Hi Luiz,
>>> This adds a debugfs entry to set msft_opcode enabling vhci to emulate
>>> controllers with MSFT extention support.
>>>
>>> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@xxxxxxxxx>
>>> ---
>>> drivers/bluetooth/hci_vhci.c | 32 ++++++++++++++++++++++++++++++++
>>> 1 file changed, 32 insertions(+)
>>>
>>> diff --git a/drivers/bluetooth/hci_vhci.c b/drivers/bluetooth/hci_vhci.c
>>> index 56c6b22be10b..ac122299bacc 100644
>>> --- a/drivers/bluetooth/hci_vhci.c
>>> +++ b/drivers/bluetooth/hci_vhci.c
>>> @@ -194,6 +194,34 @@ static const struct file_operations force_wakeup_fops = {
>>> .llseek = default_llseek,
>>> };
>>>
>>> +
>>> +static int msft_opcode_set(void *data, u64 val)
>>> +{
>>> + struct vhci_data *vhci = data;
>>> + uint16_t ogf = (val & 0xffff >> 10);
>>> +
>>> + if (val > 0xffff || ogf != 0x3f)
>>
>> I would actually just include it here to avoid any 16-bit overflow.
>>
>> if (val > 0xffff || (val & 0xffff >> 10) != 0x3f)
>
> Ack.
>
>>> + return -EINVAL;
>>> +
>>> + hci_set_msft_opcode(vhci->hdev, val);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +static int msft_opcode_get(void *data, u64 *val)
>>> +{
>>> + struct vhci_data *vhci = data;
>>> +
>>> + hci_dev_lock(vhci->hdev);
>>> + *val = vhci->hdev->msft_opcode;
>>> + hci_dev_unlock(vhci->hdev);
>>> +
>>> + return 0;
>>> +}
>>> +
>>> +DEFINE_DEBUGFS_ATTRIBUTE(msft_opcode_fops, msft_opcode_get, msft_opcode_set,
>>> + "%llu\n");
>>> +
>>> static int __vhci_create_device(struct vhci_data *data, __u8 opcode)
>>> {
>>> struct hci_dev *hdev;
>>> @@ -259,6 +287,10 @@ static int __vhci_create_device(struct vhci_data *data, __u8 opcode)
>>> debugfs_create_file("force_wakeup", 0644, hdev->debugfs, data,
>>> &force_wakeup_fops);
>>>
>>> + if (IS_ENABLED(CONFIG_BT_MSFTEXT))
>>> + debugfs_create_file("msft_opcode", 0644, hdev->debugfs, data,
>>> + &msft_opcode_fops);
>>> +
>>
>> So my concern is that you can modify this value when the device is up and running. That will cause havoc.
>>
>> Just checking HCI_UP is kinda bad since we just removed that access from the drivers.
>
> Right but we could add a check to HCI_UP inside hci_set_msft_opcode
> and make it return an error, actually this might be a good idea anyway
> even with existing so we prevent bad usage of hci_set_msft_opcode when
> already up.
I did mean actually HCI_RUNNING, but this still won't work out since you should be able to set the opcode from hdev->setup.
You might be able to craft enough tests around HCI_INIT and HCI_SETUP to make it fail hci_set_msft_opcode. So that might be the safest way after all.
One other option is to actually just store the msft_opcode from debugfs in vhci_data. And then only set it from within hdev->setup. You need to set HCI_QUIRK_NON_PERSISTENT_SETUP for this, but that might actually work best then.
Note: you need an aosp_capable debugfs setting as well.
Regards
Marcel
