[LTP] Case settimeofday01 dose not handle TIME_INTERPOLATION properly.

[Pu, L. <lo...@in...>]

	In ltp-20041007, settimeofday01 fails on some platforms(ia64,
x86-64) occasionally.

Here is the cause:
	The testcase did not consider the situation when
CONFIG_TIME_INTERPOLATION is enabled.
	When CONFIG_TIME_INTERPOLATION is enabled, kernel dose not
update the xtime to what you specify exactly. Instead, It subtracts
time_interpolator_get_offset() nanoseconds from the time you specify and
then updates xtime.
	An immediate gettimeofday() after a settimeofday() will return
you an earlier time than you specify via settimeofday() in most cases if
CONFIG_TIME_INTERPOLATION is enable.
	The test case just uses settimeofday() to set time to a certain
timeval and then uses gettimeofday() to get back a timeval structure
immediately to see if two tv_sec in timeval structures are equal.

	I attached my patch.
 Pu Long