linux/fs, branch v2.6.26.2

vfs: fix lookup on deleted directory

2008-08-06T16:03:23Z

commit d70b67c8bc72ee23b55381bd6a884f4796692f77 upstream Lookup can install a child dentry for a deleted directory. This keeps the directory dentry alive, and the inode pinned in the cache and on disk, even after all external references have gone away. This isn't a big problem normally, since memory pressure or umount will clear out the directory dentry and its children, releasing the inode. But for UBIFS this causes problems because its orphan area can overflow. Fix this by returning ENOENT for all lookups on a S_DEAD directory before creating a child dentry. Thanks to Zoltan Sogor for noticing this while testing UBIFS, and Artem for the excellent analysis of the problem and testing. Reported-by: Artem Bityutskiy Tested-by: Artem Bityutskiy Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro Signed-off-by: Greg Kroah-Hartman

jbd: fix race between free buffer and commit transaction

2008-08-06T16:03:17Z

commit 3f31fddfa26b7594b44ff2b34f9a04ba409e0f91 upstream journal_try_to_free_buffers() could race with jbd commit transaction when the later is holding the buffer reference while waiting for the data buffer to flush to disk. If the caller of journal_try_to_free_buffers() request tries hard to release the buffers, it will treat the failure as error and return back to the caller. We have seen the directo IO failed due to this race. Some of the caller of releasepage() also expecting the buffer to be dropped when passed with GFP_KERNEL mask to the releasepage()->journal_try_to_free_buffers(). With this patch, if the caller is passing the __GFP_WAIT and __GFP_FS to indicating this call could wait, in case of try_to_free_buffers() failed, let's waiting for journal_commit_transaction() to finish commit the current committing transaction, then try to free those buffers again. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Mingming Cao Reviewed-by: Badari Pulavarty Acked-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

NFS: Ensure we zap only the access and acl caches when setting new acls

2008-08-06T16:03:16Z

commit f41f741838480aeaa3a189cff6e210503cf9c42d upstream ...and ensure that we obey the NFS_INO_INVALID_ACL flag when retrieving the acls. Signed-off-by: Trond Myklebust Signed-off-by: Greg Kroah-Hartman

romfs_readpage: don't report errors for pages beyond i_size

2008-08-06T16:03:09Z

commit 0056e65f9e28d83ee1a3fb4f7d0041e838f03c34 upstream We zero-fill them like we are supposed to, and that's all fine. It's only an error if the 'romfs_copyfrom()' routine isn't able to fill the data that is supposed to be there. Most of the patch is really just re-organizing the code a bit, and using separate variables for the error value and for how much of the page we actually filled from the filesystem. Reported-and-tested-by: Chris Fester Cc: Alexander Viro Cc: Matt Waddel Cc: Greg Ungerer Signed-of-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

VFS: increase pseudo-filesystem block size to PAGE_SIZE

2008-08-01T19:43:12Z

commit 3971e1a917548977cff71418a7c3575ffbc9571f upstream This commit: commit ba52de123d454b57369f291348266d86f4b35070 Author: Theodore Ts'o Date: Wed Sep 27 01:50:49 2006 -0700 [PATCH] inode-diet: Eliminate i_blksize from the inode structure caused the block size used by pseudo-filesystems to decrease from PAGE_SIZE to 1024 leading to a doubling of the number of context switches during a kernbench run. Signed-off-by: Alex Nixon Cc: Andi Kleen Cc: Jeremy Fitzhardinge Cc: Peter Zijlstra Cc: Ingo Molnar Cc: Ian Campbell Cc: "Theodore Ts'o" Cc: Alexander Viro Cc: Hugh Dickins Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

eCryptfs: use page_alloc not kmalloc to get a page of memory

2008-08-01T19:43:10Z

commit 7fcba054373d5dfc43d26e243a5c9b92069972ee upstream Date: Mon, 28 Jul 2008 15:46:39 -0700 Subject: eCryptfs: use page_alloc not kmalloc to get a page of memory With SLUB debugging turned on in 2.6.26, I was getting memory corruption when testing eCryptfs. The root cause turned out to be that eCryptfs was doing kmalloc(PAGE_CACHE_SIZE); virt_to_page() and treating that as a nice page-aligned chunk of memory. But at least with SLUB debugging on, this is not always true, and the page we get from virt_to_page does not necessarily match the PAGE_CACHE_SIZE worth of memory we got from kmalloc. My simple testcase was 2 loops doing "rm -f fileX; cp /tmp/fileX ." for 2 different multi-megabyte files. With this change I no longer see the corruption. Signed-off-by: Eric Sandeen Acked-by: Michael Halcrow Acked-by: Rik van Riel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

proc: fix /proc/*/pagemap some more

2008-08-01T19:43:04Z

commit ee1e6ab6056a8b9c19377257002da98b83819531 upstream struct pagemap_walk was placed on stack, some hooks are initialized, the rest (->pgd_entry, ->pud_entry, ->pte_entry) are valid but junk. Reported-by: Eric Sesterhenn Signed-off-by: Alexey Dobriyan Cc: "Vegard Nossum" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

isofs: fix minor filesystem corruption

2008-08-01T19:42:58Z

commit c0a1633b6201ef79e31b7da464d44fdf5953054d upstream Some iso9660 images contain files with rockridge data that is either incorrect or incompletely parsed. Prior to commit f2966632a134e865db3c819346a1dc7d96e05309 ("[PATCH] rock: handle directory overflows") (included with kernel 2.6.13) the kernel ignored the rockridge data for these files, while still allowing the files to be accessed under their non-rockridge names. That commit inadvertently changed things so that files with invalid rockridge data could not be accessed at all. (I ran across the problem when comparing some old CDs with hard disk copies I had made long ago under kernel 2.4: a few of the files on the hard disk copies were no longer visible on the CDs.) This change reverts to the pre-2.6.13 behavior. Signed-off-by: Adam Greenblatt Reviewed-by: Pekka Enberg Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

quota: fix possible infinite loop in quota code

2008-08-01T19:42:58Z

commit b48d380541f634663b71766005838edbb7261685 upstream When quota structure is going to be dropped and it is dirty, quota code tries to write it. If the write fails for some reason (e. g. transaction cannot be started because the journal is aborted), we try writing again and again and again... Fix the problem by clearing the dirty bit even if the write failed. (akpm: for 2.6.27, 2.6.26.x and 2.6.25.x) Signed-off-by: Jan Kara Reviewed-by: dingdinghua Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

cifs: fix wksidarr declaration to be big-endian friendly

2008-07-12T21:33:42Z

The current definition of wksidarr works fine on little endian arches (since cpu_to_le32 is a no-op there), but on big-endian arches, it fails to compile with this error: error: braced-group within expression allowed only inside a function The problem is that this static declaration has cpu_to_le32 embedded within it, and that expands into a function macro. We need to use __constant_cpu_to_le32() instead. Signed-off-by: Jeff Layton Cc: Steven French Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds