Hi Rick,
Minor comment below.
On 18 May 2017 at 03:59, Rick Altherr <raltherr@xxxxxxxxxx> wrote:
> When a hw_random device's quality is non-zero, it will automatically be
> used to fill the kernel's entropy pool. Since timeriomem_rng is used by
> many different devices, the quality needs to be provided by platform
> data or device tree.
>
> Signed-off-by: Rick Altherr <raltherr@xxxxxxxxxx>
> ---
>
> drivers/char/hw_random/timeriomem-rng.c | 7 +++++++
> include/linux/timeriomem-rng.h | 3 +++
> 2 files changed, 10 insertions(+)
>
> diff --git a/drivers/char/hw_random/timeriomem-rng.c b/drivers/char/hw_random/timeriomem-rng.c
> index a0faa5f05deb..03ff5483d865 100644
> --- a/drivers/char/hw_random/timeriomem-rng.c
> +++ b/drivers/char/hw_random/timeriomem-rng.c
> @@ -151,8 +151,15 @@ static int timeriomem_rng_probe(struct platform_device *pdev)
> dev_err(&pdev->dev, "missing period\n");
> return -EINVAL;
> }
> +
> + if (!of_property_read_u32(pdev->dev.of_node,
> + "quality", &i))
> + priv->rng_ops.quality = i;
> + else
> + priv->rng_ops.quality = 0;
> } else {
> period = pdata->period;
> + priv->rng_ops.quality = pdata->quality;
> }
>
> priv->period = ns_to_ktime(period * NSEC_PER_USEC);
> diff --git a/include/linux/timeriomem-rng.h b/include/linux/timeriomem-rng.h
> index 46eb27ddbfab..320f743bf06b 100644
> --- a/include/linux/timeriomem-rng.h
> +++ b/include/linux/timeriomem-rng.h
> @@ -13,4 +13,7 @@ struct timeriomem_rng_data {
>
> /* measures in usecs */
> unsigned int period;
> +
> + /* bits of entropy per 1000 bits read */
Your dt-bindings patch says 'entropy per 1024 bits'. Please be consistent.
> + unsigned int quality;
> };
> --
> 2.13.0.303.g4ebf302169-goog
>
Regards,
PrasannaKumar
