svn commit: r227669 - user/sbruno/mptd
Sean Bruno
sbruno at FreeBSD.org
Fri Nov 18 17:39:20 UTC 2011
Author: sbruno
Date: Fri Nov 18 17:39:20 2011
New Revision: 227669
URL: http://svn.freebsd.org/changeset/base/227669
Log:
Add the Yahoo! mpt(4) montioring utility for review.
Compiles against amd64 freebsd-current at this time.
Installs a /usr/sbin/mptd and can be started/stopped via the included
rc script.
Obtained from: Yahoo! Inc. and jhb@ in a former life
Added:
user/sbruno/mptd/
user/sbruno/mptd/Makefile
user/sbruno/mptd/mpt_cam.c
user/sbruno/mptd/mpt_cmd.c
user/sbruno/mptd/mptd.c
user/sbruno/mptd/mptd.h
user/sbruno/mptd/mptd.rc (contents, props changed)
Added: user/sbruno/mptd/Makefile
==============================================================================
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ user/sbruno/mptd/Makefile Fri Nov 18 17:39:20 2011 (r227669)
@@ -0,0 +1,12 @@
+PROG= mptd
+BINDIR= /usr/sbin
+
+SRCS= mptd.c mpt_cam.c mpt_cmd.c
+
+CFLAGS+= -g -Wall -Wunused
+
+MAN=
+
+LDADD+= -lcam
+
+.include <bsd.prog.mk>
Added: user/sbruno/mptd/mpt_cam.c
==============================================================================
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ user/sbruno/mptd/mpt_cam.c Fri Nov 18 17:39:20 2011 (r227669)
@@ -0,0 +1,139 @@
+/*-
+ * Copyright (c) 2011 Yahoo! Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ */
+
+#include <sys/param.h>
+#include <err.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include <camlib.h>
+#include <cam/scsi/scsi_message.h>
+#include <cam/scsi/scsi_pass.h>
+
+#include "mptd.h"
+
+static int xptfd;
+
+static int
+xpt_open(void)
+{
+
+ if (xptfd == 0)
+ xptfd = open(XPT_DEVICE, O_RDWR);
+ return (xptfd);
+}
+
+int
+mpt_query_disk(int fd, int unit, U8 VolumeBus, U8 VolumeID,
+ struct mpt_query_disk *qd)
+{
+ struct bus_match_pattern *b;
+ struct periph_match_pattern *p;
+ struct periph_match_result *r;
+ union ccb ccb;
+ size_t bufsize;
+ int i;
+
+ /* mpt(4) only handles devices on bus 0. */
+ if (VolumeBus != 0)
+ return (ENXIO);
+
+ if (xpt_open() < 0)
+ return (ENXIO);
+
+ bzero(&ccb, sizeof(ccb));
+
+ ccb.ccb_h.func_code = XPT_DEV_MATCH;
+
+ bufsize = sizeof(struct dev_match_result) * 5;
+ ccb.cdm.num_matches = 0;
+ ccb.cdm.match_buf_len = bufsize;
+ ccb.cdm.matches = calloc(1, bufsize);
+
+ bufsize = sizeof(struct dev_match_pattern) * 2;
+ ccb.cdm.num_patterns = 2;
+ ccb.cdm.pattern_buf_len = bufsize;
+ ccb.cdm.patterns = calloc(1, bufsize);
+
+ /* Match mptX bus. */
+ ccb.cdm.patterns[0].type = DEV_MATCH_BUS;
+ b = &ccb.cdm.patterns[0].pattern.bus_pattern;
+ snprintf(b->dev_name, sizeof(b->dev_name), "mpt");
+ b->unit_number = unit;
+ b->flags = BUS_MATCH_NAME | BUS_MATCH_UNIT;
+
+ /* Look for a "da" device at the specified target and lun. */
+ ccb.cdm.patterns[1].type = DEV_MATCH_PERIPH;
+ p = &ccb.cdm.patterns[1].pattern.periph_pattern;
+ snprintf(p->periph_name, sizeof(p->periph_name), "da");
+ p->target_id = VolumeID;
+ p->flags = PERIPH_MATCH_NAME | PERIPH_MATCH_TARGET;
+
+ if (ioctl(xptfd, CAMIOCOMMAND, &ccb) < 0) {
+ i = errno;
+ free(ccb.cdm.matches);
+ free(ccb.cdm.patterns);
+ return (i);
+ }
+ free(ccb.cdm.patterns);
+
+ if (((ccb.ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP) ||
+ (ccb.cdm.status != CAM_DEV_MATCH_LAST)) {
+ warnx("mpt_query_disk got CAM error %#x, CDM error %d\n",
+ ccb.ccb_h.status, ccb.cdm.status);
+ free(ccb.cdm.matches);
+ return (EIO);
+ }
+
+ /*
+ * We should have exactly 2 matches, 1 for the bus and 1 for
+ * the peripheral.
