[MIPS] Fix deadlock on MP with cache aliases.

A proper fix would involve introducing the notion of shared caches but
at this stage of 2.6.17 that's going to be too intrusive and not needed
for current hardware; aside I think some discussion will be needed.

So for now on the affected SMP configurations which happen to suffer from
cache aliases we make use of the fact that a single cache will be shared
by all processors.  This solves the deadlock issue and will improve
performance by getting rid of the smp_call_function overhead.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c
index 4420191..570bc4e 100644
--- a/arch/mips/mm/c-r4k.c
+++ b/arch/mips/mm/c-r4k.c
@@ -29,6 +29,27 @@
 #include <asm/war.h>
 #include <asm/cacheflush.h> /* for run_uncached() */
 
+
+/*
+ * Special Variant of smp_call_function for use by cache functions:
+ *
+ *  o No return value
+ *  o collapses to normal function call on UP kernels
+ *  o collapses to normal function call on systems with a single shared
+ *    primary cache.
+ */
+static inline void r4k_on_each_cpu(void (*func) (void *info), void *info,
+                                   int retry, int wait)
+{
+	preempt_disable();
+
+#if !defined(CONFIG_MIPS_MT_SMP) && !defined(CONFIG_MIPS_MT_SMTC)
+	smp_call_function(func, info, retry, wait);
+#endif
+	func(info);
+	preempt_enable();
+}
+
 /*
  * Must die.
  */
@@ -299,7 +320,7 @@
 	if (!cpu_has_dc_aliases)
 		return;
 
-	on_each_cpu(local_r4k_flush_cache_all, NULL, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_cache_all, NULL, 1, 1);
 }
 
 static inline void local_r4k___flush_cache_all(void * args)
@@ -320,7 +341,7 @@
 
 static void r4k___flush_cache_all(void)
 {
-	on_each_cpu(local_r4k___flush_cache_all, NULL, 1, 1);
+	r4k_on_each_cpu(local_r4k___flush_cache_all, NULL, 1, 1);
 }
 
 static inline void local_r4k_flush_cache_range(void * args)
@@ -341,7 +362,7 @@
 static void r4k_flush_cache_range(struct vm_area_struct *vma,
 	unsigned long start, unsigned long end)
 {
-	on_each_cpu(local_r4k_flush_cache_range, vma, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_cache_range, vma, 1, 1);
 }
 
 static inline void local_r4k_flush_cache_mm(void * args)
@@ -370,7 +391,7 @@
 	if (!cpu_has_dc_aliases)
 		return;
 
-	on_each_cpu(local_r4k_flush_cache_mm, mm, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_cache_mm, mm, 1, 1);
 }
 
 struct flush_cache_page_args {
@@ -461,7 +482,7 @@
 	args.addr = addr;
 	args.pfn = pfn;
 
-	on_each_cpu(local_r4k_flush_cache_page, &args, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_cache_page, &args, 1, 1);
 }
 
 static inline void local_r4k_flush_data_cache_page(void * addr)
@@ -471,7 +492,7 @@
 
 static void r4k_flush_data_cache_page(unsigned long addr)
 {
-	on_each_cpu(local_r4k_flush_data_cache_page, (void *) addr, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_data_cache_page, (void *) addr, 1, 1);
 }
 
 struct flush_icache_range_args {
@@ -514,7 +535,7 @@
 	args.start = start;
 	args.end = end;
 
-	on_each_cpu(local_r4k_flush_icache_range, &args, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_icache_range, &args, 1, 1);
 	instruction_hazard();
 }
 
@@ -590,7 +611,7 @@
 	args.vma = vma;
 	args.page = page;
 
-	on_each_cpu(local_r4k_flush_icache_page, &args, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_icache_page, &args, 1, 1);
 }
 
 
@@ -689,7 +710,7 @@
 
 static void r4k_flush_cache_sigtramp(unsigned long addr)
 {
-	on_each_cpu(local_r4k_flush_cache_sigtramp, (void *) addr, 1, 1);
+	r4k_on_each_cpu(local_r4k_flush_cache_sigtramp, (void *) addr, 1, 1);
 }
 
 static void r4k_flush_icache_all(void)