On Wed, Jan 16, 2019 at 3:30 AM Jakub Jelinek <ja...@redhat.com> wrote:
>
> On Mon, Jan 14, 2019 at 10:00:11AM -0800, H.J. Lu wrote:
> > On Mon, Jan 14, 2019 at 6:22 AM Jakub Jelinek <ja...@redhat.com> wrote:
> > >
> > > On Sun, Jan 13, 2019 at 06:54:05AM -0800, H.J. Lu wrote:
> > > > > What always matters is whether we take address of a packed structure
> > > > > field/non-static data member or whether we just read that field.
> > > > > The former should be warned about, the latter not.
> > > > >
> > > >
> > > > How about this patch? It checks if address is taken with NOP.
> > >
> > > I'd like to first understand the convert_p argument to
> > > warn_for_address_or_pointer_of_packed_member.
> > >
> > > To me it seems you want to emit two different warnings, perhaps one
> > > surpressed if the other one is emitted, but you actually from the start
> > > decide which of the two you are going to check for. That is just weird.
> >
> > convert_p is only for C.
>
> Why? What is so special about C and (implicit?) casts where the rhs isn't
> ADDR_EXPR? Aren't all casts (explicit or implicit) from one pointer type
> to another pointer and satisfy the rules something that should be warned
-Wincompatible-pointer-types is C only. In C++, incompatible pointer types
aren't allowed at all.
> about if the conditions are met?
>
convert_p is set to TRUE to handle this case for incompatible pointer types:
[hjl@gnu-cfl-1 pr51628-3]$ cat x.i
struct B { int i; };
struct C { struct B b; } __attribute__ ((packed));
extern struct C *p;
long* g8 (void) { return p; }
[hjl@gnu-cfl-1 pr51628-3]$ make x.s
/export/build/gnu/tools-build/gcc-debug/build-x86_64-linux/gcc/xgcc
-B/export/build/gnu/tools-build/gcc-debug/build-x86_64-linux/gcc/ -O2
-S x.i
x.i: In function \u2018g8\u2019:
x.i:6:26: warning: returning \u2018struct C *\u2019 from a function
with incompatible return type \u2018long int *\u2019
[-Wincompatible-pointer-types]
6 | long* g8 (void) { return p; }
| ^
x.i:6:1: warning: converting a packed \u2018struct C *\u2019 pointer
(alignment 1) to \u2018long int *\u2019 (alignment 8) may result in an
unaligned pointer value [-Waddress-of-packed-member]
6 | long* g8 (void) { return p; }
| ^~~~
x.i:2:8: note: defined here
2 | struct C { struct B b; } __attribute__ ((packed));
| ^
[hjl@gnu-cfl-1 pr51628-3]$
I am using the same function to warn both addresses and pointers. I need a way
to tell that we are converting pointers, not assigning address to pointer. This
info is readily available in the front-end.
--
H.J.
