| Message ID | 20200910060403.144524-1-josephsih@chromium.org |
|---|---|
| Headers | show |
| Series | To support the HFP WBS, a chip vendor may choose a particular | expand |
On Thu, Sep 10, 2020 at 4:18 PM Pali Rohár <pali@kernel.org> wrote: > > On Thursday 10 September 2020 14:04:01 Joseph Hwang wrote: > > It is desirable to define the HCI packet payload sizes of > > USB alternate settings so that they can be exposed to user > > space. > > > > Reviewed-by: Alain Michaud <alainm@chromium.org> > > Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@chromium.org> > > Signed-off-by: Joseph Hwang <josephsih@chromium.org> > > --- > > > > Changes in v3: > > - Set hdev->sco_mtu to rp->sco_mtu if the latter is smaller. > > > > Changes in v2: > > - Used sco_mtu instead of a new sco_pkt_len member in hdev. > > - Do not overwrite hdev->sco_mtu in hci_cc_read_buffer_size > > if it has been set in the USB interface. > > > > drivers/bluetooth/btusb.c | 45 +++++++++++++++++++++++++++++---------- > > net/bluetooth/hci_event.c | 14 +++++++++++- > > 2 files changed, 47 insertions(+), 12 deletions(-) > > > > diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c > > index fe80588c7bd3a8..651d5731a6c6cf 100644 > > --- a/drivers/bluetooth/btusb.c > > +++ b/drivers/bluetooth/btusb.c > > @@ -459,6 +459,24 @@ static const struct dmi_system_id btusb_needs_reset_resume_table[] = { > > #define BTUSB_WAKEUP_DISABLE 14 > > #define BTUSB_USE_ALT1_FOR_WBS 15 > > > > +/* Per core spec 5, vol 4, part B, table 2.1, > > + * list the hci packet payload sizes for various ALT settings. > > + * This is used to set the packet length for the wideband speech. > > + * If a controller does not probe its usb alt setting, the default > > + * value will be 0. Any clients at upper layers should interpret it > > + * as a default value and set a proper packet length accordingly. > > + * > > + * To calculate the HCI packet payload length: > > + * for alternate settings 1 - 5: > > + * hci_packet_size = suggested_max_packet_size * 3 (packets) - > > + * 3 (HCI header octets) > > + * for alternate setting 6: > > + * hci_packet_size = suggested_max_packet_size - 3 (HCI header octets) > > + * where suggested_max_packet_size is {9, 17, 25, 33, 49, 63} > > + * for alt settings 1 - 6. > > Thank you for update, now I see what you mean! > > > + */ > > +static const int hci_packet_size_usb_alt[] = { 0, 24, 48, 72, 96, 144, 60 }; > > Now the another question, why you are using hci_packet_size_usb_alt[1] > and hci_packet_size_usb_alt[6] values from this array? Will answer it per the spec in the next patch series. > > > + > > struct btusb_data { > > struct hci_dev *hdev; > > struct usb_device *udev; > > @@ -3959,6 +3977,15 @@ static int btusb_probe(struct usb_interface *intf, > > hdev->notify = btusb_notify; > > hdev->prevent_wake = btusb_prevent_wake; > > > > + if (id->driver_info & BTUSB_AMP) { > > + /* AMP controllers do not support SCO packets */ > > + data->isoc = NULL; > > + } else { > > + /* Interface orders are hardcoded in the specification */ > > + data->isoc = usb_ifnum_to_if(data->udev, ifnum_base + 1); > > + data->isoc_ifnum = ifnum_base + 1; > > + } > > + > > #ifdef CONFIG_PM > > err = btusb_config_oob_wake(hdev); > > if (err) > > @@ -4022,6 +4049,10 @@ static int btusb_probe(struct usb_interface *intf, > > hdev->set_diag = btintel_set_diag; > > hdev->set_bdaddr = btintel_set_bdaddr; > > hdev->cmd_timeout = btusb_intel_cmd_timeout; > > + > > + if (btusb_find_altsetting(data, 6)) > > + hdev->sco_mtu = hci_packet_size_usb_alt[6]; > > Why you are setting this sco_mtu only for Intel adapter? Is not this > whole code generic to USB? Please refer to the answer to the Realtek adapter below. Thanks. > > > + > > set_bit(HCI_QUIRK_STRICT_DUPLICATE_FILTER, &hdev->quirks); > > set_bit(HCI_QUIRK_SIMULTANEOUS_DISCOVERY, &hdev->quirks); > > set_bit(HCI_QUIRK_NON_PERSISTENT_DIAG, &hdev->quirks); > > @@ -4063,15 +4094,6 @@ static int btusb_probe(struct usb_interface *intf, > > btusb_check_needs_reset_resume(intf); > > } > > > > - if (id->driver_info & BTUSB_AMP) { > > - /* AMP controllers do not support SCO packets */ > > - data->isoc = NULL; > > - } else { > > - /* Interface orders are hardcoded in the specification */ > > - data->isoc = usb_ifnum_to_if(data->udev, ifnum_base + 1); > > - data->isoc_ifnum = ifnum_base + 1; > > - } > > - > > if (IS_ENABLED(CONFIG_BT_HCIBTUSB_RTL) && > > (id->driver_info & BTUSB_REALTEK)) { > > hdev->setup = btrtl_setup_realtek; > > @@ -4083,9 +4105,10 @@ static int btusb_probe(struct usb_interface *intf, > > * (DEVICE_REMOTE_WAKEUP) > > */ > > set_bit(BTUSB_WAKEUP_DISABLE, &data->flags); > > - if (btusb_find_altsetting(data, 1)) > > + if (btusb_find_altsetting(data, 