- 14 Jun, 2013 16 commits
-
-
Miao Xie authored
Cc: David Sterba <dsterba@suse.cz> Signed-off-by:
Miao Xie <miaox@cn.fujitsu.com> Signed-off-by:
Josef Bacik <jbacik@fusionio.com>
-
Miao Xie authored
In order to avoid the R/O remount, we acquired ->s_umount lock during we deleted the dead snapshots and subvolumes. But it is unnecessary, because we have cleaner_mutex. We use cleaner_mutex to protect the process of the dead snapshots/subvolumes deletion. And when we remount the fs to be R/O, we also acquire this mutex to do cleanup after we change the status of the fs. That is this lock can serialize the above operations, the cleaner can be aware of the status of the fs, and if the cleaner is deleting the dead snapshots/subvolumes, the remount task will wait for it. So it is safe to remove ->s_umount in cleaner_kthread(). Cc: David Sterba <dsterba@suse.cz> Signed-off-by:
-
Stefan Behrens authored
btrfs_read_fs_root_no_name() already checks if btrfs_root_refs() is zero and returns ENOENT in this case. There is no need to do it again in six places. Signed-off-by:
Stefan Behrens <sbehrens@giantdisaster.de> Signed-off-by:
-
Stefan Behrens authored
No need to check for NULL in send.c and disk-io.c. Signed-off-by:
-
Stefan Behrens authored
Signed-off-by:
-
Liu Bo authored
We don't need to copy it back to user side as it remains unchanged. Signed-off-by:
Liu Bo <bo.li.liu@oracle.com> Signed-off-by:
-
Andreas Philipp authored
Clean up the format of the definitions of BTRFS_BLOCK_GROUP_RAID5 and BTRFS_BLOCK_GROUP_RAID6. Signed-off-by:
Andreas Philipp <philipp.andreas@gmail.com> Signed-off-by:
-
Tsutomu Itoh authored
sctx is removed from the argument of the function that doesn't use sctx. Signed-off-by:
Tsutomu Itoh <t-itoh@jp.fujitsu.com> Signed-off-by:
-
Stefan Behrens authored
The size parameter to btrfs_extend_item() is the number of bytes to add to the item, not the size of the item after the operation (like it is for btrfs_truncate_item(), there the size parameter is not the number of bytes to take away, but the total size of the item after truncation). Fix it in the comment. Signed-off-by:
-
Jan Schmidt authored
btrfs_qgroup_wait_for_completion waits until the currently running qgroup operation completes. It returns immediately when no rescan process is in progress. This is useful to automate things around the rescan process (e.g. testing). Signed-off-by:
Jan Schmidt <list.btrfs@jan-o-sch.net> Signed-off-by:
-
Wang Shilong authored
When doing qgroup accounting, we call ulist_alloc()/ulist_free() every time when we want to walk qgroup tree. By introducing 'qgroup_ulist', we only need to call ulist_alloc()/ulist_free() once. This reduce some sys time to allocate memory, see the measurements below fsstress -p 4 -n 10000 -d $dir With this patch: real 0m50.153s user 0m0.081s sys 0m6.294s real 0m51.113s user 0m0.092s sys 0m6.220s real 0m52.610s user 0m0.096s sys 0m6.125s avg 6.213 ----------------------------------------------------- Without the patch: real 0m54.825s user 0m0.061s sys 0m10.665s real 1m6.401s user 0m0.089s sys 0m11.218s real 1m13.768s user 0m0.087s sys 0m10.665s avg 10.849 we can see the sys time reduce ~43%. Signed-off-by:
Wang Shilong <wangsl-fnst@cn.fujitsu.com> Signed-off-by:
-
David Sterba authored
We want to know if there are debugging features compiled in, this may affect performance. The message is printed before the sanity checks. Also kill version.h file that serves no purpose, we don't use any version tag for kernel module. Signed-off-by:
David Sterba <dsterba@suse.cz> Signed-off-by:
-
David Sterba authored
Signed-off-by:
-
David Sterba authored
And change the message level to KERN_INFO. Signed-off-by:
-
David Sterba authored
The 'end' value must exactly cover the end of the interval, which means one byte less than the expected block alignment, or in case of a file smaller than one block, one byte less than the inode size. Signed-off-by:
-
Henrik Nordvik authored
Code checked for raid 5 flag in two else-if branches, so code would never be reached. Probably a copy-paste bug. Signed-off-by:
Henrik Nordvik <henrikno@gmail.com> Signed-off-by:
-
- 08 Jun, 2013 5 commits
-
-
Josef Bacik authored
