Whamcloud - gitweb
landing b_cmobd_merge on HEAD
[fs/lustre-release.git] / lustre / llite / file.c
index 8efd389..aa00caf 100644 (file)
 #if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0))
 #include <linux/lustre_compat25.h>
 #endif
-
 #include "llite_internal.h"
 
-static int ll_mdc_close(struct obd_export *mdc_exp, struct inode *inode,
-                        struct file *file)
+int ll_mdc_close(struct obd_export *mdc_exp, struct inode *inode,
+                 struct file *file)
 {
         struct ll_file_data *fd = file->private_data;
         struct ptlrpc_request *req = NULL;
         struct obd_client_handle *och = &fd->fd_mds_och;
         struct obdo obdo;
-        int rc, valid;
+        int rc;
         ENTRY;
 
         /* clear group lock, if present */
-        if (fd->fd_flags & LL_FILE_CW_LOCKED) {
+        if (fd->fd_flags & LL_FILE_GROUP_LOCKED) {
                 struct lov_stripe_md *lsm = ll_i2info(inode)->lli_smd;
-                fd->fd_flags &= ~(LL_FILE_CW_LOCKED|LL_FILE_IGNORE_LOCK);
-                rc = ll_extent_unlock(fd, inode, lsm, LCK_CW, &fd->fd_cwlockh);
+                fd->fd_flags &= ~(LL_FILE_GROUP_LOCKED|LL_FILE_IGNORE_LOCK);
+                rc = ll_extent_unlock(fd, inode, lsm, LCK_GROUP,
+                                      &fd->fd_cwlockh);
         }
 
-        valid = OBD_MD_FLID;
-
-        memset(&obdo, 0, sizeof(obdo));
         obdo.o_id = inode->i_ino;
-        obdo.o_mode = inode->i_mode;
-        obdo.o_size = inode->i_size;
-        obdo.o_blocks = inode->i_blocks;
+        obdo.o_valid = OBD_MD_FLID;
+        obdo_from_inode(&obdo, inode, OBD_MD_FLTYPE | OBD_MD_FLMODE |
+                                      OBD_MD_FLSIZE | OBD_MD_FLBLOCKS |
+                                      OBD_MD_FLATIME | OBD_MD_FLMTIME |
+                                      OBD_MD_FLCTIME);
         if (0 /* ll_is_inode_dirty(inode) */) {
                 obdo.o_flags = MDS_BFLAG_UNCOMMITTED_WRITES;
-                valid |= OBD_MD_FLFLAGS;
+                obdo.o_valid |= OBD_MD_FLFLAGS;
         }
-        obdo.o_valid = valid;
         obdo.o_mds = ll_i2info(inode)->lli_mds;
         rc = md_close(mdc_exp, &obdo, och, &req);
+
         if (rc == EAGAIN) {
                 /* We are the last writer, so the MDS has instructed us to get
                  * the file size and any write cookies, then close again. */
@@ -80,7 +79,7 @@ static int ll_mdc_close(struct obd_export *mdc_exp, struct inode *inode,
                                inode->i_ino, rc);
         }
 
-        mdc_clear_open_replay_data(och);
+        mdc_clear_open_replay_data(mdc_exp, och);
         ptlrpc_req_finished(req);
         och->och_fh.cookie = DEAD_HANDLE_MAGIC;
         file->private_data = NULL;
@@ -139,17 +138,19 @@ static int ll_intent_file_open(struct file *file, void *lmm,
         if (rc == 0) {
                 if (itp->d.lustre.it_lock_mode)
                         memcpy(&itp->d.lustre.it_lock_handle,
-                                        &lockh, sizeof(lockh));
+                               &lockh, sizeof(lockh));
         } else if (rc < 0) {
                 CERROR("lock enqueue: err: %d\n", rc);
         }
+        
         RETURN(rc);
 }
 
-static int ll_local_open(struct file *file, struct lookup_intent *it)
+int ll_local_open(struct file *file, struct lookup_intent *it)
 {
         struct ptlrpc_request *req = it->d.lustre.it_data;
         struct ll_inode_info *lli = ll_i2info(file->f_dentry->d_inode);
+        struct obd_export *mdc_exp = ll_i2mdcexp(file->f_dentry->d_inode);
         struct ll_file_data *fd;
         struct mds_body *body;
         ENTRY;
@@ -172,7 +173,7 @@ static int ll_local_open(struct file *file, struct lookup_intent *it)
 
         lli->lli_io_epoch = body->io_epoch;
 
