Handle failure on context switch to

[andr2000]

1.1. General fault vcoproc/coproc handler
1.2. Do not panic on ctx_switch_to failure
1.3. Add VCOPROC_GEN_FAULT state
1.4. Remove failed coproc from the scheduler's list
1.5. Add emulated IO handlers for the state of failure

[aanisov]

1.1. +1
1.2. but handle fault in an appropriate way
1.3. +1
1.4. ... and stop its scheduler
1.5. This should be handled by platform code