Add a tracepoint to the power trace system to report GPU
frequency change. Format for this is identical to cpu_frequency
event and compatible with Android tools such as Perfetto.
Change-Id: I6207f45f506ab595bd261a297abdd8f40990c8de
Signed-off-by: Deepak Kumar <[email protected]>
Signed-off-by: Indira Biruduraju <[email protected]>
Drive FAL QACTIVE signal high before GMU boot to make sure
path to DDR accepts transactions from GMU/GPU and comes out
of low power state if it was in a low power state.
Change-Id: I238a39c5b9dbd10a8c9b2c6437aaf48df410802f
Signed-off-by: Deepak Kumar <[email protected]>
Add the debug mux offset for getting clk_measure for
csi3phytimer and camss_mclk4 clk in HOLI.
Change-Id: I8b88223b0fa0036ea83c36523da85b7f190d3009
Signed-off-by: Saurabh Sahu <[email protected]>
Adjust the 6 GHz frequency to channel conversion function,
the other way around was previously handled.
Signed-off-by: Amar Singhal <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
[rewrite commit message, hard-code channel 2]
Signed-off-by: Johannes Berg <[email protected]>
Bug: 167126230
(cherry picked from commit 2d9b55508556ccee6410310fb9ea2482fd3328eb)
Change-Id: Ieb788ea5a5ad614651c1a2c83d2c0fc16a98e00f
Signed-off-by: Vamsi Krishna <[email protected]>
Float voltage is read in the batt_health() callback.
batt_health() is called from batt_status() and batt_status()
is also called from the FV callback leading to a deadlock.
Fix it by using the unlocked version of votable while reading FV.
Change-Id: Ie6f9f683c048a2a1c64d46716433bcff359dc3b0
Signed-off-by: Anirudh Ghayal <[email protected]>
Add the interrupt service routine for apss tgu to handle tgu
interrupt.
Change-Id: Ic0128e41ed7a8e837ebab12a28c744af85109bda
Signed-off-by: Raghavendra Kakarla <[email protected]>
Enable policy_engine and pd phy drivers to support
power delivery feature on Holi.
Change-Id: I5048292c5b486a8801c9be6996ef3bff61d6f15a
Signed-off-by: Sriharsha Allenki <[email protected]>
The compilation of policy_engine.c and qpnp-pdphy.c as
loadble modules leads to the circular symbol dependency
of one module on the other.
Fix this by serializing this dependency and ensures that
no symbols from the qpnp-pdphy are directly used in
policy_engine driver.
Also, add the dependency of IIO for policy_engine driver.
Change-Id: I965f887e5c0b5a0cd372cc83353380bad5d8c839
Signed-off-by: Sriharsha Allenki <[email protected]>
qnoc-qos.c needs to be built as module, so add qnoc-qos.ko to
build infra.
Change-Id: I86ace45e2259fc382ed478c6456d55701bf974d4
Signed-off-by: Odelu Kukatla <[email protected]>
Build qnoc-qos.c as a module to make the symbols
available and avoid build warnings.
Change-Id: I62ab7a04225d71c036cd78f38259a98f099632ed
Signed-off-by: Odelu Kukatla <[email protected]>
This change reverts commit ("f1984dceeb dwc3-msm: Add support to vote
USB FORCE_MEM_CORE_ON") which is using dummy clock to update USB
master clock for keeping USB controller CSR retain while going into
CXPC. There is new API qcom_clk_set_flags() propose which supports
same functionality without USB dummy clock.
Change-Id: I0e35e4dac265a22eac2c3e95946cc9171c309508
Signed-off-by: Mayank Rana <[email protected]>
Add regmap_mmio_detach_clk API to QCOM whitelist.
Change-Id: I3eb6e4f0f96f20c41b8f2354712ec79389049aa7
Signed-off-by: Jagadeesh Kona <[email protected]>
Enable idle state failure statistics and tracing in menu governor to
debug cpuidle related issues in SVM. Enable FTRACE dependency for the
config items.
Change-Id: I635a1c06db528358113672daa1ccb78078c79bf3
Signed-off-by: Lina Iyer <[email protected]>
Commit 55adb91acc ("ANDROID: GKI: enable some USB_NET_ config
options") implicitly enabled the CONFIG_USB_NET_AX88* options
which was causing trouble with the db845C_gki.fragment, since
we try to load these (along with PHYLIB) as modules.
So drop these, and it should work out.
Fixes: 55adb91acc ("ANDROID: GKI: enable some USB_NET_ config options")
Signed-off-by: John Stultz <[email protected]>
Change-Id: Ia7e69d7030842f5136aafc064e1ba8a70bad2bb6
It's needed by MTK board for reasons below:
After working with built-in mtk timer(timer-mediatek) as
tick-broadcast devices for few weeks, we found some issues caused
by linux timer framework:
1. tick broadcast installed by insmod cannot switch to oneshot
mode correctly (BUG: 161822795)
2. rcu warning will be shown if we force to re-enabled tick-broadcast
for each cpu when new tick broadcast device is added by insmod
(timer-mediatek.c in our case)
Bug: 161675989
Change-Id: If3056df11d8256f2af4a6928820faed9f690b3db
Signed-off-by: Freddy Hsin <[email protected]>
Signed-off-by: Chun-Hung Wu <[email protected]>
Now that the power supply framework has POWER_SUPPLY_TYPE_WIRELESS
change the wireless power supply type from MAINS to WIRELESS. This
would help android healthd to show "chg=w" rather than "chg=a" in
logs when a wireless charger is connected.
Change-Id: I871779aa806f775e2cdf1e0994593fc991fe77f9
Signed-off-by: Subbaraman Narayanamurthy <[email protected]>
If the device is unable to send the mission mode execution
environment change event but has already entered mission mode
with the ability to allow ramdump collection, it can set a unique
cookie pattern to indicate the availability of ramdumps. Allow
the controller to query for this unique pattern upon any bootup
failure or timeout.
Change-Id: Id1f58b75d3f22e9e21b142cbea74d61cc65615b4
Signed-off-by: Bhaumik Bhatt <[email protected]>
Move MHI to a firmware download error state for any bootup
failures such as unable to find firmware files or load images via
BHI/BHIe. This helps detect an error state sooner and shortens
the wait for a synchronous power up timeout. It also negates the
need to provide additional callbacks due to any intermediate
bootup failures.
Change-Id: I3307e432063f59e978f1e055453155803a6475b2
Signed-off-by: Bhaumik Bhatt <[email protected]>
MediaTek UFS host now supports 2 lanes. Modify the lane count to 2.
This modification shall not impact old 1-lane host because
PA_CONNECTEDRXDATALANES and PA_CONNECTEDTXDATALANES will limit the target
lanes properly during power mode change. So we could relax the limitation
in ufs_dev_params.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Stanley Chu <[email protected]>
Reviewed-by: Avri Altman <[email protected]>
Signed-off-by: Andy Teng <[email protected]>
Signed-off-by: Stanley Chu <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
Bug: 166747994
(cherry picked from commit 460d74a0911c562f56148519a2467271271a73aa)
Change-Id: Ia3cb46d691f4498829565a70ba332cd246dc98a2
Signed-off-by: Stanley Chu <[email protected]>
In ufshcd_suspend(), after clk-gating is suspended and link is set
as Hibern8 state, ufshcd_hold() is still possibly invoked before
ufshcd_suspend() returns. For example, MediaTek's suspend vops may
issue UIC commands which would call ufshcd_hold() during the command
issuing flow.
Now if UFSHCD_CAP_HIBERN8_WITH_CLK_GATING capability is enabled,
then ufshcd_hold() may enter infinite loops because there is no
clk-ungating work scheduled or pending. In this case, ufshcd_hold()
shall just bypass, and keep the link as Hibern8 state.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Avri Altman <[email protected]>
Co-developed-by: Andy Teng <[email protected]>
Signed-off-by: Andy Teng <[email protected]>
Signed-off-by: Stanley Chu <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
Bug: 166747994
(cherry picked from commit 93b6c5db06028a3b55122bbb74d0715dd8ca4ae0)
Change-Id: I595ee3f0da3b89e11cbe5d6d31ba7ce58b1757bc
Signed-off-by: Stanley Chu <[email protected]>
