[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=whbbFHUF44At4YYOB9ZWg00rhFmyArD+KocXDX-+-8a1Q@mail.gmail.com>
Date: Wed, 28 Jun 2023 21:35:49 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Kees Cook <keescook@...omium.org>
Cc: Boris Brezillon <bbrezillon@...nel.org>,
Arnaud Ebalard <arno@...isbad.org>,
Srujana Challa <schalla@...vell.com>,
Mustafa Ismail <mustafa.ismail@...el.com>,
Shiraz Saleem <shiraz.saleem@...el.com>,
Herbert Xu <herbert@...dor.apana.org.au>,
Linux Crypto Mailing List <linux-crypto@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Jason Gunthorpe <jgg@...pe.ca>,
Leon Romanovsky <leon@...nel.org>
Subject: Re: Build error in crypto/marvell/cesa/cipher.c
On Wed, 28 Jun 2023 at 20:48, Kees Cook <keescook@...omium.org> wrote:
>
> The unexpected bit is that without -fstrict-flex-arrays=3 (i.e. the
> default since the dawn of time), the compiler treats any array that
> happens to be the last struct member as a flexible array.
Oh. Ok, that explains why it's showing up for me now, at least. It's
an odd rule, but I can see why people would have done that.
I've only seen the zero- and one-sized arrays commonly used for the
traditional "fake flex array", but I guess other sizes can easily
happen.
Linus
Powered by blists - more mailing lists