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-next>] [day] [month] [year] [list]
Message-Id: <20181127.142510.717161104715347913.davem@davemloft.net>
Date:   Tue, 27 Nov 2018 14:25:10 -0800 (PST)
From:   David Miller <davem@...emloft.net>
To:     netdev@...r.kernel.org
CC:     daniel@...earbox.net, ast@...nel.org, john.fastabend@...il.com
Subject: BPF uapi structures and 32-bit


In the linux/bpf.h UAPI header, we must absolutely avoid any
non-fixed-sized types.

Otherwise we have serious problems on 32-bit.

Unfortunately I discovered today that we have take on two such cases,
sk_msg_md and sk_reuseport_md, both of which start with two void
pointers.

I hit this because test_verifier.c on my sparc64 box was built as
a 32-bit binary and this causes a hundred or so tests to fail, and
many if not all are because of the changing struct layout.

I could built 64-bit but this absolutely should work properly.

But for fully native 32-bit it is even worse, this will really
need to be resolved because llvm/clang is always going to build
the BPF programs with 64-bit void pointers.

I would strongly suggest we try and fix this now if we can.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