BitJag / atari_jaguar_240p_test_suite

The 240p test suite is a homebrew software suite for video game consoles developed to help in the evaluation of upscalers, upscan converters and line doublers.
2 stars 0 forks source link

Video Tests / Scroll Test / Write outside DRAM #11

Closed djipi closed 1 year ago

djipi commented 1 year ago

The write outside the console DRAM has been detected via the VirtualJaguar-Rx exceptions catches options. 2022-10-31 16_20_25-Virtual Jaguar v2 1 3 Rx

BitJag commented 1 year ago

Thanks! I'll take a look at this test to see if I can narrow down the culprit. Also, I still haven't looked into virtualjaguar rx, I should start using that if I can in Linux.

BitJag commented 1 year ago

@djipi I got virtualjaguar-rx up and running. Could you give me some clear steps to reproduce the problem using virtualjaguar-rx please? Just want to make sure I'm starting at the same point of failure. Thank you!

BitJag commented 1 year ago

@djipi So I got it crash consistently in virtualjaguar-rx. Also, there has been at least one person who has had a crash on hardware. I had it crash once on hardware, but can't get it to crash anymore. (tested on 3 different consoles with skunkboard/JagGD).

I'm assuming that there is something wrong with the code at this point, but I don't understand why virtualjaguar-rx consistently detects this problem, but elsewhere it doesn't show up consistently. Could you elaborate on what virtualjaguar-rx is doing versus just a standard virtualjaguar? I'm hoping that understanding virtualjaguar-rx is doing to detect this might help me isolate the problem with my code.

BitJag commented 1 year ago

I believe I found the problem. At least virtualjaguar-rx isn't throwing it's warnings anymore. This line: https://github.com/BitJag/atari_jaguar_240p_test_suite/blob/d9a9cb59b1bb5f2a2301963b6a6d0551f78df458/tests.c#L562

The if statement is checking the pointer rather than the specific item in the array. I'll get this corrected soon. I'll let you know when I get an updated release uploaded.

BitJag commented 1 year ago

@djipi Thanks again for spotting this. There is a new release that I believe has solved the problem. Your welcome to close this issue in the next day or two, or I'll close it after a bit.

djipi commented 1 year ago

@djipi So I got it crash consistently in virtualjaguar-rx. Also, there has been at least one person who has had a crash on hardware. I had it crash once on hardware, but can't get it to crash anymore. (tested on 3 different consoles with skunkboard/JagGD).

I'm assuming that there is something wrong with the code at this point, but I don't understand why virtualjaguar-rx consistently detects this problem, but elsewhere it doesn't show up consistently. Could you elaborate on what virtualjaguar-rx is doing versus just a standard virtualjaguar? I'm hoping that understanding virtualjaguar-rx is doing to detect this might help me isolate the problem with my code.

In the Settings's Exceptions tab, there are options to handle the exceptions catches. These options are specific to VJ-Rx. At the moment, you can turn on/off: 1) Allow writes to cartridge ROM. 2) Allow writes to unknown memory location (such as outside DRAM). 3) Allow M68000 exception catch.

djipi commented 1 year ago

@djipi Thanks again for spotting this. There is a new release that I believe has solved the problem. Your welcome to close this issue in the next day or two, or I'll close it after a bit.

You welcome and congrats for the tool. I have tested the new version and seems work pretty well. No more pop-ups after several tries with VJ-Rx. I guess we can close this issue as soon as the tests on hardware will be confirmed too.

BitJag commented 1 year ago

Thanks for the brief explanation for VJ-Rx. I'm going to spend some more time with it. The breakpoint functionality is intriguing, I have never used a tool like that before, but I have heard and seen others use similar tools on other platforms. This is what I get for learning to program on the Jaguar...

So far so good with these most recent fixes, I updated the "official" release ROM on JagCorner and have gotten reports back that all current concerns have been smoothed out on hardware. I'm going to close this for now, we can reopen if the issue ends up being more persistent.