Re: bootstra.tos broken on 68040/68060 real hardware

Hi Miro,

haven't run these through checkpatch yet but at first glance they look 
good to me.

Patches 1,2 and 5 look uncontroversial enough. 3 and 4 might need more 
thorough review by fbdev maintainers.

I can add my reviewed-by and submit to Geert, unless Geert can pick 
these patches up directly from your post?

Cheers,

     Michael

On 03/08/2026 5:53 PM, Miro Kropáček wrote:
> Hi,
>
> it has been some time since my last post and in the meantime I have 
> managed to get X.org & jwm not only fully working but also support 
> SuperVidel and Svethlana at the same time. So I'm posting my findings 
> here with the hope that some of you Linux folks would take care to 
> forward it to Geert and/or to upstream. I'm spending all my Atari time 
> on Atari/MiNT things, I can't find more time to follow up on this.
>
> Feel free to change, rewrite, remove, add whatever you like, I don't 
> even insist on being the author as long as the patches are shipped. 
> All of them have been tested by me on real Atari Falcon030 + CT60 + 
> SuperVidel + Svethlana + EtherNEC + 512MB + my own fixed ataboot 
> <https://github.com/mikrosk/m68kboot>. I have verified that they apply 
> cleanly even on Geert's tree.
>
> Apart from kernel patches I have managed to create a separate 
> supervidel X.org accelerated driver (full disclosure: I did it with 
> Claude Fable), if you are interested in that as well, just say so.
>
> From user point of view, I did a few optimisations / hacks to make my 
> experience more pleasant:
>
> - Disabled apt install / apt update signature verification 
> (/etc/apt/sources.list, /etc/apt/apt.conf.d/99no-verify)
>
> - To fix the broken keyring I had to use apt update 
> --allow-insecure-repositories & apt install --allow-unauthenticated 
> debian-ports-archive-keyring
>
> - yescrypt replaced with sha256 rounds=1000 (one has to run passwd 
> afterwards to take effect)
>
> - LOGIN_TIMEOUT 0 in /etc/login.defs
>
> - Speed up login: disable systemd's [email protected] in 
> /etc/pam.d/common-session
>
> - /etc/systemd/network/10-svethlana.link with SVEthlana's MAC/DHCP IP 
> being predictable:
> [Match]
> Driver=ethoc
>
> [Link]
> MACAddress=b4:bb:c2:fa:7d:20
>
> - kernel command line to filter out the debug messages: loglevel=3
>
> After all this, jwm was quite usable but the HW acceleration wasn't 
> immediately obvious. There seemed to be quite some lag coming from 
> X.org itself so when moving windows, it's nice that they are refreshed 
> in milliseconds but registering the move event takes almost a second.
>
> -- 
> http://mikro.atari.org