+ */
+ if (ccb.cdm.num_matches != 2) {
+ warnx("mpt_query_disk got %d matches, expected 2",
+ ccb.cdm.num_matches);
+ free(ccb.cdm.matches);
+ return (EIO);
+ }
+ if (ccb.cdm.matches[0].type != DEV_MATCH_BUS ||
+ ccb.cdm.matches[1].type != DEV_MATCH_PERIPH) {
+ warnx("mpt_query_disk got wrong CAM matches");
+ free(ccb.cdm.matches);
+ return (EIO);
+ }
+
+ /* Copy out the data. */
+ r = &ccb.cdm.matches[1].result.periph_result;
+ snprintf(qd->devname, sizeof(qd->devname), "%s%d", r->periph_name,
+ r->unit_number);
+ free(ccb.cdm.matches);
+
+ return (0);
+}
Added: user/sbruno/mptd/mpt_cmd.c
==============================================================================
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ user/sbruno/mptd/mpt_cmd.c Fri Nov 18 17:39:20 2011 (r227669)
@@ -0,0 +1,409 @@
+/*-
+ * Copyright (c) 2011 Yahoo! Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ */
+
+#include <sys/param.h>
+#include <sys/errno.h>
+#include <sys/ioctl.h>
+#include <sys/mpt_ioctl.h>
+#include <sys/sysctl.h>
+#include <sys/uio.h>
+
+#include <err.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+#include "mptd.h"
+
+static const char *mpt_ioc_status_codes[] = {
+ "Success", /* 0x0000 */
+ "Invalid function",
+ "Busy",
+ "Invalid scatter-gather list",
+ "Internal error",
+ "Reserved",
+ "Insufficient resources",
+ "Invalid field",
+ "Invalid state", /* 0x0008 */
+ "Operation state not supported",
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL, /* 0x0010 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL, /* 0x0018 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "Invalid configuration action", /* 0x0020 */
+ "Invalid configuration type",
+ "Invalid configuration page",
+ "Invalid configuration data",
+ "No configuration defaults",
+ "Unable to commit configuration change",
+ NULL,
+ NULL,
+ NULL, /* 0x0028 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL, /* 0x0030 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL, /* 0x0038 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "Recovered SCSI error", /* 0x0040 */
+ "Invalid SCSI bus",
+ "Invalid SCSI target ID",
+ "SCSI device not there",
+ "SCSI data overrun",
+ "SCSI data underrun",
+ "SCSI I/O error",
+ "SCSI protocol error",
+ "SCSI task terminated", /* 0x0048 */
+ "SCSI residual mismatch",
+ "SCSI task management failed",
+ "SCSI I/O controller terminated",
+ "SCSI external controller terminated",
+ "EEDP guard error",
+ "EEDP reference tag error",
+ "EEDP application tag error",
+ NULL, /* 0x0050 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL, /* 0x0058 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "SCSI target priority I/O", /* 0x0060 */
+ "Invalid SCSI target port",
+ "Invalid SCSI target I/O index",
+ "SCSI target aborted",
+ "No connection retryable",
+ "No connection",
+ "FC aborted",
+ "Invalid FC receive ID",
+ "FC did invalid", /* 0x0068 */
+ "FC node logged out",
+ "Transfer count mismatch",
