vol: Let non-usable volumes attach for non-DAFS

Before DAFS, volumes that were not inService/blessed were not
accessible by normal clients, but were still allowed to attach. That
is, access to clients was prevented at VGetVolume-time, rather than at
attach-time. Commit 939382c5 tried to short-circuit this by detecting
this volume state at attach-time. However, volume utilities (e.g.
volserver) can give us back a volume over FSSYNC when they are done
with the volume, and for non-DAFS, we then try to attach the volume.
So, with 939382c5 that attachment will fail when volserver gives us
back a volume that is not inService/blessed (which can happen for some
normal volume operations).

This situation is not terrible, since either way the volume is not
usable by clients (since the volume didn't attach), and the volume is
still usable by volserver (since volserver is allowed to check out
nonexistent volumes). But it is a deviation from pre-DAFS behavior and
it can result in confusing error messages, so revert the 939382c5
behavior for non-DAFS.

For DAFS, this behavior is fine, since the fileserver does not attach
a volume unless it is trying to service a client request. So, leave it
for DAFS.

FIXES 131505

Reviewed-on: http://gerrit.openafs.org/8557
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 988f66c5c3bd4c4f23bfa53e69a52ed52c4e3ab9)

Change-Id: I08b55b30047fc19788e2c83ada1ecd80456f1f74
Reviewed-on: http://gerrit.openafs.org/9499
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

vol: avoid attach retry loops on SYNC_FAILED

Avoid unnecessary volume attach retry loops if a volume
salvage is unable to be scheduled due to a hard error.

Reviewed-on: http://gerrit.openafs.org/7566
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Tom Keiser <tkeiser@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit dda5cc0c5f7172aa4edf58c0f876c8bf05590889)

Change-Id: I93760c1bd337a76079bf03dcba1d9c4e32cd692e
Reviewed-on: http://gerrit.openafs.org/9483
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

vol: Avoid getting stuck in ATTACHING in attach2

Since commit 5fc2365f, a VNOVOL error early in attach2 meant that we
skipped changing the volume state to anything, and just returned
instead. When we do this, the volume is in VOL_STATE_ATTACHING for
DAFS, and so if we return, the volume will forever be in
VOL_STATE_ATTACHING. The next thing that tries to access the volume
will wait forever for the volume to come out of that state.

So, revert half of 5fc2365f, and transition to ERROR state instead.
This code path should not be hit during normal usage, since a
nonexistant volume access for the fileserver will be detected earlier.
If the volume does not appear to exist at this stage of attachment,
something is wrong with the volume, so this warrants the ERROR state.

For the volserver and other volume utilities, we may hit this when a
request just plain references a nonexistant volume for whatever
reason, but in that case the vp should go away soon. For non-DAFS,
this commit does not change much, since the difference between
error_notbroken and unlocked_error is very small.

The other half of 5fc2365f is not changed, since it is correct. For
VOFFLINE errors at this point, the volume has already been
transitioned to VOL_STATE_UNATTACHED, so it is okay to return. Add a
comment to help make this more explicit.

Reviewed-on: http://gerrit.openafs.org/7516
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 4f9c3f2be9abd5548dd52888b49c407d574b6557)

Change-Id: I8e2b9d2a114f8aea283145ca44bf8ae52c354908
Reviewed-on: http://gerrit.openafs.org/9482
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Do not expose afs_assert.h from other public headers

afs_assert.h redefines the standard assert macro, which is evil and
breaks some applications that might want to include our public headers
(e.g. some versions of Cython).  This was fixed on master by commit
cac74242728ad97e3ce9cef0a949d58c237250f6, which removes afs_assert.h
entirely and adds opr_Assert.  Since that patch may be too invasive
for 1.6.x, here’s a minimal patch that just stops exposing
afs_assert.h from our other public headers.

Change-Id: I39a7b9ae8d43cfe0059e10e47ce4b1c22e01c544
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/10096
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

ihandle: don't keep reallyclosing future fds

given that we can mark something invalid for future use, ever,
once we have done so for all fds, we ih_reallyclose is done.
don't persist the setting to the detriment of new fds

Reviewed-on: http://gerrit.openafs.org/6101
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 2332ee3aa8e21738ed4aa42b53edd4e2ae0a2984)

Change-Id: Ic77354893101acda867aa8e4c8ea7a2932666d5b
Reviewed-on: http://gerrit.openafs.org/9490
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

DAFS: Preattach, not attach, in FSYNC_Drop

FSYNC_Drop currently attaches volumes that were checked out by the
dropped fssync handler, but not checked back in, in order to make the
volume available again. For DAFS, however, a full attachment is
unnecessary; just preattach instead.

Reviewed-on: http://gerrit.openafs.org/7515
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit a43f0e4717fd72727ae325eb4437cef065419cdd)

Change-Id: Iaa2320808c4855516192b164de9778c9019a4559
Reviewed-on: http://gerrit.openafs.org/9481
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Eliminate pointless changes between nbsd50/60 params

param.nbsd60.h removed AFS_64BIT_ENV and added an empty expression
to an #if.

AFS_64BIT_ENV is required for any platform to build (on 1.6)
This is a 1.6-only change. On master, AFS_64BIT_ENV was removed entirely
in commit fc9aa428.

Change-Id: I57d7e2a2ef5ab4bf787a99083d35ba70e240710c
Reviewed-on: http://gerrit.openafs.org/10137
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Tested-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

auth: avoid excessive stat of cellservdb

The modified time resolution is one second, so only
stat the cellservdb file at most only once per second.

Reviewed-on: http://gerrit.openafs.org/5743
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jeffrey Altman <jaltman@secure-endpoints.com>
(cherry picked from commit 0e3bfa033ed230fcb46ad8e3c26c8b7aae6e00af)

Change-Id: Ia533d037cc772c085cfaad47035e2a29df64952e
Reviewed-on: http://gerrit.openafs.org/9435
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Do not use separate array for srvAddrs

The array of srvAddr structs we use in afs_LoopServers have indices
unrelated to the indices of conns, rxconns, etc. Several places were
assuming that addr[i] corresponded to conn[i], which is not
necessarily true. So instead, do not use the separate addr array
(except when populating the conn and rxconn arrays), and just get the
srvAddr structure by going through the relevant conn[i].

Reviewed-on: http://gerrit.openafs.org/5790
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit f199ac666195771a02e3ebb040c6e5fe47c58c58)

Change-Id: I70be3c518d2b1ccd51e050532d966a27cf22090f
Reviewed-on: http://gerrit.openafs.org/9434
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

libafs: initialize hard mount last errors

Initialize the values of the server last errors
introduced in commit 94a8ce970d57498583e249ea61725fce1ee53a50
to avoid logging garbage for the last error codes.

Reviewed-on: http://gerrit.openafs.org/10082
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 495d6fcbba1630b136d9c1b4256f0a71666a8a7b)

Change-Id: I4402838d761ab4bbb3e0798f219dfb825347b090
Reviewed-on: http://gerrit.openafs.org/10085
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

viced: Restrict RXAFS_FlushCPS to administrators

RXAFS_FlushCPS currently can be run by anyone, including
unauthenticated users. Forcing CPS calculation can be a relatively
resource-intensive operation, though, if done frequently enough, and
only should need to be done by administrators. Thus, only let
administrators use it.

