diff options
author | Dave Jones <davej@redhat.com> | 2006-12-12 17:41:41 -0500 |
---|---|---|
committer | Dave Jones <davej@redhat.com> | 2006-12-12 17:41:41 -0500 |
commit | c4366889dda8110247be59ca41fddb82951a8c26 (patch) | |
tree | 705c1a996bed8fd48ce94ff33ec9fd00f9b94875 /fs/ocfs2/file.c | |
parent | db2fb9db5735cc532fd4fc55e94b9a3c3750378e (diff) | |
parent | e1036502e5263851259d147771226161e5ccc85a (diff) |
Merge ../linus
Conflicts:
drivers/cpufreq/cpufreq.c
Diffstat (limited to 'fs/ocfs2/file.c')
-rw-r--r-- | fs/ocfs2/file.c | 406 |
1 files changed, 276 insertions, 130 deletions
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index d9ba0a931a0..e9a82ad95c1 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -30,6 +30,9 @@ #include <linux/highmem.h> #include <linux/pagemap.h> #include <linux/uio.h> +#include <linux/sched.h> +#include <linux/pipe_fs_i.h> +#include <linux/mount.h> #define MLOG_MASK_PREFIX ML_INODE #include <cluster/masklog.h> @@ -65,7 +68,7 @@ static int ocfs2_file_open(struct inode *inode, struct file *file) struct ocfs2_inode_info *oi = OCFS2_I(inode); mlog_entry("(0x%p, 0x%p, '%.*s')\n", inode, file, - file->f_dentry->d_name.len, file->f_dentry->d_name.name); + file->f_path.dentry->d_name.len, file->f_path.dentry->d_name.name); spin_lock(&oi->ip_lock); @@ -95,8 +98,8 @@ static int ocfs2_file_release(struct inode *inode, struct file *file) struct ocfs2_inode_info *oi = OCFS2_I(inode); mlog_entry("(0x%p, 0x%p, '%.*s')\n", inode, file, - file->f_dentry->d_name.len, - file->f_dentry->d_name.name); + file->f_path.dentry->d_name.len, + file->f_path.dentry->d_name.name); spin_lock(&oi->ip_lock); if (!--oi->ip_open_count) @@ -133,7 +136,58 @@ bail: return (err < 0) ? -EIO : 0; } -int ocfs2_set_inode_size(struct ocfs2_journal_handle *handle, +int ocfs2_should_update_atime(struct inode *inode, + struct vfsmount *vfsmnt) +{ + struct timespec now; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb)) + return 0; + + if ((inode->i_flags & S_NOATIME) || + ((inode->i_sb->s_flags & MS_NODIRATIME) && S_ISDIR(inode->i_mode))) + return 0; + + if ((vfsmnt->mnt_flags & MNT_NOATIME) || + ((vfsmnt->mnt_flags & MNT_NODIRATIME) && S_ISDIR(inode->i_mode))) + return 0; + + now = CURRENT_TIME; + if ((now.tv_sec - inode->i_atime.tv_sec <= osb->s_atime_quantum)) + return 0; + else + return 1; +} + +int ocfs2_update_inode_atime(struct inode *inode, + struct buffer_head *bh) +{ + int ret; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + handle_t *handle; + + mlog_entry_void(); + + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); + if (handle == NULL) { + ret = -ENOMEM; + mlog_errno(ret); + goto out; + } + + inode->i_atime = CURRENT_TIME; + ret = ocfs2_mark_inode_dirty(handle, inode, bh); + if (ret < 0) + mlog_errno(ret); + + ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle); +out: + mlog_exit(ret); + return ret; +} + +int ocfs2_set_inode_size(handle_t *handle, struct inode *inode, struct buffer_head *fe_bh, u64 new_i_size) @@ -162,10 +216,9 @@ static int ocfs2_simple_size_update(struct inode *inode, { int ret; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - struct ocfs2_journal_handle *handle = NULL; + handle_t *handle = NULL; - handle = ocfs2_start_trans(osb, NULL, - OCFS2_INODE_UPDATE_CREDITS); + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); if (handle == NULL) { ret = -ENOMEM; mlog_errno(ret); @@ -177,7 +230,7 @@ static int ocfs2_simple_size_update(struct inode *inode, if (ret < 0) mlog_errno(ret); - ocfs2_commit_trans(handle); + ocfs2_commit_trans(osb, handle); out: return ret; } @@ -188,14 +241,14 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb, u64 new_i_size) { int status; - struct ocfs2_journal_handle *handle; + handle_t *handle; mlog_entry_void(); /* TODO: This needs to actually orphan the inode in this * transaction. */ - handle = ocfs2_start_trans(osb, NULL, OCFS2_INODE_UPDATE_CREDITS); + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); if (IS_ERR(handle)) { status = PTR_ERR(handle); mlog_errno(status); @@ -206,7 +259,7 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb, if (status < 0) mlog_errno(status); - ocfs2_commit_trans(handle); + ocfs2_commit_trans(osb, handle); out: mlog_exit(status); return status; @@ -327,7 +380,7 @@ int ocfs2_do_extend_allocation(struct ocfs2_super *osb, struct inode *inode, u32 clusters_to_add, struct buffer_head *fe_bh, - struct ocfs2_journal_handle *handle, + handle_t *handle, struct ocfs2_alloc_context *data_ac, struct ocfs2_alloc_context *meta_ac, enum ocfs2_alloc_restarted *reason_ret) @@ -432,7 +485,7 @@ static int ocfs2_extend_allocation(struct inode *inode, u32 prev_clusters; struct buffer_head *bh = NULL; struct ocfs2_dinode *fe = NULL; - struct ocfs2_journal_handle *handle = NULL; + handle_t *handle = NULL; struct ocfs2_alloc_context *data_ac = NULL; struct ocfs2_alloc_context *meta_ac = NULL; enum ocfs2_alloc_restarted why; @@ -462,13 +515,6 @@ restart_all: (unsigned long long)OCFS2_I(inode)->ip_blkno, i_size_read(inode), fe->i_clusters, clusters_to_add); - handle = ocfs2_alloc_handle(osb); - if (handle == NULL) { - status = -ENOMEM; - mlog_errno(status); - goto leave; - } - num_free_extents = ocfs2_num_free_extents(osb, inode, fe); @@ -479,10 +525,7 @@ restart_all: } if (!num_free_extents) { - status = ocfs2_reserve_new_metadata(osb, - handle, - fe, - &meta_ac); + status = ocfs2_reserve_new_metadata(osb, fe, &meta_ac); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -490,10 +533,7 @@ restart_all: } } - status = ocfs2_reserve_clusters(osb, - handle, - clusters_to_add, - &data_ac); + status = ocfs2_reserve_clusters(osb, clusters_to_add, &data_ac); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -508,7 +548,7 @@ restart_all: drop_alloc_sem = 1; credits = ocfs2_calc_extend_credits(osb->sb, fe, clusters_to_add); - handle = ocfs2_start_trans(osb, handle, credits); + handle = ocfs2_start_trans(osb, credits); if (IS_ERR(handle)) { status = PTR_ERR(handle); handle = NULL; @@ -588,7 +628,7 @@ leave: drop_alloc_sem = 0; } if (handle) { - ocfs2_commit_trans(handle); + ocfs2_commit_trans(osb, handle); handle = NULL; } if (data_ac) { @@ -623,7 +663,7 @@ static int ocfs2_write_zero_page(struct inode *inode, struct page *page; unsigned long index; unsigned int offset; - struct ocfs2_journal_handle *handle = NULL; + handle_t *handle = NULL; int ret; offset = (size & (PAGE_CACHE_SIZE-1)); /* Within page */ @@ -667,7 +707,7 @@ static int ocfs2_write_zero_page(struct inode *inode, ret = 0; if (handle) - ocfs2_commit_trans(handle); + ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle); out_unlock: unlock_page(page); page_cache_release(page); @@ -691,6 +731,12 @@ static int ocfs2_zero_extend(struct inode *inode, } start_off += sb->s_blocksize; + + /* + * Very large extends have the potential to lock up + * the cpu for extended periods of time. + */ + cond_resched(); } out: @@ -728,31 +774,36 @@ static int ocfs2_extend_file(struct inode *inode, clusters_to_add = ocfs2_clusters_for_bytes(inode->i_sb, new_i_size) - OCFS2_I(inode)->ip_clusters; - if (clusters_to_add) { - /* - * protect the pages that ocfs2_zero_extend is going to - * be pulling into the page cache.. we do this before the - * metadata extend so that we don't get into the situation - * where we've extended the metadata but can't get the data - * lock to zero. - */ - ret = ocfs2_data_lock(inode, 1); - if (ret < 0) { - mlog_errno(ret); - goto out; - } + /* + * protect the pages that ocfs2_zero_extend is going to be + * pulling into the