This off-by-one was clobbering the stack since introduced in
a803c0db4d58.
However it only set one bit, and that generally appears to have been set
already, so this rarely affected anything. But when it did, it affected
a pointer, causing a segfault.
readcount = MIN(devc->max_samples / num_channels, devc->limit_samples) / 4;
memset(trigger_config, 0, 16);
- trigger_config[devc->num_stages - 1] |= 0x08;
+ trigger_config[devc->num_stages] |= 0x08;
if (devc->trigger_mask[0]) {
delaycount = readcount * (1 - devc->capture_ratio / 100.0);
devc->trigger_at = (readcount - delaycount) * 4 - devc->num_stages;