I think the results from the touch_atime() configure test are actually
inverted:
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_TOUCH_ATIME, 1, Define if kernel has touch_atime routine),
AC_MSG_RESULT(no),
should be:
AC_MSG_RESULT(no),
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_TOUCH_ATIME, 1, Define if kernel has touch_atime routine),
Apparently some 2.6 kernels have both touch_atime and update_atime,
because it didn't cause a problem on the first system I built on.
However, it doesn't work on 2.4 kernels.
The above switch makes this test look more like the test for
HAVE_GENERIC_GETXATTR, where the test failure actually indicates that
the symbol is present, because the definition in the AC_TRY_COMPILE
collides with the kernel symbol name.
-Phil
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
