fix gattc iOS ANCS issues #453

zhongzhijie1 · 2026-02-06T07:07:39Z

fix address translation issue and improved subscribe API return value handling.

bug: v/83024

rootcause:
This will result in the upper layer not receiving the GATTC notify.
The get_le_addr_from_conn interface correctly retrieves the RPA address for the callback,
not the identity address.

The EALREADY return value indicates that ccc has already been subscribed to, so it is not an error.

Fix whitelist API typo, use the correct addr type rules.

bug: v/85615

See adapter_service.c. When adding/removing from the whitelist,
the unknown type now is treated as public, and the random type is passed directly.
So the addr_type is rely on service device list, call adapter_get_le_remote_address_type retrive it.

The stack also needs a patch to convert random type addresses into identity addresses.

Signed-off-by: zhongzhijie1 zhongzhijie1@xiaomi.com

gzh-terry · 2026-02-06T07:38:34Z

service/stacks/zephyr/sal_gatt_client_interface.c

        }

-        if (err) {
+        if (err && err != -EALREADY) {


Call if_gattc_on_element_changed if -EALREADY

I guess you means if_gattc_on_element_subscribed

gzh-terry · 2026-02-06T07:40:23Z

service/stacks/zephyr/sal_gatt_client_interface.c

        }

-        if (err) {
+        if (err && err != -EALREADY) {


Call if_gattc_on_element_changed if -EALREADY

huangyulong3 · 2026-02-06T09:18:15Z

service/stacks/zephyr/sal_gatt_client_interface.c

-    struct bt_conn_info info;
    bt_address_t addr;

+    get_le_addr_from_conn(conn, &addr);


All APIs should start with the prefix bt_.

Not in scope of this PR discussion; this is not a newly introduced API.

huangyulong3 · 2026-02-06T09:26:28Z

service/stacks/zephyr/sal_adapter_le_interface.c


-    addr.type = req->addr_type;
+    if (req->addr_type == BT_LE_ADDR_TYPE_UNKNOWN) {
+        addr.type = adapter_get_le_remote_address_type(&req->addr);


Where does we store the addr_type?

(adapter_service.c) service device list will sotre it when call add white list API

chengkai15 · 2026-02-09T06:28:56Z

service/stacks/zephyr/sal_gatt_client_interface.c

        return false;
    }

+    bt_uuid_to_uuid128(&uuid, u2);


why only support 128 bit uuid?

not only support 128, we can handle all type, but like old stack, all uuid from stack will convert to 128 Bits
and the log print is like

16Bits uuid like [0x00000000-0000-0000-0000-00000000xxxx]
now we show it as 0000xxxx-0000-1000-8000-00805F9B34FB.

So I change it to 128 to print pretty

It's a significant drawback if uuid16 must be handled as uuid128.
If we need to correct the printing in bttool, we can just go with #414 (see changes in gatt_client.c and bt_uuid.c).

this commit can drop, not necessary

chengkai15 · 2026-02-09T06:32:47Z

service/stacks/zephyr/sal_gatt_client_interface.c

        }

-        if (err) {
+        if (err && err != -EALREADY) {


Suggest return errno status code to app

now cannot do it , no helper convert gatt err to gatt_service error status

chengkai15 · 2026-02-09T06:34:26Z

service/stacks/zephyr/sal_adapter_le_interface.c

    int err;

-    addr.type = req->addr_type;
+    if (req->addr_type == BT_LE_ADDR_TYPE_UNKNOWN) {


suggest check address type， and suggest app need to past address type

(adapter_service.c) service device list will sotre it when call add white list API

now seem unkown as public in adapter_service.c, and old stack will firstly find it in device lists to match addr type, so use adapter_service.c device lists instead when use zblue

chengkai15 · 2026-02-10T04:22:50Z

service/stacks/zephyr/sal_gatt_client_interface.c

        return false;
    }

+    bt_uuid_to_uuid128(&uuid, u2);


suggest remove.

here just does uuid copy， but not uuid—>uuid128

remove done

chengkai15 · 2026-02-10T04:25:17Z

service/stacks/zephyr/sal_adapter_le_interface.c


-    addr.type = req->addr_type;
+    if (req->addr_type == BT_LE_ADDR_TYPE_UNKNOWN) {
+        addr.type = adapter_get_le_remote_address_type(&req->addr);


suggest add error handle for get address type fail

github-actions · 2026-02-11T08:39:18Z

❌ CLA Signature Required

@zhongzhijie1 Some contributors need to sign the CLA:

zhongzhijie1@xiaomi.com ❌ Needs to sign CLA

Please:

Sign the CLA at: https://www.openvela.com/#/community/cla
After signing, comment /check-cla to recheck

📋 View detailed check results: Action Run #21898142179

💡 Tip: All contributors must sign the CLA before the PR can be merged.

gzh-terry · 2026-02-11T08:56:25Z

service/stacks/zephyr/sal_gatt_client_interface.c

+        if (err && err != -EALREADY) {
            BT_LOGE("%s, %s NOTIFY failed, err:%d", __func__,
                enable ? "subscribe" : "unsubscribe", err);
+            if_gattc_on_element_subscribed(addr, element_id, BT_STATUS_FAIL, enable);


BT_STATUS_FAIL -> GATT_STATUS_FAILURE

gzh-terry · 2026-02-11T08:56:36Z

service/stacks/zephyr/sal_gatt_client_interface.c

        }
+
+        if (err == -EALREADY) {
+            if_gattc_on_element_subscribed(addr, element_id, BT_STATUS_SUCCESS, enable);


BT_STATUS_SUCCESS -> GATT_STATUS_SUCCESS

Same for other if_gattc_on_xxxxx

chengkai15 · 2026-02-11T09:08:51Z

@zhongzhijie1 please fix ci break

gzh-terry · 2026-02-11T12:13:51Z

service/stacks/zephyr/sal_adapter_le_interface.c

-    memcpy(&addr.a, &req->addr, sizeof(addr.a));
+    if (adapter_get_remote_identity_address(&req->addr, &id_addr) == BT_STATUS_SUCCESS) {
+        memcpy(&addr.a, &id_addr, sizeof(addr.a));
+        addr.type = BT_LE_ADDR_TYPE_PUBLIC;


Note: identity address can be random.

Add TODO: or FIXME:

now we let static random addr go to this branch

Copilot

Pull request overview

Fixes BLE iOS ANCS/GATTC address translation so upper layers receive notifications correctly, and adjusts LE whitelist + GATT subscribe handling to better align with identity/RPA behavior and non-fatal return codes.

Changes:

Use get_le_addr_from_conn() in GATT client callbacks to report the resolved/identity address rather than the raw connection destination.
Treat -EALREADY from subscribe/indicate registration as non-fatal and propagate a successful subscribe result to the upper layer.
Update LE whitelist add/remove to prefer identity addresses and correct the remove path to call the proper Zephyr API.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 9 comments.

File	Description
service/stacks/zephyr/sal_gatt_client_interface.c	Updates GATT client callbacks to use resolved LE address + refines subscribe error handling.
service/stacks/zephyr/sal_adapter_le_interface.c	Adjusts whitelist add/remove to use identity address where possible and fixes remove API usage.

Comments suppressed due to low confidence (1)

service/stacks/zephyr/sal_gatt_client_interface.c:939

The return value of get_le_addr_from_conn() is ignored. If address resolution fails, notifications/indications will be reported with an invalid address. Check the return value and stop iteration on failure.

    get_le_addr_from_conn(conn, &addr);

    handle = params->value_handle;
    if (data == NULL) {
        BT_LOGE("[UNSUBSCRIBED] 0x%04X", params->value_handle);
        return BT_GATT_ITER_STOP;
    }

    if_gattc_on_element_changed(&addr, handle, (uint8_t*)data, length);
    return BT_GATT_ITER_CONTINUE;
}

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-02-11T12:18:42Z

service/stacks/zephyr/sal_gatt_client_interface.c

    get_le_addr_from_conn(conn, &addr);
-    if_gattc_on_mtu_changed(&addr, att_payload, BT_STATUS_SUCCESS);
+    if_gattc_on_mtu_changed(&addr, att_payload, GATT_STATUS_SUCCESS);
 }


The return value of get_le_addr_from_conn() is ignored. If it fails, if_gattc_on_mtu_changed() will be called with an undefined address. Check the return value and return early on failure.

Copilot · 2026-02-11T12:18:42Z

service/stacks/zephyr/sal_adapter_le_interface.c

-    memcpy(&addr.a, &req->addr, sizeof(addr.a));
+    if (adapter_get_remote_identity_address(&req->addr, &id_addr) == BT_STATUS_SUCCESS) {
+        memcpy(&addr.a, &id_addr, sizeof(addr.a));
+        addr.type = BT_LE_ADDR_TYPE_PUBLIC;


When an identity address is found, this forces addr.type = BT_LE_ADDR_TYPE_PUBLIC. BLE identity addresses can be public or random static (the service layer comment in adapter_service.c notes identity addr type may be public or random). Use the stored remote address type (e.g., adapter_get_le_remote_address_type(&req->addr)) or otherwise track identity type instead of hardcoding PUBLIC.

