Received-SPF: pass (google.com: domain of linux-crypto+bounces-4242-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Sat, 18 May 2024 06:07:39 -0700
From: James Bottomley <James.Bottomley@HansenPartnership.com>
To: Herbert Xu <herbert@gondor.apana.org.au>, Jarkko Sakkinen <jarkko@kernel.org>
CC: Eric Biggers <ebiggers@kernel.org>,
 =?ISO-8859-1?Q?N=EDcolas_F=2E_R=2E_A=2E_Prado?= <nfraprado@collabora.com>,
 Ard Biesheuvel <ardb@kernel.org>,
 Linux Crypto Mailing List <linux-crypto@vger.kernel.org>,
 linux-integrity@vger.kernel.org, keyrings@vger.kernel.org,
 regressions@lists.linux.dev, kernel@collabora.com
Subject: Re: [PATCH] crypto: api - Do not load modules until algapi is ready
User-Agent: K-9 Mail for Android
In-Reply-To: <Zkif-BI_OFnFbDhv@gondor.apana.org.au>
References: <0f68c283ff4bbb89b8a019d47891f798c6fff287.camel@HansenPartnership.com> <CAMj1kXHi4r8KY9GvX573kwqvLpMfX-J=K2hWiGAKkf5bnicwYQ@mail.gmail.com> <0d260c2f7a9f67ec8bd2305919636678d06000d1.camel@HansenPartnership.com> <CAMj1kXFE_R_x10BVkU+8vrMz0RHiX0+rz-ZL+w08FH2CLQHZXA@mail.gmail.com> <66ec985f3ee229135bf748f1b0874d5367a74d7f.camel@HansenPartnership.com> <dfb0d930-7cbe-46c5-be19-d132b4906ecf@notapiano> <D1C2NPOBHAHK.20O4IME8OK1FH@kernel.org> <20240518043115.GA53815@sol.localdomain> <ZkhS1zrobNwAuANI@gondor.apana.org.au> <D1CPYHVR94JS.1DIGZUQ2H3NCI@kernel.org> <Zkif-BI_OFnFbDhv@gondor.apana.org.au>
Message-ID: <3A150785-8EE3-4B53-9E71-4DE49C6751B0@HansenPartnership.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

On May 18, 2024 5:32:56 AM PDT, Herbert Xu <herbert@gondor=2Eapana=2Eorg=2E=
au> wrote:
>On Sat, May 18, 2024 at 02:04:18PM +0300, Jarkko Sakkinen wrote:
>>
>> Right does this mean for TPM driver that a crypto API invocation not
>> having everthing needed loaded will block until this is not the case?
>
>All this does is disable module loading by the Crypto API (because
>there is no point and it may deadlock) until such a point where
>all/most drivers have finished loading=2E
>
>So if the algorithm is missing (which shouldn't happen because of
>Kconfig selects), then it will simply fail=2E

I have a curiosity question: if Eric is right and it's looking for an opti=
onal hmac accelerator module, why don't I see this?  The only real config d=
ifference between what I tested and what Nicholas did is he's arm and I'm x=
86=2E

James

--=20
Sent from my Android device with K-9 Mail=2E Please excuse my brevity=2E