linux-kernel - Re: [PATCH] x86/dumpstack: fix address space casting in show

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: <20180831143801.dn5dgpob7xb2kqaq@ltop.local>
Date:   Fri, 31 Aug 2018 16:38:02 +0200
From:   Luc Van Oostenryck <luc.vanoostenryck@...il.com>
To:     Jann Horn <jannh@...gle.com>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        kernel list <linux-kernel@...r.kernel.org>,
        Kees Cook <keescook@...omium.org>, bp@...e.de
Subject: Re: [PATCH] x86/dumpstack: fix address space casting in
 show_opcodes()

On Fri, Aug 31, 2018 at 03:26:24PM +0200, Jann Horn wrote:
> 
> By the way, here are all 60 probe_kernel_read() callers:

... 

> 41 of these (or something like that, I counted by hand) have some sort
> of cast in the call expression.
> probe_kernel_read() is kinda special in that expected types for the
> second argument are both kernel pointers and unsigned longs. It might
> make sense to have a wrapper macro around probe_kernel_read() that
> accepts anything as long as it's as wide as a pointer...

Well, if __user pointers should not be accepted, then it's much
better, typewisely, to leave it as const void * (and cast the
ulong args to some plain generic pointer, like void* or u8*).

-- Luc Van Oostenryck