Luis R. Rodriguez [Fri, 6 Apr 2012 21:50:16 +0000 (14:50 -0700)]
compat-wireless: backport get_ts_info support
Linux 3.5 will have get_ts_info to support the Precision Time Protocol.
http://linuxptp.sourceforge.net/
http://en.wikipedia.org/wiki/Precision_Time_Protocol
We cannot backport this support given that this introduces
a data structure change on the ethtool_ops.
commit
c8f3a8c31069137fe0100e6920558f1a7487ef3c
Author: Richard Cochran <richardcochran@gmail.com>
Date: Tue Apr 3 22:59:17 2012 +0000
ethtool: Introduce a method for getting time stamping capabilities.
This commit adds a new ethtool ioctl that exposes the SO_TIMESTAMPING
capabilities of a network interface. In addition, user space programs
can use this ioctl to discover the PTP Hardware Clock (PHC) device
associated with the interface.
Since software receive time stamps are handled by the stack, the generic
ethtool code can answer the query correctly in case the MAC or PHY
drivers lack special time stamping features.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
mcgrof@flash ~/devel/compat-wireless (git::master)$ time ckmake; time ckmake
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 70m16.460s
user 204m37.315s
sys 19m56.447s
real 15m35.818s
user 33m10.020s
sys 7m19.779s
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Johannes Berg [Thu, 5 Apr 2012 12:09:13 +0000 (14:09 +0200)]
compat-wireless: add export to commented out config vars
This makes it easier to enable config items correctly.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 3 Apr 2012 18:17:38 +0000 (11:17 -0700)]
compat-wireless: provide URL for code metrics archive
This uses the Google Fusion Tables to keep track of code
metrics accross releases. What would be nice is if someone
adds the code required to upload stats automatically for
each release.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 3 Apr 2012 16:59:16 +0000 (09:59 -0700)]
compat-wireless: skip forcing .config include
We know that .config will exist but it will not exist on
the make clean target. So skip that.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 3 Apr 2012 03:30:10 +0000 (20:30 -0700)]
compat-wireless: fix pr_fmt compile warnings
This took a while to figure out... Thought this was impossible
to fix, but the fix is quite simple: fix this upstream. The
assumption by drivers when upstream is that you should #define
pr_fmt *prior* to including <linux/printk.h> but -- the assumption
is incorrect given that if a subsystem includes printk.h for
anything that the subsystem needs then the module would be defining
the pr_fmt. The proper usage then for pr_fmt then is to always
undef it first, and also include <linux/printk.h>.
When test cross compiling against 2.6.24 - 3.3 I notice that including
<linux/kernel.h> also helps, so do this for now. There may be
another reason why that is required but we can figure that out later.
This patch should be broken down into chunks and submitted upstream.
This shaves off compilation time down from 1 hour to 1/2 hour!
mcgrof@flash ~/devel/compat-wireless (git::master)$ time ckmake
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 29m17.269s
user 75m56.693s
sys 10m48.469s
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 3 Apr 2012 01:03:05 +0000 (18:03 -0700)]
compat-wireless: backport assinging the tty dev to hci dev
This backports assinging the tty device to the hci parent
pointer. The goal was to make the devices link under /sys/
but on older kernels we cannot support this. One alternative
that occurs to me is to have a compat subsystem where userspace
applications can hunt for such newer devices, but this requires
quite a bit of work and consideration.
commit
7f4b2b04c88377af30c022f36c060190182850fb
Author: Andrei Warkentin <andreiw@motorola.com>
Date: Fri Feb 11 17:19:26 2011 -0600
Bluetooth: Make hci a child of the corresponding tty device.
Make /sys/class/bluetooth/hciX a symlink to
path under corresponding tty.