+ "STS data not set",
+ "FC exchange canceled",
+ "Data offset error",
+ "Too much write data",
+ "IU too short",
+ "ACK NAK timeout", /* 0x0070 */
+ "NAK received",
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL, /* 0x0078 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "LAN device not found", /* 0x0080 */
+ "LAN device failure",
+ "LAN transmit error",
+ "LAN transmit aborted",
+ "LAN receive error",
+ "LAN receive aborted",
+ "LAN partial packet",
+ "LAN canceled",
+ NULL, /* 0x0088 */
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "SAS SMP request failed", /* 0x0090 */
+ "SAS SMP data overrun",
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "Inband aborted", /* 0x0098 */
+ "No inband connection",
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ NULL,
+ "Diagnostic released", /* 0x00A0 */
+};
+
+#if 0
+static const char *mpt_raid_action_status_codes[] = {
+ "Success",
+ "Invalid action",
+ "Failure",
+ "Operation in progress",
+};
+#endif
+
+const char *
+mpt_ioc_status(U16 IOCStatus)
+{
+ static char buffer[16];
+
+ IOCStatus &= MPI_IOCSTATUS_MASK;
+ if (IOCStatus < sizeof(mpt_ioc_status_codes) / sizeof(char *) &&
+ mpt_ioc_status_codes[IOCStatus] != NULL)
+ return (mpt_ioc_status_codes[IOCStatus]);
+ snprintf(buffer, sizeof(buffer), "Status: 0x%04x", IOCStatus);
+ return (buffer);
+}
+
+#if 0
+const char *
+mpt_raid_status(U16 ActionStatus)
+{
+ static char buffer[16];
+
+ if (ActionStatus < sizeof(mpt_raid_action_status_codes) /
+ sizeof(char *))
+ return (mpt_raid_action_status_codes[ActionStatus]);
+ snprintf(buffer, sizeof(buffer), "Status: 0x%04x", ActionStatus);
+ return (buffer);
+}
+
+const char *
+mpt_raid_level(U8 VolumeType)
+{
+ static char buf[16];
+
+ switch (VolumeType) {
+ case MPI_RAID_VOL_TYPE_IS:
+ return ("RAID-0");
+ case MPI_RAID_VOL_TYPE_IM:
+ return ("RAID-1");
+ case MPI_RAID_VOL_TYPE_IME:
+ return ("RAID-1E");
+ case MPI_RAID_VOL_TYPE_RAID_5:
+ return ("RAID-5");
+ case MPI_RAID_VOL_TYPE_RAID_6:
+ return ("RAID-6");
+ case MPI_RAID_VOL_TYPE_RAID_10:
+ return ("RAID-10");
+ case MPI_RAID_VOL_TYPE_RAID_50:
+ return ("RAID-50");
+ default:
+ sprintf(buf, "LVL 0x%02x", VolumeType);
+ return (buf);
+ }
+}
+#endif
+
+const char *
+mpt_volume_name(int fd, int unit, U8 VolumeBus, U8 VolumeID)
+{
+ static struct mpt_query_disk info;
+ static char buf[16];
+
+ if (mpt_query_disk(fd, unit, VolumeBus, VolumeID, &info) != 0) {
+ /*
+ * We only print out the bus number if it is non-zero
+ * since mpt(4) only supports devices on bus zero
+ * anyway.
+ */
+ if (VolumeBus == 0)
+ snprintf(buf, sizeof(buf), "%d", VolumeID);
+ else
+ snprintf(buf, sizeof(buf), "%d:%d", VolumeBus,
+ VolumeID);
+ return (buf);
+ }
+ return (info.devname);
+}
+
+void *
+mpt_read_config_page(int fd, U8 PageType, U8 PageNumber, U32 PageAddress)
+{
+ struct mpt_cfg_page_req req;
+ void *buf;
+ int save_errno;
+
+ bzero(&req, sizeof(req));
+ req.header.PageType = PageType;
