linux-kernel - Re: [PATCH 3/3] hwrng: imx-rngc - use polling to wait for end of seeding

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] [day] [month] [year] [list]

Date:   Wed, 23 Aug 2023 13:16:05 +0200
From:   Martin Kaiser <martin@...ser.cx>
To:     Alexander Stein <alexander.stein@...tq-group.com>
Cc:     Herbert Xu <herbert@...dor.apana.org.au>,
        linux-arm-kernel@...ts.infradead.org, linux-crypto@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] hwrng: imx-rngc - use polling to wait for end of
 seeding

Hi Alexander,

Alexander Stein (alexander.stein@...tq-group.com) wrote:

> Am Dienstag, 22. August 2023, 17:25:53 CEST schrieb Martin Kaiser:
> > Use polling to wait until the imx-rngc is properly seeded.

> What is the benefit of burning CPU cycles while waiting for hardware?

it seems to me that

readl_poll_timeout
   readx_poll_timeout
      usleep_range
         usleep_range_state
            schedule_hrtimeout_range

will not wait in the foreground and burn CPU cycles.

The comment for schedule_hrtimeout_range says

 * Make the current task sleep until the given expiry time has
 * elapsed.

> > We do this only in the init function when the imx-rngc becomes active.
> > Polling is ok at this time, there's nothing else we could do while
> > we're waiting.

> > We can now remove the code for the interrupt and the completion.

> Please split the change to polling and IRQ removal into two patches.

Good point. Will do this in v2.

> RNGC_STATUS_SEED_DONE, 1000, RNGC_TIMEOUT * 1000);
> So you want to poll for up to 3s?

According to the manual, "The initial seed takes approximately 2,000,000
clock cycles." With a 66.5MHz clock, this is approx 30ms. So that should
be

RNGC_STATUS_SEED_DONE, 20000, 100000);

The comment for readx_poll_timeout suggests <= 20ms for the poll
interval.

Best regards,
Martin