Re: Sun4/110 "Panic: Crazy Interrupts"

Hi!

I'm not sure I understand the P4 register readout - but here's what it is:
`P4 register = 0x4261699584x`

We still panic on the same line (bt->bt_addr[0]):
`cpu0: data fault: pc=0xf00122cc addr=0xfe048000 ser=0x8020<WRITE,TIMEOUT>`

No such luck commenting out the cmap lines unfortunately; still just
stray interrupts.

I appreciate all the help thus far!

Samuel

On Fri, Apr 26, 2024 at 1:25 AM Julian Coleman <[email protected]> wrote:
>
> Hi,
>
> > I got some time to test this patch out - it looks like we panic at line 327:
> > ```
> > bt->bt_addr[3] = 0;
> > ```
> >
> > And here's what it reads out on the screen:
> > ```
> > cpu0: data fault: pc=0xf0012290 addr=0xfe048003 ser=0x8020<WRITE,TIMEOUT>
> > panic: kernel fault
> > ```
>
> I was guessing which addresses we need to write to for the various registers
> and it seems that I was wrong here.  I'm assuming that we get the timeout
> because there is nothing mapped here on the CG4.  In the attached patch
> I've changed all the instances from bt_addr[3] to bt_addr[0] also moved
> the lines which check the P4 register and disable interrupts to before the
> colourmap setup.
>
> You should see a line printed with "P4 register = 0x....".  Does it have
> bit 2 set?  (https://nxr.netbsd.org/xref/src/sys/dev/sun/pfourreg.h#78)
> If it crashes in the colourmap code and the P4 register had bit 2 set, it
> might be worth just commenting out the cmap changes to see if it will get
> any further.
>
> Regards,
>
> Julian
>
> PS.  I used to have a CG4 in my 3/80, so I looked at the sun3 code to see if
> we do something different there.  The notes in sun3 cg4match code mention a
> type A CG4 with an AMD DAC found in the 3/110:
>
>   https://nxr.netbsd.org/xref/src/sys/arch/sun3/dev/cg4.c#167
>
> I wonder if your 4/110 has the type A CG4 as well, which might explain why
> the sparc CG4 driver isn't working.