From: Benjamin Herrenschmidt <benh@kernel.crashing.org> Date: 2006-11-21 07:29:17
We read your IO workaroud patch for spider. The method for disabling prefetch
is same as we suppose. And the method for dummy read may be available,
or be a good reference for us.
We found a mistake in our patch at disabling prefetch. We set the prefetch
enable bit wrongly. We'll correct it in the next patch.