@ssolie Quote:

That issue showed up during beta test but only for a minority of testers. ... Any help with nailing down what triggers that graphical glitch problem would be greatly appreciated.

I get the same problem, and would be happy to run tests or report the set-up of my Sam440 (original not Flex). So far:

* I found that disabling DMA reduced the size (length) of the red lines, but did not get rid of it.

* I also found that opening a Workbench window before Workbench has finished starting WbStartup stuff PREVENTED the red lines from appearing.

* Reloading the Workbench background pattern (using Prefs/WBPattern) also got rid of the lines.

* Removing all WbStartup programs & user-startup stuff did NOT get rid of the lines.

From this I conclude that the appearance of the red lines is related to the timing of when the background pattern is loaded, which may explain why it does not appear for everyone.

My first (uneducated) guess for WHY it is timing related is that (for those affected) the pattern is being loaded at a time of heavy CPU or HD activity, and this reveals some Sam440 driver bug. Most likely some kind of corruption of data as it is loaded from disk into memory. The fact that DMA has no effect rules-out DMA transfer, but perhaps it is something else related to the SATA driver?

Last edited by ChrisH on 22-Jun-2009 at 09:18 AM.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

Thank you for this new update and cheer up! that I waiting more.

Where do i get the newest installer?

The quickfix installer asks for it.

Edit: ok, i had an old installer, replaced
it with the one on the OS4.1 cd.

Last edited by Dirk-B on 22-Jun-2009 at 09:53 AM.

_________________
A1G3-SE + OS4.1 u1 iso (x2)

..what I need to say...Great Hyperion..Just install it on omy MicroA1 and for now all is oK!... And yeah no more Graphic Glitches for me.. hmm..sweet :D

_________________
--

Amiga x5000 ı o2o ı 4GB RAM ı RadeonRX580 | SBlaster Audigy Fx - AmigaOS4.1 FInal Edition

A1200 sandwich :)

Croatian Amiga portal

@ssolie
I found the 'cause' of the red lines appearing in my Workbench window background, it is the new "Sys:Kickstart/kernel" file.

Specifically, replacing it with the old file (and doing a hard reset) gets rid of the problem. However, I noticed that the old "kernel" file is identical to the old "kernel.debug" file, so I tried replacing the "kernel" file with the NEW "kernel.debug" file, and it STILL gets rid of the problem! THEREFORE, the real 'cause' of the red lines now appearing is that I am no-longer using a "debug" kernel! But since I suspect that the red line's appearance is timing related, the use of the debug kernel is probably a red herring, since it is likely just changing the timing of things.

So for the moment I will be using the debug kernel unless it causes more problems than it solves.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

@ChrisH

Just nitpicking: what are you saying is that the old kernel is in fact the debug one, since it has the same size as the kernel.debug file... Couldn't it be the other way around, i.e. the old version of the kernel runs fine, and the debug one is in fact the non-debug one, while with the new kernel only the debug one works fine?

Varthall

_________________
AmigaOne XE - AmigaOS 4.1 - Freescale 7457 1GHz - 1GB ram

I've got no clue about AOS4, but maybe it has got something to do with VGA text mode (debug?) output (0xb8000) ? If this is used but switched off too late (when some bitmap already has been allocated in VRAM) then maybe it's trashing VRAM used by that bitmap which happens to be around the same place in VRAM which is also used for VGA text mode.

@ChrisH
The strange part of the graphic corruption for me is that I don't get the problem when workbench first finishes loading but only after I have opened a number of Workbech windows. So It looks like the window backdrop is being loaded correctly into graphics card memory then damaged by something else later on.

_________________
Sam440ep 667mhz, 512MB, 120GB 2.5" HD, OS4.1FE
WinUae 3.0.0, OS 3.9, BB3, Catweasel MkIV
Amiga 1200, Blizzard 040/40 (BlizzardPPC 060/200 with SCSI removed at present), mediatorSX pci, Voodoo3, PCI network card os 3.9BB2
4MB Minimig with ARM addon boar

@ChrisH

I think you'll find that the problem is not "fixed" because you replaced a particular component, but because you changed the way in which the system components are loaded into memory.

The graphic corruption is most likely caused by a rogue program writing into memory that is a certain distance away from the program (or its data). If there is program code there, the rogue program will crash when it tries to write into executable memory. If there is data there, the rogue program will corrupt that data. The corruption may make the victim program crash. If there is unallocated memory there and you are using the debug kernel, the rogue program will crash. But, if the victim data is a window bitmap, you'll see corruption in the window.

