From: Tom de Vries <tdevries@suse.de>
To: gdb-patches@sourceware.org
Cc: Andrew Burgess <aburgess@redhat.com>
Subject: [PING][RFC][gdb/testsuite] Fix inferior pid in gdb.server/server-kill.exp
Date: Tue, 25 Oct 2022 14:20:00 +0200 [thread overview]
Message-ID: <e0ebd3d9-5cb7-d735-9948-d03acc9cce74@suse.de> (raw)
In-Reply-To: <20221011140150.GA15277@delia.home>
On 10/11/22 16:01, Tom de Vries wrote:
> Hi,
>
> Consider this patch in gdb.server/server-kill.exp:
> ...
> proc kill_server {} {
> global server_pid
>
> + remote_exec target "pstree -a -p $server_pid"
> remote_exec target "kill -9 $server_pid"
> }
> ...
>
> We have for kill_pid_of=server, as expected, the gdbserver killed:
> ...
> Executing on target: pstree -a -p 6969 (timeout = 300)
> builtin_spawn -ignore SIGHUP pstree -a -p 6969^M
> gdbserver,6969^M
> `-server-kill,6976^M
> Executing on target: kill -9 6969 (timeout = 300)
> builtin_spawn -ignore SIGHUP kill -9 6969^M
> ...
>
> But for kill_pid_of=inferior, we also have the gdbserver killed:
> ...
> Executing on target: pstree -a -p 6805 (timeout = 300)
> builtin_spawn -ignore SIGHUP pstree -a -p 6805^M
> gdbserver,6805^M
> `-server-kill,6812^M
> Executing on target: kill -9 6805 (timeout = 300)
> builtin_spawn -ignore SIGHUP kill -9 6805^M
> ...
>
> The proc prepare contains the pid extraction code:
> ...
> if { $::kill_pid_of == "inferior" } {
> # Get the pid of GDBServer.
> set test "p server_pid"
> set server_pid 0
> gdb_test_multiple $test $test {
> -re " = ($decimal)\r\n$gdb_prompt $" {
> set server_pid $expect_out(1,string)
> pass $test
> }
> }
> } else {
> set server_pid [exp_pid -i $::server_spawn_id]
> }
> ...
>
> The bit for $::kill_pid_of == "inferior" looks like the correct code to
> extract the gdbserver pid (it prints the parent PID of the inferior).
>
> The bit for $::kill_pid_of == "server" does work for local target, but for
> remote target using say target board remote-gdbserver-on-localhost, we have:
> ...
> Executing on target: pstree -a -p 10354 (timeout = 300)
> builtin_spawn [open ...]^M
> ssh,10354 -l vries localhost/gdbserver --once localhost:2350 s
> XYZ0ZYX
> Executing on target: kill -9 10354 (timeout = 300)
> builtin_spawn [open ...]^M
> XYZ0ZYX
> ...
> In other words, we're not killing the gdbserver, but the ssh session (a
> problem which was already fixed once in commit f90183d7e31 ("Get GDBserver
> pid on remote target")).
>
> Fix this by:
> - using the $::kill_pid_of == "inferior" bit for the gdbserver pid, and
> - using the inferior command for the the inferior pid.
>
> This introduces the following 4 FAILs:
> ...
> FAIL: gdb.server/server-kill.exp: kill_pid_of=inferior: test_tstatus: tstatus
> FAIL: gdb.server/server-kill.exp: kill_pid_of=inferior: test_unwind_nosyms: bt
> FAIL: gdb.server/server-kill.exp: kill_pid_of=inferior: test_unwind_syms: bt
> FAIL: gdb.server/server-kill.exp: kill_pid_of=inferior: test_stepi: stepi
> ...
>
> The first in more detail:
> ...
> (gdb) status-packet on
> tstatus^M
> No trace has been run on the target.^M
> Collected 0 trace frames.^M
> Trace buffer has 5242880 bytes of 5242880 bytes free (0% full).^M
> Trace will stop if GDB disconnects.^M
> Not looking at any trace frame.^M
> (gdb) FAIL: gdb.server/server-kill.exp: kill_pid_of=inferior: test_tstatus: tstatus
> ...
>
> The test-case contains the comment:
> ...
> # When KILL_PID_OF is set to 'inferior' then the pid we kill is that
> # of the inferior running under gdbserver, when this process dies
> # gdbserver itself will exit.
> ...
> but that doesn't seem to be happening.
>
> I've added a sleep 60 after the kill to rule out any timing issues, and indeed
> after 60 seconds the gdbserver is still running.
>
> I don't know gdbserver well enough to decide whether the test-case assumption
> is wrong and we need to fix the test-case, or there's a gdbserver problem
> which is known or for which we can file a PR.
>
> Tested on x86_64-linux.
>
> Any comments?
>
Ping.
Thanks,
- Tom
> [gdb/testsuite] Fix inferior pid in gdb.server/server-kill.exp
>
> ---
> gdb/testsuite/gdb.server/server-kill.exp | 12 +++++++++---
> 1 file changed, 9 insertions(+), 3 deletions(-)
>
> diff --git a/gdb/testsuite/gdb.server/server-kill.exp b/gdb/testsuite/gdb.server/server-kill.exp
> index 93daf482907..05f4f21bc0c 100644
> --- a/gdb/testsuite/gdb.server/server-kill.exp
> +++ b/gdb/testsuite/gdb.server/server-kill.exp
> @@ -75,10 +75,10 @@ proc prepare {} {
> gdb_breakpoint ${srcfile}:[gdb_get_line_number "i = 0;"]
> gdb_continue_to_breakpoint "after server_pid assignment"
>
> - if { $::kill_pid_of == "inferior" } {
> + set server_pid 0
> + if { $::kill_pid_of == "server" } {
> # Get the pid of GDBServer.
> set test "p server_pid"
> - set server_pid 0
> gdb_test_multiple $test $test {
> -re " = ($decimal)\r\n$gdb_prompt $" {
> set server_pid $expect_out(1,string)
> @@ -86,7 +86,13 @@ proc prepare {} {
> }
> }
> } else {
> - set server_pid [exp_pid -i $::server_spawn_id]
> + set test "inferior"
> + gdb_test_multiple $test $test {
> + -re -wrap "Current inferior is 1 \\\[process ($decimal)\\\].*" {
> + set server_pid $expect_out(1,string)
> + pass $test
> + }
> + }
> }
>
> if {$server_pid == 0} {
next prev parent reply other threads:[~2022-10-25 12:20 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-11 14:01 [RFC][gdb/testsuite] " Tom de Vries
2022-10-25 12:20 ` Tom de Vries [this message]
2023-03-10 15:40 ` [PING][RFC][gdb/testsuite] " Tom de Vries
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e0ebd3d9-5cb7-d735-9948-d03acc9cce74@suse.de \
--to=tdevries@suse.de \
--cc=aburgess@redhat.com \
--cc=gdb-patches@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).