On Sun, 26 Mar 2000, Andrzej Bialecki wrote:
On Sun, 26 Mar 2000, Doug Rabson wrote:
On Sun, 26 Mar 2000, Andrzej Bialecki wrote:
Well, somehow the idea of overlapping subtrees sounds nice and useful
IMHO. Any suggestions how to solve these issues?
One possible way to do it
On Fri, 24 Mar 2000, Andrzej Bialecki wrote:
Hi,
Inspired by PR kern/16928 I implemented completely dynamic
creation/deletion of sysctl trees at runtime. The patches (relative to
-current) can be found at:
http://www.freebsd.org/~abial/dyn_sysctl.tgz
Included is an example of
On Sun, 26 Mar 2000, Doug Rabson wrote:
On Fri, 24 Mar 2000, Andrzej Bialecki wrote:
Hi,
Inspired by PR kern/16928 I implemented completely dynamic
creation/deletion of sysctl trees at runtime. The patches (relative to
-current) can be found at:
On Sun, 26 Mar 2000, Andrzej Bialecki wrote:
On Sun, 26 Mar 2000, Doug Rabson wrote:
On Fri, 24 Mar 2000, Andrzej Bialecki wrote:
Hi,
Inspired by PR kern/16928 I implemented completely dynamic
creation/deletion of sysctl trees at runtime. The patches (relative to
I think that if the sysctl data was reorganized, so that the per
module or instance data was at the leaves of the tree, you could avoid
the problem entirely. This is the general approach used on MIB definitions
used for SNMP; each variable is an instance (usually the 0th) at the
leaf. You
On Sun, 26 Mar 2000, Doug Rabson wrote:
On Sun, 26 Mar 2000, Andrzej Bialecki wrote:
Well, somehow the idea of overlapping subtrees sounds nice and useful
IMHO. Any suggestions how to solve these issues?
One possible way to do it would be to keep some ID of the oid's
creator. Then,
Hi,
Inspired by PR kern/16928 I implemented completely dynamic
creation/deletion of sysctl trees at runtime. The patches (relative to
-current) can be found at:
http://www.freebsd.org/~abial/dyn_sysctl.tgz
Included is an example of KLD that creates some subtrees when loaded, and
On Fri, 24 Mar 2000, Andrzej Bialecki wrote:
I'd appreciate some feedback. Thanks!
Note this is not an actual peer review (yet), but... Good job! This
is another big step in the right direction, and the code looks good
to me :) The only problems I can see right now are just all style
bugs