23 Jan, 2007
4 commits
-
For large size DIO that needs multiple bio, one full page worth of data was
lost at the boundary of bio's maximum sector or segment limits. After a
bio is full and got submitted. The outer while (nbytes) { ... } loop will
allocate a new bio and just march on to index into next page. It just
forgets about the page that bio_add_page() rejected when previous bio is
full. Fix it by put the rejected page back to pvec so we pick it up again
for the next bio.Signed-off-by: Ken Chen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
size_t is unsigned. IO errors aren't getting through.
Cc: "Chen, Kenneth W"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
* git://git.infradead.org/mtd-2.6: (84 commits)
[JFFS2] debug.h: include for current->pid
[MTD] OneNAND: Handle DDP chip boundary during read-while-load
[MTD] OneNAND: return ecc error code only when 2-bit ecc occurs
[MTD] OneNAND: Implement read-while-load
[MTD] OneNAND: fix onenand_wait bug in read ecc error
[MTD] OneNAND: release CPU in cycles
[MTD] OneNAND: add subpage write support
[MTD] OneNAND: fix onenand_wait bug
[JFFS2] use the ref_offset macro
[JFFS2] Reschedule in loops
[JFFS2] Fix error-path leak in summary scan
[JFFS2] add cond_resched() when garbage collecting deletion dirent
[MTD] Nuke IVR leftovers
[MTD] OneNAND: fix oob handling in recent oob patch
[MTD] Fix ssfdc blksize typo
[JFFS2] replace kmalloc+memset with kzalloc
[MTD] Fix SSFDC build for variable blocksize.
[MTD] ESB2ROM uses PCI
[MTD] of_device-based physmap driver
[MTD] Support combined RedBoot FIS directory and configuration area
... -
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
ocfs2: Add backup superblock info to ocfs2_fs.h
ocfs2: cleanup ocfs2_iget() errors
ocfs2: Directory c/mtime update fixes
ocfs2: Don't print errors when following symlinks
22 Jan, 2007
4 commits
-
This synchronizes us with recent ocfs2-tools changes.
Signed-off-by: Mark Fasheh
-
Get rid of some error prints in the ocfs2_iget() path from
ocfs2_get_dentry(). NFSD can easily cause us to read stale inodes.Signed-off-by: Mark Fasheh
-
ocfs2 wasn't updating c/mtime on directories during dirent
creation/deletion. Fix ocfs2_unlink(), ocfs2_rename() and
__ocfs2_add_entry() by adding the proper code to update the struct inode and
push the change out to disk.This helps rename/unlink on nfs exported file systems in particular as those
clients compare directory time values to avoid a full re-reading a directory
which hasn't changed.ocfs2_rename() loses some superfluous error handling as a result of this
patch.Signed-off-by: Mark Fasheh
-
We shouldn't print errors returned from vfs_follow_link(). This was causing
spurious errors to show up in the logs.Signed-off-by: Mark Fasheh
18 Jan, 2007
3 commits
-
Signed-off-by: Anton Altaparmakov
-
- Fix deadlock in fs/ntfs/inode.c::ntfs_put_inode(). Thanks to Sergey
Vlasov for the report and detailed analysis of the deadlock. The fix
involved getting rid of ntfs_put_inode() altogether and hence NTFS no
longer has a ->put_inode super operation.Signed-off-by: Anton Altaparmakov
13 Jan, 2007
1 commit
-
Signed-off-by: David Woodhouse
12 Jan, 2007
2 commits
-
Revert bd_mount_mutex back to a semaphore so that xfs_freeze -f /mnt/newtest;
xfs_freeze -u /mnt/newtest works safely and doesn't produce lockdep warnings.(XFS unlocks the semaphore from a different task, by design. The mutex
code warns about this)Signed-off-by: Dave Chinner
Cc: Ingo Molnar
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
NFS: Fix race in nfs_release_page()
invalidate_inode_pages2() may find the dirty bit has been set on a page
owing to the fact that the page may still be mapped after it was locked.
Only after the call to unmap_mapping_range() are we sure that the page
can no longer be dirtied.
