| @@ -161,15 +161,14 @@ JackALSARawMidiDriver::Execute() | |||||
| // times, but, strangely, it actually slowed things down, and made | // times, but, strangely, it actually slowed things down, and made | ||||
| // the code a lot more complicated. | // the code a lot more complicated. | ||||
| // | // | ||||
| // Another alternative would be to use 'epoll' interface. The | |||||
| // problem with the 'epoll' interface is that the timeout | |||||
| // I wonder about using the 'epoll' interface instead of 'ppoll()'. | |||||
| // The problem with the 'epoll' interface is that the timeout | |||||
| // resolution of 'epoll_wait()' is set in milliseconds. We need | // resolution of 'epoll_wait()' is set in milliseconds. We need | ||||
| // microsecond resolution. Without microsecond resolution, we | // microsecond resolution. Without microsecond resolution, we | ||||
| // impose the same jitter as USB MIDI. | // impose the same jitter as USB MIDI. | ||||
| // | // | ||||
| // Of course, a bigger problem is that 'ppoll()' returns later than | |||||
| // the wait time. The problem can be minimized with high precision | |||||
| // timers. | |||||
| // Another problem is that 'ppoll()' returns later than the wait | |||||
| // time. The problem can be minimized with high precision timers. | |||||
| timeout_ptr = &timeout; | timeout_ptr = &timeout; | ||||
| jack_time_t next_time = GetTimeFromFrames(timeout_frame); | jack_time_t next_time = GetTimeFromFrames(timeout_frame); | ||||