search

RPCS3 / rpcs3

PS3 emulator/debugger
https://rpcs3.net/
GNU General Public License v2.0
14.7k stars 1.86k forks source link

RIDGE RACER 7 [BCES00009] - Game crashes instantly upon selecting Kamata Fiera Complete Machine and most Special Machines #13991

Open mrwallace888 opened 1 year ago

mrwallace888 commented 1 year ago

Quick summary

If I select the Complete Fiera, the game crashes immediately. But it's a bit more specific than that.

Details

When you've got the actual list of cars, where it shows all 4 categories above and below each other, and you pick the Complete Fiera there, it won't crash. It's when you use that left/right selection system (where you can also change transmission and tuning parts) that it crashes.

Even if you select the Complete Fiera first, and then move to the Jujak on the left or the Complete Abeille on the right, and then move BACK to the Complete Fiera, it crashes.

No idea why.

Same thing with most of the Special Machines, such as the Terrajin, Crinale, and Monstrous.

Attach a log file

RPCS3.log

Attach capture files for visual issues

No response

System configuration

Windows 10 Pro Ryzen 7 5700X GeForce GTX 1660 Ti 6 GB Corsair Vengeance DDR4 16 GB (Dual-Channel) Solid State Drive

Other details

The log might show multiple crashes, I went through the vehicle list to see which vehicles crash.

AniLeo commented 1 year ago

Setting mentioned on the wiki page for the game is not being used, needs retest with the setting mentioned in Ridge Racer 7

Game prints the following to sys_tty on the crash: ·! 0:03:05.535287 {PPU[0x1000000] Thread (main_thread) [0x004025cc]} sys_tty: sys_tty_write(): “abort() is called from 0x”

mrwallace888 commented 1 year ago

Nah, it's still doing it with that on.

mrwallace888 commented 1 year ago

What else?

SharkleBunwich commented 1 year ago

Setting mentioned on the wiki page for the game is not being used, needs retest with the setting mentioned in Ridge Racer 7

Game prints the following to sys_tty on the crash: ·! 0:03:05.535287 {PPU[0x1000000] Thread (main_thread) [0x004025cc]} sys_tty: sys_tty_write(): “abort() is called from 0x”

Hey, I actually made a profile specifically to comment on this thread. I'm having the same problem as OP. However, I can report the game worked properly fine before but I can't exactly say what version of RPCS3 broke the game (if that's even the case).

For starters I do have 'Accurate RSX reservation access' ON

All I can share is that upon trying to select the car via the left/right selection system (as OP described) it CTDs only when picking one of the more endgame cars with this one error: E SYS: Stat: 'sys_cond_destroy' failed with 0x80010005 : CELL_ESRCH

Full Log: RPCS3 Logs.zip

I'm not quick to say its RPCS3's fault because this is a backed up copy when I needed to reformat my PC. I've also changed several parts (motherboard, GPU, and CPU) during that time. If that matters, here's the difference between the 2 builds:

Old Rig (when RR7 worked):

New Rig (RR7 crashing):

mrwallace888 commented 1 year ago

I swear I also remember it working before without issues as well. I did upgrade my CPU in December but the game still worked after the fact.

Do you remember how long ago it was that you remember it last working?

SharkleBunwich commented 1 year ago

Last time I played it was March 2022. That's all I remember. RPCS3 has had many updates since then. I changed my CPU and GPU late 2022 and my mobo in Feb 2023.

Level99OCR commented 1 year ago

Chiming in to say that I am also getting the exact same error and also recall it working previously without issue. I hadn't tried RR7 in over a year so unfortunately I can't pin down the time of the regression.

Game: Ridge Racer 7 3D License Ver (NPUB30457) BIOS: 4.90 RPCS3 Version: 0.0.28-15306-d41fe80b Alpha Specific configuration settings: Accurate RSX Reservation is set to "on" E SYS: Stat: 'sys_cond_destroy' failed with 0x80010005 : CELL_ESRCH

Specs: Ryzen 3700X, 32gb DDR4, GTX 3070

Edit: I made a separate RPCS3 install and tested quite a few all the way back to v0.0.15-11838-8e4451d1 using the same savegame and I couldn't find a single version that didn't crash on getting to those specific cars. Just speculating but there's probably not a high chance that its something to do with the savegame? Someone with more time than myself might be able to check that by using a really old version and starting from scratch, but the game became incredible unstable below version 0.0.16. But then again, what's the likelihood that this has been marked playable for so long and someone just didn't try to select one of those specific cars and we're all misremembering.

Also can confirm this crash happens with both Vulkan and OpenGL renderers.

AniLeo commented 11 months ago

If it's a regression we need to know what build introduced, so we can pinpoint the change at fault

SharkleBunwich commented 11 months ago

The only thing I can advise is that the March 2022 build worked fine.

rjasi commented 11 months ago

Setting mentioned on the wiki page for the game is not being used, needs retest with the setting mentioned in Ridge Racer 7 Game prints the following to sys_tty on the crash: ·! 0:03:05.535287 {PPU[0x1000000] Thread (main_thread) [0x004025cc]} sys_tty: sys_tty_write(): “abort() is called from 0x”

Hey, I actually made a profile specifically to comment on this thread. I'm having the same problem as OP. However, I can report the game worked properly fine before but I can't exactly say what version of RPCS3 broke the game (if that's even the case).

For starters I do have 'Accurate RSX reservation access' ON

All I can share is that upon trying to select the car via the left/right selection system (as OP described) it CTDs only when picking one of the more endgame cars with this one error: E SYS: Stat: 'sys_cond_destroy' failed with 0x80010005 : CELL_ESRCH

Full Log: RPCS3 Logs.zip

I'm not quick to say its RPCS3's fault because this is a backed up copy when I needed to reformat my PC. I've also changed several parts (motherboard, GPU, and CPU) during that time. If that matters, here's the difference between the 2 builds:

Old Rig (when RR7 worked):

* Ryzen 5 2600

* MSI RTX 2060

* 32GB DDR4

* MSI X470 Gaming Plus mobo

New Rig (RR7 crashing):

* Ryzen 7 5800X (OC)

* EVGA RTX 3080

* Same RAM

* MSI B550 Tomahawk mobo

Facing the exact same problem. Viewing the Soldat crinale from the machine connector gives the exact same crash and error. Also have the same PC Specs.

ukyo200000 commented 7 months ago

Facing the exact same problem, but it seems like a cpu clock rate issue? Here is my solution: underclock the cpu when I play it on my 9400F computer, it works perfectly, but when I play it on my 9700KF one, it crashs like others. So I try to underclock the 9700KF to 3.7GHz, it does not crash!

SharkleBunwich commented 7 months ago

That would make sense, given my R5 2600 wasn't overlooked but my 5800X is. However I don't necessarily feel like underclocking my CPU just for an game I can play on my PS3 instead.

On Sat, Dec 2, 2023, 8:29 PM ukyo200000 @.***> wrote:

Facing the exact same problem, but it seems like a cpu clock rate issue? Here is my solution: underclock the cpu when I play it on my 9400F computer, it works perfectly, but when I play it on my 9700KF one, it crashs like others. So I try to underclock the 9700KF to 3.7GHz, it does not crash!

— Reply to this email directly, view it on GitHub https://github.com/RPCS3/rpcs3/issues/13991#issuecomment-1837302311, or unsubscribe https://github.com/notifications/unsubscribe-auth/BA2DVDGXGEVUFWZ7ZFLVMRLYHPIV7AVCNFSM6AAAAAAZBCL4T2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMZXGMYDEMZRGE . You are receiving this because you commented.Message ID: @.***>

mrwallace888 commented 7 months ago

My CPU is not overclocked, just stock.

DanteDraconi commented 7 months ago

Previewing any special/complete car on machine connector (main menu) doesn't crash for me but trying to change their livery does crash with the same errors. As a workaround I can preview one of those cars on the machine connector menu, back out, and it will be preselected to be used on other modes. Sometimes this still crashes the game but works fine again on a fresh boot.

Ryzen 7 7800X3D / RX 7800 XT / 32GB DDR5

Manwithpants812 commented 6 months ago

Also on Ryzen 5600G but using an EVGA RTX 2060. I think the common thread between all of us might be either an AMD CPU or an RTX GPU. I'm on latest emulator version as of today and latest NVidia driver. Accurate RSX Reservation on. Crashes on both Vulkan and OpenGL. Both with log error "E SYS: Stat: 'sys_cond_destroy' failed with 0x80010005 : CELL_ESRCH [x4]". My CPU is not overclocked.

AniLeo commented 6 months ago

Checking back on this one, the Need Information tag is still relevant, we need to know what build introduced the regression, if someone with the issue can investigate

SharkleBunwich commented 6 months ago

All I'm able to provide is that it worked on the March 2022 build without issue. I don't remember the version number. Hope this helps.

On Mon, Jan 1, 2024, 10:46 AM Ani @.***> wrote:

Checking back on this one, the Need Information tag is still relevant, we need to know what build introduced the regression, if someone with the issue can investigate

— Reply to this email directly, view it on GitHub https://github.com/RPCS3/rpcs3/issues/13991#issuecomment-1873376882, or unsubscribe https://github.com/notifications/unsubscribe-auth/BA2DVDGPVWQQJFP54X2H5M3YMLK4LAVCNFSM6AAAAAAZBCL4T2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZTGM3TMOBYGI . You are receiving this because you commented.Message ID: @.***>

AniLeo commented 6 months ago

Yes it helps, it means the build that regressed is between March 2022 and June 2023, so it's a start, but we will need the exact build to check the code change that broke it

DanteDraconi commented 6 months ago

For what is worth, I've tried these builds dating around March 2022 and had the same CTD on all of them

v0.0.19: 12990, 13089, 13136
v0.0.20: 13150, 13166, 13167, 13178, 13248, 13259, 13309
v0.0.21: 13327, 13329, 13331, 13332, 13333, 13344, 13355, 13362, 13369, 13378, 13389, 13452
v0.0.22: 13600
Manwithpants812 commented 5 months ago

BLJS10001GAME.zip

Since this involves having progressed through the game to get this crash, here's my save so anyone can test this.

