Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 569262 - sys-kernel/ck-sources version bump request, with -ck3 patchset
Summary: sys-kernel/ck-sources version bump request, with -ck3 patchset
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: All Linux
: Normal normal (vote)
Assignee: kuzetsa CatSwarm (kuza for short)
URL:
Whiteboard:
Keywords:
Depends on: 585552
Blocks:
  Show dependency tree
 
Reported: 2015-12-23 02:39 UTC by thunderrd
Modified: 2017-01-20 08:51 UTC (History)
9 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments
4.9-ck1 ebuild (ck-sources-4.9.4.ebuild,1.84 KB, text/plain)
2017-01-19 03:15 UTC, kuzetsa CatSwarm (kuza for short)
Details
Minor regression fix (ck-sources-4.9.4.ebuild,1.89 KB, text/plain)
2017-01-19 22:03 UTC, kuzetsa CatSwarm (kuza for short)
Details
Cleaned up version of the ebuild (ck-sources-4.9.4.ebuild,2.22 KB, text/plain)
2017-01-20 02:57 UTC, Göktürk Yüksek
Details

Note You need to log in before you can comment on or make changes to this bug.
Description thunderrd 2015-12-23 02:39:09 UTC
New patches have been made available for kernel 4.3: 

http://ck.kolivas.org/patches/4.0/4.3/4.3-ck3/

I have built a local overlay version, and it seems stable, as have others in this thread:

https://forums.gentoo.org/viewtopic-t-941030-start-225.html

It might be nice to get this into the tree, as Con has stated that it contains some new performance enhancements.

The maintainer yngwin is on leave ATM, so perhaps someone else could pick it up temporarily?  https://www.gentoo.org/inside-gentoo/developers/unavailable-developers.html

Thanks

Reproducible: Always
Comment 1 Laurent V 2016-01-21 17:52:15 UTC
(In reply to thunderrd from comment #0)
> New patches have been made available for kernel 4.3: 
> 
> http://ck.kolivas.org/patches/4.0/4.3/4.3-ck3/
> 
> I have built a local overlay version, and it seems stable, as have others in
> this thread:
> 
> https://forums.gentoo.org/viewtopic-t-941030-start-225.html
> 
> It might be nice to get this into the tree, as Con has stated that it
> contains some new performance enhancements.
> 
> The maintainer yngwin is on leave ATM, so perhaps someone else could pick it
> up temporarily? 
> https://www.gentoo.org/inside-gentoo/developers/unavailable-developers.html
> 
> Thanks
> 
> Reproducible: Always

Hello dears, 

I made an ebuild for 4.3.3-r1-ck (in order to get a ck ebuild which fix CVE-2016-0728 )

Check https://forums.gentoo.org/viewtopic-t-941030-postdays-0-postorder-asc-start-250.html if you want to test it.

On a side note, yngwin, I'm interested to proxy-maintain ck-sources. Would be a great learning experience for me.

Kind regards, 

Laurent
Comment 2 Pacho Ramos gentoo-dev 2016-05-28 10:39:59 UTC
we have 4.3.6 in the tree
Comment 3 Cory Bolar 2016-06-09 01:26:20 UTC
Hi,

I'm looking to be the maintainer of this package.  I've been using the software for about a year now and it is well behind current kernel versions.  I've updated most of the ebuilds locally and will look into the remaining bug listed here tonight/tomorrow.  Thanks

Cory Bolar
cory.bolar@gmail.com
Comment 4 Jonas Stein gentoo-dev 2016-06-09 07:31:32 UTC
Cory, that is nice to read.
Please look at https://wiki.gentoo.org/wiki/Project:Proxy_Maintainers for the next steps.
Comment 5 Florian Berger 2017-01-10 14:51:40 UTC
A version bump would be nice. Current version in the tree is ck-sources-4.3.6.ebuild , while upstream is apparently at 4.9 http://ck.kolivas.org/patches/4.0/ .
Comment 6 thunderrd 2017-01-10 15:06:23 UTC
FWIW, the hnaparst overlay is keeping current since 4.8.x, with Con's new incarnation of BFS called MuQSS.  He updated to a ck-sources-4.9.2.ebuild just today.

https://github.com/hnaparst/overlay
Comment 7 Florian Berger 2017-01-11 06:36:48 UTC
(In reply to thunderrd from comment #6)
> FWIW, the hnaparst overlay is keeping current since 4.8.x, with Con's new
> incarnation of BFS called MuQSS.

Thanks, that works for me.
Comment 8 Pacho Ramos gentoo-dev 2017-01-16 09:55:51 UTC
It seems that finally this became orphan again as I read in:
https://github.com/gentoo/gentoo/pull/1643

CCing treecleaners as this cannot stand in this situation without any maintainer
Comment 9 kuzetsa CatSwarm (kuza for short) 2017-01-19 03:11:52 UTC
I'm volunteering to maintain this package.
Comment 10 kuzetsa CatSwarm (kuza for short) 2017-01-19 03:15:27 UTC
Created attachment 460584 [details]
4.9-ck1 ebuild

Initial template / ebuild (attached)

^ tested locally, no obvious issues
Comment 11 kuzetsa CatSwarm (kuza for short) 2017-01-19 22:03:23 UTC
Created attachment 460668 [details]
Minor regression fix

Change #1 (still present in the updated ebuild)

[quote] - "Since experimental genpatches && we want BFQ irrespective of experimental"

^ Consistent with this particular comment from the ebuild, I had made a decision to switch to 5010_*.patch* (intead of a more-broad pattern 50*_*.patch* )

... it still pulls in the appropriate "experimental" genpatch to allow BFQ to be a usable / functional scheduler choice in the resultant patched ck-sources kernel source tree; once it gets to /usr/src/ it seems to compile and even run... at least with my chosen config settings.

