irqs and last_ipl[2] already set ...

[Ken T. <ke...@we...>]

I'm having yet another go at getting the 53c7xx.c driver to drive my A4091
reliably.  Last time I tried I came to the conclusion it was some
ZORRO3/PPC timing glitch.  As the new 53c770.c UW SCSI driver displays the
same problem, it might not be a timing glitch.

First question, are the kernel messages I see sometimes of the form:

last_ipl[2] already set to 2f, now 2d!
207863: 2 -2 2 -2 2 -2 2 -2 2 -2 2 -2 2 3 -3 2

indicative of some error in interrupt handling ?

Secondly, does each device drivers's interrupt handler need to check the
dev_id parameter that's passed to determine if the interrupt is from/for
that device ?

(Reading the two interrupt status registers of the 53c710 has side effects
with or without pending interrupts)

ken.