cache_lookup() work this week.

Hiten Pandya hmp at backplane.com
Mon Sep 8 12:21:45 PDT 2003


81803.h88I3Fs6053054 at xxxxxxxxxxxxxxxxxxxx>
In-Reply-To: <200309081803.h88I3Fs6053054 at xxxxxxxxxxxxxxxxxxxx>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Lines: 34
NNTP-Posting-Host: 213.107.166.189
X-Trace: 1063048971 crater_reader.dragonflybsd.org 267 213.107.166.189
Xref: crater_reader.dragonflybsd.org dragonfly.kernel:1057

Matthew Dillon wrote:

> :My understanding is that you can't distribute a kernel binary with
> :GPL'd code statically linked into it without GPLing the whole kernel.
> :I think this is why EXT2FS, COMPAT_LINUX, and pcm aren't in FreeBSD's
> :GENERIC kernel.
> :
> : -jre
> 
>     Ah.  That is true, I think you are right.  I'm not sure how significant
>     it is because while linux installs tend to put everyone on one or two
>     slices, BSD installs tend to partition things up more finely so having a
>     mostly read-only UFS / and, say, an ext2fs /usr, would not be a big deal.
> 
>     At the point the gpl'd filesystems can be loaded as a KLD.
> 

	My thoughts exactly.  Once the VFS layer is sorted out, it
	shouldn't be a problem to load the file system modules as
	KLDs later on.  Also, we could write something a pseudo-VFS
	module which can be used to stack a module for a root file
	system; this way, the actual root is a pseudo VFS and there
	is an on-disk file system which sit on top of it.

	This will make it easier for writing root support in file
	systems, and you wouldn't need to modify file systems so
	much for it.

	Regards,

-- 
Hiten Pandya
hmp at xxxxxxxxxxxxx






More information about the Kernel mailing list