Does a lot of work towards the PIT

Added a math library, with clamp and POW.
added printf support for printing doubles
added a few helper functions in PIT for calcaulting the time in seconds of a clock cycle based on divisor

kernel/arch/io/pit.c

@@ -1,3 +1,64 @@
+#include <math.h>
+
 #include <kernel/x86/pit.h>
+#include <kernel/x86/io.h>
+
+struct pit_state pit;
+
+uint32_t normalize_divisor(uint32_t divisor)
+{
+	if (divisor == 0)
+		return 65536;
+	return clamp_u32(divisor, 0, 65535);
+}
+
+double get_time_from_divisor(uint32_t divisor)
+{
+	uint32_t _divisor = normalize_divisor(divisor);
+		
+	double clock_rate = 3579545.0 / 3.0; // This is for a bit more precision in the hz rate
+	
+	return (_divisor / (clock_rate)) * 1000;
+}
+
+/**
+ * Im worried about an overflow
+ *
+ *
+ * What i want to do is
+ *
+ * Every time we get the Timer intterupt, i want to increment a variable, just the amount of times its fired
+ * and then we can use the time rate above to get total time elapsed, which we can use for sleep and stuff
+ *
+ * Im worried about overflows, but i can add something that will automatically flip it over? or do uints already roll over? they already rollover!!! WOOOO!!! 
+ * just need to be aware that only a maximum of 52ms * 32int max (64 int max) can be tracked before it rolls over!
+ */
+
+
+uint16_t read_pit_count(uint8_t channel) // Only in lobyte/hibyte mode
+
+{
+	uint16_t count = 0;
+
+	__asm__ volatile ("cli");
+
+
+	outb(PIT_CMD_REG, (channel << 6)); // xx000000 where x is the channel
+
+	count = inb(channel);
+	count |= inb(channel) << 8;
+
+	return count; // TODO make sure to enable interrupts after this
+}
+
+void set_pit_count(uint8_t channel, uint16_t count) // Only in lobyte/hibyte mode
+{
+	__asm__ volatile ("cli");
+
+	outb(channel, count & 0xFF); // low byte
+	outb(channel, (count & 0xFF00) >> 8); // high byte
+	
+	// TODO: make sure to set interrupts
+}