Linux ARM-MSM sub-architecture
 help / color / mirror / Atom feed
From: Jinlong Mao <quic_jinlmao@quicinc.com>
To: Suzuki K Poulose <suzuki.poulose@arm.com>,
	Mike Leach <mike.leach@linaro.org>,
	James Clark <james.clark@arm.com>, Rob Herring <robh@kernel.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: <coresight@lists.linaro.org>,
	<linux-arm-kernel@lists.infradead.org>,
	<devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<linux-arm-msm@vger.kernel.org>
Subject: Re: [PATCH v6 2/3] coresight: Add support to get static id for system trace sources
Date: Tue, 19 Nov 2024 18:38:13 +0800	[thread overview]
Message-ID: <ae1d4e9a-6273-407e-952c-aec38e5c7a4f@quicinc.com> (raw)
In-Reply-To: <ef057e19-1209-4a72-b223-ef3f1e2b4513@arm.com>



On 2024/11/18 18:18, Suzuki K Poulose wrote:
> On 18/11/2024 10:14, Jinlong Mao wrote:
>>
>>
>> On 2024/11/13 21:27, Suzuki K Poulose wrote:
>>> On 04/11/2024 11:56, Mao Jinlong wrote:
>>>> Dynamic trace id was introduced in coresight subsystem, so trace id is
>>>> allocated dynamically. However, some hardware ATB source has static 
>>>> trace
>>>> id and it cannot be changed via software programming. For such source,
>>>> it can call coresight_get_static_trace_id to get the fixed trace id 
>>>> from
>>>> device node and pass id to coresight_trace_id_get_static_system_id to
>>>> reserve the id.
>>>>
>>>> Signed-off-by: Mao Jinlong <quic_jinlmao@quicinc.com>
>>>> Reviewed-by: Mike Leach <mike.leach@linaro.org>
>>>> ---
>>>>   .../hwtracing/coresight/coresight-platform.c  |  6 +++
>>>>   .../hwtracing/coresight/coresight-trace-id.c  | 39 ++++++++++++ 
>>>> +------
>>>>   .../hwtracing/coresight/coresight-trace-id.h  |  9 +++++
>>>>   include/linux/coresight.h                     |  1 +
>>>>   4 files changed, 44 insertions(+), 11 deletions(-)
>>>>
>>>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c b/ 
>>>> drivers/hwtracing/coresight/coresight-platform.c
>>>> index 64e171eaad82..633d96b9577a 100644
>>>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>>>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>>>> @@ -796,6 +796,12 @@ int coresight_get_cpu(struct device *dev)
>>>>   }
>>>>   EXPORT_SYMBOL_GPL(coresight_get_cpu);
>>>> +int coresight_get_static_trace_id(struct device *dev, u32 *id)
>>>> +{
>>>> +    return fwnode_property_read_u32(dev_fwnode(dev), "arm,static- 
>>>> trace-id", id);
>>>> +}
>>>> +EXPORT_SYMBOL_GPL(coresight_get_static_trace_id);
>>>> +
>>>>   struct coresight_platform_data *
>>>>   coresight_get_platform_data(struct device *dev)
>>>>   {
>>>> diff --git a/drivers/hwtracing/coresight/coresight-trace-id.c b/ 
>>>> drivers/hwtracing/coresight/coresight-trace-id.c
>>>> index d98e12cb30ec..df8fe50b413f 100644
>>>> --- a/drivers/hwtracing/coresight/coresight-trace-id.c
>>>> +++ b/drivers/hwtracing/coresight/coresight-trace-id.c
>>>> @@ -12,6 +12,12 @@
>>>>   #include "coresight-trace-id.h"
>>>> +enum trace_id_flags {
>>>> +    TRACE_ID_ANY = 0x0,
>>>> +    TRACE_ID_PREFER_ODD = 0x1,
>>>> +    TRACE_ID_REQ_STATIC = 0x2,
>>>> +};
>>>> +
>>>>   /* Default trace ID map. Used in sysfs mode and for system sources */
>>>>   static DEFINE_PER_CPU(atomic_t, id_map_default_cpu_ids) = 
>>>> ATOMIC_INIT(0);
>>>>   static struct coresight_trace_id_map id_map_default = {
>>>> @@ -74,16 +80,18 @@ static int coresight_trace_id_find_odd_id(struct 
>>>> coresight_trace_id_map *id_map)
>>>>    * Otherwise allocate next available ID.
