Age | Commit message (Collapse) | Author |
|
* goog/upstream-master:
MockNuggetClient: add Reset method to mock
Define UPGRADE_EN_FW_FAIL for upgrade state
sjtag: add new command for sjtag user consent cfg
identity: Add support NDK version 3 for Andorid 12
Tweak: fix typos in debug strings
Refactor NuggetClient a bit
Bug: 170775487
Bug: 178353028
Bug: 181691419
Test: release tests, QMC
Change-Id: I03dc7c270939fc049350eac4ab4625f7f74c37be
|
|
Bug: None
Test: None
Change-Id: I922758e7c64caa339b1ce03eb5e44e3c685ead72
Signed-off-by: Bill Richardson <wfrichar@google.com>
Reviewed-on: https://nugget-os-review.googlesource.com/c/host/generic/+/40284
Presubmit-Verified: TreeHugger Robot
|
|
Added SPDX-license-identifier-Apache-2.0 to:
libnos/Android.bp
libnos/generator/Android.bp
libnos/generator/test/Android.bp
libnos/test/Android.bp
libnos_datagram/Android.bp
libnos_transport/Android.bp
nugget/proto/nugget/app/avb/Android.bp
nugget/proto/nugget/app/identity/Android.bp
nugget/proto/nugget/app/keymaster/Android.bp
nugget/proto/nugget/app/weaver/Android.bp
Added SPDX-license-identifier-Apache-2.0 SPDX-license-identifier-BSD to:
Android.bp
nugget/proto/Android.bp
Bug: 68860345
Bug: 151177513
Bug: 151953481
Test: m all
Exempt-From-Owner-Approval: janitorial work
Change-Id: I81ad1b4c58c6d57550857471dc894063619162c3
|
|
Wiping data can't be finished within 300 msec
that leads the polling timeout APP_ERROR_IO while
doing factory reset.
Bug: 163076061
Test: succeed to finish factory reset 100 times
Change-Id: Iee106ffc02db385270da605fad4119f31c8ec4cc
Signed-off-by: Chungjui Fan <chungjuifan@google.com>
|
|
This allows us to run bazel from anywhere in the nugget checkout
and have it function correctly.
Bug: 120442369
Test: bazel build ...
Signed-off-by: Kenny Root <kroot@google.com>
Change-Id: I0cc2404cc807df1a5d4b2128162c204ab07720ee
|
|
Consolidate each git repo to only have one WORKSPACE. This makes the rules
a bit more easy to manage.
Bug: 120442369
Test: bazel build ...
Signed-off-by: Kenny Root <kroot@google.com>
Change-Id: I00a8f9f8635b0e500e10c0852bc1bb52073994dd
|
|
Citadel chip sometimes give the response "APP_ERROR_TOO_MUCH", which
is to complain AP sent it a count different from what AP claimed or
more than it can accept - but this should not happen.
Give to the chip a little bit of time and retry calling again.
Also increase CRC_RETRY_COUNT from 3 to 5.
Bug: 147200739
Change-Id: I10308cecab382b29c7846d9b79ff3edf36534c8d
Signed-off-by: dybertwang <dybertwang@google.com>
|
|
We don't need to see it, and some hypothetical attacker *might*
find it slightly useful to know.
Bug: None
Test: manual (logcat | grep libnos_transport)
Change-Id: I701514b3da58dee7107340c578aa778f9dfa9bfa
Signed-off-by: Bill Richardson <wfrichar@google.com>
|
|
The new 1.8.1 release has its own BUILD.bazel and doesn't require an
overridden build file.
Bug: 123083638
Test: bazel build :all
Change-Id: Icd7b42f8ad01f4a6035b240f71f7e72a026cedc6
Signed-off-by: Garret Kelly <gdk@google.com>
|
|
If the app doesn't respond within 60 seconds then we give up on it. It
may finish at some point in the future but we'll just have to see next
time we call in as there isn't a way to have the app abort the call.
Bug: 110758319
Test: TransportTest.Timeout
Change-Id: I1448df0a266515c813868f5f9327d25ecb7e933f
|
|
Add the app ID and details about the call so we can infer what the AP is
doing with Citadel. The app ID acts as a transaction ID as there will
only be one transaction with an app at a time.
This also fixes the logging priorities. Debug is a higher priority than
verbose and verbose is off by default in the Android build. The tracing
logs should be in debug and the extra spammy details are in verbose so
they can be enabled when needed.
