Linux ARM-MSM sub-architecture
 help / color / mirror / Atom feed
From: Suzuki K Poulose <suzuki.poulose@arm.com>
To: Tao Zhang <quic_taozha@quicinc.com>,
	Mathieu Poirier <mathieu.poirier@linaro.org>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Konrad Dybcio <konradybcio@gmail.com>,
	Mike Leach <mike.leach@linaro.org>,
	Rob Herring <robh+dt@kernel.org>,
	Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>
Cc: Jinlong Mao <quic_jinlmao@quicinc.com>,
	Leo Yan <leo.yan@linaro.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	coresight@lists.linaro.org, linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, devicetree@vger.kernel.org,
	Tingwei Zhang <quic_tingweiz@quicinc.com>,
	Yuanfang Zhang <quic_yuanfang@quicinc.com>,
	Trilok Soni <quic_tsoni@quicinc.com>,
	Hao Zhang <quic_hazha@quicinc.com>,
	linux-arm-msm@vger.kernel.org, andersson@kernel.org
Subject: Re: [PATCH v4 07/11] coresight-tpdm: Add nodes for dsb edge control
Date: Fri, 2 Jun 2023 09:45:34 +0100	[thread overview]
Message-ID: <a2bd3bbf-5512-971a-95a1-3220f31814a2@arm.com> (raw)
In-Reply-To: <c5c28ab8-7d6a-f8e7-ad34-8716ac77d2dc@quicinc.com>

