Apple's open-source XNU kernel on Android phones: its own loader and drivers, Metal on the Adreno. Runs from RAM; a reboot gives the phone back to Android.
C
Write down what is known about the unexplained stall
vcvkkbb0744a697b8
diff --git a/userland/mkjobs.py b/userland/mkjobs.pyindex 90deaa3..f73f7ef 100755--- a/userland/mkjobs.py+++ b/userland/mkjobs.py@@ -104,7 +104,14 @@ import sys# One thing still unexplained: ALLOW="syncdefaultsd appstored" loses the home screen the same way (77# processes, no widgets) with no retry storm at all -- a dozen nsurlsessiond lookups, nothing spinning. So# the stall has at least one other cause, and whoever picks this up should not assume nsurlsessiond is all of-# it.+# it. What is known about it so far, to save repeating the work:+# - it is not memory: the guest has 430 MB free and two memorystatus lines, against 260 in a good run;+# - SpringBoard is not asking for anything and failing -- it makes zero failed lookups, where a good run+# has it asking 138 times for commcenter.xpc alone, so it never reaches the stage that asks;+# - the widget stage is gated late. PosterBoard starts 266 s into a good 420 s run, so there is only about+# 150 s of margin, and a long enough delay would look exactly like a block. A 700 s run tells the two+# apart -- it was done for cloudd+mobileassetd+syncdefaultsd (84 processes, still no widgets, so that+# one is a block) but not yet for syncdefaultsd+appstored alone.# syncdefaultsd (com.apple.kvsd) is out on its own merits. It serves cloudkeychainproxy3 its store, then# aborts inside changeTokenForStoreWithConfiguration:reply: with# *** Terminating app due to uncaught exception 'NSInvalidArgumentException',