Reviewed-on: http://gerrit.openafs.org/7572
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 568adf7d18eb17a42caa263aabc92a686f0ae121)

Change-Id: I715e7ede7ea92be65a134116ecb4d1b7e2ccd264
Reviewed-on: http://gerrit.openafs.org/9485
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

viced: Do error translation for InlineBulkStatus

When a host has requested universal errors, error code conversion
is performed in the CallPostamble. However, the InlineBulkStatus
errorcodes are passed as part of the data set, not as RX errors,
so this translation is not performed.

Fix this so that we also translate error codes that are part of
the InlineBulkStatus response.

Reviewed-on: http://gerrit.openafs.org/6992
Reviewed-by: Jeffrey Altman <jaltman@secure-endpoints.com>
Tested-by: Jeffrey Altman <jaltman@secure-endpoints.com>
(cherry picked from commit 9498cebd11ac26ed7986b5496440159f7ef745f2)

Change-Id: Ia5b2d95e05bed0fb89aeb6132008d58de68c96a0
Reviewed-on: http://gerrit.openafs.org/9467
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

viced: Enable NAT ping on hosts

Turn on NAT ping on the Rx connection for the callback channel for
hosts. This should help improve behavior for clients behind NATs and
stateful firewalls, even for clients that predate NAT ping
functionality.

Reviewed-on: http://gerrit.openafs.org/4646
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit aafdc08cfc49da4c23ecd91f9e690fd70e95df55)

Change-Id: I428b6648276ec49fd4f003b3cf2d88a07c8aa1d9
Reviewed-on: http://gerrit.openafs.org/9420
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

dafs: prevent corruption in large fsstate.dat files

If while writing to the fsstate.dat file, it exceeds the current
size of the file (multiples of FS_STATE_INIT_FILESIZE (8MiB)),
we call fs_stateResizeFile.  This un-mmaps, truncates, and
re-mmaps the file.  Unfortunately, fs_stateMapFile() resets the
state->mmap.offset and .cursor, so any writes in flight over
the resize will overwrite the first bytes of the file (and leave
zeros in the file where the data should have been written).

Upon return from the write that caused a file resize, the offset
is eventually corrected and the state dump continues with a
silent failure.  Eventually the state dump completes and the
file header is rewritten; this may conceal some or all of
the overwrite damage at offset 0.  However, any zeros near the 8MiB
offset (and its multiples) remain as corruption.

Add a flag to fs_stateMapFile() to allow the caller to specify if
the offset and cursor should be preserved.  Modify fs_stateResizeFile()
to use this capability.

testing note: temporarily reduced FS_STATE_INIT_FILESIZE to 256 bytes
during testing in order to make the problem easier to reproduce.
This problem would normally occur only on relatively large/active
DAFS fileservers.

Reviewed-on: http://gerrit.openafs.org/9599
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 42351db75e04e46012208d38f80dc17be7ab776d)

Change-Id: I05a01ebce15e26534361863ed0f8d3161c3f8240
Reviewed-on: http://gerrit.openafs.org/9683
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

fileserver: Fix NeverAttach support

Commit 35becabed870d4bfe49abaa499d99a3ffb0a2d31 added support for
the /vicepXX/NeverAttach. However this code only appears to work on
Linux. It fails build testing on (at least) Mac OS X, FreeBSD, and AIX.
Modify the code so that the NeverAttach call uses the same variable to
locate the path of the partition as the AlwaysAttach call does.

Reviewed-on: http://gerrit.openafs.org/7125
Reviewed-by: Jeffrey Altman <jaltman@secure-endpoints.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 709a6358e1d1f6cc0b025522e1e595c50bb2b733)

Change-Id: I2f142084fdb4e66c3e7841c9ac961cab417633a0
Reviewed-on: http://gerrit.openafs.org/9471
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: Stephan Wiesand <stephan.wiesand@desy.de>

Fileserver: Add the /vicepXX/NeverAttach flag to skip mounting a partition

FIXES 130561

LICENSE MIT

Reviewed-on: http://gerrit.openafs.org/7108
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: Jason Edgecombe <jason@rampaginggeek.com>
(cherry picked from commit 35becabed870d4bfe49abaa499d99a3ffb0a2d31)

Change-Id: If14e0145404dcb81fb5d2184c0991fa3ca64a1df
Reviewed-on: http://gerrit.openafs.org/9470
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: Stephan Wiesand <stephan.wiesand@desy.de>

libafs: fs flushall for unix cm

Implement the fs flushall command on the unix cache manager to flush
all volume data.  Uses a new common pioctl code point VIOC_FLUSHALL (14),
registered with the grand.central.org assigned numbers.

Reviewed-on: http://gerrit.openafs.org/7540
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 4197bbecd9d0b2ff0b8eaec75a0df9a64f713cf0)

Change-Id: Ia4e9eb070d279bf05dbe38ea9b99f1f2f6b9c6f0
Reviewed-on: http://gerrit.openafs.org/9390
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

libafs: avoid flushing already freed dcaches in flush volumes

When flushing volume data, skip flushing dcache entries which
are already freed.

Reviewed-on: http://gerrit.openafs.org/7539
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 82865811e49aae8b7b8ff9f5d995dc32518b9eb8)

Change-Id: I88c9940748dcf7d179fcc2327cddcf222314560d
Reviewed-on: http://gerrit.openafs.org/9389
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

libafs: use afs_ResetVCache in flush volume data

Remove some code duplication by using afs_ResetVCache
in the flush volume data pioctl. Adds a flag to
ResetVCache to avoid unneeded calls to purge dnlc
when reseting all the vcaches in a volume.

Adds freeing of vcache link data in the flush volume
data pioctl.

Reviewed-on: http://gerrit.openafs.org/7538
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 36c6c46bd081e0278611c3bb6ae25dd765f10648)

Change-Id: I0c4a001c20633a404c299359914f2368b4492901
Reviewed-on: http://gerrit.openafs.org/9388
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

Cleanup VOffline log message for non-DAFS

Commit fd592c7674d4aa44dda90998b54d7b56947f6ed8 fixed the 'Volume X
(Y) is now offline' message for DAFS, but the same problem persists
for non-DAFS. Fix the non-DAFS case.

Reviewed-on: http://gerrit.openafs.org/3213
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 138ac6577eab40e25a796f7c18f3c39add000ed2)

Change-Id: I3e0c53d93bcaad43a9fd7b20ccf32bc216601455
Reviewed-on: http://gerrit.openafs.org/9412
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

volser: Do not reset copyDate in ReClone

When we ReClone in the volserver, do not reset the clone's copyDate to
the current time. If we retain the copyDate between ReClone
operations, then we can know when the clone was first created (and
thus makes local RO clones more consistent with remote RO sites).

Reviewed-on: http://gerrit.openafs.org/3892
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 112b9b35fc701ca26ba7f99c31a73a7a1eecbdc0)

Change-Id: Ifd84a161125bd14799fab2f75f321ca164ef39d2
Reviewed-on: http://gerrit.openafs.org/9451
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

namei: force-close fd on read or write error

if we errored, just reallyclose so we'll get a new fd

Reviewed-on: http://gerrit.openafs.org/6126
Reviewed-by: Jeffrey Altman <jaltman@secure-endpoints.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 730832073a1e87654f28bfff054e24d4170c01dc)

Change-Id: I019b38b6670d3a45b774740d324d5d49e23f94d1
Reviewed-on: http://gerrit.openafs.org/9436
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

vlserver: Use correct base value when replacing

When we're removing existing address entries the code calculates
a base and index value for each entry that we're removing an address
from. However, it then _uses_ a previously calculated base value,
with the new index. This works fine if the old base and the new base
match, but if they don't, chaos will ensue.

