mmc: atmel-mci: correct data timeout computation
authorLudovic Desroches <ludovic.desroches@atmel.com>
Wed, 28 Mar 2012 10:28:33 +0000 (12:28 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 13 Apr 2012 15:33:49 +0000 (08:33 -0700)
commit 66292ad92c6d3f2f1c137a1c826b331ca8595dfd upstream.

The HSMCI operates at a rate of up to Master Clock divided by two.
Moreover previous calculation can cause overflows and so wrong
timeouts.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/mmc/host/atmel-mci.c

index 9896933..0932024 100644 (file)
@@ -480,7 +480,14 @@ err:
 static inline unsigned int atmci_ns_to_clocks(struct atmel_mci *host,
                                        unsigned int ns)
 {
-       return (ns * (host->bus_hz / 1000000) + 999) / 1000;
+       /*
+        * It is easier here to use us instead of ns for the timeout,
+        * it prevents from overflows during calculation.
+        */
+       unsigned int us = DIV_ROUND_UP(ns, 1000);
+
+       /* Maximum clock frequency is host->bus_hz/2 */
+       return us * (DIV_ROUND_UP(host->bus_hz, 2000000));
 }
 
 static void atmci_set_timeout(struct atmel_mci *host,