Re: [PATCH] logfs: remove from tree

2016-12-01 Thread Christoph Hellwig 
On Mon, Oct 24, 2016 at 08:37:15AM +0200, Christoph Hellwig wrote:
> On Sat, Oct 15, 2016 at 01:33:29AM +0100, Al Viro wrote:
> > IMO that's a plain and simple "just before -rc1" fodder, to be sent
> > straight to Linus.  No objections from me, anyway.
> 
> It's "just after -rc2" now, but maybe it's still time for this
> trivial removal?

Now we're getting close to the end of this merge window, any chance
you could queue it up for the next merge window?

Re: [PATCH] logfs: remove from tree

2016-12-01 Thread Christoph Hellwig 
On Mon, Oct 24, 2016 at 08:37:15AM +0200, Christoph Hellwig wrote:
> On Sat, Oct 15, 2016 at 01:33:29AM +0100, Al Viro wrote:
> > IMO that's a plain and simple "just before -rc1" fodder, to be sent
> > straight to Linus.  No objections from me, anyway.
> 
> It's "just after -rc2" now, but maybe it's still time for this
> trivial removal?

Now we're getting close to the end of this merge window, any chance
you could queue it up for the next merge window?

Re: [PATCH] logfs: remove from tree

2016-10-24 Thread Christoph Hellwig 
On Sat, Oct 15, 2016 at 01:33:29AM +0100, Al Viro wrote:
> IMO that's a plain and simple "just before -rc1" fodder, to be sent
> straight to Linus.  No objections from me, anyway.

It's "just after -rc2" now, but maybe it's still time for this
trivial removal?

Re: [PATCH] logfs: remove from tree

2016-10-24 Thread Christoph Hellwig 
On Sat, Oct 15, 2016 at 01:33:29AM +0100, Al Viro wrote:
> IMO that's a plain and simple "just before -rc1" fodder, to be sent
> straight to Linus.  No objections from me, anyway.

It's "just after -rc2" now, but maybe it's still time for this
trivial removal?

Re: [PATCH] logfs: remove from tree

2016-10-14 Thread Al Viro 
On Sun, Sep 11, 2016 at 03:04:22PM +0200, Christoph Hellwig wrote:
> Logfs was introduced to the kernel in 2009, and hasn't seen any non
> drive-by changes since 2012, while having lots of unsolved issues
> including the complete lack of error handling, with more and more
> issues popping up without any fixes.
> 
> The logfs.org domain has been bouncing from a mail, and the maintainer
> on the non-logfs.org domain hasn't repsonded to past queries either.
> 
> Signed-off-by: Christoph Hellwig 
> ---
>  Documentation/filesystems/00-INDEX  |2 -
>  Documentation/filesystems/logfs.txt |  241 
>  MAINTAINERS |8 -
>  fs/Kconfig  |1 -
>  fs/Makefile |1 -
[snip]

IMO that's a plain and simple "just before -rc1" fodder, to be sent
straight to Linus.  No objections from me, anyway.

Re: [PATCH] logfs: remove from tree

2016-10-14 Thread Al Viro 
On Sun, Sep 11, 2016 at 03:04:22PM +0200, Christoph Hellwig wrote:
> Logfs was introduced to the kernel in 2009, and hasn't seen any non
> drive-by changes since 2012, while having lots of unsolved issues
> including the complete lack of error handling, with more and more
> issues popping up without any fixes.
> 
> The logfs.org domain has been bouncing from a mail, and the maintainer
> on the non-logfs.org domain hasn't repsonded to past queries either.
> 
> Signed-off-by: Christoph Hellwig 
> ---
>  Documentation/filesystems/00-INDEX  |2 -
>  Documentation/filesystems/logfs.txt |  241 
>  MAINTAINERS |8 -
>  fs/Kconfig  |1 -
>  fs/Makefile |1 -
[snip]

IMO that's a plain and simple "just before -rc1" fodder, to be sent
straight to Linus.  No objections from me, anyway.

Re: [PATCH] logfs: remove from tree

2016-09-12 Thread Austin S. Hemmelgarn 

On 2016-09-12 02:55, Artem Bityutskiy wrote:

On Sun, 2016-09-11 at 15:04 +0200, Christoph Hellwig wrote:

Logfs was introduced to the kernel in 2009, and hasn't seen any non
drive-by changes since 2012, while having lots of unsolved issues
including the complete lack of error handling, with more and more
issues popping up without any fixes.

