Linux: install_session_keyring: key_alloc flags are unsigned long

Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/4034
(cherry picked from commit 01c6e6bb532be688c032871da820997f1fb5150d)

Change-Id: I501e04ec6a804bb1d1eacaab0090747c229a05dd
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
(cherry picked from commit db5c991747879f3365a8218fd40052e7d137dcc1)

Linux: Reduce key_alloc flags confusion

KEY_ALLOC_IN_QUOTA should be 0, not 1.

Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/4033
(cherry picked from commit dfcd3fcae5e2d2c22f9692c60f0b4e5858c25c9e)

Change-Id: I910e87f147a5fc5022cacb8dee799503ebf5476b
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
(cherry picked from commit 4cf4381af5130f7c7845a6a24c55a3623383581a)

Don't count root session keyrings against quota

AFS PAM modules can call setpag() as root, regardless of the UID
being authenticated. This leads to the session keyring being created
using roots quota - on some systems (RHEL5) this quota is both small,
and of a fixed size.

Modify our keyring allocation code so that when a keyring is created
by root, we don't do any quota checks.

Reported-by: Rainer Toebbicke <rtb@pclella.cern.ch>
Reviewed-on: http://gerrit.openafs.org/1577
Tested-by: Rainer Toebbicke <rtb@pclella.cern.ch>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit c4537f0442ac7ecbf8c946de45004992e17d535f)

Change-Id: Ia152db35b756abc3b3684194ea3c155f22fef537
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
(cherry picked from commit 9bd2fe688f3d1cea48a7d261db3bceed6a244a85)

Fix cherry-pick of “Linux: 2.6.38: dentry->d_count is not an atomic”

Gerrit broke this patch while cherry-picking it; see
http://gerrit.openafs.org/3999 .

Change-Id: I16c8263d4173af35e524f0ec1bc058e8c6e02ab8
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/4057
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 7350bf894444c05011a3807dd2b69b2d6c2d25c0)

Linux: 2.6.38: dentry->d_count is not an atomic

d_count is now an int protected by the dentry's d_lock.
Take the lock when we use it, instead of using an atomic_*
function.

Reviewed-on: http://gerrit.openafs.org/3883
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 281f5bf5fbb0a546edcce62ef4e097ae9bbdbf73)

Change-Id: Id30054cb110720632f6f4a7fc4d7860cb69cb1e7
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3999
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit fe8d833e53938319d1fb6e20153a50a74bed5860)

Linux: 2.6.38: deal with dcache_lock removal

dcache_lock is gone in 2.6.38, and some of the vfs locking rules
have changed.

Of interest for openafs:
- inode->i_lock protects the d_alias list
- dentry->d_lock protects d_unhashed()

Add a new configure test for dcache_lock, and replace its use by
the appropriate new lock(s).

Reviewed-on: http://gerrit.openafs.org/3771
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 2eca7aef7b2940e4ef5f9901ce28481af6edb6dd)

Change-Id: Ic544db9bcf820f01f23227b71684847d742fa41f
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3998
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit 9e614fafad635704acd7958283903ee2f88dc1c6)

linux: 2.6.38: Make d_revalidate RCU-walk aware

Signal to the vfs that we don't support path RCU walk in
d_revalidate.  It will call back in non-RCU mode.

Reviewed-on: http://gerrit.openafs.org/3759
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 52556d50298e4a636f07e9a3cf0b6557c882c82d)

Change-Id: I6662782068e317a36ca70caa0d09665ff2183c2d
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3995
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit 9a91534004f853de427a661347fe029218ea9ac2)

linux: 2.6.38: New d_op handling

In 2.6.38, the super block structure has a new field to hold the
default dentry ops.  The vfs will automatically set it for new
dentries in most cases.

Set s_d_op to our set of operations, and omit setting the dentry
ops where the vfs will already do it (and where new locking rules
prohibit it).

Reviewed-on: http://gerrit.openafs.org/3758
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 08bb83d95027bb3ac68834d12b72bdc647fa24a9)

Change-Id: I15affd47c02653b3b46a0c97e64d06ba4ecbac1c
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3994
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit 0d95e6b800044f9cfa6dd4057477394c3a9cc44d)

Linux: 2.6.38: Adjust for permission inode operation changes

The permission i_op has a new signature with a flags argument, and
must now deal with RCU path walking.
- Fix existing configure test for this i_op, it succeeds when it
shouldn't
- Add a new configure test for the new signature
- Make our permission i_op "RCU-walk aware" - return ECHILD if
called in that mode

Reviewed-on: http://gerrit.openafs.org/3770
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 5bcc0ea735ea519298c98b46c66bf1326cdee5e4)

Change-Id: Iebfab65b3442ea286873e6af96f1c05c98e5d126
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3997
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit 29855d9dc5d117ba106bdb0dd9a9bc6e0592c19c)

Linux: allow compile flags to be passed to AC_CHECK_LINUX_BUILD

Some linux autoconf tests require particular compile flags such as
-Werror.  Add a parameter to the AC_CHECK_LINUX_BUILD macro that
lets the caller pass in any needed special flags.

