xref: /illumos-gate/usr/src/cmd/fm/modules/common/disk-monitor/disk_monitor.c (revision 0244979b)

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */

/*
 * Copyright 2009 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 * Copyright 2016 Nexenta Systems, Inc.  All rights reserved.
 */

/*
 * Disk Monitor
 */
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <time.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <stdarg.h>
#include <errno.h>
#include <signal.h>
#include <unistd.h>
#include <pthread.h>
#include <libnvpair.h>
#include <fm/fmd_api.h>
#include <fm/fmd_fmri.h>
#include <sys/fm/protocol.h>
#include <sys/fm/io/disk.h>
#include <fm/libtopo.h>

#include "disk_monitor.h"
#include "hotplug_mgr.h"
#include "schg_mgr.h"
#include "topo_gather.h"
#include "dm_platform.h"

#define	THIS_FMD_MODULE_NAME "disk-monitor"

static enum disk_init_state {
	INIT_STATE_NONE = 0,
	STATE_CHANGE_MGR_INITTED = 2,
	HOTPLUG_MGR_INITTED = 4
} g_init_state = INIT_STATE_NONE;

typedef enum {
	LT_SUSPECT,
	LT_REPAIRED
} fm_list_type_t;

/*
 * Global verbosity flag -- controls chattiness of debug messages and
 * warnings.  Its value is determined by the fmd property "log-level"
 * settable in the DE's .conf file.
 */
log_class_t			g_verbose = 0;
cfgdata_t			*config_data = NULL;
fmd_hdl_t			*g_fm_hdl = NULL;

static const fmd_prop_t		fmd_props[];

static void
diskmon_teardown_all(void)
{
	cleanup_hotplug_manager();
	cleanup_state_change_manager(config_data);
	config_fini();
}

static int
count_disks(diskmon_t *disklistp)
{
	int i = 0;

	while (disklistp != NULL) {
		i++;
		disklistp = disklistp->next;
	}

	return (i);
}

static int
diskmon_init(void)
{
	/*
	 * Block the generation of state change events (generated by the
	 * hotplug manager thread) here; they will be unblocked after the
	 * state change manager thread is ready to accept state changes
	 * (shortly after it starts).
	 */
	block_state_change_events();

	if (dm_platform_init() != 0)
		goto cleanup;

	if (init_hotplug_manager() != 0)
		goto cleanup;
	else
		g_init_state |= HOTPLUG_MGR_INITTED;

	if (init_state_change_manager(config_data) != 0)
		goto cleanup;
	else
		g_init_state |= STATE_CHANGE_MGR_INITTED;

	return (E_SUCCESS);

cleanup:

	unblock_state_change_events();

	/*
	 * The cleanup order here does matter, due to dependencies between the
	 * managers.
	 */
	if (g_init_state & HOTPLUG_MGR_INITTED)
		cleanup_hotplug_manager();
	if (g_init_state & STATE_CHANGE_MGR_INITTED)
		cleanup_state_change_manager(config_data);
	dm_platform_fini();

	return (E_ERROR);
}

static void
dm_fault_execute_actions(fmd_hdl_t *hdl, diskmon_t *diskp, nvlist_t *nvl)
{
	const char		*action_prop = NULL;
	const char		*action_string;

	/*
	 * The predictive failure action is the activation of the fault
	 * indicator.
	 */
	if (fmd_nvl_class_match(hdl, nvl,
	    DISK_ERROR_CLASS "." FM_FAULT_DISK_OVERTEMP))
		action_prop = DISK_PROP_OTEMPACTION;

	if (fmd_nvl_class_match(hdl, nvl,
	    DISK_ERROR_CLASS "." FM_FAULT_DISK_TESTFAIL))
		action_prop = DISK_PROP_STFAILACTION;

	if (fmd_nvl_class_match(hdl, nvl,
	    DISK_ERROR_CLASS "." FM_FAULT_SSM_WEAROUT))
		action_prop = DISK_PROP_SSMWEAROUTACTION;

	dm_fault_indicator_set(diskp, INDICATOR_ON);

	if (action_prop != NULL &&
	    (action_string = dm_prop_lookup(diskp->props, action_prop))
	    != NULL) {

		if (dm_platform_indicator_execute(action_string) != 0) {
			log_warn("Fault action `%s' did not successfully "
			    "complete.\n", action_string);
		}
	}
}

static void
diskmon_agent_repair(fmd_hdl_t *hdl, nvlist_t *nvl, int repair)
{
	char		*uuid = NULL;
	nvlist_t	**nva;
	uint_t		nvc;
	diskmon_t	*diskp;
	nvlist_t	*fmri;
	nvlist_t	*fltnvl;
	int		err = 0;

	err |= nvlist_lookup_string(nvl, FM_SUSPECT_UUID, &uuid);
	err |= nvlist_lookup_nvlist_array(nvl, FM_SUSPECT_FAULT_LIST,
	    &nva, &nvc);
	if (err != 0)
		return;

	while (nvc-- != 0) {

		fltnvl = *nva++;

		if (nvlist_lookup_nvlist(fltnvl, FM_FAULT_RESOURCE, &fmri)
		    != 0)
			continue;

		if ((diskp = dm_fmri_to_diskmon(hdl, fmri)) == NULL)
			continue;

		log_msg(MM_MAIN, "Disk %s repaired!\n",
		    diskp->location);

		dm_fault_indicator_set(diskp, INDICATOR_OFF);

		dm_state_change(diskp, HPS_REPAIRED);
	}

	if (repair)
		fmd_case_uuresolved(hdl, uuid);

}

static void
diskmon_agent_suspect(fmd_hdl_t *hdl, nvlist_t *nvl)
{
	char		*uuid = NULL;
	nvlist_t	**nva;
	uint_t		nvc;
	diskmon_t	*diskp;
	nvlist_t	*fmri;
	nvlist_t	*fltnvl;
	int		err = 0;

	err |= nvlist_lookup_string(nvl, FM_SUSPECT_UUID, &uuid);
	err |= nvlist_lookup_nvlist_array(nvl, FM_SUSPECT_FAULT_LIST,
	    &nva, &nvc);
	if (err != 0)
		return;

	while (nvc-- != 0 && !fmd_case_uuclosed(hdl, uuid)) {

		fltnvl = *nva++;

		if (nvlist_lookup_nvlist(fltnvl, FM_FAULT_RESOURCE, &fmri) != 0)
			continue;

		if ((diskp = dm_fmri_to_diskmon(hdl, fmri)) == NULL)
			continue;

		/* Execute the actions associated with this fault */
		dm_fault_execute_actions(hdl, diskp,  fltnvl);

		/*
		 * Send a state change event to the state change manager
		 */
		dm_state_change(diskp, HPS_FAULTED);
	}

	if (!fmd_case_uuclosed(hdl, uuid)) {
		/* Case is closed */
		fmd_case_uuclose(hdl, uuid);
	}
}

/*ARGSUSED*/
static void
diskmon_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl, const char *class)
{
	diskmon_t	*diskp;
	nvlist_t	*fmri;

	if (g_verbose & MM_MAIN)
		nvlist_print(stderr, nvl);

	/*
	 * Act on the fault suspect list or repaired list (embedded agent
	 * action).
	 */
	if (fmd_nvl_class_match(hdl, nvl, FM_LIST_REPAIRED_CLASS)) {

		diskmon_agent_repair(hdl, nvl, 1);
		return;

	} else if (fmd_nvl_class_match(hdl, nvl, FM_LIST_UPDATED_CLASS)) {

		diskmon_agent_repair(hdl, nvl, 0);
		return;

	} else if (fmd_nvl_class_match(hdl, nvl, FM_LIST_SUSPECT_CLASS)) {

		diskmon_agent_suspect(hdl, nvl);
		return;
	} else if (fmd_nvl_class_match(hdl, nvl, FM_LIST_RESOLVED_CLASS)) {
		return;
	}

	/*
	 * If we get any replayed faults, set the diskmon's faulted
	 * flag for the appropriate fault, then change the diskmon's state
	 * to faulted.
	 */
	if (fmd_nvl_class_match(hdl, nvl, DISK_ERROR_CLASS ".*")) {

		if (nvlist_lookup_nvlist(nvl, FM_FAULT_RESOURCE,
		    &fmri) != 0)
			return;

		if ((diskp = dm_fmri_to_diskmon(hdl, fmri)) == NULL)
			return;

		/* Execute the actions associated with this fault */
		dm_fault_execute_actions(hdl, diskp, nvl);

		/*
		 * If the fault wasn't generated by this module, send a
		 * state change event to the state change manager
		 */
		dm_state_change(diskp, HPS_FAULTED);
		return;
	}
}

static const fmd_hdl_ops_t fmd_ops = {
	diskmon_recv,	/* fmdo_recv */
	NULL,		/* fmdo_timeout */
	NULL,		/* fmdo_close */
	NULL,		/* fmdo_stats */
	NULL,		/* fmdo_gc */
};

static const fmd_prop_t fmd_props[] = {
	{ GLOBAL_PROP_LOG_LEVEL, FMD_TYPE_UINT32, "0" },
	{ NULL, 0, NULL }
};

static const fmd_hdl_info_t fmd_info = {
	"Disk Monitor",
	DISK_MONITOR_MODULE_VERSION,
	&fmd_ops,
	fmd_props
};

void
_fmd_init(fmd_hdl_t *hdl)
{
	fmd_case_t	*cp;
	int		disk_count;

	g_fm_hdl = hdl;

	if (fmd_hdl_register(hdl, FMD_API_VERSION, &fmd_info) != 0) {
		return;
	}

	if (config_init()) {
		log_err("Could not initialize configuration!\n");
		fmd_hdl_unregister(hdl);
		return;
	}

	if (config_get(hdl, fmd_props)) {
		config_fini();
		log_err("Could not retrieve configuration from libtopo!\n");
		fmd_hdl_unregister(hdl);
		return;
	}

	/*
	 * If there are no disks to monitor, bail out
	 */
	if ((disk_count = count_disks(config_data->disk_list)) == 0) {
		config_fini();
		fmd_hdl_unregister(hdl);
		return;
	}

	if (diskmon_init() == E_ERROR) {
		config_fini();
		fmd_hdl_unregister(hdl);
		return;
	}

	log_msg(MM_MAIN, "Monitoring %d disks.\n", disk_count);

	/*
	 * Iterate over all active cases.
	 * Since we automatically solve all cases, these cases must have
	 * had the fault added, but the DE must have been interrupted
	 * before they were solved.
	 */
	for (cp = fmd_case_next(hdl, NULL);
	    cp != NULL; cp = fmd_case_next(hdl, cp)) {

		if (!fmd_case_solved(hdl, cp))
			fmd_case_solve(hdl, cp);
	}
}

/*ARGSUSED*/
void
_fmd_fini(fmd_hdl_t *hdl)
{
	diskmon_teardown_all();
	g_fm_hdl = NULL;
}