23 December 2013, 14:30 | #501 |
move.l #$c0ff33,throat
Join Date: Dec 2005
Location: Berlin/Joymoney
Posts: 6,863
|
That sounds like an interesting bug, out of curiosity, how come this doesn't happen on a real Amiga?
|
23 December 2013, 14:55 | #502 | |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
Second bug is even trickier, D channel starts from $40A40, then D channel pointer is changed to 0 (move.l #0,$dff054) and because 68000 is 16-bit, high word is cleared first and D becomes 00000Axx temporarily and blitter uses this address for one D write (because cycle mode is -D-D-D), after the D-write low word gets cleared too. Of course this does not happen on real hardware and reason is most likely already known Agnus DMA feature, DMA fetch exactly 1 cycle after channel's DMA pointer modification gets either ignored or DMA uses old address and new address become valid after DMA fetch. (EDIT: also high and low address pointers seem to work slightly differently, this needs confirmation first) Bitplane and sprite pointers have similar behavior and there are demos that only work correctly if write is ignored.. This also explains why there is no same channel DMA fetches back to back, Agnus probably has some kind of 2 stage pipeline for DMA accesses. EDIT: Confirmed. Write goes nowhere, both low and high address writes are totally ignored by Agnus. This actually means there are two blitter side-effects that both "protects" this demo from crashing. Write to DPTL or DPTH: nothing happens, blitter keeps writing to original address, no self-destruct. Then write to BLTCON0 stops the blitter from doing D writes at all. Last edited by Toni Wilen; 23 December 2013 at 16:50. |
|
23 December 2013, 16:02 | #503 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
Continuation from:
http://eab.abime.net/showpost.php?p=...&postcount=496 http://eab.abime.net/showpost.php?p=...&postcount=497 A Small Intro (Intro) by Tww A500 OCS/ECS (most common) ... strange, as on the link picture it looks different ... pic.1 !!! Last edited by amilo3438; 23 December 2013 at 16:13. |
23 December 2013, 17:02 | #504 | |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
|
|
23 December 2013, 18:42 | #505 | ||||
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
Quote:
EDIT: It is caused by weird copper feature, special/dummy copjmp second cycle disappears if it would cross scanlines (=it would be positioned immediately after first strobe/refresh slot) Quote:
Quote:
It is demo bug, chipset register contents after power up are not be guaranteed, but I still "fix" it. Of course color is also wrong if any program, for example demo pack menu is run and it uses color 15. Last edited by Toni Wilen; 23 December 2013 at 20:48. |
||||
23 December 2013, 19:29 | #506 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
JFI: Interesting, on MSVC 2013 SSE2 version no problem -> http://eab.abime.net/showpost.php?p=...&postcount=499
Bliss-Out (Demo) by Ukonx A500 ECS 1mb ... pic.1 !!! Digital Stagnation (Demo) by United Lamers A500 OCS/ECS (most common) ... on demo part called - see pic.2 - something strange appear - see pic.3 !!! Megademo (Megademo) by United Lamers A500 OCS/ECS (most common) ... a little strange: F1 part1 ... pic.4 ?! Roxette (Demo) by United Lamers A500 OCS/ECS (most common) ... pic.5 !!! Last edited by amilo3438; 23 December 2013 at 21:56. |
24 December 2013, 00:54 | #507 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
FYI ... Hi Freaks (Intro) by Supplex & The Silents -> http://eab.abime.net/showpost.php?p=...&postcount=184
Just noticed on latest (todays) winuae.zip the above demo problem returned back ... the last version where it was working fine is winuae.zip 2013-12-21 15:12 !!! |
24 December 2013, 09:01 | #508 | ||
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
Quote:
This is now fully fixed. (Remember that corruption happens when "landscape" part starts, do not attempt to test it by loading statefile saved later) |
||
24 December 2013, 10:22 | #509 | ||||
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
EDIT: Turnips-PsychoMedium4 now works better but left side of the box is wrong.
Quote:
Quote:
Quote:
Quote:
Last edited by Toni Wilen; 24 December 2013 at 10:36. |
||||
24 December 2013, 11:31 | #510 |
Amiga Lover
Join Date: Jan 2010
Location: Stockholm / Sweden
Age: 46
Posts: 612
|
Toni: Your work is amazing! Merry Christmas.
