summaryrefslogtreecommitdiffstatshomepage
path: root/src/mame/audio/hng64.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/mame/audio/hng64.c')
-rw-r--r--src/mame/audio/hng64.c10
1 files changed, 7 insertions, 3 deletions
diff --git a/src/mame/audio/hng64.c b/src/mame/audio/hng64.c
index a185b9fc44f..1f5eb542f48 100644
--- a/src/mame/audio/hng64.c
+++ b/src/mame/audio/hng64.c
@@ -14,12 +14,15 @@ buriki (#)SNK R&D Center (R) HYPER NEOGEO64 Sound Driver Ver 1.15. (#)Copyrig
The earlier revisions appear to have 2 banks of code (there are vectors at the end of the 0x1e0000 block and the 0x1f0000 block)
-Those first two revisions also spam the entire range of I/O ports with values several times on startup causing some unexpected
-writes to the V53 internal registers. The important ones are reinitialized after this however, I'm guessing this is harmless
-on real hardware, as the code flow seems to be correct.
+If the banking setup is wrong then those first two revisions also spam the entire range of I/O ports with values several times
+on startup causing some unexpected writes to the V53 internal registers.
data structures look very similar between all of them
+IRQ mask register on the internal interrupt controller is set to 0xd8
+
+so levels 0,1,2,5 are unmasked
+
*/
@@ -298,6 +301,7 @@ WRITE_LINE_MEMBER(hng64_state::tcu_tm1_cb)
{
// these are very active, maybe they feed back into the v53 via one of the IRQ pins? TM2 toggles more rapidly than TM1
// printf("tcu_tm1_cb %02x\n", state);
+ m_audiocpu->set_input_line(0, state? ASSERT_LINE:CLEAR_LINE);
}
WRITE_LINE_MEMBER(hng64_state::tcu_tm2_cb)