-o noposixpaths

-o noposixpaths

[Jeff Layton]

Is this mount option actually useful? Is there ever a case where we'd
want to use unix extensions, but not use posix-style pathnames? If not,
does it make sense to deprecate this option?

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: -o noposixpaths

[Steve French]

Yes, IIRC the apple guys mentioned plausible server scenarios for this
(where we want to mount with unix extensions for symlinks and ownership
but server can not handle posix path names)

Presumably if the server file system does not support posix path names
(FAT32, NTFS?) or if we want to restrict the characters (for interoperability
with Windows clients accessing the same share?) - might be other cases.

On Mon, Jul 11, 2011 at 1:23 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> Is this mount option actually useful? Is there ever a case where we'd
> want to use unix extensions, but not use posix-style pathnames? If not,
> does it make sense to deprecate this option?
>
> --
> Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-cifs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>



-- 
Thanks,

Steve

Re: -o noposixpaths

[Jeff Layton]

On Mon, 11 Jul 2011 13:32:54 -0500
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> Yes, IIRC the apple guys mentioned plausible server scenarios for this
> (where we want to mount with unix extensions for symlinks and ownership
> but server can not handle posix path names)
> 
> Presumably if the server file system does not support posix path names
> (FAT32, NTFS?) or if we want to restrict the characters (for interoperability
> with Windows clients accessing the same share?) - might be other cases.
> 

In that case though, shouldn't those servers just not set
CIFS_UNIX_POSIX_PATHNAMES_CAP ?


> On Mon, Jul 11, 2011 at 1:23 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > Is this mount option actually useful? Is there ever a case where we'd
> > want to use unix extensions, but not use posix-style pathnames? If not,
> > does it make sense to deprecate this option?
> >
> > --
> > Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-cifs" in
> > the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> 
> 
> 


-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: -o noposixpaths

[Steve French]

On Mon, Jul 11, 2011 at 1:41 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> On Mon, 11 Jul 2011 13:32:54 -0500
> Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
>> Yes, IIRC the apple guys mentioned plausible server scenarios for this
>> (where we want to mount with unix extensions for symlinks and ownership
>> but server can not handle posix path names)
>>
>> Presumably if the server file system does not support posix path names
>> (FAT32, NTFS?) or if we want to restrict the characters (for interoperability
>> with Windows clients accessing the same share?) - might be other cases.
>>
>
> In that case though, shouldn't those servers just not set
> CIFS_UNIX_POSIX_PATHNAMES_CAP ?

I don't think servers do that (unset the POSIX_PATHNAMES CAP for one
share and not others, even assuming the share exports all volumes
of the same file system type)  for the former case
(unless they cant support posix path at all for the whole server),
and for the latter case, not sure that the server can know enough
information (about other clients which may mount the system)
to unset the CAP unilaterally.



-- 
Thanks,

Steve

Re: -o noposixpaths

[Jeff Layton]

On Mon, 11 Jul 2011 13:48:20 -0500
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> On Mon, Jul 11, 2011 at 1:41 PM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > On Mon, 11 Jul 2011 13:32:54 -0500
> > Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> >
> >> Yes, IIRC the apple guys mentioned plausible server scenarios for this
> >> (where we want to mount with unix extensions for symlinks and ownership
> >> but server can not handle posix path names)
> >>
> >> Presumably if the server file system does not support posix path names
> >> (FAT32, NTFS?) or if we want to restrict the characters (for interoperability
> >> with Windows clients accessing the same share?) - might be other cases.
> >>
> >
> > In that case though, shouldn't those servers just not set
> > CIFS_UNIX_POSIX_PATHNAMES_CAP ?
> 
> I don't think servers do that (unset the POSIX_PATHNAMES CAP for one
> share and not others, even assuming the share exports all volumes
> of the same file system type)  for the former case
> (unless they cant support posix path at all for the whole server),
> and for the latter case, not sure that the server can know enough
> information (about other clients which may mount the system)
> to unset the CAP unilaterally.
> 

This explanation doesn't make any sense. IIUC, The server should only
set the flag if it's appropriate to use posix-style pathnames on the
share. Why should the server care at all what the clients can support?

Perhaps I should phrase this question differently: Under what
circumstances would someone want to use the "noposixpaths" mount option?

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>