| Age | Commit message (Collapse) | Author |
|---|
|
By avoiding conflicts of memory access inside Glamo, this doubles the
speed of internal memory access.
Signed-off-by: Thomas White <taw@bitwiz.org.uk>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
This fix the resume path of the accelerometer. If the accelerometer
is not power up, restore to initial status. Move the REG_CTRL1 to the
last one position avoid writing on the register when the device is
in power down mode.
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Fix compilation error due to missing semicolon
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
s/though/thought/
Reported-by: Rask Ingemann Lambertsen <ccc94453@vip.cybercity.dk>
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
This patch applies upstream feedback to the group filter.
The algorithms are equivalent, thus we will get the same
results after applying this patch.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
He is not reachable at the OM address.
We need the right email if we send upstream.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
General cleanups. Now we pass checkpatch.pl.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
I had a patch by Vasily Khoruzhick <anarsoul@gmail.com>
in the linux-arm-kernel as a guide for some of the changes.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
Remove the dependence to the OLD NET DEV operation
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
|
|
This makes wake up on RTC alarm work properly, ported from rtc-pcf50633
commits 4caf79de95c26495e7cdc8204023d97598f887d2 and
c3e4e22fb0c3e1d82f66e67f6592949e48f3995a.
Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
This reverts commit 9c4451ff31b937a478f3d3eabef30b71cbe12b12.
This commit made wifi unusable after ifconfig down and sometimes after
unbinding.
|
|
Current scheme is fragile and is likely to go off sync, especially on
batfull->adapter charging automatic MBC transition.
Query the status bit every time we need it instead.
We need to export another function to query for USB presence because
we can't read anything from PCF50633 (via I2C) inside irq context and
that is needed by usb gadgets.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
If chgmod == BATFULL, setting chgena has no effect. Datasheet says we
need to set resume instead but when autoresume is used resume doesn't
work. Clear and set chgena instead.
This enables a user to force charging by re-plugging USB even when the
charger entered Battery Full mode, might be handy before a long trip.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
After reaching Battery Full condition MBC state machine switches back
into charging mode when the battery voltage falls below 96% of a
battery float voltage. The voltage drop in Li-Ion batteries is
marginal (1-2%) till about 80% of its capacity - which means, after a
BATFULL, charging won't be restarted until 75-80%.
That is a desired behaviour recommended by battery manufacturers,
don't mess with it.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Current implementation is prone to races, this patch attempts to remove all
but one (in pcf50633_adc_sync_read).
The idea is that we need to guard the queue access only on inserting and
removing items. If we insert and there're no more items in the queue it
means that the last irq already happened and we need to trigger ADC
manually. If not, then the next conversion will be triggered by the irq
handler upon completion of the previous.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Using the default kernel "events" workqueue causes problems with
synchronous adc readings if initiated from some task on the same
workqueue.
I had a deadlock trying to use pcf50633_adc_sync_read from a
power_supply class driver because the reading was initiated from the
workqueue and it waited for the irq processing to complete (to get the
result) and that was put on the same workqueue.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Some userspace (notably Enlightment's batget) ignores "capacity"
completely if there're no other properties. Expose a fake value
(850mAh, typical for BL-5C) to make them happy.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Check if machine-specific hooks exist before calling, this allows to use
this driver with gta02 or any other device that is able to provide battery
voltage readings.
Report an estimate of capacity percentage left, based on discharge curves.
Strictly speaking, this is a violation of power_supply class specs but
since this is the only driver for dumb batteries and userspace lacks any
library to do the estimation itself, we have to stuff it here.
Please keep in mind that this estimation is by no means accurate, usually
cell phones have only 5 bars to represent charge status.
Credit goes to Joerg Reisenweber and SpeedEvil for helping with gathering
and analysing the data.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Driver data must be properly registered after allocation
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
If the battery is not present hdq_read will always return an error.
If the drivers knows that the battery is not present the correct thing to do is
to return -ENODEV instead of passing the error on.
Do this for all properties except POWER_SUPPLY_PROP_PRESENT.
The power supply sysfs expects us to do so, else it won't generate a proper
uevent file.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
|
|
F4t discovered that his freerunner panics on connect to his windows box.
This crude hack workarounds the problem.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Clean up debugging messages so that we don't see any output with loglevel=4
(default for Qi). This avoids slowing down suspend/resume by slow fb
output.
Checkpatch barks on this patch but i guess most of that debugging would
have to be changed prior to upstream submission anyway.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
This is because of commit
1f06c28eaa6e77384a366c3d2ab1ccdbb8a807e0
The BQ27000 Battery requires HDQ.
Otherwise the battery is not available at all.
Signed-off-by: Luca Capello <luca@pca.it>
|
|
As discussed on the ML, currently headset button reports "KEY_PAUSE
released" on press and "pressed" on release. For "pause" semantics it makes
sense but if someone thinks of that as of just another button, it'd be good
to have "pressed/released" events consistent with button press/release.
Luca Capello <luca@pca.it> suggested to change the button event to KEY_PLAY
and to invert the state to make it more consistent.
