On Wed, Oct 7, 2020 at 10:59 AM Christoph Hellwig <hch@xxxxxx> wrote: > On Wed, Oct 07, 2020 at 10:54:19AM +0200, Arnd Bergmann wrote: > > On Wed, Oct 7, 2020 at 9:44 AM Christoph Hellwig <hch@xxxxxx> wrote: > > > > > > There are no actual implementations of FBIOSCURSOR/FBIOGCURSOR left, so > > > don't bother with a compat handler for it, and remove the remaining > > > definitions as well. > > > > > > Signed-off-by: Christoph Hellwig <hch@xxxxxx> > > > > I had submitted a similar patch earlier, and Sam Ravnborg applied it to the > > drm-misc tree, but it doesn't seem to be in linux-next, so I don't know > > what the state is. > > > > My version only removed the compat handling, not the data structures, > > so I'm happy to see your version used instead if mine got lost. > > Oh, sorry. I thought in your summary you decided to give up on > the sbuslib ones. Here is what I have pending at the moment for set_fs() and compat_alloc_user_space(): https://git.kernel.org/pub/scm/linux/kernel/git/arnd/playground.git/log/?h=compat-alloc-user-space-2 The only one I have actually given up on is the atomisp staging driver, which uses an awful hack, copying the x86 implementation of copy_in_user()/compat_alloc_user_space() into the driver. This is based on last week's linux-next, as I plan to resubmit after the merge window. Arnd
