Age | Commit message (Collapse) | Author |
|
abox suspend debug dump timing seems to be masking a race condition
in the pm suspend flow. This requires further investigation. Enabling
the dump to prevent field issues until a proper fix is in place.
BUG:262447135
TEST: manually verified that suspend dump is enabled and executing.
Change-Id: I1dbf5c4ac2d2437fb0c803b48dc40ffad789c02a
Signed-off-by: Radu Solea <radusolea@google.com>
|
|
Assign GPIOs for the driver to recover SDA.
Bug: 245350444
Signed-off-by: Amit Bhanagay <amitbhanagay@google.com>
Change-Id: I187ccb99e2aec6706f655971185fb5a379fb6597
|
|
Bug: 228131904
Test: Manual test
Signed-off-by: Radu Solea <radusolea@google.com>
Change-Id: I156c8497a082582415df683890d20882278fec35
Signed-off-by: Radu Solea <radusolea@google.com>
|
|
Don't build dev2/proto1/proto11 overlays
Bug: 225204431
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ie1aca4517e48443a773c9f8e0c3b60aeaf5c7712
|
|
Remove the logging period for acpm_ipc in order to stops unnecessary
wakes. The acpm log is not being leveraged so is wasted work/cycles.
Bug: 218969249
Test: Manual
Signed-off-by: Matt Stephenson <stephensonmatt@google.com>
Change-Id: Ia18a55c3af03ce47f423a88317a3e5c9d68edb77
|
|
This comes after fixing the error detailed in b/223626740#comment13 to
Test: Manual
Bug: 223626740
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: Icb68d9b81b09d95ec3c515e565c8847a3c2ede16
|
|
With patch: "Don't do deferred probe in parallel with kernel_init thread",
the display was shutting down before touch. Since both touch and display
share a single power control (controlled by the display), shutting down
display before touch would cause touch i2c shutdown commands to timeout.
Bug: 201726797
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ie760411f16d56cd8de9d2e0b0b809819edadb986
|
|
Test: Manual
Bug: 217730279
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: I6deca59374739dcc529c145b87a0a01db273b5b6
|
|
This captures the dependencies betweeen display devices to guarantee the
proper probe order.
Bug: 196103381
Signed-off-by: Alice Sheng <alicesheng@google.com>
Change-Id: Ie2bf1bdf7b7d385a87108d0d5a069e6ae3c7957f
|
|
It will be used to signal Tilt To Wake to turn the display on as soon as
possible.
Bug: 222158397
Change-Id: I9f9a9753444279ba62a40ff86a77080f42cf2955
|
|
Specifically, lengthen the touch disable timeout to 10s and set the
NFC debounce to 40ms.
Signed-off-by: Matthew Bouyack <mbouyack@google.com>
Bug: 201462475
Change-Id: If39cb1b90ca6a28aff66740f166feaea85ade852
|
|
This is helpful in case of a BT crash as it dumps the last raw UART
data the controller received.
Test: Manual
Bug: 213963172
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: If305ef5fe1ffa3b7281472370b7055b8214379a1
|
|
Testing concluded and team agreed that 43C is the correct threshold to
balance battery swelling mitigation and charge time
Bug: 219552596
Change-Id: Ic5b1bf368f4edd26e2d3661ce64704129ae17dae
|
|
The userspace one was raised to 53 according to new UX and HF
guidelines. Raise backup to 54 to follow safety limits and avoid
collision with userspace one
Bug: 217940566
Change-Id: I8bd8d922aa2836e32b0f178493198000cd421346
|
|
Re-enable the S3C RTC block now that the alternate driver will be used
which only makes use of the tick conters to implement a persistent clock
and doesn't configure any timer or interrupt functionality.
Bug: 208709092
Bug: 209870351
Change-Id: I966656a0a95392754fa8b8f1477d5bce77054e6e
|
|
Prior to EVT / Dev3, the 32k used by haptics, mcu, and other external
ICs was sourced from the RTCCLKOUT pin associated with the RTC block
in the main part of the SoC and in order to enable the output of the
clock, the s3c-rtc driver needed to be loaded to configure registers.
However, now the 32k is being sourced from the PERI32K pin of the PMIC
such that no driver is needed and furthermore, it now appears that the
s3c-rtc driver is unexpectedly seeing interrupts and taking wakelocks
which is causing attempted entries into suspend to abort.
Since this driver is not necessary and is creating problems, keep it
from activating by marking the hardware declaration disabled within
the device tree.
Bug: 209870351
Test: No s3c2410-rtc in /proc/interrupts and 32k present at R212 on Dev3
Change-Id: I8b2ffbbde67076d150795c3cb14e324ff465515e
|
|
Set up the pinctrl and set the fuel_alert_soc to 0. This should
let us receive an interrupt when the SOC reaches 0 so we can
shut down in a timely manner.
Bug: 206469513
Test: drain battery to 0, see device wakeup, FG interrupt logs and
immediate shutdown
Change-Id: Idda781b9c70e93038ccb162634e2f2ece0dfc926
|
|
Test: Manual
Bug: 203843647
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: I006f987bf6a180c063a2bfa01e67c25468781e2e
|
|
Also pull in the configs set in device-tree for gs101
(can remove from BoardConfig-common.mk)
Bug: 205474615
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Idd3250b90a85373a6b0f598690cbc12d067b9508
|
|
This disables the PPG NTC based battery swelling mitigation, since we're
going to use the BATT NTC instead (pending approval of final values) and
the PPG NTC based one isn't as effective and impacts charge time more
Bug: 204942637
Test: cat /sys/class/thermal/thermal_zone6/trip_point_0_temp
Change-Id: I7cb59afccce83fbd130578a83a2a76b99fc1719b
|
|
Disable pull-down on SPI MISO pin when AP is suspended to prevent
drive contention since the MCU has a pull-up enabled.
Test: Use Kibble to verify power consumption on MCU rail is reduced
by ~70 uW when AP is suspended.
Bug: 202555263
Change-Id: Ic36ea514daa85f4947f14925235f379b760ff921
|
|
Bug: 202871562
Change-Id: I9dbf0d12749bb5cccfcbd7b0814baeeed113c4cc
|
|
This reverts commit 6dee9fae56a13c6c3cdaa774e791ec6f77151ef6.
Reason for revert: Resolved IRQ issue in pa/2056901
Change-Id: I7f26f49196e340d307422c26ae7ce69f7819535e
|
|
Test: Manual
Bug: 201737830
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: I48b90199d18b037294a987610ff8cc86789c71b2
|
|
Previously 300 mAh was used and while the pack itself is labeled with
a sticker saying 294 mAh, this number should be reduced to the derated
value of 279 mAh to be consistent with what we're using in modeling.
Note that within the Linux kernel battery capacity is discussed using
three different terms.
1. "charge" represents capacity in µAh only
2. "energy" represents capacity in µWh only
3. "capacity" represents capacity in *percents*, from 0 to 100
Thus the fuelgauge,capacity_max property should be in percents but it
is actually in tenths of a percent. Thus the 1000 number currently in
the device tree actually means 100%. More problematically, this value
is improperly sent to userspace as POWER_SUPPLY_PROP_ENERGY_FULL_DESIGN,
but luckily, Android does not appear to use this power supply property.
Bug: 202396388
Signed-off-by: Tim Kryger <tkryger@google.com>
Change-Id: Ia912e5ba668b16af1703bf8adf2113abf2693378
|
|
The code was originally copied from a previous CL that shared an
interrupt between novatek and nfc. For this particular CL, novatek is
the exclusive owner of the pin, so there is no need to put it as part of
the pinctrl.
The issue with having it part of the pinctrl is pinctrl_0 already has
the interrupt pin for haptics/smart amp, so just due to the loading
order this clobbered the haptics/smart amp interrupt causing it to be a
pull-down (on an active-low interrupt), which caused it to continuously
fire.
Bug: 203011584
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I67295b89f4914184885e71045c1803e94d398a74
|
|
This will be used to ignore false touch events
when NFC is active.
Signed-off-by: Matthew Bouyack <mbouyack@google.com>
Bug: 192429087
Change-Id: If5ea3bfe1fe25c5024d7fc59c6d1ee24128baf95
|
|
To account for the fact that we now need to track battery temperature as
well.
The exact numbers will need to be readjusted later.
Bug: 200039743
Change-Id: Ie2cf734cbe506b1d3490cfa8aad10a13024a7a70
|
|
DMA is enabled for the AP - MCU SPI bus after changes to the MCU
response paths, and flow control improvements necessary for reliability.
The MCU now uses fixed transaction timing by removing the message
framework and task-switching from the MCU response paths, and asserts
flow control when resources are exhausted, which is very rare and will
be tuned further.
See go/rh-nanohub-mcu-architecture
Bug: 185406845
Bug: 169772504
Change-Id: I7a9eeb376e42d76ce7e0e04e1874cef763ed1eec
|
|
Bug: 195664265
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Idc3f7e581dc6f30c9bf692dc4f4183e03d05ed31
|
|
This reverts commit 5e0def69842d1ca81fe99729c2138011dcedf0ba.
Reason for revert: CS35L41 hibernation mode is preventing the CS40L25 from producing vibration effects.
Change-Id: Ib7e9014069a53db022aa7dd8bf6fe577ddd2ea55
|
|
This adds a missing device tree entry
to enable hibernation.
