During Google Summer of Code 2011, Bryce Lanham added the possibility to
emulate the NeXTcube machine in QEMU, e.g. see this URL for some details:

https://wiki.qemu.org/Google_Summer_of_Code_2011#NeXT_machines_system_emulation

But since the machine requires a 68040 CPU and this was not included in
upstream QEMU in 2011 yet, the patches have never been merged to upstream.
Now in 2018, Laurent completed the full 680x0 support in upstream QEMU, so
we could finally merge the NeXTcube support, too.

The QEMU interfaces changed a lot since 2011, so I had to modify the
sources quite a bit, but with the attached patches, it is now possible
to boot up to the firmware monitor again.

Note that boot device emulation is either still missing (network), or
not working correctly yet (SCSI), so you can not boot any operating
systems with this machine yet.

Please ignore the checkpatch warnings about C99 "//" comments. There
are many of those in the sources, used to comment out alternative code
lines. I did not want to remove those yet, since some of them could
still be helpful while improving the machine emulation in the future.

Thomas Huth (4):
  m68k: Add NeXTcube framebuffer device emulation
  m68k: Add NeXTcube keyboard device
  m68k: Add NeXTcube machine
  m68k: Add an entry for the NeXTcube machine to the MAINTAINERS file

 MAINTAINERS                      |    7 +
 default-configs/m68k-softmmu.mak |    4 +
 hw/display/Makefile.objs         |    1 +
 hw/display/next-fb.c             |  152 ++++
 hw/m68k/Makefile.objs            |    5 +-
 hw/m68k/next-cube.c              | 1103 ++++++++++++++++++++++++++++++
 hw/m68k/next-kbd.c               |  289 ++++++++
 include/hw/m68k/next-cube.h      |   49 ++
 8 files changed, 1608 insertions(+), 2 deletions(-)
 create mode 100644 hw/display/next-fb.c
 create mode 100644 hw/m68k/next-cube.c
 create mode 100644 hw/m68k/next-kbd.c
 create mode 100644 include/hw/m68k/next-cube.h

-- 
2.17.1

On 30/06/18 09:33, Thomas Huth wrote:

> During Google Summer of Code 2011, Bryce Lanham added the possibility to
> emulate the NeXTcube machine in QEMU, e.g. see this URL for some details:
> 
> https://wiki.qemu.org/Google_Summer_of_Code_2011#NeXT_machines_system_emulation
> 
> But since the machine requires a 68040 CPU and this was not included in
> upstream QEMU in 2011 yet, the patches have never been merged to upstream.
> Now in 2018, Laurent completed the full 680x0 support in upstream QEMU, so
> we could finally merge the NeXTcube support, too.
> 
> The QEMU interfaces changed a lot since 2011, so I had to modify the
> sources quite a bit, but with the attached patches, it is now possible
> to boot up to the firmware monitor again.
> 
> Note that boot device emulation is either still missing (network), or
> not working correctly yet (SCSI), so you can not boot any operating
> systems with this machine yet.
> 
> Please ignore the checkpatch warnings about C99 "//" comments. There
> are many of those in the sources, used to comment out alternative code
> lines. I did not want to remove those yet, since some of them could
> still be helpful while improving the machine emulation in the future.
> 
> Thomas Huth (4):
>    m68k: Add NeXTcube framebuffer device emulation
>    m68k: Add NeXTcube keyboard device
>    m68k: Add NeXTcube machine
>    m68k: Add an entry for the NeXTcube machine to the MAINTAINERS file
> 
>   MAINTAINERS                      |    7 +
>   default-configs/m68k-softmmu.mak |    4 +
>   hw/display/Makefile.objs         |    1 +
>   hw/display/next-fb.c             |  152 ++++
>   hw/m68k/Makefile.objs            |    5 +-
>   hw/m68k/next-cube.c              | 1103 ++++++++++++++++++++++++++++++
>   hw/m68k/next-kbd.c               |  289 ++++++++
>   include/hw/m68k/next-cube.h      |   49 ++
>   8 files changed, 1608 insertions(+), 2 deletions(-)
>   create mode 100644 hw/display/next-fb.c
>   create mode 100644 hw/m68k/next-cube.c
>   create mode 100644 hw/m68k/next-kbd.c
>   create mode 100644 include/hw/m68k/next-cube.h

Looks like the most interesting patch (patch 3) has been blocked 
somewhere, presumably because it is quite large.

