[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#637082: #637082



[Cc'ing #637082 - hope you dont mind...]

On 01.11.2011 19:52, Robert Millan wrote:
> Hi Michael,
> 
> Btw, with all this mess it seems you forgot about my patch in #637082.
>  If everything is okay with it, would be cool if it can be included
> with next upload.
> 

No I didn't really forgot about it.  Whole thing needs some work.

Note that mount/umount in busybox is actually linux-specific.  There
is a patch which is used in debian to implement non-linux mount,
and the result appears to work on kfreebsd (not always, as you
observed).  But that patch has not been applied upstream, -- I
haven't checked yet why.

You change the result of this patch.

But whole approach - trying to canonicalize the path this way -
is most likely wrong.  It shouldn't be needed for the kernel
since it will do path resolution internally anyway during
mount.  And it breaks various pseudo filesystems, incl.
tmpfs, aufs, nfs, etc, where the device part of mount operation
is not really a pathname.

So we're arguing how to fix something that hasn't been accepted
upstream - that's why it needs some more work... :(

Main priority with this for me is to find out what happened with
the original patch upstream adoption.

Thanks,

/mjt



Reply to: