2007-08-13 05:09:07

by Al Boldi

[permalink] [raw]
Subject: How to make -mm a more viable testbed (was: [PATCH] [1/12] x86: Work around mmio config space quirk on AMD)

Linus Torvalds wrote:
> On Sun, 12 Aug 2007, Dave Jones wrote:
> >
> > This does make me wonder, why these weren't caught in -mm ?
>
> I'm worried that -mm isn't getting a lot of exposure these days. People do
> run it, but I wonder how many..

Well, I'm not running it because it's got too much garbage in it, which makes
it a pretty unrealistic testbed. And to make matters worse, I also stay
away from rc's as a consequence.

To make -mm more viable, it may be advisable to restructure -mm in such a way
as to be a Kconfig option to mainline. This would probably involve some
patch management functionality to apply experimental submissions on a
per-patch basis, as opposed to being a 'take it or leave it slam onto
mainline' patch.


Thanks!

--
Al


2007-08-13 20:42:22

by Zan Lynx

[permalink] [raw]
Subject: Re: How to make -mm a more viable testbed (was: [PATCH] [1/12] x86: Work around mmio config space quirk on AMD)

On Mon, 2007-08-13 at 08:08 +0300, Al Boldi wrote:
> Linus Torvalds wrote:
> > On Sun, 12 Aug 2007, Dave Jones wrote:
> > >
> > > This does make me wonder, why these weren't caught in -mm ?
> >
> > I'm worried that -mm isn't getting a lot of exposure these days. People do
> > run it, but I wonder how many..
>
> Well, I'm not running it because it's got too much garbage in it, which makes
> it a pretty unrealistic testbed. And to make matters worse, I also stay
> away from rc's as a consequence.
>
> To make -mm more viable, it may be advisable to restructure -mm in such a way
> as to be a Kconfig option to mainline. This would probably involve some
> patch management functionality to apply experimental submissions on a
> per-patch basis, as opposed to being a 'take it or leave it slam onto
> mainline' patch.

I thought most people running -mm were running klive, which shou
ld tell kernel versions, uptime and other things. I run it, anyway.

http://klive.cpushare.com/
--
Zan Lynx <[email protected]>


Attachments:
signature.asc (189.00 B)
This is a digitally signed message part

2007-08-13 21:04:06

by Valdis Klētnieks

[permalink] [raw]
Subject: Re: How to make -mm a more viable testbed (was: [PATCH] [1/12] x86: Work around mmio config space quirk on AMD)

On Mon, 13 Aug 2007 14:40:06 MDT, Zan Lynx said:

> I thought most people running -mm were running klive, which shou
> ld tell kernel versions, uptime and other things. I run it, anyway.
>
> http://klive.cpushare.com/

That page says there's a whole whopping 587 people reporting back, of which
1 is on 23-rc1-mm1, and 3 are on 23-rc2-mm2.

I know I'm on 23-rc2-mm1 at the moment (backed off from 23-rc2-mm2 due to
the sysctl breakage in iptables). I certainly *hope* there's more than
5 people trying the -mm kernels, and I'm guessing that in fact, most -mm users
are *not* running klive.


Attachments:
(No filename) (226.00 B)

2007-08-13 22:35:33

by Dave Jones

[permalink] [raw]
Subject: Re: How to make -mm a more viable testbed (was: [PATCH] [1/12] x86: Work around mmio config space quirk on AMD)

On Mon, Aug 13, 2007 at 02:40:06PM -0600, Zan Lynx wrote:
> On Mon, 2007-08-13 at 08:08 +0300, Al Boldi wrote:
> > Linus Torvalds wrote:
> > > On Sun, 12 Aug 2007, Dave Jones wrote:
> > > >
> > > > This does make me wonder, why these weren't caught in -mm ?
> > >
> > > I'm worried that -mm isn't getting a lot of exposure these days. People do
> > > run it, but I wonder how many..
> >
> > Well, I'm not running it because it's got too much garbage in it, which makes
> > it a pretty unrealistic testbed. And to make matters worse, I also stay
> > away from rc's as a consequence.
> >
> > To make -mm more viable, it may be advisable to restructure -mm in such a way
> > as to be a Kconfig option to mainline. This would probably involve some
> > patch management functionality to apply experimental submissions on a
> > per-patch basis, as opposed to being a 'take it or leave it slam onto
> > mainline' patch.
>
> I thought most people running -mm were running klive, which shou
> ld tell kernel versions, uptime and other things. I run it, anyway.
>
> http://klive.cpushare.com/

If that's the case, there are 3 people running 2.6.23-rc2-mm2.
I suspect the actual number of users is considerably higher.
Somehow I doubt that 'most' people are running klive.

Dave

--
http://www.codemonkey.org.uk

2007-08-13 23:07:50

by Zan Lynx

[permalink] [raw]
Subject: Re: How to make -mm a more viable testbed (was: [PATCH] [1/12] x86: Work around mmio config space quirk on AMD)

On Mon, 2007-08-13 at 17:09 -0400, Dave Jones wrote:
> On Mon, Aug 13, 2007 at 02:40:06PM -0600, Zan Lynx wrote:
> > I thought most people running -mm were running klive, which shou
> > ld tell kernel versions, uptime and other things. I run it, anyway.
> >
> > http://klive.cpushare.com/
>
> If that's the case, there are 3 people running 2.6.23-rc2-mm2.
> I suspect the actual number of users is considerably higher.
> Somehow I doubt that 'most' people are running klive.

I had thought Andrew asked people to do it at one point. Could have
been someone else, I suppose. Ah, it was Andrea Arcangeli.

At the time, it seemed more people were doing it. I guess not though.
Or there really aren't many people running the latest -mm. It *is*
quite a hassle. I only do it because I'm crazy. :)

I think klive is a good idea for tracking kernel testing. Perhaps
mentioning it here again will get more people running it.

I suppose that someone could try to estimate usage by counting patch
downloads off kernel.org.
--
Zan Lynx <[email protected]>


Attachments:
signature.asc (189.00 B)
This is a digitally signed message part

2007-08-13 23:10:10

by Dave Jones

[permalink] [raw]
Subject: Re: How to make -mm a more viable testbed (was: [PATCH] [1/12] x86: Work around mmio config space quirk on AMD)

On Mon, Aug 13, 2007 at 05:07:23PM -0600, Zan Lynx wrote:
> On Mon, 2007-08-13 at 17:09 -0400, Dave Jones wrote:
> > On Mon, Aug 13, 2007 at 02:40:06PM -0600, Zan Lynx wrote:
> > > I thought most people running -mm were running klive, which shou
> > > ld tell kernel versions, uptime and other things. I run it, anyway.
> > >
> > > http://klive.cpushare.com/
> >
> > If that's the case, there are 3 people running 2.6.23-rc2-mm2.
> > I suspect the actual number of users is considerably higher.
> > Somehow I doubt that 'most' people are running klive.
>
> I had thought Andrew asked people to do it at one point. Could have
> been someone else, I suppose. Ah, it was Andrea Arcangeli.
>
> At the time, it seemed more people were doing it. I guess not though.
> Or there really aren't many people running the latest -mm. It *is*
> quite a hassle. I only do it because I'm crazy. :)
>
> I think klive is a good idea for tracking kernel testing. Perhaps
> mentioning it here again will get more people running it.
>
> I suppose that someone could try to estimate usage by counting patch
> downloads off kernel.org.

The fallout from bugs found each time Andrew makes an announcement
should be a fairly good indicator :-)

Dave

--
http://www.codemonkey.org.uk

2007-08-14 09:46:26

by Tilman Schmidt

[permalink] [raw]
Subject: Re: How to make -mm a more viable testbed

Zan Lynx schrieb:
> I thought most people running -mm were running klive, which shou
> ld tell kernel versions, uptime and other things. I run it, anyway.
>
> http://klive.cpushare.com/

I don't, even though I try to at least build and boot each -mm release
once. Should I? Perhaps I should.

(downloading ... running ... failing ... installing python-twisted ... running ...)

Ok, now I do.

HTH
T.

--
Tilman Schmidt E-Mail: [email protected]
Bonn, Germany
Diese Nachricht besteht zu 100% aus wiederverwerteten Bits.
Unge?ffnet mindestens haltbar bis: (siehe R?ckseite)


Attachments:
signature.asc (250.00 B)
OpenPGP digital signature