X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fllite_internal.h;h=deee1ef2575d78bb257ee6e952ea364cb6d7a49a;hp=4803907d05055827dc20ed576f200c66a02b5143;hb=3019471086f008f45b045db424f5351fc271a1be;hpb=7796bdfb07954422e81b124ba176a5e431b43d8b diff --git a/lustre/llite/llite_internal.h b/lustre/llite/llite_internal.h index 4803907..deee1ef 100644 --- a/lustre/llite/llite_internal.h +++ b/lustre/llite/llite_internal.h @@ -1,6 +1,4 @@ -/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- - * vim:expandtab:shiftwidth=8:tabstop=8: - * +/* * GPL HEADER START * * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. @@ -28,9 +26,8 @@ /* * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. - */ -/* - * Copyright (c) 2011 Whamcloud, Inc. + * + * Copyright (c) 2011, 2013, Intel Corporation. */ /* * This file is part of Lustre, http://www.lustre.org/ @@ -39,9 +36,6 @@ #ifndef LLITE_INTERNAL_H #define LLITE_INTERNAL_H - -#include - #include #include #include /* for s2sbi */ @@ -52,6 +46,7 @@ #include #include #include +#include #ifndef FMODE_EXEC #define FMODE_EXEC 0 @@ -61,24 +56,28 @@ #define VM_FAULT_RETRY 0 #endif +/* Kernel 3.1 kills LOOKUP_CONTINUE, LOOKUP_PARENT is equivalent to it. + * seem kernel commit 49084c3bb2055c401f3493c13edae14d49128ca0 */ +#ifndef LOOKUP_CONTINUE +#define LOOKUP_CONTINUE LOOKUP_PARENT +#endif + /** Only used on client-side for indicating the tail of dir hash/offset. */ #define LL_DIR_END_OFF 0x7fffffffffffffffULL #define LL_DIR_END_OFF_32BIT 0x7fffffffUL -#ifndef DCACHE_LUSTRE_INVALID -#define DCACHE_LUSTRE_INVALID 0x4000000 -#endif - #define LL_IT2STR(it) ((it) ? ldlm_it2str((it)->it_op) : "0") #define LUSTRE_FPRIVATE(file) ((file)->private_data) struct ll_dentry_data { - int lld_cwd_count; - int lld_mnt_count; - struct obd_client_handle lld_cwd_och; - struct obd_client_handle lld_mnt_och; - struct lookup_intent *lld_it; - unsigned int lld_sa_generation; + int lld_cwd_count; + int lld_mnt_count; + struct obd_client_handle lld_cwd_och; + struct obd_client_handle lld_mnt_och; + struct lookup_intent *lld_it; + unsigned int lld_sa_generation; + unsigned int lld_invalid:1; + struct rcu_head lld_rcu_head; }; #define ll_d2d(de) ((struct ll_dentry_data*)((de)->d_fsdata)) @@ -123,19 +122,24 @@ enum lli_flags { LLIF_CONTENDED = (1 << 4), /* Truncate uses server lock for this file */ LLIF_SRVLOCK = (1 << 5), - + /* File data is modified. */ + LLIF_DATA_MODIFIED = (1 << 6), + /* File is being restored */ + LLIF_FILE_RESTORING = (1 << 7), + /* Xattr cache is attached to the file */ + LLIF_XATTR_CACHE = (1 << 8), }; struct ll_inode_info { - __u32 lli_inode_magic; - __u32 lli_flags; - __u64 lli_ioepoch; + __u32 lli_inode_magic; + __u32 lli_flags; + __u64 lli_ioepoch; - cfs_spinlock_t lli_lock; - struct posix_acl *lli_posix_acl; + spinlock_t lli_lock; + struct posix_acl *lli_posix_acl; - cfs_hlist_head_t *lli_remote_perms; - cfs_semaphore_t lli_rmtperm_sem; + cfs_hlist_head_t *lli_remote_perms; + struct mutex lli_rmtperm_mutex; /* identifying fields for both metadata and data stacks. */ struct lu_fid lli_fid; @@ -164,22 +168,22 @@ struct ll_inode_info { __u64 lli_open_fd_read_count; __u64 lli_open_fd_write_count; __u64 lli_open_fd_exec_count; - /* Protects access to och pointers and their usage counters, also - * atomicity of check-update of lli_smd */ - cfs_semaphore_t lli_och_sem; + /* Protects access to och pointers and their usage counters */ + struct mutex lli_och_mutex; - struct inode lli_vfs_inode; + struct inode lli_vfs_inode; - /* the most recent timestamps obtained from mds */ - struct ost_lvb lli_lvb; + /* the most recent timestamps obtained from mds */ + struct ost_lvb lli_lvb; + spinlock_t lli_agl_lock; - /* Try to make the d::member and f::member are aligned. Before using - * these members, make clear whether it is directory or not. */ - union { - /* for directory */ - struct { - /* serialize normal readdir and statahead-readdir. */ - cfs_semaphore_t d_readdir_sem; + /* Try to make the d::member and f::member are aligned. Before using + * these members, make clear whether it is directory or not. */ + union { + /* for directory */ + struct { + /* serialize normal readdir and statahead-readdir. */ + struct mutex d_readdir_mutex; /* metadata statahead */ /* since parent-child threads can share the same @file @@ -188,42 +192,62 @@ struct ll_inode_info { * cleanup the dir readahead. */ void *d_opendir_key; struct ll_statahead_info *d_sai; - __u64 d_sa_pos; struct posix_acl *d_def_acl; /* protect statahead stuff. */ - cfs_spinlock_t d_sa_lock; - /* "opendir_pid" is the token when lookup/revalid - * -- I am the owner of dir statahead. */ - pid_t d_opendir_pid; - } d; - -#define lli_readdir_sem u.d.d_readdir_sem + spinlock_t d_sa_lock; + /* "opendir_pid" is the token when lookup/revalid + * -- I am the owner of dir statahead. */ + pid_t d_opendir_pid; + /* directory stripe information */ + struct lmv_stripe_md *d_lmv_md; + } d; + +#define lli_readdir_mutex u.d.d_readdir_mutex #define lli_opendir_key u.d.d_opendir_key #define lli_sai u.d.d_sai -#define lli_sa_pos u.d.d_sa_pos #define lli_def_acl u.d.d_def_acl #define lli_sa_lock u.d.d_sa_lock #define lli_opendir_pid