linux-kernel - Re: [PATCH] bpf: Tweak BPF jump table optimizations for objtool compatibility

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: <20200507140733.v4xlzjogtnpgu5lc@treble>
Date:   Thu, 7 May 2020 09:07:33 -0500
From:   Josh Poimboeuf <jpoimboe@...hat.com>
To:     Alexei Starovoitov <alexei.starovoitov@...il.com>
Cc:     Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Network Development <netdev@...r.kernel.org>,
        bpf <bpf@...r.kernel.org>, LKML <linux-kernel@...r.kernel.org>,
        X86 ML <x86@...nel.org>, Peter Zijlstra <peterz@...radead.org>,
        Randy Dunlap <rdunlap@...radead.org>,
        Arnd Bergmann <arnd@...db.de>
Subject: Re: [PATCH] bpf: Tweak BPF jump table optimizations for objtool
 compatibility

On Wed, May 06, 2020 at 05:03:57PM -0700, Alexei Starovoitov wrote:
> > > > > diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
> > > > > index d7ee4c6bad48..05104c3cc033 100644
> > > > > --- a/include/linux/compiler-gcc.h
> > > > > +++ b/include/linux/compiler-gcc.h
> > > > > @@ -171,4 +171,4 @@
> > > > >  #define __diag_GCC_8(s)
> > > > >  #endif
> > > > >
> > > > > -#define __no_fgcse __attribute__((optimize("-fno-gcse")))
> > > > > +#define __no_fgcse __attribute__((optimize("-fno-gcse,-fno-omit-frame-pointer")))
> > > > > --
> > > > > 2.23.0
> > > > >
> > > > > I've tested it with gcc 8,9,10 and clang 11 with FP=y and with ORC=y.
> > > > > All works.
> > > > > I think it's safer to go with frame pointers even for ORC=y considering
> > > > > all the pain this issue had caused. Even if objtool gets confused again
> > > > > in the future __bpf_prog_run() will have frame pointers and kernel stack
> > > > > unwinding can fall back from ORC to FP for that frame.
> > > > > wdyt?
> > > >
> > > > It seems dangerous to me.  The GCC manual recommends against it.
> > > 
> > > The manual can says that it's broken. That won't stop the world from using it.
> > > Just google projects that are using it. For example: qt, lz4, unreal engine, etc
> > > Telling compiler to disable gcse via flag is a guaranteed way to avoid
> > > that optimization that breaks objtool whereas messing with C code is nothing
> > > but guess work. gcc can still do gcse.
> > 
> > But the manual's right, it is broken.  How do you know other important
> > flags won't also be stripped?
> 
> What flags are you worried about?
> I've checked that important things like -mno-red-zone, -fsanitize are preserved.

It's not any specific flags I'm worried about, it's all of them.  There
are a lot of possibilities, with all the different configs, and arches.
Flags are usually added for a good reason, so one randomly missing flag
could have unforeseen results.

And I don't have any visibility into how GCC decides which flags to
drop, and when.  But the docs aren't comforting.

Even if things seem to work now, that could (silently) change at any
point in time.  This time objtool warned about the missing frame
pointer, but that's not necessarily going to happen for other flags.

If we go this route, I would much rather do -fno-gcse on a file-wide
basis.

-- 
Josh