GitLab

Control RGB leds through sysfs entries.  Add necessary
drive lines, names, current and PWM cycles.

Change-Id: I6b6174e276c2998e25619eb3e65bfe468820b84c
Signed-off-by: Amy Maloche <amaloche@codeaurora.org>

pm8038 is the first PMIC chip to support RGB led, so
driver framework must be added to support it.

Change-Id: I8103f1e5ec37cdd8718a3daa10587f56a44729c6
Signed-off-by: Amy Maloche <amaloche@codeaurora.org>

* changes:
  usb: dwc3: USB_DWC3_MSM not auto select USB_GADGET_SUPERSPEED
  usb: dwc3: add USB_DWC3_MSM peripheral controller choice.

Enabled the Kconfig options to support msm_watchdog.

Change-Id: I4cde65bc7199b4cd5e02f4dc0f7bd3d623c6bfb5
Acked-by: Kaushik Sikdar <ksikdar@qualcomm.com>
Signed-off-by: Rohit Vaswani <rvaswani@codeaurora.org>

Move out io device code from board files of MSM7627a for easy
update of io device files.

Change-Id: I641861c271ce38267309b00b4f1baff45f321d4e
Signed-off-by: Chintan Pandya <cpandya@codeaurora.org>

Some of the qcache statistics were being calculated
without being protected by a lock. In very rare cases
this might lead to a nonzero allocation count when
all qcache pages have been freed (and thus must be 0).

Change-Id: I7d82d83d021e36a331fa688e5a7ced7ad8e2b176
CRs-Fixed: 344939
Signed-off-by: Larry Bassel <lbassel@codeaurora.org>

The new freq 192MHz support is added and Fmax for low
voltage is updated. Therefore, update the freq plan.

Change-Id: I1adefcf952723a4b58d8ad5cbbd5e5c689788c38
Signed-off-by: Tianyi Gou <tgou@codeaurora.org>

Enable feature for CPU scheduler to recognize task
groups and schedule CPU bandwidth accordingly.
This is the mechanism used by Android to prioritize
foreground and background tasks differently,
enhancing the user experience. Though no quantifiable
improvement has yet been observed this can hopefully
be improved. No negative performance impact has
been seen.

Change-Id: I5754aed90b3d95a6ebce10c1233ed861473a504e
Signed-off-by: Vince Leung <vincentl@codeaurora.org>

* changes:
  msm: pm2: Use PER CPU variable to log the idle stats
  msm: pm2: Change the initialization sequence of "msm_pm_init_done"
  msm: pm2: Use a different kernel API to get the sleep length

Add KGSL_DEV_ERR_ONCE macro to kgsl_log.h, with functionality
similar to pr_err_once() but output format of dev_err(). Add logging
for deprecated code usage, to kgsl_ioctl_sharedmem_from_vmalloc()
and KGSL_USER_MEM_TYPE_ADDR switch case from kgsl_ioctl_map_user_mem()

Change-Id: I43bbd5acfb4630b88170034d61f1d099fbe3f118
Signed-off-by: Harsh Vardhan Dwivedi <hdwivedi@codeaurora.org>

With this change, the dwc3 MSM controller will not operate as
super-speed unless the USB_GADGET_SUPERSPEED will be set explicitly.

This is useful when we explicitly want to enable super-speed feature to
customers or bypass problems due to lack of host drivers which support
super-speed.

Change-Id: I55ce8564bb282d8b2251bd76e06924df321e0980
Signed-off-by: Ido Shayevitz <idos@codeaurora.org>

We distinguish between the DWC3 generic modules, compiled
depends on USB_DWC3, to the vendor specific glue layers
as dwc3-msm.c or dwc3-omap.c complied by DWC3_MSM or DWC3_OMAP
configs. These configs automatically selects the USB_DWC3.

The USB_DWC3_MSM will be used in next commits for specific
needs of the dwc3 on MSM architecture.

Also removed USB_GADGET_DWC3 from Kconfig in order to organize
as in latest Linux kernel. This is done since the dwc3/gadget
is compiled anyway (as long as with the future dwc3/host) and only
on runtime (reading HWPARAMS) the dwc3/core decides if to load the
gadget (or the host).

Change-Id: Ifc2ebb1c9a6cfd59cd004aee5b528ba2bb364627
Signed-off-by: Ido Shayevitz <idos@codeaurora.org>

Not all LEDS are supported by all PMIC chips, so add
a check to make sure unsupported LEDS are not probed
by driver.

Change-Id: I6b2a2641fa1e9bc63b23460bc568da3e6b0bc746
Signed-off-by: Amy Maloche <amaloche@codeaurora.org>

Replace local static varaible with per cpu static varaible
to log the individual cores idle statistics information
correctly.

Change-Id: If6a9ccd66c2a9ff6bb748134814fb6b7bdb4d257
Signed-off-by: Murali Nalajala <mnalajal@codeaurora.org>

Set the "msm_pm_init_done" variable after populating the idle
statistics data for both the cores. Setting this varaible before
initializing the idle stats causes the kernel exception (division
by zero). This is happening because one core is doing the driver
init and another core finds idle time and execute the arch_idle
to do a low power mode and also is trying to log the idle stats
information where another core is not prepared yet.

Change-Id: I71d2785eb47f638019817b3f5659711592243063
Signed-off-by: Murali Nalajala <mnalajal@codeaurora.org>

Currently pm2 driver calls the "msm_timer_enter_idle" &
msm_timer_exit_idle" API's for all the power modes (i.e.
SWFI, GDFS/StandalonePC) to get next timer expiration &
time sync with modem respectively. Use these API's only
when the system decide to enter into a modem assisted PC
otherwise use a differnt API to get the sleep length.

Change-Id: Idf5d53518f75551bab432b3532c0eae1f61cdff1
Signed-off-by: Murali Nalajala <mnalajal@codeaurora.org>

Currently PM driver uses the memory given by the "memblock_alloc"
api. These initial memory bytes are programmed with warm boot entry
sequences to come out of the low power modes. Currently this API is
causing the system instability with 1GB DDR parts and HIGH MEM is
enabled. To resolve this issue use a different API to allocate the
memory i.e memblock_remove.

Change-Id: I1016d63f1154e7b4d4d4b702202878d6f93cecf0
Signed-off-by: Murali Nalajala <mnalajal@codeaurora.org>

Underrun occurs while downscaling video data on 8x60. Enable writeback
if video downscale is required.

CRs-fixed: 329630
Change-Id: Ief4b002efd6f4ad9fcfd844325c05de619d47719
Signed-off-by: Mayank Chopra <makchopra@codeaurora.org>