cvs commit: src/sys/dev/sound/isa ad1816.c ess.c mss.c sb16.c sb8.c sbc.c src/sys/dev/sound/pci cmi.c csa.c csapcm.c ds1.c emu10k1.c ich.c t4dwave.c src/sys/dev/sound/pcm sound.h

M. Warner Losh imp at
Wed Apr 14 08:56:44 PDT 2004

In message: <200404141522.i3EFM1eS049089 at>
            "Brian F. Feldman" <green at> writes:
: Scott Long <scottl at> wrote:
: > Brian Feldman wrote:
: > > green       2004/04/14 07:57:50 PDT
: > Ha! you touched it last!  Neener neener neener!
: Heh, I'll certainly do my own part to make sure the drivers I use get 
: converted.  It would suck less trying to do driver development if current 
: drivers didn't panic so much because of kldload/kldunload.  For example, 
: kldload usb stuff after it's been unloaded sees something stale and crashes, 
: and sound drivers get EBUSY when I don't see any openers, then the second
: kldunload causes a crash instead of returning EBUSY again because things
: must have been half-torn-down.

kldload/kldunload have nothing to do with it.  If the drivers are
stupid, that's their own damn fault.  usb is a known problem child
that needs to be fixed.  If there are others that need to be fixed,
then fix them.  You have the source, use it :-).

I've used kldload/kldunload extensively in the development of cardbus,
pccard and pccbb.  I'd be totally lost w/o it.


More information about the cvs-src mailing list