Overhaul struct address_space.assoc_mapping renaming it to
address_space.private_data and its type is redefined to void*. By this
approach we consistently name the .private_* elements from struct
address_space as well as allow extended usage for address_space
association with other data structures through ->private_data.
Also, all users of old ->assoc_mapping element are converted to reflect
its new name and type change (->private_data).
Signed-off-by: Rafael Aquini <aquini@redhat.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Memory fragmentation introduced by ballooning might reduce significantly
the number of 2MB contiguous memory blocks that can be used within a
guest, thus imposing performance penalties associated with the reduced
number of transparent huge pages that could be used by the guest workload.
This patch-set follows the main idea discussed at 2012 LSFMMS session:
"Ballooning for transparent huge pages" -- http://lwn.net/Articles/490114/
to introduce the required changes to the virtio_balloon driver, as well as
the changes to the core compaction & migration bits, in order to make
those subsystems aware of ballooned pages and allow memory balloon pages
become movable within a guest, thus avoiding the aforementioned
fragmentation issue
Following are numbers that prove this patch benefits on allowing
compaction to be more effective at memory ballooned guests.
Results for STRESS-HIGHALLOC benchmark, from Mel Gorman's mmtests suite,
running on a 4gB RAM KVM guest which was ballooning 512mB RAM in 64mB
chunks, at every minute (inflating/deflating), while test was running:
===BEGIN stress-highalloc
STRESS-HIGHALLOC
highalloc-3.7 highalloc-3.7
rc4-clean rc4-patch
Pass 1 55.00 ( 0.00%) 62.00 ( 7.00%)
Pass 2 54.00 ( 0.00%) 62.00 ( 8.00%)
while Rested 75.00 ( 0.00%) 80.00 ( 5.00%)
MMTests Statistics: duration
3.7 3.7
rc4-clean rc4-patch
User 1207.59 1207.46
System 1300.55 1299.61
Elapsed 2273.72 2157.06
MMTests Statistics: vmstat
3.7 3.7
rc4-clean rc4-patch
Page Ins 3581516 2374368
Page Outs 11148692 10410332
Swap Ins 80 47
Swap Outs 3641 476
Direct pages scanned 37978 33826
Kswapd pages scanned 1828245 1342869
Kswapd pages reclaimed 1710236 1304099
Direct pages reclaimed 32207 31005
Kswapd efficiency 93% 97%
Kswapd velocity 804.077 622.546
Direct efficiency 84% 91%
Direct velocity 16.703 15.682
Percentage direct scans 2% 2%
Page writes by reclaim 79252 9704
Page writes file 75611 9228
Page writes anon 3641 476
Page reclaim immediate 16764 11014
Page rescued immediate 0 0
Slabs scanned 2171904 2152448
Direct inode steals 385 2261
Kswapd inode steals 659137 609670
Kswapd skipped wait 1 69
THP fault alloc 546 631
THP collapse alloc 361 339
THP splits 259 263
THP fault fallback 98 50
THP collapse fail 20 17
Compaction stalls 747 499
Compaction success 244 145
Compaction failures 503 354
Compaction pages moved 370888 474837
Compaction move failure 77378 65259
===END stress-highalloc
This patch:
Introduce MIGRATEPAGE_SUCCESS as the default return code for
address_space_operations.migratepage() method and documents the expected
return code for the same method in failure cases.
Signed-off-by: Rafael Aquini <aquini@redhat.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Update the hugetlb_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Chris Metcalf <cmetcalf@tilera.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There was some desire in large applications using MAP_HUGETLB or
SHM_HUGETLB to use 1GB huge pages on some mappings, and stay with 2MB on
others. This is useful together with NUMA policy: use 2MB interleaving
on some mappings, but 1GB on local mappings.
This patch extends the IPC/SHM syscall interfaces slightly to allow
specifying the page size.
It borrows some upper bits in the existing flag arguments and allows
encoding the log of the desired page size in addition to the *_HUGETLB
flag. When 0 is specified the default size is used, this makes the
change fully compatible.
Extending the internal hugetlb code to handle this is straight forward.
Instead of a single mount it just keeps an array of them and selects the
right mount based on the specified page size. When no page size is
specified it uses the mount of the default page size.
The change is not visible in /proc/mounts because internal mounts don't
appear there. It also has very little overhead: the additional mounts
just consume a super block, but not more memory when not used.
I also exported the new flags to the user headers (they were previously
under __KERNEL__). Right now only symbols for x86 and some other
architecture for 1GB and 2MB are defined. The interface should already
work for all other architectures though. Only architectures that define
multiple hugetlb sizes actually need it (that is currently x86, tile,
powerpc). However tile and powerpc have user configurable hugetlb
sizes, so it's not easy to add defines. A program on those
architectures would need to query sysfs and use the appropiate log2.
[akpm@linux-foundation.org: cleanups]
[rientjes@google.com: fix build]
[akpm@linux-foundation.org: checkpatch fixes]
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Acked-by: Rik van Riel <riel@redhat.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Hillf Danton <dhillf@gmail.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There is no reason to pass the nr_pages_dirtied argument, because
nr_pages_dirtied value from the caller is unused in
balance_dirty_pages_ratelimited_nr().
Signed-off-by: Namjae Jeon <linkinjeon@gmail.com>
Signed-off-by: Vivek Trivedi <vtrivedi018@gmail.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Here's the big tty/serial tree set of changes for 3.8-rc1.
Contained in here is a bunch more reworks of the tty port layer from Jiri and
bugfixes from Alan, along with a number of other tty and serial driver updates
by the various driver authors.
Also, Jiri has been coerced^Wconvinced to be the co-maintainer of the TTY
layer, which is much appreciated by me.
All of these have been in the linux-next tree for a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
iEYEABECAAYFAlDHhgwACgkQMUfUDdst+ynI6wCcC+YeBwncnoWHvwLAJOwAZpUL
bysAn28o780/lOsTzp3P1Qcjvo69nldo
=hN/g
-----END PGP SIGNATURE-----
Merge tag 'tty-3.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Pull TTY/Serial merge from Greg Kroah-Hartman:
"Here's the big tty/serial tree set of changes for 3.8-rc1.
Contained in here is a bunch more reworks of the tty port layer from
Jiri and bugfixes from Alan, along with a number of other tty and
serial driver updates by the various driver authors.
Also, Jiri has been coerced^Wconvinced to be the co-maintainer of the
TTY layer, which is much appreciated by me.
All of these have been in the linux-next tree for a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"
Fixed up some trivial conflicts in the staging tree, due to the fwserial
driver having come in both ways (but fixed up a bit in the serial tree),
and the ioctl handling in the dgrp driver having been done slightly
differently (staging tree got that one right, and removed both
TIOCGSOFTCAR and TIOCSSOFTCAR).
* tag 'tty-3.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty: (146 commits)
staging: sb105x: fix potential NULL pointer dereference in mp_chars_in_buffer()
staging/fwserial: Remove superfluous free
staging/fwserial: Use WARN_ONCE when port table is corrupted
staging/fwserial: Destruct embedded tty_port on teardown
staging/fwserial: Fix build breakage when !CONFIG_BUG
staging: fwserial: Add TTY-over-Firewire serial driver
drivers/tty/serial/serial_core.c: clean up HIGH_BITS_OFFSET usage
staging: dgrp: dgrp_tty.c: Audit the return values of get/put_user()
staging: dgrp: dgrp_tty.c: Remove the TIOCSSOFTCAR ioctl handler from dgrp driver
serial: ifx6x60: Add modem power off function in the platform reboot process
serial: mxs-auart: unmap the scatter list before we copy the data
serial: mxs-auart: disable the Receive Timeout Interrupt when DMA is enabled
serial: max310x: Setup missing "can_sleep" field for GPIO
tty/serial: fix ifx6x60.c declaration warning
serial: samsung: add devicetree properties for non-Exynos SoCs
serial: samsung: fix potential soft lockup during uart write
tty: vt: Remove redundant null check before kfree.
tty/8250 Add check for pci_ioremap_bar failure
tty/8250 Add support for Commtech's Fastcom Async-335 and Fastcom Async-PCIe cards
tty/8250 Add XR17D15x devices to the exar_handle_irq override
...
Here's the large driver core updates for 3.8-rc1.
The biggest thing here is the various __dev* marking removals. This is
going to be a pain for the merge with different subsystem trees, I know,
but all of the patches included here have been ACKed by their various
subsystem maintainers, as they wanted them to go through here.
If this is too much of a pain, I can pull all of them out of this tree
and just send you one with the other fixes/updates and then, after
3.8-rc1 is out, do the rest of the removals to ensure we catch them all,
it's up to you. The merges should all be trivial, and Stephen has been
doing them all in linux-next for a few weeks now quite easily.
Other than the __dev* marking removals, there's nothing major here, some
firmware loading updates and other minor things in the driver core.
All of these have (much to Stephen's annoyance), been in linux-next for
a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
iEYEABECAAYFAlDHkPkACgkQMUfUDdst+ykaWgCfW7AM30cv0nzoVO08ax6KjlG1
KVYAn3z/KYazvp4B6LMvrW9y0G34Wmad
=yvVr
-----END PGP SIGNATURE-----
Merge tag 'driver-core-3.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core updates from Greg Kroah-Hartman:
"Here's the large driver core updates for 3.8-rc1.
The biggest thing here is the various __dev* marking removals. This
is going to be a pain for the merge with different subsystem trees, I
know, but all of the patches included here have been ACKed by their
various subsystem maintainers, as they wanted them to go through here.
If this is too much of a pain, I can pull all of them out of this tree
and just send you one with the other fixes/updates and then, after
3.8-rc1 is out, do the rest of the removals to ensure we catch them
all, it's up to you. The merges should all be trivial, and Stephen
has been doing them all in linux-next for a few weeks now quite
easily.
Other than the __dev* marking removals, there's nothing major here,
some firmware loading updates and other minor things in the driver
core.
All of these have (much to Stephen's annoyance), been in linux-next
for a while.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"
Fixed up trivial conflicts in drivers/gpio/gpio-{em,stmpe}.c due to gpio
update.
* tag 'driver-core-3.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (93 commits)
modpost.c: Stop checking __dev* section mismatches
init.h: Remove __dev* sections from the kernel
acpi: remove use of __devinit
PCI: Remove __dev* markings
PCI: Always build setup-bus when PCI is enabled
PCI: Move pci_uevent into pci-driver.c
PCI: Remove CONFIG_HOTPLUG ifdefs
unicore32/PCI: Remove CONFIG_HOTPLUG ifdefs
sh/PCI: Remove CONFIG_HOTPLUG ifdefs
powerpc/PCI: Remove CONFIG_HOTPLUG ifdefs
mips/PCI: Remove CONFIG_HOTPLUG ifdefs
microblaze/PCI: Remove CONFIG_HOTPLUG ifdefs
dma: remove use of __devinit
dma: remove use of __devexit_p
firewire: remove use of __devinitdata
firewire: remove use of __devinit
leds: remove use of __devexit
leds: remove use of __devinit
leds: remove use of __devexit_p
mmc: remove use of __devexit
...
If we have a read oplock and set a read lock in it, we can't write to the
locked area - so, filemap_fdatawrite may fail with a no information for a
userspace application even if we request a write to non-locked area. Fix
this by populating the page cache without marking affected pages dirty
after a successful write directly to the server.
Also remove CONFIG_CIFS_SMB2 ifdefs because it's suitable for both CIFS
and SMB2 protocols.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
This should fix a regression that was introduced when the new mount
option parser went in. Also, when the unc= and prefixpath= options
are provided, check their values against the ones we parsed from
the device string. If they differ, then throw a warning that tells
the user that we're using the values from the unc= option for now,
but that that will change in 3.10.
Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Currently the code takes care to ensure that the prefixpath has a
leading '/' delimiter. What if someone passes us a prefixpath with a
leading '\\' instead? The code doesn't properly handle that currently
AFAICS.
Let's just change the code to skip over any leading delimiter character
when copying the prepath. Then, fix up the users of the prepath option
to prefix it with the correct delimiter when they use it.
Also, there's no need to limit the length of the prefixpath to 1k. If
the server can handle it, why bother forbidding it?
Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Make sure we free any existing memory allocated for vol->UNC, just in
case someone passes in multiple unc= options.
Get rid of the check for too long a UNC. The check for >300 bytes seems
arbitrary. We later copy this into the tcon->treeName, for instance and
it's a lot shorter than 300 bytes.
Eliminate an extra kmalloc and copy as well. Just set the vol->UNC
directly with the contents of match_strdup.
Establish that the UNC should be stored with '\\' delimiters. Use
convert_delimiter to change it in place in the vol->UNC.
Finally, move the check for a malformed UNC into
cifs_parse_mount_options so we can catch that situation earlier.
Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The authority fields are supposed to be represented by a single 48-bit
value. It's also supposed to represent the value as hex if it's equal to
or greater than 2^32. This is documented in MS-DTYP, section 2.4.2.1.
Also, fix up the max string length to account for this fix.
Acked-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Not all architectures (in particular, sparc64) have empty_zero_page.
So instead of copying from empty_zero_page, use memset to clear the
inline data by signalling to ext4_xattr_set_entry() via a magic
pointer value, EXT4_ZERO_ATTR_VALUE, which is defined by casting -1 to
a pointer.
This fixes a build failure on sparc64, and the memset() should be more
efficient than using memcpy() anyway.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Flags being used by atomic operations in inode flags (e.g.
ext4_test_inode_flag(), should be consistent with that actually stored
in inodes, i.e.: EXT4_XXX_FL.
It ensures that this consistency is checked at build-time, not at
run-time.
Currently, the flags consistency are being checked at run-time, but,
there is no real reason to not do a build-time check instead of a
run-time check. The code is comparing macro defined values with enum
type variables, where both are constants, so, there is no problem in
comparing constants at build-time.
enum variables are treated as constants by the C compiler, according
to the C99 specs (see www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf
sec. 6.2.5, item 16), so, there is no real problem in comparing an
enumeration type at build time
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Ted has sent out a RFC about removing this feature. Eric and Jan
confirmed that both RedHat and SUSE enable this feature in all their
product. David also said that "As far as I know, it's enabled in all
Android kernels that use ext4." So it seems OK for us.
And what's more, as inline data depends its implementation on xattr,
and to be frank, I don't run any test again inline data enabled while
xattr disabled. So I think we should add inline data and remove this
config option in the same release.
[ The savings if you disable CONFIG_EXT4_FS_XATTR is only 27k, which
isn't much in the grand scheme of things. Since no one seems to be
testing this configuration except for some automated compile farms, on
balance we are better removing this config option, and so that it is
effectively always enabled. -- tytso ]
Cc: David Brown <davidb@codeaurora.org>
Cc: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We use kzalloc() to allocate sbi, no need to zero its field.
Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
inode_init_always() will initialize inode->i_data.writeback_index
anyway, no need to do this in ext4_alloc_inode().
Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
We have a dedicated interface to sync inode metadata. Use it to
simplify ext4's code some.
Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
If we are punching hole in a file, we will return ENOTSUPP.
As for the fallocation of some extents, we will convert the
inline data to a normal extent based file first.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Now we that store data in the inode, in case we need to store some
xattrs and inode doesn't have enough space, Andreas suggested that we
should keep the xattr(metadata) in and data should be pushed out. So
this patch does the work.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
fiemap is used to find the disk layout of a file, as for inline data,
let us just pretend like a file with just one extent.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In case we rename a directory, ext4_rename has to read the dir block
and change its dotdot's information. The old ext4_rename encapsulated
the dir_block read into itself. So this patch adds a new function
ext4_get_first_dir_block() which gets the dir buffer information so
the ext4_rename can handle it properly. As it will also change the
parent inode number, we return the parent_de so that ext4_rename() can
handle it more easily.
ext4_find_entry is also changed so that the caller(rename) can tell
whether the found entry is an inlined one or not and journaling the
corresponding buffer head.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
empty_dir is used when deleting a dir. So it should handle inline dir
properly.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Currently ext4_delete_entry() is used only for dir entry removing from
a dir block. So let us create a new function
ext4_generic_delete_entry and this function takes a entry_buf and a
buf_size so that it can be used for inline data.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Create a new function ext4_find_inline_entry() to handle the case of
inline data.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
search_dirblock is used to search a dir block, but the code is almost
the same for searching an inline dir.
So create a new fuction search_dir and let search_dirblock call it.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
For "." and "..", we just call filldir by ourselves
instead of iterating the real dir entry.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch let add_dir_entry handle the inline data case. So the
dir is initialized as inline dir first and then we can try to add
some files to it, when the inline space can't hold all the entries,
a dir block will be created and the dir entry will be moved to it.
Also for an inlined dir, "." and ".." are removed and we only use
4 bytes to store the parent inode number. These 2 entries will be
added when we convert an inline dir to a block-based one.
