| Age | Commit message (Collapse) | Author |
|---|
|
This is a safety precaution to ensure that two RPMS with the
same commit hash are not built out of differing source trees.
All changes must be commited, creating a new commit ID,
or discarded before an RPM can be built.
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
|
|
Sometimes the the HPGK touchpad will fail to properly respond
to a ps2_command() during recalibration. This patch works around
that issue be scheduling another recalibration if this happens.
This solves http://dev.laptop.org/ticket/9008
Signed-off-by: Paul Fox <pgf@laptop.org>
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
The HPGK touchpad that is found on the XO driver has
historically exhibitted eratic behaviour in various
environments (very dry, very humid, etc) that can be
worked around via some delays. This patch turns those
delays into module parameters to make testing simpler.
Signed-off-by: Paul Fox <pgf@laptop.org>
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
netif_carrier_off/on was causing packet loss during a scan.
This is not needed to stop the transmit queues and is not used in any
other drivers.
Signed-off-by: Anna Neal <anna@cozybit.com>
|
|
Returning -ENOSUP is confusing wpa_supplicant.
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
This confuses NetworkManager and causes it to mix scan results from different
networks.
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
keys.
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
wpa_supplicant issues a SIOCSIWAP with BSSID all zeros to clear the current
association. The driver will interpret that as an association request for
'any' AP. Better to do nothing instead when such request is received.
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
When the kernel is built w/o Synaptics protocol support, we configure
a PSMOUSE_PS2 protocol device during initialization; however
during suspend/resume, we report it as a PSMOUSE_SYNAPTICS protocol
device. This caused the kernel to think it is not the device that
was attached at resume time and so we end up doing a deffered
reinitialization, which takes > 2s. This patch causes to not
report a Synaptics protocol extension when protocol support
is disabled and allows us to resume w/o reinitializing the device.
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
When a rule < 4 bytes was set, the mask was applied to the wrong part of the
signature. This patch corrects that.
Signed-off-by: Anna Neal <anna@cozybit.com>
|
|
Changed the set args/get args classification to be more intuitive
for set_wol_rule, get_wol_rule and reset_wol_rule.
Signed-off-by: Anna Neal <anna@cozybit.com>
|
|
Signed-off-by: Anna Neal <anna@cozybit.com>
|
|
View README for new API.
Signed-off-by: Anna Neal <anna@cozybit.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
Based on a patch from Shailendra Govardhan <shailen@marvell.com>.
This patch allows implementation of more specific wake on lan rules than those
of ethtool.
Please note that only firmware 5.110.22.p20 and above supports this feature.
Signed-off-by: Anna Neal <anna@cozybit.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
Missed in last commit
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
Add a simple mechanism to build OLPC RPMs right out of the kernel
source files. Simply do "make olpc-kernel-rpm" and the resulting
RPMS will be in olpc/SRPMS and olpc/RPMS directory.
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
Removed vmlinux* rule because it matches too many useful files, replacing
it with rules matching filetype by filename (e.g. *.gz). Also unignored
.mailmap from the top directory. Added a comment telling the user how to
check for tracked files being ignored.
Signed-off-by: Eduard - Gabriel Munteanu <eduard.munteanu@linux360.ro>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from kernel.org commit f72e9df0e04c0ea7b5c97e6db5b114234bf3c36f)
|
|
Added !vmlinux.lds.h to .gitignore because it would otherwise be ignored.
Signed-off-by: Eduard - Gabriel Munteanu <eduard.munteanu@linux360.ro>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from kernel.org commit cd50e89244087432a70598e432ff199a009b0e73)
|
|
Sometimes I got this:
$ git-status
{snip}
# On branch master
# Untracked files:
# (use "git add <file>..." to include in what will be committed)
#
# ncscope.out
nothing added to commit but untracked files present (use "git add"
to track)
Fix it.
Signed-off-by: Jike Song <albcamus@gmail.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
(cherry picked from kernel.org commit 9723c046bd5989aa9064038ce142f498bb1870d6)
|
|
With commit 3f1b0e1f287547903f11fa1e6de7d2765597766e ".gitignore update"
Linus's current git tree starts to ignore any "*.rej" files.
So "git status" no longer shows these files, but the ones who works with
quilt patchsets, this not makes life easier as expected.
Because sometimes a work flow (at least for me) requires
"quilt push -f" followed by "git status" to see unresolved merge
conflicts, work on these conflicts to correct them and finalize
the patch with "quilt refresh".
And if there are some "*.rej" files exists in tree, for whatever
reason, this means something goes really wrong there and i think
this situation not deserves to be ignored.
