2023-06-22 20:01:59

by Luiz Augusto von Dentz

[permalink] [raw]
Subject: [PATCH v4] Bluetooth: hci_event: Fix parsing of CIS Established Event

From: Luiz Augusto von Dentz <[email protected]>

The ISO Interval on CIS Established Event uses 1.25 ms slots:

BLUETOOTH CORE SPECIFICATION Version 5.3 | Vol 4, Part E
page 2304:

Time = N * 1.25 ms

In addition to that this always update the QoS settings based on CIS
Established Event.

Signed-off-by: Luiz Augusto von Dentz <[email protected]>
---
net/bluetooth/hci_event.c | 36 +++++++++++++++++++++++-------------
1 file changed, 23 insertions(+), 13 deletions(-)

diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index b1aefe4bb751..049aa7f6a7c5 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -6822,6 +6822,7 @@ static void hci_le_cis_estabilished_evt(struct hci_dev *hdev, void *data,
{
struct hci_evt_le_cis_established *ev = data;
struct hci_conn *conn;
+ struct bt_iso_qos *qos;
bool pending = false;
u16 handle = __le16_to_cpu(ev->handle);

@@ -6846,21 +6847,30 @@ static void hci_le_cis_estabilished_evt(struct hci_dev *hdev, void *data,

pending = test_and_clear_bit(HCI_CONN_CREATE_CIS, &conn->flags);

- if (conn->role == HCI_ROLE_SLAVE) {
- __le32 interval;
+ qos = &conn->iso_qos;

- memset(&interval, 0, sizeof(interval));
+ /* Convert ISO Interval (1.25 ms slots) to latency (ms) */
+ qos->ucast.in.latency = DIV_ROUND_CLOSEST(le16_to_cpu(ev->interval) *
+ 125, 100);
+ qos->ucast.out.latency = qos->ucast.in.latency;

- memcpy(&interval, ev->c_latency, sizeof(ev->c_latency));
- conn->iso_qos.ucast.in.interval = le32_to_cpu(interval);
- memcpy(&interval, ev->p_latency, sizeof(ev->p_latency));
- conn->iso_qos.ucast.out.interval = le32_to_cpu(interval);
- conn->iso_qos.ucast.in.latency = le16_to_cpu(ev->interval);
- conn->iso_qos.ucast.out.latency = le16_to_cpu(ev->interval);
- conn->iso_qos.ucast.in.sdu = le16_to_cpu(ev->c_mtu);
- conn->iso_qos.ucast.out.sdu = le16_to_cpu(ev->p_mtu);
- conn->iso_qos.ucast.in.phy = ev->c_phy;
- conn->iso_qos.ucast.out.phy = ev->p_phy;
+ switch (conn->role) {
+ case HCI_ROLE_SLAVE:
+ qos->ucast.in.interval = get_unaligned_le24(ev->c_latency);
+ qos->ucast.out.interval = get_unaligned_le24(ev->p_latency);
+ qos->ucast.in.sdu = le16_to_cpu(ev->c_mtu);
+ qos->ucast.out.sdu = le16_to_cpu(ev->p_mtu);
+ qos->ucast.in.phy = ev->c_phy;
+ qos->ucast.out.phy = ev->p_phy;
+ break;
+ case HCI_ROLE_MASTER:
+ qos->ucast.out.interval = get_unaligned_le24(ev->c_latency);
+ qos->ucast.in.interval = get_unaligned_le24(ev->p_latency);
+ qos->ucast.out.sdu = le16_to_cpu(ev->c_mtu);
+ qos->ucast.in.sdu = le16_to_cpu(ev->p_mtu);
+ qos->ucast.out.phy = ev->c_phy;
+ qos->ucast.in.phy = ev->p_phy;
+ break;
}

if (!ev->status) {
--
2.40.1



2023-06-22 20:31:31

by bluez.test.bot

[permalink] [raw]
Subject: RE: [v4] Bluetooth: hci_event: Fix parsing of CIS Established Event

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list.
This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=759615

---Test result---

Test Summary:
CheckPatch PASS 0.77 seconds
GitLint PASS 0.38 seconds
SubjectPrefix PASS 0.14 seconds
BuildKernel PASS 34.35 seconds
CheckAllWarning PASS 37.12 seconds
CheckSparse WARNING 42.75 seconds
CheckSmatch WARNING 114.46 seconds
BuildKernel32 PASS 33.19 seconds
TestRunnerSetup PASS 473.71 seconds
TestRunner_l2cap-tester PASS 18.21 seconds
TestRunner_iso-tester FAIL 25.78 seconds
TestRunner_bnep-tester PASS 6.07 seconds
TestRunner_mgmt-tester PASS 138.21 seconds
TestRunner_rfcomm-tester PASS 9.60 seconds
TestRunner_sco-tester PASS 8.90 seconds
TestRunner_ioctl-tester PASS 10.30 seconds
TestRunner_mesh-tester PASS 7.63 seconds
TestRunner_smp-tester PASS 8.79 seconds
TestRunner_userchan-tester PASS 6.43 seconds
IncrementalBuild PASS 31.64 seconds

Details
##############################
Test: CheckSparse - WARNING
Desc: Run sparse tool with linux kernel
Output:
net/bluetooth/hci_event.c: note: in included file (through include/net/bluetooth/hci_core.h):
##############################
Test: CheckSmatch - WARNING
Desc: Run smatch tool with source
Output:
net/bluetooth/hci_event.c: note: in included file (through include/net/bluetooth/hci_core.h):
##############################
Test: TestRunner_iso-tester - FAIL
Desc: Run iso-tester with test-runner
Output:
Total: 80, Passed: 75 (93.8%), Failed: 5, Not Run: 0

Failed Test Cases
ISO QoS 8_1_1 - Success Failed 1.142 seconds
ISO QoS 16_1_1 - Success Failed 0.227 seconds
ISO QoS 24_1_1 - Success Failed 0.228 seconds
ISO QoS 32_1_1 - Success Failed 0.224 seconds
ISO QoS 44_1_1 - Success Failed 0.243 seconds


---
Regards,
Linux Bluetooth

2023-06-23 17:55:49

by Pauli Virtanen

[permalink] [raw]
Subject: Re: [PATCH v4] Bluetooth: hci_event: Fix parsing of CIS Established Event

Hi Luiz,

to, 2023-06-22 kello 12:27 -0700, Luiz Augusto von Dentz kirjoitti:
> From: Luiz Augusto von Dentz <[email protected]>
>
> The ISO Interval on CIS Established Event uses 1.25 ms slots:
>
> BLUETOOTH CORE SPECIFICATION Version 5.3 | Vol 4, Part E
> page 2304:
>
> Time = N * 1.25 ms
>
> In addition to that this always update the QoS settings based on CIS
> Established Event.
>
> Signed-off-by: Luiz Augusto von Dentz <[email protected]>
> ---
> net/bluetooth/hci_event.c | 36 +++++++++++++++++++++++-------------
> 1 file changed, 23 insertions(+), 13 deletions(-)
>
> diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> index b1aefe4bb751..049aa7f6a7c5 100644
> --- a/net/bluetooth/hci_event.c
> +++ b/net/bluetooth/hci_event.c
> @@ -6822,6 +6822,7 @@ static void hci_le_cis_estabilished_evt(struct hci_dev *hdev, void *data,
> {
> struct hci_evt_le_cis_established *ev = data;
> struct hci_conn *conn;
> + struct bt_iso_qos *qos;
> bool pending = false;
> u16 handle = __le16_to_cpu(ev->handle);
>
> @@ -6846,21 +6847,30 @@ static void hci_le_cis_estabilished_evt(struct hci_dev *hdev, void *data,
>
> pending = test_and_clear_bit(HCI_CONN_CREATE_CIS, &conn->flags);
>
> - if (conn->role == HCI_ROLE_SLAVE) {
> - __le32 interval;
> + qos = &conn->iso_qos;
>
> - memset(&interval, 0, sizeof(interval));
> + /* Convert ISO Interval (1.25 ms slots) to latency (ms) */
> + qos->ucast.in.latency = DIV_ROUND_CLOSEST(le16_to_cpu(ev->interval) *
> + 125, 100);
> + qos->ucast.out.latency = qos->ucast.in.latency;

In Set CIG Parameters:
u16 ucast.latency = Max_Transport_Latency (ms)
u32 ucast.interval = SDU_Interval (us)

Here:
u16 ucast.latency = ISO_Interval (ms)
u32 ucast.interval = Transport_Latency (us)

Currently BlueZ seems to not account for this swapping of the meanings
of the QoS fields, eg. in client/player.c it has

num = (qos.ucast.out.latency * 1000 / qos.ucast.out.interval);
...
ts.it_value.tv_nsec = qos->ucast.out.latency * 1000000

and writes num packets in each interval. In the AX210 example it seems
this would give num == 0. I guess this and other places that use these
need to be updated.

Since field meanings anyway change, would 1.25ms unit for the
ISO_Interval be better than 1ms so user side doesn't need to know how
kernel rounds the number and undo that to get the actual value?

> - memcpy(&interval, ev->c_latency, sizeof(ev->c_latency));
> - conn->iso_qos.ucast.in.interval = le32_to_cpu(interval);
> - memcpy(&interval, ev->p_latency, sizeof(ev->p_latency));
> - conn->iso_qos.ucast.out.interval = le32_to_cpu(interval);
> - conn->iso_qos.ucast.in.latency = le16_to_cpu(ev->interval);
> - conn->iso_qos.ucast.out.latency = le16_to_cpu(ev->interval);
> - conn->iso_qos.ucast.in.sdu = le16_to_cpu(ev->c_mtu);
> - conn->iso_qos.ucast.out.sdu = le16_to_cpu(ev->p_mtu);
> - conn->iso_qos.ucast.in.phy = ev->c_phy;
> - conn->iso_qos.ucast.out.phy = ev->p_phy;
> + switch (conn->role) {
> + case HCI_ROLE_SLAVE:
> + qos->ucast.in.interval = get_unaligned_le24(ev->c_latency);
> + qos->ucast.out.interval = get_unaligned_le24(ev->p_latency);
> + qos->ucast.in.sdu = le16_to_cpu(ev->c_mtu);
> + qos->ucast.out.sdu = le16_to_cpu(ev->p_mtu);
> + qos->ucast.in.phy = ev->c_phy;
> + qos->ucast.out.phy = ev->p_phy;
> + break;
> + case HCI_ROLE_MASTER:
> + qos->ucast.out.interval = get_unaligned_le24(ev->c_latency);
> + qos->ucast.in.interval = get_unaligned_le24(ev->p_latency);
> + qos->ucast.out.sdu = le16_to_cpu(ev->c_mtu);
> + qos->ucast.in.sdu = le16_to_cpu(ev->p_mtu);
> + qos->ucast.out.phy = ev->c_phy;
> + qos->ucast.in.phy = ev->p_phy;
> + break;
> }
>
> if (!ev->status) {


2023-06-23 20:03:38

by Luiz Augusto von Dentz

[permalink] [raw]
Subject: Re: [PATCH v4] Bluetooth: hci_event: Fix parsing of CIS Established Event

Hi Pauli,

On Fri, Jun 23, 2023 at 10:37 AM Pauli Virtanen <[email protected]> wrote:
>
> Hi Luiz,
>
> to, 2023-06-22 kello 12:27 -0700, Luiz Augusto von Dentz kirjoitti:
> > From: Luiz Augusto von Dentz <[email protected]>
> >
> > The ISO Interval on CIS Established Event uses 1.25 ms slots:
> >
> > BLUETOOTH CORE SPECIFICATION Version 5.3 | Vol 4, Part E
> > page 2304:
> >
> > Time = N * 1.25 ms
> >
> > In addition to that this always update the QoS settings based on CIS
> > Established Event.
> >
> > Signed-off-by: Luiz Augusto von Dentz <[email protected]>
> > ---
> > net/bluetooth/hci_event.c | 36 +++++++++++++++++++++++-------------
> > 1 file changed, 23 insertions(+), 13 deletions(-)
> >
> > diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> > index b1aefe4bb751..049aa7f6a7c5 100644
> > --- a/net/bluetooth/hci_event.c
> > +++ b/net/bluetooth/hci_event.c
> > @@ -6822,6 +6822,7 @@ static void hci_le_cis_estabilished_evt(struct hci_dev *hdev, void *data,
> > {
> > struct hci_evt_le_cis_established *ev = data;
> > struct hci_conn *conn;
> > + struct bt_iso_qos *qos;
> > bool pending = false;
> > u16 handle = __le16_to_cpu(ev->handle);
> >
> > @@ -6846,21 +6847,30 @@ static void hci_le_cis_estabilished_evt(struct hci_dev *hdev, void *data,
> >
> > pending = test_and_clear_bit(HCI_CONN_CREATE_CIS, &conn->flags);
> >
> > - if (conn->role == HCI_ROLE_SLAVE) {
> > - __le32 interval;
> > + qos = &conn->iso_qos;
> >
> > - memset(&interval, 0, sizeof(interval));
> > + /* Convert ISO Interval (1.25 ms slots) to latency (ms) */
> > + qos->ucast.in.latency = DIV_ROUND_CLOSEST(le16_to_cpu(ev->interval) *
> > + 125, 100);
> > + qos->ucast.out.latency = qos->ucast.in.latency;
>
> In Set CIG Parameters:
> u16 ucast.latency = Max_Transport_Latency (ms)
> u32 ucast.interval = SDU_Interval (us)
>
> Here:
> u16 ucast.latency = ISO_Interval (ms)
> u32 ucast.interval = Transport_Latency (us)

Yeah, it sounds like we should keep the same logic as in Set CIG
Parameters, this event is sort of weird because it does change the
units for whatever reason, anyway I will swap these around so they are
the same, anyway with the latest code it does pass all the iso-tester
tests.

> Currently BlueZ seems to not account for this swapping of the meanings
> of the QoS fields, eg. in client/player.c it has
>
> num = (qos.ucast.out.latency * 1000 / qos.ucast.out.interval);
> ...
> ts.it_value.tv_nsec = qos->ucast.out.latency * 1000000
>
> and writes num packets in each interval. In the AX210 example it seems
> this would give num == 0. I guess this and other places that use these
> need to be updated.
>
> Since field meanings anyway change, would 1.25ms unit for the
> ISO_Interval be better than 1ms so user side doesn't need to know how
> kernel rounds the number and undo that to get the actual value?
>
> > - memcpy(&interval, ev->c_latency, sizeof(ev->c_latency));
> > - conn->iso_qos.ucast.in.interval = le32_to_cpu(interval);
> > - memcpy(&interval, ev->p_latency, sizeof(ev->p_latency));
> > - conn->iso_qos.ucast.out.interval = le32_to_cpu(interval);
> > - conn->iso_qos.ucast.in.latency = le16_to_cpu(ev->interval);
> > - conn->iso_qos.ucast.out.latency = le16_to_cpu(ev->interval);
> > - conn->iso_qos.ucast.in.sdu = le16_to_cpu(ev->c_mtu);
> > - conn->iso_qos.ucast.out.sdu = le16_to_cpu(ev->p_mtu);
> > - conn->iso_qos.ucast.in.phy = ev->c_phy;
> > - conn->iso_qos.ucast.out.phy = ev->p_phy;
> > + switch (conn->role) {
> > + case HCI_ROLE_SLAVE:
> > + qos->ucast.in.interval = get_unaligned_le24(ev->c_latency);
> > + qos->ucast.out.interval = get_unaligned_le24(ev->p_latency);
> > + qos->ucast.in.sdu = le16_to_cpu(ev->c_mtu);
> > + qos->ucast.out.sdu = le16_to_cpu(ev->p_mtu);
> > + qos->ucast.in.phy = ev->c_phy;
> > + qos->ucast.out.phy = ev->p_phy;
> > + break;
> > + case HCI_ROLE_MASTER:
> > + qos->ucast.out.interval = get_unaligned_le24(ev->c_latency);
> > + qos->ucast.in.interval = get_unaligned_le24(ev->p_latency);
> > + qos->ucast.out.sdu = le16_to_cpu(ev->c_mtu);
> > + qos->ucast.in.sdu = le16_to_cpu(ev->p_mtu);
> > + qos->ucast.out.phy = ev->c_phy;
> > + qos->ucast.in.phy = ev->p_phy;
> > + break;
> > }
> >
> > if (!ev->status) {
>


--
Luiz Augusto von Dentz