[NETFILTER]: Fix OOPSes on machines with discontiguous cpu numbering.

Original patch by Harald Welte, with feedback from Herbert Xu
and testing by S�bastien Bernard.

EBTABLES, ARP tables, and IP/IP6 tables all assume that cpus
are numbered linearly. That is not necessarily true.

This patch fixes that up by calculating the largest possible
cpu number, and allocating enough per-cpu structure space given
that.

Signed-off-by: David S. Miller <[e-mail]>

arch/cris/arch-v32/kernel/smp.c

···
 #include <linux/kernel.h>
 #include <linux/cpumask.h>
 #include <linux/interrupt.h>
+#include <linux/module.h>
 
 #define IPI_SCHEDULE 1
 #define IPI_CALL 2
···
 /* CPU masks */
 cpumask_t cpu_online_map = CPU_MASK_NONE;
 cpumask_t phys_cpu_present_map = CPU_MASK_NONE;
+EXPORT_SYMBOL(phys_cpu_present_map);
 
 /* Variables used during SMP boot */
 volatile int cpu_now_booting = 0;

arch/sh/kernel/smp.c

···
 #include <linux/time.h>
 #include <linux/timex.h>
 #include <linux/sched.h>
+#include <linux/module.h>
 
 #include <asm/atomic.h>
 #include <asm/processor.h>
···
 extern void per_cpu_trap_init(void);
 
 cpumask_t cpu_possible_map;
+EXPORT_SYMBOL(cpu_possible_map);
+
 cpumask_t cpu_online_map;
 static atomic_t cpus_booted = ATOMIC_INIT(0);
 

include/linux/cpumask.h

···
 #define for_each_online_cpu(cpu)  for_each_cpu_mask((cpu), cpu_online_map)
 #define for_each_present_cpu(cpu) for_each_cpu_mask((cpu), cpu_present_map)
 
+/* Find the highest possible smp_processor_id() */
+static inline unsigned int highest_possible_processor_id(void)
+{
+	unsigned int cpu, highest = 0;
+
+	for_each_cpu_mask(cpu, cpu_possible_map)
+		highest = cpu;
+
+	return highest;
+}
+
+
 #endif /* __LINUX_CPUMASK_H */

net/bridge/netfilter/ebtables.c

···
 #include <linux/spinlock.h>
 #include <asm/uaccess.h>
 #include <linux/smp.h>
+#include <linux/cpumask.h>
 #include <net/sock.h>
 /* needed for logical [in,out]-dev filtering */
 #include "../br_private.h"
···
 		/* this will get free'd in do_replace()/ebt_register_table()
 		   if an error occurs */
 		newinfo->chainstack = (struct ebt_chainstack **)
-		   vmalloc(num_possible_cpus() * sizeof(struct ebt_chainstack));
+		   vmalloc((highest_possible_processor_id()+1) 
+				   		* sizeof(struct ebt_chainstack));
 		if (!newinfo->chainstack)
 			return -ENOMEM;
-		for (i = 0; i < num_possible_cpus(); i++) {
+		for_each_cpu(i) {
 			newinfo->chainstack[i] =
 			   vmalloc(udc_cnt * sizeof(struct ebt_chainstack));
 			if (!newinfo->chainstack[i]) {
···
 
 	/* counters of cpu 0 */
 	memcpy(counters, oldcounters,
-	   sizeof(struct ebt_counter) * nentries);
+	       sizeof(struct ebt_counter) * nentries);
+
 	/* add other counters to those of cpu 0 */
-	for (cpu = 1; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
+		if (cpu == 0)
+			continue;
 		counter_base = COUNTER_BASE(oldcounters, nentries, cpu);
 		for (i = 0; i < nentries; i++) {
 			counters[i].pcnt += counter_base[i].pcnt;
···
 		BUGPRINT("Entries_size never zero\n");
 		return -EINVAL;
 	}
-	countersize = COUNTER_OFFSET(tmp.nentries) * num_possible_cpus();
+	countersize = COUNTER_OFFSET(tmp.nentries) * 
+					(highest_possible_processor_id()+1);
 	newinfo = (struct ebt_table_info *)
 	   vmalloc(sizeof(struct ebt_table_info) + countersize);
 	if (!newinfo)
···
 
 	vfree(table->entries);
 	if (table->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(table->chainstack[i]);
 		vfree(table->chainstack);
 	}
···
 	vfree(counterstmp);
 	/* can be initialized in translate_table() */
 	if (newinfo->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(newinfo->chainstack[i]);
 		vfree(newinfo->chainstack);
 	}
···
 		return -EINVAL;
 	}
 
