netdev - Re: [PATCH net V2 0/5] net/smc: fixes 2018-11-12

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: <20181115.095427.1834989231745198763.davem@davemloft.net>
Date:   Thu, 15 Nov 2018 09:54:27 -0800 (PST)
From:   David Miller <davem@...emloft.net>
To:     ubraun@...ux.ibm.com
Cc:     netdev@...r.kernel.org, linux-s390@...r.kernel.org,
        schwidefsky@...ibm.com, heiko.carstens@...ibm.com,
        raspl@...ux.ibm.com
Subject: Re: [PATCH net V2 0/5] net/smc: fixes 2018-11-12

From: Ursula Braun <ubraun@...ux.ibm.com>
Date: Thu, 15 Nov 2018 13:11:15 +0100

> v1->v2:
>    do not define 8-byte alignment for union smcd_cdc_cursor in
>    patch 4/5 "net/smc: atomic SMCD cursor handling"

This is even worse.

The atomic64_t must be properly 8 byte aligned, else it will
crash the kernel when an atomic operation is attempted on it.

You have a situation where your struct attributes are entirely
incompatible.  If the parent struct is __packed, you absolutely
cannot align the atomic64_t in the child structure properly.

One more time, __packed makes correctness here impossible.

I've warned strongly in the past to avoid __packed.

Now you have to untangle this mess somehow.