[PATCH v4 07/11] coresight-tpdm: Add nodes for dsb edge control
Suzuki K Poulose
suzuki.poulose at arm.com
Fri Jun 2 09:05:16 PDT 2023
On 02/06/2023 15:38, Tao Zhang wrote:
>
> On 6/2/2023 4:45 PM, Suzuki K Poulose wrote:
>> 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 at 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 at quicinc.com>, Tao
>>>>> Zhang (QUIC) <quic_taozha at 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 at quicinc.com>, Tao
>>>>> Zhang (QUIC) <quic_taozha at 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;
>> }
>>
> Sure, I will update in the next patch series.
>>
>>>
>>> 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);
Since we don't need mask any longer we could even do :
val &= ~EDCR_TO_WORD_MASK(i);
val |= EDCR_TO_WORD_VAL(edge_ctrl, i);
Suzuki
More information about the linux-arm-kernel
mailing list