_lock(struct dlm_rsb *r)
+{
+	DLM_ASSERT(rsb_flag(r, RSB_NEW_MASTER), dlm_dump_rsb(r););
+
+	r->res_recover_locks_count--;
+	if (!r->res_recover_locks_count) {
+		rsb_clear_flag(r, RSB_NEW_MASTER);
+		recover_list_del(r);
+	}
+
+	if (recover_list_empty(r->res_ls))
+		wake_up(&r->res_ls->ls_wait_general);
+}
+
+/*
+ * The lvb needs to be recovered on all master rsb's.  This includes setting
+ * the VALNOTVALID flag if necessary, and determining the correct lvb contents
+ * based on the lvb's of the locks held on the rsb.
+ *
+ * RSB_VALNOTVALID is set in two cases:
+ *
+ * 1. we are master, but not new, and we purged an EX/PW lock held by a
+ * failed node (in dlm_recover_purge which set RSB_RECOVER_LVB_INVAL)
+ *
+ * 2. we are a new master, and there are only NL/CR locks left.
+ * (We could probably improve this by only invaliding in this way when
+ * the previous master left uncleanly.  VMS docs mention that.)
+ *
+ * The LVB contents are only considered for changing when this is a new master
+ * of the rsb (NEW_MASTER2).  Then, the rsb's lvb is taken from any lkb with
+ * mode > CR.  If no lkb's exist with mode above CR, the lvb contents are taken
+ * from the lkb with the largest lvb sequence number.
+ */
+
+static void recover_lvb(struct dlm_rsb *r)
+{
+	struct dlm_lkb *big_lkb = NULL, *iter, *high_lkb = NULL;
+	uint32_t high_seq = 0;
+	int lock_lvb_exists = 0;
+	int lvblen = r->res_ls->ls_lvblen;
+
+	if (!rsb_flag(r, RSB_NEW_MASTER2) &&
+	    rsb_flag(r, RSB_RECOVER_LVB_INVAL)) {
+		/* case 1 above */
+		rsb_set_flag(r, RSB_VALNOTVALID);
+		return;
+	}
+
+	if (!rsb_flag(r, RSB_NEW_MASTER2))
+		return;
+
+	/* we are the new master, so figure out if VALNOTVALID should
+	   be set, and set the rsb lvb from the best lkb available. */
+
+	list_for_each_entry(iter, &r->res_grantqueue, lkb_statequeue) {
+		if (!(iter->lkb_exflags & DLM_LKF_VALBLK))
+			continue;
+
+		lock_lvb_exists = 1;
+
+		if (iter->lkb_grmode > DLM_LOCK_CR) {
+			big_lkb = iter;
+			goto setflag;
+		}
+
+		if (((int)iter->lkb_lvbseq - (int)high_seq) >= 0) {
+			high_lkb = iter;
+			high_seq = iter->lkb_lvbseq;
+		}
+	}
+
+	list_for_each_entry(iter, &r->res_convertqueue, lkb_statequeue) {
+		if (!(iter->lkb_exflags & DLM_LKF_VALBLK))
+			continue;
+
+		lock_lvb_exists = 1;
+
+		if (iter->lkb_grmode > DLM_LOCK_CR) {
+			big_lkb = iter;
+			goto setflag;
+		}
+
+		if (((int)iter->lkb_lvbseq - (int)high_seq) >= 0) {
+			high_lkb = iter;
+			high_seq = iter->lkb_lvbseq;
+		}
+	}
+
+ setflag:
+	if (!lock_lvb_exists)
+		goto out;
+
+	/* lvb is invalidated if only NL/CR locks remain */
+	if (!big_lkb)
+		rsb_set_flag(r, RSB_VALNOTVALID);
+
+	if (!r->res_lvbptr) {
+		r->res_lvbptr = dlm_allocate_lvb(r->res_ls);
+		if (!r->res_lvbptr)
+			goto out;
+	}
+
+	if (big_lkb) {
+		r->res_lvbseq = big_lkb->lkb_lvbseq;
+		memcpy(r->res_lvbptr, big_lkb->lkb_lvbptr, lvblen);
+	} else if (high_lkb) {
+		r->res_lvbseq = high_lkb->lkb_lvbseq;
+		memcpy(r->res_lvbptr, high_lkb->lkb_lvbptr, lvblen);
+	} else {
+		r->res_lvbseq = 0;
+		memset(r->res_lvbptr, 0, lvblen);
+	}
+ out:
+	return;
+}
+
+/* All master rsb's flagged RECOVER_CONVERT need to be looked at.  The locks
+ * 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;
+
+	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
+		if (lkb->lkb_grmode == DLM_LOCK_PR ||
+		    lkb->lkb_grmode == DLM_LOCK_CW) {
+			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) {
+		/* 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_rinfo(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;
+		}
+	}
+}
+
+/* We've become the new master for this rsb and waiting/converting locks may
+   need to be granted in dlm_recover_grant() due to locks that may have
+   existed from a removed node. */
+
+static void recover_grant(struct dlm_rsb *r)
+{
+	if (!list_empty(&r->res_waitqueue) || !list_empty(&r->res_convertqueue))
+		rsb_set_flag(r, RSB_RECOVER_GRANT);
+}
+
+void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list)
+{
+	struct dlm_rsb *r;
+	unsigned int count = 0;
+
+	list_for_each_entry(r, root_list, res_root_list) {
+		lock_rsb(r);
+		if (r->res_nodeid != -1 && is_master(r)) {
+			if (rsb_flag(r, RSB_RECOVER_CONVERT))
+				recover_conversion(r);
+
+			/* recover lvb before granting locks so the updated
+			   lvb/VALNOTVALID is presented in the completion */
+			recover_lvb(r);
+
+			if (rsb_flag(r, RSB_NEW_MASTER2))
+				recover_grant(r);
+			count++;
+		} else {
+			rsb_clear_flag(r, RSB_VALNOTVALID);
+		}
+		rsb_clear_flag(r, RSB_RECOVER_CONVERT);
+		rsb_clear_flag(r, RSB_RECOVER_LVB_INVAL);
+		rsb_clear_flag(r, RSB_NEW_MASTER2);
+		unlock_rsb(r);
+	}
+
+	if (count)
+		log_rinfo(ls, "dlm_recover_rsbs %d done", count);
+}
+
+void dlm_clear_inactive(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r, *safe;
+	unsigned int count = 0;
+
+	write_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry_safe(r, safe, &ls->ls_slow_inactive, res_slow_list) {
+		list_del(&r->res_slow_list);
+		rhashtable_remove_fast(&ls->ls_rsbtbl, &r->res_node,
+				       dlm_rhash_rsb_params);
+
+		if (!list_empty(&r->res_scan_list))
+			list_del_init(&r->res_scan_list);
+
+		free_inactive_rsb(r);
+		count++;
+	}
+	write_unlock_bh(&ls->ls_rsbtbl_lock);
+
+	if (count)
+		log_rinfo(ls, "dlm_clear_inactive %u done", count);
+}
+

linux-7.2.3/fs/dlm/recover.h

@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __RECOVER_DOT_H__
+#define __RECOVER_DOT_H__
+
+int dlm_wait_function(struct dlm_ls *ls, int (*testfn) (struct dlm_ls *ls));
+uint32_t dlm_recover_status(struct dlm_ls *ls);
+void dlm_set_recover_status(struct dlm_ls *ls, uint32_t status);
+int dlm_recover_members_wait(struct dlm_ls *ls, uint64_t seq);
+int dlm_recover_directory_wait(struct dlm_ls *ls, uint64_t seq);
+int dlm_recover_locks_wait(struct dlm_ls *ls, uint64_t seq);
+int dlm_recover_done_wait(struct dlm_ls *ls, uint64_t seq);
+int dlm_recover_masters(struct dlm_ls *ls, uint64_t seq,
+			const struct list_head *root_list);
+int dlm_recover_master_reply(struct dlm_ls *ls, const struct dlm_rcom *rc);
+int dlm_recover_locks(struct dlm_ls *ls, uint64_t seq,
+		      const struct list_head *root_list);
+void dlm_recovered_lock(struct dlm_rsb *r);
+void dlm_clear_inactive(struct dlm_ls *ls);
+void dlm_recover_rsbs(struct dlm_ls *ls, const struct list_head *root_list);
+
+#endif				/* __RECOVER_DOT_H__ */
+

linux-7.2.3/fs/dlm/recoverd.c

@@ -0,0 +1,468 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2011 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "member.h"
+#include "dir.h"
+#include "ast.h"
+#include "recover.h"
+#include "lowcomms.h"
+#include "lock.h"
+#include "requestqueue.h"
+#include "recoverd.h"
+
+static int dlm_create_masters_list(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r;
+	int error = 0;
+
+	write_lock_bh(&ls->ls_masters_lock);
+	if (!list_empty(&ls->ls_masters_list)) {
+		log_error(ls, "root list not empty");
+		error = -EINVAL;
+		goto out;
+	}
+
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry(r, &ls->ls_slow_active, res_slow_list) {
+		if (r->res_nodeid)
+			continue;
+
+		list_add(&r->res_masters_list, &ls->ls_masters_list);
+		dlm_hold_rsb(r);
+	}
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+ out:
+	write_unlock_bh(&ls->ls_masters_lock);
+	return error;
+}
+
+static void dlm_release_masters_list(struct dlm_ls *ls)
+{
+	struct dlm_rsb *r, *safe;
+
+	write_lock_bh(&ls->ls_masters_lock);
+	list_for_each_entry_safe(r, safe, &ls->ls_masters_list, res_masters_list) {
+		list_del_init(&r->res_masters_list);
+		dlm_put_rsb(r);
+	}
+	write_unlock_bh(&ls->ls_masters_lock);
+}
+
+static void dlm_create_root_list(struct dlm_ls *ls, struct list_head *root_list)
+{
+	struct dlm_rsb *r;
+
+	read_lock_bh(&ls->ls_rsbtbl_lock);
+	list_for_each_entry(r, &ls->ls_slow_active, res_slow_list) {
+		list_add(&r->res_root_list, root_list);
+		dlm_hold_rsb(r);
+	}
+
+	WARN_ON_ONCE(!list_empty(&ls->ls_slow_inactive));
+	read_unlock_bh(&ls->ls_rsbtbl_lock);
+}
+
+static void dlm_release_root_list(struct list_head *root_list)
+{
+	struct dlm_rsb *r, *safe;
+
+	list_for_each_entry_safe(r, safe, root_list, res_root_list) {
+		list_del_init(&r->res_root_list);
+		dlm_put_rsb(r);
+	}
+}
+
+/* If the start for which we're re-enabling locking (seq) has been superseded
+   by a newer stop (ls_recover_seq), we need to leave locking disabled.
+
+   We suspend dlm_recv threads here to avoid the race where dlm_recv a) sees
+   locking stopped and b) adds a message to the requestqueue, but dlm_recoverd
+   enables locking and clears the requestqueue between a and b. */
+
+static int enable_locking(struct dlm_ls *ls, uint64_t seq)
+{
+	int error = -EINTR;
+
+	write_lock_bh(&ls->ls_recv_active);
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	if (ls->ls_recover_seq == seq) {
+		set_bit(LSFL_RUNNING, &ls->ls_flags);
+		/* Schedule next timer if recovery put something on inactive.
+		 *
+		 * The rsbs that was queued while recovery on toss hasn't
+		 * started yet because LSFL_RUNNING was set everything
+		 * else recovery hasn't started as well because ls_in_recovery
+		 * is still hold. So we should not run into the case that
+		 * resume_scan_timer() queues a timer that can occur in
+		 * a no op.
+		 */
+		resume_scan_timer(ls);
+		/* unblocks processes waiting to enter the dlm */
+		up_write(&ls->ls_in_recovery);
+		clear_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
+		error = 0;
+	}
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+	write_unlock_bh(&ls->ls_recv_active);
+	return error;
+}
+
+static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
+{
+	LIST_HEAD(root_list);
+	unsigned long start;
+	int error, neg = 0;
+
+	log_rinfo(ls, "dlm_recover %llu", (unsigned long long)rv->seq);
+
+	mutex_lock(&ls->ls_recoverd_active);
+
+	dlm_callback_suspend(ls);
+
+	dlm_clear_inactive(ls);
+
+	/*
+	 * This list of root rsb's will be the basis of most of the recovery
+	 * routines.
+	 */
+
+	dlm_create_root_list(ls, &root_list);
+
+	/*
+	 * Add or remove nodes from the lockspace's ls_nodes list.
+	 *
+	 * Due to the fact that we must report all membership changes to lsops
+	 * or midcomms layer, it is not permitted to abort ls_recover() until
+	 * this is done.
+	 */
+
+	error = dlm_recover_members(ls, rv, &neg);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_members error %d", error);
+		goto fail_root_list;
+	}
+
+	dlm_recover_dir_nodeid(ls, &root_list);
+
+	/* Create a snapshot of all active rsbs were we are the master of.
+	 * During the barrier between dlm_recover_members_wait() and
+	 * dlm_recover_directory() other nodes can dump their necessary
+	 * directory dlm_rsb (r->res_dir_nodeid == nodeid) in rcom
+	 * communication dlm_copy_master_names() handling.
+	 *
+	 * TODO We should create a per lockspace list that contains rsbs
+	 * that we are the master of. Instead of creating this list while
+	 * recovery we keep track of those rsbs while locking handling and
+	 * recovery can use it when necessary.
+	 */
+	error = dlm_create_masters_list(ls);
+	if (error) {
+		log_rinfo(ls, "dlm_create_masters_list error %d", error);
+		goto fail_root_list;
+	}
+
+	ls->ls_recover_locks_in = 0;
+
+	dlm_set_recover_status(ls, DLM_RS_NODES);
+
+	error = dlm_recover_members_wait(ls, rv->seq);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_members_wait error %d", error);
+		dlm_release_masters_list(ls);
+		goto fail_root_list;
+	}
+
+	start = jiffies;
+
+	/*
+	 * Rebuild our own share of the directory by collecting from all other
+	 * nodes their master rsb names that hash to us.
+	 */
+
+	error = dlm_recover_directory(ls, rv->seq);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_directory error %d", error);
+		dlm_release_masters_list(ls);
+		goto fail_root_list;
+	}
+
+	dlm_set_recover_status(ls, DLM_RS_DIR);
+
+	error = dlm_recover_directory_wait(ls, rv->seq);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_directory_wait error %d", error);
+		dlm_release_masters_list(ls);
+		goto fail_root_list;
+	}
+
+	dlm_release_masters_list(ls);
+
+	/*
+	 * We may have outstanding operations that are waiting for a reply from
+	 * a failed node.  Mark these to be resent after recovery.  Unlock and
+	 * cancel ops can just be completed.
+	 */
+
+	dlm_recover_waiters_pre(ls);
+
+	if (dlm_recovery_stopped(ls)) {
+		error = -EINTR;
+		goto fail_root_list;
+	}
+
+	if (neg || dlm_no_directory(ls)) {
+		/*
+		 * Clear lkb's for departed nodes.
+		 */
+
+		dlm_recover_purge(ls, &root_list);
+
+		/*
+		 * Get new master nodeid's for rsb's that were mastered on
+		 * departed nodes.
+		 */
+
+		error = dlm_recover_masters(ls, rv->seq, &root_list);
+		if (error) {
+			log_rinfo(ls, "dlm_recover_masters error %d", error);
+			goto fail_root_list;
+		}
+
+		/*
+		 * Send our locks on remastered rsb's to the new masters.
+		 */
+
+		error = dlm_recover_locks(ls, rv->seq, &root_list);
+		if (error) {
+			log_rinfo(ls, "dlm_recover_locks error %d", error);
+			goto fail_root_list;
+		}
+
+		dlm_set_recover_status(ls, DLM_RS_LOCKS);
+
+		error = dlm_recover_locks_wait(ls, rv->seq);
+		if (error) {
+			log_rinfo(ls, "dlm_recover_locks_wait error %d", error);
+			goto fail_root_list;
+		}
+
+		log_rinfo(ls, "dlm_recover_locks %u in",
+			  ls->ls_recover_locks_in);
+
+		/*
+		 * Finalize state in master rsb's now that all locks can be
+		 * checked.  This includes conversion resolution and lvb
+		 * settings.
+		 */
+
+		dlm_recover_rsbs(ls, &root_list);
+	} else {
+		/*
+		 * Other lockspace members may be going through the "neg" steps
+		 * while also adding us to the lockspace, in which case they'll
+		 * be doing the recover_locks (RS_LOCKS) barrier.
+		 */
+		dlm_set_recover_status(ls, DLM_RS_LOCKS);
+
+		error = dlm_recover_locks_wait(ls, rv->seq);
+		if (error) {
+			log_rinfo(ls, "dlm_recover_locks_wait error %d", error);
+			goto fail_root_list;
+		}
+	}
+
+	dlm_release_root_list(&root_list);
+
+	/*
+	 * Purge directory-related requests that are saved in requestqueue.
+	 * All dir requests from before recovery are invalid now due to the dir
+	 * rebuild and will be resent by the requesting nodes.
+	 */
+
+	dlm_purge_requestqueue(ls);
+
+	dlm_set_recover_status(ls, DLM_RS_DONE);
+
+	error = dlm_recover_done_wait(ls, rv->seq);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_done_wait error %d", error);
+		goto fail;
+	}
+
+	dlm_clear_members_gone(ls);
+
+	dlm_callback_resume(ls);
+
+	error = enable_locking(ls, rv->seq);
+	if (error) {
+		log_rinfo(ls, "enable_locking error %d", error);
+		goto fail;
+	}
+
+	error = dlm_process_requestqueue(ls);
+	if (error) {
+		log_rinfo(ls, "dlm_process_requestqueue error %d", error);
+		goto fail;
+	}
+
+	error = dlm_recover_waiters_post(ls);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_waiters_post error %d", error);
+		goto fail;
+	}
+
+	dlm_recover_grant(ls);
+
+	log_rinfo(ls, "dlm_recover %llu generation %u done: %u ms",
+		  (unsigned long long)rv->seq, ls->ls_generation,
+		  jiffies_to_msecs(jiffies - start));
+	mutex_unlock(&ls->ls_recoverd_active);
+
+	return 0;
+
+ fail_root_list:
+	dlm_release_root_list(&root_list);
+ fail:
+	mutex_unlock(&ls->ls_recoverd_active);
+
+	return error;
+}
+
+/* The dlm_ls_start() that created the rv we take here may already have been
+   stopped via dlm_ls_stop(); in that case we need to leave the RECOVERY_STOP
+   flag set. */
+
+static void do_ls_recovery(struct dlm_ls *ls)
+{
+	struct dlm_recover *rv = NULL;
+	int error;
+
+	spin_lock_bh(&ls->ls_recover_lock);
+	rv = ls->ls_recover_args;
+	ls->ls_recover_args = NULL;
+	if (rv && ls->ls_recover_seq == rv->seq)
+		clear_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
+	spin_unlock_bh(&ls->ls_recover_lock);
+
+	if (rv) {
+		error = ls_recover(ls, rv);
+		switch (error) {
+		case 0:
+			ls->ls_recovery_result = 0;
+			complete(&ls->ls_recovery_done);
+
+			dlm_lsop_recover_done(ls);
+			break;
+		case -EINTR:
+			/* if recovery was interrupted -EINTR we wait for the next
+			 * ls_recover() iteration until it hopefully succeeds.
+			 */
+			log_rinfo(ls, "%s %llu interrupted and should be queued to run again",
+				  __func__, (unsigned long long)rv->seq);
+			break;
+		default:
+			log_rinfo(ls, "%s %llu error %d", __func__,
+				  (unsigned long long)rv->seq, error);
+
+			/* let new_lockspace() get aware of critical error */
+			ls->ls_recovery_result = error;
+			complete(&ls->ls_recovery_done);
+			break;
+		}
+
+		kfree(rv->nodes);
+		kfree(rv);
+	}
+}
+
+static int dlm_recoverd(void *arg)
+{
+	struct dlm_ls *ls;
+
+	ls = dlm_find_lockspace_local(arg);
+	if (!ls) {
+		log_print("dlm_recoverd: no lockspace %p", arg);
+		return -1;
+	}
+
+	down_write(&ls->ls_in_recovery);
+	set_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
+	wake_up(&ls->ls_recover_lock_wait);
+
+	while (1) {
+		/*
+		 * We call kthread_should_stop() after set_current_state().
+		 * This is because it works correctly if kthread_stop() is
+		 * called just before set_current_state().
+		 */
+		set_current_state(TASK_INTERRUPTIBLE);
+		if (kthread_should_stop()) {
+			set_current_state(TASK_RUNNING);
+			break;
+		}
+		if (!test_bit(LSFL_RECOVER_WORK, &ls->ls_flags) &&
+		    !test_bit(LSFL_RECOVER_DOWN, &ls->ls_flags)) {
+			if (kthread_should_stop())
+				break;
+			schedule();
+		}
+		set_current_state(TASK_RUNNING);
+
+		if (test_and_clear_bit(LSFL_RECOVER_DOWN, &ls->ls_flags)) {
+			down_write(&ls->ls_in_recovery);
+			set_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
+			wake_up(&ls->ls_recover_lock_wait);
+		}
+
+		if (test_and_clear_bit(LSFL_RECOVER_WORK, &ls->ls_flags))
+			do_ls_recovery(ls);
+	}
+
+	if (test_bit(LSFL_RECOVER_LOCK, &ls->ls_flags))
+		up_write(&ls->ls_in_recovery);
+
+	dlm_put_lockspace(ls);
+	return 0;
+}
+
+int dlm_recoverd_start(struct dlm_ls *ls)
+{
+	struct task_struct *p;
+	int error = 0;
+
+	p = kthread_run(dlm_recoverd, ls, "dlm_recoverd");
+	if (IS_ERR(p))
+		error = PTR_ERR(p);
+	else
+                ls->ls_recoverd_task = p;
+	return error;
+}
+
+void dlm_recoverd_stop(struct dlm_ls *ls)
+{
+	kthread_stop(ls->ls_recoverd_task);
+}
+
+void dlm_recoverd_suspend(struct dlm_ls *ls)
+{
+	wake_up(&ls->ls_wait_general);
+	mutex_lock(&ls->ls_recoverd_active);
+}
+
+void dlm_recoverd_resume(struct dlm_ls *ls)
+{
+	mutex_unlock(&ls->ls_recoverd_active);
+}
+

