From 9369a431bce1e985597eda32992960c969b27c5b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 18 Sep 2012 20:57:08 -0400 Subject: NFSv4.1: Cleanup; add "pnfs_" prefix to put_lseg() and get_lseg() Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 53f94d915bd1..77cd1151ef0d 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -453,7 +453,7 @@ static void filelayout_commit_release(void *calldata) struct nfs_commit_data *data = calldata; data->completion_ops->completion(data); - put_lseg(data->lseg); + pnfs_put_lseg(data->lseg); nfs_put_client(data->ds_clp); nfs_commitdata_release(data); } @@ -931,7 +931,7 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio, nfs_init_cinfo(&cinfo, pgio->pg_inode, pgio->pg_dreq); status = filelayout_alloc_commit_info(pgio->pg_lseg, &cinfo, GFP_NOFS); if (status < 0) { - put_lseg(pgio->pg_lseg); + pnfs_put_lseg(pgio->pg_lseg); pgio->pg_lseg = NULL; goto out_mds; } @@ -985,7 +985,7 @@ filelayout_clear_request_commit(struct nfs_page *req, out: nfs_request_remove_commit_list(req, cinfo); spin_unlock(cinfo->lock); - put_lseg(freeme); + pnfs_put_lseg(freeme); } static struct list_head * @@ -1018,7 +1018,7 @@ filelayout_choose_commit_list(struct nfs_page *req, * off due to a rewrite, in which case it will be done in * filelayout_clear_request_commit */ - buckets[i].wlseg = get_lseg(lseg); + buckets[i].wlseg = pnfs_get_lseg(lseg); } set_bit(PG_COMMIT_TO_DS, &req->wb_flags); cinfo->ds->nwritten++; @@ -1128,7 +1128,7 @@ filelayout_scan_ds_commit_list(struct pnfs_commit_bucket *bucket, if (list_empty(src)) bucket->wlseg = NULL; else - get_lseg(bucket->clseg); + pnfs_get_lseg(bucket->clseg); } return ret; } @@ -1159,12 +1159,12 @@ static void filelayout_recover_commit_reqs(struct list_head *dst, /* NOTE cinfo->lock is NOT held, relying on fact that this is * only called on single thread per dreq. - * Can't take the lock because need to do put_lseg + * Can't take the lock because need to do pnfs_put_lseg */ for (i = 0, b = cinfo->ds->buckets; i < cinfo->ds->nbuckets; i++, b++) { if (transfer_commit_list(&b->written, dst, cinfo, 0)) { BUG_ON(!list_empty(&b->written)); - put_lseg(b->wlseg); + pnfs_put_lseg(b->wlseg); b->wlseg = NULL; } } @@ -1200,7 +1200,7 @@ alloc_ds_commits(struct nfs_commit_info *cinfo, struct list_head *list) if (list_empty(&bucket->committing)) continue; nfs_retry_commit(&bucket->committing, bucket->clseg, cinfo); - put_lseg(bucket->clseg); + pnfs_put_lseg(bucket->clseg); bucket->clseg = NULL; } /* Caller will clean up entries put on list */ -- cgit v1.2.3 From 78e4e05c643768af170e5a4b21712d9a7a26cce5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 18 Sep 2012 21:02:29 -0400 Subject: NFSv4.1: Replace get_device_info() with filelayout_get_device_info() Fix the namespace pollution issue. Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/blocklayout.c | 2 +- fs/nfs/nfs4filelayout.c | 2 +- fs/nfs/nfs4filelayout.h | 2 +- fs/nfs/nfs4filelayoutdev.c | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c index dd392ed5f2e2..329bfbfed37d 100644 --- a/fs/nfs/blocklayout/blocklayout.c +++ b/fs/nfs/blocklayout/blocklayout.c @@ -874,7 +874,7 @@ static void free_blk_mountid(struct block_mount_id *mid) } } -/* This is mostly copied from the filelayout's get_device_info function. +/* This is mostly copied from the filelayout_get_device_info function. * It seems much of this should be at the generic pnfs level. */ static struct pnfs_block_dev * diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 77cd1151ef0d..af6ee4ad3f15 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -608,7 +608,7 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, d = nfs4_find_get_deviceid(NFS_SERVER(lo->plh_inode)->pnfs_curr_ld, NFS_SERVER(lo->plh_inode)->nfs_client, id); if (d == NULL) { - dsaddr = get_device_info(lo->plh_inode, id, gfp_flags); + dsaddr = filelayout_get_device_info(lo->plh_inode, id, gfp_flags); if (dsaddr == NULL) goto out; } else diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index 43fe802dd678..11053c425a61 