[PATCH 1/2] hvc_console polling mode timer backoff

[Michael Ellerman]

On Tue, 2007-04-17 at 09:44 -0500, Will Schmidt wrote:
> Paulus, I think this is ready to go, please apply, thanks. 
> 
> 
> 
> Add a back-off mechanism to hvc_console's polling logic.   This change
> drops the timers/second ratio from ~90 to ~1/2 while the console is
> idle.
> This change is most noticeable when watching /proc/timer_stats output.
> 
> This only affects when the hvc_console is running in poll mode, i.e.
> power4 and cell systems.
> 
> I've tested on Power4, Michael Ellerman has both contributed to the
> patch and tested on cell.   
> 
> Signed-off-by: Will Schmidt <will_schmidt at vnet.ibm.com>
> cc: Michael Ellerman <michael at ellerman.id.au>

That should probably be:

Signed-off-by: Michael Ellerman <michael at ellerman.id.au>

cheers

> ---
> diff --git a/drivers/char/hvc_console.c b/drivers/char/hvc_console.c
> index a0a88aa..fc9bc77 100644
> --- a/drivers/char/hvc_console.c
> +++ b/drivers/char/hvc_console.c
> @@ -47,8 +47,6 @@ #include "hvc_console.h"
>  #define HVC_MAJOR	229
>  #define HVC_MINOR	0
>  
> -#define TIMEOUT		(10)
> -
>  /*
>   * Wait this long per iteration while trying to push buffered data to the
>   * hypervisor before allowing the tty to complete a close operation.
> @@ -550,6 +548,18 @@ static int hvc_chars_in_buffer(struct tt
>  	return hp->n_outbuf;
>  }
>  
> +/*
> + * timeout will vary between the MIN and MAX values defined here.  By default
> + * and during console activity we will use a default MIN_TIMEOUT of 10.  When
> + * the console is idle, we increase the timeout value on each pass through
> + * msleep until we reach the max.  This may be noticeable as a brief (average
> + * one second) delay on the console before the console responds to input when
> + * there has been no input for some time.
> + */
> +#define MIN_TIMEOUT		(10)
> +#define MAX_TIMEOUT		(2000)
> +static u32 timeout = MIN_TIMEOUT;
> +
>  #define HVC_POLL_READ	0x00000001
>  #define HVC_POLL_WRITE	0x00000002
>  
> @@ -642,9 +652,14 @@ #endif /* CONFIG_MAGIC_SYSRQ */
>   bail:
>  	spin_unlock_irqrestore(&hp->lock, flags);
>  
> -	if (read_total)
> +	if (read_total) {
> +		/* Activity is occurring, so reset the polling backoff value to
> +		   a minimum for performance. */
> +		timeout = MIN_TIMEOUT;
> +
>  		tty_flip_buffer_push(tty);
> -	
> +	}
> +
>  	return poll_mask;
>  }
>  
> @@ -688,8 +703,12 @@ int khvcd(void *unused)
>  		if (!hvc_kicked) {
>  			if (poll_mask == 0)
>  				schedule();
> -			else
> -				msleep_interruptible(TIMEOUT);
> +			else {
> +				if (timeout < MAX_TIMEOUT)
> +					timeout += (timeout >> 6) + 1;
> +
> +				msleep_interruptible(timeout);
> +			}
>  		}
>  		__set_current_state(TASK_RUNNING);
>  	} while (!kthread_should_stop());
> 
> 
> 
> 
-- 
Michael Ellerman
OzLabs, IBM Australia Development Lab

wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)

We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.ozlabs.org/pipermail/linuxppc-dev/attachments/20070418/cf5748a5/attachment.pgp>