From 8cef4e160d74920ad1725f58c89fd75ec4c4ac38 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 12 Nov 2009 09:33:26 +0000 Subject: Btrfs: Avoid superfluous tree-log writeout We allow two log transactions at a time, but use same flag to mark dirty tree-log btree blocks. So we may flush dirty blocks belonging to newer log transaction when committing a log transaction. This patch fixes the issue by using two flags to mark dirty tree-log btree blocks. Signed-off-by: Yan Zheng Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 94627c4cc193..4a86508ce473 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4919,8 +4919,16 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans, btrfs_set_buffer_uptodate(buf); if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) { - set_extent_dirty(&root->dirty_log_pages, buf->start, - buf->start + buf->len - 1, GFP_NOFS); + /* + * we allow two log transactions at a time, use different + * EXENT bit to differentiate dirty pages. + */ + if (root->log_transid % 2 == 0) + set_extent_dirty(&root->dirty_log_pages, buf->start, + buf->start + buf->len - 1, GFP_NOFS); + else + set_extent_new(&root->dirty_log_pages, buf->start, + buf->start + buf->len - 1, GFP_NOFS); } else { set_extent_dirty(&trans->transaction->dirty_pages, buf->start, buf->start + buf->len - 1, GFP_NOFS); -- cgit v1.2.2 From 24bbcf0442ee04660a5a030efdbb6d03f1c275cb Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 12 Nov 2009 09:36:34 +0000 Subject: Btrfs: Add delayed iput iput() can trigger new transactions if we are dropping the final reference, so calling it in btrfs_commit_transaction may end up deadlock. This patch adds delayed iput to avoid the issue. Signed-off-by: Yan Zheng Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 4a86508ce473..fcdccfa46004 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2880,9 +2880,9 @@ static noinline void flush_delalloc_async(struct btrfs_work *work) root = async->root; info = async->info; - btrfs_start_delalloc_inodes(root); + btrfs_start_delalloc_inodes(root, 0); wake_up(&info->flush_wait); - btrfs_wait_ordered_extents(root, 0); + btrfs_wait_ordered_extents(root, 0, 0); spin_lock(&info->lock); info->flushing = 0; @@ -2956,8 +2956,8 @@ static void flush_delalloc(struct btrfs_root *root, return; flush: - btrfs_start_delalloc_inodes(root); - btrfs_wait_ordered_extents(root, 0); + btrfs_start_delalloc_inodes(root, 0); + btrfs_wait_ordered_extents(root, 0, 0); spin_lock(&info->lock); info->flushing = 0; -- cgit v1.2.2 From 86b9f2eca5e0984145e3c7698a7cd6dd65c2a93f Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 12 Nov 2009 09:36:50 +0000 Subject: Btrfs: Fix per root used space accounting The bytes_used field in root item was originally planned to trace the amount of used data and tree blocks. But it never worked right since we can't trace freeing of data accurately. This patch changes it to only trace the amount of tree blocks. Signed-off-by: Yan Zheng Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 31 +++++++++++++++++++++++-------- 1 file changed, 23 insertions(+), 8 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index fcdccfa46004..44d7a322ec28 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3454,14 +3454,6 @@ static int update_block_group(struct btrfs_trans_handle *trans, else old_val -= num_bytes; btrfs_set_super_bytes_used(&info->super_copy, old_val); - - /* block accounting for root item */ - old_val = btrfs_root_used(&root->root_item); - if (alloc) - old_val += num_bytes; - else - old_val -= num_bytes; - btrfs_set_root_used(&root->root_item, old_val); spin_unlock(&info->delalloc_lock); while (total) { @@ -4049,6 +4041,21 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, return ret; } +int btrfs_free_tree_block(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + u64 bytenr, u32 blocksize, + u64 parent, u64 root_objectid, int level) +{ + u64 used; + spin_lock(&root->node_lock); + used = btrfs_root_used(&root->root_item) - blocksize; + btrfs_set_root_used(&root->root_item, used); + spin_unlock(&root->node_lock); + + return btrfs_free_extent(trans, root, bytenr, blocksize, + parent, root_objectid, level, 0); +} + static u64 stripe_align(struct btrfs_root *root, u64 val) { u64 mask = ((u64)root->stripesize - 1); @@ -4897,6 +4904,14 @@ static int alloc_tree_block(struct btrfs_trans_handle *trans, extent_op); BUG_ON(ret); } + + if (root_objectid == root->root_key.objectid) { + u64 used; + spin_lock(&root->node_lock); + used = btrfs_root_used(&root->root_item) + num_bytes; + btrfs_set_root_used(&root->root_item, used); + spin_unlock(&root->node_lock); + } return ret; } -- cgit v1.2.2 From 06b2331f8333ec6edf41662757ce8882cc1747d5 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 26 Nov 2009 09:31:11 +0000 Subject: Btrfs: don't add extent 0 to the free space cache v2 If block group 0 is completely free, btrfs_read_block_groups will add extent [0, BTRFS_SUPER_INFO_OFFSET) to the free space cache. Signed-off-by: Yan Zheng Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 44d7a322ec28..39761ca07f15 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -195,6 +195,14 @@ static int exclude_super_stripes(struct btrfs_root *root, int stripe_len; int i, nr, ret; + if (cache->key.objectid < BTRFS_SUPER_INFO_OFFSET) { + stripe_len = BTRFS_SUPER_INFO_OFFSET - cache->key.objectid; + cache->bytes_super += stripe_len; + ret = add_excluded_extent(root, cache->key.objectid, + stripe_len); + BUG_ON(ret); + } + for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { bytenr = btrfs_sb_offset(i); ret = btrfs_rmap_block(&root->fs_info->mapping_tree, @@ -255,7 +263,7 @@ static u64 add_new_free_space(struct btrfs_block_group_cache *block_group, if (ret) break; - if (extent_start == start) { + if (extent_start <= start) { start = extent_end + 1; } else if (extent_start > start && extent_start < end) { size = extent_start - start; -- cgit v1.2.2 From 83d3c9696fed237a3d96fce18299e2fcf112109f Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 7 Dec 2009 21:45:59 +0000 Subject: Btrfs: make metadata chunks smaller This patch makes us a bit less zealous about making sure we have enough free metadata space by pearing down the size of new metadata chunks to 256mb instead of 1gb. Also, we used to try an allocate metadata chunks when allocating data, but that sort of thing is done elsewhere now so we can just remove it. With my -ENOSPC test I used to have 3gb reserved for metadata out of 75gb, now I have 1.7gb. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 11 +---------- 1 file changed, 1 insertion(+), 10 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 39761ca07f15..56e50137d0e6 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4593,7 +4593,6 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans, { int ret; u64 search_start = 0; - struct btrfs_fs_info *info = root->fs_info; data = btrfs_get_alloc_profile(root, data); again: @@ -4601,17 +4600,9 @@ again: * the only place that sets empty_size is btrfs_realloc_node, which * is not called recursively on allocations */ - if (empty_size || root->ref_cows) { - if (!(data & BTRFS_BLOCK_GROUP_METADATA)) { - ret = do_chunk_alloc(trans, root->fs_info->extent_root, - 2 * 1024 * 1024, - BTRFS_BLOCK_GROUP_METADATA | - (info->metadata_alloc_profile & - info->avail_metadata_alloc_bits), 0); - } + if (empty_size || root->ref_cows) ret = do_chunk_alloc(trans, root->fs_info->extent_root, num_bytes + 2 * 1024 * 1024, data, 0); - } WARN_ON(num_bytes < root->sectorsize); ret = find_free_extent(trans, root, num_bytes, empty_size, -- cgit v1.2.2 From 11dfe35a0108097f2df1f042c485fa7f758c2cdf Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 13 Nov 2009 20:12:59 +0000 Subject: Btrfs: fix possible panic on unmount We can race with the unmount of an fs and the stopping of a kthread where we will free the block group before we're done using it. The reason for this is because we do not hold a reference on the block group while its caching, since the allocator drops its reference once it exits or moves on to the next block group. This patch fixes the problem by taking a reference to the block group before we start caching and dropping it when we're done to make sure all accesses to the block group are safe. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 32 +++++++++++++++++++------------- 1 file changed, 19 insertions(+), 13 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 56e50137d0e6..432a2da4641e 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -83,6 +83,17 @@ static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits) return (cache->flags & bits) == bits; } +void btrfs_get_block_group(struct btrfs_block_group_cache *cache) +{ + atomic_inc(&cache->count); +} + +void btrfs_put_block_group(struct btrfs_block_group_cache *cache) +{ + if (atomic_dec_and_test(&cache->count)) + kfree(cache); +} + /* * this adds the block group to the fs_info rb tree