compat-drivers: add support for unified drivers

Companies typically write Linux device drivers either from
scratch of alternatively based on some internal unified
source code tree. Typically the unified driver architecture
is a mess and porting a driver to linux is always a painful
task to the extent at times a full rewrite is required and
at times preferred.

In order to try to help with these type of unified driver
strategies one possibility that has been proposed is to
make the unified driver strategy an open source project and
to not make many compromises on the style and requirements
for Linux. The alx Linux driver [1] is not yet upstream and
is undergoing changes to try to accomodate it to a format
and use new technologies to help address unification attempts
in both an open manner and also lead by the community.

compat-drivers will support these type of driver efforts
only if they adhere to the principles of never compromising
Linux upstream requirements and are targetting getting
upstream into the Linux kernel. Once these drivers are upstream,
their respective git trees will be removed and we'll simply
cherry pick the driver as we do from upstream Linux. If a
compat-drivers releases contains drivers that are from a
unified git repository the release will be postfixed with
a "-u".

[1] http://www.linuxfoundation.org/collaborate/workgroups/networking/alx

Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: Adrian Chadd <adrian@freebsd.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: move all patch types under patches/

The number of patch directories for the different life
cycle a patch can be on its way upstream had reached 4:

  * pending-stable
  * linux-next-pending
  * linux-next-cherry-picks
  * crap

Apart from this we also had the backport related patches.
In the future I see potential for other directories for
different categories of patches. Consider how we are
breaking down the backport patches into specific documented
collateral evolutions and of those type of collateral
evolutions there may be different types of collateral
evolutions which we could likely treat / document differently.
Another type of patch that I forsee us having to deal with
are patches for drivers that are not upstrea -- like crap,
but for drivers in a unified driver tree that is *designed*
from scratch to be targetting upstream Linux inclusion using
the same coccinelle technology I forsee us using for backporting
collateral evolutions.

Since the number of patch directories can get out of hand
at the top level lets just stuff then all under patches/
and deal with patches there. This also places the backport
patches into a directory called collateral-evolutions/ to
help promote and educate on the nature of the changes
that we have to deal with when backporting.

Cc: Julia Lawall <julia.lawall@lip6.fr>
Cc: Adrian Chadd <adrian@freebsd.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20121129

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: make patches apply again

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: do not apply patches in natural sort order

Without this patch the patches are applied in the natural sort order of
numbers, which caused this order for patches/network/:
patches/network/0001-netdev_ops.patch
patches/network/02-ksize.patch
patches/network/0002-net-misc.patch
patches/network/0003-netdev-needed_headroom_tailroom.patch
patches/network/03-rfkill.patch
patches/network/04-netns.patch
patches/network/0004-wext-namespace.patch
patches/network/0005-netlink-portid.patch
patches/network/05-usb.patch

The non natural sort order of numbers is more the intended way to order
the patches as now the spdiff patches are in front of all the other
patches. This makes the patches in patches/network/ to be applied in
the flowing order:
patches/network/0001-netdev_ops.patch
patches/network/0002-net-misc.patch
patches/network/0003-netdev-needed_headroom_tailroom.patch
patches/network/0004-wext-namespace.patch
patches/network/0005-netlink-portid.patch
patches/network/02-ksize.patch
patches/network/03-rfkill.patch
patches/network/04-netns.patch
patches/network/05-usb.patch

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: run refresh for all patches

In the past refresh with quilt was just run for the last patch
directory being applied and not for all, because our script had
problems when patchRefresh() was called more then once in a single
admin-update.sh run. This was fixed in the last patch and now it is
safe to run it on all patches.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: prevent mixing patches from different directories in refresh

The refresh script is able to take an already existing quilt run and
just refresh the further needed patches. This is needed to make it
possible to manually fix a patch which is not applying any more and
then automatically integrate it into the patches directory again. With
the integration of drm it could happen that quilt is running on two
different source directories like patches/network and patches/drm in one
call of admin-update.sh. If some of the drm patches did not apply and
needed a manual fix, quilt moved the patch in the second run into the
patches/network directory, because it was taken in the network refresh
run. This patch does not run quilt if there is still an unfinished
quilt job for a different patch directory.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove copy of README in refresh

