Re: Re: aes: self-test failed

[Mori Hiroki <yamori813%yahoo.co.jp@localhost>]

Hi

This is result. I did this command 10 times.
Every time passed.

c1k$ atf-run | atf-report
Tests root: /usr/tests/sys/crypto/aes

t_aes (1/1): 1 test cases
    aes_ct_selftest: [0.063425s] Passed.
[0.073933s]

Summary for 1 test programs:
    1 passed test cases.
    0 failed test cases.
    0 expected failed test cases.
    0 skipped test cases.

> ----- Original Message -----
> 
> From: "Taylor R Campbell" <campbell+netbsd-port-arm%mumble.net@localhost>
> To: "や もり" <yamori813%yahoo.co.jp@localhost>
> Cc: "port-arm%NetBSD.org@localhost" <port-arm%NetBSD.org@localhost>
> Date: 2023/03/20 月 09:14
> Subject: Re: aes: self-test failed
> 
> 
> > Date: Mon, 20 Mar 2023 08:37:52 +0900 (JST)
> > From: Mori Hiroki <yamori813%yahoo.co.jp@localhost>
> > 
> > > From: "Taylor R Campbell" <campbell+netbsd-port-arm%mumble.net@localhost>
> > > To: "や もり" <yamori813%yahoo.co.jp@localhost>
> > > Cc: "port-arm%NetBSD.org@localhost" <port-arm%NetBSD.org@localhost>
> > > Date: 2023/03/19 日 01:38
> > > Subject: Re: aes: self-test failed
> > > 
> > > You say `sometime at boot' -- do you mean that this doesn't happen
> > > every time?  Does it sometimes boot up correctly?  Can you run the atf
> > > tests for this code when it does boot up correctly?
> > > 
> > > cd /usr/tests/sys/crypto/aes
> > > atf-run | atf-report
> > 
> > My install image don't have that directory. 
> > I don't familiar NetBSD. sorry.
> 
> This is from the tests.tgz (or tests.tar.xz) set, selected when
> installing NetBSD.  You can also just extract it later from the same
> installation medium you used, either with sysinst, or by just running:
> 
> progress -zf /path/to/tests.tgz tar -C / -xpf -
> 
> > > Can you run any kind of memory test on this machine?
> > > 
> > 
> > My kernel conf have no crypto and swcrypto then
> 
> If you mean `pseudo-device crypto' and `pseudo-device swcrypto', these
> aren't directly relevant to the AES self-test.
> 
> It looks like there is some random chance here, possibly affected by
> the different layout of kernel memory -- that's why I suspect a
> hardware failure, or possibly corruption or registers by a bug in the
> interrupt restore logic.