On Mon, Jul 5, 2021 at 8:01 AM Christoph Hellwig [off-list ref] wrote:

Just a ping if we're making some progress on this.  The m68knommu
set_fs implementation is actively misleading, which lead to a problem
this merge window so I'd really like to see this set merged.

Also your improvements to the copy_from_kernel_nofaul loop would be

I've rebased the series now, and rearranged a little in order to address
Russell's concerns. I still have one regression that I need to fix before
sending the new version.

        Arnd