Fix to always use matched base and index pairs.

Reviewed-on: http://gerrit.openafs.org/4771
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit a180b0c5d8991a14f1fbd78d81dcb4754a24f4d3)

Change-Id: I37372f020ad644f30b6dbdd1ed68b56eb1c0f2e2
Reviewed-on: http://gerrit.openafs.org/9429
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

vlserver: Access cache via vl_ctx

The vlserver application-level ubik cache (which consists of
HostAddress, ex_addr, and cheader) is currently being accessed via
global variables everywhere. Instead, access these via the new vl_ctx
struct that is passed to functions during a transaction, so we have
the ability to modify the cache without making all changes visible as
we change it.

Reviewed-on: http://gerrit.openafs.org/2105
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 1f532d099b8b084d43dd0140890448464325b602)

Change-Id: I161b6c41e104af045941663f78148037313036ee
Reviewed-on: http://gerrit.openafs.org/9422
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

vlserver: Add a struct for trans-specific data

Instead of passing a ubik_trans pointer to many functions inside the
vlserver, pass a vlserver-defined vl_ctx struct, so we can add new
things to keep track of in a transaction that are not part of ubik.

Reviewed-on: http://gerrit.openafs.org/2104
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 3bf0350218dcf4d6e086941fac394769b6b9cc4a)

Change-Id: Ic224f1fab7505b486baa0a41e5481bc615f00826
Reviewed-on: http://gerrit.openafs.org/9421
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

viced: h_SetupCallbackConn_r in removeAddress_r

removeAddress_r has some code that is identical to the
h_SetupCallbackConn_r function. Call the function instead.

Reviewed-on: http://gerrit.openafs.org/4645
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 0376255747ac17ed41afbc219f5a5978e58d8e9c)

Change-Id: I9f7b4fbc6059a15b30d675a7c78ac9ff00ba32b0
Reviewed-on: http://gerrit.openafs.org/9419
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

LINUX: Avoid multiple d_invalidate loops

Currently, in afs_linux_lookup, we put an artificial limit on how many
times we loop through all dentry aliases, trying ti d_invalidate all
of them. Instead of using an arbitrary limit, we can just go through
all of them once, by using d_prune_aliases. This should be faster, and
removes some of the logic required here.

Note that this does remove our check for DCACHE_DISCONNECTED in each
alias' d_flags.  This should not be a problem, since we will still use
any remaining DCACHE_DISCONNECTED dentry via d_splice_alias if one
still exists.

Reviewed-on: http://gerrit.openafs.org/8751
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 370aaaeafa43f804b0a5286d92b4ec5f1ccb62be)

Change-Id: I1aa70afe8268852c676f241e0189bc010ad757aa
Reviewed-on: http://gerrit.openafs.org/9288
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

DAFS: VRS_r with VOL_SALVAGE_NO_OFFLINE in attach2

One caller of VRequestSalvage_r in attach2 was not passing the
VOL_SALVAGE_NO_OFFLINE flag. This really should be passed for every
place that manually sets vp->nUsers = 0, since then the VPutVolume_r
handlers will never fire.

Reviewed-on: http://gerrit.openafs.org/8203
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 3237543afafae7191635fd8e8cada1f95e8abcff)

Change-Id: I44a9a56b849bc3c1473b290961190a47ad45f5b5
Reviewed-on: http://gerrit.openafs.org/9283
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

db_dump.o depends on budb_errs.h

Add INCLS and budb_errs.h accordingly.

In contrast to master, budb did not have this already.

Reviewed-on: http://gerrit.openafs.org/9996
Reviewed-by: Jeffrey Hutzelman <jhutz@cmu.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 7af86ecffb737b050f67908b63f247e8361eacdc)

Change-Id: Id8657c39d58a1f73e07bf5d9a5324cb343b3888b
(cherry picked from commit e6d1db4643d023eda837c8ae37b82b0dc7d5ee4d)
Reviewed-on: http://gerrit.openafs.org/10005
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

volume_inline.h: Down with assert, again

Commit 34767c6a0f914960c9a1efabe69dd9c312a2b400 replaced all assert
calls in this file with osi_Assert, but shortly thereafter, commit
db6ee95864a8fc5f33b7e95c19c8ff5058d37e92 added VTimedWaitStateChange_r
with two new assert calls.  These are precarious in a public header;
fix them to osi_Assert like the ones in VWaitStateChange_r.

Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/10094
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 30fa9480dd99ed93fa642dd8ce9746760fb42180)

Change-Id: Id0bc0e75de000cf3e4133aaf31f52d9a565c8d9f
Reviewed-on: http://gerrit.openafs.org/10095
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Handle easy GetValidDSlot errors

Many callers of GetValidDSlot currently assume they will always get
back a valid dcache, and will panic on getting NULL. However, for many
of these callers, handling the NULL case is quite easy, since the
failure to get a dcache can just result in an error directly, or
obtaining the dcache is best-effort or just an optimization.

This commit just handles the "easy" cases; some other callers require
more complex handling.

Reviewed-on: http://gerrit.openafs.org/8375
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit f74a0a7bbb37a8ab6050e833cf8d66abdff31854)

Change-Id: Ic2c463edebcb821562541004bd4181483c90c3e6
Reviewed-on: http://gerrit.openafs.org/9287
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Make last_error always useful

Currently we record last_error as the last getuerror() we got when
failing to read in a slot in UFSGetDSlot. For kernels that do not have
getuerror(), this variable is currently useless, and we do not record
anywhere what the last error received was (besides logging it via
afs_warn).

So, for non-uerror, just record what 'code' we got, so we at least
have something.

Reviewed-on: http://gerrit.openafs.org/8369
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 48b60478897e7cb22e9cc1adfa3b71c188572093)

Change-Id: I32a079934c8d3ecd3d7316cfc7eb8da01580a08e
Reviewed-on: http://gerrit.openafs.org/9286
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

salvager: Trust inode-based special data over OGM

Currently the salvaging code looks for special inodes, and infers the
volume id and inode type from the OGM data in each special inode file.
However, we can already derive this information from the inode number
itself for the special inode, so if they disagree, use the values
based off of the inode number and correct the OGM data.

The inode number should be more likely to be correct, since that is
how we look up the special inode from the header when attaching the
volume. It is also impossible to get special inode files with the same
name, so this ensures we don't get duplicates. And for people that go
snooping around /vicepX/AFSIDat even though we tell them not to, it
seems more likely that they go around 'chmod'ing or 'chown'ing rather
than 'mv'ing.

This change avoids an abort in the salvaging code when the OGM data is
wrong. If we trust the OGM data when it is incorrect, we assume the
special inode file is for a different volume. So when we go to
recreate one of the special files for the volume we're actually
working with, the IH_CREATE fails (from EEXIST) and so we abort.

Reviewed-on: http://gerrit.openafs.org/6946
Tested-by: Derrick Brashear <shadow@dementix.org>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit df522b588396aaac3ec662b516ef7287f3b7c47b)

Change-Id: I7ae826b6d6747e3ab3a1d8086ff17503a904a6d5
Reviewed-on: http://gerrit.openafs.org/9282
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

namei: Abstract out OGM functions a bit more

Add GetWinOGM and SetWinOGM for getting and setting the
Windows-equivalent of the Unix OGM data. Make those and CheckOGM use
GetFileTime/SetFileTime so we can operate just via an FD_t, without
needing the full pathname. Modify the NT namei_icreate to use
SetWinOGM.

