* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2011, 2012, Whamcloud, Inc.
+ * Copyright (c) 2011, 2013, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
#ifndef LLITE_INTERNAL_H
#define LLITE_INTERNAL_H
-
-#include <lustre_acl.h>
-
#include <lustre_debug.h>
#include <lustre_ver.h>
#include <lustre_disk.h> /* for s2sbi */
LLIF_CONTENDED = (1 << 4),
/* Truncate uses server lock for this file */
LLIF_SRVLOCK = (1 << 5),
-
+ /* File data is modified. */
+ LLIF_DATA_MODIFIED = (1 << 6),
};
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_mutex_t lli_rmtperm_mutex;
+ 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;
__u64 lli_open_fd_write_count;
__u64 lli_open_fd_exec_count;
/* Protects access to och pointers and their usage counters */
- cfs_mutex_t lli_och_mutex;
+ 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;
- cfs_spinlock_t lli_agl_lock;
+ /* 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_mutex_t d_readdir_mutex;
+ /* 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
struct ll_statahead_info *d_sai;
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;
+ 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_mutex u.d.d_readdir_mutex
#define lli_opendir_key u.d.d_opendir_key
#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_mutex_t f_write_mutex;
+ /* 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;
- cfs_rw_semaphore_t f_glimpse_sem;
+ 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;
+ /* 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
#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;
struct cl_object *lli_clob;
/* mutex to request for layout lock exclusively. */
- cfs_mutex_t lli_layout_mutex;
+ struct mutex lli_layout_mutex;
/* valid only inside LAYOUT ibits lock, protected by lli_layout_mutex */
__u32 lli_layout_gen;
};
#define LL_SBI_LAYOUT_LOCK 0x20000 /* layout lock support */
#define LL_SBI_USER_FID2PATH 0x40000 /* allow fid2path by unprivileged users */
+#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" }
+
/* default value for ll_sb_info->contention_time */
#define SBI_DEFAULT_CONTENTION_SECONDS 60
/* default value for lockless_truncate_enable */
};
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
};
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;
struct lu_fid ll_root_fid; /* root object fid */
int ll_flags;
+ int ll_umounting:1;
cfs_list_t ll_conn_chain; /* per-conn chain of SBs */
struct lustre_client_ocd ll_lco;
* 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.
struct lov_stripe_md;
-extern cfs_spinlock_t inode_lock;
+extern spinlock_t inode_lock;
extern struct proc_dir_entry *proc_lustre_fs_root;
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);
+#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);
struct lustre_handle *lockh, __u64 flags);
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,
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);
#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 inode *inode, void *arg);
+int ll_data_version(struct inode *inode, __u64 *data_version, int extent_lock);
/* llite/dcache.c */
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 dentry *dentry, struct iattr *attr);
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(cfs_page_t *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);
/* 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;
/* 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;
+ cfs_waitq_t lcq_waitq;
+ struct completion lcq_comp;
+ cfs_atomic_t lcq_stop;
};
struct ccc_object *cl_inode2ccc(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;
}
cfs_waitq_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_sent; /* entries sent out */
+ 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,
static inline int ll_glimpse_size(struct inode *inode)
{
- struct ll_inode_info *lli = ll_i2info(inode);
- int rc;
-
- cfs_down_read(&lli->lli_glimpse_sem);
- rc = cl_glimpse_size(inode);
- lli->lli_glimpse_time = cfs_time_current();
- cfs_up_read(&lli->lli_glimpse_sem);
- return rc;
+ 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
}
static inline int
-ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug)
+ll_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 -EAGAIN;
+ 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 != cfs_curproc_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 = ll_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 */
#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)
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);
+ enum cl_fsync_mode mode, int ignore_layout);
/** direct write pages */
struct ll_dio_pages {
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 %p"
+ "(%lu/%u) for remote lock "LPX64"\n", inode,
+ inode->i_ino, inode->i_generation,
+ 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 %p (%lu/%u)"
+ " for lock "LPX64"\n", inode, inode->i_ino,
+ inode->i_generation, 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_lock_dcache(struct inode *inode)
* 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)
+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_refcount(dentry));
- spin_lock(&dentry->d_lock);
+ spin_lock_nested(&dentry->d_lock,
+ nested ? DENTRY_D_LOCK_NESTED : DENTRY_D_LOCK_NORMAL);
__d_lustre_invalidate(dentry);
if (d_refcount(dentry) == 0)
__d_drop(dentry);
#warning "remove old LL_IOC_QUOTACTL_18 compatibility code"
#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);