Michael Leun reported that running parallel opens on a fuse filesystem
can trigger a "kernel BUG at mm/truncate.c:475"
Gurudas Pai reported the same bug on NFS.
The reason is, unmap_mapping_range() is not prepared for more than
one concurrent invocation per inode. For example:
thread1: going through a big range, stops in the middle of a vma and
stores the restart address in vm_truncate_count.
thread2: comes in with a small (e.g. single page) unmap request on
the same vma, somewhere before restart_address, finds that the
vma was already unmapped up to the restart address and happily
returns without doing anything.
Another scenario would be two big unmap requests, both having to
restart the unmapping and each one setting vm_truncate_count to its
own value. This could go on forever without any of them being able to
finish.
Truncate and hole punching already serialize with i_mutex. Other
callers of unmap_mapping_range() do not, and it's difficult to get
i_mutex protection for all callers. In particular ->d_revalidate(),
which calls invalidate_inode_pages2_range() in fuse, may be called
with or without i_mutex.
This patch adds a new mutex to 'struct address_space' to prevent
running multiple concurrent unmap_mapping_range() on the same mapping.
[ We'll hopefully get rid of all this with the upcoming mm
preemptibility series by Peter Zijlstra, the "mm: Remove i_mmap_mutex
lockbreak" patch in particular. But that is for 2.6.39 ]
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Reported-by: Michael Leun <lkml20101129@newton.leun.net>
Reported-by: Gurudas Pai <gurudas.pai@oracle.com>
Tested-by: Gurudas Pai <gurudas.pai@oracle.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This adds fiemap to nilfs. Two new functions, nilfs_fiemap and
nilfs_find_uncommitted_extent are added.
nilfs_fiemap() implements the fiemap inode operation, and
nilfs_find_uncommitted_extent() helps to get a range of data blocks
whose physical location has not been determined.
nilfs_fiemap() collects extent information by looping through
nilfs_bmap_lookup_contig and nilfs_find_uncommitted_extent routines.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
This applies prepared rollback function and redirect function of
metadata file to DAT file, and eliminates GCDAT inode.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
During garbage collection (GC), DAT file, which converts virtual block
number to real block number, may return disk block number that is not
yet written to the device.
To avoid access to unwritten blocks, the current implementation stores
changes to the caches of GCDAT during GC and atomically commit the
changes into the DAT file after they are written to the device.
This patch, instead, adds a function that makes a copy of specified
buffer and stores it in nilfs_shadow_map, and a function to get the
backup copy as needed (nilfs_mdt_freeze_buffer and
nilfs_mdt_get_frozen_buffer respectively).
Before DAT changes block number in an entry block, it makes a copy and
redirect access to the buffer so that address conversion function
(i.e. nilfs_dat_translate) refers to the old address saved in the
copy.
This patch gives requisites for such redirection.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
This adds optional function to metadata files which makes a copy of
bmap, page caches, and b-tree node cache, and rolls back to the copy
as needed.
This enhancement is intended to displace gcdat inode that provides a
similar function in a different way.
In this patch, nilfs_shadow_map structure is added to store a copy of
the foregoing states. nilfs_mdt_setup_shadow_map relates this
structure to a metadata file. And, nilfs_mdt_save_to_shadow_map() and
nilfs_mdt_restore_from_shadow_map() provides save and restore
functions respectively. Finally, nilfs_mdt_clear_shadow_map() clears
states of nilfs_shadow_map.
The copy of b-tree node cache and page cache is made by duplicating
only dirty pages into corresponding caches in nilfs_shadow_map. Their
restoration is done by clearing dirty pages from original caches and
by copying dirty pages back from nilfs_shadow_map.
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>
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 <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
This would fix the following failure during GC:
nilfs_cpfile_delete_checkpoints: cannot delete block
NILFS: GC failed during preparation: cannot delete checkpoints: err=-2
The problem was caused by a break in state consistency between page
cache and btree; the above block was removed from the btree but the
page buffering the block was remaining in the page cache in dirty
state.
This resolves the inconsistency by ensuring to clear dirty state of
the page buffering the deleted block.
Reported-by: David Arendt <admin@prnet.org>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Pekka Enberg advised me:
> It would be nice if BUG(), BUG_ON(), and panic() calls would be
> converted to proper error handling using WARN_ON() calls. The BUG()
> call in nilfs_cpfile_delete_checkpoints(), for example, looks to be
> triggerable from user-space via the ioctl() system call.
This will follow the comment and keep them to a minimum.
Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>
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>
This adds common routines for buffer/page operations used in B-tree
node caches, meta data files, or segment constructor (log writer).
NILFS uses copy functions for buffers and pages due to the following
reasons:
1) Relocation required for COW
Since NILFS changes address of on-disk blocks, moving buffers
in page cache is needed for the buffers which are not addressed
by a file offset. If buffer size is smaller than page size,
this involves partial copy of pages.
2) Freezing mmapped pages
NILFS calculates checksums for each log to ensure its validity.
If page data changes after the checksum calculation, this validity
check will not work correctly. To avoid this failure for mmaped
pages, NILFS freezes their data by copying.
3) Copy-on-write for DAT pages
NILFS makes clones of DAT page caches in a copy-on-write manner
during GC processes, and this ensures atomicity and consistency
of the DAT in the transient state.
In addition, NILFS uses two obsolete functions, nilfs_mark_buffer_dirty()
and nilfs_clear_page_dirty() respectively.
* nilfs_mark_buffer_dirty() was required to avoid NULL pointer
dereference faults:
Since the page cache of B-tree node pages or data page cache of pseudo
inodes does not have a valid mapping->host, calling mark_buffer_dirty()
for their buffers causes the fault; it calls __mark_inode_dirty(NULL)
through __set_page_dirty().
* nilfs_clear_page_dirty() was needed in the two cases:
1) For B-tree node pages and data pages of the dat/gcdat, NILFS2 clears
page dirty flags when it copies back pages from the cloned cache
(gcdat->{i_mapping,i_btnode_cache}) to its original cache
(dat->{i_mapping,i_btnode_cache}).
2) Some B-tree operations like insertion or deletion may dispose buffers
in dirty state, and this needs to cancel the dirty state of their
pages. clear_page_dirty_for_io() caused faults because it does not
clear the dirty tag on the page cache.
Signed-off-by: Seiji Kihara <kihara.seiji@lab.ntt.co.jp>
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>