-	countersize = COUNTER_OFFSET(table->table->nentries) * num_possible_cpus();
+	countersize = COUNTER_OFFSET(table->table->nentries) *
+					(highest_possible_processor_id()+1);
 	newinfo = (struct ebt_table_info *)
 	   vmalloc(sizeof(struct ebt_table_info) + countersize);
 	ret = -ENOMEM;
···
 	up(&ebt_mutex);
 free_chainstack:
 	if (newinfo->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(newinfo->chainstack[i]);
 		vfree(newinfo->chainstack);
 	}
···
 	up(&ebt_mutex);
 	vfree(table->private->entries);
 	if (table->private->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(table->private->chainstack[i]);
 		vfree(table->private->chainstack);
 	}

net/ipv4/netfilter/arp_tables.c

···
 	}
 
 	/* And one copy for every other CPU */
-	for (i = 1; i < num_possible_cpus(); i++) {
-		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size)*i,
+	for_each_cpu(i) {
+		if (i == 0)
+			continue;
+		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
 		       newinfo->entries,
 		       SMP_ALIGN(newinfo->size));
 	}
···
 	unsigned int cpu;
 	unsigned int i;
 
-	for (cpu = 0; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
 		i = 0;
 		ARPT_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
 				   t->size,
···
 		return -ENOMEM;
 
 	newinfo = vmalloc(sizeof(struct arpt_table_info)
-			  + SMP_ALIGN(tmp.size) * num_possible_cpus());
+			  + SMP_ALIGN(tmp.size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
···
 		= { 0, 0, 0, { 0 }, { 0 }, { } };
 
 	newinfo = vmalloc(sizeof(struct arpt_table_info)
-			  + SMP_ALIGN(repl->size) * num_possible_cpus());
+			  + SMP_ALIGN(repl->size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo) {
 		ret = -ENOMEM;
 		return ret;

net/ipv4/netfilter/ip_tables.c

···
 #include <asm/semaphore.h>
 #include <linux/proc_fs.h>
 #include <linux/err.h>
+#include <linux/cpumask.h>
 
 #include <linux/netfilter_ipv4/ip_tables.h>
 
···
 	}
 
 	/* And one copy for every other CPU */
-	for (i = 1; i < num_possible_cpus(); i++) {
-		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size)*i,
+	for_each_cpu(i) {
+		if (i == 0)
+			continue;
+		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
 		       newinfo->entries,
 		       SMP_ALIGN(newinfo->size));
 	}
···
 		struct ipt_entry *table_base;
 		unsigned int i;
 
-		for (i = 0; i < num_possible_cpus(); i++) {
+		for_each_cpu(i) {
 			table_base =
 				(void *)newinfo->entries
 				+ TABLE_OFFSET(newinfo, i);
···
 	unsigned int cpu;
 	unsigned int i;
 
-	for (cpu = 0; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
 		i = 0;
 		IPT_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
 				  t->size,
···
 		return -ENOMEM;
 
 	newinfo = vmalloc(sizeof(struct ipt_table_info)
-			  + SMP_ALIGN(tmp.size) * num_possible_cpus());
+			  + SMP_ALIGN(tmp.size) * 
+			  	(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
···
 		= { 0, 0, 0, { 0 }, { 0 }, { } };
 
 	newinfo = vmalloc(sizeof(struct ipt_table_info)
-			  + SMP_ALIGN(repl->size) * num_possible_cpus());
+			  + SMP_ALIGN(repl->size) * 
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 

net/ipv6/netfilter/ip6_tables.c

···
 #include <asm/uaccess.h>
 #include <asm/semaphore.h>
 #include <linux/proc_fs.h>
+#include <linux/cpumask.h>
 
 #include <linux/netfilter_ipv6/ip6_tables.h>
 
···
 	}
 
 	/* And one copy for every other CPU */
-	for (i = 1; i < num_possible_cpus(); i++) {
-		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size)*i,
+	for_each_cpu(i) {
+		if (i == 0)
+			continue;
+		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
 		       newinfo->entries,
 		       SMP_ALIGN(newinfo->size));
 	}
···
 		unsigned int i;
 
 		for (i = 0; i < num_possible_cpus(); i++) {
+		for_each_cpu(i) {
 			table_base =
 				(void *)newinfo->entries
 				+ TABLE_OFFSET(newinfo, i);
···
 	unsigned int cpu;
 	unsigned int i;
 
-	for (cpu = 0; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
 		i = 0;
 		IP6T_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
 				  t->size,
···
 		return -ENOMEM;
 
 	newinfo = vmalloc(sizeof(struct ip6t_table_info)
-			  + SMP_ALIGN(tmp.size) * num_possible_cpus());
+			  + SMP_ALIGN(tmp.size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
···
 		= { 0, 0, 0, { 0 }, { 0 }, { } };
 
 	newinfo = vmalloc(sizeof(struct ip6t_table_info)
-			  + SMP_ALIGN(repl->size) * num_possible_cpus());
+			  + SMP_ALIGN(repl->size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;