Re: [PATCH v4 08/11] xen/compiler: import 'fallthrough' keyword from linux

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Tue, 12 Jan 2021, Jan Beulich wrote:
> On 08.01.2021 15:46, Rahul Singh wrote:
> > -Wimplicit-fallthrough warns when a switch case falls through. Warning
> > can be suppress by either adding a /* fallthrough */ comment, or by
> > using a null statement: __attribute__ ((fallthrough))
> 
> Why is the comment variant (which we use in many places already,
> albeit with varying wording) not the route of choice?

See previous discussion:

https://marc.info/?l=xen-devel&m=160707274517270
https://marc.info/?l=xen-devel&m=160733742810605
https://marc.info/?l=xen-devel&m=160733852011023

We thought it would be best to introduce "fallthrough" and only resort
to comments as a plan B. The usage of the keyword should allow GCC to do
better checks.


> > Define the pseudo keyword 'fallthrough' for the ability to convert the
> > various case block /* fallthrough */ style comments to null statement
> > "__attribute__((__fallthrough__))"
> > 
> > In C mode, GCC supports the __fallthrough__ attribute since 7.1,
> > the same time the warning and the comment parsing were introduced.
> > 
> > fallthrough devolves to an empty "do {} while (0)" if the compiler
> > version (any version less than gcc 7) does not support the attribute.
> 
> What about Coverity? It would be nice if we wouldn't need to add
> two separate constructs everywhere to make both compiler and static
> code checker happy.

I don't think I fully understand your reply here: Coverity doesn't come
into the picture. Given that GCC provides a special keyword to implement
fallthrough, it makes sense to use it when available. When it is not
available (e.g. clang or older GCC) we need to have an alternative to
suppress the compiler warnings. Hence the need for this check:

  #if (!defined(__clang__) && (__GNUC__ >= 7))