vmamidi commented on a change in pull request #7386:
URL: https://github.com/apache/trafficserver/pull/7386#discussion_r542824574
##########
File path: proxy/http/HttpTunnel.cc
##########
@@ -1336,16 +1336,7 @@ HttpTunnel::consumer_handler(int event,
HttpTunnelConsumer *c)
c->producer->read_success = true;
// Go ahead and clean up the producer side
if (p->alive) {
- p->alive = false;
- if (p->read_vio) {
- p->bytes_read = p->read_vio->ndone;
- } else {
- p->bytes_read = 0;
- }
- if (p->vc != HTTP_TUNNEL_STATIC_PRODUCER) {
- // Clear any outstanding reads
- p->vc->do_io_read(nullptr, 0, nullptr);
- }
+ producer_handler(VC_EVENT_READ_COMPLETE, p);
Review comment:
At this point, we may still have still have data left to read from the
producer, right? How can we send an VC_EVENT_READ_COMPLETE to the producer?
##########
File path: iocore/net/UnixNetVConnection.cc
##########
@@ -91,7 +91,7 @@ read_signal_and_update(int event, UnixNetVConnection *vc)
case VC_EVENT_ACTIVE_TIMEOUT:
case VC_EVENT_INACTIVITY_TIMEOUT:
Debug("inactivity_cop", "event %d: null read.vio cont, closing vc %p",
event, vc);
- Warning("read: Closing orphaned vc %p", vc);
+ Warning("read: Closing orphaned vc vc=%p cont=%p event=%d", vc,
vc->read.vio.cont, event);
Review comment:
Is orphaned VC a new name for inactivity cop? Why are we throwing a
warning for a typical inactivity cop scenario?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