On 02/06/2023 09:21, Tao Zhang wrote:
> 
> On 6/1/2023 8:14 PM, Suzuki K Poulose wrote:
>> On 27/04/2023 10:00, Tao Zhang wrote:
>>> Add the nodes to set value for DSB edge control and DSB edge
>>> control mask. Each DSB subunit TPDM has maximum of n(n<16) EDCR
>>> resgisters to configure edge control. DSB edge detection control
>>> 00: Rising edge detection
>>> 01: Falling edge detection
>>> 10: Rising and falling edge detection (toggle detection)
>>> And each DSB subunit TPDM has maximum of m(m<8) ECDMR registers to
>>> configure mask. Eight 32 bit registers providing DSB interface
>>> edge detection mask control.
>>>
>>> Signed-off-by: Tao Zhang <quic_taozha@quicinc.com>
>>> ---
>>>   .../ABI/testing/sysfs-bus-coresight-devices-tpdm   |  32 +++++
>>>   drivers/hwtracing/coresight/coresight-tpdm.c       | 135 
>>> ++++++++++++++++++++-
>>>   drivers/hwtracing/coresight/coresight-tpdm.h       |  21 ++++
>>>   3 files changed, 187 insertions(+), 1 deletion(-)
>>>
>>> diff --git 
>>> a/Documentation/ABI/testing/sysfs-bus-coresight-devices-tpdm 
>>> b/Documentation/ABI/testing/sysfs-bus-coresight-devices-tpdm
>>> index 348e167..a57f000 100644
>>> --- a/Documentation/ABI/testing/sysfs-bus-coresight-devices-tpdm
>>> +++ b/Documentation/ABI/testing/sysfs-bus-coresight-devices-tpdm
>>> @@ -60,3 +60,35 @@ Description:
>>>           Bit[3] : Set to 0 for low performance mode.
>>>                    Set to 1 for high performance mode.
>>>           Bit[4:8] : Select byte lane for high performance mode.
>>> +
>>> +What: /sys/bus/coresight/devices/<tpdm-name>/dsb_edge_ctrl
>>> +Date:        March 2023
>>> +KernelVersion    6.3
>>> +Contact:    Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang 
>>> (QUIC) <quic_taozha@quicinc.com>
>>> +Description:
>>> +        Read/Write a set of the edge control registers of the DSB
>>> +        in TPDM.
>>> +
>>> +        Expected format is the following:
>>> +        <integer1> <integer2> <integer3>
>>> +
>>> +        Where:
>>> +        <integer1> : Start EDCR register number
>>> +        <integer2> : End EDCR register number
>>> +        <integer3> : The value need to be written
>>> +
>>> +What: /sys/bus/coresight/devices/<tpdm-name>/dsb_edge_ctrl_mask
>>> +Date:        March 2023
>>> +KernelVersion    6.3
>>> +Contact:    Jinlong Mao (QUIC) <quic_jinlmao@quicinc.com>, Tao Zhang 
>>> (QUIC) <quic_taozha@quicinc.com>
>>> +Description:
>>> +        Read/Write a set of the edge control mask registers of the
>>> +        DSB in TPDM.
>>> +
>>> +        Expected format is the following:
>>> +        <integer1> <integer2> <integer3>
>>> +
>>> +        Where:
>>> +        <integer1> : Start EDCMR register number
>>> +        <integer2> : End EDCMR register number
>>> +        <integer3> : The value need to be written
>>> diff --git a/drivers/hwtracing/coresight/coresight-tpdm.c 
>>> b/drivers/hwtracing/coresight/coresight-tpdm.c
>>> index 1bacaa5..a40e458 100644
>>> --- a/drivers/hwtracing/coresight/coresight-tpdm.c
>>> +++ b/drivers/hwtracing/coresight/coresight-tpdm.c
>>> @@ -80,7 +80,14 @@ static void set_trigger_type(struct tpdm_drvdata 
>>> *drvdata, u32 *val)
>>>     static void tpdm_enable_dsb(struct tpdm_drvdata *drvdata)
>>>   {
>>> -    u32 val;
>>> +    u32 val, i;
>>> +
>>> +    for (i = 0; i < TPDM_DSB_MAX_EDCR; i++)
>>> +        writel_relaxed(drvdata->dsb->edge_ctrl[i],
>>> +               drvdata->base + TPDM_DSB_EDCR(i));
>>> +    for (i = 0; i < TPDM_DSB_MAX_EDCMR; i++)
>>> +        writel_relaxed(drvdata->dsb->edge_ctrl_mask[i],
>>> +               drvdata->base + TPDM_DSB_EDCMR(i));
>>
>> Do all TPDM DSBs have MAX_EDCR registers ? Or some have less than that ?
>> If it is latter, do we need special care to avoid writing to inexistent
>> registers ?
>>
> You are right, not all DSB TPDMs have MAX_EDCR registers. In our design, 
> the inexistent register addresses
> 
> are not occupied and safe for accessing.
> 
> Currently we don't have a good way to know the quantity of EDCR/EDCMR 
> registers for DSB TPDMs.
> 
> The only way we can think of is to set it in device tree manually.
> 
> Do you have other suggestion for this?
> 
>>>         val = readl_relaxed(drvdata->base + TPDM_DSB_TIER);
>>>       /* Set trigger timestamp */
>>> @@ -313,6 +320,130 @@ static ssize_t dsb_mode_store(struct device *dev,
>>>   }
>>>   static DEVICE_ATTR_RW(dsb_mode);
>>>   +static ssize_t dsb_edge_ctrl_show(struct device *dev,
>>> +                       struct device_attribute *attr,
>>> +                       char *buf)
>>> +{
>>> +    struct tpdm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>>> +    ssize_t size = 0;
>>> +    int i;
>>> +
>>> +    spin_lock(&drvdata->spinlock);
>>> +    for (i = 0; i < TPDM_DSB_MAX_EDCR; i++) {
>>> +        size += sysfs_emit_at(buf, size,
>>> +                  "Index:0x%x Val:0x%x\n", i,
>>> +                  drvdata->dsb->edge_ctrl[i]);
>>
>> It may be safe, but please add a check to make sure that we don't
>> overflow. At least bail out when we hit a return of 0, indicating
>> reached the end of buffer.
>>
> Can I add the following check to replace the current code??
> 
> int ret = 0;
> 
> 
> for (i = 0; i < TPDM_DSB_MAX_EDCR; i++) {
> 
>      ret = sysfs_emit_at(buf, size, "Index:0x%x Val:0x%x\n", i, 
> drvdata->dsb->edge_ctrl[i]);
> 
>      if (!ret) {
> 
>          dev_warn(drvdata->dev, "The buffer has been overflowed\n");

You don't need this, it already triggers a WARN() in sysfs_emit_at().
So you could do:

	for (....) {
		unsigned long bytes = sysfs_emit_at(buf, size, ....);
	