linux-7.2.3/fs/dlm/recoverd.h

@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
+**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __RECOVERD_DOT_H__
+#define __RECOVERD_DOT_H__
+
+void dlm_recoverd_stop(struct dlm_ls *ls);
+int dlm_recoverd_start(struct dlm_ls *ls);
+void dlm_recoverd_suspend(struct dlm_ls *ls);
+void dlm_recoverd_resume(struct dlm_ls *ls);
+
+#endif				/* __RECOVERD_DOT_H__ */
+

linux-7.2.3/fs/dlm/requestqueue.c

@@ -0,0 +1,149 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "member.h"
+#include "lock.h"
+#include "dir.h"
+#include "config.h"
+#include "requestqueue.h"
+#include "util.h"
+
+struct rq_entry {
+	struct list_head list;
+	uint32_t recover_seq;
+	int nodeid;
+	struct dlm_message request;
+};
+
+/*
+ * Requests received while the lockspace is in recovery get added to the
+ * request queue and processed when recovery is complete.  This happens when
+ * the lockspace is suspended on some nodes before it is on others, or the
+ * lockspace is enabled on some while still suspended on others.
+ */
+
+void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
+			  const struct dlm_message *ms)
+{
+	struct rq_entry *e;
+	int length = le16_to_cpu(ms->m_header.h_length) -
+		sizeof(struct dlm_message);
+
+	e = kmalloc(sizeof(struct rq_entry) + length, GFP_ATOMIC);
+	if (!e) {
+		log_print("dlm_add_requestqueue: out of memory len %d", length);
+		return;
+	}
+
+	e->recover_seq = ls->ls_recover_seq & 0xFFFFFFFF;
+	e->nodeid = nodeid;
+	memcpy(&e->request, ms, sizeof(*ms));
+	memcpy(&e->request.m_extra, ms->m_extra, length);
+
+	list_add_tail(&e->list, &ls->ls_requestqueue);
+}
+
+/*
+ * Called by dlm_recoverd to process normal messages saved while recovery was
+ * happening.  Normal locking has been enabled before this is called.  dlm_recv
+ * upon receiving a message, will wait for all saved messages to be drained
+ * here before processing the message it got.  If a new dlm_ls_stop() arrives
+ * while we're processing these saved messages, it may block trying to suspend
+ * dlm_recv if dlm_recv is waiting for us in dlm_wait_requestqueue.  In that
+ * case, we don't abort since locking_stopped is still 0.  If dlm_recv is not
+ * waiting for us, then this processing may be aborted due to locking_stopped.
+ */
+
+int dlm_process_requestqueue(struct dlm_ls *ls)
+{
+	struct rq_entry *e;
+	struct dlm_message *ms;
+	int error = 0;
+
+	write_lock_bh(&ls->ls_requestqueue_lock);
+	for (;;) {
+		if (list_empty(&ls->ls_requestqueue)) {
+			clear_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags);
+			error = 0;
+			break;
+		}
+		e = list_first_entry(&ls->ls_requestqueue, struct rq_entry, list);
+
+		ms = &e->request;
+
+		log_limit(ls, "dlm_process_requestqueue msg %d from %d "
+			  "lkid %x remid %x result %d seq %u",
+			  le32_to_cpu(ms->m_type),
+			  le32_to_cpu(ms->m_header.h_nodeid),
+			  le32_to_cpu(ms->m_lkid), le32_to_cpu(ms->m_remid),
+			  from_dlm_errno(le32_to_cpu(ms->m_result)),
+			  e->recover_seq);
+
+		dlm_receive_message_saved(ls, &e->request, e->recover_seq);
+		list_del(&e->list);
+		kfree(e);
+
+		if (dlm_locking_stopped(ls)) {
+			log_debug(ls, "process_requestqueue abort running");
+			error = -EINTR;
+			break;
+		}
+		write_unlock_bh(&ls->ls_requestqueue_lock);
+		schedule();
+		write_lock_bh(&ls->ls_requestqueue_lock);
+	}
+	write_unlock_bh(&ls->ls_requestqueue_lock);
+
+	return error;
+}
+
+static int purge_request(struct dlm_ls *ls, struct dlm_message *ms, int nodeid)
+{
+	__le32 type = ms->m_type;
+
+	/* the ls is being cleaned up and freed by release_lockspace */
+	if (!atomic_read(&ls->ls_count))
+		return 1;
+
+	if (dlm_is_removed(ls, nodeid))
+		return 1;
+
+	/* directory operations are always purged because the directory is
+	   always rebuilt during recovery and the lookups resent */
+
+	if (type == cpu_to_le32(DLM_MSG_REMOVE) ||
+	    type == cpu_to_le32(DLM_MSG_LOOKUP) ||
+	    type == cpu_to_le32(DLM_MSG_LOOKUP_REPLY))
+		return 1;
+
+	if (!dlm_no_directory(ls))
+		return 0;
+
+	return 1;
+}
+
+void dlm_purge_requestqueue(struct dlm_ls *ls)
+{
+	struct dlm_message *ms;
+	struct rq_entry *e, *safe;
+
+	write_lock_bh(&ls->ls_requestqueue_lock);
+	list_for_each_entry_safe(e, safe, &ls->ls_requestqueue, list) {
+		ms =  &e->request;
+
+		if (purge_request(ls, ms, e->nodeid)) {
+			list_del(&e->list);
+			kfree(e);
+		}
+	}
+	write_unlock_bh(&ls->ls_requestqueue_lock);
+}
+

linux-7.2.3/fs/dlm/requestqueue.h

@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2007 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __REQUESTQUEUE_DOT_H__
+#define __REQUESTQUEUE_DOT_H__
+
+void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
+			  const struct dlm_message *ms);
+int dlm_process_requestqueue(struct dlm_ls *ls);
+void dlm_wait_requestqueue(struct dlm_ls *ls);
+void dlm_purge_requestqueue(struct dlm_ls *ls);
+
+#endif
+

