Always try to dist debian stuff, so that even on non-dpkg-using
build targets (RHEL series), the tar dist will contain all debian
stuff, and can be delivered to users who want to build lustre on
debian series servers.

Test-Parameters: trivial
Change-Id: I8493f5a5e65078c0160c5493efee6906c87db6f1
Signed-off-by: Gu Zheng <gzheng@ddn.com>
Reviewed-on: https://review.whamcloud.com/33246


Reviewed-by: Wang Shilong <wshilong@ddn.com>
Tested-by: Jenkins
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

The new Ubuntu server support is missing on last package that RHEL
currently supports. Enable Debian/Ubuntu packaging for the resource
agent package.

Test-Parameters: trivial

Change-Id: I872ae6b1b02782155db14c5cc971607ca3e93d56
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-on: https://review.whamcloud.com/33152


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

This introduces a new package lustre-client-modules-dkms that uses DKMS
to automatically recompile the client kernel modules on kernel upgrades.
The package is only created if the dkms-debs target is used, otherwise
the traditional kernel-specific package is created.

Test-Parameters: trivial
Change-Id: Ie9aeee29f7fd73938b148299d246c663a783ccd3
Signed-off-by: Michael Kuhn <michael.kuhn@informatik.uni-hamburg.de>
Reviewed-on: https://review.whamcloud.com/25328


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

This enables server build for Ubuntu 18.04 LTS, the ldiskfs
patches are based on Gael's 4.12 support,
they apply to kernel versions 4.15.0-20.21 to 4.15.0-23.25

There's also a small fix to make dpkg happy when installing
lustre packages which requires lustre-client-utils.

Test-Parameters: clientdistro=ubuntu1604 trivial
Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Signed-off-by: Gael Delbary <gael.delbary@cea.fr>
Change-Id: I65e1a5ee0d17115f23ba071ff1ab23b4fb22e78f
Reviewed-on: https://review.whamcloud.com/32613


Tested-by: Jenkins
Reviewed-by: Quentin Bouget <quentin.bouget@cea.fr>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

Remove the obsolete linux-patch since patched kernels for lustre
clients have been long gone. Place only the static libraries and
*.so symlinks for the dynamic libraries in lustre-dev. The normal
dynamic libraries are placed into the utilities packages. Add in
all the missing dependencies and fix how the lustre debs are
dependent on each other. Lastly add in the missing lustre-iokit
that is present for rpm packages. Only thing missing is a package
for lustre resources which can be done at a latter time.

Test-Parameters: trivial

Change-Id: I5fd2a23bc1ae73434cef8dcf3679b50878256ab3
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Minh Diep <minh.diep@intel.com>
Reviewed-on: https://review.whamcloud.com/31348


Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Thomas Stibor <t.stibor@gsi.de>
Tested-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

This breaks out lustre-dkms into lustre-zfs-dkms and
lustre-ldiskfs-dkms (or lustre-all-dkms) as "flavours" of lustre
server dkms.  The reason for the flavours is to prevent lustre
ldiskfs dkms build from having ZFS dependencies, and to maintain
lustre zfs dkms build ordering when rebuilding for new kernels.
This also prevents building of tests and utils when --disable-tests
and --disable-utils (respectively) are passed to configure.

Test-Parameters: trivial
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: Iba500d9830a8f57662066141a176c381151861f4
Reviewed-on: https://review.whamcloud.com/27990


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

When modules are excluded from a build there is no rule to build undef.h.
As result build fails with following:
/usr/include/stdc-predef.h:40:1: fatal error:
        /work/lustre-release/undef.h: No such file or directory

Test-Parameters: trivial
Change-Id: I4bf031933964c7d11f9e1c4a88016e1827d11762
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: https://review.whamcloud.com/30965


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

This enables compatibility with the current LTS flavours of Ubuntu.
Do note that you need the Xenial HWE Kernel for Ubuntu 14.04.5, as
that distribution originally used a 3.x series Kernel.

The patches have been developed to apply cleanly to the kernel versions
4.4.0-45.66 to 4.4.0-85.108 from the Ubuntu Xenial (and its Trusty backports).

This change also adjusts the Debian scripting to produce the
ldiskfs modules and the server utilities.

To create the server modules run "./configure" with "--enable-server"
and specify "--enable-ldiskfs" and "--with-zfs/-spl" as appropriate.

The call to "make debs" will then produce the server modules and
utils instead of their client versions.

NOTE: This contains a small hack taken from LU-9995 / #29130

Test-Parameters: trivial
Signed-off-by: Martin Schroeder <martin.h.schroeder@intel.com>
Change-Id: I02cd5e9314367ad4e1f8f3d81712f84441a8bc71
Reviewed-on: https://review.whamcloud.com/29215


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Thomas Stibor <t.stibor@gsi.de>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

This actually breaks our Ubuntu builds which prevents any sort of full testing, so I am reverting this now.

This reverts commit 86c3f90d.

Change-Id: I14d242bfde1efb0144080b882e63542fc2190465
Reviewed-on: https://review.whamcloud.com/28293


Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Tested-by: Oleg Drokin <oleg.drokin@intel.com>
Tested-by: Jenkins

This enables compatibility with the current LTS flavours of Ubuntu.
Do note that you need the Xenial HWE Kernel for Ubuntu 14.04.5, as
that distribution originally used a 3.x series Kernel.

The patches have been developed to apply cleanly to the kernel versions
4.4.0-45.66 to 4.4.0-85.108 from the Ubuntu Xenial (and its Trusty backports).

This change also adjusts the Debian scripting to produce the
ldiskfs modules and the server utilities.

To create the server modules run "./configure" with "--enable-server"
and specify "--enable-ldiskfs" and "--with-zfs/-spl" as appropriate.

The call to "make debs" will then produce the server modules and
utils instead of their client versions.

Test-Parameters: trivial
Signed-off-by: Martin Schroeder <martin.h.schroeder@intel.com>
Change-Id: Ia78702da304f735bb932738784f1346be0f39026
Reviewed-on: https://review.whamcloud.com/27323


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Gu Zheng <gzheng@ddn.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Thomas Stibor <t.stibor@gsi.de>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Building DEB packages failed with errors:
dh_install: lustre-dev missing files (debian/tmp/usr/lib/*.so.*)
and dpkg-checkbuilddeps: Unmet build dependencies.
Moreover, signing the DEB packages failed due signing problems
of the author listed in debian/changelog.
Additionally, latest DEB based distributions ship Lustre modules
from kernel staging which can conflict with modules build from
upstream. To fix the problem, build modules with
configure --with-kmp-moddir=updates and add search order
of "search updates built-in" in /etc/depmod.d/lustre.conf

Signed-off-by: Thomas Stibor <t.stibor@gsi.de>
Change-Id: I02c58103bfaa83b97f5cd207af823fb477e0f0d3
Reviewed-on: https://review.whamcloud.com/23999


Tested-by: Jenkins
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

With the landing of the patch:

  "LU-5614 build: use %kernel_module_package in rpm spec"

we neither need, nor desire, to tie a lustre build to a specific
Linux kernel version.  Instead we want to allow the weak updates
system to symlink the lustre modules into any and all kernel
directories that are ABI compatible.

Therefore, we remove the Linux version string from Lustre's
RPM release field.

Note that "%define krequires" is moved to a SLES-only section
when the 0%{?build_src_rpm} conditional is removed.  This may
seem like an odd place to put it at first glance.  But the
next line is the only remaining consumer of krequires, so
we are just moving it to the last place that it might be needed.
This could perhaps be simplified further in a future commit.

When removing the no longer needed %{fullrelease} from the Release
field, we replace it with a 1 and with the %{?dist} string.
%{?dist} is customary for Fedora, RHEL, and likely most rpm-based
distros.