Reviewed-on: http://gerrit.openafs.org/3769
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit ca38c9546e323b59189670b80740510fb9ec5269)

Change-Id: I8798e103b2aff77b87684b7fce56c730f0a74b8b
[andersk@mit.edu: Remove modifications of existing tests]
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3996
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit a9ca625ba761d28a54bf817438cb9d42eb421672)

Linux: Add general autoconf macro for Linux kernel

Add AC_CHECK_LINUX_BUILD() to do a standardised Linux build, which
takes a "checking" message, the autoconf variable to use to cache
the results, headers and code to run, preprocessor variable to
define, and a description of that variable.

Reviewed-on: http://gerrit.openafs.org/1854
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit a89d6b029d059d01b8bbfc18b9ee6e3f5a964b85)

Change-Id: I40730801f21bd058c66e475c4a5d2b67ace04cfb
[andersk@mit.edu: Remove reimplementation of existing check macros]
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3993
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit 62c7e0b01e225708f525a9e6019fa11531b06aa4)

Linux: Add autoconf macro for structure checks

Add a new autoconf macro for doing structure element checks.

Reviewed-on: http://gerrit.openafs.org/1853
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 02f2c7cb3734d44dc90b77d631909373daefacd3)

Change-Id: I3639293928a821ba41456b79fd1425711f6e36f4
[andersk@mit.edu: Remove modifications of existing structure checks]
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3992
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Russ Allbery <rra@stanford.edu>
(cherry picked from commit 166019b13bf9e9be3983533ff59ae4a23e4cc9ea)

volser-higher-thread-limit-20080327

LICENSE IPL10

with tvolser we can increase this and actually *see benefit*, so...
(cherry picked from commit 7c05046f1f097a6ded89a5928ea48aca224f520a)

Change-Id: Idcccc3d8ea5422b5ee0941a1a8a0feaf87e63661
Reviewed-on: http://gerrit.openafs.org/3967
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit f7721817c5c6607c8752004c62bd6316851354e5)

vol: Log ignored dirs that look like partitions

If we see a /vicep*-like directory when we VAttachPartitions, and we
ignore it because it lacks an AlwaysAttach, log that we ignored it.
This may make things less confusing to admins that just try to create
a /vicep* directory and don't know why it doesn't work.

Feature suggested by Jason Edgecombe.

FIXES 128221

Reviewed-on: http://gerrit.openafs.org/2893
Reviewed-by: Jeffrey Altman <jaltman@openafs.org>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 4803d50bdb8cdeae2c3f86b5c09e853573bc49eb)

Change-Id: Ibd9d5b2ba9b8b143e0c13060c1aa45adc8e2bcd1
Reviewed-on: http://gerrit.openafs.org/3572
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 15908bb13f0ae17cf5525a1644456bab16e57881)

Honour --enable-debug for userspace parts

For userspace we use ${DBG}, whereas for libafs ${DBUG} is used.
Use ${DBG} outside of src/libafs.

Change-Id: I8985e45fd8406d61b85e6e4156fa313bd056a857
Reviewed-on: http://gerrit.openafs.org/3541
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 70c0dd9e2b0a43e76219e2d8f039a08ccfe90a6d)

vos: Mark longjmp-used variables as 'volatile'

vsprocs tries to do error recovery by calling longjmp from a signal
handler. Although this is quite error-prone since we call a ton of
non-async-signal-safe functions, make it a bit more likely to work by
marking variables that are used after the longjmp as volatile. This
reduces how often (depending on the platform) these values will be
completely worthless after a longjmp since they were cached in a
register or similar.

FIXES 125535

Reviewed-on: http://gerrit.openafs.org/1557
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit fe3ff8e91237f35fed33d6d417830e0fe44c6479)

Change-Id: I6d63f516a0864c643f5fea2a711a8adea9fdcc6d
Reviewed-on: http://gerrit.openafs.org/3573
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit f6a272c69683cbb9fcd7e23ff693dd6c57c0e1ac)

vos status: actually show created time

The 'created' field in 'vos status' output was being reported as
something else. Make 'vos status' actually report the time the
transaction was created in that field.

Reviewed-on: http://gerrit.openafs.org/2336
Reviewed-by: Jeffrey Altman <jaltman@openafs.org>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 36af430c29bdbfc1b4b5c1ad7568cbffab04899f)

Change-Id: I671be1ad023fc5cffc6757c6225b5efadb4dca8c
Reviewed-on: http://gerrit.openafs.org/3574
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 2325283b31c126c4f8ed9f6c7ba4d8842e3e3b0c)

Finalize changes for 1.4.14+dfsg-1

Merge branch 'upstream'

Imported upstream tag openafs-stable-1_4_14 via tarball

Add changelog for 1.4.14 upstream release

Prepare debian/changelog for new upstream release

Add bug closer for #607903

Add $named to Should-Start for openafs-client

* Add $named to Should-Start in the openafs-client init script since the
  client may do DNS lookups for the VLDB and file servers during startup
  in some situations.  (Closes: #586226)

openafs 1.4.14

perhaps 14 is luckier than 13. sigh.

Change-Id: I07d7308cf31f0706e37494d6e421df55126c595b
Reviewed-on: http://gerrit.openafs.org/3540
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Linux: use DEFINE_MUTEX for afs_linux_alloc_sem on newer kernels

DECLARE_MUTEX is gone as of 2.6.37; afs_global_lock switches from
DECLARE_MUTEX to DEFINE_MUTEX at 2.6.16, so do the same here.

