* [PATCH] Fix PR tree-opt/63184: add simplification of (& + A) != (& + B)
@ 2021-09-06 5:30 apinski
2021-09-06 7:36 ` Richard Biener
0 siblings, 1 reply; 2+ messages in thread
From: apinski @ 2021-09-06 5:30 UTC (permalink / raw)
To: gcc-patches; +Cc: Andrew Pinski
From: Andrew Pinski <apinski@marvell.com>
These two testcases have been failing since GCC 5 but things
have improved such that adding a simplification to match.pd
for this case is easier than before.
In the end we have the following IR:
....
_5 = &a[1] + _4;
_7 = &a + _13;
if (_5 != _7)
So we can fold the _5 != _7 into:
(&a[1] - &a) + _4 != _13
The subtraction is folded into constant by ptr_difference_const.
In this case, the full expression gets folded into a constant
and we are able to remove the if statement.
OK? Bootstrapped and tested on aarch64-linux-gnu with no regressions.
gcc/ChangeLog:
* match.pd: Add simplification of pointer_diff of two pointer_plus
with addr_expr in the first operand of each pointer_plus.
Add simplificatoin of ne/eq of two pointer_plus with addr_expr
in the first operand of each pointer_plus.
gcc/testsuite/ChangeLog:
* c-c++-common/pr19807-2.c: Enable for all targets and remove the xfail.
* c-c++-common/pr19807-3.c: Likewise.
---
gcc/match.pd | 15 +++++++++++++++
gcc/testsuite/c-c++-common/pr19807-2.c | 5 ++---
gcc/testsuite/c-c++-common/pr19807-3.c | 5 ++---
3 files changed, 19 insertions(+), 6 deletions(-)
diff --git a/gcc/match.pd b/gcc/match.pd
index f920bc4b7c1..cc7809dfe0f 100644
--- a/gcc/match.pd
+++ b/gcc/match.pd
@@ -2063,6 +2063,21 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
(if (ptr_difference_const (@0, @1, &diff))
{ build_int_cst_type (type, diff); }))))
+/* (&a+b) - (&a[1] + c) -> sizeof(a[0]) + (b - c) */
+(simplify
+ (pointer_diff (pointer_plus ADDR_EXPR@0 @1) (pointer_plus ADDR_EXPR@2 @3))
+ (with { poly_int64 diff; }
+ (if (ptr_difference_const (@0, @2, &diff))
+ (plus { build_int_cst_type (type, diff); } (convert (minus @1 @3))))))
+
+/* (&a+b) !=/== (&a[1] + c) -> sizeof(a[0]) + b !=/== c */
+(for neeq (ne eq)
+ (simplify
+ (neeq (pointer_plus ADDR_EXPR@0 @1) (pointer_plus ADDR_EXPR@2 @3))
+ (with { poly_int64 diff; tree inner_type = TREE_TYPE (@1);}
+ (if (ptr_difference_const (@0, @2, &diff))
+ (neeq (plus { build_int_cst_type (inner_type, diff); } @1) @3)))))
+
/* Canonicalize (T *)(ptr - ptr-cst) to &MEM[ptr + -ptr-cst]. */
(simplify
(convert (pointer_diff @0 INTEGER_CST@1))
diff --git a/gcc/testsuite/c-c++-common/pr19807-2.c b/gcc/testsuite/c-c++-common/pr19807-2.c
index d2c010140d0..529b9c97322 100644
--- a/gcc/testsuite/c-c++-common/pr19807-2.c
+++ b/gcc/testsuite/c-c++-common/pr19807-2.c
@@ -1,5 +1,4 @@
-/* Some targets can optimize this on RTL. */
-/* { dg-do link { target { x86_64-*-* i?86-*-* } } } */
+/* { dg-do link } */
/* { dg-options "-O -fdump-tree-optimized" } */
extern void link_error(void);
@@ -12,4 +11,4 @@ int main()
return 0;
}
-/* { dg-final { scan-tree-dump-not "link_error" "optimized" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-not "link_error" "optimized" } } */
diff --git a/gcc/testsuite/c-c++-common/pr19807-3.c b/gcc/testsuite/c-c++-common/pr19807-3.c
index bb7f9827725..31c88f3b850 100644
--- a/gcc/testsuite/c-c++-common/pr19807-3.c
+++ b/gcc/testsuite/c-c++-common/pr19807-3.c
@@ -1,5 +1,4 @@
-/* Some targets can optimize this on RTL. */
-/* { dg-do link { target { x86_64-*-* i?86-*-* } } } */
+/* { dg-do link } */
/* { dg-options "-O -fdump-tree-optimized" } */
extern void link_error(void);
@@ -12,4 +11,4 @@ int main()
return 0;
}
-/* { dg-final { scan-tree-dump-not "link_error" "optimized" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-not "link_error" "optimized" } } */
--
2.17.1
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] Fix PR tree-opt/63184: add simplification of (& + A) != (& + B)
2021-09-06 5:30 [PATCH] Fix PR tree-opt/63184: add simplification of (& + A) != (& + B) apinski
@ 2021-09-06 7:36 ` Richard Biener
0 siblings, 0 replies; 2+ messages in thread
From: Richard Biener @ 2021-09-06 7:36 UTC (permalink / raw)
To: Andrew Pinski; +Cc: GCC Patches
On Mon, Sep 6, 2021 at 7:31 AM apinski--- via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> From: Andrew Pinski <apinski@marvell.com>
>
> These two testcases have been failing since GCC 5 but things
> have improved such that adding a simplification to match.pd
> for this case is easier than before.
> In the end we have the following IR:
> ....
> _5 = &a[1] + _4;
> _7 = &a + _13;
> if (_5 != _7)
>
> So we can fold the _5 != _7 into:
> (&a[1] - &a) + _4 != _13
>
> The subtraction is folded into constant by ptr_difference_const.
> In this case, the full expression gets folded into a constant
> and we are able to remove the if statement.
>
> OK? Bootstrapped and tested on aarch64-linux-gnu with no regressions.
OK.
Thanks,
Richard.
> gcc/ChangeLog:
>
> * match.pd: Add simplification of pointer_diff of two pointer_plus
> with addr_expr in the first operand of each pointer_plus.
> Add simplificatoin of ne/eq of two pointer_plus with addr_expr
> in the first operand of each pointer_plus.
>
> gcc/testsuite/ChangeLog:
>
> * c-c++-common/pr19807-2.c: Enable for all targets and remove the xfail.
> * c-c++-common/pr19807-3.c: Likewise.
> ---
> gcc/match.pd | 15 +++++++++++++++
> gcc/testsuite/c-c++-common/pr19807-2.c | 5 ++---
> gcc/testsuite/c-c++-common/pr19807-3.c | 5 ++---
> 3 files changed, 19 insertions(+), 6 deletions(-)
>
> diff --git a/gcc/match.pd b/gcc/match.pd
> index f920bc4b7c1..cc7809dfe0f 100644
> --- a/gcc/match.pd
> +++ b/gcc/match.pd
> @@ -2063,6 +2063,21 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
> (if (ptr_difference_const (@0, @1, &diff))
> { build_int_cst_type (type, diff); }))))
>
> +/* (&a+b) - (&a[1] + c) -> sizeof(a[0]) + (b - c) */
> +(simplify
> + (pointer_diff (pointer_plus ADDR_EXPR@0 @1) (pointer_plus ADDR_EXPR@2 @3))
> + (with { poly_int64 diff; }
> + (if (ptr_difference_const (@0, @2, &diff))
> + (plus { build_int_cst_type (type, diff); } (convert (minus @1 @3))))))
> +
> +/* (&a+b) !=/== (&a[1] + c) -> sizeof(a[0]) + b !=/== c */
> +(for neeq (ne eq)
> + (simplify
> + (neeq (pointer_plus ADDR_EXPR@0 @1) (pointer_plus ADDR_EXPR@2 @3))
> + (with { poly_int64 diff; tree inner_type = TREE_TYPE (@1);}
> + (if (ptr_difference_const (@0, @2, &diff))
> + (neeq (plus { build_int_cst_type (inner_type, diff); } @1) @3)))))
> +
> /* Canonicalize (T *)(ptr - ptr-cst) to &MEM[ptr + -ptr-cst]. */
> (simplify
> (convert (pointer_diff @0 INTEGER_CST@1))
> diff --git a/gcc/testsuite/c-c++-common/pr19807-2.c b/gcc/testsuite/c-c++-common/pr19807-2.c
> index d2c010140d0..529b9c97322 100644
> --- a/gcc/testsuite/c-c++-common/pr19807-2.c
> +++ b/gcc/testsuite/c-c++-common/pr19807-2.c
> @@ -1,5 +1,4 @@
> -/* Some targets can optimize this on RTL. */
> -/* { dg-do link { target { x86_64-*-* i?86-*-* } } } */
> +/* { dg-do link } */
> /* { dg-options "-O -fdump-tree-optimized" } */
>
> extern void link_error(void);
> @@ -12,4 +11,4 @@ int main()
> return 0;
> }
>
> -/* { dg-final { scan-tree-dump-not "link_error" "optimized" { xfail *-*-* } } } */
> +/* { dg-final { scan-tree-dump-not "link_error" "optimized" } } */
> diff --git a/gcc/testsuite/c-c++-common/pr19807-3.c b/gcc/testsuite/c-c++-common/pr19807-3.c
> index bb7f9827725..31c88f3b850 100644
> --- a/gcc/testsuite/c-c++-common/pr19807-3.c
> +++ b/gcc/testsuite/c-c++-common/pr19807-3.c
> @@ -1,5 +1,4 @@
> -/* Some targets can optimize this on RTL. */
> -/* { dg-do link { target { x86_64-*-* i?86-*-* } } } */
> +/* { dg-do link } */
> /* { dg-options "-O -fdump-tree-optimized" } */
>
> extern void link_error(void);
> @@ -12,4 +11,4 @@ int main()
> return 0;
> }
>
> -/* { dg-final { scan-tree-dump-not "link_error" "optimized" { xfail *-*-* } } } */
> +/* { dg-final { scan-tree-dump-not "link_error" "optimized" } } */
> --
> 2.17.1
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-09-06 7:36 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-09-06 5:30 [PATCH] Fix PR tree-opt/63184: add simplification of (& + A) != (& + B) apinski
2021-09-06 7:36 ` Richard Biener
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).