[ltt-dev] too few arguments to function 'ltt_vtrace'

[dwt@PolTec.COM (Dennis W. Tokarski)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mathieu Desnoyers wrote:
| * Mathieu Desnoyers (compudj at krystal.dyndns.org) wrote:
|> * Dennis W. Tokarski (dwt at PolTec.COM) wrote:
| Hi Matthieu,
|
| I've just tried to build a kernel based on Fedora's
| kernel-2.6.27.9-150.fc10.src.rpm. Your patch set from
| patch-2.6.27.9-lttng-0.70.tar.bz2 applied with minimal
| conflicts which were easily resolved.
|
| However the error message in the subject above occurs
| during build.
|
| In ltt_serialize.c, ltt_vtrace() is defined as requiring
| five arguments, while in ltt-ftrace.c ltt_tap_marker()
| calls it with four--you aren't providing the required
| pointer to a struct marker as the first arg. It's not
| at all clear from the surrounding code what ltt_tap_marker()
| *should* be passing, but ltt_vtrace() certainly wants to
| use the arg.
|
| Same deal with is_marker_enabled() in kernel/marker.c,
| you call it from markers_show() in fs/proc/base.c without
| the first arg.
|
| A quick look at your web repository shows the second
| bug fixed, but the first is still present in the
| head revision.
|
| So how do you get a kernel to build at all, much less
| run and actually take a trace?  What actual revisions
| are you using for your work?
|
| I'd really like to give lttng a try, but there is
| this little show-stopper.
|
|>> As a quick solution :
|>> Try disabling ftrace :)
|>>
|>> Or remove the line
|>>
|>> obj-$(CONFIG_LTT_FTRACE)                += ltt-ftrace.o
|>>
|>> From ltt/Makefile
|>>
|>> The ftrace function tracing bits are bogus in 2.6.27 anyway. Therefore
|>> you won't miss much. I'll fix it in 2.6.28 though, thanks for the
|>> report!
|>>
|
|> This should now be fixed in LTTng 0.85.
|

Your suggestion to turn off ftrace worked, Mathieu, thank you.

I've installed the kernel on a qemu-i386 VM with a very spare
selection of things from the stock Fedora 10 distro, added
the lttctl tools, and taken a few traces.

It works pretty well.  lttv on its face looks pretty much
like I remember it from a couple of years ago, though I see
a few things have changed under the hood.

A different problem popped up on the way to trying a
trace on a dual processor qemu.  I'll offer a patch in
a separate post.

- --Dennis


|> Thanks,
|
|> Mathieu
|
|>> Mathieu
|>>
| Thanks for your help!
| --Dennis
|
|>>
_______________________________________________
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 mailing list
|> ltt-dev at lists.casi.polymtl.ca
|> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
|>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFJhRvwDJucIzCImrURAoomAJ91M8KJLC7UnWVAmlCC9VeO+rXS5gCfRmny
gbppuo95aRbdov50/jOa05g=
=wVN2
-----END PGP SIGNATURE-----