13 Feb, 2014
1 commit
-
Change the size to reflect correct kernel size.
Signed-off-by: Sourav Poddar
11 Feb, 2014
20 commits
-
Add device tree nodes and pinmux for hdq/1wire on
am43x epos evm.Signed-off-by: Sourav Poddar
-
The patch adds the following to the omap hdq driver.
1. HDQ Device reset call in probe.
2. Enabling '1 wire mode' and checking for presence pulse bit.
3. Proper disabling and enabling of interrupts during read path.
4. Add re-initialization code during SKIP ROM command execution.
5. Miscellaneous cleanup(formatting, return error checks).Signed-off-by: Sourav Poddar
-
Add compatible property for omap hdq driver.
Signed-off-by: Sourav Poddar
-
Device should be build through device tree.
Remove device build from platform code, as information
will not be populated in this path. All the information
for a device build need to come from device tree.Signed-off-by: Sourav Poddar
-
These adds hwmod information for hdq/1w driver.
Signed-off-by: Sourav Poddar
-
Resize the uboot size to 1 MB, as that will be sufficient.
Increase the kernel size from 5MB to 8MB, as android kernel build images
are more than 5MB.Signed-off-by: Sourav Poddar
-
Resize the uboot size to 1 MB, as that will be sufficient.
Increase the kernel size from 5MB to 8MB, as android kernel build images
are more than 5MB.Signed-off-by: Sourav Poddar
-
In the AM335x case, the controller resumes early (noirq) and so
would need the PHY to be powered up early as well. The existing
implementation prevents us from doing that so don't manage PHY power
in suspend/resume. Instead, we let the controller driver manage the
PHY power.Signed-off-by: Roger Quadros
-
The musb core driver resumes in noirq context so we must do so as well
to ensure we are ready before musb core driver resumes.Not doing so breaks suspend/resume and causes delays upto 15 seconds
with high-speed mass storage device connected to host.Signed-off-by: Roger Quadros
-
The root hub port will be resumed by the hub driver using proper
resume signal timing. There is no need for us to resume the port
in bus_resume(). In fact, doing so breaks global suspend/resume
and causes system resume delays of upto 15 seconds.Fixes D-01518 - AM335x: Connectivity/Power:
Long resume time when there is USB MSC device connected to DUTSigned-off-by: Roger Quadros
-
Add sleep states for uart1 pins to save power during suspend/resume.
Signed-off-by: Dave Gerlach
-
Update pinmux configuration for mmc1 device nodes (add leftover pins
(mmc_dat[0-3], cmd, clk, wp) & add sleep state).Currently "mmc" module is working without complete pinmux set may be
because
- it is depending on bootloader settings
- it is working out in default mode.However, it is recommended to keep Kernel independent of above settings
Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
Update omap-serial driver to set the state of the pins to:
- "sleep" on suspend
- "default" on resumeBy optionally putting the pins into sleep state in the suspend callback
we can accomplish two things.
- minimize current leakage from pins and thus save power,
- prevent the IP from driving pins output in an uncontrolled manner,
which may happen if the power domain drops the domain regulator.Signed-off-by: Dave Gerlach
-
Add sleep states for i2c-1 to save power during suspend/resume.
Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
Add sleep states for nand pins to save power during suspend/resume.
Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
Add sleep states for ecap devices to save power during suspend/resume.
Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
Update omap2 gpmc driver to set the state of the pins to:
- "sleep" on suspend
- "default" on resumeBy optionally putting the pins into sleep state in the suspend callback
we can accomplish two things.
- minimize current leakage from pins and thus save power,
- prevent the IP from driving pins output in an uncontrolled manner,
which may happen if the power domain drops the domain regulator.Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
"gpio-matrix-keypad" driver is updated to set the state of the pins to:
- "sleep" on suspend
- "default" on resumeMake use of the same by adding the pin sleep states to the related node.
Also, move the default pin state from generic pinctrl node to
gpio-matrix-keypad node.Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
Add sleep states for leds devices to save power during suspend/resume.
Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach -
update leds-gpio driver to set the state of the pins to:
- "sleep" on suspend
- "default" on resumeBy optionally putting the pins into sleep state in the suspend callback
we can accomplish two things.
- minimize current leakage from pins and thus save power,
- prevent the IP from driving pins output in an uncontrolled manner,
which may happen if the power domain drops the domain regulator.suspend()/resume() callbacks are added which simply puts the pins to
sleep/default state.Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach
10 Feb, 2014
2 commits
-
Reprogramming the DMA after tear down is initiated leads to warning.
This is mainly seen with ISOCH since we do a delayed completion for
ISOCH transfers. In ISOCH transfers dma_completion should not reprogram
if the channel tear down is initiated.Signed-off-by: George Cherian
-
Start the channel tear down only if the channel is busy, else just
bail out. In some cases its seen that by the time the tear down is
initiated the cppi completes the DMA, especially in ISOCH transfers.Signed-off-by: George Cherian
04 Feb, 2014
2 commits
-
During resume don't touch SUSPENDM/RESUME bits of POWER register
while restoring controller context. These bits might be changed
by the controller during resume operation and so will be different
than what they were during suspend.e.g. SUSPENDM bit is set by software during USB global suspend but
automatically cleared by the controller during remote wakeup or
during resume. Setting this bit back while restoring context
causes undesired behaviour. i.e. Babble interrupt is generated
and USB is broken.Fixes D-01343 - AM33XX: CONNECTIVITY: Babble interrupt when using
keyboard without mouseSigned-off-by: Roger Quadros
-
This reverts commit 5f9da94d9fe77242379c5e7944775c5ec71353f8.
