Need some testers for ct60_vmalloc()

Discuss CT60/CT63, CTPCI, SuperVidel and EtherNAT hardware here.

Moderators: Mug UK, moondog/.tSCc., [ProToS], lp, Moderator Team

User avatar
Rajah Lone
Captain Atari
Captain Atari
Posts: 371
Joined: Wed Aug 07, 2002 12:27 pm
Location: Lyon / France
Contact:

Re: Need some testers for ct60_vmalloc()

Postby Rajah Lone » Fri Jan 29, 2016 8:26 pm

@calimero: thanks you very much for the tests.

SuperVidel is detected and vmalloc() successful. The graphics should be copied to VRAM. One only vmalloc() is done for all bitmaps and inside it, I use pointers for the different sprites. Doing maths for TTRAM to VRAM relocation may be difficult since I have to deal with negatives pointers. I have an idea and will do some tests on my setup.

DGX is a custom format, created for my Dungeon Master. There were too many sprites and images, that could make the hard disk a little mad with the FAT. Consider a DGX file as a folder of DGI files (Dungeon GEM Image). Ususally I grouped all images of same object or monster.
DGI file is a plateform-independant data format for raster (mdfb + each bitplane separate). Use vr_trnfm() to transform to a raster good for the VDI. I coded a converter (2VDI, or makedgx) to create them with .gif.

User avatar
lp
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2230
Joined: Wed Nov 12, 2003 11:09 pm
Location: GFA Headquarters
Contact:

Re: Need some testers for ct60_vmalloc()

Postby lp » Fri Jan 29, 2016 9:04 pm

I added Radeon support to the VNC client. Its at my public dropbox in the \avnc folder: (avnc_svr.prg)
https://www.dropbox.com/sh/hk65mapqd3fm ... KqQTa?dl=0

User avatar
wongck
Ultimate Atarian
Ultimate Atarian
Posts: 11926
Joined: Sat May 03, 2008 2:09 pm
Location: Far East
Contact:

Re: Need some testers for ct60_vmalloc()

Postby wongck » Sat Jan 30, 2016 12:34 am

lp wrote:I added Radeon support to the VNC client. Its at my public dropbox in the \avnc folder: (avnc_svr.prg)
https://www.dropbox.com/sh/hk65mapqd3fm ... KqQTa?dl=0


Oooooo......
The binary named avnc_svr.prg also supports the Radeon driver. That would be
CT60 + Radeon or FireBee + Radeon.

FireBee + Radeon...... you have that one??? How ???
My Stuff: FB/Falcon CT63+CTPCI_ATI_RTL8139 14+512MB 30GB HDD CF HxC_SD/ TT030 68882 4+32MB 520MB Nova/ 520STFM 4MB Tos206 SCSI
Shared SCSI Bus:ScsiLink ethernet, 9GB HDD,SD-reader @ http://phsw.atari.org
My Atari stuff for sale - click here for list

User avatar
Rajah Lone
Captain Atari
Captain Atari
Posts: 371
Joined: Wed Aug 07, 2002 12:27 pm
Location: Lyon / France
Contact:

Re: Need some testers for ct60_vmalloc()

Postby Rajah Lone » Sat Jan 30, 2016 12:59 am

No we don't. So we need kind and patient users for tests.

Here new updates: http://ptonthat.fr/files/temp/
I verified the maths for pointers changed after TTRAM to VRAM relocation, but seems OK (with a simulation TTRAM to TTRAM). Should work with GFA negative pointers%.
Changes:
- custom block move (TTRAM to VRAM), instead of BMOVE. Lonny verified the GFA internal ASM code of BMOVE, which seemed OK. So I don't think it will resolve the issue on SuperVidel.
- Offscreen VDI handle is not proposed when VRAM usage is available.

