2007-08-29 15:45:52

by Peter Lund

Subject: [PATCH] avoid negative (and full-width) shifts in radix-tree.c, take 3

From: Peter Lund <[email protected]>

Negative shifts are not allowed in C (the result is undefined).
Same thing with full-width shifts.

It works on most platforms but not on the VAX with gcc 4.0.1 (it results in an
"operand reserved" fault).

Applies to Linux 2.6.22.

Signed-off-by: Peter Lund <[email protected]>
---
Shifting by more than the width of the value on the left is also not allowed.
I think the extra '>> 1' tacked on at the end in the original code was an attempt
to work around that. Getting rid of that is an extra feature of this patch.

Here's the chapter and verse, taken from the final draft of the C99 standard
("6.5.7 Bitwise shift operators", paragraph 3):

"The integer promotions are performed on each of the operands. The
type of the result is that of the promoted left operand. If the
value of the right operand is negative or is greater than or equal
to the width of the promoted left operand, the behavior is
undefined."

Thank you to Jan-Benedict Glaw, Christoph Hellwig, Maciej Rozycki,
Pekka Enberg, and Andreas Schwab for review. Special thanks to Andreas for
spotting something I knew but overlooked due to misfiring neurons.

I have kept the spacing between the variables and their types so the variable
names and the equal signs don't line up, which should keep the majority of my
reviewers happy.

--- linux-2.6.22/lib/radix-tree.c.orig 2007-07-09 01:32:17.000000000 +0200
+++ linux-2.6.22/lib/radix-tree.c 2007-08-29 17:32:00.000000000 +0200
@@ -980,12 +980,14 @@

static __init unsigned long __maxindex(unsigned int height)
{
- unsigned int tmp = height * RADIX_TREE_MAP_SHIFT;
- unsigned long index = (~0UL >> (RADIX_TREE_INDEX_BITS - tmp - 1)) >> 1;
+ unsigned int width = height * RADIX_TREE_MAP_SHIFT;
+ int shift = RADIX_TREE_INDEX_BITS - width;

- if (tmp >= RADIX_TREE_INDEX_BITS)
- index = ~0UL;
- return index;
+ if (shift < 0)
+ return ~0UL;
+ if (shift >= 8 * sizeof(unsigned long))
+ return 0UL;
+ return ~0UL >> shift;
}

static __init void radix_tree_init_maxindex(void)

2007-08-29 17:49:22

by Christoph Lameter

Subject: Re: [PATCH] avoid negative (and full-width) shifts in radix-tree.c, take 3

On Wed, 29 Aug 2007, Peter Lund wrote:

>
> - if (tmp >= RADIX_TREE_INDEX_BITS)
> - index = ~0UL;
> - return index;
> + if (shift < 0)
> + return ~0UL;
> + if (shift >= 8 * sizeof(unsigned long))

8* sizeof(unsigned long) is the constant BITS_PER_LONG.

2007-08-30 00:46:44

by Björn Steinbrink

Subject: Re: [PATCH] avoid negative (and full-width) shifts in radix-tree.c, take 3

On 2007.08.29 10:49:12 -0700, Christoph Lameter wrote:
> On Wed, 29 Aug 2007, Peter Lund wrote:
>
> >
> > - if (tmp >= RADIX_TREE_INDEX_BITS)
> > - index = ~0UL;
> > - return index;
> > + if (shift < 0)
> > + return ~0UL;
> > + if (shift >= 8 * sizeof(unsigned long))
>
> 8* sizeof(unsigned long) is the constant BITS_PER_LONG.

Or in this context: RADIX_TREE_INDEX_BITS, which happens to be the old
check, which just needed to be extended to cover negative shifts.

Bj?rn