ext3-get-blocks support caused ~20% degrade in Sequential read
performance (tiobench). Problem is with marking the buffer boundary
so IO can be submitted right away. Here is the patch to fix it.
2.6.18-rc6:
-----------
# ./iotest
1048576+0 records in
1048576+0 records out
4294967296 bytes (4.3 GB) copied, 75.2726 seconds, 57.1 MB/s
real 1m15.285s
user 0m0.276s
sys 0m3.884s
2.6.18-rc6 + fix:
-----------------
[root@elm3a241 ~]# ./iotest
1048576+0 records in
1048576+0 records out
4294967296 bytes (4.3 GB) copied, 62.9356 seconds, 68.2 MB/s
The boundary block check in ext3_get_blocks_handle needs to be adjusted
against the count of blocks mapped in this call, now that it can map
more than one block.
Signed-off-by: Suparna Bhattacharya <suparna@in.ibm.com>
Tested-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It has been reported that ext3_getblk() is not doing the right thing and
triggering following WARN():
BUG: warning at fs/ext3/inode.c:1016/ext3_getblk()
<c01c5140> ext3_getblk+0x98/0x2a6 <c03b2806> md_wakeup_thread+0x26/0x2a
<c01c536d> ext3_bread+0x1f/0x88 <c01cedf9> ext3_quota_read+0x136/0x1ae
<c018b683> v1_read_dqblk+0x61/0xac <c0188f32> dquot_acquire+0xf6/0x107
<c01ceaba> ext3_acquire_dquot+0x46/0x68 <c01897d4> dqget+0x155/0x1e7
<c018a97b> dquot_transfer+0x3e0/0x3e9 <c016fe52> dput+0x23/0x13e
<c01c7986> ext3_setattr+0xc3/0x240 <c0120f66> current_fs_time+0x52/0x6a
<c017320e> notify_change+0x2bd/0x30d <c0159246> chown_common+0x9c/0xc5
<c02a222c> strncpy_from_user+0x3b/0x68 <c0167fe6> do_path_lookup+0xdf/0x266
<c016841b> __user_walk_fd+0x44/0x5a <c01592b9> sys_chown+0x4a/0x55
<c015a43c> vfs_write+0xe7/0x13c <c01695d4> sys_mkdir+0x1f/0x23
<c0102a97> syscall_call+0x7/0xb
Looking at the code, it looks like it's not handle HOLE correctly. It ends
up returning -EIO. Here is the patch to fix it.
If we really want to be paranoid, we can allow return values 0 (HOLE), 1
(we asked for one block) and return -EIO for more than 1 block. But I
really don't see a reason for doing it - all we need is the block# here.
(doesn't matter how many blocks are mapped).
ext3_get_blocks_handle() returns number of blocks it mapped. It returns 0
in case of HOLE. ext3_getblk() should handle HOLE properly (currently its
dumping warning stack and returning -EIO).
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Acked-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
For files other than IFREG, nobh option doesn't make sense. Modifications
to them are journalled and needs buffer heads to do that. Without this
patch, we get kernel oops in page_buffers().
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The inode number out of an NFS file handle gets passed eventually to
ext3_get_inode_block() without any checking. If ext3_get_inode_block()
allows it to trigger an error, then bad filehandles can have unpleasant
effect - ext3_error() will usually cause a forced read-only remount, or a
panic if `errors=panic' was used.
So remove the call to ext3_error there and put a matching check in
ext3/namei.c where inode numbers are read off storage.
[akpm@osdl.org: fix off-by-one error]
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Cc: Marcel Holtmann <marcel@holtmann.org>
Cc: <stable@kernel.org>
Cc: "Stephen C. Tweedie" <sct@redhat.com>
Cc: Eric Sandeen <esandeen@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Same as with already do with the file operations: keep them in .rodata and
prevents people from doing runtime patching.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Steven French <sfrench@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Convert the ext3 in-kernel filesystem blocks to ext3_fsblk_t. Convert the
rest of all unsigned long type in-kernel filesystem blocks to ext3_fsblk_t,
and replace the printk format string respondingly.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Some of the in-kernel ext3 block variable type are treated as signed 4 bytes
int type, thus limited ext3 filesystem to 8TB (4kblock size based). While
trying to fix them, it seems quite confusing in the ext3 code where some
blocks are filesystem-wide blocks, some are group relative offsets that need
to be signed value (as -1 has special meaning). So it seem saner to define
two types of physical blocks: one is filesystem wide blocks, another is
group-relative blocks. The following patches clarify these two types of
blocks in the ext3 code, and fix the type bugs which limit current 32 bit ext3
filesystem limit to 8TB.
With this series of patches and the percpu counter data type changes in the mm
tree, we are able to extend exts filesystem limit to 16TB.
This work is also a pre-request for the recent >32 bit ext3 work, and makes
the kernel to able to address 48 bit ext3 block a lot easier: Simply redefine
ext3_fsblk_t from unsigned long to sector_t and redefine the format string for
ext3 filesystem block corresponding.
