Skip to contentSkip to conversation

gdbstub: allow killing QEMU via vKill command

PATCH Superseded Max Filippov 3 participants 4 messages Jan 29, 2019 – Jan 30, 2019

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 gdbstub: allow killing QEMU via vKill command

With multiprocess extensions gdb uses 'vKill' packet instead of 'k' to kill the inferior. Handle 'vKill' the same way 'k' was handled in the presence of single process.

Fixes 7cf48f6752e5 ("gdbstub: add multiprocess support to

(f|s)ThreadInfo and ThreadExtraInfo")

---
 gdbstub.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
gdbstub.c +22 -0
index bfc7afb50968..1ef31240c055 100644
@@ -1383,6 +1383,28 @@ static int gdb_handle_packet(GDBState *s, const char *line_buf)
 
             put_packet(s, buf);
             break;
+        } else if (strncmp(p, "Kill;", 5) == 0) {
+            unsigned long pid;
+
+            p += 5;
+
+            if (qemu_strtoul(p, &p, 16, &pid)) {
+                put_packet(s, "E22");
+                break;
+            }
+            process = gdb_get_process(s, pid);
+
+            if (process == NULL) {
+                put_packet(s, "E22");
+                break;
+            }
+            if (s->process_num <= 1) {
+                /* Kill the target */
+                error_report("QEMU: Terminated via GDBstub");
+                exit(0);
+            }
+            /* TODO: handle multiprocess case */
+            goto unknown_command;
         } else {
             goto unknown_command;
         }
Signature · 1 line
2.11.0
#1 lore raw

Hi,

Le 1/29/19 à 9:32 PM, Max Filippov a écrit :

Quoted 30 lines

With multiprocess extensions gdb uses 'vKill' packet instead of 'k' to kill the inferior. Handle 'vKill' the same way 'k' was handled in the presence of single process.

Fixes 7cf48f6752e5 ("gdbstub: add multiprocess support to

(f|s)ThreadInfo and ThreadExtraInfo")

---
  gdbstub.c | 22 ++++++++++++++++++++++
  1 file changed, 22 insertions(+)
gdbstub.c +9 -0
index bfc7afb50968..1ef31240c055 100644
@@ -1383,6 +1383,28 @@ static int gdb_handle_packet(GDBState *s, const char *line_buf)
 
              put_packet(s, buf);
              break;
+        } else if (strncmp(p, "Kill;", 5) == 0) {
+            unsigned long pid;
+
+            p += 5;
+
+            if (qemu_strtoul(p, &p, 16, &pid)) {
+                put_packet(s, "E22");
+                break;
+            }

gdbstub.c static int gdb_handle_packet(GDBState *s, const char *line_buf) · line 1404

Quoted 10 lines
gdbstub.c +10 -0
+            process = gdb_get_process(s, pid);
+
+            if (process == NULL) {
+                put_packet(s, "E22");
+                break;
+            }
+            if (s->process_num <= 1) {
+                /* Kill the target */
+                error_report("QEMU: Terminated via GDBstub");
+                exit(0);

I suggest:

#ifdef CONFIG_USER_ONLY
        exit(0);
#else
        qemu_system_shutdown_request(SHUTDOWN_CAUSE_GUEST_SHUTDOWN);

#endif

instead of exit(0);

Quoted 7 lines
gdbstub.c +3 -0
+            }
+            /* TODO: handle multiprocess case */
+            goto unknown_command;
          } else {
              goto unknown_command;
          }
#2 lore raw

Hi,

Quoted 9 lines · KONRAD Frederic

Hi,

Le 1/29/19 à 9:32 PM, Max Filippov a écrit :

Quoted 6 lines

With multiprocess extensions gdb uses 'vKill' packet instead of 'k' to kill the inferior. Handle 'vKill' the same way 'k' was handled in the presence of single process.

Fixes 7cf48f6752e5 ("gdbstub: add multiprocess support to

(f|s)ThreadInfo and ThreadExtraInfo")

Thanks for the patch.

Quoted 37 lines
Quoted 24 lines

---
  gdbstub.c | 22 ++++++++++++++++++++++

  1 file changed, 22 insertions(+)
gdbstub.c +9 -0
index bfc7afb50968..1ef31240c055 100644
@@ -1383,6 +1383,28 @@ static int gdb_handle_packet(GDBState *s, const
 char *line_buf)
                 put_packet(s, buf);
               break;
+        } else if (strncmp(p, "Kill;", 5) == 0) {
+            unsigned long pid;
+
+            p += 5;
+
+            if (qemu_strtoul(p, &p, 16, &pid)) {
+                put_packet(s, "E22");
+                break;
+            }

gdbstub.c static int gdb_handle_packet(GDBState *s, const · line 1404

Quoted 10 lines
gdbstub.c +10 -0
+            process = gdb_get_process(s, pid);
+
+            if (process == NULL) {
+                put_packet(s, "E22");
+                break;
+            }
+            if (s->process_num <= 1) {
+                /* Kill the target */
+                error_report("QEMU: Terminated via GDBstub");
+                exit(0);

I suggest:

Quoted 4 lines

#ifdef CONFIG_USER_ONLY
        exit(0);
#else qemu_system_shutdown_request(SHUTDOWN_CAUSE_GUEST_SHUTDOWN);

You are missing a `break;' here.

Regarding the cause, I think it's more an 'host' cause than a 'guest' one.

