This removes any form of initialization for the sram simulation in case the default "none" is set for SimInit. This allows for memory initialization from a tb by hierarchically accessing the module in an initial block, as is done in HERO, see also https://github.com/pulp-platform/hero/commit/c88efca6c5ba6666e9a949956a3020c2bea962ef. In case this is not done, the default behavior should not change drastically, as the initial value will still be 'X, however will not be reset on rst_ni. If the current behavior is desired, a value other than "none", "zeros", "ones", or "random" should be set for SimInit.
It looks to me like this PR is implementing the same behavior as #23. However, having the if condition within the always_ff block makes the tc_sram no longer synthesizable nor verilatable. Right?
This removes any form of initialization for the sram simulation in case the default "none" is set for SimInit. This allows for memory initialization from a tb by hierarchically accessing the module in an initial block, as is done in HERO, see also https://github.com/pulp-platform/hero/commit/c88efca6c5ba6666e9a949956a3020c2bea962ef. In case this is not done, the default behavior should not change drastically, as the initial value will still be
'X
, however will not be reset on rst_ni. If the current behavior is desired, a value other than "none", "zeros", "ones", or "random" should be set for SimInit.