Test: run on a phone and view the logs
Change-Id: I0a11e04c56ec64f8018c041055fe247deb633d9d
|
|
This adds CRC-16 checksums for the call_application() function. The
command from the master includes and additional packet containing the
CRC which the slave can verify. The status from the slave after an app
has processed the command included the CRC of the reply data and the
status message itself.
The protocol is backward compatible by detecting the v0 protocol
from the status response and ignoring any CRCs.
In the event of a CRC failure when reading the status or response data,
the read operation is retried 3 times. In the event of a CRC failure of
the request, the transaction is restarted a maximum of 3 times.
Bug: 66104849
Bug: 80133226
Test: TransportTest
Test: Update Android and unlock, update Citadel and can still unlock,
reboot and unlock again
Change-Id: If0d37ee492ecddcc17c4dbde3ca992698a48a5e7
|
|
Just adding the NOTICE files next to Android.bp is enough to make
the binaries appear in the Legal Info file on the phone.
Bug: b/69868595
Test: Build, observe files appear in the list
Change-Id: Ib5cde48a4a494faaed08fa0c52bb72250e06c54a
Signed-off-by: Bill Richardson <wfrichar@google.com>
|
|
Originally it as 125ms which only gave a small margin over the expected
~100ms of a reboot. By increasing this margin we aim to improve
reliability.
Bug: 80238564
Change-Id: Ic67dace2c654f42164a4da7518b2f568bbff5552
|
|
This reverts commit 8cec2e5d378b530dcf3de66ac2d5b52602aec6e7.
We need fastboot support first
Change-Id: I67251eb7ccb8f3416218c96aa3eb9de2f5e74193
Signed-off-by: Bill Richardson <wfrichar@google.com>
|
|
This adds CRC-16 checksums for the call_application() function. The
command from the master includes and additional packet containing the
CRC which the slave can verify. The status from the slave after an app
has processed the command included the CRC of the reply data and the
status message itself.
The protocol is backward compatible by detecting the legacy protocol
from the status response and ignoring any CRCs.
In the event of a CRC failure when reading the status or response data,
the read operation is retried 3 times. In the event of a CRC failure of
the request, the transaction is restarted a maximum of 3 times.
Bug: 66104849
Change-Id: Id8c8e1ce0425af66a3a41e59cf7bb3843a4bf740
|
|
Test: harness tests passed 80 times
Bug: 73959155
Change-Id: I9a664e9584eb7ee04f0f94bbcee7763b0e3a19bb
|
|
In case of chip is asleep then Linux kernel driver returns
EAGAIN error code for get_status operation.
In this case we retry the operation again. Currently limit is set to
10 retries.
Remove retry logic from clear_status() that is duplicated now.
Test: harness tests pass
Bug: 73959155
Change-Id: I72ae3e0b7f6002a61147bbd96d668a5f690c135c
|
|
Change-Id: I6fef846a3944da91fdcc6ec33add5ea3f9dd95dd
Signed-off-by: Bill Richardson <wfrichar@google.com>
|
|
So they're easier to spot in the logcat output
Change-Id: Ia7928ba7de9b92a164af343825a2a659db51b32e
Signed-off-by: Bill Richardson <wfrichar@google.com>
|
|
Change-Id: I60d81c28946847f001a4f75bb56c9d7306a1a4fb
|
|
Change-Id: I094a80d0f4f12b6bdddf3709e30386738c853cac
|
|
Change-Id: If4959b2ca940fe4c6d1eb2202480afbddc033c7f
|
|
Instead of killing the process, it logs the error and returns an error
code.
Change-Id: Ia71ccf6fc5a93abf7a07dd7d8004d779df0b46cb
|
|
This repo is going to mostly use the Google style guidelines. This code
has been moving around between repos and the most noticable style
difference is tabs vs spaces IMO.
Change-Id: Iccfa3e3ce00c2350815c0fd9a97da933ac301f4a
|
|
It now resides in host/linux and does not depend on the driver from
core/nugget/util/poker.
Change-Id: I555631bcf659e31f647a4f12d32c10883b10412f
|
|
Oops.
Change-Id: I35bf99d564546af013dd0e7c2bce628f0fe3b8c8
Signed-off-by: Bill Richardson <wfrichar@google.com>
|
|
libnos_transport implements the transport protocol on top of
libnos_datagram which defines the datagram API.
Change-Id: Ie91e4072cac953b244a7f6fed381d67b561ff48d
|