Unfork Android.
This reverts commit 8e5e42d5ae20f0324170d01ccf374a1571e82d9b.
Perf_event_paranoid=3 is no longer needed on Android. Access control
of perf events is now done by selinux. See:
https://patchwork.kernel.org/patch/11185793/.
Bug: 120445712
Bug: 137092007
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Iba493424174b30baff460caaa25a54a472c87bd4
Git-commit: 44a6aea9c2
Git-repo: https://android.googlesource.com/kernel/common/
Signed-off-by: Srinivasarao P <spathi@codeaurora.org>
Compress the kernel with lz4 instead of gzip. On x86 this must be
controlled explicitly by kbuild, as decompressor code is prepended
to the compressed kernel artifact.
On arm64 this option is not needed as "make Image.lz4" can be used
instead, or the Image can be compressed in a separate step.
Bug: 150391496
Change-Id: I2d0ce7f937f2a952b4310f53c8b03dc8d721b050
Signed-off-by: Alistair Delva <adelva@google.com>
Android expects rtc0 to support saving the system clock, however the
CMOS rtc emulated by crosvm currently only supports 24hr alarms, so
it is not useful. As this useless rtc driver occupied rtc0 it broke
various tests and features. The older crosvm used by cuttlefish did
not simulate an RTC so it was not affected because the driver probe
failed.
Bug: 158000323
Change-Id: I54590ceb0f09e6b1ff1f7ed094dab66ed7e58672
Signed-off-by: Alistair Delva <adelva@google.com>
The helper is set to the empty string.
Bug: 156358723
Change-Id: Ia7923f80a0617814bddd514bdf01a4ceffc3eddb
Signed-off-by: Alistair Delva <adelva@google.com>
Enable LTO and CFI also in x86_64_cuttlefish_defconfig.
Bug: 145297900
Change-Id: Ibf24d4da73789f49fd6a37a540f6b6a9132902dd
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
For lots of good security reasons, this config option needs to be
enabled
Bug: 152470236
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I96a88bbee9c4d17be97ed63262dbab2ef31fee79
This module allows presenting the kernel TPM interface while proxying
the TPM commands into a file descriptor. The module was originally
implemented to support running a TPM simulator on the same host system
and exposing a kernel TPM interface to a Linux container, but it is also
a convenient incremental step while we figure out our long-term strategy
with crosvm, which does not have TPM support following the same
standards as qemu.
Implicitly enables SECURITYFS via 'selects' from these new drivers.
Bug: 148102533
Test: Build and run locally with cuttlefish, check for /dev/vtpmx
Change-Id: I568a50c2ecb7899aae70e7a20efaedc84443511d
Signed-off-by: A. Cody Schuffelen <schuffelen@google.com>
Security teams from Google and Qualcomm request disabling
due to "large attack vector".
Bug: 149659981
Signed-off-by: Todd Kjos <tkjos@google.com>
Change-Id: Iee4ace633f28cf4359cb0495be59546a8d3951cd
This config will enable the Nintendo Switch Pro controller driver.
Change-Id: I50645a611566928e20a1afd4024f71803ed5fefa
Signed-off-by: Siarhei Vishniakou <svv@google.com>
Bug: 135136477
Test: tested via custom test app
Test: atest NintendoSwitchProTest
This adds support for unicode to cuttlefish_defconfig
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Test: Boots, /data/media is case insensitive
Bug: 138322712
Change-Id: I177d0b301087733234de3d6d283af946213eebca
Support future decompression of LZ4-compressed ramdisk images.
This support is in addition to the default support for gzip.
Bug: 150391496
Test: TreeHugger
Change-Id: I8c07e9363333c53cd6ff9c7e16fc42cbe318fe83
Signed-off-by: Alistair Delva <adelva@google.com>
Build the cuttlefish kernel with CONFIG_TEST_STACKINIT=y and
CONFIG_TEST_MEMINIT=y to provide the tests for stack and heap
initialization.
Bug: 144999193
Test: run cuttlefish and observe the following lines in dmesg:
test_stackinit: all tests passed!
test_meminit: all 130 tests passed!
Change-Id: I044347880de82211d9a58449ee3e558c570bda31
Signed-off-by: Alexander Potapenko <glider@google.com>
This patch enables CONFIG_INIT_STACK_ALL=y and
CONFIG_INIT_ON_ALLOC_DEFAULT_ON=y, effectively turning on stack and heap
initialization in cuttlefish kernels.
Doing so will help us mitigate information leaks and make code that
depends on uninitialized memory execute deterministically. We'll also
get coverage for the initialization features on the existing kernel
tests.
Bug: 144999193
Change-Id: I1b5914a87a235d4f416cd64d7c9a5d410da09c35
Signed-off-by: Alexander Potapenko <glider@google.com>
The ability to detach debugfs from tracing resulted in a few defconfig
changes:
-CONFIG_SW_SYNC=y
-CONFIG_DEBUG_BOOT_PARAMS=y
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: If13241fd7555dca3ffd8d0bf8ab089de5f2bda4a
CONFIG_PM_WAKELOCKS are needed by existing driver code that uses
APIs like pm_wake_lock/unlock(). We can't disable yet just yet even
though userspace has stopped using the /sys/power/wake_lock interface.
This reverts commit b82fc46d57da841e02ced5f8e8727aff5510d9bd.
Bug: 148922129
Test: Boot
Change-Id: I8b8f4f46db3647313485abff21a5659bd881b687
Signed-off-by: Sandeep Patil <sspatil@google.com>
dm-default-key is needed for metadata encryption
(https://source.android.com/security/encryption/metadata).