User avatar
calimero
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2032
Joined: Thu Sep 15, 2005 10:01 am
Location: STara Pazova, Serbia
Contact:

Re: Need some testers for ct60_vmalloc()

Postby calimero » Sat Jan 30, 2016 11:13 am

Here is new log.txt
VRAMTEST.txt


and here is screenshot of latest 2048 from my supervidel:

IMG_2343.JPG


I also try aVNC but I get: "Error at fsocket (-32)" - does it work with Sting at all??
You do not have the required permissions to view the files attached to this post.
using Atari since 1986.http://wet.atari.orghttp://milan.kovac.cc/atari/software/ ・ Atari Falcon030/CT63/SV ・ Atari STe ・ Atari Mega4/MegaFile30/SM124 ・ Amiga 1200/PPC ・ Amiga 500 ・ C64 ・ ZX Spectrum ・ RPi ・ MagiC! ・ MiNT 1.18 ・ OS X

User avatar
calimero
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2032
Joined: Thu Sep 15, 2005 10:01 am
Location: STara Pazova, Serbia
Contact:

Re: Need some testers for ct60_vmalloc()

Postby calimero » Sat Jan 30, 2016 11:22 am

I just try aVNC on MiNT (XaAES 1.5.5, MiNT 1.18) but it crash when I click on Connect from menu. I am not sure how should I set TT flags and MemoryProtection flags...

IMG_2344.JPG
You do not have the required permissions to view the files attached to this post.
using Atari since 1986.http://wet.atari.orghttp://milan.kovac.cc/atari/software/ ・ Atari Falcon030/CT63/SV ・ Atari STe ・ Atari Mega4/MegaFile30/SM124 ・ Amiga 1200/PPC ・ Amiga 500 ・ C64 ・ ZX Spectrum ・ RPi ・ MagiC! ・ MiNT 1.18 ・ OS X

User avatar
shoggoth
Nature
Nature
Posts: 848
Joined: Tue Aug 01, 2006 9:21 am
Location: Halmstad, Sweden
Contact:

Re: Need some testers for ct60_vmalloc()

Postby shoggoth » Sat Jan 30, 2016 1:20 pm

calimero wrote:I just try aVNC on MiNT (XaAES 1.5.5, MiNT 1.18) but it crash when I click on Connect from menu. I am not sure how should I set TT flags and MemoryProtection flags...

IMG_2344.JPG


Perhaps you could try an older version (without support for "modern" video ram) and see if it works better. Helps iron out problems.
Ain't no space like PeP-space.

User avatar
Rajah Lone
Captain Atari
Captain Atari
Posts: 371
Joined: Wed Aug 07, 2002 12:27 pm
Location: Lyon / France
Contact:

Re: Need some testers for ct60_vmalloc()

Postby Rajah Lone » Sat Jan 30, 2016 3:29 pm

New archives at: http://ptonthat.fr/files/temp/

I think I found a problem. MDFB pointers were tested > 0 before raster copy. Reintroduced BMOVE usage instead of custom block move.

Sorry about bothering so much. To ease test pain, I added a "one more thing" as gem candy. WIP and very early stage of development. Only moves with arrowes and numeric pad. VRAM allocation option is disabled by default.

User avatar
lp
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2230
Joined: Wed Nov 12, 2003 11:09 pm
Location: GFA Headquarters
Contact:

Re: Need some testers for ct60_vmalloc()

Postby lp » Sun Jan 31, 2016 1:26 am

calimero wrote:I just try aVNC on MiNT (XaAES 1.5.5, MiNT 1.18) but it crash when I click on Connect from menu. I am not sure how should I set TT flags and MemoryProtection flags...


I would try the regular version at my website first. I don't think messing with the prg flags will help. For the ram it won't for sure, as it calls mxalloc(). For memory protection, there's an off chance global might help, but I suspect not. Although I wrote it using MiNT, I never tested it with memory protection enabled.

gstoll
Atari freak
Atari freak
Posts: 60
Joined: Sun Feb 17, 2008 4:29 pm

Re: Need some testers for ct60_vmalloc()

Postby gstoll » Sun Jan 31, 2016 1:26 pm

Rajah Lone wrote:The toshyp should need some update, Didier's documentation in tos_drivers.lzh is more accurate.

What file do you mean? video_xbios.txt?

User avatar
Rajah Lone
Captain Atari
Captain Atari
Posts: 371
Joined: Wed Aug 07, 2002 12:27 pm
Location: Lyon / France
Contact:

Re: Need some testers for ct60_vmalloc()

Postby Rajah Lone » Sun Jan 31, 2016 1:34 pm

@gstoll: yes, the video_xbios.txt seems to contain more informations. Especially in the "mode" paragraph.
Also:
- Availability also on SuperVidel.
- with Shoggoth's notes about the fact all vmalloc() must be "vmfree()" (because this is an XBIOS call, not GEMDOS).
- returned value can be above 0x7FFFFFFF, so user should use unsigned long for pointers.