>>>>    */
>>>>   static int coresight_trace_id_alloc_new_id(struct 
>>>> coresight_trace_id_map *id_map,
>>>> -                       int preferred_id, bool prefer_odd_id)
>>>> +                       int preferred_id, unsigned int flags)
>>>>   {
>>>>       int id = 0;
>>>>       /* for backwards compatibility, cpu IDs may use preferred 
>>>> value */
>>>> -    if (IS_VALID_CS_TRACE_ID(preferred_id) &&
>>>> -        !test_bit(preferred_id, id_map->used_ids)) {
>>>> -        id = preferred_id;
>>>> -        goto trace_id_allocated;
>>>> -    } else if (prefer_odd_id) {
>>>> +    if (IS_VALID_CS_TRACE_ID(preferred_id)) {
>>>> +        if (!test_bit(preferred_id, id_map->used_ids)) {
>>>> +            id = preferred_id;
>>>> +            goto trace_id_allocated;
>>>> +        } else if (flags & TRACE_ID_REQ_STATIC)
>>>> +            return -EINVAL;
>>>
>>> nit: EBUSY sounds like a better choice here ? Requested ID is not
>>> available.
>>>
>>> Additionally, do we need to handle a case where the preferred_id is
>>> not valid ? I think we silently allocate a new trace id in such case ?
>>>
>>> Rest looks good to me.
>>>
>>> Suzuki
>>>
>>>
>>
>> If preferred_id is not valid, a random id will be allocated for the 
>> source. For the normal source, 0 is provided here as preferred_id.
> 
> My point is : If preferred_id is not valid *and* TRACE_ID_REQ_STATIC
> flag is set, we go ahead and allocate a random different ID. Which is
> not preferred. We should return -EINVAL in that case.
> 
> Suzuki
> 
> 
Got your point. Thanks.

>>
>>>
>>>
>>>> +    } else if (flags & TRACE_ID_PREFER_ODD) {
>>>>       /* may use odd ids to avoid preferred legacy cpu IDs */
>>>>           id = coresight_trace_id_find_odd_id(id_map);
>>>>           if (id)
>>>> @@ -153,7 +161,7 @@ static int _coresight_trace_id_get_cpu_id(int 
>>>> cpu, struct coresight_trace_id_map
>>>>        */
>>>>       id = coresight_trace_id_alloc_new_id(id_map,
>>>>                            CORESIGHT_LEGACY_CPU_TRACE_ID(cpu),
>>>> -                         false);
>>>> +                         TRACE_ID_ANY);
>>>>       if (!IS_VALID_CS_TRACE_ID(id))
>>>>           goto get_cpu_id_out_unlock;
>>>> @@ -188,14 +196,14 @@ static void _coresight_trace_id_put_cpu_id(int 
>>>> cpu, struct coresight_trace_id_ma
>>>>       DUMP_ID_MAP(id_map);
>>>>   }
>>>> -static int coresight_trace_id_map_get_system_id(struct 
>>>> coresight_trace_id_map *id_map)
>>>> +static int coresight_trace_id_map_get_system_id(struct 
>>>> coresight_trace_id_map *id_map,
>>>> +                    int preferred_id, unsigned int traceid_flags)
>>>>   {
>>>>       unsigned long flags;
>>>>       int id;
>>>>       spin_lock_irqsave(&id_map->lock, flags);
>>>> -    /* prefer odd IDs for system components to avoid legacy CPU IDS */
>>>> -    id = coresight_trace_id_alloc_new_id(id_map, 0, true);
>>>> +    id = coresight_trace_id_alloc_new_id(id_map, preferred_id, 
>>>> traceid_flags);
>>>>       spin_unlock_irqrestore(&id_map->lock, flags);
>>>>       DUMP_ID(id);
>>>> @@ -255,10 +263,19 @@ 
>>>> EXPORT_SYMBOL_GPL(coresight_trace_id_read_cpu_id_map);
>>>>   int coresight_trace_id_get_system_id(void)
>>>>   {
>>>> -    return coresight_trace_id_map_get_system_id(&id_map_default);
>>>> +    /* prefer odd IDs for system components to avoid legacy CPU IDS */
