DragonFlyBSD

Hi again;

Here is an updated patch.

:Hi again;
:
:Here is an updated patch.

Thanks Pedro.  I will apply the brelse fixes, the i_next_alloc_block
    fix, and a few others to DragonFly.  We will have to forego the
    application of the ASYNC enhancements and most of the syntax changes
    (though I'll definitely keep the 'error =' syntax adjustments),
    mainly because we are too close to the release to test them and I
    don't want to create any accidental bugs.  ext2fs support is really
    just there for accessibility, we don't expect people to use ext2fs for
    a main filesystem in DragonFly.

-Matt

Matthew Dillon wrote:

:Hi again;
:
:Here is an updated patch.

Thanks Pedro. I will apply the brelse fixes, the i_next_alloc_block
fix, and a few others to DragonFly. We will have to forego the
application of the ASYNC enhancements and most of the syntax changes
(though I'll definitely keep the 'error =' syntax adjustments),
mainly because we are too close to the release to test them and I
don't want to create any accidental bugs. ext2fs support is really
just there for accessibility, we don't expect people to use ext2fs for
a main filesystem in DragonFly.

-Matt

Hi Matt;

The ASYNC enhancements are actually harmless but I agree with the idea
of minimizing unnecessary risks before a release.

Ext2fs is rather uninteresting for default use but with Aditya's port of
NetBSD's allocation code, the new BSD-licensed ext2fs is a copy cat of
UFS1. This means we could start using some enhancements from FFS
(softupdates sounds crazy but is doable) and we could check some of
linux's developments (preallocation, Htree) to see if they make sense in
UFS.

Not as fun as HAMMER, but nevertheless fun ;-).

Pedro.

Has this made it in yet? If not, do we want it in for after the release?

Cheers,
Alex Hornung

Hello;

Alex Hornung (via DragonFly issue tracker) wrote:

Alex Hornung <ahornung@gmail.com> added the comment:

Has this made it in yet? If not, do we want it in for after the release?

Cheers,
Alex Hornung

Yes, Some parts were left out (async enhancements) but Matt Dillon
committed the critical parts.

BTW, the locking would require redoing but it would be nice to see the
BSD licensed ext2fs from FreeBSD merged someday.

Pedro.