(cherry picked from commit f9409fcdffa594aff9530038a5ac81a2495da3b6)
Signed-off-by: Alejandro R. Sedeño <asedeno@mit.edu>
Reviewed-on: http://gerrit.openafs.org/3273
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Tested-by: Derrick Brashear <shadow@dementia.org>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Change-Id: I3d72f56db7ebaf40f1cdcdbf9f73783cb9d71676
Reviewed-on: http://gerrit.openafs.org/3539
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

LINUX: Use correct type of error in flock code

The flock code gets a kernel error code from afs_posix_lock_file()
but then passes this error code through afs_convert_code (which sees
that it is already negative, and returns EIO). Instead, we should just
return afs_posix_lock_file()'s code direct to our caller.

Reviewed-on: http://gerrit.openafs.org/2534
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 5abe706a7e237d3872db3638d947a14c935370d5)
Change-Id: I8e12c9804cc81c8c88f820f55abd4ce863525c9a
Reviewed-on: http://gerrit.openafs.org/3533

update ticket5 from heimdal

includes updated instructions for updating.
note that gen_glue.c in lib/asn1 has a bug which omits
a needed newline in the "const struct units * asn1_%s_units(void);",
line.

FIXES 127554

Reviewed-on: http://gerrit.openafs.org/2306
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 582878a75858a341f674f833609f08b6d3bf839a)

Change-Id: I8a7eefb19d353469376a799614005245aeb91b1f
Reviewed-on: http://gerrit.openafs.org/3522
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

make openafs 1.4.13

update versions for openafs 1.4.13

Change-Id: Ifdaa32de6896afc9efa9b05da10182a737ce1f0e
Reviewed-on: http://gerrit.openafs.org/3531
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Linux: define llseek operations

With kernel 2.6.37 it is now mandatory to define the llseek operation
for files and directories.  If these are not defined, no_llseek is
called, and any attempt to seek returns ESPIPE.

Most file systems use generic_file_llseek, but it seems safer to use
default_llseek which is what the vfs used to call for us by default.
In 2.6.37 these two functions are actually functionally identical.

Reviewed-on: http://gerrit.openafs.org/3292
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry-picked from commit fb6b22cf77039962f22f462ee602f0e4a8153817)

Reviewed-on: http://gerrit.openafs.org/3512
Tested-by: Marc Dionne <marc.c.dionne@gmail.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit d0e31ec418555354ac30516269a9748ee01fc36f)
Change-Id: If90a5ac65a9ccd1d1301884065e9583ad25f5af4
Reviewed-on: http://gerrit.openafs.org/3527

CellServDB update 13 Dec 2010

Reviewed-on: http://gerrit.openafs.org/3505
Reviewed-by: Jeffrey Altman <jaltman@openafs.org>
Tested-by: Jeffrey Altman <jaltman@openafs.org>
Change-Id: I6c27922b7877041d5f74751aa06fad469d5e92cd
Reviewed-on: http://gerrit.openafs.org/3526
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Linux: fix statfs configure test

The change to the statfs configure test that was made for 2.6.36
broke the test for older kernels.  The new test is based on a call,
and that will generate a warning but not an error when the arguments
don't match the prototype.

Take another tack, and revert to the old style test, but with the
simple_statfs function instead of vfs_statfs.

Reviewed-on: http://gerrit.openafs.org/2975
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
(cherry picked from commit 59927e20052c3f075e0269e46691bc94c3974f86)

Change-Id: Iaee4e3d62ce95c33539ba57c19ddc916b097deb2
Reviewed-on: http://gerrit.openafs.org/3521
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Linux: 2.6.35 - fsync no longer takes a dentry

In 2.6.35, the fsync file operations drops the dentry argument.
Add a configure test and cope.

Reviewed-on: http://gerrit.openafs.org/2064
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit 4d89ce0a89e9b9e2709e6f6d730f245b5bce744b)

Reviewed-on: http://gerrit.openafs.org/2549
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 58bffd741cdc917f164f19b4621aa2567e6a5b3e)
Change-Id: Ibeb609a57003fc863568e4fc619db9746aa9462f
Reviewed-on: http://gerrit.openafs.org/3520
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>

Linux: adapt to truncate sequence changes

As part of changes to the truncate sequence, inode_setattr() no
longer exists, and all filesystems have to define the setattr op
so we can assume that it is not NULL.

Introduce a compat inline function afs_inode_setattr that hides
the tests and the different versions from the main code.

Note that the existing test for the inode_setattr() return type
will fail, but the value is no longer used in that case.

This is required for 2.6.36

Reviewed-on: http://gerrit.openafs.org/2543
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit eaf3378f537935f6b9843886b43d)

Reviewed-on: http://gerrit.openafs.org/2550
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 58e56d080b8d19117b60f04ecb37af0c6dcafc1a)

Change-Id: I4cfdd191c60d7672c80aa3b3789e52c3e5e87666
Reviewed-on: http://gerrit.openafs.org/3519
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Linux: use %pI4 if NIPQUAD is not available

The definition of NIPQUAD has been removed from the kernel headers
in the mainline kernel.  Replace it by using the %pI4 format
specifier which takes the IP address as a single argument.

There should be no change in the output.

This fix is required for 2.6.36.

