ACPI / EC: Fix broken 64bit big-endian users of 'global_lock'
authorViresh Kumar <viresh.kumar@linaro.org>
Sat, 26 Sep 2015 22:04:06 +0000 (15:04 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 4 Oct 2015 10:36:07 +0000 (11:36 +0100)
global_lock is defined as an unsigned long and accessing only its lower
32 bits from sysfs is incorrect, as we need to consider other 32 bits
for big endian 64-bit systems. There are no such platforms yet, but the
code needs to be robust for such a case.

Fix that by changing type of 'global_lock' to u32.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/acpi/ec_sys.c
drivers/acpi/internal.h

index b4c216bab22b8a8836c2833e8f58eaf35d2f7311..bea8e425a8def6e81e4a7d506b2e38588f7af37b 100644 (file)
@@ -128,7 +128,7 @@ static int acpi_ec_add_debugfs(struct acpi_ec *ec, unsigned int ec_device_count)
        if (!debugfs_create_x32("gpe", 0444, dev_dir, (u32 *)&first_ec->gpe))
                goto error;
        if (!debugfs_create_bool("use_global_lock", 0444, dev_dir,
-                                (u32 *)&first_ec->global_lock))
+                                &first_ec->global_lock))
                goto error;
 
        if (write_support)
index 9e426210c2a8ea6a283bc04d9dc8bd1ac739370e..9db196de003ca575daa5430de8fe57236b35b775 100644 (file)
@@ -138,7 +138,7 @@ struct acpi_ec {
        unsigned long gpe;
        unsigned long command_addr;
        unsigned long data_addr;
-       unsigned long global_lock;
+       u32 global_lock;
        unsigned long flags;
        unsigned long reference_count;
        struct mutex mutex;