+ req.header.PageNumber = PageNumber;
+ req.page_address = PageAddress;
+ if (ioctl(fd, MPTIO_READ_CFG_HEADER, &req) < 0)
+ return (NULL);
+ if (!IOC_STATUS_SUCCESS(req.ioc_status)) {
+ warnx("Reading config page header failed: %s",
+ mpt_ioc_status(req.ioc_status));
+ errno = EIO;
+ return (NULL);
+ }
+ req.len = req.header.PageLength * 4;
+ buf = malloc(req.len);
+ req.buf = buf;
+ bcopy(&req.header, buf, sizeof(req.header));
+ if (ioctl(fd, MPTIO_READ_CFG_PAGE, &req) < 0) {
+ save_errno = errno;
+ free(buf);
+ errno = save_errno;
+ return (NULL);
+ }
+ if (!IOC_STATUS_SUCCESS(req.ioc_status)) {
+ warnx("Reading config page failed: %s",
+ mpt_ioc_status(req.ioc_status));
+ free(buf);
+ errno = EIO;
+ return (NULL);
+ }
+ return (buf);
+}
+
+int
+mpt_write_config_page(int fd, void *buf)
+{
+ CONFIG_PAGE_HEADER *hdr;
+ struct mpt_cfg_page_req req;
+
+ bzero(&req, sizeof(req));
+ req.buf = buf;
+ hdr = buf;
+ req.len = hdr->PageLength * 4;
+ if (ioctl(fd, MPTIO_WRITE_CFG_PAGE, &req) < 0)
+ return (-1);
+ if (!IOC_STATUS_SUCCESS(req.ioc_status)) {
+ warnx("Writing config page failed: %s",
+ mpt_ioc_status(req.ioc_status));
+ errno = EIO;
+ return (-1);
+ }
+ return (0);
+}
+
+#if 0
+int
+mpt_raid_action(int fd, U8 Action, U8 VolumeBus, U8 VolumeID, U8 PhysDiskNum,
+ U32 ActionDataWord, void *buf, int len, RAID_VOL0_STATUS *VolumeStatus,
+ U32 *ActionData, int datalen, U16 *IOCStatus, U16 *ActionStatus, int write)
+{
+ struct mpt_raid_action raid_act;
+
+ if (IOCStatus != NULL)
+ *IOCStatus = MPI_IOCSTATUS_SUCCESS;
+ if (datalen > sizeof(raid_act.action_data)) {
+ errno = EINVAL;
+ return (-1);
+ }
+ bzero(&raid_act, sizeof(raid_act));
+ raid_act.action = Action;
+ raid_act.volume_bus = VolumeBus;
+ raid_act.volume_id = VolumeID;
+ raid_act.phys_disk_num = PhysDiskNum;
+ raid_act.action_data_word = ActionDataWord;
+ if (buf != NULL && len != 0) {
+ raid_act.buf = buf;
+ raid_act.len = len;
+ raid_act.write = write;
+ }
+
+ if (ioctl(fd, MPTIO_RAID_ACTION, &raid_act) < 0)
+ return (-1);
+
+ if (!IOC_STATUS_SUCCESS(raid_act.ioc_status)) {
+ if (IOCStatus != NULL) {
+ *IOCStatus = raid_act.ioc_status;
+ return (0);
+ }
+ warnx("RAID action failed: %s",
+ mpt_ioc_status(raid_act.ioc_status));
+ errno = EIO;
+ return (-1);
+ }
+
+ if (ActionStatus != NULL)
+ *ActionStatus = raid_act.action_status;
+ if (raid_act.action_status != MPI_RAID_ACTION_ASTATUS_SUCCESS) {
+ if (ActionStatus != NULL)
+ return (0);
+ warnx("RAID action failed: %s",
+ mpt_raid_status(raid_act.action_status));
+ errno = EIO;
+ return (-1);
+ }
+
+ if (VolumeStatus != NULL)
+ *((U32 *)VolumeStatus) = raid_act.volume_status;
+ if (ActionData != NULL)
+ bcopy(raid_act.action_data, ActionData, datalen);
+ return (0);
+}
+#endif
Added: user/sbruno/mptd/mptd.c
==============================================================================
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ user/sbruno/mptd/mptd.c Fri Nov 18 17:39:20 2011 (r227669)
@@ -0,0 +1,871 @@
+/*-
+ * Copyright (c) 2011 Yahoo! Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ */
+
+#include <sys/types.h>
+#include <sys/errno.h>
+#include <sys/ioctl.h>
+#include <sys/param.h>
+#include <sys/time.h>
+#include <sys/uio.h>
+
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syslog.h>
+#include <unistd.h>
+#include <err.h>
+#include <fcntl.h>
+
+#include "mptd.h"
+
+#define MAX_UNIT 16
+
+#define DRIVE_FAILED(Status) \
+ ((Status).State != MPI_PHYSDISK0_STATUS_ONLINE && \
+ (Status).State != MPI_PHYSDISK0_STATUS_MISSING && \
+ (Status).State != MPI_PHYSDISK0_STATUS_NOT_COMPATIBLE && \
+ (Status).State != MPI_PHYSDISK0_STATUS_INITIALIZING)
+
+#define DRIVE_MISSING(Status) \
+ ((Status).State == MPI_PHYSDISK0_STATUS_MISSING)
+
+#define DRIVE_REBUILDING(Status) \
+ ((Status).State == MPI_PHYSDISK0_STATUS_ONLINE && \
+ ((Status).Flags & MPI_PHYSDISK0_STATUS_FLAG_OUT_OF_SYNC))
+
+#define VOLUME_DEGRADED(Status) \
+ ((Status.State) != MPI_RAIDVOL0_STATUS_STATE_OPTIMAL)
+
+static char hostname[MAXHOSTNAMELEN];
+static char *mailto = "root at localhost";
+static int notifyminutes = 720; /* send mail every 12 hours by default */
+static int dostdout;
+
+/* Maximum target_id and device_id of volumes and drives, respectively. */
+#define MPT_MAX_VOL_ID 65536
+#define MPT_MAX_PD_ID 256
+
+struct mpt_physdisk {
+ uint32_t generation;
+ U8 PhysDiskBus;
+ U8 PhysDiskID;
+ U8 PhysDiskNum;
+ RAID_PHYS_DISK0_STATUS Status;
+ uint8_t spare;
+};
+
+struct mpt_volume {
+ CONFIG_PAGE_RAID_VOL_0 *config;
+ RAID_VOL0_STATUS prev_status;
+ int sentcnt;
+ uint32_t generation;
+ U8 VolumeBus;
+ U8 VolumeID;
+ int missing_drives;
+ int prev_missing_drives;
+};
+
+struct mpt_controller {
+ int fd;
+ int unit;
+ uint32_t generation;
+ int missing_drives;
+ int prev_missing_drives;
+ int bad_drives;
+ int prev_bad_drives;
+ int sentcnt;
+ struct mpt_volume *volumes[MPT_MAX_VOL_ID];
+ struct mpt_physdisk *physdisks[MPT_MAX_PD_ID];
+};
+
+static struct mpt_controller controllers[MAX_UNIT];
+static int ncontrollers;
+
+static int
+mpt_drive_location(char *p, struct mpt_physdisk *pd)
+{
+
+ return (sprintf(p, "bus %d id %d", pd->PhysDiskBus, pd->PhysDiskID));
+}
+
+static void
+mpt_scan_drive(struct mpt_controller *c, U8 PhysDiskNum, struct mpt_volume *v)
+{
+ CONFIG_PAGE_RAID_PHYS_DISK_0 *info;
+ struct mpt_physdisk *pd;
+
+ info = mpt_pd_info(c->fd, PhysDiskNum);
+ if (info == NULL) {
+ warn("mpt%d:disk%d: failed to fetch disk info", c->unit,
+ PhysDiskNum);
+ return;
+ }
+
+ /* See if we have seen this drive before. */
+ pd = c->physdisks[PhysDiskNum];
+ if (pd == NULL) {
+ pd = calloc(1, sizeof(struct mpt_physdisk));
+ pd->PhysDiskNum = PhysDiskNum;
+ c->physdisks[PhysDiskNum] = pd;
+ pd->PhysDiskBus = info->PhysDiskBus;
+ pd->PhysDiskID = info->PhysDiskID;
+ }
+
+ /* Update generation count and other state. */
+ pd->generation = c->generation;
+ pd->Status = info->PhysDiskStatus;
