LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Ingo Molnar <mingo@elte.hu>
To: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andi Kleen <andi@firstfloor.org>,
	tglx@linutronix.de, linux-kernel@vger.kernel.org
Subject: Re: unify pagetable accessors patch causes double fault II
Date: Tue, 15 Jan 2008 14:53:17 +0100	[thread overview]
Message-ID: <20080115135317.GA13493@elte.hu> (raw)
In-Reply-To: <478BE628.7090008@goop.org>


* Jeremy Fitzhardinge <jeremy@goop.org> wrote:

> Can you try this out?  It applies after "x86: move all asm/pgtable 
> constants into one place".

and here's the patch

Subject: Re: unify pagetable accessors patch causes double fault II
From: Jeremy Fitzhardinge <jeremy@goop.org>

Andi Kleen wrote:
>> OK, I see the problem.  The problem is that the _PAGE_X defines are
>> defined with _AC(UL, 1 << _PAGE_BIT_X), which has unsigned long type.
>> This means that ~_PAGE_X also has unsigned long type, and so when cast
>> to 64-bit in pte_mkX, it ends up &ing the pte with 0x00000000ffffffxxx,
>> with predictable results.
>>
>
> Actually I fixed some of that -- see the pgtable-nx patch on firstfloor -- but
> it still doesn't work. Or maybe my patch was not complete.
>

Yeah, that looks like the right sort of thing, but I wonder if there's
other places doing an open-coded "pte_val(pte) & ~_PAGE_FOO".  My patch
changes the definition of _PAGE_FOO so it should be OK everywhere.

>> The original code just used signed constants for the _PAGE_X
>> definitions, which will sign-extend when cast to 64-bit, and so have the
>> upper bits set when masking.  (Well, actually, the old code just
>> operated on pte_low, so the problem didn't arise; however, pgtable_64.h
>> also uses integers for its _PAGE_X, which has the same sign-extended
>> 32->64 casting property).
>>
>> I'll put together a fixup patch now.
>>
>
> I'm leaving now but can test later.

and below is the fix against full x86.git#mm.

	Ingo

----------->
Subject: x86/pgtable: fix constant sign extension problem
From: Jeremy Fitzhardinge <jeremy@goop.org>

When the _PAGE_FOO constants are defined as (1ul << _PAGE_BIT_FOO), they
become unsigned longs.  In 32-bit PAE mode, these end up being
implicitly cast to 64-bit types when used to manipulate a pte, and
because they're unsigned the top 32-bits are 0, destroying the upper
bits of the pte.

When _PAGE_FOO constants are given a signed integer type, the cast to
64-bits will sign-extend so that the upper bits are all ones,
preserving the upper pte bits in manipulations.

Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Cc: Andi Kleen <ak@suse.de>

Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/asm-x86/pgtable.h |   31 ++++++++++++++++++-------------
 1 file changed, 18 insertions(+), 13 deletions(-)

===================================================================
Index: linux-x86.q/include/asm-x86/pgtable.h
===================================================================
--- linux-x86.q.orig/include/asm-x86/pgtable.h
+++ linux-x86.q/include/asm-x86/pgtable.h
@@ -19,21 +19,26 @@
 #define _PAGE_BIT_UNUSED3	11
 #define _PAGE_BIT_NX           63       /* No execute: only valid after cpuid check */
 