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -158,7 +158,7 @@ struct nfs4_pnfs_ds *nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, extern void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr); extern void nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr); struct nfs4_file_layout_dsaddr * -get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags); +filelayout_get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags); void nfs4_ds_disconnect(struct nfs_client *clp); #endif /* FS_NFS_NFS4FILELAYOUT_H */ diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index f81231f30d94..b85a29df20ae 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -690,7 +690,7 @@ decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_fl * of available devices, and return it. */ struct nfs4_file_layout_dsaddr * -get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags) +filelayout_get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags) { struct pnfs_device *pdev = NULL; u32 max_resp_sz; -- cgit v1.2.3 From 1dfed2737d8cfe2f2378fddfb3bed126ff5474e7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 18 Sep 2012 19:51:12 -0400 Subject: NFSv4.1: pNFS data servers may be temporarily offline In cases where the pNFS data server is just temporarily out of service, we want to mark it as such, and then try again later. Typically that will be in cases of network connection errors etc. This patch allows us to mark the devices as being "unavailable" for such transient errors, and will make them available for retries after a 2 minute timeout period. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 22 +++++++++++++++++++--- fs/nfs/nfs4filelayout.h | 8 ++------ fs/nfs/nfs4filelayoutdev.c | 15 +++++++-------- fs/nfs/pnfs.h | 4 ++++ fs/nfs/pnfs_dev.c | 27 +++++++++++++++++++++++++++ 5 files changed, 59 insertions(+), 17 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index af6ee4ad3f15..dac2162c3ac4 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -205,7 +205,7 @@ static int filelayout_async_handle_error(struct rpc_task *task, case -EPIPE: dprintk("%s DS connection error %d\n", __func__, task->tk_status); - filelayout_mark_devid_invalid(devid); + nfs4_mark_deviceid_unavailable(devid); clear_bit(NFS_INO_LAYOUTCOMMIT, &NFS_I(inode)->flags); _pnfs_return_layout(inode); rpc_wake_up(&tbl->slot_tbl_waitq); @@ -269,6 +269,22 @@ filelayout_set_layoutcommit(struct nfs_write_data *wdata) (unsigned long) NFS_I(hdr->inode)->layout->plh_lwb); } +bool +filelayout_test_devid_unavailable(struct nfs4_deviceid_node *node) +{ + return filelayout_test_devid_invalid(node) || + nfs4_test_deviceid_unavailable(node); +} + +static bool +filelayout_reset_to_mds(struct pnfs_layout_segment *lseg) +{ + struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg); + + return filelayout_test_layout_invalid(lseg->pls_layout) || + filelayout_test_devid_unavailable(node); +} + /* * Call ops for the async read/write cases * In the case of dense layouts, the offset needs to be reset to its @@ -613,8 +629,8 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, goto out; } else dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node); - /* Found deviceid is being reaped */ - if (test_bit(NFS_DEVICEID_INVALID, &dsaddr->id_node.flags)) + /* Found deviceid is unavailable */ + if (filelayout_test_devid_unavailable(&dsaddr->id_node)) goto out_put; fl->dsaddr = dsaddr; diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index 11053c425a61..10b0f134400b 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -140,12 +140,8 @@ filelayout_test_devid_invalid(struct nfs4_deviceid_node *node) return test_bit(NFS_DEVICEID_INVALID, &node->flags); } -static inline bool -filelayout_reset_to_mds(struct pnfs_layout_segment *lseg) -{ - return filelayout_test_devid_invalid(FILELAYOUT_DEVID_NODE(lseg)) || - filelayout_test_layout_invalid(lseg->pls_layout); -} +extern bool +filelayout_test_devid_unavailable(struct nfs4_deviceid_node *node); extern struct nfs_fh * nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j); diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index b85a29df20ae..3336d5eaf879 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -804,13 +804,14 @@ nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx) struct nfs4_pnfs_ds *ds = dsaddr->ds_list[ds_idx]; struct nfs4_deviceid_node *devid = FILELAYOUT_DEVID_NODE(lseg); - if (filelayout_test_devid_invalid(devid)) + if (filelayout_test_devid_unavailable(devid)) return NULL; if (ds == NULL) { printk(KERN_ERR "NFS: %s: No data server for offset index %d\n", __func__, ds_idx); - goto mark_dev_invalid; + filelayout_mark_devid_invalid(devid); + return NULL; } if (!ds->ds_clp) { @@ -818,14 +819,12 @@ nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx) int err; err = nfs4_ds_connect(s, ds); - if (err) - goto mark_dev_invalid; + if (err) { + nfs4_mark_deviceid_unavailable(devid); + return NULL; + } } return ds; - -mark_dev_invalid: - filelayout_mark_devid_invalid(devid); - return NULL; } module_param(dataserver_retrans, uint, 0644); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index bc8e5001203d..9735031e1e1a 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -234,6 +234,7 @@ struct nfs4_threshold *pnfs_mdsthreshold_alloc(void); /* nfs4_deviceid_flags */ enum { NFS_DEVICEID_INVALID = 0, /* set when MDS clientid recalled */ + NFS_DEVICEID_UNAVAILABLE, /* device temporarily unavailable */ }; /* pnfs_dev.c */ @@ -243,6 +244,7 @@ struct nfs4_deviceid_node { const struct pnfs_layoutdriver_type *ld; const struct nfs_client *nfs_client; unsigned long flags; + unsigned long timestamp_unavailable; struct nfs4_deviceid deviceid; atomic_t ref; }; @@ -255,6 +257,8 @@ void nfs4_init_deviceid_node(struct nfs4_deviceid_node *, const struct nfs4_deviceid *); struct nfs4_deviceid_node *nfs4_insert_deviceid_node(struct nfs4_deviceid_node *); bool nfs4_put_deviceid_node(struct nfs4_deviceid_node *); +void nfs4_mark_deviceid_unavailable(struct nfs4_deviceid_node *node); +bool nfs4_test_deviceid_unavailable(struct nfs4_deviceid_node *node); void nfs4_deviceid_purge_client(const struct nfs_client *); static inline void diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c index 73f701f1f4d3..d35b62e83ea6 100644 --- a/fs/nfs/pnfs_dev.c +++ b/fs/nfs/pnfs_dev.c @@ -40,6 +40,8 @@ #define NFS4_DEVICE_ID_HASH_SIZE (1 << NFS4_DEVICE_ID_HASH_BITS) #define NFS4_DEVICE_ID_HASH_MASK (NFS4_DEVICE_ID_HASH_SIZE - 1) +#define PNFS_DEVICE_RETRY_TIMEOUT (120*HZ) + static struct hlist_head nfs4_deviceid_cache[NFS4_DEVICE_ID_HASH_SIZE]; static DEFINE_SPINLOCK(nfs4_deviceid_lock); @@ -218,6 +220,30 @@ nfs4_put_deviceid_node(struct nfs4_deviceid_node *d) } EXPORT_SYMBOL_GPL(nfs4_put_deviceid_node); +void +nfs4_mark_deviceid_unavailable(struct nfs4_deviceid_node *node) +{ + node->timestamp_unavailable = jiffies; + set_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags); +} +EXPORT_SYMBOL_GPL(nfs4_mark_deviceid_unavailable); + +bool +nfs4_test_deviceid_unavailable(struct nfs4_deviceid_node *node) +{ + if (test_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags)) { + unsigned long start, end; + + end = jiffies; + start = end - PNFS_DEVICE_RETRY_TIMEOUT; + if (time_in_range(node->timestamp_unavailable, start, end)) + return true; + clear_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags); + } + return false; +} +EXPORT_SYMBOL_GPL(nfs4_test_deviceid_unavailable); + static void _deviceid_purge_client(const struct nfs_client *clp, long hash) { @@ -276,3 +302,4 @@ nfs4_deviceid_mark_client_invalid(struct nfs_client *clp) } rcu_read_unlock(); } + -- cgit v1.2.3 From 965938b83b19aeffdc1d16ce9947c8c127e8f59b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 20 Sep 2012 15:07:45 -0400 Subject: NFSv4.1: Get rid of pNFS layout state "NFS_LAYOUT_INVALID" In all cases where we set NFS_LAYOUT_INVALID, we also set NFS_LAYOUT_DESTROYED. Furthermore, in all cases where we test for NFS_LAYOUT_INVALID, we should also be testing for NFS_LAYOUT_DESTROYED, since the latter means that we hold no valid layout segments. Ergo the two are redundant. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 4 +--- fs/nfs/nfs4filelayout.h | 6 ------ fs/nfs/pnfs.h | 7 ++++++- 3 files changed, 7 insertions(+), 10 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index dac2162c3ac4..6cce57e7fe55 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -190,8 +190,6 @@ static int filelayout_async_handle_error(struct rpc_task *task, * i/o and all i/o waiting on the slot table to the MDS until * layout is destroyed and a new valid layout is obtained. */ - set_bit(NFS_LAYOUT_INVALID, - &NFS_I(inode)->layout->plh_flags); pnfs_destroy_layout(NFS_I(inode)); rpc_wake_up(&tbl->slot_tbl_waitq); goto reset; @@ -281,7 +279,7 @@ filelayout_reset_to_mds(struct pnfs_layout_segment *lseg) { struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg); - return filelayout_test_layout_invalid(lseg->pls_layout) || + return pnfs_test_layout_destroyed(lseg->pls_layout) || filelayout_test_devid_unavailable(node); } diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index 10b0f134400b..dca47d786710 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -128,12 +128,6 @@ filelayout_mark_devid_invalid(struct nfs4_deviceid_node *node) set_bit(NFS_DEVICEID_INVALID, &node->flags); } -static inline bool -filelayout_test_layout_invalid(struct pnfs_layout_hdr *lo) -{ - return test_bit(NFS_LAYOUT_INVALID, &lo->plh_flags); -} - static inline bool filelayout_test_devid_invalid(struct nfs4_deviceid_node *node) { diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index aa9fa1b1ff4a..aacda7fbb536 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -63,7 +63,6 @@ enum { NFS_LAYOUT_BULK_RECALL, /* bulk recall affecting layout */ NFS_LAYOUT_ROC, /* some lseg had roc bit set */ NFS_LAYOUT_DESTROYED, /* no new use of layout allowed */ - NFS_LAYOUT_INVALID, /* layout is being destroyed */ NFS_LAYOUT_RETURNED, /* layout has already been returned */ }; @@ -279,6 +278,12 @@ pnfs_test_layout_returned(struct pnfs_layout_hdr *lo) return test_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags); } +static inline bool +pnfs_test_layout_destroyed(struct pnfs_layout_hdr *lo) +{ + return test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags); +} + static inline struct pnfs_layout_segment * pnfs_get_lseg(struct pnfs_layout_segment *lseg) { -- cgit v1.2.3 From 8006bfba36d42b6976ed92979f51e5f9bef2625c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 21 Sep 2012 14:48:04 -0400 Subject: NFSv4.1: Get rid of the NFS_LAYOUT_DESTROYED state We already have a mechanism for blocking LAYOUTGET by means of the plh_block_lgets counter. The only "service" that NFS_LAYOUT_DESTROYED provides at this point is to block layoutget once the layout segment list is empty, which basically means that you have to wait until the pnfs_layout_hdr is destroyed before you can do pNFS on that file again. This patch enables the reuse of the pnfs_layout_hdr if the layout segment list is empty. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 3 +-- fs/nfs/pnfs.c | 9 +-------- fs/nfs/pnfs.h | 7 ------- 3 files changed, 2 insertions(+), 17 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 6cce57e7fe55..52d847212066 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -279,8 +279,7 @@ filelayout_reset_to_mds(struct pnfs_layout_segment *lseg) { struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg); - return pnfs_test_layout_destroyed(lseg->pls_layout) || - filelayout_test_devid_unavailable(node); + return filelayout_test_devid_unavailable(node); } /* diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index fcc72ecf2fd7..bda88a275071 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -331,8 +331,6 @@ pnfs_layout_remove_lseg(struct pnfs_layout_hdr *lo, list_del_init(&lseg->pls_list); /* Matched by pnfs_get_layout_hdr in pnfs_layout_insert_lseg */ atomic_dec(&lo->plh_refcount); - if (list_empty(&lo->plh_segs)) - set_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags); rpc_wake_up(&NFS_SERVER(inode)->roc_rpcwaitq); } @@ -463,10 +461,8 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, dprintk("%s:Begin lo %p\n", __func__, lo); - if (list_empty(&lo->plh_segs)) { - set_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags); + if (list_empty(&lo->plh_segs)) return 0; - } list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) if (!recall_range || should_free_lseg(&lseg->pls_range, recall_range)) { @@ -590,7 +586,6 @@ pnfs_layoutgets_blocked(struct pnfs_layout_hdr *lo, nfs4_stateid *stateid, (int)(lo->plh_barrier - be32_to_cpu(stateid->seqid)) >= 0) return true; return lo->plh_block_lgets || - test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags) || test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags) || (list_empty(&lo->plh_segs) && (atomic_read(&lo->plh_outstanding) > lget)); @@ -912,8 +907,6 @@ pnfs_find_alloc_layout(struct inode *ino, dprintk("%s Begin ino=%p layout=%p\n", __func__, ino, nfsi->layout); if (nfsi->layout) { - if (test_bit(NFS_LAYOUT_DESTROYED, &nfsi->layout->plh_flags)) - return NULL; pnfs_get_layout_hdr(nfsi->layout); return nfsi->layout; } diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index aacda7fbb536..92f6ce6532ba 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -62,7 +62,6 @@ enum { NFS_LAYOUT_RW_FAILED, /* get rw layout failed stop trying */ NFS_LAYOUT_BULK_RECALL, /* bulk recall affecting layout */ NFS_LAYOUT_ROC, /* some lseg had roc bit set */ - NFS_LAYOUT_DESTROYED, /* no new use of layout allowed */ NFS_LAYOUT_RETURNED, /* layout has already been returned */ }; @@ -278,12 +277,6 @@ pnfs_test_layout_returned(struct pnfs_layout_hdr *lo) return test_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags); } -static inline bool -pnfs_test_layout_destroyed(struct pnfs_layout_hdr *lo) -{ - return test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags); -} - static inline struct pnfs_layout_segment * pnfs_get_lseg(struct pnfs_layout_segment *lseg) { -- cgit v1.2.3 From 8fcdc31b3d09bc348ff9bf752ae1291828756cfa Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 11 Oct 2012 12:26:04 -0400 Subject: NFSv4.1: Kill nfs4_ds_disconnect() There is nothing to prevent another thread from dereferencing ds->ds_clp during or after the call to nfs4_ds_disconnect(), and Oopsing due to the resulting NULL pointer. Instead, we should just rely on filelayout_mark_devid_invalid() to keep us out of trouble by avoiding that deviceid. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 1 - fs/nfs/nfs4filelayout.h | 1 - fs/nfs/nfs4filelayoutdev.c | 22 ---------------------- 3 files changed, 24 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 52d847212066..816c2d0d133a 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -207,7 +207,6 @@ static int filelayout_async_handle_error(struct rpc_task *task, clear_bit(NFS_INO_LAYOUTCOMMIT, &NFS_I(inode)->flags); _pnfs_return_layout(inode); rpc_wake_up(&tbl->slot_tbl_waitq); - nfs4_ds_disconnect(clp); /* fall through */ default: reset: diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index dca47d786710..8c07241fe52b 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -149,6 +149,5 @@ extern void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr); extern void nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr); struct nfs4_file_layout_dsaddr * filelayout_get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags); -void nfs4_ds_disconnect(struct nfs_client *clp); #endif /* FS_NFS_NFS4FILELAYOUT_H */ diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index 3336d5eaf879..a8eaa9b7bb0f 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -148,28 +148,6 @@ _data_server_lookup_locked(const struct list_head *dsaddrs) return NULL; } -/* - * Lookup DS by nfs_client pointer. Zero data server client pointer - */ -void nfs4_ds_disconnect(struct nfs_client *clp) -{ - struct nfs4_pnfs_ds *ds; - struct nfs_client *found = NULL; - - dprintk("%s clp %p\n", __func__, clp); - spin_lock(&nfs4_ds_cache_lock); - list_for_each_entry(ds, &nfs4_data_server_cache, ds_node) - if (ds->ds_clp && ds->ds_clp == clp) { - found = ds->ds_clp; - ds->ds_clp = NULL; - } - spin_unlock(&nfs4_ds_cache_lock); - if (found) { - set_bit(NFS_CS_STOP_RENEW, &clp->cl_res_state); - nfs_put_client(clp); - } -} - /* * Create an rpc connection to the nfs4_pnfs_ds data server * Currently only supports IPv4 and IPv6 addresses -- cgit v1.2.3 From d527e5c15de8de813cd0a2ad0b769f68c6226938 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 11 Oct 2012 13:43:38 -0400 Subject: NFSv4.1: Do not call pnfs_return_layout() from an rpciod context Move the call to pnfs_return_layout() to the read and write rpc_release() callbacks, so that it gets called from nfsiod, which is a more appropriate context. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 18 +++++++++++++++--- fs/nfs/pnfs.h | 1 + 2 files changed, 16 insertions(+), 3 deletions(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 816c2d0d133a..e7aee566861d 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -122,12 +122,21 @@ static void filelayout_reset_read(struct nfs_read_data *data) } } +static void filelayout_fenceme(struct inode *inode, struct pnfs_layout_hdr *lo) +{ + if (!test_and_clear_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) + return; + clear_bit(NFS_INO_LAYOUTCOMMIT, &NFS_I(inode)->flags); + pnfs_return_layout(inode); +} + static int filelayout_async_handle_error(struct rpc_task *task, struct nfs4_state *state, struct nfs_client *clp, struct pnfs_layout_segment *lseg) { - struct inode *inode = lseg->pls_layout->plh_inode; + struct pnfs_layout_hdr *lo = lseg->pls_layout; + struct inode *inode = lo->plh_inode; struct nfs_server *mds_server = NFS_SERVER(inode); struct nfs4_deviceid_node *devid = FILELAYOUT_DEVID_NODE(lseg); struct nfs_client *mds_client = mds_server->nfs_client; @@ -204,8 +213,7 @@ static int filelayout_async_handle_error(struct rpc_task *task, dprintk("%s DS connection error %d\n", __func__, task->tk_status); nfs4_mark_deviceid_unavailable(devid); - clear_bit(NFS_INO_LAYOUTCOMMIT, &NFS_I(inode)->flags); - _pnfs_return_layout(inode); + set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags); rpc_wake_up(&tbl->slot_tbl_waitq); /* fall through */ default: @@ -330,7 +338,9 @@ static void filelayout_read_count_stats(struct rpc_task *task, void *data) static void filelayout_read_release(void *data) { struct nfs_read_data *rdata = data; + struct pnfs_layout_hdr *lo = rdata->header->lseg->pls_layout; + filelayout_fenceme(lo->plh_inode, lo); nfs_put_client(rdata->ds_clp); rdata->header->mds_ops->rpc_release(data); } @@ -428,7 +438,9 @@ static void filelayout_write_count_stats(struct rpc_task *task, void *data) static void filelayout_write_release(void *data) { struct nfs_write_data *wdata = data; + struct pnfs_layout_hdr *lo = wdata->header->lseg->pls_layout; + filelayout_fenceme(lo->plh_inode, lo); nfs_put_client(wdata->ds_clp); wdata->header->mds_ops->rpc_release(data); } diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 2d722dba1111..dbf7bba52da0 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -62,6 +62,7 @@ enum { NFS_LAYOUT_RW_FAILED, /* get rw layout failed stop trying */ NFS_LAYOUT_BULK_RECALL, /* bulk recall affecting layout */ NFS_LAYOUT_ROC, /* some lseg had roc bit set */ + NFS_LAYOUT_RETURN, /* Return this layout ASAP */ }; enum layoutdriver_policy_flags { -- cgit v1.2.3 From 1813badd98ce02e4b96d8997b68ddef4d4ad4ec5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 11 Oct 2012 14:36:52 -0400 Subject: NFSv4.1: Use kcalloc() to allocate zeroed arrays instead of kzalloc() Don't circumvent the array size checks. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/nfs4filelayout.c') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index e7aee566861d..2e45fd9c02a3 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -750,7 +750,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo, goto out_err; if (fl->num_fh > 0) { - fl->fh_array = kzalloc(fl->num_fh * sizeof(struct nfs_fh *), + fl->fh_array = kcalloc(fl->num_fh, sizeof(fl->fh_array[0]), gfp_flags); if (!fl->fh_array) goto out_err; -- cgit v1.2.3