[PATCH BlueZ 0/1] Free ALSA seq resources in midi_device_remove()

[PATCH BlueZ 0/1] Free ALSA seq resources in midi_device_remove()

[Jacek Konieczny]

I am having problem connecting via Bluetooth LE MID to my VOX Adio Air GT
guitar amplifier. I was not able to debug the connection problem yet,
but there was a bigger problem on hand: due to constant
connecting/disconnecting ALSA seq resources were leaking until
bluetoothd hung and ALSA backend crashed.

I have described a bit more in the github issue: https://github.com/bluez/bluez/issues/225

I was able to fix the leak with the following patch. I am not sure if
that is the right way to do this, though. As the relation between
midi_disconnect() and midi_device_remove() functions is not clear to me,
I have just copied some of the cleanup code from one to the other.


Jacek Konieczny (1):
  Free ALSA seq resources in midi_device_remove()

 profiles/midi/midi.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

-- 
2.25.1

[PATCH BlueZ 1/1] Free ALSA seq resources in midi_device_remove()

[Jacek Konieczny]

Sometimes, on connection error, this gets called and frees the device
when midi_disconnect() was not called which would leave ALSA sequencer
objects leaking.
---
 profiles/midi/midi.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/profiles/midi/midi.c b/profiles/midi/midi.c
index 737d1b5f6..f644d47e9 100644
--- a/profiles/midi/midi.c
+++ b/profiles/midi/midi.c
@@ -255,6 +255,16 @@ static void midi_device_remove(struct btd_service *service)
 		return;
 	}
 
+	if (midi->seq_handle) {
+		midi_read_free(&midi->midi_in);
+		midi_write_free(&midi->midi_out);
+		io_destroy(midi->io);
+		snd_seq_delete_simple_port(midi->seq_handle, midi->seq_port_id);
+		midi->seq_port_id = 0;
+		snd_seq_close(midi->seq_handle);
+		midi->seq_handle = NULL;
+	}
+
 	btd_device_unref(midi->dev);
 	g_free(midi);
 }
-- 
2.25.1

RE: Free ALSA seq resources in midi_device_remove()

[bluez.test.bot]

[-- Attachment #1: Type: text/plain, Size: 884 bytes --]

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list.
This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=573565

---Test result---

Test Summary:
CheckPatch                    PASS      1.43 seconds
GitLint                       PASS      0.89 seconds
Prep - Setup ELL              PASS      46.37 seconds
Build - Prep                  PASS      0.50 seconds
Build - Configure             PASS      8.57 seconds
Build - Make                  PASS      202.34 seconds
Make Check                    PASS      9.49 seconds
Make Distcheck                PASS      242.98 seconds
Build w/ext ELL - Configure   PASS      8.72 seconds
Build w/ext ELL - Make        PASS      191.05 seconds



---
Regards,
Linux Bluetooth

Re: Free ALSA seq resources in midi_device_remove()

[Luiz Augusto von Dentz]

Hi Jacek,

On Mon, Nov 1, 2021 at 7:36 AM <bluez.test.bot@gmail.com> wrote:
>
> This is automated email and please do not reply to this email!
>
> Dear submitter,
>
> Thank you for submitting the patches to the linux bluetooth mailing list.
> This is a CI test results with your patch series:
> PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=573565
>
> ---Test result---
>
> Test Summary:
> CheckPatch                    PASS      1.43 seconds
> GitLint                       PASS      0.89 seconds
> Prep - Setup ELL              PASS      46.37 seconds
> Build - Prep                  PASS      0.50 seconds
> Build - Configure             PASS      8.57 seconds
> Build - Make                  PASS      202.34 seconds
> Make Check                    PASS      9.49 seconds
> Make Distcheck                PASS      242.98 seconds
> Build w/ext ELL - Configure   PASS      8.72 seconds
> Build w/ext ELL - Make        PASS      191.05 seconds
>
>
>
> ---
> Regards,
> Linux Bluetooth

Applied, thanks.

-- 
Luiz Augusto von Dentz