On Thu, 31 Mar 2011 21:28:21 +0200, Steven Rostedt wrote:

WARN_ON() should never do anything but test. That ret++ does not belong
inside the WARN_ON() condition. If there are other locations in the
kernel that do that, then those locations need to be fixed.

Testing implies evaluating, so if we allow:

     if (++i == end) { /* ... */ }

I see no reason why not to allow:

     if (WARN_ON(++i == end)) { /* ... */ }

In both cases the condition is tested.

quoted

On Thu, 2011-03-31 at 15:16 +0200, Marek Szyprowski wrote:

quoted

+       ret = 0;
+       while (!PageBuddy(pfn_to_page(start & (~0UL << ret))))
+               if (WARN_ON(++ret >= MAX_ORDER))
+                       return -EINVAL;

On Thu, Mar 31, 2011 at 09:02:41AM -0700, Dave Hansen wrote:

quoted

In any case, please pull the ++ret bit out of the WARN_ON().  Some
people like to do:

#define WARN_ON(...) do{}while(0)

to save space on some systems.

On Thu, 31 Mar 2011 21:26:50 +0200, Steven Rostedt wrote:

That should be fixed, as the if (WARN_ON()) has become a standard in
most of the kernel. Removing WARN_ON() should be:

#define WARN_ON(x) ({0;})

This would break a lot of code which expect that testing to take place.
Also see <http://lxr.linux.no/linux+*/include/asm-generic/bug.h#L108>.

But I agree, that there should be no "side effects" inside a WARN_ON(),
which that "++ret" is definitely one.

Thus I don't really agree with this point.

At any rate, I don't really care.

-- 
Best regards,                                         _     _
.o. | Liege of Serenely Enlightened Majesty of      o' \,=./ `o
..o | Computer Science,  Michal "mina86" Nazarewicz    (o o)
ooo +-----<email/xmpp: mnazarewicz@google.com>-----ooO--(_)--Ooo--