In order to fix this, NFS has hooked the releasepage() method and tries
to write the page out between the call to unmap_mapping_range() and the
call to remove_mapping(). This, however leads to deadlocks in the page
reclaim code, where the page may be locked without holding a reference
to the inode or dentry.Fix is to add a new address_space_operation, launder_page(), which will
attempt to write out a dirty page without releasing the page lock.Signed-off-by: Trond Myklebust
Also, the bare SetPageDirty() can skew all sort of accounting leading to
other nasties.[akpm@osdl.org: cleanup]
Signed-off-by: Peter Zijlstra
Cc: Trond Myklebust
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
11 Jan, 2007
1 commit
-
Revert previous attempts at messing with the linux banner string and
simply use a separate format string for proc.Signed-off-by: Roman Zippel
Acked-by: Olaf Hering
Acked-by: Jean Delvare
Cc: Andrey Borzenkov
Cc: Andrew Morton
Cc: Andy Whitcroft
Cc: Herbert Poetzl
Signed-off-by: Linus Torvalds
10 Jan, 2007
2 commits
-
Don't use ref->flash_offset directly in debugging code, use the ref_offset macro instead.
Signed-off-by: Kyungmin Park
Signed-off-by: Artem Bityutskiy -
Make JFFS2 nicer and teach it to call cond_resched() in loops
which may be quite large.Signed-off-by: Artem Bityutskiy
07 Jan, 2007
1 commit
-
This reverts commit 59287c0913cc9a6c75712a775f6c1c1ef418ef3b.
Hugh Dickins reports that it causes random failures on x86 with SuSE
10.2, and points out"Isn't that randomization, anywhere from 0x10000 to ELF_ET_DYN_BASE,
sure to place the ET_DYN from time to time just where the comment
says it's trying to avoid? I assume that somehow results in the error
reported."(where the comment in question is the existing comment in the source
code about mmap/brk clashes).Suggested-by: Hugh Dickins
Acked-by: Marcus Meissner
Cc: Andrew Morton
Cc: Andi Kleen
Cc: Ingo Molnar
Cc: Dave Jones
Cc: Arjan van de Ven
Signed-off-by: Linus Torvalds
06 Jan, 2007
3 commits
-
Looks like this is the problem, which point Al Viro some time ago:
ufs's get_block callback allocates 16k of disk at a time, and links that
entire 16k into the file's metadata. But because get_block is called for only
a single buffer_head (a 2k buffer_head in this case?) we are only able to tell
the VFS that this 2k is buffer_new().So when ufs_getfrag_block() is later called to map some more data in the file,
and when that data resides within the remaining 14k of this fragment,
ufs_getfrag_block() will incorrectly return a !buffer_new() buffer_head.I don't see _right_ way to do nullification of whole block, if use inode
page cache, some pages may be outside of inode limits (inode size), and
will be lost; if use blockdev page cache it is possible to zero real data,
if later inode page cache will be used.The simpliest way, as can I see usage of block device page cache, but not only
mark dirty, but also sync it during "nullification". I use my simple tests
collection, which I used for check that create,open,write,read,close works on
ufs, and I see that this patch makes ufs code 18% slower then before.Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
CVE-2006-5753 is for a case where an inode can be marked bad, switching
the ops to bad_inode_ops, which are all connected as:static int return_EIO(void)
{
return -EIO;
}#define EIO_ERROR ((void *) (return_EIO))
static struct inode_operations bad_inode_ops =
{
.create = bad_inode_create
...etc...The problem here is that the void cast causes return types to not be
promoted, and for ops such as listxattr which expect more than 32 bits of
return value, the 32-bit -EIO is interpreted as a large positive 64-bit
number, i.e. 0x00000000fffffffa instead of 0xfffffffa.This goes particularly badly when the return value is taken as a number of
bytes to copy into, say, a user's buffer for example...I originally had coded up the fix by creating a return_EIO_ macro
for each return type, like this:static int return_EIO_int(void)
{
return -EIO;
}
#define EIO_ERROR_INT ((void *) (return_EIO_int))static struct inode_operations bad_inode_ops =
{
.create = EIO_ERROR_INT,
...etc...but Al felt that it was probably better to create an EIO-returner for each
actual op signature. Since so few ops share a signature, I just went ahead
& created an EIO function for each individual file & inode op that returns
a value.Signed-off-by: Eric Sandeen
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix filenames on adfs discs being terminated at the first character greater
