[ltt-dev] move marker-control file into debugfs

[ltt-dev] move marker-control file into debugfs

[Gui Jianfeng]

Hi Mathieu,

I'd like to move "/proc/ltt" into debugfs. How do
you think of that? If you don't object, i'll post
a patch later, and update ltt-arm and other relative
tools.

-- 
Regards
Gui Jianfeng

[ltt-dev] move marker-control file into debugfs

[Mathieu Desnoyers]

* Gui Jianfeng (guijianfeng at cn.fujitsu.com) wrote:
> Hi Mathieu,
> 
> I'd like to move "/proc/ltt" into debugfs. How do
> you think of that? If you don't object, i'll post
> a patch later, and update ltt-arm and other relative
> tools.
> 

Yes, this should be done. Please have a look at the discussion I had
with Zhaolei on this. It was on the ltt-dev ML.

Basically, the idea is to change the semantics a bit. The following
directory tree would be great :

/mnt/debugfs/ltt/markers/
                         <channel_names>/
                                         <marker_names>/

And within each individual <marker_names>/ directory, we would have
those files :

enabled (write, takes Y, y, 1 or N, n, 0. See ltt-trace-control.c.)
        (read, prints 1 or 0)

We may want to remove the ability to connect markers to specific probes.
Markers would always be only connected to the default serializer.
Markers using ltt-type-serializer does not use the probe pointer anyway.

We will probably want to let the
kernel/marker.c:add_marker()/remove_marker() code call into the debugfs
interface code to update the directory tree accordingly.

Regards,

Mathieu


> -- 
> Regards
> Gui Jianfeng
> 
> 
> _______________________________________________
> ltt-dev mailing list
> ltt-dev at lists.casi.polymtl.ca
> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
> 

-- 
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68

[ltt-dev] move marker-control file into debugfs

[Gui Jianfeng]

Mathieu Desnoyers wrote:
> * Gui Jianfeng (guijianfeng at cn.fujitsu.com) wrote:
>> Hi Mathieu,
>>
>> I'd like to move "/proc/ltt" into debugfs. How do
>> you think of that? If you don't object, i'll post
>> a patch later, and update ltt-arm and other relative
>> tools.
>>
> 
> Yes, this should be done. Please have a look at the discussion I had
> with Zhaolei on this. It was on the ltt-dev ML.
> 
> Basically, the idea is to change the semantics a bit. The following
> directory tree would be great :
> 
> /mnt/debugfs/ltt/markers/
>                          <channel_names>/
>                                          <marker_names>/
> 
> And within each individual <marker_names>/ directory, we would have
> those files :
> 
> enabled (write, takes Y, y, 1 or N, n, 0. See ltt-trace-control.c.)
>         (read, prints 1 or 0)
> 
> We may want to remove the ability to connect markers to specific probes.
> Markers would always be only connected to the default serializer.
> Markers using ltt-type-serializer does not use the probe pointer anyway.
> 
> We will probably want to let the
> kernel/marker.c:add_marker()/remove_marker() code call into the debugfs
> interface code to update the directory tree accordingly.

  Thanks for your comments, i'll see what i can do.

> 
> Regards,
> 
> Mathieu
> 
> 
>> -- 
>> Regards
>> Gui Jianfeng
>>
>>
>> _______________________________________________
>> ltt-dev mailing list
>> ltt-dev at lists.casi.polymtl.ca
>> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>>
> 

-- 
Regards
Gui Jianfeng

[ltt-dev] move marker-control file into debugfs

[Gui Jianfeng]

Hi Mathieu,

Lai and I will be on vocation unitl 2/2, might be untouched.
It's chinese new year.
I'll post a patch to implement debugfs marker-control then.

Gui Jianfeng wrote:
> Mathieu Desnoyers wrote:
>> * Gui Jianfeng (guijianfeng at cn.fujitsu.com) wrote:
>>> Hi Mathieu,
>>>
>>> I'd like to move "/proc/ltt" into debugfs. How do
>>> you think of that? If you don't object, i'll post
>>> a patch later, and update ltt-arm and other relative
>>> tools.
>>>
>> Yes, this should be done. Please have a look at the discussion I had
>> with Zhaolei on this. It was on the ltt-dev ML.
>>
>> Basically, the idea is to change the semantics a bit. The following
>> directory tree would be great :
>>
>> /mnt/debugfs/ltt/markers/
>>                          <channel_names>/
>>                                          <marker_names>/
>>
>> And within each individual <marker_names>/ directory, we would have
>> those files :
>>
>> enabled (write, takes Y, y, 1 or N, n, 0. See ltt-trace-control.c.)
>>         (read, prints 1 or 0)
>>
>> We may want to remove the ability to connect markers to specific probes.
>> Markers would always be only connected to the default serializer.
>> Markers using ltt-type-serializer does not use the probe pointer anyway.
>>
>> We will probably want to let the
>> kernel/marker.c:add_marker()/remove_marker() code call into the debugfs
>> interface code to update the directory tree accordingly.
> 
>   Thanks for your comments, i'll see what i can do.
> 
>> Regards,
>>
>> Mathieu
>>
>>
>>> -- 
>>> Regards
>>> Gui Jianfeng
>>>
>>>
>>> _______________________________________________
>>> ltt-dev mailing list
>>> ltt-dev at lists.casi.polymtl.ca
>>> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>>>
> 

-- 
Regards
Gui Jianfeng