multithreading - I need to trace my kernel in FreeBSD, It is getting stuck at bootup because I'm changing the scheduler, How do I do that? -
i modifying scheduler , it's causing stuck (because i'm not able implement properly, , i'm not picking stuff throw in queues). i'm trying see how far gets. how write stuff log, , how read log, when boot kernel.old
i throwing threads new set of queues, instead of traditional 3 runq's of 64 each. traverse , pick thread, i'm using tailq_foreach.
you doing wrong.
what should run hypervisor support debuggers, qemu or bhyve. attach gdb kernel boot , inspect state crashes.
in principle /could/ have log write stuff , retrieve after failed boot, don't see why this. simplest way of achieving printf , possibly extend size of buffer can fit more messages. kernel puts log @ same location , checks magic number on boot knows whether should start scratch or start appending. said log available in dmesg , /var/log/messages. noted earlier, that's not way it.
Comments
Post a Comment