httpd-bugs mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From bugzi...@apache.org
Subject [Bug 57800] High CPU usage in ap_core_output_filter
Date Fri, 10 Apr 2015 21:51:58 GMT
https://bz.apache.org/bugzilla/show_bug.cgi?id=57800

Yann Ylavic <ylavic.dev@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ylavic.dev@gmail.com

--- Comment #1 from Yann Ylavic <ylavic.dev@gmail.com> ---
(In reply to szotsaki from comment #0)
[..]
> 
> An other thread shows the same with a bit different top function call line:
> #0  0x00007fc7ca3640fa in ap_core_output_filter (f=0x7fc7bf4f1690,
> new_bb=0x7fc7bf4f1928) at core_filters.c:493
>         c = 0x7fc7bf4f1290
>         net = 0x7fc7bf4f1648
>         ctx = 0x7fc7bf4f1748
>         bb = 0x7fc7bf4f1928
>         bucket = <optimized out>
>         next = 0x7fc7bf4e9048
>         flush_upto = 0x7fc7bf4e9048
>         bytes_in_brigade = 0
>         non_file_bytes_in_brigade = 0
>         eor_buckets_in_brigade = 1
>         morphing_bucket_in_brigade = 0
>         rv = <optimized out>

Two threads using simultaneously the addresses, really?

BTW, next and flush_upto shouldn't hold the same value on a single thread
either, that points an infinite loop.

Which linux distribution is it?
Which compiler (version) was used to build httpd and mainly the lib APR?
Which options (CFLAGS) if that's available somewhere (dev package)?
Can you disassemble the loop starting at line 489?

It looks like APR rings do not play well here.

-- 
You are receiving this mail because:
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org


Mime
View raw message