linux-kernel - Re: [PATCH] x86/mm: Clean up the __[PHYSICAL/VIRTUAL]_MASK

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: <20190509172244.GA11274@yury-thinkpad>
Date:   Thu, 9 May 2019 10:22:44 -0700
From:   Yury Norov <yury.norov@...il.com>
To:     Ingo Molnar <mingo@...nel.org>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>, Andi Kleen <ak@...ux.intel.com>,
        Michal Hocko <mhocko@...e.com>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Dave Hansen <dave.hansen@...el.com>,
        Yury Norov <ynorov@...vell.com>, x86@...nel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] x86/mm: Clean up the __[PHYSICAL/VIRTUAL]_MASK_SHIFT
 definitions a bit

On Thu, May 09, 2019 at 11:01:31AM +0200, Ingo Molnar wrote:
> 
> * Yury Norov <yury.norov@...il.com> wrote:
> 
> > __VIRTUAL_MASK_SHIFT is defined twice to the same valie in
> > arch/x86/include/asm/page_32_types.h. Fix it.
> > 
> > Signed-off-by: Yury Norov <ynorov@...vell.com>
> > ---
> >  arch/x86/include/asm/page_32_types.h | 5 ++---
> >  1 file changed, 2 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/x86/include/asm/page_32_types.h b/arch/x86/include/asm/page_32_types.h
> > index 0d5c739eebd7..9bfac5c80d89 100644
> > --- a/arch/x86/include/asm/page_32_types.h
> > +++ b/arch/x86/include/asm/page_32_types.h
> > @@ -28,6 +28,8 @@
> >  #define MCE_STACK 0
> >  #define N_EXCEPTION_STACKS 1
> >  
> > +#define __VIRTUAL_MASK_SHIFT	32
> > +
> >  #ifdef CONFIG_X86_PAE
> >  /*
> >   * This is beyond the 44 bit limit imposed by the 32bit long pfns,
> > @@ -36,11 +38,8 @@
> >   * The real limit is still 44 bits.
> >   */
> >  #define __PHYSICAL_MASK_SHIFT	52
> > -#define __VIRTUAL_MASK_SHIFT	32
> > -
> >  #else  /* !CONFIG_X86_PAE */
> >  #define __PHYSICAL_MASK_SHIFT	32
> > -#define __VIRTUAL_MASK_SHIFT	32
> >  #endif	/* CONFIG_X86_PAE */
> 
> I think it's clearer to see them defined where the physical mask shift is 
> defined.
> 
> How about the patch below? It does away with the weird formatting and 
> cleans up both the comments and the style of the definition:
> 
> /*
>  * 52 bits on PAE is beyond the 44-bit limit imposed by the
>  * 32-bit long PFNs, but we need the full mask to make sure
>  * inverted PROT_NONE entries have all the host bits set
>  * in a guest. The real limit is still 44 bits.
>  */
> #ifdef CONFIG_X86_PAE
> # define __PHYSICAL_MASK_SHIFT	52
> # define __VIRTUAL_MASK_SHIFT	32
> #else
> # define __PHYSICAL_MASK_SHIFT	32
> # define __VIRTUAL_MASK_SHIFT	32
> #endif
> 
> ?

My main concern was about double definition. It pretty looks like a
bug. But if it's intentional, I'm OK. In the patch below, could you
please add some note to the comment that __VIRTUAL_MASK_SHIFT defined
twice intentionally?

Thanks,
Yury
 
> Thanks,
> 
> 	Ingo
> 
> ===============>
> From: Ingo Molnar <mingo@...nel.org>
> Date: Thu, 9 May 2019 10:59:44 +0200
> Subject: [PATCH] x86/mm: Clean up the __[PHYSICAL/VIRTUAL]_MASK_SHIFT definitions a bit
> 
> Signed-off-by: Ingo Molnar <mingo@...nel.org>
> ---
>  arch/x86/include/asm/page_32_types.h | 23 +++++++++++------------
>  1 file changed, 11 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/x86/include/asm/page_32_types.h b/arch/x86/include/asm/page_32_types.h
> index 565ad755c785..009e96d4b6d4 100644
> --- a/arch/x86/include/asm/page_32_types.h
> +++ b/arch/x86/include/asm/page_32_types.h
> @@ -26,20 +26,19 @@
>  
>  #define N_EXCEPTION_STACKS	1
>  
> -#ifdef CONFIG_X86_PAE
>  /*
> - * This is beyond the 44 bit limit imposed by the 32bit long pfns,
> - * but we need the full mask to make sure inverted PROT_NONE
> - * entries have all the host bits set in a guest.
> - * The real limit is still 44 bits.
> + * 52 bits on PAE is beyond the 44-bit limit imposed by the
> + * 32-bit long PFNs, but we need the full mask to make sure
> + * inverted PROT_NONE entries have all the host bits set
> + * in a guest. The real limit is still 44 bits.
>   */
> -#define __PHYSICAL_MASK_SHIFT	52
> -#define __VIRTUAL_MASK_SHIFT	32
> -
> -#else  /* !CONFIG_X86_PAE */
> -#define __PHYSICAL_MASK_SHIFT	32
> -#define __VIRTUAL_MASK_SHIFT	32
> -#endif	/* CONFIG_X86_PAE */
> +#ifdef CONFIG_X86_PAE
> +# define __PHYSICAL_MASK_SHIFT	52
> +# define __VIRTUAL_MASK_SHIFT	32
> +#else
> +# define __PHYSICAL_MASK_SHIFT	32
> +# define __VIRTUAL_MASK_SHIFT	32
> +#endif
>  
>  /*
>   * Kernel image size is limited to 512 MB (see in arch/x86/kernel/head_32.S)