[PKT_SCHED]: improve hashing performance of cls_fw

Calculate hashtable size to fit into a page instead of a hardcoded
256 buckets hash table. Results in a 1024 buckets hashtable on
most systems.

Replace old naive extract-8-lsb-bits algorithm with a better
algorithm xor'ing 3 or 4 bit fields at the size of the hashtable
array index in order to improve distribution if the majority of
the lower bits are unused while keeping zero collision behaviour
for the most common use case.

Thanks to Wang Jian <[e-mail]> for bringing this issue
to attention and to Eran Mann <[e-mail]> for the initial
idea for this new algorithm.

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

net/sched/cls_fw.c

···
 #include <net/act_api.h>
 #include <net/pkt_cls.h>
 
+#define HTSIZE (PAGE_SIZE/sizeof(struct fw_filter *))
+
 struct fw_head
 {
-	struct fw_filter *ht[256];
+	struct fw_filter *ht[HTSIZE];
 };
 
 struct fw_filter
···
 
 static __inline__ int fw_hash(u32 handle)
 {
-	return handle&0xFF;
+	if (HTSIZE == 4096)
+		return ((handle >> 24) & 0xFFF) ^
+		       ((handle >> 12) & 0xFFF) ^
+		       (handle & 0xFFF);
+	else if (HTSIZE == 2048)
+		return ((handle >> 22) & 0x7FF) ^
+		       ((handle >> 11) & 0x7FF) ^
+		       (handle & 0x7FF);
+	else if (HTSIZE == 1024)
+		return ((handle >> 20) & 0x3FF) ^
+		       ((handle >> 10) & 0x3FF) ^
+		       (handle & 0x3FF);
+	else if (HTSIZE == 512)
+		return (handle >> 27) ^
+		       ((handle >> 18) & 0x1FF) ^
+		       ((handle >> 9) & 0x1FF) ^
+		       (handle & 0x1FF);
+	else if (HTSIZE == 256) {
+		u8 *t = (u8 *) &handle;
+		return t[0] ^ t[1] ^ t[2] ^ t[3];
+	} else 
+		return handle & (HTSIZE - 1);
 }
 
 static int fw_classify(struct sk_buff *skb, struct tcf_proto *tp,
···
 	if (head == NULL)
 		return;
 
-	for (h=0; h<256; h++) {
+	for (h=0; h<HTSIZE; h++) {
 		while ((f=head->ht[h]) != NULL) {
 			head->ht[h] = f->next;
 			fw_delete_filter(tp, f);
···
 	if (arg->stop)
 		return;
 
-	for (h = 0; h < 256; h++) {
+	for (h = 0; h < HTSIZE; h++) {
 		struct fw_filter *f;
 
 		for (f = head->ht[h]; f; f = f->next) {