[ https://issues.apache.org/jira/browse/TS-4766?focusedWorklogId=26660&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-26660 ]
ASF GitHub Bot logged work on TS-4766: -------------------------------------- Author: ASF GitHub Bot Created on: 19/Aug/16 16:23 Start Date: 19/Aug/16 16:23 Worklog Time Spent: 10m Work Description: GitHub user shinrich opened a pull request: https://github.com/apache/trafficserver/pull/881 TS-4766: HTTP/2 frame corruption fix. Moved handler reset into the do_complete_frame_read worker. You can merge this pull request into a Git repository by running: $ git pull https://github.com/shinrich/trafficserver ts-4766 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/trafficserver/pull/881.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #881 ---- commit bd69020f2febecd500edcafc9bff7ad2f0cdb126 Author: Susan Hinrichs <shinr...@ieee.org> Date: 2016-08-19T16:20:23Z TS-4766: HTTP/2 frame corruption fix. ---- Issue Time Tracking ------------------- Worklog Id: (was: 26660) Time Spent: 10m Remaining Estimate: 0h > HTTP/2 packet corruption introduced by TS-4717 fix > -------------------------------------------------- > > Key: TS-4766 > URL: https://issues.apache.org/jira/browse/TS-4766 > Project: Traffic Server > Issue Type: Bug > Components: HTTP/2 > Reporter: Susan Hinrichs > Assignee: Susan Hinrichs > Fix For: 7.0.0 > > Time Spent: 10m > Remaining Estimate: 0h > > We observed problems with large uploads over HTTP/2 via Firefox. Tracked it > down to ATS sending a GOAWAY frame early because the frame appeared to be > corrupted. The problem was introduced in the restructuring in TS-4717 to get > rid of the unbounded recursion. The logic will return with one of two > handlers set. One to deal with the case of starting to read a fresh frame > and another to deal with the case of finishing reading a partially read > frame. There is a path where we finish reading a frame but don't reset the > handler to the new frame case. When the next data arrives the wrong handler > starts reading it resulting in ATS thinking it has received a malformed frame. -- This message was sent by Atlassian JIRA (v6.3.4#6332)