Age | Commit message (Collapse) | Author | Files | Lines |
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Joerg Albert <jal2@gmx.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
This allows compilation of compat-wireless-old on 2.6.22 kernel. I suspect
some drivers don't work despite they compile, though.
Signed-off-by: Jiri Benc <jbenc@suse.cz>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
This fixes a kernel crash that was caused by using a wrong pointer for
ieee80211_local dereference in mq compat implementation.
Signed-off-by: Jiri Benc <jbenc@suse.cz>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
using git will use .compat_autoconf_ but the tarball uses the right
version. Whatever, it builds now :)
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
ath9: Fix ath_rx_flush_tid() for IRQs disabled kernel warning message.
We'll start backport on top of this.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
We fork compat-wireless for older kernels due to the new
TX multiqueue (MQ) changes. These new TX MQ changes are
painful to backport so instead of leaving everyone dry
we'll fork it, add new drivers as we can and start fresh
a new tree for >= 2.6.27.
This adds to git history drivers/ net/ and include/
with all the compat.diff changes applied. Note that
changes to compat.[ch] and mq_compat.h can now go directly
against their files based on the tree location, the compat/
directory is being removed as this is a complete fork
*with the tree*.
Patches for drivers can now be sent directly to this tree.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Kernels 2.6.23 or above didn't get these symbols:
device_create_vargs
device_create_drvdata
register_rfkill_notifier
unregister_rfkill_notifier
rfkill_force_state
Also fix a related comment typo found checking compat/compat.h .
Signed-off-by: Gabriel de Perthuis <Gabriel.de-Perthuis@laPoste.net>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
crashes right now on IEEE80211_DEV_TO_SUB_IF() in ieee80211_i.h.
We hit the BUG_ON:
BUG_ON(!local || local->mdev == dev);
Once this bug is fixed we can re-enable compilation for 2.6.22.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
using the hacked up MQ support. Loading mac80211 now works
on 2.6.22.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
sorry. This adds:
* uintptr_t
* __cold
* device_create_release()
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Added device_create_drvdata() and friends to support mac80211_hwsim for
kernels older then 2.6.26.
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
ssh://mcgrof@master.kernel.org/pub/scm/linux/kernel/git/mcgrof/compat-wireless-2.6
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
assign the callback.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
won't call notify_rfkill_state_change() so we may want to just replace
rfkill code with compat-wireless to get it to work just as well.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
variable.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
compat section for kernels <= 2.6.26.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
in the target kernel. In compat-wireless everything is
compiled as module
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
RFKILL_STATE_SOFT_BLOCKED
RFKILL_STATE_UNBLOCKED
and add the new one:
RFKILL_STATE_HARD_BLOCKED
This one won't be handled by older kernels unless we later add
support for handling in compat.c
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Also CONFIG_MAC80211_QOS won't be in kernels until 2.6.27, but
it does have its requirements. It requires *both* CONFIG_NETDEVICES_MULTIQUEUE
and CONFIG_NET_SCHED. Lets deal with these requirements and also
handle them specially in the compat_autoconf.h
This should fix some issues with unresolved symbols when CONFIG_MAC80211_QOS
is disabled. The problem was compat_autoconf wasn't disabling it.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
enable it if MQ is supported.
Signed-off-by: Tomas Winkler <tomasw@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
things which needed porting like the rtnl->netlink migration.
This has been tested by Intel for on 11n on iwl4965.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
being redefined. We can remove this and fix this properly once RFKIL is properly
backported, and if someone does this work.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
This makes mac80211 use the old internal hacked up MQ support.
Note though that unless you are using the new MQ support
your kernel won't really be able to schedule individually
on each separate hardware queue. This applies to any old
kernel releases that don't have this patch yet, which
should be <= 2.6.26.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
2.6.26 added:
ndev->needed_headroom
ndev->needed_tailroom
Then commit SHA1 f5184d267c1aedb9b7a8cc44e08ff6b8d382c3b5
added LL_ALLOCATED_SPACE and made most users of
LL_RESERVED_SPACE make use LL_ALLOCATED_SPACE instead.
Where needed_headroom and needed_tailroom is set we
let the old kernels through as they still are using
LL_ALLOCATED_SPACE
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
c2c40449398b120434f5f95167f7ac44ab10b22e
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Michael Bernhard <michael.bernhard@bfh.ch>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
on MadWifi. We use M= instead, as SUBDIRS is old.
http://madwifi.org/changeset/3603
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
as of 2.6.26, not as part of 2.6.25.
