usb_charger_init() did not call charge_manager_update_charge() if we are
sourcing VBUS. This means we can get stuck with charge_manager_is_seeded()
never returning true, and so charging never starts, and power-on is
prevented.
Change update_vbus_supplier() so it always calls
charge_manager_update_charge(), but with current = 0 when we are sourcing
VBUS.
BUG=b:80203727
BRANCH=none
TEST=Reboot Grunt EC while one USB-C port is VBUS source.
Change-Id: I24c29dc6b9ad9c50254181614a6440d2d055cd5a
Signed-off-by: Edward Hill <ecgh@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1086113
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
Reviewed-by: Jett Rink <jettrink@chromium.org>
With this, whiskers exposes a tablet mode switch to inform the
lid when the base is flipped around.
We take this opportunity to clean up a bit whiskers/keyboard
code:
- Use tablet mode switch instead of lid switch
- Refactor usb_hid_keyboard.c to accept either assistant key
or tablet mode switch, or both.
- Remove bit-field usage in HID report struct, and instead,
generalize with an "extra" field that can be used for
additional key/switches.
BRANCH=none
BUG=b:73133611
TEST=Flash whiskers, see that tablet mode events are sent when
a magnet approaches the hall sensor.
Change-Id: Ibf43bb04fdc867d18d9f318388d1ebd17b49d47f
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1077915
Reviewed-by: Wei-Han Chen <stimim@chromium.org>
There was a recent change to save the manual setting of charge current
and voltage, however it was done so assuming that the parameters were
set via the host command interface. (CL:922069) However, there are times
where the charge voltage/current would like to be manipulated without
booting the AP. This commit simply makes the EC console command work
again.
BUG=None
BRANCH=None
TEST=make -j buildall
TEST=Flash nocturne, `chgstate idle on; charger current 256; charger
voltage 7400`; verify that the charge voltage and current is actually
changed.
Change-Id: Id250d9704f8509162518495556603950248fb267
Signed-off-by: Aseda Aboagye <aaboagye@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1081120
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: Aseda Aboagye <aaboagye@chromium.org>
Reviewed-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-by: Scott Collyer <scollyer@chromium.org>
The factory reset command can be used to enable ccd factory mode. The
command can open ccd if write protect is removed and ccd hasn't been
restricted. Right now we check FWMP and the ccd password before allowing
factory reset. Factory reset cannot be used to get around anything that
disables ccd.
This adds 72 bytes.
BUG=b:77543904
BRANCH=cr50
TEST=Try enabling factory mode using factory reset. Verify setting write
protect, setting the FWMP disable ccd bit, or setting a ccd password
prevents factory reset from enabling factory mode.
Change-Id: I6e203bf6068250f009881aa95c13bc56cb2aa9e7
Signed-off-by: Mary Ruthven <mruthven@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1069369
Commit-Ready: Mary Ruthven <mruthven@chromium.org>
Tested-by: Mary Ruthven <mruthven@chromium.org>
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
Currently, console_init calls cflush() twice, once before
"Console is enabled" string is printed, once afterwards.
The reason is that firmware_ECBootTime looks for that string,
and it may get corrupted/interleaved with others if the EC
is busy during initialization.
The problem here is that the CONSOLE task may have higher
priority than other tasks (for good reasons), but, on boot,
there are other more critical tasks that need to run (e.g.
RW image verification), rather than busy-looping waiting for
the console to be flushed.
By fixing firmware_ECBootTime to not look for the string anymore,
we do not need those 2 console flush.
BRANCH=poppy
BUG=b:35647963
BUG=chromium:687228
CQ-DEPEND=CL:1075832
TEST=Flash staff, see that RW verification starts at 0.001037
instead of 0.028087 (=> 27 ms faster).
TEST=test_that -b $BOARD $IP firmware_ECBootTime
Change-Id: I794e48eb69cc647c4595fd80265adee4a434d566
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1073180
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
We need to control the console channels for cr50 testing, so we need
access to chan even if the console is restricted. Make chan a safe
command so it is always accessible.