Quoted 3 lines

#endif

instead of exit(0);

I just tested, this is not ideal in the current shape. Not quitting immediately makes the stub sends a "W00" packet. It's better if we send nothing because GDB interprets that as the connection was closed (which is effectively the case). Otherwise in multiprocess mode, GDB thinks the other processes are still alive (see below).

with exit(0):
(gdb) k
Kill the program being debugged? (y or n) y
Sending packet: $vKill;1#6e...Ack
Remote connection closed

with qemu_system_shutdown_request:
(gdb) k
Kill the program being debugged? (y or n) y
Sending packet: $vKill;1#6e...Ack
Packet received: W00
Packet vKill (kill) is supported
[Inferior 1 (process 1) killed]

So maybe a solution (for another patch) would be to send a 'W' packet for each attached process (using the multiprocess format). Another is to prevent 'W' packets to be sent at all... they are not required in this case.

Quoted 3 lines

gdbstub.c static int gdb_handle_packet(GDBState *s, const · line 1406

Quoted 3 lines
gdbstub.c +2 -0
+            }
+            /* TODO: handle multiprocess case */

About the multiprocess case, I would vote to adopt the same behaviour (i.e. terminate QEMU), whatever the PID is or how many processes are attached. Otherwise we would have different behaviour for the GDB kill command, whether the simulated board exposes multiple processes or not. Moreover, I think considering just one PID for a kill command does not make much sense in our case.

I just tested, GDB is fine with QEMU closing the connection when having multiple processes attached if we do exit(0) directly (i.e. do not send a 'W00' packet).

Thanks.

Luc

Quoted 5 lines
Quoted 5 lines
gdbstub.c +1 -0
+            goto unknown_command;
           } else {
               goto unknown_command;
           }
#3 lore raw

Le 1/30/19 à 3:31 PM, Luc Michel a écrit :

Quoted 73 lines

Hi,

Quoted 9 lines · KONRAD Frederic

Hi,

Le 1/29/19 à 9:32 PM, Max Filippov a écrit :

Quoted 6 lines

With multiprocess extensions gdb uses 'vKill' packet instead of 'k' to kill the inferior. Handle 'vKill' the same way 'k' was handled in the presence of single process.

Fixes 7cf48f6752e5 ("gdbstub: add multiprocess support to

(f|s)ThreadInfo and ThreadExtraInfo")

Thanks for the patch.

Quoted 37 lines
Quoted 24 lines
---
   gdbstub.c | 22 ++++++++++++++++++++++
   1 file changed, 22 insertions(+)
gdbstub.c +9 -0
index bfc7afb50968..1ef31240c055 100644
@@ -1383,6 +1383,28 @@ static int gdb_handle_packet(GDBState *s, const
 char *line_buf)
                 put_packet(s, buf);
               break;
+        } else if (strncmp(p, "Kill;", 5) == 0) {
+            unsigned long pid;
+
+            p += 5;
+
+            if (qemu_strtoul(p, &p, 16, &pid)) {
+                put_packet(s, "E22");
+                break;
+            }

gdbstub.c static int gdb_handle_packet(GDBState *s, const · line 1404

Quoted 10 lines
gdbstub.c +10 -0
+            process = gdb_get_process(s, pid);
+
+            if (process == NULL) {
+                put_packet(s, "E22");
+                break;
+            }
+            if (s->process_num <= 1) {
+                /* Kill the target */
+                error_report("QEMU: Terminated via GDBstub");
+                exit(0);

I suggest:

Quoted 4 lines

#ifdef CONFIG_USER_ONLY
        exit(0);
#else
        qemu_system_shutdown_request(SHUTDOWN_CAUSE_GUEST_SHUTDOWN);

You are missing a `break;' here.

Regarding the cause, I think it's more an 'host' cause than a 'guest' one.

Quoted 3 lines

#endif

instead of exit(0);

I just tested, this is not ideal in the current shape. Not quitting immediately makes the stub sends a "W00" packet. It's better if we send nothing because GDB interprets that as the connection was closed (which is effectively the case). Otherwise in multiprocess mode, GDB thinks the other processes are still alive (see below).

with exit(0):
(gdb) k
Kill the program being debugged? (y or n) y
Sending packet: $vKill;1#6e...Ack
Remote connection closed

Quoted 7 lines

with qemu_system_shutdown_request:
(gdb) k
Kill the program being debugged? (y or n) y
Sending packet: $vKill;1#6e...Ack
Packet received: W00
Packet vKill (kill) is supported
[Inferior 1 (process 1) killed]

Ok makes sense. I didn't think of that.

Quoted 15 lines

So maybe a solution (for another patch) would be to send a 'W' packet for each attached process (using the multiprocess format). Another is to prevent 'W' packets to be sent at all... they are not required in this case.

Quoted 3 lines

gdbstub.c static int gdb_handle_packet(GDBState *s, const · line 1406

Quoted 3 lines
gdbstub.c +2 -0
+            }
+            /* TODO: handle multiprocess case */

About the multiprocess case, I would vote to adopt the same behaviour (i.e. terminate QEMU), whatever the PID is or how many processes are attached. Otherwise we would have different behaviour for the GDB kill command, whether the simulated board exposes multiple processes or not. Moreover, I think considering just one PID for a kill command does not make much sense in our case.

Quoted 3 lines

I just tested, GDB is fine with QEMU closing the connection when having multiple processes attached if we do exit(0) directly (i.e. do not send a 'W00' packet).

Yes I see the same behavior with Linux but it makes GDB crashes with MinGW at least before this patch and the multiprocess one (eg: v3.0.0).

I'll wait for this patch to land and send some trace if I can still reproduce.

Cheers,
Fred

Quoted 10 lines

Thanks.

Luc

Quoted 5 lines
Quoted 5 lines
gdbstub.c +1 -0
+            goto unknown_command;
           } else {
               goto unknown_command;
           }