Darren J Moffat wrote:
Garrett D'Amore wrote:
Hmm... okay, I understand that some believe we may need kstats to
outlast a ddi-detach.
Although I'm not 100% convinced on that matter. I still believe that
any device for which you're collecting kstats that are interesting to
admins is probably either a disk or a nic, and generally those are
only interesting to admins if they are in use and thus not eligible
for detach by modunload -i 0.
or a crypto provider and those do get unloaded sometimes when not in
use but the data over the whole time of the system is probably more
interesting. However the crypto kstats are already special since we
lie to users not asserting all privilege that they are always zero.
We have kstats for both hardware and software providers. In half the
cases of our hardware providers there is no /dev entry and nothing in
userland holding them open unless they are being actively used.
I think this represents a bug.
More specifically, this means the crypto provider for hardware is likely
_not present_ in the system and therefore unavailable for use. (How can
a kcf consumer arrange for hardware crypto acceleration if the driver
isn't in use, and there is no /dev node?)
With the old kcl on Solaris 8, I believe we had a rule that kcl
providers were excluded from autounload. They could still be manually
detached and unloaded, but they were immune to modunload -i 0.
-- Garrett
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