linux-7.2.3/fs/dlm/user.c

@@ -0,0 +1,989 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2006-2010 Red Hat, Inc.  All rights reserved.
+ */
+
+#include <linux/miscdevice.h>
+#include <linux/init.h>
+#include <linux/wait.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/poll.h>
+#include <linux/signal.h>
+#include <linux/spinlock.h>
+#include <linux/dlm.h>
+#include <linux/dlm_device.h>
+#include <linux/slab.h>
+#include <linux/sched/signal.h>
+
+#include <trace/events/dlm.h>
+
+#include "dlm_internal.h"
+#include "lockspace.h"
+#include "lock.h"
+#include "lvb_table.h"
+#include "user.h"
+#include "ast.h"
+#include "config.h"
+#include "memory.h"
+
+static const char name_prefix[] = "dlm";
+static const struct file_operations device_fops;
+static atomic_t dlm_monitor_opened;
+static int dlm_monitor_unused = 1;
+
+#ifdef CONFIG_COMPAT
+
+struct dlm_lock_params32 {
+	__u8 mode;
+	__u8 namelen;
+	__u16 unused;
+	__u32 flags;
+	__u32 lkid;
+	__u32 parent;
+	__u64 xid;
+	__u64 timeout;
+	__u32 castparam;
+	__u32 castaddr;
+	__u32 bastparam;
+	__u32 bastaddr;
+	__u32 lksb;
+	char lvb[DLM_USER_LVB_LEN];
+	char name[];
+};
+
+struct dlm_write_request32 {
+	__u32 version[3];
+	__u8 cmd;
+	__u8 is64bit;
+	__u8 unused[2];
+
+	union  {
+		struct dlm_lock_params32 lock;
+		struct dlm_lspace_params lspace;
+		struct dlm_purge_params purge;
+	} i;
+};
+
+struct dlm_lksb32 {
+	__u32 sb_status;
+	__u32 sb_lkid;
+	__u8 sb_flags;
+	__u32 sb_lvbptr;
+};
+
+struct dlm_lock_result32 {
+	__u32 version[3];
+	__u32 length;
+	__u32 user_astaddr;
+	__u32 user_astparam;
+	__u32 user_lksb;
+	struct dlm_lksb32 lksb;
+	__u8 bast_mode;
+	__u8 unused[3];
+	/* Offsets may be zero if no data is present */
+	__u32 lvb_offset;
+};
+
+static void compat_input(struct dlm_write_request *kb,
+			 struct dlm_write_request32 *kb32,
+			 int namelen)
+{
+	kb->version[0] = kb32->version[0];
+	kb->version[1] = kb32->version[1];
+	kb->version[2] = kb32->version[2];
+
+	kb->cmd = kb32->cmd;
+	kb->is64bit = kb32->is64bit;
+	if (kb->cmd == DLM_USER_CREATE_LOCKSPACE ||
+	    kb->cmd == DLM_USER_REMOVE_LOCKSPACE) {
+		kb->i.lspace.flags = kb32->i.lspace.flags;
+		kb->i.lspace.minor = kb32->i.lspace.minor;
+		memcpy(kb->i.lspace.name, kb32->i.lspace.name, namelen);
+	} else if (kb->cmd == DLM_USER_PURGE) {
+		kb->i.purge.nodeid = kb32->i.purge.nodeid;
+		kb->i.purge.pid = kb32->i.purge.pid;
+	} else {
+		kb->i.lock.mode = kb32->i.lock.mode;
+		kb->i.lock.namelen = kb32->i.lock.namelen;
+		kb->i.lock.flags = kb32->i.lock.flags;
+		kb->i.lock.lkid = kb32->i.lock.lkid;
+		kb->i.lock.parent = kb32->i.lock.parent;
+		kb->i.lock.xid = kb32->i.lock.xid;
+		kb->i.lock.timeout = kb32->i.lock.timeout;
+		kb->i.lock.castparam = (__user void *)(long)kb32->i.lock.castparam;
+		kb->i.lock.castaddr = (__user void *)(long)kb32->i.lock.castaddr;
+		kb->i.lock.bastparam = (__user void *)(long)kb32->i.lock.bastparam;
+		kb->i.lock.bastaddr = (__user void *)(long)kb32->i.lock.bastaddr;
+		kb->i.lock.lksb = (__user void *)(long)kb32->i.lock.lksb;
+		memcpy(kb->i.lock.lvb, kb32->i.lock.lvb, DLM_USER_LVB_LEN);
+		memcpy(kb->i.lock.name, kb32->i.lock.name, namelen);
+	}
+}
+
+static void compat_output(struct dlm_lock_result *res,
+			  struct dlm_lock_result32 *res32)
+{
+	memset(res32, 0, sizeof(*res32));
+
+	res32->version[0] = res->version[0];
+	res32->version[1] = res->version[1];
+	res32->version[2] = res->version[2];
+
+	res32->user_astaddr = (__u32)(__force long)res->user_astaddr;
+	res32->user_astparam = (__u32)(__force long)res->user_astparam;
+	res32->user_lksb = (__u32)(__force long)res->user_lksb;
+	res32->bast_mode = res->bast_mode;
+
+	res32->lvb_offset = res->lvb_offset;
+	res32->length = res->length;
+
+	res32->lksb.sb_status = res->lksb.sb_status;
+	res32->lksb.sb_flags = res->lksb.sb_flags;
+	res32->lksb.sb_lkid = res->lksb.sb_lkid;
+	res32->lksb.sb_lvbptr = (__u32)(long)res->lksb.sb_lvbptr;
+}
+#endif
+
+/* Figure out if this lock is at the end of its life and no longer
+   available for the application to use.  The lkb still exists until
+   the final ast is read.  A lock becomes EOL in three situations:
+     1. a noqueue request fails with EAGAIN
+     2. an unlock completes with EUNLOCK
+     3. a cancel of a waiting request completes with ECANCEL/EDEADLK
+   An EOL lock needs to be removed from the process's list of locks.
+   And we can't allow any new operation on an EOL lock.  This is
+   not related to the lifetime of the lkb struct which is managed
+   entirely by refcount. */
+
+static int lkb_is_endoflife(int mode, int status)
+{
+	switch (status) {
+	case -DLM_EUNLOCK:
+		return 1;
+	case -DLM_ECANCEL:
+	case -ETIMEDOUT:
+	case -EDEADLK:
+	case -EAGAIN:
+		if (mode == DLM_LOCK_IV)
+			return 1;
+		break;
+	}
+	return 0;
+}
+
+/* we could possibly check if the cancel of an orphan has resulted in the lkb
+   being removed and then remove that lkb from the orphans list and free it */
+
+void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
+		      int status, uint32_t sbflags)
+{
+	struct dlm_ls *ls;
+	struct dlm_user_args *ua;
+	struct dlm_user_proc *proc;
+	struct dlm_callback *cb;
+	int rv, copy_lvb;
+
+	if (test_bit(DLM_DFL_ORPHAN_BIT, &lkb->lkb_dflags) ||
+	    test_bit(DLM_IFL_DEAD_BIT, &lkb->lkb_iflags))
+		return;
+
+	ls = lkb->lkb_resource->res_ls;
+	spin_lock_bh(&ls->ls_clear_proc_locks);
+
+	/* If ORPHAN/DEAD flag is set, it means the process is dead so an ast
+	   can't be delivered.  For ORPHAN's, dlm_clear_proc_locks() freed
+	   lkb->ua so we can't try to use it.  This second check is necessary
+	   for cases where a completion ast is received for an operation that
+	   began before clear_proc_locks did its cancel/unlock. */
+
+	if (test_bit(DLM_DFL_ORPHAN_BIT, &lkb->lkb_dflags) ||
+	    test_bit(DLM_IFL_DEAD_BIT, &lkb->lkb_iflags))
+		goto out;
+
+	DLM_ASSERT(lkb->lkb_ua, dlm_print_lkb(lkb););
+	ua = lkb->lkb_ua;
+	proc = ua->proc;
+
+	if ((flags & DLM_CB_BAST) && ua->bastaddr == NULL)
+		goto out;
+
+	if ((flags & DLM_CB_CAST) && lkb_is_endoflife(mode, status))
+		set_bit(DLM_IFL_ENDOFLIFE_BIT, &lkb->lkb_iflags);
+
+	spin_lock_bh(&proc->asts_spin);
+
+	if (!dlm_may_skip_callback(lkb, flags, mode, status, sbflags,
+				   &copy_lvb)) {
+		rv = dlm_get_cb(lkb, flags, mode, status, sbflags, &cb);
+		if (!rv) {
+			cb->copy_lvb = copy_lvb;
+			cb->ua = *ua;
+			cb->lkb_lksb = &cb->ua.lksb;
+			if (copy_lvb) {
+				memcpy(cb->lvbptr, ua->lksb.sb_lvbptr,
