* [PATCH] cfgrtl: Don't try to redirect asm goto to EXIT [PR108263]
@ 2023-01-03 10:28 Jakub Jelinek
2023-01-03 11:01 ` Richard Biener
0 siblings, 1 reply; 2+ messages in thread
From: Jakub Jelinek @ 2023-01-03 10:28 UTC (permalink / raw)
To: Richard Biener, Jeff Law; +Cc: gcc-patches
Hi!
The following testcase distilled from Linux kernel on ppc64le ICEs,
because fixup_reorder_chain sees a bb with a single fallthru edge
falling into a bb with simple return and decides to redirect
that fallthru edge to EXIT. That is possible if the bb ending
in the fallthru edge doesn't end with a jump or ends with a normal
unconditional jump, but not when the bb ends with asm goto which can despite
a single fallthru have multiple labels to the fallthrough basic block.
The following patch makes sure we never try to redirect such cases to EXIT.
Bootstrapped/regtested on x86_64-linux, i686-linux and s390x-linux, ok for
trunk?
2023-01-03 Jakub Jelinek <jakub@redhat.com>
PR rtl-optimization/108263
* cfgrtl.cc (fixup_reorder_chain): Avoid trying to redirect
asm goto to EXIT.
* gcc.dg/pr108263.c: New test.
--- gcc/cfgrtl.cc.jj 2023-01-02 09:32:32.000000000 +0100
+++ gcc/cfgrtl.cc 2023-01-02 14:13:40.658338194 +0100
@@ -3910,6 +3910,7 @@ fixup_reorder_chain (void)
rtx ret_label = NULL_RTX;
basic_block nb;
edge_iterator ei;
+ bool asm_goto = false;
if (EDGE_COUNT (bb->succs) == 0)
continue;
@@ -4016,7 +4017,9 @@ fixup_reorder_chain (void)
|| e_fall->dest == EXIT_BLOCK_PTR_FOR_FN (cfun))
continue;
- /* Otherwise we'll have to use the fallthru fixup below. */
+ /* Otherwise we'll have to use the fallthru fixup below.
+ But avoid redirecting asm goto to EXIT. */
+ asm_goto = true;
}
else
{
@@ -4048,7 +4051,8 @@ fixup_reorder_chain (void)
return rather than a jump to the return block. */
rtx_insn *ret, *use;
basic_block dest;
- if (bb_is_just_return (e_fall->dest, &ret, &use)
+ if (!asm_goto
+ && bb_is_just_return (e_fall->dest, &ret, &use)
&& ((PATTERN (ret) == simple_return_rtx && targetm.have_simple_return ())
|| (PATTERN (ret) == ret_rtx && targetm.have_return ())))
{
--- gcc/testsuite/gcc.dg/pr108263.c.jj 2023-01-02 14:09:49.381614027 +0100
+++ gcc/testsuite/gcc.dg/pr108263.c 2023-01-02 14:07:16.899772608 +0100
@@ -0,0 +1,25 @@
+/* PR rtl-optimization/108263 */
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+
+int v, *p;
+
+void
+foo (void)
+{
+ int i;
+ for (i = 0; ; i++)
+ {
+ if (v)
+ {
+ __label__ l1;
+ asm goto ("" : : : : l1);
+ l1:
+ return;
+ }
+ if (p[i])
+ break;
+ }
+ asm goto ("" : : "r" (i) : : l2);
+l2:;
+}
Jakub
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] cfgrtl: Don't try to redirect asm goto to EXIT [PR108263]
2023-01-03 10:28 [PATCH] cfgrtl: Don't try to redirect asm goto to EXIT [PR108263] Jakub Jelinek
@ 2023-01-03 11:01 ` Richard Biener
0 siblings, 0 replies; 2+ messages in thread
From: Richard Biener @ 2023-01-03 11:01 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: Jeff Law, gcc-patches
> Am 03.01.2023 um 11:29 schrieb Jakub Jelinek via Gcc-patches <gcc-patches@gcc.gnu.org>:
>
> Hi!
>
> The following testcase distilled from Linux kernel on ppc64le ICEs,
> because fixup_reorder_chain sees a bb with a single fallthru edge
> falling into a bb with simple return and decides to redirect
> that fallthru edge to EXIT. That is possible if the bb ending
> in the fallthru edge doesn't end with a jump or ends with a normal
> unconditional jump, but not when the bb ends with asm goto which can despite
> a single fallthru have multiple labels to the fallthrough basic block.
>
> The following patch makes sure we never try to redirect such cases to EXIT.
>
> Bootstrapped/regtested on x86_64-linux, i686-linux and s390x-linux, ok for
> trunk?
Ok
Richard
> 2023-01-03 Jakub Jelinek <jakub@redhat.com>
>
> PR rtl-optimization/108263
> * cfgrtl.cc (fixup_reorder_chain): Avoid trying to redirect
> asm goto to EXIT.
>
> * gcc.dg/pr108263.c: New test.
>
> --- gcc/cfgrtl.cc.jj 2023-01-02 09:32:32.000000000 +0100
> +++ gcc/cfgrtl.cc 2023-01-02 14:13:40.658338194 +0100
> @@ -3910,6 +3910,7 @@ fixup_reorder_chain (void)
> rtx ret_label = NULL_RTX;
> basic_block nb;
> edge_iterator ei;
> + bool asm_goto = false;
>
> if (EDGE_COUNT (bb->succs) == 0)
> continue;
> @@ -4016,7 +4017,9 @@ fixup_reorder_chain (void)
> || e_fall->dest == EXIT_BLOCK_PTR_FOR_FN (cfun))
> continue;
>
> - /* Otherwise we'll have to use the fallthru fixup below. */
> + /* Otherwise we'll have to use the fallthru fixup below.
> + But avoid redirecting asm goto to EXIT. */
> + asm_goto = true;
> }
> else
> {
> @@ -4048,7 +4051,8 @@ fixup_reorder_chain (void)
> return rather than a jump to the return block. */
> rtx_insn *ret, *use;
> basic_block dest;
> - if (bb_is_just_return (e_fall->dest, &ret, &use)
> + if (!asm_goto
> + && bb_is_just_return (e_fall->dest, &ret, &use)
> && ((PATTERN (ret) == simple_return_rtx && targetm.have_simple_return ())
> || (PATTERN (ret) == ret_rtx && targetm.have_return ())))
> {
> --- gcc/testsuite/gcc.dg/pr108263.c.jj 2023-01-02 14:09:49.381614027 +0100
> +++ gcc/testsuite/gcc.dg/pr108263.c 2023-01-02 14:07:16.899772608 +0100
> @@ -0,0 +1,25 @@
> +/* PR rtl-optimization/108263 */
> +/* { dg-do compile } */
> +/* { dg-options "-O2" } */
> +
> +int v, *p;
> +
> +void
> +foo (void)
> +{
> + int i;
> + for (i = 0; ; i++)
> + {
> + if (v)
> + {
> + __label__ l1;
> + asm goto ("" : : : : l1);
> + l1:
> + return;
> + }
> + if (p[i])
> + break;
> + }
> + asm goto ("" : : "r" (i) : : l2);
> +l2:;
> +}
>
> Jakub
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2023-01-03 11:01 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-01-03 10:28 [PATCH] cfgrtl: Don't try to redirect asm goto to EXIT [PR108263] Jakub Jelinek
2023-01-03 11:01 ` 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).