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] [day] [month] [year] [list]
Message-ID: <CAEVVKH9eF9ifwTXZyE5LKEvUcDPAr9rcEmqdM=vwBLcrPvyAKw@mail.gmail.com>
Date:   Fri, 15 Oct 2021 10:21:58 +0800
From:   Xiongwei Song <sxwjean@...il.com>
To:     Kees Cook <keescook@...omium.org>
Cc:     Xiongwei Song <sxwjean@...com>, x86@...nel.org,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>,
        Gabriel Krisman Bertazi <krisman@...labora.com>,
        Andy Lutomirski <luto@...nel.org>,
        "Chang S. Bae" <chang.seok.bae@...el.com>,
        Al Viro <viro@...iv.linux.org.uk>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH RFC] x86/elf: Fix incorrect STACK_RND_MASK for x86_64

On Fri, Oct 15, 2021 at 12:10 AM Kees Cook <keescook@...omium.org> wrote:
>
> On Thu, Oct 14, 2021 at 09:27:36PM +0800, sxwjean@...com wrote:
> > From: Xiongwei Song <sxwjean@...il.com>
> >
> > According to the comment for stack randomization of x86, the range of
> > randomization is 1GB, which occupies 30 valid bits in binary. And in
> > x86_64, PAGE_SHIFT occupies 12 bits, then the STACK_RND_MASK should
> > occupy 18 bits for 1GB.
> >
> > The current range of randomization is :
> >       0x3fffff << PAGE_SHIFT
> > , which occupies 34 bits.
> >
> > This patch changed 0x3fffff to 0x3ffff, which makes the range of stack
> > randomization is real 1GB.
>
> Why do we want to _reduce_ entropy here? Perhaps adjust the comment
> instead?

Hi Kees,

Shouldn't x86_64 have consistent behavior like arm64, riscv or powerpc
on this? I'm not sure if 34 bits is too wasteful for stack. If 34 bits have no
side effects, then I have no objection with your suggestion.

Regards,
Xiongwei


>
> -Kees
>
> >
> > Before this patch(bit0 ~ bit33 randomized):
> >       root@...ux86-64:~# for i in {1..20};do cat /proc/self/maps | grep stack;done
> >       7ffddf971000-7ffddf992000 rw-p 00000000 00:00 0                          [stack]
> >       7ffc4e0d4000-7ffc4e0f5000 rw-p 00000000 00:00 0                          [stack]
> >       7fff5898e000-7fff589af000 rw-p 00000000 00:00 0                          [stack]
> >       7ffcf910b000-7ffcf912c000 rw-p 00000000 00:00 0                          [stack]
> >       7ffce829d000-7ffce82be000 rw-p 00000000 00:00 0                          [stack]
> >       7ffec72d9000-7ffec72fa000 rw-p 00000000 00:00 0                          [stack]
> >       7ffd6e256000-7ffd6e277000 rw-p 00000000 00:00 0                          [stack]
> >       7ffe14120000-7ffe14141000 rw-p 00000000 00:00 0                          [stack]
> >       7ffe21549000-7ffe2156a000 rw-p 00000000 00:00 0                          [stack]
> >       7ffdc9d33000-7ffdc9d54000 rw-p 00000000 00:00 0                          [stack]
> >       7ffe1ced4000-7ffe1cef5000 rw-p 00000000 00:00 0                          [stack]
> >       7ffcb0440000-7ffcb0461000 rw-p 00000000 00:00 0                          [stack]
> >       7ffc84515000-7ffc84536000 rw-p 00000000 00:00 0                          [stack]
> >       7ffe557dd000-7ffe557fe000 rw-p 00000000 00:00 0                          [stack]
> >       7ffdcb7e4000-7ffdcb805000 rw-p 00000000 00:00 0                          [stack]
> >       7ffc6f989000-7ffc6f9aa000 rw-p 00000000 00:00 0                          [stack]
> >       7ffd9322a000-7ffd9324b000 rw-p 00000000 00:00 0                          [stack]
> >       7ffea8e34000-7ffea8e55000 rw-p 00000000 00:00 0                          [stack]
> >       7ffc594ff000-7ffc59520000 rw-p 00000000 00:00 0                          [stack]
> >       7ffe8db65000-7ffe8db86000 rw-p 00000000 00:00 0                          [stack]
> >
> > After this patch(bit0 ~ bit29 randomized):
> >       root@...ux86-64:~# for i in {1..20};do cat /proc/self/maps | grep stack;done
> >       7fffd0ed2000-7fffd0ef3000 rw-p 00000000 00:00 0                          [stack]
> >       7fffdf555000-7fffdf576000 rw-p 00000000 00:00 0                          [stack]
> >       7ffffec0e000-7ffffec2f000 rw-p 00000000 00:00 0                          [stack]
> >       7fffcfa0f000-7fffcfa30000 rw-p 00000000 00:00 0                          [stack]
> >       7fffff75d000-7fffff77e000 rw-p 00000000 00:00 0                          [stack]
> >       7fffd1d11000-7fffd1d32000 rw-p 00000000 00:00 0                          [stack]
> >       7fffc7850000-7fffc7871000 rw-p 00000000 00:00 0                          [stack]
> >       7fffc956e000-7fffc958f000 rw-p 00000000 00:00 0                          [stack]
> >       7fffdf7be000-7fffdf7df000 rw-p 00000000 00:00 0                          [stack]
> >       7fffeffdf000-7ffff0000000 rw-p 00000000 00:00 0                          [stack]
> >       7fffd7d7f000-7fffd7da0000 rw-p 00000000 00:00 0                          [stack]
> >       7fffe3d23000-7fffe3d44000 rw-p 00000000 00:00 0                          [stack]
> >       7fffde4b7000-7fffde4d8000 rw-p 00000000 00:00 0                          [stack]
> >       7fffd477a000-7fffd479b000 rw-p 00000000 00:00 0                          [stack]
> >       7fffc0e95000-7fffc0eb6000 rw-p 00000000 00:00 0                          [stack]
> >       7fffdbdda000-7fffdbdfb000 rw-p 00000000 00:00 0                          [stack]
> >       7fffec737000-7fffec758000 rw-p 00000000 00:00 0                          [stack]
> >       7fffed163000-7fffed184000 rw-p 00000000 00:00 0                          [stack]
> >       7ffff9a9c000-7ffff9abd000 rw-p 00000000 00:00 0                          [stack]
> >       7fffff945000-7fffff966000 rw-p 00000000 00:00 0                          [stack]
> >
> > Cc: Thomas Gleixner <tglx@...utronix.de>
> > Cc: Ingo Molnar <mingo@...hat.com>
> > Cc: Borislav Petkov <bp@...en8.de>
> > Cc: "H. Peter Anvin" <hpa@...or.com>
> > Cc: Gabriel Krisman Bertazi <krisman@...labora.com>
> > Cc: Andy Lutomirski <luto@...nel.org>
> > Cc: Kees Cook <keescook@...omium.org>
> > Cc: "Chang S. Bae" <chang.seok.bae@...el.com>
> > Cc: Al Viro <viro@...iv.linux.org.uk>
> > Cc: linux-kernel@...r.kernel.org
> > Signed-off-by: Xiongwei Song <sxwjean@...il.com>
> > ---
> >  arch/x86/include/asm/elf.h | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/arch/x86/include/asm/elf.h b/arch/x86/include/asm/elf.h
> > index 29fea180a665..0b7a5510ebd5 100644
> > --- a/arch/x86/include/asm/elf.h
> > +++ b/arch/x86/include/asm/elf.h
> > @@ -343,7 +343,7 @@ extern unsigned long get_sigframe_size(void);
> >  #else /* CONFIG_X86_32 */
> >
> >  /* 1GB for 64bit, 8MB for 32bit */
> > -#define __STACK_RND_MASK(is32bit) ((is32bit) ? 0x7ff : 0x3fffff)
> > +#define __STACK_RND_MASK(is32bit) ((is32bit) ? 0x7ff : 0x3ffff)
> >  #define STACK_RND_MASK __STACK_RND_MASK(mmap_is_ia32())
> >
> >  #define ARCH_DLINFO                                                  \
> > --
> > 2.30.2
> >
>
> --
> Kees Cook

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