Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (1.0)
Subject: Re: [PATCH 1/2 v2] kprobe: Do not use uaccess functions to access kernel memory that can fault
From:   Andy Lutomirski <luto@amacapital.net>
In-Reply-To: <20190215210801.7f8c94cf@gandalf.local.home>
Date:   Fri, 15 Feb 2019 18:14:21 -0800
Cc:     Linus Torvalds <torvalds@linux-foundation.org>,
        Linux List Kernel Mailing <linux-kernel@vger.kernel.org>,
        Ingo Molnar <mingo@kernel.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        stable <stable@vger.kernel.org>,
        Changbin Du <changbin.du@gmail.com>,
        Jann Horn <jannh@google.com>,
        Kees Cook <keescook@chromium.org>,
        Andy Lutomirski <luto@kernel.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <1B52384E-5C02-4279-9111-34881BA3204E@amacapital.net>
References: <20190215174712.372898450@goodmis.org> <20190215174945.557218316@goodmis.org> <CAHk-=wjP4bEYCFaKZVG4kLWmYBctyfsW9jjhC91tiTk9WwB8Rg@mail.gmail.com> <20190215171539.4682f0b4@gandalf.local.home> <300C4516-A093-43AE-8707-1C42486807A4@amacapital.net> <20190215191949.04604191@gandalf.local.home> <D1A39BE5-CAB2-42D1-B87B-2E21906FE5B1@amacapital.net> <20190215210801.7f8c94cf@gandalf.local.home>
To:     Steven Rostedt <rostedt@goodmis.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On Feb 15, 2019, at 6:08 PM, Steven Rostedt <rostedt@goodmis.org> wrote:
>=20
> On Fri, 15 Feb 2019 17:32:55 -0800
> Andy Lutomirski <luto@amacapital.net> wrote:
>=20
>>> I added you just because I wanted help getting the change log correct,
>>> as that's what Linus was complaining about. I kept using "kernel
>>> address" when the sample bug used for the patch was really a
>>> non-canonical address (as Linus said, it's just garbage. Neither kernel
>>> or user space). But I pointed out that this can also bug if the
>>> address is canonical and in the kernel address space. The old code
>>> didn't complain about non-canonical or kernel address faulting before
>>> commit 9da3f2b7405, which only talks about kernel address space
>>> faulting (which is why I only mentioned that in my messages).
>>>=20
>>> Would changing all the mention of "kernel address" to "non user space"
>>> be accurate?
>>>=20
>>=20
>> I think =E2=80=9Ckernel address=E2=80=9D is right. It=E2=80=99s illegal t=
o access anything that isn=E2=80=99t known to be a valid kernel address whil=
e in KERNEL_DS.
>=20
> But an non-canonical address is not a "kernel address", and that will
> cause a bug too.

Indeed. A non-canonical address is not known to be a valid kernel address.  I=
 stand by my slightly pedantic statement :)

> This patch came about because it was changed that if
> we do a uaccess on something other than a user space address and take a
> fault (either because it was a non-canonical address, or a kernel
> address), we BUG! Where before that one patch, it would just return a
> fault.
>=20
>>=20
>> The old __copy seems likely to have always been a bit bogus.
>>=20
>> BTW, what is this probe_mem_read() thing?  Some minimal inspection sugges=
ts it=E2=80=99s a buggy reimplementation of probe_kernel_read().  Can you de=
lete it and just use probe_kernel_read() directly?
>=20
> Well, the issue is that we have trace_probe_tmpl.h in that same
> directory, which does the work for kprobes and uprobes. The
> trace_kprobes.c defines all the functions for handling kprobes, and
> trace_uprobes.c does all the handling of uprobes, then they include
> trace_probe_tmpl.h which does the bulk of the work.
>=20
> In the uprobes case, we have:
>=20
> static nokprobe_inline int
> probe_mem_read(void *dest, void *src, size_t size)
> {
>    void __user *vaddr =3D (void __force __user *)src;
>=20
>    return copy_from_user(dest, vaddr, size) ? -EFAULT : 0;
> }
>=20
> Because that is adding probes on userspace code.
>=20
>=20

Can the kprobe case call probe_kernel_read?  Maybe it does already?=