[Pkg-sysvinit-devel] Re: [uml-devel] [Pkg-uml-pkgs] How is user-mode-linux using /dev/shm/?

[Mattia Dongili]

On Thu, Sep 28, 2006 at 05:51:43PM -0400, Jeff Dike wrote:
> On Thu, Sep 28, 2006 at 09:47:27PM +0200, Mattia Dongili wrote:
> > On Thu, Sep 28, 2006 at 08:14:32PM +0200, Petter Reinholdtsen wrote:
> > > To make some writable tmpfs available to those in need of such system,
> > > and to avoid using /dev/shm/ which is reserved for the shm-functions,
> > > I just uploaded sysvinit version 2.86.ds1-26.  It will mount a tmpfs
> > > on /lib/init/rw/ that can be used instead.  If /lib/init/rw/.ramfs
> > > exist, that mount point is a tmpfs.  I'm not sure if this last change
> > > will make it into Etch or not, but I hope so, to solve any problems
> > > with packages previously using /dev/shm/ as a generic tmpfs file
> > > system.
> > 
> > Well, I'd actually prefer if you could remove the noexec flag from
> > /dev/shm. I understand the security reasons given in the bugreport but
> > I'd prefer avoid having to deal with one more Debian-only (is it?)
> > thing given the soon to come general freeze. 
> 
> UML needs a non-noexec place to keep a file that will be used as its
> physical memory.  A tmpfs mount is greatly preferred for performance
> reasons as well as tmpfs being the only filesystem supporting
> MADV_REMOVE, which is used for memory hotplug.

Thinking of a Debian only fix for the above, does simply playing with
default_tmpdir in arch/um/os-Linux/mem.c (probably in which_tmpdir())
suffice to use /lib/init/rw/.ramfs as default?

(Yes I should try it... will try to find some time next week)

Thanks
-- 
mattia
:wq!