2024-03-22 14:39:01

by Yao Xiao

Subject: [PATCH BlueZ] avctp: Remove redundant UUID checks and additions

From: Xiao Yao <[email protected]>

---
profiles/audio/avctp.c | 6 ------
1 file changed, 6 deletions(-)

diff --git a/profiles/audio/avctp.c b/profiles/audio/avctp.c
index 8ad146df1..6199ee951 100644
--- a/profiles/audio/avctp.c
+++ b/profiles/audio/avctp.c
@@ -1597,12 +1597,6 @@ static void avctp_confirm_cb(GIOChannel *chan, gpointer data)
if (session == NULL)
return;

- if (btd_device_get_service(device, AVRCP_REMOTE_UUID) == NULL)
- btd_device_add_uuid(device, AVRCP_REMOTE_UUID);
-
- if (btd_device_get_service(device, AVRCP_TARGET_UUID) == NULL)
- btd_device_add_uuid(device, AVRCP_TARGET_UUID);
-
switch (psm) {
case AVCTP_CONTROL_PSM:
avctp_control_confirm(session, chan, device);
--
2.34.1

2024-03-22 15:29:17

by Luiz Augusto von Dentz

Subject: Re: [PATCH BlueZ] avctp: Remove redundant UUID checks and additions

Hi,

On Fri, Mar 22, 2024 at 2:39 PM Xiao Yao <[email protected]> wrote:
>
> From: Xiao Yao <[email protected]>
>
> ---
> profiles/audio/avctp.c | 6 ------
> 1 file changed, 6 deletions(-)
>
> diff --git a/profiles/audio/avctp.c b/profiles/audio/avctp.c
> index 8ad146df1..6199ee951 100644
> --- a/profiles/audio/avctp.c
> +++ b/profiles/audio/avctp.c
> @@ -1597,12 +1597,6 @@ static void avctp_confirm_cb(GIOChannel *chan, gpointer data)
> if (session == NULL)
> return;
>
> - if (btd_device_get_service(device, AVRCP_REMOTE_UUID) == NULL)
> - btd_device_add_uuid(device, AVRCP_REMOTE_UUID);
> -
> - if (btd_device_get_service(device, AVRCP_TARGET_UUID) == NULL)
> - btd_device_add_uuid(device, AVRCP_TARGET_UUID);

Patch description makes no sense to me, this code is adding the UUIDs
of AVCTP if the device over its PSM because at this point perhaps SDP
discovery is not complete.

> switch (psm) {
> case AVCTP_CONTROL_PSM:
> avctp_control_confirm(session, chan, device);
> --
> 2.34.1
>
>


--
Luiz Augusto von Dentz

2024-03-22 16:41:02

by bluez.test.bot

Subject: RE: [BlueZ] avctp: Remove redundant UUID checks and additions

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=837396

---Test result---

Test Summary:
CheckPatch PASS 0.45 seconds
GitLint PASS 0.34 seconds
BuildEll PASS 26.81 seconds
BluezMake PASS 1770.17 seconds
MakeCheck PASS 13.18 seconds
MakeDistcheck PASS 176.25 seconds
CheckValgrind PASS 244.12 seconds
CheckSmatch PASS 349.27 seconds
bluezmakeextell PASS 120.00 seconds
IncrementalBuild PASS 1522.33 seconds
ScanBuild PASS 999.12 seconds



---
Regards,
Linux Bluetooth

2024-03-22 18:07:32

by Yao Xiao

Subject: Re: [PATCH BlueZ] avctp: Remove redundant UUID checks and additions

Hi,

在 2024/3/22 23:28, Luiz Augusto von Dentz 写道:

> Hi,
>
> On Fri, Mar 22, 2024 at 2:39 PM Xiao Yao <[email protected]> wrote:
>> From: Xiao Yao <[email protected]>
>>
>> ---
>> profiles/audio/avctp.c | 6 ------
>> 1 file changed, 6 deletions(-)
>>
>> diff --git a/profiles/audio/avctp.c b/profiles/audio/avctp.c
>> index 8ad146df1..6199ee951 100644
>> --- a/profiles/audio/avctp.c
>> +++ b/profiles/audio/avctp.c
>> @@ -1597,12 +1597,6 @@ static void avctp_confirm_cb(GIOChannel *chan, gpointer data)
>> if (session == NULL)
>> return;
>>
>> - if (btd_device_get_service(device, AVRCP_REMOTE_UUID) == NULL)
>> - btd_device_add_uuid(device, AVRCP_REMOTE_UUID);
>> -
>> - if (btd_device_get_service(device, AVRCP_TARGET_UUID) == NULL)
>> - btd_device_add_uuid(device, AVRCP_TARGET_UUID);
> Patch description makes no sense to me, this code is adding the UUIDs
> of AVCTP if the device over its PSM because at this point perhaps SDP
> discovery is not complete.
In the scenario encountered, the SDP service discovery has already
been completed, with the SDP record only containing AVRCP_REMOTE_UUID.
At this point, the device is added support for AVRCP_TARGET_UUID,
which leads to the following code being executed:
  controller_init(session);
    //avrcp.c:controller_init() 0x55bcdbd690 version 0x0000
    session->controller = controller;
However, avrcp_set_volume relies on session->controller, erroneously
issuing the AVRCP_SET_ABSOLUTE_VOLUME command.

Perhaps add a check to verify that SDP discovery has completed before
adding the UUID?

Additionally, there is another concern: is it appropriate to add support
for both AVRCP_REMOTE_UUID and AVRCP_TARGET_UUID before the completion
of SDP? if one of the UUIDs is not supported,how should it be handled?

>
>> switch (psm) {
>> case AVCTP_CONTROL_PSM:
>> avctp_control_confirm(session, chan, device);
>> --
>> 2.34.1
>>
>>
>