Because the release field is no longer modified between the source
rpm and binary rpm build steps, there is not longer a need for
The build_src_rpm macro.  That is removed as well.

Test-Parameters: trivial
Change-Id: I8306f48d5bf0dffcae2888505ec7f0f63e47fc08
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/19954


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Explicitly clear the LDFLAGS for kernel modules build so that
the vendor's overzealous linker flags don't create build failures.

Test-Parameters: trivial

Change-Id: I427cffb229a19a8ff0c2ba0c3ed2ac135f21e24d
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/20404


Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Replace module packages definitions with calls to the vendor defined
%kernel_module_package macro.  This results in many of the rpm packages
changing names from what we are used to in the past.  Each Linux
distribution may name the modules packages slightly differently, most
notably RHEL and SLES choose different naming schemes.  The
%kernel_module_package macro handles these details for us.

For instance, on RHEL7 the packages that contain lustre's kernel modules
become:

  kmod-lustre             (previously in "lustre-modules" package)
  kmod-lustre-osd-ldiskfs (previously in "lustre-osd-ldiskfs" package)
  kmod-lustre-osd-zfs     (previously in "lustre-osd-zfs" package)
  kmod-lustre-tests       (previously in "lustre-tests" package)

On SLES12, the same packages are instead named:

  lustre-kmp-default
  lustre-osd-ldiskfs-kmp-default
  lustre-osd-zfs-kmp-default
  lustre-tests-kmp-default

These kernel module packages give us what is known as "weak-updates".
Instead of compiling Lustre for every minor kernel that is installed,
the weak-updates will automatically symlink the Lustre modules into
and kernel's modules subdirectories if that kernel offers the required
kernel symbols with exactly the same symbol versions.

In addition to changing the package names for all of the Lustre packages
containing kernel modules, the location of the modules in /lib/modules
changes as well.  Each package will put its modules in a subdirectory
matching part of the package name itself (required by the macros on RHEL
systems).  The path name is also different on RHEL and SLES systems
(e.g. "extra" vs. "updates").

Some further path examples are:

  RHEL7, kmod-lustre package:
    /lib/modules/3.10.0-327.18.2.el7_lustre.x86_64/extra/lustre/fs/ost.ko

  SLES12, lustre-kmp-default package:
    /lib/modules/3.12.57-60.35_lustre-default/updates/lustre/fs/ost.ko

  RHEL7, kmod-lustre-osd-ldiskfs package:
    /lib/modules/3.10.0-327.18.2.el7_lustre.x86_64/extra/lustre-osd-ldiskfs/fs/osd_ldiskfs.ko

  SLES12, lustre-osd-ldiskfs-kmp-default package:
    /lib/modules/3.12.57-60.35_lustre-default/updates/lustre-osd-ldiskfs/fs/osd_ldiskfs.ko

Test-Parameters: trivial clientdistro=el6.7 mdsdistro=el6.7 ossdistro=el6.7
Signed-off-by: Stephen Champion <schamp@sgi.com>
Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Change-Id: Ic5351f4bda94eae80eaf349c8490a2c758b15c80
Signed-off-by: Minh Diep <minh.diep@intel.com>
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/12063


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

For rpm, we can't have dashes in the version string so we use
underscores.  For dpkg, the requirements are reversed.

In the "debs" make target, we convert underscores in the version
string to dashes.  We also trim down the comment that no longer
really makes sense with the new versioning scheme.

Change-Id: I6d600fc4a16be60c664ef0662b482c31877fbfea
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/19488


Tested-by: Jenkins
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

The lustre_build_version.h is no longer necessary.  With the new
versioning system the lustre version can now be retrieved from
standard autotools defines and/or lustre_ver.h.

To make compatibility easier for the upstream lustre client in
the linux kernel, we use LUSTRE_VERSION_STRING from lustre_ver.h.

Change-Id: I08233d8547bc4af7bf2fb0c15bee8e306f6fa0f0
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/18108


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Authors and reviewers of patches adding things to the contrib
subdirectory have been inconsistent with the inclusion of the
contrib tree's contents in the distribution tarball.

This patch fixes that by removing all of the Makefiles from the
contrib tree, and instead just listing contrib in EXTRA_DIST
at the top level.  That will recursively include the entire
contrib directory in the distribution tarball.  This way
there is little oportunity for inclusion of files under
contrib being inconsistent.

Change-Id: If6b92ed069a3fbeb1d0a8c8c0488180ee69597f3
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/15629


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Brian J. Murrell <brian.murrell@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

This patch overhauls how version numbers are handled.

The biggest change is removing the two independent
places where versions were being generated (LB_BUILDID and
lustre_tag*.pl) with a single script named LUSTRE-VERSION-GEN
at the top level of the tree.  LUSTRE-VERSION-GEN is based
on the GIT-VERSION-GEN script that git's own source code build
system uses.

LUSTRE-VERSION-GEN generates the lustre version string from
either git, or if git is not available from the LUSTRE-VERSION-FILE
file, or if _that_ is not available, its own internal DEFAULT_VERSION
default version number.  This allows building the lustre code under
all reasonable conditions.

The new method uses LUSTRE-VERSION-GEN to supply the correct
full lustre version (which may contain git --describe information)
to AC_INIT.  This means that the autotools actually know the
correct version number.  Some benefits of this are:

* We can use standard {AC_}PACKAGE_VERSION macros and variables
  instead of inventing our own
* ./configure --version and other commands will list the correct
  version of Lustre
* The lustre tarball created by "make dist" will contain the full
  version number

The lustre/autoconf/lustre-version.ac file was replaced with
a new config/lustre-version.m4 file.  We no longer need to work
with m4-only macros prefore AC_INIT is called, so the new
file defines a LUSTRE_VERSION_VARIABLES function for us to call
from configure.ac.  The new lustre-version.m4 derives the
derive LUSTRE_{MAJOR,MINOR,PATCH,FIX} from AC_PACKAGE_VERSION
rather than being yet another source of version numbers.

The LUSTRE_VER_ALLOWED_OFFSET and LUSTRE_CLI_VER_OFFSET_WARN
m4 macros were dropped, and the OBD_OCD_VERSION strings stored
directly in lustre_ver.h.in instead.

LB_CHECK_VERSION was removed (no longer necessary).

AC_LUSTRE_VERSION_STRING was drop.  It is no longer necessary
now that we can set LUSTRE_VERSION_STRING directly to
@PACKAGE_VERSION@ in lustre_ver.h.in.

The DOWNSTREAM_RELEASE configure argument is deprecated and all
related code removed.

The lustre/doc/VERSIONING documenation is relocated to
Documentation/versioning.txt and updated to reflect the new versioning
method and practices.

The META file and BUILDID variables and related code are removed
since they are no longer used under the new versioning method.

Note that this patch makes lustre only look at the dotted version
tags in git, so we can finally stop using three to five tags for
every release.  Now we can just use one clear tag.  The tags
with underscores are obsolete.

Change-Id: I7196f92ed4975776cc8e73fd3ebf436b5921f83e
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/18107


Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Intel needed to build Lustre for its non-self-hosted accelerator version
of the Phi processor. It turns out that the OS for that is packaged in some
weird non-standard way. Intel added support into Lustre to rpm package it.
While those additions use words like "cross_build", making it sound like it
might be generic, it really isn't.

All of that is really messy and non-standard, and it is making far too
difficult to package Lustre for normal rpm-based distributions.

Hopefully this won't be too much of a problem for Intel, because Phi
processors are now self-hosting.

