[v2] Bluetooth: Always call advertising disable before setting params

In __hci_req_enable_advertising, the HCI_LE_ADV hdev flag is temporarily
cleared to allow the random address to be set, which exposes a race
condition when an advertisement is configured immediately (<10ms) after
software rotation starts to refresh an advertisement.

In normal operation, the HCI_LE_ADV flag is updated as follows:

1. adv_timeout_expire is called, HCI_LE_ADV gets cleared in
   __hci_req_enable_advertising, but hci_req configures an enable
   request
2. hci_req is run, enable callback re-sets HCI_LE_ADV flag

However, in this race condition, the following occurs:

1. adv_timeout_expire is called, HCI_LE_ADV gets cleared in
   __hci_req_enable_advertising, but hci_req configures an enable
   request
2. add_advertising is called, which also calls
   __hci_req_enable_advertising. Because HCI_LE_ADV was cleared in Step
   1, no "disable" command is queued.
3. hci_req for adv_timeout_expire is run, which enables advertising and
   re-sets HCI_LE_ADV
4. hci_req for add_advertising is run, but because no "disable" command
   was queued, we try to set advertising parameters while advertising is
   active, causing a Command Disallowed error, failing the registration.

To resolve the issue, this patch removes the check for the HCI_LE_ADV
flag, and always queues the "disable" request, since HCI_LE_ADV could be
very temporarily out-of-sync. According to the spec, there is no harm in
calling "disable" when advertising is not active.

An example trace showing the HCI error in setting advertising parameters
is included below, with some notes annotating the states I mentioned
above:

@ MGMT Command: Add Ext Adv.. (0x0055) plen 35  {0x0001} [hci0]04:05.884
        Instance: 3
        Advertising data length: 24
        16-bit Service UUIDs (complete): 2 entries
          Location and Navigation (0x1819)
          Phone Alert Status Service (0x180e)
        Company: not assigned (65283)
          Data: 3a3b3c3d3e
        Service Data (UUID 0x9993): 3132333435
        Scan response length: 0
@ MGMT Event: Advertising Ad.. (0x0023) plen 1  {0x0005} [hci0]04:05.885
        Instance: 3

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=454833

---Test result---

##############################
Test: CheckPatch - PASS


##############################
Test: CheckGitLint - PASS


##############################
Test: CheckBuildK - PASS


##############################
Test: CheckTestRunner: Setup - PASS


##############################
Test: CheckTestRunner: l2cap-tester - FAIL
Total: 40, Passed: 31 (77.5%), Failed: 3, Not Run: 6

Failed Test Cases
L2CAP LE Server - Success                            Timed out    2.553 seconds
L2CAP LE Server - Nval SCID                          Timed out    2.001 seconds
L2CAP LE ATT Server - Success                        Timed out    1.875 seconds

##############################
Test: CheckTestRunner: bnep-tester - PASS
Total: 1, Passed: 1 (100.0%), Failed: 0, Not Run: 0

##############################
Test: CheckTestRunner: mgmt-tester - FAIL
Total: 416, Passed: 348 (83.7%), Failed: 6, Not Run: 62

Failed Test Cases
Pairing Acceptor - LE 1                              Timed out    2.266 seconds
Pairing Acceptor - LE 2                              Timed out    1.999 seconds
Pairing Acceptor - LE 3                              Timed out    2.000 seconds
Pairing Acceptor - LE 4                              Timed out    2.001 seconds
Pairing Acceptor - LE 5                              Timed out    1.998 seconds
Add Advertising - Success 3 (!Powered, Adv Enable)   Failed       0.025 seconds

##############################
Test: CheckTestRunner: rfcomm-tester - PASS
Total: 9, Passed: 9 (100.0%), Failed: 0, Not Run: 0

##############################
Test: CheckTestRunner: sco-tester - PASS
Total: 8, Passed: 8 (100.0%), Failed: 0, Not Run: 0

##############################
Test: CheckTestRunner: smp-tester - FAIL
Total: 8, Passed: 4 (50.0%), Failed: 4, Not Run: 0

Failed Test Cases
SMP Server - Basic Request 1                         Timed out    2.468 seconds
SMP Server - Invalid Request 1                       Timed out    1.984 seconds
SMP Server - Invalid Request 2                       Timed out    2.004 seconds
SMP Server - Invalid Request 3                       Timed out    1.996 seconds

##############################
Test: CheckTestRunner: userchan-tester - PASS
Total: 3, Passed: 3 (100.0%), Failed: 0, Not Run: 0



---
Regards,
Linux Bluetooth

Hi maintainers,

Please disregard this patch. While most chipsets honor the "Disabling
advertising when it is already disabled has no effect" clause in the
spec, I discovered over the weekend that some (WP2, ThP2, for
instance) controllers will return an error status if there is no
advertising active when the disable request is submitted. This patch
regresses advertising on these platforms, so I will look into a
different approach to solve this race condition.