-#define _PAGE_PRESENT	(_AC(1, UL)<<_PAGE_BIT_PRESENT)
-#define _PAGE_RW	(_AC(1, UL)<<_PAGE_BIT_RW)
-#define _PAGE_USER	(_AC(1, UL)<<_PAGE_BIT_USER)
-#define _PAGE_PWT	(_AC(1, UL)<<_PAGE_BIT_PWT)
-#define _PAGE_PCD	((_AC(1, UL)<<_PAGE_BIT_PCD) | _PAGE_PWT)
-#define _PAGE_ACCESSED	(_AC(1, UL)<<_PAGE_BIT_ACCESSED)
-#define _PAGE_DIRTY	(_AC(1, UL)<<_PAGE_BIT_DIRTY)
-#define _PAGE_PSE	(_AC(1, UL)<<_PAGE_BIT_PSE)	/* 2MB page */
-#define _PAGE_GLOBAL	(_AC(1, UL)<<_PAGE_BIT_GLOBAL)	/* Global TLB entry */
+/*
+ * Note: we use _AC(1, L) instead of _AC(1, UL) so that we get a
+ * sign-extended value on 32-bit with all 1's in the upper word,
+ * which preserves the upper pte values on 64-bit ptes:
+ */
+#define _PAGE_PRESENT	(_AC(1, L)<<_PAGE_BIT_PRESENT)
+#define _PAGE_RW	(_AC(1, L)<<_PAGE_BIT_RW)
+#define _PAGE_USER	(_AC(1, L)<<_PAGE_BIT_USER)
+#define _PAGE_PWT	(_AC(1, L)<<_PAGE_BIT_PWT)
+#define _PAGE_PCD	((_AC(1, L)<<_PAGE_BIT_PCD) | _PAGE_PWT)
+#define _PAGE_ACCESSED	(_AC(1, L)<<_PAGE_BIT_ACCESSED)
+#define _PAGE_DIRTY	(_AC(1, L)<<_PAGE_BIT_DIRTY)
+#define _PAGE_PSE	(_AC(1, L)<<_PAGE_BIT_PSE)	/* 2MB page */
+#define _PAGE_GLOBAL	(_AC(1, L)<<_PAGE_BIT_GLOBAL)	/* Global TLB entry */
 /* We redefine PCD to be write combining. PAT bit is not used */
-#define _PAGE_WC	((_AC(1, UL)<<_PAGE_BIT_PCD))
+#define _PAGE_WC	((_AC(1, L)<<_PAGE_BIT_PCD))
 #define _PAGE_CACHE_MASK (_PAGE_PCD)
-#define _PAGE_UNUSED1	(_AC(1, UL)<<_PAGE_BIT_UNUSED1)
-#define _PAGE_UNUSED2	(_AC(1, UL)<<_PAGE_BIT_UNUSED2)
-#define _PAGE_UNUSED3	(_AC(1, UL)<<_PAGE_BIT_UNUSED3)
+#define _PAGE_UNUSED1	(_AC(1, L)<<_PAGE_BIT_UNUSED1)
+#define _PAGE_UNUSED2	(_AC(1, L)<<_PAGE_BIT_UNUSED2)
+#define _PAGE_UNUSED3	(_AC(1, L)<<_PAGE_BIT_UNUSED3)
 
 #if defined(CONFIG_X86_64) || defined(CONFIG_X86_PAE)
 #define _PAGE_NX	(_AC(1, ULL) << _PAGE_BIT_NX)

  parent reply	other threads:[~2008-01-15 13:53 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20080114094814.GA28300@basil.nowhere.org>
2008-01-14 12:56 ` Andi Kleen
2008-01-14 13:06   ` Ingo Molnar
2008-01-14 13:58     ` Andi Kleen
2008-01-14 16:44       ` Jeremy Fitzhardinge
2008-01-14 16:56         ` Ingo Molnar
2008-01-14 17:08           ` Andi Kleen
2008-01-14 17:18             ` unify pagetable accessors patch causes double fault III Andi Kleen
2008-01-14 19:00               ` Ingo Molnar
2008-01-14 19:54                 ` Jeremy Fitzhardinge
2008-01-14 20:15                 ` Andi Kleen
2008-01-14 19:52           ` unify pagetable accessors patch causes double fault II Jeremy Fitzhardinge
2008-01-14 22:03       ` Jeremy Fitzhardinge
2008-01-14 22:23         ` Andi Kleen
2008-01-14 22:46           ` Jeremy Fitzhardinge
2008-01-15  1:05             ` Andi Kleen
2008-01-15  1:32               ` Jeremy Fitzhardinge
2008-01-15  1:38                 ` Andi Kleen
2008-01-15 21:03                   ` [patch] x86: lfence fix Ingo Molnar
2008-01-16  0:44                     ` Andi Kleen
2008-01-15 12:55               ` unify pagetable accessors patch causes double fault II Ingo Molnar
2008-01-15 16:53                 ` Andi Kleen
2008-01-15 13:53             ` Ingo Molnar [this message]
2008-01-15 17:16               ` Folding _PAGE_PWT into _PAGE_PCD (was Re: unify pagetable accessors patch causes double fault II) Jeremy Fitzhardinge
2008-01-15 17:23                 ` Andi Kleen
2008-01-15 17:32                   ` Jeremy Fitzhardinge
2008-01-15 17:39                     ` Andi Kleen
2008-01-15 20:30                 ` Venki Pallipadi
2008-01-15 20:45                   ` Jeremy Fitzhardinge
2008-01-15 17:36               ` unify pagetable accessors patch causes double fault II Andi Kleen
2008-01-15 19:43                 ` Jeremy Fitzhardinge

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20080115135317.GA13493@elte.hu \
    --to=mingo@elte.hu \
    --cc=andi@firstfloor.org \
    --cc=jeremy@goop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=tglx@linutronix.de \
    --subject='Re: unify pagetable accessors patch causes double fault II' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).