than 128 (adfs filenames are Latin 1). I saw this problem when using a
loopback adfs image on a 2.6.17-rc5 x86_64 machine, and the patch fixed it
there.Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
03 Jan, 2007
1 commit
-
Signed-off-by: Amit Choudhary
Signed-off-by: Andrew Morton
Signed-off-by: David Woodhouse
31 Dec, 2006
3 commits
-
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
ocfs2: export heartbeat thread pid via configfs
ocfs2: always unmap in ocfs2_data_convert_worker()
ocfs2: ignore NULL vfsmnt in ocfs2_should_update_atime()
ocfs2: Allow direct I/O read past end of file
ocfs2: don't print error in ocfs2_permission() -
ramfs doesn't provide the .set_dirty_page a_op, and when the BLOCK layer is
not configured in, 'set_page_dirty' makes a call via a NULL pointer.Signed-off-by: Dimitri Gorokhovik
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
lockdep found a AB BC CA lock inversion in retry-based AIO:
1) The task struct's alloc_lock (A) is acquired in process context with
interrupts enabled. An interrupt might arrive and call wake_up() which
grabs the wait queue's q->lock (B).2) When performing retry-based AIO the AIO core registers
aio_wake_function() as the wake funtion for iocb->ki_wait. It is called
with the wait queue's q->lock (B) held and then tries to add the iocb to
the run list after acquiring the ctx_lock (C).3) aio_kick_handler() holds the ctx_lock (C) while acquiring the
alloc_lock (A) via lock_task() and unuse_mm(). Lockdep emits a warning
saying that we're trying to connect the irq-safe q->lock to the
irq-unsafe alloc_lock via ctx_lock.This fixes the inversion by calling unuse_mm() in the AIO kick handing path
after we've released the ctx_lock. As Ben LaHaise pointed out __put_ioctx
could set ctx->mm to NULL, so we must only access ctx->mm while we have the
lock.Signed-off-by: Zach Brown
Signed-off-by: Suparna Bhattacharya
Acked-by: Benjamin LaHaise
Cc: "Chen, Kenneth W"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
29 Dec, 2006
5 commits
-
The patch allows the ocfs2 heartbeat thread to prioritize I/O which may
help cut down on spurious fencing. Most of this will be in the tools -
we can have a pid configfs attribute and let userspace (ocfs2_hb_ctl)
calls the ioprio_set syscall after starting heartbeat, but only cfq
scheduler supports I/O priorities now.Signed-off-by: Zhen Wei
Signed-off-by: Mark Fasheh -
Mmap-heavy clustered workloads were sometimes finding stale data on mmap
reads. The solution is to call unmap_mapping_range() on any down convert of
a data lock.Signed-off-by: Mark Fasheh
-
This can come from NFSD.
Signed-off-by: Mark Fasheh
-
ocfs2_direct_IO_get_blocks() was incorrectly returning -EIO for a direct I/O
read whose start block was past the end of the file allocation tree. Fix
things so that we return a hole instead. do_direct_IO() will then notice
that the range start is past eof and return a short read.While there, remove the unused vbo_max variable.
Signed-off-by: Mark Fasheh
-
Errors from generic_permission() can happen in valid cases and shouldn't be
reported.Signed-off-by: Mark Fasheh
24 Dec, 2006
2 commits
-
This also adds he required page "writeback" flag handling, that cifs
hasn't been doing and that the page dirty flag changes made obvious.Acked-by: Steve French
Acked-by: Dave Kleikamp
Signed-off-by: Linus Torvalds -
Thanks to Len Brown for testing this fix, since while they have in the
past, none of my machines run reiserfs at the moment.Cc: Vladimir V. Saveliev
Acked-by: Len Brown
Signed-off-by: Linus Torvalds
23 Dec, 2006
5 commits
-
In the current jbd code, if a buffer on BJ_SyncData list is dirty and not
locked, the buffer is refiled to BJ_Locked list, submitted to the IO and
waited for IO completion.But the fsstress test showed the case that when a buffer was already
submitted to the IO just before the buffer_dirty(bh) check, the buffer was
not waited for IO completion.Following patch solves this problem. If it is assumed that a buffer is
submitted to the IO before the buffer_dirty(bh) check and still being
written to disk, this buffer is refiled to BJ_Locked list.Signed-off-by: Hisashi Hifumi
Cc: Jan Kara
Cc: "Stephen C. Tweedie"
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Christoph Hellwig has expressed concerns that the recent fdtable changes
expose the details of the RCU methodology used to release no-longer-used
fdtable structures to the rest of the kernel. The trivial patch below
addresses these concerns by introducing the appropriate free_fdtable()
