Re: [RFC/PATCH] doc: volatile considered evil

[Randy Dunlap <randy.dunlap@oracle.com>]

On Tue, 8 May 2007 13:20:31 -0700 (PDT) David Rientjes wrote:

> On Tue, 8 May 2007, Krzysztof Halasa wrote:
> 
> > > Since 'volatile' has two different semantics depending on the context in 
> > > which it is used, this warning should be appended to include the fact that 
> > > it is legitimate to use for inline assembly.
> > 
> > I think it hasn't two semantics, it's like arguing that char has two
> > semantics.
> > 
> > Volatile does one thing - prohibits C compiler from optimizing
> > accesses to the variable. Either with (volatile *) casts and with
> > volatile var;
> > 
> 
> What the meaning of an "access" to a volatile memory-mapped I/O port or a 
> variable that can be asynchronously interrupted is implementation-defined.
> 
> You're only citing qualified versions of objects.
> 
> In an asm construct, if all your input operands are modified and specified 
> as output operands as well, volatile must be added so that the entire 
> construct is not optimized away.  Additionally, it must be added if your 
> construct modifies memory that is neither listed in inputs nor outputs to 
> the construct so that it is known to have at least one side-effect.  Then, 
> the compiler cannot delete your construct if it is reachable because it 
> may produce such side-effects.
> 
> Thus, the warning that was proposed for addition to CodingStyle should be 
> modified to explicitly state that the use of 'volatile' for asm constructs 
> is perfectly legitimate and its use as a type qualifier for objects in 
> code is inappropriate.

It's already there, isn't it?  <quote from original:>

The only acceptable uses for "volatile" are:

 - in _code_, i.e., for things like the definition of "readb()" etc, where we
   use it to force a particular access.
 - with inline asms
 - on "jiffies", for stupid legacy reasons

</quote>

or are you saying that you want to subject/header/title modified also?


---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/