Flow [NPEA00021]

Is cellPngDec required? I have redone it's HLE implementation, so I would be interested, if it works with the new implementation, if it didn't before.

rpcs3-v0.0.0.6-454-gf66468f

LLE:
cellFiber
cellSpursJq

Produces same output as above.
It was outputting

Code:

·! {PPU Thread[0x9f3] (snd_stream_service_thread)[0x00653e6c]} sys_tty: sys_tty_write(ch=0, buf=*0x20026df0, len=31, pwritelen=*0xd01d7340)
·W {PPU Thread[0x9f3] (snd_stream_service_thread)[0x007a6cc0]} cellFs: cellFsOpen(path=*0x21cb7820, flags=0, fd=*0xd01d7ce0, arg=*0x0, size=0x0) -> sys_fs_open()
·W {PPU Thread[0x9f3] (snd_stream_service_thread)[0x007a6cc0]} sys_fs: sys_fs_open(path=*0x21cb7820, flags=0, fd=*0xd01d7ce0, mode=0, arg=*0x0, size=0x0)
·W {PPU Thread[0x9f3] (snd_stream_service_thread)[0x007a6cc0]} sys_fs: *** path = ''
·E {PPU Thread[0x9f3] (snd_stream_service_thread)[0x007a6cc0]} sys_fs: [b]sys_fs_open('') failed: path is invalid[/b]

this in a loop (had to force shutdown rpcs3.exe process), the log was 400MB so I couldn't upload the log (it would be full of junk even if zipped), but I think it's pretty similar to the above one.

I see it's outputting some debug output to TTY. Could you check re-try the game and see if there's anything interesting in the TTY output?

Here's the TTY log (attached).
I'm no expert on this but maybe you could add some "if empty" check on the sys_fs_open() function so it won't try to open invalid (non existent) paths in a loop?

It isn't. That error is exactly the check to prevent opening empty paths.
Apparently the game really seems to want the file. Could you provide the compressed log? Probably there's a file system handling bug somewhere.