+ pd->spare = (info->PhysDiskSettings.HotSparePool != 0);
+ if (DRIVE_MISSING(info->PhysDiskStatus)) {
+ if (v != NULL)
+ v->missing_drives++;
+ else
+ c->missing_drives++;
+ }
+ free(info);
+}
+
+static void
+mpt_scan_volume(struct mpt_controller *c, CONFIG_PAGE_IOC_2_RAID_VOL *vol)
+{
+ CONFIG_PAGE_RAID_VOL_0 *info;
+ RAID_VOL0_PHYS_DISK *disk;
+ struct mpt_volume *v;
+ int i;
+
+ info = mpt_vol_info(c->fd, vol->VolumeBus, vol->VolumeID);
+ if (info == NULL) {
+ warn("mpt%d:%d:%d: failed to fetch volume info", c->unit,
+ vol->VolumeBus, vol->VolumeID);
+ return;
+ }
+
+ /* See if we have seen this drive before. */
+ v = c->volumes[vol->VolumeBus * 256 + vol->VolumeID];
+ if (v == NULL) {
+ v = calloc(1, sizeof(struct mpt_volume));
+ v->VolumeBus = vol->VolumeBus;
+ v->VolumeID = vol->VolumeID;
+ c->volumes[v->VolumeBus * 256 + vol->VolumeID] = v;
+
+ v->prev_status = info->VolumeStatus;
+ } else {
+ v->prev_status = v->config->VolumeStatus;
+ free(v->config);
+ }
+
+ /* Update generation count and other state. */
+ v->generation = c->generation;
+ v->config = info;
+
+ /* Scan all the drives this volume spans. */
+ v->prev_missing_drives = v->missing_drives;
+ v->missing_drives = 0;
+ disk = info->PhysDisk;
+ for (i = 0; i < info->NumPhysDisks; disk++, i++)
+ mpt_scan_drive(c, disk->PhysDiskNum, v);
+}
+
+static void
+mpt_scan_volumes(struct mpt_controller *c)
+{
+ CONFIG_PAGE_IOC_2 *ioc2;
+ CONFIG_PAGE_IOC_2_RAID_VOL *vol;
+ struct mpt_volume *mv;
+ int i;
+
+ /* Get the volume list from the controller. */
+ ioc2 = mpt_read_ioc_page(c->fd, 2);
+ if (ioc2 == NULL) {
+ warn("mpt%d: Failed to get volume list", c->unit);
+ return;
+ }
+
+ /* Scan all the volumes. */
+ vol = ioc2->RaidVolume;
+ for (i = 0; i < ioc2->NumActiveVolumes; vol++, i++) {
+ mpt_scan_volume(c, vol);
+ }
+
+ /* Throw away all the volumes that disappeared. */
+ for (i = 0; i < MPT_MAX_VOL_ID; i++) {
+ mv = c->volumes[i];
+ if (mv == NULL)
+ continue;
+ if (mv->generation != c->generation) {
+ c->volumes[i] = NULL;
+ free(mv);
+ }
+ }
+ free(ioc2);
+}
+
+static void
+mpt_scan_drives(struct mpt_controller *c)
+{
+ CONFIG_PAGE_IOC_5 *ioc5;
+ IOC_5_HOT_SPARE *spare;
+ struct mpt_physdisk *pd;
+ int i;
+
+ /*
+ * Drives from active volumes are scanned when the volumes are
+ * scanned. The only thing left for us to look at are the
+ * spare drives.
+ */
+ ioc5 = mpt_read_ioc_page(c->fd, 5);
+ if (ioc5 == NULL) {
+ warn("mpt%d: Failed to get spare drive list", c->unit);
+ return;
+ }
+
+ /* Scan all the spares. */
+ c->prev_missing_drives = c->missing_drives;
+ c->missing_drives = 0;
+ spare = ioc5->HotSpare;
+ for (i = 0; i < ioc5->NumHotSpares; spare++, i++)
+ mpt_scan_drive(c, spare->PhysDiskNum, NULL);
+ free(ioc5);
+
+ /*
+ * If a drive fails when there is a hot spare, the failing
+ * drive swaps places with the hot spare. In this case, the
+ * failed drive won't be associated with a volume, so we track
+ * them via a controller-wide bad drives count.
