Remove unnecessary tcp state transition call.
ClosedPublic
Actions

Authored by hiren on Aug 21 2015, 9:09 PM.

Details

Reviewers

markj
np
gnn
rwatson

Commits

rS287830: Remove unnecessary tcp state transition call.

Summary

This call now also tracks state transitions for DTRACE and without this change
it'd report state transition from TCPS_CLOSING to TCPS_CLOSING.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

hiren updated this revision to Diff 8109.Aug 21 2015, 9:09 PM

hiren retitled this revision from to Remove unnecessary tcp state transition call..

hiren updated this object.

hiren edited the test plan for this revision. (Show Details)

hiren added reviewers: np, markj, rwatson.

hiren added a subscriber: network.

Herald added a subscriber: imp. · View Herald TranscriptAug 21 2015, 9:09 PM

Hum, I am failing to understand/find where we set state to TCPS_CLOSED before we enter tcp_usrclosed().

bz added a reviewer: gnn.Aug 21 2015, 10:10 PM

kbowling added a subscriber: kbowling.Aug 22 2015, 12:01 AM

In D3451#70430, @hiren wrote:

Hum, I am failing to understand/find where we set state to TCPS_CLOSED before we enter tcp_usrclosed().

It's not clear to me either.

This revision is now accepted and ready to land.Aug 23 2015, 5:07 PM

jch added a subscriber: jch.Aug 24 2015, 9:44 AM

Oops. I just took a second look and noticed the FALLTHROUGH above. I presume that's the reason for the assignment.

It looks like it should be ok to move the call up, right before the fallthrough.

This revision now requires changes to proceed.Aug 24 2015, 7:54 PM

Good catch by @markj.
Move the call to track TCPS_LISTEN -> TCPS_CLOSED transition correctly.

Remove unrelated tcp_timer.c changes.

In D3451#71006, @markj wrote:

Oops. I just took a second look and noticed the FALLTHROUGH above. I presume that's the reason for the assignment.

It looks like it should be ok to move the call up, right before the fallthrough.

Good catch. :-)

I am still unsure of how we handle TCPS_CLOSE correctly. i.e in my instrumentation, I never see state transition to closed state.
May be @rwatson can help here.