>>>> +    return coresight_trace_id_map_get_system_id(&id_map_default, 0,
>>>> +            TRACE_ID_PREFER_ODD);
>>>>   }
>>>>   EXPORT_SYMBOL_GPL(coresight_trace_id_get_system_id);
>>>> +int coresight_trace_id_get_static_system_id(int trace_id)
>>>> +{
>>>> +    return coresight_trace_id_map_get_system_id(&id_map_default,
>>>> +            trace_id, TRACE_ID_REQ_STATIC);
>>>> +}
>>>> +EXPORT_SYMBOL_GPL(coresight_trace_id_get_static_system_id);
>>>> +
>>>>   void coresight_trace_id_put_system_id(int id)
>>>>   {
>>>>       coresight_trace_id_map_put_system_id(&id_map_default, id);
>>>> diff --git a/drivers/hwtracing/coresight/coresight-trace-id.h b/ 
>>>> drivers/hwtracing/coresight/coresight-trace-id.h
>>>> index 9aae50a553ca..db68e1ec56b6 100644
>>>> --- a/drivers/hwtracing/coresight/coresight-trace-id.h
>>>> +++ b/drivers/hwtracing/coresight/coresight-trace-id.h
>>>> @@ -116,6 +116,15 @@ int coresight_trace_id_read_cpu_id_map(int cpu, 
>>>> struct coresight_trace_id_map *i
>>>>    */
>>>>   int coresight_trace_id_get_system_id(void);
>>>> +/**
>>>> + * Allocate a CoreSight static trace ID for a system component.
>>>> + *
>>>> + * Used to allocate static IDs for system trace sources such as 
>>>> dummy source.
>>>> + *
>>>> + * return: Trace ID or -EINVAL if allocation is impossible.
>>>> + */
>>>> +int coresight_trace_id_get_static_system_id(int id);
>>>> +
>>>>   /**
>>>>    * Release an allocated system trace ID.
>>>>    *
>>>> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
>>>> index c13342594278..129795873072 100644
>>>> --- a/include/linux/coresight.h
>>>> +++ b/include/linux/coresight.h
>>>> @@ -662,6 +662,7 @@ void coresight_relaxed_write64(struct 
>>>> coresight_device *csdev,
>>>>   void coresight_write64(struct coresight_device *csdev, u64 val, 
>>>> u32 offset);
>>>>   extern int coresight_get_cpu(struct device *dev);
>>>> +extern int coresight_get_static_trace_id(struct device *dev, u32 *id);
>>>>   struct coresight_platform_data *coresight_get_platform_data(struct 
>>>> device *dev);
>>>>   struct coresight_connection *
>>>
>>
> 


  reply	other threads:[~2024-11-19 10:38 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-04 11:56 [PATCH v6 0/3] coresight: Add static trace id support Mao Jinlong
2024-11-04 11:56 ` [PATCH v6 1/3] dt-bindings: arm: Add arm,static-trace-id for coresight dummy source Mao Jinlong
2024-11-04 11:56 ` [PATCH v6 2/3] coresight: Add support to get static id for system trace sources Mao Jinlong
2024-11-13 13:27   ` Suzuki K Poulose
2024-11-18 10:14     ` Jinlong Mao
2024-11-18 10:18       ` Suzuki K Poulose
2024-11-19 10:38         ` Jinlong Mao [this message]
2024-11-04 11:56 ` [PATCH v6 3/3] coresight: dummy: Add static trace id support for dummy source Mao Jinlong

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=ae1d4e9a-6273-407e-952c-aec38e5c7a4f@quicinc.com \
    --to=quic_jinlmao@quicinc.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=conor+dt@kernel.org \
    --cc=coresight@lists.linaro.org \
    --cc=devicetree@vger.kernel.org \
    --cc=james.clark@arm.com \
    --cc=krzk+dt@kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mike.leach@linaro.org \
    --cc=robh@kernel.org \
    --cc=suzuki.poulose@arm.com \
    /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