Reviewed-on: http://gerrit.openafs.org/6945
Reviewed-by: Jeffrey Altman <jaltman@secure-endpoints.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 0594fc80b48152b4b03edda7e19133b59fc0670e)

Change-Id: I65491a9a82911d9c99d6238a4ea66ed400db7aa0
Reviewed-on: http://gerrit.openafs.org/9281
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

budb: Do not use garbage cellinfo

If the -servers option is given, we never initialize cellinfo or the
clones array. So, don't give the cellinfo structure or the clones
array to ubik in that case, or we may crash or do other weird things.

This issue appears to have been introduced in commit fc4ab52e.

FIXES 131706

Reviewed-on: http://gerrit.openafs.org/10115
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 6b0016dd992b3215cb17c93812e14de9e4916d25)

Change-Id: I7174d3ac596fb9ecf512aa159fdc839714f8791c
Reviewed-on: http://gerrit.openafs.org/10166
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

xstat: length check cm call info

Define the cm xstat function call counters with an xmacro to avoid
duplicating the list of cm function names.  This obviates the need
to update xstat_cm_test.c when new function names are added to the
cm xstat collection id 0.

Check the number of returned records when printing the function call
counts to avoid over-running when a newer xstat_cm_test client
receives data from an older cm.

Reviewed-on: http://gerrit.openafs.org/8741
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 09c0484fd8878797957f7ff5936c542a0f6332c4)

Change-Id: I622a4f16cbb102962199f26e5431b04ea381d5fe
Reviewed-on: http://gerrit.openafs.org/9065
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Avoid tracking file locks for RO volumes

Advisory file locks for RO volumes don't make a lot of sense, since
there are no possible writes to worry about. The fileserver already
does not track these, so don't even bother processing them in the
client.

Reviewed-on: http://gerrit.openafs.org/8197
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 0fc27471e7da0c5de4addcdec1bfbca5208072cc)

Change-Id: I308d930c94708fd8d9025886c10f40cc4d10d195
Reviewed-on: http://gerrit.openafs.org/8910
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

libafscp: NULL return value not pointer

When afscp_FindCallBack sets its return-by-reference value to NULL,
it should do so to the value itself, not the pointer to it.

Caught by coverity (#988419)

Reviewed-on: http://gerrit.openafs.org/9266
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 99929488511d13dab94491cac1d477062f9da741)

Change-Id: I046189684cef6c88c46455a90ab49d03da1f1cab
Reviewed-on: http://gerrit.openafs.org/9359
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

libafscp: Actually return callback from FindCallback

Fix FindCallback so that it actually returns the callback that it
found. This requires changing the function prototype so that the
third parameter is passed by reference, and updating the single
call site.

Caught by clang-analyzer

Reviewed-on: http://gerrit.openafs.org/9198
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit f67e4a3cdfb1856e31dda0d4741c8d34df13f8fc)

Change-Id: Ia610993535acc68beef7e8cffcfac77db2638aa4
Reviewed-on: http://gerrit.openafs.org/9973
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

dafs: preattach should wait for exclusive states

In rare circumstances an FSYNC_VOL_ON operation may fail silently,
leaving the volume in its previous state.  The only clue is a FileLog
message "volume <nnnn> not in quiescent state".

This is caused by a race condition in the volume package: an
FSYNC_VOL_ON operation is attempting to preattach a volume
(in VPreAttachVolumeByVp_r()) at the same time a fileserver RPC
(e.g. FetchStatus) is detaching the volume (in VReleaseVolumeHandles_r())
at the conclusion of attach2() logic.

The fix calls VWaitExclusiveState_r() before calling
VPreAttachVolumeByVp_r().

Change-Id: Ib66859381d29311fda3e08984dcb740eadafb340
Reviewed-on: http://gerrit.openafs.org/8814
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 1f891b622e9b32a068082087eae9d787057f7f00)
Reviewed-on: http://gerrit.openafs.org/9070
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

do not ship uncompressed tarballs

There's not much point to shipping uncompressed tarballs in 2013. Delete
them during "make-release".

Reviewed-on: http://gerrit.openafs.org/10116
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Ken Dreyer <ktdreyer@ktdreyer.com>
(cherry picked from commit 0346ed6f53a6f3fb5d75de41c7ec5a139727618f)

Change-Id: Ic1247c6a42ecb6a5c6bae1fd88b789a8b3e55d49
Reviewed-on: http://gerrit.openafs.org/10131
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Enable dvolser and tvolser for FreeBSD

The top-level Makefile did not recurse into dvolser or tvolser for
FreeBSD.  They compile just fine, so enable them.

Reviewed-on: http://gerrit.openafs.org/9598
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit e93360ecc79822dd8149990a6eb5f3235b7cca1f)

Change-Id: Ia7dfbce23453a147f2c1ba81718373f632a039b6
Reviewed-on: http://gerrit.openafs.org/10122
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

build tsalvaged, tvolser, and dvolser targets on *nbsd*

Reviewed-on: http://gerrit.openafs.org/5595
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 1ca207f18491147bb7f7de3e588fb1979ac3d0c5)

Change-Id: I8570370905fa4c3bbdd72f5535329cfab5bebf1a
Reviewed-on: http://gerrit.openafs.org/10121
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afscp: Fix -s option for writes

When writing to AFS with afscp, the -s option was sleeping before any
StoreData RPCs actually got issued to the fileserver. Move the sleep
to after we have done one rx_Read/rx_Write, so we sleep after starting
to contact the fileserver, to make sleeping while writing more
consistent with sleeping while reading.

Reviewed-on: http://gerrit.openafs.org/3762
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 380cc22a45dff7e40cd50d57a56742b59ccc4952)

Change-Id: I5bafad6958111bfb196adde79361939b00a176a1
Reviewed-on: http://gerrit.openafs.org/9417
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afscp: Add -s option

Add an -s option to afscp, to specify an amount of time to sleep in
the middle of a read or write operation. This can be helpful in
simulating a slow client.

Reviewed-on: http://gerrit.openafs.org/3291
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 00023f168e0e097c6a7e9da4450a009e4818a1a5)

Change-Id: Ifefce12d7372f9f8fe7149b91af0d0a7f128cb96
Reviewed-on: http://gerrit.openafs.org/9416
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

doc: clarify setcrypt defaults

Reviewed-on: http://gerrit.openafs.org/10111
Reviewed-by: Jason Edgecombe <jason@rampaginggeek.com>
Reviewed-by: Michael Laß <lass@mail.uni-paderborn.de>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 3a24ed9baef178d34b6d0fb3a5f6a485dea9353b)

Change-Id: I2d9cd3c9167e2d67ced609e5be2d173a82b36a28
Reviewed-on: http://gerrit.openafs.org/10136
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Ken Dreyer <ktdreyer@ktdreyer.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Use -nofork when starting bosserver via systemd

Systemd does not expect the started process to fork unless
"Type=forking" is given. Use -nofork to run BOS in foreground and allow
systemd to track its state.