To get the crash do the following (pretty simple): Main Menu: Go to "Ridge State Grand Prix"

on the race selection go to: "UFRA Single Event"

select: "Full Grid 02"

scroll to the right to select the later cars, crash.

Manwithpants812 commented 5 months ago

I suppose this is new info, the crash doesn't happen when booting the game with PPU Decoder: "Interpreter (static)" instead of "Recompiler (LLVM)" (at least while scrolling left-right to select the car on the steps I described above in my last comment). The only thing is that when trying to stop the emulation, the emulator exhibits new behavior and hangs and crashes, which I suppose is not a problem if the game is saved?

I'm on the latest build 0.0.30-15966.

This is the car that was making the emulator crash (don't know if there are more cars that make it crash): https://i.imgur.com/nvVNpwq.png

I still can't unlock this car and play with it, as I'm early-ish in the game, so I don't know if there's another crash that happens if I try racing with the car.

Edit: I tried old builds (rpcs3-v0.0.17-12522 [2021-07-24] and rpcs3-v0.0.12-11082 [2020-10-25]) just for the sake of it and this crash doesn't happen either, as long as anything else is chosen that is not Recompiler (LLVM)".

Edit 2: I guess a problem with playing while using "Interpreter (static)" is that the fps slows down (max 30fps for me) when you get to the race, but that's unrelated to this bug, not sure if it's a "my hardware is not good enough" thing or what. I can get the fps to go higher if I change the Vblank Frequency to something like 480Hz and I can reach 48fps, but the menus are super fast. Not sure what the deal is there.

ronjouch commented 3 months ago

Hi. I tested 20 builds, and for me on my machine this crash is not a regression. TL;DR: menu crash reproducible since game became well-supported enough to reach menu. Here are my results (top = recent, bottom = old):

Emphasis Behavior Build
. crashes on selecting special car v0.0.31-16276- e90b3993 _win64
. crashes on selecting special car v0.0.27-14944- 7bec3b30 _win64
. crashes on selecting special car v0.0.24-14175- 238298f6 _win64
1 crashes on selecting special car v0.0.21-13389- ee98f577 _win64 2022-03-31
1 crashes on selecting special car v0.0.21-13350- 04df3928 _win64 2022-03-16
1 crashes on selecting special car v0.0.21-13327- 6c096b72 _win64 2022-03-01
. crashes on selecting special car v0.0.19-12998- 50bf528f _win64
. crashes on selecting special car v0.0.17-12631- 53bfc6fa _win64
. crashes on selecting special car v0.0.17-12500- 6f901636 _win64
. crashes on selecting special car v0.0.17-12481- 1f6ca258 _win64
. crashes on selecting special car v0.0.17-12478- 85b33e9c _win64
2 crashes on selecting special car v0.0.17-12476- 66368deb _win64
2 stuck compiling shaders v0.0.17-12475- aaac4c1b _win64
. stuck compiling shaders v0.0.17-12447- a34201eb _win64
. stuck loading ppu modules v0.0.16-12397- c3415bcf _win64
. stuck loading ppu modules v0.0.16-12281- 1dc00cce _win64
. crashes on start v0.0.15-11936- 923ba6f3 _win64

Emphasis mine at two chunks of this table:

  1. Builds [ ee98f577 , 04df3928 , 6c096b72 ] from late / mid / start March 2022 demonstrate that March 2022 builds did crash, just like before or after March 2022. This confirms @DanteDraconi’s report that the report of a regression shortly after March 2022 might be bogus.
  2. Builds [ aaac4c1b → 66368deb ] demonstrate this bug is not a regression: 66368deb is for me the first build reaching the menus, and already in this first version, the crash does happen.

Finally, note that my bisect increments/jumps at the top of the table are large (~ 200 commits). So it’s not impossible that some commit at some point between [ 0.0.17, 0.0.31 ] did temporarily address the issue, only to be shortly re-broken (in a way that would be missed with my rough bisect). However, excluding this possibility, these results contradict the hypothesis of a regression.

The crash

Almost identical to OP’s RPCS3.log:

abort() is called from 0x0000000000405068
                  from 0x00000000003b92d0
                  from 0x00000000003b9280
                  from 0x00000000002c6f6c
                  from 0x00000000000c4e7c
                  from 0x00000000000be240
                  from 0x00000000000c4760
                  from 0x0000000000010efc
                  from 0x00000000000107d4
                  from 0x0000000000010398
                  from 0x000000000001024c

Config

ahmadoji01 commented 1 month ago

Can confirm changing to Interpreter (static) fixes the issue too. Playing it on Ryzen 9 7940HS with RTX 4070 here. Minor stuttering only on one track, the rest is fine.

Edit: Yeah, the Rave Downtown Circuit causes major stuttering on some parts of the track. I wish on a future update, we could stably play using LLVM PPU Decoder

Sierra935 commented 1 month ago

is there any case that this happens on the BLJS version or just BCES and BLUS? i figured it could be a solution but im unsure.

Sierra935 commented 1 month ago

nvm just tested this still the same. hopefully this can be fixed at some point. it mostly affects Category 2 Machines.