Skip to contentSkip to conversation

2.6.9-ck2

DISCUSSION Con Kolivas 3 participants 6 messages Oct 24, 2004

lore
Share & tools
lore
#0 lore raw

These are patches designed to improve system responsiveness with specific emphasis on the desktop, but configurable to any workload.

http://ck.kolivas.org/patches/2.6/2.6.9/2.6.9-ck2/patch-2.6.9-ck2.bz2

web:
http://kernel.kolivas.org
all patches:
http://ck.kolivas.org/patches/
Split patches and a server specific patch available.

Added:
+mwII-oc.diff
Linking hardmaplimit to overcommit is no longer required as hardmaplimit is not set by default, and there was a compile problem with modules. This backs it out.

+back-sched-net-fix-scheduling-latencies-in-__release_sock.patch
Something about this patch from the low latency hacks in combination with -ck causes a panic on nmap with some hardware (only on -ck). Remove it for now.

+ll-config1.diff
Change the default low latency config options to have the debug off by default.

+fix-bad-segment-coalescing-in-blk_recalc_rq_segments.patch
A bugfix for i/o scheduler which improves performance under heavy i/o load.

+vm-pages_scanned-active_list.patch
A nasty bug that caused kswapd to get stuck consuming heaps of cpu which was in mainline 2.6.9 was tracked down by some of my users (thanks!) and fixed by Nick Piggin (thanks!).

+269ck2-version.diff
Version

Removed:
-ll-config.diff
This disabled preempting the BKL which actually seems quite safe on further testing.

Full Patchlist:
2.6.9_to_staircase9.0.diff
schedrange.diff
schedbatch2.5.diff
schediso2.8.diff
mwII.diff
mwII-oc.diff
1g_lowmem1_i386.diff
cfq2-20041019.patch
block_fix.diff
defaultcfq.diff
269rc4-mingo_ll.diff back-sched-net-fix-scheduling-latencies-in-__release_sock.patch 269rc4-mingo-bkl.diff
ll-config1.diff
cddvd-cmdfilter-drop.patch
nvidia_compat.diff fix-bad-segment-coalescing-in-blk_recalc_rq_segments.patch vm-pages_scanned-active_list.patch 269ck2-version.diff

Cheers,
Con

#2 lore raw
In reply to Con Kolivas
Quoted 2 lines · Con Kolivas

These are patches designed to improve system responsiveness with specific emphasis on the desktop, but configurable to any workload.

Hello Con,

I have a question regarding supermount patch. Is it applied into your patch set? I tried to find it in 2.6.9-ck1/ck2, and... I think it isn't.

Thanks for reply.

Greets
rafal

#3 lore raw
Quoted 2 lines · Rafal Bujnowski

I have a question regarding supermount patch. Is it applied into your patch set? I tried to find it in 2.6.9-ck1/ck2, and... I think it isn't.

Why not just use ivman?

You know why supermount is not in the mainline kernel.

Stefan Schweizer

#4 lore raw
Quoted 1 line · Stefan Schweizer

Why not just use ivman?

Thanks, I'll try this one.

rafal

#5 lore raw
In reply to Rafal Bujnowski
Quoted 11 lines · Rafal Bujnowski
Quoted 2 lines · Con Kolivas

These are patches designed to improve system responsiveness with specific emphasis on the desktop, but configurable to any workload.

Hello Con,

I have a question regarding supermount patch. Is it applied into your patch set? I tried to find it in 2.6.9-ck1/ck2, and... I think it isn't.

As I explained previously: I will continue to maintain the supermount patch due to popular request. However since it adds no performance feature to -ck and makes the patchset more complicated making debugging more difficult I will not include it by default. Having said that, I haven't had any bug reports about it in a long time. It's a simple matter to add it to -ck if you want it.

Cheers,
Con