| Message ID | d159c57f-8490-4c26-79da-6ad3612c4a14@salutedevices.com |
|---|---|
| State | New |
| Headers | show |
| Series | [v2] Bluetooth: hci_uart: fix race during initialization | expand |
Hi, sorry i'm new in bluetooth subsystem. I get the following message from CI: https://patchwork.kernel.org/project/bluetooth/list/?series=918554 Where one of tests was failed. Where I can get more information about failure: tedd_an/TestRunner_mgmt-tester fail TestRunner_mgmt-tester: Total: 490, Passed: 483 (98.6%), Failed: 3, Not Run: 4 ? Thanks On 17.12.2024 11:12, Arseniy Krasnov wrote: > 'hci_register_dev()' calls power up function, which is executed by > kworker - 'hci_power_on()'. This function does access to bluetooth chip > using callbacks from 'hci_ldisc.c', for example 'hci_uart_send_frame()'. > Now 'hci_uart_send_frame()' checks 'HCI_UART_PROTO_READY' bit set, and > if not - it fails. Problem is that 'HCI_UART_PROTO_READY' is set after > 'hci_register_dev()', and there is tiny chance that 'hci_power_on()' will > be executed before setting this bit. In that case HCI init logic fails. > > Patch moves setting of 'HCI_UART_PROTO_READY' before calling function > 'hci_uart_register_dev()'. > > Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com> > --- > Changelog v1->v2: > * Move 'set_bit()' before 'hci_uart_register_dev()' instead of > adding new bit 'HCI_UART_PROTO_INIT'. > > drivers/bluetooth/hci_ldisc.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c > index 30192bb083549..07b9aa09bbe2e 100644 > --- a/drivers/bluetooth/hci_ldisc.c > +++ b/drivers/bluetooth/hci_ldisc.c > @@ -704,12 +704,13 @@ static int hci_uart_set_proto(struct hci_uart *hu, int id) > > hu->proto = p; > > + set_bit(HCI_UART_PROTO_READY, &hu->flags); > + > err = hci_uart_register_dev(hu); > if (err) { > return err; > } > > - set_bit(HCI_UART_PROTO_READY, &hu->flags); > return 0; > } >
Hi Arseniy, On Thu, Dec 19, 2024 at 1:42 PM Arseniy Krasnov <avkrasnov@salutedevices.com> wrote: > > Hi, sorry i'm new in bluetooth subsystem. I get the following > message from CI: > https://patchwork.kernel.org/project/bluetooth/list/?series=918554 > Where one of tests was failed. Where I can get more information > about failure: > > tedd_an/TestRunner_mgmt-tester fail TestRunner_mgmt-tester: Total: 490, Passed: 483 (98.6%), Failed: 3, Not Run: 4 These looks like a false positives, your changes should affect the tests since that run with hci_vhci driver, not the uart ones. > ? > > Thanks > > On 17.12.2024 11:12, Arseniy Krasnov wrote: > > 'hci_register_dev()' calls power up function, which is executed by > > kworker - 'hci_power_on()'. This function does access to bluetooth chip > > using callbacks from 'hci_ldisc.c', for example 'hci_uart_send_frame()'. > > Now 'hci_uart_send_frame()' checks 'HCI_UART_PROTO_READY' bit set, and > > if not - it fails. Problem is that 'HCI_UART_PROTO_READY' is set after > > 'hci_register_dev()', and there is tiny chance that 'hci_power_on()' will > > be executed before setting this bit. In that case HCI init logic fails. > > > > Patch moves setting of 'HCI_UART_PROTO_READY' before calling function > > 'hci_uart_register_dev()'. > > > > Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com> > > --- > > Changelog v1->v2: > > * Move 'set_bit()' before 'hci_uart_register_dev()' instead of > > adding new bit 'HCI_UART_PROTO_INIT'. > > > > drivers/bluetooth/hci_ldisc.c | 3 ++- > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c > > index 30192bb083549..07b9aa09bbe2e 100644 > > --- a/drivers/bluetooth/hci_ldisc.c > > +++ b/drivers/bluetooth/hci_ldisc.c > > @@ -704,12 +704,13 @@ static int hci_uart_set_proto(struct hci_uart *hu, int id) > > > > hu->proto = p; > > > > + set_bit(HCI_UART_PROTO_READY, &hu->flags); > > + > > err = hci_uart_register_dev(hu); > > if (err) { > > return err; > > } > > > > - set_bit(HCI_UART_PROTO_READY, &hu->flags); > > return 0; > > } > >
Hi Arseniy, On Tue, Dec 17, 2024 at 3:12 AM Arseniy Krasnov <avkrasnov@salutedevices.com> wrote: > > 'hci_register_dev()' calls power up function, which is executed by > kworker - 'hci_power_on()'. This function does access to bluetooth chip > using callbacks from 'hci_ldisc.c', for example 'hci_uart_send_frame()'. > Now 'hci_uart_send_frame()' checks 'HCI_UART_PROTO_READY' bit set, and > if not - it fails. Problem is that 'HCI_UART_PROTO_READY' is set after > 'hci_register_dev()', and there is tiny chance that 'hci_power_on()' will > be executed before setting this bit. In that case HCI init logic fails. > > Patch moves setting of 'HCI_UART_PROTO_READY' before calling function > 'hci_uart_register_dev()'. > > Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com> > --- > Changelog v1->v2: > * Move 'set_bit()' before 'hci_uart_register_dev()' instead of > adding new bit 'HCI_UART_PROTO_INIT'. What drivers/controllers this was tested with? I want to make sure this doesn't cause regressions to other drivers if there are perhaps some drivers assuming HCI_UART_PROTO_READY was set in a certain order. > drivers/bluetooth/hci_ldisc.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c > index 30192bb083549..07b9aa09bbe2e 100644 > --- a/drivers/bluetooth/hci_ldisc.c > +++ b/drivers/bluetooth/hci_ldisc.c > @@ -704,12 +704,13 @@ static int hci_uart_set_proto(struct hci_uart *hu, int id) > > hu->proto = p; > > + set_bit(HCI_UART_PROTO_READY, &hu->flags); > + > err = hci_uart_register_dev(hu); > if (err) { > return err; > } > > - set_bit(HCI_UART_PROTO_READY, &hu->flags); > return 0; > } > > -- > 2.30.1
On 19.12.2024 22:18, Luiz Augusto von Dentz wrote: > Hi Arseniy, > > On Tue, Dec 17, 2024 at 3:12 AM Arseniy Krasnov > <avkrasnov@salutedevices.com> wrote: >> >> 'hci_register_dev()' calls power up function, which is executed by >> kworker - 'hci_power_on()'. This function does access to bluetooth chip >> using callbacks from 'hci_ldisc.c', for example 'hci_uart_send_frame()'. >> Now 'hci_uart_send_frame()' checks 'HCI_UART_PROTO_READY' bit set, and >> if not - it fails. Problem is that 'HCI_UART_PROTO_READY' is set after >> 'hci_register_dev()', and there is tiny chance that 'hci_power_on()' will >> be executed before setting this bit. In that case HCI init logic fails. >> >> Patch moves setting of 'HCI_UART_PROTO_READY' before calling function >> 'hci_uart_register_dev()'. >> >> Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com> >> --- >> Changelog v1->v2: >> * Move 'set_bit()' before 'hci_uart_register_dev()' instead of >> adding new bit 'HCI_UART_PROTO_INIT'. > > What drivers/controllers this was tested with? I want to make sure > this doesn't cause regressions to other drivers if there are perhaps > some drivers assuming HCI_UART_PROTO_READY was set in a certain order. Hi, I tested this on: CONFIG_BT=y CONFIG_BT_HCIUART=y CONFIG_BT_HCIUART_H4=y Yes, my v1 patchset with extra INIT bit was targeted to keep original behaviour - e.g. PROTO_READY bit usage still the same, just adding extra bit to handle this specific case. Thanks > >> drivers/bluetooth/hci_ldisc.c | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c >> index 30192bb083549..07b9aa09bbe2e 100644 >> --- a/drivers/bluetooth/hci_ldisc.c >> +++ b/drivers/bluetooth/hci_ldisc.c >> @@ -704,12 +704,13 @@ static int hci_uart_set_proto(struct hci_uart *hu, int id) >> >> hu->proto = p; >> >> + set_bit(HCI_UART_PROTO_READY, &hu->flags); >> + >> err = hci_uart_register_dev(hu); >> if (err) { >> return err; >> } >> >> - set_bit(HCI_UART_PROTO_READY, &hu->flags); >> return 0; >> } >> >> -- >> 2.30.1 > > >
