/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- * vim:expandtab:shiftwidth=8:tabstop=8: * * Copyright (c) 2001-2003 Cluster File Systems, Inc. * * This file is part of Lustre, http://www.lustre.org. * * Lustre is free software; you can redistribute it and/or * modify it under the terms of version 2 of the GNU General Public * License as published by the Free Software Foundation. * * Lustre is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with Lustre; if not, write to the Free Software * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ #define DEBUG_SUBSYSTEM S_ECHO #ifdef __KERNEL__ #include #include #include #include #if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)) #include #endif #include #include #else #include #endif #include #include #include #include #include #include static obd_id last_object_id; #if 0 static void echo_printk_object (char *msg, struct ec_object *eco) { struct lov_stripe_md *lsm = eco->eco_lsm; int i; printk (KERN_INFO "Lustre: %s: object %p: "LPX64", refs %d%s: "LPX64 "=%u!%u\n", msg, eco, eco->eco_id, eco->eco_refcount, eco->eco_deleted ? "(deleted) " : "", lsm->lsm_object_id, lsm->lsm_stripe_size, lsm->lsm_stripe_count); for (i = 0; i < lsm->lsm_stripe_count; i++) printk (KERN_INFO "Lustre: @%2u:"LPX64"\n", lsm->lsm_oinfo[i].loi_ost_idx, lsm->lsm_oinfo[i].loi_id); } #endif static struct ec_object * echo_find_object_locked (struct obd_device *obd, obd_id id) { struct echo_client_obd *ec = &obd->u.echocli; struct ec_object *eco = NULL; struct list_head *el; list_for_each (el, &ec->ec_objects) { eco = list_entry (el, struct ec_object, eco_obj_chain); if (eco->eco_id == id) return (eco); } return (NULL); } static int echo_copyout_lsm (struct lov_stripe_md *lsm, void *ulsm, int ulsm_nob) { int nob; nob = offsetof (struct lov_stripe_md, lsm_oinfo[lsm->lsm_stripe_count]); if (nob > ulsm_nob) return (-EINVAL); if (copy_to_user (ulsm, lsm, nob)) return (-EFAULT); return (0); } static int echo_copyin_lsm (struct obd_device *obd, struct lov_stripe_md *lsm, void *ulsm, int ulsm_nob) { struct echo_client_obd *ec = &obd->u.echocli; int nob; if (ulsm_nob < sizeof (*lsm)) return (-EINVAL); if (copy_from_user (lsm, ulsm, sizeof (*lsm))) return (-EFAULT); nob = lsm->lsm_stripe_count * sizeof (lsm->lsm_oinfo[0]); if (ulsm_nob < nob || lsm->lsm_stripe_count > ec->ec_nstripes || lsm->lsm_magic != LOV_MAGIC || (lsm->lsm_stripe_size & (PAGE_SIZE - 1)) != 0 || ((__u64)lsm->lsm_stripe_size * lsm->lsm_stripe_count > ~0UL)) return (-EINVAL); if (copy_from_user(lsm->lsm_oinfo, ((struct lov_stripe_md *)ulsm)->lsm_oinfo, nob)) return (-EFAULT); return (0); } static struct ec_object * echo_allocate_object (struct obd_device *obd) { struct echo_client_obd *ec = &obd->u.echocli; struct ec_object *eco; int rc; OBD_ALLOC(eco, sizeof (*eco)); if (eco == NULL) return NULL; rc = obd_alloc_memmd(ec->ec_exp, &eco->eco_lsm); if (rc < 0) { OBD_FREE(eco, sizeof (*eco)); return NULL; } eco->eco_device = obd; eco->eco_deleted = 0; eco->eco_refcount = 0; eco->eco_lsm->lsm_magic = LOV_MAGIC; /* leave stripe count 0 by default */ return (eco); } static void echo_free_object (struct ec_object *eco) { struct obd_device *obd = eco->eco_device; struct echo_client_obd *ec = &obd->u.echocli; LASSERT (eco->eco_refcount == 0); obd_free_memmd(ec->ec_exp, &eco->eco_lsm); OBD_FREE (eco, sizeof (*eco)); } static int echo_create_object(struct obd_device *obd, int on_target, struct obdo *oa, void *ulsm, int ulsm_nob, struct obd_trans_info *oti) { struct echo_client_obd *ec = &obd->u.echocli; struct ec_object *eco2; struct ec_object *eco; struct lov_stripe_md *lsm; int rc; int i, idx; if ((oa->o_valid & OBD_MD_FLID) == 0 && /* no obj id */ (on_target || /* set_stripe */ ec->ec_nstripes != 0)) { /* LOV */ CERROR ("No valid oid\n"); return (-EINVAL); } if (ulsm != NULL) { eco = echo_allocate_object (obd); if (eco == NULL) return (-ENOMEM); lsm = eco->eco_lsm; rc = echo_copyin_lsm (obd, lsm, ulsm, ulsm_nob); if (rc != 0) goto failed; /* setup object ID here for !on_target and