BUG=b:80319784
BRANCH=cr50
TEST=on cr50 make sure the command is accessible no matter the console
state
Change-Id: Ia392f32c319c1acf9bb97b97d7f72c7e56427ce3
Signed-off-by: Mary Ruthven <mruthven@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1079452
Commit-Ready: Mary Ruthven <mruthven@chromium.org>
Tested-by: Mary Ruthven <mruthven@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
We're doing a bit of refactoring to break out factory mode into its own
file. Now factory reset and rma reset will be two methods of entering
factory mode. Factory mode can be disabled with the disable_factory
vendor command.
Factory mode means all ccd capabilities are set to Always and WP is
permanently disabled. When factory mode is disabled, all capabilities
are reset to Default and WP is reset to follow battery presence.
This adds 56 bytes.
BUG=none
BRANCH=cr50
TEST=verify rma reset will enable factory mode.
Change-Id: I21c6f7b4341e3a18e213e438bbd17c67739b85fa
Signed-off-by: Mary Ruthven <mruthven@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1069789
Commit-Ready: Mary Ruthven <mruthven@chromium.org>
Tested-by: Mary Ruthven <mruthven@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Using the p256 curve is beneficial, because RMA feature is currently
the only user of the x25519 curve in Cr50, whereas p256 support is
required by other subsystems and its implementation is based on
dcrypto.
The p256 public key is 65 bytes in size, appropriate adjustments are
being made for the structure storing the server public key and the key
ID.
The compact representation of the p256 public key requires 33 bytes,
including the X coordinate and one extra byte used to communicate if
the omitted Y coordinate is odd or even.
The challenge structure communicated to the RMA server allows exactly
32 bytes for the public key. To comply, the generated ephemeral public
key is used in compressed form (only the X coordinate is used).
For the server to properly uncompress the public key one extra bit is
required, to indicate if the original key's Y coordinate is odd or
even. Since there is no room for the extra bit in the challenge
structure, a convention is used where the generated ephemeral public
key is guaranteed to have an odd Y coordinate.
When generating the ephemeral key, the Y coordinate is checked, and if
it is even, generation attempt is repeated.
Some clean up is also included: even with debug enabled, generated
challenge is displayed only once as a long string, convenient for
copying and pasting.
The new feature is not yet enabled, p256 support on the RMA server
side is not yet available.
Enabling p256 curve for RMA authentication saves 5336 bytes of the
flash space.
BRANCH=cr50, cr50-mp
BUG=b:73296606
TEST=enabled CONFIG_RMA_AUTH_USE_P256 in board.h, generated challenge
and verified matching auth code generated by the rma_reset
utility.
Change-Id: I857543c89a7c33c6fc2dc00e142fe9fa6fc642cf
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1074743
Reviewed-by: Randall Spangler <rspangler@chromium.org>
ID pins are considered additional KSOs while keycode scanning works
for the existing KSI0 ~ KSI7. While diriving ID pins, the state of
interconnection between ID pins and KSI pins could be used for
identifiers to tell keyboard itself. (e.g. US, Japan,and UK keyboard)
BRANCH=master
BUG=b:80168723
TEST="make -j buildall"
TEST=Verified 5 distinct keyboard samples w/ different Language ID values
on the same reworked Coral, which VOL_UP and VOL_DOWN were reworked
for ID pins. crrev.com/c/1053617 is my experimental patch on top of
this for further verification
Change-Id: I1d6e647df74c50d60bc1264c045b2587d0bf23d8
Signed-off-by: paris_yeh <pyeh@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1068951
Commit-Ready: Paris Yeh <pyeh@chromium.org>
Tested-by: Paris Yeh <pyeh@chromium.org>
Reviewed-by: Paris Yeh <pyeh@chromium.org>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
We have determined the checks to run for board_is_first_factory_boot.
This change updates cr50 to check for those conditions and enable ccd
when the system determines that it is first boot in the factory. This
will check that the board id is erased and the inactive image is a GUC
image.
The factory updates Cr50 from the GUC image, because those GUC images
don't have support for everything they need to do in the factory. To
determine that cr50 just recovered from that factory update, it will
check that the GUC image is still in the inactive region and no board id
is set.
There are 2 images installed in GUC 0.0.13 and 0.0.22, so cr50 will
check these versions. Future GUC images will have a field in the header
declaring that they are a GUC image. I still need to create the GUC
field in the header and check that in inactive_image_is_guc_image.