page cache.. we do this before the + * metadata extend so that we don't get into the situation + * where we've extended the metadata but can't get the data + * lock to zero. + */ + ret = ocfs2_data_lock(inode, 1); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + if (clusters_to_add) { ret = ocfs2_extend_allocation(inode, clusters_to_add); if (ret < 0) { mlog_errno(ret); goto out_unlock; } + } - ret = ocfs2_zero_extend(inode, (u64)new_i_size - tail_to_skip); - if (ret < 0) { - mlog_errno(ret); - goto out_unlock; - } + /* + * Call this even if we don't add any clusters to the tree. We + * still need to zero the area between the old i_size and the + * new i_size. + */ + ret = ocfs2_zero_extend(inode, (u64)new_i_size - tail_to_skip); + if (ret < 0) { + mlog_errno(ret); + goto out_unlock; } if (!tail_to_skip) { @@ -764,8 +815,7 @@ static int ocfs2_extend_file(struct inode *inode, } out_unlock: - if (clusters_to_add) /* this is the only case in which we lock */ - ocfs2_data_unlock(inode, 1); + ocfs2_data_unlock(inode, 1); out: return ret; @@ -778,7 +828,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) struct super_block *sb = inode->i_sb; struct ocfs2_super *osb = OCFS2_SB(sb); struct buffer_head *bh = NULL; - struct ocfs2_journal_handle *handle = NULL; + handle_t *handle = NULL; mlog_entry("(0x%p, '%.*s')\n", dentry, dentry->d_name.len, dentry->d_name.name); @@ -814,7 +864,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) } } - status = ocfs2_meta_lock(inode, NULL, &bh, 1); + status = ocfs2_meta_lock(inode, &bh, 1); if (status < 0) { if (status != -ENOENT) mlog_errno(status); @@ -834,7 +884,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) } } - handle = ocfs2_start_trans(osb, NULL, OCFS2_INODE_UPDATE_CREDITS); + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); if (IS_ERR(handle)) { status = PTR_ERR(handle); mlog_errno(status); @@ -852,7 +902,7 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) mlog_errno(status); bail_commit: - ocfs2_commit_trans(handle); + ocfs2_commit_trans(osb, handle); bail_unlock: ocfs2_meta_unlock(inode, 1); bail_unlock_rw: @@ -895,19 +945,41 @@ bail: return err; } +int ocfs2_permission(struct inode *inode, int mask, struct nameidata *nd) +{ + int ret; + + mlog_entry_void(); + + ret = ocfs2_meta_lock(inode, NULL, 0); + if (ret) { + mlog_errno(ret); + goto out; + } + + ret = generic_permission(inode, mask, NULL); + if (ret) + mlog_errno(ret); + + ocfs2_meta_unlock(inode, 0); +out: + mlog_exit(ret); + return ret; +} + static int ocfs2_write_remove_suid(struct inode *inode) { int ret; struct buffer_head *bh = NULL; struct ocfs2_inode_info *oi = OCFS2_I(inode); - struct ocfs2_journal_handle *handle; + handle_t *handle; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct ocfs2_dinode *di; mlog_entry("(Inode %llu, mode 0%o)\n", (unsigned long long)oi->ip_blkno, inode->i_mode); - handle = ocfs2_start_trans(osb, NULL, OCFS2_INODE_UPDATE_CREDITS); + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); if (handle == NULL) { ret = -ENOMEM; mlog_errno(ret); @@ -940,75 +1012,29 @@ static int ocfs2_write_remove_suid(struct inode *inode) out_bh: brelse(bh); out_trans: - ocfs2_commit_trans(handle); + ocfs2_commit_trans(osb, handle); out: mlog_exit(ret); return ret; } -static inline int ocfs2_write_should_remove_suid(struct inode *inode) -{ - mode_t mode = inode->i_mode; - - if (!capable(CAP_FSETID)) { - if (unlikely(mode & S_ISUID)) - return 1; - - if (unlikely((mode & S_ISGID) && (mode & S_IXGRP))) - return 1; - } - return 0; -} - -static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, - const struct iovec *iov, - unsigned long nr_segs, - loff_t pos) +static int ocfs2_prepare_inode_for_write(struct dentry *dentry, + loff_t *ppos, + size_t count, + int appending) { - int ret, rw_level = -1, meta_level = -1, have_alloc_sem = 0; + int ret = 0, meta_level = appending; + struct inode *inode = dentry->d_inode; u32 clusters; - struct file *filp = iocb->ki_filp; - struct inode *inode = filp->f_dentry->d_inode; loff_t newsize, saved_pos; - mlog_entry("(0x%p, %u, '%.