[rtl] FPGA bkdr fix

[gautschimi]

The enum was previously defined as
typedef enum logic [$clog2(NumBkdrTgts)-1:0] and NumBkdrTgts was 12. Hence, the enum was 4bit wide.

The index check (https://github.com/lowRISC/opentitan/blob/master/hw/ip/bkdr_loader/rtl/bkdr_loader.sv#L284)
assign tgt_idx_err = !(bkdr_idx_e'(reg2hw.control.target_idx.q) inside {BkdrValidTgts});

It first casts the 8b target_idx signal to the enum size (dropping 4 bits). And then checks if the value is inside the valid targets. Hence, a target_idx of 13 was seen as a legal value. Actually vivado was interpreting it as always false and was thus always reporting an invalid target.

[gautschimi]

The first commit of this PR was already merged as part of #30543

[a-will]

I'm not sure I understand why this enum even exists. It looks like this became a top-specific IP instead of something generic.

This should probably become templated if the CSR layout is so tied to a specific top.

[vogelpi]

Thanks, LGTM!

[vogelpi]

The e2e_bootstrap_disabled_fpga_cw340_rom_with_fake_keys FPGA test failed, but I've seen this failing for other PRs as well. The failure is not related to this change.