Re: n th's try to ask for help / non-X86 arch
Marc-Christian Petersen <[email protected]> Tue, 9 Mar 2004 19:24:20 +0100
| Newsgroups | gmane.linux.wolk.devel |
|---|---|
| Organization | Working Overloaded Linux Kernel |
| Message-ID | <200403091924.20575@WOLK> |
On Monday 08 March 2004 15:55, Kerin Millar wrote: Hi Kerin, > You're asking for the right things, but perhaps it would help to broaden > the channels a little. I'm daresay no-one who actually follows this list or sure. How? ... subscribing me to gentoo-server@? ;-) > Now I know you're a Debian guy, but I was thinking maybe some sort of > assistance could be sourced from the Gentoo project also. By it's very > nature, it is a good developer's platform and attracts some interesting > people. As Portage is a highly abitious project (it's scope is not limited > to even GNU/Linux in the longer term) I believe there is a rich (albeit > understaffed) array of talent on offer. There are specific sub-projects for > sparc, ppc and so forth. Furthermore, your patchset does have several > admirers in the community at large so ... who knows? nice to know. :-) > Now, I'm not an official developer but I do have a few connections and > would be quite happy to run it by the gentoo-server crowd in the first > instance - if you wish. /me picks up his cheerleading pom-poms again ... LOL. Yeah. I am very thankful you did that for 2.4-WOLK :-) > That reminds me, I need to get your 2.6 sources into Portage so I'll see if > I can chase nerdboy up with regard to committing them (the 2.4 stuff is > fully up-to-date). Did you talk to him already? > Also, don't take this amiss - but your record on commitment to the branch > has made some people wary in the sense that they might be worried that the > project will simply die without word or warning, or that you'll just I knew that it'll came so and I also know it atm. But last year, I was really disappointed. Maybe it was a subliminal try to see if there are people using WOLK and depend on future development ;> ... Dunno. Sorry for that. > I know you don't owe anyone anthing, but the long-term silence and period > of inactivity in terms of 2.4 WOLK development did upset a few people. Even Well, I owe anyone at least to fix critical security bugs if they are using WOLK in production and have choosen WOLK for their linux kernel. That's the same like companies are buying Redhat or SuSE, using it in production and then RH || SuSE decided to be lazy w/o offering bugfixes so people have to either live with that situation or to switch over to a Distribution which offers bugfixes ASAP. I know that and it has been my motto all the time doing WOLK but dunno what has happened to me in ending of the last year ... Maybe it was just that I've been disappointed alot. > I had resigned myself to the idea that the project was dead. If I may say > so, the best thing is just to let people know what's going on, even if you Yes, again you are right. Do you remember the time when I said bad things about Marcelo not doing it? - I did the same. Kinda painful for me :( > do find that maintaining it is too much hard work one day. Believe me, if > that assurance can be made I think you'll find it a lot easier to get help > and find a few developers with whom you can form a mutually beneficial and > constructive working rela tionship. I wish Michael Gasperi would come back. He did some nice and helpful work during WOLK3.x time. > One more thing. I have had a few off-the-wall chats with a couple of people > (mostly Gentoo devs) about your 2.6 branch. You know you're on to something > when even Livewire doesn't think that they suxx0rs ;) In any case, The vast hehehe ;) > majority of them are very excited by it, but they almost unanimously wish > that it wasn't based on -mm. I certainly am not comfortable with the idea > of running -mm (or any related kernel) on a production server. Having said well. I am currently in the stage of checking out every patch from the -mm tree and rip out everything which is b0rked or not-the-way-to-do-it etc. I have to finish it up and then rip out everything that fails my tests. I already did that with 4g/4g patch from -mm and I did it again with remap_file_pages() from 2.6.4-rc2-mm1 which almost breaks every architecture build. But OTOH -mm tree has very interesting fixes and speedups I'd merge by myself too but why should I re-invent the wheel ?! :) Please let me know what people are complaining about in -mm and let's discuss this so I know about objections and prolly rip that "bad" stuff out of my tree. > that, 2.6 is beginning to mature ... as evidenced by the constantly > reducing size of the -mm patchset. Some patches in -mm _are_ very > interesting (lazy-read-ahead, vm tuneups and so forth). But many of them > seem quite experimental, and some of it is just stuff which would only be yes, -bk snapshots of almost every tree (acpi, input, ieee1394) and such. -mm tree is just a test-base for the wide so bugs can be fixed asap. Almost everything goes into Linus tree after it has been in -mm for a while. > of interest to hardcore kernel developers (for instance, one of them is a > simple debugging hook to profile a function because they suspect it isn't > used anymore - do we really need that sort of thing?) At the very least, Depends. Personally I use it sometimes, but for the masses, it is not needed, yes. > would suggest that the amount of pa tches that are included from -mm is > slimmed down somewhat. I bet you that many would be much more convinced > that these could be the ultimate 2.6 sources (for server use and otherwise) > of the future! Yes, I am open for every suggestion, complaints and such about -mm and/or -wolk. I want to know what the masses want to have so I can fit their needs. > Just my 2 pence anyway ... I count this as 100 Euro's :) ... Many thanks for your mail. ciao, Marc ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ WOLK - Working Overloaded Linux Kernel [email protected] https://lists.sourceforge.net/lists/listinfo/wolk-devel