Reviewed-on: http://gerrit.openafs.org/10087
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Tested-by: Michael Laß <lass@mail.uni-paderborn.de>
Tested-by: Ken Dreyer <ktdreyer@ktdreyer.com>
Reviewed-by: Ken Dreyer <ktdreyer@ktdreyer.com>
(cherry picked from commit e2d458c11956af6fe721f7151487cb19f07ac16f)

Change-Id: I2b66ca126dbda6c2c616d74b571908c57d1e86e4
Reviewed-on: http://gerrit.openafs.org/10093
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Ken Dreyer <ktdreyer@ktdreyer.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Avoid SetupVolume panic

Currently SetupVolume panics if it cannot successfully read a
volumeinfo entry from disk. Try to return an error instead, so we
don't panic the machine.

Reviewed-on: http://gerrit.openafs.org/9094
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 6f7ae535bbac2a5376358801b7f2c9e072f2d141)

Change-Id: Ib8ea06192bfcd6c2111444db325abc4a90190bbc
Reviewed-on: http://gerrit.openafs.org/9131
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Move SetupVolume tv initialization after loop

The fields in tv are not used by the loop looking for the given volume
on disk. If we wait until after that loop to initialize the fields in
tv, it is easier to handle errors encountered in the loop.

This should incur no functional change.

Reviewed-on: http://gerrit.openafs.org/9093
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 8f95dc9eb92cb31f9d29eb87daac747f53b5a1cc)

Change-Id: I65f3b647017aebacf28026a648c75b2d279c768e
Reviewed-on: http://gerrit.openafs.org/9130
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Check dv against localhero aincr

For operations that modify directories, we call afs_LocalHero to
determine if we can perform the directory modification in our local
cache, and avoid fetching the dir blob from the fileserver. Currently,
afs_LocalHero assumes that the DV received from the fileserver is
correct, and will update the cache DV as long as we have a valid
callback on the file.

If for any reason the client cache falls out of sync with what's on
the fileserver, this can cause the client to incorrectly believe its
cache is up to date. Since, the cached data will be marked with the
newest DV, even if the DV on the server has jumped to be larger than
we expected.

While the client cache should never fall out of sync with the
fileserver, in the past this has been possible due to other bugs
(fileserver idle dead processing and client VNOSERVICE handling).
Assuming that the given DV is correct is also just unnecesarily
fragile, since we can always check if it is correct, so just check it,
and add some comments helping explain what's going on here. Note that
regular file writes effectively already check this.

Note that this change makes use of the 'aincr' argument to
afs_LocalHero, which was previously unused. aincr appears to have been
used for a purpose similar to this before OpenAFS 1.0, but was
removed, possibly accidentally.

It is possible this change negatively affects, or even breaks
(unlikely), functionality with the AFS<->DFS translator. Although
nothing of the sort has been seen, it is difficult to know one way or
the other, due to the lack of available DFS translators.

Reviewed-on: http://gerrit.openafs.org/8864
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 004e9f25634d76be58b5a60dfc3c9a058d8bcd26)

Change-Id: Ie0a3ab5defc5dde3284ad7527fdd05835a85da7f
Reviewed-on: http://gerrit.openafs.org/8898
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Linux 3.11: Convert from readdir to iterate file operation

Convert the readdir function so that it can be used as the new
"iterate" file operation.  This new operation is passed a context
that contains a pointer to the filldir function and the offset.
The context is passed into the new dir_emit function that will
call the function specified by the context.

The new dir_emit function returns true on success, so we must be
careful about how we check for failure since this is different
behaviour from what filldir currently does.

Reviewed-on: http://gerrit.openafs.org/10051
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 902b8809f03533ffa7731919930bb5178f77df6c)

Change-Id: I12dc2517af66b66b2de6b2cb01d39e3a0cbc5fe2
Reviewed-on: http://gerrit.openafs.org/10118
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

dirpath: update keytab strings in windows

this was missed in the pullup from 33eecea7db14d06c59e1081b970d4caf0af773ca

Change-Id: Iee734bf06af4152ead5cfcf33bbbcafc39812c5a
Reviewed-on: http://gerrit.openafs.org/10109
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

Merge branch 'rxkad-kdf-1.6' into origin/openafs-stable-1_6_x

Merge the changes necessary for OpenAFS-SA-2013-003 and
OpenAFS-SA-2013-004 back into the stable branch.

ubik: Fix encryption selection in ugen

Make sure that we encrypt when requested to by the application

Change-Id: If4c2ba2257bf060d3e9169ccdbcae54f54dfe5d7

Make OpenAFS 1.6.5

Change-Id: I693297ef6e20358966930cb29116d45b9151811f

Add support for deriving DES keys to klog.krb5

(cherry picked from commit e79102e7918ce5196e870a806879135743ec3abb)

Change-Id: Ia7ebfdd10dcfd6cd164b10275016147630748bac

Reload rxkad.keytab on CellServDB modification

Make the reloading of rxkad.keytab keys occur in the same way that
KeyFile keys are reloaded. That is, we only try to reload them if the
CellServDB mtime has changed. This is intended to have exactly the
same reloading behavior as KeyFile reloads.

I would have triggered this from afsconf_Check, but that approach
has annoyances. (Calling ticket5_keytab functions directly from
cellconfig pulls in libkrb5 dependencies for everything that uses
cellconfig, and we'd have to trigger an afsconf_Check call by calling
some other cellconfig function.)

9102f49a3bdc67ed74e254349eb55b529472f45c

Avoid calling afsconf_GetLatestKey directly

Don't call afsconf_GetLatestKey to determine whether we can print our
own local tokens, since we may have keytab 'local' keys, but no DES
keys. Just try to construct them and see if it fails, using
afsconf_PickClientSecObj or afsconf_ClientAuth{,Secure} as
appropriate.

auth: Do not always fallback to noauth

Make afsconf_PickClientSecObj error out if we can't construct
localauth tokens (unless the caller explicitly requested rxnull
fallback). afsconf_ClientAuth{,Secure} still falls back, as always.

Clean up akimpersonate and use for server-to-server

Since a6d7cacfd, aklog has been able to print a krb5 ticket to
itself for an arbitrary client principal, allowing a user with
access to the cell's krb5 key to get tokens as an arbitrary user.

Now that it is possible to use native krb5 tickets with non-DES
enctypes for authentication, and akimpersonate is available from libauth,
use printed native krb5 tickets for server-to-server communication (as well
as the -localauth versions of the client utilities).

Remove the early call to afsconf_GetLatestKey() in
afsconf_PickClientSecObj() so that we do not end up picking an old DES
key before we try to find a better key to use.

Before doing so, refactor the akimpersonate code to be more usable
and readable, and eliminate some dead code.  For example, we always printed
addressless tickets, so that code could be removed.  Other code had excessive
stack usage for a library routine, which is eliminated.  Use a start time
of 0 instead of 300 so that the printed ticket will always be
detected as infinite-lifetime.

In order to ensure usability on all platforms (in particular Solaris),
provide a couple more compat shims to implement routines which are not
always available from the krb5 library, in particular encode_krb5_ticket
and encode_krb5_enc_tkt_part.  Thanks to Andrew Deason for implementing
these compatability routines.

UKERNEL doesn't need this stuff.

Move akimpersonate to libauth

Give it its own source file and header, install the header at
depinstall time, and have aklog get the akimpersonate functionality
from libauth.

Keep the linux box copyright from aklog_main.c (but strip the trailing
whitespace), as that block was added with the akimpersonate code.

Remove all calls to afs_com_err() as is fitting for library code,
to let it build.  Do not bother removing curly braces which are
no longer needed; a future cleanup commit will catch that.

Derive DES/fcrypt session key from other key types

If a kerberos 5 ticket has a session key with a non-DES enctype,
use the NIST SP800-108 KDF in counter mode with HMAC_MD5 as the PRF to
construct a DES key to be used by rxkad.

To satisfy the requirements of the KDF, DES3 keys are first compressed into a
168 bit form by reversing the RFC3961 random-to-key algorithm

Change-Id: I4dc8e83a641f9892b31c109fb9025251de3dcb27

Integrate keytab-based decryption into afsconf_BuildServerSecurityObjects

Now all servers can have it.

authcon.o grows a krb5 dependency and needs to get KRB5_CPPFLAGS.

Change-Id: I95fecb3f88c19b3d5193ea8200fa20c86ec08ad7

New optional rxkad functionality for decypting krb5 tokens

An additional, optional mechanism for decrypting krb5-format tokens
is provided that uses the krb5 api with a key from a keytab
instead of using libdes and the AFS KeyFile.

The AIX compat stub for krb5_c_decrypt is contributed by Andrew Deason.

Change-Id: I97c08122c60482b84d602d6fa6482f1d5deef142

Add rxkad server hook function to decrypt more types of tokens

Allow tokens to be encrypted with algorithms other than DES.
The security object owner must provide an implementation
by calling rxkad_SetAltDecryptProc.

Make sure plainsiz is initialized before calling the alternate decrypt
proc.

viced: Rename fs_rxstat_userok

The fs_rxstat_userok function can be used for other purposes; give it
a more general name.

Reviewed-on: http://gerrit.openafs.org/7571
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 7a875f639f61966f2d327f929305091d2678e99a)

Change-Id: I44301904cc8e7145ba13276a32b7cea475bc0574
Reviewed-on: http://gerrit.openafs.org/9484
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

make openafs 1.6.4

The wrong commit was tagged for the 1.6.3 release by accident.
Update the version strings to 1.6.4, and adapt the release notes.

Change-Id: I0c4f8e796a1017aebaaf6c219fd12e1547ad9319
Reviewed-on: http://gerrit.openafs.org/9983
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

make openafs 1.6.3

update version strings for 1.6.3, finalize release notes

Change-Id: I2f3e3dee0b1ea9e2fce522ad72e9018825b4103b
Reviewed-on: http://gerrit.openafs.org/9976
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

Update release notes for 1.6.3

As of pre3, we'll support Linux 3.10 clients (at least up to rc3).

Change-Id: I81b783ad563e4a7735638bde2aa77d11ea1ec83c
Reviewed-on: http://gerrit.openafs.org/9959
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

make 1.6.3pre3

prerelease for OpenAFS 1.6.3

Change-Id: I92e3df57e9c8d55759c317cc8d7743c23c349399
Reviewed-on: http://gerrit.openafs.org/9958
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

Linux 3.10: Replace create_proc_entry() with proc_create()

Add an afs_proc_create() compat function that uses the
appropriate kernel function based on a configure test.

Reviewed-on: http://gerrit.openafs.org/9854
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 9b24013426e03a501fcaa6334ba5a9b48a8da3d1)

Change-Id: I976ef345b1638434026e852e577e1f4474171e3d
Reviewed-on: http://gerrit.openafs.org/9948
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Linux 3.10: Include linux/aio.h directly

This file needs bits from linux/aio.h, but gets them indirectly
from the inclusion of aio.h by sched.h.  That gets removed for
Linux 3.10.

Since this header is not of general interest elsewhere, just include
it directly here.

Reviewed-on: http://gerrit.openafs.org/9912
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit fca4252621b80c5126e1ae3d84a54da5ebc677ba)

Change-Id: I7da471ee033087592b9ec3f10331baf0135cc201
Reviewed-on: http://gerrit.openafs.org/9947
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

make 1.6.3pre2

prerelease for OpenAFS 1.6.3

Change-Id: Ie6148239feae53bdaca9409c6b75714a25968c00
Reviewed-on: http://gerrit.openafs.org/9909
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Update release notes for 1.6.3

We merged another change, helping Linux 3.6+ clients. Mention it.

Change-Id: I71d2ee7f2b6fc591d767f37436d0e154809e0786
Reviewed-on: http://gerrit.openafs.org/9910
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

OpenBSD 5.3: Replace use of copyinstr for setting mount point name.

As a result of a realignment of kernel memory in OpenBSD 5.3,
the copyinstr() routine no longer works for copying the mount
point name into the internal mount table structure. It also
fails silently, so it's not noticed until someone looks at
the mount table and discovers that the mount point name for
AFS is missing.

This patch replaces the use of copyinstr() with strlcpy() for
copying the mount point name in OpenBSD 5.3.

Note that this is consistent with how other similar device
support has addressed the same issue in OpenBSD 5.3.

Reviewed-on: http://gerrit.openafs.org/9752
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit b39461fc67d4f258d31e0a13b5d8dcb3c5fae49f)

Change-Id: I5e9a3d0871360e0e3b19cdb2b4a609baeed050f8
Reviewed-on: http://gerrit.openafs.org/9908
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Linux: osi_TryEvictVCache: Don’t skip the first dentry if D_ALIAS_IS_HLIST

An hlist doesn’t begin with a sentinel like a list does, so the old
code would skip the first dentry or crash with a NULL dereference if
there wasn’t one.  Use the kernel’s list_for_each_entry or
hlist_for_each_entry macros instead of trying to do it manually.

Should fix a crash observed by Alex Chernyakhovsky on kernel 3.6 and
newer.

Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/9857
Reviewed-by: Alex Chernyakhovsky <achernya@mit.edu>
Tested-by: Alex Chernyakhovsky <achernya@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
(cherry picked from commit ad6e31d5fea221bf5eefcd35a416667bc7c1f9be)

Change-Id: Id8b31246067b5a718af91c8c206d5abd5210c558
Reviewed-on: http://gerrit.openafs.org/9898
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

make openafs 1.6.3pre1

prerelease for 1.6.3

Change-Id: I8eb600cdc0ed55d15e928a9f165f68d3ff924f82
Reviewed-on: http://gerrit.openafs.org/9858
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Ken Dreyer <ktdreyer@ktdreyer.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Preliminary release notes for 1.6.3

Change-Id: Ie4a6a984cfd9b819b0a3170acce7817998307157
Reviewed-on: http://gerrit.openafs.org/9859
Reviewed-by: Ken Dreyer <ktdreyer@ktdreyer.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

ppc64_linux26: build binaries as 64-bit by default

lwp doesnt work since ucontext/setjmp are incomplete when running 32-bit
binaries with the 64-bit kernel.

Reviewed-on: http://gerrit.openafs.org/5365
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 39b788eeafe27d21ae6b4fd3cfeb59ce5dc70077)

Change-Id: Ic4fd5c1a861db8e96dd58a79d7fb2ea73ab4f0d6
Reviewed-on: http://gerrit.openafs.org/9605
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afsio: Pass KRB5_LDFLAGS along with KRB5_LIBS

We need this in case e.g. the krb5 library we're using is in an
alternate location, so we can pass -L. aklog already does this, so do
it for afsio, too.

This is a 1.6-only change. On master, this problem went away when the
krb5 ldflags/cflags/etc were standardized in commit
6f9443e702745d1e371eb72ad7fcba5cac47f3db.

Thanks to Måns Nilsson for reporting this.