Signed-off-by: Andrei Warkentin <andreiw@motorola.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 2 Apr 2012 21:15:43 +0000 (14:15 -0700)]
compat-wireless: refresh patches for next-
20120402
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 2 Apr 2012 21:08:40 +0000 (14:08 -0700)]
compat-wireless: refresh patches for next-
20120321
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 2 Apr 2012 21:05:03 +0000 (14:05 -0700)]
compat-wireless: refresh patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 2 Apr 2012 21:01:04 +0000 (14:01 -0700)]
compat-wireless: fix patches/16-bluetooth.patch
Remove hunk 2 that modifies drivers/bluetooth/hci_ldisc.c
The hunk as trying to do this:
-+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,37)
- hdev->parent = hu->tty->dev;
-+#endif
This patch makes this not necessary anymore:
commit
68478a9a06c9336fe0e2039a658092ac25935b0d
Author: David Herrmann <dh.herrmann@googlemail.com>
Date: Fri Mar 9 15:53:42 2012 +0100
Bluetooth: Remove redundant hdev->parent field
Turns out SET_HCIDEV_DEV() is defined on older kernels
since v2.6.18-rc1~21 (via
a91f2e39). So all we have to do
is remove this hunk. For newer kernels this will covert
itself accordingly, for older kernels we have the old macro
helper.
mcgrof@tux ~/linux-stable (git::master)$ git describe --contains
a91f2e39
v2.6.18-rc1~21^2~3
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 2 Apr 2012 20:49:21 +0000 (13:49 -0700)]
compat-wireless: add more files to ignore
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 2 Apr 2012 20:47:54 +0000 (13:47 -0700)]
compat-wireless: refresh patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sun, 1 Apr 2012 04:34:15 +0000 (21:34 -0700)]
compat-wireless: fix compilation of ath6kl on 2.6.27
Some drivers rely on some other part of the kernel to include
in.h but on older kernels this proves to not work. Explicitly
include that file. This fixes compilation of ath6kl on 2.6.27.
This patch will be sent upstream though.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sun, 1 Apr 2012 04:08:22 +0000 (21:08 -0700)]
compat-wireless: fix p54 pm ops backport typo again
Really..
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sun, 1 Apr 2012 04:01:54 +0000 (21:01 -0700)]
compat-wireless: fix p54 pm ops backport typo
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sun, 1 Apr 2012 03:31:35 +0000 (20:31 -0700)]
compat-wireless: reorder config includes
We need to include first the kernel config as otherwise
we end up trying to compile drivers which were disabled
for older kernels. This fixes compilation on 2.6.28 where
carl9170 was being compiled although we had disabled it
on config.mk.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sun, 1 Apr 2012 03:31:06 +0000 (20:31 -0700)]
compat-wireless: add pm ops backport for p54
This should fix compiling on 2.6.28.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sat, 31 Mar 2012 02:53:32 +0000 (19:53 -0700)]
compat-wireless: do not use broad export on makefiles
This inspired by Johannes' patch on compat which does the
same thing.
perl -pi -ne 's|^(\s*)CONFIG|$1export CONFIG|' config.mk
And then we also address the autoconf generation.
