cvs commit: src/sys/kern kern_thread.c

[Julian Elischer]

Brian Fundakowski Feldman wrote:
> On Sun, Oct 03, 2004 at 02:33:14PM -0700, Julian Elischer wrote:
> 
>>Andrey Chernov wrote:
>>
>>>On Sun, Oct 03, 2004 at 08:06:11PM +0000, Julian Elischer wrote:
>>>
>>>
>>>>  sys/kern             kern_thread.c 
>>>>Log:
>>>>Always strt out with an initilalised ksegrp structure.
>>>
>>>
>>>Is it supposed to fix slow coming afterboot hang I see in -current already 
>>>about 2 months (last working kernel is from Aug 8)? My machine is UP with 
>>>4BSD sched running nothing more complex than mysqld.
>>>
>>
>>probably not, but there some changes coming that may make a difference
> 
> 
> Isn't this the one where I asked, "How is this uninitialized?" in any
> of the existing code?

you mean "How is this initialiased?"
I don't remember the question, but it's not a simple as it seems..
it is partly initilised during use.. teh trick is that the scheduler
private extension isn't.

(it may however be nulls by default from UMA but I don't know that).