Factory mode can't be enabled on deep sleep resume. It is only enabled
after power-on reset or hard reset.
This change also moves factory stuff into a factory_mode file instead of
keeping it in board.c
This adds 200 bytes.
BUG=b:77543904
BRANCH=cr50
TEST=Verify factory mode is only enabled when cr50 recovered from
reboot not deep sleep resume, 0.0.13 or 0.0.22 are in the inactive
region, and the board id is erased.
Change-Id: Ibece878049658493e8ad159121ada63d7a6f6b79
Signed-off-by: Mary Ruthven <mruthven@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1059864
Commit-Ready: Mary Ruthven <mruthven@chromium.org>
Tested-by: Mary Ruthven <mruthven@chromium.org>
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
There is a potential loop:
(1) We throttle AP when we see under-voltage.
(2) VBAT bumps because throttling starts. From our experiment,
AP throttling saves ~1A, and thus VBAT increases by ~80mV.
(3) VBAT hasn't hit BAT_LOW_VOLTAGE_THRESH for BAT_UVP_TIMEOUT_US,
so we stop throttling.
(4) VBAT again drops below BAT_LOW_VOLTAGE_THRESH.
(5) Go back to (1).
So let's introduce a hysteresis to under-voltage throttling.
We stop throttling only when we are confident that even if we stop
throttling, the battery voltage will stay above BAT_LOW_VOLTAGE_THRESH.
BUG=b:73050145, chromium:838754
BRANCH=scarlet
TEST=manually test on scarlet
Change-Id: Ic0c17a7d37d5d6ee38c7b19f9b65d17421e55cbc
Signed-off-by: Philip Chen <philipchen@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1070568
Commit-Ready: Philip Chen <philipchen@chromium.org>
Tested-by: Philip Chen <philipchen@chromium.org>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
Some devices have GPIO pins that control USB port power connected to
the EC, so they cannot be toggled by ACPI. This patch adds a memory
map between the EC and ACPI that can be used on such devices. It can
hold the power state of up to 8 USB ports. Currently, only dumb power
ports are supported.
BUG=chromium:833436
BRANCH=fizz
TEST=On a fizz that runs BIOS with EC_ACPI_MEM_USB_PORT_POWER mapped,
check that both reads and writes are propagated.
Change-Id: I413defcb9e4d234fea7f54d46b6b8a1a10efa31e
Signed-off-by: Emil Lundmark <lndmrk@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1069273
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
In RSA, we often need to actually compute (a*b)+c+d: provide some
assembly optimized functions for that.
With -O3, 3072-bit exponent, lower verification time from 104 ms to
88 ms on STM32F072 @48Mhz.
BRANCH=poppy
BUG=b:35647963
BUG=b:77608104
TEST=On staff, flash, verification successful
TEST=make test-rsa, make test-rsa3
TEST=make BOARD=hammer test-utils test-rsa3, test on board
Change-Id: I80e8a7258d091e4f6adea11797729ac657dfd85d
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1071411
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
Dumb USB ports do not have the same notion of charge mode as smart
ports. However, the header common/usb_charge.h declares a function for
changing charge mode that the dumb USB port power implementation does
not define. Instead, it defines a similar function with a different
name, albeit with other allowed values for its second parameter.
This patch makes the names the same so the function can be used by
simply including the aforementioned header file.
BUG=none
BRANCH=fizz
TEST=emerge-fizz chromeos-ec
Change-Id: I87863f87f32f538cc1c723d9299afcc7353e1852
Signed-off-by: Emil Lundmark <lndmrk@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1069272
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
tcpm_get_chip_info does not modify the parameter info if the
function is not implemented on the given chip.
This has a interesting side effect on Kevin, where, for whatever
reason, info's value ends up being 1, and causes unaligned access
exception:
[0.039 TCPC p0 init ready]
=== PROCESS EXCEPTION: 06 ====== xPSR: 61000000 ===
r0 :00000000 r1 :28000000 r2 :00000001 r3 :00000000
r4 :00000000 r5 :00000000 r6 :00000000 r7 :00000000
r8 :200c60f4 r9 :00000000 r10:100c0c3c r11:00000000
r12:200c52ed sp :200c5320 lr :100a9c71 pc :100abd5a
Unaligned
mmfs = 1000000, shcsr = 70008, hfsr = 0, dfsr = 0
BRANCH=none
BUG=none
TEST=On ToT make BOARD=kevin -j; flash, kevin boots
Change-Id: Ie7e758d5fb8c31180f36b073b635e54cc720a8a0
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1073179
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
We multiply 2 32-bit numbers (and not 64-bit numbers), and then add
another 32-bit number, which makes it possible to optimize the
assembly and save a few instructions.
With -O3, 3072-bit exponent, lower verification time from 122 ms to
104 ms on STM32F072 @48Mhz.
Optimized mac function from Dmitry Grinberg <dmitrygr@google.com>.
BRANCH=poppy
BUG=b:35647963
BUG=b:77608104
TEST=On staff, flash, verification successful
TEST=make test-rsa, make test-rsa3
TEST=Flash test-utils and test-rsa to hammer => pass
Change-Id: I584c54c631a3f59f691849a279b308e8d4b4b22d
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/449024
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
There are still more ifdef than can be added: this just takes out
the low hanging fruits.
BRANCH=poppy
BUG=b:35647963
TEST=make buildall -j, see that we gain from 0 to 64 bytes on many
boards.
Change-Id: Ibe85b8bfa5d5c22c160e4a6656104256067beee9
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1070948
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
Allow setting password from the AP, but not from USB. Remove the old
password control logic, which is no longer needed.
Allow open if:
- Not explicitly blocked
- Not blocked via FWMP
- One of the following is true:
- A password is set
- Battery is removed (also doesn't require physical presence)
- Dev mode is on, and request came from the AP
Reduces cr50 binary by 152 bytes.
BUG=b:79983505
BRANCH=cr50
TEST=manual, with a CR50_DEV=1 build
ccd oops
ccd lock
ccd unlock -> fails
gsctool -U -> fails from host
gsctool -t -U -> fails from AP
ccd oops
ccd password foo -> fails from console
gsctool -P -> fails from host
gsctool -t -P -> works from AP
ccd get -> confirms password set
ccd lock
ccd unlock foo -> works
ccd lock
gsctool -U -> works from host, if correct password supplied
ccd lock
gsctool -t -U -> works from AP, if correct password supplied
ccd open foo -> works
ccd lock
gsctool -O -> works from host, if correct password supplied
ccd lock
gsctool -t -O -> works from AP, if correct password supplied
ccd oops
ccd lock
(remove battery)
ccd open -> works without physical presence
(reattach battery)
ccd lock
gsctool -O -> works from host
ccd lock
gsctool -t -O -> works from AP, if dev mode is enabled
Change-Id: I364b322d03db250e7dd140767d7a22dbb3ac1eef
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1072957
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
This is needed so ccd_open() can see p->flags, for a subsequent
change. No change to existing command behavior or binary size.
BUG=b:79983505
BRANCH=cr50
TEST=gsctool -I still works
Change-Id: I614d8c410e8bc55a5045e253469b2ec222078684
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1072500
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
On hammer, we do not need the console channels, so we can just
disable them to save flash size.
BRANCH=poppy
BUG=b:35647963
TEST=make newsizes, staff image size shrinks by 704 bytes
Change-Id: I7a493ae57573814b166d45e57f1ad3d885f26086
Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1070949
Reviewed-by: Randall Spangler <rspangler@chromium.org>
This makes it easier to add params or flags for vendor commands
without changing all of the command handlers. It also reduces code
size by 56 bytes.
For now, existing command handlers continue to use
DECLARE_VENDOR_COMMAND(). Added DECLARE_VENDOR_COMMAND_P() for
handlers which take the params struct directly. The CCD command will
be the first user of that, since it will have different rules for
'open' based on where the command comes from.
No change to existing command behavior.
BUG=b:79983505
BRANCH=cr50
TEST=gsctool -I still works
Change-Id: I7ed288a9c45e381162e246b50ae88cf76e67490d
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1069538
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
Previously, calls to tpm_alt_extension() were treated as if they came
from the AP via the TPM interface, even though they actually
originated from the cr50 console, which is accessible via the USB
interface.
