/* $OpenBSD: mpath.c,v 1.34 2013/09/08 09:47:36 dlg Exp $ */ /* * Copyright (c) 2009 David Gwynne * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define MPATH_BUSWIDTH 256 int mpath_match(struct device *, void *, void *); void mpath_attach(struct device *, struct device *, void *); void mpath_shutdown(void *); TAILQ_HEAD(mpath_paths, mpath_path); struct mpath_group { TAILQ_ENTRY(mpath_group) g_entry; struct mpath_paths g_paths; struct mpath_dev *g_dev; u_int g_id; }; TAILQ_HEAD(mpath_groups, mpath_group); struct mpath_dev { struct mutex d_mtx; struct scsi_xfer_list d_xfers; struct mpath_path *d_next_path; struct mpath_groups d_groups; struct mpath_group *d_failover_iter; struct timeout d_failover_tmo; u_int d_failover; const struct mpath_ops *d_ops; struct devid *d_id; }; struct mpath_softc { struct device sc_dev; struct scsi_link sc_link; struct scsibus_softc *sc_scsibus; struct mpath_dev *sc_devs[MPATH_BUSWIDTH]; }; #define DEVNAME(_s) ((_s)->sc_dev.dv_xname) struct mpath_softc *mpath; struct cfattach mpath_ca = { sizeof(struct mpath_softc), mpath_match, mpath_attach }; struct cfdriver mpath_cd = { NULL, "mpath", DV_DULL }; void mpath_cmd(struct scsi_xfer *); void mpath_minphys(struct buf *, struct scsi_link *); int mpath_probe(struct scsi_link *); struct mpath_path *mpath_next_path(struct mpath_dev *); void mpath_done(struct scsi_xfer *); void mpath_failover(struct mpath_dev *); void mpath_failover_start(void *); void mpath_failover_check(struct mpath_dev *); struct scsi_adapter mpath_switch = { mpath_cmd, scsi_minphys, mpath_probe }; void mpath_xs_stuffup(struct scsi_xfer *); int mpath_match(struct device *parent, void *match, void *aux) { return (1); } void mpath_attach(struct device *parent, struct device *self, void *aux) { struct mpath_softc *sc = (struct mpath_softc *)self; struct scsibus_attach_args saa; mpath = sc; printf("\n"); sc->sc_link.adapter = &mpath_switch; sc->sc_link.adapter_softc = sc; sc->sc_link.adapter_target = MPATH_BUSWIDTH; sc->sc_link.adapter_buswidth = MPATH_BUSWIDTH; sc->sc_link.luns = 1; sc->sc_link.openings = 1024; /* XXX magical */ bzero(&saa, sizeof(saa)); saa.saa_sc_link = &sc->sc_link; sc->sc_scsibus = (struct scsibus_softc *)config_found(&sc->sc_dev, &saa, scsiprint); } void mpath_xs_stuffup(struct scsi_xfer *xs) { xs->error = XS_DRIVER_STUFFUP; scsi_done(xs); } int mpath_probe(struct scsi_link *link) { struct mpath_softc *sc = link->adapter_softc; struct mpath_dev *d = sc->sc_devs[link->target]; if (link->lun != 0 || d == NULL) return (ENXIO); link->id = devid_copy(d->d_id); return (0); } struct mpath_path * mpath_next_path(struct mpath_dev *d) { struct mpath_group *g; struct mpath_path *p; #ifdef DIAGNOSTIC if (d == NULL) panic("%s: d is NULL", __func__); #endif p = d->d_next_path; if (p != NULL) { d->d_next_path = TAILQ_NEXT(p, p_entry); if (d->d_next_path == NULL && (g = TAILQ_FIRST(&d->d_groups)) != NULL) d->d_next_path = TAILQ_FIRST(&g->g_paths); } return (p); } void mpath_cmd(struct scsi_xfer *xs) { struct scsi_link *link = xs->sc_link; struct mpath_softc *sc = link->adapter_softc; struct mpath_dev *d = sc->sc_devs[link->target]; struct mpath_path *p; struct scsi_xfer *mxs; #ifdef DIAGNOSTIC if (d == NULL) panic("mpath_cmd issued against nonexistant device"); #endif if (ISSET(xs->flags, SCSI_POLL)) { mtx_enter(&d->d_mtx); p = mpath_next_path(d); mtx_leave(&d->d_mtx); if (p == NULL) { mpath_xs_stuffup(xs); return; } mxs = scsi_xs_get(p->p_link, xs->flags); if (mxs == NULL) { mpath_xs_stuffup(xs); return; } memcpy(mxs->cmd, xs->cmd, xs->cmdlen); mxs->cmdlen = xs->cmdlen; mxs->data = xs->data; mxs->datalen = xs->datalen; mxs->retries = xs->retries; mxs->timeout = xs->timeout; mxs->bp = xs->bp; scsi_xs_sync(mxs); xs->error = mxs->error; xs->status = mxs->status; xs->resid = mxs->resid; memcpy(&xs->sense, &mxs->sense, sizeof(xs->sense)); scsi_xs_put(mxs); scsi_done(xs); return; } mtx_enter(&d->d_mtx); SIMPLEQ_INSERT_TAIL(&d->d_xfers, xs, xfer_list); p = mpath_next_path(d); mtx_leave(&d->d_mtx); if (p != NULL) scsi_xsh_add(&p->p_xsh); } void mpath_start(struct mpath_path *p, struct scsi_xfer *mxs) { struct mpath_dev *d = p->p_group->g_dev; struct scsi_xfer *xs; int addxsh = 0; if (ISSET(p->p_link->state, SDEV_S_DYING) || d == NULL) goto fail; mtx_enter(&d->d_mtx); xs = SIMPLEQ_FIRST(&d->d_xfers); if (xs != NULL) { SIMPLEQ_REMOVE_HEAD(&d->d_xfers, xfer_list); if (!SIMPLEQ_EMPTY(&d->d_xfers)) addxsh = 1; } mtx_leave(&d->d_mtx); if (xs == NULL) goto fail; memcpy(mxs->cmd, xs->cmd, xs->cmdlen); mxs->cmdlen = xs->cmdlen; mxs->data = xs->data; mxs->datalen = xs->datalen; mxs->retries = xs->retries; mxs->timeout = xs->timeout; mxs->bp = xs->bp; mxs->flags = xs->flags; mxs->cookie = xs; mxs->done = mpath_done; scsi_xs_exec(mxs); if (addxsh) scsi_xsh_add(&p->p_xsh); return; fail: scsi_xs_put(mxs); } void mpath_done(struct scsi_xfer *mxs) { struct scsi_xfer *xs = mxs->cookie; struct scsi_link *link = xs->sc_link; struct mpath_softc *sc = link->adapter_softc; struct mpath_dev *d = sc->sc_devs[link->target]; struct mpath_path *p; switch (mxs->error) { case XS_SELTIMEOUT: /* physical path is gone, try the next */ case XS_RESET: mtx_enter(&d->d_mtx); SIMPLEQ_INSERT_HEAD(&d->d_xfers, xs, xfer_list); p = mpath_next_path(d); mtx_leave(&d->d_mtx); scsi_xs_put(mxs); if (p != NULL) scsi_xsh_add(&p->p_xsh); return; case XS_SENSE: switch (d->d_ops->op_checksense(mxs)) { case MPATH_SENSE_FAILOVER: mtx_enter(&d->d_mtx); SIMPLEQ_INSERT_HEAD(&d->d_xfers, xs, xfer_list); p = mpath_next_path(d); mtx_leave(&d->d_mtx); scsi_xs_put(mxs); mpath_failover(d); return; case MPATH_SENSE_DECLINED: break; #ifdef DIAGNOSTIC default: panic("unexpected return from checksense"); #endif } break; } xs->error = mxs->error; xs->status = mxs->status; xs->resid = mxs->resid; memcpy(&xs->sense, &mxs->sense, sizeof(xs->sense)); scsi_xs_put(mxs); scsi_done(xs); } void mpath_failover(struct mpath_dev *d) { if (!scsi_pending_start(&d->d_mtx, &d->d_failover)) return; mpath_failover_start(d); } void mpath_failover_start(void *xd) { struct mpath_dev *d = xd; mtx_enter(&d->d_mtx); d->d_failover_iter = TAILQ_FIRST(&d->d_groups); mtx_leave(&d->d_mtx); mpath_failover_check(d); } void mpath_failover_check(struct mpath_dev *d) { struct mpath_group *g = d->d_failover_iter; struct mpath_path *p; if (g == NULL) timeout_add_sec(&d->d_failover_tmo, 1); else { p = TAILQ_FIRST(&g->g_paths); d->d_ops->op_status(p->p_link); } } void mpath_path_status(struct mpath_path *p, int status) { struct mpath_group *g = p->p_group; struct mpath_dev *d = g->g_dev; mtx_enter(&d->d_mtx); if (status == MPATH_S_ACTIVE) { TAILQ_REMOVE(&d->d_groups, g, g_entry); TAILQ_INSERT_HEAD(&d->d_groups, g, g_entry); d->d_next_path = p; } else d->d_failover_iter = TAILQ_NEXT(d->d_failover_iter, g_entry); mtx_leave(&d->d_mtx); if (status == MPATH_S_ACTIVE) { scsi_xsh_add(&p->p_xsh); if (!scsi_pending_finish(&d->d_mtx, &d->d_failover)) mpath_failover_start(d); } else mpath_failover_check(d); } void mpath_minphys(struct buf *bp, struct scsi_link *link) { struct mpath_softc *sc = link->adapter_softc; struct mpath_dev *d = sc->sc_devs[link->target]; struct mpath_group *g; struct mpath_path *p; #ifdef DIAGNOSTIC if (d == NULL) panic("mpath_minphys against nonexistant device"); #endif mtx_enter(&d->d_mtx); TAILQ_FOREACH(g, &d->d_groups, g_entry) { TAILQ_FOREACH(p, &g->g_paths, p_entry) { /* XXX crossing layers with mutex held */ p->p_link->adapter->scsi_minphys(bp, p->p_link); } } mtx_leave(&d->d_mtx); } int mpath_path_probe(struct scsi_link *link) { if (mpath == NULL) return (ENXIO); if (link->id == NULL) return (EINVAL); if (mpath == link->adapter_softc) return (ENXIO); return (0); } int mpath_path_attach(struct mpath_path *p, u_int g_id, const struct mpath_ops *ops) { struct mpath_softc *sc = mpath; struct scsi_link *link = p->p_link; struct mpath_dev *d = NULL; struct mpath_group *g; int newdev = 0, addxsh = 0; int target; #ifdef DIAGNOSTIC if (p->p_link == NULL) panic("mpath_path_attach: NULL link"); if (p->p_group != NULL) panic("mpath_path_attach: group is not NULL"); #endif for (target = 0; target < MPATH_BUSWIDTH; target++) { if ((d = sc->sc_devs[target]) == NULL) continue; if (DEVID_CMP(d->d_id, link->id) && d->d_ops == ops) break; d = NULL; } if (d == NULL) { for (target = 0; target < MPATH_BUSWIDTH; target++) { if (sc->sc_devs[target] == NULL) break; } if (target >= MPATH_BUSWIDTH) return (ENXIO); d = malloc(sizeof(*d), M_DEVBUF, M_WAITOK | M_CANFAIL | M_ZERO); if (d == NULL) return (ENOMEM); mtx_init(&d->d_mtx, IPL_BIO); TAILQ_INIT(&d->d_groups); SIMPLEQ_INIT(&d->d_xfers); d->d_id = devid_copy(link->id); d->d_ops = ops; timeout_set(&d->d_failover_tmo, mpath_failover_start, d); sc->sc_devs[target] = d; newdev = 1; } else { /* * instead of carrying identical values in different devid * instances, delete the new one and reference the old one in * the new scsi_link. */ devid_free(link->id); link->id = devid_copy(d->d_id); } TAILQ_FOREACH(g, &d->d_groups, g_entry) { if (g->g_id == g_id) break; } if (g == NULL) { g = malloc(sizeof(*g), M_DEVBUF, M_WAITOK | M_CANFAIL | M_ZERO); if (g == NULL) { if (newdev) { free(d, M_DEVBUF); sc->sc_devs[target] = NULL; } return (ENOMEM); } TAILQ_INIT(&g->g_paths); g->g_dev = d; g->g_id = g_id; mtx_enter(&d->d_mtx); TAILQ_INSERT_TAIL(&d->d_groups, g, g_entry); mtx_leave(&d->d_mtx); } p->p_group = g; mtx_enter(&d->d_mtx); TAILQ_INSERT_TAIL(&g->g_paths, p, p_entry); if (!SIMPLEQ_EMPTY(&d->d_xfers)) addxsh = 1; if (d->d_next_path == NULL) d->d_next_path = p; mtx_leave(&d->d_mtx); if (newdev) scsi_probe_target(mpath->sc_scsibus, target); else if (addxsh) scsi_xsh_add(&p->p_xsh); return (0); } int mpath_path_detach(struct mpath_path *p) { struct mpath_group *g = p->p_group; struct mpath_dev *d = g->g_dev; struct mpath_path *np = NULL; #ifdef DIAGNOSTIC if (g == NULL) panic("mpath: detaching a path from a nonexistant bus"); #endif p->p_group = NULL; mtx_enter(&d->d_mtx); TAILQ_REMOVE(&g->g_paths, p, p_entry); if (d->d_next_path == p) d->d_next_path = TAILQ_FIRST(&g->g_paths); if (TAILQ_EMPTY(&g->g_paths)) TAILQ_REMOVE(&d->d_groups, g, g_entry); else g = NULL; if (!SIMPLEQ_EMPTY(&d->d_xfers)) np = d->d_next_path; mtx_leave(&d->d_mtx); if (g != NULL) free(g, M_DEVBUF); scsi_xsh_del(&p->p_xsh); if (np == NULL) mpath_failover(d); else scsi_xsh_add(&np->p_xsh); return (0); } struct device * mpath_bootdv(struct device *dev) { struct mpath_softc *sc = mpath; struct mpath_dev *d; struct mpath_group *g; struct mpath_path *p; int target; if (sc == NULL) return (dev); for (target = 0; target < MPATH_BUSWIDTH; target++) { if ((d = sc->sc_devs[target]) == NULL) continue; TAILQ_FOREACH(g, &d->d_groups, g_entry) { TAILQ_FOREACH(p, &g->g_paths, p_entry) { if (p->p_link->device_softc == dev) { return (scsi_get_link(mpath->sc_scsibus, target, 0)->device_softc); } } } } return (dev); }