[PATCH v13 1/2] drivers/coresight: Add UltraSoc System Memory Buffer driver

[hejunhao]

On 2022/11/21 18:47, Suzuki Kuruppassery Poulose wrote:
> On 18/11/2022 12:45, hejunhao wrote:
>> Hi Suzuki ,
>>
>>
>> On 2022/11/15 19:06, Suzuki K Poulose wrote:
>>> On 14/11/2022 09:03, Junhao He wrote:
>>>> From: Qi Liu <liuqi115 at huawei.com>
>>>>
>
>>>> +static void smb_init_hw(struct smb_drv_data *drvdata)
>>>> +{
>>>> +    /* First disable SMB and clear the status of SMB buffer */
>>>> +    smb_reset_buffer_status(drvdata);
>>>> +    smb_disable_hw(drvdata);
>>>> +    smb_purge_data(drvdata);
>>>> +
>>>> +    writel(SMB_LB_CFG_LO_DEFAULT, drvdata->base + SMB_LB_CFG_LO_REG);
>>>> +    writel(SMB_LB_CFG_HI_DEFAULT, drvdata->base + SMB_LB_CFG_HI_REG);
>>>> +    writel(SMB_GLB_CFG_DEFAULT, drvdata->base + SMB_GLB_CFG_REG);
>>>> +    writel(SMB_GLB_INT_CFG, drvdata->base + SMB_GLB_INT_REG);
>>>> +    writel(SMB_LB_INT_CTRL_CFG, drvdata->base + SMB_LB_INT_CTRL_REG);
>>>
>>> Does this come with interrupt on overflow ? Do we not use this ?
>>>
>> When the buffer overflow, no interrupt will come.
>> Interrupt will upgrade SMB_LB_INT_STS_REG register status if start 
>> trace.
>> Thanks.
>>> Rest looks fine to me.
>
> What is the purpose of the "Interrupt" on the SMB ? It is not clear to
> me.
The SMB_LB_INT_CTRL_REG register control the validity of both real-time
events and interrupts. When logical buffer status changes causes to issue an
interrupt at the same time as it issues a real-time event.
Real-time events are used in SMB driver, which needs to get the buffer 
status.
Interrupts are used in debugger mode and cannot be registered in kernel.
  ..._BUF_NOTE_MASK control which events flags or interrupts are valid.

Thanks.

Best regards,
Junhao.

> Suzuki
>
> _______________________________________________
> CoreSight mailing list -- coresight at lists.linaro.org
> To unsubscribe send an email to coresight-leave at lists.linaro.org