1)) { > > set_bit(BTUSB_USE_ALT1_FOR_WBS, &data->flags); > > - else > > + hdev->sco_mtu = hci_packet_size_usb_alt[1]; > > And this part of code which you write is Realtek specific. We currently only have Intel and Realtek platforms to test with. If making it generic without proper testing platforms is fine, I will make it generic. Or do you think it might be better to make it customized with particular vendors for now; and make it generic later when it works well with sufficient vendors? > > I thought that this is something generic to bluetooth usb as you pointed > to bluetooth documentation "core spec 5, vol 4, part B, table 2.1". > > > + } else > > bt_dev_err(hdev, "Device does not support ALT setting 1"); > > } > > Also this patch seems to be for me incomplete or not fully correct as > USB altsetting is chosen in function btusb_work() and it depends on > selected AIR mode (which is configured by another setsockopt). > > So despite what is written in commit message, this patch looks for me > like some hack for Intel and Realtek bluetooth adapters and does not > solve problems in vendor independent manner. You are right that sco_mtu should be changed according to the air mode. Here are some issues to handle and what I plan to do. I would like to solicit your comments before I submit the next series. [Issue 1] The air mode is determined in btusb_work() which is triggered by hci_sync_conn_complete_evt(). So “conn->mtu = hdev->sco_mtu” should not be done in sco_conn_add() in the early connecting stage. Instead, it will be moved to near the end of hci_sync_conn_complete_evt(). [Issue 2] The btusb_work() is performed by a worker. There would be a timing issue here if we let btusb_work() to do “hdev->sco_mtu = hci_packet_size_usb_alt[i]” because there is no guarantee how soon the btusb_work() can be finished and get “hdev->sco_mtu” value set correctly. In order to avoid the potential race condition, I suggest to determine air_mode in btusb_notify() before schedule_work(&data->work) is executed so that “hdev->sco_mtu = hci_packet_size_usb_alt[i]” is guaranteed to be performed when btusb_notify() finished. In this way, hci_sync_conn_complete_evt() can set conn->mtu correctly as described in [Issue 1] above. [Issue 3] Concerning CVSD: The above flow is good when the transparent mode is selected. When it is the CVSD mode, we should set hdev->sco_mtu and conn->mtu back to the original mtu size returned by hci_cc_read_buffer_size(). This is because we do not have a reliable way to determine what size is used with CVSD. AFAIK, controllers connected through USB use 48 bytes; and controllers connected through UART use 60 bytes. It seems to me that these numbers are not recorded in the kernel(?). It seems beyond the scope of this patch to set the proper value for CVSD. So we will let hdev->sco_mtu and conn->mtu go back to their original values and are not affected by this patch. I am wondering if such directions of fixing this patch look good to you? Thanks and regards! Joseph -- Joseph Shyh-In Hwang Email: josephsih@google.com
Hello! On Monday 14 September 2020 20:18:27 Joseph Hwang wrote: > On Thu, Sep 10, 2020 at 4:18 PM Pali Rohár <pali@kernel.org> wrote: > > And this part of code which you write is Realtek specific. > > We currently only have Intel and Realtek platforms to test with. If > making it generic without proper testing platforms is fine, I will > make it generic. Or do you think it might be better to make it > customized with particular vendors for now; and make it generic later > when it works well with sufficient vendors? I understood that those packet size changes are generic to bluetooth specification and therefore it is not vendor specific code. Those packet sizes for me really seems to be USB specific. Therefore it should apply for all vendors, not only for Realtek and Intel. I would propose to fix it globally for all USB adapters and not only for Intel and Realtek USB adapters. > > > > I thought that this is something generic to bluetooth usb as you pointed > > to bluetooth documentation "core spec 5, vol 4, part B, table 2.1". > > > > > + } else > > > bt_dev_err(hdev, "Device does not support ALT setting 1"); > > > } > > > > Also this patch seems to be for me incomplete or not fully correct as > > USB altsetting is chosen in function btusb_work() and it depends on > > selected AIR mode (which is configured by another setsockopt). > > > > So despite what is written in commit message, this patch looks for me > > like some hack for Intel and Realtek bluetooth adapters and does not > > solve problems in vendor independent manner. > > You are right that sco_mtu should be changed according to the air > mode. Here are some issues to handle and what I plan to do. I would > like to solicit your comments before I submit the next series. > > [Issue 1] The air mode is determined in btusb_work() which is > triggered by hci_sync_conn_complete_evt(). So “conn->mtu = > hdev->sco_mtu” should not be done in sco_conn_add() in the early > connecting stage. Instead, it will be moved to near the end of > hci_sync_conn_complete_evt(). > > [Issue 2] The btusb_work() is performed by a worker. There would be a > timing issue here if we let btusb_work() to do “hdev->sco_mtu = > hci_packet_size_usb_alt[i]” because there is no guarantee how soon the > btusb_work() can be finished and get “hdev->sco_mtu” value set > correctly. In order to avoid the potential race condition, I suggest > to determine air_mode in btusb_notify() before > schedule_work(&data->work) is executed so that “hdev->sco_mtu = > hci_packet_size_usb_alt[i]” is guaranteed to be performed when > btusb_notify() finished. In this way, hci_sync_conn_complete_evt() can > set conn->mtu correctly as described in [Issue 1] above. If it really fixes this issue, I'm fine with it. You have Realtek and Intel HW for testing, so I think if doing some heavy tests did not trigger any issue / race condition then it should be fine. > [Issue 3] Concerning CVSD: The above flow is good when the transparent > mode is selected. When it is the CVSD mode, we should set > hdev->sco_mtu and conn->mtu back to the original mtu size returned by > hci_cc_read_buffer_size(). Beware that kernel does not support CVSD mode between (USB) bluetooth adapter and kernel yet. Therefore it does not make sense to discussion this option. MTU size affects packets between kernel (or userspace) and bluetooth adapter. Bluetooth adapter then do HW encoding/decoding if non-transparent mode is used. So I think you rather mean PCMs16 mode between bluetooth adapter and kernel because this mode is used for CVSD air codec (between two bluetooth adapters over the air). This is default mode, userspace application pass PCMs16 data to kernel which pass it over USB to bluetooth adapter. And bluetooth adapter then do HW encoding, encodes PCMs16 to CVSD codec and transmit it over the air. So over USB are sent PCM data, not CVSD. > This is because we do not have a reliable > way to determine what size is used with CVSD. AFAIK, controllers > connected through USB use 48 bytes; and controllers connected through > UART use 60 bytes. I think for USB adapters we have that size. From bluetooth specification I understood that it is that size which you calculated in that array based on alt interface. I understood that bluetooth specification that packet size on USB depends only on chosen alt interface. And choice of alt interface depends on voice codec (PCM or transparent) and on number of active transports. But correct me, if I'm wrong. Maybe I did not understand bluetooth specification correctly. And for UART I do not know, I have not looked at it it yet. > It seems to me that these numbers are not recorded > in the kernel(?). It seems beyond the scope of this patch to set the > proper value for CVSD. So we will let hdev->sco_mtu and conn->mtu go > back to their original values and are not affected by this patch. > > I am wondering if such directions of fixing this patch look good to you? > > Thanks and regards! > Joseph
On Wednesday, September 23, 2020 3:22:15 AM PST Pali Rohár wrote: > On Monday 14 September 2020 20:18:27 Joseph Hwang wrote: > > On Thu, Sep 10, 2020 at 4:18 PM Pali Rohár <pali@kernel.org> wrote: > > > And this part of code which you write is Realtek specific. > > > > We currently only have Intel and Realtek platforms to test with. If > > making it generic without proper testing platforms is fine, I will > > make it generic. Or do you think it might be better to make it > > customized with particular vendors for now; and make it generic later > > when it works well with sufficient vendors? > > I understood that those packet size changes are generic to bluetooth > specification and therefore it is not vendor specific code. Those packet > sizes for me really seems to be USB specific. > > Therefore it should apply for all vendors, not only for Realtek and > Intel. I have tried to test WBS with some different USB adapters. So far, all use these packet sizes. Tested were: Broadcom BRCM20702A Realtek RTL8167B Realtek RTL8821C CSR CSR8510 (probably fake) In all cases, WBS works best with packet size of (USB packet size for alt mode selected) * 3 packets - 3 bytes HCI header. None of these devices support alt 6 mode, where supposedly one packet is better, but I can find no BT adapter on which to test this. > +static const int hci_packet_size_usb_alt[] = { 0, 24, 48, 72, 96, 144, 60}; Note that the packet