Bug: 199932068
Change-Id: I1eb808fa7babf54db125b3d72290ec6b6f4eb848
|
|
This adds the cirrus,hiber-enable property needed
for the CS40L25 to enable hibernation when suspending.
Bug: 199470889
Change-Id: I2e24a066cfc60785e98ff63f69b620e5994d99e3
|
|
Revert submission 1986244-r11-nfc-disables-touch
Reason for revert: This causes about a 4s disabling of touch on screen off/screen on.
Reverted Changes:
I6543cc80d:Ignore touch events when NFC is active
Icd8564d4f:Add NFC idle gpio to the touchscreen DT
Bug: 192429087
Change-Id: I39723d3f3358725165ef065ac8ae26488451c009
|
|
This will be used to ignore false touch events
when NFC is active.
Signed-off-by: Matthew Bouyack <mbouyack@google.com>
Bug: 192429087
Change-Id: Icd8564d4f16a696496ac1b0df5d39ec39c781222
|
|
Want to throttle at a lower temperature at high SOC, for battery
swelling mitigation. Also, switch to the fair_share governor in order to
support multiple trip points.
Bug: 187884905
Change-Id: Ie7cf6e4cff5c4688c564dff4426dc9b79a8ce698
|
|
Test: Manual
Bug: 191826568
Signed-off-by: mullerf <mullerf@google.com>
Change-Id: Ib7c4a33bb9a8772dfb0fef46ed0de38ac39023aa
|
|
In accordance with latest test data, change the following:
- Wider Battery voltage tracking range, all the way down to 3.95V
- Start charging at 150mA before bumping up to 300
Bug: 191698991
Change-Id: I02fdbb86c3cf05db9a6ce58a904ec6ea93f64d2c
|
|
This will free up ~88MB of memory for the btwifi sku
Bug: 185294503
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ie9a6c20a64b5d9941f72b4082b8d355b8f0e692e
|
|
Bug: 192368652
Signed-off-by: Seungbin Lee <seungbin.lee@samsung.com>
Change-Id: I51389207929a37b1d8789435b55c1531f1ee62ba
|
|
Bug: 192368652
Signed-off-by: Seungbin Lee <seungbin.lee@samsung.com>
Change-Id: Ideaf78576c797c9fda470d884218918fa24a1632
|
|
Disable SKIP_ABNORMAL and DYNAMIC_SCALE, for the following reasons:
- They don't work because of the "fake charging status" necessary for
wireless charging
- We have already implemented our own version of it
Bug: 187852752
Change-Id: I3b18076f0361c75d4a7434ac5f67cc2839791c9f
|
|
evt11_hl6111 should be evt1_hl6111
Bug: 191931765
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I0b12a5724bbb83948b393c00e00c30c54515c90d
|
|
With the exception of PPG_NTC on Proto devices since the location of
that one led to very fast heating. This will reduce the average rate of
NTC polling to a more reasonable one instead of twice per second.
Changing it now because the test data shows that we no longer need such
a fast polling rate, based on how fast the device actually heats up
Bug: 188443364
Change-Id: I026359e8677aaf05a3342ecc6316100c11cdd019
|
|
According to new charge data based on expected device operational
conditions.
Bug: 191409619
Test: Checked the charging rates at higher (emulated) temperatures
and up to 4.35V battery
Change-Id: Ic5c942e5abf40522b0214983dbb933361978f925
|
|
- Disable pull-up/down for pins with external
pu/pd resistors
- Configure pins to not pull down during sleep
(samsung,pin-pud-pdn = 0 = disable up/down)
- Set correct pin-pud state for gpm8/9/16/17 gpf0-1/0-3
(samsung,pin-pud = 0 = disable pull up/down)
- Remove always on for ldo9 (vts)
Change-Id: I11df199a14b4e5ff0bd91a8e51decb0215040709
Signed-off-by: Patrik Torstensson <totte@google.com>
|
|
This should cause the kernel to automatically re-order the regulator
probe to be before the display panel probe.
Bug: 191181630
Bug: 189516155
Change-Id: Ic998522343a0019e21e68af10a4e6ce4df57d245
|
|
- Set BUCK1/LDO3/LDO4/LDO6/LDO7/LDO8/LDO9
LDO10 to be off during suspend
- Improves power of suspend (from ~12mW to ~4mW SoC)
Test: dev3 with bits
Bug: 190063591
Signed-off-by: Patrik Torstensson <totte@google.com>
Change-Id: I6d2d900f5443e137b82572e1dae3b60b5fe5e910
|
|
Bug: 190677341
Change-Id: Ic3acfdaa48aefd247351f98dbfed1796e28eba1c
|
|
Also fix nanohub spi4 cs config (was still setting up spi0 cs)
Bug: 173617548
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I0d78680a88072916c4093c2406e2127b08378ac6
|