Suggested change

addr.type = BT_LE_ADDR_TYPE_PUBLIC;

addr.type = adapter_get_le_remote_address_type(&id_addr);

Copilot · 2026-02-11T12:18:42Z

service/stacks/zephyr/sal_adapter_le_interface.c

-    memcpy(&addr.a, &req->addr, sizeof(addr.a));
+    if (adapter_get_remote_identity_address(&req->addr, &id_addr) == BT_STATUS_SUCCESS) {
+        memcpy(&addr.a, &id_addr, sizeof(addr.a));
+        addr.type = BT_LE_ADDR_TYPE_PUBLIC;


Same issue as add: identity address type is hardcoded to BT_LE_ADDR_TYPE_PUBLIC, but identity addr type may be public or random. Use the device's stored address type (or track identity type) so bt_le_filter_accept_list_remove() uses the correct type.

Suggested change

addr.type = BT_LE_ADDR_TYPE_PUBLIC;

addr.type = adapter_get_le_remote_address_type(&id_addr);

Copilot · 2026-02-11T12:18:43Z

service/stacks/zephyr/sal_gatt_client_interface.c

+    get_le_addr_from_conn(conn, &addr);
+
    if (err) {
        BT_LOGE("%s, gatt read fail err:%d", __func__, err);
-        if_gattc_on_element_read(&addr, params->single.handle, (uint8_t*)data, length, BT_STATUS_FAIL);
+        if_gattc_on_element_read(&addr, params->single.handle, (uint8_t*)data, length, GATT_STATUS_FAILURE);
        return BT_GATT_ITER_STOP;


The return value of get_le_addr_from_conn() is ignored. If it fails, addr may be uninitialized and the upper-layer callback will receive a garbage address. Check for BT_STATUS_SUCCESS and stop/return early (and optionally log) before calling if_gattc_on_element_read().

Copilot · 2026-02-11T12:18:43Z

service/stacks/zephyr/sal_gatt_client_interface.c

+    get_le_addr_from_conn(conn, &addr);

    if (err) {
        BT_LOGE("%s, exchange MTU failed err: %u", __func__, err);
-        if_gattc_on_mtu_changed(&addr, 0, BT_STATUS_FAIL);
+        if_gattc_on_mtu_changed(&addr, 0, GATT_STATUS_FAILURE);
        return;


The return value of get_le_addr_from_conn() is ignored. If address lookup fails, if_gattc_on_mtu_changed() will be invoked with an invalid address. Check for BT_STATUS_SUCCESS and return early on failure.

Copilot · 2026-02-11T12:18:43Z

service/stacks/zephyr/sal_adapter_le_interface.c

+        addr.type = BT_LE_ADDR_TYPE_PUBLIC;
+    } else {
+        memcpy(&addr.a, &req->addr, sizeof(addr.a));
+        addr.type = adapter_get_le_remote_address_type(&req->addr);


In the no-identity fallback, adapter_get_le_remote_address_type() can return BT_LE_ADDR_TYPE_UNKNOWN when the device isn't in the adapter device list. Passing UNKNOWN to bt_le_filter_accept_list_add() is likely invalid, and the PR description says UNKNOWN should be treated as PUBLIC. Add an explicit mapping for UNKNOWN -> BT_LE_ADDR_TYPE_PUBLIC here.

Suggested change

addr.type = adapter_get_le_remote_address_type(&req->addr);

addr.type = adapter_get_le_remote_address_type(&req->addr);

if (addr.type == BT_LE_ADDR_TYPE_UNKNOWN) {

addr.type = BT_LE_ADDR_TYPE_PUBLIC;

}

service/stacks/zephyr/sal_adapter_le_interface.c

service/stacks/zephyr/sal_gatt_client_interface.c

gzh-terry · 2026-02-11T12:21:10Z

service/stacks/zephyr/sal_adapter_le_interface.c

-    memcpy(&addr.a, &req->addr, sizeof(addr.a));
+    if (adapter_get_remote_identity_address(&req->addr, &id_addr) == BT_STATUS_SUCCESS) {
+        memcpy(&addr.a, &id_addr, sizeof(addr.a));
+        addr.type = BT_LE_ADDR_TYPE_PUBLIC;