-        mdc_set_open_replay_data(&fd->fd_mds_och, it->d.lustre.it_data);
+        mdc_set_open_replay_data(mdc_exp, &fd->fd_mds_och, it->d.lustre.it_data);
 
         RETURN(0);
 }
@@ -195,7 +196,8 @@ static int ll_local_open(struct file *file, struct lookup_intent *it)
 int ll_file_open(struct inode *inode, struct file *file)
 {
         struct ll_inode_info *lli = ll_i2info(inode);
-        struct lookup_intent *it;
+        struct lookup_intent *it, oit = { .it_op = IT_OPEN,
+                                          .it_flags = file->f_flags };
         struct lov_stripe_md *lsm;
         struct ptlrpc_request *req;
         int rc = 0;
@@ -210,9 +212,7 @@ int ll_file_open(struct inode *inode, struct file *file)
 
         it = file->f_it;
 
-        if (!it->d.lustre.it_disposition) {
-                struct lookup_intent oit = { .it_op = IT_OPEN,
-                                             .it_flags = file->f_flags };
+        if (!it || !it->d.lustre.it_disposition) {
                 it = &oit;
                 rc = ll_intent_file_open(file, NULL, 0, it);
                 if (rc)
@@ -282,7 +282,7 @@ int ll_lsm_getattr(struct obd_export *exp, struct lov_stripe_md *lsm,
         if (rc)
                 RETURN(rc);
 
-        oa->o_valid &= (OBD_MD_FLBLOCKS | OBD_MD_FLBLKSZ | OBD_MD_FLMTIME | 
+        oa->o_valid &= (OBD_MD_FLBLOCKS | OBD_MD_FLBLKSZ | OBD_MD_FLMTIME |
                         OBD_MD_FLCTIME | OBD_MD_FLSIZE);
         RETURN(0);
 }
@@ -399,15 +399,10 @@ void ll_pgcache_remove_extent(struct inode *inode, struct lov_stripe_md *lsm,
                          tmpex.l_extent.start, lock->l_policy_data.l_extent.end,
                          start, i, end);
 
-                ll_pgcache_lock(inode->i_mapping);
-                if (list_empty(&inode->i_mapping->dirty_pages) &&
-                    list_empty(&inode->i_mapping->clean_pages) &&
-                    list_empty(&inode->i_mapping->locked_pages)) {
+                if (!mapping_has_pages(inode->i_mapping)) {
                         CDEBUG(D_INODE|D_PAGE, "nothing left\n");
-                        ll_pgcache_unlock(inode->i_mapping);
                         break;
                 }
-                ll_pgcache_unlock(inode->i_mapping);
 
                 conditional_schedule();
 
@@ -419,14 +414,7 @@ void ll_pgcache_remove_extent(struct inode *inode, struct lov_stripe_md *lsm,
                 lock_page(page);
 
                 /* page->mapping to check with racing against teardown */
-                if (page->mapping && PageDirty(page) && !discard) {
-                        ClearPageDirty(page);
-                        LL_CDEBUG_PAGE(D_PAGE, page, "found dirty\n");
-                        ll_pgcache_lock(inode->i_mapping);
-                        list_del(&page->list);
-                        list_add(&page->list, &inode->i_mapping->locked_pages);
-                        ll_pgcache_unlock(inode->i_mapping);
-
+                if (!discard && clear_page_dirty_for_io(page)) {
                         rc = ll_call_writepage(inode, page);
                         if (rc != 0)
                                 CERROR("writepage of page %p failed: %d\n",
@@ -507,6 +495,7 @@ static int ll_extent_lock_callback(struct ldlm_lock *lock,
                 down(&inode->i_sem);
                 kms = ldlm_extent_shift_kms(lock,
                                             lsm->lsm_oinfo[stripe].loi_kms);
+               
                 if (lsm->lsm_oinfo[stripe].loi_kms != kms)
                         LDLM_DEBUG(lock, "updating kms from "LPU64" to "LPU64,
                                    lsm->lsm_oinfo[stripe].loi_kms, kms);
@@ -621,6 +610,7 @@ static int ll_glimpse_callback(struct ldlm_lock *lock, void *reqp)
 }
 
 __u64 lov_merge_size(struct lov_stripe_md *lsm, int kms);
+__u64 lov_merge_blocks(struct lov_stripe_md *lsm);
 __u64 lov_merge_mtime(struct lov_stripe_md *lsm, __u64 current_time);
 
 /* NB: lov_merge_size will prefer locally cached writes if they extend the
@@ -630,7 +620,7 @@ int ll_glimpse_size(struct inode *inode, struct ost_lvb *lvb)
         struct ll_inode_info *lli = ll_i2info(inode);
         struct ll_sb_info *sbi = ll_i2sbi(inode);
         ldlm_policy_data_t policy = { .l_extent = { 0, OBD_OBJECT_EOF } };
-        struct lustre_handle lockh;
+        struct lustre_handle lockh = { 0 };
         int rc, flags = LDLM_FL_HAS_INTENT;
         ENTRY;
 
@@ -646,9 +636,11 @@ int ll_glimpse_size(struct inode *inode, struct ost_lvb *lvb)
         }
 
         lvb->lvb_size = lov_merge_size(lli->lli_smd, 0);
+        inode->i_blocks = lov_merge_blocks(lli->lli_smd);
         //inode->i_mtime = lov_merge_mtime(lli->lli_smd, inode->i_mtime);
 
-        CDEBUG(D_DLMTRACE, "glimpse: size: "LPU64"\n", lvb->lvb_size);
+        CDEBUG(D_DLMTRACE, "glimpse: size: "LPU64", blocks: "LPU64"\n",
+               lvb->lvb_size, lvb->lvb_blocks);
 
         obd_cancel(sbi->ll_osc_exp, lli->lli_smd, LCK_PR, &lockh);
 
@@ -717,7 +709,7 @@ static ssize_t ll_file_read(struct file *filp, char *buf, size_t count,
         struct lov_stripe_md *lsm = lli->lli_smd;
         struct lustre_handle lockh = { 0 };
         ldlm_policy_data_t policy;
-        ldlm_error_t err;
+        int rc;
         ssize_t retval;
         __u64 kms;
         ENTRY;
@@ -738,14 +730,14 @@ static ssize_t ll_file_read(struct file *filp, char *buf, size_t count,
         policy.l_extent.start = *ppos;
         policy.l_extent.end = *ppos + count - 1;
 
-        err = ll_extent_lock(fd, inode, lsm, LCK_PR, &policy, &lockh,
-                             (filp->f_flags & O_NONBLOCK)?LDLM_FL_BLOCK_NOWAIT:
-                                                          0);
-        if (err != ELDLM_OK)
-                RETURN(err);
+        rc = ll_extent_lock(fd, inode, lsm, LCK_PR, &policy, &lockh,
+                                (filp->f_flags & O_NONBLOCK) ?
+                                        LDLM_FL_BLOCK_NOWAIT: 0);
+        if (rc != 0)
+                RETURN(rc);
 
         kms = lov_merge_size(lsm, 1);
-        if (policy.l_extent.end > kms) {
+        if (*ppos + count - 1 > kms) {
                 /* A glimpse is necessary to determine whether we return a short
                  * read or some zeroes at the end of the buffer */
                 struct ost_lvb lvb;
@@ -785,9 +777,8 @@ static ssize_t ll_file_write(struct file *file, const char *buf, size_t count,
         struct lustre_handle lockh = { 0 };
         ldlm_policy_data_t policy;
         loff_t maxbytes = ll_file_maxbytes(inode);
-        ldlm_error_t err;
         ssize_t retval;
-        int nonblock = 0;
+        int nonblock = 0, rc;
         ENTRY;
         if (file->f_flags & O_NONBLOCK)
                 nonblock = LDLM_FL_BLOCK_NOWAIT;
@@ -815,9 +806,9 @@ static ssize_t ll_file_write(struct file *file, const char *buf, size_t count,
                 policy.l_extent.end = *ppos + count - 1;
         }
 
-        err = ll_extent_lock(fd, inode, lsm, LCK_PW, &policy, &lockh, nonblock);
-        if (err != ELDLM_OK)
-                RETURN(err);
+        rc = ll_extent_lock(fd, inode, lsm, LCK_PW, &policy, &lockh, nonblock);
+        if (rc != 0)
+                RETURN(rc);
 
         /* this is ok, g_f_w will overwrite this under i_sem if it races
          * with a local truncate, it just makes our maxbyte checking easier */
@@ -947,7 +938,7 @@ static int ll_lov_setstripe_ea_info(struct inode *inode, struct file *file,
         if (rc < 0)
                 GOTO(out, rc);
 