We shave off compile time by more than 1/2:
Before:
real 0m34.034s
user 0m33.760s
sys 0m10.580s
After:
real 0m16.617s
user 0m22.290s
sys 0m6.180s
This was against 2.6.38.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 20 Mar 2012 01:55:17 +0000 (18:55 -0700)]
compat-wireless: fix compilation of ath6kl on 2.6.34
This fixes compilation of compat-wireless completely on 2.6.34
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 14 Mar 2012 23:50:22 +0000 (16:50 -0700)]
compat-wireless: fix typo in patches/40-netdev-hw-features.patch
This fixes compilation for 2.6.35 for ath6kl.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 14 Mar 2012 23:36:44 +0000 (16:36 -0700)]
compat-wireless: backport netdev->addr_assign_type
You cannot backport assignment of netdev->addr_assign_type
given that its part of the netdev data structure only in future
kernels, so we backport this through a new patch file:
patches/61-netdev-addr_assign_type.patch
mcgrof@tux ~/linux-next (git::master)$ git describe --contains
c1f79426
v2.6.36-rc1~571^2~104
This fixes compilation of the Atheros Ethernet drivers down to 2.6.35.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 14 Mar 2012 23:14:35 +0000 (16:14 -0700)]
compat-wireless: refresh patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 14 Mar 2012 23:13:39 +0000 (16:13 -0700)]
compat-wireless: fix patches/40-netdev-hw-features.patch
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 14 Mar 2012 23:07:46 +0000 (16:07 -0700)]
compat-wireless: fix patches/37-vsnprintk.patch
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 13 Mar 2012 00:43:39 +0000 (17:43 -0700)]
compat-wireless: refresh linux-next-pending patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 13 Mar 2012 00:41:52 +0000 (17:41 -0700)]
compat-wireless: fix linux-next-pending/0002-backport-alx.patch
The last hunk was must be placed in context, patch
is stupid and will only accept hunks in order.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 13 Mar 2012 00:25:25 +0000 (17:25 -0700)]
compat-wireless: backport addr_assign_type usage
The alx driver uses the addr_assign_type variable from
the netdev data structure which was added as of 2.6.35:
mcgrof@tux ~/linux-next (git::master)$ git describe --contains
c1f79426
v2.6.36-rc1~571^2~104
Although compat.git backported dev_hw_addr_random() the patch
in compat.git cannot address usage of addr_assign_type and as
such requires an ifdef around it.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 23:54:36 +0000 (15:54 -0800)]
compat-wireless: fix jobserver
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 22:55:01 +0000 (14:55 -0800)]
compat-wireless: enable mac80211 tracing for kernels >= 2.6.33
Turns out this compiles now that we are in the future, so just
disable tracing for older kernels. This goes unstested though.
Requested-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 22:47:01 +0000 (14:47 -0800)]
compat-wireless: fix patch for disabling ath9k RC
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 22:04:01 +0000 (14:04 -0800)]
compat-wireless: add driver-select option ath9k_ap
Prefer using Minstrel for AP for ath9k. The minstrel
rate control algorihtm has been field tested more than
the ath9k rate control algorithm. To disable ath9k's RC
we renaem the config option to with a CONFIG_COMPAT prefix
and disable it later.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 21:23:55 +0000 (13:23 -0800)]
compat-wireless: refresh patches
This runs: ./scripts/admin-update.sh refresh
This refreshes all hunk offests automatically.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 21:22:43 +0000 (13:22 -0800)]
compat-wireless: fix patches/09-threaded-irq.patch for wl12xx
Hunk offsets differ for wl12xx.h changes. This is now fixed.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 18:57:27 +0000 (10:57 -0800)]
compat-wireless: backport ath6kl hw_features / [set|fix]_features
This requires a manual backport, we throw it into
patches/40-netdev-hw-features.patch following previous
ethernet backport of Atheros Ethernet drivers.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 02:18:21 +0000 (18:18 -0800)]
compat-wireless: add Makefile to PHONY
Debugging Make shows that 18 variations of Makefile are checked
for targets on the Makefile prior to process the first real target.
Stop this nonsense.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 02:02:04 +0000 (18:02 -0800)]
compat-wireless: remove recurssive call on building modules
The recursion call to build modules is not required. It was added
due to the fact that we now included a new file, CONFIG_COMPAT
and we needed it at build time. We do not need the recursive call
given that the kernel will already call the same Makefile later
when building modules. We take advantage of this fact and simply
ensure we build CONFIG_COMPAT when processing our rules.
We do not define a direct target for CONFIG_COMPAT given that
we want to force a regeneration of the file and compat autoconf,
whenever the build environment changes. We still add a dummy
CONFIG_COMPAT target rule to avoid Make looking rules for it.
In order for this build to work we now just need to export
CFLAGS so that the variables defined on the first pass are
available later at module build time.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 01:43:26 +0000 (17:43 -0800)]
compat-wireless: use PWD for file version details
The CFLAGS were using some local files to add some definition
of some variables later used by the compat module. Lets be
careful to use only the locally defined files.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 9 Mar 2012 01:25:37 +0000 (17:25 -0800)]
compat-wireless: use PWD for file definition
Instead of using PWD in random places just use it
to defint he direct file once. This will allow us
to later move variables that use PWD to another
place.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Don Fry [Tue, 6 Mar 2012 23:45:52 +0000 (15:45 -0800)]
compat-wireless: Handle name change from iwlagn to iwlwifi in 3.2
With linux-3.2 the iwlagn driver is renamed iwlwifi. For some kernels
(like 3.1.5) this causes the non-compat iwlagn.ko to try and run against
the compat mac80211 resulting in an oops. This change prevents the
non-compat iwlagn.ko from being loaded.
Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 6 Mar 2012 17:59:44 +0000 (09:59 -0800)]
compat-wireless: refresh patches
Synchs with next-
20120306
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 29 Feb 2012 01:10:13 +0000 (17:10 -0800)]
compat-wireless: disable b44 for Atheros drivers via driver-select
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 29 Feb 2012 00:56:52 +0000 (16:56 -0800)]
compat-wireless: remove -e from depmod on modlib.sh
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 29 Feb 2012 00:48:00 +0000 (16:48 -0800)]
compat-wireless: fix make install
This issue was introduced via this patch:
commit
802ea47e6316aea285f2c30d95763dd7c320eec1
Author: Rick Farina <sidhayn@gmail.com>
Date: Fri Feb 24 21:28:40 2012 -0500
compat-wireless: replace modprobe -l with functional equivilent
Fix the regression by properly using brackets on the
conditional check and also guaranteeing we output something
upon success. Not sure how this could have worked.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 28 Feb 2012 20:10:49 +0000 (12:10 -0800)]
compat-wireless: add alx Ethernet driver to linux-next-pending/
This is now posted so we merge it. We can remove this once
David merges the driver and it gets into linux-next.git.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 28 Feb 2012 20:08:26 +0000 (12:08 -0800)]
compat-wireless: add MAINTAINERS to distribution
This adds the MAINTAINERS file to the distribution.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Felix Fietkau [Mon, 27 Feb 2012 14:51:44 +0000 (15:51 +0100)]
compat-wireless: fix directory copying in admin-update.sh
On my system the trailing slash leads to the directory contents being copied
to the target instead of the full directory, which results in breakage during
build.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Felix Fietkau [Mon, 27 Feb 2012 14:35:46 +0000 (15:35 +0100)]
compat-wireless: use portable #! for bash scripts
Some systems (e.g. FreeBSD) do not have bash in /bin
Based on a patch by Jo-Philipp Wich
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sat, 25 Feb 2012 04:43:56 +0000 (20:43 -0800)]
compat-wireless: export MAKE
This is usefulf to propagate the same MAKE
variable preference over scripts used.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sat, 25 Feb 2012 04:42:39 +0000 (20:42 -0800)]
compat-wireless: fix old kernel checks for 3.x
These ancient checks require a check against the 2 release
as the KERNEL_26SUBLEVEL is only defined when that kernel
is detected.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Rick Farina [Sat, 25 Feb 2012 02:28:40 +0000 (21:28 -0500)]
compat-wireless: replace modprobe -l with functional equivilent
modprobe -l is either deprecated or removed, so let's replace it
testing [ -z "$(DESTDIR)" ] appeared to only be need due to the fact
that *enable was run directly instead of being run with $(DESTDIR)/
Testing for module presence should now work properly on modern distros
and *enable will be run properly with or without $(DESTDIR) being set.