There is no README file in the patches folder like patches/drm/README
any more.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: print original name and path of patch in refresh

In the old version patches.orig/<patch name>.patch was printed as the
current patch quilt is working on in refresh, now e.g. patches/network
/<patch name>.patch is printed.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: do not use struct spinlock

struct spinlock does not exists on kernel version <= 2.6.32, use
spinlock_t instead.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add import to fix compile error

On kernel version <= 2.6.28, I saw the following compile error without
this patch:

/net/wireless/util.c: In function ‘cfg80211_get_p2p_attr’:
/net/wireless/util.c:1120:4: error: implicit declaration of function
‘get_unaligned_le16’ [-Werror=implicit-function-declaration]

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches

This was done by running "./scripts/admin-refresh.sh refresh".

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: build RTL8723AE

This driver is building on all supported kernel versions.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove option for alx driver

The alx driver was removed from crap and also remove the config option.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: make patches apply again

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: disable mwifiex for kernels older than 2.6.27

mwifiex got Multiqueue support and while we do port multiqueue
I'm really fucking lazy to port this driver particular changes
over given that I really don't think anyone on planet earth will
be using it for kernels older than 2.6.27.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add rtl8723ae code

Note this driver is disabled for now. I really want to start
applying SmPL for porting this driver for all known collateral
evolutions. At least for the netdev_attach_ops() one.

I'm adding this in to help build the tree as without this the
compilation was failing.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: rename drm config symbols

When config symbols are renamed they should be renamed at all places
and not just at some places.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add CONFIG_MWIFIEX_USB in config.mk

The driver at least compiles with all supported kernel versions.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add CONFIG_RT2800PCI_RT3290 in config.mk

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove CONFIG_WIL6210 from config.mk

The device driver was removed from crap some time ago.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drviers: remove CONFIG_IWM from config.mk

The device driver was removed from the kernel some time ago.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add CONFIG_BT_WILINK in config.mk

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: activate CONFIG_BT_HCIUART_3WIRE in config.mk

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove CONFIG_ATH9K_PKTLOG form config.mk

This option is not available any more.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add ar5523 driver

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Enable ATI/AMD Radeon for kernels >= 3.2

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: drm: Revert ACPI table size check for < 3.6

acpi_get_table_with_size() was exported with kernels >= 3.6. Revert the
size checking for kernels < 3.6.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Rename CONFIG_DRM_RADEON_KMS

Prefix it with CONFIG_COMPAT_ to allow setting it correctly.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Update README

Add GPU subsystem information to README.md.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Enable nouveau DRM driver for kernels >= 3.3

Enable Nouveau DRM driver for kernels >= 3.3.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Fix comments

The comments were wrong causing a misunderstanding while analyzing
the script's function.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Rename CONFIG_DRM_NOUVEAU_BACKLIGHT

This was forgotten during the generation of the renaming patch.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Add missing config check

No configuration change was retriggering config update because
of this missing line unless you give the 'bt' target.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Rename CONFIG_* vars in nouveau driver

Nouveau used 2 Kconfig defined vars to set debugging log levels.
Rename to for being able to set them from config.mk correctly.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Handle int's when generating compat_autoconf.h

Some Kconfig variables take integer values instead of y/m/N like
CONFIG_NOUVEAU_DEBUG and CONFIG_NOUVEAU_DEBUG_DEFAULT. Currently all
values other than y/m/N were escaped with double-quotes to define
them as strings. Fix this to allow integer values for Kconfig
variables.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Copy core/ subdir of nouveau driver

This was introduced with the latest nouveau code rewrite. Copy
this directory too.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: deactivate cfg80211 tracing for kernels <= 2.6.32

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: make patches apply again

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Enable CONFIG_DRM_LOAD_EDID_FIRMWARE

This was added in 3.4. Rename the option in the relevant source file
and enable CONFIG_COMPAT_DRM_LOAD_EDID_FIRMWARE in config.mk to build
the symbol into DRM_KMS_HELPER kernel module.