Change-Id: I79aeb1f4de123b7550b64c1890a2773a44ff8d1e
Reviewed-on: http://gerrit.openafs.org/9692
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

doc: quote list items in POD

Recent versions of Pod::Simple complain if we use integers or other
special characters in an =item list. We have a couple bulleted lists
that happen to have integers or other special characters as the list
values. Quote the items with C<> so that Pod::Simple can correctly parse
them again.

Reviewed-on: http://gerrit.openafs.org/9838
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Tested-by: Jeffrey Altman <jaltman@your-file-system.com>
(cherry picked from commit 32d823c52f212fb79816708418ad7d958eb32a01)

Change-Id: Ie3e5c792d64f1d2e973e3a5f68b8466986cde0d7
Reviewed-on: http://gerrit.openafs.org/9842
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jason Edgecombe <jason@rampaginggeek.com>
Tested-by: Jason Edgecombe <jason@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Do not invalidate all dcaches on startup

Commit 20b0c65a289e2b55fb6922c8f60e873f1f4c6f97 changed
afs_UFSGetDSlot to always treat a dslot entry as invalid if
'datavalid' was 0. This was to force the invalidation of the given
dslot if we were reading in a dslot from the free or discard list,
since the data in that dslot is not valid.

However, 'datavalid' is also 0 when we read in dcache entries from
disk on startup. So, this means that we invalidated all cache entries
when the client started up, effectively making our persistent cache
worthless.

Fix this by only forcing this invalidation when we are reading from a
free or discarded dcache, and not during the initial cache scan. That
is, when 'indexvalid' is 1, and 'datavalid' is 0.

The parameters for these Get*DSlot variants should maybe be changed to
be a little more clear, but for now, this is a targeted fix for this
specific issue.

FIXES 131655

Reviewed-on: http://gerrit.openafs.org/9833
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 479c34614f33859131258e330e7a3f8494e0bc80)

Change-Id: I72b7ad41658c51b81aa9844327372724a0f3151e
Reviewed-on: http://gerrit.openafs.org/9837
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Tested-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

Make ihandle sync behavior runtime-configurable

The actual behavior of FDH_SYNC has changed a bit over the years, and
some people want one behavior, and some want another. Make it possible
to make this choice at runtime with the new -sync option, instead of
making this decision by running with different patches.

Note that FDH_SYNC is not a macro anymore, nor is it an inline
function. While it could be a macro, it would look a bit complex, and
there are some oddities with trying to use vol_io_params inside the
FDH_SYNC expansion (vol_io_params is not declared for LWP, for
example). And having it be an inline function causes problems with
some odd linking dependencies. For example, vlib.a contains volume.o,
but does not contain a definition for DFlushVolume (dir/buffer.c),
which is referenced in volume.o.  'vos' uses vlib.a, but does not
bring in anything that defines DFlushVolume. Currently this appears to
not cause a problem because 'vos' uses nothing from volume.o, so the
dependencies of volume.o don't matter. Adding an inline FDH_SYNC for
platforms that don't support 'static inline' would add a dependency to
volume.o (via vol_io_params), which causes an error for the lack of a
DFlushVolume.

Those are possibly just some problems, and may not be all. So instead,
make it so we don't have to deal with that and just have a normal
function. While FDH_SYNC may be called in a performance-critical
section, the overhead of a real function call is nowhere near the
delay of an actual fsync(), so presumably any overhead doesn't matter.

Reviewed-on: http://gerrit.openafs.org/9694
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit eb5190eb4a7cd95166866a89e0a8f3a69bbc6e8f)

Change-Id: I7a6c99d00eff3400423440db91a350053ed099ea
Reviewed-on: http://gerrit.openafs.org/9695
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

LINUX: Sometimes let dentry_open handle refcounts

When Linux changed dentry_open to use a 'path' argument, they also
changed it so dentry_open handles incrementing the relevant ref
counts. So now, sometimes we need to inc the dentry and vfsmount
refcounts ourselves, and sometimes we need to leave them alone.

To accommodate this, change afs_dentry_open to also handle refcounting
itself, and 'get' the given dentry and vfsmount if necessary.

Also note that currently, afs_linux_raw_open can call afs_dentry_open
twice in the case of an error, but it does not dget(dp). This means
that dp could be undercounted, since dentry_open on older kernels will
dec the refcount on the given dentry in the case of an error. This
change should also fix this so dp is not undercounted in that case.

FIXES 131613

Reviewed-on: http://gerrit.openafs.org/9801
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
(cherry picked from commit e31240732cbe449fedea5095037ac08d1d513fa9)

Change-Id: I082063d324d99c3d02ed372a1c20462f13bb4a26
Reviewed-on: http://gerrit.openafs.org/9803
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

LINUX: Avoid duplicate mntget in afs_dentry_open

Our caller already got a ref to 'mnt'; getting another one here will
overcount refs on 'mnt'.

This can make it impossible to unmount the filesystem the cache
resides on (even after the client is stopped), since it's ref count
will be very high.

FIXES 131613

Reviewed-on: http://gerrit.openafs.org/9790
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit e019429d4548348c6ab8674305d985feee040476)

Change-Id: If88e108bb3e67602cf9e6bbe7646f9862228b684
Reviewed-on: http://gerrit.openafs.org/9792
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: Stephan Wiesand <stephan.wiesand@desy.de>

OpenBSD: Add config headers for new(er) OS versions.

The OS-specific config header files have not kept up with the
newer releases of OpenBSD. This patch corrects that by adding
header files for OpenBSD versions from 4.9 to 5.3 (the newest
version currently under development).

Reviewed-on: http://gerrit.openafs.org/8927
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 20a54b02521b40b07e4ec87df55a1e9a2549c881)

Change-Id: I9ee66dbba300d5a68bc294a14dd9463441143371
Reviewed-on: http://gerrit.openafs.org/9661
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

ubik: Do not count votes from error'd connections

If the given connection has a connection-wide error on it, the vote we
got from that site is probably not valid, and we could easily be
interpreting an error code as a vote time. So instead, treat the host
as if we got a network error from it.

Reviewed-on: http://gerrit.openafs.org/8487
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 23092acce5d2d71a5ad6402119973c7c2c2f5e89)

Change-Id: Id320b33bb6662d7867ae1f4a8d0a29808f16b59d
Reviewed-on: http://gerrit.openafs.org/8947
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

ubik: Try to detect VOTE_Beacon errors

Currently the way ubik dbsites vote for each other is via the "return
value" of the Beacon VOTE RPC. Since this is really an Rx abort, this
can easily collide with actual errors on the wire, such as rxkad
errors.

Try to detect these by detecting vote times that are very different
than the current timestamp (more than an hour in the future or past),
and treat it like a network error.

If we do not do this, a single site reporting an error can cause us to
never reach quorum, since we calculate our sync site expiration based
on the oldest 'yes' vote, which for most known Rx aborts will be far
in the past.

Reviewed-on: http://gerrit.openafs.org/8486
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 4d4668b1618a2bd5b94ed4620464787f42d11cab)

Change-Id: Iaca12506a35e924631754b638f99cb12faa84479
Reviewed-on: http://gerrit.openafs.org/8946
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

vol: remove duplicate stmp declaration

Patchset 38cf31463e3f3c675de727c1e793e117a90e6d20 added a definition of
afs_ino_str_t stmp which should have replaced the b64_string_t stmp
declaration that was already present.

Thanks to Jason Madar for noticing.

