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;
MIME-Version: 1.0
References: <20200115182816.33892-1-trishalfonso@google.com> <dce24e66d89940c8998ccc2916e57877ccc9f6ae.camel@sipsolutions.net>
In-Reply-To: <dce24e66d89940c8998ccc2916e57877ccc9f6ae.camel@sipsolutions.net>
From:   Patricia Alfonso <trishalfonso@google.com>
Date:   Wed, 15 Jan 2020 14:56:39 -0800
Message-ID: <CAKFsvU+sUdGC9TXK6vkg5ZM9=f7ePe7+rh29DO+kHDzFXacx2w@mail.gmail.com>
Subject: Re: [RFC PATCH] UML: add support for KASAN under x86_64
To:     Johannes Berg <johannes@sipsolutions.net>
Cc:     jdike@addtoit.com, richard@nod.at, anton.ivanov@cambridgegreys.com,
        aryabinin@virtuozzo.com, dvyukov@google.com,
        David Gow <davidgow@google.com>,
        Brendan Higgins <brendanhiggins@google.com>,
        linux-um@lists.infradead.org, linux-kernel@vger.kernel.org,
        kasan-dev@googlegroups.com
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Jan 15, 2020 at 10:48 AM Johannes Berg
<johannes@sipsolutions.net> wrote:
> Couple questions, if you don't mind.
>
> > +#ifdef CONFIG_X86_64
> > +#define KASAN_SHADOW_SIZE 0x100000000000UL
> > +#else
> > +#error "KASAN_SHADOW_SIZE is not defined in this sub-architecture"
> > +#endif
>
> Is it even possible today to compile ARCH=um on anything but x86_64? If
> yes, perhaps the above should be
>
>         select HAVE_ARCH_KASAN if X86_64
>
> or so? I assume KASAN itself has some dependencies though, but perhaps
> ARM 64-bit or POWERPC 64-bit could possibly run into this, if not X86
> 32-bit.
>

This seems like a good idea. I'll keep the #ifdef around
KASAN_SHADOW_SIZE, but add "select HAVE_ARCH_KASAN if X86_64" as well.
This will make extending it later easier.

> > +++ b/arch/um/kernel/skas/Makefile
> > @@ -5,6 +5,12 @@
> >
> >  obj-y := clone.o mmu.o process.o syscall.o uaccess.o
> >
> > +ifdef CONFIG_UML
> > +# Do not instrument until after start_uml() because KASAN is not
> > +# initialized yet
> > +KASAN_SANITIZE       := n
> > +endif
>
> Not sure I understand this, can anything in this file even get compiled
> without CONFIG_UML?
>

You are correct; this #ifdef was unnecessary. I will remove it. Thanks!

> > +++ b/kernel/Makefile
> > @@ -32,6 +32,12 @@ KCOV_INSTRUMENT_kcov.o := n
> >  KASAN_SANITIZE_kcov.o := n
> >  CFLAGS_kcov.o := $(call cc-option, -fno-conserve-stack -fno-stack-protector)
> >
> > +ifdef CONFIG_UML
> > +# Do not istrument kasan on panic because it can be called before KASAN
>
> typo there - 'instrument'
>

Thanks for catching that!

> > +++ b/lib/Makefile
> > @@ -17,6 +17,16 @@ KCOV_INSTRUMENT_list_debug.o := n
> >  KCOV_INSTRUMENT_debugobjects.o := n
> >  KCOV_INSTRUMENT_dynamic_debug.o := n
> >
> > +# Don't sanatize
>
> typo
>

Thanks for catching this, too!

> Very cool, I look forward to trying this out! :-)
>
> Thanks,
> johannes
>

Thank you so much for the comments!

Best,
Patricia


--

Patricia Alfonso
Software Engineer
trishalfonso@google.com