Re: [PATCH] coresight: core: Add coresight name support

[Jinlong Mao <quic_jinlmao@quicinc.com>]

On 3/17/2023 10:25 PM, Suzuki K Poulose wrote:

> On 17/03/2023 05:34, Jinlong Mao wrote:
>>
>> On 3/13/2023 5:16 PM, Suzuki K Poulose wrote:
>>> Cc: Rob
>>>
>>> On 01/03/2023 15:11, Jinlong Mao wrote:
>>>> Hi Suzuki,
>>>>
>>>> On 2/9/2023 10:16 AM, Jinlong Mao wrote:
>>>>>
>>>>> On 2/8/2023 10:26 PM, Suzuki K Poulose wrote:
>>>>>> On 08/02/2023 11:07, Mao Jinlong wrote:
>>>>>>> Apart from STM and ETM sources, there will be more sources added to
>>>>>>> coresight components. For example, there are over 10 TPDM sources.
>>>>>>> Add coresight name support for custom names which will be
>>>>>>> easy to identify the source.
>>>>>>>
>>>>>>
>>>>>> As we have previously discussed, please don't make this a generic
>>>>>> code change. If your device has a "specifici" name, use that for
>>>>>> allocating in the driver and leave the core code alone.
>>>>>>
>>>>>> Suzuki
>>>>>>
>>>>> Hi Suzuki,
>>>>>
>>>>> Not only for TPDMs. There could be dozens of CTI devices.
>>>>> It is hard for user to know which CTI device it is with current 
>>>>> names.
>>>>>
>>>>> Thanks
>>>>> Jinlong Mao
>>>>
>>>> The coresight name support is applicable to CTI and TPDM devices.
>>>> This is a generic change for the source which has dozens of devices.
>>>
>>> I took a look at the CTI situation and I agree that the situation
>>> is a bit tricky. The CTI could be connected to multiple devices,
>>> some of them may not be even CoreSight devices. Given there could
>>> be numerous of them, we need some way to make the "devices" naming
>>> a bit more intuitive.
>>>
>>> Before we go ahead and add something specific to coresight, I would
>>> like to see if there is a generic property. Ideally, the "labels"
>>> in the DTS sources would have been an ideal choice, but can't
>>> see how that is available in the FDT.
>>>
>>> Suzuki
>> Hi Suzuki,
>>
>> Shall we use the full_name of device_node struct as coresight 
>> component's name ?
>>
>>    struct device_node {
>>      const char *name;
>>      phandle phandle;
>> *    const char *full_name;
>>
>> *For component below, the full_name will be "coresight-tpdm-ipcc".
>> *
>> **coresight-tpdm-ipcc* {
>
> Does that go against the convention of naming the DT nodes ?
> I am not sure. Also, we would need a way to solve this for ACPI too.
>
> Suzuki
Hi Suzuki,

Does ACPI device also use "coresight_alloc_device_name" to get the 
device's name ?

Thanks
Jinlong Mao

>
>
>>      compatible = "qcom,coresight-tpdm", "arm,primecell";
>>      reg = <0 0x10c29000 0 0x1000>;
>>
>>      clocks = <&aoss_qmp>;
>>      clock-names = "apb_pclk";
>>
>>      out-ports {
>>              port {
>>                      tpdm_ipcc_out_tpda_dl_center_27: endpoint {
>>                              remote-endpoint =
>> <&tpda_dl_center_27_in_tpdm_ipcc>;
>>                      };
>>              };
>>      };
>> };*
>>
>> *Thanks
>> Jinlong Mao
>>>
>>>
>>>>
>>>> Thanks
>>>> Jinlong Mao
>>>>
>>>>>>
>>>>>>> Signed-off-by: Mao Jinlong <quic_jinlmao@quicinc.com>
>>>>>>> ---
>>>>>>>   drivers/hwtracing/coresight/coresight-core.c | 34 
>>>>>>> +++++++++++---------
>>>>>>>   1 file changed, 19 insertions(+), 15 deletions(-)
>>>>>>>
>>>>>>> diff --git a/drivers/hwtracing/coresight/coresight-core.c 
>>>>>>> b/drivers/hwtracing/coresight/coresight-core.c
>>>>>>> index d3bf82c0de1d..5e95d9c7f256 100644
>>>>>>> --- a/drivers/hwtracing/coresight/coresight-core.c
>>>>>>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>>>>>>> @@ -1733,28 +1733,32 @@ char *coresight_alloc_device_name(struct 
>>>>>>> coresight_dev_list *dict,
>>>>>>>   {
>>>>>>>       int idx;
>>>>>>>       char *name = NULL;
>>>>>>> +    const char *coresight_name = NULL;
>>>>>>>       struct fwnode_handle **list;
>>>>>>> +    struct device_node *node = dev->of_node;
>>>>>>>         mutex_lock(&coresight_mutex);
>>>>>>>   -    idx = coresight_search_device_idx(dict, dev_fwnode(dev));
>>>>>>> -    if (idx < 0) {
>>>>>>> -        /* Make space for the new entry */
>>>>>>> -        idx = dict->nr_idx;
>>>>>>> -        list = krealloc_array(dict->fwnode_list,
>>>>>>> -                      idx + 1, sizeof(*dict->fwnode_list),
>>>>>>> -                      GFP_KERNEL);
>>>>>>> -        if (ZERO_OR_NULL_PTR(list)) {
>>>>>>> -            idx = -ENOMEM;
>>>>>>> -            goto done;
>>>>>>> +    if (!of_property_read_string(node, "coresight-name", 
>>>>>>> &coresight_name))
>>>>>>> +        name = devm_kasprintf(dev, GFP_KERNEL, "%s", 
>>>>>>> coresight_name);
>>>>>>> +    else {
>>>>>>> +        idx = coresight_search_device_idx(dict, dev_fwnode(dev));
>>>>>>> +        if (idx < 0) {
>>>>>>> +            /* Make space for the new entry */
>>>>>>> +            idx = dict->nr_idx;
>>>>>>> +            list = krealloc_array(dict->fwnode_list,
>>>>>>> +                          idx + 1, sizeof(*dict->fwnode_list),
>>>>>>> +                          GFP_KERNEL);
>>>>>>> +            if (ZERO_OR_NULL_PTR(list))
>>>>>>> +                goto done;
>>>>>>> +
>>>>>>> +            list[idx] = dev_fwnode(dev);
>>>>>>> +            dict->fwnode_list = list;
>>>>>>> +            dict->nr_idx = idx + 1;
>>>>>>>           }
>>>>>>>   -        list[idx] = dev_fwnode(dev);
>>>>>>> -        dict->fwnode_list = list;
>>>>>>> -        dict->nr_idx = idx + 1;
>>>>>>> +        name = devm_kasprintf(dev, GFP_KERNEL, "%s%d", 
>>>>>>> dict->pfx, idx);
>>>>>>>       }
>>>>>>> -
>>>>>>> -    name = devm_kasprintf(dev, GFP_KERNEL, "%s%d", dict->pfx, 
>>>>>>> idx);
>>>>>>>   done:
>>>>>>>       mutex_unlock(&coresight_mutex);
>>>>>>>       return name;
>>>>>>
>>>>> _______________________________________________
>>>>> CoreSight mailing list -- coresight@lists.linaro.org
>>>>> To unsubscribe send an email to coresight-leave@lists.linaro.org
>>>
>