[PATCH] coresight: etm4x: fix mis-usage of nr_resource in sysfs interface
Zhouwen (Jonathan, Turing)
jonathan.zhouwen at huawei.com
Fri Sep 4 06:08:46 EDT 2020
Hi Mathieu,
Thanks for your review.
在 2020/9/4 1:22, Mathieu Poirier 写道:
> On Wed, Sep 02, 2020 at 03:52:56PM +0800, Jonathan Zhou wrote:
>> The member @nr_resource represents how many resource selector pairs,
>> and the pair 0 is always implemented and reserved.
>> So let's multiply by 2 when displaying the implemented resource seletors
>> and resetting the seletor configuration .
>> And also update the validation of the input @idx.
>>
>> Cc: Mathieu Poirier <mathieu.poirier at linaro.org>
>> Cc: Suzuki K Poulose <suzuki.poulose at arm.com>
>> Cc: Mike Leach <mike.leach at linaro.org>
>> Cc: Shaokun Zhang <zhangshaokun at hisilicon.com>
>> Cc: lizixian at hisilicon.com
>> Signed-off-by: Jonathan Zhou <jonathan.zhouwen at huawei.com>
>> ---
>> drivers/hwtracing/coresight/coresight-etm4x-sysfs.c | 15 ++++++++++-----
>> 1 file changed, 10 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-etm4x-sysfs.c b/drivers/hwtracing/coresight/coresight-etm4x-sysfs.c
>> index b673e738bc9a..635b93d20efa 100644
>> --- a/drivers/hwtracing/coresight/coresight-etm4x-sysfs.c
>> +++ b/drivers/hwtracing/coresight/coresight-etm4x-sysfs.c
>> @@ -143,8 +143,10 @@ static ssize_t nr_resource_show(struct device *dev,
>> {
>> unsigned long val;
>> struct etmv4_drvdata *drvdata = dev_get_drvdata(dev->parent);
>> -
>> - val = drvdata->nr_resource;
>> + /* @nr_resource represents how many resource selector pairs,
>> + * hence multiply by 2.
>
> Your comment said it correctly, @nr_resource present the number of _pairs_. As
> such the code is correct and there is no need to multiply by 2.
>
Sorry for not mentioned the reason of modified this. From the function
@res_ctrl_store, we can see that only one individual resource selector
can be accessed via the ->res_idx. Therefore, it may be better to show
the number of individual resource selectors.
>> + */
>> + val = 2 * drvdata->nr_resource;
>> return scnprintf(buf, PAGE_SIZE, "%#lx\n", val);
>> }
>> static DEVICE_ATTR_RO(nr_resource);
>> @@ -236,7 +238,7 @@ static ssize_t reset_store(struct device *dev,
>> }
>>
>> config->res_idx = 0x0;
>> - for (i = 0; i < drvdata->nr_resource; i++)
>> + for (i = 2; i < 2 * drvdata->nr_resource; i++)
>> config->res_ctrl[i] = 0x0;
>
> This hunk is correct - please send a patch for it.
Ok.
>
>>
>> config->ss_idx = 0x0;
>> @@ -1663,8 +1665,11 @@ static ssize_t res_idx_store(struct device *dev,
>>
>> if (kstrtoul(buf, 16, &val))
>> return -EINVAL;
>> - /* Resource selector pair 0 is always implemented and reserved */
>> - if ((val == 0) || (val >= drvdata->nr_resource))
>> + /*
>> + * Resource selector pair 0 is always implemented and reserved,
>> + * namely an idx with 0 and 1 is illegal.
>> + */
>> + if ((val == 0) || (val == 1) || (val >= 2 * drvdata->nr_resource))
>> return -EINVAL;
>
> The index, i.e ->res_idx, represent pairs an not individual resource.
>
> Thanks,
> Mathieu
>
If we want all of the ->res_ctrl accessable, the ->res_idx should
represent individual resource. That's the reason.
Best regards.
Jonathan.
>>
>> /*
>> --
>> 1.9.1
>>
> .
>
More information about the linux-arm-kernel
mailing list