[asterisk-dev] [Code Review] Mainly remove bigzaplock, teach core timing about DAHDI_CHUNKSIZE > 8, and friends.

[Shaun Ruffell]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/940/
-----------------------------------------------------------

Review request for Asterisk Developers.


Summary
-------

Collection of *mostly* unrelated changes that have been collecting in various trees.  The individual commits are listed below along with a link to see the commit message that is associated with the commit.  (I'm open for suggestions about the best way to handle these kinds of change sets...mantis issues, individual reviews, emails with links to the mailing lists, combined reviews like this one..??)

http://github.com/sruffell/dahdi-linux/commit/d7797be dahdi: Remove bigzaplock by converting chan_lock to spinlock_t.
http://github.com/sruffell/dahdi-linux/commit/945312c dahdi: process_timers is already called with interrupts disabled.
http://github.com/sruffell/dahdi-linux/commit/23b3e25 dahdi: Rename 'zaptimer' -> 'dahdi_timer'
http://github.com/sruffell/dahdi-linux/commit/3ae1abc Move test for DEFINE_SPINLOCK into include/dahdi/kernel.h
http://github.com/sruffell/dahdi-linux/commit/b1bf126 dahdi: Remove remaining #if 0 locations from dahdi-base.
http://github.com/sruffell/dahdi-linux/commit/211e8de dahdi: Add '+' on KMAKE definition to enable parallel make.
http://github.com/sruffell/dahdi-linux/commit/197a5af make_version: '[[' -> '[' since it's not a bash script.
http://github.com/sruffell/dahdi-linux/commit/4b24970 dahdi: Allow core DAHDI software timing to work when DAHDI_CHUNKSIZE > 8.
http://github.com/sruffell/dahdi-linux/commit/d30a7b9 dahdi: Only disable/enable interrupts once when iterating through channels.


Diffs
-----

  linux/trunk/Makefile 9404 
  linux/trunk/build_tools/make_version 9404 
  linux/trunk/drivers/dahdi/dahdi-base.c 9404 
  linux/trunk/drivers/dahdi/dahdi_dynamic.c 9404 
  linux/trunk/drivers/dahdi/dahdi_dynamic_eth.c 9404 
  linux/trunk/drivers/dahdi/dahdi_dynamic_ethmf.c 9404 
  linux/trunk/drivers/dahdi/dahdi_dynamic_loc.c 9404 
  linux/trunk/drivers/dahdi/dahdi_transcode.c 9404 
  linux/trunk/drivers/dahdi/tor2.c 9404 
  linux/trunk/drivers/dahdi/wct4xxp/base.c 9404 
  linux/trunk/drivers/dahdi/wcte12xp/base.c 9404 
  linux/trunk/drivers/dahdi/wcte12xp/wcte12xp.h 9404 
  linux/trunk/drivers/dahdi/xpp/mmapdrv.c 9404 
  linux/trunk/drivers/dahdi/xpp/parport_debug.c 9404 
  linux/trunk/drivers/dahdi/xpp/xbus-core.c 9404 
  linux/trunk/drivers/dahdi/xpp/xbus-pcm.c 9404 
  linux/trunk/drivers/dahdi/xpp/xpp_usb.c 9404 
  linux/trunk/include/dahdi/kernel.h 9404 

Diff: https://reviewboard.asterisk.org/r/940/diff


Testing
-------

Most of the changes are timing related and I've run them through timing tests with the kernel function_graph tracer and added comments about the results in the commit message.  I started looking at many of these changes when I was testing the chan_list changes that are posted for review.  As I started to open more than 1024 channels the master_span processing time became unusable.  There are more performance changes to come. 


Thanks,

Shaun