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Separate code metrics for each subsystem

Add a subsystem marker before code metrics to differentiate between
multiple subsystems, e.g. drm, network, etc.:

compat-drivers code metrics

   1181683 - Total upstream lines of code being pulled

   network
   ----------------------------------------
      3109 - backport code changes
      2580 - backport code additions
       529 - backport code deletions
     12754 - backport from compat module
     15863 - total backport code
    1.3424 - % of code consists of backport work

   drm
   ----------------------------------------
       301 - backport code changes
       219 - backport code additions
        82 - backport code deletions
     12754 - backport from compat module
     13055 - total backport code
    1.1048 - % of code consists of backport work

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix admin-update.sh network

Sujith reported that building after admin-update.sh network
we get a build error with drm. Fix this by creating an empy
drm Makefile.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add target path for ChangeLog file

Forgot to add the target directory for kup upload on the
ChangeLog file.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix ChangeLog file name

This will now produce something like: ChangeLog-v3.7-rc1-4

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add genlog

This will generate log files for our releases.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add / to target stable release end dir

This should fix uploading via kup using scripts/gen-release.sh

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add support for uploading stable releases

This adds support for uploading stable compat-drivers releases.
The new stable releases will go on:

http://www.kernel.org/pub/linux/kernel/projects/backports/stable/

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix sed for gen-release.sh

The replacement was removing the 's' from the compat-drivers
name, and generating 'compat-driers'. We don't yet backport
support for laundry machines so fix this.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add rfkill.h from UAPI path

The rfkill.h header was moved to the UAPI path,
copy it over. But there is one more issue we would
need to address.

The UAPI changes split up the kernel and userspace API
headers into separate directories. We provide backport
support for the new rfkill module to kernels older than
2.6.31. To allow us to dynamically provide only backport
support for those kernels we have a trick within compat
to provide its own include/linux/rfkill.h header which
then will include either your own kernel's rkfill.h or
include the one we are taking from linux-next. The one
we take from linux-next is renamed to rfkill_backport.h.
In order to allow the call to include your own kernel's
rfkill.h the compat rfkill.h uses include_next. The UAPI
changes would mean though that instead of getting your
own kernel's rfkill.h you'd end up getting the UAPI
header file. Fix this by ensuring that for new kernels
we linux_next into the user's kernels rfkill.h and if
the kernel is old we still to the rfkill_backport.h
naming scheme.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove wil6210 driver from scripts/driver-select

This can be reverted once wil6210 get into Linux upstream.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove alx driver from scripts / Makefile

This can be reverted once the driver gets onto Linux upstream.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: remove full drivers from crap

Moving forward no full drivers will be accepted
into crap/. As the alx drivers has proven this only
slows down access to getting users a proper driver.
Restricting the driver to be made available only
through compat-drivers has provided less visibility
and access to the drivers.

The purpose of adding drivers into crap/ over upstream
drivers/staging/ was that some developers may prefer to
not go through the staging development process if they
were sure they could meet the proper upstream quailty
criteria in a timely manner. Practice shows that even
if this is claimed in the end the driver just ends up
bit rotting away in crap/ over a few kernel releases.

Lets stop that. If you want your driver accessible or
automatically backported you have two options:

  * Linux upstream drivers/staging/
  * Linux upstream proper

Choose wisely.

The additional patch directories then remain useful
only for patches content that is already present on
Linux upstream, nothing more.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add uapi/dri to WSTABLE

This will let us ensure we can query these files for
stable patch generation purposes.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix handing of include/uapi

The kernel make file uses an include path of kernel/include,
kernel/include/uapi (I removed all the other unrelated paths). Now
compat-drivers should also use this path otherwise we have to patch
many files. With this patch for compat-drivers it creates a include
path like this: w-d/include, w-d/include/uapi, kernel/include,
kernel/include/uapi that way this is similar to the way the kernel does
it.
When copying all the upai headers like it was done before in compat-
drivers, some headers in c-d/include/uapi will overwrite some important
headers in the kernel, we should just copy the files we actually need.
With this patch compat-drivers compiles for me against my Ubuntu 3.2
kernel, it still fails on some other kernel versions, with as I see it,
uapi unrelated errors.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: make patches apply again

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add UAPI backport fixes

The patch was posted for inclusion upstream. The
DRM code uses some include headers and if we are
not specific about the path what ends up happening
is our old kernel header files gets used instead of
the new UAPI header files.

For details on the UAPI changes:

http://lwn.net/Articles/507794/

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: copy over uapi headers

For details on the new UAPI changes see:

http://lwn.net/Articles/507794/

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20121011

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: change condition for  release target

When making a kernel.org release you should use gen-release.sh -u,
so just checking for the usage of that variable is better.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add gen-release.sh -k option

This should be used for those who really know what
they are doing and making proper releases. This will
force using 'git clean -x -d -f'. You really should
understand what this does before you use this.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add verbose help options for gen-release.sh

Be a bit more descriptive on usage for flags.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: update gen-release.sh to use git clean

gen-release.sh is updated to force git-paranoia to run
git clean for us when uploading to kernel.org. Note
that usage of kup is only for kernel.org uploading
and as such not for the faint of heart.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: improve git-paranoia with git clean

git clean can be used to further sanitize releases.

git clean -x -d -f

Will spit shine your release from any even files
you may have lying around that you tell git to
ignore. We use the -n option to git clean to warn
the user of unwanted content. If someone is really
make releases they should read and understand
git clean man pages and also understand that
running git-paranoia with -c will wipe out clean
all content found by 'git clean -x -d -n'.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Fix typo for CONFIG_BT_HCIUART

The config value should be set to a lowercase 'm' or the config is
ignored and the associated module doesn't build.  Fix this.

Signed-off-by: Kyle Manna <kyle.manna@fuel7.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20121008

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix 98-pr_fmt.patch

This fixes patches/drm/98-pr_fmt.patch for next-20121008
for header path updates for DRM drivers.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix 04-revert-prime-support.patch

This fixes the patch patches/drm/04-revert-prime-support.patch
for next-20121008. There is quite a bit of changes with hunk
differences that required manual editing.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: copy over amp.h as a bluetooth header

Brought to us by Intel via:

commit 903e45411099ae8292f5ce637ad0c72f6fef61db
Author: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Date:   Thu Sep 27 17:26:09 2012 +0300

    Bluetooth: AMP: Use HCI cmd to Read Loc AMP Assoc

    When receiving A2MP Get AMP Assoc Request execute Read Local AMP Assoc
    HCI command to AMP controller. If the AMP Assoc data is larger than it
    can fit to HCI event only fragment is read. When all fragments are read
    send A2MP Get AMP Assoc Response.

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20121002

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix patches/drm/04-revert-prime-support.patch

Required manual intervention due to hunk differences.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix patches/network/25-multicast-list_head.patch

Hunk offsets differences required manual intervention.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: enable DRM drivers compilation

Ozan Çağlayan had done most of the work to suck in all
the DRM drivers but left out the enablement of compiling
the drm drivers. The DRM drivers were also not load
tested. This patch enables to test compilation by
renaming the DRM Kconfig options with a _COMPAT prefix
and enables compilation by default for the i915 drm
driver which I can test against at run time on my
laptop running on v3.2.

Users and developer can help enable older kernels
down to 3.0 (as this was the oldest supported target
kernel for DRM drivers) or other DRM drivers as they
are run time tested.

The admin-update.sh script is modified to now always
run by including the DRM code and patching it accordingly.
If you happen to only want networking code you can do so
by supplying the network argument to the script as follows:

./scripts/adming-update.sh network

This will disable taking in DRM code.

It should be noted that although only the i915 drm driver
is being enabled all other drivers except the following have
been test compiled to work down to v3.0.

