headers
Erik Fair | 19 Apr 2010 18:49
Picon

state of NetBSD SMP

So, what's the current state of NetBSD SMP, i.e. how many kernel  
subsystems do we still have to get out from under the big lock?

	curious,

	Erik <fair <at> netbsd.org>
Permalink | Reply | 5 comments |
headers
Martin Husemann | 19 Apr 2010 20:17
Picon

Re: state of NetBSD SMP

On Mon, Apr 19, 2010 at 09:49:16AM -0700, Erik Fair wrote:
> So, what's the current state of NetBSD SMP, i.e. how many kernel  
> subsystems do we still have to get out from under the big lock?

No idea about total count, but top three of next things to attack:

 - audio (there is a branch, supposedly ~50% done)
 - networking
 - tty (noone dares to touch it?)

Martin
Permalink | Reply | 4 comments |
headers
David Holland | 20 Apr 2010 17:47
Picon

Re: state of NetBSD SMP

On Mon, Apr 19, 2010 at 08:17:27PM +0200, Martin Husemann wrote:
 > > So, what's the current state of NetBSD SMP, i.e. how many kernel  
 > > subsystems do we still have to get out from under the big lock?
 > 
 > No idea about total count, but top three of next things to attack:
 > 
 >  - audio (there is a branch, supposedly ~50% done)
 >  - networking
 >  - tty (noone dares to touch it?)

I dare, but my hands are full :(

--

-- 
David A. Holland
dholland <at> netbsd.org
Permalink | Reply | 3 comments |
headers
David Laight | 28 Apr 2010 14:43
Picon

Re: state of NetBSD SMP

On Tue, Apr 20, 2010 at 03:47:03PM +0000, David Holland wrote:
> On Mon, Apr 19, 2010 at 08:17:27PM +0200, Martin Husemann wrote:
>  > > So, what's the current state of NetBSD SMP, i.e. how many kernel  
>  > > subsystems do we still have to get out from under the big lock?
>  > 
>  > No idea about total count, but top three of next things to attack:
>  > 
>  >  - audio (there is a branch, supposedly ~50% done)
>  >  - networking
>  >  - tty (noone dares to touch it?)
> 
> I dare, but my hands are full :(

Ditto :-)

Actually the network code is probably much harder ...

	David

--

-- 
David Laight: david <at> l8s.co.uk
Permalink | Reply | 2 comments |
headers
Erik Fair | 28 Apr 2010 17:54
Picon

Re: state of NetBSD SMP


On Apr 28, 2010, at 05:43 , David Laight wrote:

> On Tue, Apr 20, 2010 at 03:47:03PM +0000, David Holland wrote:
>> On Mon, Apr 19, 2010 at 08:17:27PM +0200, Martin Husemann wrote:
>>>> So, what's the current state of NetBSD SMP, i.e. how many kernel  
>>>> subsystems do we still have to get out from under the big lock?
>>> 
>>> No idea about total count, but top three of next things to attack:
>>> 
>>> - audio (there is a branch, supposedly ~50% done)
>>> - networking
>>> - tty (noone dares to touch it?)
>> 
>> I dare, but my hands are full :(
> 
> Ditto :-)
> 
> Actually the network code is probably much harder ...

But, I'd argue, much more urgently needed than the TTY code. How many bits get moved, as a percentage of
system total, through each subsystem?

	Erik <fair <at> netbsd.org>
Permalink | Reply | 1 comment |
headers
David Holland | 28 Apr 2010 22:21
Picon

Re: state of NetBSD SMP

On Wed, Apr 28, 2010 at 08:54:21AM -0700, Erik Fair wrote:
 > > Actually the network code is probably much harder ...
 > 
 > But, I'd argue, much more urgently needed than the TTY code. How
 > many bits get moved, as a percentage of system total, through each
 > subsystem?

indeed on both points, but I have no experience at all with the
network code and so I'm probably not doing it regardless...

--

-- 
David A. Holland
dholland <at> netbsd.org
Permalink | Reply | 0 comments |

Gmane