Message ID | 87k25vh23u.fsf@concordia.ellerman.id.au (mailing list archive) |
---|---|
State | Not Applicable |
Headers | show |
On Fri, May 5, 2017 at 6:20 AM, Michael Ellerman <mpe@ellerman.id.au> wrote: > > Also due to a screw-up on my part, we need a hunk added when you merge > the powerpc tree with the KVM tree: > > diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c > index 7807ee17af4b..ffe1da95033a 100644 > --- a/arch/powerpc/kvm/book3s_xive.c > +++ b/arch/powerpc/kvm/book3s_xive.c > @@ -24,6 +24,7 @@ > #include <asm/xive.h> > #include <asm/xive-regs.h> > #include <asm/debug.h> > +#include <asm/debugfs.h> > #include <asm/time.h> > #include <asm/opal.h> > > I haven't seen a pull request from Paolo yet, so possibly you'll need > to add that later when you merge the KVM tree. Ok. I won't catch this in my allmodconfig build due to the ppc-specific nature of it, so I'll probably forget about this and not catch it. Paolo, Radim, can you try to remind me about this conflict when you send the KVM pull request? Also: > - Larger virtual address space on 64-bit server CPUs. By default we > use a 128TB virtual address space, but a process can request access > to the full 512TB by passing a hint to mmap(). I'm assuming/hoping this uses the same logic as the upcoming 47+ bit 5-level page table x86 support? Sounds like it, but I didn't check the code. See "x86/mm: Allow to have userspace mappings above 47-bits" which hasn't been merged yet, but has been discussed on linux-mm, for example. Thanks, Linus
Linus Torvalds <torvalds@linux-foundation.org> writes: > On Fri, May 5, 2017 at 6:20 AM, Michael Ellerman <mpe@ellerman.id.au> wrote: >> >> Also due to a screw-up on my part, we need a hunk added when you merge >> the powerpc tree with the KVM tree: >> >> diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c >> index 7807ee17af4b..ffe1da95033a 100644 >> --- a/arch/powerpc/kvm/book3s_xive.c >> +++ b/arch/powerpc/kvm/book3s_xive.c >> @@ -24,6 +24,7 @@ >> #include <asm/xive.h> >> #include <asm/xive-regs.h> >> #include <asm/debug.h> >> +#include <asm/debugfs.h> >> #include <asm/time.h> >> #include <asm/opal.h> >> >> I haven't seen a pull request from Paolo yet, so possibly you'll need >> to add that later when you merge the KVM tree. > > Ok. I won't catch this in my allmodconfig build due to the > ppc-specific nature of it, so I'll probably forget about this and not > catch it. Yeah fair enough. > Paolo, Radim, can you try to remind me about this conflict when you > send the KVM pull request? I Cc'ed Paolo, and will try and keep an eye out for the KVM pull request in case he forgets to mention it. > Also: > >> - Larger virtual address space on 64-bit server CPUs. By default we >> use a 128TB virtual address space, but a process can request access >> to the full 512TB by passing a hint to mmap(). > > I'm assuming/hoping this uses the same logic as the upcoming 47+ bit > 5-level page table x86 support? Sounds like it, but I didn't check the > code. > > See "x86/mm: Allow to have userspace mappings above 47-bits" which > hasn't been merged yet, but has been discussed on linux-mm, for > example. Yep. We were tracking that discussion, and made changes to make our semantics match x86. If anything changes on the x86 side before it merges we can update our code to match, hopefully prior to 4.12 releasing. There's also a selftest for the mmap behaviour, written by one of our guys, which is coming via mm I think, so that should help make sure the semantics match. cheers
diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c index 7807ee17af4b..ffe1da95033a 100644 --- a/arch/powerpc/kvm/book3s_xive.c +++ b/arch/powerpc/kvm/book3s_xive.c @@ -24,6 +24,7 @@ #include <asm/xive.h> #include <asm/xive-regs.h> #include <asm/debug.h> +#include <asm/debugfs.h> #include <asm/time.h> #include <asm/opal.h>