The logfs.org domain has been bouncing from a mail, and the
maintainer
on the non-logfs.org domain hasn't repsonded to past queries either.

Signed-off-by: Christoph Hellwig 


Back in 2008 logfs and UBIFS were in sort of competing projects. I
remember we inspected logfs code and tested it - we did not find proper
wear-levelling and bad block handling, we did not see proper error
handling, and it exploded when we were running relatively simple tests.
We indicated this here in a very humble way to avoid the "conflict of
interest" perseption:

https://lkml.org/lkml/2008/3/31/117

I did not follow logfs since then, but I think there wasn't much
development since then and all these issue are still there. I mean,
unless I am horribly mistaken, logfs does not really have the basic
features of a flash file system and there is no point keeping it in the
tree and consuming people's time maintaining it.

FWIW, I tried testing it about a year ago, and got similar results both 
from the tests and from trying to contact the maintainer.

Re: [PATCH] logfs: remove from tree

2016-09-12 Thread Austin S. Hemmelgarn 

On 2016-09-12 02:55, Artem Bityutskiy wrote:

On Sun, 2016-09-11 at 15:04 +0200, Christoph Hellwig wrote:

Logfs was introduced to the kernel in 2009, and hasn't seen any non
drive-by changes since 2012, while having lots of unsolved issues
including the complete lack of error handling, with more and more
issues popping up without any fixes.

The logfs.org domain has been bouncing from a mail, and the
maintainer
on the non-logfs.org domain hasn't repsonded to past queries either.

Signed-off-by: Christoph Hellwig 


Back in 2008 logfs and UBIFS were in sort of competing projects. I
remember we inspected logfs code and tested it - we did not find proper
wear-levelling and bad block handling, we did not see proper error
handling, and it exploded when we were running relatively simple tests.
We indicated this here in a very humble way to avoid the "conflict of
interest" perseption:

https://lkml.org/lkml/2008/3/31/117

I did not follow logfs since then, but I think there wasn't much
development since then and all these issue are still there. I mean,
unless I am horribly mistaken, logfs does not really have the basic
features of a flash file system and there is no point keeping it in the
tree and consuming people's time maintaining it.

FWIW, I tried testing it about a year ago, and got similar results both 
from the tests and from trying to contact the maintainer.

Re: [PATCH] logfs: remove from tree

2016-09-12 Thread Artem Bityutskiy 
On Sun, 2016-09-11 at 15:04 +0200, Christoph Hellwig wrote:
> Logfs was introduced to the kernel in 2009, and hasn't seen any non
> drive-by changes since 2012, while having lots of unsolved issues
> including the complete lack of error handling, with more and more
> issues popping up without any fixes.
> 
> The logfs.org domain has been bouncing from a mail, and the
> maintainer
> on the non-logfs.org domain hasn't repsonded to past queries either.
> 
> Signed-off-by: Christoph Hellwig 

Back in 2008 logfs and UBIFS were in sort of competing projects. I
remember we inspected logfs code and tested it - we did not find proper
wear-levelling and bad block handling, we did not see proper error
handling, and it exploded when we were running relatively simple tests.
We indicated this here in a very humble way to avoid the "conflict of
interest" perseption:

https://lkml.org/lkml/2008/3/31/117

I did not follow logfs since then, but I think there wasn't much
development since then and all these issue are still there. I mean,
unless I am horribly mistaken, logfs does not really have the basic
features of a flash file system and there is no point keeping it in the
tree and consuming people's time maintaining it.

Artem.

Re: [PATCH] logfs: remove from tree

2016-09-12 Thread Artem Bityutskiy 
On Sun, 2016-09-11 at 15:04 +0200, Christoph Hellwig wrote:
> Logfs was introduced to the kernel in 2009, and hasn't seen any non
> drive-by changes since 2012, while having lots of unsolved issues
> including the complete lack of error handling, with more and more
> issues popping up without any fixes.
> 
> The logfs.org domain has been bouncing from a mail, and the
> maintainer
> on the non-logfs.org domain hasn't repsonded to past queries either.
> 
> Signed-off-by: Christoph Hellwig 

Back in 2008 logfs and UBIFS were in sort of competing projects. I
remember we inspected logfs code and tested it - we did not find proper
wear-levelling and bad block handling, we did not see proper error
handling, and it exploded when we were running relatively simple tests.
We indicated this here in a very humble way to avoid the "conflict of
interest" perseption:

https://lkml.org/lkml/2008/3/31/117

I did not follow logfs since then, but I think there wasn't much
development since then and all these issue are still there. I mean,
unless I am horribly mistaken, logfs does not really have the basic
features of a flash file system and there is no point keeping it in the
tree and consuming people's time maintaining it.

