[Lcdproc] imonlcd and 0038 breakage...

[Jarod Wilson]

On 09/10/2009 07:17 PM, Leif Hassell wrote:
> lcdproc 0.5.3 and lirc are still locking up.  Unsure as to where to go
> from here.
>
> Any logfiles that would help?

Kernel messages from the lockup would be good -- don't suppose you can 
set up serial console output?

Anyhow, I've double-checked some things in the lirc_imon code, and I do 
see at least one case where I do think a deadlock is possible... Within 
ir_open(), imon_set_ir_protocol() can get called, and when it is, its 
called without the context lock held, and that in turn calls 
send_packet(), which calls mutex_unlock() followed by mutex_lock() on 
that lock wasn't already held.

However, from what I can see, this would *only* trigger if you were 
shifting ir protocols on the fly. The initial setting of protocol 
happens with the lock properly held. I'll definitely fix this of course, 
but I'm guessing its not the cause of your lockup.



> ------------------------------------------------------------------------
> *From:* jk <fblack947 at yahoo.com>
> *To:* Leif Hassell <wyldfire00101010 at yahoo.com>
> *Sent:* Thursday, September 10, 2009 4:04:52 PM
> *Subject:* Re: [Lcdproc] imonlcd and 0038 breakage...
>
>
>     **I have an Imon LCD (model 15c2:ffdc) and it has been giving me
>     fits. It has been randomly freezing as well, and I'm unsure as to
>     the cause. Perhaps I could provide some real-world test data to assist?
>
>     I'm running ubuntu 9.04, lirc 0.8.6, and lcdproc 0.5.2-0ubuntu2. ...
>
>
> I'd suggest trying lcdproc 0.5.3. As Eric mentioned, it implements a
> frame buffer which should help with the stability (in fact, he wrote it).
>
> I have the same LCD, and it runs fine under ubuntu 8.10 and lcdproc
> 0.5.3. I have to check what version of lirc I have.



-- 
Jarod Wilson
jarod at wilsonet.com