Assaf Weizman | 8 Mar 2010 09:59
Picon

Super Root Block

Hello,
Could you please explain a few theoretical questions:
a. What's the meaning/usage of the "Super Root Block"?
b. in the Readme (v2) file it's described as "an optional super root
block (SR)". So as optional element,  who chooses to whether to
activate\use it or not?

Thanks a lot. this is a very interesting file system.

Best regards
Assaf Weizman
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Ryusuke Konishi | 8 Mar 2010 11:08

Re: Super Root Block

Hi,
On Mon, 8 Mar 2010 10:59:44 +0200, Assaf Weizman wrote:
> Hello,
> Could you please explain a few theoretical questions:
> a. What's the meaning/usage of the "Super Root Block"?
> b. in the Readme (v2) file it's described as "an optional super root
> block (SR)". So as optional element,  who chooses to whether to
> activate\use it or not?

The super root block contains inodes of checkpoint file (cpfile),
segment usage file (sufile), and disk address translation (DAT) file.

It represents the top of metadata hierarchy of NILFS at a time point.
For the metadata hierarchy, please see the page 9 of the slides:

 http://www.nilfs.org/papers/jls2009-nilfs.pdf 

I think the term "optional" in the readme looks confusing.

In nilfs, if a series of logs makes a new checkpoint, it essentially
ends with a super root block.  Otherwise, it does not entail the super
root block (there is a variant of log series used for synchronous data
write).

> Thanks a lot. this is a very interesting file system.
> 
> Best regards
> Assaf Weizman

Thank you.  I think the above slides would be helpful than the readme
(Continue reading)

Assaf Weizman | 8 Mar 2010 14:22
Picon

Re: Super Root Block

Thank you very much! Mr. Konishi.

Best wishes
Assaf Weizman

On Mon, Mar 8, 2010 at 12:08 PM, Ryusuke Konishi <ryusuke@...> wrote:
>
> Hi,
> On Mon, 8 Mar 2010 10:59:44 +0200, Assaf Weizman wrote:
> > Hello,
> > Could you please explain a few theoretical questions:
> > a. What's the meaning/usage of the "Super Root Block"?
> > b. in the Readme (v2) file it's described as "an optional super root
> > block (SR)". So as optional element,  who chooses to whether to
> > activate\use it or not?
>
> The super root block contains inodes of checkpoint file (cpfile),
> segment usage file (sufile), and disk address translation (DAT) file.
>
> It represents the top of metadata hierarchy of NILFS at a time point.
> For the metadata hierarchy, please see the page 9 of the slides:
>
>  http://www.nilfs.org/papers/jls2009-nilfs.pdf
>
>
> I think the term "optional" in the readme looks confusing.
>
> In nilfs, if a series of logs makes a new checkpoint, it essentially
> ends with a super root block.  Otherwise, it does not entail the super
> root block (there is a variant of log series used for synchronous data
(Continue reading)

Ryusuke Konishi | 13 Mar 2010 07:06

NILFS 2.0.19 and NILFS utils 2.0.16 released

Hi,

I have released nilfs-2.0.19 and nilfs-utils-2.0.16.

nilfs-2.0.19 is the latest standalone module maintained for older
kernels.  This release backported several bug fixes and some
problem-solving updates from the mainline kernel; the following
changes are included:

 * fix potential hang in nilfs_error on errors=remount-ro
 * add missing reader's lock in nilfs_ioctl_sync
 * fix potential leak of dirty data on umount
 * show missing mount options in /proc/mounts.
 * change mount option to disable write barriers; "nobarrier" is
   used instead of "barrier=off".
 * add "norecovery" mount option
 * use mnt_want_write in ioctls where write access is needed
 * apply readahead for recovery on mount
 * sync superblock by nilfs thread instead of periodic write_super
   callback
 * update mailing list address
 * coding style fixes

The changes of nilfs-utils include manpage updates, a build fix for
newer glibc, and a mount option to set protection_period.

The updated packages are available at

 http://www.nilfs.org/en/download.html

(Continue reading)

Jiro SEKIBA | 13 Mar 2010 09:04
Picon

Re: NILFS 2.0.19 and NILFS utils 2.0.16 released

Hi, 

At Sat, 13 Mar 2010 15:06:12 +0900 (JST),
Ryusuke Konishi wrote:
> 
> Hi,
> 
> I have released nilfs-2.0.19 and nilfs-utils-2.0.16.

I found that COPYING of nilfs-utils is now GPL v3.
I've checked older version.  Until 2.0.14, it was GPL v2,
however it upgraded to v3 in 2.0.15.

It looks like there is no COPYING file in the git repository.
Therefore, I guess,  it was automatically copied by automake 
when creating tarball.

