[ 
https://issues.apache.org/jira/browse/TS-4554?focusedWorklogId=26125&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-26125
 ]

ASF GitHub Bot logged work on TS-4554:
--------------------------------------

                Author: ASF GitHub Bot
            Created on: 01/Aug/16 12:14
            Start Date: 01/Aug/16 12:14
    Worklog Time Spent: 10m 
      Work Description: Github user masaori335 commented on the issue:

    https://github.com/apache/trafficserver/pull/830
  
    Yes, we don't have to follow it, but we should follow it as much as 
possible. Because it could be happen that images are sent before CSS/JS, if we 
don't follow it.
    
    How about use SETTINGS_MAX_CONCURRENT_STREAMS  as the max depth? It looks 
appropriate and we don't need new configures.
    
    I'm seeing that with latest Chrome ( 51.0.2704.106 ). Chrome had that bug 
in Feb and reverted the changes. IIUC, the bug is about the order of streams 
not depth of streams. Now (maybe from Jun), it looks like they fixed the order 
and rolled it out.


Issue Time Tracking
-------------------

    Worklog Id:     (was: 26125)
    Time Spent: 1h 40m  (was: 1.5h)

> ASAN crash (stack overflow) with H2 priorities
> ----------------------------------------------
>
>                 Key: TS-4554
>                 URL: https://issues.apache.org/jira/browse/TS-4554
>             Project: Traffic Server
>          Issue Type: Improvement
>          Components: HTTP/2
>    Affects Versions: 7.0.0
>            Reporter: Leif Hedstrom
>            Assignee: Masaori Koshiba
>             Fix For: 7.0.0
>
>          Time Spent: 1h 40m
>  Remaining Estimate: 0h
>
> I'm seeing (truncated):
> {code}
> ASAN:SIGSEGV
> =================================================================
> ==11178==ERROR: AddressSanitizer: stack-overflow on address 0x2aab63633ff0 
> (pc 0x0000007ddaa9 bp 0x2aab63634050 sp 0x2aab63633ff0 T6)
>     #0 0x7ddaa8 in 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int) 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:134
>     #1 0x7ddaa8 in 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int) 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:140
>     #2 0x7ddaa8 in 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int) 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:140
> .
> .
> .
>     #2261 0x7ddc34 in 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int) 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:140
>     #2262 0x7ddc34 in 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int) 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:140
>     #2263 0x7ddc34 in 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int) 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:140
> SUMMARY: AddressSanitizer: stack-overflow 
> /usr/local/src/trafficserver/proxy/http2/Http2DependencyTree.h:134 
> Http2DependencyTree<Http2Stream*>::_find(Http2DependencyTree<Http2Stream*>::Node*,
>  unsigned int)
> Thread T6 ([ET_NET 5]) created by T0 ([ET_NET 0]) here:
>     #0 0x2aab5b0d50c4 in __interceptor_pthread_create 
> ../../../../libsanitizer/asan/asan_interceptors.cc:179
>     #1 0xcfdb99 in ink_thread_create ../../lib/ts/ink_thread.h:147
>     #2 0xcfdb99 in Thread::start(char const*, unsigned long, void* 
> (*)(void*), void*) 
> /usr/local/src/trafficserver/iocore/eventsystem/Thread.cc:99
>     #3 0xd0562e in EventProcessor::start(int, unsigned long) 
> /usr/local/src/trafficserver/iocore/eventsystem/UnixEventProcessor.cc:140
>     #4 0x497b59 in main /usr/local/src/trafficserver/proxy/Main.cc:1746
>     #5 0x2aab5ee11b14 in __libc_start_main (/lib64/libc.so.6+0x21b14)
> ==11178==ABORTING
> {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to