Signed-off-by: Rick Farina <sidhayn@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sat, 25 Feb 2012 01:24:58 +0000 (17:24 -0800)]
compat-wireless: use compat's CONFIG_COMPAT_KERNEL_* variables
compat already deals with defining CONFIG_COMPAT_KERNEL_* variables
for us for both makefiles and C / header files. The C / header file
usage was fixed in the last commit, this removes internal usage we
had for the Makefiles in preference over the same mechanism.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sat, 25 Feb 2012 00:55:32 +0000 (16:55 -0800)]
compat-wireless: start using CONFIG_COMPAT_KERNEL_ in code
This lets use the defined CONFIG_COMPAT_KERNEL_* in C code and
header files. Next step is to remove the Makefile hackery to
defines CONFIG_COMPAT_KERNEL_* stuff as we have a generated
file for us now.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Sat, 25 Feb 2012 01:23:53 +0000 (17:23 -0800)]
compat-wireless: fix old checks on KERNEL_SUBLEVEL
KERNEL_SUBLEVEL used to be used on config.mk for th 2.6.x series
but after 3.x support was added the variable should have been
changed to KERNEL_26SUBLEVEL.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Rick Farina [Sun, 19 Feb 2012 15:49:52 +0000 (10:49 -0500)]
cleanup modprobe calls
There are a lot of needless calls to "modprobe -l <drivername>"
None of this is needed, and it errors when modprobe -l is removed
(such as when KMOD replaces modutils). Since modprobe -l is already
legacy, let's just remove it.
Signed-off-by: Rick Farina <sidhayn@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Thu, 23 Feb 2012 23:36:23 +0000 (15:36 -0800)]
compat-wireless: refresh patches for backport alx
alx is not yet upstream but it will be soon I hope.
This updates the backport of it.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Thu, 23 Feb 2012 23:21:48 +0000 (15:21 -0800)]
compat-wireless: refresh patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Thu, 23 Feb 2012 23:19:57 +0000 (15:19 -0800)]
compat-wireless: fix patches/24-pcmcia.patch
Some hunk offsets were off.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Thu, 23 Feb 2012 23:15:27 +0000 (15:15 -0800)]
compat-wireless: fix patches/16-bluetooth.patch
Some context lines were removed from the upstream driver.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Thu, 23 Feb 2012 23:01:48 +0000 (15:01 -0800)]
compat-wireless: fix 08-rename-config-options.patch
Hunk conflicts due to changes on the Makefile for the wl1251 driver.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 17 Feb 2012 23:56:07 +0000 (15:56 -0800)]
compat-wireless: refresh patches
This synchs with next-
20120217.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luciano Coelho [Fri, 17 Feb 2012 09:49:00 +0000 (11:49 +0200)]
compat-wireless: update bluetooth patch to latest wireless-testing
There is a small change in the bt_sock_unregister() function in
af_bluetooth (an extra if) that cause the patch not to apply anymore.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
John W. Linville [Wed, 15 Feb 2012 21:37:52 +0000 (16:37 -0500)]
compat-wireless: remove olpc_ec_wakeup_* calls for pre-3.1 kernels
This section of the libertas driver calls functions that simply don't
exist before the release of 3.1. This code in question was an addition,
not a change from any existing code. It is safe to simply remove it for
older kernels.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 1 Feb 2012 03:05:56 +0000 (19:05 -0800)]
compat-wireless: add alx driver and backport
This adds backport support for the alx driver.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 10 Feb 2012 20:56:34 +0000 (12:56 -0800)]
compat-wireless: refresh patches with next-
20120210
Synchs with next-
20120210
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 10 Feb 2012 20:51:10 +0000 (12:51 -0800)]
compat-wireless: remove patches/50-function-namespace.patch
This is now upstream. Boy that was fast.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 10 Feb 2012 20:45:13 +0000 (12:45 -0800)]
compat-wireless: fix patches/44-deactivate-mac80211-tracing.patch
Tracing stuff is now compiled-in if tracing is enabled in your
build, but we keep it disabled regardless for compat.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 8 Feb 2012 04:59:55 +0000 (20:59 -0800)]
compat-wireless: rename path_lookup() to mpath_lookup()
This is required for compiling compat-wireless down to 2.6.24.
Don't ask me why, I just wanted to get this done. This patch
will be sent upstream.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Rick Farina [Mon, 6 Feb 2012 15:42:16 +0000 (10:42 -0500)]
compat-wireless: add missing =y to CONFIG_RT2X00_LIB_DEBUGFS
When editing config.mk with a sed script the missing =y on this line
causes issue. Since this line is commented out by default this means no
changes to the default behavior and sed scripts will be able to edit the
config.mk file much more easily.
