diff mbox series

x86-64: Define ASM_OUTPUT_ALIGNED_DECL_LOCAL

Message ID 20200626210217.884877-1-hjl.tools@gmail.com
State New
Headers show
Series x86-64: Define ASM_OUTPUT_ALIGNED_DECL_LOCAL | expand

Commit Message

H.J. Lu June 26, 2020, 9:02 p.m. UTC 
Define ASM_OUTPUT_ALIGNED_DECL_LOCAL for large local common symbol.

gcc/ChangeLog:

	* config/i386/x86-64.h (ASM_OUTPUT_ALIGNED_DECL_LOCAL): New.

gcc/testsuite/ChangeLog:

	* gcc.target/i386/pr95620.c: New test.
---
 gcc/config/i386/x86-64.h                | 11 +++++++++++
 gcc/testsuite/gcc.target/i386/pr95620.c | 19 +++++++++++++++++++
 2 files changed, 30 insertions(+)
 create mode 100644 gcc/testsuite/gcc.target/i386/pr95620.c

Comments

H.J. Lu July 12, 2020, 12:05 p.m. UTC | #1 
On Fri, Jun 26, 2020 at 2:02 PM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> Define ASM_OUTPUT_ALIGNED_DECL_LOCAL for large local common symbol.
>
> gcc/ChangeLog:
>
>         * config/i386/x86-64.h (ASM_OUTPUT_ALIGNED_DECL_LOCAL): New.
>
> gcc/testsuite/ChangeLog:
>
>         * gcc.target/i386/pr95620.c: New test.

PING:

https://gcc.gnu.org/pipermail/gcc-patches/2020-June/549001.html
H.J. Lu July 18, 2020, 3:08 p.m. UTC | #2 
On Sun, Jul 12, 2020 at 5:05 AM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Fri, Jun 26, 2020 at 2:02 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > Define ASM_OUTPUT_ALIGNED_DECL_LOCAL for large local common symbol.
> >
> > gcc/ChangeLog:
> >
> >         * config/i386/x86-64.h (ASM_OUTPUT_ALIGNED_DECL_LOCAL): New.
> >
> > gcc/testsuite/ChangeLog:
> >
> >         * gcc.target/i386/pr95620.c: New test.
>
> PING:
>
> https://gcc.gnu.org/pipermail/gcc-patches/2020-June/549001.html
>

Any objections to check it in next Monday?
Jan Hubicka July 18, 2020, 3:15 p.m. UTC | #3 
> On Sun, Jul 12, 2020 at 5:05 AM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > On Fri, Jun 26, 2020 at 2:02 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> > >
> > > Define ASM_OUTPUT_ALIGNED_DECL_LOCAL for large local common symbol.
> > >
> > > gcc/ChangeLog:
> > >
> > >         * config/i386/x86-64.h (ASM_OUTPUT_ALIGNED_DECL_LOCAL): New.
> > >
> > > gcc/testsuite/ChangeLog:
> > >
> > >         * gcc.target/i386/pr95620.c: New test.
> >
> > PING:
> >
> > https://gcc.gnu.org/pipermail/gcc-patches/2020-June/549001.html
> >
> 
> Any objections to check it in next Monday?
Looks good to me.

Honza
> 
> 
> -- 
> H.J.
H.J. Lu July 18, 2020, 3:55 p.m. UTC | #4 
On Sat, Jul 18, 2020 at 8:15 AM Jan Hubicka <hubicka@ucw.cz> wrote:
>
> > On Sun, Jul 12, 2020 at 5:05 AM H.J. Lu <hjl.tools@gmail.com> wrote:
> > >
> > > On Fri, Jun 26, 2020 at 2:02 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> > > >
> > > > Define ASM_OUTPUT_ALIGNED_DECL_LOCAL for large local common symbol.
> > > >
> > > > gcc/ChangeLog:
> > > >
> > > >         * config/i386/x86-64.h (ASM_OUTPUT_ALIGNED_DECL_LOCAL): New.
> > > >
> > > > gcc/testsuite/ChangeLog:
> > > >
> > > >         * gcc.target/i386/pr95620.c: New test.
> > >
> > > PING:
> > >
> > > https://gcc.gnu.org/pipermail/gcc-patches/2020-June/549001.html
> > >
> >
> > Any objections to check it in next Monday?
> Looks good to me.
>
> Honza

This is the patch I am checking in.  I moved the test to
libgomp/testsuite/libgomp.c
since the test fails at link-time without the fix.

