* Re: [Kernel-janitors] mm/slab.c: linux 2.6.1 fix 2 unguarded kmalloc and a PAGE_SHIFT
[not found] <vines.sxdD+dlw=B2?= =?ISO-8859-1?Q?+A@SZKOM.BFS.DE>
@ 2004-01-29 2:35 ` Matthew Wilcox
0 siblings, 0 replies; only message in thread
From: Matthew Wilcox @ 2004-01-29 2:35 UTC (permalink / raw)
To: Walter Harms; +Cc: kernel-janitors, linux-kernel
On Sun, Jan 25, 2004 at 03:02:24PM +0100, Walter Harms wrote:
> Hi list,
> this fixes catches 2 unguarded kmallocs() and changes a statement so that PAGE_SHIFT >20 causes a warning.
> At least sparc64 is prepared for a PAGE_SHIFT >20.
> - if (num_physpages > (32 << 20) >> PAGE_SHIFT)
> + if (num_physpages > (32 << (20-PAGE_SHIFT) )
Erm, that's a bad change to make. There's nothing wrong with having
a PAGE_SHIFT of >20, and this code would still work ... but with your
change it's now undefined behaviour. Please don't make this change.
(worked examples:
1. suppose we have a PAGE_SHIFT of 24 (16MB page size), the original
code shifts 32 left by 20 then right by 24, net result a right shift of
4 => 2. Since any machine with a 16MB page size is going to have more
than 2 pages, we're fine. the new code will shift 32 left by -4 which
is undefined. bad.
2. suppose we have a PAGE_SHIFT of 26 (64MB page size), the original
code shifts right by 6, result 0. Since the number of pages is > 0
(one hopes ;-), the test will also succeed as desired.
)
--
"Next the statesmen will invent cheap lies, putting the blame upon
the nation that is attacked, and every man will be glad of those
conscience-soothing falsities, and will diligently study them, and refuse
to examine any refutations of them; and thus he will by and by convince
himself that the war is just, and will thank God for the better sleep
he enjoys after this process of grotesque self-deception." -- Mark Twain
^ permalink raw reply [flat|nested] only message in thread