Two RFC with a series patches have been posted to ext2-devel list and have
been reviewed and discussed:
http://marc.theaimsgroup.com/?l=ext2-devel&m=114722190816690&w=2http://marc.theaimsgroup.com/?l=ext2-devel&m=114784919525942&w=2
Patches are tested on both 32 bit machine and 64 bit machine, <8TB ext3 and
>8TB ext3 filesystem(with the latest to be released e2fsprogs-1.39). Tests
includes overnight fsx, tiobench, dbench and fsstress.
This patch:
Defines ext3_fsblk_t and ext3_grpblk_t, and the printk format string for
filesystem wide blocks.
This patch classifies all block group relative blocks, and ext3_fsblk_t blocks
occurs in the same function where used to be confusing before. Also include
kernel bug fixes for filesystem wide in-kernel block variables. There are
some fileystem wide blocks are treated as int/unsigned int type in the kernel
currently, especially in ext3 block allocation and reservation code. This
patch fixed those bugs by converting those variables to ext3_fsblk_t(unsigned
long) type.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Some places in ext3 multiple block allocation code (in 2.6.17-rc3) don't
handle the little endian well. This was resulting in *wrong* block numbers
being assigned to in-memory block variables and then stored on disk
eventually. The following patch has been verified to fix an ext3
filesystem failure when run ltp test on a 64 bit machine.
Signed-off-by; Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Mingming Cao recently added multi-block allocation support for ext3,
currently used only by DIO. I added support to map multiple blocks for
mpage_readpages(). This patch add support for ext3_get_block() to deal
with multi-block mapping. Basically it renames ext3_direct_io_get_blocks()
as ext3_get_block().
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
- Clean up a few little layout things and comments.
- Add a WARN_ON to a case which I was wondering about.
- Tune up some inlines.
Cc: Mingming Cao <cmm@us.ibm.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Now that get_block() can handle mapping multiple disk blocks, no need to have
->get_blocks(). This patch removes fs specific ->get_blocks() added for DIO
and makes it users use get_block() instead.
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add support for multiple block allocation in ext3-get-blocks().
Look up the disk block mapping and count the total number of blocks to
allocate, then pass it to ext3_new_block(), where the real block allocation is
performed. Once multiple blocks are allocated, prepare the branch with those
just allocated blocks info and finally splice the whole branch into the block
mapping tree.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Currently ext3_get_block() only maps or allocates one block at a time. This
is quite inefficient for sequential IO workload.
I have posted a early implements a simply multiple block map and allocation
with current ext3. The basic idea is allocating the 1st block in the existing
way, and attempting to allocate the next adjacent blocks on a best effort
basis. More description about the implementation could be found here:
http://marc.theaimsgroup.com/?l=ext2-devel&m=112162230003522&w=2
The following the latest version of the patch: break the original patch into 5
patches, re-worked some logicals, and fixed some bugs. The break ups are:
[patch 1] Adding map multiple blocks at a time in ext3_get_blocks()
[patch 2] Extend ext3_get_blocks() to support multiple block allocation
[patch 3] Implement multiple block allocation in ext3-try-to-allocate
(called via ext3_new_block()).
[patch 4] Proper accounting updates in ext3_new_blocks()
[patch 5] Adjust reservation window size properly (by the given number
of blocks to allocate) before block allocation to increase the
possibility of allocating multiple blocks in a single call.
Tests done so far includes fsx,tiobench and dbench. The following numbers
collected from Direct IO tests (1G file creation/read) shows the system time
have been greatly reduced (more than 50% on my 8 cpu system) with the patches.
1G file DIO write:
2.6.15 2.6.15+patches
real 0m31.275s 0m31.161s
user 0m0.000s 0m0.000s
sys 0m3.384s 0m0.564s
1G file DIO read:
2.6.15 2.6.15+patches
real 0m30.733s 0m30.624s
user 0m0.000s 0m0.004s
sys 0m0.748s 0m0.380s
Some previous test we did on buffered IO with using multiple blocks allocation
and delayed allocation shows noticeable improvement on throughput and system
time.
This patch:
Add support of mapping multiple blocks in one call.
This is useful for DIO reads and re-writes (where blocks are already
allocated), also is in line with Christoph's proposal of using getblocks() in
mpage_readpage() or mpage_readpages().
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The return value of this function is never used, so let's be honest and
declare it as void.
Some places where invalidatepage returned 0, I have inserted comments
suggesting a BUG_ON.
[akpm@osdl.org: JBD BUG fix]
[akpm@osdl.org: rework for git-nfs]
[akpm@osdl.org: don't go BUG in block_invalidate_page()]
Signed-off-by: Neil Brown <neilb@suse.de>
Acked-by: Dave Kleikamp <shaggy@austin.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
ext3's truncate_sem is always released in the same function it's taken
and it otherwise is a mutex as well..
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Linus points out that ext3_readdir's readahead only cuts in when
ext3_readdir() is operating at the very start of the directory. So for large
directories we end up performing no readahead at all and we suck.
