Messages in this thread |  | | Date | Thu, 17 Oct 2013 17:35:29 +0200 | From | Borislav Petkov <> | Subject | Re: [PATCH v2 0/9] Extended H/W error log driver |
| |
On Thu, Oct 17, 2013 at 11:25:41AM -0400, Steven Rostedt wrote: > On Thu, 17 Oct 2013 10:33:48 -0400 > Chen Gong <gong.chen@linux.intel.com> wrote: > > > > > Gong, can you try moving the CREATE_TRACE_POINTS line to a new file - > > > arch/x86/ras/ras.c and define it there and not anywhere else, i.e. move > > > it away from edac_mc.c. Does that help? > > > > In current kernel we haven't arch/x86/ras/ras.c. You mean I create > > a new one there and just add some trace macro definition? > > The CREATE_TRACE_POINTS will cause the TRACE_EVENT() macro to define > the functions used for tracing. If you have it defined more than once, > it will either cause a kernel compile error (variables declared more > than once), or if used in modules, will cause confusion (as well as > bloat) because different functions and variables will be defined for > the same tracepoint. > > If you need to create a separate file that you can have it defined in a > single place, then please do so!
Yes, the plan was this all along to have arch/x86/ras/ras.c or arch/x86/ras/core.c where CREATE_TRACE_POINTS is done, among other things.
Btw, Gong, there's an easier way to test this, if you wanna:
simply make sure EDAC is disabled in the kernel with your patches - i.e., CONFIG_EDAC is not set in your config. It should work then.
But, the final solution should be what Steve says with the file I'm suggesting :)
Thanks.
-- Regards/Gruss, Boris.
Sent from a fat crate under my desk. Formatting is fine. --
|  |