Hi Kars!
On 10/28/19 10:19 AM, Kars de Jong wrote:
Then Real Life happened and I didn't touch my Amiga for 10+ years...
.. hmm, is this something that's likely to change? ;)
I've dusted it off and powered it on this weekend. It still works (RTC
needs a new battery though), and the 2.6.18 kernel that I modified
many years ago still works fine with the pcnet_cs driver and my
Dynalink L10C card.
Cool!
I then tried a build of master of Geert's tree, which unfortunately
doesn't fully work due to issues with the Blizzard ESP driver. I also
Could you report these issues? Is it the zorro_esp driver that's causing
problems? Normally the kernel should build fine but if there are issues,
Michael Schmitz would probably be interested to hear about them.
had similar issues with the last 2.6.2x kernel I tested in 2008(?),
after switching to the new ESP driver.
The drive is identified, set to synchronous 10 MHz mode, some
transfers are done but then it locks up. After the ESP driver runs its
recovery procedure, the bash that I run as init process exits with
code 4.
Wait, do you have build issues or do you have issues running the driver?
The new zorro_esp has seen a lot of rewrite. So old bugs might have been
fixed now.
Kars, you sound like you're in Europe - will you be on the Chaos
Communcation Congress in Leipzig this year? Will there be a Linux m68k
table?
Yes, I'm from Hengelo (the Netherlands), but no, I won't be attending.
I have no idea whether there will be a Linux/m68k presence.
I don't think there will be any m68k folk at CCC, I'm also not attending
as the event has become too political for my taste.
But I will be attending FOSDEM and Geert is coming as well (at least he
has been to every FOSDEM so far ;)), so we can maybe have a small m68k
meeting. I'm coming with two friends who are also Linux hackers.
I don't know how well bounties (like
https://www.bountysource.com/issues/80706251-m68k-convert-the-backend-to-mode_cc-so-it-can-be-kept-in-future-releases)
work, but I'd certainly willing to chip in with money, time and
resources to get this addressed for good.
I don't need money, I need time :-) I looked at the changes I made to
2.6.18, most of it is pretty self contained and should still work on
5.x without too many changes, but the io.h changes are more hairy.
Also, someone once tested this on his A1200 which had a 68030 and I
believe it broke. Something about ioremap() not accepting physical
addresses that were already mapped by a fixed mapping (Zorro II
address space).
Doesn't this depend on the type of accelerator used? I think some of them
cause incompatibilities with the PCMCIA slot, don't they?
I think cleaning up the IRQ handling etc would also be (another?) good
showcase of linux-m68k and the Amiga platform being very much alive,
contrary to what Linus might think :D
Linus knows it's alive, he just seems to wonder why it won't die ;-)
Because the Amiga community is absolutely die-hard :). This year's Amiga
conference was even extended to two days because of the public interest.
I will try to prepare a patch for the IRQ handling.
Nice. Thank you! Really cool to see more and more people starting to work
on Linux/m68k again :). This shows that all the effort I put into Debian/m68k
was worth it!
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@xxxxxxxxxx
`. `' Freie Universitaet Berlin - glaubitz@xxxxxxxxxxxxxxxxxxx
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
