[PATCH RFC 2/8] bus: mhi: host: add new interfaces to handle MHI channels directly

[Baochen Qiang]

On 11/11/2023 1:14 AM, Jeffrey Hugo wrote:
> On 11/10/2023 3:21 AM, Kalle Valo wrote:
>> From: Baochen Qiang <quic_bqiang at quicinc.com>
>>
>> When using mhi_power_down_no_destroy() MHI hosts need to unprepare 
>> MHI channels
>> by themselves.  Similarly, MHI stack will also not create new MHI 
>> device since
>> old devices were not destroyed, so MHI hosts need to prepare channels 
>> as well.
>> Hence add these two interfaces to make that possible.
>>
>> Tested-on: WCN6855 hw2.0 PCI 
>> WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30
>>
>> Signed-off-by: Baochen Qiang <quic_bqiang at quicinc.com>
>> Signed-off-by: Kalle Valo <quic_kvalo at quicinc.com>
>> ---
>>   drivers/bus/mhi/host/main.c | 91 +++++++++++++++++++++++++++++++++++++
>>   include/linux/mhi.h         | 18 ++++++++
>>   2 files changed, 109 insertions(+)
>>
>> diff --git a/drivers/bus/mhi/host/main.c b/drivers/bus/mhi/host/main.c
>> index dcf627b36e82..9bcf8a49c000 100644
>> --- a/drivers/bus/mhi/host/main.c
>> +++ b/drivers/bus/mhi/host/main.c
>> @@ -1667,6 +1667,49 @@ int mhi_prepare_for_transfer_autoqueue(struct 
>> mhi_device *mhi_dev)
>>   }
>>   EXPORT_SYMBOL_GPL(mhi_prepare_for_transfer_autoqueue);
>>   +static int __mhi_prepare_for_transfer_autoqueue(struct device 
>> *dev, void *data)
>> +{
>> +    struct mhi_device *mhi_dev;
>> +    struct mhi_chan *ul_chan, *dl_chan;
>> +    enum mhi_ee_type ee = MHI_EE_MAX;
>> +
>> +    if (dev->bus != &mhi_bus_type)
>> +        return 0;
>> +
>> +    mhi_dev = to_mhi_device(dev);
>> +
>> +    /* Only prepare virtual devices thats attached to bus */
>
> "that are"?
>
It means MHI devices with a type of MHI_DEVICE_XFER. See also 
mhi_destroy_device();


>> +    if (mhi_dev->dev_type == MHI_DEVICE_CONTROLLER)
>> +        return 0;
>> +
>> +    ul_chan = mhi_dev->ul_chan;
>> +    dl_chan = mhi_dev->dl_chan;
>> +
>> +    /*
>> +     * If execution environment is specified, remove only those 
>> devices that
>> +     * started in them based on ee_mask for the channels as we move 
>> on to a
>> +     * different execution environment
>> +     */
>> +    if (data)
>> +        ee = *(enum mhi_ee_type *)data;
>> +
>> +    if (ul_chan && ee != MHI_EE_MAX && !(ul_chan->ee_mask & BIT(ee)))
>> +        return 0;
>> +
>> +
>> +    if (dl_chan && ee != MHI_EE_MAX && !(dl_chan->ee_mask & BIT(ee)))
>> +        return 0;
>> +
>> +    return mhi_prepare_for_transfer_autoqueue(mhi_dev);
>> +}
>> +
>> +int mhi_prepare_all_for_transfer_autoqueue(struct mhi_controller 
>> *mhi_cntrl)
>> +{
>> +    return device_for_each_child(&mhi_cntrl->mhi_dev->dev, NULL,
>> +                     __mhi_prepare_for_transfer_autoqueue);
>> +}
>> +EXPORT_SYMBOL_GPL(mhi_prepare_all_for_transfer_autoqueue);
>
> This seems broken.  It appears to configure all channels as autoqueue, 
> regardless of how the controller initially configured them.  This 
> would only be safe to use if all channels were configured for 
> autoqueue, but would silently cause issues otherwise.

Thanks for pointing that. Yes, it is not correct to treat all channels 
as autoqueue regardless of its initial configuration. So how about 
change as below:

/* The difference between mhi_prepare_for_transfer_autoqueue() and 
mhi_prepare_for_transfer() comes from how to treat downlink channel */

mhi_prepare_for_transfer_dev(struct device *dev, ...)

{

...

dl_chan = mhi_dev->dl_chan;

...

if (dl_chan->pre_alloc)

        mhi_prepare_for_transfer_autoqueue(dev);

else

        mhi_prepare_for_transfer(dev);

}

/* And then iterate all devices and call mhi_prepare_for_transfer_dev() 
for each. */

int mhi_prepare_all_for_transfer(struct mhi_controller *mhi_cntrl)
{
     return device_for_each_child(&mhi_cntrl->mhi_dev->dev, NULL,
                      mhi_prepare_for_transfer_dev);
}
EXPORT_SYMBOL_GPL(mhi_prepare_all_for_transfer);