Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.33 as permitted sender) client-ip=23.128.96.33;
Date:   Tue, 5 Dec 2023 21:57:29 +0000
From:   Al Viro <viro@zeniv.linux.org.uk>
To:     tanzirh@google.com
Cc:     Kees Cook <keescook@chromium.org>,
        Andy Shevchenko <andy@kernel.org>,
        linux-hardening@vger.kernel.org, linux-kernel@vger.kernel.org,
        Nick DeSaulniers <nnn@google.com>,
        Andrew Morton <akpm@linux-foundation.org>, llvm@lists.linux.dev
Subject: Re: [PATCH] lib/string: shrink lib/string.i via IWYU
Message-ID: <20231205215729.GG1674809@ZenIV>
References: <20231205-libstringheader-v1-1-7f9c573053a7@gmail.com>
 <20231205213807.GE1674809@ZenIV>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20231205213807.GE1674809@ZenIV>
Sender: Al Viro <viro@ftp.linux.org.uk>
Precedence: bulk

On Tue, Dec 05, 2023 at 09:38:07PM +0000, Al Viro wrote:

> It also breeds includes of asm/*.h, by the look of the output, which is
> not a good thing in general ;-/  E.g. #include <asm/uaccess.h> *anywhere*
> outside of linux/uaccess.h is a bad idea.

Let me elaborate a bit:

Consider e.g. a situation when we have linux/foo.h pulling asm/foo.h;
among the things declared there is foo_bar(), which is identical in
18 instances of asm/foo.h (out of current 21).

At some point we notice that; the obvious approach would be to have
3 unusual architectures have their asm/foo.h define _ARCH_HAS_ODD_FOO_BAR
and put the common variant in linux/foo.h under ifndef _ARCH_HAS_ODD_FOO_BAR.

Except that any file that pulls asm/foo.h will get broken by that.
And something like your tool might convert the users of linux/foo.h to
asm/foo.h - all it takes is using only the primitives that currently
happen to be in asm/foo.h on all architectures.

This is not an artifical example, BTW - look at the history of uaccess.h and
checksum.h...