*s')\n", filp, - (unsigned int)nr_segs, - filp->f_dentry->d_name.len, - filp->f_dentry->d_name.name); - - /* happy write of zero bytes */ - if (iocb->ki_left == 0) - return 0; - - if (!inode) { - mlog(0, "bad inode\n"); - return -EIO; - } - - mutex_lock(&inode->i_mutex); - /* to match setattr's i_mutex -> i_alloc_sem -> rw_lock ordering */ - if (filp->f_flags & O_DIRECT) { - have_alloc_sem = 1; - down_read(&inode->i_alloc_sem); - } - - /* concurrent O_DIRECT writes are allowed */ - rw_level = (filp->f_flags & O_DIRECT) ? 0 : 1; - ret = ocfs2_rw_lock(inode, rw_level); - if (ret < 0) { - rw_level = -1; - mlog_errno(ret); - goto out; - } - /* * We sample i_size under a read level meta lock to see if our write * is extending the file, if it is we back off and get a write level * meta lock. */ - meta_level = (filp->f_flags & O_APPEND) ? 1 : 0; for(;;) { - ret = ocfs2_meta_lock(inode, NULL, NULL, meta_level); + ret = ocfs2_meta_lock(inode, NULL, meta_level); if (ret < 0) { meta_level = -1; mlog_errno(ret); @@ -1024,7 +1050,7 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, * inode. There's also the dinode i_size state which * can be lost via setattr during extending writes (we * set inode->i_size at the end of a write. */ - if (ocfs2_write_should_remove_suid(inode)) { + if (should_remove_suid(dentry)) { if (meta_level == 0) { ocfs2_meta_unlock(inode, meta_level); meta_level = 1; @@ -1034,19 +1060,19 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, ret = ocfs2_write_remove_suid(inode); if (ret < 0) { mlog_errno(ret); - goto out; + goto out_unlock; } } /* work on a copy of ppos until we're sure that we won't have * to recalculate it due to relocking. */ - if (filp->f_flags & O_APPEND) { + if (appending) { saved_pos = i_size_read(inode); mlog(0, "O_APPEND: inode->i_size=%llu\n", saved_pos); } else { - saved_pos = iocb->ki_pos; + saved_pos = *ppos; } - newsize = iocb->ki_left + saved_pos; + newsize = count + saved_pos; mlog(0, "pos=%lld newsize=%lld cursize=%lld\n", (long long) saved_pos, (long long) newsize, @@ -1079,19 +1105,66 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, if (!clusters) break; - ret = ocfs2_extend_file(inode, NULL, newsize, iocb->ki_left); + ret = ocfs2_extend_file(inode, NULL, newsize, count); if (ret < 0) { if (ret != -ENOSPC) mlog_errno(ret); - goto out; + goto out_unlock; } break; } - /* ok, we're done with i_size and alloc work */ - iocb->ki_pos = saved_pos; + if (appending) + *ppos = saved_pos; + +out_unlock: ocfs2_meta_unlock(inode, meta_level); - meta_level = -1; + +out: + return ret; +} + +static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, + const struct iovec *iov, + unsigned long nr_segs, + loff_t pos) +{ + int ret, rw_level, have_alloc_sem = 0; + struct file *filp = iocb->ki_filp; + struct inode *inode = filp->f_path.dentry->d_inode; + int appending = filp->f_flags & O_APPEND ? 1 : 0; + + mlog_entry("(0x%p, %u, '%.*s')\n", filp, + (unsigned int)nr_segs, + filp->f_path.dentry->d_name.len, + filp->f_path.dentry->d_name.name); + + /* happy write of zero bytes */ + if (iocb->ki_left == 0) + return 0; + + mutex_lock(&inode->i_mutex); + /* to match setattr's i_mutex -> i_alloc_sem -> rw_lock ordering */ + if (filp->f_flags & O_DIRECT) { + have_alloc_sem = 1; + down_read(&inode->i_alloc_sem); + } + + /* concurrent O_DIRECT writes are allowed */ + rw_level = (filp->f_flags & O_DIRECT) ? 