Age | Commit message (Collapse) | Author |
|
Enabling system_dlkm partition support. We plan to use it to
flash and use system_dlkm.img prebuilt from ci.android.com
directly. For now this partition is just a stub or place
holder while we segregate vendor kernel modules into
vendor_ramdisk and vendor_dlkm groups.
We are re-using dlkm_vendor service, hosted by Cuttelfish
project, which will take care of modprobing kernel modules.
Change-Id: I7291fcd430ec76e4f33c9d2d23c4ee8aef159c76
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
|
|
Refactored the scripts which we use on DB845c to set
Ethernet MAC address.
Updated the .sh script to fall back to parsing bootconfig
file to lookup androidboot.serialno bootarg, if it is not
found in /proc/cmdline (boot image header v4 for example).
Fixed a typo s/db45c/db845c in .rc script which was blocking
this service to run.
Renamed the script from eth_mac_addr to set_ethaddr to
follow the nomenclature we have used everywhere else and
moved them to common shared/utils/ directory.
And updated the sepolicies accordingly.
Change-Id: Ia25e464b7c13934ffd8ab66a6aa2d892d7cd5a7b
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
|
|
The Bluetooth chip does not come with a unique Bluetooth device (MAC)
address configured out of the box. This means that we manually need
to configure the device address.
Otherwise BT will stop working from v6.5 kernel version onwards.
Reference: https://bugs.linaro.org/show_bug.cgi?id=5998
Link: https://github.com/me176c-dev/android_device_asus_K013/commit/cbb7066
[AmitP: Kanged the service from above link, updated the commit
message, built an rc and shell script around this service,
and added SELinux policies.]
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I79a364917e0a4d879e4319b45b4c0f7cc7df83d3
|
|
We have used qcom/ folder to host various services, scripts
and utilites, and naming the folder to qcom/, to begin with,
was very shortsighted from my end.
So we are renaming it to utils/ and moving it under a new
shared/ folder which should hopefully be our go to
destination for similar common tools/HALs going forward.
Change-Id: I15fe45dbe5bc4752d6a2ee87a3fbed4924a66f08
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
|
|
Switch DB845c and RB5 to using android-6.1 prebuilt GKI
binaries by default.
Change-Id: Ica022c3803eacbffc66baa93417eb8ddb931ef39
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
|
|
with the "+=" operation, instead of the overwriting
operation of ":=", to avoid unexpected build errors
when inherit-product from other files.
Bug: 265084621
Test: build db845c successfully
Change-Id: Id0e7e444af2069b34bb36245163aac9799e6e3d1
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
|
|
ABL has VERIFIED_BOOT* variables disabled during build,
so it never get around to set "ro.boot.verifiedbootstate"
property on dragonboards.
But recent system/core changes depend on
ro.boot.verifiedbootstate property checks and that broke
adb remount on dragonboards:
$ adb remount
Device must be bootloader unlocked
So we are setting that property explicitly now.
Change-Id: I58b06af41bf6957b31891f62559226c028865585
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
|
|
The default baud rate on kernel releases prior to v5.10
is 9600, which breaks serial console for db845c and AOSP
takes a loooooong time to boot.
Upstream resets default baud rate to 115200 in commit
2ec812a04ff4 ("tty: serial: qcom_geni_serial: 115.2 is a
better console default than 9600").
Fixes: a0d7e989f159 ("Enable serial console on dragonboard devices.")
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I7f1f758a5a6ad2645f7b628e0ab2eb6effe0f7ed
|
|
While running system/core/fs_mgr/tests/adb-remount-test.sh
device-mapper reports h/w logical block size to be 4096.
So updating the BOARD_FLASH_BLOCK_SIZE accordingly.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: Ia300194969da0bfb5ab47b77b130da7be3acc577
|
|
Boot image header v3 remains the default header version for
db845c-userdebug builds because android11-5.4 kernel do not
have bootconfig support. To build v4 header images set
TARGET_USES_BOOT_HDR_V3=false at build time. Also removed
v2 header support for DB845c because of lack of users.
