Re: [patch v3 2/3] percpu_counter: fix code for 32bit systems forUP

From: Eric Dumazet
Date: Wed Apr 13 2011 - 22:35:38 EST


Le jeudi 14 avril 2011 Ã 10:04 +0800, shaohua.li@xxxxxxxxx a Ãcrit :
> piÃce jointe document texte brut (percpu-counter-32bits.patch)
> percpu_counter.counter is a 's64'. Accessing it in 32-bit system is racing.
> we need some locking to protect it otherwise some very wrong value could be
> accessed.
>
> Signed-off-by: Shaohua Li <shaohua.li@xxxxxxxxx>
> ---
> include/linux/percpu_counter.h | 37 ++++++++++++++++++++++++++-----------
> 1 file changed, 26 insertions(+), 11 deletions(-)
>
> Index: linux/include/linux/percpu_counter.h
> ===================================================================
> --- linux.orig/include/linux/percpu_counter.h 2011-04-14 08:50:45.000000000 +0800
> +++ linux/include/linux/percpu_counter.h 2011-04-14 09:09:14.000000000 +0800
> @@ -89,9 +89,20 @@ struct percpu_counter {
> s64 count;
> };
>
> -static inline int percpu_counter_init(struct percpu_counter *fbc, s64 amount)
> +static inline void percpu_counter_set(struct percpu_counter *fbc, s64 amount)
> {
> +#if BITS_PER_LONG == 32
> + preempt_disable();
> fbc->count = amount;
> + preempt_enable();
> +#else
> + fbc->count = amount;
> +#endif
> +}
> +
> +static inline int percpu_counter_init(struct percpu_counter *fbc, s64 amount)
> +{
> + percpu_counter_set(fbc, amount);
> return 0;
> }
>

Please dont do this and obfuscate the code.

percpu_counter_init() cannot be racy.

Its like saying spin_lock_init() could be racy.

No other thread/cpu is supposed to use the object if not yet
initialized.



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/