sizes here are based on the max isoc packet length for the USB alt mode used, e.g. alt 1 is 9 bytes. That value is only a "recommended" value from the bluetooth spec. It seems like it would be more correct use (btusb_data*)->isoc_tx_ep->wMaxPacketSize to find the MTU. > > [Issue 2] The btusb_work() is performed by a worker. There would be a > > timing issue here if we let btusb_work() to do “hdev->sco_mtu = > > hci_packet_size_usb_alt[i]” because there is no guarantee how soon the > > btusb_work() can be finished and get “hdev->sco_mtu” value set > > correctly. In order to avoid the potential race condition, I suggest > > to determine air_mode in btusb_notify() before > > schedule_work(&data->work) is executed so that “hdev->sco_mtu = > > hci_packet_size_usb_alt[i]” is guaranteed to be performed when > > btusb_notify() finished. In this way, hci_sync_conn_complete_evt() can > > set conn->mtu correctly as described in [Issue 1] above. Does this also give userspace a clear point at which to determine MTU setting, _before_ data is sent over SCO connection? It will not work if sco_mtu is not valid until after userspace sends data to SCO connection with incorrect mtu.
On Tuesday 08 December 2020 15:04:29 Trent Piepho wrote: > On Wednesday, September 23, 2020 3:22:15 AM PST Pali Rohár wrote: > > On Monday 14 September 2020 20:18:27 Joseph Hwang wrote: > > > On Thu, Sep 10, 2020 at 4:18 PM Pali Rohár <pali@kernel.org> wrote: > > > > And this part of code which you write is Realtek specific. > > > > > > We currently only have Intel and Realtek platforms to test with. If > > > making it generic without proper testing platforms is fine, I will > > > make it generic. Or do you think it might be better to make it > > > customized with particular vendors for now; and make it generic later > > > when it works well with sufficient vendors? > > > > I understood that those packet size changes are generic to bluetooth > > specification and therefore it is not vendor specific code. Those packet > > sizes for me really seems to be USB specific. > > > > Therefore it should apply for all vendors, not only for Realtek and > > Intel. > > I have tried to test WBS with some different USB adapters. So far, all use > these packet sizes. Tested were: > > Broadcom BRCM20702A > Realtek RTL8167B > Realtek RTL8821C > CSR CSR8510 (probably fake) > > In all cases, WBS works best with packet size of (USB packet size for alt mode > selected) * 3 packets - 3 bytes HCI header. None of these devices support alt > 6 mode, where supposedly one packet is better, but I can find no BT adapter on > which to test this. > > > +static const int hci_packet_size_usb_alt[] = { 0, 24, 48, 72, 96, 144, 60}; > > Note that the packet sizes here are based on the max isoc packet length for > the USB alt mode used, e.g. alt 1 is 9 bytes. That value is only a > "recommended" value from the bluetooth spec. It seems like it would be more > correct use (btusb_data*)->isoc_tx_ep->wMaxPacketSize to find the MTU. Yea, wMaxPacketSize looks like a candidate for determining MTU. Can we use it or are there any known issues with it? > > > [Issue 2] The btusb_work() is performed by a worker. There would be a > > > timing issue here if we let btusb_work() to do “hdev->sco_mtu = > > > hci_packet_size_usb_alt[i]” because there is no guarantee how soon the > > > btusb_work() can be finished and get “hdev->sco_mtu” value set > > > correctly. In order to avoid the potential race condition, I suggest > > > to determine air_mode in btusb_notify() before > > > schedule_work(&data->work) is executed so that “hdev->sco_mtu = > > > hci_packet_size_usb_alt[i]” is guaranteed to be performed when > > > btusb_notify() finished. In this way, hci_sync_conn_complete_evt() can > > > set conn->mtu correctly as described in [Issue 1] above. > > Does this also give userspace a clear point at which to determine MTU setting, > _before_ data is sent over SCO connection? It will not work if sco_mtu is not > valid until after userspace sends data to SCO connection with incorrect mtu. IIRC connection is established after sync connection (SCO) complete event. And sending data is possible after connection is established. So based on these facts I think that userspace can determinate MTU settings prior sending data over SCO socket. Anyway, to whole MTU issue for SCO there is a nice workaround which worked fine with more tested USB adapters and headsets. As SCO socket is synchronous and most bluetooth headsets have own clocks, you can synchronize sending packets to headsets based on time events when you received packets from other side and also send packets of same size as you received. I.e. for every received packet send own packet of the same size.
On Tuesday, December 8, 2020 5:13:36 PM PST Pali Rohár wrote: > On Tuesday 08 December 2020 15:04:29 Trent Piepho wrote: > > Does this also give userspace a clear point at which to determine MTU setting, > > _before_ data is sent over SCO connection? It will not work if sco_mtu is not > > valid until after userspace sends data to SCO connection with incorrect mtu. > > IIRC connection is established after sync connection (SCO) complete > event. And sending data is possible after connection is established. So > based on these facts I think that userspace can determinate MTU settings > prior sending data over SCO socket. > > Anyway, to whole MTU issue for SCO there is a nice workaround which > worked fine with more tested USB adapters and headsets. As SCO socket is > synchronous and most bluetooth headsets have own clocks, you can > synchronize sending packets to headsets based on time events when you > received packets from other side and also send packets of same size as > you received. I.e. for every received packet send own packet of the same > size. As I understand it, the RX side from the headset is not guaranteed, so in the TX only case this will not work and we still need to be told what MTU kernel has selected for the SCO link. It seems also it would add some latency to start up, since it would be necessary to wait for packets to arrive before knowing what size packet to send. Would timing based on matching TX to RX in the case of packet loss on RX side?
On Wednesday 09 December 2020 16:19:39 Trent Piepho wrote: > On Tuesday, December 8, 2020 5:13:36 PM PST Pali Rohár wrote: > > On Tuesday 08 December 2020 15:04:29 Trent Piepho wrote: > > > Does this also give userspace a clear point at which to determine MTU > setting, > > > _before_ data is sent over SCO connection? It will not work if sco_mtu > is not > > > valid until after userspace sends data to SCO connection with incorrect > mtu. > > > > IIRC connection is established after sync connection (SCO) complete > > event. And sending data is possible after connection is established. So > > based on these facts I think that userspace can determinate MTU settings > > prior sending data over SCO socket. > > > > Anyway, to whole MTU issue for SCO there is a nice workaround which > > worked fine with more tested USB adapters and headsets. As SCO socket is > > synchronous and most bluetooth headsets have own clocks, you can > > synchronize sending packets to headsets based on time events when you > > received packets from other side and also send packets of same size as > > you received. I.e. for every received packet send own packet of the same > > size. > > As I understand it, the RX side from the headset is not guaranteed, so in > the TX only case this will not work and we still need to be told what MTU > kernel has selected for the SCO link. I'm not sure if TX-only SCO link is possible. I always thought that SCO is synchronous bidirectional link. As I said, this "workaround" is useful for classic bluetooth headsets and is it possible to use it immediately without any kernel changes. And I agree that kernel should tell userspace correct MTU value. And this should be fixed. "Workaround" is useful for immediate action to deliver at least something which works with most bluetooth headsets. > It seems also it would add some latency to start up, since it would be > necessary to wait for packets to arrive before knowing what size packet to > send. I think this startup latency is negligible in HFP profile where start needs non-trivial exchange of AT commands. > Would timing based on matching TX to RX in the case of packet loss on RX > side? That is a good question for some research. I remember that e.g. pulseaudio used this technique for synchronizing bluetooth SCO RX and TX streams.