Age | Commit message (Collapse) | Author |
|
Accidentally, the patch to remove the device removal code from earlier
kernels was sent from an intermediate version of the code, so the
patch doesn't apply properly. Fix it by updating it to the version
that was actually upstreamed.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
With commit 3d6ce86ee794 ("drivers: remove force dma flag from buses")
in kernel 4.18 of_dma_configure() got the extra parameter force_dma.
This is backported by setting the force_dma member which was added in
commit d89e2378a97f ("drivers: flag buses which demand DMA
configuration") in kernel 4.15.
In older kernel versions we just ignore this parameter as the
of_dma_configure() function internally takes care of setting this
correctly.
Since commit 7b07cbefb68d ("iommu: of: Handle IOMMU lookup failure with
deferred probing or error") in kernel 4.12 this function returns int and
not void any more.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This reverts commit 1b4e6ad4d927c7998fa578f2d29e0c4ec01a9217.
Said commit broke b43 on older kernels (<3.16).
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This is just a forward include, but earlier versions don't include
linux/types.h despite relying on it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The getrawmonotonic64() function that is used by the
ktime_get_raw_ts64() backport was only introduced in 3.19. To fix
compilation with earlier kernels, do the convertion from
getrawmonotonic() manually if the kernel is < 3.19.
Additionally, add timespec_to_timespec64() that we need for this
conversion (and which was only introduced in 3.17).
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
On 4.16, the coredump method returns int rather than void. It
would be possible to fix that up, but so far I'm too lazy.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This driver has to include linux/kmemleak.h to get kmemleak_not_leak in
kernel 4.19. I haven't seen any problem when compiling against multiple
other kernels, but didn't test all the 3.X versions.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The netdev destructor spatch did not apply to the wil6210 driver any
more, because there the calls were done in a different order compared to
the spatch. Add an additional rule which applies when the order is
changed.
Fixes: 30a378636e47 ("backports: speed up netdev destructor spatch by two orders of magnitude")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Kernel 4.16 added a new .coredump member to the driver structure which
can be implemented by drivers. Remove this member on older kernel
versions as they do not support this feature.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
__skb_pad() was only added with kernel 4.13.
The free_on_error parameter of __skb_pad() is set to constant true,
which is also the setting used by skb_pad() so it should be save to use
skb_pad() instead.
This patch should be added to the mainline Linux kernel.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The hostap and atmel wireless driver now make use of the
proc_create_single_data() function which can not easily be backported.
As these drivers are not used too often just deactivate them on older
kernel versions for now.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The WIL6210 driver now uses pci_alloc_irq_vectors() which is only
available in kernel >= 4.8, only build this driver against such kernel
versions.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
[backporting the function doesn't seem too hard, but unless
somebody complains we're too lazy]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
When RSI_COEX is activated the driver wants to link against the
rsi_bt_ops symbol from the bluetooth driver btrsi. This driver was only
added in kernel 4.17. Deactivate bluetooth coexistence support for older
kernel versions.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This is used by the rtl8723bs driver. The implementation is completely
in this header file, so it is save to copy it
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
I am not aware of any user of the NFC coded in backports, at least I
haven't seen any patches regarding this. Remove this code to reduce the
efforts needed in the backports project.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Use the versions shipped with the kernel instead.
bcma and ssb didn't change much in the last few years, most of the
changes were related to code only used on embedded devices. Removing
this reduces the number of parts we have to backport. b43, b43legacy and
brcmsmac work well with the in kernel version.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
In kernel v4.19-rc1, the activate op in struct led_trigger, changed
from void to int. To solve this, add a semantic patch to insert a
wrapper function that returns void and calls the function that returns
int.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
[use <+... ...+> as suggested by Hauke]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Replace the existing skb->xmit_more semantic patch rule with a more
generic one that uses a helper inline function.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The u64_stats_init() macro was turned into a static inline function in
v4.2 to prevent warnings that happen when its argument is not used
elsewhere.
Since we started hitting these warnings, undefine the macro and define
the static inline to prevent them.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
With kernels < 3.17, we need to include hrtimer.h to get ktime_divns()
for the ktime_get_boottime_seconds() backport. But we can't just
include htrtimer.h in the backport timekeeping.h because we run into
some cyclical inclusions that cause problems. To solve that make
ktime_get_boottime_seconds() non-inline and add it to a new
backport-4.18.c file.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
In kernels earlier than 3.18, the raw_spin_lock defininition when
CONFIG_DEBUG_LOCK_ALLOC is not enabled, was dropping one of the
parameters, which caused a few "unused variable" warnings at places
where this argument was only used in the call to
raw_spin_lock_nested(). Take the latest definition which evaluates
the argument to avoid compiler warnings.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
We were including leds.h twice in backport-4.5.c and one of them was
very early, which caused inclusion conflicts at least with < 3.16
kernels.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This was introduced in 4.18 and is needed by rhashtable.h, that we
copy. So copy it as well.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
For some reason, sometimes when the git logs contain non-ascii
characters, we get some conversion errors. This only happens if one
of the strings we use is explicitly unicode, otherwise everything is
fine. The only string we use that is of unicode type is the tree
names that comes from the configuration sections. To avoid issues,
convert the strings we get from the config file into ascii before
using it.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The iwlwifi driver has a workaround for some PCI bugs that require it
to be removed and reinserted. Unfortunately, this doesn'w work on
kernels < 3.14, so ignore the actual work function in that case and
print out a message instead.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
[rename patch to have a number prefix]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This function was introduced in v4.6 and now the iwlwifi driver uses
it. Add the function for kernels older than v4.6.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Allow using wait_event_killable_timeout function in kernel
older then 4.13
Signed-off-by: Shahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
At least in some older kernel versions (e.g. 3.10), the tcp.h header
file is not implicitly included from other headers if CONFIG_IPV6 is
not set, so we need to include it in backport-4.4.c. Also, there is
an IPv6 structure that is used, so we should ifdef it away when IPv6
is not enabled.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
wireless-drivers-next.git commit cb18e2e9ec71 (brcmfmac: fix wrong
strnchr usage) used strreplace(), which is not available in kernel prior
to v4.2. Before this commit proprgates to linux-stable.git, add
strreplace() to avoid gentree failure.
Signed-off-by: Winnie Chang <winnie.chang@cypress.com>
Signed-off-by: Chi-Hsien Lin <chi-hsien.lin@cypress.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
In newer kernels, calling cc-option will use KBUILD_CFLAGS,
which itself gets subdir-ccflags-y added to it. Thus, we get
it referencing itself, since nothing here is evaluated until
KBUILD_CFLAGS is used: KBUILD_CFLAGS contains subdir-ccflags-y,
that in turn contains cc-option call, containing KBUILD_CFLAGS.
To avoid that, use := assignment to subdir-ccflags-y, meaning
the cc-option call is done right away and we don't end up with
KBUILD_CFLAGS referencing itself.
Note that this is the first assignment to subdir-ccflags-y, so
the += was never necessary.
This fixes build of the driver on kernels 4.15 and higher.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
ktime_get_raw_ts64() uses the old function which was renamed.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
[fix typo in commit message]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
These functions are currently used by the usbnet driver.
These functions were introduced in upstream commit 2695578b896ae ("net:
usbnet: fix potential deadlock on 32bit hosts")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This is used by the mac80211_hwsim driver. The code was copied from the
mainline kernel.
This was introduced in mainline kernel commit c3ab2b4ec8f7 ("net: ipv4:
Add extack messages for route add failures")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The select queue API was changed in upstream commit 4f49dec9075a ("net:
allow ndo_select_queue to pass netdev"). This patch adds a new version
of the API for kernel < 4.19
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This adds the kvzalloc() function and allows it to remove the
lib-rhashtable.patch patch.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
[fix commit message kzalloc -> kvzalloc]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This removes parts which were only used by subsystems which are not
included in backports any more, for example media, bluetooth and
Ethernet.
The patches which are removed in this commit are not applied to the
kernel tree anyway with the default configuration because non of the
files get copied. The freezer, media, regulator and sound parts were
only used by the already removed media drivers. The flow dissector file
is not copied any more.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
As reported by Hauke, and analyzed by Julia, the netdev destructor
spatch is really slow because <... ...> means 0 or more hits, and
thus the spatch has to be applied on all files. However, we don't
really need it to be this way, so use <+... ...+> which allows it
to reject almost all files immediately, cutting down the processing
time by more than two orders of magnitude as it only has to touch a
handful of files, rather than all.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
The 1.0.6 release was actually not compatible with the
attribute name syntax we use, and the confusion around
1.0.7 finally got cleared up, and that should work.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
crypto_memneq has been backported in 3.12.59 and 3.10.106, it
shouldn't be built for these kernels.
Signed-off-by: Alban Bedel <alban.bedel@avionic-design.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
It is used by new mt76 changes
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Obsoleted by kvmalloc/kvmalloc_array backport
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Overflow macros are useful for further backports
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
Function memcmp is implicit declaration on kernel 4.9.88 ARM platform.
backport-include/keys/asymmetric-type.h uses function memcmp but its
caller compat/verification/verify.c and all its related header files
do not include string.h. The header file is usually included from
arch/arm/include/asm, but in this configuration, it doesn't. We need
to be safe and insure string.h is there.
Signed-off-by: Winnie Chang <winnie.chang@cypress.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|