[Cerowrt-devel] closer ties to ltsi?

[Cerowrt-devel] closer ties to ltsi?

[Dave Taht]

Dear Greg:

What ltsi is up to:

http://ltsi.linuxfoundation.org/what-is-ltsi

is orthogonal to what we are up to, in that they are trying to produce
a usable, stable tree that can be used in a lot of CE devices, and we
are trying to push the bleeding edge in finding, fixing, and testing
things that would be good to have deployed in a lot of CE devices at
least on the bufferbloat, security, and network - especially ipv6 -
fronts.

How to faster get stuff from "R&D" to deployment is always on my mind.
I'm very happy to be successfully tracking the 3.3 process 'live' at
the moment with cerowrt. I imagine the next 'stable'
kernel chosen will be 3.3 or 3.4 based.

How to  make bloatlab #1 more effective and useful is currently on my
mind, as I'm plan to rebuild it when I get back to california next
week. While there's a great deal of gear under test, we could do
better, and certainly having some automation and a good test suite is
something I wanted to have in operation by now, and don't, as BQL +
SFQRED etc really need some serious exercising.

ltsi seems to have a lot of hw vendor support, we have a growing
amount of ISP interest...

-- 
Dave Täht
SKYPE: davetaht
US Tel: 1-239-829-5608
http://www.bufferbloat.net

Re: [Cerowrt-devel] closer ties to ltsi?

[Greg KH]

On Fri, Mar 02, 2012 at 02:33:25AM -0800, Dave Taht wrote:
> Dear Greg:
> 
> What ltsi is up to:
> 
> http://ltsi.linuxfoundation.org/what-is-ltsi
> 
> is orthogonal to what we are up to, in that they are trying to produce
> a usable, stable tree that can be used in a lot of CE devices, and we
> are trying to push the bleeding edge in finding, fixing, and testing
> things that would be good to have deployed in a lot of CE devices at
> least on the bufferbloat, security, and network - especially ipv6 -
> fronts.
> 
> How to faster get stuff from "R&D" to deployment is always on my mind.
> I'm very happy to be successfully tracking the 3.3 process 'live' at
> the moment with cerowrt. I imagine the next 'stable'
> kernel chosen will be 3.3 or 3.4 based.

If you do all of your work upstream in the main kernel.org kernel, then
LTSI can backport your changes as needed to their kernels.  I recommend
doing that.

best of luck,

greg k-h

Re: [Cerowrt-devel] closer ties to ltsi?

[Jim Gettys]

On 03/02/2012 10:43 AM, Greg KH wrote:
> On Fri, Mar 02, 2012 at 02:33:25AM -0800, Dave Taht wrote:
>> Dear Greg:
>>
>> What ltsi is up to:
>>
>> http://ltsi.linuxfoundation.org/what-is-ltsi
>>
>> is orthogonal to what we are up to, in that they are trying to produce
>> a usable, stable tree that can be used in a lot of CE devices, and we
>> are trying to push the bleeding edge in finding, fixing, and testing
>> things that would be good to have deployed in a lot of CE devices at
>> least on the bufferbloat, security, and network - especially ipv6 -
>> fronts.
>>
>> How to faster get stuff from "R&D" to deployment is always on my mind.
>> I'm very happy to be successfully tracking the 3.3 process 'live' at
>> the moment with cerowrt. I imagine the next 'stable'
>> kernel chosen will be 3.3 or 3.4 based.
> If you do all of your work upstream in the main kernel.org kernel, then
> LTSI can backport your changes as needed to their kernels.  I recommend
> doing that.
>
>
That's what Dave has been doing, to the extent possible; it's a bit of a
PITA, since the MIPS/OpenWrt folks tend not to quite keep up with x86. 
And BQL (in 3.3) is essential to bufferbloat work.
                    - Jim