Re: [PATCH v5 6/6] mtrr, mm, x86: Enhance MTRR checks for KVA huge page mapping

From: Toshi Kani <toshi.kani@hp.com>
To: Borislav Petkov <bp@alien8.de>
Cc: akpm@linux-foundation.org, hpa@zytor.com, tglx@linutronix.de,
	mingo@redhat.com, linux-mm@kvack.org, x86@kernel.org,
	linux-kernel@vger.kernel.org, dave.hansen@intel.com,
	Elliott@hp.com, pebolle@tiscali.nl, mcgrof@suse.com
Subject: Re: [PATCH v5 6/6] mtrr, mm, x86: Enhance MTRR checks for KVA huge page mapping
Date: Mon, 18 May 2015 15:53:14 -0600	[thread overview]
Message-ID: <1431985994.21526.12.camel@misato.fc.hp.com> (raw)
In-Reply-To: <20150518205123.GI23618@pd.tnic>

On Mon, 2015-05-18 at 22:51 +0200, Borislav Petkov wrote:
> On Mon, May 18, 2015 at 02:21:08PM -0600, Toshi Kani wrote:
> > The caller is the one who makes the condition checks necessary to create
> > a huge page mapping.
> 
> How? It would go and change MTRRs configuration and ranges and their
> memory types so that a huge mapping succeeds?
> 
> Or go and try a different range?

Try with a smaller page size.

The callers, pud_set_huge() and pmd_set_huge(), check if the given range
is safe with MTRRs for creating a huge page mapping.  If not, they fail
the request, which leads their callers, ioremap_pud_range() and
ioremap_pmd_range(), to retry with a smaller page size, i.e. 1GB -> 2MB
-> 4KB.  4KB may not have overlap with MTRRs (hence no checking is
necessary), which will succeed as before.

Thanks,
-Toshi

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>