Artem.

Re: [PATCH] logfs: remove from tree

2016-09-11 Thread Christoph Hellwig 
On Mon, Sep 12, 2016 at 11:53:19AM +1000, Dave Chinner wrote:
> Wasn't the lib/btree.c implementation introduced with and only used
> by logfs? Should that go as well?

The qla2xxx SCSI target driver also uses the btree library.

Re: [PATCH] logfs: remove from tree

2016-09-11 Thread Christoph Hellwig 
On Mon, Sep 12, 2016 at 11:53:19AM +1000, Dave Chinner wrote:
> Wasn't the lib/btree.c implementation introduced with and only used
> by logfs? Should that go as well?

The qla2xxx SCSI target driver also uses the btree library.

Re: [PATCH] logfs: remove from tree

2016-09-11 Thread Dave Chinner 
On Sun, Sep 11, 2016 at 03:04:22PM +0200, Christoph Hellwig wrote:
> Logfs was introduced to the kernel in 2009, and hasn't seen any non
> drive-by changes since 2012, while having lots of unsolved issues
> including the complete lack of error handling, with more and more
> issues popping up without any fixes.
> 
> The logfs.org domain has been bouncing from a mail, and the maintainer
> on the non-logfs.org domain hasn't repsonded to past queries either.
> 
> Signed-off-by: Christoph Hellwig 
> ---
>  Documentation/filesystems/00-INDEX  |2 -
>  Documentation/filesystems/logfs.txt |  241 
>  MAINTAINERS |8 -
>  fs/Kconfig  |1 -
>  fs/Makefile |1 -
>  fs/logfs/Kconfig|   17 -
>  fs/logfs/Makefile   |   13 -
>  fs/logfs/compr.c|   95 --
>  fs/logfs/dev_bdev.c |  322 -
>  fs/logfs/dev_mtd.c  |  274 -
>  fs/logfs/dir.c  |  797 
>  fs/logfs/file.c |  285 -
>  fs/logfs/gc.c   |  732 ---
>  fs/logfs/inode.c|  428 ---
>  fs/logfs/journal.c  |  894 --
>  fs/logfs/logfs.h|  735 ---
>  fs/logfs/logfs_abi.h|  629 --
>  fs/logfs/readwrite.c| 2298 
> ---
>  fs/logfs/segment.c  |  961 ---
>  fs/logfs/super.c|  653 --

Wasn't the lib/btree.c implementation introduced with and only used
by logfs? Should that go as well?

Cheers,

Dave.
-- 
Dave Chinner
da...@fromorbit.com

Re: [PATCH] logfs: remove from tree

2016-09-11 Thread Dave Chinner 
On Sun, Sep 11, 2016 at 03:04:22PM +0200, Christoph Hellwig wrote:
> Logfs was introduced to the kernel in 2009, and hasn't seen any non
> drive-by changes since 2012, while having lots of unsolved issues
> including the complete lack of error handling, with more and more
> issues popping up without any fixes.
> 
> The logfs.org domain has been bouncing from a mail, and the maintainer
> on the non-logfs.org domain hasn't repsonded to past queries either.
> 
> Signed-off-by: Christoph Hellwig 
> ---
>  Documentation/filesystems/00-INDEX  |2 -
>  Documentation/filesystems/logfs.txt |  241 
>  MAINTAINERS |8 -
>  fs/Kconfig  |1 -
>  fs/Makefile |1 -
>  fs/logfs/Kconfig|   17 -
>  fs/logfs/Makefile   |   13 -
>  fs/logfs/compr.c|   95 --
>  fs/logfs/dev_bdev.c |  322 -
>  fs/logfs/dev_mtd.c  |  274 -
>  fs/logfs/dir.c  |  797 
>  fs/logfs/file.c |  285 -
>  fs/logfs/gc.c   |  732 ---
>  fs/logfs/inode.c|  428 ---
>  fs/logfs/journal.c  |  894 --
>  fs/logfs/logfs.h|  735 ---
>  fs/logfs/logfs_abi.h|  629 --
>  fs/logfs/readwrite.c| 2298 
> ---
>  fs/logfs/segment.c  |  961 ---
>  fs/logfs/super.c|  653 --

Wasn't the lib/btree.c implementation introduced with and only used
by logfs? Should that go as well?

Cheers,

Dave.
-- 
Dave Chinner
da...@fromorbit.com