> From: Martin Sebor <mse...@gmail.com> > Date: Wed, 8 Jul 2020 02:09:37 +0200
> On 7/6/20 10:02 PM, Hans-Peter Nilsson via Gcc-patches wrote: > > We say very little about reads and writes to aggregate / > > compound objects, just scalar objects (i.e. assignments don't > > cause reads). Let's lets say something safe about aggregate > > objects, but only for those that are the same size as a scalar > > type. > > > > There's an equal-sounding section (Volatiles) in extend.texi, > > but this seems a more appropriate place, as specifying the > > behavior of a standard qualifier. > > > > gcc: > > PR middle-end/94600 > > * doc/implement-c.texi (Qualifiers implementation): Add blurb > > about access to the whole of a volatile aggregate object, only for > > same-size as scalar object. > > --- > > gcc/doc/implement-c.texi | 4 ++++ > > 1 file changed, 4 insertions(+) > > > > diff --git a/gcc/doc/implement-c.texi b/gcc/doc/implement-c.texi > > index 692297b69c4..d64922b28ad 100644 > > --- a/gcc/doc/implement-c.texi > > +++ b/gcc/doc/implement-c.texi > > @@ -576,6 +576,10 @@ are of scalar types, the expression is interpreted by > > GCC as a read of > > the volatile object; in the other cases, the expression is only evaluated > > for its side effects. > > > > +When an object of aggregate type has the same size as a scalar type, GCC > > +handles an access to the whole of that volatile aggregate type object > > +equal to an access to that volatile same-sized scalar type object. > > The grammar is a bit off here making the sentence difficult to > parse and interpret. (Thanks for fixing that.) > Richard already pointed out the alignment > requirement But, treating scalar and aggregate types of certain sizes equal, *also* implies that alignment requirements are equal. > but I'm also wondering if the statement is meant to > apply to accesses by library functions such as memcpy. I definitely *do not* want this to cover library functions. Having said that, I think memcpy would be the only applicable one, as you'd need to cast away volatile for other library functions. Still: no. Not going there. > I suspect > it should only apply to assignments (either simple or atomic), > correct? Yes: I just want to generalized and extend to all supported integer types (certainly target-dependent), that what's goes for reading and writing a volatile 8- 16- and 32-bit integer (and 64 where sizeof(uintptr_t) == 8) also goes for an aggregate of the same size, see PR middle-end/94600. > Would something like this be more accurate? > > When an object of an aggregate type with the same size and > alignment as a scalar type S is the subject of a volatile > access by an assignment expression or an atomic function, > the access to it is performed as if the object's declared > type were volatile S. I'd say it's equal in terms of contents, but if it has better flow to it when read, I'm all for it. Thanks! Care to patchualize it? (Heh, grammar.) brgds, H-P
