Re: [PATCH v5 5/6] hppa: Add emulation of Artist graphics

[Philippe Mathieu-Daudé]

On Fri, Dec 27, 2019 at 9:58 PM Helge Deller <address@hidden> wrote:
> On 24.12.19 01:18, Philippe Mathieu-Daudé wrote:
> > On 12/23/19 6:50 PM, Sven Schnelle wrote:
> >> Hi Philippe,
> >>
> >> On Sun, Dec 22, 2019 at 01:37:48PM +0100, Philippe Mathieu-Daudé wrote:
> >>>>   +    if (vga_interface_type != VGA_NONE) {
> >>>> +        dev = qdev_create(NULL, "artist");
> >>>> +        qdev_init_nofail(dev);
> >>>> +        s = SYS_BUS_DEVICE(dev);
> >>>> +        sysbus_mmio_map(s, 0, LASI_GFX_HPA);
> >>>> +        sysbus_mmio_map(s, 1, ARTIST_FB_ADDR);
> >>>
> >>> How is this chipset connected on the board?
> >>> If it is a card you can plug on a bus, you can use a condition.
> >>> If it is soldered or part of another chipset, then it has to be mapped
> >>> unconditionally.
> >>
> >> Depends on the Model. Hp 9000 712 and 715 had it onboard, for the B160L
> >> we're emulating and others it was a GSC add-on card.
> >
> > The B160L case is unclear, do you mean this is not the chipset on the 
> > machine, but the software is happy if another chipset is available?
> >
> > Looking at hw/hppa/ I only see one machine:
> >
> >   static void machine_hppa_machine_init(MachineClass *mc)
> >   {
> >       mc->desc = "HPPA generic machine";
> >       ...
> >   }
> >   DEFINE_MACHINE("hppa", machine_hppa_machine_init)
> >
> > Are you saying this generic machine is able to run different physical hw? 
> > Why not add them? This shouldn't take long and it would be clearer, what do 
> > you think?
> >
> > Adding different machines here in QEMU mostly mean add a class which 
> > declare the different properties used by each machine. Igor Mammedov 
> > recently suggested to follow the example of aspeed_machine_types[] in 
> > hw/arm/aspeed.c.
>
> Yes, we plan to add specific machines like 712 (or 715), and maybe a
> C3000 or B2000 over time, as needed device emulations (e.g. tulip, artist)
> gets accepted.
> But for that it would be very beneficial if changes (like the Artist emulation
> here in this thread) would be accepted faster upstream....

IMHO the HPPA patches are merged quicker than various other subsystems...
>From the cover, this series contains "3311 insertions". If you do the
ratio lines per patch / time for a patch to get accepted, you are not
that bad ;)

To be constructive, how do you think we can improve?
Looking at the git history, except global refactors, I see only 3
contributors, Richard (who merges your patches), you and Sven.
2 users so far, + Richard as tester.