linux-kernel - Re: general protection fault vs Oops

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

Message-ID: <CAM_iQpXzeMjAUW0LNMpGf+bqykU8fjOJmy=CDdNEPQNpSB4raA@mail.gmail.com>
Date:   Sun, 17 May 2020 13:46:39 -0700
From:   Cong Wang <xiyou.wangcong@...il.com>
To:     Subhashini Rao Beerisetty <subhashbeerisetty@...il.com>
Cc:     Randy Dunlap <rdunlap@...radead.org>,
        Valdis Klētnieks <valdis.kletnieks@...edu>,
        kernelnewbies <kernelnewbies@...nelnewbies.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: general protection fault vs Oops

On Sat, May 16, 2020 at 9:16 AM Subhashini Rao Beerisetty
<subhashbeerisetty@...il.com> wrote:
> Yes, those are out-of-tree modules. Basically, my question is, in
> general what is the difference between 'general protection fault' and
> 'Oops' failure in kernel mode.

For your case, they are likely just different consequences of a same
memory error. Let's assume it is a use-after-free, the behavior is UAF
is undefined: If that memory freed by kernel is also unmapped from
kernel address space, you would get a page fault when using it
afterward, that is an Oops. Or if that memory freed by kernel gets
reallocated and remapped as read-only, you would get a general
protection error when you writing to it afterward.