Message ID | 20240309-sunset-v2-0-f09912574d2c@ravnborg.org |
---|---|
Headers | show |
Series | sparc32: sunset sun4m and sun4d | expand |
On 2024-04-23 20:02, Sam Ravnborg wrote: > Please let me know if you expect me to rebase this on for-next. > I have not yet tried if there are merge conflicts but can take a look in > a some days if required. My local testing branch for this patchset rebased with trivial fixups, so no immediate rebase and resubmission is needed. I do run into some strange problems on SMP with this patchset plus your CAS patchset, that I do not get when I rebase the Linux patches from our kernel release that has my CAS patchset. With no CAS at all these things fails even worse, so do I need one or the other for these tests. I will need to dig deeper into figuring out the problems seems to be due to something in any of your two patchsets in themselves or if it is something else. I do need some additional fixes from our kernel release for SMP to work properly, so it could also be that there is something with my combination of your patches and my patches adapted on top of yours. There are also some needed fixes for LEON that relies upon code removed in this patchset. Maybe the best solution for that would be if I submit those and you then rebase upon them. > That is assuming you agree with the sunset of the sun platforms... I do agree with the idea in general, but being busy with other things I have had little time to dig into this lately. Thanks, Andreas
Hi Andreas. On Fri, Apr 26, 2024 at 06:31:12PM +0200, Andreas Larsson wrote: > On 2024-04-23 20:02, Sam Ravnborg wrote: > > Please let me know if you expect me to rebase this on for-next. > > I have not yet tried if there are merge conflicts but can take a look in > > a some days if required. > > My local testing branch for this patchset rebased with trivial fixups, > so no immediate rebase and resubmission is needed. I do run into some > strange problems on SMP with this patchset plus your CAS patchset, that > I do not get when I rebase the Linux patches from our kernel release > that has my CAS patchset. With no CAS at all these things fails even > worse, so do I need one or the other for these tests. :-( > > I will need to dig deeper into figuring out the problems seems to be due > to something in any of your two patchsets in themselves or if it is > something else. I do need some additional fixes from our kernel release > for SMP to work properly, so it could also be that there is something > with my combination of your patches and my patches adapted on top of > yours. > > There are also some needed fixes for LEON that relies upon code removed > in this patchset. Maybe the best solution for that would be if I submit > those and you then rebase upon them. If you can make LEON stable before starting the code removal that would be the best way forward. Then it should be easier to identify when the removal patches breaks things. With all the surgery I had to do to remove stuff it would have been a big surprise if it just worked - so it goes as expected. qemu only verify so much - the real target (and SMP) is way better. > > That is assuming you agree with the sunset of the sun platforms... > > I do agree with the idea in general, but being busy with other things I > have had little time to dig into this lately. If this round could be used to stabilize LEON and apply other more trivial stuff, then I am happy to rebase the "sunset" patchset sometimes after next -rc1. I have accumulated a bit more on top of what I already posted, and I will include a few more patches in the next round. In other words - take your time! Sam