[ Folded in patch from Dan Carpenter to remove an unused variable. ]
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The old add_dirent_to_buf handles all the work related to the
work of adding dir entry to a dir block. Now we have inline data,
so create 2 new function __ext4_find_dest_de and __ext4_insert_dentry
that do the real work and let add_dirent_to_buf call them.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The __ext4_check_dir_entry() function() is used to check whether the
de is over the block boundary. Now with inline data, it could be
within the block boundary while exceeds the inode size. So check this
function to check the overflow more precisely.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Currently, the initialization of dot and dotdot are encapsulated in
ext4_mkdir and also bond with dir_block. So create a new function
named ext4_init_new_dir and the initialization is moved to
ext4_init_dot_dotdot. Now it will called either in the normal non-inline
case(rec_len of ".." will cover the whole block) or when we converting an
inline dir to a block(rec len of ".." will be the real length). The start
of the next entry is also returned for inline dir usage.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
For delayed allocation mode, we write to inline data if the file
is small enough. And in case of we write to some offset larger
than the inline size, the 1st page is dirtied, so that
ext4_da_writepages can handle the conversion. When the 1st page
is initialized with blocks, the inline part is removed.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
For a normal write case (not journalled write, not delayed
allocation), we write to the inline if the file is small and convert
it to an extent based file when the write is larger than the max
inline size.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Let readpage and readpages handle the case when we want to read an
inlined file.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Implement inline data with xattr.
Now we use "system.data" to store xattr, and the xattr will
be extended if the i_size is increased while we don't release
the space during truncate.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
SMB2 and later will return only 1 credit for session setup (phase 1)
not just for the negotiate protocol response. Do not disable
echoes and oplocks on session setup (we only need one credit
for tree connection anyway) as a resonse with only 1 credit
on phase 1 of sessionsetup is expected.
Fixes the "CIFS VFS: disabling echoes and oplocks" message
logged to dmesg.
Signed-off-by: Steve French <smfrench@gmail.com>
Acked-by: Jeff Layton <jlayton@samba.org>
Restructure code to make SMB2 vs. SMB3 signing a protocol
specific op. SMB3 signing (AES_CMAC) is not enabled yet,
but this restructuring at least makes sure we don't send
an smb2 signature on an smb3 signed connection. A followon
patch will add AES_CMAC and enable smb3 signing.
Signed-off-by: Steve French <smfrench@gmail.com>
Acked-by: Jeff Layton <jlayton@samba.org>
A SID could potentially be embedded inside of payload.value if there are
no subauthorities, and the arch has 8 byte pointers. Allow for that
possibility there.
While we're at it, rephrase the "embedding" check in terms of
key->payload to allow for the possibility that the union might change
size in the future.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
It was hardcoded to 192 bytes, which was not enough when the max number
of subauthorities went to 15. Redefine this constant in terms of sizeof
the structs involved, and rename it for better clarity.
While we're at it, remove a couple more unused constants from cifsacl.h.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Now that we aren't so rigid about the length of the key being passed
in, we need to be a bit more rigorous about checking the length of
the actual data against the claimed length (a'la num_subauths field).
Check for the case where userspace sends us a seemingly valid key
with a num_subauths field that goes beyond the end of the array. If
that happens, return -EIO and invalidate the key.
Also change the other places where we check for malformed keys in this
code to invalidate the key as well.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The cifs.idmap keytype always allocates memory to hold the payload from
userspace. In the common case where we're translating a SID to a UID or
GID, we're allocating memory to hold something that's less than or equal
to the size of a pointer.
When the payload is the same size as a pointer or smaller, just store
it in the payload.value union member instead. That saves us an extra
allocation on the sid_to_id upcall.
Note that we have to take extra care to check the datalen when we
go to dereference the .data pointer in the union, but the callers
now check that as a matter of course anyway.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The cifs.idmap handling code currently causes the kernel to cache the
data from userspace twice. It first looks in a rbtree to see if there is
a matching entry for the given id. If there isn't then it calls
request_key which then checks its cache and then calls out to userland
if it doesn't have one. If the userland program establishes a mapping
and downcalls with that info, it then gets cached in the keyring and in
this rbtree.
Aside from the double memory usage and the performance penalty in doing
all of these extra copies, there are some nasty bugs in here too. The
code declares four rbtrees and spinlocks to protect them, but only seems
to use two of them. The upshot is that the same tree is used to hold
(eg) uid:sid and sid:uid mappings. The comparitors aren't equipped to
deal with that.
