Re: [PATCH 1/3] Fix "run" failure with GDBserver

[Pedro Alves <pedro@palves.net>]

On 2024-02-13 15:19, Lancelot SIX wrote:
> Hi Pedro,

Hi!

> 
> On Mon, Feb 12, 2024 at 08:01:51PM +0000, Pedro Alves wrote:

>> ---
>>  gdb/testsuite/gdb.base/run-fail-twice.exp | 67 +++++++++++++++++++++++
>>  gdbserver/server.cc                       | 10 +++-
>>  2 files changed, 76 insertions(+), 1 deletion(-)
>>  create mode 100644 gdb/testsuite/gdb.base/run-fail-twice.exp
>>
> 
> Looks like you forgot the gdb/testsuite/gdb.base/run-fail-twice.c file?

Indeed...  Added now.

> 
>> diff --git a/gdb/testsuite/gdb.base/run-fail-twice.exp b/gdb/testsuite/gdb.base/run-fail-twice.exp

>> +
>> +# Test doing a "run" that fails, and then another "run".
>> +
>> +# The purpose of this testcase is to test the "run" command.  If we
>> +# cannot use it, then there is no point in running this testcase.
>> +require !use_gdb_stub

I've switched this to:

 require target_can_use_run_cmd

>> +
>> +standard_testfile
>> +
>> +if {[build_executable "failed to build" $testfile $srcfile {debug}]} {
>> +    return -1
>> +}
>> +
>> +proc test_run {testname} {
>> +    gdb_run_cmd

I switched to calling "run" directly, as that's what we're testing anyhow.

>> +    gdb_test_multiple "" $testname {
>> +	-re -wrap "During startup program exited with code 126\\." {
>> +	    # What we get on GNU/Linux.
>> +	    pass $gdb_test_name
>> +	}
>> +	-re -wrap "Error creating process.*" {
>> +	    # What we get on Windows.
>> +	    pass $gdb_test_name
>> +	}
>> +	-re -wrap "Running .* on the remote target failed" {
>> +	    # What we get with older GDBserver and other remote
>> +	    # targets.
>> +	    pass $gdb_test_name
>> +	}
>> +    }
>> +}
>> +
>> +proc_with_prefix test {} {
>> +    global gdb_prompt binfile
>> +
>> +    clean_restart $binfile
>> +
>> +    gdb_test_no_output "set confirm off"
>> +
>> +    gdb_remote_download host $binfile $binfile.nox
>> +    remote_exec target "chmod \"a-x\" $binfile.nox"
>> +    gdb_test "exec-file $binfile.nox" \
> 
> Couldn't you use gdb_test_no_output and remove the 2nd argument?

I can!  And I did.

> 
>> +	"" \
>> +	"exec-file \$binfile.nox"
>> +    gdb_test "set remote exec-file $binfile.nox" \
>> +	"" \
> 
> Same here.
> 

Ditto.

Here's the updated patch.

---- 8< ----
From 04b71816555898fa804a76aa0412b1bad1dc9692 Mon Sep 17 00:00:00 2001
From: Pedro Alves <pedro@palves.net>
Subject: [PATCH] Fix "run" failure with GDBserver

If starting the inferior process with "run" (vRun packet) fails,
GDBserver throws an error that escapes all the way to the top level.
When an error escapes all the way like that, GDBserver interprets it
as a disconnection, and either goes back to waiting for a new GDB
connection, or exits, if --once was specified.

E.g., with the testcase program added by this commit, we see:

On GDB side:

 ...
 (gdb) tar extended-remote :999
 ...
 Remote debugging using :9999
 (gdb) r
 Starting program:
 Running ".../gdb.base/run-fail-twice/run-fail-twice.nox" on the remote target failed
 (gdb)

On GDBserver side:

 $ gdbserver --once --multi :9999
 Remote debugging from host 127.0.0.1, port 34344
 bash: line 1: .../gdb.base/run-fail-twice/run-fail-twice.nox: Permission denied
 bash: line 1: exec: .../gdb.base/run-fail-twice/run-fail-twice.nox: cannot execute: Permission denied
 gdbserver: During startup program exited with code 126.
 $   # gdbserver exited

This is wrong, as we've connected with extended-remote/--multi.
GDBserver should just report an error to vCont, and continue connected
to GDB, waiting for other commands.

This commit fixes GDBserver by catching the error locally in
handle_v_run.

Change-Id: Ib386f267522603f554b52a885b15229c9639e870
---
 gdb/testsuite/gdb.base/run-fail-twice.c   | 20 +++++++
 gdb/testsuite/gdb.base/run-fail-twice.exp | 63 +++++++++++++++++++++++
 gdbserver/server.cc                       | 10 +++-
 3 files changed, 92 insertions(+), 1 deletion(-)
 create mode 100644 gdb/testsuite/gdb.base/run-fail-twice.c
 create mode 100644 gdb/testsuite/gdb.base/run-fail-twice.exp

diff --git a/gdb/testsuite/gdb.base/run-fail-twice.c b/gdb/testsuite/gdb.base/run-fail-twice.c
new file mode 100644
index 00000000000..fddf841eb3e
--- /dev/null
+++ b/gdb/testsuite/gdb.base/run-fail-twice.c
@@ -0,0 +1,20 @@
+/* Copyright 2024 Free Software Foundation, Inc.
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+int
+main (int argc, char **argv)
+{
+  return 0;
+}
diff --git a/gdb/testsuite/gdb.base/run-fail-twice.exp b/gdb/testsuite/gdb.base/run-fail-twice.exp
new file mode 100644
index 00000000000..676fc486fbf
--- /dev/null
+++ b/gdb/testsuite/gdb.base/run-fail-twice.exp
@@ -0,0 +1,63 @@
+# Copyright 2024 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# Test doing a "run" that fails, and then another "run".
+
+require target_can_use_run_cmd
+
+standard_testfile
+
+if {[build_executable "failed to build" $testfile $srcfile {debug}]} {
+    return -1
+}
+
+proc test_run {testname} {
+    gdb_test_multiple "run" $testname {
+	-re -wrap "During startup program exited with code 126\\." {
+	    # What we get on GNU/Linux.
+	    pass $gdb_test_name
+	}
+	-re -wrap "Error creating process.*" {
+	    # What we get on Windows.
+	    pass $gdb_test_name
+	}
+	-re -wrap "Running .* on the remote target failed" {
+	    # What we get with remote targets.
+	    pass $gdb_test_name
+	}
+    }
+}
+
+proc_with_prefix test {} {
+    global gdb_prompt binfile
+
+    clean_restart $binfile
+
+    gdb_test_no_output "set confirm off"
+
+    gdb_remote_download host $binfile $binfile.nox
+    remote_exec target "chmod \"a-x\" $binfile.nox"
+    gdb_test_no_output \
+	"exec-file $binfile.nox" \
+	"exec-file \$binfile.nox"
+    gdb_test_no_output \
+	"set remote exec-file $binfile.nox" \
+	"set remote exec-file \$binfile.nox"
+
+    test_run "bad run 1"
+    test_run "bad run 2"
+}
+
+test
diff --git a/gdbserver/server.cc b/gdbserver/server.cc
index 74c7763d777..14a19bc1882 100644
--- a/gdbserver/server.cc
+++ b/gdbserver/server.cc
@@ -3428,7 +3428,15 @@ handle_v_run (char *own_buf)
   free_vector_argv (program_args);
   program_args = new_argv;
 
-  target_create_inferior (program_path.get (), program_args);
+  try
+    {
+      target_create_inferior (program_path.get (), program_args);
+    }
+  catch (const gdb_exception_error &exception)
+    {
+      sprintf (own_buf, "E.%s", exception.what ());
+      return;
+    }
 
   if (cs.last_status.kind () == TARGET_WAITKIND_STOPPED)
     {

base-commit: a16034bf6417dc2259fef43fd5bcc2dd1dac562f
-- 
2.43.0