Reviewed-on: http://gerrit.openafs.org/2542
Tested-by: Derrick Brashear <shadow@dementia.org>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit d4018b2673e6532543984838290ffb891b970fea)

Reviewed-on: http://gerrit.openafs.org/2551
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit e2e8881db7757ccac5ce3d5fc5603c40c5b47a43)
Change-Id: I1bbe4f8f24f18d29fb6af98fc6e6f4707def547f
Reviewed-on: http://gerrit.openafs.org/3518
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>

Linux: switch to evict_inode

In 2.6.36, the delete_inode and clear_inode inode operations
are replaced by evict_inode.
Rename our current clear_inode to evict_inode, and add a few
things that were previously handled by the generic delete_inode.

This is required for 2.6.36.

Reviewed-on: http://gerrit.openafs.org/2544
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit fb01fbd7cd2d396df8a4605eaad9febc52ef3b61)

Reviewed-on: http://gerrit.openafs.org/2552
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit e05b5937201ec313d2c8fa863f4dacec87ab687d)
Change-Id: I42a22dbea4f880f5c8bb1cd0904f1711b3a48292
Reviewed-on: http://gerrit.openafs.org/3517
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>

Linux: Rework statfs super block operations configure test

The configure test to detect if the statfs super block operation
needs a dentry argument is based on vfs_statfs, and assumes that
its signature matches the one of the operation.  In 2.6.36 this is
no longer true and the test fails.

Rework the test to actually test the operation we're interested in.

This change is required for 2.6.36.

Reviewed-on: http://gerrit.openafs.org/2545
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit 08552ad8a94f7cc5908aabe8385711e09a6418e4)

Reviewed-on: http://gerrit.openafs.org/2553
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit b8b708fc416aa281fe1d1ce1d498cc381c81679c)
Change-Id: Ib40de2018ecec5e3c82d8fa829364c31d29526b4
Reviewed-on: http://gerrit.openafs.org/3516
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>

Linux: RedHat packaging updates for RHEL6

Update our bundled spec file and related tools so they can be used
to build OpenAFS on the RHEL6 beta.

- Make kmodtool recognise el6 as having "modern" kernel naming
  conventions
- Replace %{PACKAGE_VERSION} (which seems to have disappeared)
  with the standard %{version} macro

Thanks to billings and phalenor on IRC for their testing efforts.

Reviewed-on: http://gerrit.openafs.org/1806
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 0d0e7699c9f789214205fe6837cded1a4c95f9c0)

Reviewed-on: http://gerrit.openafs.org/2638
Tested-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 41040638e23be43e260f79a4694df528244ef38b)
Change-Id: Ia2d1d1c377d7df0a4e7bdd22ac5f86c7253ffa66
Reviewed-on: http://gerrit.openafs.org/3515
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>

RedHat: Find krb5-config in /usr/bin

In RHEL6, krb5-config is no longer in /usr/kerberos/bin/, but is in
just /usr/bin/. So, try to use /usr/bin/krb5-config if we cannot find
/usr/kerberos/bin/krb5-config and krb5config was not specified.

This is a 1.4-only change, as the 'configure' on more modern branches
should have the ability to find krb5-config automatically.

Reviewed-on: http://gerrit.openafs.org/2646
Tested-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit c2eba9d57adc79b05165fc8aebaf751ea89282f5)
Change-Id: I46c6f0e3d7c2ce427eacfbb82366a7da3156d9bd
Reviewed-on: http://gerrit.openafs.org/3514
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>

Fix the sample CellServDB entry output by afs-newcell

* Fix the sample CellServDB entry printed by afs-newcell if the cell
  isn't found in /etc/openafs/CellServDB to include the correct server
  syntax.  Thanks, Peter Palfrader.

openafs-modules-dkms must depend on libc6-dev

* Add a dependency on libc6-dev to openafs-modules-dkms.  dkms doesn't
  depend on it because most kernel modules don't need it, but openafs
  builds userspace helper programs.  Thanks, Peter Palfrader.

Use dpkg --print-architecture in openafs-fileserver.config

* Use dpkg --print-architecture, not --print-installation-architecture,
  in the openafs-fileserver config script.  The latter is deprecated.
  Thanks, Peter Palfrader.  (Closes: #594488)

Update standards version to 3.9.1

* Update standards version to 3.9.1.
  - Remove obsolete conflicts/replaces/provides on openafs-ptutil.

Fix typo in changelog

Finalize changes for 1.4.12.1+dfsg-3

Run DKMS commands in a subshell for proper logging

* Run the DKMS build commands in a subshell to work around a DKMS bug
  that caused the module build logs to be discarded.  (LP: #593509)

Include com_err.h in dumpstuff.c

src/volser/dumpstuff.c uses afs_error_message in its error reporting,
but it does not include com_err.h, so it does not get the prototype
for afs_error_message. The default return type for unprototyped
functions is 'int', so the return value of afs_error_message can get
truncated on platforms where sizeof(char*) > sizeof(int), causing a
segfault whenever the result of afs_error_message is logged.

This is a 1.4-only change. This was fixed in the 1.5 branch in the
midst of the larger warning cleanup effort.

Change-Id: Id3bcde20609130d8e4e393c6c8aae75682bcd966
Reviewed-on: http://gerrit.openafs.org/2526
Tested-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit ebb6accf18cf3eeb6a99025ccf5228f979463d0e)

vol AttachByName should mimic other volpkg errors

if later callers will deem a volume offline, note it at attach time.
additionally, log the cause.

(cherry picked from commit 939382c5c1622750665a9f543168bd3b8044583f)

Change-Id: Ie5d4a6339749c4f1920a78722966b013e78ffd98
Reviewed-on: http://gerrit.openafs.org/2465
Reviewed-by: Jeffrey Altman <jaltman@openafs.org>
Tested-by: Jason Edgecombe <jason@rampaginggeek.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Jason Edgecombe <jason@rampaginggeek.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit fd19dd095a52be4a685ac80e8533b7a4388e145c)

ubik: ntohl on reading the replay log

When attempting to read the replay log, ubik was not ntohl'ing all
integers that were read in from the log, causing the log to appear
invalid on little-endian systems. Fix it.

This problem manifests as apparent corruption in the database on top
of ubik when a commit is occurring when we are shut down, or a disk
error is encountered during a commit.

Reviewed-on: http://gerrit.openafs.org/2224
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit eec0d94f519b3e27f255b9b7a637df043951424e)

Change-Id: Ic6b3ba18c843d1f41c45fa69559bea869620439a
Reviewed-on: http://gerrit.openafs.org/2237
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 639c0e467af3d78619e11cf44895809e0b9b8edf)

Add changelog entries for cherry-picked commits

openafs-fileserver init script improvements

* The openafs-fileserver init script now depends on $named since
  apparently the volserver requires DNS during startup.  Thanks, Jaap
  Winius.  (Closes: #589783)
* Add a status command to thet openafs-fileserver init script.

Finalize changes for 1.4.12.1+dfsg-2

Remove mention of IBM release notes

* Minor update to README.Debian for openafs-doc: the old IBM release
  notes are no longer included.

Drop support for lpia, add powerpcspe

Initial changelog for 1.4.12.1+dfsg-2

Do not use AFS_PTR_FMT or %p in 1.4

The afs_snprintf in the 1.4 series did not support %p, so don't use
it when logging fileserver messages.  Replace with 0x%lx and add a
cast to unsigned long, which should serve as an acceptible substitute.

Change-Id: I4b3ca260a35bb4e09bfbbe35bfe362ccd2b8eee9
Reviewed-on: http://gerrit.openafs.org/2128
Tested-by: Russ Allbery <rra@stanford.edu>
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit f9e1b702a3cd13066afd84e6745f95b4d1ababaa)

Finalize changes for 1.4.12.1+dfsg-1

prdb_verify -rebuild with supergroups fix

This change eliminates the redundant iteration over the
supergroup[] array in DumpRecreate. By iterating over both
this array and entries[], duplicate au lines are created
and the count variable becomes incorrect. Furthermore,
the au lines created in the supergroups[] section were
incorrectly in the form of au <group> <member> instead
of au <member> <group>.

FIXES 126500

Change-Id: I087772b63934406273f153f4b8a581d965f1e4d5
Reviewed-on: http://gerrit.openafs.org/1316
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Tested-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit fa2536c02cf26f7db30bf623e95f42be88138d14)
(cherry picked from commit dfa356732fd10073c3b777ccf48f753fcd528c93)
Reviewed-on: http://gerrit.openafs.org/1945
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 3fc5c5311fc4bfe5586818fa26bdccd846d4c8cc)

viced-host-uuid-and-addr-hashing-corrections-20090530

LICENSE IPL10
FIXES 124634

only valid addr/port pairs are registered in the hash table.
add then remove when changing addresses.
make host restoral properly hash hosts.
remove should remove the address we asked for and not simply the
primary address.

Change-Id: I765a81f5e993bbc0b275ea93cd84cc67e67b14ff

(cherry picked from commit ca613599a2537756462a420ae1a632747a433226)

Change-Id: I765a81f5e993bbc0b275ea93cd84cc67e67b14ff
Reviewed-on: http://gerrit.openafs.org/1843
Tested-by: Dan Hyde <drh@umich.edu>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit c29ac4fe358eabac40d30cbe5519d6c10073ba53)

Add changelog entries for cherry-picked upstream fixes

Merge commit 'upstream/1.4.12.1+dfsg'

Imported upstream tag openafs-stable-1_4_12_1 via tarball

Adjust get-orig-source for location of patch releases

* Adjust get-orig-source to cope with tarballs for patch releases being
  in the same directory as the parent normal release.

Add initial changelog for 1.4.12.1+dfsg-1

make openafs 1.4.12.1

update configure scripts with new version number

Change-Id: I3d75cdd481369aaaaaa3de343c70360a6e20847a

Move non-executable stack assembly code to end of file

The non-executable stack annotation used for Linux works by adding
an additional section to the generated object.  The annotation added
in 9bfd03d723a9cf17673f40513a7adde1d503bcbc was added to the beginning
of each assembly file.  This caused build failures on Linux s390
because the remainder of the assembly code was then put in that
section, which was discarded during linking.