This affects the following console commands:
spi_hash - was already allowed as both a safe console command and via
the USB vendor command interface. No change.
rma_auth - was allowed as a safe console command, but not via the USB
vendor command interface. Now allowed from both. No change in
security, since anyone could already do it via the console.
Unfortunately, getting a challenge fails because commands issued via
the USB vendor command interface have a maximum payload of 32 bytes
and the challenge is bigger than that; that's tracked in b:80098603.
ccd - was already allowed as a safe console command. This directly
called ccd_command_wrapper() for lock, open, and password subcommands.
It made an extra check for password set for the unlock subcommand.
Moved the unlock check to the vendor command handler. Also changed
the order of checks so that FWMP disabling unlock and open supersedes
an existing password; this matches go/ccd-open-simple. (That has no
effect on existing systems, because CCD is disabled at a higher
level.)
Reduces code size by 8 bytes.
BUG=b:79983505
BRANCH=cr50
TEST=manual, on a CR50_DEV=1 build
Compile with DEBUG_EXTENSION defined to print extra debug output
'ccd lock' now shows as coming from USB
'ccd unlock' fails because no password is set
'ccd unlock' and 'ccd open' fail if FWMP disallows unlock
'rma_auth' prints a challenge
'gsctool -t -r' prints a challenge from AP root shell
'gsctool -r 12345678' returns error 6 (incorrect challenge), rather
than error 127 (no such command).
'gsctool -I' works from the host
'gsctool -t -I' still works from AP root shell
Change-Id: I2cd1027f5135b9c336df97ee4b1b1a15354728b4
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1068102
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
The values of CROS_EC_VERSION and VERSION variables generated by
getversion.sh are exactly the same. VERSION is used in more places
than CROS_EC_VERSION, let's keep VERSION and use it everywhere.
BRANCH=none
BUG=chromium:632937
TEST=make buildall
Change-Id: Ibec9ecdd4b67789a468dddfbc1c82565f90d48a8
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1069330
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Added flags parameter to extension_route_command(). The caller now
specifies whether the command comes from the USB interface or the AP.
Moved USB-specific shuffling of response to embed result code into
usb_upgrade.c, so extension_route_command() can be more generic.
No change to permissions/behavior for existing commands.
ccd_command_wrapper() still sends vendor commands as if they come from
the AP. That's fixed in the next CL.
Reduces code size by 128 bytes
BUG=b:79983505
BRANCH=cr50
TEST=manual
Build with DEBUG_EXTENSION defined, to turn on printing each command
'ccd lock' comes from AP and works
From host, 'gscutil -I' comes from USB and fails
From AP, 'gscutil -t -I' comes from AP and works
Change-Id: I7136bb54073de9c5951a174c308151b1871c56f3
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1068101
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
We have converted all LPC-only configs to HOSTCMD_LPC so the remaining
CONFIG_LPC defines represent the common case.
BRANCH=none
BUG=chromium:818804
TEST=Full stack builds and works on yorp (espi) and grunt (lpc)
Change-Id: Iba9a48f2cab12fadd0d9ab8eab0d5d5476eab238
Signed-off-by: Jett Rink <jettrink@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1067503
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Break the ec chip code up with the more granular
CONFIG_HOSTCMD_(X86|LPC|ESPI) options.
BRANCH=none
BUG=chromium:818804
TEST=Full stack builds and works on yorp (espi) and grunt (lpc)
Change-Id: Ie272787b2425175fe36b06fcdeeee90ec5ccbe95
Signed-off-by: Jett Rink <jettrink@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1067502
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Expose the i2c interface through usb so that we can read power rails
through servod leveraging the work being done there.
BRANCH=none
BUG=chromium:806148
TEST=manual testing
- powerlog still works
- i2c over usb using servod code works (other CLs needed)
Change-Id: I48876bc4839509a397ce77376b337c37c556ae40
Signed-off-by: Ruben Rodriguez Buchillon <coconutruben@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1051136
Reviewed-by: Nick Sanders <nsanders@chromium.org>
Previously manual_mode used the current values of voltage/current to
set the desired values for each charge_request() call. Since manual
mode is entered/exited in the host command task, this can easily lead
to a race condition where the charger gets disabled in the host
command task, the reenabled by the charger task. This in turn makes
the ectool chargecontrol idle command unreliable.
