linux-kernel - Re: [PATCH 4/9] coresight-tpdm: Add reset node to TPDM node

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <f72dd193-d801-923c-c742-44584d1779c5@quicinc.com>
Date:   Thu, 27 Oct 2022 14:52:34 +0800
From:   Tao Zhang <quic_taozha@...cinc.com>
To:     Suzuki K Poulose <suzuki.poulose@....com>,
        Mathieu Poirier <mathieu.poirier@...aro.org>,
        Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
        Konrad Dybcio <konradybcio@...il.com>,
        Mike Leach <mike.leach@...aro.org>
CC:     Jinlong Mao <quic_jinlmao@...cinc.com>,
        Leo Yan <leo.yan@...aro.org>,
        "Greg Kroah-Hartman" <gregkh@...uxfoundation.org>,
        <coresight@...ts.linaro.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>,
        Tingwei Zhang <quic_tingweiz@...cinc.com>,
        Yuanfang Zhang <quic_yuanfang@...cinc.com>,
        Trilok Soni <quic_tsoni@...cinc.com>,
        Hao Zhang <quic_hazha@...cinc.com>,
        <linux-arm-msm@...r.kernel.org>, <bjorn.andersson@...aro.org>
Subject: Re: [PATCH 4/9] coresight-tpdm: Add reset node to TPDM node

Hi Suzuki,

在 10/24/2022 6:10 PM, Suzuki K Poulose 写道:
> On 08/09/2022 09:45, Tao Zhang wrote:
>> TPDM device need a node to reset the configurations and status of
>> it. This change provides a node to reset the configurations and
>> disable the TPDM if it has been enabled.
>
> It is not clear to me *why* this is needed. Please could you
> elaborate on the use case of this ? See my questions below.
For example, we usually reset the configuration through the "reset"
node before using a TPDM, so as to avoid the previous configuration
affecting the current use. And in some scenarios, it may be necessary
to reset the TPDM configuration to complete the verification of certain
function.
>>
>> Signed-off-by: Tao Zhang <quic_taozha@...cinc.com>
>> ---
>>   drivers/hwtracing/coresight/coresight-tpdm.c | 32 
>> ++++++++++++++++++++++++++++
>>   1 file changed, 32 insertions(+)
>>
>> diff --git a/drivers/hwtracing/coresight/coresight-tpdm.c 
>> b/drivers/hwtracing/coresight/coresight-tpdm.c
>> index 69ea453..74cc653 100644
>> --- a/drivers/hwtracing/coresight/coresight-tpdm.c
>> +++ b/drivers/hwtracing/coresight/coresight-tpdm.c
>> @@ -152,6 +152,37 @@ static void tpdm_init_default_data(struct 
>> tpdm_drvdata *drvdata)
>>       }
>>   }
>>   +static ssize_t reset_store(struct device *dev,
>> +                      struct device_attribute *attr,
>> +                      const char *buf,
>> +                      size_t size)
>
> minor nit: Alignment
>
I will update this in the next patch series.
>> +{
>> +    int ret = 0;
>> +    unsigned long val;
>> +    struct tpdm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>> +
>> +    ret = kstrtoul(buf, 10, &val);
>
> So, any integer value written to the sysfs knob triggers the rest ?
> It may be better to restrict this to "1".
>
Make sense, I will update this in the next patch series.
>
>> +    if (ret)
>> +        return ret;
>> +
>> +    spin_lock(&drvdata->spinlock);
>> +    /* Reset all datasets to ZERO */
>> +    if (drvdata->dsb != NULL)
>> +        memset(drvdata->dsb, 0, sizeof(struct dsb_dataset));
>> +
>> +    /* Init the default data */
>> +    tpdm_init_default_data(drvdata);
>
> Why is this needed ? Does the DSB device configuration change
> on the fly ?
>
We usually reset the configuration through the "reset" node at the
beginning of using a TPDM. DSB related elements need to be initialized
before configuring them.
The TPDM is usually to be used by the following sequence.
Reset(initialization) -> Configure -> Enable to use
>> +
>> +    spin_unlock(&drvdata->spinlock);
>> +
>> +    /* Disable tpdm if enabled */
>> +    if (drvdata->enable)
>> +        coresight_disable(drvdata->csdev);
>
> Why is this needed ? Isn't this supposed to be triggered from the "path"
> when the trace session is complete ?
>
In some scenarios, it may be necessary to reset the TPDM configuration
in the process of verification and re-configure the TPDM. If this is the
case, we need to disable the TPDM first, and re-configure before enabling
it again.
>
> Suzuki
>
>> +
>> +    return size;
>> +}
>> +static DEVICE_ATTR_WO(reset);
>> +
>>   /*
>>    * value 1: 64 bits test data
>>    * value 2: 32 bits test data
>> @@ -192,6 +223,7 @@ static ssize_t integration_test_store(struct 
>> device *dev,
>>   static DEVICE_ATTR_WO(integration_test);
>>     static struct attribute *tpdm_attrs[] = {
>> +    &dev_attr_reset.attr,
>>       &dev_attr_integration_test.attr,
>>       NULL,
>>   };
>
Best Regards
Tao