On Fri, Sep 26, 2025 at 10:36:23AM +0000, Gaurav Jain wrote: > > >> -----Original Message----- >> From: Peng Fan (OSS) <[email protected]> >> Sent: Tuesday, September 23, 2025 2:43 PM >> To: Gaurav Jain <[email protected]> >> Cc: Horia Geanta <[email protected]>; Pankaj Gupta >> <[email protected]>; Varun Sethi <[email protected]>; Silvano Di Ninno >> <[email protected]>; Meenakshi Aggarwal >> <[email protected]>; Tom Rini <[email protected]>; Fabio >> Estevam <[email protected]>; Peng Fan <[email protected]>; Simon Glass >> <[email protected]>; Baehring, Olaf <[email protected]>; Heinrich >> Schuchardt <[email protected]>; [email protected] >> Subject: Re: [PATCH] crypto/rng: double the entropy delay interval for retry >> >> Hi Gaurav, >> >> On Fri, Sep 05, 2025 at 03:03:48PM +0530, Gaurav Jain wrote: >> >during entropy evaluation, if the generated samples fail any >> >statistical test, then, all of the bits will be discarded, and a second >> >set of samples will be generated and tested. >> > >> >the entropy delay interval should be doubled before performing retry. >> > >> >Signed-off-by: Gaurav Jain <[email protected]> >> >--- >> > drivers/crypto/fsl/jr.c | 2 +- >> > 1 file changed, 1 insertion(+), 1 deletion(-) >> > >> >diff --git a/drivers/crypto/fsl/jr.c b/drivers/crypto/fsl/jr.c index >> >8f7a821ebf3..07d336c074a 100644 >> >--- a/drivers/crypto/fsl/jr.c >> >+++ b/drivers/crypto/fsl/jr.c >> >@@ -642,7 +642,7 @@ static int rng_init(uint8_t sec_idx, ccsr_sec_t *sec) >> > */ >> > if (!inst_handles) { >> > kick_trng(ent_delay, sec); >> >- ent_delay += 400; >> >+ ent_delay = ent_delay * 2; >> >> I not get the point on why need to double, would you please update the commit >> why not 500, 800 or else. What make ent_delay * 2 the best choice? > >The typical reason a re-try is required when the entropy delay value is set >too low. >Therefore, when the re-try is attempted the entropy delay is doubled, thus >giving twice as long to generate each bit of entropy. >This is done to prevent having too many failures.
ok. I would say "double the ent_delay to give more chance to pass", I will update the commit, no need resend. Thanks, Peng > >Regards >Gaurav Jain >> >> Thanks, >> Peng >> >> > } >> > /* >> > * if instantiate_rng(...) fails, the loop will rerun >> >-- >> >2.25.1 >> >