The reason why I put dev_name() in for 2.6.25 was that git-describe
of v2.6.25-7352-g0691663 (patch with SHA1 0691663, and at that point
a tag of 2.6.25) had dev_name() added. Even the top level Makefile
has 2.6.25... this also applies to clamp_val(). Will investigage
the discrepancy.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
For 2.6.25:
Add dev_name() from include/linux/device.h
Add clamp_val() from include/linux/kernel.h
For 2.6.26:
Port new unaligned API (include/linux/unaligned/*). We do this by adding
to the compat-wireless tarball include/linux/unaligned and only include this if we
are older kernels. We deal with architecture handling in compat.h.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
driver this may break them. We may eventually just add the other
usbnet drivers here as its a small list.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
on 2.6.24, it needs testign on 2.6.22. This may have broken
compiling on 2.6.22.
Also on make unload grep for ^module.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
which makes use of is_pcie.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
select SSB_BLOCKIO when B43_PIO is used.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
|
|
strings starting with "2.6" to skip make messages (no attempt is made to
support non-2.6 kernels). Extract the third group of digits. Use it in
the test.
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
It's needed for supporting Broadcom PCI cards.
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
CONFIG_BCM43XX is gone! Add changes required to compat.diff from latest pull.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
This is for support for USB Wireless devices using Atmel at76c503,
at76c505 or at76c505a chips.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
configuration on config.mk.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
> > .../compat-wireless-2.6/drivers/net/wireless/rt2x00/rt2x00firmware.c:116:
> > = Error: redefiniton of »rt2x00lib_load_firmware«
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
skb_copy_from_linear_data() and skb_copy_from_linear_data_offset().
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
but this will be cleane dump soon (tm).
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
compat autoconf.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
CONFIG_PCI=y
CONFIG_USB=y
CONFIG_PCMCIA=y
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
git status runs cleanly, and the other adds kernel 2.6.21 support and
makes config.mk select drivers based on features like PCI, USB and the
ITU-T CRC functions.
Note that I have only tested compiling everything but PCI drivers
against the old kernel so it may well be that some PCI functions are
missing in 2.6.21, they can be added then.
I had to disable rndis wlan because it uses cancel_work_sync() which is
only available in 2.6.22 and higher and can't be ported to older
kernels.
Feel free to reject the patch, with it, however, and a hack to allow
building the crypto code as a module, I now have mac80211 on my N810 :)
Not that I have a p54spi yet, but hey, it's a start.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
make sure we do that. Otherwise we'll get the build system
to install our new modules in the kernel header's Makefile
kernel release. This is what you would expect if you are
building ontop of your own kernel sources but not ontop
of your distribution's kernel.
For example, on Ubuntu Hardy, which is on , you would get
modules installed under the directory:
/lib/modules/2.6.24-rc7/
This is because the Makefile /lib/modules/2.6.24-4-generic/build/Makefile
had defined:
VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 24
EXTRAVERSION = -rc7
What you really want is the modules to be installed instead under
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
The patch:
commit 8b5f6883683c91ad7e1af32b7ceeb604d68e2865
Author: Marcin Slusarz <marcin.slusarz@gmail.com>
Date: Fri Feb 8 04:20:12 2008 -0800
byteorder: move le32_add_cpu & friends from OCFS2 to core
moves le*_add_cpu and be*_add_cpu functions from OCFS2 to core
header and converted some existing code to it. We port
it here as later kernels will most likely use it.
Thanks to Bill Moss <bmoss@clemson.edu> for reporting this.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
is now enabled as default as 'simple' is scheduled for removal for 2.6.26.
b43 driver now gets leds support. This should get some testing.
This was tested on 2.6.25. On 2.6.24-4-generic on ubuntu I was having
issues with the modules not being installed on the
/lib/modules/`uname -r`/updates dir, still looking into this but
pushing out changes for now.
Work based on Hauke Mehrtens's patch.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
athenable and b43enable are already run if they are installed. Don't
run them unconditionally for the second time.
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
compat_autoconf.h intelligence.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
time the scripts athenable and b43enable have not yet been placed in their expected
location (/usr/sbin). This causes the build to fail when uninstall tries to use these
scripts.
Change the uninstall rule to only use these scripts if they exist in their expected
location.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
update compat.diff.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Also:
* update compat.diff to reflect the latest API changes
* running make install now runs make uninstall to ensure
we only have the latest subsystem modules added.
* ath5k is disabled, its currently broken.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* ensure we use our own config to determine which drivers to build.
Be paranoid and include config.mk once in the beginning and
another time during the build process
* Fixes issues during install target users were facing
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Marc Herbert <Marc.Herbert@gmail.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
exceeded the number of arguments possible. This was caused by the way were
were handling CONFIG_* options -- we were just adding defines on the command
line.
