cvs commit: src/sys/pci if_sis.c if_sisreg.h

[Doug Rabson]

On Thu, 2003-09-04 at 02:40, M. Warner Losh wrote:
> In message: <200309030740.h837e4qS056204 at repoman.freebsd.org>
>             Poul-Henning Kamp <phk at FreeBSD.org> writes:
> :   We can't update the device description in attach (why not ?), so
> :   we device_print() what we find.
> 
> You can.  However, we use the description after PROBE, and changing it
> in attach doesn't cause us to go back and unprint it, so it does
> little good.  We could fix this, and the resource allocation issues,
> by printing things after a successful attach, but that would likely
> confuse people.

We could also move the print to after attach. That might alleviate some
of the confusion when an attach fails - in the log, it looks like the
device exists. On the other hand, that wouldn't work - some attach
methods print stuff and many create probe and attach other stuff.

For specific drivers, if they really want to change the device
description in attach (why not do it in probe like every other driver),
they could mark the device as quiet and then explicitly call
device_print_child after they have massaged the device.