P96 Native issues on OS 3.2
 

I'm creating an enhancement pack for OS 3.2 and I'm using the Native chipset driver for P96. I'm getting these artifacts on the screen, they seem to appear in images. For example, the close icon in TextEdit is not displaying, the images inside amiga guide docs and the delete requesters are displayed mangled.


Anyone else experiencing this?

https://i.postimg.cc/ZRMNmp6J/1-delete.png

https://i.postimg.cc/Wzqk1z7g/2-text-edit.png
https://i.postimg.cc/Rqrn6mP8/3-help.png

What graphics card and what video resolution are you using?

I don't have that problem on WinUAE and UAE Zorro III

I am using WinUAE as well, no graphics card.
It only happens when the Native driver is present in Devs/Monitors.

P96 should not be used if you do not have a graphics card, can you attach your WinUAE config?

He’s using the P96 native driver. That should be fine.

Having the same issues.
P96 is up to date.
OS 3.2.1
Amiga 1200

This looks like an issue of BltMaskBitMapRastPort(), which performs a blit through a mask. Unfortunately, this call had a defect in Os 3.9 and below by expecting an interleaved mask for an interleaved bitmap, instead of a planar mask. P96 and Os 3.1.4 fixed this issue (probably almost simultaneously), but it could be that some application programs still get it wrong and assume that they need to work around a defect which no longer exist. I assume the Os is fine, but what - just to give an example - is creating this icon in the requester? This does not look like the native Os requester, but some patch on top of EasyRequest().

Identify this program, and remove it. Also, Multiview should hopefully be fine when displaying graphics, but that is another call to the Os team to check for proper usage of BltMaskBitMapRastPort() and the mask.

This was not a defect and it is broken now. There is no point in using interleaved bitmaps if you don't do interleaved blits. And the only way to do an interleaved blit through mask is to use an interleaved mask.

Olaf thinks otherwise, and I agree with him. The modulo of the mask should not depend on which type of bitmap you blit to. Actually, BltMaskBitMapRastPort() is not even implemented as "blit through a mask", but as double-xor trick as the blitter does not implement a second source for a mask.

The shots above are OS components. AsyncWB and Multiview showing a guide file with embedded image or icon.

These defects do not occur with RTG cards.

https://uploads.tapatalk-cdn.com/202...f23546fde0.jpg
Here is issue in Palete Preferences. Yes, this is not RTG. I installed Native driver for gaining some free chip mem. And 400k(in my case) is a lot.
Two versions ago of P96, there were no issues.
I tested it yesterday, when I completely uninstall latest version and then install an older one. (I think it was 2.95?)

Thanks, that reproduced the issue. It is BltMaskBitMap() passing in the wrong modulo for the mask bitmap to the low-level function. I've fixed that for the 3.2.2 release. If you want a beta, please file the bug at iComp.

Great. Thank you for quick response.
I will wait for 3.2.2 then.

Yes. 3.2.2 is much much better. But still there is an issue in AmigaGuide. Boingball in reference manual for example(look in the first post). If I open that iff in MultiView, it opens just fine.
For now, I found this glitch only in AmigaGuide.

I just tried 'Native chipset driver for P96' by ThoR on aminet last night with latest P96 from Individual Computers, A1200-030@50 and OS3.2.1 and I'm getting constant boot looping ...had to remove the Native driver in Devs:Monitors to get it to boot normally again


..I noticed the Native file didn't have an icon...not sure if that's the issue?

http://aminet.net/package/driver/moni/Native
There is no icon for native driver on aminet.

I also have had boot looping problem after updating to OS3.2.1. Didn’t know that native is the reason. I thought the new s-s is causing it.
But after using maprom function with blizzkick, boot is ok.

Normally that shouldn't be a problem but you can copy an icon from an other monitor file to the missing one and see if that makes a different.

I'm not quite understanding what the native driver really does? What's the reason for using P96 if you're not using a RTG card?

With the P96 driver you can load image data into FastMem instead of ChipMem and you can use the CPU for processing the data instead of calling blitter functions. That's faster in most cases and can save ChipMem.

just tried this and unfortunately no difference...only way to get my A1200 to boot up is to remove the Native monitor driver file from DEVS:monitors


@ deluxe 1260


interesting that solved the issue but I don't have maprom option on my M-Tec-Viper 030 card at least I don't think so?