+				       DLM_USER_LVB_LEN);
+				cb->lkb_lksb->sb_lvbptr = cb->lvbptr;
+			}
+
+			list_add_tail(&cb->list, &proc->asts);
+			wake_up_interruptible(&proc->wait);
+		}
+	}
+	spin_unlock_bh(&proc->asts_spin);
+
+	if (test_bit(DLM_IFL_ENDOFLIFE_BIT, &lkb->lkb_iflags)) {
+		/* N.B. spin_lock locks_spin, not asts_spin */
+		spin_lock_bh(&proc->locks_spin);
+		if (!list_empty(&lkb->lkb_ownqueue)) {
+			list_del_init(&lkb->lkb_ownqueue);
+			dlm_put_lkb(lkb);
+		}
+		spin_unlock_bh(&proc->locks_spin);
+	}
+ out:
+	spin_unlock_bh(&ls->ls_clear_proc_locks);
+}
+
+static int device_user_lock(struct dlm_user_proc *proc,
+			    struct dlm_lock_params *params)
+{
+	struct dlm_ls *ls;
+	struct dlm_user_args *ua;
+	uint32_t lkid;
+	int error = -ENOMEM;
+
+	ls = dlm_find_lockspace_local(proc->lockspace);
+	if (!ls)
+		return -ENOENT;
+
+	if (!params->castaddr || !params->lksb) {
+		error = -EINVAL;
+		goto out;
+	}
+
+	ua = kzalloc_obj(struct dlm_user_args, GFP_NOFS);
+	if (!ua)
+		goto out;
+	ua->proc = proc;
+	ua->user_lksb = params->lksb;
+	ua->castparam = params->castparam;
+	ua->castaddr = params->castaddr;
+	ua->bastparam = params->bastparam;
+	ua->bastaddr = params->bastaddr;
+	ua->xid = params->xid;
+
+	if (params->flags & DLM_LKF_CONVERT) {
+		error = dlm_user_convert(ls, ua,
+					 params->mode, params->flags,
+					 params->lkid, params->lvb);
+	} else if (params->flags & DLM_LKF_ORPHAN) {
+		error = dlm_user_adopt_orphan(ls, ua,
+					 params->mode, params->flags,
+					 params->name, params->namelen,
+					 &lkid);
+		if (!error)
+			error = lkid;
+	} else {
+		error = dlm_user_request(ls, ua,
+					 params->mode, params->flags,
+					 params->name, params->namelen);
+		if (!error)
+			error = ua->lksb.sb_lkid;
+	}
+ out:
+	dlm_put_lockspace(ls);
+	return error;
+}
+
+static int device_user_unlock(struct dlm_user_proc *proc,
+			      struct dlm_lock_params *params)
+{
+	struct dlm_ls *ls;
+	struct dlm_user_args *ua;
+	int error = -ENOMEM;
+
+	ls = dlm_find_lockspace_local(proc->lockspace);
+	if (!ls)
+		return -ENOENT;
+
+	ua = kzalloc_obj(struct dlm_user_args, GFP_NOFS);
+	if (!ua)
+		goto out;
+	ua->proc = proc;
+	ua->user_lksb = params->lksb;
+	ua->castparam = params->castparam;
+	ua->castaddr = params->castaddr;
+
+	if (params->flags & DLM_LKF_CANCEL)
+		error = dlm_user_cancel(ls, ua, params->flags, params->lkid);
+	else
+		error = dlm_user_unlock(ls, ua, params->flags, params->lkid,
+					params->lvb);
+ out:
+	dlm_put_lockspace(ls);
+	return error;
+}
+
+static int device_user_deadlock(struct dlm_user_proc *proc,
+				struct dlm_lock_params *params)
+{
+	struct dlm_ls *ls;
+	int error;
+
+	ls = dlm_find_lockspace_local(proc->lockspace);
+	if (!ls)
+		return -ENOENT;
+
+	error = dlm_user_deadlock(ls, params->flags, params->lkid);
+
+	dlm_put_lockspace(ls);
+	return error;
+}
+
+static int dlm_device_register(struct dlm_ls *ls, char *name)
+{
+	int error, len;
+
+	/* The device is already registered.  This happens when the
+	   lockspace is created multiple times from userspace. */
+	if (ls->ls_device.name)
+		return 0;
+
+	error = -ENOMEM;
+	len = strlen(name) + strlen(name_prefix) + 2;
+	ls->ls_device.name = kzalloc(len, GFP_NOFS);
+	if (!ls->ls_device.name)
+		goto fail;
+
+	snprintf((char *)ls->ls_device.name, len, "%s_%s", name_prefix,
+		 name);
+	ls->ls_device.fops = &device_fops;
+	ls->ls_device.minor = MISC_DYNAMIC_MINOR;
+
+	error = misc_register(&ls->ls_device);
+	if (error) {
+		kfree(ls->ls_device.name);
+		/* this has to be set to NULL
+		 * to avoid a double-free in dlm_device_deregister
+		 */
+		ls->ls_device.name = NULL;
+	}
+fail:
+	return error;
+}
+
+int dlm_device_deregister(struct dlm_ls *ls)
+{
+	/* The device is not registered.  This happens when the lockspace
+	   was never used from userspace, or when device_create_lockspace()
+	   calls dlm_release_lockspace() after the register fails. */
+	if (!ls->ls_device.name)
+		return 0;
+
+	misc_deregister(&ls->ls_device);
+	kfree(ls->ls_device.name);
+	return 0;
+}
+
+static int device_user_purge(struct dlm_user_proc *proc,
+			     struct dlm_purge_params *params)
+{
+	struct dlm_ls *ls;
+	int error;
+
+	ls = dlm_find_lockspace_local(proc->lockspace);
+	if (!ls)
+		return -ENOENT;
+
+	error = dlm_user_purge(ls, proc, params->nodeid, params->pid);
+
+	dlm_put_lockspace(ls);
+	return error;
+}
+
+static int device_create_lockspace(struct dlm_lspace_params *params)
+{
+	dlm_lockspace_t *lockspace;
+	struct dlm_ls *ls;
+	int error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	error = dlm_new_user_lockspace(params->name, dlm_config.ci_cluster_name,
+				       params->flags, DLM_USER_LVB_LEN, NULL,
+				       NULL, NULL, &lockspace);
+	if (error)
+		return error;
+
+	ls = dlm_find_lockspace_local(lockspace);
+	if (!ls)
+		return -ENOENT;
+
+	error = dlm_device_register(ls, params->name);
+	dlm_put_lockspace(ls);
+
+	if (error)
+		dlm_release_lockspace(lockspace, DLM_RELEASE_NO_LOCKS);
+	else
+		error = ls->ls_device.minor;
+
+	return error;
+}
+
+static int device_remove_lockspace(struct dlm_lspace_params *params)
+{
+	dlm_lockspace_t *lockspace;
+	struct dlm_ls *ls;
+	int error, force = DLM_RELEASE_NO_LOCKS;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	ls = dlm_find_lockspace_device(params->minor);
+	if (!ls)
+		return -ENOENT;
+
+	if (params->flags & DLM_USER_LSFLG_FORCEFREE)
+		force = DLM_RELEASE_NORMAL;
+
+	lockspace = ls;
+	dlm_put_lockspace(ls);
+
+	/* The final dlm_release_lockspace waits for references to go to
+	   zero, so all processes will need to close their device for the
+	   ls before the release will proceed.  release also calls the
+	   device_deregister above.  Converting a positive return value
+	   from release to zero means that userspace won't know when its
+	   release was the final one, but it shouldn't need to know. */
+
+	error = dlm_release_lockspace(lockspace, force);
+	if (error > 0)
+		error = 0;
+	return error;
+}
+
+/* Check the user's version matches ours */
+static int check_version(struct dlm_write_request *req)
+{
+	if (req->version[0] != DLM_DEVICE_VERSION_MAJOR ||
+	    (req->version[0] == DLM_DEVICE_VERSION_MAJOR &&
+	     req->version[1] > DLM_DEVICE_VERSION_MINOR)) {
+
+		printk(KERN_DEBUG "dlm: process %s (%d) version mismatch "
+		       "user (%d.%d.%d) kernel (%d.%d.%d)\n",
+		       current->comm,
+		       task_pid_nr(current),
+		       req->version[0],