This CL replaces manual mode with two variables, manual_voltage and
manual_current. The default values are -1 which means that they are
inactive. When the ectool command 'chargecontrol idle' is executed, it
sets both variables to 0. This then removes the race condition
possibility as each iteration of the charger loop will use
manual_voltage and/or manual_current if not -1.
BRANCH=coral
BUG=b:68364154
TEST=Manual
Executed 'ectool chargecontrol idle' and 'ectool chargecontrol normal'
numerous times and verified that the charging was disabled/resumed
each time as expected. Without this fix the problem could be
reproduced always in less than 10 attempts, typcially less than
5. With this CL charging is disabled reliably each time and I'm not
able to reproduce the problem.
Change-Id: I1ed9cdb42249cdf72ab34dd95b8f42c09d9a490c
Signed-off-by: Scott Collyer <scollyer@google.com>
Reviewed-on: https://chromium-review.googlesource.com/851419
Tested-by: Scott Collyer <scollyer@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Shawn N <shawnn@chromium.org>
Commit-Queue: Scott Collyer <scollyer@chromium.org>
(cherry picked from commit b7254f38979f274acc66330905399ff5ddf4129b)
Reviewed-on: https://chromium-review.googlesource.com/922069
Commit-Ready: Daisuke Nojiri <dnojiri@chromium.org>
Tested-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-by: Daisuke Nojiri <dnojiri@chromium.org>
When EC sees voltage drops below BAT_LOW_VOLTAGE_THRESH,
we kick off a timer and ask AP to throttle.
When the timer expires which means EC hasn't seen under-voltage
for BAT_UVP_TIMEOUT_US, we ask AP to stop throttling.
We reset the throttling status and do nothing when AP is off (S5).
BUG=b:73050145, chromium:838754
BRANCH=scarlet
TEST=manually test on scarlet, confirm EC sends
EC_HOST_EVENT_THROTTLE_START and EC_HOST_EVENT_THROTTLE_STOP
host events when entering/exiting UVP.
Change-Id: Ia760989f760f95549f7a8a8acb1d01de23feab5a
Signed-off-by: Philip Chen <philipchen@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1064983
Commit-Ready: Philip Chen <philipchen@chromium.org>
Tested-by: Philip Chen <philipchen@chromium.org>
Reviewed-by: David Schneider <dnschneid@chromium.org>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
sleepmask is really useful for debugging sleep issues. Add a read only
version of sleepmask to non-DBG images. It will only be accessible once
the console is unlocked.
BUG=none
BRANCH=cr50
TEST=make sure sleepmask can be modified in DBG images and can only be
read in prod images.
Change-Id: I31ef966f6302d4a7602a014cb08c9b972d13f41e
Signed-off-by: Mary Ruthven <mruthven@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1062804
Commit-Ready: Mary Ruthven <mruthven@chromium.org>
Tested-by: Mary Ruthven <mruthven@chromium.org>
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
When the EC is powered solely by the servo, we do not want to try to
start the AP. If we do, we will watchdog reset in a while loop waiting
for the 3300 and 5000 rails to come up (which won't come up if powering
only on the servo)
BRANCH=none
BUG=b:79606767
TEST=powering bip with servo only does not watchdog reset boot loop
Change-Id: I132312f7f08201dc58d797900df16502240ee98c
Signed-off-by: Jett Rink <jettrink@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1062502
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-by: Vijay Hiremath <vijay.p.hiremath@intel.corp-partner.google.com>
This will be used as part of the checks for when to allow CCD open.
Add check for firmware space dev mode bit, based on the similar code
which reads the FWMP. Print the state of both bits in 'ccd get'.
BUG=b:79983505
BRANCH=cr50
TEST=With dev mode off, 'ccd get' does not report TPM: dev_mode.
Turn on dev mode via the recovery screen, and it does.
Change-Id: I6af78bb104004323cd377ed996e1db94bc36fc62
Signed-off-by: Randall Spangler <rspangler@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1066391
Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
Observed VBOOT hash failure for EC_RW. Function
spi_flash_read with size > SPI_FLASH_MAX_READ_LEN
is incorrectly incrementing the offset. For example:
0, 0x100, 0x300, 0x600, all with read size = 256.
BUG=
BRANCH=any EC using SPI flash
TEST=Trigger VBOOT hash re-calculation using EC
console hash rw command. Second test program
SPI flash with known test pattern longer than
SPI_FLASH_MAX_READ_LEN and read using EC
console flashread.
Change-Id: I5fda47f132f64b12044b94663a19d889f1c2b32a
Reviewed-on: https://chromium-review.googlesource.com/1036258
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: Scott Worley <scott.worley@microchip.corp-partner.google.com>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
gcc 8.1 complains about duplicate const, and while some of these really
are duplicate, others look like they were supposed to tighten the API
contract so that variables are "const pointer to const data", but didn't
have that effect.
BUG=b:65441143
BRANCH=none
TEST=building Chrome EC as part of upstream coreboot's build with a
gcc 8.1 compiler now works (better. there are other issues left)
Change-Id: I6016c5f282516471746f08d5714ea07ebdd10331
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1039812
Commit-Ready: Patrick Georgi <pgeorgi@chromium.org>
Tested-by: Patrick Georgi <pgeorgi@chromium.org>
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
Make Esc+Refresh+Power on Grunt enter Recovery Mode.
If Power is released fast:
[0.045303 KB init state: -- 02 08 -- -- -- -- -- -- -- -- -- --]
Add CONFIG_KEYBOARD_IGNORE_REFRESH_BOOT_KEY to handle this case.
If Power is held longer:
[0.045448 KB init state: 08 0a 08 08 08 -- 08 -- 08 08 -- 08 08]
Add CONFIG_KEYBOARD_PWRBTN_ASSERTS_KSI3 to handle this case.
BUG=b:79758966
BRANCH=none
TEST=Esc+Refresh+Power gives recovery screen on Grunt
Change-Id: I43a7d485535ff7b0d9bfce59f28c0049ee989818
Signed-off-by: Edward Hill <ecgh@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1063032
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Reviewed-by: Daniel Kurtz <djkurtz@chromium.org>
CPRINTS() macro is already adding the newline character, no need to
include it explicitly.
BRANCH=none
BUG=none
TEST=vefied that Coral EC does not print this newline any more
Change-Id: I1f6b1fcb90818a8d4a2d18cf5060669fc46d38a7
Signed-off-by: Vadim Bendebury <vbendeb@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1062588
Reviewed-by: Jett Rink <jettrink@chromium.org>
The decimal converted value for i2c read does not work.
It just happens to work every other time I have seen it.
BUG=none
BRANCH=none
TEST=bip i2cxfer r works
Change-Id: I7d868e3fc79eea081867634b679120f2da6f9363
Signed-off-by: Jett Rink <jettrink@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1060167
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Add the FP_CAPTURE_RESET_TEST capture mode to be able to perform the
reset pixel values test.
Update ectool accordingly and also remove the deprecated 'fpcheckpixels'
command.
Signed-off-by: Vincent Palatin <vpalatin@chromium.org>
BRANCH=poppy
BUG=b:78597564
TEST=run 'ectool --name=cros_fp fpmode capture test_reset',
then 'ectool --name=cros_fp fpframe > test.pnm'
CQ-DEPEND=CL:*626747
Change-Id: I183f33b1cb9ba4db67219b8f7740d29dc0551f2d
Reviewed-on: https://chromium-review.googlesource.com/1061074
Commit-Ready: Vincent Palatin <vpalatin@chromium.org>
Tested-by: Vincent Palatin <vpalatin@chromium.org>
Reviewed-by: Nicolas Boichat <drinkcat@chromium.org>
charge_manager_fill_power_info can be called to fill the power info
of the dedicated port. This function might call
charge_manager_get_source_current with the dedicated port, we don't
want to use assert in that case, but just fail gracefully by returning
0.
BRANCH=None
BUG=chromium:841944
TEST=Check that the function returns 0 for dedicated and not 0 for USB.
Change-Id: I357c056647e01bdb0e77a08a6c6b492aa3dbb503
Signed-off-by: Fabien Parent <fparent@baylibre.com>
Reviewed-on: https://chromium-review.googlesource.com/1059248
Reviewed-by: Daisuke Nojiri <dnojiri@chromium.org>
The charge_prevent_power_on() function checks for battery presence and
its charge level as a condition to prevent AP power on. In order to
deal with the short (up to a few seconds) time window where the fuel
gauge can be read but the discharge FET is not yet enabled, the
status of battery_get_disconnect_state() has been added as a condition
to many board's battery_is_present() function.