This is in large part a revert of commit
  426194de:
"LU-2784 build: Provide RPMs build for Xeon Phi(TM) card"

Change-Id: Ie5d655f1da3e17eae6edebb5f4d6fd6e7d58811f
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/17488


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

This patch modifies internal scripts to strengthen
lustre[-client]-dkms package vs issues ([spl,zfs]-dkms not built,
configure errors, ...) that could be encountered during on-target
DKMS build step.

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: I2dc8c91ff5874228dde0bc7995e45a6dcc6973d4
Reviewed-on: http://review.whamcloud.com/18167


Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Jenkins
Reviewed-by: Michael MacDonald <michael.macdonald@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Move the SNMP lustre tools to use cfs_get_paths so
the tools function with the upstream client as well.

Change-Id: I18d76e10c45a9c8c582e77917a33bb2afa86aac4
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: Wang Chao <chao.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/17465


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Permit Lustre Client (only) modules DKMS RPM creation.

This patch is a follow on to the first set of patches for LU-1032
that only allowed for the creation of Lustre Server (zfs only)
modules DKMS RPM.

It also changes original behavior by allowing to dynamically
modify dkms.conf on-target. This particularly helps to change
configure and list of modules to be built/installed list, like
to configure with gss to build ptlrpc_gss.ko module when
krb5_devel is present instead to have it been a mandatory
required dependency.

Also implements feature of DKMS RPM creation from Makefile
(thanks to mjmac), now in 2 separate SRPM/RPM steps and for both
Client and Server versions.

Also use an auto-increment (Array[${#Array[@]}]=) operator in
dkms.conf modules declarations to help for future changes when
there will be a need to add/delete modules.

Change in lustre/utils Makefile has been required to allow
building of ptlrpc_gss module with --enable-gss and without the
need to specify --enable-utils which was causing an unexpected
zfs user-land dependency for DKMS Server build.

To satisfy lustre rpm requirement of a package providing
lustre-osd, provides has been added to DKMS Server RPM since
it does actually generate osd-zfs module.

Signed-off-by: Michael MacDonald <michael.macdonald@intel.com>
Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: I278d50307a17fe49a06392351890946b7dd3557a
Reviewed-on: http://review.whamcloud.com/12347


Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Starting with OFED 3.18 the OFED package started to use config.h
provided by autoconf tool. This lead to clash between PACKAGE_*
macros which are defined in OFED and Lustre headers.

Also dealing with kernels that have lustre enabled already
required to undefine the common macros that used by in-kernel
Lustre client and this sources.

This fix undefine all symbols that are generated by Lustre autoconf
to avoid conflicts with kernel defines or OFEDs. They are undefined
right before new definition in config.h. The list of symbols to undef
is automatically generated by autoconf and should not be extended in
the future.

Also undefine clashed macros in autoconf checks.

Change-Id: I0d93adac19573328e905ba536db0dbd842ed2aab
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/16418


Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

There are two problems:

Firstly, we need check whether O2IBPATH is valid before using
if not, asssign '--with-o2ib=no'instead.

Secondly, macro O2IBPATHS might be "$LINUX $LINUX/drivers/infiniband".
--with-o2ib only expect one string assignment here

Signed-off-by: Wang Shilong <wshilong@ddn.com>
Change-Id: I86172056c57ec8e649c2e56455369e66bbbe3513
Reviewed-on: http://review.whamcloud.com/16183


Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Change way extra symbols are included in build.  Use
KBUILD_EXTRA_SYMBOLS to directly included Modules.symvers
files instead of buildling Modules.symvers in buildroot
directory.

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I4a6ee59f4e4eed9f878ad52993b8b17426f19d4a
Reviewed-on: http://review.whamcloud.com/15498


Tested-by: Jenkins
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

We rework the rpms and srpm Make targets.  The srpm
target generates the source rpm package much as it did
before, but it puts the results in a known temporary
rpm build tree.  This allows the "make rpms" target to
simply depend on the srpm target, and then build the
binary rpms from the source rpm.