+		       req->version[1],
+		       req->version[2],
+		       DLM_DEVICE_VERSION_MAJOR,
+		       DLM_DEVICE_VERSION_MINOR,
+		       DLM_DEVICE_VERSION_PATCH);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+/*
+ * device_write
+ *
+ *   device_user_lock
+ *     dlm_user_request -> request_lock
+ *     dlm_user_convert -> convert_lock
+ *
+ *   device_user_unlock
+ *     dlm_user_unlock -> unlock_lock
+ *     dlm_user_cancel -> cancel_lock
+ *
+ *   device_create_lockspace
+ *     dlm_new_lockspace
+ *
+ *   device_remove_lockspace
+ *     dlm_release_lockspace
+ */
+
+/* a write to a lockspace device is a lock or unlock request, a write
+   to the control device is to create/remove a lockspace */
+
+static ssize_t device_write(struct file *file, const char __user *buf,
+			    size_t count, loff_t *ppos)
+{
+	struct dlm_user_proc *proc = file->private_data;
+	struct dlm_write_request *kbuf;
+	int error;
+
+#ifdef CONFIG_COMPAT
+	if (count < sizeof(struct dlm_write_request32))
+#else
+	if (count < sizeof(struct dlm_write_request))
+#endif
+		return -EINVAL;
+
+	/*
+	 * can't compare against COMPAT/dlm_write_request32 because
+	 * we don't yet know if is64bit is zero
+	 */
+	if (count > sizeof(struct dlm_write_request) + DLM_RESNAME_MAXLEN)
+		return -EINVAL;
+
+	kbuf = memdup_user_nul(buf, count);
+	if (IS_ERR(kbuf))
+		return PTR_ERR(kbuf);
+
+	if (check_version(kbuf)) {
+		error = -EBADE;
+		goto out_free;
+	}
+
+#ifdef CONFIG_COMPAT
+	if (!kbuf->is64bit) {
+		struct dlm_write_request32 *k32buf;
+		int namelen = 0;
+
+		if (count > sizeof(struct dlm_write_request32))
+			namelen = count - sizeof(struct dlm_write_request32);
+
+		k32buf = (struct dlm_write_request32 *)kbuf;
+
+		/* add 1 after namelen so that the name string is terminated */
+		kbuf = kzalloc(sizeof(struct dlm_write_request) + namelen + 1,
+			       GFP_NOFS);
+		if (!kbuf) {
+			kfree(k32buf);
+			return -ENOMEM;
+		}
+
+		if (proc)
+			set_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags);
+
+		compat_input(kbuf, k32buf, namelen);
+		kfree(k32buf);
+	}
+#endif
+
+	/* do we really need this? can a write happen after a close? */
+	if ((kbuf->cmd == DLM_USER_LOCK || kbuf->cmd == DLM_USER_UNLOCK) &&
+	    (proc && test_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags))) {
+		error = -EINVAL;
+		goto out_free;
+	}
+
+	error = -EINVAL;
+
+	switch (kbuf->cmd)
+	{
+	case DLM_USER_LOCK:
+		if (!proc) {
+			log_print("no locking on control device");
+			goto out_free;
+		}
+		error = device_user_lock(proc, &kbuf->i.lock);
+		break;
+
+	case DLM_USER_UNLOCK:
+		if (!proc) {
+			log_print("no locking on control device");
+			goto out_free;
+		}
+		error = device_user_unlock(proc, &kbuf->i.lock);
+		break;
+
+	case DLM_USER_DEADLOCK:
+		if (!proc) {
+			log_print("no locking on control device");
+			goto out_free;
+		}
+		error = device_user_deadlock(proc, &kbuf->i.lock);
+		break;
+
+	case DLM_USER_CREATE_LOCKSPACE:
+		if (proc) {
+			log_print("create/remove only on control device");
+			goto out_free;
+		}
+		error = device_create_lockspace(&kbuf->i.lspace);
+		break;
+
+	case DLM_USER_REMOVE_LOCKSPACE:
+		if (proc) {
+			log_print("create/remove only on control device");
+			goto out_free;
+		}
+		error = device_remove_lockspace(&kbuf->i.lspace);
+		break;
+
+	case DLM_USER_PURGE:
+		if (!proc) {
+			log_print("no locking on control device");
+			goto out_free;
+		}
+		error = device_user_purge(proc, &kbuf->i.purge);
+		break;
+
+	default:
+		log_print("Unknown command passed to DLM device : %d\n",
+			  kbuf->cmd);
+	}
+
+ out_free:
+	kfree(kbuf);
+	return error;
+}
+
+/* Every process that opens the lockspace device has its own "proc" structure
+   hanging off the open file that's used to keep track of locks owned by the
+   process and asts that need to be delivered to the process. */
+
+static int device_open(struct inode *inode, struct file *file)
+{
+	struct dlm_user_proc *proc;
+	struct dlm_ls *ls;
+
+	ls = dlm_find_lockspace_device(iminor(inode));
+	if (!ls)
+		return -ENOENT;
+
+	proc = kzalloc_obj(struct dlm_user_proc, GFP_NOFS);
+	if (!proc) {
+		dlm_put_lockspace(ls);
+		return -ENOMEM;
+	}
+
+	proc->lockspace = ls;
+	INIT_LIST_HEAD(&proc->asts);
+	INIT_LIST_HEAD(&proc->locks);
+	INIT_LIST_HEAD(&proc->unlocking);
+	spin_lock_init(&proc->asts_spin);
+	spin_lock_init(&proc->locks_spin);
+	init_waitqueue_head(&proc->wait);
+	file->private_data = proc;
+
+	return 0;
+}
+
+static int device_close(struct inode *inode, struct file *file)
+{
+	struct dlm_user_proc *proc = file->private_data;
+	struct dlm_ls *ls;
+
+	ls = dlm_find_lockspace_local(proc->lockspace);
+	if (!ls)
+		return -ENOENT;
+
+	set_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags);
+
+	dlm_clear_proc_locks(ls, proc);
+
+	/* at this point no more lkb's should exist for this lockspace,
+	   so there's no chance of dlm_user_add_ast() being called and
+	   looking for lkb->ua->proc */
+
+	kfree(proc);
+	file->private_data = NULL;
+
+	dlm_put_lockspace(ls);
+	dlm_put_lockspace(ls);  /* for the find in device_open() */
+
+	/* FIXME: AUTOFREE: if this ls is no longer used do
+	   device_remove_lockspace() */
+
+	return 0;
+}
+
+static int copy_result_to_user(struct dlm_user_args *ua, int compat,
+			       uint32_t flags, int mode, int copy_lvb,
+			       char __user *buf, size_t count)
+{
+#ifdef CONFIG_COMPAT
+	struct dlm_lock_result32 result32;
+#endif
+	struct dlm_lock_result result;
+	void *resultptr;
+	int error=0;
+	int len;
+	int struct_len;
+
+	memset(&result, 0, sizeof(struct dlm_lock_result));
+	result.version[0] = DLM_DEVICE_VERSION_MAJOR;
+	result.version[1] = DLM_DEVICE_VERSION_MINOR;
+	result.version[2] = DLM_DEVICE_VERSION_PATCH;
+	memcpy(&result.lksb, &ua->lksb, offsetof(struct dlm_lksb, sb_lvbptr));
+	result.user_lksb = ua->user_lksb;
+
+	/* FIXME: dlm1 provides for the user's bastparam/addr to not be updated
+	   in a conversion unless the conversion is successful.  See code
+	   in dlm_user_convert() for updating ua from ua_tmp.  OpenVMS, though,
+	   notes that a new blocking AST address and parameter are set even if
+	   the conversion fails, so maybe we should just do that. */
+
+	if (flags & DLM_CB_BAST) {
+		result.user_astaddr = ua->bastaddr;
+		result.user_astparam = ua->bastparam;
+		result.bast_mode = mode;
+	} else {
+		result.user_astaddr = ua->castaddr;
+		result.user_astparam = ua->castparam;
+	}
+
+#ifdef CONFIG_COMPAT
+	if (compat)