'SupV' cookie is not listed. Thank you for this documentation. TOS.HYP is my bible.


I'm still waiting tests results for the 01-30 archive. If there's no news, because I want to resume my devs, I will release the official archives (2048, Teenage Queen, DGEM, X2 project) as they are, with the use_vmalloc option selectable, but disabled by default.

User avatar
calimero
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2032
Joined: Thu Sep 15, 2005 10:01 am
Location: STara Pazova, Serbia
Contact:

Re: Need some testers for ct60_vmalloc()

Postby calimero » Sun Jan 31, 2016 4:04 pm

2048 20160130 works on SuperVidel :) graphics is ok
but speed is about same level as previous build (124) with ok graphics: 8bit chunky is ultra fast, 32bit is little bit slow.

btw
xenon 20160130 works in 16 and 32bit color mode.
In 8 bit chunky color are wrong.
also, if you move game window (in any reslotion, color depth) program die with exception!

IMG_2354.JPG
You do not have the required permissions to view the files attached to this post.
using Atari since 1986.http://wet.atari.orghttp://milan.kovac.cc/atari/software/ ・ Atari Falcon030/CT63/SV ・ Atari STe ・ Atari Mega4/MegaFile30/SM124 ・ Amiga 1200/PPC ・ Amiga 500 ・ C64 ・ ZX Spectrum ・ RPi ・ MagiC! ・ MiNT 1.18 ・ OS X

User avatar
calimero
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2032
Joined: Thu Sep 15, 2005 10:01 am
Location: STara Pazova, Serbia
Contact:

Re: Need some testers for ct60_vmalloc()

Postby calimero » Sun Jan 31, 2016 4:18 pm

@shoggoth why on SuperVidel, if you chose original Falcon color mode (2, 4, 16 or 256 color mode) and display some random txt file using standard builtin TOS viewer font are completely messed-up?

IMG_2355.JPG


is this because TOS use blitter for displaying text?

---
and how I change resolution from MiNT from GUI (for now I edit xaaes config file... :/)

---
and how I can check if memory protection is active under MiNT?
(in left-shift boot menu it is turned off but I get errors from many application under mint!)
You do not have the required permissions to view the files attached to this post.
using Atari since 1986.http://wet.atari.orghttp://milan.kovac.cc/atari/software/ ・ Atari Falcon030/CT63/SV ・ Atari STe ・ Atari Mega4/MegaFile30/SM124 ・ Amiga 1200/PPC ・ Amiga 500 ・ C64 ・ ZX Spectrum ・ RPi ・ MagiC! ・ MiNT 1.18 ・ OS X

User avatar
Rajah Lone
Captain Atari
Captain Atari
Posts: 371
Joined: Wed Aug 07, 2002 12:27 pm
Location: Lyon / France
Contact:

Re: Need some testers for ct60_vmalloc()

Postby Rajah Lone » Sun Jan 31, 2016 6:07 pm

calimero wrote:2048 20160130 works on SuperVidel [...] 32bit is little bit slow.

Thanks. I put the official release from my blog. I would like to have a SuperVidel or Radeon on my FireBee to see the difference in graphics in VRAM against TTRAM TC32 screenmode.

calimero wrote:xenon 20160130 : In 8 bit chunky color are wrong.
also, if you move game window (in any reslotion, color depth) program die with exception!

Palette is not yet supported, so TC modes only for now. Thanks for the bugreport. Fixed in dev. For further discussion, better use the my GEM games thread.

User avatar
shoggoth
Nature
Nature
Posts: 848
Joined: Tue Aug 01, 2006 9:21 am
Location: Halmstad, Sweden
Contact:

Re: Need some testers for ct60_vmalloc()

Postby shoggoth » Sun Jan 31, 2016 7:56 pm

calimero wrote:@shoggoth why on SuperVidel, if you chose original Falcon color mode (2, 4, 16 or 256 color mode) and display some random txt file using standard builtin TOS viewer font are completely messed-up?

