Hi Christophe,

On 11/24/20 6:53 AM, Christophe Leroy wrote:

Le 24/11/2020 à 01:29, Dmitry Safonov a écrit :

quoted

v2 Changes:
- Rename user_landing to vdso_base as it tracks vDSO VMA start address,
   rather than the explicit address to land (Andy)
- Reword and don't use "new-execed" and "new-born" task (Andy)
- Fix failures reported by build robot

Started from discussion [1], where was noted that currently a couple of
architectures support mremap() for vdso/sigpage, but not munmap().
If an application maps something on the ex-place of vdso/sigpage,
later after processing signal it will land there (good luck!)

Patches set is based on linux-next (next-20201123) and it depends on
changes in x86/cleanups (those reclaim TIF_IA32/TIF_X32) and also
on my changes in akpm (fixing several mremap() issues).

I have a series that cleans up VDSO init on powerpc and migrates powerpc
to _install_special_mapping() (patch 10 of the series).

https://patchwork.ozlabs.org/project/linuxppc-dev/list/?series=204396&state=%2A&archive=both


I'm wondering how we should coordinate with your series for merging.

I guess your series will also imply removal of arch_unmap() ? see
https://elixir.bootlin.com/linux/v5.10-rc4/source/arch/powerpc/include/asm/mmu_context.h#L262

I think our series intersect only in that moment where I re-introduce
arch_setup_additional_pages() parameters. So, in theory we could
minimize the conflicts by merging both series in parallel and cleanup
the result by moving to generic vdso_base on the top, what do you think?

Thanks,
          Dmitry