[Ocfs2-devel] [PATCH] ocfs2/trivial: Add default mask log for heartbeat.c

[Ocfs2-devel] [PATCH] ocfs2/trivial: Add default mask log for heartbeat.c

[Tao Ma]

I just noticed today that we have no default mask
prefix for fs/ocfs2/cluster/heartbeat.c, but we have
2 places that use "mlog(0,...)". So add the default
prefix so that these "mlog(0, ...)" can work.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
 fs/ocfs2/cluster/heartbeat.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index c452d11..edec9bf 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -40,6 +40,7 @@
 #include "nodemanager.h"
 #include "quorum.h"
 
+#define MLOG_MASK_PREFIX ML_HEARTBEAT
 #include "masklog.h"
 
 
-- 
1.5.5

[Ocfs2-devel] [PATCH] ocfs2/trivial: Add default mask log for heartbeat.c

[Sunil Mushran]

Maybe better if we changed the two mlog(0,... to mlog(ML_HEARTBEAT,...

Now I don't know why we don't have a default mask in that file but instead
have distinct mlogs for HEARTBEAT and HB_BIO. One reason could have been to
limit the logging of hb_bio. If that is correct, then adding a default mask
will defeat the purpose.

Tao Ma wrote:
> I just noticed today that we have no default mask
> prefix for fs/ocfs2/cluster/heartbeat.c, but we have
> 2 places that use "mlog(0,...)". So add the default
> prefix so that these "mlog(0, ...)" can work.
>
> Signed-off-by: Tao Ma <tao.ma@oracle.com>
> ---
>  fs/ocfs2/cluster/heartbeat.c |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
> index c452d11..edec9bf 100644
> --- a/fs/ocfs2/cluster/heartbeat.c
> +++ b/fs/ocfs2/cluster/heartbeat.c
> @@ -40,6 +40,7 @@
>  #include "nodemanager.h"
>  #include "quorum.h"
>  
> +#define MLOG_MASK_PREFIX ML_HEARTBEAT
>  #include "masklog.h"
>  
>  
>   

[Ocfs2-devel] [PATCH] ocfs2/trivial: Add default mask log for heartbeat.c

[Tao Ma]

Sunil Mushran wrote:
> Maybe better if we changed the two mlog(0,... to mlog(ML_HEARTBEAT,...
> 
> Now I don't know why we don't have a default mask in that file but instead
> have distinct mlogs for HEARTBEAT and HB_BIO. One reason could have been to
> limit the logging of hb_bio. If that is correct, then adding a default mask
> will defeat the purpose.
yeah, I just got that two mlog(0,...) and tried to find the default mask 
and failed. Don't know the real reason for it. Will redo the patch as 
you like.

But the point I don't get is that why add a default mask will have an 
effect on mlog(ML_HB_BIO,...)?

Regards,
Tao
> 
> Tao Ma wrote:
>> I just noticed today that we have no default mask
>> prefix for fs/ocfs2/cluster/heartbeat.c, but we have
>> 2 places that use "mlog(0,...)". So add the default
>> prefix so that these "mlog(0, ...)" can work.
>>
>> Signed-off-by: Tao Ma <tao.ma@oracle.com>
>> ---
>>  fs/ocfs2/cluster/heartbeat.c |    1 +
>>  1 files changed, 1 insertions(+), 0 deletions(-)
>>
>> diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
>> index c452d11..edec9bf 100644
>> --- a/fs/ocfs2/cluster/heartbeat.c
>> +++ b/fs/ocfs2/cluster/heartbeat.c
>> @@ -40,6 +40,7 @@
>>  #include "nodemanager.h"
>>  #include "quorum.h"
>>  
>> +#define MLOG_MASK_PREFIX ML_HEARTBEAT
>>  #include "masklog.h"
>>  
>>  
>>   
> 

[Ocfs2-devel] [PATCH] ocfs2/trivial: Add default mask log for heartbeat.c

[Sunil Mushran]

Tao Ma wrote:
> Sunil Mushran wrote:
>> Maybe better if we changed the two mlog(0,... to mlog(ML_HEARTBEAT,...
>>
>> Now I don't know why we don't have a default mask in that file but 
>> instead
>> have distinct mlogs for HEARTBEAT and HB_BIO. One reason could have 
>> been to
>> limit the logging of hb_bio. If that is correct, then adding a 
>> default mask
>> will defeat the purpose.
> yeah, I just got that two mlog(0,...) and tried to find the default 
> mask and failed. Don't know the real reason for it. Will redo the 
> patch as you like.
>
> But the point I don't get is that why add a default mask will have an 
> effect on mlog(ML_HB_BIO,...)?

Reverse. As in, if we make that the default, then the mlogs marked hb_bio
will also print when we only want heartbeat.

Now how does that matter? Probably very little, if at all. But we will be
revisiting tracing/logging sometime next year. So I would make minimal
changes in this area right now.

[Ocfs2-devel] [PATCH] ocfs2/trivial: Add default mask log for heartbeat.c

[Tao Ma]

Sunil Mushran wrote:
> Tao Ma wrote:
>> Sunil Mushran wrote:
>>> Maybe better if we changed the two mlog(0,... to mlog(ML_HEARTBEAT,...
>>>
>>> Now I don't know why we don't have a default mask in that file but 
>>> instead
>>> have distinct mlogs for HEARTBEAT and HB_BIO. One reason could have 
>>> been to
>>> limit the logging of hb_bio. If that is correct, then adding a 
>>> default mask
>>> will defeat the purpose.
>> yeah, I just got that two mlog(0,...) and tried to find the default 
>> mask and failed. Don't know the real reason for it. Will redo the 
>> patch as you like.
>>
>> But the point I don't get is that why add a default mask will have an 
>> effect on mlog(ML_HB_BIO,...)?
> 
> Reverse. As in, if we make that the default, then the mlogs marked hb_bio
> will also print when we only want heartbeat.
oh, I see. thanks for the explanation.
> 
> Now how does that matter? Probably very little, if at all. But we will be
> revisiting tracing/logging sometime next year. So I would make minimal
> changes in this area right now.
ok, no problem. I will regenerate the patch.

Regards,
Tao