Skip to contentSkip to conversation

btio: Handle EOPNOTSUPP from accept() to prevent busy loop

PATCH Applied Luiz Augusto von Dentz 4 participants 7 messages Jun 9 – Jun 11 Bluetooth drivers

lore
Share & tools
lore

Luiz Augusto von Dentz

author maintainer
#0 lore raw

PATCH v2 1/4 btio: Handle EOPNOTSUPP from accept() to prevent busy loop

From: Luiz Augusto von Dentz <[email protected]>

When accept() returns EOPNOTSUPP on an L2CAP SEQPACKET server socket (e.g. AVCTP browsing channel, PSM 0x1b), the error is permanent and retrying will never succeed. Previously, only EBADFD was treated as fatal, causing server_cb to return TRUE for EOPNOTSUPP. Since the fd remains readable, this creates an infinite busy loop that hangs bluetoothd.

Treat EOPNOTSUPP the same as EBADFD by returning FALSE to remove the GLib IO watch and stop the loop.

---
 btio/btio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
btio/btio.c +1 -1
index 39d4411f790b..4c69d60350f8 100644
@@ -274,7 +274,7 @@ static gboolean server_cb(GIOChannel *io, GIOCondition cond,
 
 	cli_sock = accept(srv_sock, NULL, NULL);
 	if (cli_sock < 0) {
-		if (errno == EBADFD)
+		if (errno == EBADFD || errno == EOPNOTSUPP)
 			return FALSE;
 		return TRUE;
 	}
Signature · 1 line
2.54.0
author maintainer
#1 lore raw

PATCH v2 2/4 profile: Check if bearer is enabled on registration

From: Luiz Augusto von Dentz <[email protected]>

btd_profile_register now verifies that the profile's bearer type is compatible with btd_opts.mode before registering. If the required bearer is not enabled (e.g. LE-only profile when mode is BR/EDR, or BR/EDR-only profile when mode is LE), registration is rejected with -ENOTSUP.

---
 src/profile.c | 9 +++++++++
 1 file changed, 9 insertions(+)
src/profile.c +9 -0
index 65df0f7a0969..97fffe9b4d5c 100644
@@ -36,6 +36,7 @@
 #include "dbus-common.h"
 #include "sdp-client.h"
 #include "sdp-xml.h"
+#include "btd.h"
 #include "adapter.h"
 #include "device.h"
 #include "profile.h"
@@ -802,6 +803,14 @@ struct btd_profile *btd_profile_find_remote_uuid(const char *uuid)
 
 int btd_profile_register(struct btd_profile *profile)
 {
+	if ((profile->bearer == BTD_PROFILE_BEARER_LE &&
+				btd_opts.mode == BT_MODE_BREDR) ||
+			(profile->bearer == BTD_PROFILE_BEARER_BREDR &&
+				btd_opts.mode == BT_MODE_LE)) {
+		DBG("Bearer not enabled");
+		return -ENOTSUP;
+	}
+
 	if (profile->experimental && !(g_dbus_get_flags() &
 					G_DBUS_FLAG_ENABLE_EXPERIMENTAL)) {
 		DBG("D-Bus experimental not enabled");
Signature · 1 line
2.54.0
author maintainer
#2 lore raw

PATCH v2 3/4 plugins: Check btd_profile_register return value

In reply to Luiz Augusto von Dentz

From: Luiz Augusto von Dentz <[email protected]>

Ensure all plugin init functions check the return value of btd_profile_register. If registration fails (e.g. bearer not enabled), the plugin init propagates the error instead of continuing with an unregistered profile.

---
 profiles/audio/a2dp.c      | 14 ++++++++++++--
 profiles/audio/avrcp.c     | 13 +++++++++++--
 profiles/audio/hfp-hf.c    |  4 +---
 profiles/audio/micp.c      |  7 ++++++-
 profiles/input/manager.c   |  4 +---
 profiles/network/manager.c |  5 ++++-
 src/gatt-database.c        |  4 +++-
 7 files changed, 38 insertions(+), 13 deletions(-)
profiles/audio/a2dp.c +12 -2
index c7e0fc75c09e..a5e002784c02 100644
@@ -3798,9 +3798,19 @@ static struct btd_adapter_driver media_driver = {
 
 static int a2dp_init(void)
 {
+	int err;
+
 	btd_register_adapter_driver(&media_driver);
-	btd_profile_register(&a2dp_source_profile);
-	btd_profile_register(&a2dp_sink_profile);
+
+	err = btd_profile_register(&a2dp_source_profile);
+	if (err)
+		return err;
+
+	err = btd_profile_register(&a2dp_sink_profile);
+	if (err) {
+		btd_profile_unregister(&a2dp_source_profile);
+		return err;
+	}
 
