linux-kernel - RE: [Question] Alignment requirement for readX() and writeX()

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] [day] [month] [year] [list]

Message-ID: <7353d0d46f9f4a908b73ac0ff1c070fc@AcuMS.aculab.com>
Date:   Mon, 2 Aug 2021 08:37:02 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Arnd Bergmann' <arnd@...db.de>, Boqun Feng <boqun.feng@...il.com>
CC:     linux-arch <linux-arch@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Miguel Ojeda <miguel.ojeda.sandonis@...il.com>,
        Wedson Almeida Filho <wedsonaf@...gle.com>,
        Gary Guo <gary@...yguo.net>, Hector Martin <marcan@...can.st>,
        Linus Walleij <linus.walleij@...aro.org>
Subject: RE: [Question] Alignment requirement for readX() and writeX()

From: Arnd Bergmann
> Sent: 30 July 2021 17:59
...
> I am not aware of any driver that requires unaligned access on __iomem
> pointers, and since it definitely doesn't work on most architectures, ...

Unaligned accesses into PCIe space can generate a TLP that requests
only some bytes of the first and last 32bit words be transferred.
The target is expected to honour such requests.

On the x86 systems where I've looked at a TLP trace misaligned
accesses are even atomic provided the target doesn't let a local
request interleave.

OTOH drivers are unlikely to make such requests.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)