[PATCH] [gdb] Fix segfault during inferior call to ifunc

[PATCH] [gdb] Fix segfault during inferior call to ifunc

[tdevries]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: 0001-gdb-Fix-segfault-during-inferior-call-to-ifunc.patch --]
[-- Type: text/x-patch; name=0001-gdb-Fix-segfault-during-inferior-call-to-ifunc.patch, Size: 3438 bytes --]

From 0832e20dde6ebe71a1245935f4a41308368b6094 Mon Sep 17 00:00:00 2001
From: Andrew Burgess <aburgess@redhat.com>
Date: Mon, 26 Dec 2022 09:03:19 +0100
Subject: [PATCH] [gdb] Fix segfault during inferior call to ifunc

With a simple test-case:
...
$ cat test.c
char *p = "a";
int main (void) {
  return strlen (p);
}
$ gcc -g test.c
...
we run into this segfault:
...
$ gdb -q -batch a.out -ex start -ex "p strlen (p)"
Temporary breakpoint 1 at 0x1151: file test.c, line 4.
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

Temporary breakpoint 1, main () at test.c:4
4	  return strlen (p);

Fatal signal: Segmentation fault
...

The strlen is an ifunc, and consequently during the call to
call_function_by_hand_dummy for "p strlen (p)" another call
to call_function_by_hand_dummy is used to resolve the ifunc.

This invalidates the get_current_frame () result in the outer call.

Fix this by using prepare_reinflate and reinflate.

Note that this series (
https://inbox.sourceware.org/gdb-patches/20221214033441.499512-1-simon.marchi@polymtl.ca/ )
should address this problem, but this patch is a simpler fix which is easy to
backport.

Tested on x86_64-linux.

Co-Authored-By: Tom de Vries <tdevries@suse.de>
PR gdb/29941
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=29941
---
 gdb/infcall.c                        | 3 +++
 gdb/testsuite/gdb.base/gnu-ifunc.exp | 9 +++++++--
 2 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/gdb/infcall.c b/gdb/infcall.c
index c1db3e22189..ec9669a6b6a 100644
--- a/gdb/infcall.c
+++ b/gdb/infcall.c
@@ -848,6 +848,7 @@ call_function_by_hand_dummy (struct value *function,
   bool stack_temporaries = thread_stack_temporaries_enabled_p (call_thread.get ());
 
   frame = get_current_frame ();
+  frame.prepare_reinflate ();
   gdbarch = get_frame_arch (frame);
 
   if (!gdbarch_push_dummy_call_p (gdbarch))
@@ -863,6 +864,8 @@ call_function_by_hand_dummy (struct value *function,
 	     "target calling convention."),
 	   get_function_name (funaddr, name_buf, sizeof (name_buf)));
 
+  frame.reinflate ();
+
   if (values_type == NULL || values_type->is_stub ())
     values_type = default_return_type;
   if (values_type == NULL)
diff --git a/gdb/testsuite/gdb.base/gnu-ifunc.exp b/gdb/testsuite/gdb.base/gnu-ifunc.exp
index 715f5be04cb..3b3d5cb3742 100644
--- a/gdb/testsuite/gdb.base/gnu-ifunc.exp
+++ b/gdb/testsuite/gdb.base/gnu-ifunc.exp
@@ -240,14 +240,19 @@ proc misc_tests {resolver_attr resolver_debug final_debug} {
     # Test GDB will automatically indirect the call.
 
     if {!$resolver_debug && !$final_debug} {
+	# Do the test that is supposed to succeed first, to make sure
+	# elf_gnu_ifunc_record_cache is empty.  This excercises PR28224.
+	gdb_test "p (int) gnu_ifunc (3)" " = 4"
+
 	gdb_test "p gnu_ifunc()" \
 	    "'${dot}final' has unknown return type; cast the call to its declared return type"
 	gdb_test "p gnu_ifunc (3)" \
 	    "'${dot}final' has unknown return type; cast the call to its declared return type"
-	gdb_test "p (int) gnu_ifunc (3)" " = 4"
     } else {
-	gdb_test "p gnu_ifunc()" "Too few arguments in function call\\."
+	# Do the test that is supposed to succeed first, see above.
 	gdb_test "p gnu_ifunc (3)" " = 4"
+
+	gdb_test "p gnu_ifunc()" "Too few arguments in function call\\."
     }
 
     # Test that the resolver received its argument.
-- 
2.34.1

Re: [PATCH] [gdb] Fix segfault during inferior call to ifunc

[Tom Tromey]

>>>>> tdevries via Gdb-patches <gdb-patches@sourceware.org> writes:

> From 0832e20dde6ebe71a1245935f4a41308368b6094 Mon Sep 17 00:00:00 2001
> From: Andrew Burgess <aburgess@redhat.com>
> Date: Mon, 26 Dec 2022 09:03:19 +0100
> Subject: [PATCH] [gdb] Fix segfault during inferior call to ifunc

Thanks, this looks good to me.

> Note that this series (
> https://inbox.sourceware.org/gdb-patches/20221214033441.499512-1-simon.marchi@polymtl.ca/ )
> should address this problem, but this patch is a simpler fix which is easy to
> backport.

Indeed, and I think it should be put on the GDB 13 branch.

Tom