IMG_2355.JPG


is this because TOS use blitter for displaying text?


Nope. I guess you're using NVDI; I haven't implemented the console properly in bitplane modes. People generally haven't noticed this though (with some exceptions).

and how I change resolution from MiNT from GUI (for now I edit xaaes config file... :/)


You don't; there's no SuperVidel video mode requester for XaAES. It shouldn't be rocket science to make one though.
Ain't no space like PeP-space.

User avatar
calimero
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2032
Joined: Thu Sep 15, 2005 10:01 am
Location: STara Pazova, Serbia
Contact:

Re: Need some testers for ct60_vmalloc()

Postby calimero » Mon Feb 01, 2016 8:40 am

shoggoth wrote:Nope. I guess you're using NVDI; I haven't implemented the console properly in bitplane modes. People generally haven't noticed this though (with some exceptions).

neither do I :)
only recently I notice this since I was checking different color depths for Rajah... (yes, I use NVDI)

You don't; there's no SuperVidel video mode requester for XaAES. It shouldn't be rocket science to make one though.

Could you made one? :angel: btw would it work also for MyAES?
using Atari since 1986.http://wet.atari.orghttp://milan.kovac.cc/atari/software/ ・ Atari Falcon030/CT63/SV ・ Atari STe ・ Atari Mega4/MegaFile30/SM124 ・ Amiga 1200/PPC ・ Amiga 500 ・ C64 ・ ZX Spectrum ・ RPi ・ MagiC! ・ MiNT 1.18 ・ OS X

User avatar
shoggoth
Nature
Nature
Posts: 848
Joined: Tue Aug 01, 2006 9:21 am
Location: Halmstad, Sweden
Contact:

Re: Need some testers for ct60_vmalloc()

Postby shoggoth » Mon Feb 01, 2016 8:48 am

calimero wrote:
shoggoth wrote:Nope. I guess you're using NVDI; I haven't implemented the console properly in bitplane modes. People generally haven't noticed this though (with some exceptions).

neither do I :)
only recently I notice this since I was checking different color depths for Rajah... (yes, I use NVDI)


I'll fix this one day, however I was slightly exhausted after having spent years disassembling NVDI to learn the driver format.

You don't; there's no SuperVidel video mode requester for XaAES. It shouldn't be rocket science to make one though.

Could you made one? :angel: btw would it work also for MyAES?


I left that as a coding excursive for the user ;-)

Stranger things have happened though, it's not very difficult to do, I just have to find the time and energy!
Ain't no space like PeP-space.

gstoll
Atari freak
Atari freak
Posts: 60
Joined: Sun Feb 17, 2008 4:29 pm

Descriptions ct60_vmalloc()

Postby gstoll » Mon Feb 01, 2016 4:57 pm

Rajah Lone wrote:@gstoll: yes, the video_xbios.txt seems to contain more informations. Especially in the "mode" paragraph.
Also:
- Availability also on SuperVidel.

'SupV' cookie is not listed.

Is there a docu about the cookie and what function the sv_xbios installed?

Rajah Lone wrote:- with Shoggoth's notes about the fact all vmalloc() must be "vmfree()" (because this is an XBIOS call, not GEMDOS).

I think you mean that a ct60_vmalloc (1,.. ) need a ct60_vmalloc (2,..)? I can't see a function vmfree.

Rajah Lone wrote:- returned value can be above 0x7FFFFFFF, so user should use unsigned long for pointers.

In all docu about the ct60 is the return value long. And when you look to the xbios.c in the tos_driver archive the return of the function vmalloc is -1 if the value from mode is wrong.

User avatar
Rajah Lone
Captain Atari
Captain Atari
Posts: 371
Joined: Wed Aug 07, 2002 12:27 pm
Location: Lyon / France
Contact:

Re: Descriptions ct60_vmalloc()

Postby Rajah Lone » Mon Feb 01, 2016 5:40 pm

gstoll wrote:[Availability also on SuperVidel]Is there a docu about the cookie and what function the sv_xbios installed?

Couldn't find it myself. I would be happy to read it. For now, knowledge is acquired from shoggoth's posts.

gstoll wrote:I think you mean that a ct60_vmalloc (1,.. ) need a ct60_vmalloc (2,..)? I can't see a function vmfree.

