public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r13-5965] tree-optimization/108691 - indirect calls to setjmp
@ 2023-02-13 14:57 Richard Biener
0 siblings, 0 replies; only message in thread
From: Richard Biener @ 2023-02-13 14:57 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:6436add49ed9ecf74f1f01b4fe18961df4f3e83d
commit r13-5965-g6436add49ed9ecf74f1f01b4fe18961df4f3e83d
Author: Richard Biener <rguenther@suse.de>
Date: Mon Feb 13 10:41:51 2023 +0100
tree-optimization/108691 - indirect calls to setjmp
DCE now chokes on indirect setjmp calls becoming direct because
that exposes them too late to be subject to abnormal edge creation.
The following patch honors gimple_call_ctrl_altering for those and
_not_ treat formerly indirect calls to setjmp as calls to setjmp
in notice_special_calls.
Unfortunately there's no way to have an indirect call to setjmp
properly annotated (the returns_twice attribute is ignored on types).
RTL expansion late discovers returns-twice for the purpose of
adding REG_SETJMP notes and also sets ->calls_setjmp
(instead of asserting it is set). There's no good way to
transfer proper knowledge around here so I'm using ->calls_setjmp
as a flag to indicate whether gimple_call_ctrl_altering_p was set.
PR tree-optimization/108691
* tree-cfg.cc (notice_special_calls): When the CFG is built
honor gimple_call_ctrl_altering_p.
* cfgexpand.cc (expand_call_stmt): Clear cfun->calls_setjmp
temporarily if the call is not control-altering.
* calls.cc (emit_call_1): Do not add REG_SETJMP if
cfun->calls_setjmp is not set. Do not alter cfun->calls_setjmp.
* gcc.dg/pr108691.c: New testcase.
Diff:
---
gcc/calls.cc | 10 +++++-----
gcc/cfgexpand.cc | 7 +++++++
gcc/testsuite/gcc.dg/pr108691.c | 9 +++++++++
gcc/tree-cfg.cc | 4 +++-
4 files changed, 24 insertions(+), 6 deletions(-)
diff --git a/gcc/calls.cc b/gcc/calls.cc
index 4d7f6c3d291..0242d52cfb3 100644
--- a/gcc/calls.cc
+++ b/gcc/calls.cc
@@ -506,11 +506,11 @@ emit_call_1 (rtx funexp, tree fntree ATTRIBUTE_UNUSED, tree fndecl ATTRIBUTE_UNU
if (ecf_flags & ECF_NORETURN)
add_reg_note (call_insn, REG_NORETURN, const0_rtx);
- if (ecf_flags & ECF_RETURNS_TWICE)
- {
- add_reg_note (call_insn, REG_SETJMP, const0_rtx);
- cfun->calls_setjmp = 1;
- }
+ if (ecf_flags & ECF_RETURNS_TWICE
+ /* We rely on GIMPLE setting this flag and here use it to
+ catch formerly indirect and not control-altering calls. */
+ && cfun->calls_setjmp)
+ add_reg_note (call_insn, REG_SETJMP, const0_rtx);
SIBLING_CALL_P (call_insn) = ((ecf_flags & ECF_SIBCALL) != 0);
diff --git a/gcc/cfgexpand.cc b/gcc/cfgexpand.cc
index 25b1558dcb9..ab143a6d2d3 100644
--- a/gcc/cfgexpand.cc
+++ b/gcc/cfgexpand.cc
@@ -2808,6 +2808,11 @@ expand_call_stmt (gcall *stmt)
/* Must come after copying location. */
copy_warning (exp, stmt);
+ /* For calls that do not alter control flow avoid REG_SETJMP notes. */
+ bool saved_calls_setjmp = cfun->calls_setjmp;
+ if (!gimple_call_ctrl_altering_p (stmt))
+ cfun->calls_setjmp = false;
+
/* Ensure RTL is created for debug args. */
if (decl && DECL_HAS_DEBUG_ARGS_P (decl))
{
@@ -2846,6 +2851,8 @@ expand_call_stmt (gcall *stmt)
}
mark_transaction_restart_calls (stmt);
+
+ cfun->calls_setjmp = saved_calls_setjmp;
}
diff --git a/gcc/testsuite/gcc.dg/pr108691.c b/gcc/testsuite/gcc.dg/pr108691.c
new file mode 100644
index 00000000000..e412df10f22
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr108691.c
@@ -0,0 +1,9 @@
+/* { dg-do compile } */
+/* { dg-options "-O2" } */
+
+extern int __attribute__((returns_twice)) setjmp(void*);
+
+void bbb(void) {
+ int (*fnptr)(void*) = setjmp;
+ fnptr(0);
+}
diff --git a/gcc/tree-cfg.cc b/gcc/tree-cfg.cc
index a9fcc7fd050..e23293e5cd1 100644
--- a/gcc/tree-cfg.cc
+++ b/gcc/tree-cfg.cc
@@ -2280,7 +2280,9 @@ notice_special_calls (gcall *call)
if (flags & ECF_MAY_BE_ALLOCA)
cfun->calls_alloca = true;
- if (flags & ECF_RETURNS_TWICE)
+ if (flags & ECF_RETURNS_TWICE
+ && (!(cfun->curr_properties & PROP_cfg)
+ || gimple_call_ctrl_altering_p (call)))
cfun->calls_setjmp = true;
}
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2023-02-13 14:57 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-02-13 14:57 [gcc r13-5965] tree-optimization/108691 - indirect calls to setjmp 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).