calls, which simply wrap the release RCU usage. Since free_fdtable() is a
one-liner, it makes sense to promote it to an inline helper.Signed-off-by: Vadim Lobanov
Cc: Christoph Hellwig
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Mark JFFS as broken and provide a warning to users that it is deprecated
and scheduled for removal in 2.6.21Signed-off-by: Josh Boyer
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Trevor found a file size problem in eCryptfs in recent kernels, and he
tracked it down to an fsstack change.This was the eCryptfs copy_attr_all:
> -void ecryptfs_copy_attr_all(struct inode *dest, const struct inode *src)
> -{
> - dest->i_mode = src->i_mode;
> - dest->i_nlink = src->i_nlink;
> - dest->i_uid = src->i_uid;
> - dest->i_gid = src->i_gid;
> - dest->i_rdev = src->i_rdev;
> - dest->i_atime = src->i_atime;
> - dest->i_mtime = src->i_mtime;
> - dest->i_ctime = src->i_ctime;
> - dest->i_blkbits = src->i_blkbits;
> - dest->i_flags = src->i_flags;
> -}This is the fsstack copy_attr_all:
> +void fsstack_copy_attr_all(struct inode *dest, const struct inode *src,
> + int (*get_nlinks)(struct inode *))
> +{
> + if (!get_nlinks)
> + dest->i_nlink = src->i_nlink;
> + else
> + dest->i_nlink = (*get_nlinks)(dest);
> +
> + dest->i_mode = src->i_mode;
> + dest->i_uid = src->i_uid;
> + dest->i_gid = src->i_gid;
> + dest->i_rdev = src->i_rdev;
> + dest->i_atime = src->i_atime;
> + dest->i_mtime = src->i_mtime;
> + dest->i_ctime = src->i_ctime;
> + dest->i_blkbits = src->i_blkbits;
> + dest->i_flags = src->i_flags;
> +
> + fsstack_copy_inode_size(dest, src);
> +}The addition of copy_inode_size breaks eCryptfs, since eCryptfs needs to
interpolate the file sizes (eCryptfs has extra space in the lower file for
the header). The setting of the upper inode size occurs elsewhere in
eCryptfs, and the new copy_attr_all now undoes what eCryptfs was doing
right beforehand.I see three ways of going forward from here. (1) Something like this patch
needs to go in (assuming it jives with Unionfs), (2) we need to make a
change to the fsstack API for more fine-grained control over copying
attributes (e.g., by also including a callback function for calculating the
right file size, which will require some more work on both eCryptfs and
Unionfs), or (3) the fsstack patch on eCryptfs (commit
0cc72dc7f050188d8d7344b1dd688cbc68d3cd30 made on Fri Dec 8 02:36:31 2006
-0800) needs to be yanked in 2.6.20.I think the simplest solution, from eCryptfs' perspective, is to just
remove the inode size copy.Remove inode size copy in general fsstack attr copy code. Stacked
filesystems may need to interpolate the inode size, since the file
size in the lower file may be different than the file size in the
stacked layer.Signed-off-by: Michael Halcrow
Acked-by: Josef "Jeff" Sipek
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add proper prototypes for sysv_{init,destroy}_icache() in sysv.h
Signed-off-by: Adrian Bunk
Acked-by: Christoph Hellwig
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
22 Dec, 2006
3 commits
-
XFS appears to call clear_page_dirty to get the mapping tree dirty tag
set correctly at the same time the page dirty flag is cleared. I note
that this can be done by set_page_writeback() if we clear the dirty flag
on the page first when we are writing back the entire page.Hence it seems to me that the XFS call to clear_page_dirty() could
easily be substituted by clear_page_dirty_for_io() followed by a call to
set_page_writeback() to get the mapping tree tags set correctly after
the page has been marked clean.Signed-off-by: Linus Torvalds
-
The use by FUSE was just a remnant of an optimization from the time
when writable mappings were supported.Now FUSE never actually allows the creation of dirty pages, so this
invocation of clear_page_dirty() is effectively a no-op.Signed-off-by: Miklos Szeredi
Signed-off-by: Linus Torvalds -
This patch removes some questionable code that attempted to make a
no-longer-used page easier to reclaim.Calling metapage_writepage against such a page will not result in any
I/O being performed, so removing this code shouldn't be a big deal.[ It's likely that we could have just replaced the "clear_page_dirty()"
call with a call to "cancel_dirty_page()" instead, but in the
meantime this is cleaner and simpler anyway, so unless there is some
overriding reason (and Dave implies there isn't) I'll just use this
patch as-is. - Linus ]Signed-off-by: Dave Kleikamp
Signed-off-by: Linus Torvalds
