linux-kernel - Re: [RFC v2 2/3] atomic: Specify alignment for atomic_t and atomic64

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] [thread-next>] [day] [month] [year] [list]

Message-Id: <57ac401b-222b-4c85-b719-9e671a4617cf@app.fastmail.com>
Date: Mon, 22 Sep 2025 17:21:45 +0200
From: "Arnd Bergmann" <arnd@...db.de>
To: "Finn Thain" <fthain@...ux-m68k.org>,
 "Geert Uytterhoeven" <geert@...ux-m68k.org>
Cc: "Peter Zijlstra" <peterz@...radead.org>, "Will Deacon" <will@...nel.org>,
 "Andrew Morton" <akpm@...ux-foundation.org>,
 "Boqun Feng" <boqun.feng@...il.com>, "Jonathan Corbet" <corbet@....net>,
 "Mark Rutland" <mark.rutland@....com>, linux-kernel@...r.kernel.org,
 Linux-Arch <linux-arch@...r.kernel.org>, linux-m68k@...r.kernel.org,
 "Lance Yang" <lance.yang@...ux.dev>
Subject: Re: [RFC v2 2/3] atomic: Specify alignment for atomic_t and atomic64_t

On Mon, Sep 22, 2025, at 10:16, Finn Thain wrote:
>
> Yes, I noticed that also, after I started building with defconfigs.
> I pushed a new patch to my github repo.
>
> https://github.com/fthain/linux/tree/atomic_t

I had a look at that repository, and I think the extra
annotations you added on a lot of structures have the
same mistake that I made when looking at annotating
ABI structures for -malign-int earlier:

Adding __aligned__((sizeof(long))) to a structure definition
only changes the /outer/ alignment of the structure itself,
so e.g.

struct {
    short a;
    int b;
} __attribute__((aligned(sizeof(long))));

creates a structure with 4-byte alignment and 8-byte
size when building with -mno-align-int, but the padding
is added after 'b', which is still misaligned.

In order to align members the same way that -malign-int does,
each unaligned member needs a separate attribute to force
aligning it, like

struct {
    short a;
    int b __attribute__((aligned(sizeof(int))));
};

This is different from how __attribute__((packed)) works,
as that is always applied to each member of the struct
if you add it to the end.

      Arnd