What this means is that if you change a kickstart module (eg the kernel) to another of a different size, everything loaded after it is moved in memory. You could also try swapping kickstart components around in the kicklayout or running other programs in a different order, etc. Almost anything will change the result, which is why the cause is very difficult to find and why there are so many suggested "fixes".

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

@kas1e:

BCPL programs have never been supported in OS4. Support was not carried over from OS3.

I don't know why you are only seeing a message now and not before.
Perhaps some obsolete DOS functions for BCPL programs survived until recently.

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

@Varthall Quote:

Couldn't it be the other way around, i.e. the old version of the kernel runs fine, and the debug one is in fact the non-debug one, while with the new kernel only the debug one works fine?

No, because we were told that the Sam's OS4.1 beta's kernel was always running in debug mode. For example, the debug kernel caused old DvPlayer to not work on Sam440, until DvPlayer was updated. Since the "kernel" & "kernel.debug" files are identical (not just size!), this is clearly how they made it always use debug mode.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

@Dwyloc
Thanks for your report, it seems to corroborate Tonyw's explanation. It also explains why opening a Workbench window stops the red lines for me - I assumed it forced WB to load the pattern earlier, but that seemed a bit strange (unlikely). Now it seems that opening a WB window simply changes memory layout.

@tonyw
Since opening a WB window stops the red lines (see previously for more info), it appears that the memory corruption happens while Workbench is loading. As disabling WbStartup & user-startup has no effect, I suggest that either Workbench itself is to blame, or some system component that is used by Workbench (e.g. a datatype).

Last edited by ChrisH on 22-Jun-2009 at 03:18 PM.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

Thanks,

on my Peg2 I have experienced flickering when moving windows (only with transparency) in 1920x1080, in lower resolutions the flickering goes away. I did not experience any flicker with VGA only with DVI connection, but VGA @ 1920x1080 is so blurry it is not worth it.

I am wondering if the compositing+ActivateWindow+pri_lower_than_0-bug I reported a few month ago has been fixed?

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

Thanks for the update

_________________
GDG-Entertainment
OS4Games
Iksnet
Bitplane

Nope!

USB still will not output. Input only. No Mem sticks or MIDI out. Just keyboards and mice.

Phooey!

@mechanic
No output? I don't understand - my sam440ep uses USB Mem sticks all the time, not only reading but of course writing as well...and all sticks I had here worked at once without any problem.

@portable
btw: keyboard and mouse had never lost connection to the system while in use since i had the system up and running (three months ago, used nearly every day).

_________________
X1000|II/G4|440ep|2000/060|2000/040|1000

@Comi

God hears you matey. We, forgotten "classic" OS 4.0 users bought a product which has not been updated in more than a year now. We don't even have simple (not cpu demanding) workbench functionality features like sort buttons on top of workbench list windows... not to mention JXFS which could run on classics also.. etc. Leave alone the bugs like not being able to display 1280x1024x24/32 with a BlizzardVisionPPC board which is perfectly capable of displaying such a resolution on OS3.x with CGFX. Please Hyperion, don't forget the classic users!


Just my $0.02...

Cheers,

Dragster

_________________
Powerbook G4 1139/MOS 3.9/OSX 10.5.8. Pegasos II 2B5 Mobo. 2 GB RAM. 400 GB x 2 GB ATA HD, USB 2.0 PCI. SoundBlaster Audigy 2ZS Gold. Quad boot: AmigaOS 4.1 FInal Edition, MorphOS 3.9 regged, Debian Lenny PPC, OpenSUSE 11.1. Rockin'

hmm..have problem wiht this update...when i turn on Amiga for first time I cant boot WB.. after OS4.1 logo I get Wb screem...without icon ..only empty sceet ant title bar "Workbenc screen" after reset all is OK!

What is problem ?

before this update I dont have this problem!

_________________
--

Amiga x5000 ı o2o ı 4GB RAM ı RadeonRX580 | SBlaster Audigy Fx - AmigaOS4.1 FInal Edition

A1200 sandwich :)

Croatian Amiga portal

@Deniil715 Quote:

I am wondering if the compositing+ActivateWindow+pri_lower_than_0-bug I reported a few month ago has been fixed?

Using your test program (compiled using PortablE for OS4), it did not crash/freeze, so it does seem to be fixed. (And yes, I remembered to turn on transparency of inactive windows.)

@VooDoo
I guess you will have to do some detective work, since no-one else (yet) reports this problem. e.g. Does it happen if you disable (rename) Sys:WBStartup folder & Sys:S/user-startup file?

Last edited by ChrisH on 22-Jun-2009 at 09:34 PM.
Last edited by ChrisH on 22-Jun-2009 at 09:33 PM.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...