Skip to contentSkip to conversation

Bluetooth: L2CAP: Fix not tracking outstanding TX ident

PATCH Applied Luiz Augusto von Dentz 6 participants 11 messages Jan 26 – Mar 17

Subsystems (3)
lore
Share & tools
lore

Luiz Augusto von Dentz

author maintainer
#0 lore raw

PATCH v3 1/1 Bluetooth: L2CAP: Fix not tracking outstanding TX ident

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

This attempts to proper track outstanding request by using struct ida and allocating from it in l2cap_get_ident using ida_alloc_range which would reuse ids as they are free, then upon completion release the id using ida_free.

This fixes the qualification test case L2CAP/COS/CED/BI-29-C which attempts to check if the host stack is able to work after 256 attempts to connect which would for Ident field to use the full range.

---
 include/net/bluetooth/l2cap.h |  3 +--
 net/bluetooth/l2cap_core.c    | 46 ++++++++++++++++++++++++-----------
 2 files changed, 33 insertions(+), 16 deletions(-)
include/net/bluetooth/l2cap.h +1 -2
index 00e182a22720..ec3af01e4db9 100644
@@ -655,8 +655,7 @@ struct l2cap_conn {
 
 	struct sk_buff		*rx_skb;
 	__u32			rx_len;
-	__u8			tx_ident;
-	struct mutex		ident_lock;
+	struct ida		tx_ida;
 
 	struct sk_buff_head	pending_rx;
 	struct work_struct	pending_rx_work;
net/bluetooth/l2cap_core.c +32 -14

95 lines not on this page · raw message

Signature · 1 line
2.52.0
#1 lore raw

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=1047242

---Test result---

Test Summary:

CheckPatch                    PENDING   0.48 seconds
GitLint                       PENDING   0.37 seconds
SubjectPrefix                 PASS      0.07 seconds
BuildKernel                   PASS      24.97 seconds
CheckAllWarning               PASS      27.39 seconds
CheckSparse                   PASS      30.86 seconds
BuildKernel32                 PASS      24.67 seconds
TestRunnerSetup               PASS      556.94 seconds
TestRunner_l2cap-tester       PASS      28.76 seconds
TestRunner_iso-tester         PASS      76.65 seconds
TestRunner_bnep-tester        PASS      6.66 seconds
TestRunner_mgmt-tester        FAIL      116.23 seconds
TestRunner_rfcomm-tester      PASS      9.40 seconds
TestRunner_sco-tester         FAIL      14.44 seconds
TestRunner_ioctl-tester       PASS      10.15 seconds
TestRunner_mesh-tester        FAIL      11.47 seconds
TestRunner_smp-tester         PASS      8.57 seconds
TestRunner_userchan-tester    PASS      6.58 seconds
IncrementalBuild              PENDING   0.61 seconds

Details
##############################
Test: CheckPatch - PENDING
Desc: Run checkpatch.pl script
Output:

##############################
Test: GitLint - PENDING
Desc: Run gitlint
Output:

##############################
Test: TestRunner_mgmt-tester - FAIL
Desc: Run mgmt-tester with test-runner Output:
Total: 494, Passed: 486 (98.4%), Failed: 4, Not Run: 4

Failed Test Cases

Read Exp Feature - Success                           Failed       0.111 seconds
LL Privacy - Set Flags 1 (Add to RL)                 Failed       0.154 seconds
LL Privacy - Start Discovery 1 (Disable RL)          Failed       0.174 seconds
LL Privacy - Start Discovery 2 (Disable RL)          Failed       0.186 seconds

##############################
Test: TestRunner_sco-tester - FAIL
Desc: Run sco-tester with test-runner
Output:
WARNING: possible circular locking dependency detected

BUG: sleeping function called from invalid context at net/core/sock.c:3782

Total: 30, Passed: 30 (100.0%), Failed: 0, Not Run: 0
##############################
Test: TestRunner_mesh-tester - FAIL
Desc: Run mesh-tester with test-runner Output:
Total: 10, Passed: 8 (80.0%), Failed: 2, Not Run: 0

