On Mon, Dec 21, 2009 at 10:28:39PM +0300, Alexander Beregalov wrote:
> 2009/12/21 Alexander Beregalov <a.beregalov@xxxxxxxxx>:
> > 2009/12/21 Sam Ravnborg <sam@xxxxxxxxxxxx>:
> >> On Sun, Dec 20, 2009 at 03:16:35PM +0300, Alexander Beregalov wrote:
> >>> Fix warnings which become build errors since -Werror is used.
> >>> arch/sparc/kernel/us3_cpufreq.c:60: error: 'ret' may be used uninitialized in this function
> >>> arch/sparc/kernel/us3_cpufreq.c:101: error: 'new_bits' may be used uninitialized in this function
> >>
> >> What config produces these warnings?
> >> Do you have CONFIG_BUG_VERBOSE set?
> >>
> >> Also what gcc versio?
> >>
> Ah, this is it.
>
> # CONFIG_BUG is not set
>
> Is it sane config ?
Only embedded size constrained configs would
consider this.
The reason for the warning is this snippet from
include/asm-generic/bug.h:
#define BUG() do {} while(0)
The above only comes into effect when CONFIG_BUG is not set.
I dunno what the best fix is here.
Embedded or not - I cannot see why we should continue
execution after a BUG().
Code will assume we panicked and following code not to be executed.
As the comment says:
"Don't use BUG() or BUG_ON() unless there's really no way out;"
I would have expected to see something like this:
#define BUG() do {} while(1)
A never ending loop so the watchdog on embedded systems could
force a reboot. IMO more safe than continue and do random
curruption of stuff all over.
David Howells added the do {} while(0) - so cc: on this mail.
Sam
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
