svn commit: r280495 - head/sys/kern

Hans Petter Selasky hps at selasky.org
Wed Mar 25 11:06:10 UTC 2015


On 03/25/15 11:54, Gleb Smirnoff wrote:
> On Wed, Mar 25, 2015 at 08:55:35AM +0000, Hans Petter Selasky wrote:
> H> Author: hselasky
> H> Date: Wed Mar 25 08:55:34 2015
> H> New Revision: 280495
> H> URL: https://svnweb.freebsd.org/changeset/base/280495
> H>
> H> Log:
> H>   Implement a simple OID number garbage collector. Given the increasing
> H>   number of dynamically created and destroyed SYSCTLs during runtime it
> H>   is very likely that the current new OID number limit of 0x7fffffff can
> H>   be reached. Especially if dynamic OID creation and destruction results
> H>   from automatic tests. Additional changes:
> H>
> H>   - Optimize the typical use case by decrementing the next automatic OID
> H>   sequence number instead of incrementing it. This saves searching time
> H>   when inserting new OIDs into a fresh parent OID node.
> H>
> H>   - Add simple check for duplicate non-automatic OID numbers.
>
> Why didn't you use alloc_unr(9) for that?
>

Hi Gleb,

I thought about using alloc_unr(). The problem is that sysctls have no 
clear concept of freeing. For example in some existing code sysctl are 
unregistered and registered again assuming that the oid_number will be 
preserved. I didn't want touch those parts. Also, hence we are already 
traversing a list to insert an SYSCTL object in an ordered fashion, the 
benefit of alloc_unr() is not that big. What do you think?

--HPS





More information about the svn-src-all mailing list