On 02/23/2015 08:24 PM, Maciej W. Rozycki wrote: > On Mon, 23 Feb 2015, Leonid Yegoshin wrote: > >> The same is basically for transfer D$ --> I$ because in MIPS it is done via L2 >> or memory. > > The original issue aside (I don't want to dive into it) this I believe is > left to an implementer's discretion and there are MIPS implementations > indeed that fill I$ directly from D$; IIRC Alchemy silicon and its > descendants. > > Maciej > I am tracking down even more HIGHMEM bugs, so I will rework these patches, clean-up commit messages and post again. Thanks for all the feedback. Steve