Is there a git repository for this patch series somewhere for the curious?


ATB,

Mark.

Am Sat, 30 Jun 2018 10:42:56 +0100
schrieb Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>:

> On 30/06/18 09:33, Thomas Huth wrote:
> 
> > During Google Summer of Code 2011, Bryce Lanham added the
> > possibility to emulate the NeXTcube machine in QEMU, e.g. see this
> > URL for some details:
> > 
> > https://wiki.qemu.org/Google_Summer_of_Code_2011#NeXT_machines_system_emulation
> > 
> > But since the machine requires a 68040 CPU and this was not
> > included in upstream QEMU in 2011 yet, the patches have never been
> > merged to upstream. Now in 2018, Laurent completed the full 680x0
> > support in upstream QEMU, so we could finally merge the NeXTcube
> > support, too.
> > 
> > The QEMU interfaces changed a lot since 2011, so I had to modify the
> > sources quite a bit, but with the attached patches, it is now
> > possible to boot up to the firmware monitor again.
> > 
> > Note that boot device emulation is either still missing (network),
> > or not working correctly yet (SCSI), so you can not boot any
> > operating systems with this machine yet.
> > 
> > Please ignore the checkpatch warnings about C99 "//" comments. There
> > are many of those in the sources, used to comment out alternative
> > code lines. I did not want to remove those yet, since some of them
> > could still be helpful while improving the machine emulation in the
> > future.
> > 
> > Thomas Huth (4):
> >    m68k: Add NeXTcube framebuffer device emulation
> >    m68k: Add NeXTcube keyboard device
> >    m68k: Add NeXTcube machine
> >    m68k: Add an entry for the NeXTcube machine to the MAINTAINERS
> > file
[...]
> Looks like the most interesting patch (patch 3) has been blocked 
> somewhere, presumably because it is quite large.

Weird... the mail was ca. 1200 lines ... that not small, but not so
big that I'd expect that it would get blocked somewhere. Let's wait
some more more hours, maybe the spam filter is just very slow today. If
it is then still not available on the list, I'll try to send it out
again.

> Is there a git repository for this patch series somewhere for the
> curious?

I just pushed it here:

https://gitlab.com/huth/qemu/tree/next-cube

 Thomas

On 30 June 2018 at 14:14, Thomas Huth <huth@tuxfamily.org> wrote:
> Weird... the mail was ca. 1200 lines ... that not small, but not so
> big that I'd expect that it would get blocked somewhere. Let's wait
> some more more hours, maybe the spam filter is just very slow today. If
> it is then still not available on the list, I'll try to send it out
> again.

The mailing list does just occasionally randomly eat patches,
as far as I can tell. It's not purely a size thing.

thanks
-- PMM

Am Sat, 30 Jun 2018 14:17:32 +0100
schrieb Peter Maydell <peter.maydell@linaro.org>:

> On 30 June 2018 at 14:14, Thomas Huth <huth@tuxfamily.org> wrote:
> > Weird... the mail was ca. 1200 lines ... that not small, but not so
> > big that I'd expect that it would get blocked somewhere. Let's wait
> > some more more hours, maybe the spam filter is just very slow
> > today. If it is then still not available on the list, I'll try to
> > send it out again.  
> 
> The mailing list does just occasionally randomly eat patches,
> as far as I can tell. It's not purely a size thing.

Ok, good to know. I've just sent out patch 3/4 again, let's see how it
goes.

 Thomas

Oh hi!

