Re: [freetel-oslec] [PATCH] fix Pstate growing in OSLEC

[David Rowe <david@rowetel.com>]

Hi Tzafrir,

I submitted the patch to Greg a few weeks ago (as well as checking it
into Oslec SVN).

Re gathering dust, on the contrary I understand that many people are
using Oslec from the kernel staging directory with the Asterisk DAHDI
interface.

Cheers,

David

On Thu, 2009-06-18 at 20:45 +0300, Tzafrir Cohen wrote:
> The following fix has been posted in the OSLEC mailing list by Ruslan
> Cray (and shortly afterwards acepted into the OSLEC SVN tree). Sadly it
> has not been forwarded the kernel.org tree until now.
> 
> See:
> 
> http://sourceforge.net/mailarchive/forum.php?thread_name=000b01c9d88f%24ec689690%2475a8a8c0%40rcraydvlpmnt&forum_name=freetel-oslec
> 
> As a side note, the OSLEC code is gathering dust in the staging tree.
> As-is it is mostly useless without providing an interface for its code.
> I was hoping to see the mISDN code starting to use it but so far no 
> change in that front.
> 
> ---
>  drivers/staging/echo/echo.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/staging/echo/echo.c b/drivers/staging/echo/echo.c
> index 6d7217e..faab26d 100644
> --- a/drivers/staging/echo/echo.c
> +++ b/drivers/staging/echo/echo.c
> @@ -395,7 +395,7 @@ int16_t oslec_update(struct oslec_state *ec, int16_t tx, int16_t rx)
>  		old = (int)ec->fir_state.history[ec->fir_state.curr_pos] *
>  		    (int)ec->fir_state.history[ec->fir_state.curr_pos];
>  		ec->Pstates +=
> -		    ((new - old) + (1 << ec->log2taps)) >> ec->log2taps;
> +		    ((new - old) + (1 << (ec->log2taps - 1))) >> ec->log2taps;
>  		if (ec->Pstates < 0)
>  			ec->Pstates = 0;
>  	}
> -- 
> 1.5.6.5
> 
> 
-- 
Free Telephony Project
open embedded IP-PBX hardware and software
http://www.rowetel.com/ucasterisk