Received-SPF: pass (google.com: domain of linux-kernel+bounces-167075-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Message-ID: <a88b41f4-7e53-e162-5a6a-2d470e29c0bb@quicinc.com>
Date: Thu, 2 May 2024 16:37:15 -0700
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.15.1
Subject: Re: [PATCH v20 06/41] usb: host: xhci-sideband: Expose a sideband
 interrupter enable API
Content-Language: en-US
To: Mathias Nyman <mathias.nyman@linux.intel.com>,
        <srinivas.kandagatla@linaro.org>, <mathias.nyman@intel.com>,
        <perex@perex.cz>, <conor+dt@kernel.org>, <corbet@lwn.net>,
        <lgirdwood@gmail.com>, <andersson@kernel.org>,
        <krzysztof.kozlowski+dt@linaro.org>, <gregkh@linuxfoundation.org>,
        <Thinh.Nguyen@synopsys.com>, <broonie@kernel.org>,
        <bgoswami@quicinc.com>, <tiwai@suse.com>, <robh@kernel.org>,
        <konrad.dybcio@linaro.org>
CC: <linux-kernel@vger.kernel.org>, <devicetree@vger.kernel.org>,
        <linux-sound@vger.kernel.org>, <linux-usb@vger.kernel.org>,
        <linux-arm-msm@vger.kernel.org>, <linux-doc@vger.kernel.org>,
        <alsa-devel@alsa-project.org>
References: <20240425215125.29761-1-quic_wcheng@quicinc.com>
 <20240425215125.29761-7-quic_wcheng@quicinc.com>
 <ddd682da-5cfd-db09-e316-3c54939caf90@linux.intel.com>
From: Wesley Cheng <quic_wcheng@quicinc.com>
In-Reply-To: <ddd682da-5cfd-db09-e316-3c54939caf90@linux.intel.com>
Content-Type: text/plain; charset="UTF-8"; format=flowed
Content-Transfer-Encoding: 8bit

Hi Mathias,

On 5/2/2024 4:07 AM, Mathias Nyman wrote:
> On 26.4.2024 0.50, Wesley Cheng wrote:
>> Some use cases maybe require that the secondary interrupter's events to
>> be handled by the OS.  In this case, configure the IMOD and the
>> skip_events property to enable the interrupter's events.  By default,
>> assume that the secondary interrupter doesn't want to enable OS event
>> handling.
>>
>> Signed-off-by: Wesley Cheng <quic_wcheng@quicinc.com>
>> ---
>>   drivers/usb/host/xhci-sideband.c  | 28 ++++++++++++++++++++++++++++
>>   include/linux/usb/xhci-sideband.h |  2 ++
>>   2 files changed, 30 insertions(+)
>>
>> diff --git a/drivers/usb/host/xhci-sideband.c 
>> b/drivers/usb/host/xhci-sideband.c
>> index 255feae33c6e..6fdae9840c11 100644
>> --- a/drivers/usb/host/xhci-sideband.c
>> +++ b/drivers/usb/host/xhci-sideband.c
>> @@ -237,6 +237,30 @@ xhci_sideband_get_event_buffer(struct 
>> xhci_sideband *sb)
>>   }
>>   EXPORT_SYMBOL_GPL(xhci_sideband_get_event_buffer);
>> +/**
>> + * xhci_sideband_enable_interrupt - enable interrupt for secondary 
>> interrupter
>> + * @sb: sideband instance for this usb device
>> + * @imod_interval: number of event ring segments to allocate
>> + *
>> + * Enables OS owned event handling for a particular interrupter if 
>> client
>> + * requests for it.  In addition, set the IMOD interval for this 
>> particular
>> + * interrupter.
>> + *
>> + * Returns 0 on success, negative error otherwise
>> + */
>> +int xhci_sideband_enable_interrupt(struct xhci_sideband *sb, u32 
>> imod_interval)
>> +{
>> +    if (!sb || !sb->ir)
>> +        return -ENODEV;
>> +
>> +    xhci_set_interrupter_moderation(sb->ir, imod_interval);
> 
> Is there a need to adjust the moderation after initial setup?
> 
> If not then maybe we could pass the imod_interval as a parameter to
> xhci_create_secondary_interrupter(), and avoid exporting
> xhci_set_interrupter_moderation()
> 
> 

Let me preface my comments by saying that I was trying to include some 
aspects of enabling the secondary interrupter line within the main apps 
proc.  If this gets too confusing, I can remove these mechanisms for 
now.  For example, as you mentioned below 
xhci_sideband_enable_interrupt() isn't going to be used in the offload path.

However, I decided to add it so we can have some corresponding function 
that will utilize/set skip_events = false. (as it is "true" by default) 
Again, I can remove this part and revisit later when we actually have a 
use case to handle secondary interrupts on apps.

As for the IMOD setting, depending on what you think, I can add it as 
part of xhci_create_secondary_interrupter()

>> +    sb->ir->skip_events = false;
>> +    xhci_enable_interrupter(sb->ir);
>> +
>> +    return 0;
>> +}
>> +EXPORT_SYMBOL_GPL(xhci_sideband_enable_interrupt);
> 
> I can't find the place where xhci_sideband_enable_interrupt() is called in
> this series. How is it planned to be used?
> 
> Thanks
> Mathias

Thanks
Wesley Cheng