The production tester breaks with DMA disabled, so enable it again. Any breakage due to this should get reported to TI PSP so they can fix it.
This effectively reverts aa209df889
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
There is a conflict with eglibc headers since
new headers added the defines that compiler.h
was redefining without undefining them
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Denys Dmytriyenko <denys@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
When a recipe is being parsed which will be skipped due to an incompatible
machine, no local defconfig will exist for the current machine. It seems that
the fetch localpath code doesn't error in that case, so we need to check for an
IOError on the attempted open. Without this, we can hit parse errors.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
This trades "not working" for "slightly less performance" for a lot of use cases
Acked-by: Denys Dmytriyenko <denys@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
First 3 patches are cherry picks from PSP.
Next 7 add support for gen purpose adc, MFD support, sysfs triggering etc.
Currently:
cat /sys/bus/platform/tsc/ain* will trigger sampling
Output of the 12-bit ADC will be in decimal 0-4095
TODO:
1. IIO driver support
2. Other ADC features
Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
The internal ethernet on the beagleboard is not affected by the 3.0.11 usb bug, but other usb devices might be.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
This patch is rewritten to poll for the channel FIFO state without depending on a worker thread.
The v2 version fixes all the previous issues, and remains to be seen if it also fixes the issue of
100% CPU usage with a mac.
Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
After processing one DMA completion, checking if rescheduling is
required instead of scheduling another completion request if the
FIFO is non empty.
This fixes high CPU load issue seen with g_mass_storage on a beaglebone
and could potentially fix similar issues seen with other gadget drivers
Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Networking:
CONFIG_NETFILTER=* - were enabled as modules in beagle xM
Enabled options for ARP and IP tables as modules for the bone kernel.
CONFIG_NET_SCH= - networking schedulers, enabled as modules
IR:
Infrared and dongle support, enabled as modules.
Made RFSwitch subsystem as a module
USB:
Added support for usb-serial and miscellaneous devices as modules.
I also used the kernel .config we used for the -xM for comparison purposes.
Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
This change is required to get GPIO 1_6 working for bone test init purposes
These 2 lines were dropped from the original patch.
Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Enable the PL2303 usb-to-serial driver required for serial output
from the beaglebone tester.
Signed-off-by: Joel A Fernandes <joelagnel@ti.com>
Acked-by: Jason Kridner <jdk@ti.com>
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Koen Kooi
Khem Raj
Christopher Larson
Joel A Fernandes
Denys Dmytriyenko