Hello Carlos!

On 9/26/21 14:01, Carlos Milán Figueredo wrote:

While a valid hd-media for current Unstable is being built by Adrian,
I though I could play around with Debian Sarge that was the latest
version with official m68k support. I got into my hands a D-Link
DE-660+ PCMCIA card that according [1] should work on Linux; I
already tested on AmigaOS and NetBSD/amiga and it's working like a
charm. The issue is that Debian Sarge doesn't detect it and there is
not even a kernel message referencing the pcmcia.

Did you check whether the necessary kernel modules are present on the
booted installer system? They should be in /lib/modules/$(VERSION)/*.

As you already know from the unstable hd-media image, not different
d-i images ship different module packages.

For a certain driver to be available in a debian-installer image, it first
must be enabled in the kernel configuration, so you have to check there
first.

Then it needs to be enabled as a debian-installer module in the directory debian/installer/* in the kernel packaging source [1].

Then the module udeb packages generated in [1] need to be included in
the debian-installer image package list if it's a udeb that must be
included in the initrd (for example, networking drivers for the network installer image).

The idea is that the initrd contains all the modules that are needed to
get the system up and running and access the installation medium. Thus,
the cdrom installer image needs storage drivers in its initrd while the hd-media image needs modules for various filesystems and loopback mounting.

In the past years, I have only worked on the cdrom debian-installer images
and not touched network and hd-media installation as most people use the standard installation CDs. That's why you ran into the problems with hd-media when you tested it.

But we will hopefully fix the issues with the hd-media installation images the next days. I just have to go through the module package lists and check which of these we need for HD installation and which will actually work. Firewire modules will certainly not work on m68k, for example.

Adrian

[1] https://salsa.debian.org/kernel-team/linux/-/tree/master/debian/installer/modules
[2] https://salsa.debian.org/installer-team/debian-installer/-/tree/master/build/pkg-lists

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@debian.org
`. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Carlos

On 27/09/21 04:34, Carlos Milán Figueredo wrote:

Hi Adrian,

From: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Sent: domingo, 26 de septiembre de 2021 14:51

Did you check whether the necessary kernel modules are present on the
booted installer system? They should be in /lib/modules/$(VERSION)/*.

On Debian Sarge it lists: 8390.o, a2065.o, apne.o, ariadne.o, dummy.o, hydra.o and zorro8390.o

The apne.o is the Amiga PCMCIA NE2000-compatible driver [1]. But it won't detect the DE-660+ for some reason (it ends on a kernel dump) if I force loading it.

Then it needs to be enabled as a debian-installer module in the directory
debian/installer/* in the kernel packaging source [1].

The module that contains apne driver is not on hd-media initrd, but it is in a udeb package -don't know which one- read after iso-scan got the ISO from any partition with a supported filesystem. So I can test it on Sarge hd-media, but not on recent

snapshots.

Also, I have seen there have been very recent kernel patches from Michael Schmitz to the Amiga PCMCIA driver [2], it looks like for probing 16-bit cards; so I think it will be better to make tests on a recent kernel snapshot rather than the old Sarge

distribution. I will wait for the fixed hd-media.

Please check whether your card is indeed one of the supported 8390 or compatible based cards (10 or 100 Mbit).

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Cheers,

Michael

If I am successful, I will make a install video -like the one I did for NetBSD [3]- and write updated install and FAQ documentation :)

Regards,
Carlos

[1] https://elixir.bootlin.com/linux/v5.14.7/source/drivers/net/ethernet/8390/apne.c
[2] https://lore.kernel.org/linux-m68k/1630206710-5954-3-git-send-email-schmitzmic@gmail.com/
[3] https://www.youtube.com/watch?v=hJ5fuCA5tM8

Carlos Milán Figueredo | HispaMSX System Operator | http://www.hispamsx.org | | telnet://bbs.hispamsx.org | https://calnus.com

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Michael,

On Mon, Sep 27, 2021 at 12:02 AM Michael Schmitz <schmitzmic@gmail.com> wrote:

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Thanks for your perseverance!
We're at rc3 (i.e. past merge window), so net-next should be open for submission.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 9/27/21 10:11, Geert Uytterhoeven wrote:

On Mon, Sep 27, 2021 at 12:02 AM Michael Schmitz <schmitzmic@gmail.com> wrote:

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Thanks for your perseverance!
We're at rc3 (i.e. past merge window), so net-next should be open for submission.

Yes, it would be great if these improvements could get merged. :-)

And I should finally get my X-Surf500 setup configured so we can work on
that driver as well.

Adrian

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@debian.org
`. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Adrian,

yes, and I urgently need another few weeks confined in my home office
to tackle that one ...

Cheers,
Michael

On Mon, Sep 27, 2021 at 9:28 PM John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> wrote:

On 9/27/21 10:11, Geert Uytterhoeven wrote:

On Mon, Sep 27, 2021 at 12:02 AM Michael Schmitz <schmitzmic@gmail.com> wrote:

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Thanks for your perseverance!
We're at rc3 (i.e. past merge window), so net-next should be open for submission.