+		len = sizeof(struct dlm_lock_result32);
+	else
+#endif
+		len = sizeof(struct dlm_lock_result);
+	struct_len = len;
+
+	/* copy lvb to userspace if there is one, it's been updated, and
+	   the user buffer has space for it */
+
+	if (copy_lvb && ua->lksb.sb_lvbptr && count >= len + DLM_USER_LVB_LEN) {
+		if (copy_to_user(buf+len, ua->lksb.sb_lvbptr,
+				 DLM_USER_LVB_LEN)) {
+			error = -EFAULT;
+			goto out;
+		}
+
+		result.lvb_offset = len;
+		len += DLM_USER_LVB_LEN;
+	}
+
+	result.length = len;
+	resultptr = &result;
+#ifdef CONFIG_COMPAT
+	if (compat) {
+		compat_output(&result, &result32);
+		resultptr = &result32;
+	}
+#endif
+
+	if (copy_to_user(buf, resultptr, struct_len))
+		error = -EFAULT;
+	else
+		error = len;
+ out:
+	return error;
+}
+
+static int copy_version_to_user(char __user *buf, size_t count)
+{
+	struct dlm_device_version ver;
+
+	memset(&ver, 0, sizeof(struct dlm_device_version));
+	ver.version[0] = DLM_DEVICE_VERSION_MAJOR;
+	ver.version[1] = DLM_DEVICE_VERSION_MINOR;
+	ver.version[2] = DLM_DEVICE_VERSION_PATCH;
+
+	if (copy_to_user(buf, &ver, sizeof(struct dlm_device_version)))
+		return -EFAULT;
+	return sizeof(struct dlm_device_version);
+}
+
+/* a read returns a single ast described in a struct dlm_lock_result */
+
+static ssize_t device_read(struct file *file, char __user *buf, size_t count,
+			   loff_t *ppos)
+{
+	struct dlm_user_proc *proc = file->private_data;
+	DECLARE_WAITQUEUE(wait, current);
+	struct dlm_callback *cb;
+	int rv, ret;
+
+	if (count == sizeof(struct dlm_device_version)) {
+		rv = copy_version_to_user(buf, count);
+		return rv;
+	}
+
+	if (!proc) {
+		log_print("non-version read from control device %zu", count);
+		return -EINVAL;
+	}
+
+#ifdef CONFIG_COMPAT
+	if (count < sizeof(struct dlm_lock_result32))
+#else
+	if (count < sizeof(struct dlm_lock_result))
+#endif
+		return -EINVAL;
+
+	/* do we really need this? can a read happen after a close? */
+	if (test_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags))
+		return -EINVAL;
+
+	spin_lock_bh(&proc->asts_spin);
+	if (list_empty(&proc->asts)) {
+		if (file->f_flags & O_NONBLOCK) {
+			spin_unlock_bh(&proc->asts_spin);
+			return -EAGAIN;
+		}
+
+		add_wait_queue(&proc->wait, &wait);
+
+	repeat:
+		set_current_state(TASK_INTERRUPTIBLE);
+		if (list_empty(&proc->asts) && !signal_pending(current)) {
+			spin_unlock_bh(&proc->asts_spin);
+			schedule();
+			spin_lock_bh(&proc->asts_spin);
+			goto repeat;
+		}
+		set_current_state(TASK_RUNNING);
+		remove_wait_queue(&proc->wait, &wait);
+
+		if (signal_pending(current)) {
+			spin_unlock_bh(&proc->asts_spin);
+			return -ERESTARTSYS;
+		}
+	}
+
+	/* if we empty lkb_callbacks, we don't want to unlock the spinlock
+	   without removing lkb_cb_list; so empty lkb_cb_list is always
+	   consistent with empty lkb_callbacks */
+
+	cb = list_first_entry(&proc->asts, struct dlm_callback, list);
+	list_del(&cb->list);
+	spin_unlock_bh(&proc->asts_spin);
+
+	if (cb->flags & DLM_CB_BAST) {
+		trace_dlm_bast(cb->ls_id, cb->lkb_id, cb->mode, cb->res_name,
+			       cb->res_length);
+	} else if (cb->flags & DLM_CB_CAST) {
+		cb->lkb_lksb->sb_status = cb->sb_status;
+		cb->lkb_lksb->sb_flags = cb->sb_flags;
+		trace_dlm_ast(cb->ls_id, cb->lkb_id, cb->sb_status,
+			      cb->sb_flags, cb->res_name, cb->res_length);
+	}
+
+	ret = copy_result_to_user(&cb->ua,
+				  test_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags),
+				  cb->flags, cb->mode, cb->copy_lvb, buf, count);
+	dlm_free_cb(cb);
+	return ret;
+}
+
+static __poll_t device_poll(struct file *file, poll_table *wait)
+{
+	struct dlm_user_proc *proc = file->private_data;
+
+	poll_wait(file, &proc->wait, wait);
+
+	spin_lock_bh(&proc->asts_spin);
+	if (!list_empty(&proc->asts)) {
+		spin_unlock_bh(&proc->asts_spin);
+		return EPOLLIN | EPOLLRDNORM;
+	}
+	spin_unlock_bh(&proc->asts_spin);
+	return 0;
+}
+
+int dlm_user_daemon_available(void)
+{
+	/* dlm_controld hasn't started (or, has started, but not
+	   properly populated configfs) */
+
+	if (!dlm_our_nodeid())
+		return 0;
+
+	/* This is to deal with versions of dlm_controld that don't
+	   know about the monitor device.  We assume that if the
+	   dlm_controld was started (above), but the monitor device
+	   was never opened, that it's an old version.  dlm_controld
+	   should open the monitor device before populating configfs. */
+
+	if (dlm_monitor_unused)
+		return 1;
+
+	return atomic_read(&dlm_monitor_opened) ? 1 : 0;
+}
+
+static int ctl_device_open(struct inode *inode, struct file *file)
+{
+	file->private_data = NULL;
+	return 0;
+}
+
+static int ctl_device_close(struct inode *inode, struct file *file)
+{
+	return 0;
+}
+
+static int monitor_device_open(struct inode *inode, struct file *file)
+{
+	atomic_inc(&dlm_monitor_opened);
+	dlm_monitor_unused = 0;
+	return 0;
+}
+
+static int monitor_device_close(struct inode *inode, struct file *file)
+{
+	if (atomic_dec_and_test(&dlm_monitor_opened))
+		dlm_stop_lockspaces();
+	return 0;
+}
+
+static const struct file_operations device_fops = {
+	.open    = device_open,
+	.release = device_close,
+	.read    = device_read,
+	.write   = device_write,
+	.poll    = device_poll,
+	.owner   = THIS_MODULE,
+	.llseek  = noop_llseek,
+};
+
+static const struct file_operations ctl_device_fops = {
+	.open    = ctl_device_open,
+	.release = ctl_device_close,
+	.read    = device_read,
+	.write   = device_write,
+	.owner   = THIS_MODULE,
+	.llseek  = noop_llseek,
+};
+
+static struct miscdevice ctl_device = {
+	.name  = "dlm-control",
+	.fops  = &ctl_device_fops,
+	.minor = MISC_DYNAMIC_MINOR,
+};
+
+static const struct file_operations monitor_device_fops = {
+	.open    = monitor_device_open,
+	.release = monitor_device_close,
+	.owner   = THIS_MODULE,
+	.llseek  = noop_llseek,
+};
+
+static struct miscdevice monitor_device = {
+	.name  = "dlm-monitor",
+	.fops  = &monitor_device_fops,
+	.minor = MISC_DYNAMIC_MINOR,
+};
+
+int __init dlm_user_init(void)
+{
+	int error;
+
+	atomic_set(&dlm_monitor_opened, 0);
+
+	error = misc_register(&ctl_device);
+	if (error) {
+		log_print("misc_register failed for control device");
+		goto out;
+	}
+
+	error = misc_register(&monitor_device);
+	if (error) {
+		log_print("misc_register failed for monitor device");
+		misc_deregister(&ctl_device);
+	}
+ out:
+	return error;
+}
+
+void dlm_user_exit(void)
+{
+	misc_deregister(&ctl_device);
+	misc_deregister(&monitor_device);
+}
+

