From patchwork Sun May 29 18:18:14 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Boaz Harrosh X-Patchwork-Id: 828062 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by demeter2.kernel.org (8.14.4/8.14.3) with ESMTP id p4TIH6Ec026946 for ; Sun, 29 May 2011 18:18:27 GMT Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755078Ab1E2SS0 (ORCPT ); Sun, 29 May 2011 14:18:26 -0400 Received: from daytona.panasas.com ([67.152.220.89]:50774 "EHLO daytona.panasas.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755038Ab1E2SS0 (ORCPT ); Sun, 29 May 2011 14:18:26 -0400 Received: from fs2.bhalevy.com ([172.17.33.72]) by daytona.panasas.com with Microsoft SMTPSVC(6.0.3790.4675); Sun, 29 May 2011 14:18:24 -0400 From: Boaz Harrosh To: Trond Myklebust , Benny Halevy , NFS list , open-osd Cc: Benny Halevy Subject: [PATCH 04/32] NFSv4.1: make deviceid cache global Date: Sun, 29 May 2011 21:18:14 +0300 Message-Id: <1306693094-7815-1-git-send-email-bharrosh@panasas.com> X-Mailer: git-send-email 1.7.2.3 In-Reply-To: <4DE28CFB.4060608@panasas.com> References: <4DE28CFB.4060608@panasas.com> X-OriginalArrivalTime: 29 May 2011 18:18:25.0021 (UTC) FILETIME=[CCCBE2D0:01CC1E2C] Sender: linux-nfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-nfs@vger.kernel.org X-Greylist: IP, sender and recipient auto-whitelisted, not delayed by milter-greylist-4.2.6 (demeter2.kernel.org [140.211.167.43]); Sun, 29 May 2011 18:18:27 +0000 (UTC) From: Benny Halevy Move deviceid cache from the pnfs files layout driver to the generic layer in preparation for the objects layout driver. Signed-off-by: Benny Halevy --- fs/nfs/Makefile | 2 +- fs/nfs/nfs4filelayout.c | 10 ++- fs/nfs/nfs4filelayout.h | 8 +-- fs/nfs/nfs4filelayoutdev.c | 104 ++++------------------------- fs/nfs/pnfs.h | 17 +++++ fs/nfs/pnfs_dev.c | 156 ++++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 194 insertions(+), 103 deletions(-) create mode 100644 fs/nfs/pnfs_dev.c diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index 4776ff9..7516a8a 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -15,7 +15,7 @@ nfs-$(CONFIG_NFS_V4) += nfs4proc.o nfs4xdr.o nfs4state.o nfs4renewd.o \ delegation.o idmap.o \ callback.o callback_xdr.o callback_proc.o \ nfs4namespace.o -nfs-$(CONFIG_NFS_V4_1) += pnfs.o +nfs-$(CONFIG_NFS_V4_1) += pnfs.o pnfs_dev.o nfs-$(CONFIG_SYSCTL) += sysctl.o nfs-$(CONFIG_NFS_FSCACHE) += fscache.o fscache-index.o diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 571c1b0..4c67a6f 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -421,6 +421,7 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, struct nfs4_deviceid *id, gfp_t gfp_flags) { + struct nfs4_deviceid_node *d; struct nfs4_file_layout_dsaddr *dsaddr; int status = -EINVAL; struct nfs_server *nfss = NFS_SERVER(lo->plh_inode); @@ -440,12 +441,13 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, } /* find and reference the deviceid */ - dsaddr = nfs4_fl_find_get_deviceid(NFS_SERVER(lo->plh_inode)->nfs_client, id); - if (dsaddr == NULL) { + d = nfs4_find_get_deviceid(NFS_SERVER(lo->plh_inode)->nfs_client, id); + if (d == NULL) { dsaddr = get_device_info(lo->plh_inode, id, gfp_flags); if (dsaddr == NULL) goto out; - } + } else + dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node); fl->dsaddr = dsaddr; if (fl->first_stripe_index < 0 || @@ -535,7 +537,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo, memcpy(id, p, sizeof(*id)); p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE); - print_deviceid(id); + nfs4_print_deviceid(id); nfl_util = be32_to_cpup(p++); if (nfl_util & NFL4_UFLG_COMMIT_THRU_MDS) diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index 301b955..0ace0a2 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -59,10 +59,7 @@ struct nfs4_pnfs_ds { #define NFS4_DEVICE_ID_NEG_ENTRY 0x00000001 struct nfs4_file_layout_dsaddr { - struct hlist_node node; - struct nfs_client *nfs_client; - struct nfs4_deviceid deviceid; - atomic_t ref; + struct nfs4_deviceid_node id_node; unsigned long flags; u32 stripe_count; u8 *stripe_indices; @@ -96,13 +93,10 @@ extern struct nfs_fh * nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j); extern void print_ds(struct nfs4_pnfs_ds *ds); -extern void print_deviceid(struct nfs4_deviceid *dev_id); u32 nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset); u32 nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j); struct nfs4_pnfs_ds *nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx); -extern struct nfs4_file_layout_dsaddr * -nfs4_fl_find_get_deviceid(struct nfs_client *, struct nfs4_deviceid *dev_id); extern void nfs4_fl_put_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); diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index 42e3266..eda4527 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -37,30 +37,6 @@ #define NFSDBG_FACILITY NFSDBG_PNFS_LD /* - * Device ID RCU cache. A device ID is unique per client ID and layout type. - */ -#define NFS4_FL_DEVICE_ID_HASH_BITS 5 -#define NFS4_FL_DEVICE_ID_HASH_SIZE (1 << NFS4_FL_DEVICE_ID_HASH_BITS) -#define NFS4_FL_DEVICE_ID_HASH_MASK (NFS4_FL_DEVICE_ID_HASH_SIZE - 1) - -static inline u32 -nfs4_fl_deviceid_hash(struct nfs4_deviceid *id) -{ - unsigned char *cptr = (unsigned char *)id->data; - unsigned int nbytes = NFS4_DEVICEID4_SIZE; - u32 x = 0; - - while (nbytes--) { - x *= 37; - x += *cptr++; - } - return x & NFS4_FL_DEVICE_ID_HASH_MASK; -} - -static struct hlist_head filelayout_deviceid_cache[NFS4_FL_DEVICE_ID_HASH_SIZE]; -static DEFINE_SPINLOCK(filelayout_deviceid_lock); - -/* * Data server cache * * Data servers can be mapped to different device ids. @@ -89,27 +65,6 @@ print_ds(struct nfs4_pnfs_ds *ds) ds->ds_clp ? ds->ds_clp->cl_exchange_flags : 0); } -void -print_ds_list(struct nfs4_file_layout_dsaddr *dsaddr) -{ - int i; - - ifdebug(FACILITY) { - printk("%s dsaddr->ds_num %d\n", __func__, - dsaddr->ds_num); - for (i = 0; i < dsaddr->ds_num; i++) - print_ds(dsaddr->ds_list[i]); - } -} - -void print_deviceid(struct nfs4_deviceid *id) -{ - u32 *p = (u32 *)id; - - dprintk("%s: device id= [%x%x%x%x]\n", __func__, - p[0], p[1], p[2], p[3]); -} - /* nfs4_ds_cache_lock is held */ static struct nfs4_pnfs_ds * _data_server_lookup_locked(u32 ip_addr, u32 port) @@ -207,7 +162,7 @@ nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr) struct nfs4_pnfs_ds *ds; int i; - print_deviceid(&dsaddr->deviceid); + nfs4_print_deviceid(&dsaddr->id_node.deviceid); for (i = 0; i < dsaddr->ds_num; i++) { ds = dsaddr->ds_list[i]; @@ -431,8 +386,8 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags) dsaddr->stripe_indices = stripe_indices; stripe_indices = NULL; dsaddr->ds_num = num; - dsaddr->nfs_client = NFS_SERVER(ino)->nfs_client; - memcpy(&dsaddr->deviceid, &pdev->dev_id, sizeof(pdev->dev_id)); + nfs4_init_deviceid_node(&dsaddr->id_node, NFS_SERVER(ino)->nfs_client, + &pdev->dev_id); for (i = 0; i < dsaddr->ds_num; i++) { int j; @@ -505,8 +460,8 @@ out_err: static struct nfs4_file_layout_dsaddr * decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_flags) { - struct nfs4_file_layout_dsaddr *d, *new; - long hash; + struct nfs4_deviceid_node *d; + struct nfs4_file_layout_dsaddr *n, *new; new = decode_device(inode, dev, gfp_flags); if (!new) { @@ -515,20 +470,13 @@ decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_fl return NULL; } - spin_lock(&filelayout_deviceid_lock); - d = nfs4_fl_find_get_deviceid(new->nfs_client, &new->deviceid); - if (d) { - spin_unlock(&filelayout_deviceid_lock); + d = nfs4_insert_deviceid_node(&new->id_node); + n = container_of(d, struct nfs4_file_layout_dsaddr, id_node); + if (n != new) { nfs4_fl_free_deviceid(new); - return d; + return n; } - INIT_HLIST_NODE(&new->node); - atomic_set(&new->ref, 1); - hash = nfs4_fl_deviceid_hash(&new->deviceid); - hlist_add_head_rcu(&new->node, &filelayout_deviceid_cache[hash]); - spin_unlock(&filelayout_deviceid_lock); - return new; } @@ -600,34 +548,8 @@ out_free: void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr) { - if (atomic_dec_and_lock(&dsaddr->ref, &filelayout_deviceid_lock)) { - hlist_del_rcu(&dsaddr->node); - spin_unlock(&filelayout_deviceid_lock); - - synchronize_rcu(); + if (nfs4_put_deviceid_node(&dsaddr->id_node)) nfs4_fl_free_deviceid(dsaddr); - } -} - -struct nfs4_file_layout_dsaddr * -nfs4_fl_find_get_deviceid(struct nfs_client *clp, struct nfs4_deviceid *id) -{ - struct nfs4_file_layout_dsaddr *d; - struct hlist_node *n; - long hash = nfs4_fl_deviceid_hash(id); - - rcu_read_lock(); - hlist_for_each_entry_rcu(d, n, &filelayout_deviceid_cache[hash], node) { - if (d->nfs_client == clp && !memcmp(&d->deviceid, id, sizeof(*id))) { - if (!atomic_inc_not_zero(&d->ref)) - goto fail; - rcu_read_unlock(); - return d; - } - } -fail: - rcu_read_unlock(); - return NULL; } /* @@ -675,15 +597,15 @@ static void filelayout_mark_devid_negative(struct nfs4_file_layout_dsaddr *dsaddr, int err, u32 ds_addr) { - u32 *p = (u32 *)&dsaddr->deviceid; + u32 *p = (u32 *)&dsaddr->id_node.deviceid; printk(KERN_ERR "NFS: data server %x connection error %d." " Deviceid [%x%x%x%x] marked out of use.\n", ds_addr, err, p[0], p[1], p[2], p[3]); - spin_lock(&filelayout_deviceid_lock); + spin_lock(&nfs4_ds_cache_lock); dsaddr->flags |= NFS4_DEVICE_ID_NEG_ENTRY; - spin_unlock(&filelayout_deviceid_lock); + spin_unlock(&nfs4_ds_cache_lock); } struct nfs4_pnfs_ds * diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 720bb9d..3831ad0 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -157,6 +157,23 @@ bool pnfs_roc_drain(struct inode *ino, u32 *barrier); void pnfs_set_layoutcommit(struct nfs_write_data *wdata); int pnfs_layoutcommit_inode(struct inode *inode, bool sync); +/* pnfs_dev.c */ +struct nfs4_deviceid_node { + struct hlist_node node; + const struct nfs_client *nfs_client; + struct nfs4_deviceid deviceid; + atomic_t ref; +}; + +void nfs4_print_deviceid(const struct nfs4_deviceid *dev_id); +struct nfs4_deviceid_node *nfs4_find_get_deviceid(const struct nfs_client *, const struct nfs4_deviceid *); +struct nfs4_deviceid_node *nfs4_unhash_put_deviceid(const struct nfs_client *, const struct nfs4_deviceid *); +void nfs4_init_deviceid_node(struct nfs4_deviceid_node *, + const struct nfs_client *, + 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 *); + static inline int lo_fail_bit(u32 iomode) { return iomode == IOMODE_RW ? diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c new file mode 100644 index 0000000..bf05189 --- /dev/null +++ b/fs/nfs/pnfs_dev.c @@ -0,0 +1,156 @@ +/* + * Device operations for the pnfs client. + * + * Copyright (c) 2002 + * The Regents of the University of Michigan + * All Rights Reserved + * + * Dean Hildebrand + * Garth Goodson + * + * Permission is granted to use, copy, create derivative works, and + * redistribute this software and such derivative works for any purpose, + * so long as the name of the University of Michigan is not used in + * any advertising or publicity pertaining to the use or distribution + * of this software without specific, written prior authorization. If + * the above copyright notice or any other identification of the + * University of Michigan is included in any copy of any portion of + * this software, then the disclaimer below must also be included. + * + * This software is provided as is, without representation or warranty + * of any kind either express or implied, including without limitation + * the implied warranties of merchantability, fitness for a particular + * purpose, or noninfringement. The Regents of the University of + * Michigan shall not be liable for any damages, including special, + * indirect, incidental, or consequential damages, with respect to any + * claim arising out of or in connection with the use of the software, + * even if it has been or is hereafter advised of the possibility of + * such damages. + */ + +#include "pnfs.h" + +#define NFSDBG_FACILITY NFSDBG_PNFS + +/* + * Device ID RCU cache. A device ID is unique per server and layout type. + */ +#define NFS4_DEVICE_ID_HASH_BITS 5 +#define NFS4_DEVICE_ID_HASH_SIZE (1 << NFS4_DEVICE_ID_HASH_BITS) +#define NFS4_DEVICE_ID_HASH_MASK (NFS4_DEVICE_ID_HASH_SIZE - 1) + +static struct hlist_head nfs4_deviceid_cache[NFS4_DEVICE_ID_HASH_SIZE]; +static DEFINE_SPINLOCK(nfs4_deviceid_lock); + +void +nfs4_print_deviceid(const struct nfs4_deviceid *id) +{ + u32 *p = (u32 *)id; + + dprintk("%s: device id= [%x%x%x%x]\n", __func__, + p[0], p[1], p[2], p[3]); +} +EXPORT_SYMBOL_GPL(nfs4_print_deviceid); + +static inline u32 +nfs4_deviceid_hash(const struct nfs4_deviceid *id) +{ + unsigned char *cptr = (unsigned char *)id->data; + unsigned int nbytes = NFS4_DEVICEID4_SIZE; + u32 x = 0; + + while (nbytes--) { + x *= 37; + x += *cptr++; + } + return x & NFS4_DEVICE_ID_HASH_MASK; +} + +/* + * Lookup a deviceid in cache and get a reference count on it if found + * + * @clp nfs_client associated with deviceid + * @id deviceid to look up + */ +struct nfs4_deviceid_node * +nfs4_find_get_deviceid(const struct nfs_client *clp, const struct nfs4_deviceid *id) +{ + struct nfs4_deviceid_node *d; + struct hlist_node *n; + long hash = nfs4_deviceid_hash(id); + + rcu_read_lock(); + hlist_for_each_entry_rcu(d, n, &nfs4_deviceid_cache[hash], node) { + if (d->nfs_client == clp && !memcmp(&d->deviceid, id, sizeof(*id))) { + if (!atomic_inc_not_zero(&d->ref)) + goto fail; + rcu_read_unlock(); + return d; + } + } +fail: + rcu_read_unlock(); + return NULL; +} +EXPORT_SYMBOL_GPL(nfs4_find_get_deviceid); + +void +nfs4_init_deviceid_node(struct nfs4_deviceid_node *d, + const struct nfs_client *nfs_client, + const struct nfs4_deviceid *id) +{ + d->nfs_client = nfs_client; + d->deviceid = *id; +} +EXPORT_SYMBOL_GPL(nfs4_init_deviceid_node); + +/* + * Uniquely initialize and insert a deviceid node into cache + * + * @new new deviceid node + * Note that the caller must set up new->nfs_client and new->deviceid + * + * @ret the inserted node, if none found, otherwise, the found entry. + */ +struct nfs4_deviceid_node * +nfs4_insert_deviceid_node(struct nfs4_deviceid_node *new) +{ + struct nfs4_deviceid_node *d; + long hash; + + spin_lock(&nfs4_deviceid_lock); + d = nfs4_find_get_deviceid(new->nfs_client, &new->deviceid); + if (d) { + spin_unlock(&nfs4_deviceid_lock); + return d; + } + + INIT_HLIST_NODE(&new->node); + atomic_set(&new->ref, 1); + hash = nfs4_deviceid_hash(&new->deviceid); + hlist_add_head_rcu(&new->node, &nfs4_deviceid_cache[hash]); + spin_unlock(&nfs4_deviceid_lock); + + return new; +} +EXPORT_SYMBOL_GPL(nfs4_insert_deviceid_node); + +/* + * Dereference a deviceid node and delete it when its reference count drops + * to zero. + * + * @d deviceid node to put + * + * @ret true iff the node was deleted + */ +bool +nfs4_put_deviceid_node(struct nfs4_deviceid_node *d) +{ + if (!atomic_dec_and_lock(&d->ref, &nfs4_deviceid_lock)) + return false; + hlist_del_init_rcu(&d->node); + spin_unlock(&nfs4_deviceid_lock); + synchronize_rcu(); + return true; +} +EXPORT_SYMBOL_GPL(nfs4_put_deviceid_node);