`esp_gatt_value_t` unnecessarily holds 512 byte value (IDFGH-12362)

[nebkat]

Answers checklist.

• I have read the documentation ESP-IDF Programming Guide and the issue is not addressed there.
• I have updated my IDF branch (master or release) to the latest version and checked that the issue is present there.
• I have searched the issue tracker for a similar issue and not found a similar issue.

General issue report

esp_gatt_value_t has a ESP_GATT_MAX_ATTR_LEN sized array to hold any potential response values.

esp-idf/components/bt/host/bluedroid/api/include/api/esp_gatt_defs.h

Lines 374 to 380 in 4f900ed

	typedef struct {
	uint8_t value[ESP_GATT_MAX_ATTR_LEN]; /*!< Gatt attribute value */
	uint16_t handle; /*!< Gatt attribute handle */
	uint16_t offset; /*!< Gatt attribute value offset */
	uint16_t len; /*!< Gatt attribute value length */
	uint8_t auth_req; /*!< Gatt authentication request */
	} esp_gatt_value_t;

This is unnecessary as in btc_gatts_arg_deep_copy this value gets deep copied anyway:

esp-idf/components/bt/host/bluedroid/btc/profile/std/gatt/btc_gatts.c

Lines 85 to 94 in 4f900ed

	case BTC_GATTS_ACT_SEND_RESPONSE: {
	if (src->send_rsp.rsp) {
	dst->send_rsp.rsp = (esp_gatt_rsp_t *) osi_malloc(sizeof(esp_gatt_rsp_t));
	if (dst->send_rsp.rsp) {
	memcpy(dst->send_rsp.rsp, src->send_rsp.rsp, sizeof(esp_gatt_rsp_t));
	} else {
	BTC_TRACE_ERROR("%s %d no mem\n", __func__, msg->act);
	}
	}
	break;

For esp_ble_gatts_send_response a pointer + length should be accepted rather than making a second copy.

[nebkat]

Also believe this was wrongly closed: #10206 - maximum value has always been 512. The 600 was copied over from the original Android implementation which later dropped it to the correct value as per Bluetooth spec.

[esp-zhp]

Hello, thank you for your feedback. I will make the necessary corrections.
ps:
BLUETOOTH CORE SPECIFICATION Version 5.4 | Vol 3, Part F page 1408

[Indastri]

Closing the issue as the user did not reply. If you will encounter any problems, feel free to reopen this one or create a new issues.

Cheers!

[nebkat]

@Indastri I presume this was an automated message - no reply pending here.

The ESP_GATT_MAX_ATTR_LEN issue was fixed in 20a780d, however the large stack value problem remains -esp_ble_gatts_send_response should ideally accept a pointer + size.

[Indastri]

Hi @nebkat, I will point my colleague towards this issue. @zhp0406

P.S. - the message was not automated. :)
Cheers!