Or is it indented to be GPL v3 now?

thanks

regards,

> nilfs-2.0.19 is the latest standalone module maintained for older
> kernels.  This release backported several bug fixes and some
> problem-solving updates from the mainline kernel; the following
> changes are included:
> 
>  * fix potential hang in nilfs_error on errors=remount-ro
>  * add missing reader's lock in nilfs_ioctl_sync
(Continue reading)

Ryusuke Konishi | 13 Mar 2010 09:59

Re: NILFS 2.0.19 and NILFS utils 2.0.16 released

On Sat, 13 Mar 2010 17:04:13 +0900, Jiro SEKIBA wrote:
> Hi, 
> 
> At Sat, 13 Mar 2010 15:06:12 +0900 (JST),
> Ryusuke Konishi wrote:
> > 
> > Hi,
> > 
> > I have released nilfs-2.0.19 and nilfs-utils-2.0.16.
> 
> I found that COPYING of nilfs-utils is now GPL v3.
> I've checked older version.  Until 2.0.14, it was GPL v2,
> however it upgraded to v3 in 2.0.15.
> 
> It looks like there is no COPYING file in the git repository.
> Therefore, I guess,  it was automatically copied by automake 
> when creating tarball.
> 
> Or is it indented to be GPL v3 now?

Ahhh, that is an accident.

Ok, I will add a COPYING file to resolve this problem.

Thank you for letting me know!

Ryusuke Konishi

> thanks
> 
(Continue reading)

Ryusuke Konishi | 13 Mar 2010 11:43

Re: NILFS 2.0.19 and NILFS utils 2.0.16 released

2010/3/13 Ryusuke Konishi <ryusuke@...>:
> On Sat, 13 Mar 2010 17:04:13 +0900, Jiro SEKIBA wrote:
>> Hi,
>>
>> At Sat, 13 Mar 2010 15:06:12 +0900 (JST),
>> Ryusuke Konishi wrote:
>> >
>> > Hi,
>> >
>> > I have released nilfs-2.0.19 and nilfs-utils-2.0.16.
>>
>> I found that COPYING of nilfs-utils is now GPL v3.
>> I've checked older version.  Until 2.0.14, it was GPL v2,
>> however it upgraded to v3 in 2.0.15.
>>
>> It looks like there is no COPYING file in the git repository.
>> Therefore, I guess,  it was automatically copied by automake
>> when creating tarball.
>>
>> Or is it indented to be GPL v3 now?
>
> Ahhh, that is an accident.
>
> Ok, I will add a COPYING file to resolve this problem.

Done.

I made another util release (nilfs-utils-2.0.17).

Ryusuke Konishi
(Continue reading)

David Arendt | 13 Mar 2010 21:49

cleaner: run one cleaning pass based on minimum free space

Hi,

In order to reduce cleaner io, I am thinking it could be usefull to
implement a parameter where you can specify the minimum free space. If
this parameter is set, instead of normal cleaning operation, the cleaner
would wait until there is less than minimum free space available and
then run one cleaning pass (from first to last segment). If after that,
there is again more than minimum free space available, continue waiting,
otherwise run cleaning passes until there is more than minimum free
space available.

What would you think about this idea ?

Bye,
David Arendt
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo@...
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Ryusuke Konishi | 14 Mar 2010 03:28
Picon
Gravatar

[PATCH 1/6] nilfs2: fix discrepancy in use of static specifier

Two segbuf functions, nilfs_segbuf_write and nilfs_segbuf_wait, are
declared with the static storage class specifier, but their
implementations are not.

This fixes the discrepancy.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@...>
---
 fs/nilfs2/segbuf.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c
index ab56fe4..68b08f2 100644
--- a/fs/nilfs2/segbuf.c
+++ b/fs/nilfs2/segbuf.c
 <at>  <at>  -470,8 +470,8  <at>  <at>  static int nilfs_segbuf_submit_bh(struct nilfs_segment_buffer *segbuf,
  *
  * %-ENOMEM - Insufficient memory available.
  */
-int nilfs_segbuf_write(struct nilfs_segment_buffer *segbuf,
-		       struct the_nilfs *nilfs)
+static int nilfs_segbuf_write(struct nilfs_segment_buffer *segbuf,
+			      struct the_nilfs *nilfs)
 {
 	struct nilfs_write_info wi;
 	struct buffer_head *bh;
 <at>  <at>  -514,7 +514,7  <at>  <at>  int nilfs_segbuf_write(struct nilfs_segment_buffer *segbuf,
  *
  * %-EIO - I/O error
  */
(Continue reading)

Ryusuke Konishi | 14 Mar 2010 03:28
Picon
Gravatar

[PATCH 0/6] nilfs2 trivial fixes

These are all trivial patches. typo fixes, coding style fixes, and a
fix of missing static specifiers.  I'll send these upstream.

Thanks,
Ryusuke Konishi
--
Ryusuke Konishi (6):
      nilfs2: fix discrepancy in use of static specifier
      nilfs2: fix function name typos in docbook comments
      nilfs2: fix typo "cout" -> "count" in error message
      nilfs2: fix various typos in comments
      nilfs2: remove spaces before tabs
      nilfs2: remove whitespaces before quoted newlines

 fs/nilfs2/alloc.h     |    2 +-
 fs/nilfs2/dat.c       |    2 +-
 fs/nilfs2/dir.c       |    2 +-
 fs/nilfs2/gcinode.c   |    4 ++--
 fs/nilfs2/page.c      |    4 ++--
 fs/nilfs2/segbuf.c    |   10 +++++-----
 fs/nilfs2/segment.c   |    8 ++++----
 fs/nilfs2/segment.h   |    4 ++--
 fs/nilfs2/sufile.c    |    2 +-
 fs/nilfs2/super.c     |    4 ++--
 fs/nilfs2/the_nilfs.c |    2 +-
 11 files changed, 22 insertions(+), 22 deletions(-)

--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo@...
(Continue reading)


Gmane