[asterisk-dev] [Code Review] Convert ast_channel over to astobj2

[Mark Michelson]

> On 2009-03-20 03:44:07, vadim wrote:
> > /trunk/main/autochan.c, line 1
> > <http://reviewboard.digium.com/r/203/diff/1/?file=3319#file3319line1>
> >
> >     I would suggest that instead of using one global autochan list to we could use autochan list head in the channel structure itself.
> >     
> >     This will surely be more efficient
> 
>  wrote:
>     I'll let Mark comment on this, since the autochan stuff was his work.
> 
>  wrote:
>     I agree with Vadim on this point. The main reason I didn't do that is because in general, we try not to change the ast_channel structure for ABI compatibility. I suppose that I could go with the datastore method of doing this, though. Put a datastore on the channel which contains a list of the autochan pointers which point to the channel. I'll get autochan.c corrected and get it committed to the ast_channel_ao2 branch. I'll ping the review when I have finished.
> 
>  wrote:
>     Regarding ABI, this patch already breaks ABI by removing the ast_mutex_t from ast_channel.  There is also the issue8824 patch that is going to break ABI in 1.6.3, as well.
> 
>  wrote:
>     Oh heck yeah! That means I can just make this change directly to the ast_channel, then. No need for esoteric datastore magic!

All right, the change is applied. Thanks for the idea, Vadim!


- Mark


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.digium.com/r/203/#review582
-----------------------------------------------------------


On 2009-03-20 09:47:09, Russell Bryant wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.digium.com/r/203/
> -----------------------------------------------------------
> 
> (Updated 2009-03-20 09:47:09)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> This patch converts handling of the ast_channel data structure over to the astobj2 framework.  There is a lot that could be said about this, but the patch is a bit improvement for performance, stability, code maintainability, and ease of future code development.
> 
> The channel list is no longer an unsorted linked list.  The main container for channels is an astobj2 hash table.  All of the code related to searching for channels or iterating active channels has been rewritten.  Let n be the number of active channels.  Iterating the channel list has gone from O(n^2) to O(n).  Searching for a channel by name went from O(n) to O(1).  Searching for a channel by extension is still O(n), but uses a new method for doing so, which is more efficient.
> 
> The ast_channel object is now a reference counted object.  The benefits here are plentiful.  Some benefits directly related to issues in the previous code include:
> 
> 1) When threads other than the channel thread owning a channel wanted access to a channel, it had to hold the lock on it to ensure that it didn't go away.  This is no longer a requirement.  Holding a reference is sufficient.
> 
> 2) There are places that now require less dealing with channel locks.
> 
> 3) There are places where channel locks are held for much shorter periods of time.
> 
> 4) There are places where dealing with more than one channel at a time becomes _MUCH_ easier.  ChanSpy is a great example of this.  Writing code in the future that deals with multiple channels will be much easier.
> 
> Some additional information regarding channel locking and reference count handling can be found in channel.h, where a new section has been added that discusses some of the rules associated with it.
> 
> Mark Michelson also assisted with the development of this patch.  He did the conversion of ChanSpy and introduced a new API, ast_autochan, which makes it much easier to deal with holding on to a channel pointer for an extended period of time and having it get automatically updated if the channel gets masqueraded.
> 
> One final note is that currently, app_dahdichan will not compile.  I just haven't been motivated enough to do it.  Everything else should be good to go, though.  I think the best thing to do is to just add whatever is necessary to ChanSpy to allow DAHDIChan to become a wrapper for ChanSpy.
> 
> 
> Diffs
> -----
> 
>   /trunk/apps/app_channelredirect.c 183551 
>   /trunk/apps/app_chanspy.c 183551 
>   /trunk/apps/app_dahdiscan.c 183551 
>   /trunk/apps/app_directed_pickup.c 183551 
>   /trunk/apps/app_minivm.c 183551 
>   /trunk/apps/app_mixmonitor.c 183551 
>   /trunk/apps/app_senddtmf.c 183551 
>   /trunk/apps/app_softhangup.c 183551 
>   /trunk/apps/app_voicemail.c 183551 
>   /trunk/build_tools/cflags.xml 183551 
>   /trunk/channels/chan_agent.c 183551 
>   /trunk/channels/chan_bridge.c 183551 
>   /trunk/channels/chan_gtalk.c 183551 
>   /trunk/channels/chan_iax2.c 183551 
>   /trunk/channels/chan_local.c 183551 
>   /trunk/channels/chan_sip.c 183551 
>   /trunk/funcs/func_channel.c 183551 
>   /trunk/funcs/func_global.c 183551 
>   /trunk/funcs/func_logic.c 183551 
>   /trunk/funcs/func_odbc.c 183551 
>   /trunk/include/asterisk/autochan.h PRE-CREATION 
>   /trunk/include/asterisk/channel.h 183551 
>   /trunk/include/asterisk/lock.h 183551 
>   /trunk/main/Makefile 183551 
>   /trunk/main/autochan.c PRE-CREATION 
>   /trunk/main/channel.c 183551 
>   /trunk/main/cli.c 183551 
>   /trunk/main/devicestate.c 183551 
>   /trunk/main/features.c 183551 
>   /trunk/main/logger.c 183551 
>   /trunk/main/manager.c 183551 
>   /trunk/main/pbx.c 183551 
>   /trunk/res/res_agi.c 183551 
>   /trunk/res/res_clioriginate.c 183551 
>   /trunk/res/res_monitor.c 183551 
>   /trunk/res/snmp/agent.c 183551 
> 
> Diff: http://reviewboard.digium.com/r/203/diff
> 
> 
> Testing
> -------
> 
> I've done some basic testing making calls, transfers, etc., and I know Mark has done some, too.  However, this is one of those patches that seems like you can never test it enough.
> 
> Basically, any usage of Asterisk is useful testing for this.
> 
> 
> Thanks,
> 
> Russell
> 
>