[RFC][PATCH 0/6] Allow unsharing pid namespace

Eric W. Biederman ebiederm at xmission.com
Sun Mar 11 04:29:19 PDT 2007


sukadev at us.ibm.com writes:

> From: Sukadev Bhattiprolu <sukadev at us.ibm.com>
> Subject: [RFC][PATCH 0/6] Allow unsharing pid namespace.
>
> This patchset defines a struct pid_nr and uses this to allow
> processes to unshare their pid namespace.
>
> struct pid_nr will hold [pid value, namespace] pair for each
> struct pid* in use. The pid value used will be different for
> each namespace the struct pid* is in use.
>
> A list of pid_nrs will be added to struct pid and used by the
> kernel when using pid values(through pid_nr()).
>
>
> [PATCH 1/6]:
> 	Add struct pid_nr and associated helper functions.
>
> [PATCH 2/6]:
> 	Reimplement the pid_nr() function using struct pid_nr.
> 	i.e pid_nr() should return the pid_t of a process
> 	depending on the pid namespace of the caller.
>
> [PATCH 3/6]:
> 	Attach a list of struct pid_nrs to the struct pid.
>
> [PATCH 4/6]:
> 	Statically allocate/initialize a struct pid_nr for
> 	swapper process.
>
> [PATCH 5/6]:
> 	Define helper functions used to unshare pid namespace.
>
> [PATCH 6/6]:
> 	Implement unsharing of pid namespace.

You are missing anything to do with fixing the lookup in the pid hash
table.  That is a huge omission and unacceptable.

Eric



More information about the Containers mailing list