[PATCH][RFC] fixing the irqwcpustat mess

From: Christoph Hellwig (hch@lst.de)
Date: Mon May 05 2003 - 14:25:46 EST


currently only x86_64 and ia64 don't use the generic irq_cpustat code
and both have to workaround it's brokenness for the non-default case.

x86_64 defines an empty irq_cpustat_t even if it doesn't need one and
ia64 adds CONFIG_IA64 ifdefs around all users. What about this patch
instead to make __ARCH_IRQ_STAT useable?

--- 1.11/include/asm-ia64/hardirq.h Wed Feb 5 06:03:22 2003
+++ edited/include/asm-ia64/hardirq.h Tue Apr 22 22:03:09 2003
@@ -16,6 +16,9 @@
 /*
  * No irq_cpustat_t for IA-64. The data is held in the per-CPU data structure.
  */
+
+#define __ARCH_IRQ_STAT 1
+
 #define softirq_pending(cpu) (cpu_data(cpu)->softirq_pending)
 #define syscall_count(cpu) /* unused on IA-64 */
 #define ksoftirqd_task(cpu) (cpu_data(cpu)->ksoftirqd)
--- 1.3/include/asm-x86_64/hardirq.h Tue Nov 12 18:13:39 2002
+++ edited/include/asm-x86_64/hardirq.h Tue Apr 22 22:03:10 2003
@@ -12,10 +12,6 @@
    special access macros. This would generate better code. */
 #define __IRQ_STAT(cpu,member) (read_pda(me)->member)
 
-typedef struct {
- /* Empty. All the fields have moved to the PDA. */
-} irq_cpustat_t;
-
 #include <linux/irq_cpustat.h> /* Standard mappings for irq_cpustat_t above */
 
 /*
--- 1.7/include/linux/irq_cpustat.h Mon Mar 31 18:21:49 2003
+++ edited/include/linux/irq_cpustat.h Tue Apr 22 22:03:10 2003
@@ -17,9 +17,8 @@
  * definitions instead of differing sets for each arch.
  */
 
-extern irq_cpustat_t irq_stat[]; /* defined in asm/hardirq.h */
-
-#ifndef __ARCH_IRQ_STAT /* Some architectures can do this more efficiently */
+#ifndef __ARCH_IRQ_STAT
+extern irq_cpustat_t irq_stat[]; /* defined in asm/hardirq.h */
 #ifdef CONFIG_SMP
 #define __IRQ_STAT(cpu, member) (irq_stat[cpu].member)
 #else
@@ -31,8 +30,11 @@
 #define softirq_pending(cpu) __IRQ_STAT((cpu), __softirq_pending)
 #define local_softirq_pending() softirq_pending(smp_processor_id())
 #define syscall_count(cpu) __IRQ_STAT((cpu), __syscall_count)
+#define local_syscall_count() syscall_count(smp_processor_id())
 #define ksoftirqd_task(cpu) __IRQ_STAT((cpu), __ksoftirqd_task)
+#define local_ksoftirqd_task() ksoftirqd_task(smp_processor_id())
+
   /* arch dependent irq_stat fields */
-#define nmi_count(cpu) __IRQ_STAT((cpu), __nmi_count) /* i386, ia64 */
+#define nmi_count(cpu) __IRQ_STAT((cpu), __nmi_count) /* i386 */
 
 #endif /* __irq_cpustat_h */
--- 1.189/kernel/ksyms.c Sat Apr 12 18:37:08 2003
+++ edited/kernel/ksyms.c Tue Apr 22 22:03:10 2003
@@ -400,7 +400,6 @@
 EXPORT_SYMBOL(del_timer);
 EXPORT_SYMBOL(request_irq);
 EXPORT_SYMBOL(free_irq);
-EXPORT_SYMBOL(irq_stat);
 
 /* waitqueue handling */
 EXPORT_SYMBOL(add_wait_queue);
--- 1.39/kernel/softirq.c Wed Apr 16 09:51:47 2003
+++ edited/kernel/softirq.c Tue Apr 22 22:03:10 2003
@@ -33,7 +33,10 @@
    - Tasklets: serialized wrt itself.
  */
 
+#ifndef __ARCH_IRQ_STAT
 irq_cpustat_t irq_stat[NR_CPUS] ____cacheline_aligned;
+EXPORT_SYMBOL(irq_stat);
+#endif
 
 static struct softirq_action softirq_vec[32] __cacheline_aligned_in_smp;
 
@@ -321,7 +324,7 @@
         __set_current_state(TASK_INTERRUPTIBLE);
         mb();
 
- ksoftirqd_task(cpu) = current;
+ local_ksoftirqd_task() = current;
 
         for (;;) {
                 if (!local_softirq_pending())
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Wed May 07 2003 - 22:00:23 EST