Unfortunately this is coupled to the Linux kernel Bluetooth stack since it is relying on the User Channel of the hcisocket which let's your application (NuttX in this case) have a virtual HCI device and temporarily provide it largely isolated control over it (you cannot scan from the host anymore for example). There is a good description of this feature in the patch that introduced it. https://marc.info/?l=linux-bluetooth&m=137757846631034&w=2
On Tue, Apr 25, 2023, 5:54 PM Petro Karashchenko < petro.karashche...@gmail.com> wrote: > Is this going to work on OSX? Or Linux only is verified? > > вт, 25 квіт. 2023 р. о 19:45 Brennan Ashton <bash...@brennanashton.com> > пише: > > > On Tue, Apr 25, 2023, 5:41 PM Brennan Ashton <bash...@brennanashton.com> > > wrote: > > > > > > > > On Tue, Apr 25, 2023, 5:36 PM Petro Karashchenko < > > > petro.karashche...@gmail.com> wrote: > > > > > >> Hello, > > >> > > >> I'm starting to experiment with NuttX BLE support. I have a device > that > > I > > >> can successfully pair with my phone and I would like to try pairing it > > >> with > > >> the NuttX based device. I chose ESP32 for my experiments. For now I > > >> enabled > > >> (and fixed) esp32-devkitc:ble configuration, but that is basically > > all... > > >> I > > >> do not have much knowledge in bluetooth, but definitely want to master > > it. > > >> When I'm pairing from my phone I see the device name in the list of > > >> scanned > > >> devices, so the device advertises some human readable string in data. > > Then > > >> I'm running "bt bnep0 scan start" -> "bt bnep0 scan stop" -> "bt bnep0 > > >> scan > > >> get" on my ESP32 and expecting to see some of the same human readable > > >> string in the printed output as I do not known the MAC address of my > > >> device, but there is not such data sequence in "bt bnep0 scan get" > > output. > > >> I would appreciate some guidance for the BLE dummy like me, so I can > > move > > >> forward a bit. > > >> > > >> Best regards, > > >> Petro > > >> > > > > > > > > > I have not touched it in a bit, but I found the sim support I added > with > > > the hci socket made debugging the Bluetooth stack a lot easier because > I > > > could use Wireshark/btmon on the actual host machine to show all the > hci > > > traffic to and from the NuttX stack transparently. I highly recommend > > > trying that to verify there is not anything fundamentally broken. > > > > > > --Brennan > > > > > > > I should probably have added this in more detail to the docs, but see the > > testing section of the PR where I added it. I think it shows exactly what > > you are trying to do > > > > https://github.com/apache/nuttx/pull/1655 > > > > > > --Brennan > > >
