[lttng-dev] Sync instrumentation to today's linux 3.8

[lttng-dev] Sync instrumentation to today's linux 3.8

[Francis Giraldeau]

To prepare instrumentation for the next linux release, here is the ticket I
was working on. It's mostly straight forward.

http://redmine.dorsal.polymtl.ca/issues/431

There is one issue with ext4 includes. I had to copy fs/ext4/ext4.h and
ext4_extent.h for the function ext4_ext_get_actual_len(). I added them to
instrumentation/mainline/fs/ext4 directory. I'm not sure if it's the best
solution.

Still testing to do, thought. BTW, is there a test setup for regression? I
would like to verify that it doesn't break any supported kernel and I don't
have the required setup.

With this, maybe lttng-modules 2.1 can target linux 3.8?

Cheers,

Francis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lttng.org/pipermail/lttng-dev/attachments/20130125/35d5d659/attachment.html>

[lttng-dev] Sync instrumentation to today's linux 3.8

[Mathieu Desnoyers]

* Francis Giraldeau (francis.giraldeau at gmail.com) wrote:
> To prepare instrumentation for the next linux release, here is the ticket I
> was working on. It's mostly straight forward.
> 
> http://redmine.dorsal.polymtl.ca/issues/431
> 
> There is one issue with ext4 includes. I had to copy fs/ext4/ext4.h and
> ext4_extent.h for the function ext4_ext_get_actual_len(). I added them to
> instrumentation/mainline/fs/ext4 directory. I'm not sure if it's the best
> solution.
> 
> Still testing to do, thought. BTW, is there a test setup for regression? I
> would like to verify that it doesn't break any supported kernel and I don't
> have the required setup.
> 
> With this, maybe lttng-modules 2.1 can target linux 3.8?

Please send the patch by email.

3.8 is still in rc stage. Removing ext4_ext_get_actual_len() seems like
a bad move from upstream Linux if it is needed in the instrumentation
headers. This should be fixed upstream rather than duplicated within our
own code.

Thanks,

Mathieu

> 
> Cheers,
> 
> Francis

> _______________________________________________
> lttng-dev mailing list
> lttng-dev at lists.lttng.org
> http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev


-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com

[lttng-dev] Sync instrumentation to today's linux 3.8

[Mathieu Desnoyers]

* Mathieu Desnoyers (mathieu.desnoyers at efficios.com) wrote:
> * Francis Giraldeau (francis.giraldeau at gmail.com) wrote:
> > To prepare instrumentation for the next linux release, here is the ticket I
> > was working on. It's mostly straight forward.
> > 
> > http://redmine.dorsal.polymtl.ca/issues/431
> > 
> > There is one issue with ext4 includes. I had to copy fs/ext4/ext4.h and
> > ext4_extent.h for the function ext4_ext_get_actual_len(). I added them to
> > instrumentation/mainline/fs/ext4 directory. I'm not sure if it's the best
> > solution.
> > 
> > Still testing to do, thought. BTW, is there a test setup for regression? I
> > would like to verify that it doesn't break any supported kernel and I don't
> > have the required setup.
> > 
> > With this, maybe lttng-modules 2.1 can target linux 3.8?

And FYI, ext4 tracepoints are only in lttng-modules master branch. They
were excluded from 2.1.x because they arrived too late in the 2.1 cycle.

Thanks,

Mathieu

> 
> Please send the patch by email.
> 
> 3.8 is still in rc stage. Removing ext4_ext_get_actual_len() seems like
> a bad move from upstream Linux if it is needed in the instrumentation
> headers. This should be fixed upstream rather than duplicated within our
> own code.
> 
> Thanks,
> 
> Mathieu
> 
> > 
> > Cheers,
> > 
> > Francis
> 
> > _______________________________________________
> > lttng-dev mailing list
> > lttng-dev at lists.lttng.org
> > http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev
> 
> 
> -- 
> Mathieu Desnoyers
> EfficiOS Inc.
> http://www.efficios.com
> 
> _______________________________________________
> lttng-dev mailing list
> lttng-dev at lists.lttng.org
> http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com