Hi Christian,
On 20-07-27 14:45, Christian Eggers wrote:
> When hw_random device's quality is non-zero, it will automatically fill
> the kernel's entropy pool at boot. For this purpose, one conservative
> quality value is being picked up as the default value.
IMHO your value is not conservative enough and the commit message should
explain why we should use 900. Unfortunately I had not enough time to
send my patch addressing this. However please check my commit message
why 900 is not good:
8<------------------------------------------------------------------------
From 9f047eee5e4ce8353c9b764a47e7f584b2013347 Mon Sep 17 00:00:00 2001
From: Marco Felsch <[email protected]>
Date: Thu, 7 May 2020 12:01:28 +0200
Subject: [PATCH] hwrng: imx-rngc - add quality to use it as kernel entropy
pool
The RM describes the RNGB as follow:
8<----------------------------------------------------------------
The RNGB uses the True Random Number Generator (TRNG) and a
Pseudo-Random Number Generator (PRNG) to achieve a true randomness and
cryptographic strength.
8<----------------------------------------------------------------
The RNGB has 3 operation modes: self-test, seed-generation and the final
'random number generation' mode. Befor we can retrieve random numbers
from the RNGB we need to generate the seed pool:
8<----------------------------------------------------------------
During the seed generation, the RNGB adds the entropy generated in the
TRNG to the 256-bit XKEY register. The PRNG algorithm executes 20.000
entropy samples from the TRNG to create an initial seed for the random
number generation.
8<----------------------------------------------------------------
The RNGB can generate 2^20 words (4byte) of 'random' data after the
seed pool was initialized. The pool needs to be reseeded if more words
are required. The reseeding is done automatically since
commit 3acd9ea9331c ("hwrng: imx-rngc - use automatic seeding").
We can't retrieve the TRNG values directly so we need a other way to get
the quality level. We know that the PRNG uses 20.000 entropy samples
from the TRNG to generate 2^20 words (1MiB) and the quality level is
defined as (in bits of entropy per 1024 bits of input). So the quality
level can be calculated by:
20.000 * 1024
------------- = ~ 19.5
2^20
Signed-off-by: Marco Felsch <[email protected]>
---
drivers/char/hw_random/imx-rngc.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/char/hw_random/imx-rngc.c b/drivers/char/hw_random/imx-rngc.c
index 9c47e431ce90..61c844baf26e 100644
--- a/drivers/char/hw_random/imx-rngc.c
+++ b/drivers/char/hw_random/imx-rngc.c
@@ -285,6 +285,7 @@ static int imx_rngc_probe(struct platform_device *pdev)
rngc->rng.init = imx_rngc_init;
rngc->rng.read = imx_rngc_read;
rngc->rng.cleanup = imx_rngc_cleanup;
+ rngc->rng.quality = 19;
rngc->dev = &pdev->dev;
platform_set_drvdata(pdev, rngc);
Hi Marco,
On Tuesday, 28 July 2020, 09:48:17 CEST, Marco Felsch wrote:
> Hi Christian,
>
> On 20-07-27 14:45, Christian Eggers wrote:
> > When hw_random device's quality is non-zero, it will automatically fill
> > the kernel's entropy pool at boot. For this purpose, one conservative
> > quality value is being picked up as the default value.
>
> IMHO your value is not conservative enough and the commit message should
> explain why we should use 900. Unfortunately I had not enough time to
> send my patch addressing this. However please check my commit message
> why 900 is not good:
ok, you caught me. I found the value of 900 in several other drivers and
simply took it. Even parts of my commit message were simply copied...
As I have no real idea about determining the right quality, I will leave this
task for you :-)
Thanks
Christian