RH850/U2A16: Data transfer between four PEs

Hi,

i just started to use RH850/U2A, and need Multi-core collaboration in my project. Recently i had read some official Manual,  knew several ways to do Data Synchronization and memory like local ram and cluster ram. I have few questions:

1.Now there's no rtos in my project, but rtos will be deployed later.  In some samples, they access cluster ram using the direct address, like: *((int*)0xFE000000)=1, but didn't say how to get those address from cluster ram.

Sample:

(1) RH850-U2A/16 How can I use Mutual Exclusion in C-code 

(2) in www.renesas.com/.../rh850u2a-eva-group-san-sample-program-list

I think we won't use cluster ram in such a rough way,  is there any API like "malloc" to use cluster ram?  (I used "malloc" but returned address in local ram) . 

2. I had read material about IPIR(Inter-Processor Interrupts Registers), Barrier-Synchronization, spin_lock  and some sample code of them. I guess these are helpful for Inter-core synchronization,  but what method can actually transmit data between PEs (for example,as far as i know: rpmessage in Ti's soc)? ?

Because in my mind the application scenario is like:  PE0 send the data address to PE1  by physical means in RH850  at the begining,  then PE1 know where's the useful data,  after that, PE0/1 they exchange data just by (shared memory + spin_lock).

Please help, thanks