u.d.d_opendir_pid - - /* for non-directory */ - struct { - cfs_semaphore_t f_size_sem; - void *f_size_sem_owner; - char *f_symlink_name; - __u64 f_maxbytes; - /* - * cfs_rw_semaphore_t { - * signed long count; // align u.d.d_def_acl - * cfs_spinlock_t wait_lock; // align u.d.d_sa_lock - * struct list_head wait_list; - * } - */ - cfs_rw_semaphore_t f_trunc_sem; - cfs_semaphore_t f_write_sem; - - /* for writepage() only to communicate to fsync */ - int f_async_rc; - int f_write_rc; +#define lli_lmv_md u.d.d_lmv_md + + /* for non-directory */ + struct { + struct semaphore f_size_sem; + void *f_size_sem_owner; + char *f_symlink_name; + __u64 f_maxbytes; + /* + * struct rw_semaphore { + * signed long count; // align d.d_def_acl + * spinlock_t wait_lock; // align d.d_sa_lock + * struct list_head wait_list; + * } + */ + struct rw_semaphore f_trunc_sem; + struct mutex f_write_mutex; + + struct rw_semaphore f_glimpse_sem; + cfs_time_t f_glimpse_time; + cfs_list_t f_agl_list; + __u64 f_agl_index; + + /* for writepage() only to communicate to fsync */ + int f_async_rc; + + /* volatile file criteria is based on file name, this + * flag is used to keep the test result, so the strcmp + * is done only once + */ + bool f_volatile; + /* + * whenever a process try to read/write the file, the + * jobid of the process will be saved here, and it'll + * be packed into the write PRC when flush later. + * + * so the read/write statistics for jobid will not be + * accurate if the file is shared by different jobs. + */ + char f_jobid[JOBSTATS_JOBID_SIZE]; } f; #define lli_size_sem u.f.f_size_sem @@ -231,11 +255,16 @@ struct ll_inode_info { #define lli_symlink_name u.f.f_symlink_name #define lli_maxbytes u.f.f_maxbytes #define lli_trunc_sem u.f.f_trunc_sem -#define lli_write_sem u.f.f_write_sem -#define lli_async_rc u.f.f_async_rc -#define lli_write_rc u.f.f_write_rc +#define lli_write_mutex u.f.f_write_mutex +#define lli_glimpse_sem u.f.f_glimpse_sem +#define lli_glimpse_time u.f.f_glimpse_time +#define lli_agl_list u.f.f_agl_list +#define lli_agl_index u.f.f_agl_index +#define lli_async_rc u.f.f_async_rc +#define lli_jobid u.f.f_jobid +#define lli_volatile u.f.f_volatile - } u; + } u; /* XXX: For following frequent used members, although they maybe special * used for non-directory object, it is some time-wasting to check @@ -247,20 +276,36 @@ struct ll_inode_info { * In the future, if more members are added only for directory, * some of the following members can be moved into u.f. */ - struct lov_stripe_md *lli_smd; - struct cl_object *lli_clob; + bool lli_has_smd; + struct cl_object *lli_clob; + + /* mutex to request for layout lock exclusively. */ + struct mutex lli_layout_mutex; + /* valid only inside LAYOUT ibits lock, protected by lli_layout_mutex */ + __u32 lli_layout_gen; + + struct rw_semaphore lli_xattrs_list_rwsem; + struct mutex lli_xattrs_enq_lock; + struct list_head lli_xattrs; /* ll_xattr_entry->xe_list */ }; +int ll_xattr_cache_destroy(struct inode *inode); + +int ll_xattr_cache_get(struct inode *inode, + const char *name, + char *buffer, + size_t size, + __u64 valid); + /* * Locking to guarantee consistency of non-atomic updates to long long i_size, - * consistency between file size and KMS, and consistency within - * ->lli_smd->lsm_oinfo[]'s. + * consistency between file size and KMS. * - * Implemented by ->lli_size_sem and ->lsm_sem, nested in that order. + * Implemented by ->lli_size_sem and ->lsm_lock, nested in that order. */ -void ll_inode_size_lock(struct inode *inode, int lock_lsm); -void ll_inode_size_unlock(struct inode *inode, int unlock_lsm); +void ll_inode_size_lock(struct inode *inode); +void ll_inode_size_unlock(struct inode *inode); // FIXME: replace the name of this with LL_I to conform to kernel stuff // static inline struct ll_inode_info *LL_I(struct inode *inode) @@ -271,10 +316,10 @@ static inline struct ll_inode_info *ll_i2info(struct inode *inode) /* default to about 40meg of readahead on a given system. That much tied * up in 512k readahead requests serviced at 40ms each is about 1GB/s. */ -#define SBI_DEFAULT_READAHEAD_MAX (40UL << (20 - CFS_PAGE_SHIFT)) +#define SBI_DEFAULT_READAHEAD_MAX (40UL << (20 - PAGE_CACHE_SHIFT)) /* default to read-ahead full files smaller than 2MB on the second read */ -#define SBI_DEFAULT_READAHEAD_WHOLE_MAX (2UL << (20 - CFS_PAGE_SHIFT)) +#define SBI_DEFAULT_READAHEAD_WHOLE_MAX (2UL << (20 - PAGE_CACHE_SHIFT)) enum ra_stat { RA_STAT_HIT = 0, @@ -369,6 +414,34 @@ enum stats_track_type { #define LL_SBI_SOM_PREVIEW 0x1000 /* SOM preview mount option */ #define LL_SBI_32BIT_API 0x2000 /* generate 32 bit inodes. */ #define LL_SBI_64BIT_HASH 0x4000 /* support 64-bits dir hash/offset */ +#define LL_SBI_AGL_ENABLED 0x8000 /* enable agl */ +#define LL_SBI_VERBOSE 0x10000 /* verbose mount/umount */ +#define LL_SBI_LAYOUT_LOCK 0x20000 /* layout lock support */ +#define LL_SBI_USER_FID2PATH 0x40000 /* allow fid2path by unprivileged users */ +#define LL_SBI_XATTR_CACHE 0x80000 /* support for xattr cache */ + +#define LL_SBI_FLAGS { \ + "nolck", \ + "checksum", \ + "flock", \ + "xattr", \ + "acl", \ + "???", \ + "rmt_client", \ + "mds_capa", \ + "oss_capa", \ + "flock", \ + "lru_resize", \ + "lazy_statfs", \ + "som", \ + "32bit_api", \ + "64bit_hash", \ + "agl", \ + "verbose", \ + "layout", \ + "user_fid2path",\ + "xattr", \ +} /* default value for ll_sb_info->contention_time */ #define SBI_DEFAULT_CONTENTION_SECONDS 60 @@ -383,8 +456,8 @@ struct rmtacl_ctl_entry { }; struct rmtacl_ctl_table { - cfs_spinlock_t rct_lock; - cfs_list_t rct_entries[RCE_HASHES]; + spinlock_t rct_lock; + cfs_list_t rct_entries[RCE_HASHES]; }; #define EE_HASHES 32 @@ -398,17 +471,17 @@ struct eacl_entry { }; struct eacl_table { - cfs_spinlock_t et_lock; - cfs_list_t et_entries[EE_HASHES]; + spinlock_t et_lock; + cfs_list_t et_entries[EE_HASHES]; }; struct ll_sb_info { - cfs_list_t ll_list; - /* this protects pglist and ra_info. It isn't safe to - * grab from interrupt contexts */ - cfs_spinlock_t ll_lock; - cfs_spinlock_t ll_pp_extent_lock; /* Lock for pp_extent entries */ - cfs_spinlock_t ll_process_lock; /* Lock for ll_rw_process_info */ + cfs_list_t ll_list; + /* this protects pglist and ra_info. It isn't safe to + * grab from interrupt contexts */ + spinlock_t ll_lock; + spinlock_t ll_pp_extent_lock; /* pp_extent entry*/ + spinlock_t ll_process_lock; /* ll_rw_process_info */ struct obd_uuid ll_sb_uuid; struct obd_export *ll_md_exp; struct obd_export *ll_dt_exp; @@ -416,6 +489,8 @@ struct ll_sb_info { struct lu_fid ll_root_fid; /* root object fid */ int ll_flags; + unsigned int ll_umounting:1, + ll_xattr_cache_enabled:1; cfs_list_t ll_conn_chain; /* per-conn chain of SBs */ struct lustre_client_ocd ll_lco; @@ -424,8 +499,11 @@ struct ll_sb_info { struct lprocfs_stats *ll_stats; /* lprocfs stats counter */ - unsigned long ll_async_page_max; - unsigned long ll_async_page_count; + /* Used to track "unstable" pages on a client, and maintain a + * LRU list of clean pages. An "unstable" page is defined as + * any page which is sent to a server as part of a bulk request, + * but is uncommitted to stable storage. */ + struct cl_client_cache ll_cache; struct lprocfs_stats *ll_ra_stats; @@ -447,8 +525,8 @@ struct ll_sb_info { unsigned int ll_offset_process_count; struct ll_rw_process_info ll_rw_offset_info[LL_OFFSET_HIST_MAX]; unsigned int ll_rw_offset_entry_count; - enum stats_track_type ll_stats_track_type; int ll_stats_track_id; + enum stats_track_type ll_stats_track_type; int ll_rw_stats_on; /* metadata stat-ahead */ @@ -457,12 +535,12 @@ struct ll_sb_info { * count */ atomic_t ll_sa_wrong; /* statahead thread stopped for * low hit ratio */ + atomic_t ll_agl_total; /* AGL thread started count */ dev_t ll_sdev_orig; /* save s_dev before assign for * clustred nfs */ struct rmtacl_ctl_table ll_rct; struct eacl_table ll_et; - struct vfsmount *ll_mnt; }; #define LL_DEFAULT_MAX_RW_CHUNK (32 * 1024 * 1024) @@ -478,7 +556,7 @@ struct ll_ra_read { * per file-descriptor read-ahead data. */ struct ll_readahead_state { - cfs_spinlock_t ras_lock; + spinlock_t ras_lock; /* * index of the last page that read(2) needed and that wasn't in the * cache. Used by ras_update() to detect seeks. @@ -560,25 +638,28 @@ struct ll_readahead_state { unsigned long ras_consecutive_stride_requests; }; -struct ll_file_dir { - __u64 lfd_pos; - __u64 lfd_next; -}; - -extern cfs_mem_cache_t *ll_file_data_slab; +extern struct kmem_cache *ll_file_data_slab; struct lustre_handle; struct ll_file_data { - struct ll_readahead_state fd_ras; - int fd_omode; - struct ccc_grouplock fd_grouplock; - struct ll_file_dir fd_dir; - __u32 fd_flags; - struct file *fd_file; + struct ll_readahead_state fd_ras; + struct ccc_grouplock fd_grouplock; + __u64 lfd_pos; + __u32 fd_flags; + fmode_t fd_omode; + /* openhandle if lease exists for this file. + * Borrow lli->lli_och_mutex to protect assignment */ + struct obd_client_handle *fd_lease_och; + struct obd_client_handle *fd_och; + struct file *fd_file; + /* Indicate whether need to report failure when close. + * true: failure is known, not report again. + * false: unknown failure, should report. */ + bool fd_write_failed; }; struct lov_stripe_md; -extern cfs_spinlock_t inode_lock; +extern spinlock_t inode_lock; extern struct proc_dir_entry *proc_lustre_fs_root; @@ -601,13 +682,18 @@ static inline int ll_need_32bit_api(struct ll_sb_info *sbi) #if BITS_PER_LONG == 32 return 1; #else - return unlikely(cfs_curproc_is_32bit() || (sbi->ll_flags & LL_SBI_32BIT_API)); + return unlikely( +#ifdef CONFIG_COMPAT + is_compat_task() || +#endif + (sbi->ll_flags & LL_SBI_32BIT_API) + ); #endif } #define LLAP_MAGIC 98764321 -extern cfs_mem_cache_t *ll_async_page_slab; +extern struct kmem_cache *ll_async_page_slab; extern size_t ll_async_page_slab_size; void ll_ra_read_in(struct file *f, struct ll_ra_read *rar); @@ -620,28 +706,30 @@ int lprocfs_register_mountpoint(struct proc_dir_entry *parent, struct super_block *sb, char *osc, char *mdc); void lprocfs_unregister_mountpoint(struct ll_sb_info *sbi); void ll_stats_ops_tally(struct ll_sb_info *sbi, int op, int count); -void lprocfs_llite_init_vars(struct lprocfs_static_vars *lvars); +extern struct lprocfs_seq_vars lprocfs_llite_obd_vars[]; #else static inline int lprocfs_register_mountpoint(struct proc_dir_entry *parent, struct super_block *sb, char *osc, char *mdc){return 0;} static inline void lprocfs_unregister_mountpoint(struct ll_sb_info *sbi) {} static void ll_stats_ops_tally(struct ll_sb_info *sbi, int op, int count) {} -static void lprocfs_llite_init_vars(struct lprocfs_static_vars *lvars) -{ - memset(lvars, 0, sizeof(*lvars)); -} #endif /* llite/dir.c */ -void ll_release_page(struct page *page, int remove); extern struct file_operations ll_dir_operations; extern struct inode_operations ll_dir_inode_operations; -struct page *ll_get_dir_page(struct file *filp, struct inode *dir, __u64 hash, - struct ll_dir_chain *chain); -int ll_readdir(struct file *filp, void *cookie, filldir_t filldir); - +int ll_dir_read(struct inode *inode, struct md_op_data *op_data, + void *cookie, filldir_t filldir); int ll_get_mdt_idx(struct inode *inode); + +struct lu_dirent *ll_dir_entry_start(struct inode *dir, + struct md_op_data *op_data); + +struct lu_dirent *ll_dir_entry_next(struct inode *dir, + struct md_op_data *op_data, + struct lu_dirent *ent); +void ll_dir_entry_end(struct inode *dir, struct md_op_data *op_data, + struct lu_dirent *ent); /* llite/namei.c */ int ll_objects_destroy(struct ptlrpc_request *request, struct inode *dir); @@ -649,20 +737,19 @@ struct inode *ll_iget(struct super_block *sb, ino_t hash, struct lustre_md *lic); int ll_md_blocking_ast(struct ldlm_lock *, struct ldlm_lock_desc *, void *data, int flag); +#ifndef HAVE_IOP_ATOMIC_OPEN struct lookup_intent *ll_convert_intent(struct open_intent *oit, int lookup_flags); -int ll_lookup_it_finish(struct ptlrpc_request *request, - struct lookup_intent *it, void *data); -struct dentry *ll_find_alias(struct inode *inode, struct dentry *de); +#endif +struct dentry *ll_splice_alias(struct inode *inode, struct dentry *de); +int ll_rmdir_entry(struct inode *dir, char *name, int namelen); /* llite/rw.c */ -int ll_prepare_write(struct file *, struct page *, unsigned from, unsigned to); -int ll_commit_write(struct file *, struct page *, unsigned from, unsigned to); int ll_writepage(struct page *page, struct writeback_control *wbc); +int ll_writepages(struct address_space *, struct writeback_control *wbc); void ll_removepage(struct page *page); int ll_readpage(struct file *file, struct page *page); void ll_readahead_init(struct inode *inode, struct ll_readahead_state *ras); -void ll_truncate(struct inode *inode); int ll_file_punch(struct inode *, loff_t, int); ssize_t ll_file_lockless_io(struct file *, char *, size_t, loff_t *, int); void ll_clear_file_contended(struct inode*); @@ -670,6 +757,9 @@ int ll_sync_page_range(struct inode *, struct address_space *, loff_t, size_t); int ll_readahead(const struct lu_env *env, struct cl_io *io, struct ll_readahead_state *ras, struct address_space *mapping, struct cl_page_list *queue, int flags); +int vvp_io_write_commit(const struct lu_env *env, struct cl_io *io); +struct ll_cl_context *ll_cl_init(struct file *file, struct page *vmpage); +void ll_cl_fini(struct ll_cl_context *lcc); /* llite/file.c */ extern struct file_operations ll_file_operations; @@ -681,10 +771,15 @@ extern int ll_inode_revalidate_it(struct dentry *, struct lookup_intent *, extern int ll_have_md_lock(struct inode *inode, __u64 *bits, ldlm_mode_t l_req_mode); extern ldlm_mode_t ll_take_md_lock(struct inode *inode, __u64 bits, - struct lustre_handle *lockh); + struct lustre_handle *lockh, __u64 flags, + ldlm_mode_t mode); int __ll_inode_revalidate_it(struct dentry *, struct lookup_intent *, __u64 bits); +#ifdef HAVE_IOP_ATOMIC_OPEN +int ll_revalidate_nd(struct dentry *dentry, unsigned int flags); +#else int ll_revalidate_nd(struct dentry *dentry, struct nameidata *nd); +#endif int ll_file_open(struct inode *inode, struct file *file); int ll_file_release(struct inode *inode, struct file *file); int ll_glimpse_ioctl(struct ll_sb_info *sbi, @@ -696,14 +791,14 @@ int ll_local_open(struct file *file, int ll_release_openhandle(struct dentry *, struct lookup_intent *); int ll_md_close(struct obd_export *md_exp, struct inode *inode, struct file *file); -int ll_md_real_close(struct inode *inode, int flags); +int ll_md_real_close(struct inode *inode, fmode_t fmode); void ll_ioepoch_close(struct inode *inode, struct md_op_data *op_data, struct obd_client_handle **och, unsigned long flags); void ll_done_writing_attr(struct inode *inode, struct md_op_data *op_data); int ll_som_update(struct inode *inode, struct md_op_data *op_data); int ll_inode_getattr(struct inode *inode, struct obdo *obdo, __u64 ioepoch, int sync); -int ll_md_setattr(struct inode *inode, struct md_op_data *op_data, +int ll_md_setattr(struct dentry *dentry, struct md_op_data *op_data, struct md_open_data **mod); void ll_pack_inode2opdata(struct inode *inode, struct md_op_data *op_data, struct lustre_handle *fh); @@ -714,6 +809,8 @@ int ll_getattr_it(struct vfsmount *mnt, struct dentry *de, struct lookup_intent *it, struct kstat *stat); int ll_getattr(struct vfsmount *mnt, struct dentry *de, struct kstat *stat); struct ll_file_data *ll_file_data_get(void); +struct posix_acl * ll_get_acl(struct inode *inode, int type); + #ifdef HAVE_GENERIC_PERMISSION_4ARGS int ll_inode_permission(struct inode *inode, int mask, unsigned int