SNIP_CLASSIC_BSWDISTRIBUTIONGUIDE_SECTION_2_8_Crypto_Stack_Distribution

2.8 Crypto-Stack Distribution In order to provide a load distribution amongst different partitions, the different parts of the Crypto-Stack shall be allocated to the different partitions. Hereby it shall be supported that such a partitioning happens on a crypto instance basis, i.e., the crypto driver instances shall be locatable onto different distinct partitions. In order to support such a flexible allocation, the main threads of execution in the SecOC module (namely the respective MainFunctions) can be split into different Main Functions (at least one per partition). Guide to BSW Distribution Furthermore, also the Csm module can be split into different MainFunctions (at least one per partition). In such a setup it is mandatory to assign all the Csm Queues to a dedicated Main Function to define the partition, where the respective Jobs has to be processed. This way the flow through the crypto stack stays within the scope of a single partition and therefore does not require special multi-partition capable means. The inter-partition communication between SecOC and PduR is managed by PduR. In order to manage different timing requirements each MainFunction instance defines its time base individually. This setup brings a couple of constraints for the configuration in order to run the crypto stack in an efficient way. All CsmJobs, to which a SecOC_MainFunction points to, shall be assigned to Csm_MainFunctions which are assigned to the same partition like the respective SecOC_MainFunction. I.e. Csm processing of a Job needs to be handled in the same partition like SecOC processing. In such a setup, Csm module can assume all Csm_<Service> API calls always occur in the partition, where the respective CsmJob is located in (thus no multicore extension for Csm’s service APIs required).