Re: [linus:master] [mseal] 8be7258aad: stress-ng.pagemove.page_remaps_per_sec -4.4% regression
From: Nicholas Piggin
Date: Mon Aug 05 2024 - 19:25:23 EST
On Tue Aug 6, 2024 at 5:48 AM AEST, Linus Torvalds wrote:
> On Mon, 5 Aug 2024 at 12:38, Jeff Xu <jeffxu@xxxxxxxxxx> wrote:
> >
> > I'm curious, why does ppc need to unmap vdso ? ( other archs don't
> > have unmap logic.)
>
> I have no idea. There are comments about 'perf' getting confused about
> mmap counts when 'context.vdso' isn't set up.
>
> But x86 has the same context.vdso logic, and does *not* set the
> pointer before installing the vma, for example. Also does not zero it
> out on munmap(), although it does have the mremap logic.
>
> For all I know it may all be entirely unnecessary, and could be
> removed entirely.
I don't know much about vdso code, it predated my involvedment in ppc.
Commit 83d3f0e90c6c8 says CRIU (checkpoint restore in userspace) is
moving it around. Why CRIU wants to do that, I don't know.
Can userspace on other archs not unmap their vdsos?
Thanks,
Nick