		if (bytes <= 0)
			break;
		size += bytes;
	}

		
> 
>          spin_unlock(&drvdata->spinlock);
> 
>          return size;
> 
>      } else
> 
>          size += ret;
> 
> }
> 
>>> +    }
>>> +    spin_unlock(&drvdata->spinlock);
>>> +    return size;
>>> +}
>>> +
>>> +/*
>>> + * value 1: Start EDCR register number
>>> + * value 2: End EDCR register number
>>> + * value 3: The value need to be written
>>> + * The EDCR registers can include up to 16 32-bit registers, and each
>>> + * one can be configured to control up to 16 edge detections(2 bits
>>> + * control one edge detection). So a total 256 edge detections can be
>>> + * configured. So the starting number(value 1) and ending 
>>> number(value 2)
>>> + * cannot be greater than 256, and value 1 should be less than value 2.
>>> + * The following values are the rage of value 3.
>>> + * 0 - Rising edge detection
>>> + * 1 - Falling edge detection
>>> + * 2 - Rising and falling edge detection (toggle detection)
>>> + */
>>> +static ssize_t dsb_edge_ctrl_store(struct device *dev,
>>> +                    struct device_attribute *attr,
>>> +                    const char *buf,
>>> +                    size_t size)
>>> +{
>>> +    struct tpdm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>>> +    unsigned long val, mask, start, end, edge_ctrl, edge_ctrl_shift;
>>> +    int i, reg;
>>> +
>>> +    if (sscanf(buf, "%lx %lx %lx", &start, &end, &edge_ctrl) != 3)
>>> +        return -EINVAL;
>>> +    if ((start >= TPDM_DSB_MAX_LINES) || (end >= TPDM_DSB_MAX_LINES) ||
>>> +        (start > end) || (edge_ctrl > 0x2))
>>> +        return -EPERM;
>>> +
>>> +    spin_lock(&drvdata->spinlock);
>>> +    for (i = start; i <= end; i++) {
>>> +        /*
>>> +         * There are 2 bit per DSB Edge Control line.
>>> +         * Thus we have 16 lines in a 32bit word.
>>> +         */
>>> +        reg = EDCR_TO_WORD_IDX(i);
>>> +        mask = EDCR_TO_WORD_MASK(i);
>>> +        val = drvdata->dsb->edge_ctrl[reg];
>>
>>> +        edge_ctrl_shift = EDCR_TO_WORD_VAL(edge_ctrl, i);
>>> +        bitmap_replace(&val, &val, &edge_ctrl_shift, &mask, 32);
>>
>> Could we simply do :
>>
>>         reg &= ~mask;
>>         reg |= FIELD_PREP(mask, edge_ctrl);
>>
> Perhaps "FIELD_PREP" cannot be used here since "mask" must be constant 
> in this macro.

Ah, you are right. Sorry about that.

> 
> But in our code, the variable "mask" is not constant.

Still I think using the bitmap_replace is an overkill. We could simply
do:
		val &= ~mask;
		val |= EDCR_TO_WORD_VAL(edge_ctrl, i);