Failed Test Cases

Mesh - Send cancel - 1                               Timed out    1.913 seconds
Mesh - Send cancel - 2                               Timed out    1.996 seconds

##############################
Test: IncrementalBuild - PENDING
Desc: Incremental build with the patches in the series Output:

---
Regards,
Linux Bluetooth

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

Hello:

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

Quoted 8 lines · On Mon, 26 Jan 2026 12:53:40 -0500 you wrote:

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

This attempts to proper track outstanding request by using struct ida and allocating from it in l2cap_get_ident using ida_alloc_range which would reuse ids as they are free, then upon completion release the id using ida_free.

[...]

Here is the summary with links:
- [v3] Bluetooth: L2CAP: Fix not tracking outstanding TX ident

    https://git.kernel.org/bluetooth/bluetooth-next/c/8e7c7cb168a5

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
Reviewed-by on v3 1/1
#3 lore raw
In reply to Luiz Augusto von Dentz

Dear Luiz,

Thank you for your patch.

Am 26.01.26 um 18:53 schrieb Luiz Augusto von Dentz:

Quoted 3 lines

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

This attempts to proper track outstanding request by using struct ida

proper*ly*

Quoted 7 lines

and allocating from it in l2cap_get_ident using ida_alloc_range which would reuse ids as they are free, then upon completion release the id using ida_free.

This fixes the qualification test case L2CAP/COS/CED/BI-29-C which attempts to check if the host stack is able to work after 256 attempts to connect which would for Ident field to use the full range.

The last sentence seems incomplete.

Quoted 115 lines
---
  include/net/bluetooth/l2cap.h |  3 +--
  net/bluetooth/l2cap_core.c    | 46 ++++++++++++++++++++++++-----------
  2 files changed, 33 insertions(+), 16 deletions(-)
include/net/bluetooth/l2cap.h +1 -2
index 00e182a22720..ec3af01e4db9 100644
@@ -655,8 +655,7 @@ struct l2cap_conn {
 
  	struct sk_buff		*rx_skb;
  	__u32			rx_len;
-	__u8			tx_ident;
-	struct mutex		ident_lock;
+	struct ida		tx_ida;
 
  	struct sk_buff_head	pending_rx;
  	struct work_struct	pending_rx_work;
net/bluetooth/l2cap_core.c +31 -14
index 07b493331fd7..b628b0fa39b2 100644
@@ -924,26 +924,18 @@ int l2cap_chan_check_security(struct l2cap_chan *chan, bool initiator)
  				 initiator);
  }
 
-static u8 l2cap_get_ident(struct l2cap_conn *conn)
+static int l2cap_get_ident(struct l2cap_conn *conn)
  {
-	u8 id;
+	/* LE link does not support tools like l2ping so use the full range */
+	if (conn->hcon->type == LE_LINK)
+		return ida_alloc_range(&conn->tx_ida, 1, 255, GFP_ATOMIC);
 
  	/* Get next available identificator.
  	 *    1 - 128 are used by kernel.
  	 *  129 - 199 are reserved.
  	 *  200 - 254 are used by utilities like l2ping, etc.
  	 */
-
-	mutex_lock(&conn->ident_lock);
-
-	if (++conn->tx_ident > 128)
-		conn->tx_ident = 1;
-
-	id = conn->tx_ident;
-
-	mutex_unlock(&conn->ident_lock);
-
-	return id;
+	return ida_alloc_range(&conn->tx_ida, 1, 128, GFP_ATOMIC);
  }
 
  static void l2cap_send_acl(struct l2cap_conn *conn, struct sk_buff *skb,
@@ -1773,6 +1765,8 @@ static void l2cap_conn_del(struct hci_conn *hcon, int err)
  	if (work_pending(&conn->pending_rx_work))
  		cancel_work_sync(&conn->pending_rx_work);
 
+	ida_destroy(&conn->tx_ida);
+
  	cancel_delayed_work_sync(&conn->id_addr_timer);
 
  	l2cap_unregister_all_users(conn);
@@ -4782,12 +4776,34 @@ static int l2cap_le_connect_rsp(struct l2cap_conn *conn,
  	return err;
  }
 
+static void l2cap_put_ident(struct l2cap_conn *conn, u8 code, u8 id)
+{
+	switch (code) {
+	case L2CAP_COMMAND_REJ:
+	case L2CAP_CONN_RSP:
+	case L2CAP_CONF_RSP:
+	case L2CAP_DISCONN_RSP:
+	case L2CAP_ECHO_RSP:
+	case L2CAP_INFO_RSP:
+	case L2CAP_CONN_PARAM_UPDATE_RSP:
+	case L2CAP_ECRED_CONN_RSP:
+	case L2CAP_ECRED_RECONF_RSP:
+		/* First do a lookup since the remote may send bogus ids that
+		 * would make ida_free to generate warnings.
+		 */
+		if (ida_find_first_range(&conn->tx_ida, id, id) >= 0)
+			ida_free(&conn->tx_ida, id);
+	}
+}
+
  static inline int l2cap_bredr_sig_cmd(struct l2cap_conn *conn,
  				      struct l2cap_cmd_hdr *cmd, u16 cmd_len,
  				      u8 *data)
  {
  	int err = 0;
 
+	l2cap_put_ident(conn, cmd->code, cmd->ident);
+
  	switch (cmd->code) {
  	case L2CAP_COMMAND_REJ:
  		l2cap_command_rej(conn, cmd, cmd_len, data);
@@ -5419,6 +5435,8 @@ static inline int l2cap_le_sig_cmd(struct l2cap_conn *conn,
  {
  	int err = 0;
 
+	l2cap_put_ident(conn, cmd->code, cmd->ident);
+
  	switch (cmd->code) {
  	case L2CAP_COMMAND_REJ:
  		l2cap_le_command_rej(conn, cmd, cmd_len, data);
@@ -6907,13 +6925,13 @@ static struct l2cap_conn *l2cap_conn_add(struct hci_conn *hcon)
  	     hci_dev_test_flag(hcon->hdev, HCI_FORCE_BREDR_SMP)))
  		conn->local_fixed_chan |= L2CAP_FC_SMP_BREDR;
 
-	mutex_init(&conn->ident_lock);

net/bluetooth/l2cap_core.c static struct l2cap_conn *l2cap_conn_add(struct hci_conn *hcon) · line 6937

Quoted 10 lines
net/bluetooth/l2cap_core.c +1 -0
  	mutex_init(&conn->lock);
 
  	INIT_LIST_HEAD(&conn->chan_l);
  	INIT_LIST_HEAD(&conn->users);
 
  	INIT_DELAYED_WORK(&conn->info_timer, l2cap_info_timeout);
+	ida_init(&conn->tx_ida);
 
  	skb_queue_head_init(&conn->pending_rx);
  	INIT_WORK(&conn->pending_rx_work, process_pending_rx);

Kind regards,

Paul

#4 lore raw
In reply to Luiz Augusto von Dentz

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

The symptom is that when connecting the headphones via bluetoothctl, it either hangs or appears (in pavucontrol) as if it's using the AAC codec, but there's no output. Sometimes after some time, it disconnects and then re-connects successfully but using the MSBC/CVSD codecs.

I haven't had time to troubleshoot this further so far, mainly posting if it can save someone a bisect.

Bluetooth chip: Intel Corp. AX210 Bluetooth (ThinkPad T14 Gen 2)
Device: Bose QuietComfort Ultra (bluetooth:v009Ep4066d0167) Bluez 5.86-4 (Arch Linux)
Kconfig: That of the linux-mainline AUR package, plus UBSAN

#5 lore raw
Quoted 8 lines · Joan Bruguera Micó

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

The symptom is that when connecting the headphones via bluetoothctl, it either hangs or appears (in pavucontrol) as if it's using the AAC codec, but there's no output. Sometimes after some time, it disconnects and then re-connects successfully but using the MSBC/CVSD codecs.

Quoted 7 lines

I haven't had time to troubleshoot this further so far, mainly posting if it can save someone a bisect.

Bluetooth chip: Intel Corp. AX210 Bluetooth (ThinkPad T14 Gen 2)
Device: Bose QuietComfort Ultra (bluetooth:v009Ep4066d0167) Bluez 5.86-4 (Arch Linux)
Kconfig: That of the linux-mainline AUR package, plus UBSAN

FWIW, someone reported problems due to that commit in bugzilla, too:

To quote https://bugzilla.kernel.org/show_bug.cgi?id=221120:
"""
Bug 221120 - Bluetooth headphones stop connecting properly after commit 6c3ea155e5ee3e56606233acde8309afda66d483

Bisected to this commit. Before this headphones can be connected using A2DP music profile or HSP/HFP for bidirectional (headphone and microphone mono). After this dependig on the model of the device it has only one of options available(Marshall Major iv keeps only HSP while Chinese no-name keep only A2DP ). Reverting just this commit from Linus's tree restores normal bluetooth audio.
"""

Ciao, Thorsten

author maintainer
#6 lore raw
In reply to Joan Bruguera Micó

Hi Joan,

Quoted 9 lines · Joan Bruguera Micó

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

The symptom is that when connecting the headphones via bluetoothctl, it either hangs or appears (in pavucontrol) as if it's using the AAC codec, but there's no output. Sometimes after some time, it disconnects and then re-connects successfully but using the MSBC/CVSD codecs.

Quoted 7 lines

I haven't had time to troubleshoot this further so far, mainly posting if it can save someone a bisect.

Bluetooth chip: Intel Corp. AX210 Bluetooth (ThinkPad T14 Gen 2)
Device: Bose QuietComfort Ultra (bluetooth:v009Ep4066d0167) Bluez 5.86-4 (Arch Linux)
Kconfig: That of the linux-mainline AUR package, plus UBSAN

Weird, there shouldn't be any logical difference but perhaps it's device-specific. Can you collect the btmon traces? Actually it would be great to have the before and after traces for comparison.

Signature · 1 line
Luiz Augusto von Dentz
#7 lore raw

Hi Luiz,

Quoted 5 lines
Quoted 2 lines

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

Weird, there shouldn't be any logical difference but perhaps it's device-specific. Can you collect the btmon traces? Actually it would be great to have the before and after traces for comparison.

I have attached btmon traces (bad/good) to the open Bugzilla bug.
(https://bugzilla.kernel.org/show_bug.cgi?id=221120), hopefully they are helpful in diagnosing the issue.

I have also done experiments and found that the regression goes away if the IDs are not re-used, for example by commenting out the ida_free line, or by replacing ida_alloc_range (returns smallest available ID) with idr_alloc_cyclic (returns all IDs before wrapping around).

Though I'm not sure if that is just masking a Kernel bug (e.g. IDs getting released/reused too early), or if that's necessary for my hardware to not get confused by the same IDs being quickly reused.

Regards,
- Joan

author maintainer
#8 lore raw

Hi Joan,

Quoted 9 lines · Joan Bruguera Micó

Hi Luiz,

Quoted 5 lines
Quoted 2 lines

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

Weird, there shouldn't be any logical difference but perhaps it's device-specific. Can you collect the btmon traces? Actually it would be great to have the before and after traces for comparison.

Quoted 8 lines

I have attached btmon traces (bad/good) to the open Bugzilla bug.
(https://bugzilla.kernel.org/show_bug.cgi?id=221120), hopefully they are helpful in diagnosing the issue.

I have also done experiments and found that the regression goes away if the IDs are not re-used, for example by commenting out the ida_free line, or by replacing ida_alloc_range (returns smallest available ID) with idr_alloc_cyclic (returns all IDs before wrapping around).

Interesting, I didn't consider that the implementation would care about reusing the identifier as long as it is not outstanding, but if making it cyclic resolves the issue, that's probably the case.

Quoted 3 lines

Though I'm not sure if that is just masking a Kernel bug (e.g. IDs getting released/reused too early), or if that's necessary for my hardware to not get confused by the same IDs being quickly reused.