Dave reported a panic because the extent_root->commit_root was NULL in the caching kthread. That is because we just unset it in free_root_pointers, which is not the correct thing to do, we have to either wait for the caching kthread to complete or hold the extent_commit_sem lock so we know the thread has exited. This patch makes the kthreads all stop first and then we do our cleanup. This should fix the race. Thanks, Reported-by:
-
Liu Bo authored
Commit be283b2e ( Btrfs: use helper to cleanup tree roots) introduced the following bug, BUG: unable to handle kernel NULL pointer dereference at 0000000000000034 IP: [<ffffffffa039368c>] extent_buffer_get+0x4/0xa [btrfs] [...] Pid: 2463, comm: btrfs-cache-1 Tainted: G O 3.9.0+ #4 innotek GmbH VirtualBox/VirtualBox RIP: 0010:[<ffffffffa039368c>] [<ffffffffa039368c>] extent_buffer_get+0x4/0xa [btrfs] Process btrfs-cache-1 (pid: 2463, threadinfo ffff880112d60000, task ffff880117679730) [...] Call Trace: [<ffffffffa0398a99>] btrfs_search_slot+0x104/0x64d [btrfs] [<ffffffffa039aea4>] btrfs_next_old_leaf+0xa7/0x334 [btrfs] [<ffffffffa039b141>] btrfs_next_leaf+0x10/0x12 [btrfs] [<ffffffffa039ea13>] caching_thread+0x1a3/0x2e0 [btrfs] [<ffffffffa03d8811>] worker_loop+0x14b/0x48e [btrfs] [<ffffffffa03d86c6>] ? btrfs_queue_worker+0x25c/0x25c [btrfs] [<ffffffff81068d3d>] kthread+0x8d/0x95 [<ffffffff81068cb0>] ? kthread_freezable_should_stop+0x43/0x43 [<ffffffff8151e5ac>] ret_from_fork+0x7c/0xb0 [<ffffffff81068cb0>] ? kthread_freezable_should_stop+0x43/0x43 RIP [<ffffffffa039368c>] extent_buffer_get+0x4/0xa [btrfs] We've free'ed commit_root before actually getting to free block groups where caching thread needs valid extent_root->commit_root. Signed-off-by:
Chris Mason <chris.mason@fusionio.com>
-
Josef Bacik authored
Dave reported a NULL pointer deref. This is caused because he thought he'd be smart and add sanity checks to the extent_io bit operations, but he didn't expect a tree to have a NULL mapping. To fix this we just need to init the relocation's processed_blocks with the btree_inode->i_mapping. Thanks, Reported-by:
-
Naohiro Aota authored
There is a path where btrfs_drop_inode() is called with its inode's root is NULL: In btrfs_new_inode(), when btrfs_set_inode_index() fails, iput() is called. We should handle this case before taking look at the root->root_item. Signed-off-by:
Naohiro Aota <naota@elisp.net> Reviewed-by:
-
Josef Bacik authored
We get a use after free if we had a transaction to cleanup since there could be delayed inodes which refer to their respective fs_root. Thanks Reported-by:
-
- 18 May, 2013 19 commits
-
-
-
Chris Mason authored
Btrfs has been pointer tagging bi_private and using bi_bdev to store the stripe index and mirror number of failed IOs. As bios bubble back up through the call chain, we use these to decide if and how to retry our IOs. They are also used to count IO failures on a per device basis. Recently a bio tracepoint was added lead to crashes because we were abusing bi_bdev. This commit adds a btrfs bioset, and creates explicit fields for the mirror number and stripe index. The plan is to extend this structure for all of the fields currently in struct btrfs_bio, which will mean one less kmalloc in our IO path. Signed-off-by:
Tejun Heo <tj@kernel.org>
-
Josef Bacik authored
If we fail to load the chunk tree we'll call free_root_pointers, except we may not have assigned the roots for the dev_root/extent_root/csum_root yet, so we could NULL pointer deref at this point. Just add checks to make sure these roots are set to keep us from panicing. Thanks, Signed-off-by: -
Stefan Behrens authored
The quota_tree was set up to use the empty_block_rsv before which would be problematic when the filesystem is filled up and ENOSPC happens during internal operations while the quota tree is updated and COWed (when the btrfs_qgroup_info_item items) are written. In fact, use_block_rsv() which is used in btrfs_cow_block() falls back to the global_block_rsv in this case. But just in order to make it more clear what is happening, change it to explicitly use the global_block_rsv. Signed-off-by:
-
Alexandre Oliva authored
end_bio_extent_readpage computes whole_page based on bv_offset and bv_len, without taking into account that blk_update_request may modify them when some of the blocks to be read into a page produce a read error. This would cause the read to unlock only part of the file range associated with the page, which would in turn leave the entire page locked, which would not only keep the process blocked instead of returning -EIO to it, but also prevent any further access to the file. It turns out that btrfs always issues whole-page reads and writes. The special handling of non-whole_page appears to be a mistake or a left-over from a time when this wasn't the case. Indeed, end_bio_extent_writepage distinguished between whole_page and non-whole_page writes but behaved identically in both cases! I've replaced the whole_page computations with warnings, just to be sure that we're not issuing partial page reads or writes. The warnings should probably just go away some time. Signed-off-by:
Alexandre Oliva <oliva@gnu.org> Signed-off-by:
-
Miao Xie authored
btrfs_invalidate_inodes() may sleep, so we should not invoke it in the spin lock context. Fix it. Signed-off-by:
-
Miao Xie authored
We have checked if ->node is NULL or not, so it is unnecessary to use BUG_ON() to check again. Remove it. Signed-off-by:
-
Miao Xie authored
Signed-off-by:
-
Miao Xie authored
The root node of the rb-tree may be changed, so we should get it under the lock. Fix it. Signed-off-by:
-
Miao Xie authored
inode_tree_del() will move the tree root into the dead root list, and then the tree will be destroyed by the cleaner. So if we remove the delayed node which is cached in the inode after inode_tree_del(), we may access a freed tree root. Fix it. Signed-off-by:
-
Liu Bo authored
We need to set return value explicitly, otherwise we'll lose the error value. Signed-off-by:
-
Miao Xie authored
Before applying this patch, we reserved the space for the global reserve by the minimum unit if we found it is empty, it was unreasonable and inefficient, because if the global reserve space was depleted, it implied that the size of the global reserve was too small. In this case, we shoud update the global reserve and fill it. Cc: Tsutomu Itoh <t-itoh@jp.fujitsu.com> Signed-off-by:
-
Miao Xie authored
If the type of the space we need is different with the global reserve, we can not steal the space from the global reserve, because we can not allocate the space from the free space cache that the global reserve points to. Cc: Tsutomu Itoh <t-itoh@jp.fujitsu.com> Signed-off-by:
-
Miao Xie authored
cc: Tsutomu Itoh <t-itoh@jp.fujitsu.com> Signed-off-by:
-
Miao Xie authored
It is very likely that there are lots of subvolumes/snapshots in the filesystem, so if we use global block reservation to do inode cache truncation, we may hog all the free space that is reserved in global rsv. So it is better that we do the free space reservation for inode cache truncation by ourselves. Cc: Tsutomu Itoh <t-itoh@jp.fujitsu.com> Signed-off-by:
-
Miao Xie authored
The filesystem with inode cache was forced to be read-only when we umounted it. Steps to reproduce: # mkfs.btrfs -f ${DEV} # mount -o inode_cache ${DEV} ${MNT} # dd if=/dev/zero of=${MNT}/file1 bs=1M count=8192 # btrfs fi syn ${MNT} # dd if=${MNT}/file1 of=/dev/null bs=1M # rm -f ${MNT}/file1 # btrfs fi syn ${MNT} # umount ${MNT} It is because there was no enough space to do inode cache truncation, and then we aborted the current transaction. But no space error is not a serious problem when we write out the inode cache, and it is safe that we just skip this step if we meet this problem. So we need not abort the current transaction. Reported-by: -
Andreas Philipp authored
Raid5 with 3 devices is well defined while the old logic allowed raid5 only with a minimum of 4 devices when converting the block group profile via btrfs balance. Creating a raid5 with just three devices using mkfs.btrfs worked always as expected. This is now fixed and the whole logic is rewritten. Signed-off-by:
-
Stefan Behrens authored
In replace_path(), if read_tree_block() fails, we cannot return directly, we should free some allocated memory otherwise memory leak happens. Similar to Wang's "Btrfs: fix possible memory leak in the find_parent_nodes()" patch, the current commit fixes an issue that is related to the "Btrfs: fix all callers of read_tree_block" commit. Signed-off-by:
-
Wang Shilong authored
In the find_parent_nodes(), if read_tree_block() fails, we can not return directly, we should free some allocated memory otherwise memory leak happens. Signed-off-by:
-
