perfctr-2.6.1 build fix for amd64 in 2.6 SMP kernels

perfctr-2.6.1 build fix for amd64 in 2.6 SMP kernels

[Mikael Pettersson]

This patch for perfctr-2.6.1 fixes a problem which causes a
compilation error for amd64 (aka x86-64) in 2.6 SMP kernels.

The problem is that cpu_isset(cpu,mask) requires that mask
is an lvalue (it uses test_bit()), but on amd64 I want to
pass in a constant (the mask is always empty). Oops.

Thanks to Bryan O'Sullivan for reporting the build error.

/Mikael

--- perfctr-2.6.1/linux/drivers/perfctr/cpumask.h.~1~	2003-09-07 22:15:38.000000000 +0200
+++ perfctr-2.6.1/linux/drivers/perfctr/cpumask.h	2003-10-07 20:19:52.000000000 +0200
@@ -38,6 +38,8 @@
    but cpumask_t compatibility issues forced it to be moved here. */
 #if PERFCTR_CPUS_FORBIDDEN_MASK_NEEDED
 extern cpumask_t perfctr_cpus_forbidden_mask;
+#define perfctr_cpu_is_forbidden(cpu)	cpu_isset((cpu), perfctr_cpus_forbidden_mask)
 #else
 #define perfctr_cpus_forbidden_mask	CPU_MASK_NONE
+#define perfctr_cpu_is_forbidden(cpu)	0 /* cpu_isset() needs an lvalue :-( */
 #endif
--- perfctr-2.6.1/linux/drivers/perfctr/global.c.~1~	2003-10-02 22:04:35.000000000 +0200
+++ perfctr-2.6.1/linux/drivers/perfctr/global.c	2003-10-07 20:19:52.000000000 +0200
@@ -138,7 +138,7 @@
 		return ret;
 	if( cpu_control.cpu >= NR_CPUS ||
 	    !cpu_online(cpu_control.cpu) ||
-	    cpu_isset(cpu_control.cpu, perfctr_cpus_forbidden_mask) )
+	    perfctr_cpu_is_forbidden(cpu_control.cpu) )
 		return -EINVAL;
 	/* we don't permit i-mode counters */
 	if( cpu_control.cpu_control.nrictrs != 0 )