| Commit message (Collapse) | Author | Age |
|
|
|
|
|
| |
This reverts commit d18a6e2590dcde348e7ece6f5e2d76085663413b.
Breaks cross compile in Oneiric x86 chroots.
|
|
|
|
|
|
| |
Signed-off-by: John Rigby <john.rigby@linaro.org>
Signed-off-by: Sebastien Jan <s-jan@ti.com>
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
|
|
|
|
| |
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
|
|
|
|
| |
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Allow to define DEBIAN to be run standalone
- Limit the amount of time wget waits for connections
- Prevent exposing some errors when called by python script
- Print architecture of downloaded packages
Ignore: yes
Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Acked-by: Herton Krzesinski <herton.krzesinski@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
| |
We should be extracting both the linux-image and linux-image-extras
packages when calculating the ABI and module lists.
BugLink: http://bugs.launchpad.net/bugs/1021174
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Acked-by: Stefan Bader <stefan.bader@canonical.com>
Acked-by: Herton Krzesinski <herton.krzesinski@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
| |
Create $DEBIAN/abi/$ABI/fwinfo
BugLink: http://bugs.launchpad.net/bugs/1021174
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Acked-by: Stefan Bader <stefan.bader@canonical.com>
Acked-by: Herton Krzesinski <herton.krzesinski@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Backported from a Quantal patch by Andy Whitcroft <apw@canonical.com>
We are losing the perarch perf binary etc when building, this is occuring
based on parallel execution order. This occurs because both build passes
use 'debian/build/tools' as their build directory. This is simply not
safe as execution order is not guarenteed. Move the perarch build to
tools-perarch to avoid collisions.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
| |
We want to make sure we always regenerate both control and control.stub
whenever we clean. Otherwise the buildd will use the version of
control.stub generated when we are building the source package.
BugLink: http://bugs.launchpad.net/bugs/992414
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Acked-by: Leann Ogasawara <leann.ogasawara@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
BugLink: http://bugs.launchpad.net/bugs/974403
"UBUNTU: [Config] Fix invalid linux-headers link" moved the erroneous
symlink from asm-$ARCH to asm-generic. As it turns out, the asm symlink
is superfluous, so just remove it entirely.
Acked-by: Andy Whitcroft <andy.whitcroft@canonical.com>
Acked-by: Leann Ogasawara <leann.ogasawara@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
| |
BugLink: http://bugs.launchpad.net/bugs/974403
https://lists.ubuntu.com/archives/kernel-team/2012-March/019679.html
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Acked-by: Leann Ogasawara <leann.ogasawara@canonical.com>
Acked-by: Herton Krzesinski <herton.krzesinski@canonical.com>
|
|
|
|
| |
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
| |
BugLink: http://bugs.launchpad.net/bugs/882120
Lack of postinit and postrm packaging scripts prevents proper
module registration via depmod.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Acked-by: Andy Whitcroft <andy.whitcroft@canonical.com>
Acked-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
|
|
|
|
|
|
|
|
| |
When building with high parallelism we may overlap calls to dh_gencontrol
and/or the .ddeb removal code from the debian/files list. If these happen
in parallel files may be lost or updates fail leading to build failures or
incomplete packing lists. Add primative locking to each of these scenarios
using the flock command, utilising a lock file in the debian directory.
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a new linux-image-extras package for the virtual flavours.
This package contains all of the left over kernel modules which are not
in the main linux-image package. This allows users to opt-in to a full
server style install in a virtual machine, preventing slow bloat of the
core linux-image install.
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Acked-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move module inclusion to a model when the non-included modules
are retained. This allow two things:
1) processing can now be applied iterativly to carve out any number of
sub-packages, and
2) we can package the 'remaining' modules.
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Acked-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
|
|
|
|
|
| |
Simplify the checks-% target such that ABI and modules checks are
built in parallel whilst both depend on the build stamp.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It turns out that some tossers like to shortcut the process during development.
A common sequence for a quick compile test is to preapre the tree, copy the
.config, then perform an in-tree compile. For example,
fakeroot debian/rules clean prepare-generic
cp debian/build/build-generic/.config .
make oldconfig scripts prepare
make M=`pwd`/fs/ext4
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
| |
If build-% is not declared .PHONY, then its always executed.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
| |
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
| |
It appears to be superfluous.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
| |
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
| |
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
| |
As it turns out, 'kmake modules_install' understands
concurrency, and is quite a bit faster when told to
parallelize the copy operation.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
| |
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
| |
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
| |
The module and ABI checks are not independent, so make sure
the modules check is dependent on the ABI check having been
performed. Also drop some superfluous shortcuts along the way.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
| |
Regardless of the concurrency setting, kernel preparation targets
cannot be made in parallel.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
| |
binary-udebs are always dependent on binary-debs.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
| |
When pulling the compiled kernels to extract the ABI information also
extract the compiler used for that build. Record the result in the abi
directory so it is also recorded historically.
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Acked-by: Stefan Bader <stefan.bader@canonical.com>
Acked-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Because libbfd is GPLv3 only and perf is GPLv2 only. Also this avoids
statically linking against libbfd to allow multiple versions of perf
being installed in parallel.
See: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=606050
BugLink: http://bugs.launchpad.net/bugs/783660
Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
| |
BugLink: http://bugs.launchpad.net/bugs/797556
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
| |
There is no stem '$*' with which to match the automatic variable '%'.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
|
|
|
|
|
|
|
|
|
| |
Allow you to print or insertchanges against a previous version where the
version was miss-specified, or you are jumping base version. For example:
fdr prev_release=2.6.39-3.10 insertchanges
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
|
|
| |
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
|
|
| |
Signed-off-by: Andy Whitcroft <apw@canonical.com>
|
|
|
|
|
|
|
|
| |
BugLink: http://bugs.launchpad.net/bugs/750585
Signed-off-by: Steve Langasek <slangasek@ubuntu.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Since kernel symbols are resolvable internally to the kernel, the kernel
itself has a map of the symbols. Continuing the tradition of frustrating
off-the-shelf kernel exploits, make vmlinuz unreadable for non-root, just
like has been done for System.map, etc.
Signed-off-by: Kees Cook <kees.cook@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
|
|
| |
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|
|
Ignore: yes
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
|