Re: [PATCH v8 1/4] docs: perf: Add description for Synopsys DesignWare PCIe PMU driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 2023/10/19 19:06, Jonathan Cameron wrote:

...

>>>> +
>>>> +The DWC PCIe PMU driver registers a perf PMU driver, which provides
>>>> +description of available events and configuration options in sysfs, see
>>>> +/sys/bus/event_source/devices/dwc_rootport_{bdf}.
>>>> +
>>>> +The "format" directory describes format of the config fields of the
>>>> +perf_event_attr structure. The "events" directory provides configuration
>>>> +templates for all documented events.  For example,
>>>> +"Rx_PCIe_TLP_Data_Payload" is an equivalent of "eventid=0x22,type=0x1".
>>>> +
>>>> +The "perf list" command shall list the available events from sysfs, e.g.::
>>>> +
>>>> +    $# perf list | grep dwc_rootport
>>>> +    <...>
>>>> +    dwc_rootport_3018/Rx_PCIe_TLP_Data_Payload/        [Kernel PMU event]
>>>> +    <...>
>>>> +    dwc_rootport_3018/rx_memory_read,lane=?/               [Kernel PMU event]
>>>> +
>>>> +Time Based Analysis Event Usage
>>>> +-------------------------------
>>>> +
>>>> +Example usage of counting PCIe RX TLP data payload (Units of 16 bytes)::
>>>> +
>>>> +    $# perf stat -a -e dwc_rootport_3018/Rx_PCIe_TLP_Data_Payload/
>>>> +
>>>> +The average RX/TX bandwidth can be calculated using the following formula:
>>>> +
>>>> +    PCIe RX Bandwidth = PCIE_RX_DATA * 16B / Measure_Time_Window
>>>> +    PCIe TX Bandwidth = PCIE_TX_DATA * 16B / Measure_Time_Window  
>>>
>>> Silly question (sorry I didn't raise it earlier) but can we make the interface
>>> more intuitive by just multiplying the counter value at point of read by 16?  
>>
>> Really a good suggestion, and it is very convenient for end perf users.
>> But the unit of 16 is only applied to group#1 as described in Time Based Analysis
>> section.
> 
> How hard would it be to just apply it to those events?
> Userspace doesn't care what the hardware does underneath - it just wants to get
> moderately intuitive data back. Having the end user deal with this oddity + even
> the need to document it seems to me to be unnecessary burden given how simple it
> is (I assume) to remove the oddity.

Ok. Talked me into it :)
I will multiply the counter value at point of read by 16 for group#1 events.

Thank you.

Best Regards,
Shuai



[Index of Archives]     [DMA Engine]     [Linux Coverity]     [Linux USB]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Greybus]

  Powered by Linux