Re: [PATCH][RFC] fs/exec.c: provide the correct process pid to thepipe helper

[Will Drewry]

On Fri, Sep 17, 2010 at 8:26 AM, Andi Kleen <andi@xxxxxxxxxxxxxx> wrote:
> On Thu, 16 Sep 2010 13:59:59 -0500
> Will Drewry <wad@xxxxxxxxxxxx> wrote:
>
>> format_corename uses task_tgid_vnr to provide the numeric pid of a
>> core-dumping process.  For file-based coredumps, this is perfectly
>> satisfactory.  However, when the core_pattern contains a pipe, the
>> substituted PID is invalid in the namespace of the core_pattern pipe
>> helper, the init namespace.
>
> Nasty problem. I wonder how many more similar problems name spaces
> have introduced.
>
> But wouldn't it be better to place the helper into the name space(s)
> of the executed process? I guess it would risk breaking some existing
> set ups, but it seem like the cleanest solution to me. If you want
> to move the core dump out of the name space you could still
> use a named pipe or something like that with someone outside listening.
>
> That would also fix the net namespace problem you mentioned

I agree.  This was what I did first, but I wasn't confident I'd done
it right.  I'll go ahead and post the two patches I used to accomplish
it.   I'm happy to spend some time polishing them until they are
functionally correct, if everyone is happy enough with the approach.

more to come,
will
--
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/