Re: 2.6.14-rc4-rt1

[Ingo Molnar <mingo@elte.hu>]

* Andi Kleen <ak@suse.de> wrote:

> > > I am getting similar segfault on boot problem on 2.6.14-rc4-rt1 on my 
> > > x86-64 box (with LATENCY_TRACE).
> > 
> > > INIT: version 2.86 booting
> > > hotplug[877]: segfault at ffffffff8010f588 rip ffffffff8010f588 rsp
> > > 00007fffff8bee68 error 15
> > 
> > what does the ffffffff8010f588 RIP address map to? You can find out by 
> 
> It could be any kernel address that someone injected into user space.  
> Most likely some problem with the vsyscall page with either signal 
> handling or gettimeofday. vsyscall code is tricky to hack because you 
> cannot add any new functions there, just inlines, otherwise the code 
> won't end up the right section.

ah, indeed - i completely forgot about vsyscalls - they must not be 
traced. Badari, does the patch below help?

	Ingo

Index: linux/arch/x86_64/kernel/vsyscall.c
===================================================================
--- linux.orig/arch/x86_64/kernel/vsyscall.c
+++ linux/arch/x86_64/kernel/vsyscall.c
@@ -34,7 +34,7 @@
 #include <asm/errno.h>
 #include <asm/io.h>
 
-#define __vsyscall(nr) __attribute__ ((unused,__section__(".vsyscall_" #nr)))
+#define __vsyscall(nr) __attribute__ ((unused,__section__(".vsyscall_" #nr))) notrace
 #define force_inline __attribute__((always_inline)) inline
 
 int __sysctl_vsyscall __section_sysctl_vsyscall = 1;