linux-kernel - MPX is broken for 32bit programs on a 64bit host

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20181129141746.5p5ed7c6rip5qswp@linutronix.de>
Date:   Thu, 29 Nov 2018 15:17:47 +0100
From:   Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To:     x86@...nel.org, Dave Hansen <dave.hansen@...ux.intel.com>
Cc:     Borislav Petkov <bp@...en8.de>, linux-kernel@...r.kernel.org
Subject: MPX is broken for 32bit programs on a 64bit host

so I've been testing my FPU patches and noticed that the test
mpx-mini-test_32 [0] fails on 64bit host.
I went back to v4.4 and it didn't pass there fully, it ended with:
| starting mpx bounds table test
| ERROR: siginfo bounds do not match shadow bounds for register 0

But it did way more than it does today. Since commit 1b028f784e8c3
("x86/mm: Introduce mmap_compat_base() for 32-bit mmap()") it is broken
and ends very early with

| mpx dig (  1) complete, SUCCESS (       0 /    0)
| #BR status == 2, missing bounds table,kernel should have handled!!

The 64bit cases passes. Didn't try the 32bit test case on a 32bit host.
Maybe the test case is fishy (I can't tell).

This is broken since v4.12-rc1. This is known [1] since April this year.
Should I send a removal patch for MPX or is someone actually going to
fix this? Or do we wait for gcc-9 to be released?

[0] tools/testing/selftests/x86/mpx-mini-test_32
[1] https://lkml.kernel.org/r/20180428175043.y35czmhcursyrzsy@gmail.com

Sebastian