Suzuki

>>
>>> + drvdata->dsb->edge_ctrl[reg] = val;
>>> +    }
>>> +    spin_unlock(&drvdata->spinlock);
>>> +
>>> +    return size;
>>> +}
>>> +static DEVICE_ATTR_RW(dsb_edge_ctrl);
>>> +
>>> +static ssize_t dsb_edge_ctrl_mask_show(struct device *dev,
>>> +                        struct device_attribute *attr,
>>> +                        char *buf)
>>> +{
>>> +    struct tpdm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>>> +    ssize_t size = 0;
>>> +    int i;
>>> +
>>> +    spin_lock(&drvdata->spinlock);
>>> +    for (i = 0; i < TPDM_DSB_MAX_EDCMR; i++) {
>>> +        size += sysfs_emit_at(buf, size,
>>> +                  "Index:0x%x Val:0x%x\n", i,
>>> +                  drvdata->dsb->edge_ctrl_mask[i]);
>>> +    }
>>> +    spin_unlock(&drvdata->spinlock);
>>> +    return size;
>>> +}
>>> +
>>> +/*
>>> + * value 1: Start EDCMR register number
>>> + * value 2: End EDCMR register number
>>> + * value 3: The value need to be written
>>> + */
>>> +static ssize_t dsb_edge_ctrl_mask_store(struct device *dev,
>>> +                         struct device_attribute *attr,
>>> +                         const char *buf,
>>> +                         size_t size)
>>> +{
>>> +    struct tpdm_drvdata *drvdata = dev_get_drvdata(dev->parent);
>>> +    unsigned long start, end, val;
>>> +    u32 set;
>>> +    int i, reg;
>>> +
>>> +    if (sscanf(buf, "%lx %lx %lx", &start, &end, &val) != 3)
>>> +        return -EINVAL;
>>> +    if ((start >= TPDM_DSB_MAX_LINES) || (end >= TPDM_DSB_MAX_LINES)
>>> +        || (start > end) || (val & ~1UL))
>>> +        return -EPERM;
>>> +
>>> +    spin_lock(&drvdata->spinlock);
>>> +    for (i = start; i <= end; i++) {
>>> +        /*
>>> +         * There is 1 bit per DSB Edge Control Mark line.
>>> +         * Thus we have 32 lines in a 32bit word.
>>> +         */
>>> +        reg = EDCMR_TO_WORD_IDX(i);
>>> +        set = drvdata->dsb->edge_ctrl_mask[reg];
>>> +        if (val)
>>> +            set |= BIT(EDCR_TO_WORD_SHIFT(i));
>>> +        else
>>> +            set &= ~BIT(EDCR_TO_WORD_SHIFT(i));
>>> +        drvdata->dsb->edge_ctrl_mask[reg] = set;
>>> +    }
>>> +    spin_unlock(&drvdata->spinlock);
>>> +    return size;
>>> +}
>>> +static DEVICE_ATTR_RW(dsb_edge_ctrl_mask);
>>> +
>>>   static ssize_t dsb_trig_type_show(struct device *dev,
>>>                        struct device_attribute *attr, char *buf)
>>>   {
>>> @@ -385,6 +516,8 @@ static DEVICE_ATTR_RW(dsb_trig_ts);
>>>     static struct attribute *tpdm_dsb_attrs[] = {
>>>       &dev_attr_dsb_mode.attr,
>>> +    &dev_attr_dsb_edge_ctrl.attr,
>>> +    &dev_attr_dsb_edge_ctrl_mask.attr,
>>>       &dev_attr_dsb_trig_ts.attr,
>>>       &dev_attr_dsb_trig_type.attr,
>>>       NULL,
>>> diff --git a/drivers/hwtracing/coresight/coresight-tpdm.h 
>>> b/drivers/hwtracing/coresight/coresight-tpdm.h
>>> index 79df07e..f25dcdec 100644
>>> --- a/drivers/hwtracing/coresight/coresight-tpdm.h
>>> +++ b/drivers/hwtracing/coresight/coresight-tpdm.h
>>> @@ -12,6 +12,8 @@
>>>   /* DSB Subunit Registers */
>>>   #define TPDM_DSB_CR        (0x780)
>>>   #define TPDM_DSB_TIER        (0x784)
>>> +#define TPDM_DSB_EDCR(n)    (0x808 + (n * 4))
>>> +#define TPDM_DSB_EDCMR(n)    (0x848 + (n * 4))
>>>     /* Enable bit for DSB subunit */
>>>   #define TPDM_DSB_CR_ENA        BIT(0)
>>> @@ -34,6 +36,15 @@
>>>   #define TPDM_DSB_TEST_MODE        GENMASK(10, 9)
>>>   #define TPDM_DSB_HPSEL        GENMASK(6, 2)
>>>   +#define EDCRS_PER_WORD                16
>>> +#define EDCR_TO_WORD_IDX(r)            ((r) / EDCRS_PER_WORD)
>>> +#define EDCR_TO_WORD_SHIFT(r)        ((r % EDCRS_PER_WORD) * 2)
>>> +#define EDCR_TO_WORD_VAL(val, r)    (val << EDCR_TO_WORD_SHIFT(r))
>>> +#define EDCR_TO_WORD_MASK(r)        EDCR_TO_WORD_VAL(0x3, r)
>>
>> minor nit: add a new line here please
> Sure, I will update this in the next patch series.
>>
>>> +#define EDCMRS_PER_WORD                32
>>> +#define EDCMR_TO_WORD_IDX(r)        ((r) / EDCMRS_PER_WORD)
>>> +#define EDCMR_TO_WORD_SHIFT(r)        ((r) % EDCMRS_PER_WORD)
>>> +
>>>   /* TPDM integration test registers */
>>>   #define TPDM_ITATBCNTRL        (0xEF0)
>>>   #define TPDM_ITCNTRL        (0xF00)
>>> @@ -60,14 +71,24 @@
>>>   #define TPDM_PIDR0_DS_IMPDEF    BIT(0)
>>>   #define TPDM_PIDR0_DS_DSB    BIT(1)
>>>   +#define TPDM_DSB_MAX_LINES    256
>>> +/* MAX number of EDCR registers */
>>> +#define TPDM_DSB_MAX_EDCR    16
>>> +/* MAX number of EDCMR registers */
>>> +#define TPDM_DSB_MAX_EDCMR    8
>>> +
>>>   /**
>>>    * struct dsb_dataset - specifics associated to dsb dataset
>>>    * @mode:             DSB programming mode
>>> + * @edge_ctrl:        Save value for edge control
>>> + * @edge_ctrl_mask:   Save value for edge control mask
>>>    * @trig_ts:          Enable/Disable trigger timestamp.
>>>    * @trig_type:        Enable/Disable trigger type.
>>>    */
>>>   struct dsb_dataset {
>>>       u32                mode;
>>> +    u32                edge_ctrl[TPDM_DSB_MAX_EDCR];
>>> +    u32                edge_ctrl_mask[TPDM_DSB_MAX_EDCMR];
>>
>> minor nit: Please align it with the fields below.
> 
> Sure, I will update this in the next patch series.
> 
> 
> Best,
> 
> Tao
> 
>>
>>>       bool            trig_ts;
>>>       bool            trig_type;
>>>   };
>>
>> Suzuki


