Tags: vogelpi/opentitan
Tags
WIP [fpga] Stub OBTN to reduce FPGA utilization on CW310 The utilization of the CW310 FPGA is currently ca. 82%, which could disturb SCA experiments. As OTBN is not relevant for the following experiments, this commit stubs it. Signed-off-by: Andreas Kurth <adk@lowrisc.org>
WIP [aes] Re-enable switching off the masking at run time on the CW305 On the CW305, the entropy complex isn't instantiated due to resource constraints. As a result, the magic seed required by the PRNG to produce the all-0 output cannot be loaded into AES. This commit reverts the SecAllowForcingMasks parameter to directly drive the output of the PRNG to 0 in case the FORCE_MASKS bit in the auxiliary control register is set to 1. Signed-off-by: Pirmin Vogel <vogelpi@lowrisc.org>
WIP [aes] Re-enable switching off the masking at run time on the CW305 On the CW305, the entropy complex isn't instantiated due to resource constraints. As a result, the magic seed required by the PRNG to produce the all-0 output cannot be loaded into AES. This commit reverts the SecAllowForcingMasks parameter to directly drive the output of the PRNG to 0 in case the FORCE_MASKS bit in the auxiliary control register is set to 1. Signed-off-by: Pirmin Vogel <vogelpi@lowrisc.org>
[sca, sha3, kmac] Fix SHA3 captures by properly initializing KMAC config Previously, kmac_disable_masking() did not correctly initialize the full KMAC config struct. As a result of using uninitialized memory, it could happen that endianness when reading the state got changed which broke all SHA3 captures. This commit fixes that by creating a static KMAC config which for the sake of documentation, initializes all relevant struct variables to the correct value. kmac_disable_masking() then just modifies the relevant variables. Signed-off-by: Pirmin Vogel <vogelpi@lowrisc.org>
WIP [top_englishbreakfast] Stub out SPI host to save FPGA resources Signed-off-by: Pirmin Vogel <vogelpi@lowrisc.org>
WIP [aes] Optimize SecAllowForcingMasks parameter usage Previously, if enabled this parameter would allow forcing the output of the masking PRNG to 0 requiring 160 Muxes. With this commit, the design is changed to just not advance the PRNG (keep it's output constant). The effect on the masking is the same. Signed-off-by: Pirmin Vogel <vogelpi@lowrisc.org>