Yes, with opcode 2. That's why I used quotes with the vmfree() neologism.

gstoll wrote:In all docu about the ct60 is the return value long. And when you look to the xbios.c in the tos_driver archive the return of the function vmalloc is -1 if the value from mode is wrong.

-1 (signed long variable) = 0xEFFFFFFF (unsigned long variable)

gstoll
Atari freak
Atari freak
Posts: 60
Joined: Sun Feb 17, 2008 4:29 pm

Re: Descriptions ct60_vmalloc()

Postby gstoll » Tue Feb 09, 2016 4:33 pm

Rajah Lone wrote:Couldn't find it myself. I would be happy to read it.

Too Bad, because it looks like that the driver install XBIOS(0x97) with do nothing. And the extend Setscreen will not work while the magic in rez is worng and there is an new magic 0x5301.

I will add your comments, Thank for them.

Gerhard

User avatar
shoggoth
Nature
Nature
Posts: 848
Joined: Tue Aug 01, 2006 9:21 am
Location: Halmstad, Sweden
Contact:

Re: Descriptions ct60_vmalloc()

Postby shoggoth » Tue Feb 09, 2016 4:46 pm

gstoll wrote:
Rajah Lone wrote:Couldn't find it myself. I would be happy to read it.

Too Bad, because it looks like that the driver install XBIOS(0x97) with do nothing. And the extend Setscreen will not work while the magic in rez is worng and there is an new magic 0x5301.


I'm the guilty one! I worked as a technical writer for 10 years, which means I bloody hate writing technical documentation :)

Vsetscreen() should work as defined by Atari. As soon as I find some time, I'll document it for you. It would be great to put this into toshyp (which, by the way, is great). Some of these extensions (e.g. the modecode word) is reflected in e.g. VsetMode() and VcheckMode() too.

ct60_vmalloc() should work as it does on the CTPCI, otherwise I've implemented it wrong.

The SupV cookie is used to determine the presence of these extensions; I can't guarantee that any methods of detection will work in the future.
Ain't no space like PeP-space.

gstoll
Atari freak
Atari freak
Posts: 60
Joined: Sun Feb 17, 2008 4:29 pm

Re: Descriptions ct60_vmalloc()

Postby gstoll » Sat Feb 13, 2016 5:13 pm

shoggoth wrote:I'm the guilty one! I worked as a technical writer for 10 years, which means I bloody hate writing technical documentation :)

This is a problem which have all programmer ;-)

shoggoth wrote:Vsetscreen() should work as defined by Atari.

What I mean is the extend mode from ct60. In your driver is it 0x4E56 insted of 0x564E.

I add the cookie to the tos.hyp and make a update in ct60_vmalloc.

User avatar
shoggoth
Nature
Nature
Posts: 848
Joined: Tue Aug 01, 2006 9:21 am
Location: Halmstad, Sweden
Contact:

Re: Descriptions ct60_vmalloc()

Postby shoggoth » Sat Feb 13, 2016 5:44 pm

gstoll wrote:
shoggoth wrote:I'm the guilty one! I worked as a technical writer for 10 years, which means I bloody hate writing technical documentation :)

This is a problem which have all programmer ;-)

shoggoth wrote:Vsetscreen() should work as defined by Atari.

What I mean is the extend mode from ct60. In your driver is it 0x4E56 insted of 0x564E.

I add the cookie to the tos.hyp and make a update in ct60_vmalloc.


I see. I've defined it as:

Code: Select all

#define MI_MAGIC            (('M' << 8) | 'I')
#define NV_MAGIC            (('N' << 8) | 'V')
#define SV_MAGIC(device)    (('S' << 8) | device)


... which means I've accidentally swapped them. I'll fix this in some future release (but keep compatibility with this one). Thanks for the input :)

EDIT: It should be noted that the SuperVidel is implemented to work like the Falcon video XBIOS, not like the Milan one; hence only a subset of the Milan stuff is implemented, and it's far from tested properly.
Ain't no space like PeP-space.


Social Media

     

Return to “CT60 / CT63 Area”

Who is online

Users browsing this forum: No registered users and 1 guest