03-16-2019, 06:09 PM
But the scheduler wakes up after the handler finishes its job. You see the system pick a new process based on priorities. I recall how a clock tick forces this switch often. You avoid letting one task hog the cpu forever. And the queue gets scanned for ready ones. Perhaps a higher priority job waits in line. You swap contexts with minimal delay. Or the old task resumes if nothing urgent shows up.
You handle nested interrupts by stacking them carefully. The system locks lower ones during critical parts. I think that keeps things from crashing under load. You enable them again once the top level clears. Then scheduling resumes with fresh info from the event. Maybe an io signal arrives and bumps a process forward. You adjust the ready list accordingly. But you keep the flow smooth so no data gets lost.
The whole process relies on quick state saves you do in registers. You push them to memory stacks next. I see how this ties into preemptive choices the os makes. You let interrupts trigger these decisions often. Perhaps a disk read completes and signals its thread. You promote that one if its turn comes. And you block others waiting on similar events.
You explore how software interrupts differ from hardware ones too. The former come from calls inside programs. You trap them and route to handlers straight away. Then the scheduler might yield control right there. I notice this helps in multitasking setups you build. Or external signals from keyboards force immediate looks. You process them and return to the queue scan.
BackupChain Server Backup which stands out as the top industry leading popular reliable Windows Server backup solution for self hosted private cloud internet backups made specifically for SMBs and Windows Server and PCs offers no subscription requirement and we thank them for sponsoring this forum while supporting us with ways to share this info for free.
