[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ogfs-dev]Suggestion: Reorganize the source tree



Stan you probably should have made this a new post rather than a reply.

I would first off like to say I agree with you whole heartedly that this needs 
to be done.

On Wednesday 02 July 2003 02:51 am, Stanley Wang wrote:
> 	Proposal for reorganizing source tree
>
> For historical reasons the architecture of OpenGFS source tree consists
> of two layers:
>
>
> Top layer -- VFS to GFS interface
>
> Each OS (or even major OS release) gets its own directory of code.
>
> One (big) directory of common code.
>
>
> Bottom Layer -- GFS to bufcache/memory/etc...
>
> An include file and library of abstractions that GFS uses.
>
>
> Although it works well now, we could improve it according to our current
> requirement:
>
>      1. Since the only OS that OpenGFS supports now is Linux (no further
>         plan for porting work against other OS), we could combine the
>         common codes and architecture dependent codes into one directory
>         â "ogfs". And place it into the vanilla linux kernel source
>         tree.

we talked about this before, and (I think) decided it would be better to 
combine all the code like you say, and use #if's like so:
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,3,0)

>
>      2. We should own both 2.4 and 2.5/2.6 kernel tree, each of them
>         contains the corresponding ogfs codes. The 2.5/2.6 source tree
>         should be the active development tree. All changes in 2.5/2.6
>         tree should be back ported to 2.4 tree periodically. And we
>         should maintain the third source tree for all the utilities and
>         docs.

As I said before, I think we were going to try to just have one fs src tree, 
and one utils tree. OpenGFS has a lot of work to be done before it's known as 
a truly scalable and stable fs, and having do those changes to 2 different 
src trees would be a pain. Although your idea does have merit once we get the 
fs code stabilized.

>
>      3. For the sake of maintaining source tree more easily, we could
>         move from CVS to BK. Creating a project for OpenGFS @bkbits
>         (http://www.bitkeeper.com/Hosted.Creating.html) is quite easy.
>         It would also make the OpenGFS more easily to be adopted by the
>         mainstream linux kernel.

hmm, I don't know about this. I personally have a problem with the bk license 
and owner. That would make it hard to for me to continue contributing to 
OpenGFS (which arguably hasn't been much).

>
>
>
> The reorganzing would benifit both the develop and maintain work. It
> will simplify the codes and make the building work independent on the
> current running kernel. And we could make full use of all Linux's
> advanced features.
>
> Last of all, the reorganizing work is not a trivial thing.
> We need a good schedule for achieving it.

Have any ideas/rough schedule of how this would be done. Maybe you could put 
together a roadmap of sorts. Just with weeks from the start date, and then 
maybe we can plan on a start date.

>
>
> Best Regards,
> Stan

-- 
OpenGFS -- http://opengfs.sourceforge.net
Home -- http://www.brianandsara.net



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
Opengfs-devel mailing list
Opengfs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opengfs-devel


[Kernel]     [Security]     [Bugtraq]     [Photo]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Clusters]     [Linux RAID]     [Yosemite Hiking]     [Linux Resources]

Powered by Linux