Revert "PM / Sleep: Require CAP_BLOCK_SUSPEND to use wake_lock/wake_unlock"
authorAjay Nandakumar <anandakumarm@nvidia.com>
Thu, 25 Apr 2013 08:42:29 +0000 (14:12 +0530)
committerJohn Stultz <john.stultz@linaro.org>
Tue, 14 Jan 2014 00:36:47 +0000 (16:36 -0800)
This reverts commit 11388c87d2abca1f01975ced28ce9eacea239104.

The issue is that no wake lock is held at the user space i.e by Power
Manager service.This is because the PowerManagerService fails to
acquire the Wakelock.In 3.8 the wakelock module in the kernel expects
the user process to have the capability of CAP_BLOCK_SUSPEND.Which the
powermangersevice does not have.

Bug 1274297
Bug 1384311

Change-Id: I3b696108d47278cf40abce8d5a9bd012f98f2925
Signed-off-by: Ajay Nandakumar <anandakumarm@nvidia.com>
(cherry picked from commit e8464e785027a15279a13e6e32cd1aecd22d5a00)
Reviewed-on: http://git-master/r/282698
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
Tested-by: Bharat Nihalani <bnihalani@nvidia.com>
kernel/power/wakelock.c

index 8f50de394d22b30090d1cf62cd3fd3069d451d57..c8fba3380076afd8db340855340e5886b8497dc6 100644 (file)
@@ -9,7 +9,6 @@
  * manipulate wakelocks on Android.
  */
 
-#include <linux/capability.h>
 #include <linux/ctype.h>
 #include <linux/device.h>
 #include <linux/err.h>
@@ -189,9 +188,6 @@ int pm_wake_lock(const char *buf)
        size_t len;
        int ret = 0;
 
-       if (!capable(CAP_BLOCK_SUSPEND))
-               return -EPERM;
-
        while (*str && !isspace(*str))
                str++;
 
@@ -235,9 +231,6 @@ int pm_wake_unlock(const char *buf)
        size_t len;
        int ret = 0;
 
-       if (!capable(CAP_BLOCK_SUSPEND))
-               return -EPERM;
-
        len = strlen(buf);
        if (!len)
                return -EINVAL;