[PATCH] define cpu_logical_id() always

[PATCH] define cpu_logical_id() always

[Bjorn Helgaas]

Define cpu_logical_id() even when !SMP.  I added uses of this in some
iosapic printk's, which broke the UP build.

=== include/asm-ia64/smp.h 1.17 vs edited ==--- 1.17/include/asm-ia64/smp.h	2004-05-14 20:00:12 -06:00
+++ edited/include/asm-ia64/smp.h	2004-06-21 14:44:12 -06:00
@@ -123,5 +123,9 @@
 extern void lock_ipi_calllock(void);
 extern void unlock_ipi_calllock(void);
 
+#else
+
+#define cpu_logical_id(cpuid)		0
+
 #endif /* CONFIG_SMP */
 #endif /* _ASM_IA64_SMP_H */
=== include/asm-ia64/sn/sn_cpuid.h 1.8 vs edited ==--- 1.8/include/asm-ia64/sn/sn_cpuid.h	2004-02-03 22:39:51 -07:00
+++ edited/include/asm-ia64/sn/sn_cpuid.h	2004-06-21 14:53:37 -06:00
@@ -84,7 +84,6 @@
  */
 
 #ifndef CONFIG_SMP
-#define cpu_logical_id(cpu)				0
 #define cpu_physical_id(cpuid)			((ia64_getreg(_IA64_REG_CR_LID) >> 16) & 0xffff)
 #endif
 

Re: [PATCH] define cpu_logical_id() always

[David Mosberger]

>>>>> On Mon, 21 Jun 2004 14:59:39 -0600, Bjorn Helgaas <bjorn.helgaas@hp.com> said:

  Bjorn> Define cpu_logical_id() even when !SMP.  I added uses of this
  Bjorn> in some iosapic printk's, which broke the UP build.

I'm OK with the patch per se, but since there is no existing use of
the macro for UP, I'd prefer it to be added along with the UP-user(s).
Otherwise, somebody may come along and clean up the "unused" macro.

	--david

Re: [PATCH] define cpu_logical_id() always

[Bjorn Helgaas]

On Friday 25 June 2004 6:50 pm, David Mosberger wrote:
> >>>>> On Mon, 21 Jun 2004 14:59:39 -0600, Bjorn Helgaas <bjorn.helgaas@hp.com> said:
> 
>   Bjorn> Define cpu_logical_id() even when !SMP.  I added uses of this
>   Bjorn> in some iosapic printk's, which broke the UP build.
> 
> I'm OK with the patch per se, but since there is no existing use of
> the macro for UP, I'd prefer it to be added along with the UP-user(s).
> Otherwise, somebody may come along and clean up the "unused" macro.

I added the use (both UP and SMP) here:

	http://linux.bkbits.net:8080/linux-2.5/cset%4040b2da7fhqIb6rDAZoyBMBhz_djWwg

It makes the IOSAPIC setup printk look like this:

+       printk(KERN_INFO "GSI %u (%s, %s) -> CPU %d (0x%04x) vector %d\n",
+              gsi, (trigger = IOSAPIC_EDGE ? "edge" : "level"),
+              (polarity = IOSAPIC_POL_HIGH ? "high" : "low"),
+              cpu_logical_id(dest), dest, vector);

This broke the UP build.