From 9f570b8d48b6677b5557d86fb3ca148215e295f2 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 25 Jan 2011 12:42:37 -0500 Subject: [PATCH 01/45] Btrfs: fix formatting in file.c Sorry, but these were bugging me. Just cleanup some of the formatting in file.c. Signed-off-by: Josef Bacik --- fs/btrfs/file.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index f447b783bb84..4d4975592668 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -88,9 +88,8 @@ static noinline int btrfs_copy_from_user(loff_t pos, int num_pages, total_copied += copied; /* Return to btrfs_file_aio_write to fault page */ - if (unlikely(copied == 0)) { + if (unlikely(copied == 0)) break; - } if (unlikely(copied < PAGE_CACHE_SIZE - offset)) { offset += copied; @@ -162,13 +161,14 @@ static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans, ClearPageChecked(p); set_page_dirty(p); } - if (end_pos > isize) { + + /* + * we've only changed i_size in ram, and we haven't updated + * the disk i_size. There is no need to log the inode + * at this time. + */ + if (end_pos > isize) i_size_write(inode, end_pos); - /* we've only changed i_size in ram, and we haven't updated - * the disk i_size. There is no need to log the inode - * at this time. - */ - } return 0; } From d0215f3e5ebb5803cd6ec067b10c5e00a3ad7cfc Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 25 Jan 2011 14:57:24 -0500 Subject: [PATCH 02/45] Btrfs: simplify our write path Our aio_write function is huge and kind of hard to follow at times. So this patch fixes this by breaking out the buffered and direct write paths out into seperate functions so it's a little clearer what's going on. I've also fixed some wrong typing that we had and added the ability to handle getting an error back from btrfs_set_extent_delalloc. Tested this with xfstests and everything came out fine. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/file.c | 357 ++++++++++++++++++++++++------------------------ 1 file changed, 181 insertions(+), 176 deletions(-) diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 4d4975592668..f2a80e570a6c 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -45,14 +45,14 @@ * and be replaced with calls into generic code. */ static noinline int btrfs_copy_from_user(loff_t pos, int num_pages, - int write_bytes, + size_t write_bytes, struct page **prepared_pages, struct iov_iter *i) { size_t copied = 0; + size_t total_copied = 0; int pg = 0; int offset = pos & (PAGE_CACHE_SIZE - 1); - int total_copied = 0; while (write_bytes > 0) { size_t count = min_t(size_t, @@ -129,13 +129,12 @@ static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages) * this also makes the decision about creating an inline extent vs * doing real data extents, marking pages dirty and delalloc as required. */ -static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct file *file, - struct page **pages, - size_t num_pages, - loff_t pos, - size_t write_bytes) +static noinline int dirty_and_release_pages(struct btrfs_root *root, + struct file *file, + struct page **pages, + size_t num_pages, + loff_t pos, + size_t write_bytes) { int err = 0; int i; @@ -153,7 +152,8 @@ static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans, end_of_last_block = start_pos + num_bytes - 1; err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block, NULL); - BUG_ON(err); + if (err) + return err; for (i = 0; i < num_pages; i++) { struct page *p = pages[i]; @@ -896,127 +896,38 @@ fail: } -static ssize_t btrfs_file_aio_write(struct kiocb *iocb, - const struct iovec *iov, - unsigned long nr_segs, loff_t pos) +static noinline ssize_t __btrfs_buffered_write(struct file *file, + struct iov_iter *i, + loff_t pos) { - struct file *file = iocb->ki_filp; struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; struct page **pages = NULL; - struct iov_iter i; - loff_t *ppos = &iocb->ki_pos; - loff_t start_pos; - ssize_t num_written = 0; - ssize_t err = 0; - size_t count; - size_t ocount; - int ret = 0; - int nrptrs; unsigned long first_index; unsigned long last_index; - int will_write; - int buffered = 0; - int copied = 0; - int dirty_pages = 0; + size_t num_written = 0; + int nrptrs; + int ret; - will_write = ((file->f_flags & O_DSYNC) || IS_SYNC(inode) || - (file->f_flags & O_DIRECT)); - - start_pos = pos; - - vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); - - mutex_lock(&inode->i_mutex); - - err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ); - if (err) - goto out; - count = ocount; - - current->backing_dev_info = inode->i_mapping->backing_dev_info; - err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode)); - if (err) - goto out; - - if (count == 0) - goto out; - - err = file_remove_suid(file); - if (err) - goto out; - - /* - * If BTRFS flips readonly due to some impossible error - * (fs_info->fs_state now has BTRFS_SUPER_FLAG_ERROR), - * although we have opened a file as writable, we have - * to stop this write operation to ensure FS consistency. - */ - if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { - err = -EROFS; - goto out; - } - - file_update_time(file); - BTRFS_I(inode)->sequence++; - - if (unlikely(file->f_flags & O_DIRECT)) { - num_written = generic_file_direct_write(iocb, iov, &nr_segs, - pos, ppos, count, - ocount); - /* - * the generic O_DIRECT will update in-memory i_size after the - * DIOs are done. But our endio handlers that update the on - * disk i_size never update past the in memory i_size. So we - * need one more update here to catch any additions to the - * file - */ - if (inode->i_size != BTRFS_I(inode)->disk_i_size) { - btrfs_ordered_update_i_size(inode, inode->i_size, NULL); - mark_inode_dirty(inode); - } - - if (num_written < 0) { - ret = num_written; - num_written = 0; - goto out; - } else if (num_written == count) { - /* pick up pos changes done by the generic code */ - pos = *ppos; - goto out; - } - /* - * We are going to do buffered for the rest of the range, so we - * need to make sure to invalidate the buffered pages when we're - * done. - */ - buffered = 1; - pos += num_written; - } - - iov_iter_init(&i, iov, nr_segs, count, num_written); - nrptrs = min((iov_iter_count(&i) + PAGE_CACHE_SIZE - 1) / + nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE, PAGE_CACHE_SIZE / (sizeof(struct page *))); pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL); - if (!pages) { - ret = -ENOMEM; - goto out; - } - - /* generic_write_checks can change our pos */ - start_pos = pos; + if (!pages) + return -ENOMEM; first_index = pos >> PAGE_CACHE_SHIFT; - last_index = (pos + iov_iter_count(&i)) >> PAGE_CACHE_SHIFT; + last_index = (pos + iov_iter_count(i)) >> PAGE_CACHE_SHIFT; - while (iov_iter_count(&i) > 0) { + while (iov_iter_count(i) > 0) { size_t offset = pos & (PAGE_CACHE_SIZE - 1); - size_t write_bytes = min(iov_iter_count(&i), + size_t write_bytes = min(iov_iter_count(i), nrptrs * (size_t)PAGE_CACHE_SIZE - offset); size_t num_pages = (write_bytes + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + size_t dirty_pages; + size_t copied; WARN_ON(num_pages > nrptrs); memset(pages, 0, sizeof(struct page *) * nrptrs); @@ -1025,15 +936,15 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, * Fault pages before locking them in prepare_pages * to avoid recursive lock */ - if (unlikely(iov_iter_fault_in_readable(&i, write_bytes))) { + if (unlikely(iov_iter_fault_in_readable(i, write_bytes))) { ret = -EFAULT; - goto out; + break; } ret = btrfs_delalloc_reserve_space(inode, num_pages << PAGE_CACHE_SHIFT); if (ret) - goto out; + break; ret = prepare_pages(root, file, pages, num_pages, pos, first_index, last_index, @@ -1041,11 +952,11 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, if (ret) { btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT); - goto out; + break; } copied = btrfs_copy_from_user(pos, num_pages, - write_bytes, pages, &i); + write_bytes, pages, i); /* * if we have trouble faulting in the pages, fall @@ -1061,6 +972,13 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + /* + * If we had a short copy we need to release the excess delaloc + * bytes we reserved. We need to increment outstanding_extents + * because btrfs_delalloc_release_space will decrement it, but + * we still have an outstanding extent for the chunk we actually + * managed to copy. + */ if (num_pages > dirty_pages) { if (copied > 0) atomic_inc( @@ -1071,39 +989,157 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, } if (copied > 0) { - dirty_and_release_pages(NULL, root, file, pages, - dirty_pages, pos, copied); + ret = dirty_and_release_pages(root, file, pages, + dirty_pages, pos, + copied); + if (ret) { + btrfs_delalloc_release_space(inode, + dirty_pages << PAGE_CACHE_SHIFT); + btrfs_drop_pages(pages, num_pages); + break; + } } btrfs_drop_pages(pages, num_pages); - if (copied > 0) { - if (will_write) { - filemap_fdatawrite_range(inode->i_mapping, pos, - pos + copied - 1); - } else { - balance_dirty_pages_ratelimited_nr( - inode->i_mapping, - dirty_pages); - if (dirty_pages < - (root->leafsize >> PAGE_CACHE_SHIFT) + 1) - btrfs_btree_balance_dirty(root, 1); - btrfs_throttle(root); - } - } + cond_resched(); + + balance_dirty_pages_ratelimited_nr(inode->i_mapping, + dirty_pages); + if (dirty_pages < (root->leafsize >> PAGE_CACHE_SHIFT) + 1) + btrfs_btree_balance_dirty(root, 1); + btrfs_throttle(root); pos += copied; num_written += copied; - - cond_resched(); } -out: - mutex_unlock(&inode->i_mutex); - if (ret) - err = ret; kfree(pages); - *ppos = pos; + + return num_written ? num_written : ret; +} + +static ssize_t __btrfs_direct_write(struct kiocb *iocb, + const struct iovec *iov, + unsigned long nr_segs, loff_t pos, + loff_t *ppos, size_t count, size_t ocount) +{ + struct file *file = iocb->ki_filp; + struct inode *inode = fdentry(file)->d_inode; + struct iov_iter i; + ssize_t written; + ssize_t written_buffered; + loff_t endbyte; + int err; + + written = generic_file_direct_write(iocb, iov, &nr_segs, pos, ppos, + count, ocount); + + /* + * the generic O_DIRECT will update in-memory i_size after the + * DIOs are done. But our endio handlers that update the on + * disk i_size never update past the in memory i_size. So we + * need one more update here to catch any additions to the + * file + */ + if (inode->i_size != BTRFS_I(inode)->disk_i_size) { + btrfs_ordered_update_i_size(inode, inode->i_size, NULL); + mark_inode_dirty(inode); + } + + if (written < 0 || written == count) + return written; + + pos += written; + count -= written; + iov_iter_init(&i, iov, nr_segs, count, written); + written_buffered = __btrfs_buffered_write(file, &i, pos); + if (written_buffered < 0) { + err = written_buffered; + goto out; + } + endbyte = pos + written_buffered - 1; + err = filemap_write_and_wait_range(file->f_mapping, pos, endbyte); + if (err) + goto out; + written += written_buffered; + *ppos = pos + written_buffered; + invalidate_mapping_pages(file->f_mapping, pos >> PAGE_CACHE_SHIFT, + endbyte >> PAGE_CACHE_SHIFT); +out: + return written ? written : err; +} + +static ssize_t btrfs_file_aio_write(struct kiocb *iocb, + const struct iovec *iov, + unsigned long nr_segs, loff_t pos) +{ + struct file *file = iocb->ki_filp; + struct inode *inode = fdentry(file)->d_inode; + struct btrfs_root *root = BTRFS_I(inode)->root; + loff_t *ppos = &iocb->ki_pos; + ssize_t num_written = 0; + ssize_t err = 0; + size_t count, ocount; + + vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); + + mutex_lock(&inode->i_mutex); + + err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + count = ocount; + + current->backing_dev_info = inode->i_mapping->backing_dev_info; + err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode)); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + + if (count == 0) { + mutex_unlock(&inode->i_mutex); + goto out; + } + + err = file_remove_suid(file); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } + + /* + * If BTRFS flips readonly due to some impossible error + * (fs_info->fs_state now has BTRFS_SUPER_FLAG_ERROR), + * although we have opened a file as writable, we have + * to stop this write operation to ensure FS consistency. + */ + if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { + mutex_unlock(&inode->i_mutex); + err = -EROFS; + goto out; + } + + file_update_time(file); + BTRFS_I(inode)->sequence++; + + if (unlikely(file->f_flags & O_DIRECT)) { + num_written = __btrfs_direct_write(iocb, iov, nr_segs, + pos, ppos, count, ocount); + } else { + struct iov_iter i; + + iov_iter_init(&i, iov, nr_segs, count, num_written); + + num_written = __btrfs_buffered_write(file, &i, pos); + if (num_written > 0) + *ppos = pos + num_written; + } + + mutex_unlock(&inode->i_mutex); /* * we want to make sure fsync finds this change @@ -1118,43 +1154,12 @@ out: * one running right now. */ BTRFS_I(inode)->last_trans = root->fs_info->generation + 1; - - if (num_written > 0 && will_write) { - struct btrfs_trans_handle *trans; - - err = btrfs_wait_ordered_range(inode, start_pos, num_written); - if (err) + if (num_written > 0 || num_written == -EIOCBQUEUED) { + err = generic_write_sync(file, pos, num_written); + if (err < 0 && num_written > 0) num_written = err; - - if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) { - trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) { - num_written = PTR_ERR(trans); - goto done; - } - mutex_lock(&inode->i_mutex); - ret = btrfs_log_dentry_safe(trans, root, - file->f_dentry); - mutex_unlock(&inode->i_mutex); - if (ret == 0) { - ret = btrfs_sync_log(trans, root); - if (ret == 0) - btrfs_end_transaction(trans, root); - else - btrfs_commit_transaction(trans, root); - } else if (ret != BTRFS_NO_LOG_SYNC) { - btrfs_commit_transaction(trans, root); - } else { - btrfs_end_transaction(trans, root); - } - } - if (file->f_flags & O_DIRECT && buffered) { - invalidate_mapping_pages(inode->i_mapping, - start_pos >> PAGE_CACHE_SHIFT, - (start_pos + num_written - 1) >> PAGE_CACHE_SHIFT); - } } -done: +out: current->backing_dev_info = NULL; return num_written ? num_written : err; } From 4a64001f0047956e283f7ada9843dfc3f3b5d8c8 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 25 Jan 2011 15:10:08 -0500 Subject: [PATCH 03/45] Btrfs: fix how we deal with the pages array in the write path Really we don't need to memset the pages array at all, since we know how many pages we're going to use in the array and pass that around. So don't memset, just trust we're not idiots and we pass num_pages around properly. Signed-off-by: Josef Bacik --- fs/btrfs/file.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index f2a80e570a6c..24a19c2743ca 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -108,8 +108,6 @@ static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages) { size_t i; for (i = 0; i < num_pages; i++) { - if (!pages[i]) - break; /* page checked is some magic around finding pages that * have been modified without going through btrfs_set_page_dirty * clear it here @@ -824,7 +822,6 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, return err; } - memset(pages, 0, num_pages * sizeof(struct page *)); again: for (i = 0; i < num_pages; i++) { pages[i] = grab_cache_page(inode->i_mapping, index + i); @@ -930,7 +927,6 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, size_t copied; WARN_ON(num_pages > nrptrs); - memset(pages, 0, sizeof(struct page *) * nrptrs); /* * Fault pages before locking them in prepare_pages @@ -946,6 +942,11 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, if (ret) break; + /* + * This is going to setup the pages array with the number of + * pages we want, so we don't really need to worry about the + * contents of pages from loop to loop + */ ret = prepare_pages(root, file, pages, num_pages, pos, first_index, last_index, write_bytes); From 57a45ced94fe48a701361d64230fc16eefa189dd Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 25 Jan 2011 16:30:38 -0500 Subject: [PATCH 04/45] Btrfs: change reserved_extents to an atomic_t We track delayed allocation per inodes via 2 counters, one is outstanding_extents and reserved_extents. Outstanding_extents is already an atomic_t, but reserved_extents is not and is protected by a spinlock. So convert this to an atomic_t and instead of using a spinlock, use atomic_cmpxchg when releasing delalloc bytes. This makes our inode 72 bytes smaller, and reduces locking overhead (albiet it was minimal to begin with). Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/btrfs_inode.h | 3 +-- fs/btrfs/extent-tree.c | 42 ++++++++++++++++++++++++++---------------- fs/btrfs/inode.c | 5 ++--- 3 files changed, 29 insertions(+), 21 deletions(-) diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index ccc991c542df..57c3bb2884ce 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -136,9 +136,8 @@ struct btrfs_inode { * items we think we'll end up using, and reserved_extents is the number * of extent items we've reserved metadata for. */ - spinlock_t accounting_lock; atomic_t outstanding_extents; - int reserved_extents; + atomic_t reserved_extents; /* * ordered_data_close is set by truncate when a file that used diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 7b3089b5c2df..27376c97d85f 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3996,6 +3996,7 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; u64 to_reserve; int nr_extents; + int reserved_extents; int ret; if (btrfs_transaction_in_commit(root->fs_info)) @@ -4003,25 +4004,24 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) num_bytes = ALIGN(num_bytes, root->sectorsize); - spin_lock(&BTRFS_I(inode)->accounting_lock); nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents) + 1; - if (nr_extents > BTRFS_I(inode)->reserved_extents) { - nr_extents -= BTRFS_I(inode)->reserved_extents; + reserved_extents = atomic_read(&BTRFS_I(inode)->reserved_extents); + + if (nr_extents > reserved_extents) { + nr_extents -= reserved_extents; to_reserve = calc_trans_metadata_size(root, nr_extents); } else { nr_extents = 0; to_reserve = 0; } - spin_unlock(&BTRFS_I(inode)->accounting_lock); + to_reserve += calc_csum_metadata_size(inode, num_bytes); ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1); if (ret) return ret; - spin_lock(&BTRFS_I(inode)->accounting_lock); - BTRFS_I(inode)->reserved_extents += nr_extents; + atomic_add(nr_extents, &BTRFS_I(inode)->reserved_extents); atomic_inc(&BTRFS_I(inode)->outstanding_extents); - spin_unlock(&BTRFS_I(inode)->accounting_lock); block_rsv_add_bytes(block_rsv, to_reserve, 1); @@ -4036,20 +4036,30 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) struct btrfs_root *root = BTRFS_I(inode)->root; u64 to_free; int nr_extents; + int reserved_extents; num_bytes = ALIGN(num_bytes, root->sectorsize); atomic_dec(&BTRFS_I(inode)->outstanding_extents); WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents) < 0); - spin_lock(&BTRFS_I(inode)->accounting_lock); - nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents); - if (nr_extents < BTRFS_I(inode)->reserved_extents) { - nr_extents = BTRFS_I(inode)->reserved_extents - nr_extents; - BTRFS_I(inode)->reserved_extents -= nr_extents; - } else { - nr_extents = 0; - } - spin_unlock(&BTRFS_I(inode)->accounting_lock); + reserved_extents = atomic_read(&BTRFS_I(inode)->reserved_extents); + do { + int old, new; + + nr_extents = atomic_read(&BTRFS_I(inode)->outstanding_extents); + if (nr_extents >= reserved_extents) { + nr_extents = 0; + break; + } + old = reserved_extents; + nr_extents = reserved_extents - nr_extents; + new = reserved_extents - nr_extents; + old = atomic_cmpxchg(&BTRFS_I(inode)->reserved_extents, + reserved_extents, new); + if (likely(old == reserved_extents)) + break; + reserved_extents = old; + } while (1); to_free = calc_csum_metadata_size(inode, num_bytes); if (nr_extents > 0) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 9007bbd01dbf..d97b69afbbfb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6632,9 +6632,8 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) ei->index_cnt = (u64)-1; ei->last_unlink_trans = 0; - spin_lock_init(&ei->accounting_lock); atomic_set(&ei->outstanding_extents, 0); - ei->reserved_extents = 0; + atomic_set(&ei->reserved_extents, 0); ei->ordered_data_close = 0; ei->orphan_meta_reserved = 0; @@ -6670,7 +6669,7 @@ void btrfs_destroy_inode(struct inode *inode) WARN_ON(!list_empty(&inode->i_dentry)); WARN_ON(inode->i_data.nrpages); WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents)); - WARN_ON(BTRFS_I(inode)->reserved_extents); + WARN_ON(atomic_read(&BTRFS_I(inode)->reserved_extents)); /* * This can happen where we create an inode, but somebody else also From dc89e9824464e91fa0b06267864ceabe3186fd8b Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 28 Jan 2011 17:05:48 -0500 Subject: [PATCH 05/45] Btrfs: use a slab for the free space entries Since we alloc/free free space entries a whole lot, lets use a slab to keep track of them. This makes some of my tests slightly faster. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/ctree.h | 1 + fs/btrfs/free-space-cache.c | 34 ++++++++++++++++++---------------- fs/btrfs/inode.c | 10 ++++++++++ 3 files changed, 29 insertions(+), 16 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 7f78cc78fdd0..2c98d209e6ac 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -40,6 +40,7 @@ extern struct kmem_cache *btrfs_trans_handle_cachep; extern struct kmem_cache *btrfs_transaction_cachep; extern struct kmem_cache *btrfs_bit_radix_cachep; extern struct kmem_cache *btrfs_path_cachep; +extern struct kmem_cache *btrfs_free_space_cachep; struct btrfs_ordered_sum; #define BTRFS_MAGIC "_BHRfS_M" diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index a0390657451b..0282033041e1 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -393,7 +393,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, break; need_loop = 1; - e = kzalloc(sizeof(struct btrfs_free_space), GFP_NOFS); + e = kmem_cache_zalloc(btrfs_free_space_cachep, + GFP_NOFS); if (!e) { kunmap(page); unlock_page(page); @@ -405,7 +406,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, e->bytes = le64_to_cpu(entry->bytes); if (!e->bytes) { kunmap(page); - kfree(e); + kmem_cache_free(btrfs_free_space_cachep, e); unlock_page(page); page_cache_release(page); goto free_cache; @@ -420,7 +421,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); if (!e->bitmap) { kunmap(page); - kfree(e); + kmem_cache_free( + btrfs_free_space_cachep, e); unlock_page(page); page_cache_release(page); goto free_cache; @@ -1187,7 +1189,7 @@ static void free_bitmap(struct btrfs_block_group_cache *block_group, { unlink_free_space(block_group, bitmap_info); kfree(bitmap_info->bitmap); - kfree(bitmap_info); + kmem_cache_free(btrfs_free_space_cachep, bitmap_info); block_group->total_bitmaps--; recalculate_thresholds(block_group); } @@ -1342,8 +1344,8 @@ new_bitmap: /* no pre-allocated info, allocate a new one */ if (!info) { - info = kzalloc(sizeof(struct btrfs_free_space), - GFP_NOFS); + info = kmem_cache_zalloc(btrfs_free_space_cachep, + GFP_NOFS); if (!info) { spin_lock(&block_group->tree_lock); ret = -ENOMEM; @@ -1365,7 +1367,7 @@ out: if (info) { if (info->bitmap) kfree(info->bitmap); - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); } return ret; @@ -1398,7 +1400,7 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group, else __unlink_free_space(block_group, right_info); info->bytes += right_info->bytes; - kfree(right_info); + kmem_cache_free(btrfs_free_space_cachep, right_info); merged = true; } @@ -1410,7 +1412,7 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group, __unlink_free_space(block_group, left_info); info->offset = left_info->offset; info->bytes += left_info->bytes; - kfree(left_info); + kmem_cache_free(btrfs_free_space_cachep, left_info); merged = true; } @@ -1423,7 +1425,7 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, struct btrfs_free_space *info; int ret = 0; - info = kzalloc(sizeof(struct btrfs_free_space), GFP_NOFS); + info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS); if (!info) return -ENOMEM; @@ -1450,7 +1452,7 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, link: ret = link_free_space(block_group, info); if (ret) - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); out: spin_unlock(&block_group->tree_lock); @@ -1520,7 +1522,7 @@ again: kfree(info->bitmap); block_group->total_bitmaps--; } - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); goto out_lock; } @@ -1556,7 +1558,7 @@ again: /* the hole we're creating ends at the end * of the info struct, just free the info */ - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); } spin_unlock(&block_group->tree_lock); @@ -1689,7 +1691,7 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) unlink_free_space(block_group, info); if (info->bitmap) kfree(info->bitmap); - kfree(info); + kmem_cache_free(btrfs_free_space_cachep, info); if (need_resched()) { spin_unlock(&block_group->tree_lock); cond_resched(); @@ -1722,7 +1724,7 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, entry->offset += bytes; entry->bytes -= bytes; if (!entry->bytes) - kfree(entry); + kmem_cache_free(btrfs_free_space_cachep, entry); else link_free_space(block_group, entry); } @@ -1884,7 +1886,7 @@ out: block_group->free_space -= bytes; if (entry->bytes == 0) { block_group->free_extents--; - kfree(entry); + kmem_cache_free(btrfs_free_space_cachep, entry); } spin_unlock(&block_group->tree_lock); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d97b69afbbfb..2d2e079713d7 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -50,6 +50,7 @@ #include "tree-log.h" #include "compression.h" #include "locking.h" +#include "free-space-cache.h" struct btrfs_iget_args { u64 ino; @@ -70,6 +71,7 @@ static struct kmem_cache *btrfs_inode_cachep; struct kmem_cache *btrfs_trans_handle_cachep; struct kmem_cache *btrfs_transaction_cachep; struct kmem_cache *btrfs_path_cachep; +struct kmem_cache *btrfs_free_space_cachep; #define S_SHIFT 12 static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { @@ -6761,6 +6763,8 @@ void btrfs_destroy_cachep(void) kmem_cache_destroy(btrfs_transaction_cachep); if (btrfs_path_cachep) kmem_cache_destroy(btrfs_path_cachep); + if (btrfs_free_space_cachep) + kmem_cache_destroy(btrfs_free_space_cachep); } int btrfs_init_cachep(void) @@ -6789,6 +6793,12 @@ int btrfs_init_cachep(void) if (!btrfs_path_cachep) goto fail; + btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space_cache", + sizeof(struct btrfs_free_space), 0, + SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + if (!btrfs_free_space_cachep) + goto fail; + return 0; fail: btrfs_destroy_cachep(); From a41ad394a03b802497958d7c98a9dcf607266645 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 31 Jan 2011 15:30:16 -0500 Subject: [PATCH 06/45] Btrfs: convert to the new truncate sequence ->truncate() is going away, instead all of the work needs to be done in ->setattr(). So this converts us over to do this. It's fairly straightforward, just get rid of our .truncate inode operation and call btrfs_truncate() directly from btrfs_setsize. This works out better for us since truncate can technically return ENOSPC, and before we had no way of letting anybody know. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/ctree.h | 2 +- fs/btrfs/file.c | 5 +-- fs/btrfs/inode.c | 82 ++++++++++++++++++++++-------------------------- 3 files changed, 41 insertions(+), 48 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 2c98d209e6ac..34142d5647df 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2537,7 +2537,7 @@ void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending); void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, struct btrfs_root *root); -int btrfs_cont_expand(struct inode *inode, loff_t size); +int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size); int btrfs_invalidate_inodes(struct btrfs_root *root); void btrfs_add_delayed_iput(struct inode *inode); void btrfs_run_delayed_iputs(struct btrfs_root *root); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 24a19c2743ca..3786eca2a905 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -817,7 +817,7 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, last_pos = ((u64)index + num_pages) << PAGE_CACHE_SHIFT; if (start_pos > inode->i_size) { - err = btrfs_cont_expand(inode, start_pos); + err = btrfs_cont_expand(inode, i_size_read(inode), start_pos); if (err) return err; } @@ -1330,7 +1330,8 @@ static long btrfs_fallocate(struct file *file, int mode, goto out; if (alloc_start > inode->i_size) { - ret = btrfs_cont_expand(inode, alloc_start); + ret = btrfs_cont_expand(inode, i_size_read(inode), + alloc_start); if (ret) goto out; } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2d2e079713d7..3662ffec17d9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -84,7 +84,8 @@ static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { [S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK, }; -static void btrfs_truncate(struct inode *inode); +static int btrfs_setsize(struct inode *inode, loff_t newsize); +static int btrfs_truncate(struct inode *inode); static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end); static noinline int cow_file_range(struct inode *inode, struct page *locked_page, @@ -2371,6 +2372,11 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) /* if we have links, this was a truncate, lets do that */ if (inode->i_nlink) { + if (!S_ISREG(inode->i_mode)) { + WARN_ON(1); + iput(inode); + continue; + } nr_truncate++; btrfs_truncate(inode); } else { @@ -3538,7 +3544,7 @@ out: return ret; } -int btrfs_cont_expand(struct inode *inode, loff_t size) +int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -3546,7 +3552,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) struct extent_map *em = NULL; struct extent_state *cached_state = NULL; u64 mask = root->sectorsize - 1; - u64 hole_start = (inode->i_size + mask) & ~mask; + u64 hole_start = (oldsize + mask) & ~mask; u64 block_end = (size + mask) & ~mask; u64 last_byte; u64 cur_offset; @@ -3617,27 +3623,17 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) return err; } -static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) +static int btrfs_setsize(struct inode *inode, loff_t newsize) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; + loff_t oldsize = i_size_read(inode); unsigned long nr; int ret; - if (attr->ia_size == inode->i_size) + if (newsize == oldsize) return 0; - if (attr->ia_size > inode->i_size) { - unsigned long limit; - limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur; - if (attr->ia_size > inode->i_sb->s_maxbytes) - return -EFBIG; - if (limit != RLIM_INFINITY && attr->ia_size > limit) { - send_sig(SIGXFSZ, current, 0); - return -EFBIG; - } - } - trans = btrfs_start_transaction(root, 5); if (IS_ERR(trans)) return PTR_ERR(trans); @@ -3651,16 +3647,16 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); - if (attr->ia_size > inode->i_size) { - ret = btrfs_cont_expand(inode, attr->ia_size); + if (newsize > oldsize) { + i_size_write(inode, newsize); + btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); + truncate_pagecache(inode, oldsize, newsize); + ret = btrfs_cont_expand(inode, oldsize, newsize); if (ret) { - btrfs_truncate(inode); + btrfs_setsize(inode, oldsize); return ret; } - i_size_write(inode, attr->ia_size); - btrfs_ordered_update_i_size(inode, inode->i_size, NULL); - trans = btrfs_start_transaction(root, 0); BUG_ON(IS_ERR(trans)); btrfs_set_trans_block_group(trans, inode); @@ -3676,22 +3672,22 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); - return 0; + } else { + + /* + * We're truncating a file that used to have good data down to + * zero. Make sure it gets into the ordered flush list so that + * any new writes get down to disk quickly. + */ + if (newsize == 0) + BTRFS_I(inode)->ordered_data_close = 1; + + /* we don't support swapfiles, so vmtruncate shouldn't fail */ + truncate_setsize(inode, newsize); + ret = btrfs_truncate(inode); } - /* - * We're truncating a file that used to have good data down to - * zero. Make sure it gets into the ordered flush list so that - * any new writes get down to disk quickly. - */ - if (attr->ia_size == 0) - BTRFS_I(inode)->ordered_data_close = 1; - - /* we don't support swapfiles, so vmtruncate shouldn't fail */ - ret = vmtruncate(inode, attr->ia_size); - BUG_ON(ret); - - return 0; + return ret; } static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) @@ -3708,7 +3704,7 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) return err; if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) { - err = btrfs_setattr_size(inode, attr); + err = btrfs_setsize(inode, attr->ia_size); if (err) return err; } @@ -6478,7 +6474,7 @@ out: return ret; } -static void btrfs_truncate(struct inode *inode) +static int btrfs_truncate(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; int ret; @@ -6486,14 +6482,9 @@ static void btrfs_truncate(struct inode *inode) unsigned long nr; u64 mask = root->sectorsize - 1; - if (!S_ISREG(inode->i_mode)) { - WARN_ON(1); - return; - } - ret = btrfs_truncate_page(inode->i_mapping, inode->i_size); if (ret) - return; + return ret; btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1); btrfs_ordered_update_i_size(inode, inode->i_size, NULL); @@ -6568,6 +6559,8 @@ static void btrfs_truncate(struct inode *inode) ret = btrfs_end_transaction_throttle(trans, root); BUG_ON(ret); btrfs_btree_balance_dirty(root, nr); + + return ret; } /* @@ -7367,7 +7360,6 @@ static const struct address_space_operations btrfs_symlink_aops = { }; static const struct inode_operations btrfs_file_inode_operations = { - .truncate = btrfs_truncate, .getattr = btrfs_getattr, .setattr = btrfs_setattr, .setxattr = btrfs_setxattr, From 3893e33b0bebee2f67d96b6c15259dc884523c20 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 31 Jan 2011 16:03:11 -0500 Subject: [PATCH 07/45] Btrfs: cleanup error handling in the truncate path Now that we can handle having errors in the truncate path lets make sure we return errors instead of doing BUG_ON() and such. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 64 ++++++++++++++++++++++++++++++++++-------------- 1 file changed, 45 insertions(+), 19 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3662ffec17d9..d83025063ee7 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3597,13 +3597,15 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) err = btrfs_drop_extents(trans, inode, cur_offset, cur_offset + hole_size, &hint_byte, 1); - BUG_ON(err); + if (err) + break; err = btrfs_insert_file_extent(trans, root, inode->i_ino, cur_offset, 0, 0, hole_size, 0, hole_size, 0, 0, 0); - BUG_ON(err); + if (err) + break; btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); @@ -3641,7 +3643,10 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) btrfs_set_trans_block_group(trans, inode); ret = btrfs_orphan_add(trans, inode); - BUG_ON(ret); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; + } nr = trans->blocks_used; btrfs_end_transaction(trans, root); @@ -3658,17 +3663,24 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) } trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); + btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; BUG_ON(!trans->block_rsv); + /* + * If this fails just leave the orphan item so that it can get + * cleaned up next time we mount. + */ ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); - if (inode->i_nlink > 0) { - ret = btrfs_orphan_del(trans, inode); - BUG_ON(ret); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; } + if (inode->i_nlink > 0) + ret = btrfs_orphan_del(trans, inode); nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -6478,6 +6490,7 @@ static int btrfs_truncate(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; int ret; + int err = 0; struct btrfs_trans_handle *trans; unsigned long nr; u64 mask = root->sectorsize - 1; @@ -6490,7 +6503,8 @@ static int btrfs_truncate(struct inode *inode) btrfs_ordered_update_i_size(inode, inode->i_size, NULL); trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; @@ -6517,29 +6531,38 @@ static int btrfs_truncate(struct inode *inode) while (1) { if (!trans) { trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; } ret = btrfs_block_rsv_check(trans, root, root->orphan_block_rsv, 0, 5); - if (ret) { - BUG_ON(ret != -EAGAIN); + if (ret == -EAGAIN) { ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); + if (ret) + return ret; trans = NULL; continue; + } else if (ret) { + err = ret; + break; } ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, BTRFS_EXTENT_DATA_KEY); - if (ret != -EAGAIN) + if (ret != -EAGAIN) { + err = ret; break; + } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + if (ret) { + err = ret; + break; + } nr = trans->blocks_used; btrfs_end_transaction(trans, root); @@ -6549,18 +6572,21 @@ static int btrfs_truncate(struct inode *inode) if (ret == 0 && inode->i_nlink > 0) { ret = btrfs_orphan_del(trans, inode); - BUG_ON(ret); + if (ret) + err = ret; } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + if (ret && !err) + err = ret; nr = trans->blocks_used; ret = btrfs_end_transaction_throttle(trans, root); - BUG_ON(ret); + if (ret && !err) + err = ret; btrfs_btree_balance_dirty(root, nr); - return ret; + return err; } /* From 66b4ffd110f9b48b8d8c1319ee446b53b8d073bf Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 31 Jan 2011 16:22:42 -0500 Subject: [PATCH 08/45] Btrfs: handle errors in btrfs_orphan_cleanup If we cannot truncate an inode for some reason we will never delete the orphan item associated with that inode, which means that we will loop forever in btrfs_orphan_cleanup. Instead of doing this just return error so we fail to mount. It sucks, but hey it's better than hanging. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/ctree.h | 2 +- fs/btrfs/disk-io.c | 15 +++++++++++--- fs/btrfs/extent-tree.c | 3 ++- fs/btrfs/inode.c | 47 ++++++++++++++++++++++++++++-------------- fs/btrfs/ioctl.c | 4 +++- fs/btrfs/relocation.c | 2 +- 6 files changed, 50 insertions(+), 23 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 34142d5647df..841330f3d68d 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2529,7 +2529,7 @@ int btrfs_update_inode(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode); -void btrfs_orphan_cleanup(struct btrfs_root *root); +int btrfs_orphan_cleanup(struct btrfs_root *root); void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending, u64 *bytes_to_reserve); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index e1aa8d607bc7..495b1ac45f8c 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2058,9 +2058,14 @@ struct btrfs_root *open_ctree(struct super_block *sb, if (!(sb->s_flags & MS_RDONLY)) { down_read(&fs_info->cleanup_work_sem); - btrfs_orphan_cleanup(fs_info->fs_root); - btrfs_orphan_cleanup(fs_info->tree_root); + err = btrfs_orphan_cleanup(fs_info->fs_root); + if (!err) + err = btrfs_orphan_cleanup(fs_info->tree_root); up_read(&fs_info->cleanup_work_sem); + if (err) { + close_ctree(tree_root); + return ERR_PTR(err); + } } return tree_root; @@ -2435,8 +2440,12 @@ int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info) root_objectid = gang[ret - 1]->root_key.objectid + 1; for (i = 0; i < ret; i++) { + int err; + root_objectid = gang[i]->root_key.objectid; - btrfs_orphan_cleanup(gang[i]); + err = btrfs_orphan_cleanup(gang[i]); + if (err) + return err; } root_objectid++; } diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 27376c97d85f..a8f4e8d2ba60 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -7619,7 +7619,8 @@ int btrfs_cleanup_reloc_trees(struct btrfs_root *root) reloc_root = btrfs_read_fs_root_no_name(root->fs_info, &location); BUG_ON(!reloc_root); - btrfs_orphan_cleanup(reloc_root); + ret = btrfs_orphan_cleanup(reloc_root); + BUG_ON(ret); return 0; } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d83025063ee7..0600265cb9b0 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2284,7 +2284,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode) * this cleans up any orphans that may be left on the list from the last use * of this root. */ -void btrfs_orphan_cleanup(struct btrfs_root *root) +int btrfs_orphan_cleanup(struct btrfs_root *root) { struct btrfs_path *path; struct extent_buffer *leaf; @@ -2294,10 +2294,13 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) int ret = 0, nr_unlink = 0, nr_truncate = 0; if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) - return; + return 0; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + ret = -ENOMEM; + goto out; + } path->reada = -1; key.objectid = BTRFS_ORPHAN_OBJECTID; @@ -2306,11 +2309,8 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) while (1) { ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret < 0) { - printk(KERN_ERR "Error searching slot for orphan: %d" - "\n", ret); - break; - } + if (ret < 0) + goto out; /* * if ret == 0 means we found what we were searching for, which @@ -2318,6 +2318,7 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) * find the key and see if we have stuff that matches */ if (ret > 0) { + ret = 0; if (path->slots[0] == 0) break; path->slots[0]--; @@ -2345,7 +2346,10 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) found_key.type = BTRFS_INODE_ITEM_KEY; found_key.offset = 0; inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); - BUG_ON(IS_ERR(inode)); + if (IS_ERR(inode)) { + ret = PTR_ERR(inode); + goto out; + } /* * add this inode to the orphan list so btrfs_orphan_del does @@ -2363,7 +2367,10 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) */ if (is_bad_inode(inode)) { trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } btrfs_orphan_del(trans, inode); btrfs_end_transaction(trans, root); iput(inode); @@ -2378,16 +2385,16 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) continue; } nr_truncate++; - btrfs_truncate(inode); + ret = btrfs_truncate(inode); } else { nr_unlink++; } /* this will do delete_inode and everything for us */ iput(inode); + if (ret) + goto out; } - btrfs_free_path(path); - root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; if (root->orphan_block_rsv) @@ -2396,14 +2403,20 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) if (root->orphan_block_rsv || root->orphan_item_inserted) { trans = btrfs_join_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - btrfs_end_transaction(trans, root); + if (!IS_ERR(trans)) + btrfs_end_transaction(trans, root); } if (nr_unlink) printk(KERN_INFO "btrfs: unlinked %d orphans\n", nr_unlink); if (nr_truncate) printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate); + +out: + if (ret) + printk(KERN_CRIT "btrfs: could not do orphan cleanup %d\n", ret); + btrfs_free_path(path); + return ret; } /* @@ -4156,8 +4169,10 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) if (!IS_ERR(inode) && root != sub_root) { down_read(&root->fs_info->cleanup_work_sem); if (!(inode->i_sb->s_flags & MS_RDONLY)) - btrfs_orphan_cleanup(sub_root); + ret = btrfs_orphan_cleanup(sub_root); up_read(&root->fs_info->cleanup_work_sem); + if (ret) + inode = ERR_PTR(ret); } return inode; diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 5fdb2abc4fa7..ad9b8c0e930b 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -409,7 +409,9 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, if (ret) goto fail; - btrfs_orphan_cleanup(pending_snapshot->snap); + ret = btrfs_orphan_cleanup(pending_snapshot->snap); + if (ret) + goto fail; parent = dget_parent(dentry); inode = btrfs_lookup_dentry(parent->d_inode, dentry); diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 31ade5802ae8..c863c8447015 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4209,7 +4209,7 @@ out: if (IS_ERR(fs_root)) err = PTR_ERR(fs_root); else - btrfs_orphan_cleanup(fs_root); + err = btrfs_orphan_cleanup(fs_root); } return err; } From ded5db9de78f963979e1605f859de67626f54693 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 14:09:46 -0500 Subject: [PATCH 09/45] Btrfs: make sure to remove the orphan item from the in-memory list This fixes a problem where if truncate fails the inode will still be on the in memory orphan list. This is will make us complain when the inode gets destroyed because it's still on the orphan list. So if we fail just remove us from the in memory list and carry on. Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 0600265cb9b0..3bd0ff63bf30 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6589,6 +6589,12 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_orphan_del(trans, inode); if (ret) err = ret; + } else if (ret && inode->i_nlink > 0) { + /* + * Failed to do the truncate, remove us from the in memory + * orphan list. + */ + ret = btrfs_orphan_del(NULL, inode); } ret = btrfs_update_inode(trans, root, inode); From f0cd846e9221811d87047f1428cf5226e7236efe Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 14:37:08 -0500 Subject: [PATCH 10/45] Btrfs: only add orphan items when truncating We don't need an orphan item when expanding files, we just need them for truncating them, so only add the orphan item in btrfs_truncate instead of in btrfs_setsize. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 45 ++++++++++++++++++--------------------------- 1 file changed, 18 insertions(+), 27 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3bd0ff63bf30..206b60362cec 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3649,22 +3649,6 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) if (newsize == oldsize) return 0; - trans = btrfs_start_transaction(root, 5); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - btrfs_set_trans_block_group(trans, inode); - - ret = btrfs_orphan_add(trans, inode); - if (ret) { - btrfs_end_transaction(trans, root); - return ret; - } - - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); - if (newsize > oldsize) { i_size_write(inode, newsize); btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); @@ -3675,25 +3659,15 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return ret; } - trans = btrfs_start_transaction(root, 0); + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, inode); - trans->block_rsv = root->orphan_block_rsv; - BUG_ON(!trans->block_rsv); - - /* - * If this fails just leave the orphan item so that it can get - * cleaned up next time we mount. - */ ret = btrfs_update_inode(trans, root, inode); if (ret) { btrfs_end_transaction(trans, root); return ret; } - if (inode->i_nlink > 0) - ret = btrfs_orphan_del(trans, inode); nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -6517,6 +6491,23 @@ static int btrfs_truncate(struct inode *inode) btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1); btrfs_ordered_update_i_size(inode, inode->i_size, NULL); + trans = btrfs_start_transaction(root, 5); + if (IS_ERR(trans)) + return PTR_ERR(trans); + + btrfs_set_trans_block_group(trans, inode); + + ret = btrfs_orphan_add(trans, inode); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; + } + + nr = trans->blocks_used; + btrfs_end_transaction(trans, root); + btrfs_btree_balance_dirty(root, nr); + + /* Now start a transaction for the truncate */ trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) return PTR_ERR(trans); From 930f028abe39dfd0849b53131d19c4b67aacbe67 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 14:41:41 -0500 Subject: [PATCH 11/45] Btrfs: use mark_inode_dirty when expanding the file Mark_inode_dirty will call btrfs_dirty_inode which will take care of updating the inode. This makes setsize a little cleaner since we don't have to start a transaction and update the inode in there, we can just call mark_inode_dirty. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 16 +--------------- 1 file changed, 1 insertion(+), 15 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 206b60362cec..64d57e032b4e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3640,10 +3640,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) static int btrfs_setsize(struct inode *inode, loff_t newsize) { - struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_trans_handle *trans; loff_t oldsize = i_size_read(inode); - unsigned long nr; int ret; if (newsize == oldsize) @@ -3659,18 +3656,7 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return ret; } - trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - ret = btrfs_update_inode(trans, root, inode); - if (ret) { - btrfs_end_transaction(trans, root); - return ret; - } - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + mark_inode_dirty(inode); } else { /* From 695a0d0da09e75c4475bbb303def159023ef72ca Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 15:46:53 -0500 Subject: [PATCH 12/45] Btrfs: add a comment explaining what btrfs_cont_expand does Everytime I have to deal with btrfs_cont_expand I stare at it for 20 minutes trying to remember what exactly it does and why the hell we need it. So add a comment to save future-Josef some time. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 64d57e032b4e..888dbdb3b128 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3557,6 +3557,12 @@ out: return ret; } +/* + * This function puts in dummy file extents for the area we're creating a hole + * for. So if we are truncating this file to a larger size we need to insert + * these file extents so that btrfs_get_extent will return a EXTENT_MAP_HOLE for + * the range between oldsize and size + */ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) { struct btrfs_trans_handle *trans; From 850265335f792f5d39ab24e5fb7160bac28d77e5 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 15 Mar 2011 14:52:12 -0400 Subject: [PATCH 13/45] Btrfs: return error if the range we want to map is bogus Currently if we have corrupt metadata map_extent_buffer will complain about it, but not return an error so the caller has no idea a problem was hit. Fix this. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/extent_io.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 714adc4ac4c2..1bbd26b4fc5c 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3690,6 +3690,7 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, "wanted %lu %lu\n", (unsigned long long)eb->start, eb->len, start, min_len); WARN_ON(1); + return -EINVAL; } p = extent_buffer_page(eb, i); From a826d6dcb32d811b4c81df57a5ef1367516586b0 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 16 Mar 2011 13:42:43 -0400 Subject: [PATCH 14/45] Btrfs: check items for correctness as we search Currently if we have corrupted items things will blow up in spectacular ways. So as we read in blocks and they are leaves, check the entire leaf to make sure all of the items are correct and point to valid parts in the leaf for the item data the are responsible for. If the item is corrupt we will kick back EIO and not read any of the copies since they are likely to not be correct either. This will catch generic corruptions, it will be up to the individual callers of btrfs_search_slot to make sure their items are right. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/ctree.c | 123 ----------------------------------------- fs/btrfs/disk-io.c | 90 +++++++++++++++++++++++++++++- fs/btrfs/extent-tree.c | 5 ++ fs/btrfs/extent_io.h | 1 + 4 files changed, 95 insertions(+), 124 deletions(-) diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index b5baff0dccfe..73e53009e126 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -732,122 +732,6 @@ static inline unsigned int leaf_data_end(struct btrfs_root *root, return btrfs_item_offset_nr(leaf, nr - 1); } -/* - * extra debugging checks to make sure all the items in a key are - * well formed and in the proper order - */ -static int check_node(struct btrfs_root *root, struct btrfs_path *path, - int level) -{ - struct extent_buffer *parent = NULL; - struct extent_buffer *node = path->nodes[level]; - struct btrfs_disk_key parent_key; - struct btrfs_disk_key node_key; - int parent_slot; - int slot; - struct btrfs_key cpukey; - u32 nritems = btrfs_header_nritems(node); - - if (path->nodes[level + 1]) - parent = path->nodes[level + 1]; - - slot = path->slots[level]; - BUG_ON(nritems == 0); - if (parent) { - parent_slot = path->slots[level + 1]; - btrfs_node_key(parent, &parent_key, parent_slot); - btrfs_node_key(node, &node_key, 0); - BUG_ON(memcmp(&parent_key, &node_key, - sizeof(struct btrfs_disk_key))); - BUG_ON(btrfs_node_blockptr(parent, parent_slot) != - btrfs_header_bytenr(node)); - } - BUG_ON(nritems > BTRFS_NODEPTRS_PER_BLOCK(root)); - if (slot != 0) { - btrfs_node_key_to_cpu(node, &cpukey, slot - 1); - btrfs_node_key(node, &node_key, slot); - BUG_ON(comp_keys(&node_key, &cpukey) <= 0); - } - if (slot < nritems - 1) { - btrfs_node_key_to_cpu(node, &cpukey, slot + 1); - btrfs_node_key(node, &node_key, slot); - BUG_ON(comp_keys(&node_key, &cpukey) >= 0); - } - return 0; -} - -/* - * extra checking to make sure all the items in a leaf are - * well formed and in the proper order - */ -static int check_leaf(struct btrfs_root *root, struct btrfs_path *path, - int level) -{ - struct extent_buffer *leaf = path->nodes[level]; - struct extent_buffer *parent = NULL; - int parent_slot; - struct btrfs_key cpukey; - struct btrfs_disk_key parent_key; - struct btrfs_disk_key leaf_key; - int slot = path->slots[0]; - - u32 nritems = btrfs_header_nritems(leaf); - - if (path->nodes[level + 1]) - parent = path->nodes[level + 1]; - - if (nritems == 0) - return 0; - - if (parent) { - parent_slot = path->slots[level + 1]; - btrfs_node_key(parent, &parent_key, parent_slot); - btrfs_item_key(leaf, &leaf_key, 0); - - BUG_ON(memcmp(&parent_key, &leaf_key, - sizeof(struct btrfs_disk_key))); - BUG_ON(btrfs_node_blockptr(parent, parent_slot) != - btrfs_header_bytenr(leaf)); - } - if (slot != 0 && slot < nritems - 1) { - btrfs_item_key(leaf, &leaf_key, slot); - btrfs_item_key_to_cpu(leaf, &cpukey, slot - 1); - if (comp_keys(&leaf_key, &cpukey) <= 0) { - btrfs_print_leaf(root, leaf); - printk(KERN_CRIT "slot %d offset bad key\n", slot); - BUG_ON(1); - } - if (btrfs_item_offset_nr(leaf, slot - 1) != - btrfs_item_end_nr(leaf, slot)) { - btrfs_print_leaf(root, leaf); - printk(KERN_CRIT "slot %d offset bad\n", slot); - BUG_ON(1); - } - } - if (slot < nritems - 1) { - btrfs_item_key(leaf, &leaf_key, slot); - btrfs_item_key_to_cpu(leaf, &cpukey, slot + 1); - BUG_ON(comp_keys(&leaf_key, &cpukey) >= 0); - if (btrfs_item_offset_nr(leaf, slot) != - btrfs_item_end_nr(leaf, slot + 1)) { - btrfs_print_leaf(root, leaf); - printk(KERN_CRIT "slot %d offset bad\n", slot); - BUG_ON(1); - } - } - BUG_ON(btrfs_item_offset_nr(leaf, 0) + - btrfs_item_size_nr(leaf, 0) != BTRFS_LEAF_DATA_SIZE(root)); - return 0; -} - -static noinline int check_block(struct btrfs_root *root, - struct btrfs_path *path, int level) -{ - return 0; - if (level == 0) - return check_leaf(root, path, level); - return check_node(root, path, level); -} /* * search for key in the extent_buffer. The items start at offset p, @@ -1188,7 +1072,6 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, } } /* double check we haven't messed things up */ - check_block(root, path, level); if (orig_ptr != btrfs_node_blockptr(path->nodes[level], path->slots[level])) BUG(); @@ -1798,12 +1681,6 @@ cow_done: if (!cow) btrfs_unlock_up_safe(p, level + 1); - ret = check_block(root, p, level); - if (ret) { - ret = -1; - goto done; - } - ret = bin_search(b, key, level, &slot); if (level != 0) { diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 495b1ac45f8c..9f31e110b481 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -323,6 +323,7 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, int num_copies = 0; int mirror_num = 0; + clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree; while (1) { ret = read_extent_buffer_pages(io_tree, eb, start, 1, @@ -331,6 +332,14 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, !verify_parent_transid(io_tree, eb, parent_transid)) return ret; + /* + * This buffer's crc is fine, but its contents are corrupted, so + * there is no reason to read the other copies, they won't be + * any less wrong. + */ + if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags)) + return ret; + num_copies = btrfs_num_copies(&root->fs_info->mapping_tree, eb->start, eb->len); if (num_copies == 1) @@ -419,6 +428,73 @@ static int check_tree_block_fsid(struct btrfs_root *root, return ret; } +#define CORRUPT(reason, eb, root, slot) \ + printk(KERN_CRIT "btrfs: corrupt leaf, %s: block=%llu," \ + "root=%llu, slot=%d\n", reason, \ + (unsigned long long)btrfs_header_bytenr(eb), \ + (unsigned long long)root->objectid, slot) + +static noinline int check_leaf(struct btrfs_root *root, + struct extent_buffer *leaf) +{ + struct btrfs_key key; + struct btrfs_key leaf_key; + u32 nritems = btrfs_header_nritems(leaf); + int slot; + + if (nritems == 0) + return 0; + + /* Check the 0 item */ + if (btrfs_item_offset_nr(leaf, 0) + btrfs_item_size_nr(leaf, 0) != + BTRFS_LEAF_DATA_SIZE(root)) { + CORRUPT("invalid item offset size pair", leaf, root, 0); + return -EIO; + } + + /* + * Check to make sure each items keys are in the correct order and their + * offsets make sense. We only have to loop through nritems-1 because + * we check the current slot against the next slot, which verifies the + * next slot's offset+size makes sense and that the current's slot + * offset is correct. + */ + for (slot = 0; slot < nritems - 1; slot++) { + btrfs_item_key_to_cpu(leaf, &leaf_key, slot); + btrfs_item_key_to_cpu(leaf, &key, slot + 1); + + /* Make sure the keys are in the right order */ + if (btrfs_comp_cpu_keys(&leaf_key, &key) >= 0) { + CORRUPT("bad key order", leaf, root, slot); + return -EIO; + } + + /* + * Make sure the offset and ends are right, remember that the + * item data starts at the end of the leaf and grows towards the + * front. + */ + if (btrfs_item_offset_nr(leaf, slot) != + btrfs_item_end_nr(leaf, slot + 1)) { + CORRUPT("slot offset bad", leaf, root, slot); + return -EIO; + } + + /* + * Check to make sure that we don't point outside of the leaf, + * just incase all the items are consistent to eachother, but + * all point outside of the leaf. + */ + if (btrfs_item_end_nr(leaf, slot) > + BTRFS_LEAF_DATA_SIZE(root)) { + CORRUPT("slot end outside of leaf", leaf, root, slot); + return -EIO; + } + } + + return 0; +} + #ifdef CONFIG_DEBUG_LOCK_ALLOC void btrfs_set_buffer_lockdep_class(struct extent_buffer *eb, int level) { @@ -485,8 +561,20 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, btrfs_set_buffer_lockdep_class(eb, found_level); ret = csum_tree_block(root, eb, 1); - if (ret) + if (ret) { ret = -EIO; + goto err; + } + + /* + * If this is a leaf block and it is corrupt, set the corrupt bit so + * that we don't try and read the other copies of this block, just + * return -EIO. + */ + if (found_level == 0 && check_leaf(root, eb)) { + set_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); + ret = -EIO; + } end = min_t(u64, eb->len, PAGE_CACHE_SIZE); end = eb->start + end - 1; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a8f4e8d2ba60..cd794c35a636 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4754,6 +4754,11 @@ pin: } } out: + /* + * Deleting the buffer, clear the corrupt flag since it doesn't matter + * anymore. + */ + clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags); btrfs_put_block_group(cache); } diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 9318dfefd59c..f62c5442835d 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -31,6 +31,7 @@ #define EXTENT_BUFFER_UPTODATE 0 #define EXTENT_BUFFER_BLOCKING 1 #define EXTENT_BUFFER_DIRTY 2 +#define EXTENT_BUFFER_CORRUPT 3 /* these are flags for extent_clear_unlock_delalloc */ #define EXTENT_CLEAR_UNLOCK_PAGE 0x1 From 41415730a1050499fbd63b3f7dd59b3a4c3bb91a Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 16 Mar 2011 13:59:32 -0400 Subject: [PATCH 15/45] Btrfs: check return value of btrfs_search_slot properly Doing an audit of where we use btrfs_search_slot only showed one place where we don't check the return value of btrfs_search_slot properly. Just fix mark_extent_written to see if btrfs_search_slot failed and act accordingly. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/file.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 3786eca2a905..a85b044cf39e 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -608,6 +608,8 @@ again: key.offset = split; ret = btrfs_search_slot(trans, root, &key, path, -1, 1); + if (ret < 0) + goto out; if (ret > 0 && path->slots[0] > 0) path->slots[0]--; From 22a94d44bd6876a90630338229da6c0436d46593 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 16 Mar 2011 16:47:17 -0400 Subject: [PATCH 16/45] Btrfs: add checks to verify dir items are correct We need to make sure the dir items we get are valid dir items. So any time we try and read one check it with verify_dir_item, which will do various sanity checks to make sure it looks sane. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/ctree.h | 3 +++ fs/btrfs/dir-item.c | 35 +++++++++++++++++++++++++++++++++++ fs/btrfs/inode.c | 3 +++ fs/btrfs/tree-log.c | 7 +++++++ fs/btrfs/xattr.c | 2 ++ 5 files changed, 50 insertions(+) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 841330f3d68d..6036fdb88c53 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2393,6 +2393,9 @@ struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans, struct btrfs_path *path, u64 dir, const char *name, u16 name_len, int mod); +int verify_dir_item(struct btrfs_root *root, + struct extent_buffer *leaf, + struct btrfs_dir_item *dir_item); /* orphan.c */ int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c index f0cad5ae5be7..02c97ad61b6d 100644 --- a/fs/btrfs/dir-item.c +++ b/fs/btrfs/dir-item.c @@ -377,6 +377,9 @@ struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_root *root, leaf = path->nodes[0]; dir_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dir_item); + if (verify_dir_item(root, leaf, dir_item)) + return NULL; + total_len = btrfs_item_size_nr(leaf, path->slots[0]); while (cur < total_len) { this_len = sizeof(*dir_item) + @@ -429,3 +432,35 @@ int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans, } return ret; } + +int verify_dir_item(struct btrfs_root *root, + struct extent_buffer *leaf, + struct btrfs_dir_item *dir_item) +{ + u16 namelen = BTRFS_NAME_LEN; + u8 type = btrfs_dir_type(leaf, dir_item); + + if (type >= BTRFS_FT_MAX) { + printk(KERN_CRIT "btrfs: invalid dir item type: %d\n", + (int)type); + return 1; + } + + if (type == BTRFS_FT_XATTR) + namelen = XATTR_NAME_MAX; + + if (btrfs_dir_name_len(leaf, dir_item) > namelen) { + printk(KERN_CRIT "btrfS: invalid dir item name len: %u\n", + (unsigned)btrfs_dir_data_len(leaf, dir_item)); + return 1; + } + + /* BTRFS_MAX_XATTR_SIZE is the same for all dir items */ + if (btrfs_dir_data_len(leaf, dir_item) > BTRFS_MAX_XATTR_SIZE(root)) { + printk(KERN_CRIT "btrfs: invalid dir item data len: %u\n", + (unsigned)btrfs_dir_data_len(leaf, dir_item)); + return 1; + } + + return 0; +} diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 888dbdb3b128..e010000d4bc9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4272,6 +4272,9 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, while (di_cur < di_total) { struct btrfs_key location; + if (verify_dir_item(root, leaf, di)) + break; + name_len = btrfs_dir_name_len(leaf, di); if (name_len <= sizeof(tmp_name)) { name_ptr = tmp_name; diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index a4bbb854dfd2..429cfcfadf90 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -1286,6 +1286,8 @@ static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans, ptr_end = ptr + item_size; while (ptr < ptr_end) { di = (struct btrfs_dir_item *)ptr; + if (verify_dir_item(root, eb, di)) + return -EIO; name_len = btrfs_dir_name_len(eb, di); ret = replay_one_name(trans, root, path, eb, di, key); BUG_ON(ret); @@ -1412,6 +1414,11 @@ again: ptr_end = ptr + item_size; while (ptr < ptr_end) { di = (struct btrfs_dir_item *)ptr; + if (verify_dir_item(root, eb, di)) { + ret = -EIO; + goto out; + } + name_len = btrfs_dir_name_len(eb, di); name = kmalloc(name_len, GFP_NOFS); if (!name) { diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index a5776531dc2b..e5d22f280956 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -242,6 +242,8 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size) break; di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item); + if (verify_dir_item(root, leaf, di)) + continue; name_len = btrfs_dir_name_len(leaf, di); total_size += name_len + 1; From 7d0d2e8e6b6f7da221a25238cf490a095c8c4788 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 18 Mar 2011 15:13:42 -0400 Subject: [PATCH 17/45] Btrfs: check free space in block group before searching for a cluster The free space cluster stuff is heavy duty, so there is no sense in going through the entire song and dance if there isn't enough space in the block group to begin with. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/free-space-cache.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 0282033041e1..f631df870f64 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1999,6 +1999,16 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, min_bytes = max(bytes, (bytes + empty_size) >> 2); spin_lock(&block_group->tree_lock); + + /* + * If we know we don't have enough space to make a cluster don't even + * bother doing all the work to try and find one. + */ + if (block_group->free_space < min_bytes) { + spin_unlock(&block_group->tree_lock); + return -ENOSPC; + } + spin_lock(&cluster->lock); /* someone already found a cluster, hooray */ From d0a365e84a886ce6b5b7f7a76be0bb24934ec8f0 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 18 Mar 2011 15:27:43 -0400 Subject: [PATCH 18/45] Btrfs: deal with min_bytes appropriately when looking for a cluster We do all this fun stuff with min_bytes, but either don't use it in the case of just normal extents, or use it completely wrong in the case of bitmaps. So fix this for both cases 1) In the extent case, stop looking for space with window_free >= min_bytes instead of bytes + empty_size. 2) In the bitmap case, we were looking for streches of free space that was at least min_bytes in size, which was not right at all. So instead search for stretches of free space that are at least bytes in size (this will make a difference when we have > page size blocks) and then only search for min_bytes amount of free space. Thanks, Reviewed-by: Li Zefan Signed-off-by: Josef Bacik --- fs/btrfs/free-space-cache.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index f631df870f64..63776ae72f9e 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1910,8 +1910,8 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, i = offset_to_bit(entry->offset, block_group->sectorsize, max_t(u64, offset, entry->offset)); - search_bits = bytes_to_bits(min_bytes, block_group->sectorsize); - total_bits = bytes_to_bits(bytes, block_group->sectorsize); + search_bits = bytes_to_bits(bytes, block_group->sectorsize); + total_bits = bytes_to_bits(min_bytes, block_group->sectorsize); again: found_bits = 0; @@ -2034,8 +2034,7 @@ again: if (entry->bitmap && entry->bytes > bytes + empty_size) { ret = btrfs_bitmap_cluster(block_group, entry, cluster, - offset, bytes + empty_size, - min_bytes); + offset, bytes, min_bytes); if (!ret) goto got_it; } @@ -2065,7 +2064,7 @@ again: while (1) { /* out window is just right, lets fill it */ - if (window_free >= bytes + empty_size) + if (window_free >= min_bytes) break; node = rb_next(&last->offset_index); From 32cb0840ce8e13901fe71a9a8e834a531802ffc4 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 18 Mar 2011 16:16:21 -0400 Subject: [PATCH 19/45] Btrfs: don't be as aggressive about using bitmaps We have been creating bitmaps for small extents unconditionally forever. This was great when testing to make sure the bitmap stuff was working, but is overkill normally. So instead of always adding small chunks of free space to bitmaps, only start doing it if we go past half of our extent threshold. This will keeps us from creating a bitmap for just one small free extent at the front of the block group, and will make the allocator a little faster as a result. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/free-space-cache.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 63776ae72f9e..4ab35ea0443f 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1287,9 +1287,22 @@ static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, * If we are below the extents threshold then we can add this as an * extent, and don't have to deal with the bitmap */ - if (block_group->free_extents < block_group->extents_thresh && - info->bytes > block_group->sectorsize * 4) - return 0; + if (block_group->free_extents < block_group->extents_thresh) { + /* + * If this block group has some small extents we don't want to + * use up all of our free slots in the cache with them, we want + * to reserve them to larger extents, however if we have plent + * of cache left then go ahead an dadd them, no sense in adding + * the overhead of a bitmap if we don't have to. + */ + if (info->bytes <= block_group->sectorsize * 4) { + if (block_group->free_extents * 2 <= + block_group->extents_thresh) + return 0; + } else { + return 0; + } + } /* * some block groups are so tiny they can't be enveloped by a bitmap, so From 4e69b598f6cfb0940b75abf7e179d6020e94ad1e Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 21 Mar 2011 10:11:24 -0400 Subject: [PATCH 20/45] Btrfs: cleanup how we setup free space clusters This patch makes the free space cluster refilling code a little easier to understand, and fixes some things with the bitmap part of it. Currently we either want to refill a cluster with 1) All normal extent entries (those without bitmaps) 2) A bitmap entry with enough space The current code has this ugly jump around logic that will first try and fill up the cluster with extent entries and then if it can't do that it will try and find a bitmap to use. So instead split this out into two functions, one that tries to find only normal entries, and one that tries to find bitmaps. This also fixes a suboptimal thing we would do with bitmaps. If we used a bitmap we would just tell the cluster that we were pointing at a bitmap and it would do the tree search in the block group for that entry every time we tried to make an allocation. Instead of doing that now we just add it to the clusters group. I tested this with my ENOSPC tests and xfstests and it survived. Signed-off-by: Josef Bacik --- fs/btrfs/ctree.h | 3 - fs/btrfs/free-space-cache.c | 368 ++++++++++++++++++------------------ 2 files changed, 184 insertions(+), 187 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 6036fdb88c53..0ee679b6c1b7 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -783,9 +783,6 @@ struct btrfs_free_cluster { /* first extent starting offset */ u64 window_start; - /* if this cluster simply points at a bitmap in the block group */ - bool points_to_bitmap; - struct btrfs_block_group_cache *block_group; /* * when a cluster is allocated from a block group, we put the diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 4ab35ea0443f..f03ef97c3b21 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1644,30 +1644,28 @@ __btrfs_return_cluster_to_free_space( { struct btrfs_free_space *entry; struct rb_node *node; - bool bitmap; spin_lock(&cluster->lock); if (cluster->block_group != block_group) goto out; - bitmap = cluster->points_to_bitmap; cluster->block_group = NULL; cluster->window_start = 0; list_del_init(&cluster->block_group_list); - cluster->points_to_bitmap = false; - - if (bitmap) - goto out; node = rb_first(&cluster->root); while (node) { + bool bitmap; + entry = rb_entry(node, struct btrfs_free_space, offset_index); node = rb_next(&entry->offset_index); rb_erase(&entry->offset_index, &cluster->root); - BUG_ON(entry->bitmap); - try_merge_free_space(block_group, entry, false); + + bitmap = (entry->bitmap != NULL); + if (!bitmap) + try_merge_free_space(block_group, entry, false); tree_insert_offset(&block_group->free_space_offset, - entry->offset, &entry->offset_index, 0); + entry->offset, &entry->offset_index, bitmap); } cluster->root = RB_ROOT; @@ -1790,50 +1788,24 @@ int btrfs_return_cluster_to_free_space( static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, + struct btrfs_free_space *entry, u64 bytes, u64 min_start) { - struct btrfs_free_space *entry; int err; u64 search_start = cluster->window_start; u64 search_bytes = bytes; u64 ret = 0; - spin_lock(&block_group->tree_lock); - spin_lock(&cluster->lock); - - if (!cluster->points_to_bitmap) - goto out; - - if (cluster->block_group != block_group) - goto out; - - /* - * search_start is the beginning of the bitmap, but at some point it may - * be a good idea to point to the actual start of the free area in the - * bitmap, so do the offset_to_bitmap trick anyway, and set bitmap_only - * to 1 to make sure we get the bitmap entry - */ - entry = tree_search_offset(block_group, - offset_to_bitmap(block_group, search_start), - 1, 0); - if (!entry || !entry->bitmap) - goto out; - search_start = min_start; search_bytes = bytes; err = search_bitmap(block_group, entry, &search_start, &search_bytes); if (err) - goto out; + return 0; ret = search_start; bitmap_clear_bits(block_group, entry, ret, bytes); - if (entry->bytes == 0) - free_bitmap(block_group, entry); -out: - spin_unlock(&cluster->lock); - spin_unlock(&block_group->tree_lock); return ret; } @@ -1851,10 +1823,6 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, struct rb_node *node; u64 ret = 0; - if (cluster->points_to_bitmap) - return btrfs_alloc_from_bitmap(block_group, cluster, bytes, - min_start); - spin_lock(&cluster->lock); if (bytes > cluster->max_size) goto out; @@ -1867,9 +1835,9 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, goto out; entry = rb_entry(node, struct btrfs_free_space, offset_index); - while(1) { - if (entry->bytes < bytes || entry->offset < min_start) { + if (entry->bytes < bytes || + (!entry->bitmap && entry->offset < min_start)) { struct rb_node *node; node = rb_next(&entry->offset_index); @@ -1879,10 +1847,27 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, offset_index); continue; } - ret = entry->offset; - entry->offset += bytes; - entry->bytes -= bytes; + if (entry->bitmap) { + ret = btrfs_alloc_from_bitmap(block_group, + cluster, entry, bytes, + min_start); + if (ret == 0) { + struct rb_node *node; + node = rb_next(&entry->offset_index); + if (!node) + break; + entry = rb_entry(node, struct btrfs_free_space, + offset_index); + continue; + } + } else { + + ret = entry->offset; + + entry->offset += bytes; + entry->bytes -= bytes; + } if (entry->bytes == 0) rb_erase(&entry->offset_index, &cluster->root); @@ -1899,6 +1884,11 @@ out: block_group->free_space -= bytes; if (entry->bytes == 0) { block_group->free_extents--; + if (entry->bitmap) { + kfree(entry->bitmap); + block_group->total_bitmaps--; + recalculate_thresholds(block_group); + } kmem_cache_free(btrfs_free_space_cachep, entry); } @@ -1919,6 +1909,7 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, unsigned long found_bits; unsigned long start = 0; unsigned long total_found = 0; + int ret; bool found = false; i = offset_to_bit(entry->offset, block_group->sectorsize, @@ -1941,7 +1932,7 @@ again: } if (!found_bits) - return -1; + return -ENOSPC; if (!found) { start = i; @@ -1965,11 +1956,144 @@ again: cluster->window_start = start * block_group->sectorsize + entry->offset; - cluster->points_to_bitmap = true; + rb_erase(&entry->offset_index, &block_group->free_space_offset); + ret = tree_insert_offset(&cluster->root, entry->offset, + &entry->offset_index, 1); + BUG_ON(ret); return 0; } +/* + * This searches the block group for just extents to fill the cluster with. + */ +static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + u64 offset, u64 bytes, u64 min_bytes) +{ + struct btrfs_free_space *first = NULL; + struct btrfs_free_space *entry = NULL; + struct btrfs_free_space *prev = NULL; + struct btrfs_free_space *last; + struct rb_node *node; + u64 window_start; + u64 window_free; + u64 max_extent; + u64 max_gap = 128 * 1024; + + entry = tree_search_offset(block_group, offset, 0, 1); + if (!entry) + return -ENOSPC; + + /* + * We don't want bitmaps, so just move along until we find a normal + * extent entry. + */ + while (entry->bitmap) { + node = rb_next(&entry->offset_index); + if (!node) + return -ENOSPC; + entry = rb_entry(node, struct btrfs_free_space, offset_index); + } + + window_start = entry->offset; + window_free = entry->bytes; + max_extent = entry->bytes; + first = entry; + last = entry; + prev = entry; + + while (window_free <= min_bytes) { + node = rb_next(&entry->offset_index); + if (!node) + return -ENOSPC; + entry = rb_entry(node, struct btrfs_free_space, offset_index); + + if (entry->bitmap) + continue; + /* + * we haven't filled the empty size and the window is + * very large. reset and try again + */ + if (entry->offset - (prev->offset + prev->bytes) > max_gap || + entry->offset - window_start > (min_bytes * 2)) { + first = entry; + window_start = entry->offset; + window_free = entry->bytes; + last = entry; + max_extent = entry->bytes; + } else { + last = entry; + window_free += entry->bytes; + if (entry->bytes > max_extent) + max_extent = entry->bytes; + } + prev = entry; + } + + cluster->window_start = first->offset; + + node = &first->offset_index; + + /* + * now we've found our entries, pull them out of the free space + * cache and put them into the cluster rbtree + */ + do { + int ret; + + entry = rb_entry(node, struct btrfs_free_space, offset_index); + node = rb_next(&entry->offset_index); + if (entry->bitmap) + continue; + + rb_erase(&entry->offset_index, &block_group->free_space_offset); + ret = tree_insert_offset(&cluster->root, entry->offset, + &entry->offset_index, 0); + BUG_ON(ret); + } while (node && entry != last); + + cluster->max_size = max_extent; + + return 0; +} + +/* + * This specifically looks for bitmaps that may work in the cluster, we assume + * that we have already failed to find extents that will work. + */ +static int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + u64 offset, u64 bytes, u64 min_bytes) +{ + struct btrfs_free_space *entry; + struct rb_node *node; + int ret = -ENOSPC; + + if (block_group->total_bitmaps == 0) + return -ENOSPC; + + entry = tree_search_offset(block_group, + offset_to_bitmap(block_group, offset), + 0, 1); + if (!entry) + return -ENOSPC; + + node = &entry->offset_index; + do { + entry = rb_entry(node, struct btrfs_free_space, offset_index); + node = rb_next(&entry->offset_index); + if (!entry->bitmap) + continue; + if (entry->bytes < min_bytes) + continue; + ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, + bytes, min_bytes); + } while (ret && node); + + return ret; +} + /* * here we try to find a cluster of blocks in a block group. The goal * is to find at least bytes free and up to empty_size + bytes free. @@ -1984,15 +2108,7 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, struct btrfs_free_cluster *cluster, u64 offset, u64 bytes, u64 empty_size) { - struct btrfs_free_space *entry = NULL; - struct rb_node *node; - struct btrfs_free_space *next; - struct btrfs_free_space *last = NULL; u64 min_bytes; - u64 window_start; - u64 window_free; - u64 max_extent = 0; - bool found_bitmap = false; int ret; /* for metadata, allow allocates with more holes */ @@ -2029,134 +2145,19 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, ret = 0; goto out; } -again: - entry = tree_search_offset(block_group, offset, found_bitmap, 1); - if (!entry) { - ret = -ENOSPC; - goto out; + + ret = setup_cluster_no_bitmap(block_group, cluster, offset, bytes, + min_bytes); + if (ret) + ret = setup_cluster_bitmap(block_group, cluster, offset, + bytes, min_bytes); + + if (!ret) { + atomic_inc(&block_group->count); + list_add_tail(&cluster->block_group_list, + &block_group->cluster_list); + cluster->block_group = block_group; } - - /* - * If found_bitmap is true, we exhausted our search for extent entries, - * and we just want to search all of the bitmaps that we can find, and - * ignore any extent entries we find. - */ - while (entry->bitmap || found_bitmap || - (!entry->bitmap && entry->bytes < min_bytes)) { - struct rb_node *node = rb_next(&entry->offset_index); - - if (entry->bitmap && entry->bytes > bytes + empty_size) { - ret = btrfs_bitmap_cluster(block_group, entry, cluster, - offset, bytes, min_bytes); - if (!ret) - goto got_it; - } - - if (!node) { - ret = -ENOSPC; - goto out; - } - entry = rb_entry(node, struct btrfs_free_space, offset_index); - } - - /* - * We already searched all the extent entries from the passed in offset - * to the end and didn't find enough space for the cluster, and we also - * didn't find any bitmaps that met our criteria, just go ahead and exit - */ - if (found_bitmap) { - ret = -ENOSPC; - goto out; - } - - cluster->points_to_bitmap = false; - window_start = entry->offset; - window_free = entry->bytes; - last = entry; - max_extent = entry->bytes; - - while (1) { - /* out window is just right, lets fill it */ - if (window_free >= min_bytes) - break; - - node = rb_next(&last->offset_index); - if (!node) { - if (found_bitmap) - goto again; - ret = -ENOSPC; - goto out; - } - next = rb_entry(node, struct btrfs_free_space, offset_index); - - /* - * we found a bitmap, so if this search doesn't result in a - * cluster, we know to go and search again for the bitmaps and - * start looking for space there - */ - if (next->bitmap) { - if (!found_bitmap) - offset = next->offset; - found_bitmap = true; - last = next; - continue; - } - - /* - * we haven't filled the empty size and the window is - * very large. reset and try again - */ - if (next->offset - (last->offset + last->bytes) > 128 * 1024 || - next->offset - window_start > (bytes + empty_size) * 2) { - entry = next; - window_start = entry->offset; - window_free = entry->bytes; - last = entry; - max_extent = entry->bytes; - } else { - last = next; - window_free += next->bytes; - if (entry->bytes > max_extent) - max_extent = entry->bytes; - } - } - - cluster->window_start = entry->offset; - - /* - * now we've found our entries, pull them out of the free space - * cache and put them into the cluster rbtree - * - * The cluster includes an rbtree, but only uses the offset index - * of each free space cache entry. - */ - while (1) { - node = rb_next(&entry->offset_index); - if (entry->bitmap && node) { - entry = rb_entry(node, struct btrfs_free_space, - offset_index); - continue; - } else if (entry->bitmap && !node) { - break; - } - - rb_erase(&entry->offset_index, &block_group->free_space_offset); - ret = tree_insert_offset(&cluster->root, entry->offset, - &entry->offset_index, 0); - BUG_ON(ret); - - if (!node || entry == last) - break; - - entry = rb_entry(node, struct btrfs_free_space, offset_index); - } - - cluster->max_size = max_extent; -got_it: - ret = 0; - atomic_inc(&block_group->count); - list_add_tail(&cluster->block_group_list, &block_group->cluster_list); - cluster->block_group = block_group; out: spin_unlock(&cluster->lock); spin_unlock(&block_group->tree_lock); @@ -2173,7 +2174,6 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) spin_lock_init(&cluster->refill_lock); cluster->root = RB_ROOT; cluster->max_size = 0; - cluster->points_to_bitmap = false; INIT_LIST_HEAD(&cluster->block_group_list); cluster->block_group = NULL; } From 98bc3149fad639c8f50c7110b961a2a2fe085eed Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 22 Mar 2011 11:00:46 -0400 Subject: [PATCH 21/45] Btrfs: don't allocate dip->csums when doing writes When doing direct writes we store the checksums in the ordered sum stuff in the ordered extent for writing them when the write completes, so we don't even use the dip->csums array. So if we're writing, don't bother allocating dip->csums since we won't use it anyway. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e010000d4bc9..570cd44fe91b 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5944,6 +5944,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, int nr_pages = 0; u32 *csums = dip->csums; int ret = 0; + int write = rw & REQ_WRITE; bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); if (!bio) @@ -5980,7 +5981,8 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, goto out_err; } - if (!skip_sum) + /* Write's use the ordered csums */ + if (!write && !skip_sum) csums = csums + nr_pages; start_sector += submit_len >> 9; file_offset += submit_len; @@ -6048,7 +6050,8 @@ static void btrfs_submit_direct(int rw, struct bio *bio, struct inode *inode, } dip->csums = NULL; - if (!skip_sum) { + /* Write's use the ordered csum stuff, so we don't need dip->csums */ + if (!write && !skip_sum) { dip->csums = kmalloc(sizeof(u32) * bio->bi_vcnt, GFP_NOFS); if (!dip->csums) { kfree(dip); From c0da7aa1a2d8fcafe271a7077599253c8ed94bb2 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 22 Mar 2011 11:05:07 -0400 Subject: [PATCH 22/45] Btrfs: mark the bio with an error if we have a failure in dio I noticed that dio_end_io calls the appropriate endio function with an error, but the endio functions don't actually do anything with that error, they assume that if there was an error then the bio will not be uptodate. So if we had checksum failures we would never pass back EIO. So if there is an error in our endio functions make sure to clear the uptodate flag on the bio. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 570cd44fe91b..e9813bd7d556 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5744,6 +5744,10 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) kfree(dip->csums); kfree(dip); + + /* If we had a csum failure make sure to clear the uptodate flag */ + if (err) + clear_bit(BIO_UPTODATE, &bio->bi_flags); dio_end_io(bio, err); } @@ -5845,6 +5849,10 @@ out_done: kfree(dip->csums); kfree(dip); + + /* If we had an error make sure to clear the uptodate flag */ + if (err) + clear_bit(BIO_UPTODATE, &bio->bi_flags); dio_end_io(bio, err); } From 240f62c8756df285da11469259b3900f32883168 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Wed, 23 Mar 2011 14:54:42 -0400 Subject: [PATCH 23/45] Btrfs: use RCU instead of a spinlock to protect the root node The pointer to the extent buffer for the root of each tree is protected by a spinlock so that we can safely read the pointer and take a reference on the extent buffer. But now that the extent buffers are freed via RCU, we can safely use rcu_read_lock instead. Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 27 ++++++++------------------- 1 file changed, 8 insertions(+), 19 deletions(-) diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 73e53009e126..8680110f0a5a 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -147,10 +147,11 @@ noinline void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p) struct extent_buffer *btrfs_root_node(struct btrfs_root *root) { struct extent_buffer *eb; - spin_lock(&root->node_lock); - eb = root->node; + + rcu_read_lock(); + eb = rcu_dereference(root->node); extent_buffer_get(eb); - spin_unlock(&root->node_lock); + rcu_read_unlock(); return eb; } @@ -165,14 +166,8 @@ struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root) while (1) { eb = btrfs_root_node(root); btrfs_tree_lock(eb); - - spin_lock(&root->node_lock); - if (eb == root->node) { - spin_unlock(&root->node_lock); + if (eb == root->node) break; - } - spin_unlock(&root->node_lock); - btrfs_tree_unlock(eb); free_extent_buffer(eb); } @@ -458,10 +453,8 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, else parent_start = 0; - spin_lock(&root->node_lock); - root->node = cow; extent_buffer_get(cow); - spin_unlock(&root->node_lock); + rcu_assign_pointer(root->node, cow); btrfs_free_tree_block(trans, root, buf, parent_start, last_ref); @@ -930,9 +923,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, goto enospc; } - spin_lock(&root->node_lock); - root->node = child; - spin_unlock(&root->node_lock); + rcu_assign_pointer(root->node, child); add_root_to_dirty_list(root); btrfs_tree_unlock(child); @@ -2007,10 +1998,8 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(c); - spin_lock(&root->node_lock); old = root->node; - root->node = c; - spin_unlock(&root->node_lock); + rcu_assign_pointer(root->node, c); /* the super has an extra ref to root->node */ free_extent_buffer(old); From 1abe9b8a138c9988ba8f7bfded6453649a31541f Mon Sep 17 00:00:00 2001 From: liubo Date: Thu, 24 Mar 2011 11:18:59 +0000 Subject: [PATCH 24/45] Btrfs: add initial tracepoint support for btrfs Tracepoints can provide insight into why btrfs hits bugs and be greatly helpful for debugging, e.g dd-7822 [000] 2121.641088: btrfs_inode_request: root = 5(FS_TREE), gen = 4, ino = 256, blocks = 8, disk_i_size = 0, last_trans = 8, logged_trans = 0 dd-7822 [000] 2121.641100: btrfs_inode_new: root = 5(FS_TREE), gen = 8, ino = 257, blocks = 0, disk_i_size = 0, last_trans = 0, logged_trans = 0 btrfs-transacti-7804 [001] 2146.935420: btrfs_cow_block: root = 2(EXTENT_TREE), refs = 2, orig_buf = 29368320 (orig_level = 0), cow_buf = 29388800 (cow_level = 0) btrfs-transacti-7804 [001] 2146.935473: btrfs_cow_block: root = 1(ROOT_TREE), refs = 2, orig_buf = 29364224 (orig_level = 0), cow_buf = 29392896 (cow_level = 0) btrfs-transacti-7804 [001] 2146.972221: btrfs_transaction_commit: root = 1(ROOT_TREE), gen = 8 flush-btrfs-2-7821 [001] 2155.824210: btrfs_chunk_alloc: root = 3(CHUNK_TREE), offset = 1103101952, size = 1073741824, num_stripes = 1, sub_stripes = 0, type = DATA flush-btrfs-2-7821 [001] 2155.824241: btrfs_cow_block: root = 2(EXTENT_TREE), refs = 2, orig_buf = 29388800 (orig_level = 0), cow_buf = 29396992 (cow_level = 0) flush-btrfs-2-7821 [001] 2155.824255: btrfs_cow_block: root = 4(DEV_TREE), refs = 2, orig_buf = 29372416 (orig_level = 0), cow_buf = 29401088 (cow_level = 0) flush-btrfs-2-7821 [000] 2155.824329: btrfs_cow_block: root = 3(CHUNK_TREE), refs = 2, orig_buf = 20971520 (orig_level = 0), cow_buf = 20975616 (cow_level = 0) btrfs-endio-wri-7800 [001] 2155.898019: btrfs_cow_block: root = 5(FS_TREE), refs = 2, orig_buf = 29384704 (orig_level = 0), cow_buf = 29405184 (cow_level = 0) btrfs-endio-wri-7800 [001] 2155.898043: btrfs_cow_block: root = 7(CSUM_TREE), refs = 2, orig_buf = 29376512 (orig_level = 0), cow_buf = 29409280 (cow_level = 0) Here is what I have added: 1) ordere_extent: btrfs_ordered_extent_add btrfs_ordered_extent_remove btrfs_ordered_extent_start btrfs_ordered_extent_put These provide critical information to understand how ordered_extents are updated. 2) extent_map: btrfs_get_extent extent_map is used in both read and write cases, and it is useful for tracking how btrfs specific IO is running. 3) writepage: __extent_writepage btrfs_writepage_end_io_hook Pages are cirtical resourses and produce a lot of corner cases during writeback, so it is valuable to know how page is written to disk. 4) inode: btrfs_inode_new btrfs_inode_request btrfs_inode_evict These can show where and when a inode is created, when a inode is evicted. 