Super Mario Bros (NES) Must Press Start with Player 2?

edmaul69

@obsidianspider i just thought of something. Can you look in your nes config folder too and make sure there isnt a super mario bros config in there either.

obsidianspider

@edmaul69 Which folder?

Also, the problem is present for the FDS version as well as the NES version. I've also confirmed the problem exists regardless of how I rename the file, so I would think that a config file is probably not the issue as I would assume it's looking for a file name?

obsidianspider

Today I updated to 4.3.13 (0ea0391) and tested again. The problem remains with fceumm, nestopia, and quicknes.

@meleu you're the master of multiple controllers. Do you have any suggestions of things I can do to try to debug/test this? If it was every game I would think it was something with my setup, but it's only Super Mario Bros. It's almost like it's something in the game itself.

pjft

Likely unrelated, but wasn't there something similar to this in Super Mario World for the SNES, or for the Super Mario All Stars?

Maybe it's something with the actual game.

Edit: maybe a similar approach helps with a game specific config file.

https://www.raspberrypi.org/forums/viewtopic.php?f=78&t=79083

obsidianspider

I'd not heard of that with the All-Stars version.

What's got me stumped is why it started recently (I obviously don't play the game every single time I boot up my Pi). Something must have changed with Raspbian or RetroPie, but I can't figure out what. :-/

pjft

@obsidianspider I had updated the post. See if that could be a potential workaround.

obsidianspider

Hmm, maybe I could tell it just to use start from player 2. I'll have to read up more on how those config files work.

meleu

@obsidianspider I only read the OP, then sorry if I'm going to suggest something already said...

First thing to do to help us to help you: launch the game with verbose mode and then paste the runcommand.log somewhere.

After that, I would check is if there's a .cfg file for this game. It would override the configs at .../nes/retroarch.cfg and .../all/retroarch.cfg.

obsidianspider

I updated RetroPie this morning to 4.3.13 (6f5ada6) and then ran with verbose logging.

This is the log when I first tried the game

This is the log after I flipped to use Player 2 to press start, exited the game, then restarted it, and was able to use Player 1 to start the game

I used diffchecker.com to look for anything different and aside from some hex values due to memory mapping, they seem the same. :-/

I don't see any game specific configs and I'm pretty sure the logs also confirm that.

obsidianspider

Updating to 4.3.15 (609bbaa) did not resolve the issue.

I wish I could figure out what is changing when I plug a controller into "Player 2" and press Start that is somehow getting things working correctly for this game. :-/

mediamogul

@obsidianspider

Have you checked to see if there's anything out of the ordinary at /opt/retropie/configs/nes/retroarch.cfg?

obsidianspider

@mediamogul Here's the contents of that file. I don't recall ever messing with it.

pi@sfc-pi:~ $ cd /opt/retropie/configs/nes/
pi@sfc-pi:/opt/retropie/configs/nes $ ls
content_history.lpl  emulators.cfg  retroarch.cfg
pi@sfc-pi:/opt/retropie/configs/nes $ cat retroarch.cfg
# Settings made here will only override settings in the global retroarch.cfg if placed above the #include line

input_remapping_directory = "/opt/retropie/configs/nes/"

#include "/opt/retropie/configs/all/retroarch.cfg"
pi@sfc-pi:/opt/retropie/configs/nes $

mediamogul

@obsidianspider

That seems to be OK. The only two settings I can think of that would be limited to NES, yet affect all the different NES cores would be at /opt/retropie/configs/nes/retroarch.cfg, or any remapping files at /opt/retropie/configs/nes/. I believe I read above that you've already eliminated the latter. Very strange indeed.

Edit: You might also try looking at /opt/retropie/configs/all/retroarch/config/remaps/. Nothing should be outputting there, but it's worth a look.

obsidianspider

@mediamogul said in Super Mario Bros (NES) Must Press Start with Player 2?:

/opt/retropie/configs/all/retroarch/config/remaps/

Nothing in there.

pi@sfc-pi:~ $ cd /opt/retropie/configs/all/retroarch/config/remaps
pi@sfc-pi:/opt/retropie/configs/all/retroarch/config/remaps $ ls
pi@sfc-pi:/opt/retropie/configs/all/retroarch/config/remaps $

dirthurts

I had this issue a while back and found that running the older version of the default emulator (sorry can't remember what the default was) worked just fine for me.

obsidianspider

@dirthurts I'm not opposed to that, but I don't know how to regress the emulator or to prevent it from upgrading when I upgrade other things.

dirthurts

@obsidianspider
right after you launch the game, simply hit any button to launch the configuration menu (you'll see the prompt). From there, simply choose a different emulator for that rom and you're good to go.

mediamogul

@dirthurts

I believe he's wanting to regress the emulator version rather than select another available option.

@obsidianspider

That shouldn't be necessary, as the current versions of the various NES cores don't exhibit this behavior for most people. However, I'm at a loss as to why it is happening to you. If you were wanting to give regression a shot and have it available for continuous use, you could make a copy of the NES script modules at ~/RetorPie-Setup/scriptmodules/libretrocores/. Using lr-fceumm as an example you would first alter the existing module ID with something unique like:

rp_module_id="lr-fceumm regression"

also, (someone correct me if I'm wrong here) changing it's selection classification to 'experimental' seems to be the trigger for disallowing a binary installation option, which would yield whatever the most current version is.

rp_module_section="exp"

From there, look for the line beginning with gitPullOrClone and add your desired version number to the end rather than the zeroes in the example below:

    gitPullOrClone "$md_build" https://github.com/libretro/libretro-fceumm.git v0.0.0

Afterward, when you're selecting the option to install from the RetroPie-Setup menu, if the option to install as binary is still present, you'd obviously need to install from source to avoid installing whatever is most current.

Again, this whole process shouldn't even be necessary, but if it is something you're looking to experiment with, this should allow that experimentation without interfering with the existing modules. Reverting back to the default versions would be no more difficult than selecting the install options from the original default script modules, which would overwrite what was done here. You could then delete your custom script modules and be back where you were.

obsidianspider

@mediamogul I'm really not wanting to regress the emulator, especially since it's only one game (well, three, SMB on NES, Famicom, and FDS), but it's happening on all three emulators, fceumm, nestopia, and quicknes. I've tried different variants of the ROM, and since versions for "different systems" also do it, leads me to believe something else is up. I am not sure if anyone else here is using the 4-player Raphnet adapter.

I'm not the only one it's happened to, ( I found a post on another forum about it in 2012 on Windows ) but I don't know what caused this to break.

I've determined that the game worked before I reimaged my SD card. Before that I was running whatever version of RetroPie was out in 2016 and had been updating every week or so. When I borked my install a few months ago I did a fresh reinstall of 4.3 from the website. I didn't test this game immediately after doing the fresh install, but the problem arose after it happened, and it worked just fine before.

mediamogul

@obsidianspider said in Super Mario Bros (NES) Must Press Start with Player 2?:

I am not sure if anyone else here is using the 4-player Raphnet adapter.

As nice as Raphnet adapters are, that's a good candidate for your issues. My personal experience is that multi-port adapters and Linux don't mesh perfectly without a dedicated driver. If I'm remembering right, there's a certain part of the USB HID "standard" that makers of multi-port adapters employ that is meant to delineate a single controller entry in multiple entries. This element is not fully implemented in Linux and causes confusion in some scenarios between the multiple controllers.