flags); #else @@ -723,35 +820,47 @@ int ll_inode_permission(struct inode *inode, int mask, struct nameidata *nd); int ll_inode_permission(struct inode *inode, int mask); # endif #endif + int ll_lov_setstripe_ea_info(struct inode *inode, struct file *file, - int flags, struct lov_user_md *lum, - int lum_size); + __u64 flags, struct lov_user_md *lum, + int lum_size); int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename, struct lov_mds_md **lmm, int *lmm_size, struct ptlrpc_request **request); int ll_dir_setstripe(struct inode *inode, struct lov_user_md *lump, int set_default); -int ll_dir_getstripe(struct inode *inode, struct lov_mds_md **lmm, +int ll_dir_getstripe(struct inode *inode, struct lov_mds_md **lmmp, int *lmm_size, struct ptlrpc_request **request); +#ifdef HAVE_FILE_FSYNC_4ARGS +int ll_fsync(struct file *file, loff_t start, loff_t end, int data); +#elif defined(HAVE_FILE_FSYNC_2ARGS) +int ll_fsync(struct file *file, int data); +#else int ll_fsync(struct file *file, struct dentry *dentry, int data); +#endif int ll_do_fiemap(struct inode *inode, struct ll_user_fiemap *fiemap, int num_bytes); -int ll_merge_lvb(struct inode *inode); +int ll_merge_lvb(const struct lu_env *env, struct inode *inode); int ll_get_grouplock(struct inode *inode, struct file *file, unsigned long arg); int ll_put_grouplock(struct inode *inode, struct file *file, unsigned long arg); -int ll_fid2path(struct obd_export *exp, void *arg); +int ll_fid2path(struct inode *inode, void *arg); +int ll_data_version(struct inode *inode, __u64 *data_version, int flags); +int ll_hsm_release(struct inode *inode); + +struct obd_client_handle *ll_lease_open(struct inode *inode, struct file *file, + fmode_t mode, __u64 flags); +int ll_lease_close(struct obd_client_handle *och, struct inode *inode, + bool *lease_broken); /* llite/dcache.c */ -int ll_dops_init(struct dentry *de, int block, int init_sa); -extern cfs_spinlock_t ll_lookup_lock; + +int ll_d_init(struct dentry *de); extern struct dentry_operations ll_d_ops; void ll_intent_drop_lock(struct lookup_intent *); void ll_intent_release(struct lookup_intent *); -int ll_drop_dentry(struct dentry *dentry); -void ll_unhash_aliases(struct inode *); +void ll_invalidate_aliases(struct inode *); void ll_frob_intent(struct lookup_intent **itp, struct lookup_intent *deft); void ll_lookup_finish_locks(struct lookup_intent *it, struct dentry *dentry); -int ll_dcompare(struct dentry *parent, struct qstr *d_name, struct qstr *name); int ll_revalidate_it_finish(struct ptlrpc_request *request, struct lookup_intent *it, struct dentry *de); @@ -763,15 +872,12 @@ void ll_lli_init(struct ll_inode_info *lli); int ll_fill_super(struct super_block *sb, struct vfsmount *mnt); void ll_put_super(struct super_block *sb); void ll_kill_super(struct super_block *sb); +struct inode *ll_inode_from_resource_lock(struct ldlm_lock *lock); struct inode *ll_inode_from_lock(struct ldlm_lock *lock); void ll_clear_inode(struct inode *inode); -int ll_setattr_raw(struct inode *inode, struct iattr *attr); +int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import); int ll_setattr(struct dentry *de, struct iattr *attr); -#ifndef HAVE_STATFS_DENTRY_PARAM -int ll_statfs(struct super_block *sb, struct kstatfs *sfs); -#else int ll_statfs(struct dentry *de, struct kstatfs *sfs); -#endif int ll_statfs_internal(struct super_block *sb, struct obd_statfs *osfs, __u64 max_age, __u32 flags); void ll_update_inode(struct inode *inode, struct lustre_md *md); @@ -780,15 +886,16 @@ void ll_delete_inode(struct inode *inode); int ll_iocontrol(struct inode *inode, struct file *file, unsigned int cmd, unsigned long arg); int ll_flush_ctx(struct inode *inode); -#ifdef HAVE_UMOUNTBEGIN_VFSMOUNT -void ll_umount_begin(struct vfsmount *vfsmnt, int flags); -#else void ll_umount_begin(struct super_block *sb); -#endif int ll_remount_fs(struct super_block *sb, int *flags, char *data); +#ifdef HAVE_SUPEROPS_USE_DENTRY +int ll_show_options(struct seq_file *seq, struct dentry *dentry); +#else int ll_show_options(struct seq_file *seq, struct vfsmount *vfs); +#endif +void ll_dirty_page_discard_warn(struct page *page, int ioret); int ll_prep_inode(struct inode **inode, struct ptlrpc_request *req, - struct super_block *); + struct super_block *, struct lookup_intent *); void lustre_dump_dentry(struct dentry *, int recur); void lustre_dump_inode(struct inode *); int ll_obd_statfs(struct inode *inode, void *arg); @@ -799,10 +906,14 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data, const char *name, int namelen, int mode, __u32 opc, void *data); void ll_finish_md_op_data(struct md_op_data *op_data); +int ll_get_obd_name(struct inode *inode, unsigned int cmd, unsigned long arg); +char *ll_get_fsname(struct super_block *sb, char *buf, int buflen); /* llite/llite_nfs.c */ extern struct export_operations lustre_export_operations; __u32 get_uuid2int(const char *name, int len); +struct inode *search_inode_for_lustre(struct super_block *sb, + const struct lu_fid *fid); /* llite/special.c */ extern struct inode_operations ll_special_inode_operations; @@ -818,11 +929,11 @@ extern struct inode_operations ll_fast_symlink_inode_operations; /* llite/llite_close.c */ struct ll_close_queue { - cfs_spinlock_t lcq_lock; - cfs_list_t lcq_head; - cfs_waitq_t lcq_waitq; - cfs_completion_t lcq_comp; - cfs_atomic_t lcq_stop; + spinlock_t lcq_lock; + cfs_list_t lcq_head; + wait_queue_head_t lcq_waitq; + struct completion lcq_comp; + cfs_atomic_t lcq_stop; }; struct ccc_object *cl_inode2ccc(struct inode *inode); @@ -865,20 +976,7 @@ struct vvp_io { /** * locked page returned from vvp_io */ - cfs_page_t *ft_vmpage; -#ifndef HAVE_VM_OP_FAULT - struct vm_nopage_api { - /** - * Virtual address at which fault occurred. - */ - unsigned long ft_address; - /** - * Fault type, as to be supplied to - * filemap_nopage(). - */ - int *ft_type; - } nopage; -#else + struct page *ft_vmpage; struct vm_fault_api { /** * kernel fault info @@ -889,7 +987,6 @@ struct vvp_io { */ unsigned int ft_flags; } fault; -#endif } fault; } u; /** @@ -916,9 +1013,7 @@ struct vvp_io_args { union { struct { -#ifndef HAVE_FILE_WRITEV struct kiocb *via_iocb; -#endif struct iovec *via_iov; unsigned long via_nrsegs; } normal; @@ -1012,19 +1107,24 @@ struct ll_lock_tree_node * ll_node_from_inode(struct inode *inode, __u64 start, __u64 end, ldlm_mode_t mode); void policy_from_vma(ldlm_policy_data_t *policy, struct vm_area_struct *vma, unsigned long addr, size_t count); -struct vm_area_struct *our_vma(unsigned long addr, size_t count); +struct vm_area_struct *our_vma(struct mm_struct *mm, unsigned long addr, + size_t count); static inline void ll_invalidate_page(struct page *vmpage) { - struct address_space *mapping = vmpage->mapping; - loff_t offset = vmpage->index << PAGE_CACHE_SHIFT; - - LASSERT(PageLocked(vmpage)); - if (mapping == NULL) - return; - - ll_teardown_mmaps(mapping, offset, offset + CFS_PAGE_SIZE); - truncate_complete_page(mapping, vmpage); + struct address_space *mapping = vmpage->mapping; + loff_t offset = vmpage->index << PAGE_CACHE_SHIFT; + + LASSERT(PageLocked(vmpage)); + if (mapping == NULL) + return; + + /* + * truncate_complete_page() calls + * a_ops->invalidatepage()->cl_page_delete()->vvp_page_delete(). + */ + ll_teardown_mmaps(mapping, offset, offset + PAGE_CACHE_SIZE); + truncate_complete_page(mapping, vmpage); } #define ll_s2sbi(sb) (s2lsi(sb)->lsi_llsbi) @@ -1068,9 +1168,10 @@ static inline struct obd_export *ll_i2mdexp(struct inode *inode) static inline struct lu_fid *ll_inode2fid(struct inode *inode) { struct lu_fid *fid; + LASSERT(inode != NULL); fid = &ll_i2info(inode)->lli_fid; - LASSERT(fid_is_igif(fid) || fid_ver(fid) == 0); + return fid; } @@ -1093,8 +1194,8 @@ ssize_t ll_listxattr(struct dentry *dentry, char *buffer, size_t size); int ll_removexattr(struct dentry *dentry, const char *name); /* llite/remote_perm.c */ -extern cfs_mem_cache_t *ll_remote_perm_cachep; -extern cfs_mem_cache_t *ll_rmtperm_hash_cachep; +extern struct kmem_cache *ll_remote_perm_cachep; +extern struct kmem_cache *ll_rmtperm_hash_cachep; cfs_hlist_head_t *alloc_rmtperm_hash(void); void free_rmtperm_hash(cfs_hlist_head_t *hash); @@ -1102,7 +1203,7 @@ int ll_update_remote_perm(struct inode *inode, struct mdt_remote_perm *perm); int lustre_check_remote_perm(struct inode *inode, int mask); /* llite/llite_capa.c */ -extern cfs_timer_t ll_capa_timer; +extern struct timer_list ll_capa_timer; int ll_capa_thread_start(void); void ll_capa_thread_stop(void); @@ -1194,73 +1295,105 @@ struct ll_statahead_info { unsigned int sai_miss_hidden;/* "ls -al", but first dentry * is not a hidden one */ unsigned int sai_skip_hidden;/* skipped hidden dentry count */ - unsigned int sai_ls_all:1, /* "ls -al", do stat-ahead for - * hidden entries */ - sai_in_readpage:1;/* statahead is in readdir()*/ - cfs_waitq_t sai_waitq; /* stat-ahead wait queue */ - struct ptlrpc_thread sai_thread; /* stat-ahead thread */ - cfs_list_t sai_entries_sent; /* entries sent out */ + unsigned int sai_ls_all:1, /* "ls -al", do stat-ahead for + * hidden entries */ + sai_in_readpage:1,/* statahead is in readdir()*/ + sai_agl_valid:1;/* AGL is valid for the dir */ + wait_queue_head_t sai_waitq; /* stat-ahead wait queue */ + struct ptlrpc_thread sai_thread; /* stat-ahead thread */ + struct ptlrpc_thread sai_agl_thread; /* AGL thread */ + cfs_list_t sai_entries; /* entry list */ cfs_list_t sai_entries_received; /* entries returned */ cfs_list_t sai_entries_stated; /* entries stated */ + cfs_list_t sai_entries_agl; /* AGL entries to be sent */ cfs_list_t sai_cache[LL_SA_CACHE_SIZE]; - cfs_spinlock_t sai_cache_lock[LL_SA_CACHE_SIZE]; - cfs_atomic_t sai_cache_count; /* entry count in cache */ + spinlock_t sai_cache_lock[LL_SA_CACHE_SIZE]; + cfs_atomic_t sai_cache_count; /* entry count in cache */ }; int do_statahead_enter(struct inode *dir, struct dentry **dentry, int only_unplug); void ll_stop_statahead(struct inode *dir, void *key); +static inline int ll_glimpse_size(struct inode *inode) +{ + struct ll_inode_info *lli = ll_i2info(inode); + int rc; + + down_read(&lli->lli_glimpse_sem); + rc = cl_glimpse_size(inode); + lli->lli_glimpse_time = cfs_time_current(); + up_read(&lli->lli_glimpse_sem); + return rc; +} + static inline void ll_statahead_mark(struct inode *dir, struct dentry *dentry) { - struct ll_inode_info *lli = ll_i2info(dir); - struct ll_statahead_info *sai = lli->lli_sai; - struct ll_dentry_data *ldd = ll_d2d(dentry); + struct ll_inode_info *lli = ll_i2info(dir); + struct ll_statahead_info *sai = lli->lli_sai; + struct ll_dentry_data *ldd = ll_d2d(dentry); - /* not the same process, don't mark */ - if (lli->lli_opendir_pid != cfs_curproc_pid()) - return; + /* not the same process, don't mark */ + if (lli->lli_opendir_pid != current_pid()) + return; - if (sai != NULL && ldd != NULL) - ldd->lld_sa_generation = sai->sai_generation; + LASSERT(ldd != NULL); + if (sai != NULL) + ldd->lld_sa_generation = sai->sai_generation; } static inline int -ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug) +d_need_statahead(struct inode *dir, struct dentry *dentryp) { - struct ll_inode_info *lli; - struct ll_dentry_data *ldd; - - if (ll_i2sbi(dir)->ll_sa_max == 0) - return -ENOTSUPP; + struct ll_inode_info *lli; + struct ll_dentry_data *ldd; + + if (ll_i2sbi(dir)->ll_sa_max == 0) + return -EAGAIN; + + lli = ll_i2info(dir); + /* not the same process, don't statahead */ + if (lli->lli_opendir_pid != current_pid()) + return -EAGAIN; + + /* statahead has been stopped */ + if (lli->lli_opendir_key == NULL) + return -EAGAIN; + + ldd = ll_d2d(dentryp); + /* + * When stats a dentry, the system trigger more than once "revalidate" + * or "lookup", for "getattr", for "getxattr", and maybe for others. + * Under patchless client mode, the operation intent is not accurate, + * which maybe misguide the statahead thread. For example: + * The "revalidate" call for "getattr" and "getxattr" of a dentry maybe + * have the same operation intent -- "IT_GETATTR". + * In fact, one dentry should has only one chance to interact with the + * statahead thread, otherwise the statahead windows will be confused. + * The solution is as following: + * Assign "lld_sa_generation" with "sai_generation" when a dentry + * "IT_GETATTR" for the first time, and the subsequent "IT_GETATTR" + * will bypass interacting with statahead thread for checking: + * "lld_sa_generation == lli_sai->sai_generation" + */ + if (ldd && lli->lli_sai && + ldd->lld_sa_generation == lli->lli_sai->sai_generation) + return -EAGAIN; + + return 1; +} - lli = ll_i2info(dir); - /* not the same process, don't statahead */ - if (lli->lli_opendir_pid != cfs_curproc_pid()) - return -EAGAIN; +static inline int +ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug) +{ + int ret; - ldd = ll_d2d(*dentryp); - /* - * When stats a dentry, the system trigger more than once "revalidate" - * or "lookup", for "getattr", for "getxattr", and maybe for others. - * Under patchless client mode, the operation intent is not accurate, - * which maybe misguide the statahead thread. For example: - * The "revalidate" call for "getattr" and "getxattr" of a dentry maybe - * have the same operation intent -- "IT_GETATTR". - * In fact, one dentry should has only one chance to interact with the - * statahead thread, otherwise the statahead windows will be confused. - * The solution is as following: - * Assign "lld_sa_generation" with "sai_generation" when a dentry - * "IT_GETATTR" for the first time, and the subsequent "IT_GETATTR" - * will bypass interacting with statahead thread for checking: - * "lld_sa_generation == lli_sai->sai_generation" - */ - if (ldd && lli->lli_sai && - ldd->lld_sa_generation == lli->lli_sai->sai_generation) - return -EAGAIN; + ret = d_need_statahead(dir, *dentryp); + if (ret <= 0) + return ret; - return do_statahead_enter(dir, dentryp, only_unplug); + return do_statahead_enter(dir, dentryp, only_unplug); } /* llite ioctl register support rountine */ @@ -1316,34 +1449,41 @@ void ll_iocontrol_unregister(void *magic); #define cl_inode_mode(inode) ((inode)->i_mode) #define cl_i2sbi ll_i2sbi -static inline void cl_isize_lock(struct inode *inode, int lsmlock) +static inline struct ll_file_data *cl_iattr2fd(struct inode *inode, + const struct iattr *attr) +{ + LASSERT(attr->ia_valid & ATTR_FILE); + return LUSTRE_FPRIVATE(attr->ia_file); +} + +static inline void cl_isize_lock(struct inode *inode) { - ll_inode_size_lock(inode, lsmlock); + ll_inode_size_lock(inode); } -static inline void cl_isize_unlock(struct inode *inode, int lsmlock) +static inline void cl_isize_unlock(struct inode *inode) { - ll_inode_size_unlock(inode, lsmlock); + ll_inode_size_unlock(inode); } static inline void cl_isize_write_nolock(struct inode *inode, loff_t kms) { - LASSERT_SEM_LOCKED(&ll_i2info(inode)->lli_size_sem); - i_size_write(inode, kms); + LASSERT(down_trylock(&ll_i2info(inode)->lli_size_sem) != 0); + i_size_write(inode, kms); } static inline void cl_isize_write(struct inode *inode, loff_t kms) { - ll_inode_size_lock(inode, 0); - i_size_write(inode, kms); - ll_inode_size_unlock(inode, 0); + ll_inode_size_lock(inode); + i_size_write(inode, kms); + ll_inode_size_unlock(inode); } #define cl_isize_read(inode) i_size_read(inode) -static inline int cl_merge_lvb(struct inode *inode) +static inline int cl_merge_lvb(const struct lu_env *env, struct inode *inode) { - return ll_merge_lvb(inode); + return ll_merge_lvb(env, inode); } #define cl_inode_atime(inode) LTIME_S((inode)->i_atime) @@ -1352,6 +1492,9 @@ static inline int