linux-kernel - Re: [PATCH v5 03/15] x86/mtrr: replace some constants with defines

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: <20230405202633.GFZC3ZeVWtmjkODYW7@fat_crate.local>
Date:   Wed, 5 Apr 2023 22:26:33 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     Juergen Gross <jgross@...e.com>
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        "H. Peter Anvin" <hpa@...or.com>
Subject: Re: [PATCH v5 03/15] x86/mtrr: replace some constants with defines

On Wed, Apr 05, 2023 at 09:55:59AM +0200, Juergen Gross wrote:
> On 03.04.23 18:03, Borislav Petkov wrote:
> > On Sat, Apr 01, 2023 at 08:36:40AM +0200, Juergen Gross wrote:
> > > @@ -643,10 +646,12 @@ static bool set_mtrr_var_ranges(unsigned int index, struct mtrr_var_range *vr)
> > >   	unsigned int lo, hi;
> > >   	bool changed = false;
> > > +#define BASE_MASK	(MTRR_BASE_TYPE_MASK | (size_and_mask << PAGE_SHIFT))
> > > +#define MASK_MASK	(MTRR_MASK_VALID | (size_and_mask << PAGE_SHIFT))
> > 
> > No, "MASK_MASK" is too much. :-)
> 
> Any better suggestion for the name? :-)

Looking at this again, what this is actually doing is masking out the
reserved bits. But in an unnecessarily complicated way.

What it should do, instead, is do that explicitly:

	/* Zap the reserved bits and compare only the valid fields: */
	if (((vr->base_lo & ~RESV_LOW) != (lo & ~RESV_LOW)) ||
	    ((vr->base_hi & ~RESV_HI) != (hi & ~RESV_HI)))

where

#define RESV_LOW	GENMASK_ULL(8, 11)
#define RESV_HI		GENMASK(phys_addr - 1, 63)

and then we can get rid of that size_or_mask and size_and_mask
stupidity.

I think that would simplify this variable ranges handling code a lot
more and make it pretty straightforward...

> No. The "48" is the _number_ of physical address bits, so the 64 bit address
> mask will be 0000ffff.ffffffff (48 bits set).

Uff, sorry about that. I got confused by that SIZE_OR_MASK_BITS() where
phys_addr and not phys_addr - 1 works because it the arithmetic works
with starting bit 0.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette