1003591 - [MADAI][Bluetooth] Incoming a2dp connection failed

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Bluetooth, defect, P3)

Description

[ILBEOM KIM]

User Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.1; Trident/4.0; CNS_UA; AD_LOGON=4C47452E4E4554; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR 3.0.30729; Media Center PC 6.0; .NET4.0C; InfoPath.3; Tablet PC 2.0; .NET4.0E; CNS_UA; AD_LOGON=4C47452E4E4554)

Steps to reproduce:

1. Connected to the bluetooth headset
2. Disconnected the bluetooth headset
3. Try to connection by the bluetooth headset side


Actual results:

HFP connection was succeeded but A2DP connection has failed.


Expected results:

HFP & A2DP connection should success when Incoming connection request.

Comment 1

[ILBEOM KIM]

Attachment: a2dp_connection_failed log

Comment 2

[ILBEOM KIM]

the assertion log was occurred.
Could you check this case?

556	01-01	12:51:05.309	V	217	558	AudioFlinger:		 setParameters(io 21, keyvalue closing=true, calling pid 217
557	01-01	12:51:05.309	E	221	2969	bt-btif		 : bta_av_rc_create ACP handle exist for shdl:0
558	01-01	12:51:05.309	V	217	558	AudioFlinger:		 ThreadBase::setParameters() closing=true
559	01-01	12:51:05.309	E	221	844	BTIF_AV		 : ### ASSERT : external/bluetooth/bluedroid/main/../btif/src/btif_av.c line 370 Callback is NULL (0) ###

Thank you for your support.

Comment 3

[ILBEOM KIM]

In the BluetoothA2dpManager.cpp

btav_callbacks_t callback function defines the below element.

static btav_callbacks_t sBtA2dpCallbacks = {
  sizeof(sBtA2dpCallbacks),
  A2dpConnectionStateCallback,
  A2dpAudioStateCallback
};

But, bluedroid uses connection_priority_cb.
(bt_av.h)
/** BT-AV callback structure. */
typedef struct {
    /** set to sizeof(btav_callbacks_t) */
    size_t      size;
    btav_connection_state_callback  connection_state_cb;
    btav_audio_state_callback audio_state_cb;
    btav_connection_priority_callback connection_priority_cb;
    btav_audio_focus_request_callback audio_focus_request_cb;
} btav_callbacks_t;

(btif_av.c)
        case BTA_AV_RC_OPEN_EVT:
            /* Check if connection allowed with this device */
            BTIF_TRACE_DEBUG0("Check A2dp priority of device");
            if (idle_rc_event != 0)
            {
                BTIF_TRACE_DEBUG0("Processing another RC Event ");
                return FALSE;
            }
            idle_rc_event = event;
            memcpy(&idle_rc_data, ((tBTA_AV*)p_data), sizeof(tBTA_AV));
            if (event == BTA_AV_RC_OPEN_EVT )
            {
                bdcpy(btif_av_cb.peer_bda.address, ((tBTA_AV*)p_data)->rc_open.peer_addr);
            }
            else
            {
                bdcpy(btif_av_cb.peer_bda.address, ((tBTA_AV*)p_data)->pend.bd_addr);
            }
            HAL_CBACK(bt_av_callbacks, connection_priority_cb, &(btif_av_cb.peer_bda));
            break;

Could you check this case?
Thanks.

Comment 4

[Ben Tian (inactive)]

Which branch of hardware/libhardware do you fetch from codeaurora? The assertion failure results from extra callback functions added by qcom [1][2] that mismatches the version used by B2G. However the assertion failure is not necessarily relevant to the bug symptom.

Can you capture sniffer/hcidump log as [3] to help us investigate the symptom? Also is the occurrence rate 100%?

[1] https://www.codeaurora.org/cgit/external/gigabyte/platform/hardware/libhardware/commit/include/hardware/bt_av.h?h=caf/kk_3.5_rb4.1&id=03b487823f473bbed527282cfb5e19bb3c13eb0a
[2] https://www.codeaurora.org/cgit/external/gigabyte/platform/hardware/libhardware/commit/include/hardware/bt_av.h?h=caf/kk_3.5_rb4.1&id=d59af8ede363c8c39f624a4b581785ffbe2d14f7
[3] https://wiki.mozilla.org/B2G/Bluetooth-bluedroid#Debugging:

Comment 5

[ILBEOM KIM]

Attachment: btsnoop_hci.log

Comment 6

[ILBEOM KIM]

Attachment: a2dp_conn_fail_hci.log

Comment 7

[ILBEOM KIM]

I attached snoop, and hci log.

I requested the used branch name.
If I receive the branch information then I will update the comment.

Thanks.

Comment 8

[Preeti Raghunath(:Preeti)]

QA,

Please help with reproducing on Open C.

Comment 9

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Preeti Raghunath(:Preeti) from comment #8)
> QA,
> 
> Please help with reproducing on Open C.

Testing on OpenC is useless, because it uses JellyBean + bluez bluetooth stack. Bug reporter uses Kitkat + bluedroid bluetooth stack (probably CAF modified version, not AOSP version). Our reference design is all based on bluedroid (AOSP version).

Comment 10

[Preeti Raghunath(:Preeti)]

Eric,

Can we please check if this is reproducible in house?

Comment 11

[Jocelyn Liu [:jocelyn] [:joliu]]

Hi Preeti,

We couldn't reproduce it right now since we don't have MADAI on hand.
Also, we need hardware/libhardware branch that was used as Comment4 mentioned.
If we get the branch, we can try to build a similar environment to try to reproduce the bug on nexus5.

Hi ILBEOM,
Any updates about the branch name?

Thanks,
Jocelyn

Comment 12

[bhavana bajaj [:bajaj]]

Blocking this is this is basic STR for a shipping device and we progressing via logs etc to debug this since we don't have a device in hand.

Comment 13

[ILBEOM KIM]

I'm sorry that update is so late.
I received the answer by the project team.

The branch name is the confidential information.
So I requested this case to the Qualcomm.

Thank you for your support.

Comment 14

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to ILBEOM KIM from comment #13)
> I'm sorry that update is so late.
> I received the answer by the project team.
> 
> The branch name is the confidential information.
> So I requested this case to the Qualcomm.
> 
> Thank you for your support.

Since it might be related to proprietary bluedroid modification based on Comment 13. So i suggest to have more concrete information for us to debug further, if it's really a gecko bug. Please feel free to re-nom 1.4+ if needed.

Comment 15

[ILBEOM KIM]

I had processed in the Qualcomm case and received the answer.

the connection_priority_cb already was added in the bluedroid, because MADAI uses the latest Blueroid stack.

But Gecko didn't handle the connection_priority_cb, So callback function has the null pointer.

I attach the Qualcomm answer to below.

------------------------------------------------------------------------------------------------------
Based on the functionality of connection_priority_cb, it just performed that connection is allowed from framework in checking mStateMachine when BTA_AV_PENDING_EVT/BTA_AV_RC_OPEN_EVT happened and then go ahead and connect by calling allow_connection in bta_av.c 
I believe that the connection priority function should be handled in jni and framework side because the stack code of MADIA is coming from latest bluedroid stack almost. Otherwise, the code should be regressive. 


Thanks.

Comment 16

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

AFAIK connection_priority_cb is not part of AOSP HAL, Bruce can you help to provide suggestion?

Comment 17

[howie [:howie]]

There is no branch name to reproduce the issue. Also, the bluedroid bluetooth stack is probably CAF modified version. Although QC reported the stack code is coming from latest bluedroid stack "almost", it doesn't indicate connection_priority_cb is part of AOSP HAL as Comment 16 mentioned. Once again, our reference design is all based on the "official AOSP version". Please provide indication that the stack is AOSP version then re-nom if needed. Thanks.

Comment 18

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

connection_priority_cb changes from here, https://www.codeaurora.org/cgit/quic/la/platform/external/bluetooth/bluedroid/commit/btif/src/btif_av.c?h=LNX.LA.3.7&id=457739021acb9a67aa2f2e80afd0e11f584141b4

Comment 19

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to ILBEOM KIM from comment #15)
> I had processed in the Qualcomm case and received the answer.
> 
> the connection_priority_cb already was added in the bluedroid, because MADAI
> uses the latest Blueroid stack.
> 
> But Gecko didn't handle the connection_priority_cb, So callback function has
> the null pointer.
> 
> I attach the Qualcomm answer to below.
> 
> -----------------------------------------------------------------------------
> -------------------------
> Based on the functionality of connection_priority_cb, it just performed that
> connection is allowed from framework in checking mStateMachine when
> BTA_AV_PENDING_EVT/BTA_AV_RC_OPEN_EVT happened and then go ahead and connect
> by calling allow_connection in bta_av.c 
> I believe that the connection priority function should be handled in jni and
> framework side because the stack code of MADIA is coming from latest
> bluedroid stack almost. Otherwise, the code should be regressive. 
> 
> 
> Thanks.

Thanks for sharing this information. I think there are two things we can discuss here:
(1). Does Mozilla want to remain Android HAL neutral? If we can accept non-neutral HAL changes we can have changes like on Bug 1004896 to introduce flags into gecko. If we don't want to have this kind of special flag to check bluedroid implementation, bluedroid btif_av shall have internal flag to enable/disable checking whether priority required to be considered. Although for me, it's like a bug, see (2).

(2). This device connection priority mechanism is for Android only. Default priority for devices that we try to auto-connect to and and allow incoming connections for the profile. Android bluedroid in JB/Kitkat seems does not check this part.
Because Android let users allow an incoming hfp/a2dp connection from Bluetooth Settings menu. If the user unselected a2dp profile (priority=0), bluedroid stack shall block the incoming a2dp connection. It's clear that bluedroid will auto-accept this a2dp connection even if a2dp profile had been un-selected with patch in Comment 18. However, b2g Settings application does not have this UX design. Eventually, we will receive connection_priority_cb callback and we have to directly call |allow_connection()| to accept incoming connection anyway inside gecko. Although this is not necessary since we don't have connection priority for Settings app.

So if we can accept things to allow gecko has a specific CAF bluedroid flag to choose the different HAL interface, we can do (2) stuff.
I hope this can help decision making and please correct me if there is anything wrong.
Thanks.

Comment 20

[ILBEOM KIM]

Dear Mozilla Engineer.

I agree your opinion.
The connection priority check is used in the android scenario.
So FFOS will always try to connect when receive the connection_priority_cb.

For resolved this case, we need to use (2) method because MADAI uses the bluedroid modified by qualcomm.

Could you release the flag patch?

Comment 21

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Attachment: Add CAF bluedroid flag and expose env variables from gonk-misc (Patch 1)

Comment 22

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

You need to add flag in your BoardConfig.mk with Patch 1, BOARD_HAVE_BLUETOOTH_BLUEDROID_CAF := true

Comment 23

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Attachment: Add vendor specific flag for customized bluetooth HAL (Gecko) (Patch 2)

Comment 24

[Preeti Raghunath(:Preeti)]

Please help understand if CAF is looking at this?

Comment 25

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Attachment: Bug 1003591 - Allow a2dp incoming connection (Patch 3)

Comment 26

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #25)
> Created attachment 8419948 [details] [diff] [review]
> Bug 1003591 - Allow a2dp incoming connection (Patch 3)

oh...you probably need to replace allowConnection to allow_connection, because I tested CAF on JB branch.

Comment 27

[ILBEOM KIM]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #26)
> (In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #25)
> > Created attachment 8419948 [details] [diff] [review]
> > Bug 1003591 - Allow a2dp incoming connection (Patch 3)
> 
> oh...you probably need to replace allowConnection to allow_connection,
> because I tested CAF on JB branch.

I will check the patch and update to result.

Thanks.

Comment 28

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to ILBEOM KIM from comment #27)
> (In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #26)
> > (In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #25)
> > > Created attachment 8419948 [details] [diff] [review]
> > > Bug 1003591 - Allow a2dp incoming connection (Patch 3)
> > 
> > oh...you probably need to replace allowConnection to allow_connection,
> > because I tested CAF on JB branch.
> 
> I will check the patch and update to result.
> 
> Thanks.

Please make sure you delete objdir-gecko to rebuild gecko.

Comment 29

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Comment on attachment 8419886 [details] [diff] [review]
Add vendor specific flag for customized bluetooth HAL (Gecko) (Patch 2)

Review of attachment 8419886 [details] [diff] [review]:
-----------------------------------------------------------------

It seems like we cannot assign nullptr here.

::: dom/bluetooth/bluedroid/BluetoothServiceBluedroid.cpp
@@ +495,1 @@
>  

+#ifdef MOZ_B2G_BT_BLUEDROID_CAF
+static void
+WakeStateChangedCallback(bt_state_t state)
+{
+}
+#endif

@@ +630,5 @@
>    RemoteDevicePropertiesCallback,
>    DeviceFoundCallback,
>    DiscoveryStateChangedCallback,
> +#ifdef MOZ_B2G_BT_BLUEDROID_CAF
> +  nullptr,

WakeStateChangedCallback,

Comment 30

[Inder]

This issue has been fixed in CAF for LF builds, see:[1a][1b]
If you sync to AU86+ [2] then you shouldn't see this issue.
So, the changes in this bug are not needed.

Thanks Preeti for checking.

[1a] https://www.codeaurora.org/cgit/quic/la/platform/hardware/libhardware/commit/?h=LNX.LF.3.5.2.2&id=b11effcb128e22dc3090aed9afdf7afa2a270005
[1b] https://www.codeaurora.org/cgit/quic/la/platform/external/bluetooth/bluedroid/commit/?h=b2g_kk_3.5&id=4494d0952636c490b9bfd8c71105be8f9edc13ba
[2] https://www.codeaurora.org/cgit/quic/lf/b2g/manifest/tag/?id=AU_LINUX_GECKO_B2G_KK_3.5.01.04.00.113.086

Comment 31

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Inder from comment #30)
> This issue has been fixed in CAF for LF builds, see:[1a][1b]
> If you sync to AU86+ [2] then you shouldn't see this issue.
> So, the changes in this bug are not needed.
But in btif_av.c, Does connection_priority_cb still need to be added?
Can we have a flag for btif_av.h and btif_av.c?

Comment 32

[Inder]

Greg -- can you please comment.

Comment 33

[Greg Grisco]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #31)
> (In reply to Inder from comment #30)
> > This issue has been fixed in CAF for LF builds, see:[1a][1b]
> > If you sync to AU86+ [2] then you shouldn't see this issue.
> > So, the changes in this bug are not needed.
> But in btif_av.c, Does connection_priority_cb still need to be added?
> Can we have a flag for btif_av.h and btif_av.c?

Yeah, this doesn't cause a compilation error because, I guess, the new callback was added to the end of the list.  But if it's causing runtime issue then I think we need to conditionally compile it like we did with the other bluetooth.h callbacks.  I'll take a look.

Comment 34

[ILBEOM KIM]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #28)
> (In reply to ILBEOM KIM from comment #27)
> > (In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #26)
> > > (In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #25)
> > > > Created attachment 8419948 [details] [diff] [review]
> > > > Bug 1003591 - Allow a2dp incoming connection (Patch 3)
> > > 
> > > oh...you probably need to replace allowConnection to allow_connection,
> > > because I tested CAF on JB branch.
> > 
> > I will check the patch and update to result.
> > 
> > Thanks.
> 
> Please make sure you delete objdir-gecko to rebuild gecko.

Dear Mozilla Engineer.

I tested the patch, and modified some codes.

1. I added the below macro in the configure.in
AC_SUBST(MOZ_B2G_BT_BLUEDROID_CAF)
If remove the above macro, then "if CONFIG['MOZ_B2G_BT_BLUEDROID_CAF']:" condition isn't adapted.

2.  if test -n "${BOARD_HAVE_BLUETOOTH_BLUEDROID_CAF}"; then
    ===>  if test -n "${BOARD_HAVE_BLUEDROID_CAF}"; then

I tested the patch, and a2dp connection failed wasn't reproduced.

Thank you for your support.

Comment 35

[Thomas Zimmermann [:tzimmermann] [:tdz]]

Comment on attachment 8419886 [details] [diff] [review]
Add vendor specific flag for customized bluetooth HAL (Gecko) (Patch 2)

Review of attachment 8419886 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/bluetooth/bluedroid/BluetoothServiceBluedroid.cpp
@@ +630,5 @@
>    RemoteDevicePropertiesCallback,
>    DeviceFoundCallback,
>    DiscoveryStateChangedCallback,
> +#ifdef MOZ_B2G_BT_BLUEDROID_CAF
> +  nullptr,

Simply assign the fields by value, like this

static struct s {
  .field = FieldFunc
};

All unassigned fields will be set to zero automatically. As long as you're not going to use them, you won't need these preprocessor defines and related build support.

Comment 36

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Hi Michael,

https://www.codeaurora.org/cgit/quic/la/platform/hardware/libhardware/tree/include/hardware/bt_av.h?h=LNX.LF.3.5.2.2&id=b11effcb128e22dc3090aed9afdf7afa2a270005#n61

Can we have a Q_BLUETOOTH flag to disable btav_connection_priority_callback? This may also need to change bluedroid stack.

Comment 37

[Michael Vines [:m1] [:evilmachines]]

Yeah we could add a Q_BLUETOOTH flag here.  This is not a high priority bug though, based on the current flags so we are unlikely to prioritize it.

Comment 38

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Michael Vines [:m1] [:evilmachines] from comment #37)
> Yeah we could add a Q_BLUETOOTH flag here.  This is not a high priority bug
> though, based on the current flags so we are unlikely to prioritize it.
I wonder CAF has additional patch on Gecko for this bug. Otherwise, a2dp incoming connection shall always fail. Honestly, I really hope you can add Q_BLUETOOTH for btav_connection_priority_callback. Without the flag, we have to introduce CAF flag ourself. It really doesn't make sense to fail CAF build.

Comment 39

[Michael Wu [:mwu]]

CAF flags or any vendor specific compile time flags are an automatic r-. Please assign fields by value as mentioned in comment 35. If you have any future need for this and can't find a good solution, let me know.

Comment 40

[Michael Wu [:mwu]]

(In reply to Thomas Zimmermann [:tzimmermann] [:tdz] from comment #35)
> Simply assign the fields by value, like this
> 
> static struct s {
>   .field = FieldFunc
> };
> 

Unfortunately designated initializers aren't supported in c++, but the equivalent can still be done. Just without the nice syntax.

Comment 41

[Michael Vines [:m1] [:evilmachines]]

Bhargav is checking comment 38 over here.  If this is causing issues we will change the CAF HAL for B2G to avoid the need for a Gecko CAF flag

Comment 42

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Michael Wu [:mwu] from comment #39)
> CAF flags or any vendor specific compile time flags are an automatic r-.
> Please assign fields by value as mentioned in comment 35. If you have any
> future need for this and can't find a good solution, let me know.
Thanks mwu. It's clear that we will talk to :m1 if there is any non-AOSP interface. :)

Comment 43

[bhargavg1]

(In reply to Michael Vines [:m1] [:evilmachines] from comment #41)
> Bhargav is checking comment 38 over here.  If this is causing issues we will
> change the CAF HAL for B2G to avoid the need for a Gecko CAF flag

Shawn, our internal test teams dont see any issue while testing the STR mentioned in comment 0

Let me loop back to you on the changes needed for CAF

Comment 44

[bhargavg1]

Moving ni to Vasanth who is working on this feature

Comment 45

[vasanth]

Sorry for the confusion. 
I see this issue is reproducible from our side as well. 
Based on that what we need to do?

Shawn,
I guess you are adding support for A2dpConnectionPriorityCallback with a Gecko CAF flag?
That should fix this issue right?
Then why do we need Q_BLUETOOTH flag? 
If needed where to add it?

Comment 46

[Michael Vines [:m1] [:evilmachines]]

(In reply to vasanth from comment #45)
> I guess you are adding support for A2dpConnectionPriorityCallback with a
> Gecko CAF flag?
> That should fix this issue right?
> Then why do we need Q_BLUETOOTH flag? 
> If needed where to add it?

A2dpConnectionPriorityCallback should be wrapped with a Q_BLUETOOTH flag so that adding a CAF flag to Gecko can be avoided.

Comment 47

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Per Comment 46 answer, cancel ni flag.

Comment 48

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Any target date for this bug? Which branch will be fixed this issue?

Comment 49

[bhargavg1]

(In reply to Shawn Huang [:shawnjohnjr] from comment #48)
> Any target date for this bug? Which branch will be fixed this issue?

for v2.0 the change would land @https://www.codeaurora.org/cgit/external/gigabyte/platform/external/bluetooth/bluedroid/log/?h=caf/b2g_kk_3.5

Change is merged internally it should show up externally in a day/two

Comment 50

[bhargavg1]

(In reply to Shawn Huang [:shawnjohnjr] from comment #48)
> Any target date for this bug? Which branch will be fixed this issue?

Change is available at [1]  for v2.0

[1] https://www.codeaurora.org/cgit/external/gigabyte/platform/external/bluetooth/bluedroid/commit/?h=caf/b2g_kk_3.5&id=5a71145c850dda353b7d93805f53a732f15febdf

Comment 51

[bhargavg1]

Fix is available at [1]  for v2.0

for v2.1 would be available [2], wait for the changes to land externally

[1] https://www.codeaurora.org/cgit/external/gigabyte/platform/external/bluetooth/bluedroid/commit/?h=caf/b2g_kk_3.5&id=5a71145c850dda353b7d93805f53a732f15febdf 

[2] https://www.codeaurora.org/cgit/quic/la/platform/external/bluetooth/bluedroid/log/?h=LNX.LF.3.5.1_RB1.2

Comment 52

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

How long usually will this patch set be merged take?

Comment 53

[bhargavg1]

(In reply to Shawn Huang [:shawnjohnjr] from comment #52)
> How long usually will this patch set be merged take?

v2.1 Changes at, 

platform/external/bluetooth/bluedroid - https://www.codeaurora.org/cgit/quic/la/platform/external/bluetooth/bluedroid/commit/?h=LNX.LF.3.5.1_RB1.2&id=949f774c77d25185d664ab4e7d8abf7fc78fe94a

platform/hardware/libhardware - https://www.codeaurora.org/cgit/quic/la/platform/hardware/libhardware/commit/?h=LNX.LF.3.5.1_RB1.2&id=dbbce4d60675b973aa6d40a07994c578e2067082