5) sync: btrfs_sync_file btrfs_sync_fs These show sync arguments. 6) transaction: btrfs_transaction_commit In transaction based filesystem, it will be useful to know the generation and who does commit. 7) back reference and cow: btrfs_delayed_tree_ref btrfs_delayed_data_ref btrfs_delayed_ref_head btrfs_cow_block Btrfs natively supports back references, these tracepoints are helpful on understanding btrfs's COW mechanism. 8) chunk: btrfs_chunk_alloc btrfs_chunk_free Chunk is a link between physical offset and logical offset, and stands for space infomation in btrfs, and these are helpful on tracing space things. 9) reserved_extent: btrfs_reserved_extent_alloc btrfs_reserved_extent_free These can show how btrfs uses its space. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 3 + fs/btrfs/ctree.h | 1 + fs/btrfs/delayed-ref.c | 6 + fs/btrfs/extent-tree.c | 4 + fs/btrfs/extent_io.c | 2 + fs/btrfs/file.c | 1 + fs/btrfs/inode.c | 12 + fs/btrfs/ordered-data.c | 8 + fs/btrfs/super.c | 5 + fs/btrfs/transaction.c | 2 + fs/btrfs/volumes.c | 16 +- fs/btrfs/volumes.h | 11 + include/trace/events/btrfs.h | 667 +++++++++++++++++++++++++++++++++++ 13 files changed, 727 insertions(+), 11 deletions(-) create mode 100644 include/trace/events/btrfs.h diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 8680110f0a5a..465b5d7d6b48 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -535,6 +535,9 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans, ret = __btrfs_cow_block(trans, root, buf, parent, parent_slot, cow_ret, search_start, 0); + + trace_btrfs_cow_block(root, buf, *cow_ret); + return ret; } diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 0ee679b6c1b7..9d0f59142afa 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -28,6 +28,7 @@ #include #include #include +#include #include #include "extent_io.h" #include "extent_map.h" diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index e807b143b857..bce28f653899 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -483,6 +483,8 @@ static noinline int add_delayed_ref_head(struct btrfs_trans_handle *trans, INIT_LIST_HEAD(&head_ref->cluster); mutex_init(&head_ref->mutex); + trace_btrfs_delayed_ref_head(ref, head_ref, action); + existing = tree_insert(&delayed_refs->root, &ref->rb_node); if (existing) { @@ -537,6 +539,8 @@ static noinline int add_delayed_tree_ref(struct btrfs_trans_handle *trans, } full_ref->level = level; + trace_btrfs_delayed_tree_ref(ref, full_ref, action); + existing = tree_insert(&delayed_refs->root, &ref->rb_node); if (existing) { @@ -591,6 +595,8 @@ static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans, full_ref->objectid = owner; full_ref->offset = offset; + trace_btrfs_delayed_data_ref(ref, full_ref, action); + existing = tree_insert(&delayed_refs->root, &ref->rb_node); if (existing) { diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index cd794c35a636..86ea471d3801 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5412,6 +5412,8 @@ again: dump_space_info(sinfo, num_bytes, 1); } + trace_btrfs_reserved_extent_alloc(root, ins->objectid, ins->offset); + return ret; } @@ -5433,6 +5435,8 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) update_reserved_bytes(cache, len, 0, 1); btrfs_put_block_group(cache); + trace_btrfs_reserved_extent_free(root, start, len); + return ret; } diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 1bbd26b4fc5c..77c65a0bea34 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2192,6 +2192,8 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, else write_flags = WRITE; + trace___extent_writepage(page, inode, wbc); + WARN_ON(!PageLocked(page)); pg_offset = i_size & (PAGE_CACHE_SIZE - 1); if (page->index > end_index || diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index a85b044cf39e..656bc0a892b1 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1205,6 +1205,7 @@ int btrfs_sync_file(struct file *file, int datasync) int ret = 0; struct btrfs_trans_handle *trans; + trace_btrfs_sync_file(file, datasync); /* we wait first, since the writeback may change the inode */ root->log_batch++; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e9813bd7d556..eaa271484199 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1787,6 +1787,8 @@ out: static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, struct extent_state *state, int uptodate) { + trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); + ClearPagePrivate2(page); return btrfs_finish_ordered_io(page->mapping->host, start, end); } @@ -3718,6 +3720,8 @@ void btrfs_evict_inode(struct inode *inode) unsigned long nr; int ret; + trace_btrfs_inode_evict(inode); + truncate_inode_pages(&inode->i_data, 0); if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 || root == root->fs_info->tree_root)) @@ -4510,6 +4514,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, return ERR_PTR(-ENOMEM); if (dir) { + trace_btrfs_inode_request(dir); + ret = btrfs_set_inode_index(dir, index); if (ret) { iput(inode); @@ -4584,6 +4590,9 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, insert_inode_hash(inode); inode_tree_add(inode); + + trace_btrfs_inode_new(inode); + return inode; fail: if (dir) @@ -5261,6 +5270,9 @@ insert: } write_unlock(&em_tree->lock); out: + + trace_btrfs_get_extent(root, em); + if (path) btrfs_free_path(path); if (trans) { diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index 083a55477375..a1c940425307 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -202,6 +202,8 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, INIT_LIST_HEAD(&entry->list); INIT_LIST_HEAD(&entry->root_extent_list); + trace_btrfs_ordered_extent_add(inode, entry); + spin_lock(&tree->lock); node = tree_insert(&tree->tree, file_offset, &entry->rb_node); @@ -387,6 +389,8 @@ int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) struct list_head *cur; struct btrfs_ordered_sum *sum; + trace_btrfs_ordered_extent_put(entry->inode, entry); + if (atomic_dec_and_test(&entry->refs)) { while (!list_empty(&entry->list)) { cur = entry->list.next; @@ -420,6 +424,8 @@ static int __btrfs_remove_ordered_extent(struct inode *inode, spin_lock(&root->fs_info->ordered_extent_lock); list_del_init(&entry->root_extent_list); + trace_btrfs_ordered_extent_remove(inode, entry); + /* * we have no more ordered extents for this inode and * no dirty pages. We can safely remove it from the @@ -585,6 +591,8 @@ void btrfs_start_ordered_extent(struct inode *inode, u64 start = entry->file_offset; u64 end = start + entry->len - 1; + trace_btrfs_ordered_extent_start(inode, entry); + /* * pages in the range can be dirty, clean or writeback. We * start IO on any dirty ones so the wait doesn't stall waiting diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index d39a9895d932..2edfc039f098 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -52,6 +52,9 @@ #include "export.h" #include "compression.h" +#define CREATE_TRACE_POINTS +#include + static const struct super_operations btrfs_super_ops; static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno, @@ -620,6 +623,8 @@ int btrfs_sync_fs(struct super_block *sb, int wait) struct btrfs_root *root = btrfs_sb(sb); int ret; + trace_btrfs_sync_fs(wait); + if (!wait) { filemap_flush(root->fs_info->btree_inode->i_mapping); return 0; diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 3d73c8d93bbb..5b4bc685bb0e 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1389,6 +1389,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, put_transaction(cur_trans); put_transaction(cur_trans); + trace_btrfs_transaction_commit(root); + mutex_unlock(&root->fs_info->trans_mutex); if (current->journal_info == trans) diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index dd13eb81ee40..8ba3c9ebff93 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -33,17 +33,6 @@ #include "volumes.h" #include "async-thread.h" -struct map_lookup { - u64 type; - int io_align; - int io_width; - int stripe_len; - int sector_size; - int num_stripes; - int sub_stripes; - struct btrfs_bio_stripe stripes[]; -}; - static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_device *device); @@ -1923,6 +1912,8 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, BUG_ON(ret); + trace_btrfs_chunk_free(root, map, chunk_offset, em->len); + if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) { ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset); BUG_ON(ret); @@ -2650,6 +2641,8 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, *num_bytes = chunk_bytes_by_type(type, calc_size, map->num_stripes, sub_stripes); + trace_btrfs_chunk_alloc(info->chunk_root, map, start, *num_bytes); + em = alloc_extent_map(GFP_NOFS); if (!em) { ret = -ENOMEM; @@ -2758,6 +2751,7 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans, item_size); BUG_ON(ret); } + kfree(chunk); return 0; } diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 7fb59d45fe8c..7b38d0668b51 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -145,6 +145,17 @@ struct btrfs_device_info { u64 max_avail; }; +struct map_lookup { + u64 type; + int io_align; + int io_width; + int stripe_len; + int sector_size; + int num_stripes; + int sub_stripes; + struct btrfs_bio_stripe stripes[]; +}; + /* Used to sort the devices by max_avail(descending sort) */ int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2); diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h new file mode 100644 index 000000000000..f445cff66ab7 --- /dev/null +++ b/include/trace/events/btrfs.h @@ -0,0 +1,667 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM btrfs + +#if !defined(_TRACE_BTRFS_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_BTRFS_H + +#include +#include + +struct btrfs_root; +struct btrfs_fs_info; +struct btrfs_inode; +struct extent_map; +struct btrfs_ordered_extent; +struct btrfs_delayed_ref_node; +struct btrfs_delayed_tree_ref; +struct btrfs_delayed_data_ref; +struct btrfs_delayed_ref_head; +struct map_lookup; +struct extent_buffer; + +#define show_ref_type(type) \ + __print_symbolic(type, \ + { BTRFS_TREE_BLOCK_REF_KEY, "TREE_BLOCK_REF" }, \ + { BTRFS_EXTENT_DATA_REF_KEY, "EXTENT_DATA_REF" }, \ + { BTRFS_EXTENT_REF_V0_KEY, "EXTENT_REF_V0" }, \ + { BTRFS_SHARED_BLOCK_REF_KEY, "SHARED_BLOCK_REF" }, \ + { BTRFS_SHARED_DATA_REF_KEY, "SHARED_DATA_REF" }) + +#define __show_root_type(obj) \ + __print_symbolic(obj, \ + { BTRFS_ROOT_TREE_OBJECTID, "ROOT_TREE" }, \ + { BTRFS_EXTENT_TREE_OBJECTID, "EXTENT_TREE" }, \ + { BTRFS_CHUNK_TREE_OBJECTID, "CHUNK_TREE" }, \ + { BTRFS_DEV_TREE_OBJECTID, "DEV_TREE" }, \ + { BTRFS_FS_TREE_OBJECTID, "FS_TREE" }, \ + { BTRFS_ROOT_TREE_DIR_OBJECTID, "ROOT_TREE_DIR" }, \ + { BTRFS_CSUM_TREE_OBJECTID, "CSUM_TREE" }, \ + { BTRFS_TREE_LOG_OBJECTID, "TREE_LOG" }, \ + { BTRFS_TREE_RELOC_OBJECTID, "TREE_RELOC" }, \ + { BTRFS_DATA_RELOC_TREE_OBJECTID, "DATA_RELOC_TREE" }) + +#define show_root_type(obj) \ + obj, ((obj >= BTRFS_DATA_RELOC_TREE_OBJECTID) || \ + (obj <= BTRFS_CSUM_TREE_OBJECTID )) ? __show_root_type(obj) : "-" + +TRACE_EVENT(btrfs_transaction_commit, + + TP_PROTO(struct btrfs_root *root), + + TP_ARGS(root), + + TP_STRUCT__entry( + __field( u64, generation ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + __entry->generation = root->fs_info->generation; + __entry->root_objectid = root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), gen = %llu", + show_root_type(__entry->root_objectid), + (unsigned long long)__entry->generation) +); + +DECLARE_EVENT_CLASS(btrfs__inode, + + TP_PROTO(struct inode *inode), + + TP_ARGS(inode), + + TP_STRUCT__entry( + __field( ino_t, ino ) + __field( blkcnt_t, blocks ) + __field( u64, disk_i_size ) + __field( u64, generation ) + __field( u64, last_trans ) + __field( u64, logged_trans ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + __entry->ino = inode->i_ino; + __entry->blocks = inode->i_blocks; + __entry->disk_i_size = BTRFS_I(inode)->disk_i_size; + __entry->generation = BTRFS_I(inode)->generation; + __entry->last_trans = BTRFS_I(inode)->last_trans; + __entry->logged_trans = BTRFS_I(inode)->logged_trans; + __entry->root_objectid = + BTRFS_I(inode)->root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), gen = %llu, ino = %lu, blocks = %llu, " + "disk_i_size = %llu, last_trans = %llu, logged_trans = %llu", + show_root_type(__entry->root_objectid), + (unsigned long long)__entry->generation, + (unsigned long)__entry->ino, + (unsigned long long)__entry->blocks, + (unsigned long long)__entry->disk_i_size, + (unsigned long long)__entry->last_trans, + (unsigned long long)__entry->logged_trans) +); + +DEFINE_EVENT(btrfs__inode, btrfs_inode_new, + + TP_PROTO(struct inode *inode), + + TP_ARGS(inode) +); + +DEFINE_EVENT(btrfs__inode, btrfs_inode_request, + + TP_PROTO(struct inode *inode), + + TP_ARGS(inode) +); + +DEFINE_EVENT(btrfs__inode, btrfs_inode_evict, + + TP_PROTO(struct inode *inode), + + TP_ARGS(inode) +); + +#define __show_map_type(type) \ + __print_symbolic(type, \ + { EXTENT_MAP_LAST_BYTE, "LAST_BYTE" }, \ + { EXTENT_MAP_HOLE, "HOLE" }, \ + { EXTENT_MAP_INLINE, "INLINE" }, \ + { EXTENT_MAP_DELALLOC, "DELALLOC" }) + +#define show_map_type(type) \ + type, (type >= EXTENT_MAP_LAST_BYTE) ? "-" : __show_map_type(type) + +#define show_map_flags(flag) \ + __print_flags(flag, "|", \ + { EXTENT_FLAG_PINNED, "PINNED" }, \ + { EXTENT_FLAG_COMPRESSED, "COMPRESSED" }, \ + { EXTENT_FLAG_VACANCY, "VACANCY" }, \ + { EXTENT_FLAG_PREALLOC, "PREALLOC" }) + +TRACE_EVENT(btrfs_get_extent, + + TP_PROTO(struct btrfs_root *root, struct extent_map *map), + + TP_ARGS(root, map), + + TP_STRUCT__entry( + __field( u64, root_objectid ) + __field( u64, start ) + __field( u64, len ) + __field( u64, orig_start ) + __field( u64, block_start ) + __field( u64, block_len ) + __field( unsigned long, flags ) + __field( int, refs ) + __field( unsigned int, compress_type ) + ), + + TP_fast_assign( + __entry->root_objectid = root->root_key.objectid; + __entry->start = map->start; + __entry->len = map->len; + __entry->orig_start = map->orig_start; + __entry->block_start = map->block_start; + __entry->block_len = map->block_len; + __entry->flags = map->flags; + __entry->refs = atomic_read(&map->refs); + __entry->compress_type = map->compress_type; + ), + + TP_printk("root = %llu(%s), start = %llu, len = %llu, " + "orig_start = %llu, block_start = %llu(%s), " + "block_len = %llu, flags = %s, refs = %u, " + "compress_type = %u", + show_root_type(__entry->root_objectid), + (unsigned long long)__entry->start, + (unsigned long long)__entry->len, + (unsigned long long)__entry->orig_start, + show_map_type(__entry->block_start), + (unsigned long long)__entry->block_len, + show_map_flags(__entry->flags), + __entry->refs, __entry->compress_type) +); + +#define show_ordered_flags(flags) \ + __print_symbolic(flags, \ + { BTRFS_ORDERED_IO_DONE, "IO_DONE" }, \ + { BTRFS_ORDERED_COMPLETE, "COMPLETE" }, \ + { BTRFS_ORDERED_NOCOW, "NOCOW" }, \ + { BTRFS_ORDERED_COMPRESSED, "COMPRESSED" }, \ + { BTRFS_ORDERED_PREALLOC, "PREALLOC" }, \ + { BTRFS_ORDERED_DIRECT, "DIRECT" }) + +DECLARE_EVENT_CLASS(btrfs__ordered_extent, + + TP_PROTO(struct inode *inode, struct btrfs_ordered_extent *ordered), + + TP_ARGS(inode, ordered), + + TP_STRUCT__entry( + __field( ino_t, ino ) + __field( u64, file_offset ) + __field( u64, start ) + __field( u64, len ) + __field( u64, disk_len ) + __field( u64, bytes_left ) + __field( unsigned long, flags ) + __field( int, compress_type ) + __field( int, refs ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + __entry->ino = inode->i_ino; + __entry->file_offset = ordered->file_offset; + __entry->start = ordered->start; + __entry->len = ordered->len; + __entry->disk_len = ordered->disk_len; + __entry->bytes_left = ordered->bytes_left; + __entry->flags = ordered->flags; + __entry->compress_type = ordered->compress_type; + __entry->refs = atomic_read(&ordered->refs); + __entry->root_objectid = + BTRFS_I(inode)->root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), ino = %llu, file_offset = %llu, " + "start = %llu, len = %llu, disk_len = %llu, " + "bytes_left = %llu, flags = %s, compress_type = %d, " + "refs = %d", + show_root_type(__entry->root_objectid), + (unsigned long long)__entry->ino, + (unsigned long long)__entry->file_offset, + (unsigned long long)__entry->start, + (unsigned long long)__entry->len, + (unsigned long long)__entry->disk_len, + (unsigned long long)__entry->bytes_left, + show_ordered_flags(__entry->flags), + __entry->compress_type, __entry->refs) +); + +DEFINE_EVENT(btrfs__ordered_extent, btrfs_ordered_extent_add, + + TP_PROTO(struct inode *inode, struct btrfs_ordered_extent *ordered), + + TP_ARGS(inode, ordered) +); + +DEFINE_EVENT(btrfs__ordered_extent, btrfs_ordered_extent_remove, + + TP_PROTO(struct inode *inode, struct btrfs_ordered_extent *ordered), + + TP_ARGS(inode, ordered) +); + +DEFINE_EVENT(btrfs__ordered_extent, btrfs_ordered_extent_start, + + TP_PROTO(struct inode *inode, struct btrfs_ordered_extent *ordered), + + TP_ARGS(inode, ordered) +); + +DEFINE_EVENT(btrfs__ordered_extent, btrfs_ordered_extent_put, + + TP_PROTO(struct inode *inode, struct btrfs_ordered_extent *ordered), + + TP_ARGS(inode, ordered) +); + +DECLARE_EVENT_CLASS(btrfs__writepage, + + TP_PROTO(struct page *page, struct inode *inode, + struct writeback_control *wbc), + + TP_ARGS(page, inode, wbc), + + TP_STRUCT__entry( + __field( ino_t, ino ) + __field( pgoff_t, index ) + __field( long, nr_to_write ) + __field( long, pages_skipped ) + __field( loff_t, range_start ) + __field( loff_t, range_end ) + __field( char, nonblocking ) + __field( char, for_kupdate ) + __field( char, for_reclaim ) + __field( char, range_cyclic ) + __field( pgoff_t, writeback_index ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + __entry->ino = inode->i_ino; + __entry->index = page->index; + __entry->nr_to_write = wbc->nr_to_write; + __entry->pages_skipped = wbc->pages_skipped; + __entry->range_start = wbc->range_start; + __entry->range_end = wbc->range_end; + __entry->nonblocking = wbc->nonblocking; + __entry->for_kupdate = wbc->for_kupdate; + __entry->for_reclaim = wbc->for_reclaim; + __entry->range_cyclic = wbc->range_cyclic; + __entry->writeback_index = inode->i_mapping->writeback_index; + __entry->root_objectid = + BTRFS_I(inode)->root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), ino = %lu, page_index = %lu, " + "nr_to_write = %ld, pages_skipped = %ld, range_start = %llu, " + "range_end = %llu, nonblocking = %d, for_kupdate = %d, " + "for_reclaim = %d, range_cyclic = %d, writeback_index = %lu", + show_root_type(__entry->root_objectid), + (unsigned long)__entry->ino, __entry->index, + __entry->nr_to_write, __entry->pages_skipped, + __entry->range_start, __entry->range_end, + __entry->nonblocking, __entry->for_kupdate, + __entry->for_reclaim, __entry->range_cyclic, + (unsigned long)__entry->writeback_index) +); + +DEFINE_EVENT(btrfs__writepage, __extent_writepage, + + TP_PROTO(struct page *page, struct inode *inode, + struct writeback_control *wbc), + + TP_ARGS(page, inode, wbc) +); + +TRACE_EVENT(btrfs_writepage_end_io_hook, + + TP_PROTO(struct page *page, u64 start, u64 end, int uptodate), + + TP_ARGS(page, start, end, uptodate), + + TP_STRUCT__entry( + __field( ino_t, ino ) + __field( pgoff_t, index ) + __field( u64, start ) + __field( u64, end ) + __field( int, uptodate ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + __entry->ino = page->mapping->host->i_ino; + __entry->index = page->index; + __entry->start = start; + __entry->end = end; + __entry->uptodate = uptodate; + __entry->root_objectid = + BTRFS_I(page->mapping->host)->root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), ino = %lu, page_index = %lu, start = %llu, " + "end = %llu, uptodate = %d", + show_root_type(__entry->root_objectid), + (unsigned long)__entry->ino, (unsigned long)__entry->index, + (unsigned long long)__entry->start, + (unsigned long long)__entry->end, __entry->uptodate) +); + +TRACE_EVENT(btrfs_sync_file, + + TP_PROTO(struct file *file, int datasync), + + TP_ARGS(file, datasync), + + TP_STRUCT__entry( + __field( ino_t, ino ) + __field( ino_t, parent ) + __field( int, datasync ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + struct dentry *dentry = file->f_path.dentry; + struct inode *inode = dentry->d_inode; + + __entry->ino = inode->i_ino; + __entry->parent = dentry->d_parent->d_inode->i_ino; + __entry->datasync = datasync; + __entry->root_objectid = + BTRFS_I(inode)->root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), ino = %ld, parent = %ld, datasync = %d", + show_root_type(__entry->root_objectid), + (unsigned long)__entry->ino, (unsigned long)__entry->parent, + __entry->datasync) +); + +TRACE_EVENT(btrfs_sync_fs, + + TP_PROTO(int wait), + + TP_ARGS(wait), + + TP_STRUCT__entry( + __field( int, wait ) + ), + + TP_fast_assign( + __entry->wait = wait; + ), + + TP_printk("wait = %d", __entry->wait) +); + +#define show_ref_action(action) \ + __print_symbolic(action, \ + { BTRFS_ADD_DELAYED_REF, "ADD_DELAYED_REF" }, \ + { BTRFS_DROP_DELAYED_REF, "DROP_DELAYED_REF" }, \ + { BTRFS_ADD_DELAYED_EXTENT, "ADD_DELAYED_EXTENT" }, \ + { BTRFS_UPDATE_DELAYED_HEAD, "UPDATE_DELAYED_HEAD" }) + + +TRACE_EVENT(btrfs_delayed_tree_ref, + + TP_PROTO(struct btrfs_delayed_ref_node *ref, + struct btrfs_delayed_tree_ref *full_ref, + int action), + + TP_ARGS(ref, full_ref, action), + + TP_STRUCT__entry( + __field( u64, bytenr ) + __field( u64, num_bytes ) + __field( int, action ) + __field( u64, parent ) + __field( u64, ref_root ) + __field( int, level ) + __field( int, type ) + ), + + TP_fast_assign( + __entry->bytenr = ref->bytenr; + __entry->num_bytes = ref->num_bytes; + __entry->action = action; + __entry->parent = full_ref->parent; + __entry->ref_root = full_ref->root; + __entry->level = full_ref->level; + __entry->type = ref->type; + ), + + TP_printk("bytenr = %llu, num_bytes = %llu, action = %s, " + "parent = %llu(%s), ref_root = %llu(%s), level = %d, " + "type = %s", + (unsigned long long)__entry->bytenr, + (unsigned long long)__entry->num_bytes, + show_ref_action(__entry->action), + show_root_type(__entry->parent), + show_root_type(__entry->ref_root), + __entry->level, show_ref_type(__entry->type)) +); + +TRACE_EVENT(btrfs_delayed_data_ref, + + TP_PROTO(struct btrfs_delayed_ref_node *ref, + struct btrfs_delayed_data_ref *full_ref, + int action), + + TP_ARGS(ref, full_ref, action), + + TP_STRUCT__entry( + __field( u64, bytenr ) + __field( u64, num_bytes ) + __field( int, action ) + __field( u64, parent ) + __field( u64, ref_root ) + __field( u64, owner ) + __field( u64, offset ) + __field( int, type ) + ), + + TP_fast_assign( + __entry->bytenr = ref->bytenr; + __entry->num_bytes = ref->num_bytes; + __entry->action = action; + __entry->parent = full_ref->parent; + __entry->ref_root = full_ref->root; + __entry->owner = full_ref->objectid; + __entry->offset = full_ref->offset; + __entry->type = ref->type; + ), + + TP_printk("bytenr = %llu, num_bytes = %llu, action = %s, " + "parent = %llu(%s), ref_root = %llu(%s), owner = %llu, " + "offset = %llu, type = %s", + (unsigned long long)__entry->bytenr, + (unsigned long long)__entry->num_bytes, + show_ref_action(__entry->action), + show_root_type(__entry->parent), + show_root_type(__entry->ref_root), + (unsigned long long)__entry->owner, + (unsigned long long)__entry->offset, + show_ref_type(__entry->type)) +); + +TRACE_EVENT(btrfs_delayed_ref_head, + + TP_PROTO(struct btrfs_delayed_ref_node *ref, + struct btrfs_delayed_ref_head *head_ref, + int action), + + TP_ARGS(ref, head_ref, action), + + TP_STRUCT__entry( + __field( u64, bytenr ) + __field( u64, num_bytes ) + __field( int, action ) + __field( int, is_data ) + ), + + TP_fast_assign( + __entry->bytenr = ref->bytenr; + __entry->num_bytes = ref->num_bytes; + __entry->action = action; + __entry->is_data = head_ref->is_data; + ), + + TP_printk("bytenr = %llu, num_bytes = %llu, action = %s, is_data = %d", + (unsigned long long)__entry->bytenr, + (unsigned long long)__entry->num_bytes, + show_ref_action(__entry->action), + __entry->is_data) +); + +#define show_chunk_type(type) \ + __print_flags(type, "|", \ + { BTRFS_BLOCK_GROUP_DATA, "DATA" }, \ + { BTRFS_BLOCK_GROUP_SYSTEM, "SYSTEM"}, \ + { BTRFS_BLOCK_GROUP_METADATA, "METADATA"}, \ + { BTRFS_BLOCK_GROUP_RAID0, "RAID0" }, \ + { BTRFS_BLOCK_GROUP_RAID1, "RAID1" }, \ + { BTRFS_BLOCK_GROUP_DUP, "DUP" }, \ + { BTRFS_BLOCK_GROUP_RAID10, "RAID10"}) + +DECLARE_EVENT_CLASS(btrfs__chunk, + + TP_PROTO(struct btrfs_root *root, struct map_lookup *map, + u64 offset, u64 size), + + TP_ARGS(root, map, offset, size), + + TP_STRUCT__entry( + __field( int, num_stripes ) + __field( u64, type ) + __field( int, sub_stripes ) + __field( u64, offset ) + __field( u64, size ) + __field( u64, root_objectid ) + ), + + TP_fast_assign( + __entry->num_stripes = map->num_stripes; + __entry->type = map->type; + __entry->sub_stripes = map->sub_stripes; + __entry->offset = offset; + __entry->size = size; + __entry->root_objectid = root->root_key.objectid; + ), + + TP_printk("root = %llu(%s), offset = %llu, size = %llu, " + "num_stripes = %d, sub_stripes = %d, type = %s", + show_root_type(__entry->root_objectid), + (unsigned long long)__entry->offset, + (unsigned long long)__entry->size, + __entry->num_stripes, __entry->sub_stripes, + show_chunk_type(__entry->type)) +); + +DEFINE_EVENT(btrfs__chunk, btrfs_chunk_alloc, + + TP_PROTO(struct btrfs_root *root, struct map_lookup *map, + u64 offset, u64 size), + + TP_ARGS(root, map, offset, size) +); + +DEFINE_EVENT(btrfs__chunk, btrfs_chunk_free, + + TP_PROTO(struct btrfs_root *root, struct map_lookup *map, + u64 offset, u64 size), + + TP_ARGS(root, map, offset, size) +); + +TRACE_EVENT(btrfs_cow_block, + + TP_PROTO(struct btrfs_root *root, struct extent_buffer *buf, + struct extent_buffer *cow), + + TP_ARGS(root, buf, cow), + + TP_STRUCT__entry( + __field( u64, root_objectid ) + __field( u64, buf_start ) + __field( int, refs ) + __field( u64, cow_start ) + __field( int, buf_level ) + __field( int, cow_level ) + ), + + TP_fast_assign( + __entry->root_objectid = root->root_key.objectid; + __entry->buf_start = buf->start; + __entry->refs = atomic_read(&buf->refs); + __entry->cow_start = cow->start; + __entry->buf_level = btrfs_header_level(buf); + __entry->cow_level = btrfs_header_level(cow); + ), + + TP_printk("root = %llu(%s), refs = %d, orig_buf = %llu " + "(orig_level = %d), cow_buf = %llu (cow_level = %d)", + show_root_type(__entry->root_objectid), + __entry->refs, + (unsigned long long)__entry->buf_start, + __entry->buf_level, + (unsigned long long)__entry->cow_start, + __entry->cow_level) +); + +DECLARE_EVENT_CLASS(btrfs__reserved_extent, + + TP_PROTO(struct btrfs_root *root, u64 start, u64 len), + + TP_ARGS(root, start, len), + + TP_STRUCT__entry( + __field( u64, root_objectid ) + __field( u64, start ) + __field( u64, len ) + ), + + TP_fast_assign( + __entry->root_objectid = root->root_key.objectid; + __entry->start = start; + __entry->len = len; + ), + + TP_printk("root = %llu(%s), start = %llu, len = %llu", + show_root_type(__entry->root_objectid), + (unsigned long long)__entry->start, + (unsigned long long)__entry->len) +); + +DEFINE_EVENT(btrfs__reserved_extent, btrfs_reserved_extent_alloc, + + TP_PROTO(struct btrfs_root *root, u64 start, u64 len), + + TP_ARGS(root, start, len) +); + +DEFINE_EVENT(btrfs__reserved_extent, btrfs_reserved_extent_free, + + TP_PROTO(struct btrfs_root *root, u64 start, u64 len), + + TP_ARGS(root, start, len) +); + +#endif /* _TRACE_BTRFS_H */ + +/* This part must be outside protection */ +#include From db5b493ac78e46c7b6bad22cd25d8041564cd8ea Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Wed, 23 Mar 2011 08:14:16 +0000 Subject: [PATCH 25/45] Btrfs: cleanup some BUG_ON() This patch changes some BUG_ON() to the error return. (but, most callers still use BUG_ON()) Signed-off-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 3 ++- fs/btrfs/disk-io.c | 5 ++++- fs/btrfs/extent-tree.c | 25 ++++++++++++++++++------- fs/btrfs/file-item.c | 3 ++- fs/btrfs/inode-map.c | 3 ++- fs/btrfs/ioctl.c | 5 ++++- fs/btrfs/root-tree.c | 6 ++++-- fs/btrfs/transaction.c | 12 +++++++++--- fs/btrfs/tree-log.c | 15 +++++++++------ 9 files changed, 54 insertions(+), 23 deletions(-) diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 465b5d7d6b48..4edcbe915736 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -3709,7 +3709,8 @@ int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root unsigned long ptr; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size); if (!ret) { leaf = path->nodes[0]; diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 9f31e110b481..00cbb41af660 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1248,7 +1248,10 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, root, fs_info, location->objectid); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + kfree(root); + return ERR_PTR(-ENOMEM); + } ret = btrfs_search_slot(NULL, tree_root, location, path, 0, 0); if (ret == 0) { l = path->nodes[0]; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 86ea471d3801..a6a8159c5d1e 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5463,7 +5463,8 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; path->leave_spinning = 1; ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, @@ -6457,10 +6458,14 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID); path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; wc = kzalloc(sizeof(*wc), GFP_NOFS); - BUG_ON(!wc); + if (!wc) { + btrfs_free_path(path); + return -ENOMEM; + } btrfs_assert_tree_locked(parent); parent_level = btrfs_header_level(parent); @@ -6918,7 +6923,11 @@ static noinline int get_new_locations(struct inode *reloc_inode, } path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + if (exts != *extents) + kfree(exts); + return -ENOMEM; + } cur_pos = extent_key->objectid - offset; last_byte = extent_key->objectid + extent_key->offset; @@ -7442,7 +7451,8 @@ static noinline int replace_extents_in_leaf(struct btrfs_trans_handle *trans, int ret; new_extent = kmalloc(sizeof(*new_extent), GFP_NOFS); - BUG_ON(!new_extent); + if (!new_extent) + return -ENOMEM; ref = btrfs_lookup_leaf_ref(root, leaf->start); BUG_ON(!ref); @@ -7647,7 +7657,8 @@ static noinline int init_reloc_tree(struct btrfs_trans_handle *trans, return 0; root_item = kmalloc(sizeof(*root_item), GFP_NOFS); - BUG_ON(!root_item); + if (!root_item) + return -ENOMEM; ret = btrfs_copy_root(trans, root, root->commit_root, &eb, BTRFS_TREE_RELOC_OBJECTID); @@ -7673,7 +7684,7 @@ static noinline int init_reloc_tree(struct btrfs_trans_handle *trans, reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root, &root_key); - BUG_ON(!reloc_root); + BUG_ON(IS_ERR(reloc_root)); reloc_root->last_trans = trans->transid; reloc_root->commit_root = NULL; reloc_root->ref_tree = &root->fs_info->reloc_ref_tree; diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 4f19a3e1bf32..a2134195a85e 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -48,7 +48,8 @@ int btrfs_insert_file_extent(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; file_key.objectid = objectid; file_key.offset = pos; btrfs_set_key_type(&file_key, BTRFS_EXTENT_DATA_KEY); diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index c56eb5909172..c05a08f4c411 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -30,7 +30,8 @@ int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid) int slot; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; search_key.objectid = BTRFS_LAST_FREE_OBJECTID; search_key.type = -1; diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index ad9b8c0e930b..88d3cb2eaf75 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2350,12 +2350,15 @@ static noinline long btrfs_ioctl_start_sync(struct file *file, void __user *argp struct btrfs_root *root = BTRFS_I(file->f_dentry->d_inode)->root; struct btrfs_trans_handle *trans; u64 transid; + int ret; trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) return PTR_ERR(trans); transid = trans->transid; - btrfs_commit_transaction_async(trans, root, 0); + ret = btrfs_commit_transaction_async(trans, root, 0); + if (ret) + return ret; if (argp) if (copy_to_user(argp, &transid, sizeof(transid))) diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index 6a1086e83ffc..29b2d7c930eb 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -88,7 +88,8 @@ int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, search_key.offset = (u64)-1; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0); if (ret < 0) goto out; @@ -332,7 +333,8 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *leaf; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; ret = btrfs_search_slot(trans, root, key, path, -1, 1); if (ret < 0) goto out; diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 5b4bc685bb0e..ce48eb59d615 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -57,7 +57,8 @@ static noinline int join_transaction(struct btrfs_root *root) if (!cur_trans) { cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS); - BUG_ON(!cur_trans); + if (!cur_trans) + return -ENOMEM; root->fs_info->generation++; cur_trans->num_writers = 1; cur_trans->num_joined = 0; @@ -195,7 +196,11 @@ again: wait_current_trans(root); ret = join_transaction(root); - BUG_ON(ret); + if (ret < 0) { + if (type != TRANS_JOIN_NOLOCK) + mutex_unlock(&root->fs_info->trans_mutex); + return ERR_PTR(ret); + } cur_trans = root->fs_info->running_transaction; cur_trans->use_count++; @@ -1156,7 +1161,8 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, struct btrfs_transaction *cur_trans; ac = kmalloc(sizeof(*ac), GFP_NOFS); - BUG_ON(!ac); + if (!ac) + return -ENOMEM; INIT_DELAYED_WORK(&ac->work, do_async_commit); ac->root = root; diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 429cfcfadf90..f9425e33e358 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -1828,7 +1828,8 @@ static int walk_log_tree(struct btrfs_trans_handle *trans, int orig_level; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; level = btrfs_header_level(log->node); orig_level = level; @@ -3114,9 +3115,11 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree) .stage = 0, }; - fs_info->log_root_recovering = 1; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + + fs_info->log_root_recovering = 1; trans = btrfs_start_transaction(fs_info->tree_root, 0); BUG_ON(IS_ERR(trans)); @@ -3124,7 +3127,8 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree) wc.trans = trans; wc.pin = 1; - walk_log_tree(trans, log_root_tree, &wc); + ret = walk_log_tree(trans, log_root_tree, &wc); + BUG_ON(ret); again: key.objectid = BTRFS_TREE_LOG_OBJECTID; @@ -3148,8 +3152,7 @@ again: log = btrfs_read_fs_root_no_radix(log_root_tree, &found_key); - BUG_ON(!log); - + BUG_ON(IS_ERR(log)); tmp_key.objectid = found_key.offset; tmp_key.type = BTRFS_ROOT_ITEM_KEY; From 7e75bf3ff3a716d7b21d8fb43bf823115801c1e9 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 18 Mar 2011 22:56:43 +0000 Subject: [PATCH 26/45] btrfs: properly access unaligned checksum buffer On Fri, Mar 18, 2011 at 11:56:53AM -0400, Chris Mason wrote: > Thanks for fielding this one. Does put_unaligned_le32 optimize away on > platforms with efficient access? It would be great if we didn't need > the #ifdef. (quicktest: assembly output is same for put_unaligned_le32 and direct assignment on my x86_64) I was originally following examples in Documentation/unaligned-memory-access.txt. From other code it seems to me that the define CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS is intended for larger portions of code. Macros/wrappers for {put,get}_unaligned* are chosen via arch//include/asm/unaligned.h accordingly, therefore it's safe to use put_unaligned_le32 without the ifdef. dave Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 00cbb41af660..2bdb124333ab 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -29,6 +29,7 @@ #include #include #include +#include #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -198,7 +199,7 @@ u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len) void btrfs_csum_final(u32 crc, char *result) { - *(__le32 *)result = ~cpu_to_le32(crc); + put_unaligned_le32(~crc, result); } /* From 97d9a8a420444eb5b5c071d4b3b9c4100a7ae015 Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Thu, 24 Mar 2011 06:33:21 +0000 Subject: [PATCH 27/45] Btrfs: check return value of read_tree_block() This patch is checking return value of read_tree_block(), and if it is NULL, error processing. Signed-off-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 3 +++ fs/btrfs/extent-tree.c | 6 ++++++ fs/btrfs/relocation.c | 6 ++++++ 3 files changed, 15 insertions(+) diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 4edcbe915736..84d7ca1fe0ba 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -682,6 +682,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, if (!cur) { cur = read_tree_block(root, blocknr, blocksize, gen); + if (!cur) + return -EIO; } else if (!uptodate) { btrfs_read_buffer(cur, gen); } @@ -4087,6 +4089,7 @@ find_next_key: } btrfs_set_path_blocking(path); cur = read_node_slot(root, cur, slot); + BUG_ON(!cur); btrfs_tree_lock(cur); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a6a8159c5d1e..5bc658a9d85c 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6067,6 +6067,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, if (reada && level == 1) reada_walk_down(trans, root, wc, path); next = read_tree_block(root, bytenr, blocksize, generation); + if (!next) + return -EIO; btrfs_tree_lock(next); btrfs_set_lock_blocking(next); } @@ -7937,6 +7939,10 @@ static noinline int relocate_one_extent(struct btrfs_root *extent_root, eb = read_tree_block(found_root, block_start, block_size, 0); + if (!eb) { + ret = -EIO; + goto out; + } btrfs_tree_lock(eb); BUG_ON(level != btrfs_header_level(eb)); diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index c863c8447015..58250e09eb05 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -1724,6 +1724,7 @@ again: eb = read_tree_block(dest, old_bytenr, blocksize, old_ptr_gen); + BUG_ON(!eb); btrfs_tree_lock(eb); if (cow) { ret = btrfs_cow_block(trans, dest, eb, parent, @@ -2513,6 +2514,10 @@ static int do_relocation(struct btrfs_trans_handle *trans, blocksize = btrfs_level_size(root, node->level); generation = btrfs_node_ptr_generation(upper->eb, slot); eb = read_tree_block(root, bytenr, blocksize, generation); + if (!eb) { + err = -EIO; + goto next; + } btrfs_tree_lock(eb); btrfs_set_lock_blocking(eb); @@ -2670,6 +2675,7 @@ static int get_tree_block_key(struct reloc_control *rc, BUG_ON(block->key_ready); eb = read_tree_block(rc->extent_root, block->bytenr, block->key.objectid, block->key.offset); + BUG_ON(!eb); WARN_ON(btrfs_header_level(eb) != block->level); if (block->level == 0) btrfs_item_key_to_cpu(eb, &block->key, 0); From fc0e4a314e361af3b13d9320e92c64118f9a3e61 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Thu, 24 Mar 2011 11:41:21 +0000 Subject: [PATCH 28/45] btrfs: use GFP_NOFS instead of GFP_KERNEL In the filesystem context, we must allocate memory by GFP_NOFS, or we may start another filesystem operation and make kswap thread hang up. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5bc658a9d85c..7922f296420d 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -471,7 +471,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, if (load_cache_only) return 0; - caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_KERNEL); + caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); BUG_ON(!caching_ctl); INIT_LIST_HEAD(&caching_ctl->list); @@ -1743,7 +1743,7 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, static void btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len) { - blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL, 0); + blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0); } static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, From 32471f6e1983922473573da62cbee58699574aa4 Mon Sep 17 00:00:00 2001 From: liubo Date: Mon, 21 Mar 2011 08:54:27 +0000 Subject: [PATCH 29/45] Btrfs: add datacow flag in inode flag For datacow control, the corresponding inode flags are needed. This is for btrfs use. v1->v2: Change FS_COW_FL to another bit due to conflict with the upstream e2fsprogs Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- include/linux/fs.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/linux/fs.h b/include/linux/fs.h index e38b50a4b9d2..de9dd8119b71 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -364,6 +364,8 @@ struct inodes_stat_t { #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ #define FS_EXTENT_FL 0x00080000 /* Extents */ #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */ +#define FS_NOCOW_FL 0x00800000 /* Do not cow file */ +#define FS_COW_FL 0x02000000 /* Cow file */ #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */ #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ From 75e7cb7fe0c391561bd3af36515be3f3c64a04c6 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 22 Mar 2011 10:12:20 +0000 Subject: [PATCH 30/45] Btrfs: Per file/directory controls for COW and compression Data compression and data cow are controlled across the entire FS by mount options right now. ioctls are needed to set this on a per file or per directory basis. This has been proposed previously, but VFS developers wanted us to use generic ioctls rather than btrfs-specific ones. According to Chris's comment, there should be just one true compression method(probably LZO) stored in the super. However, before this, we would wait for that one method is stable enough to be adopted into the super. So I list it as a long term goal, and just store it in ram today. After applying this patch, we can use the generic "FS_IOC_SETFLAGS" ioctl to control file and directory's datacow and compression attribute. NOTE: - The compression type is selected by such rules: If we mount btrfs with compress options, ie, zlib/lzo, the type is it. Otherwise, we'll use the default compress type (zlib today). v1->v2: - rebase to the latest btrfs. v2->v3: - fix a problem, i.e. when a file is set NOCOW via mount option, then this NOCOW will be screwed by inheritance from parent directory. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 1 + fs/btrfs/disk-io.c | 6 ++++++ fs/btrfs/inode.c | 31 ++++++++++++++++++++++++++++--- fs/btrfs/ioctl.c | 41 +++++++++++++++++++++++++++++++++++++---- 4 files changed, 72 insertions(+), 7 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 9d0f59142afa..8302ecd4197f 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1282,6 +1282,7 @@ struct btrfs_root { #define BTRFS_INODE_NODUMP (1 << 8) #define BTRFS_INODE_NOATIME (1 << 9) #define BTRFS_INODE_DIRSYNC (1 << 10) +#define BTRFS_INODE_COMPRESS (1 << 11) /* some macros to generate set/get funcs for the struct fields. This * assumes there is a lefoo_to_cpu for every type, so lets make a simple diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 2bdb124333ab..125639ddaffe 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1854,6 +1854,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY); + /* + * In the long term, we'll store the compression type in the super + * block, and it'll be used for per file compression control. + */ + fs_info->compress_type = BTRFS_COMPRESS_ZLIB; + ret = btrfs_parse_options(tree_root, options); if (ret) { err = ret; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index eaa271484199..7a7a202b82ab 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -384,7 +384,8 @@ again: */ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) && (btrfs_test_opt(root, COMPRESS) || - (BTRFS_I(inode)->force_compress))) { + (BTRFS_I(inode)->force_compress) || + (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); @@ -1256,7 +1257,8 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page, ret = run_delalloc_nocow(inode, locked_page, start, end, page_started, 0, nr_written); else if (!btrfs_test_opt(root, COMPRESS) && - !(BTRFS_I(inode)->force_compress)) + !(BTRFS_I(inode)->force_compress) && + !(BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS)) ret = cow_file_range(inode, locked_page, start, end, page_started, nr_written, 1); else @@ -4584,7 +4586,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, if ((mode & S_IFREG)) { if (btrfs_test_opt(root, NODATASUM)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; - if (btrfs_test_opt(root, NODATACOW)) + if (btrfs_test_opt(root, NODATACOW) || + (BTRFS_I(dir)->flags & BTRFS_INODE_NODATACOW)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW; } @@ -6866,6 +6869,26 @@ static int btrfs_getattr(struct vfsmount *mnt, return 0; } +/* + * If a file is moved, it will inherit the cow and compression flags of the new + * directory. + */ +static void fixup_inode_flags(struct inode *dir, struct inode *inode) +{ + struct btrfs_inode *b_dir = BTRFS_I(dir); + struct btrfs_inode *b_inode = BTRFS_I(inode); + + if (b_dir->flags & BTRFS_INODE_NODATACOW) + b_inode->flags |= BTRFS_INODE_NODATACOW; + else + b_inode->flags &= ~BTRFS_INODE_NODATACOW; + + if (b_dir->flags & BTRFS_INODE_COMPRESS) + b_inode->flags |= BTRFS_INODE_COMPRESS; + else + b_inode->flags &= ~BTRFS_INODE_COMPRESS; +} + static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry) { @@ -6999,6 +7022,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, } } + fixup_inode_flags(new_dir, old_inode); + ret = btrfs_add_link(trans, new_dir, old_inode, new_dentry->d_name.name, new_dentry->d_name.len, 0, index); diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 88d3cb2eaf75..32c980ae0f1c 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -138,6 +138,24 @@ static int btrfs_ioctl_getflags(struct file *file, void __user *arg) return 0; } +static int check_flags(unsigned int flags) +{ + if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ + FS_NOATIME_FL | FS_NODUMP_FL | \ + FS_SYNC_FL | FS_DIRSYNC_FL | \ + FS_NOCOMP_FL | FS_COMPR_FL | \ + FS_NOCOW_FL | FS_COW_FL)) + return -EOPNOTSUPP; + + if ((flags & FS_NOCOMP_FL) && (flags & FS_COMPR_FL)) + return -EINVAL; + + if ((flags & FS_NOCOW_FL) && (flags & FS_COW_FL)) + return -EINVAL; + + return 0; +} + static int btrfs_ioctl_setflags(struct file *file, void __user *arg) { struct inode *inode = file->f_path.dentry->d_inode; @@ -153,10 +171,9 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) if (copy_from_user(&flags, arg, sizeof(flags))) return -EFAULT; - if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ - FS_NOATIME_FL | FS_NODUMP_FL | \ - FS_SYNC_FL | FS_DIRSYNC_FL)) - return -EOPNOTSUPP; + ret = check_flags(flags); + if (ret) + return ret; if (!is_owner_or_cap(inode)) return -EACCES; @@ -201,6 +218,22 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) else ip->flags &= ~BTRFS_INODE_DIRSYNC; + /* + * The COMPRESS flag can only be changed by users, while the NOCOMPRESS + * flag may be changed automatically if compression code won't make + * things smaller. + */ + if (flags & FS_NOCOMP_FL) { + ip->flags &= ~BTRFS_INODE_COMPRESS; + ip->flags |= BTRFS_INODE_NOCOMPRESS; + } else if (flags & FS_COMPR_FL) { + ip->flags |= BTRFS_INODE_COMPRESS; + ip->flags &= ~BTRFS_INODE_NOCOMPRESS; + } + if (flags & FS_NOCOW_FL) + ip->flags |= BTRFS_INODE_NODATACOW; + else if (flags & FS_COW_FL) + ip->flags &= ~BTRFS_INODE_NODATACOW; trans = btrfs_join_transaction(root, 1); BUG_ON(IS_ERR(trans)); From 3ab3564f018b9b265d0258e4a231794bacd5ad85 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 22 Mar 2011 17:20:26 +0000 Subject: [PATCH 31/45] btrfs: return EXDEV when linking from different subvolumes btrfs_link returns EPERM if a cross-subvolume link is attempted. However, in this case I believe EXDEV to be the more appropriate value. >From the link(2) man page: EXDEV oldpath and newpath are not on the same mounted file system. (Linux permits a file system to be mounted at multiple points, but link() does not work across different mount points, even if the same file system is mounted on both.) This matters because an application may have different behaviors based on return codes. Signed-off-by: Mark Fasheh Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 7a7a202b82ab..67fd6e9552d3 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4817,7 +4817,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, /* do not allow sys_link's with other subvols of the same device */ if (root->objectid != BTRFS_I(inode)->root->objectid) - return -EPERM; + return -EXDEV; btrfs_inc_nlink(inode); inode->i_ctime = CURRENT_TIME; From b4d00d569a49fcef02195635dbf8d15904b1fb63 Mon Sep 17 00:00:00 2001 From: Li Dongyang Date: Thu, 24 Mar 2011 10:24:25 +0000 Subject: [PATCH 32/45] Btrfs: make update_reserved_bytes() public Make the function public as we should update the reserved extents calculations after taking out an extent for trimming. Signed-off-by: Li Dongyang Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 2 ++ fs/btrfs/extent-tree.c | 16 +++++++--------- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 8302ecd4197f..9e21176cdf57 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2157,6 +2157,8 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, u64 root_objectid, u64 owner, u64 offset); int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len); +int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve, int sinfo); int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 7922f296420d..0671f5b77eb8 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -36,8 +36,6 @@ static int update_block_group(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, int alloc); -static int update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo); static int __btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, u64 parent, @@ -4233,8 +4231,8 @@ int btrfs_pin_extent(struct btrfs_root *root, * update size of reserved extents. this function may return -EAGAIN * if 'reserve' is true or 'sinfo' is false. */ -static int update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo) +int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve, int sinfo) { int ret = 0; if (sinfo) { @@ -4714,10 +4712,10 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)); btrfs_add_free_space(cache, buf->start, buf->len); - ret = update_reserved_bytes(cache, buf->len, 0, 0); + ret = btrfs_update_reserved_bytes(cache, buf->len, 0, 0); if (ret == -EAGAIN) { /* block group became read-only */ - update_reserved_bytes(cache, buf->len, 0, 1); + btrfs_update_reserved_bytes(cache, buf->len, 0, 1); goto out; } @@ -5206,7 +5204,7 @@ checks: search_start - offset); BUG_ON(offset > search_start); - ret = update_reserved_bytes(block_group, num_bytes, 1, + ret = btrfs_update_reserved_bytes(block_group, num_bytes, 1, (data & BTRFS_BLOCK_GROUP_DATA)); if (ret == -EAGAIN) { btrfs_add_free_space(block_group, offset, num_bytes); @@ -5432,7 +5430,7 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) ret = btrfs_discard_extent(root, start, len); btrfs_add_free_space(cache, start, len); - update_reserved_bytes(cache, len, 0, 1); + btrfs_update_reserved_bytes(cache, len, 0, 1); btrfs_put_block_group(cache); trace_btrfs_reserved_extent_free(root, start, len); @@ -5634,7 +5632,7 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, put_caching_control(caching_ctl); } - ret = update_reserved_bytes(block_group, ins->offset, 1, 1); + ret = btrfs_update_reserved_bytes(block_group, ins->offset, 1, 1); BUG_ON(ret); btrfs_put_block_group(block_group); ret = alloc_reserved_file_extent(trans, root, 0, root_objectid, From fce3bb9a1bd492793170e117c60d5718b7896af4 Mon Sep 17 00:00:00 2001 From: Li Dongyang Date: Thu, 24 Mar 2011 10:24:26 +0000 Subject: [PATCH 33/45] Btrfs: make btrfs_map_block() return entire free extent for each device of RAID0/1/10/DUP btrfs_map_block() will only return a single stripe length, but we want the full extent be mapped to each disk when we are trimming the extent, so we add length to btrfs_bio_stripe and fill it if we are mapping for REQ_DISCARD. Signed-off-by: Li Dongyang Signed-off-by: Chris Mason --- fs/btrfs/volumes.c | 150 ++++++++++++++++++++++++++++++++++++++------- fs/btrfs/volumes.h | 1 + 2 files changed, 129 insertions(+), 22 deletions(-) diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 8ba3c9ebff93..c440c89a470a 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2956,7 +2956,10 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, struct extent_map_tree *em_tree = &map_tree->map_tree; u64 offset; u64 stripe_offset; + u64 stripe_end_offset; u64 stripe_nr; + u64 stripe_nr_orig; + u64 stripe_nr_end; int stripes_allocated = 8; int stripes_required = 1; int stripe_index; @@ -2965,7 +2968,7 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, int max_errors = 0; struct btrfs_multi_bio *multi = NULL; - if (multi_ret && !(rw & REQ_WRITE)) + if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD))) stripes_allocated = 1; again: if (multi_ret) { @@ -3011,7 +3014,15 @@ again: max_errors = 1; } } - if (multi_ret && (rw & REQ_WRITE) && + if (rw & REQ_DISCARD) { + if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | + BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_DUP | + BTRFS_BLOCK_GROUP_RAID10)) { + stripes_required = map->num_stripes; + } + } + if (multi_ret && (rw & (REQ_WRITE | REQ_DISCARD)) && stripes_allocated < stripes_required) { stripes_allocated = map->num_stripes; free_extent_map(em); @@ -3031,12 +3042,15 @@ again: /* stripe_offset is the offset of this block in its stripe*/ stripe_offset = offset - stripe_offset; - if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10 | - BTRFS_BLOCK_GROUP_DUP)) { + if (rw & REQ_DISCARD) + *length = min_t(u64, em->len - offset, *length); + else if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | + BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10 | + BTRFS_BLOCK_GROUP_DUP)) { /* we limit the length of each bio to what fits in a stripe */ *length = min_t(u64, em->len - offset, - map->stripe_len - stripe_offset); + map->stripe_len - stripe_offset); } else { *length = em->len - offset; } @@ -3046,8 +3060,19 @@ again: num_stripes = 1; stripe_index = 0; - if (map->type & BTRFS_BLOCK_GROUP_RAID1) { - if (unplug_page || (rw & REQ_WRITE)) + stripe_nr_orig = stripe_nr; + stripe_nr_end = (offset + *length + map->stripe_len - 1) & + (~(map->stripe_len - 1)); + do_div(stripe_nr_end, map->stripe_len); + stripe_end_offset = stripe_nr_end * map->stripe_len - + (offset + *length); + if (map->type & BTRFS_BLOCK_GROUP_RAID0) { + if (rw & REQ_DISCARD) + num_stripes = min_t(u64, map->num_stripes, + stripe_nr_end - stripe_nr_orig); + stripe_index = do_div(stripe_nr, map->num_stripes); + } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { + if (unplug_page || (rw & (REQ_WRITE | REQ_DISCARD))) num_stripes = map->num_stripes; else if (mirror_num) stripe_index = mirror_num - 1; @@ -3058,7 +3083,7 @@ again: } } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { - if (rw & REQ_WRITE) + if (rw & (REQ_WRITE | REQ_DISCARD)) num_stripes = map->num_stripes; else if (mirror_num) stripe_index = mirror_num - 1; @@ -3071,6 +3096,10 @@ again: if (unplug_page || (rw & REQ_WRITE)) num_stripes = map->sub_stripes; + else if (rw & REQ_DISCARD) + num_stripes = min_t(u64, map->sub_stripes * + (stripe_nr_end - stripe_nr_orig), + map->num_stripes); else if (mirror_num) stripe_index += mirror_num - 1; else { @@ -3088,24 +3117,101 @@ again: } BUG_ON(stripe_index >= map->num_stripes); - for (i = 0; i < num_stripes; i++) { - if (unplug_page) { - struct btrfs_device *device; - struct backing_dev_info *bdi; - - device = map->stripes[stripe_index].dev; - if (device->bdev) { - bdi = blk_get_backing_dev_info(device->bdev); - if (bdi->unplug_io_fn) - bdi->unplug_io_fn(bdi, unplug_page); - } - } else { + if (rw & REQ_DISCARD) { + for (i = 0; i < num_stripes; i++) { multi->stripes[i].physical = map->stripes[stripe_index].physical + stripe_offset + stripe_nr * map->stripe_len; multi->stripes[i].dev = map->stripes[stripe_index].dev; + + if (map->type & BTRFS_BLOCK_GROUP_RAID0) { + u64 stripes; + int last_stripe = (stripe_nr_end - 1) % + map->num_stripes; + int j; + + for (j = 0; j < map->num_stripes; j++) { + if ((stripe_nr_end - 1 - j) % + map->num_stripes == stripe_index) + break; + } + stripes = stripe_nr_end - 1 - j; + do_div(stripes, map->num_stripes); + multi->stripes[i].length = map->stripe_len * + (stripes - stripe_nr + 1); + + if (i == 0) { + multi->stripes[i].length -= + stripe_offset; + stripe_offset = 0; + } + if (stripe_index == last_stripe) + multi->stripes[i].length -= + stripe_end_offset; + } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { + u64 stripes; + int j; + int factor = map->num_stripes / + map->sub_stripes; + int last_stripe = (stripe_nr_end - 1) % factor; + last_stripe *= map->sub_stripes; + + for (j = 0; j < factor; j++) { + if ((stripe_nr_end - 1 - j) % factor == + stripe_index / map->sub_stripes) + break; + } + stripes = stripe_nr_end - 1 - j; + do_div(stripes, factor); + multi->stripes[i].length = map->stripe_len * + (stripes - stripe_nr + 1); + + if (i < map->sub_stripes) { + multi->stripes[i].length -= + stripe_offset; + if (i == map->sub_stripes - 1) + stripe_offset = 0; + } + if (stripe_index >= last_stripe && + stripe_index <= (last_stripe + + map->sub_stripes - 1)) { + multi->stripes[i].length -= + stripe_end_offset; + } + } else + multi->stripes[i].length = *length; + + stripe_index++; + if (stripe_index == map->num_stripes) { + /* This could only happen for RAID0/10 */ + stripe_index = 0; + stripe_nr++; + } + } + } else { + for (i = 0; i < num_stripes; i++) { + if (unplug_page) { + struct btrfs_device *device; + struct backing_dev_info *bdi; + + device = map->stripes[stripe_index].dev; + if (device->bdev) { + bdi = blk_get_backing_dev_info(device-> + bdev); + if (bdi->unplug_io_fn) + bdi->unplug_io_fn(bdi, + unplug_page); + } + } else { + multi->stripes[i].physical = + map->stripes[stripe_index].physical + + stripe_offset + + stripe_nr * map->stripe_len; + multi->stripes[i].dev = + map->stripes[stripe_index].dev; + } + stripe_index++; } - stripe_index++; } if (multi_ret) { *multi_ret = multi; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 7b38d0668b51..cc2eadaf7a27 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -126,6 +126,7 @@ struct btrfs_fs_devices { struct btrfs_bio_stripe { struct btrfs_device *dev; u64 physical; + u64 length; /* only used for discard mappings */ }; struct btrfs_multi_bio { From 5378e60734f5b7bfe1b43dc191aaf6131c1befe7 Mon Sep 17 00:00:00 2001 From: Li Dongyang Date: Thu, 24 Mar 2011 10:24:27 +0000 Subject: [PATCH 34/45] Btrfs: adjust btrfs_discard_extent() return errors and trimmed bytes Callers of btrfs_discard_extent() should check if we are mounted with -o discard, as we want to make fitrim to work even the fs is not mounted with -o discard. Also we should use REQ_DISCARD to map the free extent to get a full mapping, last we only return errors if 1. the error is not a EOPNOTSUPP 2. no device supports discard Signed-off-by: Li Dongyang Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 2 +- fs/btrfs/disk-io.c | 5 ++++- fs/btrfs/extent-tree.c | 43 +++++++++++++++++++++++++----------------- 3 files changed, 31 insertions(+), 19 deletions(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 9e21176cdf57..a18b7bc2b22c 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2229,7 +2229,7 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo); int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end); int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, - u64 num_bytes); + u64 num_bytes, u64 *actual_bytes); int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 type); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 125639ddaffe..8ecc5419d8b6 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3054,7 +3054,10 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root, break; /* opt_discard */ - ret = btrfs_error_discard_extent(root, start, end + 1 - start); + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_error_discard_extent(root, start, + end + 1 - start, + NULL); clear_extent_dirty(unpin, start, end, GFP_NOFS); btrfs_error_unpin_extent_range(root, start, end); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 0671f5b77eb8..e990d2d1ba4a 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -1738,39 +1738,45 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, return ret; } -static void btrfs_issue_discard(struct block_device *bdev, +static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len) { - blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0); + return blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0); } static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, - u64 num_bytes) + u64 num_bytes, u64 *actual_bytes) { int ret; - u64 map_length = num_bytes; + u64 discarded_bytes = 0; struct btrfs_multi_bio *multi = NULL; - if (!btrfs_test_opt(root, DISCARD)) - return 0; /* Tell the block device(s) that the sectors can be discarded */ - ret = btrfs_map_block(&root->fs_info->mapping_tree, READ, - bytenr, &map_length, &multi, 0); + ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD, + bytenr, &num_bytes, &multi, 0); if (!ret) { struct btrfs_bio_stripe *stripe = multi->stripes; int i; - if (map_length > num_bytes) - map_length = num_bytes; for (i = 0; i < multi->num_stripes; i++, stripe++) { - btrfs_issue_discard(stripe->dev->bdev, - stripe->physical, - map_length); + ret = btrfs_issue_discard(stripe->dev->bdev, + stripe->physical, + stripe->length); + if (!ret) + discarded_bytes += stripe->length; + else if (ret != -EOPNOTSUPP) + break; } kfree(multi); } + if (discarded_bytes && ret == -EOPNOTSUPP) + ret = 0; + + if (actual_bytes) + *actual_bytes = discarded_bytes; + return ret; } @@ -4371,7 +4377,9 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, if (ret) break; - ret = btrfs_discard_extent(root, start, end + 1 - start); + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_discard_extent(root, start, + end + 1 - start, NULL); clear_extent_dirty(unpin, start, end, GFP_NOFS); unpin_extent_range(root, start, end); @@ -5427,7 +5435,8 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) return -ENOSPC; } - ret = btrfs_discard_extent(root, start, len); + if (btrfs_test_opt(root, DISCARD)) + ret = btrfs_discard_extent(root, start, len, NULL); btrfs_add_free_space(cache, start, len); btrfs_update_reserved_bytes(cache, len, 0, 1); @@ -8765,7 +8774,7 @@ int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) } int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, - u64 num_bytes) + u64 num_bytes, u64 *actual_bytes) { - return btrfs_discard_extent(root, bytenr, num_bytes); + return btrfs_discard_extent(root, bytenr, num_bytes, actual_bytes); } From f7039b1d5c32241f87a513e33120db36bf30264d Mon Sep 17 00:00:00 2001 From: Li Dongyang Date: Thu, 24 Mar 2011 10:24:28 +0000 Subject: [PATCH 35/45] Btrfs: add btrfs_trim_fs() to handle FITRIM We take an free extent out from allocator, trim it, then put it back, but before we trim the block group, we should make sure the block group is cached, so plus a little change to make cache_block_group() run without a transaction. Signed-off-by: Li Dongyang Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 1 + fs/btrfs/extent-tree.c | 50 +++++++++++++++++++- fs/btrfs/free-space-cache.c | 92 +++++++++++++++++++++++++++++++++++++ fs/btrfs/free-space-cache.h | 2 + fs/btrfs/ioctl.c | 46 +++++++++++++++++++ 5 files changed, 190 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index a18b7bc2b22c..93a0191aded6 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2232,6 +2232,7 @@ int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, u64 num_bytes, u64 *actual_bytes); int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 type); +int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range); /* ctree.c */ int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key, diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index e990d2d1ba4a..1efeda3b2f6f 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -440,7 +440,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, * allocate blocks for the tree root we can't do the fast caching since * we likely hold important locks. */ - if (!trans->transaction->in_commit && + if (trans && (!trans->transaction->in_commit) && (root && root != root->fs_info->tree_root)) { spin_lock(&cache->lock); if (cache->cached != BTRFS_CACHE_NO) { @@ -8778,3 +8778,51 @@ int btrfs_error_discard_extent(struct btrfs_root *root, u64 bytenr, { return btrfs_discard_extent(root, bytenr, num_bytes, actual_bytes); } + +int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_block_group_cache *cache = NULL; + u64 group_trimmed; + u64 start; + u64 end; + u64 trimmed = 0; + int ret = 0; + + cache = btrfs_lookup_block_group(fs_info, range->start); + + while (cache) { + if (cache->key.objectid >= (range->start + range->len)) { + btrfs_put_block_group(cache); + break; + } + + start = max(range->start, cache->key.objectid); + end = min(range->start + range->len, + cache->key.objectid + cache->key.offset); + + if (end - start >= range->minlen) { + if (!block_group_cache_done(cache)) { + ret = cache_block_group(cache, NULL, root, 0); + if (!ret) + wait_block_group_cache_done(cache); + } + ret = btrfs_trim_block_group(cache, + &group_trimmed, + start, + end, + range->minlen); + + trimmed += group_trimmed; + if (ret) { + btrfs_put_block_group(cache); + break; + } + } + + cache = next_block_group(fs_info->tree_root, cache); + } + + range->len = trimmed; + return ret; +} diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index f03ef97c3b21..0037427d8a9d 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -2178,3 +2178,95 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) cluster->block_group = NULL; } +int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, + u64 *trimmed, u64 start, u64 end, u64 minlen) +{ + struct btrfs_free_space *entry = NULL; + struct btrfs_fs_info *fs_info = block_group->fs_info; + u64 bytes = 0; + u64 actually_trimmed; + int ret = 0; + + *trimmed = 0; + + while (start < end) { + spin_lock(&block_group->tree_lock); + + if (block_group->free_space < minlen) { + spin_unlock(&block_group->tree_lock); + break; + } + + entry = tree_search_offset(block_group, start, 0, 1); + if (!entry) + entry = tree_search_offset(block_group, + offset_to_bitmap(block_group, + start), + 1, 1); + + if (!entry || entry->offset >= end) { + spin_unlock(&block_group->tree_lock); + break; + } + + if (entry->bitmap) { + ret = search_bitmap(block_group, entry, &start, &bytes); + if (!ret) { + if (start >= end) { + spin_unlock(&block_group->tree_lock); + break; + } + bytes = min(bytes, end - start); + bitmap_clear_bits(block_group, entry, + start, bytes); + if (entry->bytes == 0) + free_bitmap(block_group, entry); + } else { + start = entry->offset + BITS_PER_BITMAP * + block_group->sectorsize; + spin_unlock(&block_group->tree_lock); + ret = 0; + continue; + } + } else { + start = entry->offset; + bytes = min(entry->bytes, end - start); + unlink_free_space(block_group, entry); + kfree(entry); + } + + spin_unlock(&block_group->tree_lock); + + if (bytes >= minlen) { + int update_ret; + update_ret = btrfs_update_reserved_bytes(block_group, + bytes, 1, 1); + + ret = btrfs_error_discard_extent(fs_info->extent_root, + start, + bytes, + &actually_trimmed); + + btrfs_add_free_space(block_group, + start, bytes); + if (!update_ret) + btrfs_update_reserved_bytes(block_group, + bytes, 0, 1); + + if (ret) + break; + *trimmed += actually_trimmed; + } + start += bytes; + bytes = 0; + + if (fatal_signal_pending(current)) { + ret = -ERESTARTSYS; + break; + } + + cond_resched(); + } + + return ret; +} diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h index e49ca5c321b5..65c3b935289f 100644 --- a/fs/btrfs/free-space-cache.h +++ b/fs/btrfs/free-space-cache.h @@ -68,4 +68,6 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, int btrfs_return_cluster_to_free_space( struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster); +int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, + u64 *trimmed, u64 start, u64 end, u64 minlen); #endif diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 32c980ae0f1c..649f47d2afb4 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -40,6 +40,7 @@ #include #include #include +#include #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -258,6 +259,49 @@ static int btrfs_ioctl_getversion(struct file *file, int __user *arg) return put_user(inode->i_generation, arg); } +static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) +{ + struct btrfs_root *root = fdentry(file)->d_sb->s_fs_info; + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_device *device; + struct request_queue *q; + struct fstrim_range range; + u64 minlen = ULLONG_MAX; + u64 num_devices = 0; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + mutex_lock(&fs_info->fs_devices->device_list_mutex); + list_for_each_entry(device, &fs_info->fs_devices->devices, dev_list) { + if (!device->bdev) + continue; + q = bdev_get_queue(device->bdev); + if (blk_queue_discard(q)) { + num_devices++; + minlen = min((u64)q->limits.discard_granularity, + minlen); + } + } + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + if (!num_devices) + return -EOPNOTSUPP; + + if (copy_from_user(&range, arg, sizeof(range))) + return -EFAULT; + + range.minlen = max(range.minlen, minlen); + ret = btrfs_trim_fs(root, &range); + if (ret < 0) + return ret; + + if (copy_to_user(arg, &range, sizeof(range))) + return -EFAULT; + + return 0; +} + static noinline int create_subvol(struct btrfs_root *root, struct dentry *dentry, char *name, int namelen, @@ -2426,6 +2470,8 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_setflags(file, argp); case FS_IOC_GETVERSION: return btrfs_ioctl_getversion(file, argp); + case FITRIM: + return btrfs_ioctl_fitrim(file, argp); case BTRFS_IOC_SNAP_CREATE: return btrfs_ioctl_snap_create(file, argp, 0); case BTRFS_IOC_SNAP_CREATE_V2: From c622ae6085d0c6ad834213bbf1477eb311359078 Mon Sep 17 00:00:00 2001 From: liubo Date: Sat, 26 Mar 2011 08:01:12 -0400 Subject: [PATCH 36/45] btrfs: make inode ref log recovery faster When we recover from crash via write-ahead log tree and process the inode refs, for each btrfs_inode_ref item, we will 1) check if we already have a perfect match in fs/file tree, if we have, then we're done. 2) search the corresponding back reference in fs/file tree, and check all the names in this back reference to see if they are also in the log to avoid conflict corners. 