You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
// !FIXME! Time since previous, for jitter measurement?
572
571
573
-
// Use gather-based send. This saves one memcpy of every payload
574
-
iovec gather[2];
575
-
gather[0].iov_base = pkt;
576
-
gather[0].iov_len = p - pkt;
577
-
gather[1].iov_base = const_cast<void*>( pChunk );
578
-
gather[1].iov_len = cbChunk;
579
-
580
-
int cbSend = gather[0].iov_len + gather[1].iov_len;
572
+
int cbSend = out.Finish();
581
573
Assert( cbSend <= sizeof(pkt) ); // Bug in the code above. We should never "overflow" the packet. (Ignoring the fact that we using a gather-based send. The data could be tiny with a large header for piggy-backed stats.)
582
574
583
575
// !FIXME! Should we track data payload separately? Maybe we ought to track
0 commit comments