Luckily gmail brought this to the top since I don’t pay attention to the
list. I’m away from my computer at the moment, but I had more than this
working, including interrupts, Ethernet, and SCSI for sure. (
https://i.imgur.com/Py0FO.png) I’d love to dive back into this now that
68040 support is complete and merged.


On Sat, Jun 30, 2018 at 8:34 AM, Thomas Huth <huth@tuxfamily.org> wrote:

> Am Sat, 30 Jun 2018 14:17:32 +0100
> schrieb Peter Maydell <peter.maydell@linaro.org>:
>
> > On 30 June 2018 at 14:14, Thomas Huth <huth@tuxfamily.org> wrote:
> > > Weird... the mail was ca. 1200 lines ... that not small, but not so
> > > big that I'd expect that it would get blocked somewhere. Let's wait
> > > some more more hours, maybe the spam filter is just very slow
> > > today. If it is then still not available on the list, I'll try to
> > > send it out again.
> >
> > The mailing list does just occasionally randomly eat patches,
> > as far as I can tell. It's not purely a size thing.
>
> Ok, good to know. I've just sent out patch 3/4 again, let's see how it
> goes.
>
>  Thomas
>

Am Sat, 30 Jun 2018 12:07:01 -0500
schrieb Bryce Lanham <blanham@gmail.com>:

> Oh hi!
> 
> Luckily gmail brought this to the top since I don’t pay attention to
> the list. I’m away from my computer at the moment, but I had more
> than this working, including interrupts, Ethernet, and SCSI for sure.
> ( https://i.imgur.com/Py0FO.png) I’d love to dive back into this now
> that 68040 support is complete and merged.

 Hi Bryce,

great to hear that you are still around! If you like to take care of
this again, that would be very welcome! So I'll keep my hands off these
patches for now, and let you handle this again :-)

For what it's worth, I've pushed my final state of the patches here:

 https://gitlab.com/huth/qemu/tree/next-cube

I've also managed to get the next-nic.c code compilable again, but as I
mentioned earlier, it's currently not working anymore. Not sure, but I
think there is some wrong overlap in the memory mapping of the 0x110 and
0x150 DMA registers between next-nic.c and next-cube.c in there.

Anyway, feel free to use my patches as a base for your new work, or
simply ignore them if you prefer to start with your old codebase again.
Either way I'm just glad if we finally get an additional 68k machine in
QEMU :-)

 Cheers,
  Thomas

Now you've made me subscribe again to qemu-devel ;)

On 30/06/18 18:07, Bryce Lanham wrote:
> Oh hi! 
> 
> Luckily gmail brought this to the top since I don’t pay attention to the
> list. I’m away from my computer at the moment, but I had more than this
> working, including interrupts, Ethernet, and SCSI for sure.
> (https://i.imgur.com/Py0FO.png) I’d love to dive back into this now that
> 68040 support is complete and merged.
> 
> 
> On Sat, Jun 30, 2018 at 8:34 AM, Thomas Huth <huth@tuxfamily.org
> <mailto:huth@tuxfamily.org>> wrote:
> 
>     Am Sat, 30 Jun 2018 14:17:32 +0100
>     schrieb Peter Maydell <peter.maydell@linaro.org
>     <mailto:peter.maydell@linaro.org>>:
> 
>     > On 30 June 2018 at 14:14, Thomas Huth <huth@tuxfamily.org <mailto:huth@tuxfamily.org>> wrote:
>     > > Weird... the mail was ca. 1200 lines ... that not small, but not so
>     > > big that I'd expect that it would get blocked somewhere. Let's wait
>     > > some more more hours, maybe the spam filter is just very slow
>     > > today. If it is then still not available on the list, I'll try to
>     > > send it out again.  
>     > 
>     > The mailing list does just occasionally randomly eat patches,
>     > as far as I can tell. It's not purely a size thing.
> 
>     Ok, good to know. I've just sent out patch 3/4 again, let's see how it
>     goes.
> 
>      Thomas
> 
>