mbox series

[0/2] tree_nonzero_bits vs vector and complex types

Message ID 1667425595-2654-1-git-send-email-apinski@marvell.com
Headers
Series tree_nonzero_bits vs vector and complex types |

Message

Li, Pan2 via Gcc-patches Nov. 2, 2022, 9:46 p.m. UTC 
  From: Andrew Pinski <apinski@marvell.com>


While looking at older unconfirmed bug reports, I noticed there was
an ubsan found issue and noticed tree_nonzero_bits was being called with
a vector type. How ubsan found it was at the end of tree_nonzero_bits,
did "return wi::shwi (-1, TYPE_PRECISION (TREE_TYPE (t)));" and
it was with a vector of 1 elements which meant precision was 0
as precision stores the log2 of the number of elements in a vector.

Anyways we want to catch these kind of errors of calling tree_nonzero_bits
with a vector or a complex type. And fix the places where it is called.

Thanks,
Andrew Pinski


Andrew Pinski (2):
  Fix PR 105532: match.pd patterns calling tree_nonzero_bits with vector
    types
  Add assert for type on tree_nonzero_bits

 gcc/fold-const.cc                             |  3 +++
 gcc/match.pd                                  | 25 +++++++++++--------
 .../gcc.c-torture/compile/vector-shift-1.c    |  8 ++++++
 3 files changed, 25 insertions(+), 11 deletions(-)
 create mode 100644 gcc/testsuite/gcc.c-torture/compile/vector-shift-1.c