Change #2 (regression fix)

The [un]removal of K_SECURITY_UNSUPPORTED // K_DEBLOB_AVAILABLE - as per disussion on freenode IRC, those ought to have been considered a regressions issue (for the sake of compatibility)

I'm about 96% certain adding back in these particular changes won't do anything different in my configuration, but I've rebuilt the kernel anyway just in case it unexpectedly changes something subtle [but important] in the kernel source tree:

K_SECURITY_UNSUPPORTED="1"
K_DEBLOB_AVAILABLE="1"
Comment 12 Göktürk Yüksek archtester gentoo-dev 2017-01-20 02:57:42 UTC
Created attachment 460678 [details]
Cleaned up version of the ebuild

I'd like to make a few clarifications here:

- Beginning with 4.8, CK sources unconditionally include the BFQ: http://ck.kolivas.org/patches/4.0/4.8/4.8-ck8/patches/ The experimental genpatches 5000_* series also add support for BFQ. This introduces a collision during patching, so we mask them using 5010_* (although it's a rather poor way, since we also exclude any "potential" 5020_* etc.).

- Regarding the question about K_EXP_GENPATCHES_NOUSE, gentoo-sources pull in the experimental genpatches when USE=experimental is provided. Setting this to 1 unconditionally pulls in the patchset.

- Regarding K_EXP_GENPATCHES_PULL, we need this so that we can selectively apply the experimental patches. We do this by using the variable K_EXP_GENPATCHES_LIST="5010*_*.patch*".

- On K_DEBLOB_AVAILABLE="1", deblob is the process of removing binary-only objects from the tree if you are not familiar with it. You can read about it more here: http://www.fsfla.org/svn/fsfla/software/linux-libre/scripts/README.

I'll spin up a VM to test a build here. I think we are mostly good. I've created a cleaner version of the ebuild using yours as the basis. That's what I will be committing, let me know if you have any concerns about it.
Comment 13 kuzetsa CatSwarm (kuza for short) 2017-01-20 03:09:37 UTC
Your new version looks like it will probably work the same as mine.

Though I don't see any reason for two of the changes:

-# Copyright 1999-2017 Gentoo Foundation
+# Copyright 1999-2016 Gentoo Foundation

-K_EXP_GENPATCHES_LIST="5010*_*.patch*"
+K_EXP_GENPATCHES_LIST="5010_*.patch*"

Shouldn't hurt anything though.

Other than that, it looks mostly like rephrasing a comment about genpatches, and moving things (but not changing) for some layout reason, and re-inclusion of some unused="" (blank) variables which aren't doing anything.

Since 100% of these the changes appear to be trivial, or for cosmetic reasons, and/or for legacy (looking like previous versions) layout purposes, I have no reason to believe this version of the ebuild will produce a different unpacked+patched kernel source tree.
Comment 14 kuzetsa CatSwarm (kuza for short) 2017-01-20 03:17:57 UTC
Now that I think about it - I'm having a funny behavior with BFQ (CONFIG_DEFAULT_BFQ=y seems to be a placebo) for my version of the patch. I was investigating this issue today, so I probably should have a closer look to compare the differences:

> K_EXP_GENPATCHES_LIST="50_*.patch*"

> K_EXP_GENPATCHES_LIST="5010_*.patch*"

> K_EXP_GENPATCHES_LIST="5010*_*.patch*"
Comment 15 thunderrd 2017-01-20 03:38:49 UTC
Sarah, is the 'funny' behavior you are speaking of what you mentioned on IRC earlier about BFQ not becoming active on your drives even though it was selected in the .config?  I saw your comments, but had no time to respond.

Are you running SSDs?  I do, and I had to write a small script to start BFQ on my SSDs, because it started automatically only on the mechanical drives.
Comment 16 Göktürk Yüksek archtester gentoo-dev 2017-01-20 04:57:36 UTC
commit b4a883ed61410fcdb4d846e7ec604fa8e27ef1e5
Author:     Sarah White <kuzetsa@gmail.com>
AuthorDate: Thu Jan 19 23:25:09 2017 -0500
Commit:     Göktürk Yüksek <gokturk@gentoo.org>
CommitDate: Thu Jan 19 23:55:24 2017 -0500

    sys-kernel/ck-sources: bump to 9.4.9 #569262
    
    Package-Manager: portage-2.3.0
Comment 17 kuzetsa CatSwarm (kuza for short) 2017-01-20 08:00:07 UTC
As per #15 / @thunderrd

Yes, I ran into the default behavior of SSD (set noop as IO scheduler) and assumed it was because of a misconfiguration. Embarrassing that I don't have any spinning rust to verify this only happens with SSD.

The version bumps should be taken care of, and I'll keep an eye on any bugs for sys-kernel/ck-sources so this particular bug #569262 is RESOLVED / FIXED.
Comment 18 thunderrd 2017-01-20 08:51:18 UTC
Good job.  I'll help however I can to keep this going.