iio: hid-sensor-attributes: Fix divisions for 32-bit platforms
authorAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Thu, 5 Sep 2019 11:27:59 +0000 (14:27 +0300)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 6 Sep 2019 06:09:34 +0000 (08:09 +0200)
The commit 473d12f7638c

  ("iio: hid-sensor-attributes: Convert to use int_pow()")

converted to use generic int_pow() helper. Though, the generic one returns
64-bit value and, in cases when it is used as divisor, it compels 64-bit
division from compiler.

In order to fix this, introduce a temporary 32-bit variable to hold the result
of int_pow() and use it as divisor afterwards.

In couple of cases, replace int_pow() with a predefined unit factors for time
and frequency.

Fixes: 473d12f7638c ("iio: hid-sensor-attributes: Convert to use int_pow()")
Reported-by: kbuild test robot <lkp@intel.com>
Reported-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://lore.kernel.org/r/20190905112759.13035-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/iio/common/hid-sensors/hid-sensor-attributes.c

index b9dd19b34267d6667831c507b5bf7931ad45cab0..442ff787f7afab97e73ba94abad9464fc0b74c96 100644 (file)
 #include <linux/irq.h>
 #include <linux/kernel.h>
 #include <linux/slab.h>
+#include <linux/time.h>
+
 #include <linux/hid-sensor-hub.h>
 #include <linux/iio/iio.h>
 #include <linux/iio/sysfs.h>
 
+#define HZ_PER_MHZ     1000000L
+
 static struct {
        u32 usage_id;
        int unit; /* 0 for default others from HID sensor spec */
@@ -93,8 +97,10 @@ static void simple_div(int dividend, int divisor, int *whole,
 
 static void split_micro_fraction(unsigned int no, int exp, int *val1, int *val2)
 {
-       *val1 = no / int_pow(10, exp);
-       *val2 = no % int_pow(10, exp) * int_pow(10, 6 - exp);
+       int divisor = int_pow(10, exp);
+
+       *val1 = no / divisor;
+       *val2 = no % divisor * int_pow(10, 6 - exp);
 }
 
 /*
@@ -129,6 +135,7 @@ static void convert_from_vtf_format(u32 value, int size, int exp,
 
 static u32 convert_to_vtf_format(int size, int exp, int val1, int val2)
 {
+       int divisor;
        u32 value;
        int sign = 1;
 
@@ -136,10 +143,13 @@ static u32 convert_to_vtf_format(int size, int exp, int val1, int val2)
                sign = -1;
        exp = hid_sensor_convert_exponent(exp);
        if (exp < 0) {
+               divisor = int_pow(10, 6 + exp);
                value = abs(val1) * int_pow(10, -exp);
-               value += abs(val2) / int_pow(10, 6 + exp);
-       } else
-               value = abs(val1) / int_pow(10, exp);
+               value += abs(val2) / divisor;
+       } else {
+               divisor = int_pow(10, exp);
+               value = abs(val1) / divisor;
+       }
        if (sign < 0)
                value =  ((1LL << (size * 8)) - value);
 
@@ -202,12 +212,12 @@ int hid_sensor_write_samp_freq_value(struct hid_sensor_common *st,
        if (val1 < 0 || val2 < 0)
                return -EINVAL;
 
-       value = val1 * int_pow(10, 6) + val2;
+       value = val1 * HZ_PER_MHZ + val2;
        if (value) {
                if (st->poll.units == HID_USAGE_SENSOR_UNITS_MILLISECOND)
-                       value = int_pow(10, 9) / value;
+                       value = NSEC_PER_SEC / value;
                else if (st->poll.units == HID_USAGE_SENSOR_UNITS_SECOND)
-                       value = int_pow(10, 6) / value;
+                       value = USEC_PER_SEC / value;
                else
                        value = 0;
        }
@@ -296,6 +306,7 @@ EXPORT_SYMBOL(hid_sensor_write_raw_hyst_value);
 static void adjust_exponent_nano(int *val0, int *val1, int scale0,
                                  int scale1, int exp)
 {
+       int divisor;
        int i;
        int x;
        int res;
@@ -309,9 +320,10 @@ static void adjust_exponent_nano(int *val0, int *val1, int scale0,
                        return;
                }
                for (i = 0; i < exp; ++i) {
-                       x = scale1 / int_pow(10, 8 - i);
+                       divisor = int_pow(10, 8 - i);
+                       x = scale1 / divisor;
                        res += int_pow(10, exp - 1 - i) * x;
-                       scale1 = scale1 % int_pow(10, 8 - i);
+                       scale1 = scale1 % divisor;
                }
                *val0 += res;
                *val1 = scale1 * int_pow(10, exp);
@@ -321,13 +333,15 @@ static void adjust_exponent_nano(int *val0, int *val1, int scale0,
                        *val0 = *val1 = 0;
                        return;
                }
-               *val0 = scale0 / int_pow(10, exp);
-               rem = scale0 % int_pow(10, exp);
+               divisor = int_pow(10, exp);
+               *val0 = scale0 / divisor;
+               rem = scale0 % divisor;
                res = 0;
                for (i = 0; i < (9 - exp); ++i) {
-                       x = scale1 / int_pow(10, 8 - i);
+                       divisor = int_pow(10, 8 - i);
+                       x = scale1 / divisor;
                        res += int_pow(10, 8 - exp - i) * x;
-                       scale1 = scale1 % int_pow(10, 8 - i);
+                       scale1 = scale1 % divisor;
                }
                *val1 = rem * int_pow(10, 9 - exp) + res;
        } else {