Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20200803044024.GA6429@gondor.apana.org.au> <20200830223304.GA16882@gondor.apana.org.au>
 <20201026011159.GA2428@gondor.apana.org.au> <20201227113221.GA28744@gondor.apana.org.au>
 <20210108035450.GA6191@gondor.apana.org.au> <20210708030913.GA32097@gondor.apana.org.au>
 <20210817013601.GA14148@gondor.apana.org.au> <20210929023843.GA28594@gondor.apana.org.au>
 <20211029041408.GA3192@gondor.apana.org.au>
In-Reply-To: <20211029041408.GA3192@gondor.apana.org.au>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Fri, 29 Oct 2021 10:39:35 -0700
Message-ID: <CAHk-=whEFkHvtecaUF5MEZMxW_UVUMn-bLmfAhcvx4t9Qia90A@mail.gmail.com>
Subject: Re: [GIT PULL] Crypto Fixes for 5.15
To:     Herbert Xu <herbert@gondor.apana.org.au>
Cc:     "David S. Miller" <davem@davemloft.net>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Linux Crypto Mailing List <linux-crypto@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Oct 28, 2021 at 9:14 PM Herbert Xu <herbert@gondor.apana.org.au> wrote:
>
> This push fixes a build-time warning in x86/sm4.

Hmm..

> Tianjia Zhang (1):
>       crypto: x86/sm4 - Fix invalid section entry size

So I do wonder why the crypto code asm tends to use such complex
".section" directives.

Almost everythign else in the kernel just uses

     .section <name>,"a"

for data (and "ax" for text).

The crypto code uses that very subtle "M" flag which allows the linker
to merge data, but how much of a win is that added complexity? Now you
need to get the sizes exactly right, and you hit tooling issues.

Plus, I get the feeling that some people have just copied-and-pasted
those things, and don't necessarily realize just _how_ subtle 'M'
sections are.

How much of a data savings is it to have this complexity? Particularly
since I suspect most of the time these things end up being individual
modules, and never actually get linked together at all?

So I get this very strong feeling that the crypto code is doing
complicated and fragile things that cause tooling issues - and does so
for no good reason and no actual real-life upside.

I've pulled this, but I do suspect that all the fancy

   .section .rodata.xyz, "aM", @progbits, abc

could just as well be

    .section .rodata, "a"

instead, and you guys are only causing problems for yourself by trying
to be clever in pointless ways.

            Linus