In message <[EMAIL PROTECTED]> Peter Wemm writes:
: > aha0: status reg test failed ff
: > unknown9: <SCSI Host Adapter> at port 0x330-0x333 irq 11 drq 5 on isa0
: ^^^^^^^^^^^^^^^^^^^^^
: Warner, what does this mean? the status register is 0xff? Could
: that mean that the 1542CP didn't get it's mapping at 0x330
: activated?
Likely. The register status 0xff means that we got back garbage from
the aha device. There are three bits we test for specifically, DIAG
active, command register busy and reg_rscvd (which may be more than
one bit). If any of these read 1, that almost always means that we're
not talking to real hardware. I surmise that the mapping to 0x330
didn't actually happen. I've had some reports of oddball aha
compatible EISA boards that had this problem in the past, but that was
due to a minor eisa bug, iirc.
We do all these tests mondo early in the probe process. We do it just
after the bus_space_alloc(...RF_ACTIVE); happens.
: Also:
: > ADP1542: adding memory range 0xc8000-0xdc03f, size=0x40, align=0x4000
: Could this be causing a failed allocation somewhere? (pr doesn't it matter
: since the ROM isn't being activated?)
That I can't tell you. We're definitely failing the probe because of
what the probe's talking to hardware results in its concluding that
the hardware isn't there.
Warner
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
