W1: w1_io.c reset comments and msleep
David Fries [Thu, 16 Oct 2008 05:05:01 +0000 (22:05 -0700)]
w1_reset_bus, added some comments about the timing and switched to msleep
for the later delay.  I don't have the hardware to test the sleep after
reset change.  The one wire doesn't have a timing requirement between
commands so it is fine.  I do have the USB hardware and it would be in big
trouble with 10ms interrupt transfers to find that the reset completed.

Signed-off-by: David Fries <david@fries.net>
Signed-off-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/w1/w1_io.c

index 97b338a..f4f82f1 100644 (file)
@@ -293,12 +293,24 @@ int w1_reset_bus(struct w1_master *dev)
                result = dev->bus_master->reset_bus(dev->bus_master->data) & 0x1;
        else {
                dev->bus_master->write_bit(dev->bus_master->data, 0);
+               /* minimum 480, max ? us
+                * be nice and sleep, except 18b20 spec lists 960us maximum,
+                * so until we can sleep with microsecond accuracy, spin.
+                * Feel free to come up with some other way to give up the
+                * cpu for such a short amount of time AND get it back in
+                * the maximum amount of time.
+                */
                w1_delay(480);
                dev->bus_master->write_bit(dev->bus_master->data, 1);
                w1_delay(70);
 
                result = dev->bus_master->read_bit(dev->bus_master->data) & 0x1;
-               w1_delay(410);
+               /* minmum 70 (above) + 410 = 480 us
+                * There aren't any timing requirements between a reset and
+                * the following transactions.  Sleeping is safe here.
+                */
+               /* w1_delay(410); min required time */
+               msleep(1);
        }
 
        return result;