FIXES 131620

Reviewed-on: http://gerrit.openafs.org/9608
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit dd903ce729c24c9d1ed78b227a261b2549482461)

Change-Id: Ia0a43f64183b47b6df30495673907d176d95799a
Reviewed-on: http://gerrit.openafs.org/9771
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

vos: Restore some VNOVOL error messages

Many places in vos/vsprocs have code to delete a volume. Commit
f4e73067cdef990262c69c38ac98761620a63f25 tried to refactor them by
consolidating the common "delete" code into DoVolDelete. However, not
all of the removed code had exactly the same behavior, and some of
these variants were not handled by DoVolDelete.

One such variation is that DoVolDelete always printed an error message
if the target volume did not exist. But for some call sites this
condition is not an error, and prior to the refactoring they did not
print such an error message. Commit
1092cbe34fc8519826b3fa0565505b7bd81bc922 tried to correct this by
suppressing the error message if the target volume does not exist.
However, this means that all DoVolDelete calls do not print such an
error, where some should and some should not print an error. This
means that in some edge cases when we encounter an unexpected VNOVOL
error, we now skip printing the specific error we got and instead go
right to cleanup/recovery/exit. For a few other cases, we used to
print an error and continue (because it is a non-fatal error or a
warning), but now we print nothing when we encounter a VNOVOL error.

Fix this by specifically printing an error for the VNOVOL error for
DoVolDelete call sites that used to print such an error. Do this for
all such sites except ones where we obviously print an error
immediately afterwards anyway.

This is just a quick targeted fix. A future more robust fix should
involve altering DoVolDelete to handle all of the different behaviors
expected by its various callers.

Reviewed-on: http://gerrit.openafs.org/9704
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 0a063815cc27ed065bcc9e4a91fb3a7b280818bb)

Change-Id: Ia333b2257b9fdc39f7bcc9ba8ed3c98254d2e76a
Reviewed-on: http://gerrit.openafs.org/9770
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

vos: noise messages when attempting to delete non-existent volume

With vos refactor commit f4e73067cdef990262c69c38ac98761620a63f25,
some formerly conditional volume deletes are now unconditional.
This regresses 'vos move' output with harmless "error" messages
when AFSVolTransCreate() returns VNOVOL:
  "Failed to start transaction on <volume>"
  "Volume not attached, does not exist, or not online"

Modify DoVolDelete() to return early (and silently) with VNOVOL
in this case, allowing the caller to handle this appropriately.

Reviewed-on: http://gerrit.openafs.org/9596
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 1092cbe34fc8519826b3fa0565505b7bd81bc922)

Change-Id: I77e701378a999c8ad9413bfefcace0ee06aecb07
Reviewed-on: http://gerrit.openafs.org/9597
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>

Rework set_header_word macros

Rework the set_header_word macros so that all compilers are
happy:
- the use of offsetof() is avoided, as it has an issue on IRIX
when the result is not constant
- the assignment within the macro is explicitely sequenced before
the function call to avoid a gcc sequence-point warning

Reviewed-on: http://gerrit.openafs.org/8816
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 7f96084969d3082842a9575ee8ff7da8c52a0796)

Change-Id: Iefdf7f6dbb60ffc569b67b3c89893fb43527eb02
Reviewed-on: http://gerrit.openafs.org/8818
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Simon Wilkinson <simonxwilkinson@gmail.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

macos: get more packets if requested

analogous to eead07418f13cd87a90770f03ea5118c546d8f1a for ukernel,
this is for macos' upcall procedure

Reviewed-on: http://gerrit.openafs.org/7788
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit 33a761ebd11e33b5e111e67204218e8f67bef431)

Change-Id: I1b59c802717393758370c376f95569467dc61049
Reviewed-on: http://gerrit.openafs.org/9747
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

viced: Sanity check file link count during CoW

A few ihandle bugs in the past have caused the CopyOnWrite code to
open cached file handles for files which have been deleted. When we
CoW, both of the files we're dealing with had better actually be on
disk, so bail out and flag an error if either of them appear unlinked.

FIXES 131530

Reviewed-on: http://gerrit.openafs.org/8839
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 752514bd1c8eaae487d25116dc59853552f77d12)

Change-Id: I2aa70145b12e9babd1a1eb881bf7f48151c20f6c
Reviewed-on: http://gerrit.openafs.org/9508
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

ihandle: Add FDH_ISUNLINKED

Add the FDH_ISUNLINKED functionality to ihandle. This lets the caller
know if the file for the underlying file descriptor has been deleted
out from under us. This is useful for sanity checks in some callers.

Reviewed-on: http://gerrit.openafs.org/8838
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 9cf9a0e978ece2b0afb8ba5947455f307a424cab)

Change-Id: If5cde825a2e7413c47409c69f15507d6df1934c3
Reviewed-on: http://gerrit.openafs.org/9507
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

volser: Check vnode length on dump

Commit aadf69eabb1962496fa93745ab560a5b48cacd61 adds length checks on
vnodes during fileserver read/write operations. Do the same thing when
we dump volume data from the volserver, to ensure that we don't
transmit incorrect data e.g. to other RO sites when releasing.

FIXES 131530

Reviewed-on: http://gerrit.openafs.org/8836
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit eb938bd18d9806f3529b7a2cbccdce5340c0e39f)

Change-Id: Ia49e7310f2e143d4eed965a5090e687be55864e4
Reviewed-on: http://gerrit.openafs.org/9505
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

afs: Add a little more info on SLVC loop panic

If we panic due to a perceived infinite loop, log a little more info
about our loop iterations.

Reviewed-on: http://gerrit.openafs.org/8850
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 2f087dc434f62cfa7888434427eaac6803b563ab)

Change-Id: Ie4a02a40fafa88f8b54b05b8a0926f258ae7cf7a
Reviewed-on: http://gerrit.openafs.org/9509
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

afs: Apply VLRU safety check for Linux too

This invariant should apply to all platforms, not just those with
dynamic vcaches. Since this prevents an infinite loop if the list os
corrupt or something, having this around everywhere seems useful. So,
drop the check for afsd_dynamic_vcaches.

Reviewed-on: http://gerrit.openafs.org/8555
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
(cherry picked from commit 696db8667d12b1c5639515840f69c9a46d2b866d)

Change-Id: I81e4c7c7678a9f4a3e02952e243735d59b3af35c
Reviewed-on: http://gerrit.openafs.org/9502
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>

viced: fix log message for MapName_r

Do not log a code of zero if the name plus instance length exceeds
the max pr name buffer.

Reviewed-on: http://gerrit.openafs.org/6861
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Tested-by: Derrick Brashear <shadow@dementix.org>
(cherry picked from commit ef6ae1e7eda56e1d6ff7b9ade25b290c855e5780)

Change-Id: If644670ed522eaba575e7246084f106231649991
Reviewed-on: http://gerrit.openafs.org/9474
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>

viced: Ignore client loopback alternate addresses

When we receive interface information from a client to determine its
alternate IP addresses, ignore any addresses that look like loopback.

Reviewed-on: http://gerrit.openafs.org/2368
Reviewed-by: Jeffrey Altman <jaltman@secure-endpoints.com>
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Tested-by: Jeffrey Altman <jaltman@secure-endpoints.com>
(cherry picked from commit 2a6ddae62e4fac5eecdf4634746e7294405393d6)

Change-Id: Idcdaabd3aad62f0e84efbaf1d8523338bc683924
Reviewed-on: http://gerrit.openafs.org/9466
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@your-file-system.com>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Tested-by: BuildBot <buildbot@rampaginggeek.com>