On Thu, 2005-09-15 at 21:54 +1000, ibm-main wrote: > ----- Original Message ----- > From: "MAEDA Naoaki" > > > Hi Chandra, > > > > I believe that simplifying CKRM is the right thing to do. > > At first glance, I have some questions and comments. > > Yeah, me too. > Humour me a little here folks - what the hell's going on ???.
Hi Shane, Thanks for the great question. > > Are we forking or what ???. No, we are not forking. Here is the scoop... Concerns the lkml community has with CKRM are: - Code size is too big - over engineered - multiple hooks all over the kernel - complex to understand - not using existing features So, we started looking at different ways to attack the above mentioned concerns. The first step of that was the moving rbce to userland. With connectors it became feasible to move that functionality to userland and thereby removing most of the hooks, reducing complexity and removal of lot of code in the kernel. The second step was to remove the classtype level of indirection which adds indirection and complexity. The next is to move the filesystem interface to configfs instead of using our own filesystem(rcfs). I will start a different thread for an another item that might concern a few to see if it is worth it pursuing. At the end of all of this CKRM code would be at least 50% (or more) smaller and less complex. > > I built a 2.6.13.system (at work) when I saw Chandras posts - clean start, > sounds good. > Unfortunately, I had to pick the patches up from the (web-interface) > archive, and it appears there must be a limit on the number of lines. Hence > the patches were incomplete - I had to junk that build. > O.K., that happens. Will roll up the f series and post it at the project website. > > Then I get home to a truck-load of patches from Gerrit. > So tell me - what the hell am I supposed to apply to what ???. > I *really* would like to be involved here, but I'm new, and I just can't > work out where you're all heading. > > - do I go with Chandra on a 2.6.13 build, and look at that as the future > direction > - do I patch (some/any of ???) Gerrits updates on top of that > - do I add (some of) Gerrits patches to the (incomplete) e18 I am currently > fighting with. > - do I give up on this altogether, and go look at LKCD as something to > occupy my idle hours ???. NOOOOOOOOOOOOO.... :) Here is what i suggest: If you looking for a stable tree with all the controller support use e19 and then e20. If you want to look into future and building features, use the f series, which is currently the development tree and will become production tree when stable. Do note that many of us using/reviewing the f series would help getting it stable earlier. > > Shane ... > > > > ------------------------------------------------------- > SF.Net email is sponsored by: > Tame your development challenges with Apache's Geronimo App Server. Download > it for free - -and be entered to win a 42" plasma tv or your very own > Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php > _______________________________________________ > ckrm-tech mailing list > https://lists.sourceforge.net/lists/listinfo/ckrm-tech > -- ---------------------------------------------------------------------- Chandra Seetharaman | Be careful what you choose.... - [EMAIL PROTECTED] | .......you may get it. ---------------------------------------------------------------------- ------------------------------------------------------- SF.Net email is sponsored by: Tame your development challenges with Apache's Geronimo App Server. Download it for free - -and be entered to win a 42" plasma tv or your very own Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech
