check compatibility for `2026062100` build

[pixincreate]

https://grapheneos.org/releases#2026062100

this release might break zygisk. have to check.

[ApolloDawnstar]

Wow, they put this one out ridiculously fast. I appreciate you for keeping up with this Magisk fork! I'd be pretty much screwed if you weren't doing this for so long, thanks 🙂

[pixincreate]

topjohnwu#9888

[pixincreate]

no, ideally, this shouldn't break anything. but i'll keep this open for now.

[adcape]

Build 20260623 bootloops, but I'm not sure whether it is caused by Magisk or something else.

[pixincreate]

Build 20260623 bootloops, but I'm not sure whether it is caused by Magisk or something else.

@adcape, without logs, i can't help. did you try disabling modules? did you try booting in safe mode? without any information, i can't help.

[adcape]

@pixincreate Currently I don't use modules. No, I didn't try booting in safe mode. I'll try to get more info for troubleshooting when I'll be updating the OS next time some time next week.

[pixincreate]

consider creating a separate issue for this. i'll be locking this.

[pixincreate]

https://grapheneos.org/releases#2026062800 has zygote related changes too.

^^ ps, i just saw the term zygote being mentioned in the release and copy pasted the link. yet to go through the changes.

[pixincreate]

https://github.com/pixincreate/Magisk/actions/runs/28891775617/artifacts/8150039826

not sure if this fixes magisk not working in android 17.

https://github.com/pixincreate/Magisk/tree/issue-30-zygisk-android17-sepolicy

[pixincreate]

zygote jni signatures against grapheneos 17 and they appear to match the grapheneos hooks already present in this fork. the only relevant change i found for android 17 is the memfd selinux policy update, which is what the above mentioned test build includes.

[pixincreate]

https://github.com/GrapheneOS/platform_frameworks_base/blob/d881c58713b9a8af651915f59c3a469043ffd245/core/java/com/android/internal/os/Zygote.java#L374

current grapheneos 17:

nativeForkAndSpecialize:
([JII[II[[IILjava/lang/String;...ZZZ)I

nativeSpecializeAppProcess:
([JII[II[[IILjava/lang/String;...ZZZ)V

current hooks:

nativeForkAndSpecialize_grapheneos_b:
(II[II[[IILjava/lang/String;...ZZZ[J)I

nativeSpecializeAppProcess_grapheneos_u:
(II[II[[IILjava/lang/String;...ZZZ[J)V

add new variants for current grapheneos 17:

nativeForkAndSpecialize:
([JII[II[[IILjava/lang/String;Ljava/lang/String;[I[IZLjava/lang/String;Ljava/lang/String;ZZ[Ljava/lang/String;[Ljava/lang/String;ZZZ)I

same as grapheneos_b, but move jlongArray from the end to the first parameter.

nativeSpecializeAppProcess:
([JII[II[[IILjava/lang/String;Ljava/lang/String;ZLjava/lang/String;Ljava/lang/String;Z[Ljava/lang/String;[Ljava/lang/String;ZZZ)V

same as spec_grapheneos_u, but move jlongArray from the end to the first parameter.

no new nativeForkSystemServer variant needed; grapheneos still matches aosp there: (II[II[[IJJ)I

[pixincreate]

ec478f6 should fix it.

artifact will be generated here. would appreciate if anyone running magisk on grapheneos could test this.