GitLab

commit 957ed60b

 upstream.

Each inode of nilfs2 stores a root node of a b-tree, and it turned out to
have a memory overrun issue:

Each b-tree node of nilfs2 stores a set of key-value pairs and the number
of them (in "bn_nchildren" member of nilfs_btree_node struct), as well as
a few other "bn_*" members.

Since the value of "bn_nchildren" is used for operations on the key-values
within the b-tree node, it can cause memory access overrun if a large
number is incorrectly set to "bn_nchildren".

For instance, nilfs_btree_node_lookup() function determines the range of
binary search with it, and too large "bn_nchildren" leads
nilfs_btree_node_get_key() in that function to overrun.

As for intermediate b-tree nodes, this is prevented by a sanity check
performed when each node is read from a drive, however, no sanity check
has been done for root nodes stored in inodes.

This patch fixes the issue by adding missing sanity check against b-tree
root nodes so that it's called when on-memory inodes are read from ifile,
inode metadata file.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

nilfs_btree_delete function does not terminate part of virtual block
addresses when shrinking the last remaining child node into the root
node.  The missing address termination causes that dead btree node
blocks persist and chip away free disk space.

This fixes the leak bug on the btree node deletion.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_btree_delete function wrongly terminates virtual block address
of the btree node held by its parent at index 0.  When concatenating
the index-0 node with its right sibling node, nilfs_btree_delete
terminates the block address of index-0 node instead of the right
sibling node which should be deleted.

This bug not only wears disk space in the long run, but also causes
file system corruption.  This will fix it.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This replaces nilfs_mdt_mark_buffer_dirty and nilfs_btnode_mark_dirty
macros with mark_buffer_dirty and gets rid of nilfs_mark_buffer_dirty,
an own mark buffer dirty function.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This records the number of used blocks per checkpoint in each
checkpoint entry of cpfile.  Even though userland tools can get the
block count via nilfs_get_cpinfo ioctl, it was not updated by the
nilfs2 kernel code.  This fixes the issue and makes it available for
userland tools to calculate used amount per checkpoint.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Cc: Jiro SEKIBA <jir@unicus.jp>

This applies read-ahead to nilfs_btree_do_lookup and
nilfs_btree_lookup_contig functions and extends them to read ahead
siblings of level 1 btree nodes that hold data blocks.

At present, the read-ahead is not applied to most btree operations;
only get_block() callback function, which is used during read of
regular files or directories, receives the benefit.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_btree_get_block() now may return untested buffer due to
read-ahead.  This adds a new flag for buffer heads so that the btree
code can check whether the buffer is already verified or not.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This adds __nilfs_btree_get_block() function that can issue a series
of read-ahead requests for sibling btree nodes.

This read-ahead needs parent node block, so nilfs_btree_readahead_info
structure is added to pass the information that
__nilfs_btree_get_block() needs.

This also replaces the previous nilfs_btree_get_block() implementation
with a wrapper function of __nilfs_btree_get_block().
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This adds mode argument to nilfs_btnode_submit_block() function and
allows it to issue a read-ahead request.

An optional submit_ptr argument is also added to store the actual
block address for which bio is sent.  submit_ptr is used for a series
of read-ahead requests, and helps to decide if each requested block is
continous to the previous one on disk.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This removes all inline uses from btree.c.  Gcc now agressively apply
inline expansion even for the functions declared without the keyword;
the inline use in btree.c looks excessive.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

The patch "reduce repetitive calculation of max number of child nodes"
gathered up the calculation of maximum number of child nodes into
nilfs_btree_nchildren_per_block() function.  This makes the function
get resultant value from a private variable in bmap object instead of
calculating it for each call.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

The current btree implementation repeats the same calculation on the
maximum number of child nodes.  This is because a few low level
routines use the calculation for index addressing in a btree node
block.

This reduces the calculation by explicitly passing the maximum number
of child nodes (ncmax) through their argument.

This changes parameter passing of the following functions:

 - nilfs_btree_node_dptrs
 - nilfs_btree_node_get_ptr
 - nilfs_btree_node_set_ptr
 - nilfs_btree_node_init
 - nilfs_btree_node_move_left
 - nilfs_btree_node_move_right
 - nilfs_btree_node_insert
 - nilfs_btree_node_delete, and
 - nilfs_btree_get_node

The following functions are removed:

 - nilfs_btree_node_nchildren_min
 - nilfs_btree_node_nchildren_max

Most middle level btree operations are rewritten to pass a proper
ncmax value depending on whether each occurrence of node is "root" or
not.

A constant NILFS_BTREE_ROOT_NCHILDREN_MAX is used for the root node,
whereas nilfs_btree_nchildren_per_block() function is used for
non-root nodes.  If a node could be either root or a non-root node, an
output argument of nilfs_btree_get_node() is used to set up ncmax.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_btree_node_nchildren_max() and nilfs_btree_node_nchildren_min()
functions switch return value depending on whether target node is the
root or a node block.  In most uses of these functions, however, the
node type is fixed, and moreover the same calculation is repeatedly
performed in loop.

This unfold these functions depending on context and move them outside
loops wherever possible.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_bmap_lookup and its variants are supposed to take a valid
pointer argument to return a block address, thus pointer checks in
nilfs_btree_lookup and nilfs_direct_lookup are needless.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This unifies two similar functions nilfs_btree_set_target_v and
nilfs_direct_set_target_v into one, nilfs_bmap_set_target_v.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This replaces all uses of nilfs_btree struct in implementation of
btree mapping with nilfs_bmap struct.

Name of local variable "btree" is kept not to bloat amount of change.
And, a part of local variables "bmap" is renamed to "btree" to uniform
naming rule.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

The first argument of bops->bop_propagate operation takes a constant
qualifier, and causes compilation error when removed cast to pointer
of nilfs_btree structure type.  This fixes the issue to prepare for
succesive removal of nilfs_btree struct.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

Will remove nilfs_bmap_key_to_dkey(), nilfs_bmap_dkey_to_key(),
nilfs_bmap_ptr_to_dptr(), and nilfs_bmap_dptr_to_ptr() for simplicity.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This inserts sanity checks soon after read btree node from disk.  This
allows early detection of broken btree nodes, and helps to narrow down
problems due to file system corruption.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

According to the report titled "problem with nilfs_cleanerd" from
Łukasz Wójcicki, nilfs_btree_lookup_dirty_buffers or
nilfs_btree_add_dirty_buffer got memory violation during garbage
collection.

This could happen if a level field of given btree node buffer is
incorrect, which is a crucial internal bug.

This inserts a sanity check to figure out the problem.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This cleanup patch gives several improvements:

 - Moving all kmem_cache_{create_destroy} calls into one place, which removes
 some small function calls, cleans up error check code and clarify the logic.

 - Mark all initial code in __init section.

 - Remove some very obvious comments.

 - Adjust some declarations.

 - Fix some space-tab issues.
Signed-off-by: Li Hong <lihong.hi@gmail.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_btree_release_path() and nilfs_btree_free_path() are bound into each other
tightly. Make them into one procedure to clearify the logic and avoid some
misusages.
Signed-off-by: Li Hong <lihong.hi@gmail.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_btree_alloc_path() and nilfs_btree_init_path() are bound into each other
tightly. Make them into one procedure to clearify the logic and avoid some
misusages.
Signed-off-by: Li Hong <lihong.hi@gmail.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

`make CONFIG_NILFS2_FS=m M=fs/nilfs2/` will give the following warnings:

fs/nilfs2/btree.c: In function 'nilfs_btree_propagate':
fs/nilfs2/btree.c:1882: warning: 'maxlevel' may be used uninitialized in this function
fs/nilfs2/btree.c:1882: note: 'maxlevel' was declared here

Set maxlevel = 0 to fix it.
Signed-off-by: Li Hong <lihong.hi@gmail.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This applies list_splice_tail (or list_splice_tail_init) operation
instead of list_splice (or list_splice_init, respectively) to append a
new list to tail of an existing list.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This removes the obsolete nilfs_btnode_get() function and makes
nilfs_btree_get_block() directly call nilfs_btnode_submit_block().

This expansion will provide better opportunity for code optimization.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This displaces nilfs_btnode_get() use to create new btree node block
with nilfs_btnode_create_block.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

lock_buffer() and unlock_buffer() uses in btree.c are eliminable
because btree functions gain buffer heads through nilfs_btnode_get(),
which never returns an on-the-fly buffer.

Although nilfs_clear_dirty_page() and nilfs_copy_back_pages() in
nilfs_commit_gcdat_inode() juggle btree node buffers of DAT, this is
safe because these operations are protected by a log writer lock or
the metadata file semaphore of DAT.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

The current btree lookup routines make a kernel oops when detected
inconsistency in btree blocks.  These routines should instead return a
proper error code because the inconsistency usually comes from
corruption of on-disk metadata.

This fixes the issue by converting BUG_ON calls to proper error
handlings.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

The current btree code is written so that btree functions call dat
operations via wrapper functions in bmap.c when they allocate, free,
or modify virtual block addresses.

This abstraction requires additional function calls and causes
frequent call of nilfs_bmap_get_dat() function since it is used in the
every wrapper function.

This removes the wrapper functions and makes them available from
btree.c and direct.c, which will increase the opportunity of
compiler optimization.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

The btree path object is cleared just before it is freed.

This will remove the code doing the unnecessary clear operation.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

Even though many btree functions take a btree object as their first
argument, most of them are not used in their functions.

This sticky use of the btree argument is hurting code readability and
giving the possibility of inefficient code generation.

So, this removes the unnecessary btree arguments.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

Although get_block() callback function can return extent of contiguous
blocks with bh->b_size, nilfs_get_block() function did not support
this feature.

This adds contiguous lookup feature to the block mapping codes of
nilfs, and allows the nilfs_get_blocks() function to return the extent
information by applying the feature.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

will remove indirect function calls using nilfs_btree_operations
table.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

Previously, the bmap codes of nilfs used three types of function
tables.  The abuse of indirect function calls decreased source
readability and suffered many indirect jumps which would confuse
branch prediction of processors.

This eliminates one type of the function tables,
nilfs_bmap_ptr_operations, which was used to dispatch low level
pointer operations of the nilfs bmap.

This adds a new integer variable "b_ptr_type" to nilfs_bmap struct,
and uses the value to select the pointer operations.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This will cut off 16 bytes from the nilfs_bmap struct which is
embedded in the on-memory inode of nilfs.

The b_high field was never used, and the b_low field stores a constant
value which can be determined by whether the inode uses btree for
block mapping or not.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

This indirect function is set to NULL only for gc cache inodes, but
the gc cache inodes never call this function.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

Two get block function for btree nodes, nilfs_bmap_get_block() and
nilfs_bmap_get_new_block(), are called only from the btree codes.
This relocation will increase opportunities of compiler optimization.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_bmap_delete_block() is a wrapper function calling
nilfs_btnode_delete().  This removes it for simplicity.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>

nilfs_bmap_put_block() is a wrapper function calling brelse().  This
eliminates the wrapper for simplicity.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>