On 19.12.2024 22:01, Luiz Augusto von Dentz wrote: > Hi Arseniy, > > On Thu, Dec 19, 2024 at 1:42 PM Arseniy Krasnov > <avkrasnov@salutedevices.com> wrote: >> >> Hi, sorry i'm new in bluetooth subsystem. I get the following >> message from CI: >> https://patchwork.kernel.org/project/bluetooth/list/?series=918554 >> Where one of tests was failed. Where I can get more information >> about failure: >> >> tedd_an/TestRunner_mgmt-tester fail TestRunner_mgmt-tester: Total: 490, Passed: 483 (98.6%), Failed: 3, Not Run: 4 > > These looks like a false positives, your changes should affect the > tests since that run with hci_vhci driver, not the uart ones. Got it, Thanks > >> ? >> >> Thanks >> >> On 17.12.2024 11:12, Arseniy Krasnov wrote: >>> 'hci_register_dev()' calls power up function, which is executed by >>> kworker - 'hci_power_on()'. This function does access to bluetooth chip >>> using callbacks from 'hci_ldisc.c', for example 'hci_uart_send_frame()'. >>> Now 'hci_uart_send_frame()' checks 'HCI_UART_PROTO_READY' bit set, and >>> if not - it fails. Problem is that 'HCI_UART_PROTO_READY' is set after >>> 'hci_register_dev()', and there is tiny chance that 'hci_power_on()' will >>> be executed before setting this bit. In that case HCI init logic fails. >>> >>> Patch moves setting of 'HCI_UART_PROTO_READY' before calling function >>> 'hci_uart_register_dev()'. >>> >>> Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com> >>> --- >>> Changelog v1->v2: >>> * Move 'set_bit()' before 'hci_uart_register_dev()' instead of >>> adding new bit 'HCI_UART_PROTO_INIT'. >>> >>> drivers/bluetooth/hci_ldisc.c | 3 ++- >>> 1 file changed, 2 insertions(+), 1 deletion(-) >>> >>> diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c >>> index 30192bb083549..07b9aa09bbe2e 100644 >>> --- a/drivers/bluetooth/hci_ldisc.c >>> +++ b/drivers/bluetooth/hci_ldisc.c >>> @@ -704,12 +704,13 @@ static int hci_uart_set_proto(struct hci_uart *hu, int id) >>> >>> hu->proto = p; >>> >>> + set_bit(HCI_UART_PROTO_READY, &hu->flags); >>> + >>> err = hci_uart_register_dev(hu); >>> if (err) { >>> return err; >>> } >>> >>> - set_bit(HCI_UART_PROTO_READY, &hu->flags); >>> return 0; >>> } >>> > > >
diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c index 30192bb083549..07b9aa09bbe2e 100644 --- a/drivers/bluetooth/hci_ldisc.c +++ b/drivers/bluetooth/hci_ldisc.c @@ -704,12 +704,13 @@ static int hci_uart_set_proto(struct hci_uart *hu, int id) hu->proto = p; + set_bit(HCI_UART_PROTO_READY, &hu->flags); + err = hci_uart_register_dev(hu); if (err) { return err; } - set_bit(HCI_UART_PROTO_READY, &hu->flags); return 0; }
'hci_register_dev()' calls power up function, which is executed by kworker - 'hci_power_on()'. This function does access to bluetooth chip using callbacks from 'hci_ldisc.c', for example 'hci_uart_send_frame()'. Now 'hci_uart_send_frame()' checks 'HCI_UART_PROTO_READY' bit set, and if not - it fails. Problem is that 'HCI_UART_PROTO_READY' is set after 'hci_register_dev()', and there is tiny chance that 'hci_power_on()' will be executed before setting this bit. In that case HCI init logic fails. Patch moves setting of 'HCI_UART_PROTO_READY' before calling function 'hci_uart_register_dev()'. Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com> --- Changelog v1->v2: * Move 'set_bit()' before 'hci_uart_register_dev()' instead of adding new bit 'HCI_UART_PROTO_INIT'. drivers/bluetooth/hci_ldisc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)