Thanks.
H.J. Lu July 24, 2020, 11:12 a.m. UTC | #5 
On Sat, Jul 18, 2020 at 8:55 AM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Sat, Jul 18, 2020 at 8:15 AM Jan Hubicka <hubicka@ucw.cz> wrote:
> >
> > > On Sun, Jul 12, 2020 at 5:05 AM H.J. Lu <hjl.tools@gmail.com> wrote:
> > > >
> > > > On Fri, Jun 26, 2020 at 2:02 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> > > > >
> > > > > Define ASM_OUTPUT_ALIGNED_DECL_LOCAL for large local common symbol.
> > > > >
> > > > > gcc/ChangeLog:
> > > > >
> > > > >         * config/i386/x86-64.h (ASM_OUTPUT_ALIGNED_DECL_LOCAL): New.
> > > > >
> > > > > gcc/testsuite/ChangeLog:
> > > > >
> > > > >         * gcc.target/i386/pr95620.c: New test.
> > > >
> > > > PING:
> > > >
> > > > https://gcc.gnu.org/pipermail/gcc-patches/2020-June/549001.html
> > > >
> > >
> > > Any objections to check it in next Monday?
> > Looks good to me.
> >
> > Honza
>
> This is the patch I am checking in.  I moved the test to
> libgomp/testsuite/libgomp.c
> since the test fails at link-time without the fix.
>

OK to backport to GCC 10 branch?
diff mbox series

Patch

diff --git a/gcc/config/i386/x86-64.h b/gcc/config/i386/x86-64.h
index 88db428f592..0c5b8af5a13 100644
--- a/gcc/config/i386/x86-64.h
+++ b/gcc/config/i386/x86-64.h
@@ -59,6 +59,17 @@  see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
 #define ASM_OUTPUT_ALIGNED_DECL_COMMON(FILE, DECL, NAME, SIZE, ALIGN)		\
   x86_elf_aligned_decl_common (FILE, DECL, NAME, SIZE, ALIGN);
 
+#undef  ASM_OUTPUT_ALIGNED_DECL_LOCAL
+#define ASM_OUTPUT_ALIGNED_DECL_LOCAL(FILE, DECL, NAME, SIZE, ALIGN)  \
+  do								      \
+    {								      \
+      fprintf ((FILE), "%s", LOCAL_ASM_OP);			      \
+      assemble_name ((FILE), (NAME));				      \
+      fprintf ((FILE), "\n");					      \
+      ASM_OUTPUT_ALIGNED_DECL_COMMON (FILE, DECL, NAME, SIZE, ALIGN); \
+    }								      \
+  while (0)
+
 /* This is used to align code labels according to Intel recommendations.  */
 
 #define SUBALIGN_LOG 3
diff --git a/gcc/testsuite/gcc.target/i386/pr95620.c b/gcc/testsuite/gcc.target/i386/pr95620.c
new file mode 100644
index 00000000000..ac9b40383e7
--- /dev/null
+++ b/gcc/testsuite/gcc.target/i386/pr95620.c
@@ -0,0 +1,19 @@ 
+/* { dg-do assemble { target lp64 } } */
+/* { dg-require-effective-target fopenmp } */
+/* { dg-require-effective-target lto } */
+/* { dg-options "-flto -fopenmp -mcmodel=medium" } */
+
+double a[353783808];
+int b, c, d;
+
+int
+main()
+{
+  for (; b;)
+#pragma omp parallel
+    a[c] = 1;
+  for (;; b++)
+    if (a[c])
+      d++;
+  return 0;
+}