[Bug middle-end/41992] New: ICE on invalid dereferencing of void *

Hi

This piece of invalid code (the dereference shouldn't be there) triggers an
ICE. The crash happens on 4.3.2, 4.4.1 and 4.4.2. It happens with or without
optimizations.

static void MONITOR(void *ptr)
{
__asm__ volatile ("                                     \n\
XORL    %%ECX, %%ECX                                    \n\
XORL    %%EDX, %%EDX                                    \n\
MONITOR                                                 \n\
"::"a"(*ptr):"cx","dx","cc","memory");
}

s.e: In function 'MONITOR':
s.e:7: warning: dereferencing 'void *' pointer
s.e:7: internal compiler error: in gimplify_expr, at gimplify.c:7074
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.


--
           Summary: ICE on invalid dereferencing of void *
           Product: gcc
           Version: 4.4.1
            Status: UNCONFIRMED
          Severity: minor
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: mikulas at artax dot karlin dot mff dot cuni dot cz
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41992

------- Comment #1 from rguenth at gcc dot gnu dot org  2009-11-09 10:16 -------
Which is because the gimplifier thinks this is an error for the frontend to
pass through:

Breakpoint 1, fancy_abort (
    file=0x10b1ab8 "/space/rguenther/src/svn/trunk/gcc/gimplify.c", line=7173,
    function=0x10b2ca4 "gimplify_expr")
    at /space/rguenther/src/svn/trunk/gcc/diagnostic.c:728
728       internal_error ("in %s, at %s:%d", function, trim_filename (file),
line);
(gdb) up
#1  0x00000000007fa0a5 in gimplify_expr (expr_p=0x7ffff6f07a48,
    pre_p=0x7fffffffced8, post_p=0x7fffffffc558,
    gimple_test_f=0x7b9486 <is_gimple_asm_val>, fallback=1)
    at /space/rguenther/src/svn/trunk/gcc/gimplify.c:7173
7173          gcc_assert (!VOID_TYPE_P (TREE_TYPE (*expr_p)));
(gdb)
#2  0x00000000007e9450 in gimplify_asm_expr (expr_p=0x7ffff7ede980,
    pre_p=0x7fffffffced8, post_p=0x7fffffffc558)
    at /space/rguenther/src/svn/trunk/gcc/gimplify.c:4896
4896              tret = gimplify_expr (&TREE_VALUE (link), pre_p, post_p,

the C standard doesn't claim dereferencing a void pointer is invalid, so
the gimplifier should deal with this.

Before tree-ssa we ICEd with

t5.c:3: internal compiler error: in int_mode_for_mode, at stor-layout.c:289
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

or with 2.95 with

t5.c:3: Internal compiler error in `emit_move_insn_1', at expr.c:2835

so, not a regression.


--

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |ice-on-invalid-code
   Last reconfirmed|0000-00-00 00:00:00         |2009-11-09 10:16:57
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41992

------- Comment #2 from joseph at codesourcery dot com  2009-11-09 13:16 -------
Subject: Re:  ICE on invalid dereferencing of void *

On Mon, 9 Nov 2009, rguenth at gcc dot gnu dot org wrote:

> the C standard doesn't claim dereferencing a void pointer is invalid, so
> the gimplifier should deal with this.

It's not invalid in general, but if the value gets used (NB conditional
expressions involving such dereferences, whose values don't get used, are
OK; DR 106) then it is invalid and there should be an error; "void value
not ignored as it ought to be" is used in some places.  Though I note that
if you change the dereference in the present testcase to use the return
value of a function returning void, the resulting error "using result of
function returning 'void'" comes from the gimplifier and not the front
end, so making the gimplifier deal with this case as well would at least
be consistent.


--


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41992

------- Comment #3 from mikulas at artax dot karlin dot mff dot cuni dot cz  2009-11-11 21:06 -------
You can dereference void * in asm arguments --- i.e.
void *p; ... asm volatile ("prefetch %0"::"m"(*p));

gcc warns in this case about the dereference and maybe it shouldn't (but it's
trivial to supress the warning with a cast to char *).

If you change "m" constraint to "mr", you get an ICE.


--


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41992