Boot image header v4 is the default header version for
rb5-userdebug builds.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: Icdfef0d348a87fb10a0f56f1a98490070077aa55
|
|
A recent android-mainline kernel commit 01e249da0a52 ("ANDROID: GKI:
Enable CONFIG_NULL_TTY by default") enabled CONFIG_NULL_TTY in the
GKI, which caused the serial console to be disabled by default. This
is a reasonable default for production devices, but for development
boards it makes more sense to have it be enabled. Do so by passing
the appropriate console flags in the boot image.
Change-Id: I21e2a53a4f46a698060297dbff717a75816117df
|
|
android13-5.15 commit 98f1e6b5a0db ("ANDROID: gki_defconfig: Disable
SERIAL_QCOM_GENI_CONSOLE_DEFAULT_ENABLED") broke console access on
DB845c and RB5. Renable the console access by adding
qcom_geni_serial.con_enabled=1 to the kernel cmdline as suggested in
the commit message.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: Ibce09eecd4cbb44d7b3d2841956eec0b6bf042d1
|
|
Move TARGET_USES_BOOT_HDR_V3 build flag from db845c/BoardConfig.mk
to db845c.mk, so that TARGET_USES_BOOT_HDR_V3 flag can be used in
vendor/ firmware package' .mk files.
Needed for certain conditional PRODUCT_COPY_FILES in vendor package
to work correctly.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: Ia81e4cad55d7a8de20d5353e7cfd3d297fa6482c
|
|
"master-I2a659a980d32912bec5a1c0869c88e06a29cf9ce", "master-I6852683ad14a8fa6b0fc22f39e0c9c7823d1da2c", "master-I8f8b467a4dfad68f6e2e53954b5944c6155de782", "master-I9ef8e57207aa96ec1f8193e8d8cd5dfc47567d38", "master-Id17d162d142a2595e54e163f5faaf7dcf9b68e35"
* changes:
Add a sm8450_mini target.
Create a mini configuration for the dragonboards.
Factor out the full device configuration.
Move the architecture definitions into the device-specific BoardConfigs.
Remove pixel3_mainline target.
|
|
In an upcoming patch we are going to add "mini" device configurations
that avoid most of the usual driver dependencies. This requires
splitting out the device-specific configuration from the configuration
required for a full device image.
Bug: 217455793
Change-Id: Id17d162d142a2595e54e163f5faaf7dcf9b68e35
|
|
Newer SOC generations may have different levels of CPU feature support
(e.g. SM8450 which is ARMv9), so let's allow the individual boards to
specify their own sets of features.
Bug: 217455793
Change-Id: I6852683ad14a8fa6b0fc22f39e0c9c7823d1da2c
|
|
Re-enable vendor_boot by default for the db845c build
This was earlier enabled and then reverted. But should
be ok to re-enable now, as long as everyone is using
the flash-all.sh script that flashes vendor_boot.
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I819532a5a4d0e1ca41ac4947dd18f4932ecb2a78
|
|
* Enabled multiple appended DTBs support to allow db845c
boot.img/vendor_boot.img to boot on rb5 as well.
* Introduced a vendor.hw property which will be used to run
device specific services.
* Folded Caleb's HDMI audio routing fix.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I779d01946ae359d5c405d959419e29460083e5df
|
|
Moving all the vendor binaries (including bootloader, qdl
and firmware binaries) to an external vendor.tgz package
to be downloaded from:
https://releases.linaro.org/android/aosp-linaro-vendor-package/extract-linaro_devices-20220106.tgz
BUILD_BROKEN_ELF_PREBUILT_PRODUCT_COPY_FILES build flag is
used to copy a few vendor firmware files to ramdisk or
vendor_ramdisk to work-around the dependency of their
drivers on FW_LOADER_USER_HELPER_FALLBACK kernel config.
