On Wed, Dec 11, 2002 at 09:06:35AM -0800, Jun Sun wrote: > Part of you patches fixes an off-by-one problem in flushing > cache for a range of addresses. I have a different version > for that part which is a little more visually pleasing. > > See attachment. In case you were wondering why the whole loop construction is looking so odd, take a look at the compiler output. The loop construction I've choosen is the only one that's producing bearable code. Sad to see that gcc is producing crap code from such a trivial loop ... Ralf
