Re: +core_pattern-set-core-helpers-root-and-namespace-to-crashing-process .patchadded to -mm tree

[Neil Horman]

On Wed, Dec 19, 2012 at 05:22:30PM +0100, Oleg Nesterov wrote:
> On 12/18, Neil Horman wrote:
> >
> > On Tue, Dec 18, 2012 at 09:06:04PM +0100, Oleg Nesterov wrote:
> > > > Perhaps its best just to restrict this patch to adjusting the root fs location
> > > > for the chroot case.
> > >
> > > Probably... at least for the start.
> > >
> > > BTW. Of course this is subjective, but personally I think that "||"
> > > looks strange. Perhaps it would be better to add something like
> > > --croot argument?
> > >
> > The || is ambiguous with its simmilarity to a shell 'or' command,
> 
> Ah, I didn't mean this.
> 
> I meant, this is not flexible. We can implement --croot, then we can
> (may be) add --switch_ns. If you use "||" now for chroot, what will you
> do for switch_ns?
> 
> > but I don't
> > think the --croot argument is much better on that front, as that then becomes
> > ambiguous with arguments supplied to the pipe reader directly.
> 
> Not sure I understand... why?
> 
> > The token should
> > be leading the pipe_reader string in core_pattern to indicate a change in
> > environment independent of the executable path.
> 
> Do you mean that || at the front is more "visible" ? True, but I am
> not sure this is that important.
> 
All I really meant was that placing the token for croot at the front of the
pattern would be more readable as it disambiguates its meaning from the rest of
the command.

> But I won't insist.
> 
I wouldn't worry about it.  After looking over the changes Eric just had merged
for 3.8 I'm becomming more convinced that this isn't really needed anymore. With
his changes, we can migrate a process between all available namespaces
dynamically in user space.  With that functionality we can just write a setns
administrative utility to make this all work. I've started that work here:
http://marc.info/?l=util-linux-ng&m=135594402801895&w=2

all thats arguably left after that is to make core_pattern a per-container
(possibly per mount namespace?).  My patch doesn't do that. But its something I
can look into.

Neil
 
> Oleg.
> 
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/