Mai-Hime Senretsu (舞-HiME 鮮烈！真 風華学園激闘史！！) - ULJS00062

[serio]

r3101
game freezes after the opening video.

it worked before (and was in fact pretty playable), i managed to track the moment it stopped working to sometime between 2800 and 2810. can't get anymore specific, since the buildbot doesn't host all of the old versions anymore, just every 10th one.

i ran both versions with default settings with everything set to off to find the culprit.

attached logs for both versions, though not sure how helpful that'll be, since there's no error message or anything. the game just stops after the video's over/skipped. logs are on info, since anything above that freezes jpcsp for me. if there are any specific debug log codes i can use to check in both versions, i can do that.

edit: dunno how useful that'll be, in 3101 i managed to enable debug mode while the ending video was fading out and let it run for a while. the game was going at around 1 frame per 30 seconds or so, i waited until the video and everything was gone off screen, and it seemed to get stuck.

[gid15]

Could add the following lines to LogSettings.xml

Code:

<logger name='hle.sceMpeg'> <level value='debug' /> </logger>
<logger name='hle.scePsmfPlayer'> <level value='debug' /> </logger>
<logger name='hle.scePsmf'> <level value='debug' /> </logger>
<logger name='compiler'> <level value='debug' /> </logger>

and run with 2810 and 2800?
Almost all the changes between these 2 revisions were affecting these modules...

[serio]

i had to split the logging into compiler only and the other 3, otherwise it choked up.

one of the non-compiler ones is so long because i couldn't skip the video for a while because of the lag i got from all the logging at once.

_c is compiler logs.

[nash67]

here is a jpcsp trace log

[gid15]

2800:

Code:

sceMpegAvcDecodeFlush mpeg=0x9AD21C4
sceMpegRingbufferAvailableSize ringbufferAddr=0x09AD2350
sceMpegRingbufferAvailableSize returning 0
sceMpegFlushAllStream mpeg=0x9AD21C4
sceMpegUnRegistStream mpeg=0x9AD21C4, streamUid=0x4D
sceMpegUnRegistStream mpeg=0x9AD21C4, streamUid=0x48
sceMpegFreeAvcEsBuf mpeg=0x9AD21C4, esBuf=0x1
sceMpegDelete mpeg=0x9AD21C4
sceMpegRingbufferDestruct ringbufferAddr=0x09AD2350

2810:

Code:

sceMpegAvcDecodeFlush mpeg=0x9AD21C4
sceMpegRingbufferAvailableSize ringbufferAddr=0x09AD2350
sceMpegRingbufferAvailableSize returning 352

After a call to sceMpegAvcDecodeFlush(), the new version does no longer return 0 for sceMpegRingbufferAvailableSize(). This is probably the reason for the problem. I will correct this behavior.

[gid15]

2800:

Code:

sceMpegAvcDecodeFlush mpeg=0x9AD21C4
sceMpegRingbufferAvailableSize ringbufferAddr=0x09AD2350
sceMpegRingbufferAvailableSize returning 0
sceMpegFlushAllStream mpeg=0x9AD21C4
sceMpegUnRegistStream mpeg=0x9AD21C4, streamUid=0x4D
sceMpegUnRegistStream mpeg=0x9AD21C4, streamUid=0x48
sceMpegFreeAvcEsBuf mpeg=0x9AD21C4, esBuf=0x1
sceMpegDelete mpeg=0x9AD21C4
sceMpegRingbufferDestruct ringbufferAddr=0x09AD2350

2810:

Code:

sceMpegAvcDecodeFlush mpeg=0x9AD21C4
sceMpegRingbufferAvailableSize ringbufferAddr=0x09AD2350
sceMpegRingbufferAvailableSize returning 352

After a call to sceMpegAvcDecodeFlush(), the new version does no longer return 0 for sceMpegRingbufferAvailableSize(). This is probably the reason for the problem. I will correct this behavior.

I've tried a fix in r3105. Could you retest?

Thank you!

[serio]

works now. good job.

[nash67]

this can be moved to playable now