Avoid a division in {u,n}delay

arch/riscv/lib/delay.c

@@ -16,6 +16,62 @@
 #include <linux/timex.h>
 #include <linux/export.h>
 
+/*
+ * This is copies from arch/arm/include/asm/delay.h
+ *
+ * Loop (or tick) based delay:
+ *
+ * loops = loops_per_jiffy * jiffies_per_sec * delay_us / us_per_sec
+ *
+ * where:
+ *
+ * jiffies_per_sec = HZ
+ * us_per_sec = 1000000
+ *
+ * Therefore the constant part is HZ / 1000000 which is a small
+ * fractional number. To make this usable with integer math, we
+ * scale up this constant by 2^31, perform the actual multiplication,
+ * and scale the result back down by 2^31 with a simple shift:
+ *
+ * loops = (loops_per_jiffy * delay_us * UDELAY_MULT) >> 31
+ *
+ * where:
+ *
+ * UDELAY_MULT = 2^31 * HZ / 1000000
+ *             = (2^31 / 1000000) * HZ
+ *             = 2147.483648 * HZ
+ *             = 2147 * HZ + 483648 * HZ / 1000000
+ *
+ * 31 is the biggest scale shift value that won't overflow 32 bits for
+ * delay_us * UDELAY_MULT assuming HZ <= 1000 and delay_us <= 2000.
+ */
+#define MAX_UDELAY_US	2000
+#define MAX_UDELAY_HZ	1000
+#define UDELAY_MULT	(2147UL * HZ + 483648UL * HZ / 1000000UL)
+#define UDELAY_SHIFT	31
+
+#if HZ > MAX_UDELAY_HZ
+#error "HZ > MAX_UDELAY_HZ"
+#endif
+
+/* RISC-V supports both UDELAY and NDELAY.  This is largely the same as above,
+ * but with different constants.  I kept the number of milliseconds the same
+ * between the two calls.
+ *
+ * NDELAY_MULT = 2^27 * HZ / 1000000000
+ *             = (2^27 / 1000000000) * HZ
+ *             = 1.34217728 * HZ
+ *             = 1 * HZ + 342177280 * HZ / 1000000000
+ */
+#define MAX_NDELAY_NS   2000000UL
+#define MAX_NDELAY_HZ	MAX_UDELAY_HZ
+#define NDELAY_MULT	((unsigned long)(1ULL * HZ + 342177280ULL * HZ / 1000000000ULL))
+#define NDELAY_SHIFT	27
+
+#if HZ > MAX_NDELAY_HZ
+#error "HZ > MAX_NDELAY_HZ"
+#endif
+
 void __delay(unsigned long cycles)
 {
 	u64 t0 = get_cycles();
@@ -26,18 +82,26 @@ void __delay(unsigned long cycles)
 
 void udelay(unsigned long usecs)
 {
-	u64 ucycles = (u64)usecs * riscv_timebase;
+	unsigned long ucycles = usecs * lpj_fine * UDELAY_MULT;
 
-	do_div(ucycles, 1000000U);
-	__delay((unsigned long)ucycles);
+	if (usecs > MAX_UDELAY_US) {
+		__delay((u64)usecs * riscv_timebase / 1000000ULL);
+		return;
+	}
+
+	__delay(ucycles >> UDELAY_SHIFT);
 }
 EXPORT_SYMBOL(udelay);
 
 void ndelay(unsigned long nsecs)
 {
-	u64 ncycles = (u64)nsecs * riscv_timebase;
+	unsigned long ncycles = nsecs * lpj_fine * NDELAY_MULT;
+
+	if (nsecs > MAX_NDELAY_NS) {
+		__delay((u64)nsecs * riscv_timebase / 1000000000ULL);
+		return;
+	}
 
-	do_div(ncycles, 1000000000U);
-	__delay((unsigned long)ncycles);
+	__delay(ncycles >> NDELAY_SHIFT);
 }
 EXPORT_SYMBOL(ndelay);