I think we'd be best off to remove a layer of caching in this code. If
this was originally done for performance reasons, then that really seems
like a premature optimization.
This patch does that -- it removes the rbtrees and the locks that
protect them and simply has the code do a request_key call on each call
into sid_to_id and id_to_sid. This greatly simplifies this code and
should roughly halve the memory utilization from using the idmapping
code.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The direct-IO write path already had the i_size checks in mm/filemap.c,
but it turns out the read path did not, and removing the block size
checks in fs/block_dev.c (commit bbec0270bdd8: "blkdev_max_block: make
private to fs/buffer.c") removed the magic "shrink IO to past the end of
the device" code there.
Fix it by truncating the IO to the size of the block device, like the
write path already does.
NOTE! I suspect the write path would be *much* better off doing it this
way in fs/block_dev.c, rather than hidden deep in mm/filemap.c. The
mm/filemap.c code is extremely hard to follow, and has various
conditionals on the target being a block device (ie the flag passed in
to 'generic_write_checks()', along with a conditional update of the
inode timestamp etc).
It is also quite possible that we should treat this whole block device
size as a "s_maxbytes" issue, and try to make the logic even more
generic. However, in the meantime this is the fairly minimal targeted
fix.
Noted by Milan Broz thanks to a regression test for the cryptsetup
reencrypt tool.
Reported-and-tested-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
by using cifs_invalidate_mapping rather than invalidate_remote_inode
in cifs_oplock_break - this invalidates all inode pages and resets
fscache cookies.
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
We don't need to permit a write to the area locked with a read lock
by any process including the process that issues the write.
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
I've legally changed my name with New York State, the US Social Security
Administration, et al. This patch propagates the name change and change
in initials and login to comments in the kernel source as well.
Signed-off-by: Nadia Yvette Chambers <nyc@holomorphy.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
The caller doesn't do anything with the dentry, so there's no point in
holding a reference to it on return. Also cifs_prime_dcache better
describes the actual purpose of the function.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
This can reduce the size of the module by ~120KB which
could be useful for embedded systems.
$ size fs/cifs/built-in.o*
text data bss dec hex filename
388567 34459 100440 523466 7fcca fs/cifs/built-in.o.new
495970 34599 117904 648473 9e519 fs/cifs/built-in.o.old
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Make the compilation work again when CIFS_DEBUG is not #define'd.
Add format and argument verification for the various macros when
CIFS_DEBUG is not #define'd.
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
We were checking incorrectly if signatures were required to be sent,
so were always sending signatures after the initial session establishment.
For SMB3 mounts (vers=3.0) this was a problem because we were putting
SMB2 signatures in SMB3 requests which would cause access denied
on mount (the tree connection would fail).
This might also be worth considering for stable (for 3.7), as the
error message on mount (access denied) is confusing to users and
there is no workaround if the server is configured to only
support smb3.0. I am ok either way.
CC: stable <stable@kernel.org>
Signed-off-by: Steve French <smfrench@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
It uses an undefined KERN_EVENT and is itself unused.
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Currently, the code relies on the callers to do that and they all do,
but this will ensure that it's always done.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Now that the smb_vol contains the destination sockaddr, there's no need
to pass it in separately.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Take advantage of accelerated strchr() on arches that support it.
Also, no caller ever passes in a NULL pointer. Get rid of the unneeded
NULL pointer check.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Passing this around as a string is contorted and painful. Instead, just
convert these to a sockaddr as soon as possible, since that's how we're
going to work with it later anyway.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Otherwise, "ls -l" will simply show the ownership of the files as
the default mnt_uid/gid. This may make "ls -l" performance on large
directories super-suck in some cases, but that's the cost of cifsacl.
One possibility to make it suck less would be to somehow proactively
dispatch the ACL requests asynchronously from readdir codepath, but
that's non-trivial to implement.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The option to have a blank "pass=" already exists, and with
a password specified both "pass=%s" and "password=%s" are supported.
Also, both blank "user=" and "username=" are supported, making
"password=" the odd man out.
Signed-off-by: Jesper Nilsson <jesper.nilsson@axis.com>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
This patch enables optional for original SMB2 (SMB2.02) dialect
by specifying vers=2.0 on mount.
Reviewed-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
If we netogiate mandatory locking style, have a read lock and try
to set a write lock we end up with a write lock in vfs cache and
no lock in cifs lock cache - that's wrong. Fix it by returning
from cifs_setlk immediately if a error occurs during setting a lock.
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
that reacquires byte-range locks when a file is reopened.
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
because the is no difference here. This also adds support of prefixpath
mount option for SMB2.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Error out with a clear error message if there is no unc= option. The
existing code doesn't handle this in a clear fashion, and the check for
a UNCip option with no UNC string is just plain wrong.
Later, we'll fix the code to not require a unc= option, but for now we
need this to at least clarify why people are getting errors about DFS
parsing. With this change we can also get rid of some later NULL pointer
checks since we know the UNC and UNCip will never be NULL there.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
If we're using cifsacl, then we don't want to override the uid/gid with
the current uid/gid, since that would prevent you from being able to
upcall for this info.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
...and make those symbols static in cifsacl.c. Nothing outside
of that file refers to them.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The format specifiers are for signed values, but these are unsigned.
Given that '-' is a delimiter between fields, I don't think you'd get
what you'd expect if you got a value here that would overflow the sign
bit.
The version and authority fields are 8 bit values so use a "hh" length
modifier there. The subauths are 32 bit values, so there's no need to
use a "l" length modifier there.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
According to several places on the Internet and the samba winbind code,
this is hard limited to 15 in windows, not 5. This does balloon out
the allocation of each by 40 bytes, but I don't see any alternative.
Also, rename it to SID_MAX_SUB_AUTHORITIES to match the alleged name
of this constant in the windows header files
Finally, rename SIDLEN to SID_STRING_MAX, fix the value to reflect
the change to SID_MAX_SUB_AUTHORITIES and document how it was
determined.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
...and lift the restriction in id_to_sid upcall that the size must be
at least as big as a full cifs_sid.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
..nothing outside of cifsacl.c calls it. Also fix the incorrect
comment on the function. It returns 0 when they match.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
...instead of hardcoding in '5' and '6' all over the place.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Add a label we can goto on error, and get rid of some excess indentation.
Also move to kernel-style comments.
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Most of these are unsigned ints, so we should be passing "uint" to
module_param. Also, get rid of the extra "(bool)" in the description
of enable_oplocks.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
We had planned to upgrade to ntlmv2 security a few releases ago,
and have been warning users in dmesg on mount about the impending
upgrade, but had to make a change (to use nltmssp with ntlmv2) due
to testing issues with some non-Windows, non-Samba servers.
The approach in this patch is simpler than earlier patches,
and changes the default authentication mechanism to ntlmv2
password hashes (encapsulated in ntlmssp) from ntlm (ntlm is
too weak for current use and ntlmv2 has been broadly
supported for many, many years).
Signed-off-by: Steve French <smfrench@gmail.com>
Acked-by: Jeff Layton <jlayton@redhat.com>
READ is zero so the "rw & READ" test is always false. The intended test
was "((rw & RW_MASK) == READ)".
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The inline data feature will need some inline xattr functions, so
export them from fs/ext4/xattr.c so that inline.c can use them.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The block device access simplification that avoided accessing the (racy)
block size information (commit bbec0270bdd8: "blkdev_max_block: make
private to fs/buffer.c") no longer checks the maximum block size in the
block mapping path.
That was _almost_ as simple as just removing the code entirely, because
the readers and writers all check the size of the device anyway, so
under normal circumstances it "just worked".
However, the block size may be such that the end of the device may
straddle one single buffer_head. At which point we may still want to
access the end of the device, but the buffer we use to access it
partially extends past the end.
The 'bd_set_size()' function intentionally sets the block size to avoid
this, but mounting the device - or setting the block size by hand to
some other value - can modify that block size.
So instead, teach 'submit_bh()' about the special case of the buffer
head straddling the end of the device, and turning such an access into a
smaller IO access, avoiding the problem.
This, btw, also means that unlike before, we can now access the whole
device regardless of device block size setting. So now, even if the
device size is only 512-byte aligned, we can read and write even the
last sector even when having a much bigger block size for accessing the
rest of the device.
So with this, we could now get rid of the 'bd_set_size()' block size
code entirely - resulting in faster IO for the common case - but that
would be a separate patch.
Reported-and-tested-by: Romain Francoise <romain@orebokech.com>
Reporeted-and-tested-by: Meelis Roos <mroos@linux.ee>
Reported-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge 'block-dev' branch.
