Menu no longer re-appears if reset was held for too long (1-5 minutes)

[rianhunter]

If you press reset while the sd2snes is on the menu and hold it for a long time (1-5 minutes), it no longer reappears. I'm not 100% sure why but I think it's because the SRAM where the menu ROM is stored becomes corrupted. I haven't read the data sheet on the SRAM but perhaps it's because no SRAM accesses happen during reset.

[mrehkopf]

Spot-on! Sadly the datasheet does not seem to have definitive information on that. We specifically chose the part variant that has no maximum cycle time limitation but the chips don't seem to adhere to that very well.

Degradation seems to occur when a chip is selected for a prolonged amount of time with no address changes occurring (probably needing to cross burst page boundaries). To prevent degradation during normal use I word-interleaved the chips so that even small wait loops would not cause a single chip to be selected for too long.

This is also why I recommend turning off the boot logo animation and boot directly to cartridge when a Super Nt is used; otherwise the duration where the cartridge is held in reset before boot can be enough to cause degradation of the menu ROM after it has been loaded.

[rianhunter]

Thanks Ikari for the response. Does this issue personally motivate you to fix this or are you fine with the behavior how it is? I would like to fix this. Would you support an external contribution to fix this or would you prefer to keep the code simple? If you would support a fix, is there a specific way that you'd like this fixed? If you have no opinions, I'm happy to investigate a simple way and uninvasive fix for this on my own.

[rianhunter]

Also Happy new year!

[mrehkopf]

Thanks, Happy New Year to you, too! :)

Actually I am quite unhappy with the situation as it is ^^;
I have a rewrite of the time sharing / decoding part of the FPGA gateware in mind but so far I did not get around to doing it. So that will only happen some time in the future.

In the meantime I'd gladly accept a fix for this issue! From the top of my head I think it may be enough to gate the ROM*_CE signals using the SNES_DEAD signal so that no chip would be selected at least when held in reset. However an exception would be necessary for MCU access cycles, since the MCU needs the RAM to be accessible even when the SNES is held in reset.

[rianhunter]

Good pointers, I'll look into this. I'll need to set up an environment to compile the firmware and HDL to get started. I'll probably be able to get a potential fix out in February or March.

[rianhunter]

Just a quick update. Predictably life got in the way but this is still on my todo list in the near term. Feel free to take over this task if you have time but otherwise I still plan on getting to it!