1   3.0.101             [  OK  ]
2   3.1.10              [  OK  ]
3   3.2.72              [  OK  ]
4   3.3.8               [  OK  ]
5   3.4.110             [  OK  ]
6   3.5.7               [  OK  ]
7   3.6.11              [  OK  ]
8   3.7.10              [  OK  ]
9   3.8.13              [  OK  ]
10  3.9.11              [  OK  ]
11  3.10.92             [  OK  ]
12  3.11.10             [  OK  ]
13  3.12.49             [  OK  ]
14  3.13.11             [  OK  ]
15  3.14.56             [  OK  ]
16  3.15.10             [  OK  ]
17  3.16.7              [  OK  ]
18  3.17.8              [  OK  ]
19  3.18.23             [  OK  ]
20  3.19.8              [  OK  ]
21  4.0.9               [  OK  ]
22  4.1.12              [  OK  ]
23  4.2.5               [  OK  ]
24  4.3-rc7             [  OK  ]

Bigger, manual changes were needed on
patches/collateral-evolutions/media/0002-no_dmabuf/v4l2.patch

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This adds the new driver CPTCFG_RTL8XXXU and removes the config option
for CPTCFG_B43_PCMCIA. pcmcia support in b43 is now done in ssb.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

wpan_dev_hard_header() accesses ieee802154_ptr which is not available
on older kernel versions, remove this function on such kernel versions.
We do not compile cfg802154 on such kernel versions any way, but the
6lowpan stack uses this header file and we need 6lowpan for bluetooth.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Normally this forward declaration is coming from netdevice.h, but it
was only added in more recent kernel versions there. This patch fixes
the build with older kernel versions.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

When CONFIG_OF is not set kernel < 3.5 does not provide an empty
implementation of this function, more recent kernels do so.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

The return type changed from void to int to return some error code.
Some code is now checking the return value which fails on older kernel
versions when the function returns void.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

inet_frags_uninit_net() was added in commit 1d6119ba

 "net: fix
percpu memory leaks" and is used by the 6lowpan part of ieee802154 in
backports.

percpu_counter_destroy() was added in kernel 3.9, so do not add this
function on older kernel versions. This is currently only used by the
6lowpan part of ieee802154 which is only activated for kernel >= 3.18.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

saa7164-encoder.c uses clamp() with constant values which fails on
kernel version < 3.18 because the code wants to creates values of the
same type and wants to assign variable values to them. The new code
works with this kind of usage, use the new code also on older kernel
versions to fix this problem.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Upstream commit 3b47d303

 "net: gro: add a per device gro flush timer"
added napi_complete_done(). On older kernel versions we will just use
the old napi_complete() function, this should work without any
problems, we only do not get the new improvements which need a more
recent net subsystem.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This driver uses of_find_property() and devm_gpio_free() which are not
available in this kernel versions.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This driver now uses the gpio consumer api.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This driver makes use of the device_property..() functions which were
introduces in kernel 3.18 with commit b31384fa

 "Driver core:
Unified device properties interface for platform firmware"

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

The realtek drivers were reorganized and are now located under realtek
directory.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

commit 2e98a32a

 ("backport: add
of_property_read_u64_array()") added a call
to of_find_property_value_of_size(), which is a
static function, causing compilation error.

Implement it locally as well.

Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
[remove of_find_property_value_of_size() from backport-3.10.c]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

1   3.0.101             [  OK  ]
2   3.1.10              [  OK  ]
3   3.2.72              [  OK  ]
4   3.3.8               [  OK  ]
5   3.4.110             [  OK  ]
6   3.5.7               [  OK  ]
7   3.6.11              [  OK  ]
8   3.7.10              [  OK  ]
9   3.8.13              [  OK  ]
10  3.9.11              [  OK  ]
11  3.10.92             [  OK  ]
12  3.11.10             [  OK  ]
13  3.12.49             [  OK  ]
14  3.13.11             [  OK  ]
15  3.14.56             [  OK  ]
16  3.15.10             [  OK  ]
17  3.16.7              [  OK  ]
18  3.17.8              [  OK  ]
19  3.18.23             [  OK  ]
20  3.19.8              [  OK  ]
21  4.0.9               [  OK  ]
22  4.1.12              [  OK  ]
23  4.2.5               [  OK  ]
24  4.3-rc7             [  OK  ]

manual changes done to the following patches:
 * patches/collateral-evolutions/network/0059-uuid-in-mei/nfc.patch
 * patches/collateral-evolutions/network/0013-fix-makefile-includes/brcmfmac.patch
 * patches/collateral-evolutions/network/0062-acpi-gpio-remove/hci_bcm.patch

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

VIDEO_MT9P031 uses the GPIO consumer and should only be build on
kernel version >= 3.13

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This function was added in commit b31384fa

 "Driver core: Unified
device properties interface for platform firmware".

