[PATCH] Bluetooth: use hdev lock for accept_list and reject_list in conn req

[PATCH] Bluetooth: use hdev lock for accept_list and reject_list in conn req

[Niels Dossche]

All accesses (both read and modifications) to
hdev->{accept,reject}_list are protected by hdev lock,
except the ones in hci_conn_request_evt. This can cause a race condition
in the form of a list corruption.
The solution is to protect these lists in hci_conn_request_evt as well.

I was unable to find the exact commit that introduced the issue for the
reject list, I was only able to find it for the accept list.

Note:
I am currently working on a static analyser to detect missing locks
using type-based static analysis as my master's thesis
in order to obtain my master's degree.
If you would like to have more details, please let me know.
This was a reported case. I manually verified the report by looking
at the code, so that I do not send wrong information or patches.
After concluding that this seems to be a true positive, I created
this patch. I have both compile-tested this patch and runtime-tested
this patch on x86_64. The effect on a running system could be a
potential race condition in exceptional cases.
This issue was found on Linux v5.17.1.

Fixes: a55bd29d5227 ("Bluetooth: Add white list lookup for incoming connection requests")
Signed-off-by: Niels Dossche <dossche.niels@gmail.com>
---
 net/bluetooth/hci_event.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index abaabfae19cc..b9038f24f46f 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -3222,8 +3222,11 @@ static void hci_conn_request_evt(struct hci_dev *hdev, void *data,
 		return;
 	}
 
+	hci_dev_lock(hdev);
+
 	if (hci_bdaddr_list_lookup(&hdev->reject_list, &ev->bdaddr,
 				   BDADDR_BREDR)) {
+		hci_dev_unlock(hdev);
 		hci_reject_conn(hdev, &ev->bdaddr);
 		return;
 	}
@@ -3236,14 +3239,13 @@ static void hci_conn_request_evt(struct hci_dev *hdev, void *data,
 	    !hci_dev_test_flag(hdev, HCI_CONNECTABLE) &&
 	    !hci_bdaddr_list_lookup_with_flags(&hdev->accept_list, &ev->bdaddr,
 					       BDADDR_BREDR)) {
+		hci_dev_unlock(hdev);
 		hci_reject_conn(hdev, &ev->bdaddr);
 		return;
 	}
 
 	/* Connection accepted */
 
-	hci_dev_lock(hdev);
-
 	ie = hci_inquiry_cache_lookup(hdev, &ev->bdaddr);
 	if (ie)
 		memcpy(ie->data.dev_class, ev->dev_class, 3);
-- 
2.35.1

RE: Bluetooth: use hdev lock for accept_list and reject_list in conn req

[bluez.test.bot]

[-- Attachment #1: Type: text/plain, Size: 1096 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=628619

---Test result---

Test Summary:
CheckPatch                    PASS      1.54 seconds
GitLint                       PASS      1.01 seconds
SubjectPrefix                 PASS      0.89 seconds
BuildKernel                   PASS      30.05 seconds
BuildKernel32                 PASS      27.33 seconds
Incremental Build with patchesPASS      38.38 seconds
TestRunner: Setup             PASS      458.46 seconds
TestRunner: l2cap-tester      PASS      15.27 seconds
TestRunner: bnep-tester       PASS      6.04 seconds
TestRunner: mgmt-tester       PASS      99.50 seconds
TestRunner: rfcomm-tester     PASS      7.31 seconds
TestRunner: sco-tester        PASS      7.24 seconds
TestRunner: smp-tester        PASS      6.75 seconds
TestRunner: userchan-tester   PASS      5.87 seconds



---
Regards,
Linux Bluetooth

Re: [PATCH] Bluetooth: use hdev lock for accept_list and reject_list in conn req

[Marcel Holtmann]

Hi Niels,

> All accesses (both read and modifications) to
> hdev->{accept,reject}_list are protected by hdev lock,
> except the ones in hci_conn_request_evt. This can cause a race condition
> in the form of a list corruption.
> The solution is to protect these lists in hci_conn_request_evt as well.
> 
> I was unable to find the exact commit that introduced the issue for the
> reject list, I was only able to find it for the accept list.
> 
> Note:
> I am currently working on a static analyser to detect missing locks
> using type-based static analysis as my master's thesis
> in order to obtain my master's degree.
> If you would like to have more details, please let me know.
> This was a reported case. I manually verified the report by looking
> at the code, so that I do not send wrong information or patches.
> After concluding that this seems to be a true positive, I created
> this patch. I have both compile-tested this patch and runtime-tested
> this patch on x86_64. The effect on a running system could be a
> potential race condition in exceptional cases.
> This issue was found on Linux v5.17.1.

this doesn’t belong in the commit message.

> 
> Fixes: a55bd29d5227 ("Bluetooth: Add white list lookup for incoming connection requests")
> Signed-off-by: Niels Dossche <dossche.niels@gmail.com>
> ---
> net/bluetooth/hci_event.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> index abaabfae19cc..b9038f24f46f 100644
> --- a/net/bluetooth/hci_event.c
> +++ b/net/bluetooth/hci_event.c
> @@ -3222,8 +3222,11 @@ static void hci_conn_request_evt(struct hci_dev *hdev, void *data,
> 		return;
> 	}
> 
> +	hci_dev_lock(hdev);
> +
> 	if (hci_bdaddr_list_lookup(&hdev->reject_list, &ev->bdaddr,
> 				   BDADDR_BREDR)) {
> +		hci_dev_unlock(hdev);
> 		hci_reject_conn(hdev, &ev->bdaddr);
> 		return;
> 	}
> @@ -3236,14 +3239,13 @@ static void hci_conn_request_evt(struct hci_dev *hdev, void *data,
> 	    !hci_dev_test_flag(hdev, HCI_CONNECTABLE) &&
> 	    !hci_bdaddr_list_lookup_with_flags(&hdev->accept_list, &ev->bdaddr,
> 					       BDADDR_BREDR)) {
> +		hci_dev_unlock(hdev);
> 		hci_reject_conn(hdev, &ev->bdaddr);
> 		return;
> 	}

Can we use an exit goto-label instead. I am not a big fan of "unbalanced” _unlock statements.

> 	/* Connection accepted */
> 
> -	hci_dev_lock(hdev);
> -
> 	ie = hci_inquiry_cache_lookup(hdev, &ev->bdaddr);
> 	if (ie)
> 		memcpy(ie->data.dev_class, ev->dev_class, 3);

Regards

Marcel

Re: [PATCH] Bluetooth: use hdev lock for accept_list and reject_list in conn req

[Niels Dossche]

On 05/04/2022 19:18, Marcel Holtmann wrote:
> Hi Niels,
> 
>> All accesses (both read and modifications) to
>> hdev->{accept,reject}_list are protected by hdev lock,
>> except the ones in hci_conn_request_evt. This can cause a race condition
>> in the form of a list corruption.
>> The solution is to protect these lists in hci_conn_request_evt as well.
>>
>> I was unable to find the exact commit that introduced the issue for the
>> reject list, I was only able to find it for the accept list.
>>
>> Note:
>> I am currently working on a static analyser to detect missing locks
>> using type-based static analysis as my master's thesis
>> in order to obtain my master's degree.
>> If you would like to have more details, please let me know.
>> This was a reported case. I manually verified the report by looking
>> at the code, so that I do not send wrong information or patches.
>> After concluding that this seems to be a true positive, I created
>> this patch. I have both compile-tested this patch and runtime-tested
>> this patch on x86_64. The effect on a running system could be a
>> potential race condition in exceptional cases.
>> This issue was found on Linux v5.17.1.
> 
> this doesn’t belong in the commit message.

Hi Marcel
I'll remove it from the commit message.
I can write it in below the --- in the future such that it won't be included.

> 
>>
>> Fixes: a55bd29d5227 ("Bluetooth: Add white list lookup for incoming connection requests")
>> Signed-off-by: Niels Dossche <dossche.niels@gmail.com>
>> ---
>> net/bluetooth/hci_event.c | 6 ++++--
>> 1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
>> index abaabfae19cc..b9038f24f46f 100644
>> --- a/net/bluetooth/hci_event.c
>> +++ b/net/bluetooth/hci_event.c
>> @@ -3222,8 +3222,11 @@ static void hci_conn_request_evt(struct hci_dev *hdev, void *data,
>> 		return;
>> 	}
>>
>> +	hci_dev_lock(hdev);
>> +
>> 	if (hci_bdaddr_list_lookup(&hdev->reject_list, &ev->bdaddr,
>> 				   BDADDR_BREDR)) {
>> +		hci_dev_unlock(hdev);
>> 		hci_reject_conn(hdev, &ev->bdaddr);
>> 		return;
>> 	}
>> @@ -3236,14 +3239,13 @@ static void hci_conn_request_evt(struct hci_dev *hdev, void *data,
>> 	    !hci_dev_test_flag(hdev, HCI_CONNECTABLE) &&
>> 	    !hci_bdaddr_list_lookup_with_flags(&hdev->accept_list, &ev->bdaddr,
>> 					       BDADDR_BREDR)) {
>> +		hci_dev_unlock(hdev);
>> 		hci_reject_conn(hdev, &ev->bdaddr);
>> 		return;
>> 	}
> 
> Can we use an exit goto-label instead. I am not a big fan of "unbalanced” _unlock statements.

I'll make a v2 and send it.

> 
>> 	/* Connection accepted */
>>
>> -	hci_dev_lock(hdev);
>> -
>> 	ie = hci_inquiry_cache_lookup(hdev, &ev->bdaddr);
>> 	if (ie)
>> 		memcpy(ie->data.dev_class, ev->dev_class, 3);
> 
> Regards
> 
> Marcel
> 

Thanks
Niels