[ 
https://issues.apache.org/jira/browse/TS-559?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

John Plevyak resolved TS-559.
-----------------------------

    Resolution: Fixed

Committed revision 1038809.


> Segfault when forward proxy HTTPS to a host not listening on port 443
> ---------------------------------------------------------------------
>
>                 Key: TS-559
>                 URL: https://issues.apache.org/jira/browse/TS-559
>             Project: Traffic Server
>          Issue Type: Bug
>          Components: Network
>            Reporter: Leif Hedstrom
>            Priority: Blocker
>             Fix For: 2.1.5
>
>
> When doing a forward proxy setup, and a request like this is processed:
> GET https://127.0.0.1 HTTP/1.0
> Host: 127.0.0.1
> where 127.0.0.1:443 is not listened on, then two things will happen:
> 1) The client connection "hangs".
> 2) When client disconnects, we segfault (see trace below).
> It is understood that ATS does not support forwarding proxy of HTTPS like 
> this, instead we support CONNECT which is what (most) clients will use to 
> forward proxy HTTPS. But, obviously it's not acceptable to segfault under 
> these conditions. Note that this problem is only there if the server is setup 
> to handle forward proxy, i.e. remap.required is set to '0'.
> #0  0x0000000000000000 in ?? ()
> #1  0x0000000000519cd7 in HttpServerSession::do_io_close 
> (this=0x7fffe01ccf20, alerrno=<value optimized out>)
>     at HttpServerSession.cc:128
> #2  0x000000000052e8b8 in cleanup_entry (this=0x7fffec6c7c30) at HttpSM.cc:261
> #3  cleanup_all (this=0x7fffec6c7c30) at HttpSM.cc:272
> #4  HttpSM::kill_this (this=0x7fffec6c7c30) at HttpSM.cc:6060
> #5  0x000000000052f3f8 in HttpSM::main_handler (this=0x7fffec6c7c30, 
> event=104, data=0x7fffdc013740)
>     at HttpSM.cc:2550
> #6  0x000000000068a7c1 in handleEvent (event=<value optimized out>, 
> nh=0x7ffff526d638, vc=0x7fffdc013660)
>     at ../../iocore/eventsystem/I_Continuation.h:147
> #7  read_signal_and_update (event=<value optimized out>, nh=0x7ffff526d638, 
> vc=0x7fffdc013660)
>     at UnixNetVConnection.cc:146
> #8  read_signal_done (event=<value optimized out>, nh=0x7ffff526d638, 
> vc=0x7fffdc013660)
>     at UnixNetVConnection.cc:176
> #9  0x000000000068d20d in read_from_net (nh=0x7ffff526d638, 
> vc=0x7fffdc013660, thread=<value optimized out>)
>     at UnixNetVConnection.cc:308
> #10 0x0000000000684460 in NetHandler::mainNetEvent (this=0x7ffff526d638, 
> event=<value optimized out>, 
>     e=<value optimized out>) at UnixNet.cc:411
> #11 0x00000000006b0494 in handleEvent (this=0x7ffff526c010, e=0xf5dd70, 
> calling_code=5) at I_Continuation.h:147
> #12 EThread::process_event (this=0x7ffff526c010, e=0xf5dd70, calling_code=5) 
> at UnixEThread.cc:149
> #13 0x00000000006b0e23 in EThread::execute (this=0x7ffff526c010) at 
> UnixEThread.cc:271
> #14 0x00000000006af31a in spawn_thread_internal (a=0xf50720) at Thread.cc:85
> #15 0x00007ffff7bca761 in start_thread (arg=0x7ffff4e67710) at 
> pthread_create.c:301
> #16 0x00007ffff63544fd in clone () at 
> ../sysdeps/unix/sysv/linux/x86_64/clone.S:115

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to