+ */
+ c->prev_bad_drives = c->bad_drives;
+ c->bad_drives = 0;
+ for (i = 0; i < MPT_MAX_PD_ID; i++) {
+ if (c->physdisks[i] == NULL)
+ continue;
+ if (!DRIVE_FAILED(c->physdisks[i]->Status))
+ continue;
+ if (!c->physdisks[i]->spare)
+ continue;
+ c->bad_drives++;
+ }
+
+ /* Throw away all the drives that disappeared. */
+ for (i = 0; i < MPT_MAX_PD_ID; i++) {
+ pd = c->physdisks[i];
+ if (pd == NULL)
+ continue;
+ if (pd->generation != c->generation) {
+ c->physdisks[i] = NULL;
+ free(pd);
+ }
+ }
+}
+
+static void
+mpt_scan_controller(struct mpt_controller *c)
+{
+
+ /* Bump the overall generation count. */
+ c->generation++;
+
+ mpt_scan_volumes(c);
+ mpt_scan_drives(c);
+}
+
+static void
+mpt_scan_all(void)
+{
+ int i;
+
+ for (i = 0; i < ncontrollers; i++)
+ mpt_scan_controller(&controllers[i]);
+}
+
+static int
+mpt_open(void)
+{
+ CONFIG_PAGE_IOC_2 *ioc2;
+ char path[MAXPATHLEN];
+ int fd, unit;
+
+ ncontrollers = 0;
+ for (unit = 0; unit < MAX_UNIT; unit++) {
+ snprintf(path, sizeof(path), "/dev/mpt%d", unit);
+ fd = open(path, O_RDWR);
+ if (fd < 0)
+ continue;
+
+ /*
+ * Don't bother monitoring controllers that don't
+ * support RAID volumes. The emulated mpt(4)
+ * controllers in VMWare crash the VM when queried for
+ * a list of hot spare drives via IOC page 5, so this
+ * test lets us avoid them altogether.
+ */
+ ioc2 = mpt_read_ioc_page(fd, 2);
+ if (ioc2 == NULL || ioc2->MaxPhysDisks == 0) {
+ if (ioc2)
+ free(ioc2);
+ close(fd);
+ continue;
+ }
+ free(ioc2);
+ controllers[ncontrollers].fd = fd;
+ controllers[ncontrollers].unit = unit;
+ ncontrollers++;
+ }
+ if (ncontrollers == 0)
+ return (ncontrollers);
+
+ mpt_scan_all();
+
+ return (ncontrollers);
+}
+
+static FILE *
+mailer_open(void)
+{
+ FILE *fp;
+
+ if (dostdout)
+ fp = stdout;
+ else
+ fp = popen("/usr/sbin/sendmail -t", "w");
+ fprintf(fp, "To: %s\n", mailto);
+ return fp;
+}
+
+static void
+mailer_close(FILE *fp)
+{
+
+ if (dostdout == 0)
+ pclose(fp);
+ else
+ fflush(fp);
+}
+
+static void
+mailer_write(FILE *fp, const char *fmt, ...)
+{
+ va_list ap;
+ char *mfmt, *pfmt = NULL;
+
+ pfmt = mfmt = strdup(fmt);
+
+ va_start (ap, fmt);
+ vfprintf (fp, fmt, ap);
+ va_end (ap);
+
+ /* XXX: Hack for Subject: */
+ if (strncmp(fmt, "Subject: ", 9) == 0) {
+ char *p;
+ pfmt += strlen("Subject: ");
+ if ((p = strchr(pfmt, '\n')) != NULL)
+ *p = '\0';
+ }
+
+ if (dostdout == 0) {
+ va_start (ap, fmt);
+ vsyslog(LOG_CRIT, pfmt, ap);
+ va_end (ap);
+ }
+
+ if (mfmt)
+ free(mfmt);
+}
+
+/* Look for any failed disks in this volume. */
+char *
+mpt_show_failed(struct mpt_controller *c, struct mpt_volume *v)
+{
+ RAID_VOL0_PHYS_DISK *disk;
+ struct mpt_physdisk *pd;
+ int i, comma = 0, instate;
+ char *str, *p;
+
+ instate = 0;
+ disk = v->config->PhysDisk;
+ for (i = 0; i < v->config->NumPhysDisks; disk++, i++) {
+ pd = c->physdisks[disk->PhysDiskNum];
+ if (pd == NULL)
+ continue;
+ if (DRIVE_FAILED(pd->Status))
+ instate++;
+ }
+
+ if (instate == 0)
+ return (NULL);
+
+ str = calloc(instate * 64, sizeof(char));
+ if (str == NULL)
+ return (NULL);
+
+ p = str;
+ *p++ = '(';
+
+ disk = v->config->PhysDisk;
+ for (i = 0; i < v->config->NumPhysDisks; disk++, i++) {
*** DIFF OUTPUT TRUNCATED AT 1000 LINES ***
More information about the svn-src-user
mailing list