Porting PewPew

[Neelix]

I voted a 5 for this game. What controls does it use? I'm not gonna bother downloading if it uses the touchscreen, and it looks like it might from the previews...

Navigating the menu interface does require the touchscreen.


Gameplay in the games I've played so far is exclusively on the nubs and it plays like Super Geometry Dust.


- Neelix

 

[sebt3]

I voted a 5 for this game. What controls does it use? I'm not gonna bother downloading if it uses the touchscreen, and it looks like it might from the previews...

Like SGD, double nubs but you can also use d-pad, action buttons and the screen if you prefer (althrough, nubs are the best here)

 

[ZXDunny]

I voted a 5 for this game. What controls does it use? I'm not gonna bother downloading if it uses the touchscreen, and it looks like it might from the previews...

I know Seb was a little upset about getting rated 1 star, but you've just rated a game you haven't played??

Yes, of course I did - there's nothing to stop me! And besides, Sebt3 has been a very good friend to me and I don't like to see him upset.


D.

 

[fusion_power]

Hm, so no chance to play this game on Hotfix 6? And at standard clocking? ^^"

 

[mcobit]

Not before the sgx problems are worked around I guess.

 

[sebt3]

ok I managed to reproduce the lockups.


Here is my dmesg.

[280513.334869] pvr: pewpew2: IOCTL 69 failed (13)

[280513.370025] SGX Hardware Recovery triggered (from SGX_MISRHandler)

[280513.370178] PID = 16070, process name = pewpew2

[280513.370178] EVENT_STATUS =	 0x20000000

[280513.370208] EVENT_STATUS2 =    0x00000000

[280513.370208] BIF_CTRL =		 0x00000000

[280513.370208] BIF_INT_STAT =	 0x00000000

[280513.370239] BIF_MEM_REQ_STAT = 0x00000000

[280513.370239] BIF_FAULT  =	   0x00000000

[280513.370239] CLKGATECTL =	   0x00212120

[280513.370269] HW Recovery dump generated at 1340456384568761

[280513.670471] pvr: pewpew2: IOCTL 69 failed (13)

[280513.694671] SGX Hardware Recovery triggered (from SGX_MISRHandler)

[280513.694824] PID = 16070, process name = pewpew2

[280513.694854] EVENT_STATUS =	 0x20000000

[280513.694854] EVENT_STATUS2 =    0x00000000

[280513.694854] BIF_CTRL =		 0x00000000

[280513.694885] BIF_INT_STAT =	 0x00000000

[280513.694885] BIF_MEM_REQ_STAT = 0x00000000

[280513.694885] BIF_FAULT  =	   0x00000000

[280513.694885] CLKGATECTL =	   0x00212120

[280513.694915] HW Recovery dump generated at 1340456384893407

[280528.983276] pvr: pewpew2: IOCTL 69 failed (13)

[280529.018035] SGX Hardware Recovery triggered (from SGX_MISRHandler)

[280529.018188] PID = 16070, process name = pewpew2

[280529.018218] EVENT_STATUS =	 0x20000000

[280529.018218] EVENT_STATUS2 =    0x00000000

[280529.018218] BIF_CTRL =		 0x00000000

[280529.018249] BIF_INT_STAT =	 0x00000000

[280529.018249] BIF_MEM_REQ_STAT = 0x00000000

[280529.018249] BIF_FAULT  =	   0x00000000

[280529.018249] CLKGATECTL =	   0x00212120

[280529.018280] HW Recovery dump generated at 1340456400216771

 

[mcobit]

http://test.maemo.or...bug.cgi?id=9150


seems like the nokia people had similar problems, no word, if they were solved though.


Is it possible, that those are triggered by a combination of our inputs and a sgx driverbug, that is not addressed with the nokia drivers, as they only worked around their device specific stuff?


Edit: Never experienced this kind of stuff in android. Could it be, that those workarounds, that notaz mentioned in the androidthread could be applied to our driver or other firmware?

 

[notaz]

http://test.maemo.or...bug.cgi?id=9150


seems like the nokia people had similar problems, no word, if they were solved though.


Is it possible, that those are triggered by a combination of our inputs and a sgx driverbug, that is not addressed with the nokia drivers, as they only worked around their device specific stuff?

Could be.

Edit: Never experienced this kind of stuff in android. Could it be, that those workarounds, that notaz mentioned in the androidthread could be applied to our driver or other firmware?

No, the problem with android was it asking the driver "do you support some visual X" and SGX driver responding "no" even when it does and Android switching to software rendering for some operations as a result. The workaround was just to hack android to not ask that question and make it assume SGX supports everything.

 

[_wb_]

Good job, sebt3!


One small remark: it slightly distracts me to see the touch-screen "nubs" moving while I'm playing; since nobody would play this game with the stylus anyway if you have actual nubs instead, perhaps it is better to just remove the touch controls altogether. My guess is that it shouldn't be too difficult to just not draw them - and perhaps that will even help a tiny bit to make it run smoother on lower cpu freqs.

 

[sebt3]

Ok a new release trying to work around lockups.


I also disabled the onscreen joystick.


if you get a lockup, go into setting and disable : "Play again" screen, Auto-pause. and retry. If you still get one, report again i'll try to see what else

 

[mcobit]

As we are in the dev section: May I ask what you suspect to cause the lockups?

 

[sebt3]

As we are in the dev section: May I ask what you suspect to cause the lockups?

What I fixed was a lockup from the game code (something I introduce).


I've yet to reproduce the lockup Neelix shown in his screenies.

 

[mcobit]

I have the lockups with the new version, too. Also the onscreen joysticks are still visible.

 

[sebt3]

I have the lockups with the new version, too

hum, so until I manage to reproduce, I cant help

 

[mcobit]

Are you sure, that you uploaded a new version?


I don't see any difference.

 

[sebt3]

Are you sure, that you uploaded a new version?


I don't see any difference.

The arraw on the top here shouldnt be there anymore :

else you havent the last one

 

[Neelix]

Well I've yet to have another lockup, but it started after playing for a while last time, so I may have to sacrifice myself for sake of troubleshooting and keep playing


Also I just rebooted which may be a factor. Before I rebooted the SGX driver had been disabled entirely after it crashed when I tried running the old PNDBuilder PND and recovery failed, so I had to reboot just to get pewpew2 started this time.


I'm beginning to think the lockups are a bug in the SGX driver itself and pewpew2 was just working it hard enough and long enough to trigger it.


No arrow symbol appearing at the top anymore but the on screen controls are still showing up as a distraction.


Online features still not working though I wasn't expecting that to be fixed in this build.


Some definite slow downs when there is a lot of debris on the screen from multiple enemies but not enough to make it any less fun. (I run at 800 Mhz)


- Neelix

 

[mcobit]

Slowdowns seem to be quite random for me. Not only, when much stuff is on the screen, but as said before by sebt3, maybe slowdowns are triggered by the sgx recoveries.

 

[Neelix]

I just quit pewpew2 and tried PNDBuilder again just to see if it would work, which it did.


Out of interest, I checked dmesg afterwards to see if there were any messages generated and I noticed this:


[ 4674.546386] pvr: pewpew2: cleaning up 2147 unfreed resources


Could that mean there's a resource leak, or is that just normal?


- Neelix

 

[mcobit]

A quick google turns up this:


https://bugs.maemo.org/show_bug.cgi?id=9320#c14


May be helpful.