linux-kernel - Re: [PATCH 08/32] dmaengine: ste_dma40: Optimise local MAX() macro

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 18 Apr 2013 12:00:24 +0100
From:	Russell King - ARM Linux <linux@....linux.org.uk>
To:	Arnd Bergmann <arnd@...db.de>
Cc:	Lee Jones <lee.jones@...aro.org>, Rabin Vincent <rabin@....in>,
	linus.walleij@...ricsson.com, Vinod Koul <vinod.koul@...el.com>,
	linux-kernel@...r.kernel.org,
	Per Forlin <per.forlin@...ricsson.com>,
	Dan Williams <djbw@...com>,
	linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH 08/32] dmaengine: ste_dma40: Optimise local MAX() macro

On Thu, Apr 18, 2013 at 12:46:03PM +0200, Arnd Bergmann wrote:
> On Thursday 18 April 2013, Lee Jones wrote:
> > The current implementation of the DMA40's local MAX() macro evaluates
> > its arguments more times than is necessary. This patch strips it
> > optimises it to only evaluate what's appropriate.
> 
> No, it does not.
> 
> > index b21a8a3..7b451b2 100644
> > --- a/drivers/dma/ste_dma40.c
> > +++ b/drivers/dma/ste_dma40.c
> > @@ -53,7 +53,7 @@
> >  #define D40_ALLOC_PHY          BIT(30)
> >  #define D40_ALLOC_LOG_FREE     BIT(0)
> >  
> > -#define MAX(a, b) (((a) < (b)) ? (b) : (a))
> > +#define MAX(a, b) ((a > b) ? a : b)
> 
> This just makes the macro worse in case you pass a complex expression
> in, which can now have additional unwanted effects. Just drop this patch.

Never got the original patch...

A much better idea is to get rid of that buggy MAX() macro altogether
and use the macros already provided by the kernel, which are safe from
side effects - but more importantly are _type_ _safe_.  The above goes
wrong when you consider 'a' and 'b' may have different signed-ness.

Consider:

	int val_in = -5;
	unsigned val = MAX(val_in, 5U);

The resulting value is (unsigned)-5, not (unsigned)5.

Best use the kernel's max() or max_t() _everywhere_.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/