0 : 1; + ret = ocfs2_rw_lock(inode, rw_level); + if (ret < 0) { + rw_level = -1; + mlog_errno(ret); + goto out; + } + + ret = ocfs2_prepare_inode_for_write(filp->f_path.dentry, &iocb->ki_pos, + iocb->ki_left, appending); + if (ret < 0) { + mlog_errno(ret); + goto out; + } /* communicate with ocfs2_dio_end_io */ ocfs2_iocb_set_rw_locked(iocb); @@ -1117,8 +1190,6 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, } out: - if (meta_level != -1) - ocfs2_meta_unlock(inode, meta_level); if (have_alloc_sem) up_read(&inode->i_alloc_sem); if (rw_level != -1) @@ -1129,19 +1200,90 @@ out: return ret; } +static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe, + struct file *out, + loff_t *ppos, + size_t len, + unsigned int flags) +{ + int ret; + struct inode *inode = out->f_path.dentry->d_inode; + + mlog_entry("(0x%p, 0x%p, %u, '%.*s')\n", out, pipe, + (unsigned int)len, + out->f_path.dentry->d_name.len, + out->f_path.dentry->d_name.name); + + inode_double_lock(inode, pipe->inode); + + ret = ocfs2_rw_lock(inode, 1); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + + ret = ocfs2_prepare_inode_for_write(out->f_path.dentry, ppos, len, 0); + if (ret < 0) { + mlog_errno(ret); + goto out_unlock; + } + + /* ok, we're done with i_size and alloc work */ + ret = generic_file_splice_write_nolock(pipe, out, ppos, len, flags); + +out_unlock: + ocfs2_rw_unlock(inode, 1); +out: + inode_double_unlock(inode, pipe->inode); + + mlog_exit(ret); + return ret; +} + +static ssize_t ocfs2_file_splice_read(struct file *in, + loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, + unsigned int flags) +{ + int ret = 0; + struct inode *inode = in->f_path.dentry->d_inode; + + mlog_entry("(0x%p, 0x%p, %u, '%.*s')\n", in, pipe, + (unsigned int)len, + in->f_path.dentry->d_name.len, + in->f_path.dentry->d_name.name); + + /* + * See the comment in ocfs2_file_aio_read() + */ + ret = ocfs2_meta_lock(inode, NULL, 0); + if (ret < 0) { + mlog_errno(ret); + goto bail; + } + ocfs2_meta_unlock(inode, 0); + + ret = generic_file_splice_read(in, ppos, pipe, len, flags); + +bail: + mlog_exit(ret); + return ret; +} + static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, const struct iovec *iov, unsigned long nr_segs, loff_t pos) { - int ret = 0, rw_level = -1, have_alloc_sem = 0; + int ret = 0, rw_level = -1, have_alloc_sem = 0, lock_level = 0; struct file *filp = iocb->ki_filp; - struct inode *inode = filp->f_dentry->d_inode; + struct inode *inode = filp->f_path.dentry->d_inode; mlog_entry("(0x%p, %u, '%.*s')\n", filp, (unsigned int)nr_segs, - filp->f_dentry->d_name.len, - filp->f_dentry->d_name.name); + filp->f_path.dentry->d_name.len, + filp->f_path.dentry->d_name.name); if (!inode) { ret = -EINVAL; @@ -1176,12 +1318,12 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, * like i_size. This allows the checks down below * generic_file_aio_read() a chance of actually working. */ - ret = ocfs2_meta_lock(inode, NULL, NULL, 0); + ret = ocfs2_meta_lock_atime(inode, filp->f_vfsmnt, &lock_level); if (ret < 0) { mlog_errno(ret); goto bail; } - ocfs2_meta_unlock(inode, 0); + ocfs2_meta_unlock(inode, lock_level); ret = generic_file_aio_read(iocb, iov, nr_segs, iocb->ki_pos); if (ret == -EINVAL) @@ -1209,11 +1351,13 @@ bail: struct inode_operations ocfs2_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, + .permission = ocfs2_permission, }; struct inode_operations ocfs2_special_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, + .permission = ocfs2_permission, }; const struct file_operations ocfs2_fops = { @@ -1227,6 +1371,8 @@ const struct file_operations ocfs2_fops = { .aio_read = ocfs2_file_aio_read, .aio_write = ocfs2_file_aio_write, .ioctl = ocfs2_ioctl, + .splice_read = ocfs2_file_splice_read, + .splice_write = ocfs2_file_splice_write, }; const struct file_operations ocfs2_dops = { |