[Ecls-list] Locking strategy (best so far)

[Matthew Mondor]

On Fri, 6 Apr 2012 16:24:49 +0200
Juan Jose Garcia-Ripoll <juanjose.garciaripoll at googlemail.com> wrote:

> On Fri, Apr 6, 2012 at 2:32 PM, Matthew Mondor <mm_lists at pulsar-zone.net>wrote:
> 
> > I confirm that these changes allow the HTTPd to run here as well,
> > thanks!  However, I didn't stress-test it much or try performance tests
> > with it yet, but that should happen soon.
> 
> 
> I am now running some stress tests and trying to hit some corner cases.
> There are already some subtle race conditions in the lisp part of the code.
> Hopefully will be easy to fix...

With some tests outside of Slime as I have trouble with Slime on recent
HEAD:

I can use ab(8) a few runs of -c16 -n1000 with good performance (2k+
requests served per second).  After a few runs, one of the threads (the
first try it was the children manager one, at the second try it was one
of the server thread) gets stuck into a busy loop inside the GC.  I
also saw a bunch of stack overflow errors at the REPL.

A backtrace of the busy-looping thread:

#0  handle_or_queue (the_env=0x7f7ff7fc6000, signal_code=0x7f7ff7a3ccf8, code=11) at /home/mmondor/work/ecl-git/ecl/src/c/unixint.d:500
#1  <signal handler called>
#2  0x00007f7ff5e1211c in GC_mark_from () from /usr/pkg/lib/libgc.so.1
#3  0x00007f7ff5e135fd in GC_mark_some () from /usr/pkg/lib/libgc.so.1
#4  0x00007f7ff5e0bc3f in GC_stopped_mark () from /usr/pkg/lib/libgc.so.1
#5  0x00007f7ff5e0c2ea in GC_try_to_collect_inner () from /usr/pkg/lib/libgc.so.1
#6  0x00007f7ff5e0cc20 in GC_collect_or_expand () from /usr/pkg/lib/libgc.so.1
#7  0x00007f7ff5e1088c in GC_alloc_large () from /usr/pkg/lib/libgc.so.1
#8  0x00007f7ff5e11207 in GC_generic_malloc_ignore_off_page () from /usr/pkg/lib/libgc.so.1
#9  0x00007f7ff7774089 in ecl_alloc (n=16384) at /home/mmondor/work/ecl-git/ecl/src/c/alloc_2.d:717
#10 0x00007f7ff776403e in cl__make_hash_table (test=<optimized out>, size=<optimized out>, rehash_size=0x2b38280, rehash_threshold=<optimized out>)
    at /home/mmondor/work/ecl-git/ecl/src/c/hash.d:888
#11 0x00007f7ff7764312 in cl_make_hash_table (narg=<optimized out>) at /home/mmondor/work/ecl-git/ecl/src/c/hash.d:745
#12 0x00007f7ff281031e in L47make_http_request (narg=0) at test-httpd.c:5769
#13 0x00007f7ff282125e in L68http_request_parse (V1=0x10270d71, V2=0x1ea58c0) at test-httpd.c:6845
#14 0x00007f7ff281a407 in L85http_serve (V1=0xb7d5f90) at test-httpd.c:9136
#15 0x00007f7ff241ebd3 in L57accept_loop_thread (V1=0xc409e40) at ecl-mp-server.c:6497
#16 0x00007f7ff771d04d in cl_apply (narg=<optimized out>, fun=0x1f27200, lastarg=<optimized out>) at /home/mmondor/work/ecl-git/ecl/src/c/eval.d:166
#17 0x00007f7ff7770842 in thread_entry_point (arg=<optimized out>) at /home/mmondor/work/ecl-git/ecl/src/c/threads/process.d:256
#18 0x00007f7ff5e17fb5 in GC_inner_start_routine () from /usr/pkg/lib/libgc.so.1
#19 0x00007f7ff5e14b60 in GC_call_with_stack_base () from /usr/pkg/lib/libgc.so.1
#20 0x00007f7ff7209dbd in pthread__create_tramp (cookie=0x7f7fea800000) at /usr/src/lib/libpthread/pthread.c:492
#21 0x00007f7ff6a76ef0 in ___lwp_park50 () from /usr/lib/libc.so.12
#22 0x00007f7feac00000 in ?? ()
#23 0x00007f7ff7ff60c0 in ?? ()
#24 0x0000000111110001 in ?? ()
#25 0x0000000033330003 in ?? ()
#26 0x0000000000000000 in ?? ()

Thanks,
-- 
Matt