tangentially - last I checked metalava had hardcoded heap sizes and I still see hardcoded Xmx vars in soong - on low memory systems you could set them differently if they can’t be overridden with env vars
You’re right, it is set to 8GB and under normal circumstances it should not even matter, but sadly with my HW it overruns the allocated size, but it does not matter when producing a system image so I haven’t done anything with it
My plan is to make a large patch of the codebase changes needed after I get this to boot past logo, in theory the change to trebleapp should be the last change needed but I want to keep it as close to stock as possible for bringup to Android 15 when that update comes
do you mean something like e-2.7-u-20240107-UNOFFICIAL-alpha-treble_arm64_bvN.img by @Null
awaiting the 3 years awaited /e/ team GSI upgrade....
last was e-1.17-q from 2022 12 XX
- Arm A (system-arm-a.img.xz)
- Arm A/B (system-arm-ab.img.xz)
- Arm64 A (system-arm64-a.img.xz)
- Arm64 A/B (system-arm64-ab.img.xz)
- Binder 64 A (system-a64-a.img.xz)
- Binder 64 A/B (system-a64-ab.img.xz)
https://doc.e.foundation/support-topics/understand-and-use-gsi
https://doc.e.foundation/support-topics/install-GSI
[HOWTO] install eOS GSI on a Treble device - new “Pie” GSI added 2020 June 14th
It is important to remember that this is just 0.3 of this build, there is bound to be some breakage so please report errors everyone!
On my phone there is one error:
The app_switcher button which brings up windows to switch between does not function, but I’m unsure if this is a code issue or related to the fact that my device has no GSI overlays so I eagerly await feedback so I can address this if it is a problem for others.
Ok, I think the Navbar being broken is related to the following line of code in:
Lineage21/device/lineage/gsi/init
In the file:
init-gsi.rc
The offending code seems to be:
on property:persist.sys.treble.mainkeys=*
setprop qemu.hw.mainkeys ${persist.sys.treble.mainkeys}
I have added # to comment it out, I will push a new build and see if this resolves the issue.
I think that /e/OS handles the Nav bar in-house through the systemUI, it is just a guess and I have not investigated it but I think that this is causing the nav-bar to appear twice.
Test N°6
Nothing Phone (1) A063 spacewar
SoC: Qualcomm SM7325-AE Snapdragon 778G+ 5G
NothingOS 2.6 (A14) Security: Nov. 2024
No custom recovery, no LineaageOS; no /e/OS
system-e-2.7_0.3-u-20240107-by-NULL-arm64_bvN.img by @Null
The /e/OS-U GSI installation on my NP1 went smoothly straight away: The system boots up and can be set up in the usual way. A functioning WLAN Wi-Fi connection is quickly established. Further tests will follow …
Bravo!!! @Null
You have developed an executable /e/OS-U 2.7_0.3 for A/B devices in a very short time. You have proven it: With existing know-how and good will (almost) anything is possible. You have my respect and esteem for your excellent achievement.
S u g g e s t i o n
Compress your system.img into a common format such as *.gz, *.xz, or *.zip.
The download of your 2.9GB system.img file took me over two hours despite a fast fiber optic connection, because the data transfer rate was only a few hundred KB/s.
Photos system start in the attachment
Test N°7
Gigaset GX6 (5G rugged) GX6
SoC: MediaTek Dimensity 900 5G
Stock Android 13 (A13) Kernel: 4.19.191
No custom recovery, no LineaageOS; no /e/OS
The /e/OS-U GSI installation on my GX6 went smoothly straight away: The system boots up and can be set up in the usual way. A functioning WLAN Wi-Fi connection is quickly established. Further tests will follow …
Great work!!! I could do not this.
@ronnz98 you can do anything I can do.
In any event, I wanted to ask you about helping to make generic overlays for /e/OS because it seems there is an issue with the SystemUI on my device, but it presently takes too long for me to build this so I was going to put overlay troubleshooting on pause until I’m back to my HW.
If you would throw your weight behind this project though, it could be brought up a lot, lot more quickly.
Test N°8
Samsung Galaxy A23 5G (a23xq) SM-A236B
SoC: Qualcomm SM6375 Snapdragon 695 5G
Stock Android 14 One UI 6 ::: Kernel: 5.4.254 ::: latest firmware [ U9 ] BIT-9 (11.2024)
TWRP 3.7.1_12-0-a23xq; no LineageOS ROM; no /e/OS ROM;
system-e-2.7_0.3-u-20240107-by-NULL-arm64_bvN.img by @Null
The /e/OS-U GSI e-system does not boot but the reboot ends up in the Samsung Android logo and blocks the system so that I cannot run adb logcat
.
lineage-21.0-20241118-UNOFFICIAL-arm64_bvN.img by @AndyYan
The system boots up and can be set up in the usual way. A functioning WLAN Wi-Fi connection is quickly established.
iode-5.8-20241214-arm64_ab.img (with build-in microG) by @iode.tech
The system boots up and can be set up in the usual way. A functioning WLAN Wi-Fi connection is quickly established.
You should be able to use adb logcat while it is booting before the system even loads, or are you saying that it crashes before even reaching that point?
Perhaps this is an issue with HW overlays for Samsung Galaxy A23 5G
Please help me to learn how to use adb logcat correctly.
Let’s take a classic GSI installation as an example:
…
… fastboot reboot fastboot
… fastboot erase system
… fastboot flash system system.img
… fastboot -w
… fastboot reboot
…
Where, when and how do I enter the adb logcat command?
Use it as soon as the /e/ bootlogo starts coming up, at first it will say ‘waiting for device’ but adb is one of the first things loaded on the phone
Test N°9
Gigaset GS4 Gigaset_GS4
SoC: MediaTek Helio P70 MT6771T
Stock Android 11::: 12/2023 ::: Kernel: 4.14.186
No custom recovery; no LineageOS ROM; no /e/OS ROM;
iode-5.8-20241214-arm64_ab.img (with build-in microG) by @iode.tech
The system boots up and can be set up in the usual way. A functioning WLAN Wi-Fi connection is quickly established.
system-e-2.7_0.3-u-20240107-by-NULL-arm64_bvN.img by @Null
The /e/OS-U GSI e-system does not boot, no e-logo is displayed - the screen remains black. However, I can boot into Android Recovery via hardware keys and from there back into FASTBOOT (to successfully flash iodéOS 5.8 GSI)
I could run the command adb logcat
and see the message waiting for device
, until after a few seconds several thousand lines ran from top to bottom in PC terminal …
... following this pattern
01-08 00:47:11.917 1268 1664 I ActivityManager: Killing 3874:com.google.android.apps.docs/u0a139 (adj 975): empty #17
01-08 00:47:11.917 496 496 I [mTEE] : come into the TZ_VFS_OPEN function
01-08 00:47:11.917 496 496 D [mTEE] : pathname = /vendor/thh/ta/d91f322ad5a441d5955110eda3272fc0.ta
01-08 00:47:11.917 496 496 D [mTEE] : flags = 0 mode = 0 retval = 5
01-08 00:47:11.918 496 496 D [mTEE] : come into the TZ_VFS_LSEEK function
01-08 00:47:11.919 496 496 D [mTEE] : come into the TZ_VFS_LSEEK function
01-08 00:47:11.919 1268 1296 I libprocessgroup: Successfully killed process cgroup uid 10139 pid 3874 in 0ms
However, I’m sure that I could not copy the complete terminal output into a text file, because a large amount of text is missing from the beginning of my terminal input - including the initially visible
adb logcat
--- waiting for device ---
Any good advice is welcome!
Use this command:
adb logcat > logcat.txt
To dump all the contents of logcat to a text file, actually
Thank you for doing this by the way, currently I’m trying to enable root access in this ROM as well as set up the overlay system using Trebleapp.
I’ve removed the Navbar and systemui overlay information because I believe it is causing navbar breakage but I’m not positive so it is very much trial and error at this point in regards to that, but when it comes to fatal errors I can resolve them because I’m more familiar with source code at that level than SystemUI mechanics
Dear @Null I don’t understand anything about what you are coding, but I am thrilled with what you are creating as a one-man show. Although I have already gained some experience in installing GSI’s on different systems, I don’t always succeed in everything, because on the one hand the devil is in the detail and on the other hand I am anything but a professional. You have to tell me what information you need and I try to provide it - with your additional help if necessary.
Ok, just
adb logcat > logcat.txt
to get the logcat text to be put into a text file, then upload the text file to
I will be able to find the errors in it if I have it
I also think I fixed the nav bar, hello moto’s navbar was overiding the system navbar and this issue should be fixed in the next build.
Here is a build with root enabled, I’m interested in knowing if the far right button that brings up the recent windows is broken.
I think the developers of /e/OS did a hilarious thing and exported it to the vendor.img partition
Yes, hilarious the developers of /e/OS have exported the window manager to the vendor partitions framework-res.apk LOL
Ok ok, I will do my best here
Seems the issue is this value needs to be loaded into the ROM
<string name="config_recentsComponentName">foundation.e.blisslauncher/com.android.quickstep.RecentsActivity</string>
<string name="config_secondaryHomePackage">foundation.e.blisslauncher</string>
Okkkkkkkkkk