[PATCH] DECLARE_PER_CPU/DEFINE_PER_CPU patch

This old __per_cpu_data define wasn't enough if an arch wants to
use the gcc __thread prefix (thread local storage), which needs to
go *before* the type in the definition.  So we have to go for a
DECLARE macro, and while we're there, separate DECLARE and DEFINE,
as definitions of per-cpu data cannot live in modules.  This also
means that accidental direct references to per-cpu variables will
be caught at compile time.

arch/ia64/kernel/perfmon.c

@@ -375,8 +375,8 @@ static pmu_config_t	pmu_conf; 	/* PMU configuration */
 static pfm_session_t	pfm_sessions;	/* global sessions information */
 static struct proc_dir_entry *perfmon_dir; /* for debug only */
 static pfm_stats_t	pfm_stats;
-int __per_cpu_data pfm_syst_wide;
-static int __per_cpu_data pfm_dcr_pp;
+DEFINE_PER_CPU(int, pfm_syst_wide);
+static DEFINE_PER_CPU(int, pfm_dcr_pp);
 
 /* sysctl() controls */
 static pfm_sysctl_t pfm_sysctl;

arch/ia64/kernel/setup.c

@@ -58,7 +58,7 @@ extern char _end;
 unsigned long __per_cpu_offset[NR_CPUS];
 #endif
 
-struct cpuinfo_ia64 cpu_info __per_cpu_data;
+DECLARE_PER_CPU(struct cpuinfo_ia64, cpu_info);
 
 unsigned long ia64_phys_stacked_size_p8;
 unsigned long ia64_cycles_per_usec;

arch/ia64/kernel/smp.c

@@ -80,7 +80,7 @@ static volatile struct call_data_struct *call_data;
 #define IPI_CPU_STOP		1
 
 /* This needs to be cacheline aligned because it is written to by *other* CPUs.  */
-static __u64 ipi_operation __per_cpu_data ____cacheline_aligned;
+static DECLARE_PER_CPU(__u64, ipi_operation) ____cacheline_aligned;
 
 static void
 stop_this_cpu (void)

include/asm-generic/percpu.h

 #ifndef _ASM_GENERIC_PERCPU_H_
 #define _ASM_GENERIC_PERCPU_H_
+#include <linux/compiler.h>
 
 #define __GENERIC_PER_CPU
-#include <linux/compiler.h>
+#ifdef CONFIG_SMP
 
 extern unsigned long __per_cpu_offset[NR_CPUS];
 
+/* Separate out the type, so (int[3], foo) works. */
+#ifndef MODULE
+#define DEFINE_PER_CPU(type, name) \
+    __attribute__((__section__(".percpu"))) __typeof__(type) name##__per_cpu
+#endif
+
 /* var is in discarded region: offset to particular copy we want */
-#define per_cpu(var, cpu) (*RELOC_HIDE(&var, __per_cpu_offset[cpu]))
+#define per_cpu(var, cpu) (*RELOC_HIDE(&var##__per_cpu, __per_cpu_offset[cpu]))
 #define __get_cpu_var(var) per_cpu(var, smp_processor_id())
 
+#else /* ! SMP */
+
+/* Can't define per-cpu variables in modules.  Sorry --RR */
+#ifndef MODULE
+#define DEFINE_PER_CPU(type, name) \
+    __typeof__(type) name##__per_cpu
+#endif
+
+#define per_cpu(var, cpu)			var##__per_cpu
+#define __get_cpu_var(var)			var##__per_cpu
+#endif
+
+#define DECLARE_PER_CPU(type, name) extern __typeof__(type) name##__per_cpu
+
 #endif /* _ASM_GENERIC_PERCPU_H_ */

include/asm-ia64/percpu.h

@@ -8,7 +8,7 @@
 
 #ifdef __ASSEMBLY__
 
-#define THIS_CPU(var)	(var)	/* use this to mark accesses to per-CPU variables... */
+#define THIS_CPU(var)	(var##__per_cpu)	/* use this to mark accesses to per-CPU variables... */
 
 #else /* !__ASSEMBLY__ */
 
@@ -16,8 +16,14 @@
 
 extern unsigned long __per_cpu_offset[NR_CPUS];
 
-#define per_cpu(var, cpu)	(*(__typeof__(&(var))) ((void *) &(var) + __per_cpu_offset[cpu]))
-#define __get_cpu_var(var)	(var)
+#ifndef MODULE
+#define DEFINE_PER_CPU(type, name) \
+    __attribute__((__section__(".data.percpu"))) __typeof__(type) name##__per_cpu
+#endif
+#define DECLARE_PER_CPU(type, name) extern __typeof__(type) name##__per_cpu
+
+#define per_cpu(var, cpu) (*RELOC_HIDE(&var##__per_cpu, __per_cpu_offset[cpu]))
+#define __get_cpu_var(var)	(var##__per_cpu)
 
 #endif /* !__ASSEMBLY__ */
 

include/asm-ia64/processor.h

@@ -134,7 +134,7 @@ struct ia64_psr {
  * CPU type, hardware bug flags, and per-CPU state.  Frequently used
  * state comes earlier:
  */
-extern struct cpuinfo_ia64 {
+struct cpuinfo_ia64 {
 	/* irq_stat must be 64-bit aligned */
 	union {
 		struct {
@@ -175,7 +175,9 @@ extern struct cpuinfo_ia64 {
 	__u64 prof_counter;
 	__u64 prof_multiplier;
 #endif
-} cpu_info __per_cpu_data;
+};
+
+DECLARE_PER_CPU(struct cpuinfo_ia64, cpu_info);
 
 /*
  * The "local" data pointer.  It points to the per-CPU data of the currently executing

include/linux/percpu.h

 #ifndef __LINUX_PERCPU_H
 #define __LINUX_PERCPU_H
-#include <linux/config.h>
-
-#ifdef CONFIG_SMP
-#define __per_cpu_data	__attribute__((section(".data.percpu")))
+#include <linux/spinlock.h> /* For preempt_disable() */
 #include <asm/percpu.h>
-#else
-#define __per_cpu_data
-#define per_cpu(var, cpu)			var
-#define __get_cpu_var(var)			var
-#endif
 
 #define get_cpu_var(var) ({ preempt_disable(); __get_cpu_var(var); })
 #define put_cpu_var(var) preempt_enable()
+
 #endif /* __LINUX_PERCPU_H */

kernel/softirq.c

@@ -151,8 +151,8 @@ struct tasklet_head
 
 /* Some compilers disobey section attribute on statics when not
    initialized -- RR */
-static struct tasklet_head tasklet_vec __per_cpu_data = { NULL };
-static struct tasklet_head tasklet_hi_vec __per_cpu_data = { NULL };
+static DEFINE_PER_CPU(struct tasklet_head, tasklet_vec) = { NULL };
+static DEFINE_PER_CPU(struct tasklet_head, tasklet_hi_vec) = { NULL };
 
 void __tasklet_schedule(struct tasklet_struct *t)
 {