To correct this we now embrace our own autoconf header file, just as the kernel
has its own, with defines of the options specified in config.mk. Our header file
is self-generated by scripts/gen-compat-autoconf.sh which parses any
kernel-like config file and outputs a header file with proper defines.
The compat wireless autoconf file is include/linux/compat_autoconf.h and
should not be edited by hand.
The compat-wireless top Makefile makes uses of gen-compat-autoconf.sh through
scripts/check_config.sh. This script and the top level makefile
also check for updates to config.mk and if changes were made it forces
a regeneration for the compat autoconf file. To check for changes we are using
md5sum, which may be overkill, but should work well for now.
Additionally the top level makefile checks for the version of compat-wireless, if
it detects a new release has been pulled it will force a regeneration of the
compat autoconf as well.
This new build scheme been tested on 2.6.22 and 2.6.23.
External build has also been tested with 2.6.23:
make KLIB=/home/mcgrof/devel/kernels/linux-2.6.23.12 \
KLIB_BUILD=/home/mcgrof/devel/kernels/linux-2.6.23.12
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
add backport support for its use by a helper, compat_is_pcie()
to reduce the size of the compatibility code.
* Update compat.diff to reflect latest changes
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* Switching to using simple rate algorithm over pid as pid still
has some usability issues with some drivers.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Note: ath5k has a bug upstream which will be fixed soon
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* usb_endpoint_num compat support added for libertas
* We have a new rate control algorithm, the rate PID controller.
We are selecting this as default for now to get users to help test this
as this is what we will be moving to use as default soon anyway.
* Added missing files from include/net for the Intel ipws: ieee80211.h ieee80211_crypt.h
* Adds rtl8180_grf5101.c rtl8180_grf5101.h
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* Fix load scripts to only try to unload modules loaded,
this prevents errors to users which they don't need to
see
* Handle zd1211rw name change, we missed a few scripts to
fix
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
this wasn't being picked up. Now it should be.
* Lets you build external kernels now, usage:
make KLIB=/home/mcgrof/kernels/linux-2.6.23.9 KLIB_BUILD=/home/mcgrof/kernels/linux-2.6.23.9
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* Update compat.diff to reflect new wireless-2.6 git tree changes
* Do not continue if the compat.diff fails, this will prevent us
from updating the compat tarball with invalid content.
* Use defines now in config.mk, get rid of stupid defines in compat.h
* Lets you specify your own GIT_TREE variable for admin-update.sh
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
anyone please let me know. Not adding old softmac as that will die very
soon.
* Adds ipw2100, ipw2200
* Adds net/ieee80211/ for Intel's ipw drivers and libertas. This
required porting sg_init_table().
* I thought I had added b43legacy, I hadn't this adds it for sure
* Update docs, added TODO list on README
* Fix uninstall, wasn't removing all modules as claimed
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
ethtool
stuff was renamed.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Include linux/autoconf.h before other headers.
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
adm8211
iwl4965
p54pci
p54usb
rt2400pci
rt2500pci
rt61pci
rt2500usb
rt73usb
rtl8180
rtl8187
b43
b43legacy
Tested successfully these:
p54pci (on FullMAC card)
rt2500pci
rtl8180
b43
As with athload we add a b43load script lets you:
b43load bcm43xx
or
b43load b43
libertas was tested but I have 31.02 on bcdDevice and the new
device fails with:
usb8xxx: URB in failure status: -71
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb 4-1: reset high speed USB device using ehci_hcd and address 3
usb8xxx: probe of 4-1:1.0 failed with error -12
usbcore: registered new interface driver usb8xxx
Adds some more backport code.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
drivers available if there were disabled before.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
loaded,
either MadWifi or ath5k.
Usage:
athload ath5k
athload madwifi
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
release they are using, this might come in handy later. Release
added to compat-release.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* rm -rf net/ upon clean now
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
* Use a config.mk, so we get to strip down the compat.diff
* Make compat.diff match patches for upstream
* include <linux/version.h> on <net/cfg80211.h> as most
wireless works uses that.
* Lets unlaod MadWifi now automaticaly. Provide scipts to switch back
to MadWifi easily if its present. The script is athenable. You can either:
athenable ath5k
athenable madwifi
* Borrow from MadWifi madwifi-unload
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|
|
Backporting tested on 2.6.22. Fixed my git repository, updated ipw3945d
detection.
Signed-off-by: Luis R. Rodriguez <mcgrof@winlab.rutgers.edu>
|