Signed-off-by: S.Çağlar Onur <caglar@pardus.org.tr>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
(cherry picked from kernel.org commit 1f5d3a6b6532e25a5cdf1f311956b2b03d343a48)
|
|
(which is autogenerated by kbuild)
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from kernel.org commit e4c576b911e364737b1bf4f5bfdab1c440713f26)
|
|
Add some autogenerated files to various .gitignore files
Signed-off-by: Matthew Wilcox <willy@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from kernel.org commit 2cfed60cc24676d65e01278dbf10d0069de02592)
|
|
wpa_supplicant will do a scan before associating. After reasonable period of
time, it will try to read the results via a get_scan_results ioctl
(SIOCGIWSCAN). If no results are available (for instance if the driver
returns -EAGAIN) the supplicant will request another scan via the set_scan
ioctl (SIOCSIWSCAN).
When receiving the -EAGAIN error the caller should just retry a SIOCGIWSCAN.
But wpa_supplicant does not check the return error and just re-issues a
SIOCSIWSCAN in all cases, which worsens the problem. So a better fix for this
problem would probably be in wpa_supplicant.
But for now we can address this problem in the driver by not scheduling new
scans when one is in progress.
Also, modify lbs_get_scan so that -EAGAIN errors show up in debug logs.
|
|
This reverts commit 215f858d98978d7fa331a628bcc5a5e9a676de25.
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Remove __init from a bunch of functions; they're exported to modules,
and thus should not be marked as such. This fixes an oops that happens
when promfs is built as a module.
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
set CONFIG_MTD_REDBOOT_DIRECTORY_BLOCK to 0; block 0 is where OFW writes
the partition information.
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
WPA key installation must happen as soon as possible after the 4-way handshake
or the EPOL-GTK frame cannot be decrypted and gets dropped. This patch executes
WPA key installation in ioctl context instead of deferring the task to the
association worker.
In a previous patch I had tried to reduce the WPA key installation delay by
forcing immediate execution of the association thread, but that may cause
incomplete association requests.
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
This reverts commit 93970329291a9ff73301a96a27c4af274f096835.
The original commit contains changes to psmouse-base.c and
olpc.c that mention an upstream fix and fixing a low power
toggle bug but the commit does not point to an upstream source
or describe what the bug is. Upon searching lkml, it looks
like this commit is a merge of several patches sent upstream.
As this has not received wide-spread testing yet, I am
reverting for now and we can add the changes back in
properly as separate commits after the 8.2 release.
Signed-off-by: Deepak Saxena <dsaxena@laptop.org>
|
|
Commit abe7a66b36f344a822 (in olpc-2.6 repository) addressed the issue of slow
WPA key installation, but could break the ordering of iwconfig commands like:
iwconfig eth0 essid foobar key 235232363632 mode managed channel 11
This patch adds additional guards to ensure that the fast key installation
only applies to WPA.
Thanks to Dan Williams for pointing that out.
Note to maintainer: This patch and abe7a66b36f344a822 have been submitted as
a single patch to wireless-testing: http://marc.info/?l=linux-wireless&m=122161439327313&w=2
Signed-off-by: Javier Cardona <javier@cozybit.com>
|
|
PCI_GOANY was broken without a backported patch..
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Previously, one would have to specifically choose CONFIG_OLPC and
CONFIG_PCI_GOOLPC in order to enable PCI_OLPC. That doesn't really work
for distro kernels, so this patch allows one to choose CONFIG_OLPC and
CONFIG_PCI_GOANY in order to build in OLPC support in a generic kernel (as
requested by Robert Millan).
This also moves GOOLPC before GOANY in the menuconfig list.
Finally, make pci_access_init return early if we detect OLPC hardware.
There's no need to continue probing stuff, and pci_pcbios_init
specifically trashes our settings (we didn't run into that before because
PCI_GOANY wasn't supported).
Signed-off-by: Andres Salomon <dilinger@debian.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
|
|
WPA requires that the PTK is installed immediately after the 4-way handshake
in order to properly decrypt the subsequent incoming EAPOL-GTK frame. If the
PTK is not enabled by the time the EAPOL-GTK frame arrives, the frame is
dropped and the supplicant does not receive the group key.
This will happen with fast Access Points that send the EAPOL-GTK frame before
the suplicant has successfully installed and enabled the PTK. To mitigate
this situation, this patch simplifies and accelerates the SIOCSIWENCODEEXT
execution.
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
The RTC code that was tacked in there defines an arch_initcall;
meanwhile, the code also has a module_init function that conflicts w/
the RTC initcall. The RTC code should probably be separated or..
something. For now, just don't allow modular building of olpc-pm.
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
...plus xo.config, of course.
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
This config has the config snippets that an XO requires. Everything's
modular, as well. This can be combined w/ a generic distro config, or
to a minimalist, super-optimized config for a tiny kernel.
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
|
|
|
Signed-off-by: Andres Salomon <dilinger@debian.org>
|
