Re: NFS bug ?!

Harald Koenig (koenig@tat.physik.uni-tuebingen.de)
Sun, 23 Feb 1997 23:35:17 +0100 (MET)


> > I would like to
> > - know what's the reason and idea that this open() call succeeds for root
>
> Because it came from cache

root can open the file even if nothing is cached at all. just create a file
on the server which "nobody" can't access and try to open it as root on the
client; it will succeed.

Greg Alexander wrote:

> The client specification is fairly obvious...it caches stuff and has normal
> file access rights. The server specification is a kludge that doesn't agree
> with standard.

I had the same "problem" with a SGI IRIX 5.3 NFS server (this was where I
detected it...)

> The client side does an open...it gets the rights and whatnot from
> the server and sees, "hrm...this file is owned by someone, but yer root, so
> I'll let you fiddle" (client should let root do everything). The open
> _ALWAYS_ succeeds, because open is decided by client-side perms and does not
> result in a check to see if the server wants this specific user opening that
> file.

and the NFS server never really checks permissions for the open ?

thanks for answering,

Harald

--
All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \
koenig@tat.physik.uni-tuebingen.de                     ^^^^^       ^^^^^