The problem with this approach is that the return value of
battery_is_present() is also used by charge_state_v2 as a condition to
force the charge state to ST_IDLE. Then, if the config option
CONFIG_CHARGER_MAINTAIN_VBAT is not defined, the value of requested
current and voltage will be forced to 0 as long as the state remains
in ST_IDLE. When the battery is dead or has been cutoff its discharge
FET will be disabled. In order to for the discharge FET to be enabled
the charger must provide at least a precharge level of current to the
battery. But, if the FET status is used as a condition for
battery_is_present which in turn forces the charge state to ST_IDLE
which can lead to requested_current being forced to 0.
This CL enables a way to remove battery_get_disconnect_state() as a
condition from battery_is_present and instead call
battery_get_disconnect_state() directly in the function
charge_prevent_power_on. Therefore AP power can still be gated by the
battery FET status, but the charge state will not be stuck in ST_IDLE.
This new check is guarded by CONFIG_BATTERY_REVIVE_DISCONNECT. Boards
which currently condition battery_is_present on the value of
battery_get_disconnect_state() don't change at all, but a board which
needs to remove that condition on battery_is_present can still use the
FET check to prevent AP power on.
BUG=b:79133101
BRANCH=none
TEST=Tested on Yorp and verifed that when using either a dead battery
or a battery that had been cutoff, prevent_power_on was set to 1 until
the FET status was correct for the battery to provide power.
Change-Id: Ic27f42610a7b751394b29a013c4dd17030a3df31
Signed-off-by: Scott Collyer <scollyer@google.com>
Reviewed-on: https://chromium-review.googlesource.com/1053095
Commit-Ready: Scott Collyer <scollyer@chromium.org>
Tested-by: Scott Collyer <scollyer@chromium.org>
Reviewed-by: Aseda Aboagye <aaboagye@chromium.org>
When the EC sends an interrupt to the AP notifying it of new
accelerometer data we need to make sure the spot we record the timestamp
of the event is virtually identical to the spot the AP records the same
point in time.
Therefore a better spot for that is right next to the gpio toggling of
the interrupt line.
BUG=b:67743747
TEST=In the kernel, fifo_info->info.timestamp still has sane values.
TEST=CTS should still pass
BRANCH=master
Change-Id: Ic77101a045123e779f576c46b401c765304976fd
Signed-off-by: Alexandru M Stan <amstan@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/802976
Reviewed-by: Gwendal Grignou <gwendal@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
Instead of fetching incoming CEC messages using a specific read
command, extend the standard mkbp event so the CEC message can
be delivered directly inside the event.
Signed-off-by: Stefan Adolfsson <sadolfsson@chromium.org>
BUG=b:76467407
BRANCH=none
TEST="ectool cec read" still working with a kernel that has support
for the increased mkbp size.
CQ-DEPEND=CL:1046186,CL:1051085
Change-Id: Id9d944be86ba85084b979d1df9057f7f3e7a1fd0
Reviewed-on: https://chromium-review.googlesource.com/1051105
Commit-Ready: Stefan Adolfsson <sadolfsson@chromium.org>
Tested-by: Stefan Adolfsson <sadolfsson@chromium.org>
Reviewed-by: Randall Spangler <rspangler@chromium.org>
the EC should not clear the AC_PRESENT flag in the ACPI flags when we
know that we're on a charger. ACPI rules on the AP side do special
things when they see AC_PRESENT off, like forcing the EC to limit the
charge current to 512mA.
BUG=b:78509594
BRANCH=none
TEST=chargestate current no longer stuck at 512mA
Change-Id: I54ddea6d2661a9d6b1593431a5912a8a60db572d
Signed-off-by: Caveh Jalali <caveh@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1050917
Commit-Ready: Caveh Jalali <caveh@google.com>
Tested-by: Caveh Jalali <caveh@google.com>
Reviewed-by: Duncan Laurie <dlaurie@google.com>
Reviewed-by: Jett Rink <jettrink@chromium.org>
