Re: [PATCH 6/7] mmc_spi: convert timeout handling to jiffies andavoid busy waiting

[Pierre Ossman]

On Wed, 11 Mar 2009 14:35:03 +0100
Wolfgang MÃes <wolfgang.mues@xxxxxxxxxxxx> wrote:

> @@ -816,8 +819,11 @@ mmc_spi_data_do(struct mmc_spi_host *hos
>  	else
>  		clock_rate = spi->max_speed_hz;
>  
> -	timeout = ktime_add_ns(ktime_set(0, 0), data->timeout_ns +
> -			data->timeout_clks * 1000000 / clock_rate);
> +	timeout = data->timeout_ns +
> +		  data->timeout_clks * 1000000 / clock_rate;
> +	timeout = usecs_to_jiffies((unsigned int)(timeout / 1000));
> +	if (!timeout)
> +		timeout = 1;
>  

Does this round properly upwards? Having a timeout 6 ms shorter than it
should be can be a real problem (although the huge timeout your other
patch introduces means we probably won't be affected by it).

Rgds
-- 
     -- Pierre Ossman

  WARNING: This correspondence is being monitored by the
  Swedish government. Make sure your server uses encryption
  for SMTP traffic and consider using PGP for end-to-end
  encryption.

Attachment: signature.asc
Description: PGP signature