netdev - Re: [PATCH v2] wcn36xx: Add RX frame SNR as a source of system entropy

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YyJqfsXLESDWDBvR@zx2c4.com>
Date:   Thu, 15 Sep 2022 00:57:50 +0100
From:   "Jason A. Donenfeld" <Jason@...c4.com>
To:     Bryan O'Donoghue <bryan.odonoghue@...aro.org>
Cc:     loic.poulain@...aro.org, kvalo@...nel.org, davem@...emloft.net,
        edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
        wcn36xx@...ts.infradead.org, linux-wireless@...r.kernel.org,
        netdev@...r.kernel.org
Subject: Re: [PATCH v2] wcn36xx: Add RX frame SNR as a source of system
 entropy

On Wed, Sep 14, 2022 at 10:28:41PM +0100, Bryan O'Donoghue wrote:
> The signal-to-noise-ratio SNR is returned by the wcn36xx firmware for each
> received frame. SNR represents all of the unwanted interference signal
> after filtering out the fundamental frequency and harmonics of the
> frequency.
> 
> Noise can come from various electromagnetic sources, from temperature
> affecting the performance hardware components or quantization effects
> converting from analog to digital domains.
> 
> The SNR value returned by the WiFi firmware then is a good source of
> entropy.
> 
> Other WiFi drivers offer up the noise component of the FFT as an entropy
> source for the random pool e.g.
> 
> commit 2aa56cca3571 ("ath9k: Mix the received FFT bins to the random pool")
> 
> I attended Jason's talk on sources of randomness at Plumbers and it
> occurred to me that SNR is a reasonable candidate to add.

Neat!

> Cc: Jason A. Donenfeld <Jason@...c4.com>
> Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@...aro.org>
> ---
>  drivers/net/wireless/ath/wcn36xx/txrx.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/net/wireless/ath/wcn36xx/txrx.c b/drivers/net/wireless/ath/wcn36xx/txrx.c
> index 8da3955995b6e..b73229776af8b 100644
> --- a/drivers/net/wireless/ath/wcn36xx/txrx.c
> +++ b/drivers/net/wireless/ath/wcn36xx/txrx.c
> @@ -16,6 +16,7 @@
>  
>  #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
>  
> +#include <linux/random.h>
>  #include "txrx.h"
>  
>  static inline int get_rssi0(struct wcn36xx_rx_bd *bd)
> @@ -297,6 +298,8 @@ static void wcn36xx_update_survey(struct wcn36xx *wcn, int rssi, int snr,
>  	wcn->chan_survey[idx].rssi = rssi;
>  	wcn->chan_survey[idx].snr = snr;
>  	spin_unlock(&wcn->survey_lock);
> +
> +	add_device_randomness(&snr, sizeof(s8));

Won't this break on big endian? Just have an assignment handle it:

    u8 snr_sample = snr & 0xff;
    add_device_randomness(&snr_sample, sizeof(snr_sample);

That & 0xff is redundant, but it doesn't hurt to be explicit.

Jason