Received-SPF: pass (google.com: domain of linux-kernel+bounces-129652-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Feedback-ID: i56a14606:Fastmail
User-Agent: Cyrus-JMAP/3.11.0-alpha0-333-gbfea15422e-fm-20240327.001-gbfea1542
Precedence: bulk
MIME-Version: 1.0
Message-Id: <80d7ab36-2746-43c7-a9b5-184f59981b81@app.fastmail.com>
In-Reply-To: 
 <CAPyNcWdPJYOtDbun-qUJApm9ndsVkEj-sHDfbMWK2yCT6FOwdA@mail.gmail.com>
References: 
 <CA+G9fYtsoP51f-oP_Sp5MOq-Ffv8La2RztNpwvE6+R1VtFiLrw@mail.gmail.com>
 <CAPyNcWeu+bzyQg9S3wDb43jbfk95Su5XcSRFPzUbS2ofZ=+5Fg@mail.gmail.com>
 <347bb21e-98db-4bd5-9ca1-550eac5be9f8@app.fastmail.com>
 <277decf7-d9cb-451c-a105-6ecc998f9f46@intel.com>
 <CAPyNcWeXpidnVsXgt4XpwXiA0CRFdZYFN9O2ys5oq9FQKXNxsw@mail.gmail.com>
 <CAPyNcWdPJYOtDbun-qUJApm9ndsVkEj-sHDfbMWK2yCT6FOwdA@mail.gmail.com>
Date: Wed, 03 Apr 2024 13:17:05 +0200
From: "Arnd Bergmann" <arnd@arndb.de>
To: "Anton Protopopov" <aspsk@isovalent.com>,
 "Alexander Lobakin" <aleksander.lobakin@intel.com>
Cc: "Naresh Kamboju" <naresh.kamboju@linaro.org>,
 "open list" <linux-kernel@vger.kernel.org>, lkft-triage@lists.linaro.org,
 "Linux Regressions" <regressions@lists.linux.dev>,
 "Anders Roxell" <anders.roxell@linaro.org>,
 "Dan Carpenter" <dan.carpenter@linaro.org>,
 "Daniel Borkmann" <daniel@iogearbox.net>, "David Ahern" <dsahern@kernel.org>,
 "Alexei Starovoitov" <ast@kernel.org>,
 "Russell King" <rmk+kernel@armlinux.org.uk>,
 "Ard Biesheuvel" <ardb@kernel.org>,
 "Linus Walleij" <linus.walleij@linaro.org>
Subject: Re: include/linux/build_bug.h:78:41: error: static assertion failed: "struct
 bpf_fib_lookup size check"
Content-Type: text/plain;charset=utf-8
Content-Transfer-Encoding: quoted-printable

On Wed, Apr 3, 2024, at 12:09, Anton Protopopov wrote:
> On Wed, Apr 3, 2024 at 11:57=E2=80=AFAM Anton Protopopov <aspsk@isoval=
ent.com> wrote:
 end.
>>
>> Yeah, I am sending a patch for this right now. Better not to depend on
>> compiler options
>
> One __packed__ was not enough though. The problem was also with the
> union of two __u16's which is padded to be 32 bits when AEABI=3Dn and
> the whole structure is packed (so total size is 66 in this case).

The __packed attribute is easy to misunderstand, in this case you
would need to mark every internal union and struct as well, since
you otherwise run into one or both of these problems:

- an inner aggregate with explicit packing still requires
  32-bit alignment (and padding) for each member, even if the
  outer struct puts it at an unaligned position

- You get a compiler warning if an internal structure
  is follows the normal alignment constraints but is located
  at an unaligned offset, since that violates the alignment
  constraints of the C standard.

      Arnd