Re: -current panic on amd64

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

tron%zhadum.org.uk@localhost wrote:

> On my HP Proliant ML110 G4 problems stopped after the pullup of
> ticket #10 into the "netbsd-5" branch. But it seems there are
> more bugs lurking around.

It looks ipmi.c still uses CFATTACH_DECL() (not CFATTCH_DECL_NEW())
even after device_t/softc split.

On the other hand, there is a report that sometimes ipmi is not
probed properly (and then it doesn't panic during attach2).
Maybe we have to rethink how long delay is required
in bmc_io_wait_spin()?
---
Izumi Tsutsui