I was going to just mark everything here for stable and leave it to the
3.8 merge window, but having decided on doing another -rc, I migth as
well merge it now.
This removes the bd_block_size_semaphore semaphore that was added in
this release to fix a race condition between block size changes and
block IO, and replaces it with atomicity guaratees in fs/buffer.c
instead, along with simplifying fs/block-dev.c.
This removes more lines than it adds, makes the code generally simpler,
and avoids the latency/rt issues that the block size semaphore
introduced for mount.
I'm not happy with the timing, but it wouldn't be much better doing this
during the merge window and then having some delayed back-port of it
into stable.
* block-dev:
blkdev_max_block: make private to fs/buffer.c
direct-io: don't read inode->i_blkbits multiple times
blockdev: remove bd_block_size_semaphore again
fs/buffer.c: make block-size be per-page and protected by the page lock
Not a bug as such, just warning noise from the xlog_cksum()
returning a __be32 type when it should be returning a __le32 type.
On Wed, Nov 28, 2012 at 08:30:59AM -0500, Christoph Hellwig wrote:
> But why are we storing the crc field little endian while all other on
> disk formats are big endian? (And yes I realize it might as well have
> been me who did that back in the idea, but I still have no idea why)
Because the CRC always returns the calcuation LE format, even on BE
systems. So rather than always having to byte swap it everywhere and
have all the force casts and anootations for sparse, it seems simpler to
just make it a __le32 everywhere....
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Ben Myers <bpm@sgi.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>
Currently, in ext4_iget we do a simple check to see whether
there does exist some information starting from the end
of i_extra_size. With inline data added, this procedure
is more complicated. So move it to a new function named
ext4_iget_extra_inode.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Pull vfs fixes from Al Viro:
"A bunch of fixes; the last one is this cycle regression, the rest are
-stable fodder."
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
fix off-by-one in argument passed by iterate_fd() to callbacks
lookup_one_len: don't accept . and ..
cifs: get rid of blind d_drop() in readdir
nfs_lookup_revalidate(): fix a leak
don't do blind d_drop() in nfs_prime_dcache()
Pull CIFS fixes from Steve French:
"Two low risk, small fixes, that fix cifs regressions introduced in
3.7."
* 'for-linus' of git://git.samba.org/sfrench/cifs-2.6:
CIFS: Fix wrong buffer pointer usage in smb_set_file_info
cifs: fix writeback race with file that is growing
Noticed by Pavel Roskin; the thing in his patch I disagree with
was compensating for that shite in callbacks instead of fixing
it once in the iterator itself.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
We are leaking fattr and fhandle if we decide that dentry is not to
be invalidated, after all (e.g. happens to be a mountpoint). Just
free both before that...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Commit fa77dcfafe introduces block bitmap checksum calculation into
ext4_new_inode() in the case that block group was uninitialized.
However we brelse() the bitmap buffer before we attempt to checksum it
so we have no guarantee that the buffer is still there.
Fix this by releasing the buffer after the possible checksum
computation.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Acked-by: Darrick J. Wong <darrick.wong@oracle.com>
Cc: stable@vger.kernel.org
Remove a level of indentation by moving the DIO read and extending
write case to the beginning of the file. This results in no actual
programmatic changes to the file, but makes it easier to
read/understand.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We really don't want to look at the block size for the raw block device
accesses in fs/block-dev.c, because it may be changing from under us.
So get rid of the max_block logic entirely, since the caller should
already have done it anyway.
That leaves the only user of this function in fs/buffer.c, so move the
whole function there and make it static.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since directio can work on a raw block device, and the block size of the
device can change under it, we need to do the same thing that
fs/buffer.c now does: read the block size a single time, using
ACCESS_ONCE().
Reading it multiple times can get different results, which will then
confuse the code because it actually encodes the i_blksize in
relationship to the underlying logical blocksize.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When we fail to get a dquot lock during reclaim, we jump to an error
handler that unlocks the dquot. This is wrong as we didn't lock the
dquot, and unlocking it means who-ever is holding the lock has had
it silently taken away, and hence it results in a lock imbalance.
Found by inspection while modifying the code for the numa-lru
patchset. This fixes a random hang I've been seeing on xfstest 232
for the past several months.
cc: <stable@vger.kernel.org>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ben Myers <bpm@sgi.com>