Re: OT: Fixed Roland'd Hurd EA ext2 patch for Linux

[cascardo]

On Sun, Jan 29, 2006 at 04:28:23PM +0100, Alfred M. Szmidt wrote:
> I don't know what the procedure to include a patch in Linux is.  If
> there are some problems including the patch (copyright, functionality,
> ...) then this is up to the maintainers of Linux to point out.  I
> think there is no harm in submitting it to where ever patches go for
> Linux.
> 

I think we should leave that to Roland, since he is the author of the
patch. But I will be glad if it's accepted. Since the patch I've
submitted to the list fails to apply (because of the added lines
indicating the date of change), an incremental patch is available from
http://people.debian.org/~mbanck/xattr-hurd/cascardo.interdiff.

>    Another solution I am thinking in working on is modifying debugfs
>    to allow people to set translators in an umounted filesystem. This
>    would require no linux building, but only a e2fsprogs update.
> 
> What is debugfs?
> 

It is a tool written by Theodore T'so that manipulates an ext2
filesystem. There is also a debugfs in Linux, which I don't know
well. I'm not referring to it.

I've written a simple tool using the e2fslibs from the same Theodore
T'so that sets and shows the translator from userspace in an unmounted
filesystem. I considered they should be included as a command in
debugfs instead. Howeve, reading the code of debugfs, it already
supports reading the block of the translator, but it seems one cannot
overwrite the contents of a block using debugfs. I'm not really sure
which approach would be better: using my simple tools, writing support
for debugfs to write arbitrary data to arbitrary blocks (and then
writing a debugfs script to set the translator) or writing support for
debugfs to directly set the translator.

Regards,
Thadeu Cascardo.
--

signature.asc
Description: Digital signature