Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 2 Jun 2020 11:41:41 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Marco Elver <elver@google.com>
Cc:     Dmitry Vyukov <dvyukov@google.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        syzbot <syzbot+dc1fa714cb070b184db5@syzkaller.appspotmail.com>,
        LKML <linux-kernel@vger.kernel.org>,
        syzkaller-bugs <syzkaller-bugs@googlegroups.com>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        the arch/x86 maintainers <x86@kernel.org>,
        Oleg Nesterov <oleg@redhat.com>,
        kasan-dev <kasan-dev@googlegroups.com>
Subject: Re: PANIC: double fault in fixup_bad_iret
Message-ID: <20200602094141.GR706495@hirez.programming.kicks-ass.net>
References: <000000000000d2474c05a6c938fe@google.com>
 <CACT4Y+ajjB8RmG3_H_9r-kaRAZ05ejW02-Py47o7wkkBjwup3Q@mail.gmail.com>
 <87o8q6n38p.fsf@nanos.tec.linutronix.de>
 <20200529160711.GC706460@hirez.programming.kicks-ass.net>
 <20200529171104.GD706518@hirez.programming.kicks-ass.net>
 <CACT4Y+YB=J0+w7+SHBC3KpKOzxh1Xaarj1cXOPOLKPKQwAW6nQ@mail.gmail.com>
 <CANpmjNP7mKDaXE1=5k+uPK15TDAX+PsV03F=iOR77Pnczkueyg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CANpmjNP7mKDaXE1=5k+uPK15TDAX+PsV03F=iOR77Pnczkueyg@mail.gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jun 01, 2020 at 02:40:31PM +0200, Marco Elver wrote:
> I think Peter wanted to send a patch to add __no_kcsan to noinstr:
> https://lkml.kernel.org/r/20200529170755.GN706495@hirez.programming.kicks-ass.net
> 
> In the same patch we can add __no_sanitize_address to noinstr. But:
> 
> - We're missing a definition for __no_sanitize_undefined and
> __no_sanitize_coverage.

Do those function attributes actually work? Because the last time I
played with some of that I didn't.

Specifically: unmarked __always_inline functions must not generate
instrumentation when they're inlined into a __no_*san function.

(and that fails to build on some GCC versions, and I think fails to
actually work on the rest of them, but I'd have to double check)

> - We still need the above blanket no-instrument for x86 because of
> GCC. We could guard it with "ifdef CONFIG_CC_IS_GCC".

Right; so all of GCC is broken vs that function attribute stuff? Any
plans of getting that fixed? Do we have GCC that care?

Does the GCC plugin approach sound like a viable alternative
implementation of all this?

Anyway, we can make it:

KASAN := SANITIZER_HAS_FUNCTION_ATTRIBUTES

or something, and only make that 'y' when the compiler is sane.

> Not sure what the best strategy is to minimize patch conflicts. For
> now I could send just the patches to add missing definitions. If you'd
> like me to send all patches (including modifying 'noinstr'), let me
> know.

If you're going to do patches anyway, might as well do that :-)