|
24 December 2013, 12:35 | #511 | ||
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
Quote:
Now only one problem remains in "most likely CPU emulator timing" issues ... thats really great !!! Quote:
p.s. Hope will manage to finish verifying all planed demos/intros for this year (now only letters V,W,X,Y,Z remains) ... so will continue for two days. Merry Christmas !!! EDIT: To Moderator, it is reached character limit on page 1 post 2 ... could be created another empty page 3 below page 2 to continue with updating. TIA Last edited by amilo3438; 24 December 2013 at 13:15. |
||
24 December 2013, 19:59 | #512 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
Just noticed that from "Demos with "sprite related" issues":
Laurent & Anna (Demo) by Saturne & Trilogy -> http://eab.abime.net/showpost.php?p=...&postcount=384 no anymore vertical line appears in middle of the screen ... hm, does this has been fixed or it happens by accident ?! (p.s. It seems to be "fixed" from winuae.zip test version 2013-12-15.) EDIT: It applies also for "Demos with "scanline emulation optimization" issues": Blue Venture (Demo) by Anxious -> http://eab.abime.net/showpost.php?p=...&postcount=400 (p.s. It seems to be "fixed" from winuae.zip test version 2013-12-22.) But cant find anywhere written confirmation that above 2 demos has really been fixed or that is a side effect from some other fix. Last edited by amilo3438; 24 December 2013 at 20:18. |
25 December 2013, 00:55 | #513 |
Global Moderator
Join Date: Aug 2008
Location: Sidcup, England
Posts: 10,300
|
@amilo3438: I have inserted a blank post #3 in the thread following the first two posts for you to continue adding links.
|
25 December 2013, 19:03 | #514 |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
It does not need confirmation, if it works now, it is fixed (related problem) or something else hide the problem. Looking for possible hidden problems from working programs is not worth the trouble
|
25 December 2013, 20:07 | #515 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
Yeah, now look and work good so no need for fixing anything.
P.s. There is one motto ... If it aint broke, don't fix it. Last edited by amilo3438; 25 December 2013 at 20:31. |
26 December 2013, 19:23 | #516 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
Conformism (Trackmo) by Vega
A500 ECS (most common) ... part title screen (pic.1) -> some holes on a dots flag screen -> pic.2 !!! First Intro (Demo) by Violent Wave A500 OCS/ECS (most common) ... dash & dots ... pic.3 !!! Instant Thing (Intro) by Visdom A500 OCS/ECS (most common) ... scrolltext twitching on ECS !!! Vertical Pulse (Demo) by Visdom A500 OCS/ECS (most common) ... pic.4 !!! Ace Ball (Crack Intro) by Wanted Team A500 OCS/ECS (most common) ... rotating cube occasionally disappears from screen in CE mode !!! (It doesnt happen if CE=OFF.) First Attempt (Demo) by Wave A500 OCS/ECS (most common) ... end scroller ... pic.5 !!! Last edited by amilo3438; 27 December 2013 at 14:17. |
27 December 2013, 14:42 | #517 |
Amiga 500 User
Join Date: Jun 2013
Location: EU
Posts: 1,502
|
First Intro (Intro) by Weird Science 2662
A500 OCS KS1.2/1.3 512 chip ... red dash ... pic.1 !!! Aegis Sonix v1.3 - import (Crack Intro) by Wizards A500 OCS KS1.2 512 chip ... it doesnt work ?! (Note: Its released 1987.) Last edited by amilo3438; 27 December 2013 at 16:11. |
27 December 2013, 17:59 | #518 | |
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
This was really good test case. There is small delay before BPLCON1 write and before output changes. Image is correct only if delay is exactly right. |
|
27 December 2013, 18:31 | #519 | ||
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
Quote:
Code:
MOVE.L #my_copper_list,$000025c0 |
||
27 December 2013, 20:02 | #520 | ||||||
WinUAE developer
Join Date: Aug 2001
Location: Hämeenlinna/Finland
Age: 49
Posts: 26,502
|
Quote:
Quote:
Quote:
Quote:
Quote:
Quote:
I think I can finally say that chipset and 68000 CPU is now practically 100% (minus stupid bugs) cycle-exact and biggest remaining problems are: blitter register modifications while active and reading write-only registers that isn't correctly emulated yet. Thanks for finding HUGE amounts of test cases (the more there is similar test cases, the easier it is to solve). (Of course next problem you find is something unexpected..) |
||||||
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Problematic GVP A530 Turbo | Mogsy | support.Hardware | 22 | 13 November 2012 19:12 |
Problematic thread? | prowler | project.EAB | 5 | 28 October 2011 00:06 |
Problematic Holiday Lemmings 94 demo disk | MethodGit | support.Games | 19 | 08 November 2010 03:56 |
Problematic slaves list | keropi | project.Killergorilla's WHD packs | 44 | 28 November 2008 12:46 |
Looking for two demos... | eternalcrusader | request.Demos | 3 | 01 August 2006 08:57 |
|
|