Assassin's Creed [NPUB30451]: Save games do not load [Linux]
Quick summary
Attempts to load a save game do not work.
Details
Sometimes when a game is saved, and then an attempt is made to load that game from the main menu, RPCS3 freezes. For example, this was observed immediately after the opening scene when Altair is killed and the player is returned to the modern age. If you quit from this area and re-load, the save doesn't work. Written in the log is the following, in case it's relevant:
·F 0:00:58.591798 {PPU[0x100000a] Thread (TaskThread1) [0x00e151f4]} VM: Access violation reading location 0x0 (unmapped memory)
~~Given that #13531 had been save-related, I tested a build prior to that PR's merge to master to ensure it wasn't a regression from this, though that build also failed to load a save.~~ Actually, I can't rule out this PR yet as I was using an older build to load a save game, but not to save a game and then load it, which might have succeeded.
Attach a log file
Log: rpcs3.log.zip
Attach capture files for visual issues
N/A
System configuration
- OS: Linux Mint 21.1 / kernel 5.15
- CPU: Intel i7-10870H
- GPU: Nvidia GTX 3070 / driver 525.85
- Commits tested: cf5346c263111760752cabb94767c07c501207c4 and 53cc067c174a67556f0c0752aa93f8eca5252a5b
- Save data: BLUS30089-ACSAVE.zip
Other details
No response
I failed to replicate the issue RPCS3.log
I mentioned that it doens't always fail. However, as I said the one place it's failed for me every time is right after Altair is killed and the player returns to the modern age. If you could please get to that point, save and exit, then reload, and let me know what you find, that would be great.
Can you go to manage -> savedata inside rpcs3 then view in folder and upload the save data here?
Hey thanks, I should done that from the start, it was late, I was tired, yeah that's it. :)
I've uploaded the save data that crashes for me in the description, curious what you find. Interestingly, the save is labeled as the disc version (BLUS30089) but I don't think that's out of the ordinary.
Yea your save says to start a new game instead of continue memory block 1. Where my save says continue memory block 3. It's not an encryption or decryption issue. Isn't a ps3 vs rpcs3 issue. You're not missing any files. But it is capable of seeing your issue is there with your data.
Not sure what you mean. With that save, I have two options: either New or Continue, with "Memory Block 1" below that. I guess you're not seeing that? When I select Continue, the screen fades to white and remains there. This seems to be an rpcs3 issue unless I'm missing something.
Also @Darkhost1999, can you post the save data you used above, from Memory Block 3? I'll see if it works here and let you know (I expect that it should). Thanks.
Thanks, as I expected I was able to load your save without issue. Can you please let me know what happens when you try to continue from Memory Block 1 on my save?
Can you please let me know what happens when you try to continue from Memory Block 1 on my save?
I retried your saved data and can replicate the issue Memory block 1 waking up as Desmond. RPCS3-Log.zip I have to kill task rpcs3 after closing the emulator when the game crashes.
That's odd. From your screenshot above it's clear you're using Windows, maybe this is a Linux-specific issue. Also, what version (or commit) of rpcs3 are you testing? Thanks.
What commit of rpcs3 are you testing?
RPCS3 v0.0.27-14822-7000ef44 as indicated in the log I shared above. And my whole config is inside that log too.
Okay, and I missed that you had posted a log but thanks. Also I just tested with v0.0.27-14823 and still encounter the issue.
So given we're using the same game patch level, and basically the same rpcs3 version with different results I'm going to mark this as linux-specific as I'm pretty sure it is. If anyone else with linux can confirm this issue with the save data I posted that would really help.
Confirming Issue occurs around hacking mission memory block 4 MyData-BLUS30089-ACSAVE.zip
F {PPU[0x100000a] Thread (TaskThread1) [0x00e1547c]} VM: Access violation reading location 0x0 (unmapped memory)
@Darkhost1999 not sure if the reported issue was due to a corrupted savegame. Please can you provide an update or retest the issue?
Saves are likely being corrupted when creating a save on RPCS3 while being Desmond Miles. I don't wish to test if this is every time you're Desmond or only specific missions. Yet there hasn't been any code changes that touches those files since last test. So you're asking me to use the same tools to prove the same scenario that's already been proven without altering the tools in any way.
Long story short this issue can be "resolved" by only saving the game while you're Altair and never while Desmond but that's ridiculous as Desmond is genuine break or stopping points people intentionally use to stop playing the game as developer intended stopping points.