Is it possible to get a remote HCI error from g_dbus_connection_call_sync()?

[jaklug (John Klug)]

What I did was connect to a Bluetooth device:

g_dbus_connection_call_sync()

This resulted in the following HCI error:

> HCI Event: Command Status (0x0f) plen 4                #1997 [hci0] 60.389250
        LE Create Connection (0x08|0x000d) ncmd 1
          Status: Memory Capacity Exceeded (0x07)

I found the following error in the error parameter of g_dbus_connection_call_sync() :

error.message:
GDBus.Error:org.bluez.Error.Failed: Software caused connection abort

I then tried:

 if(g_dbus_error_is_remote_error(error)) {

This turned out to be true.

I then did:

msg = g_dbus_error_get_remote_error(error);

But msg was NULL.

Is there anything I can do to get a more specific error?

[fmuellner (Florian Müllner)]

@jaklug

msg = g_dbus_error_get_remote_error(error);

But msg was NULL.

NULL is unexpected, but that doesn’t really matter here: The expected result is org.bluez.Error (read: the error name returned by the service), which doesn’t tell you anything you don’t know already.

TIL that the D-Bus protocol allows passing arbitrary arguments in messages of type ERROR. If bluez uses that, then you can use the lower-level API to examine the arguments:

  message = g_dbus_connection_send_message_with_reply_sync ()

  if (g_dbus_message_get_message_type (message) == G_DBUS_MESSAGE_TYPE_ERROR)
    {
      GVariant *body = g_dbus_message_get_body (message);
     // ...
    }

But that’s a big if, it’s much more likely than not that the generic “Software caused connection abort” is all bluez is giving you. If there was a more detailed error, I’d expect a more specific error name, so callers don’t have to jump through hoops to get to the details.

[jaklug (John Klug)]

Florian Müllner wrote:

TIL that the D-Bus protocol allows passing arbitrary arguments in messages of type ERROR. If bluez uses that, then you can use the lower-level API to examine the arguments:

I am using the org.bluez.Device1 D-Bus API:

https://git.kernel.org/pub/scm/bluetooth/bluez.git/tree/doc/device-api.txt

I do not see any error passing interface listed.

Thanks for your help.

John Klug