* [PATCH BlueZ] device: Fix device removal on le-connection-abort-by-local
@ 2025-08-21 20:45 J. Nick Koston
2025-08-21 22:06 ` [BlueZ] " bluez.test.bot
0 siblings, 1 reply; 3+ messages in thread
From: J. Nick Koston @ 2025-08-21 20:45 UTC (permalink / raw)
To: linux-bluetooth; +Cc: luiz.dentz, J. Nick Koston, J . Nick Koston
When a LE connection fails with le-connection-abort-by-local, temporary
devices are incorrectly removed from D-Bus, preventing clients from
retrying the connection.
The issue occurs when:
1. A connection attempt fails with ECONNABORTED (including the
le-connection-abort-by-local case)
2. The device is added to the connect list for automatic retry
3. The temporary timer (30s by default) is not restarted
4. If the timer expires before the retry completes, the device is
removed from D-Bus
5. Subsequent retry attempts fail with "device not found"
This commonly happens with devices that abort connections during
pairing or when incomplete GATT discovery triggers an abort.
The fix restarts the temporary timer when ECONNABORTED occurs, ensuring
the device remains available on D-Bus for the retry attempt. This
matches the behavior when new connection attempts are initiated, where
the timer is restarted to allow time for connection/pairing to complete.
Reproducer logs show:
- 14:19:37.374 - le-connection-abort-by-local, retry queued
- 14:19:56.343 - Device removed from D-Bus (~19s later)
- 14:19:56.606 - Retry fails: device not found on D-Bus
With this fix, the device remains available for the retry, allowing
the connection to succeed on subsequent attempts.
Fixes: https://github.com/bluez/bluez/issues/1489
Signed-off-by: J. Nick Koston <nick@koston.org>
Signed-off-by: J. Nick Koston <nick@home-assistant.io>
---
src/device.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/device.c b/src/device.c
index 0179c3dab..28803af38 100644
--- a/src/device.c
+++ b/src/device.c
@@ -6271,8 +6271,12 @@ static void att_connect_cb(GIOChannel *io, GError *gerr, gpointer user_data)
if (gerr) {
DBG("%s", gerr->message);
- if (g_error_matches(gerr, BT_IO_ERROR, ECONNABORTED))
+ if (g_error_matches(gerr, BT_IO_ERROR, ECONNABORTED)) {
+ /* Restart temporary timer to prevent device removal */
+ if (device->temporary)
+ set_temporary_timer(device, btd_opts.tmpto);
goto done;
+ }
if (device_get_auto_connect(device)) {
DBG("Enabling automatic connections");
--
2.39.3 (Apple Git-145)
^ permalink raw reply related [flat|nested] 3+ messages in thread* [PATCH BlueZ] device: Fix device removal on le-connection-abort-by-local
@ 2025-08-21 20:17 J. Nick Koston
0 siblings, 0 replies; 3+ messages in thread
From: J. Nick Koston @ 2025-08-21 20:17 UTC (permalink / raw)
To: linux-bluetooth; +Cc: luiz.dentz
From fe7e071643a8de84bdc4079551cc0065890c0cbb Mon Sep 17 00:00:00 2001
From: "J. Nick Koston" <nick@koston.org>
Date: Thu, 21 Aug 2025 15:12:39 -0500
Subject: [PATCH] device: Fix device removal on le-connection-abort-by-local
When a LE connection fails with le-connection-abort-by-local, temporary
devices are incorrectly removed from D-Bus, preventing clients from
retrying the connection.
The issue occurs when:
1. A connection attempt fails with ECONNABORTED (including the
le-connection-abort-by-local case)
2. The device is added to the connect list for automatic retry
3. The temporary timer (30s by default) is not restarted
4. If the timer expires before the retry completes, the device is
removed from D-Bus
5. Subsequent retry attempts fail with "device not found"
This commonly happens with devices that abort connections during
pairing or when incomplete GATT discovery triggers an abort.
The fix restarts the temporary timer when ECONNABORTED occurs, ensuring
the device remains available on D-Bus for the retry attempt. This
matches the behavior when new connection attempts are initiated, where
the timer is restarted to allow time for connection/pairing to complete.
Reproducer logs show:
- 14:19:37.374 - le-connection-abort-by-local, retry queued
- 14:19:56.343 - Device removed from D-Bus (~19s later)
- 14:19:56.606 - Retry fails: device not found on D-Bus
With this fix, the device remains available for the retry, allowing
the connection to succeed on subsequent attempts.
Fixes: https://github.com/bluez/bluez/issues/1489
Signed-off-by: J. Nick Koston <nick@koston.org>
---
src/device.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/device.c b/src/device.c
index 0179c3dab..28803af38 100644
--- a/src/device.c
+++ b/src/device.c
@@ -6271,8 +6271,12 @@ static void att_connect_cb(GIOChannel *io, GError *gerr, gpointer user_data)
if (gerr) {
DBG("%s", gerr->message);
- if (g_error_matches(gerr, BT_IO_ERROR, ECONNABORTED))
+ if (g_error_matches(gerr, BT_IO_ERROR, ECONNABORTED)) {
+ /* Restart temporary timer to prevent device removal */
+ if (device->temporary)
+ set_temporary_timer(device, btd_opts.tmpto);
goto done;
+ }
if (device_get_auto_connect(device)) {
DBG("Enabling automatic connections");
--
2.39.3 (Apple Git-145)
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2025-08-21 22:06 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-08-21 20:45 [PATCH BlueZ] device: Fix device removal on le-connection-abort-by-local J. Nick Koston
2025-08-21 22:06 ` [BlueZ] " bluez.test.bot
-- strict thread matches above, loose matches on Subject: below --
2025-08-21 20:17 [PATCH BlueZ] " J. Nick Koston
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 NNTP newsgroup(s).