cl_merge_lvb(struct inode *inode) struct obd_capa *cl_capa_lookup(struct inode *inode, enum cl_req_type crt); +int cl_sync_file_range(struct inode *inode, loff_t start, loff_t end, + enum cl_fsync_mode mode, int ignore_layout); + /** direct write pages */ struct ll_dio_pages { /** page array to be written. we don't support @@ -1395,43 +1538,101 @@ static inline int ll_file_nolock(const struct file *file) static inline void ll_set_lock_data(struct obd_export *exp, struct inode *inode, struct lookup_intent *it, __u64 *bits) { - if (!it->d.lustre.it_lock_set) { - CDEBUG(D_DLMTRACE, "setting l_data to inode %p (%lu/%u)\n", - inode, inode->i_ino, inode->i_generation); - md_set_lock_data(exp, &it->d.lustre.it_lock_handle, - inode, &it->d.lustre.it_lock_bits); - it->d.lustre.it_lock_set = 1; - } - - if (bits != NULL) - *bits = it->d.lustre.it_lock_bits; + if (!it->d.lustre.it_lock_set) { + struct lustre_handle handle; + + /* If this inode is a remote object, it will get two + * separate locks in different namespaces, Master MDT, + * where the name entry is, will grant LOOKUP lock, + * remote MDT, where the object is, will grant + * UPDATE|PERM lock. The inode will be attched to both + * LOOKUP and PERM locks, so revoking either locks will + * case the dcache being cleared */ + if (it->d.lustre.it_remote_lock_mode) { + handle.cookie = it->d.lustre.it_remote_lock_handle; + CDEBUG(D_DLMTRACE, "setting l_data to inode "DFID + "(%p) for remote lock "LPX64"\n", + PFID(ll_inode2fid(inode)), inode, + handle.cookie); + md_set_lock_data(exp, &handle.cookie, inode, NULL); + } + + handle.cookie = it->d.lustre.it_lock_handle; + + CDEBUG(D_DLMTRACE, "setting l_data to inode "DFID"(%p)" + " for lock "LPX64"\n", + PFID(ll_inode2fid(inode)), inode, handle.cookie); + + md_set_lock_data(exp, &handle.cookie, inode, + &it->d.lustre.it_lock_bits); + it->d.lustre.it_lock_set = 1; + } + + if (bits != NULL) + *bits = it->d.lustre.it_lock_bits; } -static inline void ll_dentry_rehash(struct dentry *dentry, int locked) +static inline void ll_lock_dcache(struct inode *inode) { - if (!locked) { - cfs_spin_lock(&ll_lookup_lock); - spin_lock(&dcache_lock); - } - if (d_unhashed(dentry)) - d_rehash_cond(dentry, 0); - if (!locked) { - spin_unlock(&dcache_lock); - cfs_spin_unlock(&ll_lookup_lock); - } +#ifdef HAVE_DCACHE_LOCK + spin_lock(&dcache_lock); +#else + spin_lock(&inode->i_lock); +#endif } -static inline void ll_dentry_reset_flags(struct dentry *dentry, __u64 bits) +static inline void ll_unlock_dcache(struct inode *inode) { - if (bits & MDS_INODELOCK_LOOKUP && - dentry->d_flags & DCACHE_LUSTRE_INVALID) { - lock_dentry(dentry); - dentry->d_flags &= ~DCACHE_LUSTRE_INVALID; - unlock_dentry(dentry); - } +#ifdef HAVE_DCACHE_LOCK + spin_unlock(&dcache_lock); +#else + spin_unlock(&inode->i_lock); +#endif } -#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2,7,50,0) +static inline int d_lustre_invalid(const struct dentry *dentry) +{ + struct ll_dentry_data *lld = ll_d2d(dentry); + + return (lld == NULL) || lld->lld_invalid; +} + +static inline void __d_lustre_invalidate(struct dentry *dentry) +{ + struct ll_dentry_data *lld = ll_d2d(dentry); + + if (lld != NULL) + lld->lld_invalid = 1; +} + +/* + * Mark dentry INVALID, if dentry refcount is zero (this is normally case for + * ll_md_blocking_ast), unhash this dentry, and let dcache to reclaim it later; + * else dput() of the last refcount will unhash this dentry and kill it. + */ +static inline void d_lustre_invalidate(struct dentry *dentry, int nested) +{ + CDEBUG(D_DENTRY, "invalidate dentry %.*s (%p) parent %p inode %p " + "refc %d\n", dentry->d_name.len, dentry->d_name.name, dentry, + dentry->d_parent, dentry->d_inode, d_count(dentry)); + + spin_lock_nested(&dentry->d_lock, + nested ? DENTRY_D_LOCK_NESTED : DENTRY_D_LOCK_NORMAL); + __d_lustre_invalidate(dentry); + if (d_count(dentry) == 0) + __d_drop(dentry); + spin_unlock(&dentry->d_lock); +} + +static inline void d_lustre_revalidate(struct dentry *dentry) +{ + spin_lock(&dentry->d_lock); + LASSERT(ll_d2d(dentry) != NULL); + ll_d2d(dentry)->lld_invalid = 0; + spin_unlock(&dentry->d_lock); +} + +#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 7, 50, 0) /* Compatibility for old (1.8) compiled userspace quota code */ struct if_quotactl_18 { __u32 qc_cmd; @@ -1447,6 +1648,21 @@ struct if_quotactl_18 { /* End compatibility for old (1.8) compiled userspace quota code */ #else #warning "remove old LL_IOC_QUOTACTL_18 compatibility code" -#endif /* LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2,7,50,0) */ +#endif /* LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 7, 50, 0) */ + +enum { + LL_LAYOUT_GEN_NONE = ((__u32)-2), /* layout lock was cancelled */ + LL_LAYOUT_GEN_EMPTY = ((__u32)-1) /* for empty layout */ +}; + +int ll_layout_conf(struct inode *inode, const struct cl_object_conf *conf); +int ll_layout_refresh(struct inode *inode, __u32 *gen); +int ll_layout_restore(struct inode *inode, loff_t start, __u64 length); + +int ll_xattr_init(void); +void ll_xattr_fini(void); + +int ll_page_sync_io(const struct lu_env *env, struct cl_io *io, + struct cl_page *page, enum cl_req_type crt); #endif /* LLITE_INTERNAL_H */