  reply	other threads:[~2023-06-02  8:46 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-04-27  9:00 [PATCH v4 00/11] Add support to configure TPDM DSB subunit Tao Zhang
2023-04-27  9:00 ` [PATCH v4 01/11] dt-bindings: arm: Add support for DSB element size Tao Zhang
2023-04-27 12:59   ` Rob Herring
2023-04-27  9:00 ` [PATCH v4 02/11] coresight-tpda: Add DSB dataset support Tao Zhang
2023-05-23 10:07   ` Suzuki K Poulose
2023-05-23 14:48     ` Suzuki K Poulose
2023-05-25  7:20       ` Tao Zhang
2023-05-25  7:16     ` Tao Zhang
2023-05-25  9:08       ` Suzuki K Poulose
2023-05-26  3:22         ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 03/11] coresight-tpdm: Initialize DSB subunit configuration Tao Zhang
2023-05-23 13:42   ` Suzuki K Poulose
2023-05-25  8:12     ` Tao Zhang
2023-05-25  9:09       ` Suzuki K Poulose
2023-05-26  3:46         ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 04/11] coresight-tpdm: Add reset node to TPDM node Tao Zhang
2023-05-23 14:53   ` Suzuki K Poulose
2023-05-25  8:36     ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 05/11] coresight-tpdm: Add nodes to set trigger timestamp and type Tao Zhang
2023-06-01  9:05   ` Suzuki K Poulose
2023-06-02  2:29     ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 06/11] coresight-tpdm: Add node to set dsb programming mode Tao Zhang
2023-06-01  9:23   ` Suzuki K Poulose
2023-06-02  2:58     ` Tao Zhang
2023-06-02  8:25       ` Suzuki K Poulose
2023-06-02  8:31         ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 07/11] coresight-tpdm: Add nodes for dsb edge control Tao Zhang
2023-06-01 12:14   ` Suzuki K Poulose
2023-06-02  8:21     ` Tao Zhang
2023-06-02  8:45       ` Suzuki K Poulose [this message]
2023-06-02  9:00         ` Suzuki K Poulose
2023-06-05  9:12           ` Tao Zhang
2023-06-02 14:38         ` Tao Zhang
2023-06-02 16:05           ` Suzuki K Poulose
2023-04-27  9:00 ` [PATCH v4 08/11] coresight-tpdm: Add nodes to configure pattern match output Tao Zhang
2023-06-01 13:28   ` Suzuki K Poulose
2023-06-02  8:29     ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 09/11] coresight-tpdm: Add nodes for timestamp request Tao Zhang
2023-06-05 10:19   ` Suzuki K Poulose
2023-06-06 10:55     ` Tao Zhang
2023-04-27  9:00 ` [PATCH v4 10/11] dt-bindings: arm: Add support for DSB MSR register Tao Zhang
2023-04-27  9:00 ` [PATCH v4 11/11] coresight-tpdm: Add nodes for dsb msr support Tao Zhang
2023-06-05 10:24   ` Suzuki K Poulose
2023-06-06 12:45     ` Tao Zhang
2023-04-27 16:53 ` [PATCH v4 00/11] Add support to configure TPDM DSB subunit Suzuki K Poulose
     [not found]   ` <725b6ccd-ff70-a3d2-fe44-797c0509e643@quicinc.com>
2023-06-01  8:17     ` Tao Zhang
2023-06-01  8:36       ` Suzuki K Poulose

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=a2bd3bbf-5512-971a-95a1-3220f31814a2@arm.com \
    --to=suzuki.poulose@arm.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=andersson@kernel.org \
    --cc=coresight@lists.linaro.org \
    --cc=devicetree@vger.kernel.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=konradybcio@gmail.com \
    --cc=krzysztof.kozlowski+dt@linaro.org \
    --cc=leo.yan@linaro.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.poirier@linaro.org \
    --cc=mike.leach@linaro.org \
    --cc=quic_hazha@quicinc.com \
    --cc=quic_jinlmao@quicinc.com \
    --cc=quic_taozha@quicinc.com \
    --cc=quic_tingweiz@quicinc.com \
    --cc=quic_tsoni@quicinc.com \
    --cc=quic_yuanfang@quicinc.com \
    --cc=robh+dt@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox