search

ramapcsx2 / gbs-control

GNU General Public License v3.0
771 stars 110 forks source link

remove "harmless" framesync frequency warnings, when debugging is turned off? #473

Open nyanpasu64 opened 1 year ago

nyanpasu64 commented 1 year ago
ramapcsx2 commented 1 year ago

The warnings represent stuff I could think of or saw happening when developing stuff at the time. Since a lot of code changed, and if those warnings don't show, it makes sense to remove and replace with anything new :p

Edit: Not a general rule, but this clockgen stuff in particular.

nyanpasu64 commented 1 year ago

The problem is that "Error: trying to tune external clock frequency while clock frequency uninitialized!" does show even in normal operation, because the existing frame sync (and my clockgen frame sync) begins running before externalClockGenSyncInOutRate() synchronizes the external clockgen. So this warning may falsely alarm viewers, and perhaps should be disabled by default. Of course ideally you'd run externalClockGenSyncInOutRate() before runFrequency(), but I don't want to change the order of execution of code where I don't understand the current system for when tasks run.

If externalClockGenSyncInOutRate() fails (due to faulty hardware etc.), then this warning will loop infinitely. Maybe this is an annoyance for end users, and removing these prints won't hurt debugging(?) since externalClockGenSyncInOutRate() itself will print an error message. But I'm not entirely sure that's a good idea.

Perhaps the error should be skipped on its first print of a "session", or suppressed until externalClockGenSyncInOutRate() fails, etc. I don't know how long a "session" is, and when to count it as the "first print" again. I think I could've found out a month or two ago, but I'm again quite rusty on the codebase and state machines after a long period of not working on the software side of things. In the meantime I could disable the print, which should be a less-bad user experience than before.

ramapcsx2 commented 1 year ago

Hm, sounds like it'd be best to actually go in and make sure the order of operations is so that the frame sync skips until the clock gen has initialized.