Yes, it would be great if these improvements could get merged. :-)

And I should finally get my X-Surf500 setup configured so we can work on
that driver as well.

Adrian

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@debian.org
`. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Geert,

thanks for your words of encouragement!

On 27/09/21 21:11, Geert Uytterhoeven wrote:

Hi Michael,

On Mon, Sep 27, 2021 at 12:02 AM Michael Schmitz <schmitzmic@gmail.com> wrote:

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Thanks for your perseverance!
We're at rc3 (i.e. past merge window), so net-next should be open for submission.

Quick inspection of DaveM's netdev status page confirms that it indeed is.

I've so far sent APNE patches to net, not net-next, as these are
improvements to an existing driver. Anything wrong with that?

Cheers,

Michael

Gr{oetje,eeting}s,

Geert

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Michael,

On Tue, Sep 28, 2021 at 6:06 AM Michael Schmitz <schmitzmic@gmail.com> wrote:

On 27/09/21 21:11, Geert Uytterhoeven wrote:

On Mon, Sep 27, 2021 at 12:02 AM Michael Schmitz <schmitzmic@gmail.com> wrote:

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Thanks for your perseverance!
We're at rc3 (i.e. past merge window), so net-next should be open for submission.

Quick inspection of DaveM's netdev status page confirms that it indeed is.

I've so far sent APNE patches to net, not net-next, as these are
improvements to an existing driver. Anything wrong with that?

Fixes should have "[net]", new development "[net-next]".
I usually just ignore all of that, as I mainly send fixes to the netdev maintainers ;-)

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Carlos,

On 28/09/21 10:44, Carlos Milán Figueredo wrote:

Hi Michael!

From: Michael Schmitz <schmitzmic@gmail.com>
Sent: lunes, 27 de septiembre de 2021 0:00

Please check whether your card is indeed one of the supported 8390 or
compatible based cards (10 or 100 Mbit).

Taking a look at the Windows drivers, I can confirm it is indeed NE2000-compatible, but I was not able to find a reference to the 8390. It works on AmigaOS with cnet.device and on NetBSD/amiga out of the box.

NE2000 compatible means 8390 (or compatible) chipset.

Your card works with the cnet.device - that means 8-bit PCMCIA
interface, not 16 bit? (You can probably tell that my knowledge of
AmigaOS matters is limited ...)

I wouldn't hold my breath waiting for the 16 bit / 100 Mbit patches to
be accepted into the mainstream kernel. Haven't had any feedback from
netdev, and the review process on linux-m68k wasn't too encouraging
either. That said, I have another respin of the patch series in the
pipeline (need to check that this is a good time to submit patches to
netdev first).

Thanks for letting me know. Am I correct to assume that the patch is to improve "possible" Linux support for cards that under AmigaOS would need the cnet16.device on AmigaOS?

If 'cnet16.device' means 16-bit access on the PCMCIA interface, then
yes. And that would mean that patch series isn't going to be any help to
you.

You will have to log the kernel crash message you get when loading the
apne module, maybe someone can derive a hint on improving the existing
(8-bit) driver from that. A good look at the NetBSD driver source may be enlightening as well.

Cheers,

Michael

Regards,
Carlos

Carlos Milán Figueredo | HispaMSX System Operator | http://www.hispamsx.org/ | telnet://bbs.hispamsx.org/ | https://calnus.com

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Geert,

On 28/09/21 20:12, Geert Uytterhoeven wrote:

We're at rc3 (i.e. past merge window), so net-next should be open for
submission.

Quick inspection of DaveM's netdev status page confirms that it indeed is. >>
I've so far sent APNE patches to net, not net-next, as these are
improvements to an existing driver. Anything wrong with that?

Fixes should have "[net]", new development "[net-next]".

OK - patch has hopefully made it to netdev now.

I usually just ignore all of that, as I mainly send fixes to the netdev maintainers ;-)

The perks of a maintainer, I'm sure. But I may poke Arnd this time if no
review is forthcoming. He may remember ...

Cheers,

Michael

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Michael!

On 9/28/21 09:19, Michael Schmitz wrote:

Fixes should have "[net]", new development "[net-next]".

OK - patch has hopefully made it to netdev now.

I'm not seeing it here, am I missing something?

https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/log/

Adrian

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@debian.org
`. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Adrian,

On 28/09/21 21:17, John Paul Adrian Glaubitz wrote:

Hi Michael!

On 9/28/21 09:19, Michael Schmitz wrote:

Fixes should have "[net]", new development "[net-next]".

OK - patch has hopefully made it to netdev now.

I'm not seeing it here, am I missing something?

https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/log/

Not sure - I would think that is where it'll end up once reviewed and
accepted. Wish it was that fast.

Netdev use a patchwork queue to coordinate reviews - haven't had time to
look there yet.

Cheers,

Michael

Adrian

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)