Best,
Daniel Winkler


On Thu, Mar 25, 2021 at 10:27 AM Daniel Winkler
<danielwinkler@google.com> wrote:
>
> Hi all,
>
> It looks like my change breaks the expectations of one mgmt-tester
> test, which uses an Adv Enable (True) as a test condition. It is
> surprised to first see an Adv Enable (False) in the HCI traffic, and
> fails. I think my suggested approach here is the simplest and most
> robust to solve this race condition, so if the maintainers are happy
> with it, I can look into changing the test expectations to suit the
> new scenario. Please advise.
>
> Thanks in advance,
> Daniel
>
> On Wed, Mar 24, 2021 at 11:47 AM Daniel Winkler
> <danielwinkler@google.com> wrote:
> >
> > In __hci_req_enable_advertising, the HCI_LE_ADV hdev flag is temporarily
> > cleared to allow the random address to be set, which exposes a race
> > condition when an advertisement is configured immediately (<10ms) after
> > software rotation starts to refresh an advertisement.
> >
> > In normal operation, the HCI_LE_ADV flag is updated as follows:
> >
> > 1. adv_timeout_expire is called, HCI_LE_ADV gets cleared in
> >    __hci_req_enable_advertising, but hci_req configures an enable
> >    request
> > 2. hci_req is run, enable callback re-sets HCI_LE_ADV flag
> >
> > However, in this race condition, the following occurs:
> >
> > 1. adv_timeout_expire is called, HCI_LE_ADV gets cleared in
> >    __hci_req_enable_advertising, but hci_req configures an enable
> >    request
> > 2. add_advertising is called, which also calls
> >    __hci_req_enable_advertising. Because HCI_LE_ADV was cleared in Step
> >    1, no "disable" command is queued.
> > 3. hci_req for adv_timeout_expire is run, which enables advertising and
> >    re-sets HCI_LE_ADV
> > 4. hci_req for add_advertising is run, but because no "disable" command
> >    was queued, we try to set advertising parameters while advertising is
> >    active, causing a Command Disallowed error, failing the registration.
> >
> > To resolve the issue, this patch removes the check for the HCI_LE_ADV
> > flag, and always queues the "disable" request, since HCI_LE_ADV could be
> > very temporarily out-of-sync. According to the spec, there is no harm in
> > calling "disable" when advertising is not active.
> >
> > An example trace showing the HCI error in setting advertising parameters
> > is included below, with some notes annotating the states I mentioned
> > above:
> >
> > @ MGMT Command: Add Ext Adv.. (0x0055) plen 35  {0x0001} [hci0]04:05.884
> >         Instance: 3
> >         Advertising data length: 24
> >         16-bit Service UUIDs (complete): 2 entries
> >           Location and Navigation (0x1819)
> >           Phone Alert Status Service (0x180e)
> >         Company: not assigned (65283)
> >           Data: 3a3b3c3d3e
> >         Service Data (UUID 0x9993): 3132333435
> >         Scan response length: 0
> > @ MGMT Event: Advertising Ad.. (0x0023) plen 1  {0x0005} [hci0]04:05.885
> >         Instance: 3
> >
> > === adv_timeout_expire request starts running. This request was created
> > before our add advertising request
> > > HCI Event: Command Complete (0x0e) plen 4         #220 [hci0]04:05.993
> >       LE Set Advertising Data (0x08|0x0008) ncmd 1
> >         Status: Success (0x00)
> > < HCI Command: LE Set Scan.. (0x08|0x0009) plen 32  #221 [hci0]04:05.993
> >         Length: 24
> >         Service Data (UUID 0xabcd): 161718191a1b1c1d1e1f2021222324252627
> > > HCI Event: Command Complete (0x0e) plen 4         #222 [hci0]04:05.995
> >       LE Set Scan Response Data (0x08|0x0009) ncmd 1
> >         Status: Success (0x00)
> > < HCI Command: LE Set Adver.. (0x08|0x000a) plen 1  #223 [hci0]04:05.995
> >         Advertising: Disabled (0x00)
> > > HCI Event: Command Complete (0x0e) plen 4         #224 [hci0]04:05.997
> >       LE Set Advertise Enable (0x08|0x000a) ncmd 1
> >         Status: Success (0x00)
> > < HCI Command: LE Set Adve.. (0x08|0x0006) plen 15  #225 [hci0]04:05.997
> >         Min advertising interval: 200.000 msec (0x0140)
> >         Max advertising interval: 200.000 msec (0x0140)
> >         Type: Connectable undirected - ADV_IND (0x00)
> >         Own address type: Public (0x00)
> >         Direct address type: Public (0x00)
> >         Direct address: 00:00:00:00:00:00 (OUI 00-00-00)
> >         Channel map: 37, 38, 39 (0x07)
> >         Filter policy: Allow Scan Request, Connect from Any (0x00)
> > > HCI Event: Command Complete (0x0e) plen 4         #226 [hci0]04:05.998
> >       LE Set Advertising Parameters (0x08|0x0006) ncmd 1
> >         Status: Success (0x00)
> > < HCI Command: LE Set Adver.. (0x08|0x000a) plen 1  #227 [hci0]04:05.999
> >         Advertising: Enabled (0x01)
> > > HCI Event: Command Complete (0x0e) plen 4         #228 [hci0]04:06.000
> >       LE Set Advertise Enable (0x08|0x000a) ncmd 1
> >         Status: Success (0x00)
> >
> > === Our new add_advertising request starts running
> > < HCI Command: Read Local N.. (0x03|0x0014) plen 0  #229 [hci0]04:06.001
> > > HCI Event: Command Complete (0x0e) plen 252       #230 [hci0]04:06.005
> >       Read Local Name (0x03|0x0014) ncmd 1
> >         Status: Success (0x00)
> >         Name: Chromebook_FB3D
> >
> > === Although the controller is advertising, no disable command is sent
> > < HCI Command: LE Set Adve.. (0x08|0x0006) plen 15  #231 [hci0]04:06.005
> >         Min advertising interval: 200.000 msec (0x0140)
> >         Max advertising interval: 200.000 msec (0x0140)
> >         Type: Connectable undirected - ADV_IND (0x00)
> >         Own address type: Public (0x00)
> >         Direct address type: Public (0x00)
> >         Direct address: 00:00:00:00:00:00 (OUI 00-00-00)
> >         Channel map: 37, 38, 39 (0x07)
> >         Filter policy: Allow Scan Request, Connect from Any (0x00)
> > > HCI Event: Command Complete (0x0e) plen 4         #232 [hci0]04:06.005
> >       LE Set Advertising Parameters (0x08|0x0006) ncmd 1
> >         Status: Command Disallowed (0x0c)
> >
> > Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
> > Signed-off-by: Daniel Winkler <danielwinkler@google.com>
> > ---
> >
> > Changes in v2:
> > - Added btmon snippet showing HCI command failure
> >
> >  net/bluetooth/hci_request.c | 6 ++++--
> >  1 file changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/net/bluetooth/hci_request.c b/net/bluetooth/hci_request.c
> > index 8ace5d34b01efe..2b4b99f4cedf21 100644
> > --- a/net/bluetooth/hci_request.c
> > +++ b/net/bluetooth/hci_request.c
> > @@ -1547,8 +1547,10 @@ void __hci_req_enable_advertising(struct hci_request *req)
> >         if (!is_advertising_allowed(hdev, connectable))
> >                 return;
> >
> > -       if (hci_dev_test_flag(hdev, HCI_LE_ADV))
> > -               __hci_req_disable_advertising(req);
> > +       /* Request that the controller stop advertising. This can be called
> > +        * whether or not there is an active advertisement.
> > +        */
> > +       __hci_req_disable_advertising(req);
> >
> >         /* Clear the HCI_LE_ADV bit temporarily so that the
> >          * hci_update_random_address knows that it's safe to go ahead
> > --
> > 2.31.0.291.g576ba9dcdaf-goog
> >