Messages in this thread | | | From | Arnd Bergmann <> | Date | Mon, 12 Mar 2018 21:21:25 +0100 | Subject | Re: [GIT PULL] CRIS: Drop support for the CRIS-port |
| |
On Sun, Mar 11, 2018 at 12:05 PM, Jesper Nilsson <jesper.nilsson@axis.com> wrote: > Hi Arnd, > > As promised, pull the below tag for the removal of the CRIS-port. > There are still a few references to the CRIS port left, and they > can be grouped into two categories: > > - Examples or references to the CRIS port as provenance. > These I believe can be left as is. > > - The workarounds for CRIS not having "unpacked" structs. > When I looked at it 10 years ago, CRIS was the only port > that needed explicit alignment of some structs, > due to pointers being valid at any byte offset in CRIS, > which would wreak havoc with code that assumed that the > lower bits of a pointer are free to use as flags. > I'm leaving these in until we can say that no newer > port has the same problem. > > Best regards,
Hi Jesper,
Thanks for the patch! I had done a slightly different approach, but am merging your changes into my tree now. I already have separate patches for each driver subsystem, so I'll keep my versions, having already made sure that your changes were identical.
For Documentation/features and the Kconfig files that have lists of architectures, I also have separate patches that do the changes for all architectures, to avoid touching those files (up to) eight times for one trivial change each. I could redo those, but since your patch conflicts with David Howell's patch for mn10300, it's easier to keep what I already have.
This leaves basically the removal of arch/cris along with the MAINTAINER file update. I'm keeping the patch I did, but change the author and changelog text to what you sent, with my Signed-off-by and a list of my modifications) added below.
Hope that works for you.
Arnd
| |