You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Each Device can have an interrupt address, which you configure as the first w bits you send it.
Then, if an interrupt occurs, the device will throw you to that address.
To return from an interrupt, the address can appear on 5w, on an address configured in the device manager. It can be a command from the device manager ("output 1111 for iret"). It can be a input from the device manager (that way you can save it for later and do "context switches").
Any way is great.
I love the last way, as it allows you to both save it for later, or jump to it right now. It will take w@ ops to read the address (or w@/4 for hex If I choose that), which is 320 at best, which is pretty expensive.
The one before the last - feels a bit like cheating. Yet, the first two feels too - as the address "magically appears" there each time.
A bit of cheating is important.
A cheating can also be "configure a hex.vec w/4 address to be configured to be filled by the return value each time.
Maybe what triggers me is the 1/100 timer interrupt. Flipjump might not be built for such a short interval.
Need to think this through.
The text was updated successfully, but these errors were encountered:
Each Device can have an interrupt address, which you configure as the first w bits you send it.
Then, if an interrupt occurs, the device will throw you to that address.
To return from an interrupt, the address can appear on 5w, on an address configured in the device manager. It can be a command from the device manager ("output 1111 for iret"). It can be a input from the device manager (that way you can save it for later and do "context switches").
Any way is great.
I love the last way, as it allows you to both save it for later, or jump to it right now. It will take w@ ops to read the address (or w@/4 for hex If I choose that), which is 320 at best, which is pretty expensive.
The one before the last - feels a bit like cheating. Yet, the first two feels too - as the address "magically appears" there each time.
A bit of cheating is important.
A cheating can also be "configure a
hex.vec w/4
address to be configured to be filled by the return value each time.Maybe what triggers me is the 1/100 timer interrupt. Flipjump might not be built for such a short interval.
Need to think this through.
The text was updated successfully, but these errors were encountered: