Hi Barry,
On Sat, Feb 04, 2006 at 12:54:15PM -0800, Barry K. Nathan wrote:
> On 2/4/06, Willy TARREAU <[email protected]> wrote:
> > The last O(1) patch I've seen does not apply to kernels more recent than 2.4.19
> > (it's on Ingo's site). Do you have any up to date pointer that I could use ?
> > However, I have a local rediff of the lowlat patch that I will include.
>
> (This e-mail ended up being a bit longer than I intended. The most
> relevant stuff may be in the last couple of paragraphs, but I've
> included the whole message for the sake of completeness.)
>
> Red Hat ships it in a 2.4.21-based kernel. Here's their latest source RPM:
> ftp://ftp.redhat.com/pub/redhat/linux/updates/enterprise/3AS/en/os/SRPMS/kernel-2.4.21-37.0.1.EL.src.rpm
>
> However, I just remembered that the RHEL 3 kernel patch series starts
> with an -ac patch and builds up from there. I think it gets the O(1)
> scheduler from there but it might apply further patches to it (but I'm
> not 100% sure my memory is correct here).
>
> Here's the last -ac patch:
> http://kernel.org/pub/linux/kernel/people/alan/linux-2.4/2.4.22/patch-2.4.22-ac4.bz2
That's what I found too, but -ac is too much different. I tried applying
some patches designed for 2.4.21-ac and 2.4.22-ac on their non-ac respective
equivalents, and they rejected a lot of stuff.
> If you want to see what Red Hat/Fedora did against 2.4.22, this is
> what the final Fedora Core 1 kernel shipped:
> http://cvs.fedora.redhat.com/viewcvs/rpms/kernel/FC-1/
> (There are newer kernels for FC1 from Fedora Legacy, but I think those
> just add security fixes.)
>
> There's also 2.4.27-pre2-pac1, which has the O(1) scheduler. I don't
> know if it introduces any bugs into the O(1) scheduler though. (It did
> introduce a bug into the overcommit accounting, because part of it was
> missing.)
> http://kernel.org/pub/linux/kernel/people/bero/2.4/2.4.27/patch-2.4.27-pre2-pac1.bz2
>
> Finally, 2.4.31-lck1 has the O(1) scheduler. This is the "base" patch
> for 2.4.31-lck1, which has O(1) but also has "kernel preemption, low
> latency and CK interactivity":
> http://www.plumlocosoft.com/kernel/patches/2.4/2.4.31/2.4.31-lck1/components/010-lckbase.diff.bz2
>
> It's probably the most recent forward-port of the O(1) patch, and it's
> probably going to be the smallest diff to look through as well, if you
> want to cherry-pick it out and make it work on 2.4.32 or 2.4.33-pre.
I have not time to spend digging through the diff, but I can link to
it. It clearly is the most recent version amongst all those we have
seen.
> (I don't think I'll be doing this, however. The boxes I manage that
> would greatly benefit from O(1) will probably move to kernel 2.6 soon
> for other reasons anyway.)
> --
> -Barry K. Nathan <[email protected]>
Thanks for your investigation,
Willy
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