 	return 0;
 }
profiles/audio/avrcp.c +11 -2
index b6823753fe68..f63acd47091a 100644
@@ -4987,8 +4987,17 @@ static struct btd_profile avrcp_controller_profile = {
 
 static int avrcp_init(void)
 {
-	btd_profile_register(&avrcp_controller_profile);
-	btd_profile_register(&avrcp_target_profile);
+	int err;
+
+	err = btd_profile_register(&avrcp_controller_profile);
+	if (err)
+		return err;
+
+	err = btd_profile_register(&avrcp_target_profile);
+	if (err) {
+		btd_profile_unregister(&avrcp_controller_profile);
+		return err;
+	}
 
 	populate_default_features();
 
profiles/audio/hfp-hf.c +1 -3
index c91b16426898..8de2d7a62d68 100644
@@ -507,9 +507,7 @@ static struct btd_profile hfp_hf_profile = {
 
 static int hfp_init(void)
 {
-	btd_profile_register(&hfp_hf_profile);
-
-	return 0;
+	return btd_profile_register(&hfp_hf_profile);
 }
 
 static void hfp_exit(void)
profiles/audio/micp.c +6 -1
index 475f32daf75c..3d39ef5e147f 100644
@@ -318,12 +318,17 @@ static unsigned int micp_id;
 
 static int micp_init(void)
 {
+	int err;
+
 	if (!(g_dbus_get_flags() & G_DBUS_FLAG_ENABLE_EXPERIMENTAL)) {
 		DBG("D-Bus experimental not enabled");
 		return -ENOTSUP;
 	}
 
-	btd_profile_register(&micp_profile);
+	err = btd_profile_register(&micp_profile);
+	if (err)
+		return err;
+
 	micp_id = bt_micp_register(micp_attached, micp_detached, NULL);
 
 	return 0;
profiles/input/manager.c +1 -3
index 0fcd6728c2fc..1fd82d82f500 100644
@@ -118,12 +118,10 @@ static int input_init(void)
 
 	}
 
-	btd_profile_register(&input_profile);
-
 	if (config)
 		g_key_file_free(config);
 
-	return 0;
+	return btd_profile_register(&input_profile);
 }
 
 static void input_exit(void)
profiles/network/manager.c +4 -1
index 693547d45fbc..a5f28a99ebfd 100644
@@ -180,7 +180,10 @@ static int network_init(void)
 	if (server_init(conf_security) < 0)
 		return -1;
 
-	btd_profile_register(&panu_profile);
+	err = btd_profile_register(&panu_profile);
+	if (err)
+		return err;
+
 	btd_profile_register(&gn_profile);
 	btd_profile_register(&nap_profile);
 
src/gatt-database.c +3 -1
index 680a52952b16..30e25b6f41ca 100644
@@ -3624,7 +3624,9 @@ static void add_profile(void *data, void *user_data)
 {
 	struct btd_adapter *adapter = user_data;
 
-	btd_profile_register(data);
+	if (btd_profile_register(data))
+		return;
+
 	adapter_add_profile(adapter, data);
 }
 
Signature · 1 line
2.54.0
author maintainer
#3 lore raw

PATCH v2 4/4 bearer: Check btd_opts.mode on btd_bearer_new

In reply to Luiz Augusto von Dentz

From: Luiz Augusto von Dentz <[email protected]>

Only create the bearer interface if the corresponding transport is enabled. Return NULL if BREDR bearer is requested in LE-only mode or LE bearer in BREDR-only mode, so the D-Bus interface is never registered for unsupported bearers.

---
 src/bearer.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
src/bearer.c +14 -0
index 02267c711431..ac3a22030770 100644
@@ -34,6 +34,7 @@
 
 #include "log.h"
 #include "error.h"
+#include "btd.h"
 #include "adapter.h"
 #include "device.h"
 #include "profile.h"
@@ -278,6 +279,19 @@ struct btd_bearer *btd_bearer_new(struct btd_device *device, uint8_t type)
 {
 	struct btd_bearer *bearer;
 
+	switch (btd_opts.mode) {
+	case BT_MODE_LE:
+		if (type == BDADDR_BREDR)
+			return NULL;
+		break;
+	case BT_MODE_BREDR:
+		if (type != BDADDR_BREDR)
+			return NULL;
+		break;
+	case BT_MODE_DUAL:
+		break;
+	}
+
 	bearer = new0(struct btd_bearer, 1);
 	bearer->device = device;
 	bearer->type = type;
Signature · 1 line
2.54.0
#4 lore raw
In reply to Luiz Augusto von Dentz

