[Ocfs2-devel] xattr fixes branch

[Ocfs2-devel] xattr fixes branch

[Joel Becker]

Folks,
	To keep track of what we want to upstream in the 2.6.28-rc
cycle, I've put together a branch called 'xattr-28' with almost all of
the patches merged.  You can see it and pull it from my linux-2.6.git
repository.

View:
http://oss.oracle.com/git/?p=jlbec/linux-2.6.git;a=shortlog;h=xattr-28
Pull:
git://oss.oracle.com/git/jlbec/linux-2.6.git xattr-28

	The branch contains the following changes:

Joel Becker (5):
      ocfs2: Check xattr block signatures properly.
      ocfs2: Don't return -EFAULT from a corrupt xattr entry.
      ocfs2: Check errors from ocfs2_xattr_update_xattr_search()
      ocfs2: Specify appropriate journal access for new xattr buckets.
      ocfs2: Don't repeat ocfs2_xattr_block_find()

Tao Ma (1):
      ocfs2: Remove unused ocfs2_restore_xattr_block().

Tiger Yang (4):
      ocfs2: fix license in xattr
      ocfs2: fix function declaration and definition in xattr
      ocfs2: remove duplicate definition in xattr
      ocfs2: add handler_map array bounds checking

	These fixes and cleanups I think are candidates for 2.6.28-rc.
The last patch I'm waiting for is Tao's "ocfs2: Proper hash collision
handling in xattr bucket division."  That's also a bugfix, and I think a
candidate for 2.6.28-rc.  Once we get that nailed down, I think we're
good, and we'll send this stuff along upstream.
	Tristan, I'd love it if you could test this branch to make sure
these fixes and cleanups didn't break anything.

Joel

-- 

"Egotist: a person more interested in himself than in me."
         - Ambrose Bierce 

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127

[Ocfs2-devel] xattr fixes branch

[Tao Ma]

Joel Becker wrote:
> Folks,
> 	To keep track of what we want to upstream in the 2.6.28-rc
> cycle, I've put together a branch called 'xattr-28' with almost all of
> the patches merged.  You can see it and pull it from my linux-2.6.git
> repository.
Thanks for that. So I can base my work on your branch for my future work.

Regards,
Tao

[Ocfs2-devel] xattr fixes branch

[Joel Becker]

On Fri, Oct 24, 2008 at 08:57:32AM +0800, Tao Ma wrote:
> Joel Becker wrote:
> > Folks,
> > 	To keep track of what we want to upstream in the 2.6.28-rc
> > cycle, I've put together a branch called 'xattr-28' with almost all of
> > the patches merged.  You can see it and pull it from my linux-2.6.git
> > repository.
> Thanks for that. So I can base my work on your branch for my future work.

	Yes.  What I put there will go to Mark as-is.  But the only
thing I'm planning to add to xattr-28 is the hash collision fix.  On top
of that we'll start our .29 work, which includes my read-metadata
branch, the xattr_set patch, some other xattr cleanups I'm doing, the
ACL code, and much much more :-)

Joel

-- 

"If the human brain were so simple we could understand it, we would
 be so simple that we could not."
	- W. A. Clouston

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127

[Ocfs2-devel] xattr fixes branch

[Mark Fasheh]

On Thu, Oct 23, 2008 at 05:22:46PM -0700, Joel Becker wrote:
> Folks,
> 	To keep track of what we want to upstream in the 2.6.28-rc
> cycle, I've put together a branch called 'xattr-28' with almost all of
> the patches merged.  You can see it and pull it from my linux-2.6.git
> repository.

Joel,

	Thanks for doing this. I've got the patches all pulled into
ocfs2.git now, and they're ready to go. I'm thinking we should try something
new this time around. I'd like to get yours and everyone elses opinions on
this.

	I'm thinking to keep fixes that will go upstream in the current
release in a 'fixes' branch. For 2.6.28, I started this from your xattr-28
branch, put my signoffs on the patches, and added an mmap fix from Tao which
I had already. This will get rebased every time I send fixes upstream.

	For new features, I started a merge_window branch. This worked
really well for us recently imho. Early in the cycle, this might get rebased
as large-ish fixes are pushed upstream. By the end of the development cycle
though (when fixes are smaller), I suspect it won't have to be rebased as
often, or at all.

	If people wouldn't mind developing against some version of
merge_window, it would make getting all this stuff together much easier. As
always though, if the choice is between rebasing or doing some PITA merge
versus getting a high quality feature going, feel free to push the merge
pain onto me :)

	Right now, merge_window is based on upstream. I'm thinking perhaps
we should base it on the current 'fixes' branch, since some of them are on
the large-ish side - that would make getting our patches together more
smoothly.

	'linux-next' and 'ALL' branches would stay the same - a merge of
everything we've got, intended to go into the linux-next and -mm trees.


	Tiger - you had asked me to keep the previous merge_window branch
around. It's now in ocfs2.git as 'merge_window_28'. Let me know if I can
remove this.
	--Mark

--
Mark Fasheh

[Ocfs2-devel] xattr fixes branch

[Joel Becker]

On Sat, Oct 25, 2008 at 02:19:00PM -0700, Mark Fasheh wrote:
> 	Thanks for doing this. I've got the patches all pulled into

	No problem.

> 	I'm thinking to keep fixes that will go upstream in the current
> release in a 'fixes' branch. For 2.6.28, I started this from your xattr-28
> branch, put my signoffs on the patches, and added an mmap fix from Tao which
> I had already. This will get rebased every time I send fixes upstream.
> 
> 	For new features, I started a merge_window branch. This worked
> really well for us recently imho. Early in the cycle, this might get rebased
> as large-ish fixes are pushed upstream. By the end of the development cycle
> though (when fixes are smaller), I suspect it won't have to be rebased as
> often, or at all.

	Works for me.  I'll even take on one of the two branches if that
makes your life easier.  Or if you want me to take them at moments your
busy (like we just did), just shout.

> 	If people wouldn't mind developing against some version of
> merge_window, it would make getting all this stuff together much easier. As
> always though, if the choice is between rebasing or doing some PITA merge
> versus getting a high quality feature going, feel free to push the merge
> pain onto me :)

	I think merge_window is great, because our ALL doesn't
distinguish "going to next version" from "marinating".  I wonder,
though, if "merge_window" and "linux-next" don't overlap?  What would be
in "merge_window" that wouldn't also be in "linux-next"?

> 	Right now, merge_window is based on upstream. I'm thinking perhaps
> we should base it on the current 'fixes' branch, since some of them are on
> the large-ish side - that would make getting our patches together more
> smoothly.

	This is probably your call, but this sounds like a good way to
do it.

Joel

-- 

"You don't make the poor richer by making the rich poorer."
	- Sir Winston Churchill

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127