-        rc = mdc_req2lustre_md(req, 1, exp, NULL, &md);
+        rc = mdc_req2lustre_md(ll_i2mdcexp(inode), req, 1, exp, &md);
         if (rc)
                 GOTO(out, rc);
         ll_update_inode(f->f_dentry->d_inode, &md);
@@ -973,7 +964,7 @@ static int ll_lov_setea(struct inode *inode, struct file *file,
 {
         int flags = MDS_OPEN_HAS_OBJS | FMODE_WRITE;
         struct lov_user_md  *lump;
-        int lum_size = sizeof(struct lov_user_md) + 
+        int lum_size = sizeof(struct lov_user_md) +
                        sizeof(struct lov_user_ost_data);
         int rc;
         ENTRY;
@@ -985,8 +976,7 @@ static int ll_lov_setea(struct inode *inode, struct file *file,
         if (lump == NULL) {
                 RETURN(-ENOMEM);
         }
-        rc = copy_from_user(lump, (struct lov_user_md  *)arg, 
-                            lum_size);
+        rc = copy_from_user(lump, (struct lov_user_md  *)arg, lum_size);
         if (rc) {
                 OBD_FREE(lump, lum_size);
                 RETURN(-EFAULT);
@@ -1028,7 +1018,7 @@ static int ll_lov_getstripe(struct inode *inode, unsigned long arg)
                             (void *)arg);
 }
 
-static int ll_get_cwlock(struct inode *inode, struct file *file,
+static int ll_get_grouplock(struct inode *inode, struct file *file,
                          unsigned long arg)
 {
         struct ll_file_data *fd = file->private_data;
@@ -1037,11 +1027,10 @@ static int ll_get_cwlock(struct inode *inode, struct file *file,
         struct lustre_handle lockh = { 0 };
         struct ll_inode_info *lli = ll_i2info(inode);
         struct lov_stripe_md *lsm = lli->lli_smd;
-        ldlm_error_t err;
-        int flags = 0;
+        int flags = 0, rc;
         ENTRY;
 
-        if (fd->fd_flags & LL_FILE_CW_LOCKED) {
+        if (fd->fd_flags & LL_FILE_GROUP_LOCKED) {
                 RETURN(-EINVAL);
         }
 
@@ -1049,45 +1038,45 @@ static int ll_get_cwlock(struct inode *inode, struct file *file,
         if (file->f_flags & O_NONBLOCK)
                 flags = LDLM_FL_BLOCK_NOWAIT;
 
-        err = ll_extent_lock(fd, inode, lsm, LCK_CW, &policy, &lockh, flags);
-        if (err)
-                RETURN(err);
+        rc = ll_extent_lock(fd, inode, lsm, LCK_GROUP, &policy, &lockh, flags);
+        if (rc != 0)
+                RETURN(rc);
 
-        fd->fd_flags |= LL_FILE_CW_LOCKED|LL_FILE_IGNORE_LOCK;
+        fd->fd_flags |= LL_FILE_GROUP_LOCKED|LL_FILE_IGNORE_LOCK;
         fd->fd_gid = arg;
         memcpy(&fd->fd_cwlockh, &lockh, sizeof(lockh));
 
         RETURN(0);
 }
 
-static int ll_put_cwlock(struct inode *inode, struct file *file,
+static int ll_put_grouplock(struct inode *inode, struct file *file,
                          unsigned long arg)
 {
         struct ll_file_data *fd = file->private_data;
         struct ll_inode_info *lli = ll_i2info(inode);
         struct lov_stripe_md *lsm = lli->lli_smd;
-        ldlm_error_t err;
+        int rc;
         ENTRY;
 
-        if (!(fd->fd_flags & LL_FILE_CW_LOCKED)) {
+        if (!(fd->fd_flags & LL_FILE_GROUP_LOCKED)) {
                 /* Ugh, it's already unlocked. */
                 RETURN(-EINVAL);
         }
 
         if (fd->fd_gid != arg) /* Ugh? Unlocking with different gid? */
                 RETURN(-EINVAL);
-        
-        fd->fd_flags &= ~(LL_FILE_CW_LOCKED|LL_FILE_IGNORE_LOCK);
 
-        err = ll_extent_unlock(fd, inode, lsm, LCK_CW, &fd->fd_cwlockh);
-        if (err)
-                RETURN(err);
+        fd->fd_flags &= ~(LL_FILE_GROUP_LOCKED|LL_FILE_IGNORE_LOCK);
+
+        rc = ll_extent_unlock(fd, inode, lsm, LCK_GROUP, &fd->fd_cwlockh);
+        if (rc)
+                RETURN(rc);
 
         fd->fd_gid = 0;
         memset(&fd->fd_cwlockh, 0, sizeof(fd->fd_cwlockh));
 
         RETURN(0);
-}       
+}
 
 int ll_file_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
                   unsigned long arg)
@@ -1124,7 +1113,7 @@ int ll_file_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
         case LL_IOC_LOV_SETSTRIPE:
                 RETURN(ll_lov_setstripe(inode, file, arg));
         case LL_IOC_LOV_SETEA:
-                RETURN( ll_lov_setea(inode, file, arg) ); 
+                RETURN(ll_lov_setea(inode, file, arg));
         case LL_IOC_LOV_GETSTRIPE:
                 RETURN(ll_lov_getstripe(inode, arg));
         case LL_IOC_RECREATE_OBJ:
@@ -1132,10 +1121,10 @@ int ll_file_ioctl(struct inode *inode, struct file *file, unsigned int cmd,
         case EXT3_IOC_GETFLAGS:
         case EXT3_IOC_SETFLAGS:
                 RETURN( ll_iocontrol(inode, file, cmd, arg) );
-        case LL_IOC_CW_LOCK:
-                RETURN(ll_get_cwlock(inode, file, arg));
-        case LL_IOC_CW_UNLOCK:
-                RETURN(ll_put_cwlock(inode, file, arg));
+        case LL_IOC_GROUP_LOCK:
+                RETURN(ll_get_grouplock(inode, file, arg));
+        case LL_IOC_GROUP_UNLOCK:
+                RETURN(ll_put_grouplock(inode, file, arg));
         /* We need to special case any other ioctls we want to handle,
          * to send them to the MDS/OST as appropriate and to properly
          * network encode the arg field.
@@ -1164,17 +1153,16 @@ loff_t ll_file_seek(struct file *file, loff_t offset, int origin)
 
         lprocfs_counter_incr(ll_i2sbi(inode)->ll_stats, LPROC_LL_LLSEEK);
         if (origin == 2) { /* SEEK_END */
-                ldlm_error_t err;
-                int nonblock = 0;
+                int nonblock = 0, rc;
                 ldlm_policy_data_t policy = { .l_extent = {0, OBD_OBJECT_EOF }};
 
                 if (file->f_flags & O_NONBLOCK)
                         nonblock = LDLM_FL_BLOCK_NOWAIT;
 
-                err = ll_extent_lock(fd, inode, lsm, LCK_PR, &policy, &lockh,
+                rc = ll_extent_lock(fd, inode, lsm, LCK_PR, &policy, &lockh,
                                      nonblock);
-                if (err != ELDLM_OK)
-                        RETURN(err);
+                if (rc != 0)
+                        RETURN(rc);
 
                 offset += inode->i_size;
         } else if (origin == 1) { /* SEEK_CUR */
@@ -1413,34 +1401,28 @@ int ll_getattr(struct vfsmount *mnt, struct dentry *de,
 #endif
 
 struct file_operations ll_file_operations = {
-        read:           ll_file_read,
-        write:          ll_file_write,
-        ioctl:          ll_file_ioctl,
-        open:           ll_file_open,
-        release:        ll_file_release,
-        mmap:           generic_file_mmap,
-        llseek:         ll_file_seek,
-        fsync:          ll_fsync,
-        //lock:           ll_file_flock
+        .read           = ll_file_read,
+        .write          = ll_file_write,
+        .ioctl          = ll_file_ioctl,
+        .open           = ll_file_open,
+        .release        = ll_file_release,
+        .mmap           = generic_file_mmap,
+        .llseek         = ll_file_seek,
+#if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,5,0))
+        .sendfile       = generic_file_sendfile,
+#endif
+        .fsync          = ll_fsync,
+        //.lock           ll_file_flock
 };
 
 struct inode_operations ll_file_inode_operations = {
-        setattr_raw:    ll_setattr_raw,
-        setattr:        ll_setattr,
-        truncate:       ll_truncate,
+        .setattr_raw    = ll_setattr_raw,
+        .setattr        = ll_setattr,
+        .truncate       = ll_truncate,
 #if (LINUX_VERSION_CODE > KERNEL_VERSION(2,5,0))
-        getattr_it:     ll_getattr,
+        .getattr_it     = ll_getattr,
 #else
-        revalidate_it:  ll_inode_revalidate_it,
+        .revalidate_it  = ll_inode_revalidate_it,
 #endif
 };
 
-struct inode_operations ll_special_inode_operations = {
-        setattr_raw:    ll_setattr_raw,
-        setattr:        ll_setattr,
-#if (LINUX_VERSION_CODE > KERNEL_VERSION(2,5,0))
-        getattr_it:     ll_getattr,
-#else
-        revalidate_it:  ll_inode_revalidate_it,
-#endif
-};