[
https://issues.apache.org/jira/browse/TS-2939?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14070406#comment-14070406
]
ASF subversion and git services commented on TS-2939:
-----------------------------------------------------
Commit fd7365db20cda8d0ef120723f22baed0afb64288 in trafficserver's branch
refs/heads/master from [~nottheoilrig]
[ https://git-wip-us.apache.org/repos/asf?p=trafficserver.git;h=fd7365d ]
TS-2939 Metalink: Fix crash when checking the digest of a file that wasn't
cacheable
> Metalink: Fix crash when checking the digest of a file that wasn't cacheable
> ----------------------------------------------------------------------------
>
> Key: TS-2939
> URL: https://issues.apache.org/jira/browse/TS-2939
> Project: Traffic Server
> Issue Type: Bug
> Reporter: Jack Bates
> Assignee: Jack Bates
> Fix For: sometime
>
>
> TSCacheRead() is reentrant particularly in case of a cache miss so its
> handler might execute sooner than expected particularly when the Metalink
> plugin is checking the digest of a file that wasn't cacheable. This
> TSCacheRead() handler will clean up an earlier TSVConnRead() buffer but
> CacheVC::openReadMain() will continue operating on the buffer and crash.
> Close the virtual connection to stop it from operating on the buffer.
--
This message was sent by Atlassian JIRA
(v6.2#6252)