[Simh] Fwd: 4.3BSD Quasijarus0c on the VAX-11/730

[Ethan Dicks]

On Wed, Jul 2, 2014 at 9:57 AM, Ethan Dicks <ethan.dicks at gmail.com> wrote:
> On Wed, Jul 2, 2014 at 8:44 AM, Cory Smelosky <b4 at gewt.net> wrote:
>>> root device? rb0*         <--- Enter "rb0*" here.
>>> idc0: lost interrupt       <--- This error repeats infinitely.
>
> One should be able to attach an RA81 to the emulated 11/730, run from
> that, then use the running system to debug problems with the RB730
> code.  It would at least allow a controlled poking at the driver, and
> even adding driver debug code.

OK... so getting caught up on this, I see that device debugging
support is already included in recent versions of SIMH.  Very nice.
I've enabled it ("set rb debug") and gone through the boot process
described previously and there aren't that many register writes up to
the "lost interrupt" point...

root device? rb0*
DBG(319687933)> RB REG: reg 0 write, value = 100C6
DBG(319687934)> RB REG: reg 3 write, value = 250000
DBG(319687940)> RB REG: reg 0 write, value = 46
DBG(319687957)> RB REG: reg 0 read, value = 4000046
DBG(319688004)> RB REG: reg 0 read, value = 4000046
DBG(319688051)> RB REG: reg 0 read, value = 4000046
DBG(319688090)> RB CMD: Seek, CYL=37, TRK=0, SECT=0
DBG(319688090)> RB CMD: Done, INT
DBG(319688098)> RB REG: reg 0 read, value = 50100C6
DBG(319688104)> RB REG: reg 0 read, value = 50100C6
DBG(319688104)> RB REG: reg 0 write, value = 50000C6
DBG(319688105)> RB REG: reg 0 read, value = 40000C6
DBG(319688140)> RB REG: reg 0 read, value = 40000C6
DBG(319688143)> RB REG: reg 0 write, value = C0
DBG(319688440)> RB CMD: Done, INT
DBG(319688558)> RB REG: reg 0 read, value = 50000C1
DBG(319688561)> RB REG: reg 0 write, value = C0
idc0: lost interrupt
DBG(523421810)> RB REG: reg 0 read, value = 40000C1
DBG(523421813)> RB REG: reg 0 write, value = C0
idc0: lost interrupt
DBG(729448369)> RB REG: reg 0 read, value = 40000C1
DBG(729448372)> RB REG: reg 0 write, value = C0
idc0: lost interrupt

I haven't traced out every bit of every transaction there (I see the
disk address (DAR) write, but no bus address (BAR) or byte count (BCR)
writes yet), but once it's in the loop, the CSR is reading back
R80+CRDY+IE+DRDY (R80 selected, controller ready, interrupts enabled,
and drive ready) and the BSD driver is writing into it a "Get status"
function (010 in F2:F0) and clearly not seeing what it is expecting.

So from thee top, I see the BSD driver put in a Seek command, the SIMH
rb handler process that seek and interrupt the BSD driver, which then
fiddles a few things (ATTN0 and IR) to acknowledge the interrupt, then
sends a status request, which appears to trigger the second interrupt,
but the BSD driver must not be seeing what it likes because it loops
on sending that status request every few seconds.

Probably time to hit the idc driver in 4.3BSD to see what _it_ thinks
should be happening.

-ethan