LOV hint */ if ((oa->o_valid & OBD_MD_FLID) != 0) eco->eco_id = lsm->lsm_object_id = oa->o_id; if (lsm->lsm_stripe_count == 0) lsm->lsm_stripe_count = ec->ec_nstripes; if (lsm->lsm_stripe_size == 0) lsm->lsm_stripe_size = PAGE_SIZE; idx = ll_insecure_random_int(); /* setup stripes: indices + default ids if required */ for (i = 0; i < lsm->lsm_stripe_count; i++) { if (lsm->lsm_oinfo[i].loi_id == 0) lsm->lsm_oinfo[i].loi_id = lsm->lsm_object_id; lsm->lsm_oinfo[i].loi_ost_idx = (idx + i) % ec->ec_nstripes; } } else { OBD_ALLOC(eco, sizeof(*eco)); eco->eco_device = obd; lsm = NULL; } if (oa->o_id == 0) oa->o_id = ++last_object_id; if (on_target) { oa->o_gr = FILTER_GROUP_ECHO; oa->o_valid |= OBD_MD_FLGROUP; rc = obd_create(ec->ec_exp, oa, &lsm, oti); if (rc != 0) goto failed; /* See what object ID we were given */ eco->eco_id = oa->o_id = lsm->lsm_object_id; oa->o_valid |= OBD_MD_FLID; LASSERT(eco->eco_lsm == NULL || eco->eco_lsm == lsm); eco->eco_lsm = lsm; } spin_lock (&ec->ec_lock); eco2 = echo_find_object_locked (obd, oa->o_id); if (eco2 != NULL) { /* conflict */ spin_unlock (&ec->ec_lock); CERROR ("Can't create object id "LPX64": id already exists%s\n", oa->o_id, on_target ? " (undoing create)" : ""); if (on_target) obd_destroy(ec->ec_exp, oa, lsm, oti); rc = -EEXIST; goto failed; } list_add (&eco->eco_obj_chain, &ec->ec_objects); spin_unlock (&ec->ec_lock); CDEBUG (D_INFO, "created %p: "LPX64"=%u#%u@%u refs %d del %d\n", eco, eco->eco_id, eco->eco_lsm->lsm_stripe_size, eco->eco_lsm->lsm_stripe_count, eco->eco_lsm->lsm_oinfo[0].loi_ost_idx, eco->eco_refcount, eco->eco_deleted); return (0); failed: echo_free_object (eco); return (rc); } static int echo_get_object (struct ec_object **ecop, struct obd_device *obd, struct obdo *oa) { struct echo_client_obd *ec = &obd->u.echocli; struct ec_object *eco; struct ec_object *eco2; int rc; if ((oa->o_valid & OBD_MD_FLID) == 0 || oa->o_id == 0) /* disallow use of object id 0 */ { CERROR ("No valid oid\n"); return (-EINVAL); } spin_lock (&ec->ec_lock); eco = echo_find_object_locked (obd, oa->o_id); if (eco != NULL) { if (eco->eco_deleted) /* being deleted */ return (-EAGAIN); /* (see comment in cleanup) */ eco->eco_refcount++; spin_unlock (&ec->ec_lock); *ecop = eco; CDEBUG (D_INFO, "found %p: "LPX64"=%u#%u@%u refs %d del %d\n", eco, eco->eco_id, eco->eco_lsm->lsm_stripe_size, eco->eco_lsm->lsm_stripe_count, eco->eco_lsm->lsm_oinfo[0].loi_ost_idx, eco->eco_refcount, eco->eco_deleted); return (0); } spin_unlock (&ec->ec_lock); if (ec->ec_nstripes != 0) /* striping required */ return (-ENOENT); eco = echo_allocate_object (obd); if (eco == NULL) return (-ENOMEM); eco->eco_id = eco->eco_lsm->lsm_object_id = oa->o_id; spin_lock (&ec->ec_lock); eco2 = echo_find_object_locked (obd, oa->o_id); if (eco2 == NULL) { /* didn't race */ list_add (&eco->eco_obj_chain, &ec->ec_objects); spin_unlock (&ec->ec_lock); eco->eco_refcount = 1; *ecop = eco; CDEBUG (D_INFO, "created %p: "LPX64"=%u#%u@%d refs %d del %d\n", eco, eco->eco_id, eco->eco_lsm->lsm_stripe_size, eco->eco_lsm->lsm_stripe_count, eco->eco_lsm->lsm_oinfo[0].loi_ost_idx, eco->eco_refcount, eco->eco_deleted); return (0); } if (eco2->eco_deleted) rc = -EAGAIN; /* lose race */ else { eco2->eco_refcount++; /* take existing */ *ecop = eco2; rc = 0; LASSERT (eco2->eco_id == eco2->eco_lsm->lsm_object_id); CDEBUG (D_INFO, "found(2) %p: "LPX64"=%u#%u@%d refs %d del %d\n", eco2, eco2->eco_id, eco2->eco_lsm->lsm_stripe_size, eco2->eco_lsm->lsm_stripe_count, eco2->eco_lsm->lsm_oinfo[0].loi_ost_idx, eco2->eco_refcount, eco2->eco_deleted); } spin_unlock (&ec->ec_lock); echo_free_object (eco); return (rc); } static void echo_put_object (struct ec_object *eco) { struct obd_device *obd = eco->eco_device; struct echo_client_obd *ec = &obd->u.echocli; /* Release caller's ref on the object. * delete => mark for deletion when last ref goes */ spin_lock (&ec->ec_lock); eco->eco_refcount--; LASSERT (eco->eco_refcount >= 0); CDEBUG(D_INFO, "put %p: "LPX64"=%u#%u@%d refs %d del %d\n", eco, eco->eco_id, eco->eco_lsm->lsm_stripe_size, eco->eco_lsm->lsm_stripe_count, eco->eco_lsm->lsm_oinfo[0].loi_ost_idx, eco->eco_refcount, eco->eco_deleted); if (eco->eco_refcount != 0 || !eco->eco_deleted) { spin_unlock (&ec->ec_lock); return; } spin_unlock (&ec->ec_lock); /* NB leave obj in the object list. We must prevent anyone from * attempting to enqueue on this object number until we can be * sure there will be no more lock callbacks. */ obd_cancel_unused(ec->ec_exp, eco->eco_lsm, 0, NULL); /* now we can let it go */ spin_lock (&ec->ec_lock); list_del (&eco->eco_obj_chain); spin_unlock (&ec->ec_lock); LASSERT (eco->eco_refcount == 0); echo_free_object (eco); } static void echo_get_stripe_off_id (struct lov_stripe_md *lsm, obd_off *offp, obd_id *idp) { unsigned long stripe_count; unsigned long stripe_size; unsigned long width; unsigned long woffset; int stripe_index; obd_off offset; if (lsm->lsm_stripe_count <= 1) return; offset = *offp; stripe_size = lsm->lsm_stripe_size; stripe_count = lsm->lsm_stripe_count; /* width = # bytes in all stripes */ width = stripe_size * stripe_count; /* woffset = offset within a width; offset = whole number of widths */ woffset = do_div (offset, width); stripe_index = woffset / stripe_size; *idp = lsm->lsm_oinfo[stripe_index].loi_id; *offp = offset * stripe_size + woffset % stripe_size; } static void echo_client_page_debug_setup(struct lov_stripe_md *lsm, struct page *page, int rw, obd_id id, obd_off offset, obd_off count) { char *addr; obd_off stripe_off; obd_id stripe_id; int delta; /* no partial pages on the client */ LASSERT(count == PAGE_SIZE); addr = kmap(page); for (delta = 0; delta < PAGE_SIZE; delta += OBD_ECHO_BLOCK_SIZE) { if (rw == OBD_BRW_WRITE) { stripe_off = offset + delta; stripe_id = id; echo_get_stripe_off_id(lsm, &stripe_off, &stripe_id); } else { stripe_off = 0xdeadbeef00c0ffeeULL; stripe_id = 0xdeadbeef00c0ffeeULL; } block_debug_setup(addr + delta, OBD_ECHO_BLOCK_SIZE, stripe_off, stripe_id); } kunmap(page); } static int echo_client_page_debug_check(struct lov_stripe_md *lsm, struct page *page, obd_id id, obd_off offset, obd_off count) { obd_off stripe_off; obd_id stripe_id; char *addr; int delta; int rc; int rc2; /* no partial pages on the client */ LASSERT(count == PAGE_SIZE); addr = kmap(page); for (rc = delta = 0; delta < PAGE_SIZE; delta += OBD_ECHO_BLOCK_SIZE) { stripe_off = offset + delta; stripe_id = id; echo_get_stripe_off_id (lsm, &stripe_off, &stripe_id); rc2 = block_debug_check("test_brw", addr + delta, OBD_ECHO_BLOCK_SIZE, stripe_off, stripe_id); if (rc2 != 0) { CERROR ("Error in echo object "LPX64"\n", id); rc = rc2; } } kunmap(page); return rc; } static int echo_client_kbrw(struct obd_device *obd, int rw, struct obdo *oa, struct lov_stripe_md *lsm, obd_off offset, obd_size count, struct obd_trans_info *oti) { struct echo_client_obd *ec = &obd->u.echocli; obd_count npages; struct brw_page *pga; struct brw_page *pgp; obd_off off; int i; int rc; int verify = 0; int gfp_mask; verify = ((oa->o_id) != ECHO_PERSISTENT_OBJID && (oa->o_valid & OBD_MD_FLFLAGS) != 0 && (oa->o_flags & OBD_FL_DEBUG_CHECK) != 0); gfp_mask = ((oa->o_id & 2) == 0) ? GFP_KERNEL : GFP_HIGHUSER; LASSERT(rw == OBD_BRW_WRITE || rw == OBD_BRW_READ); if (count <= 0 || (count & (PAGE_SIZE - 1)) != 0 || (lsm != NULL && lsm->lsm_object_id != oa->o_id)) return (-EINVAL); /* XXX think again with misaligned I/O */ npages = count >> PAGE_SHIFT; OBD_ALLOC(pga, npages * sizeof(*pga)); if (pga == NULL) return (-ENOMEM); for (i = 0, pgp = pga, off = offset; i < npages; i++, pgp++, off += PAGE_SIZE) { LASSERT (pgp->pg == NULL); /* for cleanup */ rc = -ENOMEM; pgp->pg = alloc_pages (gfp_mask, 0); if (pgp->pg == NULL) goto out; pgp->count = PAGE_SIZE; pgp->disk_offset = pgp->page_offset = off; pgp->flag = 0; if (verify) echo_client_page_debug_setup(lsm, pgp->pg, rw, oa->o_id, off, pgp->count); } rc = obd_brw(rw, ec->ec_exp, oa, lsm, npages, pga, oti); out: if (rc != 0 || rw != OBD_BRW_READ) verify = 0; for (i = 0, pgp = pga; i < npages; i++, pgp++) { if (pgp->pg == NULL) continue; if (verify) { int vrc; vrc = echo_client_page_debug_check(lsm, pgp->pg, oa->o_id, pgp->page_offset, pgp->count); if (vrc != 0 && rc == 0) rc = vrc; } __free_pages(pgp->pg, 0); } OBD_FREE(pga, npages * sizeof(*pga)); return (rc); } #ifdef __KERNEL__ #if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)) static int echo_client_ubrw(struct obd_device *obd, int rw, struct obdo *oa, struct lov_stripe_md *lsm, obd_off offset, obd_size count, char *buffer, struct obd_trans_info *oti) { struct echo_client_obd *ec = &obd->u.echocli; obd_count npages; struct brw_page *pga; struct brw_page *pgp; obd_off off; struct kiobuf *kiobuf; int i; int rc; LASSERT (rw == OBD_BRW_WRITE || rw == OBD_BRW_READ); /* NB: for now, only whole pages, page aligned */ if (count <= 0 || ((long)buffer & (PAGE_SIZE - 1)) != 0 || (count & (PAGE_SIZE - 1)) != 0 || (lsm != NULL && lsm->lsm_object_id != oa->o_id)) return (-EINVAL); /* XXX think again with misaligned I/O */ npages = count >> PAGE_SHIFT; OBD_ALLOC(pga, npages * sizeof(*pga)); if (pga == NULL) return (-ENOMEM); rc = alloc_kiovec (1, &kiobuf); if (rc != 0) goto out_1; rc = map_user_kiobuf ((rw == OBD_BRW_READ) ? READ : WRITE, kiobuf, (unsigned long)buffer, count); if (rc != 0) goto out_2; LASSERT (kiobuf->offset == 0); LASSERT (kiobuf->nr_pages == npages); for (i = 0, off = offset, pgp = pga; i < npages; i++, off += PAGE_SIZE, pgp++) { pgp->disk_offset = pgp->page_offset = off; pgp->pg = kiobuf->maplist[i]; pgp->count = PAGE_SIZE; pgp->flag = 0; } rc = obd_brw(rw, ec->ec_exp, oa, lsm, npages, pga, oti); // if (rw == OBD_BRW_READ) // mark_dirty_kiobuf (kiobuf, count); unmap_kiobuf (kiobuf); out_2: free_kiovec (1, &kiobuf); out_1: OBD_FREE(pga, npages * sizeof(*pga)); return (rc); } #else static int echo_client_ubrw(struct obd_device *obd, int rw, struct obdo *oa, struct lov_stripe_md *lsm, obd_off offset, obd_size count, char *buffer, struct obd_trans_info *oti) { #warning "echo_client_ubrw() needs to be ported on 2.6 yet" LBUG(); return 0; } #endif #endif struct echo_async_state; #define EAP_MAGIC 79277927 struct echo_async_page { int eap_magic; struct page *eap_page; void *eap_cookie; obd_off eap_off; struct echo_async_state *eap_eas; struct list_head eap_item; }; #define EAP_FROM_COOKIE(c) \ (LASSERT(((struct echo_async_page *)(c))->eap_magic == EAP_MAGIC), \ (struct echo_async_page *)(c)) struct echo_async_state { spinlock_t eas_lock; obd_off eas_next_offset; obd_off eas_end_offset; int eas_in_flight; int eas_rc; wait_queue_head_t eas_waitq; struct list_head eas_avail; struct obdo eas_oa; struct lov_stripe_md *eas_lsm; }; static int eas_should_wake(struct echo_async_state *eas) { unsigned long flags; int rc = 0; spin_lock_irqsave(&eas->eas_lock, flags); if (eas->eas_rc == 0 && !list_empty(&eas->eas_avail)) rc = 1; spin_unlock_irqrestore(&eas->eas_lock, flags); return rc; }; static int ec_ap_make_ready(void *data, int cmd) { /* our pages are issued ready */ LBUG(); return 0; } static int ec_ap_refresh_count(void *data, int cmd) { /* our pages are issued with a stable count */ LBUG(); return PAGE_SIZE; } static void ec_ap_fill_obdo(void *data, int cmd, struct obdo *oa) { struct echo_async_page *eap = EAP_FROM_COOKIE(data); memcpy(oa, &eap->eap_eas->eas_oa, sizeof(*oa)); } static void ec_ap_completion(void *data, int cmd, struct obdo *oa, int rc) { struct echo_async_page *eap = EAP_FROM_COOKIE(data); struct echo_async_state *eas; unsigned long flags; eas = eap->eap_eas; if (cmd == OBD_BRW_READ && eas->eas_oa.o_id != ECHO_PERSISTENT_OBJID && (eas->eas_oa.o_valid & OBD_MD_FLFLAGS) != 0 && (eas->eas_oa.o_flags & OBD_FL_DEBUG_CHECK) != 0) echo_client_page_debug_check(eas->eas_lsm, eap->eap_page, eas->eas_oa.o_id, eap->eap_off, PAGE_SIZE); spin_lock_irqsave(&eas->eas_lock, flags); if (rc && !eas->eas_rc) eas->eas_rc = rc; eas->eas_in_flight--; list_add(&eap->eap_item, &eas->eas_avail); wake_up(&eas->eas_waitq); spin_unlock_irqrestore(&eas->eas_lock, flags); } static struct obd_async_page_ops ec_async_page_ops = { .ap_make_ready = ec_ap_make_ready, .ap_refresh_count = ec_ap_refresh_count, .ap_fill_obdo = ec_ap_fill_obdo, .ap_completion = ec_ap_completion, }; static int echo_client_async_page(struct obd_export *exp, int rw, struct obdo *oa, struct lov_stripe_md *lsm, obd_off offset, obd_size count, obd_size batching) { obd_count npages, i; struct echo_async_page *eap; struct echo_async_state eas; struct list_head *pos, *n; int rc = 0; unsigned long flags; LIST_HEAD(pages); #if 0 int verify; int gfp_mask; /* oa_id == 0 => speed test (no verification) else... * oa & 1 => use HIGHMEM */ verify = (oa->o_id != 0); gfp_mask = ((oa->o_id & 1) == 0) ? GFP_KERNEL : GFP_HIGHUSER; #endif LASSERT(rw == OBD_BRW_WRITE || rw == OBD_BRW_READ); if (count <= 0 || (count & (PAGE_SIZE - 1)) != 0 || (lsm != NULL && lsm->lsm_object_id != oa->o_id)) return (-EINVAL); /* XXX think again with misaligned I/O */ npages = batching >> PAGE_SHIFT; memcpy(&eas.eas_oa, oa, sizeof(*oa)); eas.eas_next_offset = offset; eas.eas_end_offset = offset + count; spin_lock_init(&eas.eas_lock); init_waitqueue_head(&eas.eas_waitq); eas.eas_in_flight = 0; eas.eas_rc = 0; eas.eas_lsm = lsm; INIT_LIST_HEAD(&eas.eas_avail); /* prepare the group of pages that we're going to be keeping * in flight */ for (i = 0; i < npages; i++) { struct page *page = alloc_page(GFP_KERNEL); if (page == NULL) GOTO(out, rc = -ENOMEM); page->private = 0; list_add_tail(&PAGE_LIST(page), &pages); OBD_ALLOC(eap, sizeof(*eap)); if (eap == NULL) GOTO(out, rc = -ENOMEM); eap->eap_magic = EAP_MAGIC; eap->eap_page = page; eap->eap_eas = &eas; page->private = (unsigned long)eap; list_add_tail(&eap->eap_item, &eas.eas_avail); } /* first we spin queueing io and being woken by its completion */ spin_lock_irqsave(&eas.eas_lock, flags); for(;;) { int rc; /* sleep until we have a page to send */ spin_unlock_irqrestore(&eas.eas_lock, flags); rc = wait_event_interruptible(eas.eas_waitq, eas_should_wake(&eas)); spin_lock_irqsave(&eas.eas_lock, flags); if (rc && !eas.eas_rc) eas.eas_rc = rc; if (eas.eas_rc) break; if (list_empty(&eas.eas_avail)) continue; eap = list_entry(eas.eas_avail.next, struct echo_async_page, eap_item); list_del(&eap->eap_item); spin_unlock_irqrestore(&eas.eas_lock, flags); /* unbind the eap from its old page offset */ if (eap->eap_cookie != NULL) { obd_teardown_async_page(exp, lsm, NULL, eap->eap_cookie); eap->eap_cookie = NULL; } eas.eas_next_offset += PAGE_SIZE; eap->eap_off = eas.eas_next_offset; rc = obd_prep_async_page(exp, lsm, NULL, eap->eap_page, eap->eap_off, &ec_async_page_ops, eap, &eap->eap_cookie); if (rc) { spin_lock_irqsave(&eas.eas_lock, flags); eas.eas_rc = rc; break; } if (oa->o_id != ECHO_PERSISTENT_OBJID && (oa->o_valid & OBD_MD_FLFLAGS) != 0 && (oa->o_flags & OBD_FL_DEBUG_CHECK) != 0) echo_client_page_debug_setup(lsm, eap->eap_page, rw, oa->o_id, eap->eap_off, PAGE_SIZE); /* always asserts urgent, which isn't quite right */ rc = obd_queue_async_io(exp, lsm, NULL, eap->eap_cookie, rw, 0, PAGE_SIZE, 0, ASYNC_READY | ASYNC_URGENT | ASYNC_COUNT_STABLE); spin_lock_irqsave(&eas.eas_lock, flags); if (rc && !eas.eas_rc) { eas.eas_rc = rc; break; } eas.eas_in_flight++; if (eas.eas_next_offset == eas.eas_end_offset) break; } /* still hold the eas_lock here.. */ /* now we just spin waiting for all the rpcs to complete */ while(eas.eas_in_flight) { spin_unlock_irqrestore(&eas.eas_lock, flags); wait_event_interruptible(eas.eas_waitq, eas.eas_in_flight == 0); spin_lock_irqsave(&eas.eas_lock, flags); } spin_unlock_irqrestore(&eas.eas_lock, flags); out: list_for_each_safe(pos, n, &pages) { struct page *page = list_entry(pos, struct page, PAGE_LIST_ENTRY); list_del(&PAGE_LIST(page)); if (page->private != 0) { eap = (struct echo_async_page *)page->private; if (eap->eap_cookie != NULL) obd_teardown_async_page(exp, lsm, NULL, eap->eap_cookie); OBD_FREE(eap, sizeof(*eap)); } __free_page(page); } RETURN(rc); } static int echo_client_prep_commit(struct obd_export *exp, int rw, struct obdo *oa, struct lov_stripe_md *lsm, obd_off offset, obd_size count, obd_size