svn commit: r295768 - head/usr.sbin/iostat

[Gleb Smirnoff]

On Fri, Feb 19, 2016 at 04:50:40PM -0800, Conrad Meyer wrote:
C> > It is not quite as simple as this would make it sound.  The elements or
C> > members of an aggregate (e.g.) structure type are initialized as if it were
C> > an object of static storage duration (i.e., to zero) if the initializer list
C> > does not contain enough initializers for all members of the aggregate type,
C> > per item 21 of section 6.7.8 of n1256.pdf.  However, such initialization
C> > does not necessarily need to zero any padding bytes that are present, which
C> > may take unspecified values.  Personally, I think this particular clang
C> > warning can be too aggressive, especially for complex structs, but on the
C> > other hand given the indeterminateness of padding, bzero/memset are often a
C> > better choice anyway.
C> 
C> By definition, padding byte contents are unused.  There is no reason
C> their values matter one way or another, so why do we care about the
C> distinction between bzero and member zero initialization?

Well, one can run bcmp() or a hashing function over entire structure,
in that case contents of padding matter.

But I still don't want to see this warning in -Wall, since it kills
very useful C99 initializers.

The code, that runs bcmp() or hashes over structs, should take care, and
we shouldn't pessimize the entire build for its sake.

-- 
Totus tuus, Glebius.