From: "Kewen.Lin" <linkw@linux.ibm.com>
To: GCC Patches <gcc-patches@gcc.gnu.org>
Cc: Segher Boessenkool <segher@kernel.crashing.org>,
Bill Schmidt <wschmidt@linux.ibm.com>,
Richard Biener <richard.guenther@gmail.com>
Subject: [PATCH] Fix PR94401 by considering reverse overrun
Date: Thu, 2 Apr 2020 15:15:42 +0800 [thread overview]
Message-ID: <f8d6e4fd-b668-b63f-a737-dec7610af14d@linux.ibm.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 710 bytes --]
Hi,
The commit r10-7415 brings scalar type consideration
to eliminate epilogue peeling for gaps, but it exposed
one problem that the current handling doesn't consider
the memory access type VMAT_CONTIGUOUS_REVERSE, for
which the overrun happens on low address side. This
patch is to make the code take care of it by updating
the offset and construction element order accordingly.
Bootstrapped/regtested on powerpc64le-linux-gnu P8
and aarch64-linux-gnu.
BR,
Kewen
-----------
gcc/ChangeLog
2020-04-02 Kewen Lin <linkw@gcc.gnu.org>
PR tree-optimization/94401
* tree-vect-loop.c (vectorizable_load): Handle VMAT_CONTIGUOUS_REVERSE
access type when loading halves of vector to avoid peeling for gaps.
[-- Attachment #2: PR94401_REVERSE.diff --]
[-- Type: text/plain, Size: 2277 bytes --]
diff --git a/gcc/tree-vect-stmts.c b/gcc/tree-vect-stmts.c
index 12beef6978c..3d27f59ba22 100644
--- a/gcc/tree-vect-stmts.c
+++ b/gcc/tree-vect-stmts.c
@@ -9590,11 +9590,20 @@ vectorizable_load (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi,
if (new_vtype != NULL_TREE)
ltype = half_vtype;
}
+ tree offset = dataref_offset
+ ? dataref_offset
+ : build_int_cst (ref_type, 0);
+ if (ltype != vectype
+ && memory_access_type == VMAT_CONTIGUOUS_REVERSE)
+ offset = size_binop (
+ PLUS_EXPR,
+ build_int_cst (ref_type,
+ DR_GROUP_GAP (first_stmt_info)
+ * tree_to_uhwi (
+ TYPE_SIZE_UNIT (elem_type))),
+ offset);
data_ref
- = fold_build2 (MEM_REF, ltype, dataref_ptr,
- dataref_offset
- ? dataref_offset
- : build_int_cst (ref_type, 0));
+ = fold_build2 (MEM_REF, ltype, dataref_ptr, offset);
if (alignment_support_scheme == dr_aligned)
;
else if (DR_MISALIGNMENT (first_dr_info) == -1)
@@ -9607,16 +9616,27 @@ vectorizable_load (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi,
TYPE_ALIGN (elem_type));
if (ltype != vectype)
{
- vect_copy_ref_info (data_ref, DR_REF (first_dr_info->dr));
+ vect_copy_ref_info (data_ref,
+ DR_REF (first_dr_info->dr));
tree tem = make_ssa_name (ltype);
new_stmt = gimple_build_assign (tem, data_ref);
- vect_finish_stmt_generation (stmt_info, new_stmt, gsi);
+ vect_finish_stmt_generation (stmt_info, new_stmt,
+ gsi);
data_ref = NULL;
vec<constructor_elt, va_gc> *v;
vec_alloc (v, 2);
- CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, tem);
- CONSTRUCTOR_APPEND_ELT (v, NULL_TREE,
- build_zero_cst (ltype));
+ if (memory_access_type == VMAT_CONTIGUOUS_REVERSE)
+ {
+ CONSTRUCTOR_APPEND_ELT (v, NULL_TREE,
+ build_zero_cst (ltype));
+ CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, tem);
+ }
+ else
+ {
+ CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, tem);
+ CONSTRUCTOR_APPEND_ELT (v, NULL_TREE,
+ build_zero_cst (ltype));
+ }
gcc_assert (new_vtype != NULL_TREE);
if (new_vtype == vectype)
new_stmt = gimple_build_assign (
next reply other threads:[~2020-04-02 7:15 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-02 7:15 Kewen.Lin [this message]
2020-04-02 8:28 ` Jakub Jelinek
2020-04-02 10:07 ` Kewen.Lin
2020-04-02 10:39 ` Jakub Jelinek
2020-04-02 12:05 ` Segher Boessenkool
2020-04-02 9:21 ` Richard Biener
2020-04-02 10:33 ` Kewen.Lin
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=f8d6e4fd-b668-b63f-a737-dec7610af14d@linux.ibm.com \
--to=linkw@linux.ibm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=richard.guenther@gmail.com \
--cc=segher@kernel.crashing.org \
--cc=wschmidt@linux.ibm.com \
/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).