We have a better fix so revert this one.Signed-off-by: Roger Quadros
03 Feb, 2014
15 commits
-
Remove cpu OPPs else cpufreq will try to modify to allowed frequency
with boot voltage resulting in potential device hang.Reported-by: Aparna Balasubramanian
Signed-off-by: Nishanth Menon
Tested-by: Tom Rini
Tested-by: Mugunthan V N -
This advisory is applicable only for AM335x PG1.0 only.
By reverting this workaround the MSC read performance increases
by 50%. We are no longer using the Tranparent mode of CPPI,
for RX, instead we are using generic RNDIS mode for both RX and TXSigned-off-by: George Cherian
-
This patch
- prefix all NAND flash partition names with "NAND" to differentiate them from
QSPI and NOR MTD partitions.
- increases the size of kernel partition for NAND
- re-order NAND partitions to keep all SPL MTD partitions together.
- re-order DT bindings to keep 'ti,xx' specific binding togetherSigned-off-by: Pekon Gupta
-
This patch
- prefix all NAND flash partition names with "NAND" to differentiate them from
QSPI and NOR MTD partitions.
- increases the size of kernel partition for NAND
- re-order NAND partitions to keep all SPL MTD partitions together.
- re-order DT bindings to keep 'ti,xx' specific binding togetherSigned-off-by: Pekon Gupta
-
This patch
- prefix all NAND flash partition names with "NAND" to differentiate them from
QSPI and NOR MTD partitions.
- increases the size of kernel partition for NAND
- re-order NAND partitions to keep all SPL MTD partitions together.
- re-order DT bindings to keep 'ti,xx' specific binding togetherSigned-off-by: Pekon Gupta
-
This patch
- prefix all NAND flash partition names with "NAND" to differentiate them from
QSPI and NOR MTD partitions.
- increases the size of kernel partition for NAND
- re-order NAND partitions to keep all SPL MTD partitions together.
- re-order DT bindings to keep 'ti,xx' specific binding togetherSigned-off-by: Pekon Gupta
-
This patch tweaks GPMC interface signal timings for better NAND throughput.
It should be noted that signal timings are NAND part specific, and should be
referred from device datasheet.Signed-off-by: Pekon Gupta
-
This patch tweaks GPMC interface signal timings for better NAND throughput.
It should be noted that signal timings are NAND part specific, and should be
referred from device datasheet.Signed-off-by: Pekon Gupta
-
This patch tweaks GPMC interface signal timings for better NAND throughput.
It should be noted that signal timings are NAND part specific, and should be
referred from device datasheet.Signed-off-by: Pekon Gupta
-
Add mmc1 pincontrol sleep state for power savings in suspend state.
Signed-off-by: Balaji T K
-
Update omap_hsmmc driver to set the state of the pins to:
- "default" on resume
- "idle" after runtime suspend
- "sleep" on suspendThis should make it possible to optimize energy usage for the pins both
for the suspend/resume cycle, and for runtime cases in-between mmc
transfers.By optionally putting the pins into sleep state in the suspend callback
we can accomplish two things.
- minimize current leakage from pins and thus save power,
- prevent the IP from driving pins output in an uncontrolled manner,
which may happen if the power domain drops the domain regulator.Signed-off-by: Hebbar, Gururaja
Signed-off-by: Balaji T K -
PING token is not mandatory in status phase of control transfer
so some high speed USB sticks doesn't support this. If such devices
are connected to MUSB then they would not respond to PING token
causing delayed or failed enumeration.[Roger Q] This also fixes Enumeration issues with some Super-Speed USB hubs
e.g. Dlink DUB-1340Fixes D-01330 - AM33XX: CONNECTIVITY: USB3.0 hubs do not numerate successfully
Signed-off-by: Ajay Kumar Gupta
Signed-off-by: Roger Quadros
Acked-by: Felipe Balbi -
Adding a delay in musb_runtime_resume() prevent the spurious
Babble interrupt from happening during a remote wakeup resume.Fixes D-01343 - AM33XX: CONNECTIVITY: Babble interrupt when using
keyboard without mouseSigned-off-by: Roger Quadros
-
Static variable in a function fails big time with multiple instances.
This will cause musb_runtime_resume to restore invalid context for the
2nd musb instance during probe.Add a suspended flag to struct musb {} to track whether the controller
is really suspended or not. Restore context only if it was suspended.Signed-off-by: Roger Quadros
-
Update matrix-keypad driver to set the state of the pins to:
- "sleep" on suspend
- "default" on resumeBy optionally putting the pins into sleep state in the suspend callback
we can accomplish two things.
- minimize current leakage from pins and thus save power,
- prevent the IP from driving pins output in an uncontrolled manner,
which may happen if the power domain drops the domain regulator.When using any of the GPIOs in a matrix-keypad for wakeup the pins must
be set to the appropriate state to allow this dependent on the board.Signed-off-by: Hebbar, Gururaja
Signed-off-by: Dave Gerlach
