[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-10-31 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

Andrew Pinski  changed:

   What|Removed |Added

   Target Milestone|--- |12.0

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-25 Thread mpolacek at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

Marek Polacek  changed:

   What|Removed |Added

 Resolution|--- |FIXED
 Status|ASSIGNED|RESOLVED

--- Comment #8 from Marek Polacek  ---
Fixed.

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-25 Thread cvs-commit at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

--- Comment #7 from CVS Commits  ---
The trunk branch has been updated by Marek Polacek :

https://gcc.gnu.org/g:1ba397e9f93d3abc93a6ecbabc3d873489a6fb7f

commit r12-8248-g1ba397e9f93d3abc93a6ecbabc3d873489a6fb7f
Author: Marek Polacek 
Date:   Fri Apr 22 19:40:27 2022 -0400

c++: __builtin_shufflevector with value-dep expr [PR105353]

Here we issue an error from c_build_shufflevector while parsing a template
because it got a TEMPLATE_PARM_INDEX, but this function expects
INTEGER_CSTs
(except the first two arguments).  It checks if any of the arguments are
type-dependent, if so, we leave the processing for later, but it should
also check value-dependency for the 3rd+ arguments, so as to avoid the
problem above.

This is not a regression -- __builtin_shufflevector was introduced in
GCC 12, but it looks safe enough.

PR c++/105353

gcc/cp/ChangeLog:

* typeck.cc (build_x_shufflevector): Use
instantiation_dependent_expression_p except for the first two
arguments.

gcc/testsuite/ChangeLog:

* g++.dg/ext/builtin-shufflevector-3.C: New test.

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-22 Thread mpolacek at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

--- Comment #6 from Marek Polacek  ---
Good point, I suppose this is better:

--- a/gcc/cp/typeck.cc
+++ b/gcc/cp/typeck.cc
@@ -6315,7 +6315,9 @@ build_x_shufflevector (location_t loc, vec
*args,
   if (processing_template_decl)
 {
   for (unsigned i = 0; i < args->length (); ++i)
-   if (type_dependent_expression_p ((*args)[i]))
+   if (i <= 1
+   ? type_dependent_expression_p ((*args)[i])
+   : instantiation_dependent_expression_p ((*args)[i]))
  {
tree exp = build_min_nt_call_vec (NULL, args);
CALL_EXPR_IFN (exp) = IFN_SHUFFLEVECTOR;

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-22 Thread jakub at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

--- Comment #5 from Jakub Jelinek  ---
Maybe, but for i <= 1 IMHO type_dependent_expression_p is right, that is why we
build_non_dependent_expr, c_build_shufflevector oesn't care about those exact
values.

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-22 Thread mpolacek at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

--- Comment #4 from Marek Polacek  ---
My fix is just

--- a/gcc/cp/typeck.cc
+++ b/gcc/cp/typeck.cc
@@ -6315,7 +6315,7 @@ build_x_shufflevector (location_t loc, vec
*args,
   if (processing_template_decl)
 {
   for (unsigned i = 0; i < args->length (); ++i)
-   if (type_dependent_expression_p ((*args)[i]))
+   if (instantiation_dependent_expression_p ((*args)[i]))
  {
tree exp = build_min_nt_call_vec (NULL, args);
CALL_EXPR_IFN (exp) = IFN_SHUFFLEVECTOR;


so what we leave the IFN_SHUFFLEVECTOR to be processed when instantiating. 
fold_non_dependent_expr wouldn't know what to substitute T with yet.

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-22 Thread jakub at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

Jakub Jelinek  changed:

   What|Removed |Added

 CC||jakub at gcc dot gnu.org

--- Comment #3 from Jakub Jelinek  ---
As c_build_shufflevector requires that the 3rd+ args are all INTEGER_CSTs that
fit into shwi, I think we need to avoid calling that function if any of those
isn't such and processing_template_decl.  Perhaps we can
fold_non_dependent_expr them first.

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-22 Thread mpolacek at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

Marek Polacek  changed:

   What|Removed |Added

   Assignee|unassigned at gcc dot gnu.org  |mpolacek at gcc dot 
gnu.org
 Status|NEW |ASSIGNED
 CC||mpolacek at gcc dot gnu.org

--- Comment #2 from Marek Polacek  ---
Right.  I have a fix.

[Bug c++/105353] __builtin_shufflevector with value dependent constant

2022-04-22 Thread pinskia at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105353

Andrew Pinski  changed:

   What|Removed |Added

Summary|__builtin_shufflevector |__builtin_shufflevector
   |with template parameter |with value dependent
   |fails to compile on GCC 12  |constant
   |but compiles on clang   |
 Ever confirmed|0   |1
   Last reconfirmed||2022-04-22
 Status|UNCONFIRMED |NEW

--- Comment #1 from Andrew Pinski  ---
Confirmed, it is all value dependent constants really. e.g. even sizeof fails:



#include 

typedef std::uint8_t Simd128U8VectT __attribute__((__vector_size__(16)));

#define ShuffleIndex sizeof(T)

template
static inline Simd128U8VectT ShufFunc(Simd128U8VectT vect) noexcept {
if constexpr(unsigned(ShuffleIndex) >= 16)
return Simd128U8VectT { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
};
else if constexpr(ShuffleIndex == 0)
return vect;
else
return __builtin_shufflevector(vect, vect, ShuffleIndex, ShuffleIndex +
1,
ShuffleIndex + 2, ShuffleIndex + 3, ShuffleIndex + 4, ShuffleIndex
+ 5,
ShuffleIndex + 6, ShuffleIndex + 7, ShuffleIndex + 8, ShuffleIndex
+ 9,
ShuffleIndex + 10, ShuffleIndex + 11, ShuffleIndex + 12,
ShuffleIndex + 13,
ShuffleIndex + 14, ShuffleIndex + 15);
}