3) recover the logged inode refs to fs/file tree. In current btrfs, however, - for 2)'s check, once is enough, since the checked back reference will remain unchanged after processing all the inode refs belonged to the key. - it has no need to do another 1) between 2) and 3). I've made a small test to show how it improves, $dd if=/dev/zero of=foobar bs=4K count=1 $sync $make 100 hard links continuously, like ln foobar link_i $fsync foobar $echo b > /proc/sysrq-trigger after reboot $time mount DEV PATH without patch: real 0m0.285s user 0m0.001s sys 0m0.009s with patch: real 0m0.123s user 0m0.000s sys 0m0.010s Changelog v1->v2: - fix double free - pointed by David Sterba Changelog v2->v3: - adjust free order Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/tree-log.c | 37 ++++++++++++------------------------- 1 file changed, 12 insertions(+), 25 deletions(-) diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index f9425e33e358..c50271ad3157 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -799,12 +799,12 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans, struct inode *dir; int ret; struct btrfs_inode_ref *ref; - struct btrfs_dir_item *di; struct inode *inode; char *name; int namelen; unsigned long ref_ptr; unsigned long ref_end; + int search_done = 0; /* * it is possible that we didn't log all the parent directories @@ -845,7 +845,10 @@ again: * existing back reference, and we don't want to create * dangling pointers in the directory. */ -conflict_again: + + if (search_done) + goto insert; + ret = btrfs_search_slot(NULL, root, key, path, 0, 0); if (ret == 0) { char *victim_name; @@ -886,37 +889,21 @@ conflict_again: ret = btrfs_unlink_inode(trans, root, dir, inode, victim_name, victim_name_len); - kfree(victim_name); - btrfs_release_path(root, path); - goto conflict_again; } kfree(victim_name); ptr = (unsigned long)(victim_ref + 1) + victim_name_len; } BUG_ON(ret); + + /* + * NOTE: we have searched root tree and checked the + * coresponding ref, it does not need to check again. + */ + search_done = 1; } btrfs_release_path(root, path); - /* look for a conflicting sequence number */ - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, - btrfs_inode_ref_index(eb, ref), - name, namelen, 0); - if (di && !IS_ERR(di)) { - ret = drop_one_dir_item(trans, root, path, dir, di); - BUG_ON(ret); - } - btrfs_release_path(root, path); - - - /* look for a conflicting name */ - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, - name, namelen, 0); - if (di && !IS_ERR(di)) { - ret = drop_one_dir_item(trans, root, path, dir, di); - BUG_ON(ret); - } - btrfs_release_path(root, path); - +insert: /* insert our name */ ret = btrfs_add_link(trans, dir, inode, name, namelen, 0, btrfs_inode_ref_index(eb, ref)); From dac97e516c617f9c797f64b0224050b70aea30c7 Mon Sep 17 00:00:00 2001 From: Yoshinori Sano Date: Tue, 15 Feb 2011 12:01:42 +0000 Subject: [PATCH 37/45] Btrfs: fix uncheck memory allocations To make Btrfs code more robust, several return value checks where memory allocation can fail are introduced. I use BUG_ON where I don't know how to handle the error properly, which increases the number of using the notorious BUG_ON, though. Signed-off-by: Yoshinori Sano Signed-off-by: Chris Mason --- fs/btrfs/compression.c | 6 ++++++ fs/btrfs/extent-tree.c | 4 ++++ fs/btrfs/inode.c | 2 ++ 3 files changed, 12 insertions(+) diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 4d2110eafe29..992a4b92083e 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -340,6 +340,8 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, WARN_ON(start & ((u64)PAGE_CACHE_SIZE - 1)); cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS); + if (!cb) + return -ENOMEM; atomic_set(&cb->pending_bios, 0); cb->errors = 0; cb->inode = inode; @@ -354,6 +356,10 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS); + if(!bio) { + kfree(cb); + return -ENOMEM; + } bio->bi_private = cb; bio->bi_end_io = end_compressed_bio_write; atomic_inc(&cb->pending_bios); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 1efeda3b2f6f..cd0b69f57375 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6978,6 +6978,10 @@ static noinline int get_new_locations(struct inode *reloc_inode, struct disk_extent *old = exts; max *= 2; exts = kzalloc(sizeof(*exts) * max, GFP_NOFS); + if (!exts) { + ret = -ENOMEM; + goto out; + } memcpy(exts, old, sizeof(*exts) * nr); if (old != *extents) kfree(old); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 67fd6e9552d3..f739b256967e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -290,6 +290,7 @@ static noinline int add_async_extent(struct async_cow *cow, struct async_extent *async_extent; async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS); + BUG_ON(!async_extent); async_extent->start = start; async_extent->ram_size = ram_size; async_extent->compressed_size = compressed_size; @@ -388,6 +389,7 @@ again: (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); + BUG_ON(!pages); if (BTRFS_I(inode)->force_compress) compress_type = BTRFS_I(inode)->force_compress; From 2d4e6f6ad2b9f84f568d07dae4bdbc7f48a70ad1 Mon Sep 17 00:00:00 2001 From: liubo Date: Thu, 24 Feb 2011 09:38:16 +0000 Subject: [PATCH 38/45] Btrfs: fix return value of setflags ioctl setflags ioctl should return error when any checks fail. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/ioctl.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 649f47d2afb4..6b70e0e2bd1e 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -247,9 +247,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) btrfs_end_transaction(trans, root); mnt_drop_write(file->f_path.mnt); + + ret = 0; out_unlock: mutex_unlock(&inode->i_mutex); - return 0; + return ret; } static int btrfs_ioctl_getversion(struct file *file, int __user *arg) From 9f7c43c96727a53bea45f7f2549d897f0a6117b8 Mon Sep 17 00:00:00 2001 From: liubo Date: Mon, 7 Mar 2011 02:13:33 +0000 Subject: [PATCH 39/45] Btrfs: fix memory leak of empty filesystem after balance After Josef's patch(commit 3c14874acc71180553fb5aba528e3cf57c5b958b), btrfs will exclude super bytes when reading block groups(by marking a extent state UPTODATE). However, these bytes do not get freed while balance remove unused block groups, and we won't process those removed ones any more, when we do umount and unload the btrfs module, btrfs hits a memory leak. This patch add the missing free operation. Reproduce steps: $ mkfs.btrfs disk $ mount disk /mnt/btrfs -o loop $ btrfs filesystem balance /mnt/btrfs $ umount /mnt/btrfs $ rmmod btrfs Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index cd0b69f57375..a561060f5ffb 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -8669,6 +8669,12 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, BUG_ON(!block_group); BUG_ON(!block_group->ro); + /* + * Free the reserved super bytes from this block group before + * remove it. + */ + free_excluded_extents(root, block_group); + memcpy(&key, &block_group->key, sizeof(key)); if (block_group->flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | From c59021f846881a957ac5afe456d0f59d6a517b61 Mon Sep 17 00:00:00 2001 From: liubo Date: Mon, 7 Mar 2011 02:13:14 +0000 Subject: [PATCH 40/45] Btrfs: fix OOPS of empty filesystem after balance btrfs will remove unused block groups after balance. When a empty filesystem is balanced, the block group with tag "DATA" may be dropped, and after umount and mount again, it will not find "DATA" space_info and lead to OOPS. So we initial the necessary space_infos(DATA, SYSTEM, METADATA) to avoid OOPS. Reported-by: Daniel J Blueman Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 1 + fs/btrfs/disk-io.c | 6 ++++++ fs/btrfs/extent-tree.c | 23 +++++++++++++++++++++++ 3 files changed, 30 insertions(+) diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 93a0191aded6..d47ce8307854 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2234,6 +2234,7 @@ int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 type); int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range); +int btrfs_init_space_info(struct btrfs_fs_info *fs_info); /* ctree.c */ int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key, int level, int *slot); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 8ecc5419d8b6..b3fc8475870f 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2065,6 +2065,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->metadata_alloc_profile = (u64)-1; fs_info->system_alloc_profile = fs_info->metadata_alloc_profile; + ret = btrfs_init_space_info(fs_info); + if (ret) { + printk(KERN_ERR "Failed to initial space info: %d\n", ret); + goto fail_block_groups; + } + ret = btrfs_read_block_groups(extent_root); if (ret) { printk(KERN_ERR "Failed to read block groups: %d\n", ret); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a561060f5ffb..f619c3cb13b7 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -8778,6 +8778,29 @@ out: return ret; } +int btrfs_init_space_info(struct btrfs_fs_info *fs_info) +{ + struct btrfs_space_info *space_info; + int ret; + + ret = update_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM, 0, 0, + &space_info); + if (ret) + return ret; + + ret = update_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA, 0, 0, + &space_info); + if (ret) + return ret; + + ret = update_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA, 0, 0, + &space_info); + if (ret) + return ret; + + return ret; +} + int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) { return unpin_extent_range(root, start, end); From c2db1073fdf9757e6fd8b4a59d15b6ecc7a2af8a Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Tue, 1 Mar 2011 06:48:31 +0000 Subject: [PATCH 41/45] Btrfs: check return value of btrfs_alloc_path() Adding the check on the return value of btrfs_alloc_path() to several places. And, some of callers are modified by this change. Signed-off-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/compression.c | 11 +++++++---- fs/btrfs/dir-item.c | 10 ++++++---- fs/btrfs/file-item.c | 2 ++ fs/btrfs/inode.c | 18 ++++++++++++------ 4 files changed, 27 insertions(+), 14 deletions(-) diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 992a4b92083e..41d1d7c70e29 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -663,8 +663,9 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, atomic_inc(&cb->pending_bios); if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) { - btrfs_lookup_bio_sums(root, inode, comp_bio, - sums); + ret = btrfs_lookup_bio_sums(root, inode, + comp_bio, sums); + BUG_ON(ret); } sums += (comp_bio->bi_size + root->sectorsize - 1) / root->sectorsize; @@ -689,8 +690,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0); BUG_ON(ret); - if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) - btrfs_lookup_bio_sums(root, inode, comp_bio, sums); + if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) { + ret = btrfs_lookup_bio_sums(root, inode, comp_bio, sums); + BUG_ON(ret); + } ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0); BUG_ON(ret); diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c index 02c97ad61b6d..c62f02f6ae69 100644 --- a/fs/btrfs/dir-item.c +++ b/fs/btrfs/dir-item.c @@ -151,7 +151,7 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root ret = PTR_ERR(dir_item); if (ret == -EEXIST) goto second_insert; - goto out; + goto out_free; } leaf = path->nodes[0]; @@ -170,7 +170,7 @@ second_insert: /* FIXME, use some real flag for selecting the extra index */ if (root == root->fs_info->tree_root) { ret = 0; - goto out; + goto out_free; } btrfs_release_path(root, path); @@ -180,7 +180,7 @@ second_insert: name, name_len); if (IS_ERR(dir_item)) { ret2 = PTR_ERR(dir_item); - goto out; + goto out_free; } leaf = path->nodes[0]; btrfs_cpu_key_to_disk(&disk_key, location); @@ -192,7 +192,9 @@ second_insert: name_ptr = (unsigned long)(dir_item + 1); write_extent_buffer(leaf, name, name_ptr, name_len); btrfs_mark_buffer_dirty(leaf); -out: + +out_free: + btrfs_free_path(path); if (ret) return ret; diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index a2134195a85e..a6a9d4e8b491 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -170,6 +170,8 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; if (bio->bi_size > PAGE_CACHE_SIZE * 8) path->reada = 2; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index f739b256967e..04e9fffee8cc 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1467,8 +1467,11 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, if (bio_flags & EXTENT_BIO_COMPRESSED) { return btrfs_submit_compressed_read(inode, bio, mirror_num, bio_flags); - } else if (!skip_sum) - btrfs_lookup_bio_sums(root, inode, bio, NULL); + } else if (!skip_sum) { + ret = btrfs_lookup_bio_sums(root, inode, bio, NULL); + if (ret) + return ret; + } goto mapit; } else if (!skip_sum) { /* csum items have already been cloned */ @@ -1903,10 +1906,10 @@ static int btrfs_io_failed_hook(struct bio *failed_bio, else rw = READ; - BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, + ret = BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, failrec->last_mirror, failrec->bio_flags, 0); - return 0; + return ret; } /* @@ -5943,9 +5946,12 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, __btrfs_submit_bio_start_direct_io, __btrfs_submit_bio_done); goto err; - } else if (!skip_sum) - btrfs_lookup_bio_sums_dio(root, inode, bio, + } else if (!skip_sum) { + ret = btrfs_lookup_bio_sums_dio(root, inode, bio, file_offset, csums); + if (ret) + goto err; + } ret = btrfs_map_bio(root, rw, bio, 0, 1); err: From 92986796d84ef939e304099dece32572a755b280 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 4 Mar 2011 17:14:37 +0000 Subject: [PATCH 42/45] btrfs: don't mess with i_nlink of unlocked inode in rename() old_inode is not locked; it's not safe to play with its link count. Instead of bumping it and calling btrfs_unlink_inode(), add a variant of the latter that does not do btrfs_drop_nlink()/ btrfs_update_inode(), call it instead of btrfs_inc_nlink()/ btrfs_unlink_inode() and do btrfs_update_inode() ourselves. Signed-off-by: Al Viro Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 36 +++++++++++++++++++++++++----------- 1 file changed, 25 insertions(+), 11 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 04e9fffee8cc..4822b3132784 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2664,10 +2664,10 @@ failed: * recovery code. It remove a link in a directory with a given name, and * also drops the back refs in the inode to the directory */ -int btrfs_unlink_inode(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct inode *dir, struct inode *inode, - const char *name, int name_len) +static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, struct inode *inode, + const char *name, int name_len) { struct btrfs_path *path; int ret = 0; @@ -2739,12 +2739,25 @@ err: btrfs_i_size_write(dir, dir->i_size - name_len * 2); inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME; btrfs_update_inode(trans, root, dir); - btrfs_drop_nlink(inode); - ret = btrfs_update_inode(trans, root, inode); out: return ret; } +int btrfs_unlink_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, struct inode *inode, + const char *name, int name_len) +{ + int ret; + ret = __btrfs_unlink_inode(trans, root, dir, inode, name, name_len); + if (!ret) { + btrfs_drop_nlink(inode); + ret = btrfs_update_inode(trans, root, inode); + } + return ret; +} + + /* helper to check if there is any shared block in the path */ static int check_path_shared(struct btrfs_root *root, struct btrfs_path *path) @@ -6999,11 +7012,12 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, old_dentry->d_name.name, old_dentry->d_name.len); } else { - btrfs_inc_nlink(old_dentry->d_inode); - ret = btrfs_unlink_inode(trans, root, old_dir, - old_dentry->d_inode, - old_dentry->d_name.name, - old_dentry->d_name.len); + ret = __btrfs_unlink_inode(trans, root, old_dir, + old_dentry->d_inode, + old_dentry->d_name.name, + old_dentry->d_name.len); + if (!ret) + ret = btrfs_update_inode(trans, root, old_inode); } BUG_ON(ret); From c055e99eea6e4f614267632fac546e7896c0227b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 4 Mar 2011 17:15:18 +0000 Subject: [PATCH 43/45] btrfs: check link counter overflow in link(2) Signed-off-by: Al Viro Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4822b3132784..04babaf31a33 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4837,6 +4837,9 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, if (root->objectid != BTRFS_I(inode)->root->objectid) return -EXDEV; + if (inode->i_nlink == ~0U) + return -EMLINK; + btrfs_inc_nlink(inode); inode->i_ctime = CURRENT_TIME; From 1561deda687eef0e95065f1268d680ddc5976ee7 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Sun, 27 Mar 2011 16:07:36 +0800 Subject: [PATCH 44/45] btrfs: fix possible deadlock by clearing __GFP_FS flag Using the GFP_HIGHUSER_MOVABLE flag to allocate the metadata's page may cause deadlock. Task1 open() ... btrfs_search_slot() ... btrfs_cow_block() ... alloc_page() wait for reclaiming shrink_slab() ... shrink_icache_memory() ... btrfs_evict_inode() ... btrfs_search_slot() If the path is locked by task1, the deadlock happens. So the btree's page cache is different with the file's page cache, it can not allocate pages by GFP_HIGHUSER_MOVABLE flag, we must clear __GFP_FS flag in GFP_HIGHUSER_MOVABLE flag. Reported-by: Itaru Kitayama Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 2 ++ fs/btrfs/inode.c | 3 ++- 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index b3fc8475870f..5cf3aa7b125c 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1724,6 +1724,8 @@ struct btrfs_root *open_ctree(struct super_block *sb, goto fail_bdi; } + fs_info->btree_inode->i_mapping->flags &= ~__GFP_FS; + INIT_RADIX_TREE(&fs_info->fs_roots_radix, GFP_ATOMIC); INIT_LIST_HEAD(&fs_info->trans_list); INIT_LIST_HEAD(&fs_info->dead_roots); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 04babaf31a33..06274186b290 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2536,6 +2536,8 @@ static void btrfs_read_locked_inode(struct inode *inode) BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item); alloc_group_block = btrfs_inode_block_group(leaf, inode_item); + if (location.objectid == BTRFS_FREE_SPACE_OBJECTID) + inode->i_mapping->flags &= ~__GFP_FS; /* * try to precache a NULL acl entry for files that don't have @@ -4084,7 +4086,6 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, BTRFS_I(inode)->root = root; memcpy(&BTRFS_I(inode)->location, location, sizeof(*location)); btrfs_read_locked_inode(inode); - inode_tree_add(inode); unlock_new_inode(inode); if (new) From d9d04879321af570ea7285c6dad92d9c3cd108a1 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Sun, 27 Mar 2011 21:23:21 -0400 Subject: [PATCH 45/45] Btrfs: fix __btrfs_map_block on 32 bit machines Recent changes for discard support didn't compile, this fixes them not to try and % 64 bit numbers. Signed-off-by: Chris Mason --- fs/btrfs/volumes.c | 26 ++++++++++++++++++++------ 1 file changed, 20 insertions(+), 6 deletions(-) diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index c440c89a470a..8b9fb8c7683d 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3126,13 +3126,19 @@ again: if (map->type & BTRFS_BLOCK_GROUP_RAID0) { u64 stripes; - int last_stripe = (stripe_nr_end - 1) % - map->num_stripes; + u32 last_stripe = 0; int j; + div_u64_rem(stripe_nr_end - 1, + map->num_stripes, + &last_stripe); + for (j = 0; j < map->num_stripes; j++) { - if ((stripe_nr_end - 1 - j) % - map->num_stripes == stripe_index) + u32 test; + + div_u64_rem(stripe_nr_end - 1 - j, + map->num_stripes, &test); + if (test == stripe_index) break; } stripes = stripe_nr_end - 1 - j; @@ -3153,11 +3159,19 @@ again: int j; int factor = map->num_stripes / map->sub_stripes; - int last_stripe = (stripe_nr_end - 1) % factor; + u32 last_stripe = 0; + + div_u64_rem(stripe_nr_end - 1, + factor, &last_stripe); last_stripe *= map->sub_stripes; for (j = 0; j < factor; j++) { - if ((stripe_nr_end - 1 - j) % factor == + u32 test; + + div_u64_rem(stripe_nr_end - 1 - j, + factor, &test); + + if (test == stripe_index / map->sub_stripes) break; }