/** TODO: consider random (static) identity address */

done, add it

… handling. bug: v/83024 rootcause: This will result in the upper layer not receiving the GATTC notify. The `get_le_addr_from_conn` interface correctly retrieves the RPA address for the callback, not the identity address. The EALREADY return value indicates that ccc has already been subscribed to, so it is not an error. Signed-off-by: zhongzhijie1 <zhongzhijie1@xiaomi.com>

bug: v/85615 See adapter_service.c. When adding/removing from the whitelist, the unknown type is treated as public, and the random type is passed directly. The stack also needs a patch to convert random type addresses into identity addresses. Signed-off-by: zhongzhijie1 <zhongzhijie1@xiaomi.com>

zhongzhijie1 requested review from chengkai15, gzh-terry, huangyulong3, hyson710 and zhangyuan376 as code owners February 6, 2026 07:07

gzh-terry reviewed Feb 6, 2026

View reviewed changes

huangyulong3 reviewed Feb 6, 2026

View reviewed changes

chengkai15 reviewed Feb 9, 2026

View reviewed changes

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from e7c573c to e8794dc Compare February 9, 2026 13:26

chengkai15 reviewed Feb 10, 2026

View reviewed changes

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from e8794dc to 58b1597 Compare February 11, 2026 08:38

zhongzhijie1 requested review from chengkai15, gzh-terry and huangyulong3 February 11, 2026 08:39

huangyulong3 approved these changes Feb 11, 2026

View reviewed changes

huangyulong3 previously approved these changes Feb 11, 2026

View reviewed changes

gzh-terry reviewed Feb 11, 2026

View reviewed changes

chengkai15 previously approved these changes Feb 11, 2026

View reviewed changes

zhongzhijie1 dismissed stale reviews from chengkai15 and huangyulong3 via 3ae5674 February 11, 2026 09:38

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from 58b1597 to 3ae5674 Compare February 11, 2026 09:38

zhongzhijie1 requested a review from huangyulong3 February 11, 2026 09:40

zhongzhijie1 requested review from chengkai15 and gzh-terry February 11, 2026 09:40

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from 3ae5674 to 1d88ee6 Compare February 11, 2026 10:04

chengkai15 previously approved these changes Feb 11, 2026

View reviewed changes

gzh-terry previously approved these changes Feb 11, 2026

View reviewed changes

zhongzhijie1 dismissed stale reviews from gzh-terry and chengkai15 via 7fe810a February 11, 2026 11:55

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from 1d88ee6 to 7fe810a Compare February 11, 2026 11:55

zhongzhijie1 requested review from chengkai15 and gzh-terry February 11, 2026 11:59

gzh-terry reviewed Feb 11, 2026

View reviewed changes

gzh-terry previously approved these changes Feb 11, 2026

View reviewed changes

gzh-terry requested a review from Copilot February 11, 2026 12:15

Copilot started reviewing on behalf of gzh-terry February 11, 2026 12:15 View session

Copilot AI reviewed Feb 11, 2026

View reviewed changes

gzh-terry reviewed Feb 11, 2026

View reviewed changes

zhongzhijie1 dismissed gzh-terry’s stale review via e4eea5c February 11, 2026 12:23

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from 7fe810a to e4eea5c Compare February 11, 2026 12:23

gzh-terry previously approved these changes Feb 11, 2026

View reviewed changes

chengkai15 previously approved these changes Feb 11, 2026

View reviewed changes

zhongzhijie1 added 2 commits February 11, 2026 22:17

zhongzhijie1 dismissed stale reviews from chengkai15 and gzh-terry via b65504c February 11, 2026 14:17

zhongzhijie1 force-pushed the gatt_client_fix_version2 branch from e4eea5c to b65504c Compare February 11, 2026 14:17

gzh-terry approved these changes Feb 11, 2026

View reviewed changes

huangyulong3 approved these changes Feb 11, 2026

View reviewed changes

chengkai15 approved these changes Feb 11, 2026

View reviewed changes

	addr.type = BT_LE_ADDR_TYPE_PUBLIC;
	addr.type = adapter_get_le_remote_address_type(&id_addr);

fix gattc iOS ANCS issues #453

Are you sure you want to change the base?

fix gattc iOS ANCS issues #453

Conversation

zhongzhijie1 commented Feb 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

chengkai15 Feb 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 11, 2026

❌ CLA Signature Required

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

chengkai15 commented Feb 11, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 11, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 11, 2026

Choose a reason for hiding this comment

zhongzhijie1 commented Feb 6, 2026 •

edited

Loading

chengkai15 Feb 10, 2026 •

edited

Loading