This trivial patch inverts button events for GTA03 KEY_PLUS and KEY_MINUS,
i'm not sure how undesirable that is. I can prepare an alternative that
preserves current behaviour on GTA03 per request.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
Using HAL for E's battery gadget highlighted an oddity: the kernel exposed
last full charge property but didn't expose current charge property. This
resulted in the wrong computation of current battery capacity by E's gadget
(and probably other programs as well).
This patch exposes a corresponding bq27000 register to make E battery
gadget happy (it is still not showing correct values because of bugs in HAL
resulting in 3 batteries (apm emulation and usb supply being bogus here)
instead of one).
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
|
|
This patch adds a call to cancel_delayed_work before a call
to schedule_delayed_work.
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
Signed-off-by: Daniel Willmann <daniel@totalueberwachung.de>
|
|
Machines that do not route the USB vbus signal to a GPIO can use this
mechanism as an alternate way for the udc driver to detect the usb
connect/disconnect state, which is useful (for example) to enable the
link up/down state to be determined for the g_ether gadget.
Signed-off-by: Mike Westerhof <mwester@dls.net>
|
|
The g_ether USB gadget driver currently decides whether or not there's a
link to report back for eth_get_link based on if the USB link speed is
set. The USB gadget speed is however often set even before the device is
enumerated. It seems more sensible to only report a "link" if we're
actually connected to a host that wants to talk to us. The patch below
does this for me - tested with the PXA27x UDC driver.
Signed-off-by: Jonathan McDowell <noodles@earth.li>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Avoid a kernel panic during an early call
to the pcf50633_mbc_get_status function.
|
|
This patch fixes a linking failure in the Glamo GPIO SPI bitbang support.
Signed-off-by: Rask Ingemann Lambertsen <rask@sygehus.dk>
|
|
This patch fixes some indentation which was off by one and converts
spaces to tabs.
Signed-off-by: Rask Ingemann Lambertsen <rask@sygehus.dk>
|
|
Small patch to add /sys/class/power_supply/battery/present.
You must not run a GTA01 without battery thus it is always 1.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
resources
Call unbind() hook of the gadget so that it can properly release its
resources. This fixes bug http://docs.openmoko.org/trac/ticket/2240
("removing/reloading g_ether not working"). All other instances of
usb_gadget_unregister_driver in Linux already call unbind() so it is
bit odd s3c2410_udc.c did not do so.
Signed-off-by: Timo Juhani Lindfors <timo.lindfors@iki.fi>
|
|
This patch adds the call to the worker in
bq27000_battery_external_power_changed. Now (un)plugging the USB cable
effects the battery status soon. I don't know if it is possible call
the status change directly.
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
Signed-off-by: Daniel Willmann <daniel@totalueberwachung.de>
|
|
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
|
|
Distributions that use Xorg will not need the workaround at
all and now they can disable it.
Updated configurations to enable it by default.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
~ Cleanup of a redundant call I introduced before.
~ Removed useless assignation.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
|
|
To disable the workaround at run-time run:
echo 0 > /sys/class/i2c-adapter/i2c-0/0-0073/pcf50633-regltr.9/glamo3362.0/glamo-fb.0/xglamo_hack
I guess I should add a compile option now to completely disable the workaround but let's call
this an enhancement and let's close the bug now.
By default the workaround is enabled.
Signed-off-by: Nelson Castillo <arhuaco@freaks-unidos.net>
Reported-by: Timo Juhani Lindfors <timo.lindfors@iki.fi>
|
|
issue""
This reverts commit aa2dbc74bed8cf60e591c654ba1882011b9a8ece.
Reverting the revert. Sorry for the noise.
Check:
http://lists.openmoko.org/pipermail/openmoko-kernel/2009-April/010023.html
|
|
This reverts commit bd4b7e8e84ab43a13a4620b001d52d373c2122b3 which brought
WSOD back for some devices.
It was reported that the problem does not happen with Qi and it
seem it is good news. See:
https://docs.openmoko.org/trac/ticket/2274
Conflicts:
drivers/video/display/jbt6k74.c
|
|
Disabled additional debugging introduced
in 70b36104ab8c03f5be6d03344214d09447d9d4a8
|
|
Changing dev_info => dev_dbg.
Signed-off-by: Tim Niemeyer <tim.niemeyer@mastersword.de>
|
|
This closes bug #2267.
Signed-off-by: Mike Westerhof <mwester@dls.net>
|
|
Cleaned unnecessary debugging output.
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@gmail.com>
|
|
Re-enable usbmode attribute.
It was disabled by 70b36104ab8c03f5be6d03344214d09447d9d4a8.
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
|
|
Cleanup code of ohci-s3c2410.
Signed-off-by: Michael Trimarchi <michael@panicking.kicks-ass.org>
|
|
Space transition between sleep-in and sleep-out of at least
120 ms as mentionned in the spec. Also, made sure the LCM
is moved into sleep state before the pixel clock is turned off.
To do so, I've modified glamo-fb to send FB_BLANK_POWERDOWN before
turning off the pixel clock. Also fixed various violation of coding
style. This patch should reduce the risk of white screen.
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@gmail.com>
|
|
A network device is supposed to disable its internals in the "close"
function. Doing so affords us protection against various races,
including the ioctl vs. rfkill conflict reported by Michael.
Signed-off-by: Werner Almesberger <werner@openmoko.org>
Reported-by: Michael Trimarch <michael@panicking.kicks-ass.org>
|
