 No.03173
Luigi30 Tester
Nov 30, 2008, 06:43
|
From the diff:- cpunum_set_input_line_and_vector(machine, 1, 0, HOLD_LINE, 0x30); + cpu_set_input_line_and_vector(device, 0, HOLD_LINE, 0x30); It's passing sound IRQs to the wrong CPU and hanging up because of it, I think? It won't take space->machine->cpu[1]. |
|---|---|
 No.03176
Mamesick Senior Tester
Nov 30, 2008, 07:13
|
The CPU number should be already stored in "device" object so I wouldn't say a stupidity but it could be a core bug. (cpunum is always zero?) |
|
No.03183
Luigi30 Tester
Nov 30, 2008, 18:06
edited on: Nov 30, 2008, 18:19
|
I have it printing "MSM interrupt" to the error log every time it triggers one in the driver. When it freezes up, MAME dumps a couple kilobytes of "MSM interrupt" to the error log in a few seconds. u3 to u4: - cpunum_set_input_line_and_vector(machine, 1, 0, HOLD_LINE, 0x38); + cpu_set_input_line_and_vector(machine->cpu[cpunum], 0, HOLD_LINE, 0x38); Changing cpu[cpunum] to cpu[1] as it was in u3 seems to fix the problem. It was trying to send the IRQ to the active CPU (0, executing the program) instead of CPU 1. Fix sent. |
 No.03205
Fujix Administrator
Dec 3, 2008, 08:51
|
Fix submitted to Devs. |