for the block group * cache @@ -156,7 +167,7 @@ block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr, } } if (ret) - atomic_inc(&ret->count); + btrfs_get_block_group(ret); spin_unlock(&info->block_group_cache_lock); return ret; @@ -407,6 +418,8 @@ err: put_caching_control(caching_ctl); atomic_dec(&block_group->space_info->caching_threads); + btrfs_put_block_group(block_group); + return 0; } @@ -447,6 +460,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache) up_write(&fs_info->extent_commit_sem); atomic_inc(&cache->space_info->caching_threads); + btrfs_get_block_group(cache); tsk = kthread_run(caching_kthread, cache, "btrfs-cache-%llu\n", cache->key.objectid); @@ -486,12 +500,6 @@ struct btrfs_block_group_cache *btrfs_lookup_block_group( return cache; } -void btrfs_put_block_group(struct btrfs_block_group_cache *cache) -{ - if (atomic_dec_and_test(&cache->count)) - kfree(cache); -} - static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info, u64 flags) { @@ -2582,7 +2590,7 @@ next_block_group(struct btrfs_root *root, if (node) { cache = rb_entry(node, struct btrfs_block_group_cache, cache_node); - atomic_inc(&cache->count); + btrfs_get_block_group(cache); } else cache = NULL; spin_unlock(&root->fs_info->block_group_cache_lock); @@ -4227,7 +4235,7 @@ search: u64 offset; int cached; - atomic_inc(&block_group->count); + btrfs_get_block_group(block_group); search_start = block_group->key.objectid; have_block_group: @@ -4315,7 +4323,7 @@ have_block_group: btrfs_put_block_group(block_group); block_group = last_ptr->block_group; - atomic_inc(&block_group->count); + btrfs_get_block_group(block_group); spin_unlock(&last_ptr->lock); spin_unlock(&last_ptr->refill_lock); @@ -7395,9 +7403,7 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) wait_block_group_cache_done(block_group); btrfs_remove_free_space_cache(block_group); - - WARN_ON(atomic_read(&block_group->count) != 1); - kfree(block_group); + btrfs_put_block_group(block_group); spin_lock(&info->block_group_cache_lock); } -- cgit v1.2.2 From 7a7965f83e89f0be506a96769938a721e4e5ae50 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 1 Feb 2010 02:41:17 +0000 Subject: Btrfs: Fix oopsen when dropping empty tree. When dropping a empty tree, walk_down_tree() skips checking extent information for the tree root. This will triggers a BUG_ON in walk_up_proc(). Signed-off-by: Yan Zheng Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 432a2da4641e..559f72489b3b 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5402,10 +5402,6 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans, int ret; while (level >= 0) { - if (path->slots[level] >= - btrfs_header_nritems(path->nodes[level])) - break; - ret = walk_down_proc(trans, root, path, wc, lookup_info); if (ret > 0) break; @@ -5413,6 +5409,10 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans, if (level == 0) break; + if (path->slots[level] >= + btrfs_header_nritems(path->nodes[level])) + break; + ret = do_walk_down(trans, root, path, wc, &lookup_info); if (ret > 0) { path->slots[level]++; -- cgit v1.2.2 From 2ac55d41b5d6bf49e76bc85db5431240617e2f8f Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 3 Feb 2010 19:33:23 +0000 Subject: Btrfs: cache the extent state everywhere we possibly can V2 This patch just goes through and fixes everybody that does lock_extent() blah unlock_extent() to use lock_extent_bits() blah unlock_extent_cached() and pass around a extent_state so we only have to do the searches once per function. This gives me about a 3 mb/s boots on my random write test. I have not converted some things, like the relocation and ioctl's, since they aren't heavily used and the relocation stuff is in the middle of being re-written. I also changed the clear_extent_bit() to only unset the cached state if we are clearing EXTENT_LOCKED and related stuff, so we can do things like this lock_extent_bits() clear delalloc bits unlock_extent_cached() without losing our cached state. I tested this thoroughly and turned on LEAK_DEBUG to make sure we weren't leaking extent states, everything worked out fine. Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 559f72489b3b..1727b26fb194 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6561,6 +6561,7 @@ static noinline int invalidate_extent_cache(struct btrfs_root *root, struct btrfs_key key; struct inode *inode = NULL; struct btrfs_file_extent_item *fi; + struct extent_state *cached_state = NULL; u64 num_bytes; u64 skip_objectid = 0; u32 nritems; @@ -6589,12 +6590,14 @@ static noinline int invalidate_extent_cache(struct btrfs_root *root, } num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - lock_extent(&BTRFS_I(inode)->io_tree, key.offset, - key.offset + num_bytes - 1, GFP_NOFS); + lock_extent_bits(&BTRFS_I(inode)->io_tree, key.offset, + key.offset + num_bytes - 1, 0, &cached_state, + GFP_NOFS); btrfs_drop_extent_cache(inode, key.offset, key.offset + num_bytes - 1, 1); - unlock_extent(&BTRFS_I(inode)->io_tree, key.offset, - key.offset + num_bytes - 1, GFP_NOFS); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, key.offset, + key.offset + num_bytes - 1, &cached_state, + GFP_NOFS); cond_resched(); } iput(inode); -- cgit v1.2.2 From 5a0e3ad6af8660be21ca98a971cd00f331318c05 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 24 Mar 2010 17:04:11 +0900 Subject: include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo Guess-its-ok-by: Christoph Lameter Cc: Ingo Molnar Cc: Lee Schermerhorn --- fs/btrfs/extent-tree.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 1727b26fb194..fa2907d531a2 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -22,6 +22,7 @@ #include #include #include +#include #include "compat.h" #include "hash.h" #include "ctree.h" -- cgit v1.2.2 From 90d2c51dbb4db05c040bc7db264bb7ab35e35455 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Thu, 25 Mar 2010 12:37:12 +0000 Subject: Btrfs: add NULL check for do_walk_down() btrfs_find_create_tree_block() may return NULL, so we must check the returned value, or we will access a NULL pointer. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 1727b26fb194..503a18eaef52 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5205,6 +5205,8 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, next = btrfs_find_tree_block(root, bytenr, blocksize); if (!next) { next = btrfs_find_create_tree_block(root, bytenr, blocksize); + if (!next) + return -ENOMEM; reada = 1; } btrfs_tree_lock(next); @@ -5417,7 +5419,8 @@ static noinline int walk_down_tree(struct btrfs_trans_handle *trans, if (ret > 0) { path->slots[level]++; continue; - } + } else if (ret < 0) + return ret; level = wc->level; } return 0; -- cgit v1.2.2 From 1b1d1f6625e517a08640ddb4b8f8a0e025243fe3 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 19 Mar 2010 20:49:55 +0000 Subject: Btrfs: fail to mount if we have problems reading the block groups We don't actually check the return value of btrfs_read_block_groups, so we can possibly succeed to mount, but then fail to say read the superblock xattr for selinux which will cause the vfs code to deactivate the super. This is a problem because in find_free_extent we just assume that we will find the right space_info for the allocation we want. But if we failed to read the block groups, we won't have setup any space_info's, and we'll hit a NULL pointer deref in find_free_extent. This patch fixes that problem by checking the return value of btrfs_read_block_groups, and failing out properly. I've also added a check in find_free_extent so if for some reason we don't find an appropriate space_info, we just return -ENOSPC. Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 503a18eaef52..4c910359c807 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4170,6 +4170,10 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, ins->offset = 0; space_info = __find_space_info(root->fs_info, data); + if (!space_info) { + printk(KERN_ERR "No space info for %d\n", data); + return -ENOSPC; + } if (orig_root->ref_cows || empty_size) allowed_chunk_alloc = 1; @@ -7372,7 +7376,6 @@ static int find_first_block_group(struct btrfs_root *root, } path->slots[0]++; } - ret = -ENOENT; out: return ret; } -- cgit v1.2.2 From 287a0ab91d25ca982f895a76402e5893b47ed7a6 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 19 Mar 2010 18:07:23 +0000 Subject: Btrfs: kill max_extent mount option As Yan pointed out, theres not much reason for all this complicated math to account for file extents being split up into max_extent chunks, since they are likely to all end up in the same leaf anyway. Since there isn't much reason to use max_extent, just remove the option altogether so we have one less thing we need to test. Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 4c910359c807..5f6bc283c0c8 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2846,7 +2846,7 @@ int btrfs_unreserve_metadata_for_delalloc(struct btrfs_root *root, } spin_unlock(&BTRFS_I(inode)->accounting_lock); - BTRFS_I(inode)->reserved_extents--; + BTRFS_I(inode)->reserved_extents -= num_items; BUG_ON(BTRFS_I(inode)->reserved_extents < 0); if (meta_sinfo->bytes_delalloc < num_bytes) { @@ -3111,7 +3111,7 @@ again: return -ENOSPC; } - BTRFS_I(inode)->reserved_extents++; + BTRFS_I(inode)->reserved_extents += num_items; check_force_delalloc(meta_sinfo); spin_unlock(&meta_sinfo->lock); -- cgit v1.2.2 From b5cb160084fad438c513d0952849e597ffe9e3d9 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 12 Mar 2010 19:28:18 +0000 Subject: Btrfs: fix small race with delalloc flushing waitqueue's Everytime we start a new flushing thread, we init the waitqueue if there isn't a flushing thread running. The problem with this is we check space_info->flushing, which we clear right before doing a wake_up on the flushing waitqueue, which causes problems if we init the waitqueue in the middle of clearing the flushing flagh and calling wake_up. This is hard to hit, but the code is wrong anyway, so init the flushing/allocating waitqueue when creating the space info and let it be. I haven't seen the panic since I've been using this patch. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5f6bc283c0c8..101041d4d2b2 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2676,6 +2676,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, INIT_LIST_HEAD(&found->block_groups); init_rwsem(&found->groups_sem); + init_waitqueue_head(&found->flush_wait); + init_waitqueue_head(&found->allocate_wait); spin_lock_init(&found->lock); found->flags = flags; found->total_bytes = total_bytes; @@ -2944,12 +2946,10 @@ static void flush_delalloc(struct btrfs_root *root, spin_lock(&info->lock); - if (!info->flushing) { + if (!info->flushing) info->flushing = 1; - init_waitqueue_head(&info->flush_wait); - } else { + else wait = true; - } spin_unlock(&info->lock); @@ -3011,7 +3011,6 @@ static int maybe_allocate_chunk(struct btrfs_root *root, if (!info->allocating_chunk) { info->force_alloc = 1; info->allocating_chunk = 1; - init_waitqueue_head(&info->allocate_wait); } else { wait = true; } -- cgit v1.2.2 From ab6e24103cbd215e922938a4f58c75194761a60e Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 19 Mar 2010 14:38:13 +0000 Subject: Btrfs: fix data enospc check overflow Because we account for reserved space we get from the allocator before we actually account for allocating delalloc space, we can have a small window where the amount of "used" space in a space_info is more than the total amount of space in the space_info. This will cause a overflow in our check, so it will seem like we have _tons_ of free space, and we'll allow reservations to occur that will end up larger than the amount of space we have. I've seen users report ENOSPC panic's in cow_file_range a few times recently, so I tried to reproduce this problem and found I could reproduce it if I ran one of my tests in a loop for like 20 minutes. With this patch my test ran all night without issues. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/extent-tree.c') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 101041d4d2b2..786899da3eb9 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3234,7 +3234,8 @@ int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode, u64 bytes) { struct btrfs_space_info *data_sinfo; - int ret = 0, committed = 0; + u64 used; + int ret = 0, committed = 0, flushed = 0; /* make sure bytes are sectorsize aligned */ bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1); @@ -3246,12 +3247,21 @@ int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode, again: /* make sure we have enough space to handle the data first */ spin_lock(&data_sinfo->lock); - if (data_sinfo->total_bytes - data_sinfo->bytes_used - - data_sinfo->bytes_delalloc - data_sinfo->bytes_reserved - - data_sinfo->bytes_pinned - data_sinfo->bytes_readonly - - data_sinfo->bytes_may_use - data_sinfo->bytes_super < bytes) { + used = data_sinfo->bytes_used + data_sinfo->bytes_delalloc + + data_sinfo->bytes_reserved + data_sinfo->bytes_pinned + + data_sinfo->bytes_readonly + data_sinfo->bytes_may_use + + data_sinfo->bytes_super; + + if (used + bytes > data_sinfo->total_bytes) { struct btrfs_trans_handle *trans; + if (!flushed) { + spin_unlock(&data_sinfo->lock); + flush_delalloc(root, data_sinfo); + flushed = 1; + goto again; + } + /* * if we don't have enough free bytes in this space then we need * to alloc a new chunk. -- cgit v1.2.2