Previously the rpm target _also_ generated a source rpm
package, which could be redudant, and failed to verify that
the source rpm package was itself buildable.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Change-Id: I9b2e83e000aca62b5d4841e2ebe9e91fa1db300f
Reviewed-on: http://review.whamcloud.com/12961


Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Debian seems compiling with fixed configurations, if users
want to make debian package with external IB, it won't work.

Signed-off-by: Wang Shilong <wshilong@ddn.com>
Change-Id: I0ecb5c79d5948c0e8ab821c175ca89afee9cb97d
Reviewed-on: http://review.whamcloud.com/13949


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Blake Caldwell <blakec@ornl.gov>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

After passing a few configuration parameters in "--with/--without"
option to rpmbuild some code become useless.

Don't pass options through configure_args that can be passed through
rpmbuild options. This allows to avoid unexpected behavior during
the build from source rpm.

Change module-dist-hook: target according coding guidelines.

Remove obsolete liblustre.{a,so} from .spec file that were actually
removed in commit cdfbc722.

Don't add the version of kernel to .src.rpm.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: Ib5f50d257b5d95efe9c45d1865f9dab9ccc3c19a
Reviewed-on: http://review.whamcloud.com/12987


Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Propagate values from autoconf (LINUX, LINUXRELEASE, LINUX_OBJ)
to the debian build system to correctly identify kernel sources
(--with-linux) and objects (--with-linux-obj). The 'make debs'
target previously used KERNEL_VERSION, which does not get set
anymore.

Other fixes for building debs with Ubuntu 12.04+:
 1) Add module-assistant to Build-Depends in debian/control
 2) Omit copying lustre-*-orig.tar.gz and lustre-*.diff.gz files
    since this is the upstream source
 3) Change debhelper compatibility level to 7 (10.04+)
 4) Use dh_prep instead of dh_clean -k
 5) avoid non-functional --disable-zerocopy configure option
 6) do not try to install lc_common script that is
    only in server builds

Signed-off-by: Blake Caldwell <blakec@ornl.gov>
Change-Id: I9787bac687d26b03c4d1204347240624bfaced96
Reviewed-on: http://review.whamcloud.com/13129


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Brian J. Murrell <brian.murrell@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

- remove unused utilities: lustre_createcsv, lustre_config, lustre_up14
- move req_layout, loadgen, wirecheck and wiretest to tests RPM
- add dependency tests from utils (multiop from liblustreapi)
- make corrent mapping for configure parameters
  "--enable-*/--disable-*" to "--with/--without *"
- fix find command for empty results
- fix utils building (add ifdef UTILS and fix .spec file)
- don't install few server scripts and utils for client build

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: I947ea4739526f74a18427dd853c4198871dd0e21
Reviewed-on: http://review.whamcloud.com/10654


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Headers that are generated during kernel build and placed
in various include/generated directories are always in
the kernel build dir. There are several places in which the kernel
source dir is used instead. This causes configure/build failures
on systems using separate build and source directories.

This patch cleans those paths up.

Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Change-Id: I7335a8088d25f1e3bf64d8c131b542dfce38fff0
Reviewed-on: http://review.whamcloud.com/10160


Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Remove the libsysio and liblustre tests from config/lustre-build.m4,
lnet/autoconf/lustre-lnet.m4, and lustre/autoconf/lustre-core.m4,
leaving a temporary AM_CONDITIONAL(LIBLUSTRE, false) in their
place. Remove related bits from lustre-spec.in and debian/rules.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: I9483bc9114fa0682b84e6a9014123e5838f1c1a7
Reviewed-on: http://review.whamcloud.com/10201


Tested-by: Jenkins
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Brian J. Murrell <brian.murrell@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

liblustre is not used, nor is it tested, so don't build it by default.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: Ie2369d0163963d6e691b3787495b1b2155fe6111
Reviewed-on: http://review.whamcloud.com/9756


Tested-by: Jenkins
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

The "rpms" and "srpm" targets were unnecessarily complicated.  The rpms
target in particular has a very long shell script embedded in the
autoMakefile, which is not especially desirable.  Because of the embedded
shell script with its associated backslashes, we didn't use standard
autoconf/automake macros because we didn't want shell comments to appear
after line continuation.  To get around that, we need another layer of
variables to convert autoconf/automake variables into Make variables.