The new version of dm-default-key is vendor-independent and can be used
both with and without inline encryption hardware.
Bug: 137270441
Bug: 147814592
Change-Id: I92ce45c4b1543ff8bc5cdb5ebe4ddcdb740c2dc7
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Satya Tangirala <satyat@google.com>
Enable blk-crypto's kernel crypto API fallback, so that devices
without inline encryption hardware can continue to use the inlinecrypt
mount option and have file content crypto handled by the block layer.
Bug: 137270441
Test: xfstests, cuttlefish boot and stress test, pixel 4 boot and stress
test
Change-Id: I26376479ee38259b8c35732cb3a1d7e15f9b05a3
Signed-off-by: Satya Tangirala <satyat@google.com>
enable CONFIG_BLK_INLINE_ENCRYPTION and
CONFIG_FS_ENCRYPTION_INLINE_CRYPT
Bug: 137270441
Test: Test cuttlefish boots both with and without inlinecrypt mount
option specified in fstab, while using both F2FS and EXT4 for
userdata.img. Also tested by running gce-xfstests on both the
auto and encrypt test groups on EXT4 and F2FS both with and
without the inlinecrypt mount option. The UFS changes were
tested on a Pixel 4 device.
Change-Id: Ie1b77f7615d6a7a60fdc9105c7ab2200d17636a8
Signed-off-by: Satya Tangirala <satyat@google.com>
SCSI_VIRTIO and VIRTIO_BALLOON are not needed to run cuttlefish.
Test: Treehugger
Bug: 135937364
Change-Id: Ie65598c1adc834d2a88799d73eee55bc86d329ae
Signed-off-by: Ram Muthiah <rammuthiah@google.com>
Without this, the virtio_pmem driver does not seem to work.
Bug: 146400078
Change-Id: I2689aa4739b83a0a5a0e9930dc50b57e4f612525
Signed-off-by: Alistair Delva <adelva@google.com>
Options needed for the virtual platform and physical platforms via DT
to define PMEM regions for resume-on-reboot feature
Bug: 146400078
Change-Id: Icaffbd3a7425ac3d9914378fb356d32f9393eec0
Signed-off-by: Kenny Root <kroot@google.com>
Unfork Android.
This reverts commit 8e5e42d5ae20f0324170d01ccf374a1571e82d9b.
Perf_event_paranoid=3 is no longer needed on Android. Access control
of perf events is now done by selinux. See:
https://patchwork.kernel.org/patch/11185793/
Bug: 120445712
Bug: 137092007
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Iba493424174b30baff460caaa25a54a472c87bd4
Quota logging is needed to enable data metering.
Bug: 147203196
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I7219a3abd922b31249a3bccb94bc0dcff197788e
These configs will now be required in aosp/1204022.
Change-Id: Ib2961a9abf545f483cd5691f04b93cffd011ec82
Signed-off-by: Siarhei Vishniakou <svv@google.com>
Bug: 136263708
Test: none
Fix to re-enable wifi on cf + kernel 4.14. Remove mac80211_hwsim
virtual device on cuttlefish kernels as we want to use virt_wifi
now instead.
Test: Local boot with cuttlefish
Bug: 145836345
Change-Id: Ifac04724afef7f5ad5aff46bae9c1ea0e77f892c
Signed-off-by: Ram Muthiah <rammuthiah@google.com>
Fix conflict between jemalloc/scudo and MADV_NOHUGEPAGE by disabling the
transparent hugepage support. It has also been suggested that this
feature can make VM behavior less predictable.
Bug: 131119917
Change-Id: I17556838fbf1f893e26c5658ee95b4e3b16b10ad
Signed-off-by: Alistair Delva <adelva@google.com>
Enables a GNSS serial driver for cuttlefish.
Bug: 146517987
Change-Id: Ife42306a3052e49365e3f0004f220424f0be7782
Signed-off-by: Alistair Delva <adelva@google.com>
The gnss serial core utilizes the serdev API, so we need to enable it.
Bug: 146062677
Change-Id: Id6477f5b1eb0d9cfd240b62b337643f46f490a63
Signed-off-by: Alistair Delva <adelva@google.com>
fs-verity will be used for APK verification in R.
Bug: 142494008
Change-Id: I51626da9a9975d337c44096f8443c3713dfcd6eb
Signed-off-by: Eric Biggers <ebiggers@google.com>
/proc/uid_time_in_state has no data on cuttlefish because its cpu
frequency tables are empty. Because time in state & concurrent time
accounting are intertwined this causes the
/proc/uid_concurrent_{policy,active}_time files to also not contain
any data.
Add a minimal, fake cpufreq driver that creates a freq table with 2
frequencies per policy, to allow testing time in state functionality.
Test: all 3 proc files show reasonable data on cuttlefish
Test: log shows no errors from bad /proc/uid_time_in_state format
Bug: 139763108
Bug: 140796321
Bug: 141206930
Change-Id: I8c7fe1007a80c21a9bcba9455bf837947cf42963
Signed-off-by: Connor O'Brien <connoro@google.com>
The snapshot dm target is required to implement the Virtual-AB
mechanism.
Introduce CONFIG_DM_SNAPSHOT in arm64 and x86 cuttlefish defconfigs to
enable this feature.
Bug: 142527064
Test: kernel build
Signed-off-by: Alessio Balsini <balsini@google.com>
Change-Id: I69bc614509eaff259a2aa9195e7e1d406b36bbb2
Signed-off-by: Alessio Balsini <balsini@google.com>
Simon1511