headers
OGAWA Hirofumi | 4 Jun 04:15
Picon
Gravatar

inode dirty management and reference count

Hi,

Those patches add the inode dirty management and reference count. I
think those can be used for both of good writeback flush and
atomic-commit.

I guess the issues of this patchset are, iget() is not finding the all
in-core inode (reopen is not allowed, it's same with current though).
And tux3fuse patches are almost untested, sorry.

	static-http://userweb.kernel.org/~hirofumi/tux3/

Please review, and pull if ok.

Thanks.
--

-- 
OGAWA Hirofumi <hirofumi <at> mail.parknet.co.jp>
Permalink | Reply | 0 comments |
headers
OGAWA Hirofumi | 19 Jun 20:21
Picon
Gravatar

inode dirty management and reference count

Hi,

Those patches are bug fixes in xattr.c and writeback.c.

And those add the deferred ileaf update for inode number allocation.  It
defers the ileaf update (this is only for inum allocation) until backend
for atomic-commit. To do it, this allows the inode doesn't have the data
btree, and at the first write, allocate btree.

	static-http://userweb.kernel.org/~hirofumi/tux3/

Please review, and pull if ok.

Thanks.
--

-- 
OGAWA Hirofumi <hirofumi <at> mail.parknet.co.jp>
Permalink | Reply | 0 comments |
headers
OGAWA Hirofumi | 29 Jun 03:56
Picon
Gravatar

start the logging for atomic commit

Hi,

Those patches are start to logging for atomic commit, and fix the
stage_delta()/flush_log() to flush correctly. Also, this fixes the
deferred bfree stuff.  And some bug fixes.

The problem of this patchset is, first of all, I worked only for
creation path as start.

I copied code from kernel to utility.*, so we would need to think about
license of those, or remove the code from this patch.

And test is not enough, some programs/code-path would be untested at all.

Main known problems are: It will still use writeback stuff, so, we need
more code to switch to commit stuff. But, before switch, we will need to
add to flush bitmap inode.

And This is first one to start logging, so, there may be bugs on around
those.  Kernel is compile test only.

And this may change the disk format without changing revision. Well,
there would be many problems.

But, it starts the atomic-commit more or less.

	static-http://userweb.kernel.org/~hirofumi/tux3/

Please review, and pull if ok.
(Continue reading)

Permalink | Reply | 2 comments |
headers
Nandan Yen | 30 Jun 11:52
Picon

Updating my calendar

Hello

I am creating a birthday calendar for myself.  Can you please click on the link below and enter your birthday
for me?

http://www.birthdayalarm.com/bd2/85334167a876025448b1472437903c808070343d1386

Nandan
Permalink | Reply | 0 comments |

Gmane