attrib +A doesn't work

attrib +A doesn't work

[Tom Vier]

i thought 3.6 support attribs (or at least some of them), but A isn't
working. i ran touch test; chattr +A test; stat test; cat test; stat test;
the second stat showed an updated atime.

is there some undocumented mount option to enable attribs?

-- 
Tom Vier <tmv@comcast.net>
DSA Key ID 0xE6CB97DA

Re: attrib +A doesn't work

[Nikita Danilov]

Tom Vier writes:
 > i thought 3.6 support attribs (or at least some of them), but A isn't
 > working. i ran touch test; chattr +A test; stat test; cat test; stat test;
 > the second stat showed an updated atime.
 > 
 > is there some undocumented mount option to enable attribs?

Yes. 

Support for inode attributes was introduced between 2.4.18 and
2.4.19. Take latest reiserfsprogs from ftp.namesys.com, run
/sbin/reiserfsck --clean-attributes /your-device, and after that
finishes mount with -attrs option.

 > 
 > -- 
 > Tom Vier <tmv@comcast.net>
 > DSA Key ID 0xE6CB97DA

Nikita.

Re: attrib +A doesn't work

[Oleg Drokin]

Hello!

On Thu, Sep 05, 2002 at 01:57:34PM +0400, Nikita Danilov wrote:

> Support for inode attributes was introduced between 2.4.18 and
> 2.4.19. Take latest reiserfsprogs from ftp.namesys.com, run
> /sbin/reiserfsck --clean-attributes /your-device, and after that
> finishes mount with -attrs option.

mount -o attrs in fact.

Bye,
    Oleg

Re: attrib +A doesn't work

[Allen H. Ibara]

> mount -o attrs in fact.

I'm not sure if this is the right place to report this...

The 'attrs' mount option is broken with kernels 2.4.20-pre2 through -pre5 (the
latest as of this email).

See "license cleanup" patch found here:

ftp://ftp.namesys.com/pub/reiserfs-for-2.4/2.4.19.pending/01-license_cleanup.diff

It contains a re-work of reiserfs_parse_options() in super.c that removes
the 'attrs' option. Perhaps a merge error or simple omission.

Trivial fix:

diff -ur linux-2.4.19-orig/fs/reiserfs/super.c linux-2.4.19/fs/reiserfs/super.c
--- linux-2.4.19-orig/fs/reiserfs/super.c	Sat Sep  7 17:29:12 2002
+++ linux-2.4.19/fs/reiserfs/super.c	Sat Sep  7 13:44:27 2002
@@ -560,6 +560,7 @@
     char * pos;
     opt_desc_t opts[] = {
 		{"tails", 't', tails, -1},
+		{"attrs", 0, 0, REISERFS_ATTRS},
 		{"notail", 0, 0, -1}, /* Compatibility stuff, so that -o notail for old setups still work */
 		{"conv", 0, 0, REISERFS_CONVERT}, 
 		{"nolog", 0, 0, -1},

--Allen Ibara

Re: attrib +A doesn't work

[Oleg Drokin]

Hello!

On Sat, Sep 07, 2002 at 05:35:26PM -0700, Allen H. Ibara wrote:
> > mount -o attrs in fact.
> I'm not sure if this is the right place to report this...

Sure it is.

> The 'attrs' mount option is broken with kernels 2.4.20-pre2 through -pre5 (the
> latest as of this email).
> See "license cleanup" patch found here:
> ftp://ftp.namesys.com/pub/reiserfs-for-2.4/2.4.19.pending/01-license_cleanup.diff
> It contains a re-work of reiserfs_parse_options() in super.c that removes
> the 'attrs' option. Perhaps a merge error or simple omission.

Thanks a lot for letting us know.
Somehow this omission got past our testing.

Bye,
    Oleg

Re: attrib +A doesn't work

[Paul Jarc]

[Old thread.]

Oleg Drokin <green@namesys.com> wrote:
> On Thu, Sep 05, 2002 at 01:57:34PM +0400, Nikita Danilov wrote:
>> Support for inode attributes was introduced between 2.4.18 and
>> 2.4.19. Take latest reiserfsprogs from ftp.namesys.com, run
>> /sbin/reiserfsck --clean-attributes /your-device, and after that
>> finishes mount with -attrs option.
>
> mount -o attrs in fact.

Should -o attrs be used just once, or every time afterwards?


paul

Re: attrib +A doesn't work

[Nikita Danilov]

Paul Jarc writes:
 > [Old thread.]
 > 
 > Oleg Drokin <green@namesys.com> wrote:
 > > On Thu, Sep 05, 2002 at 01:57:34PM +0400, Nikita Danilov wrote:
 > >> Support for inode attributes was introduced between 2.4.18 and
 > >> 2.4.19. Take latest reiserfsprogs from ftp.namesys.com, run
 > >> /sbin/reiserfsck --clean-attributes /your-device, and after that
 > >> finishes mount with -attrs option.
 > >
 > > mount -o attrs in fact.
 > 
 > Should -o attrs be used just once, or every time afterwards?
 > 

On each mount. Does attribute support actually works for you?

 > 
 > paul

Nikita.

Re: attrib +A doesn't work

[Paul Jarc]

Nikita Danilov <Nikita@Namesys.COM> wrote:
> Paul Jarc writes:
>> Should -o attrs be used just once, or every time afterwards?
>
> On each mount. Does attribute support actually works for you?

I just did a quick test with chattr +A, and it seemed to work.  Out of
curiosity, what would happen if I forgot to use attrs once?  I'm
guessing attributes would not take effect and could not be changed
until I remounted, but once I did, the stored attributes wouldn't be
lost, right?  Also, is there any way to make attrs and notail show up
in /proc/mounts?


paul

Re: attrib +A doesn't work

[Nikita Danilov]

Paul Jarc writes:
 > Nikita Danilov <Nikita@Namesys.COM> wrote:
 > > Paul Jarc writes:
 > >> Should -o attrs be used just once, or every time afterwards?
 > >
 > > On each mount. Does attribute support actually works for you?
 > 
 > I just did a quick test with chattr +A, and it seemed to work.  Out of
 > curiosity, what would happen if I forgot to use attrs once?  I'm
 > guessing attributes would not take effect and could not be changed
 > until I remounted, but once I did, the stored attributes wouldn't be
 > lost, right?  Also, is there any way to make attrs and notail show up
 > in /proc/mounts?

Only after Halloween, I am afraid.

 > 
 > 
 > paul

Nikita.