It gets rather difficult to read and modify.

Instead we move the scripting into autoconf m4 files, where scripting
is much easier (little line continuations necessary, far fewer escapes needed).
We also have direct access to the original variables, so we don't need
to hop through two or three files before we eventually find where
a variable gets set.

All of the decisions are made at configure time anyway, so constructing
the command line options for rpmbuild at configure time is the Right Thing
to do.

A nice side effect of this change is that one can now easily look at
the autoMakefile after running "./configure" and see exactly the command
line that will be passed to rpmbuild.

Change-Id: I10fcfa740d9e901805615c2262263cc1ea8552bf
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/8426


Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Building ldiskfs is enabled by default, so we need to disable it
if --disable-ldiskfs option is specified to ./configure

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: I7ffb013b976e870de32d38b669a1437f8388bbda
Reviewed-on: http://review.whamcloud.com/8883


Tested-by: Jenkins
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Christopher J. Morrone <chris.morrone.llnl@gmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

* Adopt lbuild script for new version of MPSS with x.x.x notation.
* Remove dependency from MPSS package to avoid renaming issue in
  the future. The name of package which was used for dependency
  was renamed in MPSS.
* Use new server with MPSS released packages for download.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: Ie4407ad00177ad6d22770230a4dc6bde967d91ef
Reviewed-on: http://review.whamcloud.com/8836


Tested-by: Jenkins
Reviewed-by: Brian J. Murrell <brian.murrell@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Fix absence dependency modules from ldiskfs sources. Otherwise the
build of modules begins early than actual sources prepared in ldiskfs
directory in case of parallel build (with -j <n> option).

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: I24d3630d9132235a2dc5433eb5b8d3379c04acaa
Reviewed-on: http://review.whamcloud.com/8404


Tested-by: Jenkins
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Christopher J. Morrone <chris.morrone.llnl@gmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

This patch is a first pass at cleaning up version_tag.pl.

Primarily, it removes the strange "make_META.pl" symlink behavior.  It
is not at all clear to me why the symlink was being generated on the fly,
and then version_tag.pl changed its behavior based on its name.  Odd.

Instead version_tag.pl now take "--make_META" as the first option on
its command line.

Further, we haven't use CVS in a long time, so this patch removes
the cvs-related file and checks.

The is_git and is_cvs variables weren't even used, so those are removed
as well.

Change-Id: I651234bdf1d2d32db16151f332f2b3785ee123a0
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/8427


Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

* Remove OFED backport headers for all components except lnet.
* Move all OFED related autoconf testing into lnet/autoconf.
* After OFED built is done, there should only be one -devel rpm.
  We don't really need to filter out any version string.

Change-Id: If983ede2606e4c5c397f676cf16807f6e00318bc
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-on: http://review.whamcloud.com/8451


Reviewed-by: James Simmons <uja.ornl@gmail.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Shuichi Ihara <sihara@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Add --with-kmp-moddir option to configure.

Distributions vary in the installation directory for kernel modules.

The RHEL standard installation directory is
        /lib/modules/$(uname -r)/extra
while the SLES standard is
        /lib/modules/$(uname -r)/updates

Adding the option --with-kmp-moddir option to configure allows users
to select the appropriate installation target.  With this change, it
is neccessary to support both options in the test framework.

Signed-off-by: Stephen Champion <schamp@sgi.com>
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: Iec3137e0e5039dd43622c2e285030a5339fa6fd3
Reviewed-on: http://review.whamcloud.com/8065


Reviewed-by: James Simmons <uja.ornl@gmail.com>
Tested-by: Jenkins
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>