Add the annotation to the end of each assembly file instead, following
the instructions at:

    http://www.gentoo.org/proj/en/hardened/gnu-stack.xml

Change-Id: I5bd2d8d0ac31932b59757dcec0a942c459d61cac
Reviewed-on: http://gerrit.openafs.org/1615
Tested-by: Russ Allbery <rra@stanford.edu>
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 9dbe48ac2d14742d2acefb9cb4138b810a1a128e)

fix other oldtvix typo

i failed to push this to commit 0dac8f93af5262472bc5bdf94dbb605a24da59d0,
and so, well, here's the real fix. sorry.

Change-Id: I6799ecb207f246ac3e4c48b1de55268d2fba4d23
Reviewed-on: http://gerrit.openafs.org/1837
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit 51c1eb362da866e28a9fb9a6f5d9d412aac7e8c2)

Change-Id: If42f31b47d50aecb147fa82040a6c3ab48f7a7ce
Reviewed-on: http://gerrit.openafs.org/2009
Tested-by: Derrick Brashear <shadow@dementia.org>
Reviewed-by: Derrick Brashear <shadow@dementia.org>

Initialize oldvtix

Initialize the variable to avoid compiler complaints that it
could be used uninitialized.  Use a sentinel value and issue
a warning if we try to use the variable and we haven't set
it to a useful value.

Change-Id: I037be2ad8e0b55d06b50aafdeca5012764640f16
Reviewed-on: http://gerrit.openafs.org/1810
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit 0dac8f93af5262472bc5bdf94dbb605a24da59d0)

Change-Id: I3311ad5a73e21f7fb44f9dcf0cda251bae96eb0f
Reviewed-on: http://gerrit.openafs.org/2008
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Recover from afs_GetVolSlot errors

afs_GetVolSlot can panic in a few different ways, such as failing to
read from or write to VolumeInfo. Instead of panic'ing, return an
error to the application. Adjust callers to deal with getting a NULL
volume returned.

Based on a patch by Mike Meffie.

Change-Id: Ibb301fed795ee9ee5906b8e7973945a06218b8f1
Reviewed-on: http://gerrit.openafs.org/1801
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry-picked from commit 80f05473428969ff71d1791cda8cb09d30223724)

Change-Id: Ie89034cb0fcc8f92868f338996d62edfe150b1b2
Reviewed-on: http://gerrit.openafs.org/2004
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>

Linux: replace invalidate_inode_pages

This helper has been deprecated for a while, and gets removed
with 2.6.34.
Replace it with invalidate_remote_inode, which has been around for
all of 2.6's life, according to Chaskiel in RT #124377.

Change-Id: I8658b454e1c997c041f654ba55d06f713ea25459
Reviewed-on: http://gerrit.openafs.org/1523
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 80481fb14ea9301b0f6b3503d07ecaef39aef0d8)
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-on: http://gerrit.openafs.org/1988

macos 32 bit platform user address transform

make the ioctl32 translator work as expected, e.g., create
proper user addresses.

Change-Id: Ia690dee5f9978d679e942a9954f2123df4a96bc2
Reviewed-on: http://gerrit.openafs.org/1687
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit ec74299c7c0f6e705e6ba2329001b92f9941868e)

Check if AFS is mounted before killing processes with open files

* Skip killing processes with files open in AFS if AFS does not appear
  to be mounted according to /etc/mtab.  Otherwise, we may call lsof
  without a specific mount point and kill far more processes than we
  intend to.  (This code is disabled by default, so this problem would
  only be seen by people who enabled it.)

Finalize changes for 1.4.12+dfsg-4

Parse /etc/mtab instead of the mount command to find the AFS mount

The /etc/mtab syntax doesn't contain an English word and is therefore
less likely to change based on localization.  Hopefully this will be
equivalent.

In the openafs-client init script, don't assume AFS is mounted at /afs

* In the openafs-client init script, don't assume that AFS is mounted on
  /afs when unmounting it or killing processes with AFS files open.
  Instead, parse the output from mount to find the AFS mount point.

Read all members of previous cacheinfo file

If we don't have three destination variables, the cachedir gets the
cache size appended to it.  Add a cachesize variable whose contents
we then ignore.

Linux Keyrings: don't ignore error code from session keyring creation

Creating a session keyring can fail, for instance if the user's
keyring quotas are reached (number of keys/keyrings or data size).

Instead of ignoring errors, return them so they can be passed back
to the caller.

FIXES 126230

Reviewed-on: http://gerrit.openafs.org/1151
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from 0caf14224a9153bb488be9e52d67892a2c441a5a)

Change-Id: Ibffa5ff33f756c9365b1148448039e668d39fb92
Reviewed-on: http://gerrit.openafs.org/1581
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 0bec77e1f9dfdb7927926953b808603647910cc0)

Adjust afs_lockctl to compensate for byte-range lock fixes

As of commit 49b7bbdd3b45df694fadbef48f9ed99d9bfe07b9, l_len is no
longer 1 less than it should be.  Update afs_lockctl to expect this.

Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Change-Id: I3fb7665457d489e56564b14933795084a0958d6f
Reviewed-on: http://gerrit.openafs.org/1353
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 99e704c791bbedd9aa3fc1a0feb9493b0eaefd7d)
Reviewed-on: http://gerrit.openafs.org/1373
(cherry picked from commit 89193695da6dc8946e4a03ed40081a91050ca5dc)

VOL_LOCK needed when traversing DiskPartitionList

VLockPartition sets VOL_LOCK, opens (perhaps creating) the lock file
for a partition, and flock's it.  VUnlockPartition, sets VOL_LOCK, and
closes the lock file fd.

VLockPartition is called from the salvager.  Nothing ever calls
VUnlockPartition.

VPFullUnlock is called every 30 seconds, does NOT set VOL_LOCK, and
happily closes any lock file fd it finds, breaking any lock set with
flock.

Other routines use the VOL_LOCK when traversing DiskPartitionList.

Change-Id: I28aa02488a6b53dc9f9aa9b2053c71d29222bdaf
Reviewed-on: http://gerrit.openafs.org/1213
Reviewed-by: Dan Hyde <drh@umich.edu>
Tested-by: Dan Hyde <drh@umich.edu>
(cherry picked from commit 105481d3348a93d33695fb705f9fa5c8af210e7e)
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
Reviewed-on: http://gerrit.openafs.org/1595
(cherry picked from commit 3c1d60a19ccdc23ecd9eb0ca79bd72816abcd82e)

Linux: don't count pag keys against root's keyring quotas

Keys associated with PAGs are created with root ownership for
security reasons, which means that they count agains root's
keyring quotas.  The default configuration used by most distros
restricts root to the same quotas as a regular user, so a single
user can potentially fill up the quota and prevent new pags from
getting created system-wide.

This can also be an issue for busy multi-user systems where the
default maximum number of keys (200 currently) can easily be reached.

Reviewed-on: http://gerrit.openafs.org/1167
Reviewed-by: Marc Dionne <marc.c.dionne@gmail.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit a3812f211a56c0d6e0a7ff8a97f157707d3d8c28)

Change-Id: I5c12e55d98d5f887c60a7ae1cf1aa64f37282ce3
Reviewed-on: http://gerrit.openafs.org/1580
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit a27015099a1dc1e41001e4a4946848f958c6a09b)

Add changelog for upstream cherry-picks

Preserve AFS mount point and cache directory in cacheinfo

* Preserve the AFS mount point and cache directorys set in
  /etc/openafs/cacheinfo if the file already exists rather than
  overwriting them with the defaults.  Thanks, Liam Healy.
  (Closes: #580077)

Finalize changes for 1.4.12+dfsg-3

Add changelog entry for upstream cherry-picks

Print rxdebug statistics as unsigned values

bytesSent and bytesReceived are unsigned.  Print them as such
to avoid printing negative numbers for large values.  This is
already the case in the master branch.

Bug reported by jakllsch on IRC

Change-Id: If2c5d9cfebf11691d73b3d6bfd73deb6b61ada8e
Reviewed-on: http://gerrit.openafs.org/1657
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit a123d4ab988c6a8cc51b864ea1d02a70b2770bf3)

Increase the maximum number of sysnames

Change-Id: I8438ca19649730260370d8b00fb0157e808abe31
Reviewed-on: http://gerrit.openafs.org/778
Reviewed-by: Jeffrey Altman <jaltman@openafs.org>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit b772bd7e3c1305710eb3c18b2fe4108e467bd383)
Reviewed-on: http://gerrit.openafs.org/1650
(cherry picked from commit 812dcc2c1b4e0f333dd0a85f46066c8018aad79b)

volmonitor keep vtrans lock

VolMonitor must maintain the VTRANS lock the whole time it is walking
the allTrans list.  Failure to do so can cause core dumps trying to
access memory that has already been free'd.

Two versions of this change were coded and tested.  The other version
used the VTRANS lock only around a THOLD, but needed a TRELE, too.
Timing tests were run counting the number of vos status, vos listvol,
and vos backupsys operations that could be performed during a fixed
number of vos status operations.  The THOLD/TRELE version caused other
vos operations to run about 5% slower.

FIXES 126110

Change-Id: I7e749d30c955867faacafa978d7d643dee648ca3
Reviewed-on: http://gerrit.openafs.org/1098
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit d4f6ece19b05f2233e53f2d3c59f92316ec5ec3a)
Change-Id: Ic12bb6687a9c34e85abc2023c02264de6ae27614
Reviewed-on: http://gerrit.openafs.org/1588
Reviewed-by: Dan Hyde <drh@umich.edu>
Tested-by: Dan Hyde <drh@umich.edu>
(cherry picked from commit 190ef2cb29577f2fcf1853149cf210fdf0736ae3)

Create missing root directory when ORPH_ATTACH

When we are salvaging with ORPH_ATTACH, orphans are normally attached to
a volume's root directory. If the volume is missing a root directory,
however, nothing is attached, and the volume can appear empty or
unusable.

So, to make it possible to get a useful volume out of a volume that lost
(only) its root directory, create a new root dir, and attach orphans to
that root.

FIXES 94658

Reviewed-on: http://gerrit.openafs.org/1235
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 32ebc44f42aad7dd77d42d07bfd891e8be6f1fff)

Change-Id: Ib74507e28630ce409f412b67f19d6008d215c6b0
Reviewed-on: http://gerrit.openafs.org/1613
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 135e196b6145d5d2ebddc1d65b3193d2b5c7cffb)

Don't add CellServDB records if AFSDB is in use

* If the user configures openafs-client to use AFSDB records for VLDB
  server location, don't prompt the user for VLDB servers for the local
  cell even if they're not present in CellServDB and don't try to add an
  entry for the local cell to CellServDB.  (Closes: #575299)

Finalize changes for 1.4.12+dfsg-2

Build with -fno-strict-aliasing

* Build with -fno-strict-aliasing.  The upstream development branch has
  a better fix that selectively enables this for specific files with
  known problems, but be conservative for the stable release.

Move non-executable stack assembly code to end of file

The non-executable stack annotation used for Linux works by adding
an additional section to the generated object.  The annotation added
in 9bfd03d723a9cf17673f40513a7adde1d503bcbc was added to the beginning
of each assembly file.  This caused build failures on Linux s390
because the remainder of the assembly code was then put in that
section, which was discarded during linking.

Add the annotation to the end of each assembly file instead, following
the instructions at:

    http://www.gentoo.org/proj/en/hardened/gnu-stack.xml

Change-Id: I5bd2d8d0ac31932b59757dcec0a942c459d61cac
Reviewed-on: http://gerrit.openafs.org/1615
Tested-by: Russ Allbery <rra@stanford.edu>
Reviewed-by: Simon Wilkinson <sxw@inf.ed.ac.uk>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 9dbe48ac2d14742d2acefb9cb4138b810a1a128e)
Reviewed-on: http://gerrit.openafs.org/1618
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit eb799d07e7dd37f379a4e84df5eb3397b16f5a41)

volmonitor copy link before calling free

Copy tt->next before TRELE(tt) calls free(tt).

We have a core file from a VTRANS_OBJ_LOCK(tt) assert failure, with tt
pointing into glibc's malloc data structures.
(cherry picked from commit 81ca1e19600681a2c5696610130b9a39809bb1c1)

Change-Id: I5ba81f9e3a76f52578ab3b9dec394b6d95b03934
Reviewed-on: http://gerrit.openafs.org/1586
Reviewed-by: Dan Hyde <drh@umich.edu>
Tested-by: Dan Hyde <drh@umich.edu>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 0583af324ca2cbc556722b9e84d85cad49cfc622)

h_TossStuff_r: make sure host does not go away

When h_TossStuff_r h_NBLock_r's a host, it is not only possible for
someone else to grab a hold on the host, but in theory it's also
possible for someone to hold a host, release it, and for the host to be
deleted again (assuming some callers hold HOSTDELETED hosts, which they
should not be doing).

To make this safety check a bit more robust, hold the host in
h_TossStuff_r before h_NBLock_r'ing, to ensure that it does not go away
while we're waiting for H_LOCK.

Reviewed-on: http://gerrit.openafs.org/1330
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 380bd77c328d8d83a007cd97f6564fcb5b990a73)

Change-Id: I19d001dba8b0c96b694aefdba2475f0103d33e44
Reviewed-on: http://gerrit.openafs.org/1368
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit b78eeb0cca48c65d4466052d207fa22f08a8fa70)

h_TossStuff_r: check held-ness after lock

h_TossStuff_r checks if a host is held or locked by another thread
before trying to delete the host. Unfortunately, it checks if it is
locked before checking if it is held, and the lock check drops H_LOCK.
Thus, another thread could hold the host while we don't have H_LOCK, and
we could delete a host that is being held.

Although it is a bug if any thread holds a host that is being deleted,
some instances of this still exist, so make the check more robust.
Reverse the order of the tests, so we detect if someone held the host
while the lock check dropped H_LOCK.

Also log when this happens, as it indicates a bug occurring.

FIXES 126454

Reviewed-on: http://gerrit.openafs.org/1312
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 45aceee9842265fb0ccdb5e8f3f6d32c8d2b99ea)

Change-Id: Ibce716b0d57d02d1ad912276dfa9f4b169294947
Reviewed-on: http://gerrit.openafs.org/1367
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 94a439668b8ef90027f922e9fb85fe1f3e983100)

Allow GetSomeSpace_r to select an optimal host

Previously GetSomeSpace_r would never find an 'ideal' host for which to
clear callbacks, since lih0_r and lih1_r required a non-NULL rock to do
anything. Remove the requirement for the passed-in host rock to be
non-NULL, and make lih*_r more threadsafe, by passing in a parameter
struct for the rock.

Also attempt to make the GSS_r code a bit more clear with some
descriptive variable names and such.

FIXES 126451

Reviewed-on: http://gerrit.openafs.org/1317
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit eb7d90fd0db68ea49ec1820adda599b5922c7de2)

Change-Id: I1c283153977f47cb21c57bf6e65305068c6fd3be
Reviewed-on: http://gerrit.openafs.org/1370
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit f3899ac3b342ad065dd9484720a4d721a3fb549a)

Add changelog entries for upstream cherry-picks

Remove lih_r

Nothing references the function lih_r anymore. Remove the dead code, and
be a little more similar to what's on master.

Change-Id: I00262cd27156fc370cd74c35d2e7a0309e1037f9
Reviewed-on: http://gerrit.openafs.org/1369
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Alistair Ferguson <alistair.ferguson@mac.com>
Reviewed-by: Derrick Brashear <shadow@dementia.org>
Tested-by: Derrick Brashear <shadow@dementia.org>
(cherry picked from commit 4ca7b6fcc124f473ed5b86bea804a9df5d2c8833)

Hopefully finalize changes for 1.4.12+dfsg-1

Merge branch 'upstream'