-obj-$(CONFIG_COMPAT_DRM_TDFX)  += tdfx/
-obj-$(CONFIG_COMPAT_DRM_R128)  += r128/
-obj-$(CONFIG_COMPAT_DRM_MGA)   += mga/
-obj-$(CONFIG_COMPAT_DRM_SIS)   += sis/
-obj-$(CONFIG_COMPAT_DRM_SAVAGE)+= savage/
-obj-$(CONFIG_COMPAT_DRM_EXYNOS) +=exynos/
-obj-$(CONFIG_COMPAT_DRM_UDL) += udl/

Apart from the above drivers we should be able to enable the
other drivers as reports come in for run time testing on them.
As it is right now only the i915 goes run time tested -- by me.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: backport usage of shmem_truncate_range() on i915

Backport e2377fe0b -- this could likely be implemented
on compat.git but I'd much prefer a better review of that.
For now this stays as a patch.

commit e2377fe0b65e3c7577ff6df1701c56ef477d336f
Author: Hugh Dickins <hughd@google.com>
Date:   Mon Jun 27 16:18:19 2011 -0700

    drm/i915: use shmem_truncate_range

    The interface to ->truncate_range is changing very slightly: once "tmpfs:
    take control of its truncate_range" has been applied, this can be applied.
     For now there is only a slight inefficiency while this remains unapplied,
    but it will soon become essential for managing shmem's use of swap.

    Change i915_gem_object_truncate() to use shmem_truncate_range() directly:
    which should also spare i915 later change if we switch from
    inode_operations->truncate_range to file_operations->fallocate.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Keith Packard <keithp@keithp.com>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: address usage of intel_enable_gtt() for i915

Kernels older than 3.6 don't have intel_enable_gtt() exported
and most Linux distributions have CONFIG_AGP_INTEL as built-in
(y) and as such we don't bother to make it available as
modular here. The change to export intel_enable_gtt() was
made through commit 8ecd1a66 but previous to this the driver
never called intel_enable_gtt() so we can safely assume we
can do the same for older kernels.

There should be a way to export an intel_enable_gtt() for
kernels with CONFIG_AGP_INTEL as built-in but that itself
is a project.

commit 8ecd1a6615f0d9de6759aafe229bc1cc4ee99c7b
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Jun 7 15:56:03 2012 +0200

    drm/i915: call intel_enable_gtt

    When drm/i915 is in control of the gtt, we need to call
    the enable function at all the relevant places ourselves.

Reviewed-by: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This now goes with i915 run time tested against 3.2.0-3-amd64
and code extrated from linux-next next-20120926.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: force using include/drm for header namespace

Turns out DRM drivers use this:

 #include "drmP.h"

Instead of:

 #include <drm/drmP.h>

Withtout this change we are using the older kernel's header
file and not the one we are cherry pickng over from the kernel
we want to help backport.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: separate NOSTDINC_FLAGS include dir lines

This will make it easier to review changes to this variable
later.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh drm patches for next-20120925

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20120925

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add a few header BT header files to WSTABLE

I noted while porting a stable fix down that the patch did
not carry over a few header file updates. This was caused by
the fact that the WSTABLE variable did not carry over the
header files. Fix this.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix kup mkdir on gen-release.sh

kup does not expect an ending slash for mkdir, otherwise
it fails and does not even create the directory for you.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix space damage on git-paranoia

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: add kup support with git-paranoia support

kup [0] is the kernel.org uploader. This integrates support for
usage of kup with the gen-release.sh script. When kup is used
we also use git-paranoia which will bail out if either we have
content on a git tree that lets us produce content that is not
GPG signed, that has not yet been commited, or integrated into
a git tree. To use kup just pass on the -u argument to
gen-release.sh.

For generic releases, that do not use kup (without -u) we still
run git-paranoia but in interactive mode, allowing releases to
be made that are not targeted for kernel.org with random junk
or not GPG signed junk.

[0] https://korg.wiki.kernel.org/index.php/Userdoc:using_kup

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: fix gen-release.sh to keep tar file

Without -k option to bzip2 the original tar file will be
deleted.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: gen-release.sh update to reflect usage of linux-next

This removes the force update parameter option given that
we're not using it, and also updates the release script to
account for using linux-next.git on a daily basis for future
releases.

Usage of kup, the kernel uploader, will be addressed next.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: gen-release.sh rename COMPAT_WIRELESS to COMPAT_DRIVERS

Simple search and replace.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: rename gen-stable-release.sh to gen-release.sh

We'll use this same script to make daily linux-next.git based
releases.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: update gen-stable-release.sh to make bzip2 again

The kernel uploader expects you to sign the tar file but you can
pass along and upload a bzip2 archive. At the other end the
daemon will uncompress and then verify the tarball.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh crap patches for next-20120920

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20120920

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: adjust scripts to support daily linux-next releases

We'll use these for daily linux-next.git and linux-stable.git releases.

