[ofw] RE: [PATCH] make kal_cep_config safe

[Sean Hefty]

>Currently, kal_config_cep assumes that the CID passed in as input is always
>valid.  However, users can destroy CEPs at will using the UAL_DESTROY_CEP IOCTL
>in IBAL.  This means that a CEP could be destroyed while other kernel code is
>attempting to configure it, even though the CID was never explicitly exposed to
>user-mode.
>
>This patch traps an invalid CID in kal_config_cep.

Where does a race occur in the current code?