batch, struct obd_trans_info *oti) { struct obd_ioobj ioo; struct niobuf_local *lnb; struct niobuf_remote *rnb; obd_off off; obd_size npages, tot_pages; int i, ret = 0; ENTRY; if (count <= 0 || (count & (PAGE_SIZE - 1)) != 0 || (lsm != NULL && lsm->lsm_object_id != oa->o_id)) RETURN(-EINVAL); npages = batch >> PAGE_SHIFT; tot_pages = count >> PAGE_SHIFT; OBD_ALLOC(lnb, npages * sizeof(struct niobuf_local)); OBD_ALLOC(rnb, npages * sizeof(struct niobuf_remote)); if (lnb == NULL || rnb == NULL) GOTO(out, ret = -ENOMEM); obdo_to_ioobj(oa, &ioo); off = offset; for(; tot_pages; tot_pages -= npages) { if (tot_pages < npages) npages = tot_pages; for (i = 0; i < npages; i++, off += PAGE_SIZE) { rnb[i].offset = off; rnb[i].len = PAGE_SIZE; } /* XXX this can't be the best.. */ memset(oti, 0, sizeof(*oti)); ioo.ioo_bufcnt = npages; ret = obd_preprw(rw, exp, oa, 1, &ioo, npages, rnb, lnb, oti); if (ret != 0) GOTO(out, ret); for (i = 0; i < npages; i++) { struct page *page = lnb[i].page; /* read past eof? */ if (page == NULL && lnb[i].rc == 0) continue; if (oa->o_id == ECHO_PERSISTENT_OBJID || (oa->o_valid & OBD_MD_FLFLAGS) == 0 || (oa->o_flags & OBD_FL_DEBUG_CHECK) == 0) continue; if (rw == OBD_BRW_WRITE) echo_client_page_debug_setup(lsm, page, rw, oa->o_id, rnb[i].offset, rnb[i].len); else echo_client_page_debug_check(lsm, page, oa->o_id, rnb[i].offset, rnb[i].len); } ret = obd_commitrw(rw, exp, oa, 1, &ioo, npages, lnb, oti, ret); if (ret != 0) GOTO(out, ret); } out: if (lnb) OBD_FREE(lnb, npages * sizeof(struct niobuf_local)); if (rnb) OBD_FREE(rnb, npages * sizeof(struct niobuf_remote)); RETURN(ret); } int echo_client_brw_ioctl(int rw, struct obd_export *exp, struct obd_ioctl_data *data) { struct obd_device *obd = class_exp2obd(exp); struct echo_client_obd *ec = &obd->u.echocli; struct obd_trans_info dummy_oti; struct ec_object *eco; int rc; ENTRY; rc = echo_get_object(&eco, obd, &data->ioc_obdo1); if (rc) RETURN(rc); memset(&dummy_oti, 0, sizeof(dummy_oti)); data->ioc_obdo1.o_valid &= ~OBD_MD_FLHANDLE; data->ioc_obdo1.o_valid |= OBD_MD_FLGROUP; data->ioc_obdo1.o_gr = FILTER_GROUP_ECHO; switch((long)data->ioc_pbuf1) { case 1: if (data->ioc_pbuf2 == NULL) { // NULL user data pointer rc = echo_client_kbrw(obd, rw, &data->ioc_obdo1, eco->eco_lsm, data->ioc_offset, data->ioc_count, &dummy_oti); } else { #ifdef __KERNEL__ rc = echo_client_ubrw(obd, rw, &data->ioc_obdo1, eco->eco_lsm, data->ioc_offset, data->ioc_count, data->ioc_pbuf2, &dummy_oti); #endif } break; case 2: rc = echo_client_async_page(ec->ec_exp, rw, &data->ioc_obdo1, eco->eco_lsm, data->ioc_offset, data->ioc_count, data->ioc_plen1); break; case 3: rc = echo_client_prep_commit(ec->ec_exp, rw, &data->ioc_obdo1, eco->eco_lsm, data->ioc_offset, data->ioc_count, data->ioc_plen1, &dummy_oti); break; default: rc = -EINVAL; } echo_put_object(eco); RETURN(rc); } static int echo_ldlm_callback (struct ldlm_lock *lock, struct ldlm_lock_desc *new, void *data, int flag) { struct ec_object *eco = (struct ec_object *)data; struct echo_client_obd *ec = &(eco->eco_device->u.echocli); struct lustre_handle lockh; struct list_head *el; int found = 0; int rc; ldlm_lock2handle (lock, &lockh); /* #ifdef this out if we're not feeling paranoid */ spin_lock (&ec->ec_lock); list_for_each (el, &ec->ec_objects) { found = (eco == list_entry(el, struct ec_object, eco_obj_chain)); if (found) break; } spin_unlock (&ec->ec_lock); LASSERT (found); switch (flag) { case LDLM_CB_BLOCKING: CDEBUG(D_INFO, "blocking callback on "LPX64", handle "LPX64"\n", eco->eco_id, lockh.cookie); rc = ldlm_cli_cancel (&lockh); if (rc != ELDLM_OK) CERROR ("ldlm_cli_cancel failed: %d\n", rc); break; case LDLM_CB_CANCELING: CDEBUG(D_INFO, "cancel callback on "LPX64", handle "LPX64"\n", eco->eco_id, lockh.cookie); break; default: LBUG (); } return (0); } static int echo_client_enqueue(struct obd_export *exp, struct obdo *oa, int mode, obd_off offset, obd_size nob) { struct obd_device *obd = exp->exp_obd; struct echo_client_obd *ec = &obd->u.echocli; struct lustre_handle *ulh = obdo_handle (oa); struct ec_object *eco; struct ec_lock *ecl; int flags; int rc; if (!(mode == LCK_PR || mode == LCK_PW)) return -EINVAL; if ((offset & (PAGE_SIZE - 1)) != 0 || (nob & (PAGE_SIZE - 1)) != 0) return -EINVAL; rc = echo_get_object (&eco, obd, oa); if (rc != 0) return rc; rc = -ENOMEM; OBD_ALLOC (ecl, sizeof (*ecl)); if (ecl == NULL) goto failed_0; ecl->ecl_mode = mode; ecl->ecl_object = eco; ecl->ecl_policy.l_extent.start = offset; ecl->ecl_policy.l_extent.end = (nob == 0) ? ((obd_off) -1) : (offset + nob - 1); flags = 0; rc = obd_enqueue(ec->ec_exp, eco->eco_lsm, LDLM_EXTENT, &ecl->ecl_policy, mode, &flags, echo_ldlm_callback, ldlm_completion_ast, NULL, eco, sizeof(struct ost_lvb), lustre_swab_ost_lvb, &ecl->ecl_lock_handle); if (rc != 0) goto failed_1; CDEBUG(D_INFO, "enqueue handle "LPX64"\n", ecl->ecl_lock_handle.cookie); /* NB ecl takes object ref from echo_get_object() above */ spin_lock(&ec->ec_lock); list_add(&ecl->ecl_exp_chain, &exp->exp_ec_data.eced_locks); ulh->cookie = ecl->ecl_cookie = ec->ec_unique++; spin_unlock(&ec->ec_lock); oa->o_valid |= OBD_MD_FLHANDLE; return 0; failed_1: OBD_FREE (ecl, sizeof (*ecl)); failed_0: echo_put_object (eco); return (rc); } static int echo_client_cancel(struct obd_export *exp, struct obdo *oa) { struct obd_device *obd = exp->exp_obd; struct echo_client_obd *ec = &obd->u.echocli; struct lustre_handle *ulh = obdo_handle (oa); struct ec_lock *ecl = NULL; int found = 0; struct list_head *el; int rc; if ((oa->o_valid & OBD_MD_FLHANDLE) == 0) return -EINVAL; spin_lock (&ec->ec_lock); list_for_each (el, &exp->exp_ec_data.eced_locks) { ecl = list_entry (el, struct ec_lock, ecl_exp_chain); found = (ecl->ecl_cookie == ulh->cookie); if (found) { list_del (&ecl->ecl_exp_chain); break; } } spin_unlock (&ec->ec_lock); if (!found) return (-ENOENT); rc = obd_cancel(ec->ec_exp, ecl->ecl_object->eco_lsm, ecl->ecl_mode, &ecl->ecl_lock_handle); echo_put_object (ecl->ecl_object); OBD_FREE (ecl, sizeof (*ecl)); return rc; } static int echo_client_iocontrol(unsigned int cmd, struct obd_export *exp, int len, void *karg, void *uarg) { struct obd_device *obd; struct echo_client_obd *ec; struct ec_object *eco; struct obd_ioctl_data *data = karg; struct obd_trans_info dummy_oti; struct oti_req_ack_lock *ack_lock; struct obdo *oa; int rw = OBD_BRW_READ; int rc = 0; int i; ENTRY; unlock_kernel(); memset(&dummy_oti, 0, sizeof(dummy_oti)); obd = exp->exp_obd; ec = &obd->u.echocli; switch (cmd) { case OBD_IOC_CREATE: /* may create echo object */ if (!capable (CAP_SYS_ADMIN)) GOTO (out, rc = -EPERM); rc = echo_create_object (obd, 1, &data->ioc_obdo1, data->ioc_pbuf1, data->ioc_plen1, &dummy_oti); GOTO(out, rc); case OBD_IOC_DESTROY: if (!capable (CAP_SYS_ADMIN)) GOTO (out, rc = -EPERM); rc = echo_get_object (&eco, obd, &data->ioc_obdo1); if (rc == 0) { oa = &data->ioc_obdo1; oa->o_gr = FILTER_GROUP_ECHO; oa->o_valid |= OBD_MD_FLGROUP; rc = obd_destroy(ec->ec_exp, oa, eco->eco_lsm, &dummy_oti); if (rc == 0) eco->eco_deleted = 1; echo_put_object(eco); } GOTO(out, rc); case OBD_IOC_GETATTR: rc = echo_get_object (&eco, obd, &data->ioc_obdo1); if (rc == 0) { rc = obd_getattr(ec->ec_exp, &data->ioc_obdo1, eco->eco_lsm); echo_put_object(eco); } GOTO(out, rc); case OBD_IOC_SETATTR: if (!capable (CAP_SYS_ADMIN)) GOTO (out, rc = -EPERM); rc = echo_get_object (&eco, obd, &data->ioc_obdo1); if (rc == 0) { rc = obd_setattr(ec->ec_exp, &data->ioc_obdo1, eco->eco_lsm, NULL); echo_put_object(eco); } GOTO(out, rc); case OBD_IOC_BRW_WRITE: if (!capable (CAP_SYS_ADMIN)) GOTO (out, rc = -EPERM); rw = OBD_BRW_WRITE; /* fall through */ case OBD_IOC_BRW_READ: rc = echo_client_brw_ioctl(rw, exp, data); GOTO(out, rc); case ECHO_IOC_GET_STRIPE: rc = echo_get_object(&eco, obd, &data->ioc_obdo1); if (rc == 0) { rc = echo_copyout_lsm(eco->eco_lsm, data->ioc_pbuf1, data->ioc_plen1); echo_put_object(eco); } GOTO(out, rc); case ECHO_IOC_SET_STRIPE: if (!capable (CAP_SYS_ADMIN)) GOTO (out, rc = -EPERM); if (data->ioc_pbuf1 == NULL) { /* unset */ rc = echo_get_object(&eco, obd, &data->ioc_obdo1); if (rc == 0) { eco->eco_deleted = 1; echo_put_object(eco); } } else { rc = echo_create_object(obd, 0, &data->ioc_obdo1, data->ioc_pbuf1, data->ioc_plen1, &dummy_oti); } GOTO (out, rc); case ECHO_IOC_ENQUEUE: if (!capable (CAP_SYS_ADMIN)) GOTO (out, rc = -EPERM); rc = echo_client_enqueue(exp, &data->ioc_obdo1, data->ioc_conn1, /* lock mode */ data->ioc_offset, data->ioc_count);/*extent*/ GOTO (out, rc); case ECHO_IOC_CANCEL: rc = echo_client_cancel(exp, &data->ioc_obdo1); GOTO (out, rc); default: CERROR ("echo_ioctl(): unrecognised ioctl %#x\n", cmd); GOTO (out, rc = -ENOTTY); } EXIT; out: /* XXX this should be in a helper also called by target_send_reply */ for (ack_lock = dummy_oti.oti_ack_locks, i = 0; i < 4; i++, ack_lock++) { if (!ack_lock->mode) break; ldlm_lock_decref(&ack_lock->lock, ack_lock->mode); } lock_kernel(); return rc; } static int echo_client_setup(struct obd_device *obddev, obd_count len, void *buf) { struct lustre_cfg* lcfg = buf; struct echo_client_obd *ec = &obddev->u.echocli; struct obd_device *tgt; struct lustre_handle conn = {0, }; struct obd_uuid echo_uuid = { "ECHO_UUID" }; int rc; ENTRY; if (lcfg->lcfg_inllen1 < 1) { CERROR("requires a TARGET OBD name\n"); RETURN(-EINVAL); } tgt = class_name2obd(lcfg->lcfg_inlbuf1); if (!tgt || !tgt->obd_attached || !tgt->obd_set_up) { CERROR("device not attached or not set up (%s)\n", lcfg->lcfg_inlbuf1); RETURN(-EINVAL); } spin_lock_init (&ec->ec_lock); INIT_LIST_HEAD (&ec->ec_objects); ec->ec_unique = 0; rc = obd_connect(&conn, tgt, &echo_uuid, NULL, FILTER_GROUP_ECHO); if (rc) { CERROR("fail to connect to device %s\n", lcfg->lcfg_inlbuf1); return (rc); } ec->ec_exp = class_conn2export(&conn); RETURN(rc); } static int echo_client_cleanup(struct obd_device *obddev, int flags) { struct list_head *el; struct ec_object *eco; struct echo_client_obd *ec = &obddev->u.echocli; int rc; ENTRY; if (!list_empty(&obddev->obd_exports)) { CERROR("still has clients!\n"); RETURN(-EBUSY); } /* XXX assuming sole access */ while (!list_empty(&ec->ec_objects)) { el = ec->ec_objects.next; eco = list_entry(el, struct ec_object, eco_obj_chain); LASSERT(eco->eco_refcount == 0); eco->eco_refcount = 1; eco->eco_deleted = 1; echo_put_object(eco); } rc = obd_disconnect(ec->ec_exp, 0); if (rc != 0) CERROR("fail to disconnect device: %d\n", rc); RETURN(rc); } static int echo_client_connect(struct lustre_handle *conn, struct obd_device *src, struct obd_uuid *cluuid, struct obd_connect_data *data, unsigned long flags) { struct obd_export *exp; int rc; rc = class_connect(conn, src, cluuid); if (rc == 0) { exp = class_conn2export(conn); INIT_LIST_HEAD(&exp->exp_ec_data.eced_locks); class_export_put(exp); } RETURN (rc); } static int echo_client_disconnect(struct obd_export *exp, unsigned long flags) { struct obd_device *obd; struct echo_client_obd *ec; struct ec_lock *ecl; int rc; ENTRY; if (exp == NULL) GOTO(out, rc = -EINVAL); obd = exp->exp_obd; ec = &obd->u.echocli; /* no more contention on export's lock list */ while (!list_empty (&exp->exp_ec_data.eced_locks)) { ecl = list_entry (exp->exp_ec_data.eced_locks.next, struct ec_lock, ecl_exp_chain); list_del (&ecl->ecl_exp_chain); rc = obd_cancel(ec->ec_exp, ecl->ecl_object->eco_lsm, ecl->ecl_mode, &ecl->ecl_lock_handle); CDEBUG (D_INFO, "Cancel lock on object "LPX64" on disconnect " "(%d)\n", ecl->ecl_object->eco_id, rc); echo_put_object (ecl->ecl_object); OBD_FREE (ecl, sizeof (*ecl)); } rc = class_disconnect(exp, 0); GOTO(out, rc); out: return rc; } static struct obd_ops echo_obd_ops = { .o_owner = THIS_MODULE, .o_setup = echo_client_setup, .o_cleanup = echo_client_cleanup, .o_iocontrol = echo_client_iocontrol, .o_connect = echo_client_connect, .o_disconnect = echo_client_disconnect }; int echo_client_init(void) { struct lprocfs_static_vars lvars; lprocfs_init_vars(echo, &lvars); return class_register_type(&echo_obd_ops, NULL, lvars.module_vars, OBD_ECHO_CLIENT_DEVICENAME); } void echo_client_exit(void) { class_unregister_type(OBD_ECHO_CLIENT_DEVICENAME); }