2.6.10-ck4

2.6.10-ck4

[Con Kolivas]

[-- Attachment #1: Type: text/plain, Size: 1197 bytes --]

These are patches designed to improve system responsiveness. It is 
configurable to any workload but the default ck4 patch is aimed at the 
desktop and ck4-server is available with more emphasis on serverspace.

http://ck.kolivas.org/patches/2.6/2.6.10/2.6.10-ck4/

web:
http://kernel.kolivas.org
all patches:
http://ck.kolivas.org/patches/

I recommend all ck3 users upgrade.

2.6.10-ck3 was a brown paper bag release. A poorly considered last 
minute change made for some odd starvation problems. For this release I 
rewrote a large section of the staircase code that had been troubling me 
and been getting steadily worse. In the process I've made the semantics 
of resuming an old timeslice much simpler and more predictable.


Changed:
-cfq-ts-19g.diff
+cfq-ts-20.diff
Jens' latest incarnation of the cfq-timeslices patch with i/o priority 
support for read and write has much smoother read vs write 
characteristics now.


Added:
+s10_test1.diff
+s10_s10.1.diff
+s10.1_s10.2.diff
+s10.2_s10.3.diff
Staircase updates

+1504_vmscan-writeback-pages.patch
A fix for more oom-kill problems.


Thanks to the many people involved in testing the staircase changes and 
reporting back.

Cheers,
Con


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 256 bytes --]

Re: 2.6.10-ck4

[Alexey Dobriyan]

Problems rebooting here. Don't have them in 2.6.10 and 2.6.11-rc1.

If you want complete messages, just ask. I will write them down into 
electronic form.

First time there was an unstoppable infinite loop of quickly moving USB 
related messages. I only saw "[uhci_hcd]" string reliably. Can't reproduce.

Third time:

[<c0113160>] do_page_fault+0x0/0x562
[<c010398f>] error_code+0x2b/0x30
[<c0113160>] do_page_fault+0x0/0x562
[<c01133fe>] do_page_fault+0x97/0x562
	[repeat several times to fit console]
Code: [snip]
 <0> Kernel panic - not syncing: Fatal exception in interrupt.

Second time:

[<c0113160>] do_page_fault+0x0/0x562
[<c0113160>] do_page_fault+0x0/0x562
[<c01133fe>] do_page_fault+0x29e/0x562
	[many tty related functions snipped]
[<c0113160>] do_page_fault+0x0/0x562
[<c0113160>] do_page_fault+0x0/0x562
[<c01133fe>] do_page_fault+0x29e/0x562
Code: [snip]
 <0> Kernel panic - not syncing: Fatal exception in interrupt.

Other than that everything seems to be ok. 

	Alexey

Re: 2.6.10-ck4

[Christian Axelsson]

Are there any plans on putting back reiser4 in -ck?
Im looking for a patchet that has reiser4 and that is more friendly 
towards ati-drivers and other 3rd-party modules (no massive api-changes).

-- 
Regards,
Christian

Re: 2.6.10-ck4

[Con Kolivas]

[-- Attachment #1: Type: text/plain, Size: 407 bytes --]

Christian Axelsson wrote:
> Are there any plans on putting back reiser4 in -ck?
> Im looking for a patchet that has reiser4 and that is more friendly 
> towards ati-drivers and other 3rd-party modules (no massive api-changes).

The -cko and -nitro patches use -ck as their base and add reiser4. Wait 
till they sync up with the latest ck and use those (the url is available 
from my web page).

Cheers,
Con

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 256 bytes --]

Re: 2.6.10-ck4

[Con Kolivas]

[-- Attachment #1: Type: text/plain, Size: 739 bytes --]

Alexey Dobriyan wrote:
> Problems rebooting here. Don't have them in 2.6.10 and 2.6.11-rc1.
> 
> If you want complete messages, just ask. I will write them down into 
> electronic form.
> 
> First time there was an unstoppable infinite loop of quickly moving USB 
> related messages. I only saw "[uhci_hcd]" string reliably. Can't reproduce.
> 
> Third time:

>  <0> Kernel panic - not syncing: Fatal exception in interrupt.
> 
> Second time:
> 
> [<c0113160>] do_page_fault+0x0/0x562

>  <0> Kernel panic - not syncing: Fatal exception in interrupt.

This is funny:

> Other than that everything seems to be ok. 

Is that like saying "the operation was a success but the patient died" ?

Sorry I have no idea what those are.

Cheers,
Con

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 256 bytes --]