Re: [PATCH] Linux Kernel Markers 0.7 for 2.6.17 (with type checking!)

[Mathieu Desnoyers]

* Jeremy Fitzhardinge (jeremy@xxxxxxxx) wrote:
> Mathieu Desnoyers wrote:
> >+#ifdef CONFIG_MARK_SYMBOL
> >+#define MARK_SYM(name) \
> >+	do { \
> >+		__label__ here; \
> >+		here: asm volatile \
> >+			(MARK_KPROBE_PREFIX#name " = %0" : : "m" (*&&here)); 
> >\
> >+	} while(0)
> >+#else 
> >+#define MARK_SYM(name)
> >+#endif
> 
> BTW, this won't work if you put the MARK_SYM in a loop which gcc 
> unrolls; you'll only get the mark in the last unrolled iteration 
> (because the symbol assignments will override each other).
> 
> Do make this work properly, you really need to put the mark entries into 
> a separate section, so that if gcc duplicates the code, you get 
> duplicated markers too.
> 

Good point, I will change it to :

#define MARK_SYM(name) \
        do { \
                __label__ here; \
                volatile static void *__mark_kprobe_##name \
                        asm (MARK_CALL_PREFIX#name) \
                        __attribute__((unused)) = &&here; \
here: \
                do { } while(0); \
        } while(0)

Which fixes the problem. Some tests showed me that the compiler does not unroll
an otherwise unrolled loop when this specific macro is called. (test done with
-funroll-all-loops).

Regards,

Mathieu


OpenPGP public key:              http://krystal.dyndns.org:8080/key/compudj.gpg
Key fingerprint:     8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/