And for that we needed to bump up the boot and vendor_boot
partition sizes as well.
Also folded John's installer script changes into this patch.
Change-Id: I1b46ee17f4a218c735e2ac8d81759364b3bf7332
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
[Reworked to handle versioned vendor/linaro/<dev> paths]
Signed-off-by: John Stultz <john.stultz@linaro.org>
|
|
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I8ec9dbd971d7d7590d0342ca3e8bc3ddaa116c9c
|
|
Upstream commit eb9b7bfd5954 ("arm64: dts: qcom: Harmonize DWC
USB3 DT nodes name") (v5.14-rc1) changed the DTS USB node names,
breaking the sys.usb.controller property hardcoded in the
platfrom specific init.usb.common.rc
This script will get rid of the static/hardcoded proprty name
and set the property to the available on-board USB controller
from /sys/class/udc instead.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I1255cd9a3456510d93437989e7e378d9e6bc0c8b
|
|
Update PRODUCT_SHIPPING_API_LEVEL and FCM for db845c so they are
at 31 and 6 respectively.
This also dropps the dummy memtrack HAL, as its no longer required.
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I96fb44dbb54d8f51dee917175342ee734c074512
|
|
Set ro.sf.lcd_density to 160 at build time because
the default density (FALLBACK_DENSITY=213) is a bit
too high for the device.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I33fb0f028199aa236a775347ab442ba1d384da18
|
|
Copy in the latest firmware files from linux-firmware as of
commit ecdfcf8e2ca1 ("Merge branch 'adlp_dmc_firmware'
of git://anongit.freedesktop.org/drm/drm-firmware into main")
from:
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/
This fixes the use of new firmware names which was causing
regressions with upstream kernels.
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I419a15c6f2ce64fa68560000eb273a861e238e17
|
|
This adds the properties:
* ro.soc.model
* ro.soc.manufacturer
Bug: 181845848
Signed-off-by: Will McVicker <willmcvicker@google.com>
Change-Id: Ia8f75103d32a2679931f792be2ccb08a918663be
|
|
Copy in the latest a630_sqe.fw from linux-firmware as of commit
3f026a2f13a8 ("rtl_bt: Update RTL8822C BT(UART I/F) FW to
0x59A_76A3") from:
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/
This resolves boot issues caused by kernels 5.12+ requiring
newer seq ucode.
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I45ff1d87a1686fe1afa7ed8df275abb2837ef40e
|
|
This reverts commit 63f8e16a54cbeffac39255f6b60b5bfbfcb070fa.
Reason for revert: Appears to break boot in lab (odd since firmware was updated)
Bug: 183952621
Change-Id: I98d5a6e651c421ab7dd41c1b4b5b0d8d79d76033
Signed-off-by: Todd Kjos <tkjos@google.com>
|
|
enable vendor_boot by default for the db845c build
Test: boot tested with the bootv3 images
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
Change-Id: I792eefc4b825cb936ccd0d942fe2b94457d1e06c
|
|
The maximum size of all dynamic partition groups should not exceed
the super size - DAP metadata size. Today the configuration of
the device doesn't take the metadata into acount.
Since we recommend 4M to partners, reserve 4M here.
Bug: 182431975
Test: build
Change-Id: I648017e4737a96d663449b8b27f1c6c19730e6cd
|
|
To enable boot image header v3 support, set
"TARGET_USES_BOOT_HDR_V3=true" while building AOSP images
for db845c. Build will fall back to boot image header v2
otherwise.
Boot image header v3 support depends on the latest
bootloader binaries. Run
device/linaro/dragonboard/installer/db845c/flash-all-aosp.sh
script to update the bootloader binaries on db845c.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: Iaf565fc25c731a9d42c8fbc550686b588caba5df
|
|
All dynamic partitions are A/B.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I9650626b5a463ff9f368680ae8c32b8432fee1e8
|
|
so that it will be possible to work with GSI image
Test: boot to homescreen
console:/ # getprop init.svc.vendor.boot-hal-1-1
running
console:/ #
VtsHalBootV1_0TargetTest and VtsHalBootV1_1TargetTest
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
Change-Id: I147f04de1c8bc7fd506d351f514498341ef7388c
|
|
Super partition size is not set correctly. I rounded off
to 10G during initial porting days. Reset it to the
correct size mentioned in the bootloader ptable.
It fixes the following warning we see during adb remount:
[liblp]Device size does not match (got 12437225472, expected 10737418240)
[liblp]Block device super size mismatch (expected10737418240, got 12437225472)
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I89d74070893c14796e24cc8286fd794725b48717
|
|
The overlayfs remount handling of superpartitions expects the
fstab to have all the partitions in it.
This means we can't use the split between the ramdisk fstab
and the vendor fstab that we have been previously.
So this patch merges the two fstabs together.
With this change "adb root; adb remount" works now.
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: Iad408d5d3b1e4c9291a947cd44ee65f4aaef5c45
|
|
This reverts commit 2cf92a4173d774ad491709f80fe8fe3aa1698e6a.
Needed to get WiFi working on android-5.4 kernel. Otherwise
we need to backport a chunk of in-kernel qrtr ns patches
from upstream (v5.7-rc1) kernel. Userspace qrtr-ns tool is
compatible with upstream kernel as well, check:
https://android-review.googlesource.com/c/1286419
Change-Id: Icf9654ed415fe77c8f22592feccd7c68f46bd765
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
|
|
This reverts commit 05e94c35cb0cdba07582fa0055805f9553ab0bf3.
Reason for revert: No longer necessary as default fw_devlink=on works
Change-Id: I906341051416c3e622f29cf1df4a4e54368de83c
|
|
android-mainline commit
("ANDROID: GKI: driver core: Set fw_devlink=on by default"),
https://android-review.googlesource.com/c/1305327
broke db845c.
Set fw_devlink=permissive while we debug it further.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: If49ed679632a0a5c3ffc26bcc4fff62619f474d4
|
|
No need to run qrtr-ns and qrtr-cfg services
explicitly. They are moved to kernel now
starting from v5.7-rc1.
We are keeping qrtr-lookup around for debugging.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I02d72e73ef3d64fed1c024a97cec3560241aff22
|
|
Downloaded from:
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/qcom/sdm845
at commit 64dba0fedb22 ("Merge branch 'a630-zap' of https://github.com/andersson/linux-firmware")
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I95e8bd0ef30c7aa58db4fc8e269f13fa62727022
|
|
db845c development boards do not have unique
Ethernet MAC address. All boards share same
MAC address and it cause problems while
running CTS tests on db845c in a lab with
multiple db845c boards connected.
Set the Ethernet MAC address based on board's
unique serial number as a workaround for now.
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Change-Id: I1c5b63b7d6cbf5e55ac40980b5823cc6bb4f44d3
|
|
as it is required by the GetFstabPath funtion defined here:
https://android.googlesource.com/platform/system/core/+/refs/heads/master/fs_mgr/fs_mgr_fstab.cpp#411
and the fstab.ramdisk.common does not need to be copied to vendor partition.
Test: boot up db845c with android-5.4 and android-mainline kernels, and tested with vts-kernel -m VtsKernelEncryptionTest
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
Change-Id: Ib2ec7277e3b4e912546841174bdcb5638578ed33
|
|
With changes landing in 5.7-rc, adding deferred_probe_timeout=30
argument on the boot line allows modules dependencies to be able
to be resolved after init.
Otherwise we may see some modules fail to load due to missing
dependencies and then if the dependencies are loaded after init
begins, the modules waiting will fail to get EPROBE_DEFER errors
and will just fail for good.