Signed-of-By: Rick Farina <sidhayn@gmail.com>
Rick Farina [Mon, 6 Feb 2012 15:38:14 +0000 (10:38 -0500)]
compat-wireless: Calling make directly is not permitted
make was being called directly, fix to $(MAKE)
Signed-of-By: Rick Farina <sidhayn@gmail.com>
Luis R. Rodriguez [Wed, 1 Feb 2012 03:36:46 +0000 (19:36 -0800)]
compat-wireless: add Ethernet drivers to driver-select
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 1 Feb 2012 02:24:59 +0000 (18:24 -0800)]
compat-wireless: update language on scripts/admin-update.sh
Use the actual git tree name. The term "linux-next.git based"
does more justice than "bleeding edge based". It also helps educate
users / developers about the source of the code.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 1 Feb 2012 00:08:46 +0000 (16:08 -0800)]
compat-wireless: refresh patches
This applies on top of next-
20120131.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 1 Feb 2012 00:03:18 +0000 (16:03 -0800)]
compat-wireless: fix patches/99-change-makefiles.patch
CONFIG_USB_NET_QMI_WWAN was added and therefore the
respective Makefile, drivers/net/usb/Makefile, needs
to be adjusted to remove that from our list of targets.
If we user our own generated Makefiles for drivers the
advantage is that we would not face this issue when upstream
gets a new driver that our hunk needs to ignore but -- the
disadvantage is that any changes upstream would require a
respective change into compat-wireless. Driver Makefiles
though typically are just copied as-is and never modified.
Its subsystem Makefiles that we tend to have to patch on
this file.
One option would be to somehow annotate upstream targets as
compat-wireless maintained so that we can script out what
targets to leave in.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 31 Jan 2012 23:39:12 +0000 (15:39 -0800)]
compat-wireles: remove ath9k packet log crap patch
This has been sitting in crap/ directory for ages. Time
to go. Too bad. Someone will have to get this upstream.
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 26 Jan 2012 22:55:44 +0000 (23:55 +0100)]
compat-wireless: fix some config options
Adds some missing config options interceded in the last time and remove
some old now unused config options.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Sat, 21 Jan 2012 13:43:36 +0000 (14:43 +0100)]
compat-wireless: rename path_lookup
path_lookup in net/mac80211/mesh_pathtbl.c consolidates with
path_lookup in include/linux/namei.h used in old kernel versions.
This should also go into compat-wireless-stable-3.3.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Sat, 21 Jan 2012 13:43:35 +0000 (14:43 +0100)]
compat-wireless: use skb_get_queue_mapping() in b43 and b43legacy
Use skb_get_queue_mapping() for getting the queue_mapping member of
skb. Some old kernels do not have the member queue_mapping with all
configuration options, but this function always returns something.
This should also go into compat-wireless-stable-3.3.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Sat, 21 Jan 2012 13:43:34 +0000 (14:43 +0100)]
compat-wireless: backport dev_pm_ops for bcma
bcma now also uses dev_pm_ops which does not work with older kernel
versions. Just backport it like it was done for the other drivers.
This should also go into compat-wireless-stable-3.3.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Eliad Peller [Mon, 23 Jan 2012 17:02:34 +0000 (19:02 +0200)]
compat-wireless: fix wl12xx driver-select
wl12xx is a possible choice for both group drivers and
singular modules (making it impossible to build only
the wl12xx module). change the group name to "ti" instead.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 20 Dec 2011 02:33:42 +0000 (18:33 -0800)]
compat-wireless: use linux-stable instead of linux-2.6-allstable
I used to use hpa's linux-2.6-allstable.git tree to make
stable releases of compat-wireless. This allowed me to
make updates to the stable releases whenever a 2.6.x.y
extraversion got out -- but also to make updates whenever
Linus updated his own tree so I can make RC releases
based on the latest RC release. Greg has taken over of
that tree and its now called linux-stable.git. What I am
seeing though is that the linux-stable git tree now seems
to not update as often to Linus' tree. This is fine as all
I have to do is add Linus' tree as a remote and fetch
those objects and then reset the tree to say, v3.2-rc6
and that'll get what Linus' tree looks like today.
This patch just changes the path where we pick up stable
releases from but also documents this change in case
anyone that wants to make stable release must also add
Linus' tree as a remote and fetch those objects.
Cc: greg@kroah.com
CC: kvalo@qca.qualcomm.com
Cc: linville@tuxdriver.com
Cc: senthilb@qca.qualcomm.com
Cc: johannes@sipsolutions.net
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
John W. Linville [Tue, 3 Jan 2012 15:28:36 +0000 (10:28 -0500)]
Enable CONFIG_RT2800{PCI,USB}_RT53XX
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luciano Coelho [Wed, 18 Jan 2012 08:13:41 +0000 (10:13 +0200)]
compat-wireless: check if modprobe exists before using it
If you're cross-compiling compat-wireless on a different host, you
can't reliably use modprobe. The host system may not even have
modprobe, so it doesn't make sense to run modprobe, unless you're
installing the modules on the target machine.
The modprobe command was being used unconditionally to set the MADWIFI
and OLD_IWL variables. This commit adds a check to see whether
modprobe exists before trying to use. It also adds a dependency on
the modprobe commands to all recipes that rely on it, so we fail more
gracefully if modprobe doesn't exist in the system.
Reported by: William List <wlist@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Dominique Martinet [Tue, 17 Jan 2012 00:08:12 +0000 (01:08 +0100)]
compat-wireless: build: Throw error if space in path
Because of make limitations, the build will fail if there are spaces in
the path leading to the compat-wireless directory.
I tried to make it work anyway for a while and, while it probably is
possible, it would involve rewriting part of kbuild and well...
This at least throws up nicely if there is a space in the path.
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Mon, 16 Jan 2012 14:35:12 +0000 (06:35 -0800)]
compat-wirless: refresh patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 10 Jan 2012 00:56:33 +0000 (16:56 -0800)]
compat-wireless: refresh patches for next-
20120109
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Wed, 28 Dec 2011 15:38:28 +0000 (07:38 -0800)]
compat-wireless: refresh patches
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 23 Dec 2011 15:12:52 +0000 (07:12 -0800)]
compat-wireless: clarify patches/46-use_other_workqueue.patch
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Fri, 23 Dec 2011 15:10:08 +0000 (07:10 -0800)]
compat-wireless: clarify patches/47-no_trans_start_on_netdev_queue.patch
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 22 Dec 2011 20:47:30 +0000 (21:47 +0100)]
compat-wireless: no trans_start on netdev_queue
struct netdev_queue does not have the attribute trans_start in kernel <
2.6.31. trans_start on struct net_device does the same on older kernels.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 22 Dec 2011 20:47:29 +0000 (21:47 +0100)]
compat-wireless: use other workqueue
Old kernel versions do not support WQ_HIGHPRI and WQ_MEM_RECLAIM so we
should use create_singlethread_workqueue() which was used at this
position before.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 22 Dec 2011 20:47:28 +0000 (21:47 +0100)]
compat-wireless: make patches apply again
* sk_add_backlog() is no longer used in net/bluetooth/l2cap_core.c so
the patch is not needed any more.
