From: NeilBrown Whitespace cleanup, fix one dprintk, remove superfluous casts of NULL. Signed-off-by: J. Bruce Fields Signed-off-by: Neil Brown Signed-off-by: Andrew Morton --- 25-akpm/fs/nfsd/nfs4state.c | 72 ++++++++++++++++++++++---------------------- 1 files changed, 36 insertions(+), 36 deletions(-) diff -puN fs/nfsd/nfs4state.c~knfsd-trivial-cleanup-of-nfs4statec fs/nfsd/nfs4state.c --- 25/fs/nfsd/nfs4state.c~knfsd-trivial-cleanup-of-nfs4statec 2004-09-02 20:28:54.366427264 -0700 +++ 25-akpm/fs/nfsd/nfs4state.c 2004-09-02 20:28:54.373426200 -0700 @@ -236,7 +236,7 @@ static struct nfs4_client * create_client(struct xdr_netobj name) { struct nfs4_client *clp; - if(!(clp = alloc_client(name))) + if (!(clp = alloc_client(name))) goto out; INIT_LIST_HEAD(&clp->cl_idhash); INIT_LIST_HEAD(&clp->cl_strhash); @@ -268,7 +268,7 @@ copy_cred(struct svc_cred *target, struc static int cmp_name(struct xdr_netobj *n1, struct xdr_netobj *n2) { - if(!n1 || !n2) + if (!n1 || !n2) return 0; return((n1->len == n2->len) && !memcmp(n1->data, n2->data, n2->len)); } @@ -399,7 +399,7 @@ parse_ipv4(unsigned int addr_len, char * return 0; } cbaddr |= (temp << shift); - if(shift > 0) + if (shift > 0) shift -= 8; } *cbaddrp = cbaddr; @@ -411,7 +411,7 @@ parse_ipv4(unsigned int addr_len, char * return 0; } cbport |= (temp << shift); - if(shift > 0) + if (shift > 0) shift -= 8; } *cbportp = cbport; @@ -423,7 +423,7 @@ gen_callback(struct nfs4_client *clp, st { struct nfs4_callback *cb = &clp->cl_callback; - if( !(parse_ipv4(se->se_callback_addr_len, se->se_callback_addr_val, + if ( !(parse_ipv4(se->se_callback_addr_len, se->se_callback_addr_val, &cb->cb_addr, &cb->cb_port))) { printk(KERN_INFO "NFSD: BAD callback address. client will not receive delegations\n"); cb->cb_parsed = 0; @@ -431,9 +431,9 @@ gen_callback(struct nfs4_client *clp, st } cb->cb_netid.len = se->se_callback_netid_len; cb->cb_netid.data = se->se_callback_netid_val; - cb->cb_prog = se->se_callback_prog; - cb->cb_ident = se->se_callback_ident; - cb->cb_parsed = 1; + cb->cb_prog = se->se_callback_prog; + cb->cb_ident = se->se_callback_ident; + cb->cb_parsed = 1; } /* @@ -812,7 +812,7 @@ alloc_init_file(unsigned int hashval, st alloc_file++; return fp; } - return (struct nfs4_file *)NULL; + return NULL; } static void @@ -825,7 +825,7 @@ release_all_files(void) while (!list_empty(&file_hashtbl[i])) { fp = list_entry(file_hashtbl[i].next, struct nfs4_file, fi_hash); /* this should never be more than once... */ - if(!list_empty(&fp->fi_perfile)) { + if (!list_empty(&fp->fi_perfile)) { printk("ERROR: release_all_files: file %p is open, creating dangling state !!!\n",fp); } release_file(fp); @@ -839,20 +839,20 @@ alloc_stateowner(struct xdr_netobj *owne struct nfs4_stateowner *sop; if ((sop = kmalloc(sizeof(struct nfs4_stateowner),GFP_KERNEL))) { - if((sop->so_owner.data = kmalloc(owner->len, GFP_KERNEL))) { + if ((sop->so_owner.data = kmalloc(owner->len, GFP_KERNEL))) { memcpy(sop->so_owner.data, owner->data, owner->len); sop->so_owner.len = owner->len; return sop; } kfree(sop); } - return (struct nfs4_stateowner *)NULL; + return NULL; } /* should use a slab cache */ static void free_stateowner(struct nfs4_stateowner *sop) { - if(sop) { + if (sop) { kfree(sop->so_owner.data); kfree(sop); sop = NULL; @@ -867,7 +867,7 @@ alloc_init_open_stateowner(unsigned int unsigned int idhashval; if (!(sop = alloc_stateowner(&open->op_owner))) - return (struct nfs4_stateowner *)NULL; + return NULL; idhashval = ownerid_hashval(current_ownerid); INIT_LIST_HEAD(&sop->so_idhash); INIT_LIST_HEAD(&sop->so_strhash); @@ -921,7 +921,7 @@ release_stateowner(struct nfs4_stateowne while (!list_empty(&sop->so_perfilestate)) { stp = list_entry(sop->so_perfilestate.next, struct nfs4_stateid, st_perfilestate); - if(sop->so_is_open_owner) + if (sop->so_is_open_owner) release_stateid(stp, OPEN_STATE); else release_stateid(stp, LOCK_STATE); @@ -960,7 +960,7 @@ release_stateid(struct nfs4_stateid *stp list_del_perfile++; list_del(&stp->st_perfile); list_del(&stp->st_perfilestate); - if((stp->st_vfs_set) && (flags & OPEN_STATE)) { + if ((stp->st_vfs_set) && (flags & OPEN_STATE)) { release_stateid_lockowner(stp); nfsd_close(stp->st_vfs_file); vfsclose++; @@ -991,8 +991,8 @@ move_to_close_lru(struct nfs4_stateowner list_del_init(&sop->so_strhash); list_del_init(&sop->so_perlockowner); - list_add_tail(&sop->so_close_lru, &close_lru); - sop->so_time = get_seconds(); + list_add_tail(&sop->so_close_lru, &close_lru); + sop->so_time = get_seconds(); } void @@ -1030,7 +1030,7 @@ find_openstateowner_str(unsigned int has struct nfs4_stateowner *local = NULL; list_for_each_entry(local, &ownerstr_hashtbl[hashval], so_strhash) { - if(!cmp_owner_str(local, &open->op_owner, &open->op_clientid)) + if (!cmp_owner_str(local, &open->op_owner, &open->op_clientid)) continue; *op = local; return(1); @@ -1078,7 +1078,7 @@ set_access(unsigned int *access, unsigne *access = 0; for (i = 1; i < 4; i++) { - if(test_bit(i, &bmap)) + if (test_bit(i, &bmap)) *access |= i; } } @@ -1089,7 +1089,7 @@ set_deny(unsigned int *deny, unsigned lo *deny = 0; for (i = 0; i < 4; i++) { - if(test_bit(i, &bmap)) + if (test_bit(i, &bmap)) *deny |= i ; } } @@ -1278,7 +1278,7 @@ nfsd4_process_open2(struct svc_rqst *rqs /* Search for conflicting share reservations */ status = nfserr_share_denied; list_for_each_entry(stq, &fp->fi_perfile, st_perfile) { - if(stq->st_stateowner == sop) { + if (stq->st_stateowner == sop) { stp = stq; continue; } @@ -1484,7 +1484,7 @@ find_openstateowner_id(u32 st_id, int fl dprintk("NFSD: find_openstateowner_id %d\n", st_id); if (flags & CLOSE_STATE) { list_for_each_entry(local, &close_lru, so_close_lru) { - if(local->so_id == st_id) + if (local->so_id == st_id) return local; } } @@ -1888,16 +1888,16 @@ find_stateid(stateid_t *stid, int flags) if ((flags & LOCK_STATE) || (flags & RDWR_STATE)) { hashval = stateid_hashval(st_id, f_id); list_for_each_entry(local, &lockstateid_hashtbl[hashval], st_hash) { - if((local->st_stateid.si_stateownerid == st_id) && - (local->st_stateid.si_fileid == f_id)) + if ((local->st_stateid.si_stateownerid == st_id) && + (local->st_stateid.si_fileid == f_id)) return local; } } if ((flags & OPEN_STATE) || (flags & RDWR_STATE)) { hashval = stateid_hashval(st_id, f_id); list_for_each_entry(local, &stateid_hashtbl[hashval], st_hash) { - if((local->st_stateid.si_stateownerid == st_id) && - (local->st_stateid.si_fileid == f_id)) + if ((local->st_stateid.si_stateownerid == st_id) && + (local->st_stateid.si_fileid == f_id)) return local; } } else @@ -1967,7 +1967,7 @@ find_lockstateowner(struct xdr_netobj *o for (i = 0; i < LOCK_HASH_SIZE; i++) { list_for_each_entry(local, &lock_ownerid_hashtbl[i], so_idhash) { - if(!cmp_owner_str(local, owner, clid)) + if (!cmp_owner_str(local, owner, clid)) continue; return local; } @@ -1980,7 +1980,7 @@ find_lockstateowner_str(unsigned int has struct nfs4_stateowner *local = NULL; list_for_each_entry(local, &lock_ownerstr_hashtbl[hashval], so_strhash) { - if(!cmp_owner_str(local, owner, clid)) + if (!cmp_owner_str(local, owner, clid)) continue; *op = local; return(1); @@ -2005,7 +2005,7 @@ alloc_init_lock_stateowner(unsigned int unsigned int idhashval; if (!(sop = alloc_stateowner(&lock->lk_new_owner))) - return (struct nfs4_stateowner *)NULL; + return NULL; idhashval = lockownerid_hashval(current_ownerid); INIT_LIST_HEAD(&sop->so_idhash); INIT_LIST_HEAD(&sop->so_strhash); @@ -2068,7 +2068,7 @@ int check_lock_length(u64 offset, u64 length) { return ((length == 0) || ((length != ~(u64)0) && - LOFF_OVERFLOW(offset, length))); + LOFF_OVERFLOW(offset, length))); } /* @@ -2210,7 +2210,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struc status = posix_lock_file(filp, &file_lock); if (file_lock.fl_ops && file_lock.fl_ops->fl_release_private) file_lock.fl_ops->fl_release_private(&file_lock); - dprintk("NFSD: nfsd4_lock: posix_test_lock passed. posix_lock_file status %d\n",status); + dprintk("NFSD: nfsd4_lock: posix_lock_file status %d\n",status); switch (-status) { case 0: /* success! */ update_stateid(&lock_stp->st_stateid); @@ -2459,7 +2459,7 @@ nfsd4_release_lockowner(struct svc_rqst nfs4_lock_state(); - status = nfs_ok; + status = nfs_ok; local = find_lockstateowner(owner, clid); if (local) { struct nfs4_stateid *stp; @@ -2469,7 +2469,7 @@ nfsd4_release_lockowner(struct svc_rqst status = nfserr_locks_held; list_for_each_entry(stp, &local->so_perfilestate, st_perfilestate) { - if(stp->st_vfs_set) { + if (stp->st_vfs_set) { if (check_for_locks(stp->st_vfs_file, local)) goto out; } @@ -2496,7 +2496,7 @@ alloc_reclaim(int namelen) kfree(crp); return NULL; } - return crp; + return crp; } /* @@ -2566,7 +2566,7 @@ nfs4_find_reclaim_client(clientid_t *cli strhashval = clientstr_hashval(client->cl_name.data, client->cl_name.len); list_for_each_entry(crp, &reclaim_str_hashtbl[strhashval], cr_strhash) { - if(cmp_name(&crp->cr_name, &client->cl_name)) { + if (cmp_name(&crp->cr_name, &client->cl_name)) { return crp; } } _