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I0715ae5099ec5511e0154b7c88c098a16fde9fb8
|
|
which should be soc@0/1d84000.ufshc,
as the soc item defined in the dts is "soc: soc@0" like here:
https://android.googlesource.com/kernel/common/+/refs/heads/android-mainline/arch/arm64/boot/dts/qcom/sdm845.dtsi#705
otherwise links like /dev/block/by-name/super won't be created during the boot,
and will get problem like this:
[ 8.807222] init: Loaded kernel module /lib/modules/xhci-plat-hcd.ko
[ 8.814077] init: [libfs_mgr]ReadFstabFromDt(): failed to read fstab from dt
[ 8.821719] init: Using Android DT directory /proc/device-tree/firmware/android/
[ 8.852403] init: realpath failed: /dev/block/by-name/super: No such file or directory
[ 8.860633] init: Failed to mount required partitions early ...
[ 8.867386] init: InitFatalReboot: signal 6
and vts test failures for following test cases:
VtsKernelLiblpTest#DeviceTest.BlockDeviceInfo
VtsKernelLiblpTest#DeviceTest.ReadSuperPartitionCurrentSlot
VtsKernelLiblpTest#DeviceTest.ReadSuperPartitionOtherSlot
Test: checked that /dev/block/by-name/super is there,
and the failures of VtsKernelLiblpTest#DeviceTest passed
Signed-off-by: Yongqin Liu <yongqin.liu@linaro.org>
Change-Id: I33e3a4dbd7c0435d561f694dc3103736382791e0
|
|
Include the db845c/firmware directory in build only for db845c
target.
This allows us to add locally the firmware for pixel3_mainline target.
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
Change-Id: I4f3ebcb290243be7e168b42c24a826b994791020
|
|
Apparently I mistyped the filename and didn't notice
it until I recently ran checkbuild against aosp/master.
This fixes it up.
Signed-off-by: John Stultz <john.stultz@linaro.org>
Change-Id: I09df293e7d5ba7e8b3e4e8d8e99d0fb7481c3b2b
|
|
Adds the init.qcom.rc file to the build so the qcom remoteproc
userland code can be started and wifi can be initialized.
This pach was split out from earlier patch by Amit Pundir
<amit.pundir@linaro.org>
Change-Id: I2032e7c2d3d5d3a10b7afd387f97a93235a908cf
Signed-off-by: John Stultz <john.stultz@linaro.org>
|
|
Add Qcom userspace tools and their respective sepolicy rules.
Userspace tools are downloaded from following github:
To trigger loading of wlan firmware on SDM845
git clone https://github.com/andersson/pd-mapper
Userspace reference for net/qrtr in the Linux kernel
git clone https://github.com/andersson/qrtr
Qualcomm Remote Filesystem Service Implementation
git clone https://github.com/andersson/rmtfs
Trivial File Transfer Protocol server over AF_QIPCRTR
git clone https://github.com/andersson/tqftpserv
Change-Id: Ic466af6fef010a9b71c90e38205f49a876b001e2
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
|
|
We've previously not used the remoteproc, modem or wifi firmware
files we've installed, so we haven't noticed that we are placing
them to the wrong path, as the kernel expects them to be under
firmware/qcom/sdm845/.
Fix this so the kernel can find the files when it looks for them.
Change-Id: Iea8415ed53f05b22f446274bc6d73d4267988ec2
Signed-off-by: John Stultz <john.stultz@linaro.org>
|
|
Adds board-2.bin file from:
https://github.com/andersson/linux-firmware/tree/db845c-wifi/ath10k/WCN3990/hw1.0
Change-Id: I5d5a46e02503b9c63066ffe3b4d9c52353e98944
Signed-off-by: John Stultz <john.stultz@linaro.org>
|
|
Update ath10k/WCN3990 firmware files to match files here:
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/ath10k/WCN3990/hw1.0
Change-Id: I7e9571ca12bb94dd1d5a166488ae389743a6489c
Signed-off-by: John Stultz <john.stultz@linaro.org>
|