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=1108786

---Test result---

Test Summary:

CheckPatch                    PASS      1.22 seconds
GitLint                       PASS      0.78 seconds
BuildEll                      PASS      20.04 seconds
BluezMake                     PASS      656.72 seconds
MakeCheck                     PASS      18.90 seconds
MakeDistcheck                 PASS      249.96 seconds
CheckValgrind                 PASS      295.94 seconds
CheckSmatch                   PASS      350.89 seconds
bluezmakeextell               PASS      181.16 seconds
IncrementalBuild              PASS      673.06 seconds
ScanBuild                     WARNING   1036.74 seconds

Details
##############################
Test: ScanBuild - WARNING
Desc: Run Scan Build
Output:
1 warning generated.
tools/btgatt-client.c:1822:2: warning: Value stored to 'argv' is never read tools/check-selftest.c:42:3: warning: Value stored to 'ptr' is never read

        argv += optind;
                ptr = fgets(result, sizeof(result), fp);
        ^       ~~~~~~
                ^     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1 warning generated.

https://github.com/bluez/bluez/pull/2206

---
Regards,
Linux Bluetooth

#5 lore raw
In reply to Luiz Augusto von Dentz
This series was applied to bluetooth/bluez.git (master) by Luiz Augusto von Dentz <[email protected]>:

Hello:

This series was applied to bluetooth/bluez.git (master) by Luiz Augusto von Dentz <[email protected]>:

Quoted 10 lines · On Tue, 9 Jun 2026 14:53:10 -0400 you wrote:

From: Luiz Augusto von Dentz <[email protected]>

When accept() returns EOPNOTSUPP on an L2CAP SEQPACKET server socket (e.g. AVCTP browsing channel, PSM 0x1b), the error is permanent and retrying will never succeed. Previously, only EBADFD was treated as fatal, causing server_cb to return TRUE for EOPNOTSUPP. Since the fd remains readable, this creates an infinite busy loop that hangs bluetoothd.

[...]

Here is the summary with links:
- [BlueZ,v2,1/4] btio: Handle EOPNOTSUPP from accept() to prevent busy loop

    https://git.kernel.org/pub/scm/bluetooth/bluez.git/?id=8c580d30d265
  - [BlueZ,v2,2/4] profile: Check if bearer is enabled on registration
    https://git.kernel.org/pub/scm/bluetooth/bluez.git/?id=abf0911e25ef
  - [BlueZ,v2,3/4] plugins: Check btd_profile_register return value
    https://git.kernel.org/pub/scm/bluetooth/bluez.git/?id=7b5895d80635
  - [BlueZ,v2,4/4] bearer: Check btd_opts.mode on btd_bearer_new
    https://git.kernel.org/pub/scm/bluetooth/bluez.git/?id=8c821a150b8f

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
#6 lore raw
In reply to Luiz Augusto von Dentz
Quoted 11 lines · Luiz Augusto von Dentz

From: Luiz Augusto von Dentz <[email protected]>

When accept() returns EOPNOTSUPP on an L2CAP SEQPACKET server socket (e.g. AVCTP browsing channel, PSM 0x1b), the error is permanent and retrying will never succeed. Previously, only EBADFD was treated as fatal, causing server_cb to return TRUE for EOPNOTSUPP. Since the fd remains readable, this creates an infinite busy loop that hangs bluetoothd.

Treat EOPNOTSUPP the same as EBADFD by returning FALSE to remove the GLib IO watch and stop the loop.

This was the infinite loop that was happening on startup, right?

Were there specific circumstances for it to hang like this? I was wondering whether this was something I could reproduce in the integration tests I put together for the meson port.

Quoted 18 lines

---
 btio/btio.c | 2 +-

 1 file changed, 1 insertion(+), 1 deletion(-)
btio/btio.c +0 -0
index 39d4411f790b..4c69d60350f8 100644
@@ -274,7 +274,7 @@ static gboolean server_cb(GIOChannel *io,
 GIOCondition cond,


  cli_sock = accept(srv_sock, NULL, NULL);
  if (cli_sock < 0) {

btio/btio.c +1 -1
-		if (errno == EBADFD)
+		if (errno == EBADFD || errno == EOPNOTSUPP)

  return FALSE;
  return TRUE;
  }