* ath6k got some now functions and the patches did not apply any more.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 22 Dec 2011 20:47:27 +0000 (21:47 +0100)]
compat-wireless: CONFIG_BT_L2CAP and CONFIG_BT_SCO removed
The config options CONFIG_BT_L2CAP and CONFIG_BT_SCO were removed from
the kernel and this patch removes them from compat-wireless.
commit
f1e91e1640d808d332498a6b09b2bcd01462eff9
Author: Ulisses Furquim <ulisses@profusion.mobi>
Date: Wed Dec 21 01:32:09 2011 -0200
Bluetooth: Always compile SCO and L2CAP in Bluetooth Core
The handling of SCO audio links and the L2CAP protocol are essential to
any system with Bluetooth thus are always compiled in from now on.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 22 Dec 2011 20:47:26 +0000 (21:47 +0100)]
compat-wireless: remove platform_device_id for kernel < 2.6.30
Kernel < 2.6.30 do not support struct platform_device_id, so remove it
for them and hope that the probing will reject an invalid device.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Hauke Mehrtens [Thu, 22 Dec 2011 20:47:25 +0000 (21:47 +0100)]
compat-wireless: compile fix for commit patches/09-threaded-irq.patch
This commit had some compile problem with kernel < 2.6.32
commit
c486a4fcc5ddfb3232edfd0489fe63ffc1fa8aea
Author: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Date: Tue Dec 13 10:40:55 2011 -0800
compat-wireless: fix patches/09-threaded-irq.patch
There was a semicolon missing at the end of a line and wl1271_hardirq
has to be renamed to wl12xx_hardirq.
This is now compile tested with all maijor kernel version from 2.6.27.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <mcgrof@frijolero.org>
Luis R. Rodriguez [Tue, 13 Dec 2011 19:33:32 +0000 (11:33 -0800)]
compat-wireless: copy driver files only if present
This modifies admin-update.sh to only copy files over
from a driver target directory if files are found. This
should avoid a bunch of complaints of copying files
from directories that are just placeholders for other
folders, as happens with drivers/net/wireless/brcm80211/
which has folders a Makefile, Kconfig but no files there
itself:
brcmfmac/
brcmsmac/
brcmutil/
include
Kconfig
Makefile
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Luis R. Rodriguez [Tue, 13 Dec 2011 18:59:55 +0000 (10:59 -0800)]
compat-wireless: refresh patches
automatic refresh
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Luis R. Rodriguez [Tue, 13 Dec 2011 18:40:55 +0000 (10:40 -0800)]
compat-wireless: fix patches/09-threaded-irq.patch
This fixes patches/09-threaded-irq.patch hunk breakage
caused due to changes in the upstream wireless-testing
commit
a390e85:
Author: Felipe Balbi <balbi@ti.com>
Date: Thu Oct 6 10:07:44 2011 +0300
wl12xx: move common init code from bus modules to main
Move all common parts from sdio.c and spi.c to main.c, since they now
can be handled as part of the platform driver.
Signed-off-by: Felipe Balbi <balbi@ti.com>
[forward-ported, cleaned-up and rephrased commit message]
[added a bunch of fixes and a new pdata element]
[moved some new code into main.c as well]
Signed-off-by: Luciano Coelho <coelho@ti.com>
This goes not compile tested, apart from *all* the things
a390e85 changes one minor change was instead of using
DRIVER_NAME in the request_threaded_irq() it now uses
pdev->name, and pdev is struct platform_device. Too lazy
to git checkout v2.6.30 to test compile against that
kernel but at least:
git blame include/linux/platform_device.h v2.6.30
seems to show that the pdev->name char was added by
Russell King via
bbbf508d on v2.6.15 so this should
compile down to that kernel.
mcgrof@tux ~/linux-2.6-allstable (git::master)$ git describe --contains
bbbf508d
v2.6.15-rc1~59^2~6^2~2
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Eliad Peller [Mon, 12 Dec 2011 17:14:16 +0000 (19:14 +0200)]
compat-wireless: fix patches/08-rename-config-options.patch
wl12xx_sdio_test module was removed
Signed-off-by: Eliad Peller <eliad@wizery.com>
Luis R. Rodriguez [Mon, 5 Dec 2011 19:59:13 +0000 (11:59 -0800)]
compat-wireless: refresh patches
Synchs with linux-next tag
20111205
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Hauke Mehrtens [Sat, 3 Dec 2011 14:46:25 +0000 (15:46 +0100)]
compat-wireless deactivate mac80211 tracing for old kernels
Activating tracing on kernel <= 2.6.32 causes build errors.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>