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Instead of using the kernel version to check if we have to add a
backport for ACPI_COMPANION check for it directly. RedHat 7 backported
this change to their kernel and writes a warning without this change.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

These functions were added in commit c648a013

 "netlink: add nla_get
for le32 and le64"

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This was added to the mainline kernel in commit ac4f832eb2 "mm,
page_alloc: distinguish between being unable to sleep, unwilling to
sleep and avoiding waking kswapd". Instead of __GFP_KSWAPD_RECLAIM the
old mm code did not got any flag.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

We also have to add the prototype of ktime_t ktime_get(), because
including linux/hrtimer.h caused some strange build errors.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Commit d49dc5e7

 "mei: bus: use mei_cldev_ prefix for the API
functions" changed the prefix for these functions. On older kernel
versions they are still exported under their old name, name our code
use the old names on older kernel versions.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This include is missing in kernel version <= 3.5.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Commit 38cf595b

 "ipv6: remove unused neigh parameter from ndisc
functions" removed one parameter from the callback function
ndisc_send_na(). This semantic patch makes the code add NULL for that
parameter on older kernel versions. Null was passed in most of the
callers before.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Since commit 84b00607

 ("mac80211: use ktime_get_seconds")
mac80211 uses ktime_get_seconds(). This patch provide a backport
using ktime_get_ts() for it.

[commit message and time64_t taken from Arend's patch]

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Handle the recent genl_notify() prototype change.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Currently, a Makefile line like this:

    obj-$(CONFIG_IPW2100) += ipw2x00/

will get mangled into

    #obj-$(CONFIG_IPW2100+= ipw2x00/

when disabled by the gentree script. Fix the regular
expression to contain the entire match as the first
group, so it gets all preserved when replaced later
by "#\1" using expression.sub().

This then makes it come out as

    #obj-$(CONFIG_IPW2100) += ipw2x00/

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

The prototype for this function is changing upstream, so backport
a copy of the new one that has a bool * argument instead of u32 *.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
[debugfs_{read,write}_file_bool() is in kernel 4.3, do not add it there]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Since commit 69b5f4da

 ("bcma: add support for population subnodes
also when build as module") the bcma driver uses the function
of_platform_default_populate(). This patch provides the stub function.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

1   3.0.101             [  OK  ]
2   3.1.10              [  OK  ]
3   3.2.72              [  OK  ]
4   3.3.8               [  OK  ]
5   3.4.110             [  OK  ]
6   3.5.7               [  OK  ]
7   3.6.11              [  OK  ]
8   3.7.10              [  OK  ]
9   3.8.13              [  OK  ]
10  3.9.11              [  OK  ]
11  3.10.92             [  OK  ]
12  3.11.10             [  OK  ]
13  3.12.49             [  OK  ]
14  3.13.11             [  OK  ]
15  3.14.56             [  OK  ]
16  3.15.10             [  OK  ]
17  3.16.7              [  OK  ]
18  3.17.8              [  OK  ]
19  3.18.23             [  OK  ]
20  3.19.8              [  OK  ]
21  4.0.9               [  OK  ]
22  4.1.12              [  OK  ]
23  4.2.5               [  OK  ]
24  4.3-rc7             [  OK  ]

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

The following upstream commit added a new header files used by usbnet:

commit c40a2c88


Author: Oliver Neukum <oneukum@suse.com>
Date:   Mon Sep 7 16:05:38 2015 +0200

    CDC: common parser for extra headers

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

In order to make it possible to use GIT_TRACE, ignore stderr
in the bpgit scripting.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

The upstream commit that added it was backported in 4.1.9

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

1   3.0.101             [  OK  ]
2   3.1.10              [  OK  ]
3   3.2.71              [  OK  ]
4   3.3.8               [  OK  ]
5   3.4.108             [  OK  ]
6   3.5.7               [  OK  ]
7   3.6.11              [  OK  ]
8   3.7.10              [  OK  ]
9   3.8.13              [  OK  ]
10  3.9.11              [  OK  ]
11  3.10.88             [  OK  ]
12  3.11.10             [  OK  ]
13  3.12.47             [  OK  ]
14  3.13.11             [  OK  ]
15  3.14.52             [  OK  ]
16  3.15.10             [  OK  ]
17  3.16.7              [  OK  ]
18  3.17.8              [  OK  ]
19  3.18.21             [  OK  ]
20  3.19.8              [  OK  ]
21  4.0.9               [  OK  ]
22  4.1.7               [  OK  ]
23  4.2.0               [  OK  ]

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

This removes the old driver IEEE802154_FAKEHARD and adds two new
drivers.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

DVB_TDA10071 now uses regmap_bulk_write() which was added in kernel 3.4.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

VIDEO_TC358743 now uses the GPIO driver subsystem which is not
backported, make it depend on kernel 3.13.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>