public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Richard Biener <rguenther@suse.de>
To: gcc-patches@gcc.gnu.org
Subject: [PATCH] tree-optimization/98381 - fix live bool vector extract
Date: Tue, 5 Jan 2021 15:54:35 +0100 (CET)	[thread overview]
Message-ID: <nycvar.YFH.7.76.2101051554150.905@elmra.sevgm.obk> (raw)

This fixes extraction of live bool vector results for the case of
integer mode vectors.

Bootstrapped and tested on x86_64-unknown-linux-gnu (and i386.exp with 
SDE), pushed.

2021-01-05  Richard Biener  <rguenther@suse.de>

	PR tree-optimization/98381
	* tree.c (vector_element_bits): Properly compute bool vector
	element size.
	* tree-vect-loop.c (vectorizable_live_operation): Properly
	compute the last lane bit offset.
---
 gcc/tree-vect-loop.c | 5 ++---
 gcc/tree.c           | 9 +++++++--
 2 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/gcc/tree-vect-loop.c b/gcc/tree-vect-loop.c
index 830531f48b8..965cc164f6e 100644
--- a/gcc/tree-vect-loop.c
+++ b/gcc/tree-vect-loop.c
@@ -8494,7 +8494,7 @@ vectorizable_live_operation (vec_info *vinfo,
 {
   loop_vec_info loop_vinfo = dyn_cast <loop_vec_info> (vinfo);
   imm_use_iterator imm_iter;
-  tree lhs, lhs_type, bitsize, vec_bitsize;
+  tree lhs, lhs_type, bitsize;
   tree vectype = (slp_node
 		  ? SLP_TREE_VECTYPE (slp_node)
 		  : STMT_VINFO_VECTYPE (stmt_info));
@@ -8637,7 +8637,6 @@ vectorizable_live_operation (vec_info *vinfo,
   lhs_type = TREE_TYPE (lhs);
 
   bitsize = vector_element_bits_tree (vectype);
-  vec_bitsize = TYPE_SIZE (vectype);
 
   /* Get the vectorized lhs of STMT and the lane to use (counted in bits).  */
   tree vec_lhs, bitstart;
@@ -8661,7 +8660,7 @@ vectorizable_live_operation (vec_info *vinfo,
       vec_lhs = gimple_get_lhs (vec_stmt);
 
       /* Get the last lane in the vector.  */
-      bitstart = int_const_binop (MINUS_EXPR, vec_bitsize, bitsize);
+      bitstart = int_const_binop (MULT_EXPR, bitsize, bitsize_int (nunits - 1));
     }
 
   if (loop_vinfo)
diff --git a/gcc/tree.c b/gcc/tree.c
index 421a2b4bc02..e0a1d512019 100644
--- a/gcc/tree.c
+++ b/gcc/tree.c
@@ -14021,8 +14021,13 @@ vector_element_bits (const_tree type)
 {
   gcc_checking_assert (VECTOR_TYPE_P (type));
   if (VECTOR_BOOLEAN_TYPE_P (type))
-    return vector_element_size (tree_to_poly_uint64 (TYPE_SIZE (type)),
-				TYPE_VECTOR_SUBPARTS (type));
+    {
+      if (VECTOR_MODE_P (TYPE_MODE (type)))
+	return vector_element_size (tree_to_poly_uint64 (TYPE_SIZE (type)),
+				    TYPE_VECTOR_SUBPARTS (type));
+      else
+	return 1;
+    }
   return tree_to_uhwi (TYPE_SIZE (TREE_TYPE (type)));
 }
 
-- 
2.26.2

                 reply	other threads:[~2021-01-05 14:54 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=nycvar.YFH.7.76.2101051554150.905@elmra.sevgm.obk \
    --to=rguenther@suse.de \
    --cc=gcc-patches@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).