So take it all out and use the core VM's page_cache_readahead(). This means
that ext3 directory reads will use all of readahead's dynamic sizing goop.
Note that we're using the directory's filp->f_ra to hold the readahead state,
but readahead is actually being performed against the underlying blockdev's
address_space. Fortunately the readahead code is all set up to handle this.
Tested with printk. It works. I was struggling to find a real workload which
actually cared.
(The patch also exports page_cache_readahead() to GPL modules)
Cc: "Stephen C. Tweedie" <sct@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
One can do "chattr +j" on a file to change its journalling mode. Fix
writeback mode with "nobh" handling for it.
Even though, we mount ext3 filesystem in writeback mode with "nobh" option,
some one can do "chattr +j" on a single file to force it to do journalled
mode. In order to do journaling, ext3_block_truncate_page() need to
fallback to default case of creating buffers and adding them to transaction
etc.
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Migrate a page with buffers without requiring writeback
This introduces a new address space operation migratepage() that may be used
by a filesystem to implement its own version of page migration.
A version is provided that migrates buffers attached to pages. Some
filesystems (ext2, ext3, xfs) are modified to utilize this feature.
The swapper address space operation are modified so that a regular
migrate_page() will occur for anonymous pages without writeback (migrate_pages
forces every anonymous page to have a swap entry).
Signed-off-by: Mike Kravetz <kravetz@us.ibm.com>
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch fixes lost referrence on ext3 current handle in
ext3_journalled_writepage().
Signed-Off-By: Denis Lunev <den@sw.ru>
Cc: Kirill Korotaev <dev@sw.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch adds tests for the return value of sb_getblk() in the ext2/3
filesystems. In fs/buffer.c it is stated that the getblk() function never
fails. However, it does can return NULL in some situations due to I/O
errors, which may lead us to NULL pointer dereferences
Signed-off-by: Glauber de Oliveira Costa <glommer@br.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I noticed some problems while running ext3 with the debug flag set on.
More precisely, I was unable to umount the filesystem. Some investigation
took me to the patch that follows.
At a first glance , the lock/unlock I've taken out seems really not
necessary, as the main code (outside debug) does not lock the super. The
only additional danger operations that debug code introduces seems to be
related to bitmap, but bitmap operations tends to be all atomic anyway.
I also took the opportunity to fix 2 spelling errors.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
- ->releasepage() annotated (s/int/gfp_t), instances updated
- missing gfp_t in fs/* added
- fixed misannotation from the original sweep caught by bitwise checks:
XFS used __nocast both for gfp_t and for flags used by XFS allocator.
The latter left with unsigned int __nocast; we might want to add a
different type for those but for now let's leave them alone. That,
BTW, is a case when __nocast use had been actively confusing - it had
been used in the same code for two different and similar types, with
no way to catch misuses. Switch of gfp_t to bitwise had caught that
immediately...
One tricky bit is left alone to be dealt with later - mapping->flags is
a mix of gfp_t and error indications. Left alone for now.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Update the file systems in fs/ implementing a delete_inode() callback to
call truncate_inode_pages(). One implementation note: In developing this
patch I put the calls to truncate_inode_pages() at the very top of those
filesystems delete_inode() callbacks in order to retain the previous
behavior. I'm guessing that some of those could probably be optimized.
Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
Acked-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
We are not using the in-inode space for xattrs in reserved inodes because
mkfs.ext3 doesn't initialize it properly. For those inodes, we set
i_extra_isize to 0. Make sure that we also don't overwrite the
i_extra_isize field when writing out the inode in that case. This is for
cleanliness only, and doesn't fix an actual bug.
Signed-off-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Use improved credits estimates for quota operations. Also reserve a space
for a quota operation in a transaction only if filesystem was mounted with
some quota options.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This had a fatal lock ranking bug: we do journal_start outside
mpage_writepages()'s lock_page().
Revert the whole thing, think again.
Credit-to: Jan Kara <jack@suse.cz>
For identifying the bug.
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The extra race-with-truncate-then-retry logic around
ext3_get_block_handle(), which was inherited from ext2, becomes unecessary
for ext3, since we have already obtained the ei->truncate_sem in
ext3_get_block_handle() before calling ext3_alloc_branch(). The
ei->truncate_sem is already there to block concurrent truncate and block
allocation on the same inode. So the inode's indirect addressing tree
won't be changed after we grab that semaphore.
We could, after get the semaphore, re-verify the branch is up-to-date or
not. If it has been changed, then get the updated branch. If we still
need block allocation, we will have a safe version of the branch to work
with in the ext3_find_goal()/ext3_splice_branch().
The code becomes more readable after remove those retry logic. The patch
also clean up some gotos in ext3_get_block_handle() to make it more
readable.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Suparna Bhattacharya
Badari Pulavarty
Neil Brown
Christoph Hellwig
Mingming Cao
Andrew Morton
Arjan van de Ven
Christoph Lameter
Denis Lunev
Glauber de Oliveira Costa
Al Viro
Mark Fasheh
Andreas Gruenbacher
Jan Kara
Linus Torvalds