The wp_match_n vector is off by one. Which results in undefined states, at least in simulation.
Signed-off-by: Michael Walle <mich...@walle.cc> --- cores/lm32/rtl/lm32_debug.v | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/cores/lm32/rtl/lm32_debug.v b/cores/lm32/rtl/lm32_debug.v index 2cb21a7..d697fab 100644 --- a/cores/lm32/rtl/lm32_debug.v +++ b/cores/lm32/rtl/lm32_debug.v @@ -167,7 +167,7 @@ wire [0:breakpoints-1]bp_match_n; // Indicates if a h/w instructio reg [`LM32_WPC_C_RNG] wpc_c[0:watchpoints-1]; // Watchpoint enable reg [`LM32_WORD_RNG] wp[0:watchpoints-1]; // Watchpoint address -wire [0:watchpoints]wp_match_n; // Indicates if a h/w data watchpoint matched +wire [0:watchpoints-1]wp_match_n; // Indicates if a h/w data watchpoint matched wire debug_csr_write_enable; // Debug CSR write enable (from either a wcsr instruction of external debugger) wire [`LM32_WORD_RNG] debug_csr_write_data; // Data to write to debug CSR -- 1.7.2.5 _______________________________________________ http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org IRC: #milkymist@Freenode