Even if it is a bug in the remote side I guess we don't want to just break support, we will just need to make sure using idr_alloc_cyclic doesn't break L2CAP/COS/CED/BI-29-C qualification test.

Quoted 2 lines

Regards,
- Joan

Signature · 1 line
Luiz Augusto von Dentz
#9 lore raw

CCing the net maintainers because this is a regression that Linus afaics would like to have seen fixed by now.

Quoted 4 lines · Luiz Augusto von Dentz
Quoted 2 lines · Joan Bruguera Micó
Quoted 2 lines
Quoted 2 lines

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

Side note: there is now a third report where the reporter confirmed that reverting this patch (6c3ea155e5ee3e ("Bluetooth: L2CAP: Fix not tracking outstanding TX ident") [v7.0-rc1]) fixes the issue:
https://bugzilla.kernel.org/show_bug.cgi?id=221177

But I'm writing for another reason:

Luiz, does it look like a fix will become available really soon now? If not, I wonder if we should just revert the culprit at this point (and reapply a improved change for 7.1), as the regression was first reported on 2026-02-22 – so we are way beyond the ""rule of thumb should generally be "[fix regressions] within a week", preferably before the next rc."" that Linus recently mentioned:
https://www.kernel.org/doc/html/next/process/handling-regressions.html#on-how-quickly-regressions-should-be-fixed

Ciao, Thorsten

Quoted 24 lines
Quoted 4 lines
Quoted 3 lines

Weird, there shouldn't be any logical difference but perhaps it's device-specific. Can you collect the btmon traces? Actually it would be great to have the before and after traces for comparison.

Quoted 8 lines

I have attached btmon traces (bad/good) to the open Bugzilla bug.
(https://bugzilla.kernel.org/show_bug.cgi?id=221120), hopefully they are helpful in diagnosing the issue.

I have also done experiments and found that the regression goes away if the IDs are not re-used, for example by commenting out the ida_free line, or by replacing ida_alloc_range (returns smallest available ID) with idr_alloc_cyclic (returns all IDs before wrapping around).

Interesting, I didn't consider that the implementation would care about reusing the identifier as long as it is not outstanding, but if making it cyclic resolves the issue, that's probably the case.

Quoted 3 lines

Though I'm not sure if that is just masking a Kernel bug (e.g. IDs getting released/reused too early), or if that's necessary for my hardware to not get confused by the same IDs being quickly reused.

Even if it is a bug in the remote side I guess we don't want to just break support, we will just need to make sure using idr_alloc_cyclic doesn't break L2CAP/COS/CED/BI-29-C qualification test.

author maintainer
#10 lore raw

Hi Thorsten,

Quoted 17 lines · Thorsten Leemhuis

CCing the net maintainers because this is a regression that Linus afaics would like to have seen fixed by now.

Quoted 4 lines · Luiz Augusto von Dentz
Quoted 2 lines · Joan Bruguera Micó
Quoted 2 lines
Quoted 2 lines

I had trouble connecting my Bluetooth headphones on Linux 7.0-rc1, and a bisect landed me on this commit 6c3ea155e5ee. A revert fixes it.

Side note: there is now a third report where the reporter confirmed that reverting this patch (6c3ea155e5ee3e ("Bluetooth: L2CAP: Fix not tracking outstanding TX ident") [v7.0-rc1]) fixes the issue:
https://bugzilla.kernel.org/show_bug.cgi?id=221177

But I'm writing for another reason:

Quoted 7 lines

Luiz, does it look like a fix will become available really soon now? If not, I wonder if we should just revert the culprit at this point (and reapply a improved change for 7.1), as the regression was first reported on 2026-02-22 – so we are way beyond the ""rule of thumb should generally be "[fix regressions] within a week", preferably before the next rc."" that Linus recently mentioned:
https://www.kernel.org/doc/html/next/process/handling-regressions.html#on-how-quickly-regressions-should-be-fixed

idr_alloc_cyclic is not compatible with struct ida, and changing it to use struct idr will probably require reintroducing a lock since the idr_ functions appear unsafe to run in any context as the ida versions.

Signature · 1 line
Luiz Augusto von Dentz