This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: stap fails if !CONFIG_FRAME_POINTER && !CONFIG_UTRACE


On Thu, Jun 10, 2010 at 01:49:17PM +0200, Mark Wielaard wrote:
> On Thu, 2010-06-10 at 13:41 +0200, Mark Wielaard wrote:
> > I must admit that I am slightly confused how/why
> > #if defined(CONFIG_UPROBES) || defined(CONFIG_UPROBES_MODULE)
> > succeeds in that case though. How gets that defined if utrace isn't
> > there?
> 
> O wait, I see, it doesn't get defined, we catch the #else:

Right.

> #if defined(CONFIG_UPROBES) || defined(CONFIG_UPROBES_MODULE)
> #include <linux/uprobes.h>
> #else
> #include "uprobes/uprobes.h"
> #endif
> 
> [... actual autoconf-test ...]
> 
> 
> So, in that case, the correct order of tests seems to be:
> 
> 
> #if defined(CONFIG_UPROBES) || defined(CONFIG_UPROBES_MODULE)
>   /* kernel built-in uprobes found */
>   #include <linux/uprobes.h>
> #else
>   /* use systemtap local copy, but only if we also have utrace
>   #if defined(CONFIG_UTRACE)
>     #include <linux/uprobes.h>
>   #else
>     #error "CONFIG_UTRACE not enabled for systemtap uprobes"
>   #endif
> #endif
> 
> [... actual autoconf-test ...]

This works also, lmk if you want a revised patch.

Tony


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]