Message-ID: <CALCETrU2V5KBA97g3O-yHiUu1acmM_K9b2a5ATmKSqGwdd-+Dw@mail.gmail.com>
Date:   Fri, 22 Feb 2019 15:59:30 -0800
From:   Andy Lutomirski <luto@...nel.org>
To:     Jann Horn <jannh@...gle.com>
Cc:     Nadav Amit <namit@...are.com>,
        Alexei Starovoitov <alexei.starovoitov@...il.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        Ingo Molnar <mingo@...nel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Changbin Du <changbin.du@...il.com>,
        Kees Cook <keescook@...omium.org>,
        Andy Lutomirski <luto@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Network Development <netdev@...r.kernel.org>,
        "bpf@...r.kernel.org" <bpf@...r.kernel.org>,
        Rick Edgecombe <rick.p.edgecombe@...el.com>,
        Dave Hansen <dave.hansen@...el.com>,
        "Peter Zijlstra (Intel)" <peterz@...radead.org>,
        Igor Stoppa <igor.stoppa@...il.com>
Subject: Re: [PATCH 1/2 v2] kprobe: Do not use uaccess functions to access
 kernel memory that can fault

On Fri, Feb 22, 2019 at 3:02 PM Jann Horn <jannh@...gle.com> wrote:
>
> On Fri, Feb 22, 2019 at 11:39 PM Nadav Amit <namit@...are.com> wrote:
> > > On Feb 22, 2019, at 2:21 PM, Nadav Amit <namit@...are.com> wrote:
> > >
> > >> On Feb 22, 2019, at 2:17 PM, Jann Horn <jannh@...gle.com> wrote:
> > >>
> > >> On Fri, Feb 22, 2019 at 11:08 PM Nadav Amit <namit@...are.com> wrote:
> > >>>> On Feb 22, 2019, at 1:43 PM, Jann Horn <jannh@...gle.com> wrote:
> > >>>>
> > >>>> (adding some people from the text_poke series to the thread, removing stable@)
> > >>>>
> > >>>> On Fri, Feb 22, 2019 at 8:55 PM Andy Lutomirski <luto@...capital.net> wrote:
> > >>>>>> On Feb 22, 2019, at 11:34 AM, Alexei Starovoitov <alexei.starovoitov@...il.com> wrote:
> > >>>>>>> On Fri, Feb 22, 2019 at 02:30:26PM -0500, Steven Rostedt wrote:
> > >>>>>>> On Fri, 22 Feb 2019 11:27:05 -0800
> > >>>>>>> Alexei Starovoitov <alexei.starovoitov@...il.com> wrote:
> > >>>>>>>
> > >>>>>>>>> On Fri, Feb 22, 2019 at 09:43:14AM -0800, Linus Torvalds wrote:
> > >>>>>>>>>
> > >>>>>>>>> Then we should still probably fix up "__probe_kernel_read()" to not
> > >>>>>>>>> allow user accesses. The easiest way to do that is actually likely to
> > >>>>>>>>> use the "unsafe_get_user()" functions *without* doing a
> > >>>>>>>>> uaccess_begin(), which will mean that modern CPU's will simply fault
> > >>>>>>>>> on a kernel access to user space.
> > >>>>>>>>
> > >>>>>>>> On bpf side the bpf_probe_read() helper just calls probe_kernel_read()
> > >>>>>>>> and users pass both user and kernel addresses into it and expect
> > >>>>>>>> that the helper will actually try to read from that address.
> > >>>>>>>>
> > >>>>>>>> If __probe_kernel_read will suddenly start failing on all user addresses
> > >>>>>>>> it will break the expectations.
> > >>>>>>>> How do we solve it in bpf_probe_read?
> > >>>>>>>> Call probe_kernel_read and if that fails call unsafe_get_user byte-by-byte
> > >>>>>>>> in the loop?
> > >>>>>>>> That's doable, but people already complain that bpf_probe_read() is slow
> > >>>>>>>> and shows up in their perf report.
> > >>>>>>>
> > >>>>>>> We're changing kprobes to add a specific flag to say that we want to
> > >>>>>>> differentiate between kernel or user reads. Can this be done with
> > >>>>>>> bpf_probe_read()? If it's showing up in perf report, I doubt a single
> > >>>>>>
> > >>>>>> so you're saying you will break existing kprobe scripts?
> > >>>>>> I don't think it's a good idea.
> > >>>>>> It's not acceptable to break bpf_probe_read uapi.
> > >>>>>
> > >>>>> If so, the uapi is wrong: a long-sized number does not reliably identify an address if you don’t separately know whether it’s a user or kernel address. s390x and 4G:4G x86_32 are the notable exceptions. I have lobbied for RISC-V and future x86_64 to join the crowd.  I don’t know whether I’ll win this fight, but the uapi will probably have to change for at least s390x.
> > >>>>>
> > >>>>> What to do about existing scripts is a different question.
> > >>>>
> > >>>> This lack of logical separation between user and kernel addresses
> > >>>> might interact interestingly with the text_poke series, specifically
> > >>>> "[PATCH v3 05/20] x86/alternative: Initialize temporary mm for
> > >>>> patching" (https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Flkml%2F20190221234451.17632-6-rick.p.edgecombe%40intel.com%2F&amp;data=02%7C01%7Cnamit%40vmware.com%7Cf2513009ef734ecd6b0d08d69913a5ae%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C0%7C0%7C636864707020821793&amp;sdata=HAbnDcrBne64JyPuVUMKmM7nQk67F%2BFvjuXEn8TmHeo%3D&amp;reserved=0)
> > >>>> and "[PATCH v3 06/20] x86/alternative: Use temporary mm for text
> > >>>> poking" (https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Flkml%2F20190221234451.17632-7-rick.p.edgecombe%40intel.com%2F&amp;data=02%7C01%7Cnamit%40vmware.com%7Cf2513009ef734ecd6b0d08d69913a5ae%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C0%7C0%7C636864707020821793&amp;sdata=vNRIMKtFDy%2F3z5FlTwDiJY6VGEV%2FMHgQPTdFSFtCo4s%3D&amp;reserved=0),
> > >>>> right? If someone manages to get a tracing BPF program to trigger in a
> > >>>> task that has switched to the patching mm, could they use
> > >>>> bpf_probe_write_user() - which uses probe_kernel_write() after
> > >>>> checking that KERNEL_DS isn't active and that access_ok() passes - to
> > >>>> overwrite kernel text that is mapped writable in the patching mm?
> > >>>
> > >>> Yes, this is a good point. I guess text_poke() should be defined with
> > >>> “__kprobes” and open-code memcpy.
> > >>>
> > >>> Does it sound reasonable?
> > >>
> > >> Doesn't __text_poke() as implemented in the proposed patch use a
> > >> couple other kernel functions, too? Like switch_mm_irqs_off() and
> > >> pte_clear() (which can be a call into a separate function on paravirt
> > >> kernels)?
> > >
> > > I will move the pte_clear() to be done after the poking mm was unloaded.
> > > Give me a few minutes to send a sketch of what I think should be done.
> >
> > Err.. You are right, I don’t see an easy way of preventing a kprobe from
> > being set on switch_mm_irqs_off(), and open-coding this monster is too ugly.
> >
> > The reasonable solution seems to me as taking all the relevant pieces of
> > code (and data) that might be used during text-poking and encapsulating them, so they
> > will be set in a memory area which cannot be kprobe'd. This can also be
> > useful to write-protect data structures of code that calls text_poke(),
> > e.g., static-keys. It can also protect data on that stack that is used
> > during text_poke() from being overwritten from another core.
> >
> > This solution is somewhat similar to Igor Stoppa’s idea of using “enclaves”
> > when doing write-rarely operations.
> >
> > Right now, I think that text_poke() will keep being susceptible to such
> > an attack, unless you have a better suggestion.
>
> A relatively simple approach might be to teach BPF not to run kprobe
> programs and such in contexts where current->mm isn't the active mm?
> Maybe using nmi_uaccess_okay(), or something like that? It looks like
> perf_callchain_user() also already uses that. Except that a lot of
> this code is x86-specific...

This sounds like exactly the right solution.  If you're running from
some unknown context (like NMI or tracing), then you should check
nmi_uaccess_okay().  I think we should just promote that to be a
non-arch-specific function (that returns true by default) and check it
the relevant bpf_probe_xyz() functions.

Alexei, does that seem reasonable?

Powered by blists - more mailing lists