linux-7.2.3/fs/dlm/user.h

@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2006-2010 Red Hat, Inc.  All rights reserved.
+ */
+
+#ifndef __USER_DOT_H__
+#define __USER_DOT_H__
+
+void dlm_purge_lkb_callbacks(struct dlm_lkb *lkb);
+void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
+		      int status, uint32_t sbflags);
+int dlm_user_init(void);
+void dlm_user_exit(void);
+int dlm_device_deregister(struct dlm_ls *ls);
+int dlm_user_daemon_available(void);
+
+#endif

linux-7.2.3/fs/dlm/util.c

@@ -0,0 +1,66 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#include "dlm_internal.h"
+#include "rcom.h"
+#include "util.h"
+
+#define DLM_ERRNO_EDEADLK		35
+#define DLM_ERRNO_EBADR			53
+#define DLM_ERRNO_EBADSLT		57
+#define DLM_ERRNO_EPROTO		71
+#define DLM_ERRNO_EOPNOTSUPP		95
+#define DLM_ERRNO_ETIMEDOUT	       110
+#define DLM_ERRNO_EINPROGRESS	       115
+
+/* higher errno values are inconsistent across architectures, so select
+   one set of values for on the wire */
+
+int to_dlm_errno(int err)
+{
+	switch (err) {
+	case -EDEADLK:
+		return -DLM_ERRNO_EDEADLK;
+	case -EBADR:
+		return -DLM_ERRNO_EBADR;
+	case -EBADSLT:
+		return -DLM_ERRNO_EBADSLT;
+	case -EPROTO:
+		return -DLM_ERRNO_EPROTO;
+	case -EOPNOTSUPP:
+		return -DLM_ERRNO_EOPNOTSUPP;
+	case -ETIMEDOUT:
+		return -DLM_ERRNO_ETIMEDOUT;
+	case -EINPROGRESS:
+		return -DLM_ERRNO_EINPROGRESS;
+	}
+	return err;
+}
+
+int from_dlm_errno(int err)
+{
+	switch (err) {
+	case -DLM_ERRNO_EDEADLK:
+		return -EDEADLK;
+	case -DLM_ERRNO_EBADR:
+		return -EBADR;
+	case -DLM_ERRNO_EBADSLT:
+		return -EBADSLT;
+	case -DLM_ERRNO_EPROTO:
+		return -EPROTO;
+	case -DLM_ERRNO_EOPNOTSUPP:
+		return -EOPNOTSUPP;
+	case -DLM_ERRNO_ETIMEDOUT:
+		return -ETIMEDOUT;
+	case -DLM_ERRNO_EINPROGRESS:
+		return -EINPROGRESS;
+	}
+	return err;
+}

linux-7.2.3/fs/dlm/util.h

@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2005 Red Hat, Inc.  All rights reserved.
+**
+**
+*******************************************************************************
+******************************************************************************/
+
+#ifndef __UTIL_DOT_H__
+#define __UTIL_DOT_H__
+
+int to_dlm_errno(int err);
+int from_dlm_errno(int err);
+
+#endif
+