commit 8e342c9c65c602b2fc85c80b8a75c5d515c48059
author Yong Li <yong.b.li@linux.intel.com> Tue Nov 19 13:58:01 2019 +0800
committer Vernon Mauery <vernon.mauery@linux.intel.com> Mon Dec 16 19:57:59 2019 +0000
tree 40e5c9f112a3e8e7deb1b04a71c0d539b5cd2f23
parent 0f63e01ce6bb11920d78d999267558500ca9a272

Do not stop session in deactivate payload call

According to section 24.2 of the spec, the Deactivate Payload command
should not cause the session to be terminated.

Also during SOL looptest, there is only one time startSession call,
but multiple stopSessions calls.
This causes the looptest will fail if there is any new session comes in,
needs to remove the stopSession call.

Tested:
Start the loop test in a terminal:
ipmitool -H $BMCIP  -Uroot -P 0penBmc -I lanplus sol looptest 500 200

Then start another session in another terminal:
ipmitool -H $BMCIP  -Uroot -P 0penBmc -I lanplus raw 6 1

The looptest still works

sol on web still works
ssh -p2200 still works

Signed-off-by: Yong Li <yong.b.li@linux.intel.com>
Change-Id: I15729fa220fae18adb318bdae98d5d9ccfec9ae1

command/payload_cmds.cpp

diff --git a/command/payload_cmds.cpp b/command/payload_cmds.cpp
index 219f2d5..c5d64fb 100644
--- a/command/payload_cmds.cpp
+++ b/command/payload_cmds.cpp
@@ -173,13 +173,6 @@
              */
             return outPayload;
         }
-
-        auto check =
-            std::get<session::Manager&>(singletonPool).stopSession(sessionID);
-        if (!check)
-        {
-            response->completionCode = IPMI_CC_UNSPECIFIED_ERROR;
-        }
     }
     catch (std::exception& e)
     {