math/float/thread_code.c, revert previous patch which added check for NaN. Here is a patch that reverts the previous patch applied to thread_cod.c. Some of the float tests were reporting a failure where it detected the difference between the result and expected result was NaN(not a number). As some of the results and expected results resolved to inf(infinity), the difference calculation would result in NaN. With the previous patch, float_power and float_bessel tests were consistently failing on all architectures tried. Signed-off by: Henry Yei <hyei@mvista.com>.
1 file changed