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20120918

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: backport netlink pid to portid rename

The patch:

commit 15e473046cb6e5d18a4d0057e61d76315230382b
Author: Eric W. Biederman <ebiederm@xmission.com>
Date:   Fri Sep 7 20:12:54 2012 +0000

    netlink: Rename pid to portid to avoid confusion

    It is a frequent mistake to confuse the netlink port identifier with a
    process identifier.  Try to reduce this confusion by renaming fields
    that hold port identifiers portid instead of pid.

    I have carefully avoided changing the structures exported to
    userspace to avoid changing the userspace API.

    I have successfully built an allyesconfig kernel with this change.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Changed the struct members:

struct netlink_notify->pid    to
struct netlink_notify->portid

struct genl_info->snd_pid     to
struct genl_info->snd_portid

To help backport this and not have to #ifdef around it against
kernel versions compat has introduced two helpers for us to
simply do the backport with three macro helpers:

genl_info_snd_portid()
netlink_notify_portid()
NETLINK_CB_PORTID()

This takes care of the work for us requiring only one
single line change. This patch now goes into its own
new filename following the lead of the other 4-digit
prefixed patches which try to separate collateral
evolutions to be used for spdiff to further backport
uses in other subsystems / drivers.

mcgrof@garbanzo ~/devel/compat-drivers (git::master)$ time ckmake
Trying kernel 3.5.0-030500-generic [OK]
Trying kernel 3.4.4-030404-generic [OK]
Trying kernel 3.3.7-030307-generic [OK]
Trying kernel 3.2.2-030202-generic [OK]
Trying kernel 3.1.10-030110-generic [OK]
Trying kernel 3.0.18-030018-generic [OK]
Trying kernel 2.6.39-02063904-generic [OK]
Trying kernel 2.6.38-02063808-generic [OK]
Trying kernel 2.6.37-02063706-generic [OK]
Trying kernel 2.6.36-02063604-generic [OK]
Trying kernel 2.6.35-02063512-generic [OK]
Trying kernel 2.6.34-02063410-generic [OK]
Trying kernel 2.6.33-02063305-generic [OK]
Trying kernel 2.6.32-02063255-generic [OK]
Trying kernel 2.6.31-02063113-generic [OK]
Trying kernel 2.6.30-02063010-generic [OK]
Trying kernel 2.6.29-02062906-generic [OK]
Trying kernel 2.6.28-02062810-generic [OK]
Trying kernel 2.6.27-020627-generic [OK]
Trying kernel 2.6.26-020626-generic [OK]
Trying kernel 2.6.25-020625-generic [OK]
Trying kernel 2.6.24-020624-generic [OK]

real    26m46.734s
user    140m26.659s
sys     13m26.194s

Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Stephen Hemminger <shemminger@vyatta.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Use GIT_TREE variable.

In gen-stable-release.sh stable git tree path is hardcoded.
Use GIT_TREE environment variable if specified.

Signed-off-by: Bala Shanmugam <bkamatch@qca.qualcomm.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh in synch with next-20120911

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Update 53-pr_fmt.patch

Fix ath5k/debug.c redefinition warning for pr_fmt macro

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: Sync with linux-next 20120906

Refresh patches for next-20120906

Signed-off-by: Ozan Çağlayan <ozancag@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: copy .compat_base on scripts/refresh-compat

Reported-by: Xiong Huang <xiong@qca.qualcomm.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: refresh patches for next-20120904

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>

compat-drivers: backport dma-buf ops begin_cpu_access

This was added as of the original dma-buf implementation via
the 3.4 kernel but through a separate set of patches. Tuck this
into the patch that already deals with dma buf ops.

mcgrof@frijol ~/linux-next (git::master)$ git describe --contains \
fc13020e086bfedf2afb95c91c026d5af1f80107
v3.4-rc1~57^2~3

Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>