[Devel] Re: [PATCH 2/2] exec: move core_pattern pipe helper into the crashing namespace

[Oleg Nesterov]

On 09/17, Will Drewry wrote:
>
> Instead, this change implements the more complex option two.  It
> migrates the ____call_usermodehelper() thread into the same namespaces
> as the dumping process.  It does not assign a pid in that namespace so
> the collector will appear to be pid 0.

Hmm... You mean, it won't visible in that namespace? Afacis, it should
have the correct pid in the init ns, no?

I am a bit worried task_active_pid_ns() != nsproxy->pid_ns, but perhaps
this is OK... Say, sys_getpid() returns 0, strange.

> +		/* Run the core_collector in the crashing namespaces */
> +		if (copy_namespaces_unattached(0, current,
> +			&pipe_params.nsproxy, &pipe_params.fs)) {
> +			printk(KERN_WARNING "%s failed to copy namespaces\n",
> +			       __func__);
> +			argv_free(helper_argv);
> +			goto fail_dropcount;
> +		}

This looks overcomplicated to me, or I missed something.

I do not understand why should we do this beforehand, and why we need
copy_namespaces_unattached().

Can't you just pass current to umh_pipe_setup() (or another helper) as
the argument? Then this helper can copy ->fs and ->nsproxy itself.

In fact, I do not understand why create_new_namespaces() is used. It
is called with flags == 0 anyway, can't we just do

	ns = coredumping_task->nsproxy;
	get_nsproxy(ns);	
	switch_task_namespaces(current, ns);

?

Oleg.

_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers